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

-------