United States
Environmental Protection
risl Erwironmsrta!
tory
Cincinnati OH 45268
EPA-600 7-78-2T7
November 1978
Development of
Sorbent
Distribution and
Recovery Sys
a
ov
i
Energy/Environ
R&D Program
Report
ent
-------
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 interlace 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 INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service. Springfield. Virginia 22161.
-------
EPA-600/7-78-217
November 1978
DEVELOPMENT OF A SORBENT
DISTRIBUTION AND RECOVERY SYSTEM
by
Sidney H. Shaw, Richard P. Bishop, and Robert J. Powers
Seaward International, Inc.
Falls Church, Virginia 22044
Contract No. 68-03-2138
Project Officer
J. S. Dorrler
Oil and Hazardous Materials Spills Branch
Industrial Environmental Research Laboratory
Edison, New Jersey 08817
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental
Research Laboratory, U. S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the con-
tents necessarily reflect the views and policies of the U. S.
Environmental Protection Agency, nor does mention of trade names
or commercial products constitute endorsement or recommendation
for use.
11
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and increasingly more efficient pollution con-
trol methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (IERL-Ci) assists in developing and demonstrating new and
improved methodologies that will meet these needs both efficiently and
economically.
This report describes the development and testing of a versatile
system for collecting spilled oil from the water's surface. The system
described makes use of particles of reticulated polyurethane foam which are
both olephilic and hydrophobic. Because of these properties, the particles
are able to preferentially absorb the oil from the slick. In addition to
the sorbent foam, the system includes equipment for storing and distributing
the foam, and for recovering the oil from the foam. The results of this
report will be useful both to those responsible for the operation of the
prototype system and to others working in the field of waterborne oil spill
recovery. Further information may be obtained through the Resource
Extraction and Handling Division, Oil & Hazardous Materials Spills Branch
in Edison, New Jersey.
David G. Stephan
Di rector
Industrial Environmental Research Laboratory
Cincinnati
m
-------
ABSTRACT
This report describes the design, fabrication, and test of a
prototype system for the recovery of spilled oil from the surface
of river, estuarine, and harbor waters. The system utilizes an
open cell polyurethane foam in small cubes to absorb the floating
oil. The system is highly mobile and can be transported in two
pickup trucks.
The sorbent is transported and distributed over the water
surface by means of a pneumatic broadcaster. An inclined, open
wire mesh belt conveyor is used to remove the saturated sorbent
from the water. The recovered oil and water is removed from the
sorbent by squeezing in a converging belt press or regenerator.
After regeneration, the foam is reapplied to the oil slick. The
foam can be reused for a great many cycles.
Tests of the system, using both diesel fuel and lubricating
oil, were conducted at EPA's OHMSETT test tank. The sweep speeds
ranged up to 5 knots in both calm water and waves. Oil collection
rates of 10.5 m3/h were achieved. The oil content of the re-
covered liquid varied from 38% to 79%.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables vii
Metric Conversion Factors viii
Abbreviation and Symbols ix
Acknowledgement x
1. Introduction 1
2. Conclusions and Recommendations 2
3. Modes of Operation
Batch processing 4
Continuous processing 6
4. Development Program
Background 8
Component selection and design 9
Reuseable foam sorbent 9
Sorbent broadcasting unit 11
Harvesting conveyor 16
Sorbent regenerator unit 21
5. System Testing at OHMSETT
Test program 34
Test results 41
References 52
Appendix 53
-------
FIGURES
Number
3.1 Sorbent oil recovery system deployed at a stream. . . 5
4.1 Regenerated sorbent density vs. squeezing force ... 10
4.2 Discharge nozzle of sorbent braodcaster 12
4.3 View of sorbent broadcaster's blower 13
4.4 Overall view of sorbent braodcaster 14
4.5 Open material handling wheel of broadcaster 14
4.6 Harvesting conveyor at stream bank 18
4.7 Harvesting conveyor on catamaran 18
4.8 Harvesting conveyor's power supply 22
4.9 Overall view of sorbent regenerator 23
4.10 Detail of sorbent regenerator 25
4.11 Sorbent regenerator with chain guard removed 26
4.12 Sorbent cubes tumbling down regenerator's
discharge chute 27
4.13 Sorbent regenerator's power supply 28
4.14 Post-test modifications to the regenerator:
plumbing to combine the two discharge ports 31
4.15 Sorbent regenerator operating in the back of
a pickup 32
5.1 OHMSETT test tank arrangement 36
5.2 Nonlinear regression of volumetric efficiency . . . .48
5.3 Nonlinear regression of oil fraction in
recovered liquid 50
5.4 Nonlinear regression of throughput efficiency .... 51
A-l 13%" Dayton blower characteristic at standard
conditions (0.075 lb/ft3) 56
A-2 Blower characteristic curves 59
A-3 Velocity of single particles (unloaded) 60
A-4 Particle Velocities (loaded, full throttle) 61
A-5 Pressure losses in the system as a function of
the sorbent loading coefficient, f 64
VI
-------
TABLES
Number
Page
4.1 Characteristics of the Sorbent Broadcaster 15
4.2 Characteristics of the Harvesting Conveyor 19
4.3 Characteristics of the Sorbent Regenerator 33
5.1 Test Conditions and Results 42
VII
-------
METRIC CONVERSION FACTORS
H-
H-
Symkal
m
ft
yd
mi
in'
II1
w
Ib
up
Trap
II 01
c
P<
41
gal
yd'
Approximate
Waia Vw Kaew
inche*
leet
yard*
mllaa
Convitiioni to Mitric
M.,,,,.,k,
LENGTH
•2.6
10
0.9
1.6
AREA
•Ojuara inche* 6.6
equate feel O.M
aquara yard* O.I
square mile* 2.6
acres
ounce*
pound*
shot Ions
12000 Ib)
teaspoons
lableapoon*
lluid ounce*
cup*
pint*
quart*
gallon*
cubic feat
cubic yarda
0.4
MASS (wiight)
28
0.45
0.9
VOLUME
6
16
10
OJ4
0.47
0.96
1.1
0.01
0.76
M6l8urii
T* Find
centimeter*
cenlimaler*
malar*
kilomaiara
square centimeter*
square meter*
squire maters
square kilometers
hectares
gram*
kilogram*
tonne*
milliliier*
millllitara
milliliters
lilur*
liters
liter*
liter*
cubic meter*'
cubic meter*
Symbal
cm
cm
m
km
m'
m2
km>
ha
g
kg
i
ml
ml
ml
1
1
1
|
m1
m'
TEMPERATURE (•wet)
•»
FsHrentujil
temperature
Umtfc ol tteitfht* and UvBMVat. Piiee t?
S/9 (altar
subtract ing
121
M. bO OUluv No. CU.10 r»6
Celsius
temperature
•c
-------
LIST OF ABBREVIATION AND SYMBOLS
ABBREVIATION
r —correlation coefficient
SYMBOLS
E —throughput efficiency
F —oil fraction
R —sorbent broadcast rate
T —oil slick thickness
U —viscosity
V —velocity
VE —volumetric efficiency
IX
-------
ACKNOWLEDGMENT
This report summarizes a development effort carried out by
Seaward International, Inc. Sidney H. Shaw was project director.
Working with him were Robert L. Beach, Richard P. Bishop, Louis
S. Brown, David W. Durfee, Michael Krenitsky, Frank A. March,
and Robert J. Powers.
Hydronautics, Incorporated, developed the sorbent broad-
caster under the direction of William T. Lindemuth.
Garth D. Gumtz of Morris Knowles, Inc., designed the sor-
bent regenerator.
The Office of Research and Development, U. S. Environmental
Protection Agency, supported this project. J. Stephen Dorrler
was project officer.
-------
SECTION 1
INTRODUCTION
Spilled oil and other liquid contaminants with specific
gravities less than 1.0 tend to float on water. Numerous devices
have been developed for collecting spilled oil. Most use a skim-
ming device to pick up a thin slice of oil from the top of the
water column. The performance of these devices tends to be ser-
iously degraded by waves and currents. Collection of floating
oil by means of a sorbent material, such as straw, which is then
mechanically removed from the water is another frequently used
technique. The sorbent approach is less affected by waves and
current, but has the disadvantages of producing large amounts of
oily sorbent, which is difficult to dispose of, and of presenting
logistical problems in supplying the volume of raw sorbent.
In an effort to foster the development of new techniques for
oil recovery, the U. S. Environmental Protection Agency (EPA)
funded the development of an oil recovery system employing reuse-
able sorbent material. Reuseable sorbent retains the advantages
of other sorbent techniques, being less sensitive to wave and
current action, and also significantly reduces the sorbent supply
and disposal problems. Further, the oil can be recovered in a
useful form.
The objectives of this program were to design, build, and
test a reuseable sorbent system for recovering spilled oil from
the surface of river, estuarine, and harbor waters. The system
was based on the results of previous EPA contracts. The system
was to be easily transportable in standard pickup trucks. Each
system unit was to be individually powered and operated either as
part of a continuous process or independently. Each unit, other
than the sorbent regenerator, was to be manually transportable
over rough terrain. The complete system was to be operable from
vessels of opportunity or from a combination of one or more small
boats and a dock or shore.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
Of necessity, the OHMSETT testing of the prototype sorbent
oil recovery system was limited in its extent. Tests were
limited to a small number of speeds, wave conditions, and slick
thicknesses and, because of physical limitations, testing was
not done in a continuous mode. Still, the system demonstrated
that the concept is sound and that the prototype equipment is a
useful tool in the cleanup of spilled oils and selected floating
hazardous substances, at a minimum, for the range of conditions
under which it was tested. These conditions were; relative
velocities of up to 5 knots; waves up to 0.3 in height; and
slick thicknesses from 0.24 mm to 0.84 mm.
Under the best conditions the system was able to collect
virtually all of the oil encountered. Better than 40 percent
of the oil encountered was collected even under the most adverse
conditions of high speed and waves.
In the best test case, oil was collected at a rate of 10.5
m3/hr, from a 0.5 mm slick at a sweep speed of 2.5 knots with a
sorbent distribution rate of 17.4 m3/hr.
The recovery system is highly, mobile, being transportable in
only two pickup trucks. With the exception of the foam regenera-
tor, all system elements can be transported manually over rough
terrain.
Sorbent losses due to wind effects with wind velocities up
to 15 knots are negligible. More significant foam losses occur
due to waves washing over the containment boom. During the test
program, the maximum losses occurred at 5-knot sweep speeds in
0.3-m breaking waves. These losses were less than 2%.
Regenerated foam densities were significantly higher than
were predicted from prior laboratory tests. The design regen-
erated foam density was 140 kg/m3 using a roller squeeze force
of 7 N/mm. During testing, the regenerated foam density aver-
aged 160 kg/m3 using an average squeeze force of 16 N/mm. This
probably resulted from oil clinging to the outside of the foam
during squeezing and then being reabsorbed into foam when the
-------
squeeze force was removed. During testing, the oil concentration
varied from 38% to 79%. The recovered liquid was not highly
emulsified and gravity separation of the oil and water was rapid.
Complete removal of all oil, or "sheen polishing," is not
possible using recycled foam in this system because of oil
remaining in the foam cubes after regeneration.
System throughput efficiency could be improved by operating
with a sorbent logjam in front of the harvester. This would
increase contact time which, as the test results indicate, would
increase throughput efficiency. Operating with a logjam was not
done during the test program, in order to reduce the number of
test variables.
It is recommended that the system be operated on a variety
of actual oil spills to gain additional performance information
under more realistic conditions.
It is recommended that the system be operated in a contin-
uous mode in order to prove its capabilities in this mode since
physical limitations prevented testing in this mode at OHMSETT.
It is recommended that in order to improve the tracking of
the regenerator upper squeeze belt and to eliminate the slippage
between the belt and the drive roller, that standard drive roller
lagging be applied in narrow strips with oil resistant adhesive
and should be applied in a helical or chevron pattern from either
end of the roller.
-------
SECTION 3
MODES OF OPERATION
The two modes of operation of the sorbent oil recovery
system are batch processing and continuous processing. The major
difference between the two modes is that transfer conveyors are
used in continuous processing. These conveyors carry the sorbent
from harvester to regenerator, and from regenerator to broad-
caster.
The following two scenarios illustrate these two modes of
operation. The first scenario is for batch processing as con-
ducted on the bank of a stream where a spill has occurred. In
the second, the system has been mounted on a barge to recover
an oil spill from harbor waters.
BATCH PROCESSING
An accidental spill has occurred in a small stream. The
cleanup crew of 6 or 8 transfers the entire sorbent oil recovery
system onto two available pickup trucks. The regenerator nearly
fills the back of one pickup, while the second transports the
sorbent broadcaster, harvester, and power supplies.
The trucks are driven to within 15 m of the stream. The
cleanup crew forms three teams, or crews, one to set up each
major component (Figure 3.1).
One crew consisting of one or two people wheels the sorbent
broadcaster over the rough terrain to the bank of the stream.
The flexible duct is attached directly to the discharge of the
blower, and the nozzle is clamped onto the other end of the duct.
Once the fuel tank is filled, the broadcaster is ready for
operation.
The other people of the cleanup crew, meanwhile, ready the
sorbent harvester. It is offloaded from the truck and wheeled
to the bank, downstream of the broadcaster. Assembly is rapid.
The diverter plates are bolted onto the frame, the floats are
attached, and the lower end is slipped into the water. The upper
end of the harvester is supported on the uneven ground by the
telescoping legs. These are adjusted until the conveyor reaches
-------
REGENEKHTOK.
INPUT BOX
Figure 3.1. Sorbent oil recovery system deployed at a stream.
-------
an angle of approximately 0.52 rad from horizontal. Another crew
has brought the diverter boom and attaches it to the harvester's
downstream diverter plate. The other end of the diverter boom
is anchored upstream to herd the sorbent into the mouth of the
harvester. The power supply is wheeled close to the harvester
and is positioned out of the way. The hydraulic hoses are
attached. The sorbent harvester is also ready for operation.
The third team which consists of one or two people readies
the regenerator. They drive the pickup truck to fairly level
ground near the sorbent harvester. The discharge chute is
attached. The discharge hose is connected to a storage tank.
The team wheels the power supply near the regenerator and con-
nects the hydraulic hoses. Now the entire system is ready to
recover the spill.
Sorbent cubes are broadcast into the spill. The cubes
absorb oil and are guided to the harvester by the diverter boom.
The cubes exit the harvester into a bin placed beneath the screw
conveyor. When the bin is nearly full, one of the cleanup crew
replaces it with an empty one and carries the saturated sorbent
to the regenerator.
The operator feeds the cubes into the regenerator. Nearly
free of oil, they tumble down the discharge chute into another
bin. This bin is carried to the stream bank where they will
once again be applied to the spill.
The members of the team not operating the equipment are
used to carry bins of sorbent cubes between the three major com-
ponents, providing an almost steady flow of sorbent around the
system.
CONTINUOUS PROCESSING
There has been an oil spill in a harbor, and the sorbent
oil recovery system will be used. A barge is available as the
working platform.
The system is set up in much the same manner as for batch
processing, except that transfer conveyors (not part of the
system described in this report) replace the bins and the people
to carry them. A screw conveyor carries the saturated sorbent
from the harvester to the regenerator. This conveyor, a commer-
cially available item, bolts onto the end of the harvester's screw
conveyor. A belt-type conveyor, also commercially available,
carries the oil-free cubes to the broadcaster. The use of con-
veyors to transport the sorbent allows operation of the system by
a crew of 3 or 4 operators. The broadcaster's nozzle is mounted
near the bow of the barge. The harvester is attached to the side
of the hull. A diverter boom on one side, and the hull on the
other, funnel the saturated sorbent into the harvester. The cubes
-------
are lifted by the harvester belt and dropped into the screw con-
veyor, which takes them to the regenerator to be squeezed free
of oil. To complete the cycle, the belt conveyor carries the
squeezed sorbent back to the broadcaster.
-------
SECTION 4
DEVELOPMENT PROGRAM
BACKGROUND
In June 1971, the U. S. Environmental Protection Agency
(EPA) awarded five research contracts to evaluate the feasibility
of the recovery of floating oil using reuseable foam sorbent
material. The preliminary design of equipment to distribute, to
harvest, and to remove the oil from the foam was included in the
scope of these contracts. The complete results of these con-
tracts are contained in references 1 through 5. In general,
these reports concluded that such an oil recovery system was
feasible, and they recommended the construction of a prototype
system.
On October 1, 1974, the EPA awarded a contract to Seaward
International, Inc., of Falls Church, Virginia, for the design,
construction, and test of a sorbent oil recovery system. The
system was to consist of the following component elements:
1. Reuseable foam sorbent
2. Sorbent broadcasting unit
3. Sorbent Harvesting unit
4. Oil-sorbent separator unit
DESIGN FEATURES
The sorbent distribution and recovery system allows rapid
recovery of spilled oil with sorbent cubes that can be reused
over 100 times, according to the manufacturer. As a result of
separating the oil from the sorbent, disposal problems are kept
to a minimum. During the OHMSETT test program, the same sorbent
was used repeatedly with both oils and other hazardous materials
during three weeks of testing. No perceptible change in the
sorbent was noted during this period.
-------
Due to the design of the system, moderate currents and
waves have little influence on collection efficiency. This fact
was substantiated during the test at OHMSETT.
Each system component has debris-handling features. On the
harvesting conveyor, a cross brace is provided to block large
debris. If small debris should jam the screw conveyor, it can
be reversed to free the jam. The sorbent regenerator's rollers
can be raised to pass debris. The broadcaster's duct easily
disassembles for cleaning blockages. No testing, however, was
done with debris present.
The system is highly mobile and maneuverable. The total
system is capable of being transported entirely in two pickup
trucks. All units other than the regenerator can be transported
manually over rough terrain. The regenerator can be operated
while in the bed of a pickup truck. Furthermore, each system
component can be operated independently of the other components,
since each is powered by a separate power supply. These cap-
abilities were demonstrated during simulated field tests without
oil.
COMPONENT SELECTION AND DESIGN
Reuseable Foam Sorbent
The earlier work conducted under the auspices of the EPA
and described in References 2 and 4, optimized the sorbent mate-
rial and configuration. It was concluded that the sorbent should
be a high tensile (340 to 1400 kPa), industrial, open cell poly-
urethane foam with a density of 16 to 32 kg/m3. For use with oil
viscosity up to 1 Pa-s, a pore size of 3 pores/mm was recom-
mended. A pore size of 1 pore/mm was recommended for oils with
viscosity above 1 Pa-s. The optimum sorbent configuration was
concluded to be cubes of 19 run on an edge. The previous test-
ing showed that this foam was capable of absorbing oil up to 30
times its weight. When squeezed, up to 90% of the oil could be
removed from the foam leaving it as light as three times its
original weight. When reused, the foam, prewetted with residual
oil, had more absorbency than the virgin foam. Tests also
showed that the foam could be recycled over 100 times with only
slight loss of absorbency.
-------
300
250
200
c
•J
IH
cn
0)
OJ
2 150
ffl
0)
o>
i
50 -
O
E
Reference 2
Reference 4
(Diesel Fuel)
Design Squeezing
Force = 7
I
5 10 15 20
Squeezing Force, N/mm of Belt Width
25
Figure 4.1. Regenerated sorbent density vs. squeezing force.
10
-------
Following these recommendations, sorbent cubes of open-
cell, reticulated, polyurethane foam were procured. The foam
density is 32 kg/m3, and each cube measures 19 mm on an edge.
Pore size is 3 pores/mm.
Figure 4.1 is a plot of the regenerated sorbent density
versus squeeze force that was derived from data in References 2
and 4. As shown in the figure, a design squeeze force of 7 N/mm
was selected to give a reasonable low regenerated density at a
moderate squeezing force level.
Sorbent Broadcasting Unit
Preliminary Design and Development of the Sorbent Broadcaster—
The sorbent broadcaster was developed and tested by
Hydronautics, Inc., under a subcontract. Their report on the
development of this work is the Appendix to this report.
Post-test Modifications to the Sorbent Broadcaster—
Two desirable modifications to the sorbent broadcaster
became apparent during the testing at OHMSETT--a swivel in the
discharge nozzle (Figure 4.2) would allow the operator to change
the discharge direction easily, and an idler pulley would alle-
viate slippage problems and the tendency for the belts to come
off during operation. Both of these modifications were incor-
porated prior to delivery of the system to the E.P.A.
The idler pulley is mounted on the broadcaster's base, and
it engages the V-belts between the engine and the blower. The
pulley adjusts to provide sufficient tension to prevent the
V-belts from slipping. Another benefit of the third pulley is
the prevention of the V-belts jumping off the pulleys.
As tested, the broadcaster nozzle was not free to turn about
the axis of the flexible ducting. This made it awkward for the
operator to broadcast the sorbent. Therefore, a swivel was
added between the ducting and the nozzle. The design is simple:
pins engaged in slots allow the nozzle to turn relative to the
ducting, but the sorbent still encounters a smooth interior.
11
-------
Final Design of the Sorbent Broadcaster—
Principal characteristics of the sorbent broadcaster are
listed in Table 4.1. Two men can transport this unit, since
the heaviest component, the blower assembly, (Figure 4.3) has a
mass of 80 kg. Its compact size is suitable for transportation
in a pickup truck along with other system components.
An overall view of the sorbent broadcaster is shown in
Figure 4.4, taken during the tests at OHMSETT. In this case, the
sorbent cubes were fed to the blower from a feed hopper by a
screw conveyor (at the right). The flexible ducting allows the
operator to aim the cubes in any direction.
The blower assembly is constructed of commercially available
parts. The blower itself is a Dayton "High-pressure blower" with
a 340-mm open material handling wheel (Figure 4.5). It is powered
by a 5.3-kW gasoline engine through a V-belt drive. The complete
assembly is mounted on an integral frame with detachable wheels
for easy deployment.
The sorbent cubes are transported from the blower to the
nozzle through standard 8-inch sheetmetal ducting. The length
and geometry of the ducting may be adjusted by using standard
straight and adjustable elbow sections as required, although
ducting lengths over 7.5 m is not recommended. The number and
sharpness of the bends in the ducting should also be minimized
for best performance. The last 1.8 m of the duct, between the
rigid sheetmetal and the nozzle, is standard flexible ducting.
The combination of the duct flexibility and the axial nozzle
swivel permits the operator to aim the sorbent stream easily
onto the oil.
Figure 4.2. Discharge nozzle of sorbent broadcaster
12
-------
A - engine
B - screw conveyor (not part of system)
C - inverted-cone hopper
D - blower
E - rigid duct
Figure 4.3- View of sorbent broadcaster's blower,
13
-------
A - discharge nozzle
B - flexible duct
C - sorbent broadcaster
Figure 4.4. Overall view of sorbent broadcaster,
Figure 4.5. Open material handling wheel of broadcaster.
14
-------
TABLE 4.1. CHARACTERISTICS OF THE SORBENT BROADCASTER
Capacity
Length of Blower Assembly
Width of Blower Assembly
Height of Blower Assembly
Mass of blower assembly
Mass of pipe, hose, and nozzle
Blower type and size
Blower speed (unloaded)
Duct type and size
Power
Frame
1.1 kg/s
1.43 m overall
0.61m overall
0.99 m
80 kg
23 kg
Dayton "High-pressure blower"
No. 4C131 with a 340 mm
open material handling
wheel
2900 rpm maximum
200 mm diameter; 7.3m long,
including 1.8 m of flex-
ible duct, with swiveling
discharge nozzle
5.2-kW gasoline engine with
vertical shaft, driving
the blower with V-belts
Contains engine and blower;
integral handles and de-
tachable wheels
15
-------
Harvesting Conveyor
Preliminary Design of Harvesting Conveyor—
The preliminary design of the sorbent harvester was based
on previous design work done (Reference 2). The preliminary
design called for a small diesel engine to drive the belt
through a variable speed drive, permitting the belt speed to be
varied to match the relative velocity of the water. With a
small diesel committed to the harvester, independent operation
would be possible.
The belt was a round wire mesh of approximately 13 mm mesh.
It was to meet the water surface at an angle of approximately
0.52 rad to prevent the foam cubes from tumbling.
Also included in the preliminary design for a vessel-mounted
sorbent system was a transfer conveyor to collect the sorbent
from the harvester and transport them to the regenerator. The
belt was to have a speed range of 0.5 to 2.7 m/s and was to be
driven by a right-angle drive from the harvester.
Development of Harvesting Conveyor—
At the outset of the development program, it was decided to
stress the design goals of mobility and independent operation of
each component of the sorbent system. Emphasis on these goals
dictated much of the prototype design of the sorbent harvester.
In order to minimize individual component weights and to
provide greater operating flexibility, the decision was made to
utilize a gasoline engine, rather than a diesel engine, and to
drive the harvester hydraulically. The hydraulic drive motor
was small, lightweight, and was capable of providing a wide range
of harvester belt speeds. A commercially available gasoline
powered hydraulic power supply was selected. It was rated at
7.8 kW and had an output of 1.4 m3/h at 10,000 kPa. This unit
is lightweight and is easily transported since it is mounted on
rubber tires. The gasoline engine is equipped with both an
electric starter, and a manual backup for reliability and for use
in explosive conditions.
The unit connected to the harvester motor by rubber hydrau-
lic hoses, which provided flexibility in positioning the power
supply and harvester during operations. The hoses allowed the
power unit to be placed in a position remote from any explosive
vapors. Thus, the flexibility provided by using a hydraulic
drive system compensates for the lesser degree of safety of the
gasoline engine as compared to a diesel engine.
16
-------
The harvesting conveyor structure was fabricated in aluminum
to keep its weight to a minimum. A stainless steel open mesh belt
was used for the conveyor. The openings in this belt were select-
ed so as to be large enough to permit water to flow easily through
the mesh, at the same time preventing the foam cubes from falling
through.
In order to keep the length of the harvester to a minimum,
it was desirable to operate the conveyor at the steepest angle
practical. Reference 2 indicated that an angle of 0.52 rad above
the horizontal was about the steepest inclination possible without
incurring tumbling of the cubes down the conveyor face; therefore,
this angle was selected. A conveyor length of 3.7 m was selected.
At the 0.52 rad angle, this provided for a vertical lift of about
1.5 m between the waterline and the sorbent discharge.
An aluminum channel was placed across the upper end of the
conveyor to prevent large debris from being carried all the way up
the conveyor. The sorbent cubes discharged from the conveyor belt
into a 230-mm screw conveyor mounted transversely below the top of
the conveyor belt. The conveyor screw, driven by its own hydraulic
motor, was designed to be capable of being stopped and reversed
to clear jams between the screw and conveyor trough. The screw
speed was controlled by the hydraulic power unit engine speed.
Provision was made to extend the screw conveyor length by the
attachment of additional trough lengths and a longer screw. The
screw conveyor is a standard, commercially available size.
During shop testing of the completed harvester unit, it was
found that the inception of sorbent cube tumbling down the con-
veyor face commenced at about 0.4 rad inclination. To prevent
this tumbling when operating at the design angle of 0.52 rad,
flights were added to the conveyor belt. The flights were made
by brazing spare conveyor belt wires at right angles to conveyor
belt surface. The zig-zag shape of the wires provided 13 mm high
flights across the belt face.
In order to support the conveyor, two pillow blocks were
located under the side rails just above the center of gravity
where the support shafts connected to foundations could be in-
serted. The lower end of the unit was supported by two articu-
lated foam-filled pontoons mounted to the outsides of the frame
at the water line. These pontoons allowed the conveyor to follow
the undulating water surface.
For ease of overland transportation of the harvester,
pneumatic tires were provided. They attached to the screw con-
veyor trough assembly at the upper end of the unit and were easily
removable. (Figure 4.6.)
17
-------
A - float
B - diverter plate
C - collection bin
D - screw conveyor
Figure 4.6. Harvesting conveyor at stream bank.
Figure 4.7. Harvesting conveyor on catamaran.
18
-------
TABLE 4.2. CHARACTERISTICS OF THE HARVESTING CONVEYOR
Length
Width
Height
Mass
Conveyor width
Conveyor angle
Conveyor length
Belt speed
Belt type and size
Belt drive
Power
Support
3.7 m overall at operating
angle of 0.52 rad
1.4m overall
2.2 m overall, at operating
angle of 0.52 rad
140 kg
0.91 m
0.52 rad from horizontal
3.7 m
to 2 m/s
heavy duty, stainless steel
round-wire belt, 13 mm mesh
76-mm sprockets
hydraulic pump (1.4 m3/h at
10,000 kPa) driven by a
7.8-kW gasoline engine
Integral support frame with
wheels and with telescoping,
adjustable legs; elastomer-
covered foam floats support
the lower end of the har-
vester in the water at its
proper draft
19
-------
Post-Test Modifications to the Harvesting Conveyor—
During testing at OHMSETT, two modifications to the har-
vester seem desirable—removing the expanded metal cover, and
adding adjustable legs for support.
As tested, the harvester had an expanded metal cover over
the belt, from the deflector plates at the lower end to near the
drive sprockets (Figure 4.6). The cover prevented easy access to
remove debris, and since it was evident that there would be
little attendant danger, the expanded metal cover was removed.
When deployed from a vessel, the harvester was well support-
ed as tested, but there was no provision to account for uneven
ground if the harvester were to be deployed from a stream bank.
Adjustable telescoping legs were added to the harvester with sup-
ports that rotated, allowing the legs to be adjusted in order to
support the harvester belt at an angle of approximately 0.5 rad
from horizontal.
Final Design of Harvesting Conveyor—
The characteristics of the sorbent harvester are listed in
Table 4.2. Its size allows it to be transported in the back of
a standard pickup truck. With a mass of 140 kg, two men can
remove it from the truck and roll it over rough terrain to the
spill site. Once there, telescoping, adjustable legs enable the
operators to support the harvester at the desired angle of 0.52
rad.
Another mode of operation is shown in Figure 4.7. This
photograph, taken during the tests at OHMSETT, shows the har-
vester mounted on a catamaran. The lower end is supported by
two foam floats. Each float is covered with abrasion-resistant
elastomer and is articulated, allowing the floats to conform to
the water surface.
The conveyor itself is 0.91 m wide and 3.7 m long, permit-
ting a vertical lift of 1.5 m from the water surface. Hydraulics
power both the belt and the screw conveyor. Figure 4.6, taken
before the post-test modifications were made, shows the belt's
hydraulic motor and controls, as well as the screw conveyor with
pneumatic wheels attached. The screw conveyor's separate hyd-
raulic motor is mounted at the hidden end of the shaft.
Sprockets drive the belt directly. The belt speed is
adjustable to 2 m/s, allowing the operators to control the speed
as conditions dictate. If the belt is driven slowly, the sor-
bent cubes logjam in front of the harvester. This increases the
contact time,which should improve recovery of the oil. During
the test at OHMSETT, the belt was always driven fast enough to
prevent logjamming, in order to allow the sorbent residence time
to be easily determined, thus keeping the number of test para-
meters manageable.
20
-------
Detachable aluminum side plates, which bolt onto the frame,
channel the sorbent cubes into the harvester. These plates ex-
tend above and below the water surface to prevent escape of the
sorbent, even in waves. The design of the plates also allows for
the connection of diverter booms.
The power supply (Figure 4.8) is a separate component, add-
ing further to mobility of the system. This unit is commercially
available. The hydraulic pump is driven by a 7.8-kW gasoline
engine, and it delivers 1.4 itr/h at 10,000 kPa. Hydraulic hoses
carry the fluid to the harvester, providing flexibility in place-
ment of the power supply.
Sorbent Regenerator Unit
Preliminary Design of Sorbent Regenerator—
The preliminary design of the regenerator was based on the
converging-belt sorbent regenerator of earlier EPA contracts
(References 2 and 4). One major change of the preliminary design
was to use springs to produce the compression load, another was
to select a squeeze force of 7 N/mm, a compromise between the
squeeze forces used in the two contracts referenced.
Figure 4.1 shows data from the earlier work. At low squeeze
forces, approximately 1 N/mm, regenerated sorbent density ranges
up to nearly 300 kg/m3, or about 75% removal of the oil. Oil
removal can be increased to 95% at high squeeze forces approach-
ing 25 N/mm. At a squeeze force of 7 N/mm the regenerated sor-
bent density is approximately 130 kg/m3. This corresponds to 90%
removal of the oil from the sorbent.
Selecting a squeezing force of 7 N/mm makes the regenerator
more tolerant of debris. The lower squeezing force also requires
a less massive structure, resulting in a weight saving.
Friction was chosen to drive the belts because of the prob-
lems Hydronautics experienced with sorbent jamming between the
chain and sprocket. The belt width was 61 cm although3only 36 cm
are required for the design oil recovery rate of 5.7 m /h. The
sorbent was to be fed onto the center area of the belt, with
shrouds to prevent sorbent losses off the sides of the belt.
Power was to be provided by a 2.6-kW diesel engine. Roller
chain and sprockets were to transmit the power to the two drive
rollers, one for each belt. (This preliminary engine is much
smaller than the final design because there was no hydraulic
system.)
The preliminary design called for a rotating brush scraper
to remove the sorbent cubes from the lower belt, to reduce the
loss of cubes to the lower collection pan.
21
-------
A - hydraulic hoses
B - hydraulic oil filter
C - hydraulic oil filter pressure gauge
D - hydraulic oil reservoir
E - hydraulic oil pressure gauge
Figure 4.8. Harvesting conveyor's power supply.
22
-------
Figure 4.9. Overall view of sorbent regenerator.
Development of Sorbent Regenerator--
Early in the program, it was decided to change the design
to utilize hydraulic power. The switch to hydraulic power pro-
vided several advantages: the regenerator could be made lighter
and more compact; flexibility in positioning the power supply
relative to the regenerator was possible; a commercial power
supply could be used; drive power, squeeze pressure, and belt
tension force could all be provided by the same power source;
squeeze force could be adjusted over a wide range and yet held
nearly constant as the squeeze rollers moved up and down with
varying sorbent thickness; conveyor belt speed was easily con-
trolled over a wide range; conveyor belt jams could be cleared
by powering open the squeeze rollers.
The prototype sorbent regenerator is shown in Figure 4.9.
Saturated sorbent is fed into the feed hopper at the input and
directly onto the lower mesh belt. Any free oil or water drains
through the belt and into the primary collection trough. The
motion of the lower belt transports the foam into the squeezing
zone (Figure 4.10). A leveling bar at the exit from the feed
hopper prevents any large debris from moving to the squeeze
rollers and also acts to level off the sorbent layer.
23
-------
The two regenerator belts are friction driven by a drive
roller. Sprockets on the drive roller ends (Figure 4.11) are
connected by an idler chain to the low speed, radial piston hy-
draulic motor. The hydraulic motor is located directly under
the feed hopper.
As the belts approach the squeeze zone, they converge,
sandwiching the sorbent foam cubes between them. The squeeze
rollers are each 254 mm in diameter. The lower two are fixed
and the upper two are free to move vertically. In order to
accommodate uneven sorbent flow, these rollers are connected to
gimballed bearing blocks which allow the rollers to rise uneven-
ly. Each of the four bearing blocks is connected to the rod of
a double-acting hydraulic cylinder.
Three manual three-position hydraulic flow control valves
are used to control the regenerator functions. One controls the
conveyor drive motor, another directs the hydraulic flow to the
cylinders on the upper belt tension roller to control tension.
The third control valve controls the squeeze roller cylinders'
force.
The hydraulic fluid pressure to the tension roller and
squeeze roller cylinders is controlled by the adjustable pres-
sure control valves located over the corresponding directional
control valves. In order to keep the hydraulic pressure con-
stant on the tension and squeeze cylinders, the "down" side of
the cylinders are connected to 0.0004 m3 hydraulic accumulators
charged with nitrogen to 1400 kPa. During operation, the gas
volume in the accumulators keeps the squeeze and upper belt
tension force nearly constant, at the level set by the pressure
control valve, regardless of squeeze and tension roller vertical
movement.
In order to keep the upper and lower conveyor belts centered
on the rollers, the drive rollers are coated with elastomer and
ground with a 6-mm crown in the center. Bearing blocks on the
shaft of the tail roller for the lower belt have adjustment
screws which are used both to tension lower belt and to prevent
the conveyor belts from running to one side or the other.
Most of the oil squeezed from the sorbent drains through
the lower mesh belt and drains to the bottom of the collection
trough and out through the 76-mm drain fitting. To improve the
flow of cold or viscous oil, the bottom wall of the collection
trough is fitted with a baffled chamber filled with stainless
steel wool. The collection trough can be heated by passing
steam or the power supply engine exhaust gases through this heat
exchange chamber. A drip pan with a separate drain fitting is
located under the complete length of the regenerator unit to
catch any oil dripping off the lower belt or other parts of the
unit.
24
-------
A - lower belt
B - upper belt
C - upper belt tension roller
D - tension accumulator
E - tension accumulator pressure gauge
F - squeeze accumulator
G - squeeze accumulator pressure gauge
Figure 4.10. Detail of sorbent regenerator.
25
-------
B
A - discharge chute
B - squeeze hydraulic cylinder
C - upper belt tension hydraulic cylinder
D - chain-and-sprocket drive
E - chain guard (removed)
F - bearing adjustment screw
Figure 4.11.
Sorbent regenerator with chain
guard removed.
26
-------
Figure 4.12. Sorbent cubes tumbling down
regenerator's discharge chute.
Counterrotating cylindrical bristle brushes sweep the sor-
bent cubes off the upper and lower belts at the discharge end of
the regenerator, causing the sorbent to fall by gravity down the
removable aluminum discharge chute (Figure 4.12). At the bottom
of the chute, the sorbent can be fed onto a conveyor or into tote
pans or a collection hopper.
The regenerator unit weighs approximately 820 kg. It can be
lifted either by a fork lift from the discharge end or by means
of a wire rope bridle attached to the two eyebolts on the frame
and the shackle on the end of the feed hopper. When used in the
field, the regenerator can be operated in the bed of a pickup
truck.
Power for the regenerator is provided by a commercially avail-
able, gasoline engine-driven 12-kW hydraulic power supply mounted
on a rubber-wheeled portable frame (Figure 4.13). The unit has
an output capacity of 2 m3/h at 14,000 kPa. The power supply
engine is equipped with a manual start for maximum reliability.
A flow-control valve is provided to unload the hydraulic pump
during starting and to provide a remote means for stopping the
regenerator if necessary. Three hoses connect the power supply to
the regenerator unit by means of quick-connect fittings. Two
13-mm hoses are the main hydraulic supply and return, while .the
smaller hose is the case drain return from the regenerator drive
motor.
27
-------
A - throttle
B - control valve
C - hydraulic oil filter pressure gauge
D - hydraulic oil filter
E - hydraulic hoses
F - hydraulic oil reservoir
Figure 4.13. Sorbent regenerator's power supply.
28
-------
The reasoning behind replacing the diesel engine and diesel
drive with a hydraulic power supply driven by a gasoline engine
was the same as for the harvesting conveyor. The benefits were a
weight saving, increased operational flexibility, and greater
availability of parts. The operational flexibility in positioning
the power supply compensates for whatever loss of safety may ac-
company the replacement of a diesel engine with a gasoline engine.
Post-Test Modifications to the Sorbent Regenerator—
The major desirable modification to the sorbent regenerator
that became evident during the tests at OHMSETT was to provide
more positive drive and tracking for the belts. Several other
minor items were also modified.
Throughout most of the tests at OHMSETT, the only problem
with the belts was that they needed occasional adjustment to track
properly. Approximately every 30 minutes a technician would make
the small adjustment. In later tests with an oil of high lub-
ricity, one of the belts would occasionally stop because the
drive roller lost friction, and subsequently the other belt would
stop. The addition of a slide bed with centering strips and
tracking guides solved the tracking problems for the lower belt,
even if the regenerator is not level.
As tested at OHMSETT, several small rollers supported the
lower, stainless steel belt between the drive roller and the
first squeeze roller. Crowned rollers provided tracking, but it
was apparent that more positive tracking was needed. When the
regenerator was not level, the belt would track to one side.
This problem was solved by installing a slide bed to replace
the small support rollers. Nylon strips, riveted to a 1.4-m
long aluminum plate in a chevron pattern, center the belt when the
regenerator is nearly level. Crowned rollers augment this track-
ing. When the regenerator is not level, guides along the sides
of the belt prevent it from walking too far to one side or the
other. The guides are made of seasoned oil-soaked oak.
Although no modifications were made to the unit, prior to
its delivery, to eliminate the belt slipping encountered with
high lubricity oil, it is felt that applying conventional lagging
to the upper belt drive roller in a chevron pattern will both
improve the driving reliability and the belt tracking.
Other modifications to the sorbent regenerator include adding
a splash guard, and plumbing two discharge lines into one line
(Figure 4.14).
29
-------
As tested, the regenerator used two hoses to remove the oil,
one connected to the primary oil collection sump and the other to
the drip pan. To simplify this, both the upper and the lower
collection pans were brought to a common discharge connection
through a 2-way, 3-port valve. In the normal valve position, the
pump takes suction from the main upper collection pan. When the
lower pan begins to fill, which only happens occasionally, the
operator momentarily changes the valve position and the pump
takes suction from the lower pan. A male quick-connect fitting,
compatible with the 76-mm discharge hose, attaches to the elbow.
The regenerator's hydraulic motor can be seen opposite the 3-way
valve in Figure 4.14.
Final Design of the Sorbent Regenerator—
The regenerator squeezes oil from the sorbent cubes by means
of converging belts. The lower belt, made of stainless steel
wire cloth, is porous, allowing the oil to drain through. The
upper belt has polyurethane covers which make it highly abrasion-
resistant and impervious to oil. Three rollers move vertically
by means of hydraulic cylinders: one roller tensions the upper
belt, while the other two provide the squeeze force. Character-
istics of this unit are contained in Table 4.3.
With a mass of 820 kg, the sorbent regenerator is the only
component of the system that two people cannot maneuver. However,
its size allows it to be carried and operated in the back of a
pickup truck (Figure 4.15). The unit's dimensions are 2.44 m
long by 0.94 m wide by 1.77 m high overall.
The power supply can also be seen in Figure 4.15. The
hydraulic pump driven by a 12-kW gasoline engine delivers
2 m3/h of hydraulic fluid at a pressure of 14,000 kPa. The
hydraulic system is fully described later in this section.
Post-test modifications to the sorbent regenerator include
improved drive and tracking for the belts, piping and a valve
to consolidate the discharge lines into one, and a splash guard.
30
-------
A - 3-port, 2-way valve
B - discharge port
C - hydraulic motor
D - upper collection pan
E - lower collection pan
Figure 4.14.
Post-test modifications to the regenerator:
plumbing to combine the two discharge ports,
31
-------
Figure 4.15. Sorbent regenerator operating
in the back of a pickup.
32
-------
TABLE 4.3. CHARACTERISTICS OF THE SORBENT REGENERATOR
Length
Width
Height
Mass
Upper belt type
and size
Upper belt drive
Lower belt type and
size
Lower belt drive
Belt speed
Squeeze force
Number of squeezings
Power unit
2.44 m overall
0.94 m
1.77 m overall
820 kg
Polyester conveyor belting
with polyurethane coating,
610 mm wide, 2.7m long
Friction
Balanced weave, stainless
steel wire cloth, 610 mm
wide, 6.1 m long, 7.1 mm
thick
Friction; tracking provided
by herringbone slide bed,
and tracking guides
To .61 m/s
Adjustable, nominally 7 N/mm
2
3
Hydraulic pump (2 m /h at
14,000 kPa) driven by a 12-kW
gasoline engine
33
-------
SECTION 5
SYSTEM TESTING AT OHMSETT
Extensive tests were conducted at U. S. Environmental
Protection Agency's OHMSETT test facility. The results showed
that the sorbent oil recovery system is an effective means of
recovering accidental spills, particularly in choppy waters
that defeat many other types of cleanup devices. The results
also indicated several desirable modifications to the system.
TEST PROGRAM
Test Layout
The test layout simulated field conditions and allowed a
realistic evaluation of the performance envelope of the system.
Factors involved in selecting the test layout were:
1. There had to be adequate contact time between
the sorbent and the oil. This dictated the
distance between the broadcast and recovery
units as a function of speed.
2. The broadcast and recovery of the sorbent
had to be carried out in a test area in
which side losses of both oil and sorbent
were controlled. This dictated the maximum
funneling angle at a given test speed.
3. The area of sorbent coverage had to be large
enough to minimize "over-spray" losses of the
sorbent past the funneling boom, but small
enough to be manageable. This dictated the
boom "mouth" opening.
4. Provision had to be made for transfer of
personnel and recovered sorbent after the
completion of a test.
5. The test layout had to be compatible with
existing EPA tank procedures and equipment
inasmuch as possible.
34
-------
In line with the above factors, a preliminary test layout
was made as shown in Figure 5.1. With the planned test layout
the following contact times were calculated:
Velocity Contact Time
Knots m/s s
1
2
4
6
.5
1.0
2.1
3.1
36
18
9
6
A 15-second or greater contact time between sorbent and oil
was considered desirable from earlier work with absorption, and
this could be obtained at the 1- and 2-knot speeds. Reasonable
contact times could also be obtained at the higher speeds.
A funneling angle of 0.18 rad is obtained with the planned
test layout. This gives a velocity normal to the boom of 0.7
knots at a 4-knot tow speed and 1.1 knots at a 6-knot tow speed.
This was adequate to prevent significant oil or sorbent losses
while funneling.
The 4.9-m boom spacing allowed adequate room for broadcast-
ing the sorbent. It also allowed in excess of 5.7 m3/h oil
encounter rate for a 0.75-mm slick at a 1-knot tow speed.
35
-------
• Tow Carriage
Tow Direction
D.-.1 Discharge
—Orifices
n—ia = ,
Regen
on T
Crane
rator
ailer
D
Figure 5.1. OHMSETT test tank arrangement.
36
-------
Test variables
Tests were run over a range of sweep speeds, oil thick-
nesses, wave heights, and oil types. The independent variables
(parameters to be varied) and dependent variables (variables to
be measured) were:
Independent Variables
Dependent variables
1. Sweep speed
(1,2,4,6 knots)
2. Oil thickness
(0.75, 1.5 mm)
3. Waves (calm to 0.6 m
wave height)
4. Wind speed (variable,
depending on weather
during test)
5. Oil type (light,
heavy)
6. Sorbent broadcast
rate
1. Recovery rate
2. Oil fraction (ratio of
collected oil to total
collected liquid volume)
3. Volumetric efficiency
(volume of oil recovered
per unit volume of
sorbent recovered)
4. Throughput efficiency
(percentage oil en-
countered which was
recovered)
5. Sorbent losses
The independent variables were controlled or measured as
follows:
1. Sweep speed
The sweep speed was determined by the rate at which the
tow carriage was pulled down the length of the tank. Sweep
speed was set for each run in accordance with the predeter-
mined test matrix and held constant for the entire run.
2. Oil thickness
The test matrix specified an average slick thickness
for each test run. Using the test run advance speed and the
fixed width at the mouth of the diverter booms, the required
flow rate to obtain the desired slick thickness was cal-
culated. The oil distribution pump and control valves were
37
-------
then adjusted to this flow rate prior to each test run.
During the run, the oil was sprayed onto the water from the
distribution manifold nozzles on the leading edge of the
tow carriage. The actual slick thickness was calculated
after each test from the total oil volume distributed,
sweep speed and boom width. For the purpose of this cal-
culation, it was assumed that the slick was uniform in
thickness.
3. Wave conditions
The wave type desired for each run to match the pre-
determined test matrix was generated using the OHMSETT tank
wave maker. Using previously calibrated settings for the
wave flap stroke, speed and phase, the type of wave (regular
or chop), wave height and period were adjusted. The wave
maker was not operated during calm water test runs.
4. Wind speed
No control could be exercised over wind speed or
direction. The wind speed and direction was measured and
recorded for each test run.
5. Oil type
The type oil laid down on the water was in accordance
with the predetermined test matrix. It was controlled by
connecting the appropriate oil storage container to the oil
distribution system.
6. Sorbent broadcast rate
The sorbent broadcast rate was the most difficult in-
dependent variable to control. Although the test matrix was
designed around discrete sorbent broadcast rates, it was
not possible in practice to accurately preselect the sorbent
rate. Instead, the total sorbent volume broadcast during
a test run was determined by calculation after each run,
from the weight of saturated sorbent collected and the
average density of the saturated sorbent. The sorbent broad-
cast rate was then obtained by dividing the total sorbent
collected by the broadcast time.
It had been planned to control the sorbent broadcast
rate by controlling the speed of the screw conveyor which
carried the sorbent from the feed hopper to the broadcaster.
However, good speed control was not obtainable and the sor-
bent transport rate was not constant at a fixed screw speed,
because of sorbent bunching and surging in the conveyor.
38
-------
Measurement of the dependent variables were made as follows:
1. Recovery rate
The total operating time of the recovery device during
a test was measured. The sorbent was weighed before and
after the tests, and the recovered oil and water was
measured by difference. This weight was converted to a
volume and the recovery rate was calculated as collected
volume divided by operating time. Also, when the sorbent
was regenerated, the collected volume of oil and water was
measured, and a second determination of the recovery rate
was made using this information.
2. Oil fraction
The liquid from the regenerator was collected, and
the oil content was measured by determining the oil-water
interface. The oil fraction was calculated by dividing the
volume of collected oil by the total volume of liquid
recovered.
3. Volumetric efficiency
The volumetric efficiency is the ratio of the oil
collected to the volume of sorbent recovered. The oil
recovered was determined by direct measurement. The
recovered sorbent volume was determined by sampling the
recovered saturated sorbent and measuring its density and
dividing the weight of the recovered sorbent by this
density.
4. Throughput efficiency
Throughput efficiency was defined as the percentage of
the oil encountered which was recovered. It was calculated
by dividing the recovered oil volume by the volume of oil
deployed.
5. Sorbent losses
Sorbent lost past the system was swept back after a
test run with the sweep boom on the tow bridge. The number
of sorbent cubes were estimated and compared to the volume
broadcast during the test run. Losses were negligible on
all runs.
39
-------
Test Procedure
Before the actual testing began, all equipment was assem-
bled and checked out. Several dry runs were made to check out
test procedures.
The general test procedure for these tests was as follows:
A. Before test
1. Set tow speed
2. Set up sorbent broadcast and recovery hoppers
3. Turn on broadcast and recovery equipment
4. Set oil flow rate
5. Set wave height and period and begin wave
generation
6. Set sorbent screw conveyor speed to desired
broadcast rate
7. Set up still and motion picture photographic
equipment
8. Measure water temperature, air temperature,
wind speed, and record general weather conditions
9. Start broadcaster and harvester
B. During test
1. Tow apparatus down tank
2. Release oil at controlled rate
3. Broadcast sorbent
4. Recover sorbent and catch in hoppers
5. Sample recovered sorbent
6. Photograph test
C. After test
1. Weigh recovered sorbent
2. Estimate volume of lost sorbent
3. Regenerate collected sorbent
4. Measure volume of collected oil
5. Measure water content of collected oil
6. Measure density of both the saturated and the
regenerated sorbent.
A total of 33 individual test runs were conducted. On seven
of these runs the run was aborted because of jamming of the
broadcaster and no meaningful data were collected. This jamming
resulted from choking the broadcaster by supplying sorbent to it
at a faster rate than it could handle. The jamming was just one
manifestation of the more general problem of inability to closely
control the sorbent broadcast rate.
40
-------
The test plan matrix was based on having the sorbent
broadcast rate as a controllable independent variable. It was
assumed that this could be done by using a screw conveyor with
a variable speed drive to supply the sorbent to the broadcaster.
In preliminary tests with dry. lightweight sorbent, a linear
relationship was found between the sorbent delivery rate from
the conveyor and the screw speed. However, when heavier, pre-
wetted sorbent was used, the conveyor delivery rate ceased to be
constant or linearly proportional to the screw speed. The sor-
bent in the feed hopper and screw conveyor no longer behaved like
a fluid, instead it compressed and bunched causing surging and
uneven delivery from the conveyor. As a result, the sorbent
broadcast rate was not a well controlled variable. The average
rate for each run had to be determined at the conclusion of the
run from the recovered sorbent density and total weight measure-
ments .
The inability to accurately control the sorbent broadcast
rate precluded the following of the test matrix which specified
the broadcast rate to be used for each run. It also prevented
the duplication of test runs to check their repeatability.
The test conditions and the results for all test runs are
contained in Table 5.1.
TEST RESULTS
General Conclusions
Because of the relatively limited number of test runs, the
lack of repeatability and the number of variables associated
with the system, the test results have more qualitative meaning
than quantitative. There were many sources of inaccuracy in the
measured data. Slick thickness and sorbent distribution were
not uniform across the slick surface, broadcast rates were not
steady, collected sorbent and regenerated sorbent densities were
determined by samples which may not have been truly representa-
tive. The fact that on two tests throughput efficiencies in
excess of 100% were obtained, indicates the danger of putting
too much reliance on the actual numerical results. The through-
put efficiency of 117% for run 16 resulted from the volume of
oil recovered being measured as 0.01 m3 greater than the amount
applied. A 0.01 m3 error could easily have resulted from
residual oil in the regenerator and drain hoses, or from an
error of a few millimeters in measuring the thickness of the
oil in the collection tank.
Several general conclusions or trends can be derived from
the test results.
41
-------
TABLE 5.A. TBST CONDITIONS AND RESULTS
to
Test 1
1
2
3
4 Ul
5
6
7
8
9 (1)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25 (1)
26
27
28 (1)
29 (1)
30 (1)
31 (1)
32
33
Run Speed
0.76
0.51
1.02
1.27
0.76
0.76
0.76
1.02
.02
.27
.27
.02
.02
.02
.02
.02
.02
1.02
1.02
1.52
2.54
2.03
2.54
1.02
1.02
1.02
1.02
0.51
0.51
0.51
0.51
2.03
1.52
Slick
Thickness
(mm)
0.54
0.81
0.54
0.50
0.54
0.54
0.54
0.54
0.56
0.50
0.51
0.79
0.24
0.84
0.55
0.21
0.29
0.54
0.80
0.41
0.28
0.30
0.22
0.53
0.55
0.54
0.28
0.69
0.78
0.65
0.40
0.32
0.40
Videos! ty
(mra'/s)
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.3
6.
6.
6.
6.
6.
6.
295
295
295
295
295
295
295
295
295
295
Sorbent Have (2)
Rate Height
(m'/s) «.Vs)
0.0032 calm
0.0050 0.3-n R
0.0054 calm
0.0024
0.0021
0.0057
0.0031
0.0048
0.0031
0.0031
0.0024
0.0035 0.3-m R
0.0029
0.0032
0.0031 0.3-M H
0.0038
0.0036
0.0042
0.0074
0.0043 calm
0.0057
0.0052
0.0051
0.0052
0.0052
0.0056 0.3-i« ti
Recovery
Rate
(m'/h)
6.3
B.I
9.6
5.4
5.5
7.3
8.5
10.5
6.0
8.5
4.6
7.9
7.9
4.8
5.5
9.1
9.2
9.2
6.8
5.4
5.2
8.5
7.8
4.4
7.0
8.7
Throughput
Efficiency
84
10B (3)
82
71
62
86
7fl
90
50
52
88
46
70
117 (3)
96
79
60
72
48
42
43
81
76
84
57
80
Oil
Fraction
0.62
0.48
0.57
0.66
0.74
0.44
0.68
0.63
0.70
0.79
0.62
0.73
0.78
0.56
0.55
0.70
0.75
0.67
0.38
0.51
0.38
0.76
0.60
0.63
0.74
0.75
Vol. Liq.Rcvd
Vol. Sorb. Rcvd
0.79
0.79
0.74
0.80
0.82
0.63
0.93
0.78
0.66
0.80
0.70
0.81
0.81
0.60
0.73
0.79
0.78
0.74
0.55
0.49
0.49
0.51
0.65
0.45
0.39
0.49
Volumetric
Efficiency
-in
38
42
53
60
28
63
50
46
63
44
59
63
34
40
55
58
50
20
25
19
38
39
29
29
36
(1) Broadcaster jammed - Test aborted
(2) R - Random Harbor Chop
w - Regular Nave, 1.5-s Period
(3) Experimental error, these values should not exceed 100%.
-------
Throughput efficiency tends to be inversely proportional
to run speed. The most likely cause for this is that residence
time of the sorbent in the slick was reduced as speed increased,
This is because the distance involved remained fixed throughout
the tests. These distances were from the broadcaster to the
harvester, and the length of the harvesting conveyor. As a
result, the time the cubes spent absorbing oil and draining
decreased with increasing speed. This trend occurred for both
light and viscous oils. As evidence, compare the results of
tests 13 v. 23 (light oil, calm water), 17 v. 21 (light oil,
regular waves), and 27 v. 32 (viscous oil, calm water):
Speed Throughput
Test No. (m/s) Efficiency
13
23
17
21
27
32
1.02
2.54
1.02
2.54
1.02
2.03
88
43
96
48
84
57
These pairs of tests were run with nominally the same slick
thicknesses. (See Table 5.1 for complete test data.)
With the light oil, throughput efficiency tended to be the
same or higher in waves than in calm water, probably due to
greater mixing. On the other hand, throughput efficiency with
the viscous oil does not appear to be affected by waves. Com-
parison of test 13 with tests 16 and 17 (run speed 1.02 m/s),
and test 21 with test 23 (run speed 2.54 m/s) show this weak
trend.
43
-------
Throughput
Test No.
13
16
17
0.
0.
Wave
Condition
calm
3-m regular wave
3-m random wave
Efficiency
(%)
88
117*
96
23 calm 43
21 0.3-m regular wave 48
* experimental error, this value should not be greater
than 100%
Volumetric efficiency (volume of oil recovered per unit
volume of sorbent recovered) increased with increasing slick
thickness. This was true for both light and viscous oils and
was probably due to the fact that with thinner slicks, more
sorbent cubes competed for the same amount of oil. Tests 17,
18, and 19 were all with the light oil at a run speed of 1.02 m/s
with nearly constant broadcast rate, and were all conducted in
0.3-m regular waves. Tests 27 and 26 were viscous-oil tests in
calm water at the same speed, also with nearly the same sorbent
rate.
Volumetric
Slick Thickness* Efficiency
Test No. (mm) (%)
17
18
19
27
26
.29
.54
.80
.28
.54
40
55
58
29
39
* calculated from oil distribution rate and width of
sweep
44
-------
Throughput efficiency is proportional to sorbent broadcast
rate, while volumetric efficiency is inversely proportional to
sorbent broadcast rate at low speed. With less sorbent in the
slick (low rate). each cube absorbs a larger amount of oil,
which increases volumetric efficiency. On the other hand,
throughput efficiency suffers at lower sorbent rates because
there is not enough sorbent in the slick. Test run 1, 5, 6, and
7 illustrate these points. These tests were conducted at the
same tow speed, with the same slick thickness, and with light
oil in calm water. Only the sorbent broadcast rate varied.
Run Sorbent Volumetric Throughput
Test speed broadcast rate efficiency efficiency
No. (m/s) (m3/s) (%) (%)
7 0.78 .0057 28 86
1 0.78 .0032 48 84
5 0.78 .0024 53 71
6 0.78 .0021 60 62
At higher speeds, this same trend for throughput efficiency
holds, but volumetric efficiency appears to be independent of
sorbent broadcast rate. Compare the results of tests 10 v. 11
and 32 v. 22. Each pair of tests was conducted in calm water
with very nearly the same amount of oil; only the sorbent broad-
cast rate varied.
Run Sorbent Volumetric Throughput
Test speed broadcast rate efficiency efficiency
No. (m/s) (m3/s) (%) (%)
10 1.27 .0048 50 90
11 1.27 .0031 46 50
32 2.03 .0052 29 57
22 2.03 .0043 25 42
45
-------
Volumetric efficiency for the light oil tended to be higher
with a longer residence time (as evidenced by slower speed),
but was unaffected by residence time for the viscous oil. Tests
8 and 11 were conducted with the light oil, and with the same
nominal slick thickness and sorbent broadcast rate.
Volumetric
Run speed efficiency
Test No. (m/s) (%)
8 1.02 63
11 1.27 46
On the other hand, compare the results for the viscous oil. The
only significant difference in parameters between tests 27 and
32 was residence time.
Volumetric
Run speed efficiency
Test No. (m/s) (%)
27 1.02 29
32 2.03 29
These results appear anomalous. It was witnessed during the
OHMSETT tests that the cubes absorbed the light oil far faster
than the viscous oil; therefore, it would seem that the above
trends should be the other way around.
The last general conclusion is that winds up to 15 knots
have negligible effect on system performance. This was the most
severe environmental condition during the tests at OHMSETT, and
the wind did not blow any sorbent cubes away. Of more signi-
ficance were waves breaking over the boom; but even at 2.57 m/s
in breaking waves, sorbent cube losses were only 2%.
46
-------
Data Correlation
In an attempt to evaluate the performance of the system
with the wide range of variables considered, the data was
entered into a computerized nonlinear multivariable regression
analysis. The program can handle linear combinations of non-
linear functions of the forms:
y = k-x + b y = k.log (x) + b
Y = k-x*5 + b y = k-ex + b
y = k-x + b log (y) = k-log (x) + b
In summary, system performance is a trade-off between
throughput efficiency, which requires high sorbent rates, and
the oil fraction in the recovered liquid, which required low
sorbent rates. Sorbent rates in the range of 0.004 to 0.006
m3/s appear to be best for the ranges of flow velocity, slick
thickness, and viscosity that were used in the OHMSETT tests.
Discussion of the nonlinear regressions follows.
Volumetric efficiency (Figure 5.2)—
The volumetric efficiency is a significant function of
sorbent rate and slick thickness. It is not significantly
affected by water flow velocity, indicating that residence
times in these tests were adequate. Viscosity does not signi-
ficantly affect volumetric efficiency.
The nonlinear regression is:
log VE = 0.0523-0.0789R + 0.0263 V~ + 0-385 log T
r = 0.8659
where VE = volumetric efficiency
R = sorbent broadcast rate (m3/s)
V = velocity (m/s)
T = slick thickness (mm)
47
-------
>,
o
0)
•H
U
•H
M-l
<4-l
W
U
•H
I*
-P
-------
Oil Fraction in Recovered Fluid (Figure 5.3)—
The fraction of oil in the oil-water mixture varies as a
function of all of the independent variables. Most significant-
ly, low sorbent rates and high viscosity improve the proportion
of oil in the recovery stream. Of secondary importance, but
still significant, are the effects of flow velocity (higher
velocity improves performance) and slick thickness (higher
proportion of oil with increasing thickness). Higher flow
velocity increases the oil fraction in the recovered liquid, in
general, because the contact time is less. The sorbent cubes
land on the oil and absorb it first. With shorter contact times,
there is less time to absorb water. A thicker slick also
results in a higher oil fraction for the same reasons. Only
with the high viscosity lubricating oil at low sorbent rates
did the fraction of oil in the recovered fluid exceed 90%.
The nonlinear regression of oil fraction is:
F = 0.0101-0.0641R + 0.621V + 0.221 log T + 0.0064 /TT
r = 0.7445
where F = oil fraction
U = viscosity (mm2/s)
Throughput Efficiency (Figure 5.4)
Throughput efficiency is significantly affected by all
variables. In general, the thinner the slick and the higher
the sorbent rate, the better the performance. Increasing water
flow rate severely reduces the amount of oil recovered.
The nonlinear regression of oil recovery throughput
efficiency is:
E = 128.8 + 7.023R - 42.8V - 61.63T - 3.23 log U
r = 0.854
49
-------
1.0
0.9
0.8
0.7
0.6
•o
3 0.5
C1
•H
0)
o
u
2
C
-H
O
•H
+>
U
-------
A
B
C
D
E
F
G
H
100 .,,,,.., ,
= :!S
90 ;;||;;
«« Htt::
ighput Efficiency (%)
<* m ^j a
o o o o c
g 30 :g:s:
% ii;i:|:!
20 iia!
10 i|||
0 ]
Flow
Velocity
0.51
1.02
1.02
0.51
1.02
1.02
1.02
2.54
ll§£^HHt£
git-t^^ttrr?
(-.». «— * ..W-.-«. f&
r.:r ^ ^r: rrp ?r
2 3
Sorbent
(m/s)
r—A—Jf
i
Rat
Th
^
t-^ii
xi — — •*-»
1
}
e (
Sli
ick
Tt ~£*T
n
F
5
n3/s
ck
ness
0.54
0.25
0.25
0.80
0.54
0.54
0.80
0.25
t P F F
w
X
(i
^
«•••
6
1C
nm)
tr4*
ftfl
t 1 1 i-
E~
)00
T7|t~^
1=JH
=*J
*
7 8
)
Vi
(IT
G
H
scosity
m2/s)
6.3
6.3
295
6.3
6.3
295
6.3
6.3
Figure 5.4. Nonlinear regression of throughput efficiency.
51
-------
REFERENCES
1. Cochran, R. A.; Eraser, J. P.; Hemphill, D. P.; Oxenham,
J. P.: and Scott, P. R. (October 1973). An Oil Recovery
System Utilizing Polyurethane Foam - A feasibility Study.
U. S. Environmental Protection Agency Report No.
EPA-670/2-73-084.
2. Gumtz, Garth D.; and Meloy, T. P. (September 1973). Oil
Recovery System Using Sorbent Material. U. S. Environ-
mental Protection Agency Report No. EPA-670/2-73-068.
3. Henager, C. H.; and Smith, J. D. (September 1972).
Concept Evaluation: Recovery of Floating Oil Using
Polyurethane Foam Sorbent. U. S. Environmental Protection
Agency Report No. EPA-R2-73-156.
4. Miller, E.; Stephens, L.; and Ricklis, J. (January 1973).
Development and Preliminary Design of a Sorbent-Oil
Recovery System. U. S. Environmental Protection Agency
Report No. EPA-R2-73-156.
5. Sarter, James D.; Foget, Carl R.; and Castle, Robert W.
(April 1973). Oil/Sorbent Harvesting System for Use on
Vessels of Opportunity. U. S. Environmental Protection
Agency Report No. EPA-R2-73-166.
52
-------
APPENDIX
DESIGN RATIONALE FOR THE PNEUMATIC OIL-SORBENT
TRANSPORT/BROADCASTING SYSTEM
by William T. Lindemuth
HYDRONAUTICS, Incorporated was subcontracted to provide
technical support to Seaward International, Inc., in the perform-
ance of its contract No. 68-03-2138 with the Environmental Pro-
tection Agency. HYDRONAUTICS1 primary responsibility was for the
design, prove-out and documentation of the pneumatic oil-sorbent
transport/broadcasting subsystem.
PRELIMINARY DESIGN
The preliminary design of the pneumatic broadcasting system
was predicated on the design method developed by HYDRONAUTICS,
Incorporated under contract to EPA and described in Reference 1 .
An important consideration in the design of this subsystem was
the requirement that it be self-contained, capable of being
carried by two men (i.e., less than 91 Kg). This constraint led
directly to the selection of the largest available blower compat-
ible with the weight limitation and a gasoline engine to drive the
blower.
The blower selected was a Dayton "High Pressure Blower"
No. 4C131 which has an open material handling (OMH) wheel (as
recommended in Reference 1 ) of 34 cm (13%") diameter. The inlet
diameter is 20 cm (8") and the outlet is 18 x 15 cm (7-1/8" x
5-3/4"). The blower is mounted on its side with the inlet facing
upward to allow gravity feed of the sorbent cubes. A cone-shaped
hopper is affixed over the inlet, both to increase the effective
feed area and to minimize inlet losses to the blower. An adapter
is added to the outlet to join with 20 cm (8") diameter transport
ducting.
53
-------
The galvanized transport duct is nominally 7.3 m (24 feet)
long including a 1.8 m section of flexible duct which allows the
sheet metal nozzle to be aimed manually to direct the sorbent
broadcasting as desired. The nozzle is incorporated to direct
the sorbent downward and to widen the broadcast pattern.
The design point for sorbent transport was based on a
sorbent density of 9.13 g/cc and recommendations of Reference 1,
namely: duct velocity =13.7 m/sec, and sorbent weight/air
weight, Wc/Wa = 1. The preliminary calculations for the blower
power follow.
Duct Area = ir/4 x 202 = 314 cm2
Air Volume, Va = 314 x 13'7 =0.43 m3/sec (910 cfm)
10"
Air Density = 1.2 x 10" 3 g/cc (.075 lb/ft3)
Air Weight, Wa = 1.2 x .43 = .52 Kg/sec
Sorbent Volume at 0.13 g/cc =
-43 = 4 x 10~3 m3/sec (8.5 cfm)
Head Loss in Duct (Reference 1, p. 131)
Note: English units hereafter.
W C W L
h = - ^ - 5. . f-L/D + 0.5 + ==£• + w 6Q
0.445 x 10* Wa Wa69.
f = air friction factor = 0.013
0.5 = nozzle loss coefficient (estimate for fixed
nozzle)
54
-------
C = sorbent friction factor =1.5 (estimate)
h = 0'4-?- (0.013 • 36 + 0.5 + 1.0) + 1>fQ'yl2
= 1.4 in. H20
= 0.3 (Ref. 1, p. 139) (conservative for sorbent
cubes)
Apparent Head = h/F, =4.65 in. H2O
3,
V / h = 910/-/4.65 = 422
Read off blower characteristic curve (Figure A-l)
^N = 0.365
.*. N = 910/0.365 = 2500 RPM
Air Horsepower = V x H/(n x 6350)
910 x 4.65 , n
~ 0.7 x 6350
W x V2 tip
"Chip power" = x 55Q
w - 2500 13.5 _ ,47 f
vtip ~ ^o^ x v x TT^ ~ 147 tps
w = - = 1>14 lbs/sec
C bU
T, , . _ 1.14 x 1472 _ n
Ideal hpchip - 64.4 x 550 -0
55
-------
0.7
0.6
& 0.5
J
u
0.4
0.3
0 1000 2000
V-A/TT. CfmXin.H.O*
O 2
Figure A-l. 13V Dayton blower characteristic at standard
conditions (0.075 lb./ft3).
56
-------
F (Ref. 1, p. 139) = 3 (Conservative Estimate)
h actual =3x0.7=2.1
P
+ air hp
Required total hp =
1.0
at 2500 RPM
A 5.0 hp (at 3600 rpm) gasoline engine was selected to
drive the blower through a 1:1 V-belt system. Thus, 3.47 hp
is nominally available at 2500 rpm which allows a 12% margin
for drive losses.
DEMONSTRATION TESTS
The sorbent characteristics (19 mm cubes with density range
from 0.13 to 0.24 g/cc) are not sufficiently similar to the
sorbents used in Reference 1 to allow complete confidence in the
design method. Hence, a demonstration test series was performed
to validate the system's performance and establish its charac-
teristics and operational limitations as part of the design
process.
The components identified in the preliminary design were
assembled and instrumented at HYDRONAUTICS, Incorporated. Sor-
bent cubes (19 mm) having two different densities were used in
the tests:
Type B at 0.156 g/cc.
Type G at 0.197 g/cc.
Measurements included the torque, T, and speed, n, of the
engine; static head at the blower outlet, pi; static head
developed by the blower, SH, i.e., between the outlet and inlet;
the dynamic head, q = 'spV 2; and sorbent particle velocities.
Independent variables in the test series were the sorbent type,
sorbent flow rate (from 0 to 0.78 kg/sec), and various duct/
nozzle configurations.
57
-------
Blower Characteristics—
The normalized blower characteristics were obtained by
running the system at several (engine) throttle settings and
varying (throttling) the discharge airflow. The results are
shown in Figure A-2 and compared with the manufacturer's values.
The data for the ratio of static to dynamic head with no sorbent
are self-consistent, but indicate that the measured values of
static head (SH) are low compared with published values. Of
course, the placement of the inlet and outlet static taps was
arbitrary and inflow conditions were far from the ideal con-
ditions normally used to establish blower curves.
The blower efficiency curve has some scatter which may be
attributed to inaccuracies in the measurement of engine torque,
T.
Sorbent Velocity--
The sorbent particle velocity through a transparent section
of the transport duct was deduced from slow-motion pictures
(400 frames/sec). The results are shown in Figures A-3 and A-4
for isolated particles (unloaded) and for randomly selected par-
ticles when loaded, respectively. The open symbols in Figure A-3
show the mean velocity of seven isolated particle cubes, Vc, at
each of three air speeds, Va = 14, 18 and 23.5 m/sec. The test
at the lowest speed was repeated for both sorbent types. The
closed symbols give the highest speed among the seven particles
in each test run. The error bars on the open symbols indicate
the standard deviation ± a) of the measured particle velocity
about the mean.
The symbols in Figure A-4 are similar, but approximately 25
particles among a population of several hundred were randomly
selected to obtain the mean particle velocity during a test run
with sorbent loading. The maximum particle velocity was
obtained from inspection of the entire population.
The air velocity in the unloaded case was derived from the
dynamic head measured by a pitot probe placed in the transport
duct. The data in Figure A-3 show that a sorbent particle will
occasionally approach the air velocity in the duct, but the
mean velocity of isolated cubes is lower than the air velocity
by a constant increment or slip velocity as shown by the dashed
line through the data and parallel to the solid line for V =
Va- The slip velocity is nominally 8.5 m/sec for both sorbent
types.
58
-------
e
•
CM
0.6
0.4
0.2
UNLOADED- O NO SCR8ENT (f = 1 .0)
_ ( Q TYPE 8
I 4
LOADED
TYPEG
0.8
1.0
1.2
X
09
d
V
*">
>"
II
0.8
0.6
0.4
0.2
/OPERATING POINT FOR
/ UNLOADED SYSTEM AT
FULL THROTTLE
LITERATURE
FOR BLOWER
( NO INLET
LOSSES)
0.8
1.0
^ 2gSH /n«D
1.2
Figure A-2. - Blower characteristic curves.
59
-------
30
20
\
E
10
O TYPE B SORBENT CUBES
ATYPE G SORBENT CUBES
Filled in symbols represent maximum value
Open symbols represent average value
XRepresent! twice the standard deviation
10
20
30
V , m /sec
a
Figure A-3. - Velocity of single particles (unloaded).
60
-------
20
10
Q TYPEBSORBENT
A TYPE G SORBENT
Filled in symbols represent maximum value
Open symbols represent average value
I Represents twice the standard deviation
^-DESIGN POINT V&OCITY
10 20
V , m /sec
30
Figure A-4. - Particle Velocities (loaded, full throttle).
61
-------
Loaded Air Velocity—
The air velocity in the loaded case was deduced from the
blower characteristic curve (Figure A-2) , because the pitot probe
could interfere with the flow of sorbent particles and its
operation would, in turn, be affected by the cubes. The compu-
tation was based on the assumption that the ratio of effective
dynamic head, fq, to static head is a function only of the
static head normalized by the blower speed, where f = 1 + Wc/Wa
and q = Va2/2g (note, SH is in meters of air). The results are
shown in the following table.
TABLE A-l. LOADED AIR VELOCITY
(Deduced from Blower Characteristic Curve)*
Sorbent
Type
B
G
None
w
c
Kg/sec
0.39
0.52
0.78
0.26
0.39
0.52
0.0
Va
m/sec
19.1
17.2
15.2
19.4
17.6
14.6
27
VWa
0.54
0.80
1.35
0.35
0.58
0.94
0.0
f
1.54
1.80
2.35
1.35
1.58
1.94
1.0
n, rps
35.8
36.7
32.3
39.7
36.5
34.7
47.3
Hp**
2.78
2.85
2.51
3.08
2.83
2.69
3.67
*Assumed that -Jfq/SH vs -^2g SH/mrD is the same for loaded
and unloaded condition.
**Engine torque is nearly constant ~6.8 ft-lbs (horsepower
is given in English units) .
The blower efficiency based on the measured static head
and the combined weight of air and sorbent flow is seen in
Figure A-2 to be reduced by a factor of about 0.8 relative to
the unloaded (air only) case.
The deduced air velocities appear to be consistent with the
measured maximum particle velocities in Figure A-4. The mean
sorbent velocity seems nearly independent of the air velocity,
62
-------
in the limited range of these tests, however; and the heavier
Type G sorbent moved decidedly slower, Vc ~ 6 m/sec compared to
~10 m/sec for Type B.
System Losses—
The effects of the sorbent loading on the head losses are
shown in Figure A-5, where pj/q is the normalized head loss in
the duct and nozzle and (SH-Pl)/q represents the losses on the in-
let side of the blower. Although there is considerable scatter
in the data, the inlet losses are relatively independent of the
loading factor, f. The duct and nozzle losses increase with
increased sorbent loading, especially with the Type G foam.
When these losses are normalized by the effective dynamic head,
fq, the resulting loss coefficient is nearly constant for the
Type G sorbent, but decreases with loading factor for Type B.
The higher losses for Type G may be cavsed by their greater
density and/or their greater resiliency which seems to cause a
noticeable increase in the number of collisions with the duct
walls; hence, the lower mean velocities that were noted earlier.
The nozzle contributes about 60 percent of pl for the unloaded
system. Hence, the nozzle loss coefficient is about 0.8.
Limiting Factors—
For both types of sorbent the limiting factor for sorbent
loading was the available torque of the 5-hp gasoline engine.
The maximum sorbent loadings that could be condinuously main-
tained were:
Type B = 0.78 Kg/sec or 5.0 liter sec.
Type G = 0.52 Kg/sec or 2.6 liter/sec.
The maximum transport rate was not measurably affected by the
nozzle or even small radius 180°-bends in the transport ducting.
Greater rates could be sustained for short (^2-second) intervals
if the mean loading rate was not increased.
The minimum air velocity for satisfactory sorbent transport
was nominally 15 m/sec.
The mechanism by which the sorbent transport failed may be
described as follows. As sorbent feed rate increased, the
engine speed decreased (at nearly constant torque) until the
63
-------
1.0
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
3.0
2.0'
1.0
f-(1 * W /W }
c a
I
I
I
I I
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
f
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4
Figure A-5. - Pressure losses in the system as a function of
the sorbent loading coefficient, f.
64
-------
air velocity in the duct was insufficient to maintain sorbent
transport. The duct would fill up and become clogged and/or
the engine would stall. Sorbent lying in the blower, hopper,
and/or duct was readily cleared when the engine was restarted
except when the duct was fully clogged, i.e., a short section
of duct was full from top to bottom so that the flow of air was
completely shut off.
Final Design
The only major change to the preliminary design of the
sorbent transport/broadcasting system arising from the
demonstration tests was to substitute a 7-hp gasoline engine
for the 5-hp drive. The design point performance, 0.52 Kg/sec,
was achieved by the system as tested with Type G sorbent and
exceeded by 50% with Type B sorbent. Since the only limiting
factor in any of the tests was attributable to the available
engine torque, we feel that the added weight (4.5 kg) to the
system is justified by the expected increase in performance
(up to 1.1 Kg/sec) capability. Estimated weight of the final
system design follows.
Dayton No. 4C131 Blower plus hopper
and duct flange 39 kg
7 hp 4-cycle Briggs and Stratton
gasoline engine with vertical
shaft 15 kg
Base for engine and blower 23 kg
Total Blower Assembly 77 kg (170 Ibs)
Nozzle 4 kg
7.3 meters of 8-in. dia. stove pipe
and "flexflyte" hose 2_1 kg
Total System 102 kg (225 Ibs)
Reference:
1. Miller, E; Stephens, L.; and Ricklis, J. (January 1973.)
Development and Preliminary Design of a Sorbent-Oil
Recovery System. U. S. Environmental Protection Agency
Report No. EPA-R2-73-166.
65
-------
TbCHNICAL HEPORT DATA
(flease read laaruciions on ilie revent before completing)
\ HLPOR r NO.
_EPA-600/7-78-2T7
4 Tl 1 Lt AND SUBTll LE
Development of a Sorbent Distribution and Recovery
System
3 RECIPIENT'S ACCESSION NO.
5 REPORT DATE
November 1978 issuing date
6. PERFORMING ORGANIZATION CODE
7 AUTMOHIS)
Sidney H. Shaw, Richard P. Bishop and
Robert J. Powers
8. PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Seaward International, Inc.
6269 Leesburg Pike, Suite 204
Falls Church, Virginia 22044
10 PROGRAM ELEMENT NO.
EHE623
11. CONTRACT/GRANT NO.
68-03-2138
12 SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Lab.-Cinn, OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report describes the design, fabrication, and test of a prototype system for the
recovery of spilled oil from the surface of river, estuarine, and harbor waters. The
system utilizes an open cell polyurethane foam in small cubes to absorb the floating
oil. The system is highly mobile and can be transported in two pickup trucks.
The sorbent is transported and distributed over the water surface by means of a
pneumatic broadcaster. An inclined, open wire mesh belt conveyor is used to remove
the saturated sorbent from the water. The recovered oil and water is removed from
the sorbent by squeezing in a converging belt press or regenerator. After regenera-
tion, the foam is reapplied to the oil slick. The foam can be reused for a great
many cycles.
Tests of the system, using both diesel fuel and lubricating oil, were conducted at
EPA's OHMSETT facility. The sweep speeds ranged up to 5 knots in both calm water
and waves. Oil collection rates of 10.5 m3/h were achieved. The oil content of the
recovered liquid varied from 38% to 79%.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Oils
Skimmers
Water pollution
Performance tests
Estuaries
Harbors
b. IDENTIFIERS/OPEN ENDED TERMS
Equipment development
Oil spill cleanup
COSATI Field/Group
68D
DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
76
20. SECURITY CLASS {This page J
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
66 . U. $. GOVERNMENT PIINTINC OfFICE: 1978-657-060/1529 Region No. 5-11
------- |