United States
         Environmental Protection
         Agency  	
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-453/R-97-010C
May 1997
         Air
5EPA    Hazardous Air Pollutant
         Emissions from the Production
         of Polyether Polyols-
         Supplementary Information Document
         for Proposed Standards
                             U.S. Environmental Protection Agency
                                 5 Library (RH2J)

-------
                    HAZARDOUS AIR POLLUTANT
                EMISSIONS FROM THE PRODUCTION
                      OF POLYETHER POLYOLS
Ni
                  Supplementary Information Document
                          for Proposed Standards
                           U.S. Environmental Protection Agency
                            egion 5, Library (PL-12J)
                           Emission Standards Division
                  U.S. ENVIRONMENTAL PROTECTION AGENCY
                           Office of Air and Radiation
                     Office of Air Quality Planning and Standards
                     Research Triangle Park, North Carolina  27711

                                  May 1997

-------
                                   DISCLAIMER

       This report has been reviewed by the Emission Standards Division of the Office of Air
Quality Planning and Standards, EPA, and has been approved for publication.  Mention of
trade names or commercial products is not intended to constitute endorsement or
recommendation for use.
                                          11

-------
                    ENVIRONMENTAL PROTECTION AGENCY
Hazardous Air Pollutant Emissions from the Production of Polyether Polyols - Supplementary
                     Information Document for Proposed Standards
1.     The standards regulate hazardous air pollutant emissions from the production of
      polyether polyols.  Only polyether polyols production facilities that are part of major
      sources under Section 112(d) of the Clean Air Act (Act) will be regulated.

2.     For additional information contact:

      Mr. David Svendsgaard
      Organic Chemicals Group
      U.S. Environmental Protection Agency (MD-13)
      Research Triangle  Park, NC 27711
      Telephone:  (919)  541-2380

3.     Paper copies of this document may be obtained from:

      U.S. Environmental Protection Agency Library (MD-36)
      Research Triangle  Park, NC 27711
      Telephone:  (919)  541-2777

      National Technical Information Service (NTIS)
      5285 Port Royal road
      Springfield, VA 22161
      Telephone:  (703)  487-4650
                                         111

-------
                                    OVERVIEW



      This Supplementary Information Document (SID) contains memoranda providing

rationale and information used to develop the Polyether Polyols proposal package.  These

memoranda were written by EC/R Incorporated under contract to the U.S. Environmental

Protection Agency (EPA). The data and information contained in these memoranda were

obtained through literature searches, industry meetings, plant visits, and replies to

questionnaires sent to industry.

      The memoranda included in this SID are referred to in the Basis and Purpose

Document and in the preamble to the proposed rule.  These memoranda were compiled into

this single document to allow interested parties more convenient access to this information.

The memoranda included herein are also available from the docket (Docket A-96-38).

      The memoranda included in this SID are listed below with their document numbers.


   Document No.                              Description


 II-B-3             Caldwell, M., Caldwell Environmental, and Seaman, J., EC/R
                   Incorporated, to Svendsgaard, D., U.S. Environmental Protection
                   Agency.  March 26, 1997.  Memorandum.  Polyether Polyols
                   Production Industry Description.

 II-B-2             Seaman, J., EC/R Incorporated, to Svendsgaard, D., U.S.
                   Environmental Protection Agency.  February 20, 1997.
                   Memorandum. Maximum Achievable Control Technology Floor
                   Determination and Regulatory Alternative Consideration.

-------
  Document No.                              Description


II-B-5              Seaman, J., EC/R Incorporated, to Svendsgaard, D., U.S.
                   Environmental Protection Agency.  April 29, 1997. Memorandum.
                   Summary of Nationwide Baseline Emissions for Polyether Polyols
                   Production Facilities.

II-B-4              Hendricks, D. and Seaman, J.,  EC/R Incorporated, to Svendsgaard,
                   D., U.S. Environmental Protection Agency.  April 25, 1997.
                   Memorandum. Estimated Impacts for the Polyols NESHAP.

II-B-7              Seaman, J., EC/R Incorporated, to Svendsgaard, D., U.S.
                   Environmental Protection Agency.  May 16, 1997. Memorandum.
                   The Definition of an Extended Cookout as a Control Technique for the
                   Polyether Polyols Production Industry.

II-B-17            Chappell, L., U.S. Environmental Protection Agency, to Svendsgaard,
                   D., U.S. Environmental Protection Agency.  May 21, 1997.
                   Economic Analysis for the National Emissions Standards for
                   Hazardous Air Pollutants (NESHAP) for the Polyether Polyols
                   Industry.

-------
                                                                 A-96-38
                                                                 H-B-3
JCt '/R' Incorporated	7996 EPA Outstanding Small Business Contractor
    MEMORANDUM

    Date:     March 26, 1997

    Subject:  Polyether Polyols  Production Industry Characterization

    To:       David Svendsgaard,  EPA/QAQPS/OCG

    From:  '   Mary-Jo Caldwell,  Caldwell Environmental,  Inc.
              Joanne Seaman, EC/R Inc.
         The purpose of this memorandum is^ to describe the polyether
    polyols manufacturing source  category as it  is  being defined by
    the U.S. Environmental Protection .Agency (EPA)  for purposes of
    developing a national emission standard for  hazardous air
    pollutants (NESHAP) .   In this memorandum, the source category is
    described, both existing and  projected facilities are discussed,
    a brief discussion of the end uses of polyether polyols is
    provided, and the polyether p.olyol manufacturing process,
    emissions, and controls are .described.

    INTRODUCTION     .              .

         For regulatory development purposes, the EPA has defined the
    polyether polyol•source category as facilities  that, manufacture
    polyether polyols,  which are  defined.as follows:

         Polyether polyols are defined as the products formed by the'
         reaction of ethylene oxide (EO),  propylerie oxide 
-------
Additionally,  the first group performs the reaction  on a batch
basis, while  the second group reacts on a continuous basis.

MANUFACTURING FACILITY DISTRIBUTION AND DESCRIPTION

     The distribution and description of existing polyether
polyol manufacturing facilities is discussed  in this section. SPI
representatives report that no new plants are expected to be
built within  the next five years,  but modifications  can make an
existing source become a new source.

Existing Facilities

     The trade association for polyether polyol manufacturers,
the Society of the .Plastics Industry (SPI) , and the  EPA project
team worked together to gather information on the polyether
polyol manufacturing industry.  The SPI supplied to  the EPA
project team  a list of plants that produce polyether polyols in
the United  States, and this list is provided  in Table 1.   Two
more facilities were added to this list; one  facility sent a
"questionnaire response to the EPA, and the second facility was
listed in the Chemical Economics Handbook (CEH).  The EPA project
team determined that,  of the list of 84 plants, there were only
two facilities in the list that polymerize THF exclusively, and
the remaining 82 produced polyols with epoxides.  One of the
facilities  indicated that although they produce polyols with
epoxides, they also have a THF polymerization line.   Therefore,
the project team tentatively estimated that the source category
consists of three facilities that polymerize  THF and 82
facilities  that polymerize epoxides into polyether polyols.2

     The types of polyether polyols produced  at each of the 84
facilities  are also presented in Table I3.  In Table  1, polyether
polyols are broken down according to application into three main
categories: urethane,  non-urethane, and surfactants.   As shown in
Table 1, some facilities produce only one type of polyether
polyol and  others produce several.  Even at those facilities that
only produce  one type of polyether polyol,  multiple  polyols
within the  general classification are often produced.  Sixty five
percent of  the facilities in the database manufacture a
surfactant, and 20 percent of the facilities  reported urethane
production  while 35 percent reported non-urethane production.3

Projected Facilities

     The number of facilities that use epoxides to produce
polyether polyols that are expected to be major sources subject
to the rule is 76.  This number was calculated by first
determining the percentage of the facilities  in the  database that
are major sources  (92.3 percent) .  This percentage was multiplied
by the total  number of facilities in the database that use
epoxides in Table 1 (84 facilities), to estimate the number of

-------
facilities  using epoxides  to produce polyether polyols that are
expected to be  subject  to  the rule.  The three facilities that
use THF  to  produce  polyether polyols are all expected to be major
sources  subject to  the  rule.   Because one of the facilities that
uses THF also produces  polyether polyols  using epoxides,  the total
number of facilities  expected to be subject to the rule is
estimated as 78 (76+2).

POLYETHER POLYOLS PROPERTIES AND END USES

     The end use of a polyether polyol is determined by the
properties  of the polymer.   The linear or slightly branched
polyether polyols serve in flexible applications, such as in
flexible slab and molded foam,  reaction injection molding,  and in
other elastomer,  sealant,  and coating applications.  The branched
polyether polyols serve in applications requiring rigidity such
as rigid foams,  solid or microcellular plastic, and hard,
solvent-resistant coatings.  As noted in the discussion of Table
1, polyether polyols  fall  into two main classifications:  high-
molecular-weight, linear or slightly branched polyether polyols
(urethanes), and low-molecular-weight, highly branched polyether
polyols  (non-urethanes).4

Urethanes

     Properties.  There are three important factors that
determine the properties of an polyol product:  its
functionality,  its  average molecular weight, and its degree of
steric hindrance.

     The functionality  of  the polyether polyol is determined by
the functionality of  the initiator, and is defined by the number
of terminal hydroxyl  groups on each polymer molecule.  The number
of terminal hydroxyl  groups influences the degree to which
urethanes produced  from the polyol will be cross-linked; that is,
that larger the number  of  terminal hydroxyl groups, the more
cross-linking.   Therefore,  the functionality of the polyol is
important in determining the  properties of the ultimate
polyurethanes. 5

     The molecular  weight  is  controlled by the total amount of
monomer  fed to  the  reaction,  and the amount of initiator used.  A
high ratio  of monomer to initiator results in a comparably high
average  molecular weight.   Higher molecular weights also require
longer reaction times.   The average molecular weight and
functionality of an polyol determine its "hydroxyl  number," which
is the number of terminal  hydroxyl groups per unit weight.5

-------
4-4
 a;
 o
 3
73
 O
 ^j.
Cu

 tn
4-1
 U
 3
•o
 o
T3
 C
 CO
,— <
 O
 o
Qu

 i-i
 Q)
x:
-P
 a>
 o
Cu

 cn
 C
-H
 U
 3
•o
 O
 (U
•H

4-)
-H
 O
 (0
 CU
.—I
JQ
•o
o ta
jj jJ
IB 0 C
i-4 -H Id
>i C JJ
x o o
O -H IB
O. 1 t O U
X -H id
O 1 ^

Eb
(B
JJ
fl
cn




























































u
K

K
Q























X



























M
U
M





tt)

^j
m
to
O
d
S





























































^<
H
OS
o























X














X





o
C
0)
3
O
CU
1
0)
o
K









(U
T3
C
s





























































A
H
a
a:
o
ta
C3












X





































n
m
a
0)
jj
CO









01
T3
C
-rn
s




























































01
H
§
H
iJ
H























X














X










0)
C
(U
01
1-1
m
o









0)
•n
c























X


























c
o
JJ
i







•a
o
i
01
CX






































X











C
o
Ll
I— 1
(U
Oi









0]
c























X














X












o
a.









a
c
s-
c












X






























X






c
n
a
jj
CO









TJ
O
c
5


-------
 OJ
EC
 tl)
 U
 a
•a
 o
 S-l
cu

 to
-U
 u
 D
T5
 O
•a
 c
 to
 O
a>
 0)
J3
4-t
 C
•H
 U
 =3
T3
 O
 ^
a,

 w
 0)
•H
4-)
 u
CO

D
 0)
1—(
X}
•o
41 ca
a o a
i-l -rt 18
£ 0 0
O -rt *
Q, i . 0 U
X -H «
O 1 • o u
X -H «
O 1 -<
0 3
43 0
a: a.



cu

g
•H
i-l
m
CQ






















































1
o
H
X
o
I-l
s











































X



b
en
CQ



(U
JJ
JJ
o
T!
C
(t
s

-------
  a
<0 U C
•H -H 18
X 0 U
O i-l •, 0 U
X -H (d
O 1 
Z









X









X
























1-1
o
0
cu
E






ison
•a









X









X

















X









b
co




c
o
en
-H
ca

r-t
(U
•H









X




































c
Q
a)
CO





73
O
O
id






















































tt
a
o
X
s
ta



















X


























4)
JJ
X
11)
a






X
c
o
a











































X


c
o
Oi
3
Q
en
,— i
f—i
id
Cu
as
ns
Cn
IB
-H
S

















































z
H
a
o
OS
s
H
i











































X

c
03
JJ
a
1C



O
J-
C
eens
c

-------
 0)
OS
•o
 a>
 u
 D
•o
 o
 >-l
a,

 CO
4-1
 U
 3
T3
 O
 S-l
CU

T3
 C
 CO
•— i
 o
 O
Cu
 Q)
x:
JJ
 0)
 O
cu
 o
 3
•a
 o
 s-i
Cu

 to
 OJ
-H
 O
 to
 fO
E-
TJ
o a
jj jj
« O C
r-l -H Id
>! C JJ
XO o
O -H id
a ' , 0 0
X -H *
O 1 IU.
a a n
JJ O 3
H a M
•o n
O JJ
jj u c
« -H id
•-» C JJ
SN O 0
X -H n)
O i «w
M a n
r-l O 3
«« a co
































o
c
m
c o
0 H
ss a
ID
c
0)
kl
D




^
c
«J

6
o
u

^
4J
•H
O

•v.

ID
JJ
m
TO












































4J
C
o
^r
i
a
•r4
•H
m



o
J_J
0

M
c
(U
01
0




















X






















Ol
u_t
s

o
0
01
1)
i-l
o


0)
1— 1
I-l
-H
J>
CO
m
0
J3




















X















X










r-l

y
c
0)
K




(U
JJ
JJ
o
f-l
M
m
u




















X






















JJ
c

o
o<

rj
Ol
>p^
EC



JJ
c
-r<
o
Cu

si
Ol
-I-l




















X

























J t
0]

0
0)
o
as


^
p.)
r-l
o
X

JJ
c
o


























































f*
*3
o

§




































X











^J
dl

m
0)

D]
Ol
c

^JJ
a
w

c
(fl
w






















































1^
H

1^
^
J
X
D]
Ed
OK










X









X



























rC
T5
o

u




1— 1
I—I
m
K

1— 1
i-H
•H




















X


























^

-------
 ro
 (U
 a:
 T3
 cu
 u
 O

 r-l

 Ol
 4-1

 U

 3

 73
 O
 r-l
•Ol
 C
 03

 CO
 rH
 o
 >t
 1— I
 o
 a,

 M
 CU
 x:
 o
 CU
 •r-l
 O
 3
 •o
 o
 i-l
 to
 CU
 u
 to
 cn
  CU
  fO
•a
o n
jj JJ
id u C
i-l -rl 10
>, C JJ
X 0 0
0 -H 10
0* c: tJ
rl 0 3
0< S5 01
73 a
0 JJ
jj u c
18 -rl 10
r-l fl JJ
>i 0 0
X -^ 10
0 1 «W
JJ 0 3
H Z 01
73 a
ID JJ
jj u c
id ft a
r-l C JJ
>i 0 0
X-rf 18
0 1 «W
^ C rl
r-l 0 3
«< Z ca


































o
c
81
1 JJ
C ID
0 M
S 3
O
c
10
A
JJ
O
D






^,
a
a
§*
0
o

-H
o

-X

ID
JJ
10
JJ
W


















X




























fl
0)
w*
•H
r-l
•H
2






C
(8
c
M



































X














u
(U
^
o





>1

C
m
jj
m
a


















X






























N
O
T>
C

W





a
-H
JJ
(8
























































m
u
w
w
H

H









































X

m
r-l
«
u

e
0)
£
u
o
a




m
-H
JS
a
£
01




























































01
X
H
H









































X









8

9


O









































X




J^
0)

c
0)
a
m
u



(8
C
0)
-a
18
it
a,



































X





X










o
Q



jj
^
O
a
(U
u
£









































X






C
m
E
jj
ca
id
a






-------
 cu
05
 u
 3
-o
 o
 M
a<

 w
4J
 U
 3
•o
 O
 S-I
a,

TS
 c
 (0
 O
cu
 OJ
 o
cu
 u
 a
TJ
 O
 ^
CU
 cu
•H
-P
 U
 fO
en

o


<—i

 0)
1—I
J3
 (T3
•o
o a
« o c
r-( -H B>
S C -u
xo o
O -H «
0, i «w
0 C t4
M 0 3
0. Z M
•a a
O X)
-u o c
(8 -rl «
t-4 C -U
SOU
X -H IS
0 ' 
ID
M
D




^,
c

Oi
3
o
u



^
^J
•rf
O

•x^

o
X)
a)
XI
01




































X










0)

-r(
r-l
o
^4
4J
<1)
O4







ttJ
c


•rt
s
0)

f^
•rt
^
0}
C
•rt
4-1
^1
m
s

s
0)
z




















X









X





X








0)
•a
•g
1C
u

c
o
-rt
c
D






(U
4J
3
.U
•rt
U
CO
C
M




































X








cu
T3
•g
flj
O

C
o
•H
c


0
m
(U
^
^4
m
t^
a

4;
4-1
3
0
w






























































Z
H
01
Z
o
CJ
01
H
S




















X





























u

H
H
rn
^






-------
      The degree of steric  hindrance of an polyol governs  its
 reactivity in the final  urethane polymerization reaction.   Steric
 hindrance,  or reactivity,  is determined by the relative amounts
 of  EO and PO used in the reaction, and by the timing of their
 use.   The steric effect  is most important at the end of the
 polyol chain.   When PO is  the last monomer added to a  chain,  the
 terminal hydroxyl group  is termed a "secondary" hydroxyl,  because
 of  the presence of the methyl side group.  When EO is  the last
 monomer added,  the terminal hydroxyl group is "primary" .   A
 primary hydroxyl group at  the end of an polyol chain reacts more
 quickly than a secondary group, because the secondary  hydroxyl is
 sterically hindered by the methyl side group.  Thus, the
 resulting polyol is more reactive.5

      Polyether polyols for polyurethane use must fulfill  some
 strict purity criteria.  Traces of catalyst residue or salts  can
 cause problems in the end-use.  Catalyst residue concentration
 must  usually be below 15 parts per million (ppm) .  Other
 impurities that may have a significant effect at a very low level
 are dissolved gases,  oxidation products, and materials that
'affect surface tension.  Because of the need to control these
 effects,  the polyether polyol manufacturing process has evolved
 to  a  high level of complexity to assure product quality and
 consistency."

      The most important  nonEO/nonPO polyether polyol is
 polytetramethylene glycol (PTMEG) , derived from tetrahydrofuran.4
 The high-performance elastomers made from PTMEG are superior  to
 castable polyurethane elastomers in their resilience,  hydrolytic
 stability,  thermal stability, ultraviolet-light resistance,
 compression set, and adhesion characteristics; and they are at
 least comparable to polyurethane elastomers in abrasion
 resistance.   However, they are poorer with respect to  ultimate
 tensile-elongation values, tear strength, and solvent
 resistance.6

      End Uses.   Most polyether polyols are produced for
 polyurethane application.4'7  The largest world market  of
 polyether polyols for urethane applications is in the  production
 of  polyurethane foam, primarily flexible foams, and in lesser
 quantities,  for rigid foams.  The remainder of the polyether
 polyols for urethane applications are used in a variety of
 markets,  including microcellular, cast, and thermoplastic
 elastomers;  adhesives and  sealants; surface coatings and
 polyurethane fibers.7    Polyurethane polyols are also  used in
 some  instances to produce  reaction injection molding plastics,
 which are used mainly in automotive applications.5

      Polytetrahydrofuran can be prepared so that it has  reactive
 end-groups (e.g., hydroxyl or amino), or relatively inert end
 groups (e.g.,  alkoxy).   The most commercially important
 polytetrahydrofurans are probably the hydroxy-terminated

                                10

-------
polymers,  i.e., the PTMEG, which are used in injection molding
and extrusions and to prepare nonfiber polyurethanes,  spandex,
and polyesters.8'9   Polytetrahydrofuran can'also  be prepared so
that it has a preponderance of primary amino end  groups.   The
resulting  elastomers display a high degree of resilience,
abrasion resistance, and insensitivity to water.8

     According to a report by Stanford Research Institute (SRI)
International, in the beginning of 1994,  the world capacity for
polyether  polyols for urethanes was approximately 8.5 billion
pounds  (3.9 million metric tons).  The United States accounted
for 34 percent of the total capacity.   Actual production in
Western Europe, the United States, Japan,  the Republic of Korea,
Canada, Taiwan and Mexico in 1992 was approximately 5.2 billion
pounds, representing a 1992 operating rate of about 69 percent in
those regions.  The projected consumption growth  rate for
urethane polyether polyols in the United States is 2.7 percent
for 1992-1997.7

Nonure thanes

     Properties.  Polyether polyols for nonurethane applications
are high-molecular-weight polyols that are nonvolatile and range
from viscous fluids to waxy solids at room temperature.
Nonurethane polyols are used safely in many personal-care
applications because they possess a very low degree of oral
toxicity,  are relatively nonirritating to the skin, and have low
potential  for eye injury.  Nonurethane polyether  polyols can also
be thermoplastic.8

     The pseudo-plastic behavior of nonurethane polyether polyols
is used to impart leveling and flowability to waterborne coatings
and helps  to build a film and reduces spatter and misting in
water-based lubricants.  Nonurethane polyether polyols are used
as a dispersive medium for a variety of applications;  they can
serve as a wetting agent and a thickening agent,  allowing the
cleanser or lubricant to cling to the surface.  At
concentrations higher than normal for friction reduction, they
produce a  coherent needlelike stream from a normal spray which
can be used to cut semisoft solids such as cardboard,  leather,
and cheese.8

     End Uses.  The major markets for nonurethane polyether
polyols are surfactants, functional fluids,  lubricants, and
personal-care products.  Sizeable quantities of nonurethane
polyether  polyols are also used as chemical  intermediates.9  ,
According  to a report by SRI International,  in 1994, consumption
of nonurethane polyether polyols in the United States was as
follows: surface-active agents 38 percent; functional fluids and
lubricants 28 percent; other (animal health products,  herbicides,
rubber chemicals, cosmetics and toiletries,  mold  release agents,
and adhesive formulations) 34 percent.6    In  1994,  nonurethane

                                11

-------
polyether polyol production in the United States, Western Europe,
and Japan was approximately 1,050 million pounds  (475,000 metric
tons), according to an SRI International report.  In 1994-1997,
the United States is expected to exhibit consumption growth of
approximately 2.7 percent per year.6

PROCESS DESCRIPTION

     A description of the manufacturing process and chemistry
used to produce polyether polyols from epoxides and from THF are
presented in this section.

Polyols from Epoxides

     Polyether polyols are produced by the addition
polymerization of cyclic ethers.  Most polyether polyols are
produced by base catalysis.  Exceptions are polyether polyols
based on amine initiators which are self-catalyzed or require
additional catalysts only in the later stages of the process, and
polytetrahydrofuran which is produced by acid catalysis.4

     A generalized block diagram for polyether polyol processing
is shown in Figure 1.  The polymerization reaction takes place in
a reactor made of carbon or stainless steel, and begins with an
initiator, which is a starter polyol, amine, glycerin, or other
material with one or more reactive hydrogen sites.  The starting
material used will depend on what functionality is sought in the
final product.4'5

     A catalyst is added to the reactor containing the initiator;
potassium hydroxide or sodium hydroxide are typically used, but
boron trifluoride (BF3) ,  and other acids and bases are also used.
The catalyst is used to activate the initiator by abstracting the
active hydrogen to produce an anion.  Then the monomer feed is
started.  During the formation of the alkoxide ion and the
addition of the monomer, oxygen needs to be rigorously excluded
because the alkoxide ion is very sensitive to oxidation.  The
oxygen is purged from the vessel by replacing any air in the
vessel with nitrogen.4'5

     Once the monomer feed is started, the anion reacts with the
cyclic ether, cleaving the ether ring.  Cleavage of the ether
ring leaves an oxygen ion at the end of the new chain, which
reacts with another molecule of monomer.  This process continues
until the desired molecular weight is reached and then the
monomer feed is terminated.  In some cases, if a higher molecular
weight product is required, a facility will start reacting a
small amount of epoxide  (known as an initiator) in a  small
reactor.  Once a certain molecular weight is achieved, the
material is transferred to the main reactor.  The initiation and
polymerization reactions are illustrated in Figure 2.5  The


                                12

-------
monomers most commonly used to produce  polyether polyols are PO
and EO  (called  "epoxides") .4'5

     During the addition of the monomer,  the heat of reaction is
very high, and heat removal is necessary and is accomplished by
external or internal heat  exchange.   Extensive process control
and instrumentation are required to operate safely within a given
temperature range in order to produce consistent products.  The
monomer feed rate may be limited by heat-removal capacity or
pressure limitations.  The reaction rate depends on catalyst
concentration and reaction temperature,  as well as the type of
monomer feed.  The reaction temperature not only affects the
reaction rate, but also the composition of the polyether polyol
produced.4

     After the desired molecular weight is reached and the
monomer feed stops, the product may have subsequent refining or
undergo other reactions to make the final product.  In polyol
manufacturing, refining usually refers  to catalyst removal from
the polyol and, if a solvent is used, removal of the solvent from
the polyol prior to sale.  The degree  of catalyst removal from the
crude reactor product depends on the  application and can range
from simple neutralization for polyether polyols used in rigid
foams to removal to less than S ppm potassium for most other
applications, where deleterious effects of salts have been shown.
The choice of the catalyst removal process depends on the
required purity of the polyether polyol and may involve
crystallization and filtration of salts and/or treatment with
absorbents.  The catalyst  can either  be neutralized with an acid
or stripped with a solvent.  If an acid is used, a solid waste
results which must be treated.  If a  solvent is used, solid waste
is avoided and the solvent is often recovered and recycled.4'5
The catalyst separation solvents typically used are the organic
HAPs toluene or hexane.

     In the final purification step,  water and volatile organic
compounds are removed by stripping in the reactor or in falling
or wiped film evaporators.  The polyether polyols are stabilized
with proprietary inhibitors  (antioxidants), saturated with
nitrogen, and stored hot.4

     Polyether polyol manufacturing industry representatives
indicated that organic HAPs other than  EO and PO are used for
more than just obtaining the purity needed.  Reactivity,
viscosity (which affects mixing), solubility  (in water versus
organic solvents), and stability  (against oxidative and thermal
degradation), are all important factors in manufacturing
polyether polyols to realize needed end use properties, and other
organic HAPs are necessary in achieving and maintaining the
desired levels of each of  these characteristics.10
                                13

-------
              initiator^
              catalyst->
        >>water
                acid
              solvent
             absorbent^
Refining
           > excess acid
         -> salts
         ^absorbants

          .water
       > solvent
           volatiles
        antioxidants
               N
                 2
stabilization
storage
 Figure 1. Process Flow Diagram for Poly ether Alcohol Manufacturing (Ref. 4)
                              14

-------
           R-OH
         INITIATOR
          ALCOHOL
           R-0'
           AN I ON
Alkaline
catalyst

(KOH OR
NaOH)
  / \
H2C	CH
    I
    R1
                          (CYCLIC
                          ETHER
                          MONOMER)
   0
  / \
H2C	CH
                                                                  (1)
                                         R-O-
                R-0-«2-CH-0-
                     I
                     R1
                                                                  (2)
                R-f 0-H2-CH
      R-/0-H2-CH

       \    I
           R1
      NOTE:   R-OH in equation (1} can be replaced with an amine, and the epoxide monomer in
            equation (2) can be replaced with a larger cyclic ether ring, such as
            tetrahydrofuran.

      FIGURE 2. Reactions in Polvether Polyol Production      	
(3)
Polyols made from  Tetrahydrofuran

      The polymerization of  THF involves cleaving  the THF molecule
by using an acid catalyst,  and then  reacting the  open THF with
unreacted  THF to build a longer chain.   The  basic requirement  for
polymerization is  that a THF molecule  must be broken by some
mechanism;  assuming that a  suitable  counter-ion is present,
polymerization follows.  Once initiated,  propagation occurs as a
result of  collisions between the THF monomer and  the broken THF's
chain end.8

      The polymerization of  THF is represented by  the following
scheme8:
                          Initiator-i
                                    15

-------
     The synthesis of polytetrahydrofuran involves the use of
strong acid initiators.   Therefore, the polymer is stable against
attack by bases,  but somewhat acid sensitive.  Like all polyether
polyols, it is subject to oxidative attack and subsequent
breakdown.  The addition of common antioxidants (e.g.. amines
orpyrocatechol) ,  inhibits these reactions and imparts adequate
stability to the polymer for commercial purposes.8

     Because THF polymerizes only by an ionic mechanism,
requirements for successful polymerization include pure, dry
reagents, dry apparatus,  and either high vacuum or an inert
atmosphere (e.g.,  dry nitrogen).  Use of a suitable temperature
and a limited amount of solvent, depending on the temperature,
are also critical factors,  because the polymerization of THF is
an equilibrium polymerization.8

     The most commercially important polytetrahydrofuran is
PTMEG.  The traditional,  as well as the simplest and most direct,
method of preparing PTMEG is to initiate the polymerization of
THF with protonic acid,  which leads to the formation of an -OH
head group and termination of the polymerization with water,
which also produces an -OH end group.   This  process is shown by
the following:
EMISSION SOURCES AND CONTROL

     There are four primary potential sources of organic HAP
emissions in the polyether polyol manufacturing process: process
vents, storage vessels,  equipment leaks, and wastewater.  The
emissions from these sources, and the types of controls being
used, are discussed in the following sections.

     Another potential emission source considered  in  other
NESHAPs is unloading operations, and these were investigated for
this industry as well.  Industry representatives indicated that
all organic HAP unloading operations are conducted in a closed
vent system due to the explosive nature of EO.  Therefore, no
emissions are expected to result from this portion of the
process.2

Emissions and Controls on Production Units that Use Epoxides

     Process Vents.   The most significant organic  HAP loss from
the polyether polyol manufacturing process occurs  at  the end of
the batch when the unreacted epoxide is vented.  Ethylene oxide
reacts quite quickly, so normally the residual EO  in  the vent is
quite low.  Propylene oxide reacts an order of magnitude slower
at a given catalyst level, temperature, and pressure.

                                16

-------
Consequently, the uncontrolled level of PO in the vent is usually
substantially higher  than EO.10

     Information on process vent emissions was obtained via a
questionnaire.  The EPA/SPI questionnaire was written by the EPA
project team with input  from industry representatives, and was
then submitted to SPI members and nonmembers.  Additional process
vent data, limited to annual emissions and control efficiencies,
was received from the SPI.  The process vent emissions and
control device data for  28 facilities, with the facility names
coded, are presented  in  Table 2.1   The  emissions  data in Table 2
indicate that combined EO and PO controlled emissions range for
these select facilities  from 20 to 80,966 pounds per year
(Ib/year) .  As can be seen in Table 2, a variety of controls are
used to reduce process vent emissions.  The process vent controls
used by the 28 facilities, listed in order of frequency of use,
are: scrubbers, flares,  extended cookout, incinerators, and
condensers.  Over 40  percent of the facilities that submitted
control device data used scrubbers to control process vent
emissions, and 25 percent used flares.1   Some facilities  using
scrubbers use a condenser upstream of the scrubber to recycle the
epoxides back into the batch being processed.  The unreacted
epoxides not condensed are converted to glycols in the scrubber
by using an acid or base medium.  This glycol is often sold as a
byproduct.3

     Some facilities  use flares or incineration to reduce process
vent organic HAP emissions.  Because the polyether polyol
manufacturing process using epoxides is a batch process,
incineration is not always the most economical control
alternative, according to industry representatives.  The industry
representatives reported that the batches can take anywhere from
12 to 36 hours, and this would mean that if a combustion device
were used, there is a time period for each batch when the
incinerator or combustion device would simply burn natural gas or
fuel oil to keep the  combustion device in operation.3   It should
be noted that incinerators are used at two of the facilities in
the control device database.:

     Extended cookout (ECO)  is also used to reduce process vent
organic HAP emissions from polyether polyol production that occur
when the reactor is opened.  Extended cookout reduces the amount
of unreacted EO and/or PO (epoxides) in the reactor by allowing
it to react for a longer time period, thereby reducing epoxides
emissions that may have  otherwise been routed to a control
device.  Because the  epoxides  are converted to product,  the
producer's raw material  usage decreases.  The point in time where
the costs for emissions  reduction resulting from a decreased
production rate exceed the benefits from improved raw material
efficiency is the starting point for an ECO.  This point was
determined based on average raw material costs and "typical"
product profit margins.  The starting point for calculating

                                17

-------
emission reduction from ECO is when the amount of epoxides in the
reactor liquid has decreased to  25  percent of the amount in the
vessel after the epoxides addition  phase.

     The use of ECO as an emission  control technology has an
economic penalty,  since ECO must be done in the reactor system,
lengthening batch time and reducing the reactor system annual
production capacity.   This penalty  is offset to some extent by
the improvement in raw material  efficiency due to converting more
epoxide to product.   Once ECO is complete,  the remaining
emissions are low.9

     Storage.  The storage vessels  used in the polyether polyol
manufacturing process are a potential source of organic HAP
emissions.  Most of the storage  vessels used in polyether polyol
manufacturing are pressurized, and  of this group all of the EO
and PO storage vessels are pressurized as  a precaution due to the
explosive nature of EO.2

     Equipment Leaks.  Emissions of organic HAPs may occur around
pumps, valves,  and other components,  and are referred to as
"equipment leaks".  In some facilities,  pumps may have double
mechanical seals,  with a fluid such as propylene glycol as a seal
barrier fluid.   In addition,  69  percent of the facilities have
leak detection and repair (LDAR)  programs.   The LDAR programs
allow the facilities  to be aware of any leaks immediately so the
problem can be fixed, minimizing emissions.5

     Wastewater.  Another potential source of organic HAP
emissions in polyether polyol production facilities is the
wastewater from the manufacturing process.   The majority of
wastewater emissions  from this industry are a result of the steam
ejectors used on these batch reactors to pull a vacuum on the
vessel.  However,  although seventeen percent of the facilities in
the database have Group 1 wastewater streams that require
controls according to the HON, none of the facilities for which
data were received control air emissions from the wastewater
streams.5  A Group 1 wastewater  stream is defined by the HON as
having a minimum volatile organic HAP (VOHAP) concentration of
10,000 ppm and a minimum stream  flowrate of 10 liters per minute.

Emissions and Controls on Production Units that use
Tetrahydrofuran

     Even though THF is not an organic HAP, there are organic HAP
emissions from one of the facilities in the database that
polymerize THF.  At this facility the catalyst extraction process
resulted in organic HAP emissions.2   An  organic HAP solvent is
                               18

-------
14-1
 Q)
 m
 cu
•H
 O
 (C
n
 0)
u
 u
 QJ
r-t
 V
Cfl
 CO
r-l
 O
 M
JJ
 C
 o
u


T3
 C
 (0

 03
 OJ
JJ
 (0
 E
-H
4J
 CO
 w
 c
 o
•H
 cn
 CQ
-H
CM


 QJ
r-l
XI
 to
ABATEMENT
CONTROL EFFICIENCY
(PERCENT REDUCTION)
(LB/YR)
AFTER CONTROL EMISSIONS
SMISSIONS (LB/YR)
:ONTROL I
BEFORE (
Other-
HAP
EO&PO
2
0
N
Other-
HAP
EO&PO
O
S
Other-
HAP
EO&PO
0
Cu
O
H
FACILITY
Incinerator |
2
cn
cn
cn
en
cn
cn
cn
cn
„.
f
r-
in

a
K
z
K
z
z
•£
Scrubber |
o
en
cn
cn
cn
o
r-
CN
rH
m
O
CN
cn
CN
H
H
O
P-
CN
CN
cn
vo
rH
in
CO
in
rH
O
a
M
CO
u
o
01
CO
cn
CO
cn
CO
cn
H
CO
in
m
vo
O
in
cn
m
H
o
H
co
vo
ro
10
VD
m
CN
in
o
(N
0
vo
r-

Scrubber |
Z
CO
cn
<
CO
2
o
fN
Z
o
CN
2
tt
a
tt
2
in
rH
a.
Extended
Cookout |
<
CO
cn
CO
cn
CO
cn
Z
cn
CO
rH

«
a,
z
CO
M
a.

rH
O
CN
r-l
CO
Cti
2
tt
2
«
2
2
r-
I-H
0|
CO
Extended
Cookout 1
,<
CO
01
^
CO
cn
2
in
ro
i-H
g
in
H



a
2
ID
HH
Ai
CO
Shell & tube
Condenser
o
in
on
in
cn
in
en
rH
in
rH
Cl
VO
CN
r-
P-
VO
H
vo
in


tt
2
g

M
ffl
O
<
S
m
cn
r-
cn
2
vo
r-l
in
(N
r-
rH
CO
m
CN
CO
rH
2
i-H
rH
cn
ro
vo
CO
rH
in
r*>
00
CO
cn
in

Extended
Cookout
o
rH
n
CO
CO
on
o
(N
CO
in
CO
cn
0
0
CO
cn
CO
in
X
Z
ca
z
01
Z
a
2
SPI13

-------
g
w
33
w
F*
n
rt




si
M C^
H CJ
0 O
H Q
Ib U
Ed OS
H
O CJ
S S
o «-
CJ


X
^x,
CO
£•


CD
§
H
CD
m
H
Aft

O
u
g
o
CJ

«
u
H
b.
^

X
09
~—
01
§
H
CQ
CQ
H

•
O

§

H
§
b
U
ffl
EH
j
H
CJ
tC.
b
3
f-4
O
U U
(U
73 Q
0) XI
.* 3
U M
m o
& co

<
0
ON



g
o
0\





z





Cti
z




o:
Z






ft!
Z


§

|




OS
z



z


^*
M
CU
CO



^|
U 0)
ra 3

m u
u to
0

in
00

in
CO




„.
Z
TT


vo
vo
en
0
00


vo
vo
a\
o
CO






rt

•*


t-i
ro
r-
vo
in
in

r**
j^

IO
in
in



g





"^



^j
(U
3

O
CO

z
in
CO



g
in
CO





Z
in
p^
vo
rH







g
in
r^
vo
r-(






§

Z




§



a
z

in
H
M
&.
CO



) >
^
0
0
o
u

z


1


g



Pi
S


z
en
vo
H




cn
vo
<-i








rf
Z


Z

g




Z



z





X

o

(0
^i
0)
-H
0
c
HI

g


06
z


a:
Z



Q!
Z


Z
o
^*
in




o
\r>
*f


o
cn








Z

a
z




Qi
z



z





o



^1
a]
XI
^
o
co

z


a
z


§



o:
z


z
vo
(-4
rH
H
nj o
a co

z
00

a\
01


g
CO
•
S



<<
Z
o
^*.
H








Z
o
r-
H







1

g




\,
Z



eti


rn
H
CU
CO
















































01
r-l
T3 X)
0) (Q
4J 0
J-4 -H
O r-l
O* OT

-------
used for the catalyst  extraction, and none of these emissions to
the atmosphere are controlled.2

     Process Vents.  The polymerization of THF occurs on a
continuous basis.  THF is  not an organic HAP and organic HAPs are
not typically used in  the  actual synthesis of the polymer.

     Stora,gg.  Only one of the THF polymerizers stored an organic
HAP.  The storage vessel was a Group 2  (not requiring controls
based on the HON criteria)  fixed roof storage vessel.

     Equipment Leaks.   Potential organic HAP emissions may occur
around pumps, valves,  and  other components in facilities
manufacturing polyether polyols from THF.  Based on data received
from industry, it appears  that monitoring of equipment leaks is
not generally conducted at such facilities.2

     Wastewater.  Neither  facility that polymerized THF generate
Group 1 wastewater streams.  Therefore, no wastewater controls
were reported.2

REFERENCES

1. Meeting Agenda, by  Seaman, J.C., EC/R, Inc. MACT Update.
July 11, 1996.

2.  Memorandum from J.  Seaman, EC/R Inc.,. to D. Svendsgaard. U.S.
EPA. February 20, 1997.  Maximum Achievable Control Technology
Floor Determination and Regulatory Alternative Consideration.

3.  Letter from M. Healey,  The Society of the Plastics Industry,
Inc., to D. Svendsgaard, U.S. EPA.  November 22,  1995.
Presumptive MACT for Polyether Polyols Production.

4.  Gum, W. ,  et. al. ,  Editor.  Reaction Polymers: Polyurethanes,
Epoxies, Unsaturated Polyesters, Phenolics, Special Monomers, and
Additives.  Hanser publishers.  1992.

5.  Memorandum from W.  Battye, EC/R Inc., to D. Svendsgaard, U.S.
EPA. October 24, 1993  (finalized January 7, 1997) .  Site Visit
Report, Dow Chemical Polyether Polyols Unit,  Freeport, Texas.

6.  Stanford Research  Institute  (SRI) Consulting.  "Chemical
Economics Handbook Abstract, Polyalkylene Glycols".  Chemical
Industries Newsletter  July-August 1995.  Online.   Available:
http://www.sri.com.  7 February 1997.

7.  Stanford Research  Institute  (SRI) Consulting.  "Chemical
Economics Handbook Abstract, Polyether Polyols for Urethanes".
Chemical Industries Newsletter Jan.-Feb. 1995.  Online.
Available: http://www.sri.com.  7 February 1997.


                                21

-------
E
                                                            A-96-38
                                                            II-B-2

C/                            '              "            '
/R Incorporated	1996 EPA Outstanding Small Business Contractor
    MEMORANDUM

    Date:          February  20,  1997

   •Subject:       Maximum Achievable Control  Technology Floor
                   Determination and Regulatory Alternative
                   Consideration             .
    From:          Joanne C. Seaman,  EC/R

    To:            David Svendsgaard,  EPA'/OAQPS/ESD/OCG


         The purpose of this memorandum is to present the approach
    used to develop maximum achievable control  technology (MACT)
    floors and regulatory alternatives for new  and existing sources
    in the polyether .polyols industry.  -This memorandum is organized
    in four 'sections.  First, a brief background of the project and
    regulatory process is provided.   Second, a  discussion is provided
    regarding the approach used to  determine' the MACT floor levels of
    control.  Third, MACT floors calculations are described, followed
    by. a presentation of the results  of the. analysis.'  The final
    section is a discussion of the  development  of regulatory
    alternatives' more stringent than  the MACT floors. '

    BACKGROUND                                       .

    Clean Air Act Requirements

         Section 112 of the Clean Air Act (CAA) ,  as amended in 1990,
    defines a minimum level of control-,  referred to as the. "MACT
    floor, " for standards established under Section 112 (d) .   For new
    sources, emission standards -"shall not be less stringent than  the
    emission control that is achieved in practice by the best
    controlled similar source."  For  existing sources,  the "emissions .
    standards must be at -least as stringent as  either "the average
    emission limitation achieved by the best performing 12 percent of
    the existing sources (for which the Administrator. has emissions
    information) ... in the category or subcategory for categories  or
    subcategories with 30 or more sources" or "the average emission
    limitation achieved .by the best performing  5. sources "' for.
    categories or subcategories with  less than  30. sources.

    Subcategoriiation of the Source Category

         There are -several fundamental decisions  that were required
    to b*e made before the MACT floor  could be determined.   These
            South Square Office             .              Research Triangle Park Office
   3721-D University Drive • Durham. North Carolina 27707       2327 Englert Drive. Suite 100 • Durham. North Carolina 27713
     Telephone: (919) 493-6099 • Fax: (919) 493^393              Telephone: (919) 484-0222 • Fax: (919) 484-0122

-------
decisions are discussed in this section.  The EPA decided it was
appropriate to subcategorize the source category for purposes of
analyzing the MACT floors and regulatory alternatives.  The
subcategories are polyether polyols made from the polymerization
of epoxides and polyether polyols made from the polymerization of
tetrahydrofuran (THF).   An "epoxide" is a chemical compound
consisting of a three-membered cyclic ether.  Ethylene oxide  (EO)
and propylene oxide (PO)  are the only epoxides that are listed as
HAP, and, therefore regulated by this rule.  Subcategorization
was necessary due to the distinctively different nature of the
epoxide and THF processes and the effect of these differences on
the applicability of controls.  One noteworthy distinction
between the two subcategories is that the first group, polyols
made with epoxides, uses HAP as the monomer (s) , whereas the
second group does not use a HAP monomer.  Additionally, the first
group performs the reaction primarily on a batch basis, while the
second group performs the reaction on a continuous basis.
Separate MACT floors were developed for each subcategory and are
presented in this memorandum.

Major Source Determination

       Of the 26 facilities. in the largest portion of the
database, the process vent portion, two facilities were reported
to be area sources.  It was assumed that the remaining 24
facilities are major sources or synthetic minor sources.  This
determination was made because,  while some type of control is» in
place to reduce emissions at many facilities, there is no
evidence that these controls are "federally enforceable."  This
means that the potential to emit is probably above major source
levels if the controls used are not considered when calculating
emissions.  Further,  there are no federal regulations limiting
emissions from polyols production at this time.

Database

     The joint information gathering effort between the U.S.
Environmental Protection Agency (EPA) and the Society of the
Plastics Industry (SPI)  (EPA/SPI questionnaire) was used in
determining MACT floors for the following emission sources at
polyether polyols production facilities:  storage vessels,
equipment leaks,  and wastewater.  For emissions from process
vents,  additional information was received from the SPI, in the
form of emission and control efficiency data  (SPI database).1
There was a good deal of variation in the level of detail of the
process vent information provided by the EPA/SPI questionnaire
and the SPI database.   Specifically, the responses to the EPA/SPI
questionnaire included the following:  detailed descriptions of
     '  Letter and attachment from Maureen Healey of the SPI, to
David Svendsgaard,  EPA/OAQPS/ESD/OCG.  May 2, 1996.

-------
Che processes;  pre-control  and post-control emission rates; and
process vent  stream characteristics  for some streams, including
flowrate, duration,  and heat  content.  The SPI data consisted of
limited information on 16 facilities, where total process vent
post-control  emission rates were  provided for only 12 facilities;
and control efficiencies were provided for all 16 facilities.

     The EPA/SPI  questionnaire was written by the EPA project
team with input from industry representatives.  The SPI was then
responsible for submitting  the questionnaire to SPI members and
nonmembers.   The  SPI representatives indicated that they had sent
questionnaires  to anyone who  voiced  an interest in responding,
and that they did not select  the  best performers when soliciting
questionnaire responses.  Therefore, the EPA project team
concludes that  the responses  received are not skewed towards any
control performance range,  and are representative of the entire
industry.

Number of Sources

     The EPA  received a list  of 82 plants that produce polyether
polyols in the  United States  from the SPI2.  Two  additional
facilities were also included on  this list:  one facility sent a
questionnaire response to the EPA, and the second facility was
listed in the Chemical Economics  Handbook (CEH) .  The EPA project
team determined that,  of the  list of 84 plants, there were only
two facilities  in the list  that polymerize THF exclusively, and
the remaining 82  produced polyols with epoxides.  One of the
facilities indicated that,  although  they produce polyols with
epoxides, they  also have a  THF polymerization line.  Therefore,
the project team  estimated  that the  source category consists of
three facilities  that polymerize  THF and 32 facilities that
polymerize epoxides into polyether polyols.

     Seventy-six  facilities that  use epoxides to produce
polyether polyols are expected to be major sources subject to the
rule.  This number was calculated by first determining the
percentage of the facilities  in the  database that are major
sources  (92.3 percent) .  This percentage was multiplied by the
total number  of facilities  in the database that use epoxides, to
estimate the  number of facilities using epoxides to produce
polyether polyols.   The three facilities that use THF to produce
polyether polyols are all expected to be major sources subject to
the rule.  Because one of the facilities that uses THF also
produces polyether polyols  using  epoxides, the total number of
facilities in the entire source category expected to be subj-ect
to the rule is  estimated as 78 (76+2).
     2 SPI Enterprise System: May 9, 1996 printout of "Facilities
for Polyether  Polyols  Production"  for  1994.

-------
APPROACH TO DETERMINE THE MACT FLOOR LEVELS OF CONTROL

     For all emission sources with  the  exception of process
vents, the MACT floor approach was  to analyze the level of
control in place at the facilities, in  order to establish the
MACT floor.  For process vents,  the EPA project team investigated
a couple of different MACT floor formats.

Use of the HON in the MACT Floor Determination

     The EPA studied methods to simplify the MACT floor analysis,
and decided to use the Hazardous Organic NESHAP, or HON (40 CFR
part 63,- subparts F, G, and H,  59 FR 19402), in the MACT floor
analysis.  The rationale for this conclusion is provided in this
section.

     The HON contains emission limitations  for five emission
source types:  process vents,  storage vessels,  transfer
operations, wastewater, and equipment leaks.  Because of the
similarities between the synthetic  organic  chemical manufacturing
industry (SOCMI)  and polyether polyols  industries,  the EPA
concluded that the HON requirements for storage vessels,
wastewater, and equipment leaks were appropriate to use in
defining the MACT floor for the polyether polyols production
industry.  Another practical reason for using the HON
requirements is that the HON provides   "ready-made" alternatives.
That is, the HON analysis takes into account equipment type,
equipment size, equipment contents, stream  characteristics, and
other important aspects of the emission source that should be
considered in the floor determination.

     The intent of this approach is to  determine how controls at
existing polyols facilities compare to  the  level of control that
would be required by the HON for all emission types except
process vent emissions.  The HON-based  type of analysis does not
provide specific numeric values for the MACT floor.  Rather, the
conclusion of each floor analysis using this HON-based approach
is whether the MACT floor is less stringent, than, more stringent
than, or equal to,  the HON-level of control.  For each facility
in each subcategory, the existing controls  were identified for
each emission point.  The existing  level of control was then
compared to the level of control that would be required by the
HON, and the emission point was characterized as being controlled
at a level less stringent than the  HON  requirements (less
than HON) ,  a level equivalent to the HON requirements (equal to
HON) , or a level more stringent than the HON requirements
(greater than HON) .

     The HON requirements noted above apply to three of the four
emission source types and part of the fourth: storage vessels,
wastewater, and equipment leaks,  as well as process vents from
continuous unit operations that use nonepoxide HAP to make or

-------
modify the product.   For process vents from batch unit operations
that make or modify the product, the "Control of Volatile Organic
Compound Emissions from Batch Processes:  Alternative Control
Techniques Information Document" (EPA-450/R-94-020)  (Batch ACT)
was used.  Finally,  for process vent emissions of epoxides and
for process vent emissions of nonepoxides from catalyst
extraction, an aggregated emission reduction approach was used.

Use of the Batch ACT for Process Vents from Batch Unit Operations

     The Batch ACT provides guidance to State and local air
pollution regulatory agencies on the development of regulations
for air emissions from batch processes, and is intended to apply
to all batch operations.  Due to the similarities between the
processes studied in the Batch ACT, and the general nature of the
applicability criteria, the EPA concluded that these criteria
were appropriate to use in defining the MACT floor for process
vents from batch unit operations in the polyether polyols
industry.  The determination of the MACT floor using the Batch
ACT is also described in the procedures section of this
memorandum.

MACT Approach for Process Vents

     In the early stages of the presumptive MACT  (P-MACT)
process, the EPA project team proposed an average emission factor
as the format for the MACT floor for process vents.  This
emission factor was calculated by dividing total process vent
emissions by actual production for the year 1993 .  The EPA
project team's intent for this MACT floor format was to allow the
facilities to take credit for reacting their product to
completion in an extended cookout.  The emission factor approach
would not require definitions or calculations for these extended
cookouts.  However,  industry representatives criticized this
format, saying that the emission factor MACT floor was
unrealistically low and that it restricted the facilities from
producing certain polyols classes.  They added that the EPA
project team did not understand the chemistry and diversity of
the polyols products, and stated that, in an effort to meet this
standard, they would only be able to produce certain polyols
classes that were not "capped" with EO.  Industry representatives
then proposed a control efficiency format, along with a
definition of all the controls used at polyols facilities, which
included extended cookout.

     As a result of the industry recommendations, and due to the
lack of data for a MACT floor based on stream characteristics,
the project team selected a control efficiency format.  Each
facility's control efficiency was calculated by  subtracting the
total controlled process vent emissions from the total
uncontrolled process vent emissions.  This difference was then
divided by the total uncontrolled process vent emissions, and

-------
divided by the total  uncontrolled process vent emissions, and
multiplied by 100 percent.  The MACT  floor was calculated as the
median of the top twelve  percent of the  control efficiencies at
28 facilities  (or 3.36  facilities).

     The MACT floor for existing sources that polymerize THF to
produce polyols was determined by averaging the emission
information available from  the two facilities in the database.
Similar to the new source standard for polyols with epoxides, the
new source MACT floor for THF polymerization was identified as
the level of control  at the best performing facility.

MACT FLOOR LEVELS OF  CONTROL CALCULATIONS

     Two basic procedures were used to determine the MACT floors
for the polyether polyols subcategories. The first, the HON-based
approach, compared existing levels of control with the level of
control that would be required at polyether polyols facilities if
the HON requirements  were applied.  This approach was used for
storage vessels, wastewater, and equipment leaks.  For process
vent emissions of nonepoxides in making  or altering the product,
the same approach was used, except that  the 90 percent control
level from the Batch  ACT  was. used for process vents from batch
unit operations, and  the  HON process  vent provisions were used
for process vents from  continuous unit operations.  The 90
percent Batch ACT control level was selected, because the
estimated cost-effectiveness for this level was comparable to the
cost-effectiveness of the HON continuous vent provisions
(approximately $3,000 per Megagram) .  A  second approach was used
to assess the average emission limitation for process vent
emissions of epoxides and for nonepoxide emissions from catalyst
extraction.  The approach used for these process vent emissions
was an aggregated emission  reduction  percentage for all process
vents within the polyether  polyols manufacturing process unit
(PMPU).

     The control level  for  both existing and new sources was
determined by comparing the controls  in  place at each facility
with the existing source  requirements in the HON.  This section
describes the general HON-based approach used for all emission
source types (i.e., storage vessels,  process vent nonepoxide
emissions from making or  altering the product, wastewater
streams, equipment leaks) ,  followed by specifics of the
individual emission source  types.

General Approach

     After each emission  point at each facility was
characterized,  all emission points of a  given emission source
type were grouped together  and a facility-wide determination was
made for each emission  source type.   For instance, if a storage
vessel was controlled at  a  level less stringent than the HON, and

-------
no other storage vessel was controlled at a level more stringent
than the HON, the facility was classified as "less than HON" for
storage vessels.  If all controls at the facility were equivalent
to the HON levels, the facility was classified as "equal to HON."
If one or more points was controlled at a level more stringent
than the HON, and no point of the same type was controlled at a
level less stringent than the HON at that facility,  the facility
was classified as "greater than HON. "

     It is important to note, however, that if an emission point
was uncontrolled, and the HON would not require control for that
point, the level of control was classified as equal to the HON
level of control.  Therefore, the floor for a subcategory could
be the HON, when in fact all emission points of that particular
emission source type were uncontrolled.

     If a facility reported different levels of control (in
comparison to the HON) within one emission source type, an
additional analysis was necessary to classify the facility.  In
these situations, the existing emission level was compared to the
emission level that would be required if HON controls were
applied.  If the existing emissions were less than the HON-level
emissions, the facility was classified "greater than HON," but if
the HON-level emissions were lower than the existing emissions,
the facility was classified "less than HON."

     The floor for all emission source types except process vents
was defined for both subcategories as less than,  equal to, or
greater than, the HON level of control.  The HON-based approach
used for new sources was similar to the existing source approach.
This determination was based on the majority of individual
facility classifications for the subcategory.

     Storage Vessels

     The applicability of the existing HON storage tank
provisions is based on tank size and vapor pressure of the total
organic HAP in the storage tank.  Vessels meeting the
capacity/vapor pressure criteria must be controlled using one of
three control techniques:  (1) an internal floating roof,  (2) an
external floating roof, or (3) a closed vent system to control
device  (with a 95 percent control efficiency) .  If the maximum
true organic HAP vapor pressure of the vessel contents is high
enough, the control options are limited to a closed vent system
vented to a control device.  Pressure tanks  (at a pressure
greater than 23 psi) are exempt form the HON requirements.  ,

     In order to evaluate the applicability of the HON provisions
to polyether polyols storage vessels, the maximum true vapor
pressure of the HAP were estimated, and the level of control that
would be required determined.  For each facility, the existing
storage vessel controls were noted and summarized by individual

-------
HAP.  The existing level of control  was  then compared to the HON-
level of control for each vessel,  and the MACT floors were
determined as discussed above in the section entitled "General
Approach."


     Modifying the Product

     The HON process vent provisions apply to continuous process
vents emitting streams containing more than 0.005 weight-percent
HAP.  Control (98 percent emission reduction)  is required for
each process vent with a flow rate greater than or equal to 0.005
standard cubic meters per minute,  an organic HAP concentration
greater than or equal to 50 ppmv,  and a  total resource
effectiveness (TRE)  index value less than or equal to 1.

     The Batch ACT process vent provisions apply to volatile
organic compound (VOC)  emissions from process vents from batch
unit operations.  In this analysis,  only the organic HAP were
considered.  The first level of applicability in the Batch ACT is
based on annual emissions.  If annual HAP emissions, calculated
before a control or recovery device,  are less than specified
levels, control is not required.  For the 90 percent control
level, these "cutoff" levels are between 7,300 and 11,800
kilograms per year,  depending on the volatility of HAP's emitted.
If annual emissions were greater than the cutoff level,  emission
were input into an equation from the ACT to determine a cutoff
flow rate.  If the actual flow rate  of the batch vent stream is
less than the cutoff flow rate,  90 percent control is required.

     EC/R applied the HON or Batch ACT criteria to each process
vent for which the necessary vent stream parameters were
provided.  In some instances,  EC/R was able to estimate the
necessary parameters.  In situations where a process vent from a
continuous unit operation was analyzed,  EC/R applied the HON
criteria to the stream prior to the  control device, to assess
whether the stream would have required control based on the HON
criteria.  Similarly, in situations  where a process vent from a
batch unit operation was controlled  using a control or recovery-
device, the Batch ACT criteria were  applied prior to the device.
The existing level of control was then compared to the level of
control that would be required by the HON or Batch ACT.

     Process Vent Emissions of Epoxides  and of Nonepoxides from
     Catalyst Extraction

     For process vent emissions of epoxides and of nonepoxides
from catalyst extraction, an average aggregated emission
reduction was calculated for each PMPU.   To do this the following
calculation was used:

-------
     where:
     R = Emission reduction, percentage;
     Eu = Uncontrolled epoxide process vent  emissions, pounds
     per year, (Ib/yr); and
     Ec =  Controlled epoxide process vent emissions, Ib/yr.

     Wastewater

     The HON requires wastewater streams at  existing sources that
have a total organic HAP concentration of 1,000 ppmw or greater
and a flow rate of 10 liters per minute or-greater  (at the point
of determination), or a total organic HAP concentration of 10,000
ppmw or greater at any flow rate, to be managed in  controlled
equipment and treated to reduce the HAP concentration.

     EC/R applied these criteria to uncontrolled wastewater
streams at each facility. The stream characteristics were then
compared to  the HON applicability criteria to assess whether the
HON would require control.  Each wastewater  stream, and in turn
each facility, was classified as "less than  HON," "equal to HON,"
or "greater  than HON."  The floor for wastewater for each
subcategory  was then determined as discussed above  in the section
entitled "Use of HON in the MACT Floor Determination."

     Equipment Leaks

     The HON equipment leak provisions require a. combination of
leak,  detection,  and repair  (LDAR)  procedures and equipment
requirements. In order to compare the existing control levels to
the HON levels, baseline control levels were assessed for each
type of component at each facility.  A qualitative  assessment was
then made for each facility to determine whether the existing
level of control was less than, equal to, or greater than the HON
level of control.

Maximum Achievable Control Technology Floor  Conclusions

     Table 1 presents a summary of the MACT  floor levels of
control for  all emission types from both subcategories.  Based on
existing EPA policy, information from the EPA/SPI questionnaires,
and information from the SPI data submittal3, the EPA is  making
recommendations for MACT floors for all emission source types
identified in the two subcategories.  The majority  of the data
received were based on 1993 information.  The sections below
present the  MACT floors first for polyols made with epoxides, and
then for polyols made with THF.
     3  Letter  and attachment from Maureen Healey of the SPI,  to
David Svendsgaard, EPA/OAQPS/ESD/OCG.  May 2, 1996.

-------
MACT Floor Conclusions  for  Polyether Polyols Made with Epoxides

     Unloading

     Industry representatives  reported that all unloading of HAP
was conducted in a  closed vent system;  therefore, no emissions
are expected from this  portion of  this process.

     Storage Vessels

     The MACT floor level of control for  storage vessels at
existing sources that make  polyether polyols with oxides was
determined to be the HON level of  control and applicability.  The
rationale for this  determination follows.  The majority of the
vessels were pressurized vessels  (which are exempt from the HON
requirements), with only four  facilities  reporting vessels that
were HON Group 1 storage vessels.   It was determined that these
four facilities have storage vessels that would require control
under the HON.  Three of these four facilities with HON Group 1
storage vessels have HON reference control technologies on their
HON Group 1 storage vessels; therefore, the MACT floor level of
control for storage vessels at existing sources was determined to
be equal to the level of control in the HON.  No facility had
controls more stringent than the HON level of control for their
Group 1 vessels,- therefore, the MACT floor level of control for
storage vessels at  new  sources is  the HON existing source storage
vessel level of control.

     Equipment Leaks

     The database of 13 facilities reporting equipment leak
information indicated that  nine out of 13 facilities have LDAR
programs in place.  Of  the  nine LDAR programs in place, six were
reported to be equal to the level  of stringency required by the
HON (i.e., in terms of  monitoring  frequency and leak definition).
No facility reported an equipment  leak emission control program
more stringent than the HON existing source level of control.
Therefore, the MACT floor level of control for equipment leaks
                                10

-------
 0)
rH
A
 ffl
EH

















tn
O
o
.H
Cu
c-i
U
















1
*
4-1
c
iu
S
a,
3
tr
u
M
i -H
•—< >;
o w
04

o o
u u
4-1 4-1
C C
o o
u o
0 0
Z Z


1-1 1-1
o o
4J 4J
c c
0 0
o u

o o
2 Z
percent
percent
0 O




— \
2 Z





rf* rf*
R, R,
2 Z




control
control
o o
Z Z
jS Cu o) tn
4J X 0) CU
•H tn 0 O
3 M U
3 3
(I) O O
73 CO CO
ID
S tn 3
C di
m -HZ
^H 4->
O in
>, -H
•-I X
o u
Cb


-------
for new and existing sources was determined to be the  level of
control required by the HON  (40 CFR 63, subpart H) .

     Process Vents

     Three sets of MACT floor levels of control were established
for process vent emissions:  one for epoxide emissions;  a second
for nonepoxide HAP emissions from catalyst extraction; and a
third for emissions of nonepoxide HAP used to make or  modify the
product.  For the subcategory that uses epoxides, the  MACT floor
levels were established by determining the average emission
control for the best performing 12 percent of the facilities with
respect to that emission point, which translated to 3.36
facilities for the database of 28 facilities.  The median
approach was used where there was a wide range of values within
the best performing facilities.  .For new sources, the  MACT floor
levels were established by determining the emission control for
the best controlled facility for that emission point in  the
subcategory.
     The MACT floor for epoxide emissions from process  vents at
existing facilities that produce polyether polyols with epoxides
was- calculated as the median of the top twelve percent  of the
database.  The MACT floor was calculated to be a 98.1 percent
control efficiency, and was rounded to 98 percent since the
standard is technology based and it is believed that facilities
would not use a different control technology to meet a  standard
that is a tenth of a percent more stringent.

     The MACT floor level of control for epoxide emissions from
process vents at new sources was determined to be a control
efficiency of 99.9 percent.  This new source MACT floor was based
on a facility that reported a control efficiency of 99.9 percent.

     tToTigpo-x-ide ff^P Tftnjasions from Catalyst Extraction

     For nonepoxide HAP emissions from catalyst extraction, the
MACT floor was determined as the median of the data in  the
database, which was 90 weight percent aggregated emission
reduction for existing sources.  The MACT floor level of control
for new sources was determined to be 98 weight percent  aggregated
emission reduction.

     Nonepoxide HAP From Making or Modifying the Product

     For nonepoxide HAP process vent emissions from making or
modifying the product at existing sources, the MACT floor level
of control was calculated using the median of the top twelve
percent.  The existing source MACT floor level of control for
nonepoxide HAP emissions from making or modifying the product was

                                12

-------
determined to be no control.  The new source MACT floor level of
control for nonepoxide HAP process vent emissions was determined
to be 39.0 percent control efficiency.

     Wastewater

     The MACT floor level of  control for wastewater for new and
existing sources was identified to be less than the HON level of
control.  In fact, the level  of control is no control.  The
database indicated that only  two facilities in the database
reported Group 1 wastewater streams that require controls
according to the HON, and neither facility controlled air
emissions from these wastewater streams.  Further, of the Group 2
wastewater streams that were  reported, none had air emission
controls .


Polyols Made with. Tetrahydrofuran

     Only one of the two facilities in the database for the THF
subcategory uses and emits organic HAP.  Therefore, the MACT
floor analysis was based on that one facility.  This one source
also sets the level of control for the new source MACT floor for
THF polymerization.

     Storage Vessels

     The MACT floor level of  control for new and existing sources
was identified as the HON level of control.  Only one of the THF
facilities had a storage vessel for HAP.  This storage vessel was
a Group 2 vessel that does not require controls and did not have
a control.

          ss Vents
     The  facility  that uses  organic HAP in its process does not
control the  process vent  emissions of the organic HAP. The second
facility  in  the database  emits  hydrogen fluoride, well below the
HON's  definition of a halogenated stream, and has minimal
controls  (approximately 20 percent emission reduction) on these
emissions.   Therefore, the MACT floor level of control for
process vents  from polyether polyols produced using THF at
existing  sources was identified to be no control.

     Equipment Leaks

     No control of emissions from equipment leaks was reported at
the  facility that  polymerizes THF.  Therefore, the MACT floor for
equipment leaks at existing  and new facilities was determined to
be less than the HON level of control.  More specifically, the
MACT floor level of control  is no control.
                                13

-------
     Wastewater

     The MACT floor  for wastewater for existing and new
facilities  that produce polyols by polymerization of THF was
determined  to be equal to the HON level of control.  No Group 1
wastewater  streams were reported and no controls were reported
for the Group 2 wastewater streams.  The MACT floor level of
control is  equal to  the HON level of control (i.e., no control) .

DEVELOPMENT OF REGULATORY ALTERNATIVES MORE STRINGENT THAN THE
MACT FLOOR

     Only one regulatory alternative more stringent than the MACT
floor level of control was developed and considered for each
subcategory.  Table  2 presents the MACT floor and the regulatory
alternative for existing sources.or each subcategory.  The
rationale for the level of this alternative is discussed below.

Polyether Polyols Made with Epoxides

     The MACT floor  level of control for storage vessels and
equipment leaks was  equal to the HON level of control.
Therefore,  the regulatory alternative included the HON level of
control for these emission types.

     The MACT floor  level of control was determined to be less
stringent than the HON level of control (i.e.,  no control) for
wastewater  emissions.  The HON level of control was considered
for the regulatory alternative for wastewater because it had
received extensive evaluation during the development of the HON,
at which time the EPA concluded that the cost and other impacts
associated  with the  HON-level of control were reasonable.
Therefore,  the regulatory alternative included the HON level of
control for wastewater.

     During the development of the HON, alternatives more
stringent than the promulgated levels were considered and
rejected by the EPA.  Therefore, it was unnecessary to consider
controls more stringent than the HON levels, since this industry
closely mirrors the  SOCMI, and the EPA had previously considered
them unacceptable for the HON.

     For the process vent regulatory alternatives, an emission
reduction format was chosen and applied to the three groups of
HAP process vent emissions:  epoxide emissions; nonepoxide HAP
emissions from the making or modification of the product; and
nonepoxide  HAP emissions from catalyst extraction.  The MACT
floor level of control for epoxide process vent emissions was
considered  to be sufficiently stringent, since it mirrored the
highest level of control in the Batch ACT (without performing a
                                14

-------
 0)
iH
A
 a

















in

-H
X
o
a
w

,
M
O
CTI
0)
4-J
(0
u
,Q
3
W





2 2
0 0
33 a




<-i
0
M
C^
0 °
8 =

o
2


90 percent
90 percent


EH
a
c*
8«
§5
a*
0 °
o
X

4J JJ
c c
0)  en

2 2
0 0
x :r


._ MM
s §-
3 ^ ^
•* h <
•§ ^ *
« tf a
6 S a
in i X
^H 0
o a
Oi (U



rH
o
M
8 1
0 ^
o



•-(
o
Ij
J-1 -z
g§
8==

o
2


JJ
. C
HI
u z
u O
m K
a
o



ft
2




--<
«i
>X,
2



rH
O
M
e 2=
0 °
S31
O

c. M M
5 g.
S iH r-l
•* Eu rt
W e. _
•rt *"* :
* ^ *
H § *

in
—i
O
>1
rH b
O =
a. H

-------
cost effectiveness analysis to determine applicability).
Therefore, no levels of control more stringent than the floor
were evaluated.

     The MACT floor level of control for nonepoxide HAP emissions
from making or altering the product was determined to be an
aggregated control efficiency of 0 percent.  The EPA determined
that the applicability criteria from either the HON or the Batch
ACT could be applied for process vents from continuous or batch
unit operations, respectively.  The HON Group determination
approach was deemed appropriate because in the equation in the
HON for determining Group I/Group 2 applicability, the total
resource effectiveness index  (TRE) has an inherent cost
effectiveness value in it.  After determining the Group status,
the control requirement from the HON, a 98 percent control
efficiency, was used for this regulatory alternative, as well.
If the process vent was from a batch unit operation, the Group
determination was based on equations first developed in the Batch
ACT.  The 90-percent control level from the Batch ACT was
selected because the facilities that did report controls on these
streams reported condensers, and 90 percent was the lowest
control efficiency achieved.  Further, the estimated cost-
effectiveness for this level was comparable to the cost-
effectiveness of the HON continuous vent provisions.  Based on
these analyses, the EPA determined that it was acceptable to
consider a single regulatory alternative beyond the MACT floor
level.

Polyether Polyols Made with Tetrahydrofuran

     The MACT floor level of control for all the emission types
in this subcategory was determined to be less stringent than the
HON level of control.   For these same reasons as presented above
for the polyether polyols made with epoxides subcategory, the
regulatory alternative included the HON level of control for
storage vessels, equipment leaks, and wastewater.  For the
process vent emissions,  similar to the other subcategory, the HON
or Batch ACT Group determination was included in the regulatory
alternative.
                                16

-------
CO
 0)
















Alternatives
^i
M
0
nj
rH
3
0)
OS
^
M
O
01
CU
4J
It)
0
£1
3
co



2 2
0 O
X 33



i-l
0
M
XI g.
U *

O
2
98 percent
98 percent




EH
C £
CU U
O JJ
M 10
4) CQ
a
M
O^ O
PI
1

4J 4J
C C
V 
0%  O
"2 •—('-<
4)
•o g U
t8 cf ^
£ 5>
to *•
to cu
^ T3
0 -H
>^ X
<-* O
o a
04 (U

^
o
4J z
n O
CJ ^
O
2

.-t
O
u
§1
o *•

o
2
0 percent
HON 	




\
2





tfT*
R,
2



O
M
C2
°S
y *
o
2
5 QM
•H 5 *i
3 ""* ^
CU tj rn
13 0 W
TJ jJ* *y
E § *

CO
r-^
O
>,
r-1 tu
0 S3
0! E-<

-------
 R Incorporated
                                                           A-96-38
                                                           II-B- 5


                                1996 EPA Outstanding Small Business Contractor
 MEMORANDUM

 Date:   .

 Subject:


 From: '

 To:
                April  29, .1996

                Summary  of  Nationwide Baseline Emissions  for
                Polyether Polyols Production Facilities
                Joanne  C/ Seaman,  EC/R Incorporated

                David Svendsgaard,  EPA/OAQPS/ESD/OCG ^
      The purpose of this  memorandum is to present the projected
 nationwide baseline emissions of  hazardous air pollutants  (HAP)
 from polyether polyols production . facilities .   Nationwide
 baseline emissions were calculated by first calculating  the
 average HAP emissions from the .database 'for each emission  source
 within the polyols production process (i.e.,. process, vents,
 equipment leaks, storage,  and wastewater) .  Nationwide baseline
 HAP emissions were estimated by extrapolating average emissions
 from these emission sources to nationwide ' emissions based  on the
.number of actual sources  in this  industry. •

      This approach was used instead of developing "model plants"
 and fchen scaling up emissions from each possible model plant to
 nationwide emissions.  The rationale behind this 'decision  was
.that there would have been an extremely large number of  possible
 model plants-, considering the different combinations of  scenarios
 from the various emission 'sources in the 'polyols production
 process.                                                '  '

 SOURCE CATEGORY:-
      Polyether .polyols are  defined as the products formed by the
 reaction of ethylene oxide  (EO) ,  propylene oxide (PO) , or other
 cyclic ethers with compounds • having one or more reactive
 hydrogens  (i.e.,. a compound having a hydrogen terminally  bounded
 with a nitrogen, sulfur,  oxygen,  phosphorous atom,  etc.) ^   This
 definition excludes .materials regulated as glycols or glycol
 ethers under the Hazardous  Organic National Emission Standard for
 Hazardous Air Pollutants  (HON) .-  A major source is any
 stationary source or group  of stationary sources located  within a
 contiguous area and under common  control that emits or has the
 potential to emit, considering controls,  in the aggregate,  10
 tons per year of any HAP  or 25 tons per year of any combination
 of HAP .   The source category  has  been divided into two
 subcategories : polyether  polyols  made with epoxides, which is the
 majority of the source category;  and polyols made from the
         South Square Office
3721-D University Drive • Durham. North Carolina 27707
  Telephone: (919) 493-6099 • Fax: (919) 493-6393
                                            Research Triangle Park Office
                                    2327 EngJen Drive. Suite 100 » Durham. North Carolina 27713
                                       Telephone: (919) 484-0227 • Fax: (919) 484-0122

-------
polymerization of tetrahydrofuran (THF) .   The source category was
subcategorized due to the  inherently different nature of the two
types of processes.   Epoxide  means  a chemical compound consisting
of a three-membered cyclic ether.   For the purposes of this rule,
epoxides include EO and PO.

     The Society of the Plastics  Industry (SPI)  supplied to the
EPA project team a list of 82 plants that produce polyether
polyols in the United States.  Two  more  facilities were added to
this list:  one facility sent a questionnaire response to the
EPA, and the second facility  was  listed in the Chemical Economics
Handbook (CEH) .  The EPA project  team determined that, of the
list of 84 plants, there were only  two facilities in the list
that polymerize THF exclusively,  and the remaining 82 produced
polyols with epoxides.   One of the  facilities indicated that,
although they produce polyols with  epoxides, they also have a THF
polymerization line.  Therefore,  the project team estimated that
the source category consists  of three facilities that polymerize
THF and 82 facilities that polymerize epoxides into polyether
polyols.2

     The number of facilities that  use epoxides to produce
polyether polyols that are expected to be major sources subject
to the rule is 76 .  This number was calculated by first
determining the percentage of the facilities in the database that
are major sources (92.3 percent).   This percentage was multiplied
by the total number of facilities in the database that use
epoxides,  to estimate the  number  of facilities using epoxides to
produce polyether polyols  that are  expected to be subject to the
rule.  The three facilities that  use THF to produce polyether
polyols are all expected to be major sources subject to the rule.
Because one of the facilities that  uses THF also produces
polyether polyols using epoxides, the total number of facilities
in the entire source category expected to be subject to the rule
is estimated as 78  (76+2).

     A description of the  calculation approach and data used,
along with the resulting nationwide emissions for each emission
source type, is presented  in the  appropriate sections below.

NATIONWIDE BASELINE EMISSIONS

     The polyether polyols process  vent database was developed
based on two main sources: (1) responses to a questionnaire that
was a joint information gathering effort between the U.S.
Environmental Protection Agency (EPA) and the SPI  (EPA/SPI
questionnaire) ; and  (2) emission and control efficiency data
provided by the SPI  (SPI database)  -1  For equipment leaks,
     1 Letter and attachment  from Maureen Healey of the SPI, to
David Svendsgaard, EPA/OAQPS/ESD/OCG.  May 2,  1996.

-------
wastewater and storage  vessels,  baseline emissions were
calculated from the EPA/SPI  questionnaire responses .  For process
vent baseline emissions,  the EPA/SPI  questionnaire data was
supplemented by the SPI database.
         Vents
     Polyols Made jwith  Epoxides

     The  responses  to the EPA/SPI  questionnaire included the
following:   detailed descriptions  of  the processes; pre-control
and post -control emission rates; and  process vent stream
characteristics,  such as flowrate,  duration and heat content.
The SPI data consisted  of limited  information on 16 facilities,
where total  process vent post -control emission rates were
provided  for only 12 facilities, and  control efficiencies were
provided  for all 16 facilities.  Both the SPI database and the
EPA/SPI questionnaire database were used to estimate baseline
emissions.   The SPI database  reported total epoxide and other-HAP
emissions and did not distinguish  where in the process these
emissions occurred.  Therefore, due to the lack of detailed
information,  it was assumed that all  epoxide emissions from the
SPI database were from  the reactor step of the process, and all
other-HAP emissions were from the  finishing step of the process.
In performing these calculations,  it  was assumed that the
information  in the  SPI  database and the EPA/SPI questionnaire
database  are representative of the entire industry.
     The process  of producing polyols is performed in two basic
steps: the  reaction step, where  the polyol is polymerized; and
the finishing  step, where the crude polyol from the reactor is
further treated.   Table 1 presents the summary of emissions from
both databases for the reactor step of the process.  The EO and
PO emissions from the reactors ranged from 20 pounds per year
(Ib/yr) to  80,966 lb/yr.  The EO and PO emission rates were
separated into two emissions scenarios: a low epoxide emissions
scenario and a high epoxide emissions scenario.  In the low
epoxide emissions scenario, emissions were less than or equal to
one ton per year  (tpy) , with an  average emission rate of 474
lb/yr.  This scenario represents 74 percent of the facilities in
the database.   In the high epoxide emissions scenario, emissions
were greater than 1 tpy and the  average epoxide emission rate was
21,200 lb/yr.   This scenario represents 26 percent of the
facilities  in  the database.

     Five of the  polyols producers using epoxides in the EPA/SPI
questionnaire  database reported  emissions of HAP other than EO
and PO  (other-HAP) from their reactors.  Similar to the EO and PO
data, emissions of other HAP were separated into two scenarios:
low other-HAP  emitters, with emissions less than 36 lb/yr for

-------
Table 1.  Summary of Baseline HAP Emissions  from the Reactor Step
         of the Polyols Production with Epoxides Process
FACILITY
-,
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
TOTAL
Low Emitters'
Averaae
Hign Emitters'
V/eraae
REPORTED EPOXIDE
EMISSIONS 'pounds
oer vear)
20
49
110
125
144
144
160
169
196
224
272
303
540
860
1329
1675
1735
2,116
2,633
6,325
9,858
15,602
80,966
135,000
474
21,200
Epoxide
EO
EOSPO
EOSPO
EOSPO
EO
£0
EOSPO
PO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
EO
EO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
EOSPO
REPORTED OTHER-HAP
EMISSIONS (pounds
oer vear)









36

25,732
27





1,751



4
27,500
22
13,700
POLLUTANT









1, 3-Butadiene

Toluene s Toluene
Diamine
Diethylene Glycol
S Ethylene Glvcol





Methanol &
Ethvlene Glycol



Toluene
Other-HAP
Other-HAP
Other-HAP
other-HAP; and high other-HAP emitters, with  emissions greater
than 1,751 Ib/yr.  Of the facilities  in the database, 13 percent
were low other-HAP emitters from the  reactor  step,  with an
average emission rate of 23 Ib/yr.  Eight percent of the
facilities were calculated to be high other-HAP emitters from the
reactor step, with an average emission rate of 13,700 Ib/yr.

-------
Procepg T*nji?gions from Finishing

     The EPA/SPI questionnaire responses indicated that there
were three basic finishing options:   (1)  neutralization of the
catalyst only,  with the resulting salt remaining in the polyol;
(2) catalyst neutralization followed by mechanical removal (such
as filtration or centrifugation)  of  the salt formed during the
neutralization step; and (3)  catalyst neutralization followed by
solvent removal of the salt that forms from the neutralization
step.  The first two finishing procedures described do not result
in HAP emissions.  It is only in the third finishing procedure,
when solvents are used to remove the salt, that HAP emissions
occur from the solvent use.   Table 2 summarizes the emissions
reported from the finishing steps in the process.  Other-HAP
emissions ranged from 20 Ib/yr to 116,812 Ib/yr.  Due to the wide
range of emissions,  finishing step emissions were also separated
into two emissions scenarios:  low other-HAP emitters; and high
other-HAP emitters.   The low other-HAP emission scenario is
characterized by emissions less than or equal to 2,740 Ib/yr,
with an average emission rate of 1,020 Ib/yr.  This scenario
represents 22 percent of the facilities in the database.  The
high other-HAP emitting scenario,  consisted of facilities with
emissions greater than 2,740 Ib/yr,  with an average emission rate
of 53,400 Ib/yr.  This scenario represents thirteen percent of
the facilities in the database.  Although one facility reported
EO and PO emissions from the finishing step of the process, the
data indicated that this is not typical at any other polyols
production facilities, so these emissions were not included in
the calculation of national baseline emissions from process vent
finishing operations.

                        lisa ions
     Nationwide process vent  baseline emissions of HAP from the
reactor and finishing steps of the  process were calculated by
multiplying the average emission rate of the HAP from that
emission scenario by the percent occurrence of the scenario in
the database,  and by the nationwide estimate of polyols producers
with epoxides,  82.  For example:
         Ib/yr = EAVG * Od * 82 = 474 *  0.74  *  82 = 28,800
     Where
     ELO =
     82 =
Emissions from the low epoxide emissions  scenario,
(Ib/yr);
Average epoxide emissions (lb/yr*facility);
     Occurrence of facilities in the database  that
     were characterized by this scenario,  0.74;
     and
Number of facilities in the nation.

-------
Table 3 summarizes the nationwide baseline emissions from the
reactor and finishing steps of the process.

  Table 2.  Summary  of  Baseline HAP  Emissions from the  Finishing
       Step of the Polyols Production with Epoxides Process
FACILITY
1
2
3
4
5
6
7
8
9
TOTAL
Low Emitters '
Average
High
Emitters '
Average
REPORTED
OTHER- HAP
EMISSIONS
(pounds per
year)
20
247
639
1,450
2,740
9,596
33,681
116,812

165,185
1, 020
53,400
OTHER- HAP
POLLUTANT
Other
(unspecified)
Other
(unspecified)
Other
(unspecified)
Other
(unspecified)
Hexane
Other
(unspecified)
Hexane & Toluene
Toluene

Other-HAP

Other-HAP
REPORTED
EPOXIDE
EMISSIONS
(pounds per
year)








140,768
140,768

140,768
EPOXIDE








EO&PO
EO&PO

EO&PO
After the nationwide baseline emissions were calculated for each
emissions scenario,  they were totaled.  This total is the
nationwide process vent emissions of HAP from polyols facilities
with epoxides.   Therefore,
               (28,800 + 452,000
               569,000) lb/yr
               579 tons/yr.
+ 235 + 89,900 + 18,400 +

-------
    Table  3.   Summary
       Emissions  from
of Baseline Nationwide Process  Vent HAP
Polyether PolyoIs Made with Epoxides
REACTOR EMISSIONS '
EMISSION
SCENARIO
Low Epoxide
Emitters
High Epoxide
Emitters
Low Other-
HAP Emitters
High Other-
HAP Emitters

PERCENT OF
FACILITIES
IN
DATABASE
74%
26%
13%
8%
TOTAL'
AVERAGE
EO&PO
EMISSIONS
(pounds
per year)
474
21,200
NA
NA

NATIONWIDE
EO&PO
EMISSIONS
( pounds
per year)
28,800
452,000
NA
NA
480,800
AVERAGE
OTHER-HAP
EMISSIONS
(pounds
per year)
NA
NA
22
13,700

NATIONWIDE
OTHER- HAP
EMISSIONS
(pounds per
year)
NA
NA
235
89,900
90,100
FINISBING EMISSIONS
EMISSION
SCENARIO
Low Other-
HAP Emitters
High Other-
HAP
Emitters
TOTAL
PERCENT OF
FACILITIES
IN
DATABASE
22%
13%

AVERAGE
EO&PO
EMISSIONS
(pounds
per year)
NA
NA

NATIONWIDE
EO&PO
EMISSIONS
(pounds
per year)
NA
NA

AVERAGE
OTHER-HAP
EMISSIONS
(pounds
per year)
1,020
53,400

NATIONWIDE
OTHER-HAP
EMISSIONS
(pounds per year
18,400
569, 000
587,000
     Polyols Made by  Polymerizing Tetrahydrofuran

     As stated above,  there  are  only  three facilities in the
nation that presently polymerize THF  in order to make polyether
polyols.  One of the  facilities  does  not emit HAP; therefore,
nationwide process vent  emissions are the total of the process
vent emissions reported  in the EPA/SPI questionnaire from the
other two facilities.  These emissions are presented in Table 4.

-------
Table 4.  Summary of Baseline Nationwide  Process Vent HAP
Emissions from Polyether Polyols Producers  that Polymerize
Te trahydr o f uran
FACILITY
1
2
TOTAL
REPORTED EMISSIONS
(lb/yr)
3,533
43,497
47,030
POLLUTANT
Hydrogen fluoride
Toluene
Total HAP
Equipment Leaks

     Polyols Made with Spoxides

     The nationwide baseline emissions  for equipment leaks were
estimated by first calculating the emissions  for each component
per each control option (component /control)  for each facility in
the EPA/SPI questionnaire database.   Control  options for
equipment leaks were reported as any of the following: LDAR with
the frequency and leak definition specified;  or the percent of
components with any of the three control options presented in the
EPA/SPI questionnaire.  Table 5 summarizes the control options
that were presented in the EPA/SPI questionnaire.

     The equipment leak emissions for each component at each
control level reported were calculated  in the following manner:
EELC/C  db/yr)

Where:
      (lb/yr)
EF

CE



uc'
               = Nc/c *  EF *  (100  - CE)  *
                                                  *  UC
                       Emissions from each equipment component/
                    control option combination,  lb/yr;
                    =  Number or equipment components with the
                    specific control  option;
                    =  Emission factor,  for SOCMI uncontrolled
                    equipment components,  kg/(source*hr)
                    =  Control efficiency  assigned to the control
                    option, from the  EPA's Protocol Document for
                    Estimating Equipment Leaks 2,  percent;
                    =  Concentration  of  the HAP in the line; and
                    =  Unit conversion,  8760 hr/yr * 2.204 Ib/kg.
     2  Hausle,  K.J.  (Radian Corporation) .  Protocol For
Equipment Leak Emission Estimate.  Prepared for U. S.
Environmental Protection Agency.  Research Triangle Park,
Publication No. EPA-453/R-93-026. June 1993.
                                                     NC.
                                8

-------
   Table 5.   Summary  of  the Control Options  for Equipment Leaks
              Presented  in the EPA/SPI Questionnaire
Control
Option
A
B
C
Equipment
All
Valves
Pumps
Compressors
Open-ended
lines
Sampling
connections
Pressure
Relief Valves
Pumps
Sampling
connections
Controls
Closed vent system
Sealless
Dual mechanical seal with
barrier fluid
Mechanical seals with
barrier fluid
Capped, plugged, blind -
f lagged
In- situ sampling
Rupture disk
Sealless
Closed loop sampling
This calculation was completed for each component/control type  at
each facility.  These calculations are not presented  in this
document.  The equipment leaks database was then sorted by each
component/control, and the average emission rate of each
component/control per facility was calculated.  Also, the percent
occurrence of each component/control in the database  was
calculated.

     Finally, emissions were scaled to nationwide emissions by
multiplying the average emissions per component/control by the
percent occurrence of each component/control, and by  the number
of facilities in the country,  then dividing by the number of
facilities in the database,  11.  That is:
ENAT, c/c>

Where:
ENAT, c/c


EAVG, c/c
     82
     11
             lb/yr  =  E,
                        >AVG, C/C
                 82/11
=  Nationwide emissions from each equipment
component with each control option,  lb/yr;
=  Average emissions,  from the database, of  each
equipment component with each control  option,
lb/yr;
=  Number of facilities in the nation,  and;
   Number of total facilities in the database.

-------
Table 6 presents the nationwide baseline emission for each
component/control.
 Table 6.  Estimate of Baseline Nationwide HAP
from Equipment Components on Polyether Polyols
                 Lines that Use Epoxidea
Emissions
Production
EQUIPMENT
COMPONENT
Valves -
vapor
Valves -
vanor
Valves -
vapor
Valves -
light
liquid
Valves -
light
liquid
Valves -
light
liquid
Valves -
heavy
liquid
Valves -
heavy
liquid
Connectors
Connectors
CONTROL
HON LDAR
Quarterly
LDAR
No controls
HON LDAR
Quarterly
LDAR
No controls
LDAR, no
specified
frequency or
leak
definition
No controls
HON LDAR
Weekly LDAR
with 10,000
leak
definition
CONTROL
EFFICIENCY
(percent)
92
67
0
88
61
0
7
0
93
Emission
factor
from PES
memorandum
AVERAGE
CONTROLLED
EMISSIONS PER
FACILITY
(lb/yr)
790
30,000
5,900
910
89,000
47,000
290
1,200
4,600
7,500
NATIONWIDE
EMISSIONS
(lb/yr)
5, 900
220,000
44,000
6,800
670,000
• 350,000
2,100
8,800
35,000
56,000
     3   Memorandum,  from Meardon,  K.,  Pacific Environmental
Services,  Inc., to Group IV Resins  Docket No. A-95-45.  March 24,
1995.  Determination of MACT Floors for Equipment Leaks.
                                10

-------
 Table 6.  Estimate of Baseline Nationwide HAP
from Equipment Components on Folyether Polyols
                 Lines that Use Epoxides
Emissions
Production
EQUIPMENT
COMPONENT
Connectors
Connectors
Pumps -
light
liquid
Pumps -
light
liquid
Pumps -
light
liquid
Pumps -
light
liquid
Pumps -
light
liquid
Pumps -
heavy
liquid
Pressure
Relief
Valves
Pressure
Relief
Valves
Pressure
Relief
Valves
Pressure
Relief
Valves
CONTROL
Quarterly
LDAR with
10,000 leak
definition
No controls
100 percent
with control
option B
HON LDAR
Monthly
monitoring
at 10, 000
ppm
Quarterly
LDAR at
varying leak
definitions
No controls
No controls
100 percent
with control
option B
Quarterly
LDAR with
500 ppm leak
63 percent
with control
option B
Quarterly
LDAR with
10,000 leak
CONTROL
EFFICIENCY
(percent)
Emission
factor
from PES
memorandum
0
100
75
69
45
0
0
100
Emission
Factor
from PES
memorandum
63
Emission
Factor
from PES
memorandum
AVERAGE
CONTROLLED
EMISSIONS PER
FACILITY
(Ib/yr)
7,300
30,000
0
53
2,300
10,000
1,900
100
0
16,000
3,000
41,000
NATIONWIDE
EMISSIONS
(Ib/yr)
54,000
220,000
0
394
17, 000
78,000
14,000
750
0
120,000
22,000
304,000
                               11

-------
 Table 6.   Estimate of  Baseline Nationwide HAP
from Equipment Components  on Polyetner Polyols
                 Lines  that  Use Epoxides
Emissions
Production
EQUIPMENT
COMPONENT
Pressure
Relief
Valves
Compressors
Sample
Connectors
Sample
Connectors
Sample
Connectors
Open Ended
Lines
Open Ended
Lines
Open Ended
Lines
TOTAL
EMISSIONS
(tons per
year)
CONTROL
Varying
percentages
with control
option B
Quarterly
LDAR w/ 500
ppm
100 percent
with control
option B or
C
38
percent with
control
option C
No controls
100 percent
with control
option B
86 percent
with control
option B
No controls

CONTROL
EFFICIENCY
(percent)
15-22
Emission
factor
from PES
memorandum
100
38
0
100
86
0

AVERAGE
CONTROLLED
EMISSIONS PER
FACILITY
(lb/yr)
180,000
4,000
0
500
2,200
0
180
86

NATIONWIDE
EMISSIONS
(lb/yr)
1,300, 000
30,000
0
3, 700
16,000
0
1, 400
640
1,800 tons/yr
             Made b   Polmerizing Tetrahdrofmran
     There are three  facilities in the nation that use THF to
make polyether polyols,  but  only one of these three use organic
HAP in the process.   Therefore, nationwide equipment leak
emissions for polyols made by the polymerization of THF are the
emissions from the one facility in this subcategory that uses
organic HAP.   Further,  it should be noted that the facility that
                               12

-------
reported emissions did not  report  any  controls  for equipment  leak
emissions.  Equipment leak  emissions from polyols made by the
polymerization of THF are summarized in  Table 7.

   Table 7.   Estimate of  Baseline Nationwide HAP Emissions from
    Equipment Components at Polyether  Polyols Facilities that
                    Polymerize Tetrahydrofuran
EQUIPMENT COMPONENT
Valves - vapor
Connectors
Pumps - heavy liquid
Pressure Relief Valves
Sample Connectors
Open Ended Lines
TOTAL (tons per year)
NATIONWIDE ESTIMATE
EMISSIONS (pounds per
OF
year)
23,000
212,000
4,660
28,100
2, 900
656
136
Storage Vessels

Polyols Made with Epoxides

Polyols production facilities  store epoxides  and other-HAP
onsite.  The nationwide baseline  emissions estimate of  HAP
emissions from the storage of  these substances  was made on  a per
pollutant per control option basis  (pollutant/control).
Nationwide baseline emissions  were estimated  by first calculating
the average emissions from each pollutant/control for each
storage vessel at each facility in the EPA/SPI  database.  To
yield a nationwide emissions estimate, the following calculation
was made for each pollutant/control:
     EEO,p  (Ib/yr)
                    = 1.4
                    = 195
   AVG, EO, P
1.7 * 82
          82
     Where:
     EEOfP  =  Nationwide,  baseline emissions from EO stored in
     pressurized vessels, Ib/yr;
     NV,EO,P=  Average  number of EO, pressurized vessels per
     facility;
     EAVG.EO.P  =   Average emissions per vessel from EO pressurized
     vessels,  Ib/yr;  and
     82 =  Number of facilities  in  this  subcategory  in the
     nation.
                                13

-------
Table 8 summarizes the inputs into the calculation for nationwide
emissions and the resulting emissions estimates from storage
vessels for each HAP stored per each control option for polyether
polyols facilities.

Table 8.  Baseline Nationwide HAP Emissions Estimate from Storage
    Vessels at Facilities that Produce Polyether Polyols with
                             Epoxides
POLLUTANT
EO
PO
PO
PO
Toluene
Toluene
Ethylene Glycol
Formaldehyde
Hexane
Methanol
Acrylonitrile
Diethanolamine
TOTAL (Ib/yr)
AIR
POLLUTION
CONTROL
Pressurized
Pressurized
Internal
Floating
Roof
Fixed Roof
venting to
Incinerator
Pressurized
Fixed Roof
Fixed Roof
Fixed Roof
Pressurized
vessel to
vent
condenser
Fixed Roof
Pressurized
Fixed Roof

AVERAGE
NUMBER OF
VESSELS PER
FACILITY
1.4
1.6
0.10
0.20
0.10
0.10
0.10
0.10
0.10
0.10
0.10
0.20

AVERAGE
EMISSIONS
PER VESSEL
(pounds per
year)
1.7
1.5
1510
0
74
0 (seems
low)
0.055
89
320 (seems
high)
6.4
0
10

NATIONWIDE
EMISSIONS
(pounds per
year)
200
200
12,000
0
600
0
0.41
740
2,600
53
0
160
17,000
     Polols Made b  Polmerizing
     Nationwide baseline storage emissions from facilities that
polymerize THF equals the emissions from the one facility that
uses and  stores organic HAP.  Emissions were reported to be 349
Ib/yr  from one fixed-roof toluene storage vessel.
                                14

-------
I
I
         Wastewater

              Polyols Made with Epoxides

              Organic-containing wastewater streams result from  (1)  the
         direct contact of water with organic compounds during chemical
         processing and  (2) contamination of " indirect -contact"  water.
         There were five wastewater sources in the database, which were
         (1)  scrubber effluent; (2) equipment washing; (3) direct  contact
         with organics from "other process equipment" (the "other  process
         equipment" category consisted of decanters, solvent recovery,
         condensers, and cooling tower blowdown)  (4) vacuum system
         effluent; and  (5) filter washing.  The average flowrate and
         organic HAP concentration of the wastewater streams in  the
         EPA/SPI database were calculated for each wastewater source
         listed above.  Also calculated from the database were the average
         number of each model wastewater stream per facility.  These
         average flow rates and organic HAP concentrations for each
         wastewater source make up the model wastewater streams  for this
         industry.

              Emissions were calculated from each of the five wastewater
         model streams in the following manner:

              EWW/MODEL (lb/yr)  = FAVG/MODEL * C^P/MODEL *  Fe * TMODEL * UC

              ESCRUSBER         =  29 * 490 * 0.5 * 32,640 * 1/453,600
                             =  511 lb/yr

              Where .-
              EWW/MOOEL    =  Emissions from each model wastewater  stream,
                        lb/yr;
              FAVG/MODEL    =  Average flowrate of each wastewater model
                        stream, liters per minute;
              CHAP/MODEL    =  Average organic HAP concentration of  each
                        wastewater model stream,  milligram per liter;
              Fe        =  Fraction emitted, from the HON;
              TMODEL      -  Duration of time the model wastewater stream
                        flows per year,  minutes per year; and
              UC        =  Unit conversion, 1 mg/453,600 Ib.

              The duration of each wastewater model stream flow  was
         estimated from information provided in the process vent section
         of the EPA/SPI questionnaire responses.  Specifically,  the
         average number and average duration of vent stream occurrences
         reported for process vents from the reactors and vacuum systems
         were calculated.  The average number of reaction runs per line
         was calculated to be 544 events per year, with the average
         reaction duration calculated to be 103 minutes.  Similarly, the

                                         15

-------
average number of evacuation events per year was calculated to be
375 with an average duration of 137 minutes.

     Also,  in order to estimate the duration of the wastewater
flows per year,  the following assumptions were made:
     1}   Scrubber effluent  flowed for one hour after the
     completion of each reaction;
     2)   Equipment is  washed down once each day, with wastewater
     flow lasting one  hour;
     3)   Other process equipment wastewater events  (decanters,
     condensers,  and cooling tower blowdown) occur once per
     reaction and last 30 minutes each;
     4)   The wastewater from vacuum systems flows for the same
     amount of time that the vents from the vacuum systems
     release to the atmosphere;
     5)   Filter washing occurs once each day,  with wastewater
     flow lasting one  hour.
The durations of the wastewater streams from each wastewater
source are listed in Table  9.

 Table 9.  Baseline Nationwide HAP Emissions  from Wastewater  at
     Facilities that Produce Polyether Polyols with Epoxides
WASTEWATER
MODEL
STREAM
Scrubber
Effluent
Equipment
Washing
other
Process
Equipment
Vacuum
System
Filter
Washing
TOTAL
(TONS/YR)
NUMBER OF
STREAMS
PER
FACILITY
1.33
4.00
2.50
1.83
0.83

AVERAGE
FLOWRATE
(liters/
min)
29
21
200
5.6
17

AVERAGE
ORGANIC HAP
CONCENTRATION
(parts per
million)
490
500
1,300
70,000
850

DURATION
(min/yr)
32,640
21,900
16,320
51,375
21,900

PREDOMINANT
HAP
EO
PO
Toluene
PO
Toluene

NATIONWIDE
EMISSIONS
(Ib/yr)
32,000
58,000
890,000
2,300,000
22,000
1,700
     Scale-up to Nationwi<^«;»

     The emissions from each wastewater model stream were scaled
up to nationwide emissions  by multiplying the organic HAP
emission rate from each wastewater model stream by the percent of
facilities that reported data for their wastewater streams  (that
                                16

-------
I
I
        is, 7  facilities reporting wastewater information/12 facilities
        in the database = 58 percent) .   The percentage of facilities with
        wastewater data in the database was needed in the calculation
        because, unlike the other  emission sources,  the questionnaire
        instructed facilities not  to respond to this portion of the
        questionnaire if their wastewater streams were dilute enough or
        had a  small enough flow  to have "insignificant" emissions.
        Nationwide wastewater organic HAP emissions  from scrubber
        effluent streams  (ENrWW/ScRUBBER)  were calculated in the following
        manner:

             EN,HW/SCRUBBER (lb/yr)  = EWH/MODEL  *  0-583  *  82  *  1.33
                               = 511.5  lb/yr * 0.583 * 82 * 1.33
                               = 32,525 lb/yr
             where:
             EWW/MODEL    =  Nationwide emissions from  the scrubber model
                       wastewater  stream,  lb/yr;
             0.583     =  Portion  of facilities in the total EPA/SPI
                       database  with significant wastewater emissions;
             82        =  Nationwide estimate of polyol facilities; and
             1.33      =  Number of scrubber streams per facility,  from
                       the database.

        The nationwide baseline  emissions of wastewater emissions from
        polyols facilities that  use epoxides were then totaled for the
        five wastewater model streams.   These emissions are presented in
        Table  9.

             Polyols Made by Polymerizing1 Tetrahyd.rofmran

             Nationwide baseline emissions from wastewater at polyols
        facilities that polymerize THF  are the total emissions from the
        one facility that uses an  organic HAP in the process.  The
        calculation approach and duration assumptions used to calculate
        organic HAP emissions from the  model wastewater streams were also
        used to calculate the organic HAP emissions  from this facility's
        wastewater streams.  Wastewater organic HAP  emissions for the
        facility that polymerizes  THF were 213 lb/yr from two streams
        with a total flowrate of 115 liters per minute and an organic HAP
        concentration of 302 ppm.

        SUMMARY

             Total nationwide baseline  HAP emissions from polyols
        production with epoxides is 4,100 tons per year.  The majority
        of the emissions, 44 percent and 42 percent,  are from equipment
        leaks  and wastewater, respectively.  The nationwide baseline
        emissions from polyols production with THF are 160 tons per year,
        which  is four percent of the total nationwide emissions from the
        source category; however,  THF polymerizers are only 2.4 percent

                                         17

-------
of the number of  total  facilities in the source category.
Further, the majority  (85  percent) of emissions from the THF
polymerizers come from  equipment leaks.

     The summary  of nationwide baseline emissions from all
emission sources  in the polyols production process are presented
in Table 10, for  polyols that use epoxides, and in Table 11 for
polyols made from the polymerization of THF.

   Table 10.   Summary of Nationwide  Baseline BAP  Emissions  from
                    Polyols Made with, Epoxides
SOURCE
Process Vents
Equipment Leaks
Storage Vessels
Wastewater
TOTAL
NATIONWIDE BASELINE
EMISSIONS (Total HAP tons
/year)
580
1,800
8.5
1,700
4,100
   Table 11.  Summary of Nationwide Baseline HAP Emissions from
           Polyols Made by Polymerizing Tetrabydrofuran
SOURCE
Process Vents
Equipment Leaks
Storage Vessels
Wastewater
TOTAL
NATIONWIDE
BASELINE EMISSIONS (Total
HAP tons /year)
24
136
0.18
0.11
160
                                18

-------
                                                                      IA-96-38
                                                                      H-B-4
     Z7C
I   Jd /R Incorporated	1996 EPA Outstanding Small Business Contractor
        MEMORANDUM

        Date:     April  25,  1997    .        '        '           "

        Subject:  Estimated  Impacts for .the Polyether Polyols NESHAP

        From:     David  Hendricks,  EC/R Incorporated T^pC
                  Joanne Seaman,  EC/R Incorporated     (j    '

        To':       David  Svendsgaard,  EPA/OAQPS/ESD/OCG


             The purpose of  this memorandum is  to  present the estimated
        impacts of the hazardous air pollutant  (HAP)  control options for
        existing sources under  the  proposed polyether polyols
        manufacturing standards.  This memorandum• is  divided into six
        sections, beginning  with a  description  of  regulated emission
        sources-and the  regulatory  requirements  for each source.  The
        second section presents  the methodology used  to develop the  model
        plants for which the impacts were estimated.  .'The remaining
        sections present the primary environmental, impacts, secondary
        environmental impacts (air-  pollution, water pollution, and.solid
        and hazardous waste  impacts)',  energy impacts,  and costs.

        DESCRIPTION OF EMISSION SOURCES AND REGULATORY REQUIREMENTS

             There are four  primary potential sources of organic HAP
        emissions in the polyether  polyols manufacturing process:
        process vents, storage  vessels,  equipment  leaks,  and wastewater.
        The emissions from these sources,  and the  types of controls  being
        used, are discussed  in  the  following sections.

             Another potential  emission source  considered in other
        NESHAPs is unloading operations,  and these were investigated for
        this industry as  well. "  Industry representatives indicated that
        all organic HAP  unloading operations are conducted in a closed
        vent system .due  to the  explosive'nature  of ethylene oxide  (EO) .
        Therefore,  no emissions  are expected to  result from this portion
        of the process.   .    .     '

        Emissions and Controls on Production Units that Use Epoxides

             Process  Vents.   The  most  'significant  organic HAP loss from
        the polyether polyols manufacturing process occurs' at the 'end of
        the batch when the unreacted epoxide is  vented.   Ethylene oxide
        reacts quite  quickly, so  normally the residual EO in the vent is
        quite low.   Propylene oxide (PO)  reacts  an order of .magnitude
        slower at a given catalyst  level,  temperature,  and pressure.
                South Square Office                           Research Triang£ Park Office
       3721-D University Drive • Durham. North Carolina 27707        2327 Englen Drive. Suite 100 • Durham. North Carolina 27713
         Telephone: (919) 493-6099 • Fax: (919) 493-6393              Telephone: (919) 484-0222 • Fax: (919) 484-0122

-------
manufacturing polyether polyols from THF.  Based on data received
from industry,  it  appears that monitoring  of equipment leaks is
not generally conducted at such facilities.

     Wastewater.   Neither facility  that polymerized THF generate
Group 1 wastewater streams.  Therefore, no wastewater controls
were reported.

Regulatory Requirements

     The proposed  regulation for polyether polyols production
will address HAP emissions from five emission source types:  (1)
process vents  (epoxide emissions) ,  (2) process vents (nonepoxide
HAP emissions),  (3)  storage tanks,  (4) equipment leaks, and (5)
wastewater treatment.  The regulatory requirements for existing
and new sources are summarized in Tables 1 and 2, respectively.

DEVELOPMENT OF  MODEL PLANTS

     The estimation of primary impacts, secondary impacts, and
costs required  the development of model plants that divided the
industry into "typical" groups based on similar characteristics.
Three such model plants were developed.  The small model plant
category defines the segment of the industry that uses an epoxide
(i.e.,  EO and PO)  and has a production capacity of 50 million
pounds per year (MMlb/yr) .  This category  also includes one
additional facility to account for  the polyether polyols made
from the polymerization of THF subcategory.  The large model
plant category  also uses an epoxide-based  process, but has a
production capacity of 200 MMlb/yr.   The final model plant
category uses a HAP solvent in the  catalyst extraction step.
This category is referred to as the catalyst extraction model
plant and has a production capacity of 100 MMlb/yr.

     The model  plant production capacities were established using
the production  capacities listed in the Chemical Economics
Handbook1 (CEH)  for each facility in the EPA database.   Other
characteristics used to define the  model plants were process vent
exhaust stream  characteristics, storage tank capacity and
throughput, total  number of each type of equipment covered in the
HON equipment leaks regulations  (40 CFR 63, subpart H) and
existing level  of  control, and wastewater  characteristics.  The
model plant parameters are summarized in Table 3, and the
development of  each parameter is discussed in the following
sections.

Process Vents

     The primary process vent exhaust stream characteristics
needed to size  control devices were flow rate, temperature,
pressure, and pollutant concentration. Information on each of

-------
reactor system's annual production capacity.  This penalty is
offset  to some extent by the improvement in raw material
efficiency due to converting more epoxide to product.  Once ECO
is complete,  the remaining emissions  are low.

      Storage.   The storage vessels used in the polyether- polyols
manufacturing process are a potential source of organic HAP
emissions.   Most of the storage vessels used are pressurized, and
all of  the reported EO and PO storage vessels are pressurized as
a precaution due to the explosive nature of EO.

      Equipment Leaks.   Emissions of organic HAPs may occur around
pumps,  valves,  and other components,  and are referred to as
"equipment leaks."  Sixty-nine percent of the facilities in the
database have  leak detection and repair (LDAR) programs.  The
LDAR  program alerts the facility of any leaking components so the
problem can be fixed,  thereby minimizing emissions.

      Wastewater.   Another potential source of organic HAP
emissions is  the processing of wastewater from the manufacturing
process.   The  majority of wastewater  emissions from this industry
are a result  of the steam ejectors used on the batch reactors to
pull  a  vacuum  on the vessel.   Although 17 percent of the
facilities  in  the database have Group 1 wastewater streams that
require controls according to the HON,  none of the facilities for
which data were received control air  emissions from the
wastewater streams.   A Group 1 wastewater stream is defined by
the HON as  having a minimum organic HAP concentration of 10,000
parts per million (ppm)  and a minimum stream flowrate of 10
liters  per minute.

Emissions and  Controls on Production  Units that use
Tetrahydrofuran

      Even though tetrahydrofuran (THF)  is not an organic HAP,
there are organic HAP emissions from  one of the facilities in the
database that  polymerizes THF.   At this facility, the catalyst
extraction process resulted in organic HAP emissions.  An organic
HAP solvent is used for the catalyst  extraction, and none of
these emissions to the atmosphere are controlled.

      Process Vents.   The polymerization of THF occurs on a
continuous  basis.   THF is not an organic HAP and organic HAPs are
not typically  used in the actual synthesis of the polymer.

      Storage.   Only one of the THF polymerizers stored an organic
HAP.   The storage vessel was  a Group  2 (not requiring controls
based on the HON criteria)  fixed-roof storage vessel.

     Equipment Leaks.   Potential organic HAP emissions may occur
around  pumps,  valves,  and other components in facilities

-------
manufacturing polyether polyols from THF.   Based on data received
from industry, it appears that monitoring of equipment leaks is
not generally conducted at such facilities.

     Wastewater.  Neither facility that polymerized THF generate
Group 1 wastewater streams.  Therefore,  no wastewater controls
were reported.

Regulatory Requirements

     The proposed regulation for polyether polyols production
will address HAP emissions from five emission source types:  (1)
process vents (epoxide emissions) ,  (2)  process vents (nonepoxide
HAP emissions),   (3) storage tanks,  (4)  equipment leaks, and  (5)
wastewater treatment.  The regulatory requirements for existing
and new sources are summarized in Tables 1 and 2,  respectively.

DEVELOPMENT OF MODEL PLANTS

     The estimation of primary impacts,  secondary impacts, and
costs required the development of model plants that divided the
industry into "typical" groups based on similar characteristics.
Three such model plants were developed.   The small model plant
category defines the segment of the industry that uses an epoxide
(i.e., EO and PO) and has a production capacity of 50 million
pounds per year  (MMlb/yr) .  This category also includes one
additional facility to account for the polyether polyols made
from the polymerization of THF subcategory.   The large model
plant category also uses an epoxide-based process, but has a
production capacity of 200 MMlb/yr.  The final model plant
category uses a HAP solvent in the catalyst extraction step.
This category is referred to as the catalyst extraction model
plant and has a production capacity of 100 MMlb/yr.

     The model plant production capacities were established using
the production capacities listed in the Chemical Economics
Handbook1 (CEH)  for each facility in  the EPA database.   Other
characteristics used to define the model plants were process vent
exhaust stream characteristics, storage tank capacity and
throughput, total number of each type of equipment covered in the
HON equipment leaks regulations (40 CFR 63,  subpart H)  and
existing level of control, and wastewater characteristics.  The
model plant parameters are summarized in Table 3,  and the
development of each parameter is discussed in the following
sections.

Process Vents

     The primary process vent exhaust stream characteristics
needed to size control devices were flow rate, temperature,
pressure, and pollutant concentration.   Information on each of

-------
w
>J
CQ

is







































•
u
u
o
CO
a
0
-*4
•
•
1




































a»
XX
&a
e
Hd
• U
44 C
• U
« 0
X »


































(
u
c
C

^
•
8
M
Oi
































JJ
•
>,
r4
a
AJ
0
u

a
«4

£u

*
• a
v o
•HX
SJJ
U
aa
• u
a u
o x
X •










JJ
u
3
4)
0
a
a
B
x
X
a
JJ
X
4
tJ
0
a
a
X
X

a
~*

g
e
u
x
x
o
A
•
a
o
Z

•
•
fl
X
X
&
M
€
a
0
u
o
a
to
•>•
o
>> <*
u a
C O u
U 01 •
U C U
3 JJ J7
0 « 3
MOW
Z
0
z


z
0



c
o
to
01

~6
V

0
u
fl
O)
0)
V4
a
OI
a
jj
c e
01 0
o —
Lj JJ
II U
a 3
•o
0 U
m u
n
.C 01 •" O
u£ o .c a
jj o JJ CU «*J
E CD 0% CO J3 M «
0 0 01 K O u -1
ki JJ 3 II 0 CB JJ
^COJJCD  o a> D oo u o« .u
au O3h
o. in ai « TI 3 o
^0jjjj otoonoijc
>ij:c^aviu£3j oe 3 e M to
o a j: o o-x£ ax o —
kj£U ajJfjjjjcoE

iaj£**4x . oa«c
^Q. Eamua x>u
jjai<-* JJ co>*4 10
Ur^OCUCUUxu « 01 C ti
jj IMO* ain i >i£ X
^* ^ 4J O
O O •** Q
a. a. 3 tu
z
o
X


2
o

0 01
!•• JC
<0 H

•0
U 0) 4»
4» C CD
u o m

W AJ 01
U « --»
(N or u •
0) C C
Q. a) O
aco E-
O (X 41 JJ
U H U 0
O -^3
^. Z 3T5
-^ O U 01
X 0) ^
a u
3 ai c:
O J=*-i O
W JJ O •«
o u u
S J-» W
O O C -^t
-C ^ O 6
e-«M o *


































<£
Z






<
z




1

o
•a
» >JS

O O -*4
a. cu 2
                                                                                                                          in

-------
CO

























*
o
o
h
g
O
n

e
o
*
a"

































o
|| Sourc
a
*•*
f*
H
e
•
a




















a
a

!>
•
e
0
o
a.
























0
a
14
O
JJ>
a

&

w c
O i
a a
0 U
JJ £
55
m
«*•
o

w
JJ
a



t|
^

0
g
O

a
i
e a
•0 0
X U
0 0
A •*
fl u
ss



*J
u
£
o
w
a

a
lj
e
u
IB
b
o
a
a
•J,
a
a
•H
o.
s
Q
qj

X
0
1
2


e

^4
X
o
Ck
H











Z
O
z

o
X



(U
JJ
(C C
Ol 0
•! •«•*
(U to
a. ca
~H
oo E
u u
0) O O
k* ai a B o>
U JC-H « U C JL-< M
aa a m i> »« a
— EBOI U 0) > 6 0 HI
•^OIJJOCOI U «J 0) O C Ol
UAJCJJOIOI U 3JJO)O>
o>-c ai > « ao > Of Sij^0-15
u u 01 C u QpJJ O C
O - a cn«j ai o> o B Z O " tj O-^
n x; 3 01 u o £ 3 m jj
C .C~40 33 J="oW3
> «« &JC 01 3 f Ckf 01
Q*OT) jJk4CjJO*O jJh

uo) JJJC-CJJE jjf-e
J-ff'OQJJ'HOCO'OAjJi-^O
MoaajjO'^utfQiajjo-^
b.uaaaciaib.aaa'auoi
o e
JJ C O
JJ (0 O -*4
C O)~H JJ
01 01 a) O
0» O kJ M 3
n ai Di E Ol
« a «» oi u
z
o
ox
c
-H 01
JJ U
-i 3
X O
u en
b J=
0) jj ta
f m •*< a;
JJ- 3-D
01 O — i
>i >. 01 X
"i •-< "O O
O O RJ O.
a. a. e u

2
O


i
X
O gj
UJC
n
u a «
01 C S
*J O O»
*H jj ra
O 10 -H
3
(N a* jj •
01 C C
Q, Q) O
3U E~4
O K 01 JJ
M f- U O
U -43
"Z 3-D
« O fr 0)
X 01 K
a u
3 01 C
O f *H O
o " Sin
S JJ B
01 O C-4
j= u o e
H. > 0)
rH ^H -O [L.
O 0 fl I

-------
TABLE 3 .  MODEL PLANT PARAMETERS
Parameter
Production Capacity
(MMlb/yr)
Process Vent Exhaust Stream
Characteristics
Flow (acfm)
Heat Content (Btu/scf)
Temperature ( °F }
Pressure (atm)
Pollutant
Concentration (Ib/scf)
Fixed-roof Storage Tanks
Number Per Facility
Capacity (gal)
Throughput (gal/yr)
Equipment Leaks - Number of
Uncontrolled Components
Compressors
Open-ended Lines
Sampling Connections
Pressure Relief Devices
Valves in Gas Service
Valves in Light Liquid
Service
Pumps in Light Liquid
Service
Connectors
Wastewater
Group I Wastewater Stream?

Small
50

250
8.
150
1
0.02

1
38,667
. 134,500

1
0
12
8
125
166
2
409
No
Model Plant
Large
200

250
8
150
1
0.02

1
38,667
134,500

1
7
2
11
3
392
13
385
Yes

Solvent
Extraction
100

250
8
150
1
0.02

1
38,667
134,500

0
58
6
34
57
218
.5
219
No

-------
these parameters was obtained from the EPA database for
facilities that generally matched the specifications  (process
type and production rate)  of each model plant.   An analysis of
this data showed that there was no apparent trend in the exhaust
stream characteristics from one model plant to  another.  For
example, the reported flow rate for facilities  corresponding to
the large model plant category was not significantly larger than
that for the small model plant category.   Consequently, the
process vent exhaust stream data for all three  model plant '
categories were used to develop a single "model" process vent
exhaust stream.  This one exhaust stream was used to size all
control devices for all model plants.

Storage Vessels

     Information in the EPA database was used to determine the
capacity and throughput of non-pressurized tanks that would be
applicable to the HON storage requirements.  As there was no
general trend in the capacity or throughput of  the tanks with the
size of the facilities, an average capacity and throughput was
calculated for all applicable tanks in the database.  This
"model" tank was used for all three model plant categories.  The
database also showed no more than one tank per  facility that
would require control; therefore,  it was assumed that each model
plant would have only one such tank.

Equipment Leaks

     Extensive LDAR program information was provided in the
questionnaire responses used to develop the EPA database.  This
information was analyzed to determine the average number of each
equipment type  (i.e, compressors,  open-ended lines, sampling
connectors, pressure relief devices, valves in gas service,
valves in light liquid service, pumps in light  liquid service,
and connectors) for each model plant category.   The
questionnaires showed that most facilities currently have some
type of LDAR program in place.  While some of these LDAR programs
met the HON requirements,  others did not.  For those programs
that did not meet the HON requirements, the model plants must
address  (for the purpose of estimating impacts) this intermediate
level of control.

     The total number of each equipment type was first estimated.
This was done by selecting facilities from the database that
closely corresponded to the model plant categories in terms of
production capacity and manufacturing, type  (with or without
catalyst extraction) .  Next, using the entire database of •
facilities, the fraction of each component type controlled at a
level equal to or more stringent than the HON was determined.

-------
For those components controlled at  a level less  than the HON,  a
control efficiency was assigned to  each type of  control based  on
the control strategy reported,  and  the fraction  of each component
type at each of  these levels  of control was determined.  These
two numbers were then multiplied together,  and the resulting
number was used  as the portion  of components with a control
efficiency calculated to be equivalent to the HON level of
control.  Finally,  the percentage of components  at the HON level
of control and HON equivalent level of control were subtracted
from the total number of components.   The resulting number of
uncontrolled components was used to estimate the cost impacts
(see Attachment  7) .

     In two instances (open-ended lines for small model plants
and compressors  for catalyst  extraction model plants) , there were
no components reported for  the  facilities selected from the
database.  Based on the database as a whole and  knowledge of the
industry, there  is. no indication that this  should be the case.
For example,  catalyst extraction model plants have valves in
vapor service, so it is expected that there should be at least
one compressor.   Further analysis of the database, however, did
not provide adequate information to accurately estimate the
number of these  components  per  model plant.   As  a result, the
data was used as  is  with the  understanding  that  impacts may be
slightly underestimated.

Wastewater

     Wastewater  streams reported in the EPA/SPI  questionnaire
were compared to  the HON Group  1 wastewater characteristics.
Only two facilities  reported  wastewater streams  that met the
Group 1 requirements,  and both  of these facilities fell into the
large model plant  category.   No wastewater  emission controls were
reported for either  facility.   For  the purposes  of the impacts
analysis, it was  assumed that small and catalyst extraction model
plants have no Group 1 wastewater streams.

PRIMARY ENVIRONMENTAL IMPACTS

     This section  presents  the  nationwide primary air pollution
impacts (reduction of HAP emissions)  resulting from the
implementation of  the proposed  standards.   The primary impacts
were calculated by applying sufficient controls  to each emission
source at each model plant  type to  bring them into compliance
with the standards.   The control technologies and efficiencies
used in these calculations  are  shown in Table 4.  The model plant
emission reductions  were then scaled up to  nationwide levels
based on the number  of model  plants and the existing level of
control.  For example,  it was estimated that there are 38
facilities nationwide that  would be characterized by the large
model plant,  but  only 11 of these facilities would require

-------




Q
W
CO
o
04
o
QS
QJ

W
33
CLj
=* ,
iH
as o
EH OS
H EH
3 2
o
KJ r \
>l CJ
COMPL
IRING
O D
EH O
u
Q (£
Cd
EH CO
0 EH
W 2
iJ 3
H. T
»H
CO 04

CO J
u u
H Q
O O
OS
hJ
O &.
2 0
IOL TEC
NUMBER
tt.
S§
0 §
O ri.
U
CO
Cb Q
O OS
<
>H Q
OS 2
S E"
1 W
cn

•1
•So
8^
jj °
«.§
5*5
0 0)
EH







0)
u
M
3
O
CO
c
0
•H
to
to
-H



0 „ ,
^0
-H ° . 01
•> jj
S !-i 0 "£
*-• rti rH «
o_g^ ic
-H-9 SrH
i i c ^-* r\>
•AJ ^M m
!§S
2



rH 4-)
(U C
rrt /rt
W 'U
O_j
^^
S cu






rt!
vo - ° -








^f«
oo in „ «
-l -U CO
3 0) «C
M jj o s <:
U C rH O ^>.
CO H Dt4 33 2










10
y
U CO (C
4J M <1)
C C J
(U (C ^
> E-" -U 0)
C -U
CO 0 0) (C
w o) e &
0 m a 0
o w -H .u
O O 3 to
^ 4J O1 (0
O4 CO W S




10
O
cs







rH
rH
(C
E
CO





rH 0 00 ^
T-l r^ CO w









oo in r, oo
cr> -i
> EH JJ 0
C JJ
to 00 (C
to D) e 5
0 tc a 0
U H -H JJ
0 O3. co
M JJ & <0
cu co w s




00
m







0
Dl
ki
(C
•J




^,^4
« s •» s |








^-^
oo oo in .. ^
CT> CTv  W 2




^ X
O
O JJ
OS M
(C
rH DI a
OJ C XJ
C-H 3
0 0 H JJ CO
W i-i 0 (C
(C (0 JJ O 2 <
rH rH C rH O ^
Cb Cb H b X Z



CO
01
w
TJ
w
>r{
X


i i to tn
04 .y
to to < to tc
JJ JJ 33 X 0
C C C J
000 (C M
> > T3 E" JJ 0
•H C JJ
to to x 0 0 n)
to to o Di e 3
0 0 a ic a 0
U U 0 M -H JJ
O O C O 3 tQ
^ ^ o jj ty cc
04 04 2 co w S





rH


c
o
J_l ,_J
*J *("!
CO JJ
>i U
rH (0
1C M
JJ JJ
03 X
U Cd

4J
c
13
^
Q.
01
c
o
•o
c
a
>,
C
O -H
Dl (0
4) 4)
u n
10'
U >,
.a w
3 n
WJ
— i Dl
0 C
a -"
u
b u
01 •*<
j: u
jj
to ^3
>, «
^4 r^
o »^
a. . o
tb ^*
0 g jj
5* §
iu U
?o .
C *
-< e rt
jj g io
n ^ «
01 u W
S* >,
2^ &
&^ e
V AI C
u g -^
g, k4
01 ;? a
Si*
o °- ai
rH m JJ
a.g c
JJ 3
o» o
"H E 6
- 2 «
Q ^ -*^
Ovu -
•o o cr
o> ^ -H
a , ^
•^ ^
14 = 01
Oi n OI 0)
6.2 -0 &
o o 3 S
o »L *^ *J
.? 0
£ R -S c
« a §
.. •" C
o 5 o o
= •=21
C tj W K
2 « o, S
a- c „
_. ° i 2
'^ Q. «J 4J
ai "• a
*T3 « O OI
0 ? -i C
e" - 5
-> 01 rt U
-H J? -0 0
10 ^ C 0
E "u w °
u ~ 0) fl
m JJ
0 S = S
£ °
«! a » b
£ Sj -c a
t- Hi &• ?»

10 .a o
o

-------
I
         control of epoxide emissions from process  vents (see Attachment 1
         for further details of the number of model plants requiring
         control) .   This information is also presented in Table 4.

              Table 5 presents the emission reductions by emission source
         for each model plant.  Table 6 summarizes  this information by
         emission source for all model plant types, along with the
         nationwide baseline emissions2 and percent reduction  from
         baseline.   As shown in Table 6,  the proposed  standards are
         expected to reduce HAP emissions by 1,810  Mg/yr,  or 42.5% from
         baseline.   The vast majority of emission reductions are achieved
         through the control of process vents and equipment leaks.  These
         two source types account for approximately 85 percent (1,550
         Mg/yr)  of total emission reductions.  The  methodologies used to
         calculate the emission reductions is presented in Attachment 1.

         SECONDARY ENVIRONMENTAL IMPACTS

              The  intent of the proposed standards  is  to reduce HAP
         emissions;  however,  the application of control technologies can
         lead to other secondary environmental impacts.   These secondary
         impacts can have a positive environmental  impact,  such as the
         improvement of wastewater quality through  the implementation of
         wastewater controls,  or a negative impact.  'Negative impacts may
         result,  for example,  from air pollutants emitted as a result of
         combustion sources (e.g.,  flares)  or the generation of a solid or
         liquid  waste stream.   The secondary impacts on air pollution,
         water pollution,  solid waste,  and hazardous waste are discussed
         in this section for each emission source.

         Air Pollution

              The  secondary air pollution impacts associated with the
         proposed  standards are the increased criteria and toxic pollutant
         emissions  resulting from the on-site combustion of organic HAPs
         and fuels.   The combustion of organic HAPs and natural gas in a
         flare generates emissions of nitrogen oxides  (NOX) , carbon
         monoxide  (CO) ,  and volatile organic compounds (VOC) .   The
         combustion of fuel oil generates these same criteria pollutant
         emissions,  as well as particulate matter (PM)  and sulfur oxides
         (SOX) .  In addition,  fuel oil combustion generates emissions of
         the following HAPs:   arsenic (As), beryllium  (Be),  cadmium (Cd) ,
         chromium  (Cr) ,  lead (Pb),  manganese (Mn) , mercury (Hg) ,  and
         nickel  (Ni).

              Criteria pollutant emissions will also occur as a result of
         the combustion of coal,  oil,  or  natural gas used to generate the
         additional  energy needed by the  control equipment.   These off-
         site air impacts  were not included in this analysis,  although an
         estimation of the additional energy requirements is presented
         later in this section.
                                        11

-------
           TABLE 5.   EMISSION REDUCTION BY MODEL  PLANT
Model
Plant
Small
Large
Catalyst
Extraction
Emission Reduction (Mg/yr)
Process
Vents
32. 2a
350
64.2
Storage
Tanks
3.1
6.2
3.1
Equipment
Leaks
104a
794
199
Wastewater
0
249
0
Total
139
1400
266
a Includes  emission reduction from the THF facility.
  TABLE 6.   EMISSION REDUCTION FROM BASELINE BY EMISSION SOURCE
Emission Source
Process Vents
Storage Tanks
Equipment Leaks
Wastewater
Total
Total Emission
Reduction
(Mg/yr)
447
12.4
1100
249
1810
Baseline
Emissions
(Mg/yr)
604
22.2
1940
1700
4260
Reduction
From Baseline
(%)
74.0
55.9
56.7
14.6
42.5
                                12

-------
     There is no  on-site combustion associated with the required
control technologies  for storage tanks or equipment leaks.
Therefore, no secondary air impacts are expected from these
technologies.

     Table 7 presents the estimated criteria air pollutant
impacts for process vent and wastewa-ter control technologies.
The total criteria air pollutant emission increases are estimated
to be 79.3 Mg/yr.  Table 8 presents the toxic emissions resulting
from the fuel oil combustion.   Total toxic pollutant emissions
are estimated to be 25.7 kg/yr.   The following sections briefly
describe the methodologies used to calculate the emissions  shown
in Tables 7 and 8.  More detailed information is presented  in
Attachment 2.

     Process vents.   Secondary air impacts from process vent
controls are a. result of the combustion of organic  HAPs and
supplemental fuel in  flares,  as well as the combustion of fuels
to generate steam.  Emission factors for flares were obtained
from Section 13.5 of  AP-42.3  The emission  factors  (Ib/MMBtu)
were multiplied by the heat content of the exhaust  stream to
obtain the emission rate.

     In order to conservatively estimate secondary  air emissions
resulting from the process vent control requirements,  it was
assumed that the steam used by the flares is generated on site.
Emissions were calculated for a small industrial boiler with a
thermal efficiency of 80 percent burning No.  2 fuel oil (0.5
percent sulfur by weight).

     The steam requirements of the flares were determined by the
OAQPS Control Cost Manual (OCCM)4 spreadsheet.  The steam
requirements were converted to gallons of fuel oil  by dividing by
140,000 Btu/gal, then dividing by 0.80 to account for the boiler
efficiency.   The fuel oil usage was used to estimate emissions
from steam generation.

     Criteria and toxic air pollutant emissions from the
combustion of fuel oil were estimated using emission factors from
Section 1.3  of AP-42.   Fuel oil  usage was multiplied by the AP-42
emission factors to obtain emissions per model plant.

     The EPA database was used to determine the percentage  of
facilities that are not controlled at the level of  the proposed
standard (98 percent  control) .   Analysis of the database showed
that approximately 30 percent  of the facilities are either,,
uncontrolled or controlled at  less  than 98 percent.   This
percentage was then multiplied by the total number  of each  model
plant type to obtain  the number of  uncontrolled facilities  per
model plant  type.  The emissions per model plant were multiplied
by the number of uncontrolled facilities in each model plant
type,  then added together to obtain nationwide emissions.

                                13

-------


























CO
UJ
g
rs
O
CO
z
o
H
CO
3
CO
2
O
0
tc.
o
u.
CO
1^
<
1
cr
tr
O
o
UJ
CO
•
UJ
CO





ca
23
1



,
i


o
0
UJ
in
CM




i
•



o
UJ
CO
• k-
O) ^ "S ^E
2^ « O O
a>
(UJ
























































O
UJ
CM
**
^^
O
Ul
CM
^

O
o
UJ
^B


T»
0
UJ
in
V*


O
o
LU
CM



(O
in
in
d

CM




in
CM
r>»


§

_
CM





^
CM
d
O
S
u.
























































0
LU
CM
^*
^v
0
LU
CO
CD
^H
010
J4
LU

"^

O
o
UJ
V*


^w
o
u
e^

^
O
LU
CO
CO


Y*
a
UJ

^w


S
UJ
CM
^


g
UJ
^

^r co
cviio
oio
UJIm

•o

a
SI
•a
c
a
co
1
(D

co
5
u
£
S
£
«
U)

a
X
ID
"o
to
3
CO
1

S

-------
                                   LUi
           !9
           LU
  8;

  LU lUil
  inieoil
    cc

    li-



    en

    s?
    in

    o
   ai


   u.

   CM

   O


   O
  DC


  CD

  CO
  LU

  o
  cc


  o
  CO



  o
  00
 o

 cc
 O
 u.
 o

 0.
g
o
III
CO

CO

III
_1
CD
                                   LU
        s

        LU
        (O
        c\i
       LU
                                  LU
           LUi
           (V
                                                   ill
                                                      LUI

                                                    '(W!
 sl;
 LUI
    UJII
 OlTT-l
                       LU
                         'LU!
      s

      LU
      CU
          ev
          CO
                                    LU
                         Sin1

                       , 19!
                      LUILUJ!
LU
                                CM
                                          O)
        in
        a
        cu

        a
        XI
        a
       S    S.
       evil

-------
     Wastewater.  Steam stripping was  the  control  technology that
was analyzed for control of air emissions  from wastewater.
Secondary air impacts associated with  steam  stripper operation
can occur from combustion of fuels  for steam generation and the
handling or combustion of the recovered organic  compounds.   For
the purpose of this evaluation, it  was assumed that recovered
organic compounds are handled properly and that  no significant
emissions occur from the handling process.

     As with process vents, it was  assumed that  additional  steam
must be generated on site by a small industrial  boiler.  The
boiler emission calculations and methodology for scaling to
nationwide emissions were completed in a manner  similar to  that
described above for process vents.

Water Pollution

     Potential water pollution impacts may result  from each of
the control technologies evaluated, with the exception of flares.
Positive effects will result from wastewater and equipment  leak
'controls, while negative effects will  be seen from process  vent
and storage tank controls.

     Steam strippers remove organic compounds from wastewater,
thereby improving the quality of the wastewater  being discharged
to the wastewater treatment plant,  or  discharged from the
facility.  As this wastewater will  have a  reduced  level of
organics, water quality impacts will be positive.

     Control of equipment leaks will reduce  the  amount of organic
material that may eventually be contained  in a wastewater stream.
For example, the leaking material may  enter  a wastewater stream
through equipment washdowns or from stormwater runoff.   However,
the nature of these organic materials  is such that the majority
of the material will evaporate before  or soon after entering the
wastewater stream, so that the final effects will  be minimal.

     The potential for water pollution is  also present with
storage tank improvements.  Before  an  internal floating roof can
be installed or upgraded,  the tank  must be emptied and cleaned.
A small amount of wastewater will be generated during tank
cleaning, but it is not expected that  this source  of water
pollution will be large enough to have a measurable impact  on
water quality.

     The most significant impact on water  quality  is associated
with process vent control.  Control of organic HAP emissions
using combustion does not result in any water quality impacts, as
no water effluents are generated through the use of a flare.
However, the use of a scrubber for  organic HAP control results in
increased water consumption and an  effluent  wastewater stream.


                                16

-------
I
              In a scrubber control  system, water Is used to remove water-
         soluble pollutants contained in the exhaust air stream.   The
         amount of wastewater generated is equal to the amount of water
         needed by the scrubber  to absorb the pollutants (assuming there
         is no recycle of scrubbing  water) .  The scrubber water flow rate
         used in the scrubber cost analysis was 17 gallons per 1000 ft3  of
         exhaust gas flow, which is  the basic..design assumption used in
         the OCCM.  The exhaust  flow rate to the scrubber was 220 ft3,, and
         the scrubber was assumed to operate 8,000 hours per year.   This
         equates to an annual wastewater generation of approximately 1.8
         million gallons.

              Nationwide water quality impacts  were then calculated by
         multiplying the annual  scrubber water  usage rate by the number of
         facilities that will install such a system.  Based on an analysis
         of the EPA database,  it was assumed that only small facilities
         will install a scrubber (large and catalyst extraction facilities
         are expected to install flares) .   The  6 small facilities which
         will require process vent control (see discussion in the model
         plants section)  are estimated to produce a total of 10.8 million
         gallons of wastewater annually.

         Solid and Hazardous Waste

              In general, there are  few solid or hazardous waste impacts
         associated with the implementation of  the proposed standards.
         There are no significant solid or hazardous wastes generated as a
         result of storage tank control by tank improvements,  or as a
         result of process vent control using a flare or scrubber.

              Solid waste from equipment replacement includes seals,
         packing,  rupture disks,  and other used equipment components,  such
         as pumps and valves.   Metal solid wastes such as mechanical
         seals,  rupture disks,  and valve parts  could be sold to metal
         recyclers.   Although additional monitoring of equipment may
         result in a greater rate of replacement of faulty equipment,  it
         may also reduce equipment failure.   These solid waste impacts are
         considered not to be significant.

              Solid and hazardous waste could be generated from the use of
         steam strippers  to control wastewater  emissions.  The possible
         sources include  organic compounds recovered in the steam stripper
         overheads condenser,  solids removed during feed pretreatment,  and
         wastes  generated in the control of system vent emissions.   System
         vent emissions,  if not sent to a combustion control device,  may
         be collected on  a sorbent medium that  requires either disposal or
         regeneration.   If the sorbent is  disposed,  it creates additional
         solid waste.

              Although waste generation can increase for any nonrecyclable
         organics that cannot  be used as supplemental fuel,  these organic
         waste would most likely have been removed otherwise from the

                                        17

-------
wastewater via the air (volatile  organics only) or via an
oil/water separator.   Similarly,  solids removed from the
wastewater in cases where pretreatment is necessary would have
likely been removed in a clarifier or activated sludge unit.


ENERGY IMPACTS

     The energy demands associated with the control technologies
for the proposed regulation include the need for additional
electricity,  natural gas,  and  fuel oil.  The storage tank and
equipment leak controls are not expected to require any
additional energy consumption.  The total .nationwide energy
demands that would result from implementing the process vent and
wastewater controls are presented in Table 9.  Overall, the
controls will require approximately 2.2 x 1011 Btu annually.
Attachment 3 provides more detailed information on the
calculation of the energy impacts.

     The electricity requirements for the pumps and blowers in
the scrubber systems were calculated using the procedure detailed
in the OCCM.  This procedure takes into account the pressure drop
throughout the system, and the pump requirements to recirculate
the water.  The steam stripper electricity requirements were
based on the horsepower requirements of the pumps used in the
system.

     The use of a flare generally results in an increased natural
gas usage for device start-up, supporting combustion of the vent
stream, or to promote flame stability if the heat content of the
vent stream is too low.  The fuel impacts are equal to the design
fuel requirements.  The fuel requirements are dependent on the
exhaust flow rate and heat content.

     Wastewater steam strippers require additional energy to
generate the steam.  As discussed in the secondary air pollution
impacts section, it was assumed that this energy was obtained
from the combustion of No. 2 fuel oil in a boiler.  The fuel
usages are based on the steam  stripper design, and the boiler
characteristics that were discussed previously.

COST IMPACTS

     The costs to the affected industry due to the application of
the requirements of- the proposed  standards include the costs of
any control equipment that must be purchased, along with the
costs of the installation of that control equipment.  There are
also costs for the operation of control equipment.  There may
also be costs associated with  certain work practices and other
programs that reduce HAP emissions.  Finally, there are costs
associated with the required reporting, recordkeeping, and
monitoring.  This section presents the methodologies used to

                               18

-------
                  TABLE 9. CALCULATION OF ENERGY REQUIREMENTS
                      ENERGY REQUIREMENTS FOR SCRUBBERS
                                  ELECTRICITY
Control Device/
Model Plant
Scrubber
Small
Electricity Cost
(SW
77.04
Electricity
Unit Cost
(S/kW-hr)
0.059
Conversion
Factor
(Btu/kW-nr)
3,415
Number of
Model Plants
6
Total Energy.
Requirement
(Btu/yr)
2.68E+07
         ENERGY REQUIREMENTS FOR WASTEWATER TREATMENT
                           ELECTRICITY
Control Device/
Model Plant
Steam Stripper
Large
Electricity
Requirement
(kW-hr/yr)
15.462
Conversion
Factor
(Btu/kW-hr)
3415
Number of
Model
Plants
6
Total Energy
Requirement
(Btu/yr)
3.17E+08
         ENERGY REQUIREMENTS FOR WASTEWATER TREATMENT
                             STEAM
  Control Device/
   Model Plant
  Energy
Requirement
  (Btu/yr)
 Boiler
{Efficiency
Number ot
  Model
  Plants
Total Energy
Requirement
  (Btu/yr)
Steam Stripper
  Large	
                     0.80
                   6
              2.56E+10
                       ENERGY REQUIREMENTS FOR FLARES
                                 NATURAL GAS
Natural Gas Natural Gas Total Energy
Control Device/ Natural Gas Cost Unit Cost Heat Content Number of Requirement
Model Plant ($/yr) ($/Mscf) (Btu/tt3) Model Plants (Btu/yr)
Rare
Large
CE-Oxides
CE-Oxides and
Other-HAP
1,889
1,889
1,921
3.03
3.03
3.03
1,000
1,000
1,000
11
9
6
6.86E+09
5.61 E+09
3.80E+09
                       ENERGY REQUIREMENTS FOR FLARES
                                    STEAM
Steam steam total Energy
Control Device/ Steam Cost Unit Cost Heat Content Boiler Number of Requirement
Model Plant ($/yr) ($/1000lb) (Btu/lb) Efficiency Model Plants (Btu/yr)
Rare
Large
CE-Oxides
CE-Oxides and
Other-HAP
18,166
18,166
33,317
4.65
4.65
4.65
1,206
1,206
1,206
0.80
0.80
0.80
11
9
6
6.48E+10
5.30E+10
6.48Ef10
                                            19

-------
 estimate  the cost impacts,  as well as a summary of the estimated
 costs.

     In general,  the methodologies used to estimate the capital
 and annual  control costs are the same as the methodologies used
 to estimate the costs of the HON.  These methodologies are
 described in detail in Volume IB of the HON Background
 Information Document5 (BID)  and are hot reproduced here.   The
 OCCM was  also used extensively in the cost estimates.   Parameters
 used in the costing analysis for process vents,  storage tanks,
 equipment leaks,  and wastewater treatment are shown in Tables  10,
 11, 12, and 13,  respectively.  Model plant costs are presented in
 Tables 14,  15,  and 16 for small, large, and catalyst extraction
 model plants,  respectively.  Nationwide costs are summarized in
 Table 17.

 Process Vents

     Cost estimates were developed for two different control
 devices,  depending on the size of the model plant.  A scrubber
'was used  as the basis for the cost estimate for small model
 plants, and a flare was used for large and catalyst extraction
 model plants.   As described in the secondary air pollution
 impacts section,  a single "model" exhaust stream was developed
 for costing the control devices.  A detailed presentation of the
 scrubber  cost analysis is presented in Attachment 4, and the
 flare cost analysis is presented in Attachment 5.

     While the HAP emissions from small and large model plants
 are assumed to be contained in a single exhaust stream (the
 "model" exhaust stream discussed above) , the catalyst extraction
 model plants have two exhaust streams.  The first is from the
 reactor vents and is identical to the small and large model
 plants.   The second is the exhaust from the catalyst extraction
 process.   This second exhaust stream is considered to be
 identical to the "model" exhaust stream; therefore, there are two
 "model" exhaust streams to be controlled at the catalyst
 extraction model plants.

     As discussed in the secondary air pollution impacts section,
 it was estimated that there are six catalyst extraction model
 plants that will require process vent control.  This same
 analysis  showed that 15 catalyst extraction model plants will
 require control of the catalyst extraction process  (referred to
 as nonepoxide HAP emissions) .  For the six facilities that will
 require process vent control, it was assumed that they will
 combine the process vent and nonepoxide HAP emission streams into
 a single  stream for control purposes.  Consequently, a larger
 flare was costed for these six facilities (i.e., the flare was
 sized to  handle twice the flow rate as the other flares) .  The
 remaining nine model plants that require only nonepoxide HAP


                                 20

-------
       TABLE 10.   PARAMETERS USED IN  THE  POLYETHER POLYOLS
                      COST ANALYSIS FOR PROCESS VENTS
Parameter
Base Year
     Scrubber
     Flare
Reporting Year
Chemical Engineering  Plant  Cost Index
     July 1989
     July 1995
     August  1996
Electricity  Cost  (scrubber)
Water Cost  (scrubber)
Steam Cost  (flare)
Natural Gas  Cost  (flare)
Operating Labor Rate
     Scrubber
     Flare
Maintenance  Labor Rate
     Scrubber
     Flare
Interest Rate
Raw Chemical Cost Used for  Recovery Credit
Equipment Life
     Scrubber
     Flare
    Value

   July 1989
   July 1995
 August 1996

    356.0
    381.4
    382.1
 $0.059/kW-hr
$0.22/1000  gal
$4.65/1000 Ib
 $3.03/1000 £t

   $13.20/hr
   $16.00/hr

   $14.50/hr
   $17.20/hr
      7%
     N/A

   10  years
   15  years
                             '  21

-------
      TABLE 11.  PARAMETERS USED IN THE POLYETHER  POLYOLS
                      COST  ANALYSIS  FOR STORAGE TANKS
Parameter                                            Value
Base Year                                          July 1989
Reporting Year                                    August 1996
Chemical Engineering Plant Cost Index
     July 1989                                       356.0
     August 1996                                     382.1
Electricity Cost                                     N/A
Water Cost                                           N/A
Steam Cost                                           N/A
Natural Gas Cost                                     N/A
Operating Labor Rate                                 N/A
Maintenance Labor Rate                               N/A
Interest Rate                                         7%
Raw Chemical Cost Used for Recovery Credit
(hexane)                                           $0.40/kg
Equipment Life	10 years
                              22

-------
       TABLE 12.  PARAMETERS USED IN THE  POLYETER POLYOLS
                     COST ANALYSIS FOR  EQUIPMENT LEAKS
 Parameter
 Base  Year
 Reporting Year
 Chemical Engineering Plant Cost  Index
      July 1989
      August  1996
 Electricity  Cost
 Water Cost
 Steam Cost
 Natural Gas  Cost
 Operating Labor Rate
 Maintenance  Labor Rate
 Interest Rate
 Raw Chemical Cost Used for Recovery Credit
 (propylene oxide)
 Equipment Life
      Pump Seals
      Rupture Disk
      Monitoring Instrument
	All Other Equipment	
   Value
 July 1989
August 1996

   356.0
   382.1
    N/A
    N/A
    N/A
    N/A
    N/A
   $22.50
     7%

  $1.41/kg

  2 years
  2 years
  6 years
  10  years
                               24

-------
       TABLE 13.
PARAMETERS USED IN THE POLYETER POLYOLS
     COST ANALYSIS FOR WASTEWATER
Parameter
Base Year
Reporting Year
Chemical Engineering  Plant Cost Index
     July 1989
     August 1996
Electricity Cost

Water Cost
Steam Cost
Natural Gas Cost
Operating Labor Rate
Maintenance Labor Rate
Interest Rate
Raw Chemical Cost Used for Recovery Credit
Eauicment Life
                                  Value
                                July 1989
                               August 1996

                                  356.0
                                  382.1
                              $0.059/kW-hr
                              $0.0580/1000
                                 liters
                                $7.68/Mg
                                  N/A
                                $13.20/hr
                                $14.50/hr
                                   7%
                                  N/A
                                15 years
                               24

-------
 en
 E-*
 cu
 W
 Q
 O
 2
CO
w
CO
                       03
                       JJ
r  ®  <°
°n  w
u  o  3
^S Jc
S^S
                       w
                                     ! T3 2
                                     i « ~
                                     1 OS
                              -S'
                                 ^ 4J
                                 <1» -H -
    T3

     (0
     CO

     O

    JJ
                        o,
                         0)
                        •o
                         o
                        s
                         o
                        u
                                   c  o
                                   0  U <
                                   2  0)
                                      M CO
                                     -H O
                                     •o o
                 JJ
                 U JJ
                 HI CO
                 M O
                                  rH «!
                                   10
                                   JJ •
                                  •H W 2
                                  C 0 2
                                  3 CJ ^

                                                      co

                                                      s
                                            s
                                                     00
                                                     in
                                                   VO
                                                   CN
                                                               s
                                 O

                                 CD
                                            01
                                                     U
                                                             «
                                                             aiw
                                                                       (N
                                                                       in
                                                   in

                                                   o
                                                                    a) ai
                                                                    EJ^
                                                                    ara
                                                                    -33
                                                                                                                              in
                                                                                                                              CN
       CO
       JJ

       10
                                                                        0)
                                                                       •a
                                                                        o
                                                                                            (0

                                                                                            CO
       CO

     •  10
    to  cu
   JJ  M
    CO JJ
    O  CO
    o
       w
  • i-<  cu
 co  
-------
CO
W
W
J
co
w
*c
o |jj _
^ rH
ro &
5-B
S-SS
0 0 3
rH= 0
10 m 5
JJ 5-ii
0 g1
^ io
•J

• a
2 w
c °
<°
**«
JJ
_. c
rH*S I
10 JJ K
•"•^ m
0 Q. X
H <0 g
0 C
M
e c
J-SM
So£
.2 3 0>
g-os
«§£-
Jo.S|
5 ,, -H O
l*5-£
il?&§
1-5|,S0

•w-
(thousand


XJ
0
rH
Q.
rH
0>
•o
O
2
M
CD
a

jj
to
o
o








0)
JJ
m
o
o
iH
(0
3
C
<












t-H
10
JJ
o
H
&l^
C D, u
Monitor!
Recordkee
and Repor 1


jj
0
0) JJ
>H en
-H 0
•D CJ
C
M


JJ
U JJ
0) m
i4 O
-< O
Q


jj
c
rH **
-.o i
10 JJ =
JJ-H tJ
^
0 C
^


JQ
AJ ^J ^H
•H tQ 2
COS
D 0 ^




C
o ,,
-H -y
oi 5
01 "o
-rt 0
E 0-
U


0 0 °
m . CM vo
O _- CTl VO
5 - co
rH " rH


0 O
vo Q IT a\
n« vo m *3
CM rH
CM CM


O CM TJI O\
in a\ *r
n vo r- CM

rH S S *
"> <• •* *
s * s *

in o vo m
rH ! rH m
M ^ rH n


in ^ vo ro
rH ' rH VO
H CD  b CKWE- CdJ 36-
































.
m
jj
m
O
y
• i-t
m a
U 3
a c
rH C
rH (0
O
T3 <»-<
0
vo
en jj
a\ C
rH 0>
U
JJ U
en 0)
3 O.
Ol
3 0
< m
1 jO







                                                                                                                      VO

                                                                                                                      CN

-------
en
w
W
J
ffl
<<
E-
1C
0
*u
0 £ en —
114 «%«>•
u »-•
w r| 10 "o
JJ % r* C
W M °* <°
o w w
CJ jj •-: 3
DO
tO ?T3 J=
jj *** O JJ
o ^ s —
£ "3
^ JJ
(0
o

1—1 r
II



J.
. C
^ f
r-l (0 c
10 JJ £
JJ-rt ~
o a^
E-> (0 lr
0 C
M
c c
Io ° —
•32 &
S ° *
.2 3 01
c T3 2
E At ^
M«
<1> M, 01 _,
S°.2?*







v>
"O
c
10
CO
o
X
JJ
^~
JJ
c
10
1— 1
a,

i— i
0)
•a
o
2
i_i
0)
Q.
jj
CO
Q
0








SS!-'5S
rt w 	 ' *H -U
•2 •§?&§
|ii|S0







CQ
JJ
01
o
0

r^
(0
3
<









1— 1
10
JJ
o
E-

.g^
c?-!-5
C Q4jj
•H Q) i.
M 0) o
Q _^ X
"* ^^ Qi
•"•o a,
•-* ^ CX
CO05
0 Orrj
s « C
os fl
JJ
o
0) JJ
i-l (0
•fH O
•a o
c
M


JJ
O JJ
cu in
M 0
3°

xj
•H S5
rH 10 g
10 jj 5
•elf
o z
w c
1— 1

««»
J3
JJ jj r-i
-H CO 2
C 0 2
a cj -^

^^



c
o .
.H "
w c
"2
E °*
U



CN O ° CO
oo •«» _, m
3 rt-












«H (N ° m
m o 0 o
•"* N S '
<*i [


^ ^ t-( Cft
• , *^ wi
. • o%
(XI r-l
^* CJ


«J «n in 2
m « •» "
r-l ' . '
r-l m «l ^
" (T> - ** f^



CN in 0 r-
VO r-4 " VO
CN M <~l r-l


en in ^ rr
f_^ ^^ * ^L j
CN CN •*> n




— •.
in m in m
CN
in o • ^
vo in o CN













«1
JJ
c
10
1— 1
a
>H
cu
•o
o
E
o
•H
JJ
u
<0
V4
JJ
X
.
1— 1
10
JJ
(0
y

JJ
10
CO
• (0
CO 01
JJ u
10 JJ
O CO
0
^ ^ . . rH OJ
0) Ol r-l " CO 10 JJ
1-1 3 (0





_. ^K ^* . -.
•-4 a o vo
r-l r^ Q 0





<0 C 3
r-l C 0)
f™^ id xJ
O ' CO
T3 -t O
w a> u
3 ao
01
0) 3 0 O
•ncu TJ •" W-JJJ
(0 -H X >i O C JJ C « « "
V) X OT O r-l MO) 0) 03 0)11
to OO 0)  jj II
ujj-^ai— 10 ujjc (0 a) M x. •* *z u 10
OCOCO.--I OCOCUr-i XOC 310 100)
U 0) Cd O •£ Ur 'H  — ex a, > — x tiMEi uj 2H

-------
On
<

cn
Cd
o
><
J
o
OS
Cd

EH
O
04

o
o
o
M
E-

z
Ed
J
CQ
2-
«J
O ^j
E1* „_



\ Annual
•H
10
JJ
•f-l
a
CO
o
'm ^ >•
|||

•wv
•0
ousan
J3
Plant

^H
0)
•a
o
s
>,
CD

ca
jj
CQ
O
O
0)
T3
-H
3
O
-H
JJ
(0
z






C
JJ O
CQ -H
r-l U

10









i— i
F-H
ca
VI





Annual


r-l
CO
JJ .
•i-l
a
CO
u
r-l
TO
3
C
*•


i^
CO
jj
•H
a
10
o


_4
Annua!
i— i
(0
jj
•H
a
(0
u



c*
.2 -y
J*o

^
lu


o
o
en
CN

O
CN
CO


S
O
CN
in
r-l


in
m

o
m
o
^




vo
fl1
CN





m
in
m

00
CN
CN


to
JJ
c
01
>
CQ
CQ
0)
0
O
M
0<
o
CO
00
o
CN
n
fl"
CN
r-l
O
CN
CN
O
O
CO
o
'
fl"

o
vo

0
CN
CN

O
0
CO
co
^
C
IM a
0 EH
O
•H fl)
I Dl
•O fl
0) ^
X O
•H JJ
[L, CO
0
CN
CO
m
0
CN
VO
0
o
r-l
O
CO
m
r-l
0
pn
0
m
o
CN
QT\
^
O
fl1
n
CN
CN
CN
in

0
in
0
r-l




JJ
C
01
£
a M
-*-i ^
3 (8
cr a)
H J
vo
vo
CO
o
en
CN
CD
CN






vo
vo
CO
o
on
fl"
CN








1-1
0) JJ
JJ C
(0 
CN
0
vo
fl"
n
o
vo
CO
m
0
fl"
CN
in
r-
CO

o
VO
n








r-j
10
JJ
o
EH




















CQ
10
O
T3
vo
a\
^H
jj
ca
3
Ol
3
rfj

*





                                                                                                                         CO
                                                                                                                         OJ

-------
 control will use a smaller  flare sized -for a  single exhaust
 stream.

      The THF facility  is characterized by the small model plant
 category.   Therefore,  process vent control costs  for the THF
 facility are estimated to be equal to that for a  single small
 model plant.  From Table 14, the total capital investment for the
 THF  facility is $38,000, and the total annual costs (including
 direct; indirect;  and  monitoring, recordkeeping,  and reporting
 costs)  are $58,900.

 Storage Tanks

      The storage tank  cost estimates directly followed the
 methodology presented  in the HON BID.  The primary  input
 parameters for this methodology were tank dimensions (height and
 diameter)  and product  recovery due to implementation of controls.
 Tank dimensions were calculated based on the  capacity and
 assuming that the  diameter is twice the height.   The product
 recovery is equal  to the emission reduction,   which  was calculated
 using the  EPA's TANKS3 program.  A detailed analysis of the
 storage tank cost  analysis is presented in Appendix 6.

 Equipment  Leaks

      The equipment leaks cost estimates were  calculated using a
 spreadsheet developed by EC/R.   The spreadsheet was based on the
 HON  methodologies  and incorporates credit for  recovery of
 product.   The procedure used to determine the  number of
 uncontrolled components to be monitored is discussed in the model
 plants  section'.  Attachment 7 presents the cost calculation
 methodology in more detail.

      The equipment leak control costs associated  with the THF
 facility is represented by the costs estimated for  a small model
 plant.   From Table 14,  the total capital investment for the THF
 facility is $52,700, and the total annual costs (including
 direct;  indirect;  and monitoring,  recordkeeping,   and reporting
 costs)  are $26,100.

Wastewater

      Steam stripper cost were calculated using the  HON wastewater
 spreadsheet,  which is presented in Attachment  8.

Monitoring,  Recordkeeping,  and Reporting

      In addition to the control costs,  the facilities will incur
monitoring,  recordkeeping,  and reporting (MRR) costs.   Due to the
 similarity of the MRR requirements in the proposed  standards  with
 those for  elastomer production facilities,  the same methodology


                               29

-------
used in the Polymers and Resins I NESHAP will be used to  estimate
the MRR costs for these standards as well.

     The Polymers and Resins I NESHAP methodology used
information from the HON SF-83 analysis to develop  a  ratio of
annual MRR costs to the total annual control costs  (sum of the
direct and indirect control costs).  This analysis6 showed that
the MRR costs are approximately 30 percent of the total annual
control costs.

     To estimate the MRR cost on a model plant basis,  the total
annual costs were multiplied by 0.30.  For example, the process
vent annual control cost for a small model plant was  estimated to
be $45,300.  The MRR costs would then be $13,600, and the total
annual cost would be $58,900.  On a model plant basis,  this
analysis probably overstates the MRR costs for those  facilities
where control is already required, and understates  the MMR costs
for facilities  that are essentially uncontrolled.   However, this
analysis does provide a reasonable nationwide estimate for the
entire Polyether Polyols project.

Estimated Cost  of Compliance

     As previously noted.  Table 17 summarizes the total capital
investment and  total annual costs by model plant and  nationwide
totals.
                                30

-------
REFERENCES

1.   Chinn, H.  CEH Marketing Research Report - Polyether Polyols
for Urethanes.  Chemical Economics  Handbook - SRI International.
1995.

2.   Memorandum from Seaman,  J.,  EC/R Inc.,  to Svendsgaard, D. ,
United States Environmental  Protection Agency.   Summary of
Nationwide Baseline Emissions for Polyether Polyols  Production
Facilities.  April  29,  1997.

3.   Compilation of Air Pollutant Emission Factors.  Volume I:
Stationary Point and Area Sources.   AP-42  Fifth Edition.  United
States Environmental Protection Agency,  Research Triangle Park,
North Carolina.

4.   OAQPS Control  Cost Manual.   EPA-450/3-90-006.   United States
Environmental Protection Agency,  Research  Triangle Park, North
Carolina.  January  1990.

5.   Hazardous Air  Pollutant  Emissions from Process  Units in the
Synthetic Organic Manufacturing Industry - Background Information
for Proposed Standards.   Volume IB:   Control Technologies.  EPA-
453/D-92-016b.  United  States Environmental Protection Agency,
Research Triangle Park,  North Carolina.  November 1992.

6.   Memorandum from Norwood,  P., EC/R Inc.,  to Evans, L. , United
States Environmental Protection Agency.  Estimated Monitoring,
Reporting,  and Recordkeeping  Costs  for Polymer and Resins I.
August 9, 1994.
                                31

-------
           ATTACHMENT 1




CALCULATION OF EMISSION REDUCTIONS

-------
               Calculation Methodology Description

                        Emission Reduction

Process Vents

     The EPA database was  used to determine the number of
facilities already achieving the level of control specified .in
the proposed regulation  (98 percent control) .   Of the 28
facilities in the database,  19 facilities (approximately 70
percent) were achieving  98 percent control or greater.  It was
assumed that the percentage of controlled facilities would be the
same for all model plant categories.

     The number of uncontrolled facilities for each model plant
type was then calculated.   For large and solvent extraction model
plants, the total estimated number of major source facilities
(76) was multiplied by the fraction represented by each model
plant category to calculate the number of model plants.   This
number was then multiplied by the percent of facilities
controlled at a level below the standard (30 percent).  The
following example is presented for solvent extraction model
plants:

   (76 major facilities)x(0.25 fraction sol. ext. )x(l-0.70 fraction controlled below standard)
                             = 6 facilities

     Similar calculations  were performed to determine the number
of large model plants (11)  controlled at a level below the
proposed standard.

     For small model plants,  the total estimated number of major
source facilities was increased by one (for a total of 77) to
account for the HAP-emitting facility that produces polyether
polyols by polymerizing  tetrahydrofuran (THF).   The same
calculation as shown above was  then used to determine the number
of small model plants (6)  controlled at a level below the
proposed standard.

     Next,  the existing  level  of control was evaluated for those
facilities in the database at  less than 98  percent control.  Four
categories were developed  based on the existing level of control:
95-98 percent control,  90-95 percent  control,  80-90 percent
control, and no control.   Of  those facilities  at less than 98
percent control,  12.5 percent  were in the 95-98 percent control
category,  37.5 percent in  the  90-95 percent capacity, 25.0  .
percent in the 80-90 percent category,  and 25.0 percent in the no
control category.

     Based on data reported in the Section 114  questionnaires,
uncontrolled process vent  emission rates were  estimated for each
model plant category.  These emission rates were 24.3 Mg/yr for
small model plants,  97.3 Mg/yr  for large model  plants,  and 48.6
Mg/yr for catalyst extraction  model plants.   These uncontrolled
emission rates,  along with the  number of facilities at less than

-------
98 percent control were used  to  estimate  the  emission  reductions.
For  the four categories of existing  level of  control listed
above,  95 percent, 92 percent, 85 percent,  and 0  percent control
were used in the emission reduction  calculations.   The following
example is presented for small model plants:

       (24.3 Mg/yr)x(0.98-0.95 percent additional control)x(l facility)
     + (24.3 Mg/yr)x(0.98-0.92 percent  additional controljx (3 facilities)
      +  (24.3 Mg/yr)x(0.98-0.85 percent additional control)x(l facility)
       -t- (24.3 Mg/yr)x(0.98-0 percent  additional control)x(l facility)
                             =32.2 Mg/yr

      Similar calculations were performed  to determine  the
emission reductions for large and catalyst  extraction  model
plants  (350.2 Mg/yr and 64.2  Mg/yr,  respectively).

      For the THF facility represented  in  the  small  model plant
category,  the emission reduction is  estimated to  be the average
emission reduction per small  model plant:

             (32.2 Mg/yr)/(6 model plants)  = 5.4 Mg/yr

Storage Tanks

      As described in the model plant section  of the memorandum,
an analysis of the EPA database  showed that there was  typically
one  uncontrolled fixed-roof storage  tank  at each  model plant.  It
was  also determined that the  average capacity of  the uncontrolled
tanks was 38,667 gallons and  the average  annual throughput was
134,500 gallons.

      The EPA's TANKS3 program was used to calculate annual
emissions for the tank first  without any  control, then with an
internal floating roof.  Hexane  was  used  as the stored liquid.
The  difference between these  two annual emission  rates (0.62
Mg/yr)  was used as the estimated emission reduction per tank.

      The number of model plants  with tanks  requiring control was
determined using the EPA database.   An analysis showed that 3 out
of 12 facilities  (25 percent) have uncontrolled fixed-roof tanks.
This percentage was multiplied by the  total number  of  each type
model plant to determine the  number  with  tanks requiring control.

      The emission reduction per  tank was  then multiplied by the
number  of facilities with tanks  requiring control to obtain the
total emission reduction.  The following  example  is presented for
small model plants:

  (20 model plants)x(0.25 fraction uncontrolled)x(0.62 Mg/yr-model plant) = 3.1 Mg/yr

      Similar calculations were performed  to determine  the
emission reductions for large and catalyst  extraction  model
plants  (6.2 Mg/yr and 3.1 Mg/yr, respectively).

-------
I
         Wastewater

              Two  facilities  in the  EPA database reported Group 1
         wastewater streams.   Uncontrolled emissions for each facility
         were calculated by multiplying the average VOC concentration in
         the wastewater  stream by the annual flowrate.  The fraction
         removed  (FJ for each HAP (HON,  Table 9)  was applied  to  the
         uncontrolled emission rates to determine the emission reduction
         for each  facility.   These two values were averaged, and a value
         of 41.5 Mg/yr was  used as the emission reduction per facility.

              Both of the facilities reporting Group 1 wastewater streams
         are large facilities;  therefore, it was assumed that only large
         model plants have  Group 1 wastewater streams.  To determine the
         number of large model  plants with wastewater streams requiring
         control,  the two facilities with Group 1 wastewater streams were
         divided by 12,  the total  number of large facilities in the EPA
         database.   Consequently,  17 percent of the large model plants, or
         6 facilities, were estimated to have Group 1 wastewater streams
         -requiring control.

              The  total  emission reduction was calculated by multiplying
         the emission reduction by the number of facilities:

                     (41.5 Mg/yr) x (6 facilities)  = 249.0 Mg/yr

-------
Wastewater

     Two facilities in the EPA database reported Group  1
wastewater streams.  Uncontrolled emissions for each  facility
were calculated by multiplying the average VOC concentration in
the wastewater stream by the annual flowrate.  The  fraction
removed (Fr)  for each HAP  {HON, Table  9)  was applied to the
uncontrolled emission rates to determine the emission reduction
for each facility.  These two values were averaged, and a value
of 41.5 Mg/yr was used as the emission reduction per  facility.

     Both of the facilities reporting Group 1 wastewater streams
are large facilities; therefore,  it was assumed that  only large
model plants have Group 1 wastewater streams.  To determine the
number of large model plants with wastewater streams  requiring
control, the two facilities with Group 1 wastewater streams were
divided by 12, the total number of large facilities in  the EPA
database.   Consequently, 17 percent of the large model  plants,  or
6 facilities, were estimated to have Group 1 wastewater streams
-requiring control.

     The total emission reduction was calculated by multiplying
the emission reduction by the number of facilities:

           (41.5 Mg/yr) x (6 facilities) = 249.0 Mg/yr

-------
en
vo
o w
vo <
0 O.
CQ
CJ
H
CQ
H
K
H
S H
3
O rt|


^J
O »4
• H J

H CJ
2 Q CQ
0 i 13
O 0,
K £«
wll

u a
J2 Ql 55"
3 0
H CQ H
55H
0 •<
H O
CQ H
CQ fa
H H
2 H
CM 25
H
Q
H
s
2
fn




































































8
OS

•a a) ai

x . . . . y> . o -.* ~« r> o 01 oo

Cb ^fM^CO^ SZ 000 00

-i ^ -» 0) 4) 0
u* c  —
-^ « « *o -Z
•« >» "o ce -» -*
-* ^. o en CB
^ as a> a •*« O.
••-^.*j flo- ccn«*-
o -^ u jc •• cJi u o> -• a aoca.
*-i *- -^ 01 UTJC-O- -^ O — ' *J C
c «.. DC u cojco*ac"u a *JCTI~*
O - ^iJSO 3 ~« en *• JC O « — — 4)C*J
-« j£ oijjujc — » a. u "--. ^j en -• ** • u as — * *j


-*U H ^ -• 41 3 m to 3 O — C O "O JJ -*-i •*-• iJ C 4J
nj«*j >, MJ CSO) -H 0) u u CJ O O O flJ " > w
U--* -CO 4) U'OiJ4l>jS fl CJCJ ti »J« £3
-i u -- 41  i-i 3 O
*J C X *J D. 4J -t -^ E 3 • 3 C CJ'-*— tiM«-i £ 4) CT> -« Q. 4)30
^» QJ iJ fl E Qi QtlfljtTOl*^U*J 4)4>OO Up.«-*'OO JCUCJ
C»-«OCOUH jtfc/)Q-J<>f-2 ccowosce *- H — cc co  a.
4) C -0 W
T3 2 C5 ° ±









— .
T-t
at
Q*

^
•
u
u
3
a
a
ti
a.
u
4>
g
<
01
^






a
X

5-

-------
r-
o\
*»«. es
\o
o ca
•V.C3
\o <
O 04
                                  O  u
                                  •u  fl
                                  m  o
                                  O  01
                                  Z  2
                                              a
                                              O
                               3  n  u
                               a* 01  10
                               n  nj  u
                               J  Z  6u
                                  a  X
                                  a  a
                                 —  s
                                  w
                                  o   •
                                  a.  01
                           3 jc  a.  oi
                           o- «  E  v
                           •»  3  « -a
                           J CD t-i —
                              «"  O»  X
                              V*  O  (0
                              3 -O  Z
                              w —
                               >.  01
                              «  a.   •
                              ~>  e  en
                               
                              Q t-  <

-------
o

c
o
01 0
pi, CO
• CO
ov us
P* (N











*J
«u

U


~ 4)
.0 E
— ~*
O
n >
4)
m it
01 U
2 0.
cn
i1,,
3 8.
C 1)
•8 >
in vo
n oi
• O\
O 0










U
o
.. (J
« . u
Jj
-. Of
c S
4* CL
Q CO

O O
Q. D.
> >
M
U)
in
in









u
0
u
U
h.

c
Q

iJ


T3
C
>


















4)

3

g

4)
U
18
a
CO
w
o
a
>
c
o

CD
to
n











u

2
U



e
^,
o


41
U
a.
ca

o
a
>
«>
m




















*••


h*
«

4)
E
<8

O
X
C
•8
m 0 (TV tn
• t~t m













^

^ .- —
"*-t ^ JJ

— Ol
41 -*

u cn -t
3 -« T3 *-
0 4) - —
X 3

O — * •** Cl
a.  t- < cc
in co o co
• P* (N •-*
O O
O














_^

O
o ~ -•
a: •- •- *j --
4) *J *J *"- *••
C •*•* *•! U ~—
O 	 «w
U — ^
4J
0)
c
41
Q
u
O
a
Ul O
n o
• f-
o •-«

O)

..







^*.
J3

u

3 Ol
(j •-.
^^ 41


*8
>» -•
u 3
-~ 0
n o
4> O
a z:

o o
a a
> •>





•a
3


J

0)
Ol
w
41
<

>>


Q

w

41
W
3
m
cn
4)
id
0.
w
o
a
TJ
>
CD
V










^
tj

a



3

*8

a
E
4J
H

a>
u
«

ki
3
to

0

m


K

Ol

^


81
4)

0)
o
<8

3
to

^
3
O*

J


Ol
>
<
X
•"*
i
p*

in



j—fc
o:


Ol
4t



1

H
^
c ce
V
«4 *J
"i c
< u
09
U C
Ol O
<8 U
w
0) n
> TJ
< O
>« ^H
-H 
O) V
0) *—
13

-. 3
O j-»
£  M
J3 4)
*- £

H


3 ~+
U 3
0

•** ^
a —
a 3
— o*
j
i* r-





r-+ ••
^^ -— •

01 O
£ 0
co ce

4) 0)
o o
c c
IQ fl)

a a
O Q
a a





o o


u u
c c

a. a.
c c














c
0
"2

^
o
c


Vi

o
cn


*8
Q
H
>

I

O
in
m
















>*
"TJ


O*
CO


*J
a)

u
o
u
Lu















0

u
b.

C
o

c

a
x


4)

a
w
o
a.
«
n
o\
o









„

O
u

u.
c
o

o
c
Q.
X
fd



in T
. to
01 r*
u»



o7

o» ••

2 "*

n
4) Q*
O» —
C
B£ Ol
C
01 *8
3


U 3

E 4>

t* ^*


O O
cu &

>. >>
2 ^
a a
Q



"«
-*H
CL

01
01
c

a
o>
c
JJ

4)


CO
CO
4)
a.

ij


>
M

5
4)
03
O
£
r-


t3
3




0)
Ol -
hi *8
41 •*»
< a

>•
••^ u



4J 1-1
a.

±4 QJ


m 41
W rq

U 3
o to
CL
>





T3
a
o*

,j

E
I

C
2

>•
.^

a

*-•


u

(A
a
w

u
O
a.
5
^
0
m









^
4

a


^

4J
<8

a

41


41
•fl

^
to






•a
3
a*

j


I

X
z

>,
.„

a

*-»


u

a
01
i^

w
a
5
CO
Q
r*
»^
eg









Z
«8

a


^4

Jj

**


4)


01
Jo

3
to

0- OV

01 r*
in in


K as
01 o»
4> 4)
•o -o



aa
4) U
H H
Ol 4)
O U
ij . >,
•^ ^
Q Q
•-t 0

m r<
m


tf
Ol
4)
•O -

flC

t?
41 -o
H —
4) 4)
O Ol
<8 C
i* a:

cn

T3 E


•-4 -J


X O
J5 I

>. >-
^ *n
a 2

-------
f-
ov
»>^
\0
o
^v.
\o
0





o
9
m

JE
3
0
O
K
CU

05
^
§
H









































*»«••

ca
o
<
cu
H
K cs
O ^
&4 1
ft
J •<
H "-
•<
H CO
H S
a o
i H

H t?
« 5
O U
0. J
H ««;
K 0
CO J
2 H
0 <
H H
ca I*]
CQ a
&
H



































































*n eo in


m in *A
in r*
.
O N
U
3
cr
j


u m ••

U 4J ^4 10
o u n —
a u m ? a>
c u u. A a
o « — -
— a. c >. ~
JJ O 1-1 «» •"
-« O u * 3 —
3 — ro O a
U JJ K >0 «l
-< IB 3 *J k< Ol
10 w jj 10 ft) 10
a 3 10 a u
JJ 0) H E 3
C 1 W A) O
O (A u 3 H
~* O O CU
0) u Q. 09 0) U
-> a > w  TJ "
0 W 3 K
— -a jj o (o o
(0 CJ C G> O<
z jj aj fl  > >
~ ZJ
< >


































VO O 01
en o -*
CD O «
o o in
- r- r*
a>
VO \O CM
GO
^
•• 3
— O1
«-« i3
o
E 4)
• a> -
J3 * —
« k4 «J
\ 0) -«
A > 0
^ < 3-


4J <-4 C
23 2
-ai a
— sun
U n u
— « h< a>
— 









^




s


3
a

1
o

H

^
ID
3
C
C
<




































O
o
o
o




















JJ
0
u
IB
h.
at
1
V4
3
H




































^
n
r-













jj


U


1

s
13
3
(T
E
3
E
X
(0
£




































CO

o















AJ
IM


J=
0>

z
•o
3
cr
E
1
X
11
X




































\D O ^
. o CD

M fl 
0 = 3
JE jt:
C U -H
Si 2
o
£

-------
r-
o\
—in
vo
o ca
  .o
vo <
O 04
                   ~ 2
                   5 u
                   — in
               o
               a.
               u
                     8

-------
J\
*x.H
tO
o W
*«^ ri
£S
O Oi
CO
u
EH
CO
H
fy*1
H
H
SB
«•• tj
O ^
U
0 J
' y 'f!
ro < rt!
H U
S H H
« Q CQ
C3 t S
o cu
n> C-i
Pi C*
Di BJ Q
o 5
CO Oi «4
rrf W
5 os s
•4 O
EH CO H
tr C-i
»<•« CH
0 «*•
H 0
CO H
CO b
H H
W §
H
Q
H

.

5

H




























































§
C£
£

"H «
« «
O *•"
« V 4t C
-j« [^ ttO r* 41 .u  ^J 3
2 - \fl - o ^ _ O
^ f*i««^ 3JC£ £
£1. 2 flj ntD K 7 2 ' *•»
MO C ^ "•"i*'1^ flT3<0
2H ^ UCICU •* U U
rfOl W ^*JT3iJ*O 3*O -*
X3
= §HW-5 JZOZO JZI l^


C >•
0 ^
^0
o »* fcj en
o u v «. .. n u
2 ,, .. -o -a u UJJ
« 4JC-9C-D-- - — <8
C •• O B) C — 0 TJ C ^^ JJCJ
O •- ^-C -«*-*tV)CJ£0 « -IB 03
.^ ^ ODjJO i*4J*s.W-- E-»« rtC>
C .J C C >*--* -^ a -« U JJ '-^ *J OJ4CA lj C
^H o H --4 15 O UC<-'-• u P- O *J
mu >*•*-• CO— OJ WUUO.OO W>. »0-—
u-« -co a>jj «! uu u*OH uu,
.« 4j .. o «o EOJEO i:^-i*-« ^«t»c ID
— . OJ jj «0 £ D. Q fl ^ w UVVOQ 4» -« U U U O
ij-n.-^iJO> -^03 u^^^roo cn^C4* o
c •-• U w tj H Ji Q > t- ccQcncnaca: - o. en Q jc Q
i> C -^ £ u
— ? >n T3 -»4 U
H H a a: c





















>•
u
•«* i-t ^M ^»
4J
C
<0





Ji
a
!
•a u
V T3 41
JJ 01 >
0) AJ O
Jtf <0 O
a j£
« CO O>
en « c
c o~«
= §2
n" - 3
u u i
> it •
o > -.
u o o
u u
•0
a •a a.
*J 0) D
*-4 U 1
O -. 4J
^ O r^
£ 4 ~*
3 C 3
^^ 3 ffi
— *•* ^
e « ~
ffl 01 6
•H Z IB
Q •»*
u a
o C O •
3 -x « C
JJ • b. ^*
tfl «•
u n 0) v
CO — Ol M
^» 3 *^
OI J£ «
U (QUO)
U = -I I
••4 U
b. 01 «j e
01 E E
J£ D O 3
U u u rH
OJ U 3 O
a rf < o






















M 0V r4 »H








c ji
0) a
a. n
o o
0 C
-4 O
-*4
T] *1 4J
u n n
u ft) 3
ft) CO A*
.* 0
CD O <
10 -1
OI K •
C J3 £
3 a U
b. ft)
z
U 4J
«> -« T3
> •-! ft)
O 0) u
Cl -^ £
ft) — O)
cn *^ • -^
c .a E ft)
-• n n x
•O JJ -H -^
•MOO —
r« 2
<0 1-1 • E
nS-5 JS
• ~ . Q
6 rl »
 C
r^ ffl 14 k4
"^ ZH O 4)
J£
« 0 Q. S
01 -< W
I o> a. m
4)
^ J « E
V -. Z!
•a »w D. 3
•a o E u








«B
•^
CD
a
r-
^
•
0)
3
m
GO
w
b
&

u
^4
b
ft)
JC
a.
m
g
e
jj
<
?
>
^






CO
(Q
X
£
Houston, '


a
c
o
ft
u
(0
3
U
1-4
3
C
O
CD
CD
-*
E
U

C
**
•o
0)

o
Q
W
0
V
Jj
u
£

-------
to
o:
o



















o
•
3
IK
P£
o
o
«
Pi


ji
jjj










































H
Pi













EH
M
33 ^*^
f* (N
O-*
Pi
^•J rt
H *-
EH 01
H 8
Q O
H
i EH

H §

O U
Izl <
K U
09 J
2 H
0 <
H EH
CO H
CO Q
H













































































o o o **
01 O O to
at o o *
CD to rt O


^







c


.c ••
o. c
E O

Li U U
>. >. c
n > 3
C 4J JJ b.
0 •-, iu
•-« ^ ^ 01
u 01 01 Li
10 -• « 3
-H o o a
3 •• E E n
0 — . . 01
— a a a LI
,0 — ~i -4 a.
u 	
= a < m o
o 01 a
•-4 n u LI in
n a o 0 >
n o *J u
—i _3 U o «*J
E u 10 0
bl ~i U, b.
nj 01
-. 01 -4 -> 3
>o to 10 >o — '
= S W M >

< Bi










































CD \O O O
n -00
» r* O O
in n o o
r- r-

0







3
O1 —
-I 01
« i
Ol ••
« — .Q
M 1C **

> » i



fl 4t *J

fl 3 m
a U •• «*
(0 — 4)
ij u jj 3=
«T» W »-»
a — n ••
4) E fl h
W 4) U ~« O
3 H 41 3 JJ
m u u u
« 4> 4> 41 10
4) U £ -• £••
M <0 « O
a iu -^ z t>
U Q U
M 3 U 3
0 M * O "O
a c a o
 H > 0-











































o
(T» O O a? tn O IP r^ •-* O O O O O OOU>OO
o o o vo -H r^ • mtDooo o oo-oo
0} O O •«> 0 (N «*i CMVOOm O O O *-> O O
O O O rN O Ul , -^ Ui-"' Ol ^o *-N
*--»-. m C-^AJ c*j"^
-. C 3J3U at •• « J3

4) Ol Q — U. "u ^
w — -41 JJ >» W
4) --O --WU (0 - -* . O 4J
Euu-»< J33£ 0) ^>C^J*J £
-~- Q3O3 ^^ 05 Ol Q *• jJ C3 *M O Ol
J3 -*O,AJCT" —-CD-* .J -*»M -^ 10 •• **4
»-i » O £ U -M — > 4» 4) J3 — 1-. 4) U. — 4)
— OS OliQ^jJ 0)^2 Ol -^ 4)<-i UX
CC3h. *w 4)0. C *- J= D. O J= ^
CBEzEO O— W I- - -* U £ iJ — U •-
« 3 3 u tt -^ O u nj u .u en O m • CT>  at  a» m
^4»4> — oe C -it...* JEEUEE-.U.
•— « O > 2 ^^ fQ -»^ »j 0 U <0 D U 3
U^4)3^fc,jtf &.3O13 • UJJCJCtt.jejfQ'a
•OE-WCOJOIC ^aOJJ UU OCQ.Q
£ 3 »" C — > fl Ji ifl TJ l-i O J^UU 4l(QigU
*J Z W < CD < £- u > > Q. f- OQQ Q H > O.
•^ QJ K
Z Q Q


JD

w
CD
m
j

E

'c
1
V
o
E
m •-«
0 ^
jj .a
Q *^
u.
o
n <~
3^
en 01
C ~-i
i i
22
4) U,
Q ^
*n














•o
41
JJ


m
Ol
c
3


41
>
o
o

T5
41

O
£t
C
3



£
«0

a

c

o*
cess Match
u
<




«

vO
in
in

o
0
o





o
0
o






o
o
^,
1-1















TJ
4)

0)

rtj
I


k4
4)
>
O
U

T3
41
I
C




4)
2


|Q
o
f+
u.
4)
tomatic Gau<
3
<




5
o
r-
w>
a

o
o
o





o
0
o






o
o
r-













0)
1
J
41
>
O
CJ

Ol
c
•o

CO



o
o

a.

-^
3
03



E
<0

a

c
CN
01
s
s




in
o
tn
,^
o

o
o
o





0
o
o






o
o
v>
r-


















4)
jj
4)
B
«
?
s



01
>
o
CJ

'!
•o


CO



E
(0

Q
c
%O
r^
4>
2
4)
T>
T3
m




o

v
^
00

o
o
o





o
o
o






o
f-
































4)
n
01
3

t3
<
"S.


4)
2
U
u
c
m
o
en
4)
j
o
o




S> Si
u> n
u> vo
p^ *•
«• n

0 0
0 0
0 0





o o
o o
0 0






0 0
O d
r* \o














C JC
II
0 C
« o


n >o
01 3
U
u <
a s:
10 O
b. 01
z

— TJ
** 01
0) u

~ Ol
E D




-------
or-
V.CT
0 &
vo <
0 ft













M
?3

« r>j
O •*»*
b i
O.

• H *—
CO (<
EH 05
S n o
wj M VJ
K H
O I H
O 4
K H t|
A « 5
0 0
CO 0t J

2 PS U
H CO J
»H
O <
H H
CO H
CO Q
H
H












































































0 0 0 -«
CM O O *O
a> u> m O

»








C


I §

•— > u
Li Lt O
>• >• C
(D > 3

•-H "x. ^ Ot
S -2^3
« O O tn
3 •• E E a
U _ i . ft)
ft] ^- ^^ i-t Ot
o — « —
c n < CD o
o u a
•< JJ Li Li 
« J U U «M
E TJ TJ O
u ft a. a.
3 U 01 TJ
C £ 01 W >
< fX


























































•a

3
cr

J
0)

TJ

4)
>


•"•
a
Vapor Pressure at
















































(O VO O O
«-t - O O
tn CN O O
r- r-
 O
JJ -ta-
01 •- -0
E JJ Lt -^
— * *n 3 o 3
*-• •• O J= CJ •? «
~ ffl tJJ TJ J JJ
c s: 3 a. —
» £ E O U —
0) 3 3 Li 4) •*«
0) « « £ 31 C Li
CO O O t* TJ T) 4)
30 o cn cn jj
*J C Li 4)
- o S S - ° 1
fO -« CJ 0) -t
3 U ,_, _ OT Q
TJ W U *9 ,-• (TJ
W. .Q 0> = ~, Li .*
*O £ «-» c ai 4) c
£ = •" C J= > T>
*J 2 iJ < CO < t-
z














































o
r>t *4 O O O OO O O VO O O
in VO O O O O O O • O O
(N 0 O O • O O O (N O O
. p> - p* . . r-
* • CN
t-t O3 OD






4)


01 E ..«
c "o xi 3 o
o E ^ *- e
•.* . ^- £ cr -
jj J3 JJ a J3
u ^ cn •*»* «-•
C \ JJ C JJ ***
3 X) U 4) - «w J3
— £ "w *"" —
-0) JJ >. Ll
--^.ujj eo •• •»* • o *J
J3 3 -C CD --CJJU £
*- 0) -« J — . *4 -^ IQ .. -*
41 4) .a— 'LiOfb*.*— 4)
(DLiZ O1 ^ 1J — U2
aicu c — ' j^ o* o j; *M
n LI •• — . jj £ jj — • u ••
'us LI ns w jj maia>0) O'-«0)4)3JJ
TJUUU. »JJ — JuOO
01 > ai aj O ua)(Q
c -. u. j«; ^EEUEE^-O.
-««i-iO U 'O'flOnjioO
JJ O Z *J flJ E4)4)JJO)— »E-U
jj o Q -fl w 05 o cn a u
-• 4> Li 2  > D. H U Q Q QH=>0.
U 4>
Q a




J3

"—
CD
U
a
CO
3
E
C
1
Lt
>,
-*^
V
A
at •-*

n K
b.

a
a .—
3 >.
's^
C i"i
-^ o
Jj t
£5
j<
Si


N
2
u
C
(0
3
a


































a
3
jj
n
jj
01
-*v
cn
c
u
jj
tl4
Ji
U
0)
a





r-
r-
o
o
fi
^



o
o




o
o
0









Q
o
VO
f^




^










•o

Jj
Jjj
CO
1

„
Lt
41
>
o
u

•o
41

O

3

„.
£


a
c
rw
£
U
tJ
1)
X
CO
CO
u
u





~
1-4
vO
in
m



o
o




o
o
0









o
o
**




^












u
0)

cn
c
3

W
41
>
O
o

•a
41
•3

C

•^
^
0)
2

0
Automatic Gauge Fl





in M ™ ft
v ui o  r- «-«




*4 r* O\ r*







jd
CO
•a c
cn 0)
c a
3 O
•»
Ll O
4> «-«
>
S*O ^
4) t)
JJ D
C S.
— m u
•a  ^.
O CO
a u ^
3 0) — -
25% E

3 -^ to Q
ffl i-i 3
^. 03 1-1 -
~- ^ -o c
E ~T -x. ?
tt E ^ -v
Q •« '3 —
a 2
C C L. ^
*^ •« 4) U
v vO C
N ri 
-. -• o a.
0) ft) -i
3 z cn a.
at
C Lt _1 4)
3 t3 -M "Q.
^ t3 O E
O tJ O TJ
U -3 OS CO





SI
CN
VO
^.
n



o
o




o
o
o









0
VO




^








X
m
as
O
C
0

T)
3
O
4

jj
U
41
2

T3
0)


cn


„_.

E



C
•-*
O
Li
4)
4)
Li
E
3
O
TJ
















































O
r*
(N
Ok
^





















2
n
u
a
0)
3
fl
u
£

-------
I
         a\
         ->»'*••
         10
         a H:
         *^O
         vo <
         O 04
                                      IB c
                                      jj nj
                                      O JJ
                                        Cu



                                        O
                                        ca

                                        6
                                  £    -.
                                  ~l    IQ
                                  —    3
                                        
-------
              ATTACHMENT 2

CALCULATION OF SECONDARY AIR IMPACTS FOR
   PROCESS VENT AND WASTEWATER CONTROL

-------
               Calculation Methodology Description

                      Secondary Air Impacts


     Methodologies and example calculations are presented below
for secondary air pollution emissions, from the use of  control
devices on process vent exhaust streams and wastewater streams.
Calculations for all model plants are presented in Tables 5  and 6
of the memorandum.

Process Vents

     No secondary air pollutant emissions are expected from  the
use of scrubbers on process vent exhaust streams.   Flares,
however, will generate secondary air emissions from two sources:
combustion of organics and supplemental fuel in the flare itself,
and combustion of fuel to generate the steam used  by the flare.

     To obtain the total Btu content of the stream burned by the
flare, the Btu content of the process vent exhaust stream was
added to the Btu content of the supplemental fuel.   This total
Btu content was then multiplied by the emission factors for
flares in Section 13.5 of AP-42 to obtain criteria pollutant
emissions.

     The Btu content of the supplemental fuel was  calculated by
first dividing the annual fuel costs (calculated by the OCCM
spreadsheet) by the unit cost of the fuel.  The resultant fuel
usage was multiplied by the Btu content per cubic  foot of natural
gas to obtain the Btu content of the supplemental  fuel.

     The following example is presented for NOX emissions from a
large model plant:

Btu content of the process vent exhaust:
 (8 Btu/ft3)x(220 ft3/min)x(60 min/hr)x(8760  hr/yr)  = 925 MMBtu/yr

Btu content of the supplemental fuel:
       ($1889/yr)/ ( $3 . 03/Mf t3)x(1000 Btu/ft3)  =  623 MMBtu/yr

Total Btu content of the exhaust stream:
           925 MMBtu/yr +  623 MMBtu/yr =  1,548  MMBtu/yr

NOX emissions:
  (1548 MMBtu/yr) x( 0.068 lb/MMBtu)x(l Mg/2200 Ib)  = 0.048 Mg/yr

     It was assumed that the steam required by the flare was
generated by a small industrial boiler with an efficiency of 80
percent burning No.  2 fuel oil.   The fuel oil requirements were
calculated by taking the steam costs from the OCCM spreadsheet,
dividing by the unit cost of the steam,  then multiplying by  the
Btu content per pound of steam (saturated steam at 400 psig).
The resultant Btu/yr was divided by the Btu content of No. 2 fuel

-------
oil  (140,000 Btu/gal)  to  obtain  the  fuel  oil usage.   Emission
factors from Section 1.3  of AP-42 were  then used to  calculate
criteria and toxic pollutant  emissions.   The following  example  is
presented for NOX emissions from a large model plant:

Energy required to produce the steam:
   ($18,166/yr)/($4.65/Mlb)x(1206 Btu/lb)70.80 =  5.9xl09 Btu/yr

Fuel oil requirement:
        (5.9xl09 Btu/yr)/(140,000 Btu/gal) = 42,066 gal/yr

NOX emissions:
     (42.1 Mgal/yr)x(20 lb/Mgal)x(l  Mg/2200 Ib)  = 0.38  Mg/yr

Wastewater

     As with flares, it was assumed  that  the required steam is
generated on-site through the use of a  small industrial boiler
burning No. 2 fuel oil.   Starting with  the  steam costs  calculated
by the HON wastewater  spreadsheet, fuel oil usage and secondary
emissions were calculated as  described  above for process  vents.

-------
        ATTACHMENT  3




CALCULATION OF ENERGY IMPACTS

-------
               Calculation Methodology Description

                         Energy Impacts


     Methodologies and example calculations are presented below
for energy requirements from electricity, natural gas,  and steam
requirements.   Calculations for all model plants are presented in
Table 7 of the memorandum.

Electricity

     Annual electricity costs for the scrubber were calculated by
the OCCM spreadsheet.   The  annual electricity requirements (kW-
hr/yr)  for the steam stripper were calculated by the HON
wastewater spreadsheet.  Unit cost used by both spreadsheets  is
$0.059/kW-hr.   An energy conversion "factor of 3,415 Btu/kW-hr was
used to convert electricity requirements to Btu/yr.  The
following example is presented for a small model plant  using  a
scrubber:

   ($77.04/yr)/($0.059/kW-hr)x(3415 Btu/kW-hr) = 4.46xl06 Btu/yr

Natural Gas

     Annual natural gas costs for the flares were calculated  by
the OCCM spreadsheet.   Unit cost used by .the spreadsheet  is
$3.03/1000 ft3.  Natural gas  usage was  converted to Btu/yr using
the factor 1,000 Btu/ft3.  The  following example is presented for
a large model  plant:

   ($1889/yr) / ($3.03/1000 ft3)x(1000  Btu/ft3) =  6.23xl08 Btu/yr

Steam

     Annual steam costs for the flares were calculated  by  the
OCCM spreadsheet,  and annual steam costs for the steam  strippers
were calculated by the HON  wastewater spreadsheet.  Unit cost
used by the OCCM spreadsheet is $4.65/1000 Ib steam, and
$3.49/1000 Ib  steam by the  HON wastewater spreadsheet.  Steam
usage was  converted to Btu/yr using the factor 1,206 Btu/lb
steam.   This conversion factor was taken from the HON wastewater
spreadsheet and assumes that the steam is saturated at  400 psig.
It was  assumed that the steam would be produced by a small
industrial boiler with an efficiency of 80 percent.  The
following example is presented for a flare at a large model,,
plant:

($18,166/yr)/($4.65/1000 lb)x(1206 Btu/lb)/O.80 = 5.89xl06 Btu/yr

-------
         ATTACHMENT 4




CALCULATION OF  SCRUBBER COSTS

-------
                Calculation Methodology Description

                          Scrubber- Costs

     The  following procedure was  taken from Volume IB of the HON
BID, and  most  of the same assumptions  concerning vapor-liquid
equilibrium were followed here.   The ..process vent exhaust stream
characteristics (e.g.,  flow rate  and pollutant concentration)
were determined from the Section  114 questionnaire responses, and
is discussed further in the model plant section of the
memorandum.

     The  average flow rate of the process vent exhaust streams
reported  in the Section 114 questionnaire responses was 250
actual cubic feet per minute (acfm) ; however, the input to the
HON costing procedure requires the  flow to be measured at
standard  conditions.   The flow was  corrected to standard
conditions  (77°F was  used as  standard  temperature in order to be
consistent  with other physical property values used in various
equations of the costing procedure) , and 220 standard cubic feet
per minute  (scfm)  was used to size  the scrubber.  The majority of
other input parameters were consistent with the that used in the
HON BID,  although three require further discussion.  The
calculation of the absorption factor requires the slope of the
vapor-liquid equilibrium line.  The HON analysis assumed a value
of 0.1 for  the slope,  which is valid if the solute is highly
soluble in  the solvent.   As propylene  oxide is very soluble in
water, this assumption is valid here.  The vapor Schmidt number
was assumed to be equal to 1,  which is generally valid for very
dilute streams.   The liquid Schmidt number was assumed to be 350,
as diffusivity data for propylene oxide and water was not
available.   However,  the liquid Schmidt number has very little
effect on the  overall design of the column.  Using a liquid
Schmidt number as high as 1000 changes the column height from
11.2 feet to only 11.4 feet.

-------
Scrubber  Capital  Investment
     The  scrubber design  and capital investment is based on a
packed bed, countercurrent scrubber tower.  The scrubber capital
investment  includes  the cost of the scrubber, packing, platform,
stack, and  any  associated ducts and fans.  The capital investment
is directly related  to the size of the scrubber and the  pressure
drop across the scrubber, which is dependent on the liquid and
vapor flow  rates  through  the column.  The absorbing liquid is
water, and  the  vapor consists mainly of air.  The design and
costing procedures closely follow the procedure presented in
Chapter 9 of  the  OCCM (Gas Absorbers) , and the Handbook  on
Control Technologies for  Hazardous Air Pollutants (HAP Manual)1.

     1.   Calculate the flow rate of gas to the scrubber.
     The  amount of vapor  flowing to the scrubber is equal to the
average exhaust flow rate from the reactor vents reported in the
questionnaires.   The vapor stream is predominantly air.


       Mols Vapor             /    •  \   (     , ,.
        (  lb-moll = (220.0 scfm) *  60 ™   J	LJ
        [   hr  J              I   hr )   (392 scf (at T = 77°F)J

                =     Jb-mol
                         hr


     2.   Calculate the liquid flow rate through the scrubber.
     It is  assumed that the liquid and vapor flow rates  through
the column  are  essentially constant.  Therefore,  the liquid to
vapor flow  rate ratio throughout the column is also constant and,
as used in  the  HON analysis, equal to 17 gpm/1000 scfm.   By
converting  both quantities to Ib-moles per hour and multiplying
the ratio by  the  vapor flow rate,  the liquid flow rate is found.
     '     Handbook - Control Technologies for Hazardous Air
Pollutants.  Publication Number EPA-625/6-86-014.   United  States
Environmental Protection Agency, Air and Energy Engineering
Research Laboratory, Research Triangle Park, North Carolina.
September 1986.

-------
           17 gpm  _
         1000 scfm ~
                     17
-2L
min
 hr j
                      gal
                                  lb"mo1'
                                   18 lb
1000
472.6
153.06
                             min
                  1 Ib-mol
                  392 scf
                                                 60
                           hr
                               water
                           hr
                                vapor
Liquid
Flow Rate
Through =
Scrubber
(Ib-mol/hr)
Ib-mol
4726 Water
hr
Ib-mol
15306 Vap°r
1 hr J
                                 __ fi7 Ib-mol Vapor
                                           hr
                               = 103.96
                                  Ib-mol
                                   hT~
     3.   Calculate the diameter of  the tower.
     As  explained in Section 3.1.2.1.5 of Volume IB  of the HON
BID, a correlation for randomly packed towers based  on flooding
considerations is used to determine the tower diameter.  The
vapor flow rate per cross-sectional area of the column is first
determined by the correlation.   By  knowing the vapor flow rate,
the cross-sectional area of  the column can then be found.
                   Liquid Flowrate
        ABSCISSA =
                    Vapor Flowrate  —
                      F          I hrj
                             hr
hrj    ( Density of Vapor (py)
      \ Density of Liquid (PL)


     18 lb }
 33.67
                                       .5
                  Ib-mol
       hr
                                 ,29

                                       Ib-mol,
                                                 0.0739
                               lb
                               ft*
                          62.2
                                                       lb
                                                           .5
                  = 0.0661

-------
      ORDINATE = -0.9809237 * (ABSCISS A)(~°-0065226 * ln (ABSCISSA))

                  + (ABSCISSA)(-°-021897)

                = 0.126
      The vapor flow rate per cross-sectional area is calculated
by  the following equation.   The  column is assumed to operate at
60  percent of flooding.
             Vapor Flowrate
            Per Cross-Sectional

                    lb  )
                  	-»
                  sec-ft^ J
where:

     pv

     PL
      uL/2.42

      a/e3
             ORDINATE * pv * pL * gc
                          ML  .2
                         2.42 J
                                    1/2
                                       * f
    density of the  vapor = 0.0739  lb/ft3

    density of water = 62.2 lb/ft3

    gravitational constant =
    32.2  fflbm/lbf-sec2

    viscosity of solvent = 0.85 cp

    the void space  of packing per  surface area to
    volume ratio of packing2  = 69.1 ft

    flooding factor = 0.6
         Vapor
        Flowrate
       per Cross -
        Sectional
         Area
sec-ft
= I 0.126 * 0.0739 * 62.2 * — *	L™-11/2
                    69.1   (0.85)1
                                          *0.6
                                              lb
     Vapor Flowrate
   per Cross-Sectional = 0.3168
         Area              sec'ft"
     -     Treybal,  R.E.  Mass-Transfer Operations,  Third Edition.
New York,  McGraw-Hill Book Company.   1980.   p. 196.

-------
                   Area of Tower
                        (ft2)
     Vapor Flowrate	
    (	secj
     Vapor Flowrate per
    Cross-Sectional Area
          Area of Tower
                         33.67
              (ft2)
Ib mol
  hr
                                                      Ihr
3600 sec,
                                    0.3168
              Ib
                                          sec-
               ft2
                      = 0.856 ft2
                       A      /  j-  %2      ( Diameter 17
                       A = n  * (radius)-' = it *  	
                                          I   2   J
                                 N
                 Diameter _ f 4    ] -
                    (ft)   ~  «

                 Diameter = | — * 0;856| 2 = 1.044 ft
      4.   Determine the number of  transfer units.
      The number  of transfer units  represent the number of
theoretical equilibrium stages required to absorb the vapor
pollutant.  This number is determined from a quantity called the
absorption factor which is calculated by dividing the liquid to
vapor ratio by the slope of the equilibrium line.  The slope of
the equilibrium  line is represented by the difference in the
vapor mole fractions divided by the difference in the liquid mole
fractions from the top to the bottom of the column.  As used in
the HON analysis,  the slope is equal to 0.1.
                                                      0.1
Absorption _
Factor ~
1 51 IK
47° 6 Ib-mol/hr water *
Ib-mol

Ib-mol ,
                    = 19.16

-------
            Number of

             Transfer  _ ^ \(    Hal—Cone   }   ,

              Units       I 0.02 * Hal—Cone J  *
             (NOG)       LV               '
                             1
                        ( 1  -  1/AF;


                NOG = 4.072
      5.   Determine the height of a  transfer unit.
      The height  of a transfer unit  is the  height of  one
theoretical equilibrium stage.   This  number is determined  from
the vapor and liquid flow rates per cross-sectional  area of the
tower and from the Schmidt numbers of the  vapor and  the liquid.



               Vapor Flowrate

             per Cross -Sectional   (      _Jb  }    ( 3600 sec
                             =        — •
               Area  -~            f^sec         hr
                   { ft2 • hr]
                             = 1140.48   Ib
                                      ft2 • hr
              Liquid Flowrate

            per Cross-Sectional

              Area [ -  lb .]  ~           0.856 ft2
                  I ft2 - hrj
                            = 2186.1  - Ib
                                        hr
                     Height of a
                       weight of a         f   1

                      Transfer Unit ~ HG + | TE



where:
            HG = b * ^Vapor Rowrate per Area)c   I _V.fPor |0.5

                     (Liquid Flowrate per Area)d * (

-------
HL =
                      / Liquid Flowrate per Area^s
                      \  Viscosity of the Liquid J
                            Liquid
                            Schmidt
                           ^ Number)
                                         0.5
     The values b, c, d,  Y,  and s and the Schmidt numbers are all
constants which are dependent on the packing  and the  liquid and
vapor flow  rates through  the tower.  The vapor  Schmidt number is
assumed to  be equal to  1,  which is valid for  very dilute gas
streams at  atmospheric  pressure,  and the liquid Schmidt number is
assumed to  be 350.  The height (HOG) of a transfer unit is
calculated  below.
3.82 *  1140.48
              HG =
                                 ft2 • hr
                                       0.41     0.5
                                           * (1)
                               = 2.15
                           2,86.1
                                     hr
                  HOG = 2.15 +
                    1.08
                               = 2.206 ft
     6.  Determine the  height of the column.
     The height of the  column is determined from the number of
transfer units and the  height of a transfer unit.


              Height (ft) s  (HOG * NOG)  + 2 * (0.25 * Diameter)

              Height (ft) =  (2.206 * 4.072) + 2 > (0.25 * 1.044)

                       =  11.24 ft

     7.  Determine the  pressure drop across the tower.
     The pressure drop  is related to the  height of  the tower, the
liquid  and vapor flow rates per cross-sectional area,  and the
liquid  and vapor densities.

-------
                   Pressure  NOG * HOG * g * (10~8)
                    Drop  ~          5.2

                             (10)(r * L"/pL)
                                             Pv
Where:
     NOG  =    Number of Transfer Units
     HOG  =    Height of a  Transfer Unit  (ft)
     pL   =    Density of the  liquid
     pV   =    Density of the  vapor
     L"    =    Liquid Flow  rate per Area  (Ib/f t2«hr) )
     G"    =    Vapor Flow rate per Area  (Ib/ft2»hr) )
     5.2  =    Conversion between lb/ft2  and  inches of H2O
The values of g (11.13)  and r  (0.00295) are constants dependent
on the  tower design parameters.

                *ST • (2-065:24'072) • "-» • co-8) •

                           f  0.00295 * 2186.1]
                        (10)1     62.2     ;

                        ( (1140.48)2)   . ,n .  ,   „ _
                         -i	'—\ = 4.30 inches H?O
                        (  0.0739 J            z

     8.   Calculate the weight  of the tower.
     As  explained in Section 3.1.2.2 of Volume IB  of  the  HON BID,
the weight  of the tower is dependent on the height and diameter
of the tower.

-------
           Weight = (48 * Diameter (ft) * Height (ft)) + 39  * (Diameter)2

           Weight = (48 * 1.044 * 11.24) + 39 * (1.044)2 = 605.8 Ib


      9.   Calculate the capital cost of the scrubber.
      The capital cost  of the tower  and associated equipment .is
calculated from the design parameters.  All  costs are  reported in
July 1989 dollars,  except for total capital  costs, total  annual
direct costs,  total annual indirect costs, and total annual
costs,  which are also  reported in August 1996  dollars.



       Tower Cost = ( 1.900604 * [  wt (Ib) ] Q'93839] „       , Cost Index
                  (         (  1000 IbJ      )            $298.2
       Tower Cost =  1.900604 *     '             * $1000 *
                               1000 IbJ       ;           $298.2

                 = $1417.27



      The volume of  packing is needed to determine packing costs.



                = it  * (Radius (ft))2 * HOG (ft) * NOG
                 = * *  Diameter (ft) 2 . HQQ (ft)



                 = it * [ L044 (ft) 12 * (2.206) ft * (4.072)
                      I    2    /

                 = 7.690 ft3

          Packing  $9.7   $392.8  ..  .      .... ,.  , ,Qn  
-------
                    = [(210 « (24)0-839) + (2 . 4.52 .  (24)1-43)]
                       3355.9
                       $352.4
                   = $3910.80
               Cost
                   = 79.1239 * (24)0-5612 „  $355.9  =
               Cost                      I S342.5
      The platform cost  is dependent on  the diameter of  the
column.

             Platform _ j(10)(0.7884 * In (diameter) + 3.325)] *


                      ( $355.9 }
                      ( $298.2 J
                    - r(10)(°-7884* In (1-044) * 3.325)1 * f S355.9'
                      1         -                  J   I $298.2,
                    = $2727.52

      For consistency with  the HON  BID,  the stack  cost was assumed
to be $5000.00.   The basic equipment cost (EEC) is the  sum of  the
tower,  packing,  duct, fan,  platform,  and stack  cost.


               EEC = $1417.27 + $90.36 + $3910.80 + $489.27 *

                     $2727.52 + $5000.00 = $13,635.22

      The purchased equipment cost  includes the  EEC and
instrumentation,  freight,  and sales tax.
                                   10

-------
          Instrumentation = 0.1  * BEC

                Freight = 0.05 * BEC

              Sales Tax = 0.03 * BEC
                  PEC = 1.18  * BEC = 1.18 * $13,635.22 = $16,089.56
     The  total capital cost of the scrubber  system  incorporates
an installation factor of  2.2.


              TCI  = 2.2 * PEC = 2.2 * $16,089.56 = $35,397.03

                   = $37,992.15  (August 1996$)

Scrubber  Annual Cost
     1.   Calculate the direct annual cost.
     The  direct annual cost associated with  the scrubber includes
the operating and supervisory labor costs, the maintenance
material  and labor costs,  and the utility costs.  It  is assumed
that the  scrubber operates for 8,000 hours per year.   Labor rates
are reported in July 1989  dollars, except where noted otherwise.

     la.   Calculate the operating and supervisory labor.
     It is assumed that the supervisory  labor cost  is 15 percent
of the operating labor cost.


                Operating/          /  05 hrs \
               Supervising = (1.15) *           *
                 Labor            V 8 hr shift )

                            Operating       / $ N
                             Hours   » „,    —
                             (hr/yr)    Wa^e I hrj
                        « , 15 .       «, 8000    *
                                 8 hr        yr    hr

                        = $7590.00/yr


     Ib.   Calculate  the maintenance material  and  maintenance
labor cost.
     It  is assumed that maintenance material  is equal to
maintenance labor .
                                  11

-------
Maintenance = f  °'5 hr } * f °PeratinS *} * (Labor _$_
                          8 hr shift]  * (  Hours   yr
                                    8000^
                         (  8 hr )

      Maintenance Labor         =     $7250.00/yr
      Maintenance Material      =     $7250.00/yr


      Ic .   Calculate utility costs.
      The  scrubber utility  costs  include water and electricity
costs.   The electricity costs depend on the vapor flow and
pressure  drop through the  column.


             Water _ water (Ib/hr)  ^   $0.22   ^ g{)00 hr
              Cost    8.34 Ob/gal)  *  1000 gal        yr
103.96 J*22L * 18 -£_
( hr Ibmol J
8^4 lb
- $°-22 r °000 hr
1000 gal yr
                                 gal

                   = $394.90/yr


                             Vapor

          Electricity  nonni   in.    j.   Pressure ,. ,    „ _..
                   = 0.0002  * Through *  n     (inches H^O)  *
                            Scrubber    UTOp
                            (SCFM)

                      Operating (ji£\    $0.0509
                       Hours  I y,. I *  KW •  hr
                  = 0.0002 * 220.0 SCFM * 4.30 inches *

                      8000 *£ *  $0-0509
                          yr    KW • hr

                  = $77.04/yr
      The direct annual cost  of the scrubber is the  sum of the
costs presented above.
                                   12

-------
          Scrubber
        Direct Annual = Operating Supv.Labor + Maintenance Labor and
            Cost

                      Materials -*- Utilities


                   = $7590.00 + $7250.00 + $7250.00 + $394.90 + $77.04


                   = $22,561.94/yr

                   = $24,216.06/yr (August 1996$)


2.   Calculate  the  scrubber indirect annual cost.
     The scrubber indirect annual  costs  include  the overhead,
taxes,  insurance, administrative costs,  and capital recovery.

     2a.   Calculate  the overhead cost.
     The overhead cost is a function of  the operating and
supervisory labor costs,  and the maintenance labor and materials
costs.

            Overhead = (Operating/Supv. Labor + Maintenance Labor and

                      Materials)  * 0.6


                   = $7590.00 + $7250.00 + $7250.00) *  0.6


                   = $13,254.00/yr


     2b.   Calculate  the taxes,  insurance,  and  administrative
costs,  and the capital recovery costs.   These  costs are dependent
on the  total capital cost of the scrubber.   Capital recovery
isbased on equipment life of 10 years and an interest rate of
7 percent.
                                  13

-------
                         Taxes = 0.01 * TCI

                      Insurance = 0.01 * TCI

                 Administration = 0.02 * TCI

           Capital Recovery Cost = 0.14245 * TCI
                          Cost = °-18245 * TCI = °-18245 * $35,397.03

                              = $6458.19/yr
The  indirect annual  cost  is the sum  of the above costs.

                 Scrubber
                 Annual    _   ,   ,   _      T
                 T ,.   .  = Overhead  + Taxes + Insurance +
                 indirect
                  Cost

                            Administration  +  Capital Recovery


                         = $13,254.00 +  $6458.19 = $19,712.19/yr

                         = $21,157.38/yr (August 1996$)

The  total  annual  cost of  the scrubber  is equal  to the  sum of the
direct annual cost and the indirect  annual cost.

                Scrubber

                Annual  = 522,561.94 + $19,712.19 =  $42,274.13/yr
                 Cost

                        = $45,373.44/yr  (August 1996$)
                                     14

-------
nonepoxide HAP exhaust  steams into a single 500 acfm stream  for
control purposes.  The  remaining 9 catalyst extraction model
plants would control a  single 250 acfm exhaust stream for
nonepoxide HAP emissions from the finishing process.

-------
       ATTACHMENT 5




CALCULATION OF FLARE COSTS

-------
               Calculation Methodology Description

                           Flare Costs
     The OCCM spreadsheet was used to calculate the flare costs
for process vents at large model plants (oxide emissions only)
and catalyst extraction model plants (oxide and nonepoxide HAP
emissions).   The process vent exhaust stream characteristics
(e.g., flow rate and pollutant concentration) were determined
from the Section 114 questionnaire responses and is discussed
further in the model plant section of the memorandum.

     The average flow rate of the process vent exhaust streams
reported in the Section 114 questionnaire responses was 250
actual cubic feet per minute (acfm) .  However, the flow rate
entered into the spreadsheet must be the flow at the flare tip.
As suggested in the OCCM,  the exhaust pressure at the collection
point (1 atmosphere, gauge) is assumed to drop to 1 psig at the
flare tip.   Accounting for this pressure drop reduces the flow
from 250 acfm to 17.0 acfm, which is the value entered into the
spreadsheet.  All other spreadsheet input parameters were based
on physical property data or were taken from tables in the OCCM.

     It was assumed that all process vent exhaust streams are
collected into a. single stream for control purposes  (this stream
is represented by the 250 acfm exhaust stream discussed above) .
Catalyst extraction model plants, however, have oxide emissions
from process vents and nonepoxide HAP emissions from the
finishing steps.  For cost purposes, it was assumed that the
nonepoxide HAP emissions constitute another 250 acfm exhaust
stream.

     The EPA database was used to determine the number of
facilities already achieving the level of control specified in
the regulation for nonepoxide HAP emissions  (90 percent control)
(see Attachment 1 for a discussion of oxide emissions from
process vents).  Of the 9 facilities reporting nonepoxide HAP
emissions,  2 facilities (22 percent) were achieving 90 percent
control or greater.

     The number of uncontrolled catalyst extraction model plants
was then calculated.  The total number of catalyst extraction
model plants  (19) was multiplied by the fraction of  facilities
control at a level below the standard  (78 percent):

    (19  facilities)x(0.78)  = 15 model plants with uncontrolled
                                  nonepoxide HAP emissions

     In Attachment 1, it was shown that there are  6  catalyst
extraction model plants which will require controls  for oxide
emissions from process vents.  The flare costing procedure
assumed that these 6 facilities combined the oxide and

-------
TOTAL ANNUAL COST SPREADSHEET PROGRAM-FLARES [1]

SMALL FACILITY WITH A PRODUCTION RATE OF 50 MM Ibs/yr

COST BASE DATE: March 1990 [2]

VAPCCI (Third Quarter 1995): [3]                                    109

                            INPUT PARAMETERS

- Vent flowrate (acfm):                                17.0
         (Ib/hr):                                   1114.9
- Vent heat content (BTU/scf):                          8.0
- Fuel heat content (BTU/scf):                        1000
- Inlet gas temperature (oF):                            150
- Vent stream density (Ib/scf):                     0.08446
- System pressure (psig):                            14.70
- Liquid density (Ib/ft3):                              51.74

                            DESIGN PARAMETERS

- Gas velocity, max. (ft/sec):                          60.00
- Auxii. fuel requirement (scfrn):                        7.09
- Total gas flowrate (scfrn):                            24.1
- Rare tip diameter (in):                               1.24
- Heat release rate (BTU/hr):                        105603
- Flare height (ft):                                     1.8
- KO drum max. velocity (ft/see):                       4.95
- KO drum min. diameter (in):                           3.2
- KO drum height (in):                                 9.7
- KO drum thickness (in):                              0.25
- No. of pilot burners:                                   1

                            CAPITAL COSTS

Equipment Costs ($):

- Rare/self-supported:                               8,220
- Rare/guy-supported:                                  0
- Flare/derrick-supported:                               0
    Minimum flare cost:                              8,220
    Knockout drum cost:                               78
- Total equipment (base):                            8,298
          (escalated):                               9,880

Purchased Equipment Cost ($):                      11,658

Total Capital Investment ($):                         22,383

-------
                             ANNUAL COST INPUTS
Operating factor (hr/yr):
Operating labor rate ($/hr):
Maintenance labor rate ($/hr):
Operating labor factor (hr/yr):
Maintenance labor factor (hr/sh):
Steam price ($71000 Ib):
Natural gas price ($/mscf):
Annual interest rate (fraction):
Control system life (years):
Capital recovery factor
Taxes, insurance, admin, factor
   Item

Operating labor
Supervisory labor
Maintenance labor
Maintenance materials
Natural gas
Steam
Overhead
Taxes, insurance, administrative
Capital recovery

Total Annual Cost (July 1995$)
         8760
           16
         17.20
          630
          0.5
         4.65
         3.03
         0.07
           15
       0.1098
         0.04

ANNUAL COSTS

    Cost ($/yr)     Wt. Factor    W.F.(cond.)
        9,853
        1,478
        9,419
        9,419
        1,889
       18,166
       18,102
          895
        2,458

       71,679
0.137
0.021
0.131
0.131
0.026
0.253
0.253
0.012
0.034

1.000
 0.673
IMB^
 0.047


 1.000
Costs Revised to August 1996 Dollars:

Total Capital Investment:             $22,425
Total Annual Direct Costs:            $50,317
Total Annual Indirect Costs:          $21,494
Total Annual Costs:                 $71,811
[1] Data used to develop this spreadsheet were taken from Chapter 7 of
the 'OAQPS Control Cost Manual' (4th edition).

[2] Base equipment costs reflect this date.

[3] VAPCCI = Vatavuk Air Pollution Control Cost Index (for flares)
corresponding to year and quarter shown. Base equipment cost, purchased
equipment cost, and total capital investment have been escalated to this
date via the VAPCCI and control equipment vendor data.

-------
TOTAL ANNUAL COST SPREADSHEET PROGRAM-FLARES [1]

SMALL FACILITY WITH A PRODUCTION RATE OF 50 MM Ibs/yr

COST BASE DATE: March 1990 [2]

VAPCCI (Third Quarter 1995): [3]                                    109

                            INPUT PARAMETERS

- Vent flowrate (acfm):                               34.0
         (Ib/hr):                                  2044.8
- Vent heat content (BTU/scf):                        8.0
- Fuel heat content (BTU/scf):                       1000
- Inlet gas temperature (oF):                         150
- Vent stream density (Ib/scf):                    0.08446
— System pressure (psig):                           14.70
- Liquid density (Ib/ft3):                             51.74

                            DESIGN PARAMETERS

- Gas velocity, max. (ft/sec):                        60.00
- Auxii. fuel requirement (scfm):                      14.18
- Total gas flowrate (scfm):                           48.2
- Rare tip diameter (in):                             1.75
- Heat release rate (BTU/hr):                      193682
- Rare height (ft):                                   2.5
— KO drum max. velocity (ft/sec):                      4.95
- KO drum min. diameter (in):                         4.6
- KO drum height (in):                                13.7
- KO drum thickness (in):                             0.25
— No. of pilot burners:                                  1

                            CAPITAL COSTS

Equipment Costs ($):

-Flare/self-supported:                              9,184
— Flare/guy-supported:                                 0
— Rare/derrick-supported:                              0
    Minimum flare cost:                            9,184
    Knockout drum cost:                             129
— Total equipment (base):                          9,313
   '   '    (escalated):                             11,088

Purchased Equipment Cost (S):                      13,084

Total Capital investment (S):                        25,121

-------
                            ANNUAL COST INPUTS
Operating factor (hr/yr):
Operating labor rate ($/hr):
Maintenance labor rate ($/hr):
Operating labor factor (hr/yr):
Maintenance labor factor (hr/sh):
Steam price ($71000 Ib):
Natural gas price ($/mscf):
Annual interest rate (fraction):
Control system life (years):
Capital recovery factor
Taxes, insurance, admin, factor
   Item

Operating labor
Supervisory labor
Maintenance labor
Maintenance materials
Natural gas
Steam
Overhead
Taxes, insurance, administrative
Capital recovery

Total Annual Cost (July 1995$)
          8760
            16
         17.20
           630
           0.5
          4.65
          3.03
          0.07
            15
        0.1098
          0.04

ANNUAL COSTS

    Cost ($/yr)     Wt. Factor    W.F.(cond.)
         9,853
         1,478
         9,419
         9,419
         1,921
        33,317
        18,102
         1,005
         2,758

        87,272
0.113
0.017
0.108
0.108
0.022
0.382
0.207
0.012
0.032

1.000
 0.553
«••
 0.043

 1.000
Costs Revised to August 1996 Dollars:

Total Capital Investment:             $25,167
Total Annual Direct Costs:            $65,528
Total Annual Indirect Costs:          $21,905
Total Annual Costs:                 $87,433
[1] Data used to develop this spreadsheet were taken from Chapter 7 of
the 'OAQPS Control Cost Manual* (4th edition).

[2] Base equipment costs reflect this date.

[3] VAPCCI = Vatavuk Air Pollution Control Cost Index (for flares)
corresponding to year and quarter shown.  Base equipment cost, purchased
equipment cost, and total capital investment have been escalated to this
date via the VAPCCI and control equipment vendor data.

-------
               ATTACHMENT 6




CALCULATION  OF STORAGE  TANK CONTROL COSTS

-------
               Calculation Methodology Description

                    Storage  Tank Control  Costs


     The responses to the EPA/SPI questionnaires were used to
determine the following parameters for the uncontrolled  fixed-
roof tanks:  size, throughput, number per facility, and  chemical
stored.  In addition, the EPA's TANKS3 program  (see Attachment  1)
was used to determine the emission reduction.

     Of the 12 facilities that reported information on storage
tanks, 3 had uncontrolled fixed-roof tanks large enough  to be
covered by proposed regulation.  An analysis of the information
supplied by these 3 facilities showed that there was no  general
trend that related the size and throughput of the tanks  with the
size of the facility.  Therefore, a single "model tank"  for use
with all three model plants was developed.

     The average tank volume was 38,667 gallons, with a
throughput of 134,500 gallons.  As none of the facilities
reported more than one uncontrolled fixed-roof tank, it  was
assumed that there was only one such tank per model plant.
Hexane was used as the chemical stored in the model tank.

     The following procedure for calculating the cost of adding
an internal floating roof to a fixed-roof tank was taken from
Volume IB of the HON BID.

-------
                   COSTS FOR THE INSTALLATION OF
                   AN INTERNAL FLOATING ROOF IN
                         A FIXED-ROOF TANK


     The following calculations- axe used to  estimate the total
annual cost for  the installation and operation of an internal
floating roof in an existing fixed-roof tank.

     The model tank stores toluene and has a capacity of
38,667 gallons.   Additional design parameters  for the example
tank are presented in Table 1.   The internal floating roof,
installed in this model tank, will have a liquid-mounted primary
seal and controlled deck fittings.

     1.  Before  an internal floating roof can  be  installed in the
example tank,  the tank  must be  cleaned and degassed.   The cost
for cleaning and degassing a tank (cstdegas)  is based on its
capacity.

     Cstdegas  =     7.61 ($/gal)  *  [tank_size,  gal]0-5132
     Cstdegas  =     7.61 * [38,667  gal]°-513.2
     Cstdegas  =     $1,720 per  tank

     2.  Determine the  installed capital cost  for a  new internal
floating roof (cstroof)  having  a primary liquid-mounted seal and
controlled deck  fittings.

     Cstroof   =     [509  ($/ft)  * tank_dia(ft} ] + 1160 ($)
     Cstroof   =     [509  ($/ft)  * 23.6 (ft)] + 1160  ($)
     Cstroof   =     $13,172

     3.  The total  capital investment  (TCI)   for the  installation
of an internal floating roof  in  the model tank is the sum of the
tank preparation  costs  and the  installed capital  cost for the
roof.

               Total    _  ,  .   .      Installed
                  -  ,    Tank cleaning       .  ,
              . Capltal  =     and     +   Capl^
              investment   ,   .         cost for
               rr,-.*    degassing cost  ,.   .      ,.
               (TCI)      6    6      floating roof


                        TCI (July 1989S) = 51,720  + $13,172


                        TCI (July 1989S) = 514,892


                        TCI (August 19965) = 515,984

-------
     4.   Indirect  annual costs  (ann_fix) are estimated as a
percentage  of the  total capital investment  (TCI) and include
capital recovery;  maintenance charges (5 percent) ;  inspection
charges  (1  percent);  and taxes, insurance, and administrative
charges  (4  percent) .   The capital  recovery factor  (0.14245) is
based on  7  percent interest over 10  years.


         ann.fix (July 1989$) = $14,892 * [0.14245 + 0.05 + 0.01 + 0.04]


                 Indirect Annual Cost (July 1989$) = $3,611/yr

                 Indirect Annual Cost (August 1996$) = $3,876/yr


     5.   The direct  annual cost (ann_var) is a cost savings equal
to the value of the  recovered product.   This product recovery
credit is based on the HAP emission  reduction achieved by the
internal  floating  roof and the market value of the  HAP.
              Direct Annual     HAP emissions
                 Cost     =    reduction
             (August  1996 $)      (Mg/yr)
-Average market
 value of HAP
    ($/Mg)
            ann_var = 0.62 (Mg/yr) * [-0.40 ($/kg)] * 1,000 (kg/Mg)


                 Direct Annual Cost (August 1996$) = -$248/yr
     6.  The total  annual cost  (TAG)  for the installation and
operation of an internal floating roof is the sum of  the indirect
annual cost  and direct annual cost.


           Total annual cost = Indirect annual cost + Direct annual cost
                Total annual  cost = $3,876  + [-$248]
                 (August 1996$)

                   Total annual cost  = $3,628/yr
                     (August 1996$)

-------
TABLE 1.  MODEL TANK DESIGN  PARAMETERS
Parameter Description
Tank capacity (tank_size)
Annual tank throughput
Tank orientation and type
Tank diameter (tank_dia)
Tank height
Average tank vapor space
height
Adjustment factor for small
diameter tanks
Tank paint factor - white
roof and aluminum color shell
Average ambient diurnal
temperature change
Product factor - organic
liquid other than crude oil
Stored product - HAP
Product molecular weight
Product specific gravity
Product vapor pressure at
25 °C
Product Antoine coefficients
A
B
C
Product average market price
Units
gal
gal/yr
	
feet
feet
feet
dimensionless
dimensionless
OF
dimensionless
	
Ib/lb mol
	
mmHg
psia

dimensionless
dimensionless
dimensionless
August
1996$/kg
Value
38,667
134,500 ..
vertical
fixed roof
23.6
11.8
5.9
1.0
1.3
20
1.0
hexane
86.17
0.659
151.3
2.92

6.87776
1171.530
224.366
0.40

-------
            ATTACHMENT 7




CALCULATION OF EQUIPMENT LEAK COSTS

-------
,,M™™    SMALL MODEL PLANT
UNCONTROLLED COMPONENTS FOR LDAR PROGRAM
i
1 '
Equipment
i Tvne
Compressors
Open-ended
Lines
Sampling
Connections
rressure
Relief
Devices
vaives in
Vapor
Service
vaives in
Light Liquid
Service
Pumps in
Light Liquid
Service
Connectors
Total Number
1
0
25
25
695
690
8
3716
Total with
HON Level of
Control
0
0
0
0
56
14
' 1
1152
Total with
HON
Equivalent
1 Control
0
0
2
10
389
345
3
1747
Total Number
Uncontrolled
1
0
24
15
250
331
4
818
Average
Number
Uncontrolled
1
0
12
8
125
166
2
409

-------
            SMALL MODEL PLANT
UNCONTROLLED COMPONENTS FOR LDAR PROGRAM
Equipment
Tvoe
Compressors
Open-ended
Lines
Sampling
Connections
pressure
Relief
Devices
vaives in
Vapor
Service
vaives in
Light Liquid
Service
Pumps in
Light Liquid
Service
Connectors
Total Number
1
0
25
25
695
690
8
3716
Total with
HON Level of
Control
0
0
0
0
56
14
1
1152
Total with
HON
Equivalent
Control
0
0
2
10
389
345
3
1747
Total Number
Uncontrolled
1
0
24
15
250
331
4
818
Average
Number
Uncontrolled
1
0
12
8
125
166
2
409

-------
             SMALL MODEL PLANT
    INPUT TABLE FOR LDAR PROGRAM COSTS
Eauioment Type
Compressors
Open-ended Lines
Sampling Connections
Pressure Relief Devices
Valves in Gas/Vapor Service
Valves in Light Liquid Service
Pump Seals in Light Liquid Service
Connectors
Total Components
Total
Uncontrolled
1
0
12
8
125
166
2
409
723
Control Equipment
Monitoring Instrument
Replacement Pump Seal
Compressor
Pressure Relief Device
Rupture Disk
Holders, Valves, Installation, etc.
Open-ended Lines
Sample Connections
Base Cost
(July 1 989$)
6,500
180
6,240
78
3,852
102
408
    Initial Leak Detection and Repair
   	   Proaram
Value
Monitoring Fee ($/component)
Initial Leak Frequency (fraction)
   Valves in Gas/Vapor Service
   Valves in Light Liquid Service
   Pump Seals in Light Liquid Service
   Connectors
Leaks Requiring Further Repair
   Valves
   Connectors
   Pumps
Hours for Repairs
   Valves
   Pumps
   Connectors
     2.5

   0.114
   0.065
     0.2
   0.021

    0.25
    0.25
    0.75

       4
      16
       2

-------
             SMALL MODEL PLANT
   INPUT TABLE FOR LDAR PROGRAM COSTS
  Annual Leak Detection and Repair
  	Program	
 Value
Monitoring Fee ($/component)
Additional Pump Monitoring Time (hr)
Initial Leak Frequency (fraction)
   Valves in Gas/Vapor Service
   Vaives in Light Liquid Service
   Pump Seals in Light Liquid Service
   Connectors
Leaks Requiring Further  Repair
   Valves
   Connectors
   Pumps	___	
        2
   0.0083

     0.02
     0.02
      0.1
    0.005

     0.25
     0.25
     0.75
        Maintenance Program
 Value
Calibration/Maintenance for Monitoring
Instrument (July 1989$)
Annual Maintenance Charge for
Compressors, Pressure Relief Devices,
Open-ended Lines, and Sampling
Connections (percent of capital costs)
Replacement Pump Seal Cost (July
1989$)
Miscellaneous Charges for
Compressors, Pressure Relief Devices,
Open-ended Lines, and Sampling
Connections (percent of capital costs)
Miscellaneous Charges for Pump Seals
(percent of capital costs)	
$4,280.00



        5

  $180.00



        4

       80
           Economic Data
 Value
Interest Rate
Economic Life (years)
   Pump Seals
   Rupture Disks
   Monitoring Instrument
   All Other Equipment
CRF Values
   2 Years
   6 Years
   10 Years
Chemical Engineering Plant Cost Index
   July 1989
   August 1996
Recovery Credit
   Average Emission Reduction (Mg/yr)
   Raw Material Cost (July 1989$/Mg)
Labor Rate ($/hour)	
     0.07

        2
        2
        6
       10

   0.5531
   0.2099
   0.1425

    356.0
    382.1

      5.2
   $1,164
   $22.50

-------
CO
CO
o"
CM
CO

CO
N.
c\i
^^
CM

O
0
co
o


col co
CO! O
co'i co'
en


CO


CO
0
CO
CM
in
CM"
 ca
jo

CO
 en
^.
 CO
j£
cq
o
               en
               m
               CO
 CO
^£.
 CO
_2

co
co
 en
J£
 CO
_£

CO
CO
 V)
JC
 CO
_CD

N;

CM
 CO
jc:
 CO
                                                                en
        CO
        CD
3
O
3
O
3
O
CO     T-

d     CM
                                                                                  o      o     o     o
                                                                                  in      in     in     in
                                                                                  CM      CM     c\i     CM
                                                                                  CM      CM     CM     CM
SMALL MODEL PLANT
LEAK DETECTION AND REPAIR ANNUAL COSTS

"O
o
Li-
en
m
ec
CO
kj*
"co
Q.
CD
QC
.12
'E
it
CD
15
3
-«
CO
_CD
•*f
^-~
T»
O
X
to
0)
>
Gas/Vapor Valves
125 gas/vapor va
Light Liquid Valves
e x 0.065 leaks/valve =
ips x 0.200 leaks/pump
> fc
166 light liquid va
Light Liquid Pumps
2 light liquid pu
Connectors
u
&_
o
"o
CD
0
.0
co
x.
ca
_CD
§
o
X

409 connectors


v_
'co
Q.
CD
OC
O)
cz
*w
CT
CD
GC
CO
J£
03
CD
_1
II II II II
in ia to m
CM CM f- CM
o" d o" o
X X X X
S CD |
> > E
Gas/Vapor Valves
14.25 gas/vapor val
Light Liquid Valves
10.79 light liquid val
Light Liquid Pumps
0.4 light liquid pu
Connectors
8.589 connectors



-------
                                                                                       CO
                                                                                       CO
                                                                        CO
                                                                                                     co
                                                                                                     i^
                                                                                                     CM*
                                                                                                     ^f-
                                                                                                     CM
                                                                                                     O
                                                                                                     o
                                                                                                     CO'
                                                                                                     o

                                                                                                     ta-
                                                                                                             CO
                                                                                                             cbi
                                                                                                             o»
                                                                                                      co
                                                                                                      o

                                                                                                      CO
                                                                                                      CO
    CO

    CO
    o
    O
                   CO
                  ^
                   CO

                  .52

                  co
                  CO
                  J£
                  CO
                  _CD


                  CO

                  O
           CO
          ^
           CO
                                co
                                jsc
                                co
                                j»

                                co
                                CO
                                          CO
                                         .*
                                          co
                                         _G>

                                         cq

                                         CO
                                             co
                                             XL
                                             CO
                                             .2

                                             N.

                                             cvi
                                      CO
                                      jz:
                                      co
                                      _£


                                      CO

                                      O
                     CO
                    ^
                     CO
                     CD
                                                                             CM
                  3

                  O
                                                                        O
                                                                        iq
                                                                        cvi
                                                                        CM
           o


           o
           in
           cvi
                                                                                        o
                                                    o
                                                    in
                                                    c\i
                                                    CM
 o


o
in
cvi
CM
_J  O.
UJ  UU
Q  CC
CO
    o
    UJ

    UJ
    o
    UJ
 CD


"co


"to

 CO
 CD
                          CD


                         "5


                         "co
                         ^.
                          CO
                  in
                  co
                  q
                  d
        O


I"     I
3      O
P-    .P
                 CO


                 CO
                 CD
        CO

        CO
        V
              O     -r-
              O     CM
              CM     O

              O     O
                                                                                       CO
                                                                                      J2


                                                                                      1
                                                                                       o
                                                                                CO
                                                                                CD
                                   in
                                   CVJ
                                                        m
                                                        CM
                                      m

                                      d
                                                             in
                                                             CM
                                                             CD
                                                                                                        .
                                                                                                      ta
                                                                                                      CD

                                                                                                     1

                                                                                                      o


                                                                                                     CO
                                                                                                     m
                                                                                                     CD
                                                                                               3
                                                                                               O
                                                                                              CM
•o


 o
u.

 CO

 CO
 CD
 a
 CD
cc
       _     >
   S     CD     |

   .>     ^     E
   CO     CO     3
   >     >     Q.

   o     2     2
   0.    3     3

   CO     2"    2"
,- >  co —  53  —
CO -^5  CD •—  Q-T^

   ffl  >     p




i_ CM T3 CO "O      CO
O T- '5 T- '5      £
Q.     —     —
CO

                     CO
                     CD



                     8

                     O5
                     O
       °"    O
              CD
        CO
        CO
        CO
       o
 t .     t  .     J^


 O3    O)    O

J     J     O
               CO
               0.
               CD
              CC

               O)
                                       cr
                                       CD
                                       DC

                                       co

                                       CO
                                       CD
                                                     OT

                                                     >
                                                     •
 CO


 O
 Q.


1
 CO
 o>

in
                                                               CO
                                                               >
                                                               3
                                                               .0-
                                                        "oj  E
                                                        :=  3
                                                               3
                                                               or
                                                "§>
                                                               d '=

                                                              ""  °"
                                      CO
                                      co

                                      O
                                                            O5
                                                                   O)
                                          CO

                                         2
                                         13
                                          CD
                                          c.
                                          c.
                                          o
                                         O
I

CD
00


00
                                                               tn
                                                              Ta
                                                               o
                                                              O
 w  CO  CO  CO  CO  {/j

 5 To  > "co  p "io
^  CD ~m  QJ  g  CD
 « — -5 ~~ n  ~
> eo •> oo "- co
 i_ CO 2 O* ~° «
 o m  3 CD -5 »

 9-co .SfCM .2"
                                                                                                             CO
                                                                                                             ^£
                                                                                                             CO
                                                                  CO    <
                                                                  Q.     CO
                                                                  CD     CO

                                                                 DC    O
                                                                                                  05
                                                                                                         o>
                                                         co
                                                         o
                                                        "o
                                                         CD
                                                         c

                                                         o
                                                        o
       CO

       CO
       o
       o

       DC


       D_
       01
       DC

-------
    CO

    co
    O
    O
o
0
0
o
o
co"
II
CD
>
CO
"CD
E
o
o
cvi

8
^
00
CO
co"
II
CD
>
CO
"CD
E
o
0
cvj

o
o
CO

f&

II
CD
>
CO
"CD
E
o
0
cvi

o
0
CO
T*
CO
Cff
II
CD
>
CO
"CD
E
o
0
cvi

0
0
CO

t&

II
CD
>
CO
15
E
o
o
cvi

CVJI
•
                                                                                                      CO

                                                                                                     •s
                                                                                                     3£.
                                                                                                      
slli|i Is
> «-^ 0) = 03 0 »_
-=O}CO:=3;= OCD
CO ^ /» •—
>m>*^cocM OJ_CVJT3CO-O S2OOC
2 """ 3 *~ 3 ° ^ m
1 = - s |
"^ — •*- c 'S
« £ £ c «
co en en o £
O 13 13 O Q.
co
CD
15
"5
fe.
8 pressure
,ps (visual)
1
Q_


co
a.
Q.
CVJ


CO
CO
o
Q

AL MONITORING
5



•a
I
CO
ca
CD
"ca
c
<


A/apor Valves
CO
CO
CD
CO
CD
^w
O
Q.
CO
in
CVJ


CO
CD
"co
•a
gp
~o co
'3 *-
cr
en
LJ
co
a.
o.
,~o
' 3
V™
3
.g-
0)
_j
ca
k.
O
"5
CD
o
o
a>
CO O
0 "**"
"o
CD
c
o
O

-------
  UJ
o
o
in
r«-
CO
if}

o
•"
s
   CO

  .CO
   o
   o
             CO

             £

             in
      tn
     ^.
      CD
     J2

     q

     o
eo
^c
co
0>
                           CO
                           J£
                           co
                           co
CO   T^


i-'   CD
                                        o
                                 o
                                   o
                                   in
                            o
                            in

                            cvi    CM
                            CM    CM
                            if}    if}
                                                  o
                                                  o
                                                  in

                                                  CM
                                                  CM
                                                  if}
                                                       o
                                                       o
                                                       in
                                                       cvi
                                                       CM
                                                       if}
_l Q.
LU LU
Q DC

ii
Wc3
  UJ


  UJ
  Q
                            CO
                            0)




                           1
                            O
                      .
                       CO
                      _0>




                      1
                       O
in    in
CM    CM

o    o
                      tn
                           in
                           CM

                           o
                                      CO
                            o


                           CO
                                                       CO
                                                      _2


                                                       £

                                                       o
                                CM
                                                    Q_
                                                    UJ
                                                    DC

                                                     +



                                                    i
                                                    CC

                                                    O
     co
     Q.
     0>
     CC

     O)
     cr
     O)

     DC

     (O
     JC
     CO
co
CO



I
^
o
Q.
CO
                 CO    "


                 I    E


                 2    §.

                 •o
     "3
     CT
   co"   .§•   .2"
,„ -> « ~~ S "~

m*f*
T= Cn CO := 3 :^
            •5   S
                           CO CO CO
   2 «? 1 CM £ »
»_ CO _
O   3 en 3

9-   cfco cr
CO   ™i     i
              CM
          CO

          O
               O)
                    03
  CD



  O

  O
                    co

                   "co
                    a.
                    at
                   CC
                    o
                   O
                         co to
                           .^
                           co
> m

fe^
CL
CO


1
CO
CD
                                            CD
                                            O)
l"
cr
                                          to
                                          JC
                                          ca
                                m
                              «2 co
                                ^~
                              CO
                                                    CD
                                                    c

                                                    o
                                                    O
                                                           CO

                                                           &
                                                           o
                                                           o
                                                           gc
                                                           <
                                                           a.
                                                           UJ
                                                           GC
                                          co

                                          CO
                                          O
                                          O
                                          CC

                                          s

-------
CO

co
O
      o
      co
         CM
ol

°.i

CMI
COI
o
GO

CM
CO
r-

co"
0
o
o
co
CM

0
CO
CO
^f
CM

co
CO
^
CM


-^
co
CM
CO

5
rj-
00
in
CO

S
o
CM
co
in
CM

Tf
CM
CM
CM
CM
CM"
CM 00
Tf ^

cd in
oo en
oo T_

CD" CM"
i- 
CD
C
o
0
~(L
£
A
CO

in co
o ,_•
0



X X
0 0
0 0
CO O
CO CO
00 ••—
"CO
CO
CO
Q.
0.
I
CO
o
"o.
CD
DC





CO
CO
O
O
LU
0
^^
LU
z
<
1
«*J
z
z
<
<
o





co;
LU
a
CC
o;
3n
g
^2
coi
Z3.1
o:
LU
<0
_t
LU
o:
COI












CD
S
_c
O)
Monitor!

^
0
0



X
o
o
o"
o
m
co"



o
CO
CO
£
a.
o
O

^
0
0



X
o
o
o
CM
co"
«5-



CO
Q
CD
Q.
CC

^.
o
o



X
o
q
CM
CO
<*

QJ
co"
"ca
S
CD
•a
o

rf
0
O



X
O
0
co'
ao
o"
CO

CO
o
"o
CD
C
o
O
CD
a.
ca
CO

^3.
0
o



X
o
o
cd

-------
          §
       ol
       o
   CO

   CO
   o
   o
T     CMI
CM     COI
69-     69-1
             O
             CO

             CM
             CO
             f^

             CD"
             if*
          O     O
          q     co

          O     03
          co     **•
          CM     CM
             CO
             03


             CM
                                                               CO

                                                               OJ
                                                               CO
                                                               CM
                   CO


                   Oi
                    O
                    CM

                    oS
                                                                     •»-     CM
                    CM

                    CM
                    CM
                    CM

                    CM"
                              CM  CO

                              co  in
                              co  en
                              co  «sr
                              CD" CM"
                                                                                                    II   II
co
m

in
co

d
co
                                                                                                           CM
Q. <
-1 D-
Ill 111
Q CC
CO
   o
   UJ

   LU
   Q
   UJ
in
p
o
       co
                                 o
                                 CD
                 0
                 d
      q
      d
       q
       d
       T    o
       q    co
       d    d
          o
          q
          co
          O5
      o
      q

      oo
             CO

             fc
             O
             O
             111
             O
                 ~    <
•s
6


Cfl
O
"o
CD
C
C
o
O
_.
*W
Q.
E
CO
CO
^ JO
CO
CD
CO
a.
|
Q.
^^
c
CD
E
CD
O
CO
a.
CD
CC
UJ
t-
•z.
<
*£
<
^3
•ZL

4—.

^J|
^^
o
H-
en;
ui
a
cc
     CM
CO    00     CO     CO

              £*"
                 CO
                 CD

    d     »    *
   .2     C     D-
   UJ     .2     E
                                     12
                                     o
                                     CO
                                     CO
                                     CD

                                     a.


                                     o
                                    O
                                                     CO
                                                     J£
                                                     co
                                                     CD
                                                            en
                                                            CO
                                                  CO

                                                  CD
                    I.    2
                    3    O
                    CC    X
                       o
                       CD

                       c
                       o
                      O
                      JD
                       Q.

                       E
                       CO
                      CO
                       CD


                       CD
                       O
                       JO

                       OL
                       CD
                       CC
UAL CHARGES
                                                 CO

                                                 O
                                                 UJ

                                                 <


                                                 UJ
                                                 o
                                                 CO
                                                                        g
                                                                  o
                                                                  ^T
                                                                  d
o
03

g
                                                                                  Cffl
                                                                                  lil
                                                                                  OI
                                                                                  CD
                                            O.

                                           CO
                                           •a

                                    *     I
                                     o     c

                             =    °^|
                             Oj     g5 o  S

                             CD    '= ° .22

                             01    -i 'S -S
                             «v%l    --«—• lu  ._
                             Cg     c Q. E
                             <|     o CD "a
                             -I!    2 OC <
                                               CO
                                               LU
                                               CJ
                                               CC


                                               O

                                               CC
                                               o
                                               m

-------
SMALL MODEL PLANT
LEAK DETECTION AND REPAIR ANNUAL COSTS

CO!
&
o;
o;
_ji
*s
3
T,
CO
k
0
o
-J
»
O)
CM
U5



ANNUAUCQSTS
_j

>s
>•!
.ts *5
Cost per MMIb Capac
on a capacity of 50 Mf
"co •£
3 CO
C »
C. *"

-------
            LARGE MODEL PLANT
UNCONTROLLED COMPONENTS FOR LDAR PROGRAM
Equipment Type
Compressors
lOpen-ended Lines
Sampling
Connections
Pressure Relief
Devices
Valves in Vapor
Service
Valves in Light
Liquid Service
Pumps in Light
Liquid Service
iConnectors
Total Number
3
85
10
71
36
3270
72
7005
Total with
HON Level of
Control
0
0
0
0
3
65
13
2172
Total with
HON
Equivalent
Control
0
58
1
28
20
1635
9
3292
Total Number
Uncontrolled
3
27
9
43
13
1570
50
1541
Average
Number
Uncontrolled
11!
7!
2
11
3
392
13
385 1

-------
             LARGE MODEL PLANT
   INPUT TABLE FOR LDAR PROGRAM COSTS
          Equipment Type
   Total
Uncontrolled
Compressors
Open-ended Lines
Sampling Connections
Pressure Relief Devices
Valves in Gas/Vapor Service
Valves in Light Liquid Service
Pump Seals in Light Liquid Service
Connectors
Total Components  	
         1
         7
         2
        11
         3
       392!
        13
       385 j
       814
Control Equipment
Monitoring Instrument
Replacement Pump Seal
Compressor
Pressure Relief Device
Rupture Disk
Holders, Valves, Installation, etc.
Open-ended Lines
Sample Connections
Base Cost
(Julv1989$)
6,500
180)
6,240
78|
3,852
102!
408
   Initial Leak Detection and Repair
   	Program	
   Value
Monitoring Fee ($/component)
Initial Leak Frequency (fraction)
   Valves in Gas/Vapor Service
   Valves in Light Liquid Service
   Pump Seals in Light Liquid Service
   Connectors
Leaks Requiring Further Repair
   Valves
   Connectors
   Pumps
Hours for Repairs
   Valves
   Pumps
   Connectors
        2.5 i

     0.114
     0.065
        0.2
     0.021

      0.25
      0.25 j
      0.751

         4
        16!
         2

-------
             LARGE MODEL PLANT
   INPUT TABLE FOR LQAR PROGRAM COSTS
  Annual Leak Detection and Repair
  	    Program 	
 Value
Monitoring Fee ($/component)
Additional Pump Monitoring Time (hr)
Initial Leak Frequency (fraction)
   Valves in Gas/Vapor Service
   Valves in Light Liquid Service
   Pump Seals in Light Liquid Service
   Connectors
Leaks Requiring  Further Repair
   Valves
   Connectors
   Pumps           	
        2
   0.0083

     0.02
     0.02
      0.1
    0.005

     0.25
     0.25
     0.75
        Maintenance Program
 Value
Calibration/Maintenance for Monitoring
Instrument (July 1989$)
Annual Maintenance Charge for
Compressors, Pressure Relief Devices,
Open-ended Lines, and Sampling
Connections (percent of capital costs)
Replacement Pump Seal Cost (July
1989$)
Miscellaneous Charges for
Compressors, Pressure Relief Devices,
Open-ended Lines, and Sampling
Connections (percent of capita] costs)
Miscellaneous Charges for Pump Seals
(percent of capital costs)	
$4,280.00



        5

  $180.00



        4

       80
           Economic Data
 Value
Interest Rate
Economic Life (years)
   Pump Seals
   Rupture Disks
   Monitoring Instrument
   All Other Equipment
CRF Values
   2 Years
   6 Years
   10 Years
Chemical Engineering Plant Cost Index
   July 1989
   August 1996
Recovery Credit
   Average Emission Reduction (Mg/yr)
   Raw Material Cost (July 1989$/Mg)
Labor Rate ($/hour)      	
     0.07

        2
        2
        6
       10

   0.5531
   0.2099
   0.1425

    356.0
    382.1

     20.9
   $1,164
   $22.50

-------
   CO

   CO
   o
   o
      o
      q
      o
      o
      in

      co"
       o
       q

       o
       rr
       CM

       CD"
      o o
      q q
      oo' CM*
      in r>-
      oo co
      <& CM"
                   II  II
                                     o
                                     q
                                     co

                                     co
               o
               o
            o
            q
            o
            o
            in_

            in
                         o
                         in
                                                  o
                                                  co
                                                  o>
               o
               in
               CM
               co
o
in

CM
CD
o
in
cvi
oo
-IE

_! Q.
LU LU
Q DC


ii
LU <

CDz
   o
   LU

   LU
   Q
   LU
                     co


            -    .* i
            O    CO -s
            co    sr CO
            to    ^l >
            CD    CD jj-
            Q.    3 CD



            I    II
                     ,0

                     ts
                     CD


                     o
                     o

                     S.
                     Q.

                     co
                     CD




                    •o
                     CD
                    T3


               CD     CD
               ""^"     i


                     CD
                     Q.
            O
            q

            o
            •

CO ?8


2 2" "

!«s

tis
      c

T- T- O
      O
      _2
      Q.
      E
      CO
      CO
(O
c
O    CO

3    g
CD    ~

c    -o

o    -g
o    ^

CD    Q}

°-co c
E CD CD
CO C Q.
«0 Jj O

cvi TJ r*-
   CO
   •o

   CD
£
co
O
O

<

E
<
o
                                                          O
                                                          LU
                                                          _ll
                                                          oi
                                                          z
                                                           co ^* CM   co     in
 »_    S en 2 i—  w co
 o    "3 co "3     o co

 §•   -2"    &    t3
>    -i    _i     o

5    £    £     i
 fO    C3)    OJ    O
CD    Zj    j     o
                          CO


                          CO

                          O
                          O

                          CD

                          Z

                          DC

                          O

-------
A
0
h-
o
CO
CO
in
0
o
c\i
o
col
01
O3I
te-i
in
03
co'
co_
in
rr
CO*
in
CO
co"
   co

   co
   O
   o
                      CO
                     J£
                      CO
                     .CD

                     CO

                     o
 CO
J£
 CO
_CD

in

in
CM
 CO

 CO
£

CO

cvi
CO
^
co
CD
                                        CO
                                                       CO
                                                      ^£.
                                                       co
                                                       CD
CO
^
CO
CO
                                                            <
 CO
-X
 CO
                           q
                          .c\i
0}
^:
CO
5

q
c\i
                     o


                    i
                    CM
                    £
                           o
                                                                                             o
                                                                                             in
                                                                                             CM
                                                                                             CM
3
O
                                      o
                                      in
                                      CM
                                      CM
3
O
o
in
CM
CM
_l Q.
LU LU
Q EC

ii
LU <"
CD2


*z
   LU
   LU
   Q


5
CO
>
1
CO
_CD
•*

^-
d


0)
"5
>
1
CO
_CD
m
CO
o
d


Q.
3
Q.
1
CO
JD
0
O
CM
d
o
"5
CD
c
o

"33
CO
j»
T—
CM
O
d
                                                      in    in    in
                                                      CM    CM    h-
                                                      CD    d    d
                                             CJ

                                             d
                                                                                       CO
                                                                                       JD

                                                                                       "2j


                                                                                       O
                                                                  co
                                                                  CO

                                                                 I
                                                                  o

                                                                 Tf
                                                            CO
                                                           JD

                                                            2

                                                            o
                                                                       co
                                             CO


                                            "co

                                             3
                                             o


                                            CM
                                                                             Q.
                                                                             LU
                                                                             £E
                                                                                                                     a
                                                                                                                     z
                                                                                                                     cr
   LU

•a
c
3
O
u_
CO
CO
CD
"co
Q.
CD
_
:•!
£
X X X X
III
CO CO 3
,I|I|J 1
-ig)!^!^ s
i^ T3 O5 T3 i"" 05 GO
O "5 <^ *5 O ro
& -2" -2" o
•~> — 1 — 1 CD
CO D5 CT O
O "J "J O


•ing Repair
D-
CD
o:
CO
CO
CD



_>
O
Q.
CO

CO
O
X X
CD (D
> ->
CO CO
gas/vapor \
/alves
light liquid \
^ = in
O 5" CM
_J
O)
LJ
X X
CO
Q.
o
'urnps
light liquid |
connectors
co in
T3 -: W 00
'= 0 °.
"-T "o co
_J (D
o> o
J O



JO
to
0
O
'«
Q.
CD
cr


CO
CD
"co
0
^
(O
CO
O
X

CO CO CO CO
Ji CD JC Q.
CO > CO C
5 "5 S 3
q "3 d 3
d ""•"* f
LJ Ll
S "c
O> O}
LJ _J


CO CO
-s: Jai
CO CO
03 _CD
LO "|"^*
O) ^^ ^^
1
-------
                                                                                                 co
                                                                                                 J£
                                                                                                 CO
   CO

   CO
   O
   O
o
o
CM
j^
ifi'

O
0
CO
0

of
o
0
CM

CO

o
0
C3

CM
of
o
0
CM
CM


^j.
CM
CO
,3.
CM
CJ
CMI CO

S*
. of
 CD

"co


"CD
o
p
CM
 CO

"5


I
       o
       o
 CD


"co
       O
       O


£    £
                     CD




                    I
       O
       O

       CM
              CO

             I
                           o
                           p
                           CM"
CD
                     o


                    CO
                    CO
                    o
                    p
                    o
                                                                                                I

                                                                                                I
                                                                                                 to
                                                                                                ^
                                                                                                 CO
                                                                                                 CM
                                                                                                 O

                                                                                                 O
                                                                                       CO
                                                                                      J£

                                                                                       CO
                                                                                       CO
                                                                                                       OJ
                                                                                       >

                                                                                      I
                                                                                       CO
                                                                                      ^
                                                                                       CO
                                                                                      JD

                                                                                      CM
                                                                                      p
                                                                                      O
                                                 CO
                                                 J£
                                                 CO

                                                 -2

                                                 CO

                                                 iri
                                                  ll


                                                 OL
                                                 E


                                                 I
                                                 ^
                                                 CO
                                                 CO
                                                                                   CO
                                                                                  Ji£
                                                                                   CO
                                                                                   CD
                                                                                                                    CO
                                                                                                                    CM
                                                                                   C

                                                                                   o
                                                                                  .0
                                                                                   co
                                                                                  ^
                                                                                   CO
                                                        in
                                                        o
                                                        p
                                                        o
_i a.
in LJJ
S01

is
01 *c
02

> to CD E CM CO CO CD E CO CO -^ co CD E CM : CM CO CO ^ _ o CL CO .? CO lo JO O5 CO "S 3 .CT 05 _J X CO 5 CO .«. .5" gj CM OJ CO X co CL | Q. •Q '5 o- CO — II $2 3 CT _J ra _j J2 2 o CD C 0 O X CO o "o CD O O in CO CO co § > Is "55 CL CD 3 co to £ a. X CO CD 1 Is £ k. 3 CO 03 £ Q. T" ^« 03 3 to co CL E 3 a. X X CO CL a. CO CO % O O O •z CL O O 2 < •o £ CO ^c CO CO CO c eo CO co > CO o - CL 3 co ^ > co "^ CM "• co „ m >- "2 o> 2 •«- £ oo O '3 CO '3 O ^ - cr " to CO O . O3 O) o o


-------
   CO

   £
   O
   o
    01
0
CM
CD*

4&


o
CO
CO

*~
CM"

0
p
c\i

CM
^J."
*9-
00
CO
o>
in
CM
t&

CO
CD
^f
0
CO
CD"

CM
^
in
CO
r>-
in"
CM
                        to
                       ^
                        CD
                       J)

                       CM
                       O
          w
         .*:
          03
         j£
         in
         crj
         CM
          CO
          J£
          CO
          CD
                    (A

                    CO
                   JD

                   00
                   in
3
0
.C
0
m
CM'
CM
3
O
£
O
in
CM
CM
3
0
€
o
in
CM
CM
3
o
^


CM
CM
_i n.
UJ UJ
Q IT
oz
   o
   UJ
   01
   o
                                                            CO
                                                            CD
                                 3
                                 O
                                              CO
                                              CD
                                               .
                                              3
                                              O
   in    in     in     in
   CM    CM     r*-_     CM
   O    O     O"     CD
                                                  (0
                                                  CD

                                                 1
                                                  3
                                                  O


                                                 CD
                              (0
                              J3


                              1
                              O
                                                                                CM
                                                                                              1X1
                                                                                              cc
        >     E
   co     co     5
   >     >     Q.

              Q.
              CD
              CC
              O)
               CO
              CC
               O
 o    S    S     o
 Q.    3     3     *r
 §    S^S"    §
 >      0.
                                    > CO    CM
                                    i_ T- TJ in 2
                                             -      -
                                           5^ CM .?T
    co

T-  O
   "o
    CD
       
-------
   CO

   CO
   O
   O
           ol   o
           PI   p
           cdi   co
           O!   CO
           t—;   O5
                       O
                       O

                       O
                       CO
                       CM
O
cq
05

CM
CM
CO


CO
tfl-
00
CO


O5
co
cq
CM
CO
      o!
      001
CMI
      CO
      COI
      coj
      i-l   co*
CM
co
m
as
TJ- co
CM CO
d •*
CO O
T- CO
of co*

~£


 II II
mi    i*-
O!    T-;
      O5
      O)
CO

O*
                                                       CO
                                                       CO
si
_l Q.
LU LU
QDC

ii
LU <
O Z

SQ

JB
   LU

   LU
   Q
   UJ
r^
T—
X
0
p
CO

"5
CD
CO
o.
c
CD
CD
CO
Q.
CD
CC
CO
CO
VJ
o
LU
O
^
Aw
<;
z
LU
•z.
<
1
^•J
<
•z.
<
_J
<
S

co:
Ul
O!

O
z:
u;
.}
^y*
z
z
 CC <
                                                                                   CO
                                                                                   LU
                                                                                   O
                                                                                   CC

                                                                                   I
                                                                                   o
                                                                                   CC
                                                                                   o
                                                                                   ffl

                                                                                   3

-------
                    O
                    O
   CO

   CO
   O
   O
O!
Oj
cd I    co
O!    00
*-J    °l
CMI    CO
o
q

8
CM
O
CO

O>
^J-
CM
CM
CO
Tf
CO
co
op
•*
en
co
                                                       CO
o
co!
             CM

CM     Tl-l    CO
00     GDI    in
«•     coi    o>

       •P-l    CO
Tj- 00
C\J CO

O rf
CO O
T- CO

ro'co
                                                                                      u  u
                                                                                               m
                                                                                               o
                                                                                                         CO
                                                                                                         o"
                                                                                                      O5

                                                                                                      o
                                                                                                      co"
                                                                                                      00

_l Q.
LU ai
D DC
OQ
2z
UJ <
oz
   o
   UJ

   UJ
   o
   UJ
             O
             O
             d
             co
jn
 CO
 0)
CO

 Q.
           0)

           o
           u
          jo
           a.
           CD
          CC
          cyj

          co
          O
          O
          UJ
          O
                    UI
                    K
                    Z
          O
^CHARGES
*4J
3;
z
z
ISCELLANEQUS ^
s
*t if if -*r •* o
o o o o o co
CD 0 0 CD O 6
X X X X X X
§ 8 § S S §
§O CO CM CO CO'
TT in r«- t- o
in CM co co co i-
co" co" v
v* <&
onitoring Instrumen
2
ompressors
o
CO
a
CO
5
£
_3
"5.
ir
CM ** m CM
T J2 t»
6* CO
CD
rn
6
UJ
co"
CD
"to
co
CD
T3
O
I
ample Connections
CO
epiacement Pump {
cc
ANNUAL CHARGES
CO
D
O
LU
OTAL MISCELLAN
i-


(01
UI
a
CD
o
33
a
0!
m
o
_j


lonitoring Cost
epair Costs
^ cc
o
•^>
d
X
CM
in
CO
r-
•c in
|»
dministration and S
<

CO
UJ
o
cc
OTAL LABOR CHA
h-

-------
   MODEL PLANT WITH CATALYST EXTRACTION
UNCONTROLLED COMPONENTS FOR LDAR PROGRAM
Equipment
Tvoe
Total Number
Total with
HON Level of
Control
Compressors i 01 0
Open-ended
Lines
Sampling
Connections
Pressure
Relief
Devices
valves in
Vapor
Service
vatves in
Light Liquid
Service
Humps in
Light Liquid
Service
263
13
110
315
906
24
Connectors 1 1988
0
0
0
25
18
4
Total with
HON
Equivalent
Control
Total Number
Uncontrolled
01 0
147
1
43
176
453
9
616.281 934
116
12
67
113
435
11
437
Average
Number
Uncontrolled
o:
58
6
34
57
218
5
2191

-------
   MODEL PLANT WITH CATALYST EXTRACTION
UNCONTROLLED COMPONENTS FOR LDAR PROGRAM
Equipment
Tvoe
Compressors
Open-ended
Lines
Sampling
Connections
Pressure
Relief
Devices
valves in
Vapor
Service
valves in
Light Liquid
Service
Pumps in
Light Liquid
Service
Connectors
Total Number
0
263
13
110
315
906
24
Total with
HON Level of
Control
Total with
HON
Equivalent
Control
Total Number
Uncontrolled
l
Average
Number
Uncontrolled
oi oi oi o:
Oj
0
0
25
18
4
1988! 616.28
147
1
43
176
453
9
116
12
67
113
435
11
934 i ' 437
58 i
1
6
34!
57\
218
5
2191

-------
  MODEL PLANT WITH CATALYST EXTRACTION
   INPUT TABLE FOR LDAR PROGRAM COSTS
           Equipment Type
   Total
Uncontrolled
Compressors
Open-ended Lines
Sampling Connections
Pressure Relief Devices
Valves in Gas/Vapor Service
Valves in Light Liquid Service
Pump Seals in Light Liquid Service
Connectors
Total Components	
         0
        58
         6
        34
        57
       218
         5
       219
       597
Control Equipment
Monitoring Instrument
Replacement Pump Seal
Compressor
Pressure Relief Device
Rupture Disk
Holders, Valves, Installation, etc.
Open-ended Lines
Sample Connections
Base Cost
(Juiv1989$)
6,500
180
6,240
78
3,852
102J
408!
   Initial Leak Detection and Repair
   	   Program     	
  Value
Monitoring Fee ($/component)
Initial Leak Frequency (fraction)
   Valves in Gas/Vapor Service
   Valves in Light Liquid Service
   Pump Seals in Light Liquid Service
   Connectors
Leaks Requiring Further Repair
   Valves
   Connectors
   Pumps
Hours for Repairs
   Valves
   Pumps
   Connectors
       2.5

     0.114
     0.065
       0.2
     0.021

      0.25
      0.25
      0.75

         4
        16
         2

-------
  MODEL PLANT WITH CATALYST EXTRACTION
   INPUT TABLE FOR LDAR PROGRAM COSTS
   Annual Leak Detection and Repair
  	      Program
 Value
Monitoring Fee ($/component)
Additional Pump Monitoring Time (hr)
Initial Leak Frequency (fraction)
   Valves in Gas/Vapor Service
   Valves in Light Liquid Service
   Pump Seals in Light Liquid Service
   Connectors
Leaks Requiring  Further Repair
   Valves
   Connectors
   Pumps	
        2
   0.0083

     0.02
     0.02
      0.1
    0.005

     0.25
     0.25
     0.75
        Maintenance Program
 Value
Calibration/Maintenance for Monitoring
Instrument (July 1989$)
Annual Maintenance Charge for
Compressors, Pressure Relief Devices,
Open-ended Lines, and Sampling
Connections (percent of capital costs)
Replacement Pump Seal Cost (July
1989$)
Miscellaneous Charges for
Compressors, Pressure Relief Devices,
Open-ended Lines, and Sampling
Connections (percent of capital costs)
Miscellaneous Charges for Pump Seals
(percent of capital costs)         	
$4,280.00



        5

  $180.00



        4

       80
           Economic Data
 Value
Interest Rate
Economic Life (years)
   Pump Seals
   Rupture Disks
   Monitoring Instrument
   All Other Equipment
CRF Values
   2 Years
   6 Years
   10 Years
Chemical Engineering Plant Cost Index
   July 1989
   August 1996
Recovery Credit
   Average Emission Reduction (Mg/yr)
   Raw Material Cost (July 1989$/Mg)
Labor Rate ($/hour)     	     	
     0.07

        2
        2
        6
       10

   0.5531
   0.2099
   0.1425

    356.0
    382.1

     10.4
   $1,164
   $22.50

-------
              o
              q
              ci
              o
              in
              co"
O co
P O
00
< _l
CE <
I- D
8
     o o
     q o
     CM cd
     in co
     CO O)
     c\T o"
     «* CD
      II  II
o
q
co


CM"



 II
ol   o
qi   q

Cfii   TT
T-.   CO
Oil   TT

ml   co"
o
in

CM
o
o
in

in
o
m
c\i
ol    o
mi    in
                                                               in
                                                                     CM
I- <
COgg
li<
< Q.
\- UJ


Is
? z
b <
§5

BIB
-j^
UJ Q

S^
O <
        (D


>-    .* 5
O    

£    £ w'
OL   3 CD

E    1.2

8    55
o
q
d
Tj-
CM
co"
     o o
     q q
     CO CM
     h- m
     €£• CO
        co"
                         X X
 o

7j    ®
 (D    ~


 *    ^
      TJ

 0>    §
              o
              o
0.


CO
-CO


O
q

co
o
                                       CO
O


w
01
o;
£2
^^
a:
Oi


ing Instrumen
Q

O
^>


£
o
CO
CO
CD
w
a.
o
O
CO
o
CO
CO
£
Q.
O
u
o





CO
(D
(D
Q
"3
CE
ffl
3

£
ct
CO
CO
TJ
1
"5.
CO




"o
co"
CD
1
co"
CD
T>
o
CO






Connections
 r^
u. in
o
a.
CO

CO
CO
X
CD
1
cr
^ ^j
^n °°
'5 CM
cr
*_j
S
CO
_j
X X
CO
Q.
E
I 1
s= 1
^L *^ *•

^ w
'5
or
LJ
««*
j;
O)
_i
o
o
CO ^
o Si
"o
CD
c
O
o
CO
O
O
CD
Z
CE
O
H;
O
|
<
y—
Q

-------
                                                                                  CM

                                                                                  CO
                                                                                         CO
                                                                                         CO

                                                                                         CO


                                                                                         CO
                                                                                         o
                                                                                         q
                                                                                         d

                                                                                         CM
                                                                                                    *-•   CO
                                                                                                    inl   co
                                                                                               CM


                                                                                               CO
                                                                                               o

                                                                                               CM"
   CO

O CO


00
< -J
CC <
K D
                co
               j£
                CO
               JOJ


               in

               CD
co

CD

.2

CM

rr
                           co
                           jc
                           ca
                           CD
                            co
                           j£
                            CD

                           ^

                           CO
                               en
                               ^
                               CO
                               _co


                               co
                                                       CO
                                                       x.
                                                       co
                                                       2.

                                                       in

                                                       co
                CO
               ^c
                CO

               SL

               oq

               C5
CO
.*:
co
CD
O
o
m
CM
CM
3
o
o
in
CM
CM
o
0
in
CM
CM
o
o
in
CM
CM
K <


§i
< Q.
j-UJ
< CC
i-S
56
_j UJ
UJ Q

Q ^

O<
I

 co

 CD
 CD
 CO
 a.
 CD
CC
               0)

               "5

               "co
               jx:
               CD
               o
                     I
                     CO
                     co
                     in
                     co
                     o
                     d
          CO
          CD


         "CD
                     CO
                     a.



                     I
                     ^
                     CD
               I


               1
               JC
               co
               CD
                     o    f
                     O    CM
                     CM    q
                     d    d
                                                 in    in
                                                 CM    CM
                                                 d    d
                                           in    in
                                           r-.    CM

                                           d    d
                           •D    g
                                       'CD
                                       a.
                                       CD
                                       cc

                                       D)
                               co

                               >
                               CD
                                                 o
                                                 S
                                                       J
                                                       CO
                                                             1
                                                             3
                                                                   CO
                                                    «-

                                                 » 52 g-2    i
                                                 SL-= 0) = 05    o

            CD

            CD
3 CM 3
cr    cr
j    13

£    £
_o>    as
_i    LJ
                              5 ™


                              CD



                              O

                              O
                                       CD
                                       QC
                                       co
                                       CD
> CO ^
i- en jo
o-*  ^

&«-^
                                                                   05
                                                          s:    -a
                                                          ^
                                        S    £    £    i
                                        co    en    en    o
                                        O    Z3 '   Ii    O
















CO
to
O
O
&••
CO
a.
CD
CC
nj co co
© 05 O
~ ~ «=
CO CO CO
333
. O 0 O
.C -C .C
"3- ^ CO




X





,„ CO CO CO 52 CO
CO ^» m _^ Q. _v>
S> CO > CO P CO
•= ^ "co 2, 3 ^i
CD »^ ft
> in > co°- m
*- CM 2 ^ S Is-
0 CD 3 in 3 -J
8--5COS-0

CO .C JC
co o) o
CD Zi 13
CD
.£
"«
o
CM










CO
' CD
J2r
CD
C
o
0













CO
CO
O
0
cc
CL
UJ
CC
^f
5


DC
TJJ
a.
UJ
cc

0
•z.
o
t;
2
O
CO
O
c
cc
c
p
^
c
\-

-------
o
cq
|x^
o
CO
i/>
O
CM
^^
f^»
i—
T-"
O
o
d
CM
CD
i-T
CO
CO
|s»i
CO
co
CM
^r in
•r-! CM_
*^i co"
CO
CO
in
*^f
CO
in
               w
              -*:
               CO
              co
                     CO
                     CD
            CO
            .*:
            co
            CO
      i-    m
      co    -"fr
CO
.X
CO

J£

cq

co
o
o
m
CM
CM
o
o
in
CM
CM
o
o
in
CM
CM
o
o
in
CM
CM
O co
po
00
< -I
CE <
I- D
x z
LJJZ
I- <
< a.
§5


II
< O
-
LU O

8*
o <
2 Hi
                               co
                              5


                              1
                               o
                   .
                   co
                   CD
                   O
un    in   in    in
CM    CM   r*.    CM
o'    o   o    ci
CO
CD
o


CO
                                                             ca
o


CVJ
Q.
LU
CE
                                                                       O


w
CO
Q.
CD
cc
Requiring
CO
CQ
CD



CO
apor Valve
CO
CO
CD
x
1
>
O
Q_
CO
-^ CO
°^ II-
ZI
O5
Z3
X
CO
0)
>
cr
'•= CO
**^ C^L
^3i £•
CM ~"
CO |O
CM g.
Hj
O)
"_J
X X
Q.
|
°- «
1 8
~ CO
H c
o> o
= o
CO ^f
2 co
o *~
Q
C
o
O



u.
"CO
Q.
CD
CC
"o
••*
CO
O
O
X


^ CO
co —
> CM
is
CO
(O
CO
o
X

CQ ^
> «
15^
l§
o-™
i£
01
'Jj
X

Q_ OT
£ CO
"^ 0}
1^^
^2 •
Of
LJ
C3)
*_j
X

CO
CO
J£
0 "N
CD
0
O

CO
O
O
- REPAIR
|

OSTS (MONITOR!
O
cc
o
<
z.
<
i
<
H

-------
O co
PO
00
< _J
cc <
h- D
x z
UJ^
I- <
                                            o
                                            cq

                                            o
                                            CO
                                      o
                                      CM
                                            O
                                            O

                                            O
                                            CM
                                            CO
                                                CO
                                                CO
                                                       CO

                                                       CM
                                                       in
                                                       CM

                                                       co-
                                             co
                                             cq
                                             in

                                             co
                                             in
              


                 CO

                 CO
1
o
in
CM
CM
O
O
in
c\i
CM
o
o
m
c\i
CM
o
o
in
CM
CM
H UJ
< DC


II


II
< O
-J UJ
O. h-
-j s
UJ Q
Q v-
0<
2 UJ
 in
 CM
        in
        CM
             in
                   in
                   CM
                   d
                                 .
                                CO
                                SL


                                1
                                o
                                      co
                                      Si


                                      1
                                      O
                                          co
                                         .2
                                         is
                           
              g    co
              CO
              Q.

              E
              3
              Q.
                              CO
CO
 8.   '§    =
 o    ~ M~

 li^ts
 O) CO := 3 :=
 CO > CM Z! CD
 CO ^3 CO ^^
 C«5  3 cvj 3
' ^~ .2" m .2"
co
O
           C3)
   o
   CD


   I
   u


«2

O CO
o *-
CD


O
O
                        CO
                             CD
                                CO
                                ^
                                CO
                                   CO
                                   CD
                                     CO CO
                        j-j    •*- S.  co J2 3 ^    g»    =
CO
Q.
CD

CC
                        o
                        o
                             > CM
                                     co
                                            co
                                            ®

                                           in
                         co

                         co
                                                m
                            CO
                            CO

                           O
                                   O)
                                        O)
                      o
                      O
CO

fe
O
O
cc
<
CL
UJ
cc
CO

CO
o
o
cc
<
Q


<

•z.

<
     g

-------
   CO

O 03

F O
00
o
o
d
CO
CM
*
o
0
d



o
CD
CM
CO
T—

o

CO
^^
in
co"
o
CO
in
en
CM

o
TT
CM
CM


01 0
o
d
CM
oi
i-> CO
COI i-

CM
                                                               o
                                                               q
                                                               ci
                                                               CO
                                                               CM
q
d
     CO
CM
h--

03
CO
CM

in"
CM
o>
ol
Pi
CO!
CM
N-

d
in
CO

CD"
t <
K-

15
-
LU Q
Qi<:
0<








COI
o:
0!
ui
O!
z
o
z
UK
L^_
z
i
«ji
3
z
z
o












"c
CD
to
c
O)
_c
2
0
2
If)
o
o"




X

§
d




Cfl
o
CO
CO
0.
E
o
o
in in in
o o o
odd




XXX

000
o o o
CM CO CO
in co T-
CO O> O5
CM" o" in"
ff± prj ^5.

fl /n
0 "
m co o
CD -JS
co" £ 0
^ > -1 ®
£ 1 « §
Q > -0 Q
CD CO o o
3 CD c Q.
Q. 2 CD £
a o Q. «
OC X O CO
in in
°. TT
o




X X

o o
q q
co d
^ CO

^
CD
CO
a.
a.
CD
CD
O
cn
a.
CD
DC



CO

CO
O
o
UJ
^^
•z.
UJ
1-
-z.
i
1
^m*
13
Z.
•z.
_J
i





COI
LU
a

o
X!
^
O
z
z
coi
=x
q
<0
IU!
a
COj
s
o
o"




X

O
o
§
in
co"

"c
CD
<^B
s
1o
0)
o
1
2
o
d




X

o
0
d




CO
&_
o
CO
CO
CD
a.
0
O
^ Tf Tj-
000
odd




XXX

0 O O
o o q
CM CO CO
in co ^r
co o> ^r
CM" o" CM"
<&• CO tf*

r< CO
O j?Z
UJ O
co" o'
J2 5 S
.2 « o
Q > 0
0} OJ" CD
3 CD Q.
a. 2 £
3 O CO
QC 31 CO
o
CO
d




X

o
o
d
00
J2
03
CD
CO
a.
E
— »
^
a.
c
CD
CD
0
CO
a.
CD
CC
CO
in
UJ
0

2
CO

Ui
_1
UJ
o
CO
2
I
H-

-------





z f—
O CO
00
< -J
cr. <
ml
H <
^J <
< Q.
1- UJ
< DC
x°
t <
H!
Z H-
< o

Q- 1—
_j LU
LU O
Q N£
o<
iii





















-

















»-j
C/7;
o:
a
LU
o;
z
 in
CO
CM"




CO £
o «
8 Q
CD CD
Q. 3
E a.
0 3
0 CL

































.
££
UJ
co"
CD
"co
2
CD
TJ
O
O
^
CO
^f
in
CD"
V*
II










in
0
d




X

o
o
CO
CO
O)
o"
CO

1&








































CO
CD
_c
T3
CD
T3
CD
CD
a.
O
o
CO
in
a>
CM
O3-

II










in
o
d




X

o
o
CO
OJ
in"








o
0
0
"* qi CM


CM
CM
Ol 0)
^»
oo
^ COI t^




























CO

Connectio
CD
Q.
CO
CO
t«-

II










in
o
d




X

o
o
cd

«£







««•

II










in






X

o
o
d
CO
CO ***
CO
CD
CO
a.
|
£
CD
E
CD
O
CO
a.
(D
QC
CM
W
















CO
CO
0
o
UJ
o

<
•^
111
H
z
^
•^
ANNUAL h
_j
. O
                 o
          (O
          cvj
                 CO
                 q

                 cd
                 o
          CM
          N.

          CO
          CO
          CM

          in
                                     CM
                                     O)
                            01
                            q

                            cd

                            S!
                         CM
                         r*-.

                         o
                         m
                         CO
                         co"
          rf
          O
q
d
                 q
                 d
          q
          d
           q
           d
cm
LU!
a
m
          o
          q
          d
          o
          m
          CD"
          o
          o
   o
   o
   o
   o
                 CM     co
                 in     CD
                 co     a>
                    o
                    q
                    cd
    co     co

    =     3
           CC


           a
    X     _j

    o     S
    q     5


    -     \
    CO     <
                 CM
                        8
                                     CM
Z

-------
O CO

K O
bo
< _1
CC <
H- D
X 2
            O
            CO

            in
            o
            T—

            cvf
                           a
                           UJ
                           cc
                           o
                           2
                           O
                                            in
< OL
h- LU
?2
2 H
< O
-l UJ
0. K

-J W
UJQ a
Q ^ O!
            O)
o
o
rr



            rf
            d
                    cm
                    i^
                    O!
                    a
                           G
                           UJ
                           GC


                           O
                           CO
                           CO

                           3.
                           UJ
                           I
                           CO

                           CO
                           o
                           o
                           II
                          CO
                          fe
                          o
                          o
                    
-------
   CO

O co
P O
00
< _J
CC <
h- D
x z
CO
< 0.
I- UJ

is
t <
           o
           CD
           IT)
           O

           CM"
           0)
     D
     UJ
     CC
     o


     o

     o
     I)
     D
     111
     CC


     O
     CO
     CO
                     in
                     in
< O
-J UJ
Q- K
-*£
UJ Q z

§ * —

2 S}8
   -*=>!
     Q
     IU1
     CC

     z
     O!
     Ull
     d
      a
      cc
      u;
           o
           q
           rf
           (O
     UJ
                             CO
                             o
           O)
     co    U
                        co
                        o
                        o
co   co
S2   co
8   8
*   *
I   5
i   1
           in
           in
           v*
           ll
                             CO

                             &
                             O
                             O
                                co
                                O)
                                OJ
                             -J CO
                             < 3
                             I- O)
                                        O -2


                                        II
                                        o H
                                        _Q  O
                                        ^  8

                                        !«
                                        ^" °
                                        to  to
                                        85

                                        1?

-------
           ATTACHMENT  8




CALCULATION OF STEAM STRIPPER COSTS

-------
                Calculation Methodology Description

                       Steam Stripper Costs


     From the  EPA/SPI  questionnaires  responses, it was determined
that only two  facilities  have Group 1 wastewater streams.  As
both of  these  facilities  would fall under  the classification, of a
large model  plant,  it  was assumed that small and catalyst
extraction model  plants do not have Group  1 wastewater streams.
Consequently,  the average of  the wastewater flow rates reported
by the two large  facilities was used to estimate steam stripper
costs for large model  plants.

     The HON wastewater spreadsheet was used to calculate the
cost of  the  steam stripper.

-------




















a.
O
CO
s
m
a
CO
in
m
a>
^*
f—
CD













(X
^
s
3
CO
o
o
z
2
a.
^
a
UJ
o
C/3
O



^_
O
•50
a> ui
H Q-
.°
"3
™"5


8|


| §






Equipment
Component



r^
O)



a>
OJ
CM_
r--"
CO
«/»



OJ
^*
V
^^
V*


03
tn

^™






CO
in
CO







OJ
0
CO
CM


OJ
a-
Tn






i&
TJ
cu
CO
u.
CO »-
fs. CO
co



T 0
r^ o
CO O
^ CM
»• f»-
*^


•*r OJ
f»- OJ
CO CO
V co"
T™ ^r
6^ <^


O 00
o ••-
o •«•
q in

*~ '""






CO CO
in m
CO CO







CO OJ
% 3
CM


a>
en f--
00 w
OJ o-
>• '**





as
w»
^•0) O)
.
3



^_
f^
3
« J:
^^"
o o

CQ a>
•- 1
ol 2.
CO
CM

•

00
OJ
CN
^jT
V*




CN

CM"
tf^



OJ
in
m

^^






CO
m
CO







OJ
in
CM
CM



OJ
O)
4



Si
c
Collection Deca
73
CO
CO
cu
0
OJ
in
0
a



CM

^
<^




O
o

V*




o
(0

^™






co
in
CO







OJ
CO




CO
oo
OJ
CO
(a





S
to
03

CD
E
(TJ
u.
CM
T—



CO
CO
CO
CO
CM
*^


OO

CVI
CO"
CM



CO
CO
«M
a

^^






CO
in
CO







oo
5-
CO



CD
00
OJ
I






"^
at
a.
a.

-------
A. CAPITAL COSTS
  1) Steam Stopper Column
                   Di = inside diameter, ft =                                     1.3
                 Lt TI = tangent-to-tang length. ft=                              29.0
                   Ts = est wall thickness, ft =                               0.0521
                   rno(ss304), Jb/ft*3 =                             501

                 Weight of the column:

                   Ws = pi'Di( Tl * 0.8116*Di)Ts*mo

                 Ws. Ib =           3.286

                 BE SURE WEIGHT. DIAMETER. AND TOWER HEIGHT ARE WITHIN THE LIMITS
                  FOR ALL EQUATIONS
                 The weight equation for calculating steam stnpper column cost
                 is the same for carbon and stainless steel, but the densities of
                 the two metals are different.
                 - density of carOon steel = 490 lb/ft*3
                 - density of stainless steel = 501 fb/ft*3


   a. Chemical Engineering Magazine. December 28. 1981. pg. 180.
                   COST INDEX =                  230.9

                 -Shell, skirts, nozzles, and manholes

                          Cbs = exp{6.823 + 0.14178 (In Ws) •»• 0.02468 (In Ws)*2]
                            9020 Ib < Ws < 2.470.000 Ib
                                       WEIGHT TOO SMALL

                          Cbs. S =                 .           14,608
                          Cbs(304ss). S = Cbs'Xcol =            43971
                          Xcol.304ss = 3.01 (Based on Vendor data)
                 -Platforms and Ladders
                          Cbp = 151.81*01*0.63316^1*0.80161
                            3 ft < Oi < 24 ft  : 57.5 ft< U < 170 ft
                                       DIAMETER TOO SMALL
                                       LENGTH TOO SHORT

                          Cbp, $ *                2707

                 -Cost of Trays - 304 stainless sieve trays

                          Cbt = (Number of trays)'278.38*exp(0.1739*0]
                            2ft
-------
STEAM STRIPPER COSTING PROCEDURE: BASED
ON ASPEN PRINTOUT - 300LPM (80GPM) CASE
STEAM/FEED RATIO. KG STEAM/KG WW-
WASTEWATER FEED RATE. GPM:
                     TAG:
FLOW RATES           TCI:
  FEED FLOW. KG/HR:
BOTTOMS FLOW. KG^WR:
 STEAM FLOW, KG^R:
OVERHEAD FLOW.KGMR:
 PRI COND H2O, KG/S:
                    0.040
                  16.7750
                 S111.150
                 S415,586
                0.333818375
   3802.33
3766.19215
152.182596
188.318694
3.82909831
 GPM:
 GPM:
tCWhr
 GPM:
 GPM:
    16.8
    16.6
426415.6
     0.8
    60.8
TDAC       S34.830
TIAC       $76.320

         DENSITY H2O  = 8.33lb/gal
         DENSITY Steam = 8.00lb/gal
         Kilogram. Kg - 2.205 Ib
EQUIPMENT

PREHEATER AREA. M*2:       50.7502412
PRI COND AREA. MA2:        4.58713858

COLUMN DIAMETER. M:       0.34893212
                NTU:            16.25

NO. OF COLUMN TRAYS:              10
 ACTIVE HEIGHT. FT:                 15
 TOTAL HEIGHT. FT:                 29
           FTA2:
           FT*2:

           FT:
                     546.0
                      49.4

                       1.3
                                              WATT=9.486E-04 8TU/S
                                              1,750 based on teiecon with
                                              Rob Goldman of Edwards Eng.
                 0.13 ft»2 area based on
                 teiecon with Rob Goldman of
                 Edwards Engineering.

-------
                                   (53.83 -i- 40.71 -(tower wall thicknessOn))
                                0.375 inches < Ts < 2 inches
                                           THICKNESS OKAY

                              no. of manholes =                         10
                              mt diam, in =                             w  (THIS IS INHERENT IN THE CURVE)

                              Manhole cost S =                     14.269

                   -Nozzles Fig 15-25 (SUGGEST USING: ONE.6 inch(ovemeads); ONE.4 incft(feed):
                      TWO.4 incn(steam); and ONE.4 inch(bottoms)

                              Nozzles cost a SUM[(no. of nozzlesCfincnes of nozzle)!*
                                   (24.57 * 35.94'(towerwall thickness, in))
                                0.375 inches < Ts < 2 inches
                                          THICKNESS OKAY

                              sum[(no. noz)(in noz.)] = [(4*4"}+(TS")] =                           22

                              Nozzles cost, 5 =                      1.035

                   -Trays Fig 15-26 304 stainless only

                              Trays cost = (no. of trays)'214.54-exp(0.2075*Di)
                                2 ft < Oi < MS ft
                                          DIAMETER TOO SMALL

                              Trays cost, 5 =                         2.828

                   -Udders Tables
                              S.43/lb of ladder. 30lb/ft of ladder
                              Ladder height = column height =                       29
                              Ladder cost, S =                        374.1

                   -Platforms and Handrails Table 9
                              5.43/Ib  ( d=4ft. 1700th )
                              Assume a linear relationship - 425 ib/tt of diameter
                              P and H cost, S =                     243.22

                   -Insulation Fig 15-27
                              Use foamglass (ave)
                              Assume 3" thick (ave)
                              S10/R2
                              A (ft*2)=            142.28125

                              Insulation cost $ =                  1422.8125

                   Total column cost. S -                           42.944
• TOTAL COLUMN COST, S =                          46.699

-------
 2) Feed TanKs - Chemical Engineering Magazine January 25.198Z P6 144-46.
                  COST INDEX =                    230.9 (sum CE index (Jan-Aprt4))

                Cb = exp(2.331 + 1.3673 (In V) - 0.063088 (In V)*2]
                   i.300gal
-------
5) Primary condensers-Chemical Engmeermg Magazine November 21,
              1988. pg. 66 - 75. Ftgs. 1 & 2. cost index =                                 343

              Assume fixed-tube, single shell, single pass exchangers and
              12 foot length tubes.

              Pri Cond cost » 2228.8*exp(0.0041 rA,ft*2)
                        150ft*2*(122 ft H2O)'(8.33 Ib/gaD*
                (min/60 s)*(lbf/lb)/[0.64n(0.00l341 hp)/(0.7376 ft Ibl/s)J

              Overhead hp =                       0.0
              GROUP A               FLOW TOO SMALL            GROUP A: HP TOO SMALL
                                                                 GROUPS: HP TOO SMALL
              OVERHEAD PUMP. S =             2.256

              Total Cost of Pumps = sum of (feed.bottoms.overhead)

              TOT PUMP COST. S =             26.218

              Conversion from hp to kw-hr/yr is (W)/(0.001341hp)'(KW/1000W>*
               (8424 hr/yr)

-------
   7) Flame arrestors - Casts gwen by Penny Lassiter from a tetecon
                  by  A. Giteiman (RTT).  Eacft arrester is estimated at $100.  Cost
                  is in September 1986 dottars. COST INDEX =                               319
                  From an RT1 report from Paul Peterson to Susan Thomloe. EPA7CPB.
                  dated January 18,  1988: it is assumed that vent lines from the storage
                  tanks, condensers, and decanters each had flame arresters in place.
                  The current approach assumes that these vent lines are routed bade to the
                  feed storage tank for the steam stnpper and a that there is a single
                  flame arrester on the vent line from the storage tank to the control device.


                  Cost of 1  arrester, S =                                too
8. ANNUAL COSTS - Utility costs BASIS: 8424 hr/yr operation

   Utility costs consist of electricity, steam, and water costs

   1)  Electricity - Electncrty is needed for pumps and for
                  refrigeration.

                  a. Pumps - electricity use is calculated under the capital costs

                  Pump Electncrty use * (Total pump use in hp)"0.7457KW(mech.)/hp-(8424 hr/yr)

                  Tot Pump Use. HP *                   2.46

                  PUMP ELECT USE. KW-H/yr =                    15,462


                  TOTAL ELECT USE. KW-H/YR=                  15.462
  2) Steam - Steam is needed for stripping the organics.
                 Steam cost is estimated at $9.26/Mg (S4.20/lbm).
                 This cost is presented in July 1989 dollars. There is no cost index.
                 The steam is 400 psig sard steam with a heat content of 2802 KJ/kg
                 (1206 BTU/lnm).
                 The heat loss through the column is 5% of the heat input
                 Telecon. Kristine Scott, Radian with Rob Stepian, APV Crepaco.

                 steam use = (steam ftow.kg/hr)'(Mg/10A3 kg)'(1.05)*(8424hr/yr)

                 Steam use.  Mg/yr =                  t .345

  3) Water - Cooling water is used in the primary condenser.

                 Water use = (Pri cond H2O flow.gpm)'(60 min/hr)-(3.785Iiters/gal>*
                    (8424hr/yr)*(fraction lost to cooling tower evaporation and
                    slowdown - 0.04)
                 WATER USE, L/YR =            4 653 791

-------
E9R
                                                                 A-96-38
                                                                 n-B-7
Incorporated	   1996 EPA Outstanding Small Business Contractor
    MEMORANDUM  .                    .               -

    DATE:      May 16, 1997

    SUBJECT:  The Definition of an  Extended Cookout as a Control
               Technique for the Polyether Alcohol  Production Industry

    FROM: .     Joanne C. Seaman, EC/R  Incorporated

    TO:        David Svendsgaard, EPA/OAQPS/ESD/OCG


         This memorandum defines an extended cookout (ECO)  as a
    control option for ethylene oxide  (EO) and propylene oxide (PO)
     (epoxide)  emissions -from the production of polyether alcohols
    made with epoxides.  Also included  in this memorandum  is the
    description of the default- onset point for the ECO which can be
    used by facilities that prefer  to accept this  value in lieu of
    calculating their own onset point for the'ECO.

    BACKGROUND

        . 'Polyether alcohols made with epoxides are predominately
    produced in a.batch process. Epoxide emissions from the'batch
    reactor vent result, in part, from  the release of unreacted '
    epoxide when the reactor is opened  during any  stage' of the batch
    reaction cycle.   Typically, during  a batch where an extended
    cookout (ECO)  is not practiced,, the operator finishes  adding the
    epoxide to the reactor, and the reaction is stopped by either
    cooling the reactor or neutralizing the catalyst.  The reactor is
    then opened in order to vent excess gas pressure, and  transfer
    the .reacted product to other equipment for additional  handling
     (such as catalyst removal or product purification)  or  transfer
    the reacted product directly to product storage if further
    processing is not required.  Facilities maximizing production
    from a  reactor would typically  empty the reactor as' quickly as
    possible to start a new batch cycle.  The reactor is emptied
    despite raw material economics  that indicate that it 'would be
    advantageous to keep the product in the reactor,  and the reactor
    emissions are typically vented  to a control device.

         Facilities that practice ECO use this as  an alternative to,
    or in addition to add-on control devices to control epoxide
    emissions from process vents.   Extended cookout reduces the
    amount  of unreacted epoxide in  the  reactor by  allowing it to
    react for a longer time period, thereby reducing epoxide
    emissions.  For an ECO, the product continues  to react in the
            South Square Office                          Research Triangle Park Office
   3721-D University Drive • Durham. North Carolina 27707       2327 Englert Drive. Suite 100 • Durham. North Carolina 27713
     Telephone: (919) 493-6099 • Fax: (919) 493-6393             Telephone: (919) 484-0222 • Fax: (919) 484-0122

-------
reactor for longer than  is  calculated to  be  economically
advantageous.  This additional  reaction period (or ECO) reduces
the emissions of epoxide that otherwise would vent to  a control
device or may make an add-on control  device  unnecessary to
achieve the required epoxide emission reduction.   Since the
epoxide that would have  been vented is converted  to product,  the
producer's raw material  usage decreases.

     A facility using ECO,  however, assumes  the disadvantage  of
decreased annual throughput, because  the  reactor  is used for
emission reduction cookout  rather  than to react another batch
immediately.  Additionally, there  is  the  potential for decreased
product quality, due to  the possibility of secondary reactions
occurring during the ECO.   When a  producer uses ECO, it is going
beyond the economic balance of  raw material  savings versus
production capacity.  The ECO is being used  to reduce  emissions
and the lost capacity is not compensated  by  the recovered raw
material.

CALCULATION FOR THE ONSET OP AN EXTENDED  COOKOUT

     The calculation for the onset of  an  ECO assumes that a
producer will operate a cookout as long as it is  economically
viable; that is, as long as the value  of  the epoxide recovered as
product exceeds the value of the production  lost  due to the ECO
batch time. Beyond this point,  the producer  has no economic
incentive to continue to cookout the  remaining epoxide.  At this
economic breakpoint the cookout becomes an emission control .
Therefore,  that economic breakpoint is  considered the  onset of
ECO for the purpose of evaluating  the  emission reduction.

     For the purpose of determining a  "default" ECO onset,
representatives from Union  Carbide Corporation performed a
calculation for four product classes.  This calculation required
the following confidential  information: the  batch size,- the
epoxide concentration at the end of the epoxide feed step; the
batch time without the extended cookout,  the amount of time the
reactor runs on line per year;  the cost of the epoxide; and,  the
product variable margin  (profit of the  product, dollar per pound
     The calculation takes into account only the liquid phase
epoxide.  Some epoxide in the vapor phase is also recovered and
reacted into useful product during cookout, but this  is normally
small and may be ignored without affecting the break  point.
Also, this analysis does not take into account the capital
charges for the reactor itself/ it only takes into account 'the
annual lost production due to longer batches.  A capital
utilization charge per hour of reactor batch time could be
included in the calculation; this would reduce the time required
to reach the economic breakpoint .  This was not done  in this
calculation for the sake of simplicity.

-------
Equations

     Union Carbide has data to show that a first order reaction
rate fits  the base-catalyzed alcohol alkoxylations its facilities
conduct using either a monol or polyol starter.   Based on this
knowledge,  the rate constant was calculated for each product used
in the evaluation of ECO.   The rate constant was determined .by
solving the rate equation for K,  and then inserting the epoxide
concentration at the end of the epoxide addition (C0) and  at any
given time (Ct) ,  as  indicated  in Equation  1.


                         c    =c    *e''*•''
                          £.pox, c   Epox.,a


                                              Equation 1.
     where:

           C spox.t = the concentration of the epoxide in the
           reactor liquid at a  given time,  weight percent  (wt
           percent);
           C EPOX.O  - the concentration of  the  epoxide  in  the
           reactor liquid after the usual end of  the reaction
           phase,  wt  percent;
           K = reaction rate constant,  I/hours;
           t = time,  hours.

     Next,  the amount of epoxide converted to product (pounds per
batch)  as  a result of an ECO was calculated as  presented in
Equation 2.

                         E con = E to - E cf

                                              Equation 2.

     where:
           E con  =  the amount of epoxide  converted to product as a
           result  of  the  cookout,  pounds per batch (Ib/batch) ;
           E co = the amount of epoxide in the reactor at the end
           of the  usual batch time,  Ib/batch. Calculated by
           multiplying the concentration from Equation 1
           (CEPOX.O) with the batch size;
           E tf = the amount of epoxide in the reactor at the end
           of the  cookout interval,  Ib/batch. Calculated by
           multiplying the concentration from Equation 1 (CEpdx,t)
           with the batch size.

Next, the  value of the epoxide converted into additional product
was calculated.   This savings  was calculated with Equation 3 .

-------
      where:
                        .co = E con * B/yr * Cost Epox

                                              Equation 3.
           Val rpox-co = Value of the epoxide converted to product,
           dollars per year ($/yr);
           E con  =  the  amount of epoxide converted to  product as a
           result  of the cookout,  pounds  per batch  (Ib/batch') ,
           from Equation 2;
           B/yr =  number of batches per year at  the extended
           cookout rate,
           Cost  Epox = cost of the epoxide, dollars per pound.

     Then,  the  amount of production  lost,  in dollars lost per
year, as  a result of  the reactor  being used longer for an ECO was
calculated by multiplying the difference in the number of batches
made without an ECO and the number made  with an ECO,  at that
given cookout time, by the product variable margin ($/lb) .
Equation  4  illustrates this calculation.

            PL =  (B w/o CO/yr -  B w/CO/yr) * BS * Prof

                                             Equation 4.

     where:
           PL =  production lost due to the cookout,  $/yr;
          B  w/o CO/yr =  number of batches that  would have
          occurred without the cookout,  batches/yr;
          B  w/CO/yr = number of batches  at  the  cookout rate,
          batches/yr;
          BS =  batch  size, Ib/batch, and;
          Prof  =  the  profit made on  the  product, $/lb.

     Finally, the  incremental economic analysis was  calculated by
first determining  the cost analysis  at time  2.  The  cost  analysis
was calculated  as  the difference between the value of epoxide
converted minus the lost  production, -as  presented  in Equation 5.

                       CAt = Val Epox, co, t -  LPe

                                         Equation 5.

     where:
          CA =  cost analysis at a given  cookout time;
          Val Epox, Co = value of the epoxide converted  to product
          for that given  time, $/yr, from Equation 3,  and,-'
          LP  =  lost production for that  given time,  $/yr  from
          Equation 4.

To be conducted incrementally, the cost  analysis conducted  at
time 2 is subtracted  from the cost analysis done at  time  1.

-------
Extended Cookout Onset Determination

     These calculations were made on a spreadsheet at one  tenth
of an hour intervals,  and the incremental cookout economics
calculated.  The break-even point,  or the "onset" of  an  ECO,  was
identified as the point in time when the incremental cookout
economics changed from a positive to--a negative value.

     The break-even point was identified for four different
classes of products made by Union Carbide.  For all four classes
of products,  it was determined that the break-even point occurred
prior to the point in time when the combined unreacted epoxide
concentration in the reactor liquid is equal to 25 percent of the
concentration of the epoxide at the end of the epoxide feed  step.
The variables used in these calculations are considered
confidential business information,  and therefore cannot be
reproduced for this memorandum.   However, the individual facility
can recreate this calculation for the products made at that
facility and document a different ECO point of onset.  The rule
will be written to reflect flexibility in setting a point of
onset for an ECO.

-------
MEMORANDUM

TO:          David Svendsgaard

FROM:       Linda Chappell

SUBJECT:   Economic Analysis for the National Emissions Standards for Hazardous Air
              Pollutants (NESHAP) for the Polyether Polyols Industry

DATE:       May 21,1997
       A brief discussion of the economic analysis and regulatory flexibility analysis conducted
for the Polyether Polyols NESHAP follows.

I.  ECONOMIC IMPACTS

       The goal of the economic impact analysis is to estimate the market response of the
polyether polyols industry to the emission standards and to determine any adverse effects that may
result from the regulation.  Polyether polyols are classified as a thermoset resin and typically
produced as an intermediate product or as an input into other products. The majority of polyether
polyols are used for the manufacture of urethanes, surface-active agents, functional fluids, and
synthetic lubricants. Approximately 79 facilities owned by 36 different companies produce
polyether polyols domestically, and 72 of these facilities may potentially be affected by the
regulation.  The polyether polyol facilities are dispersed throughout the country in 22 different
states with the largest concentration of 18 facilities in Texas. Of the 36 companies producing
polyether polyols, seven are classified as small businesses.

       Since the nationwide annualized cost of this regulation of $7.7 million represents
approximately 0.06 percent of the estimated 1996 sales revenues for domestically produced
polyether polyols, the EPA determined that the regulation is not likely to have a significant impact
on this industry as a whole.  For this reason, a streamlined economic analysis is performed.  The
goal of this streamlined analysis is to determine whether individual facilities producing polyether
polyols and companies owning those facilities are likely to be adversely impacted by the
regulation. Facility-specific impacts are examined to assess the likelihood of facility closures and
employment impacts.

       A. Analytical Approach and Control Cost Estimates

       Three different model plants (small, large, and catalyst extraction model plants) are
developed to estimate facility and nationwide annualized and capital control costs for the
regulation. The capital and annualized costs for each of the model plants, as well as estimates of
the nationwide costs are shown on Table 1. The capital costs for the regulation are estimated to

                                            1

-------
be $10.1 million and the annualized costs $7.7 million.  (All values are shown in 1996 dollars.)
                                          Table 1
     Model Plant and Nationwide Control Cost Estimates - Polyether Polyols NESHAP
                                (Thousands of 1996 dollars)
Model Plant/ Nationwide
Small1
Large1
Catalyst Extraction1
Nationwide Totals
Capital Costs
$106.7
516.1
222.6
$10,060.0
Annualized Costs
$89.4
292.1
283.1
$7,670.0
  Capital and annualized model plant control cost estimates represent the maximum costs of compliance that any model plant will incur as a result of
the regulation, and this estimate may overstate costs for some affected facilities.

       Compliance costs are assigned to each facility assuming that each facility must control all
emission points. For example, the small model plant category includes control cost estimates for
three emission points that consist of process vent scrubbers, fixed-roof storage tanks, and
equipment leaks. Specific facilities may not require controls for all emissions points. However,
the analysis assumes that each facility incurs the maximum costs of compliance for plants within a
model plant category.   This assumption results from insufficient facility-specific data and is
necessary to ensure that the costs incurred by any affected facility are not understated.  The
Agency recognizes that, by using this assumption, it is overstating the costs incurred by many
affected facilities.

       Capital costs are annualized at a seven percent discount rate to compute the annual costs
of capital.  Then, the annualized capital costs are combined with annual operating and
maintenance costs, recordkeeping, monitoring, and reporting costs, and other annual costs to
compute the total annualized costs to comply with the proposed rule.

       Model plant annualized control costs are compared to facility-specific revenues for each of
the 72 facilities that will be impacted by the regulation.  A cost-to-sales ratio is developed to
estimate the impact each facility is likely to experience as a result of the regulation. Since
polyether polyols are frequently produced at multi-product facilities and in some cases by firms
owning multiple facilities, revenue data specific to the polyether polyols production of each
facility are not readily available.  For this reason, revenues are estimated  using the production
and/or capacity  data available for facilities in this industry and an estimate of the market price per
pound paid for polyether polyols during 1996 of $1.05.  Actual production data are available for
only 12  companies that responded to an Information Collection Request administered through the
Society of Plastics Institute for the EPA. Based upon the production and capacity data for the 12
surveyed facilities and capacity information obtained from other data sources, an estimate of
annual production for each of the 72 facilities is obtained. Facility-specific revenues are
calculated by multiplying the price per pound of polyether polyols of $1.05 by the estimated

-------
production for each facility.
       The cost-to-sales ratios are computed for each facility in the three model plant categories
by dividing model plant costs by the estimated facility-specific revenues. A cost-to-sales ratio
exceeding one percent is determined to be an initial screening criteria for a significant facility-
specific impact.

Facility-Specific Impacts

       Descriptive statistics including the estimated minimum, maximum, mean, and median cost-
to-sales ratios for the 72 facilities are shown for each model plant category on Table 2. While the
mean cost-to-sales ratios for each model plant category are well below one percent and as such
determined not to be significant, the maximum cost-to-sales ratio in the catalyst extraction model
plant category exceeds one percent. To examine the impacts more closely, a frequency
distribution of cost-to-sales ratios is developed.  This distribution is shown on Table 3.

                                          Table 2
                Facility Impacts of the Proposed Polyether Polyols NESHAP

Model Plant Size/ Statistic
Small Model Plant:
Minimum
Maximum
Mean
Median
Catalyst Extraction Model Plant:
Minimum
Maximum
Mean
Median
Large Model Plant:
Minimum
Maximum
Mean
Median
Total Annual Cost / Estimated Sales
Revenue ("/o)1

0.028
0.881
0.280
0.192

0.169
1.542
0.448
0.373

0.139
0.232
0.145
0.139
' Assumes that the mean sales for facility with data available are the sales levels for facilities for which data are unavailable.

       Only one facility out of the 72 facilities impacted is expected to experience a cost-to-sales
ratio exceeding one percent.  The facility for which costs exceed one percent of sales is estimated
to produce about 23 million pounds of polyether polyols per year. Total annualized compliance
costs are estimated to be about 1.5 percent of annual sales for this facility, which is classified in

-------
the catalyst extraction model plant category.  This facility is owned by a large, financially strong
company.  Company sales were more than $2 billion in 1996, with net income more than $200
million.  The compliance costs are an insignificant share of those resources, so it is probable that
the company will choose to comply with the regulation, rather than shutting down their polyether
polyol production.

       Based on the foregoing, the EPA concludes that the proposed NESHAP for polyether
polyols will impose costs that are negligible on the majority of facilities. For only one facility out
of 72 in the industry do costs exceed one percent of sales.  Based on an analysis of the costs of
compliance compared to facility and company financial data, the Agency finds it unlikely that the
company owning this facility will choose to close it, because the company is financially robust and
the costs are a small share of the company sales and net income. The generally  small scale of the
impacts suggests that there will be no significant impacts on markets for the products made using
polyether polyols, such as polyurethanes.

       Costs do not exceed one percent of company sales for any of the companies owning
facilities producing polyether polyols.  Thus,  the Agency concludes that no company will be made
likely to incur bankruptcy as a result of this regulation.

REGULATORY FLEXIBILITY ANALYSIS

       The Regulatory Flexibility Act (RFA) provides that, whenever an agency promulgates a
final rule under 5 U.S.C. (MARK) 553, after being required to publish a general notice of
proposed rulemaking, an agency must prepare a final regulatory flexibility analysis unless the head
of the agency certifies that the final rule will not have a significant economic impact on a
substantial number  of small entities.  Pursuant to section 605(b) of the Regulatory Flexibility Act,
5 U.S.C. 605 (b), it is certified that this rule will not have a significant impact on a substantial
number of small entities.

       The EPA analyzed the potential impact of the rule on small entities  and  determined that
only seven of the 36 polyether polyol producing firms are small entities — not substantial number
of entities.  Of these seven, no small companies will experience an increase in costs as a result of
the promulgation of this rule that is greater than one percent of revenues.  Therefore, the  Agency
did not prepare an initial regulatory flexibility analysis.

       Although the statute does not require the EPA to prepare an RFA because the
Administrator has certified that the rule will not have a significant economic impact on a
substantial number of small entities, the EPA did  undertake a limited assessment, to the extent it
could, of possible outcomes and the economic effect of these on small polyether polyol producing
entities as previously discussed.

-------
                                          Table 3
             Frequency Distribution:  Total Annual Compliance Cost/Facility
             	Sales by Model Plant Category	
    Model Plant/ Frequency Distribution
Total Number of Facilities
 Small Model Plant Cost-to-Sales Ratio1:
  0 to 0.2 percent
  0.2 to 0.5 percent
  0.5 to 1 percent
  1 to 5 percent
  over 5 percent
 Total Small
            16
            3
            3
            0
            0
            22
 Catalyst Extraction Model Plant Cost-to-Sales
 Ratio1:
   0 to 0.2 percent
   0.2 to 0.5 percent
   0.5 to 1 percent
   1 to 5 percent
   over 5 percent
 Total Catalyst Extraction Model Plant
             1
            12
             1
             1
             0
            15
 Large Model Plant Cost-to-Sales Ratio1:
  0 to 0.2 percent
  0.2 to 0.5 percent
  0.5 to 1 percent
  1 to 5 percent
  over 5 percent
 Total Large Model Plant         	
            29
            4
            2
            0
            0
            35
1 Assumes that the mean sales for facility with data available are the sales levels for facilities for which data are unavailable.

-------


-------