United States Industrial Environmental Research EPA-600/ 2-79-108
Environmental Protection Laboratory May 1979
Agency Research Triangle Park NC 27711
Research and Development
Reduction of Air
Emissions from Gasoline
Storage Tanks
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-108
May 1979
Reduction of Air Emissions from
Gasoline Storage Tanks
by
Arnold Gunther
Stop-Los Company
29 Lorelei Road
West Orange, New Jersey 07052
Contract No. 68-02-2679
Program Element No. 1AB604
EPA Project Officer: Bruce Tichenor
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
The purpose of the project was to determine the technical and economic
feasibility of the use of flexible plastic membranes as a mean to control
emissions emanating from gasoline storage tanks. The emission rates and the
expected life of the membranes were to be established.
A demonstration pilot unit was built and operated. The emission rates
were determined, as well as the life expectancy of the membranes.
The results obtained indicate that emission control of better than 99$
when compared to uncontrolled tanks can readily be achieved and that the
life expectancy of the membranes is in the order of 20 years of continuous
service.
The installed cost of these devices for commercial size applications
has been estimated and found to be very competitive with the conventionally
used floating roofs.
This report was submitted in fulfillment of contract No. 68-02-2679 by
the Stop-Los Company under the sponsorship of the Environmental Protection
Agency. Work was completed as of January 26, 1979.
ii
-------
CONTENTS
Page
List of Figures iv
List of Tables iv
I Conclusions 1
II Introduction 2
III Equipment and Operation 3
IV Method Used for Vapor Losses Determination 11
V Results of the Vapor Loss Determinations 16
VI Accuracy of the Data and the Results 25
VII Interpretation of the Results Obtained 26
VIII Life Expectancy of Liners 29
IX Economics 30
iii
-------
LIST OF FIGURES
No.
1 Schematic Elevation View of Pilot Unit 4
2 Time History of Operational Parameters 5
3 Vapor Loss Determination Apparatus 12
4 Loss Rate for Tank 2 20
5 Loss Rate for Tank 1 23
Laminated Liner
6 New and Old Technologies Cost Comparison 31
LIST OF TABLES
No.
5-1 Vapor Loss Determinations at Tank No. 2 17
5-2 Presaturation Run 18
5-3 Vapor Loss Determinations at Tank No. 1 18
Flouropolymeric Liner
5-4 Vapor Loss Determinations at Tank No. 1 19
Laminated Liner
iv
-------
SECTION I
CONCLUSIONS
1. Prom the data obtained during the operation of the demonstration
pilot unit, it has been concluded that emission control efficiencies of
better than 99$ can readily be achieved when compared to uncontrolled tanks.
2. The emission rates remained essentially constant and independent
of time.
3. The emission rates for dynamic operation of the storage tanks were
found to be larger than the emission rates for static storage operation.
A storage tank whose contents are subjected for extended periods of time to
successive and contiguous cycles of filling and emptying is said to be
under dynamic operation. A storage tank whose contents are withdrawn and
refilled infrequently is said to be under static operation. The emission
rate under dynamic operation is due to gas permeation through the membranes
and to gas leakage through fissures and other imperfections in the attach-
ment areas of the membranes to the tank. The emission rate under static
operation is essentially controlled by gas permeation.
4. The life expentancy for the membranes is in the order of 20 years
of continuous service.
5. The installed cost of these devices for commercial size applications
are very competitive with the conventionally used floating roofs.
-------
SECTION II
INTRODUCTION
The purpose of the project was to determine the technical and economic
feasibility of the use of flexible plastic membranes as a means to control
emissions emanating from storage tanks, holding gasoline and similar pro-
ducts. One of the goals set at the outset of this program was to provide an
emission control greater than 95$ when compared to uncontrolled tanks.
Another goal was to establish the expected life of the flexible membranes
under actual operating conditions, which is closely related to the economical
feasibility. The technology required for the project at hand was made avail-
able by the Stop-Los Company, who is the originator of this technology, and
in which it owns propietcyry rights.
The emission rates from gasoline storage to be expected were known be-
forehand for the materials to be utilized in the construction of the flexible
membranes. However, this knowledge was restricted to a static system, i.e.-
a device built according to the principles of this technology where the
liquid stored would be kept in the tank without adding or withdrawing liquid
at frequent intervals. The effect on the flexible membranes of a frequent
cycling, for an extended period of time, and its consequences, if any, on the
emission rates was one of the paramount objectives of this project.
With the previously cited objectives and goals in mind, a demonstration
pilot unit was built and operated by the Stop-Los Company under a contract
with the U.S. Environment Protection Agency. The following Sections of this
report contain a detailed description of this effort.
-------
SECTION III
EQUIPMENT AND OPERATION
The demonstration unit consisted of two vertical tanks of about 1300
gallons capacity each, fitted internally with flexible membranes or liners,
a circulating pump for transfering the liquid between the tanks, piping,
control valves, and two bladders.
The tanks were filled approximately half-way with water (saturated with
gasoline) on which a layer of about 1" thick of gasoline was added. The
water was transferred back and forth between the tanks by the circulating
pump, thus raising the liquid level in one tank while simultaneously lowering
it in the other.
The upper and lower operating levels were maintained by level controls
electrically interlocked with the control valves. The filling and emptying
of the tanks, between the selected liquid levels, was performed automatically
and continuously at a relatively high frequency. The time required for one
cycle (filling and emptying) was 36 minutes. The flexible membranes were
thus subjected to the same range of cyclic stresses while exposed to the
gasoline and its vapors as they would be in any commercial size storage tank.
Under the assumption of a weekly turnover for a large commercial tank
operation, which is more frequent than normal in industrial practice, each
week of operation for the demonstration unit is approximately equivalent to
one arid a half years of commercial tank operation. (Ten hours per day for
five days per week in the demo-unit operation will subject the liners to
approximately 80 cycles).
The demonstration unit will now be described in detail with the aid of
Fig. 1 and Fig. 2. The tanks Tl and T2 were built out of 0.476 cm (3/16")
thick steel plate. They are vertical cylinders with an open top and a dished
bottom, measuring 1.83 m (61) in diameter by 1.83 m (6 feet) between tangent
lines, supported on three I beam legs, equally spaced along their lower
periphery. The top rim stands at 3.35 m (11*) from the floor. Two sight
levels covering a span of 33 cm (13") each, with the center of the sight
tubes located approximately at the lower and upper operating liquid levels
are provided in each tank. Tank T2 carries two float-type level controls
LCI and LC2 located at about the same level as the sight glasses. Block
valves Bl, B2, B3 and B4 permit isolation of the level controls from the
tank's contents for maintenance and repair purposes. The differential bet-
ween the controlled upper and lower liquid levels in the tank was set at
2.06 m (48").
Each tank carries a glass apparatus, Gl and G2. They are used for the
determination of the gasoline loss and they are mounted in the tank's upper
portions. Their assembly and operation will be described in detail in the
section dealing with the gasoline loss determination. The circulating pump
P, which delivers 178 dm^/min (47gpm) is a vertical shaft centrifugal type,
-------
Gl G2
g g
•KM* r> WT*>
LC2
32
ST2
LL
V6 ,VU f
BL2
Tl P f!2
Figure 1. Schematic elevation view of pilot unit.
-------
BP
ABE
M» •
HL
>J
•
&
TIME AXIS
Figure 2. Time history of operational parameters.
5
-------
totally enclosed and explosion-proof. Its suction and discharge ports are
2" IPS. Its suction port is connected to the tanks' bottom discharge
nozzles and its discharge port to the tanks' side inlet nozzles. The block
valves B5, B6, B7, and B8 permit isolation of the pump and its associate
piping from the tanks.
Strainers ST1 and 3T2 are provided in the station lines to protect the
pump and the solenoid valves from rust, scale, and occasional debris that
may accumulate at the bottom of the tanks.
Solenoid valves S1 and S3 are of the normally open type (they close
when energized) and the valves S2 and 34 are of the normally closed type
(open when energized).
During the operation of the system, the transfer pump is kept always
running. When liquid is being transferred from T2, into T1, valves 32 and
S4 are open and valves 31 and S3 are closed. As soon as the liquid level
in tank T2 reaches the low level, valves 32 and 34 are closed and valves
31 and S3 become open. Liquid now flows from tank T1 into T2. The auto-
matic operation just described is achieved by the proper electrical inter-
lock between the level controls and the solenoid valves. Manual switches
are also provided to override the automatic control when so required.
Valves V1 and V2 are provided for the filling and the draining of the
tanks.
A circular trough TR, whose cross section is depicted in Fig. 1, floats
on the surface of the liquid. It is built of 0.3175 cm (1/8") thick steel
plate, with its outside diameter measuring about 2.5 cm (1") less than the
tank's inside diameter in order to avoid any undue friction or binding with
the tank's shell at any point of its vertical displacement. The inside
diameter of the trough is 1.47 m (5£").
A circular membrane M, built out of a 0.0254 cm (10 mils) thick poly-
propylene sheet and having one of its sides coated with a 0.00254 cm (1 mil)
thick film of an epoxy resin, is bolted to the internal rim of the circular
trough. A neoprene gasket between the rim and the membrane is kept under
compression by bolting a flat steel ring to the rim. This ring lays over the
membrane's periphery and exerts an uniform pressure over the gasket.
A vertical liner L, affecting the shape of a cylinder with a diameter
slightly smaller than the tank's inside diameter, is bolted to the upper
portion of the shell and to the external rim of the floating trough.
Here again, a neoprene gasket 0.3175 cm thick by 5 cm width, which is
located between the liner and the shell, is kept under compression by seg-
mental circular steel strips pressed against the liner by bolts. The seal-
ing of the liner over the trough is done in an identical manner as described
for the circular membrane M. The liner L is made from 0.0254 cm (10 mils)
thick polyurethane coated on one of its sides with a 0.00254 cm thick film
of an epoxy resin. The epoxy film, which provides most of the barrier
-------
against the permeation of the gasoline vapors, faces the tank's shell. The
height of the fully stretched liner is 1.83 m (6').
The vapor space between the liquid surface and the circular membrane M
is kept in communication with the annular vapor space comprised between the
shell and the liner L by the two conduits C1 and C2. These conduits are
built out of two short lengths of 2" pipe (about 15 cm long each) that
traverse the cross section of the trough and are welded to it at their ends.
In order to avoid corrosion of the tanks and troughs, they were coated
with an epoxy paint which has outstanding resistance to attactf by gasoline.
The vapor phase always contains gasoline vapors in a concentration
higher than the one corresponding to the upper flammability limit (about 8$
by volume), and consequently the danger of fire or explosion is practically
non-existing. Nevertheless, it was consided convenient (mainly from a
corrosion viewpoint) to ground the tanks and all their associated piping.
Also a jumper was provided between the circular trough and the tank. An
aluminum rod of 2.5 m (8') long was driven into the soil and wired to the
tanks, thus serving as an anodic protection.
An overhead hoist travelling over an I beam was provided at each tank
to aid in the installation of the liners.
During the operation of the tanks, the floating circular trough, which
follows the continuous change of the liquid level, induces the folding or
unfolding of the vertical liner L. Consequently, the annular space between
the liner and the tank's shell, which is filled with air and the vapors of
gasoline at their prevailing vapor pressure, changes continuously in volume.
It decreases in volume for a rise in the liquid level and viceversa. If
these volume changes would remain uncontrolled they would tend to induce
pressure changes in the vapor space. Pressure would tend to increase for a
raise in the liquid level and to decrease in the reverse case. An increase
in pressure will tend to "balloon-up" the liner with the consequence that
its orderly folding would be impaired and it would not allow the full utili-
zation of the storage tank capacity due to a restriction in the available
travel of the floating trough. The partially folded and "ballooned" liner
would make contact with the tank's roof before the liquid level could attain
its maximum height. On the other hand, a decrease in pressure would tend to
create a sizable vacuum in the vapor space with the consequence that the
liner would be pushed and pressed against the shell by the atmospheric
pressure. The frictional forces so developed might become large enough to
hold the liner so firmly pressed against the shell that the floating trough
literally hangs-up. This in turn may stretch the liner beyond its safe limit
and produce its failure. All of the above problems are avoided by providing
a variable capacity container properly connected to the vapor space.
The bladders BL fulfill this requirement. Each is built out of the
same flexible material as the vertical liner, and is essentially a flat
cylinder, with both ends closedj it collapes and expands when very small
-------
pressure differentials are applied across its surface. Its enclosed volume,
when fully expanded, is larger than the tank's annular vapor space
volume. Although in industrial practice the bladders would be located inside
the storage tank and attached to the tank's roof, in the present case, be-
cause the tanks are open top, the bladders are located outside the tanks.
The top of each bladder is cemented to a 0.9144 m (3*) diameter disc of
plywood and hangs freely from it. The valves V3 and V4 permit isolation of
the bladders' content from their respective tanks, whose vapor spaces are
connected to the bladder by a 2.54 cm (1") I.D. vinyl flexible tubing.
Valves V5 and V6 serve for bleeding of the bladders into the atmosphere,
which is done when the tanks are initially charged with the gasoline layer.
In any position of the freely hanging bladder, except for the totally
expanded one, the weight of the hanging portion is balanced by the pressure
differential between the atmospheric pressure and the pressure inside the
bladder, which consequently is under a slight vacuum. This vacuum is of an
order of 0.5 mm of water column for the presently used size and shape of
bladders. Very slight imperfections in the sealing areas of liners may allow
some air to leak in which will eventually collect in the bladders.
In order to help in the understanding of the mechanism of the leakage of
air and its eventual expulsion from the vapor space back into the atmosphere,
a sequential operation of a tank will now be described. Let us begin with
the liquid level located at or near its lowest working level. Let us further
assume that the tank is now starting to be filled after being kept at the low
level for a substantial period of time, during which the bladder has reached
its fully expanded position due to air leaking in. The rise in the liquid
level will now tend to compress the gas contained in the annular vapor space.
The gas, now unable to flow into the bladder which is completely filled,
will flow from.the annular space,into the space comprised between the mem-
brane M and the surface of the liquid. The membrane will displace upwards,
from its initial position M1 to a final position M% which is attained when
practically all of the gas in the annulus has been transferred, which in
turn occurs when the higher operating liquid level has been reached. The
gas enclosed by the liners and the bladder is now under a slight pressure
caused by the weight of the membrane M, which is held suspended by the gas
pressure. For the membranes utilized in this project, the pressure exerted
is on the order of 0.3 mm of water column. If the tank is kept for a sub-
stantial time without unloading, the pressure differential across the liners
will eventually cause a leak-out of the gas contained under the membrane M.
A withdrawal of liquid now will cause the gas contained in the bladder to
flow into the annular space.
Let us now consider the case where the tank is being filled and emptied
continuously, which constitutes the other extreme of the variety of possible
operational modes. In this case, a steady state is obtained in which the
mass rate of air leaking in is equal to the mass rate of air leaking out.
The air leaking out is essentially saturated with the vapors of gasoline.
At the steady state the gas bounded by the liners undergoes cyclic changes
of pressure, with the period of the pressure cycles (time elapsed between
two crests or two troughs of the pressure wave) equal to the period of the
8
-------
liquid level. Fig. 2 shows the liquid level change between the high (HL)
and the low (LL) levels and the pressure wave, both as a function of time.
It also indicates the positions of the flexible membrane M, between its two
extremes M1 and M", and the expansion and contraction of the bladder,
between its fully expanded and its fully empty positions: BF and BE. It
may be noticed that all of them are periodic functions of time with the
same period 6.
The pressure wave is a step function and indicates that the gas which
is under a slight pressure during a fraction of the period 9, undergoes a
sudden change in pressure to become under a slight vacuum for the remaining
of the period 6. This is an oversimplification which does not take place in
practice, the change in pressure is gradual but is effected in a relatively
short time span.
Leakage out occurs during over-pressurization of the gas by the weight
of the membrane M, which takes place for all the positions of the membrane
above its lowest one: M'. It may be noticed that the bladder remains at
its fully expanded position during the over-pressurization periods. The
bladder is then under the slight pressure exerted by the membrane M.
Leakage in takes place during the under-pressurization periods, which
occurs when the bladder is ina partially contracted position and therefore
under a slight vacuum. The membrane M is at lowest position, with its
weight supported by its attachment to the trough TR. In the present case,
where the pressure differential -dPg created by the bladder is larger, in
absolute value, than the pressure differential +dPj, the gas remains under
a vacuum for a shorter time (tj) than when under pressure (t2). The ratio
t}/t2 depends on the ratio <3P]/<8>2 °* the pressure differential and on the
partial pressure of the gasoline in the gas.
The most important consequence of all of the above is that a storage
tank subjected to a continuous turnover will experience a leakage loss
larger than the one at any other operating mode. This consequence is a
simple reflection of the fact that under continuous cycling, a sizable
fraction of the total operating time is under conditions when the air is
leaking out.
In the majority of the commercial storage applications, the cycling
time (or in other terms, the time expended in filling and emptying the
tanks) is a relatively small fraction of the total time the product is kept
under storage, and consequently the losses due to leakage will be a small
fraction of the total loss. The total loss being the sum of the loss due to
gas permeation through the liners and the loss due to vapor leakage.
In summary it should be mentioned that the losses found in the present
demo-unit would thus represent the upper boundary of the expected losses in
commercial applications.
The start-up operation of the demo-unit will now be described in detail.
-------
The first step consisted of the filling of the tanks with the required
volume of water. A flexible hose was connected between a water faucet and
the valve V1, and with all the block valves closed except B5,, the tank T1
was filled to its upper operating level, which was ascertained by viewing
the upper sight glass tube. The valves V1 and B5 were then closed and the
flexible hose reconnected to valve 72. Valve U was now open and tank T2
filled to its lower operating level, as indicated by its lower sight tube.
During the filling of the tanks, the air contained therein was displaced
through the bladders by keeping the valves V3, VA, V5 and V6 open. The gas-
oline was introduced into the tank by pouring it into the top opening of the
glass apparatus G2, after removal of the vinyl tube, and by opening the ball
valve BV2, care being taken to maintain the liquid level in the tank below
this valve.
In order to minimize the effect of the solubility of gasoline in water
on the determinations of gasoline losses, the water was presaturated with
gasoline by introducing about 8 da? (2 gallons) of the same into the tank T2
and operating the plant for 1 week. By measuring the content of the gasoline
layer at consecutive time intervals, it was ascertained that no or very
negligible amounts of gasoline were transferred into the water and conse-
quently the system attained equilibrium.
Following the presaturation run, 37.85 dm3 (10 gallons) of gasoline
were added to each tank which provides two layers of gasoline about 2.54 cm
(1H) thick each. The plant was then run for about one month, and weekly
gasoline losses were determined. After ascertaining that the losses were
practically identical in both tanks, it was decided that a better utiliza-
tion of the demo-unit could be made by testing different liner materials in
one of the tanks while keeping the other unchanged.
Tank T1 was then fitted with a vertical liner made out of a 0.005 cm
(2 mils) thick film of a fluoropolymer, and the unit was run for another two
months. Then the liner was removed and a new vertical liner, built out of a
laminate of 0.005 (2 mils) thick of the same fluoropolymer and a 0.015 cm
(6 mils) thick polyethylene, was installed. This was run for the remaining
three months of the demonstration phase of the program.
In summary, an epoxy coated polyurethane liner was run for five consecu-
tive months and kept standing during the sixth month, undergoing approximate-
ly 1800 cycles (each cycle is equivalent to one tank turnover), a fluoro-
polymeric liner was run for two months through about 650 cycles, and finally,
a laminated liner was run two months and kept standing during the following
month, which coincided with the sixth month of the demonstration phase of
the project.
10
-------
SECTION IV
METHOD USED FOR VAPOR LOSSES DETERMINATION
*
The vapors of the stored gasoline escape from the space enclosed by the
liners into the surrounding atmosphere by permeating across the flexible
membranes and by leaking through fissures, due to imperfect sealing at the
attachment areas to the floating trough and to the tank. Also, but very
seldom, leakage occurs through pinholes or other imperfections in the liners,
such as defective seals between the adjoining modular panels that constitute
the finished liner.
While the loss due to permeation is a known and predictable quantity
that depends on the particular material used in the liner construction, its
thickness, and the nature and temperature of the stored liquid, the leakage
loss is unpredictable and it will depend on the quality control of the
materials, workmanship and inspection procedures for leakage detection.
Inasmuch as the pressure differentials across the liners are of very small
order, the absolute values for the leakage is usually relatively small.
Leakage due to pressure differentials occur essentially during the filling
and emptying phase of the storage tank operation. Because the time expended
in filling and emptying commercial size storage tanks is usually a small
fraction of the total time the liquid is kept under static storage, the in-
fluence of the leakage loss on the total loss is minimal, and the total loss
is therefore controlled by the permeation loss. The pre-estimated loss
(prior to the initiation of the present project) was on the order of a few
liters (gallons) for a six month plant operation or about 0.4 dnK (0.1
gallon) per week. The determination of such a small loss when dealing with
relatively large tanks, made it impractical to determine it with any pre-
cision by measuring the level differentials in the tank. The precision of
level readings in the sight glasses is about ± 0.16 cm (± 1/J.6") which
represents in a 1.83 m (6') diameter tank + 4.16 dm3 (± 1.1 gallons).
Therefore this method would be unacceptable.
The method used in this project measures the volume decrease between
successive determinations for a small area of the gasoline layer. The ratio
of the small area, to the total area in the tank covered by the layer is known,
and it is relatively simple to translate the measured volume decrease into
the tank layer loss for the period of time between two successive determina-
tions .
Fig. 3 illustrates schematically the apparatus used for the vapor loss
determinations. The glass decanter G, fabricated out of 0.635 cm (l/4M)
thick Pyrex glass, is supported by a clamp on a rod frame attached to the
tank (not shown in the illustration). It has a side tube facing a 2" IPS
steel nipple which is threaded into a bronze gate valve (GV). A polyurethane
sleeve was tightly slid and clamped over the nipple and the side tube, assur-
ing a leakproof connection. A three-way glass petcock P allows the flow of
liquid between the tank and the decanter, in one case; the drainage of the
decanter through the elbowed tubing, in another case; or the flow from the
11
-------
GASOLINE LAYER
Figure 3. Vapor loss determination apparatus,
12
-------
tank and through the elbowed tubing, in the third case. A polyvinyl tube
PV2 provides a flexible connection between the decanter and the tank. The
gas phases of the decanter and the tank are connected by a 2.54 cm ID
polyvinyl tubing PV1.
In order to determine the volume of the gasoline layer held in the tank
at any given time, a fraction of the layer is collected in the decanter and
its volume determined. The knowledge of the ratio of the tank area to the
decanter area permits the calculation of the total Trolume of gasoline con-
tained in the tank's layer.
To obtain the establishment of a layer in the decanter with the same
thickness as the tank's layer, it is necessary that the liquid level selected
be such as to allow the flow of the gasoline from the tank into the decanter
when the gate valve GV is open. This requirement implies that the gasoline
layer should partially or totally fill the circular duct connecting the tank
to the decanter.
The procedure followed for the volume determination consisted of the
following steps:
1. At the start; the valves GV, BV and BV are closed and the decanter
G is empty.
2. A liquid level is set and marked on the glass tube of the sight
gage, in compliance with the requirement that the gasoline layer should
partially or totally fill the connecting duct.
3. The liquid level is now brought to an intermediate level which lays
below the above mentioned set level but remains within the glass tube.
4. The gage!svalve VI,which communicates it with the tank contents, is
now closed and the contents of the glass tube are drained through the
petcock PI. This will remove any slugs of gasoline that entered the tube
while raising the liquid level.
5. PI is closed and VI is opened which allows the glass tube to be
filled only with water.
6. The liquid level in the tank is now brought back to the set level
by manually operating the pump's controls and observing the glass gage. This
will insure that the gasoline-water interface is always at nearly the same
level in all determinations. The liquid in the tank is left standing still
overnight to insure that any gasoline droplets that may have been dispersed
in the liquid phase will collect in the top layer.
7. With valve GV still closed, valves BV and BV are now opened. Water
from the tank fills the decanter until hydrostatic equilibrium is attained.
Valve GV is then opened, allowing the gasoline from the tank's layer to flow
into the decanter. After a period of two hours, which has proven ample for
13
-------
attainment of equilibrium, valves BV and GV are closed. The decanter now
contains a gasoline layer of exactly the same thickness as the one in the
tank.
8. The water layer in the decanter is slowly decanted off into a beaker
by manipulating the three-way petcock. Decantation proceeds until the gas-
oline layer reaches the petcock, care being taken to leave a small water
column between the petcock and the gasoline.
9. The gasoline layer is drained off the decanter into a beaker to-
gether with the small amount of water that remained in the three-way petcock
and the 90° glass tube.
10. The collected gasoline is poured into a burette, where the accom-
panying water that Settles at the bottom is drained off. The remaining gas-
oline is measured and recorded. The burettes used in these determinations
could be read to 0.2 cm3.
The area covered by the gasoline layer collected in the decanter was
calculated by measuring the height of the layer and its volume. It was found
to be 160 cm*. The area covered by the gasoline and occupying the tank's
available internal cross section was found to be 18,600 car. The ratio of the
tank area to the decanter area is therefore 18,600/160 = 116.
There are two systematic errors associated with the above described
volume determinations. One error is caused by the lack of coincidence of the
liquid level with the mark set in the glass tube. The smallest division in
the glass tube is 0.16 cm (1/16") and consequently the liquid level can only
be read within ± 0.16 cm of the set mark. Consequently, the gasoline layer
that collects in the decanter during step 7 of the determination procedure
may be located within 0.16 cm above or below its assumed position. This un-
certainty in the location in turn causes an uncertainty in the volume of the
portion of the layer comprised within the connecting circular duct, because
of the variation of the layer's cross-section with its elevation within the
duct. From the dimensions of the duct and the thickness of the gasoline
layer, it has been calculated that the maximum error to be expected from this
source at eachvolume determination is ±0.65 cm3. The other error is due to
the readings of the volume of gasoline in the measuring burette. The burette
utilized had a capacity of 100 cm3, and the volume collected in the decanters
were of the order of 300 cm3. Three readings were thus required which intro~
duced an error of ± 0.6 cm3 per determination. The total systematic error is
thus equal to or less than ±1.25 cm3 per determination.
The gasoline loss experienced in the tank between two successive deter-
minations is calculated as follows:
First define:
Vj[ = the gasoline volume in the decanter determined at test No. i
-------
— the gasoline volume in the decanter determined at test Mo. i+1
dVj} = the loss of gasoline in the decanter
dVf — the loss of gasoline in the tank
136 = the ratio between the areas covered by the layers in the tank
and in the decanter.
We have then:
dVD = Vj. - Vi+1 ± 2.5 cm3 (4-1)
dVT = 116. dVD - Vi ± 291.25 cm3 (4-2)
The volume collected in the decanter is discarded, since it is not an operat-
ing loss, it has to be substracted from the actual loss.
When the gasoline loss is desired to be found between determinations
i and i + n, equations 4-1 and 4-2 become:
dVD = Vi - Vi+n±2.5 cm3 (4-3)
i+n-1
dVT - 116. dVD - ^Z vi ± (290 + 1.25 (n-1)) (4-O
15
-------
SECTION V
RESULTS OF THE VAPOR LOSS DETERMINATIONS
GENERAL
Table 5-1 contains the data obtained for the six months operation of
tank No. 2. As described in Section III, tank No. 2 was fitted with an
epoxy coated polyurethane liner. This liner was continuously operated for
5 months at an average of 10 hours per day and 5 days per week. It was thus
subjected to about 1800 cycles. During the sixth month, the system was left
standing still to determine the static storage loss.
Table 5~2 contains the data for the presaturation run.
Table 5-3 contains the data obtained for the fluoropolymeric liner that
was fitted in tank 1 and operated for 2 months continuously. Only 2 sets of
data were taken in the last 5 weeks of operation.
Table 5-4 contains the data obtained for the liner built out of a
laminate of a 0.005 cm thick film of a fluoropolymer and a 0,015 cm thick
polyethylene film, which was fitted in the tank 1. The liner was run con-
tinuously for 2 months and kept standing still for the 3rd month.
A plot of the cumulative loss against time for the tank No. 2 is repre-
sented' in Fig. 4. The curve shown was calculated from the data in the table
5-1> by the least squares method. The curve has a sharp change of slope
that occurs at the beginning of the sixth month of operation, coinciding
with the change in the operation regime, i.e. from dynamic to static. Fig. 5
represents a similar plot for the data shown in table 5-4.
DYNAMIC LOSS RATE FOR TANK No. 2
From table 5-1, the total loss of gasoline during the,5 months period
of continuous operation (dynamic storage condition) is calculated as follows:
Volume decrease in the decanter between the determinatiqn No. 20 and No. 0 :
dVD = 347.6 - 251.8 = 95.8 cm3 (5-1)
The sum of the volumes of gasoline collected in the decanter and
discarded from the twenty determinations is:
- 6068.8 cm3 (5-2)
Applying the formula (4-4; gives:
dVT *= 116. dVD -J"1, Vt ± (116 x 2.5 + 1.25 x 19) cm3 (5-3)
16
-------
Table 5-1. VAPOR LOSS DETERMINATIONS AT TANK No. 2
Date,
1978
6/29
7/7
7/14
7/21
7/28
8/4
8/11
8/18
8/25
9/1
9/8
9/15
9/22
9/29
10/6
10/13
10/20
10/27
11/3
11/17
12/1
12/15
12/29
Determi-
nation
number,
i
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
IS
19
20
21
22
Volume of
gasoline
collected
in the
decanter,
cm3
347.6
341.0
338.4
333.6
327.0
324.2
318.8
312.2
310.4
305.2
299.6
298.4
295.6
288.8
284.2
278.8
272.4
270.4
265.2
257.0
251.8
250.0
246.4
dVD
Loss at
decanter,
cm3
6.6
2.6
4.8
6.6
2.8
5.4
6.6
1.8
5.2
5.6
1.2
2.8
6.8
4.6
5.4
6.4
2.0
5.2
8.2
5.2
1.8
3.6
dVT
Loss
at tank,
cm3
417
-40
219
431
-2
308
447
-103
293
344
-160
26
493
245
342
464
-41
333
686
346
-43
168
Cumulative
loss at tank,
cm3
417
377
596
1027
1025
1328
1775
1672
1965
2309
2149
2175
2668
2913
3255
3719
3678
4011
4697
5043
5000
5168
17
-------
Table 5-2. PRESATURATION RUN
Date,
1978 .
6/22
6/26
6/28
Determi-
nation
number,
i
1
2
3
Vi
Volume of
gasoline
collected
in the
decanter,
ca3
64.8
51.4
50.8
dVD
Loss at
decanter,
cm3
13.4
0.6
Gasoline
transferred
to water,
cm3
'
1589.6
18.2
Table 5-3. VAPOR LOSS DETERMINATIONS AT TANK No.l
Fluoropolymeric liner
Date,
• 1978
8/25
9/29
Determi-
nation
number,
i
0
1
Volume of
gasoline
collected
in the
decanter,
cm3
273.2
258.4
-
dVp
Loss at
decanter,
cm3
14.8
dVr
Loss
at tank,
cm3
1443
Cumulative
loss at tank,
cm3
1443
18
-------
Table 5-4. VAPOR LOSS DETERMINATIONS AT TANK No. 1
Laminated Liner
Date,
1978
10/6
11/3
12/1
12/15
12/29
Determi-
nation "
number,
i
0
1
2
3 '
4
Volume of
gasoline
collected
in the
decanter,
cm3
306.6
295.8
287.4
284.8
282.2
dVD
Loss at
decanter ,
cm3
10.8
8.4
2.6
2.6
dVf
Loss
at tank,
cm3
946
678
14
17
Cumulative
loss at tank,
cm3
946
1624
1638
1655
dVT = 116 x 95.8 - 6068.8 ± 314 - 5044 ± 314 cm3
The relative error (er) attached to dVf is:
314 x 100
5044
6.3*
(5-4)
(5-5)
(5-6)
The yearly loss rate (Ly) is then:
L- = dV? x 12 mo. = 12106 cm3 (3.2 gallons)
^ 5 mo. year year
with a precision of ± 6.3$
STATIC LOSS RATE'FOR TANK No. 2
The loss experienced in tank No. 2 at standstill (static storage condi-
tion) is calculated from determination Nos. 20, 21, and 22. Using the proce-
dure just outlined, we obtain:
dVD = 251.8 - 246.4 * 5.4 cm?
(5-7)
19
-------
STATIC OPERATION
LOSS RATE sS0.10£ dm3
week
DYNAMIC OPERATION
LOSS RATEi 0.22 ±0.015 dffl3/week
8
30 32 ait
TIME, weeka
Figure 4. Loss rate for tank 2.
-------
V. = 251.8 + 250 = 501.8 cm3 (5-S)
i=20
dVT = 116 x 5.4 - 501,8 i (116 x 2.5 < 1.25 x 2) em? (5-9)
dVT = 124.6 ± 295 cm3 (5-10)
is thus comprised between -170 cm3 and +420 cm3. The negative number
would indicate a gain in the system, which is absurd and therefore discarded.
The positive number indicates that the loss is between zero and 420 era3.
The yearly loss rate is then equal or less than:
Ly^ 420 x 12 mo. - 5040 cm3 (1.33 gal ) (5-11)
year year
If it is assumed for the moment that this loss rate is due to gas
permeation through the liners, which have a total surface of 12.54 m^
(135 ft2), then the permeation coefficient (Cp) should be equal or less than:
5040 = 1.1 cm3 or 0.77 g (5-12)
365 x 18.54 day.m^ day.m^
Permeation determinations carried out by the pouch method (i.e., in pouches
fabricated with the same material as the liner) indicated a coefficient of
0.53 g/day.m2.
It may then be concluded than during static storage conditions the loss
rate is essentially due to the gas permeation through the liners. The larger
loss rate found for dynamic storage conditions is attributed to leakage in
addition to permeation as explained in Section III.
SCALE-UP OF LOSS RATES TO COMMERCIAL SIZE TANKS
In order to calculate the emission control efficiencies obtainable with
the present technology when applied to commercial size tanks, a tank measur-
ing 100 ft. in diameter and 40 ft. in height is considered. The loss rate
for a tank of any size is here assumed to be proportional to the area of the
liners fitted into the tank. This area is the sum of the cross-sectional and
lateral areas of thf tank. For the size mentioned above, the total lined
area measures 1897 *&• (20,420 ft^), The ratio of the areas of the commercial
tank to the demo-unit tank is 1897/12.54 = 152. The yearly loss (LDy) for
the commercial tank operating under dynamic storage condition would then be:
LDy = 12.106 dm3/year x 152 = 1840 dmVyear (11 Bbls/year) (5-13)
The static storage yearly loss is equal or less than:
5.04 dm3/year x 152 = 766 dm3/year (4.8 Bbls/year) (5-14)
21
-------
In summary, 100" $ x 40' H storage tank holding gasoline and fitted with
an epoxy coated polyurethane liner is expected not to emit more than 11
3bls/year under dynamic storage and 5 Bbls/year under static storage condi-
tions.
DYNAMIC LOSS RATE FOR TANK No. 1
The loss rates experienced in tank No. 1 when fitted with the laminated
liner are calculated from the data in table 5-4. For the loss during dynamic
storage condition we have: (determinations 0, 1 and 2).
dVD = 306.6 - 287.4 = 19.2 cm3 (5-15)
1
306.6 + 295.8 = 602.4 cm3 (5-16)
dVT = 19,2 x 116 - 602.4 ± (116 x 2.5 + 1.25 x 2) cm3 (5-17)
dVT « 1625 ± 295 cm3 (5-18)
The yea.-ly loss rate is:
LDy « 1625 x 52 weeks » 10562 cm3 or 2.8 gal (5-19)
8 weeks year year
with a precision of + 18$.
STATIC LOSS RATE FOR TANK No. 1
For the loss rate under static storage conditions we have from
determinations 2, 3 and 4:
dVD = 287.4 - 282.2 = 5.2 cm3 , (5-20)
3 T
287.4 + 284.8 = 572.2 cnH (5-21)
dVT - 5.2 x 116 - 572.2 ± 295 cm3 = 31 ± 295 Cm3 (5-22)
The loss will thus be between -264 cm3 and + 326 cm3. The negative
number indicates a gain in the system, which is absurd and is therefore
discarded. The loss will thus be between zero and + 326 cm.3. The maximum
yearly rate loss to be expected is then:
326 x 12 mo. = 3912 cm3 (l gal ) (5-23)
year year
22
-------
-2
x
<
a:
H
STATIC OPERATION
LOSS RATE^0.08 dm3/we«k
S3
M
03
CO
O
•DYNAMIC OPERATION
^^
LOSS RATE : 0.197±0.035 dm3/wcek
|
o
t
12
TIME,weeks
Figure 5. Loss rate for tank 1.(Laminated liner)
23
-------
The maximum gas permeation coefficient to be expected would be:
cp = 3912 = 0.85 cm3 or 0.6 g
365 x 12.54 day.m2 day.m2
Gas permeation tests conducted with a material identical to the one of the
liner, indicated a zero permeation loss. The pouch method was also used
here. It can be concluded that the loss rate for this type of liner is
essentially nil under static storage conditions, within experimental error.
The sharp change in the rates when passing from dynamic to static
storage conditions, clearly indicate that the mechanism of loss is quite
different in each regime. As mentioned previously, the higher rate obtained
for dynamic conditions is due to additional losses by leakage through im-
perfections in the seals or other portions of the liners. This leakage
is caused by pressure differentials created across the liners, as explained
in detail in Section III of this report.
-------
SECTION VI
ACCURACY OF THE DATA AND THE RESULTS
It was shown in Section IV that each data point (i.e., each determin-
ation of the volume of gasoline collected in the decanter) was affected with
a possible error of + 1.25 cm-?. It was also shown that the probable error
associated with the calculation of the gasoline loss in the tank between any
two determinations was:
eT ^ (116 x 2.5 + 1.25 (n-1)) cm3 (6-1)
where:
e-p is the error attached to the calculated value
of the gasoline loss between the determinations
numbered i and i + n
The error committed increases very slowly with the number n, while the
calculated loss in the tank increases very nearly proportionally with the
number n. Consequently, the relative error of the calculated value dV-j-
(volume lost between the two determinations) decreases very rapidly with n.
The relative error er is defined as:
e_ « fT_ (6-2)
dVT
where: er is the relative error
e«p is the absolute error
dV-p is the loss of gasoline in the tank between determinations. The
error eT is of the order of 300 cm?.
In order to obtain a relative error er less than 10$, dV.p should be
larger than 3000 cm3. This requirement was fulfilled in tank No. 2 where
the total calculated loss was close to 5000 cm3 for the dynamic storage
operation. The accuracy of the loss found for the tank No. 1 when fitted
with the 2 mils thick fluropolymeric liner and operating under dynamic
storage conditions is :
dVT = 1443 cnp/month or 4.6 gal/year (6-3)
eT^ 292 cm3 (6-4)
(6-5)
This low precision is due to the short time span between determina-
tions and to the low loss rate. For the tank No. 1 when fitted with the
laminated liner and for dynamic storage operation the accuracy was found
to be ± 18#. (See Section V, formula 5-18).
25
-------
SECTION VII
INTERPRETATION OF THE RESULTS OBTAINED
The results obtained for the emission rates of gasoline in the demo-
unit can be compared with emission rates from commercial size uncontrolled
storage tanks. The correlations and formulas established by the American
Petroleum Institute have been determined from data obtained on commercial
size tank operations. In order to establish a comparison on the same basis,
the results from the demo-unit will be scaled-up to commercial size tanks.
The scale-up factor is assumed to be proportional to the area of the liners.
The justification for this assumption rests on the fact that vapor perme-
ation losses are proportional to the area of the permeating membrane.
Although it would seem more justifiable to use a ratio of the lengths of the
sealing surfaces of the liners for the leakage losses, the use of the areas
ratio will provide a more conservative approach. This latter ratio is the
square of the former one.
¥e will now consider the scaled-up results obtained from the 3 different
types of liners utilized in the demo-unit and compare them with the data
obtained from the API correlations for an uncontrolled commercial sized tank.
A commercial tank measuring 30.48 m (1001) in diameter by 12.2 m (40*) in the
height /ill be considered for comparison purposes. The scale-up factor be-
tween this size tank and the demo-unit tank size is 152.
EXPECTED CONTROL EFFICIENCIES
Folyurethane-Epoxy Coated Liner In Tank No. 2
Dynamic Storage Condition-
The dynamic rate loss for the demo-unit size found at (5-6) was
12.1 dm3/year and the scale-up dynamic loss rate (l.Dy) is:
LDy ~ 17.1 x 152 = 1840 dm3/year or 11 Bbla
year
Number of cycles undergone by the liner: 1800
The loss rate for the uncontrolled commercial size tank is found from
Fig. 12 in API Bulletin 2518*. For a gasoline having a vapor pressure of
3516 Pa (5 psia) the loss is: 1,589,700 dm3/year (10,000 Bbls/year). The
emission control efficiency (EFF) thus provided by the Stop-Los Company new
technology is then:
EFF = 100 x (1 - 11/10,000) - 99.91 (7-1)
* American Petroleum Institute, Evaporation loss from fixed-roof
tanks, API bull-tin 2518, June 1962,
26
-------
The maximum number of cycles or turnovers listed in this nomograph is
400/year. " h '
Static Storage Condition-
The scaled-up static loss rate (L3V) for the demo-ur.it was found at
(5-14) to be: y
LSy^T 766 dm3 /year (4.8 3bl/year)
The equivalent of static storage is found in the API Bulletin 2518
under "Breathing Loss". Breathing losses are a function of the average
daily temperature change, the tank outage, the vapor pressure of the gaso-
line and the tank diameter, and a paint factor, which is assumed, conser-
vatively, to be 1, The average daily temperature change will be taken for
the same geographical area where the demo-unit is located. The choice of
the tank outage is somehow arbitrary, a tank about 2/3 full will be here
considered. The average daily temperature change in the Newark, N.J. area
is 8.33 K (15°F), the tank outage: 4.2? m (14* ), the gasoline vapor pressure:
3516 Fa (5 psia). From Fig. 8 of API Bulletin 2518:
Breathing loss = 95382 dm3/year (600 Bbls./year)
The emission control efficiency provided by the Stop-Los Company technology
is then:
(1 - 766/95382)= 99.2$ (7-2)
Polyethylene-Fluoropolymer Laminate Liner In Tank No. 1
Dynamic Storage Condition-
The value found at (5-19) for the demo-unit was:
LDy = 10.562 dm3 /year
3caling-up this value we obtain:
LDy = 10.562 x 152 = 1605 dmVyear or 10 3bl/year
Number of cycles undergone by the liner: 700
The loss rate for the uncontrolled commercial size tank is the same
as found above, or:
Loss = 1,589,700 dm3 /year (10,000 3bl/year)
The emission c ontrol efficiency is then:
EFF= 100 (1 - 10/10,000) = 99. 9#
27
-------
Polyethylene-Fluoropolymer Laminate Liner in Tank No..1
Static Storage Condition-
It was shown before that the loss rate was essentially nil and
consequently it should provide a control emission efficiency of nearly
10056.
28
-------
SECTION VIII
LIFE EXPECTANCY OF LINERS
It is known that the liners' materials of construction are essentially
inert to attack by the stored gasoline and similar hydrocarbons such as
crude oil, kerosene, jet fuel, etc., when they are in a stress free state.
The effect of the application of repetitive alternating stresses, such as the
ones developed during the folding and unfolding of the vertical liners are
also known for the cases where the environment surrounding the liners is at-
mospheric air. Folding endurance tests which are routinely performed and
reported* indicate that the materials will withstand a large number of al-
ternating stresses before failing. For instance, polyurethane has virtually
infinite folding endurance; polyethylene can endure about 100,000 cycles be-
fore failing and the fluoropolymer utilized in the laminated liner will with-
stand about 250,000 cycles. But, the endurance of these materials when
cyclicly stresses in an atmosphere containing hydrocarbon vapors was not
known with certainty prior to the initiation of the present program.
The constancy of the loss rate throughout time, as shown by the slopes
of the curves plotted in Figures 4 and 5> would indicate that no cracks or
other types of mechanical failures were sustained by the liners. The inspec-
tion of the liners during and after the completion of the test periods tends
to corroborate this assumption.
In a survey of oil companies conducted by the API (API Bulletin 2518,
page 19), it was found that the great majority of storage tanks had a yearly
turnover of 10 or less. This implies that on the average, a storage tank is
filled and emptied once every 5 weeks. Yearly turnovers of the order of 30
to 40 were infrequently found. In this report a weekly turnover was assumed
for a commercial type operation.
The liners used in the demo-unit underwent between 700 to 1800 cycles,
each cycle corresponding to a tank turnover. The minimum life expectancy
would then be in the order of 14 years for the laminated liner, and about 34
years for the polyurethane liner. It may then be concluded that under normal
operating conditions the liners should render satisfactory service for a 20
years period, which would be an acceptable standard to the industry.
* McGraw Hill, Modern Plastics Encyclopedia, 1974-75*
Pages 730, 732
29
-------
SECTION IX
ECONOMICS
An economic comparison between the new technology dealt with in this
report and the conventional technology for emissions control of stored
petroleum and its derivatives, clearly shows the new technology to be highly
competitive.
The yardstick used in this comparison is the installed cost of the
devices offered by both technologies. Installed costs for the new technology
were extracted from recent quotations made by the Stop-Los Company to pros-
pective customers. The installed costs for the conventional technology were
estimated from the figure 8-5 of the Petroleum Processing Handbook*.
Because the data in this reference are for the years 1964-1965, an
escalation factor of 2 was used to arrive at present day costs. The escala-
tion factor was taken from the Business Week Magazine's Index **: Consumer
Price Index/3LS. The datum (100) of this Index is for the year 1967. For
January 1979 the Index stood at 204.7.
The installed cost for the conventional devices, i.e., double deck and
pontoon types of floating roofs, as well as for the devices of the new
technology, are plotted against tank capacities in figure 6.
* William F. Bland, Robert L. Davidson. Petroleum Processing
Handbook, KcGraw Kill Book Co., 1967> page 8-24.
*•* Business rfeek, McGraw Hill Inc., March 12, 1979, page 2.
30
-------
-70
2
a
8
o
-60
-50
-45
-40
-35
-30
-25
-20
-19
-18
20
DOUBLE DECK FLOATING ROOFS
PONTOON FLOATING ROOFS
STOP-LOS VAPOR SEALS
30
y
60
TANK CAPACITY, 1000 barrels
Figure 6. New and old technologies cost comparison.
31
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
I. REPORT NO.
EPA-600/2-79-108
3. RECIPIENT'S ACCESSION-NO.
TITtE AND SUBTITLE
Reduction of Air Emissions from Gasoline Storage
Tanks
5. REPORT DATE
1979
May
6. PERFORMING ORGANIZATION CODE
. AUTHOR(S)
Arnold Gunther
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Stop-Los Company
29 Lorelei Road
West Orange, New Jersey 07052
10. PROGRAM ELEMENT NO.
1AB604
11. CONTRACT/GRANT NO.
J-02-2679
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final: 1/78 - 1/79
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES IERL-RTP project officer is Bruce Tichenor, Mail Drop 62,
919/541-2547.
16. ABSTRACT
The report gives results of a project to determine the technical and econ-
omic feasibility of using flexible plastic membranes to control emissions from gaso-
line storage tanks. The emission rates and the expected life of the membranes were
to be established. A demonstration pilot unit was built and operated. The emission
rates were determined, as well as the life expectancy of the membranes. The results
indicate that emission control of 99-plus percent, compared to uncontrolled tanks',
can be achieved readily, and that the life expectancy of the membranes is in the or-
der of 20 years of continuous service. The estimated installed cost of these devices
for commercial application is very competitive with conventionally used floating
roofs.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Gasoline
Storage Tanks
Emission
Membranes
Plastics
Air Pollution Control
Stationary Sources
Flexible Plastic Mem-
branes
13B
21D
13D
14B
11G
111
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
36
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
32
------- |