EPA-670/2-73-068
September 1973
OIL RECOVERY SYSTEM USING SORBENT MATERIAL
By
Garth D. Gumtz
Thomas P. Meloy
Project 15080 HET
Program Element 1BB041
Project Officer
Kurt Jackobson
Agricultural and Spills Branch
Environmental Protection Agency
Washington, D.C. 20460
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
-------
EPA Re-view Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents
necessarily reflect the views and policies of the
Environmental Protection Agency, nor does mention
of trade names or commercial products constitute
endorsement or recommendation for use.
ii
-------
ABSTRACT
Using a combination of dimensional analysis and ;mathemat-
ical analysis as well as experiments, a mathematical
model of oil sorbent systems has been created which de-
scribes the efficiency and capacity of virtually any oil
sorbency system. Following the analytical experimental
stage, a scale model system was set up and run in a wave
tank; the results of which substantiate the conclusions
from the analysis. Because of the low efficiency of air
broadcasting of sorbents, an alternate dispersal system
was built and tested.
One inch.cubes of recticulated foam which rapidly saturate
with oil, are readily picked up and the oil recovered from
the foam sorbents. Broadcasting and systems losses are
low. Capacities of 3,500 to 14,000 GHP are readily achiev-
able at 2 knots per hour within the wave and wind conditions
specified. Operating and capital costs of the system are
$59,500 per year and $216,500; yielding a cost of 6<: per
gallon recovered, including depreciation. Because of the
system's high practicality, simplicity, reliability and
economy, it is recommended that a full scale system be
built and put into service.
iii
-------
CONTENTS
Section Pag e
I Conclusions 1
II Recommendations 3
III Introduction ' 5
IV Basic Sorption Properties 7
Introduction 7
Dimensional Analysis 9
Particle Size and Shape 12
Pore Size 13
Slick Thickness 15
Oil Viscosity and Surface Tension 18
Sorption to Less Than Saturation 20
Conclusions 22
V Sorbent Broadcasting . 23
Introduction 23
A Random Model 23
Dimensional Analysis 26
Laboratory Tests 27
Broadcast Velocity 31
Wind Velocity 34
Sorbent Oiliness 35
Sorbent Size and Shape 35
Broadcast Equation 36
Air Blown or Gausian vs Rectilinear
Broadcast System 36
Behavior of Sorbent in Slick 39
Rectilinear Broadcast System 39
Conclusions 40
VI Sorbent Herding 41
Description of System 41
Boom Design problems 42
Hydrodynamics of Slick Flow in Channel 44
Hydrodynamics of Sorbent Flow -au 47
Choking of Sorbent in Channel 48
Contact Time 48
Oil Pickup Rate 48
Sorbent Loss 50
Conclusions 50
VII Sorbent Pickup 53
Sorbent and Oil Losses 53
Mechanical Constraints 54
-------
CONTENTS
Section Page
VII Conclusions 54
VIII Recovery of Oil from Sorbent 55
Introduction 55
Sorbent Cycle 55
Sorbent/Oil Recovery 57
Alternatives 60
Conclusions 61
IX Total System Concept 63
Introduction 63
Systems Specifications 67
Broadcasting • 68
Herding 74
Pickup and Delivery to Recovery Unit 80
Oil/Sorbent Recovery 86
Recovery-to-Broadcast and Aft-to-Fore
Transport 92
Sorbent Selection 92
Conclusions 96
X Vessel Requirements 97
Introduction 97
Available Ships 97
Operations 98
Deployment of Recovery System 98
Oil Storage 99
Conclusions 99
XI Preliminary Performance Estimates 101
Introduction 101
Gross Oil Pickup 101
Oil Throughput 102
Oil Type 102
Sorbent Losses and Degradation 103
Preliminary Cost Estimates 103
CapitaI9Eoats 104
Operating Costs 104
Conclusions 105
XII Appendices 107
Appendix A 109
Appendix B 157
vi
-------
FIGURES
No. Page
1 Sorbent Geometry 14
2 Pore Size ve Time 16
3 Slick Thickness vs Time 17
4 Sorption Time vs Viscosity 19
5 Dimensionless Sorption Time 21
6 Percent Random Coverage vs Surface
Area of Broadcast Sorbent 25
7' Blower/Injector Subsystem 32
8 Nozzle/Diffuser Subsystem 33
9 Relationship between Slick Velocity
in Channel and Boom Geometry 45
10 Relationship between Oil Pile-up and
Current Velocity 46
11 Pilot (%) Scale Spilled Oil Recovery
System 64
12 Sorbent Oil Recovery System Layout 70-71
13 Sorbent Broadcast System 72-73
14 Herding Boom-Support Arm Interface 76-77
15 Herding Boom-Support Arm 78-79
16 Herding Boom-Pickup Conveyor Interface 82-83
.... .tir -.
17 Sorbent Pickup Conveyor & Delivery Conveyor 84-85
18 Sorbent Pickup Conveyor Support 88-89
19 Oil/Sorbent Recovery Unit 90-91
20 Transfer Conveyor 94-95
vii
-------
TABLES
No. Page
1 Sorbent Broadcast Data 28-30
2 Urethane Foam Sorbent Recycling. 56
3 Oil Removal as a Function of Conveyor
Belt Speed 58
4 Oil Removal as a Function of Sorbent
Size for Several Oils 59
5 Oil Removal as a Function of Wringer
Belt Compression 60
6 Sorption Wave Tank Test Data 65-66
viii
-------
SECTION I
CONCLUSIONS
Summary
It has been said that a camel is a horse designed by a
committee. Air broadcast oil sorbency systems fit well
that description. Through careful experimental theoretical
work, we have shown that the conventional concept of air
dispersal of sorbents on oil and then picking them up is
not effective and in its place we have recommended an
efficient and more durable system.
The System
Recommended in this report is a system which consists of a
shrouded screw fed gravity drop sorbent dispersing system.
The sorbent is dropped on the slick in front of the narrows
of the recovery channel. Slick velocity in the channel is
slowed in relation to the free stream velocity by the
presence of oil and sorbents. In the channel the slick
deepens and under certain conditions will build a head
wave. Under certain conditions of high sorbency flow, the
channel chokes resulting in sorbents and oil losses under
the boom.
With proper operating and design conditions the sorbent has
ample time to absorb the oil slick and is readily recovered
by a wire mesh belt lifting the oil saturated sorbent to
the deck of the ship. Very little oil passes through the
recovery system back onto open water.
Once onboard the oil is recovered from the sorbents by
passing the oil saturated sorbent cubes between two belts
through rollers. After recovering 75% of the oil from the
sorbent, the reuseable sorbent is transported forward and
rebroadcast on the slick. The chunky reticulated foam
particles can be used repeatedly without appreciable de-
gradation.
Conclusions
In the study the following conclusions were reached:
(1) For durability and efficacy, 1 inch cubes are a
very effective size.
(2) Plate-like foam particles are less effective.
-------
(3) Air broadcast nozzles give non uniform distribution
of sorbent, air broadcast particles are subject to
wind velocity and gusting scatter, and the system has
inherently high losses.
(4) A shrouded rectilinear screw fed distribution system
eliminates all of the problems mentioned in number
3, above.
(5) Foam particles rapidly saturate with oil and sorption
time is usually not a problem.
(6) A four-to-one compression of the slick by an angled
boom is desirable.
(7) Bridging or choking can occur in the channel under
improper operating conditions.
(8) Oil recovery works well with a dual belt wringer
sys tern.
(9) Capacity of 3500 to 14000 gallons per hour can readily
be achieved with no. 4 fuel oil and marine diesel fuel
oil.
(10) Proposed design will work effectively with a variety
of sorbent material and shapes.
(11) Commercial polyurethane foam - ester-type - with 100
pores per linear inch, cut in one inch cubes is
suggested with 20 PPI foam as an option for heavier oils
(12) Wave action does not materially alter system behavior.
(13) Oil-sorbent-systern ship requirements are met by a
number of ships in the private and public domain.
(14) Capital and operating costs of the system are re-
spectively, $216,500 and $59,500 per year.
(15) Recovery cost including capital depreciation are 6c
per gallon.
(16) Salvage value of recovered oil could actually pay for
capital and operating costs of the system depending
on spill frequency and spill thickness.
-------
SECTION II
RECOMMENDATIONS
Initially the authors undertook thi.s project with a strong
conviction that a sorbency recovery system was both in-
feasible and uneconomical. It was believed the h.igh loss
rates and long sorbency times would preclude this system
ever being designed effectively. The proposal was designed
so that the systems analysis would preclude money being
spent for full scale system.
While initial studies on sorbency rates were encouraging
the definitive work on air broadcasting of sorbents showed
that uniform coverage would be poor and loss rates high.
The invention of the shrouded-rectilinear-screw-fed-
distribution-system eliminated the pessimism on the fea-
sibility of the sorbency oil recovery system. It now can be
said unequivocally that the system is economical and tech-
nically feasible.
It is recommended that:
(1) The full scale system proposed in this report be de-
signed, built and field operated.
(2) The approach delineated in this report be advertised
through papers, presentations and other information
channels to the oil industry as a cost effective as
well as technically effective method of recovering of
oil in thin slicks. The system delineated in this
paper shows not only will it be cheaper than straw for
oil recovery, but far cheaper than straw as a system.
-------
SECTION III
INTRODUCTION
Spillage of oil on estuarian and marine waters has become
an international problem. Standard operating procedures,
to date, call for dumping straw on the slick and sub-
sequently recovering the oil/straw sludge. Subsequent
disposal problems of the straw sludge as well as the vast
quantity of straw used has raised the question of the
possibility of using recyclable sorbents. Furthermore,
slicks often become quite thin before recovery operations
can be mounted. Conventional methods such as skimmers
and straw are ineffective on these thin slicks.
This study was made to determine if it was possible to
design and operate an oil recovery system which broadcast
oil absorbent, subsequently collected the sorbent, re-
covered the oil and reused the sorbent. This study in-
dicates that such a system is feasible but not in the
manner originally conceived.
The main parameters to be considered in such a study are:
(1) Is the sorbency absorption rate great enough
to pick up the oil in the time the sorbent
is on the slick?
(2) Can the sorbent be broadcast onto the slick
with a uniform pattern without suffering
prohibitive losses?
(3) Can the slick with the sorbent in it be com-
pressed and herded to a pickup channel without
undue loss of either sorbent or oil?
(4) Can the sorbent-oil complex be recovered from
the surface of the sea without undue loss of
either oil or sorbent?
(5) Once recovered, can the oil be recovered from
the sorbent in sufficient quantity to make the
system practical? Further, does the sorbent
have a sufficient lifetime in recycling through
the recovery system to make it economically
feasible when compared to the standard sorbent -
straw?
-------
The method of approach in each of the areas .was to set up
a dimensional analysis of the system and then run experi-
ments to fill in the system parameters. This approach was
successful and has resulted in the development of a number
of equations that describe various operations in the oil-
sorbent system over a wide range of variables and oper-
ating conditions. These tests were run in' small to large
containers and the system was proof tested in a wave tank.
In the following sections, the experiments and dimensional
analysis techniques are delineated.
-------
SECTION IV
BASIC SORPTION PROPERTIES
For an engineering study, the characteristics of a process
material must be specified so as to avoid major errors in
the design of a full scale system. In a scientific study,
these characteristics would themselves be of primary im-
portance; here, however, the final system is the primary
aim, and, therefore, a semiquantitative or qualitative
grasp of sorption characteristics can be entirely suffi-
cent. Most sorbent properties were characterized in a
quantitative fashion during the study, but, after com-
pletion of the wave tank tests, semiquantitative or
qualitative relations were sought for reasons described -in
detail later on in this report. Basically, factors pe-
culiar to the system as a whole (as compared to the sorbent
as a distinct entity) were fo'und to be of such overwhelming
importance that extreme care in defining sorbent properties
became wasteful.
Introduction
A considerable abundance of physical properties serve to
describe the sorption process, even in an extremely sim-
plified situation. If we consider a single sorbent
particle placed onto a quiescent oil slick of infinite
extent, the sorption process would seem ideal enough to
permit "complete" description. Even here, however, the
following relatively large family of parameters must be
considered:
t = the time over which sorption has occurred
t = the time necessary for the sorbent particle
s to saturate with oil
T = the fraction of the saturated oil volume or
mass sorbed at time t
V = the bulk volume of the sorbent particle or
one of the factors characterizing shape as
illustrated with the data presented below
C = the wetted perimeter of the sorbent particle
or the second factor which characterizes shape
-------
d = the average pore diameter of the sorbent foam
(note that in the form 1/d this parameter
represents the average pores per unit length,
a parameter commonly in use in the foam pro-
duction industry)
e = the void fraction for the sorbent foam
H = the fraction of the sorbent void which is
open celled
g = the acceleration due to gravity
1 = slick thickness
yo = oil viscosity
OQ = oil/air interfacial surface tension
p = oil density
6Q = contact angle between oil and sorbent matrix
9OW = oil/water interfacial surface tension
yw = water viscosity
aw = water/air interfacial surface tension
pw = water density
0W = contact angle between water and sorbent matrix
T = temperature of the entire system
Now, this set of parameters, although 20 in number, does,
indeed represent an ideal or much simplified system. No
attempt has been made (in this particular development, not
in the study as a whole) to include the effects wind, waves
or particle-particle interactions on the sorption process.
Obviously, effects of oil aging (e.g., the formation of
oil/water dispersions and the infall of particulates on a
slick), velocity of impact of sorbent particles on a slick,
buffeting of sorbent by herding boom a|nd conveyor pickup,
and reuse of sorbent are also not included in this analysis,
Careful consideration of an entire sorbent oil recovery
system, indicates that there are many more variable;; which
are not being accounted for in the present analysis. The
-------
above approach does, however, account for essentially
all the relevant materials parameters (not what may be
called the systems parameters or those conditions imposed
on the materials by the environment within which they
operate in a particular instance). This development will,
therefore, provide only indications of the effect and im-
portance of the various materials parameters for the system
(general approach) under investigation.
Dimensional Analysis
Recognizing that temperature may be expressed in units of
kT where k is Boltzmann's constant and is, therefore, a
measure of energy, the 20 parameters of the previous sec-
tion can be expressed in terms of 3 basic dimensions (mass,
length, and time). There are, therefore, 17 dimensionless
parameters which describe the system in question. This is
jar too many to be tractable in an engineering study of
limited scope. Three rationales may be used in reducing
the number of parameters which have to be considered in
this study.
First, implicit dependence is apparent for the various
materials properties on temperature; that is, if the 19
other parameters are functions of temperature and are
given quantitatively at a specific temperature, then no
explicit dependence is needed for temperature in the
dimensional analysis. 19 parameters, therefore, remain.
Second, certain parameters may effectively and practically
be considered as constants. The void fraction or porosity
(inherently a dimensionless number) will be very close to
one for most foam materials of interest, and, furthermore,
if a family of sorbents is selected which is geometrically
similar, e is a constant by definition. A similar argument
can be made for n, the open cell fraction of the voids;
furthermore, n should be set at a value of one (for
maximum oil pickup capacity) unless unforeseen factors
mitigate drastically against same. The contact angles
(00 and 9W) as expressed in radians (a dimensionless format)
can also be assumed to be constant for "normal" oil/water
systems. Since economic considerations demand that the
sorbent foam be reused, the sorbent should be prewetted
with oil, and, therefore, 9o = ir and 6W=0. Other parameters
in the main set may also be considered as constants, but
this is better done in light of the discussipn in the
following paragraph. With the above 4 parameters set
constant, 15 remain for consideration in the dimensional
analysis.
-------
Third, certain of the remaining parameters may be judged to
produce only second-order effects on the sorption process.
For instance, the relative effect due to the yigcosity of
water should he small since 1) the oil is the material which
undergoes a primary bulk motion and is sorbed and 2) the
viscosity of water is a constant and can eventually be com-
bined with other constant factors Cset parameters) resulting
in. a dimensionless parameter which is itself a constant.
Changes of viscosity due to temperature are much less
critical for water than for oils (although of the same
general form); since the effect of water viscosity is
second-order, the effect of temperature on this parameter
is of third-order and, therefore, of little concern.
Densities for both water and oils are little different in a
fairly general sense and are somewhat weak functions of
temperature in the range of interest CO to 40°C). Therefore,
the domensionless parameter pw/p0 can be reasonably postu-
lated to have little effect on the sorption phenomena
realizing that, even for an entirely gravitational (Froude)
effect, the geometry of a slick due to thinning around a
sorbent particle and the effect of oil viscosity on flow
to a particle are overriding. Likewise, surface tensions
are 1) relatively weak functions of temperature, 2) es-
sentially constant for water, and 3) are of the same general
magnitude for oils of, .a "similar" nature (see the section
below on viscosity and surface effects; this is the case for
both ao and oow) . Therefore, in the light of previous con-
siderations, the sorption process may be assumed to be a
relatively weak function of both the surface behavior of
the supporting water phase (aw or aw/ao) and the inter-
action of the oil and water phases (aow or aow/ao). In
this paragraph 4 parameters (yw, Pw» ^w and aOw) have been
shown to be of secondary consequence to the sorption pro-
cess; 11 important parameters remain for inclusion in the
following dimensional analysis.
Since the 11 parameters are expressed in terms of 3 basic
dimensions, a set of 8 dimensional parameters will serve to
describe the sorption process. The following set was
selected (and eventually modified) for study:
3 ' 2
-
_ _
ts c i vi -2- ~2
P0d g 8d Po
Still assuming an implicit temperature dependence, the
following dimensionless parameters must be kept in mind as
possibly effecting the sorption phenomena if they deviate
substantially from the assumptions posited above:
10
-------
fl
W»
°pw
Now, among the 8 dimensionless parameters relevant to this
study T goes from 0 to 1.0 as t/ts does the same. That is,
T, the dependent variable, depends on t/ts, an independent
variable, within a closed (finite) space whose extent is
determined entirely by the magnitude of ts. Since modeling
of the dependence of T on t/ts is relatively straightfor-
ward (as a two stage process: 1) sinking of the sorbent
into an infinitely thick slick followed by 2) transfer of
oil from a slick of infinite extent but finite thickness
to the sorbent "sink"), basic engineering laboratory
studies may just involve determining the dependence of
gt /C on the remaining independent dimensionless variables.
s
In general, the following functional relationship is to be
determined experimentally,
. . » * » UQ ,
C 1 V 1 23
P0d 8
but, based on the assertions of the previous paragraph, of
primary, initial importance is determination of the func-
tional relationship,
3 2
= h[ C , C , d , yo ,
V 1 23
0
Pd
Classically in engineering stud'ies, such a functional re-
lation is assumed to be separable into power relations for
the various independent parameters, or
That separability is an assumption must be kept in mind
since the general equation implies that the superscripts
in the separated equation may be functions of the in-\
dependent variables; experiment or appropriate modeling may
determine such functional relationships.
11
-------
Particle Size and Shape
Characterization of particles, given the present state of
the art, is best done as related to particle use. Size and
shape are abstract concepts, not amenable to quantification
in a general sense. Fundamentally, the use of particles to
sorb oil involves two critical factors related to particle
geometry: oil sorption capacity and transfer of oil from
slick to particle. Given these factors, the following
postulates account for the effects of size and shape. First,
sorption capacity relates directly to the volume of the
sorbent particle (more exactly to n e V) that is, for the
purpose of oil sorption V may be considered as analogous to
particle size. Second, transfer of oil from slick to par-
ticle depends (especially for the case of a relatively thin
slick) on the wetted perimeter available for movement of oil
into the sorbent; therefore, C may be used as a measure of
shape for the process. Note, however, that other parameters
may also be important in characterizing particle shape.
For example, the area of the particle which initially con-
tacts the slick may relate to the initial stage of the
sorption process while the height of the particle could be
important in describing buoyancy effects; however, for an
engineering study of a limited number of sorbent config-
urations, use of V and C should be adequate which can, of
course, be checked to a limited degree by experiment. The
possibility must be kept in mirid that, in particular situ-
ations, V and C may not be sufficient to describe sorbent
geometry particle as related to the sorption phenomena;
the following pages illustrate that for the purposes of this
study, however, these two parameters are adequate.
The plot on the following page summarizes the results
of more than 100 sorption tests. Sorbent particles (fully
reticulated polyurethane foam with 100 pores per inch)
were dropped onto a 1.5 millimeter slick of No. 4 fuel oil
in a 10 foot diameter tank. For the basic sorption tests
(preoiled sorbent), 6 different sorbent sizes and shapes
were used; considering the averaged data points on the plot
from left to right along the abscissa, the geometries were:
(1) 1/2 x 1/2 x 1/2 inch rectangular parallelpipeds
(2) 4 x 1/2 x 1/2
(3) 1 x 1 x 1/2
(4) Ixlxl
inch rectangular parallelpipeds
inch rectangular parallelpipeds
inch rectangular parallelpipeds
12
-------
(5) 2 x 2 x 1/2 inch rectangular parallelpipeds
(6) 2x2x1 inch rectangular parallelpipeds
The form of the relation between reduced saturation time
and reduced sorbent geometry was determined by performing
repetitive least squares fits on the two dimensionless
parameters in a log-log format; the curve fit pertains to
only 5 of the data collections. The second sorbent
geometry above (or the blackened circle on the plot) was
not considered in the curve fitting procedure but is
plotted for the purpose of comparison to the data which was
This comparison is very good (in fact, excellent when the
geometry of the 4 x 1/2 x 1/2 inch strands is compared to
the blocky nature of the other five sorbent geometries).
Figure 1 illustrates the following relation for saturation
time as a function of sorbent geometry,
-3.56
where D is a constant whose value depends on all the
systems parameters other than geometry. Note that satu-
ration time depends approximately on the inverse of the
wetted perimeter; this agrees well with the mass transfer
postulate made above. However, saturation time does not
vary in direct proportion to the volume of the sorbent
particle as might have been expected; the variation is with
the 1.78 power of the volume instead. Several factors
probably account for the added volume dependence: 1) the
length through which oil must move into a particle by
capillary flow effects the rate of the sorption process,
2) the buoyancy or rate of settling into the oil/water
system depends on the height of the particle, and 3)
(especially for relatively thick slicks and small particles)
the initial area of contact between the particle and the
slick should be of importance. Since parameters related to
these three effects are significantly different only for the
4 x 1/2 x 1/2 inch strands and since the experimental results
for the strands compare very favorably with prediction, the
conclusion must be that for the purposes of an engineering
study the reduced sorbent geometry dependence as illustrated
in Figure 1 is adequate.
Pore Size
Although of possible operational importance, pore size is
13
-------
10C
10
(M
J*
bfi
I
•l-l
H
§
0)
U
-3
0)
10
10"
10
MBJ.
10
10
Reduced Sorbent Geometry, (C/l)7< 35 (C3/V)~3> 56
Figure 1
10
14
-------
not a very crucial parameter in the sorption process.
Figure 2 is a log-log plot of reduced sorption time (to
saturation) versus reduced pore size in which the former
parameter includes an "abstracting out" of the effects of
sorbent particle size and shape. The dependence of
saturation time on pore size is relatively weak. The form
of this dependence (according to a least squares fit) is
as follows,
(g/C)1/2(C3V)1'78(C/l)~3'68t0 = E (d/l)~°-337
s
where E is a constant whose value depends on all the
systems parameters other than sorbent geometry and pore
size. The fractional power dependence can be explained as
a consequence of 1) sorption dep-endence on factors other
than capillary flow of oil within the sorbent and 2) anN
actual net like structure of the sorbent as opposed to
capillaries.
Slick Thickness
Oil sorption depends critically upon the thickness of the
slick upon which it occurs. Figure 3 summarizes the
results of a large number of sorption tests with slicks of
varying thickness. Reduced sorption time has been defined
for the purposes of this plot so as to "abstract out" the
effects of particle size, shape and pore size. The plot
recognizes the fact that, slick as thickness becomes very
large, sorption dependence on thickness disappears; that
is, the sorbent behaves as if the slick were infinitely
thick. The data bears out this fact very well. Note the
steeply sloped straight line on the plot whose dependence
on 1 cancels out the thickness dependence of the reduced
sorption time (see the ordinate). There are obviously
two regions of interest for the phenomena of sorption:
zones in which 1) dependence on slick thickness is severe
(the horizontal line on the left hand size of Figure 3)
and 2) the sorbent interacts with the slick as if it were
infinite (the line with a 3_. 34 slope on the right of the
Figure). Note that a least squares fit to the five points
on the left of Figure 3 leads to the relation illustrated
by the broken line; there is, however, enough scatter in
the data to make the horizontal approximation sufficient
for the purposes of an engineering study. The functional
equation for the sorption time now takes the form,
3.34 (g/c)l/2(C3/v)1.78(c/d)-3.68t =
15
-------
oo
CD
00
eo
•SS
0*
s
H
o
•i-H
-*->
&
O
CO
-8
13
0
K
7 _
3 .
2.5
0.125
O
O
0.25 0.5
Reduced Pore Size, d/l
Figure 2
1.0
2.0
16
-------
o
73
0)
200
I I I
X
CQ
oo
co
CO
I
100
u
oo
t-
60
40
3.34 -
20
I
a
t-i
o
ra
10
4
o;
2 0.3
I
I I
o
I
0.5 0.75 1.0 1.5
Slick Thickness, 1 (mm)
Figure 3
17
-------
where F is a constant and where for a. sufficiently thick
slick -3mi-2m2=3 . 34 and for a sufficiently thin slick
3mi = -2m2« What constitutes sufficiently thin and thick
slicks will be quantified in the next section; suffice it
to say here that something like 3 millimeters for the
conditions of Figure 3 is a critical slick thickness.
Oil Viscosity and Surface Tension
From fluid mechanics (see any basic text) sorption or flow
rate is known to vary inversely with Newtonian viscosity
for both capillary and sheet flow; sorption time should,
therefore, vary in direct proportion to viscosity. Ad-
mittedly, oils are, in general, not Newtonian liquids; for
the purposes of an engineering study, however, they may
reasonably be assumed Newtonian. Figure 4 is a plot of
reduced sorption time versus reduced viscosity; note that
the reduced viscosity is expressed in terms of pore size,
as opposed to slick thickness, in this plot. The direct
proportionality between sorption time and viscosity is
good although a least squares fit would give a log-log
proportionality slightly different than one.
Given the results of the previous section and paragraph,
the equation describing the sorption
becomes
process to saturation
0.50
0. 75+1.670
where
a = 0 when 7T < 0.329
and
for
0 = 1, F,
0.50
0.329
2.42
andT(lc) = 0.329
and FQ = 3.77 x 10°
This value for F0 was obtained from the extensive data
gathered during the studies of the effects of sorbent size
18
-------
10
i r i
3
«
I
EH
I
&
O
O
-3
o>
4
2
10S
4
2
102
10
I I
10-
Reduced Viscosity, (>»/9l3/2g1/2) a/d)3/2
Figure 4
19
-------
and shape and is, therefore, representative of very real-
istic calm water conditions. The relation for sorption
saturation time is relatively complete and can be used as
a base for estimates under other than calm water conditions
given that enough data is available to allow a reestimate
of F0.
Sorption To Less Than Saturation
Less than sorption to saturation has been discussed pre-
viously. Sorption in this more general sense can be
modeled by postulating a two stage process with the con-
dition that only preoiled sorbent be considered. The
latter condition is consistent with the economics for using
artificially produced sorbent materials; that is, since
these sorbents (e.g., polyurethane) must be used many times
(on the order of one hundred) to successfully compete with
straw, the vast majority of particles sorbing oil within a
steady state system will have been effectively preoiled.
The two stage process is as follows:
1) Initial sorption through the thickness of the slick at
a rate equivalent to that estimated for an infinitely thick
slick. The figure on the following page illustrates this
stage which takes place up to a time t*. The total satu-
ration times(both for an infinitely thick slick, tg(l-»-<»),
and for the slick of interest, tg(l) are determined using
the relation presented in the previous section. t* is
obtained as the 1A/V fraction of the saturation time for an
infinitely thick slick. The reduced sorption, T=v(t)/V,
is a function of the reduced sorption time, t/ts, where
v(t) is the volume of oil sorbed at time t; note that v(t*)
is estimated as 1A where A is the area of approach of the
sorbent particle to the slick.
2) The final step in the sorption process involves mass
transfer of oil from the slick proper to the sorbent
particle. This stage is modeled Dimply by making, a linear-
interpolation between the point [t*, T(t*)| and jl.OO, l.OOJ
on a plot such as that in Figure 5.
Figure 5 contains two examples of typical sorption curves.
Curve A is for a relatively large sorbent particle in a
relatively thin slick; such a system is more typical for
the design approach as conceived in this study. Curve B
is typical of a very thin sorbent particle or of a relatively
thick slick of a very low viscosity oil; such conditions
cannot be deemed realistic from a design standpoint although
they could, of course, occur as an extreme. Note that
20
-------
1.00
n>
a.
c
o
(D
a
o
i-t
O
3
0.75
0.50
0.25
t*
B
t* =
v(t*)/Vf = 1A/V
0.25 0.50
Reduced Sorption Time, t/ts
Figure 5
0.75
1.00
-------
although curve B would superficially appear to represent a
sorption process more favorable than, that of curve A, this
is not necessarily the case. First, since the curves
refer to relative fractions of oil sorption, actual oil
sorption must still be determined or estimated. Second,
factors attendant to sorbent broadcasting, harvesting, re-
covery and transport may weigh against selecting a sorbent
with a more favorable relative oil sorption rate.
Conclusions
A very complete description of the sorption of oil by
polyurethane (ester-type) foams has been developed. Due
to the importance of other design factors; however, this
model cannot be used entirely on its own to specify an
optimum sorbent.
22
-------
SECTION V
SORBENT BROADCASTING
Introduc t ion
As will be shown below, a random model of the broadcasting
of sorbent particles points to severe problems with a
real coverage. That is, large excesses (over and above the
theoretically ideal quantity) of sorbent must be broadcast
to effectively contact and, therefore, sorb oil. Under con-
ditions of randomly broadcasting small sorbent particles
onto a large slick with no sorbent losses, the fraction of
the area effectively contacted (¥) is given as 1 - 10~z Io8 2
where z is the fraction of the ideal amount of sorbent
broadcast. According to this relation 99% effective cov-
erage necessitates the use of about 6.5 times the ideal
anrount of sorbent; obviously, any technique approaching
that of random broadcasting should not be used.
A more ordered approach to the broadcasting of sorbent
materials is required. Sorbent can be broadcast to more
accurately cover a slick in many ways. Ideally, each
sorbent particle should be carefully placed in a predeter-
mined spot in a slick so that maximum sorption occurs; in
terms of system complexity, this approach must be ruled
out. Note that oil sorbing belts provide such exact
placement of/sorbent material. Another technique for
sorbent placement is directing a stream of sorbent onto
the surface of a slick. This particular approach was
subjected to a great deal of attention throughout the
course of the study being discussed here.
The results of any conclusions from studies on air conveying
and broadcasting of sorbents from nozzles are presented in
the following paragraphs. Combined with the results of the
sorption studies and hydrodynamic testing, these results
and conclusions led to the broadcasting subsystem as
presented in the preliminary design section of this report
(Section IX).
A Random Model
The broadcasting subsystem presents some inherent difficul-
ties as regards sorption efficiency. Obviously, sorbent
losses may occur during a broadcasting operation. Further-
more, broadcasting, since it is by nature a relatively un-
controllable operation, will result in a significant
"piling up" of sorbent material; in the case of absorption
23
-------
this would hinder the oil recovery whereas for adsorption
(the mechanical occlusion of oil) it may help. Note,
however, that even for the mechanical interaction of sor-
bent material (the corralling effect) a random model still
has import; that is, it is quite possible that-a significant
amount of the randomly broadcast material will result in a
deposition pattern where the necessary interaction is not at
all probable. Either way, broadcasting is a process which
lends itself quite readily to mathematical modeling, the
subject of the following paragraphs.
Consider the deposition of "n" sorbent units randomly onto
an area which "N" sorbent units could effectively cover;
that is, if placed exactly, "N" sorbent units would be the
exact number necessary to sorb all the oil within the de-
position area. There are,
M = Nn
total ways to place the sorbent units in the deposition area.
Note also that this model assumes randomness in two di-
mensions and rigidity in the third, a not very likely sit-
uation; however, the model is being used as an indicator, not
as a substitute for the real world.
Now, the probability, P^, that only "i" unit areas are
covered when "n" sorbent units are deposited randomly on
a total area of "N" sorbent units is,
1. -Li
J^ f nk + l
P = fri = 1 N I \ (~l)
fr Nn (N-l)! i I
k=l
for i = n or N whichever is less and
where fr. = the number of ways in which just "i" unit
areas can be covered when "n" sorbent units
are deposited on an area of "N" total
sorbent units.
Now, the expected fractional coverage is just.
EC
-E j. ..
N
where m = n or N whichever is less
24
-------
D
U>
K
Ul
o
u
z
Ul
£
K
Ul
Ul
i
IDEM. FACTIONAL
COVERAGE
Figure 6
25
-------
Although expected fractional coverages may be calculated
for a large number of values for "n" and "N", the inter-
esting case is the limit for "N" approaching infinity.
The result of this limiting process is presented in Fig.
6 as a semi-log plot of the expected percent -uncovered
versus the ideal fractional coverage. Note that (n = N)
results in 50% of the area remaining uncovered whereas
(n * 2N) still results in only 75% coverage.
The analysis just presented points to severe problems with
sorbent broadcasting; these are basically: (.1) inherent
and unavoidable inefficiency unless mechanical means are
used to place sorbent on an oil slick in a more determin-
istic fashion; (2) enhancement of the importance of the
basic sorption processes which cannot be overcome to any
great degree by selection of a particular broadcast
technique; and (3) sorbent losses which occur "independent"
of basic considerations of broadcast efficiency and may be
looked upon as the broadcast area outside the enclosed area
considered in the above analysis. Resolution of the above
problems depends entirely on realistic- testing of broad-
cast subsystems plus tank testing to determine the effects
of herding and corralling floating sorbent and oil. Such
testing is considered in the following sections of this
report. For the purpose of emphasis,•without additional
mechanical or physical constraints, iip_ broadcast subsystem
can be more efficient than the one postulated above with a
rectilinear statistical model in which sorbent losses have
been assigned a value of zero.
Dimejcisional Analysis
As was the case for basic sorption properties, reduction
to dimensionless variables vastly simplifies the basic
analysis. The effects of a broadcasting subsystem can be
expressed in terms of 9 parameters (given conditions of
constant temperature and pressure) which, in turn, lead
to 6 dimensionless variables. Assuming separability of
the dimensionless variables for the purpose of an engi-
neering study leads to the following expression,
H
where d. = any of several length parameters which serve
to describe the broadcast pattern including
•width and dispersion
26
-------
H = a lineal measurement describing the size of the
broadcast subsystem; best thought of as the
height of the broadcast discharge from the mean
water line for the purposes of this study
Vd = air velocity at the broadcast discharge
Vw = wind velocity; in general, a vector quantity
although usually directed with the discharge
velocity
g = the acceleration due to gravity
ps = the bulk density of the sorbent particles
pa = the density of air
ya = the viscosity of air
c = length of parameter (in general, an infinite
family of same) describing the geometry of the
sorbent particles
A = a constant for the particular system in
question
2
S = c_ or surface area of a particle for a given
family of particles
Note that for a given family of sorbents both p8/Pa
Cp/H may be considered constant as long as sorbent size is
scaled with subsystem size. Discharge .and wind velocity
and their relevant dimensionless parameters are, therefore,
of primary interest. Consideration of the effects of these
two parameters points to severe problems with the concept
of broadcasting by using an open air carrier stream. These
problems are dealt with in the following paragraphs.
Laboratory Tests
Table 1 gives the compiled results of an extensive series
of broadcasting experiments run in the laboratory. Un-
fortunately, the general problem of broadcasting sorbent
materials is so complex that, these data permit only a semi-
quantitative analysis. As will be seen, however, such an
analysis is sufficient to highlight the difficulties with
pneumatic conveyance as a broadcast technique. Sorbent
size, shape and condition (again using fully reticulated
27
-------
TABLE 1.
SORBENT BROADCAST DATA__
Test
No.
*1
*2
*3
*4
*5
*6
*7
*8
*9
10
11
12
13
14
*15
16
17
*30
*31
*32
Vd
feet/minute
1,050
1,050
1,750
625
660
1,200
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1, 050
1,050
1,050
V (x , y components)
w
(feet/minute)
Particle Size & Condi
0
0
0
0
0
0
0
0
300 (300, 0)
0
300 (300, 0)
300 (150,260)
1,300 (662, 1,125)
1,500 (0, 1,500)
1,500 (0, 1,500)
£,500 (0, 1,500)
0
0
0
0
Ycg
(inches)
tion - 1/2
+40
+39
+79
+21
+17
+65
+43
+44
+37
+24
+23
+38
+33
+84
+ 113
+79
+17
+46
+46
+43
Ysd
(feet/
standard
deviation)
' Cubes, I
0.42
0.44
0.93
0.40
0.36
0.79
0.52
0.47
0.33
0.28
0.30
0.36
0.55
0.73
0.86
1.06
0.24
0.63
0.55
0.55
X
eg
(inches)
ry
-1
+2
+9
+5
+4
+8
+4
-7
-29
++A
-20
+16
+20
-12
-1
-19
+3
+8
+8
+9
x .
sd
(feet/
standard
deviation)
0.22
0.16
0.30
0.18
0. 17
0.26
0.19
0.26
0.41
0.23
0.31
0.26
0.34
0.44
0.47
0.50
0.20
0.21
0.22
0.23
* Diffuser/Deflector Not Used
28
-------
TABLE 1.
SORBENT BROADCAST DATA (Cont.)
Test
No.
33
34
35
*36
37
38
39
*18
19
20
*21
*22
23
24
*25
*26
J
va
feet/minute
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1,050
1, 050
ffuser/Defle
V (x,y components)
AV
(feet/minute)
Particle Size & Cor
1,500 (0, 1,500)
300 (0, 300)
1,200 (0, 1,200)
Particle Size & Cor
300 (0,300)
1,200 (0,1,200)
1,500 (0, 1,500)
0
Particle Size & Coi
0
0
1,500 (0, 1,500)
1,500 (0, 1,500)
Particle Size & Coi
0
0
1,500 (0, 1,500)
1,500 (0, 1,500)
Particle Size & Coi
0
ctor Not Used
Ycg
(inches)
dition - 1,
+52
+24
+50
dition - 1,
+24
+41
+44
+18
dition - 1
+54
+17
+65
+ 103
dition - 1
+62
+ 15
+75
+ 118
idition - 1
+59
Ysd
(feet/
standard
deviation)
2" Cubes,
0.83
0.24
0.47
'2", Preoil
0.40
0.43
0.46
0.40
'2 x 1/2 x -
0.83
0.38
1.13
0.90
x 1 x 1/2"
0.60
0.38
0.85
0.87
' Cubes, D
0.52
X
eg
(inches)
Dry
-11
+ 1
-7
jd
+1
-4
-4
+1
:", Dry
+ 10
+5
-6
-5
Dry
+9
+5
-9
-6
ry
+6
X ,
sd
(feet/
standard
deviation'
0.48
0.24
0.32
0.33
0.27
0.37
0.32
0.41
0.42
0.58
0.62
0.22
0.36
0.48
0.35
0.32
29
-------
TABLE J.
SORBENT BROADCAST DATA (Cont.)
Test
No.
"27
28
29
40
41
42
43
* Dii
vd
feet/minute
1,050
1,050
1,050
1,250
1,250
1,250
1,250
fuser/Deflecto
V (x , y components)
(feet/minute)
1,500 (0, 1,500)
1,500 (0, 1,500)
0
Particle Size & Con<
0
300 (0,300)
1,200 (0, 1,200)
1,500 (0, 1,500)
r Not Used
Y
eg
(inches)
+93
+49
+ 19
ition - 1'
+23
+19
+29
+41
Ysd
(feet/
standard
deviation)
0.49
0.65
0.40
Cubes, Pi
0.57
0.52
0.48
0.66
Xog
(inches)
-3
-5
0
eoiled
+2
-1
-2
-4
Xsd
(feet/
standard
deviation)!
0.60
0.40
0.32
0.42
0.49
0.50
0.49
30
-------
polyurethane foam with 100 pores per inch) broadcast air
velocity, wind velocity Cnormal to the broadcast direction,
x, and the para-llel, y), plus the specific system used in
the tests are the independent variables. The dependent,
measured, variables are as follows:
Ycg = the mean distance (first moment) in the broad-
cast direction to which a sorbent particle was
carried during a test (inches)
YS(J = the standard deviation or second moment of the
distribution of sorbent particles in the direc-
tion of broadcast (feet)
Y = the mean distance (first .moment) normal to.the
broadcast direction at which a sorbent particle
was deposited (inches)
Y , = the standard deviation or first moment of the
distribution of sorbent particles normal to the
direction of broadcast (feet)
Measurement was-carried out by literally counting particles
on a six inch grid after a broadcast run had been made.
Initial observations verified that-the distributions were
close to Gausian in shape and, therefore, that a "normal"
statistical analysis and description pertained.
The Figures (Figure 7 and Figure 8) on the following two
pages illustrate the basic configuration of the laboratory
broadcast system and the diffuser deflector called out in
Table 1. The system consisted essentially of an air
blower with damper, a screw feeder which injected sorbent
into the blower discharge stream, and galvanized duct
which guided the air/sorbent stream to the point of. exit
from the system. Initial tests (compare numbers 8 and 10)
confirmed that free discharge from an open duct led to
relatively large broadcast range (44 inches) and high
scatter (standard deviation of about 0.47 feet). On this
basis a broadcast diffuser/deflector was designed in an
attempt to (1) separate carrier air and sorbent from one
another after exit from the duct; and (2) direct the
sorbent down onto the receiving surface rather than
allow excess scatter over the surface. The results
comparing these two basic approaches are'discussed later.
Broadcast Velocity
Inspection of the data of Table 1 indicates that both
31
-------
Co
KJ
V \ \ \ Air and
k >- >
/ \ \ \ X Sorbent *"
Tapered Screw Feeder
To Nozzle/
' Diffuser
Subsystem
Figure 7. BLOWER/INJECTOR SUBSYSTEM
(plan view cross section)
-------
Mesh Diffuser
To Blower/
Injector
Subsystem
Nozzle
Main
Conduit
Nozzle
/ Opening
Typical Sorbent
Deposition Area
Diffused
Air
Figure 8. NOZZLE/DIFFUSER SUBSYSTEM
(side view)
-------
broadcast distance and standard deviation vary in pro-
portion to broadcast velocity in the direction of said
velocity. That is, the exponent a^ in the general dimen-
sionless equation is approximately %. Normal to the
broadcast velocity, its effect (on, say X and Xg(j) is
less important and for the purpose of this work can be
considered negligible.
Broadcast velocity, as to be expected, is, therefore, a
very important parameter- which, as it is increased, not
only increases deposition range but also decreases control
over deposition pattern. In this sense, there is no scal-
ing advantage, and broadcasting by air blowing will be in-
efficient for a\large scale system just as has been demon-
strated for small scale systems in the laboratory and a
rectilinear random model on paper.
Wind Velocity
For estimating purposes, wind velocity may be considered
to provide proportional increases in both broadcast distance
and deviation as does broadcast velocity. Its functional
relation to these length parameters can, however, not be
represented quite so simply as that for broadcast velocity.
Under reasonable conditions (and for the purposes of the
study here) the kinematic viscosity of air (ya/pa) should
have a minimal effect on the performance of the broadcast
system and thus Vw and H which appears also in dimension-
less groups 2 and 6 need not be considered in dimensionless
group 3. Therefore, the exponent c is zero. With the
effect of dimensionless group 3 being nullified the effect
of Vw is in group 2 with the exponent 6.
Although the wind velocity may drop to zero, the various
length parameters describing the broadcast distribution do
not necessarily also go to zero (due to the existence of
a finite broadcast velocity). The implication is that the
parameter b^ in the general distribution equation cannot be
a constant. However, since the parameter a^ is approx-
imately constant and since for the range of experimental
investigation wind velocity had much the same effect as
broadcast velocity, the dimensionless function (V^/Hg)^
may be equally well represented by the dimensionless form,
1 + B Vw
34
-------
where B is approximately constant. The new form of di-
mensionless group 2 in the above equation conforms to all
the requirements of dimensional analysis and.represents the
physical system more accurately. For the laboratory scale
system under consideration, (Kg)^ has.a value of about 11.4
ft/sec. The data shows that for wind velocities of 0, 300,
1200 and 1500, ft/min (dry, \ inch cubes and broadcast
velocity of 1050 ft/min) in the direction of broadcast,
Y was approximately 17, 24, 50 and 8,0 inches respectively.
Therefore, B has an average value of about 1.3 which is
sufficiently precise for preliminary design purposes.
Sorbent Oiliness
Sorbent oiliness relates to two basic phenomena: .(1) the
effect of sorbent density on broadcast pattern and,(2) the
hindering of broadcasting by particle interactions due to
"tackiness." In a strictly ballistic sense, the former
phenomena would be of little importance; however, ejection
of material from a nozzle does depend critically on material
density. -On the other hand, the latter phenomena would seem,
at least superficially, to be of more importance, but the
data does not appear to bear this out. Unless the sorbent
has a considerable excess of very tacky (viscous) oil, the
turbulence of the broadcast air stream apparently imparts
enough energy to the .particles to permit only collisional
interactions, not surface-to-surface adhesion.
Typically, oiled sorbent which has been wrung out has a
density of 3 to 4 times that of dry sorbent. The data of
Table 1 indicates a decrease in the broadcast length
parameters of 10 to 20% due to this 'increased sorbent
particle density. The parameter d_ in the basic broadcast
has, therefore, an apparent value of about -0.13.
Sorbent Size and Shape
Size and shape of the sorbent particles should and do have
a very significant effect on broadcast pattern. Surface
or "sail" area would seem to be the parameter of most
relevance; defining such a parameter in quantitative terms
can, however, be a considerable problem. It is best to con-
sider particle size and shape effects on an individual
basis, but with some caution in its application the analysis
in the following paragraph can be of use in preliminary
design considerations.
Consider the sorbent length or size parameter, Cp, in the
basic broadcast distribution equation to be the square root
of the particle surface area, S, or
35
-------
eP = &•
Now, four basic particle sizes.and shapes were.considered
in the laboratory tests: % inch cubes, 1 inch cubes,
1 x 1 x % inch plates and h x ^ x 4 inch parallelpipeds with
surface areas of 1^, 6, 4 and 8% square inches respectively.
Considering the data for a 1,050 ft/min broadcast velocity,
1500 ft/min wind velocity, without the diffuser/deflee tor
and dry sorbent, the effect of particle surface area is
represented by a value of e_ of about -0.14. This implies,
for instance, that broadcast distance decreases with ine
creasing sorbent particle surface area; similarly the
spread of the particle distribution also decreases slightly
with 'increasing surface. Again, this development must be
used with great care since the effects it attempts to re-
present are very complex. Anyway, the effect of particle
surface area on broadcast deposition is not very large;
that is, it is not nearly as important as the effect due
to a similar parameter-on the basic sorption process.
Broadcast Equation
The basic broadcast equation has been reduced to the
approximate form:
di = A Vd
H (Kg)*2
1+1.3 Vw
(Bg)
Combined with the rest of the basic data on Table 1, this
relation can be used, judiciously, in the preliminary
design of an air carrier broadcasting system.
Air Blown or Gausian vs Rectilinear Broadcast System
In "A Random Model"- section, a mathematical analysis.was
made of the percent coverage that could be obtained by a
random 2 dimensional rectilinear distributed system for
the distribution of sorbent particles. Experimentally, a
less optimal system broadcasting of sorbent by Air carrier
was used. From the air carrier system a Gausian distribution
at best would be expectad. An a priori evaluation of the
two systems is that the rectilinear system, for the same
percent of even coverage and sorbent loss, will require the
broadcasting of approximately half the sorbent.
Ideally, particle-by-particle placement on a slick (with
exactly the right spacing between the positioning of
particles) would be preferred. In a practical sense,
36
-------
however, such a goal is not realizable.
This study lead to the conceptualization of a screw
conveyor deposition system which.,. although not ideal in
the sense of the previous paragraph, provides for positive
placement of sorbent in a boom channel with a minimum of
effects due to the environment and the behavior of the rest
of the oil recovery system.
In the rectilinear system the probability of a particle
being placed across the width of the slick is equally un-
likely. This means that the edges of the channel will be
equally well covered with sorbent as the center. In the
air carrier system a Gausian distribution occurs in the
channel cross section. In the Gausian distribution, the
cross section density of particles in the channel is the
familiar "bell shaped curve," meaning that a large percent-
age of the particles are clustered in the middle of the
channel and little on the edges. If one wishes to even out
the coverage in the air carrier system, then the losses may
reach percentages as high as 30% while in the center there
will still be a peaking of sorbent density. - One may con-
sider an air carrier system to be analogist with that of a
shot gun blast into a target where the center is more
densely covered than the periphery with particles strag-
gling out beyond the ddges.
The rectilinear system is somewhat analogous to a sewing
machine with a loose needle thus the probability of a
stitch at any point along the extant of the sewn part is
equally likely.1 Since the designing of a rectilinear
dispersion system is easier and more reliable than that of
a blower system, it is obvious why we have chosen the more
novel of the two.
Furthermore, the sensitivity of the broadcast distribution
parameters to the many parameters of interest weighs
heavily against the use of air carrier approach. For in-
stance, variations in discharge velocity cannot be: tolerated
since broadcast distance and spread are approximately pro-
portional to same. Sorbent losses with a fixed boom or
containment system would,. therefore, increase with some-
thing like the square of an increase in broadcast velocity.
The effectiveness of broadcast deposition pa,t.tern. C"evenness")
would decrease in a similar fashion wi .h .decrease in broad-
cast velocity.
Wind velocity would have a similar effect but, more im-
portantly, is an uncontrollable design factor. Ousting
37
-------
winds of from 0 to 10 knots would produce fluctuating
changes in broadcast distribution..of from 0. to 200.%. This
represents about an order-of-magnitude changes in sorbent
losses or deposition effectiveness which is generally un-
acceptable in process design.
Furthermore, sorbent density Cwringing efficiency or
changes in same) has a significant effect on broadcast
deposition. This also is deleterious from the viewpoint of
a relatively steady-state process design.
In conclusion, broadcasting of sorbent particles using an
air carrier stream is an unacceptable technique except
under the most idealized environmental and system oper-
ational conditions.
The only drawback to a rectilinear dispersion system is its
Inflexibility as to width. A blower system can readily be
adjusted in all of its inefficiencies to handle the width
of 10-20 feet. However, to go from 10-20 feet with a recti-
linear dispersion system requires an entirely new dispersal
design. Since the channel width is fixed by the system,
this is not considered a drawback.
In the above discussion it has been assumed that quiesent
air and other optimable conditions are present. For vary-
ing winds, turbulence, rocking motion and variations in
density of particles due to inefficiencies in the recovery
system or degradation of particle size, the rectilinear
system becomes even more effective. For example, consider
the most commonly expected variable, namely wind velocity.
Since -the air carrier system uses air velocity to disperse
the particles, turbulence and wind velocity will disperse
the particles where they are not desired. This means higher
losses and even less uniform distribution. The rectilinear
system on the other hand, because it can be shrouded close
to the water level will be little affected by air velocity.
A shroud leading from the dispersion nozzle to within a.foot
of the water will allow the random distribution of the
particles to fall unaffected by turbulence, dust,.or steady
wind.
Other variables such as the roll of the vessel, change in
the air pressure density.. temperature will have no affect.
Furthermore, strong variations in the density of the.par tittle
due to retaining either water or oil will likewise not
affect the rectilinear nozzle. Variations in particle size
will not affect the volume distribution -across the recti-
linear nozzle - it will affect statistically where the
38
-------
smaller particles land. In other words, we expect some
segregation of the finer particles to occur -but ;where
smaller particles are deposited more.of.them will be
deposited so that the 'volume of material distributed
across the inlet will remain roughly rectilinear.
In the air blown system, these smaller particles will be
carried much further and will be subject to dusting, tur-
bulence, and steady wind resulting in much higher losses.
Behavior of Sorbent in Slick
As a slick is compressed, distribution of the. sorbent
particles becomes better than random. If the particles
are small enough and the initial distribution were thor-
oughly effective, then there will be a mechanical re-
arrangemeht of sorbent distribution as the slick is com-
pressed by the herding booms. The distributional re-
arrangement will be short ranged and thus voids in the
center if the particles are dense in the center will tend
to disappear whereas voids in the periphery if the dis-
tribution is scant may tend to decrease but not signif-
icantly. Thus, as the slick is compressed it is important
that the initial particle distribution be as uniform as
possible. . In the case of platey material, particles
falling on top of each other will act as one particle
because of tne stability of the stacking. Thus, the re-
arrangement phenomenon argues.for square qr round particles
rather than platey- ones. Tfurth'ermore, bridging is less
likely to occur with squa'i'e" .or'chunky particles/than with
platejr or stringy particles. Once bridging occurs,
effectiveness of the sorbent recovery system is.comprised
and lost rate becomes exhorbant as will ,be .discussed in
Section VI, Choking of Sorbent in Channel.
Rectilinear Broadcast System
In the above paragraph it was
-------
screw conveyor. The sides of the conveyor are sloping a
triangle arranged so that the amount of volume of material
spilling over the edge of the conveyor is equal at all
points. This arrangement means that particle size and
shape do not materially affect the-volume of material being
dropped at a given point in the nozzle.
The material then drops straight down towards the surface
of the sea, shrouded from the wind by two skirts of metal
parallel to the direction of flow of the cotiveyor and
perpendicular to the surface of the sea. Essentially,
thesesskirts are simply wind shields that allow material
to drop vertically.
The system is easy to construct, easy to run, it is rugged
and reliable and should work under all but hurricane
level winds. The system has the further advantage that
foreign body damage is unlikely to occur for a wide variety
of particles in both size, hardness and density can be
handled. It is, in short, the well known screw conveyor
of the chemical engineering field adapted to this use.
Conclusions
In this section we have described and developed a dimension-
al analysis equation shown earlier in this section under,
Broadcast Equation, which fits the experimental data for
air carrier broadcast systems. Furthermore, based on these
results we have determined an alternate approach. The
following conclusions can be drawn.
(1) The above mentioned equation adequately describes
air carrier broadcast systems.
(2) Air carrier broadcast systems result in high loss
and poor coverage.
(3) Particle shape should be chunky rather than stringy
or blocky.
(4) Rectilinear distribution systems will yield lower
sorbent loss, uniform dispersion and greater oil
pickup*.
40
-------
SECTION VI
SORBENT HERDING
The sorbent recovery system consists of a boom, channel,
a recovery conveyor and a broadcast system. The layout,
Figure 12, shows a ship and attached to the ship, the
herding boom, which is. inclined out at 30° from the di-
rection of the ship. This boom compresses 4/1 the slick
into a channel four feet wide. The absorbent is broadcast
in the area and is also entrained in the channel. The oil
sorbent and water move down the channel and are picked up
by a sorbent recovery conveyor.
Description of System
The system attached to the ship consists of a boom support
arm jotting out at right angles from the ship terminating
in the support float wh^Lch is the lead end of the herding
boom. The herding boom is a barrier extending 2.2 feet
below the nominal surface of the water. The boom is
diamond shaped in cross sectioned with the water level
being at the widest point of the diamond. The boom is
described in greater detail on Page 125 in Appendix A.
The boom itself is articulated and has rapid wave following
characteristics and at the design speeds is effective in
containing the oil. The boom's function is to compress the
slick into the channel.
Sorbent is broadcast in back of the channel mouth where the
slick is being compressed. The slick and the absorbent
enter the channel. The juncture between the boom and the
channel is supported from the ship by the aft boom support
arm. The outer barrier to the channel is identical to the
herding boom in construction but its orientation is parallel
to the center line of the ship. In the channel the slick
is four or more times thicker than the ambient slick due
to the compression from the herding boom and from the ship's
bow. The speed of the slick moving through the channel is
significantly less than the free stream speed. This causes
further thickening of the slick and increased contact time
between sorbent and oil. The reasons for the lower channel
speed of the slick are complex but are due in part to drag
forces and also the forward motion imparted in the channel
from the recovery conveyor belt. The bottom part of the
recovery conveyor belt acts as paddles, tending to create a
general motion forward in the channel and at ship speeds
below half a knot material tends to flow out of the channel
rather than back towards the pickup conveyor belt.
41
-------
The pickup of the sorbent is done by an open mesh belt
whase openings are small enough to trap the sorbent
particles, but large enough, to allow water and smaller
debris to flow through. Essentially it is a sieve sep-
arating the sorbent from the sea water. As the sorbent rides
up the conveyor belt, mechanically entrapped water drains
from the system. The pickup conveyor is described in
greater detail in Section VII.
Boom Design Problems
Other than durability, suitability and wave following
characteristics of the boom, the main problems unique to
the oil sorbent recovery system are: design strength, loss
of bulk oil and loss of oil by entrainment. Calculations
based on area drag show that at 4 knots the strength safety
factor for this sytem is tenfold, indicating that mechanical
ally this system may be operated for extended periods of time
at 4 knots without mechanical failure.
Loss of bulk oil under the boom when herding is governed by
a dimensional number called the densiometric Froude number.
This "run-under" the boom or "drainage" is a gravity
phenomenon and the point where drainage will occur is at
Frpude numbers in excess of 1.1.
N = V
' "— ' (Froude Number)
~
.
I/
1
L8
Po
Where V = Free stream velocity perpendicular to boom
L = Depth of barrier below the slick
g = Gravity constant
pw = Density of water
p = Density of oil
This equation may be modified to calculate the depth:
42
-------
NFRD = 1'1
g - 32.2
Pw~Po = .135 (#4 Fuel Oil)
pw
= 1 knot (1.685 ft/sec)
At the design speed of 2 knots per second the velocity at
right angles to the boom inclined at 30° is not 2 knots,
but one knot because Sin 30° is 1/2.
6.3
.55 feet
At a free stream velocity of 4 knots the perpendicular
velocity of the boom is 2 knots and the depth from equation
6.2 to maintain a critical Froude number is 2.2 feet. This
is the design depth of the boom and it means that with a
nominal #4 oil the system will not suffer oil bulk oil
losses. In actual operation, however, when the boom tends
to bend or the density of the oil is increased, oil will
begin sliding under the boom at lower velocities. However,
at the designed operating speed of 2 knots, little or no
loss is anticipated.
Sorbent loss under the boom, due to the Froude effect, is
expected to take place at even higher Froude numbers due
to mechanical properties of the sorbent and the fact that
it will consistently have a lower density than the oil.
Another form of oil loss is called entrainment, a phenomenon
in which the oil is broken into droplets by turbulence and
behaves as does the water column. Thus, since the bulk of
the water flows under the boom, entrained oil will also.
For a boom of the depth indicated here, entrainment losses
begin at approximately 2 knots. Entrainment is the result
of the turbulence due to the water in contact with the boom
and thus the amount of oil lost is the function of the
square of the velocity. At lower speeds, the percent
lost is small.
Sorbent losses due to entrainment will be quite small due
to the fact that sorbent is not broadcast in the region of
43
-------
contact with the boom and will in this region have sig-
nificantly lower densities than the oil, thus remaining
on the surface, easily herded.
Hydrodynamics of Slick Flotf in Channel
When the slick enters the mouth of the boom it is compressed
into the channel. Basically, two effects take place, the
slick is thickened and the velocity of the slick in the
channel changes. Obviously if the slick velocity remains
the same as the free stream velocity, a significant thicken-
ing will occur of the slick in the channel. If the width
ratio of the mouth to the channel is four to one, then the
slick thickness in the channel will be four times the free
stream thickness. However, due to drag, forces on the slick
and the sorbent recovery belt the slick velocity in the
channel is significantly reduced.
Figure 9 is a dimensionless plot of channel width divided
by mouth width vs slick velocity divided by free stream
velocity. This plot indicates two important phenomenon;
first, that (for the oils studied) when the channel width
to the mouth width ratio is 0.25 (1-4), the slick velocity
is approximately one-half the free stream velocity. Below
.25 the slick velocity rapidly decreases with decreasing
channel to mouth width ratio. Thus, the design parameter
chosen for the system was a channel width to mouth width
ratio was .25.
The second important phenomenon is that increasing oil
viscosity means decreasing slick channel velocity. For
example, the #2 fuel oil has a velocity of approximately
55% of that of the free stream velocity whereas for #4
fuel oil the slick channel velocity decreased to 45%. This
means that residence time in the channels for the more
viscous, heavier oils increases thereby increasing contact
time with the sorbent and also increasing depth of slick
in the channel.
Assuming no loss of oil under the boom, the simple com-
pression means an increase in slick thickness. However,
there is a further thickening of the slick due to the
decrease in slick velocity in relation to the free stream
slick velocity. Figure 10 shows the slick velocity as a
function of distance along the channel. In Figure 10, the
mouth width to channel width is 3/1 and simple compression
would dictate that the slick thickness should be three
times as great in the channel as in the free stream. But
actual measurement shows that it is over six times as thick.
44
-------
H
S3
1.0
0.9
0.8
0.7
E* 0.6
0.5
0.4
0.3
0,2
0.1
Model: 1/4 scale
ChannGl width: 6"
Mouth width: 6U - 3'0"
Channel length: 4'
Mouth length: 3.5'
Current velocity: 0.43 knt (0.726 ft/sec)
No wave
0 0.1 0.2 0.3 0.4 0.5 O.C 0.7 0.3 0.9 1.0
CHANNEL WIDTH
MOUTH WIDTH
Figure 9. RELATIONSHIP BETWEEN SLICK VELOCITY
IN CHANNEL AND BOOM GEOMETRY
45
-------
Oil type: #4 Fuel Oil
10
s
o
o 9
w
H 8
g 7
o
ei
p*
en
en t-
W D
o
CO
«/
Current velocity
0.76 ft/sec.
(0.45 knt)
Current velocity = 0.524 ft/sec,/_ 0,
M v _^_ ' (U.jJ.
1 . 1 » « * » «—
0 0.2 0.4 0.6 0.8
DISTANCE ALONG THE CHANNEL
TOTAL LENGTH
Figure 10. RELATIONSHIP BETWEEN OIL PILE-UP
AND CURRENT VELOCITY
46
1.0
-------
This additional doubling of the thickness is due to the
channel slick velocity being half the free streajn velocity.
Hydrodynamics of STJrberit Flow
In the previous section the behavior of oil in the channel
was describ.d without sorbent being present. When a sor-
bent is added there is a further increase in slick thick-
ness and decreased slick-sorbent velocity. This relation-
ship is linear for a wide variety of conditions and is de-
scribed by Vs= 0.57 V0 where Vs is the velocity of the
sorbent in the channel and Vo is the velocity of the oil
slick in the channel.
Thus, the addition of sorbent of the type used in the
system cuts the channel velocity flow in half again (Vs=
.57 Vo). Therefore, the ratio of sorbent velocity in the
channel to the free stream velocity is approximately one-
fourth, half the reduction being due to the presence of the
oil - Figure 9 .- and the other half being due to the
presence of sorbent resulting in the equation in the previ-
ous paragraph. From straight mass balance considerations,
it follows that if the velocity of the sorbent and slick in
the channel is reduced to one-fourth the free stream
velocity then the thickness of the slick'must be fourfold
in order for the same amount of oil to pass a given point
in a given time.
There is a further enhancement of the slick thickness at
the mouth of the channel due to the fact that the sorbent
has not absorbed oil. Thus, the additional thickness of
the slick with sorbent will decrease as one moves down the
channel and oil absorbs into the sorbent material. This
enhanced slick thickness at the mouth of the channel is
due to the free space in the sorbent before it becomes
saturated with oil.
Waves adversely affect sorbent velocity. Interestingly
enough, there is at first a marked decrease in slick channel
velocity with increasing wave height and then as wave height
further increases, sorbent velocity in the channel climbs
back asymptotically to the wave free velocity.
Sorbent size also affected channel velocity. As sorbent
size increased, sorbent velocity in the channel decreased.
47
-------
Choking of Sorbent in Channel
Under a variety of conditions, choking was observed to
occur in the channel. Once choking occurs there is a con-
tinued build-up in front of the channel and both sorbent
and oil are lost. Sorbent build-up in the channel has the
typical shape of head waves similar to those formed by an
oil slick in front of a barrier. The head wave migrates as
does oil slick head wave and it sheds individual sorbent
particles similar to the way an oil head wave sheds oil
droplets. Choking is dependent on sorbent broadcast rates
and size as well as boom geometry and current velocity.
Choking occurs as slick or sorbent channel velocity de-
creases and a packing of the sorbent occurs where particles
not only are touching each.other on the surface but also
underneath. Literally, bridging occurs which has sufficient
physical strength to resist the velocity pressure head. In
actual operation, choking can be controlled by vessel speed
and sorbent broadcasting rates and the choke channel can
be broken up mechanically by agitation supplied from on-
board the vessel*
Contact Time
Sorbent particles one inch on an edge will reach operational
saturation in under five seconds if the slick thickness is
great enough. Thus, the required contact time in the channel
is five seconds or longer. At 2 knots, sorbent particles
cover 20 feet in 5.88 seconds. Since sorbent velocity in
the channel is approximately 1/4 of free stream speed, the
residence time in the channel is four times longer than the
free stream velocity. Thus, for a free stream velocity of
approximately 2 knots the contact time is on the order of
24 seconds and at free stream velocity of 4 knots, contact
time is on the order of 12 seconds. Thus, the absorption
or contact time for sorbent particles in the oil is more
than sufficient. Actually, the channel length could be
cut by ten feet and still meet design specifications.
Oil Pickup Rate
A series of scale model tests were run in a wave tank to de-
termine the efficacy of the system. Basically, a 1/4 scale
model system was set up, the sorbent was broadcast, picked
up, cleaned and rebroadcast. This was done under a variety
of conditions to determine the dynamics of the system de-
scribed earlier in this section. After choosing the final
configuration a further series of tests was run and then
48
-------
based on the data obtained, an equation was set up de-
scribing what sorbent flow rates should be and what the
oil pickup rate would be. It is as follows:
(1) QF = (C) CW) (h) Ckil (k2)
(2) Q0 = Q CEi) Cg2> C53)
o
Where Q,y = Sorbent flow rate down channel (ft /min)
C = Free stream velocity (ft/min)
W = Channel width (ft)
h = Cube height (ft)
k^ = Area density ratio of cubes in the channel
k£ = Ratio of sorbent velocity to current
velocity
QQ = Oil pickup rate (ft/min)
C^ = % sorbent retained after each cycle
£9 = 5S of oil and water recovered by volume
5 = % of oil recovered by volume
The constants k^ and k2 are estimated to be respectively
0.5 and 0.3.
Assuming a free stream velocity of 2 knots, a channel width
of 4 feet and sorbent cubes 1 inch on edge, the sorbent flow
rate is 10 cubic feet per minute.
It is estimated that £-^ the sorbent loss rate will be less
than 0.5% or g^ equal to .995. £2 M the percent of the
cube volume that is filled with liquid - oil or water. g2
is assumed to be one due to the long contact times in the
channel relative to that needed for operational saturation.
£o - the percent of oil by volume recovered from the sorbent
cube after it is picked up, is estimated to be 95%.
Substituting the figures in the above equation, number (2):
Q0 = *p
-------
with QF - 10 ft
£l = .995
52 = -95
£3 - 1.0
QQ = 9.45 ft3/min
= 70.5 gal/min
- 4230 gal/hr (maximum)
The results indicate that with the experimental data and
the design parameters that the 3000 gal/hr condition can
readily be met. The assumptions made include several about
slick characteristics. It was assumed that the slick thick-
ness was 1.5 mm, that it was continuous and that the ship
was continuously in the slick. If there are variations in
slick thickness or patchiness of the slick, but these are
of short duration and the overall volume of oil remains
constant, then the system performance will still be ap-
proximately the same. However, if there are deep slicks in
spots with extended patches of clear water between the oil
then the system's performance will be seriously degraded.
Sorbent Loss
As designed, sorbent loss is not expected to be a sig-
nificant problem. Loss of sorbent during the broadcasting
process should be minimal. From Froude number considerations
and operating at 2 knots, loss under the boom should not
occur. Entrainment losses should be a minimum. For an
operational system, harvesting losses should be quite small.
Condition of bleeding sorbent under the boom, choking in the
channel, degradation of the boom or of the pickup system
will lead to rapid massive sorbent loss. Sorbent can also
be lost if the system does not shut down during operations
involving sharp turns or backing. Degradation of the
mechanical pickup system will obviously lead to sorbent
loss.
Conclusions
The sorbent herding system will be able to meet design
specifications with relative ease and minimal losses. Con-
tact time is sufficient,, loss of oil and sorbent in the
operational system is minimal.
50
-------
Operating at the maximum system speed, 4 knots free stream
velocity, some oil loss is anticipated but sorbent contact
time is more than adequate and sorbent loss should be
minimal. However, should choking of the channel occur,
rapid degradation of the system's performance will result
with massive losses of sorbent. The situation would be
exacerbated with the presence of debris.
51
-------
SECTION VII
SORBENT PICKUP
The oil saturated sorbent is picked Tip by an inclined open
mesh wire belt and then dropped on a horizontal belt trans-
porting the particles to the recovery system. The open
mesh screen of the pickup belt allows water to freely flow
through the system but the sorbent particles are too large
to go through the mesh of the belt and hence are lifted on
the deck. During the lifting of the sorbent from the water
surface to shipboard, the open mesh allows mechanically
entrained water to drain back into the ocean. Drainage is
important as water would become entrained in the oil during
processing, degrading the oil product with excess water.
The sorbent belt extends two feet below the surface of the
water. The belt is permanently attached to the ship and
its actual depth in the water will vary with the ships roll,
This will not be a serious problem.
As the belt moves through the water it tends to decrease
the thickness of the oil-sorbent slick. Material in front
of the channel pushes the slick to the rear, towards the
belt, and the mechanically absorbed oil absorbent is lifted
up by the belt's action. The water passes freely through
the mesh of the belt. Losses around the edge of the
conveyor are prevented by the use of a flexible skirt
sealing the edge of the conveyor to the ship* The con-
veyor is the full width of the channel and sealed to the
outer boom by a mechanical linkage.
The required theoretical power to lift the sorbent is less
than 3/10 of a HP but the design specifies a 2 HP motor
with a variable speed drive. The added HP is used to over-
come belt and the variable speed motor friction and other
binds in the system which may result from prolonged
operation.
Sorbent and Oil Losses
Operating at the designed condition of 2 knots an hour,
little sorbent or oil loss is anticipated. Oil which is
not absorbed may pass readily through the screen and hence
constitute a loss but it is a system loss not a mechanical
one. Sorbent is unlikely to pass under the end of the
conveyor unless ships roll becomes a problem. Build-up
of sorbent in front of the conveyor to the point where it
would pass under the conveyor is not considered a problem
53
-------
since there is a significant thinning of the slick sorbent
area near the conveyor. Debris too large to be picked up
by the conveyor can build up and force sorbent underneath
the boom resulting in a loss not only of sorbent but also
oil.
Mechanical Constraints
The recovery of sorbent from the water is considered a
simple straightforward operation and should not produce
undue loss of sorbed oil or severe problems of operation.
Heavy debris in the system will degrade not only the
system's performance but also the belt by mechanically
banging into it and widening screen openings. When the
sorbent is picked up, particles on the edge may fall or be
blown off the belt under certain conditions particles
could tumble and gain enough energy so that, if deflected,
they may fall off the belt and land outside the barrier
area. This is not considered a major deficiency.
When the sorbent is lifted up to the recovery belt it is
then dropped on a horizontal belt leading to the processing
equipment. Due to variations in ship designs, the material
dropping from the recovery belt to the transport belt can
be lost. A very simple mechanical shroud will alleviate
this situation.
Conclusions
Sorbent pickup is a relatively simple, straightforward
problem. Since it is a mechanical system of a certain
degree of fragility operating in the sea, operational
problems may develop.
54
-------
SECTION VIII
RECOVERY OF OIL FROM SORBENT
Introduction
Due to the judicious initial selection of sorbent with the
overall project goals in mind, sorbent and oil recovery
(separation) after sorption proved to be a relatively
simple matter. At first, a simple wringer system was used
on oil soaked sorbent; operational difficulties (.i«e.,
maintenance of sorbent feed through the rollers, imposition
of adequate pressure between the rollers, and excessive
power requirements) eventually led to the abandonment of
this approach. A converging conveyor belt wringer was then
selected as a more desirable alternative. Laboratory tests
bore this out, and, therefore, this approach was selected
for the preliminary design of the total system.
jior_bent Recycle
Selection of fully reticulated polyurethane (ester-type)
foam has several obvious advantages. In this case,
"Scottfoam" is being referred to. Not only do such foams
have an excellent capability for oil sorption, but they
can also be wrung out successfully in that they maintain
their sorption efficiency with reuse and do not degrade
mechanically. Table 2, shown on the next page, presents a
summary of the results of a recycle experiment illustrating
just these points.
For these 100 sorption cycles, the average oil carried by
the polyurethane sorbent was 30 grams per gram of dry foam
whereas the average oil recovered was 28 grams per gram.
The data indicates that sorption and recovery actually
improve slightly as a function of number of cycles (for
up to the 100 cycles considered). This can only be due to
conditioning of the foam material. The polyurethane foam
selected for use in the system being developed here are
very effective, can be recycled efficiently, and stands up
well to the mechanical abuse required by such a system.
These positive results point to the relative importance of
other factors on sorbent effectiveness (.e.g., sorbent
losses).
Storability is a factor related to the above considerations.
Although it was not considered in as great detail, two
useful pieces of information were obtained during the »
course of the development program. First, storage of dry
55
-------
Table 2.
Urethane Foam Sorbetrt Recycling
Cycle
No.
1
2
3
4
5
10
15
20
25
35
50
76
100
Oil* Sorbed
(gm/gm dry foam)
30
31
31
32
31
31
30
30
31
31
31
32
32
% Oil Removed
Uy Wringing
89
92
93
93
93
93
93
92
96
94
94
93
93
% Oil Retained
After Wringing
11
8
7
7
7
7
7
8
4
6
6
7
6
*No. 4 fuel oil at ambient temperatures.
56
-------
sorbent in a compressed state (for the sack of storage
volume conservation) leads initially to decreased sorbent
effectiveness, but upon recycling the sorbent quickly re-
covers to nearly full effectiveness. Initially only about
20 grams of oil are sorbed ami released by a gram of
sorbent; after about 20 cycles recovery is nearly complete
such that 27 grams of oil sorbed and released by a gram of
sorbent. On the other hand, pre-oiled sorbent which was
stored in a compressed state for an extended period (about
4 months) exhibited both relatively poor effectiveness and
recovery properties. Sorption was reduced to about 16
grams oil per gram dry sorbent and increased only slightly
upon recycling. This effect can be explained quite simply
by a permanent structural change in polyurethane upon
prolonged exposure to hydrocarbons while in a deformed state.
In conclusion, dry polyurethane foam sorbent may be stored
in a compressed state prior to use; pre-oiled (used)
sorbent can be stored for further use, but such storage
should involve sorbent which has been wrung to as dry a
state as possible and remains uncompressed while in
storage.
It should be noted that the above development holds for
oils other than no. 4 fuel oil with only minor variations.
For instance, marine diesel fuel and no. 6 fuel oil exhibit
sorption of 32 and 29 grams oil per gram dry foam respec-
tively; the minor differences from no. 4 fuel oil can be
attributed entirely to differences in the specific gravities
of the hydrocarbons involved. Release of oil from the
sorbent foam (as opposed to the consistency of the physical
sorption) depends much more importantly on the nature of
the oil. This is considered in more detail in the following
section.
Sorbent/Oil Recovery
Although the wringer roller recovery system worked fairly
well during the bench work discussed above, considerable
difficulties arose during tank testing. This led to the
conceptual development and bench testing of a converging,
dual belt wringer system which will be discussed in the
paragraphs which follow.
First, however, it is best to highlight the problems en-
countered with the wringer/roller system during tank test-
ing:
(1) Most important, constant and positive feed through
the rollers was almost impossible to maintain under
operational conditions.
57
-------
(2) Sufficient compression between rollers was achieved
only intermittently.
C3) Related to C2$, power requirements for the wringer/
roller were both highly -variable and deemed ex-
cessive .
(4) Oil recovery efficiency decreased under operational
conditions to 75 to 80% which is comparable to con-
verging, dual belt wringer performance with otherwise
much more positive performance Csee the discussion
below).
In light of the observations just mentioned, a 6 inch wide
model of a converging, dual belt wringer system was con-
structed and tested on a bench scale. Sorbent size (all
cubes), belt speed, oil type and belt compression were all
factors which were studied. In general,, the results of
these studies were very heartening, and, therefore, the
preliminary design of a full scale system was undertaken.
Surprisingly, little difficulty was incurred during these
studies with gravity drainage of the oil wrung from the
sorbent; this bodes well for full scale operation.
Belt speed has a significant effect on system performance
although it is not nearly as severe as might be expected.
Table 3 below presents data on belt speed as it affected
the recovery of no. 4 fuel oil.
Table 3.
Oil Removal as a Function of
_ Conveyor Belt Speed*
Belt Speed Oil Removed
(FPS)
0.17 83
0.33 82
0.67 78
1.00 76
1.67 74
*No . 4 fuel oil sorbed and 1" polyurethane cubes with a
belt compression of 57 Ibs. per foot.
58
-------
Note that a tenfold increase in belt speed results in only
about a 10% decrease in removal efficiency.
Table 4 illustrates the effects of both sorbent size and
oil type.
Table 4.
Oil Removal as a Function of..
Sorbent Size for Several Oils*
Sorbent
Size Marine No. 4 No. 6
Inch Diesel Fuel Fuel Oil Fuel Oil
0.25 58 59 54
0.50 72 70 53
1.00 80 77 55
*Cubic polyurethane sorbent, 0.67 Fps belt speed and
57 Ibs/ft belt compression.
Sorbent size is seen to effect removal of the lighter oils
more drastically than the heavier oils. Removal is good
for the more nearly full size (one inch) sorbent in all
cases: about 80, 75 and 50% removal for marine diesel
fuel, no. 4 fuel oil and no. 6 fuel oil respectively. The
differences are attributable to the ease of drainage of
lighter oils combined with re-sorption after passing
through the wringer.
Compression between converging conveyor belts, like belt
speed, has only a minimal effect on oil removal; this, of
course, applies only to compression within reasonable
limits. Table 5 illustrates the phenomena.
i -,
59
-------
Table 5.
Oil Removal as a Function, of
Wringer Belt Compares sip ri*
Belt Compression Oil Removal
(Ibs/ft) C%)
10 66
20 69
30 72
57 77
86 79
*Marine diesel fuel, V sorbent cubes and a 2" per
second belt speed.
A factor of 9 increase in belt compressive loading leads
to only about a 20% increase in removal efficiency. There-
fore, within reasonable limits, this factor should be de-
signed with an eye on sorbent mechanical degradation and
system power requirements as opposed to oil removal effi-
ciency .
Sorbent/oil recovery using a converging, dual belt wringer
device .has been shown to be both feasible and a remarkably
flexible unit operation.
Alternatives
Two obvious alternative approaches to oil/sorbent recovery
present themselves:
First, oily sorbent may just be collected and disposed of.
Due to the cost of polyurethane foams, however, this is
prohibitively expensive. It is, furthermore, limited to
small scale, constrained, labor-intensive spill situations.
Obviously a cheaper sorbent (.e.g., straw) should be used.
The main benefit of the foam is its reusability.
Secondly, batchwise (e.g., wringer press) recovery of oil
and sorbent could be used. This could be ideal for the
small scale operations mentioned in the previous paragraph
but would not be suitable for high capacity operations which
should not be labor-intensive.
60
-------
Conclusions
A converging, dual conveyor belt wringer should he used in
a full scale oil recovery system.
Such a device has been shown to work well for a range of
oils and to be highly adjustable in terms of throughput.
That is, compressive loading and belt speed have only minor
effect on wringer efficiency. The nature of the oil re-
covered and sorbent size are the more important "independent"
variables.
Capacities will be more than sufficient with such recovery
devices. Using the data of the previous section, the
following design capacity is indicated for a 2 foot wide
conveyor with a 120 pound compression load and operating
at l^s feet per second on 1" sorbent cubes at 50% belt
coverage (something approaching 100% is possible), 2"
depth, and saturated with no. 4 fuel oil at ambient tem-
perature :
Q = 1.5 ft/sec x 3,600 sec/hr x 1/6 ft x 2 ft x
0.50 (coverage) x 0.75 (removal) x 7.5 gal/ft3
= 5,000 gallons/hour.
Similar capacities for marine diesel fuel and no. 6 fuel
oil would be 5,300 and 3,300 gallons/hour respectively.
61
-------
SECTION IX
TOTAL SYSTEM CONCEPT
Introduction
The preceding sections have outlined in detail the prin-
ciples and unit operations required in the prod-uction of a
full scale, high capacity oil recovery system. A much more
complete understanding of basic sorptidn phenomena Cf*°m an
engineering viewpoint) is now possible. Broadcasting of
sorbent materials has been elevated from a speculative art
to one containing elements of applied engineering. Sorbent
herding and pickup have been described and studied quanti-
tatively. Sorbent/oil recovery has been placed in quanti-
tative relation to the necessary independent variables and,
from this, preliminary design and capacity estimates have
been made. Tied together, all these factors now permit
the design of a full scale system. The preliminary design
of such a system is presented in this and the following two
sections of this report. This preliminary design provides
the base for the final design of a full scale system.
The basic system as conceived during the course of the
study under consideration was actually tested on a pilot
(H) scale during an extensive series of tank tests. The
basic test system is shown in Figure 11 on the following
page. Table 6 on the subsequent two pages, summarizes the
results of a series of 32 tank tests aimed at the study of
sorption under relatively realistic conditions. This data
shows a remarkable variation in performance under similar
conditions. This is due to the myriad phenomena which were
discussed in detail in Section VI on Sorbent Herding.
Nevertheless, recovery effectiveness (amount of the swept
slick picked up) ranged from about 50 to 100% for relatively
thin slicks {maximum thickness of 0.51 mm as compared to a
nominal full scale design value of 1.5 mm). Furthermore,
100% effectiveness could be sustained in the tank tests
when operating conditions were selected (.reached) properly.
Unfortunately, the duration of these tests was such that
minor adjustment, in conditions was all that was possible
once a test had started. Since oil was only very infrequently
noticed to bypass the sorbent/boom/pickup system, generally
a 100% effectiveness should be possible except in the cases
where channel overloading of oil (with respect to possible
sorbent throughput) occurs.
The three basic test series, % inch cubes and no. 4 fuel
oil, 1 x 1 x % inch blocks and no. 4 fuel oil, and % inch
63
-------
OIL/SORBENT
RECOVERY DEVICE
SORBENT FEED HOPPER
SCREW CONVEYOR/FEEDER
AIR BLOWER
BROADCASTER NOZZLE
BOOM MOUTH
HERDING BOOM
OILY CHANNEL
CLEAN WATER SURFACE
SORBENT BED & OIL
PICKUP CONVEYOR
FIGURE 11. PILOT (1/4) SCALE SPILLED OIL RECOVERY SYSTEM
-------
ON
Table 6.
Sorption Wave Tank Test Data
Test
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Broadcast
Rate
(ft3/min)
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.67
0.14
0.33
0.67
1.33
0.67
0.67
0.67
Slick
Thickness
(mm)
No. 4 Fuel
0.51
0.34
0.25
0.17
0.08
0.04
0.17
0.08
0.17
0.08
0.08
0.08
0.08
0.08
0.08
0.17
0.34
Surface
Current
Velocity
(knots)
Oil Sorbed with
0.40
0.59
0.79
0.40
0.79
1.58
0.40
0.79
0.40
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
Wave
Height
(inches)
1/2 inch
0
0
0
0
0
0
2
2
4
3.5
2
2
2
2
2
2
2
Contact
Time
(sec)
Polyurethane
61.2
21.6
31.0
63.0
20.0
6.2
81.0
17.4
39.0
9.6
11.9
11.0
17.4
21.2
17.4
12.6
10.1
Oil
Recovery
Rate
(gpm)
Foam Cubes
0.805
0.895
0.751
0.265
0.212
0.487
0.527
0.542
0.311
0.481
0.350
0.525
0.542
0.456
0.542
0.583
1.015
Recovery
Effectiveness
(%)
54
60
50
53
42
97
100
100
62
96
70
100
100
91
100
58
51
No. 4 Fuel Oil Sorbed with 1 x 1 x 1/2 inch Polyurethane Foam Blocks
1« 0.50 0.17 0.40 0 21.9 0.231 46
-------
Table 6.
Sorption Wave Tank Test Data (Cont'd.)
Test
No.
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Broadcast
Rate
(ft3/min)
0.50
0.67
0.67
0.67
0.50
0.67
0.50
0.50
0.67
0.67
0.33
0.33
0.33
0.33
Slick
Thickness
(mm)
0.08
0.08
0.17
0.11
0. 11
0.11
0.11
0.34
0.34
SAE 30 Motor
0.08
0.08
0.11
0.23
0.08
Surface
Current Wave Contact
Velocity Height Time
(knots) (inches) (sec)
0.79
0.79
0.40
0.59
0.59
0.59
0.59
0.59
0.59
Oil Sorbed with
0.79
0.79
0.59
0.59
0.79
0
0
0
2
2
4
4
4
4
1/2
0
4
4
0
0
12.4
10.4
21.8
11.9
4.5
14.5
26.1
18.5
8.2
inch Polyurethane
16.4
7.9
15.8
22.4
9.6
Oil
Recovery Recovery
Rate Effectiveness
(gpm) (%)
0.235
0.239
0.251
0.289
0.343
0.314
0.290
0.550
1.140
Foam Cubes
0.252
0.339
0.404
0.582
0.284
47
48
50
58
69
63
58
37
76
50
68
81
58
57
-------
cubes and SAE motor oil, resulted in average recovery
effectivenesses of 76, 55 and 63% respectively. The
implications of this are obvious: (J.) cubes are more
effective than blocks in sorbing and (.2) heavier oil are
sorbed less effectively. Both these points were also
demonstrated on a bench scale as discussed in Section IV
of this*report. Basic wave tank test data indicates also
that waves and apparent contact times do not significantly
affect sorption under actual conditions, in fact, waves may
enhance sorption slightly. Anyway, the system composed of
the basic unit operations as presented in the previous
sections works.
Systems Specifications
The following elements must be included in a full scale
system:
(1) Broadcasting
(2) Herding
(3) Pickup and transfer
(4) Recovery
(5) Onboard transport
Based on the considerations discussed previously, cut-off
screw conveyor broadcasting of sorbent has been selected.
See Figure 11.
Sorbent herding will be achieved using existing barrier
materials held in place by floats and articulated boom
supports.
Sorbent pickup can be effectively achieved using a standard
wire mesh conveyor placed in the water at approximately a
30° inclination.
Traasfer of sorbent to the deck of the vessel will take
place with a similar conveyor.
Sorbent/oil recovery can most positively be achieved using
a converging, dual belt conveyor wringing device.
Onboard transport is best done using both a flight conveyor
(for high speed, excess lift capability) and a screw con-
veyor ({or basic aft-to-fore transport and connection to
the broadcast subsystem).
67
-------
Figure 12 on the following pages illustrates th.e total
system in plan view. Although all subsystems, are con-
strue table using off-the-shelf hardware, some special fab-
rication will obviously be required. This is especially
true for the pickup conveyor and the oil/sorbent recovery
device.
Broadcasting
Figure 13 is composed of preliminary drawings of the sor-
bent feed hopper, aft-to-fore transport conveyor and cut-
off screw conveyor broadcaster. The following list
identifies the basic elements of this subsystem:
(1) Storage and feed hopper
(2) Leveling screw flights
(3) Tines
(4) "Aft" screw conveyor
(5) U-trough discharges
(6) "Fore" screw conveyor
(7) Cut-off trough discharger
(8) Basic broadcasting screw conveyor
(9) Discharge trough
(10) Screw conveyor sprocket drive
(11) Variable speed drive motor
(12) Right angle transfer drive
(.13) Tines and leveling screw device
(14) Structural channel
CIS) Remote control panel
(16) Chain guard
(17) Sorbent removal chute
68
-------
MMN DECK
I
Wr. ____T
SORBENT OIL RECOVERY SYSTEM LAYOUT
Figure 12.
69
-------
v_ »V_.,. V ^l^tA^^^=
wev/ c-c
SECTION JB-& 3CM.E J
3IDC CLeVAT/ON \H£\J
71
SORBENT BROADCAST SYSTEM
Figure 13.
-------
Herding
The overall barrier subsystem for sorbent herding is ad-
equately presented in the sorbent oil recovery system
layout drawing (.Figure 12). Some details of this sub*
system are not, however, readily apparent from this over-
all view. These are clarified in the following paragraphs
Details of the herding boom-support arm interface are
presented in Figure 14. The relevant descriptors for this
preliminary drawing are:
(1) Float
(2) Shackle
(3) Pin
(4) Herding boom support arm
(5) Post
(6) Belting
(7) D-ring
(8) Herding boom
(9) Cover
The herding boom support arm is detained in Figure 15:
(1) Plate (8) Shaft
(2) Pipe (_9) Disc
(3) Plate Q.O) Tube
(4) Plate (_n) Eye
(5) Support (12) Collar
(6) Spacer (_13) Boit
(7) Bracket
73
-------
HERDING BOOM-SUPPORT ARM INTERFACE
Figure 14.
WATER
I
75
-------
I I
HERDING BOOM-SUPPORT ARM
Figure 15.
77
-------
The herding boom must also interface with the pickup
conveyor in an articulated fashion. Figure 16 details
this interface with the following key:
(1) Bar
(2) Support
(3) Wall
(4) Skirt
(5) Herding Boom
(6) Catch
C71 Bolt
C8) D-ring
C9) Belting
CIO) Bolt
(11) Flat washer
(12) Hex nut
Pickup and Delivery to Recovery Unit
Figure 17 details the pickup and delivery conveyors which
interface to each other and the herding boom and oil/
sorbent recovery device respectively. The following items
comprise this subsystem:
(1) Channel
(2) Angle
(3) Wire Belt
(4) Wire Belt
(5) Gear
(6) Gear blank
(7) Adjustable speed drive
(8) Sprocket
(9) Chain
CIO) Motor
(11) Rubber pad
(12) Rod
(13) Shaft
(14) Shaft
(15) Pin
(16) Pipe
(17) Plate
(18) Bolt
(19) Bolt
(20) Shaft
(21) Shaft
(22) Shaft
(23) Shaft Coupling
(24) Motor support
(25) Guard
C26) Small guard
{.27) Bottom plate
C28) Scraper
C29) Cover
C30) Cover
(31) Pad eye
(32) Flange bearing
79
-------
HERDING BOOM-PICKUP CONVEYOR INTERFACE
Figure 16.
V\EW A-A
\
A
\
A
WATER LINE
cue new
WATER '-IN.
YICV
81
-------
RE t O >J F P. . Co 'J V T V5R
SORBENT PICKUP CONVEYOR AND DELIVERY CONVEYOR
Figure 17.
83
-------
The pickup conveyor must be supported from onboard the
vessel with a type of platform. This is illustrated in
Figure 18 on the following pages. Cthe detail items called
out refer to the numbers on the above list).
Oil/Sorbent Recovery
The dual converging belt oil/sorbent recovery unit is
detailed in Figure 19. The basic, off-the-shelf components
which comprise this subsystem are numbered on this pre-
liminary drawing:
(1) Main Conveyor Belt
(2) Feed box
(3) Head pulley
(4) Tail pulley
(5) Take-up pulley
(6) Electric Motor
(7) Tail pulley
(8) Brackets
(9) Heavy duty shaft
(10) Speed reducer
(11) Roller Chain
(12) Roller Chain
(13) Idler sprockets
(14) Side plate
(15) Idler sprocket
(16) Hinge shaft
(17) Belt drive
(18) Pressure belt
(19) Oil sump
(20) hose coupling & shut-off valve
(21) Main frame
(22) Drive sprockets
(23) Drive sprocket
(24) Support angle
(25) Pillow blocks
(.26) Discharge chute
(.27) Chain guard
(.28) Head pulley
Note that a key in the design of this unit is maintenance of
equal belt speeds (.to minimize sorbent degradation) while
still allowing the wringer belts to move apart in the
presence of "excess" sorbent or foreign objects.
85
-------
SORBENT PICKUP CONVEYOR SUPPORT
Figure 18.
87
-------
OIL/SORBENT RECOVERY UNIT
Figure 19.
L_ _ IT
^ , , , _^_
t-
-------
Recovery-to-Broadcast 'and Aft-to-Fore Transport
Transfer of sorb;ent from the oil/sorbent recovery sub-
system can most readily be achieved using a simple flight
conveyor. Figure 20 illustrates such a transfer conveyor.
The following detail applies:
(1) Wire belt
(2) Main support frame
(3) Receiving hopper
(4) Canvas cover
(5) Housing
(6) Drive motor
(7) Sprocket drive
(8) Belt head shaft
(9) Discharge chute
(10) Canvas Cover
(11) Upper support bracket
(12) Retaining chain
(13) Lower support bracket
(14) Belt tail shaft
(15) Chain guard
Aft-to-fore soirbent transport is accomplished by screw
conveyance as has already been described in this section,
under Broadcasting (See Figure 13). This unit -operation is
essentially integral to the broadcast unit operation.
Sorbent Selection
The sorbent for use in the oil recovery system outlined
above has essentially been selected in previous sections of
this report. To reiterate:
91
-------
TRANSFER CONVEYOR
Figure 20.
SIDE.
93
-------
(1) The sorbent must be oleophilic, not degrade upon
handling and wringing, and must be capable of ex-
tensiye reuse (.1 • e . , sorption through 100 to 1000
cycles). Polyurethane foam Cester-type) complies
with these requirements in excellent fash.ion.
(2) Sorbent particles must be sufficiently small to
promote rapid oil sorption (.with contact times of
15 to 90 seconds) but must be large enough to still
be handleable. In isolation, high surface area to
volume ratios for sorption make sense. With slick
buildup and headwave formation in a herding channel,
however, minimum surface area to volume is accept-
able. Handleability and movement of sorbent mass are
enhanced by using equidimensional particles.
Section IV, Particle Size and Shape, demonstrated the
adequacy of 1 inch sorbent cubes for a full scale
system. This sorbent size and shape promotes maximum
sorption efficiency, system effectiveness and
sorbent handleability.
(3) Compared to other factors which effect oil recovery
using sorbent materials, pore size is of relatively
minor importance (see Particle Size and Shape,
Section- IV). Foam with 100 pores per linear inch is
entirely adequate for oils ranging from SAE 30 motor
oil to marine diesel fuel. It may be of benefit,
however, for very viscous oils or extreme cold
weather operations to have coarser foam available:
20 PPI foam is suggested.
Conclusions
A preliminary design for a complete system to recover
spilled oil using sorbent materials has been presented.
The system includes unit operations for sorbent broad-
casting, sorbent/oil mass herding and pickup, oil/sorbent
recovery and onboard transport (.transfer and aft-to-fore).
The proposed system could handle a range of sorbent materi-
als with varying degrees of success. However, based on
other considerations, commercial polyurethane Cester-type)
foam with 100 pores per linear inch C2Q PPI alternate) and
cut in 1 inch cubes is included as the system's basic
sorbent material.
95
-------
SECTION X
VESSEL REQUIREMENTS
The ideal boat to meet the requirements for the sorbent
oil recovery system is an offshore oil supply boat ranging
in length from 125-200 feet. Such, boats have the speed,
size, maneuverability, depth, deck space, personal accom-
modations and free board required to meet th.e specifications
outlined in the RFP and generated throughout the study.
Introduc tion
Boats meeting the following specifications are satis-
factory for use with the proposed sorbent oil recovery
system:
Length ... 80-200 feet Width ... 20-40 feet
Draft .... 2-4 feet Displacement..50-400 ton
Propulsion .. low speed, Twin screw .. preferably
preferably diesel variable pitch
Maneuverability Single screw .. large
rudder ships acceptable
HP 200-2000 Speed ... 12 knots
Endurance... 8 hours Deck Space .. 1000 sq ft
Stability... accept 10,000 Accommodation .. 8 men
Ibs of deck weight in
combinations Free Board .. 2-4 feet
Electrical Power ...
preferably 110 or 220 volts,
AC and approximately 20 HP.
A large uninterrupted deck area is ideal for the oil re-
covery system. A lateral passage on the side of the ship
is required for the screw conveyor moving dry sorbent to t,he
broadcast nozzles. In ships not ha-ving this free area, it
should be possible to cut through deck houses or bulk heads
to accommodate the system.
Av a i1a b1e Ships
The following ships meet the sorbent recovery systems
specifications.
97
-------
Offshore oil supply boats, 125' to 200'
Powered oil barges 80' or more
Commercial construction tenders 80' or more
Small commercial tugs 100' or more
Fishing boats 80' or more
USCG 100' or larger buoy boats
USCG 100' or larger inland buoy tenders
USCG 115' river buoy tenders
USCG 75' or larger inland construction tenders
Army or Seabee construction or special purpose
craft 80' oz more
Operations
Not only will the ship operate primarily during daylight
hours but it will make long passes to scoop the heart out
of an oil slick. Normal procedure will be to stream up or
down the current due to the fact that oil slicks orient
their long axis parallel to the direction of current flow.
Operating at full design efficiency of 3,000 gallons per
hour, a maximum of 80 tons of oil would be taken on board
per shift. Due to the patchiness and variability of oil
slicks thickness during most operations only 10-20 tons of
oil may actually be recovered.
Deployment of Recovery System
Once a vessel from which to deploy the recovery system has
been chosen, booms and conveyor belts would be mounted on
one or preferably both sides of the ship. The other equip-
ment would be installed on the deck and the oil storage
capacity for the oil would be maximized. Th.e recovery
ship would then move at maximum speed to the slick area
where the booms and channels would be deployed followed by
system operation.
Since this system is designed to operate at 2 knots where
maneuverability is a problem for all ships, twin screws
with variable pitch are preferred. The system is designed
to operate at up to 4 knots and will be effective in re-
98
-------
covering lighter oils at these higher speeds.
While wind is expected to be a probleju at th.e low recovery
speeds, two foot waves are not. Normal turns should not
result in sorbent losses. However, sharp turns or backing
will result in sorbent loss. Obviously, sorbent dispers-
ment must be stopped and the sorbent either recovered from
the channel before a maneuver is begun or at least all the
sorbent must be kept in the channel during the maneuver.
Oil Storage
Oil should be stored, when possible, onboard the ship, in
towed tanks or in towed bags rather than in a tanker lashed
alongside. This would allow the recovery system to be
deployed on both sides of the vessel, thereby doubling the
capture area. It would also increase maneuverability and
speed of deployment.
The problems of onboard storage are cited in Appendix A,
Page 147and are mainly centered around the fact that
200,000 gallons of 670 tons is difficult to store in the
recovery vessel itself. While that is true, it must also
be remembered that to recover that much oil would require
100-200 operating hours for a sorbent system. The problem
of towed bags not being liftable out of the water can be
solved by having barges go to the bags and pump them out
until they can be lifted from the water. In an oil
recovery project the most important thing is to get the
slick contained quickly before it disperses into'a
relatively unrecoverable state.
Conclusion
The sorbent system recovery requirements can be met by a
number of ships in the private and public domain.
99
-------
SECTION XI
PRELIMINARY PERFORMANCE ESTIMATES,
In t r o due t'ton
The previous sections of this report led from basic con-
siderations of the phenomena and operations involved in
sorbent recovery of spilled oil to wave tank testing.
Based on the results of these tests and studies, consider-
ation can be given to the preliminary design of a full
scale system. The components of this f-ull scale system
were specified on the basis of the results of the test
program to comply with, the overall program goals. At
steady-state each of thes-e, jcomponent.s must operate in
sequence with each of the other components. Their gross
performance is, therefore, interdependent. This section
recaps the performance estimates -of the components and re-
presents their performance in terms of the overall system.
Furthermore, preliminary figures are presented on system
capital and operating costs. As will be seen, these
figures not only demonstrate a reasonableness for final
system development but also provide a possibility of an
economic payout for the system based on oil recovered.
The latter could prove very helpful in promoting widespread
adoption of the proposed approach.
The presentation below does not attempt to preclude more
refined estimation of system performance during final de-
sign. Rather, feasibility is demonstrated and a brief
framework presented within which final design estimates
can be made.
Gross Oil Pickup
Oil pickup and recover is limited substantially by the
amount of oil which can be swept into and contained by the
herding boom of the proposed system as well as by the
po ssible rate of sorbent throughput within the herding
channel. Sorbent throughput relates critically to the
possible .effectiveness of the system in recovering the oil
Which is swept. The efficiency of the system Ci.e., the
relative degree to which sorbent is saturated with oil), on
the other hand, relates more critically to the other system's
parameters.
101
-------
Oil Throughput
At a design speed of 2 knots with a 16 foot boom mouth and
a 1.5 mm slick of no. 4 fuel oil, about 7000. gallons of oil
will be swept into the herding channel each hour. The wave
tank test data CSection IX, Table 6) indicates that, de-
pending on the exact operation of the system, anywhere from
50 to 100% of this oil will be picked up: 3500 to 7000
gallons per hour. It has been remarked previously that
proper system operation (.e.g., building up a sorbent/oil
"head waveM)enhances sorbent efficiency. Therefore, under
design conditions the system will be very effective (_±.e.,
will recover the required amounts of oil) whereas sorbent
efficiency may vary widely within the range of effective-
ness indicated. For preliminary purposes;
System effectiveness:
3500 to 7000 gallons recovered/hour
System efficiency:
50 to 100% of swept oil recovered.
Now, the amount of oil swept by the system is proportional
to vessel speed. Recovery, therefore, could almost be
double the nominal values called out above since the maximum
system speed (based on the hydrodynamic wave tank tests) has
been estimated to be about 4 knots. However, the sorption
efficiency could be low at higher vessel speeds if little
or no sorbent bed can be maintained in the herding channel.
This depends on a multitude of factors including (a:id per-
haps most importantly) the competence of the operating
crew. Assuming "proper" operation efficiencies could range
from 35 to 10-0% at the maximum vessel speed depending on
the presence or absence of waves, nature of the oil, etc.
Oil Type
Both the wave tank tests (.Table 6) and benchwork (.Section
IV, Oil Viscosity and Surface Tension) considered system
and sorption with different oil types. For a physical
system as that being discussed here, oils can be character-
ized on the basis of viscosity, density and surface tension.
In general,
(1) Higher viscosity oils sorb less readily.
(2) Dense oils sorb less readily.
(3) High surface tension oils sorb more readily.
102
-------
The results of Sections IV, Oil Viscosity and Surface
Tension, and Sorpti.on to Less Than Saturation; and Section
IX can be combined during final design work to give esti-
mates of performance as a function of different physical
parameters. Note that the interaction of oil and water
Ce. g. , emulsibility). is not covered by the above parameters.
In some cases, for instance highly aromatic oils, this
could lead to somewhat faulty estimates. Temperature
effects are also not included explicitly In the above; in
terms of the oil, however, they are included implicitly
since the three physical parameters of interest are functions
of temperature.
Sorbent Losses' and Degradation
A very favorable result of the study under discussion is
that, for the proposed system, sorbent losses and degra-
dation are of essentially negligible consequence. Sorbent
degradation was not observed during either bench (Section
VIII, Sorbent Recycle and Sorbent/Oil Recovery) or wave
tank tests (Table 6). Likewise, sorbent losses (given
broadcast within the herding channel and not ahead of or
in the boom mouth) were negligible except under conditions
of very careless system operation - using the non air
broadcasting system. A liberal estimate of combined de-
gradation and losses would be an effective 1% loss of
sorbent every 100 cycles (5 to 10 hours of operation)
through the system. Depending on the relative importance of
degradation and actual losses, adjustment for this effective
loss could be either just make—up with fresh sorbent or com-
bined sorbent blow-down and make-up. Conservatively, the
proposed system would, therefore, be expected to require
only about 2 bulk cubic feet of sorbent make-up for every
8 hours of operation.
Preliminary Cost Estimates
The following two sections, Capital Costs and Operating
Costs, summarize preliminary estimates of capital and
operating cost estimates. Field demonstration of the pro-
totype oil recovery system is not considered although it
would be desirable. It should be kept in mind that capital
costs for additional systems would be considerably reduced
since all engineering design would have been completed on
the prototype development. Some elaboration is presented on
the operating cost estimate to give an indication of cost
effectiveness.
103
-------
Capital Costs
Labor, material and .miscellaneous costs represent tb_e
capital necessary to design and construct a full scale
prototype oil recovery system.
Labor costs include engineering, construction, supervision,
overhead, G & A, fees and -major subcontracts.
Material costs include principle equipment items plus
minor materials necessary for construction.
Miscellaneous costs include reports, engineering repro-
duction, travel, per diem, etc.
Subsystem
Herding/Pickup
Subtotal
Broadcast
Subtotal
Oil Extraction
Subtotal
Transfer
Subtotal
Sorbent Material
Element
Labor $42,500
Material 16,500
Miscellaneous 1,000
Labor $39,500
Material 11,300
Miscellaneous 500
Labor $56,200
Material 12,500
Miscellaneous 1,500
Labor $ 4,000
Material 900
Miscellaneous 100
Estimated Cost
Labor $ 1,500
Material 1,000
Miscellaneous 500.
Subtotal
Power Generator
Flatbed Trailer
Shipping and Freight Charges
TOTAL
$ 60,000
$ 51,300
$ 70,200
$ 5,000
$ 3,000
$ 10,000
$ 7,000
$ 10.000
$216.500
Operating Costs
Assuming 2 recovery operations each at 10 days (.10 hours/
day), plus 2 days for start-up and shut-down each year, the
following analysis holds assuming a nominal oil recovery
rate of 5,000 gallons/hour.
104
-------
Charge Yearly Cost
Work boat rental C$3.00/day x 22 days). $ 6,600
Work boat crew C$2007day x 22 days) 4,400
Dockside labor C50 hours @ $10/hour) 500
Oil storage barge rental C$200/day x 22 days) 4,400
System transport (4 trips @ $50jO) 2,000
Sorbent make-up 100
Maintenance (1% of capital cost) 2,200
Burdened recovery system labor (3 men @
$100/day x 22 days) 6,600
Field costs (per diem @ $1,500, travel
@ $1,200) 2,700
Miscellaneous field expenses 2,000
Recovery system charge (20 year payout at 8%) 25,000
System management (30 days @ $100/day) 3,000
TQTAL $59.500
Now, the total oil recovered by this assumed operation is:
Q = 5,000 x 10 x 20 = 1,000,000 gallons
The recovery cost is, therefore; about 6<: per gallon which
is very feasible compared to fuel oil costs (about 12 to
20<:/gallon depending on grade) and just about a break even
based on crude oil values (ten to fifteen dollars per ton),
Given the assumptions involved, it seems probable that the
full scale recovery system as presently conceived could
actually pay for itself based on the salvage value of re-
covered spill oil. This becomes even more likely if
service charges are assessed for the use of the system
based on avoidance of environmental impact.
Conclusions
In terms of the above discussion, the following have been
demonstrated:
(1) In line with the design goals, the proposed system
will be highly effective: 3,500 to 14,000 gallons/
hour oil recovery.
(2) Depending on operating conditions, the system will be
from 35 to 100% efficient.
(3) In addition to the above, the capital and operating
costs for the system appear to be reasonable:
$216,500 and $59,500 per year respectively.
105
-------
(4) Also, there appears to be a chance that such a system
can be economically self-sufficient based on the value
of the oil spilled.
106
-------
SECTION XII
APPENDICES
107
-------
APPENDIX
A
109
-------
AFFILIATE OF UNION CARBIDE CORPORATION AND THE SINGER COMPANY
11440 ISAAC NEWTON INDUSTRIAL SQUARE NORTH, RESTON, VIRGINIA 22070 TELEPHONE (703) 471-1310
TELEX 899442
May 2, 1972
Meloy Laboratories, Inc.
6631 Iron Place
Springfield, Virginia 22150
Attention: Mr. Garth Gumtz
Subject: Subcontract A3669 under EPA Contract No. 68-01-0068
Enclosure: (1) Preliminary Design of an Oil Recovery System
Using Sorbents
Gentlemen:
Ocean Systems, Inc., submits enclosure (1) in accordance with Article in,
Reports, of the subject contract. The five (5) drawings called out in the
table of contents are not included with this submission as they were delivered
previously.
The budgetary price estimate called out in the scope of work will be delivered
on May 3, 1972. At that time all work required by the subject contract will
have been accomplished.
It has been a pleasure working with you on this program and we look forward
to the possibility of working with you in the near future.
Very truly yours,
OCEAN SYSTEMS, INC.
Dale E. Bellovich
Contract Administrator
DEB:ds
cc: R. Pasquale
110
OFFICES: NEW YORK. N.Y. • WASHINGTON. D. C. « TARRYTOWN, N.Y. • RIVIERA BEACH, FLA. • MORGAN CITY. LA.
SANTA BARBARA, CALIF. • SEATTLE, WASH. • LONDON, ENGLAND • HOUSTON, TEXAS • NASSAU, BAHAMAS • STAVANGER, NORWAY
-------
A JOINT MELOY LABORATORIES AND OCEAN
SYSTEMS, INC. PRELIMINARY DESIGN
OF AN OIL RECOVERY SYSTEM
USING SORBENTS
Submitted by
Ocean Systems, Incorporated
11440 Isaac Newton Industrial Square, North
Reston, Virginia 22070
May 2, 1972
In response to Subcontract A 3669 to
Meloy Laboratories Under EPA Contract 68-01-0068
111
-------
TABLE OF CONTENTS
Section
Page
1 INTRODUCTION 113
2 GENERAL SYSTEM DESCRIPTION 114
3 PRINCIPAL SYSTEM DIMENSIONS AND 119
ASSUMPTIONS
4 COMPONENT PRELIMINARY DESIGN DESCRIPTION 124
AND RATIONALE
4.1 Herding Boom 125
4.2 Recovery and Delivery Conveyor 136
4.3 Recovery Vessel 141
4.4 oil Storage and Sorbent Disposal 147
4.5 oil Transfer System 152
4.6 Flotsam Fence 153
Applicable Original OSI Drawings Delivered Separately
Title
Sorbent Oil Recovery System Layout -
Herding Boom-Support Arm Interface -
Herding Boom-Support Arm -
Herding Boom-Recovery Conveyor Interface
Sorbent Recovery Conveyor and Delivery
Conveyor, Sheets 1 and 2 -
Size
E
D
D
D
Number
001569
001570
001571
001572
001573
112
-------
SECTION 1
INTRODUCTION
Ocean Systems, Inc./ in accordance with Subcontract
Number A3669 to Meloy Laboratories, under EPA Contract 68-01-
0068, was tasked with developing a preliminary system design
for harvesting sorbents, and developing the vessel or platform
and the storage of 200,000 gallons of oil and the disposal of
the sorbents used in the recovery system. In addition, Ocean
Systems, Inc. provided technical support to Meloy Laboratories,
Inc. in the overall development of a preliminary design of a
system for recovering oil from the surface of protected waters
using sorbents. The interim report submitted in January 1972
covered the overall progress and results of Tasks 1, 2, 3, and
4 which included model tests in the OSI test tank under various
current and sea conditions.
The overall objective of the oil recovery system was
to provide a system capable of recovering 1500 gal/hr of oil
with 10% or less water content, with the potential to increase
this to 3,000 gafl/hr. Oil to be recovered will have a viscosity
range from that of light diesel oil to near water density of
heavy asphalt at 20°C.
The'interim report covering Tasks 1, 2, 3, and 4 has
been used in completing Task 5, which is the preliminary design
of the harvesting sub-system, the deploying vessel, the storage
of 200,000 gallons of oil and the disposal of the sorbents used
in the recovery system. These elements have been interfaced and
integrated with the preliminary design of the sorbent broadcasting,
oil sorbent separation and sorbent reuse sub-systems which were
developed by Meloy Laboratories.
113
-------
SECTION 2
GENERAL SYSTEM DESCRIPTION
The complete oil recovery system consists of the
herding boom, the sorbent recovery conveyor, the delivery
conveyor, the oil extraction unit, the transfer conveyor,
the feed hopper, the broadcasting conveyor, the vessel and
the oil storage system. These are shown on Drawing 001569.
The general description of the herding boom, the
recovery and delivery conveyor, the recovery vessel, the
oil storage and sorbent sub-system, the oil transfer system
and the flotsam fence will be covered in this section. The
detailed description of these components is found elsewhere
in this report.
Herding Boom
The herding boom consists of a triangular floating
surface section which is one foot above the waterline and
a triangular subsurface.section which is two feet deep.
The latter absorbs the water and acts as a "dynamic keel".
The combined action of the two sections causes the herding
boom to conform closely to the surface of the water. .The
herding boom gathers and contains the sorbent cubes which
absorb the oil as they travel toward the recovery conveyor.
The forward end of the herding boom is attached to a float
that is maintained approximately 16 feet from the vessel,
by a rigid arm. The after end of the herding boom is attached
to the outboard side of the recovery conveyor which is four
feet from the side of the vessel. The herding boom gradually
converges to a distance of four feet from the ship and continues
at this distance until it joins the recovery conveyor.
Recovery and Delivery Conveyors
The recovery conveyor picks up the oil soaked
cubes and deposits them to the delivery conveyor which in turn
delivers them to the oil extraction unit. The conveyor consists
114
-------
of a four foot wide-stainless steel flexible wire belt driven
by a variable speed electric powered unit. It is 23 feet
long and pivots on a rotating foundation with two degrees of^
freedom. The conveyor terminates two feet below the waterline
and is secured to the side of the vessel with an adjustable
hogging line.
The delivery conveyor is constructed in a similar
manner, but is driven with a constant speed electric drive.
It is 13 feet long and is mounted on a foundation with two
degrees of freedom.
Recovery Vessel
The recovery vessel must be large enough to accom-
modate the equipment and long enough to permit the recovery
system to operate effectively. At the same time the draft
must riot be excessive so that it can operate in shallow waters.
These are conflicting requirements. In addition, the vessel
must be capable of operations at low speeds, i.e., about 1.5
to 2 knots.
The oil recovery system is approximately 60 feet
long. An additional 10 feet is required at each end for
installation and working space. Therefore, the minimum
length of the ship is 80 feet. To accommodate the equipment
and to provide operating space, the vessel must have at least
1000 square feet of clear deck area. Most of it should be
aft since most of the equipment will be located there.
The vessel also must have a clear area inboard
to accommodate the broadcasting conveyor. Normally passage-
ways alongside the deck house will be adequate. If there are
no passageways, deck compartments can be cut out to accommodate
the equipment.
115
-------
A number of vessels of opportunity can be used to
deploy the system. One of the best vessels is the offshore
supply ship which has a long uninterrupted deck area and a
relatively shallow draft. The characteristics and type of
vessels that are suitable to act as deployment vessels are
given in Section 4.3.
Oil Storage and Sorbent Disposal
Oil Storage
Oil storage is divided into two parts. The
first is oil storage during recovery operations and the second
is storage of the oil ashore or afloat after recovery.
The storage of oil during recovery operations
aboard the recovery vessel is not feasible except in the case
where a powered tanker barge is used. A small oil barge will
normally be used to receive the recovered oil. The barge will
be moored alongside the vessel during operations. When it is
full or if the draft becomes excessive, it would be replaced
by another barge. In some cases the barge may be kept along-
side until the recovery operations are completed. The storage
of 200,000 gallons of recovered oil would require a relatively
small barge, i.e., 670 tons.
fStorage of the recovered oil after the operation
is best accomplished afloat in an oil barge that can be moored
near to the scene of operations. In nearly all instances, the
storage of the contaminated oil at available shore facilities
is not possible because this oil would contaminate the instal-
lation. The contamination of an oil barge is acceptable
because it can be readily cleaned after the recovered oil
is transferred to an appropriate oil processing activity.
A full discussion of oil storage appears in
Section 4.4.
116
-------
Sorbent 'Di'sp'o's'al
Sorbents will be stored in strong polyethylene
bags. These can be utilized for storage of used as well as
new cubes. The bags have the advantages of low cost/ light
weight, resistance to oil/ occupy very little space when
empty and can be readily stored in irregular spaces.
It is recommended that used cubes be stored
and reused unless operational experience proves otherwise.
Oil Transfer System
It will be necessary to pump the recovered oil
from the sump of the oil extraction unit to the barge along-
side.
The pump selected was a Blackmer Multi-Vane GXS2-1/2
pump driven by a 5 HP electric motor.
The hose that will be used is a Heliflex 2-1/2 inch
hose.
Both the pump and the hose are suitable for handling
petroleum products.
A full description of the oil transfer system
appears in Section 4.5. oar'-
Flotsam Fence
A flotsam fence will be required in areas where
flotsam«=-can be introduced into the recovery system.
The flotsam fence is made of Flat-Flex wire belt
3/4" mesh x .092 wire. The fence will be 24 inches wide of
117
-------
which 8 inches will be above the water and 16 inches will
be beloyr^the water.
A full description of the flotsam fence appears
in Section 4.6.
118
-------
SECTION 3
PRINCIPAL SYSTEM DIMENSIONS AND ASSUMPTIONS
To arrive at a preliminary design for the oil
recovery system using sorbents, certain initial base values
or assumptions were made. These assumptions were based on
data obtained during the earlier phases of the contract.
Specifically, the following assumptions were made:
Sorbent size = 1 inch cubes
Sorbent loss = .5% (per cycle)
Slick Thickness = 1.5 mm (#4 fuel oil @ 20°C)
Initial calculations indicated that a volume flow
) bulk ft /min should satisfy t]
minimum of 1500 gal/hr of oil recovered.
rate of 20 bulk ft /min should satisfy the requirement of a
In determining the principal system dimensions
the following two equations were used. These equations
predict the quantity of oil recovered as a function of
several parameters .
[1] Qp = (C) (W) (h) (k^ (k2)
12] QQ = Qp (^ U2) U3)
Where Q,, = Sorbent flow rate down channel (ft /min)
r
C = Current velocity (ft/min)
W = Channel width (ft)
h = Cube height (ft)
k, = Area density ratio of cubes in the channel
k,, = Ratio of sorbent velocity to current
velocity
Q_ = Oil pick-up, rate (ft /min)
£, = % sorbent retained after each cycle
£2 = % of oil and water recovered by volume
C = % of oil recovered by volume
119
-------
Equation [1] gives the sorbent flow rate down the
channel and up the conveyor. Equation 12] gives the oil pick-up
rate as a function of the sorbent flow rate. The parameters
C, W, and h are fixed with the remaining parameters a function
of the operating conditions.
Specifically, the variable parameters k,, k~, £,/
£~, and £, are a function of the channel geometry, current
« -J
speed, sea state, oil properties, sorbent properties, and
the contact time.
The hydrodynamic tests conducted by Meloy Labs, Inc.
and Ocean Systems, Inc. in Ocean Systems' wave and current
tank were accomplished in order to gain an understanding of
the functional relationships of the parameters involved. With
these relationships known it is possible to determine the
physical dimensions of a system which will provide a
specified oil recovery rate.
The interim report documented the results of the
hydrodynamic tests and these results are used to determine
the physical dimensions of the system. Specifically,
all the parameters above are determined for a desired oil
flow rate using the data contained in the interim report.
In determining the physical dimensions of the
sorbent recovery system, some initial basic assumptions were
made based on the experience gained during the testing phase.
These assumptions fixed specific parameters in order to
determine the remaining parameters, i.e.,
h = 1 inch (.0833 ft) > (Specified by
QF = 10 ft /rain (20 bulk ft /min) J Meloy
k = 5
1 • * k(Estimate from
k~ = .3 J interim Report)
120
-------
From Equation II]
(C) F
v l
. . . . _ —
(h) (k.^ (k2) .0833 x .5 x .3
C (W) = 780
C1 (W) = 7.9 (C1 = current speed in knots)
The above relationship indicates that for a channel
width of 4 feet, the current velocity must be equal to or
greater than 2 knots to provide a foam flow rate of 10 ft /min.
In determining the oil flow rate the values of S-ifSo
£3 must be determined. Based on the results of the test program
£, = .995 and £3 = .95 by assumption. £-' however, is
dependent on the slick thickness in the channel, contact time,
and the oil's properties. For #4 fuel oil, Figure 2.3.5 of the
interim report indicates the slick thickness in the channel is
on the order of 5 times the undisturbed slick thickness for a
model current speed of approximately 1/2 knot. This represents
a prototype current velocity of approximately 1 knot. For a
prototype current velocity approaching 2 knots the ratio of the
channel slick thickness to the undisturbed slick thickness will
be greater than 5. Meloy Labs., .Inc. has determined that for
slick thickness in the order of 4.5mm or greater (a ratio of 3)
that 100% sorbtion of oil will occur in less than 5 seconds.
(#4 fuel oil @ 20 °C) . Selecting a channel length of 20 feet,
the contact time in the channel will be a minimum of 6 seconds
with the sorbent particles flowing at the same speed as the
current for a maximum current speed of 2 knots. This insures
that 100% sorbtion will occur resulting in £2 equaling approxi-
mately 1.0. In reality, the sorbent velocity will be less than
the current velocity, thus increasing the contact time and further
insuring that 100% absorption occurs. Substituting these
values back in Equation [2] we have:
Q0 = QF (?x) U2) U3)
121
-------
with QF = 10 ft /min (maximum)
q = .995
£2 = =95
£3 = i.o
QQ = 9.45 ft3/min
= 70.5 gal/min
= 4230 gal/hr (maximum)
These calculations indicate that a channel width
of 4 feet and a channel length of 20 feet will be more than
adequate to provide an oil recovery rate of 3000 gal/hr for
#4 fuel oil @ 20°C at a vessel speed of 2 knots.
In the above calculations the values for k, and k2
were based on results contained in the interim report. The
value k, represents the percentage of water surface area covered
by the sorbent particles in the channel and k« represents the
ratio of the sorbent velocity to the current velocity. Figure
2.3.7 of the interim report indicates that k_ will
equal 0.3 for geometries where the mouth width of the boom is
4 times the channel width. This represents a prototype system
with a mouth opening of 16 feet and a channel width of 4 feet
for #4 fuel oil. The results of Figure 2.3.7 were obtained
from tests where the model current speeds were 1/2 knot or
less representing prototype current speeds of 1 knot or less.
It was noted, however, that for higher current speeds the
value of k2 increases indicating that the estimated value of k_
is a minimum and in practice should be greater. The value
of k, is difficult to predict but on the average it will be
equal to or greater than 0.5.
122
-------
With the values of k, and k» being minimums/ and
the oil pick-up rate (Qn) being larger than required, the
ship's speed or current speed required for the desired oil
recovery rate will be equal to or less than 2 knots. In
addition a mouth width equal to 4 times the channel width
is desirable.
In summary, the system with the following
characteristics should provide an oil recovery rate greater
than 3000 gal/hr with §4 fuel oil at 20°C providing that
sufficient oil is available to be recovered.
Channel width = 4 feet
Channel length = 20 feet
Mouth width = 16 feet
Ship's speed range = 1 to 2 knots
123
-------
SECTION 4
COMPONENT PRELIMINARY DESIGN DESCRIPTION AND RATIONALE
124
-------
SECTION 4.1
HERDING BOOM
The most effective means for "herding" both sorbents
and oil is through the use of an oil containment barrier.
An oil containment barrier designed by Ocean Systems, Inc.
is particularly suited to the proposed system and has been
chosen as the most viable herding boom.
The barrier concept is based on the use of flexible
polyurethane foam, with a "dynamic keel" that imparts high
static and dynamic stability. The barrier consists of a
non-water-absorbing foam package that provides buoyancy and
a surface barrier and a water-absorbing foam package that
provides a submerged barrier and serves as a "dynamic keel".
The two packages are connected into an integral unit forming
the barrier. Figure 1 shows the construction of the barrier.
The unique feature of the barrier is the achievement
of stability and buoyancy utilizing readily available seawater.
Although the "dynamic keel" is essentially weightless in water,
the large displaced volume of the trapped water and the large
waterplane area provide excellent sea conformance characteristics.
Both the top and bottom sections have triangular
shaped cross sections. The top section is an equilateral
triangle with a one foot altitude and the bottom section an
isosceles triangle with a two foot altitude. The triangular
shape offers a large waterplane area which is essential for
good dynamic response. Also, the triangular cross section is
an optimum structural shape and is relatively easy to fabricate.
The physical dimensions reflect both functional and operational
requirements.
The flotation section is constructed
of flexible polyether foam to which is applied a flexible
urethane elastomer coating. The resultant structure is flexible,
125
-------
BARRIER SCHEMATIC
URETHANE ELASTOMER COATING
POLYETHER FOAM
ELASTOMER LAYER
TRENGTH BELT (DACRON)
.ETICULATED POLYESTER FOAM
CENTER
OF
PRESSURE
CORING BELT
-------
highly buoyant, and has a very high resistance to oils and
abrasion.
The bottom section is constructed of a specially
treated reticulated polyester foam, which absorbs water and
serves as a "dynamic keel". Polyester foam is highly flexible
and has very good resistance to oils. The bottom section is
joined to the top section with the same type of urethane
elastomer used for the top coating.
The structural belt that runs longitudinally in the
center of the lower section is made of dacron webbing. Dacron
was selected for the belt material because of its high strength
and low stretch properties. Low stretch in the belt material
is required to reduce the amount of differential elongation
between top and bottom sections of the modules when the barrier
is under axial load. The structrual belt is joined to the
foam bottom sections with urethane elastomer.
The cross sectional properties of the boom are as
follows:
Dimensions
L =12 inches
6 = 30 degrees
t = 40 mil (avg)
t CTYP)
127
-------
Weight
Dry weight = 14 Ibs/ft
Wet weight = 75 Ibs/ft
(bottom section fully saturated)
The most important feature of the "dynamic keel"
barrier is its ability to conform to the sea surface and
thereby channel the sorbents. Figure 2 shows the heave
response for full scale wave heights up to 3 feet based on
model tests and theoretical calculations. It can be seen in
the figure that the heave phase angle is in good agreement
with theory while the heave response operator is not. The
disagreement in heave response is thought to be primarily
due to the assumption of linearity in the theory. The model
tests results indicate that the heave response of the barrier
is much better (with respect to surface conformance) than
predicted by theory.
For protected waters with two foot seas, wave periods
lie mostly between 3 and 4 seconds or wave frequencies between
1.5 and 2 radians per second. Near this end of the scale the
"dynamic keel" closely follows the sea surface as Figure 2
indicates enabling sorbent "herding" with little or no loss
due to sorbents washing over the surface barrier.
The interface between the herding boom and the
sorbent recovery conveyor is critical to the system performance
as far as sorbent/loss is concerned. OSI Drawing # 001572 shows
this interface. The herding boom is permitted to pivot around
Item 6. Item 4 is a Herculite fabric skirt which provides a
flexible seal between the herding boom and the conveyor wall.
Herculite fabric was chosen for the seal because of its high
resistance to oils. The conveyor is fixed rigidly to the ship
128
-------
+50
w
^
o
I
w
CO
-100°
.5
1.5
rt
O
B
&
w
di
o
w
CO
2
O
OH
CO
§
5 .
1.0
1.5
2.5
3.0
. 1.0
1.5
2.0
2.5
3.0
OJ (rad/sec)
O 1/4 Scale Model Test
A 1/2 Scale Model Test
HEAVE RESPONSE OPERATOR & HEAVE PHASE ANGLE
vs.
WAVE FREQUENCY FOR PROTOTYPE
FIGURE 2
129
-------
with the waterline lying in the middle of the conveyor wall.
The conveyor wall is sized to permit a 4 foot high wave to
pass by without sorbent loss. The interface allows free
movement of the herding boom, yet contains the sorbent in
the channel.
The shape of the herding boom is controlled by the
boom support arms. The forward boom support arm supports
the drag load on the herding boom and controls the mouth width.
The aft boom support arm contols the width of the channel.
The forces on the forward boom arm may be calculated
using a drag coefficient of 1.5 for the herding boom.
Drag = 1/2 p C-. AV
3
Where p = Density of water (1.99 slugs/ft )
CD = Boom's drag coefficient (1.5)
A = Area normal to the current velocity
(24 feet2)
V . = Current speed (2 knots)
Drag = 1/2(1.99)(1.5)(12 x 2)(2 x 1.689)^
= 410 Ibs.
The forces at the end of the forward support arm
are as follows:
MM
I
130
-------
Balancing forces, force F and P are found,
F = 810 Ibs.
P = 700 Ibs.
Treating the boom as a pinned cantilever beam, the
critical buckling load can be found using Euler ' s formula:
P = v EI
cr " ~~
Where E = Modulus of elasticity (30 x 10 psi)
A
I = Moment of inertia (in )
£ = Length of column (21 feet)
For a 4" steel pipe, 1/4" wall I = 6.47 in
p _ TT (30 x 10 ) (6.47)
cr (21 x 12) 2
P = 30,166 Ibs.
C 3T
Safety factor = 3°QQ66 = 43
For a current speed of 4 knots the current drag on
the boom is 4 times greater or 1640 Ibs. This reduces the
safety factor to approximately 10. This gives the capability
of running the system at double the predicted maximum operating
speed with a sufficient margin of safety.
The interface between the herding boom and the
forward support arm is detailed in Drawing # 001570. The
float forward of the herding boom provides buoyancy for the
added weight of the support arm. The end of the forward support
arm provides the means for joining the herding boom, boom
support arm, boom support line, and the flotsam fence.
131
-------
The aft support arm aids in controlling the shape
of the mouth and channel of the herding boom. It is similar
in design to the forward support arm but lighter. The loads
the aft support arm sees will be minimal.
In determining the cross-sectional dimensions of the
"herding boom", consideration was given to sorbent loss, oil
loss, and cost.
When "herding" the oil with the boom, the oil wants
to run under the barrier. This run under or drainage is a
gravity phenomenon and the point where drainage will occur
can be approximated by a dimensionless number. The dimensionless
number that is used in connection with the drainage phenomenon
is called the densiometric Froude number given below.
NFRD
w
pw
Where V = Free stream velocity
L = Depth of barrier below the slick
g = Gravity constant
p = Density of water
W
PQ = Density of oil
Previous work in this area and model testing done
at Ocean Systems, Inc. established this critical Froude number
to be 1.1. At values less than this number oil is contained
and at values greater than this number oil is lost due to
drainage under the barrier.
Given the critical Froude number is 1.1 and a current
velocity of 2 knots, the value L may be computed.
132
-------
L =
Where V =2 knots (3.4 ft/sec)
NFRD = 1.1
g = 32.2
pw" po = .135 (#4 Fuel Oil)
pw
-135
L = 2.2 feet
The free stream velocity (V) is considered normal
to the barrier. Considering the shape of the mouth of the
system/ (see Figure 3) , the boom angle (8) to the current
(mouth angle) can range from 0° to 90°. During the hydro-
dynamic testing of the "herding" boom the optimum boom angle
appeared to be 30°. This angle reduces the normal velocity
to the "herding" boom to 1/2 the current velocity resulting
in L equaling .55 feet. However, to allow for herding
heavier oils whose density can reach as high as .95 (requires
L to be 3 times larger than the example) a barrier depth of
2 feet was chosen.
1
At moderate current speeds oil is also lost under
the barrier by the formation of droplets. This phenomenon
is known as entrainment. Entrainment begins"'with a 2 foot
•-JR'-
deep barrier at velocities roughly equivalent to 2 knots.
Further increases in depth produce little benefit.
Consideration was given to providing screens both
top and bottom to further protect the system from sorbent
loss. This idea was rejected for two reasons.
133
-------
MOUTH
x ^-SAGGING
Q.
Ul
FIGURE 3
134
-------
1. The screens would have a detrimental effect on
the sea conformance characteristics of the "dynamic keel"
boom.
2. The sorbent "herding" portion of the boom occurs
near the channel with the majority of the sorbents never seeing
the portion of the boom which is angled across the current.
The major function of the "herding" portion of the boom is to
"herd" the oil into the channel/ not the sorbent. The channel
portion of the boom "herds" both the sorbent and the oil with
the normal current velocity equalling approximately 0. Further
protection against sorbent loss using screens would be ineffective,
A summary of the characteristics of the herding boom
are as follows:
Mouth width
Mouth length
Mouth angle
Channel width
Channel length
Herding boom length
Herding boom weight
Forward support arm
weight
Aft support arm
weight
= 16 feet (does not include
channeling effect
of the ship)
= 20 feet
= 30° nominal
= 4 feet
= 20 feet
= 20 feet
= 280 Ibs (dry)
= 1500 Ibs (bottom section fully
saturated)
= 295 Ibs
= 110 Ibs
135
-------
SECTION 4.2
SORBENT RECOVERY SYSTEM
In the design of the sorbent recovery system, the
objective was the transfer of sorbents from the channel
to the oil extraction unit located on the deck of the ship.
In the conceptual stage of the design a conveyor system was
chosen which would combine both the trommel concept and the
transfer of sorbents to the oil extraction unit.
Drawings # 001569 and 001573 detail the sorbent recovery
system. It consists of two (2) conveyors; a sorbent recovery
conveyor and a delivery conveyor. The sorbent recovery
conveyor's function is to transfer the sorbent from the water
surface to the delivery conveyor. The delivery conveyor's
function is to transfer the sorbent to the oil extraction unit.
A two conveyor system was chosen for its flexibility in positioning
on board the recovery vessel.
The width of the sorbent recovery conveyor is deter-
mined by the channel width. A conveyor width smaller than
the channel would tend to clog/ or impede the flow of sorbents
from the channel to the conveyor. The overall length of the
conveyor is governed by the recovery ship's freeboard, the
desired inclination of the conveyor, and the interface require-
ment with the oil extraction unit. The typical or average
freeboard of the recovery vessel is 3 feet. The maximum
inclination angle of the conveyor which will permit a smooth
transfer of sorbents onto the ship is 40°. The interface
requirement with the oil extraction'unit required the delivery
conveyor to be four feet off the deck of the recovery ship.
Drawing # 001569 shows the sorbent recovery system
(which meets the above requirements) mounted on the recovery
ship. The conveyors may be positioned in a variety of ways
-------
to accommodate varying ships of opportunity. This versatility
is accomplished by the pivot assembly which secures the con-
veyor to the ship. The pivot assembly (see Drawing # 001573
allows the conveyor to pivot up and down, side to side, arid
may be moved inboard or outboard on the structural frame
connecting the pivot assembly to the ship. This versatility
allows use of the system on ships with freeboards up to six
feet -and with varying shapes. It also allows the conveyor to
be stored inboard while enroute to the spill area.
Drawing # 001569 shows a side view of the sorbent
recovery conveyor mounted on the recovery ship. The view
shows the conveyor extending two (2) feet below the water
surface. This is required to allow the system to operate in
2 foot seas with no sorbent loss under the conveyor. At one
point during the design a conveyor which floated on the water
surface was considered. This proved unsatisfactory due to its
large natural period. The floating end of the conveyor could
not respond quickly enough to conform to the sea surface.
Consequently, the conveyor was rigidly fastened to the recovery
ship with allowance made for a maximum of 2 foot seas. For
added flexibility, the ability to further extend the conveyor
under the water surface was designed into the conveyor, By
bolting Item 1 (see Drawing #001572, point A) to the conveyor
at point B, the end of the conveyor will lie three feet below
the water surface.
The sorbent recovery conveyor is sealed to the side
of the ship by a flexible skirt (Item 11, Drawing # 001573).
The conveyor is drawn tight against the side of the recovery
ship by a hogging line attached to padeyes (Item 31) located
on the end of the conveyor. Once the conveyor is positioned
on the ship, the pivot assembly is secured in place such that
137
-------
the recovery conveyor is only free to move up and down.
The delivery conveyor is positioned below the
sorbent recovery conveyor and transfers the recovered sorbents
to the oil extraction unit. The only requirements for the
delivery conveyor are the ability to receive and deliver the
sorbents to a bin whose top is located four feet o.ff the deck
without loss of sorbents.
The power requirements for the conveyor are small.
The system is limited to a process rate of 20 bulk ft /min
of sorbent. If the sorbents are totally saturated this will
be a maximum of 10 ft /min of oil recovered or 74.8 gal/min.
The maximum expected height required to raise the oil by the
sorbent recovery conveyor is 15 feet. The theoretical power
required may be calculated as follows assuming 100% efficiency.
Power = Recovery Rate x Height of Oil Raised
Power = 10 ft3/min x 60 lbs/ft3 x 15 ft
= 9000 ft Ibs/min
= .27 HP
Assuming at worst an efficiency of 25% the power
required equals 1 HP. However, to allow for an increase in
the process rate and increased loads due to the system fouling,
a 2 HP unit was chosen to drive the sorbent recovery conveyor.
This unit is a three-phase variable speed motor which will
allow the operator to control the speed of the conveyor.
In addition a remote electrical control is provided for
remote control of the conveyor at a centrally located control
station. The conveyor speeds can vary from 20 ft/min to
200 ft/min. An increase or decrease in these limits, if
desired, may be accomplished using gears. The maximum required
138
-------
speed for the sorbent recovery conveyor is estimated to be
100 ft/rain with a current velocity of 2 knots.
The variable speed drive for the sorbent recovery
conveyor is an important feature of the system. The overall
flow rate of the system is controlled by the belt speed of
the recovery conveyor. _.The_jvelocity of the belt controls the
sorbent flow which in turn controls the contact time which
governs the percentage of oil absorbed. Consequently/ the
efficiency of the total system is directly dependent upon
the speed of the recovery conveyor.
The delivery conveyor is driven by a constant speed,
three-phase, 1 HP motor. The speed of the belt is fixed at
93 ft/min. The speed of the delivery belt need not be variable.
The delivery conveyor's only function is to transfer the
sorbent from the recovery conveyor to the oil extraction unit,
the flow rate being fixed by the recovery conveyor. At the
maximum sorbent flow rate of 20 bulk ft ,
be loaded with 1 layer of sorbent cubes.
maximum sorbent flow rate of 20 bulk ft /min, the belt will
A summary of the characteristics of the sorbent
recovery system are as follows:
Sorbent Recovery Conveyor
/
Length = 23 feet
Width = 4 feet
Weight = 950 Ibs (1280 Ibs including support
frame)
Power = 2 HP
Belt Speed = 20 ft/min to 200 ft/min
Lift Height = 15 ft maximum
139
-------
Delivery Conveyor
Length = 13 feet
Width = 3 feet
Weight = 580 Ibs (920 Ibs including support
frame)
Power = 1 HP
Belt Speed = 93 ft/min
140
-------
SECTION 4.3
VESSEL PRELIMINARY DESIGN DESCRIPTION AND RATIONALE
Ship Characteristics
The ship characteristics developed were based on
the following oil recovery system technical and operational
considerations:
System Considerations
The physical size of the system.
The weight of the system.
The volume of sorbent that is required.
The deck space required for the system.
The principal components of the oil recovery
system are the herding boom, the conveyor, the oil extraction
device and the sorbent transfer and broadcasting equipment.
Other additional components are the sorbent containers, the
recovered oil reservoir, the pump and the portable power unit.
It is assumed that a portable electric power supply will be
required.
Vessel Operational Requirements
and Considerations
The vessel operational requirements (for
protected waters) of the RFP are as follows:
Vessel Speed 12 knots
Maximum Draft 3 feet
Waves 2 feet
Current 6 knots
Wind 20 mph
Temperature 20°C
Recovery (minimum/desirable) 1500/3000 gal/hr
141
-------
Other requirements are:
It must have an 8 hour supply of fuel.
It must have adequate deck space and stability,
It must be able to tow (alongside) the oil
storage craft during operations.
It must have sufficient maneuverability to
operate in restricted areas.
It must be capable of operating the system
, without assistance from other craft.
It should have relatively low freeboard,
i.e., not over four feet.
General Deployment Vessel Principal Characteristics
The following deployment vessel characteristics
are given below based on the technical and operational
requirements of the oil recovery system using sorbents.
These characteristics are not based with a specific vessel
in mind, but rather on general requirements. Later, certain
vessels are identified which are compatible with these require-
ments. Some of the characteristics are given in minimum and
maximum terms.
Length
Width
Draft
Displacement
Type of Propulsion
Screws
80-200 feet
20-40 feet
2-6 feet
50-400 tons
Preferably diesel with
low speed capability
Preferably twin screws
with low speed capability
such as afforded by con-
trollable pitch screws.
Single screw ships with a
large rudder are satis-
factory.
142
-------
Horsepower
Speed
Endurance
Maneuverabili ty
Deck Spatce
Stability
Personnel Accommodations
Freeboard
200-2000
12 knots
8 hours
Must be good in restricted
waters.
2
1000 feet minimum
Must be capable of taking
the weight of system
excluding oil.
Should have capability to
accommodate crew and
working party for a total
of 8 men.
2-4 feet
NOTE; It should be noted that a range of 2 to 6 feet
is given for the draft. The 3 foot maximum draft in
the RFP is very difficult to attain. Some vessels
can come close to this draft with a reduced fuel load.
Recommended Deployment Vessels and Craft
The vessels and craft listed can meet most of the
vessel requirements except for the draft.
USCG 100' or larger buoy boats
USCG 100' or larger inland buoy tenders
USCG 115' river buoy tenders
USCG 75' or larger inland construction tenders
Commercial construction tenders 80' or more
Fishing boats 80' or more
Small commercial tugs 100' or more
143
-------
Powered oil barges 80' or more
Army or Seabee construction or special purpose
craft 80' or more
Offshore oil supply boats, 125' to 200'
It should be recognized that a large uninterrupted
deck area is ideal for the oil recovery system. The ship
best meeting this requirement is the offshore oil vessel.
These are plentiful on the Gulf coast and some around Santa
Barbara, California.
Normally most ships have a passage on each side
of the deck house that is 2 to 3 feet wide. This is necessary
so that the screw conveyor which also acts as the broadcaster
can have a clear run. Where this is not possible, deck houses
can be cut to accommodate the system. If the ship has solid
bulwarks, these can also be cut out. After the operation,
the areas cut out can be welded up.
Operational Scenario
The normal operational mode for operations would be
with the system deployed to the starboard side and the oil
barge tied to the vessel on the port side. It is not feasible
for the oil to be stored aboard the vessel except when the
vessel is an oil barge. The gain in topside weight which
affects stability precludes this. In addition, the unloading
of the oil would interfere with oil recovery operations. An
alongside barge has the advantages of providing steering
symmetry and it can be replaced by another barge when it
is full without disrupting operations. Normally the oil
recovery system should be installed and removed with the
vessel anchored so that both sides are accessible.
The vessel will normally travel at 2 knots during
recovery operations. It is recognized that accurate steering
is difficult at such low speeds. This can be improved by the
144
-------
use of engines if the vessel is twin screwed or by using a
drogue if it is single screwed.
During normal oil recovery operations, the vessel
will operate with or against the current because oil slicks
tend to form their long axis parallel to the current. It is
not always possible to operate in this manner because of lack
of sea space. In addition, currents in excess of 2 knots will
cause the ship to move astern relative to land. In such a
situation it will often be necessary to operate at an angle
to the prevailing current. This in general should not present
problems although currents other than those parallel to the
longitudinal axis of the vessel may change the configuration
of the boom and cause eddies.
The effect of wind on the direction of the vessel
is significant at low speeds. This can be compensated for
by the rudder and/or engines to keep the vessel parallel to
the current where feasible. This is more difficult when
heading into the wind and particularly with a high bow.
It can be expected that wind and current will not be in the
same direction in inland waters.
Normal maneuvers and turns should be done gradually
to maintain steady state oil recovery operations. However,
sharp turns or backing down should not affect or damage the
oil recovery system providing the speed does not exceed 4 knots,
Each vessel will have a different effect on the
oil recovery system. A vessel with fine lines will steer
better but will probably roll more. Obviously a larger vessel
is more desirable from an operational viewpoint, but will have
an undesirable larger draft. The single characteristic that
has the greatest effect on the system will be the number of
screws. A single screw ship will be more difficult to control
under adverse conditions than a twin screw type. It may be
145
-------
necessary to use a small boat on the bow ofL. the simgle screw
ship to enable it to steer effectively.
An important characteristic that affects the oil
recovery system is slow speed control. Controllable pitch,
screws, or a diesel-electric system are best. If speed is
excessive, drogues or holdback craft may be required.
The effect of 2 foot seas on the vessel will be
minimal. The roll angle should never exceed 5 degrees. An
80 foot vessel is not greatly affected by two foot seas. In
addition, the damping action of the barge alongside as well as
the conveyor and herding boom will reduce the roll. The
pitching angle will be insignificant.
The weight of the oil recovery system will be
approximately 10,000 Ibs. Such a deck weight should not
degrade the stability of the vessel to any significant degree.
However, it should be recognized that such weights will require
weight handling as well as industrial facilities for the
installation of the system on ships of opportunity.
146
-------
SECTION 4.4
OIL STORAGE AND DISPOSAL SYSTEM PRELIMINARY
DESIGN DESCRIPTION AND RATIONALE
Introduction
To effect support of the oil recovery system, it is
necessary to move and store 200,000 gallons of recovered oil.
In nearly all cases, the storage capacity of the vessel that
is deploying the oil recovery system is limited. The only
exception would be if a small powered tanker barge were used.
In all other cases the periodic removal of recovered oil is
mandatory. It is further assumed that in nearly all cases
the use of the vessel fuel tanks are not available for storage
of recovered oil. The limited fuel tank capacity as well
as the contamination of fuel tanks would preclude such a
course of action. In addition, the increase in draft of the
vessel and the necessity for unloading would cause operational
interruption.
The providing and disposal of the agglomerates used
in oil recovery has also been considered.
Oil Storage During Recovery Operations
Topside oil storage on the vessel is not feasible
for the following reasons:
As oil is recovered the vessel will increase in
draft and the center of gravity of the ship will rise. The
latter effect will in general govern the oil storage capacity.
The weight of the oil will be added at a point significantly
above the center of gravity of the ship. This will raise
the center of gravity of the deploying vessel and reduce its
stability. The limiting capacity is different for each ship
and must be evaluated accordingly. A noticeable reduction
in the "stiffness" of the vessel as tested by a person walking
from one side to the other is a rough measure of reduced
stability. The vessel must then be unloaded. Such a procedure
147
-------
is not efficient or feasible. Therefore/ oil storage during
operations is best achieved by the use of an oil barge moored
alongside the deployment vessel. The recovered oil will be
pumped from the temporary storage tank aboard the deploying
vessel to the oil storage barge. When the oil storage barge
is full or when the draft becomes excessive, the barge will
be replaced by another one or it will be unloaded to ashore
or afloat oil storage facilities.
Shore and Afloat Storage for 200/000 Gallons of Recovered Oil
Afloat storage is considered to be the most
important in oil recovery operations. The reason for this is
that areas subject to spills are normally within range for
afloat storage facilities. Fuel oil facilities normally
consist of fuel oil barges and small powered tankers as part
of the fuel oil distribution system. Such craft transport
the oil from the large tanker terminals to the points of use.
This method is used where possible since it is the most
economical mode of transportation. The oil is then transferred
to fuel oil tank storage facilities on shore for further dis-
tribution, usually by tank truck.
It is important to recognize that such facilities do not
have significant shore storage facilities for polluted oil.
In general such facilities are available only at refineries
and related oil processing plants. Therefore, polluted oil
could not be offloaded at fuel company facilities. Thus the
most feasible method of storing polluted oil would be on oil
barges or small powered oil barges such as are normally found
in most cities that have water transportation available. The
recovered oil could then be offloaded at the large tanker
terminals for transport to oil processing facilities.
It would take a barge of approximately 665 ton
capacity to accept 200,000 gallons of oil. This is feasible
since most local oil barges or small powered tankers have a
148
-------
capacity of 500 to 2000 tons. The great advantage of oil
barges or tankers is that they can be brought to the scene
of operations. In many cases a single barge moored alongside
the oil recovery vessel would provide the requisite oil storage
capacity. However, in shallow waters, the draft limitation
would require an additional barge before the capacity of the
first barge is reached.
Small powered tankers are normally available in
coastal areas from companies that are in the oil transport
business.
Where barges or water transport are not available,
tanker truck transportation to a suitable offloading facility
would be required. This would be a slow and disruptive step
in oil recovery operations since this would mean that oil
storage during recovery could only be provided aboard in
small portable tanks or with craft of opportunity.
In view of the above, the use of afloat oil storage
facilities is recommended where possible. Such facilities
usually have adequate capacity and can be provided at the
site of operations.
149
-------
Sorbent Storage, Transporation and Disposition
The storage, transportation and disposition of
the sorbent cubes will be based on the following operational
requirements:
Total Volume of Cubes 200 feet
Working Volume of Cubes 120 feet
Volume of Feed Hopper 80 feet
Makeup Volume 80 feet
Container Selection
Due to the relatively large volume of
sorbent, the containers must be light, be capable of being
stacked, occupy a minimum of space when empty, be oil
resistant and rugged. In addition, the cost must not be
excessive.
A number of containers were considered.
The most promising two appeared to be the 20 gallon poly-
ethylene garbage can and heavy polyethylene bags.
A comparison of the two containers indicated
that the polyethylene bag was superior for the following
reasons:
It is less expensive
It can be transported easily
It can be stored very efficiently even
in cramped and irregular areas.
When empty they occupy very little space.
After an operation, the sorbent cubes can be
bagged directly from the feed hopper.
-------
The sorbent cubes can be stored in the bags
or the bags can be placed in a confined
container for permanent storage.
Bag Specifications
The bag specifications are as follows:
Material - Polyethylene, 6 mils thick
Size - 20 to 30 gallons
Number Required - 100
Including Spares
Volume Requirements
The approximate space requirements for the 200 feet
of all the sorbent is a space 41 x 5' x 10'. For the makeup
volume of 80 feet , a space of 4' x 5' x 4' will be required.
These are relatively small volumes and are particularly
advantageous when the limited space aboard the vessel is
considered.
Sorbent Disposal After Use
There are two courses of action open concerning the
disposition of the sorbents after use. They can be disposed
of by burying or burning them, or saving them for the next
operation. It is believed that the latter course of action
isl feasible and especially if they can be used several times
a year. It is recommended that the sorbents be stowed after
use either in the polyethylene bags or placing the polyethylene
bags in a confined container. These can be stored in the
weather.
151
-------
SECTION 4.5
OIL TRANSFER SYSTEM PRELIMINARY DESIGN AND RATIONALE
Previously it has been established that the most
feasible method of the storage of oil that is being recovered
is by means of a barge that is being towed alongside the
recovery vessel. This will require an oil transfer system.
This system will take suction from the oil recovery sumpf
through an oil pump and discharge it into the oil barge.
A number of oil transfer systems were examined.
The system best suited is as follows:
Pump - Blackmer Multi-Vane GXS2-1/2 Pump
driven by a 5 HP electric motor
This pump is capable of transferring 3400 gallons
per hour. It has advantages of high efficiencies at various
viscosities of oil, it is designed to handle petroleum
products, can handle oils with a wide range of viscosities,
can pass de,bris without damage and is self adjusting.
Hose - Heliflex Blue 2-1/2" Hose for suction •
(25') and discharge (2001).
It is lightweight, flexible and is designed to
carry oils, chemicals and acids. Fittings compatible with
this hose are available.
The pump selected is a positive displacement vane
£•
pump. This type can"pump approximately 3400 gallons per hour
with an oil viscosity range of 100 to 25,000 SSU. In our range
of differential pressure, there is little effect of viscosity on
pumping volume. For comparison at 20°C, 100 SSU is the mid-range
of No 4 fuel oil and 20,000 SSU is the mid-range of No. 6 fuel oil.
152
-------
SECTION 4.6
FLOTSAM FENCE PRELIMINARY DESIGN DESCRIPTION
In the course of oil recovery operations, it can be
expected to encounter flotsam and debris. This would be
inimical to the operation of the oil recovery system. The
amount and type of flotsam could vary widely. Large floating
debris such as logs and trees can usually be seen and avoided.
The danger from these is minimal since they cannot be picked
up by the oil recovery system. The greatest danger is from
smaller objects such as bottles, cans and small pieces of
wood. These can be picked up by the conveyor and introduced
into the system.
When operating in areas where flotsam is encountered,
a flotsam fence should be used. The fence will be deployed
as shown in Drawing # 001569. One end will be attached to the
float and the other to the ship. When flotsam is captured,
it must be removed. This can be accomplished either by
removing the objects physically from a small boat or detaching
the fence from the vessel and dumping it by swinging it outboard.
This latter operation would also be done with a small boat.
The flotsam fence would normally assume a parabolic
configuration and the flotsam would tend to be collected in the
apex. For this reason, the flotsam fence normally would not
shed the flotsam outboard of the boom.
The flotsam fence is made of 3/4" Flat Flex wire belt
.: an7-'"
mesh which is made of .092" stainless wire. The total width is
24 inches of which 8 inches is above the water and 16 inches
below. Floats are attached at the surface and 1 Ib. ballast
weights are at the bottom. A sketch of the fence is shown on
the following page.
155
-------
It should be recognized that the flotsam fence
when filled with debris will act as an oil barrier and will
significantly affect oil sorbtion. This is particularly true
when the Froude number exceeds 1.1. At this point all the
collected oil will be dumped under the fence at once.
154
-------
SKETCH OF FLOTSAM FENCE
WT
_^
1 __^
1
. ., .... ^.
'
1
8" ,. |
A '
T 2
16" 1
\ \
3" x 3"
CLOSED CELL
FOAM FLOATS
BALLAST
1 LB/FT
TOWING BRIDLE
(2) 1/4" x 1" BACK TO BACK ANGLES
FOR TOWING ATTACHMENT
FLAT FLEX WIRE BELT
3/4" MESH .092 STAINLESS WIRE
WEIGHT 2 LBS PER FOOT
LENGTH-
DEPTH—
WEIGHT-
VOLUME-
PRINCIPAL CHARACTERISTICS
TOTAL WEIGHT-
DRAG AT 2 KNOTS-
-35 FEET
-*16" IN WATER 8" ABOVE WATER
-3 LBS PER FOOT INCLUDING BALLAST
-APPROX 5 FT INCLUDING FLOATS
-APPROX 125 LBS
-130 LBS
FIGURE 4
155
-------
APPENDIX
157
-------
.. >. r--rv ^/~j
sa Z ~"4 Ll "VJ Vi*M V c4ii) Li L^a Li \JX U
AFFILIATE OF UNION CARBIDE CORPORATION AND THE SINGER COMPANY
11440 ISAAC NCWTON INDUSTRIAL SQUARE NORTH, RESTON, VIRGINIA 22070 TELEPHONE (703) 471-1310
TaEX 899442
May 17, 1972
Meloy Laboratories, Inc.
6631 Iron Place
Springfield, Virginia 22150
Attention: Mr. R. Pasquale, Contracting Officer
Subject: Budgetary Estimate to Construct a Prototype Oil
Recovery System Using Sorbants
Reference: (1) OSI letter dated May 3,- 1971
(2) Telephone Conversation between R. Pasquale and
D. Bellovich on May 11, 1972
Gentlemen:
Ocean Systems, Inc., herein submits the following budgetary figures for the
major cost categories in support of our $60,000 budgetary estimate provided
in the reference (1) letter.
1. Direct Labor Costs including Overhead $42,500
2. Material 16,500
a. Flotsam Fence $ 700
b. Transfer Pump & Hoses 3,500
c. Herding Boom 7,200
d. Recovery & Delivery 5,100
Conveyor
3. Other Costs 1,000
Total Budgetary Estimate $60,000
This estimate includes costs for doing detailed design work necessary for pre-
paring working drawings required for fabrication, fabricating the flotsam fence,
forward hurding boom, and recovery and delivery conveyors, providing the
transfer pump and hoses, and supplying the boom support arm interface.
This does not include providing a power supply, oil extractor, or system
interface components.
158
-------
Mr. R. Pasquale
May 17, 1972
Page -2--
It is hoped that this additional breakdown is sufficient to assist you in
preparing your final report for EPA.
Very truly yours,
OCEAN SYSTEMS, INC.
Dale E. Bellovich
Contract Administrator
DEB:ds >'
159
-------
1
Accession Number
w
5
« Subject Field & Group
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization , .
Meloy Laboratories, Inc.
6715 Electronic Drive
Springfield, Virginia 22151
Title
Oil Recovery System Using Sorbent Material
•J^J Authors)
Garth D.
Thomas P
Gumtz
. Meloy
ix Project Designation
EPA, Contract NO.
68-01-0068
21 Note
22
Citation
Environmental Protection Agency report number,
EPA-670/2-73-068, September 1973.
23
Descriptors (Starred First)
*Oil slick Recovery, *S/orbent Oil Recovery, *Shipboard Oil Recovery;
sorbent broadcasting, sorbent herding, sorbent pickup, sorbent
recycle, vessel requirements.
25
Identifiers (Starred First)
*Oil Recovery by Sorbents
27 Abstract
1 The feasibility of recovering oil in slicks 1mm & thicker by the use of
recycled sorbents has been shown in Laboratory & wave tank tests. Sorbents
made of recticulated foam are broadcast on the sea, herded by a boom, picked
up by a porous belt and the oil squeezed out of the sorbents by a wringer.
The sorbents are then rebroadcast on the sea for further oil recovery,.
General equations were developed for basic sorption properties, sorbent
broadcasting, sorbent herding, sorbent pickup, recovery of oil from the
sorbent & for the total system. Based on the laboratory modeling & general
equations, the total system concept was developed. It was concluded that
one inch cube sorbent particles distributed in a shrouded rectilinear screw
fed system was optimal. A 4/1 compression ratio of the slick by a boom
hording the sorbent & oil to the channel would work under virtually any
wave condition.
Capital costs for this system are $216,500 & operating costs $59,500
per year. Economic studies made indicating that recovery costs is about
6$ per gallon, assuming a system is used 22 days per year & a 20 year
depreciation. For large spills the salvage cost of the oil could pay for the
operation. The system has a nominal capacity of 5,000 gallons per hour.
Abstractor
-
Institution
Meloy Laboratories, Inc.
WR:102 (REV. JULY 1969)
WBfil r
if U.S. GOVERNMENT PRINTING OFFICE: 1973— •
SEND. WITH COPY OF DOCUMENT. TO: WATER RESOURCES SCIENTIFIC IN FORM AT lOtTTTENT ER
U.S. DEPARTMENT OF THE INTERIOR
lii WASHINGTON. D. C. 20240 /•, /
------- |