-------
j
•o
O)
_E
•£>
O
oo
CO
LlJ
CQ
?
C
• «• fc
= .£ !
Q. If)
CC
t. CO — > —
0*
*J C
.- O <1»
_j -o
i *
*J C
.£7 41
(C
fc %
P- O
O *•>
O t- *J
u ro o
r- A
ro
11 as
5S
JO 0
I- Q>
' P
o>
«J tft
C CJl
t- "- C
•=r
OJ Ol
"1
0)
c
t. m
•o
(J 4-> >
OJ 3 C •*- (0
.* "O O (J *D
o o "- ro -^
*"* a. re z
0 ^-^
«
1
1
c
o.
£.
0.0.0.^0.0.0.0.0.0.0.0.0.0. o. 0.0.0.
°" 2°"zz °"z "" °" zezz z z
z "• z °" z z z z z z: zz z zz
0.0; 0.0, £0.0,0. o-o. §:§:££ =- g^g
(0
ZZZZ Z ZZZZZZ2 Z ZZZ
zzzzSzzzzz^zzz
,
gsi§§§issaISsK S Sis
^f-TcJ *^ CM" c»7 m in r-i « IH CM CM t-i
fH
ar ^
° z o a- ft. °-
o 5 - x E" g* § «*
*Jfc.-->"l/)l-C «JfRJO, ow^"
S 0 «-^ (S *" ^ lW> ^"S E Er-"5"Q.'a> - 3 *S "w ^
S wi £ w -J- «F- *^ •»- *J J: .c ei w QJ n cjj tn *J *J c
O-J2 fc. F~ i— ^ t— tO 0» O)4J t-IDf— C*JtOlrt"-
e > «J S Q. Q. O. Q.to55_J Q.^£ (Jjottowi QJ
tfoeoeQ-oe=3CDCOoa^Q. to 3 = S
OrH tM« J^lOlDp-COOJO jJeMW ^ Sffir^
°-
z
o.
in
in
fi
5
Monessen, PA
. Chattanooga Coke & Cheol
"•
z
z
a.
^
g
fH
Chattanooga, TN
. Lone Star Steel ,
0.0.
z
0.0.
z z
O. Q.
Z
z z
CM CT1
CM CM
r—
«1 -
4-* 0)
C
. ra
3 Z
m in
z
o.
z
o.
a.
z:
0.
z
o.
o
ro
S
Is"
o o
m
z
0.
a.
^
=
z
o.
o.
z
O-
tM
^-
Weirton, WV
. Wheeling-Pittsburgh,
if:
c_
z
z
z
o.
z
z
o.
z
a.
z
O_
o.
s
E. Steubenville, WV
. Milwaukee Solvay,
in
^
z
1
°-
o.
0_
•Si
o.
o.
a.
o.
z
^
o.
g
CC
^
«
Milwaukee, WI
. J&L Steel, Pittsburgh, P
in
«n
a>
o
0.
in
in o
3 T3
C C
d. O.
r— H-
It 11
a. 6.
z
ndirect final cooler.
ource is controlled.
H* l/l
ro jD
3-7
-------
g2
U «
&
o ^37
tai
M
s :
i *
S5
£>
j«
„ ., 5 j
' •
i>
w
a
ta.
i§ W-
5o
^
3
o E
i !
§*
.*
J OESULFOnilATION
*] PLAMT
1
|
^ ,
CC
1 D >. u
1 < a C »-
£ 3 ?! *
I COW
• f' •• » CL U O
V\ - e
^j^ ^^QJ § j^
z §5
3iu
|5 V Ȥ
s| s
t
ee
n
-* z 2^ o
5^ o L^ o
r A V £>|s
6> §
0
tv. o
fe> 5
^ !/>
2
(/>
wg
^il
a.
L^ K
«
e <>
£ 1 ^ < ^ fe> <
5S 1 8 H *~
*">• u 5
— r^ •> 5 L. . .. .,3
-« 1° *-
c u o L^ o: o
-. ^j nr^T-* S J« r^ s
U- g x ^ S i?> o
U. ZK1 *| 0 < °
-- •..,, ' ,4 1 g J
o i o n_ ' \y ^ o u^ x
"* K u. « w 3
_l £ tn
i *
x r^
J>_
a 2
O U £ 3
P _^ «
** D J
w z
0.
cn_
i
Q.
k_
O
u
O O ew 0 « o £
- " 2 t 8 - +•
M «n w O -rj
. • 1
c.
oOMON O Ji
s «•« 2" Sg° o
ooonONOw—o Sf
S ?K |
s* «
o o D.
S BON; gj
CTJ
o ... — ... _ o
05 < o o o
- Of. K. o — -
*- K o nt
o n - o X
"" — o> — CD
"co
•- — * * O ---
O — — «1 — — — Q Q)
2 5 g | to
0»N- g01 J§~ * C
*- CO
o o c C
« Jj: QJ ^£0
h» f- (t) Q_
| 0 0 S g
o. «i S -o O
O eT 4_* JT"
' x rf -
0" a s «7
u 5 C £ •- f*5
° tp * O Q C Q)
Se^S " i S.0.,,,^ ^
^o«= 2« » i- s S *g c w
Z ZO^^^^Z^Sw 3* §
oaiooQSeiio^^Su0^ a!0 O"^
tiJ
O
3-8
-------
TABLE 3-4. FATE OF COKE OVEN BY-PRODUCTS
Component
Route
H2, CH4, light hydrocarbons,
N2, 02, CO, and C02
Ammonia
Water
H2S, HCN
Benzene, toluene, xylene (BTX)
HC1
Tar bases (such as CsH5N)
Tar acids (such as phenol)
Naphthalenes
Heavy organics (boiling point >200° C)
Remain in gas; used as fuel gas.
Via gas to ammonia scrubber or
via liquor to ammonia still;
then back to gas and to the.
ammonia scrubber. Most ammonia
converted to ammonium sulfate.
Via liquor to ammonia still; re-
mains as waste ammonia liquor.
Via gas or liquor to free ammonia
still and into gas to desulfurizer.
Via gas to light-oil scrubbers.
Via liquor to waste ammonia
liquor as CaCl2 (lime still).
Condensed into tar or via gas
to ammonia scrubber.
Via liquor to dephenolizer or
condensed as tar.
Condensed in tar or via gas
and condensed in final cooler.
Condensed as tar (small fraction
to light oil).
3-9
-------
recovery step. The liquor traditionally is steam-stripped with the addition
of a caustic to return the ammonia to the gas stream for recovery. Ammonium
sulfate crystals that result from an acid contact procedure are separated
from the saturated liquor.
The exhauster is a fan that provides motive power for the gas. A
collection device removes the remaining tar from the gas, generally as a
particulate; both gas scrubbers and electrostatic precipitators are used as
collection devices in the industry.
The final cooler is a pretreatment step for light-oil recovery. In
the process, naphthalene is condensed from the gas and separated from the
cooling water by absorption in tar or by flotation. Light oil is recovered
by absorption in a petroleum fraction wash oil. The light oil is steam-
stripped from the wash oil, and the wash oil is recirculated. Desulfuriza-
tion, which removes hydrogen sulfide from the coke oven gas, is not in
widespread use.
The following subsections further describe the individual processes.
The reader should be aware that (1) today's by-product plants often have
evolved over 20 to 50 years of maintenance, design, and operational changes;
(2) the technology is mature, providing many options for coal chemical
recovery; and (3) the market for coal chemicals is uncertain, and economic
pressure has led to operational changes at the plants. This situation
results in a substantial plant-to-plant process variability.
3.2.2 Tar Processing
3.2.2.1 Tar Decanter. Figure 3-2 outlines the tar separation opera-
tions. Tar condensation initially occurs by direct contact with flushing
liquor in the collecting and suction mains. The gas mains are sprayed and
Vigorously flushed with recycled liquor to quench the gas and to avoid
buildup of tar deposits. Approximately 80 percent of the tar is separated
from the gas in the mains and is flushed to a tar decanter. Twenty percent
of the tar is condensed and collected in a primary cooler along with a
significant amount of water. Tar continues to be removed in the exhauster,
which provides motive power for the gas, and a collector (often an electro-
static precipitator or gas scrubber) removes most of the residual, entrained
tar particulate.
In a tar decanter, the tar is separated from the flushing liquor by
gravity. Typical residence times are about 10 minutes for liquor and
3-10
-------
E
CO
w
o>
s o
3§
is-
= 3 Z
IS*
LL
EXCESS-
AMMONIA
LIQUOR
TANK
2
EC
u cc
UJ LU
c a
il
cc x
LU uj
O uj
i|
i
z o
O 2
il
si
|!
i!
2 NJ
o
£ o
S2
CO
Q.
ID
Cv
CO
I
o>
3-11
-------
40 hours for tar. The degree of separation achieved is highly variable
because of coal type and operating differences between plants. Liquor from
the decanters is recycled to the gas mains.
Tar decanters are often elongated, multicompartment, rectangular tanks
that collect tar on the bottom of the tank and remove flushing liquor at
the top. In addition to these two primary streams, sludge accumulates and
may be collected by a drag conveyor from the bottom of the decanter. The
temperature of the flushing liquor in the decanters is approximately 80° C.
Decanter coal tar generally is stored in vented cylindrical tanks maintained
at 70° to 90° C. Decanted flushing liquor also is stored in tanks that are
vented to the atmosphere.
Multiple decanting stages may be used to reduce the tar's moisture
content. These decanters, which may be covered, commonly are vented to. the
atmosphere. If the tanks are covered, they have hatches to allow access to
the decanter interior. Industry's common target for water in coal tar is
approximately 1 to 2 percent.10
The "heavy" tar that condenses in the mains when the raw coke oven gas
is hit with flushing liquor tends to be richer in pitch and high-boiling
compounds and collects the coal and coke fines entrained out of the ovens
by the gas. In contrast, the "light" or "primary cooler" tar that condenses
in the primary cooler tends to be cleaner, relatively lower boiling, less
viscous, and less dense.
Depending on the plant scale and the design philosophy, these two
streams of tar may be merged or separated. In the latter case, at least
two kinds of decanters are required. One, often called the flushing-liquor
decanter, separates the heavy tar and sludge from the flushing liquor,
which is cooled and recirculated. A second, called the primary-cooler
decanter or primary-cooler tar-intercepting sump, accepts the light tar and
condensate. Some of the condensate is used as makeup to the flushing
liquor and some is forwarded (perhaps through a third kind of decanter) to
ammonia recovery or waste treatment.
"Tar decanter" means the decanter type that accepts either all the tar
or only the heavy tar. The tar decanter may be equipped with a mechanical
device to remove coal tar sludge, coal and coke fines, and adhering tar.
3-12
-------
The tar and liquor that come to either decanter will have been in
recent contact with raw coke oven gas at about the same temperature (60° to
80° C) and pressure (~1 atm) and will be saturated with the components of
that gas. If separation from the gas were perfect, there would be no tar
fog to be removed from the gas and no froth from the liquid. Separation is
never perfect; therefore, any coke oven gas mechanically entrained with the
descending tar and liquor will be delivered to the decanter at a slightly
higher pressure and will build up in the decanter if it is not vented.
If the contents of the decanter are permitted to cool, some of the
mechanically entrained gas will dissolve. However, no reasonable amount of
cooling will dissolve all the gas, and hydrogen is especially difficult to
dissolve. Therefore, the minimum venting rate is related to the design of
the gas/liquid separator upstream; the venting rate will vary necessarily
from plant to plant even when expressed per unit of production.
If the decanter is heated, perhaps in the belief that heating helps
separate the tar from the liquor, some of the dissolved species will'revert
to the gas phase. Thus, heating augments the volume of emissions and
alters their composition. For example, the total amount of benzene emitted
will increase even though the concentration of benzene per unit volume of
emissions may be reduced.
Tar decanter emissions are sensitive to two variables that are not
narrowly limited: residence time in a gas-liquid separator, and optional
heating. The rates recorded by VanOsdell9 and those developed in this
document should be viewed in light of this sensitivity.
Air emissions from a vented decanter depend on the composition and
temperature of the flushing liquor, possible presence of a dispersed light-
organic phase floating on top of the flushing liquor, size and location of
the vents, interior design of the decanter, and wind effects. The emissions
contain significant amounts of benzene and PAH's.9
The estimated rate of benzene emissions from a tar decanter at U.S.
Steel, Fairfield, Alabama, was 15.6 g/Mg of coke.9 The benzene emission
rate measured at a tar decanter at a Pennsylvania steel plant was 1.2 kg/h
(2.6 Ib/h).11 This decanter was one of two for a coke battery. Emissions
from the two decanters are assumed to be twice the emissions from the
single decanter, or 2.4 kg/h (5.2 Ib/h). The corresponding benzene emission
3-13
-------
factor for this decanter would be 84.7 g/Mg coke. One of three tar decan-
ters was tested at a steel plant in Indiana,12 where the average benzene
emission rate was 4.4 kg/h (9.7 Ib/h). The corresponding emissions for
three decanters are 13.3 kg/h (29 Ib/h), which yields a benzene emission
factor of 69.6 g/Mg of coke at this decanter. The average benzene emission
factor from these two decanters is 77.2 g/Mg of coke. The emission factor
is designated as 77 g of benzene per megagram of coke to estimate emissions
from tar decanters.
3.2.2.2 Ball Mill. The tar decanter collects sludge at a rate of
approximately 600 g/Mg of coke produced.13 Recent hazardous waste regula-
tions will tend to discourage disposal of this tar decanter sludge. One
method of recycling the sludge is to process it in a ball mill and recycle
it to the coke ovens. A ball mill is a revolving mill that achieves size
reduction through mechanical impact.
Emissions from the ball mill will depend on temperature and air flow
from the ball mill. A ball mill was observed at the Bethlehem Steel plant
at Bethlehem. The operating temperature was low enough so that benzene and
benzo(a)pyrene (BaP) emissions measured during a pretest screening estimate
were not considered significant.14
Emissions from a ball mill processing tar-decanter sludge apparently
can be controlled if the ball mill is operated at a relatively low tempera-
ture, but excessive temperatures drive off benzene and tar components from
the sludge. Emissions from the ball mill processing tar sludge are believed
to be relatively small at current operating conditions, and the ball mill
is therefore not considered to be a major source.
3.2.2.3 Flushing-Liquor Circulation Tanks. The water that separates
from the tar in the tar decanters is transferred to the flushing-liquor
circulation tanks, as illustrated in Figure 3-2. The cooled flushing
liquor is used to reduce the temperature of the gas leaving the coke oven.
Because water is driven off the coal during the coking process and most of
this water is condensed into the flushing liquor, water must be removed
from the circulating flushing-liquor. This excess flushing liquor is
stored in the excess-ammonia liquor tank.
The emission factor for the flushing-liquor circulation tank (9 g/Mg
of coke) and excess-ammonia liquor tank (9 g/Mg of coke) was obtained from
3-14
-------
a test where the fugitive emissions from a primary-cooler condensate tank
were measured.9 This tank was assumed to be similar to a flushing-liquor
circulation tank and contained liquids similar to those in the excess-
ammonia liquor tank.
Ammonia liquor is produced at a rate of about 7 percent of the coal
rate, or 100,000 g/Mg of coke. If the flushing liquor contained 600 ppm
benzene, the maximum benzene emission rate would be 60 g/Mg of coke. The
benzene emission rate at a particular plant from the storage of flushing
liquor is thought to depend on the number of tanks, the number of vents,
the geometry of the tank, and other factors.
3.2.2.4 Tar Dewatering. The tar-dewatering process reduces the water
content of the tar more efficiently than does the decanting process.
Depending on the plant, the tar-dewatering process may consist of additional
storage time with or without chemical emulsion breakers, centrifugal separa-
tion, steam heating in tar dehydrators, or a combination of these methods.
Centrifugal dewatering should not produce air emissions directly, although
fugitive emissions are possible if any storage vessels are required for
centrifugal dewatering.
In many existing plants, the coal tar is not refined onsite but is
sold to tar refiners. As mentioned previously, a common specification is
that this sold tar should contain no more than 2 percent water; however,
much more than this amount of water usually is mixed into the tar underflow
from the tar decanter. Accordingly, plants dewater the crude coal tar
usually by heating it in tar dehydrators to reduce its viscosity and
providing residence time for water droplets to coalesce and rise to the
surface of the denser tar. Ordinarily, the temperature is maintained above
90° C, and the combined vapor pressures of hydrocarbons over the tar phase
and water over the aqueous phase can exceed 1 atm. The result is a plume
of steam and hydrocarbons from the vent if the tank is vented to the
atmosphere.
Some of the by-product plants dewater tar by heating it with steam
coils to a temperature beyond the boiling point of water.15 The benzene
emissions could depend on the quantity of water vapor or steam driven off
during the dewatering process.
3-15
-------
Emissions from tar dewatering were evaluated at three by-product
plants.12 16 17 The emissions data for tar dewatering at the Fairless
Hills Works (Appendix C) showed higher emissions from the West tank
(3.2 kg/h) than from the East tank (1.1 kg/h). These tanks are operated
in series rather than in parallel, and the wet tar enters the West tar
dehydrator first. Consequently, the emissions from the West tar dehydrator
are expected to be higher than emissions from the East tar dehydrator. The
daily benzene emission rates from the two tar-dewatering tanks at this
first plant were 27 and 76 kg,-respectively. Daily benzene emissions from
tar dewatering at the second plant were 43 kg. The tar is dewatered in
storage at the third plant, where benzene emissions were 24 kg/day. The
benzene emission factors from these three plants were 41, 9.5, and
12.9 g/Mg of coke, respectively. These were averaged to obtain a benzene
emission factor for tar dewatering of 21 g/Mg of coke.
The tar-dewatering tanks contained tar with 200 to 2,000 ppm benzene
in the liquid. Tar, as collected from the flushing liquor and the primary
cooler, can contain greater than 0.2 percent benzene or 2,000 ppm at a rate
of 40 kg/Mg of coke produced. The maximum potential for benzene loss from
tar dewatering and storage calculated from these values is greater than
80 g/Mg of coke. The benzene emissions from tar dewatering and storage
probably will be less than 80 g/Mg of coke and will depend on the method of
operating these processes.
S.2.2.-5 Tar Refining. Emissions from tar refining are essentially
fugitive vapor emissions from vented tanks. Tar-refining plants are rela-
tively unique because each plant has been built and operated to meet spe-
cific market conditions. The basic operations are shown in Figure 3-3.
Emissions from a product storage tank were estimated as 0.008 g of benzene
per megagram of coke and 0.015 g of nonbenzene aromatic hydrocarbons per
megagram of coke, based on measured concentrations and estimated working
losses.9 Benzene emissions from these sources are therefore believed to be
relatively little, and tar-refining emissions are not considered a major
source when compared to others in the by-product plant.
3.2.2.6 Pitch Prilling. The tar recovered in a by-product plant can
be refined by distillation, which separates the tar components into various
fractions according to the relative vapor pressures. The high-boiling
3-16
-------
fraction, which includes some BaP, is called pitch and can be formed into
prills or pellets by prilling.
Approximately 2 million Mg of coal tar pitch are produced annually in
the United States. The pitch is used in the production of carbon elec-
trodes and synthetic graphite, for roofing and paving, and as a binder for
composites such as foundry cones and refractory bricks.
3-16a
-------
co
T3
O)
CO
CO
CO
£
g>
o o
«-
-------
The pitch may be shipped in molten form in tank cars, as cast packages
of convenient size, as lumps after it solidifies, or as extruded pencils or
beads. The latter are known as "prills" and, at one of the very few plants
performing this operation,18 they are glassy spheroids of perhaps 1 mm
(1/16 in.) diameter.
At this plant, the pitch is preheated, filtered, and pumped to a head
tank where it is maintained at a controlled temperature and depth. From
the head tank it drains by gravity through a steel plate perforated so the
individual streams break up into droplets. These droplets, falling into a
stream of recirculated water, are rapidly quenched and rarely agglomerate.
The temperature in the head tank is crucial to proper priller operation.
In times of reduced demand, the priller must be shut down. Restarting
is a nuisance and may constitute an exceptional pollutant source if live
steam must be used to thaw pitch residues.
At the plant discussed above, the hot pitch tank head space is vented
to avoid the buildup of explosive concentrations of hydrocarbons in air.
The venting may be passive (i.e., a fan or steam ejector is not used) and
still at some risk of explosion, or active. If the venting is active, air
is pulled into the head space at various vents or at the edges of the lid,
and a stream of air and hydrocarbons is exhausted. A steam or air ejector
is preferable to a fan because of tar condensation in the vent lines.
The emissions from a pitch prilling operation were measured at a large
tar refinery.19 The flow rate of BaP was 0.00035 g/Mg of coke, considerably
less than BaP emissions from coke batteries at a large plant—approximately
0.11 g/Mg of coke.20
3.2.2.7 Tar and Tar Product Storage. Tar and tar products are stored
in tanks in coke oven by-product plants. The primary cooler tar and the
flushing-liquor tar contain benzene, which can evaporate into the air over
the surface of the fluid inside the tank. Some of the tar products contain
the light components of the tar, which also contains benzene. Each of the
tar and tar products storage tanks can contain BaP and other PAH's. If the
tank is heated, the PAH vapor pressures may be significant.
The vapors from the surface of the liquid enter the head space of the
tank, where they can be emitted into the atmosphere by air displacement
3-18
-------
when the tank is filled. A lesser emission is contributed from tank
"breathing" (volume displacement caused by temperature changes). Emissions
from storage tanks are influenced to some extent by the tank design, which
can influence the amount of benzene and PAH's in the tank head. Storage
tank design for emission control is discussed in Subsection 4.1.11.
Benzene emissions from tar storage were measured at a smaller plant21
as 5.4 kg/day. Another plant17 had benzene emissions of 24 kg/day from tar
storage, but the second plant practiced tar dewatering in the heated tar
storage area. The emissions factors from these two plants were 11 and
12.9 g of benzene per megagram of coke, respectively. The benzene emission
factor for estimating emissions from tar storage, 12 g/Mg of coke, is
obtained when the two emission factors are averaged. Uncontrolled emis-
sions of BaP (before control with a venturi scrubber) were measured as
6.8 g/day from pitch storage at a large plant.19
3.2.3 Ammonia Wastewater Processing
This subsection describes the processes used to recover ammonia and
phenols from wastewater. No significant benzene emission sources have been
identified in ammonia recovery from wastewater.
The ammonia produced in a coke oven is approximately 0.2 percent of
the weight of the coal fed to the ovens. Flushing liquor sprayed into the
collecting mains to cool the gas absorbs some of the ammonia, and more
ammonia is absorbed in the water condensed in the primary cooler (see
Figure 3-1). Flushing liquor contains around 5 to 6 g of ammonia per
liter. Along with ammonia, compounds such as hydrogen sulfide, phenolic
compounds (tar acids), and cyanides dissolve in the flushing liquor. The
distribution of ammonia between the gas and liquid phases depends on operat-
ing conditions and coal composition. Figure 3-1 assumes a phase split
where 75 percent of the ammonia remains in the gas phase.
Several processing options have been developed to recover the ammonia.
The ammonia-handling route shown in Figure 3-1 is known as the semi direct
process and is the option commonly used in the United States. All of the
ammonia eventually is recovered from the gas stream, but a portion enters
the flushing liquor first and is later stripped out.
3-19
-------
For the semi direct process, three alternatives are used for the liquor:
no treatment, free-still ammonia stripping, and free- and fixed-still
ammonia stripping. Based on a recent Environmental Protection Agency (EPA)
survey of the by-product coking industry, all three alternatives are used.22
Out of 52 plants surveyed, 33 plants (53 percent) used or were planning to
use free and fixed stills; 4 plants (8 percent) used only free stills; and
the remainder did not attempt to recover ammonia from excess-ammonia liquor.
3.2.3.1 Ammonia Liquor Treatment. Aqueous ammonia solutions are
decanted from the tar in a variety of processing vessels. Much of this
solution is recycled as flushing liquor, and a portion is constantly drawn
off to additional decanters as excess-ammonia liquor. The ammonia in the
excess-ammonia liquor must be put into the gas phase for recovery via the
acid contactor. The traditional removal technique is steam stripping as
shown in Figure 3-4.
Ammonia removal from the coke oven gas traditionally has been by
contact with sulfuric acid and recovery of crystalline ammonium sulfate.
®
The Phosam process involves the absorption of ammonia in circulating
aqueous ammonium hydrogen phosphate (monobasic) solution, the stripping of
ammonia from this medium, and the condensation of the concentrated ammonia.23
Distillation of the product, either cryogenically or under pressure, yields
a substantially pure ammonia that is more readily marketable than are the
salts.
3.2.4 Tar Acid (Phenol) Processing
Phenol removal is practiced as a part of wastewater treatment and is
not believed to be a significant benzene source. The term phenol is often
used to refer to all the tar acids in the excess-ammonia liquor stream.
However, tar acids consist of approximately 60 to 80 percent phenol, and
the remainder is mostly cresol with small amounts of some higher phenolic
homologs.24 2S Phenol is a minor constituent of coke oven gas, whose
concentration varies according to coking practice and coal composition.
During 20 years of operation, one operator has reported phenol concentra-
tions in the excess-ammonia liquor between 500 and 4,500 ppm and coking
times of 13 and 22 hours.26 Waste ammonia liquor phenol concentrations of
1,000 to 2,000 ppm are cited commonly as design values.
3-20
-------
AMMONIA TO
GAS STREAM
ii COOLING WATER
f \ DEPHLEGMATOR (PARTIAL CONDENSER)
*/t
* ~ 100° C VAPOR
EXCESS-AMMONIA
LIQUOR
WASTEWATER
FREE-
AMMONIA
STILL
FIXED-
AMMONIA
STILL
LIME
LEG
(DISSOLVER)
IF NaOH USED
NO DISSOLVER NEEDED
STEAM
Figure 3-4. Ammonia stills.9
3-21
-------
Several phenol removal/recovery techniques are practiced. The tradi-
tional process techniques are solvent extraction and steam stripping. In
both cases, the phenol-rich stream, once extracted, is treated with caustic
to make sodium phenol ate.
The solvent extraction dephenolization process generally uses light
oil or benzene to extract phenol from the excess-ammonia liquor. A flow
diagram of a solvent extraction dephenolization process is shown in Fig-
ure 3-5. The excess-ammonia liquor flows through an absorber column, which
may be a packed tower, a tray tower, a mechanically agitated column, or a
series of mixer-settlers. The solvent rate is generally 1.2 volumes of
solvent per volume of excess-ammonia liquor, although wide variations
occur.
Dephenolization generates wastewater after the tar acids are removed
from the sodium salts and springing gas. The wastewater will be saturated
with light oil, and the springing of the phenols with high-carbon dioxide
gas will tend to strip benzene from the water. These emissions are not
considered to be significant nationally with respect to other by-product
benzene sources because only a few plants are known to remove phenols with
light-oil extraction and the solubility of benzene in the water is expected
to be low.
3.2.5 Final Cooler and Naphthalene Recovery
The basic function of the final cooler is to reduce the temperature of
the coke oven gas from approximately 60° C to approximately 25° C to improve
light-oil absorption in the light-oil scrubber. As the gas is cooled, some
water and most of the naphthalene in the coke oven gas are condensed into
the cooling medium. Both must be removed from the gas to prevent problems
downstream.
Three forms of final coolers and naphthalene recovery technologies are
used in the domestic by-product industry. These forms of recovery are:
direct cooling with water—naphthalene recovery by physical separation;
direct cooling with watei—naphthalene recovery in the tar bottom of the
final cooler; and direct cooling with wash oil--naphthalene recovery in the
wash oil. Of the 55 plants listed in Table 3-3, 23 use direct-water final
coolers, 18 use tar-bottom final coolers, and 5 use a wash-oil final cooler.
3-22
-------
SOLVENT EXTRACTION
SPRINGING
EXCESS-
AMMONIA
LIQUOR
PHENOLIZED LIGHT OIL
»
DEPHENOLIZED-
AMMONIA
LIQUOR
10% CAUSTIC
LEAN ABSORBENT
WASTE GAS
TAR ACID
PRODUCT
*•
WASTE (Na2, C03,
WATER, PHENOL, ETC.)
HIGH CQ2 GAS
STEAM STRIPPING DEPHENOLIZATION (VAPOR RECIRCULATION)
EXCESS-
AMMONIA
LIQUOR
1
DEPHENOLIZED-
AMMONIA
LIQUOR
STEAM -
J t
CAUSTIC, 10%
SODIUM PHENOLATE (TO PROCESSING AS ABOVE)
Figure 3-5. Solvent extraction and steam-stripping dephenolization processes.9
3-23
-------
The circulating water absorbs hydrogen cyanide and benzene from the
coke oven gas and liberates them to the atmosphere if, as in many plants,
the same water is cooled against air in an open tower. An indirect cooler;
i.e., a large shell-and-tube exchanger for the coke oven gas, prevents
cooling tower emissions. The wash oil used to cool the coke oven gas in a
wash-oil final cooler is cooled indirectly in a heat exchanger. This
cooling eliminates naphthalene fouling of the heat exchanger surface, which
would occur if hot water from a direct-water final cooler were cooled in
the heat exchanger. Naphthalene is soluble in wash oil.
In plants that use a direct-water final cooler, cooling the coke oven
gas causes condensation of naphthalene crystals and small amounts bf liquid
hydrocarbons. This condensation occurs because at that point the system
pressure is higher9 and the temperature is often lower than in other parts
of the process. Crude naphthalene that condenses in the final cooler must
be removed periodically or it will clog tubes, vents, and meters. Removing
and processing the naphthalene for sale leads to benzene emissions, as
discussed in Subsection 3.2.5.5.
An alternative method is to introduce tar into the final cooler.
Several plants have tar-bottom final coolers in which the water, after it
has cooled the coke oven gas and entrained the condensed hydrocarbons, is
forced through a pool of tar. The tar removes most of the naphthalene from
the water and is recirculated to tar storage tanks.9 In another variation
of a tar-bottom final cooler, the water contacts tar in an external device
consisting of one or more mixing zones and as many settling zones, and
light tar can be sprayed into a lower section of the final cooler if a
decanter is provided to separate water and tar.18
These methods for dissolving the naphthalene in a hydrocarbon liquid
eliminate naphthalene processing and the benzene emission from that step.
However, these methods do not eliminate benzene from the final-cooler
cooling tower. If light tar from the primary cooler decanter is used
because the heavy tar is too viscous and has suspended solids, the light
tar already contains significant quantities of benzene. Water brought near
equilibrium with coke oven gas at about 45° C cannot be expected to give up
much benzene to a tar that was in equilibrium with the same gas at about
35° C and a slightly lower pressure. If the tar is supplied intermittently
3-24
-------
or only at a rate required to keep the naphthalene from clogging, the tar
will shortly come to equilibrium with the water and accept no more benzene.
3.2.5.1 Direct-Water Final Cooler--Physical Separation of Naphthalene.
Figure 3-6 is a flow diagram of a final cooler and recirculating water
system with naphthalene collection by physical separation. After contact-
ing the coke oven gas in the final cooler, the water is pumped through a
sealed outlet to a separation device. Naphthalene, entrained tar, and
vapor-phase gums condense in the separation device by gravity in a sump
operation or flotation unit. The emissions from naphthalene separation are
discussed in Subsections 3.2.5.4 and 3.2.5.5.
After separation of the naphthalene, the water is cooled in an atmos-
pheric cooling tower and recirculated to the final cooler. The water
contains soluble compounds such as chlorides and cyanides from the cooling
operation, as well as benzene and other hydrocarbons from the coke oven
gas. The individual draft water cooling tower transfers heat from the
water to the air by atmospheric water-spray cooling. Water cooling is
affected by the air circulation in the tower and ambient temperature. A
blowdown stream may be bled from the recirculation water to prevent buildup
of honevaporated water, chloride, and cyanide.
The final cooler may be designed as a once-through water flow unit.
However, recirculation is preferred because of resource conservation and
water pollution constraints.
3.2.5.2 Tar-Bottom Final Cooler—Naphthalene Recovery in Tar. Another
common way of handling the final-cooler water is to pass the water through
tar in the bottom of the final cooler and allow the naphthalene to dissolve
in the tar. The naphthalene is then included with the tar in any additional
refining operations.
Figure 3-7 is a flow diagram of a tar-bottom final cooler. Sufficient
water must exist above the tar bottom to force the water through the distrib-
uter and into the tar. The water then separates by gravity and is recircu-
lated. The tar can be recirculated continuously to the tar storage tanks
and may be sold as a final product or refined. The final-cooler water is
cooled in a cooling tower and recirculated to the final cooler. A blowdown
operation may be used.
3-25
-------
,05
CO
a
(0
c
O
a
$
CO
o
>
a
o
8
CO
c
CO
u
0)
{£>
CO
3
_O)
iZ
3-26
-------
en
o>
o
-------
In the tar-bottom final cooler, the water that descends from the zone
of contact with the coke oven gas carries the solid and liquid hydrocarbons
that were condensed out of the gas. The condensed hydrocarbons are col-
lected by the tar and the water that disengages from the tar is essentially
free of entrained naphthalene, although some naphthalene is dissolved
(solubility 0.003 g/100 g of water at 25° C). If the tar is furnished
batchwise to the tar-bottom cooler, it eventually becomes saturated with
naphthalene as evidenced by a "silvery" irridescence or light scattering by
the crystals. Tar in this condition cannot remove suspended naphthalene,
crystals from the cooling water and may become difficult to transfer.
Consequently, the operator usually changes the tar batch when it appears
(visually) to be saturated with naphthalene. Estimates based on the vapor
pressures of solid and liquid naphthalene suggest that the tar becomes
saturated with naphthalene when the concentration is about 30 mole percent
or roughly 15 percent by weight, about twice the usual percentage.
The tar-bottom cooler method not only eliminates naphthalene handling
and attendant benzene emissions but also has implications for benzene
emissions from the final cooler. In this design, the water that picked up
benzene when it cooled the gas and went to the atmospheric cooling tower
may lose some of its benzene to the tar. The amount of benzene the water
loses depends on the source of the tar and the tar-to-water ratio. The
primary or heavy tar that is condensed in the gas mains by quenching at
about 80° C contains very little benzene, perhaps 0.1 percent by weight.
If all of this tar (about 40 kg/Mg of coke) contacted all of the cooling
water (about 4,200 kg/Mg), which contains benzene in equilibrium with coke
oven gas, some of the benzene would separate into the tar.
However, there are operating debits. The heavy tar is viscous and is
normally stored and handled while it is warm; therefore, cooling it to
35° C in this contact may be inadvisable. Using the smaller amount of
light tar, which is richer in benzene, might solve the naphthalene problem
but probably would not affect the benzene concentration. Using the whole
tar is an intermediate case. In any event, achieving close contact between
a viscous tar and an aqueous slurry of naphthalene crystals may invite
emulsification, clogging of nozzles, or both. This process seems to work
3-28
-------
as a naphthalene-handling method but probably should not be expected to
reduce benzene emissions from the final-cooler cooling tower significantly.
3.2.5.3 Wash-Oil Final Cooler—Naphthalene Recovery in Wash Oil.
Traditionally, water has been used to cool gases in the final cooler,
but wash oil also can be used. The petroleum wash oil normally used for
the cooling medium has a boiling range of 270° to 350° C, a specific gravity
of 0.830, and a flash point of 150° C.27 Naphthalene and some light oil
will dissolve in the wash oil, and the water that condenses must be removed
in a decanter. The wash oil normally is cooled by indirect heat exchange
and recirculated to the final cooler. A slipstream of the wash oil contain-
ing naphthalene is routed to the light-oil recovery plant for removal of
both the naphthalene and light oil. A lean wash-oil makeup stream is
provided to the final-cooler recirculation tank. Figure 3-8 is a process
flow diagram of the wash-oil final cooler.
In principle, benzene emissions from naphthalene handling and from the
direct final cooler can be eliminated by the wash-oil final cooler. Because
the oil's heat capacity is about half that of water, the circulation rate must
be approximately doubled to maintain the same temperature pattern found in
direct-water fi.nal coolers. If the column is the spray type, more pump
work per pound of coolant is required to break the oil into droplets of a
suitable size distribution.28 If a packed or baffled column is chosen, the
more viscous oil runs through the column more slowly, and allowance must be
made for the increased quantity of oil in the column. Because the wash oil
removes heat from the gas, it must be cooled by cooling water (the normal
process) or possibly ambient air for much of the year.
Cooling the gas to any temperature above its dew point would not be a
problem. However, the purpose of this unit is to cool and dehumidify the
gas; the cooling required for moisture condensation is the greater part,
perhaps 80 percent, of the unit's capacity. In a direct-water-cooled
column, the heavy hydrocarbons remaining in the gas, naphthalene espe-
cially, also tend to condense. These hydrocarbons, partly solid because
naphthalene melts at 80° C and crystallizes from the condensate at lower
temperatures, form on or in the water and create a slight separation problem.
3-29
-------
1
o
oc
I
O>
o
•g
o
I
CO
9
J=
CO
CO
£
3-30
-------
Similarly, in a wash-oil final cooler, the condensing water will form
small droplets on or in the oil. The problem is analogous to that of the
hydrocarbon and naphthalene but is often more difficult to solve. In both
cases, the dispersed phases are substantially less than is the continuous
phase. In the wash-oil final cooler, separation is hindered by the vis-
cosity of the oil and possibly other factors.
3.2.5.4 Naphthalene Processing. Naphthalene collected by physical
separation is impure, has a dirty yellow-brown appearance, and contains a
high percentage of water (approximately 50 to 60 percent). The naphthalene
slurry is commonly dewatered by gravity separation. Crystallized naphtha-
lene may be refined through drying when the crystals are melted in a rectan-
gular tank equipped with coils for either cold water or steam circulation.
After 24 hours in the vessel, an upgraded naphthalene with a greater than
78° C crystallization point is generated.29 The crude naphthalene also may
be dissolved in coal tar after physical separation and sold as a commercial
feedstock.
With a direct-water final cooler, crude naphthalene is recovered from
the hot well of the direct final cooler. The naphthalene crystals are wet
with a film of mixed hydrocarbons, often of a brownish color, which suggests
that some tar fog bleeds through the electrostatic precipitator and the
ammonia saturator. This unpredictable amount of liquid hydrocarbon medium
is a solvent for benzene. At these conditions, a liquid hydrocarbon would
contain about 3 moles of benzene per 100 moles of liquid, perhaps 6 percent
by weight. The naphthalene made at this step might be 1 kg/Mg; the liquid
hydrocarbon would not be more than 2 kg/Mg to prevent the naphthalene from
dissolving and is probably less than 0.5 kg/Mg. Thus, the dissolved ben-
zene might be as much as 30 g/Mg, much of which would be evaporated during
naphthalene handling and processing. For example, the naphthalene is
conveyed some distance in open troughs, heated and dissolved in the accom-
panying hydrocarbon to disengage water, and stored while it is hot for
convenient handling.
Crude naphthalene has little market value; therefore, approximately
one-third of all plants (see Table 3-3) eliminate the nuisance by some
variant of the tar-bottom cooler. However, about 40 percent of the plants
handle naphthalene in some manner. Because more naphthalene is in the tar
3-31
-------
than is recovered at the final cooler and some of this naphthalene can be
recovered during tar refining, the tar-bottom final cooler does not elimi-
nate the production and sale of naphthalene.
3.2.5.5 Emissions from the Final Cooler and Naphthalene Processing
Units. Whether the tower is a direct-water once through, direct-water
recycle, direct water with a tar bottom, or wash-oil operation, the final-
cooler unit does not generate air emissions because it is a closed system.
However, air emissions may be emitted from the induced-draft cooling tower
used in conjunction with the direct-water and tar-bottom final coolers. In
this unit, light components such as benzene and cyanide contained in the
recirculating water will be stripped out.
Air emissions from a direct-water final cooler cooling tower were
evaluated at three by-product plants.9 1J 16 The air stream directly above
the cooling tower at the first plant contained 51.6 g of benzene per megagram
of coke produced based upon a measured concentration and an assumed gas
flow rate.9 An analysis of the cooling tower blowdown showed it also
contained 22 to 43 g of cyanide and 10 to 16 g of phenol per megagram of
coke produced.9 Cyanide was emitted into the atmosphere from this cooling
tower at a rate of 280 g/Mg of coke. Benzene emissions were measured from
the direct-water final-cooler cooling tower from a second large by-product
plant at a rate of 800 kg/day, or 292 Mg/ yr.11 This rate corresponds to a
benzene emission factor of 230 g/Mg of coke. The third plant emitted
benzene at a rate of 764 kg/day, or 280 Mg/ yr.16 This benzene emission
factor is 300 g/Mg of coke, based upon capacity. Another benzene emission
factor from a direct-water final-cooler cooling tower was estimated as 69 g
of benzene per megagram of coke produced, based on emission data provided
by a large steel company.30 This emission factor is not inconsistent with
the measured benzene emissions, although the emissions are expected to vary
to some extent from plant to plant as well as with time at the same plant.
The benzene emission factor from cooling towers for direct-water final
coolers is 270 g/Mg of coke, the average of the two emission factors identi-
fied from actual measurements of benzene concentrations and volumetric gas
flow rates.
3-32
-------
The emissions from a cooling tower for a tar-bottom final cooler were
measured at another by-product plant.17 The rate of benzene emissions was
130 kg/day, or 47 Mg/yr. The benzene emission factor, based on an assumed
capacity, was 70 g/Mg of coke. Even considering the relative size of the
plants, emissions from the cooling tower were less than those from the
direct-water final-cooler cooling tower.
The wash oil is cooled indirectly with heat exchangers; therefore,
benzene emissions are not anticipated from the cooling tower of a wash-oil
final cooler; however, a wash-oil decanter and circulation tank are associ-
ated with a wash-oil final cooler. These are potential sources of benzene
emissions similar to a wash-oil decanter and circulation tank used with a
wash-oil scrubber; therefore, potentially significant benzene emissions are
likely if these sources are not controlled. The benzene emissions from a
wash-oil decanter used for light-oil recovery were measured at a by-product
plant at a rate of 9.5 kg/ day, 3.8 g/Mg of coke, or 3 Mg/yr.16
Emissions are generated from the majority of the naphthalene separa-
tion, handling, and processing operations. Naphthalene separation, when
conducted in open air dip tanks or vented storage tanks, is a potential
emission source of benzene, naphthalene, and other aromatic hydrocarbons.
These emissions increase when the crude naphthalene is refined by drying
with steam and/or melting.
Air emissions from a flotation separation and naphthalene-refining
tank have been assessed. The separator was approximately 8 m (25 ft) long,
3 m (10 ft) wide, and 3 m (10 ft) deep. The refining tank was lined with
steam coils and had a 5-m vent stack. Despite no measurable emission flow
rate from the separation tank, a vapor emitted from the vent was found to
consist primarily of benzene, benzene homologs, aromatic hydrocarbons,
fused polycyclic hydrocarbons, and fused nonalternant polycyclic hydrocarbons.
The naphthalene emission rate from the refining tank was estimated at 1.56
kg/Mg of coke produced. The benzene emission rate was not estimated.9
Naphthalene is separated in a Denver flotation unit and processed in a
naphthalene drying tank and melt pit at a by-product plant in Pennsylvania.11
The benzene emission rate from the Denver flotation unit was 300 kg/day, or
110 Mg/yr. Benzene emissions from the naphthalene melt pit were as great
as 216 kg/day, and the emission benzene rates from the two tests at the
3-33
-------
drying tank were 17 kg/day and 0.44 kg/day. The slurry recovered from the
Denver separation is transferred to the melt pit with an initial emission
rate of 1.5 to 3 kg/h (3 to 6 Ib/h) As the liquid level in the pit rises,
the emission rate increases to approximately 5 kg/h (10 Ib/h). As the
slurry in the pit melts, emissions increase to approximately 10 kg/h (25 Ib/h),
The average emission rate is assumed to be 3 kg/h (7 Ib/h), or an emission
factor of 20 g of benzene per megagram of coke. The benzene emission
factors for the Denver flotation unit, the naphthalene melt pit, and the
naphthalene drying tank were 87, 20, and 0.12 g/Mg of coke, respectively.
The emifsions from the drying tank varied highly, depending upon the fraction
of benzene evolved in the previous step, the melt pit. The order of magni-
tude of these combined naphthalene processing emissions was consistent with
emission estimates of Subsection 3.2.5.4. The emission factor for both
naphthalene recovery and processing—107 g of benzene per megagram of
coke—was obtained when emission factors for the individual steps in the
process were summed.
Other potential emission sources from the final-cooler system are:
(1) the heated tanks used to store the naphthalene-rich and lean tar of the
tar-bottom final cooler; (2) the wash-oil collecting tank, circulation
tank, decanter, and storage tank of the wash-oil final cooler; and (3) the
storage tanks, sumps, and/or lagoon where the decanted wastewater and
blowdown are piped for separation and storage. Emissions from these poten-
tial sources were not measured, although emissions from a wash-oil decanter
that was a part of the wash-oil scrubber system were measured and are
discussed in Subsection 3.2.6.1. Most of these sources are considered
small, compared with the major benzene emission sources at by-product
plants, such as the final-cooler cooling tower and tar decanters.
3.2.6 Light-Oil Processing
Light oil is a clear yellow-brown oil composed primarily of benzene
(60 to 85 percent), toluene (6 to 17 percent), xylene (1 to 7 percent),
solvent naphtha (0.5 to 3 percent), and over 100 minor constituents that
boil between 0° to 200° C. The recovered quantity averages slightly less
than 1 percent of the coal charge. Light-oil processing at by-product
plants can consist of only light-oil recovery or light-oil recovery followed
3-34
-------
by light-oil refining. About two-thirds of the by-product plants sell
crude light oil, while the other third further refine the light oil.31
3.2.6.1 Light-Oil Recovery. Light oil is recovered from the coke
oven gas in a wash-oil scrubber. The wash oil is petroleum straw oil with
a boiling point above 200° C to allow effective separation from the light
oil. This wash oil resists degradation, has a high absorptive capacity for
light oil, has a low specific gravity (0.88 maximum) to aid in water separa-
tion, and does not react with the gas.24 The wash oil is pumped to the top
of a scrubbing tower and flows countercurrent to the coke oven gas entering
from the bottom. These towers may be either tray, packed, or gravity spray
towers that are operated as a single unit or with two or more in series.
The wash oil is kept above the coke oven gas temperature to prevent water
condensation and emu!sification problems. The wash oil is circulated at
1.5 to 2.5 £/m3 of gas and will remove approximately 95 percent of the
light oil. A variation of this process is to substitute a coal tar fraction
for the petroleum wash oil.27 -._ . •
The benzolized wash oil (wash-oil and light-oil mixture) is separated
by steam stripping. Live steam is injected into the bottom of a plate
tower and the more volatile light oil is stripped-overhead. The wash oil
is recycled to the scrubber. This process, shown in Figure 3-9, includes a
rectifier that separates the recovered light oil into two fractions: inter-
mediate and secondary. The flow scheme would not include the rectifier if
the crude light-oil fraction were the final product.
Emission sources in the light-oil recovery plant include atmospheric
vents on light-oil storage tanks, process decanters, condenser vents,
intercepting sumps, and contaminated sumps. These emission rates depend on
the operating temperature and process design parameters.
Data for one light-oil storage tank indicate the following emission
levels:9
Benzene, 17.4 g/Mg coke;
Toluene, 0.6 g/Mg coke; and
Hydrogen sulfide, 0.5 g/Mg coke.
The benzene emissions from a light-oil storage tank at another by-product
plant were measured as less than 12 kg/day, or about 25 g/Mg of coke.20
The head space concentration in this tank was 110,240 ppm, indicating a
3-35
-------
a 1°
> a
~ o _
O CO UJ
u < -i
= o *
. S S •-
IS s e
n
Jl
J]
i
i
j
A
i
< w
VI
1
1
1
1
1
J
-1
1
1
1
1
1
1
1
J
T
1
|
t
1
1
1
1
J
1 i
1
1
|
1 -i
-i
1
1
1
1
0)
c
o
1
Ȥ
£
'o
-------
potential benzene emission from working losses.21 The emissions from the
tank vent are thought to be relatively low from breathing losses.
If the head space of a storage tank containing 75 mole percent benzene
is permitted to attain equilibrium at 26° C, the vapor concentration of
benzene would be 100,000 ppm (derived from 13,330 Pa/101,308 Pa x 106 x
0.75). This estimated value of vapor concentration can be used to estimate
that the benzene emissions from,working losses are 5.8 g/Mg of coke. These
emissions are greater if the benzene-containing liquid is stored at a
higher temperature.
In the light-oil system described by Wilson and Wells,24 the coke oven
gas rises through a wash-oil scrubber, and the effluent benzolized wash oil
is preheated and stripped. The stripped vapors are partially condensed and
the uncondensed vapor passes to a light-oil rectifier where the overhead
consisting of benzene, toluene, and xylene (BTX), water vapor, and noncon-
densibles goes to a water-cooled condenser. The noncondensibles, which are
saturated with BTX at temperatures up to 35° C in the summer, are vented to
the air.
The noncondensibles cannot come from air leakage into the distillation
system because the system is under complete, positive pressure. The feed
of benzolized wash oil is not commonly stored in contact with air--the
source of noncondensibles in many distillations. The most probable source
seems to be coke oven gas dissolved in the wash oil at the scrubber.
The amount of noncondensibles to be vented can be estimated from the
solubilities and the wash-oil rate. In Subsection 3.2.5, the solubility of
coke oven gas in coal tar has been estimated to determine the amount of
noncondensibles released in tar dewatering. Assuming the wash oil is
chemically similar to tar in its ability to absorb benzene, the same estima-
tion scheme applies in this case. The solubility at 25° C and 1 atm, a
conservative estimate, is about 1 mole of gas per 1,000 moles of oil. The
mean molecular weight of the oil is assumed to be 200.
According to Wilson and Wells, a rule-of-thumb circulation rate is 1.6
to 2.5 £/stdm3 of gas; 1 Mg of coal gives 160 kg (16,000 moles) of dry
gas.24 The corresponding wash oil is about 700 £, 600 kg, or 3,000 moles.
Approximately 3 moles of gas dissolve if the gas is at atmospheric pressure
and 4 moles dissolve at 34,000 Pa (5 psig). At worst, the vent gas is
3-37
-------
saturated with benzene at 35° C where its vapor pressure is about 19,000 Pa
(140 mm Hg). Thus, the vent gas carries with it no more than 1 mole or
80 g of benzene per megagram of coal, or 110 g/Mg of coke. This benzene
emission rate is clearly greater than that existing at the tar decanter
because the amount of liquid exposed to the gas is much greater and the gas
temperature is lower and its pressure is higher.
The emissions from the light-oil condenser vent were evaluated at a
steel plant in Pennsylvania.11 The benzene emission rate was 314 kg/day,
or 115 Mg/yr. The emission factor of other light-oil condenser vents is
assumed to be 89 g/Mg of coke produced, which is not inconsistent with the
theoretical estimate presented earlier in this subsection.
The benzene that condenses in the light-oil condenser is collected in
the light-oil decanter. If the light-oil decanter is open, significant
benzene emissions can result, since the benzene concentration is high in
the decanter. The light-oil decanter can vent through the light-oil condenser
vent if it is enclosed and sealed.
Emissions from a wash-oil decanter used for light-oil recovery were
measured at a by-product plant at a rate of 9.5 kg/day, or 3 Mg/yr.16
This emission rate corresponds to an emission factor of 3.8 g of benzene
per megagram of coke. Similar emissions are expected from the wash-oil
circulation tank, which contains wash oil separated in the wash-oil decanter.
The emission factor from the wash-oil circulation tank is assumed to be
3.8 g of benzene per megagram of coke.
3.2.6.2 Light-Oil' Refining. Light-oil refining involves the use of
fractional distillation to separate the crude light oil into its various
components. Initial processing produces an intermediate light oil composed
primarily of crude heavy solvent and naphtha. The light-oil vapors are
condensed, and the forerunnings (cyclopentadiene, carbon disulfide, hydrogen
sulfide, and other components boiling below benzene) are removed by distilla-
tion in another column. The light oil must be desulfurized before sale;
this process is accomplished by a sulfuric acid wash to remove impurities,
followed by neutralization and decanting of the aqueous waste. The washed
BTX mixture is then distilled in a series of steam stills to separate the
components.27
3-38
-------
Light-oil refining onsite is often batch or semicontinuous because
this practice increases the unit's flexibility. Products include the
forerunnings, benzene of various purities, toluene and xylene, washed
solvent naphtha, and crude solvent naphtha.
Emission sources in the light-oil refining plant include the atmospheric
vents on the decanters and product storage. These emissions are likely to
include benzene and its homologs and result from working and breathing
losses of the tanks. Condenser vents are another source of emissions of
noncondensibles as well as the vapor from benzene and its homologs.
3.2.7 Wastewater Processing
Depending on the coal type and coking practice, the flow of wastewater
originating from the coke ovens and by-product plant is 100 to 200 £/Mg of
coke produced. Initially, the water is in the form of water vapor generated
from vaporizing surface moisture on the coked coal and bound water in the
coked coal. Water is also formed from the ultimate coke oven gas combustion,
which is used to underfire the battery.
Most of the water vapor is condensed into the flushing liquid. This
blowdown is the primary wastewater stream. Other sources of wastewater in
the by-product plant are:
Barometric condenser water from steam jets used to draw vacuum
on the ammonia crystal!izer,
Steam stripping waste from wash-oil and light-oil decanters,
and
Blowdown from the final cooler.
In one sense, ammonia recovery and phenol recovery from excess-ammonia
liquor are wastewater cleanup operations. However, for this document they
are treated as by-product recovery processes.
Barometric condenser water from vacuum ammonia crystal!izers is a
high-volume wastewater (1,000 £/Mg of coke). The waste can be greatly
reduced in volume through use of surface condensers rather than barometric
condensers. This step has led to an order of magnitude reduction in rate.32
No literature reference has been found to suggest that this waste can be
nearly eliminated through the use of vacuum pumps to draw the low pressure
3-39
-------
on the crystal!izer. Presumably, the service is thought to be too severe.
An attempt has been made to use recycled water in a cooling tower, but this
system had problems with corrosion and pH control.
Final-cooler blowdown is necessary to control the buildup of chlorides
in the cooling water. A recycle system is recommended to minimize the
wastewater volume. The final-cooler blowdown generally is combined with
the excess-ammonia liquor for treatment.
Wash-oil and light-oil decanters generate approximately 300 £ of
wastewater per megagram of coke produced. This waste results from steam
stripping the wash oil to recover light oil. One firm has published plans
to put its light-oil separator water into the final-cooler makeup. This
wastewater also can be blended with the excess-ammonia liquor and treated
at the wastewater facility.
Wastewater emissions are difficult to quantify. Benzene may be emitted
from wastewater by aeration or evaporation from lagoons, sewers, and ditches.
The waste steam may be combined with benzene-saturated wastewater with the
release of benzene vapors into the atmosphere. Information about these
wastewater sources is limited.
Sumps are one source of benzene emissions for which information
is limited. The wastewater contained in a sump may emit benzene that is
entrained or dissolved in the water. Benzene-containing liquids also may
be present on the surface of wastewater in various sumps. Tar is recovered
in common tar-intercepting sumps, and oil may be recovered from a light-oil
sump.
Sump is defined here as a wastewater separation device containing one
or several streams that flow into a decanter, pit, or tank. There, some,of
the organic materials may float to the top for separation and recovery.
Many potential sources of benzene-containing water could be treated in a
sump. Light oil is recovered by distillation from the wash oil and the
condensate contains water. The water may be separated from the light oil
in a process decanter and may then flow to another decanter or sump.
Because of the many conceivable combinations of process water flows
and because of the absence of a detailed industry survey of sumps, benzene
emission estimates from sumps are possibly one of the least reliable of the
3-40
-------
various sources considered in this chapter. The sumps may be deep and
narrow, or shallow and wide and may differ according to contents, degree of
enclosure, and method of venting emissions.
Often a sump is open to the atmosphere and has oil containing benzene
on the surface. Benzene diffuses into the atmosphere from a surface at a
rate depending on the thickness of the boundary layer, the diffusion coef-
ficient, and the concentration. An increased wind speed across the sump
will tend to decrease the boundary layer and increase emissions. Also, the
rate of transport is increased by temperature and benzene concentration
activity at the surface. The shape of a sump is important because its
area also can influence emissions. Partial enclosure can reduce emissions
because it increases the boundary layer of air.
Benzene that is not emitted from an open sump and that remains in the
water eventually can enter the atmosphere by evaporation during wastewater
treatment or after discharge to a receiving body of water. The common
light-oil intercepting sumps at two by-product plants emitted 41 and 56 kg
of benzene per day.19 3S These measurements correspond to benzene emission
factors of 3 and 27 g/Mg of coke, respectively. The emission factor used
for estimating emissions from light-oil-intercepting sumps is 15 g of
benzene per megagram of coke, based upon the average emission factor obtained
from the two sumps that were sampled. Measurements of the emissions from a
common tar-intercepting sump of another by-product plant indicated 45 kg of
benzene emissions per day.21 The benzene emission factor for a common
tar-intercepting sump for this plant is 95 g/Mg of coke. Each of these
sumps emitted approximately 16 Mg of benzene per year. Emissions from the
common tar-intercepting sump are the most important from these three sumps
because of the potential for emitting tar components.
3.2.8 Fugitive Emissions from Leaking Equipment Components
Leaking valves, flanges, pumps, exhausters, sampling connections,
pressure relief valves, and open-ended lines are potential sources of
fugitive benzene emissions from coke oven by-product plants. Defective
seals on valves, pumps, and other equipment can permit benzene to leak out
of the process and evaporate into the air. Personnel exposure to these
3-41
-------
types of fugitive emissions has been reduced by the use of respirators,
benzene hazard signs, and building evacuation fans; but these cannot be
considered environmental controls.
Benzene emissions from leaks can be significant when the benzene
content of the leaking liquid is high or when quantities of leaking coke
oven gas enriched with benzene are significant. Most of the benzene liquids
are found in the light-oil recovery and refining parts of the by-product
plants. The exhausters are potential sources of coke oven gas and benzene
emissions since the benzene has not been recovered from the gas at that
stage of the process.
Emission factors of volatile organic compounds (VOC's) from potentially
leaking process units were obtained from an extensive investigation of
petroleum refineries.34 A source survey was also carried out at three
by-product plants to determine whether emissions from leaking process units
at coke oven by-product plants were similar to leaking process units at
petroleum refineries.35 The valves, pump seals, and exhausters were screened
at each of these by-product plants and emissions were measured when the
leaking sources were enclosed in a Mylar® bag and an equilibrium flow of
air through the enclosure was analyzed. From the screening value distribu-
tions and the measured emission rates from most leaking sources, emissions
from the by-product plants were estimated. These results are presented in
Table 3-5. Emission factors from the petroleum refineries, also presented
in Table 3-5, are lower than are emissions at by-product plants except for
exhauster emissions, which were lower at by-product plants. The emission
factors from the petroleum refinery surveys are believed to be more repre-
sentative of leaking units because they are consistent with the by-product
data and were developed from a larger data base than were by-product source
data. Therefore, emission factors from the refinery data will be used to
estimate emissions from leaking by-product equipment. It should be noted
that the expected emissions from various by-product plants have a considerably
greater range of variability than does the difference between the emission
factors that were determined at by-product plants and at petroleum refineries.
Table 3-6 presents benzene emission factors for coke by-product plants
that were obtained from the VOC emission factors of petroleum refineries.
3-42
-------
TABLE 3-5. A COMPARISON OF EMISSIONS FROM LEAKS FROM BY-PRODUCT
PLANTS TO THOSE FROM PETROLEUM REFINERIES
Source
Nonmethane
organic
emission factor
petroleum and
refinery34
(kg/source day)
Nonmethane
organic
emission factor
by-product plants35
(kg/source day)
Benzene
emission factor
by-product plants35
(kg/source day)
Valves
(Light liquid)
Pump seals
(Light liquid)
Exhausters
0.26
2.7
1.236
0.43
5.2
0.37
0.25
4.0
0.087
TABLE 3-6. BENZENE EMISSION FACTORS DERIVED FROM
VOC EMISSION FACTORS
Percent of
sources
leaking
initially
Valves
Pumps
Exhausters
Pressure relief
devices
Sampling
connnections
Open-ended lines
11
24
35
d
d
d
VOC emission
factor
(kg/source
day)
0.26
2.7
1.2
3.9
0.36
0.055
Benzene emission factor
(kg benzene/source day)
Plant A,a
light
oil, BTX
0.18
1'9 c
0.28C
2.7
0.25
0.038
Plant B,b
refined
benzene
0.22
2.3
0.28C
3.4
0.31
0.047
a70 percent benzene in light oil.
86 percent benzene average in light oil and refined benzene.
C23.5 percent benzene in nonmethane hydrocarbon. (From Table 3-5, 0.087
0.37).
f\
This type of information would not be appropriate for relief valve over
pressure, sampling connections, and open-ended lines.
3-43
-------
Two different types of plants were assumed to estimate these emission
factors. Plant A had light-oil and BTX recovery with an average of 70 per-
cent benzene in the benzene-containing light liquids. Plant B produced
refined benzene in addition to the light oil with an average of 86 percent
benzene in the light liquids. The estimated benzene emission factor at
by-product plants was obtained by multiplying the VOC emission factor by
the fraction of benzene in the liquid. Emission factors for exhausters
were obtained by multiplying the VOC emission factor from compressors in
hydrogen service by 0.235, because this was the measured ratio of benzene
to nonmethane hydrocarbons present in the coke oven gas at the exhausters.35
The benzene emission factors from potentially leaking units in Table 3-6
can be used to estimate industry emissions. The number of units of each
type at the different by-product plants was estimated and the emission
factors for each unit were multiplied by the number of appropriate units at
the plant.36 This model plant approach is discussed in Chapter 6. The
benzene emissions from leaking process units estimated by this procedure
are a significant part of the overall emissions at coke oven by-product
recovery plants.
3.2.9 Summary of Emissions
A summary of the major benzene air emission sources is provided in
Table 3-7. The estimated emission rate for benzene is given for each
source with the annual emissions from all by-product plants.
3.3 BASELINE REGULATIONS
The States listed in Table 3-8 have rules that govern the storage of
VOC's and may be applicable to the storage of benzene and light oil. These
States generally require vapor controls on storage tanks that hold more
than 150 m3 (40,000 gal) of organics with a vapor pressure greater than
10,000 Pa (1.5 psia). The vapor control must be a pressure tank with no
vapor emissions, an edge-sealed floating roof, or a vapor recovery system.
These States regulate 21 by-product recovery plants, which produce about
42 percent of U.S. coke capacity.
Six of these States also require vapor controls on organic compound
water separators. This control is applicable to any separator that decants
a light-oil/water mixture or a benzene/water mixture. Except for California's
3-44
-------
TABLE 3-7. UNCONTROLLED BENZENE EMISSION FACTORS
FOR COKE BY-PRODUCT PLANTS
Source
Emission factor
(g benzene/Mg coke)
Industry emissions
(Mg/yr)36
Cooling tower
Direct-water
Tar-bottom
Light-oil condenser vent
Naphthalene separation
Naphthalene processing
Tar- intercept ing sump
Tar dewatering
Tar decanter
Tar storage
Light- oil sump
Light-oil storage
BTX storage
Benzene storage
Flushing-liquor circulation tank
Excess-ammonia liquor tank
Wash-oil decanter
Wash-oil circulation tank
Pump seals
Valves
Pressure-relief devices
Exhausters
Sample connections
Open-ended lines
Total (rounded)
270
70
89
87
20
95
21
77
12
15
5.8
5.8
5.8
9
9
3.8
3.8
a
a
a
a
a
a
6,340
1,090
4,080
2,040
470
5,360
1,090
4,350
680
780
300
80
80
510
510
180
180
600
400
270
30
50
20
29,000
Emissions were estimated on the basis of number of potentially leaking
units. Emission factors are listed in Table 3-6.
3-45
-------
TABLE 3-8. STATES REQUIRING VAPOR CONTROLS ON STORAGE TANKS
AND SEPARATORS
Minimum tank
size
State
California
Colorado
Kentucky
Mary 1 and
Michigan
Missouri
Pennsylvania
Wisconsin
(m3)
150
150
150
240
150
150
150
150
(gal)
40,000
40,000
40,000
65,000
40,000
40,000
40,000
40,000
Minimum vapor
pressure separators
(Pa)
10,000
10,000
10,000
10,000
10,000
12,000
10,000
10,000
(psia)
1.5
1.5
1.5
1.5
1.5
1.8
1.5
1.5
included
Yes
Yes
Yes
Yes
Yes
No
Yes
No
Minimum
separator flow
(A/day )(gal /day)
760
760
760
760
760
201
200
200
200
No .
Minimum
—
200
~"
TABLE 3-9. CALIFORNIA REGULATIONS FOR COKE OVEN BY-PRODUCT PLANTS
Rule 462 -
Rule 463 -
Rule 464 -
Rule 466
Rule 466.1 -
Required to install, maintain, and operate a vapor contain-
ment or collection system on transfer of light oil (BTX)
from storage tanks (12- to 600,000-£ or 3- to 150,000-gal)
to railroad cars.
Required to install, maintain, and operate an approved vapor
containment collection system on (12- to 600,000-2 or
3- to 150,000-gal) light-oil storage tanks.
Required to cover all wastewater separators (eight tar
decanters)
Required to install and maintain approved mechanical seals or
equivalent on all pumps or compressors handling VOC's (11,000 Pa
or 1.55 psi Reid or greater). Also inspect three times daily.
Required to inspect, record, and maintain all valves and
flanges handling VOC's (11,000 Pa or 1.55 psi Reid vapor
pressure or greater.)
3-46
-------
regulations, no State regulations were found that would apply specifically
to tar-decanter, tar-dewatering, tar storage, or cooling tower emissions.
Table 3-9 lists relevant California regulations that can reduce benzene
emissions from by-product plants.
3.3.1 Baseline Regulatory Requirements
The solid waste disposal guidelines are written with broad definitions
of "solid waste" and "disposal" so they may be interpreted to include coke
oven by-product plant emissions. For example, disposal is defined as
including the placement of liquids or solids so any component may enter the
environment, including fugitive air emissions.37 The EPA Office of Solid
Waste Management has not promulgated specific standards for by-product
plant fugitive emissions, and there is no indication that they plan to
provide specific standards.
In 1978, the Occupational Safety and Health Administration (OSHA)
promulgated an exposure limit on airborne concentrations of benzene of
1 part benzene per million parts of air, regulated dermal and eye contact
with benzene solutions, and imposed monitoring and medical testing.require-
ments on employers whose workplaces contain 0.5 ppm or more of benzene.38
The regulation originally applied to benzene emissions from any source in
the plant but was amended to exempt benzene emissions from mixtures contain-
ing less than 1 percent benzene (i.e., storage tanks). However, the regula-
tion subsequently was remanded to OSHA in 1980 because of an incomplete
administrative record, coupled with the question of the cost/benefit associ-
ated with the standard.39
By-product recovery operations currently are subject to a benzene
worker exposure limit of 10 ppm, based on an 8-hour time weighted average
for a 40-hour week. A ceiling concentration of 25 ppm, with a maximum peak
of 50 ppm (with a maximum duration of 10 minutes) for each 8-hour shift
also is permitted. Engineering or administrative (work practice) controls
could be required, if feasible, to meet the 10-ppm limit but usually are
not necessary. If controls are not feasible to achieve full compliance,
OSHA may require protective equipment or other measures.40 For example,
OSHA may require the use of a respirator for an employee repairing a leaking
pump. The current regulation applies to benzene emissions from any source
in the plant. It is anticipated that this regulation will be enforced for
3-47
-------
at least 1 to 2 years while the more stringent benzene standard undergoes
further Agency review.41 The current OSHA standard is expected to have no
influence on the baseline regulatory requirements because there are no
equipment requirements.
3.4 REFERENCES
1. Sheridan, E. T. Supply and Demand for United States Coking Coals and
Metal!urgial Coke. Bureau of Mines, U.S. Department of the Interior.
1976. ,p. 18.
2. Telecon. Murphy, M>, U.S. Department of Energy, with Lough, C.,
Midwest Research Institute. February 27, 1979.
3. Energy Information Administration, U.S. Department of Energy. Coke
and Coal Chemicals in 1979. Energy Data Report. Washington, DC.
October 31, 1980. p. 1.
4. Telecon. Sheridan, E. T., U.S. Department of Energy, with Lough, C.,
Midwest Research Institute. February 23, 1979.
5. Energy Information Administration, U.S. Department of Energy. Coke
and Coal Chemicals in 1976. Coke and Coal Chemicals, Annual. Energy
Data Report. Washington, DC. May 11, 1978. p. 5.
6. Reference 3, p. 2.
7. Reference 3, p. 6.
8. Reference 3, p. 4 and 5.
9. VanOsdell, D. W., et al. Environmental Assessment of Coke By-Product
Recovery Plants. U.S. Environmental Protection Agency. Research
Triangle Park, NC. Publication No. EPA-600/2-79-016. January 1979.
10. Baldwin, V. H., and D. W. Coy. Study to Develop Retrofit Information
and Other Data for Use in Setting Standards for Coke Oven Emissions.
EPA Contract No. 68-02-2612, Task 39. March 1978.
11. U.S. Environmental Protection Agency. Benzene Coke Oven By-Product
Plants—Emission Test Report, Bethlehem Steel Corporation, Bethlehem,
Pennsylvania. Research Triangle Park, NC. EMB Report No. 80-BYC-l.
March 1981.
12. U.S. Environmental Protection Agency. Benzene Coke Oven By-Product
Plants—Emission Test Report, Bethlehem Steel Corporation, Burns
Harbor, Indiana. Research Triangle Park, NC. EMB Report No. 80-BYC-5.
March 1981. ;;
3-48
-------
13. Baldwin, V. H. Environmental and Resource Conservation Considerations
of Steel Industry Solid Waste. U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-600/2-79-074. April
1979.
14. Roberson, R., and D. Marsland. Memorandum regarding emission testing
at Bethlehem Steel Corporation, Bethlehem, Pennsylvania. July 28, 1980.
15. Allen, C. C. Trip Report to Republic Steel Corporation, Cleveland, Ohio.
January 21, 1982, Research Triangle Institute. Research Triangle Park,
N.C. January 27, 1982.
16. U.S. Environmental Protection Agency. Benzene Coke Oven By-Product
Plants—Emission Test Report, U.S. Steel Corporation, Fairless Hills,
Pennsylvania. Research Triangle Park, NC. EMB Report No. 80-BYC-8.
March ,1981.
17. U.S. Environmental Protection Agency. Benzene Coke Oven By-Product
Plants—Emission Test Report, CF&I Steel Corporation, Pueblo, Colorado.
Research Triangle Park, NC. EMB Report No. 80-BYC-6. March 1981.
18. Halberstadt, P. Trip Report to U.S. Steel Corporation, Clair-
ton, Pennsylvania, April 30, 1979 and May 1, 1979. Research Triangle
Institute. Research Triangle Park, N.C. April 10, 1980.
19. U.S. Environmental Protection Agency. Benzene Coke Oven By-Product
Plants—Emission Test Report, United States Steel Corporation, Clairton,
Pennsylvania. Research Triangle Park, NC. EMB Report No. 80-BYC-2.
March 1981.
20. Coke Oven Emissions from By-Product Coke Oven Charging, Door Leaks,
and Topside Leaks on Wet-Coal Charged Batteries. Background Informa-
tion Document (Draft). Research Triangle Institute. Research Triangle
Park, NC. March 1981.
21. U.S. Environmental Protection Agency. Benzene Coke Oven By-Product
Recovery Plants—Emission Test Report, Wheeling-Pittsburgh Steel
Corporation, Monessen, Pennsylvania. Research Triangle Park, NC. EMB
Report No. 80-BYC-3. March 1981.
22. Carbone, W. F. Phenol Recovery from Coke Wastes. Sewage and Indus-
trial Waste. 22(2):200. 1950.
23. United States Steel Engineers and Consultants, Inc. (a subsidiary of
U.S. Steel). United States Steel Phosam Process. Bulletin 2-01.
24. Wilson, P. J., Jr., and J. H. Wells. Coal, Coke, and Coal Chemicals.
New York, McGraw-Hill, 1950.
25. T. Nicklin, et al. U.S. patent no. 3,035,889. Assigned to Clayton
Aniline Co., Ltd. United Kingdom.
3-49
-------
26. Wilks, F. Phenol Recovery from By-Product Coke Waste. Sewage and In-
dustrial Waste. 22(2):196. 1950.
27. McGannon, H. E. (ed.). The Making, Shaping, and Treating of Steel.
9th edition. Section 4. U.S. Steel Corporation. Pittsburgh, PA.
1971.
28. Perry, R. H., and C. H. Chilton. Chemical Engineer's Handbook. 5th
edition. New York, McGraw-Hill, 1973.
29. McNeil, D. The Separation and Purification of Naphthalene, Anthra-
cene, and Other Polynuclear Hydrocarbons. Coal Carbonization Products.
Great Britain, Pergammon Press, 1966.
30. Letter from Thorpe, J. S., Bethlehem Steel Corporation, to Goodwin, D. R.,
U.S. Environmental Protection Agency. September 26, 1979. Response
to Section 114 questionnaire "Current and Planned Emission Controls
for Coke Oven By-Product Recovery Plants."
31. Energy Information Administration., U.S. Department of Energy. Coke
and Coal Chemicals. Energy Data Report. Washington, DC. DOE/EIA-0122/1.
October 1978.
32. Traubert, R. M. Weirton Steel Division Brown's Island Coke Plant.
Iron and Steel Engineer. 55(1):61. 1978.
33. U.S. Environmental Protection Agency. Benzene Coke Oven By-Product
Plants—Emission Test Report, Republic Steel Corporation, Gadsden,
Alabama. Research Triangle Park, N.C. EMB Report No. 80-BYC-4.
March 1981.
34. Mesich, F. G. Results of Measurement and Characterization of Atmos-
pheric Emissions from Petroleum Refineries. In: Proceedings from
Symposium on Atmospheric Emissions from Petroleum Refineries. U.S.
Environmental Protection Agency. Publication No. EPA-600/9-80-013.
March 1980. p. 139.
35. U.S. Environmental Protection Agency. Benzene Fugitive Leaks—Leak
Frequency and Emissions Factors for Fittings in Coke Oven By-Product
Plants. Research Triangle Park, N.C. EMB Report No. 81-BYC-12.
January 1982.
36. Allen, C. C. Memorandum regarding Method of Estimating Coke Oven By-
Product Plant Industry Emissions. Research Triangle Institute.
Research Triangle Park, NC. May 28, 1981.
37. U.S. Congress. Resource Conservation and Recovery Act of 1976. Public
Law 94-580. Washington, DC. U.S. Government Printing Office.
October 21, 1976, as amended December 11, 1980.
3-50
-------
38. Permanent Standard for the Regulation of Benzene. U.S. Occupational
Safety and Health Administration. 43 Federal Register 5918, February 10,
1978, as amended, 43 Federal Register 27962, June 27, 1978.
39. U.S. Supreme Court. Slip Op. No. 78-911. Industrial Union Department,
AFL-CIO vs. American Petroleum Institute, et al. July 2, 1980.
40. U.S. Occupational Safety and Health Administration Regulation for
Benzene Exposure. 29 Code of Federal Regulations. Part 1910.1000.
Office of the Federal Register, General Services Administration.
July 1, 1980.
41. Telecon. Scott, M., Research Triangle Institute, with Martonic, J.,
Occupational Safety and Health Administration. April 7, 1981.
3-51
-------
4.0 EMISSION CONTROL TECHNOLOGY
This chapter discusses the technology that has been or could be
used to control benzene emissions from the by-product plant sources
discussed in Chapter 3. A few of these controls have been demon-
strated in by-product recovery plants; others, such as controls for
product storage tanks, are obvious candidates for technology transfer
from other industries with similarly controlled emission sources.
The major emphasis in this chapter is on emission controls that
have been demonstrated for by-product recovery sources. The emission
sources at most plants are uncontrolled, but a few plants have imple-
mented and demonstrated control techniques for selected sources. Gas
blanketing is the most widely demonstrated control technique and one
of the simplest and most effective for by-product recovery plants.
Various options exist for gas blanketing and are discussed in detail
in the following subsections. In general, the principles of gas
blanketing require sealing all of the source's openings to the atmos-
phere, supplying a constant-pressure gas blanket, and providing for
the recovery or destruction of displaced vapor emissions.
To understand the operating principles of gas blanketing, consider
the three cases of vapor flow for the schematic in Figure 4-1. The
first case is for vapor flow out of the source's vent line, from
pumping liquid into the tank, breathing losses, or from the continuous
evolution of gas dissolved in the liquid. As the pressure in the
vapor space increases above the constant pressure setpoint of the
controller, the controller opens and relieves the excess pressure by
venting the vapors to a recovery or destruction system. A second case
occurs when liquid is pumped out of the tank. Then the blanketing gas
4-1
-------
PRESSURE CONTROLLER (OPEN)
CONSTANT-
rnL-ooUnt: \
GAS SUPPLY
VAPOR SPACE — ^>
LIQUID *.
r
t
^ ,—, .., TO VAPOR
»• \/ ->• r
"^2
—L—
/_x, ; KbCUVhKY OK
VAPOR
• VENT LINE DESTRUCTION
•*"= — LIQUID LEVEL
EMISSION SOURCE
Case I. Emissions generated from pumping liquid, breathing losses,
or evolution of dissolved gases.
CONSTANT-
PRESSURE
GAS SUPPLY
VAPC
LIQUID
PRESSURE CONTROLLER (CLOSED)
c ^ ___ (NO FLOW)
J '"
IR SPACE -^_
C •«
\
1
•+?-
\
A
VENT LINE
^ LIQUID
3
LEVEL
TO VAPOR
RECOVERY OR
VAPOR
DESTRUCTION
EMISSION SOURCE
Case II. Maintaining constant pressure when pumping liquid out.
CONSTANT- (NO FLOW)
PRESSURE <:
GAS SUPPLY
NO
LIQUID
FLOW
PRESSURE CONTROLLER (CLOSED) TO VAPOR
(NO FLOW) /• RECOVERY OR
—: j VAPOR
DESTRUCTION
CONSTANT PRESSURE
EMISSION SOURCE
Case III. Static condition with no flow.
Figure 4-1. Vapor flow for a gas blanketing control system.
4-2
-------
flows through the vent line into the vapor space to maintain a constant
pressure, to relieve the partial vacuum, and to prevent the enclosed
vessel from collapsing inward. The third case represents the static
condition when the liquid level remains constant and there is no net
evolution of gas or vapor from the liquid. For this case, there is no
flow of blanketing gas or emissions, and the system remains pressurized
at the constant supply pressure.
In by-product recovery plants, gas blanketing takes advantage of
several unique characteristics of the by-product processes because the
major elements of the system shown in Figure 4-1 are already in place.
A constant-pressure gas supply is provided by raw coke oven gas in the
collecting main or clean coke oven gas in the gas holder. A pressure
controller is also in place because the As-kania regulator controls the
collecting main pressure, and gas holders have pressure controllers
that maintain a constant pressure for underfiring the battery. For
gas blanketing from the collecting main, vapor recovery systems are in
place in the form of by-product recovery processes that remove organics
from the raw coke oven gas (e.g., light-oil scrubbers). For gas
blanketing from the gas holder, a vapor destruction system is in place
because the clean coke oven gas is burned and the fuel value is recovered
when the gas is used to underfire the coke ovens. Therefore, major
requirements for gas blanketing are already in place and would not be
purchased and retrofitted. Major cost items for the gas blanketing
system in by-product recovery plants would be piping, valves, insula-
tion, and equipment modifications for leak-tight enclosure.
This chapter also discusses other controls that have been demon-
strated in by-product recovery plants or similar industries. For
example, a wash-oil scrubber is used in by-product plants to absorb
organics from gas streams in the light-oil recovery operation. Another
demonstrated control is a processing equipment change to control
emissions from cooling towers and naphthalene handling by altering the
final-cooler process. Candidates for technology transfer include
adsorption, vapor condensation, other forms of gas blanketing, other
forms of vapor destruction, alternative controls for storage tanks,
and controls for leaking equipment components.
4-3
-------
Because a control technique may be applicable for several emission
sources, for each control the applicable sources are described. For
easy reference, Table 4-1 lists each source, the applicable control
technique, and the subsection where the control is discussed.
4.1 GAS BLANKETING FROM THE COLLECTING MAIN
4.1.1 Applicable Sources
A coke oven gas blanket from the collecting main can be used to
control emissions from the tar decanter, tar-intercepting sump, tar-
dewatering tanks, tar storage tanks, flushing-liquor circulation
tanks, and weak ammonia liquor storage tanks. The emission sources
were chosen as a group because they are in close proximity to each
other. In addition, all of these vessels are associated with the
recovery of tar and ammonia liquor in the initial step of the by-
product recovery process.
The close proximity allows the use of a common large header to
supply coke oven gas to the area from the collecting main; smaller
branches of piping connect the individual vent lines to the header.
Because the liquid contents of these tanks come from water contact
with the raw coke oven gas and subsequent separation of tar and flushing
liquor, no contamination problems are expected from a raw coke oven
gas blanket. An advantage in using coke oven gas from the collecting
main for these sources is that additional organics are recovered in
the tar and light oil instead of being vented to the atmosphere.
4.1.2 Description of Technology
A gas blanket from the collecting main is provided by making a
pressure tap on the main, piping the gas to the by-product plant, and
connecting the enclosed sources to the blanketing line. Vapor emissions
from the sources would flow back into the collecting main and would be
processed with the raw coke oven gas. If liquid were removed from an
emission source, coke oven gas would fill the vapor space and maintain
a constant pressure.
Gas blanketing from the collecting main has been implemented in
the by-product recovery plant of Armco, Inc., in Houston, Texas.1 The
system at Armco was designed and installed by Koppers Company, Inc., a
4-4
-------
TABLE 4-1. EMISSION SOURCES AND CONTROL TECHNIQUES
Emission source
Tar decanter
Flushing-liquor circulation
Tar- intercepting sump
Tar storage and dewatering
Ammonia liquor storage
Light-oil plant3
Light-oil sump
Light-oil storage
Pure benzene storage
Cooling tower and naphthalene
hand! i ng
Equipment leaks
Control
technique
COG- CM
COG- CM
COG- CM
COG- CM
WOS
COG-CM
WOS
COG-GH
Enclosure
COG-GH
WOS
GB-GH
WOS
TBFC
WOFC
Varies
Subsection
4.1
4.1
4.1
4.1
4.4
4.1
4.4
4.2
4.5
4.2
4.4
4.3
4.4
4.6.1
4.6.2
4.8
COG-CM = coke oven gas blanket from the collecting main.'
WOS = wash-oil vent scrubber.
COG-GH = coke oven gas blanket from the gas holder or underfire
system.
GB-GH = nitrogen or natural gas blanket vented to the gas holder.
TBFC = tar-bottom final cooler.
WOFC = wash-oil final cooler.
^Includes the light-oil condenser and decanter, wash-oil decanter, and
circulation tank.
4-5
-------
major builder of coke batteries and by-product recovery plants. The
following discussion describes the control system's design and require-
ments in general and is based primarily on the design demonstrated at
the Houston plant. A simplified schematic of the Armco system is
provided in Figure 4-2 for reference to the general discussion.
Specific details on the Armco system are provided following the general
discussion.
An explanation of collecting main operation is needed to describe
how the control system works. Coke oven gas is generated from the
coking of coal in the ovens and is removed through a series of stand-
pipes on each oven. The standpipes are connected to a common, large
duct called the collecting main that routes the coke oven gas to the
by-product recovery plant. The pressure in the collecting main is
very carefully controlled at 5 to 10 mm (0.2 to 0.4 in.) of water
pressure by the battery operator because of the direct impact of
collecting main pressure on the back pressure in the coke ovens. Coke
plant operators have explained that pressure control in the collecting
main is inherently reliable and must be reliable for the safe operation
of the battery.1 2 3 Pressure control is provided by the Askania
regulator, and because of the importance of precise pressure-control,
a battery worker controls the pressure manually when any problems are
experienced.
Excessive pressures and pressure excursions usually are controlled
by a bleeder control valve that vents the excess pressure through a
stack. A high collecting main pressure causes the battery operator
many problems; e.g., unseating charging port lids, blowing standpipe
caps or damaging standpipes, and causing voluminous emissions from
coke oven doors. Negative collecting main pressures also are avoided
because of more serious effects. Oxygen infiltration from the oven
doors or topside can produce an explosive mixture in the collecting
main, suction main, and every coke oven gas main (and associated
process equipment) in the by-product plant. Negative collecting main
pressure also causes serious heat damage to doors, door seals, jambs,
and other parts of the battery structure. Because of the existing
4-6
-------
15 cm
Collecting
Main
1
Offtake
Main (61 cm)
15 cm
Atmospheric
Vent
15 cm
~l—eg
-#-
Flushing-Liquor Decanters
(Tar Decanters)
15 cm
Flushing-Liquor
Collecting Tank
SYMBOLS
t^[ Three-way valve
)( Steam-out connection
^ Gate valve
M Butterfly valve
Figure 4-2. Gas blanketing of tar decanters and flushing-liquor tank
from the collecting main.1
4-7
-------
emphasis on precise pressure control, the collecting main is considered
a reliable source for a gas blanketing control system.
Many design features and modifications to the emission sources
must be considered for the gas blanketing to work effectively and
safely. Each emission source must be enclosed to accept a slightly
positive pressure without leaks to the atmosphere. For most storage
tanks, enclosure would involve closing atmospheric vent lines and
connecting the tank's vent line to the gas blanketing line. For
riveted vertical tanks in poor condition, more extensive modifications
may be required. For example, the roof may need to be replaced,
welded, or sealed in some manner to avoid leakage of coke oven gas
from existing gaps where the roof contacts the perimeter of the tank
shell.
Tar decanters and tar-intercepting sumps may require more exten-
sive modifications before a gas blanket can be applied. Tar decanter
tops usually have a rectangular surface where the liquid is either
exposed to the atmosphere or partially covered with concrete slabs set
on steel support beams. For many plants, the decanter top must be
removed, a water seal and metal cover installed, and gasket material
added to provide a tight seal for the metal cover. A water seal for
the tar decanter is illustrated in Figure 4-3.4 The seal is a heavy
plate structure suspended from the roof of the decanter near the
sludge discharge chute that allows the major portion of the liquid
surface to be blanketed at a small positive pressure. The remaining
13 percent5 of the liquid surface provides clearance for the sludge
conveyor and is open to the atmosphere. In summary, the following
items are required to prepare the tar decanter for a positive-pressure
gas blanket:
Remove the existing cover,
Blank pipelines,
Clean and inspect the tank, and repair leaks;
Install the steel plate cover, water seal, steel support
beams, and gaskets;
4-8
-------
ec
LU
O
u
_1 Ul U4
-i a x
Si-
ec o
OJ
<-"
CO
O
03
TJ
cu
0)
O)
CO
I
-a
co
"to
a>
CO
CO
en
O LU
_j a.
u- a
4-9
-------
Weld; and
Add access openings and vent pipe.
The tar-intercepting sump requires the same modifications listed
for the tar decanter except for the water seal. Because no sludge
conveyor is used, the entire surface of the sump can be covered with
metal plate and sealed with gasket material.
Heat tracing and insulation are important design considerations
for this application. The vented emissions and the raw coke oven gas
contain tar and naphthalene that can condense and plug lines and
valves. Although heat tracing and insulation should prevent this
condensation and accumulation in the vent lines vent and drain connec-
tions are included in the design for steaming out lines should the
need arise.
Each vessel would be equipped with three-way lubricated plug
valves to avoid sticking because of tar deposits. Valve connections
are arranged so that in one position the tank is vented to the collecting
main and in the other position the tank is vented to the atmosphere.
This arrangement permits the blanketing line and the tank(s) to be
isolated for maintenance or visual inspections and ensures that the
tank is vented at all times. In either position, the plug valve
provides a clear opening for the passage of vapors and prevents pockets
where tar may accumulate and interfere with the opening and closing of
the valve.
4.1.3 Demonstration of Gas Blanketing from the Collecting Main
Gas blanketing from the collecting main was installed at Armco's
Houston Works between 1976 and 1977 and was operated successfully
until the coke battery shut down in 1981. The gas blanket was applied
to two tar decanters and a flushing-liquor circulation tank as shown
in Figure 4-2. The tops of the tar decanters were enclosed up to the
sludge conveyor with a 6-mm (0.25-in.) steel plate and sealed with
gasket material. Access hatches on the decanter and circulating tank
4-10
-------
also were covered and sealed. A vertical manifold of small valves was
installed to allow the operator to determine the level of tar and
flushing liquor in the tar decanter.1
The gas blanketing line was a 15-cm (6-in.) pipe connected to the
61-cm (24-in.) offtake main upstream of the Askania regulator (butterfly
control valve). Three-way valves, atmospheric vents, and steam-out
connections for line cleaning were installed; all of these lines were
steam traced and insulated. The blanketing pressure was typically
controlled at 6 mm of water with a range of 4 to 8 mm of water. No
significant operating problems were experienced with the control
system. -1
The system at the Houston Works was not extended to control
emissions from ammonia liquor or tar storage. (Armco installed a
wash-oil scrubber for these sources, as discussed in Subsection 4.4.)
However, the same gas blanket could be applied to these storage tanks
if the tanks were enclosed and connected to the gas blanket lines, as
described in the general discussion. Armco personnel indicated that
three tar-collecting tanks, which were connected to a negative-pressure
vent system (see Subsection 4.7.1), also could have been controlled by
gas blanketing from the collecting main.1
4.1.4 Control Efficiency
The benzene control efficiency of the gas blanketing system
depends upon three major factors: leakage, the efficiency of benzene
removal in the light-oil scrubbers, and the efficiency of the underfire
combustion systems. Approximately 90 to 95 percent of the benzene in
coke oven gas is removed in the light-oil recovery process,6 and 5 to
10 percent remains with the gas and is incinerated. Incineration
efficiencies up to 99 percent have been reported for control of gasoline
vapors,7 8 and similar or higher efficiencies are expected in the
combustion of coke oven gas because of higher operating temperatures
and longer residence times. Assuming a periodic inspection and main-
tenance program prevents leaks, a control efficiency in excess of
4-11
-------
99 percent would be expected for the gas blanketing system. However,
considering that a leak might develop and would require some time to
repair, a more conservative estimate of 98 percent control efficiency
is reasonable.
These control efficiency estimates apply only to emissions collected
within the gas blanketing system. Because the tar decanter would not
be covered completely (to allow sludge removal), control efficiency
for the tar decanter emissions is estimated to be 95 percent.
4.2 GAS BLANKETING WITH CLEAN COKE OVEN GAS
4.2.1 Applicable Sources
A coke oven gas blanket from the gas holder or battery underfire
system has been used to control emissions from the light-oil condenser,
decanter, and storage tank; wash-oil decanter and circulation tank;
and benzene-toluene-xylene (BTX) storage.1 3 9 These emission sources
are generally in close proximity to each other in an area called the
light-oil pi antj and all are associated with the recovery of light oil
(70 percent benzene). The close proximity allows the use of a common
large header to supply coke oven gas to the area from the gas holder;
smaller branches of piping connect the individual vent lines to the
header. No contamination problems are expected because this gas
blanketing control has been demonstrated for these sources with both
desulfurized and undesulfurized coke oven gas. Collected emissions
from all of the sources would be added back to the coke oven gas to
recover their fuel value in the gas combustion system.
4.2.2 Description of Technology
A positive-pressure blanket of clean coke oven gas is provided by
making a pressure tap at the gas holder or underfire gas supply,
piping the gas to the light-oil plant, and connecting the enclosed
sources to the blanketing line. Vapor emissions from the sources
would flow back into the clean gas system and ultimate control would
be provided by combustion of the coke oven gas.
Available data indicate that at least three by-product recovery
plants have implemented gas blanketing of emission sources in the
light-oil plant.1 3 9 One plant installed such a system as early as
4-12
-------
1954 and has since continued operation without difficulties.3 The
following discussion describes the control system and provides details
on the three demonstrated applications. A simplified schematic con-
taining the major design details is given in Figure 4-4 for reference
to the general discussion.
After by-product removal, the clean coke oven gas is used to
underfire the coke ovens and to provide a fuel source for other combus-
tion processes. A few plants have desulfurization facilities and most
do not; however, both sulfur-containing and desulfurized coke oven
gases have been demonstrated in this application. The clean coke oven
gas is maintained at a constant pressure, typically 36 to 46 cm (14 to
18 in.) of water by a gas holder. The gas holder has an existing
pressure controller, and pressure excursions are prevented by a bleeder
control valve on the gas holder. The bleeder control valve vents at
about 51 cm (20 in.) of water, and in addition, many gas holders have
a water seal that will blow at about the same or slightly higher
pressure.2 A continuous supply of blanketing gas is available because
the gas is required for underfiring the battery. Most plants have a
source of natural gas that is used to supplement or replace the coke
oven gas in the gas holder or underfire system in the event that the
supply of coke oven gas is interrupted.1 2 3
Several design features and modifications to the emission sources
must be considered for positive-pressure blanketing with clean coke
oven gas to work effectively and safely. Each emission source must be
enclosed to accept the positive gas pressure without leaks to the
atmosphere. For most vessels in the light-oil plant, enclosure includes
closing all vents to the atmosphere and connecting the vessel's vent
line to the gas blanketing line. The light-oil condenser and horizontal
tanks require few modifications to withstand a pressure of 36 to 46 cm
(14 to 18 in.) of water.1 3 9 However, old storage tanks, particularly
riveted vertical tanks in poor condition, may require extensive modifi-
cations to withstand the pressure without leakage to the atmosphere.
Because of gaps in the roofs of these tanks, extensive repairs, sealing
gaps, or replacing the roof would be required.3
4-13
-------
00
00
rgH&l X
o
in
-------
Heat tracing and insulation are recommended for all of the blanketing
lines to avoid condensation, accumulation, and plugging in the linfes.
As shown in Figure 4-4, steam-out connections are provided for line
cleaning if necessary. Three-way lubricated plug valves are installed
so the blanketing line or vessels can be isolated for maintenance,
line cleaning, or visual inspection. The valve arrangement ensures
that the emission source is vented at all times, either to the atmosphere
or to the coke oven gas main. Flame arrestors are installed in the
atmospheric vent lines to prevent flame propagation into the tank
should emissions ignite while they are vented to the atmosphere. Many
plants already use flame arrestors in this application.
Gas blanketing of vessels containing light oil or benzene reduces
the fire and explosion hazard associated with these vessels when they
are vented to the atmosphere. Currently, the vast majority of by-product
plants do not use gas blanketing and the vents on light-oil storage
tanks are open to the atmosphere. When the atmospheric vent is open,
oxygen can enter the vapor space when the tanks are emptied periodically
or when significant cooling takes place. This oxygen infiltration can
cause the vapor in the tank to be within the explosive limits of
vapor. Applying a positive-pressure blanket eliminates oxygen infiltration
and maintains the vapor space in the tank above its upper explosive
limit. Eliminating oxygen also reduces sludge formation in the tanks
and process equipment that contain wash oil and light oil. The sludge
results from the oxidation reaction between oxygen in the air and wash
oil or light oil.
4,2.3 Demonstration of Gas Blanketing with Clean Coke Oven Gas
Gas blanketing of the light-oil plant has been demonstrated at
Bethlehem Steel Corporation's Sparrows Point plant;3 Republic Steel
Corporation's Cleveland plant;9 and the Armco, Inc., plant in Houston.1
At Sparrows Point, undesulfurized coke oven gas from the gas holder is
used to blanket wash-oil decanters, circulation tanks, collecting
tanks, and wastewater storage tanks in Plants A and B. The system was
installed in Plant B in 1954, and a similar system was installed in
Plant A as part of the conversion to a wash-oil final cooler.3
4-15
-------
The main supply for the gas blanket is a 20-cm (8-in.) line
connected to the coke oven gas line exiting the wash-oil scrubbers.
The various tanks are connected with a 15-cm (6-in.) line that runs
from the 20-cm (8-in.) supply to the top of each tank. An isolation
valve is installed in each tank's vent, and steam-out connections are
provided for line cleaning. Each tank is also equipped with 5-cm
(2-in.) atmospheric vent lines and flame arresters, but these lines
are closed during normal operation. None of the gas blanketing lines
are heated or insulated. Water U-seals are placed in the 20-cm (8-in.)
line to help remove condensate and to protect the system from excessive
pressures. No safety relief valves, pressure controllers, pressure or
flow monitors, alarms, or explosive limit detectors are on the tanks.3
The Bethlehem Steel personnel indicated no problems with the gas
blanketing system and minimal maintenance requirements. The cost of
the installation was justified because it prevented oxidation, sludge
formation, and fouling of lines and equipment. The gas blanket prevents
oxygen in the air from contacting the wash oil and light oil, which
react with the oxygen to produce a sludge. When sludge formation is
avoided, there is a large savings in labor ,to clean the final cooler,
heat exchangers, and piping.3 In addition, solid waste disposal costs
are not incurred for the potentially hazardous sludge.
A gas blanketing system was installed in Republic Steel's Cleveland
plant in 1960. In Plant 1, desulfurized gas from the battery underfire
system is used to blanket the wash-oil decanters, circulation tanks,
rectifier separators, primary light-oil separators, secondary light-oil
separators, light-oil condensers, and final-cooler circulation tanks.
In Plant 2, an undesulfurized gas blanket is applied to the primary
and secondary light-oil separators, rectifier separators, and wash-oil
circulation tanks.9
The main supply line for the coke oven gas is a 15-cm (6-in.)
line with 5-cm (2-in.) lines connecting separators and 10-cm (4-in.)
lines connecting the decanters to the supply line. The gas blanketing
lines to each source are steam traced and insulated to minimize conden-
sation and fouling; in addition, four drip points are installed so any
4-16
-------
condensate could be drained from the lines.9 Other design features of
the system are similar to those described previously for the Sparrows
Point plant.
Plant personnel stated that routine maintenance on the gas blanketing
system was minimal. Routine inspections include a monthly check of
the seals in the flame arresters and quarterly inspections of piping
and other equipment. When line cleaning is necessary, a steam supply
is connected and the lines are steamed out. The purpose of the blanketing
system is to reduce sludge formation (as described for Sparrows Point),
and the system was reported to work well in performing this function.9
The Houston plant of Armco, Inc., installed gas blanketing in the
light-oil plant between 1976 and 1977 and used the system until the
coke batteries shut down in 1981. A schematic of the Armco system is
provided in Figure 4-4. A blanket of undesulfurized coke oven gas
from the gas holder was used to control emissions from the wash-oil
decanter, circulation tank, storage tank, two light-oil storage tanks,
three light-oil condensers, and two light-oil separators. Each of
these emission sources was equipped with three-way valves, flame
arresters, steam-out connections, steam tracing, and insulation as
discussed previously in the general description. No major modifications
or repairs were required to pressurize the emission sources.1 I
A 15-cm (6-in.) line from the gas holder supplied the gas blanket
to the light-oil plant. Vent connections to the supply line were
10 cm (4 in.) in diameter for the wash-oil tanks, 5 cm (2 in.) for the
light-oil storage tanks, and 8 cm (3 in.) for the light-oil condensers
and separators. The gas blanket was maintained at a pressure of 38 cm
(15 in.) of water by the gas holder. Plant personnel reported no
significant operating difficulties with the system.1
4.2.4 Control Efficiency
The benzene control efficiency of the gas blanketing system
depends upon the amount of leakage and the efficiency of combustion in
the underfiring system. The temperature and residence time of the
coke oven gas in the combustion system are expected to result in
efficiencies of 99 percent or greater. (Incineration efficiencies of
4-17
-------
99 percent have been reported for gasoline vapors.) Assuming periodic
inspection and maintenance minimize leaks, an estimated 98-percent
control efficiency for the gas blanketing system is reasonable.
4.3 NITROGEN OR NATURAL GAS BLANKETING
4.3.1 Applicable Sources
A gas blanket of nitrogen or natural gas can be used to control
emissions from pure benzene storage tanks. By-product plant operators
have claimed that coke oven gas should not be recommended as the
blanketing medium because of product quality specifications for the
pure benzene and the possibility of contamination from components
(e.g., sulfur compounds) in the coke oven gas.2 9 For pure benzene
storage tanks, emissions from breathing or working losses would be
routed to the coke oven gas main and burned in the gas combustion
system. Alternatively, emissions may be routed to the gas main before
light-oil removal and recovered in the wash-oil scrubbing operation.
4.3.2 Description of Technology
The choice of blanketing gas depends upon existing gas supplies
in the plant, proximity of the supply to the tank, and reaction or
contamination considerations between the blanket gas and the liquid in
the tank. Nitrogen or natural gas was considered for blanketing pure
benzene storage tanks'because most by-product plants have an existing
supply of one or both. For example, coke plants that are part of an
integrated steelmaking complex may have access to nitrogen from their
oxygen plant associated with steelmaking.9 Most coke plants have a
source of natural gas used to supplement the coke oven gas; to replace
the coke oven gas in emergency situations; or to underfire the coke
ovens during startup, idle, or controlled shutdown of the coke
battery.* 2 3
The major elements of a nitrogen or natural gas blanketing system
must be purchased and installed, whereas the coke oven gas blanketing
systems use many elements already in place. The gas must be purchased
or routed to the by-product recovery plant, a pressure controller
installed to control supply pressure, and another pressure controller
installed where emissions are vented to maintain the blanketing pressure.
A schematic of a gas blanketing system is given in Figure 4-5.
4-18
-------
PRESSURE
CONTROLLER
GAS
SUPPLY
FLAME
ARRESTOR
ATMOSPHERIC
VENT
LIQUID f
LINE r
•Z-
A
i
PRESSURE
CONTROLLER
r TO GAS HOLDER
"^ ORUNDERFIRE
SYSTEM
THREE-WAY VALVE
VENT LINE
BENZENE STORAGE
TANK
VAPOR SPACE
LIQUID LEVEL
Note: Dashed arrows show emission's flow.
Figure 4-5. Schematic of a nitrogen or natural gas blanketing system.
4-19
-------
The pressure controller or pressure reducer controls the supply
pressure of the gas at 38 to 46 cm (15 to 18 in.) of water. Because
displaced vapors are vented to the gas holder, which is maintained at
46 cm (18 in.) of water, another pressure controller is installed to
prevent backflow of coke oven gas and to maintain the blanket pressure.
When the pressure in the tank's vapor space increases to above 46 cm
(18 in.) of water, the pressure controller opens and vents the vapors
to the gas holder. When liquid is removed from the tank, more blanket-
ing gas is provided through the pressure controller on the gas supply
to maintain a constant pressure. Under static conditions with no
liquid or vapor flow, the system remains pressurized with no net flow
of the blanketing gas or vapor emissions.
The benzene storage tanks must be enclosed to accept a positive-
pressure gas blanket without leaking. For some storage tanks, enclosure
is accomplished when the tank's atmospheric vent line is connected to
the gas blanketing line. Modifications may be required for old riveted
storage tanks that are not currently leak tight. The extent of the
modifications will depend upon the tank's condition and may include
sealing and repairing the roof, replacing the roof, or replacing the
tank.
Heat tracing and insulation would be required for the gas blanket-
ing line from the benzene storage tank to the vapor destruction system.
Line heating would be most important for winter operations because
benzene freezes at 5.5° C (42° F). Three-way valves would be installed
on each storage tank to allow the tank to be vented at all times,
either to the control system or to the atmosphere. The ability to
vent to the atmosphere is necessary to isolate the tank from the gas
blanket, to perform maintenance or visual inspections of the inside of
the tank, and to prevent loss of the blanketing gas if the tank is
emptied or taken out of service. Flame arrestors would be installed
in the atmospheric vent lines to reduce the fire and explosion hazard
when the tank is vented to the atmosphere.
Nitrogen blanketing of benzene storage tanks has been applied at
the Aliquippa Works of J&L Steel, but displaced emissions are not
4-20
-------
controlled.10 Currently, nitrogen is used to blanket crude light-oil
storage tanks to prevent sludge formation. However, emissions are
vented to the atmosphere and are not vented to a vapor recovery or
destruction device.
The control efficiency of a nitrogen blanketing system that is
vented to a vapor recovery or destruction device depends upon the same
factors as that of a coke oven gas blanketing system: extent of
leakage and combustion efficiency. Assuming the gas blanketing lines
are well maintained with little leakage, an efficiency of 98 percent
or greater should be obtained with this control system.
4.4 WASH-OIL SCRUBBERS
4.4.1 Applicable Sources
A wash-oil scrubber can be used to control emissions from the
various storage tanks in the by-product recovery plant. The wash-oil
scrubber has been applied to weak ammonia liquor tanks, tar storage
tanks, and tar-dewatering vessels.1 Other potential applications
include light-oil storage tanks, BTX storage tanks, and pure benzene
storage tanks.
The applicability of a wash-oil scrubber as an efficient control
device to sources with heated vapors (e.g., tar-dewatering and tar
storage tanks) depends upon the temperatures of the vapors in the
scrubber. The vapors must be cooled for the scrubber to be effective,
either by a condenser or by a sufficiently high flow rate of cool
wash-oil spray.
An advantage of the wash-oil scrubber over gas blanketing is the
applicability to old storage tanks in poor condition. The pressure
drop through the wash-oil scrubber is negligible; therefore, modifi-
cations to old tanks are minimal because the tanks are not subjected
to pressures significantly higher than the normal operating conditions.
4.4.2 Description of Technology
The wash-oil scrubber would be installed on the side of the
storage tank or in a centralized location to control emissions from
several storage tanks. The emissions enter the bottom of the scrubbing
chamber and contact a spray of wash oil that is introduced into the
4-21
-------
top of the spray chamber. The wash-oil spray absorbs benzene from the
vent vapors. After passing through the scrubber, the benzolized wash
oil is routed to the light-oil recovery plant for removal of benzene
and other organics from the wash oil. The debenzolized wash oil is
then recycled to the wash-oil scrubber.
The process of absorbing benzene from a gas stream with a wash-oil
scrubber is not new to the by-product recovery industry. The coke
oven gas leaving the final cooler contains about 2 percent benzene
that is removed in a wash-oil scrubbing operation. Most by-product
recovery plants remove the light oil (primarily BTX) from the coke
oven gas by contacting the gas with liquid petroleum wash oil in a
scrubbing tower (absorber). The inlet wash oil, containing about
0.2 percent light oil, is sprayed into the top of the wash-oil scrubber
and flows through spray nozzles to contact the g'as stream. The outlet
wash oil contains 2 to 3 percent light oil and removes 90 to 95 percent
of the light oil from the coke oven gas.6
Recent designs of wash-oil scrubbers are not fitted with hurdles
or packing to accomplish gas-liquid contact. Contact is accomplished
by the use of single conical sprays placed at two or three elevations
in the tower. Restrictions to gas flow by accumulated residues commonly
found in packed scrubbers are minimized or eliminated in scrubbers of
this design.6 Wash-oil scrubbers currently used for light-oil removal
are large towers designed to handle high volumes of coke oven gas.
Applying this scrubbing operation to the vented emissions from storage
tanks results in a much smaller scale design for the scrubbing chamber
and a lower wash-oil circulation rate.
The Houston plant of Armco, Inc., used a wash-oil scrubber to
control the vented emissions from two tar storage tanks, an ammonia
liquor storage tank, and an ammonia liquor sump. A simplified schematic
of the control system for the Houston plant is given in Figure 4-6.
The system was installed between 1976 and 1977 and was operated without
difficulty until the coke battery was shut down in 1981.l
The two tar storage tanks shown in Figure 4-6 have capacities of
1.6 million £ (425,000 gal) and 280,000 & (75,000 gal). Tar is dewatered
in the larger tank by steam heating for 6 days and settling for 1 day.
4-22
-------
O £
n.
O)
«,
a
3
CO
c e e
„„„
i
1
O
CO
i
u.
O
II
s
0.
I
in
CO
O
CT
c
O
E
E
as
Hi
C
O
05
0)
.a
u
CA
'5
j=
I
(D
4
2
I
il
4-23
-------
The tar is then transferred to the smaller tank and sold locally. The
ammonia liquor storage tank has a capacity of 280,000 £ (75,000 gal),
and the ammonia liquor sump has dimensions of 3.7 m by 6.7 m (12 ft by
22 ft). These sources were enclosed when the access manways were
covered and sealed and the atmospheric vent lines were connected to
the scrubber entrance. The sump was enclosed with a 1-cm (0.375-in.)
metal cover and gasket, and access openings that were installed in the
sump cover were also sealed with gasket material.1
The scrubber is a metal chamber with a diameter of 0.3 m (1 ft)
and a length of 3.7 m (12 ft). Debenzolized wash oil is supplied to
the top of the scrubber through a 2.5-cm (1-in.) supply line at 0.1 £/s
(1.6 gal/min) through a spray nozzle. (The design and operating gas
flow rates were not available.) The scrubbed vent gases exit the
scrubber through a 20-cm (8-in.) vent line, and the wash oil is removed
from the scrubber by gravity drain through a 7.6-cm (3-in.) drain
line. The wash-oil drain runs to an enclosed sump that routes the
wash oil to the wash-oil decanter in the light-oil recovery system.
Organics are removed in the light-oil recovery system, and a slipstream
of debenzolized wash oil is recirculated to the top of the spray
scrubber.1 The debenzolized wash oil is removed from the hot side of
the wash-oil heat exchanger at about 110° C (230° F) and enters the
scrubber as hot wash oil. Plant personnel could not explain why hot
wash oil was used instead of cooled wash oil at 32° C (90° F). Hot
wash oil has a much lower solubility for benzene (boiling point =
80° C) and other volatile compounds than cool wash oil has.
The diameter of the vent lines range from 7.6 cm (3 in.) to 15 cm
(6 in.). The 7.6-cm (3-in.) vents from the ammonia liquor storage
tank and sump combine at a 10-cm (4-in.) line that enters the base of
the scrubber. Each of the tar tanks has a 15-cm (6-in.) vent line
that enters the base of the scrubber.i
4.4.3 Control Efficiency
No emission test results or estimates were available for the
control efficiency of the wash-oil scrubber at the Houston plant. The
low solubility of benzene in the hot wash oil, which is above the
4-24
-------
boiling point of benzene, indicates that this scrubber was not designed
for control of benzene emissions. The solubility of benzene in hot
wash oil at 110 to 130° C is only 5 to 10 percent of the solubility of
benzene in cool wash oil at 25 to 30° C. The hot wash oil that enters
this scrubber is near the temperature that is used to strip (remove)
benzene, toluene, and xylene from the benzol ized wash oil in the
wash-oil still. These factors lead to the conclusion that the scrubber
with hoj: wash oil would not control benzene emissions.
Many factors in the design and operation of a scrubber affect its
performance. The rate and efficiency of absorption at constant pressure
depend on (1) the chemical and physical properties of the solvent
(wash oil) and the solute (benzene or light oil), (2) the operating
temperature, (3) the contacting efficiency of the column, and (4) the
gas and liquid flow rates.
The type of scrubber and packing also affect control efficiency.
In unpacked scrubbers, the gas is in contact with droplets of wash oil
sprayed into the top of the chamber. These spray scrubbers have the
advantage of a very low pressure drop, and they do not foul by sludge
accumulation on packing or bubble trays. Demisters often are added at
the top of the spray chamber to remove liquid droplets entrained in
the countercurrent gas flow. Packing could be used in the lower part
of a spray chamber to increase the surface area available for mass
transfer and reduce the backmixing due to turbulent air currents.
Packed-bed scrubbers are more suitable for storage vessels that do not
contain tar in the gas than for dirty gases that could foul the packing.
Packed-bed scrubbers can be designed with very low pressure drops,
depending on the type of packing, the gas and liquid flow rates, and
the required benzene removal efficiency.
Two important factors influence the rate and efficiency of benzene
absorption in a spray chamber. The first factor is the amount of
benzene vapor absorbed by the wash oil at equilibrium. This quantity
can be represented by the partition factor, K, which has been expressed
in the literature as the concentration of benzene in the wash oil
divided by the concentration of benzene in the vapor at equilibrium,
where the units of concentration are the same for both phases. Parti-
4-25
-------
tion factors for benzene and xylene in wash oil are given in Table 4-2
as a function of temperature.11 The partition factor for benzene
decreases with increasing temperature; i.e., benzene is less soluble
in wash oil at higher temperatures than at lower temperatures. The
fairly high values of K shown in Table 4-2 indicate that benzene is
quite soluble in wash oil. Other light-oil components such as xylene
are more soluble than is benzene in wash oil at a given temperature;
i.e., they are more strongly partitioned (separated) from the gas into
the liquid.
The second major factor affecting control of benzene emissions is
scrubber's contacting efficiency. One measure of this efficiency is
the number of theoretical equilibrium stages provided by the scrubber.
A theoretical stage is an operation in which liquid and gas phases are
brought into contact with each other such that the two phases are in
equilibrium after the operation. A number of theoretical stages may
be required to attain a specified separation or removal efficiency.
The number of theoretical stages is thus a measure of a particular
scrubber's effectiveness for benzene removal. For example, the benzene
concentration in the vapor leaving the top of the scrubber would be in
equilibrium with the wash oil leaving the bottom of the scrubber if
the scrubber were equivalent to only one theoretical stage. For a
scrubber with a performance greater than that obtained with one theore-
tical stage, the vapor phase benzene concentration leaving the scrubber
would be lower. The number of theoretical stages in a particular
scrubber design is a function of the four factors previously listed.
Table 4-3 illustrates the percent control of benzene in a wash-oil
spray chamber using the theoretical relationship developed by Lowry.11
The parameter KL/G is the product of the partition factor (K), the
liquid rate (L), and the gas rate (G), in consistent units. Table 4-3
indicates that benzene removal efficiency increases when KL/G increases,
even with a low number of theoretical stages. Scrubber design can be
optimized through cooling the wash oil or gas (increases K), increasing
the wash oil flow rate (increases L), or modifying the design and
adding packing (increases the number of theoretical stages).
4-26
-------
TABLE 4-2. PARTITION FACTORS FOR BENZENE
AND XYLENE IN WASH OIL11
Temperature
(°C)
25
80
130
Partition factor, K
(liquid concentration/gas concentration)3
Benzene
650
114
36.4
Xylene
7,570
716
170
aSame concentration units must be used (e.g., g benzene/L wash oil
and g benzene/L gas)
TABLE 4-3. PERCENT CONTROL OF BENZENE IN A
WASH-OIL SPRAY CHAMBER11
KL
G
0.5
1.0
1.5
2.0
5.0
10.0
20.0
1
33.3
50.0
60.0
66.0
83.3
90.9
95.2
Number of theoretical
2
42.9
66.0
78.9
85.0
96.8
99.1
99.8
equil ibrium
5
49.2
83.0
95.2
98.0
100.0
100.0
100.0
stages
.10
50.0
91.0
99.4 '
99.9
100.0
100.0
100.0
K = partition factor, liquid concentration/gas concentration.
L = wash-oil flow rate, in units consistent with K and G.
G = vent gas flow rate, in units consistent with K and L.
4-27
-------
A small-scale study of variables affecting benzene absorption
from air by petroleum wash oil in a spray tower has been reported with
approximately 30 to 90 percent benzene recovery.i2 The benzene removal
was found to be a function of the gas flow rate, the liquid flow rate,
and the height of the spray chamber. In practical wash-oil spray
systems for by-product plant applications, higher recovery rates can
be obtained when scrubber design is altered. For example, the diameter
of the wash-oil droplet could be decreased from the 1.5 to 2.0 mm in
diameter in the study, the length of the scrubbing chamber can be
increased from the 1.4-m reported length, and the wash-oil flow rate
can be increased.
Engineering design calculations were performed to examine the
potential application of wash-oil scrubbers to storage tanks holding
light oil, BTX, benzene, and ammonia liquor.13 14 The calculations
were based on the following worst case assumptions: (1) maximum gas
feed rate to the scrubber of 19 £/s (40.1 ftVmin) resulting from a
maximum anticipated liquid displacement rate of 19 H/s (300 gal/min);
(2) a maximum gas phase benzene concentration of 17 percent by volume
(corresponding to storage of pure benzene liquid at 90° F); and
(3) maximum scrubber operating temperature of 90° F. Two other design
parameters assumed, not falling in the category of "worst case," were
the following: (1) the spray nozzle that distributes wash oil within
the column produces a mean droplet diameter of 1 mm; and (2) the
smallest droplet produced by the same nozzle has a diameter of 0.2 mm.
These calculations indicated that a wash-oil scrubber with an 8-in.
inner diameter, an active height of 13 ft, and a wash-oil (solvent)
feed rate of 0.5 gal/min will achieve a continuous benzene control
efficiency of at least 90 percent from these sources.
For sources with gas phase benzene concentrations of less than
17 percent and for smaller gas phase (vent system) flow rates, smaller
scrubbers with correspondingly lower wash-oil feed rates can be designed.
However, a scrubber of the design summarized above will ensure that
90 percent efficiency is achieved at design (worst-case) conditions
and that the benzene concentration in the absorber offgas stream can
be maintained at or below the design level.
4-28
-------
The previous discussion indicates that high control efficiencies
can be obtained and have been demonstrated for wash-oil scrubbers.
Based on data presented by Lowry, properly designed and operated
wash-oil scrubbers theoretically can provide benzene control effi-
ciencies of 95 percent or greater; however, the highest known control
efficiency demonstrated so far is 90 percent. Supplemental cooling
may be required to obtain a 90-percent control efficiency for sources
with heated vapors. The cooling may be supplied by indirect heat
exchange (e.g., shell in tube condenser) or by using a sufficiently
high flow rate of cool wash oil.
4-28a
-------
4.5 ENCLOSURE
Control of emissions from the light-oil sump can be accomplished
by covering the sump to reduce evaporative losses. Most sumps in
by-product plants are pits that receive liqyid streams from various
processing steps. The liquid surface for most sumps is uncovered and
completely open to the atmosphere; however, a few plants have covered
or partially covered sumps. Enclosure is accomplished by installing a
steel cover, sealing the cover, and adding access manways and a vertical
vent. In such an installation, the edge of the sump cover would rest
in a trough around the edge of the sump, and a gasket material in the
trough would prevent emissions from the edge of the sump cover.
This enclosure procedure is the same as that described in Subsec-
tion 4.4.2 for the Armco, Inc., plant's ammonia liquor sump. At this
plant, the sump was covered with a 1-cm (0.375-in.) steel cover and
gasket. Access manways were installed in the steel cover to provide
ready access for maintenance, cleaning, and visual inspection.1
The purpose of the sump cover is to protect against wind that
might blow benzene vapors out of the sump into the environment. For
example, emissions from an open light-oil sump were measured as 56 kg
of benzene per day, suggesting that the equivalent of approximately
146,000 £ per day of saturated benzene vapors are blown from the
sump.15 Enclosing the sump would limit emissions primarily to working
losses (from increasing the liquid level in the sump) and breathing
losses (from increasing the temperature of the liquid in the sump).
The control efficiency of a sump cover is difficult to determine and
depends upon many factors, such as wind speed, temperature, benzene
concentration, and liquid throughput. For sumps operated at or near a
constant liquid level, a 98-percent control efficiency is estimated
for a tight cover compared to the uncontrolled situation with wind
blowing across the exposed liquid surface.
4.6 CONTROL OF COOLING TOWER AND NAPHTHALENE-HANDLING EMISSIONS
By-product plants that recover light oil cool the coke oven gas
from 60° C to 25° C in an operation called final cooling. The purpose
of the final cooler is to lower the gas temperature before the coke
4-29
-------
oven gas enters the wash-oil scrubbers to improve absorption efficiency
and to optimize light-oil recovery. Three forms of final cooling
generally are used by the industry and, depending on the type, the
nature and quantity of benzene emissions are quite different.
Approximately 23 plants with about 43 percent of the total U.S.
coke capacity use a process called direct-water final cooling. In
this process, the coke oven gas is cooled by direct contact with
water, naphthalene and other organics condense in the water, naphtha-
lene is removed by physical separation, and the water is recycled
through a cooling tower back to the final cooler. Because some benzene
condenses and is removed with the direct-contact water, benzene emissions
result from the naphthalene/water separation and from the cooling
tower as air strips the residual benzene from the cooling water. The
direct-water final cooler produces much greater benzene emissions than
do the other two processes; for this reason, the direct-water final
cooler will represent the uncontrolled case for benzene emissions from
the cooling tower and naphthalene handling.
The demonstrated control technology for these emissions is based
on the other two major final cooling processes; i.e., the tar-bottom
final cooler and the wash-oil final cooler. These two final cooling
processes will be discussed as control alternatives for the uncontrolled
case represented by the direct-water final coolers.
4.6.1 Tar-bottom Final Cooler
The tar-bottom final cooler is used by approximately 18 by-product
recovery plants. The coke oven gas is cooled by direct contact with
water, but the water is then sent through tar in the bottom of the
final cooler. The tar removes naphthalene and some other organics
from the water, the tar and water are separated, and the water is then
cooled in a cooling tower. The tar-bottom cooler does not eliminate
benzene emissions from the cooling tower, but it does eliminate benzene
emissions from the physical separation of naphthalene and water. The
naphthalene remains with the tar and is sold, or it may be removed in
a tar-refining operation.
A plant would not need to replace the direct-water final cooler
with a tar bottom to obtain the benefits of a tar-bottom final cooler.
4-30
-------
A one-stage mixer-settler containing tar could be inserted into the
final cooling process to remove naphthalene from the direct-contact
water. At a scale of 4,000 Mg of coke per day, with a 20° C increase
in water temperature through the final cooler, approximately 4,800 Mg
of cooling water per day is required for final cooling and should
contact a roughly comparable quantity of tar. The daily production of
whole tar for this size plant is about 160 Mg, about 30 Mg of which is
light tar. Because light tar is cleaner and less viscous than whole
tar is, light tar is more desirable for use in a tar mixer-settler.
If the light tar is recirculated from the settler at a rate 100 times
the production rate, the effective tar circulation rate is 3,000 Mg/day.
The combined stream of 4,800 Mg/day of water and 3,000 Mg/day of tar
could be forced through an orifice-plate mixer and into a tar settler
or decanter. The settler should provide a residence time of 30 minutes,
with a vent back to the gas exiting the final cooler. The water will
be circulated from the settler to the cooling tower in the usual way.
A sketch of this retrofit design for a tar-bottom final cooler is
presented in Figure 4-7.
In Chapter 3, benzene emissions from the cooling tower were
estimated as 270 g/Mg coke for a direct-water final cooler and
70 g/Mg coke for a tar-bottom final cooler. A control efficiency of
74 percent is thus estimated for cooling tower emissions through the
installation of a tar mixer-settler or tar-bottom process. Naphthalene
handling and processing are eliminated; therefore, the control efficiency
is estimated as 100 percent for these emission sources.
4.6.2 Wash-oil Final Cooler
Available data indicate that five by-product recovery plants use a
wash-oil final cooler. The coke oven gas is cooled by direct contact
with cool wash oil, which also removes the naphthalene. The wash oil
is circulated through an indirect heat exchanger, cooled, and then
returned to the final cooler. A slipstream of the wash oil containing
naphthalene is routed to light-oil recovery, and a makeup stream of
lean wash oil is added back to the final cooler circulation loop.
4-31
-------
Q
LU
GO
cu
LU
O
LU
8
H
LU
LU
O
X
U
2
is
O W
O >
H V)
en
Ul
X
LU
ux
w
.2
1
.i
u.
^ CC
_J 01
§<
u 5
.
O > <
O O O
CC
H CC
££
X
CO
LU CC
CC <
U. J-
4-32
-------
The wash-oil final cooler eliminates emissions from naphthalene
handling because the naphthalene is removed in the wash oil. Benzene
emissions from the cooling tower of a direct-water or tar-bottom final
!
cooler also are eliminated. This final cooling process effectively
eliminates the benzene emissions associated with a direct-water final
cooler by cooling the wash oil with indirect (noncontact) heat exchange
and eliminating the need for a cooling tower.
A wash-oil final cooler has been retrofitted at the Sparrows
Point plant of Bethlehem Steel Corporation.3 Figure 4-8 contrasts the
process flow diagram of a direct-water and wash-oil final cooler.
Although some existing process lines could be used, conversion of a
direct-water final cooler to a wash-oil final cooler would require the
installation of new process equipment. In addition, the final cooler
probably would have to be retrofitted with new spray nozzles, pumps,
and piping.
The control efficiency of a wash-oil final cooler compared to the
uncontrolled case of a direct-water final cooler is estimated as
100 percent for emissions from both the cooling tower and naphthalene
handling. This efficiency is obtained by eliminating the cooling
tower and the physical separation of naphthalene in the final cooling
process.
4.7 ALTERNATIVE CONTROL TECHNIQUES
This section will discuss control techniques that have been
demonstrated in a few specific applications in by-product recovery
plants and others that are candidates for technology transfer from
other industries. These controls are discussed separately, and the
applicability of two controls operating in series is also discussed as
a method for improving overall control efficiency.
4.7.1 Venting to the Suction Main
The suction main is that part of the coke oven gas main between
the Askania regulator and exhausters that is maintained at a negative
pressure of -200 to -300 mm of water. The exhausters provide the
motive force for the coke oven gas by pulling the gas (negative pressure)
4-33
-------
WATER
WASH OIL
1
.t 1 , COKE OVEN GAS TO
j | LIGHT-OIL SCRUBBER
j r^
: FINAL j
! COOLER j
1 •
COKE OVEN !
GAS FROM |
AMMONIA ! j
ABSORBER
NAPTHALENE
WATER
SLURRY
1 r
NAPTHA
J"~ HANOI
1
NAPHTH
i t
COLD HOT
WELL WELL
A
> '
ATMOSPHERIC
COOLING
TOWER
WASH OIL
j ' | 1 t
1 1 INDIRECT j WASH j CIRCULATION | WASH OIL FROM
"[""-*{ EXCHANGER l"""! TANK T~ ^HT-O.L STILL
_, i r r~~~
j BLEED STREAM 1
LENE ' >•
.ING TO LIGHT-OIL STILL
CONDENSATE !
L_n I
i SfflSS "! M?!U .
i ___r°11
i
WATER TO
WASTEWATER
TREATMENT
WASH-OIL FINAL COOLER
DIRECT-WATER FINAL COOLER
COMMON UNITS
Figure 4-8. Conversion of a final cooler from water to wash oil cooling medium.
4-34
-------
through the suction main and primary coolers ^nd by pushing the gas
(positive pressure) through the by-product recovery processes downstream
of the exhausters. Emission control could be accomplished by enclosing
the source to accept a negative pressure without leakage inward and
then connecting the vent line to the suction line at the primary
coolers. Emissions would enter the by-product recovery process, and
pollutants would be removed with the by-products or incinerated with
the coke oven gas.
The Houston plant of Armco, Inc., used a negative-pressure system
to control emissions from three tar-collecting tanks. The system was
installed between 1976 and 1977 and was operated without difficulty
until the coke battery shut down in 1981. A simplified schematic of
the system is shown in Figure 4-9. Vent lines on each of the horizontal
tanks were 10 cm (4 in.) in diameter and were connected to a common
vent line that was 15 cm (6-in.) in diameter. The 15-cm common vent
was connected to the 91-cm (36-in.) suction main at the primary coolers
where the normal operating pressure was -200 to -300 mm of water.
Atmospheric vents, three-way valves, and steam-out connections were
installed at each tank, and all of the vent lines were steam traced
and insulated.1
Armco personnel indicated no problems with the negative pressure
system but expressed reservations about the potential for oxygen
infiltration and the resulting explosion hazard.1 For example, if a
significantly large leak developed in the tank or if the atmospheric
vent were inadvertently left open, air could mix with the coke oven
gas in the main. Normally the coke oven gas is maintained above its
upper explosive limit; however, if a significant quantity of air were
introduced, the coke oven gas might be between the upper and lower
explosive limits. This would result in an explosive mixture exposed
to continuously arcing tar precipitators located downstream of the
exhausters. The operator's preference would have been to blanket
these tar-collecting tanks with positive-pressure gas from the collecting
main.1
4-35
-------
-*•
15 cm
Atmospheric
Vent
Suction Main
at Primary Coolers (91 cm)
10 cm
Steam-Out
Connection
Tar-Collecting Tanks
SYMBOLS
Cf[ Three-way valve
Steam-out connection
Gate valve
Atmospheric vent
Figure 4-9. Negative-pressure system from tar-collecting tanks to suction main?
4-36
-------
Many industry commenters have expressed concern about the safety
hazard associated with negative-pressure systems. However, the use of
negative pressure on tanks is not unusual. For example, every coke
plant has a primary cooler, which is in effect a large tank, and each
primary cooler operates at a negative pressure. The concern is not
the existence of negative pressure in the tank, but rather that the
tank be designed for safe operation under negative pressures. The
choice of a positive- or negative-pressure system is probably best
made by the operator who must consider the condition and operation of
a specific vessel, the costs, and the safety aspects of each system.
The control efficiency of a negative-pressure control system is analogous
to that of positive-pressure systems, which is approximately 98 to
99 percent.
4.7.2 Vapor Condensation
Although vapor condensation is not typically used for air pollution
sources at by-product plants, benzene vapors that escape from storage
tanks and process vents conceptually can be recovered with a condenser.
It is not anticipated that many of the by-product plant benzene sources
will be controlled through vapor condensation because condensation is
only moderately effective, and supplemental systems such as carbon
adsorption would be required for the 98+ percent control achievable
with other control techniques.
Two types of condensers are shown in Figures 4-10 and 4-11.
Figure 4-10 shows a simple surface condenser, and Figure 4-11 illus-
trates a two-state condenser that can be operated at a lower temper-
ature and consequently a higher control efficiency.
Condensation occurs when the condensible's partial pressure and
vapor pressure are equal. Removal efficiencies depend on the inlet
concentrations of condensibles. When the gas is saturated with hydro-
carbons; e.g., light-oil condenser vent gas, refrigeration up to
-73° C may yield removal efficiencies up to 96 percent.7 Complete
condensation is not possible because performance is limited by the
equilibrium partial pressure of the vapor stream. Consequently,
condensers often are used in combination with other control equipment
such as incinerators, carbon adsorption units, or absorption units.
4-37
-------
HARM ORGANIC
LIQUID STREAM
CONOENSATE
RETURN
Figure 4-10. Surface condenser unit used on a tank handling
warm volatile, organic materials.16
CONDENSER
AIR
(PRECOOLER)
PRECOOLEH
DISCHARGE FROM UNIT
CONDENSER
AIR
_J L_
PHECOOLER
REFRIGERATION
UNIT
}
I
i
REFRIGERATION
SYSTEM
; j
ti
._^_^ Jt_
COOLING
-RECOVERY
SECTION
VAPOR
CONDENSED
SECTION
y^
HYDRO-
CARBON
—*• WATER
Figure 4-11. Refrigeration vapor recovery unit.17
4-38
-------
The presence of noncondensible gases in storage tanks, sumps, and
the tar decanter is a major factor affecting condenser performance
when the condenser is applied to these sources. Air or nitrogen can
blanket the condenser surface so the added thermal resistance reduces
the condensation coefficients up to 50 percent.16 Factors that often
affect condenser performance include sizing (surface area, coolant
flow rate, and temperature), variation in vapor temperature and partial
pressure, and fouling from particulate matter such as tar or a frozen
component. For example, tar and naphthalene are expected in the tar
decanter and dewatering emissions. In addition, benzene freezes at
5.5° C; therefore, the condenser must include a means for removing
frozen benzene from the condenser if high separation efficiencies are
to be obtained by very low operating temperatures.
Benzene vapor can be removed from a vapor stream at an estimated
60-percent efficiency by a surface condenser operating at 7° C, assuming
the vapor inlet and outlet are saturated with benzene. This operating
temperature prevents freezing of the benzene vapor. A two-stage
system that combines a preliminary condenser operating at 6° C, followed
by a final condenser operating at -73° C, can increase the overall
benzene control efficiency to 99 percent.18 Benzene and water vapor
are collected on the condenser fins and can be removed by reheating to
6° C. An emission level of 10 ppm benzene vapor is possible if the
inlet vapor concentration is 1,000 ppm benzene. These systems have
not been demonstrated in by-product recovery plants. Because of the
presence of both noncondensibles and readily condensible components
(e.g., tar, naphthalene) in by-product plant emissions, the control
efficiency is expected to be less than that stated above for by-product
plant sources.
4.7.3 Adsorption
Hydrocarbons in a gas or vapor may be adsorbed and retained on
the surface of a granular solid. Organic vapor recovery by adsorption
is used widely by industry, and complete turn-key adsorption systems
are available from many manufacturers. Activated carbon is useful for
4-39
-------
benzene recovery from moisture-laden by-product plant emissions because
it can adsorb organic gases and vapors when water vapor is in the gas
stream.14
Adsorbers can have fixed, moving, or fluidized beds.7 The simplest
fixed-bed adsorber is a vertical, cylindrical vessel fitted with a
perforated supporting carbon screen (see Figure 4-12). The cone-shaped
carbon bed allows more surface area for gas contact and accommodates
high gas flow rates at a lower pressure drop than does a flat-bed
adsorber. If more than one carbon bed in a single unit is used, the
beds usually are arranged as shown in Figure 4-13. For a continuous
process operation, a minimum of two fixed-bed units in parallel operation
is recommended so that one unit is adsorbing while the other is being
steam stripped of solvent and regenerated (see Figure 4-14). Moving-bed
adsorbers move the adsorbent in and out of the adsorption zone, thus
increasing the adsorbent's efficiency. However, disadvantages of this
system include wear on moving parts, attrition of the adsorbent, and
lower steam utilization efficiency that results from the shorter beds.
The fluidized bed adsorber contains a number of shallow fluidized beds
where the organic vapor fluidizes the activated adsorbent. A high
loading of the solvent into the adsorbent can be maintained in this
unit, thus reducing the steam requirements for regeneration. Desorbed
material can be vented to the gas main, collected by a condenser, or
eliminated by any of the other control techniques discussed in this
chapter.
Vacuum regeneration can be used instead of steam regeneration to
eliminate the problem of disposal of a wastewater stream created by
steam regeneration.19 In this application, the carbon bed is under a
vacuum caused by a liquid ring seal pump. Desorbed organic vapor is
condensed by indirect cooling.
When an air vapor mixture is passed over carbon, adsorption is
100 percent at the beginning, but when the retentive capacity (ratio
of the weight of the adsorbate retained to the weight of the carbon)
is reached, traces of benzene appear in the exit air. In the control
of a benzene atmospheric discharge, the adsorption cycle should stop
4-40
-------
VIPOB 10 COMHNSIH
»f /"'.'
* i '
HI
: < 1
• i 1
s 1 . t
i. I 2
[II
f 7 i
I'll
nl\
f f f \-Cll
i'' •/' '•$
•f i"'.
> I '-
i i ••
-I;
".'?4
ift
X * ;
\ ^':
o» \\A
\\\
^-cuiHoaicii
SHUL HQUSIKC
VirOR FR1E
1IB OUI
Figure 4-12. Sketch of a vertical adsorber with
two cones.16
Figure 4-13. Cross-section of an adsorber with
four beds of adsorbed carbon.16
Figure 4-14. Sketch of a two-unit, fixed-bed adsorber.16
4-41
-------
at the first break point as determined by the detection of benzene
discharge. In general, fixed-bed adsorbers are not installed to
remove organics when the vapor-laden stream contains less than 3.2 kg
of solvent per 1,000 stdm3 of gas (0.2 Ib per 1,000 dry stdft3) or
when the organic concentration is greater than 25 percent of the lower
explosive limit of the mixture.20
Carbon adsorption is not known to be used at by-product recovery
plants for control of vapor emissions. For vapor emissions from
light-oil or benzene stqrage tanks, the technology transfer should be
straightforward. Other by-product emission streams containing tar and
naphthalene may not be suitable candidates for technology transfer
because of potential fouling and regeneration difficulties.
4.7.4 Absorption
The application of a wash-oil scrubber to absorb benzene from
vented vapors was discussed in Subsection 4.4. This subsection
discusses alternative absorption systems that have not been demon-
strated in by-product recovery plants. These systems are candidates
for technology transfer and offer alternative techniques that may
achieve a result similar to the wash-oil scrubber. In general, the
factors discussed in Subsection 4.4 that affect control efficiency of
absorption are applicable here and will not be repeated.
The discussion of wash-oil scrubbers was based primarily on an
unpacked scrubber with a spray of wash oil. The gas-liquid contact in
other scrubber designs has been accomplished by several types of
equipment, including packed towers (see Figure 4-15), spray towers,
bubble cap tray towers (see Figure 4-16), jet scrubbers, and venturi
absorbers (see Figure 4-17). The majority of industrial applications
absorb gas with a packed or plate tower instead of an agitated vessel
(gas dispersed by a sparger system into a liquid-filled vessel), spray
chamber, or venturi scrubber. Collecting efficiencies depend on the
absorber type and scrubbing liquor.16
For emission streams that would not foul the packing, these
scrubber designs could provide a higher control efficiency than would
a simple, unpacked spray scrubber. Potential applications in the
by-product recovery plant include light-oil and benzene storage tanks.
4-42
-------
CO
u
CD
I
CD
Q.
s
_cu
.a
O
E
O1
CD
CD
0)
i
CD
.5?
iZ
o
in
CO
CD
O
+•"
•a
cu
X.
u
CD
Q.
3
,S"
iZ
4-43 '
-------
CLEAN GAS
Figure 4-17. Venturi absorber with cyclone-type liquid separator
(Chemical Construction Corp., New York, N.Y.).16
4-44
-------
Another control option is to combine absorption with another
control technique. The Vapor Control Company of Houston, Texas,
markets a unit that can use a combination of absorption and refriger-
ation for benzene removal. The liquid absorber is chilled and contacts
the vapor steam in a countercurrent packed scrubber. The system would
strip the benzene in a regenerator, using a heat exchanger to reduce
energy requirements. Levels of hydrocarbon vapors as low as 1,000 ppmv21
can be obtained in the exit gas, which would provide 99 percent removal
of benzene if the inlet gas contained 100,000 ppmv benzene. The only
by-product plant source where benzene concentrations of this magnitude
were measured is the light-oil condenser vent. In addition, pure
benzene storage tanks would have an equilibrium vapor pressure of
130,000 ppmv at 26° C.
4.7.5 Vapor Destruction
The discussion of gas blanketing in Subsection 4.2 included the
use of the coke oven gas combustion system for vapor destruction. If
the coke plant operator chooses not to use the existing combustion
system, an incineration device may be retrofitted for vapor destruction.
A thermal afterburner can be installed to incinerate benzene
vapors. The process exhaust system or a blower delivers the organic
vapor stream to a refractory-lined burner area. The gases are mixed
thoroughly with the burner flames and are passed through the remaining
part of the chamber where combustion is completed.7 This technology
has been demonstrated for gasoline vapors and has been tested for
benzene vapors.17 One significant advantage is that a wide range of
hydrocarbons can be destroyed; a disadvantage is that potentially
valuable compounds are not recovered. However, the fuel value of the
hydrocarbons is recovered when heat recovery is practiced.
The major factors affecting afterburner performance are residence
time to complete combustion, temperature, and vapor velocity. A
minimum residence time of 0.3 to 0.5 s is recommended with vapor veloc-
ities of 7.6 to 16.2 m/s to ensure good mixing without quenching the
flame. The required discharge temperature varies depending on the
organic compound, but it is usually between 538° to 816° C.16 If
4-45
-------
combustion is inhibited by low temperature, low residence time, or poor
mixing, carbon monoxide, aldehydes, and other products of incomplete
combustion result. Maximum efficiency occurs when all combustible
matter passes through the burner at the proper temperature.7
Properly designed and operated thermal afterburners usually
achieve organic vapor removal efficiencies in excess of 95 percent,
and efficiencies of up to 99 percent have been reported for gasoline
vapors
7 8
In general, efficiency improves with increasing operating
temperature, flame contact, and residence time. An afterburner rarely
attains 90 percent efficiency in removing combustibles below 700° C if
there is residual carbon monoxide.16 Thermal incineration of benzene
vapors at temperatures of 760°' to 816° C reportedly can limit benzene
emissions in the tail gas to as little as 10 ppmv.19
4.7.6 Vapor Balance Systems
A vapor balance system uses a variable vapor space to contain the
vapors produced in storage tanks. For example, the vents from product
storage tanks with similar products can be combined into a vapor
reservoir tank. The vapor reservoir tank can be either a lifter-roof
type or an internal diaphragm type that accumulates displaced vapors.16
When liquid is pumped into a storage tank, the displaced vapors are
collected in the vapor reservoir by increasing the vapor space in the
reservoir (i.e., the roof is lifted or the diaphragm is raised). If
the pressure limitations of the storage tank and vapor reservoir are
exceeded, the vapors are vented through a pressure relief device.
These vented emissions must be recovered or destroyed to obtain a
control efficiency analogous to gas blanketing.16
The equipment modifications, three-way valves, heat tracing and
insulation, and other requirements for gas blanketing would also be
required for a vapor balance system. The emission sources must be
enclosed to accept the slight, positive pressure .of the system. If
provision is made to handle excessive vapors that exceed the capacity
of the balance system, a control efficiency equal to that of gas
blanketing could be obtained. In a by-product plant, the excess
4-46
-------
vapors could be returned to the by-product recovery process or to the
gas combustion system. No details are available on the use of vapor
balance systems in by-product recovery plants.
4.8 CONTROLS FOR FUGITIVE EMISSIONS FROM EQUIPMENT COMPONENTS
In Chapter 3, fugitive emissions from leaking process equipment
are discussed. These equipment items include valves, pumps, exhausters,
open-ended lines, sampling systems, safety relief valves, and flanges.
Techniques for controlling emissions from these sources include leak
detection and repair programs and equipment specifications. In some
cases, the techniques for controlling these emissions in by-product
recovery plants are based on transfer of control technology as applied
to related industries, such as petroleum refineries and chemical
plants. This approach is possible because the related industries use
similar types of equipment. There may be differences between by-product
recovery plants and related industries in average line temperatures,
product composition, and other parameters. However, these differences
do not significantly influence the applicability of the techniques
used in controlling the fugitive emissions.
The major reference for the following discussions is the preliminary
draft of the background information document (BID) for volatile organic
compounds (VOC's) in the petroleum refinery industry.22 When a reference
number appears in the title of a particular subsection, the entire
discussion in that subsection is attributed to that reference.
4.8.1 Leak Detection and Repair Methods22
Leak detection and repair methods can be applied in order to
reduce fugitive emissions from by-product plant sources. Leak detec-
tion methods are used to identify equipment components that are emit-
ting significant amounts of benzene. Emissions from leaking sources
may be reduced by three general methods: repair, modification, or
replacement of the source.
4.8.1.1 Leak Detection Techniques. Various monitoring techniques
that can be used in a leak detection program include individual component
surveys, unit area (walk-through) surveys, and fixed-point monitoring
systems. These emission detection methods would yield qualitative
indications of leaks.
4-47
-------
4.8.1.1.1 Individual component survey.22 Each fugitive emission
source (pump, valve, compressor, etc.) is checked for leakage in an
individual component survey. The source may be checked for leakage by
visual, audible, olfactory, soap solution, or instrument techniques.
Visual methods are good for locating liquid leaks, especially pump
seal failures. High-pressure leaks may be detected when the escaping
vapors are audible, and leaks of odorous materials may be detected by
smell. Predominant industry practices are leak detection by visual,
audible, and olfactory methods. However, in many instances, even very
large leaks are not detected by these methods.
Applying a soap solution on equipment components is one individual
survey method. If bubbles are seen in the soap solution, a leak from
the component is indicated. The method requires that the observer
subjectively determine the rate of leakage based on formation of soap
bubbles over a specified time period. The method is not appropriate
for very hot sources, although ethylene glycol can be added to the
soap solution to extend the temperature range. This method is also
not suited for moving shafts on pumps or compressors, since the motion
of the shaft may cause entrainment of air in the soap solution and
indicate a leak when none is present. In addition, the method cannot
generally be applied to open sources such as relief valves or vents
without additional equipment.
The use of portable hydrocarbon detection instruments is the best
known individual survey method for identifying leaks of VOC's from
equipment components because it is applicable to all types of sources.
The instrument is used to sample and analyze the air in close proximity
to the potential leak surface by traversing the sampling probe tip
over the entire area where leaks may occur. This sampling traverse is
called "monitoring" in subsequent descriptions. A measure of the
hydrocarbon concentration of the sampled air is displayed in the
instrument meter.
4.8.1.1.2 Unit area survey.22 A unit area or walk-through
survey entails measuring the ambient concentration within a given
distance; e.g., one meter, of all equipment located on ground levels
4-48
-------
and other accessible levels within a processing area. These measure-
ments are performed with a portable VOC detection instrument utilizing
a strip chart recorder.
The instrument operator walks a predetermined path to assure
total available coverage of a unit on both the upwind and downwind
sides of the equipment, noting on the chart record the location in a
unit where any elevated VOC concentrations are detected. If an ele-
vated VOC concentration is recorded, the components in that area can
be screened individually to locate the specific leaking equipment.
It is estimated that 50 percent of all significant leaks in a
unit are detected by the walk-through survey, provided that there are
only a few pieces of leaking equipment, thus reducing the VOC back-
ground concentration sufficiently to allow for reliable detection.
The major advantages of the unit area survey are that leaks from
accessible leak sources near the ground can be located quickly and
that the leak detection manpower requirements can be lower than those
for the individual component survey. Some of the shortcomings of this
method are that VOC emissions from adjacent units can cause false leak
indications; high or intermittent winds (local meteorological condi-
tions) can increase dispersion of VOC, causing leaks to be undetected;
elevated equipment leaks may not be detected; and additional effort is
necessary to locate the specific leaking equipment, i.e., individual
checks in areas where high concentrations are found.
4.8.1.1.3 Fixed-point monitors.22 This method consists of
placing several automatic hydrocarbon sampling and analysis instru-
ments at various locations in the process unit. The instruments may
sample the ambient air intermittently or continuously. Elevated
hydrocarbon concentrations indicate a leaking component. As in the
walk-through method, an individual component survey is required to
identify the specific leaking component in the area. Leaks from
adjacent units and meteorological conditions may affect the results
obtained. The efficiency of this method is not well established, but
it has been estimated that 33 percent of the number of leaks identi-
fied by a complete individual component survey could be located by
4-49
-------
fixed-point monitors. These leaks would be detected sooner by fixed-
point monitors than by use of portable monitors, because the fixed-
point monitors operate semi-continuously. Fixed-point monitors are
more expensive, multiple units may be required, and the portable
instrument is also required to locate the specific leaking component.
Calibration and maintenance costs may be higher. Fixed-point monitors
have been used to detect emissions of hazardous or toxic substances
(such as vinylchloride) as well as potentially explosive conditions.
Fixed-point monitors have an advantage in these cases, since a partic-
ular compound can be selected as the sampling criterion.
4.8.1.1.4 Visual inspections.22 Visual inspections can be
performed for any of the leak detection techniques discussed above to
detect evidence of liquid leakage from plant equipment. When such
evidence is observed, the operator can use a portable VOC detection
instrument to measure the VOC concentration of the source. In a
specific application, visual inspections can be used to detect the
failure of the outer seal of a pump's dual mechanical seal system.
Observation of liquid leaking along the shaft indicates an outer seal
failure and signals the need for seal repair.
4.8.1.2 Repair Techniques.22 When leaks are located by the leak
detection methods described in this subsection, the leaking component
can then be repaired or replaced. Many components can be serviced
on-line. This is generally regarded as routine maintenance to keep
operating equipment functioning properly. Equipment failure, as
indicated by a leak not eliminated by servicing, requires isolation of
the faulty equipment for either repair or replacement.
4.8.1.2.1 Pumps. Most critical service process pumps a,re backed
up with a spare so that they can be isolated for repair. Of those
pumps that are not backed up with spares, some can be corrected by
on-line repairs (e.g., tightening the packing). However, most leaks
that need correction require that the pump be removed from service for
seal repair.
4.8.1.2.2 Valves. Most valve leaks can be reduced on-line by
tightening the packing gland for valves with packed seals or by lubrication
4-50
-------
for plug valves, for example. Various valve maintenance programs have
been performed by EPA and refinery personnel. Union Oil Company and
Shell Oil Company each conducted studies at their California refineries
on maintenance of leaking valves. Emission rates were estimated based
on screening value correlations. EPA studied the effects of maintenance
on fugitive emissions from valves at four refineries. Each valve was
sampled to determine emission rates before and after maintenance to
evaluate emission reductions. In a separate study, EPA examined
maintenance effectiveness on block valves at an ethylene production
unit based on screening valves alone. In a subsequent study, routine
on-line maintenance achieved a 70-percent reduction in mass emissions.
In each of these studies, maintenance consisted of routine proce-
dures, such as adjusting the packing gland while the valve was in
service. In general, the programs concluded that (1) a reduction in
emissions may be obtained by performing maintenance on valves with
screening values above 10,000 ppmv; (2) for valves with screening
valves (before maintenance) below 10,000 ppmv, a slight reduction in
emissions after maintenance may result; moreover, emissions from these
valves may increase; and (3) directed maintenance (emissions measured
during repair until VOC concentration drops to a specified level) is
preferable to undirected maintenance (no measurement of the effect of
repair).
Valves that need to be repacked or replaced to reduce leakage
must be isolated from the process. While control valves can usually
be isolated, block valves, which are used to isolate or bypass equipment,
normally cannot be isolated. One refiner estimates that 10 percent of
the block valves can be isolated.
When leaking valves can be corrected on-line, repair servicing
can be done immediately after detection of the leak. When the leaks
can be corrected only by a total or partial shutdown, the temporary
emissions could be larger than the continuous emissions that would
result from not shutting down the unit until it was time for a shutdown
for other reasons. Simple maintenance procedures, such as packing
gland tightening and grease injection, can be applied to reduce emissions
4-51
-------
from leaking valves until a shutdown is scheduled. Leaks that cannot
be repaired on-line can be repaired by drilling into the valve housing
and injecting a sealing compound. This practice is growing in acceptance,
especially for safety concerns.
4.8.1.2.3 Flanges. One refinery field study noted that most
flange leaks could be sealed effectively on-line by simply tightening
the flange bolts. For a flange leak that requires off-line gasket
seal replacement, a total or partial shutdown of the unit would probably
be required because most flanges cannot be isolated.
For many of these cases, temporary flange repair methods can be
used. Unless a leak is major and cannot be temporarily corrected, the
temporary emission from shutting down a unit would probably be larger
than the continuous emissions that would result from not shutting down
the unit until time for a shutdown for other reasons.
4.8.1.2.4 Relief valves. In general, relief valves that leak
must be removed in order to repair the leak. In some cases of improper
reseating, manual release of the valve may improve the seat seal. In
order to remove the relief valve without shutting down the process, it
is necessary to install a block valve on a three-way valve upstream of
the relief valve if the relief valve system is to be isolated and
repaired on-line without shutting down the unit.
Flares can also be used as a means of handling emergency releases
from various processes within the plant. According to the current
knowledge of flare design, the best available flare design or state-of-
the-art flare design is the smokeless flare. Smoking flares are
environmentally less desirable because they emit particulates.
There are a number of techniques currently in use which help
flares achieve smokeless operation. One technique involves the use of
staged elevated flare systems, where a small diameter flare is operated
in tandem with a large diameter flare. ; The system is designed such
that the small flare takes the continuous low flow releases and the
larger flare accepts emergency releases. A second technique involves
the use of a small, separate conveyance line to the flare tip in order
to maintain a high exit velocity for the continuous low pressure gas
4-52
-------
flow. A third technique, sometimes used in conjunction with either of
the above techniques, involves the use of continuous flare gas recovery.
In the third technique, a compressor is used to recover the continuously
generated flare gas "base load." The compressor is sized to handle
the "base load," and any excess gas is flared. These techniques can
be used to help provide smokeless operation of a flare which is used
to reduce fugitive emissions of VOC (including benzene) that are
captured and transported by closed vent systems.
In recent tests, smokeless steam-assisted flares, smokeless
air-assisted flares, and smokeless flares with no assist were found to
be as efficient as enclosed combustion devices in destroying VOC over
a broad range of operating conditions if the heat content of the
flared gas is maintained above a certain minimum, and the velocity of
the gas at the flare tip is maintained below a certain maximum. Based
on the test data and a comparison of vent stream characteristics
between the test data and equipment leaking VOC, EPA believes that the
destruction efficiency of smokeless flares would be at least 98 percent.
Enclosed combustion devices can be designed and operated to
achieve VOC (including benzene) emission reductions of at least 98 percent.
Vapor recovery systems can be readily designed and operated to achieve
VOC (including benzene) emission reductions of at least 95 percent.
Existing enclosed combustion devices and vapor recovery systems may
not achieve the VOC emission reduction efficiencies that new control
devices achieve. However, existing control devices achieve a VOC
reduction efficiency of at least 95 percent.
An emission reduction efficiency of 95 percent is considered
appropriate for control devices used to reduce equipment leaks of VOC,
including benzene. The use of enclosed combustion devices and flares
achieving a 98 percent emission reduction is too costly to add to a
source solely to control VOC leaks in light of the presence of existing
control devices that can achieve 95 percent control. Because flares
with no assist, steam, or air assist can achieve at least 98 percent
VOC (including benzene) reduction efficiency if designed for smokeless
operation and existing control devices, such as enclosed combustion
4-53
-------
devices and vapor recovery systems, will achieve at least 95 percent
VOC (including benzene) reduction efficiency, a VOC reduction efficiency
of 95 percent is appropriate.
Recommended design and operation requirements for flares include
smokeless operation and the presence of a flame. The presence of a
flame can be ensured by monitoring the flare's pilot light with a
thermocouple or some other heat sensor connected to an alarm. Smokeless
operation of the flare can be ensured through visible emission require-
ments. Many plants currently comply with State limits similar to this
requirement. In addition, only steam-assisted flares, air-assisted
flares, or flares with no assist could be used. Steam-assisted flares
would have to be operated with exit velocities less than 18 m/sec
(60 ft/sec), under standard conditions, combusting gases with heating
values of 11.2 MJ/scm (300 Btu/scf) or greater. Air-assisted flares
would have to be operated with heating values of 11.2 MJ/scm (300 Btu/scf)
or greater and with exit velocities equal to, or less than, the actual
velocity. The actual velocity would be calculated by dividing the gas
flow (in standard units), by the unobstructed (free) cross section
area of the flare tip. Flares operated without assist would have to
be operated with exit velocities less than 18 m/sec (60 ft/sec), under
standard conditions, combusting gases with heating values of 7.4 MJ/scm
(200 Btu/ scf) or greater. For enclosed combustion devices that do
not use catalysts to aid in combustion of organic vapor streams,
provisions for a minimum vapor residence time of 0.75 seconds at a
minimum temperature of 816° C are considered equivalent to at least a
95 percent emission reduction efficiency.
4.8.1.2.5 Exhausters. Leaks from exhauster seals may be reduced
by the same repair procedure that was described for pumps (i.e.,
tightening the packing). Other types of seals, such as a labyrinth
seal, may require that the exhauster be taken out of service for
repair. Coke plants have spare exhauster capacity because of the
importance of continuous exhauster operation to the safe and efficient
operation of both the coke battery and the by-product recovery plant.
The spare exhauster capacity could be used while the leaking exhauster
is repaired without shutdown of the gas removal system.
4-54
-------
4.8.1.3 Emission Control Effectiveness of Leak Detection and
Repair.22 The control efficiency achieved by a leak detection and
repair program is dependent on several factors, including the leak
definition, inspection interval, and the allowable repair time.
4.8.1.3.1 Definition of a leak. The first step in developing a
monitoring plan for fugitive VOC emissions is to define an instrument
meter reading that is indicative of an equipment leak. The choice of
the meter reading for defining a leak is influenced by several consid-
erations. The percent of total mass emissions that can potentially be
controlled by the leak detection and repair program can be affected by
varying the leak definition. Table 4-4 gives the percent of total
mass emissions affected at various leak definitions for a number of
component types. From the table, it can be seen that, in general, a
low leak definition results in larger potential emission reductions.
Other considerations are more source specific. For valves, the
selection of an active level for defining a leak is a tradeoff between
the desire to locate all significant leaks and to ensure that emission
reductions are possible through maintenance. Although test data show
that some valves with meter readings less than 10,000 ppm have signif-
icant emissions rates, most of the major emitters have meter readings
greater than 10,000 ppm. Maintenance programs on valves have shown
that emission reductions are possible through on-line repair for
essentially all valves with nonzero meter readings. There are, how-
ever, cases where on-line repair attempts result in an increased
emission rate. The increased emissions from such a source could be
greater than the emission reduction if maintenance is attempted on low
leak valves. These valves should, however, be able to achieve essen-
tially 100 percent emission reduction through off-line repair. Gener-
ally, the emission rates from valves with meter readings greater than
or equal to 10,000 ppm are significant enough so that an overall
emission reduction is likely for a leak detection and repair program
with a 10,000-ppm leak definition. Therefore, 10,000 ppm seems to be
the most reasonable leak definition to direct maintenance effort at
the bulk of the valve emissions while still having confidence that an
overall emission reduction will result.
4-55
-------
For pump and exhauster seals, the rationale for selection of an
action level is different because the cause of leakage is different.
As opposed to valves, which generally have zero leakage, most seals
leak to a certain extent while operating normally. These seals would
tend to have low instrument meter readings* With time, however, as
the seal begins to wear, the concentration and emission rate are
likely to increase. At any time, catastrophic seal failure can occur
with a large increase in the instrument meter reading and emission
rate. As shown in Table 4-4, over 90 percent of the emissions from
compressor seals and pump seals are from sources with instrument meter
readings greater than or equal to 10,000 ppm. Because properly designed,
installed, and operated seals should have low instrument meter readings
and because the bulk of the pump and exhauster seal emissions are from
seals that have worn out or failed such that they have a concentration
equal to or greater than 10,000 ppm, this level was chosen as a reasonable
action level.
4.8.1.3.2 Inspection interval. The length of time between
inspections should depend on the expected occurrence and recurrence of
leaks after a piece of equipment has been checked and/or repaired.
This interval can be related to the type of equipment and service
conditions, and different intervals can be specified for different
pieces of equipment. Monitoring may be scheduled on an annual, quar-
terly, monthly, or weekly basis. The choice of the interval affects
the emission reduction achievable because more frequent inspection
intervals will result in earlier detection and repair of leaking
sources.
4.8.1.3.3 Allowable repair time. If a leak is detected, the
equipment should be repaired within a certain time period. The allow-
able repair time should allow the plant operator sufficient time to
obtain necessary repair parts and maintain some degree of flexibility
in overall plant maintenance scheduling. The determination of this
allowable repair time will affect emission reductions by influencing
the length of time that leaking sources are allowed to continue to
emit.
4-56
-------
4.8.1.3.4 Estimation of reduction efficiency for valves and
pumps.22 2S A mathematical model was developed to approximate the
behavior of fugitive emissions from equipment. The leak detection and
repair (LDAR) model can be used to evaluate programs requiring leak
detection and repair of leaking sources at regular intervals (1 month,
3 months, 6 months, 9 months, or 1 year). The model also includes an
option to evaluate a program requiring quarterly inspection of all
valves, attempted repair of leaking valves, reinspection of repaired
valves monthly until they are determined not to be leaking for two
successive months, and repair of leaking valves including those that
could not be repaired within 15 days during a process turnaround. In
addition, the model allows a variable input for repair effectiveness,
process unit turnaround frequency, leak occurrence, and leak frequency.
The model can also incorporate the uncertainty of the inputs and
calculate approximate confidence intervals. A description of the
methodology and data used to develop the LDAR model can be found in
Reference 23.
For leaks in by-product recovery plants, the emission factors and
percent of initial leakers shown in Table 3-6 were used as inputs to
the LDAR model. The overall emission reduction of the leak detection
and repair program for various monitoring; intervals was estimated with
the LDAR model and is shown in Table 4-5.
4.8.1.3.5 Estimation of reduction efficiency for safety relief
devices and exhausters.22 The estimated reduction efficiencies for
safety relief devices and exhausters are given in Table 4-6 and are
based on a leak definition of 10,000 ppmv. The first column in
Table 4-6 represents the percentage of total mass emissions that can
be expected from these sources with concentrations at the source
greater than 10,000 ppmv. If a leak detection and repair program
resulted in repair of all such sources to 0 ppmv, elimination of all
sources over 10,000 ppmv between inspections, and instantaneous "repair
of those sources found at each inspection, then emissions could be
expected to be reduced by the amount represented by the first column
in Table 4-6 (see Item A). However, because these conditions are not
4-57
-------
o
i - i
o
H-4
1 —
S
CM
to
V^ 0
31
-J ex.
LU Q
0
I— 1 LU
O — 1
=3
LU
0 _
to ,
f
4-1
C
i
r—
01
IB
3
cr
01
01
u
s-
Zl
o
in ro
CM ro
, r» co
r-. en
CM ci
U3 r-.
CM CO
rH i-i
CM CM
U>
01 in
> CL
•— E
re 3
> O-
-0 LU <
C llMUl "^
*S t~~* _i_
* LU x:
01 O LU
5 '*£ Q
£ LU <=C
m — 1
Q. tO
0) U_ LU
0 >
01 _l
•re 0 >
£ O LU
re ^ i — i
>> Q -J
re LLJ LU
t
*°
10 = 2Lu
u>.£ tO U_
«.•*-• to «sc
§ £ l-< to
en «r- SI
« "S LU S
•a O
j^ ' —
"3 S LU s:
"§- «g
El LUO
§ a "^s
O I ^~
.~J O . fl.
0
§ °. to cc:
t- o 1 t— i
u- i-n «-t- O.
c
•»- c
c
S- 0
55
u u
re 01
U. Q.
u)
C
,>>
JC
-u
o
01
<0
o-
•o
s-
o
o
ro
u
o
s~
0
o
re
u-
"ro
i- i.
O 01
c
CQ -i-
E_ C
O O
U 4->
re cj
1 1 QJ
o.
c:
2*
JZ
o
s:
_>>
Ol
S-
re
3
O*
re
^
0
0
re
O)
u
s-
3
O
l/>
CM «3-
LT) l£>
«f ir>
r-. r-.
^ "^
CTl CTl
o o
C3 O
co co
o' o
in tn
en en
o o
0 0
en en
o ci
r^- 01
o o
01
••—
Ol S-
s- in a)
>, > in
-t-> i — 3
01 ttj re
t- > .rz
re x
10 UJ
c
- 01 o -a
t/) w) C -r- 01
01 C •!- 4-3 I-
O *r~ ^ U -r-
s- ±£. re ai re
3 re ci -*J o.
O 01 r— Ol 01
i- a) o c: 01
O J3 W 01 S-
**- 01 re
2"c S
iz re 01 re
O C •- J3
> U 3 -— > !-
B u o at 3 in
CL re t- u u 01
a. tr o o
T3 Ol 01 O S-
O C S- i- 3
o re ro s- 4J o
O 3 re in
• "O 4-> U -C
o 01 re a> 4-1 at
rH E •= «- S
All 3 4-> '*-' U) O
in cm
tn in o
o u ••- in re TH
•r- 01 S- 4-> V) • JZ
03 fC O O) E E
&- m o. fll *i~ 4-> • LU
4-» in in 4-> u
c 01 s- c s- re x
01 3 O •— O t- V-
U i— M— I*- «r- o
c re •*-* re ai
o > •- x tno.^ x
O *--• 01 4-3 01 4J
01 C E !_ 0
0 • U 3 S.
ot-izoi o>»o x
4J s- 4-> cj -a m
JZ tj i. < i in
4-> re 3 ai m 4-> x
•i- If- U t. i-H C
3 CJ 0 .3 «C
C O 4- £- re O
in o ^ ai o o u
u-Pm uaij
in t- -p cn"a c in t- ro ai
oureai o -r-
E U 01 ^ -i- 4-> CJ
0 CL Ji -i- 4-> CJ t. .1-
£-oimrereij-re o **-
c -r- i — ai t_ -i- a>
m ai t- s- re c ^
CS-CO OCLO Ut—
o t-ai-poi uoi'i— re o
•r-3 X3 S-4J.OI-
inu24-> s- uo£-4^
in oi4->i_4->»r-4-> aii- CLC
•r- s-oireo rec j-ai CLO
E ^)4->C CLOl (-M re U
01 T3 in Ol 3 O *-^
c-ii -a £-cruoo
i— re ro "o i — 01 -uoc
-*-> oil — reo 3-o *i— TJ <; -r-
•*•> co- aimcLCJC£cj
01-P-O4-) C OI3re re"O S-T3O"O
O S-4->i-J^:4->C
-------
met in practice, the fraction of emissions from sources with concen-
trations over 10,000 ppmv represents the theoretical maximum reduction
efficiency. The approach to estimation of emission reduction presented
here is to reduce this theoretical maximum control efficiency by
accounting quantitatively for those factors outlined above.
This approach can be expressed mathematically by the following
equation:24
Reduction efficiency =AxBxCxD,
where:
A = Theoretical maximum control efficiency = fraction of total
mass emissions from sources with concentrations greater
than 10,000 ppmv.
B = Leak occurrence and recurrence correction factor = correc~
tion factor to account for sources that start to leak
between inspections (occurrence); for sources that are
found to be leaking, are repaired, and start to leak again
before the next inspection (recurrence); and for known
leaks that could not be repaired.
C = Noninstantaneous repair correction factor = correction
factor to account for emissions that occur between detec-
tion of a leak and subsequent repair, since repair is not
instantaneous.
D = Imperfect repair correction factor = correction factor to
account for the fact that some sources that are repaired
are not reduced to zero. For computational purposes, all
sources that are repaired are assumed to be reduced to an
emission level equivalent to a concentration of 1,000 ppmv.
An implicit assumption here is that the leak detection program detects
all of the sources with concentrations greater than 10,000 ppmv that
are present at the time of the inspection. As an example of this
technique, Table 4-6 gives values for the "B," "C," and "D" correction
factors for various possible inspection intervals and allowable repair
times.
Recent results of the LDAR model indicate that the ABCQ approach
slightly overestimates the emission reduction achieved by the inspection
program. The emission reduction for valves in gas service was estimated
4-59
-------
using both approaches and revealed a ratio of the LDAR/ABCD emission
reduction of 0.69 for quarterly monitoring and 0.77 for monthly monitoring.
To put all of the emission reductions on approximately the same basis
(i.e., LDAR model), the percent reductions for safety relief valves
and exhausters in Table 4-6 were adjusted by the LDAR/ABCD ratio,
which is listed as Factor E in the table. For safety relief valves,
the resulting emission reductions are 44 and 52 percent for quarterly
and monthly monitoring, respectively. For exhausters, the corrected
emission reductions are 55 and 64 percent for quarterly and monthly
monitoring, respectively.
4.8.2 Preventive Programs22
An alternative approach to controlling fugitive emissions from
by-product plant operations is to replace components with leakless
equipment. This approach is referred to as a preventive program.
This subsection will discuss the kinds of equipment that could be
applied in such a program and the advantages and disadvantages of this
equipment.
4.8.2.1 Pumps. As discussed in Chapter 3, pumps can be potential
fugitive emission sources because of leakage through the drive-shaft
sealing mechanism. This kind of leakage can be reduced to a negligible
level through the installation of improved shaft sealing mechanisms,
such as dual mechanical seals, or it can be eliminated entirely by
installing seal!ess pumps. Another control option is to enclose the
seal area, collect the emissions, and transport the emissions to a
control device or return them to the process.
4.8.2.1.1 Dual mechanical seals. As discussed in Chapter 3,
dual mechanical seals consist of two mechanical sealing elements
usually arranged in either a back-to-back or a tandem configuration.
In both configurations a nonpolluting barrier fluid circulates between
the seals. The barrier fluid system may be a circulating system, or
it may rely on convection to circulate fluid within the system. While
the barrier fluid's main function is to keep the pumped fluid away
from the environment, it can serve other functions as well. A barrier
fluid can provide temperature control in the stuffing box. It can
4-60
-------
also protect the pump seals from the atmosphere, as in the case of
pumping easily oxidizable materials that form abrasive oxides or
polymers upon exposure to air. A wide variety of fluids can be used
as barrier fluids. Some of the more common ones that have been used
are water (or steam), glycols, methanol, oil, and heat transfer fluid.
In cases in which product contamination cannot be tolerated, it may
also be possible to use a clean product, a product additive, or a
product diluent.
Emissions from barrier fluid degassing vents can be controlled by
a closed-vent system, which consists of piping, and, if necessary,
flow-inducing devices to transport the degassing emissions to a control
device, such as a process heater, or vapor recovery system. Control
effectiveness of a dual mechanical seal and closed-vent system is
dependent on the effectiveness of the control device used and the
frequency of seal failure. Failure of both the inner and outer seals
can result in relatively large emissions at the seal area of the pump.
Pressure monitoring of the barrier fluid may be used in order to
detect failure of the seals. In addition, visual inspection of the
seal area also can be effective for detecting failure of the outer
seals.
An alternative to venting the barrier fluid to a control device
is to operate the barrier fluid system such that the barrier fluid
pressure is greater than the stuffing box pressure. For dual mechan-
ical seals in a back-to-back arrangement, the higher pressure of the
barrier fluid will result in some leakage of the barrier fluid across
the inboard face of the seal into the stuffing box and subsequently
into the pumped liquid. The pressure of the barrier fluid prevents
outward leakage from the process stream and any leakage will be from
the barrier fluid into the process stream. Barrier fluid going across
the outboard face of the seal will exit to the atmosphere. Therefore,
the barrier fluid must be compatible with the process liquid as well
as with the environment. This control option is not suitable for dual
mechanical seals in a tandem arrangement. In the tandem arrangement,
the barrier fluid is at a pressure lower than that in the stuffing
4-61
-------
box; therefore, any leakage from the stuffing box will be into the
barrier fluid. Control of emissions from the barrier fluid's reservoir
for seals in the tandem arrangement must provide for the collection of
the emissions and transport to a control device.
Another control option for pumps is to purge the barrier fluid to
an appropriate by-product recovery process. The barrier fluid may be
circulated through the seal and transported to an appropriate point in
the process for removal or destruction of any benzene in the barrier
fluid. Alternatively, the barrier fluid may be recirculated through a
closed system with removal of a slipstream of the barrier fluid to the
process to prevent accumulation of benzene in the fluid. For either
case, clean barrier fluid must be added to the system on a continuous
basis to replace any barrier fluid that is removed.
Dual mechanical seals are used in many by-product plant process
applications; however, there are some conditions that preclude the use
of dual mechanical seals. Their maximum service temperature is usually
limited to less than 260° C, and mechanical seals cannot be used on
pumps with reciprocating shaft motion.
4.8.2.1.2 Seal less pumps. The seal!ess or canned-motor pump is
designed so that the pump casing and rotor housing are interconnected.
The impeller, motor rotor, and bearings are completely enclosed and
all seals are eliminated. A small portion of process fluid is pumped
through the bearings and rotor to provide lubrication and cooling.
Standard single-stage canned-motor pumps are available for flows
up to 160 m3 per second and heads up to 76 m. Two-stage units are
also available for heads up to 183 m. Canned-motor pumps are widely
used in applications where leakage is a problem.
The main design limitation of these pumps is that only clean
process fluids may be pumped without excessive bearing wear. Since
the process liquid is the bearing lubricant, abrasive solids cannot be
tolerated. Also, there is no potential for retrofitting mechanical or
packed seal pumps for seal!ess operation. Use of these pumps in
existing plants would require that existing pumps be replaced.
4-62
-------
4.8.2.2 Exhausters. Exhausters can be potential fugitive emis-
sion sources because of leakage through the drive-shaft sealing mech-
anism. This kind of leakage can be reduced to a negligible level
through the use of improved shaft-sealing mechanisms, which are analo-
gous to those described for pumps.
Many exhausters have mechanical seals called a labyrinth seal,
which may also incorporate a barrier fluid. Control options for this
type of system are similar to those described in the previous subsection.
For example, emissions from the barrier fluid's reservoir may be piped
to a control device or back to the process. The barrier fluid system
may be operated at a higher pressure than the stuffing box pressure so
that any leakage would be the inward leakage of the barrier fluid.
Alternatively, emissions from the reservoir vent may be added back to
the process stream. For example, a closed loop from the reservoir
vent to the exhauster inlet may be installed to add the emissions back
to the coke oven gas.
4.8.2.3 Valves. As in the case of pumps, valves can be sources
of fugitive emissions because of leakage through the packing used to
isolate process fluids from the atmosphere (see Chapter 3). This
source of emissions, however, can be eliminated by isolating the valve
stem from the process fluid. Sealed-bellows valves are designed to
perform in this manner. The stem in a sealed-bellows valve is isolated
from the process fluid by metal bellows. The bellows is generally
welded to the bonnet and dish of the valve, thereby isolating the
stem.
There are two main disadvantages to these valves. First, they
are only available in globe and gate valve configurations. Second,
the crevices of the bellows may be subject to corrosion under severe
conditions if the bel-lows alloy is not carefully selected.
The main advantage of these valves is that they can be designed
to withstand high temperatures and pressures so that leak-free service
can be provided at operating conditions beyond the limits of diaphragm
valves.
4-63
-------
4.8.2.4 Safety/Relief Valves. A rupture disk can be used up-
stream of a safety/relief valve so that under normal conditions it
seals the system tightly but will break when its set pressure is
exceeded, at which time the safety/relief valve will relieve the
pressure. The rupture disk installation is arranged to prevent disk
fragments from lodging in the valve and preventing the valve from
being reseated if the disk ruptures. It is important that no pressure
be allowed to build in the pocket between the disk and the safety/relief
valve; otherwise, the disk will not function properly. A pressure
gauge and bleed valve can be used to prevent pressure buildup. With
the use of a pressure gauge, it can be determined whether the disk is
properly sealirig the system against leaks. It is also necessary to
install a block valve or a three-way valve upstream of the rupture
disk if the disk/relief valve system is to be isolated and repaired
on-line without shutting down the unit.
Use of a rupture disk upstream of a safety/relief valve would
eliminate leaks due to improper seating of the relief valve. Also,
the disk can extend the life of a safety/relief valve by protecting it
against system materials that could be corrosive and thereby cause
seal degradation.
Another control option would be to install o-rings in the pressure
relief device to improve the sealing mechanism. 'The o-rings could
provide a tighter seat for the metal disk and could alleviate poor
seating caused by corrosion or deposits on the metal-to-metal seal.
No data are available to estimate the control effectiveness of instal-
ling o-ring seals.
A closed-vent system can also be used to collect and dispose of
emissions from the relieving or leaking of safety/relief valves. The
vent on the relief valve could be connected to a control device or to
an appropriate point in the process to recover or to destroy the
vented emissions. The efficiency of a closed-vent system would be
determined by the control efficiency of the control device that is
used to destroy or recover the emissions.
4-64
-------
4.8.2.5 Open-Ended Lines.22 Caps, plugs, and double block and
bleed valves are devices for closing off open-ended lines. When
installed downstream of an open-ended line, they are effective in
preventing leaks through the seat of the valve from reaching atmosphere.
In the double block and bleed system, it is important that the upstream
valve be closed first. Otherwise, product will remain in the line
between the valves, and expansion of this product can cause leakage
through the valve stem seals.
The control efficiency will depend on such factors as frequency
of valve use, valve seat leakage, and material that may be trapped in
the pocket between the valve and cap or plug and lost on removal of
the cap or plug. Annual emissions from a leaking open-ended valve are
approximately 100 kg.23 Assuming that open-ended lines are used an
average of 10 times per year, that 0.1 kg of trapped organic material
is released when the valve is used, and that all of the trapped organics
released are emitted to atmosphere, the annual emissions from closed
off open-ended lines would be 1 kg. This would be a 99 percent reduc-
tion in emissions. Due to the conservative nature of these assumptions,
a 100 percent control efficiency has been to estimate the emission
reductions of closing off open-ended lines.
4.8.2.6 Closed-Purge Sampling.22 Emissions from purging sampling
lines can be controlled by a closed-purge sampling system, which is
designed so that the purged material is returned to the system or sent
to a closed disposal system and so that the handling losses are mini-
mized. An example of a closed-purge sampling system is one where the
purged material is flushed from a point of higher pressure to one of
lower pressure in the system and where sample-line dead space is
minimized. Other sampling systems are available that.utilize partially
evacuated sampling containers and require no line pressure drop, and
nonextractive sampling is possible.
Reduction of emissions by applying closed-purge sampling is
dependent on many highly variable factors, such as frequency of sampling
and amount of purge required before the closed-purge system is applied.
For emission calculations, it has been assumed that closed-purge
4-65
-------
sampling systems will provide 100 percent control efficiency for the
sample purge from uncontrolled sampling systems.
4.9 REFERENCES
1. Branscome, M. R. Trip Report to Armco, Incorporated, Houston,
Texas. Research Triangle Institute. Research Triangle Park, NC.
March 4, 1982.
2. Branscome, M. R. Trip Report to U.S. Steel Corporation Fairless
Hills, Pennsylvania. Research Triangle Institute. Research
Triangle Park, NC. March 8, 1982.
3- Allen, C. C. Trip Report to Bethlehem Steel Corporation, Sparrows
Point, Maryland. Research Triangle Institute. Research Triangle
Park, NC. January 20, 1982.
4. Jablin, R. A., et al. Cost to Control Emission of Benzene from
Coke Oven By-Product Plants. R. Jablin & Associates. EPA Contract
No. 68-02-3056. February 13, 1979.
5.
6.
7.
8.
9.
10.
11.
VanOsdell, D. W., et al. Environmental Assessment of Coke By-Product
Recovery Plants. U.S. Environmental Protection Agency. Washington,
DC. Publication No. EPA-600/2-79-016. January 1979. 387 p.
McGannon, H. E. (ed.). The Making, Shaping, and Treating of
Steel. U.S. Steel Corporation. Pittsburgh, PA. 1971. p. 174-175.
Control Techniques for Volatile Organic Emissions from Stationary
Sources. Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency. Research Triangle Park, NC.
Publication No. EPA-450/2-78-022. May 1978.
Organic Chemical Manufacturing: Combustion Control Devices.
Volume 4. U.S. Environmental Protection Agency. Research Triangle
Park, NC. Publication No. EPA-450/3-80-026. December 1980.
p. II-3 to 11-10.
Allen, C. C. Trip Report to Republic Steel Corporation, Cleveland,
Ohio. Research Triangle Institute. Research Triangle Park, NC.
January 21, 1982.
Letter from Lucas, A. W., J&L Steel Corporation, to D. R. Goodwin,
U.S. Environmental Protection Agency. August 17, 1979. Response
tq Section 114 questionnaire "Current and Planned Emission Controls
for Coke Oven By-Product Recovery Plants."
Lowry, H. H. (ed.). Chemistry of Coal Utilization.
Volume. New York, John Wiley and Sons, Inc., 1963.
Supplementary
4-66
-------
12. Hixon, A. W., and C. E. Scott. Absorption of Gases in Spray
Towers. Ind. and Eng. Chem. 27(3):307-314. 1935
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
Wood, J. P., and J. J. Spivey. Methodology for Spray Absorber
Design and Performance Assessment: Benzene Removal from Light
Oil Storage Tank Surge Vent Gases. Research Triangle Institute.
Research Triangle Park, NC. April 27, 1983.
Memorandum from Wood, J. P., Research Triangle Institute, to D. W.
Coy, Research Triangle Institute. November 14, 1983. 6 pp.
Wash-Oil Scrubbers—Guidelines for Equipment Specifications.
Benzene Coke Oven By-Product Plants—Emission Test Report, Republic
Steel Corporation, Gadsden, Alabama. U.S. Environmental Protection
Agency. Research Triangle Park, NC. EMB Report No. 80-BYC-4.
March 1981.
Air Pollution Engineering Manual. Danielson, John A. (ed.).
Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency.1 Research Triangle Park, NC. Publication
No. AP-40. May 1973.
Standard Support and Environmental Impact Statement for Control
of Benzene from the Gasoline Marketing Industry (Draft). Office
of Air Quality Planning and Standards, U.S. Environmental Protection
Agency. Research Triangle Park, NC. June 21, 1978.
Control of Refinery Vacuum Producing Systems, Wastewater Separators,
and Process Unit Turnarounds. U.S. Environmental Protection
Agency. Research Triangle Park, NC. Publication No. EPA-450/2-
77-025. October 1977.
Duravent, S. W., D. Gee, and W. M. Talber. Evaluation of Control
Technology for Benzene Transfer Operations. Office of Air Quality
Planning and Standards, U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-450/3-78-018.
April 1978.
Control Techniques for Hydrocarbon and Organic Solvent Emissions
from Stationary Sources. National Air Pollution Control Adminis-
tration, U.S. Department of Health, Education, and Welfare.
Washington, D.C. Publication No. AP-68. March 1970.
Letter from Schkade, Otto, Vapor Control Company, to C. Allen,
Research Triangle Institute. April 7, 1982. Enclosing product
information on vapor recovery systems.
VOC Fugitive Emissions in Petroleum Refining Industry-Background
Information for Proposed Standards. U.S. Environmental Protection
Agency. Research Triangle Park, NC. Publication No. EPA-450/3-81-
015a. November 1982.
4-67
-------
23. Fugitive Emission Sources of Organic Compounds—Additional Infor-
mation on Emissions, Emission Reductions, and Costs. U.S. Environ-
mental Protection Agency. Research Triangle Park, NC. Publication
No. EPA-450/3-82-010. April 1982.
24. Tichenor, B. A., K. C. Hustvedt, and R. C. Weber. Controlling
Petroleum Refinery Fugitive Emissions Via Leak Detection and
Repair. Symposium on Atmospheric Emissions from Petroleum Refineries.
Austin, TX. Publication No. EPA-600/9-80-013. November 6, 1979.
4-68
-------
-------
5. MODIFICATIONS
5.1 BACKGROUND
This chapter identifies and discusses possible modifications to sources
in coke by-product plants. The purpose of this chapter is to present what
changes are potential modifications, not to define what changes would be
judged as a modification. Determination of a modification is made by the
Administrator.
"Modification"' is defined in 40 CRF Part 61, Section 61.02, as:
"any physical change in, or change in the method of operation of,
a stationary source which increases the amount of any hazardous air
pollutant emitted by such source or which results in the emission of
any hazardous air pollutant not previously emitted, except that:
(1) Routine maintenance, repair, and replacement shall not be con-
sidered physical changes, and
(2) The following shall not be considered a change in the method of
operation:
(i) An increase in the production rate, if such increase does not
exceed the operating design capacity of the stationary source;
(ii) An increase in hours of operation.1
The owner or operator of any source must notify EPA of changes that
could increase emissions of an air pollutant for which a NESHAP applies.2
Such changes are not considered modifications if the owner or operator
demonstrates that no increase in applicable emissions would result from the
alteration, in which case, the existing source would not have to meet the
emission standards for a new source.
5.2 PROCESS MODIFICATIONS
The by-product coke industry is a mature industry with a mode of
operation that has been developed by over 50 years of experience. A new
by-product recovery process probably will not be commercially available
within the next 10 years. Some companies have experimented with using
inferior coking coals by either coal briquetting or formed coke processes,
5-1
-------
but large-scale commercial use of these is not expected in the near future.3
Thus, any process modifications will be within the process description
explained in Chapter 3.
One example of a process variation that would not be considered a
process modification is inconsistent variation in naphthalene processing.
There is substantial potential for temperature variability in naphthalene
melting operations, and thi ^variability leads to emission variability.
The temperature variability probably would not be considered a process
modification, but if the method of naphthalene melting consistently results
in greater emissions, such a change may constitute a process modification.
5.2.1 Tar Dewatering
Thermal dewatering of tar is a variation of tar dewatering by decanting
in storage tanks. Water is driven off as water vapor. Higher temperatures
are used in thermal dewatering than are used in other dewatering processes;
therefore, the implementation of thermal dewatering could increase benzene
emissions and might be a process modification.
5.2.2 Tar Storage
Increases in the storage temperature and changes in the method of
filling the tank are examples of process modifications that could increase
emissions.
5.3 EQUIPMENT MODIFICATIONS
Combined with the definition of modification that excludes routine
maintenance, repair, and replacement of equipment, it is not expected that
equipment changes would be potential modifications. Any discontinuance of
a control or control technique on a source that, does not offset the increased
emissions by implementing an alternate control technique on that source
would be considered a modification.
5.4 REFERENCES
1.
2.
3.
National Emission Standards for Hazardous Air Pollutants.
40 CFR 61.02.
National Emission Standards for Hazardous Air Pollutants.
40 CFR 61.05 (44 Federal Register 55174).
Subpart A.
Subpart A.
Hogan, W. T., and F. T. Koebkle. Analysis of the U.S. Metallurgical
Coke Industry. Industrial Economics Research Institute, Fordham
University. October 1979. (Prepared for U.S. Department of Commerce
under EDA Project 99-26-09886-10.)
5-2
-------
6. MODEL PLANTS AND CONTROL OPTIONS
The impact of various options to control benzene emissions from
coke by-product recovery plants is determined, in part, through analysis
of model plants. Subsection 6.1 defines three model coke by-product
recovery plants that typify processes that might be present at a small
by-product plant, a medium by-product plant, and a large by-product
plant. Discussed in Subsection 6.2 of this chapter are the control
options considered for the benzene emission sources present in coke
by-product recovery plants.
6.1 MODEL PLANTS OVERVIEW
Model plants for this industry are parametric descriptions of the
processes that may be practiced at an actual plant in a given size
range. Model plants are used primarily to estimate costs for each
control option as a function of plant size. Specific production
capacity, processes, and emission sources first were identified for
each actual plant to develop estimates of total nationwide impacts. A
cost function for each process and emission source was then developed
from the model plant analysis in terms of production capacity and
applied to each actual plant. Actual plant costs are summed for all
55 plants to estimate total nationwide costs. This method of analysis
accounts for variations in the processes used at individual plants and
the differences in cost caused by these variations. Nationwide emission
estimates are based on the type of process or emission source at a
particular plant, the associated emission factor for the emission
source in terms of grams of benzene per megagram of coke capacity, and
the plant's capacity. The estimated nationwide environmental and
energy impacts of each control option are presented in Chapter 7.
6-1
-------
Information regarding estimated control costs and costing methodology
is presented in Chapter 8, and an economic impact analysis of the
control options is presented in Chapter 9.
6.1.1 Selection of Model Plant Size
Three model plants were developed to represent typical process
combinations for a small plant (Model Plant 1), a medium plant (Model
Plant 2), and a large plant (Model Plant 3). The approximate distribu-
tion of actual plant sizes as a function.of coke capacity is shown in '.
Figure 6-1. Based on the distribution indicated in Figure 6-1, 25 (of
55 existing plants) plants produce between 300 Mg/day of coke (330
ton/day) and 2,000 Mg/day of coke (2,200 ton/day), accounting for
17 percent of total domestic coke capacity. For the model plant
analyses, a small model plant is defined as a plant producing
1,000 Mg/day of coke (1,100 ton/day), slightly less than the midpoint
of the actual production range. A total of 26 plants produce between
2,000 Mg/day of coke (2,200 ton/day) and 6,000 Mg/day of coke (6,600
ton/day). These medium-sized .plants account for 59 percent of total
domestic capacity. The production range midpoint of 4,000 Mg/day
(4,400 ton/day) was selected to define the size of a medium-sized
model plant (Model Plant 2). According to the distribution shown in
Figure 6-1, four plants produce between 6,000 Mg/day (6,600 ton/day)
and 13,000 Mg/day (14,300 ton/day) of coke. These large plants account
for 24 percent of total domestic capacity. For model plant analyses,
a large plant (Model Plant 3) is defined as a model plant producing
9,000 Mg/ day (9,900 ton/day), the midpoint of the actual production
range.
No construction of new plants is expected during the next 5 years.
However, if a new plant were constructed, it most probably would fall
within the size ranges for Model Plant 1, 2, or 3.
6.1.2 Selection of Model Plant Emission Sources
A total of 55 coke by-product recovery plants currently operate
throughout the United States. These plants vary widely in size, age,
design, equipment, products, and degree of control. Other factors
such as space requirements; availability of public water treatment
6-2
-------
40,000 r-
35,000
30,000
j: 25,000
e/j
z
o
u
u
uu
20,000
15,000
10,000
5,000
15
>
H;
U
a
I-
u.
O
u
tr
10
300- 1001- 2001- 3001- 4001- 5001- 6001- 7001- 8001- 9001-10,001-
1000 2000 3000 4000 5000 6000 7000 8000 9000 10,00012,000
COKE CAPACITY (Mg/DAY)
NOTE: Numbers above bars indicate number of plants in a givan size range.
12,001-
13,000
Figure 6-1. Distribution of plant size as a function of coke capacity.
6-3
-------
facilities for waste disposal; and the plant's physical location in
relationship to sensitive environmental areas, such as wetlands, also
contribute to the site-specific nature of by-product plant processes
and operational characteristics.
Many different process combinations are used throughout the coke
by-product recovery industry because of the site-specific nature of
the plants. For this reason, typical representations of actual pro-
cesses and process combinations were assigned to the appropriate model
plant size. The process combinations used are similar to those widely
used at actual plants. By-product recovery processes associated with
the emission sources considered for regulation for each model plant
are presented in Table 6-1.
The presence of an emission source at a plant depends on the
processes practiced at that plant. Benzene emission sources associated
with the model plant processes are shown in Table 6-2. Coke by-product
recovery process flow diagrams for the three model plants are presented
in Figure 6-2, Figure 6-3, and Figure 6-4. These flow diagrams are
intended to represent the typical products, processes, and emission
sources for each model plant size. Table 6-3 indicates the estimated
number of process units for each model plant size. The number of
process units and storage tanks at the model plants was derived from
plant trips, emission test reports, and responses to Section 114
questionnaires. The number of units and the processes practiced at
specific plants are variable because various sizing options are avail-
able. For example, a small plant could have one large light-oil
storage tank or two smaller light-oil storage tanks. The numbers in
Table 6-4 represent typical numbers of sources according to plant size
and span the range of the available data for the number of units at
specific plants.
As indicated in Figures 6-2 through 6-4, crude tar production is
practiced at Model Plants I, 2, and 3. Benzene emission sources
associated with crude tar production considered for regulation include
tar decanters, tar-intercepting sumps, flushing-liquor circulation
tanks, tar-dewatering tanks, excess-ammonia liquor storage tanks, and
tar storage tanks (including tar-collecting tanks).
6-4
-------
TABLE 6-1. COKE BY-PRODUCT RECOVERY PLANT PROCESSES
Model plant
Size (Mg/day)
Range represented (Mg/day)
Number of plants within represented
range
Percent of total coke capacity
Crude tar production
Direct-water final cooler
Tar-bottom final cooler
Wash-oil final cooler
Naphthalene processing
Light-oil recovery
Light-oil rectification
Light-oil refining
1
1,000
300-
2,000
25
17
Yes
Yes
No
No
Yes
Yes
No
No
2
4,000
2,000-
6,000
26
59
Yes
No
Yes
No
No
Yes
Yes
No
3
9,000
6,000-
13,000
4
24
Yes
No
No
Yes
No
Yes
Yes
Yes
Based on the distribution presented in Figure 6-1.
'includes naphthalene separation, drying, and handling.
6-5
-------
IS
O
O
O
_w
O.
"5
"O
o
0)
i l
I Q.
§
5 8
8 o
(O
I u.
6-6
-------
2|
•&1
Wash-Oil
Storage
^ '
^\
L
Wash-Oil
Circulation
CIT)
(2^
>
OJ.
"I
:'S
— *1
0
re
O
u
CC
n
1-
'
L
.1
§P-
S §£t
£CO
>s
•1 o
O ™
1^^
I?5S
Si |K
8.2-
fij
.rv!'
CO
T3
O)
O
O
O
CM
JO
Q.
"55
T3
o
I
0)
U
•s
o
i
OJ
^
o
o
CO
O)
iZ
6-7
-------
CO
O)
2.
CO
OJ
-
(6
6-8
-------
TABLE 6-2. EMISSION SOURCES FOR COKE BY-PRODUCT RECOVERY
MODEL PLANTS
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Source
Tar decanter
Tar- intercepting sump
Flushing- liquor circulation tarik
Tar- dewate ring tank
!
Tar storage tank
Excess-ammonia liquor storage tank
Direct-water final -cooler cooling tower
Naphthalene processing
Light-oil condenser and light-oil decanter vent
Light-oil storage tank
Light-oil sump
Tar-bottom final -cooler cooling tower
Benzene mixtures (BTX) storage tank
Wash-oil decanter
Wash-oil circulation tank
Benzene storage tank
f*
Equipment components
Model plant
1,2,3
1,2,3
1,2,3
1,2
1,2,3
1,2,3
1
1
1,2,3
1,2,3
1,2,3
2
2,3
1,2,3
1,2,3
3
1,2,3
Corresponds to sources indicated in Figures 6-2, 6-3, and 6-4.
Includes naphthalene separation, drying, and handling.
'Pumps, valves, exhausters, pressure-relief devices, sampling connection
systems, and open-ended lines.
6-9
-------
TABLE 6-3. NUMBER OF 'PROCESS UNITS AT COKE BY-PRODUCT RECOVERY
MODEL PLANTS
Process equipment
Tar decanter
Flushing-liquor circulation tank
Tar-intercepting sump
Tar dewatering tank
Tar-storage tanka
Light-oil and BTX storage tank
Light-oil condenser
Light-oil sump
Light-oil decanter
Wash-oil decanter
Wash-oil circulation tank
Excess-ammonia liquor storage tank
Benzene storage tank
Model
Plant 1
2
1
1
1
4
2
3
1
1
1
1
1
0
Number of
Model
Plant 2
3
2
1
2
8
6
3
1
I
2
2
3
0
units
Model
. Plant 3
6
3
2
4
12
9
3
2
2
4
4
6
3
Includes tar-collecting tanks.
6-10
-------
The final-cooler cooling tower, generally uncontrolled throughout
the industry, is usually the largest source of benzene emissions at a
plant equipped with a direct-water final cooler. This process is
practiced by approximately 23 plants. Benzene emissions are released
when the water from the final cooler is cooled against air in the
direct-water final-cooler cooling tower.
Plants within the size range of Model Plant 1 (300 to 2,000 Mg/day)
account for about half of the direct-water final coolers and a similar
proportion of tar-bottom final coolers in the industry. For the model
plant analyses, Model Plant 1 has been assumed to have a direct-water
final cooler and Model Plant 2 a tar-bottom final cooler.
At Model Plant 1, naphthalene is separated from the process
stream by a direct-water final cooler. Naphthalene is removed from
the well of the final cooler and may be transported to facilities for
steam drying. Naphthalene processing (including separation, drying,
and handling) may result in significant quantities of benzene emissions.
Model Plant 2 has a tar-bottom final cooler. This process is
used by approximately 18 plants, or 33 percent of the industry.
Although benzene emissions are still released when water is cooled
against air in the final-cooler cooling tower, emissions are substan-
tially less than are emissions from the direct-water final-cooler
cooling tower. When naphthalene is separated by a tar-bottom final
cooler, the naphthalene remains in the tar. The tar in which the
naphthalene is entrained may be recirculated by pipeline to tar storage
tanks or sold as a final product. Thus, benzene emissions from
naphthalene separation, drying, and handling are not attributed to
Model Plant 2.
At Model Plant 3, a wash-oil final cooler is assumed to be present.
Five plants currently are equipped with this system. Because the
wash oil is cooled in an indirect heat exchanger, there are no benzene
emissions from the cooling tower. In this system, naphthalene dissolves
in the wash oil, which is then indirectly cooled and recirculated to
the final cooler. Although emissions are not released from the cooling
tower, some emissions occur from the wash-oil decanter and wash-oil
circulation tank associated with the wash-oil final cooler.
6-11
-------
Light-oil recovery processes are attributed to Model Plants 1, 2,
and 3. Benzene emission sources associated with light-oil recovery
include the common vent for light-oil condensers and light-oil decant-
ers, light-oil storage tanks, light-oil sumps, wash-oil decanters, and
wash-oil circulation tanks. At Model Plant 3, wash-oil decanters and
wash-oil circulation tanks occur in conjunction with both the wash-oil
final-cooler system and light-oil recovery operations. However,
light-oil rectification to obtain benzene-mixture products such as BTX
is attributed only to Model Plants 2 and 3. Storage tanks used to
hold benzene mixtures are the emission sources associated with light-oil
rectification at these model plants. Because light-oil refining is
usually practiced at large plants, benzene storage tanks are attributed
only to Model Plant 3.
Fugitive emission sources at coke by-product recovery plants
include equipment components such as pumps, valves, exhausters,
pressure-relief devices, sampling connection systems, and open-ended
lines. This equipment is prevalent among all plants and is attributed
to Model Plants 1, 2, and 3. Benzene emissions and the associated
control costs for this equipment depend on the number of pieces of
equipment at the plant, and not on plant capacity. Plants that practice
benzene refining would have more pieces of equipment than do plants
that recover light-oil and BTX. Thus, Model Plant 3, which practices
benzene recovery, is credited with more pieces of equipment than are
Model Plants 1 and 2.
Table 6-4 presents the estimated number of leaking equipment
components in benzene service for each model plant size. The number
of equipment components was derived from emission test reports and
responses to Section 114 questionnaires. The data on number of
exhausters ranged from 2 to 6 for 8 plants, and 1 plant had 25
exhausters. Because exhausters can be sized to handle different
capacities, the number of exhausters was not a function of capacity;
therefore, the number chosen for the model plants (six) represents an
average of available data. Sample connections were defined as a
6-12
-------
TABLE 6-4. NUMBER OF EQUIPMENT COMPONENTS AT COKE BY-PRODUCT
RECOVERY MODEL PLANTS
Equipment item
Exhausters
Pump seals
Valves
Relief valves
Sample connections
Open-ended lines
Model Plant 1
6
15
105
5
10
22
Number of units
Model Plant 2
6
15
105
5
10
22
Model Plant 3
6
30
210
9
21
45
6-13
-------
subset of open-ended lines; therefore, the 22 open-ended lines for
Model Plant 2 includes 10 sampling connections and 12 open-ended lines
that are not sampling connections.
6.2 CONTROL OPTIONS OVERVIEW
In Subsection 6.1, the emission sources considered for regulation
are identified in association with typical processes that may be
practiced at each size model plant. These emission sources are dis-
cussed further in Chapter 3. Several options are available for the
control of benzene emissions from these sources. The control options
considered for "best available technology" (BAT) for each emission
source and the associated benzene control efficiencies are presented
in Table 6-5. Further information regarding each control technique is
contained in Chapter 4. Detailed cost information is presented in
Chapter 8 for each emission source and associated control option. The
environmental and energy impacts of the control options are discussed
in Chapter 7, while the economic impacts are presented in Chapter 8.
6.2.1 Final-Cooler Cooling Tower
As shown in Table 6-5, three options are considered to control
emissions from final-cooler cooling towers. At plants operating a
direct-water final cooler, naphthalene could be collected by a tar-
bottom final cooler or a wash-oil final cooler. Both systems would
eliminate benzene emissions resulting from naphthalene separation,
handling, and drying and would reduce emissions from the cooling tower
substantially. Use of the tar-bottom final cooler would achieve an
overall benzene emission reduction of about 81 percent, while use of a
wash-oil system would achieve an emission reduction of 100 percent.
At a medium plant operating a tar-bottom final cooler, a benzene
100-percent control efficiency also would be achieved with a wash-oil
final cooler.
6.2.2 Gas Blanketing System
Gas blanketing has been demonstrated as an effective control
technique for removing hydrocarbon vapors; e.g., benzene, from process
vessels and product storage tanks. The basic principles of gas blanket-
ing require sealing all the openings on a vessel or tank, supplying a
6-14
-------
TABLE 6-5. COKE BY-PRODUCT PLANT BENZENE EMISSIONS
SOURCES AND CONTROL OPTIONS
Control
efficiency
Emission source
Control option
1. Final-cooler cooling towers
a. Direct-water final-cooler
cooling tower
b. Tar-bottom final-cooler
cooling tower
2. Tar decanters
3. Tar-intercepting sump
4. Flushing-liquor circulation tanks
5. Tar-dewatering tanks
6. Light-oil condenser and light-oil
decanter vents
7. Wash-oil decanters
8. Wash-oil circulation tanks
9. Tar-storage tanks
10. Excess-ammonia liquor storage
tanks
11. Light-oil storage tanks
12. Benzene-mixture storage tanks
13. Benzene storage tanks
1. Use tar-bottom final cooler 81
2. Use wash-oil final cooler 100
1. Use wash-oil final cooler 100
Coke oven gas blanketing from collecting main 95
Coke oven gas blanketing from collecting main 98
Coke oven gas blanketing from collecting main 98
1. Coke oven gas blanketing from collecting main 98
2. Wash-oil scrubber 90
Coke oven gas blanketing from gas holder 98
Coke oven gas blanketing from gas holder 98
Coke oven gas blanketing from gas holder 98
1. Coke oven gas blanketing from collecting main 98
2. Wash-oil scrubber 90
1. Coke oven gas blanketing from collecting main 98
2. Wash-oil scrubber go
1. Coke oven gas blanketing from gas holder 98
2. Wash-oil scrubber 90-
1. Coke oven gas blanketing from gas holder 98
2. Wash-oil scrubber 90
1. Nitrogen or natural gas blanketing 98
system
14.
15.
16.
17.
18.
19.
20.
Light-oil sumps
Pumps
Valves
Exhausters
Pressure-relief devices
Sampling connection systems
Open-ended lines
2. Wash-oil scrubber
Source enclosure
1. Quarterly inspections
2. Monthly inspections
3. Equip with dual mechanical
seals
1. Quarterly inspections
2. Monthly inspections
3. Equip with sealed bellows
valves
1. Quarterly inspections
2. Monthly inspections
3. Equip with degassing
reservoir vents
1. Quarterly inspections
2. Monthly inspections
3. Equip with rupture disc
system
Closed-purge sampling
Plug or cap
90
98
71
83
100
63
72
100
55
64
100
44
52
100
100
100
Includes a 100-percent emission reduction for naphthalene processing and a 74-percent emission
reduction for the direct-water final-cooler cooling tower.
6-15
-------
constant-pressure gas blanket with coke-oven gas, nitrogen or natural
gas, and providing for recovery or destruction of displaced vapor
emissions. Depending on the source to be controlled, displaced vapors
from the enclosed source can be transported through a piping system to
the collecting main, to the battery gas holder, or to another point in
the by-product recovery process.
With gas blanketing from the collecting main, a vapor recovery
system is in place in the form of the by-product recovery process that
removes organics from the raw coke oven gas. Emission sources that
can be blanketed with raw coke oven gas from the collecting main
include tar decanters, tar-intercepting sumps, flushing-liquor circula-
tion tanks, tar storage tanks, tar-dewatering tanks, and excess-ammonia
liquor storage tanks. With gas blanketing from the gas holder, a
vapor destruction system is in place because the clean oven gas is
burned to recover the fuel valve. Emission sources that can be
blanketed with clean coke oven gas from the battery gas holder include
light-oil condensers and decanters, wash-oil decanters and circulation
tanks, light-oil storage tanks, and benzene-mixture storage tanks. To
prevent product contamination, nitrogen or natural gas can be used to
blanket storage tanks containing refined benzene. Emissions could be
routed to the collecting main and burned in the gas combustion system
or routed to the gas main before light-oil removal and recovered in
the wash-oil scrubbing operation.
With source enclosure, the blanketing system's benzene control
efficiency is essentially 100 percent. Because the deterioration of
piping occasionally can result in leaks, the benzene control efficiency
for gas blanketing is estimated at 98 percent for each source except
tar decanters. A lower control efficiency (95 percent) is estimated
for tar decanters because a portion of this vessel's surface area must
be left open to the atmosphere to allow for sludge removal operations.
6.2.3 Wash-Oil Scrubber
A wash-oil scrubber also can be used to absorb organics from tar
dewatering tanks and from storage tanks containing tar, excess ammonia
liquor, light-oil, BTX, or refined benzene. In some cases, a wash-oil
scrubber could be less expensive than gas blanketing would be. Wash-oil
6-16
-------
scrubbers currently used for light-oil removal are large towers designed
to handle high volumes of coke oven gas. This technology can be
applied to these storage tanks based on a smaller scale design for the
scrubbing chamber and a lower wash-oil circulation rate. In an unpacked
wash-oil scrubber, emissions enter the bottom of the scrubbing chamber
and contact a spray of wash oil, which is introduced into the top of
the spray chamber. The wash-oil spray absorbs benzene from the vented
vapors. After passing through the scrubber, benzolized wash oil is
routed to the light-oil recovery plant for removal of benzene and
other organics; the debenzolized wash oil is then recycled to the
scrubber. The benzene control efficiency of this technique is estimated
to be 90 percent.
6.2.4 Light-Oil Sump
Source enclosure has been demonstrated as an effective method for
reducing benzene emissions from this source. The enclosure (i.e., a
roof) need not be permanently affixed so the roof could be removed to
allow for maintenance or sludge removal. A gasket seal could be
installed around the rim of the sump cover to form a closed system to
contain the emissions. In addition, a vertical vent could be added to
the sump cover so that excess pressure does not build up in the sump.
Emissions from the vertical vent could be controlled by means of a
water leg seal, a pressure-relief device, or a vacuum relief device.
The' control efficiency of the sump cover, including the vertical vent,
is estimated at 98 percent.
6.2.5 Pumps
Three options are considered to control fugitive emissions from
leaking pumps. These options include implementing a leak detection
and repair program based on .quarterly or monthly inspection intervals.
As indicated in Table 6-5, quarterly inspections would achieve about a
72-percent benzene control efficiency, while monthly inspections would
achieve about an 83-percent benzene control efficiency. A third
option would require that pumps be equipped with dual mechanical seal
systems. This equipment requirement would achieve a benzene control
efficiency estimated at 100 percent.
6-17
-------
6.2.6 Valves
Three options also are considered to control fugitive benzene
emissions from leaking valves. These options'include implementing'a
leak detection and repair program based on inspections made at quarterly
or monthly intervals. A third option would require installing sealed-
bellows valves. Quarterly monitoring valves result in about a
63-percent control efficiency. A leak detection and repair program
based on monthly monitoring intervals would achieve a benzene control
efficiency estimated at 73 percent. Equipping each existing valve at
a medium-sized plant with sealed bellows valves would result in about
a 100-percent benzene control efficiency.
6.2.7 Exhausters '
Control options similar to those for pumps and valves are consid-
ered for application to exhausters. Implementing a leak detection and
repair program with monitoring at quarterly intervals would achieve
about 42 percent benzene control efficiency, while monitoring at
monthly intervals would result in a 52-percent benzene control effi-
ciency. An estimated benzene control efficiency of 100 percent would
be achieved if each exhauster were equipped with degassing reservoir
vents. Emissions from the degassing reservoir vents could be vented
to a control device or back to the process. For example, a closed
loop could be installed to route emissions from the degassing reservoir
vent to the exhauster inlet and back into the coke oven gas.
6.2.8 Pressure-Relief Devices
The control options considered for pressure-relief devices include
quarterly inspections, monthly inspections, and equipment requirements.
The equipment requirements considered include the use of a rupture
disc system (block valve or a three-way valve). A leak detection and
repair program with monitoring at quarterly inspections would achieve
a benzene control efficiency of about 44 percent, while an estimated
benzene control efficiency of 52 percent would result from a monthly
inspection program. Equipping each device with a rupture disc system
would achieve a benzene control efficiency estimated at 100 percent.
6-18
-------
6.2.9 Sampling Connection Systems and Open-Ended Lines
Benzene emissions from open-ended lines can be eliminated by
capping or plugging the end of the line. Closed-purge sampling tech-
niques can eliminate benzene emissions from a sampling connection
system. As shown in Table 6-5, the benzene control efficiency for
both control options is estimated at 100 percent.
6-19
-------
-------
7.0 ENVIRONMENTAL IMPACT
This chapter discusses the environmental impacts from imple-
menting the control options presented in Chapter 6. The primary
emphasis is a quantitative assessment of benzene emissions that would
result from each of the control options. The emissions of organic
compounds other than benzene also are estimated. Both beneficial and
adverse environmental impacts are assessed in terms of water quality,
solid waste, energy, and other environmental concerns.
7.1 BENZENE AIR POLLUTION IMPACT
7.1.1 Emission Source Characterization
The emission sources at coke oven by-product plants are discussed
in Chapter 3. The emission sources, emission factors, and
uncontrolled industry emissions are presented in Table 7-1. These
uncontrolled emissions are characteristic of existing conditions and
are considered baseline. They are estimated under the regulatory
alternative of no national by-product plant benzene emission standard.
Table 3-3 presents assumptions about the emission source distribution
among the various coke oven by-product plants. Chapter 6 describes
the model plant approach used to characterize the various emission
sources for different sized plants. The emission factors presented in
Chapter 3, the capacity of the plants identified in Table 3-3, and the
types of emission sources present at the different plants also
identified in Table 3-3 are used to estimate industry emissions.
7.1.2 Development of Benzene Emission Levels
Emission factors for the model units were determined for each
control option to estimate the impacts of the control options on
benzene emission levels. The control technology discussed in
Chapter 4 is applied to the model plants and to the industry model to
7-1
-------
TABLE 7-1. ESTIMATED NATIONAL BASELINE BENZENE EMISSIONS FROM COKE
OVEN BY-PRODUCT RECOVERY PLANTS
Number
of plants
uncontrolled
Direct-water final -cooler
cooling tower
Tar-bottom final -cooler
cooling tower
Light-oil condenser vent
Naphthalene separation
Naphthalene processing
Tar- intercepting sump
Tar dewatering
Tar decanter
Tar storage
Light-oil sump
Light-oil storage
Benzene-tol uene-xyl ene
storage
Benzene storage
Flushing- liquor circulation
tank
Excess-ammonia liquor tank
Wash-oil decanter
Wash-oil circulation tank
Pump seals
Valves
Pressure relief devices
Exhausters
Sample connections
Open lines
23
18
44
23
23
55
54
55
55
46
46
20
7
55
55
44
44
46
46
46
46
46
46
Capacity
uncontrolled
(Mg/day)
64,376
42,790
125,724
64,376
64,376
154,680
142,000
154,680
154,680
143,203
143,203
39,479
35,720
154,680
154,680
131,340
131,340
143,203
143,203
143,203
143,203
143,203
143,203
Emission National
factor emissions
(g/Mg) (Mg/yr)
270
70
89
87
20
95
21
77
12
15
5.8
5.8
5.8
9
9
3.8
3.8
a
a
a
a
a
a
6,340
1,090
4,080
2,040
470
5,360
1,090
4,350
680
780
300
80
80
510
510
180
180
600
400
270
33
50
18
Emissions were estimated on the
leaking units. Emission factors
basis of the number of potentially
in kg/day are listed in Table 3-6.
7-2
-------
estimate the reduction in benzene emissions below baseline levels.
For example, the controlled emission factor for tar decanters was 5
percent of the uncontrolled emission factor because the control was
assumed to be 95 percent effective.
Controlled benzene emission factors were also developed for
sources that would be controlled by implementation of a leak detection
and repair program. These factors for pressure relief devices and
exhausters were calculated by multiplying the uncontrolled emission
factor for each type of equipment by a set of correction factors (see
Appendix F). The factors for pump seals and valves were obtained from
the leak detection and repair (LDAR) model discussed in Subsection
4.8.1.3. Plugs for open-ended lines and closed sampling lines were
assumed to be 100 percent effective.
The resulting controlled benzene emissions are listed in Table 7-2
by source. Where the control options require an equipment specifica-
tion to control leaks, it is assumed that there are no subsequent
emissions from the controlled source.
7.1.3 Impact oh Benzene Emissions from New Sources
Over a 5-year period from 1982 to 1986, no new by-product plants
are expected to be operated. Therefore, the control options are
estimated to affect only existing emissions.
7.2 IMPACT OF THE CONTROL OPTIONS ON VOLATILE ORGANIC COMPOUND (VOC)
EMISSIONS
VOC emissions were estimated by using emission factors derived
from coke oven by-product plant sampling.1 The bases for derivation
of the emission factors are detailed in a separate report.2 The
emission factors are used in Table 7-3 to estimate the .national
emissions of VOC's. The atmospheric emissions are estimated as
approximately 194,000 Mg/yr of VOC's. Table 7-4 presents the effect
of the control options on national VOC emissions from each of the
plant sources.
The estimated 194,000 Mg of VOC's emitted each year from coke
oven by-product plants are a significant part of the estimated national
VOC emissions (1,400,000 Mg/yr from the processing of over 100 different
organic chemicals).3 The organic materials emitted from by-product
7-3
-------
•
LU
s*:
o
o
1
/— .
1 — S-
•r ,n^2^
C O •*-> O*
o in c c
•r- OH)
IA Q) O M
u» jr c
— *J t. 0)
E oija
tu m •*-»
0 t- C!
i ra£
V)
O •->
'« >1
"> "V
O »r- (U
Is!
in ro o>
m c: ja
m o
CO v^
0)
J- U «
III
0
c
o
CL
O
e
e
o
o>
U
3
Q
CO
1:
O°|OOO|CO Sc\i 'D013'3"0000 -HOIDOIOOOI-^ OOOCTl'-DOOO
CT> CM
CM
T
§22222 SS ooooooo oooooooo ooororontvjco
OOiaiSSo OO ^OOOOOOOrH rHrHlDlDOOOlD r-.r-~r-nnrOLOrH
en en o> m «i- «a- I-H I-H
S c3 S IS H? Si rS 7S *° mm us ID ID in mm ID ID us ID us us us us
Uo4..Z.E. S -r- O Ot-
UO4-'4^>4J c C'l— Ctt»4J C 'F— C r— *J 'I— C »r- f-.,—
§,_g_2_g^ fejIE tjiojJouia) t-rot- 4-ioffloi +JO +->oo
[J!.El54245^2 "sm •§U)CLtJI0'0 -0 JDcncLUuim D.L) CL u tu
S d *r" !«»-•.« U CD L) CJ) »r- UiC UOU-fJ-r-olf— •r-wt/J'r-loI-C
0 S S S £= S-"*-^111^^ W W^D)D10) r-S OJf-S £*"" "5 g) ^- S ^- "c r^'^" ^ S £*"" 4*'1" ? 'a.
,2 | »C > > S ^ fil *J 1 4->^I jCOJCJ 4JJZ~-rJX:wtOC
*J t-w o o o \n ^£ "Ow^ m'c'fo o "w-S:~5)Lo m"c7c o ro'c'o.m'cra-n''"
f "S 1 1 1 55 | 5<5 ^^ | 5
-------
TABLE 7-3. ESTIMATED NATIONAL BASELINE VQC EMISSIONS FROM COKE
OVEN BY-PRODUCT RECOVERY PLANTS
Number
of plants
uncontrol led
Direct-water final -cooler
cooling tower
Tar-bottom final -cooler
cooling tower
Light-oil condenser vent
Naphthalene separation
and processing
Tar- intercepting sump
Tar dewatering
Tar decanter
Tar storage
Light-oil sump
Light-oil storage
BTX storage
Benzene storage
Flushing- liquor circulation
tank
Excess-ammonia liquor tank
Wash-oil decanter
Wash-oil circulation tank
Pump seals
Valves
Pressure relief devices
Exhausters
Sample connections
Open 1 i nes
Total (rounded)
23
18
44
23
55
54
55
55
46
46
20
7
55
55
44
44
46
46
46
46
46
46
Capacity Emission National
uncontrolled factor emissions
(Mg/day) (g/Mg) (Mg/yr)
64,376
42,790
125,724
64,376
154,680
142,000
154,680
154,680
143,203
143,203
.39,47.9
35,720
154,680
154,680
131,340
131,340
143,203
143,203
143,203
143,203
143,203
143,203
4,239
1,100
127
168
202
492
164
281
21.4
8.3
8.3
5.8
12.9
12.9
5.4
5.4
— b
— b
— b
— b
— b
— b
99,600
17,200
5,830
3,950
11,400
25,500
9,260
15,900
1,120
430
120
76
730
730
260
260
850
570
390
140
,76
26
194,400
Benzene and other VOC.
Emissions were estimated on the basis
divided by 0.7; i.e., the fraction of
of benzene emissions in Table 7-1
benzene in light oil.
7-5
-------
V)
t/5
LU
o
o
o
U)
to
Z V5
O I—
O Q.
O O
ce: rs
O Qi
CJ CL.
tui
Z CO
UJ
M Z
Z UJ
UJ >
CO O
UJ
u_
U-
UJ
UJ
UJ
__I
CD
'*" ** ^>
ift u<:
:-ifg
§ ™
v> p c<^
§•
g gg g
o o o o o o o ooo in
mo to I-H r-
ooo r- o o
^
ch
• o t-i coin cMin »-t
8 gg §
> oo o oo ooo
_ ) o o CM (^ o in in in
CM co in in t-i 01 en cococo
OO OO OOO
*
ooo ooo
---
IDU3VJ3 IJ3UOCO
• -
c »— t- 01 01
c ~t- c o)
. RJ *J O W)
) f— U •!-
1/3 (U U r—
13 U) 13 C
.
O) J i —
-
U O) -r- U) C
- -'f- co-o
u en u -4-> T- w i—
c CL u CL c <
•i- > > -r-
U Dl r— OJ r- C f—
-
oo QJ (— E
I +J .C
.CO) t- 4-> i—
.
(OO QJ (DO D O =3
.* O r- ^ CJ 0-E O
••- > -i- (C t- >» I
O O Of— 0> *— XJ
O) F—
>>••- TJ "- >, CO ^-
. -
OJCO. C05O1 T3 r-
f-l CM U O U r-i CM
rHCM r-ICMiHCMro
rH CM CO iH CM 0
i-
•r- O* r—
i— ra «—
r- T- TJ
•r- W>
§•
O •—
I 4-> tO
7-6
-------
plants can participate in a wide variety of reactions in the atmosphere,
including singlet oxygen formation4 and formation of ozone-hydrocarbon
reaction products.5 6
7.3 WATER POLLUTION IMPACT
Most of the control options for the major emission sources do not
increase the water pollution of the plants. The preferred technique
for most major emission sources is coke oven gas blanketing, which
results in essentially complete control of the emission source. Any
emissions that are vented from the process are returned to the process
at a different location. Thus, no water pollution problems are associ-
ated with recycling benzene vapors.
A possible exception is the increased cyanide concentration in
wastewater due to indirect heat exchange. Presently, cyanide is
emitted from the final-cooler cooling tower at some plants by air
stripping of the wastewater. Measured HCN air emissions and calcula-
tions based on once-through cooling water indicate that about 200 g/Mg
of coke could be added to wastewater for treatment, if indirect cooling
were substituted for direct cooling.1 The actual amount of additional
cyanide in the wastewater would depend on cooling water temperature,
degree of recycle practiced, and numerous other factors.
7.4 SOLID WASTE DISPOSAL IMPACT
None of the control options will adversely impact either solid
waste generation or disposal. The blanketing control techniques not
only would result in more complete control of the source but would
eliminate some of the potential solid waste problems due to sludge
formation in light-oil plant process equipment.
Potential solid waste sources include replaced mechanical seals,
seal packing, rupture discs, and valves. Neither the volume of this
waste nor its degree of potential environmental hazard is expected to
be significant.
7.5 ENERGY IMPACT
The blanketing and venting systems are essentially passive control
techniques; the only energy required for their operation is heat to
7-7
-------
prevent vapor freezing in some of the blanketing and vent lines. The
energy to heat these pipes could come from electrical heating tape or
steam tracing. The pipes would be insulated to reduce the energy
requirements.
Table 7-5 summarizes the energy requirements that were assumed
for the gas blanketing and wash-oil scrubber control options described
in Chapter 6 and costed in Chapter 8. Steam estimates include amounts
needed for pipe heat tracing. The modest amount of steam could be
available from low-pressure waste steam currently vented.
A major energy impact for the control technology is the electrical
power for the wash-oil final cooler. Alternatively, if tar-bottom
final cooling is used, the electrical consumption is much lower. This
altered consumption results from differences between wash oil's and
water's heat capacities and heat transfer coefficients and because a
tar bottom (mixer-settler) is add-on instead of entire replacement
equipment.
The major energy impact of the control options is the potential
for recovering large amounts of benzene and other organic compounds
that otherwise would be released to the atmosphere. The light aromat-
ics are important because their uses include fossil fuel replacement
and gasoline additives.
Several of the coke oven by-product plant sources that emit
benzene also emit coke oven gas (methane and hydrogen). The amount of
coke oven gas emitted could be substantially greater than the amount
of benzene emissions. Table 7-6 summarizes process unit coke oven gas
emissions that could be recovered as a result of recycle of these
gases back to the coke oven gas.
Table 7-6 does not include an estimate of the coke oven gas lost
from other potential sources at by-product plants. If recovery of
21.3 H of gas/min/Mg of coke/day (see Table 7-6) is assumed, the
national energy savings from the recovered coke oven gas would be
approximately 36,100 TJ/yr (0.034 quad/yr).
7-8
-------
TABLE 7-5. ENERGY USE AT A MODEL BY-PRODUCT PLANT
(4,000 Mg coke/day)
User
Steam
(Mg/yr)
Electricity
(MWh/yr)
Gas blanketing
Tar decanter, tar-intercepting sump,
and flushing-liquor circulation tank
Tar dewatering, tar storage
Light-oil storage
Excess-ammonia liquor tank
Condenser, light-oil decanter, wash-oil
decanter, and circulation tank
Wash-oil scrubber
Excess-ammonia liquor tank
Light-oil storage
Final cooler
Tar-bottom final cooler
Wash-oil final cooler
350
440
107
128
176
60
60
10
10
85
2,020
TABLE 7-6. EMISSIONS OF COKE OVEN GAS FROM
SELECTED COKE OVEN BY-PRODUCT PLANT SOURCES1
Sources
Emissions
(£ gas/min/Mg coke/day)
Tar decanter
Light-oil condenser
Tar dehydrater
Tar storage
11.0
0.2
4.6
5.5
7-9
-------
7.6 OTHER ENVIRONMENTAL IMPACTS
The control options would have improve the general appearance of
by-product plant operations because they would tend to eliminate
aesthetically displeasing phenomena such as water vapor plumes from
process vents and naphthalene precipitation from air, and they might
reduce some of the odors emitted from some process steps. Other
environmental considerations, such as noise level, are not expected to
be influenced by the control options.
7.7 IRREVERSIBLE AND IRRETRIEVABLE COMMITMENT OF RESOURCES
The control options do not involve a tradeoff between short-term
environmental gains at the expense of long-term environmental losses.
The control options, do not result in irreversible and irretrievable
commitment of resources. As a result of the control options, resources
such as light aromatic hydrocarbons are recovered, and emissions from
affected sources are essentially eliminated.
7.8 IMPACT OF DELAYED STANDARDS
Delay of the standard will not significantly impact water pollution,
solid waste disposal, or energy. A delay will result in continued air
pollution at or above the level of national baseline benzene emissions
(see Table 7-1). The health impact from control at this level (described
in Appendix E) would continue throughout the delay.
7.9 REFERENCES
1. VanOsdell, D.W. Environmental Assessment of Coke By-Product
Recovery Plants. U.S. Environmental Protection Agency. Research
Triangle Park, NC. Publication No. EPA-600/2-79-016. January
1979.
2.
3.
4.
Branscome, M. R., Summary of VOC and Total Organic Estimates for
Coke By-Product Recovery Plants. Research Triangle Institute.
Research Triangle Park, NC. July 11, 1983.
Hydrocarbon Pollutants from Stationary Sources.
Protection Agency. Research Triangle Park, NC.
EPA-600/7-77-110. September 1977.
U.S. Environmental
Publication No.
Geacintov, N. E. Reactivity of Polynuclear Aromatic Hydrocarbons
with 02 and NO in the Presence of Light. U.S. Environmental
Protection Agency. Research Triangle Park, NC. Publication No.
EPA-650/1-74-010. 1973.
7-10
-------
5.
6.
7.
Bufalini, J., and A. Altshuller. Kinetics of Vapor-Phase
Hydrocarbon-Ozone Reactions. Canadian Journal of Chemistry 43
1965. pp. 2243-2250. —'
Tebbens, B., J. Thomas, and M. Mukai. Fate of Arenes Incorporated
with Airborne Soot. American Industrial Hygiene Association Journal
September-October 1966. pp. 415-422.
Coke By-product Emissions Evaluation Briefing.
tal Services. November 1980.
Scott Environmen-
7-11
-------
-------
8. COSTS
8.1 COST ANALYSIS OF REGULATORY ALTERNATIVES
This chapter contains cost estimates of implementing various
controls for benzene at existing and new by-product plants. Costs of
process modifications and add-on controls are presented for each of
the hazardous pollutant sources considered for regulation. The cost
analysis assumes that each source is uncontrolled and applies the
controls to the sources at each model plant. Not all by-product
plants will incur all of the costs described in this section because
the types of pollution sources differ among the various plants.
Control costs are presented in terms of total capital cost and total
annualized cost and their components.
Control costs for a particular plant are estimated according to
its coke capacity; a linear correlation between control cost and coke
capacity was obtained from the cost estimates for the three model
plants. Nationwide control requirements are estimated on an individual
plant basis, according to available information oh process sources and
coke capacity. Nationwide capital and annualized control costs are
presented in Subsection 8.1.4 for existing coke oven by-product plants.
Controls were selected for major air emission sources for coke
oven by-product plants and are described in Chapter 4. The special
process characteristics of the by-product plants were used to identify
cost-effective controls through implementation of various recycle
techniques. By-product plants have sources of gas for blanketing and
existing pressure control on the collecting main and gas holder for
the blanket gas. These characteristics permit implementation of
relatively inexpensive and effective controls.
8-1
-------
Subsection 8.1.1 gives the cost analysis for control of benzene
sources for existing by-product plants. The control method most
frequently advocated for the sources is blanketing with raw coke oven
gas from the collecting main or blanketing with clean coke oven gas
from the gas holder, both under a slight, positive pressure. As
discussed in Chapter 4, gas blanketing generally achieves essentially
complete control at less cost .than do condensers, absorbers, or
incinerators, which achieve only partial control.
8.1.1 Existing Facilities
8.1.1.1 Rationale. The number of process units, tanks, and
other emission points for the three model plants is given in Table 8-1
and was estimated from industry surveys and plant trips. A range of
cost estimates is provided for each model plant and is based on a
range of plant types and layouts. Because piping incurs a major
portion of cost for gas blanketing systems, a range of piping distances
is used for each model plant. These piping distances are based on
plant layout data from four plants and include two relatively compact
plants (Armco, Inc., in Houston and Bethlehem Steel Corporation in
Sparrows Point) and two plants that are comparatively spread out (U.S.
, /
Steel Corporation in Fairfield and Fair!ess Hills).
Costs of gas blanketing and wash-oil scrubber systems are based
on designs that have been applied by the industry (see Chapter 4).
Much of the design data were obtained from systems installed at Armco,
Inc., Houston Works, by the Engineering and Construction Division of
Koppers Company, Inc., a major builder of coke ovens and by-product
recovery plants.
To consider site-specific factors, EPA visited several by-product
plants that had installed some form of gas blanketing. In addition,
personnel visited the U.S. Steel Fairless Works to examine potential
difficulties in retrofitting a gas blanketing system in a plant with
long pipe runs. EPA consultants toured the plant and examined its
layout, existing piping and supports, operating parameters, and relevant
construction blueprints. Extensive data on tank dimensions, pumping
rates, piping distances, and pipe supports were obtained to develop a
8-2
-------
TABLE 8-1. NUMBER OF UNITS AT THE MODEL PLANTS
Process equipment
Tar decanter
Flushing- liquor circulation tank
Tar- intercepting sump
Tar-dewatering tank
Tar-collecting, storage tank
Light-oil storage tank
Light-oil condenser
Light-oil sump
Light-oil decanter
Wash-oil circulation tank
Wash-oil decanter
Excess-ammonia liquor storage tank
Pure benzene storage tank
Model
Plant 1
2
1
1
1
4
2
3
1
1
1
1
1
0
Number of units
Model
Plant 2
3
2
1
2
8
6
3
1
1
2
2
3
0
Model
Plant 3
6
3
2
4
12
9
3
2
2
4
4
6
3
8-3
-------
detailed construction cost estimate. The Fair!ess Works' estimate is
provided in Appendix F and was used to derive unit costs for many of
the items required for the model plants.
Major capital cost items, their unit cost, and the origin of the
estimate are summarized in Table 8-2. Annualized cost items are
listed in Table 8-3. In the following subsections, these unit costs
are applied to model plants for each emission point, or group of
emission points, to generate a range of capital and annualized costs
for each model. Annual light-oil recovery credits are subtracted from
annualized costs to determine net or total annualized costs. Recovery
credit is based on recovering additional light oil or, for cases of
venting to the gas holder, light oil's fuel value. Recovery credits
are expected to be conservative because no credit is estimated for
recovery or additional fuel value for organics other than light oil.
Available data are too sparse to estimate accurately the quantity,
composition, and value of these organics; but the limited data show
that the quantity of other organics could be significant. These other
organics vary in composition with the emission point and include such
compounds as hydrogen, methane, ethane, toluene, xylene, naphthalene,
and tars. Best available estimates of the quantity and value of these
other organics based on available data are summarized in Appendix F.3.
8.1.1.2 Tar Decanter, Tar-Intercepting Sump, and Flushing-Liquor
Circulation Tank. The costs of controlling these sources were calcu-
lated by grouping the sources because they are generally located close
to each other. The costs include covering and sealing the tar decanter
and sump and blanketing all of these vessels with coke oven gas from
the collecting main. Pressure control would be provided by the Askania
regulator, which maintains collecting main pressure at 5 to 10 mm of
water. Discussions with plant operators indicated that pressure
control in the collecting main is inherently reliable. High-pressure
excursions sound an alarm and open emergency bleeder stacks to vent
the excess. Low pressure is avoided because of potential damage to
the coke ovens and oxygen infiltration. When necessary, an operator
will control the Askania manually to maintain the desired collecting
main pressure.
8-4
-------
TABLE 8-2. CAPITAL COST ITEMS
(1982 Dollars)
Item
Capital cost factors
Cover decanter and sumps
Fittings0
Flame arrestors ,
Instrumentation
Performance test
Pipe (straight)
Pipe columns
Pressure controller
Pressure reducer
Pressure tap
Pump
Scrubber shell
Description
Construction fee
Contingency
Engineering
Startup
Clean, cover, and seal 22 m2 (240 ft2)
for $7,800
Clean, cover, and seal 52 m2 (560 ft2)
for $16,000
20-cm (8-in. ) pipe
15-cm (6-in. ) pipe
15-cm (6-in.) pipe, light-oil plant
10-cm (4-in.) pipe, light-oil plant
7.6-cm (3-in.) pipe
7.6-cm (3-in.) pipe, light-oil plant
For 15-cm (6-in.) vent
For 7.6-cm (3-in.) vent
Flow rate, pressure, temperature
200 to 300 person-hours
20 cm (8 in.)
15 cm (6 in. )
15 cm (6 in.), light-oil plant .
10 cm (4 in.), light-oil plant
7.6 cm (3 in.)
7.6 cm (3 in.), light-oil plant
8.1 cm (2 in.)
2.5 cm (1 in: )
For piping support
For 7.6-cm (3-in.) line
For ,15-cm (6-in.) line, with backup
For 7.6-cm (3-in.) line
Equipment rental
Labor and materials
2.2 H/s (35 gal/min), 2 hp
7.2 m2 (77 ft2) at $1,530
22 m2 (240 ft2) at $5,000
• .Cost/unit
10% of capital
15% of capital
15% of capital
1% of capital
355/m2
(32.5/ft2)
308/m2
(28.6/ft2)
20/m
(6. I/ft)
16/m
(5.0/ft)
26/m
(7.8/ft)
15/m
(4.6/ft)
6.9/m
(2. I/ft)
13/m
(4. I/ft)
1,870
920
1,300
8,000
138/m
(42. I/ft)
109/m
(33.2/ft)
185/m
(56.4/ft)
126/m
(38.4/ft)
46.6/m
(14.2/ft)
102/m
(31.0/ft)
30.7/m'
(9.36/ft)
20.2/ra
(6.17/ft)
1,500
3,400
12,600
525
4,500/15 days
1,750/tap
2,570
214/m2
(20/ft2)
226/m2
(20.8/ft2)
Reference
1
1
1
1
2,3a
2,3b
d
d
d
d
d
d
e
6
f
g
2h
2h
2h
2h
2h
2h
U
2h
2h
i
j
k
1
m
m
n
0
'P
Footnotes on last page of table.
(continued)
8-5
-------
TABLE 8-2. (continued)
Item Description
Steam trace, insulation 20-cm (8-in.) line
15-cut (6-in. ) line
15-cm (6-in.) line, light-oil plant
10-cm (4- in.) line, light-oil plant
7.6-cm (3-in.) line
7.6-cm (3-in.) line, light-oil plant
Valves 20-cm (8-in.) plug
15-cm (6-in. ) plug
15-cm (6-in.) 3-way, light-oil plant
7.6-cm (3-in.) 3-way, light-oil plant
5.1-cm (2-in. ) gate
2.5-cm (1-in. ) vent
2.5-cm (1-in.) gate
1.3-cm (0.5-in.) gate
Cost/unit
122/m
(37. I/ft)
68/m
(20.7/ft)
86/m
(26. I/ft)
70/m
(21.3/ft)'
45.6/m
(13.9/ft)
62/m
(18.9/ft)
1,020
620
1,770
730
157
170
75
24
Reference '
2q
2q
2q
2q
2q
2q •
2r
2r
2r
2r
2r
2r
2r
2r
Derived from Appendix F. Includes installing seal plate, gaskets, welds, access openings, blanking
lines, removing existing cover, and cleaning tank.
Derived from Appendix F. Based on replacing 52-m2 (560-ft2) primary cooler (tar) decanter top and
includes blanking lines; removing concrete cover; cleaning; installing steel plate, supports, access
openings, and vent pipe; and welding.
cFittings include els, tees, reducers, and -flanges.
Cost of fittings derived from Appendix F, which contains detailed construction estimate for one
plant. Based on costs of fittings per meter of pipe for this design.
From Grotn Equipment Company; see Appendix F.
^Includes flowmeter with low flow alarm with 2.5-cm (1-in) flange connections ($787), stainless
steel pressure indicator ($90), temperature gauge ($164) from distributor for Brooks Instruments
Division, Charlotte, NC. Installation cost of $130 per instrument is used.
^Includes presurvey, setup, laboratory preparation, analysis, report preparation, travel, and per
diem expenses.
Installed capital cost derived from Reference 2 with details in Appendix F. Includes installation
premium for area where continuing operations may interfere with work progress. For the light-oil
plant, includes cost premium for flanged pipe and installation premium for work in a hazardous area.
Costs for 2.5-cm (1-in.) and 5.1-cm (2-in.) pipe include fittings. All pipe is Schedule 40.
From Appendix F.
••Includes a pressure sensor, control valve, and alarm; from BGV Controls, Inc., distributor for
Fisher Controls, Charlotte, NC.
kFrom Appendix F. Includes two Garlock "Gar-Seal" 100 butterfly valves, Teflon-coated surfaces
including disc and valve liner, two General Torque valve actuators, chemical seal, and Robertshaw
digital control modules with electronic differential pressure transmitter and electropneumatic
relays.
Reduces gas supply pressure to 380 to 460 mm (15'to.lS in.) of water; from BGV Controls, Inc.,
distributor for Fisher Controls, Charlotte, NC.
BFrom Appendix F. Estimate provided by the Mueller Company.
"From Appendix F. Pumps rated as 2.2 i/s at 23-m head (35 gal/min at 75 ft) with a 2-hp motor.
includes pump ($1,350), foundation ($390), and electrical ($830).
°The design is for a flow of 0.15 mVs (310 ftVm) and wash-oil rate of 0.3 je/s (4.3 gal/min).
shell is a pipe with a 0.5-m (1.5-ft) diameter and a 4.9-m (16-ft) length. See Appendix F.
pThe design is for a flow of 0.6 m3/s (1,200 ftVm) and a wash-oil rate of 2.24 £/s (35 gal/min).
The shell is a pipe with a 0.9-m (3-ft) diameter and a 7.3-m (24-ft) length. See Appendix F.
^Derived from Reference 2 in Appendix F. Includes 1.3-cm (0.5-in.) Schedule 80 pipe, valves, steam
traps, insulation, and stainless steel jacket. Hazardous area installation premium included for the
light-oil plant. Insulation is 5.1 cm (2 in.) thick for 15-cm (6-in.) pipe, 3.8 cm (1.5 in.) thick
for 7.6- or 10-cm (3- or 4-in.) pipe, and 2.5 cm (1 in.) thick for the steam supply line.
rDerived from Reference 2 in Appendix F. Includes hazardous area installation premium in the light-
oil plant.
Cost
The
8-6
-------
TABLE 8-3. ANNUALIZED COST ITEMS
(1982 Dollars) ....
Item
Benzene credit, as fuel
Benzene credit, recovered
Capital recovery (10 years at 10%)
Electricity
Light-oil credit
Maintenance
Nitrogen (storage and supply)
Overhead
Steam
Taxes, insurance, and administration
Cost
$0.15/kg
$0.47/kg
16.3% of capital
$0.04/kWh
$0.33/kg
5% of capital
$0.27/m3
(0.76/100 ft3)
80% of labor
$17.6/Mg
4% of capital
Reference
4,5a
6
7
4b
c
8
d
d
8
4e
8
Fuel value is based on underfire gas at $2.76 per million Btu's from
Reference 4 in 1979 dollars ($4.00 per million Btu's in 1982 dollars);
a fuel content of 17,500 Btu's/lb in Reference 5.
Adjusted from value of $0.027/kWh (1979 dollars) in Reference 4.
A light-oil credit equal to 70 percent of the benzene value is used.
In Reference 4, the 1979 value of light oil was given as $0.77/gal
and the value of benzene as $1.15/gal.
Includes rental of 5.7-m3 (1,500-gal) liquid nitrogen storage tank,
vaporizer, controls, and nitrogen. Estimate provided by National
Welders Supply Company, Inc., Raleigh, NC.
Adjusted from value of $12/Mg ($5.44/1,000 Ib) in 1979 dollars in
Reference 4.
8-7
-------
The major cost elements for the blanketing system include covering
and sealing the decanter and sump, installing a pressure tap upstream
of the Askania, and adding the steam-traced and insulated piping to
route emissions to the collecting main. Costs for covering and sealing
include removing the existing concrete top; blanking the lines; clean-
ing, inspecting, and repairing the tank; installing steel plate,
supports, and gaskets; welding; and adding access openings and a vent
pipe. The tar decanter is sealed by a water seal plate near the
sludge conveyor discharge, as illustrated in Figure 8-1. The majority
of the liquid surface is blanketed with gas from the collecting main
and the remainder (approximately 13 percent) provides clearance for
the sludge conveyor and is open to the atmosphere.
A 20-cm (8-in.)-diameter vent line is used to carry the blanketing
gas and to route displaced emissions to the collecting main. The
large-diameter line is used to lower the pressure drop in the vent
line and, consequently, to minimize pressure on the tar decanter.
Included with the vent line is a 1.3-cm (0.5-in.) line for steam
tracing, 5.1 cm (2 in.) of fiberglass insulation, and a stainless
steel protective jacket. The steam tracing should avoid condensation
and accumulation in the vent lines. However, vent and drain connections
are provided for steaming out the line should the need arise.
Each vessel is equipped with three-way cast iron lubricated plug
valves to prevent sticking because of tar deposits. Valve connections
are arranged so that in one position the tank is vented to the collect-
ing main and in the other position the tank is vented to the atmosphere.
This arrangement permits the blanketing line and/or the tank(s) to be
isolated for performing maintenance and ensures that the tank is vented
at all times. In either position, the plug valve provides a clear
opening for the passage of vapors and does not have pockets where tar
may accumulate and interfere with the opening and closing of the valve.
Capital and annualized cost estimates are summarized in Table 8-4.
A range of piping distances is given for each model plant to represent
variations in plant layouts. Some plants will be able to use an
8-8
-------
CO
u
9)
•a
v.
TO
co
£
_0)
u_
8-9
-------
TABLE 8-4. COSTS FOR GAS BLANKETING OF TAR DECANTER, TAR-INTERCEPTING
SUMP, AND FLUSHING-LIQUOR CIRCULATION TANK
(All Costs in 1982 Dollars)
Model Plant 1
Cost element
Pressure taps
20-CB (8-in. ) pipe, m
(ft)
7.6-on (3-1n. } pipe, m
(ft)
Pipe supports
Three-way valves
20-cm (8-1n.) plug valve
Clean, cover, seal decanter, m2
(ft2)
Clean, cover, seal sump, m2
(ft2)
Capital cost
Total capital costf
Annual iz«d costs
Maintenance, overhead9 (9%)
Utilities'1
Taxes, insurance (4%)
Capital recovery1 (16.3%)
Total annual ized cost
Light-oil credit3
Annual ized cost
Benzene reduction (Hg/yr)
Cost effectiveness ($/Mg)
Minimum
1
61
(200)
46
(150)
0
4
1
0
(0)
3.0
(32)
34,200
48,300
4,300
1,900
1,900
7.900
16,000
30,100
(14,100)
63.9
(220)
Maximum
1
122
(400)
91
(300)
11
4
1
149
(1,600)
3.0
(32)
121,000
171,000
15,000
3,800
6,800
28.000
53,600
30.100
23,500
63.9
370
Model Plant 2
Minimum
1
91
(300)
46
(150)
0
6
1
0
(0)
23
(250)
52,700
74,300
6,700
2,600
3,000
12,000
24,300
121. QOO
(96,700)
256
(380)
Maximum
1
366
(1,200)
91
(300)
21
6
1
223
(2,400)
23
(250)
239,000
337,000
30,000
9,700
13,000
55,000
; 108, 000
121,000
(13,000)
256
(50)
Model Plant 3
Minimum
1
183
(600)
91
(300)
0
10
1
0
(0)
46
(500)
97,100
137,000
12,000
5,300
5,500
22,000
44,800
271.000
(226,000)
575
(390)
Maximum
1
457
(1,500)
183
(600)
32
10
1
446
(4,800)
46
(500)
377,000
532,000
48,000
12,800.
21,000
87,000
169,000
271.000
(102,000)
575
(180)
Cost per
• unit
. 4,000a
280b.
(85.3)°
99. lc^
(30.2) '
l,500d
1,660
1,020
327e
(30.5)e
328eo
(30.5)e
From Table 8-2; one-half of rental ($2,250) plus labor and materials ($1,750).
From Table 8-2; includes installed pipe ($138/m or $42.I/ft), fittings ($20/m or $6.10/ft), steam tracing,
and insulation ($122/m or $37.I/ft).
cFro* Table 8-2; includes installed pipe ($46.6/m or $14.2/ft), fittings ($7.0/m or $2.13/ft), steam
tracing, and insulation ($45.6/m or $13.9/ft).
Assumes some plants may add pipe supports for 25 percent of pipe; one column each 6.1 m (20 ft) for 20-cm
(8-in.) pipe and each 3.7 m (12 ft) for 7.6-cm (3-in.) pipe.
eAssum«s some plants have existing covers and others do not. The cost is averaged from Table 8-4 ($346/m2
or $32.5/ft2 and $308/m2 or $28.6/ft2).
Total capital cost includes construction fee (10 percent), contingency (15 percent), engineering (15 per-
cent), and startup (1 percent).
^Maintenance and overhead are 5 and 4 percent of capital, respectively.
hSteam at $17.6/Mg.
Capital recovery factor for 10-year lifetime at 10 percent.
•'Light-oil credit at $0.33/kg ($0.15/lb).
8-10
-------
existing cover on the decanter and sump, while others must install a
new cover and seal. For some plants, the piping may be run on the
racks supporting the flushing-liquor line, and in other cases new pipe
supports may be required. Both of these conditions are included in
minimum and maximum estimates for the model plants.
8.1.1.3 Excess Ammonia Liquor Tanks. Two control options were
considered for emissions from the excess ammonia liquor storage tanks:
gas blanketing and wash-oil scrubbers. Depending upon the location of
the storage tanks, a blanket of coke oven gas from either the collecting
main or gas holder can be used to control emissions. The cost estimate
provided in Table 8-5 includes a range of piping distances to generate
a range of costs for each model plant. The system's design features
are similar to those described in Subsection 8.1.1.2.
The cost of a wash-oil vent scrubber is provided in Table 8-6.
The design is based on each tank venting at a rate of 0.013 m3/s
(200 gal/min) and the scrubber shell requirements discussed in Appen-
dix F. For Model Plant 2, the wash-oil rate would be approximately
0.1 £/s (1.6 gal/min) and the scrubber shell would be 0.3 m (1 ft) in
diameter and 3.7 m (12 ft) in length. Wash oil would be supplied
through an uninsulated 2.5-cm (1-in.) line and would be removed through
a 5.1-cm (2-in.) drain line. A range of piping distances is given for
each model plant. In addition, pumps may be required at some plants
to move the wash oil, and other plants may use existing wash-oil pumps
and gravity drain to recycle the wash oil.
8.1.1.4 Light-Oil Plant. The light-oil plant processes benzolized
wash oil from the wash-oil scrubbers, recovers the light oil, and
recycles the wash oil. Some plants produce only the crude light oil,
others refine the light oil into primary and secondary light oil, and
a few plants refine it further to produce pure benzene. The major
equipment items emitting benzene in the light-oil plant are the light-
oil condenser, wash-oil decanter, and wash-oil circulation tank.
(Product storage tanks are discussed separately in Subsection 8.1.1.5).
The control technology discussed in Chapter 4 for the light-oil
plant is gas blanketing with clean coke oven gas from the gas holder
8-11
-------
TABLE 8-5. COSTS FOR GAS BLANKETING AMMONIA LIQUOR STORAGE TANKS
(All Costs in 1982 Dollars)
Model Plant 1
Cost element
15-cra (6-in. ) vent pipe, m
(ft)
Three-way valves
15-cm (6-in.) plug valve
Pip* supports
Capital cost
Total capital cost0
Annuali zed costs
Maintenance, overhead (9%)d
Utilities"
Taxes, insurance (4%)
Capital recovery (16.3%)f
Total annual i zed cost
Light-oil credit9
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Hg)
Minimum
46
(150)
1
1
0
11,100
15,700
1,400
840
630
2,600
5,470
1,500
3,970
3.22
1,200
Maximum
•152
(500)
1
1
7
42,200
59,500
5,400
2,800
2,400
9,700
20,300
1,500
18,800
3.22
5,800
Model Plant 2
Minimum
61
(200)
3
1
0
17,400
24,500
2,200
1,100
1,000
4,000
8,300
6,000
2,300
12.8
180
Maximum
183
(600)
3
1
9
54,400
76,800
6,900
3,400
3,100
12,500
25,900
6,000
19,900
12.8
1,600
Model Plant 3
Minimum
91
' (300)
6
1
0
28,300
39,800
3,600
1,600
1,600
6.500
13,300
13,700
(400)
29.0
(14)
Maximum unit
305 193*
(1,000) (58.9)a
6 1,660
1 620
15 l,500b
92,000
130,000
12,000
5,600
5,200
21,000
43,800
13,700
30,100
29.0
1,040
aFroa Table 8-2; includes installed pipe ($109/m or $33.2/ft), fittings ($16/m or $5.00/ft), steam tracing,
and insulation ($68/m or $20.7/ft).
Assumes some plants may add pipe columns for 25 percent of pipe; one column each 5.2 m (17 ft) for 15-cm
(6-in.) pipe.
Total capital cost includes construction fee (10 percent), contingency (15 percent), engineering (15 per-
cent), and startup (1 percent).
Maintenance and overhead are 5 and 4 percent of capital, respectively.
eSteam at $17.6/Mg.
Capital recovery factor for a 10-year lifetime at 10 percent.
9Light-oil credit at S0.33/kg ($0.15/lb).
8-12
-------
TABLE 8-6. COSTS FOR WASH-OIL VENT SCRUBBER FOR AMMONIA LIQUOR STORAGE TANKS
(All Costs in 1982 Dollars)
Model Plant 1
Cost element
Scrubber shell , ra2
(ft2)
7.6-cm (3-in.) vent pipe, m
(ft)
2.5-cm (1-ih.) wash-oil line, m
(ft)
5.1-crn (2-in.) wash-oil drain, m
(ft)
7.6-cm (3-in.) vent valves
Pumps
Instrumentation
Performance test
Capital cost
Total capital costd
Annual i zed costs
Maintenance, overhead (9%)e
f
Utilities
Taxes, insurance (4%)
Operating labor^
Capital recovery (16.3%)h
Total annual ized cost
Light-oil credit1
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Hg)
Minimum
2.6
(28)
9.1
(30)
30.5
(100)
30.5
(100)
1
0
1
1
12,600
17,800
1,600
-
710
4,200
2,900
9,400
1,400
8,000
2.96
2,700
Maximum
2.6
(28)
9.1
(30)
152
(500)
152
(500)
1
2
1
1
24,000
33,800
3,000
150
1,400
4,200
5,500
14,300
1,400
12,900
2.96
4,400
Model Plant 2
Minimum
3.4
(37)
46
(150)
61
(200)
61
(200)
3
0
1
1
17,500
24,700
2,200
-
990
4,200
4,000
11,400
5,600
5,800
11.8
490
Maximum
3.4
(37)
46
(150)
152
(500)
152
(500)
3
2
1
1
27,300
38,500
3,500
290
1,500
4,200
6,300
15,800
5,600
10,200
11.8
860
Model Plant 3
Minimum
4.6
(50)
91
(300)
122
(400)
122
(400)
6
0
1
1
25,200
35,500
3,200
-
1,400
4,200
5,800
14,600
12,500
2,100
26.6
79
Maximum
4.6
(50)
91
(300)
305
(1,000)
305
(1,000)
6
3
1
1
42,200
59,500
5,400
510
2,400
4,200
9,700
22,200
12,500
9,700
26,6
360
Cost per
unit
226
(21)
46.6
(14.2)
20 2a
£-W. t-
(6.17)a
30. 7a
(9.36)a
, 730
2,570b
1,300C
8,000
Includes fittings.
Assumes some plants use existing wash-oil supply and gravity drain; other plants require pumps.
''Includes flowmeter with alarm ($920), pressure gauge ($120), and temperature gauge ($290).
lent) Candtstartut (l^ercent)"51™0110" *** (1° percent)> Cont1'n9ency (15 percent), engineering (15 per-
p
Maintenance and overhead are 5 and 4 percent of capital, respectively.
Electricity at $0.04/kWh.
aFor 30 min/day at $23/hr.
_Capital recovery factor for 10-year lifetime at 10 percent.
1 Light-oil credit at $0.33/kg ($0.15/lb).
8-13
-------
or battery underfire system. The gas blanketing technology has been
demonstrated in'the light-oil plant for at least three by-product
recovery plants. Pressure control is provided at 380 to 460 mm (15 to
18 in.) of water by the existing pressure controller on the gas holder.
Excess pressure in the gas holder is prevented by a bleeder control
valve and, in addition, many gas holders have a water seal that will
blow at about 500 mm (20 in.) of water.
The blanketing system consists of a 15-cm (6-in.) header from the
gas holder to the light-oil plant with 10-cm (4-in.) vent lines connect-
ing the equipment to the header. All lines are heat traced, insulated,
and provided with steam-out connections and drains. Three-way valves
allow the tanks to be vented either to the blanket line or to the
atmosphere for isolating and maintaining the equipment. Flame arresters
are included in the atmospheric vent lines, although some plants
already have flame arresters in place and others operate routinely
without them. A pressure tap will be made either at the gas holder or
on the battery underfire gas line.
A range of costs for gas blanketing the light-oil plant is given
in Table 8-7 for a range of piping distances at the model plants.
Light-oil credit for this system is based on the light oil's fuel
value because the light oil is returned to the coke oven gas, which is
burned. Some plants with gas blanketing of the light-oil plant have
observed decreased sludge formation, which occurs from oxidation
reactions with oxygen in the air. No estimates of the credits associ-
ated with reduced fouling, reduced maintenance, and reduced hazardous
waste disposal costs are available.
8.1.1.5 Light-Oil and BTX Storage Tanks. Two control options
were evaluated for emissions from light-oil and BTX product storage
tanks: gas blanketing and wash-oil scrubbers. Light-oil storage
tanks can be blanketed with clean coke oven gas from the gas holder or
battery underfire as described for the light-oil plant (see Subsection
8.1.1.4). For storage tanks that are sufficiently close to the light-
oil plant, the same header line from the gas holder may be used for
both the light-oil plant and the storage tanks.
8-14
-------
TABLE 8-7. COSTS FOR GAS BLANKETING OF LIGHT-OIL CONDENSER, LIGHT^IL DECANTER,
>• lt WJBOBRi Af»D ftkCtftftftt AN
(All. Costs in 1982 Dollars)
Cost element
Pressure tap
10- to 15-cm (4- to 6-in.) pipe,
Plug valve
Three-way valves
Flame arrestors
Capital costs
Total capital costs0
Annual ized costs
Maintenance, overhead (9%)
Utilities8
Taxes, insurance (4%)
Capital recovery (16.3%)
Total annual ized cost
Light-oil credit*'
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)
Model Plant 1
Minimum
1
m 61
(ft) (200)
1
6
6
30,800
43,500
3,900
1,000
1,700
7,100
'13', 700
7,400
6,300
34.6
180
Maximum
1-
183
(600)
1
6
6
64,400
90,700
8,200
3,100
3,600
14,800
29,700
7.400
22,300
34.6
640
Model Plant 2
Minimum
1
122
(400)
I
8
8
50,900
71,800
6,500
2,100
2,900
11,700
23,200
29,600
(6,400)
138
(46)
Maximum
1
244
(800)
1
8
8
84,400
119,000
10,700
4,100
4,800
19,400
39,000
. 29,600
' 9,400
138
68
Model Plant 3
Minimum
1
183
(600)
I
13
13
75,900
107,000
9,600
• 3,100
4,300
17,400
34,400
66,600
(32,200)
311
(100)
Maximum un.it
1 3,550a
305 275b,
(1,000) (83.8)D
1 620
13 730
13 920
109,000
154,000
13,900
5 , 100
6,200
25,100
50,300
66,600
(16,300)
311
(52)
From Table 8-2; equipment rental for 6 days ($1,800) plus labor and materials ($1,750).
bAssumes 75 percent of pipe is 15-cm (6-in.) header and 25 percent is 10-cm (3-in.) vent lines. Cost
includes installed pipe ($170/m or $51.9/ft), fittings ($23/m or $7.0/ft), steam tracing, and insulation
($81.7/m or $24.9/ft).
cTotal capital cost includes construction fee (10 percent), contingency (15 percent), engineering (15 per-
cent), and startup (1 percent).
Maintenance and overhead are 5 and 4 percent of capital, respectively.
eSteam at $17.6/Mg.
Capital recovery factor for 10-year lifetime at 10 percent.
9Light-oil credit of $0.15/kg as fuel.
8-15
-------
Gas blanketing costs for light-oil storage tanks are given in
Table 8-8 for a range of piping distances at the model plants. The
design features are the same as those described for the light-oil
plant. In addition, pipe columns are added for the maximum case
because a storage tank occasionally may be in a remote location without
existing overhead pipe racks. Light-oil credit again is based on its
fuel value instead of on the value of recovering light oil.
Costs of a wash-oil vent scrubber are provided in Table 8-9.
Design is based on a maximum vent rate of 0.013 ms/s (200 gal/min)
generated from pumping light oil into the tank and the scrubber shell
requirements discussed in Appendix F. Wash-oil scrubbers may be an
appropriate control for old, vertical storage tanks with a riveted
construction. Extensive modifications, such as replacing the roof on
the entire tank, may be required to rehabilitate the old, vertical
tanks to accept a positive-pressure gas blanket. However, a wash-oil
scrubber has a negligible pressure drop and could be installed as a
vent control without major tank modifications. The scrubber would be
installed beside the tank or mounted on the side of the storage tank.
Wash oil is supplied through an uninsulated 2.5-cm (1-in.) line
and would be removed through a 5.1-cm (2-in.) drain line. A range of
piping distances is given for each model plant. In addition, pumps
may be required at some plants to move the wash oil, and other plants
may use existing wash-oil pumps and gravity drain to recycle the wash
oil. Wash oil leaving the scrubber would be routed through the light-
oil plant for light-oil recovery.
8.1.1.6 Tar-Collecting, Tar Storage, and Tar-Dewatering Tanks.
Costs for two control options—gas blanketing and a wash-oil scrubber—
were evaluated for tar collecting, storage, and dewatering tanks. A
blanket of coke oven gas from the collecting main can be used to
control emissions from tar tanks, as described in Subsection 8.1.1.2
for tar decanters. Cost estimates for the model plants are given in
Table 8-10 for a range of piping distances. The operational and
design features (insulated and heated line, pipe supports, and three-
way valves) are the same as those described for the tar decanter. The
8-16
-------
TABLE 8-8. COSTS FOR GAS BLANKETING OF LIGHT-OIL AND BTX STORAGE TANKS
(All Costs in 1982 Dollars)
Model Plant 1
Cost element
10- to 15-cm (4- to 6-in.) pipe,
Three-way valves
Pipe supports
Flame arresters
Capital costs
Total capital costs
Annual i zed costs
Maintenance, overhead0 (9%)
Utilities'1
Taxes, insurance (4%)
Capital recovery8 (16.3%)
Total annual i zed cost
Light-oil credit
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)
Minimum
m 18
(ft) (60)
2
0
2
8,300
11,700
1,100
290
500
1,900
3,790
500
3,290
2.08
1,600
Maximum
152
(500)
2
8
2
57,200
80,700
7,300
2,500
3,300
13,200
26,300
500
25,800
2.08
12,000
- Model Plant 2
Minimum
18
(60)
6
0
6
14,900
21,000
1,900
290
840
3,400
6,400
3,600
2,800
16.6
170
Maximum
213
(700)
6
11
6
85,100
120,000
10,800
3,500
4,800
19,600
38,700
3,600
35,100
16.6
2,100
Model Plant 3 .
Minimum
61
(200)
9
0
9
31,600
44,600
4,000
1,000
1,800
7,300
14,100
8,000
6,100
. 37.3
160
Maximum
244
(800)
9
12
9
99,900
141,000
12,700
4,100
5,600
23,000
45,400
8,000
37,400
37.3
1,000
Cost per
unit
275*
(83.8)a
730
l,500b
920
aAssumes 75 percent of pipe is 15-cm (6-in.) header and 25 percent is 10-cm (3-in.) vent lines. Cost
includes installed pipe ($170/m or $51.9/ft), fittings ($23/m or $7.0/ft), steam tracing, and insulation
($81.7/m or $24.9/ft).
Assumes some plants may add pipe columns for 25 percent of pipe. One column each 5.1 m (17 ft) for 15-cm
(6-in.) pipe and each 4.3 m (14 ft) for 10-cm (4-in.) pipe.
Maintenance and overhead are 5 and 4 percent of total capital cost, respectively.
dSteam at $17.6/Mg.
eCapital recovery factor for 10-year lifetime and 10 percent.
fLight-oil credit of $0.15/kg as fuel.
8-17
-------
TABLE 8-9. COSTS OF WASH-OIL VENT SCRUBBER FOR LIGHT-OIL AND BTX STORAGE TANKS
(All Costs in 1982 Dollars)
Model Plant 1
Cost element
Scrubber shell, m2
(ft2)
ID-cm (4-in. ) vent pipe, m
(ft)
2.5-cra (1-i n.) wash-oil line, m
(ft)
5.1-cro (2-in.) wash-oil drain, m
(ft)
Pump
Vent valves
Instrumentation
Performance test
Capital cost
Total capital cost6
Annual i zed costs
Maintenance, overhead (9%)
Utilities9
Taxes, insurance (4%)
Operating labor
Capital recovery (16. 3*)1
Total annual i zed cost
Light-oil creditj
Annual i zed cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Hg)
Minimum
3.0
(32)
15
(50)
30.5
(100)
30.5
(100)
0
2
1
1
16,000
22,600
2,000
190
900
4,200
3,700
11,000
900
10,100
1.91
5,300
Maximum
3.0
(32)
15
(50)
183
(600)
183
(600)
2
2
1
1
28,900
40,700
3,700
380
1,600
4,200
6.600
16,500
900
15,600
1.91
8,200
Model Plant 2
Minimum
4.6
(50)
91
(300)
30.5
(100)
30.5
(100)
0
6
1
1
34,200
48,200
4,300
1,100
1,900
4,200
7,900
19., 400
7,200
12,200
• 15.2
800
Maximum
4.6
(50)
91
(300)
213
(700)
213
(700)
2
6
1
. 1
48,700
68,700
6,200
1,400
2,700
4,200
11,200
25,700
7.200
18,500
15.2
1,200
Model Plant 3
Minimum
5.9
(64)
137
(450)
61
(200)
61
(200)
0
9
1
1
47,200
66,600
6,000
1,600
2,700
4,200
10,900
25,400
16,200
9,200
34.3
270
Maximum
5.9
(64)
137
(450)
244
(800)
244
• (800)
2
9
1
1
61,600
86,900
' 7,800
2,200
3,500
4,200
14,200
31,900
16.200
15,700
34.3
460
Cost per
unit
226
(21)
196a
(59.7)a
K
20. 2b,
(6.17)b
h
30. 7D,
(9.36)b
2,570C
730
,t
1,300°
8,000
-
Includes installed pipe ($126/m or $38.4/ft) and steam tracing with insulation ($70/m or $21.3/ft)
blncludes fittings.
cAssu«es that some plants use existing wash-oil supply and gravity drain and that other plants require
pumps.
dlnc1udes flowmeter with alarm ($920), pressure gauge ($120), and temperature gauge ($290).
eTotal capital cost includes construction fee (10 percent), contingency (15 percent), engineering (15 per-
cent), and startup (1 percent).
^Maintenance and overhead are 5 and 4 percent of capital, respectively.
9Steam at $17.6/Mg and electricity at $0.04/kWh.
hFor 30 min/day at $23/h.
i
Capital recovery factor for 10-year lifetime at 10 percent.
•^Light-oil credit at $0.33/kg.
8-18
-------
TABLE 8-10,
COSTS FOR GAS BLANKETING OF TAR COLLECTING, STORAGE, AND
DEWATERING TANKS
('All Costs in 1982 Dollars)
Model Plant 1
Cost element
15-cm (6-in.) pipe, m
(ft)
Pipe supports
Three-way valves
Capital cost
Total capital cost
Annual ized costs
Maintenance, overhead (9%)
Utilities6
Taxes, insurance (4%)
Capital recovery (16.3%)
Total annual ized cost
Light-oil credit9
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)
Minimum
61
(200)
0
5
20,100
28,300
2,500
1,100
1,100
4,600
9,300
5.600
3,700
11.8
310
Maximum
152
(500)
7
5
48,300
68,000
6,100
2,800
2,700
11,100
22,700
' 5.600
17,100
11.8
1,500
Model Plant 2
Minimum
91
(300)
0
10
34,300
48,300
4,300
1,600
1,900
7.900
15,700
22,200
( 6,500)
47.2
(140)
Maximum
762
(2,500)
37
10
219,000
309,000
27,800
13,900
12,400
50,400
104,500
22,200
82,300
47.2
1,700
Model Plant 3
Minimum
122
(400)
0
16
50,100
70,700
6,400
2,200
2,800
11,500
22,900
50,000
(27,100)
106
(260)
Maximum Unit
914 193aa
(3,000) (58.9)a
h
44 1,500°
16 1,660
269,000
380,000
34,200
16,700
15,200
61,900
128,000
50,000
78,000
106
740
Includes installed pipe ($109/m or $33.2/ft), fittings ($16/m or $5.0/ft), steam tracing, and insulation
($68/m or $20.77ft).
bAssumes some plants may add pipe supports for 25 percent of pipe; one column each 5.2 m (17 ft) for 15-cm
(6-in.) pipe.
cTotal capital cost includes construction fee (10 percent), contingency (15 percent), engineering (15 per-
cent), and startup (1 percent). „
"^Maintenance and overhead are 5 and 4 percent of capital, respectively.
eSteam at $17.6/Mg.
fCapital recovery factor for 10-year lifetime at 10 percent.
9Light-oil credit of $0.33/kg ($0.15/lb).
8-19
-------
cost of a pressure tap is not included because the 15-cm (6-in.)
header for the tar tanks will tie into the gas blanketing line for the
tar decanter.
The cost of a wash-oil scrubber for control of emissions from
tar-dewatering and tar storage tanks was also examined. Because the
vapors from these sources are hot, the vapors must be cooled to obtain
a reasonable control efficiency from absorption in the wash oil. A
high flow rate of once-through wash oil was considered for these
sources to effect both cooling and absorption, but this design could
require increasing the existing wash-oil still capacity at some plants.
The high wash-oil flow rate would be required because of the heat
content of the vapors, primarily from removal of the latent heat of
water that is present in the emissions.
An alternate design is presented in Figure 8-2, which is a
conceptual design of a wash-oil condenser and scrubber that would
require a relatively low usage of wash oil. The design includes a
two-zone scrubber in which initial cooling and absorption are accom-
plished in the bottom zone and additional absorption is accomplished
in the top zone. On the scale of Model Plant 2, cooled wash oil would
be sprayed into the bottom zone at 16 SL/s (250 gal/mi n), and the wash
oil and condensed water would enter the separator. Water would be
separated and sent to wastewater treatment. The wash oil from the
separator would be circulated through an indirect contact heat exchanger
for cooling and then recirculated to the bottom spray zone. A slip-
stream o'f wash oil at 0.3 iL/s (5 gal/mi n) would be sent to the light-oil
recovery process for removal of organics. Fresh wash oil would be
sprayed into the top zone of the scrubber at 0.3 iL/s (5 gal/min) to
remove benzene vapors, which pass through the cooling section of the
scrubber.
The capital cost for this design as applied to Model Plant 2 is
given in Table 8-11. Annualized costs for the three model plants are
given in Table 8-12. The capital costs for Model Plants 1 and 3 were
estimated from Model Plant 2 by scaling the capital cost on the basis
of capacity to the 0.6 power.
8-20
-------
CC —
ol U
01
CC
CC
01
CO
@
O
UJ W
1
XC 1
i i CC
i *e
1 1
o
CO
i
<-
->1^ —
X
^
-'--
**
> CO
< 01
gri
CO N
0
1 2
1 ff>
\
L.__J
_i
0
CC
\-
o
u
_l
01
s 5
ca -1
03
3
CC
o
CO
L
O
o
o>
en
^ H S @
"{ •£ s
dco" o.
£ S "
°2 •* S
< "8
0 5£
CM
+j
CL
"33
"D
o
S
15
09
CO
o
t;
CO
03
+J
CO
1
•a
I
o
C
•a
"co
3
S-
u
o
o
CM'
06
1
D)
IT
8-21
-------
CD
W
•a o
u
ro
•P -P
UJ •!-
c a.
H CO
u
o
CM
ro
un
ro ro ,
O O
CD O
O 00
CM O
CO rH
u -o -a
000
en CM CM
0
00
CM CM
0
CM
CM
a
o
ID
r--
CM
CO
o
co
CO
a>
CO
UJ
CQ
ca
o
to
UJ CM
a
CO
a>
i— in o
CO •!-
_e CD -p
in c u
o co
a>
a)
•r- C
s- o
O -P CD CO
to o 3 e
DO S- U
co u i
U) ro
10
cu co
u
en
c
CO
u
X
CO
(0
CO
10
en
ro •> in CM
rH O
V> 01
o
-p
ro
ro
a.
01
- O
S- O
>> -P CD OO
O C •!-
ro
C tO N •!-
I I
o
(fl CO
CO
to s- o in a. -P
+> 3 S rH ro ro
to in I— v^ c_5 a;
O CO
(0
E -P
CM CO CO
to
cu
Q.
a.
co co
a. >
•i- CO
a. i—
CM
O)
CD
O)
CO
CM
CO
a. -P
u
to
Q.
CO O)
CO
o
to
CO r— O. -P <0
a> at
to
S-
<0
CO
to
•I—1 »r~ (O
co
o co to
TO U
a. 3:
Q. -P
CO CO
ro •!- i—
-p co co
•a -p •!-
3: >
001—
•p -a
o c
to ro
•r- U
•P to
ro co
tO CM
•P rH
•a co
c en
tO 3
tO
cr> en
O) CO
O S-
C 3
co 3 O) CM i-H
••- r— CO -fag- -tfl-
S- O 3
CD c to -p -P
H-1 CO
ro .a
u
8-22
-------
8.1.1.7 Light-Oil Sump. Emissions from the light-oil sump are
controlled by providing a steel sump cover with a vertical vent. The
edge of the sump cover rests in a trough around the sump's edge and is
sealed with gasket material.
Costs for covering the sump are estimated in Table 8-13 for
different sizes of sumps at the model plants. The unit cost for the
cover installation is derived in Appendix F and includes replacing the
existing cover; blanketing lines; cleaning; adding gasket material;
installing the sump cover, supports, and access hatches; and welding.
8.1.1.8 Pure Benzene Storage Tanks. A coke oven gas blanketing
system was considered for pure benzene storage tanks, but plant opera-
tors indicated that contamination may result from contact of the coke
oven gas with pure benzene. This cost analysis is based on supplying
a nitrogen or natural gas blanket to pure benzene storage tanks and on
returning vented emissions to the gas holder or battery underfire
system. Some coke plants that are part of an integrated steel plant
may have excess nitrogen available from the oxygen plant associated
with steel making. Most coke plants have a source of clean natural gas
that is used to supplement the coke oven gas; to replace the coke oven
gas in emergency situations; or to underfire the coke ovens during
startup, idle, or controlled shutdown of the coke battery. The cost
analysis also recognizes that a few plants may have neither nitrogen
nor natural gas available and would incur an annual expense for
purchasing nitrogen.
Costs of gas blanketing controls for pure benzene storage tanks
are summarized in Table 8-14 for Model Plant 3. The system design
includes a pressure reducer to supply the gas blanket at a pressure of
380 to 460 mm (15 to 18 in.) of water, a pressure controller that will
open and vent to the gas holder at pressures over 460 mm (18 in.) of
water, three-way valves for isolating tanks, and flame arresters.
When liquid was pumped out of the storage tank, nitrogen or
natural gas would fill the vapor space in the tank. When liquid was
pumped into the tank, excess pressure in the vapor space would be
vented through the pressure controller to the gas holder. The pressure
8-23
-------
TABLE 8-12. ANNUALIZED COST ESTIMATES FOR A WASH-OIL CONDENSER
AND SCRUBBER FOR TAR STORAGE AND DEWATERING
(1982 dollars)
Cost element
Capital cost
Annual ized costs
El ectri ci ty
Cooling water9
Maintenance, overhead (9%)
Taxes, insurance (4%)
Operating labor
Capital recovery (16.3%)c
Total annual ized cost
Light-oil credit
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)
Model
Plant 1
118,000
500
5,500
10,600
4,700
4,200
19,200
44,700
5,100
39,600
10.8
3,700
Model
Plant 2
275,000
1,000
22,000
24,800
11,000
4,200
45,000
108,000
20,500
87,500
43.4
2,000
Model
Plant 3
448,000
2,300
50,000
40,300
17,900
4,200
73,000
188,000
46,000
142,000
97.6
1,500
Based on 13 £/s (200 gal/min) for Model Plant 2 at $0.055/1,000 £
($0.21/1,000 gal) from Reference 4 in 1982 dollars. Flow rates for
Model Plants 1 and 3 were scaled from Model Plant 2 based on coke
capacity.
bFor 30 min/day at $23/h.
cCapital recovery factor for 10-year lifetime at 10 percent.
dLight-oil credit at $0.33/kg ($0.15/lb).
8-24
-------
1
_,
•1-H
O
J— /*"•%
3: co
O 1-
I-H 10
r—
C3 O
•z. o
Qi CM
UU 00
^> en
CD rH
CJ
o "~
LJ_ U)
-p
co in
1— 0
CO C-3
CD
O r—-
CO
00
LLJ
—i
?9
'^•_
r~"
j_
a
£
•P
V
r
E
^S
c
•r-
sz
ximum
10
nimum
•I—
,
-p
cu
cu
r—
cu
•p
in
o
CJ
tO
ro < — > <•"•»
CM LO ID CM
CD ID «*
OO ^J- rH
ID O rH O O O
OO O 'OO CD O
rH O Cn v^/ "4- ID
CM rH ID
v^/ 1 f*t OQ
t-^. CM rH O 00
. f^. -CO O O
CM 00
OO O ID LO CD O
Cn O • rH O O
CD "5^ *^-^ I*** OO
r4" 0 00
v-^ OO ^*
OO ID ID LO O O
. CO • rH O LO
CO v_^ «4" *— •* OO 00
t-T rH
>»*^ ^*N
cn LO ID LO o o
• CM . I-H 0 CD
O CM <*• -*-* rH O
CM ^^^ ** **
r~- o
rH
CO ID ID LO O O
.00 • rH 0 UO
OO v> «tf- v_/ OO OO
*v H
rH rH
^*N
/*^s 4-J
,M P It-
VN C ^^
SU« f,
«|H "
S^^ Q)
« Q.
rd CL
CD
cn -P -Q
c: -P
*O OJ V)
c > o to
m o -P
*» . r™ O
£_ E -P IO U
CU -i- co -P
> 1 O T- "O
o co (j a. cu
O v — • tO N
1 O T-
E 10 i—
E O 4J i— tO
tO 1 «r- to 3
CU ID O. -P E
r- . 10 0 C
cj r-. cj f— 0
cn «* ID
«
rH
CD O O
1^- r-» O
rH 00
o
^S TD
^— / 5<
or
TU ^*^N
tO ^ C£
CU *>!" r~
f— v^y s^.
J_
U t
o E a
tO 5
- s- c
CU 3 L
(j in a
C E t
tO -r-
c: i—
cu n
-P in +.
E CU -r
•r- X C
tO 10 n
2E 1— C.
0 0
o o
•* CO
LO CM
CM CM
0 0
S §
rn" CM"
CM
0 0
O O
r** rH
CM 0
rH rH
0 0
•* CD
in rH
O
i — 1
I 0 0
CD CD
cn uo
J ., .
1 CM CM
1 0 0
•3- 0
1 un in
*
CM
s
)
)
. -P
> in
1 0
•> o
0)
^ -a +•
CU *r-
> N "C
> •!- a
> i— i-
) tO I.
J 3
E i-
to c
3 I
J i— 4-
tO X
i. -P C
3 0 T
31—-
CD OO ^t"
O • LO
ID 00
CM
O CO O
O - LO
f^s. 00 ^d™
rH
CM
CD LO O
O • CM
ID rH i— 1
CM
( 1 /^N
O LO O
0 • LO
ID rH «*
- CM «— '
cn
"^^
i o r-. •*
1 o ro r-»
| LO
^ ^
10 r-- o
o oo r-.
o • oo
^ LO ^
CM
^^
I**"** *•"""%
S- D)
**^Si *"*"*>•
cn -fee-
**^s
co
E CO
> O cu
•P *r— a
J in -P cu
> o o >
- 03-1-
; "d -P
T3 CU 0
cu s- cu
N f-
> •!- CO <«-
r- E CU
5 tO CU
: 3 N -P
7) C. C tO
c
m «
Q) Q/
v35 5*
J_3
-p u
c cu
0 O.
U in
cu
S- 4J
-P » CU
C r- U
cu to s-
U -P CU
i- T- CL
CU CL
Q. to O
U rH
CD
*~J 0 tO
OJ -P CU
CU E E
•*- CU T-
U -P
C S- CU
o cu i- /-•»
«r— p> *r" .O
u -^i- ' \
3 S- LO
S- -O 10 rH
•P coj.
in to >> O
E . i *e-
O s*"*' LO CD *— >
0 -P rH
E CD C35
m cu s- J- -*:
CU U tO O "V.
-as- it- oo
3 cu -a oo
. r— O. tO J-
CM O CU O CD
i c rH x: -p •fee-
00 -r- s_/ S_ U
cu to -P
cu -p a. > <*- to
i— LO 3 O
JD 0 -P >> -P
to o s- -a i. T-
i — to c a> ~o
i— .p to > a)
e to m o s-
o -P cu o o
t- •<- -a u cu
4- Q. C E S- i—
tO tO tO *t-~
CU U E F— 0
O) » CU to 1
(0 r- /-> +J ^-J -P
s- to -p c -i- j=
cu -p E T- a. o)
> O CU 10 10 -i—
«C 1— U S 0 — 1
tO JO 0 T3 CU
8-25
-------
TABLE 8-14. COSTS FOR NITROGEN OR NATURAL GAS
BLANKETING OF PURE BENZENE STORAGE TANKS
(All Costs in 1982 Dollars)
Model Plant 3 rn*+ „„„
Cost element
2.5-cm (1-in. ) gas supply, m
(ft)
7.6-cffl (3-in.) vent pipe, m
Pressure controller
Pressure reducers
Three-way valves
Flame arresters
Pipe supports
Capital costs
Total capital costs6
Annuali zed costs
Maintenance, overhead (9%)
Utilities9
Taxes, insurance (4%)
Capital recovery (16.3%)
Total annual i zed cost
Benzene credit1
Annuali zed cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)
Minimum
30.5
(100)
61
(200)
1
2
3
0
14,300
20,100
1,800
-
800
3,300_
5,900
2,900 '
3,000
18.7
170
Maximum unit
91.4 20.2
(300) (6.17)d
244 53.5bb
(800) (16.3)D
1 4 400C
2 525
3 730
3 920
O wt-w
H
10 1,500°
40,300
56,800
5,100
15,000
2,300
9,300
31,700
2.900
28,800
18.7
1,500
Includes fittings.
Includes installed pipe ($46.6/m or $14.2/ft) and fittings ($7.0/m or
$2.13/ft).
cFrom Table 8-2; includes pressure sensor, control valve, and alarm.
Assumes some plants may add pipe columns and others may use existing
pipe supports.
6Total capital cost includes construction fee (10 percent), contingency
(15 percent), engineering (15 percent), and startup (1 percent).
Maintenance and overhead are 5 and 4 percent of capital, respectively.
9Nitrogen at $0.27/m3 ($0.76/100 ft") Includes rental of 5.7-m3
(1 500-gal) liquid nitrogen storage tank, vaporizer, and gas usage.
Some plants are assumed to have a nitrogen source and others must
purchase nitrogen.
hCapital recovery factor for 10-year lifetime at 10 percent.
Benzene credit of $0.15/kg as fuel.
8-26
-------
setpoint for the pressure controller would be slightly higher than
the pressure in the gas holder would be. The benzene vapors would be
returned to the coke oven gas that is used as fuel.
Costs of applying a wash-oil vent scrubber to benzene storage
tanks are summarized in Table 8-15. The system is analogous to the
wash-oil scrubbers previously described. Debenzolized wash oil is
sprayed into the top of the scrubber, and the wash oil is drained and
returned to the light-oil recovery system.
8.1.1.9 Final Cooler. In standard descriptions of by-product
plants, crude naphthalene is recovered from the hot well of the direct-
water final cooler.11 In a new plant, the tar-bottom final cooler
might be in one piece. Retrofit costs for an existing plant are based
upon the design of a one-stage mixer-settler expending pump work
comparable to the extra lift work of the one-piece design. This
system also would be suitable for new applications. The following
paragraph describes the parameters chosen for the cost estimation and
a rationale for their selection.
At a scale of 4,000 Mg of coke per day, with a 20° C increase
through the final cooler, approximately 4,800 Mg of water per day
contacts a comparable amount of tar. Daily production of whole tar is
about 160 Mg; for light tar, which is cleaner and less viscous, daily
production is approximately 30 Mg. If the light tar is recirculated
from the settler at a rate 100 times the throughput, the effective tar
rate is 3,000 Mg/day. If the combined stream is forced through an
orifice-plate mixer at a pressure drop of 70 kPa (10 psi), the
theoretical pump work is about 5.7 kW (7.6 hp). The electrical load
will be about 10 kW and a 15-hp motor should suffice. The settler
should provide a residence time of 30 minutes, requiring 300 m3
(10,000 ft3), with a vent back to the gas exiting the final cooler.
The water will be circulated from the settler to the cooling tower in
the usual way, but a pair of small circulating pumps and extra piping
are required for the tar circuit. Cost estimates, scaled to the three
model plants, are shown in Table 8-16.
8-27
-------
TABLE 8-15. COSTS OF WASH-OIL VENT SCRUBBER FOR
BENZENE STORAGE TANKS
(All Costs in 1982 Dollars)
Model Plant 3
Cost element
Scrubber shell , m2
(ft2)
2.5-cm (1-in.) wash-oil line, m
(ft)
5.1-cm (2- in.) wash-oil drain, m
(ft)
10-cm (4-in.) vent pipe, m
(ft)
Vent valves
Flame arresters
Pump
Instrumentation
Performance test
Capital cost
Total capital cost
Annuali zed costs
Maintenance, overhead (9%)
Utilitiesf
Taxes, insurance (4%)
Operating labor^
Capital recovery (16.3%)h
Total annual ized cost
Benzene credit1
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)
Minimum
3.4
(37)
61
(200)
61
(200)
45.7
(150)
3
3
0
1
1
23,900
33,700
3,000
-
1,300
4,200
5.500
14,000
5.700
8,300
17.2
480
Maximum
3.4
(37)
244
(800)
244
(800)
45.7
(150)
3
3
2
I
1
38,400
54,100
4,900
510
2,200
4,200
8.800
20,600
5,700
14,900
17.2
870
Cost per
unit
226
(21)
20. 2a
(6.17)a
30. 7a
(9.36)a
126
(38.4)
730
920
2,570b
1,300C
8,000
Includes fittings.
bAssumes some plants use existing wash-oil supply and gravity drain
while other plants require pumps.
clncludes flowmeter with alarm ($920), pressure gauge ($120), and
temperature gauge ($290).
dTota1 capital cost includes construction fee (10 percent), contingency
(15 percent), engineering (15 percent), and startup (1 percent).
Maintenance and overhead are 5 and 4 percent of capital, respectively.
Electricity at $0.04/kWh.
9For 30 min/day at $23/h.
hCapital recovery factor for 10-year lifetime at 10 percent.
Benzene credit at $0.33/kg ($0.15/lb).
8-28
-------
TABLE 8-16. COSTS FOR INSTALLING A TAR-BOTTOM FINAL COOLER
(All Costs in 1982 Dollars)
Cost element
Settler9
Mixer pumps, drivers3
Circulating pumps, drivers
Installed capital cost
Annual i zed costs
Maintenance, overhead (9%)
Utilities0
Taxes, insurance (4%)
Capital recovery (16.3%)d
Total annual ized cost
Light-oil credit6
Annual ized cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)
Model
Plant 1
76,000
16,000
9,300
101,000
9,100
900
4,000
16,500
30,500
52,800
(22,300)
112
(200)
Model
Plant 2
173,000
32,000
11,000
216,000
19,400
3,400
8,600
35,200
66,600
211,000
(144,000)
448
(320)
Model
Plant 3
280,000
43,000
12,000
335,000
30,200
7,700
13,400
54,600
106,000
475,000
(369,000)
1,010
(370)
of capital, respectively.
Installed costs, derived from Reference 12.
Maintenance and overhead are 5 and 4 percent
GElectricity at $0.04/kWh.
Capital recovery factor for 10-year lifetime at 10 percent.
eLight-oil credit of $0.33/kg ($0.15/lb).
8-29
-------
8.1.1.10 Wash-Oil Final Cooler. In principle, benzene emissions
from naphthalene handling and the direct final cooler can be eliminated
by one device: the wash-oil final cooler. As described in Chapter 4,
the cooling fluid is a suitable wash oil directly contacting the coke
oven gas. It is as assumed that the use of a suitable wash oil,
coupled with the use of appropriate additives and proper operating
conditions would permit easy separation of the condensed water from
the circulating oil in the system.
The cost estimation for a new system of this kind was furnished
by Wilputte in 1977, as reported by VanOsdell.13 Those numbers,
scaled to the three sizes of model plants and escalated to 1982 dollars,
are the basis of Table 8-17.
The least certain and, at the largest scale, the most significant
cost is for the wash-oil makeup. Although 0.1 percent loss is arbitrary
and sounds trivial, at the larger scale it tends to overwhelm the
annualized cost.
8.1.1.11 Fugitive Emissions from Equipment Components. This
subsection summarizes costs associated with controlling benzene emissions
from equipment components that service or contain materials having a
benzene concentration of 10 percent or more by weight. Exhausters
that handle coke oven gas with over 1 percent benzene also are included.
The light-oil recovery and refining processes at by-product recovery
plants use pumps, valves, pressure-relief devices, sampling connections,
and open-ended lines in benzene (or light-oil) service. Costs are
determined by following the methodology established to control volatile
organic compounds (VOC's) from the petroleum refinery industry.
Details are provided in Appendix F.
Two types of model plants were derived to estimate control costs
for equipment components in benzene service. Model Plants 1 and 2
represent the majority of by-product plants that produce light oil
(about 70 percent benzene), and Model Plant 3 represents plants that
not only recover light oil but also refine it into benzene. The
number of equipment items for each model plant is given in Table 8-18
and was derived from plant surveys and questionnaires.
8-30
-------
TABLE 8-17. COSTS FOR INSTALLING A WASH-OIL FINAL COOLER
(All Costs in 1982 Dollars)
Cost element
Model
Plant 1
Model
Plant 2
Model
Plant 3
Total capital cost, millions'
Annualized costs
2.1
4.8
7.9
Additional operating labor
/•»
Maintenance, overhead (9%)
Makeup wash oil
Utilities6
Taxes, insurance (4%)
Capital recovery (16.3%)
Total
Light-oil credit^
Annual i zed costg
Benzene reduction (Mg/yr)g
Cost effectiveness ($/Mg)g
Light-oil credit11
Annual i zed cost
Benzene reduction (Mg/yr)
Cost effectiveness ($/Mg)h
40,000
144,000
84,000
20,200
84,000
340,000
712,000
65,000
647,000
138
4,700
12,000
700,000
26
27,000
40,000
430,000
335,000
80,700
190,000
780,000
1,860,000
260,000
1,600,000
550
2,900
48,000
1,810,000
102
18,000
40,000
710,000
755,000
181,700
320,000
1,290,000
3,300,000
580,000
2,720,000
1,240
2,200
108,000
3,190,000
230
14,000
Updated and scaled from information by Wilputte Corporation in Refer-
ence 13.
bLabor in addition to that currently used for direct-water or tar-
bottom final cooler.
Maintenance and overhead are 5 and 4 percent of capital, respectively.
dAt 0.1 percent of circulation ($0.11/kg).
eElectricity at $0.04/kWh.
fCapital recovery factor for 10-year lifetime at 10 percent.
9Replaces direct-water final cooler; light-oil credit is $0.33/kg
($0.15/lb).
hReplaces tar-bottom final cooler; light-oil credit is $0.33/kg
($0.15/lb).
8-31
-------
TABLE 8-18. MODEL PLANTS FOR FUGITIVE BENZENE EMISSIONS FROM
EQUIPMENT COMPONENTS
Number of items at each model plant
Equipment item
Model
Plants 1 and 2C
Model Plant 3"
Exhausters
Pump seals
Valves
Pressure-relief devices
Sampling connections
Open-ended lines
6
15
105
'5
10
22
6
30
210
9
21
45
Model Plants 1 and 2 represent plants that produce light oil only.
DModel Plant 3 represents plants that produce light oil and pure
benzene.
8-32
-------
The cost analysis that was applied to the model plants evaluates
inspections, leak detection, repair, and equipment modifications as
controls for the equipment in benzene service. Capital cost items are
listed in Table 8-19 and were inflated to 1982 dollars at a 10-percent-
per-year rate. Total capital costs for the model plants in Table 8-19
were generated by multiplying the cost per item by the number of items
for each model as listed in Table 8-18.
Annualized costs for each equipment item and control option are
summarized in Table 8-20. Development of these cost estimates is
described in detail in Appendix F in 1979 dollars, which were scaled
to 1982 dollars at 10 percent per year. Annualized cost of control
(inspection, repair, and equipment) was estimated and an annual recovery
credit was subtracted to calculate total annualized cost per item
shown in Table 8-20 (see Appendix F for calculations). Annualized
costs for the two model plants are also summarized in Table 8-20.
Total annualized cost was obtained by multiplying the annualized cost
per item by the number of items at each plant (Table 8-18). • Control
techniques expected to save money are denoted as credit by parentheses.
In addition to costs shown in Table 8-20, each model plant would
be expected to incur an expense for the monitoring instrument that
cannot be attributed to each equipment item. Annualized cost of the
monitoring instrument is estimated as $5,000 per year (1979 dollars)
and is based on a capital cost of $8,500. Annualized cost includes
capital recovery ($2,000 for a 6-year lifetime at 10 percent), mainte-
nance and calibration ($2,700), and other annual expenses ($300 or
4 percent of capital).
8.1.2 New Facilities
The installed capital and annualized costs associated with the
control options in terms of new facilities may be less than the pro-
jected cost for existing facilities. The controls may be incorporated
into the design of a new facility to take advantage of optimum plant
layout to minimize piping distances. However, the annualized and
capital costs for new facilities are expected to fall within the range
estimated for existing facilities with the lower end of the range
8-33
-------
GO
^ i
c-f 1
^ fc. H
UJ |
— ' 1
n
1 — H
z 1
UJ Q
«E I
Q. 1
•— ' i
=> i
uT I
U_
0
1
o
oc
I—
!iE
o
o
:
C£
0
u_
£
GO
0
0
£
t— 1
a.
•a:
o
of
1— »
1
CO
III
' ' '
' 1
CO
«=c
1—
Jj
UI-
o
u ro
r— 01 J->
ra T3 a
*J 6 re
••-£'—
a. a.
ro
0
a.
re
1"
ai
=
1 o
1 "
o o o o o
§C5 O O 0
o o T-I o
o o co to ro
CO O ID **"
O CM
I-l
o o o o o
00000
m o O CD LO
r^. o co m i-*
rH CO tf> CM
in i-l
o o o o r*-
CM rO O CM ID
^- oo 01 in
•a
*3
c
e_
Si
'£
r
m
X)
x:
+J U)
•r- 4J
*. c
QJ
in >
* <~
ttl -r-
tn O
>
ui la ai
^ u ui .O
O -r- O) 3t
r— c t- tn
r— ro *r-
a< jc O) x>
-° S 5 «
T3 E -W S-
n QJ 3
to o o or
tn
ftl
u
'GJ
"113
q-
0)
!•"•
O)
W J-
t- 1
o a
4-> S-
Ifl W> =J
a) ui 13 in
> o. ro ui
£ E s: v
ro 3 X s-
>• a- UJ &.
o o
00
00 rj
?— I rH U
1 1
10 rt
0 0
s s
3
L)
V
c
3
t-
4J
OJ •!-
en
Q- *r~
t.
"S 5
o. o 'c
ro r— o
j o s:
•
U)
c
0
tn -u
a) u
.= g
i— c
o tn
•o u oi
0) U
T3 O5 t-
C C 3
0) v- C
c "a.
01 E i—
CS. ra r—
O o <£
• ro
1 1
"• t
.2 £
•o
a> rB
a. t-
§: g1
E a.
g u
*f- •!-
m
in a.
> S—
ro "O
> g
>.
i? c
? ^0
-*J
'«.*«*-
in u o
fO O
01 *™~ "^
O) -Q i-
•o ra
»*- o.
-o o
c «
ra (u to
w
E* c
QJ S- ro
w <*- 'a.
in aj J=
O) U
.*- t. in
t_ 3
I- OJ
ro t- a)
jo ra xi
- E -o
in a) r—
r- +> 3
ra tn o
OJ >> 5
1 t/> in
4->
II r— ^ C
ro tn a>
.r- -a =J
c J-
1 ra a) -»J
1 j^ s_ in
11 U 3 C
=J en
II ro
C O JZ
t-( 0 t—
11 ra xi o
in
£
5
«^
i
D.
o-
ai
8-34
-------
S
LU
-^^~
CD
§r
O
|~
1 1 1
D-
i — l
— >
CD"
LU
^-
O
C£.
u_
co
^•^
0
CO
CO
1 — 1
•z.
LU
LU
t — 1 ^— v
1 — in
HH S-
CD cd
U_ r—
O
U_ Q
CD
CM
—J CO
CD CTl
C£ t— 1
[ *— ^
r^
0
O
o
LJL.
CO
1 — '
CO
o
o
Q
LU
r^*j
r-H
I
IT)
^
CD
CM
1
OO
LU
I
CO
<^ O
*«— ' \^^ CD
CM
rH
t/1
C QJ
o tn >
*J O 'ro
U •<" >
CU -P
a. u to
Ul QJ 5
c a. o
s- >> i
0) r— T3
4-> x: QJ
j_ .p i —
ro c ro
=J 0
o o o
o o o
o
\.D
0 O O
tn co o
O CM •?!•
rH rH rH
CO
O O O
CM CM rH
O
CM
o un o
r- co en
o
CM
.a
C r—
o tn ro
•f- C CD
'•P O (0
L) *r-
OJ -P r—
Q. U rO
tn i OJ
OJ r— E
t. 4J r-
ro C TO
U 0 3
cxs: o
10
o.
a.
o o o
U> 0 0
**• o 1/1
rH ^T
rH
o o o
*S" O LO
rH «^-
rH
(D O1 O
r^. ID rH
CM
f — ID rH
CM
10
C
CD
to >
C
O (/> £-
•f- C -i-
•P 0 O
(J •«- >
cu -p i-
Q. O QJ
tn OJ tn
c a. ai
•i- tn S-
S- >i-r-
CU r— tn
I- P ro
ro c O)
3 O OJ
crs: o
to
ai
IO
x
UJ
o o o
o o o
O CO ID
CM rH r-
^~s *^^
0 0 O
r- tn o
""^ ^
CTJ LO CM
rH O «fr
CM CM CO
rH rH C7>
t/)
C
o in
•r- C
•P O
(J *r-
CU -P
D. U
c a.-^
•r- U) U)
C *i-
>••- -a
T >, cu
CD r- £_
4-> J= =3
i- P +J
ro C Q.
13 O 3
O* 3E CC
W
QJ
U
OJ
•a
0)
•r-
o;
I-
I
OJ
s-
10
CO
0)
t-
a.
o o
rH 0
CM ro
CM
0 O
•3- O
CM CM
rH
*3"
0"i Cn
o
rH
CO CO
i-H
O rH
rH
C
D)
S-
Q.
T3
CU
ui tn
a. o
(0 ^~
O 0
in
C
o
tn P
QJ t
C <
•r- C.
r— C
O
-o u
OJ
-a en
c c
0) *i-
C CL
OJ B
Q. TO
o wo
y^
ro
CO
rH
||
O
•P
(J
i.
v— /
U-
X
•a
c
cu
CL
«^
C
.f_
U)
^
o
T3
cn
r--.
en
t.
s_
ro
>>
a
Q.
-P
C
0)
u
cu
a.
o
rH
ro
•a
a>
ro
c
r-H
to
CD
C
>
O)
I-
o
^J
•a
£
u
ai
.p
o
c
J3
O) <*-
C 0
10 OJ
in to
TO 3
a>
OJ i.
-a ^
C 10
rd P
I/)
- 0
E U
QJ
to o
in OJ
D)
TJ to
5 cu
i*- fO
J- CU
CU J=
s_ aj
ro t-
^D ro
- E
10 QJ
'ro tn
QJ >,
r- ^
ra in
U T-
c
TO QJ
U D
OJ -P
E 0.
3
TO
3 ro
"O
S-
in O
QJ q-
•o
3 in
tj i/i
c o
i— 1 C_3
_Q (J
8-35
-------
being more appropriate for new plants. The costs for controlling a
new facility would be a function of the plant layout, piping distances,
and the number and sizes of the various emission points. Costs for
new,sources are not presented separately because those elements that
may be less expensive cannot be clearly identified and quantified, and
because any estimate would be a function of an assumed plant layout.
8.1.3 Modified Sources
The analysis presented in Subsection 8.1.1 for existing sources
is applicable to sources that are modified.
8.1.4 Summary of Estimated Control Costs
Cost estimates are provided in the previous sections for the
benzene sources at by-product plants, including groups of emission
sources and the leak detection and repair program. Not all by-product
recovery plants have all of the emission sources for which cost esti-
mates have been provided. All plants are assumed to have tar recovery
and handling sources, but a distribution of process types exists for
final coolers and light-oil recovery. Table 8-21 shows this distribu-
tion with 23 plants (42 percent of total capacity) having a direct-water
final cooler and 32 plants (58 percent of total capacity) having another
type or no final cooler. Nine plants (7 percent of total capacity) do
not recover light oil, and 7 plants (23 percent of capacity) refine the
light oil into pure benzene.
The capital and annualized costs for each control option for each
model plant are summarized in Tables 8-22 through 8-24. Also presented
in the tables for each source are uncontrolled benzene emissions,
benzene reductions achieved by the controls, and VOC reductions.
Average cost effectiveness is calculated by dividing the annualized
cost by the benzene emission reduction achieved. For sources with
more than one control option, an incremental cost effectiveness is
also given. Incremental cost effectiveness for a particular control
option is calculated by subtracting the annualized cost for the next
less stringent option from that particular option, and dividing the
difference in cost by the difference in emission reduction between the
two options.
8-36
-------
TABLE 8-21. ESTIMATED DISTRIBUTION OF TYPES OF COKE PLANT
EMISSION SOURCES
Source
Final cooler:
Direct- water
Tar-bottom
Wash- oil
Other
Light-oil storage
Benzene storage
Number of plants
23
18
6
8
46
7
Percent of
total capacity
42
28
14
8
* 93
23
8-37
-------
UJ
CO
CO
LU
U_
LL.
CO
8
CO
CO
o
CO
o
CO
CO
UJ
CSJ
CM
CO
CJ O
N .=
«f- *J
.
-o re
jj e r-
T3 t-
r- C «J
re ra »—
»— a» c «*•*
o c o s-
:§
88!
gc
(
CO r-H CM
) 00
J ro
in o o o
m o o o
r— if)
'r- C -r- 'i- C
S* to
r— i— r—
t. (0 »—
a. u « re a. u
C O.T3 U) C O-
O.UD1-P Q. U Q. r-
trt O) 3 Q.
CQ.-OOI CQ.1- 6
n- 0) r- .^ > -c o. t. ••-» .c
> •(- 0) t- -P •
- .
r— m c D
~ "
. O QJ 4->
JO •— O. t-
Dl C *—
c-^- m
r~ t. 'f f-
i o s
I 4-* O
— I- 1C
v- to t~ ••- .
••- ns nj
I— U 4J
011— •-- cr
in -I- 4J •<-
3 4-> -P ra
«= i •? c T3 -^ ra c r- -
T ! T !» C *J t T T t •-
»»- O n3 o (U •—
P-OOH-U T3 -3
!•§ 2-
iS 2.
8-38
-------
CM
1—
^C
— 1
D-
1
UJ
t — i
O
C£
O
\ '
CO
co
UJ
UJ
•""^
r-H
t—
o
LU
U_
1 1
UJ
I—
CO
o
o
t — \
<:
CO
I1 "
CO
o
o
ION REDUCTIONS,
co
CO
1 — 1
UJ
oo
CVI
co
UJ
ca
1—
r— in
CO W
c u c *-^
QJ 4j QJ QJ cn
e m 4- > s:
t 5 "S s ^
c
I-H
in
QJ i in
cn o a) -->
ro -P QJ C cn
OJ O f- > -*•,
1 CM '— s
TJ 3 CO t-
oj o cn >>
»r- 4-> Ul
r— ^^ W) t-
ro *O co
3 4J C i-
c in ro i—
c: o UJ i —
•r- U) 3 O
0.00*0
ro u x:
tJ -P CM
cn
rH
CCO
O -O *~*
•f— QJ t.
•P > >>
8 y-S'S.
> -D x: :E
OJ (J s-^
t- co
QJ O "D /•"%
C«- QJ J-
QJ 4J > >>
N U QJ -v
C ZJ -r=- CO
QJ T3 XT El
CO OJ O ^— '
TJ
Q)
r- tfl
r- QJ C /-*
O C O t-
i~ - >,
4-> N in "v.
c cr m cn
O O) •<- E
o xi E -^
C Q)
Control option
IS
y
o
U)
tr
o
in
in
3
Soo o g§ d §g gg s§ s §§§ sss §§§ Iss s
$§§ S oo S3 §§ 5° S*S° ei1-0. m.°.°. S""- c
^t-V ^ ~« m -HN ~ "^ >D S^SS >M '
rH rH rH
/-S VN
s°,°, § §§ ^ §s §§ ^s s °-°§ sS§ s§§ §°s s
$i§ ^ gco Sw "S^S £00 5^^.
^2"_.._ _.; r*!r-^ rH COrH LDCn rHIXt COrHrHrH OOIX> OO«* OO^T
5§°- S £« rH rHrH rHrH ^ CO -^ ^ CM rH ^ ^
rH rH
«3-r-ininin>X) o to
sgg s ss a ss ss ss si °°§ °°g °°s °°s
CMCOCOCMCMrH ,'~H m
000 CO 00 r-. UDCO rHcO r^CO ddcMin ixir^O rHrHCM COCOP-
§SS CO CMrH Cn rHrH CMCM rHrH cOrHrHrH- rH
CQ *f l£> in O rH . rH
^trTrH i-Tr-T
CO CO LO
,3-cvi coco CM«X> cMr^. iniocnr^ mcMCM coroixi eMixicn
COOCM tD COr- CO rHCM in^ ^2 £ ^^S ^^^ 000 CM CM *
^ in Q UD "d" ^" fO rH rH rH t— 1 r— 1 r— i OJ m
•* in rH CM rH
in in in
ojesj rHrH cncn rHrH cnr-r-vr- CMCMCM 10 m ID cncncn
ggg S S3 S 33 SS S3 fl- SSS ^'"^ °°° ""f'r
in in i-t oj r^
t. E E
™ t. i. ™ " S '
8-">, "l5.fi" 1 »
^ss^ | J i 11 i. l»!gl»g>'lg:5
^si i if l il il i! 1L- Hi i!Ll5s
,. "- 7~- -^ -S-0 •? S S S3 Q.un)roo.uro o.uoi4->D.oa.
E1*-1^ £ -2 S "^.n "^^ 3 oiaJsQjw^Qjaj wojaictnaj^
S,.^ * SS " SS £S £ «. SS.-oScS.in ea-oa; CQ.S-
I?? : a? : s- «: as •;i£1-;.=5 .isr^is
•J|| ' J -slef iiiiii^h lr.i^|Sl|I|||l|
ill i 11! i ' I!! Ill lit . ill! ill! iilf ills
OJ Q> " >
TJ
C
i — CO
t_ 0) -r- QJ «
S c o - cn ^
r- t- — i S. CO -0 C
t_OQ) >• 4JQJin^ CCO
aJor-t-i-x:4J^OfOP _
|88iJS.?SS^-»o,
aj, — u O" i — u-P ^ccn
(0 -r-TD QJS- COJin
gl ™ 7 § C'g 3 5fe ^
^-u^u ? » S^S •? -- 5
gfc?1"? S ^ ^ 5 Sf^- S1 « «>
u-Soso uimroc T3£3io s-i- o> g.
I04JO4J 34J+Jrt) CtrtU-'- OZJ 10 E
fefo^D-Cca,"' SS-t | . -S^ | S
^.jjcoc orara ui '£ •£ u
ou.^j3"--a.^«c:r--e ^-em ,— ^ u
S£oLo S«oT -ofc^™ o « g o ^. ^ -
^•£S,= S -S^5 i^f S iS S i-°u, S |
1 &Si -S -SS-S g^ SS S •§> | | |
-r- . . =3 •,- (O 0) -^ <1J rt X fO 'f- OJ QJ u l— -o _I-OSLU*J -ixi ca _ia- > uj a.
rt N^^in^^cda; 0 rt cJ
8-39
-------
r~ tft I
m IA
g^j O ftt > a_
£o •*- •*- **•.
o "to4"*^
- ca <
>o<
CD CO O O
> O C
- CD C
J O O
J 00 O
O CO GO O
oo
t_U
Q
f—
) O
•&
a;
o
u_
co
CO
»— C (O
«o ro i—
4^ *J 0> I—
a. c
IB t
i o
>13
> CD C
) 01 c
in o o o
Ul
O
LU
o ~a ^
"S >
> u «*
'•o2l_
01 (J ^^
t- m
« i
oo c I-H in
4 CNJ CM rH rH r
V)
te
U Ol *r- >t
J-» M W^s,
^
§^
§§ 15
01 in in in
•
co
i1
-r- *O
«*- c
•s
co
co
>._._ £ ^
i. « « en
ro TO ns c
•»- C ••- ••- C
p— O) »— (U r— C
U *•— I/) (J *r- O> O *i— U) O •*-
QJ -M r— r— O) -P •— OJ +J TO CO *D 4->
CO <1J 3 Q) CO OJ CJ CO (U QJ C CO QJ „
C O.~O CO C Q. CO C Q.*D •(- U)
e.cr- .e ^ ej= cx:
>,.,- +J (O >,.^ 4_> >,.r- +J J_ >,.^ +J
§• p.
.
fl O >> o
I
I
.? I
•S
o. -a
3-40
-------
Estimates of nationwide costs, emission reductions, and cost
effectiveness are presented in Table 8-25. Nationwide costs (excluding
the leak detection and repair program, LDAR) were estimated from the
model battery analysis with the use of linear cost functions and data
on plant-specific capacities and processes. For each control option,
a midrange capital and annualized cost was determined for each model .
plant. The midrange costs for the model plants were then used to
express the control cost as a linear function of coke capacity. The
cost function for a source was then applied to each real plant that
has the given emission source by using the real plant's capacity in
the cost function. Nationwide costs were determined by summing the
costs for all plants. To estimate the nationwide costs of the LDAR
program, the costs for each type of model battery shown in Tables 8-19
and 8-20 were multiplied by the number of each type of battery currently
existing. For exhausters, a total of 55 plants was used. A total 'of
46 plants produce light oil, and 7 of»these refine it to benzene.
Therefore, a total of 39 plants fall into the Models 1 and 2 category,
and 7 plants are represented by Model Plant 3.
Regulatory alternatives were developed from the control options
in Table 8-25 for the purpose of determining the economic impact
(Chapter 9) of differing control strategies. Regulatory Alternative I
represents baseline control with no national emission standard.
Based upon the average and incremental cost effectiveness in
Table 8-25, several options were chosen as Regulatory Alternative II
for the economic impact analysis. The controls for Regulatory Alterna-
tive II include the tar-bottom final cooler; gas blanketing for Sources
No. 2 through No. 7 (tar decanter, tar-intercepting sump, flushing-
liquor circulation tank, tar storage and dewatering, light-oil condenser,
light-oil decanter, wash-oil decanter, wash-oil circulation tank, the
excess-ammonia liquor tank, light-oil and benzene mixture storage tanks,
and benzene storage tanks); a sealed cover for the light-oil sump;
monthly monitoring for pumps and valves in benzene service (at least
10-percent benzene by weight); quarterly monitoring for exhausters in
benzene service (at least 1-percent benzene by weight); and equipment
8-41
-------
CO
CO
LU
LU
>
t-t
O
LU
U_
U_
ii i
I—
co
o
o
o
U C O1
E m cj 01 a:
U 0 V- >^1
- u •*- •fj^
5
«LJ VI
S?** L *3
£. VI 01 eX
a) O **•" (U ^^
5uls^a
**
« i»^
•(•- 4-* rH VI
§""" w n>
•U-pr-
e vi c *-
^ o w *o
^^
« u
*re 1£
4-> ** to r—
-r- (A V> O
CL O 3 T3
(O U O
^
ew^^
gill
CS"
nI^T
V *» > >>
03 w tj v^
ei
r— in
*•» M U» ***.
WJ= g'-'
§
1
O
i
•
a
c
VI
t=
c
V
V
S
O O O O
O CD O CO
CO O O CM
rH rH
§O CD O
O CD CO
£i°-°* ^-
ro cn
' x-s* /— %
rH CM rH f-.
CM ID rH CM
>-*CM CM ^
O O O O
r*» o CD *$*
r- rH ID O
ro"co"aC 0^
co m
§CD O O
CD O CD
CD rH O
^0 rH Cg
rH
CD O O CO
o in cn co
CM CO O CO
r*» co rH cn
Soo o
in cn CD
CO CO O CM
co~co*f-r o"
aj o
o QJ & vi
0»— r- >>
U O O W
r- 8 S CD
el! '•£
§<*-«»- c:
t. ro
+* r- r- O» •—
o o o o
_Q I t 4-» V)
i XTJC re
t- vi in CD cn
£ g ?£ c
r- CU
at a
in in in o O
3 3 23 O
-* — ' w QJ
t CT
t| u -^
Is § s
*3 f— u t-
oil ^ ,2
.05: g .=
t. x -S *-
« i CD+J a>
r-4-> C 0 +>
§O *r- jQ C
o r— i re
(J t- 0 t- U
" .r- 0 «
r*- o
g§
in r-
r-Ti-T
CD O
rH rH
§
Wash-oil scrubber
Coke oven gas blanketi
,_ L
§ 5
cr
•r- TJ
V (0
CO
J= C C ^
35 3S
«4- C O)
o cn cr
T3 *r- Id C
C 4-> t- f-
ta to o f-
5''" -5-c
CO
o o o
CM O O
rH rH Cn
rH CM
O CD O
CM O O
rH rH CM
rH rH
SCO CD
O CM
in tn ID
o o o
«r ID to
r-. *D m
ro" «-H CM
Svo tn
in rH
CM
o cn CD
s ss
"•"
? ss
<- in in
**
01
>, i"
system
oven gas blanketing s
Wash-oil scrubber
Coke oven gas blanketi
S
*O
CS' a,
ft «
*> OI V) J-
x: +*.* o
cn c c *J
•r- OS tO V>
f— U -M ,
O) t-
-T3 C O
oi .2 o-
V) i- 4J 'r-
c o to •—
OJ | »—
•0 JC 3 «
C M U ••-
o re i- c
"*-" i
1 4-> O tO
CD U W U C
•r- cu re x re
—ITD * LU 4->
•d- in
11
88
CO CM
rH CM
CD CO
CO O
*£> CO
O O
ss
C-TcM
CD CM
CO CO
^r m
CM CM
ro cO
11
cn
system
Wash-oil scrubber
Coke oven gas blanketi
U)
•o c
c to
03 +>
in a
H O)
!1
cn
££
0 3
r- 6
•r* OJ
— J J3
to
CD CD
CD O
CM CO
r-TcO*
O O
O CD
CM^CO
rH rH
CO CO
o cn
CO CM
CO *f
ID f->
!SS
2>
system
Wash-oil scrubber
Nitrogen gas blanketi r
U1
1
g,
2
S
V)
cu
c
N
C
o>
CD
^
O O O O
CM rH CM 0
CM rH rH O
"" a
o o o o
CM rH rH O
CM rH rH CO
CM
' CM rH
,-*. . .
<• in •ft- o
ID f m in
rH *D
**^ r-T
rH O O O
r* co
ID
o co cn CM
cn o o in
o 10 r*- co
rH
S CM ID ID
CM cn cn
r- ^- ^- m
O ID
t
rH
^t CM
f*C rH O
m co CD
*£>
000
o
r-"
CM
o m co
ID rH f-
ro «* tn
CM O rH
tn cn o
CM CM «3-
rH rH rH
O O O
(A
g
mechanical seals
Quarterly inspections
Monthly inspections
Equip with sealed-bel'
valves
.0
o>
5
S
§00 o o o
CD CD O P*- O
ro o o «j- CM CM
rH O CM CM
rH (D
§00 o o o
O CD O O CO
CO ID O "t- CO CO
CM
^*M— *
co o cn cn
ro **• co r*- rH r-
CM in cn tt- *? co
r^. • — '*— 'CM
o o o o o o
O CM
CM ro
CO rH
r^. o o CM co ID
r- cn «tf r- o CO
rH rH CM ro
CO rH CO O CM O
rH CM CO ' CM *f r-^
rH rH CM
ro co co o o o
ffj ro co r** P»« r^»
CM CM CM
U
(A
Quarterly inspections
Monthly inspections
Equip with degassing
reservoir vents
Quarterly inspections
Monthly inspections
Equip with rupture di:
X)
U)
, 'i
T3
0)
*«, 2
^ 0>
^ i_
tn 3
3 V>
re v>
_pr Q>
X *-
iu d-
rH CM
rH rH
O O
O Cn
CM CD
rH
O O
o cn
CM ID
Cn ro
CM CM
**;
ro cn
CO
tn to
in in
r- CM
CO cn
CM r^
in rH
co cn
c\i r^
in rH
cn
c
Q.
E
re
CO
Si cn
Q) 3 r—
4-> o. a.
& i £
1 &
1
01
c
re
5
'*
01
f-
3
cr
t-
o
V)
re
u,
4J' ' '
C
IA
C
CD '
tl
O
C
o
E
re
!
CT
O
i-
o.
t-
ai a.
a "
C "O
* £•
>o ro
u> C
U CU **x
C +J CD
•r- O> O
TJ O
c
O «i>i U3
•f~ m ID
4-> • O> CO
2
T3 U- (J
8-42
-------
controls for pressure-relief devices, sampling connections, and open-ended
lines in benzene service (at least 10-percent benzene by weight).
Regulatory Alternative III was chosen as a more stringent combina-
tion of controls that would yield a greater emission reduction than
that achieved by Regulatory Alternative II. The options chosen for
analysis as Regulatory Alternative III include a wash-oil final cooler;
equipment specifications for pumps, valves, and exhausters in benzene
service'(at least 1-percent benzene for exhausters and at least
10-percent benzene for pumps and valves, by weight); and for other
sources the same controls as listed for Regulatory Alternative II.
8.1.5 Comparison of Actual and Estimated Capital Costs
Because only a few gas blanketing systems have been installed,
any comparison of actual and estimated costs is limited. Armco, Inc.,
personnel estimated the cost of their gas blanketing system as $130,000
(1975 dollars) but warned that the estimate was approximate. The
system was part of a larger multimillion-dollar construction project,
which made it difficult to extract only gas blanketing costs.14
Inflating this estimate to 1982 dollars at a rate of 10 percent per
year yields an estimate of $250,000 (1982 dollars).
The Armco system controlled tar decanters; flushing-liquor circu-
lation tanks; ammonia-1iquor storage tanks; tar-collecting, tar-
dewatering, and tar storage tanks; and the light-oil plant, sump, and
storage tanks. A capacity of 837 Mg coke/day puts the plant into the
Model Plant 1 category. Capital cost estimates for Model Plant 1 for
sources controlled at Armco total $121,000 to $412,000 (1982 dollars)
with a midrange estimate of $267,000. The Model Plant 1 estimate
appears reasonable compared to the actual plant estimate.
Bethlehem Steel Corporation at Sparrows Point estimated the cost
of gas blanketing the light-oil plant in by-product Plant B as $44,000
(1982 dollars). Plant personnel indicated that this blanketing system
was also part of a larger project, and all costs could not be identified
clearly.15 The two by-product plants at Sparrows Point are designed
for a total coke capacity of 7,100 Mg/day, which roughly equals two
Model-2-type plants. Costs for gas blanketing the light-oil plant for
Model Plant 2 were estimated as $72,000 to $120,000 with a midrange of
$96,000 (1982 dollars). Comparison with Bethlehem's estimate of
8-43
-------
$44,000 indicates this estimate may be high. However, the estimate
encompasses a wide range of piping distances for the model plants, and
another specific plant with a different layout may incur greater
expenses than the relatively compact Sparrows Point plant would.
Two other companies submitted cost estimates for their own design
of a gas blanketing system. These designs were more sophisticated
than were gas blanketing systems that have been used in the industry
because of elaborate pressure controllers, alarms, blanketing tech-
niques, and redundant controls. The two companies suggested they
might choose to use nitrogen or natural gas instead of coke oven gas
to blanket the emission sources. These designs have not been applied
by the industry, and because the blanketing technology differs from
that recommended in Chapter 4 and analyzed in this chapter, direct-cost
comparisons would not be valid. However, in total capital costs,
estimates for the theoretical designs of the undemonstrated systems
were significantly higher than was the simpler coke oven gas blanketing
"system, which has been applied in at least three by-product recovery
plants.
8.2 OTHER COST CONSIDERATIONS
By-product coke plants have incurred a number of regulations that
relate to atmospheric and environmental emissions of solid waste and
water. The Occupational Safety and Health Administration (OSHA) has
developed occupational health rules that restrict personal exposure of
workers to 10 ppm benzene. (8-hour time-weighted average). It is
presently unclear which by-product emission sources are covered by
OSHA benzene standards and which mandated equipment and equipment
performance could be required. The environmental control alternatives
could effectively lower the emissions of the affected sources and help
attain the personal exposure standard, but the converse is not neces-
sarily true. OSHA controls could reduce worker exposure and have
little environmental benefit; e.g., venting of emissions into the
atmosphere away from the workers. For these reasons, the cost of OSHA
compliance is assumed to have no influence on potentially mandated
environmental controls.
8-44
-------
The coke oven by-product plants also have occupational health
requirements for exposure to benzene-soluble particulate materials
from the coke oven battery. Atmospheric emission controls are required
for charging, doors, pushing, quenching, and oven leaks. The costs
for OSHA regulations and other air and water regulations have been
included in the baseline costs, which will be analyzed in Chapter 9,
Economic Impact.
8.3 REFERENCES
1. Neveril, R. B. Capital and Operating Costs of Selected Air
Pollution Control Systems. U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. EPA-450/5-80-002.
December 1978. p. 3-2-3-8.
2. Mossman, M. J. Mechanical and Electrical Cost Data: 1982, 5th
ed. Robert Snow Means Company, Inc. 1981.
3. National Construction Estimator: 1982, 30th ed. Craftsman Book
Company, 1982. p. 194-197.
4. PEDCo Environmental, Inc. Technical Approach for a Coke Production
Cost Model. (Prepared for U.S. Environmental Protection Agency.)
EPA Contract No. 68-02-3071, Task 1. December, 1979. p. 12.
5. Perry, R. H. (ed). Chemical Engineer's Handbook. (Fifth Edition).
McGraw-Hill, Inc., 1973. p. 9-16.
6. Current Prices of Chemicals and Related Materials. Chemical
Marketing Reporter. April 9, 1982. p. 40-41.
7. Reference 1, p. A-9.
8. Reference 1, p. 3-15-3-18.
9. Hall, R. S., J. Matley, and K. J. McNaughton. Current Costs of
Process Equipment. Chemical Engineering. April 5, 1982. 89(11):
80-116.
10. Happell, J. , and D. G. Jordan. Chemical Process Economics. 2nd
ed. Marcel Dekker. NY. 1975. pp. 213-255.
11. Wilson, P. J., Jr., and J. H. Wells. Coal, Coke, and Coal Chemi-
cals. New York, McGraw-Hill, 1950. p. 44.
12. Guthrie, K. M. Process Plant Estimating, Evaluation, and Control.
Sol ana Beach, Craftsman Book Company, 1974.
8-45
-------
13. VanOsdell, D. W., et.al. Environmental Assessment of Coke
By-Product Recovery Plants. U.S. Environmental Protection Agency.
Research Triangle Park, NC. Publication No. 600/2-79-016.
January 1979. 387 p.
14. Branscome, M. R. Trip Report to Armco, Incorporated, Houston,
Texas. March 4, 1982. Research Triangle Institute.
15. Letter from McMullen, R. M., Bethlehem Steel Corporation, to
Cuffe, S. T. , U.S. Environmental Protection Agency. March 12,
1982. 5 Section 114 response.
8-46
-------
9. ECONOMIC IMPACT
This chapter addresses the economic impacts of the regulatory alterna-
tives for coke oven by-product plants. These alternatives are described in
Chapter 6 and apply to new and existing coke oven by-product plants.
Regulatory Alternatives II and III would have neglible impacts on the price
and production level of furnace and foundry coke but the regulatory alter-
natives are not expected to result in any closings of furnace or foundry
coke batteries, plants, or companies.
Section 9.1 presents a profile of the coke industry. Section 9.2
contains an analysis of the impacts of the regulatory alternatives, which
also addresses the potential impacts of compliance with a comprehensive
list of environmental and other controls. These impacts are measured
against the existing state of control for all sources. Section 9.3 presents
potential socioeconomic and inflationary impacts.
9.1 INDUSTRY PROFILE
9.1.1 Introduction
Coke production is a part of Standard Industrial Code (SIC) 3312—
Blast Furnaces and Steel Mills. Coke is principally used in the production
of steel and ferrous foundry products, which are also part of the output of
SIC 3312. Thus coke is both produced and principally consumed within
SIC 3312. Furthermore, many producers of furnace coke are fully integrated
iron- and steel-producing companies. Any regulation on coke production is
expected to have some impact on the entire blast furnaces and steel mills
industry with special emphasis on coke producers.
This profile has two purposes: (1) to provide the reader with a broad
overview of the industry and (2) to lend support to an economic analysis by
assessing the appropriateness of various economic models to analyze the
industry. Further, the profile provides some of the data necessary to the
analysis itself.
9-1
-------
The industry profile comprises six major sections. The remainder of
this introduction, which constitutes the first section, provides a brief,
descriptive, and largely qualitative look at the industry. The remaining
five sections of the profile conform with a particular model of industrial
organizational analysis. This model maintains that an industry can be
characterized by its basic conditions, market structure, market conduct,
and market performance.
The basic conditions in the industry, discussed in the second and
third sections of this profile, are believed to be major determinants of
the prevailing market structure. Most important of these basic conditions
are supply conditions, which are largely technological in nature, and
demand conditions, which are determined by the attributes of the products
themselves.
The market structure and market conduct of the blast furnaces and
steel mills industry are examined in the fourth section. Issues addressed
include geographic concentration, firm concentration, integration, and
barriers to entry. Market structure is believed to have a major influence
on the conduct of market participants. Market conduct is the price and
nonprice behavior of sellers. Of particular interest is the degree to
which the industry pricing behavior can be approximated by the competitive
pricing model, the monopoly pricing model, or some model of imperfect
competition.
The fifth section of the industry profile addresses market perform-
ance. The historical record of the industry's financial performance is
examined, with some emphasis on its comparison with other industries. The
sixth section of the industry profile presents projections of key variables
such as coke production and steel production. The seventh section discusses
market behavior.
9.1.1.1 Definition of the Coke Industry. Coke production is a part
of SIC 3312—Blast Furnaces and Steel Mills, which includes establishments
that produce coke and those that primarily manufacture hot metal, pig iron,
silvery pig iron, and ferroalloys from iron ore and iron and steel scrap.
Establishments that produce steel from pig iron, iron scrap, and steel
scrap and establishments that produce basic shapes such as plates, sheets,
and bars by hot rolling the iron and steel are also included in SIC 3312.1
The total value of shipments from SIC 3312 in 1980 was $50,303,900,0002 and
9-2
-------
an approximate value for total coke production in 1980 was $4,648,413,000,3
or less than 10 percent of the total value of shipments.
Coke is produced in two types of plants: merchant and captive.
Merchant plants produce coke to be sold on the open market, and many are
owned by chemical or other companies. The majority of coke plants in the
United States are captive plants which are vertically integrated with iron
and steel companies and use coke in the production of pig iron. At the end
of 1979, 17 plants were merchant and 43 were captive, and merchant plants
accounted for only 9 percent of total coke production.4
9.1.1.2 Brief History of the Coke Industry in the Overall Economy.
Traditionally, the value of coke produced in the United States has con-
stituted less than 1 percent of the gross national product (GNP).5 6
During most of the 1950's, coke production was about 0.30 percent of GNP,
and during the 1960's and until the mid-19701s, coke production was only
about 0.20 percent or less of GNP. However, in 1974, coke production as a
percent of GNP rose to above 0.30 percent. This trend continued for the
next 2 years. By 1979, coke production was about 0.2 percent of GNP.7 8
Previously, U.S. .coke exports have been greater than imports, but that
trend may be changing. The values of all U.S. imports and exports and U.S.
coke imports and exports are shown in Table 9-1. From 1950 to 1972, coke
exports were much greater than coke imports, but after 1973, this trend was
reversed. The same pattern applies to the percentages of coke imports and
exports within total U.S. imports and exports. From 1950 to 1972, coke
exports were a larger percentage of total U.S. exports than coke imports
were of total U.S. imports. Again, from 1973 to 1979, this trend reversed,
and coke imports were a larger proportion of total U.S. imports than coke
exports were of total U.S. exports.
U.S. coke production has always been a substantial portion of world
coke production. This share has decreased during the past 30 years, as
indicated in Table 9-2. From 1950 to 1977, world coke production generally
increased while U.S. coke production decreased. This trend explains the
decline in the U.S. percentage of world coke production.
9.1.1.3 Size of the Iron and Steel Industry. The value of shipments
of SIC 3312 has increased since 1960. There have been a few fluctuations
in this growth; for example, as shown in Table 9-3, the 1965 value of
9-3
-------
<0
in
_jj
UJ
a
S- i —
O O
X
a) c
o
0) •!-
O r—
p-
v*<
ro
In"
S-
m ro
S- r^
O O
0-73
0) C
to i^
ID ^
v-'
if- in
in o 4->
-P s- /— .
S_ 0) 0 4J
o t- a. t=
trO E G)
J=-^ 0
- ••. t
•I™* in >—
i— 0)
CU ro ro Q.
O in O
o ro 4->
ro
c in
O 1-
in «i- to
•P -P i—
J- Q.>—
O E O
Q. ^ 73
E in
•r- C C
O O
a> o T-
0 J- 1^
O O •!-
vS
inro
•P r-*,
s- in
0 S-
Q. ro
E i —
•i— ^
O
. 73
. C
__i O
'ia ^
4J -i-
o .a
S-
tO
O)
,C3rHOOOOOO'C3C3OOOC3OOOOC3r-lrHrHOOOOO
OOOOOOOOOOOOOOOOOOOOOOOOOOO
cvj^r^coeMe\iLn^rHr^crie>j^corHco<^-Lni£>Lnoioor->-rHiDr»~i^-
U3l^rOCritDCOrH«d-r~~OOlJDOOI>~-COOlDfOli3OOOOCO'd-OrOfO^flj3
r~l rH |-H rH iH rH CVJ rH rH CO f**» ^t" fO CO ^t" f^« CO
c*^oe\ioorHinrHo^u5ii5or^roLnLncoLniJDOinLn'sj"«i-roiHrv-.
• t i i *^ ^**
OOOOOOOOOOOOOOOOOOOOOOOOrHrHO
OOOOOOOOOOOOOOOOOOOOOOOOOOO
rooiLnr^ro^inLncD'd-LnLoo^OLn^oor-.ai^-LnoirjrocviLOrH
OO CT> LO rH
rH rH rH
CTior^cric\!^t^oOTcsjr^r^^rHr^'=*-LnoocNioOTi^coinr----oo
X3
C3 iH C\I CO ^* LO IO f^»» 00 CTl C3 rH C\J CO ^" 1X5 CO f**» OO CT) O iH C\J OO o^ cri rj> CT> CD cri cji CTV rj> cji CD en o^v CT^ cr> cr> en en cr> cr> en en en
rHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHiHrHrHrH
.
in
E
J-«
cu
4->
'T*
ro
r—
r-™
O
73
'*-'
0)
3
r— •
to
4-5
0
C
73
C
ro
in
E
s-
O)
4-3
^
0
4->
4->
•r"
C
|T3
C
*r—
eu
t-
o
D-
O)
S_
0)
eu
3
in
4-J
J_
O
ex
X
cu
73
C
(0
in
5-
O
E
•r-1
CU
-^
o
U
00
in r-.
fO rH
r— 73
O C.
73 ro
d r^
cu en
S- rH
t-
3 c:
O hH
tJ -Q
9-4
-------
TABLE 9-2. COKE PRODUCTION IN THE WORLD2 4 7
Year
World production
(million megagrams)
U.S. production
(million megagrams)
U.S. production
as a share of
world production
(percent)
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
182.3
204.1
208.9
225.6
211.5
242.3
256.8
266.1
255.0
260.4
279.7
272.0
272.9
281.7
298.5
310.3
310.4
303.9
315.8
335.8
350.5
342.7
340.5
365.8
367.4
363.3
367.2
373.5
364.7
341.0
65.9
71.9
62.0
71.5
54.4
68.3
67.6
69.0
48.6
50.7
51.9
46.9
47.1
49.3
56.4
60.7
61.2
58.6
57.8
58.8
60.3
52.1
54.9
58.4
55.9
51.9
52.9
48.5
44.5
48.0
36.1
35.2
29.7
31.7
25.7
28.2
26.3
25.9
19.1
19.5
18.6
17.2
17.3 _
17.5
18.9
19.6
19.7
19.3
18.3
17.5
17.2
15.2
16.1
16.0
15.2
14.3
14.4
13.0
12.2
14.1
Oven and beehive coke combined.
9-5
-------
TABLE 9-3. VALUE OF SHIPMENTS, SIC 33128 9
Year
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
Current dollars
(millions)
15,783.8
14,873.3
15,571.6
16,418.0
18,840.1
20,841.7
21,193.9
19,620.6
21,161.1
22,299.0
21,501.6
21,971.3
23,946.7
30,365.5
41,671.7
35,659.8
39,684.1
41,897.8
1972 Dollars
(millions)
22,981.7
21,468.4
22,071.7
22,933.4
25,914.9
28,043.2
27,610.6
24,829.9
25,628.1
25,713.8
23,535.0
22,882.0
23,946.7
28,700.9
35,917.7
28,038.8
29,643.8
29,645.4
9-6
-------
shipments of SIC 3312 was the highest value between 1960 and 1972. Since
1972, the value of shipments has remained around $30 million, with the
highest value being $35 million (1972 dollars) in 1974.
For SIC 3312, Table 9-4 shows the value added by manufacture, the
total number of employees, and the value added per employee. Current and
constant (1972) dollar figures are included. Both the total value added by
manufacture and the value added per employee peaked in 1974, the same year
in which the value of shipments for this industry was the highest. The
increasing value added per employee might indicate that this industry is
changing to a more capital-intensive production process.
9.1.2 Production
9.1.2.1 Product Description. Two types of coke are produced: fur-
nace coke and foundry coke. Furnace coke is used as a fuel in blast
furnaces; foundry coke is used as a fuel in the cupolas of foundries. Coke
is also used for other miscellaneous processes such as residential and
commercial heating. In 1978, only 2 percent of all coke used in the United
States was used for these miscellaneous purposes, 93 percent was used in
blast furnaces, and the remaining 5 percent was used in foundries.14
Time-series data for the percent of total U.S. consumption attributable to
each use are shown in Figure 9-1.
9.1.2.2 Production Technology. Coke is typically produced from coal
in a regenerative type of oven called the by-product oven. The type of
coal used in coke production and the length of time the coal is heated
(coking time) determine the end use of the coke. Both furnace and foundry
coke are usually obtained from the carbonization of a mixture of high- and
low-volatile coals. Generally, furnace coke is obtained from a coal mix of
10 to 30 percent low-volatile coal and is coked an average of 18 hours, and
foundry coke is obtained from a mix of 50 percent or more low-volatile coal
and is coked an average of 30 hours.
The first by-product oven in the United States was built in 1892 to
produce coke and to obtain ammonia to be used in the production of soda
ash. In such ovens, the by-products of carbonization (such as ammonia,
tar, and gas) are collected instead of being emitted into the atmosphere as
they were in the older, beehive ovens.
9-7
-------
TABLE 9-4. VALUE ADDED, SIC 33128 9
Year
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
Value added by
Current dollars
(millions)
6,844.4
6,546.3
6,620.9
7,506.4
8,479.6
9,379.8
9,643.6
8,910.1
9,275.8
9,853.2
9,350.5
9,563.1
10,304.7
12,769.4
17,425.8
13,356.2
14,755.5
15,021.4
manufacture
1972 Dollars
(millions)
9,965.6
9,449.0
9,384.7
10,485.3
11,663.8
12,620.8
12,563.3
11,275.8
11,233.9
11,362.1
10.234.8
9,959.5
10,304.7
12,069.4
15,019.7
10,501.8
11,022.3
10,628.6
Employees
(thousands)
550.0
503.4
502.2
500.5
532.9
565.4
559.4
533.1
533.1
537.7
526.5
482.2
469.1
502.1
518.0
451.3
451.9
441.4
Value added
per employee —
1972 dollars
(thousands)
18.1
,18.8
18.7
20.9
21.9
22.3
22.5
21.2
21.1
21.1
19.4
20.7
22.0
24.0
29.0
23.3
24.4
24.1
9-8
-------
LU
o
o
00
CO
CO
r-
CM
oo
to
CO
CO
CO
to
CM".
o
I
E
O)
DC
42
LU >r
/
\
LJJ
O
cc
D
I . I
eo
LU
C/5
cr
LU
I
O
) I
CM
CO
O
CO
§
s
CM
LO
O
u
I
en
CM
cn
o
cn
oo
oo
CO
00
00
00
CO
CM
0380 3MOO dO lN30«3d
9-9
-------
The total amount of coke that can be produced each year is restricted
by the number of ovens in operation for that year, and not all ovens are in
operation all of the time. Oven operators try to avoid closing down a
group of ovens for any reason because of the time and energy lost while the
ovens cool and reheat and because of the oven deterioration that results
from cooling and reheating. However, it is estimated that at any time,
approximately 5 to 10 percent of existing coke oven capacity is out of
service for rebuilding or repair.20 In a report written for the Department
of Commerce, Father William T. Hogan estimated the potential annual maximum
capacity of U.S. oven coke plants as of July 31, 1979.21 His estimates are
shown in Table 9-5. Hogan assumes that almost 10 percent of his estimate
of total capacity will be out of service at any given time; therefore, he
subtracts the out-of-service capacity from total capacity to obtain maximum
annual capacity.
Within the limits of the number of ovens available for coking, both
furnace and foundry coke production levels vary. Some ovens that produce
furnace coke can be switched to produce foundry coke by changing the coal
mix and increasing the coking time. Furthermore, some ovens that produce
foundry coke could be changed to produce furnace coke by changing the coal
mix and decreasing the coking time. Also, some variation in the combina-
tion of flue temperature and coking time is possible for either type of
coke. A shorter coking time results in greater potential annual produc-
tion.
9.1.2.3 Factors of Production. Table 9-6 provides a typical labor
and materials cost breakdown for furnace coke production. Coal is the
major material input in the production of coke. In 1979, greater than 61
percent of the coal received by coke plants was from mines that were
company owned or affiliated.23 In this same year, 14 States shipped some
coal to coke plants outside their borders.24 Of the coal received by
domestic coke plants, over 81 percent came from West Virginia, Kentucky,
Pennsylvania, and Virginia.25 Any potential adverse impact on the coke
industry probably will have some impact in these States. A total of 69.9
million megagrams of bituminous coal was carbonized in 1979.26
Table 9-7 shows employment in the by-product coke industry from 1950
to 1970 and the percentage of total SIC 3312 employees in the by-product
9-10
-------
TABLE 9-5. POTENTIAL MAXIMUM ANNUAL CAPACITY OF OVEN COKE
PLANTS IN THE UNITED STATES ON JULY 31, 197912
In existence
Furnace plants
Merchant plants
Total
Out of service3
Furnace plants
Merchant plants
Total
In operation
Furnace plants
Merchant plants
Total
Number of
batteries
169
30
199
(18)
(2)
(20)
151
28
179
Number of
ovens
10,076
1,337
11,413
(1,026)
(117)
(1,143)
9,050
1,220
10,270
Capacity
(Mg)
53,095,381
4,400,691
57,496,072
(5,255,001)
(460,599)
(5,715,600)
47,840,380
3,940,092
51,780,472
Batteries and ovens down for rebuilding and repair.
9-11
-------
Percent of cost
77.1
9.4
6.6
6.9
TABLE 9-6. TYPICAL COST BREAKDOWNS: FURNACE COKE PRODUCTION AND
HOT METAL (BLAST FURNACE) PRODUCTION13
Furnace coke production
Labor and materials
Coking coal
Coal transportation
Labor (operation and maintenance)
Maintenance materials
Total labor and material costs
Hot metal production
Charge metal!ics
Iron ore
Agglomerates
Scrap
Fuel inputs
Coke
Fuel oil
Limestone fluxes
Direct labor
Maintenance
General expenses
Total labor and material costs
100.0
Percent of cost.
100.0
9-12
-------
TABLE 9-7. EMPLOYMENT IN THE BY-PRODUCT COKE INDUSTRY15
Year
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971a
1972
1973
1974
1975
1976
1977
Number of employees
20,942
22,058
21,919
21,011
17,944
19,595
19,318
19,203
15,654
15,865
15,779
13,106
12,723
12,696
13,021
14,003
13,745
13,662
14,136
13,617
13,997
11,955
11,127
11,121
11,207
12,109
11,047
10,196
Percentage of all
employees in SIC 3312
NA
NA
NA
NA
NA
NA
NA
NA
3.06
3.13
2.87
2.60
2.53
2.54
2.44
2.48
2.46
2.56
2.65
2.53
2.66
2.48
2.37
2.21
2.16
2.68
2.44
2.31
NA = not applicable.
aFigures for 1971-1977 are estimates. See text for more detail.
9-13
-------
coke industry. This table shows decreasing employment in the by-product
coke industry. A similar decline in employment has occurred in SIC 3312.
Unfortunately, employment data for the by-product coke industry are not
available after 1970; however, these figures can be estimated by regressing
employment in the by-product coke industry on total iron and steel industry
employment and on the ratio of coke used in steel production.* These
estimates are also shown in Table 9-7.
9.1.3 Demand and Supply Conditions
Domestic consumption of coke from 1950 to 1980 is graphed in Figure
9-2. In the early 1950's, the amount of coke consumption was fairly large;
an average of 65 million megagrams was consumed annually between 1950 and
1958. The late 1950's and early 1960's showed a sharp decrease in coke
consumption, with an average of only 48 million megagrams consumed
annually. Domestic consumption of coke increased during the mid-1960's to
mid-1970's to an annual figure of 57 million megagrams but it did not reach
the 1950 to 1957 level. The late 1970's show another slump in coke
consumption.
The variation in coke consumption shown in Figure 9-2 has both cyclic
and trend components. The demand for coke is derived from demands for iron
and steel products, and these demands are sensitive to the performance of
the overall economy. Cycles in coke demand are linked to cycles in aggre-
gate demand or cycles in demand for particular products such as automobiles.
The trend component in coke consumption results from changes in blast
furnace production techniques. Coke is used as a fuel in blast furnaces,
but it is not the only fuel that can be used. Coke oven gas, fuel oil, tar
and pitch, natural gas, and blast furnace gas have all been used as supple-
ments to coke in heating the blast furnaces. The increased use of these
supplemental fuels over the past 20 years has caused the amount of coke
used per ton of pig iron produced (the coke rate) to decrease. Other
causes of the decline in coke rate are increased use of oxygen in the blast
*Regressions performed by Research Triangle Institute, 1980.
9-14
-------
o
00
00
r-
CM
O)
oo
CD
CD
CM
to
o
CD
CD
in
in
CM
in
o
in
cc
<
UJ
o>
1
M—
O
o
Q.
O
o
*•"
I
s.
Q.
03
0)
CM
O)
£
O)
o
r*.
it)
(O
o
CD
in
in
in
CO
3>IOO JO S1AIVUDV03IAI NOinilA!
9-15
-------
furnaces and use of higher metallic content ores. Table 9-8 shows U.S. pig
iron production, coke consumed in the production of pig iron, and the coke
rate for 1950 to 1980. (Data limitations make it difficult to calculate
the foundry coke rate in cupola production.)
Recently, there has been some concern about the ability of the United
States' coke-making capacity to support domestic steel product!"on--the
major source of coke demand. The study conducted by Hogan and Koelble of
the Industrial Economics Research Institute at Fordham University indicates
that in 1978, U.S. production of coke was 14.1 percent below domestic
consumption.31 Imports increased dramatically in that same year. Hogan
and Koelble attribute this decline in coke production to the abandonment of
coke ovens for environmental reasons and predict a severe coke shortage by
1982.32 This prediction is disputed in a Merrill Lynch Institutional
Report by Charles Bradford. The Bradford report attributes the lack of
adequate U.S. coke production in 1978 to two factors: (1) a coal miner's
strike, which caused the drawing down of stocks of coke when they should
have been increasing, and (2) the premature closing because of EPA regula-
tion of some coke ovens that normally would have been replaced before they
were closed.33 The Bradford report states that a survey of U.S. steel .
producers revealed that all of the major steel producers are or soon will
be self-sufficient with regard to coke-making capacity.34 The Bradford
explanation of 1978 coke imports seems more reasonable because 1979 coke
imports decreased about 1.6 million megagrams compared to the 1978 level.
Furthermore, it seems unlikely that a severe shortage of coke capacity will
occur in 1982 because currently there are no signs of a major shortfall in
capacity.
9.1.4 Market Structure
Market power, the degree to which an individual producer or groups of
producers can control market price, is of particular economic importance.
Market structure is an important determinant of market power. Pricing
behavior is relevant to the choice of the methodology used in assessing the
potential impacts of new regulations. It is important to determine if the
competitive pricing model (price equal to marginal cost) adequately
describes pricing behavior for coke producers.
9-16
-------
TABLE 9-8. COKE RATE2 4 7
Year
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
Pig iron production
(thousand megagrams)
58,514
63,756
55,618
67,906
52,570
69,717
68,067
71,128
51,851
54,622
60,329
58,834
59,546
65,173
77,527
80,021
82,815
78,744
80,529
86,186
82,820
73,829
80,628
91,915
86,616
72,322
79,788
73,931
79,552
Coke used in
blast furnaces
(thousand megagrams)
51,403
55,362
49,386
58,880
46,861
60,675
58,279
60,861
42,898
44,107
46,462
42,855
42,298
44,596
51,076
53,576
54,653
51,300
51,399
55,065
54,754
48,269
50,214
54,791
51,154
44,375
47,678
44,292
47,889
Coke rate
0.86
0.87
0.89
0.87
0.89
0.87
0.86
0.86
0.83
0.81
0.77
0.73
0.71
0.68
0.66
0.67
0.66
0.65
0.64
0.64
0.66
0.65
0.62
0.60
0.59
0.61
0.60
0.60
0.60
9-17
-------
Any analysis of market structure must consider the characteristics of
the industry. This analysis addresses the number of firms producing coke;
the concentration of production in specific firms; the degree of inte-
gration in coke production; the availability of substitutes for coke; and
the availability of substitutes for the commodities for which coke is an
input to production. Also, some information on past pricing in the coke
industry is presented. These topics will be considered together with
financial performance (Section 9.1.5) and growth (Section 9.1.6) in asses-
sing market behavior (Section 9.1.7).
9.1.4.1 Concentration Characteristics and Number of Firms. This
section describes various concentration measures that can be computed for
the furnace and foundry coke industries. Normally, concentration ratios
are used as an indication of the existence of market power. While concen-
tration ratios are a useful tool for describing industry structure, concen-
tration should not be used as an exclusive measure of market power. Many
other factors (e.g., availability of substitutes, product homogeneity, ease
of market entry) determine a firm's ability to control market price.
As of December 1982, 30 companies operated by-product coke ovens.35
Fourteen companies are integrated iron and steel producers; 16 companies are
merchant firms. These companies owned and operated a total of 55 coke
plants; 37 of these plants were captive and 18 of them were merchant. A
list of these companies, their plant locations, the major uses of coke at
each plant, and plant coke capacities is given in Table 9-9.
Reported capacities in Table 9-9 are maximum, nominal figures, which
do not include any allowance for outage like that estimated for the overall
industry in Table 9-5. All of the largest plants are captive, and most of
the merchant plants have very small capacities. Furnace coke production is
concentrated in captive plants. Virtually all of the coke used in foundries
and in other industries was produced by merchant plants. If coke plant
sites were ranked according to capacity, the top five plant sites and top
ten plant sites would have 30.9 percent and 45.8 percent of total coke
capacity, respectively. (A plant site or location may include more than
one complete plant.)
By-product coke plants are concentrated in the States bordering on the
Ohio River, probably because of the coal in that area. Figure 9-3 shows
9-18
-------
i-t
M
O
0
CO
en
rH
>>
re
c
re
CO
1(1
1 LI
<:
co
Q
LU
f-
1— l
Z.
ID
LU
1—
1— )
CO
1—
1
Q-
LU
O
u
en
en
LU
_i
CO
£
re
>> s-
-P CO
•r- re >-->
u o> s-
re cu re
0. E CU
o -a
C. S-
> >
•p -P -P
a. Q. a.
(O ttJ fO
O 0 O
-Q^
CM
;!-
O
Hamilton, 0
Houston, TX
Middletown,
•
(J
C
t— 1
•s
o
o
£
i-
-. rH rH CO CM
ID CM CO CM 00
o cn co vo ci
CM rH rH CO
cu cu cu cu cu
u u u u u
c c c c c
3 3 3 U 33
H- t*-. t*-. t|_ t^
4J 4J 4-> 4^ +J
tn w w w> v>
(tj (Tj fO fO CO
/^ OQ CO ^Q fyn
o oj cu cu o>
a. ex ex ex ex
re re re re re
o o o o o
o
•z.
HH ••
>- -P
<: »<: -z. c.
Bethlehem,
Burns Harbo
Johnstown,
Lackawanna,
Sparrows Po
CX
i.
o
o
r—
CU
cu
-P
CO
E
CU
XI
cu
r—
XI
-p
cu
CO
c§
CO
cu
o
re
c
&-
3
4-
•P
in
re
CQ
cu
>
•p
CX
re
o
o
o
o"
cu
3
Q.
CX
i-
o
o
r—
CU
cu
•p
CO
t— 1
o&
LJ_
0
•st-
rH
CU
U
re
c.
&~
3
4J
re
CO
cu
>
CX
re
o
CX
re
o
-,_
o
Portsmouth,
/—^
•p
o
t-
• -p
ex cu
S- CD
O 1
0 CU
i.
in •!-
Q. CX
0 E
i— LU
CJ ^— '
^^
O
CD
O
rH
CU
U
re
C
5-
3
q-
in
re
CO
cu
>
CX
re
o
t — i
Dearborn, M
,
o
o
S-
o
-p
o
-a
o
u_
rH
rH
CO
CU
U
re
• C
£-
3
4J
in
re
CO
cu
>
•p
CX
re
o
CO
^
I-H
O
D)
re
o
•r-
O
LU
d
o
r—
CU
cu
-p
CO
•a
c
re
c
»— i
rH
CO
LO
cu
u
re
c
£_
3
q-
-P
re
r*—
CO
cu
>
•p
CX
re
0
i
o
co
re
•r—
x:
o
.
u
c
«t
at
re
i-
cu
•p
c
1 — 1
tn CD r**- co
rH CO Is* CO
CM o co r>-
rH rH rH rH
CU 4J -P
in t/> tn in
re re re re
ca ca ca ca
cu cu cu cu
> > > >
4J 4J +J +J
a. cx cx cx
re re re re
o o o o
z
CO
1
o^i
•-)
-a
cu
3
£2
.p
o c
•* o
m u
ri ^-i ^
,— i
cu
u
re
c
£-
3
S-
•P
re
CO
CU
Q.
re
o
-
-------
in
>> s-
+? D)
•r— re x*^»
u en $-
re cu re
0. E >
0 TJ
C S-
cu re cu
Jxf Ul Q.
O 3
O O
JC
p
re
O)
o
u
«[ ,
o
U)
cu
in
3
S-
_0
re
s:
c
/"s O
CU P P
3 re c
c (_> re
P *<£ "a.
C •!-
o in >
a) re
•a a.
o o
i- CJ
re
o
o
CO
i-H
r—
03
S—
p
in
^
•a
^> •*"-
£-«
13 S-
C 0)
3 JC
0 P
Ll_ 0
P
C
re
JC
o
o>
z.
*,
re
en
o
o
c:
re
•p
re
o
•a
c
re
cu
o •
o o
CJ
re
O>r-
o re
0 O
C -i-
re P~
P CU
P JC
re o
JC
o
}
y
TJ
3
O
P
C
re
f~*
u
CU
•z.
1— 1
**
CO
•r—
p..
O
CL.
re
re
•r—
•a
c
^
0)
NX
O
o
-a
c
re
in
re
CD
^^
in p
C •!-
0) r-
N •!—
••- p
•P Z3
o
l^D
1-1
^-
5-
T3
C
3
O
n
0)
u
re
c
S-
3
l^-
p
in
re
CQ
•P
JC
re
f~
u
QJ
s:
i— i
s:
«v
[ ^
IP—
o
P
a)
a
0)
o
o
P
•r-
O
£_
P
CU
0
00
cu
0
re
c
S-
3
if»
P
(A
re
ca
p
c
re
r~
O
CU
s:
^_
z
*,
o
"re
<»-
t-
3
oa
cL
j-
o
o
OJ
NX
O
C_3
re
c
r"
re
a;
i
s-
cu
c:
C
O
0 01
rH rH
*^> ***>
g S—
•a -o
c c
3 3
O O
u. u_
p p
c. c
re re
r* f—
O U
S- i-
cu cu
2 21
o
»»
** G)
P C
^- «r—
o re
3C O-
in
r—
re
u
•r-
E
ai
o
o
o -o
n
a) re
NX
o au
o -^
o
cu o
S-
•r- CU
Q. T-
E S~
JJJ LU
CM
rH
r-«
re
«r—
J-
p
in
3
T3
*. f*
'•^ CJ
1C
1— <
00
in
00
^•j
i.
-a
c
3
O
E:
^|
o
«*
!C
o
p
c
o
s-
1—1
•
CL<^^
S- r—
O CU
o cu
-p
«U v
s_
73
C
3
O
M-
*i
0)
U
re
c
j_
3
>> *!—
y
•o p
c in
3 re
O i—
u_ m
p P
C 1C
re re
.d ^cr
O 0
S- S-
cu cu
y y
_j
ca
.
o
0
cu
NX
O
C_3 S-
01
0) P
C r-
o re
> E
tt) -r-
r^x.
CM
J-
0)
JC
p
0 r-
re
*>«|—
^>) S-
$— *p
13 in
C 3
3 TJ
O C
U_ •(-
P
C
re
r"
U
S-
a>
y
a*
3
C
•r—
f1*** CM *t*^
tO i.O d
r-l UTJ O
(J
Nw/
^>k
s_
"Q
c
3
0
M-
*t
0)
U
re
c
s_
3
^>j 4—
5_
TJ P
c in
3 re
O i—
u_ ca
.-.
p P
c c:
re re
JC JC
u u
S- i.
CU 01
E: 2:
— i
«^
a:
O "
•a
•• &-
o re
TJ 3c
01 TJ
r— O
O O
h- 3:
9-20
-------
/••%,
•o
cu
3
•r-
-P
C
0
CJ
v~ '
ai
i
cr>
LlJ
oa
t^
(A
s
tO
•P C3)/~%
•r- 10 S-
u cn to
to cu cu
a. E >>
10
u T3 s-
Cflj
vw
a»
c
to
r—
a.
cu
E
to
C
>}
c
tO
ex
E
O
CJ
co
o
in
cu
u
(O
c
3
4—
•P
in
r^
oa
cu
_>
•P
ex
10
o
X
1—
*»
i-
10
+J
CO
CO
c
o
1
•
o
o
I—
cu
CO
-p
CO
i-
to
•p
CO
cu
c
o
— 1
CM 00 rH in
rH CM rH
f** CT> rH O
i-Tr-Tr-T
0) 01 CO CO
U U U O
to to to to
c c c c
SH Ci. £_ £»
3333
4— 4— 4— 4—
^•3 ^> ^J ^_>
cn in in cn
(0 to to tO
oa oa oo CQ
CU CO O) 0)
> > > >
ex ex a. a.
tO fO (0 tO
O O CJ O
2>
_, *
•a
- c
>> to
-P M > I—
•r- y ^g CA
CJ I-H
*\ **
CU -P C CA
4-> T- 0 -
•r- O -P C
c i. i. jt
tO -P -r- O
s- a> a> s-
C3 Q 3: 00
ex
s-
o
CJ
r—
CU
CO
-p
co
1—
fO
c
o
•r-
-p
ro
z
in f — co cn m co in
in in CD co rH <^- r->
r>~ r-. rH »* co cn oo
,-f
cu cu cu co co co co
U U O CJ U O O
to to to to to to to
c c c c c c c
3333333
*l 1| 'l ll ti t!Trl tl^
+J +J +J +J +> +3 +1
cn CA in in w CA in
to to to to to to to
oa OQ oa oa oa oa oa
CO CO CO CU CU CU CU
>>>>>>>
+j +j -P +J -P -P -P
ex ex ex ex ex ex ex
fO to tO to tO fO to
0 CJ CJ O CJ O CJ
CM"
^_,
— J 1C
a: a: H-I o
o o
_i »,
«<: •> o — i x: c
-a E oxc o 3
c - o to o
10 C r— CJ » "-P
i— CU r- -r- (A C in
CU T3 -r- JZ (0 CU O)
> cn in CJ E J- c:
0) T3 in 0 i- 3
i — to fO • .£! to O
O C3 S CO r- 3: >-
ex
5-
O
CJ
r—
CO
CO
co
o
•r-
I—
f">
3
ex
CO
O£.
f*^
rH
O
CD
CU
O
to
C
3
4-
1 1
cn
to
ca
cu
>
-P
ex
o
CO
^^
Q_
ft
C.
o
-p
5-
m
r™
CJ
.
ex
s_
o
o
J— -
cu
cu
•p
CO
.
CO
ID
00 <3- CD
rH rH CM
00 CD CM
t™"l ^j"
CO CU CU
u u u
tO to to
c c c
333
t- 4- -P
in in cn
to to to
CQ CQ CQ
CO CO CO
> > >
•p -p -p
ex ex ex
fl3 03 03
o o o
>
•1— -e— t-
(0 tO tO
LL. LL. CD
rH OO
CD m
-=f rH
rH rH
CU CO
O O
to to
C. C
f f
>— S—
3 3
4-4-
•P -P
cn IA
to ra
ca oa
cu cu
> >
•p -P
CL ex
0 0
"T*
O r-
n
c
•r- O
fO >
S- O
0 5-
_i a.
r*- o^
o oo
^
cu cu
u u
to to
c c
3 3
•P -P
in CA
to O-
c
CO -
^Q C
3 CO
CO in
•P in
CO CO
c:
• o
LU S
r*
O)
i-
3
f^
in
-p
•r—
CU
1
O)
c
r«.
CU
O)
s
00
o
u
r*~
10
ti
4J
in
3
"O
- C
^'"~
73 S-
C CU
3 x:
O_t_^
•T"^
u_ o
1 ^
c
to
u
S-
cu
_J
c
to
S-
S-
tO
r-
in
U
T3
O
s_
Q_
^^
ca
to
E ex
to s-
J3 O
(0 CJ
«I
rH
CD
CT>
CU
CJ
to
C
3
4-
-P
cn
to
CQ
.j
^
to
CJ
cu
^_
•^x*
„
-a
c
to
r—
cn
ex
i_
o
CJ
^«
tO
o
p;
CU
x:
o
TJ
CU
,«
^
<£
^-^
TS
co
c
i \
c
o
o
9-21
-------
to
••- re s-
o ra re
re cu cu
Q. E >>
u -o s-
c cu
cu re a.
££ IO
O 3
O O
in
en
CD
CM
CO
00
CT>
CM
cu cu
o
03
CU
U
a>
ta
ra
JC
u
ra
cu
ra
ra
u
ra
ra
cu
•p
ro
u
cu
o i— ra
i- T-
cu -p s-
o. ra P
cu
•P
ra
u
ra
CL.
rO
cu a.
CD r-
^
ra <
•i~*
-C C
D. S- /~s
i— a> .
cu -P a.
73 10 S-
ra ra o
r- LU O
O
C33
C
ra
to
o
cu
ra
ra
ra
cu
re
•a
cu
S- CU
CU T3
ja 3
3 T)
cu
cu
S- (/) i—
cu cu
TJ in -P
•i- CO «
C -P -P
o c
u cu -p
s- ra
in ra
•r- CL O)
P
co c: u)
U) CU
J-
cu -P
ra
ra c cu
E o o
ra -5 -P r-
ra re
u o
e ra
a.
cu
o
ra
ra
cu
cu .a c T-
O)
9-22
-------
8
d
oo
s
i
c
D
o>
JO
CL
0
CO
cn
•C
i,
il
9-23
-------
the number of coke plants in each State. Pennsylvania and Ohio contain 10
coke plants each, and Indiana has 8 plants.
Table 9-10 divides the United States into 11 coke-consuming and
coke-producing regions and shows the amount of coke produced in each region
arid the locations of coke consumption. Most of the regions produce the -v
bulk of the coke they consume; only three regions produce less than 80
percent of their own consumption arid only one produces more than it needs
for its own consumption. Transportation of coke across long distances is
avoided whenever possible to reduce breakage of the product into smaller,
less valuable pieces and to minimize freight charges.
The concentration of production or capacity in specific firms may have
economic importance. Table 9-11 presents the percent of total capacity
owned by the largest four (of 30) firms. The four-firm concentration ratio
for the coke industry has changed little over the years. In 1959, the
four-firm concentration ratio was 53.5 (the top four firms own 53.5
percent of total capacity)40; in 1980 it was 54.4 percent.
In the preceding discussion, furnace and foundry coke production are
considered jointly. However, each existing coke battery may be considered
a furnace or foundry coke producer, based on the battery's primary use.
Separate capacity-based concentration ratios for the two types of coke are
calculated based on this allocation. The 1980 four-firm concentration
ratio for furnace coke is 60.0; the 1980 four-firm ratio for foundry coke
is 57.8.
Concentration in the steel industry has economic relevance because a
large fraction of all furnace coke is produced by integrated iron and steel
companies. Historically, the eight largest steel producers have been
responsible for approximately 75 percent of industry production. However,
from 1950 to 1976, the share of production attributable to the top four
firms declined from 62 percent to 53 percent.41
In summary, concentration exists in the production of both types of
coke and in steel production. However, the concentration is not sufficient
to guarantee market power, and many companies are involved in the pro-
duction of both coke and steel products. Other factors must be considered
in any final assessment of market power.
9-24
-------
1 s
°
II
II
S!
S
^
5
1
£
5 §
^ r
<-> c
LU cn 1 a,
^§
z
a.
^
o
S
s:
»— i
*»
0 X
n 2 r-
)
^i g i_ •> »
CD fO 11 1 ^— *2Z
US- 0» i«i r-
Di C
—i tO \\ T-
: 2:
3 I-H
3
t— 4
aV
5
„
. *• ID
1.0 ^~ CM I***
VO •
=3 5
O)
*» •» 2E
(tf O
•i- TJ
£ £ "c •" «
o o to o c
it- i — i — c. m
•r- 0 >> T- -r-
r- 0 S- r- -0
TO "5 < — *—
o 2; >— i •— i
co co I-H r-*- o
co in r- rH CM
rH
in -jr _,
•r- - 3 ™
"ST" QJ **~
"*— TT f—
^ u> (O "3
>, in C O >
I? 0) tO 4-> •—
o c O) in .>>
4J QJ X- C 0 C
C 1— U C '<- C
ai «i— «r- x: a)
m s: s: o Q-
9-25
cn cn
cn CO
-------
TABLE 9-11. PERCENT OF COKE CAPACITY OWNED BY TOP FIRMS
(JANUARY 1980)35
Firm
U.S. Steel, Inc.
Bethlehem Steel Corp.
J&L Steel Corp.
Republic Steel Corp.
Sum of largest four firms
Capacity
(103 Mg)
14,002
7,651
5,469
5,250
32,372
Percent of
total capacity
23.52
12.85
9.19
8.82
54.38
9-26
-------
9.1.4.2 Integration Characteristics. When one firm carries out
activities that are at separate stages of the same productive process,
especially activities that might otherwise be performed by separate firms,
that firm is said to be vertically integrated. Through vertical integra-
tion, the firm substitutes intrafirm transfers for purchases from suppliers
and/or sales to distributors. A firm may seek to supply its own materials
inputs to ensure a stable supply schedule or to protect itself from
monopolistic suppliers. The firm may seek to fabricate further or
distribute its own products to maintain greater control over the consuming
markets or to lessen the chance of being shut out of the market by large
buyers or middlemen. Therefore, the presence of vertical integration may
constitute a firm's attempt to control costs or ensure input supplies.
Vertical integration does not guarantee market power (control over market
price).
Many coke-producing firms, especially furnace coke producers, are
vertically integrated enterprises. As previously mentioned, 45 of the
existing coke plants are captive; i.e., they are connected with blast
furnaces and/or steel mills. In addition, many coke firms own coal mines,
and greater than 61.0 percent of the coal used in ovens was from captive
mines in 1979.23 Assurance of coal supply to coke production and coke
supply to pig iron production appears to be the motivation behind such
integration.
One implication of vertical integration is that much of the furnace
coke used in the United States never enters the open market—it is consumed
by the producing company. Accordingly, the impact analysis for furnace
coke (Section 9.2.2) uses an implied price for furnace coke based on its
value in producing steel products, which are transferred on the open market.
9.1.4.3 Substitutes. Substitutes for a given commodity reduce the
potential for market power in production of the commodity. The substitu-
tion of other inputs for coke in blast furnaces is somewhat limited, but
not totally unfeasible. In addition, electric arc furnaces, which do not
require coke, are becoming increasingly important in steel production. The
recent trend toward electric arc furnaces and mini-milIs has eased entry
into the iron and steel industry, which in turn reduces market power.
9-27
-------
Imported coke can also be substituted for domestically produced coke.
In fact, although U.S. iron and steel producers prefer to rely on domestic
sources of coke, coke imports have increased recently. If the cost of
domestic coke increased substantially compared to the cost of imported
coke, U.S. iron and steel producers might attempt to increase imports even
more.
Furthermore, substitutes exist for the final products (iron and steel)
to which coke is an input. Increases in the price of coke and the result-
ing increases in the price of iron and steel products can lead to some
substitution of other materials for iron and steel, which also reduces
market power in the production of coke. Analagous substitutions for
foundry coke are possible, and cupola production of ferrous products, which
uses foundry coke, has competition from electric arc furnaces that do not
use coke. Hence, there is a technological substitute for foundry coke in
the manufacture of ferrous products. Furthermore, imported foundry coke
can be substituted for domestic foundry production. In conclusion, some
substitution for coke is possible in the manufacture of both steel and
ferrous products.
9.1.4.4 Pricing History. As previously indicated, a significant
portion of all U.S. coke production is not traded on the market. However,
the Bureau of Mines collects annual data on coke production and consumption
and gives the quantity and the total value of coke consumed by producing
industries, sold on the open market, and imported. Dividing total value by
quantity yields an average price for each of these categories. Time-series
data on these three average values are given in Table 9-12. (Furnace and
foundry coke are combined in these figures.)
Also shown in Table 9-12 are data on the average value of coal that is
carbonized in coke ovens. An examination of coke and coal prices reveals
that increases in coal prices generally coincide with increases in coke
prices. In fact, only 3 years show an increase in the price of coal that
was not accompanied by an increase in the price of the two categories of
coke. Although it is impossible to conclude from this trend that
individual firms have market power, it indicates that the industry can pass
through some increases in costs.
9-28
-------
TABLE 9-12. COMPARISON OF COAL PRICES AND DOMESTIC AND IMPORTED
COKE PRICES2 4 7
Average value of
coal
Average value of Average value of
carbonized. oven coke used oven coke sold Average value of
in coke ovens ' by producers commercially imported coke '
($/Mg)
($/Mg)
($/Mg)
($/Mg)
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
9.56
9.85
10.17
10.19
9.92
9.74
10.31
10.92
10.90
10.89
10.90
10.79
10.86
10.46
10.23
10.48
10.78
11.05
11.03
11.49
13.46
15.43
17.34
20.19
40.22
48.73
48.68
50.99
57.37
14.26
14.50
15.11
15.36
17.33
17.90
19.39
19.98
19.82
19.16
19.92
19.12
19.53
18.88
19.17
17.89
18.40
18.58
19.57
21.54
30.30
32.86
35.76
41.34
82.32
92.84
93.83
90.57
105.79
14.54
15.72
17.63
17.96
18.95
18.52
20.27
21.51
21.90
23.03
22.32
23.30
23.36
23.24
22.85
23.90
24.49
24.99
24.25
27.01
33.04
41.29
44.87
47.31
72. 47
96.61
104.01
111.95
118.03
13.34
13.17
15.96
12.02
11.98
12.26
12.38
14.43
14.25
12.89
. 13.06
13.44
14.42
14.78
16.10
16.95
20.60
20.41
22.31
21.36
25.46
31.93
27.70
40.16
60.14
94.84
93.35
--
•• —
Both furnace and foundry coke and the coals used to produce furnace and
foundry coke are included in these figures.
Market value at the oven (current dollars).
cGeneral customs value as reported by the Department of Commerce (current
dollars).
9-29
-------
9.1.4.5 Market Structure Summary. Although there is no perfect
method for measruing the extent of market power, the preceding sections
addressed four characteristics used to measure the potential for market
power—concentration, integration, substitution, and historical price
trends. Concentration statistics indicated that some potential for market
power exists in the coke industry, yet, these statistics are not conclusive
proof. Similarly, vertical integration in the steel industry is not ^
conclusive in identifying the presence of market power because vertical
integration is a method of controlling the cost and ensuring the quality
and supply of inputs. Finally, the possibility of substitution represents
a strong argument against the existence of extensive market power in the
coke-making industry.
9.1.5 Financial Performance
Financial data on many of the coke-producing firms or their parent
firms, including captive and merchant furnace and foundry producers, are
shown in Table 9-13. (Data for other firms were not available.) From the
financial data in Table 9-13, three ratios have been calculated for each
company (Table 9-14). The first, a liquidity ratio, is a measure of a
firm's ability to meet its current obligations as they are due. A
liquidity ratio above 1.0 indicates that the firm is able to pay its
current debts with its current assets; the higher the ratio, the bigger the
difference between current obligations and the firm's ability to meet them.
All of the coke-producing firms have liquidity ratios between 1.0 and 3.0.
These figures are consistent with liquidity ratios for firms in a wide
variety of manufacturing industries.
The second ratio, a coverage ratio, gives an indication of the firm's
ability to meet its interest payments. A high ratio indicates that the
firm is more likely to be able to meet interest payments on its loans.
This ratio can also be used to determine the ability of a firm to obtain
more loans. The coverage ratio of the coke-producing firms ranged from 1.5
to 15.5. Such ratios are equal to or higher than the coverage evidenced in
most manufacturing industries.
The last of the ratios, a leverage ratio, indicates the relationship
between the capital contributed by creditors and that contributed by the
owners. Leverage magnifies returns to owners. Aggressive use of debt
9-30
-------
s
03
rH
S
O£.
U.
C3
Z
J-H
CJ
^>
0
o
O-
tu in
^ s-
f*i rO
O i—
r—
Z 0
0 T3
Z 00
0£-
•-» cn
sE-o
o fc
tn
•P CU
C 'I—
CU -P
5- •<-
J- r-
3 T-
•t—
4->
c tn
CD -P
S- CU
1 s- tn
3 tn
CJ tO
"
15
*» "
£
*
z
1
(0
CJ
t-H 0
x: xi
— i -P I —
1 t-t
S !! CO
O
z
- ""-r"_T.J>_rr^r COOOVOinrH VDrHrHCMOOCn
orHijnin^-ooot^corH §^S-*vD cMr-omcMr-.
^§rHCOCM^COC^VOCMOO°Cn^CO » m CM H in rH
rn'H cJ rH CO rH
l*""f
oocnrHooor-ooooo S^oSS SocnL75S~
SSS^SS^^SmOO H « « «« VO^O^^^in rH
_T ^" _T «T 10" CM vo" CO o" VD 00 CO 0 CO CO CO Cn O O rH CM CO
SS^S^ScXrHrH^l? ? U « "3 °° -~ ?3 ^ 5 -8 *
CO" ' rH
-i r, I** C*J in 0 tV 0 0 !** C> CO ID 0 CM <• «* CO r-j O CM Cn
siSsssss§3S assss ss.ss^^
ft!f"sga|gag3| i"a|a» ~SSSSS
CO tO T~l Oi VNI * * r*
CM" . .. r-
cnoin^inorHooooo gwogjo ^^^J^^g
rHSo^SSSrHcninvo rH cn § oo o vo co us r- c» •*
^^sascsassa ss^sa ^"aags;
Cn^^rH^O^^^^^cX ScM^ COCM CMrH
CM"
llsissiSSil S81SS
sasiWgssas issss ssssss
^"uT rH" CM* CO CM CM COCOrHrH COrH rHrH
• o.
oT . . Q- J-
s- y • .* e-e- -§.5 o3 •
35 &^33.33°^ o|
JS5,:^ip !l!I! !!l.«
Iis:"ls|»s^:o$^e^:2»^
, J= -P •— vn o «o
-------
0)
C -P
to o>
r- £=
CO
OTO
CO
«*• CM I
•el- iH
CSJ
T-H UJ rH CO
CO CD
in CO
r->. vo »-i in CM CM
CO
w
rd
en o o co '
rH 00 t-1 I
in CM en co i
in co CM co
O)
CO
CM CSJ
CSJ
t-i iH iH CO CSJ O
co csj
CO
CO
CO
i— CU
n>
3 -P
C U)
C 0)
O
. . O U)
O Q.O 3 i—
c s- — -
.
a. s-
S- U O
.33 U.
Q.-Q T3 •— U
T3COO)O-»->S-US-r—
U
C S CO
I—< CO CO CO
I/) O S- CO -
O -P CO.
U CO i
CO
-P -P -r- I
to co
•p -p
a. >»cux:.p
. O)
u -^
{=. O
i-i O
. rd
• •(—
. O -C
: I-H *o •— -o rd co
CO I
CO
o
.p r-
«r- E O
U
S-r-4J-JT- S--P Q-tO CO rd r- rd -J E O-'r;
— ---
O
CO
.
m co
U -Q
•r- rd
-------
TABLE 9-14. FINANCIAL RATIOS FOR COKE-PRODUCING FIRMS
Company name Liquidity ratio9 Coverage ratio
Armco, Inc
Bethlehem Steel Corp.
CF&I Steel Corp.
Crucible Steel, Inc.
Cyclops Corp.
Ford Motor Company
Inland Steel Co.
Inter! ake, Inc.
J&L Steel Corp.
Kaiser Steel Corp.
Northwest Industries, Inc.
National Steel Corp.
Republic Steel Corp.
U.S. Steel Corp.
Wheel ing- Pittsburgh
Alabama By- Products
Allied Chemical Corp.
Diamond Shamrock Corp.
McLouth Steel
Jim Walter
Koppers Co. , Inc.
Philadelphia Coke
1 •? m IT rM -t-w v*a"t* i r* — -^—
1.99
1.60
1.57
2.51
1.75
1.33
1.67
1.92
1. 27
1.43
2.30
1.71
2.03
1.67
1.63
2.21
1.43
1.96
1.54
1.98
2.24
1.54
Current assets
6.95
4.59
3.02
6.46
9.82
15.26
5.75
2.17
2.63
2.70
5.73
4.83
8.15
2.31
2.51
2.95
4.90
4.36
1.88
3.46
10.76
1.73
Leverage ratio0
1.97
2.09
1.92
2.24
2.22
2.28
2.07
2.15
2.48
2.17
2.38
2.33
1.83
2.00
2.22
2.30
2.54 ,.
2.34
2.51
3.02
1.99
2.48
Current 1iabi1ities
EBIT
Coverage ratio = Annual interest expense
'Leverage ratio =
Total liabilities
Tangible net worth
9-33
-------
increases the chance of default and bankruptcy. The chance of larger
returns must be balanced with the increased risk of such actions. The
leverage ratio indicates the vulnerability of the firm to downward business
cycles. Also, a high ratio reveals a low future debt capacity, i.e. addi-
tions to debt in the future are less likely. The firms with coke-making
capacity had leverage ratios that ranged from 1.8 to 3.0. These figures
are relatively high among leverage ratios for firms in many manufacturing
industries. Firms with coke-making capacity are engaged in substantial
amounts of debt financing.
Another measure of financial performance is the rate of return on
equity. Studies of the iron and steel industry show low rates of return on
equity. In an analysis performed by Temple, Barker, and Sloane, Inc.
(TBS), the real (net of inflation) rate of return in the steel industry was
estimated to be 0.2 percent for the period 1970 to 1980. The TBS analysis
projects a rate of return on equity of 1.0 percent for 1980 to 1990.44
These estimates of historical and projected return on equity compare very
poorly with estimates of the required return on investment in the steel
industry. A difference between realized and required returns implies that
equity financing of capital expenditures may be difficult.
As noted, low rates of return on equity affect common stock prices and
have implications for future investment financing, including environmental
control expenditures. For the steel industry, issuing new stock to raise
investment capital is unlikely under current circumstances. If environmen-
tal and other control investments cannot be financed through new equity,
another source of funds must be found. Increased debt is one potential
source. However, firms with coke-making capacity already have incurred
substantial amounts of debt. The TBS analysis concludes that to avoid
deterioration in its financial condition, the steel industry is likely to
reduce expenditures to modernize productive facilities rather than increase
its external financing.45
9.1.6 Projections
The demand for coke is derived from the demand for steel produced by
processes that utilize coke. Hence, projection of the future production of
steel by process type is a necessary precursor to the development of
projections of coke production and coke capacity requirements.
9-34
-------
In the initial study, steel industry projections developed by Data
Resources Incorporated (DRI) were used. However, the DRI projections
developed during 1979 are for a short time period (up to 5 years), whereas
projections of the economic impact of the quench tower standard are re-
quired for years up to 1995. A revised projection of steel production by
process type (basic oxygen furnace, open hearth, and electric arc) for
1985, 1990, and 1995 has been developed and is presented in Table 9-15.
This projection is based on two sources:
1. "Environmental Policy for the 1980's: Impact on the American
Steel Industry," Arthur D. Little, Inc. 1981.
2. Memorandum from Don Anderson, Economics Department, Research
Triangle Institute, to Dave McLamb, U.S. Environmental Protection
Agency. November 20, 1981.
In developing the projections for 1995, it is assumed that the growth of
the projected variables between 1990 and 1995 will be the same as the
growth pattern between 1980 and 1990.
Table 9-16 presents the projection of furnace-coke consumption,
furnace-coke production, and furnace-coke imports for 1985, 1990, and 1995.
The projected furnace-coke consumption is based on a continuation of
historical trends of furnace-coke consumption in hot-steel production
(steel produced in basic oxygen and open-hearth furnace processes) and the
projected steel production presented in Table 9-15. Furnace coke capacity
requirements are projected assuming a capacity utilization rate of 85
percent by the coke producers during the period.
Coke capacity projection is sensitive to the level of coke imports.
Hogan and Koelble46 assert that coke suppliers in western Europe and Japan,
which are the major foreign coke suppliers to the U.S. steel industry, are
not likely to export substantial additional quantities to the United
States, in spite of the fact that U.S. coke imports have been growing
steadily. If so, coke imports for 1985 and 1995 are more likely to remain
at about the 1985 level of imports of 3.5 million Mg during the 1985-95
period. In Table 9-17, the coke capacity (furnace coke plus foundry coke)
is projected under two scenarios: Scenario 1 is the long-run capacity
projection of Table 9-15, and Scenario 2 is the capacity projection,
assuming coke imports at the projected 1985 level through 1995.
9-35
-------
TABLE 9-15. SUMMARY OF THE PROJECTIONS FROM THE LINEAR MODEL*
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
B!
Variable
World coke production
Furnace coke production
Foundry coke production
Furnace coke consumption
Foundry coke consumption
Coke imports
Coke exports0
Coke capacity
Capacity utilization6
Coal rate
Capital /output ratio
VHP of cokef (1979 $/Mg)
1985
427,015
39,893
2,959
43,440
2,536
3,547
423
49,115
84.37
1.40
14.00
208.27
Projections
1990
460,270
35,933
2,914
40,660
2,491
4,727
423
49,115d
80.52
1.36
18.59
262.55
Note: Figures for variables 1-8 are in thousand megagrams.
Figures for variable 9 are in percent.
Figures for variable 10 are in megagrams of coal per megagrams of coke.
Figures for variables 11-12 are in current dollars per.megagram.
aThe projection methodology includes no explicit assumption of additional
controls like those assessed in this report. Projections are based on nor-
mal growth and intended to represent long-run trends in the industry.
U.S coke production = coke demand + coke exports - coke imports.
cAssumed constant throughout the decade.
d!990 coke capacity is assumed to equal 1985 coke capacity.
eCapacity utilization = U.S. coke production/coke capacity.
fVMP stands for value of marginal product. This is a measure of the implied
price of furnace coke based on its value in the production of steel pro-
ducts. Historical estimates of VMP were based on econometric analysis of
production functions for steel.
9-36
-------
TABLE 9-16. SUMMARY OF STEEL INDUSTRY PROJECTIONS
Projections
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Variable
U.S. steel production3
Proportion, basic oxygen furnace
Proportion, electric arc furnace3
Proportion, open hearth furnace3
U.S. steel consumption
Steel imports
Steel exports
Labor productivity
Producer price index of steel
mill products (1967 = 100)
Producer price index of ferrous
scrap (1967 = 100)
1985
132,723
64.25
27.83
7.92
154,442
23,913
2,194
347.04
287.5
324.7
1990
137,713
66.50
30.98
2.52
162,534
27,298
2,477
411.39
323.5
383.2
Note: Figures for variables 1 and 5-7 are given in thousand megagrams.
Figures for variables 2-4 are in percent.
Figures for variable 8 are in thousand megagrams per employee.
3Based on estimates by Data Resources, Inc.28
Steel consumption = steel production + steel imports - steel exports.
9-37
-------
TABLE 9-17. PROJECTIONS OF COKE CAPACITY REQUIREMENTS'
1985, 1990, and 1995
Capacity requirements
scenario
1
2
Projections (10s Mq/yr)
1985
49,587
[46,106]
49,587
[46,106]
1990
48,578
[45,149]
49,966
[46,538]
1995
47,558
[44,182]
50,334
[46,959]
aCoke capacity includes furnace and foundry coke. Figures in brackets
represent furnace coke capacity.
Note: Scenario 1 assumes imports to grow at the long-term trend;
Scenario 2 assumes the imports for 1990 and 1995 at the 1985 trend
1 eve!.
9-38
-------
Forecasts of U.S. coke demand are very sensitive to forecast steel
production and technology. Other projections have been made of domestic
coke needs in 1985. In a Merrill Lynch Institutional Report, Charles
Bradford forecast furnace coke consumption for 1985 at between 38.1 and
43.5 million megagrams.43 Blast furnace consumption assumed to be 92 to
93 percent of total coke consumption (figures for recent years) corre-
sponds to a forecast of total coke demand of 41 to 46 million megagrams. .,
The projection presented in this report is at the high end of that range.
However, Hogan and Koelble and Lawrence R. Smith (Koppers) forecast a much
higher coke demand for 1985. They project the demand for furnace coke
alone at 51.7 to 53.5 million megagrams.28 These sources do not directly
address foundry coke demand; consequently, the projections for foundry coke
production cannot be compared.
9.1.7 Market Behavior: Conclusions
Market structure, financial performance, and potential growth
influence the choice of a methodology to describe supply responses in the
coke-making industry. Although some characteristics of this industry
indicate a potential for market power, other characteristics belie it.
Some concentration exists in coke-making capacity and steel produc-
tion; however, many firms produce coke and iron and steel products.
Vertical integration is substantial; however, integration appears to result
primarily from a desire for increased certainty in the supply of critical
inputs. Furthermore, substitution through alternative technologies and
coke imports is feasible, and some substitutes for the industry's final
products (iron and steel) are available. In any industry, the potential
for substitution is a major factor leading to competitive pricing.
Certainly, the financial profile of coke-making firms is not indicative of
monopoly profits. Prospects for industry growth are limited. An indi-
vidual firm must actively compete with other firms in the industry to
improve its profit position.
No industry matches the textbook definition of perfect competition.
The important issue is whether or not the competitive model satisfactorily
captures major behavioral responses of firms in the industry. Based on the
factors outlined in this section, the competitive pricing model adequately
describes supply responses for coke-making firms.
9-39
-------
9.2 ECONOMIC IMPACT OF REGULATORY ALTERNATIVES
9.2.1 Summary
Economic impacts have been projected for the baseline and for each
regulatory alternative. Furnace and foundry coke impacts are examined
separately because their production costs and markets differ. All cost and
price impacts are in third-quarter 1979 dollars. To convert to 1982 dollars,
multiply by 1.25 which is the ratio of the 1982 producer price to the same
index for the third quarter of 1979, as updated in the Survey of Current
Business.46 47 When measured on a per-unit basis, the costs of meeting
baseline regulations for foundry coke plants tend to be greater than those
for furnace coke plants for two reasons. First, some economies of scale
are present for some of the controls. Foundry plants tend to be smaller
than furnace plants, thus, they have higher per-unit control costs. Second,
for a given battery, foundry coke output will be less than furnace coke
output because foundry coke coking time is about two-thirds longer than
furnace coke coking time.
Recovery credits cause Regulatory Alternative II to result in annualized
costs of $-1 million for furnace and foundry coke producers combined.
Regulatory Alternative II requires capital expenditures of $23 million for
furnace and foundry coke producers combined. Regulatory Alternative III
would result in annualized costs of $42 million and capital costs of $161
million for the combined furnace and foundry coke sectors. Full compliance
with baseline regulations measured against the existing state of control
results in annualized costs of $436 million and capital costs of $1,159
million.
Price impacts are estimated under the empirically supported assumption
that furnace coke demand is responsive to higher coke prices. Foundry coke
demand is also assumed to respond to price. Regulatory Alternative II
would have negligible impacts of $0.02/Mg and $0.19/Mg (less than 0.10 percent
change) on the prices of furnace and foundry coke, respectively (1979 dollars).
Regulatory Alternative III would result in furnace and foundry coke price
increases of $0.70/Mg (0.05 percent) and $1.44/Mg (0.77 percent), respectively.
Compliance with baseline regulations measured against the existing state of
control increases the furnace coke price by 6.4 percent and the foundry coke
price by 15.4 percent.
9-40
-------
Regulatory Alternatives II and III would have little impact on the
production of either furnace or foundry coke. Complete compliance with
baseline regulations measured against existing compliance would decrease
furnace production by 6.6 percent and foundry production by 12.2 percent.
Complete compliance with baseline regulations produces three potential
furnace battery closures and five potential foundry battery closures. The
regulatory alternatives are not projected to result in any battery, plant,
or company closures.
9.2.2 Methodology
The following approach focuses on the long-run adjustment process of
furnace and foundry coke producers to the higher costs of coke production
that the baseline and the regulatory alternatives will create. These
long-run adjustments involve investment and shutdown decisions. Short-run
adjustments, such as altering coking times, to meet the fluctuations in the
demand for coke are not the subject of this analysis.
Because of differences in production costs and markets, furnace and
foundry coke producers are modeled separately. Both are assumed to behave
as if they were competitive industries selling coke in a market. This
assumption is somewhat more realistic for foundry than for furnace coke
producers because most furnace coke is produced in plants captive to the
steel industry. However, interfirm and intrafirm shipments of coke are not
uncommon, as can be inferred from Table 9-10. A pi ant-by-plant review of
the coke industry by Hogan and Koelble also confirms the existence of such
exchanges.48
A set of programmed models has been developed to produce intraindustry
and interindustry estimates of the economic impacts of the alternative
regulations. The models are applied to both furnace and foundry coke, and
the sectors included are coke, steel, and ferrous foundries. The rest of
the economy is incorporated into the interindustry portion of the analysis.
The analytical approach incorporates a production cost model of the
coke industry based on engineering data, an econometric model of the steel
industry, and an input-output model of the rest of the economy and final
demand. The interrelationships of these models for furnace coke are shown
in Figure 9-4. The upper portion of Figure 9-4 encompasses the supply side
impacts of the regulatory alternatives; the lower portion contains the
9-41
-------
OO
—1 0
oo <: i—i
I— Z OJ
CJ •—< Q-
< U-
o. ra
2: z z
• *•-
t
L'J
o
_J
UJ
OD LU LLJ
I— *z 1—
o o co
-!
i-< t—
Q-
<: s
Z =3
UJ CO
O3 CO
t
UJ
:*£
o
CJ >•
2= O.
UJ 1-0
C£
O
CJ
to ~&Z CL.
UJ O C-
S O •=>
o to
a
gS
UJ
o
C- O
O Z
Z UJ
P z
WO H-1
to
o
UJ U.
z z
o
C-9
O
Z UJ
«=c s*:
s: o
UJ C_3
o
1/5
Z =1
CJ
i—« ce
UJ O
Di U.
o
o
*J
o
CO
a
o
o
UJ
CT)
£
.1
u_
O —I
U. UJ
UJ
a i—
z co
-< Z O
X UJ t-i
' * ' y f^
^^ <^
U. O — 1
O Di =3
i— < c;
to > uj
t— z cs
oo uj
0
i
oo
UJ
u. >- >
O OS >— <
o t—
OO 1— <
1— < Z
CO _l Qi
o => uj
O O H-
UJ —I
ce. <
o
z -J
h—
CO OS
UJ O
S u.
o
o
9-42
-------
demand side impacts. In the synthesis step, the two sides are brought
together and the equilibrium price and quantity relationships are deter-
mined. An analogous diagram for foundry coke would substitute ferrous
foundry products for steel. The methodology is described further in the
following subsections.
9.2.2.1 Supply Side. A production cost model that incorporates
technical relationships and engineering cost estimates is used with
plant-specific information to compute separate industry supply functions,
with and without additional controls.49 Supply functions are estimated on
a year-by-year basis for furnace and foundry coke plants projected to be in
existence between 1980 and 1990. Both coke production costs and the costs
that plants incur to meet existing environmental regulations are computed
to estimate the industry supply curve before any additional controls are
applied. Estimates of the costs of control for further compliance with the
baseline regulations and the regulatory alternatives are used to compute
the projected upward shifts in that supply function. All costs are in 1979
dollars.
This approach provides a method of estimating the industry supply
curve for coke, which shows the alternative coke quantities that will be
placed on the market at alternative prices. When the supply curve is
considered in conjunction with the demand curve, an equilibrium price and
coke output rate can be projected. Supply curve shifts caused by the
regulatory alternatives can be developed from the compliance cost estimates
made by the engineering contractor. These new supply functions, along with
the demand curve, can then be used to compute the equilibrium price and
output rate under each regulatory alternative.
9.2.2.1.1 Data base. PI ant-by-plant data on over 60 variables for
furnace and foundry coke plants in existence in 1979 were compiled from
government publications, industry contacts, and previous studies of the
coke industry. The data base was sent to the American Iron and Steel
Institute, which submitted it to their members for verification, correc-
tions, and additions,50 and to the American Coke and Coal Chemicals Insti-
tute.
9-43
-------
9.2.2.1.2 Output relationships. For a given battery, the full capac-
ity output of coke, measured in megagrams per year, is dependent on the
nominal coal charge (megagrams of coal per charge) per oven, the number of
ovens, and the effective gross coking time (net coking time plus decarboni-
zation time). The following values for effective gross coking time were
used except where pi ant-specific values were available.49 50
Wet coal
Preheated coal
Furnace
coke
18 hours
13 hours
Foundry
coke
30 hours
24 hours
An age-specific outage rate that varies from 4 to 10 percent is assumed to
allow for normal maintenance and repair. Thus, the model assumes some
increase in such costs as plants age.
The quantities of by-products produced are estimated from engineering
relationships. These quantities depend on the amount of coal carbonized,
percentage of coal volatile matter, coking time, and configuration of the
by-product facility at a plant. The by-products included in the model are
coke breeze, coke oven gas, tar, crude light oil, BTX, ammonium sulfate,
anhydrous ammonia, elemental sulfur, sodium phenol ate, benzene, toluene,
xylene, naphthalene, and solvent naphtha. All plants are assumed to
produce breeze and coke oven gas.
9.2.2.1.3 Operating costs. The major costs of operation for a coke
plant are expenditures for coal, labor, utilities, and chemicals. The
activities within the coke plant were allocated to five production and ten
environmental control cost centers (Figure 9-5) to facilitate the develop-
ment of the operating cost estimates.
Coal is the major operating cost item in coke production. Plant-
specific estimates of the delivered price of coal were developed by identi-
fying the mine that supplies each plant and estimating transportation costs
from the mine to the plant. When it was not known which coal mine supplied
a particular plant, it was assumed that the coal came from the nearest
mines supplying coal of the same volatile matter and ash content as that
used by the plant. Transportation cost estimates were based on the dis-
tances traveled and the transport mode (barge or rail) employed.
9-44
-------
I.
O
u
^-»
c
JO
a
u
sx
o
o
in
i
E
3
9-45
-------
Maintenance labor and supervision requirements were estimated for 69
jobs within the coke plant. Primary variables that determine the number of
maintenance labor and supervision man-hours needed include type of plant
(merchant or captive), number of battery units, number of plants at a site,
size of by-product plant, type of coal charge (wet or preheated), and coke
production. The labor rates used for captive plants were $17.04/h for
supervisory positions and $15.70/h for production labor. For merchant
plants, rates of $15.80/h and $14.40/h were assumed.
The major utilities at a coke plant are steam, electricity, water, and
other fuels. Utility requirements were estimated from the data on the
plant configuration and output rates for coke and the by-products. The
prices used for the utilities are $5.44/103 Ib steam; $0.027/kWh electric-
ity; $0.16/103 gal cooling water; and $2.76/106 Btu underfire gas.
9.2.2.1.4 Capital costs. Although no net additions to industry
coke-making capacity are anticipated during the 1980 to 1990 period, a
number of producers have plans to rebuild or replace existing batteries.
Such actions will alter the long-run industry supply curve because the new
batteries will typically have lower operating costs per unit of output than
the batteries they replace and, most importantly, their capital costs will
be reflected in the new supply curve. Hogan and Koelble43 have identified
pi ant-by-plant rebuild/replacement intentions. These plans are included in
the data base. The cost of building a model new coke battery and the cost
of major rehabilitation of an existing battery have been estimated for the
affected facilities. It has been assumed that new furnace construction
will be 6-meter batteries and new foundry construction will be 4-meter
batteries. Pad-up rebuilds are assumed to leave the battery size unchanged.
Pad-up rebuild costs were estimated as a function of battery size. A zero
salvage value is assumed for existing batteries.
The capital cost breakdown for new plants is shown in Table 9-18. For
such plants, the major capital cost items are the battery, quench tower,
quench car, pusher machine, larry car, door machine and coke guide, by-
product plant, coal handling system, and coke handling system. A 60-oven
battery is assumed. Pipeline charging can increase the coke-making
capacity of a given oven by about 25 percent by reducing gross coking time.
9-46
-------
TABLE 9-18. ESTIMATED CAPITAL COSTS OF NEW BATTERIES51
Conventionally
charged battery
Capital costs by element
(106 1979 dollars)
Pipeline
charged battery
Capacity (103 Mg/yr)
4-metera 6-metera
450 720
4-metera 6-metera
560 900
Coke battery
Quench tower with baffles
Quench car and pushing
emissions control
Pusher machine
Air-conditioned larry car
Door machine and coke guide
By-product plant
Coal -hand! ing system
Coke-handling system
Off sites
Total
34.20
2.45
6.58
2.50
1.72
1.80
32.50
18.20
6.85
1.60
$108.40
48.90
2.85
7.92 .
3.20
2.28
2.10
39.75
23.60
8.80
1.80
$141.20
64.60
2.45
6.58
2.40
0
1.80
35.76
20.62
7.74
1.69
$143.74
83.70
2.85
7.92
3.20
0
2.10
43.74
26.70
10.00
1.91
$182. 12
In the production cost model, new foundry batteries were assumed to be
4-meter batteries and new furnace batteries were assumed to be 6-meter
batteries.
9-47
-------
Consequently, the per-unit operating cost is reduced. The capital costs
show economies of scale, i.e., larger plants have smaller per-unit-of-
capacity capital costs. The capital cost per unit of capacity is higher
for pipeline-charged batteries than for conventionally charged batteries.
Periodically, batteries must undergo major rehabilitation or
rebuilding because of performance deterioration. The costs of pad-up
rebuilds will vary from site to site depending on battery maintenance, past
operating practices, and other factors. Average estimates of the cost of
rebuilding were developed for this study and are shown in a report by
PEDCo.49 The economic life of coke-making facilities is subject to
considerable variation depending upon past maintenance and operating
practices, which also affect current operating costs. For this study, 25
years was used as the average preferred life of a new coke-making facility;
however, many batteries are operated for 35 to 40 years. If 35 to 40 years
is a more reasonable battery lifetime, use of a 25-year lifetime will
result in some overestimation of the annual costs of new or rebuilt facili-
ties. However, firms will probably not plan or expect to wait 35 to
40 years to recoup an investment in coke-making capacity.
9.2.2.1.5 Environmental costs. Plant-specific estimates of the
installed capital and operating costs for current and proposed environ-
mental regulations and the regulatory alternatives under consideration in
this study were incorporated in the model. The current and proposed
regulations include workplace standards (Occupational Safety and Health
Administration) [OSHA], water quality regulations best practicable
technology [BPT] and best available technology [BAT], State implementation
plan (SIP) requirements, and proposed air quality regulations for topside,
charging, and door leaks (National Emission Standard for Hazardous Air
Pollutants) [NESHAP] and quench towers (New.Source Performance Standards
[NSPS]. Compliance expenses already incurred for all plants in the data
base for each of the current and proposed regulations (existing control
costs) were estimated. Therefore, it was possible to estimate the remain-
ing environmental costs to plants to meet current and proposed regulations
(baseline control costs). It has been assumed that costs to comply with
9-48
-------
OSHA and BPT water requirements under the Federal Water Pollution Control
Act will have to be incurred by 1981. Costs for all other existing
environmental regulations are assumed to be incurred by 1983.
The scatter diagrams in Figures 9-6 and 9-7 show estimates from the
coke supply model of average total cost of production in 1980, including
environmental costs, for all furnace and foundry coke plants. A weak,
inverse relationship between the average cost of production and the size of
the plant is evident in Figures 9-6 and 9-7. However, a number of other
factors create variability in the average cost of production across coke
plants. The most important of these factors are the delivered price of
coal, the age of the plant, and the by-products recovered.
9.2.2.1.6 Coke supply function—existing facilities. The operating
and capital cost functions were used to estimate the cost of production,
including relevant environmental costs, for all plants in the data base.
This cost does not include a return on investment for existing facilities.
The capital costs for these facilities have already been incurred and do
not affect operating decisions.
Capital costs that have not yet been incurred are annualized at 6.2
percent, which is estimated to be the real (net of inflation) cost of
capital for the coke industry. (This percentage is an after-tax estimate.)
This figure, which was estimated from data on the capital structure for
publicly owned steel companies, has been used in this study as the minimum
acceptable rate of return on new facilities.52
The capital costs of controls affixed to coke oven batteries are
annualized under two different assumptions. For scenario A, it is assumed
that when a battery reaches the end of its useful life, it is rebuilt or
replaced by a battery of the same height. If this situation occurs, most
of the control equipment is salvageable.53 Accordingly, under scenario A,
each annualization is performed over the life of the control equipment.
However, not every battery is rebuilt.or replaced at the end of its
useful life. Similarly, some old batteries are replaced by new batteries
that are not comparable in size (height). In such cases, capital expendi-
tures for affixed,controls must be recouped by the time battery retirement
occurs. Under scenario B, this control equipment is assumed to be com-
pletely unsalvageable. Annualizations are performed over the remaining
9-49
-------
o
CD—r
00-
1
M O
»— OJ-r
O —
i
a.
£
CO
o
o
A A
A
.i
i
o
o —
8-7
* * 4
* A A
O.
oo
—,—f—,—I—I
300 600 300 1,200 1,500 1,800
PRODUCTION CAPACITY CTONNES*1'000/YR5
Figure 9-6. Estimated average cost of furnace coke production as a function
of plant production, 1980.
2, 100
9-50
-------
o
I—
o
(O —
LJ
o
r—
s
I—
O
o
£ i
o
o
CVI-
o
o-
I
t.
100 200 SCO
PRODUCTION CAPACITY (TONNES*1000/YR)
Figure 9-7. Estimated average cost of foundry coke production as a function of
plant production, 1980.
9-51
4CO
-------
life of the battery or over the control life, whichever is smaller.
Estimates of remaining life for existing batteries are based on a long
total life (40 years) because some batteries are being kept in operation
for 40 years. While 40 years is longer than the preferred average life of
a battery, it is not necessarily longer than the battery's realized life.
The regulatory alternatives for coke oven by-product plants involve
control equipment that is not affixed to batteries. Accordingly, the
equipment is not affected by battery age or size (height) of the battery
replacement. The capital costs of the regulatory alternatives are annual-
ized over the life of the control equipment (10 years). This action is
tantamount to assuming either that all by-product plants have a remaining
life of at least 10 years or that the control equipment is salvageable.
The supply function for each plant is estimated as follows: the
average cost of production is computed for each battery in the plant; these
batteries are arranged in increasing order of their average costs of
production and the output for each battery is accumulated to produce a
stepped marginal cost function for the plant; plant overhead costs are
averaged for all relevant plant output rates; and average'total costs are
computed for each output rate by summing the average costs for plant
overhead and the battery. Each plant's supply function is the portion of
the marginal cost function above the average total cost function. For -
existing plants where the average total cost exceeds marginal cost over the
entire range of output, the supply function is the point on the plant's
average total cost function represented by capacity output (after allowing
for outages). The aggregate long-run supply function for all currently
existing coke plants and batteries is obtained by horizontally summing the
supply function for each plant. The 1980 industry marginal cost (supply)
curves for existing furnace and foundry coke plants are presented in
Figures 9-8 and 9-9, respectively.
9.2.2.1.7 Coke supply function—new facilities. The cost of coke
production for new furnace and foundry batteries was estimated from the
engineering cost model, assuming the new model plants described previously.
These costs include the normal return on investment and allowances for
9-52
-------
O
oo-
o
CO
o
o
o
o-
I
J.
o.
00
AVERAGE COST
MARGINAL COST
0 -
40,000
Figure 9-8. Marginal and average cost functions for furnace coke, 1980.
10,000 20,000 30,000
PRODUCTION CTONNES*1000/YR}
SO,COO
9-53
-------
o
r—r
o
U5-T
i.i JL
O
•^