United States     Office of Air Quality       EPA-450/3-79-024
          Environmental Protection  Planning and Standards     April 1979
          Agency       Research Triangle Park NC 27711
          __
vvEPA     Guidance for Lowest
          Achievable Emission
          Rates from 18 Major
          Stationary Sources
          of Particulate, Nitrogen
          Oxides, Sulfur Dioxide,
          or Volatile Organic
          Compounds

-------
                                                                    JAN
                                     ERRATA for

                                 EPA-450/3-79-024

             "Guidance for Lowest Achievable Emission Rates from
             18 Major Stationary Sources of Particulate, Nitrogen
             Oxides, Sulfur Dioxide, or Volatile Organic Compounds"


1.  On page 3.1-7, Table 3.1-1, the uncontrolled emission factors for sulfur
    dioxide and nitric oxides under g/kWh and lb/103 hp-h should be changed:

    from:

                  g/kWh           lb/103 hp-h

    S02         5.35-7.75          8.75-13.5

    N0y          7.3-4.0             12-66
      A
    to:
                                  lb/103 hp-h
    S02      5.35(S)-7.75(S)    8.75(S)-13.5(S)

    N0y         0.73-4.0            1.2-6.6
      A
2.  On page 3.2-7, Table 3.2-1, under "Ratio of flue gas rate to pulp production,"
    the symbols for metric and English units should be changed:

    from:

    m3, dry/Mg ADP/day |  (scfm/ton ADP)
    dscm/min/Mg ADP/day |  (scfm/ton ADP/day)



3.  On page 3.2-15, line 10 should be changed:

    from:

    available.   Table 3.2-3 reflects the better AIP levels  that have

    to:

    available.   Table 3.2-2 reflects the better AIP levels  that have

-------
4.  On page 3.4-3, line 7 should be changed:
    from:
    lyst to burn the coke, which reheats the catalyst.  The resulting
    tp_:
    lyst to burn the coke, which regenerates the catalyst.  The resulting
5.  On page 3.4-13, Reference 2 should be changed:
    from:
    2.  Data compiled from files of Greene & Associates, Inc.
    to:
    2.  Data supplied by Greene & Associates, Dallas, Texas.

6.  On page 3.9-10, Table 3.9-3, columns 4 through 7 should be changed
    from:
     to:
Coverage ,
liters/
103m2
119
21
56
7
49
(qal/103ft )
(1.7)
(0-3)
(0.8)
(0.1)
(0.7)
Range of VOC emissions
o
g/m coated
31 - 62
3.6 - 11.3
25 - 56
2.5 - 5.9
26 - 38.7
Total 88 - 174
lb/103ft2
(3.7 - 7.5)
(0.4 - 1.4)
(3.0 - 6.9)
(0.3 - 0.7)
(3.1 - 4.6)
(11 - 21)
Coverage ,
liters/
103m2
69
12.2
32.6
4
28.5
(gal/103ft?)
(1.7)
(0.3)
(0.8)
(0.1)
(0.7)
Range of VOC emissions
g/m coated
18 - 37
2.1 - 6.6
14 - 33
1.5 - 3.4
15 - 23
lb/103ft2
(3.7 - 7.5)
(0.4 - 1.4)
(3.0 - 6.9)
(0.3 - 0.7)
(3.1 - 4.6)
                     Total
51 - 103    (11 - 21)

-------
                                         3
 7.  On page 3.9-11, line 2 should be changed:
     from:
     range from 88 to 174 g/km2 (11 to 21 lb/1000 ft2) of flat wood coated.
     to:
     range from 51 to 103 g/m3 (11 to 21 lb/1000 ft2)  of flat wood coated.
 8.  On page 3.9-14, in Table 3.9-5, the words "Print ink" in the first
     column should be changed to "Grain ink."
 9.  On page 3.9-16, the last sentence in Section 3.9.3.5 should be changed:
     from:
     The major deterrent to its use is the highest of both
     the EB system and the costing materials.
     to:
     The major deterrent to its use is the high cost of both the EB system and  the
     coatinq materials.

10.  On page 3.9-20, Reference 16 should be added:
     16.  Personal Communication with M. Kay,  Southcoast Air Quality Management
          District, El Monte, California.
11.   On page 3.12-3,  in  Table 3.12-1,  the second  line  under  note  d  should  be
     changed:
     from:
     light  liquid means  a liquid  lighter  than  kerosene;  liquid means  a  liquid equal to
     to:
     light  liquid means  a liquid  lighter  than  kerosene;  heavy liquid  means a liquid
     equal  to

-------
12.  On page 3.13-3, in the first sentence of Section 3.13.3, the words
     "as summarized in Table 3.13-2" should be deleted.
13.  On page 3.15-6, line 15 of the first paragraph of Section 3.15.3 should
     be changed:

     from:

     have demonstrated 93 percent solvent reduction.*  A more complete descrip-


     tor

     have demonstrated 93 percent solvent reduction.    A more complete descrip-



14.  On page 3.15-6, the footnote at the bottom of the page should be deleted.



15.  On page 3.15-8, line 2 should be changed:

     from:

     produced an 87 percent reduction at some large appliance plants.*  Temper-


     to:
                                                                     o
     produced an 87 percent reduction at some large appliance plants.   Temper-



16.  On page 3.15-8, line 14 should be changed:

     from:

     into the bed.*^  For metal  furniture or large  appliance coating lines  the


     to:
                  g
     into the bed.    For metal  furniture or large  appliance coating lines  the
17.   On page 3.15-8,  the two footnotes  at  the  bottom  of  the  page  should  be
     deleted.

-------
                                         5

18.  On page 3.15-9, line 9 should be changed:

     from:

     powder coating by electrostatic spray methods is 99 percent.*  Several limi-


     to:
                                                                 g
     powder coating by electrostatic spray methods is 99 percent.   Several limi-



19.  On page 3.15-9, line 20 should be changed:

     from:

     The average achievable reduction is 78 percent.*  Higher-solids coatings are


     to:
                                                    Q
     The average achievable reduction is 78 percent.   Higher-solids coatings are



20.  On page 3.15-9, the footnote at the bottom of the page should be deleted.



21.  On page 3.15-12, the following references should be added:

     8.  Information supplied by Larry F. Nonemaker, E. I.  Dupont Demours & Co.,  Inc.,
     Winnewood, Pa., April 14, 1978.

     9.  Letter from W. C. Moses, Technical Manager, Chemical  Plant Division,
     Suttcliffe, Speakman & Company, Limited,  March 10, 1978.
22.  On page 3.2-7, in the fourth line under "explanation of abbreviations,"
     "m3, dry/Mg ADP/day" should be changed to "dscm/min/Mg ADP/day."

-------

-------
                                      EPA-450/3-79-024
Guidance for  Lowest Achievable Emission
 Rates from 18 Major  Stationary Sources
  of Particulate, Nitrogen Oxides,  Sulfur
 Dioxide,  or Volatile Organic Compounds
                          by
                    PEDCo Environmental, Inc.
                       Chester Towers
                      11499 Chester Road
                     Cincinnati, Ohio 45246
           EPA Project Officers: John H. Haines and Gary D. McCutchen
                        Prepared for

               U.S. ENVIRONMENTAL PROTECTION AGENCY
                  Office of Air, Noise, and Radiation
               Office of Air Quality Planning and Standards
               Research Triangle Park, North Carolina 27711          _  ,,0-,,^.

                        APT.M979     V) S. &*•*"«**£$"     'HCy
                        Mprn ia/a     u.<>; w c , ;^r-,rv ^rL-J.^-J'  .„.. ciQQr

-------
This report is issued by  the Environmental  Protection Agency to report
technical data of interest to a  limited number of readers.  Copies are
available free of charge to  Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from the
Library Services Office (MD-35),  U.S. Environmental Protection Agency,
Research Triangle Park,  North Carolina 27711; or, for a fee,  from the
National Technical Information  Service, 5285 Port Royal Road, Springfield,
Virginia 22161.
This report was furnished to the Environmental Protection Agency
by  PEDCo  Environmental, Inc., Chester  Towers, 11499 Chester Road,
Cincinnati, Ohio 45246, in fulfillment of Contract No.  68-0.1-4147.
Mention of company or product names is  not to be  considered  as an
endorsement  by the Environmental Protection Agency.
                   Publication No. EPA-450/3-79-024
               n S Environmental  Protection
               WjW-
                              ii

-------
                                     ERRATA for

                                 EPA-450/3-79-024

             "Guidance for Lowest Achievable Emission Rates from
             18 Major Stationary Sources of Particulate, Nitrogen
             Oxides, Sulfur Dioxide, or Volatile Organic Compounds"


1.  On page 3.1-7, Table 3.1-1, the uncontrolled emission factors for sulfur
    dioxide and nitric oxides under g/kWh and lb/103 hp-h should be changed:
    from:
                  g/kWh           lb/103 hp-h
    S02         5.35-7.75          8.75-13.5


    N0v          7.3-4.0             12-66
      A
    to:
                  g/kWh           lb/103 hp-h
    S02      5.35(S)-7.75(S)    8.75(S)-13.5(S)

    N0v         0.73-4.0            1.2-6.6
      X
2.  On page 3.2-7, Table 3.2-1, under "Ratio of flue gas rate to pulp production,"
    the symbols for metric and English units should be changed:

    from:

    m3, dry/Mg ADP/day |  (scfm/ton ADP)

    _to:

    dscm/min/Mg ADP/day |  (scfm/ton ADP/day)
3.  On page 3.2-15, line 10 should be changed:

    from:

    available.  Table 3.2-3 reflects the better AIP levels that have

    to:

    available.  Table 3.2-2 reflects the better AIP levels that have

-------
4.  On page 3.4-3, line 7 should be changed:
    from:
    lyst to burn the coke, which reheats the catalyst.  The resulting
    to:
    lyst to burn the coke, which regenerates the catalyst.  The resulting

5.  On page 3.4-13, Reference 2 should be changed:
    from:
    2.  Data compiled from files of Greene & Associates,  Inc.
    to:
    2.  Data supplied by Greene & Associates, Dallas, Texas.
6.  On page 3.9-10, Table 3.9-3, columns 4 through 7 should be changed:
    from:
     to:
Coverage ,
liters/
103m2
119
21
56
7
49
(qal/103ft )
(1.7)
(0.3)
(0.8)
(0.1)
(0.7)
Range of VOC emissions
p
g/m coated
31 - 62
3.6 - 11.3
25 - 56
2.5 - 5.9
26 - 38.7
Total 88 - 174
lb/103ft2
(3.7 - 7.5)
(0.4 - 1.4)
(3.0 - 6.9)
(0.3 - 0.7)
(3.1 - 4.6)
(11 - 21)
Coverage ,
liters/
103m2
69
12.2
32.6
4
28.5
(gal/103f£)
(1.7)
(0.3)
(0.8)
(0.1)
(0.7)
Range of VOC emissions
0
g/m coated
18 - 37
2.1 - 6.6
14 - 33
1.5 - 3.4
15 - 23
lb/103ft2
(3.7 - 7.5)
(0.4 - 1.4)
(3.0 - 6.9)
(0.3 - 0.7)
(3.1 - 4.6)
                     Total
51 - 103    (11 - 21)

-------
                                         3
 7.   On page 3.9-11,  line 2 should be changed:
     from:
     range  from 88 to 174 g/km2 (11 to 21 lb/1000 ft2)  of flat  wood  coated.
     to:
     range  from 51 to 103 g/m3 (11 to 21 lb/1000 ft2)  of flat wood coated.
 8.   On page 3.9-14,  in  Table 3.9-5,  the words  "Print  ink"  in  the  first
     column should be changed to "Grain  ink."
 9.   On page 3.9-16,  the last sentence in  Section  3.9.3.5  should  be  changed:
     from:
     The major deterrent to its  use is the highest of  both
     the EB system and the costing materials.
     to:
     The major deterrent to its  use is the high  cost of both  the  EB  system  and  the
     coating materials.

10.   On page 3.9-20,  Reference 16 should be added:
     16.  Personal Communication with  M. Kay,  Southcoast Air  Quality Management
          District, El Monte, California.
11.   On page 3.12-3,  in Table 3.12-1,  the second  line  under  note  d  should  be
     changed:
     from:
     light  liquid means a liquid lighter than  kerosene;  liquid  means  a  liquid  equal  to
     to:
     light  liquid means a liquid lighter than  kerosene;  heavy liquid  means a liquid
     equal  to

-------
12.  On page 3.13-3,  in the first sentence of Section  3.13.3,  the words
     "as summarized in Table 3.13-2"  should be deleted.
13.  On page 3.15-6,  line 15 of the first paragraph of Section 3.15.3 should
     be changed:


     from:

     have demonstrated 93 percent solvent reduction.*  A more complete descrip-


     to:
                                                    g
     have demonstrated 93 percent solvent reduction.   A more complete descrip-
14.  On page 3.15-6, the footnote at the bottom of the page should be deleted.




15v  On page 3.15-8, line 2 should be changed:


     from:

     produced an 87 percent reduction at some large appliance plants.*  Temper-



     to.:
                                                                     o
     produced an 87 percent reduction at some large appliance plants.   Temper-




16.  On page 3.15-8, line 14 should be changed:


     from:

     into the bed."f* For metal furniture or large appliance coating lines the


     to:

     into the bed.9  For metal furniture or large appliance coating lines the
17.  On page 3.15-8, the two footnotes at the bottom of the page should be
     deleted.

-------
                                         5


18.  On page 3.15-9, line 9 should be changed:


     from:


     powder coating by electrostatic spray methods is 99 percent.*  Several  limi-


     to:

                                                                 o
     powder coating by electrostatic spray methods is 99 percent.    Several  limi-




19.  On page 3.15-9, line 20 should be changed:


     from:


     The average achievable reduction is 78 percent.*  Higher-solids coatings are


     to:
                                                    p
     The average achievable reduction is 78 percent.   Higher-solids coatinqs are




20.  On page 3.15-9, the footnote at the bottom  of the page should be deleted.




21.  On page 3.15-12, the following references should be added:


     8.  Information supplied by Larry F.  Nonemaker, E. I.  Dupont  Demours &  Co.,  Inc.,
     Winnewood, Pa., April 14, 1978.


     9.  Letter from W. C. Moses, Technical Manager, Chemical  Plant Division,
     Suttcliffe, Speakman & Company, Limited,  March 10, 1978.
22.  On page 3.2-7, in the fourth line under "explanation of abbreviations,"
     "m3, dry/Mg ADP/day" should be changed to "dscm/min/Mg ADP/day."

-------

-------
                            CONTENTS
Acknowledgment                                                v

1.   Introduction and General Discussion                     1-1
     1.1  Introduction                                       1-1
     1.2  Relationship of NSPS, BACT, and LAER               1-2
     1.3  Legislative Basis                                  1-4
     1.4  LAER Guidance Documents                            1-5

2.   Methodology:  Approach and Procedures                   2-1
     2.1  Introduction                                       2-1
     2.2  State Implementation Plans (SIP) Limitations       2-1
     2.3  Achieved-In-Practice (AIP) Limitations             2-4
     2.4  New Source Performance Standards (NSPS)
          Limitations                                        2-5
     2.5  LAER Finalization                                  2-6

3.   LAER Guidelines                                         3-1
     3.1  Stationary Gas Turbines/Electric Utilities       3.1-1
     3.2  Kraft Pulp Mills                                 3.2-1
     3.3  Electric Arc Furnaces at Steel and Gray
          Iron Foundries                                   3.3-1
     3.4  Petroleum Refineries—Catalytic Cracking
          and Fuel Burning                                 3.4-1
     3.5  Fabric Coating                                   3.5-1
     3.6  Large Industrial Boilers                         3.6-1
     3.7  Primary Aluminum Plants                          3.7-1
     3.8  Bulk Gasoline Terminals                          3.8-1
     3.9  Flat Wood Paneling                               3.9-1
     3.10 Petroleum Liquids Storage                       3.10-1
     3.11 Petroleum Refineries—Wastewater Separators     3.11-1
     3.12 Petroleum Refineries—Fugitive Emissions        3.12-1
     3.13 Graphic Arts Printing                           3.13-1
     3.14 Automobile and Light Truck Coating              3.14-1
     3.15 Metal Furniture and Large Appliance Coating     3.15-1
     3.16 Can Coating                                     3.16-1
     3.17 Metal Coil Coating                              3.17-1
     3.18 Paper Coating                                   3.18-1
                               111

-------
                      CONTENTS (continued)
4.   Cost Estimating Methodology                             4-1
     4.1  Introduction                                       4-1
     4.2  Factors Affecting the Cost of Emission
          Control                                            4-2
     4.3  Capital Cost Estimates                             4-3
     4.4  Methodology for Estimating Annualized Costs        4-12
     4.5  Cost-effectiveness                                 4-23

5.   Financial and Economic Analysis Techniques              5-1
     5.1  Introduction                                       5-1
     5.2  The Analysis Techniques                            5-2
     5.3  Information Needed to Perform Analysis
          Techniques                                         5-lf

Appendix A                                                   A-l

Appendix B                                                   B-l
                                IV

-------
                         ACKNOWLEDGEMENT

     This report was prepared under the direction of Mr. Jack A.
Wunderle.  Principal authors within PEDCo Environmental, Inc.,
were Messrs. Joseph Carvitti and Jack A. Wunderle.  The JACA
Corporation prepared Chapter 5.
     Project Officers for the Environmental Protection Agency
were Messrs. John H. Haines and Gary D. McCutchen.  The authors
appreciate the contributions made by the EPA project officers and
their associates.

-------
                           SECTION 1
               INTRODUCTION AND GENERAL DISCUSSION

1.1  INTRODUCTION
     The  Clean  Air  Act as  amended  in  1977  (CAAA  1977)  contains  three
technology-based limitations  affecting the location and construction  of new or
modified air pollution sources:   (1) New Source Performance  Standards (NSPS);
(2)  Best  Available  Control  Technology  (BACT); and (3)  Lowest Achievable
Emission   Rate   (LAER).   Although  NSPS and  BACT are  relatively familiar
concepts   that  have  been  incorporated  into  statutes, regulations,  and imple-
mentation  plans, LAER  is relatively  new.   The  1977 amendments  established
LAER as  a  rtatutory  requirement, and EPA's  Emission  Offset  Interpretative
Ruling incorporated  LAER  as a  regulatory demand in  the preconstruction
review of major  stationary sources   which would contribute to a violation of
an  NAAQS.   Additionally, the  CAAA  1977  (§178)  requires that the Admin-
istrator issue  guidance  documents   for  the purpose  of assisting  states  in
implementing the requirements of §129 and §173 pertaining to lowest  achievable
emission rate.
     The  purpose  of  this document is  to  provide technical  assistance and
guidance  to  those who  must  prepare  and  submit  applications  for  proposed
                                                   *
construction  or modification  of sources and  facilities  and to those who must
approve or deny such applications.
  Throughout  this document the use of certain terms--major new
  source,  facility, major modification, potential  emissions,
  allowable emissions, NSPS, BACT, LAER--is intended  to  reflect
  the meanings  imparted to those terms by the CAAA 1977.
                              1-1

-------
     The  following sections  deal with the  interrelationships of the new source
requirements  embodied  in  the  CAAA  1977;  the  legislative  basis  of  LAER
(including the factors that must be  considered in determining LAER); some of
the limitations and  constraints involved in LAER determination;  and, finally,
the scope,  intent,  applicability,  use,  and limitations  of  the source-specific
guidance  in Section  3.

1.2  RELATIONSHIP OF NSPS, BACT, AND LAER
     Because  LAER is a relatively new requirement, it should be considered
in relation  to the  other requirements  applicable  to  new or modified major
sources or facilities, that is, to NSPS and BACT.
     The  NSPS are applicable wherever a new source intends to  locate  (01  in
the case  of  a modified  source—is located).   They are limited to the specific
sources and  facilities  that are the subject of a Federal  promulgation  (Title 40,
Chapter  I,  Subchapter  C, Part 60 CPU—Statutory authority,  §111).  An NSPS
generally embodies  specific  emission  limitations and may include operational
and   performance  standards.   The  statute  directs  that the  "standard  of
performance"  shall  be  the  best technological system  of  continuous emission
reduction  adequately  demonstrated,  considering cost,  energy requirements,
and other effects, such as socioeconomic impacts.
     The BACT  requirement is  to be  applied  to  new and  modified sources
subject  to  Prevention   of   Significant  Deterioration   (PSD)  review  and  is
required  for  each pollutant subject to  regulation under the Act.  The BACT
review is relevant to  major  new or modified stationary sources or facilities.  '
The BACT  and  NSPS definitions are similar in that BACT must be technically
achievable and must  also reflect  consideration of  cost, energy requirements,
and other possible  impacts.   The BACT  requirement however is to be applied
on a  case-by-case  basis and is also to consider any alternatives  (production,
process  modification,  control methods,  systems, and techniques) that either
singly or in combination could lead to further reduction in emissions if
                               1-2

-------
applied.   Further,  BACT is  to  be at  least as  stringent  as  any  applicable

NSPS  or   National   Emission   Standards   for   Hazardous  Air   Pollutants

(NESHAPS).  In  a  given case,  therefore, the imposition of BACT  limitations

may be more stringent than NSPS or NESHAPS.

     The  LAER requirements are imposed in the review process on a proposed

major  new  source or  proposed major modification  in  a nonattainment area.

As  applied  to a  modification,  LAER means the lowest achievable  emission rate

for the new or modified facilities  within  the source.

     Section II.A.7  of  the  proposed  Emission   Offset  Interpretive  Ruling

defines LAER as  follows:


          "Lowest achievable  emission rate" means  for  any source,
     that rate  of emissions based on the  following, whichever is
     more stringent:
               (i) the most stringent emission limitation which
          is contained in the  implementation plan  of any State
          for such class or category of  source, unless the owner
          or operator  of the proposed source demonstrates that
          such  limitations are  not  achievable, or

               (ii)  the most  stringent emission limitation which
          is achieved  in practice by such class or category of
          source.

     This term, applied to a modification, means the LAER
     for the new  or  modified  facilities within the source.
     In no event  shall the application of this term permit a
     proposed new or modified  source to  emit any pollutant in
     excess  of  the amount allowable under applicable new source
     standards  of performance.

     Although  transfer  of technology had only a minor role in  the determi-
nation  of LAER in this document,  such  decisions  were guided  by a pertinent
discussion   in  EPA's  proposed  Emission  Offset  Interpretive Ruling.    This

discussion   ("Technology  transfer   in  determining LAER,"  p.  3208) states:


          It has  been  EPA's interpretation that in determining
     the  lowest achievable emission rate  (LAER), the reviewing
     authority may consider transfer of  technology from one
     source  type  to  another where such technology  is applicable.
     Although Congress  changed  the  definition of LAER, EPA con-
     tinues  *:o believe  that technology transfer may be considered
     in dete .mining LAER.  Congress  intended to require new
                              1-3

-------
     sources in nonattainment areas to apply the "maximum  feas-
     ible pollution control," even if this involves
     "technology-forcing."  Therefore, the Agency does  not feel
     that the phrase "achieved in practice by such class or
     category of source"  [under Section 171(3)]  prohibits
     technology transfers from other types of sources.   If
     pollution-control technology can feasibly be transferred
     from one type of source to another, then for purposes of
     determining LAER, EPA will consider both types  of  source  to
     be in  the same "class or category of source."

     Of course,  technology transfer need not  be  considered in  determining

LAER  if it is  not feasible  for  the  specific  application  under  consideration.

Further, NSPS,  BACT,  and  LAER  all  must be  achievable  before they  are

applied.  The LAER  provisions are  generally more  stringent  than BACT, and

BACT  may  be more  stringent  than  NSPS.   Neither BACT nor  LAER can  be

less  stringent than  an applicable  NSPS  for  the specific source  or facility and

the specific pollutant emitted.


1.3  LEGISLATIVE  BASIS

     Insight as  to Congressional intent  in adopting the LAER requirement and

guidance to be  followed  in developing LAER limits are found  in the Congress-

ional Record (House,  August  3,  1977,  p. H 8551).   The  following quotation is

from conference agreements  between House and Senate.

          The House definition  of  "lowest achievable emission
     rate" is adopted for purposes of this section.  In determin-
     ing whether an emission rate  is achievable,  cost  will have
     to be taken into account,  but cost factors in the nonattain-
     ment content will have somewhat less weight than  in  deter-
     mining new source performance standards  under Section 111.
     Of course, health considerations are of prime importance.
     Facilities seeking to locate  or expand in  areas not  meeting
     air quality health standards  should be required to use the
     best control technology and processes available.  The defi-
     nition is intended to describe the lowest  rate  which is
     actually, not theoretically,  possible.   If the  cost  of a
     given control strategy is so  great that a  new major  new
     source could not be built or operated, then such  a control
     would not be achievable and could not be required by the
     Administrator.
                               1-4

-------
     These  citations  show clearly  that  LAER is intended to require more than
add-on  control  devices.   It calls  for the  consideration, selection, and appli-
cation  of alternative production  procedures, modifications  of  unit processes,
and  control  techniques  in a  combination  that results in minimal pollutant  re-
lease.   In  so doing, it  establishes environmental concern as a prominent con-
sideration in the early stages of planning of new and modified  sources.

1.4   LAER  GUIDANCE DOCUMENTS

1.4.1  Purpose

     The CAAA 1977  calls for the issuance of guidance documents to assist  the
state in  implementing Part  D, Title 1,  of  the Act as it pertains to application
of the  lowest  achievable  emission rate to  major  new  or  modified  stationary
sources  (§178).   In satisfying  this   requirement,   EPA  intends  to  provide
guidance  in the  form of reference  material that is useful to engineers  in
industry and in  state or local  agencies,  to  those  who  prepare and those who
review permit applications, to those regulated by the LAER requirement, and
to those  who implement that requirement.

1.4.2  Scope and  Limitations
     The user  must  recognize that the guidelines  presented herein are limited
as to scope  and  subject.  This document pertains to selected major stationary
sources, to certain  emitting facilities  within those  sources,  and to  specific
emissions from  those sources or  facilities.  In the  section on primary aluminum
reduction  plants,  for  example,  the  LAER  guidance  addresses only  sulfur
dioxide  emissions  from anode bake ovens  and reduction cell facilities.   It does
not address LAER for  grinding and mixing  of materials,  casting and  tapping
of metal, or other emission points.  Nor does it address other  emissions (such
as hydrocarbons, fluorides, and particulates) from the  anode bake and reduc-
tion  cell operations.
     It  is  important  to emphasize  that  the LAER criteria apply to all proposed
major new  or modified sources.  The fact  that no  guidance  document has been
                               1-5

-------
issued for a source-facility  emission grouping does not relieve the appropriate
authority  from  the obligation  to apply LAER  criteria in the  permit  review
process;  nor  does it relieve  the applicant from seeking  to  attain the lowest
achievable emission rate.
     The  LAER requirement, like BACT and  NSPS,  is considered to be evolu-
tionary and subject to  change.   Achievable levels  of  air pollution control will
improve with  technological  advances.    As  new  and  improved technologies or
processes are conceived, demonstrated,  and practiced,  they will be recognized
and  used.  Today's  state of the art  will be  replaced by that of  tomorrow.
     Finally,  the technical  information and guidance presented  in this report
are  not intended  to  apply  to  all  situations  and  specific conditions  of the
industrial  processes  and  pollutants described  here.  This is termed a guid-
ance  document  because it provides useful reference  materials, but  it should
serve  only as a starting point in a LAER investigative review.
     In preparing  these initial LAER guidelines, the time  constraints specified
in the CAAA  1977 and the need  to allocate technical and professional  resources
within that time frame  limited  the  scope  and detail of the document.   These
time and  manpower constraints dictated  the following limitations  on preparation
of the guidance documents:

     0    No visits were made to sources or facilities  to  confirm
          technology and/or performance.
     0    The technical data base was  drawn from telephone or
          written communication with equipment vendors,  facility
          owners, governmental control agencies, and other
          groups, coupled with limited perusal of published
          literature.
     0    A cut-off date was established for  data acquisition to
          provide sufficient time to catalog  and analyze the
          information before preparing the document.
     0    A "current as of" date was fixed for review of regula-
          tions set forth in the State Implementation Plans
           (SIP), which undergo frequent revision.
     Acquisition of technical information and  SIP limitations through query by
telephone or  mail  is subject to the celerity and generosity of the respondent.
                               1-6

-------
Although  every effort  was made to maximize  the quality of the  data,  the need
to impose  deadlines  on  data acquisition  may  have  caused  the  exclusion  of
certain appropriate and relevant information.
                               1-7

-------
                           REFERENCES
     Appendix S—Emission Offset Interpretive Ruling Federal
     Register, Vol.  44,  No.  11,  January 16,  1979.   p.  3282.
     (Although this  was  designated in the Federal  Register as a
     final rule,  the Administrator did invite comment on certain
     portions of  the Ruling,  indicating the  possibility of
     change).

     Title 40, part  51 of the Code of Federal Regulations,
     §51.24(b), (1)  and (2),  FR Vol.  43, No. 118,  June 18, 1978,
     p.  26382 and as specified in §169(1), CAAA 1977.
3.   Reference 2.

4.   Reference 1,

5.   Reference 1.

6.   Reference 1.
     p. 3283.
§51.24(b)(17) p.  26383.

(II)(A) Definitions.  (4) and (5) p. 3282.

(II)(A) Definitions.  (6).  p.  3282.

(IV, Condition 1).   p. 3284 and footnote 2,
                               1-8

-------
                            SECTION  2
              METHODOLOGY:   APPROACH AND PROCEDURES

2.1  INTRODUCTION
     The   CAAA  1977   defines   LAER  as   the   more   stringent   of  two
emission   limitations   for    a    class    or    category    of   source—that
contained   in    the   implementation    plan   of   any    state   or   that
achieved  in   practice.    The  Act  further  requires   that   LAER  be  at
least   as   stringent  as   an   applicable   NSPS   limitation.    Thus,   the
Act    gives    clear    direction   as   to    the   major    considerations   in
determining    LAER.     The    methodology,   approach,    and    procedures
utilized   v.ere   geared   to    discovering,    for   each    source/facility/
pollutant     grouping,     the    most    stringent    emission     limitation
embodied    in    a    State   Implementation   Plan    (SIP),    achieved   in
practice  (AIP),   or  required   by   an   applicable  New   Source  Perfor-
mance   Standard   (NSPS).    Figure   2-1,   LAER  Determination   Decision
Tree,  illustrates the approach and procedures used.

2.2  STATE IMPLEMENTATION PLAN (SIP) LIMITATIONS
     The   first   step  in   the   review   of  SIP   limitations   was  to  ac-
quire   an   up-to-date   file   of   current  air  pollution  control  regula-
tions    of   the   major   state   and   local  air   pollution   control   agen-
cies.     The   acquisition   and    verification   procedures   were   as   fol-
lows:

     1.   Letters   were   mailed   to   all   state    agencies   and   the  most
          significant   city,    county,    and   regional   agencies    within
          each    state.     This    letter    advised    the   agency    of   the
          purpose   of  the   LAER  study   and   requested   that  the   agency
          send,   as   soon   as   possible,   a   complete   and   current  copy
          of its air pollution control  laws and regulations.
                               2-1

-------
REG.
REVIEW
                                     QUERY CONTACTS:
                                    STATE AND LOCAL,
                                    CONTROL VENDORS,
                                     TRADE ASSOC.,
                                    SOURCE OWNERS,
                                         ETC.
       SELECT
        MOST
     STRINGENT
   E.L.(TENTATIVE)
NEDS
RETRIEVAL
*
EXAMINE
FOR AIP


  SOURCE RESIDENCE,
 REG. INTERPRETATION,
 ACHIEVED IN PRACTICE
AND CONTROL TECHNOLOGY
      QUERIES
                                                                               ASSESS  EMISSION
                                                                               REDUCTION METHODS
                                                                               SELECT TENTATIVE
                                                                               BERS COMBINATION
                                                            LEGEND:
                                                            SIP  -
                                                            AIP  -
                                                            NSPS -
                                                            E.L. -
                                                            REG  -
                                                            LAER -
                                                            BERS -
                                                            NEDS -
STATE IMPLEMENTATION PLAN
ACHIEVED-IN-PRACTICE
NEW SOURCE PERFORMANCE STANDARDS
EMISSION LIMIT
REGULATION
LOWEST ACHIEVABLE EMISSION RATE
BEST EMISSION REDUCTION SYSTEM
NATIONAL EMISSION DATA SYSTEM
Figure  2-1.    LAER  Determination  "Decision   Tree"

                                         2-2

-------
     2.    Approximately   2    weeks    after    the    letters    were    sent,
          follow-up   calls   were  made   to   jurisdictions   that   had   not
          yet    responded.     The   agency   was    again   advised    of    the
          study  purpose   and   requested   to  send   a   copy   of   current
          regulations.

     3.    Two    cross-checks    were   made   to    verify   that   the   state
          regulations    provided   were,    in   fact,   the   most    current:
          the    regulations   given   in   the  Environmental    Reporter
          State   Air   Laws,   published    by  the   Bureau    of    National
          Affairs,    Inc.,    were   compared   with   those   sent    by    the
          states;    the  PEDCo   files   of   state   and   local   air  pollu-
          tion  regulations,    last   updated   in    June   1976,   were  also
          examined.

     Each   response,   as   received,   was   dated   and   filed.    An   ac-
counting    of   those    jurisdictions    not    responding   was   maintained.
Final    tabulations     indicated   that    the    files    contained     current
regulations    for    all   states   and   for   54  city,   county,    and  local

jurisdictions.     Generally,     all    regulations     were   current     as    of

January  31, 1978.
     In   review  of   the   regulations   of   each   jurisdiction   to   ascer-

tain  the   lowest   emission   limitation   required   in   an   SIP,   the   regu-

lations were grouped into sets and subsets:


     General Fuel Burning Regulations

          - Particulate

          - Sulfur Dioxide

          - Oxides of Nitrogen

     General Manufacturing Process Regulations

          - Particulate

          - Sulfur Dioxide

          - Volatile Organic Compounds

     Source-Specific Regulations

          - By the  specific major source types named in the
            guidance document
                                2-3

-------
The    regulations   of   each    jurisdiction   were   reviewed   and    placed
into   the   above  groupings.    Where  a   particular  jurisdiction   had  no
applicable    regulation   specific   for   a    given   source    (e.g.,    Kraft
Pulp    Mills   or   Primary    Aluminum    Plants)    an   applicable    general
process   regulation  was  cited.    Each   category  was   then  examined  to
ascertain    the   most    stringent   SIP    emission   limitation.    The    limi-
tations   were   viewed   in    conjunction   with   the  specified   compliance
determination   method   (source   test   or    other)    because   the    two
parameters--emission      limitation      and      compliance      determination
method—together  determine   the  relative   degree  of  stringency  of   a
limitation.
     This    process    served    to    identify    the   most    stringent    SIP
limitation     and   the   responsible    jurisdiction.      The    final    steps
were   to    determine   whether  a   source   is   resident  in  the   jurisdic-
tion   (or   whether  the   regulation   was   designed  to  preclude   entrance
of   such    source),   to   resolve    any   questions  involving   interpreta-
tion    of    a   regulation,    and    to    acquire   information   regarding
sources that had attained emission levels within the limitation.

2.3  ACHIEVED-IN-PRACTICE (AIP) LIMITATIONS
     Three  methods  were    used    to   obtain   information  regarding   the
lowest  emission   rate  achieved   in  practice   and  the   attendant   con-
trol    technology:     queries   were   made   to   state   and  local   control
agencies,     control    equipment    vendors,    trade   associations,    source
owners,    and   other;    published   literature   was    reviewed;   and   a
limited    "quick   look"   retrieval    was    obtained   from   the   National
Emissions   Data   System   (NEDS).    This   retrieval   covered   the  entire
nation   and   listed    for    each    pollutant/facility/source   grouping   of
interest    the    source    and    location,    source    classification    code,
control   device  used,   and   emission   reduction   achieved.    To   reduce
the    volume   of    this    listing   the   retrieval   logic   incorporated   a
control   efficiency    cut-off   for    each    source-facility    so   that   only
those   attaining   a   control   efficiency   greater   than   the   specified
cut-off value were listed.  Information gleaned from the several
                               2-4

-------
methods    was   then   examined   to   determine    the   lowest   emission   rate

achieved  in  practice.


2.4   NEW  SOURCE PERFORMANCE STANDARDS  (NSPS) LIMITATIONS

      EPA   regulations    adopted   under    statutory   authority   of   Section
111,   CAAA   1977,   and   promulgated  under  Title   40,  Chapter  I,   Sub-
chapter   C  Part  60  CFR  are  known  as  New Source Performance  Standards
(NSPS).    Such   standards   are   required   by   the   Act   to   reflect   the
degree     of    emission    limitation    achievable    by    application    of    the
best   system  of   emission   reduction  that  has   been  adequately  demon-
strated.     NSPS   have    been    promulgated   for    some   of   the   sources
addressed   in   this   document.    The   NSPS   citation   in   the   Code   of
Federal    Regulations    (CFR)   and   the   companion   LAER   guidance   cate-
gory  (subsection of Section 3) are listed below.
                   LAER CATEGORY

              3.1   Stationary Gas Turbines
                   (utility power plants -
                   oxides of nitrogen, sulfur
                   dioxide)

              3.2   Kraft Pulp Mills  (lime
                   kilns - particulate; re-
                   covery furnaces - sulfur
                   dioxide; and power boilers
                   - particulate)
              3.3  Electric Arc Furnaces
                  (steel  and gray  iron found-
                  dries - particulate)

              3.3  Electric Arc Furnaces
                  (steel  and gray  iron
                  foundries - particulate)

              3.4  Petroleum Refiners (cata-
                  lytic crackers - sulfur
                  dioxide)
     NSPS - 60 CFR SUBPART
     (AFFECTED FACILITY)

GG  Stationary Gas Turbines
    (>1000 Hp - oxides  of nitro-
    gen, sulfur dioxide) - pro-
    posed only - Oct. 1977

BB  Recovery furnaces (particulate),
    Smelt dissolving tanks (partic-
    ulate),  Lime kilns  [particulate,
    total reduced sulfur (TRS)
    limited to 5 ppmv for seven
    facilities].

Z   Ferroalloy Production Facilities*
    (electric submerged arc furnaces  -
    particulate).

AA  Steel Plants*  (electric arc fur-
    naces -  particulates)
    Petroleum Refineries (catalytic
    cracker - particulate,  CO)
              3.6  Industrial Boilers         D
                  (particulate,  sulfur di-
                  oxide, oxides  of nitrogen)
              3.7  Primary Aluminum Reduction  S
                  (reduction cells and anode
                  bake  oven - sulfur dioxide)
              3.10 Petroleum Liquids
                  Storage (hydrocarbons)
   Fossil-Fuel-Fired Steam Gen-
   erators  (>250 million Btu/h -
   particulate, sulfur dioxide,
   oxides of nitrogen)

   Primary  Aluminum Reduction
   Plants (reduction cell pot-
   rooms and anode bake plant -
   fluorides)

   Storage  Vessels for Petroleum
   Liquids  (>40,000 gal capacity
   hydrocarbons)
              * NSPS are planned specific to particulates  from electric arc
               furnaces in steel and gray iron foundries.
                                    2-5

-------
     The   NSPS   specifies   an   affected   facility   and   the   pollutants
covered   by   the   standard.    The   LAER  guidance   also   pertains   to
specific   facilities   and   pollutants.    The   two   NSPS  and   LAER  cate-
gories    were    examined    for    a   match   of   source/facility/pollutant
groupings.    Where   that  match   occurred,  the   NSPS  emission  limita-
tion   was   recorded   for   comparison  with  the  AIP-   and   SIP-derived
LAER values.

2.5  LAER FINALIZATION
     The   tentative   LAER  limits   from  review   of   AIP   and   SIP   were
then   compared   and   the  more   stringent  of   the  two  was   selected.
Where   an  NSPS   was  applicable,   the  AIP-SIP  selection  was   compared
with   the  NSPS   limitation   and   again   the   more  stringent   was  se-
lected.    Finally   consideration   was  given   to  changes   in  raw   ma-
terial   input,    to   production    and   process   modifications,   and    to
technology   transfer  in  a   combination   that   would   represent  a   best
system   of   emission  reduction   that   could   be  applied   to   a  given
source/facility/pollutant    set.     A    final    LAER   value    was    then
selected,    and   the   technology    for   achieving   the   emission   limita-
tion   was   described.    Factors   that  could   limit  application  of  the
technology were addressed.
     The  draft  document  was   then  organized  in  standard  format  of
the   EPA  Office  of  Research   and  Development   (ORD),   subjected  to
editorial   scrutiny,   and  offered   to    affected   and   interested    par-
ties  for  review  and  comment.   The  comment  period  was  intended  to
allow   for   discovery   and  correction  of   any  errors   and   to   permit
 consideration   of   comments   and   other   information   that  would  en-
 hance  the   clarity  and   technical  accuracy  of  the   document.   Appro-
 priate   modifications   based  on   these   and   internal   review   comments
 have been incorporated into the final  document.
                                2-6

-------
                            SECTION 3
                         LAER GUIDELINES

     This Section consists of  18  subsections that deal with LAER
for  pollutants  emitted  from  the  facilities  of certain  major
sources.  Not  all pollutants  that may be emitted from a facility
are covered, nor are  all the facilities at  a  major  source.   The
guidance pertains to  a limited number  of  stationary sources,  to
selected facilities within those sources, and to specific emis-
sions from  those facilities.  The  format  of  each guidance  con-
sists of Process  Description,  Emissions,  Control Measures, Emis-
sion Limits, and Determination of LAER.
     This  is  a  technical document,  for  guidance  only,  and  it
should not  be  construed as regulatory in nature.  Its purpose is
to provide  technical  information  and reference materials for use
as  a starting point  in  making  a  LAER  determination,  not for
rigidly prescribing  LAER.  The information  presented is limited
in  scope  and  is not intended  to  apply  to  all  situations and
conditions  that may be encountered in the review of permit appli-
cations.  The  guidances  do not replace the  individualized atten-
tion  and  consideration to be  afforded an entity that seeks ap-
proval  to  install  a  new--or modify an  existing—source.   The
review  authority is  encouraged to recognize the individuality of
each  permit application  and the desirability of a  case-by-case
approach  that  ultimately  leads to  a LAER  determination based on
and specifically tailored  to a given  set of  circumstances.
     Caution should  be exercised  in utilizing this  document be-
cause  of  the  anticipated changes  in the  CAAA  1977  specified
criteria  upon  which  LAER determinations  are to  be made,  i.e.,
State   Implementation  Plan   (SIP)   limits,  achieved-in-practice
                               3-1

-------
(AIP)  levels,  and  New  Source  Performance  Standards  (NSPS).
Changes  in  SIP-related limitations  are  imminent in  response to
CAAA 1977 requirements  that  SIP's  be revised for areas where the
National  Ambient Air  Quality  Standards  (NAAQS)  have not  been
attained.  Also, advances  in control technology and in equipment
performance can  be  expected to  result in AIP  levels lower than
those  reported  here.   Furthermore,  an expanded list  of  sources
are  to  be the subject  of  NSPS promulgation in the near  future,
and  consideration is being given to  revising  some  NSPS.   There-
fore, the user  should  verify that the SIP, AIP, and  NSPS limits
in this  document are currently  applicable  and have  not  changed
since document publication.
     Finally,  it is  important to note that these guidelines apply
to both  new and  modified  facilities.  With regard  to modified or
reconstructed  facilities,  the  reviewer  may  give  case-by-case
consideration to any  special economic  or physical  constraints
that might limit the application of certain control techniques to
a modification project,  i.e., the level of control  required for a
process  undergoing  modification  or  reconstruction  may not be as
stringent as that required if the same process  were a grass-roots
construction project.
                               3-2

-------
3.1  MAJOR  SOURCE  CATEGORY:   STATIONARY  GAS  TURBINES/ELECTRIC
     UTILITIES—NITRIC OXIDE AND SULFUR DIOXIDE EMISSIONS
3.1.1  Process Description
     A gas  turbine  is a rotary engine, of which a common example
is  the  jet aircraft  engine.   Compressed  air is rapidly expanded
by  the  combustion of  a  fuel in a  combustion chamber.   The high
velocity  and  high  temperature gases  rotate a turbine  fan that
drives  a power-output  shaft.   Figure  3.1-1 presents  a cutaway
view of  a stationary gas turbine.   Turbines range  in size from
less than  30  kW (40 hp) to  over 75  MW (105 hp).2  Manufacturers
continue  to increase  turbine  capacity,  and turbines  are  often
installed  in  groups,  so that the combined  power  output from one
location may  exceed  1.12  GW (1.5 x 106 hp).3  Over 90 percent of
the horsepower sold in the U.S. goes to utilities and that margin
                   4
continues to widen.
     Three  basic  types of gas turbines are  used  in  the electric
power industry:   simple cycle, regenerative cycle,  and combined
cycle.    A  simple  cycle gas  turbine  consists typically  of one or
more compressor stages,  one or more  combustion chambers  where
liquid or  gaseous  fuels are burned, and  one or more turbines to
drive the  compressor  and the  load.   These  can  be  arranged  in
various  configurations.   Figure 3.1-2  is  a block  diagram  of a
typical  simple  cycle gas turbine.   The  turbine  is  started with
an  electric motor,  diesel  engine,  or  other  energy source  to
rotate the compressor that provides compressed air to the combus-
tors.  Fuel is  then  introduced into the combustors and burned to
produce  hot gases,  which expand across the  first  set(s) of tur-
bine blades, providing the driving force to continue rotating the
compressor(s)  mounted on  the  same shaft.   The hot gases  are
further  expanded  across  the power turbine blades that  drive the
electrical generators.  The exhaust gases, containing pollutants,
exit to the atmosphere at temperatures ranging from 430° to 600°C
(800° to 1100°F).6
                              3.1-1

-------
                                                                                    EXHAUST
I
N)
                                                              EXHAUST
                                                               DUCT
                                               HIGH
                                             COMPRESSOR
                                                                HIGH
                                                             COMPRESSOR
                                                              TURBINE
   LOW
COMPRESSOR
 TURBINE
 POWER
TURBINE
               Figure 3.1-1.   Cutaway  view  of a  typical  stationary gas turbine.

-------
U)
I
u>
                                   STATIONARY GAS TURBINE
                                                                         Vy
                                                                         ROTARY
                                                                         ENERGY
                                                                     J
LOAD
                        Figure  3.1-2.   Typical  simple  cycle gas  turbine.

-------
     The regenerative  cycle  gas turbine is  essentially a simple
cycle  gas  turbine  with  an  added  heat exchanger,  as  shown  in
Figure  3.1-3.    Thermal  energy is  recovered  from  the  430°  to
                                      f.
600°C  (800°  to  1100°F)  exhaust gases   and  used to  preheat the
compressed air.   Since less  fuel is required to heat the com-
pressed air  to the design turbine inlet temperature,  the regen-
erative  cycle  improves   the  overall  efficiency  of  the simple
cycle.
     The combined cycle gas turbine also recovers waste heat from
the turbine  exhaust gases.   It  is essentially a simple cycle gas
turbine with the  hot  gases vented  to  a  waste heat  boiler,  as
                      Q
shown in Figure 3.1-4.    Steam generated by the waste heat boiler
can  be  used  to  generate  electricity  with  conventional  steam
turbines.   Some  waste heat  boilers  are designed   to  generate
additional steam by the firing of conventional  fuels in a fire-
box.   Such  systems are  known  as supplementary-fired  combined
cycle gas turbines.

3.1.2  Emissions
     The pollutants generated by gas turbines  are those common to
all  combustion processes:  NOX,  HC,  CO,  S02,  particulates, and
visible emissions.  Table 3.1-1 summarizes the typical pollutant
emissions,  which  are   dependent  on  such variables  as turbine
firing  temperature,  turbine pressure  ratio,  turbine load, com-
bustor  design,  fuel  characteristics,  and atmospheric  conditions.
This section discusses only S02 and N0x, for which lowest achiev-
able emission  rates (LAER) are  developed.

3.1.2.1  Nitric  Oxides (NO )--
                          J\
     Nitric  oxides produced  by  combustion of  fuels  in stationary
gas turbines are formed by the  combination of  nitrogen and  oxygen
in  the  combustion air  (thermal  NO ) and from  the reaction  of the
                                  X
nitrogen  in  the  fuel with  the  oxygen  in  the  combustion air
(organic  NO ).  Formation  mechanisms  are complex,  and  detailed
            X
discussion is  beyond the  scope  of  this document.

                              3.1-4

-------
U)
•

M
I
Ul
                                             EXHAUST
                                 STATIONARY GAS TURBINES
                                                                                     LOAD
                   Figure 3.1-3.   Typical  regenerative cyclone gas turbine,

-------
             FUEL
                                      WASTE  HEAT
                                        BOILER
 COMPRESSED
     AIR —»

AIR
COMBUSTION
 CHAMBER  I
                              -EXHAUST-H—
     COMPRESSOR

AIR
          STATIONARY GAS TURBINE
                                                      LOAD
                                            ROTARY
                                            ENERGY
           Figure 3.1-4.   Typical combined cycle gas  turbine,

-------
                         TABLE 3.1-1.   UNCONTROLLED  POLLUTANT  EMISSIONS
                              FROM TYPICAL STATIONARY GAS TURBINES9
u>
Pollutant
Carbon
monoxide
Sulfur b
dioxide
Nitric
oxides
Hydro
carbons
Particulate
Symbol
CO
S02
NO
X
HC

Uncontrolled emissions
g/kWh
0.13-16
5.35-7.75
7.3-4.0
0.22-1.1
NR
lb/103 hp-h
0.2-26
8.75-13.5
12-66
0.36-1.8
NR
ppmv @ 15% O2
2.3-160
NR
50-350
1-5 (as hexane)
0.002-0.1 gr/scf
               Highly dependent  on  combustion efficiency.
               S = sulfur content of  fuel  in weight percent.
               Highly dependent  on  combustion temperature,  combustor
               design,  fuel nitrogen, and other factors as  discussed.
             d
               Highly  dependent on type of  fuel and combustion  efficiency.
             NR =  not  reported.

-------
     Several  major  factors  limit  the  formation  of  NC>X:   the
availability of 02/ the combustor residence time and temperature,
the  amount  of moisture  in the  inlet air,  and  combustor  pres-
sure.10'11'12  The quantity of  NO   emissions  generated by  gas
                                   X
turbines is limited by the residence time of the hot gases in the
engine combustors  and by the  rapid quenching  of these  gases by
dilution air.  Therefore, very high combustion efficiencies (high
temperature)  can be  attained  without generating the very high
equilibrium  quantities  of NOX.  Humidity  in the  inlet  air will
decrease NO   formation  by reducing the combustion flame tempera-
     12    x
ture.
     Organic  NO   is formed  during  combustion by  the  chemical
combination  of the nitrogen atoms contained in the fuel molecule
with   oxygen  in    the   air.    The   exact  mechanism   is  not
known.13'14'15   Generally,  organic  N0x  is a  problem  only in
burning  of  residual oils,  blends,  some  crude  oils,  or  heavy
distillates  that have high  nitrogen contents.    '    Table 3.1-2
indicates the nitrogen content of various  fuels.
     NO   emissions  from  gas  turbines were measured and  reported
for  over 50  source tests.20  Uncontrolled  NOX emissions  ranged
from about  40  to  500  ppm at 15  percent  02 for all  fuels, and
averaged  90  ppm  for natural gas,  130  ppm for distillate  fuel, and
190  ppm  for jet-A  type  fuel.   Table 3.1-3 gives the range and
average  of  uncontrolled NOX emissions from  the sources  that  were
tested which  include combustor  rigs (models)   as well  as  full-
scale  turbines.

3.1.2.2   Sulfur  Dioxide (S02)--
     The  formation  of S02  in stationary  turbine operations  is
strictly  a  function of the sulfur content  of the fuel.  Generally
100  percent of the sulfur is  converted to  SO2.   Table  3.1-2 shows
the  typical  concentrations of  sulfur  in  common  fuels.   Turbine
operators generally use  low-sulfur  natural  gas and light distil-
 lates,21  although crudes and residuals can also be used.
                               3.1-8

-------
TABLE 3.1-2.   NITROGEN AND SULFUR CONTENTS OF COMMON FUELS
Fuel
Natural gas
Distillate oil
Crude oil
Residual oil
Content, percent by weight
Nitrogena
nil
<0.015
<0.2
<2.0
Sulfurb
0-0.1
0.01-0.48
0.06-3.0
0.5-3.2
References 15 and 16.


References 17 and 18.
         TABLE 3.1-3.  N0v EMISSIONS BY TYPE OF FUEL
                         X
                                                    20
                Uncontrolled NO  emissions,
                               X

                        ppmv @ 15% 09
Fuel
Natural gas
Distillate fuel
Jet-A
Range
40-150
50-240
40-500
Average
90
130
190
Number of tests
18
30
9
                             3.1-9

-------
     The  SO_  concentration is  calculated at  180  to 260  S  ppm,
depending upon the sulfur content (S) of the fuel and the exhaust
gas  rate.   The sulfur  content  is  expressed  in weight percent.
The  mass rate  emission  factor  is  determined to be  0.02  S  kg
S02/kg fuel  (0.02  S  Ib  SO2/lb fuel), where S  is  the sulfur con-
tent in  percent by weight.   This  is approximately equivalent to
5.3 S to 7.7 S g S02/kWh (8.7 S to 13 S Ib SO2/103 hp-h).

3.1.3  Control Measures
3.1.3.1  NO  Control--
           X
     Wet control techniques--Because  formation  of  NO   is  ex-
      •             —LJ		 -                            ,/V
tremely  sensitive  to  flame  temperature,  injecting water or  steam
                                                           22
into  the reaction zone  will reduce  production  of NO .     In
                                                         X.
full-scale  field  operations  reductions  of up  to 70 to 90 percent
have  been  achieved  at  water-to-fuel  (w/f)   ratios  of  1.0 and
reductions  of 50  to 70 percent  have been observed at w/f ratios
from  0.5 to  0.7,  as  shown in Figure  3.1-5.23  Industry readily
                                             24
accepts  this technique  for  control of  NO .     One manufacturer
                                           X
guarantees  an NO  emission  limit of 75  ppm at 15 percent  stack
            25    X
gas  oxygen.
     For distillate-fuel-fired turbines,  one  EPA  test shows the
highest  reduction of 85 percent, from 315 ppm to  58 ppm N0x (w/f
ratio  of l.l).26  Another EPA test showed the lowest  concentra-
tion of  about  26 ppm NO  ,  inlet concentration 163  ppm,  and 84
                         J*.
percent  efficiency in the stack gas of a liquid-fuel-fired tur-
                                        97
bine at  a  water-to-fuel ratio  of  1.1.     For natural-gas-fired
turbines the highest reduction  in  one  source test was from 110
ppm  to  13  ppm NO  (88 percent)  at a  w/f  ratio of 1.0.   This
                   X
facility also yielded the least NO  emissions (13 ppm) with wet
    .   ,   28                        X
controls.
     Water-to-fuel ratios  above  1.0 do not decrease  the formation
of NO   substantially.  Water  and steam injection have essentially
     J\
the  same effect  on  NO  emissions.   The overall  gas turbine ef-
                       2^
ficiency is  reduced  by  about 1  percent at  a w/f ratio of 1.0,
                               3.1-10

-------
 o
 -=>
 o
    90
    80
     70
     60
     50
     40
     30
     20
     10
              D NATURAL GAS


              O LIQUID FUEL
              /

                o


             o0o
       /  o
      /
     /DO     o
/
                          o
                     D
       0      0.2
  0.4     0.6     0.8


   WATER/FUEL RATIO
                        o
  o
   o
1.0     1.2
Figure  3.1-5.   Effectiveness of water/steam injection

             in reducing NOX emissions.1°
                         3.1-11

-------
                                                               29
corresponding to about a 5 percent increase in fuel  consumption.
     This control technique is considered to be  the best system of emission
reduction to  achieve  NSP  standards  for  NO   on turbines with  capacities
                                             J\.
greater than  10.7  gigajoules per  hour of heat input.  This is  equivalent to
                          o
approximately 0.75 MW (10  hp) output capacity.  Manufacturers estimate that
                                                                    30
3 years will be required to incorporate wet controls  for  smaller units.
     Dry control techniques--Dry  control  techniques consist of  operational or
design modifications  that  govern  the conditions of  combustion to reduce NO
                                                                           2\.
formation.   In full-scale  turbine applications, some  dry methods  have reduced
                                       31
NO  emissions by more than 40 percent.
   A.
     Although  dry NO   control techniques have  not been adequately demon-
                      .K-
strated  on  full-scale  turbines,   research   and   development   efforts  with
combustor rigs  (models)  indicate  NO  reductions  of up to 94 percent may be
                                    .X.
possible  by  various  combustion modifications.   These methods have  not been
incorporated in  production turbines.
     Table 3.1-4 indicates the degree  of NO  reduction achieved with various
                                           X.
dry  methods  on  combustor  rigs.   Some  of  these methods  have reduced
emissions on  full-scale  engines  by  about  40 percent.  These data represent
the potential of  efficient dry  control methods which may not be  developed  for
                              30
production turbines until 1982.
     Combined wet and dry  control methods—Emission tests  showed  the best
performance  of  combined wet  and dry  controls  with  a peaking gas turbine
using  distillate fuel.   Uncontrolled NOV was measured at 173 ppm at 15 percent
                                     yC
O-.   By combustion  modification  (lean  primary  zone), emissions  were reduced
to  82 ppm  (53  percent reduction).   By  water  injection, emissions  w^re
reduced  to  34  ppm  (80  percent  reduction).   By  a  combination  of  lean
combustion  in  the primary zone  and water  injection, NO   emissions were
                                                         32
reduced  to 16 ppm, corresponding  to 91 percent reduction.
     Catalytic control methods--Catalytic  exhaust   control   consists of NOX
reduction by ammonia in  the presence  of a catalyst.
                              3.1-12

-------
Laboratory tests  have demonstrated reductions  of up to  98  per-
cent.     Because  this  technique  has  not  been  demonstrated  on
                    30
full-scale turbines,
method.
it cannot be considered  as  a  LAER  control
      TABLE 3.1-4.  N0x EMISSION REDUCTIONS BY DRY CONTROL
           TECHNIQUES ON EXPERIMENTAL COMBUSTOR RIGS32
Technique
Lean burn, fuel-air mixing in full-
size combustor rig
Lean primary in half-size combustor
rig
Exhaust gas recirculation, half-
size combustor rig
Rig tests; premix, prevaporization,
staged combustion, lean burn
Lean primary, reduced residence
time in full-size combustor rig
Lean burn, premix, staged fuel and
air; full-size rig
Vortex air blast rig
NO reduction, %
X
12-44
15-20
30-38
35-61
40
51-60
94
3.1.3.2  S02 Control—
     Because SO  emissions from gas turbines are strictly a func-
tion of  the fuel  sulfur  content (and  essentially  all sulfur is
converted  to  SO2), the only technique  being  used  to control SO2
emissions  from gas turbines is use of low-sulfur fuels.  Flue gas
desulfurization  (stack gas  scrubbing)  systems  are economically
unattractive compared  to  the cost of  low-sulfur fuel because of
the  large  gas  volumes  to  be treated  and  the low SO2 concentra-
tions .
     Sulfur  content of distillate  fuels   used  in  gas  turbines
commonly ranges  from 0.01 to  0.48  percent by  weight; sulfur in
                              3.1-13

-------
crudes  ranges   from  0.06  to  3.0  percent,  and   in   residual  oils  from
                        17 1ft
0.5   to    3.2    percent.  '     Some   residual   oils  have   much   higher
sulfur  contents,  but   these  are   unusable  in   gas   turbines.   Sulfur
                                                      17  18
content of natural gas may range from 0 to 0.1 percent.  '
3.1.4  Emission Limits
3.1.4.1  NO  Limits--
           A.
     Although   a  few   state   and   local  control  agencies  apply   regu-
lations    specifically    to    stationary    gas    turbines,    most    states
apply  general   standards  to  gas   turbines.    The   most  stringent  SIP
limits   on   NO    emissions   applicable   to   gas   turbines   are    86   g
NO /GJ   (0.2   Ib   NO  /106   Btu)  input  for  gas-fired   burners  and   128
   X                  X  c                                   04 qc
g/GJ   (0.3    Ib    NO  /10    Btu)    for   oil-fired  burners.  '      These
                      X.
standards,   approximately   equivalent   to    50   and   75   ppmv,  respec-
tively,   are   applied   in   eight   states.    Apparently   the  most   strin-
gent   of   all   state   regulations    is   that   of   San   Diego   County,
California,  which   limits  emissions  to  75   ppmv   and   42  ppmv  at   15
percent   oxygen  when   burning   liquid   and   gaseous   fuels,  respec-
tively.36
     The   New   Source   Performance  Standards  (NSPS)   for   stationary
gas   turbines   are  applicable   to   turbines   whose   peak   load  is  equal
to  or  greater  than 10.7  GJ/h   (10  x  106   Btu/h)  of  heat  input.   The
emissions   limit  for  NOx  is  75  ppm  by   volume  at   15  percent   oxygen
and    International   Standard   Organization   (ISO)    ambient    atmospheric
conditions.     The   standard  also   includes  an   adjustment   factor   for
gas   turbine   efficiency   and   a   fuel-bound  nitrogen   allowance.
emissions  would be limited  according  to the following equation:37
                                                                         NO
                              3.1-14

-------
     STD = (0.0075 E)  + F
     STD = allowable NO  emissions (percent by volume at
           15 percent oxygen)
       E = efficiency adjustment factor
         = 14.4 kJ/Wh                        (13,600 Btu/kWh)
           Actual ISO                        Actual ISO heat
           heat rate                               rate
       F = Fuel-bound nitrogen allowance
           Fuel-bound nitrogen                     F
           (percent by weight)          (NO ,  percent by volume)
                                           X
                   N <_ 0.015               0
           0.015 < N <^ 0.1              0.04 (N)
           0.1   < N <_ 0.25             0.004 + 0.0067 (N-0.1)
                   N > 0.25             0.005

The  NSPS  would  allow an  additional  50 ppm NOx  attributable to
fuel N0x.
     The most  stringent  of all NO  regulations applicable to gas
                                  X
turbines appears  to be the 42 ppm limit of  San Diego County for
gaseous fuels.
     The lowest  NO   emission  rate achieved in practice, based on
available  data,  is  13 ppm  [21 g NO /GJ  (0.05  Ib  NO /106 Btu)
                                                           O fi
input]  by  an 88  percent reduction using  water injection.     An
NO  concentration of 16 ppm has been achieved by combined wet and
  X
dry control techniques with an efficiency of 91 percent.

3.1.4.2  S02 Limits--
     State  implementation plans  limit the  amount of  sulfur in
fuels  to  0.3 to  2.6  percent by  weight,  with  an  average  of 1.0
        38
percent.     The  most  stringent  0.3  percent limit corresponds to
about  130 g S02/GJ  (0.3  Ib S02/106 Btu) input.
     New  Source  Performance  Standards require  SO-  reduction to
150  ppm by volume,  corrected to  15  percent  oxygen,  or a maximum
                              3.1-15

-------
                                               39
fuel  sulfur content of 0.8  percent by  weight.     The  150  ppm limit corre-
sponds to about 346 g SO2/GJ (0.8 lb/106 Btu) input.
     The  most  stringent  current  SO^ regulation applicable  to  gas turbines
appears to  be an SIP limit  of 0.3  percent  sulfur  content  of  the fuel, corre-
sponding  to 130 g  SO2/GJ  (0.3  Ib  of  SO2/106  Btu) input or about 56 ppm by
volume.

3.1.4.3 Most  Stringent Limits--
     The  most stringent NO  limit is  the  San  Diego 42  ppm limit,  and the
most stringent SO2 limit  is  the  SIP and NSPS maximum  of 0.3 weight percent
sulfur  content in liquid fuels and essentially zero for  natural gas.

3.1.5  Determination of LAER

     The  recommended  limitations  are based  on  SIP's  and  on performance
information  available in early 1979.  It is  anticipated that several additional
SIP regulations covering these sources will be promulgated and/or modified  in
1979  and 1980  and  that appreciable new performance data will become available
in the near term.   Conceivably,  some  SIP  regulations may be more  stringent
than  the   LAER suggested  herein.   Furthermore,  performance  testing  may
show that more  stringent  limits  than those suggested are feasible or it may
show that  the suggested  limits are  appropriate or that they are not achievable
for some specific subcategories.   In any case,  the basis  for determining LAER
for many  source categories  is expected to  change frequently.  Since LAER is
near the  vanguard of control technology,  a more detailed analysis is partic-
ularly  necessary when addressing  modified or  reconstructed  facilities  subject
to the provisions  of  Section  173 of  the  Clean  Air  Act.  Emission limitations
reasonable for new sources may in  some instances  be  economically or  tech-
nically unreasonable when  applied  to modified  or reconstructed sources of the
same type.
                              3.1-16

-------
3.1.5.1  LAER for NO  —
                    .X.

     The  recommended LAER guidelines for NO  emissions  from  stationary gas
                                              .A.
turbines  are  based on the  lower NO   emission  concentrations  achieved in
                                      X
practice  by wet control  techniques.   In addition  a turbine efficiency  factor

and  an  allowance  for  the  fuel nitrogen content is  incorporated  into the LAER

guideline.
     The  controlled NO  concentrations  achieved  by wet techniques  are indi-
                       X
cated in  Figure 3.1-6.   The  broken  line indicates  the baseline  LAER at 50

ppmv NO  at 15 percent oxygen.   Since  NO   emissions increase with turbine
         X                                 X
efficiency, a turbine efficiency factor  is included  so  as not to penalize  energy

efficient operations.   This  efficiency  factor  was  developed under  the NSPS

studies.    It is  computed as follows,


     „„.  .      ,.  .    p              14.4 kJ/watt-h
     Efficiency  factor, E  =   ^	r—T~-	=j	rTLnA .—.
                             lower heating value  (.LHVJ neat
                             input per unit of power output

     The  fuel-bound nitrogen allowance is incorporated in the LAER since fuel

NO   is  not  effectively controlled by wet control  techniques.  Hence,  if a
   X
turbine firing low nitrogen  content fuel and  having a control efficiency of 90

percent was  required  to  switch  to a  fuel having  a higher nitrogen content,

the  NO   control  efficiency  would decrease.   The fuel-bound nitrogen  allow-
       X
ance  concept was developed under the NSPS  and  would equally apply for the
LAER guideline.  A  maximum increase of 50  ppmv  NO  above  the  50 ppmv
                                                       X
NO   LAER baseline is suggested.  This allowance  is  computed according to
   X
the following method:

              Fuel-Bound  Nitrogen      Allowance
               (percent by weight)          F

                       N   <                 0
              0.015 < N   <  0.1          0.04  (N)
              0.1  < N   <  0.25     0.004 + 0.0067 (N-0.1)
                       N   >  0.25         0.005
                             3.1-17

-------
                                       LEGEND

                                 o  WATER/FUEL RATIO
                                 •  NO WATER USED IN COMBUSTION
                                 X  WATER/FUEL RATIO UNKNOWN
J^U
300
280
260
240
c^ 220
»a
ir>
Z 200
(0
1 ^0
CO
§ 160
t— •
t/?
£ 140
X
0
z 120


100


80

60
40
20
n
GAS TURBINE
f








f

t
t
t
i
T


-1-
t

•+-
i"1.
To x'^
-•- To '

j- ! ^-3
rr^ i
o •
1 o 1
! i










«X>
' "1
O
1C
Vd
T^

t














r--
o-
^
















— •

—
—
—
—
—


—

—
t


—


• 	
CNJ •
O • T
To f I
I ~j

i !
! "i
O • II
' o i A
I LTI 0
i— 'd nr
" "5° \~*>[if> \ ~\i i"
o r^ 1 v 9
i !^ ! !
. Lrt....o^r^in' i — — ...
SIZE, MW °^---^"""g ^^5
FUEL TYPE [*- LIQUID FUEL (DISTILLATE) 	 4"~NATURAL GAS~H
                                                    LAER VALUE
               3 Tests made on model  combustor rigs rather than in the field
      Figure  3.1-6.   Summary of NOX emission  data  .„
   from gas turbines  using  wet control  techniques.
(Taken  from field or  engine tests  except where noted.)
                             3.1-18

-------
     Like  the  proposed NSPS, the measured NOx emission rate is corrected to
15  percent oxygen  and  ISO conditions.  ISO  conditions are defined as stan-
dard ambient  conditions  of  1  atmosphere  pressure,  60 percent relative  humid-
ity,  and 288°K.
     The  recommended  LAER guidelines  for NOv emissions are stated  as the
                                              .A.
following:
          NOX =  [0.0050 (E) + F]
   where: NO  =  NO  emission LAER at 15 percent O~
             XX                              "
            E =   Efficiency  factor, above
            F =   Fuel-nitrogen allowance, above

3.1.5.2  LAER for S02~-
     The  LAER  for  SO2 is recommended to  be  the  emission rate  associated
with firing  liquid fuels containing a maximum of 0.3 percent by weight  sulfur.
This is approximately equivalent  to 50 ppmv SO2 or 130 g/GJ (0.3  lb/10  Btu)
and  is  based on the most  stringent SIP regulation.   The LAER for gaseous
fuels would be essentially no SO2 emissions.
                              3.1-19

-------
                         REFERENCES
1.  Standard   Support  and   Environmental  Impact  Statement  Volume   I:
    Proposed   Standards  of  Performance  for  Stationary   Gas  Turbines.
    EPA-450/2-77-017a,  U.S.  Environmental  Protection Agency.  Office of Air
    Quality  Planning  and   Standards.   Research  Triangle  Park,  North
    Carolina.   September 1977.  p. 3-2.

2.  EPA-450/2-77-017a, p. 3-1.

3.  Analysis of the Cost and Benefits of Nitrogen Oxide Emission Control  in
    Proposed  Johnsonville  and  Gallatin  Gas  Turbines.   Tennessee  Valley
    Authority.  Muscle Shoals, Alabama.   November 1974.

4.  EPA-450/2-77-017a, p. 3-13.

5.  EPA-450/2-77-017a, p. 3-37.

6.  Scott  Research Laboratories, Inc.  Turbine  Exhaust Emissions Measured
    at  Facilities of  New  York  Power  Pool.   Prepared for   General  Applied
    Science  Laboratories.   Report  No.  SRL  1378-01-0374.   March  1974.   p.
    9-28.

7.  EPA-450/2-77-017a, p. 3-43.

8.  EPA-450/2-77-017a, p. 3-45.

9.  EPA-450/2-77-017a, Table 3.8. and pp.  3-46  to 3-49.

10.  Dilmore,   J.A.,   and  W.  Rohrer.    Nitric  Oxide   Formation  in  the
    Combustion of Fuels  Containing  Nitrogen  in a  Gas  Turbine Combustor.
    ASME 74-GT-37.   April 1974.  p.  3.

11.  General  Electric  Corporation.    Advanced  Combustion   Systems  for
     Stationary  Gas  Turbines  Proposal  in   response to  EPA,  RFP  Number
     DU-75-A182.  May 8, 1975.   p.  15.

12.   EPA-450/2-77-017a,  p.  3-73.
                             3.1-20

-------
13.   Dilmore and Rohrer (1974),  p.  2.

14.   General Electric  (1975), pp. 18-19.

15.   Johnson,  R.H.,  and F.C.  Wilhelm.   Control of our Turbine  Emissions in
     the World  Environment.  General Electric  Company.  1974.   p. 11.

16.   General Electric  (1974), p. 19.

17.   Personal  communication  from  E.W.  Zeltmann,  General  Electric,  to D.
     Walters, EPA, June 15,  1973.

18.   Burnes  Fuel  Oils.    Mineral   Industry   Survey.   Petroleum  Products
     Survey  Number  71.   U.S.  Department  of  Interior,   Bureau  of  Mines.
     1972.

19.   Johnson,   R.  and  C.  Wilkes.   Comments  of the  General Electric  Gas
     Turbine Products  Division on the  Impact of Fuel-Bound  Nitrogen  on the
     Formation  of Oxides of Nitrogen from Gas  Turbines.   January  28, 1974.
     p. 5.

20.  EPA-450/2-77-017a, Appendix C.

21.   EPA-450/2-77-017a, p. 4-7.

22.  EPA-450/2-77-017a,

23.  EPA-450/2-77-017a, Table 4-14 and  p.  4-27.

24.  EPA-450/2-77-017a, p. 4-24.

25.  Personal  communications between R.H. Gaylord,  Turbodyne Corporation,
     to D.R.  Goodwin,  EPA.  Section F.  December 19,  1975.

26.  EPA-450/2-77-017a, Appendix C, Table 45,
     Facility Code FA.  p.  C-69.

27.  EPA-450/2-77-017a, Appendix C, Table 47,
     Facility Code HAl.  p.  C-71.

28.  EPA-450/2-77-017a, Appendix C, Table 48,
     Facility Code HA2.  p.  C-72.

29.  EPA-450/2-77-017a, p. 4-36.

30.  Stationary  Gas Turbines.   Standards of  Performance  for  New  Stationary
     Sources.    Federal  Register,   Part III.   October 3,   1977.   p. 53,784.
                             3.1-21

-------
31.   EPA-450/2-77-017a, p. 4-96.
32.   EPA-450/2-77-017a, Appendix C.  Table 37.  Facility Code Z2.  p. C-57.
33.   EPA-450/2-77-017a, Table 4-11.   p.  4-91.
34   Dibelius   N.R.,   and R.J.  Ketterer.   Status  of  State  Air  Emissions
     Affecting   Gas    Turbines.     General   Electric.    ASME   Publication.
     73-Wa/GT-8.   1973.
35.   Duncan,  L.J.  Analysis  of Final  State Implementation Plans - Rules and
     Regulations.   APTD-1334.  pp.  69-65.
36.   San  Diego  Air Pollution  Control District.   Rules and Regulations.   Rules
     68, 50, 52,  and 62.   February  1972.
37.  EPA-450/2-77-017a, pp.  1-1 to 1-2.
38.  Duncan,  APTD-1334,  pp. 58-63.
39.  EPA-450/2-77-017a, p. 1-3.
40.  EPA-450/2-77-017a, p. 8-21.
                              3.1-22

-------
3.2  MAJOR SOURCE CATEGORY:  KRAFT PULP MILL LIME KILNS, POWER
     BOILERS (BARK AND COMBINATION) FOR PARTICULATE ONLY; RE-
     COVERY FURNACE FOR SULFUR DIOXIDE ONLY
3.2.1  Process Descriptions^
3.2.1.1  Kraft Pulp Mill—
     The Kraft chemical wood pulping process involves the extrac-
tion of  cellulose fibers,  or "wood pulp,"  from  the  wood by dis-
solving  and  removing the  lignin  that  binds the  cellulose fibers
together.  The pulp is suitable for making paper, paperboard, and
building materials.   More than 80  percent of the  chemical wood
pulp made  in the United States is produced by the Kraft method,
which  can  be used with almost any wood species,  requires a rela-
tively  short time  period to complete  delignification,  does not
degrade  the  valuable  cellulose and hemicelluloses in the wood as
badly  as other  chemical  pulping  processes, and  permits recovery
                                              2
of a high percentage of the cooking chemicals.
     Figure  3.2-1  is  a  flow  sheet of  typical  Kraft  pulp mill
operations  showing the  recovery  and  recycling  of  the valuable
sodium  salts.    Pulpwood  logs  are  debarked and  chipped  (not
shown),  and  the  chips  are  fed  into  a continuous  digester counter-
current  to  a  fresh chemical  stream  (called  white  liquor)  con-
taining  about 21  percent  active  chemicals, of which three  quar-
ters  is sodium  hydroxide and one quarter  is sodium sulfide  in
water  solution.   The digester  is  held  at  7.03  to 9.49 kg/cm2  (100
to  135 psig) and  170°  to  175°C  (338°  to 347°F).3   Time required
for  the  cooking  cycle  is  from  1 hour  for  unbleached  brown pulp  to
                                                            3
as much as 5 or  6  hours  for pulps  that are to be  bleached.   The
cooking  process  causes  formation  of  malodorous  sulfide  gases,
such as  hydrogen sulfide,  methyl  mercaptan,  and  dimethyl sulfide.
Venting  of these gases gives  a kraft  mill  its typical  sour odor.
     The contents  of  the digester exit  through  a "blow  tank,"
where  steam and  noncondensibles are flashed-off, and cooked chips
                                3.2-1

-------
                                                                                                  H2S, CH3SH, CH3SCH3,
                                                                                                  AND HIGHER COMPOUNDS
               CHIPS
                                             CH3SH, CH3SCH3, H2S
ac.
UJ
I—
tx>
U)
 I
NJ
RELIEF
T s-u.cu ru-cru-, Ho<;
HEAT
EXCHANGER
NONCONDENSABLES
,i t , ,
                                   ,
                                NONCONDENSABLES
                                                  SEPARATOR
                            BLOW
                            TANK







1/1
z
Ul
Q 1— .
z
o
o
H.W.
ACCUM.
                                                                     Y
                TURPENTINE


   CONTAMINATED WATER


STEAM, CONTAMINATED WATER,
                                                 CONTAMINATED    *H2S, AND CH3SH
                                                  • WATER
                                                 AIR
                         PULP
                       SPENT AIR,
                          AND CH3SSCH3
                                            OXIDATION
                                             TOWER
ON
!

1
i






ac.
o
I—
o
D-
UJ
t
                                                 BLACK LIQUOR
                                                  50% SOLIDS,,
                                                                DIRECT CONTACT
                                                                 EVAPORATOR
                                                                                              WATER
                                             RECOVERY
                                             FURNACE
                                                                                       OXIDIZING
                                                                                            ZONE

                                                                                       REDUCTION
                                                                          CAUSTICIZER
\




t t
GREEN
LIQUOR
LSMELT ZONE]
NazS + Na2C(
                                                                                                                 IH™
                    Figure  3.2-1.   Typical  kraft sulfate pulping and  recovery  process.

-------
are sent  to  a filter that  separates  the pulp from the spent cooking  liquor,
now  called  "black  liquor."   The  pulp  passes  on for further  refining  and
possibly bleaching before it  is pressed, dried,  and sold as pulp or made  into
paper or  other products.
     Satisfactory  economics for the Kraft process  require efficient recovery of
sodium  and  sulfur values  from  the  black  liquor,  as depicted  on the  flow
sheet.  The  organic  sulfides, also called  "reduced  sulfur"  or  "mercaptans,"
are often oxidized as an  air  pollution  control  measure to render  them  less
volatile and  thus  diminish  loss  when  a direct  contact evaporator is used in
subsequent steps.  The  black liquor is then concentrated to 50  percent solids
in a multiple-effect evaporator and pumped to the recovery furnace.
     At the  recovery  furnace the black liquor  is concentrated  to 70 percent
combustible  solids, in the  case  of direct  contact evaporation,    by counter-
current   flow  against  hot  combustion  gases  from  the  furnace.    If  good
oxidation  is  obtained  upstream,  this  unit will  emit  only  small  quantities of
volatile reduced  sulfur compounds.
     The   black   liquor  concentrate  is  sprayed  into  the  recovery furnace,
where  the  carbon  from  the  wood   is  burned,  the remaining   water  is
evaporated,  and  the  sodium is  changed to molten sodium carbonate or sodium
sulfide.   These   molten  salts, or "smelt," are  redissolved  in water to  form
"green liquor," then are clarified and causticized  with  lime.
     The  calcium  carbonate  resulting  from causticizing is  filtered from the
"white  liquor" and is passed on  to  an oil- or  gas-fired  kiln.    Entering the
kiln  at  35   percent  moisture,   the   calcium  carbonate  is   dried  and   then
decomposes  at  about  1300°C (2370°F)  to  calcium oxide and carbon  dioxide.
The "white liquor" is recycled to the  digester.

3.2.1.2   Lime Kiln—
      The  lime kiln is essential  to the  system  of recycling  caustic soda for
reuse in  digestion.   It receives lime  "mud" or  calcium  carbonate  and burns  it
to  quicklime  (calcium  oxide), which,  after being slaked,  is allowed to react
with  sodium carbonate in the green liquor  to make the caustic soda.
                               3.2-3

-------
     Most  lime  kilns  used  by  pulp mills  are  of  the  rotary  type  and  are
constructed of refractory-lined  steel.   They  are  2.4 to 4  m (8  to  13 ft) in
diameter and  38  to  122 m (125  to  400 ft) long.4  They  are  inclined, and  the
lime  mud is  fed into the elevated end,  where it  contacts  the  counterflow of
combustion  gases from  an  oil- or  gas-fired  burner  located at  the opposite
end.   The  kiln  rotates slowly, at about 0.5  to 1.0  rpm.  As  it moves down
the  inclined  kiln,   the  mud dries,  agglomerates into pellets,  and  finally is
calcined into  calcium oxide  in the  hot  zone of the  kiln.  The fresh, hot  lime
pellets  are discharged at the  lower end  of the kiln and  usually are slaked
immediately.
     The lime kiln has several major limitations.  It is not efficient in the  use
of heat; the  burnt  lime product may vary widely in its subsequent, reactivity;
and  unless the  sodium content of  the  mud  is kept at 0.25  percent or higher,
large loose balls or  rings of lime adhere to the inner surface of the kiln  and
cause  stoppages.   Efforts have been made  to improve the energy efficiency,
the  reactivity of the lime product, and  the operability  of  the  unit by  flash-
drying the mud ahead  of  the  kiln  or by  calcining  it in a fluid-bed unit in
which  time-temperature relationships can be  controlled  closely.   Flash-dried
mud enters the  kiln at about 10 percent moisture.   Tests  run at a  kraft mill
in  Albany,  Oregon,  indicated  that feeding  "dry mud" doubled the  drying
capacity of the kiln,  produced a more  reactive lime product,  and reduced  lime
emission losses to  10 percent.7   Fluid-bed calciners are in limited  use,  but
their  production  rates  are  relatively  low,   about 20  to  140 Mg/day  (25  to 150
tons/  day).4  Another  development, the use  of oxygen  to boost  kiln output,
should decrease the amount of  dust carryover per  ton  of  burnt lime produc-
tion, because it is  claimed  that  lime production increases 25 percent in a  kiln
retrofitted for  oxygen  enrichment.  The fuel requirements per increased  ton
of  lime  output are  about half that  required  with  air alone.  Advantages  are
reduced fuel requirement and  reduced gas volume per  ton of  burnt lime  and
the  attendant reduction of emissions.
                               3.2-4

-------
3.2.1.3  Power Boilers  (Bark  and Combination)--
     Kraft pulp mills use about 5  Mg of steam per megagram of air-dried pulp
(ADP) (5 tons steam per ton of ADP).9  Before 1960 nearly all of this steam
was made at the mill in conventional oil-,  gas-,  or coal-fired boilers.    Wood
waste  materials,  including bark,  shavings,  sander dust,  log  yard  cleanup,
and  sawdust  (often called "hogged fuel"),  were  burned  in a tepee burner,
which  gave no heat recovery,  had  no  combustion controls,  and emitted large
amounts  of  particulate.   By 1973,  88 of 273 mills had  at least one power boiler
that could burn wood bark.
     Bark is  difficult  to  burn.   Depending  on its  source, it may  contain  a
large  percentage of ash  and abrasive sand and as much as  50 percent water.
Bark  requires  a  higher temperature than  conventional fuels for  good combus-
tion;  if it is  charged wet, it must remain in  the combustion zone long enough
for the water and volatiles to boil  off before it ignites.
     Bark char is very  light,  about  1/6  the weight of water.   It occurs as
round,  flat  flakes, which  have  a large  surface  to  mass  ratio and readily
                                              12
become entrained with  the  combustion gases.     Recent  theoretical studies
indicate  that at normal flue  gas velocities, char of certain particle sizes does
                                                                 13
not burn  out in a conventional  furnace  and is always  carried over.

3.2.1.4  Recovery Furnace--
     Figure 3.2-1 shows  the  position of the recovery furnace in the chemicals
recovery  process.   Black liquor  of 40 percent to 55  percent solids content
from the  multiple-effect evaporators is  concentrated to as high  as  70 percent
solids  in a direct-contact evaporator  heated  by  recovery furnace  flue gas.
This  direct-contact  evaporator may be a  cyclonic or venturi-type liquid-gas
contactor, or  it  may  be a  cascade evaporator.  Depending on the pH and
temperature  of  the  black   liquor,  this  contactor  can  remove  appreciable
quantities of residual SO^ from the flue gas.
     The  black liquor  concentrate  is sprayed  into the  combustion zone of the
furnace,   where  the  organic materials  burn.   The chemicals,  chiefly sodium
salts,  melt and  are  accumulated  on   a  shallow hearth at  the  base of the
furnace.   The molten  salts,  or smelt,  are  subsequently dissolved in water to
                              3.2-5

-------
form green liquor,  which  is  causticized to white liquor and  recycled  to  the
digesters.
     The  recovery furnace  in  a  large,  modern kraft mill  can supply a major
fraction of the process steam  needed [as much as 5 Mg/Mg ADP  (10,000 Ib/ton
ADP)].   It  is  also  a major  source of  emissions,,  both  particulate  and  the
malodorous reduced  sulfur compounds.

3.2.2  Emissions
3.2.2.1  Lime Kiln Particulate Emissions--
     The  rolling  and tumbling action of lime mud in  a rotary kiln and  the
vaporization  of  sodium compounds in the high-temperature zone cause most of
the particulate emissions.   Lime dust is  made up of particles ranging from  1 to
100 urn in  diameter;  soda  fume is very  small and less  than 1  [m in diameter.
Thus,  lime  particulate  is  relatively easy  to  remove and  soda  fume is very
difficult  to remove.   Because  sodium is  held  at less than 0.5 percent of  the
feed  material,  soda fume  constitutes  a minor  percentage  of  the  particulate
generated,   but  its  small  particle  size  makes  it  a  major  contributor to
emissions.
     Stack  emission  data  for  66  controlled and  uncontrolled  lime  kilns were
reported  by 35 mills in 1973.16  The data have limited value since the methods
used to  determine emissions are  unknown.   Averages for the 10 lowest and 10
highest emitters  are given in Table 3.2-1.   Reference 1, the source of most of
the  data,  does  not  specify the type of fuel or the control equipment,  if any,
                               3.2-6

-------
                                  TABLE  3.2-1.   LIME  KILN  PARTICULATE  EMISSIONS
U)
•
to
I
-J

Lower 10 emitters1
Average
Range
Higher 10 emitters17
Average
Range
NSPS, Sept. 24, 1976
Gas fuel
Oil fuel
Emission factor
Untreated
Scrubber
Emission concentration.
g/m'1 , dry

0.08a
0.02-0.21a

2.70a
0.57-9.163

0.15
0.30



(gr/dscf )

(0.037)a
(0.01-0.09)3

(1.18)a
(0.25-4.0)3

(0.065)
(0.130)



Weight rate, of
emission ,
kg/Mg ADP

0.1
0.037-0.26

5.9
2.59-21.5




22.5
1 .5
Ratio of flue gas rate
to pulp production,
m!, dry/Mg ADP/day

0.91


1 .72







(scfm/ton ADP)

(29)


(55)







                   Explanation  of abbreviations:
                      g/m3,  dry is grams of particulate per dry standard cubic meter of  exhaust gas.
                      gr/dscf is grains of particulate per dry standard cubic foot of exhaust  gas.
                      kg/Mg  ADP is kilograms per  megagram of air-dried pulp.
                      m3 ,  dry/Mg ADP/day is dry standard cubic meters per minute per megagram  of air-dried pulp per day.

                   Method  of stack test unknown.

                   Equivalent to pounds per 1000  pounds ADP.

-------
used by each  of  the  reporting mill operators.   It can be assumed
that many of the large emitters were older mills using inadequate
control equipment.   These  data are now about 5  years  old;  prob-
ably the larger  emitters have now modified their kiln systems to
reduce emissions.  Table  3.2-1 also shows that  the  higher emit-
ters vent nearly twice  the flue gas per unit weight of air-dried
pulp produced.

3.2.2.2  Power Boilers (Bark and Combination),  Particulate
         Emissions--
     Emissions from  bark-fired  (hogged-fuel) boilers and boilers
burning combinations of bark plus oil, gas, or  coal are chiefly
ash  and  unburned wood particulate.  Emissions  also  include some
"tramp" sand  caught  in the bark  during logging  and  transport of
the  pulpwood.  This  sand has little visibility in the stack gas,
but  it does contribute weight.   Coal  used as  an auxiliary fuel
will  contribute  fly ash  the same  as if the  coal were burned
alone.  Neither  oil  nor wood contains appreciable ash.   Because
of  the low density  and  airfoil shape of wood  fuels,  it is easy
for  certain sizes of wood-carbon particulate,  perhaps incandes-
cent,  to be swept out the  stack.
     The type of auxiliary fuel used probably affects particulate
emissions:   tests at  one  mill showed that  the  combination of
natural gas and  bark gave  the lowest emissions, followed  in order
                 18
by oil and  coal.
     Review of  the  recent  literature on hog-fueled boilers indi-
cates  that  although  no single design optimizes emissions, certain
design and operational  practices can  help  reduce  them.  These
include predrying of wood fuel,  control  of  fly ash reinjection,
and  proper adjustment of  overfire and underfire  air,  which are
discussed more fully in the Controls section.
     Uncontrolled  emissions from  power boilers  range  from 25 to
37.5 g/kg of  bark fired (50 to  75  Ib/ton).
                               3.2-8

-------
3.2.2.3  Recovery Furnace, S02 Emissions—
     As shown in Figure  3.2-1,  flue gases from the recovery fur-
nace in a  typical  kraft  pulp mill go to a direct-contact evapor-
ator and the combined vapors then go to an electrostatic precipi-
tator prior  to  discharge.  To date,  operators  have shown little
concern  about S02  emissions  from  the  recovery  furnace.   Even
though concentrations of  S02  from the recovery furnace may be as
high as  700 to 800  ppm,  the contact  evaporator usually reduces
this to 50  to 100  ppm before the  gas  reaches  the stack.  Atten-
tion has been focused on emissions of total reduced sulfur (TRS)
and  of particulate,  which  apparently  are  considered  much  more
serious.   Most  of  the literature  on control of  recovery boiler
effluents deals with these two kinds of emissions.
     Emissions of  S0« from recovery furnaces  are  a  function of
operating parameters, as discussed in the Controls section.

3.2.3  Control Measures
3.2.3.1  Lime Kiln Particulate—
     Wet scrubbers,  usually  venturi or impingement-type, are the
most common devices used in  the  kraft pulp  industry to control
                                 19
lime kiln  particulate emissions.     The venturi  scrubber is more
efficient  (97 to  99%),  but  it  requires a higher pressure drop
[2.5  to 7.5  kPa  (10 to  30  in.  H,0)]  than the  impingement-type
                                       20
scrubber [1 to 2 kPa  (4 to 8 in. H20)].
     Three  process  changes  (described briefly at  page 3.2-4) are
reported to be effective in reducing lime kiln emissions.   The
predrying  of lime mud before  calcination,  the  use of oxygen to
boost  lime  kiln  output  and  fluid-bed  calcination  reportedly
reduce  lime losses per  unit of throughput, provide  energy sav-
ings,  and  increase production.   None of the literature  indicates
that  all  of  these  steps have been  applied together in a single
operating  mill.  Such modifications  would be most feasible at new
installations  and  are worthy  of  consideration  in  devising an
overall control methodology  for a  new  lime  kiln  facility.
                                3.2-9

-------
     EPA Method  5  tests  on lime kilns have  shown  that emissions
can be controlled to 0.07 g/m3,  dry (0.03 gr/dscf)  on a gas-fired
unit21  and to  0.21  g/m3,  dry  (0.09 gr/dscf)  on  an oil-fired
unit.22   The  control devices  in  both  instances were  venturi
scrubbers with pressure  drops in  the range of 5 to  6  kPa (20 to
24 in. of water).

3.2.3.2  Power Boilers (Bark and Combination) Particulate
         Control--
     In  general,  current practice  is to control  bark-fired and
combination-fired boilers  by  the use  of dry mechanical cyclones
as the  primary  control  device,  followed by a scrubber, baghouse,
electrostatic precipitator or granular filter (dry scrubber) for
secondary  control.   Because the  use of  bark-burning  boilers is
relatively new  to  the paper  industry   and involves some unusual
                     12
combustion problems,    the variety  of secondary control devices
in use  is  not surprising.   In combination boilers, the auxiliary
fuel  and furnace design  also  affect selection and performance of
the secondary control device.
      In  addition to  the  add-on control devices, certain practices
in the  design of the bark furnace and in the preparation of bark
fuel  can reduce  the  loading in the exhaust stream:

      0    Design for temperature of  the  combustion zone to be
          maintained at  or above 1093°C  (2000°F) to  improve
          burning of the wood carbon.
      0    Reduce flue gas velocity,  and  hence the  entrainment of
          very  light wood particles, by  careful control of excess
           air.   Predrying the wood also  reduces particle  reen-
           trainment,  and it reduces  the  moisture content  of  the
           fuel.
      0     Control particle-size range  of freshly hogged wood fuel
           and of reinjected fly ash  so that  these  materials  will
           not pass  through the furnace without burning.   The
           quantity  of particles sized  below  0.3 mm should be
           limited.
                                3.2-10

-------
     As noted earlier,  the  furnace design must take into account
the control  devices  to be used.  For instance,  if  a baghouse is
proposed,   care   must  be  taken  to  ensure against  incandescent
particles  reaching  the  bags  and  damaging them.   Electrostatic
collectors  may   achieve  desired performance  only  when  coal  is
fired  with the  bark.   The  average  of  emissions  from  the  five
best-controlled  combination  bark-fuel boilers  (data for 26 units
reporting  in 1973)  was  3.9  g/Mg  (7.8  Ib/ton)  of  bark fired.
These values are equivalent to 0.17 g/GJ  (0.4 lb/10  Btu) or 0.4
g/m3,   dry  (0.18 gr/dscf).   As  indicated in Table  3.2-2, various
control technologies have reduced emission rates  markedly since
1973.   Table  3.2-2  also  illustrates the general  effect on emis-
sions when gas,   oil,  and coal are used as  auxiliary fuels.

3.2.3.3  Recovery Furnace, Sulfur Dioxide  Emissions Control--
     A  number  of variables  in recovery  furnace  operation can
affect SO,.,  emissions.   One  investigator has shown how control of
turbulence, secondary  air,  and spray-droplet  size can reduce SO,.,
                                            27
in  recovery furnace  flue gas  to near zero.    Another has found
that simply holding  excess oxygen to 3 percent reduces S0~ levels
           28
to  25  ppm.    One theoretical  study points out  that very small
changes in sulfur or sodium  emissions inside the furnace can give
                                                     29
rise to  very large  excursions  in  SO.,  concentration.    A series
of  tests  showed that SO, emissions from  a recovery furnace could
                          30
be  held at 50  to 100 ppm.    Concentrations of SO2 ranged from 5
to  100 ppm  in   combined  exhausts  from   the recovery furnace and
                                                31
contact  evaporator  at a  North Carolina  mill.     Measured SO2
emissions  in the  furnace  exhaust ranged  from 0 to 200 ppm  (55 ppm
                              32
average)  at an  Alabama mill.     A summary of  these literature
citations  indicates  that  the following operating conditions tend
to  minimize S02  emissions:

     0    Holding excess  oxygen at or above 3 percent
     0     Injection  of large spray droplets of black  liquor in
           the furnace
                               3.2-11

-------
TABLE 3.2-2.  PARTICULATE EMISSIONS FROM POWER BOILERS (BARK AND COMBINATION)
Fuel
Bark and coal
Bark and gas
Bark and oil
Bark and coal
Bark
Bark
Bark
Bark and oil
Bark
Bark and oil
Barkb
Controls
Cyclone, ESP
Multicyclones
and
wet scrubber
Fabric filter
Multicyclones/scrubber
Multicyclones/venturi
Cyclone/wet scrubber
Granular filter
(dry scrubber)
Granular filter
(dry scrubber)
Granular filter with
electrostatic mode
Average emissions,
g/m3 , dry
0.026
0.032
0.126
0.137
0.069
0.032
0.037
0.059
0.105
0.158
0.018
(gr/dscf )
(0.012)
(0.014)
(0.055)
(0.06)
(0.03)
(0.014)
(0.016)
(0.026)
(0.046)
(0.069)
(0.008)
Reference
(date)
23
(1978)
18
(1974)
24
(1978)
25
(1978)
26
(1978)
 60  to 65  percent coal,  35  to  40 percent bark.

 Full-sized demonstration unit operated on  a  portion  of  the  exhaust stream
 from a 100 percent bark-fired boiler.

-------
     0    A high percentage of solids in the liquor
     0    A ratio of sulfur to sodium below 0.5:
     0    High turbulence at the secondary air inlet ports
     Recent information on factors that affect SC>2 emissions from
kraft recovery  furnaces  indicate  that sulfidity plays a dominant
role.33'34  Lower  sulfidity levels  (20  to 24 percent)  have ad-
verse  effects  on  pulp  quality  and  increase safety  problems;
medium  levels  (25  to  30  percent)  allow  these  problems  to  be
averted,  and  sulfidity  levels  in  excess  of  31  percent sharply
increase  S09  emissions.  This  information indicates that optimum
           ^
process  control—limiting  sulfidity  levels  to  30 percent and
keeping furnace  operating  conditions close to optimum—will keep
the  24-hour  average emission rate below 250  ppm  and the monthly
average rate to about 100 ppm.
     The  possibility  of using  flue  gas desulfurization (FGD)  to
                                                               34
reduce SO9 emission levels from recovery furnaces was examined.
          ^
It  was  concluded  that  available  FGD systems  have successfully
controlled  sources  having  gas  volumes  and  S02  concentrations
comparable to  those of kraft recovery furnaces; FGD is therefore
considered  "available technology."   Although  no  detailed cost
analysis  was  made,  capital  cost  was  estimated  to  be   about
$5000/ton per  day of furnace capacity, and annual cost was  esti-
mated to  be $1600/ton per day of furnace capacity.
      In  summary, recovery  furnace  operating  conditions  can in-
fluence  SO2  emissions,  and  optimization of operating conditions
will  curtail  these  emissions.   The  sulfidity  level plays a  major
role:   the  higher  the  sulfidity,  the greater the SO2 emissions.
The  type  of  pulp  or paper  manufactured  affects  the  sulfidity
level  in the process.   For lower-sulfidity pulp, attentive con-
trol  of  sulfidity  levels   (optimum  process  control)  results  in
minimum S02 emissions.   Flue gas desulfurization has been used to
control sources  comparable  to recovery furnaces in  gas volume and
                                3.2-13

-------
concentration;  therefore  EPA  considers  FGD  to  be  "available
technology" for  the  control of S09 emissions  from recovery fur-
                                  ^
naces.  On recovery  furnaces  operating at high sulfidity levels,
FGD offers a means of reducing S02 emissions.

3.2.4  Emission Limits
     This section summarizes  emission  limitations  categorized by
state  implementation plans  (SIP),  new source  performance stan-
dards  (NSPS),  and   achieved-in-practice  (AIP)  levels  for  lime
kilns, power boilers, and recovery furnaces.

3.2.4.1  Lime Kiln,  Particulate Emissions—
     The NSPS  for lime  kilns in Kraft pulp mills are set at 0.15
and  0.30  g/m3, dry  (0.067  and 0.134  gr/dscf)  for gas- and oil-
fired units, respectively.  These are approximately equivalent to
0.27 and 0.59 kg/Mg ADP  (0.54 and 1.08 Ib/ton ADP).
     In  six states  having major  Kraft pulp  producers,  the SIP
limit  for  particulate emissions from  lime  kilns  is the same and
is the most stringent of all the SIP  limits.   This limit is 0.5
kg/Mg  ADP  (1.0 Ib/ton ADP),  equivalent to  0.28 g/m ,  dry  (0.122
gr/dscf),  with  distinction  regarding  the type  of  fuel  fired.
Thus,  the  SIP limit  for  oil-fired  kilns  is slightly more  strin-
gent than the NSPS.
     The  lowest AIP  emission levels are  well  below SIP and NSPS
limits.  Emissions from  gas-fired lime kilns have been  controlled
to   0.07  g/m3,   dry   (0.03  gr/dscf).   Oil-fired kilns  have been
controlled  to  0.21 g/m3,  dry  (0.09  gr/dscf).

3.2.4.2  Power Boilers,  Particulate—
     No NSPS  have been promulgated  for bark-fired boilers  serving
kraft  pulp  mills.
      The most stringent  SIP limitation is that of  Florida:   0.047
kg/GJ  input (0.1 lb/106  Btu input), which is roughly approximated
as 0.09 g/m3,  dry (0.04  gr/dscf).
                                3.2-14

-------
     The BACT  level proposed by  EPA Region X  is  0.09 g/m ,  dry
(0.04 gr/dscf).   This BACT proposal is based on a fabric filter
installation on  a hogged-fuel power boiler  for which tests have
                             ns
                             22
                                            o
indicated particulate emissions of 0.069 g/m ,  dry (0.03 gr/dscf)
with 100 percent bark-firing.
     The AIP  levels  vary with the type of  fuel  or fuel combina-
tions  fired  and the  control  device  applied.   Information suffi-
cient  for  full  definition  of the  relationship  and  impact  of
varying  fuels and  fuel  combinations  on  emission levels  is not
available.   Table 3.2-3  reflects  the better AIP levels that have
been reported.  These AIP levels appear to be significantly lower
than SIP or NSPS limitations.

3.2.4.3  Recovery Furnace, SO_ Emissions—
                             ^
     No  NSPS  are applicable to SO2  emissions  from recovery fur-
naces.
     Although SIP's  do not  specifically limit S02 emissions from
recovery furnaces,  they  do contain  regulations  covering all new
sources.  The most  stringent  SIP  limits are 300 ppm SO2, typical
of California and Oregon regulations.
     The BACT level  proposed by  EPA's Region VI  is  250 ppm SO
                                                                £*
for the paper mill at Morrilton, Arkansas, including the recovery
furnace.
     The lowest  AIP  levels  from  recovery  furnaces,  achieved  by
optimizing process conditions,  have  been  0  to 25 ppm SO,,; excur-
sions  to 100 to  200 ppm are  common,  however, even  under these
optimum conditions.
     The AIP  levels  are significantly  lower than  SIP  or  BACT
limits.  It is not known whether SO2  emissions from most recovery
furnaces can  be  reduced to the  lowest SIP levels.   AIP levels
cannot be  compared  directly with  SIP  or BACT  limits,  since the
latter include all emissions from the paper mills.
                               3.2-15

-------
3.2.5  Determination of  Lowest Achievable Emission Rate (LAER)
     The  recommended  limitations  are  based  on SIP's  and  on performance
information  available in early 1979.  It is anticipated  that  several  additional
SIP  regulations covering these sources will be  promulgated and/or modified in
1979  and 1980 and  that  appreciable new performance data will become available
in the near  term.   Conceivably,  some  SIP regulations  may be  more stringent
than  the  LAER suggested  herein.   Furthermore,  performance testing may
show that more  stringent  limits  than  those  suggested are feasible  or it may
show that the  suggested limits are  appropriate  or that they are  not achievable
for some  specific subcategories.  In any case,  the basis for  determining LAER
for many source categories is expected to change frequently.   Since LAER is
near the vanguard of control technology,  a more detailed analysis  is  partic-
ularly necessary when  addressing  modified  or reconstructed facilities  subject
to the  provisions  of Section 173  of  the Clean Air  Act.  Emission  limitations
reasonable  for  new  sources may  in  some  instances  be  economically or tech-
nically  unreasonable when  applied  to modified  or reconstructed  sources of the
same type.
      The LAER  levels  recommended  for each  process  pollutant are based on
AIP  emissions.  These recommendations are summarized in Table 3.2-3.

   TABLE 3.2-3.  RECOMMENDED LAER  FOR  KRAFT PULP MILL  PROCESSES
      Process
 Pollutant
                                         Recommended  LAER
 Lime Kilns
  Gas-fired
  Oil-fired

 Power  Boilers
  Auxiliary  gas-fired
  Auxiliary  oil-fired
  Auxiliary  coal-fired
  100%  bark-fired

 Recovery Furnace
Particulate
Particulate
Sulfur
dioxide
             0.07  g/nu,  dry  (0.03 gr/dscf)
             0.21  g/m ,  dry  (0.09 gr/dscf)
0.032 g/m.,  dry (0.014  gr/dscf)
0.059 g/rru,  dry (0.026  gr/dscf)
0.026 g/nu,  dry (0.012  gr/dscf)
0.032 g/m ,  dry (0.014  gr/dscf)

50 ppmv daily average
100 ppmv - 3 hour maximum
 3.2.5.1  Lime Kiln,  Particulate--
      The  recommended LAER values are based  on achieved-in-practice levels.
 The  suggested value for gas-fired  kilns is  0.07  g/m3, dry (0.03 gr/dscf) and
                                3.2-16

-------
for oil-fired units,  0.21  g/m3, dry (0.09 gr/dscf).   The  literature  reports
process  changes  such  as  predrying  of  lime  muds,  use  of  oxygen,  and
fluidized-bed  calcination  could  provide  benefits  in the form of energy  conser-
vation,  production increase, and  emission reduction.  Such process changes
should  be  considered  along with  a high-pressure-drop  venturi  or  another
equally effective  device  when  devising an overall control strategy to attain
the lowest achievable emission rate.

3.2.5.2  Power Boilers  (Bark and Combination), Particulate--
     The  fuel or  fuel  combination  fired affects  the quantity  of  paniculate
emissions  from power boilers,  the type of control used,  and the performance
of that control.   Although  available stack test  data  are  insufficient to define
this relationship  fully,  the reported  AIP levels provide a basis  for LAER
determinations.   The suggested LAER  values  are  as  follows:  for  100  percent
bark  fired boilers,  0.032  g/m3,  dry  (0.014  gr/dscf);   for  gas/bark-fired,
0.032 g/m3, dry  (0.014  gr/dscf);  for  oil/bark-fired, 0.059 g/m3,  dry (0.026
gr/dscf);   and  for  coal/bark-fired,   0.026   g/m3,  dry  (0.012   gr/dscf).
Depending  on fuel composition  and fuel mix,  a LAER  value more or less strin-
gent than that suggested may be appropriate on a case by case  basis.

3.2.5.3  Recovery Furnace, S02"
     It is  recommended  that LAER  be  set at a maximum  value  of  50  ppm  by
volume on  a  daily average  basis, with excursions allowed to  100  ppm over a
period  of  no  more than  3 hours  in  any 24-hour period.  Pulp or paper manu-
facturing  of  the low-sulfidity  type can meet  this  or a  lower limitation  by
attentive  control  of  process variables.  Operation at high sulfidity levels may
require  the  application  of  FGD  technology   in  addition  to optimum  process
control to meet the LAER limitation.
                               3.2-17

-------
                           REFERENCES


1.   Atmospheric Emissions from the Pulp and Paper Manufacturing
     Industry.   EPA-450/1-73-002,  U.S.  Environmental Protection
     Agency,  Research Triangle Park,  North Carolina.  September
     1973.  p.  5.

2.   Casey, J.P.  Pulp and Paper,  Chemistry and Chemical Tech-
     nology;  Volume I:  Pulping and Papermaking.  Interscience
     Publishers, Inc., New York.  1952.  pp. 133-167.

3.   Casey (1952).  p. 145.

4.   EPA-450/1-73-002.  p. 53.

5.   Casey (1952).  pp. 172-3.

6.   Wenzl, H.F.J.  Kraft Pulping, Theory and Practice.  Lockwood
     Publishing Co.,  Inc., New York.  1967.  pp. 124-126.

7.   Wenzl (1967), p. 125.

8.   Oxygen Boosts Kiln Output.  TAPPI,60(11).  November 1977.
     pp.  37-8.

9.   Shreve, R.N.  The Chemical Process Industries, McGraw-Hill
     Book Company, Inc., New York.   1945.  p. 705.

10.  Effenberger, H.K., et al.  Control of Hogged-Fuel Boiler
     Emissions,  A Case History.  TAPPI,56(2):   111-115.  February
     1973.  p.  111.

11.  EPA-450/1-73-002.  pp. 77-79.

12   Barren, A.J.  Studies on  the  Collection of Bark Char
     Throughout the  Industry.   TAPPI, 53(8 ) -.1441-1448.  August
     1970.  p.  1441.

13   Adams, T.N.   Particle Burnout in  Hog  Fuel  Boiler Furnace
     Environments.   TAPPI,60(2):123-125.   February 1977.   p.  135.

14.  EPA-450/1-73-002.  p. 21.
                                3.2-18

-------
15.  Compilation of Air Pollutant Emission Factors, 2nd Edition,
     AP-42, Part A.  U.S. Environmental Protection Agency, Re-
     search Triangle Park, North Carolina.  May 1974.  pp. 10,
     1-5.

16.  EPA-450/1-73-002.  p. 55.

17.  EPA-450/1-73-002.  p. 92-93.

18.  Kutyna, A.G.,  et al.  Combination Fuel-Boiler Particulate
     Emission Control Pilot Studies.  TAPPI,57(9):139-143.  Sep-
     tember 1974.  pp. 141-143.

19.  EPA-450/1-73-002.  p. 12.

20.  EPA-450/1-73-002.  p. 54.

21.  Air Pollution Emission Test.  Project No. 74-KPM-17.  U.S.
     Environmental Protection Agency.  Office of Air Quality
     Planning and Standards, Emission Measurements Branch, May
     1974.

22.  Air Pollution Emission Test.  Project No. 74-KPM-19.  U.S.
     Environmental Protection Agency.  Office of Air Quality
     Planning and Standards, Emission Measurements Branch,
     September 1974.

23.  Personal communication with Nick Burkholtz.  State of
     Virginia EPA Office, Richmond,  Virginia.  March 13,  1978.

24.  Personal communication with Larry Sims.   EPA Region X
     Office, Seattle, Washington.  March 13,  1978.

25.  Wood Residue-Fired Steam Generator Particulate Matter Con-
     trol Technology Assessment.  EPA 450/2-78-044, Office of Air
     Quality Planning and Standards, U.S.  Environmental Protec-
     tion Agency, Research Triangle Park,  North Carolina.
     October 1978.   p. 16.

26.  Combustion Power Company.  Stack test data on dry scrubber
     installations.  1346 Willow Road, Menlo  Park, California.
     January 1979.

27.  Thoen, G.N., et al.   Effect of Combustion Variables on the
     Release of Odorous Compounds from a Kraft Recovery Furnace.
     TAPPI, 51(8):329-333.  August 1968.

28.  Walther,  J.E., and H.R. Amberg.  A Positive Air Quality
     Control Program at a New Kraft Mill.   Air Pollution Control
     Association, 20(1):9-18.   January 1970.
                               3.2-19

-------
29.  Borg,  A.,  et al.   Inside a Kraft Recovery Furnace - Studies
     on the Origins of Sulfur and Sodium Emission.   TAPPI,
     57(1):126-128.  January 1974.

30.  Blue,  J.D.,  and W.F.  Llewellyn.   Operating Experience of a
     Recovery System for Odor Control.  TAPPI, 54(7):1143-1147.
     July 1971.

31.  Fluharty,  J.R.  Progress in Air and Water Pollution Abate-
     ment in a 65-Year Old Mill.  TAPPI, 58(5):83-85.   May 1975.

32.  Lange, H.B.  Jr.,  et al.  Emissions from a Kraft Recovery
     Boiler - the Effects of Operational Variables.  TAPPI,
     57(7):105-109.  July 1974.

33.  Improved Air Pollution Control for a Kraft Recovery Boiler:
     Modified Recovery Boiler No. 3.   EPA 650/2-74-071a, U.S.
     Environmental Protection Agency, Research Triangle Park,
     North Carolina.  August 1974.

34.  Internal EPA memo, E.J. Vincent to R.O. Pfaff, Region IV,
     regarding BACT for SO9 from Kraft mills.  June 8, 1977.
                               3.2-20

-------
3.3  MAJOR SOURCE CATEGORY:   ELECTRIC ARC FURNACES AT STEEL
     FOUNDRIES AND GRAY IRON  FOUNDRIES; PARTICULATE, ONLY:
     CHARGING (AND BACKCHARGING AT  STEEL FOUNDRIES) AND TAPPING.
3.3.1  Process Description
     Castings made of iron  or steel are solid solutions of iron and carbon to
which  various amounts of alloying elements have  been  added.   The  carbon
content of the  finished  casting  is  what distinguishes  iron  from  steel; irons
typically  contain  2 to nearly 4  percent carbon,  whereas steels  contain less
than 2 percent carbon.
    ' The  direct  arc  furnace is  widely  used  today in  both  iron  and steel
foundries.   The furnace  is  usually charged with solid scrap,  iron,  or steel,
although  molten  pig  iron from  a  blast furnace  or prereduced  iron  pellets
sometimes form part  of the charge.  An electric arc furnace  (EAF) is a cylin-
drical  refractory-lined vessel, above which  are  suspended three  carbon elec-
trodes that can  be lowered  through the furnace roof to  a  position  above  the
charged  materials.  These electrodes can  be retracted,  and  the  furnace roof
can be rotated aside  to permit charging.   Alloying agents and  slagging mate-
rials are  usually added through  doors  in  the side  of the furnace.  Once  the
charge is in place, the electrodes are lowered into  the furnace  and current is
switched  on.   Arcing of current from  the  electrodes into the scrap generates
heat to melt  the  scrap.
     When the charge  is  melted and the temperature  is adjusted,  the slagging
operation  is begun.  In slagging, the carbon and oxygen  contents are adjusted
to the values desired for the iron  or steel  product.   Molten  slag  is withdrawn
from  the  furnace  and  various  elements are added—carbon  to the iron and
possibly iron oxide to the steel.
     When the desired chemistry has been achieved, the electrodes are lifted
out and  the furnace  is  tilted as much as 45 degrees  towards the  ladle that
receives  the  furnace  charge.  The hot  steel or iron is  then poured from  the
ladle  into holding furnaces  or  directly  into  sand molds,  from  which, after
cooling,  the  rough  casting  is  taken  for  cleaning,  possibly annealing, and
other finishing operations.   Figure  3.3-1 shows  the process  flow  in a  typical
                               3.3-1

-------
                                                           2
iron  foundry  that uses  an  EAF as  the melting  vessel.    Except for  the

cupola, which  is not used in a steel  foundry, Figure 3.3-1 also represents the

process flow in a steel foundry.


3.3.2  Emissions

     Production  of steel or of gray  iron  for  castings  from an EAF is  a batch

process in which the following steps  generate significant emissions:


     1.   Charging   the   furnace.    Scrap   metal   is   the   major   mate-
          rial,    plus   additives   such    as   "carbon    raiser,"    ferro
          silicon   (for   gray   iron),   and   limestone   and   coke   (for
          steel).  The roof is open.

     2.   Meltdown    operations.     These     include    unscheduled    occur-
          rences   such  as   "cave-ins"  of  unmelted   material   from  the
          furnace   walls   into   the   molten   pool   and   "pulling   bot-
          tom,"   which  is   undesired   boring  into   the   furnace   bottom
          by   electrodes   lowered   too   far.    This   category   also  in-
          cludes     scheduled     operations     such     as     "backcharging"
          (usually   for   steel),   which  is   the   addition   of  a   large
          quantity   of   additional   scrap   to   the   molten   bath   imme-
          diately   after    the   first    charge   is    partially   melted.
          The    roof   is   closed   during   meltdown;   charging   doors,   or
          the  roof itself, is opened  during "backcharging."

      3.   Oxygen  lancing,   in  steelmaking   furnaces  only.     The   roof
          is closed.

      4.   Slagging,     refining,     and    "working    the    heat,"     during
          which  the   steel   or   iron  is   brought  to   the   proper   com-
          position  and   temperature.    Some   alloy    steels    require    a
          second   or   reducing   slag,   which   must   be   accomplished
          with  minimum   induction   of  air.    The  roof   is   closed;  the
          furnace is tilted to  remove slag.

      5.   Reestablishment   of    the   arc    after   an    interruption.    The
           roof is closed.

      6.   Tapping  of  the   metal   into   a  ladle.    The  roof   is   closed;
          the  furnace  is tilted.
                                3.3-2

-------
               METALLICS
U)

OJ
I
               FLUXES
                                                                                        FINISHING

                                                                                               DUST
                                                          PARTICIPATE
                                                     l\0 \ O) EMISSIONS
                                   METAL
                                  MELTING -L
                                                                   •A'(./i  EMISSIONS
 CASTING
SHAKEOUT
           COOLING AND
            CLEANING
                                                      CORE
                                                    MAKING
                   SAND
               PREPARATION
               Figure 3.3-1.   Iron foundry  process  flow and emission sources.

-------
     In the following discussion the distinction between iron and
steel foundries  is  made only when the EAF  emissions  differ sig-
nificantly.  During  melting and refining,  uncontrolled particu-
late emissions from  iron  furnaces  range  from 2 to 20 kg/Mg (4 to
40  Ib/ton)  of iron  charged and  average  7.0 kg/Mg  (14 Ib/ton).
The  emission  rates  from  steel  furnaces  are reported  to average
8.0 kg/Mg (16 Ib/ton) charged.   Emission factors for charging and
tapping  are  not  available.   Based  on  extrapolation  of limited
emission  test  data,  charging  and  tapping  emissions  together
account  for  10 percent of  the  total uncontrolled emissions (0.7
kg/Mg for  iron and 0.8 kg/Mg  for  steel  furnaces).   These values
assume  alloying  takes place in  the  ladle.   Charging and tapping
are  estimated  as  about 5 percent of  total  furnace emissions when
no  alloys  are  added to the ladle.3   It is  to be noted that emis-
sions may  vary considerably as a result of furnace type and age,
the  kinds  of scrap  processed,  the additives to the melt, and the
types of iron  or  steel products.
     Rotation  of the  furnace  roof  to  the side during charging
renders  ineffective all  of  the close-coupled evacuation systems
during  this period of  intense emissions.  Most of these  emissions
are due to oil on the  scrap,  sand embedded in the recycled cast-
ings, and  miscellaneous organic  materials and dirt.  Tests of the
effects  of cleanliness  and  quality  of scrap on intensity of emis-
sions  have shown that emissions doubled when dirty,  low-quality
scrap was  used.4  When a  5-ton-per-heat gray-iron EAF  was charged
with scrap wetted  to  1  weight  percent with oil, more than 9  kg
 (20 Ib) of soot  and dust was  emitted in  a  short time.   Emissions
during  charging  and tapping  of a  steel  furnace are perhaps  10
percent of the total.   In contrast,  emissions  during  charging  and
tapping of a gray iron EAF  can be  5 percent of  the total.  Emis-
 sions  during  tapping are usually  negligible  unless  alloying  in
                        2
 the ladle  is practiced.
     During  meltdown the charge is  rapidly brought  up to tempera-
 ture;  all remaining oil  and some  volatile nonferrous metals  such
                                3.3-4

-------
as lead,  zinc,  and magnesium are  expelled.  "Cave-ins"  and "pulling bottom"
can  exacerbate  these meltdown  emissions  beyond  the volume  capacity  of  the
ventilation system.   Since the  roof  covers the furnace  during  this period,
however,  adjustment of electrode  positioning and  of  power input should hold
the emissions to a level that is  controllable by  a well-designed ventilation and
particulate capture  system.   In  "backcharging,"   which  is common  practice
with steel EAF's, the  cold metal  scrap produces   a  violent eruption when it
hits  the  molten  pool, and the amount  of  pollutants generated at this  time is
probably  higher than at  any other time during the  heat.   Because charging
doors  must  be   opened,   or  the entire roof rotated  away,  capture of these
intense emissions is  incidental.
     Oxygen lancing,  used in steel furnaces,  helps remove carbon  and accel-
erates  the melting process.  Although  oxygen  generates  gaseous emissions of
carbon monoxide and carbon dioxide and increases  the potential for  particulate
emissions,  the  practice  of careful control  of  oxygen rate (to  maintain  the
correct chemistry of the  heat)  and  keeping the  furnace roof closed during
lancing should  minimize escape  of  particulates.  Near the end  of the  lancing
period, some iron oxidizes to a  highly visible,  fine red  fume  that is most
difficult  to capture.   During the  peak of oxygen lancing,  the  emission rate
has  been  measured  at  2  to  3  times that  occurring  during average furnace
operation.
     Slagging is  done once  per  heat  in  producing  gray  iron,  but may  be
required  twice  to make certain alloy steels.   This second  slagging  must be
done  with minimum  inleakage  of  air,  i.e.,  ventilation  must be stopped or
heavily restricted within  the  furnace.   During  this  period, fumes   escape
upward through the  electrode   holes and  into the foundry  building bay area.
Emissions   during the first slagging are  less  pronounced  (via  the electrode
holes).   They  are more  pronounced in furnaces with direct shell  evacuation
than on furnaces with  side draft hood  evacuation.   The operations of refining
and  working the heat,  which are combined with the second  slagging, are done
principally in  producing  steel  castings where standards for  soundness  and
strength of the casting are very high.   Because this  step involves only slight
temperature  changes  and  very little reaction, emissions are  of moderate  inten-
sity.   During the period  of no furnace ventilation, slagging presents  emission
problems.

                              3.3-5

-------
     At any  time  during the operation the arc may be lost, for such  reasons
as  shutoff  due to  a preset  "demand  limiter,"  mechanical failure,  or faulty
positioning  of  the electrodes.   If  the  arc can be  reestablished quickly, this
loss has slight  effect on emission level,  since the  roof is closed and the melt
is hot.   If  the bath must  cool, however, striking  the arc again will vaporize
metal and cause  a  surge  of  very  high particulate.   Minimizing arc  inter-
ruptions is  the best  means of preventing such emissions.
     Tapping iron and  steel is done with the power off,  the  electrodes lifted,
and the furnace tilted.  During the pouring of iron or steel,  very fine partic-
ulate escapes as sparks or  fume along  the flowing stream of hot metal.   These
emissions are  negligible in  gray iron production, but perhaps because of the
much higher temperatures  needed for steel  [200°C or (390°F)],  they can be
at  least as severe as those during charging.  Treatments such as deoXidation
or  special alloying are often done while the molten  metal is poured or is in the
ladle and can  cause  violent emissions of short duration.

 3.3.3.   Control Measures
      Control of air  pollution at an iron or steel foundry is  a function of the
 efficiency of two  operations:
      1    Capture   and    containment   of   particulate-laden    gases    as
           they are generated at the  furnace or at the ladle.
      2    Treatment   of    the  captured,   particulate-laden   gases   in   a
           control system  to remove the particulate material.
      Methods   and  equipment for  capture  of pollutants  are  described   in
  Reference  2.  Important  among  these methods  is  the  use  of  large  canopy
  hoods   which  are positioned over  the  open furnace and over an open  ladle
  during charging and during  tapping/alloying in the  ladle.  While the lid  is  on
  the furnace,  either a  roof or side-draft hood  on the furnace roof or a direct
  furnace evacuation system on the  furnace  roof is  used.   The  canopy hoods
  must  be located precisely  so as to maximize pollutant capture and minimize  the
  total  amount  of  air pulled into the hood with minimal interference to operation
  of the overhead  crane,  charging buckets,  and other equipment.   Some means
  of eliminating crosscurrents of air, such as a shroud enclosure around part of
  the furnace,  improves the efficiency of the  canopy hoods.
                                 3.3-6

-------
     Hoods must be designed to maximize pollutant capture with low air inleak-
age  and  without creating  significant operational  and  maintenance problems.
Direct shell  evacuation,  or "fourth-hole,"  systems  are effective and  simple,
allowing  minimal  inleakage.   Roof hoods  and direct shell evacuation are alter-
native controls,  although  roof  hoods are  no  longer being  installed.   These
systems  for pollutant  capture  have been available and  in  use  for  many years
in both iron and steelmaking.
     Side-draft  hoods mounted  on  the  furnace  roof are used  with gray  iron
                                  2
and  smaller steel foundry furnaces.    One side is open  to provide maintenance
access  to the  electrodes.   These  hoods  are  effective in collecting  fumes
escaping  via  the  electrode holes.    This  type  of hood does require  larger
exhaust  volumes than  do roof  hoods  or direct furnace evacuation systems, but
these  larger  volumes  reduce   temperature  at the  control device  and assure
complete combustion of carbon monoxide.
     Fabric filter  collectors (baghouses) are  regarded  as  the most  efficient
and  versatile device  for removal  of  particulates in the exhaust from EAF's.
Major control  system components consist of  an exhaust fan, a cooler to reduce
the  temperature  of the  hot gases  so that  they will not damage bag  fabrics,
and  a baghouse.   Filter  velocities are as low  as 1.1 cm/s  (2.1 ft/min) and as
                               t;
high as 1.3 cm/s (2.55 ft/min).   Wet  scrubbers  and electrostatic precipitators
appear  applicable  to  EAF  particulate  control,  but neither device has been
                               2
widely used by  U.S. foundries.
     A number  of  process  and/or  equipment changes have been proposed  that
                                                          o
would  reduce emissions  and  in some  cases save   energy.    These include a
closed charging  system,  preheating  of scrap to  drive  off  oil and  moisture, a
hooded  charge bucket, a  hooded tapping ladle,  degreasing  of scrap, and the
use  of  an enclosure  around the furnace.   Of  these,  only  the enclosure or
shroud  has been tried in the  United  States, and  this  installation  is on steel-
                                  c
making furnaces  of  60-ton capacity.

3.3.4  Emission  Limits
     1.   Standards of  Performance  for New  Stationary Sources
(NSPS),  Electric Arc  Furnaces in the  Steel  Industry; Federal Register,
                               3.3-7

-------
Tuesday, September 23,  1975.   These  standards apply  to steelmaking only,
and  not  to  foundries  using  electric arc furnaces.   The  following  pertinent
sections  of  this NSPS  for steelmaking provide a  reference point  for  emission
limits from foundries.

     "...no owner or operator....shall cause to be discharged...
     from an electric arc furnace any gases  which:
     (1)  Exit from a control device and contain  particulate
          matter in excess of 12 mg/dscm (0.0052  gr/dscf)
     (2)  Exit from a control device and exhibit  3 percent
          opacity or greater.
     (3)  Exit from a shop and, due solely to operations of any
          EAF's, exhibit greater than zero percent shop opacity
          except:
          (i)  Shop opacity greater than zero percent, but less
               than 20 percent, may occur during  charging peri-
                ods.
          (ii)  Shop opacity greater than zero percent, but less
                than 40 percent, may occur during  tapping peri-
                ods."
     Identical standards  were adopted by  California  South  Coast (12-3-1976)
 and California  Bay Area for  facilities modified or built after  October 21,  1974,
 again  for electric arc steelmaking and not for  steel  or iron foundries.
     Figure 3.3-2  illustrates  the  relationship of emission limits  on  EAF's as
 specified in  NSPS  and  in  the state regulations of  California,  Pennsylvania,
 and New Jersey.   For comparison purposes, it shows the achieved-in-practice
 (AIP)  emission levels  at six  steel  and  six  iron  foundries. '    The values
 represent particulate  concentration  (g/m3,  dry)  in the exhaust from the con-
 trol device.   In Figure 3.3-3 the  data  are  presented on the  basis of mass
 emission rate  (kg/h).   The  figures  indicate that most  AIP values, especially
 at production rates of 10 Mg/h, are  lower  than  either the SIP requirements or
 NSPS for steelmaking.
                                3.3-8

-------

0.048
Of\A C
.U4o
-o 0.044
CTM-iJ ^
* O-.
o<0.022
>-. CO
t- cz
z o
SoO.016
itA nJ
££0.014
-J Od
Z3 LU
O*^ n mo
M 2 0.012
. ^,
Sc 0.010
°-
0.008

0.006
0.004

0.002

	 1 	 1 	 T- 	 1 	 1
O
0 0458 SIP LIMITS FOR CALIF.. N.J., AND PA.
-
7

^_

_
o
A
NSPS; U.S. AND CALIFORNIA FOR STEELMAKING EAF'S "
0.012 (DOES NOT APPLY TO FOUNDRIES)

^ O
O
A O
-
A
A A
A O STEEL FOUNDRY AIP .
° A IRON FOUNDRY AIP

	 1 	 1 	 __!_ 	 1 	 1 	 — 	
                      10      15     20
                    HOT METAL POURED, Mg/h
    25
    30
   Figure 3.3-2.  Particulate concentration
             versus  production rate.
   15,
   10
 CO
 CO
 CJ
 I—
 ce.
                               1
                             O STEEL FOUNDRY AIP
                             A I RON FOUNDRY AIP
                                       PROPOSED LAER, EQUIVALENT TO
                                         0.05 kg/Mg METAL POURED
                 10     15      20
                HOT METAL POURED, Mg/h
25
30
Figure 3.3-3.  Particulate emissions by weight
              versus production  rate.

                           3.3-9

-------
     The lowest AIP  emission levels discovered in preparing this  guideline for
both iron and steel  foundries  were those reported in Reference 7.   All test
results  reported here were  based on EPA Method  5..   This reference reports
an  AIP value of 11.0 mg/m3,  dry (0.0048  gr/dscf)  for  a new iron  foundry.
At a second and newer foundry having  an EAF and control system design that
was  based on the experience of the first, the test results showed 8.9 mg/m  ,
dry  (0.0039  gr/dscf)  at the  control  device  outlet.  In  both  instances the
EAF's  were  equipped  with  side-draft  hoods  and hoods  above  the  pouring
spout  and  slag  door.   Reference 7 reports  the  lowest  achieved  values for
steel EAF  control systems to be 5.74 mg/m3, dry (0.0025  gr/dscf)  and 6.63
mg/m3,  dry  (0.0029  gr/dscf).   The capture system  for  both  steel  foundry
EAF's  was by direct  shell evacuation.  The control equipment at the iron and
steel foundry EAF's reported here was  a  fabric filter  device.

3.3.5  Determination of Lowest Achievable Emission Rate (LAER)
     The  recommended  limitations are  based  on  SIP's  and on  performance
information available  in  early  1979.  It  is anticipated that several  additional
 SIP regulations  covering these sources will be promulgated and/or modified in
 1979 and 1980 and that appreciable new performance data will become available
 in  the near term.  Conceivably,  some  SIP regulations may be more stringent
 than  the  LAER  suggested herein.    Furthermore,  performance  testing may
 show  that  more stringent  limits  than  those  suggested are feasible  or  it may
 show  that the suggested limits are appropriate or  that they are not achievable
 for some specific subcategories.   In  any case, the basis for determining LAER
 for many  source categories  is expected  to change frequently.  Since LAER  is
 near  the vanguard of control technology,  a  more detailed analysis is particu-
 larly  necessary when addressing modified or reconstructed facilities  subject  to
 the provisions  of Section 173  of the Clean Air Act.  Emission  limitations reason-
 able  for  new  sources  may in some instances  be economically  or technically
 unreasonable when  applied to modified  or reconstructed sources  of the same
 type.
       On the basis  of  the  reported AIP  levels,  the suggested  LAER ^s 9.2
 mg/m3,  dry  (0.004  gr/dscf)  for  iron  foundry  EAF's  and 6.9  mg/m ,  dry
                                3.3-10

-------
(0.003 gr/dscf) for steel foundry  EAF's.   These values represent the average
emission   limit  over  a  complete  furnace  cycle,  which  includes  charging,
melting,  slagging,  and  tapping.   Because  the volume  of  exhaust per ton of
molten metal varies  with the capture device used, the  suggested LAER limits
are in terms of control  device outlet concentration rather  than mass  emission
rate.
                               3.3-11

-------
                           REFERENCES
1.    McGannon,  H.E.   The Making,  Shaping and Treating of Steel,
     United States Steel, Ninth Edition.   1971.

2.    Georgieff, N.T., and F.L.  Bunyard.   An Investigation of the
     Best Systems of Emission Reduction for Electric Arc Furnaces
     in the Gray Iron Foundry Industry.   (Draft)  U.S. Environ-
     mental Protection Agency,  Research Triangle  Park, North
     Carolina.   October 1976.  Ch.  3,4,6.

3.    Air Pollution Control Techniques for Electric Arc Furnaces
     in the Iron and Steel Foundry Industry.  Guideline Series.
     EPA 450/2-78-024, U.S. Environmental Protection Agency,
     Office of Air Quality Planning and Standards.  June 1978.
     p. 2-16.

4.    Coulter, R.S.  Smoke, Dust,  Fumes Closely Controlled in
     Electric Furnaces.  The Iron Age, 173:107-110, January 14,
     1954.

5.    Georgieff, N.T.  Addendum to:  An Investigation of the Best
     Systems of Emission Reduction for Electric Arc Furnaces in
     the Gray Iron Foundry Industry to Include Electric Arc
     Furnaces in the Steel Foundry Industry.  (Draft) U.S. Envi-
     ronmental Protection Agency, Research Triangle Park, North
     Carolina.  1976.  pp. 3-6, 4-1 to 4-7, Cl-4.

6.   Attachment to trip  report, Lone Star Steel Company, Lone
     Star, Texas.  EPA Contract No. 68-01-4143.  Sept. 14, 1977.

7.   EPA 450/2-78-024, pp. A-l, A-ll.
                                3.3-12

-------
3.4  MAJOR SOURCE CATEGORY:  PETROLEUM REFINERIES CATALYTIC
     CRACKING UNIT AND LIQUID FUEL BURNING--SULFUR DIOXIDE
     EMISSIONS
3.4.1  Process Description
3.4.1.1  Catalytic Cracking—
     Catalytic cracking is  a  high-temperature,  low-pressure pro-
cess that converts  certain heavier portions of crude oil primar-
ily  into gases,  gasoline  blend  stocks,  and  distillate fuels.
Feedstocks to  catalytic cracking  units  are  gas oils from atmos-
pheric  or  vacuum crude oil  distillation units, thermal cracking
units,  lube  oil extraction and  dewaxing units, coking units, or
deasphalting units.   Catalytic  cracking units are normally oper-
ated to produce a maximum of gasoline blend stocks, but the units
are  very  flexible  and  operating  conditions   can  be  varied to
maximize other products.
     As  of  January 1,  1978,  catalytic  cracking  capacity in the
                                                                3
United  States as reported in the Oil & Gas Journal was  788,050 m
(4,956,682 bbl) per stream day.    Two types of  catalytic  cracking
processes  are in use  today:   (1)  the  fluid process, which  uses
powdered  catalyst,  and  (2) the  Houdry  and  Thermafor Catalytic
Cracking  (TCC) processes  (in limited use), which  use  a  pelletized
catalyst.   Of the  285 U.S.  petroleum  refineries,  approximately
124 contain  fluid catalytic cracking units,  17  contain  TCC  facil-
ities,  and 3  contain Houdriflow  units.
     A  fluid catalytic cracking unit  is composed of three basic
sections:   cracker,  regenerator,  and  fractionator.   As  shown in
Figure  3.4-1,  the  cracking  reactions  take place continuously in
the  cracking section,  with the  spent  catalyst  being continuously
regenerated   and  returned  to  the cracking section.   Both the
cracking and  regeneration sections operate on  the  fluidization
principle,  which makes possible a continuous flow of catalyst as
well  as hydrocarbon  feed.   Gas oil feedstock  is mixed  with the
hot  catalyst and introduced into  the  cracker.   Steam is  added at
the  base of the cracker  to  strip and  purge  the spent catalyst of

                               3.4-1

-------
                                                        FRACTIONATOR
u>
I
to
                                                                      GAS
                        CRACKER
STRIPPING
 STREAM
                 RECYCLE FEED
            FRESH EAS/OIL FEED.
                                                                       GASOLINE
                                                                  LIGHT CYCLE OIL
                                                                    RECYCLE FEED.
TOWER BOTTOMS _



PARTICULATE
CONTROL
DEVICE
^


CO BOILER

1

                                               GAS
                                               COOLER
                                 COMBUSTION AIR
          Figure  3.4-1.    Generalized  schematic  of  a fluid  catalytic cracking  unit.

-------
adsorbed hydrocarbons.   The hydrocarbon  products are  withdrawn
from the top of the cracker and sent to a fractionator for separ-
ation into product streams.
     The coke-laden spent catalyst  is  withdrawn from the base of
the  cracker  and  transferred  to the  regenerator.  A  controlled
amount of  air  is introduced into the  regenerator with  the cata-
lyst to burn the coke,  which reheats the catalyst.  The resulting
combustion  gases  are  channeled  through  a  series  of  cyclone
separators located  inside the regenerator  to remove  most of the
entrained  catalyst  fines.  The  remaining catalyst fines  can be
removed  from  the  gases by   electrostatic precipitators  or  a
third-stage  cyclone  separator  outside  the  regenerator.   The
regenerated  catalyst   is   withdrawn   from  the  bottom  of  the
regenerator vessel  and returned to  the cracker  to  complete the
cycle.
     The hot flue gases,  at about 538°C (1000°F) contain 5 to 10
percent  carbon  monoxide (CO),  which generally  is  burned in a CO
or waste heat boiler  to recover a considerable amount of energy.
The  CO boiler  is  located  either  upstream or  downstream  of the
electrostatic precipitator.   If the  CO boiler  is  downstream of
the  electrostatic  precipitator,  the flue  gas  must be  cooled to
less than  315°C  (600°F) in  a cooler before entering the electro-
static precipitator.   Gases leaving the CO boiler are discharged
to the atmosphere and can be a source of air pollution.
     The hydrocarbon  product  vapors leaving  the fluid catalytic
cracker  are  sent to a fractionator for  the  first separation of
products into gases, gasoline, and cycle oils.  These streams are
further  separated  in  the refinery  as  needed.   Typical operating
ranges and yields for  a  fluid  catalytic  cracker are as follows:
                              3.4-3

-------
Temperature,  °C (°F)                460° to 524°C   (860° to 975°F)
Pressure, kPa (psig)  by         68.9  to 172.4 kPa  (10 to 25 psig)
Catalyst/oil  ratio by wt.                          4:1 to 20:1
Gasoline yield, vol.  % of  feed                    35 to 50
Coke formation, wt. % of feed                      4 to 12
Dry gas formation, wt. %                           7 to 11
Conversion of feed to lighter
 products, vol. %                                 60 to 90
Coke content of spent catalyst,
 wt. %                                          0.25 to 2.3
Regenerator:
Temperature,  °C (°F)             566°  to  740°C   (1050° to 1300°F)
Pressure, kPa  (psig)          6.9 to 172.4 kPa  (1 to 25  psig)
Coke content of regenerated
 catalyst, wt. %                               0.10 to  0.50
     Hydrocarbon cracking  deposits  a   small  portion  of   the  feed  on
the   catalyst   in  the   form  of   coke.    Likewise,   some   sulfur   origi-
nally   present   in    the   feed   is  deposited   on   the   catalyst.    The
amount   of   sulfur   varies  with  the  type   of  feed,  rate   of   recycle,
steam   stripping  rate,   type   of   catalyst,   cracking   temperature,   and
other factors.
     The  amount  of  hydrocarbon  that  remains  on   the  spent  catalyst
as  it   leaves   the    cracker   is   important   for   safe   operation   of
catalytic   cracking     units.     Essentially   all    coke-forming   compounds
are  removed  in   the  regenerator   by   air   oxidation.    The  oxidation
reaction   is   highly  exothermic  and   results  in  a   temperature   in-
crease    in   the  regenerator.   Excessively  high  temperatures   in   the
regenerator    can be  detrimental  to   the   catalyst  and  also  can   cause
afterburning   downstream   of   the   regenerator,   which   may  lead   to
 severe   damage   of   cyclones   and   auxiliary  equipment.   Oxidation   of
 coke   is  accomplished   with   an   amount   of   air  that   is   insufficient
 for  complete  combustion   so   that  only a  portion   of  the  carbon  is
 oxidized  to   carbon   dioxide.     Usually  the   volume  ratio   of   carbon
 dioxide  to   carbon   monoxide   is  maintained  between  1  and  2.    The
 carbon   monoxide   in   the   flue   gas   from  the   regenerator   is   then
 burned   to   carbon  dioxide   to  recover   the   remaining   energy   for
 steam generation.  When  CO boilers are used to recover  heat in
                               3.4-4

-------
conjunction with  the catalytic cracking  unit,  supplemental fuel
is usually provided.  The type and amount of fuel used to supple-
ment  the carbon  monoxide  will  affect  the  type  and  amount  of
pollutants emitted.

3.4.1.2  Liquid Fuel Burning—
     When a  refinery process  requires  a  temperature higher than
that obtainable from the steam supply or any other available hot
stream,  direct-fired furnaces  are  used.   These furnaces or fired
heaters burn liquid  fuels, refinery gases, or commercially avail-
able  natural  gas.   The  liquid fuels  are  reduced  crude  oils,
Bunker  fuels,  vacuum tower bottoms,  No. 4 or No.  2 fuel oil,  or
components comprising  these  fuels.   Operators commonly  limit the
use  of the lighter  distillate fuel  oils  because of their market
value  and the need  for  special equipment for safe and  efficient
combustion.
     The  number   of  applications  for liquid  fuel  burning varies
among  individual  refineries,  depending on processing complexity.
Liquid  fuels  are used  typically  in the furnaces  for  crude oil
atmospheric   distillation,  vacuum   distillation,   visbreaking,
catalytic cracking,  coking,  thermal cracking, solvent  deasphalt-
ing,  hydrotreating,  catalytic reforming,  and asphalt  stripping.
Although refinery furnaces vary in  shape and form, with various
burner  and tube  arrangements,  they usually  consist  of two main
sections:  the radiant section and the convection  section.  Some
unique  problems  may be associated with  certain furnace designs
and  special  fuels such as  those of low molecular weight.

3.4.2   Emissions
3.4.2.1  Catalytic Cracker—
      Atmospheric   emissions  from  catalytic  cracking  operations
have been measured  by several investigators.  The primary  emis-
sions   from  catalytic  cracking  include  sulfur  oxides,  nitrogen
                               3.4-5

-------
oxides,   carbon  monoxide,   carbon   dioxide,   oxygen,   water,   nitrogen,
hydrocarbons,  ammonia, cyanides, and participates.
     Concentrations of sulfur oxides in  regenerator flue gases normally range
from 150 to 3500 ppm.  The concentration of sulfur oxides  is a function of the
amount  of  sulfur in the coke present on the catalyst and the  amount of air
used  for  regeneration.   Because the amount  of sulfur present on  the  spent
catalyst in  the regenerator is a  function of the  sulfur level in  the  crude oils
and the  processing of the feedstock before it reaches the cracker,  the  levels
of  sulfur oxides emitted from the  cracker  can vary  widely.   Generally  about
10 percent of  the total sulfur in the reactor feedstocks remains with the coke
on  the  catalyst.  This percentage  increases to  about 20 percent if the feed-
stock has  been hydrodesulfurized  or  cracked in another process before being
                            2
fed to the catalytic cracker.

3.4.2.2  Liquid Fuel Burning-^
     Potential  pollutants  from liquid fuel burning  include  nitrogen oxides,
carbon  monoxide, particulate,  and  sulfur dioxide.  The formation and emission
of  pollutants in a furnace depend on the fuel, the  operation of the furnace,
and the design of the  firebox  and burners.   Because  emissions  of SO2 are
essentially  determined  by  the sulfur in the fuel,  they  can be reduced by
firing fuels  containing  lower levels  of sulfur  but  not  by  redesigning the
burners or the firebox.
      A high  sulfur fuel oil  containing  4.0  percent  sulfur by weight,  when
fired  in  a  furnace with  40  percent  excess  air,  could  produce  a  flue gas
 containing  about 0.18  volume percent SO2 or 1800 ppm by volume.   A furnace
 firing 0.3  percent  sulfur  fuel oil  with  40  percent excess  air would discharge
 0.015 volume percent SO2 or  150  ppm by volume in the flue gas.
      Either improper atomization of the  fuel oil in the burner  or provision of
 insufficient air  to  burn the  fuel oil completely will produce unburned carbon,
 which is emitted as particulate  matter.  Excess air is required in  all furnaces
 to assure  complete combustion,  and thereby  minimize carbon monoxide in the
 flue gas,  but high  excess  air  rates  reduce  thermal efficiency.   When  fired
 with liquid fuel oils, refinery  furnaces operate  with 30  to  50 percent  air in
                               3.4-6

-------
excess  of  that  theoretically required.   When fired with natural gas,  they use
only 10 to  25  percent  excess air.

3.4.3.  Control Measures
3.4.3.1  Catalytic Cracker Units--
     Two  types  of  control  measures  have been applied to control of  SO- at
catalytic  cracking operations:   desulfurization  of  the  cracker  feed  and  flue
gas desulfurization (FGD) of the effluent from the regenerator.
     A  large  research effort has  been  directed toward  development of  hydro-
desulfurization   technology,   and   many   processes   applicable  to  catalytic
cracking feedstocks are  available.   Although hydrodesulfurization  does  reduce
the emission of SO2/  its  use  is limited  for certain  feedstocks.   The degree of
desulfurization  required  to  sufficiently reduce  SO2 emissions  involves  such
severe  conditions and such Jorge amounts of hydrogen  that  complete  hydro-
desulfurization of feedstocks is not possible.
     FGD  is  based  on  intimate contact between the flue  gas and the liquid
droplets of the scrubbing solution.  Over  the  past 15 years,  many absorption
mediums and  contactors  have been tried.   A system developed  by a  major oil
company consists of  a  dilute caustic  stream  in conjunction with a venturi
scrubber,  which efficiently  captures both  the  sulfur oxides  and particulates.
The  process  is in commercial  application  at four  major refineries located in
New  Jersey,  Louisiana,  and  Texas.  The scrubber can  be  used  on  regen-
erators downstream  of  the  waste heat  boilers  or  after  heat removal  from  a
high-temperature  regenerator,  providing adequate pressure drop is available.
The  technology is offered  for license,  and  reportedly two  other  commercial
installations are  planned.   No other  scrubbing  systems are  known  to be in
commercial   use  on  catalytic  cracker  regenerators,  but  several  FGD's  are
installed on oil and coal fired boilers.
     The   mixture  of   gas  and  scrubbing  solution  droplets,  after  passing
through the venturi  contactor,  passes  to  a separator where clean  flue gases
are separated from the scrubbing liquid.   The  vent gas is discharged to the
atmosphere.   A purge liquid stream is  removed to keep  the  circulating liquid
stream  suitable  for scrubbing.  Purge stream treatment facilities are  required
prior  to effluent disposal.
                             3.4-7

-------
     A  commercial-scale  FGD system on  a cracker  regenerator has  been in
                                                                          3
operation since March 1974.   The  design  capacity  of  this  unit is 141.6  m'/s
(300,000 cfm)  of  flue gas,  and the  unit has  achieved  an SO2 collection  effi-
ciency  above  95  percent.   It  has  completed  a  29-month  run  with  no
problems.
                                                                    3
     To  achieve  the  same levels  of pollutant  removal  on a 12,719  m  .(80,000
bbl) per day  Gulf Coast  fluid catalytic   cracking  unit processing  a gas oil
containing  3.2 percent  sulfur,  the  FGD  system  requires an investment .only
two-thirds of that required  for combined  hydrodesulfurization of the feed and
electrostatic  precipitation of the  flue  gas.  Annual operating costs  are  esti-
mated  at half  of the costs  of  hydrodesulfurization and electrostatic  precip-
itation.  The benefits to  a refiner  of  products with lower sulfur contents are
not considered in these  economic comparisons.

3.4.3.2  Liquid Fuel Burning--
     The two methods available to reduce  SO2 emissions are desulturization  of
the fuel before combustion and FGD.
     Reducing the sulfur content of fuel is practiced in some petroleum refin-
eries.    To  prepare  low-sulfur-content  liquid fuel acceptable to  the  home
heating  market,  refiners often desulfurize a  distillate stream  by the use  of
hydrogen  (hydrodesulfurization).   To prepare heavier liquid fuels,  marketed
with  higher  sulfur content, the  same  desulfurized distillate stream is  often
blended with a residual stream to yield the desired products.   Without blend-
ing,   the  high-sulfur-content  residual  stream  would not  be  marketable.
Although the  costs vary widely depending on the  crude oil  type, it is very
costly to prepare  heavy fuel oils with  less than 0.3 percent  sulfur by weight
from  any crude  oil;  thus  most  heavy  fuel oils produced have sulfur contents
of  0.3 percent  or higher.  Only  a small portion  of  the  total  production  of
liquid fuels  is burned  in  the refinery.  By the burning of hydrodesulfurized
gas,  oils and  other distillate streams,  a refiner can reduce  the level  of sulfur
oxides emitted from the furnaces.   At the present time no FGD systems have
been  utilized  in  the  U.S.  to  control  sulfur  dioxide emissions  from  petroleum
refinery  heaters  and  boilers.   In  Japan,  FGD's are used  extensively  to
control SO2  emissions from oil-fired  boilers and heaters.
                              3.4-8

-------
3.4.4  Emission Limits
3.4.4.1  New Source Performance Standards  (NSPS)  Limits--
     NSPS require  control of carbon  monoxide and participate emissions from
fluid  catalytic cracking  unit  catalyst  regenerators and from FCC incinerator
waste  heat  boilers;  there is  no limit  on  SO-  emissions from such units.  In
regard  to liquid  fuel burning,  NSPS  have been  enacted  for  SO0  emissions
                                                          fi
from fossil-fuel-fired boilers of more than  73 MW (250 x 10  Btu/h) heat  input
rate.   The  NSPS  limit  for  SO9  emissions from units  of this  size  that fire
                                        fi
liquid  fossil  fuels is 340  ng/J (0.80 lb/10   Btu) heat  input.   Refinery combus-
tion  devices  are  generally exempt from this NSPS because  of their smaller size
and/or because they are  not used to generate steam.

3.4.4.2  State Implementation Plan  (SIP) Limits--
     Catalytic  crackers--A search  of  SIP  regulations and  state laws  revealed
no  SO9   emission  limits   that  apply specifically  to  catalytic cracker regen-
        4
erators.    In  the absence of source-specific regulations for catalytic cracker
regenerators,  applicable  regulations are  those that fall  under  such general
classifications  as "other  process  sources,"  "noncommercial fuel,"  or "emissions
from fuel combustion operations."   The most stringent state and local general
process  regulations  applicable  to  catalytic  cracker  regenerators  are a  New
Jersey regulation limiting emissions from noncommercial  fuel  to 312  ppm SO9 by
      5
volume   and a regulation of the Bay  Area  Air Pollution Control District, San
Francisco, California,  (Regulation  No.  2   3122) limiting concentrations to 300
ppm  SO2  by  volume.
     Liquid fuel burning—Review   of   state   regulations   has   shown   wide
variation  in  the  types of regulations,  degree  of stringency, and methods of
enforcing the  limitation.   In general SO9 emissions are limited by a regulation
restricting the quantity  of  SO9 emitted per unit quantity of heat input or by
one limiting  the  sulfur  content  of  the fuel.   In  some states the regulations
specify maximum  allowable  ground-level concentrations  resulting  from  emis-
sions.   Many  states  have several  forms  of regulations,  each  applying  to a
different  fuel  or type of source.   Limits on the  sulfur content in fuel oils for
                              3.4-9

-------
all purposes range  from  a  low of 0.15 percent  to a  high of 4.4 percent by
weight.  Most  states  differentiate  the various grades  of  oil, allowing  higher
sulfur contents in residual oils  than in distillate oils.   The most restrictive
regulation  of fuel  oil sulfur  content  is  about 0.15  percent  sulfur; this is  a
regulation  of Clark  County, Nevada,  which, however,  has no petroleum  refin-
ery.   The  most restrictive regulation applicable to an  area  (New Jersey) in
which  refineries are  operated  is  0.2 percent  sulfur  in  No.  2  and  lighter
commercial fuel oils, and  0.3  percent sulfur in Nos. 4,  5,  and 6 residual fuel
oils.

3.4.4.3  Achieved in Practice  (AIP) Limit--
     A commercial-scale FGD  unit has  demonstrated 95 percent removal  of the
                                                                2
SO2  component in exhaust from a  catalytic cracker regenerator.    The  diver-
sity  of liquid  fuel burning  units  and of the fuels they  burn precludes  deter-
mining an  AIP  limit for refinery furnaces.  A general AIP limit would be 0.3
weight percent sulfur in fuel.
3.4.5  Determination of LAER
     The  recommended limitations are  based  on  SIP's and on  performance
information  available  in  early 1979.   It  is  anticipated  that  several additional
SIP  regulations covering  these sources will be promulgated and/or modified  in
1979 and 1980  and  that appreciable new performance data  will become available
in the near term.   Conceivably,  some SIP regulations may be more  stringent
than  the  LAER  suggested  herein.    Furthermore, performance  testing  may
show  that  more stringent limits  than those suggested  are feasible or  it may
show  that the  suggested  limits are appropriate or that they are not achievable
for  some specific subcategories.   In  any case,  the  basis for  determining LAER
for  many  source  categories is expected  to change frequently.  Since LAER is
near  the  vanguard of control technology, a more  detailed analysis is partic-
ularly necessary when addressing modified or reconstructed facilities subject
to  the  provisions  of Section  173 of  the  Clean  Air Act.   Emission limitations
reasonable for new sources  may in  some instances be economically  or tech-
nically unreasonable when applied to modified or reconstructed sources of the
same  type.
                               3.4-10

-------
3.4.5.1   Catalytic Cracking Unit--

     The San Francisco Bay Area  limitation of 300 ppm is  the most stringent
regulation  adopted to date.  Compliance  with  this regulation can be  achieved
with FGD on  even the highest concentration of SO-  anticipated in regenerator
flue gases.   Additionally,  the  period  of  operation between turnarounds on  an
FGD system  has been  demonstrated  to be  as  long  as  the  period between
turnarounds on  catalytic  cracker  units, or longer.   An  FGD system can  be
used  on a  waste gas  stream  from a high-temperature  regenerator or a CO
boiler.
     An  alternative  control measure applicable  to  most catalytic cracker feed-
stocks is hydrodesulfurization or  the processing of  low-sulfur  feed  stocks.
Application  of this process is  limited,  however, because the degree of desul-
furization needed to  conform with strict emission  regulations is extremely  high
relative to normal feedstock desulfurization  requirements.
     Therefore,  it is concluded  that the LAER for SO- emissions  from cata-
lytic cracker  regenerators is  300  ppm or a 95  percent by volume reduction of
the uncontrolled  concentration from  the  source,  whichever  is  less stringent;
this degree  of  control  may  be achieved by  application  of  FGD or by  other
equally effective methods.

3.4.5.2   Liquid Fuel  Burning--
     Although  some  industries are installing  FGD systems  on industrial-size
boilers and  furnaces, none are yet installed in  petroleum refineries.  A petro-
leum refiner's primary  business  is production of marketable  fuels  that  meet
Federal,  state,  and  local regulations  on sulfur or SO^  content.   Because  he
burns  for  for his own  use  only  2 to  6 percent of  the  fuel produced,  the
refiner may be  in a  position to achieve  a low  emission rate by preparing and
burning  low-sulfur   fuels.  Most  other  industries,  lacking the  capability  to
produce  their own fuel  and to ensure a low-sulfur  fuel supply,  may  have to
attain LAER by alternative means.
     Alternatively, refiners are also  in  a better position than their customers
to install an FGD system  and  burn high sulfur residual oil.   The furnaces at
                             3.4-11

-------
refineries are located near the process they serve and therefore are sometimes
scattered throughout  the refinery  property.   The  distance  between furnaces
could  require  that  separate  systems be  installed  if FGD  were  chosen for
reduction of SO- emissions.   New refineries can  generally incorporate  plans  to
install fuel  oil  desulfurization  equipment on  a single stream, produce low-
sulfur fuel  oil,  and  overcome  the constraints  involved  in  FGD application.
     In summary, SO2 emission  from liquid fuel burning is directly related  to
the sulfur content of the fuel.  The LAER for liquid  fuel burning  is that SO2
emission level equivalent to the use  of fuel having  0.3 weight percent sulfur,
corresponding to approximately 150  ppmv  SO-.
                               3.4-12

-------
                          REFERENCES
1.  Annual Refining Survey.   The Oil and Gas Journal,  March 20,
    1978.

2.  Data compiled from files of Greene & Associates,  Inc.

3.  Fluid Catalytic Cracking Unit Flue Gas Scrubbing.   Exxon
    Technology Sales Bulletin,  April 1977.

4.  National Summary of State Implementation Plan Reviews (Sec-
    tion 4 ESECA).  EPA-450/3-75-053b.  July 1975.

5.  State Air Laws.  Environmental Reporter, published by the
    Bureau of National Affairs.
                             3.4-13

-------
3.5  MAJOR SOURCE CATEGORY:  FABRIC COATING--VOLATILE ORGANIC
     COMPOUND (VOC) EMISSIONS
3.5.1  Process Description
     Fabric   coating   involves   the    coating   of   a   textile   substrate
to   impart   properties   that   are   not    initially   present,    such   as
strength,    stability,    water    or    acid   repellency,   or    appearance.
The   fabric   coating   industry   is   diverse,    with  wide  variations  in
both   products   and   plant  sizes.    The  coated   textiles  are  used  in
industrial    and   electrical    tapes,    tire    cord,    utility   meter    seals,
imitation    leathers,    tarpaulins,    shoe    materials,    upholstery    fab-
rics,  and rubber-coated fabrics.
     Coating   solutions   may   be   either   aqueous   or   organic   based.
The   latter  produces   organic   emissions.    It   is   estimated  that  36
Gg/h  (80   x   106  Ib/yr)   of  VOC  is  emitted  in  the   United  States  by
the vinyl-coated fabric  segment of the industry.
     Figure    3.5-1    shows    a    typical    fabric   coating     operation.
Milling   and  mixing   of   coatings    are   primarily   restricted   to    coat-
ings   containing   rubber,   and   emissions  are  not   considered   signif-
icant.
     Fabric   is   usually   coated  with   a   knife   or   a   roller  coater.
Both   are   spreading   devices  for   high-speed  application   of  coatings
to   flat   surfaces;   the   operations   are   very   similar   to   the   paper
coating   techniques    shown   in    Section   3.18,    Figures    3.18-2    and
3.18-3.
     In   knife    coating,   probably    the   least    expensive   method,   the
substrate   is  held   flat  by  a  roller  and   is   drawn  beneath  a   knife
that    spreads   the  viscous   coating    evenly    over   the   full   width  of
the   fabric.    Knife    coating  may   not   be  appropriate   with  materials
such   as   certain   unstable   knit   goods   or   in  applications   requiring
great precision  in the coating thickness.
     In   roller   coating,    the   coating  material  is  applied   to  moving
fabric,  in a  direction  opposite to the movement of  the substrate,
                               3.5-1

-------
             RUBBER
    PIGMENTS
CURING AGENTS
                                        SOLVENT
             MILLING
                     MIXING
           DRYING AND
             CURING
                     COATING
                   APPLICATION
                                                          FABRIC
                         COATED PRODUCT
Figure 3.5-1.   Typical fabric  coating  operation,
                           3.5-2

-------
by hard rubber or steel rolls.  Roller coaters apply a coating of
constant thickness without regard to fabric irregularities.
     Rotogravure  printing is  widely  used in  vinyl  coating  of
fabrics and is  a  large source of solvent emissions.  Rotogravure
printing involves  a roll coating technique in  which the pattern
to be  printed  is  etched on  the  coating roll with  thousands  of
tiny recessed  dots.   The recessed  dots  pick up  ink from a res-
ervoir and transfer it to the fabric surface.
     After  being  coated,  the  fabric  is  sent  to  drying ovens.
Typical drying  ovens process  fabric continuously, operating with
a web  or  conveyor  feed  system.   Ovens  can be  enclosed or semi-
enclosed;  depending on size,  they exhaust from a few thousand to
tens of thousands of cubic feet per minute of air.  Newer instal-
lations are reported to operate with exhaust concentrations up to
40 percent  of  the lower explosive  limit (LEL).  The  oven heat
accelerates evaporation  of  the solvent  and  can produce chemical
changes within  the coating solids  to  give desired properties to
the  product.   Many operators  control  evaporation rates  to give
desired properties to  the  coated  fabric.   High  air  velocities
distribute  heat  uniformly  over  the fabric  surface,  facilitate
heat  transfer  to  the  coating  and  substrate (by minimizing the
laminar zone  next to  the the  coated surfaces),  and remove evap-
orated  solvents  from  the oven at a  rate that will prevent their
buildup to explosive levels.

3.5.2  Emission of Pollutants
     The coating line, consisting of the coating  application area
and  drying  oven,  is the largest source  of solvent emissions in a
fabric  coating  plant.   It is also the most readily controllable.
The  coater  and  the oven are both considered significant emitting
facilities.  Some  coating plants report that over 70  percent of
the  solvents  used within the plant  are  emitted from the coating
line.   Other  plants,   especially  those  using   vinyl  coatings,
report  that only 40 to 60 percent  of  the solvents purchased are
                               3.5-3

-------
emitted from  the  coating line.  The remaining  solvents  are lost
as fugitive emissions  from  other  stages of processing and clean-
up.   Control   techniques  for  fugitive  sources include  tightly
fitting covers  for open tanks, collection hoods  for areas where
solvent is  used for cleanup,  and closed containers  for solvent
wiping cloths.
     Solvent emissions from the coating applicator account for 25
to 35  percent of  all solvent emitted  from a  coating line.  This
solvent may be collected by totally enclosing the appliccitor in a
small room or booth and exhausting the booth to a control device.
Another method  is to  cover  the applicator with  a  hood  that can
collect most  of  the  solvent emissions.  Solvent  emissions from
the  ovens  account  for 65  to  75  percent of  all  solvent emitted
from  a coating line.   In most  ovens,  almost all  the solvent
emissions are captured and vented  with exhaust gases.   On some
coating lines,  emissions  from the  coating  applicator  hood are
ducted to the oven and included with the oven exhaust.

3.5.3  Control Measures
     Although  few fabric coating  facilities  have elected to con-
trol  organic  emissions,   several  technically  feasible  control
systems are available.  These are carbon adsorption and incinera-
tion.   Another  approach  to  reducing organic  emissions  is  to
switch  to  coatings with lower organic solvent content,  such as
aqueous emulsion  coatings.
     As  in the paper  coating industry (Section  3.18),  carbon
adsorption  systems  on fabric coating  lines have been shown  to be
97 to  98  percent  efficient in  controlling organic solvent vapors
that are  drawn into the carbon bed.2'  Control efficiencies are
limited somewhat  by the inability  to  capture all emissions from
the  coating  application  area.   In  paper   coating operations,
recovery  of the solvent introduced  to the  coating line has been
documented  in the  range of 96 percent.  The similarity between
                               3.5-4

-------
fabric and  paper coating lines suggests that  this  efficiency is
also achievable  in  fabric  coating operations.  '    The FACT docu-
ment pertaining  to  fabric  coating suggests a  minimum  of  90 per-
cent efficiency for collection of coating line emissions.
     Both catalytic incinerators and thermal incinerators  (after-
burners) can  destroy 95 to  99  percent of  the organic emissions
introduced  to them.   As  stated  earlier,  the overall  facility
control is  dependent  on the  solvent emission capture efficiency.
Although the use of afterburners in fabric coating plants  has not
been  documented,  afterburner efficiencies  of 98  to  99  percent
                                        /:
have  been  obtained  across  the device.    The same  efficiencies
should also be achievable in fabric coating operations.
     Although incineration consumes energy,  recovery of heat can
eliminate or minimize  this disadvantage.   Fuel costs can  also be
reduced by  increasing the organic level in exhaust  gases,  i.e.,
by reducing dilution air.
     As shown in Table 3.18-1  (paper  coating), an overall reduc-
tion  of  80 to  100  percent  can be attained through  the  use of
coatings with inherently  low  levels   of  organic solvents.   The
degree of  reduction depends  on  the  organic solvent  contents of
the  coating used originally and the  new  coating.   No  industry
contacts reported information from plants using low-solvent coat-
ings.  Although  some  plants  have  converted to use of low-solvent
coatings,   this   action cannot  be considered  a universally  ap-
plicable control measure.   Coating line  operations  and  fabric
specifications vary widely.
     Several  considerations  affect  the technical   and  economic
feasibility of  organic  emission  control  in  the fabric  coating
industry.    Although the larger  facilities may  specialize in  a
specific product, many plants produce  a  variety of  products or
operate under  contract to coat products  to a customer's  speci-
fications.   The  latter operators, often called "commission coat-
ers," must  use  a variety of coating formulations to comply with
                               3.5-5

-------
the  customer's  specifications.   The  resulting  variations  in
emissions  present problems  in  the  design  of  control  systems.
Even  if the  operator  knows  the  solvent compositions,  exhaust
volume and controls must  be  based on the most critical or diffi-
cult  situation.   The number  of  solvents  used  also  affects the
owner's ability to recover and reuse the solvent.  Thus,  the type
of  coating is  an important  factor  in  the  cost  of  controlling
emissions from a fabric coating plant.

3.5.4  Emission Limits
     The initial  criterion for  defining LAER for a surface coat-
ing  industry  is the  degree  of emission control required by the
most  stringent  regulation adopted and successfully enforced by a
state or local air pollution control agency.
     As  reported  elsewhere,  most  regulations  of organic solvent
emissions  are  patterned after what is now Rule  442  of the South
                                                     7
Coast  (California)  Air Quality  Management District.    Review of
regulations in the 16 states that contain about  85 percent of all
surface coating industries showed them to be essentially the same
            Q
as  Rule  442.    Indiana  has  the most stringent regulation in that
it  limits  organic solvent emissions to  1.4  kg/h (3  Ib/h) or 6.8
kg/day  (15  Ib/day)  unless such emissions are reduced by at least
85  percent,  regardless of the  reactivity or  temperature  of the
solvent.  Organic solvents that have been determined to be photo-
chemically  unreactive or that  contain  less  than specified per-
centages of photochemically reactive organic materials are exempt
from  this regulation.
      The California Air Resources Board recently  adopted a model
rule  for  the  control  of VOC  emissions  from  paper  and fabric
                   Q
coating  operations.   This model rule, which must be met 3 years
from  the date  of adoption,  limits VOC emissions  from the coating
line  to  120 g solvent/liter  (1.0  Ib/gal) of coating minus water.
This  is to be  accomplished by the  use  of add-on  control equipment
                               3.5-6

-------
unless the  solvent  content  of the  coating  used is no  more than  265  g/liter
(2.2 Ib/gal) of coating minus water.
     In  the definition  of  LAER  for  surface  coating  emissions,   it  is  not
appropriate to  exempt  solvents  based on  their  reactivity.   Recent  research
has  indicated  that  substituting  low-reactivity  solvents  for  higher-reactivity
solvents  may  improve  photochemical oxidant  air  quality  in one  city while
worsening it in downwind regions.10  Accordingly, EPA has adopted a policy
emphasizing  the  need  for  "positive   reduction   techniques"   rather  than
                          Y
substitution of compounds.
     Emission   controls  achieved   in   practice  for  fabric  coating   exceed
regulatory  requirements by  a wide  margin.  Therefore, it is  concluded that
LAER  for  fabric coating is  a function of controls  achieved  in  practice rather
than  controls required by current  regulations.

3.5.5  Determination of LAER
     The  recommended  limitations  are  based  on  SIP's and  on  performance
information available  in early 1979.   It  is  anticipated  that several additional
SIP regulations covering  these  sources  will be  promulgated and/or modifed in
1979 and 1980 and  that appreciable new performance data will become available
in  the  near term.   Conceivably,  some SIP  regulations may  be  more  stringent
than   the  LAER  suggested  herein.  Furthermore,  performance  testing  may
show  that  more stringent limits  than those suggested  are  feasible or it may
show that the suggested limits are appropriate or that they are not achievable
for some specific subcategories.   In any case,  the basis for determining LAER
for many  source  categories  is expected  to  change frequently.   Since LAER is
near   the  vanguard   of  control  technology,   a  more  detailed  analysis  is
particularly necessary  when addressing modified  or  reconstructed facilities
subject  to the  provisions  of Section  173  of  the  Clean Air  Act.   Emission
limitations  reasonable  for new sources  may in  some instances  be economically
or  technically unreasonable  when  applied to modified  or reconstructed  sources
of  the  same type.
      Control  efficiencies  greater  than   95  percent across  the  control device
have been documented on fabric coating operations  using  thermal or catalytic
                               3.5-7

-------
                                           *J  *\ 4  f\
incineration or  carbon adsorption systems.  '  '  '    At least  90 percent of the
VOC emissions  from  fabric coating can  be sent to  a control device that yields
an overall control  efficiency  of 85 percent,  which  is  equivalent to the uncon-
trolled  emissions  from a low-solvent coating containing 310  g of  solvent/liter
(2.6 Ib/gal) of coating minus water.  The following  shows this derivation:

     Assuming a typical coating contains 22 percent solids,
     100 gallons of coating contains  22  gal solids
                                 and  78  gal solvent
     Reducing the solvent emitted yields:  (78)  - [0.85 (78)]  =
     11.7 gal permitted
     Assuming a solvent density of 7.36  Ib/gal,  11.7 gal =86.1
     Ib solvent
     Equivalent coating required = 86.1  Ib solvent/(22 + 11.7)
                                   gal of coating minus water
                                 = 2.6 Ib solvent/gal coating
                                   minus water
                                 = 310 g solvent/liter coating
                                   minus water
     Although  conversion   to   a   waterborne   or   higher-solids   coating
will    significantly   reduce    VOC    emissions,   the   310  g/liter   (2.6
Ib/gal)   limitation  may  not  be   achievable,   in  which  case   control
of   part  of   the  VOC  emissions   is  still   recommended.    The  recom-
mended   LAER   limitation   for   vinyl   coating   is  370    g   solvent/liter
(3.0   Ib/gal)   minus   water,   based   on  solids   content   of  15   percent
and    a    solvent   density    of   826    g   solvent/liter   (6.7    Ib/gal).
Complete   control  of   VOC   emissions    using   add-on   control   devices
providing   at   least   85    percent    overall    plant    control    is    also
acceptable.
                               3.5-8

-------
                           REFERENCES
1.    Control of Volatile Organic Emissions for Existing Station-
     ary Sources,  Volume II:   Surface Coating of Cans,  Coils,
     Paper,  Fabrics,  Automobiles,  and Light-duty Trucks.
     EPA-450/2-77-008,  U.S.  Environmental Protection Agency,
     Research Triangle Park,  North Carolina.   May 1977.

2.    Letter from W. C.  Moses.  Technical Manager, Chemical Plant
     Division, Sutcliffe, Speakman & Company, Limited.   March 10,
     1978.

3.    Personal communications with R. W.  Laundrie.  Manager-
     Ecology, Safety and Health Engineering,  Chemical/Plastic
     Division, General Tire and Rubber Company, Akron,  Ohio.
     March 1978.

4.    Personal communications with N. Glazer.   Philadelphia Air
     Management Services.  February 21 and 28, 1978.

5.    Personal communication with S. Bruntz.   Division of Air
     Pollution, Kentucky Department of Environmental Protection.
     March 1, 1978.

6.    Vincent, E.J., et al.   Are Afterburners  Obsolete?  Presented
     at Air Pollution Control Equipment Seminar, APCA/National
     Association of Corrosion Engineers, Altanta, Georgia.
     January 17-19, 1978.

7.    Recommended Policy of Control of Volatile Organic Compounds.
     Federal Register,  42 FR 131,   U.S.  Environmental Protection
     Agency.  July 8, 1977.

8.    Source Assessments:  Prioritization of Air Pollution from
     Industrial Surface Coating Operations.   EPA 650/2-75-0192,
     U.S. Environmental Protection Agency.  February 1975.

9.    Lam, J.Y., et al.   Consideration of a Proposed Model Rule
     for the Control of Volatile Organic Compounds from Paper and
     Fabric Coating Operations.  Prepared for California Air
     Resources Board, Sacramento,  California.  August 23, 1978.
                               3.5-9

-------
10.   Control Strategy Preparation Manual for Photochemical
     Oxidant.  OAQPS 1.2-047,  U.S. Environmental Protection
     Agency.  January 1977.

11.   Recommendation by U.S.  Environmental Protection Agency,
     ESED.  Raleigh-Durham,  North Carolina,   August 8,  1978.
                                3.5-10

-------
3.6  MAJOR SOURCE CATEGORY:  LARGE INDUSTRIAL BOILERS--PARTICULATE,
     SULFUR DIOXIDE, AND NITROGEN OXIDE EMISSIONS
3.6.1  Process Description
     Industrial  boilers  are fired  with coal, natural  gas,  oil,
and industrial wastes.   The  resulting heat of combustion is then
used -to  produce  steam in the boilers.   Liquids  other than water
may be  heated in  the boiler  for  use  in  subsequent industrial
processes.   The  hot  flue gases,  after producing  steam  in  the
boiler,  pass  through  an economizer  and sometimes an air heater
and are finally discharged through a stack.
     This section deals with coal-fired boilers in the heat input
range of 58 to 175 MW (200 to 600 x 106 Btu/h), including stokers
in the range  of  88  to 117 MW (300  to  400  x 106 Btu/h); and with
oil-fired boilers in  the range  of 73 to 102 MW (250 to 350 x 10
Btu/h).

3.6.1.1  Pulverized-Coal-Fired Boilers--
     For  combustion in  a  pulverized-coal-fired boiler,  coal  is
pulverized so that  at least  70  percent passes through a 200-mesh
sieve.  This  finely ground  coal  is conveyed pneumatically  to  a
burner located  in  the  furnace.  The  system operates  as  a  con-
tinuous process;  within  specified  design  limitations,  the  coal
feed  can  be  varied as  required  by boiler load.  Figure  3.6-1
depicts a pulverized-coal-fired boiler system.
     A small portion of the air required for combustion (15 to 20
percent in  current  installations) is used  to  transport the  coal
to the  burner.   This  is known  as  primary combustion  air.   The
primary air is also used to  dry the coal in the pulverizer.   The
remainder of  the combustion  air (80 to 85 percent)  is introduced
at the burner and is  known as secondary air.  The control system
regulating the flow of  both  coal and primary  air is  so designed
that a predetermined  air/coal ratio is  maintained  for  any given
load.1
                              3.6-1

-------
   PENTHOUSE
   PENDANT
  REHEATER
                                                  STEAM DRUM
 SECONDARY
SUPERHEATER
            ELECTROSTATIC
            PRECIPITATOR
    PRIMARY
   SUPERHEATER
   ECONOMIZER
   STEAM COIL
   AIR HEATER
       FORCED DRAFT FAN
                                         PRIMARY AIR FAN
Figure 3.6-1.   Radiant boiler  for pulverized  coal firing,
                 (The Babcock  &  Wilcox  Co.)
                              3.6-2

-------
     Burners  are  characterized by  firing position,  i.e.,  wall-
fired  or  tangential.     Arrangements   for  the  introduction  of
primary, secondary,  and,  in some cases,  tertiary air  vary with
burner manufacturers.  One manufacturer uses an adjustable burner
that  can be  tilted upward  or  downward  to  control  the  furnace
outlet  temperature  so  that  steam  temperature  can be  regulated
over a wide range.
     Pulverized-coal-fired boilers  may  be designed as either wet
or dry bottom units.  In a wet bottom furnace, the temperature is
maintained above the ash fusion temperature and the ash is melted
so that it can be removed from the  furnace as a liquid.  In a dry
bottom  furnace,   the  temperature   is  maintained  below the  ash
fusion temperature so that the ash will not fuse.

3.6.1.2  Stoker-Fired Boilers—
     Stoker-fired boilers  are  used  in  the small  to  medium size
ranges from 9 to 117 MW (30 to 400 x 10  Btu/h).  The stokers are
designed to feed  coal  onto a grate  in  the  furnace and to remove
the  ash residue.   Higher  rates  of  combustion  are possible,  and
the  continuous process  of stoker firing permits good control and
high  efficiency.   Stokers  are  often preferred over pulverizers
because  of  their  greater  operating  range (i.e., operation at low
loads)  and  lower  power requirements, and because they can burn a
variety of solid fuels.
     The grate area required for a  given stoker type and capacity
is determined from  rates  established by experience.  Table 3.6-1
lists  maximum recommended coal burning rates for three types of
stokers, which are discussed in the  following paragraphs.
                              3.6-3

-------
     TABLE 3.6-1.   MAXIMUM ALLOWABLE COAL BURNING RATES  FOR
                     THREE TYPES OF STOKERS1
Type of stoker
Multiple-retort
underfeed stokers
Chain- or traveling-
grate stokers
Spreader stokers
Coal burning rate
MW/mz
1.89
1.58
2.36
(Btu/ft'-h)
(600,000)
(500,000)
(750,000)
     Multiple-retort  underfeed  stokers—In  a  multiple-retort,
rear-end-cleaning stoker the retort  and grate are inclined 20 to
25  degrees.   These  units  usually  consist  of  several  inclined
retorts side  by side, with  rows  of tuyeres  between each retort
(Figure 3.6-2a).   Coal  is  worked from the  front hopper  to the
rear ash-discharge  mechanism by pushers.  The forced-air system
is zoned beneath  the  grates by means of air dampers, and combus-
tion control  is fully modulated.   In   larger  furnaces  the walls
are water-cooled,  as  are the grate  surfaces  in  some units.   Use
of  multiple-retort underfeed  stokers   is  declining.   Capacities
generally range from 7 to 146 MW (25 to 500 x 106 Btu/h).

     Chain-grate or traveling-grate  stokers—A chain-grate  (Fig-
ure 3.6-2b) or traveling-grate unit  consists essentially of grate
sections that move from the front to the rear, carrying coal from
the hopper in front into the combustion zone.  The fuel bed moves
progressively  to  the  rear,  where  the ash  is  continuously dis-
charged.   Modern  units  have  zone-controlled  forced  draft.
Complete  combustion-control systems are  used,  and overfire air,
especially in the  front wall, aids combustion of the volatiles in
the  fuel.   Capacities  range from  6 to 88  MW (20 to  300 x 10
Btu/h) heat input.

     Spreader stokers—The  spreader  stoker  combines  suspension
and  fuel  bed firing  by the stoker  mechanism,  which throws coal
into  the  furnace  over  the  fire with  a uniform spreading action
(Figure 3.6-2c).   Because coal is  burned partly  in suspension and

                              3.6-4

-------
                                           COAL HOPPER
                                                | COAL RAMS
               ASH DISCHARGE PLATE
         FUEL
      DISTRIBUTORS
             a.   Multiple-retort underfeed  stoker.
                                    OVERFIRE-AIR
                                     NOZZLES
                        COAL HOPPER
                   COAL GATE!
             -RETURN
               BEND
DRAG
PLATE
STOKER
CHAIN
 DRIVE
SPROCKET
HYDRAULIC
  DRIVE
b.  Traveling-grate  spreader stoker with  front ash  discharge.-1
                     COAL
                    HOPPER
                 FEEDER
                 STOKER
                 CHAIN

               ASH  HOPPER
   OVERFIRE
    AIR

   OVERTHROW OVFRFTR[r
     ROTOR  OV!?"RE'
             Al K
                    c.   Chain-grate  stoker.

                 Figure  3.6-2.  Types  of  stokers.

                                  3.6-5

-------
partly  on  the grate,  the coal  forms  a thin,  fast-burning bed.
This method  of firing provides  rapid  response to  load fluctua-
tions.   The  grates  are  either  stationary  or  move  continuously
from the rear to the front.
     Partial  suspension burning of  coal  in  a  spreader  stoker
results  in  greater carry-over of particulate  matter  in the flue
gas.   Spreader  stokers  therefore  must  be  equipped  with dust
collectors,  and  the larger  carbon-bearing  particles  are often
recirculated  to  the furnace  for further burning.  Recirculation
of  the  larger particulate  into the  furnace  can result  in an
increase  in boiler  efficiency  of  2  to 3 percent.   Capacity of
spreader  stokers  range  from  2 to 146 MW  (6  to 500 x  106  Btu/h)
heat  input.  '

3.6.1.3 Oil-Fired Boilers—
      Fuel  oil is  atomized and burned  in suspension.  Atomization
produces fine oil droplets that expose  a  large surface area  per
unit  of oil  volume  to  the hot  furnace and promotes combustion.
      The oil  burners  are normally  located in  the vertical walls
 of the furnace,  as shown in Figure  3.6-3.  Before the oil  reaches
 the burner it is  passed through  a strainer or filter  that  removes
 sludge.   This  filtering process  prolongs  pump  life,  reduces
 burner wear, and increases combustion efficiency.
      For proper  atomization,  oil of  a  grade heavier  than No^ 2
 must be heated  to  reduce  its viscosity to  26 to 30 x 10   m /s
 (130 to  150 Saybolt Universal).   Steam  or  electric heaters  are
 required  to raise  the  oil  temperature  to  the  required  degree:
 approximately 57°C  <135°F) for  No.  4 oil, 85°C (185°F) for No. 5
 oil,  and 93° to 12l°C (200° to 250°F) for No. 6 oil.

 3.6.2  Emissions
 3.6.2.1  Emissions  from Coal-Fired Boilers—
      Flue  gases  from  coal-fired  boilers  contain   particulate
 matter and gaseous products of combustion,  including oxides  of
 sulfur and  nitrogen.

                               3.6-6

-------
                        CENTERING   AIR REGISTER DOOR     OIL
                         SUPPORT     (SECONDARY AIR)    ATOMIZER
                   LIGHTER
                REGISTER
                DRIVE ROD
                IMPELLER    REFRACTORY THROAT   WATER-COOLED
                          WITH STUDDED TUBES   FURNACE WALL
Figure 3.6-3.   Oil  firing  burner  with water-cooled throat.'
                            3.6-7

-------
     Particulates—The  quantity   of  uncontrolled   particulate
emissions depends  primarily on the type of  combustion  unit,  the
ash content of the coal,  the fuel rate,  and the degree of fly ash
reinjection.  On stoker-fired units,  the  grate heat-release rate
and coal size also affect emissions.
     When pulverized coal is burned, nearly all the ash particles
are  formed  in suspension  and about 80 percent  of  the ash leaves
the  furnace entrained in  the  flue gas.   In a slag-type  or wet
bottom furnace, however,  as much as 50 percent of the ash may be
retained in the furnace.
     In  a  properly operated stoker burning  coal,  the passage of
air  and  the agitation of  the fuel bed on the grate serve to keep
ash  accumulations  more or less  porous,  and the ash is discharged
to an ashpit in fairly large pieces.
     With  a spreader stoker, some of the  fuel  is  burned in sus-
pension  and a considerable quantity Of ash particles, containing
some unburned  fuel,  is  consequently carried over with the gases.
This material  is  usually  collected in hoppers and  may be rein-
jected into the furnace for  further burning.

     Sulfur oxides—In  a  coal-fired  furnace,   about  90  to  95
percent  of  sulfur  in  the  coal  is  converted to  sulfur oxides
(SO  ).4  The  balance of  the sulfur is emitted  in  the fly ash or
   •"•                                                          >
combines with the  slag or  ash in the furnace and is  removed with
them.  Sulfur  dioxide is the principal oxide of sulfur; only 2 to
3  percent  of the sulfur content of the coal is emitted as sulfur
trioxide.   Rates  of sulfur oxides emissions depend on the sulfur
content  of  the coal  and not on the  type of furnace.

     Nitrogen  oxides—Emissions  of  nitrogen   oxides   (NOX)   are
caused  by high-temperature  reaction  of  atmospheric nitrogen  and
oxygen  in the combustion zone (called "thermal N0x") and  also by
partial  combustion  of  nitrogenous compounds  in  the fuel  ("fuel
nitrogen").   The  important factors  that affect  NOx production  are
flame and  furnace  temperature, residence  time  of combustion  gases
                               3.6-8

-------
at  the flame  temperature,  rate of  cooling  of  the  gases,  and
amount of excess air in the flame.
     Emission  factors  for  coal-fired boilers are  presented  in
Table S.6-2.5

3.6.2.2  Emissions from Oil-Fired Boilers—
     Emissions  from  fuel  oil combustion depend on  the grade and
composition  of the  fuel,  the type and  size of the  boiler,  the
firing and  loading practices, and  the level  of  equipment main-
tenance.    Table 3.6-3  presents  emission   factors  for  fuel  oil
combustion in industrial boilers without control equipment.   The
emission  factors  for  industrial boilers  are  grouped  into  dis-
tillate  and residual  oil  categories  because  the  combustion  of
each produces  significantly  different emissions  of particulates,
SO  , and NO  .
  X        X
     Particulates—Particulate  emissions  are  most  dependent  on
the  grade  of  fuel  fired.    The lighter  distillate  oils  cause
significantly  lower particulate formation  than  do  the heavier
residual  oils.  In  boilers  firing Grade 6 oil, particulate emis-
sions  generally can be  considered as a  function  of the sulfur
content of the oil.  This is because Grade 6 oil,  whether refined
from naturally  occurring  low-sulfur crude  oil or desulfurized by
one of several processes currently in practice, has substantially
lower  viscosity than  other  grades,  and  also  lower asphaltene,
ash, and  sulfur contents, all of which lead to better atomization
and cleaner  combustion.
     Boiler  load can affect particulate emissions in units firing
Grade  6  oil.  At low  loads  particulate  emissions  may be reduced
by  as  much  as  60  percent.  No significant particulate reductions
at  low loads have been noted in boilers firing any of the lighter
grades.

     Nitrogen oxides—Emissions  of  NO  formed from fuel nitrogen
                                      X
are primarily  a function of the nitrogen  content of the fuel and
the  available  oxygen.   Emissions of  thermal  NO   are largely  a
                                                 X
                              3.6-9

-------
    TABLE 3.6-2.   EMISSION  FACTORS  FOR  BITUMINOUS-COALgFIRED
          INDUSTRIAL  BOILERS WITHOUT  CONTROL  EQUIPMENT
Furnace capacity,
MW (106 Btu/h)
heat input
Greater than 29 (100)
Pulverized
Wet bottom
Dry bottom
Spreader stoker
Particulates,
kg/Mt (Ib/ton)
coal burned


6.5A (13A)C
8.5A (17A)
6.5A (13A)e
Sulfur *
oxides,
kg/Mt (Ib/ton)
coal burned


19S (38S)
19S (38S)
19S (38S)
Nitrogen
oxides,
kg/Mt (Ib/ton)
coal burned


15 (30)
9 (18)
7. .5 (15)
a The letter A on all units indicates that the weight percentage
  of ash in the coal should be multiplied by the value given.
  Example:  If the factor is 8 and the ash content is 10 per-
  cent,  the particulate emissions before the control equipment
  would be 10 times 8,  or 80 kg of particulate per Mt of coal
  (10 times 16, or 160 pounds of particulate per Mt of coal).

b S equals the sulfur content (see footnote a above).

0 Without fly-ash reinjection.

d For all other stokers use 5A for particulate emission factor.
  Emission factor data for stokers with capacities greater than
  29 MW are not documented.

e Without fly-ash reinjection.  With fly-ash reinjection from
  first-staged collector, use 20A.
                               3.6-10

-------
     TABLE 3.6-3.  EMISSION FACTORS FOR FUEL OIL COMBUSTION'
Pollutant
      Residual oil,
   kg/kl (lb/103 gal)
  Distillate oil,
kg/kl (lb/103 gal)
Particulate

Sulfur dioxide

Nitrogen oxides
(total as N02)
     19.25 (158.65)

      7.5C (60)C
    0.25 (2)

   17.2S (143.68)

    2.8  (22)
  Particulate emission factors for residual oil combustion are
  best described, on the average, as a function of fuel oil
  grade and sulfur content, as shown below.
     Grade 6 oil
kg/kl = 1.25(S) + 0.38
[lb/103 gal = 10,(S) + 3]
          where:  S is the percentage, by weight, of sulfur
                  in the oil

     Grade 5 oil:  1.25 kg/kl (10 lb/103 gal)

     Grade 4 oil:  0.88 kg/kl (7 lb/103 gal)

  S is the percentage, by weight, of sulfur in the oil.

  Nitrogen oxides emissions from residual oil combustion in
  industrial and commercial boilers are strongly dependent on
  the fuel nitrogen content and can be estimated for accurately
  by the following empirical relationship:

          kg N02/kl = 2.75 + 50(N)2,

          [Ib N02/103 gal = 22 + 400(N)2]

          where:  N is the percentage, by weight, of nitrogen
                  in the oil.  Note:  For residual oils having
                  high (>0.5%, by weight) nitrogen contents,
                  one should use 15 kg N02/kl (120 Ib N02/103
                  gal) as an emission factor.
                               3.6-11

-------
function of peak  flame  temperature  and available oxygen,  factors
that depend  on boiler  size,  firing  configuration,  and operation
practices.  Fuel NO  are the predominant NO  emissions in boilers
                   X                       X
firing residual oil.  Thermal  NO  emissions predominate in units
                                 *\
firing distillate oils,  primarily because the nitrogen content of
these lighter oils is negligible.

3.6.3  Control Measures
3.6.3.1  Particulate—
     Problems  of  ash  removal  and disposal are significant, prin-
cipally where  solid fuels are burned.   Fuel oil contains little
ash,  and  any  ash  formation  primarily  affects  the  furnace and
boiler interiors.   Mechanical  dust collectors are used occasion-
ally.
     With  the  early  methods of burning   coal  on  grates  with
natural draft, most of  the coal  ash remained on the grate  and was
ultimately discharged into a hopper for disposal.  With the newer
boilers,  such as  the  spreader  stoker and  the  pulverized-coal-
fired  boilers, part or all of the  burning occurs in  suspension,
leading  to greater carry-over of particulate matter in the flue
gas.
     Achieving a  low emission  rate requires  some form  of particu-
late  control equipment to remove the  fly ash from flue gases of
units  that burn solid  fuels.   In addition,  careful operation and
use of the  fuel specified  for  the boiler  are required  to minimize
visible  emissions.   The   commercially available  high-efficiency
particulate  removal equipment includes electrostatic precipita-
                                         £i
tors,  fabric filters, and  wet  scrubbers.

     Electrostatic  precipitators—Precipitators   are   the  most
widely used particulate control device on  pulverized-coal-fired
boilers.   Electrostatic precipitators  impart an electric charge
to the  particles to be  collected  and then  propel  the  charged
particles by  electrostatic  force  to  the collecting  electrodes.
                               3.6-12

-------
Collection efficiency of an electrostatic precipitator depends on
the  time of  particle exposure  to the  electrostatic field, the
strength  of  the field,  and the resistivity of the dust particle.
Efficiency  above 99  percent can  be  achieved  when  the  unit is
properly  designed.   Table 3.6-4 presents  selected operating and
design data for  a spreader-stoker  coal-fired boiler.

     Fabric filters—These devices are being used increasingly on
coal-fired stokers, but  to a lesser extent on those  burning pul-
verized  coal.   Use  of  fabric  filters  is  favored  when sulfur
content  of  the  coal  is  very low  and when  carbon content of the
particulate is high (as in spreader stokers).
     Fabric  filters  trap  dust particles by  impingement  on the
fine fibers  of the  fabric.  As the collection of dust continues,
an accumulation  of  dust  particles  adheres to the fabric surface,
forming   a  highly  efficient  filter  cake.   The  fabric  filter
achieves  maximum efficiency during this  period  of dust buildup.
After  a   fixed  operating period,  which  depends  on  the  pressure
drop, the bags are cleaned by passage of a reverse flow of air or
by mechanical  vibration.   Filtering efficiency  is slightly lower
after  cleaning  until  the  collected  dust  again forms  a filter
cake.
     The  fabric  filter can be applied to any boiler for which dry
collection is desired and maximum temperatures are  lower  than
about  288°C   (550°F).7   Coated fiberglass  filters  are generally
used at  the  upper temperature  limit.  Efficiencies  greater than
99 percent can be achieved, as shown in Table 3.6-5.

     Wet scrubbers—Wet scrubbers  remove dust from  a gas stream
by collecting it with a suitable liquid.   A good wet  scrubber can
effect intimate contact between the gas stream and liquid for the
purpose of transferring suspended particulate matter  from the gas
to the  liquid.   Collection  efficiency,  dust-particle size,  and
pressure  drop  are  closely  related  in  the operation  of a  wet
scrubber.  The  required  operating  pressure  drop  varies inversely
with dust-particle size at a given collection efficiency; or, for

                              3.6-13

-------
        TABLE 3.6-4.   EXAMPLE OF AN ESP APPLICATION ON A
                       COAL-FIRED BOILER6
Plant name
Heskett Unit 1
Location
Boiler capacity
Boiler type
Fuel type
     Sulfur content
     Ash content
     Moisture content
Flue Gas
  Temperature
  Volume
  Velocity in ESP
Collecting surface3
Inlet  loading
Outlet loading
Efficiency
Mandan, North Dakota
25 MW  (output); about 75 MW input
Spreader stoker
Coal,  lignite
0.3 to 1.4 %
6.7 %
36.1 %

214°C  (418°F)
180,000 Nm3/h  (189,000  acfm)
1.16 m/s  (3.8  ft/s)
6180 m2  (66,500  ft2)
5.7  to 9.4  g/m3  (2.5  to 4.1  gr/ft3)
0.023  g/m3  (0.01 gr/ft3)
99.68% (designed for  99.45%)
 a Specific collection area is  1154  m2/1000  irfVmin (352  ft /1000
   acfm).
                                3.6-14

-------
      TABLE 3.6-5.  EXAMPLE BAGHOUSE DESIGN PARAMETERS AND
                       OVERALL EFFICIENCY6
Plant name
Location

Boiler type

Boiler capacity


Gas flow rate

Compartments

Fabric

Air-to-cloth ratio,
 6 compartments

Air-to-cloth ratio,
 5 compartments
   (cleaning or maintenance
    on one compartment)

Pressure drop (normal)

Pressure drop (cleaning)
                  Nucla Station
                  Nucla, Colorado

                  Spreader stoker

                  12.65 MW (one of three boilers);
                  about 40 MW input

                  146,000 m3/h (86,240 acfm)

                  6

                  Fiberglass with graphite finish

                  51 m3/h per m2 (2.8 acfm/ft2)
                  61 m3/h per m2 (3.35 acfm/ft2)
                  1120 Pa (4.5 in. H20)

                  Up to 1490 Pa (6 in. H20)
     Load,
      MW
Overall efficiency,
Total particulate emission,
     ng/J (lb/106 Btu)
       6

      11

      12
       99.98

       99.97

       99.92
      3.0   (0.007)

      8.6   (0.02)

     17.2   (0.04)
                               3.6-15

-------
a  given   dust-particle   size,   collection  efficiency   increases  as  operating
pressure  drop increases.  Although scrubbers are not widely used  on boilers,
they  can achieve collection efficiencies in the range  of  80  to  99  percent, as
                      C
shown in Table 3.6-6.

3.6.3.2  Sulfur Oxides-- ••
     Sulfur  oxides  can  be controlled by burning  low-sulfur  fuel  (naturally
occurring  or pretreated for  sulfur  removal) or by removing the sulfur  oxides
from  the combustion gases before  they are released (flue gas  desulfurization,
FGD).   The  pretreating  techniques include physical  and chemical coal  clean-
ing,  coal  liquefaction  and coal gasification.  Physical coal  cleaning  is  limited
to  removal of  pyritic sulfur  and can achieve  up to 50  percent SO2  removal.
Chemical  coal  cleaning,  gasification,  and  liquefaction  are not  in  commercial
operation  at this time.  FGD technologies include:

      1.    Injection of  materials such as  limestone or dolomite
           into the furnace.
      2.    Wet scrubbing of flue gases.
      3.    Use of dry  sorbent  systems.
 Of these  processes, only wet  scrubbing is  widely used  with industrial boilers.

      Wet_scrubbinq of  flue gases-Wet   scrubbing  FGD  systems  applied  on
 coal-fired  industrial  boilers   in the United  States include lime or  limestone
 scrubbing,  the  dilute  or concentrated  double alkali process, and  scrubbing
 with sodium carbonate  or sodium hydroxide solutions.8   In  these systems, flue
 gases  contact the scrubbing solution  after fly ash  is  removed in an  ESP  or
 prescrubber.   Sulfur  dioxide reacts with the slurry in a scrubber-absorber
 and  forms large quantities of sludge.  The  sludge is  separated and disposed
 of,  while  the scrubbing  solution is recirculated  with the  makeup  slurry.
 These  FGD  units  have  achieved  efficiencies  above  90  percent,  as  shown  in
 Table 3.6-7,  but  long-term  monitoring  data  are lacking.   Many  other sulfur
 oxide removal systems  are   used  on  large utility  boilers and  on  industrial
 oil-fired  boilers.
                                3.6-16

-------
     TABLE 3.6-6.  EXAMPLE OF A WET SCRUBBER APPLICATION FOR
           PARTICULATE REMOVAL ON A COAL-FIRED BOILER6
Plant name
Lewis and Clark Station
Montana-Dakota Utilities
Boiler capacity
Boiler type
Fuel type
     Sulfur content
     Ash content
Wet scrubber
     Design
     Vendor
L/G ratio, inlet
L/G ratio, outlet
Pressure drop
Open or closed loop
Electric power requirement
Particulate removal
 efficiency
S02 removal efficiency
55 MW; about 170 MW input
Pulverized coal-fired
Coal, lignite
0.45 %
9.0 %

Venturi, flooded disc
Research-Cottrell
1.755 liters/Nm3 (13 gal/1000 ft3)
2.295 liters/Nm3 (17 gal/1000 ft3)
3235 Pa (13 in. H2O)
Closed
0.5 MW

98%  [0.064 g/m3 (0.028 gr/scfd)]
15%  (minimum) at 0.45% sulfur
content
                               3.6-17

-------
                   TABLE  3.6-7.   EXAMPLES  OF FGD APPLICATIONS ON  INDUSTRIAL  BOILERS8
u>
•
CTi
I
I—'
CO
Plant name,
location

Caterpiller
Tractor,
Joliet, 111.
FMC Soda Ash
Green River,
Wyo.
General Motors
Parma, Ohio
Rickenbacker AF
Columbus, Ohio


Kerr-McGee
Chem. Corp.
Trona, Calif.
(under con-
struction)

No. of
boilers

2


2

4
B 7


2



Total boiler
capacities,
MW
18


200

32
20


64


Total gas 1
flow rates,
Nm3/h
152,500


705,600
@160°C

217,600
84,750


776,200
@160°C


Fuel,
(% s)

Coal
(3.2)


Coal
(1)

Coal
(2.5)
Coal
(3.6)


Coke, coal
or oil
(0.5-5)


FGD system
and vendor

Double alkali
(dilute), Zurn
Industries

Sodium scrub-
bing , FMC

Double alkali
(dilute), GM
Environmental
Limestone
scrubbing,
Research-
Cottrell/
BAHCO
Sodium scrub-
bing, C.E.A.


S02 removal
efficiency, /

90+


87-94

90
90


98+



             Note-   The S02 removal efficiencies shown were obtained during specific test periods
                    test periods and are not  long-term averages.

-------
 3.6.3.3   Nitrogen  Oxides--
      Techniques  for  controlling  NO   from  large  oil-fired  and
                                     X
 pulverized-coal-fired  boilers  include  flame temperature modera-
 tion  by  two-stage  combustion, low-excess-air firing,  and furnace
 modifications  such  as  recirculation  of flue  gas  through  the
 combustion  zone.   These approaches involve reduction  of peak  gas
 temperatures  and  changes  in  the  time-temperature conditions of
            g
 combustion.
      Control  of  NOx  by  scrubbing  techniques  is  under  active
 investigation, but has  not yet been used  on industrial boilers in
 the United  States.10

 3.6.4  Emission Limits
 3.6.4.1   State Implementation  Plan  (SIP)  Limits—
     Most  state  regulations  include  emission  limitations   for
 industrial  boilers  under  general  regulations  for  fuel  burning
 equipment.   Emission  limits  vary  with  the  type  of  fuel  fired
 (coal,  oil, or  other)  and  with  boiler  size,  but they do not
 usually specify boiler  type.
     Particulate emission regulations are commonly set at 43 ng/J
 (0.1  lb/10   Btu),  the  most  stringent limitations  are imposed by
 the District of Columbia (coal) and New Mexico (oil) as follows:

          District of Columbia (see Figure 3.6-4)  —Range is
          about 16 to 24 ng/J  (0.04 to 0.06 lb/106  Btu).
          New Mexico-- 2.7 ng/J (0.006 lb/106 Btu)  on the basis
          of 7000 hours per year.
     Sulfur  dioxide  emission  regulations are frequently set  at
516 ng/J (1.2 lb/106 Btu) for coal-fired boilers;  the most strin-
gent  limitations are imposed  by  the  State of Wyoming and  Clark
County,  Nevada,  as  follows:

          Wyoming - 86 ng/J (0.2 lb/106 Btu).
          Clark County,  Nevada-- 65 ng/J (0.15 lb/106 Btu).

                              3.6-19

-------
                                             BOILER HEAT  INPUT, MW


                                        2.93   	29.3
                                                                       293
LO
CTi
I
          IS)
          •z.
          O
CO

o


-------
     Nitrogen oxide emission regulations are generally set at 301
and 129 ng/J  (0.7  and 0.3 lb/10  Btu)  for  coal  and oil,  respec-
tively.  The  most  stringent  regulation  is imposed by New York
City:   103  ng/J (0.24 lb/106 Btu) for both coal and oil.

3.6.4.2  New Source Performance Standards--
     Federal regulations  of particulate,  SO«,  and NO  emissions
                                            £         X
apply to all types of coal- and oil-fired boilers with heat input
greater than 73 MW (250  x 10  Btu/h) .  The following limitations
apply to all sources that were built after August 17, 1971:

          Particulate—43  ng/J  (0.1  lb/106 Btu);  SO2—520  ng/J
          (1.2 lb/106 Btu) for coal and 344 ng/J (0.8 lb/106 Btu)
          for oil;  NO --301 ng/J  (0.7  lb/106 Btu)  for coal  and
                      X
          129 ng/J (0.3 lb/106 Btu) for oil.

     Revisions that  would reduce  these emission limits  are cur-
rently  being studied,  but no  definite  values  have  been prom-
ulgated.

3.6.4.3  Achieved-in-Practice (AIP) Limits--
     Data on particulate  emissions from industrial-sized  boilers
are often  obtained under  constant operating  conditions with all
combustion and control system parameters  set at a level to mini-
mize emissions.  Under these  conditions,  emissions as low as 6.9
    3
mg/m   (0.003  gr/dscf) were measured at  the  Caterpillar  Tractor
Co.,  in Decatur,   Illinois, utilizing a  fabric  filter system.
This rate is  approximately equal to 2.6 ng/J (0.006 lb/10  Btu).
Emissions with fabric filter systems and high-efficiency electro-
static  precipitators  range more typically  from 13  to  17.2  ng/J
(0.03 to 0.04 lb/106  Btu) range.  Emissions from distillate-oil-
fired  boilers  are  about  5.6  ng/J (0.013  lb/  106  Btu)  with no
control.  Well-operated residual-oil-fired  boilers emit particu-
lates in the range of 21  to 43 ng/J (0.05 to 0.1 lb/106 Btu) with
no controls.
                              3.6-21

-------
     Sulfur dioxide  emissions  vary  directly with  fuel  sulfur  content.   The
coal -fired  boiler  achieving best  control  is  the FMC plant  in  Green River,
Wyoming,  which reportedly  has measured  up to 94 percent SO2 removal while
burning a  1  percent sulfur content coal (see Table 3.6-7).   This 3-stage tray
scrubber  utilizes  sodium carbonate at liquid-to-gas ratios of  1.3 to 2  liters/m
(10 to 15  g/1000 ft3) and  a pressure  drop of 2.8 to 3.2 cm (7 to 8 in.) of
water.
     Lowest  achievable  NOX   emissions  are .not  well  documented,  but EPA
programs  to acquire NOX emission data for large industrial boilers are sched-
uled for completion during 1979.

3.6.5  Determination of LAER
      The   recommended  limitations are based  on  SIP's  and on  performance
information available in early 1979. It is anticipated that several additional SIP
regulations covering these  sources will be promulgated and/or modified in 1979
and  1980  and  that appreciable new  performance data will become  available  in
the  near  term.   Conceivably, some  SIP  regulations may be  more  stringent
than  the  LAER   suggested herein.    Furthermore,  performance  testing  may
 show  that  more  stringent  limits  than those suggested  are  feasible  or  it may
 show  that the suggested limits are appropriate or that they are not achievable
 for  some specific  subcategories.  In  any case, the basis for  determining LAER
 for  many  source categories is  expected  to change frequently.   Since LAER  is
 near  the  vanguard of control  technology,  a more detailed  analysis  is  partic-
 ularly necessary  when  addressing modified or reconstructed facilities  subject
 to the provisions  of  Section  173  of the Clean  Air  Act.  Emission  limitations
 reasonable for new sources  may  in some  instances  be economically or tech-
 nically unreasonable when applied to  modified or reconstructed sources of the
 same type.
       The following discussion pertains to coal-fired boilers in the  58 to 175 MW
 (200  to 600 x 106 Btu/h)  heat input  range including stokers  of 88 to 117 MW
 (300  to 400 x 106 Btu/h)  and to oil-fired boilers  in the range  of  73 to 102 MW
 (250 to 350 x 106 Btu/h) heat input.  Based on particulate  control efficiencies
 of  high-efficiency fabric  filter and  ESP  systems,  an emission  limit  of  13 ng/J
                               3.6-22

-------
           £
(0.03 lb/10  Btu) can  be achieved.   This level of particulate emission will also
require  controls on some  residual-oil-fired  boilers depending on  fuel compo-
sition  and  firing  efficiency.   Control  efficiencies  in the range of 40  to  70
percent  will be required.
     A  sulfur dioxide emission reduction  of 90 percent  can  be achieved  on
coal-fired  units  by  use of  FGD.   The  economic feasibility  of using FGD  on
smaller  boilers, especially those that already burn low-sulfur fuel,  is question-
able.
     Nitrogen  oxide  emissions from  packaged oil-fired boilers can be limited to
130 ng/J (0.3 lb/106 Btu).  With pulverized-coal-fired  boilers,  a  limit  of  260
               £•
ng/J  (0.6  lb/10  Btu) is achievable;  this value is based on studies performed
for NSPS revisions.   A  limit of 2.7 ng/J (0.5 lb/106  Btu) can be achieved  by
subbituminous coal-fired boilers.   NC>  levels for stoker  fired  boilers  will  be
                                      X
determined  by  a  current  study on industrial boilers.   The  above limits  for
SO2 and NOx are on a 30 day average basis  using a continuous monitor,  while
the particulate  level is based upon the average  of  three or more  runs  using
EPA Method  5.
                             3.6-23

-------
                         REFERENCES


1.  Smith,  W.  S.,  and C.  W.  Gruber.  Atmospheric Emissions from
    Coal Combustion - An Inventory Guide.   Public Health Service
    Publication No. 999-AP-24,  U.S. Department of Health,  Edu-
    cation, and Welfare.   April 1966.

2.  The Babcock & Wilcox Co.  Steam/Its Generation and Use,  38th
    edition.  1975.

3.  Smith,  W.  S.  Atmospheric Emissions from Fuel Oil Combustion
    - An Inventory Guide.  Public Health Service Publication No.
    999-AP-2,  U.S. Department of Health, Education, and Welfare.
    November 1962.

4.  Weisburd,  M.  I., A. Stein,  R. J. Bryan, L. G. Wayne, and A.
    Kokim.   Air Pollution Control Field Operations Manual, A
    Guide to Inspection and Enforcement.  Vol. II, U.S. Environ-
    mental Protection Agency Contract No..  CPA 70-122, Task Order
    1.  February  1972.

5.  Compilation of Air Pollutant Emission Factors, 2nd Edition.
    Publication No. AP-42, U.S. Environmental Protection Agency.
    February 1976.

6.  Szabo,  M. F.,  and R. W. Gerstle.  Operation and Maintenance
    of Particulate Control Devices on Coal-fired Utility
    Boilers.  EPA-600/2-77-129, U.S. Environmental Protection
    Agency.  July 1977.

7.  Danielson, J.A., ed.  Air Pollution Engineering Manual, 2nd
    Edition.  AP-40, U.S. Environmental Protection Agency.  May
    1973.

8.  'Tuttle, J., A. Patkar,  and N.  Gregory.  Environmental Pro-
    tection Agency Industrial Boiler FGD Survey:   First Quarter
    1978.   EPA  600/7-78-052a.  March 1978.

9.  Air Quality Criteria  for Nitrogen Oxides, Air  Pollution
    Control Office.  AP-84, U.S. Environmental Protection
    Agency.  January 1971.
                              3.6-24

-------
10.   Ando,  J.   Status of S02 and NO  Removal Systems in Japan.
     Presented at FGD Symposium,  Hollywood,  Florida.  November
     8-11,  1977.

11.   Air Pollution Emission Test - Caterpillar Tractor Co.
     Project 77-SPP-18,  Emission Measurement Branch, U.S.
     Environmental Protection Agency,  Research Triangle Park,
     North Carolina.  April 1977.
                              3.6-25

-------
3.7  MAJOR SOURCE  CATEGORY:   PRIMARY ALUMINUM PLANT REDUCTION
     CELLS AND ANODE  BAKE  OVENS--SULFUR DIOXIDE EMISSIONS
3.7.1  Process Descriptions
3.7.1.1  Primary Aluminum  Plant--
     The  base  ore for  primary aluminum production  is  bauxite,  a hydrated
oxide of  alumina  consisting  of  30 to  70 percent alumina (Al-CO,  and lesser
                                                            c*  O
amounts  of iron, silicon, and titanium.  The bauxite ore is  purified to alumina
by  the Bayer process and then  transported to the primary  aluminum reduction
plant.
     At the  reduction plant  the alumina  is electrolytically  reduced to metallic
aluminum  in  a  bath  of  molten  cryolite  (the  electrolyte) by the  Hall-Heroult
process.    The  heavier  molten  aluminum  settles  beneath the  cryolite.  It is
periodically  decanted, transferred by crucible to holding furnaces, and  then
cast into  ingots,  billets,  slabs, and   other  bulk  shapes  for  shipment to
customers.
     Figure 3.7-1  is a schematic of a  primary aluminum  reduction plant.

3.7.1.2  Reduction Cell—
     The  electrolytic reduction  of alumina takes  place  in shallow, rectangular
carbon-lined  steel shells (pots)  arranged  in series  to  form a  "pot line."
Cryolite   serves as  both the  electrolyte and the  solvent for the  alumina.
Carbon blocks  (anodes) are suspended  in the  pots (cathodes),  and  the  two
are  connected  electrically to accomplish the  electrolytic reaction.   The  heat
generated  by  electrical  resistance  to  high-amperage,  low-voltage direct
current applied across  the  electrodes creates operating temperatures  between
950° and  1000°C (1728°  and   1832°F).   The carbon anodes are  depleted by the
reaction  of  oxygen formed  in  the process  (2A12O3  -»  4A1  + 2O3)  with anode
carbon  (2C  +  Cs  -> CO  + CO-).  Because the process  is continuous,  both the
                              3.7-1

-------
 POWER PLANT
RECEIVE, RECTIFY
DISTRIBUTE POWER
V- «/ a T=s
_J -f^"®


/ t
ALUMINA CR
Y^
POT CHARC
MATERIAL!
1
RAW MATERIALS
RECEIVE, STORE
DISTRIBUTE

/ ^ X ,
(PITCH I /COKE ]
YOLITE """-^-^-"^
ANODE PLANT
. ./ GRIND, MIX
y^ '. PITCH/COKE
T , ,
JE JANODE PASTE

OR

CATHODE PLANT
BUILD CARBON
LINED STEEL POT
1
CATHODE CELL
(POT)
\
TO
| SODERBERG ] ] PREBAKE | POT HOUSE

|
TO MOLD/B
POT HOUSE ANOD

AKE
ES
TO
POT HOUSE
            rau
                     POT HOUSE
REFINE, ALLOY,
CAST METAL


SHIP
PRODUCT
      Figure  3.7-1.   Schematic of  primary  aluminum
                      reduction plant.
                              3.7-2

-------
anodes  and   the  cryolite-alumina  bath  components  that  are  consumed  or
removed in the process must  be  replenished periodically.
     Aluminum reduction  cells  are  of  two main  types,  prebake  (PB)  and
Soderberg.   Soderberg cells are designated according to the manner of mount-
ing the stud  in the  carbon anodes:  vertical stud  Soderberg  (VSS) or hori-
zontal  stud Soderberg (HSS).  The prebake and  Soderberg  processes differ
in the preparation of the anodes.   In the  Soderberg process the anode paste
mixture is formed (baked)  in  place by the pot heat.   In the prebake process,
as the  name  implies,  the anode is baked  (usually in a  facility separate from
the pot room) before it  is inserted into the pot.   Although it requires more
electrical  power,  the Soderberg cell  has been favored  by the  industry because
no  separate   facility is  needed  for  manufacture of  anodes.    The  trend in
recently constructed plants,  however, is to prebake  anodes.   One reason for
this  change  in  preference is the lower power requirements  of the PB cell;
another is that Soderberg cells  generate  volatile  pitch vapors  that must be
captured  and  treated,  often  leading to plugging  of  ductwork  and  control
devices by organic condensibles.

3.7.1.3  Prebake Anode Manufacturing--
     Pitch and  petroleum  coke,  including  recycled  anode butts, are  mixed
with a pitch  binder to form a paste used in the cathodes (pot  liners) and the
"green" anodes  for  prebake cells.   The approximate  blend is  75 percent coke
and  25  percent  pitch  binder.    Anode  making  at  the  "green  mill"   (paste
preparation plant)  includes  crushing,  grinding,  screening,  and sizing  the
coke, then  blending the  sized  coke fractions  with binder in  heated mixers.
     For prebake  anodes the  paste is transferred to molds and densified by a
hydraulic press  or  by  mechanical  vibration.    The  green anodes are then
baked in furnaces to develop  thermal stability,  strength, and electrical  conduc-
tance properties.    Anodes are  generally  baked  in  a  series  of  sunken  pits
(called ring-type furnaces) served by a flue system that circulates hot  combus-
tion  gases from  the  heated   pit  through  preceding  sections  to  preheat  the
anodes.   Anodes are packed  into  the pits,  and a layer  of coke  is placed over
and around them.   The pits  are fired with gas or oil through  mobile manifold
                             3.7-3

-------
burners  at a  temperature of  approximately  1200°C  (2190°F).  The complete
cycle—charging  of  pits,  preheating,  firing,  cooling,  and anode  removal-
requires approximately 28  days.
     A  recent  development is use of an indirect-fired tunnel kiln fitted with
air locks  and an inert atmosphere to preclude  oxidation of the carbon anodes.
This  system is  more  complex and  is  subject to mechanical problems.   The
advantages are  a  shorter  and  more  uniform baking  cycle,  reduced space
requirements,  and recycling  of hydrocarbon emissions to  the firebox as a fuel
supplement.
     The  final step is to  fit the  baked anodes with a metal rod yoke assembly
that supports  the  anodes  in  the  reduction cell and provides electrical conduc-
tivity.  The baked anodes are airblasted or brushed to remove surface fines;
then  the  rod  yoke assembly is  mated to the anodes and  cemented in place,
usually with molten iron.

3.7.2  Emissions
3.7.2.1   Prebake Anode Manufacturing Emissions—
     Anode paste preparation emissions--Material  handling operations in anode
paste  preparation  generate  airborne particulate  matter  (coke dust).   Spent
anode butts recycled to the plant have surface  deposits of pot materials that
can be a source of coarse  particulate  fluorides in the regrinding and mixing
process.   Small amounts  of  volatile hydrocarbons are released  during paste
mixing.   Paste preparation generates no sulfur dioxide  (SO2).
      Anode bake oven emissions—Materials   entering  the  anode  bake  oven
 (furnace  or kiln)  consist of  the green anodes,  coke or  anthracite packing,
 and combustion fuel,  either natural gas or  oil.   Emissions  from the  bake oven
 include  the products of fuel combustion;  burned and unburned  hydrocarbons
 consisting principally of  high-boiling-point organics formed  by the cracking,
 distillation,  and oxidation of the paste binder pitch; sulfur dioxide from the
 carbon   paste;  fluorides  from  recycled anode  butts;  and  other  particulate
 matter.
      Sulfur dioxide emissions  result from oxidation of sulfur contained in the
 raw  materials  used in anode  manufacture:   high-grade coke  (petroleum and
                               3.7-4

-------
pitch coke) and pitch.   Before calcining, the coke portion of the anode paste
has  already been subjected at the refinery  coke operation to temperatures  as
high  as  that  in  the  bake oven or higher; therefore,  the SO- emissions in the
bake exhaust are primarily from the sulfur  in the pitch (0.5  percent sulfur)
and  from  combustion  gases.   Data  pertaining  to  SO-  emissions  from anode
bake stacks  are very  sparse.   Uncontrolled  emissions  in the  exhaust are
reported as follows:   5 to 47 ppm;   0.7  to  2  kg SO0/Mg  aluminum produced*
                         2
(1.4  Ib  to  4 Ib  SO9/ton).   Results  of source  tests reported by the National
                  £i
Emission Data System  (NEDS) indicate that emissions range from 0.09 to  1.7
kg/Mg (0.18 to 3.4 Ib/ton).

3.7.2.2  Reduction Cell Emissions--
     Emissions from  the reduction cell  include  (1) particulates from  the peri-
odic   addition  of alumina  and cryolite  and  from  condensation  of  vaporized
materials at the bath and anode surfaces;

(2) carbon  monoxide and  carbon dioxide  from oxidation of carbon anodes; (3)
organics (tar  fog)   from  volatilization  of the  anode  materials by  the high-
temperature  bath  in  the  cell;  and  (4) oxides  of  sulfur  from the  anode
materials.  Emissions of organics and of  SO2 are greater from  Soderberg cells
than  from  prebaked  cells  because the lower-boiling-point  organics  and some
sulfur are  driven off in the anode prebaking oven.   In  all types of  reduction
cells  the  "primary"  emissions are those  captured  by  the pot hood exhaust
system  and conveyed to a control device.   The "secondary"  (roof)  emissions
are those that escape the exhaust system and exit through the  roof  monitors.
     Few  measurements  of  SO-   emissions  from reduction cells  have been
reported.   One  report  indicates up  to  80 ppm  in the exhaust from Soderberg
plants.   Reference  2  reports  "sulfur  dioxide data ranged from 5 ppm for a
prebake  plant to 80 ppm for a  vertical stud  Soderberg  plant.   No  sulfur
  Emission values  given throughout this subsection are  in terms
  of the quantity  of aluminum produced; e.g.,  the notation kg/Mg
  denotes kilograms SCL per megagram of aluminum produced.
                             3.7-5

-------
dioxide  data were  obtained on  a  horizontal  stud  Soderberg  plant."   Values
reported for VSS plants include  a range of 200 to 300 ppm  (basis of 2 percent
sulfur)  or  17.5 to  25  kg/Mg (35 to  50  lb/ton),3  and a  200-ppm  average.
Reference 2  reports  emissions from  prebake  plants  in  the  range of 30 kg/Mg
(60 lb/ton)  on the basis of 3  percent  sulfur in coke, and Reference 3 gives a
range  of 20  to  30  ppm  SO2.   NEDS data  from  source  tests at  PB plants
indicate  a range of 20.9 to 23.4 kg/Mg (41.7 to 46.8  lb/ton)  and an average
of 22.4 kg/Mg (44.8 lb/ton).   NEDS gives no data  on VSS  or HSS facilities,
and no  data on HSS  plants were discovered.
3.7.3  Control Measures for SO2

     The two  methods available  for reducing SO2  emissions from  anode bake
ovens  and  reduction cells  at  a  primary aluminum plant are (1) flue  gas  desul-
furization  (FGD)  systems  that  remove SO2  from the  exhaust  stream and  (2)
limitations on the  sulfur  content of coke used in anode manufacture.

3.7.3.1  Flue Gas Desulfurization  (FGD)--
     Use of  FGD  systems  has  been  associated primarily  with the  relatively
strong  concentrations  of SC>2  in combustion gases from fossil-fuel-fired boilers
at  utility  plants.    To a lesser extent,  FGD has  been applied  to  industrial
combustion   and process  sources.   Because  most applications  have been  for
control  of  significant  SO2 concentrations, relatively little  information is  avail-
able  regarding efficiency or  operation of FGD on weak SO2 streams similar to
 those from reduction  cells  and  anode  bake ovens.   Even though  FGD  systems
 designed specifically  for  SO2  control have  not been demonstrated  at these
 primary aluminum  plant operations,  technology transfer is  possible.
      Use of  FGD  systems  has  been  successful on the exhausts  from  boilers
 fired  with  low-sulfur  oil  and  gas,  from foundry cupolas  using low-sulfur
 coke,  and  from  certain  industrial  processes.5'6   The  SO2  concentrations in
 the exhaust streams from such sources approximate those from reduction cells
 and  anode bake ovens.   Collection  efficiencies cited in Reference  5 range from
 70  to  98  percent  for boilers,  47  to 99 percent  for cupolas,  and  50 to 90
 percent  for  industrial processes.   As regards SO2  control  at a  VSS  plant,
                               3.7-6

-------
Reference 6 cites measured efficiencies  of 53 to 83 percent for systems consid-
ered to be  technology transfer  candidates and indicates a design efficiency of
95 percent  as achievable for a  suggested hypothetical pot room  SCU control
system.   The  authors  conclude that  the  greatest degree of  control  being
adequately demonstrated  by technology  transfer  candidates  is approximately 80
to 85  percent.   Considering  the  uncertainties,  they  believe  that  70 percent
SO-  collection efficiency is  achievable by  transfer of technology  at  a VSS
plant.
     In the potential  application of  FGD  to  control  pot room SO2 emissions at
a  primary  aluminum  plant,  it  is  important  to  recognize  the  differences in
operation  and in pot  room exhaust volumes associated  with  VSS, HSS, and PB
plants.  The primary system exhaust volumes  at PB and HSS  plants  are
generally  higher than those  at  VSS  plants by 6 to 7  times and 8 to 10 times,
respectively.   Accordingly the  larger  volumes  and weaker concentrations at
HSS  and PB plants would reguire proportionally larger  control system capacity
for anodes  of a  given sulfur content.   Additionally, the anticipated trend to
prebaked  cell  plants  could require  provision of  SO-  control on  anode  bake
oven emissions either separately or as  added  primary  control system capacity.
     As  an  alternative  to  use  of  two systems--a dry fluoride/  particulate
control followed by an FGD system--it  would  seem technically feasible to use a
wet control system to capture both  contaminants.   Reference 6  reports that a
wet scrubber/wet electrostatic precipitator system installed  on a VSS  plant for
purposes  of fluoride  and particulate control  collects approximately  70 percent
of the  SO9  emissions  from the primary  system.   The application of any wet
          £*
system would  require provision of  treatment facilities to  satisfy  wastewater
discharge requirements.
     As  discussed in Section 1,  the  determination of LAER   is  primarily  a
technology-oriented  consideration,   in  which economics   is  relegated   to  a
relatively  minor  role.   As  regards  the  economic impact of  environmental
controls  in the  primary aluminum  industry,  it has been  stated that controls
(associated  with  fluorides) more stringent than  the NSPS will tend  to discrim-
inate  against the small  market entrant that is without existing  capacity, tend
to encourage a greater proportion of imports  of primary ingot,  and encourage
                              3.7-7

-------
higher  domestic  prices  for  primary   aluminum  and  fabricated  products.
Although  these  observations  refer  to the  imposition of an NSPS  standard for
fluorides,  they  should  also  be  considered when  developing  LAER standards
for S02 where  the control costs are substantial.   Substantial costs for SO2
control could result where the fluoride control  system  selected  to  meet the
NSPS  requirement  would not also provide  the required  SO2 control.    Sub-
stantial costs could also be associated with wastewater treatment.

3.7.3.2  Limitation of Sulfur in Petroleum Coke--
      Sulfur  dioxide emissions from anode  bake ovens and reduction cells stem
from  the sulfur content of the  coke (usually petroleum coke)  and the coal tar
pitch  binder.   The emissions relate  directly to the amount  of sulfur in  these
raw materials,  much  as the  SO2  emissions from  fuel  combustion  relate to the
sulfur  content  of fossil  fuels.   Limitation  of  sulfur  content,  as  is  often
applied to fuels,  can  be similarly  applied  to anode materials.
      Good quality feedstocks for  anode coke  include  thermal  tar,  cat cracker
 slurry,  decanted oil,  and  coal  tar pitch.   Poor feedstocks include vacuum
 residuals  and  derivatives from high-sulfur  crudes.7  To some  extent  sulfur
 can be removed from coke by calcination.   If  the coke is to be used  for  anode
 manufacturing,  however,  caution must be exercised.   The  usual calcination
 temperature  (1370° to  1425°C,  or 2500° to  2600°F) has  no  deleterious effect,
 but  high-temperature calcining (at 1590°  to  1650'C,  or 2900°  to  3000'F)  tends
 to cause  expansion  ("popcorning")  of  coke and  to  reduce  its density.   This
 effect is  just the  opposite  of  the   intended  purpose-densification.   The
 expanded  coke  requires  addition  of   more  pitch binder at the bake  step.
 Moreover, the baked  anode becomes porous  and brittle, with the deleterious
 effect of' increased  electrical resistivity  and  an attendant power requirement
 penalty.8
       Pitch  ordinarily  contains  about 0.5  percent  sulfur;  petroleum coke
 usually contains 2.5 to 5 percent sulfur but  may have  as little  as 1.5 percent
 and  as  much  as  7  percent.8   The  sulfur  content of  coke  depends  on  the
 crude stock and  the  sulfur distribution in the  crude,  i.e.,  the tendency of
 sulfur  to concentrate  or not  to  concentrate  in the bottoms and thus  in  the
 coke.
                               3.7-8

-------
Marketing  factors  related  to  availability  of low-sulfur  coke are  extremely
complicated and  uncertain,  and  the  trend  appears to be  toward  coke with
higher  sulfur  content;  however,  it is reported that for the near term, coke
containing 3 percent sulfur or less should be available.   One petroleum coke
manufacturer indicates  that very  little low-sulfur coke (2.5 percent or less) is
now available,  that 4 to 5 percent coke is  abundant, and  that low-sulfur coke
(when available) commands  a  price  four  to five times that of the high-sulfur
coke.

3.7.4  Emission Limits
3.7.4.1  New Source Performance Standards  (NSPS) Limits-
     No NSPS limitations are applicable to SO? emissions from anode  bake oven
and electrolytic reduction cell facilities at primary aluminum plants.

3.7.4.2  State  Implementation Plan (SIP)  Limits--
     No SIP contains SC^ emission limitations that pertain  specifically to anode
bake oven or  electrolytic reduction  cell  operations.   Such  limitations  are
embodied  in general regulations  intended to limit SO2  emissions  from process
sources.   The  most frequent  SIP limitation on process sources limits  emissions
to  no greater than 500  ppm S02 on  a volume  basis.   The  most  stringent
limitation  is Regulation No.  2, Sec.  3122 of the Bay Area Air Pollution Control
District,  San Francisco,  which  provides that SG>2 emissions shall not  exceed
300  ppm.   Because the  volumes  of exhaust from VSS,  HSS, and PB  plants are
variable,  it is  difficult  to judge  compliance with a 300-ppm limitation.  Limited
stack  test  data  indicate,  however, that  SO9 concentrations in uncontrolled
                                             c.t
exhaust from reduction  cells  and  anode bake ovens  remain below  this level
without the  use  of control equipment specifically designed for SO2 removal.

3.7.4.3 Achieved-in-Practice  (AIP)  Limits—
     As  mentioned earlier,  no  FGD  systems specifically designed for SO,,
control  on HSS,  VSS,  and  PB pot  lines  have been installed at primary alum-
inum plants  in  the United  States.  At  a VSS plant the wet systems used to
control  particulate  and  fluoride  reportedly also  capture  70 percent  of the SO?
                              3.7-9

-------
in the primary  exhaust  and 45  percent of the SO2 in the secondary  (roof)
system.   (Emissions are  approximately  6.3  kg/Mg  or 12.6 Ib/ton,  with a basis
of 2 percent sulfur coke.  )
     One  aluminum producer reportedly will install FGD systems  at  two VSS
plants.3'6  The  proposed  system will  consist  of  a wet  scrubber following  a
dry  fluoride/particulate control  device  serving  the pot exhaust and an exist-
ing wet scrubber for secondary  (roof)  emissions.   With  changes  in  cell tech-
nology,  95 percent  capture and transport of cell emissions  to  the primary
system (5  percent  to  secondary   system),   and  attentive  cell operational
procedures,  it is expected  that overall  SO2 control efficiency  will be about 84
percent.   Resultant  SO2  emissions,  based  on use  of 2.8 percent sulfur coke
(and  an  alumina sulfur content of  0.045  percent), are projected  to  be 4.1
kg/Mg (8.1 Ib/ton).    The  installation  is  in response  to a  PSD-BACT deter-
mination that gave approval subject  to  an  emission limitation of 9.5  kg/Mg (19
Ib SO2/ton).6
      No achieved-in-practice  information was discovered relative to ancillary
control of  SO2  by  wet  fluoride/particulate control systems  on   PB, HSS, or
anode bake  plants.   Several PSD-BACT review  determinations  on  PB  plants
have  been  approved with  an SO2 limitation based  on 3 percent  sulfur in the
coke.  In the absence of other specific information,  the achieved-in-practice
limitations are those  values given in  Section 3.7.2, Emissions.
      In  summary,  for  both anode bake ovens and reduction cells, the  applied
 controls  based  on wet methods were  designed and installed  for the primary
 purpose  of controlling  fluoride  and  particulates;  reduction  of  SO2 emissions  is
 generally  secondary  and incidental;  emission measurements are oriented toward
 fluoride/particulates; and  SO2 emissions are largely unreported.   Because SO2
 emissions are directly  related to sulfur  content of the coke, the best achieved-
 in-practice levels are  usually use of low-sulfur coke rather  than  application  of
 control technology.

 3.7.5  Determination of LAER
       The recommended limitations   are based on  SIP's and on performance
 information  available in  early  1979.   It is anticipated that several  additional
                               3.7-10

-------
SIP regulations  covering these sources will be  promulgated and/or modified in
1979  and 1980 and that  appreciable new performance data will become available
in the near  term.   Conceivably,  some  SIP regulations may be  more stringent
than  the  LAER suggested  herein.   Furthermore,  performance  testing may
show  that more  stringent  limits  than those  suggested are feasible  or  it may
show  that the suggested limits are appropriate  or  that they are not achievable
for some  specific subcategories.  In any case,  the basis for determining LAER
for many source categories is  expected to change freguently.  Since LAER is
near  the  vanguard  of control technology,  a more detailed analysis particularly
is necessary when  addressing modified or  reconstructed facilities subject to
the provisions of Section 173 of the Clean  Air Act.   Emission limitations  reason-
able  for  new  sources  may  in some  instances  be  economically or  technically
unreasonable when  applied to modified or reconstructed sources of  the same
type.
     No  NSPS limitations are applicable  to sulfur dioxide emissions from either
anode bake  ovens  or  reduction  cell  pot lines at  primary aluminum  plants.
Limited  stack  test  data indicate  that  uncontrolled  emissions  are  within the
most  stringent  limitation embodied in a State  Implementation Plan.   For both
anode bake  oven  and reduction cell  emissions,  the  achieved-in-practice levels
are highly variable and depend primarily on the  coke sulfur  content.   Emis-
sion  data are insufficient,  particularly for  anode  bake ovens or HSS  or  PB
cell plants,  to allow definition  with any  degree  of  confidence of a best achieved
-in-practice  level.
      The most practical method of limiting SO2 emissions from primary alumi-
num  plants  is   to  utilize  low-sulfur  materials in  the  anodes.   The   sulfur
content  of coke is directly related to the  guality of  the  crude  from which it is
produced.   Supplies  of 3  percent sulfur coke derived largely from domestic
crude are expected to  be  available for the  next  5  to 10 years.  Since future
supplies  of  coke will  depend  primarily on  foreign crudes, the  future avail-
ability  of low-sulfur  coke  is  uncertain.   Given the  domestic  dependence on
and the  competing  demands  for various petroleum products, it is unadvisable
to  suggest  a single  value that  would  purport to  represent  LAER  when  the
determining  emission parameter—coke sulfur  content—is  uncertain.
                              3.7-11

-------
     It  is  appropriate  to  suggest factors  to  be  considered in deriving LAER
once a  specific case is in hand and  the unknowns  are resolved.   The  fol-
lowing  are therefore presented as general  guidance  to be considered in deter-
mining LAER for SO2 emissions from primary aluminum reduction plants.

3.7.5.1  Sulfur-in-Coke Limitation--
     For  all  plant types  consider a limitation on sulfur content in coke  that
reflects  the  lowest-sulfur coke  available.   The   limitation  should  allow no
greater  than  3  percent  sulfur in coke and  lesser-sulfur  coke should be
required  unless it  can  be adequately  demonstrated that  it  is not available.
Coke blending  should be  considered  where  supplies  of low-sulfur  coke are
available but limited in quantity.

3.7.5.2  Application of  FGD--
     Consider  the application  of FGD  on a case-by-case basis,  [n examining
the  merits of applying FGD technology  to aluminum reduction plant  emissions
the  following  general observations are  worthy of consideration:
     At  this  time FGD systems have not  been applied  to primary  aluminum
plant facilities  for  the sole and specific purpose of SO2 control.  Wet  systems
installed  for  particulate  arid  fluoride  control at a  VSS  plant have,  however,
reportedly  effected a  70  percent SO2 removal  in  the  secondary system.   In
addition, proposed FGD systems at  two  VSS plants are  to  be installed  as  a
result of  BACT  determinations  and are expected  to  provide an  overall SO2
control efficiency of approximately 84  percent.   The use of FGD appears to  be
technically feasible at aluminum reduction plants,  although costs  are expected
to  be high  compared  with those associated  with  FGD  use  at power  plants.
     Because  of  the lower  exhaust  volumes and  higher  SO2 concentrations
resulting from  coke of  a  given  sulfur content,  FGD is more  readily  applicable
to  VSS  than  to  HSS  or  prebake facilities.  The  larger exhaust volumes and
attendant dilute  SO2 concentrations at HSS and  PB plants  would  necessitate
much  larger FGD  system capacity than for a VSS  plant.   New  plants are
 expected to be of the PB type.
                              3.7-12

-------
     Wet systems  used to control  particulates  and fluorides may  be utilized to
achieve SO9  control as well.  Where wet systems cannot be adapted to achieve
           £j
adequate SO2 control,  two systems could be  used to satisfy  LAER and NSPS
limitations.   Depending  on  the  sulfur  content  of the  coke used and  the
resultant  SO2  emission  rate,  consideration could  be  given  to specifying a
LAER  limitation that could be met  by excellent capture and FGD treatment of
primary  emissions  only.   In such  case neither  the secondary emissions from
reduction  cells nor the  bake  plant  emissions would require FGD  treatment.
     Wet scrubber  effluent must  be  handled  in a manner that conforms with
wastewater treatment standards and water quality requirements.
     Costs associated with  installation of FGD systems and  attendant waste-
water  treatment facilities  would exercise  economic  constraints on the  viability
of constructing a  new (modified)  primary aluminum reduction facility.
                              3.7-13

-------
                          REFERENCES
1.   Background Information for Standards of Performance:   Pri-
    mary Aluminum Industry, Volume 1:   Proposed Standards.   EPA
    450/2-74-020a.   U.S.  Environmental Protection Agency,
    Research Triangle Park, North Carolina.  October 1974.

2.   Singmaster and Breyer.  Air Pollution Control in the Alumi-
    num Industry, Volume I.  Contract No. CAP-70-21, Environ-
    mental Protection Agency.  July 23, 1973.

3.   Personal communication with Fred Fenske, Washington State
    Department of Ecology, May 10-11, 1978.

4.   Personal communication with Paul Boys, Region X, U.S.  Envi-
    ronmental Protection Agency, April 13, 1978.

5.   PEDCo Environmental, Inc.  Summary Reports on SC^ Control
    Systems for Industrial Combustion and Process Sources -
    Volumes I and IV.  Industrial Environmental Research Labor-
    atory, U.S. Environmental Protection Agency, Research
    Triangle Park, North Carolina.  December 1977.

6.  Final Determination Analysis Document; Prevention of Signif-
    icant Deterioration, Approval of Modifications  to Martin
    Marietta Aluminum Plants.  Region X, iJ.S. Environmental Pro-
    tection Agency.  January 11,  1977.

7.  Internal Memo -  Office of Air Quality  Planning  and Stan-
    dards, Research  Triangle Park, North Carolina.  Reid Iverson
    to  Gordon Rapier re: Determination  of  BACT  for  S02 Emissions
    from Primary Aluminum  Plant.  December 19,  1977.

8.  Personal  communication with  Richard  Albrecht, Sohio, Cleve-
    land, Ohio,  April  12,  1978.
                              3.7-14

-------
3 8  MAJOR SOURCE CATEGORY:  BULK GASOLINE TERMINALS TRUCK
     LOADING OPERATIONS—VOLATILE ORGANIC COMPOUNDS (VOC)
3.8.1  Process Description
     Gasoline and other liquid petroleum products are distributed
from  the  refinery  to  the  consumer  by  an extensive  network of
pipelines, tank trucks, railroad tank cars, and marine tankers or
barges.   The  bulk gasoline terminal is  an integral part of this
network,  serving as the primary storage and distribution facility
for  a  regional  marketing  area.   Within  this area  it provides
products  to  smaller but  similar  distribution facilities,  called
bulk  plants,  that  serve  smaller,   localized  areas.   In  this
report,  a bulk  terminal  is  defined as  a distribution facility
with  an average  daily gasoline throughput of  greater than 76,000
liters  (-20,000  gallons)  and a bulk plant as  one with an average
daily gasoline throughput of 76,000  liters  or  less.   The bulk
terminal  receives  gasoline  by pipeline,  ship,  rail,  or  barge;
stores  it in tanks; then redistributes  it by tank-truck to bulk
plants, commercial  accounts,  or retail outlets.
      Equipment  and structures  at the typical bulk terminal  in-
clude storage tanks, loading  (unloading)  equipment,  liquid lines,
tank  trucks,  parking  and access roadways, and business offices.
The gasoline  is usually  stored above grade in  floating-roof  tanks
that   have   large   storage  capacities   [generally   greater   than
250,000  liters   (65,000   gallons)  at  new  installations].    The
loading  facility  consists  of  equipment  to  meter  and deliver
gasoline  into tank trucks from the  storage  tanks.   It is located
 at a  central  island  (loading  rack)  accessible to tank trucks.
The   loading  rack may be at  grade  level  to  accommodate bottom
 filling  of  tank-trucks   or  above grade  for loading through  top
hatches.   Liquid  lines  provide  the  link between  storage tanks,
 the  loading rack, and the tank  truck.   This guidance pertains to
 emissions of  voltaile organic compounds  (VOC)  during the  truck
 loading operations at bulk terminals.
                               3.8-1

-------
3.8.2  Emissions
     Volatile organic  compound emissions  at bulk  terminals  can
occur from  storage  tanks when  the  contents are  at rest,  during
transfer (loading) of gasoline from storage tank into tank truck,
from the tank truck,  and at points along  the  gasoline liquid or
vapor plumbing  lines.   Evaporative emissions  from  these sources
are  categorized  as  standing storage  or  "breathing" loss,  liquid
transfer or "working" loss,  and miscellaneous or "fugitive" loss.
VOC  emissions  from  storage tanks are  discussed in  Section 3.10,
"Gasoline and Crude Oil Storage."
     Working loss from  tank trucks  results from the active move-
ment of  liquids,  most commonly when vapors  are displaced from" a
vessel as  liquid is added.   Working  loss includes both filling
and  emptying  or drainage   losses,  i.e.,  the vapors displaced
during filling  and  those generated after  draining  by the inter-
action of residual liquid and the air introduced into  the vessel.
The  quantity and composition  of emissions  are  related  to  the
physical and  chemical characteristics of  the  old  (residual)  and
the  new  (loading)  cargo;   the  rate  and amount of  unloading and
loading;  leakage from  the  vessels;  the  liquid  lines and their
connections   and  fittings;  temperature  differentials  between
vessels  and liquids; and,  most importantly,  the loading method,
whether splash or submerged fill.
     Figure 3.8-1 depicts the several tank truck loading methods:
overhead,  either by  splash- or  submerged-fill pipe,  and bottom
fill.   Top-splash fill  generates relatively  greater amounts of
vapors because  of turbulence and the opportunity  for contact of
vapor,  air,  and  liquid  during  the liquid free-fall.   Both  sub-
merged  fill,  wherein the  fill  tube  is  always near or below the
liquid level, and bottom fill,  wherein the inflow is  always below
the  liquid surface,  minimize  vapor  generation,  turbulence,  and
VOC  emission.
     Fugitive  losses,  largely  preventable,  result  from  improper
operation  and  maintenance,  faulty  equipment, and human error.

                              3.8-2

-------
U)
oo
I
           LEGEND:
                                           4
                                        SPLASH
                             — TOP LOADING-
                                      f
                              CLOSED  !
                              HATCH   'VAPOR VENT
             SUBMERGED  PIPE
                     y
sr
S*
8

	

-. 1
U* 1 * "
jl\ | VAPOR ZONE \ 1
TO | I
j
1 L10UID ZONE ,
1 - XX I

^r- 4
	 ** /
•*~ /
/

. 	 ,
-4-
-••^ ^*" *




r
.s





•^_
                                                                                    LIQUID LEVEL

                                                                                     Y
                                                                                       < 15 cm (6 in.)
->• LIQUID FLOW
BOTTOM LOADING
                  	+- VAPOR FLOW
                                3.8-1.  Loading methods.

-------
Such   losses   include   leaks   from   liquid   and   vapor   lines,   connec-
tions,   hatch   covers,   and   fittings-   caused   by  improper   mating   or
deterioration   of   components;   defective   or   maladjusted  sensors   and
relief    valves;    vessel    overfills;    open    hatches;     backflow    and
drainage; and similar losses.
     Typical   values    for   uncontrolled   emissions  from   truck   loading
of  gasoline  at  bulk   terminals  are   presented   in   Table  3.8-1.    The
emission   values   do    not   include   fugitive   emissions   due  to   inept
loading or faulty equipment.

3.8.3  Control Measures
     Effective    control   of    tank-truck   loading   operations   at    bulk
terminals   includes    measures   to    suppress    vapor    generation;    a
vaportight   and   properly   sized    system   for   capturing,   collecting,
and   conveying  the   vapors;  and   an   efficient  means   of   vapor  dis-
posal.    Such   measures,  together  with   design  features  that  enhance
vapor   capture,   disposal,    and   spill   prevention;   good    operational
procedures;     preventive     maintenance     practices;     and     attentive
housekeeping provide effective control.
      The   use   of   submerged  fill   rather  than   splash  loading   reduces
vapor  generation  by  approximately   58   percent.    Of   the  two   sub-
merged-fill   methods,   bottom   loading   is   preferred    over   a    sub-
merged-fill   pipe    because    the    installation    is   much   simpler,   the
inflowing   gasoline   is    always   below   the    liquid   level,   and   the
independent    vapor    extraction   and    gasoline   filler   lines   facilitate
vapor collection.
      The   vapor   collection   system  captures   the  vapors  and   conveys
them  to  a  vapor  processing  system.    The   system  must be maintained
vaportight   throughout,    with    particular    attention    to    leak-prone
points   such   as   top-hatch   closures,   vapor   holders,   knockout  (con-
densate)    tanks,     backflow    valves,    and    pressure-vacuum   relief
valves.     Proper   setting   of  fill   meters   and   valves   is  important,
particularly during rapid loading.  The system should incorporate
                              3.8-4

-------
           TABLE 3.8-1.   HYDROCARBON EMISSION FACTORS  .
         FOR LOADING GASOLINE INTO TANK CARS AND TRUCKS*
         Emission source
Emission factor
Submerged loading—normal service
     lb/103 gal transferred
     kg/103 liters transferred

Splash loading—normal service
     lb/103 gal transferred
     kg/103 liters transferred

Submerged loading—balance service
     lb/103 gal transferred
     kg/103 liters transferred

Splash loading—balance service
     lb/103 gal transferred
     kg/103 liters transferred
       5.0
       0.6
      12.0
       1.4
       8.0
       1.0
       8.0
       1.0
  Reference 1, p. 4.4-8.

b The gasoline used in this example has a Reid vapor pressure of
  10 psia.

0 Emission factors are calculated for a dispensed fuel
  temperature of 60°F.
                               3.8-5

-------
design  features  that prevent  vapor  backflow  (drainage)  when
connections are removed,  maintain  a  vaportight connection as the
tank vehicle settles with increasing load,  and eliminate the need
for open  "topping,"  i.e.,  visual observation  as liquid level is
adjusted to tank  capacity.   Gasoline loading lines should incor-
porate  similar features  to  prevent liquid  leaks and drainage
spills.  Tests at bulk terminals have shown that 30 to 70 percent
                                              2
of the vapors can escape  capture at the truck.
     A  recent EPA  publication on  reasonably  available  control
technology (RACT) states  the importance of maintaining vapor- and
liquid-tight  systems  and the techniques  for detecting, control-
ling,  and  minimizing  leaks  from tank trucks and vapor collection
systems.    This  publication defines  leak-tight  conditions  for
RACT;  describes compliance test methods and procedures for ascer-
taining the degree of leak-tightness and for detecting leaks; and
suggests useful record keeping,  inspection, and reporting methods
for  ensuring  that  leak-tight  conditions  are maintained.   The
equipment  performance criteria  and the regulation recommended in
this publication  are briefly discussed below.
     Both  the  compliance  test  procedure  and the  recommended
regulation in this publication define a "leak-tight condition" as
one  that is equivalent  to  99 percent  capture efficiency during
vapor  transfer from  a  truck tank.   The  publication points out
that  this capture efficiency will  decrease because some  sources
(e.g.,  pressure  and vacuum valves  and   hatch  seals)  may  leak
shortly  after maintenance.   The suggested  control  approach of the
publication  is to encourage  more frequent and effective  mainte-
nance  procedures  and adherence  to the  test  and monitoring proce-
dures  described  below.   The recommended  regulation  states  that
gasoline  tank  trucks  and  their vapor  collection  systems  (which
are  tested annually)  should  not sustain a  pressure change  of  more
than  750 pascals (3 in.  of  water)  in 5 minutes when pressurized
to 4500  pascals  (18  in.  of water) or evacuated to  1500  pascals  (6
in.  of water).   For  tank  trucks  and vapor collection  systems,  a
combustible  gas  detector  is used  as  a  monitoring procedure  to

                              3.8-6

-------
ascertain leak-tightness.   During loading operations,  the  recommended regu-
lation  requires that no reading be greater than or equal to 100  percent of the
lower  explosive limit  (LEL, measured as  propane) at 2.5 cm  (1  in.)  around
the perimeter of a potential leak source in the system.  (The vapor  collection
system includes  all piping, seals,  hose connections,  pressure-vacuum vents,
and  other  possible leak sources between the truck and  the  vapor processing
unit and/or the storage tanks  and vapor holder.)
     The recommended  regulation further requires that  the  vapor  collection
and  vapor  processing  system be  designed  and  operated to  prevent gauge
pressure in the tank  truck from exceeding 4500 pascals  (18 in.  of water) and
prevent  vacuum  from  exceeding 1500 pascals  (6  in.  of water).   No  visible
liquid  leaks that can be avoided are allowed from either the tank truck or the
vapor  collection system.
     Two general types  of vapor processing  systems are applicable to  bulk
terminal  truck  loading:    systems  that  thermally destroy the  vapors,   and
systems  that recover the vapors as useful product.  It is  emphasized that the
values for mass  emission rates  at the  outlet of  the  vapor  processing  units
discussed below will vary with the leak-tight condition of the tank trucks and
vapor  piping.
     Thermal  oxidizers  (destructors, afterburners, or incinerators)  are  com-
monly  used at some industrial processes for the  control of combustible aero-
sols,  vapors,  gases,  and  odors.   At gasoline terminals the thermal oxidizer
(TO)  converts gasoline  vapors to  essentially carbon dioxide  and water rather
than recovering them  as liquid gasoline.   The  most effective type of TO unit
consists  of a  vapor  holder,  propane  tank,  burner  (flare),  and associated
piping.  Gasoline  vapors are  directed to the vapor holder, where propane is
added  when necessary to  maintain the  VOC/air  ratio above  the  flammability
limit.   Vapors  are drawn  from  the  vapor  holder,  pass  through the burner,
mix with metered air,  and are  combusted.  Later models of TO systems  do not
require  a  vapor  holder; the  vapors  from tank trucks are vented  directly to
the TO unit.   Requisite safety features  include flashback and safety  interlock
flame  guards  and automatic malfunction  shutdown;  heat  recovery  is  optional.
The  TO  units  have been used for the destruction of VOC vapors from  bulk
                             3.8-7

-------
terminal operations.  The  EPA  test results  on four TO units—two with vapor
holders and  two without--indicated average mass  emission  rates at the outlet
of 8.7,  1.3,  28.5 and 31 mg/liter loaded.4'8
     Vapor  recovery systems  used at bulk terminals  include those employing
compression-refrigeration-absorption  (CRA), compression-refrigeration-conden-
sation  (CRC),  straight refrigeration (RF), and carbon adsorption  (CA).  The
technology  of  these systems  is  well  understood;  the reliability  and control
efficiencies are  good and are being improved.
     The  CRA  unit is  based  on the absorption  of  gasoline  vapors under
pressure  with  cool  gasoline from storage.   The  system generally includes a
saturator,  vaposphere,  chilled  gasoline  absorber,  compressor,  pumps,  and
instrumentation.   Incoming  vapors  are  passed through  the   saturator and
sprayed with gasoline as  a  safety measure to ensure that the vapor concen-
tration  is  above  the  explosive  level.   The  vapors are  then  compressed,
cooled, and  passed to the absorber, where  they are brought into  contact with
chilled  gasoline.  The bottoms  from  the absorber containing gasoline and the
absorbed  vapors are returned  to  storage, and the  remaining air  is vented to
the  atmosphere.  The  EPA  test results on four CRA units indicated average
mass emission  rates at  the  outlet of  59, 43,  40, and 31 mg/liter  of gasoline
       Q-13
loaded.  1J
     The  straight refrigeration (RF) system condenses vapors by refrigeration
at atmospheric  (ambient)  pressure.   Since  vapors  are treated  on demand,  no
vapor  holder  is  required.   The  vapors are  fed  directly to  the  condenser,
cooled  to  minus  73°C  (minus  100°F), and  condensed.  Condensate  is with-
drawn from  the condenser bottom and air is vented from the top.   The stored
brine  is cooled by a two-stage refrigeration unit.  Cooling for  condenser coils
is  supplied  by  methyl  chloride.   The EPA  test  results  on  three  RF  units
indicated average  mass  emission  rates at the outlet of 63,  37,  and 34 mg/liter
                   14-17
of gasoline loaded.
     Carbon adsorption (CA)  systems are based on  the affinity  of  activated
carbon for hydrocarbon vapors.  Equipment  consists  of twin activated carbon
beds,  a vacuum regeneration  unit,  a condenser-separator, and pumps.  Inlet
vapors are passed  through  a carbon  bed,  the gasoline  vapor  is  adsorbed on
                              3.8-8

-------
the  activated  carbon,  and  air  is  vented  to  the  atmosphere.   The  adsorbed
vapors  are  then  vacuum-stripped  and  passed  to the  condenser-separator.
The recovered gasoline  is  pumped to storage.   The  air  and any  remaining
hydrocarbons  are  recycled  from the  condenser-separator through the  carbon
bed to  the atmosphere.  During operation one carbon  bed is in the adsorption
cycle and  one is  in  the  desorbtion  or  regeneration cycle.   The   EPA test
results  on  two CA units indicated  average mass  emission  rates  at the  outlet
(during normal periods  of  operation*) of  3,  10,  and 2 mg/liter of gasoline
loaded.18"20

3.8.4  Emission Limits
3.8.4.1  New Source Performance Standards (NSPS) Limits--
     No NSPS  are applicable  to VOC emissions from the loading of tank  trucks
at bulk gasoline terminals.

3.8.4.2  State  Implementation Plan (SIP) Limits—
     Air pollution control regulations  of some  state and local agencies contain
provisions specifically  applicable  to'gasoline transfer operations.  Table 3.8-2
lists major  provisions  of the  more  restrictive  of  such regulations  as they
apply to gasoline loading at  bulk  terminals.
  The emission  rates during abnormal unit operations were not cal-
  culated for the average mass emission rate at the outlet;  e.g.,
  emission rates during incorrect timer settings (Reference  18),
  and while the unit was purposely overloaded (Reference 19).
                             3.8-9

-------
   TABLE 3.8-2.   STATE IMPLEMENTATION PLAN (SIP)  REQUIREMENTS.
  APPLICABLE TO GASOLINE LOADING OPERATIONS OF BULK TERMINALS'
     SIP agency
       Regulatory requirement
California
  San Francisco Bay
  area
  South Coast
Colorado
New Mexico
  Alburquerque

Missouri
  St. Louis
Organic emissions with vapor pressure
greater than 10.3 kPa (1.5 psia) shall
be reduced by at least 90% of the amount that
would be emitted without controls.

Submerged fill pipe.

Vapor collection and disposal system or
equal required; vapor disposal system
(if absorber or condensation system)
must recover 90% by weight of vapors.

Vapor collection and disposal system or
equal required; drainage prevention;
vaportight lines at. all times; all vapors
to vapor recovery or disposal unit;
final emissions not to exceed 0.155
kg/liter (1.24 lb/103 gal) loaded.
Similar to Colorado regulations.
Vapor recovery system or equal required;
adsorber, condensation, or equal systems
must limit discharge of hydrocarbons to
0.5 g/gal (0.13 kg/kl or 1.1
lb/103 gal) loaded.
a Gasoline loading facilities having an average daily through-
  put greater than 76,000 liters  (20,000 gal).

b The 90 percent requirement in the California area regulations
  would be approximately equivalent to an emission limit  of
  0.14 kg/kl  (1.2 lb/103 gal), i.e., (1-0..9) x uncontrolled
  VOC emission  factor of 1.4 kg/kl  (12 lb/103 gal).
                               3.8-10

-------
     Table 3.8-2 illustrates the similarity of formats and emission limitations in
SIP regulations.   The  most stringent limitation  is that of Missouri for the St.
Louis area, which  requires  that  vapor recovery system or equal be  provided,
displaced  vapors  be  vented to  vapor  recovery  system,  drainage spill  be
prevented,  and  the  vapor  disposal  system meet an  emission limit  of  130
mg/liter  (1.1  lb/103  gal).   Some  state  SIP regulations  are  presently  being
revised  to reflect  the stricter requirements (an emission limit  of 80 mg/liter
loaded) achievable  with RACT.

3.8.4.3  Achieved-in-Practice (AIP) Limits--
     Available  test data  indicate the  control efficiency achieved in practice
with some  types of TO and CA units approaches 99.9 percent.  Slightly lower
control efficiencies  have  been reported for other vapor recovery units.  It is
emphasized that these values pertain  to  the  emission  reduction  across  the
control device only and  therefore assume  100 percent collection and transport
of  vapors to  the  disposal  unit—a  condition  that  has  not  been observed in
practice.

3.8.5  Determination of LAER
     The  recommended  limitations  are based  on  SIP's and  on performance
information available  in  early  1979.   It  is  anticipated  that several additional
SIP  regulations  covering  these sources will be promulgated  and/or  modified in
1979 and 1980 and  that appreciable  new performance data will  become available
in the near  term.   Conceivably, some SIP regulations may  be more  stringent
than  the  LAER  suggested  herein.   Furthermore,  performance testing  may
show that more  stringent limits than  those suggested  are  feasible or it  may
show that the suggested limits are appropriate or that they are not  achievable
for some specific subcategories.  In any case,  the  basis for determining LAER
for many source categories is  expected  to change  frequently.  Since LAER is
near  the vanguard of control  technology,  a more  detailed  analysis is  partic-
ularly necessary when addressing  modified or reconstructed  facilities  subject
to  the  provisions  of Section 173 of the Clean Air  Act.  Emission  limitations
reasonable for  new  sources may in some  instances be  economically or tech-
nically  unreasonable when applied  to modified or reconstructed sources of the
same type.
                              3.8-11

-------
     No  NSPS limitations are  applicable  to VOC  emissions from  tank-truck

loading  at  bulk  gasoline  terminals.    The  most  stringent SIP  imposes  an

emission  limit of  130  mg/liter  (1.1 lb/103  gal) loaded.  Emission values  less

than this SIP limitation have  been achieved  in practice.
     LAER  for  tank-truck loading operations at bulk  terminals is  determined

to be  an emission limit  of  30 mg/liter (0.25 lb/10  gal) of  gasoline loaded and

reguires an overall  system  efficiency  of 97 percent.   On the basis  of  EPA

emission  test  data,   this  value  appears  achievable  with thermal  oxidizer,

carbon adsorption, or  other vapor  recovery units  operated at optimal effi-

ciency in  conjunction with  a  highly effective  capture-collection  system  that

provides virtually leakproof transport of vapors to the control unit.

     It  should  be  emphasized  that  effective vapor collection  and  leak-tight

delivery to the processing unit  are  required to maintain the integrity of  the

vapor  control system and  to minimize VOC  emissions.  The RACT performance

criteria, compliance  test methods, and  other related requirements  are  recom-

mended for attaining  and maintaining a  leak-tight  system.
     The  following   measures  are  designed  to  reduce   or   preclude VOC

emissions and to  support LAER  controls.

     1.   System   design    features   that    provide  vaportight    connec-
          tions    at   all   times,   incorporate   closed   hatch   "topping,"
          accomodate    vehicle    settling,    prevent    venting    of    the
          relief   valve   during   loading    at    maximum   rate,    prevent
          all   backflow   or   drainage   when   connections    are   made   or
          disengaged,    and    provide   maximum    legal   release    pressure
          for all relief valves.

     2.   Supervised     standard     operating     practices    that     curtail
          VOC     loss    due    to    poor   housekeeping    and/or    operating
          procedures.      Examples     of    poor     operating     practices
          include   excessive    fill    rates   that   increase   vapor   gen-
          eration  and   pressures,    improper   setting   of   fill   meters
           (causing   overfills),    careless   or     improper    connections,
          improperly    set   pressure/vacuum   relief  valves,    and   open
          hatches.

     3.   Supervised     programs     of     preventive     maintenance    and
           scheduled    inspections   with   the   objective   of   maintaining
           liquid-    and    vapor-tight    systems,    preventing    leaks    in
           liquid/vapor    plumbing    and    fittings,    and    repairing    or
           replacing  defective or  malfunctioning hardware.
                               3.8-12

-------
                           REFERENCES


1    Supplement No.  7 for Compilation of Air Pollution Emission
     Factors,  Second Edition.   Office of Air Quality Planning and
     Standards, U.S. Environmental Protection Agency.   April
     1977.

2    Control of Hydrocarbons from Tank Truck Gasoline Loading
     Terminals, Guidelines Series.  EPA-450/2-77-026,  Office of
     Air Quality Planning and Standards, U.S. Environmental
     Protection Agency.  October 1977.  pp. 2-3.

3.   Control of Volatile Organic Compound Leaks from Gasoline
     Tank Trucks and Vapor Collection Systems,  EPA-450/2-78-051,
     Office of Air Quality Planning and Standards No. 1.2-119,
     U.S. Environmental Protection Agency, Research Triangle
     Park, N.C.  December 1978.

4.   EPA-450/2-77-026, Table 3-2.  pp. 3-5.

5.   Report on Performance Test of Vapor Control System at the
     AMOCO Terminal, Baltimore, Maryland.  Contract No.
     68-01-4145, Task  12, U.S. Environmental Protection Agency,
     Region III and Division of Stationary Source Enforcement.
     September 1978.

6.   Demonstration  of  Reduced Hydrocarbon Emissions from Gasoline
     Loading Terminals.  EPA-650/2-75-042, Office of Research  and
     Development, U.S. Environmental  Protection Agency.  June
     1975.  Table VI,  p. 18.

 7.   Report on Performance Test of Vapor  Control System at
     Belvoir Terminal, Newington, Virginia.   Contract No.
     68_0i-4145, Task  12, U.S. Environmental Protection Agency,
     Region III and Division of Stationary  Source Enforcement.
     September 1978.

 8.   Air Pollution  Emission Test.  Project  No.  78-BEZ-5, Office
     of Air Quality Planning and  Standards,  Emission  Standards
     and Engineering Division, Emission Measurement Branch,  U.S.
     Environmental  Protection  Agency.  May 12,  1978.
                               3.8-13

-------
9.    EPA-450/2-77-026,  Table 3-2.   pp.  3-5.

10.   Report on Performance Test of Vapor Control  System at Crown
     Central Terminal,  Baltimore,  Maryland.   Contract No.
     68-01-4145,  Task 12,  U.S.  Environmental Protection Agency,
     Region III and Division of Stationary Source Enforcement.
     September 1978.

11.   Air Pollution Emission Test.   Project No.  76-GAS-17,  Office
     of Air Quality Planning and Standards,  Emission Standards
     and Engineering Division,  Emission Measurement Branch, U.S.
     Environmental Protection Agency.   September  1976.

12.   Report on Performance Test of Vapor Control  System at Texaco
     Terminal, Coraoplis,  Pennsylvania.  Contract No. 68-01-4145,
     Task 12, U.S. Environmental Protection Agency, Region III
     and Division of Stationary Source Enforcement.  September
     1978.

13.   Air Pollution Emission Test.   Project No.  75-GAS-10,  Office
     of Air Quality Planning and Standards,  Emission Standards
     and Engineering Division,  Emission Measurement Branch, U.S.
     Environmental Protection Agency.   December 1974.

14.  EPA-450/2-77-026, Table 3-2.   pp. 3-5.

15.  Air Pollution Emission Test.   Project No.  77-GAS-18,  Office
     of Air Quality Planning and Standards, Emission Standards
     and Engineering Division, Emission Branch, U.S.
     Environmental Protection Agency.   November 1976.

16.  Air Pollution Emission Test.   Project No.  75-GAS-8, Office
     of Air Quality Planning and Standards, Emission Standards
     and Engineering Division, Emission Measurement Branch, U.S.
     Environmental Protection Agency.  December 1975.

17.  Air Pollution Emission Test.  Project No. 76-GAS-16,  Office
     of Air Quality Planning and Standards, Emission  Standards
     and Engineering Division, Emission Measurement Branch, U.S.
     Environmental  Protection Agency.   September  1976.

18.  Air Pollution  Emission Test.  Project No. 77-GAS-19,  Office
     of Air Quality Planning and Standards, Emission  Standards
     and Engineering Division, Emission Measurement Branch, U.S.
     Environmental  Protection Agency.   October 1977.
                               3.8-14

-------
19.  Report on Performance Test of Vapor Control System at
     Phillips Fuel Terminal,  Hackensack, New Jersey,   Draft
     Report.  Contract No. 68-01-4145,  Task 12,  U.S.
     Environmental Protection Agency,  Region II  and Division of
     Stationary Source Enforcement,  January 1979.

20.  Report on Performance Test of Vapor Control System at
     British Petroleum Terminal,  Finksburg,  Maryland.   Contract
     No. 68-01-4145,  Task 12,  U.S. Environmental Protection
     Agency, Region III and Division of Stationary Source
     Enforcement.   September 1978.
                             3.8-15

-------

-------
3.9  MAJOR SOURCE CATEGORY:   FLAT  WOOD  PANELING--VOLATILE
     ORGANIC COMPOUNDS (VOC)  EMISSIONS
     Flat  wood  products  include  a variety  of  materials  such  as  plywood,
particleboard,   hardboard,   panelboard,  fiberboard,  insulation  board,   and
molding.  In 1976 approximately 640 flat wood plants  were in  operation; 247 of
these were hardwood plywood and 240 were softwood  plywood  plants.
     Plants  that apply coatings to their  wood  products generate emissions of
volatile organic  compounds  (VOC),  which is the subject of this  section.  On
the basis of membership in national  associations  of flat wood producers, it is
estimated that  40 percent  of hardwood plywood  plants,  10 percent of the
softwood  plywood plants,  and under 15 percent of the particleboard plants
apply  coatings.    For  this  reason, the major emphasis  in  this  guidance is
directed toward  the hardwood plywood industry.  It is  intended to apply to
printed interior  wall  panels made from hardwood  plywood and thin particle-
board,  natural  finish hardwood  plywood  panels,  and  Class  II finishes for
hardboard  paneling.    Segments  of  the  flat  wood products  industry  not
addressed are exterior siding,  tileboard, and  particleboard  used as a furni-
ture component.

3.9.1  Process Description
     Figure  3.9-1  is  a general flow diagram  of flat wood  coating processes
utilizing conventional coatings with  organic  solvents.   As the  diagram  indi-
cates, the  first operation has an optional  step,  the sanding of  particleboard
before it enters the brushing unit.  All other flat wood products  go directly
to  the brushing  operation.   After  the  brushing  operation,  the  stock is
directed to  a reverse-roll  coater  for filler application.   Normally, filler  is not
                                                           2
applied to hardboard before the application  of a base  coat.    If the panel is
not to  be filled, it bypasses the filler  and drying oven and  proceeds to a
direct-roll coater for application of a sealer  or first  base  coat.  Drying  ovens
can  be   gas-fired  or  electrically  heated.    If  ultraviolet-curable   coatings
(commonly called UV coatings) are  applied, a UV  oven  is  used for curing.
                               3.9-1

-------
                                       FILLER
OJ
>£>
I
                OFF-

            T.OADING
                                                                              SEALER OR FIRST
                                                                              BASE COAT
                                                                              FIRST BASE COAT
                                                                                   OR
                                                                              SECOND PASE COAT
PRINT
                              Figure  3.9-1.  General process  flow.

-------
     The  most  common  flat wood products coated  by the  industry are  ply-
wood,  particleboard,   and  hardboard.   A few  plants  coat  pine  and cedar
siding.   Coatings that can be factory-applied  include  filler, sealer,  base coat,
ink, and  topcoat.  Fillers  are  used  to fill voids and cracks in the wood  to
provide a  smooth surface.  Sealers seal off pores and substances in the wood
that may  affect  subsequent  finishes.   Base  coats are  used  as the primary
coating/coloring  of the  panels,  and inks are  used  only  for  decorative  or
simulated   grain  panels.    Topcoats provide both  a protective  coating  and  a
finished  appearance.   Not all factory-prefinished  wood products undergo the
complete series of coatings.   For instance, a sealer is not required after the
application  of  certain  fillers;  also,  some  builders  are losing interest  in  top-
coat-finished materials because they damage easily during installation.
     Following application of the sealer  or  first base coat,  a second direct-roll
coater is   used to apply  an initial  or a second  base  coat to the panel.  The
filled,  sealed,  base-coated board can be shipped as is after drying, or it can
be  given   wood grain  or  topcoats.   The  grainprinter is a direct-roll  coater.
Usually, three or four such  printers are  used  in  series to provide different
colors  of  grain.   After it is printed,   the panel  goes  through  a  direct-roll
coater or a curtain coater (often  used by  plywood coating plants) to receive a
protective   coat.   The  panel then  goes  through  a  drying  oven  for  curing
before shipment.
     Similar processes are  used  by  plants with waterborne  coating systems.
The main  differences  are longer ovens, lower operating  rates, or higher  oven
temperatures  (for proper  curing).
     Among the several methods of applying coatings to flat wood, the  pre-
ferred  ones are  roll  coating and  curtain coating.   In roll  coating  (Figure
3.9-2) the coating material is applied  to  the wood by cylindrical rollers.  If
the cylinder rotates in the same direction as panel movement, the applicator is
called  a  direct-roll coater,  generally  used to apply base coat, print, and top-
coat.   If  the  cylinder rotates  in  the  opposite  direction  of  panel movement
(reverse-roll   coater), the  coating is  forced  into voids  and  cracks  in the
panels, fills these depressions,  and provides a smooth surface.
                               3.9-3

-------
     In  curtain  coating   (Figure  3.9-3),  used  mostly  for  topcoating,  the
coating material  forms a  curtain  through  which the  panel  passes.  Coating
material  is  metered into  a pressure  head and  is forced  through a calibrated
slit between two knives to form  a continuous,  uniform curtain.   The  rate  of
panel  movement  determines  the  coating  thickness.   All  excess  coating  is
caught in a trough and recirculated.

3.9.2  Emissions
     Volatile organic  compounds  are  emitted  by  the  evaporation  of  volatile
organic solvents  used in  conventional coatings  applied to flat wood products.
Small  quantities  of dust  and smoke  also  may  be emitted to the atmosphere.

3.9.2.1  Nationwide Emissions--
     No  more  than one-quarter  of the  U.S.  plants that turn  out flat wood
products  apply  coatings.   Most of  these are  primarily  hardwood  plywood
plants, and in some cases they  coat  only  a  small percentage of production.
     In  1975 the total VOC  emissions from all  flat wood  plants  was estimated
to  be 67,000  Mg  (74,000 tons).   This  estimate is based on annual  coating
                  •3           c       o
usage of 132  Mm  (35  x 10  gal),   an  average  emission  factor  of  0.5  kg
VOC/liter  (4.2  Ib/gal)    of  coating,  and  no  emission control.   In the same
year,  VOC  emissions  from stationary and automotive sources were estimated to
be 17.3  Tg  (19  x 106 tons) and  10.8 Tg  (12  x 106 tons).5   Therefore, VOC
emissions from  flat  wood  products manufacturing  account for  less than  0.4
percent  of  emissions from stationary  sources,  or about 0.2 percent of  all VOC
emissions.

3.9.2.2  Sources and Quantity of VOC  Emissions  --
      Emissions of  volatile organic solvents at  flat  wood coating plants  occur
primarily at the  coating lines.  Oven  exhausts  are discrete point sources, and
coaters  and rollers  are  termed  fugitive emission sources.   Solvents  used in
organic-based  coatings are normally  multicomponent mixtures that may include
methyl ethyl ketone,  methyl isobutyl ketone,  toluene, xylene, butyl acetates,
propanol, ethanol, butanol, VM and P naphtha,  methanol, amyl acetate,
                               3.9-4

-------
              COATING
                                        APPLICATOR
                                                           (a)
              APPLICATOR
              ROLL
                                           COATING
                                             PANEL
                                                           (b)
               (ARROWS SHOW DIRECTION OF  ROLLER AND PANEL
                             MOVEMENTS)
Figure 3.9-2.   Simplified schematic  of roll  coaters.   (a)  direct-
                 roll coater,  (b) reverse-roll coater.
                                3.9-5

-------
                           COATING
                          RESERVOIR
Figure 3.9-3.   Pressure head curtain  coater,
                      3.9-6

-------
mineral  spirits, SoCal I and  II,  glycols, and glycol ethers.  Organic solvents
most  often  used  in  waterborne  coatings  are  glycol,  glycol ethers  (such as
butyl cellosolve),  propanol, and  butanol.  Contents of volatile  organics in the
different  types of  conventional  coatings  supplied  to  the flat  wood  coating
industry  are  shown  in  Table 3.9-1  with  estimated  emission  factors.   The
composition  of the solvent determines  the type of VOC emitted.  Waterborne
coatings are  discussed under control measures.

3.9.2.3  Factors  Influencing VOC  Emissions--
     Table  3.9-2  lists  the  common  organic  solvents  used   in  conventional
coatings  and  their  vapor  pressures  and  relative evaporation  rates.   The
evaporation   rate  indicates the rate of  VOC emissions relative  to  each com-
pound.   For  example, ethanol evaporates three times faster than iso-butanol
at constant  temperature,  pressure, and humidity.   Coating mixtures contain a
number of these solvents  and  vary with each operation  (filling,  sealing, base
coating, topcoating)  and plant.  The VOC emissions from each  operation and
plant therefore vary  widely  in  mass  rate per unit production, in mass  rate
per  unit weight  of  coating  used,  and in concentration.  In  addition, the
distribution   of   solvent   emissions  from  solvent  mixing,   handling,  and
application (workroom emissions exhausted through  roof vents and windows),
and  from drying ovens  (point  sources)  can  vary widely.   For  example,  a
plant that uses highly volatile  solvents  such as methanol,  ethanol, and methyl
ethyl ketone  will  emit much greater amounts of VOC  through  roof vents and
windows  (say  70  percent)   than  through  drying  oven exhaust  (say 30
percent).   In contrast,   a  plant that  uses relatively  low-volatility  solvents
such  as amyl  acetate, butanol,  and VM  and  P  naphtha  will  emit much less
VOC  from  handling,  mixing,  and application  (say  20 percent)  than from
drying  oven  exhaust  (say 80  percent).  The  first example  is  probably the
more  typical.

3.9.2.4  Summary--
     Table  3.9-3  presents  estimates  of  potential  VOC  emissions  from each
                                       Q
operation using  conventional   coatings.   At  plants  applying  filler,  sealer,
                              3.9-7

-------
U)
I
oo
                  1.9-1.  VOC CONTENT OF CONVENTIONAL FLAT WOOD COATINGS3
J. t\D .i-U-J —
Paint type
Filler
Sealer
Base coat
Grain ink
Topcoat
Number
of companies
4
3
7
6
8
Density ,
kg/liter
1.7
1.1
1.4
1.2
1.1
(Ib/gal)
(14.5)
(9)
(11.5)
(10)
(8.8)
Volatile
weight
15 -
15 -
40 -
30 -
50 -
organics ,
percent
30
50
75
70
75
    a Source:  Miscellaneous paint companies and Mr. Martin  Kay,  South Coast Air  Quality

               Management District California.  Reference 7.

-------
TABLE 39-2   SOLVENTS USED IN COATINGS AND THEIR VAPOR PRESSURE
                     AND EVAPORATION RATEb
Compound
Butanol, iso
Butanol, n
Butanol, sec
Ethanol, anhydrous
Propanol, anhydrous
Methanol
Methyl ethyl ketone
Methyl isobutyl ketone
Toluene
Xylene
Butyl acetate, sec
Butyl acetate, iso
Butyl acetate, n
VM and P naphtha
Amyl acetate (primary)
Glycols
Glycol ethers
Vapor pressure,
at 20°C, mm Hg
8.8
4.4
12.7
44.0
31.2
96.0
70.6
16.0
38.0
9.5
4.0
12.5
7.8
2.0
4.0
<0.01
<1.0
Evaporation rate
0.63
0.46
0.90
1.9
1.7
3r"
. 5
4.6
If
. 6
IP
.5
0.75
1r\
. 9
1.45
1f\
. 0
OA r"
. 45
0.4

<0. 01
Relative  to  that  of  butyl  acetate,  1.0.
                               3.9-9

-------
          TABLE 39-3   POTENTIAL VOC EMISSIONS FROM FLAT WOOD OPERATIONS USING
                                  CONVENTIONAL COATINGS

Operation coating
Filler
Sealer
Base coat
Grain ink
Topcoat

Range of VOC emissions ,
kg/liter
coating
0.26 - 0.53
0.17 - 0.54
0.44 - 1.0
0.36 - 0.84
0.53 - 0.79
Ib/gal
coating
(2.2 - 4.4)
(1.4 - 4.5)
(3.7 - 8.6)
(3.0 - 7.0)
(4.4 - 6.6)
b
Coverage ,
liters/
103m2
119
21
56
7
49
(gal/103ft)
(1.7)
(0.3)
(0.8)
(0.1)
(0.7)
Range of VOC emissions
g/m2 coated
31 - 62
3.6 - 11.3
25 ~ 56
2.5 - 5.9
26 ~ 38.7
	 . 	
Total 88 - 174
lb/103ft2
(3.7 - 7.5)
(0.4 - 1.4)
(3.0 - 6.9)
(0.3 - 0.7)
(3.1 - 4.6)
(11 - 21)
*Table 3.9-1 presents typical VOC contents.
DPaint coverage information is from Reference 10.

-------
base  coat,  grain  ink,  and  topcoat,  the estimated VOC  emission  factor  may
range  from  88 to 174  g/km2 (11 to 21 lb/1000  ft2) of flat wood  coated.
     Because  the  volatile  fraction of the coatings  contributes essentially 100
percent  of  the  VOC emissions,  the total emissions  from a plant are essentially
the product  of the  weight fraction  of volatile organics  in  the  coatings  and
coating usage.  Composition  of the VOC emissions  depends upon the types  of
solvents  used.    Exhausts  from  the  ovens  release  practically  all  of  the
incoming  volatile   compounds,  and the  fraction  of  total  plant emissions  that-
comes  from  the dryer ovens depends on the  types of solvents used, or more
specifically  their  relative  volatility or  evaporation  rate.   Apparently  no
emission test  data  are available.

3.9.3  Control Measures
     Technology for  reduction of VOC emissions  from flat wood operations
includes use  of low-solvent,  ultraviolet-curable coatings; waterborne  coatings;
and incinerators.   Generally, adsorption or add-on control devices other than
incinerators   are   not  considered  feasible   or  demonstrated  technology.1
High-solids  coatings are not practical for current  use; however, their use as
fillers  and  sealers  has been  demonstrated  and  may be  further  developed.
Electron-beam (EB)  curing systems  may  have  potential application  for  some
product  lines; however, only one  commercial  facility in the United States has
installed such a system.   '

3.9.3.1  Incineration--
     The  use  of   add-on control  devices such  as  direct-flame and  catalytic
incinerators (afterburners) is  very  limited  in the  flat wood industry.   Two
plants,  both  in southern California,  operate  afterburners, but one has peti-
tioned   to   stop  operations  of  the   burner  because  they  are  using  Rule
66-approved  solvents.   Data  are  not  available on  control  efficiency or  fuel
requirements  at that plant.15  The afterburner at  the second  plant  has been
tested  several  times,  but it  has  not  met  control  efficiency  requirements.
     No  adsorption system is  known to be used in the  flat wood industry.
Use of  multicomponent  solvents and  the use  of different  coating formulations
                               3.9-11

-------
at various steps along the coating line are not conducive to the application of
adsorption  techniques for  emission  control.    Specific  applications  may  be
found  in  redwood  surface treatment  facilities,  where over  90  percent of the
coating is volatile and can be  recycled by carbon adsorption.

3.9.3.2  Waterborne Coatings--
     The  primary emission reduction  technique used  in the  flat wood industry
is conversion from  conventional high-solvent coatings to waterborne coatings.
Paint manufacturers have developed and are continuing to  develop  waterborne
coating formulations  to  replace  conventional  coatings  for  many  factory flat
wood applications.   The use  of  applicable  waterborne  coating in  place of a
conventional organic  solvent-borne  coating  can  reduce volatile organic  emis-
sions by at least 70 percent.17  Table 3.9-4 presents typical values of volatile
organics  in  conventional  and waterborne coatings.   Table  3.9-5  presents an
estimate of  VOC emissions in weight per  area of surface covered.
     Major  changes  are  not  required for use  of  waterborne  coatings.   The
primary use of  waterborne coatings  is  in the filler  and base  coat applied  to
printed interior  paneling.   Limited  use  has been made of waterborne materials
for  inks,  groove  coats, and  topcoats on printed  paneling, and for inks and
groove coats on  natural hardwood panels.  Waterborne coatings  can reduce
fire hazards,  fire  insurance  costs,  and air pollution.   Problems with water-
borne coatings  include possible grain raising, wood swelling, and poor quality
finish.17  A major  complaint  is that  the  waterborne  coatings available to date
require longer cure times.
                           19
3.9.3.3   Ultraviolet Curing   --
      Ultra violet-curable  coatings,  where  applicable,  effect  almost  100  percent
reduction of VOC  emissions.   In the flat  wood  industry,  UV coatings have
found use  as  clear  to semitransparent fillers  and  topcoats  for  interior printed
paneling  and cabinetmaking products.
      Ultraviolet-curable  coatings  are  a  combination of resin, prepolymers and
monomers,  and  a  photosensitizer  (which serves  as a  catalyst).   Polyester,
 acrylics, methane,  and  alkyds are common coating   materials.  Applied as  a
                                3.9-12

-------
                    TABLE 3.9-4.  VOLATILE ORGANICS IN FLAT WOOD COATINGS
                                                                         18
Paint
category
Filler

Sealer

Base coat

Grain ink

Topcoat

Paint
typea
C
W
C
W
C
W
C
W
C
W
Density
kg/
liter
1.7
1.7
1.1
1.1
1.4
1.4
1.2
1.3
1.1
1.1
(Ib/gal)
(14.5)
(14.5)
(9)
(9)
(11.5)
(12)
(10)
(10.5)
(8.8)
(9)
Weight
percent
non-
volatile
75
75
60
55
45
55
40
50
40
45
Typical VOC
content
kg/
liter
0.43
0.05
0.43
0.12
0.76
0.08
0.72
0.18
0.64
0.17
(Ib/gal)
(3.6)
(0.4)
(3.6)
(1.0)
(6.3)
(0.7)
(6.0)
(1.5)
(5.3)
(1-4)
VOC content
(less water)
kg/
liter
0.43
0.07
0.43
0.23
0.76
0.19
0.72
0.38
0.64
0.32
( Ib/gal )
(3.6)
(0.60)
(3.2)
(1.9)
(6.3)
(1.6)
(6.0)
(3.2)
(5.3)
(2.7)
VOC reduction
for equivalent
coverage, %
-
90
-
70
-
90
-
80
-
80
I
I—I
CO
       C = conventional paint with organic solvent.
       W = waterborne, i.e., at least 80 percent of the volatile portion of the coating
           is water.


       Data received indicate that all companies providing information were able to meet
       the VOC content given for waterborne coatings.

-------
                                                                 -1 p

                           TABLE  3.9-5.  ESTIMATED3 VOC EMISSIONS
Paint
Filler
Sealer
Base coat
Print ink
Topcoat
Total
b
Coverage,
liters/103m2 (gal/1000 ft2)
Waterborne
112 (1.6)
25 (0.35)
46 (0.65)
7 (0.1)
46 (0.65)
236 (3.4)
Conventi onal
119 (1.7)
21 (0.3)
56 (0.8)
7 (0.1)
49 (0.7)
252 (3.6)
Potential VOC emissions,
kg/100 m2 coated (lb/1000 ft2 coated)
Waterborne
0.3 (0.6)
0.2 (0.4)
0.2 (0.5)
0.1 (0.2)
0.4 (0.9)
1.2 (2.6)
Conventional
3.0 (6.1)
0.54 (1.1)
2.4 (5.0)
0.3 (0.6)
1.8 (3.7)
8.0 (16.5)
UV
nil
0
0.24 (0.5)
0.1 (0.2)
nil
0.4 (0.8)
vo
I
     a Table 3.20-4 presents typical VOC content.


     b Paint coverage based on information from Reference 11.   Adjustments made for

       waterborne and conventional paints are based on typical nonvolatiles content.


     c UV line uses no sealer; uses waterborne base coat and ink.  Total is adjusted

       to cover potential emissions from the UV coatings.

-------
liquid, the  coating  is  cross-linked and hardened  on  exposure to UV.  The
curing is  extremely fast.   Although  there have  been  attempts to  develop
opaque UV coatings,  none is yet available.
     The  advantages  of UV coatings  include  reduced  power  requirements,
space savings through reduced storage and oven size, very little emission of
VOC, and  the  essentially  100 percent  usable coating  (since all components of
the coating normally react and  become part of  the  coating).  Moreover,  the
short cure times can be measured in seconds, and a  superior product results.
Since little or  no curing takes  place  after  the  panel  leaves the oven,  proper
cure  times must be  carefully established.  Safety  precautions  must be taken
to minimize exposure to  UV radiation and  to avoid contact with  the coating,  as
                                                  20
some  of the raw materials can cause chemical burns.

3.9.3.4  Electron Beam Curing--
     One commercial  facility in  the  United States uses an electron beam (EB)
system.  Opaque coatings can be cured to  a depth  of approximately 15 mils.
Three to 5 mils  of EB-cured coating  produce a  smooth,  wear-resistant finish
with performance comparable to that of many plastic laminates.   '    Although
emissions from  the system are still unknown, some airborne acrylics have been
                                                                13
reported, and  monomers  and ozone  emissions  are  possibilities.     Over  99
percent control of VOC is expected  from EB  systems.
3.9.3.5  Summary of Control Measures--
     The  highest reduction  of  VOC emissions,  about  97 percent,  can  be
achieved by  use  of UV-curable coatings.  This control measure, however,  can
be  only  applied  to certain products  and operations.   The use  of waterborne
coatings  can achieve  up to 90 percent  reduction over conventional coatings,
although use  is  limited somewhat by  the  necessity  of maintaining acceptable
surface and appearance quality.   Incineration of dryer exhaust
                              3.9-15

-------
gases can reduce overall plant emissions by about 75 percent.   At
some  installations  incineration might  also be  used to  control
coating  line  emissions  and  provide an  overall reduction  of  95
percent  or  more.  Electron  beam curing  can provide  both a  99
percent  control  of VOC  emissions  and  excellent  surface  finish
quality.  The major deterrent to its use is the  highest  of both
the EB system and the costing materials.

3.9.4  Emission Limits
     Review of  State  Implementation Plans  (SIP)  showed  no  VOC
emission limitations specific to the flat wood industry.  Rather,
states  have  adopted   general hydrocarbon  emission  regulations
patterned after Los Angeles Rule 66.  The South Coast Air Quality
Management  District  Rule 442   (formerly  Los  Angeles  Rule  66)
places  limitations  on  emissions   from  equipment  using  organic
solvents  or  organic materials containing  organic solvents.  For
organic  materials  that  come  into   flame  contact or are  baked,
heat-cured, or heat-polymerized the  limitation is 1.4 kg (3.1 Ib)
per  hour,  not to  exceed 6.5 kg  (14.3  Ib)  per  day.  For organic
materials emitted  from use of photochemically reactive solvents,
the  limitation is 3.6  kg  (7.9 Ib) per day.  For organic materials
emitted  from  use of photochemically reactive solvents,  the limi-
tation is 3.6 kg (7.9  Ib) per hour,  not to exceed 18 kg (39.6 Ib)
per  day  if not in flame contact, baked, heat-cured, or heat-poly-
merized.   The above-mass emission  rate  limitations do  not apply
if the emissions are reduced by at  least 85 percent.
      Recent  research  has  indicated that  substituting  low-reac-
tivity  solvents for  high-reactivity solvents  may reduce photo-
chemical  oxidant levels  locally while increasing them in downwind
regions.   Accordingly,  EPA's  current  policy  on  VOC  emissions
emphasizes reduction of  all hydrocarbons rather than substitution
of exempt for nonexempt  solvents.
                                3.9-16

-------
     The  EPA  has set  no NSPS for this  source category.  It appears  LAER
will  be determined by  estimated  reduction  levels achieved in practice by  use
of waterborne coatings  in  combination with solvent-borne coatings.

3.9.5  Determination  of  LAER
     The  recommended  limitations  are  based  on SIP's and on  performance
information  available  in early  1979.  It is anticipated  that  several  additional
SIP  regulations covering these sources will be  promulgated and/or modified in
1979  and 1980 and  that  appreciable new performance data will become available
in the near  term.  Conceivably,  some  SIP regulations may be more stringent
than  the  LAER  suggested  herein.   Furthermore,  performance  testing mah
show that  more  stringent  limits  than those  suggested are feasible  or it may
show that the  suggested limits are  appropriate  or that they are not achievable
for some specific  subcategories.  In any case, the basis for  determining  LAER
for many  source categories is  expected to change frequently.   Since LAER is
near  the  vanguard of  control  technology,  a  more detailed analysis  is partic-
ularly necessary  when  addressing  modified or  reconstructed facilities subject
to the  provisions  of Section  173 of  the Clean  Air  Act.  Emission limitations
reasonable for new sources  may in  some instances be  economically  or  tech-
nically unreasonable  when  applied  to modified or reconstructed sources of  the
same type.
     A  study was  made by EPA  to determine the lowest emission levels  being
                                                   21
achieved  by manufacturers  of  flat wood paneling.     Since  no state  regu-
lations  are  more  restrictive,  the  emission  limitations  recommended  in this
document  are considered representative of LAER.   Table 3.9-6  presents  these
limitations.
     The  recommended  emission  limits  are  stated in terms of  kg of VOC  per
      2                                      2
100  m   of  coated  surface  (Ib  per 1000  ft ) to give  operators   necessary
flexibility in adjusting  the VOC  content of the various coatings  applied to  a
given panel.   Because  practices  vary,  it would be  difficult to  set a  VOC limit
for  each  type  of  coating.   By balancing the VOC content  and properties of
the various coats, it is possible  to achieve acceptable VOC reductions without
sacrificing product quality.
                              3.9-17

-------
               TABLE  3.9-6.  RECOMMENDED EMISSION
            LIMITATIONS  EOR FACTORY-FINISHED PANELING
Product category
Printed interior wall panels
made of hardwood plywood and
thin particleboard
Natural-finish hardwood
plywood panels
Class II-finished hardboard
paneling
Recommended limitation
kg of VOC
per 100 m2 of
coated surface
2.9
5.8
4.8
Ib of VOC
per 1000 ft2 of
coated surface
6.0
12.0
10.0
     For  printed  interior panels,  emission  limits  are based on  partial use  of
waterborne  and  solvent-borne  coatings.   Waterborne  coatings  that  produce
acceptable quality are not available for all coatings, particularly clear  topcoats
and  printing  inks.   For natural-finish  paneling,  the limits are based on use
of solvent-based coatings of lower solvent content than conventional coatings.
The  number of coats  and coverage of coatings  vary, but  (for  typical usage)
the  recommended limitations  are  equivalent to  usage of  coatings  that have
average  VOC  contents  of  0.20 kg/liter  (1.7  Ib/gal)  for printed hardwood
paneling,  0.38 kg/liter (3.2  Ib/gal)  for natural-finish  paneling,  and  0.32
kg/liter  (2.7 Ib/gal) for Class II-finish  hardboard paneling.
                               3.9-18

-------
                           REFERENCES


1.    Pacific Environmental Services,  Inc.   Control of Volatile
     Organic Emissions from Existing Stationary Sources.   Volume
     VII Factory Surface Coating of Flat Wood Paneling.   EPA
     450/2-78-032,  U.S.  Environmental Protection Agency OAQPS.
     Research Triangle Park,  North Carolina.   June 1978,  pp. 1-2,
     3.

2.    Personal communication between F.T. Fisher, Pacific Finish-
     ing Company,  Paramount,  California, and Pacific Environ-
     mental Services,  Santa Monica, California.

3.    Cohen, R.W.  Napko Uses Aqueous Coating Systems for Panel
     Refinishing.   Plywood and Panel Magazine, June 1977.

4.    Pacific Environmental Services,  Inc.   (1978).  p. B-l.

5.    Control of Volatile Organic Emissions from Existing Station-
     ary Sources-Volume I:  Control Methods for Surface Coating
     Operations.  EPA-450/2-76-028, U.S. Environmental Protection
     Agency OAQPS,  Research Triangle Park, North Carolina.
     November 1976.  p.  1.

6.    EPA 450/2-78-032, p. 1-15.

7.    EPA 450/2-78-032, p. 1-16.

8.    Physical Properties of Common Organic Solvents.  The Sol-
     vents and Chemicals Companies.  (Promotional Literature).
     1976.

9.    EPA 450/2-78-032, p. 2-4.

10.  Personal communication with P. Russel, Abitibi Corporation,
     Cucamonga, California.

11.  EPA 450/2-78-032, p. 2-2.

12.  Price, M.B.  Reliance Universal, Inc., The Future of High-
     Solids Coatings.   In:  Proceedings of the Fourth Waterborne
     and Higher-Solids Coatings Symposium, New Orleans, 1977.  p.
     155.
                               3.9-19

-------
13.   Personal communications with B.  Christopherson,  and D.B.
     Carnagey,  Williamette Corporation,  Bend,  Oregon.

14.   Space-Age Coating of Particleboard Offers Enduring Surface.
     Furniture Methods and Materials,  May 1977.

15.   Personal communication with J. Davis,  Masonite Fabrication
     Corporation, City of Industry, California.

17.   EPA 450/2-78-032, p. 2-3.

18.   EPA 450/2-78-032, pp. 2-4,5.

19.   EPA 450/2-78-032, p. 2-6.

20.   EPA 450/2-78-032, p. 4-2.

21.   EPA 450/2-78-032, pp. iv-v.
                                3.9-20

-------
3 10  MAJOR SOURCE CATEGORY:  PETROLEUM LIQUIDS STORAGE-
      VOLATILE ORGANIC COMPOUNDS (VOC) EMISSIONS
3.10.1  Process Description
     In this  report,  petroleum liquids are defined as crude oil,
condensate,  and any  finished or  intermediate  products manufac-
tured  or  extracted  in a petroleum  refinery.   These definitions
are  intended  to  apply  to  tanks  with  capacities  greater than
150,000  liters  (40,000  gallons)  for  the storage  of petroleum
liquid  with  a  true vapor  pressure  greater  than 10.5  kPa (1.5
psi).
     The  three  major types of storage  tanks  for petroleum  liquid
are  fixed-roof, external  (open-top)  floating-roof,  and internal
 floating-roof.   Figures  3.10-1,   3.10-2,  and  3.10-3  are  cross-
 sectional  drawings  of  these tanks.
     Fixed-roof tanks are  generally cylindrical,  with a conical
 roof permanently attached  at the  top.  These  tanks are commonly
 equipped  with a pressure-vacuum vent that allows  them to operate
 at a  slight internal pressure  or  vacuum.   The  pressure-vacuum
 valves (vents) prevent  release  of vapors only during  very small
 changes in  temperature,  pressure,  or  liquid level.   These tanks
 are generally considered  to  represent  the  minimun  acceptable
 standard for the storage of petroleum liquids.
      The  external   floating-roof  tank  has  a  steel  cylindrical
 shell  fitted with  a  deck  or roof  that floats  on  the  liquid sur-
 face  and rises  and  falls with  the  liquid  level.   The  liquid
 surface is  completely covered by  the floating roof except in the
 small annular  space between the  roof and the tank shell.   A seal
 attached  to the  roof  contacts  the tank shell  and  covers  the
 annular  space.   The  seal  slides  against the  tank shell  as  the
 roof is raised or lowered.
      Internal  floating-roof tanks are fixed-roof tanks  in which
 an  internal roof or  cover  floats  on the  liquid surface.   Opera-
 tion  is  analagous  to the  external  floating  roof.  The two kinds


                               3.10-1

-------
                         CONSERVATION
                       (PRESSURE VACUUM)
                           VENT
                      GAUGE
                      HATCH
. MANHOLE
                             NOZZLE
                                                                      MANHOLE
                  Figure 3.10-1.   Fixed-roof  (cone-roof)  storage tank.
                   Figure  3.10-2.   Single deck,  external floating-roof storage
                                                    tank.2
                                                               FIXED ROOF
                         I
                           FIXED-ROOF
                           SUPPORTS-^
                                       FLOATING ROOF


                                     \ SLEEVE AND
\
                                              MANHOLE
 6UIOE AND   /
RUCTURAL COLUMNS
                                                                      MANHOLE
                   Figure 3.10-3.   Covered floating-roof storage tank with
                                         internal  floating cover.
                                              3.10-2

-------
of  internal  covers  commonly  used  are  pan-type  steel  covers  (known as
covered  floating roofs) and nonferrous  (plastic  or aluminum) covers  (known
as  internal floating  covers).   The  fixed roof shields  the internal roof  from
weather  effects  and  from  structural damage  or sinking  due to accumulated
rain or snow.

3.10.2  Emissions
     VOC  emissions  from  fixed-roof  storage  tanks  are  classified either as
breathing  losses or  working  losses.  These  losses occur when  the  working
limit of  the  pressure-vacuum vent  is exceeded.   Breathing losses consist of
vapors  expelled  from the  tank  because  of  the  thermal expansion/contraction
caused by diurnal temperature and barometric  pressure  changes.
     Working losses consist  of filling  and emptying losses.  Filling loss results
from vapor displacement by the  input of liquid.   Emptying  loss  results from
expulsion  of  vapor  after  product  withdrawal and  is  attributable to  vapor
growth as  the newly  inhaled air  becomes saturated with  VOC.
     Factors   affecting  the  emission  rate  are  the  true   vapor  pressure
(volatility) of  the  product  stored,  the  temperature  change  (diurnal) in the
tank vapor space, height of vapor space, tank diameter, schedule of emptying
and filling, and mechanical condition of the tank.
     Because  its  design  dictates  operation  over  a  small  pressure-vacuum
range, the fixed-roof  tank is  subject to appreciable breathing and  working
losses.   Pressure-vacuum valves  restrict vapor release during small  temper-
ature-pressure fluctuations, but  they allow  venting  during filling,  emptying,
and breathing.
     The  main source of emissions from  external floating-roof tanks is attri-
butable to wind-induced loss by the seal.  Wind-induced emissions occur when
the air  flow across the  tank creates  pressure differences  around the  circum-
ference of the  floating  roof  and  causes air to  flow  by the seal into  the
annular vapor  space on  the leeward  side  and air plus VOC to flow out on
                             3.10-3

-------
the windward  side.   Improper or  loose  fit of the  seal can  create  gaps  or
openings  between  the   seal  and   tank   shell,  which  can  cause  increased
wind-induced  losses.  Other VOC losses  are associated  with  evaporation  from
the wetted wall after emptying  and leakage of VOC  vapor  through the fabric
cover that bridges the space between  the  shoe seal and  floating roof.
     Internal  floating-roof tanks are  usually  equipped  with special air  vents
in the  fixed  roof or at  the  top  of the  shell.  These  provide ventilation to
minimize  the  possibility  of VOC  reaching  the flammable  range in the  tank
vapor   space.   Any VOC that flows  by  the seal   because  of wind-induced
pressure  differentials or  by other  seals  in  the floating  cover escapes  through
these vents.   The fixed roof reduces the  wind-induced pressure  differential
around  the  seal,  and  emissions  are  lower than  from external floating-roof
tanks equipped with single  seals.
     Reference  3  presents  formulas and  tables   of  emission  factors   for
calculating emissions from fixed-roof  tanks,  external floating-roof  tanks,  and
internal  floating-roof   tanks.   Because   of  the  many  variables   that  effect
emissions from these tanks, it is recommended that  these formulas be used to
calculate  emissions.  For external floating-roof tanks with secondary seals and
for all  internal floating-roof  tanks, standing storage losses  are computed by
using  Equation  (3),  Section  4.3.2.2  of  Reference 3, and  then reducing  that
value by 75 percent (i.e.  25%  Ls).   No  additional  reduction is recognized at
this time for  secondary seals  on  internal floating roofs.  The  basic equations
for  calculating emissions from floating roof  tanks  are being  revised by  the
American  Petroleum Institute  based on  recently  completed studies and  tests.
Appropriate revisions  will  be  made to Reference 3 when this work  is com-
pleted .

 3.10.3   Control Measures
      Besides  improving  air  quality  and eliminating  (or  at  least reducing)
 safety  (fire)   hazards,  the   control   of  VOC  emissions  effects  an  economic
 savings  through retention  of  valuable products.   The  most feasible and tech-
 nically  sound  methods  of  control  are  floating  roofs  and  vapor  disposal
 systems.
                               3.10-4

-------
 3.10.3.1  External Floating-Roof Tanks--
      External  floating  roofs  virtually  eliminate the  vapor  space  above the
 stored  liquid and  essentially  eliminate the working  and breathing losses  asso-
 ciated with fixed-roof tanks.   The critical  point is the sliding seal or closure
 covering  the  annular space between the tank  wall  and the roof.   The seal  is
 basically  either a  metallic  shoe seal or a nonmetallic foam-filled or liquid-filled
 seal.   Use  of  a  primary  (single)   seal  can  be  expected  to  reduce  VOC
 emissions by more  than 90 percent over uncontrolled (fixed-roof) tanks.
      The metallic shoe  seal is characterized by a 75  to  135  cm (30 to 51 in.)
 high long metal sheet  held against the  tank shell.   The shoe is attached  to
 the  roof  by braces and held  against  the wall  by springs or weighted levers.
 A flexible,  impervious fabric  (envelope) from  the top of the shoe seal to the
 roof  top  closes  the annular space  between  the  roof and the seal.   The vapor
 space is  restricted  to the  small  area  between the  liquid  surface  and the
 envelope.  (See Figure  3.10-4a).
      Emissions  occur  from the  exposed liquid  surface of  the  gap  spaces
 between the shoe  and  tank  wall,  and  through openings in  the  envelope  or
 shoe.   The envelope  and shoe conditions affect emissions since holes, tears,
 or  other  openings  provide a direct  path  between the  annular  vapor  space
 (bounded by the  shoe  liquid surface,  envelope,  and  roof)  and  the  atmo-
 sphere.
     The  nonmetallic seal,  usually  a  flexible tube filled with resilient foam  or
liquid,  is attached to the outer  periphery of the  floating roof and covers the
annular  space  between  the   roof  and   shell  (Figure  3.10-4,  b-d).   The
liquid-filled  seal  is  liquid-mounted  and   the  foam-filled  seal   is  either
liquid-mounted or vapor-mounted.  When the vapor-mounted seal is suspended
                             3.10-5

-------
    SECONDARY  SEAL
     (WIPER TYPE)
                            TANK
                            WALL
     METALLIC  WEATHER GUARD
                                              ^FLOATING ROOF
                                                SCUFF BAND
                                                 LIQUID FILLED
                                                     TUBE
-TANK
 WALL
     METALLIC WEATHER GUARD
                     =4-
            FLOATING ROOF
            SEAL FABRIC
            RESILIENT FOAM
                 LOG
           VAPOR SPACE
TANK
WALL
                                          METALLIC WEATHER GUARD
      Figure  3.10-4.   Primary seals.
                       3.10-6

-------
 above the liquid surface, an annular vapor space (bounded by the tank shell,
 seal, roof, and liquid surface) exists and any gap between the  seal and shell
 provides  direct  access  of  vapors to the  atmosphere.  Liquid-mounted  seals
 rest  on the liquid  surface  in  the annular space between  the shell and  roof.
 Thus the  annular vapor space is essentially eliminated.   The  fluid used  in
 liquid seals  should be  compatible with the stored products to avoid  contami-
 nation  if the  tube  ruptured.   Because of their flexibility,  nonmetallic  seals
 have some ability to adapt  themselves to  imperfections in  tank walls, thereby
 reducing  the gap  space somewhat.
      The use of a secondary seal  that is located above  and completely covers
 the  primary  seal  can  effectively control the VOC emissions  that escape by the
 primary seal  or  through the  shoe seal  envelope.   Two types  of  secondary
 seals are  commonly installed--rim-mounted (Figure 3.10-5) and  shoe-mounted
 (Figure  3.10-4a).   The rim-mounted secondary seal  is preferred since the
 shoe-mounted secondary seal does  not provide protection against VOC leakage
 through the primary shoe seal  envelope.
     Wind-induced air  flow  around the primary  seal system is the main cause
 of VOC losses  through  primary seals.  Improper fit and leakage through the
 shoe  seal envelope and shoe can also contribute  to VOC  emissions.  Reference
 4  states   that rim-mounted  secondary  seals  are  effective in  controlling
 emissions  from liquid  and vapor-mounted  primary seals.  The references  cited
 as the basis for this  statement are various industry sponsored test programs
 (Reference listings 5 through 10).
     In reference  4 the  reductions that would occur from installing  secondary
 seals over primary  seals have  been calculated for various  primary-secondary
 seal  type  combinations.   The  calculations indicate  that the  use of a  tight
 fitting,  rim-mounted  secondary seal over  primary  seals  effectively  curtails
 emissions  and  that  the  quantity  of  emissions  curbed increases  as  the  true
vapor pressure of the stored liquid increases.
                             3.10-7

-------
          RIM-MOUNTED
         SECONDARY  SEAL
                                        TANK
                                        WALL
    • TANK
    WALL
             RIM-MOUNTED
            SECONDARY SEAL
            ^-FLOATING ROOF
             SEAL FABRIC
              RESILIENT FOAM
                   LOG
              VAPOR SPACE
                                                 ^-FLOATING ROOF'

                                                SCUFF BAND
                                                LIQUID-FILLED
                                                    TUBE
                                      -TANK
 RIM-MOUNTED
SECONDARY SEAL
Figure  3.10-5.    Rim-mounted  secondary seals
                          3.10-8

-------
3.10.3.2  Internal Floating Roof Tanks--
     The  internal  floating-roof  tanks  described  in   3.10.1  are  currently
considered  equivalent to external floating-roof tanks equipped  with  secondary
seals (as discussed in 3.10.3.1).

3.10.3.3  Vapor Recovery/Disposal Systems—
     Two  general  types  of  these  systems  are  used  most  frequently  to
effectively  reduce  emissions  from  fixed-roof storage  tanks:   systems  that
thermally  destroy  collected  vapors  (disposal)   and  systems  that  recover
collected  vapors as useful product (recovery).   The thermal oxidizers (e.g.,
afterburners,  flares) combust  the  vapors  and  result  in  emissions  that  are
essentially  water  and  carbon  dioxide.   The vapor  recovery  systems  can
employ  compression,  adsorption, refrigeration,  and absorption principles to
recover  the VOC in the vapor as liquid product.  Control efficiencies range
from  85  to 98  percent,  depending  on the  physical properties  of the stored
liquids and the design of the equipment.
     Systems   can  be  designed  to  collect,   transport,  and  dispose of both
working  and breathing emissions from a series of interconnected tanks.  Such
systems  should  include  the  ability  to  isolate each  tank and  prevent vapor
backflow  from  the  entire  system  during  sampling,  gauging,  or pressure
change  at  an  individual tank.   Vapor  recovery/disposal systems  are most
often  found  at  gasoline  marketing  terminals;   they   have not  been  used
extensively on  large  tanks  or  at tank farms.   Figure 3.10-6  is  a  schematic
layout of  a  hypothetical  system.

3.10.4  Emission Limits
3.10.4.1   New Source  Performance  Standards--
     Present NSPS require  that storage vessels  with capacities greater than
151,412 liters  (40,000 gallons)  containing petroleum  liquid with a true vapor
pressure  greater  than 10.5  kPa (1.5 psia)   be equipped with an  external  or
internal  floating roof,  a vapor recovery system  (collection and disposal),  or
their  equivalents.  Tanks in which crude oil  and  condensate are stored before
                             3.10-9

-------
                                   VAPOR MANIFOLD HEADER
               RECOVERED PRODUCT RETURN
                          TANK FARM
                                           VENT
                    STRIPPER

                          ~
           PUMP
                                      ADSORBERS I
                                             S^
                                                            VAPOR
                                                            HOLDER
                                   VAPOR RECOVERY  SYSTEM6
Figure 3.10-6.   General schematic  of vapor  recovery   system.
 The principle,  design, mode of  operation, and components of
 systems vary.   Systems principles  qenerally include, one.or
 more of the  following: 'adsorption,  absorption, combustion,

 and refrigeration.
                             3.10-10

-------
custody  transfer are  specifically exempted.   A revised NSPS proposed in May
1978 has  not  yet  been promulgated and is therefore not applicable until final
promulgation.  Since  promulgation  of the standard may occur in  late 1979, the
main  features  of  the proposed  standard  are  relevant  and worthy  of mention
here.   The  proposed standards  would require  a  floating  roof; an  external
floating  roof with double  seals that meet  specified  gap  limitations or a  vapor
recovery  system;  or the equivalent of either  device.   The exemption on  crude
oil  and  condensate   tanks  was  removed.   A  final  review of  the  comments
received on the proposal and the most recent emission data is being completed.
The promulgation  will reflect changes to the proposal as  a  result of the final
review.

3.10.4.2  State Implementation Plans--
     The most stringent  SIP  requirement discovered was  the  South  Coast
(California)  Air  Quality  Management   District  Regulation  463  (Storage  of
Organic  Liquids).    This  SIP  limitation  requires   floating  roofs  or   vapor
recovery  systems (or the equivalent)  for liquids  of a true  vapor pressure
between  78  to  570 mm Hg  (1.5 to 11.1 psia),  inclusive, stored in tanks with a
capacity  greater  than 150,000 liters  (39,630  gal);  double  seals on  external
floating-roof tanks;   specific allowable  gaps  between  seal  and   tank  wall  in
terms of width and percent of circumference;  and internal  floating-roof covers
on  fixed-roof  tanks   or  vapor recovery/disposal systems  of  equal reduction
efficiency.

3.10.5  Determination of LAER
     The  recommended  limitations  are  based  on SIP's and  on performance
information  available   in early 1979.  It is anticipated  that  several additional
SIP regulations covering these sources will be  promulgated and/or modified in
1979 and 1980 and that appreciable new performance data will become available
in the near term.  Conceivably,  some  SIP regulations may be more stringent
than  the LAER  suggested  herein.   Furthermore,  performance  testing may
show  that more stringent limits  than those suggested are feasible or it may
show  that the  suggested limits are  appropriate or that they are not achievable
                             3.10-11

-------
for some specific subcategories.  In any case, the basis  for determining LAER
for many  source categories  is  expected  to  change frequently.  Since LAER is
near  the  vanguard  of  control  technology,  a  more detailed analysis is partic-
ularly necessary when addressing  modified or  reconstructed  facilities subject
to the provisions of Section  173 of  the Clean Air  Act.  Emission limitations
reasonable for new sources  may in  some  instances be  economically  or tech-
nically unreasonable when applied  to modified or reconstructed sources of the
same  type.
     The  LAER recommendations below  are intended for application  to petro-
leum  liquid storage  tanks greater than 150,000 liters (40,000  gallons) contain-
ing petroleum  liquids with a true  vapor pressure greater than  10.5  kPa  (1.5
psi).   The suggested  requirements  are in the form  of devices that contain
and prevent emissions rather than  a  specified emission limit or rate.
     The  recommended  LAER for petroleum  liquids storage in  external floating-
roof tanks is the application of double-seal technology as briefly discussed in
Section 3.10.3.   (An  expanded treatment of  double-seal technology can  be
found in   Reference  4.)  Double-seal  technology consists of a  primary  seal
(either a metallic-type shoe  seal;  a  liquid-mounted foam-filled  seal;  a liquid-
mounted liquid-filled type seal;  or  another  equally effective closure device)
and  a continuous secondary  seal extending from the floating roof to the  tank
wall  (a rim-mounted secondary).  The seal closure  devices should  have  no
holes, tears, or other visible  openings in  the seal or seal fabric; should  be
intact;  and have  a  uniform, tight  fit between  the roof  and  tank wall around
the  roof  circumference.   All  openings in the  roof,  except rim  space-vents
should  project below  the liquid  surface.   Openings,  except for rim space-
vents,  automatic bleeder vents, leg sleeves,  and emergency  drains,  should be
equipped  with a cover or lid  that is closed  at  all time  except when in actual
use.
      The  importance of minimizing  the gap  between the rim-mounted  secondary
seal  and  the  tank  wall in preventing  VOC emissions  is emphasized.  At this
time  specific  gap  requirements  have  not  been  promulgated  by  EPA.   It is
suggested that the gap  area specified  in Reference 4,  page 5-2, be used as
an interim guide until  more  definitive requirements become available.
                              3.10-12

-------
Reference 4 suggests  that the gap  area of gaps exceeding 0.32 cm  (1/8 inch)
in width between the secondary seal and the  tank  wall not exceed 6.5 cm  per
0.3 M of tank diameter (1.0 square  inch per  foot of  tank  diameter).
     For fixed roof  tanks,  the recommended  LAER is an internal floating-type
cover with a  single  continuous closure device  between the tank shell and the
floating-cover edge.
     Vapor recovery/disposal  systems of equivalent VOC control capability can
also be used  to meet  the LAER  requirements,  as  can  other systems  that have
satisfactorily demonstrated an equal reduction performance.
     The   reader   is   reminded  that  should  the  revised   NSPS  sched-
uled   for   promulgation   in   late-1979   contain    any   provisions   more
stringent   than    this   LAER   determination,    then,    under    Section
171(3)  of  the  Clean  Air  Act   as  Amended   August  1977,  these  pro-
visions  of the NSPS  would apply.
                             3.10-13

-------
                           REFERENCES
1.   Control of Volatile Organic Emissions from Storage of Petro-
     leum Liquids in Fixed-Roof Tanks.   Office of Air Quality
     Planning and Standards,  U.S.  Environmental Protection
     Agency, Research Triangle Park,  North Carolina.

2.   Background Information for Proposed New Source Performance
     Standards:  Asphalt Concrete Plants,  Petroleum Refineries,
     Storage Vessels, Secondary Lead  Smelters and Refineries,
     Brass or Bronze Ingot Production Plants, Iron and Steel
     Plants, Sewage Treatment Plants,  Volume 1.  Office of Air
     Quality Planning and Standards,  U.S.  Environmental Protec-
     tion Agency, Research Triangle Park,  North Carolina.   June
     1973.  Main Text pp. 32-33.

3.   Compilation of Air Pollutant Emission Factors.  U.S.  Envi-
     ronmental Protection Agency,  Research Triangle Park,  North
     Carolina.  April 1977.  Section  4.3.

4.   Control of Volatile Organic Emissions from Petroleum Liquid
     Storage in External Floating Roof Tanks.  EPA-450/2-78-047,
     Guideline Series, Office of Air  Quality Planning and Stan-
     dards, U.S. Environmental Protection Agency, Research
     Triangle Park, North Carolina.  December 1978.  p. 3-1.

5.   SOHIO/CBI Floating-Roof Emission Program, Interim Report.
     October 7, 1976.

6.   SOHIO/CBI Floating-Roof Tank Emission Program, Final Report.
     November 1976.

7.   Chicago Bridge & Iron Company.  Western Oil and Gas Associ-
     ation, Metallic Sealing Ring Emission Test Program, Interim
     Report.  January 1977.

8.   Chicago Bridge & Iron Company.  Western Oil and Gas Associ-
     ation, Metallic Sealing Ring Emission Test Program, Final
     Report.  March 1977.

9.   Pittsburgh-Des Moines .Steel Company.  Measurement of Emis-
     sions  from a Tube Sear^ Equipped Floating Roof Tank.
     October 9, 1978.

10.  Western Oil and Gas Association, Metallic Sealing Ring
     Emission Test Program, Supplemental Report.  June 1977.


                              3.10-14

-------
3.11  MAJOR SOURCE CATEGORY:  PETROLEUM REFINERIES—WASTEWATER
      SEPARATORS,  PROCESS UNIT TURNAROUNDS,  AND VACUUM PRODUCING
      SYSTEMS—VOLATILE ORGANIC COMPOUNDS (VOC).
3.11.1  Process Description
3.11.1.1  Wastewater Separators—
     Petroleum refineries operate wastewater treatment facilities
to meet standards for process effluent.  The wastewater separator
is  the  first stage of  water treatment.   Contaminated wastewater
is  collected from  process  drains,  storm sewers,  and equipment
cleaning  operations.    Additionally,  water  from  leaks,  spills,
pump  and  compressor sealing, flushing,  and  sampling may go into
the wastewater  system.   Because  rains can cause sudden surges in
the  volume of  water  handled, the storm sewer  water  is usually
segregated and  retained in  a separate holding pond before treat-
ment in the wastewater  separator.  The wastewater separator skims
oil from the water  and  returns it  to the process.  The wastewater
then  undergoes  additional  treatment  to  meet effluent standards.
     Refinery wastewater drains and treatment facilities can emit
volatile  organic  compounds  (VOC)  by  evaporation.   Volatile or-
ganics may be  emitted at any place that wastewater is exposed to
the  atmosphere.   Sealing  sewer  openings by  the use  of liquid
traps or  other  devices minimizes VOC emissions.  Refinery opera-
tors have used  such sealing devices for many years because of the
safety hazard associated with hydrocarbon-air mixtures.
     Wastewater  separators,  the   most  common  of which  are API
separators,  are possibly the largest source of VOC emissions from
wastewater treatment.   An API separator  is a gravity differential
device consisting of a  series of baffles  and weirs in  a container
operated  at  atmospheric pressure.   Because the  input  is by grav-
ity flow  from  an  underground  sewer,  the  tank or  container is
underground.  Although  early practice did not include  covering of
                              3.11-1

-------
these tanks, some refiners  now  cover the forebays of API separa-
tors to control VOC  emissions  and a few cover the entire separa-
tor area.

3.11.1.2  Process Unit Turnarounds—
     Periodic  process unit  turnarounds  are  required  to  ensure
safe  and  efficient  operation  of equipment.   Insurance require-
ments,  safety  regulations,  and industry  standards require peri-
odic  inspection  and  maintenance  of all  vessels.   Thus, process
unit turnarounds are an essential operation in a refinery.
     A  turnaround  entails emptying  the  equipment for inspection
and  maintenance.   A  typical   procedure  involves draining  the
liquids from the vessel;  depressuring the vessel;  flushing the
vessel  with  water,  steam,  or inert gases to remove all  hydrocar-
bons; and  then air  blowing to provide oxygen for  breathing.  The
operations  provide  many  opportunities  for VOC  emissions  to the
atmosphere.  The vapor content of vessels is usually vented to  a
flare  system,   through a  furnace  firebox,  to   a  vapor recovery
system, or  directly  to the  atmosphere,  depending  on  operating
practice and refinery  configuration.

3.11.1.3   Vacuum-Producing  Systems—
     Use  of vacuum-producing   systems  is principally  at vacuum
towers  for  distillation  of the  very  high-boiling components of
crude oil  from residuum fractions.   Other uses include  evacuation
of processing  equipment for maintenance  and  for removing  noncon-
densables  from condensers that  process steam  from turbines.
      This  separation of high-boiling fractions  (340°  to 566°C or
650°  to 1050°F  gas  oils)  from  residuum is  required because it
renders the gas  oils  more valuable  for subsequent  processing.
Separation is  accomplished  by  distillation under vacuum  because
the oil will  crack  at temperatures required to  distill it  under
atmospheric pressure.   Distillation  under  vacuum  greatly  in-
creases the overall yield  of gas oil  fractions  of the crude oil.
                               3.11-2

-------
     Distillation  involves  heating  the  so-called  reduced  crude
oil  to  about 400°C  (750°F)  and sending  it to a  tower operated
under vacuum.   The flash zone  at  which the  vapors and residuum
are  separated  typically operates   above  370°C  (700°F)  and  at
absolute pressures of  1.3 to 16 kPa (10 to  120 mm Hg).  Instru-
ments  on  the  vacuum-producing system  control  the  flash  zone
pressure and,  indirectly,  the pressure at the  top of the tower.
Effective pressure  in  the flash zone  may be  further reduced by
introduction  of steam  with  the  feed or  at  the  bottom of  the
tower, the  latter  providing  a  stripping action on the residuum
product before it leaves the bottom of the tower.
     The vapor  portion of the  feed flows upward  through an en-
trainment  removal  section and  then through  distillation trays,
which intimately mix the vapors with cooled heavy-vacuum gas oil
pumped back  into the tower  at  a higher  level as  reflux  and as
cooling fluid.  The remaining vapors then proceed through a simi-
lar section with light-vacuum gas oil as reflux and coolant.   The
vapors  that  leave  the  top  of the tower  are exhausted by  the
vacuum-producing system.
     The basic types of vacuum-producing systems are ejectors and
mechanical pumps (compressors).  An ejector system is a gas com-
pressor that  uses  the  high-pressure energy of one fluid  to  com-
press another.   The  high-pressure  fluid  is  almost always steam,
which is "jetted"  from  a nozzle through a vacuum chamber at high
velocity,  entraining the vapors from the tower.   The steam  from
the  ejector  is  then  condensed by contact  with  water  (either
directly or through tubes of a water cooler) or by an air cooler.
The  noncondensables  are  vented to the  atmosphere,  to  another
stage of jet  ejection,  or to a blower.   The vacuum towers in oil
refineries usually require three ejectors in series.
     For  the  large  ejector systems  required on  large  vacuum
units, the  steam  is condensed  and  the  gas  is cooled.   The  re-
sultant mixture  of water, hydrocarbons,  and  dissolved gases  is
sent  to  a  ground-level  accumulator called  a  "hotwell."   Water
                              3.11-3

-------
from the hotwell requires treatment to remove oil,  and vapors from the hotwell
may  require  containment and  disposal.  Use of a  surface  condenser, either
water-  or  air-cooled,  greatly  decreases the  volume of water  that  must  be
treated,  because  the cooling  water is  not  contaminated  with  hydrocarbons.
     Although refineries use steam ejectors almost exclusively because of their
initial  low  cost,  lack of moving  parts,  operating simplicity,  and  reliability,
vacuum  pumps  also  can produce  the  necessary vacuum.   Such  pumps  are
being  considered  more  and more  because  of increasing  concern with  water
pollution and because  of  rising  energy costs, which favor pumps  over  the
very  thermally  inefficient  steam ejectors.   No  refinery  is  known  to have
installed vacuum pumps.  They are particularly expensive for use at  very low
vacuums below  50  mm Hg absolute  pressure.  Although the  fuel savings
offered by  vacuum  pumps  is  substantial,  the pumps  are  much less reliable
than  steam  injectors  and  cost  much  more  to install and maintain.   It is
anticipated  that serious consideration  will  be  given to  vacuum  pumps  for
application to new vacuum towers on the basis of  energy cost savings.

3.11.2  Emissions
3.11.2.1  Wastewater Separators--
     Because the  VOC at  wastewater  separators  come  from many  refinery
sources  served by  the collection system, volatility  of  the  compounds  emitted
ranges widely.   Typically, the compounds include  Cj to  Cg paraffins, olefins
and aromatics,  hydrogen, hydrogen sulfide,  and ammonia.   None of the data
adequately  quantified  emissions  from wastewater separators,  although  it  is
estimated  that  the  emissions  from uncovered wastewater  separators  can be
large.

3.11.2.2 Process Unit Turnarounds—
      Potential  emissions from  turnarounds  include  the full range  of hydro-
carbons in  a refinery.  Emissions also may include hydrogen, ammonia, carbon
monoxide,  hydrogen sulfide,  sulfur oxides, and particulates.
                              3.11-4

-------
3.11.2.3  Vacuum-Producing Systems—
     The barometric  condenser hotwell accumulators  and the last
stage of the steam ejector system emit light, noncondensed hydro-
carbon  compounds  (VOC),  which would  generally include hydrogen,
methane, ethane,  propane, butanes,  pentanes, and  hydrogen sul-
fide, together with low concentrations of olefins and uncondensed
gases (nitrogen, oxygen,  and CO-).

3.11.3  Control Measures
3.11.3.1  Wastewater Separators—
     Volatile organic emissions from wastewater separators can be
reduced and  well  controlled by  covering the  separators  and all
openings to  enclose  completely the liquid  contents.   A floating
cover equipped  with  closure  seals between the edges of the cover
and  the separator can minimize VOC emissions.   Use of separator
covers  and liquid trap seals for sewer drains should minimize VOC
emissions from wastewater systems.

3.11.3.2  Process Unit Turnarounds—
     Control  of VOC  emissions during  process  unit  turnarounds
starts  with  the pumping  of liquid contents from the process unit
to storage.   This  operation now  involves  enough economic incen-
tive  for complete  compliance to  be anticipated.   Collecting the
vapors  from  the depressuring of vessels is  the  next  most effec-
tive  step.   The purged vapors along with vapors  forced from the
system  by  flushing  of the  process  unit with water,  steam,  or
inert gas can be  directed to a vapor control  system,  to  a flare
header  system, or to a firebox for incineration.

3.11.3.3  Vacuum-Producing Systems—
     Vapors  from the  last stage  of a  vacuum-producing system and
from  the hotwell  can be  collected and sent  to a furnace  firebox
or to a vapor recovery unit with the assistance  of a mechanical
                              3.11-5

-------
blower (compressor).  Both  recovery  and use-as-fuel systems pro-
vide  incentives  related  to  the concern  with  air quality  and
energy costs and the  trend  to larger vacuum units in refineries.

3.11.4  Emission Limits
     No New Source Performance Standards (NSPS) have been promul-
gated for wastewater  separators,  process turnarounds,  or vacuum-
producing  systems   at refineries, nor is  information available
upon which to base an achieved-in-practice emission limit.

3.11.4.1  State Implementation Plan  (SIP) Limits—
     Specific  regulations  applicable to vacuum-producing systems
and  wastewater  separators  are  embodied  in  California's  South
Coast Basin  rules; these  are the most  restrictive found in any
SIP.   No  regulation  specific  to process  unit  turnarounds was
found.
     Under  Rule 465  of the  California South  Coast Air Quality
Management District (SCAQMD), the discharge from vacuum-producing
devices  is  limited to  no  more  than 1.36  kg  (3.0 Ib) of organic
materials  in  any   1  hour  from  any  vacuum-producing  devices or
systems,  including hotwells  and accumulators,  unless such  dis-
charge  has been reduced by  at  least 90 percent.  In  other  areas
of the  country VOC emissions are limited under a general  rather
than  a  source-specific  regulation.   For  example,  the Bay Area Air
Pollution  Control  District,  San Francisco,  California,  (Regula-
tion  3101)  states "....a  person shall not  discharge  into the
atmosphere an  effluent  containing a  concentration of more than 50
ppm or organic compounds calculated as hexane (or 300 ppm  total
carbon)."
      Wastewater separators  are  addressed  specifically in  SCAQMD
Rule  464 of  the California  South Coast Basin,  which states  that  a
person  shall  not  use  any  compartment  of  any  vessel or  device
 operated for the recovery  of oil from effluent water,  unless such
                               3.11-6

-------
compartment is  equipped with  one of the  following vapor-loss control devices:


     a.   A solid cover on which all  openings  are  sealed, thereby
          totally enclosing the liquid contents  of the  compart-
          ment.

     b.   A floating pontoon or double-deck type cover  equipped
          with closure seals between  the cover's edge and the
          compartment wall.  [Gaps between the compartment wall
          and seal are limited to 0.32 cm  (1/8 in.)  for an accum-
          ulative length of 97 percent of  tank perimeter, and are
          not to exceed 1.3 cm (1/2 in.) for an  accumulative
          length of the remaining 3 percent of tank perimeter.
          No gap is to exceed 1.3 cm  (1/2  in.) in  width.]

     c.   A vapor recovery system that reduces the emission of
          all hydrocarbon vapors and  gases into  the  atmosphere by
          at least 90 percent by weight.

     d.   Other equipment having an efficiency equal to or
          greater than a, b, or c (listed  above) and approved by
          the Air Pollution Control Officer.

3.11.5  Determination of LAER;  Controls to Achieve LAER

3.11.5.1  Wastewater Separators—

     In  the  absence  of  achieved-in-practice  data and  NSPS,   the  lowest

achievable emission  rate  is determined  to be  that degree  of  VOC emission

reduction required  by the  South Coast Air Quality Management District  Rule

464 (see  Section 3.11.4.1).  The LAER is therefore a  vapor-recovery  system

that reduces  all VOC  by at least 90 percent  by  weight,  or  certain  specified

equipment,  or other equipment of equal or greater control efficiency.  Control

equipment includes  the covering of separators,  forebays, and  other openings

with seal-equipped  floating roofs; use of  liquid trap  seals for  sewer drains;

and a monitoring and inspection program to detect any  excessive VOC releases
to the  drainage  system.


3.11.5.2  Process Unit Turnarounds--

     LAER for  process unit turnarounds  is that  emission rate resulting  from

collecting and subsequently either incinerating or  recovering the vapors
                              3.11-7

-------
released   during  the  depressuring  and  purging  of  process  units.    It  is
estimated  that  LAER  is an  overall  reduction of  90  percent  by  weight,  or
greater in potential emissions from the turnaround process.   A piping system
would be  required to convey  the VOC  emissions to the  control device.

3.11.5.3   Vacuum-Producing Systems--
     The  recommended  limitations  are based on SIP's and on  performance
information  available in early 1979.   It is  anticipated  that  several  additional
SIP regulations covering these sources will be promulgated and/or modified in
1979  and  1980 and that  appreciable new performance data will become available
in the near  term.   Conceivably,  some  SIP  regulations may be more stringent
than  the LAER  suggested  herein.   Furthermore,  performance  testing may
show  that more  stringent  limits  than those suggested are feasible  or it may
shew  that the  suggested limits are appropriate or that they are not achievable
for some  specific subcategories.   In any case, the basis for  determining  LAER
for many  source categories is expected to  change frequently.  Since LAER is
near  the  vanguard  of  control technology,  a  more detailed analysis  is partic-
ularly necessary  when  addressing  modified or  reconstructed facilities  subject
to the provisions of  Section  173 of  the  Clean Air  Act.  Emission  limitations
reasonable for  new sources  may in  some instances be economically or  tech-
nically unreasonable when  applied to modified or reconstructed sources of the
same  type.
     LAER for vacuum-producing  systems is a 90 percent  by weight reduction
of the VOC emissions  from  the  last  stage  of  the system and the  associated
hotwell.   This  reduction anticipates  effective capture,  collection, and trans-
port  of VOC to  the control unit by  the  collection  system.  Available controls
to achieve LAER  include an  effective  collection system and  vapor disposal  by
either incineration (e.g., a furnace firebox) or vapor  recovery.
                              3.11-8

-------
                           REFERENCES
1.   Monroe, E.S.  Energy Conservation and Vacuum Pumps.  Chem-
     ical Engineering Progress, Vol. 71, No. 10.  October 1975.
     pp. 69-73.
                              3.11-9

-------

-------
 3.12  MAJOR  SOURCE  CATEGORY:   PETROLEUM  REFINERIES—FUGITIVE
      EMISSIONS—VOLATILE  ORGANIC  COMPOUNDS  (VOC)
 3.12.1   Process Description
     Equipment  leaks provide  numerous  sources of fugitive vola-
 tile  organic compound  (VOC)  emissions  in petroleum  refineries.
 Among the  equipment items that are  subject  to  leaks are pump  and
 compressor  seals,  valve  stems  and bonnets,  flanges, pressure-
 relief devices, process  drains, sample valves,  and other open-end
 valves.
     The proper  mating  of two sealing  surfaces  is  essential if
 pump  seals,   compressor  seals,  valve  stem  seals,  valve  seats,
 bonnets,  flanges,  and  other  connections  are   to  maintain their
 sealing  ability.    These  sealing  surfaces  are  sometimes  finely
 machined  surfaces  and   sometimes  compressed  packing   (gaskets).
 Proper design and  maintenance  are required  to minimize leakage.
     Properly seated pressure  relief devices will provide a good
 seal  until  they  have been  opened and  closed  several times  and
 foreign matter (such  as corrosion products,  carbon particles,  and
 gum) collects on  the  seats, after which  the  seal may no longer be
 free of leaks.
     The open-end  valves  used for  sampling and  bleeding lines
 become a source of  fugitive emissions if they are not completely
 closed or are not seating properly.

 3.12.2  Emissions
     Although  a  wide range   of  volatile organic compounds  is
 emitted  as  a result of  equipment leaks, most  are  hydrocarbons
with one to  six  carbon atoms,  such as  olefins,  paraffins,  and
 aromatics.   Their composition  depends on the  kinds  of crude  oil
processed in  the  refinery, the complexity of  the  refining proc-
ess, and the specific processing  units  involved.  The volume of
pollutants  emitted  depends on these same  items in addition to
maintenance practices and refinery age.
                              3.12-1

-------
     Available data regarding emissions from fugitive sources are
limited,   and  EPA  is  currently  conducting  studies to  quantify
these emissions.   Data regarding leak rates and VOC emissions for
fugitive refinery  sources  recently became  available as  a result
of one  such study.1  These are presented  in Table 3.12-1,  which
indicates the  percentage  of  all  sources  of  a  given type  that
exhibited leaks  (estimated percent leaking) and  the average VOC
emissions from  those  that were leaking  (emission  factor  esti^
mate).   The data  were derived  from  studies  at  nine  refineries
(large and  small,  new and  old) and from an examination of 500 to
600 sources at each of the refineries studied.

3.12.3  Control Measures
     Many  design  and  operational measures   are  available  for
reduction of  emissions at petroleum  refineries.   A well planned
and  executed  monitoring and maintenance  program is one of the
most effective methods.  For many years refineries have used such
programs  to prevent the formation of  vapor clouds and to elimi-
nate  the possibility  of  explosion;  however,  only recently have
refiners  used  portable  flame  ionization detectors  and  other
analytical  instruments to  determine total hydrocarbon levels near
major  process  equipment.   Analytical  instruments  capable  of
measuring hydrocarbon concentrations  in  a range  from 1  or 2 to
several  thousand parts per  million  are necessary for  this type
monitoring.
     Because  fugitive emissions  at pump seals, compressor seals,
and  valve  stems  occur at  random and cannot be predicted, a peri-
odic  walk-through inspection and monitoring program are  required
to  detect  leaks  if  emissions  are to be reduced  significantly.
Although initially expensive, a properly designed and permanent
monitoring  system would serve  to  discover, relay to,  and  alert
refinery personnel to the presence of VOC and the need  for cor-
rective  action.
                               3.12-2

-------
                          TABLE  3.12-1.  ESTIMATED EMISSION FACTORS.
                        FOR SELECTED PETROLEUM REFINERY VOC SOURCES
              Source type
                                   Estimated
                                    percent,
                                    leaking
                                                      Emission  factor  estimate
           g/h per source (Ib/h per source)
U)
NJ
I
U)
Valves             j
  Gas/vapor service
  Light liquid/two-phase service
  Heavy liquid service

Pump seals
  Light liquid service
  Heavy liquid service

Compressor seals
  Hydrocarbon service

Flanges

Drains

Relief valves
                                         29.3
                                         36.5
                                          6.7
                                         63.8
                                         22.6
70.3

 3.1

19.2

39.2
                     21.0   (0.047)
                     10.0   (0.023)
                      0.3   (0.0007)
                    118.0   (0.26)
                     20.0   (0.045)
440.0   (0.98)

  0.26  (0.00058)

 32.0   (0.07)

 86.0   (0.19)
      Adapted from Table 1-1,  p.  2,  Reference 1.

      Estimate of percent of total sources found to be leaking.   A leaking source is
      defined as a source with measured leakage values equal to  or greater than 200 ppmv
      or sources with measured leak rates greater than 0.00454 g/h (0.00001 Ib/h).

      Emission factor estimate for leaking sources only,  i.e., those found to leak as
      defined in "a" above.

      Gas/vapor service means the hydrocarbon stream was a vapor at process conditions;
      light liquid means a liquid lighter than kerosene;  liquid means a liquid equal to
      or heavier than kerosene.

-------
     Design changes  that can  reduce fugitive  emissions at  new
facilities  include  pumps  and compressors  equipped with  fluid-
flushed,   double mechanical  seals  and  a  vapor  collection  and
recovery system.   (Pumps  with double seals and various  kinds of
flushing  fluids  are currently used  by refineries  when  handling
toxic  fluids.)   Other design  changes include  enclosures  around
flanges and  the use of  double  pipes; however, the high initial
cost and  even higher maintenance costs  render both impractical.

3.12.4  Emission Limits
     No New  Source Performance Standards  (NSPS)  limits  apply to
fugitive  emissions  from refineries,  and  Achieved in  Practice
(AIP) limits have not been adequately quantified.

3.12.2.1  State Implementation Plan  (SIP) Limits—
     In many states  VOC emission regulations require  that all
persons use reasonable  care  to avoid discharge,  leaking,  spil-
ling, seeping, pouring, or dumping of compounds, or that they use
known  and existing  vapor control emission devices  or  systems.
     California's  South  Coast  Air  Quality Management  District
(SCAQMD) Rule No. 466  (adopted May 7, 1976, and amended  September
2,  1977)  prohibits the  use  of any  pump or compressor to handle
organic  materials with  a Reid vapor pressure of  80  mm mercury
(1.55  lb/in.2)  or greater  unless  such pump  or  compressor is
equipped  with  a  mechanical  seal  in good  working  order or  some
other  device of equal or greater  efficiency and approved by the
Executive Officer.   The  rule  specified that mechanical seals be
maintained so  that  there is  (1)  no  leakage  greater than  three
drops  per minute,  (2)  no visible mist  from liquid being pumped
where  such liquids  do  not condense at ambient  conditions,  and (3)
no visible  indication of leakage  evident at  or  near  the  seal/
shaft  interface  of gas  compressors.  An inspection  for  visible
 leaks  at  pumps  and compressors  is required  once every 8-hour
period unless the refinery is  located more than 3  miles
                               3.12-4

-------
from  a continuously  manned  control  center.   In this case an inspection  is
required once every 24 hours.
     SCAQMD Rule  No.  467  (adopted May 7,  1976) prohibits  the  use of any
safety pressure relief valve on any equipment that handles organic materials
                                                                   o
with  an  absolute  pressure of more  than 776 mm mercury  (15  Ib/in. ) unless
the relief valve  is vented to a vapor recovery  or disposal system, is protected
by  a  rupture disc,  or is maintained by an inspection system approved by the
Air Pollution Control  Officer.   SCAQMD Rule No. 466.1 (adopted November 3,
1978)  pertains to valves  and flanges.  The general requirements are record
keeping, periodic  leak inspection, leak  repair within 2 days (some  exceptions
allow  a greater  repair time), and the use of a seal  (cap,  plug, or  flange) on
each  valve at the end of a pipe or line containing  VOC when  the  line is not
in use (certain valves exempted).

3.12.5  Determination  of  LAER
     The  recommended  limitations  are  based  on SIP's and  on  performance
information  available in  early 1979.   It is anticipated  that several additional
SIP regulations  covering these sources will be promulgated and/or modified in
1979 and 1980 and that appreciable new  performance data will become available
in the near term.   Conceivably,  some SIP regulations may be  more  stringent
than  the LAER  suggested herein.   Furthermore,  performance  testing  may
show  that  more stringent  limits  than those suggested  are feasible or it may
show  that the suggested  limits are appropriate or that they are not  achievable
for some specific subcategories.   In any case,  the basis for determining LAER
for many  source categories is expected  to change frequently.   Since LAER  is
near  the  vanguard  of control technology, a more detailed  analysis  is  partic-
ularly necessary when addressing modified or reconstructed facilities  subject
to the provisions  of  Section 173 of the Clean Air  Act.   Emission  limitations
reasonable  for new  sources may in some instances be economically or  tech-
nically unreasonable when applied to modified  or reconstructed  sources of the
same  type.
                             3.12-5

-------
     The  complexity of processing units used in  refineries varies,  depending
on the different crude oils  processed and products  produced.   These differ-
ences  among  refineries and refinery  operations  and  the fact that fugitive
emissions  occur randomly,  originate  from  numerous  sources,  and  are  not
precisely  quantified,  make it  difficult  to   establish  an  acceptable national
standard  for limiting mass emissions.
     The  lowest achievable emission rate (LAER) for  fugitive VOC emissions at
petroleum  refineries  is  therefore  determined  to  be  the  emission  rate  that
occurs  after  equipment   designed  to  minimize  emissions  has  been  properly
installed  and  a  program of practical  housekeeping, monitoring, routine inspec-
tion,   preventive  maintenance,  and  equipment repair/replacement  has been
established.   Where  technically  feasible,  pumps and  compressors should have
double mechanical seals that are fluid-flushed,  and the flushing fluid facilities
should be equipped with  a vapor collection/recovery  system.  Certain types of
compressors  can be equipped with appropriate purge  systems.   The provisions
of  California's   SCAQDM   Rule  466   (see  3.14.4)   in  regard  to   inspection
schedule,  leakage rate,   and visible  emissions are  considered applicable as
LAER  guidelines  for  pumps  and  compressors.  The inspection-maintenance
program  should include a periodic walk-through  inspection using visual tech-
niques and,  where  appropriate,  be supplemented  by hydrocarbon  detection
equipment  to  provide   early  detection  of  existing or developing  leaks.
Particular attention should be directed to pipeline  valves,  sampling  equipment,
and  pressure  relief valves.   All   leaking  or defective   components  should
command  early  attention  and be isolated  and  repaired or replaced  to  protect
air quality  and safety.
                               3.12-6

-------
                      REFERENCES
Emission Factors and Frequency of Leak Occurrence for Fit-
tings in Refinery Process Units.  EPA-600/2-79-044,  U.S.
Environmental Protection Agency, Industrial Environmental
Research Laboratory, Research Triangle Park, North Carolina
February 1979.
                        3.12-7

-------
3.13  MAJOR SOURCE CATEGORY:  GRAPHIC ARTS PRINTING--VOC
      EMISSIONS
     In 1970 there were  more than 40,000 printing and publishing
establishments  in the United  States.    Total  solvent  usage  in
                                                                2
this industry  in 1976 was  estimated at  340  Gg (380,000 tons).
With an  overall degree of emission  control estimated  at 30 per-
cent, the  total  of  national organic compound emissions was esti-
mated at 240 Gg  (270,000 tons).  The approximate contributions of
each major  graphic  arts  process are gravure,  41 percent; lithog-
raphy, 28  percent;  letterpress,  18  percent;  and flexography,  13
        2
percent.

3.13.1  Process Descriptions
     High-volume  printing operations  utilize rotary  presses  in
which  the  image carrier  is  curved and  mounted on  a rotating
cylinder.   In  gravure,  the  image  cylinder  rotates  in  the  ink
trough  (ink  fountain).   In  other  processes a  second cylinder
rotates  in an ink trough and  delivers  ink to the plate cylinder
(image),  usually through  a series  of  distribution  rollers.   In
direct printing  operations, the image carrier transfers the image
directly to the  print  surface.  In offset  printing operations, an
intermediate  surface  transfers the  image to  the print surface.
     Printing methods  are classified by the principles upon which
the  printing  image carriers are  based.   Two  types are discussed
herein:  rotogravure and  flexography.
     Printing  presses  are  fed by sheet or roll paper.  Roll-fed
operations  require  dryers to evaporate the solvent from the ink.
Dryers  are high-velocity hot-air  dryers,  direct-flame dryers, or
indirect steam-heated  dryers.

3.13.1.1   Rotogravure—
      In  gravure printing,  the image areas are recessed relative
to  the nonimage  areas.   The gravure  cylinder  rotates in an ink
                               3.13-1

-------
trough  or  fountain.  Excess  ink  is  removed by  a  steel  doctor
blade.  The paper is pressed against the cylinder as it turns, by
use  of  a  rubber-covered  impression  roll.   When the  process is
roll-fed it  is known as  rotogravure.   Sheet-fed gravure  is not
widely used.
     Rotogravure requires very  fluid  inks,  with solvent contents
ranging  from  55  to  75 percent  and  higher.   Typical  solvents
include  alcohols,   aliphatic  naphthas,  aromatic  hydrocarbons,
esters,   glycol  ethers,  ketones,   nitroparaffins,   and  water.
Solvent is evaporated in  low-temperature dryers [(23° to 120°C),
(70°  to 250°F)],  usually indirectly heated  by  steam or hot  air.
Steam drum dryers can be used.

3.13.1.2  Flexography—
      In flexographic printing, as in letterpress, the  image  areas
are  raised above the nonimage  surface.  The distinguishing  fea-
ture  is  that the image  carrier is made of rubber and  other  elas-
tomeric  materials.   Flexographic presses  are  usually rotary web
presses,  i.e., roll-fed.   The  flexographic printing market in-
cludes  flexible packaging  and  laminates,  multiwall  bags,   milk
cartons,   folding   cartons,   corrugated paper,  paper cups and
plates,  labels, tags, tapes,  envelopes,  and  gift wrap.
      Flexography uses  very fluid  inks with organic solvent con-
tent ranging  from  50  to  85 percent.   The inks  dry by  solvent
evaporation,  usually in high-velocity air dryers at  air  temper-
atures  below  120°C  (250°F).  Solvents that do not  damage rubber
must be used.   Typical solvents  are  alcohols, aliphatic hydro-
carbons, glycols, esters, glycol ethers, glycol ether  esters, and
water.

3.13.2   Emissions
      The major air pollutants from printing operations  are  those
 from evaporation of organic solvents used  in ink dilution and
                               3.13-2

-------
cleanup.  Minor  quantities  of dust  (from  workroom  activity)  and
diffuse fumes (from dryers)  may be also vented to the atmosphere.
     The primary sources of organic emissions are from the drying
of inks.  Types  and  quantities of emissions depend mainly on the
inks and  solvents  used, the  type and size  of printing process,
the requirements of  the job being printed (ink consumption),  and
the degree  of  emission reduction achieved  by  control  devices.
     Printing inks  are  composed  of  materials  similar  to  those
used in surface coatings:  pigments,  vehicles, and solvents.  The
specifications  for  an ink vary widely according to the applica-
    2
tion   and influence  the quantity  and  composition  of the  VOC
emissions.
                                                                3
     The  following materials  used as solvents in printing inks,
usually  in  combinations, determine  the composition of the  VOC
emissions:

          toluene                  methanol
          xylene                   ethanol
          heptane                  propanol
          hexane                   isopropanol
          isooctane                butanol
          mineral spirits          ethylene glycol
          Stoddard solvent         glycol ether esters
          naphthas                 glycol esters
          heavy naphthas           acetone
          methyl ethyl ketone      ethyl acetate
          methyl isobutyl ketone   isopropyl acetate
          ethyl acetate            normal propyl acetate
     Inks  incorporating these solvents  are  considered conven-
tional  inks.  Ultraviolet and electron-beam-curable inks,  water-
borne  inks,  and heat-reactive inks  are discussed  under control
measures.   In  gravure  and  flexography  more  solvent  is  added
directly  to  the ink troughs,  and  emissions  are not directly re-
lated to  the ink composition.

3.13.3  Control Measures
     Emissions  can  be reduced by add-on control devices  and by
the use of low-solvent inks  as summarized in Table 3.13-2.  The
                              3.13-3

-------
applicability of  each  control method to  each  printing technique
is discussed below.

3.13.3.1  Add-On Control Equipment—
     Fume  incinerators  (direct  flame  and catalytic)  and carbon
adsorbers  are  the  only devices  with  proven high  efficiency in
controlling hydrocarbon vapors  from rotogravure  and flexographic
printing operations.
     Incineration--Incineration  is  a technically feasible method
of  controlling  emissions  from  all  printing  operations.   Both
direct-flame and  catalytic  incinerators  are  potential methods of
controlling emissions from flexography and package gravure print-
ing.   In  a  direct-flame  incinerator  a  temperature of  600° to
680°C (1100° to 1250°F) and a residence time of 0.3 to 0.5 second
are generally  sufficient  to achieve 90 percent oxidation of most
organic  vapors  passing  through the device.  Temperatures of  760°
to  820°C (1400° to  1500°F) may  be  necessary to oxidize aromatics
such as  toluene and xylene.
     Heat  recovery can substantially  reduce the  cost of incin-
eration.   Heat recovery  equipment  that  uses the hot  incinerator
gases  to preheat dryer exhaust gases prior to  incineration re-
duces  incinerator  fuel requirements.   In some cases heat equip-
                                                                 4
ment can be  used  to supply  the heated air  required  at  the dryer.
     A  third type  of  incinerator,  the  pebble-bed,  has been  sug-
gested   as  an  applicable  control  for  graphic  arts processes.
Pebble-bed incinerators combine  the functions  of  a  heat exchanger
and a  combustion  device,  as  shown in Figure  3.13-1.   The  solvent-
laden  exhaust from the dryers and  floor  sweeps  enters one of the
pebble  beds,  which  has  been heated by combustion  chamber  exhaust.
Oxidation  of the vapors  starts  in the  preheated bed  and is  com-
pleted in  the combustion  chamber.   The exhaust gases exit through
 a second  pebble  bed,  transferring heat to  the pebbles.  Pebble-
bed systems  are  designed to  achieve  a  heat  recovery efficiency
                               3.13-4

-------
                                                                                 TO ATMOSPHERE
I
en
      AND
DAMPE
DAMPER ':
rfH?

DRYERS
npsuFTPs

ft,
.K _^^_
v:W

•M^H
CERAMIC
.'•; BED VI
\ "••';.••
•/•. • •


•
X^~^
1,400°F
ct,
~-
GAS OR OIL


(\MPER
> t
*
; DAMPER
1 A A
T W






TO D
200-
[
RYERS
600° F
      Figure 3.13-1.  Ceramic-bed regenerative-type incineration and  heat recovery system,

-------
of 85 percent.   Use of pebble-bed incinerators appears to be best
suited to continuous printing operations  having rich VOC exhaust
streams.
     Carbon adsorption--Recovery  of  solvents  by  use of  carbon
adsorption  systems  has been  successful  at  several  large  roto-
gravure plants.  Rotogravure presses used a  single, water-immis-
cible  solvent  (toluene)  or else a mixture which  is recovered in
approximately  the  proportions  used in the ink.   Three such sys-
tems  are  reported  to  recover 23 to  26 kl per  day (6000 to 7000
gal/day)  of solvent  at  an average  efficiency of  90  percent or
        c a *"7
greater.       These recovery systems  were installed for regula-
tory and economic reasons.  Solvent is evaporated from the web in
indirect  steam-heated  ovens, which preclude  any solvent decompo-
sition.   Regeneration of  the  carbon  is  accomplished by  use of
steam, followed by condensation  and decantation.
     Some rotogravure  operations, such as printing and coating of
packaging  materials,   use  inks  and  coatings  containing complex
solvent  mixtures.   Many  of the  solvents are  water-soluble.   A
folding  carton  operation  for  example,   requires  at  least five
solvents,  some of which  are  soluble in  water.   Also, when fre-
quent  product changes call for different solvent combinations,
solvent  recovery is  virtually impossible.  Reformulation of inks
may  offer a possible method of avoiding these difficulties.
     A new  type of  carbon adsorption  system,  a  fluidized bed
system developed in Japan, reportedly offers  a  method of avoiding
the  problems associated with the use of water-soluble solvents in
conventional  fixed-bed systems.   This new system  utilizes  nitro-
gen  gas  as the  desorbent.  Because the  solvent is  condensed in
indirect heat exchangers  and the nitrogen is recycled, there is
no  mixing  with  water  at desorption.   The  advantages  claimed
include  better thermal efficiency,  lower power  consumption, and
regeneration  at higher temperatures to remove  high boiling mate-
rials.  The disadvantages include  possible  higher  capital  cost
and  a requirement  that  relatively constant  air volume be  main-
tained.

                              3.13-6

-------
     Where  add-on  devices  are  used,  the  overall VOC  reduction
(control) efficiency  is  dependent upon  the  extent to  which the
capture system collects,  contains, and delivers the VOC vapors to
the control device  inlet.   Capture  system efficiencies vary with
the complexity  and configuration of the printing operation and
with  the difficulty  encountered in accessing emission  points.
The efficiency of capture systems at existing plants  are reported
to be  75 to 85  percent  for publication  rotogravure,  75  percent
for packaging rotogravure,  and  70 percent for flexography.  Such
capture  systems,  coupled  with  the  90  percent  or  greater  VOC
reduction  achievable  with  incineration  or  carbon  adsorption
systems,  provide an  overall VOC reduction  of  75  percent  for
publication rotogravure,  65 percent for packaging  rotogravure,
and 60 percent for flexography.

3.13.3.2  Use of Low-Solvent Inks—
     Low-solvent  inks are of  three  types:   waterborne,  high-
solids,  and radiation  curable.  Only waterborne  inks  are  now
widely  used  for packaging gravure  and flexographic  printing.
Waterborne  inks  are  not  completely  solvent-free  because  the
volatile portion contains  up to 35  percent water-soluble organic
compounds.   Although  waterborne inks  are  used  extensively  in
printing  corrugated paperboard  for  containers,   multiwall  bags,
and other packaging materials  made  of paper  and  paper products,
their use is somewhat limited because they absorb into thin paper
stocks and seriously weaken the paper.
     Flexographic and rotogravure packaging  printing  operations
with less demanding quality requirements  can use waterborne inks
to achieve  emission levels comparable to  those  attained  by the
application of add-on control devices.   A waterborne  ink consist-
ing of 75  volume  percent  water and  25  volume  percent  organic
solvent  in  the  solvent  portion of   the  ink  is considered  to  be
equivalent  in  control effectiveness to  either carbon  adsorption
                9
or incineration.
                              3.13-7

-------
3.13.4  Emission Limits
     Review  of  existing  State  Implementation  Plans  (SIP)  showed  no VOC
emission limitations specific to the graphic arts industry.  Rather,  states have
adopted  general  hydrocarbon  emission regulations  patterned  after  the Los
Angeles  Rule 66,10 which  has  been modified pursuant to an EPA ruling that
regards  all hydrocarbons as photochemically  reactive.
     The South Coast Air  Quality Management District Rule  442 (formerly Los
Angeles  Rule 66) places limitations on emissions from  equipment using organic
solvents  or   organic  materials  containing  organic   solvents.    For  organic
materials that  come into  flame  contact  or  are  baked,  heat-cured,  or heat-
polymerized,  the  limitation is 1.4 kg  (3.1 Ib) per hour, not to exceed 6.5  kg
(14.3  Ib)  per  day.  For organic materials emitted  from the  use  of photo-
chemically reactive  solvents,  the limitation is 3.6 kg (7.9 Ib) per hour, not to
exceed 18 kg (39.6 Ib)  per day if not  in flame contact, baked,  heat-cured,  or
heat-polymerized.   The  above mass  emission  rate  limitations  do not  apply  if
the  emissions are reduced  by at  least 85 percent.
     At this  time  no  NSPS  have  been  promulgated  for  the  graphic arts
industry.    It  is  anticipated  that   such  standards  will  be  proposed  for
publication  rotogravure in late 1979,  after  the  completion of emission surveys
and control  technology  studies.

3.13.5  Determination of LAER
      The  recommended  limitations  are  based  on  SIP's  and  on  performance
information  available  in early 1979.   It is  anticipated that several  additional
 SIP regulations covering these sources will be promulgated  and/or modified in
 1979 and 1980 and  that appreciable new performance data will become available
 in the near  term.  Conceivably, some  SIP  regulations may  be more stringent
 than  the LAER  suggested  herein.   Furthermore,  performance  testing  may
 show  that  more  stringent limits than  those suggested are  feasible  or  it may
 show  that the  suggested limits are appropriate or that they are not achievable
 for some specific subcategories.  In any  case, the basis for determining LAER
 for many source categories  is  expected to  change  frequently.  Since LAER is
                               3.13-8

-------
near  the vanguard of control technology, a more detailed analysis is  partic-
ularly necessary  when addressing modified or reconstructed  facilities  subject
to  the  provisions  of Section 173 of the  Clean Air  Act.   Emission limitations
reasonable  for  new  sources may in some instances be  economically or tech-
nically  unreasonable when applied to modified  or  reconstructed  sources of the
same  type.
     The  most  stringent  regulations   in   State  Implementation  Plans  are
patterned   after  Los  Angeles   Rule 66.   Emission  reductions achieved  in
practice  exceed  those  required  by  the  most  stringent  SIP regulations.
Incineration,  carbon adsorption,  and the use  of  low-solvent-content inks are
proven techniques  for  reducing  VOC  emissions  from  rotogravure  and flexo-
graphic printing.
     The use of incineration may be constrained where natural gas  is in short
supply.   Incorporating heat  recovery  to the  extent possible will  reduce the
total  fuel  requirements.   Generally, incineration  is technically feasible when
carbon  adsorption  with solvent  recovery is  not possible.   The  use of  carbon
adsorption may  increase energy  requirements,  but it also may  enable recycling
of  solvents.    Waterborne  ink   can generally  be  used  for  packaging  and
specialty  rorogravure  and   flexographic  printing  operations with relatively
sturdy  substrate.   The feasibility of ink substitution must  be  determined  on a
case-by-case  basis, depending  on  product  specifications and the type  of
process  used.   Low-solvent inks   may  be   developed  and   used  to yield
reductions  equal  or close  to those achieved  with  control system  hardware.
     In actual operation,  direct-flame and catalytic  incinerators have  demon-
strated emission reductions  of 90 to 95+ percent  on certain types  of printing
operations.   Carbon adsorption   units can reduce emissions by  90  percent  or
more  on  rotogravure  processes  using solvents that are insoluble  in  water.
These  values pertain to the  efficiency of  the device  only.
     This  guideline  relies substantially  on  information provided in  the  EPA
Control   Techniques   Guideline   (CTG),  Volume   VIII:    Graphic   Arts--
Rotogravure and Flexography (Reference  2).   The overall  emission  reductions
reported as  achievable  in the CTG  pertain  to retrofit application of  control
strategies.   Since  LAER  applies only  to  new (modified)  facilities,  greater
                             3.13-9

-------
emission reduction can be expected.  Unlike retrofit systems, new installations
offer  an  opportunity to  maximize plant layout, process methods,  and capture-
control system designs to obtain the most effective VOC control.   Considering
these factors,  it  appears the  following overall  reductions  of  VOC  emissions
from  solvent-borne  inks  are attainable with effective capture-control systems
and  are  presented  as  suggested  LAER values:   publication rotogravure, 80
percent;  packaging  rotogravure,  70  percent; and flexography,  65 percent.
     Comparable  emission  reductions  achieved by use of water-borne and/or
low-solvent (high-solids)  inks  rather  than solvent-borne inks are  considered
as meeting the LAER values.
                               3.13-10

-------
                           REFERENCES


1.   Gadomski,  R.R.,  et al.   Evaluations of Emissions and Control
     Technologies in the Graphic Arts Industries,  Phase I.
     Graphic Arts Technical Institute.   August 1970.

2.   Control of Volatile Organic Emissions from Existing Sta-
     tionary Sources, Volume VIII:  Graphic Arts-Rotogravure and
     Flexography, Guideline Series.  EPA 450/2-78-033,  U.S.
     Environmental Protection Agency, Office of Air Quality
     Planning and Standards.  December 1978.  pp.  2-7,  2-8.

3.   EPA 450/2-78-033.  p. 2-6.

4.   EPA 450/2-78-033.  p. 3-4.

5.   George, H.F.  Gravure Industry's Environmental Program.
     EPA-5601/1-75-005.  In:  Proceedings of Conference on Envi-
     ronmental Aspects of Chemical Use in Printing Operations,
     Office of Toxic Substances, Environmental Protection Agency.
     January 1976.  pp. 204-216.

6.   Watkins, E.G., and P. Marnell.  Solvent Recovery in a Modern
     Rotogravure Printing Plant.  EPA-5601/1-75-005.   In:  Con-
     ference on Environmental Aspects of Chemical Use in Printing
     Operations, Office of Toxic Substances, Environmental Pro-
     tection Agency.   January 1976.  pp. 344-355.

7.   Harvin, R.L.  Recovery and Reuse of Organic Ink Solvents.
     EPA-5601/1-75-005, In:  Proceedings of Conference on Envi-
     ronmental Aspects of Chemical Use in Printing Operations,
     Office of Toxic Substances, Environmental Protection Agency.
     January 1976.

8.   EPA 450/2-78-033.  p. 3-5.

9.   EPA 450/2-78-033.  pp. 3-9, 3-10.

10.  Rule 66 of the Rules and Regulations of the County of Los
     Angeles Air Pollution Control District.  Los Angeles,
     California.  January 1973.

11.  Personal communication with E. Vincent, Industrial Studies
     Branch, U.S. Environmental Protection Agency, Research
     Triangle Park, North Carolina.  March 12, 1978.
                              3.13-11

-------
3.14  MAJOR SOURCE CATEGORY:  AUTOMOBILE AND LIGHT TRUCK COATING--
      VOLATILE ORGANIC COMPOUND (VOC) EMISSIONS
3.14.1  Process Description
     Automobile and light  truck  coating is a multistep operation
carried out on a conveyor system where each vehicle body receives
several coats of  paint to  enhance appearance and protect against
adverse  weathering.    In  some   coating  operations,  hoods  and
fenders are coated  separately from the main body line and joined
to the body after coating.
     Although  no  automobile  or   light  truck  assembly line  is
"typical," features common to all are shown in Figure 3.14-1.  As
the process begins, an automobile body emerges from the body shop
and undergoes a  solvent wipe and metal treatment step (usually a
phosphate  wash  cycle)  to  improve paint adhesion  and  corrosion
resistance.  The first coat, a primer, is applied by dip or spray
methods;  then  the unit is baked.  A  second prime coat (surfacer
or guide coat) is sometimes applied by spraying.  After the prime
coats have been baked, the topcoat is applied by a combination of
manual and automatic spray devices in a spray booth.  The topcoat
is applied in one  to  three steps with  a  bake  (cure)  step after
each.  Additional coats  of different colors may also be applied.
The coated body then  goes  to the trim  shop,  where the interior
and exterior trim is applied.
     Touchup  coating  is applied  at  various  stages of  the top-
coating line  and  the  touchup sprays  are cured  in the  line oven.
After  the final  touchup,  the coating  must be  cured in  a low-
temperature oven  to  protect heat-sensitive plastics  and  rubber
parts  built into  the  vehicle.    Organic-solvent-borne  coatings
have  been used in  conjunction with  the  low-temperature  curing.

3.14.2  Emission of Pollutants
     Emissions of volatile organic compounds (VOC) occur at three
significant facilities  in  a  vehicle  finishing operation:   1) the
                              3.14-1

-------
 FROM BODY SHOP
                                   METAL
                                PRETREATMENT
                               DRY-OFF OVEN
     PRIME
   APPLICATION
     AREA
PRIME CURE OVEN
            FIRST TOPCOAT
           APPLICATION AREA
  FIRST TOPCOAT
   CURE OVEN
 SECOND TOPCOAT
APPLICATION AREA
    (IF ANY)
            SECOND TOPCOAT
              CURE OVEN
               (IF ANY)
                                          COATED PARTS FROM
                                             OTHER LINES
   THIRD TOPCOAT
 APPLICATION AREA
     (fF ANY)
 THIRD TOPCOAT
   CURE OVEN
    (IF ANY)
1
 TRIM APPLIED
 (SEATS. RUGS,
DASH. TIRES. ETC.)
  REPAIR TOPCOAT
 APPLICATION AREA
  REPAIR TOPCOAT
      OVEN
(LOW TEMPERATURE)
                                                                FINISHED
                                                                PRODUCTS
Figure  3.14-1.   General flow diagram for  automotive and light
          truck assembly plants.   Main bodies may be on
             separate lines  from  hoods and  fenders.
                                  3.14-2

-------
prime  coating  line, 2) the topcoating line,  and 3) the  final repair  area.   In
each of these  facilities,  emissions occur  in  the  application area,  the flash-off
area, and the  cure oven.
     The  prime coat serves  the  dual function  of  protecting  the  surface  from
corrosion  and  providing  for  good adhesion  of the topcoat.  Organic-solvent-
borne  primer  is  usually applied by a  combination  of  manual and  automatic
spray  methods,  with or  without the use  of  electrostatic  techniques.   When
manual  methods  are used,   health  regulations  require  that  solvent concen-
trations be kept  low.   Waterborne  primers may be  applied by spray  or dip
methods.   When organic-solvent-borne  spray is  used, 85 to 90 percent of the
solvent evaporates  in  the booth and  flash-off area; the  remaining 10 to 15
percent evaporates in the  oven.   Typical  solvent  emissions  from  the  prime
coating operation  are 193  kg/h (425 Ib/h) with a spray applied two-coat prime
containing 32 percent solids  enamel, and 18 kg/h  (40 Ib/h) with EDP coating.
The  EDP  is   followed  by a prime surfacer  coat  that  has typical  solvent
emissions  ranging  from  (120  to  300 Ib/h)   depending on  the type of coating
used.
     The  topcoat  is applied in one  to three steps, each followed by a  curing
oven.   The  topcoat colors   used on-line vary  in accordance with  consumer
requests.   Metallic and  "tutone"  finishes are also applied  on the  topcoat line.
The  coating  is applied  by manual  and/or  automatic  spraying in an enclosed
spray  booth.   Solvent concentrations in the  manual  booth must comply  with
health  regulations.   Because  of  the length of time that the  auto body is in
the spray  booth,  85 to 90 percent  of the solvent  evaporates in the  booth  and
its  flash-off  area.   Typical  solvent emissions  from the topcoat  line are  386
kg/h  (850 Ib/h)  with a 32 percent  solids enamel,  1.3 Mg/h (3000  Ib/h)  with a
12  percent solids  lacquer, and 77 kg/h  (170 Ib/h) with waterborne  coatings.
     Areas damaged  during  trim  application or imperfections not detected in
the main  paint shop are  repainted in a final repair step.  Because the vehicle
now  contains  heat-sensitive materials,  coatings  used for repair are  generally
limited to solvent-borne materials that can be  dried in low-temperature ovens.
Production in the repair  area is intermittent.  Typical emissions from the  final
repair  area are 13.3 kg/h (29.3 Ib/h) with solvent-borne coatings.
                             3.14-3

-------
     Uncontrolled  organic  emissions  from coating  of  vehicles  with organic-
solvent-borne coatings  can range  from less  than 272 kg/h  (600 Ib/h) to more
than 1.8 Mg/h (4000 Ib/h) over an  entire assembly line.  This wide range is
caused  by  variations in  the  surface area coated  on  different vehicles,  the
number of vehicles  coated per hour,  the  coating application  method,  and.  the
solvent content of the coatings.   A plant  may operate  more than one assembly
line.
     Sources  of organic  emissions from  a  vehicle assembly  plant  that are  not
considered  here  include  the  application  of   adhesives  and  soundproofing
materials.   These account for about 10 to  30 percent of total organic emissions
from the plant.   Other  fugitive  emissions  occur   during  the initial  solvent
wipe, plant cleanup, and solvent storage.

3.14.3  Control Measures
3.14.3.1  Prime Coating Line--
     The  use  of waterborne  spray primers   and/or  electrocoating  systems
provides  substantial reductions  in  VOC emissions and is  the  most common
control measure  now used in the industry.   The  greatest emission reduction
occurs when primer is  applied  by electrophoretic (electrodeposited) dip  of the
total  body.   Only  waterborne coatings can   be   applied  by  this  process.
 Typical solvent contents of  the coatings  (not including water) range from 0.8
 to  2.0  Ib  solvent/gal  coating.  The  object to be  coated  is immersed in a
 waterborne  coating  and an electric potential difference is  induced between the
 object  and  the  coating bath.   As  the  object emerges  from  the bath,  its
 coating is  90  percent  solids  (by volume), 9  percent water,  and 1 percent
 organic cosolvent.  Excess  coating is  returned to the bath by washing  the
 object with makeup  and ultra-filtered  water.   Because of  the extremely
 low-solvent  usage  [about  3  kg/h  (7 Ib/h)], the  exhaust from this  oven
 requires no further emission control unless it  presents an odor problem.  The
 electrophoretic  dip process  is used at over^ 40   percent of U.S.  assembly
 plants and  is very  widely used in Europe.  '  '
      The  percentage  reduction  achieved   by  a  change   to  electrophoretic
 coatings depends on the original system.  For example, if the  change  is  from
                               3.14-4

-------
a  32-percent solids  primer  [about  2.4  kg  (5.3 Ib) of  organic  solvent  per
gallon  of coating] to  electrophoresis [about 0.45 kg  (1.0 Ib) of organic solvent
per gallon  of coating], the reduction is 80  percent.   Reduction in  emissions of
95 percent by use of this priming method have been reported.
     The  major  limitations of electrophoretic  dip coating are that  it can  be
used only  directly over  metal or  other  conductive surfaces and only one coat
can be applied.   Electrical  requirements increase by  about 1400  kW with this
method;  this  represents  a 12 percent increase over the 12,000 kW required for
applying organic-solvent-borne  primers.   Additionally, the electrophoretic dip
coatings may contain  amines that  are  driven  off during the  curing step;
incineration  of  the  oven  exhaust  is then required  to eliminate  the  visible
emissions and malodors associated  with amines.
     Spraying of waterborne primers is possible, but such operations are not
readily automated.   Some plants  do spray waterborne primers,  and  this is a
viable  option for many plants,  especially when more  than one primer  coat is
required.
     The use of organic-solvent-borne  coatings with  relatively  high  solids
content  inherently  reduces  VOC  emissions   because  of  the  lower-solvent
content.   The achievable reduction depends  on a  comparison of  the higher-
solids  coating with the coating  that would  otherwise be  used.   For example, a
coating with  50  percent solids  achieves  an 86 percent  reduction  of  emissions
compared with a  coating with a lacquer  with  12  percent solids; the reduction
is only 53  percent, however, with respect  to use of an enamel with 32 percent
solids.   Obviously,  even further  reductions  can be  achieved if the transfer
efficiency  is  improved or an add-on control device is  also installed.  Topcoats
containing 37-40  volume  percent solids  and nearly 50 percent solids have been
used by Volkswagen  and American Motors,  respectively.
     Add-on  control  measures available  for use  on  prime coating lines  include
activated  carbon adsorption  and  incineration.   Pilot  studies on  the  use of
carbon adsorption in primer spray booths and  flash-off areas have  indicated
                                                      fi  7
potential  for reductions of greater  than 85 percent.  '    Presently,  no full-
scale  carbon adsoption systems are in operation on  any automobile or truck
coating line.  Carbon  adsorption  is technically feasible, however,  and General
                              3.14-5

-------
Motors  has acknowledged  that  activated  carbon can  be used effectively on
spray booths and ovens to reduce solvent  emissions by 90 to 95 percent if the
                                                       fi 7
system  is properly engineered and regularly maintained.  '
     Both  catalytic and thermal incinerators  could be  used  to  control  VOC
emissions from  prime  and  topcoating  spray booths  and  ovens.   Incinerators
operated at high temperatures can almost completely destroy organic vapors.
Reductions  in VOC emissions  of  90 to 98 percent have been reported with the
                                                         159
use  of  thermal  incinerators on primer and topcoat  ovens. '  '    The fuel  con-
sumption  by incinerators of oven exhaust need not be  excessive if the  oven
operates at a  relatively high  proportion  of the lower  explosive limit (LEL).
Heat recovery  systems become  attractive with  increased VOC concentration.
With  spray booth  and flash-off  area  exhausts,  however, the  high flow rates
and   very  low  organic  vapor  concentrations  require  that an  incinerator
consumes  huge  amounts of  fuel.  Moreover,  the  opportunity for more  than
primary energy recovery  is  restricted by the. limited need for  the  large
amount of energy  available.   For these  reasons, although technically feasible,
incineration of  spray  booth and  flash-off  area exhaust is not  practiced at any
plant.

3.14.3.2  Topcoating Line--
      Many  of  the  control  measures applicable to the  prime  coating line can
also  be used on the  topcoating  line.  Reductions in organic solvent emissions
of up to 92 percent  from  topcoat spray booths  and ovens are possible by use
of  waterborne  topcoats.    As before,  the exact  reduction  depends on the
original coating  and  the  substitution.   Waterborne  coatings  are currently
being used at  two General Motors automobile  assembly  plants in California on
a full-scale basis.  These plants have  reported 88 percent reductions in VOC
emissions  as a result of the coating substitution.5  As with waterborne primer
systems, however, use of  waterborne  topcoats increases  electrical usage  (in
this  case,  by 42  percent).   Difficulties in  precipitation and dewatering of the
collected overspray  increase  the  solid waste  disposal problem.   Waterborne
coatings  are  sensitive to  humidity,  and  it has  been  reported that suitable
waterborne coatings do not exist for van  interiors.
                              3.14-6

-------
     The   use  of  higher-solids  coatings,  activated  carbon  adsorption,  or
thermal and catalytic incineration  on  the  topcoat line  can achieve  the same
reductions  as  presented  for  prime  coatings  and are  subject  to  the same
limitations  described  earlier.
     Recently developed powder  coatings  may have possible  application in  the
automobile and  light-duty  truck  coating  industry.  To  date, their use in this
industry has  been limited  by the need to change  colors often  and the lack of
availability  in  metallic  colors.   Preliminary  studies  indicate  that  powder
coating systems reduce  energy requirements, have acceptable durability, offer
95  to  99  percent  utilization,  and  essentially  eliminate  all organic  solvent
emissions.5'    Within  the last  year,  the major auto  producers have resumed
investigation of powder  coatings.

3.14.3.3  Final Repair--
     Control of emissions from the final repair  spray  booth  and oven  has  not
been practiced  because the intermittent operations make add-on control equip-
ment less cost-effective than in other areas.   These  emissions can be collected
and sent  to  an  activated  carbon adsorption  unit or  incinerator at  another
location  within  the  plant.   This system,  however,  would not provide cost-
effective  control.   Considerable reductions in  emissions  can be  accomplished
by the use of  higher-solids repair  coatings,  but this measure  has not been
practiced at any plant.

3.14.4  Emission Limits
     The initial  criterion for defining  lowest  achievable emission  rate  (LAER)
for  a  surface coating industry  is the  degree of  emission  control required by
the  most stringent  regulation adopted  and successfully enforced by  a  state or
local air pollution control agency.
     Most  organic solvent  emission regulations are patterned  after what is now
Rule 442 of  the  South  Coast (California)  Air  Quality Management  District.
Review of  the  regulations  applicable  in  the  16 states that  contain about 85
percent  of all surface  coating industries  showed  that these are essentially the
same as  Rule 442.    Indiana has the most  stringent regulation, which limits
                              3.14-7

-------
organic solvent emissions  to  1.4 kg/h (3 Ib/h)  or  6.8  kg/h (15  Ib/day)  unless
such  emissions are reduced by  at least 85  percent,  regardless  of the reactiv-
ity  or temperature  of  the solvent.   Organic  solvents that  have been  deter-
mined  to be  photochemically unreactive or  that  contain  less  than specified
percentages of photochemically reactive organic materials are  exempt from this
regulation.   Recent research,  however,  has indicated that substituting low-
reactivity  solvents  for  higher-reactivity solvents may improve photochemical
                                                                           13
oxidant  air  quality in  one  city  while worsening it  in  downwind regions.
Accordingly,  EPA  has  adopted a  policy  emphasizing the need for "positive
reduction  techniques"  rather than  substitution  of compounds.    Thus,  the
"low reactivity" solvents will  no longer be exempt.
     California Air Resources Board has recently adopted a model rule for the
control  of  VOC  emissions  from  light- and medium-duty  vehicle assembly
plants.  This  rule  limits  VOC  emissions  from  the  prime  or topcoat lines to
0.275  kg/liter  (2.29 Ib/gal) of coating  as applied,  excluding water, unless the
emissions  are  reduced  by 90  percent through  treatment  of  the  exhaust.
Other recommended emission  limitations represent rates  achievable by  appli-
cation of reasonably available control  technology  (RACT).   It  has been pro-
posed that these rates  also  be  adopted as new source performance standards
(NSPS)  for the industry.  The recommended VOC emission  limitations achiev-
able  by application  of RACT are:   1) 0.23 kg/liter (1.9  Ib/gal)  of coating
minus water for the prime coating line,  2) 0.34 kg/liter (2.8  Ib/gal) for the
topcoating line,  and 3) 0.58 kg/liter  (4.8 Ib/gal)  for the  final repair area.
The recommended  limitation  for prime coating  is  based on the use of  water-
borne electrophoretic dip  primer, 0.15 kg/liter (1.2 Ib/gal)  of coating  minus
water,   followed by water-borne primer  surfacer  with  0.34  kg/  liter (2.8
Ib/gal) of  coating minus water.

3.14.5   Determination of LAER
      The  recommended limitations  are based  on  SIP's  and  on  performance
information available in early 1979.   It  is  anticipated that several additional
SIP regulations covering  these  sources will be promulgated and/or modified in
1979 and 1980  and  that appreciable  new performance data will become available
                              3.14-8

-------
in the near term.   Conceivably,  some SIP  regulations may be more  stringent
than  the  LAER  suggested herein.   Furthermore,  performance  testing  may
show  that  more  stringent  limits  than those suggested  are feasible or it  may
show  that  the suggested limits are  appropriate or that they are not achievable
for some specific subcategories.   In any case, the basis for determining LAER
for many source categories  is expected  to  change frequently.  Since LAER is
near  the vanguard  of  control technology,  a  more detailed analysis is partic-
ularly necessary when  addressing  modified or  reconstructed  facilities subject
to the  provisions of Section  173 of  the Clean Air  Act.  Emission limitations
reasonable for new sources may in  some  instances be economically  or tech-
nically  unreasonable when  applied  to modified or  reconstructed sources of  the
same  type.
     LAER for automobile and light-duty truck coating  is determined  to be a
combination of limitations represented by  RACT  and  the emission  reductions
achieved-in-practice  using  current  technology and control  measures.    For
prime  coat  application,  the use of waterborne  coatings  in the  electrophoretic
dip process followed by water-borne primer surfacer represents LAER.  This
reduction is comparable to the RACT limitation of 230  g of solvent/liter  (1.9
Ib/gal) of coating  minus water and can also be achieved using higher solids
coatings in conjunction with an oven incinerator.
     For  the  topcoating line,  LAER  can  be  met by the use of waterborne
coatings with  achievable  reductions  of 88  to 92  percent,   which  is again
comparable  to  the  RACT limitation of  320  g  of  solvent/liter (2.8 Ib/gal) of
coating  minus  water.   Similar reductions  are achievable when  using other
low-solvent coatings such as powders or higher solids  coatings, although oven
exhaust incineration is  required  when  higher solids   coatings are  used  to
achieve the 88 to 92 percent reduction.
     For the final repair area, higher solids coatings would represent LAER if
this technology had been demonstrated  in  the  industry.  LAER is  considered
to be  the  recommended VOC limitation capable  of being  met by the  application
of  RACT,  i.e.,  580 g  of  solvent/liter  (4.8 Ib/gal) of  coating minus  water.
                              3.14-9

-------
                           REFERENCES
1.   Much of the material that follows is extracted almost
     directly from:   Control of Volatile Organic Emissions for
     Existing Stationary Sources.   Volume II:   Surface Coating of
     Cans, Coils, Paper, Fabric, Automobile,  and Light-duty
     Trucks.  U.S. Environmental Protection Agency, Research
     Triangle Park,  North Carolina.  Publication Number EPA
     450/2-77-008.  May 1977.

2.   Letter from V.H. Sussman, Ford Motor Company,  Dearborn,
     Michigan to J.  McCarthy regarding draft of reference 1.
     August 6, 1976.

3.   Report of trip by J.A. McCarthy, U.S. Environmental Protec-
     tion Agency, Research Triangle Park, North Carolina to
     General Motors assembly plants in South Gate and Van Nuys,
     California.  November 17, 1975.

4.   Report of trip by J.A. McCarthy, U.S. Environmental Protec-
     tion Agency, Research Triangle Park, North Carolina to
     assembly plant in Framingham, Massachusetts.  November 17,
     1975.

5.   Carson, W., et al.  Proposed Model Rule for the Control of
     Volatile Organic Compounds from Automobile Assemblyline
     Coating Operations.  Prepared for California Air Resources
     Board, Sacramento, California.  January 26, 1978.

6.   Evaluation of a Carbon Adsorption Incineration Control
     System for Auto Assembly Plants.  Radian Corporation,
     Austin, Texas.  Prepared for U.S. Environmental Protection
     Agency.  EPA Contract No. 68-02-1319, Task No. 46.  January
     1976.

7.   Letter from W.R. Johnson, General Motors Corporation,
     Warren, Michigan to Radian Corporation commenting on
     Reference 6.  March 12, 1976.

8.   Vincent, E.J., et al.  Are Afterburners Obsolete?   (Pre-
     sented at Air Pollution Control Equipment Seminar
     APCA/National Association of Corrosion Engineers.  Atlanta,
     Georgia.  January 17-19, 1978.)
                              3.14-]0

-------
9.    Taback,  H.J.,  et al.   Control of Hydrocarbon Emissions from
     Stationary Sources in the California South Coast Air Basin.
     KVB Consultants, Tustin,  California.  Prepared for
     California Air Resources  Board.   KVB Interim Report
     5804-714.   November 1976.

10.  Levinson,  S.B.  Electrocoat,  Powder Coat,  Radiate.  Which
     and Why?  Journal of Paint Technology.  44(570):42, July
     1972.

11.  Recommended Policy on Control of Volatile Organic Compounds.
     U.S. Environmental Protection Agency.  Federal Register July
     8, 1977 (42 FR 131).

12.  Data on the geographical  distribution of surface coating
     industries can be found in:  Sources Assessments:  Priori-
     tization of Air Pollution from Industrial Surface Coating
     Operations.  U.S. Environmental Protection Agency.  Publi-
     cation Number EPA 650/2-75-0192.  February 1975.

13.  A series of documents which are referenced and summarized
     in:  Control Strategy Preparation Manual for Photochemical
     Oxidant.  U.S. Environmental Protection Agency.  Publication
     Number OAQPS 1.2-047.  January 1977.
                              3.14-11

-------
3.15  MAJOR SOURCE CATEGORY:   METAL FURNITURE AND LARGE
      APPLIANCE COATING - VOLATILE  ORGANIC  COMPOUND  (VOC)
      EMISSIONS
3.15.1  Process Description
     Metal furniture is  manufactured for both  indoor  and outdoor use by two
general  consumer  categories:   business/institutional  and  residential.   Metal
furniture  products  include  tables,   chairs,   wastebaskets,   beds,  desks,
lockers,   benches,   shelving,  file cabinets, lamps,  room dividers, and  many
other  similar items.   Size  of the  plants  varies  with the  type  of furniture
manufactured,  the number of  manufacturing and coating lines,  and the amount
of assembly required.
     Large  appliance products include  doors,  cases, lids,  panels and interior
support  parts  of  residential and commercial washers,  dryers,  ranges, refrig-
erators,  freezers,  water  heaters,  dishwashers, trash compactors, air condi-
tioners and similar  products.   A typical large  appliance  plant is much bigger
than  a typical  metal  furniture  plant,  but may manufacture only one or  two
different  types  of appliances and contain as few as  two coating lines.   Most of
the  coatings  used  in  these  industries  are  enamels,  although  quick-drying
lacquers  are sometimes  applied to repair scratches  and  nicks  that  occur
during assembly.   Some  coatings are metallic.   The coatings applied to  metal
furniture and  large  appliances  must  protect  the  metal  from the  corrosive
action of  agents such as  heat,  water,  detergents,  and  the outdoor elements.
They must  have  good  adhesion properties to  prevent peeling or  chipping,
must  be  durable,  and  must meet customer  standards of appearance.   Coatings
may  contain mixtures of 2 to  15  different solvents.   Prime and interior single
coat materials  are  typically 25  to  36  percent  solids  by volume, and topcoat
and  exterior single coats are  30  to  40  percent solids.   The underside of many
exterior  large appliance  parts are  sprayed with gilsonite to provide  additional
moisture resistance and  sound-deadening properties.   The gilsonite is  typi-
cally  sprayed at about  25 to 30 percent solids by volume.
     A typical metal furniture line  is depicted  in Figure 3.15-1;  a typical large
appliance  line is shown in Figure 3.15-2.  Unassembled,  partly assembled, or
                              3.15-1

-------
en
 I
NJ
              FROM
           MACHINE SHOP
                     CLEANSING AND
                     PRETREATMENT
PRIME COAT FLASHOFF AREA
      AND OVEN
      (OPTIONAL)
                                                                       ELECTROSTATIC. OR
                                                                      CONVENTIONAL AIR OR
                                                                      AIRLESS SPRAY COATING
                                                                         FLOW COATING
                                                                       TOfCOATM SINGLE
                                                                       HAT AffllCATION
                                     Fiqure  3.15-1.    Coating of  metal  furniture,

-------
                                                          DIRECT TO METAL TOPCOAT
                                                                                                          FLASHOFF
                                                                                                       (OPEN OR TUNNELED)
                 FROM SHEET METAL MANUFACTURING
                    EXTERIOR PARTS
                 (CASES. LIDS AND DOORS)
U)
INTERIOR  CLEANSING AND
 PARTS   PRETREATMENT
           SECTION
                                                                      FLASHOFF
                                                                   (OPEN OR TUNNELED)
                                                       PRIME DIP
                                                                                     TO ASSEMBLV
                                   Figure 3.15-2.    Coating  of  large  appliances.

-------
totally assembled furniture or  appliance  pieces are  first  transported  on  a
conveyor  through an  alkaline  cleansing  and pretreatment process.  Sometimes
the  wash  section  is  omitted  and  the pieces are  cleaned  in  a shot-blasting
chamber  or organic  solvent cleaning operation.
     Most  metal furniture and some large  appliance pieces  are finished with a
single coat.  Other surfaces,  however,  require a prime coat because of the
topcoat formulation  or the intended use.   The  prime  coat  may be  applied  by
electrostatic or  conventional  spray,  dip,  or flowcoating  techniques   (coating
techniques are  discussed  below).  The substrate  with  the prime coat  then
goes through  a  flash-off period  to  prevent  popping  of  the film when the
coating is baked.   The  prime coat  is usually baked in  an oven  at about 160°
to 200°C  (300° to 400°F).
      The  topcoat or single exterior coat may be applied to metal furniture by
spraying,  dipping,  and flowcoating.   Large  appliances  are usually  sprayed.
A touch-up operation  or application of  highlighting tone follows.   The coated
part passes through a  flash-off zone  and into a baking  oven.    The baking
oven  usually  contains  several  temperature  zones  ranging from  160° to  230°
 (300° to  400°F).  Some  metal furniture parts are  air dried, but all  appliance
parts are heat cured.
      The  heart  of  a coating  line is  the method by  which  the coating  is
 applied.   Regardless of the type of coating, there are basically three possible
 modes  of  application:   dip coating,  flow  coating, and spraying.   Each  are
 briefly described below.
      Dip  coating is the  immersion of pieces into  a  coating bath.   The  tank
 that contains  the  coating  is continuously agitated.   As the parts move  on a
 conveyor, they  are  immersed  in the  coating and  withdrawn;  the excess
 coating  is then allowed  to drain back into the  tank.   Dip  coating, in addition
 to  being  one of the  most efficient  coating methods, also provides  the  best
 coverage  of cavities and crevices.  Consequently, it  is used largely  for  prime
 or  undercoating but may also be used  for top  coating  in  cases where the
 plant uses one or two colors.
      Flow coating  is also used  when only  one  or  two colors  are  applied.   As
 the parts are  moved  by a conveyor  through an enclosed  booth, stationary or
                               3.15-4

-------
oscillating  nozzles at various angles  emit streams  of  coating, which flow over
the part.   Excess  coating,  which drains  into  a  sink on  the  bottom  of  the
booth, is filtered and recycled.
     Spraying  is the most used  of all coating  application techniques.  There
are three natural categories  of  spray coating based upon the way the coating
is  atomized and  transported from  the delivery device to  the  metal surface.
Air spray  utilizes  a  high velocity air  gun  to  atomize  the coating into  fine
droplets and  carry the  droplets  to  the substrate.  Airless spray  is  accom-
plished by forcing  a  liquid  coating through  a  small  orifice  at  very high
pressures.   In electrostatic  spraying, the coating is atomized  by one  of  the
above  methods or  by  feeding  the coating to a disk or  bell rotating at high
rpm.   The paint particles  are  given a negative  charge  by  the atomization
device and are attracted  to the  grounded  metal parts.   Any of the  spraying
methods can be used manually  or with automatic controls.  Most spray coating
in  metal  furniture  facilities is  done manually, whereas  in large  appliance
facilities it  is done automatically.   The  important  difference among  spray
coating methods  is  transfer or application  efficiency.   Conventional  air  or
airless  sprays are  typically 30  to 50 percent  efficient.   Electrostatic spray
guns  can  be as high  as 65 percent  efficient and  high  speed  disks  and  mini
bells  can achieve 90  percent efficiency.
     Spray coating  and, frequently,  other  coating methods  are  performed in a
booth  to contain any over-spray or drippings,  to prevent dirt from contacting
the paint,  and to control the temperature  and  humidity  at the point of appl-
ication.   Air  flow  rates  through  spray booths vary  depending  on whether
they  are occupied  and  on their  size.  OSHA prescribes minimum  air velocities
to  assure  capture of  overspray and to keep  VOC  concentrations  below  the
threshold limit values (TLV).
     As has already been mentioned,  the baking oven usually contains  several
zones  at temperatures  ranging  from 160° to  230°C  (300° to 450°F).  The
exhaust air  flow  rate  depends  on the  size  of oven openings  through which
parts  enter and  exit.   Insurance underwriter  requirements typically limit the
atmosphere within industrial baking ovens to  25 percent  of  the  lower explosive
limit  (LEL); however the use of continuous monitoring equipment  changes the
requirements  and  allows  concentrations as high  as  40  percent  of the LEL.
                              3.15-5

-------
3.15.2  Emission of Pollutants
     The  significant  emitting facilities in a metal furniture or large appliance
coating plant are the  prime and  topcoating  lines.   On each line,  VOC's  are
emitted from the coating area, flash-off area, and oven.1  As shown in  Table
3.15.1, it is  estimated that 65 to 80  percent of the uncontrolled VOC emissions
are  released from  the spray booth and flash-off area in spray  applications,
and  the  remaining  20 to  35 percent from the oven.   In a  dip or flowcoat
application,  an  estimated 50  to 60 percent of  the VOC emissions come from the
coating and  flash-off area,  and the  other 40  to 50  percent from  the  oven.
Typical uncontrolled VOC  emissions from the coating of metal file cabinets are
6.8  kg/h (15 Ib/h)  from the entire plant.   Typical uncontrolled emissions from
the  coating  of  automatic  washers  are 13.2 kg/h  (29 Ib/h) from the coating
area  and oven, and 11.8 kg/h (26 Ib/h) from the sound-deadener application.

3.15.3  Control Measures
      Operators  of metal furniture and  large appliance coating  processes use
several  measures  to reduce VOC  emissions at the  point of  application or
remove  them  from  the  exhaust  stream.   Waterborne  coating  is  a feasible
control measure.   Waterborne coating can  be applied by  spray,  dip, flowcoat,
or electrodeposition (EDP).  Electrodeposition  is limited to  waterborne coat-
ings used for the primer  or single coat application.  EDP coatings are applied
from  an aqueous  bath,  which  contains   about 10  to 15  percent  solids  (by
volume)  and 2  to 4 percent organic solvents.   Applying direct  current in the
 bath  causes the   solids  to become  attached  to the  grounded  metal  piece.
 Electrodeposition  can be  performed  either  anodically  or cathodically.  The
 metal parts  emerge from the bath with a  coating containing about 90 percent
 solids  (by  volume), 1  to 2  percent  organic solvent,  and the  balance  water.
 Two automobile finishing  plants  using this application  method  report  reduc-
 tions in VOC  emissions  of 95 percent.2   Large appliance plants  using EDP
 have demonstrated  93  percent solvent  reduction.*   A more  complete descrip-
 tion of  this process is  presented in Section 3.14.3.
      Emission reductions of 60 to 90 percent may be obtained by use of water-
 borne coatings on topcoat lines  and  primer or single coat lines where the
   Data  on metal furniture and large appliance operations supplied by
   paint company.   April 14, 1978.
                               3.15-6

-------
   TABLE 3.15-1.   DISTRIBUTION OF VOC EMISSIONS FROM METAL
          FURNITURE AND LARGE APPLIANCE COATING LINES3

                             (percent)
Application
method
Electrostatic spray
Conventional air or
airless spray
Dip
Flow
Application and
flash-off area
65
80
50
60
Oven
35
20
50
40
Source:  Reference 1, Volume III.


a The base case coating is applied at 25 volume percent solids
  and 75 volume percent organic solvent, which is equivalent
  to a VOC emission factor of 0.66 kg of organic solvent
  per liter of coating  (5.5 Ib/gal) minus water.
                              3.15-7

-------
coating  is applied by spray,  dip,  or flowcoat.   Use  of these  coatings  has
produced  an 87 percent  reduction  at  some large appliance plants.*  Temper-
ature,  humidity, gun-to-metal  distance, and  flash-off time affect the appear-
ance and  other characteristics of the coating.
     Activated carbon adsorption, although technically  feasible, has not been
used  in  the  metal  furniture  or large appliance coating industries.   Carbon
adsorption is  a feasible control option for the application and  flash-off areas
because  exhaust  gases are  at ambient temperatures and contain  only small
amounts  of particulate matter that  could foul the carbon  bed.   (Some  partic-
ulate  removal would  be necessary at spray booths).  Carbon adsorption could
reduce  emissions from these  areas  by  70  to  90 percent of the vapors that are
drawn into the bed.1  Use of activated carbon adsorption on paper and fabric
coating lines  has provided up to  98  percent removal of VOC vapors drawn
into the   bed.t   For metal  furniture  or large  appliance  coating lines  the
mixture of solvents  that are  recovered cannot be used without further treat-
ment.   Collected  mixed  solvents can be  used  as   fuel  for  ovens  or other
heating processes.
     No  serious  technical  problems  are associated with the  use of either
catalytic   or   thermal incinerators  at  these  facilities.   Documentation  of
achievable reductions of 96  to  99 percent across the control device for incin-
erators used  at automobile, can, coil,  and paper coating  facilities is presented
in  Sections 3.14,  3.16, 3.17,  and 3.18,  respectively.  As discussed in  Section
3.14.3,  incineration  of  exhausts from the  spray  booth and  flash-off areas
requires   auxiliary   fuel.   Fuel input  can  be  reduced  with  heat  recovery
equipment.
      Emissions  from  topcoat and  single  coat application and  curing  can  be
 controlled by use of powder coatings.   These may be applied electrostatically
 by spraying  or dipping,  or  by dipping the preheated  metal  into a fluidized
 bed.  Electrostatic spraying  is more  widely  used than  dipping because it can
 apply thinner  films  of coating.  Powder spray coating requires a booth,  as
 *Data on metal furniture and large  appliance operations supplied
  by paint company.  April 14, 1978.
 letter from W. C.  Moses, Technical Manager, Chemical Plant
  Division,  Suttcliffe, Speakman & Company,  Limited.  March 10, 1978.
                               3.15-8

-------
does spray coating with  conventional coatings.  Ventilation requirements  are
greatly  reduced because  the booths  are not occupied.  Electrostatic dipping is
limited  to  simple  shapes.   The  fluidized  bed  dipping  method  applies  the
powder  only in thick films.  Use of powder  coating  reduces energy  require-
ments  in the  application area  and  greatly  reduces  VOC emissions.   These
coatings have  acceptable  durability and  offer  more  than 95 percent  utiliza-
     o
tion.    Reduction  in  VOC emissions may range  from 95 to 99+ percent over
conventional  systems.   Data reveal  that the  average  solvent  reduction  for
powder  coating by electrostatic spray methods is 99 percent.   Several limi-
tations  are associated with the use  of powder  coatings  in the metal furniture
coating  industry.   Color  changes require about  half an hour of downtime to
evacuate the  spray booth and purge  the former  color from the  application
device.    Color  matching  during  manufacture of  the   powder  is  difficult;
metallic  coatings are  not presently available;  powder  films have appearance
limitations;  they   do   not coat  well  within  small  recesses;   and  excessive
humidity during storage or application can affect  performance.
     Reductions  of VOC  emissions  by  choosing higher-solids   coatings over
conventional  organic-solvent-borne  coatings may  range from 50  to  82 percent,
depending  on the  solids  content of the coating that would otherwise be used.
                                                *
The  average achievable reduction is 78  percent.    Higher-solids coatings  are
applied  most  effectively  by automated electrostatic spraying, but  manual  and
conventional  spraying techniques can also be used.   Some minimal increase in
energy  may be required  to raise the pressure of the spray  gun  or heat  the
more viscous  coating  so  that it can be pumped and  atomized.   Emissions from
gilsonite application in the  large  appliance  industry can also  be controlled
through use of  higher-solids coatings;   no  data  are available on  the  solvent
reductions  that can be obtained.

3.15.4  Emission Limits
     The initial criterion  for defining LAER  for  a surface coating industry is
the  degree of  emission   control  required  by  the  most stringent regulation
*
 Data supplied by paint company.   April  14, 1978.
                              3.15-9

-------
adopted  and  successfully  enforced by  a state or  local air pollution  control
agency.
     Most organic solvent emission regulations are patterned after what is now
Rule  442 of the  South Coast  (California) Air  Quality  Management District.
Review  of the  regulations applicable  in the  16  states  that  contain  about  85
percent  of all surface coating industries  showed that they were  essentially  the
same  as  Rule 442.5  Indiana has the most stringent regulation in that it  limits
organic solvent emissions  to 1.4 kg/h or  6.8 kg/h (3 Ib/h or 15  Ib/day) unless
such  emissions  are reduced by at least 85  percent, regardless of the  reac-
tivity or temperature of the solvent.   Organic  solvents that have been deter-
mined to be photochemically  unreactive or  that  contain  less  than specified
percentages  of  photochemically reactive organic materials are exempt from this
regulation.
     In  defining  LAER  for surface coating emissions, it is not appropriate to
exempt  solvents  on  the  basis  of  their reactivity.    Recent  research  has
indicated  that   substituting   low-reactivity   solvents   for  higher-reactivity
solvents  may   improve  photochemical  oxidant  air  quality  in  one  city  while
                                   £
worsening it in  downwind regions.    Accordingly, EPA has adopted a  policy
emphasizing the  need  for "positive  reduction techniques"  rather  than  sub-
                        A
stitution of compounds.
      California  Air Resources  Board has recently adopted a model rule for the
control  of  VOC  emissions from metal  parts  and product coating operations.
This rule must be  met within  3 years  from the date of  adoption.  It limits the
VOC emissions  from baked coatings  to 275 g solvent/liter (2.3  Ib/gal) coating
minus water,   and  from  forced-air-dried  coating  to 340 g solvent/liter (2.8
Ib/gal)  of coating minus water.
      A  recent  EPA publication presents emission limits achievable  through the
application of  reasonably available control technology  (RACT).   The  recom-
mended  limitation for  metal furniture  coating  is  0.36  kg  of organic solvent/
 liter of  coating minus water (3.0 Ib/gal); for large appliance coating it is 0.34
 kg/liter  (2.8  Ib/gal).   These  limitations  are  comparable  to  an  80  percent
 reduction  in  solvent  emissions  from  each  affected  facility.   There   are
 currently  existing facilities  which meet or  exceed the  RACT  limitations  for
                               3.15-10

-------
both categories;  therefore,  it is concluded that LAER for metal furniture and
large appliance coating is a function of controls  achieved in practice rather
than controls  required  by existing regulations.

3.15.5  Determination of LAER
     The  recommended limitations  are  based  on  SIP's  and  on performance
information  available in  early 1979.   It is anticipated  that several  additional .
SIP  regulations covering these  sources  will be  promulgated and/or modified in
1979  and 1980  and that    appreciable   new   performance   data   will  become
available  in the  near  term.  Conceivably,  some SIP regulations  may be more
stringent than the LAER suggested herein.  Furthermore, performance testing
may  show that more stringent  limits that those suggested are feasible or it
may  show that  the suggested  limits  are  appropriate or that they  are  not
achievable for some  specific subcategories.  In any  case,  the  basis for deter-
mining  LAER  for many source  categories is expected to  change frequently.
Since  LAER  is  near  the  vanguard  of control technology,  a  more  detailed
analysis is  particularly necessary when addressing  modified  or reconstructed
facilities  subject to the  provisions  of Section 173  of  the  Clean  Air  Act.
Emission  limitations  reasonable  for   new sources  may  in  some  instances  be
economically  or technically  unreasonable when applied to  modified  or recon-
structed sources of the same type.
     The  lowest  achievable  emission   rate  for  metal   furniture  and  large
appliance coating is  determined  to  be a combination  of limitations represented
by  RACT  and emission reductions  achieved in practice.   For metal furniture
coating operations,  LAER can  be met by  the use of waterborne coatings,
which  can achieve reductions  of 88  to 92 percent.   This  is comparable to the
RACT limitation  of 360  g  solvent/liter  (3.0  Ib/gal) of coating  minus water.
For  large  appliance  coating,  LAER  is  also  equal to RACT that requires the
use  of  a coating containing 340 g  solvent/liter (2.8 Ib/gal)  of coating minus
water.  Similar reductions for  these  operations can  be  achieved by the use of
other low-solvent coatings (such as  powders) or higher-solids coatings.  The
application of add-on control devices  capable  of providing  equivalent reductions
is also  an acceptable way of meeting  these limitations.
                              3.15-11

-------
                           REFERENCES


1.   Much of the material on emissions is extracted almost
     directly from:   Control of Volatile Organic Emissions for
     Existing Stationary Sources.   Volume III:   Surface Coating
     of Metal Furniture, and Volume V:  Surface Coating of Large
     Appliances.  EPA 450/2-77-032 and 034,  U.S. Environmental
     Protection Agency,  Research Triangle Park, North Carolina.
     December 1977.

2.   Taback, H. J.,  et al.  Control of Hydrocarbon Emissions from
     Stationary Sources in the California South Coast Air Basin.
     KVB Consultants, Tustin, California.  Prepared for Cali-
     fornia Air Resources Board.  KVB Interim Report 5804-714.
     November 1976.

3.   Levinson, S. B.  Electrocoat, Powder Coat, Radiate.  Which
     and Why?  Journal of Paint Technology 44(570):42, July 1972.

4.   Recommended Policy on Control of Volatile Organic Compounds.
     U.S. Environmental Protection Agency.  (42 FR 131) Federal
     Register, July 8, 1977.

5.   Data on the geographical distribution of surface coating
     industries can be  found in:  Sources Assessments:  Prioriti-
     zation of Air Pollution from Industrial Surface Coating
     Operations.  EPA 650/2-75-0192, U.S. Environmental Protec-
     tion Agency.  February 1975.

6.   A  series of documents referenced and summarized in:  Control
     Strategy Preparation Manual  for Photochemical Oxidant.
     OAQPS  1.2-047, U.S. Environmental Protection Agency.
     January 1977.

7.   Fugita, E., et  al.  Consideration of a Proposed Model Rule
     for the Control  of Volatile  Organic Compound Emissions  from
     the Surface Coating  of Manufactured Metal  Parts and  Prod-
     ucts.  Presented to  California Air Resources Board,
     Sacramento, California.   September 27, 1978.
                               3.15-12

-------
 3.16  MAJOR  SOURCE CATEGORY:  CAN COATING—VOLATILE  ORGANIC
      COMPOUND  (VOC) EMISSIONS
 3.16.1   Process Description
      Cans  are manufactured by  two  different processes.1  The three-piece
 can is  made from a rectangular sheet and two  circular ends.   The metal sheet
 is  rolled into  a cylinder and soldered, welded, or cemented at the seam.  One
 end is  attached  during manufacturing and the other  during  packaging of the
 product.  The two-piece  can is drawn and wall-ironed from a shallow cup and
 requires only  one  end, which is attached after the can is filled  with product.
      Two-piece can manufacturing is  a high-speed process that  combines  the
 fabricating  and  coating  operations.   These  cans  are most commonly  used by
 the beverage  industry.    Figure  3.16-1 presents  a  process  flow diagram  of
 fabricating  and coating two-piece cans.
     Metal for two-piece cans is received in coil form and is  fed continuously
 into a  press  (cupper) that  stamps and forms a  shallow cup.    The  cups go
 through  an extrusion  process that,  in a  lubricating  solution, draws  and
 wall-irons them  into   cans  in a  lubricating solution  and  trims the  uneven
 edges.   The formed cans are then cleansed to remove  the lubricating solution,
 rinsed with hot water, and dried.
     The exteriors of  the  cans  are  sometimes roller coated with  a base coat.
 The base coat is  transferred from a feed tray, over  a series of rollers, and
 onto the can,  which rotates  on a mandrel.   The  coating  is cured or baked at
 177°  to  204°C (350°  to  400°F)  in single or multipass continuous ovens at a
 rate of  500  to  2000 cans per minute.
     Designs and lettering  are transferred to the  cans from a  printing blanket
 on  a  rotary printer.   A  protective varnish is sometimes roll coated  directly
 over  the inks on  the same  line.  The decorative coating  and varnish are
 cured or baked  in  single  or multipass continuous,  high-production  ovens at
 163° to  204°C (325°  to 400°F).
     After printing, the  cans are  necked, flanged,  and tested.   The cans are
spray coated on  the  interior and  spray and/or roll coated  on the exterior of
the bottom  end.   The coating  is  usually  cured or  baked in  a continuous,
                             3.16-1

-------
                                                                 CANS

8)
COIL
^
COPPER


n
WALL
IRONER



WASHER
                                                                               OVEN
                                                                    iSE COAT TRAY

                                                             EXTERIOR BASE COATER,
I
to
                      $%
OVEN    INTERIOR BODY SPRAY
      AND EXTERIOR END SPRAY
       AND/OR ROLL COATER
                                                       LEAK
                                                       TESTER
                                                                 PRINTER AND OVER
                                                                                    COLOR 4
                                                                                      COLOR 3

                                                                                      COLOR 2
                                                                                    COLOR 1


VARNISH

^•~^
COATER ft y-N VA
J




RNISH
FRAY
NECKER AND
 FLANGER
                                                                                  OVEN
                 Figure 3.16-1.   Two-piece can fabricating and coating operation

-------
singlepass  oven  at  temperatures of  107°  to  204°C  (225°  to 400°F).  Coated
cans are stacked on pallets for shipment to users.
     The   three-piece  can   manufacturing  process   consists  of  two  major
operations:   sheet coating and  can fabricating.  The sheet coating operation
consists  of base  coating  and printing, which  includes  the  overvarnish  coat.
     The  sheets are roll coated on one side only  by transfer of the coating
through  a  series  of rollers from the  coating tray onto the sheets.  Sheets are
then picked  up by  preheated wickets  and transported through  a continuous,
multizone oven at rates of 50 to 150 sheets per minute, depending on the type
of coating.   The coating  is cured at temperatures up to  218°C (425°F); sheets
are  air  cooled  in the  last  zone of the  oven.  Oven  exhaust rates usually
range  between 0.94 and  6.6 Nm3/s  (2000 and 14,000 scfm).
     The sheet printing  or  lithograph  operation  usually involves applying one
or two colors of ink on the exterior base  coat,  or directly on the metal.   Inks
are  applied by  a series  of rollers  that transfer  the design first  to a blanket
cylinder, then  onto the  metal  sheet, as shown in  Figure 3.16-2.  Varnish is
applied  directly over the  wet  inks  by a  direct-roll  coater.  Inks  and  over-
varnish  are  cured in a  wicket  oven similar to,  but  usually smaller than, the
base coat  oven; exhaust rates  are  0.7 to 3.8  Nm3/s (1500  to  8000  scfm).  If
the  design requires  more than  two colors, the first  set of inks is dried in an
oven before additional color  inks are applied.   After all inks  are applied, the
sheets are overvarnished and then cured  in another  oven.
     The  three-piece cans  are  fabricated by  forming  the can  bodies from
coated sheets.   This process  includes  a  side-seam  spray  to protect the
soldered  seam  and  an   inside  spray to  protect the inside surfaces.   These
processes  are not considered in this guidance document.

3.16.2 Emission of Pollutants
     Two  significant  emitting   facilities at can  manufacturing  plants are the
base  coating  line and   the  overvarnish  coater/oven portion of  the  printing
line.  On  each  of these  lines emissions are generated from both a roller  coater
and an  oven.   Uncontrolled VOC  emissions  from the line for  base  coating of
sheets (three-piece cans) average  50.8 kg/h (112 Ib/h); emissions from
                              3.16-3

-------
                                    INK
                                APPLICATORS
                BLANKET
               CYLINDER
U)
(T)
I
                SHEET  (PLATE)
                   FEEDER
LITHOGRAPH
  COATER
OVER VARNISH
   COATER
                                                                           WICKET OVEN
SHEET (PLATE)
   STACKER
                                   Figure 3.16-2.   Sheet  printing  operation,

-------
 printing of sheets average  29.5 kg/h  (65 ib/h).   Typical combined  emissions
 from  base coating and  overvarnishing  on a two-piece  can  coating  line are 39
 kg/h (86 Ib/h).1  These emissions may range from 9.1 to 270 kg/h (20  to 600
 Ib/h)  uncontrolled,  depending  on the  line size and speed  and  the type of
 coating.
     When solvent-borne  sheet  coatings are applied by roller, 8 to 27 percent
 of the coating line emissions occur in the coater area and 73  to 92 percent are
 from  the oven.   It  has  been suggested  that  as little as 20  percent of the
 solvent is emitted in the  oven.   (In  two-piece exterior coating lines, up  to 88
 percent  of the emissions  occur  in the  coater area before the can bodies  enter
 the  oven.)  The oven  emissions are low in concentration of solvent per unit
 of exhaust  (most can coating ovens  are designed to operate at 25 percent of
 the lower explosive limit (LEL) and can be  incinerated.
     The spray  coating of the  side seams  and can interior,  end sealing, end
 spraying  or  roll coating  (exterior),  coating storage, and cleaning  operations,
 are  not considered  in  this  document.   Waterborne  interior  and  end sealing
 coatings  are   available  for  some  applications.   The  contact  between  these
 coatings and the canned product must be taken into account.

 3.16.3 Control Measures
     The  two  principal  control  measures utilized in  the  can  industry coating
 lines are low solvent coating materials and fume incinerators.
     The  two types of fume incinerators are thermal incinerators, operating in
 the  temperature  range  of  649°  to  760°C  (1200°  to  1400°F),  and  catalytic
 incinerators,  operating  in the  temperature  range  of 343°  to 510°C  (650° to
 950°F).   Thermal  incinerators  operate  typically at  90 percent efficiency of
 hydrocarbon  control  (across  the  device) with reported values up  to 98  per-
 cent efficiency,  dependent upon the  specific installation  and  coating  materials
being  used.  '  '  '    Catalytic incinerators  have  reported  control  efficiencies
 (across the device) of 90  percent.
     Fume  incinerators  have  been used with coating  materials and can coating
lines with high solvent to solids content ratios.
                             3.16-5

-------
     If  the  coater  area  (the  point  of  application of the coating material) were
enclosed and  the emissions combined  with those from the  oven, a high  per-
centage of  the  coating  line  emissions  would be directed to the fume incin-
erator.   This has  reportedly been attempted at only one  plant, because the
industry feels that enclosures  hinder control  and operation  of the coaters.
     Reduction  of   organic emissions   from  the point of  application  of  the
coating material  by the  use  of  low solvent coatings  (waterborne and/or  high
solids)  is 60 to  90  percent, dependent  upon the particular type  of coating line
and the solvent content of the coating  materials that would  otherwise be used.
The use of low  solvent  coating materials, as compared to high solvent  coating
materials,  eliminates some of the concern  over controlling  fugitive  emissions.
Powder coating  systems  have  limited potential application  to  exterior  base
coats,  interior  body  sprays,  and  overvarnish.   They  are  essentially  100
percent solids and therefore produce no organic solvent emissions.
     Waterborne coatings contain a polymer  or resin  base, water,  an  organic
cosolvent,  and  a solubilizing agent.   The organic cosolvent improves stability
and flow-out, depresses  foaming action,  and controls the  drying rate.  High-
solids  coatings, with  70 to 80  percent  solids  by  weight,  may  be difficult to
apply because  the material is highly  viscous.   A heating unit may be  used to
 raise  the  application  temperature and  thereby  reduce  viscosity.   Powder
 coatings  require a different type  of application equipment.   Powder coating
 technology has not yet been developed to the  point that  thin, continuous  films
 can be produced  at high line speeds, as with  solvent or waterborne coatings.
      Although waterborne, high-solids, and  powder coatings are comparable in
 performance  to solvent-borne coatings in  some applications in the can industry,
 they  are  not available  to replace many  of the present solvent-borne formu-
 lations.   Therefore,  this control  option  is not  universal.   The availability of
 low-solvent  coatings  should  increase  substantially in  the  next several years,
 however.   The can  coating  industry will  require extensive testing  of these
 new coatings to  determine  their  effects on the manufacturing and packaging
 processes  and  on  the canned product.
                               3.16-6

-------
     Ultraviolet-curable  inks  are  cured  with a  UV  light  source  such  as
mercury  vapor lamps.   Although these are totally organic,  very little vapor-
ization occurs  during the almost instantaneous curing.   Therefore,  UV-curing
can reduce  VOC emissions by nearly 100 percent (there may be slight volatil-
ization of  low-molecular-weight  compounds).    This technology  is  currently
limited to  application of  thin,  semitransparent coating  films  and  operations
that would  normally reguire oven  drying between applications of ink colors.
UV curing  is  being investigated  for  single-pass  curing of  base coat,  inks,
and overvarnish  coat for both sheet and two-piece beverage cans.
     Inks  reguiring no overvarnish are currently  in  use.  These "No-Var"
inks  are cured  by exposure to  ultraviolet  light.  When  "No-Var"  inks  are
used,  solvent  emissions from overvarnish  are eliminated and a great deal of
energy is  saved in comparison  to  conventional oven curing of  inks and  over-
varnish .

3.16.4  Emission Limits
     The initial  criterion  for defining LAER  for a surface  coating industry is
the  degree  of emission control  required  by the  most  stringent state  regu-
lation .
     Most organic  solvent emission  regulations are patterned after what is  now
Rule  442 of the  South  Coast  (California)  Air Quality  Management District.
Review of  regulations applicable in the 16 states that contain about 85 percent
of all  surface  coating industries showed that they are essentially the same as
Rule  442.8  Indiana has  the most stringent regulation, which limits organic
solvent  emissions  to  1.4  kg/h  or  6.8  kg/day  (3 Ib/h  or  15  Ib/day) unless
such  emissions are reduced  by  at least 85  percent, regardless of  the reactiv
ity or temperature of  the  solvent.   Organic  solvents that have been deter-
mined to  be  photochemically unreactive  or  that contain  less than  specified
percentages of photochemically  reactive organic materials are exempt  from  this
regulation.
     California Air Resources Board has recently  adopted a model rule for the
control  of  VOC  emissions  from  can and  coil  coating operations.    This  model
rule, which must be met 3 years from the date of  adoption, limits VOC
                              3.16-7

-------
emissions  from 3-piece  sheet  base coating and overvarnish to 180  g solvent/
liter (1.5  Ib/gal)  of  coating minus water, and from 2-piece base coating and
overvarnish  to 250  g  solvent/  liter  (2.1  Ib/gal)  of  coating  minus  water.
These  limits  can  be  achieved  by  the  use  of  low solvent coatings  or  with
add-on  control eguipment.
     In defining  LAER  for surface coating emissions,  it is not appropriate to
exempt  solvents  on  the  basis  of  their  reactivity.    Recent  research  has
indicated   that  substituting   low-reactivity  solvents  for  higher-reactivity
solvents  may  improve  photochemical oxidant  air  quality  in  one  city while
worsening it  in downwind  regions.1   Accordingly,  EPA has adopted  a policy
emphasizing  the  need  for  "positive reduction techniques" rather  than  sub-
stitution of compounds.
     Emission  controls achieved in practice  for  can coating exceed regulatory
requirements  by  a  wide margin.   Therefore,  it is concluded that LAER  for
can coating is a  function of controls achieved in practice rather  than  controls
required  by current regulations.

3.16.5  Determination of LAER
     The  recommended  limitations  are  based  on  SIP's and  on  performance
information available  in early 1979.   It  is  anticipated  that several additional
SIP regulations covering  these sources  will be promulgated and/or  modified in
1979 and  1980 and that appreciable new performance data will  become  available
in  the  near term.  Conceivably,  some SIP  regulations may be more  stringent
than the LAER  suggested  herein.  Furthermore, performance  testing  may
show that  more  stringent limits  than those suggested  are feasible or it may
show that the suggested  limits are appropriate or that they are not achievable
for some specific subcategories.  In any case, the basis for determining  LAER
for many source  categories is  expected  to  change frequently.  Since LAER is
near the  vanguard  of  control  technology,  a  more detailed analysis is partic-
ularly  necessary when addressing modified or  reconstructed  facilities subject
 to  the provisions of  Section  173  of  the  Clean Air Act.  Emission limitations
 reasonable for new sources  may  in  some  instances  be economically  or  tech-
 nically unreasonable  when applied to modified or reconstructed sources of the
 same type.
                               3.16-8

-------
     The lowest achievable  emission rate for  sheet  facecoating and overvarnish
and  two-piece  can  (exterior) basecoating and  overvarnish is determined  to be
a  combination  of limitations  represented  by  RACT  and  emission  reductions
achieved in  practice.   The LAER can  be met by  the use of waterborne coat
ings, which  can achieve  reductions  of  88 to 92 percent.  This  is  comparable
to the RACT limitation of  340 g  solvent/liter  Ib/gal) of coating minus water.
Similar  reductions  can be  achieved by the  use of other low-solvent coatings
(such  as  powder)  or  higher-solids   coatings.    The  application  of  add-on
control   devices  capable   of  providing  equivalent   reductions  is  also  an
acceptable  way  of meeting these  limitations.
                              3.16-9

-------
                          REFERENCES
1.   Control of Volatile Organic  Emissions  for  Existing  Station-
    ary Sources.   Volume II:   Surface Coating  of  Cans,  Coils,
    Paper,  Fabrics,  Automobiles,  and Light-duty Trucks.
    EPA-450/2-77-008,  U.S.  Environmental  Protection Agency,
    Research Triangle  Park, North Carolina.  May  1977.

2.   Listing of Surface Coating Industry Sources  from National
    Emission Data System (NEDS).   Unpublished. U.S. Environ-
    mental Protection Agency,  Research Triangle Park, North
    Carolina.  February 1978.

3.   Incinerator Efficiency Tests.  (Confidential  report provided
    by can coating company.)  May 1975.

4.   Source Listed in State of Indiana Emission Inventory System
    (EIS).  Communication with M. Knudson, Indiana Air  Pollution
    Control Division.   February 20, 1978.

5.   Personal  communication with J. Helgeson, Continental Can
    Company,  Portage,  Indiana.  February 20, 1978.

6.  Levinson, S.B.  Electrocoat, Powder Coat,  Radiate.   Which
    and Why?  Journal  of Paint Technology.  44(570):42, July
    1972.

7.  Recommended  Policy of  Control of Volatile Organic Compounds.
    U.S. Environmental Protection Agency.  Federal  Register (42
    FR 131).  July  8,  1977.

8.  Source Assessments:   Prioritization of Air Pollution From
    Industrial Surface Coating Operations.  EPA  650/2-75-0192.
    U.S. Environmental Protection Agency.  February 1975.

9.  Pantalone, J.A.,  and  L. Shepard.   Consideration of  Model
    Rule  for the Control  of Volatile  Organic  Compound Emissions
    From  Can and Coil Coating Operations.   Prepared for
    California Air  Resources  Board.   July 26, 1978.

10.  Control Strategy  Preparation Manual  for Photochemical  Oxi-
    dant.   OAQPS 1.2-047,  U.S.  Environmental  Protection Agency.
    Januarv 1977.
                              3.16-10

-------
3.17 MAJOR SOURCE CATEGORY:  METAL COIL COATING—VOLATILE
     ORGANIC COMPOUND (VOC) EMISSIONS.
3.17.1  Process Description
     Coil coating is the coating of any flat metal sheet or strip
that  comes  in  rolls  or  coils.    The metal  is  typically  roll
coated on one or both sides on a continuous production line.   The
metal may  also  be  printed or embossed.  The coated metal is slit
and  fabricated  by  drawing,  stamping,  roll  forming,   or  other
shaping operations  into finished products  to  be  used  for cans,
appliances,  roof  decks,   shelving,   industrial  and  residential
siding,  cameras,  culvert  stock,  cars, gutters,  and many other
items.
     On some lines,  the metal is uncoiled at one end of the line
and  recoiled  at  the  opposite  end.  On  other  lines,  called
"wraparound" lines,  the metal is uncoiled  and recoiled at about
the  same   point  on  the line.   Some  coil  coating lines  have  a
single  coater   and  one  curing   or  baking  oven;   others,  called
"tandem" lines, have several successive coaters, each of which is
followed by  an  oven so that several different coatings  may be
applied in a  single pass.   Figure  3.17-1 is a  schematic  of  a
tandem coil coating line.
     The metal is moved through the line by power-driven rollers.
It is uncoiled  as  the  process begins and goes through a splicer,
which joins  one coil  of  metal  to the end of another  coil for
continuous, nonstop production.   The metal is then accumulated so
that, during a  splicing  operation,  the accumulator  rollers can
descend to  provide  a  continuous  flow of  metal  throughout the
line.   The  metal  is  cleaned at  temperatures of 49°C to  71°C
(120°F to 160°F), brushed,  and rinsed to remove dirt,  mill scale,
grease,  and  rust before coating  begins.   Pretreatments, varying
with the type of metal being coated and the type of coating being
applied, are then  applied to the metal to protect  against cor-
rosion and provide proper coating adhesion.
                              3.17-1

-------
aCCUWVlATOM
                                                                      ACCUMUIATON
                                           MME
                                          COATEM
METAL CLEAHINB


M


IETMEATMEI


T
D

MME
OVEN
PRME
OUENCM
                                                                   3
                                                                              SHEAR

                                                                             D
                                          TOTCOAT
                                          COATED
                                                        TOPCOAT
            TOPCOAT
            QUENCH
                U    u
                  MECOIlflIC
                   METAL
       Figure 3.17-1.   Diagram  of  coil coating  line.

-------
     The first or prime  coat may be applied on one or both sides
of the metal by a set of three or more power-driven rollers.  The
pickup  roll,  partially  immersed in  the  coating,  transfers  the
coating to  the applicator roll.  The metal  is  coated,  typically
in a  reverse roll  fashion,  as  it passes  between the applicator
roll and the  large  backup roll.  A third roll,  called a "doctor"
roll,  may be  used to control film thickness when applying a high
viscosity coating by making contact with the pickup roll.
     The applied  coating  is usually dried  or baked in  a con-
tinuous,  multizone,  high  production  catenary,   flotation,   or
double-pass  oven.   The  temperatures  of the preheat,  drying,  or
baking  zones  may  range  from  38°C  to  538°C  (100°F to 1000°F)
depending on  the type  and film thickness of coating used and the
type of metal being coated.   The flow  rates of exhaust from the
                                     3             3
ovens range from approximately 1.9 Nm /s to 12.3 Nm /s (4000 scfm
to 26,000 scfm).  Many of these ovens are designed for operation
at 25 percent of  the room-temperature lower explosive level when
coating at  rated solvent input.  As the metal exits the oven,  it
is cooled in  a quench chamber by a spray  of water or a blast of
air and then  is water-cooled.
     A  second coat or  topcoat may  be applied  and  cured  in  a
manner  similar to the  prime  coat.   The topcoat oven,  however,  is
usually longer than the  prime coat oven and contains more zones.
     Another method of applying a prime coat on aluminum coils or
a  single  coat on steel  coils is to  electrodeposit a waterborne
coating to  one or  both  sides  of the  coil.  The  coil  enters  a
V-shaped electrocoating  bath that  contains a roll on the bottom.
As the metal  goes around the roll,  electrodes on each side can be
activated to  permit coagulation of the paint particles on one or
both surfaces of  the  coil.   The coated coil is  then rinsed and
wiped by  squeegees  to remove  the  water and excess paint parti-
cles.    With  steel  coils,  the  electrodeposited  coating  must  be
baked in  an oven.   With  aluminum coils, however,  the prime coat
is stable  enough  that the metal can be passed immediately over
                              3.17-3

-------
rolls to  the topcoat  coater without  destroying the  finish and
then can be baked as a two-coat system.
     After  cooling,  the  coated metal  passes  through  another
accumulator,  is  sheared  at  the  spliced  section,  waxed,  and
finally recoiled.  The accumulator rolls rise during the shearing
process,   collecting the coated  metal  to ensure  continuous pro-
duction.

3.17.2  Emissions of Pollutants
     Emissions  from a  coil  coating line  come  from  the coating
area, the preheat  and  baking  zones of  the  oven,  and the  quench
area.   These emissions  are  mainly  volatile  organics  and  other
compounds,  such as  aldehydes,  that result from  thermal  degra-
dation  of volatile  organics.   Emissions from  the  combustion of
natural,  gas,  the fuel used most  commonly to  heat the ovens, are
carbon  monoxide,  unburried fuel,  nitrogen  oxides, and aldehydes.
When  fuel  oil   is  used  to  heat the   ovens,   sulfur  oxides and
greater  quantities  of  nitrogen oxides  and  particulates will be
emitted.
     Of  the  uncontrolled  organic  vapors  emitted  from  the  coil
coating  line, the  oven  emits approximately  90 percent, the  coater
area approximately 8 percent,  and the  quench  area the  remaining  2
percent.   Only  these organic vapors are of concern in this docu-
     2  3
ment.z'J

3.17.3   Control Measures
     Three types of control measures  have  been widely applied  to
coil  coating operations:   thermal incineration,  catalytic  in-
cineration,  and reformulation  from solvent-based  to  water-based
or high-solids  coatings.
      Thermal incinerators, successfully used in many  coil coating
 facilities,  have achieved organic emission  reduction  efficiencies
of 90  to 99 percent, depending on the  specific operation.  '   The
                               3.17-4

-------
 overall  plant reduction  in  emissions  depends on the VOC capture
 efficiency  at the coating application  area.   The coating area can
 be  enclosed, and  essentially  all  of  the organic  vapors  can be
 captured  and vented to the incinerator.  As a minimum, 96 percent
 of  the  vapors from the coating  line should be subject to collec-
 tion—100 percent  of  oven emissions plus 60  percent of emissions
 from  the  coater and quench areas.  Table 3.17-1 presents typical
 emissions based on  assumed control efficiencies for various coil
 coating operations.
      Many coil  coating facilities  are currently using catalytic
 incinerators  to reduce organic emissions and are  achieving re-
 duction  efficiencies  of  85  to 95   percent,  depending on the
                   4 5
 specific operation. '
      Reformulation  of  coatings  is   from  organic  solvents  to
 waterborne  or  high-solids coatings.   Waterborne  coatings  have
 been  successfully  applied,   within limits,  to   several  coating
 lines  and have  reduced   organic  emissions  by 70 to  95  percent,
 depending on the processes and the solvent level of the original
 solvent-borne  coating.  '     High-solids  coatings  have only re-
 cently  seen  significant  use in  the   coil coating  industry, but
 progress  is  being  made in commercializing  coatings with medium-
 high  to  high-solids content.   Table   3.17-2  lists  the potential
 percentage  reduction,   in pounds  of  organic solvent per  unit
 volume of coating,  that  can be  realized  by  converting to water-
 borne and high-solids coatings.
     The  options,  however,  are  limited.   There  is  a  lack  of
 waterborne   and  high-solids   coatings  equivalent   to  organic-
 solvent-borne coatings  for  many applications,  especially  where
 resistance  to corrosion  or  wear is  critical,  or  where  forming
 operations are  severe.   Some coatings used  in the  industry can
poison  incinerator  catalysts.   Incineration,   especially  non-
 catalytic, will  increase the  use of natural  gas or other fuels if
no nearby operations  can use the recovered  energy.   This  latter
                              3.17-5

-------
U)
-J
I
en
                  TABLE 3.17-1.  EMISSIONS FROM COIL COATING OPERATIONS
                                                            4
                              [Emission rates in kg/h (lb/h)]
Type of operation
Duct work
Canopies and
awning
Fencing
Screening
Gutters
Metal doors, exclud-
ing garage doors
Typical
uncontrolled
145.6
(320.6)
83.99
(185.0)
29.5
(64.9)
20.6
(45.4)
12.9
(28.5)
7.63
(16.8)
Incineration
1.5
(3.2
0.86
(1.9
0.32
(0.7
0.23
(0.5
0.14
(0.3
0.091
(0.2
-14.6
- 32.1)
- 8.40
- 18.5)
- 2.9
- 6.5)
- 2.0
- 4.5)
- 1.3
- 2.9)
- 0.77
- 1-7)
Controlled
catalytic,
combustion
7.3 - 21.8
(16.0 - 48.1)
4.2 - 12.6
(9.3 - 27.8)
1.5 - 4.4
(3.2 - 9.7)
1.0 - 3.1
(2.3 - 6.8)
0.6 - 2.0
(1.4 - 4.3)
0.4 - 1.1
(0.8 - 2.5)
               Based on 90 to 99 percent control.


               Based on 85 to 95 percent control.

-------
TABLE 3.17-2.   POTENTIAL REDUCTIONS FROM USE,OF
    WATERBORNE AND HIGH-SOLIDS COIL COATINGS
Coating
formulation,
by volume
Waterborne
32% solids
54.4% water
13.6% organic
solvents
Organic
solvent-borne
20% solids
80% solvent
50% solids
50% solvent
70% solids
30% solvent
kg of organic
splvent per
m of coating
minus water




264



708

444

264
(Pounds of organic
solvent per gallon
of coating minus
water)




(2.2)



(5.9)

(3.7)

(2.2)
Potential reduction
by using waterborne
coatings, %








90

58

0

-------
limitation  has  been  overcome through  the use  of total  energy
     , .    6,8,10
recycling.

3.17.4  Emission Limits
     The initial criterion for defining  LAER  for a surface coat-
ing industry  is the degree  of  emission control  required  by the
most stringent  regulation adopted  and  successfully enforced by a
state or local air pollution control agency.
     As reported  elsewhere,  most regulations  of organic solvent
emissions  are patterned  after what is now Rule  442  of the South
Coast (California) Air Quality Management  District.     Review of
regulations in the 16 states that contain about 85 percent of all
surface coating industries showed them to be essentially the same
as Rule 442.12  Indiana has the most stringent regulation in that
it limits  organic  solvent emissions to 1.4 kg/h or 6.8 kg/day (3
Ib/h or 15 Ib/day)  unless such emissions are reduced by at least
85 percent,  regardless of the  reactivity  or  temperature  of the
solvent.   Organic solvents that have been determined to be photo-
chemically unreactive or that  contain less  than specified per-
centages  of photochemically reactive organic materials are exempt
from this  regulation.
     California Air  Resources Board has recently  adopted a model
rule  for  the control of VOC  emissions  from  can and coil coating
operations.13   This  model rule, which must be  met  3 years from
the  date   of  adoption,  limits VOC  emissions  from the coil prime
and topcoating  or single  coating line  to 120  g solvent/liter (1.0
Ib/gal) of coating minus water through the use  of add-on control
equipment,  unless the solvent content  of  the coating used is no
more than 200 g/liter  (1.7 Ib/gal)  of  coating minus water.
      In defining LAER for  surface coating emissions,  it is  not
appropriate  to  exempt   solvents   on  the  basis  of  reactivity.
Recent  research  has indicated  that  substituting low-reactivity
solvents  for higher-reactivity solvents may improve  photochemical
                               3.17-8

-------
 oxidant air  quality  in  one  city  while worsening it in downwind  regions.
 Accordingly,  EPA  has  adopted a  policy  emphasizing  the  need  for  "positive
 reduction  techniques" rather than  substitution of compounds.11
      Emission  controls  achieved  in  practice for  metal coil coating  exceed
 regulatory requirements  by  a  wide margin.   Therefore, it is  concluded that
 the LAER for metal coil  coating is a function of controls achieved in practice
 rather than controls required by existing regulations.

 3.17.5  Determination  of  LAER
      The  recommended  limitations  are  based on SIP's  and on  performance
 information  available  in  early 1979.  It is  anticipated  that several  additional
 SIP regulations covering  these  sources will  be  promulgated  and/ or modified
 in  1979 and  1980  and  tht  appreciable  new performance  data  will  become
 available in the  near term.  Conceivably,  some  SIP  regulations may be more
 stringent  than the  LAER  suggested herein.   Furthermore, performance testing
 may  show that more   stringent limits than  those suggested are  feasible or it
 may  show  that the  suggested  limits   are  appropriate   or  that  they are  not
 achievable  for some   specific subcategories.   In  any  case,  the  basis  for
 determining  LAER for many source categories is  expected to change frequen-
 tly.  Since LAER is near the vanguard of control technology, a more detailed
 analysis is particularly necessary  when addressing modified  or  reconstructed
 facilities  subject  to  the  provisions  of  Section  173  of the  Clean  Air Act.
 Emission limitations reasonable  for  new  sources  may  in  some  instances  be
 economically or technically unreasonable  when applied  to  modified or recon-
 structed sources of the same  type.
     Incineration  efficiencies  of  98 to  99 percent have  been documented and
 are attainable  for coil coating  operations  where  thermal incinerators  (after-
                    r-  Q                                                v
 burners)  are  used.  '     Incineration   efficiencies of 95 percent  have  been
attained for operations utilizing  catalytic incineration.   As  discussed earlier,
overall  control  efficiency  also depends  on  VOC   collection  efficiency.   All of
the oven exhaust  can be directly  vented to the control  device,  and  if the
coating area is  enclosed,  essentially all  of the VOC emissions from this facility
can also be  captured  for  incineration.   Switching  to waterborne  or highsolids
                              3.17-9

-------
coatings will eliminate  some of the concern over  collection of solvent emissions
since the  organic  content  is substantially  reduced.   The derivation  of  an
equivalent waterborne  coating followings:

     Overall control =  96% collection, 97% reduction
                     = 93%
     Assuming a typical  coating contains 25 percent solids,
        100 gallons of  coating contains 25 gallons solids
                                       75 gallons solvent
     Reducing the solvent by 93% = (75) - [0.93 (75)] =5.3
     gallons of solvent allowed
     Assuming a solvent density of 7.36 Ib/gallon, 5.3 gallons =
     39 Ibs of solvent.
     Equivalent coating required = 39 Ib of solvent/(25 +  5.3)
                                   gallons coating munus  water
                                 = 1.3 Ib of  solvent/gallon
                                   coating minus water
                                 = 155  g of  solvent/liter
                                   coating minus water

      The  best controlled  systems  consist of  an  enclosed  coating area and a
 thermal incinerator.   The  use of  a heat  recovery  system  on  this  type of
 operation  has the  additional capability of actually reducing the  fuel  usage in
 the plant up  to  87 percent.6  Therefore, it may  be concluded that the lowest
 achievable emission rate for  a coil coating line (coating  application  area  and
 oven)  is  an  efficient VOC collection scheme in  conjunction  with a  97 to 99
 percent reduction in  emissions, or  use of the  equivalent  low-solvent coating.
                                3.17-10

-------
                           REFERENCES


 1.   Much of the material that  follows is extracted almost
     directly  from:  Control of Volatile Organic Emissions  from
     Existing  Stationary Sources.  Volume II:  Surface Coating of
     Cans, Coils, Papers, Fabrics, Automobiles, and Light-Duty
     Trucks.   U.S. Environmental Protection Agency, Research
     Triangle  Park, North Carolina.  Publication Number EPA
     450/2-77-008.  May 1977.

 2.   Cosden, W.B.  The Ecology  of Coil Coating, Metal Finishing.
     November  1974.  p. 55-58.

 3.   A Study of Gaseous Emissions from the Coil Coating Process
     and Their Control.  Scott  Research Laboratories,
     Plumsteadvilie, Pennsylvania.  (Prepared for the National
     Coil Coaters Association.  October 1971.)

 4.   Hughes, T.W., D.A. Horn, C.W. Sandy, and R.W. Serth.  Source
     Assessment:  Prioritization of Air Pollution from Industrial
     Coating Operations.  Monsanto Research Corporation, Dayton,
     Ohio.  Prepared for U.S. Environmental Protection Agency,
     Research  Triangle Park, North Carolina.  Publication Number
     EPA 650/2-75-019.  February 1975.

 5.   Telephone conversation with Dr. F. Graziano,  Pre-Finished
     Metals, Incorporated,  Elk  Grove,  Illinois.  January 23,
     1978.

 6.   E.J. Kurie, Manager,  Ross Air System Division, Midland Ross
     Corporation,  New Brunswick, New Jersey.  INERTAIR Drying
     Dramatically Reduces Fuel  Consumption on Coil Coating Line.
     (Prepared for National Coil Coaters Association Fall Tech-
     nical Meeting.  Chicago, Illinois.  October 1977.)

 7.   E.J. Vincent,  et al.   Are Afterburners Obsolete?  (Presented
     at Air Pollution Control Equipment Seminar,  APCA/National
     Association of Corrosion Engineers.   Atlanta,  Georgia
     January 17-19,  1978.)

8.   W.  Parsons,  Project Engineer,  Inyrco,  Incorporated,
     Milwaukee, Wisconsin.   Thermal Oxidation System Actually
     Saves Energy for Coil  Coater.   Reprint,  Pollution Engineer-
     ing.  May 1977.
                              3.17-11

-------
9.   Telephone conversation with J.C.  Magdich,  Plant Engineer,
     Kaiser Aluminum and Chemical Corporation,  Toledo,  Ohio.
     January 25,  1978.

10.  Telephone conversation with W.B.  Cosden,  Vice-President,
     National Coil Coaters Association,  PPG,  Incorporated,
     Beverly, New Jersey.  January 30, 1978.

11.  Recommended Policy on Control of Volatile Organic Compounds.
     U.S. Environmental Protection Agency.   Federal Register July
     8, 1977  (42 FR 131).

12.  Data on the geographical distribution of surface coating
     industries can be found in:  Source Assessments:  Prioriti-
     zation of Air Pollution from Industrial Surface Coating
     Operations.  U.S. Environmental Protection Agency.  Publi-
     cation Number EPA 650/2-75-0192.  February 1975.

13.  A series of documents which are referenced and summarized
     in:  Control Strategy Preparation Manual for Photochemical
     Oxidant.  U.S. Environmental Protection Agency.  Publication
     Dumber OAQPS 1.2-047.  January 1977.
                               3.17-12

-------
3.18  MAJOR SOURCE CATEGORY:  PAPER COATING— VOLATILE ORGANIC
      COMPOUND (VOC) EMISSIONS
3.18.1  Process Description
     Paper  is  coated for a  variety  of decorative and functional
purposes  by applying waterborne,  organic-solvent-borne,  or sol-
ventless  extrusion  materials.    Products  with  coatings  that
incorporate organic  solvents include  adhesive  tapes and labels,
decorated paper,  waxed paper,  book  covers,  office paper, carbon
paper, typewriter ribbons, and photographic films.
     Figure  3.18-1  shows  a  typical  paper  coating  line.   Com-
ponents  include  an unwind  roll,  a coating applicator,  an oven,
various tension and chill rolls, and a rewind roll.  Coatings may
be  applied  to paper in  several ways.   The  main applicators are
knives, reverse rollers, or rotogravure  devices.  A knife coater
consists  of a blade that scrapes  excess coating from the paper,
as  shown in  Figure  3.18-2.   The  position  of  the  knife can be
adjusted  to control  the thickness of the coating.   The reverse
roll coater applies a constant thickness of coating to the paper,
usually  by  means  of three  rollers,  each rotating  in  the  same
direction,  as  shown  in Figure 3.18-3.   A  transfer roll  picks up
the coating solution from  a trough and transfers it to a coating
roll.  A  "doctor" roll removes  excess material  from the coating
roll, thereby determining the thickness of the coating.  A rubber
backing roll supports  the  paper web at the point of contact with
the coating roll,  which is  rotating in a direction opposite to
that of the paper.   Rotogravure is usually considered a printing
process—the coating is picked up  in a recessed area of the roll
and transferred directly to the substrate.
     Ovens  range  from  6.1  to  61 m (20  to 200  ft)  in length and
may be  divided  into  two to  five  temperature zones.  The first
zone, where most of the  solvent  is evaporated,  is  usually at a
low temperature  43.3°C (£llO°F).  Other zones  are maintained at
progressively higher temperatures  to  cure  the coating after most
                              3.18-1

-------
                                          ZONE1
                                         EXHAUST
                                                        ZONE 2
                                                       EXHAUST
CO
I
             HEATED AIR
            FROM BURNER
           REVERSE ROLL
              COATER
UNWIND
                  \
^ — , _ .
•^^1 1
-*"O O

| 	 If 	 1 1 	 -
0 0
                                           OVEN
                                                                            HOT AIR NOZZLES
                                                                                       TENSION ROLLS
                                                                                           REWIND
                           Figure  3.18-1.   Typical  paper  coating line.

-------
                EXCESS COATING
                                     BLADE     COATED WEB
                            PAPER WEB
   Figure J.18-2.   Knife application  technique  for coating paper,
                                        DOCTOR ROLL
              METERING GAP
            TRANSFER ROLL
                                                   COATED PAPER WEB
                                                     BACKING ROLL
                                     COATING RESERVOIR
Figure  3.18-3.   Reverse roll application technique for  coating  paper,
                                  3.18-3

-------
of the  solvent has  evaporated.   Exhaust streams  from  different
zones may  be  discharged independently  or  collected in  a  common
header in front of the air pollution control device.  The average
exhaust temperature is about 93.3°C (200°F).

3.18.2  Emissions of Pollutants
     Many different organic solvents are used, including toluene,
xylene,  methyl  ethyl  ketone,  isopropyl alcohol,  methanol,  ace-
tone, and  ethanol.   Solvent emissions  from an individual coating
plant  vary with  the  size  and  number  of  coating lines.    In a
typical  paper coating  plant,  about 70 percent of all solvents
used  are emitted from  the  coating line.   Uncontrolled emissions
from  a  single  line may vary  from 2.3 to  454 kg/h  (50  to 1000
Ib/h), depending  on the line size.  A coating line consists of a
coatir"  applicator  and a drying  oven,  the  two significant emit-
ting  facilities.
      The remaining  30 percent of plant emissions are from sources
such  as  solvent transfer, storage, and mixing operations.

3.18.3   Control Measures
      Reductions  in  solvent emissions  of  80  to  99  percent  are
achievable  by  use   of  low-solvent coatings,  as  shown  in Table
3.18-1.  These  coatings form organic resin films  with properties
equal to  those  of typical solvent-borne  coatings.   Waterborne
coatings are  two-phase systems  in which water is  the continuous
phase and  a polymer resin is the  dispersed,  phase.   When  the water
is evaporated,  the  polymer  coating is cured and  forms a  film with
properties  similar  to  those  of  organic-solvent-borne  coatings.
Plastisols usually  contain  little or no solvent, although  solvent
is  occasionally added to improve  flow  characteristics.  Hot-melt
coatings contain no  solvent.   The polymer resins  are applied to
the  paper  surface  in a molten  state, and  all materials  deposited
on  the paper remain as part of the coating.   A  plastic  extrusion
                               3.18-4

-------
  TABLE 3.18-1.  SOLVENT EMISSION REDUCTIONS ACHIEVABLE BY USE
      OF LOW-SOLVENT COATINGS IN THE PAPER COATING INDUSTRY
Type of low-solvent coating
Reduction achievable,  %c
Waterborne coatings

Plastisols

Extrusion coatings

Hot-melts

Pressure-sensitive adhesives

   Hot-melt
   Waterborne
   Prepolymer

Silicone release agents

   Waterborne emulsions
   100 percent nonvolatile coatings
         80-99

         95-99

            99+

            99+
            99
         80-99
            99
         80-99
            99+
  Based on comparison with a conventional coating containing
  35 percent solids  by volume and 65  percent organic solvent
  by volume.
                             3.18-5

-------
coating is  a  type  of hot-melt coating in which  a  molten thermo-
plastic  sheet is  discharged  from  a  slotted dye  onto  a  paper
substrate.  The moving  substrate  and  molten plastic are combined
in  a  nip between  a  rubber  roll and  a chill roll.   Prepolymer
adhesive  coatings  are applied as a  liquid composed  of monomers
containing  no solvent.  The  monomers  are polymerized  by either
heat or radiation.   Although  these  prepolymer systems show prom-
ise, they are only in a developmental stage.
     Carbon adsorption  systems can  be 97 to 98 percent efficient
in  controlling organic  solvent  vapors  that  are  drawn into the
carbon bed.2'3   Carbon adsorption has been used since the 1930's
for  collecting solvents  emitted from paper  coating operations,
mainly because it is  profitable  to  reclaim the  emitted  solvent.
The efficiency of  control equipment operation depends  largely  on
inlet  solvent concentration;  with an inlet concentration of 1000
ppm  toluene,   the  achievable control  efficiency  is  97  percent,
whereas with  an  inlet concentration of 3000 ppm  it  is practicable
to  achieve control  efficiencies of 98  percent.2   These control
efficiencies  pertain  to  the  control  equipment  only  and do not
reflect  the  efficiency  of  solvent  capture and  emission  delivery
to  the  control  device.  Essentially  all  of  the  emissions from
this  area  can  be  captured with  hoods.   Solvent recovery in the
range  of 96 percent  of the  solvent  introduced  to the  coating line
                       14
has been demonstrated.  '
      Thermal  incinerators may be used to  control  organic vapors.
 Catalytic incinerators have  rarely been  applied in paper coating
 operations using roll  or knife  coaters,  but  certainly are  appli-
 cable.   Incinerators operating  at high  temperatures  can be  up  to
 99 percent efficient  in controlling  organic vapors  directed  to
 the incinerator.5   Although  no  documentation has been acquired,
 industry  sources  have  stated  that high  efficiencies  (98  to  99
 percent  across  the  device)  are achievable by  use of  an  after-
 burner in  a  paper coating  plant.  Overall facility control would
 be less  because of emissions that escape capture.
                               3.18-6

-------
     Incinerators  and  carbon   adsorbers   are  the   two  proven   add-on
 control    devices    for   controlling    organic   solvent   emissions   from
 paper    coating   lines.     Both    have   been   retrofitted   onto   paper
 coating    lines    and   are   being    operated   successfully.     The   main
 constraint  on   the   use  of  incinerators   is  the  possible  shortage   of
 natural    gas.     Often   the   combination    of   afterburner   and   oven,
 however,  uses   no   more  fuel  than  the  oven  alone,  with  proper  heat
          /?
 recovery.     The   major   disadvantage   of   carbon   adsorption   is  that
 some   solvent    mixtures   may    not   be    economically    recoverable    in
 usable   form.   If  the   solvent   can  be   recovered  for   reuse,   carbon
 adsorption represents an economic advantage.
     Control   of  solvent   emissions   from    such   sources   as  transfer,
 storage,    and    mixing    operations    requires    that   solvent-containing
 vessels    be     equipped    with   tight-fitting   covers    that    are   kept
 closed.    Areas   frequently   cleaned   with   solvent   should  be   equipped
 with   hoods  to   capture  solvent   fumes,   which  are  then   ducted  to  a
 control   device.    Sol vent-soaked    rags    should   be   kept   in   closed
 containers.

 3.18.4  Emission Limits
     Initial    criterion    for    defining    lowest     achievable     emission
 rate   (LAER)   for    a  surface   coating    industry   is    the   degree    of
 emission   control   required   by   the   most   stringent   regulation   adop-
 ted   and   successfully  enforced   by   a   state   or   local   air   pollution
 control agency.
     As    reported  elsewhere,   most   organic  solvent   emission    regu-
 lations   are  patterned  after   what   is   now  Rule   442  of  the   South
 Coast    (California)    Air   Quality    Management    District.7    Review    of
 regulations   applicable   in   the   16   states   that   contain   about   85
 percent    of   all    surface    coating   industries   showed   them   to   be
                                        o
 essentially  the   same   as   Rule   442.     Indiana   has   the  most  strin-
 gent   regulation   in    that    it   limits   organic   solvent   emissions    to
1.4   kg/h   (3   Ib/h)   or   6.8  kg/day (15  Ib/day)  unless  such  emissions
are reduced by at  least 85 percent,  regardless of the reactivity
                              3.18-7

-------
or temperature of the solvent.  Organic  solvents that have been  determined to
be photochemically unreactive or that contain less  than specified percentages
of photochemically reactive organic materials  are exempt  from this regulation.
     California Air Resources Board has recently adopted a model rule for the
control  of VOC  emissions  from paper  and fabric  coating operations.    This
model rule, which must be met 3 years from the date of adoption,  limits VOC
emissions from  the coating line to 120 g  solvent/liter  (1.0 Ib/gal) of coating
minus water through the  use of add-on control equipment, unless the solvent
content of the coating used is  no  more than 265 g/liter (2.2  Ib/gal) of coating
minus water.
     In defining  LAER  for surface coating emissions, it  is not  appropriate to
exempt   solvents  on the  basis   of  their  reactivity.   Recent  research  has
indicated  that   substituting   low  reactivity   solvents  for  higher  reactivity
solvents  may  improve  photochemical oxidant  air  quality  in one  city  while
worse^ng it  in  downwind regions.10  Accordingly, EPA has  adopted a policy
emphasizing the need  for  "positive reduction techniques" rather than  substi-
tution of compounds.
     Emission  controls  achieved  in practice  for paper coating exceed  regula-
tory requirements by a  wide  margin.   Therefore,  it is  concluded that  LAER
for  paper  coating is  a  function  of  controls achieved  in practice rather than
 controls required by existing regulations.

 3.18.5   Determination of  LAER
      The  recommended  limitations  are based  on  SIP's and on  performance
 information  available in  early 1979.   It is anticipated  that  several  additional
 SIP regulations  covering  these  sources will be promulgated  and/  or modified
 in  1979  and  1980   and   tht  appreciable  new  performance   data  will  become
 available  in  the near term.   Conceivably,  some SIP  regulations may be more
 stringent than the LAER suggested  herein.  Furthermore,  performance  testing
 may show that more  stringent  limits  than those  suggested  are feasible or  it
 may show that  the suggested  limits  are appropriate  or  that they  are  not
 achievable for  some specific subcategories.  In any case, the basis for deter-
 mining  LAER for many  source  categories is  expected to change frequently.
                               3.18-8

-------
Since LAER is near  the  vanguard of  control  technology,  a more detailed
analysis  is  particularly necessary  when  addressing modified or reconstructed
facilities  subject  to  the  provisions  of  Section  173  of  the  Clean  Air  Act.
Emission  limitations reasonable for  new  sources  may  in some  instances  be
economically or technically unreasonable  when  applied  to modified  or  recon-
structed sources of the same type.
     Control efficiencies of greater than  95  percent across-the-control  device
have  been  documented  and are  achievable for  paper  coating  operations  in
which  thermal  or  catalytic incineration or  carbon  adsorption  systems  are
     >2 O  C
used. '  '    A paper  coating operation  can deliver at least 90 percent of the
VOC emissions to a control device  that  yields an  overall control  efficiency of
85 percent;  this is equivalent  to the uncontrolled emissions from a low-sol vent
coating  containing 310 g of solvent/liter  (2.6  Ib/gal) of  coating minus  water.
This derivation is shown in the following  calculation:

     Assuming a typical coating contains 22 percent solids,
     100 gal of coating contains 22 gal solids
                                 78 gal solvent
     Reducing the solvent emitted yields:  (78) - [0.85 (78)]  =
     11.7 gal permitted
     Assuming a solvent density of 7.36 Ib/gal, 11.7 gal =
     86.1 Ib solvent
     Equivalent coating required = 86.1 Ib solvent/(22 + 11.7)
                                   gal of coating minus water
                                 = 2.6 Ib solvent/gal coating
                                   minus water
                                 = 310 g solvent/liter coating
                                   minus water
     Although  conversion  to   a  waterborne  or  higher-solids  coating  will
significantly reduce VOC  emissions,  it  may  not achieve the  310  g/liter (2.6
Ib/gal)  limitation,  in  which case control of part of the VOC  emission  is still
recommended.    Control  of VOC  emissions  using  add-on   control  devices
providing at least  85  percent overall plant control  is also acceptable.
                              3.18-9

-------
                           REFERENCES


1.   Control of Volatile Organic Emissions for Existing Station-
     ary Sources,  Volume II:   Surface Coating of Cans,  Coils,
     Paper,  Fabrics,  Automobiles,  and Light-Duty Trucks.
     EPA-450/2-77-008,  U.S. Environmental Protection Agency,
     Research Triangle Park,  North Carolina.   May 1977.

2.   Personal communication with W.  C. Moses.  Technical  Manager,
     Chemical Plant Division,  Sutcliffe,  Speakman & Company,
     Limited.  March 10, 1978.

3.   Taback, H. J., et al.  Control of Hydrocarbon Emissions  from
     Stationary Sources in the California South Coast Air Basin.
     Prepared for California Air Resources Board, KVB Interim
     Report 5804-714,  KVB Consultants, Tustin, California.
     November 1976.

4.   Personal communication with S.  Bruntz.   Division of Air
     Pollution, Kentucky Department of Environmental Protection.
     March 1, 1978.

5.   E. J. Vincent, et al.  Are Afterburners Obsolete?  Presented
     at Air Pollution Control Equipment Seminar, APCA/National
     Association of Corrosion Engineers,  Atlanta, Georgia.
     January 17-19, 1978.

6.   Young,  R.A. (ed.).  Heat Recovery:  Pays for Air Incinera-
     tion and Process Drying.  Pollution Engineering,  7^:60-61.
     September 1975.

7.   Recommended Policy on Control of Volatile Organic Compounds.
     Federal Register, 42 FR 131, U.S. Environmental Protection
     Agency.  July 8, 1977.

8.   Source Assessments:  Prioritization of Air Pollution  from
     Industrial Surface Coating Operations.  EPA 650/2-75-0192,
     U.S. Environmental Protection Agency.  February 1975.

9.   Lam, J. V., et  al.  Consideration of a Proposed Model Rule
     for the Control of Volatile Organic Compounds from Paper and
     Fabric Coating  Operations.  Prepared for California Air
     Resources Board, Sacramento, California.  August 23,  1978.
                              3.18-10

-------
10.  Control Strategy Preparation Manual for Photochemical Oxi-
     dant.  OAQPS 1.2-047,  U.S.  Environmental Protection Agency.
     January 1977.

11.  Recommendation by U.S. Environmental Protection Agency,
     ESED.  Raleigh-Durham, North Carolina.   August 8,  1978.
                              3.18-11

-------

-------
                            SECTION 4
                   COST ESTIMATING METHODOLOGY

4.1  INTRODUCTION
4.1.1  Purpose
     The  purpose of  this  chapter  is  to  provide  air  pollution
control officials  and industry personnel with  a  method of esti-
mating preliminary costs  of air pollution control measures.  The
method  presented  herein should  provide  estimates accurate  to
within +  20  to 30 percent  for  specific  cases when adequate data
are available.

4.1.2  Scope
     A methodology  for  developing  total installed capital cost
estimates and total annualized cost estimates is presented.  This
is  a  generalized approach;  specific process or  equipment costs
are not included herein.  Emphasis  is on costs of add-on control
systems;   however,  by following the general procedure  and basic
principles, costs can be  developed  for process modifications and
material or fuel substitution.
     Design parameters  and  other  factors affecting the overall
cost of  installing and operating control  systems are  discussed.
Capital costs  may be estimated for both new plants and retrofit
situations.  Annualized  cost components are separately computed
as  operating  costs,  maintenance   costs,  capital  charges,  and
product  and  energy  recovery  credits.   Secondary  environmental
costs for  wastewater and solid waste  treatment and disposal can
also be developed.  An example of a cost estimation for a control
system  is presented,  and  forms  are  provided  for actual  cost
estimation.
                               4-1

-------
4.2  FACTORS AFFECTING THE COST OF EMISSION CONTROL
     The  major gas  stream  characteristics  affecting  costs  and
selection of equipment  are  the exhaust volume,  pressure, temper-
ature,  moisture  content, and  corrosiveness.   Accurate  data  are
necessary  for  accurate  cost  estimation.  Variations  of  these
parameters, especially  exhaust volume, must be  studied and con-
sidered  in the control design.   Fluctuation  in  gas  flow rates
affect  the  performance  of some control devices more than others.
Gas conditioning equipment such as precleaners,  cooling chambers,
and humidifiers  are required  for some control  options.  Certain
gaseous  components  in  the gas  stream,  especially oxides of sul-
fur,  require   that  corrosion-resistant materials be used  in the
control  system.
     The physical properties of the pollutant and the gas stream
affect  the control system  design  and hence  the  cost.   These
properties  include  particle  size, pollutant concentration, elec-
trical  properties  of the particles,  and  moisture content of the
gas stream.   Accurate  data  on only  the basic variables, however,
will  enable manufacturers or suppliers to furnish costs that can
be used to  develop  preliminary estimates of capital  costs.
      Process  operating methods and  cycles necessarily  influence
system  design, which may affect  overall  costs.   Age,  type,  and
size  of facility;  availability of space  for locating  the  control
device;  and ease of  retrofitting existing facilities all affect
system  selection and ultimate  costs.
      Plant parameters  such as the requirements  for  an availabil-
ity  of  utilities  and  wastewater  and solid waste treatment/dis-
posal facilities  affect the design  and cost of  a control  system.
Plant  location,  local  climate,   geography,  and  demography may
determine more stringent emission limitations.
                                4-2

-------
 4.3   CAPITAL COST ESTIMATES
      Total  capital costs include equipment  costs,  direct capital
 costs,   indirect  capital  costs,  and  contingencies.   The  cost
 estimate for  a  complete  system  will  include  the  cost of  the
 control  device  and  auxiliary equipment such as fans, pumps,  and
 duct  work.    Capital costs associated  with  wastewater and  solid
 waste treatment/disposal facilities  can also be  estimated.

 4.3.1 Limitations,  Restrictions, and Uncertainties
      The accuracy of a cost  estimate  for an APC system may  vary
 from  +  20  to  30  percent (when  based  on preliminary design of
 major equipment)  to +  5  percent  (when prepared  from  complete
 drawings,  specifications,   and  site  surveys).     The procedure
 outlined in this  section  will provide predesign estimates  accu-
 rate  to  within + 20 to 30 percent  of  actual costs if sufficient
 exhaust  gas,  emission, and operating data are available.
      The use of direct  cost  factors without  giving  proper  con-
 sideration  of  their applicability to  site  specifics  lends un-
 certainty  to  the  final  estimates.    The factors  represent the
 average  or typical installation, and these factors will vary  with
 individual  installations.   Because 10  to 15 equipment items may
 be involved in  a  total  cost estimate,  however,  the  errors  will
 tend  to  average out.  The use  of other factors  such  as material
 cost  factors,  cost-scaling exponents,  and cost  indices will add
 to the error and uncertainty of cost estimates.

 4.3.1.1  Retrofitted Systems—
     Because  retrofitted  control  systems may cost considerably
 more  (or sometimes less) than a complete system incorporated into
 the design of a new plant,  the effect of retrofitted systems must
be considered.   This requires  site-specific considerations such
 as existing  duct work,  physical obstruction,  and  process equip-
ment configurations.  The  cost estimating procedure assumes that
                               4-3

-------
actual  retrofitting  requirements are  known  for the  sources  for
which costs are being developed.

4.3.1.2  Process Modifications and Fuel/Material Substitution--
     In some cases a process modification and/or fuel or material
substitution is necessary to reduce emissions to a desired level.
Examples of  such  process  changes include combustion modification
to control NO   emissions  from power boilers; the substitution of
ultraviolet  dryers for  hot-air dryers to eliminate VOC emissions
from printing  operations;  or the conversion of dry rock grinding
to wet  rock  grinding to eliminate particulate emissions.
     The  cost  methodology  does  not  specifically   give  direct
information  required  tc  estimate costs of process modifications.
By  following the  general  procedures, however,  estimates  can be
developed.   For  example,   the  —ices for  major equipment  are
obtained  and cost  factors are  "-pplied  in  the same  way as cost
factors for  add-on control devices.

4.3.2   Capital  Cost Components
     Total   installed   cost,   or total  capital  investment,  is
developed  from the  base  purchase  prices of  equipment items by
multiplying  the unit costs  by individual  direct labor  and mate-
rial cost  factors  and by indirect  cost factors.

4.3.2.1 Equipment Cost—
      Sources of cost  information—An  index  to  industrial  equip-
ment  and manufacturers  ~an  be consulted to obtain  companies  and
contacts   for  acquiring   cost  information   on  each  individual
component.     Chemical  Engineering Equipment Buyers'  Guide    and
other   indexes,   advertisements  in technical   journals,  and  the
yellow pages  in  telephone  books  provide  an excellent means  of
 locating   sales   representatives.    Capital  and Operating Costs
 of Selected Air Pollution Control  Systems,  (EPA-450/3-76-014,  May
 1976)2 is  an  excellent  publication   source  of equipment  costF.
                                4-4

-------
     Past purchase  orders are also  useful  in determining equip-
ment costs, even  though  they do  not readily provide exact equip-
ment  size and  current  costs.   Cost  indexes  and cost  scaling
methods can be used to determine proper costs.
     A third  source of cost information is  actual or model cost
studies for similar processes  given in technical journals, trade
publications,   EPA  documents,  etc.   The  disadvantage of  using
these sources is  that they are sometimes general in their treat-
ment of processes and control systems.   Many times, however, this
can be a reasonable method of estimating costs, with some adjust-
ment of the data.

     Cost indexes—Since prices and  labor  costs vary widely from
one  time  period  to  another,  cost  indexes  are  used  to  update
equipment  costs  from  those  indicated  in  a  previous  year.   One
source of data  on  wage  and price  fluctuations  is  the  Monthly
Labor Review  published by U.S. Bureau  of  Labor Statistics.   The
Chemical Engineering plant index (CE Index) is frequently used to
extrapolate entire system costs to current or future dollars,  and
the Marshall  and Stevens  (M&S) Equipment  Index can  be used  for
specific  equipment  items.   Table   A-l  in  Appendix   A  presents
suitable cost indexes for various  types of equipment.   Figure  4-1
shows  the trend of  plant  costs  and  construction   labor  cost
indexes from 1957 through 1977.
     EXAMPLE:   The purchase  price  of a 500-hp fan was $20,000 in
1962.  The price of  the  same  fan in 1975 would be estimated by
multiplying  the   1962 price by  the  ratio  of 1975  Fabricated
Equipment Cost Index  in  Appendix  A to the 1962 index as follows:

     (1975 price)   =    (1962 price)  (^5 index}
                         $20,000   x
                         $38,000
                               4-5

-------
t/*>
o
     57  58  59  60
                                                                                  76  77
      Figure  4-1.
Chemical Engineering plant cost and  construction labor index.

    (Information obtained from Reference  2.)

-------
     Scaling exponents — When  the purchase price  is available  on
 an  identical  type of  equipment  of a  different  size,   scaling
 exponents  can  be used to adjust the price to apply to equipment
 of  the desired size or capacity.   Tables  4-1,  4-2, and 4-3 pre-
 sent  scaling  exponents  found  in  or  derived from various cost
 sources.   Size/price  relationships are generally exponential and
 follow the equation:

                       CA x
          P  = P      (— — }
           A   rB     
-------
       TABLE  4-1.   COST FACTORS AND  SCALING  EXPONENTS  FOR AIR POLLUTION CONTROL  DEVICES
I
oo
Control device
Fabric filter
Electrostatic precipitator
Wet collector
Cyclones
Absorption units
Carbon adsorber
Catalytic incinerator
Thermal incinerator
with heat exchanger
without heat exchanger
Condensers
Direct cost
factor3
2.0
2.0
3,0
1.7f
3.0
1.6d
1.6
1.6
1.6
3.0
Stainless ,
steel factor
1.5d
2.4
2.4e
?.0e
2.5g
2.5h



Scaling exponents0
Unit
ft2 cloth
ft2 plate
acfm
acfm
acfm
Ib carbon
acfm
acfm
acfm
acfm
Exponent
0.95
0.37
0.70
0.80f
0.62g
0.90d
0.60
0.70
0. 4-0. 5
0.70g
               Consi dpr<=-d an approximate average;  cost factors  could  vary  by up  to + 40 to 50 percent or
               more in some cases.  Includes equipment,  material  and  labor costs.  Does not  include indirect
               cost.  Developed from PEDCo experience  unless  otherwise  noted.
               Suitable for approximation of 304  or 316  type  stainless  steel.
               To escalate cost of a similar unit of a different  size.
               Reference 2.
               Reference 3.
               Reference 4.
               Reference 5.
               Reference 1.

-------
s  4-2.
COST FACTORS  AND  SCALING EXPONENTS FOR AUXILIARY EQUIPMENT  ITEMS
Equipment item
Fan system
Pump system g
Reciprocating
Centrifugal
e
Drivers
e
Chemical injection
Heat exchanger (S/T)
Air cooler

Quench tower
Spray chamber
Reheater
Ductwork
Exhaust stack
e
Hoppers
Transformers
Screw conveyors
Direct cost
factor3
2.0
2.4
2.4
2 A
. *4
1.6
2.3
1.6

3.0
2.5
3.5
2.6
1.5
11
. 1
1.2
1.6e
Stainless b
steel factor
2.8d
2.4
2.0


2.0
CS/SS 2.5
SS/SS 3.1
3.0

3.3
3.3
2.5d
3.3g
3.0




Scaling exponents
Unit
BHP
BHP
BHP
BHP

BHP
ft2s.a.
ft2s.a.

acfm
acfm
acfm
ft c.s.a.
ft
ft3

kVa
length, ft
Exponent
0.96
0.52
0.52
1.00

0.60
0.70
0.70
f
0.85
0.43f
0.78
0.55e'h
1.0
0.68

0.33
0.8e'f
  Considered an approximate  average;  cost  factors  could  vary by up to  +  40 to  50 percent
  or more in some cases.   Includes  equipment,  material and  labor costs.  Does  not include
  indirect cost.  Developed  from PEDCo  experience  unless otherwise noted.
b Suitable for approximation of 304 or  316 type  stainless steel.
c To escalate cost of a similar unit of a  different  size.
  Reference 1.
  Reference 4.
  Reference 2.
  Reference 6.
h Resulting cost is expressed as dollars per linear  foot of ductwork installed.

-------
                    TABLE 4-3.   COST FACTORS AND  SCALING  EXPONENTS  FOR WASTEWATER
                                     AND SOLID  WASTE  TREATMENT  EQUIPMENT
I
h-1
o
Equipment item
Storage tanks
Less than 40,000 gal
More than 40,000 gal
Clar if ier
Vacuum t liter
Sludge dryer
Centrifuge
Rotary drum filter
Ion exchangers
Carbon adsorption
Reverse osmosis units
Direct cost
factor3
2.0
2.0
2.5
4.0
4.0
3. 5
1.6
1.7f
1.5f
1.7
Stainless ,
steel factor
3.25d
3.35d





2.5h
2.5h

^
Scaling exponent
Unit
gal
gal
ft2c.s.a
gprn
gpm
gpm
ft2s.a.



Exponent
0.29d'e
0.63d'e
0.63f
0, 41 f
0.45g
0.44f
0. 63f



               a Considered an approximate average;  cost  factors could vary by up to +_ 40 to 50 percent
                 or more in some cases.   Includes  equipment, material and labor costs.  Does not include
                 indirect cost.  Developed from  PEDCo  experience unless otherwise noted.
                 Suitable for approximation of  304 or  316  type  stainless steel.
               C To escalate cost of a similar unit of a  different size.
                 Reference 7.
               e Exponents are for total installed costs.
                 Reference 8.
                 Reference 4.
                 Reference 1.

-------
carbon steel prices to  stainless  steel prices on control devices
and  auxiliary  equipment.  These  factors  are suitable  for esti-
mating 304 or  316  type  stainless  prices.   Cost factors for other
materials can be found in the cited literature.
     EXAMPLE:  The 1959 price of a 10-hp carbon steel centrifugal
pump was  $400.  The  stainless  steel  factor  is  2.0  (Table 4-2).
The  1977  price  of a 30-hp  stainless steel  centrifugal  pump is
estimated as follows:

     (1977 price) = $400 x ()      x (°° ) x 2.0 = $2800
The accuracy  of  this 1977 price is lessened because it is neces-
sary to use multiple cost factors.

4.3.2.2  Direct Capital Costs —
     Direct costs  are  those associated with the direct installa-
tion  of  equipment.   They  include materials  and labor  for the
installation  of  instrumentation,  electrical  work,  foundations,
structural  work,  site  work,  insulation,  painting,  and piping.
     The  direct  cost  factors  in  Tables  4-1  through  4-3  were
obtained  from PEDCo  experience,   cost  handbooks,  and  the  cited
literature.   These  cost  factors  are multiplied  by  the purchase
price  of each equipment  item  required  to  obtain the direct cost
of  installing that equipment item,  including material, labor, and
equipment cost.
     The direct  cost factors in these tables are considered  to be
average  or  typical;  they can vary  substantially  from job  to job.
To  ensure accuracy  when  using these factors,  consideration must
be  given to site-specific conditions.
                                4-11

-------
4.3.2.3  Indirect Capital Costs--
     Indirect  costs include engineering, field and labor  expenses, contractor's
fee,  freight,  off site,  spares, sales  tax,  and allowance for shakedown.  These
costs are determined  by multiplying the  total direct cost by the indirect cost
factors presented  in  Table  4-4.   These  factors,  which were  developed from
the cited references,  are. considered  typical for air pollution control systems.
Consideration must be  given to their use for site-specific case conditions  to
ensure accuracy.

4.3.2.4  Contingencies--
     Contingencies incorporate  additional unforeseen  expenses  resulting from
equipment breakage,  inclement  weather,  strikes, shipment delays,  and related
items.   Generally, these costs  run  between 10 and 20 percent of the combined
direct  and indirect costs.   For the  purpose of this cost estimating procedure,
a contingency factor of 15 percent will be used.

4.3.2.5  Total Capital Cost---
     The total  capital  cost  is  the  sum  of the  direct  capital  costs,  indirect
capital costs,  and  contingencies.   If possible, estimated  production losses,
research  and development,  and other related  costs  should be  added directly
to  this  total.   Table  4-5  illustrates  the  complete  capital   cost  estimating
procedure.

4.4  METHODOLOGY FOR ESTIMATING ANNUALIZED COSTS
     Annualized   costs  include  direct operating  costs  (labor, maintenance,
utilities,  and  raw materials)  and  fixed  costs  (overhead,  property  taxes,
insurance, capital recovery  costs).   The annualized costs for operating solid
waste  and wastewater treatment/disposal systems associated with the  emission
control facilities  are  included.   Some  pollutants  can be recycled to  the
                               4-12

-------
TABLE 4-4.  INDIRECT CAPITAL COST FACTORS
                                         1,10
Cost item
Interest during construction
Field labor and expenses
Contractor's fee
Engineering
Offsite
Spares
Sales tax
Start-up and shakedown
Total indirect cost
Percent
of direct cost
5.0
10.0
5.0
10.0
1.3
0.5
1.5
5.0
38
                     4-13

-------
             TABLE 4-5.   CAPITAL COST ESTIMATION
 Equipment
   item
                   Purchase
                    price
   Direct
cost factor'
Direct
                  Equipment
                    costc	
                                                Total
                                                direct
A.   Total direct costs

B.   Total indirect costs at 38% (A)

C.   Contingencies at 15% (A + B)

     Total capital costs
          (A + B + C)
a

b
From Tables 4-1, 4-2, and 4-3.

Includes equipment cost and installation, instrumentation,
electrical work, foundations, structural work, site work,
painting, piping, and labor.

Sum of equipment purchase prices.

Sum of individual direct costs.

Includes engineering, contractor's fee, interest during
construction, shakedown, spares, freight, taxes, offsite,
and field and labor expenses.
                            4-14

-------
system, used as a component of another product, or used  as  a  fuel
supplement.  Whenever  an economic value can be attributed  to the
captured  pollutant,  an  appropriate credit  must be  included in
computing the annualized costs.

4.4.1  Limitations, Restrictions, and Uncertainties
     The estimated  annualized cost may be somewhat less accurate
than the estimated capital cost because the accuracy of the fixed
costs depends largely on the  accuracy of the estimates of capital
cost and equipment life.  When adequate operating and exhaust gas
data are available, the estimate of the annualized cost will fall
within +  30 to 40 percent of actual  costs.   The annualized cost
factors in  this section are  considered to be  average or typical
and  consideration should  be  given to  site-specific conditions.

4.4.2  Annualized Cost Components
     A discussion of  the  individual annualized  cost components
and  methodology for estimating  total  annualized costs  are pre-
sented in this section.

4.4.2.1  Utilities-
     Utilities  include  electricity,  water,  steam,  fuel  oil,  and
natural  gas  requirements.    Annual  incremental  quantities  are
determined  through  the use of fan  equations,  heat  balances,  and
the like.   These  quantities  are  then multiplied by the unit cost
for  the  local  area.   Unit costs  for  utilities vary  widely  de-
pending on  the geographical  location,  user classification,  and
consumption rates.

4.4.2.2  Raw Materials—
     Raw materials and special chemicals are required for certain
control systems,  e.g.,  activated carbon  for adsorption  units,
adsorption  chemicals,   limestone  for  flue  gas  desulfurization
                               4-15

-------
units,  and  sulfur  trioxide for ESP  flue  gas  conditioning.   Cost
data are  not presented for  these  materials because  of the wide
variety of  materials  used.   Prices can be obtained from chemical
suppliers  or price catalogs.   Annual consumption  is determined
and is multiplied  by  the  unit cost to obtain  the  annual cost of
raw materials.   Incremental  costs  for  fuel  and material substi-
tution are  computed in the same manner.
     EXAMPLE:   A continuous  source operation emits  16  kg/h  (35
Ib/h)  of  toluene.   The adsorption  rate of  activated carbon is 7
percent,  and the  attrition  rate  1 percent.   The unit  cost is
assumed to be  $1.01/kg.   The carbon requirements  and costs  are
estimated  as  follows:
     -, ,- ,   ^  a.  i      .  ,0.07 kg toluene,  _ ,~n ,,_ ^aT-Knn/h
     16 kg/h toluene  -r  (1>(? kg\arbon   )  - 230 kg carbon/h
     annual consumption =  230 kg carbon/h x 0.01 x  8000  h/yr
                        =  18,400 kg
     annual carbon cost =  $1.01/kg x 18,400 kg
                        =  $18,580  per year
4.4.2.3   Operating Labor—
     Operating  labor  depends  on the type, size,  and complexity of
the control system.   A large,  complex  system  such  as  the flue gas
desulfurization system may  require the continuous effort  of two
or three  operators  (2 to 3 hours  of  labor/hour  of operation),
whereas  a  small wet  collector may require  only occasional  obser-
vation (0.05 to 0.1 hour  of  labor/hour  of operation).  The  degree
 of system  automation and the continuity  of operation  also in-
 fluence   operating labor  requirements.   Labor demands must  be
 determined  on  a  case-by-case  basis,  considering  each equipment
 component  and  the  overall system  operation.   The total man-ho>urs
 per hour  of operation is multiplied by  the  annual system  opera-
 ting  time  and labor  rate  to determine annual  operating labor
 costs.   Labor  rates  vary widely  according  to the  geographical
 location.
                                4-16

-------
4.4.2.4  Maintenance--
     Maintenance costs  include the material,  labor,  replacement
parts, and  supervision necessary to maintain the control system
at the  desired level of performance.   These  costs are sometimes
very  difficult to predict  since they  vary with the  age  of the
system and modes of  operation.  Table  4-6 presents the ranges of
maintenance costs  for various types of control  devices in terms
of  dollars  per  m3/s  (cents  per  acfm).   These  data  generally
correlate with data  from  other cost sources.   The estimator must
determine whether  maintenance will be  low,  typical,  or high,
based on the  complexity  of  a specific application in order to
estimate  the  annual  maintenance cost.   An  alternate  method is
given in Table 4-7  where maintenance  costs  are  expressed  as  a
percentage of capital investment.
     The  exhaust   gas  conditions and  system  complexity largely
determine the degree  of  maintenance   of control  systems.   For
example,   a flue gas  desulfurization system is a highly automated
and complex system that has  problems  associated with scaling and
corrosion.  High maintenance costs are  expected.
     A high-temperature gas  stream from an electric  arc furnace
may  have  to  be  cooled  before  it  enters   a  fabric  filter.
Occasional  bag   failure   and   replacement   necessitate  normal
maintenance.
     A  process with ambient  operating  conditions and  exhaust
temperatures  controlled by  a  medium-energy wet  collector  is an
example  of  a  simple  system with  little mechanization  and low
maintenance  costs.
                               4-17

-------
TABLE 4-6.  ANNUAL MAINTENANCE COSTS OF CONTROL DEVICES'

            [dollars per m /s (cents/acfm)]
Generic type
Gravitational and dry
centrifugal collectors
Wet collectors
Electrostatic precipitators
High voltage
Low voltage c
Fabric filters
Afterburners
Direct flame
Catalytic
Low
17 (0.8)
68 (3.2)
34 (1.6)
17 (0.8)
68 (3.2)
102 ( 4.8)
237 (11.2)
Typical
51 (2.4)
136 (6.4)
68 (3.2)
47 (2.2)
170 (8.0)
203 ( 9.6)
680 (32.0)
High
93 (4.0)
203 (9.6)
102 ( 4.8)
68 ( 3.2)
270 (12.8)
340 (16.0)
1190 (56.0)
 Includes material,  labor,  and  supervision.
 Reference  11;  escalated  to 1977  prices  by Chemical  Engineering
 Labor  Cost Index.
 Add  bag replacement cost:   Assume  1-  to 2-year life and
 $18/m2 ($0.50/ft2)  cloth.
                             4-l8

-------
             TABLE  4-7.   PROCESS  MAINTENANCE  COSTS
Type of system
Simple
Normal0
Complex
Annual maintenance cost
as a percent of capital investment
Labor
1-3
2-4
3-5
Materials
1-3
3-5
4-6
   Adapted from Reference 1,  p.  134.
   Atmospheric conditions,  little mechanization.
   Normal operating condition;  average system.
   Corrosive and severe operating conditions or extensive
   instrumentation.
4.4.2.5  Fixed Cost—
     Fixed  costs   include   capital  charges,   property  taxes,
insurance, overhead,  and administrative costs.   Capital charges
are  determined by  multiplying  the total  capital  cost of  the
system  by the  capital  recovery factor.   The  capital  recovery
factor  is  a  function of the interest rate  and overall equipment
life.   Capital  recovery factors  for various  interest  rates  and
equipment  lives  are shown in Appendix A.   A 10 percent interest
rate  can  be  used  in most  cases.    Insurance,  property  taxes,
overhead,  and other administrative costs are usually estimated at
                                          12
4 percent  of the total capital investment.
     Equipment life  of  pollution control devices is generally 10
to  15  years,  depending upon the application.   Better estimates
can  be obtained  through equipment  manufacturers and suppliers.
     EXAMPLE:  The  following procedure  is  used to determine the
total  fixed cost of a control system requiring  an initial capital
investment of  $1.2  million  at an interest  rate  of 10 percent and
having an  overall  equipment  life of  22  years.    The capital
recovery  factor for 22  years and 10 percent is  0.11401.
                               4-19

-------
Therefore,

Capital recovery cost    = 0.11401 ($1.2 million) = $136,800
Taxes, insurance, etc.   = 0.04 ($1.2 million)    =   48,000
     Total fixed costs   =                          $184,800
     Product recovery and energy credits—When   the   capture  of
pollutants results in an economically recyclable, recoverable, or
salable  product,  product  recovery  credits   should  be  applied
against  annual  costs.   Sometimes  such  credits  substantially
reduce  operating  costs.   First,  the value  (unit cost)  of the
pollutant  must  be determined.   If  the  collected pollutant is
recycled, the actual value can be assumed to be  the final product
price  less  processing costs.   The  annual  amount of pollutant
capture  is  estimated  and multiplied  by the  net unit  value to
determine the product  recovery credit.
      Similarly,  credits  for  energy  (heat)  recovery  should be
estimated  when  applicable.   Heat recovery from  the high tempera-
ture   exhausts   from  incineration  controls  can  greatly   reduce
operating  costs.

      Solid/liquid  waste disposal—It is  necessary to dispose of
pollutants  that cannot be recycled,  recovered,  or reused.   Solid
waste disposal  techniques include several methods of  landftiling.
Liquid waste disposal methods include  landfilling, incineration,
dry-well injection,  and  ocean dumping.  Selection of  a disposal
technique  depends on the  nature and  quantity  of the waste  and
geographical location.   Annual  costs  of each  method  are  not
readily available,  and  a detailed  discussion of these disposal
methods is beyond the scope  of  this document.  Annual costs can,
however, be estimated by  considering the following items:

      0    Expenses,  capital recovery,  and maintenance  on  capital
           items (trucks,  land,  equipment, etc.)
      0    Fuel  costs
                                4-20

-------
     0    Electricity costs
     0    Operating labor
               stationary equipment
               mobile equipment
     0    Materials (barrels,  chemicals,  etc.)
     0    Landfill fees
     0    Administrative costs
The cost of operating  landfills  ranges from  $2 to $5/Mg of waste  for  small
                                               10
operations to $1  to $3/Mg  for  large operations.    If  solid waste materials are
hazardous,  they  may  require  special handling,  such  as disposal in sealed
barrels.   Barrels alone cost $20 each, and hold up to 250 kg of waste.  This
would make disposal  cost at least $80/Mg.

     Wastewater treatment- -Wastewater  treatment facilities  at  a plant may  have
sufficient capacity to  handle  any additional volume generated by an  emission
control  system.   The  unit  treatment cost  in dollars/103 gallons  is determined
from operating records,  EPA  documents,  or other  sources, then multiplied by
the incremental volume to  be treated  annually to  estimate incremental costs.
     When new or expanded wastewater  treatment  facilities are  required, the
cost of  utilities,  raw  materials, chemicals, operating labor,  and maintenance
and fixed costs are  incorporated in the annualized cost  estimate for the over-
all  system.   These costs  can  be separated out from among the direct  emission
control costs.

4.4.3   Total Annualized Costs
     Direct  operating costs, fixed  costs,  product credits, and energy credits
for the operation  of  air  pollution control  systems and secondary facilities can
be summarized  in  a format  given in Table  4-8.
                               4-21

-------
                 TABLE 4-8.  SUMMARY  OF PREDESIGN ESTIMATE  OF ANNUALIZED COSTS
I
to
ro
Item
DIRECT OPERATING COST
I. Utilities
Raw materials
Electricity
Water
Steam
Gas
Fuel oil
II. Operating labor
Direct labor
Supervision
III. Maintenance (see Tables 4-6
Labor and materials
Supplies
Replacement parts
IV. Sludge disposal
V. Wastewater treatment
FIXED COSTS
I. Taxes, insurance, over-
head, etc.
II. Capital recovery charges

PRODUCT RECOVERY CREDIT

Unit
$
S
$
$
$
$
$
15%
, 4-7
$
$
cost Quantity Cost
/ton tons/yr
/kWh kWh
/103 gal 103 gal/yr
,/103/lb 103 Ib/yr
/106 Btu 106 Btu/yr
10J gal 10J gal/yr
/man-hour man-hours/h
direct labor operation
)
/ton tons/yr
/103 gal 103 gal/yr
TOTAL DIRECT OPERATING
4% of total capital investment
% of total capital investment

$
TOTAL FIXED COSTS
/ton tons/yr
TOTAL ANNUALIZED COSTS

-------
4.5  COST-EFFECTIVENESS
     The  cost-effectiveness  of a control technique is arrived at
by comparing the degree of emission reduction with the total cost
of achieving the  reduction.   It is expressed in dollars per unit
weight of pollutant removed and is computed by dividing the total
annualized  cost by  the total  annual quantity of  pollutant re-
moved.  As control efficiency improves, the quantity of emissions
is reduced,  but the  cost  of control  increases.   Cost-effective-
ness relationships are useful in emission control decision making
since  several  feasible  systems  usually  are  available  for  con-
trolling  each  source of emission.  Cost-effectiveness relation-
ships  vary  from   industry  and  from  plant  to plant within  an
industry.
     EXAMPLE:  Assuming  the  annualized cost of an  air pollution
control  system  is  estimated at $126,000, the  system  captures  99
percent  of  the  39 kg/h  of emissions  generated,  and  total oper-
ating time  is  8000 hours per  year, cost-effectiveness is  deter-
mined as follows:

  Amount of pollutant captured =0.99  (39 kg/h) (8000  h/yr)
                               = 3.09  x 105  kg/yr
                           jt. -i • ,
  Cost-effectiveness = -
                       3.09 x 105 kg/yr

                     = $0.41/kg pollutant removed
                               4-23

-------
                          REFERENCES


1.  Peters,  M.S.,  and K.D. Timmerhaus.   Plant Design and
    Economics for Chemical Engineers.   McGraw-Hill,  New York,
    New York.  1968.

2.  Kinkley, M.L., and R.B. Neveril.  Capital and Operating
    Costs of Selected Air Pollution Control Systems.
    EPA-450/3-76-014, U.S. Environmental Protection Agency,
    Research Triangle Park, North Carolina.  May 1976.

3.  Manual published by Richardson Engineering Service, Inc.
    Downey, California.  1974.

4.  Guthrie, K.M.  Capital Cost Estimating.  Chemical Engineer-
    ing Magazine.  March 24, 1969.  p. 114.

5.  Control Techniques for Hydrocarbon and Organic Solvent
    Emissions from  Stationary Sources.  Publication No. AP-6B.
    U.S. Dept. H.E.W., Public Health Service, Washington, D.C.
    March 1970.

6.  Robert  Snow Means  Company,  Inc.  Building Construction  Cost
    Data, 1977.   Duxbury,  Massachusetts.   1977.

7  Guthrie, K.M.   Process Plant  Estimating  Evaluation and
    Control.  Craftsman Book  Co.,  Solano Beach,  California.
    1974.

8  Blecker, H.C.,  et. al.  Wastewater Equipment.   Chemical
    Engineering  Deskbook  Issue.   October 21,  1974.  p.  115.

9  Chemical Engineering  Equipment Buyers  Guide,  1977-78.
    Chemical Engineering  Magazine.  McGraw-Hill,  New  York,  New
    York.   July  1977.

10.  Perry,  J.H.,  ed.  Chemical  Engineer's  Handbook,  4th edition.
    McGraw-Hill,  New York, New  York.   1963.

11   Control Techniques for Particulate Air Pollutants.  Publi-
     cation No.  AP-51.   Office of Air  Programs,  Environmental
     Protection Agency, Research Triangle  Park,  North Carolina.
     January 1969.
                               4-24

-------
12.   Memo from Bill Hamilton of Economic Analysis Branch,  U.S.
     Environmental Protection Agency,  Research Triangle Park,
     North Carolina, to staff.  Subject:  EAP Input to RACT/CTD
     Documents.  March 14,  1977.

13.   Lund, H.F.,  ed.  Industrial Pollution Control Handbook.
     McGraw-Hill  Book Co.,  New York City.  1971.  Chapter 26,  p,
     7-46.
                                4-25

-------

-------
                            SECTION 5
           FINANCIAL AND ECONOMIC ANALYSIS TECHNIQUES

5.1  INTRODUCTION
     This section of the guidance document is intended as general
reference for state  and local  agencies,  to assist them in deter-
mining  the  economic impact  upon major  stationary  sources  that
may result from application of pollution controls.
     The costs incurred by  the source as a result of control are
assumed to be directly obtainable from the source or obtained by
the procedures discussed in Section 4.
     Subject to this assumption,  generalized methodology will be
described for evaluating the  economic impact of these costs upon
the source  (i.e.,  the  individual plant)  that is being subjected
to the control requirements.
     Six techniques  for performing relevant  economic and finan-
cial  analysis  will  be presented,  with  comments on the factors
that might affect  the  selection of one or another of these tech-
niques  (or  combination of  techniques)   for  a  given situation.
These  factors  influencing  selection  of  an  appropriate analysis
technique include  the  impact  to be measured, the available data,
the personnel and other resources available to the state or local
agency  for  processing the  data,  and  interaction  with certain
special  incentive  arrangements that  are  available  for financial
needs arising from pollution control.
     In  conjunction  with the  description of each analysis tech-
nique,  the data  that must be  available with  respect to the par-
ticular source in order to make proper use of that technique will
be  defined.   This  is  a "feedback"  situation:  initial selection
of  a  particular analysis technique  will define the particular
                               5-1

-------
data needed.  This will lead to a determination of the availabil-
ity of these  data,  and that in turn may necessitate  a  change in
the selection of  the  analysis  technique.   An important objective
in  what  follows,   therefore,  is  to  clarify the  relationships
between the various analysis techniques and their respective data
requirements.
     In addition to the foregoing, consideration will be given to
whether  the  capital  needed  for compliance  is  available to  a
particular  source.  This is only one facet  of the various poten-
tial economic impacts of the costs of compliance.  However, it is
sufficiently fundamental to call for special highlighting.
     The  intent  here is not  to  provide a  textbook on economics
and finance,  to  address a  readership of financial experts, or to
assume the  place of  a  qualified  analyst.   Rather,  the objective
is  to provide  uncomplicated,  straightforward  explanations  (in
sufficient detail that  the information can serve as the framework
on  which to  construct practical  applications)  for  readers  not
having specific background in finance.
     The  discussion  is  presented in two parts:   description of
the  analysis  techniques (5.2) and  information needed to perform
analysis  techniques  (5.3).

5.2  THE  ANALYSIS TECHNIQUES
     Six  techniques  for  analysis of the economic impact of pollu-
tion  control  upon  individual  plants  are  presented  below  and
detailed  in Table 5-1:

     1.   Computation of Debt Service Coverage Ratio
     2.   Computation of Return on  Investment (ROI) Ratio
     3.   Computation of Discounted Cash Flow
     4.   Reference  to  Previous  Impact Studies
     5.   Reference  to  Financing Decisions by Others
     6.   Computation of  Increase  in  Operating Costs  or Assets
                                5-2

-------
                                Table 5-1.  COMPARISON OF ANALYSIS  TECHNIQUES



Techniques
Debt Service Covei ngo
Ratio








Return on Jn vestment
(ROD Ratio









Discounted Cash Flow
(t)CF) Computation






Impact
Being
Analyzed
Capital Availability









Prof i tabili ty










Profitability






Availability of Data


Sources (Prioritized)
Plant
Other Studies
FPA
OSHA
References
Census of Manufacturers
Robert Morris Associates
Ti oy Almanac
Trade Associations
Federal Trade Commission
Per i odicals
Plant
Other Studies
F,PA
OSHA
References
Robert Morris Associates
Troy Almanac
Federal Trade Commission
Equipment Suppliers
Census of Manufacturers
Trade Associations
Peri odical s
PI ant








Availability
Existing Source
- Moderately Easy
New Source
- Difficult






Exi sti ng Source
- Moderately Easy
New Source
- Difficult







Existing Source
•- Moderately
Di Ff icu 1 t
New Source
- Difficult


Agency Level
" Of Effort "~
Required
Moder ate

Ex tens i ve
Forecasting
Required





Moderate

Extensive
Forecasting
Required






Extensive
Forecasting
Required
Extensive
Forecasting
Required




Limitations
1) Veriti cat inn of
forecasts Cor new
or expanded source.
2) Considerable
error possible if
measuring at: arm's
length.



1) Verification of
forecasts for new
or expanded source.
2) Indirectness of
measure .






1) Verification of
forecasts for all
sources .
2) Considerable
error possible if
measuring at arm's
1 ength.
(Jl
I
L/J
      (continued)

-------
                                          Table 5-1  (continued).
Techniques 	
Heliance Upon Previous
Impact Studies









Financing Decision By
Others



rncrcase in Operating
Cor. ts or Assets








Impact
Doing
	 Ana ly zed
Whatever Impacts
Studied Previously









Capital Availability


Availability of Data
Sources (Prioritized)
EFA: Rconomic Analysis
Branches of offices of
Air Quality Planning &
Standards (Durham, NC)
and Office of Water
Programs (Washington,- IT)
OSHA







Plant
Dun and Bradstreet
Local Industrial Directory
SBA (Size Standard)



Prof i tability








Plant








Availability
Relatively Easy









Moderately Easy
If Working With
Source
Difficult If Doing
Arm's Length Analysi



Existing Source
- Moderately
Difficult

New Source
- Difficult





Agency Level
Of Effort
Required
Minimum









Minimum
-S



Moderate

Extensive
Forecasting
Required




Limitations
1) Control costs
magnitude may be
so different that
makinq judgments
may be difficult.
2^ If control costs
being analyzed are
additive to ones
studied (rather
than substituted)
then judgments may
be dif f icu 1 t .
3) Industry may not
have been studied.
1) If analysis IR
performed without
having source apply
then it is difficult
to analyze existing
source and very di f-
ficult to analyze
new or expanded
source .
1) Verification of
forecasts for now
sources .

2) Accuracy of
allocation of
overhead expenses
from another part
of company.
3) Indirectness of
measure.
Ui
I

-------
     This  list does not  represent  an exhaustive set of analysis
techniques.  There  are others,  but these six were chosen because
they cover a variety of possible circumstances.   The first three
listed techniques are  conventional tools for analysis of business
financial  conditions.   The next  two  techniques are  not  so con-
ventional, but are particularly  applicable  to pollution control
impact financial conditions.  The sixth technique was created for
this guidance  and is  intended  for  special  circumstances  when it
is impractical to establish net income figures  for a plant.
     In  all  these techniques it  is most important  to determine
when the impact crosses  the  threshold from minimal  to signifi-
cant.  The  definition given  to  "significant"   is that  the plant
will close  because it  is no longer  able to  conduct business or
because  continuing  operation  is  not  as  profitable  as  other
                                                                  i
options.  In several of these techniques the criteria for signif-i
icant  impact  are well  defined,  e.g.,  necessary funds  cannot be
obtained  for  the  pollution control  expenditure,  or  the  owners
would be better off liquidating the plant and investing the money
elsewhere  rather  than buying  pollution  control equipment.   In
other techniques (including some of the conventional ones), there
are no  widely recognized  specific  numerical  criteria for deter-
mining significant impacts, for which the analyst may be required
to use  extensive  judgment.   Some  judgments  will be  relatively
simple  and others will  be difficult.  Further,  some techniques
are more direct than others in measuring significant impact.  For
example,  although  the  ROI and discounted  cash flow  techniques
both measure profitability, ROI does not necessarily indicate the
precise  level  at which the plant would choose  to  close,  whereas
the discounted  cash flow  technique  does  indicate such a  level.
     Before the  six  techniques  are  described,  a definition  of
individual plant source  is  provided.   A  plant is  the  single
physical location where  the  pollution control  equipment is to be
                               5-5

-------
installed.   It is analyzed on  a  stand-alone basis without regard
to other operations of  the company at other locations.   Analysis
is relatively  straightforward  when a plant  is the  whole company
or a separate subsidiary of a larger company.  Analysis  may still
be  fairly  straightforward  when  the  plant  is a  division or  a
specific profit  center  of  the  company.   Analysis is  much more
complicated when  the plant  is part of  a  division and  buys  and
sells material  from and to other plants or  divisions within the
corporation.   The main problems in the latter case are validating
company-submitted data  that  have not  been independently audited
or are  not publicly available and  assigning a value  or transfer
price to  the intracompany goods  in order to  derive  net income.
     In  the  subsections  that  follow, the  individual  techniques
are  first  briefly discussed.  Following that  is  a comparison of
the  techniques that  will aid an analyst in  selection of tech-
niques for a given situation.

5.2.1  Analysis Technique 1;  Computation of Debt Service
       Coverage Ratio
     The debt  service coverage technique involves  an analysis of
whether or not a source  can meet  debt  (i.e., interest and princi-
pal),  lease,  and long-term  payment obligations,  including those
for  the new pollution control expenditure,  at the level of cash
flow that  exists at the  source.   It  is  therefore also  a measure
of the  capital availability  impact  of  a pollution control regula-
tion.
     The  lending community  looks  to  this  ratio  in longer-term
borrowing  situations  to determine whether adequate funds will be
available  to repay  debt  obligations.
     Cash  flow  from  the operation  includes net  profit  before
taxes,  lease,   and  interest payments  plus  depreciation.   It is
only out of this  cash  flow that payments  can be  made for  such
obligations  unless  the company has "unproductive"  assets  that it
                                5-6

-------
could convert  to  cash to buy  control  equipment.   "Unproductive"
designates assets  unrelated to the  nature  of the business,  for
example,  long-term  marketable  securities  or  certificates  of
deposit.
     This technique can be applied as a verification of a financ-
ing inability  determined  under Technique  5 or as a prediction of
whether financing may be obtainable.  It can be used for a new or
an existing plant.   It  is used primarily when the agency and the
source  are  engaged in deliberations, since  the  level  of obliga-
tions that are a crucial part of the analysis is highly dependent
upon the particular circumstances of the plant.
     The results of  the  debt service coverage analysis will be a
ratio,  the  numerator of  which is cash flow from operations and
the  denominator of  which  is   future  debt  and  other  long-term
obligations.   The  ratio needs  to  be computed on a  "before" and
"after" pollution control basis.  In order for a  financing source
to loan money to  a  plant,  the  "after" ratio should exceed 1.0.
Most  financing sources,  however, would require  an ample cushion
greater than 1.0 to reflect risk over time.   If the "after" ratio
exceeds  2.0,   it  can  be assumed that there is  no  significant
impact.   If  the ratio  is between 1.5 and  2.0,  chances are that
the  financing  is  still  obtainable  but the  company  is  in a risk
situation unless it  is  accustomed to operating at that ratio.  A
ratio  between 1.0  and  1.5, for  the purposes  of this analysis,
will be considered to indicate  significant impact.
     The  banking  industry,  although using  this technique exten-
sively,  does  not  specify a criterion to  be  used in making judg-
ments.   For  example,  a bank may  grant financing in  a case where
the  ratio is  less than  1.5  because the  company has historically
operated  at that level and has  been  able  to  meet  its obligations.
Thus  it is  important to do both before and  after analyses of the
ratio  to indicate the  extent  of change  of  the  ratio as well as
the  absolute level after  pollution  control impacts.
                               5-7

-------
     In this technique and  several  others,  the analyst must con-
sider the extent to which a price increase may occur to cover the
pollution control expenditures.  This  consideration is discussed
here  and  is not  repeated  in  the  next three  technique descrip-
tions.  The  determination  of  impact from any  pollution control
situation  is highly  sensitive  to   whether  a  plant  source  can
increase the price  of its products  and therefore pass on to the
consumers  the  costs  of  pollution  control.   To the  extent that
this can occur,  the impact of pollution controls does not have to
be absorbed by the company.
     The  increase  of price,  however,  can  sometimes lead  to  a
slight  reduction  in  total  products purchased  from the company.
Therefore, although the  revenue of  the company goes up to cover
the  pollution  control costs,  it can come back down slightly if
fewer  products  of  the  company  are  purchased.   The quantitative
measure of this effect of a change in quantity purchased due to a
change  in price  is  referred  to as elasticity.  It is typical,
however, that the quantitative determination of elasticity is not
available  in an analysis.  Therefore the  impact over a range of
possible  price  behaviors  can  be approximated  by performing the
analysis as  if  there were no price  increase and as if there were
a  full  price increase without any change in quantity  of products
sold.   (The analytical  tables  presented  later do  include pro-
visions  for  cases  in which an  elasticity  figure  is  available.)
Then the  analyst  must use  judgment to  determine  where  in the
range the  impact is  likely  to  occur.  Such a judgment  can be made
only if  the analyst can  determine at least  qualitatively the
possible price  increase that the company would  be able  to set for
its  products.   The  analyst should  consult economic textbooks to
determine  how to assess elasticity.
      Performing this analysis  for   an existing plant requires a
moderate amount of  effort.  Performing this  analysis  for  a new or
expanded  plant requires  a  considerable amount of  forecasting by
the  source,  not  only of profit and loss  but also  of  equity and
debt financing  practices.

                                5-8

-------
5.2.2  Analysis Technique 2:  Computation of Return on
       Investment (ROI) Ratio
     ROI is a measure of how efficiently the investments (assets)
of a  plant are being  utilized and thus is  a  measure of profit-
ability.   It  is not a  direct measure of  capital  availability.
     The ROI  measure is a ratio:  the numerator  is  net profits,
and the  denominator is  the  assets value of  the  plant including
net working capital and the net book value  of fixed assets.   No
single criterion is recognized for determining when the ROI ratio
of a  plant is  significantly impacted.   The  analyst,  therefore,
may perform  "before" and "after"  ROI  impact evaluations,  taking
into consideration a possible price increase.
     The preferred ratio with which to compare the "after" ROI is
the plant's  threshold rate.   By  threshold  is meant  the plant's
minimum acceptable  rate  of  profit per each new investment.  Most
plants have  such a  threshold  ROI rate,  which may or may not be
obtainable.  The analyst should remember that the minimum thresh-
old rate for  a  new investment may be higher than the average ROI
for existing investments.
     If the threshold  rate  is not obtainable, the analyst may be
able  to  obtain the industry average ROI  for comparison with the
plant's  "after"  ROI.  Sources of an industry average  ROI value
include trade associations,  Robert Morris Associates, the Federal
Trade  Commission,   and  previous  industry  impact studies.   The
analyst  must  use judgment  to  determine  how much deviation from
the industry average constitutes a significant impact.
     If the above two rates are not obtainable, the analyst could
compare  the  plant  ROI  with the  marginal cost of capital, i.e.,
the plant's cost of the  next source of funds for  its investments.
Significant  impact would occur  when  the  ROI  is less  than  the
marginal cost of capital.
                               5-9

-------
     This technique can be applied  to  new or existing plants and
to plants with which an agency is in deliberation or is analyzing
at arm's length,  i.e.,  predicting impact without being engaged in
deliberations.
     In  analysis  of a new plant the future  financial conditions
must be  estimated in order  to  perform the  ROI  calculation.   It
will be necessary for the source to  forecast at least the first 2
or 3 years  of operation  so  that the ROI impact  for  more  than 1
year can be  viewed.  This  measurement  requires  a  moderate  to
extensive level of effort for the agency analyst.

5.2.3  Analysis Technique 3:   Computation of Discounted Cash Flow
       (DCF)*
     The discounted  cash flow technique  involves  looking  at the
future   operations of a  plant to determine  whether  it  would be
better   for  the plant owners  to  liquidate assets  or  to continue
operating  the plant  with the  pollution  control  equipment  in-
stalled.  The  discounting refers  to a  consideration of the value
of money over  time.   This technique is a direct measure of prof-
itability.
     The technique  is  applicable to analysis  of new  or existing
plants   and  is  more  applicable to plants with which the agency is
in  deliberations  than to those being analyzed  at arm's  length
without benefit of the plant's financial data.
     The  general  procedure  for  performing  this analysis  is to
predict the future financial conditions of the company during the
useful   life of the pollution control equipment.  The  value of
future  net  cash flows from the plant  operations plus the future
terminal value of the plant (appropriately discounted) at the end
of  the  useful life  of  the  pollution  control  equipment are com-
puted to obtain present value.

* Reference for this discussion and for several of  the tabular
  formats presented later was Appendix A:  Draft Guidance on
  Nonferrous Smelter Orders Under Section 119 of the  Clean Air
  Act,  EPA Division of Stationary Source Enforcement.
                               5-10

-------
     The value of  future  cash flows must then be compared to the
present value of the plant if it were immediately liquidated.  If
the value of future operations is greater than the present value,
the  firm,  according  to  rational  decision-making  theory,  would
purchase the pollution control equipment.  If the future value is
less than  the  present value,  the  plant would  be liquidated and
the money invested elsewhere at a greater return.
     The  discount  rate  is  of particular  significance  to  this
technique.  The discount rate used in this analysis is a weighted
average cost of  capital.   Capital  is defined as sources of long-
term funds for the company and typically includes long-term debt,
preferred  stock,  common  stock,   and  earnings  retained  in the
business.  Most  going  concerns  are  already earning money  at a
rate greater than their cost of capital.  This technique is based
on the premise that when the plant earns less than what the  funds
cost, the plant will choose to close.
     The discounted cash flow technique requires forecasts in all
situations.  One  of the difficult aspects of forecasting  is that
projections  of greater than  2  or 3  years  are often inaccurate.
In addition, because  forecasts are not subject to  an independent
audit  and because they are supplied  by the  company,  the  analyst
must verify  or test  the  reasonableness of  the forecast.   It is
therefore  appropriate  for an analyst to request that the  company
not only make  a  financial  forecast but  also  describe the physical
inputs  that  went  into  it.   The physical  inputs consist of the
amounts  of  raw  materials, labor,  and other  supplies  needed to
manufacture  the  product.   The analyst then  must  obtain  (from the
source  or others) price  factors to use with the physical inputs
in  deriving   financial  data  that  can be  compared  with  those
supplied  by  the  plant.
     Since  forecasts  are  realistic  only  for periods  up  to  3
years,  the analyst must extrapolate net cash flow for the  remain-
ing  years to the end of  the  useful life of  the pollution control
equipment.
                                5-11

-------
     The analysis  of discounted  cash flow should  be done on  a
"before" and "after" pollution control basis  and with or without
a price increase to cover the pollution control costs.
     This particular method is the most highly labor-intensive of
those  presented.   The  extrapolations  require  a  considerable
amount  of  time, as  does determination  of  the proper  labor  and
cost factors for the physical inputs.

5.2.4  Analysis Technique 4:  Reference to Previous Impact
       Studies
     Often an  analyst  can  obtain  an EPA economic impact study of
the  entire  industry of which the plant  under consideration is a
part.   Although such  a  study  may  pertain  to  compliance  with
pollution control  regulations other  than the  one in question, it
may provide applicable information on pollution control costs and
economic impacts.
     Economic  impact analyses have  been  done  for New Source Per-
formance Standards (NSPS) and NESHAPS regulations by the Economic
Analysis Branch of the Office of Air Quality Planning and Stand-
ards.   To  date  approximately two  dozen such  studies  have been
performed.   These  studies  are  available  through  the  Economic
Analysis Branch  Office in Durham, North Carolina.  Many addition-
al  industry impact  studies relating to  water  pollution control
regulations  have been performed  by  the  Economic Analysis Branch
EPA  Office  of Water Programs, Washington, D.C.   A state or local
agency  analyst may wish to  contact these two branches of EPA to
obtain  a list  of available  studies.
     Application of  an earlier study  is most  straightforward when
the  control cost  magnitudes are similar,  the regulatory impact
under  consideration is  not in  addition  to  the regulatory impact
of  the  other  study,  and  the other  study reaches  a definitive
impact  decision.
      If the control costs  are similar in magnitude  and  the other
study  concludes that a  significant impact will occur, then it is
                                5-12

-------
likely that  such a conclusion also holds  for  the present situa-
tion.  When  the  conclusion  is  that  no significant  impact will
occur, the  analyst must  use  judgment  to  determine how  far the
control costs  of the other  study  could expand while maintaining
the  same  conclusion.    The  expansionary consideration  can cover
circumstances  where  the control  costs  of the  present  situation
are  greater  than  those previously studied  or when  the  present
regulatory impact  is  in addition to the previous one.  No clear-
cut  criteria can be  presented here to  aid the  analyst  in calcu-
lating the  expansionary limits.   The judgment depends upon the
particular facts of the previous study.
     To  the  extent   that  another  impact  study  is  useful  for
gauging  the  impact  of  the  situation  under  consideration,  the
amount of resources required of an agency is minimal.

5.2.5  Analysis Technique 5;   Reference to Decisions by Others
     One impact  of pollution control  regulations to consider for
a  source  is  the  availability  of capital to purchase  the equip-
ment.  The inability  for  a  plant source to obtain needed financ-
ing  capital is  indicative of significant impact, since the plant
may not be able to continue  to conduct business.
     For a small business as  defined by the  Small Business Ad-
ministration (SBA), the skill  of SBA  can be utilized to evaluate
the  capital  availability impact of a pollution control expendi-
ture  by  having  a source  make  application.   If  the  financing
request is rejected  by the  SBA,  sometimes called  lender of the
last  resort, then the source  is unlikely to obtain financing by
any  conventional  means.  SBA has specific  programs for pollution
control  expenditures;   currently  these  programs  are  adequately
funded.
     Although  no  similar  institution   is  applicable  to  large
businesses,  rejections by two or more  financing sources provide a
similar  indication  of  the  unavailability  of  financing.   To
attribute financing inability to the pollution control regulation
                               5-13

-------
in  question,  the  analyst  must  know  that disapproval  of  the
financing  request  was  caused by  the impact  of the  regulation
itself and  not  other factors such  as improper  documentation  or
lack of management expertise.
     This technique  is  particularly relevant when the  agency  is
engaged  in  deliberations  with the source and can ask  the opera-
tions  to apply  for  financing  (the  debt service coverage  ratio
allows  the analyst  to estimate  the  availability  of  financing
without  the source having to  apply).   This technique can be used
for both new and existing sources.  It calls for a relatively low
level  of  effort,  since  it  enlists  the  skills of an existing
institution set up to do this type of work.

5.2.6  Analysis Technique 6;  Computation of Increase in
       Operating Expenses or Assets
     In  some  multiplant  situations  the  net  income for  a plant
will  be either  extremely difficult  to  determine or  so artifi-
cially  determined that  it  does  not  represent the true profit
picture  of that company  if  it  were  independent.  The  importance
of  determining  income  is  that  all of the  above-presented tech-
niques  require  it  for  the  analysis.   Where  income   cannot  be
determined  with  an acceptable level of accuracy, the analyst may
wish to  examine  the  operating expenses or asset  replacement value
of  the plant  and to compare  them to  the  magnitude  of  the pollu-
tion control costs as an  indirect evaluation technique.
     This  technique is applicable to  new and  existing sources
with which the agency is  in deliberations or those being  analyzed
at  arm's length.   This  technique  is an indirect  measure of  impact
because  it provides no  definitive indication  that the  increase
involved would  cause  the  plant  to liquidate  or that the plant
could  not  obtain  financing.  As  with ROI,  there are no  criteria
for determining  when an  increase in  cost constitutes  significant
recognized economic impact.  A level  of 30 to  40 percent  (after
                                5-14

-------
price  increase considerations)  is  a reasonable criterion to  use

when related  to expenses  and  greater than 40  to  50 percent of  the
asset  replacement value.

     This  analysis  technique  requires a relatively moderate level
of effort.


5.2.7  Comparison of the  Analysis Techniques

     The  six  techniques  are  compared in the accompanying matrix

according  to  salient  characteristics.  The matrix describes what

each  technique measures,  the data  required for  performing   the

technique,  the sources  of such  data and  the   relative  ease  of

acquisition,  the amount of effort required by the agency, and  the
limitations of the  technique.

     Examination of the  matrix reveals several  pertinent points:


     0     For  analysis  of sources at  arm's  length,  the two most
           appropriate  methods  are  use  of  the ROI   ratio   and
           reliance  upon previous impact studies.

     0     The  discounted cash  flow  technique requires the  most
           work on  the  part of the agency,  followed  by ROI,  debt
           service  coverage,   and  increase  in operating costs  or
           assets.

     0     Forecasting is required for all analyses of new sources
           except those based  on  decisions  by others and previous
           impact studies.

     0     Financing decisions by others,  debt  service coverage,
           and perhaps  reliance upon previous  impact studies   are
           the techniques  for  directly determining capital avail-
           ability.

     0    Discounted cash flow  is  the  most  direct  measure  of
          potential for  plant closure due  to low profitability,
           followed by the indirect measure of ROI and increase in
          operating costs or assets.

     It is emphasized that  on some occasions the agency analysts
will not  be  able  to  perform  an  analysis,  or judgments  will  be

very difficult.   These  techniques  will accommodate most situa-

tions,  however.   In the  event  that  an analyst  needs  advice,  he
                               5-15

-------
should contact  EPA Division of  Stationary  Source  Enforcement in
Washington or the Economic Analysis Branch in Durham.

5.3  INFORMATION NEEDED TO PERFORM ANALYSIS TECHNIQUES
     This section  describes more  fully  the information required
for  performing  each  technique  and  the  alternative  sources  for
obtaining the necessary  information.   Sample  tables performing
the  necessary  calculations are  presented  for each technique.
     Prior  to a description of  the information needed  for each
technique,  the  subject  of  special  Federal  pollution  control
financial assistance programs will be  addressed.
     In  some cases  the benefits  of  these  programs need  to be
woven into  the calculations of the various  techniques.

5.3.1  Special Pollution Control Financial  Assistance Programs
     The  Federal  government operates four major Federal programs
to  ease  the  financial  burden  of pollution control  expenditures:

     0    Industrial Development Bonds
     0    SBA Direct Compliance  Loans
     0    SBA Lease Guarantees
     0    Rapid Amortization

     Two  of these programs are particularly  important to sources
that have  difficulty  obtaining  financing  for pollution control
and are provided  by  SBA.
     The  other two programs  reduce pollution control costs;  one
makes   available   low-cost  financing  to   companies  considered
credit-worthy;  the other provides  for rapid recovery of portions
of the pollution  control  costs through tax  reduction provisions
of the  Federal  corporate  income  tax laws.
     This section will briefly describe  the four programs,  focus-
 ing on  eligible users,  benefits,  and access to the program.
                                5-16

-------
Where  it  is expected  that a source  would use any  program,  the
pertinent information on interest rates and terms is provided for
entry  to  the  tables  pertaining  to  the  various  analysis  tech-
niques.
     Additional  information  on these four programs  can be  ob-
tained  from  reference:   Choosing Optimum Financial Strategies:
Pollution Control Systems,  JACA  Corp.,   Environmental  Research
Information Center, U.S. EPA, Cincinnati,  Ohio.

5.3.1.1  Industrial Development Bonds—
     Industrial Development Bonds (IDE's)  are a type of municipal
bond with  lower interest rates for the company  than a corporate
bond and  a repayment  period  significantly longer than  that  ob-
tainable  from  a bank.   The  reason for lower costs is  that  the
interest is tax-free to the bond buyers,  which makes them willing
to accept interest  rates  that are two to three points lower than
those  on  taxable corporate bonds.  IDE's  are  issued by a public
entity  on behalf of  an industrial source  and are  issued  for  a
public  purpose.   Pollution   control  equipment  is  considered  a
public purpose and is therefore eligible for this type of financ-
ing.
     The  amount  of  an  issue  is usually  quite  large,  i.e.,
$500,000 and greater.  The reasons for such large amounts are the
high  fixed costs  of  obtaining such  bonds,  due  to advertising
expenditures,   bond counsels,  investment  bankers,  and printing.
The  interest  savings  accruing  to  a  company  from use  of  these
bonds  must  therefore be of  sufficient magnitude  to surpass such
fixed costs.
     The  bonds  are also referred to  as revenue  bonds  as opposed
to general  obligation bonds, which  means  that only  the revenue
from  the  company  is  used to repay  the  bonds.   Therefore,  the
benefiting  company  must  be  credit-worthy.   Also,  because  the
bonds are often sold nationwide, the reputation of the benefiting
company must be fairly high.  All of  these  factors  point to the
                               5-17

-------
fact that  only  financially healthy companies and companies  with
large pollution control expenditures  can make use  of the benefits
of this financing program.
     Although small  companies  generally are not able  to benefit
from this  program,  companies with good credit  position  and ex-
penditure needs of  less  than $500,000  can obtain  the benefits of
the program  in  some states.   This  occurs when the  state allows
the private  placement  of bonds whereby the bonds are  sold  to a
single  entity,  such  as a bank  or an insurance company.   The
effect  of  this  mechanism is  to  lower   the  amount of fixed costs
incurred and  therefore permit pollution  control  expenditures in
amounts  of less  than  $500,000  to be  beneficially financed by
industrial development bonds.
     The determination of whether a source is likely to use IDB's
can be based on 1) an indication from operators  of the source, 2)
previous use  of IDB's  by the source plus  an expenditure greater
than $500,000,  or 3) an expenditure  greater than  $500,000 plus a
good credit or bond rating.  In the latter case, if the company's
rating  is  not  readily available  through  Dun and  Bradstreet,
Moody1s,  or  Standard  &  Poor's,  then  for  simplification  the
analyst may assume that an IDE will not be used.
     When  it  is likely  that  an  IDE will be  used,  the values of
interest rate and repayment  length to  be entered  into Table 5-2,
line 2(a)  are the  current municipal  A-rated bond  yield* and 25
years.    Because the choice of an A-rating  and  25 years is some-
what arbitrary,  these  values  should not be used if the analyst
knows  a more  exact bond  rating  and  repayment  period  for the
company.

5.3.1.2  SBA Direct Compliance Loans—
     The SBA makes loans directly (without having a bank also put
up  money)  to "small businesses"  for pollution control expendi-
tures.    The  small business must fit the numerical  definition of

* Economic Indicators,  Council of Economic Advisors or Survey
  of Current Business,  U.S. Department  of Commerce.

                               5-18

-------
         Table  5-2.    POLLUTION CONTROL  COSTS
 1.   Capital costs including installation
 2.   Interest Rate (i)

     a.   When anticipating impact or
         specific financing source
         cannot be determined

     b.   If IDE rate  possible

     c.   If SBA rate  possible
                                                      10
 3.   Tax depreciable life (n)
                                                                   years
10.
     Annualized costs (capital
         recovery factor equation
         times capital costs)
         Line 1 x
              id
                         i)
                  (1 + i
                         n
                           - 1
     Average annual depreciation expense

                   1
                   n
                                        (1)
    Line 1 x
 6.   Average annual interest expense

         (Line 4  -  Line 5)

 7.   Annual O&M costs,  including
         insurance, & G&A<2'and
         excluding  depreciation
         and interest

 8.   Average annual by-product value

 9.
Annual control costs
    (Lines 4 + 7 - line 8)

Investment tax credit
    (.1 x Line 1)
 (1)   Straight line depreciation is  assumed  for  purposes  of
      simplifying the calculations
 (2)   Assume 4% of capital costs is  insurance  and G&A if  no
      other figures readily available.
                               5-19

-------
small as  established by SBA, face  a  compliance  requirement that
can be  traced  to  a Federal requirement, and be  unable  to obtain
the loan  or able to  obtain it only  at unreasonable  terms.  The
numerical  qualification is  usually  in terms of  number  of  em-
ployees, ranging between 250  and  1500 depending on the industry.
By a telephone call to SBA, the analyst should obtain the specif-
ic  numerical   limit  for the  Standard  Industrial  Classification
Code  (SIC)  applicable to  the plant's line of business.   Deter-
mination of size  for SBA  qualifications is based  on the entire
company and not on only the plant being analyzed.
     This SBA program :.s not to be confused with SBA's tradition-
al loan programs for businesses that often have limited funds and
often only guarantee bank loans having high interest rates.
     Presently, the  SBA interest  rate on direct compliance loans
is 6-5/8 percent.   Loans  are allowed to 30 years,  although typi-
cally they  are granted for  a shorter  period  that is reasoncible
for the company to repay.  These loans are available for purchase
of  add-on  pollution control equipment  as  well  as  process change
equipment.
     To  obtain an SBA  loan  the company should  first be refused
financing by  a conventional  source  or offered  financing  at un-
reasonable  terms.  The  source should then determine from SBA the
numerical employee  limits  for their  SIC;  if  the number employed
is under  that  limit,  the source may apply.  Documentation of the
bank  situation will be required,  along with a  statement by the
control  agency of  the  adequacy  and  necessity  of  the proposed
control  equipment.   The application process  can  be very time-
consuming and  may require  several  exchanges  of  information with
the SBA.
      If  the analyst  can determine by calling SBA that the source
is  small,  and if  there is some  indication that the source will
have  difficulty  obtaining  conventional  financing,  then the ana-
lyst  should use  the current  SBA terms  as  data  for entry into
                               5-20

-------
Table 5-2.   The  current SBA interest  rate  (obtainable from SBA)
should be  entered,  and the  repayment  life should  be  assumed to
equal the equipment's useful life.

5.3.1.3  SBA Lease Guarantees—
     Some small businesses will also be eligible for conducting a
lease of pollution  control  equipment under  terms the  SBA will
guarantee.   In the  pollution  control field,  a  lease typically
occurs  with  Industrial Development  Bond  (IDE)  financing,  ex-
plained  above.  Until  this  program began,  small businesses could
not  "afford"  IDE's  because of the high  fixed  costs of obtaining
them.  This  program therefore groups  the needs  of  several small
businesses  under  an IDB  issue  guaranteed 100  percent by SBA.
     Interest  rates  and repayment terms are similar to those of
IDE's.   The  eligibility criteria  for  this  program  are different
from  those  criteria of  a  direct compliance  loan program.   A
company is eligible for this program if:

     together  with  its affiliates,  (it) is  independently owned
     and  operated,  is  not dominant in  its field  of  operation,
     does not have assets exceeding $9 million, does not have net
     worth in  excess of $4 million,  and does not have an average
     net  income,  after Federal  income taxes,  for  the preceding
     two  years  in excess  of $400,000  (average net income to be
     computed without benefit of any carryover loss).*
     To participate  in this program the small business must first
obtain the  sponsorship  of a bank.  The bank prepares a statement
of opinion  that the company is credit-worthy.  Then the require-
ments of the SBA  direct compliance loan program apply, i.e., the
company  must obtain a  certificate from  the control  agency that
the  equipment is  adequate and  necessary,  and must  complete an
application.
     If  there  is  some indication  that this program will be used
by  a source,  then   the interest  rates  and repayment  period as
* Part 121.3-11(a), Chapter 1, Title 13.
                               5-21

-------
determined under  SBA's compliance  loan program  will provide  a
sufficient degree of accuracy for analysis.

5.3.1.4  Rapid Amortization—
     This is  the  only  one of the four  Federal programs  that in-
volves tax deductions  for pollution control  equipment.   Although
it is a  special  program for purchase of pollution control equip-
ment, very few companies  use it,  and those that  do  use  it do so
under certain circumstances explained below.
     The  rapid  amortization  program  allows  expenditures  for
pollution control equipment to be deducted for Federal income tax
purposes  over a  60-month period, which is usually  shorter than
the  normal  tax-depreciable  life.   Along  with this  accelerated
write-off,  the  source  is  entitled to  one-half  the normal  10
percent investment tax credit, or 5 percent.
     To  determine  whether rapid  amortization  is  beneficial, the
analyst must  compare,  for example,  using net present value  (NPV)
techniques,  the value of 60 months'  depreciation plus a 5 percent
investment tax credit versus the normal tax life of the equipment
and a 10 percent investment tax credit.
     The  results of such a  comparison are presented  in Figure
5-1. A  company with equipment of 15 years taxable life  and a 10
percent discount  rate  would prefer  the rapid amortization,  since
the point at  which those two figures meet on the graph is to the
right of  the  line.   If the equipment had a 10-year life and a 10
percent  discount rate,  however,  a  normal  accelerated rate would
be preferred.  The  special circumstances under which rapid  amor-
tization  is  preferred,  then, are a  relatively  long  (12  years or
more) normal  useful life of the equipment or a high discount rate
coupled  with  equipment that  has  a  slightly  shorter  useful  life.
     To  simplify the impact  analysis,  it  is  suggested that this
program  not  be considered as beneficial to  the  source.   Even if
the  source  does  use rapid  amortization,  the difference  in the
                               5-22

-------
Ul
I
to
U)
                  20
               f-  IS _
               a:
               UJ
 1
               w
               H
               c<  10
o
u>
00
t—<
Q
                      TRADITIONAL DEPRECIATION
                            YIELDS GREATER
                            TAX BENEFITS
                                                                RAPID AMORTIZATION
                                                                  YIELDS  GREATER
                                                                   TAX BENEFITS
                                         10          1 b         20

                                                ASSET LIFE - YEARS
                    Source:  Economic Analysis Division,  Office of Planning and Evaluation, Environmental
                             Protection Agency.
        Figure 5-1.   Indifference  curve for  amortization options  under the  Tax Reform Act of
                           1976 traditional vs.  rapid amortization.

-------
results will  be  minor.   The analytical tables  are  formulated to
reflect  normal  taxable  useful  lives  of  equipment  and the  10
percent investment tax credit.

5.3.2  Information Required
5.3.2.1  Pollution Control Costs—
     In most impact analysis techniques described here, the
pollution control  costs  must be numerically imposed on financial
information  supplied  by  or  constructed  for  the   source.   The
following  format  for control  costs,  if completed,  will suffice
for  all  techniques that require this information.   (As mentioned
earlier, it is assumed that  control costs are available.)

     0    Capital  costs  including installation
     0    Yearly interest  rate  on those capital costs
     0    Depreciation period  in years (taxable life)
     0    Annual   operating  and  maintenance  costs,   including
          insurance,  general  and administrative,  but  excluding
          interest and depreciation
     0    Byproduct recovery yearly value
     0    Investment  tax credit
     Values  for  the above  cost items should be  entered into  Table
 5-2  by either the source operator or  the control  agency staff.
 If the depreciation  period cannot  be  given,  then  the Industrial
 Revenue  Code asset depreciation range (ADR)  can  be obtained from
 the  Federal  tax manual.   Since a range of figures  is presented,
 the  midpoint of  the range  should be selected.
     The method  of  computing annual depreciation and  interest
 expenses utilizes  the capital  recovery  factor,  which determines  a
 uniform annual  series of  payments for  a given  interest rate,  and
 life equals the  initial  capital expenditure.
                                5-24

-------
5.3.2.2  Debt Service Coverage Ratio—
     The data required  to  perform the debt burden coverage anal-
ysis are specified  in  Tables  5-2 through 5-6.  Table 5-2, pollu-
tion control  cost,  has been  discussed.   Tables 5-3  and  5-4 are
the profit  and  loss summaries  for an existing and  a new or ex-
panded  plant,  respectively.  The portion  of Table  5-5  that is
required for this analysis consists of lines 10 through 13; this
information,  to  be  obtained  from  the plant  (or  a  suitable
secondary source),  consists of  the future obligations, including
principal,  interest, lease payments, and other long-term contract
obligations.  Table 5-6 is the  analytical  table based on infor-
mation  from Tables  5-2  through 5-5, and  is  used in the coverage
ratio computation.
     The sources  of  data  for   Tables  5- 2  through  5-6  are as
follows.  The data  for Table 5-2 are assumed given.  The data for
Tables 5-3  and 5-4  should come primarily from the plant.  Second-
ary sources  of  information for  constructing Tables  5-3  and 5-4
can be  found  in other studies  and  references.    Other  studies
include  those previously conducted by  the  Environmental Protec-
tion Agency and  the  Occupational Safety and Health Administra-
tion.    References  for  secondary source  data would  include the
Robert  Morris Associates,  the Troy Almanac, trade associations,
the Census  of Manufacturers,  the Federal Trade Commission, vari-
ous periodicals,  and equipment  suppliers to the  industry.  When
these studies and references will not provide all of the elements
of  Tables  5- 3  and  5- 4,  the analyst  must use judgment  or ask
members  of  industry  or trade  associations where  to  obtain the
necessary data.   In some  cases,  rough  estimates may be required
for such completions.
     Table  5- 3,  the  profit  and loss  summary  for  an existing
plant,   consists  of  16  line  items.  The  first eight are for the
basic net income  data  prior to the impact of pollution controls.
The first line calls for revenue data broken down by revenue  from
                               5-25

-------
          Table  5-3.  PROFIT  AND  LOSS  SUMMARY FOR
                        EXISTING PLANT
                                                   Most Recent  5 Years
                                               "19	 19	 19	 19	 19	
    Revenue
    a.  From operations                         		
    b.  From Non-operations                      	 	
       c.  Total Revenue                        	 	
2.  Cost of Sales
    a. Cost of Sales                        m		
    b. Pollution control expenses (existing) ^' 	 	
       c. Total Cost of Sales	
3.  Gross Operating Profits
       (l(c)-2(c))                              	 	
4.  Depreciation & Amortization
    a. Pollution control (existing)(1)          	 	
    b. Building and iirprovements               		
    c. Transportation                          		
    d. Machinery and equipment                 		
    e. Other                                   		
       f. Total Depr. & Amort.                 	 	
5.  Other Expenses
    a. Selling, general & administrative       	 	
    b. Taxes other than income tax	
    c. Research                                		
    d. Interest, existing obligations          	 	
    e. Bad debt allowance                      	 	
    f. Long term lease payments (z>             		
       g.  Total Expenses                      	 	
6.  Net  Income Before Tax
        (3-(4(f)+5(g))                           		
7.  Income Taxes
    a. State                                    		
    b. Federal                                  		
       c. Total  Income  Taxes                    	  	
 8.  Net  Income After Tax,  Before Pollution
      Control6-7(c)	  	

 (Computations below are only  performed  for column of most recent year unless
  the  source is  in a cyclical  business)

 9.  Annual Control Costs (from Table 5-2, Line 9)               	

10.  Line 9 x  (1-T)                                              	
    T =  (.48  if line 6 >  $50,000)
    T =  (.22  if line 6 < $50,000 but *  $25,000)
    T = (.20  if line 6 <• $25,000)
  (continued)
                                   5-26

-------
                     Table  5-3  (continued).
11.   Net Income After Tax and After Pollution
     Control Without Product Price Increase (Line 8-Line 10)
12.  Product Price Increase Under Full Recovery
     (Annual Control Costs 4- Number of Units
      Sold per year)
13.  % of Full Price Recovery Possible x Line 10
14.  Net Income After Tax, and After Pollution
     Control With Product Price Increase
       Lines 11 + 13)

If elasticity coefficient known and price increase expected

15.(3^A Profits   =  QAP x  {I+T3---~)

16.  Net Income After Tax and After Pollution Control
     With Product Price Increase   (Line 11 + (l-T)Line  15)
 (1)  Existing means expenses already being incurred for oonpliance to any air,
        water, solid waste, radiation, etc. pollution control regulations
 (2)  Greater than one year
 (3)  Q = Quantity Products Sold
     P = Avg. Product Selling Price
         Revenue Line1  (c)
         Quantity Products Sold  (Q)
    AP = increase in price (line 12)
     E = elasticity coefficient
     C - Unit variable costs = Lines 2+5vQ
                                     5-27

-------
        Table 5-4.
                   PROFIT AND LOSS  SUJYLARY FOR  NEW OR
                   EXPANDED EXISTING PLANT
                                                 First 3 Years of Operation
                                                 19         19         19
                                                                           (1)
1.  Revenue
   a)  Number Units Sold
   b)  Price Per Unit
                      (2)
      c)  Total Revenue

2.  Cost of Sales
   a)  Purchased Materials
      #     @
   (use supplementary sheets  for each year
    and for more materials if necessary)

   b) Fuel Oil
      Gals	 @ $	
   c) Natural Gas
      cf@ $
d) Electricity
   kwh      @ $	

e> I^or              wage
   Categories  flhrs x Rate
                             (3)
      f) Total Cost of Sales
 3. Gross Operating Profits
    (l(c) - 2(f))
 4. Depreciation and Amortization
                     Depr.
   Categories  Cost  Rate
       Total Depreciation & Amortization
                  (4)
 5. Other Esqr—
    Category
       Total Other Expenses
  (continued)
                                       5-28

-------
                            Table 5-4   (continue^) .
                                                   First 3 Years of Operation
                                                   19         19	       19

 6. Net Income Before Taxes
    (3 - (4+5))                                   	       	       	
 7. Income Taxes
                    Rates
                 Stipulated  Avg.
    a) State     	  	                   	       	       	
    b) Federal (5)	  	                   		       	
       c)  Total Income Taxes                       	       	       	

 8. Net Income After Taxes
    (6 - 7(c))                                     	                  __
 9. Annual Control Costs
    (from Table  5-2 Line 9                         	       	       	
10. Line 9 x  (1 - T)                               	       	       	
       T = (.48 if line 6> $50,000)
       T = (.22 if line 6< $50,000 but> $25,000)
       T = (.20 if line 6< $25,000)
11. Net Income After Tax and After Pollution
     Control Without Price Increase
      (Line 8 - Line 10)                            	       	       	
12. Product Price Increase Under Full Recovery
    (Annual Control Costs * number units sold
     per year)                                     	       	       	
13. Percent of Full Price Recovery
    'Possible x Line 10                            	       	       	
14. Net Income After Tax and After Pollution
     Control With Product Price Increase
      (Line 11+13)                                           	       	
If elasticity coefficient known and price increase expected:

15. (6)A Profits = Q A P x (1 + E-^p)               	

16. Net Income After Tax and After Pollution
     Control With Product Price Increase
     (Line 11 + a ~ T) x Line 15)                   	
(1) After initial construction period.
(2) Include supplementary sheets for multi-product plant.
(3) Including benefits.
(4) If allocated then describe the basis and calculations for allocation on
    separate sheet.
(5) Provide separate schedule that reconciles difference between stipulated
    and average rate.
(6) Q = Quantity Products sold;         Revenue Line l(c)
    P = Avg. ProductSelling Price = Quantity Products Sold(Q)
    AP = Increase in Price (line 12)         J
    E = Elasticity Coefficient
    C = Unit Variable Cost = Lines Q. + 5)* Q

                                      5-29

-------
               Table 5-5.   INVESTED  CAPITAL SUMMARY
                   NEW, EXPANDED OR EXISTING PLANT
                                             Exist
                                              Plant (^    New or Expanded Plant
                                                             3 Years of Operation
                                               19        19       19	       19
1.  Property,  Plant, Equipment and
    Other Assets (Book Value)

    (See bottom of Table D  for details)

    a.  Land and minerals
    b.  Buildings and  improvments
    c.  Machinery and  equipment  facilities
    d.  Pollution control  (existing)
    e.  Transportation equipment
    f.  Other fixed assets  (including  long
         term marketable securities)

        h.  Total Book Value

2.  Current Assets
    a.  Cash on hand and deposit
    b.  Temporary cash investments
    c.  Trade receivables,  net
    d.  Inventories
    e.  Other current assets
        f.  Total Current Assets

3.  Current Liabilities
    a.  Trade accounts payable
    b.  Other expense accruals
    c.  Notes payable, current
    d.  Other current liabilities
        e.  Total current  liabilities

4.  Total Net Working Capital (Line 3 minus 2)

5.  Total Investment  (l(h)+4) Before Controls)

6.  Total Invested Capital Without
    Pollution ControT
    a.   Debt    	% $	
    b.   Equity  (Preferred  & Camon Stock)
                 	% $	

         c.Total    % $
 7.   Pollution Control Capital Costs,
     Book Value (From Table 5-2, Line 1)
   (continued)
                                       5-30

-------
                         Table  5-5  (continued).
                                              Existing
                                               Plant/1)   New or Expanded Plant
                                                        First 3 Years of Operation
                                                19	      19	     19


 8.   Total Investment After Controls (5+7)                 	
 9.   Total Invested Capital With Pollution
     Control
     a.  Debt          %  $
     b.  Equity (Preferred  & Cannon Stock
                 	%  $	
        c.   Total   100%  $
 (Figures Below Required Over Life of Pollution Control Equipment)

10.  Current Installments on Long Term  Debt
     Without Pollution Control
     a.  Interest                               	 	 	 	  	  etc.
     b.  Principal                              	 	 	 	  	  etc.

11.  Current Installments on Pollution
     Control Debt
     a.  Interest                               	 	 	 	  	  etc.
     b.  Principal                              	 	 	 	  	  etc.

12.  Lease Payments and Long Term Contracts     	 		  etc.

13.  Lines 10 + 11 + 12                         	 	 	 	  	  etc.


List of Major Property, Plant and Equipment Items
     Item     Cost   % Rate of Depreciation   Item    Cost  % Rate of Depreciation
 (1)  Figures for existing plant do not have to include physical descriptions of
     facilities if obtained from audited statements.
                                          5-31

-------
                                               Table  5-6.    DEBT  SERVICE  ANALYSIS
                   Numerator of Debt Service Ratio
                                                                             Denominator of Debt Service Ratio
 I
OJ
               FDR EXISTIN3 SOURCE BEFORE POLLUTION OONTROL
               1.  Net income, before taxes
                   and interest plus 1/3 x
                   lease payments, most
                   recent  5 years            	  	  	

               2.  Depreciation, most
                   recent  5 years            	  	  	

               3.  Sub-total  (Cash Flow Fran
                   Operations)               		  	
 4.  Principal Payments, Next
    5 years T  (1-T)                 	
 5.  Interest Payments, Next 5 years 	

 6.  Lease and Long Term Contract
    Payments                        	
 7.  Capital Expenditures for Re-
    placement, Next 5 years -r(l-T)  	

 8.  Sub-Total
                FOR EXISTING SOURCES AFTER POLLUTION OONTROL, WITH3UT PRICE INCREASES
                9.  Pollution Control,  O&M
                    Costs (Line 7 - Line 8,
                    Table  5-2

               10.  Pollution Control Equip-
                    ment Depreciation
               11.  Line 3 (above)+Line 10
                    Line 9
12.   Pollution Control  Interest
     Payments (Line  11(a) Table 5-5)

13.   Pollution Control  Principal
     Payments (Line  1Kb) Table 5-5)
     T (1-T)

14.   Line 12  + 13
15.   Line 8 (above)  + 14
               FOR EXISTING SOURCES AFTER POLLUTION COmTOL, WITH PRICE INCREASE
               16.  Line 13 or Line 15,
                    Table  5-3 or 5-4 x(l-T)

               17.  Line 11  (above) + Line 16
18.  Line 14 (above)

-------
the operations of the basic business and additional revenues that
are ancillary  to  the main business of the  plant  or are sporadic
operations.  The  cost  of  sales category,  line 2,  consists of all
of the costs  of producing the product.  The  value  for  line 3 is
the difference between lines  1 and 2.   Line 4  consists  of the
noncash  flow,  depreciation,  and/or amortization  amounts for the
various items of land, equipment, and buildings.  Line 5 consists
of  all  of  the   other  nonproduction  expenses  incurred by  the
company,   broken  down  into  various categories  as  shown.   Where
these  or  any  other  costs  are  allocated  from other  plants  or
headquarters  offices,  the  basis  for  the  allocation  should  be
explained  and  quantified.   Line 6 is  the net income before tax,
which  consists of  line  3 minus  lines 4 and 5.   Line  7  is the
income tax burden for the net income before tax and line 8 is the
net income after tax but before the impact of pollution controls,
     Lines 9 through 16  call  for three different net income com-
putations.   The computation on lines 10 and  11 assumes that the
pollution control costs are  totally absorbed by the company with
no  product  price  increase and therefore  that profits  are  de-
creased by the amount of the annual control costs times one minus
the  marginal  income  tax  rate.   The  computation  on  lines  12
through  14  assumes  that  there is a price increase  of a percent-
age, perhaps 100,  of the product  price  increase  that  would be
necessary  to  fully  recover  the cost of  the  pollution control
equipment.   This  computation  assumes  that  an elasticity coeffi-
cient is not available and that the analyst would use judgment as
to what  percent  of the full product price  recovery is possible.
Lines 15 and 16 of Table 5-3 determine profits when an elasticity
coefficient is known and when a price increase is expected.
     Table  5-4 is  a profit  and loss summary  for  a new  or  an
expanded  existing plant.    The line items  1  through 16  are the
same  as   those of  Table  5-3.   The  difference in  Table  5-4  is
largely  in  lines  1 through 7, which present  the  various revenue
                               5-33

-------
and cost  items.   The primary  difference  is that  the  entries in
Table 5-4 require physical input data,  which are then factored by
costs to determine the financial result.   This method can be used
by itself  as  the sole  financial  forecast,  or it  can  be used to
verify any strictly financial forecast provided by the source for
the various line items.
     The primary  source for these line item  inputs  is the plant
source  under  consideration.   Secondary  sources  of  information
would  be  the various  studies  and references  mentioned earlier.
     Table 5-5,  applicable to an existing or new plant, calls for
invested  capital  data  for such  plants.   The  lines  that  are
necessary for the performance  of the debt service coverage anal-
ysis are lines 10 through 13, explained above.
     Table 5-6  is for the  debt  service  coverage analysis, which
brings  together  the  elements  of  the previous  tables.   As  ex-
plained in Section 5.2, the debt service coverage  is a ratio, the
numerator  of which  is  the cash  flow  from  operations  and the
denominator  of  which  is the  future  debt burden.  Table  5-6 is
divided  into  three  parts:   the  ratio before  pollution  control
impacts, the  ratio  after pollution control impacts but without  a
price  increase,  and  the ratio after pollution control and with  a
price  increase.
     The  table  is to  be read as  if there  are two columns.  The
column on  the left represents  the numerator  of the debt  service
coverage  ratio  and  the  column on  the  right,  the denominator.
Lines  1 through 3  of the  left-hand column in the before-pollu-
tion-control  situation show that  the  cash flow  consists  of the
net  income before  taxes  and  interest  payments  plus  one-third
times  any  lease payments that might be present  and depreciation.
The  reason  for  multiplying the lease  payments by  one-third is to
approximate  the interest that is  part of  those  lease  payments.
     Lines  4  through  8  of Table  5-6 in  the before-pollution-
control  situation lists the obligations, consisting of  principal
                                5-34

-------
payments,  interest  payments,   lease  payments,  other  long-term
contracts, and the capital expenditures necessary for replacement
items.   The  principal  and capital  expenditures  payments  are
divided  by  one minus  the  marginal tax rate  in order to compute
the amount of  after-tax  cash flow that pays for capital expendi-
tures  and principal  payments.   The  division  by  one  minus  the
marginal  tax  rate gives the  appearance of making  the principal
payments  larger  than they  actually are.   The division is neces-
sary,  however,  to make  both  sides comparable  from  a tax stand-
point.   If  the  left-hand part  of  the  table  could  have  been
presented on an after-tax basis, the division would not have been
necessary.
     A  question  arises  in  this  analysis  in conjunction with the
cash  flow  from  depreciation  and  the  capital  expenditures  for
replacement.   The question is  whether  the agency would wish to
force  the  operators  of source to  forego replacement  expenditures
and use  the cash  flow  that  would have gone  for  these  expenditures
to meet  the principal  interest  and lease payment obligations.  If
the  agency would not  like to make that  assumption,  then it may
want  to consider  making  the  future  yearly  capital   expenditure
levels equal to the depreciation  funds.  The  effect of this is to
make  lines 2 and  7 cancel  each  other.
      The second  section  of Table  5-6,  left hand column,  computes
the  cash flow from  operations  after  the occurrence  of  pollution
control  costs.  This cash  flow  is  then  compared with  the adjusted
cash  flow,  right hand column,  from line.  8 above,  which consists
of adding  the  pollution  control  interest payments and the princi-
pal  payments,  the  latter  divided by  1  minus the  marginal tax
rate.
      The next  section provides a  similar  analysis  for the  impact
after pollution control where  a price increase does  occur.  The
left-hand column adjusts  the  cash flow  from operations for the
additional revenue  from the price increase,  and the  right-hand
                                5-35

-------
column does not  adjust  the payments,  but has them  equal  to line
14 of the second part of the table.
     The coverage  ratios  can be  computed  in more  than one way.
The analyst may  wish  to average the last 5 years with the next 5
years.  This has the effect of dampening out any cyclical effects
in operation  of  the source.  Or,  the analyst may wish to choose
the funds  flow from  operations  for the most recent  year,  if he
considers  it  more  representative than  the  previous  years,  and
compare that level with future years.
     The analyst may wish  to account  for  inflation in adjusting
the ratio.   If it appears  that the plant  under consideration is
increasing earnings on a real basis (i.e.,  after inflation), then
the  analyst may wish to  adjust  the  cash  flows of the previous
years by a  rate  of inflation to more accurately compare past and
future cash flows.
     Concerning  the  decision criteria,  if the  ratio is greater
than  2.0,  it is considered that cash  flow  is  adequate to cover
the future obligations.  A  ratio of 1.5 to 2.0 indicates that the
cash  flow  is probably  adequate to  cover future obligations,  but
that some risk is inherent.  As a criterion for this  technique, a
numerical  figure of less  than 1.5 was  chosen  as the level below
which  the  cash  flow is  not considered adequate to  cover debt-
burden  and therefore  as  the level that constitutes  significant
impact.  The benefit of  performing this  calculation for condi-
tions  before  pollution control is  that the  analyst can identify
those  cases  where  a significant impact may not occur  even though
the  ratio  after pollution  control  is  less  than 1.5 but greater
than  1.0.   This situation  would  occur  where the plant source is
accustomed to operating at  such a ratio.
     The analyst also needs  to  determine  whether the source has
"unproductive"  long-term  assets  that can  be sold to  finance the
pollution  control  equipment.  Such assets would appear in Table
5-5,  line   l(f)  (see next  section  for  the  explanation of Table
5-5).
                                5-36

-------
     The analyst  should note that the values  developed for many
lines of the various tables are not entered into the final calcu-
lations in application  of  an analysis technique.  In some cases,
therefore,  it  is  not necessary to know  the  breakdown of certain
line items,  or the  analyst  may wish to  obtain information only
for critical line items of the tables, as long as the information
is  highly  accurate.   Sometimes,   the "extra"  lines  serve  as  a
validation  or  as  input  for compiling  values  for other lines.

5.3.2.3  Return on Investment (ROI) Ratio--
     Tables  5-2  through 5-5 and Table 5-7 are  necessary to per-
form  the  Return  on  Investment   (ROI)   technique.   Tables  5-2
through  5-4  are  to  be  completed  as described above.   Lines  1
through  9  of  Table  5-5 are to be completed for  this analysis.
These  lines  consist  of the asset values  of  the plant source,  to
be reported  at book  value.   The best source for these data would
be other  EPA or  OSHA  impact studies performed on the industry,
plus the  sources  referenced earlier.   Line 1  is  for long-term
assets; lines 2 and 3 are current assets  and current liabilities,
respectively.  Line  4 is the  difference between  lines  2 and 3,
which is called Net Working Capital.  Line 5 is the total invest-
ment value of the plant without the pollution control expenditure
                                                *
under consideration.  Line 5 is the sum of lines 1 and 4.  Line 6
calls for a  listing of the invested capital of  the source divided
into  the debt and  equity  portions.    (In  calculation  of  debt
amounts, deferred taxes should be  treated as  a debt item if it
appears  that in  the next few years  the  amount of deferred taxes
will decrease;  otherwise they are to be treated as equity.)  In
addition to  dividing debt and equity by  their  percentage contri-
butions  to  the total,  the  absolute  dollar  value  should also be
listed  as  a subcategory of  line 6.   Line 7  is pollution control
capital  costs taken from  Table  5-2.   Line 8  is the  total  of
assets before pollution control.  The debt to equity breakdown in
                               5-37

-------
  Table 5-7.  ROI SUMMARY  FOR NEW, EXPANDED OR EXISTING PLANT
1.  Net Income After Taxes, Before Controls
    a".   Existing (Line 8,  Table 5-3)
    b.   New or Expanded (Line 8, Table 5-4)

2.  Investment: Fixed Assets  Plus Net
    Working Capital (Lines 1  + 4,Table 5-5^

3.  ROI Before Controls (1 4  2)

4.  Total Investment with Controls
    (Line 2 (above) + Line 1  (Table 5-2)

5.  Net Income After Taxes, After Controls
    Without Price Increase
    a. Existing (Line 11, Table  5-3)
    b. New  (Line 11, Table 5-4)
6.  ROI After Controls Without Price
    Increase  (5 - 4)
7.  Net Income After Controls With Price
    Increase
    a. Existing (Line 14 or 16,  Table 5-3)
    b. New  (Line 14 or 16, Table  5-4)

8.  RQI After Controls With Price
    Increase   744)
                                              Existing
                                               Plant
                                                 19
xxxxxx
xxxxxx
xxxxxx
         Year 1,2 or 3 of Operations
            New or Expanded Plant
                  19
                 xxxxxx
                 xxxxxx
                 xxxxxx
                                       5-38

-------
percentage  and dollar  amounts  should be  presented  here as  a
subcategory of line 9.
     Table 5-7 is  the analytical table for which the analysis of
ROI is  performed.   ROI is essentially a  fraction,  the numerator
of  which is  an  income  figure,  the  denominator  of which is an
asset figure.  Table 5-7 presents the ROI  before pollution con-
trol in  lines 1  through  4  for existing  and  for  new or expanded
plants.    Lines 5  and  6 are  for  computing  the  ROI  after  the
occurrence of the  pollution control  expenditure  without a price
increase  to  cover  the cost of pollution  control.   Lines 7 and 8
are for the ROI calculation with a price increase.
     There are no  established criteria for determining when the
percentage change  in  ROI  or  the absolute ROI  is to be considered
as  being significantly impacted by pollution controls.   As des-
cribed  earlier,  the  analyst  should  first  compare  ROI  with the
plant's  threshold  rate  of ROI  for new  investments.  If that rate
is  not  obtainable, the analyst should obtain an  average ROI for
the industry or,  lastly, should obtain the firm's  cost of funds
for the next  financing.

5.3.2.4  Discounted Cash Flow (DCF)—
     To  perform  the Discounted  Cash  Flow technique, the  analyst
must complete Tables  5-2 through  5-4 plus  Tables  5-8  and 5-9.
Table 5-8 is  a  preliminary to  the  analysis  conducted  in Table
5-9.
     The  first line  of  Table  5-8 calls  for  an extrapolation of
net  income  after  tax,  pollution controls, and price considera-
tions plus  interest payments  over a  period of  years equal to the
useful  life  of the control equipment.  The analyst should use a
text on  extrapolation to determine which extrapolation procedure
would be most applicable to  the data at hand.  Line 2 of Table
5-8 calls  for the  liquidated value of the  plant.  The liquidated
value of a plant can be  considered as the  current book value of
                               5-39

-------
      Table  5-1
INFORMATION  FOR  VALUE  ANALYSIS  FOR  NEW,  EXPANDED,OR EXISTING  PLANT
Ul
 I
              1.  Extrapolation of Net Inccme After Tax After Controls,
                  With or Without Price Increase Plus Interest Payments
                  for Number of Years Bqual to Control Bquipment Life
                  (Lines 11, 14 or 16, Tables 5-3 or 5-4 + Line  5, Table 5-3 or 5-4
                                                                                                etc.
              2.  Liquidated Value
                  a.  New Source                 =
                  b.  Expanded or Existing Source =


              3.  Terminal Value

              4.  Cost of Equity = 15% (I-T)  or  =
                    JCurrent book value of assets (Line 1 (h)
                    ~ unless appraisal obtained frcm source.)

                     (Require source to document estimate)
                                                                                Line 4 , liable 5-5)
                             Net Incone after Controls with or without price  increase
                             (Line 11,  14 or 16, Table 5-3 or Line  11,14 or 16,TabloJ5--4--
                               Market value of Equitey (See financial  text tor computation
               5.  Weighted Average Cost of Capital (C)
                                                       McLI.Js.cT_ Vd-HJc ui- 04 vu. i-^y  \ ^.K-n—  J-J-IILUI^.^"^	        i           .
                                                       or if pvtolicly owned  company  use eaminqs oer share t  stock once)
                                       (1)
                                                  (2)     (3)
                                            (4)
                                          Cost or
                                         Interest
                                         RateilOO
                                                                                         (5)
                  Qjuity
                  Debt or
                  Preferred
                  Stock
        II
        #2
        #3
        In
J  f   100
        100
        100
        100
        100
                                       100%
 Cost
 (From
Line 4)
                                                                            Total
                                                                            = (C)  Weighted Avg.
                                                                                  Cost of Capital

-------
Table  5-9.    PRESENT  VS.   FUTURE  VALUE  ANALYSIS  FOR  NEW,   EXPANDED, OR EXISTING  PLANT
           Pollution Control Equipnent            #1	|2	|3	|4	|5      ...       #n
             Depreciable Life	Years (n)       19	19	19	19	   19      ...       19
           1.  Net Inocme After Tax After
              Controls, With or Without
              Price Increase Plus Interest
              Payments (Line 1, Table 5-8)

           2.  Plus:  Pollution Control
              Depreciation (Line 5, Table 5-2)

                            Subtotal
           3.  Less:  Each year's current
              Portion of long-term debt
              plus leases (Line 13, Table 5-5)

           4.  Less:  Future new enviromental
              Annualized Pollution Control
              Expenditures x (1-T) + Depr.

           5.  Net Cash Flow

           6.  Discount Rate (l+C)n (C is from
              Line 5, Column 5, Table  5-8)

           7.  Discounted Cash Flow  (5 4- 6)
                                                Total of
                                                Line 7
          (1)   (1+C) numbers under each line are not
                divisors, they are descriptive of
                number to be entered on line.)
 8.   Terminal Value T (l+C)n
 9.   Investment Tax Credit (line 10,Table 5-2)
10.   Sub-total (7+8+9)                       ~
11.   Less Liouidated Value                  ~
12.   Present Value                          ~~

-------
the fixed  assets  plus  the net working value, both  figures  being
taken from Table  5-5.   If the agency so desires,  a  source can be
asked to provide its own appraisal of the liquidated value of its
assets and net working  capital.   Such an appraisal  would require
documentation as to the basis for the amount provided.   Line 3 of
Table 5-8  is the  terminal value,  that is,  the value of the plant
at the end of the useful life of the pollution control  equipment.
If it is  difficult for  the source to estimate such  a  value,  the
analyst may  wish  to  take the  present  liquidated  value of  the
plant  and increase  the  figure by  an  appropriate  inflationary
rate.  The analyst may use  discretion as to whether  to use the
results of the figure  shown  in the second part of  line 4  or to
use an arbitrarily high figure,  such as 15 percent before tax if
the plant is in  financial difficulty.   The arbitrary  figure is
based on  the  fact that the inherent cost of equity capital for a
marginal  plant  is really  at  least 15 percent before tax in order
to attract capital.  Line 5 of Table 5-8 calls for calculation of
the  weighted average costs of  capital.   The capital consists of
the  equity cost plus the cost  of  the  various  debt  or preferred-
stock  issues.   These are  to  be computed  on an after-tax basis.
The various volumns of  line 5 indicate the calculations  needed, to
derive  the weighted average  costs of capital.  In the event  that
deferred  taxes are  included in  the capital items, the cost of
such capital is to be considered zero.
     Table 5-9 contains  the  calculations  for computing the  dis-
counted  cash flow technique.   The period  over which the calcula-
tions are to be performed is  equal to the  depreciable life of the
pollution control  equipment.   The  future value  of  the plant
consists  of  the net income over  the useful  life of  the  pollution
control  equipment plus  the cash flow that will be  available  from
depreciation  of the pollution control equipment.   To be deducted
from  those amounts are  the obligations  of  the company plus  future
pollution control expenditures.   These  pollution control expendi-
tures  are other than the  one under  consideration and  should  only
                                5-42

-------
be those  for which the  company has already  entered  into a con-
tract.  Additional pollution  control  expenditures  not contracted
for could be  added  at the discretion of the  analyst  if they are
verifiable.   The results of these calculations thus far will be a
net  cash  flow  for  each  of  the  years  of the  pollution control
equipment's  useful  life.   These  future  cash flows must  be dis-
counted to  a present value  by dividing each yearly  figure by a
compounded  weighted  average  cost  of  capital,   as  shown  in the
appropriate line,  The results are discounted cash flows for each
year, which are summed to the right of Table 5-9.  Added to these
discounted cash flows are the terminal value of the plant, appro-
priately  discounted,   and the  investment  tax  credit  from the
pollution control investment under consideration.  To be deducted
from  this  amount  is  the liquidated value  of  the  plant  as  if
liquidated  today.   If  the  result  is  greater than 0,  the plant
would  decide to continue operations  with the  pollution control
equipment  installed.   If the value  is  less than  0,  the plant
would  decide,  in rational  decision-making  theory,  that greater
profits are  to  be  gained elsewhere.  Thus, significant impact is
considered to occur when the value is less than 0.

5.3.2.5  Previous Impact Studies—
     The  analyst should obtain  a list of the  industry economic
impact studies  available  from EPA's Office of Water Programs and
Office of Air Quality  Planning and Standards.   The  analyst may
also wish to contact OSHA for  industry impact studies.
     Since  the  sizes of the plants  analyzed  in  such economic
impact studies  are  likely to be  different from that of the plant
under  consideration,  it will  be necessary to  scale  the control
costs  for the plants in  the  economic impact studies  to the size
of the plant being analyzed or  vice versa.   Scaling  can be done
with the  aid of  another EPA publication (available from the
                               5-43

-------
Office  of  Air  Quality  Planning  and  Standards),   Capital and
Operating Costs of Selected Air Pollution Control Systems,  EPA-
450/3-76-014,  May 1976.
     In  addition  to scaling  the control  costs the  analyst may
also need  to  update  the  control  costs, since  some of the EPA
economic impact  studies  may be  up  to 5 years  old.   Appropriate
indices  for  updating those  control  costs  are the  Marshall and
Swift  (M&S)  indices published  by McGraw-Hill in  Chemical  Engi-
neering.
     As  mentioned  earlier,  the analyst  must  use   judgment  to
determine whether  the  two sets  of  control costs  being compared
justify a judgment concerning the impact of the proposed expendi-
tures on the plant in question.  There are no established criter-
ia for the analyst to follow in making such judgment.
     The earlier EPA economic impact studies will also be helpful
in application of several of the other analysis techniques.

5.3.2.6  Financing Decisions by Others—
     This analysis  technique is applicable  only to a source that
is  alleging  difficulty  in making the pollution control expendi-
tures because  of the inability to  obtain  financing.   The  tech-
nique cannot  be used to  anticipate the impact  of air pollution
control costs on a source.
     First a determination of  the size  (number of employees) and
the business Standard Industrial Classification code  (SIC)  of the
company  (not solely the  plant) must be made to determine whether
the  source  is  eligible  for  an  SBA loan.   The  plant is the best
source  for  this information on number of  employees  and SIC.  A
phone call  to the nearest SBA  office will  determine whether the
company  is  indeed a small business.   If the  company is involved
in  more  than  one line of business  the percent  of  sales for each
line must be supplied to SBA.
                               5-44

-------
     If the  business is  small by  SBA standards,  it  should  be
directed to  apply for one  of  the pollution  control  programs  of
the  Small  Business  Administration,  described in  Section 5.3.1.
If a determination  is made  by  the SBA that the source is refused
financing solely because  of the  pollution control expenditure  in
question,   then  significant  impact   has  been  determined.   The
necessity for attributing the  rejection to the pollution control
expenditure  is  that  there  are many  other reasons why  any loan
could  be  refused  by any  type  of  lending  organization.   Even
without the impact the company would not have received financing;
other possible reasons for refusal are lack of SBA funds, improp-
er  application  documentation,  management  difficulties,  or  the
inability to make  a determination.   The entire  SBA  application
process may require  60 to 120 days.
     If the company  does not meet the SBA definition of small,  it
should be  directed  to obtain  at least two  letters  of financing
refusal  from commercial  financing organizations.  Again,  these
organizations should state  the refusal is attributable to pollu-
tion control costs.
     This method  does not  adhere to the  definition  of "plant"
                                                   #
given  earlier because any  financing  organization looks  to  the
entire company, and not  just the plant, as security repayment of
a loan.

5.3.2.7  Increase in Operating Expenses or Assets—
     For  analysis  of cost  increases attributable to pollution
control that could  lead  to  significant impact,  the  analyst must
complete Tables 5-2 through 5-5, explained earlier.   The cost of
pollution control,  i.e.,  annual  control costs, is to be compared
with  lines   2,  4,   and 5  of Tables   5-3  or  5-4,  depending upon
whether the  plant is existing or  new.   These line items consist
of plant expenses.   If  the  annual control costs are greater than
30  to  40 percent after accounting for price increases, the sig-
nificant impact may  be thought to occur.
                               5-45

-------
     In the  event  that the analyst  wishes to compare  the  fixed
capital costs  increase  from pollution control expenditures,  the
first line  of  Table 5-2  should  be compared with the  first line
[l(h)j of  Table  5-5.   The percentage increase deemed to  be sig-
nificant in  this case  is  higher,  i.e.,  40 to 50  percent,  because
of the normal  situation in  which a new plant may have a signifi-
cant portion of  its cost devoted  to pollution  control expendi-
tures.  The  figures of 30 to  40 percent  for  an  expense increase
and 40 to  50 percent  for  an asset increase are  not  based on any
recognized  criteria.   The  analyst  may wish  to  use  judgment to
modify these expenditure levels.
                               5-46

-------
APPENDIX A
   A-l

-------
                Table  A-l.   CHEMICAL ENGINEERING PLANT AND EQUIPMENT COST INDEXES'
Index
CE plant index
Engineering and
supervision
Building
Construction labor
tquipment, machinery
supports
Fabricated equipment
Process machinery
Pipe, valves, and
fittings
Process instruments
and controls
Pumps and compressors
Electric equipment
and materials
Structural supports,
insulation, and paint
1969
Annual
119.0
110.9
122.5
128.3
116.6
115.1
116.8
123.1
126.1
119.6
92.8
112.6
1968
Annual
113.6
108.6
115.7
120.9
111.5
109.9
112.1
117.4
120.9
115.2
91.4
105.7
1967
Annual
109.7
107.9
110.3
115.8
107.7
106.2
108.7
113.0
115.2
111.3
90. 1
102.1
1966
Annual
107.2
106.9
107.9
112.5
105.3
104.8
106.1
109.6
110.0
107.7
86.4
101.0
1965
Annual
104.2
105.6
104.5
109.5
102.1
103.4
103.6
103.0
106.5
103.4
84.1
98.8
1964
Annual
103.3
104.2
103.3
108.5
101.2
102.7
102.5
101.6
105.8
101.0
85.5
98. 3
1963
Annual
102.0
103.4
102.1
107.2
100.5
101.7
102.0
100.7
105.7
100.1
87.6
97. 3
1962
Annual
101.5
102.6
101.4
105.6
100.6
101.0
101.9
100.6
105.9
101.1
89.4
99.2
1961
Annual
102.0
101.7
100.8
105.1
100. 2
100. 1
101. 1
101. 1
105.9
100. 8
92. 3
99.8
1960
Annual
101.8
101.3
101.5
103.7
101.7
101.2
108.1
104.1
105.4
101.7
95.7
101.9
1959
Annual
101.8
102.5
101.4
101.4
101.9
100.9
101.8
103.3
102.9
102. 5
101.0
101.6
1958
Annual
99.7
99. 3
99.5
100.0
99.6
99.6
100.1
98.8
100.4
100. 0
100.6
100.4
1957
Annual
98.5
98.2
99.1
98.6
98.5
99. 5
98.1
97.9
96.7
97.5
98.4
98.0
I
to
         From Card, Inc.,  Reference 2.  More details on other cost indexes are available from this manual.
            (continued)

-------
Table A-l  (continued).
Index
CE plant index
Engineering and
supervision
Building
Construction labor
Equipment, machinery
supports
Fabricated equipment
Process machinery
Pipe, valves, and
fittings
Process instruments
and controls
Pumps and compressors
Electric equipment
and materials
Structural supports,
insulation, and paint
1977
Annual
204.1

161.7
197.2
176.8

218.8
212.8
210.1

247.3

201. 5
238.8

159.2

223.6
1976
Annual
. 191.1

150.7
185.3
172.9

205.0
198.5
196.9

235.1

192.5
221.2

147.6

206.7
1975
Annual
182.3

141.8
176.9
168.4

194.7
192. 2
184.7

217.0

181.4
208.3

143.0

198.6
1974
Annual
165.4

134.4
165.5
163.4

171.2
170.1
160.3

192.2

164.8
175.7

126.4

172.4
1973
Annual
144.1

122.8
150.6
157.9

141.9
142.5
137.6

151.3

147.1
139.5

104.2

140.9
1972
Annual
137.2

111.9
142.0
152.2

135.4
136.3
132.1

142.9

143.8
135.9

99.1

133.6
1971
Annual
132.3

111.4
135.5
146.2

130.4
130.3
127.1

137.3

139.9
133.2

98.7

126.6
1970
Annual
125.7

110.6
127.2
137.4

123.8
122.7
122.9

132.0

132.1
125.6

99.8

117.9
            From Card, Inc.,  Reference 2.
            this manual.
More details on other cost indexes are available  from

-------
                       Table A-2.   CAPITAL RECOVERY FACTORS
Equipment
life, yr
1
2
3
4
~
5
6
\J
7
8
U
9
10
11
_I_ ~L
12
_1_ £i
13
14
15
16
17
18
19
20
21
22
23
24
25
Annual compounded interest, %
5
1. 05000
0. 53780
0. 36721
0. 28201
0.23097
0. 19702
0. 17282
0. 15472
0. 14069
0.12950
0. 12039
0.11283
0.10646
0.10102
0.09634
0. 09227
0. 08870
0.08555
0.08275
0.08024
0.07800
0.07597
0.07414
0.07247
0.07095
6
1.06000
0.54544
0.37311
0.28859
0.23740
0.20336
0.17914
0.16104
0.14702
0.13587
0.12679
0.11928
0.11296
0.10758
0.10296
0.09895
0.09544
0.09236
0.08962
0.08718
0.08500
0.08305
0.08128
0.07968
0.07823
7
1.07000
0.55309
0.38105
0.29523
0.24389
0.20980
0.18555
0.16747
0.15349
0.14238
0.13336
0.12590
0.11965
0.11434
0.10979
0.10586
0.10342
0.09941
0.09675
0.09439
0.09229
0.09041
0.08871
0.08719
0.08581
8
1.08000
0.56077
0.38803
0.30192
0.25046
0.21632
0.19207
0.17401
0.16008
0.14903
0.14008
0.13270
0.12652
0.12130
0.11683
0.11298
0.10963
0.10670
0.10413
0.10185
0.09983
0.09803
0.09642
0.09498
0.09368
10
1.10000
0.57619
0.40211
0.31547
0.26380
0.22961
0.20541
0.18744
0.17464
0.16275
0.15396
0.14676
0.14078
0.13575
0.13147
0.12782
0.12466
0.12193
0.11955
0.11746
0.11562
0.11401
0.11257
0.11130
0.11017
12
1.12000
0.59170
0.41635
0.32923
0.27741
0.24323
0.21912
0.20130
0.18768
0.17698
0.16842
0.16144
0.15568
0.15087
0.14682
0.14339
0.14046
0.13794
0.13576
0.13388
0.13224
0.13081
0.12956
0.12846
0.12750

15
1.15000
0.61512
0.43798
0.35027
0.29832
0.26424
0.24036
0.22285
0.20957
0.19925
0.19107
0.18448
0.22526
0.21689
0.17102
0.16795
0.16537
0.16319
0.16134
0.15976
0.15842
0.15727
0.15628
0.15543
0.15470

20
1.20000
0.65455
0.47473
0. 38629
0.33438
0.30071
0.27742
0. 26061
0.24808
0. 23852
0.23110
0. 22526
0.17911
0.17469
0.21388
0.21144
0. 20944
0.20781
0.20646
0.20536
0. 20444
0.20369
0.20307
0.20255
0.20212
(continued)

-------
                              Table A-2 (continued).
i
Ui
Equipment
lire, yr
26
27
28
29
30
31
32
33
34
35
40
45
50

5
0.06956
0.06829
0.06712
0.06605
0.06505
0.06413
0.06328
0. 06249
0. 06176
0.06107
0.05828
0. 05626
0.05478
Annual compounded interest, %
6
0.07690
0.07570
0.07459
0.08385
0.07265
0.07179
0.07100
0.07027
0.06960
0. 06897
0.06646
0.06480
0.06344
7
0.08456
0.08343
0.08239
0.08145
0.08059
0.07980
0.07907
0.07841
0.07780
0.07723
0.07501
0.07350
0.08246
8
0.09251
0.09145
0.09049
0.08962
0.08883
0.08811
0.08745
0.08685
0.08630
0.08580
0.08386
0.08259
0.08174
10
0.10916
0.10826
0.10745
0.10673
0.10608
0.10550
0.10497
0.10450
0.10407
0.10369
0.10226
0.10139
0.10086
12
0.12665
0.12590
0.12524
0.12466
0.12414
0.12369
0.12328
0.12292
0.12260
0.12232
0.12130
0.12074
0.12042
15
0.15407
0.20176
0.20122
0.20102
0.20085
0.15200
0.15173
0.15150
0.15131
0.15113
0. 15056
0.15028
0.15014
20
0.20176
0.15353
0.15306
0.15265
0.15230
0.20070
0.20059
0. 20049
0 20041
0.20034
0. 20014
0.20005
0.20002

-------
             Table  A-3   CAPITAL  COST  ESTIMATION
 Equipment
   item
Purchase
 price
   Direct
cost factor'
Direct
 costb
                  Equipment
                    cost0	
                             Total
                             direct
                             costd _
A.   Total direct costs

B.   Total indirect costs at 38% (A)

C.   Contingencies at 15% (A + B)

     Total capital costs
          (A + B + C)
  From Tables 4-1, 4-2, and 4-3.
  Includes equipment cost and installation, instrumentation,
  electrical work, foundations, structural work, site work,
  painting, piping, and labor.
  Sum of equipment purchase prices.

  Sum of individual direct costs.
  Includes engineering, contractor's fee, interest during
  construction, shakedown, spares, freight, taxes, offsite,
  and field and labor expenses.
                             A-6

-------
     Table A-4.   SUMMARY OF PREDESIGN  ESTIMATE OF  ANNUALIZED  COSTS

DIRECT
I.


II.
Item
OPERATING COST
Utilities
Raw materials
Electricity
Water
Steam
Gas
Fuel oil
Operating labor
Direct labor
Supervision
Unit

$
$
$
$
$
$
$
15%
cost

/ton
/kWh
/103 gal
/103/lb
/106 Btu
103 gal
/man-hour
direct labor
Quantity Cost

tons/yr
kWh
103 gal/yr
103 Ib/yr
106 Btu/yr
103 gal/yr
man-hours/h
operation
 III.   Maintenance (see Tables  4
        Labor and materials
        Supplies
        Replacement parts

  IV.   Sludge disposal

   V.   Wastewater treatment

FIXED  COSTS

   I.   Taxes, insurance, over-
        head, etc.

  II.   Capital recovery charges
PRODUCT RECOVERY CREDIT
-6,  4-7)





   $

   $
/ton

/103 gal
tons/yr

103 gal/yr
   4% of total  capital investment

      % of total capital investment

                 TOTAL FIXED COSTS

   $      /ton              tons/yr

             TOTAL ANNUALIZED COSTS

-------
           PLANT AND PROCESS PARAMETERS
Type of process:.
Production capacity, tons/h.

Throughput capacity, tons/h

Annual production, tons/h	

Operating time, h/yr	
Emission limitation, Ib/h
                     gr/scf
Estimated length of ductruns, ft

Space limitations:	
Other limitations/restrictions:
                        A-8

-------
                     EMISSION DATA SUMMARY
 Parameter
High
Variation
  average
                                                             Low
 Gas flow rate,  acfm

 Temperature,  °F

 Moisture content,  % V

 Grain  loading,  gr/scf

 Particle size distribution
  % W <  vim-


 Gas analysis

     % V  02

     % V  C02

     % V  CO

     % V  SO
           x
     % V N,
Uncontrolled emission
 rate, Ib/h
       gr/scf
Emission limitation,
       Ib/h
       gr/scf
                            A-9

-------
                    SUMMARY OF PREDESIGN CAPITAL COST ESTIMATES
>
I
Equipment
item

Footnotes :
Design
basis


Capacity/
size

Cost sources
and basis


Bare
cost

Installation
factor

Direct
installed
cost

A. Total direct cost
B. Total indirect cost
% A
C. Contingency 20% (A+B)
Total capital cost

-------
            APPENDIX B




       SAMPLE COST ESTIMATES




FOR AN AIR POLLUTION CONTROL SYSTEM




USING THE COST ESTIMATING PROCEDURE
               B-l

-------
        DESIGN DATA FOR THE CONTROL SYSTEM


                Process Parameters
Type:  Reverberatory furnace
Production capacity:  45 Mg/h (50 TPD)
Operating time:  6000 h/yr
Emission limit:  3.6 kg/h (7.9 Ib/h)
Ductwork:  110 feet total
Solid waste disposal:  landfill in sealed barrels
             Emission Characteristics


Exhaust volume:  18.4 m3/s  (39,000 acfm)
Temperature:  1090°C  (2000°F)
Emission rate:  83 kg/h3(183 Ib/h) particulate matter
Grain loading:  5.7 g/m   (2.5 gr/scf)
Particle size:  Majority between  0.03 -  0.5 ym
               Control System Design
Shaker-type fabric  filter
Superficial velocity,  1.2 cm/s  (2.5  ft/min)
System pressure drop,  1.8 kPa  (7  in.  H20)
Efficiency, greater than 99 percent
Annual labor,  2000  h
                         B-2

-------
PUMP FAN
30ftLINED 300 gpm '
39,000 acfm / ' \
t 2000°F *" QUENCH
TOWER

CD *"
w 30 gpm
Mni/rim
RbVLRBLRAIOkY MAKLUI »•
FURNACE
' 20 ft 20 ft A-X ^20 ft
18,000 acfm / J INSULATION
250°F / 1
SHAKER-TYPET
FABRIC-FILTER DISPQSAL

300-GAL
HOLD TANK
A/C  = air-to-cloth ratio.
      Figure B-l.  Control  system diagram for  reverberatory furnace.

-------
                    CAPITAL COST ESTIMATE

A.   QUENCH TOWER
     A carbon steel quench tower is required to cool the
exhaust gases from 2000° to 250°F.  The following design
specifications and purchase costs were obtained from a
Croll-Reynolds representative:
     Exhaust volume -         39,000 acfm
     Diameter of tower -      10 ft
     Depth of packing -        3 ft
     Pressure drop -          2 in. H20
     Height of tower -        16 ft high
     Liquid flow rate -       300 gpm water
     Purchase cost -          $26,000
Prices are in mid-1977 dollars and include vessel, spray
heads, controls, arid supports.
     Refractory lining for the tower is estimated at $50 per
ft  installed.  The lining is 4 in. thick, and a total of
220 ft  of refractory brick is needed at an installed cost
of $11,000.
     Direct cost factor =3.0 (Table 4-2)
     Purchase cost      = $26,000
          Total direct costs
               Tower      $78,000
               Lining	11,000
               Total      $89,000
B.   PUMP SYSTEM
     A cast iron centrifugal pump capable of handling 300
gpm at 100 ft head is required for the gas cooling system.,
A 20-hp pump  system was previously purchased from an Ingersoll-
                             B-4

-------
Rand supplier in 1971 for $1000.  The price included the
base, couples, and motor.  This application requires a 15-hp
pumps system, arrived at in the following manner:
                   (gpm)  (ft head)  _  300  (100)
     horsepower = 3g87  (efficiency)~3387  (0.5)

                = 15 hp
The price of the 15-hp pump in 1977 is estimated as follows:

     Purchase price =    (^TR^ '    x $100° x irllf = $130°
     The direct cost factor is 2.4 (Table 4-2).
          Direct cost =2.4 ($1300) = $3100
C.   HOLD TANK
     It is estimated that 30 gpm of water will be lost and
must be made up on a continuous basis.  A retention tank of
10-min capacity  (300 gallons)  is required.  Carbon steel
construction is adequate.  A 1968 purchase price of $4270 is
obtained from a technical journal.  The CE indexes yield an
estimated 1977 price of  $6900.  The direct cost factor of
2.0 is applied and results in a total direct cost of $13,800.
D.   FABRIC FILTER
     A shaker-type fabric filter is selected as the desired
control.  The exhaust gas volume at the inlet  is 18,000 acfm
at 250°F.  Purchase costs were obtained from Fisher Kloster-
inann for the following  specifications:
     Air-to-cloth ratio  of 2.5
     Dacron bags
     7500 ft2 cloth area
     Insulation, shaker  units, bags,  ladder,
          supports, hopper, factory assembly
     Purchase price = $22,500
                             B-5

-------
     Direct cost factor =2.0 (Table 4-1)
     Direct cost = $45,000
E.   FAN SYSTEM
     The system pressure drop is predicted at 7 in. E^O
during normal operation.  At 18,000 acfm capacity, the fan
is sized at 35 hp in the following manner:
     ,             (acfm) (pressure drop, in. H0O)
     horsepower = -- 6356 (efficiency) - 2~
                ~ 6356 (0.6)
     From a compilation of fan system purchase prices, the
1977 price is estimated to be $2600 for a motor, starter,
and drive of carbon steel construction.
     A direct cost factor of 2.5 is applied and results in  a
total direct cost of $6500.
F.   DUCTWORK
     Approximately 50 ft of 1/4 in. carbon steel ductwork
lined with refractory is required for the duct run from the
furnace to the quench tower.  Recommended velocity through
the ductwork is 400 ft/min.  A price of $1.50/lb of  installed
carbon steel ductwork was obtained from a construction cost
manual .
     An additional 60 feet of ductwork is needed between the
quench tower and stack.
                       i       f4-2   39,000 acfm _ Q  .,,- f 2
     A = cross-sectional area, ft  = "4000 fpm -- 9.75 ft
                           , .1/2      , ,Q  -,r\
                   ri-      4 A         4(9.75)       _   „
     D = diameter, ft =    — =—    =    -- ^ -    =  3.52  ft
A 42-in. -diameter duct of  1/4  in. carbon  steel  is  selected.
                       2
The weight is 5.5 Ib/ft  and the length is  50 feet.
                              B-6

-------
      Total surface area,  ft  = (length)  (diameter)  n
                              = 50 (42/12)  n = 550 ft2

 Because the temperature in this duct run is 2000°F, a refrac
 tory lining is installed  at a cost of $50/ft3.   A 4-in.
 thickness is recommended  by the supplier.
      Installed cost = 550 ft2  (~- ft)  $50/ft3  = $9200
 The  60-ft duct run from the quench chamber to the stack  is
 1/4-in.  carbon steel.   An exhaust volume of about 18,000
 acfm is handled at a velocity of  400 fpm.
      Dieters,  ft  =      -foi          -2.5ft

      Total  surface  area,  ft2 =  60  (2.5)  n =  471  ft2
      The direct  costs  for ductwork  installation  is calculated
at  $1.50/lb.
                            2                      ••)
      Total  surface  area,  ft =  471  +  550  = 1021  ft .
      Total  weight,  Ib  = 1021 ft2  (5.5  lb/ft2)  =  5600  Ib
      Direct cost for ductwork = 5600  ($1.50) = $  8,400
      Direct cost for refractory =                 9,200
          Total direct cost =                  $17,600
      The material portion of ductwork  is  $3200.
G.    INSULATION
      The ductwork from the  quench tower to the fabric filter
and the fabric filter itself must be insulated.  Mineral
wool  insulation 4 in.  thick and an aluminum casing are
needed for about 3000 ft2 of surface area  (estimated from
ductwork and fabric filter dimensions).  The direct costs
are estimated from a construction cost manual at $7.50/ft2.
                            B-7

-------
           Table B-l.  SUMMARY OF CAPITAL COSTS
Item
A. Quench tower and
refractory lining
B . Pump system
C. Hold tank
D. Fabric filter
E. Fan system
F. Ductwork
G. Insulation
H. Refractory
Equipment and
material
purchase cost
$ 26,000
1,300
6,900
22,500
2,600
3,200a
8,700a
5,100b
Direct cost
factor
3.0
3.5
2.0
2.0
2.5
2.6
2.6
4.0
Direct
cost
$ 78,000
4,550
13,800
45,000
6,500
8,400
22,500
20,200
                      Total  direct cost (rounded)  $199,000
     Total direct cost =
                        Q
     Total indirect cost  =
                d
     Contingency
     Total capital cost =
$199,000
  75,600
  41,200
$316,000 (rounded)
a Back-calculated from an estimated direct cost factor of 2.6
  to approximate purchase cost.
b Back-calculated from an estimated direct cost factor of 4.0
  Included is refractory for ductwork and quench chamber.
c Estimated at 38 percent of direct cost.
d Estimated at 15 percent of direct and indirect cost.
                             B-8

-------
     Direct cost = $7.50/ft2 (3000 ft2) = $22,500.
The material portion is $8700.
H.    REFRACTORY
     The refractory is $9200 for the ductwork and $11,000
for the quench tower.  The total installed cost is $20,200
The material portion is $5100.
                             B-9

-------
                  ESTIMATE OF ANNUAL COSTS

A.   ELECTRICITY
     Electricity is required to operate the shaker motors,
fan, pump, conveyor, and lights.  Total kilowatt hours are
computed as follows :
     Item           Horsepower     Kilowatts      MWh
Fan system              35            26          156
Pump system             15            11           66
Screw conveyor           5             4           24
Shaker motors            7             5           30
Lighting                              10           30
          Total                                   306
     The price of electricity in this area is 5 cents/kWh.
Therefore the total annual cost of electricity is:
     306,000 kWh/yr ($0.50/kWh) = $15,300
B .   WATER
     Makeup water is required for the gas cooling system at
the rate of 1.9 liters/s  (30 gpm) .  The price of water in
the area is 6.6 cents/m   ($0.2
total annual cost of water is:
the area is 6.6 cents/m  ($0.25/10  gal).  Therefore the
                                            *0.066/m3 - "700
C.   SOLID WASTE DISPOSAL
     A total of 490 Mg/yr (540 tons/yr)  of collected dust
will be disposed of by landfilling in sealed barrels.  This
will cost $95 /Mg, including labor, shipment, barrels, and
disposal charge at the landfill.  Therefore the annual cost
of solid waste disposal is:
     490 Mg/yr ($95 }  = $46,600.

                              B-10

-------
D.    LABOR
     It was estimated that about 2000 hours of labor would
be required to operate the entire system.  Labor costs
$10/man-hour.  Therefore the annual labor cost is:
     2000 h/yr ($10/man-hour) = $20,000.
E.    MAINTENANCE
     Labor and materials, including bag replacement, are
estimated at 6 percent of the capital costs.  An additional
15 percent of the labor and material is added to cover
supplies.  The total maintenance expense is:
     Labor and materials 0.06 ($316,000) = $18,960
     Supplies            0.15 ($ 18,960) =   2,840
          Total                          = $21,800
F.  FIXED COSTS
     Annual capital charges are estimated on the basis of a
10-year equipment life and a compound annual interest rate
on the capital of 10 percent.  The capital recovery factory
is 0.16275.  Therefore the capital charges are:
     0.16275 ($316,000)  = $51,400
     Property taxes, insurance,  overhead, and other admin-
istrative costs amount to 4 percent of the capital investment,
or 0.04  ($316,000)  = $12,600.
                             B-ll

-------
              SUMMARY OF ESTIMATED ANNUAL COSTS






Direct operating costs



  A. Electricity                        $ 15,300



  B. Water                                 2,700



  C. Solid waste disposal                 46,600



  D. Labor                                20,000




  E. Maintenance                          21,300



         Total direct operating costs = $106,000




Fixed costs



  F. Capital charges                    $ 51,400



     Property taxes, insurance, etc.      12,600



          Total fixed costs           = $ 64,000



          Direct operating            = $106,000



          Fixed costs                 =   64,000



     Total annualized costs           = $170,000
                              B-12

-------
                                   TECHNICAL REPORT DATA
                            jl'ti'UM- read //;w/i/< Han* mi ilu- >
-------

-------