WATER POLLUTION CONTROL RESEARCH SERIES • 15080 DJP 10 70
OIL/WATER
SEPARATION SYSTEM
WITH SEA SKIMMER
U.S. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Reports describe
the results and progress In the control and abatement
of pollution in our Nation's waters. They provide a
central source of Information on the research; develop-
ment, and demonstration activities In the Environmental
Protection Agency, through Inhouse research and grants
and contracts with Federal, State and local agencies,
research Institutions, and industrial organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications Branch,
Research Information Division, Research and Monitoring,
Environmental Protection Agency, Washington, D.C. E0460,
-------
OIL/WATER SEPARATION SYSTEM
WITH SEA SKIMMER
.The Garrett Corporation
AiResearch Manufacturing Division
Los Angeles, California 90009
for the
OFFICE OF RESEARCH AND MONITORING
ENVIRONMENTAL PROTECTION AGENCY
Project #.15080 DJP
Contract #14-12-524
October 1970
For rale by the Superintendent of Documents, U.S. Government Printing Office
Washington, D.C., 20402 - Price $1.60
-------
EPA Review 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
An oil separation and skimming barge capable of processing up to 500 gpm
was designed, constructed, and operated in open ocean from Sea State 0 through
3. Test data was collected for the centrifuge on land and for the total system
at Sea States 0 and I. The major components are: (I) entrance paddle wheel,
(2) louvered quiet pond, (3) self-adjusting skimming weirs, (4) storage/surge
tank, (5) centrifuge plus auxiliary equipment.
The primary target performance for the centrifuge was a water discharge
containing less than 100 mg/1 oil and secondly an oil discharge containing
less than 5 percent water starting with an influent oil emulsion of up to
40,000 mg/1. The centrifuge efficiency was 'dependent upon flow rate, emulsion
concentration, and the gravity (°API) of the oil. For oils between 21 and 31°
API, and centrifuge operating between 2750 and 3350 rpm, the primary perfor-
mance was met at 100 gpm flow rate and 30,000 mg/1 oil emulsion or at 500 gpm
flow rate and 1000 mg/1 oil emulsion. Starting with 23,000 to 48,000 mg/1
oil emulsions flowing at 235 gpm it is possible to reduce the oil concentra-
tion in the discharge water to an average value of 210 mg/1. Recycling this
discharge through the centrifuge reduced the oil concentration to less than
40 mg/1.
Skimmer efficiency was measured by spilling 35° API crude oil on the
ocean and comparing the amount of oil recovered to the amount dumped. At
1.2 knots and Sea State 0 the skimmer displayed an efficiency of 90 percent,
dropping to 75 percent at 2.3 knots and Sea State I.
This report was submitted in fulfillment of Project No. 15080 DJP,
Contract No. 14-12-524, under the sponsorship of the Water Quality Office,
Environmental Protection Agency.
-------
CONTENTS
Sect ion
I CONCLUSIONS '
II RECOMMENDATIONS 3
III INTRODUCTION 5
IV OBJECTIVE 9
V DESCRIPTION OF SYSTEM "
VI SKIMMER BARGE l5
VII OIL/WATER CENTRIFUGE 6l
VIII SYSTEM TESTING 95
IX DISCUSSION l23
X ACKNOWLEDGMENTS '55
XI SELECTED BIBLIOGRAPHY '57
XII PUBLICATIONS '59
161
XIII GLOSSARY AND ABBREVIATIONS
XIV APPENDIXES
163
-------
Barge Model as Modified for January 6 Tests
ILLUSTRATIONS
Figure
I Simplified Oil Skimmer Barge Piping Schematic '2
2 Oil /Water Separator Installation on Skimmer '3
3 Oil /Water Separator u
4 Areas Swept for Various Speeds and Sweep Widths '6
5 Paddle Drive Mechanism '8
6 Relationship of Model Speed to Full -Scale Speed 20
7 Flow of Water through the Catamaran if no 21
Bottom is Used
8 Flow of Water through the Catamaran if the 22
Louvered Bottom is Installed
9 Water Flow through the Catamaran when a Louvered 23
Bottom and a Forward Section is Used
10 Water Flow through the Catamaran when the 24
Forward Section Is Used without the Louvered
Bottom Section
26
12 Oil Skimmer Barge [/I2th-Scale Model with 27
Paddle Wheel
13 Oil Skimmer Barge l/!2th-Scale Model with 28
Paddle Wheel
14 Configuration of Model for Converging Passage 30
'Type Wave Attenuator (January 6 Tests)
15 Oil Skimmer Barge l/!2th-Scale Model with 31
Sloping Board Replacing Paddle Wheel
16 Floating Weir Model (I/I2th-Scale) 32
17 Oil Skimmer Barge I/12th Scale Model Paddle 33
Wheel Drive and Weir Pumping Systems
18 Model Configuration during the Tests of 34
January 12, 1970
VI
-------
ILLUSTRATIONS (Continued)
Figure Page
19 051 Skimmer Barge l/!2th-Scale Model 35
20 Oil Skimmer Barge !/!2th-Sca1e Model 36
21 Oil Skimmer Barge l/!2th-Scale Model 37
22 Skimmer Barge Basic Hull 39
23 Skimmer Displacement Curve 40
24 Skimmer Hull Compartments *'
25 Individual Compartment Displacements 42
26 Oil Skimmer Barge Piping Schematic 43
27 Fluid Flow Processing Diagram 45
28 Wave Gate *6
29 Wave Fences A7
30 Louvered Bottom *8
31 Self-Adjusting Floating Weir *9
32 Weir Pumping 5l
33 Weir Pumping Unit Performance 52
(J. W. Stang Model 3CRI8EL)
34 Surge Tank 53
35 Paddle Wheel Installation 55
36 Side View of Paddle Wheel 56
37 Weir Winch 59
38 Centrifuge Inlet Screen Installation in 60
Surge Tank
39 Cross Section of Oil/Water Centrifuge 62
40 Oil/Water Separator PN 585010-1-I 64
vii
-------
ILLUSTRATIONS (Continued)
Figure Page
41 Schematic of Centrifuge Test Setup 65
42 AiResearch Oil/Water Centrifuge 66
43 Centrifuge Speed and Torque Converter Slip 68
44 Pump Output Head and Separator Inlet 69
Pressure Requirements
45 Abnormal Centrifuge Performance 70
46 Crossplot of Original Performance Measurements 72
47 End Cap Seawater Corrosion (March 6, 1970) 74
48 End Cap Seawater Corrosion (March 27, 1970) 74
49 Corroded Discharge Vane, PN 585014 75
50 Corroded Discharge Housing, PN 585024 75
51 Cross Section of Original and Modified Designs 77
52 Inlet Configuration of Modified Centrifuge 78
53 Corrosion Protection for End Cap Surfaces 79
54 Outlet Flow Splitter and Oil Discharge Tube 80
Assembly
55 Outlet Flow Splitter and Drum Assembly 80
56 Oil Discharge Pressure Regulator 81
57 Performance Improvement 83
58 Oil/Water Separator Inlet Pressures 84
59 Preliminary Centrifuge Power Requirements 85
(Steady-State)
60 Variable Speed Performance 86
61 Detailed Performance Plot at Reduced 88
Centrifuge Speed
VI I
-------
ILLUSTRATIONS (Continued)
Figure
62 Separation as a Function of Oil Specific Gravity 89
63 Emulsion Breaker Evaluation, Tretolite JN9045 90
64 Influence on Separation by Increased Temperatures 90
65 Solids Distribution within Centrifugation System 91,
66 General Map Representing all of the Modified 92
Unit Performance Data
67 Sweeping Oil with Single Headrlck Boom 97
68 Head rick Rapidly Deplovable Boom 98
69 Headrick Boom Harness (Original Configuration) 99
70 Skimmer with Single 500-ft Section of Gates Boom 100
71 Gates Rubber Company Boom 101
72 Battelle-Northwest Water Spray Boom 102
73 Battel1e Spray Boom during Skimming 103
74 Test Areas 104
75 Barge Towing Forces 107
76 Dynamometer Setup for Measuring Towing Forces 108
77 Centrifuge Inlet and Discharge Samples 110
78 Skimmer Efficiency Test Setup 112
79 Setup for Determination of System Efficiency 113
80 Skimmer during Efficiency Test 114
81 Skimmer Efficiency Test 115
82 Turbulence at Forward End of Quiet Pond 125
83 Possible Solutions to Quiet Pond Turbulence 126
ix
-------
ILLUSTRATIONS (Continued)
Figure Pacie
84 Oil Drop Terminal Velocity as a Function of '28
Drop Diameter and Centrifugal Force
85 Time Required for an Oil Drop to Reach 131
99 Percent of Terminal Velocity
86 Travel Time as a Function of Drop Diameter '32
and Rotational Speed
87v Oil/Water Centrifuge Oil Drop Residence Time 133
as a Function of Flow
88 Photoelectric Particle Counter Schematic '34
89 Schematic 6f the Particle Sampling System '37
90 Data • Redact I on Prog ram '40
9( Sample Performance Test Data '42
92 Cost of Removing Oil from Ocean Surface 147
33 Wave Height Measuring Device '68
-------
TABLES
Table Page
I Program Summary 7
2 Sea State Table I0
3 Chronology of Tests l06
4 General Crude Oil Tests on Oil Samples Submitted ''7
by A1 Research Manufacturing Company
5 Modified Hempel Distillation of Oils Submitted 118
by AiResearch Manufacturing Company
6 Summary of Evaporation Loss Calculation 120
7 Summary of Results of Skimmer Efficiency 121
Test, July 27, 1970
8 Effect of Time on Particle Counts in Well Water, 136
May 14, 1970
9 Dilutant Particle Counts, May 25, 1970 138
10 Well Water Particle Counts, May 26, 1970 138
II Waste Water Droplet Counts, June 2, 1970 143
12 Effect of Flow Rate on Exit Oil Concentration, 144
June 2, 1970
13 Cost Summary - Method I 150
14 Cost Summary - Method II 151
15 Cost Summary - Method III 152
16 Sea Dragon Instrumentation 166
XI
-------
SECTION I
CONCLUSIONS
The following conclusions were reached as a result of the oil skimming
and separating system operation: *
I. The oil skimming barge, with its paddle wheel and floating weirs,
combines with the centrifuge to make a technically feasible method
of removing oil from the surface of the sea.
2. The total system (skimmer and centrifuge) can recover and process
up to 500 gpm (30,000 gallons/hour) of oil /water mixture.
(a) Inlet emulsions of up to 1000 mg/1 of 21° API or greater
crude oil can be separated by the centrifuge operating at
3350 rpm to produce an effluent containing 100 mg/1 or less
of oil. Degrees API are given by the following:
Degrees API = sp gr ' l31'5
(b) Inlet emulsions of more than 1000 mg/1 of oil will result In
the production of effluents containing more than 100 mg/1 of
oil. These effluents are returned to the quiet pond in the
skimmer barge for reprocessing. By recycling the effluents
back through the centrifuge, oil concentrations in the pro-
duced water can be reduced to less than 100 mg/1.
3. System overall effjcjency (Figure 79) for recovery of 35° API oil
is 90 percent when operated at low velocity (1.2 knots) in smooth
water (Sea State 0). The skimmer displayed an efficiency of 75 per-
cent at 2.3 knots and Sea State I. When the skimmer was operated
under Sea State 3 conditions, the efficiency visually appeared to
decrease although no quantitative data was taken.
4. The performance of the centrifuge when operating at constant speed
is dependent upon the flow rate, emulsion concentration, specific
gravity of the oil, and temperature.
(a) Lowering the inlet rate to 100 gpm allows processing an emulsion
with up to 33,000 mg/1 of 21° API oil so that less than 100 mg/1
remains in the discharge water.
(b) Processing emulsions containing 10,000 mg/1 of 21 to 25° API
oil at 200 gpm and 2750 rpm results in discharge water with
200 mg/1 of entrained oil. At the same operating conditions,
but using 28 to 31° API oil, the oil concentration In the
discharge water is 80 mg/1.
-------
(c) At the same operating conditions as above with 21 to 25° API
oil, raising the inlet emulsion temperature from 68 to !40°F
results in a reduction of oil concentration in the effluent
water from 200 mg/1 to 55 mg/1.
5. The skimmer barge can be used to recover weathered oil slicks. The
weathered oil is collected in the quiet pond from where it is
removed by hand and stored in drums- During one 4-hour period,
eighteen 55-gallon drums of tar were collected in this manner.
6. The centrifuge is capable of separating solid particles from the
inlet emulsions- Starting with 9-5 mg/1 of suspended oil-free
sol ids, over 80 percent were removed from the effluent when the
centrifuge was operated at 2750 rpm-
7. A standard commercially-available emulsion breaker increased the
separation ability by 50 percent and its use is an economically
feasible method of upgrading performance. As an example, it would
cost $0.005 (l/2-cent) per bbl of emulsion to use Tretolite
JN9045 ($2.5l/gal.) at the tested rate of fifty parts Tretolite
per million parts of emulsion.
8. The centrifuge does not attain the theoretical separation efficiency
expected from Stokes1 Law considerations- Oil droplets of ten
microns and larger should be removed from the effluent water.
However, analysis of the water samples shows that some of these
droplets are not being removed.
9. A multiple-staged centrifuge would have a greater overall efficiency
than the single-stage unit. When the effluent from the centrifuge
was recycled, the oil concent rat ion was further reduced.
10. This recovery/separation system can handle crude oils with gravities
ranging from 21 to 35° API. The thicker materials, such as tar,
will separate in the quiet pond or in the surge tank but will not
progress as far as the centrifuge.
-------
SECTION II
RECOMMENDATIONS
The following recommendations are made as a result of the tests performed.
OIL SKIMMER BARGE
A seagoing oil skimmer barge worth approximately $100,000 is now avail-
able for oil spill cleanup or, if properly maintained, for additional develop-
ment of components. Since oil spills can contain a wide range of properties,
and can occur under a variety of conditions, it is recommended that' this piece
of equipment be maintained and that the modifications described below be incor-
porated and evaluated to obtain even better overall performance.
I. It has been found that weathered oil slicks coagulate into large
masses of tar as they are drawn into the skimmer. '.These cannot be
handled by pumping systems. A chain-belt type of conveyor could
lift these masses directly from the quiet pond to a storage.bin on
deck. This device could also be effective in recover ing oi J-sb'aked
straw or other adsorbents.
2. The three small weirs should be replaced with a single large weir for
handling heavier oils. A larger pump, such as a Wejlpoint 6-in. by
4-in. centrifugal pump with a vacuum unit, could be used.
3. As an alternate method to replace the weir pumps, build and'test a.
vacuum tank system for lifting the oiI/water mixture from the quiet
pond. This would eliminate the weir suction pumps and the resultant
emulsification. The vacuum tank would be pumped by the centrifuge
inlet pump. Another alterative is to use a diaphragm pump in place
of the centrifugal pump to reduce emulsification in the pumping
process.
4. Modify the forward section of the quiet pond to reduce the present
turbulence and the minor loss of oil through the louvers at this
point. This may be done by modifying the aft end pf the spillway to
reduce vertical circulation, modifying the wave fen'cesj 'or closing
some of the forward louvers.
5- The present skimmer could be divided Into sect ions compatible with
truck, rail, or air transporatIon. This'would make th.is!;plrece 6f
equipment available to any area of the United .States within fine oi"
two days-
6. Tests reported in this study were for treatment above 68°F. More
information is needed on the efficiency at lower temperatures to
determine the system efficiency in colder climates.
-------
7. Future systems should be built with a permanent power supply in the
form of a diesel engine-driven hydraulic system located in the
engine compartment. All other equipment would then be driven by
hydraulic motors or actuators.
8. Conduct additional testing. A limited amount of testing was done
at Sea States 0 and I, and more quantitative data should be taken
at these states as well as under more severe conditions-
CENTRIFUGE
I. The present centrifuge could be modified to incorporate a multiple-
stage design. This would allow the unit to improve the separation
of higher Inlet emulsion concentrations while at the same time re-
ducing the concentration of oil in the discharge water.
2. Incorporate the centrifuge feed pump Into the design of the Inlet
section of the centrifuge. The centrifuge discharge section would
have to be designed to be compatible with the inlet section charac-
teristics. This would eliminate one of the undesirable emulsifying
components.
-------
SECTION III
INTRODUCTION
Spills of crude oil, refined products, vegetable and fish oils on marine
and inland waters occur from many sources and pose a constant threat to the
beneficial use of these waters and the adjacent shoreline. Sources of oil
spills include vessels, industrial establishments, pipelines, terminals, and
offshore drilling operations. Usually in the recovery of those spills, large
quantities of water are also picked up with the oil. In order to reduce the
amount of liquid to be transported from the spill site to the eventual disposal
area, it is desirable to remove as much of this excess water as possible at
the recovery site-
In the past, this separation step had not been done or it had been accomp-
lished by allowing the recovered product to settle in large tanks prior to
drawing off the excess water. Finely divided emulsions can require a prohibi-
tive settling time-
In the program described in this report, an investigation was made to
evaluate the feasibility of using a centrifuge to speed the removal of oil
from the recovered oil/water mixture- From previous testing on a prototype
5-gpm centrifuge, it was expected that oil/water mixtures containing one
percent oil (10,000 mg/1) could be separated so that less than 0.01 percent
(100 mg/1) oil remained in the water. The centrifuge is not designed to sepa-
rate mixtures in which the oil has weathered into tar. Based on these pre-
liminary tests, a program was initiated by the EPA/Office of Research and
Monitoring for AIResearch to design, build and test a centrifuge of 500 gpm
capacity to separate oil/water mixtures collected during recovery of oil
spilIs from the sea.
The program was later modified to include the design, construction, and
testing of an ocean-going oil skimmer capable of removing 500 gpm of oil/water
mixtures from the surface of the ocean. This equipment was used to test the
system under actual operating conditions and environment.
The separator is a horizontal centrifuge with axial inlet and discharge
ports that allows a throughput rate of 500 gpm with a power consumption of
60 horsepower. A thin annular channel at the periphery of the centrifuge
resulted in a small displacement distance for the oil droplets migrating
under the influence of the high centrifugal forces.
The oil skimmer was designed as an independent seaworthy craft capable
of supporting the required skimming equipment, the centrifuge, and the men
required to operate the system. The skimmer also contained an entrance paddle
wheel, self-adjusting weirs, storage/surge tanks, and a large quiet pond with
a louvered bottom that carried skimmed water and oil along within the hull of
the skimmer. This allowed preliminary settling) the resultant thickened
oil film was then skimmed by means of a weir system. Before being returned
to the ocean, the excess water skimmed with the oil was passed through the
centrifuge.
-------
A summary of the program is shown in Table I. Approximately 48 percent
of the effort was expended in the fabrication of the'"centrifuge and 22 percent
in the construction of the skimmer barge. As originally planned, testing was
limited to five days at sea. However., 20 additional at-sea days were accu-
mulated with the equipment under a concurrent contract with the American
Petroleum Institute.
-------
TABLE I
PROGRAM SUMMARY
TASK
TRADE-OFF STUDIES
CENTRIFUGE LAYOUT DRAWING
CENTRIFUGE DETAIL DRAWINGS
CENTRIFUGE FAB. AND ASSEMBLY
CENTRIFUGE TESTS
MODEL TESTS
SKIMMER DESIGN
SKIMMER FABRICATION
SYSTEM ASSEMBLY
SYSTEM TEST ( 5 DAYS OF
AT- SEA TESTS)
SYSTEM MADE AVAILABLE TO
THE AMERICAN PETROLEUM
INSTITUTE FOR ADDITIONAL
TESTING WITH SKIMMING BOOMS
AND OIL CONTAINER
FINAL REPORT
1969
J
-
-
F
-
M
—
A|M|J
•
-
F
-
-
j
•
—
A
•
—
S
1
—
0
.
N
'
D
1
1970
J
—
F
'
.
M
•
•
A
•
•
M
•
.
J
—
.-.
.•
J
—
•
-'
A
— "
'
S
-
'
i 1 1
0
-
•
N
-
•
D
-
•
TOTAL
PERCENT
OF
EFFORT
2
3
5
48
7
1
2
22
2
3
0
3
100
S-65E68
-------
SECTION IV
OBJECTIVE
The objective of the oil spill recovery program was to demonstrate
recovery and separation of an oil slick from the surface of the ocean under
conditions up to and including Sea State 3 (Table 2). The program consisted
of two tasks:
I. Design., build, and test a prototype model of an oil/water centri-
fuge with these objectives:
(a) Capable of separating a wide range of mixtures and emulsions
of crude oils and water.
(b) Capable of separating oil/water mixtures at a rate of up to
500 gpm-
(c) Capable of minimizing level of oil contamination in the effluent
water. The target performance was a minimum concentration of
100 mg/1 of oil in the water effluent and 5 percent water in
the oi1 phase.
(d) Capable of minimizing the water content in the recovered oil
to the extent that it does not compromise (c) above.
(e) Capable of being readily transported.
2. Design, build, and test a seaworthy skimmer to be used with the
centrifuge to recover oil slicks from harbors and open seas, the
skimmer to incorporate a paddle wheel, skimming well, self-adjusting
floating weirs for skimming the oil, and a surge tank to ensure
a submerged inlet to the centrifuge. The objectives for the
skimmer are:
(a) Capable of recovering thin films, light crude oil, refined
products, and/or heavy films of lower gravity and/or weathered
crudes and crude emulsions.
(b) Capable of functioning in a quiet harbor, around piers, docks,
vessels, with trash and debris present, and/or in open waters,
including the high seas, under Sea State conditions 0, I, 2,
and 3.
(c) Capable of sweeping up to 30 acres/hr using auxiliary side
attachments (booms).
-------
TABLE 2
SEA STATE TABLE*
Sea
State
0
1
2
3
4
5
6
7
8
9
Sea Description
Ripples
Small wavelets, still short
Large wavelets, crests
begin to break
Small waves, becoming
larger
Moderate waves, taking a
more pronounced long form
Large waves begin to form
Sea heaps up and whi te foam
from breaking waves begins
to f o rm
Moderately high waves of
greater length
High waves; sea begins to
roll; visibility affected
Very high waves with long
overhanging crests
Exceptionally high waves;
Visibility affected
Air filled with foam and
spray
Wind
Description
Light airs
Light breeze
Gentle breeze
Moderate breeze
Fresh breeze
Strong breeze
Moderate gale
Fresh gale
Strong gale
Whole gale
Storm
Hurricane
Wind
Velocity,
knot
2
5
8.5
10
12
13. S
14
16
18
19
20
22
24
24.5
26
28
30
30.5
32
34
36
37
38
40
42
44
46
48
50
51.5
52
54
56
59.5
> 64
Average
Wave Height
ft
0.05
0. IB
0.6
0.88
1.4
1.8
2.0
2.9
3.8
4.3
5.0
6.4
7.9
8.2
9.6
1 1
14
16
16
19
21
23
25
28
31
36
40
44
49
52
54
59
64
73
> 80
Significant
Wave Height,
ft
0.08
0.29
1.0
1.4
2.2
2.9
3.3
4.6
6.1
6.9
8.0
10
12
13
15
18
22
?•*
26
30
35
37
40
45
50
58
64
71
78
83
87
95
103
116
> 128
Approximate
Period Range,
sec
up to 1.2
sec
0.4 to 2.8
0.8 to 5.0
1.0 to 6.0
1.0 to 7.0
1.4 to 7.6
1.4 to 7.8
2.0 to 8.8
2.5 to 10.0
2.8 to 10.6
3.0 to 1 1. 1
3.4 to 12.2
3.7 to 13.5
3.8 to 13.6
4.0 to 14.5
4.5 to 15.5
4.7 to 16.7
4.8 to 17.0
5.0 to 17.5
5.5 to 18.5
5.8 to 19.7
6 to 20.5
6.2 to 20.8
6.5 to 21.7
7 to 23
7 to 24.2
7 to 25
7.5 to 26
7.5 to 27
8 to 28.2
8 to 28.5
8 to 29.5
8.5 to 31
10 to 32
10 to 35
S-60636
*From Handbook of Ocean and Underwater Engineering, John J. Meyers, ed., McGraw-Hill, New York, 1969.
-------
SECTION V
DESCRIPTION OF SYSTEM
The system shown schematically in Figure I and pictorially in Figure 2
consists of the following two major components:
I. Skimmer Barge--This vessel contains weirs, pumps, and the necessary
oil/water processing equipment to recover oil from the ocean sur-
face. Initial separation of the oil. and water is accomplished by
draining excess water out through the louvered bottom of the quiet
pond. A second separation step occurs as the weirs skim the oil-
rich surface from the quiet pond. A third separation is done in
the surge tank where the pumped oil floats to the surface. The
remaining water in the surge tank is passed through the centrifuge
to remove the smaller oil droplets still remaining in suspension.
The design and development of the barge is described in Section VI.
2. 500-gpm Centrifuge—The separation of the subsurface oil/water
emulsion from the recovered oil is accomplished by means of an oil/
water centrifuge. The design objective was to produce a centrifuge
capable of discharging oil with less than 5 percent water and water
with less than 100 mg/1 of oil. The centrifuge is shown in Figure 3
and the design and development is described in Section VII.
For the sea tests the centrifuge was mounted on the skimmer barge and
became an integral part of the overall recovery/separation system. Although
this equipment performs the critical function of recovery and separation,
certain additional equipment is required for a complete ocean oil spill re-
covery system, namely, containment booms and an oil storage tank. These were
supplied through a concurrent contract with the American Petroleum Institute.
I I
-------
SURGE TANK
FLOATING
WEIRS
WEIR PUMPS
FVJ
CENTRIFUGE
|
I
TO OIL
STORAGE
OIL/WATER MIXTURE
CLEAN WATER
OIL
WATER
OVERBOARD
OIL
RECEIVER
S-61060
Figure 1. Simplified Oil Skimmer Barge Piping Schemat
1C
-------
Figure 2. Oil/Water Separator Installation on Skimmer
13
-------
Figure 3. Oil/Water Separator
14
-------
SECTION VI
SKIMMER BARGE
INTRODUCTION
The major function of the skimmer barge was to suppress wave action and
obtain a quiet pond from which the surface oil may be skimmed. The overall
size was determined by considering the potential sea state and the surface
area to be swept. These in turn determined the size of the quiet pond, the
rate of processing of skimmed fluid, the barge speed, and the size of process-
ing equipment. The goal for the skimmer barge was the capability to sweep by
the booms 30 acres of ocean surface and to process 500 gal/min of surface fluid.
Barge speed versus sweep width is shown in Figure 4. The vessel was to be
capable of operating from Sea State 0 to Sea State 3 (see Table 2). From the
foregoing problem statement and consultations with shipbuilders, the optimum
size of the vessel was determined to be 45 ft long with a 26-ft beam.
In order to operate effectively the self-adjusting floating weirs used
to skim the oil had to be located in an environment of still water. If the
skimmer barge is moving through the water to increase its area of operation,
the water in the immediate vicinity of the weir has to be moving with the
barge to obtain a zero relative current. This was to be obtained by the in-
stallation of suitable baffles or other devices between the two hulls of the
barge. Since the cost of building, evaluating, and changing full-scale com-
ponents of the skimmer was prohibitive, a one-twelth-scale model was built
for evaluation purposes.
MODEL DESIGN, CONSTRUCTION, AND TEST RESULTS
A one-twelth-scale model of the oil skimmer barge was built and tested
with these initial objectives:
I. To determine whether a simple U-shaped hull with two side pontoons
and a closure across the stern would be sufficient to stop relative
water motion on top of the quiet pond.
2. To determine if the addition of a forward spillway, or weir, would
suffice to stop relative water motion on the top of the quiet pond.
3. To determine if a louvered bottom between the hulls would suffice
to stop relative water motion on top of the quiet pond.
4. To determine if a combination of all of these devices is necessary.
After these questions were resolved, the model was further developed to
evaluate the effect of the paddle wheel and other wave-damping devices. Even-
tually a self-adjusting floating weir and pumping equipment also were installed.
15
-------
CO
fe
a.
to
100 200 300 500
SWEEP WIDTH, FT
1000
2000
Figure 4. Areas Swept for Various Speeds and Sweep Widths
5000 10000
S-61885
-------
The model was constructed with removable components so that these com-
ponents could be tested in various configurations. Test data consisted
primarily of visual observations and an 8-mm color movie record of the tests.
These were taken at 64 frames/sec, which produced a full-scale effect when
viewed at 16 frames/sec. Details of important phenomena were also examined
frame by frame on a viewer. Small wooden floats were attached to the bridge
structure across the quiet pond by means of strings. These, along with powder
floating on the water surface, assisted in determining the direction and
approximate velocity of the water currents.
First Model Version
The model was built of l/4-in. plywood to the scale of one inch to a
foot. (All dimensions in this section are model dimensions. For full-scale
dimensions change inches to feet.) The model was divided into several water-
tight compartments so that various conditions of ballast could be simulated.
As originally constructed, the model consisted of the engine compartment
and the twin hulls. A simulated 6-in.-wide paddle drive bridge was installed
immediately behind the 8-in.-diameter paddle wheel. The paddle drive mechanism
shown in Figure 5 was mounted upon this bridge. Also constructed were the
louvered bottom and the forward bottom sections. These parts were removable
to permit evaluation of their effect. An 8-bladed, 8-in.-diameter paddle
wheel was used.
These tests were conducted on November 24, 1969 by manually towing the
mode] alongside a long dock at Alamitos Bay. The primary purpose of this test
series was to determine the necessity of the louvered bottom and forward spill-
way. The model was towed at various depths and speeds to simulate a wide set
of conditions.
The fluid forces that are relevant in ship model testing are the inertial
and gravitational forces. The relationship between these must be identical for
both the full scale and model unit:
Fx = Ma = PL3 ^j- = pV2L2
FG = Mg = pL3g
ft), • fej, • (4 • (4
where F, = inertial force, Ib
2
M = Mass of object, Ib-sec /ft
2
a = Acceleration, ft/sec
17
-------
-FIXED LINE UNWINDS FROM DRUM
CAUSING DRIVE TO TURN
V-BELT DRIVE
LINE TO
FIXED POINT
CD
TOWLINE
S-60919
Figure 5. Paddle Drive Mechani
sm
-------
2
p = Fluid density, Ib-sec /ft
L = Length, ft
V = Velocity, ft/sec
Fr = Gravity force
2
g = gravitational acceleration, 32.2 ft/sec
Subscri pt
P = Full-scale prototype
M = Model
therefore
V,,
Using this relationship, the equivalent speed for a one-twelth-scale model is
shown in Figure 6.
Without a bottom in the barge, the water through which the model was being
towed did not deflect until just in front of the pond aft bulkhead, as shown
in Figure 7. This was an unsatisfactory condition because the floating weirs
were being pushed through the water at essentially boat speed. This upset the
operation of the weir.
Adding the louvered bottom tended to slow the water at the forward end of
the quiet pond, (Figure 8); however, there was still considerable velocity at
the we i r locat ion.
Addition of the spillway reduced the amount of water that entered the
quiet pond; the flow pattern is shown in Figure 9. Floats, tied to the paddle
drive bridge by strings, floated idly in the region of the weirs, showing ideal
current conditions in the quiet pond.
When the louvered bottom was removed but the spillway retained, a reverse-
flow eddy was created, as shown in Figure 10, which carried the indicating
floats forward over the spillway. This condition would cause the surface oil
to move away from the weir inlet.
Evaluation of the paddle wheel was unsuccessful because the drive mechanism
was too elastic to drive the paddle at a constant speed. Also, there were no
waves at this test site to test the ability of the paddle in attenuating in-
coming waves.
19
-------
o
LU
CO
FROUDE NUMBER RELATIONSHIP
y ~ , ~ i ~ 3.46
m
K
j l.69Kp=3.46
K = 2.05 V
p m
WHERE V = VELOCITY, FPS AND
L = LENGTH, FT
o
o
o
o
1.69 FT/SEC = 1.0 KNOT
SUBSCRIPT "P" REFERS TO
FULL-SCALE PROTOTYPE
SUBSCRIPT "M" REFERS TO
I/I2-SCALE MODEL
Kp IS VELOCITY IN KNOTS
EQUIVALENT FULL SCALE VELOCITY KNOTS, K s-609lo'-A
Figure 6. Relationship of Model Speed to Full-Scale Speed
20
-------
FLOATING WEIR-
ro
NOTE: LENGTH OF ARROW INDICATES
APPROXIMATE VELOCITY,
RELATIVE TO CATAMARAN,
OF WATER FLOW.
S-60930
Figi.re 7. Flow of Water through the Catamaran
if no Bottom is Used
-------
NOTE: LENGTH OF ARROW INDICATES
APPROXIMATE VELOCITY, RELATIVE
TO CATAMARAN, OF WATER FLOW
FLOATING WEIR
4-^
LOUVERED BOTTOM
S-60929
Figure 8. Flow of Water through the Catamaran if the
Louvered Bottom is Installed
-------
NOTE: LENGTH OF ARROW INDICATES
APPROXIMATE VELOCITY, RELATIVE
TO CATAMARAN, OF WATER FLOW
FLOATING WEIR
LOUVERED BOTTOM
FORWARD SECTION
(SPILLWAY)
S-60928 -A
Figure 9. Water Flow through the Catamaran when a Louvered
Bottom and a Forward Section is used
-------
NOTE: LENGTH OF ARROW INDICATES
APPROXIMATE VELOCITY, RELATIVE
TO CATAMARAN, OF WATER FLOW.
.FLOATING WEIR
.-J—,
I
I
J
•FORWARD SECTION (SPILLWAY)
S-6092/
Figure 10. Water Flow through the Catamaran when the Forward Section
is Used without the Louvered Bottom Section
-------
The mode] was also towed backwards at simulated speeds up to 8.2 knots
in the unballasted (lightest) configuration. It towed very well, although
observation of the floats indicated some eddy current in the quiet pond.
From the results of these tests the following conclusions were made:
I. The skimmer barge must have a bottom to the quiet pond if it is to
operate at speeds greater than l/2-knot. A suitable quiet pond can
be produced if a spillway and a louvered bottom is used.
2. The skimmer barge must have a spillway at the bow to minimize the
quantity of water flowing through the quiet pond.
3. If a sloping spillway is used, the skimmer barge must have a near-
vertical bottom at the forward section to prevent the tendency of
the bow to rise out of the water as speed is increased.
4. The quiet pond should be moved forward to the center of the hull to
reduce the effects of barge pitching.
5. The free surface effect of the bottom area will help the barge lift
over large waves.
6. The long sloping spillway was a wave generator and should be greatly
shortened.
7. The skimmer barge will tow backwards very well; this method should
be used for to-site towing.
Second Model Version
After completion of the tests in November, the model was modified to make
improvements and incorporate additional features as shown in Figures II, 12,
and 13. The engine room forward bulkhead was moved forward 6 in. to provide
for a larger engine room, more deck space, and to position the quiet pond
closer to the center of the barge. The 19-1/2-in.-long louvered bottom was
also moved forward an equivalent distance. The IO-l/2-in. spillway with the
10-deg diffuser was cut short to 6-1/2 in. and moved aft so that the aft
vertical bulkhead was in line with the forward end of the louvered bottom. A
six-bladed 7-in.-diameter paddle wheel was substituted in place of the 8-in.-
diameter eight-bladed paddle wheel and its axis was moved aft to the 10-in.
station. Also, the original paddle drive bridge was removed and a new one,
forward of the paddle wheel, was installed.
The paddle drive was mechanized by means of a 12-volt electric motor
driving through a gearbox and a 65-in. V-belt to the paddle wheel. The motor
drive and its 12-volt motorcycle battery were housed in the cabin that was
added to the model. Paddle speed changes were made by changing drive pulley
sizes and a variable resistance in the motor drive circuit.
25
-------
-ENGINE HOUSE
7-FT, 6-BLADED PADDLE V/HEEL
LOUVERED BOTTOM MOVED FORWARD
— NEW BRIDGE
BULKHEAD MOVED
FORWARD-T
HI ^»Ui\-» I I I I J. U
IT r;° PLACESI
SHORTER SPILLWAY AND MOVED AFT
SLOT TYPICAL
NOTE: DIMENSIONS ARE IN INCHES FOR MODEL
(ONE MODEL INCH = ONE FOOT FULL SCALE)
S-60918-A
Figure II, Barge Model as Modified for January 6 Tests
-------
Figure 12. Oil Skimmer Barge I/I2th-Scale Model with
Paddle Wheel
27
-------
Figure 13. Oil Skimmer Barge I/I2th-Scale Model with
Paddle Wheel
28
-------
After modifications, the model was towed in the open waters of Newport
Harbor to evaluate the paddle wheel as a wave attenuator and a surface oil
mover. The tests were only partially successful because the tow boat could
not be operated at a speed slow enough for good model testing. The ability
of the paddle to attenuate waves was not established, but it was a definite
help in bringing the oil surface within the catamaran at the lower towing
speeds.
The configuration was then changed to that shown in Figures 14 and 15,
and the model was tested in rougher water than previously observed. The use
of a sloping surface instead of the paddle wheel appeared to be a better
method of attenuating the wave action. It did not appear to cause as much
disturbance in the quiet pond.
Later the same day the model was taken to Alamitos Bay and towed along
the dock in the same manner the tests of November 24th were run. Per lite
floating on the water was used as an indicator during these tests. Visual
observations indicated that the paddle wheel aided in collecting the Perlite
into the quiet well during operation in quiet water.
The following conclusions were derived from these tests:
(a) The action of the paddle wheel generated a choppy wave condition in
quiet pond.
(b) The paddle wheel, if driven at the proper speed, is an effective
device for moving the oil surface towards the weir inlet. The
speed must be adjusted so that the paddle blade does not cause a
bow wave nor unduly disturb the water surface. This represents a
peripheral speed approximately equal to the forward motion of the
skimmer. At skimmer speeds of less than one knot, the paddle can
be operated at a peripheral speed of approximately one knot without
unduly disturbing the water surface.
•x
(c) The sloping-surface convergent passage of the spillway is an effec-
tive wave attenuator.
Thi rd Model Version
A floating weir as shown in Figure 16 was added to the skimmer model. An
automotive electric fuel pump was used to pump the oil and water from the wei.r
to an adjacent receiver. This pump, shown in Figure 17, was also powered by
the 12-volt motorcycle battery and had a pumping capacity of 0.5 gpm. In
order to reduce the sloshing in the quiet pond, two vertical gratings, called
wave fences, were installed as shown in Figure 18.
The model as shown in Figures 19 through 21 was tested by towing it
behind a small skiff in Marina del Rey. The original intent of the test was
to recover oil from the quiet pond by means of the floating weir while the
model was being towed through a simulated sea condition. The pump, however,
failed to work properly so this function was tested during only a few inter-
mittent periods of operation. These results were sufficient to indicate
feasibility of the operation.
29
-------
POSSIBLE NEW LOCATION FOR
SMALL PADDLE WHEEL
FLOATING WEIR
ADJUSTABLE SURFACE FOR
ATTENUATING WAVES
Cx
o
S-60926
Figure 14. Configuration of Model for Converging Passage
Type Wave-Attenuator (January 6 Tests)
-------
•
•
Figure 15.
Oil Skimmer Barge l/!2th-Scale Model with
Sloping Board Replacing Paddle Wheel
31
-------
MATERIAL: 0.040 ALUMINUM ALLOY SHEET
FLOTATION CHAMBER
SUCTION PIPE WELL
1.83
S-60921
TRANSFER TUBE
Figure 16. Floating Vlelr Model (l/12-ScaVe)
-------
Figure 17. Oil Skimmer Barge l/!2th-Scale Model Paddle
Wheel Drive and Weir Pumping Systems
33
-------
FLOATING WEIR
SUCTION LINE
ADJUSTABLE SURFACE FOR
ATTENUATING WAVES
TRASH SCREEN
WAVE ATTENUATING FENCES
LOUVERED BOTTOM
FORWARD SECTION
S-60925
Figure 18. Model Configuration during the Tests of January 12, 1970
-------
Figure 19. Oil Skimmer Barge I/I2th-Scale Model
35
-------
\
Figure 20. Oil Skimmer Barge l/12-Scale Model
36
-------
Figure 21. Oil Skimmer Barge l/!2th-Sca1e Model
37
-------
The model was towed under a number of conditions and it was observed by
film review that the waves around the weir in the quiet pond were smal lei-
after the wave fences were installed.
The following conclusions were made:
(a) The wave attenuating fences/ as shown in Figures 18 and 21, visually
appeared to reduce the waves and surges in the quiet pond.
(b) The floating weir will operate in the environment of the quiet pond
under sea conditions through Sea State 3 as the model was tested in
4-in. waves, which, to scale, are equivalent to Sea State 3.
FULL-SCALE SKIMMER BARGE DESIGN AND CONSTRUCTION
The general configuration of the full-scale skimmer barge was determined
from the model tests described in the previous section. The vessel is 45 ft
long, 26 ft across the beam, and 8 ft from keel to main deck. It was designed
in accordance with U.S. Coast Guard requirements for steel tank barges and was
constructed primarily of 1/4-in. steel plates at Todd Shipyards, San Pedro,
California. While the original design concept included permanently installed
pumping equipment with a single diesel engine power source below the main deck
in an engine room, the barge was built without these refinements to reduce
cost. Instead, all pumping was accomplished by means of rental diesel engine-
driven units mounted on the main deck. All of the piping was, therefore,
placed above the main deck.
The basic hull shown in Figure 22 weighed 59,000 Ib at launch and when
fully equipped with all the wave suppressors, weirs, and processing equipment,
weighed approximately 90,000 Ib. At this weight, it would float at nominally
the 2-ft 10-in. waterline. A displacement curve for the barge is shown in
Figure 23. To sink to the nominal operational waterline of 5 ft 0 in., approx-
imately 80,000 Ib of ballast water had to be added to bring the vessel dis-
placement up to 170,000 Ib.
The hull was divided into II watertight compartments as shown in Figure
24. Four of these (3P, 3S, 4P, and 4S) were used as reserve buoyancy compart-
ments and check valves were installed in their bilge pipes so that they could •
not be inadvertently filled with water. The engine room was similarly protected
The remaining compartments could be filled or emptied at a rate of 150 gpm to
trim the vessel properly. The displacement of each of the individual compart-
ments is shown in Figure 25.
A simplified piping schematic diagram was shown in Figure I and the de-
tailed pumping arrangements are shown in Figure 26. Any one of the three
weir pumps can also serve for bilge pumping and ballast pumping through con-
nections with the bilge and ballast manifolds. The oil transfer pump is also
connected to the seawater inlet so that it can be used as a source of high-
pressure (60 psi) water for general purposes such as washdown. Underwater
viewing ports were installed in compartments 3$ and 4P.
38
-------
Figure 22. Sk:mmer Barge Basic Hull
39
-------
300
250
tn
Q
'
•
. ,
CO
I I
!
LU
CJ
00
I-H
o
200
50
100
50
:
i;
^:i:i=EE = iS
/
:
•
/
/
;
/
!;!==IiI
E
APPROXIMATELY LINEAR
AT 34,200 LB/FT
tttt
I
p;t=
ffl
FT
Hi
DRAFT, FT
Figure 23. Skimmer Displacement Curve
-------
S-60501 -A
Figure 24. Skimmer Hull Compartments
-------
80,000
CQ
3
o.
C/>
70, 000
60, 000
50,000
40,000 -A
30,000 :-
20,000
I 0,000
37,400 LB
••[• 22,550 LB
DRAFT, FT
S-60531
Figure 25. Individual Compartment Displacements
4 2
-------
VENT
PORT
V40 V39Y V38 V37
V35Z V34T V33T V32T V3!
FLOATING
WEIRS V2
"1
1
1 I
BLEED
r
V56 |
_ WATER OUT
•_ ._- —
V20
CENTRIFUGE
V57
V22
VI9
V 47
TOIL
I PUMP
CHEST
V23
TO
STORAGE
S-59667 -A
Figure 26. Oil Skimmer Barge Piping Schematic
-------
Figure 27 is a fluid-flow process diagram showing various items used in
the skimming process; their functions are described below.
Wave Gate (Figure 28)
The purpose of the wave gate was to help maintain a smooth surface in the
quiet pond by blocking waves from entering. The amount of water permitted to
enter was controlled by the gate opening. Wave buildup against the gate in-
creased the velocity through the opening, and this velocity was dissipated in
the wave fences described below.
Wave Fences (Figure 29)
The wave fences were designed to reduce the sloshing within the quiet
pond by restricting the flow of the water and dissipating the energy in a
multitude of minor turbulences.
Louvered Bottom (Figure 50)
The bottom of the quiet pond had a series of slotted openings (louvers)
to permit the water taken in over the spillway to pass out through the bottom
of the quiet pond. This created an environment of still water in the region
of the floating weirs, since the water in the quiet pond, especially that near
the surface at the aft end, was carried along within the skimmer at skimmer
velocity. If there were no bottom, water passing under the barge would well
up and cause currents in the quiet pond.
Self-Adjusting Floating Weir (Figure 5l)
Oil was skimmed from the surface of the quiet pond by means of three
identical self-adjusting, floating weirs. These weirs were self-adjusting
in the sense that they would follow the surface of the water and skim at what-
ever rate the skimmed liquid was pumped from them. This feature enabled a
weir to be operated at any flow rate, from zero to the full flow of 170 gpm
and still effectively skim the surface oil. Under very light oil and smooth
water conditions the weir can be operated at approximately one percent of the
maximum rate, thereby increasing the weir efficiency (oil/water ratio) by
these two phenomena:
(a) At the lower flow rates the weir is self-adjusting to skim only a
few thousandths of an inch of the top surface of the quiet pond.
The lower limit is dependent upon the surface tension and the con-
dition of the quiet pond surface (waves and currents). At low flow
rates the flow over the weir may be intermittent if the effect of
surface tension or oil viscosity exceeds the average depth of sub-
mergence of the weir.
(b) A second effect is that with a lower flow rate over the weir for a
given flow rate of oil into the skimmer, a thicker layer of oil will
accumulate in the quiet pond. This will increase the oil/water rati<*
of the top surface as skimmed by the weir.
44
-------
TO OIL oil
STORAGE PUMP
SURGE
TANK
I WATER/OIL
-—-^\ OVERFLOW
FLOATING WEIR
r—WAVE FENCES
WAVE GATE
QUIET POND —
SPILLWAY
WATER
OUT
LOUVERED BOTTOM
S-60812
Figure 27. Fluid Flow Processing Diagram
-------
FORWARD BRIDGE v
3500 GPM PUMP BRIDGE
SKIMMER KEEL LINE
S-60562 -B
Figure 28. Wave Gate
46
-------
6 FT
S-60561
Figure 29. Wave Fences
-------
FWD
00
ENGINE ROOM
SPILLWAY
NOMINAL WATERLINE
QUIET POND
I-BEAM
12 INr
JU
Z
6 IN.
T
12 IN-H h*-2 IN.
S-60537 -A
Fvcxure 30. \_ouvered Bottom
-------
TO'PUMP
SUCTION HOSE
WEIR
BRIDGE
STRUCTURE
WEIR
TRUNNION
COUNTERBALANCE WEIGHT
LWEIR SUPPORT PIVOT S-60564
Figure 31. Self-Adjusting Floating Weir
-------
Weir Pumping
Each weir was pumped by means of a diese] engine-driven 3-in. centrifugal
trash pump of 170 gpm capacity. These pumps had a self-priming feature and
would reprime if the weir were momentarily pumped dry. The weir was connected
to the pump inlet by means of a 6-ft-long section of flexible hose that allowed
the weir complete freedom of movement (Figure 32). During pumping this hose
would automatically seek a minimum volume condition, since the interior was
at a negative pressure. This caused the hose to run in a straight line between
the weir attach point and the pump inlet pipe, while at the same time it was
completely flexible as to changes in length.
The three weir pumping units were rented from John W. Stang Corporation
of Orange, California and are more completely described in Figure 33.
Surge Tank (Figure 34)
All of the discharge from the weir pumps was fed into the surge tank.
The inlet was located tangentially at the 4-ft level at a nominal 18 in. be-
low the surface.
The liquid level in the surge tank was controlled by an overflow weir
that was fed from the bottom of the tank. This permitted overflow of only
the cleanest water.
The surface of the liquid in the tank tended to remain level as the
barge pitched and rolled beneath it. Since the roll and pitch rates were so
slow and the roll and pitch angles were of such small magnitude, no sloshing
over the tank walls occurred. Also, in normal operation the tank surface was
covered with a thick layer of viscous oil that further reduced sloshing.
Oil was drawn from the circular oil overflow weir on the vertical axis
of the tank where the influence of pitch and roll had the minimum effect.
The centrifuge inlet water was drawn from the bottom of the surge tank
at rates up to 500 gpm. The centrifuge effluent water was discharged into
the bottom of the quiet pond so that any entrained oil could be reprocessed
back through the system.
Oi1 Transfer Pump
The surge tank oil overflow and the centrifuge discharge oil were col-
lected in an oil transfer barrel at the inlet of the oil transfer pump. This
pump, described below, transferred the oil to forward end of the skimmer
(for closed-circuit skimmer testing), or to oil-receiving barrels or tanks.
Oil transfer pump characteristics are as follows;
Prime mover Petter Type ABI Diesel Engine
Pump Roper Model 3600 GHB
50
-------
Figure 32. Weir Pumping
51
-------
B. PERFORMANCE TABLE*
A. CHARACTERISTICS
3-IN. SELF-PRIMING TRASH PUMP
HANDLES' UP TO I-I/2-IN. DIA SOLIDS
NONCLOGGING VOLUTE PRIME SYSTEM
3-IN. THREADED MALE SUCTION AND DISCHARGE
SELF-LUBRICATING, STAINLESS STEEL,
MECHANICAL SEAL
BASE MOUNTED WITH LIFTING EYE
HEAVY DUTY IMPELLER AND VOLUTE
REMOVABLE ELBOW AND COVER FOR ACCESS TO
ALL WORKING PARTS FOR CLEANOUT AND
MAINTENANCE
RENEWABLE WEAR PLATE
3-IN. SUCTION LINE STRAINER
SUCTION CHECK VALVE
HAND OPERATED, 1-1/2 IN. FILLER PLUG
CAPACITY, U. S. GPM
TOTAL
HEAD
FT
30
40
50
60
70
80
90
100
PSI
13.0
17.3
21.6
26.0
30.3
34.6
39.0
43.3
TOTAL STATIC SUCTION LIFT,
FT
5
372
334
290
248
200
159
113
70
10
333
332
290
248
200
159
113
70
15
278
276
275
248
200
159
1 13
70
20
217
217
216
216
200
159
113
70
25
-
150
150
150
150
149
1 13
70
*Continuous service (governed throttle)
2
o
120
100
80
60
40
20
PERFORMANCE CURVES
50 100 150 200 250 300 350 400 450
CAPACITY, U.S. GPM
"BASED ON NOMINAL SIZE SUCTION LINE 5 FT
LONGER THAN STATIC SUCTION LIFTS SHOWN.
S-6060I
Figure 33. Weir Pumping Unit Performance
(J. W. Stang Mode] 3CRI8EL)
52
-------
•CENTRIFUGE
INLET SCREEN
WATER/OIL TO
CENTRIFUGE
OIL
OVERFLOW
TO OIL
RESERVOIR
OIL OVERFLOW WEIR
WATER
OVERFLOW
WEIR
WATER OVERFLOW
AND RETURN TO
QUIET POND
SEDIMENTS
DRAIN TO
O.UIET POND
Figure 34. Surge Tank
S-60563 -
53
-------
Type He 1ical gear
Size 2-in. inlet and discharge
Capacity 60 gpm at 500 rpm
Operating speed 500 rpm
Relief valve setting 60 psig
Paddle Wheel
On September 16, 1970 the wave gate was replaced with the paddle wheel as
shown in Figure 35. This device was designed to provide a pumping action at
the entrance to the skimmer to sweep in oil when the skimmer was stopped or
traveling at very low speeds and to act as a check valve for oil already in
the quiet pond. It also reduced the effect of waves entering the skimmer by
breaking them up.
The paddle wheel was driven by means of a small diesel engine through a
V-belt drive, a 30:1 worm gear speed reducer; and a chain drive. The V-belt
sheave sizes could be changed for coarse speed selection, and a fine adjust-
ment of A:I could be made by adjustment of the engine governor. The centerline
was located 26 in. aft of the lip of the spillway, as shown in Figure 36. The
paddle wheel was vertically adjustable with a nominal clearance of I in.
between it and the top of the spillway.
SKIMMER BARGE TESTS
The skimmer barge, incorporating the centrifuge described in Section VII>
was tested at sea as described in Section VIII. The tests demonstrated the
feasibility of skimming oil from the ocean surface. The following is a dis-
cussion of the test results (see Section VIIl) as they pertain to the skimmer
barge, and describes notable features, problem areas, and possible improvements
that could be made.
Wave Gate
Prior to installation of the wave gate on September 16, 1970 conditions
in the quiet pond were such that when the barge was towed upwind at 2 to 5
knots in Sea State 3, the weirs sloshed and rocked to such an extent as to be
ineffective as skimming devices. After installation of the wave gate the
weirs could be operated going upwind at speeds up to 5 knots under any sea
conditions experienced during the test program (Sea State 0 to Sea State 3).
When closed, the wave gate also afforded complete protection to the quiet
pond during high speed towing to the deployment site. Under many other con-
ditions, however, the wave gate tended to increase the disturbances in the
forward end of the quiet pond and also was capable of causing a bow wave from
wave reflections off the submerged gate. This gate should be readily adjust-
able under any sea conditions and forward velocity to obtain the maximum benefit
which would require a hydraulic cylinder, rack and pinion, jackscrews, or some
54
-------
WORM GEAR
SPEED REDUCER
01
cn
DIESEL ENGINE
S-6064S -A
Figure 35. Paddle Wheel Installation
-------
PADDLE WHEEL DIAMETER
PADDLE WHEEL LENGTH
NUMBER OF PADDLES
PADDLE HEIGHT
36 IN.
18 FT. 10 IN.
6
8 IN.
PADDLE WHEEL
SPILLWAY
SKIMMER
KEEL
LINE
S-60649 -A
Figure 36. Side View of Paddle Wheel
56
-------
similar device. The present model had a worm-gear cable hoist to lift the
gate; it is then bolted in the desired position. This could not be done when
waves were striking the gate. Possibly the addition of devices to break up
the wave so it is not reflected from the smooth surface would enhance its
performance.
Wave Fences
During the first at-sea test with the skimmer barge (skimming tow tests
on June 3., 1970) it was observed that waves entering the skimmer or generated
within the skimmer by "sloshing" reflected off the aft bulkhead of the quiet
pond and increased the surface disturbance in the quiet pond. Also, waves
directed at the aft corners of the quiet pond would splash up over the deck.
After the wave fence was installed just ahead of the aft bulkhead, the wave
reflections no longer occurred and water was not observed to splash up on the
afterdeck from waves in the quiet pond.
Use of the wave fences in the forward part of the quiet pond markedly
increased the local disturbance, but the overall effect was the reduction in
size of the waves.
Louvered Bottom
Water in the quiet pond was carried along with the barge, and it was
repeatedly observed that objects or oil floating on the surface of the quiet
pond had practically no velocity with respect to the barge, even when the
barge was traveling through the water at speeds above 5 knots. It was
observed through the aft underwater viewing port that submerged particles had
very low velocity and a single particle could be observed for several seconds
within the limited range of view of the port. The oil on the water surface
in front of the floating weirs moved slowly toward the weirs as the surface
was drawn off. These observations were in agreement with observations made
during the model test with the louvered bottom.
Self-Adjust ing Floating Weir
Proper operation of the-weir was dependent upon a rather delicate initial
balance, which was adjusted by means of a weight bar (Figure 31). Once ad-
justed for a particular weir, the adjustment never needed to be changed. A
second adjustment was the counterbalance weight, which tended to make the
weir assembly float higher or lower in the water. This adjustment did have to
be changed occasionally, primarily to allow for changes in average wave condi-
tions in the quiet pond.
Although a weir operated for hours unattended, an operator was stationed
on the weir bridge to tend the weirs. His principal occupation was to remove
kelp and other floating objects from the quiet pond and occasionally to change
the trash screens in-the weir inlet.
57
-------
When a weir was not in use it was hoisted free of the water by means of
a permanently attached hand winch assembly (Figure 37). This was done to
prevent excessive movement during high speed towing and rough sea conditions.
Wei r Pumps
During the entire test program, there was never a need to disassemble
any of these pumps for cleaning, although they did have an easily removable
face plate to facilitate cleaning. (The diesel engines that drove these pumps
also performed without malfunctions for the entire test program.)
This type of centrifugal pump, although ideally suited as far as trans-
ferring the oil/water mixture from the weir to the surge tank, tended to
emulsify the oil. This was not a significant problem with the thicker oils,
but when 35° API gravity oil was used, a large amount of oil was observed in
the surge tank overflow. Although this was still readily removed by the
centrifuge, it would be more convenient to remove all the oil possible in the
surge tank.
Surge Tank
The location of the inlet caused mild circular circulation within the
tank, with the oil migrating to the center top and the cleaner water to the
bottom outside. This circulation was later obstructed somewhat by the in-
stallation of screens across the outlet pipes (Figure 38) to prevent the flow
of kelp pieces and other undesirable objects into the centrifuge. This tank
inlet also seemed to be a little too close to the top and unnecessarily dis-
turbed the surface water.
The overflow weir allowed only the cleanest water to be returned to the
quiet pond. In fact, under low flow conditions (50 gpm) the surge tank over-
flow was visually judged to contain less than 25 mg/1 since samples of the water
taken in 4- or 16-oz jars showed no oil film on the water surface. This water was
returned to the quiet pond, as was the centrifuge discharge water, so in the
event that excessive oil were present it would be reprocessed and removed.
During operation, the flow demands of the centrifuge and the output of
the weir pumps had to be adjusted so that the weir pumps always pumped at a
slightly higher rate than required by the centrifuge in order to assure a full
surge tank. The excess was handled by the surge tank overflow weir.
When a significant amount of oil was collected on top of the water surface
in the surge tank, it was drawn off by means of a centrally located conical
overflow weir. This oil drained through a 3-in. pipe into the oil transfer
barrel.
58
-------
Figure 37. Weir Winch
59
-------
Figure 38. Centrifuge Inlet Screen Installation in Surge Tank
60
-------
SECTION VII
OIL/WATER CENTRIFUGE
ASCRIPTION
The oil/water centrifuge, mounted on the skimmrner deck as shown in Figure
'> was used to separate the oil/water emulsions collected in the surge tank
Or> the skimmer barge. A centrifuge is ideally suited for separating substances
°f differing densities because it uses the centrifugal forces produced by
r°tational motion to create an intense gravitational field- The higher the
""otational speed, the larger the gravity field, and hence the larger the
*orce causing separation of the different density particles- The larger the
SeParation force, the faster the separation of these particles takes place-
The oil/water centrifuge was designed for the present application and
operated on a continuous basis utilizing an annulus-type centrifuge barrel.
'he annulus-type centrifuge barrel consists of two concentric cylinders that
Rotate as a single assembly, as shown in the simplified cross section of
F'9ure 39.
As the oil/water mixture is pumped into the centrifuge the fluid encounters
end plate of the inner barrel- The end plate directs the mixture radially
°utward to the annulus where it flows axially to the discharge end of the
Centrifuge barrel. At the same-time the-flow is also rotating with the barrel.
While traveling the length of the annulus, the oil/water mixture is subjected
to the tremendous forces produced by the high rotational speed of the centri-
fuge barrel. These forces separate the two fluids before the fluids reach
the discharge end of the barrel; then the two fluids are physically separated
bV the flow splitter. The oil flows radially Inward between the flow splitter
ar|d the inner barrel. Water, which gravitates to the outside, flows outside
of the flow splitter and then radially Inward between the flow splitter and
the outer barrel. The fluids are discharged along the axial centerline in
SeParate concentric pipes.
For this type of centrifuge, control of the oil quality Is accomplished
bV adjusting the back pressure at the oil outlet with a pressure regulator.
DESIGN AND CONSTRUCTION
The oil/water centrifuge was designed to process oil/water emulsion at
a rate of 500 gpm under a centrifugal force field 4000 times the normal
9l"avltat ional force. The horizontal drum, which consisted of two closed
concentric cylinders, was supported at each end by ball bearings and weighed
8°0 lb dry. The drum was rotated by means of a belt drive from a 100-hp
djesel engine, which also drove a 500-gpm centrifugal feed pump, a 7.4-cfm
a'i" compressor, and other minor accessories, making the separator a self-
contained package. The entire package, including the drive, weighed 8000 lb
required approximately 64 sq ft of deck space.
61
-------
BUBBLE VENT CONTROL
ROTARY SEALS
BEARING »
OIL CONTAMINATED
BALLAST WATER
(FROM PUMP)
INNER
BARREL
EMULSION
OUTER
BARREL
OIL
WATER
FLOW SPLITTER
WATER OUT
ROTARY SEALS
BACK PRESSURE CONTROL
RECOVERED
OIL OUT
S-54 298-A
Figure 39. Cross Section of Oil/Water Centrifuge
62
-------
When mounted aboard the skimmer barge, the centrifuge package admitted
the oil/water emulsion through a 4-in. flanged pipe connection, and the clean
water was discharged back into the quiet pond through a 6-in- pipe- The
recovered oil was discharged into the oil reservoir through a I-1/2-in. pipe.
PERFORMANCE TESTS OF CENTRIFUGE
Figure 40 shows the first version of the centrifuge, which was tested at
the Signal Oil and Gas Company field laboratory in Huntington Beach- The test
objective was to determine performance over a variety of conditions of flow
rate, centrifuge speed, and inlet feed emulsion concentration to provide the
oasis for any required design improvements.
The test setup included instrumentation as shown schematically in Figure
*' and in the photographs in Figure 42.
A 4-in. line supplied the centrifuge with seawater from the Signal Oil
a"d Gas Company water wells at a rate in excess of 500 gal/min. This water
Contained no oil, but it did contain a small amount of iron sulfide particu-
'ate matter that could be easily dissolved by addition of a small amount of
hydrochloric acid. Oil was added to the seawater by pumping it from an open-
top 15,000-gal Baker tank through a wobble-disk flowmeter and throttling
^alve system, and then into the seawater line downstream of the seawater
urbine flowmeter. As an alternate the oil could be injected through'the air
b'eed line that led directly into the centrifuge barrel. The air bleed line
Was located downstream of the feed pump and throttling valve.
After the oil was separated from the seawater, it was discharged from
the centrifuge through a wobble-disk flowmeter into a sump. A floating weir
^claimed the oil from the surface of the sump and transferred it to a second
5,000-gal Baker tank. The seawater supernate was discharged from the centri-
U9e directly into a waste water sump.
During operation all fluid lines were instrumented for temperature,
Pressure, and flow monitoring, whereas the centrifuge was monitored through
Panel gages and additional temporary instrumentation. Panel instrumentation
'^eluded inlet and outlet pressure on the centrifuge plus the oil discharge
Pressure for the centrifuge. In addition to the normal operating instruments
°r the engine, an oil mist pressure gage displayed the manifold pressure
or the oil mist bearing lubrication system- In conjunction with the bearing
'ubrication system, a bearing temperature monitoring unit formed a part of
the temporary instrumentation. This unit was a chromel-constantan thermo-
jjouple transducer that read out bearing temperatures directly in degrees
^ahrenheit- Each bearing had one thermcouple placed adjacent to the outer
ing race and welded to the upper portion of each pillow block. The
t bearing pillow block contained a total of two thermocouples; however,
second thermocouple was positioned to sense the outboard angular contact
ing.
63
-------
*
'
Figure 40. Oil/Water Separator PN 585010-1-1
64
-------
BEARING TEMPERATURE
SEAWATER
o
01
CENTRIFUGE SPEED
BEARING
TEMPERATURE
FLOWMETER
CRUDE OIL
OIL PUMP
WATER
DISCHARGE
SAMPLE
PORT
PRESSURE
I REGULATOR
FLOWMETER
S-60920
Figure 41. Schematic of Centrifuge Test Setup
-------
68039-19
Figure 42. AiResearch Oil/Water Centrifuge
66
-------
Centrifuge speed and seawater flow were sensed by magnetic pickups and
displayed on a counter in Hertz. A simple conversion facto.- of 1.25 times the
speed frequency gave the rpm of the centrifuge. A calibrated graph was required
to convert the seawater flow frequency to gpm-
After completing the maintenance and startup procedures, a given set of
conditions was selected for each test run. The seawater flow was adjusted to
the preselected rate by first opening the supply line valve until 5 to 10 psi
^9s recorded on the feed pump inlet pressure gage. Next the main pump dis-
charge valve on the test panel was opened until the pump inlet pressure
decreased to less than I psi- This procedure, which increased the flow rate
of seawater through the pump, was performed in reverse order to decrease the
flow rate. Parallel to seawater flow rate adjustment, the oil flow rate set
Point was established. A 50-psi pressure was arbitrarily established for the
discharge from the oil supply pump and maintained through the test program.
^°st of the oil was bypassed through the centrifugal oil pump while a small
Percentage was drawn through a throttling valve, metered, then injected into
the seawater line to form the test emulsion. The entire process took from
^to 15 minutes depending upon the emulsion concentrations- Low oil concentra-
tions took longer to establish than high ones. After the oil and seawater
fjow rates were established, a stabilization period of approximately 2 to 3
111111 passed before samples were taken-
Three seawater effluent samples were taken during a 3-tnin period using
separatory funnels. After collecting three samples, an established
v°lume of trichloroethylene was added to each. The oil was extracted from
the water by the trichloroethylene, which was then drawn from the funnel,
''tered, and colorimetrically measured. (See Appendix 2.) Contamination
evels were expressed as parts by weight of oil per million of water-
DISCUSSION OF PERFORMANCE RESULTS
The centrifuge was placed into the field laboratory on March 2, 1970 and
Was subjected to a comprehensive performance test program. After a few tests
Wei"e performed it became apparent that an insufficient oil discharge flow
Passage within the centrifuge was restricting the oil discharge. The oil
svel in the centrifuge was too high and was being carried over the discharge
|ow splitter with the clean water. On March 6, 1970, the centrifuge was
^'sassembled in the field and a larger diameter flow splitter was installed.
"is modification was made to increase the oil discharge pressure and flow
^te by increasing the pressure drop across the flow splitter- As the follow-
If19 test results indicate, however, a second more extensive modification was
^quired to bring the unit up to its present performance level. Figures 43
hrough 45 display the performance of the originally designed unit.
Centrifuge speed as a function of engine speed is shown in Figure 43 with
^second curve showing the point of minimum torque converter slip- Based on
V\'s curve the most economical operation speed would be 2,000 rpm for the
system and 3,200 rpm for the centrifuge. This does not limit the
to 3,200 rpm; it merely points to the optimum power conversion
°r this sheave and drive system-
67
-------
4UUU
3500
•zf\(\r\
5UUU
X
OL
* 2500
•\
0
UJ
UJ
a.
CO
LU
£ 2000
t— <
a:
h-
z
UJ
1500
1000
500
PRIMAR'
CENTRIF
/ SHEAVE OD
JGE SHEAVE
, /
x.
= 14.0 IN.
3D = 7.1 IN.
/
/^
/
/
/
V
\
\
^
^
s
\ 1
- IV> CM J> t/1 C> -~
3 O O O O O C
TORQUE CONVERTER SLIP, PERCENT
1 V
500 1000 1500
DIESEL ENGINE SPEED, RPM
2000 2500
S-60909 -A
Figure 43. Centrifuge Speed and Torque Converter Slip
68
-------
TEST DATE 3/18/70
AVERAGE CENTRIFUGE SPEED, 2750 RPM
INLET EMULSION CONCENTRATION 0.5 to 3.0 PERCENT OIL
LARGEST WEIR 585037-5
OIL GRAVITY = 2I.5°API
100
90
30
70
60
S 50
ui
DC
LU
o:
a.
30
20
PUMP OUTPUT PRESSURE
REQUIRED CENTRIFUGE INLET PRESSURE
200
300
400 500 600
EMULSION FLOW RATE, GPM
S-60908 -A
Figure 44. Pump Output Head and Separator Inlet
Pressure Requirements
69
-------
CJ
ai
ouu
200
100
0
400 gpr
300 c
^
L3
200
ORIGII
CONFK
OIL G
AT CO
INLET
n. /
pm .x^
) gpm
| 0
ML CENTRIFU
DURATION (58
CAVITY =21.
NSTANT 5,000
CONCENTRATI
IE
5010-1-1)
mg/1
ON
2000
2500
3000
3500
4000
CENTRIFUGE SPEED, RPM
S-61880
Figure 45. Abnormal Centrifuge Performance
70
-------
Figure 44 shows the pump output head and centrifuge inlet pressure as a
function of the flow rate through the unit. The centrifuge feed pump, when
driven at 1800 rpm, delivered an average pressure of 73 psig at flow rates from
200 to 500 gpm. The centrifuge inlet pressure requirement shown in this figure
Was well below the output of the pump and the throttling valve was used to
drop the pump discharge pressure to meet the centrifuge inlet requirements.
One of the most important facts obtained from this plot is that 500 gpm would
flow through the centrifuge at less than 50 psi inlet pressure.
Figure 45 is a plot of the data taken with the original centrifuge. These
data indicate that the centrifuge system does not provide the expected or
calculated performance in that the unseparated oil increases as speed is
increased- The basic equations governing gravitational forces within the
centrifuge and the rate of particle settling within these force fields indi-
cate that higher rotational speeds would give better separation if resident
times were held constant. Inspection of Figure 45* however, revealed the
°Pposite was true for the original unit. Better separation occurred at the
400-gpm fiow rate when the lower centrifuge speed of 2750 rpm was used. Two
Possible explanations for this anomaly were considered and either one, or a
combination of these, could have produced the measured results:
(a) Since the centrifugal feed pump was directly driven by the diesel
engine, as was the centrifuge, it was not usually matched with
flow rate requirements- Figure 44 shows this to be true by com-
paring the pump discharge pressure and centrifuge inlet pressure.
The pump produced roughly twice the required head- The pressure
head produced by the feed pump was created by restricting the
throughflow with the throttling valve-
(b) When dissimilar fluids are passed through a restriction such as a
partially opened valve, a mixing process occurs- Large droplets
are broken into numerous small droplets, which are well mixed
within the seawater.
The combination feed pump and throttling valve apparently created
'arge numbers of small particles that were not being separated by the centri-
fuge- When the centrifuge speed was increased to effect a better separation,
Poorer separation occurred and more oil was discharged with the effluent
Water." Apparently the increased centrifugal speed could not overcome the
detrimental effect of increased feed pump speed. The separating force was
^ore than offset by the creation of the smaller droplets-
In Figure 46, quantitative data was taken at a constant centrifuge speed
of 2750 rpm and the inlet emulsion was held constant at 5,000 mg/1 oil in the
Seawater. At a water flow rate of 400 gpm the centrifuge discharge water
contained approximately 200 mg/1 oil when the oil was injected upstream of
the feed pump. When the oil was injected at the air bleed line, which is
downstream of the feed pump system, the discharge water contained approxi-
^ately 120 mg/1 oil. These results, plotted in Figure 46, indicate that a
^0 percent decrease in the contamination level of the discharge water occurred
when the oil was injected downstream of the pumping system.
71
-------
400
0.5 PERCENT INLET OIL EMULSION
LARGEST WEIR 585037-6
CENTRIFUGE AT 2750 RPM
OIL GRAVITY = 2I.5°API
300
UJ
C3
£ 200
Q
O
_ 100
OIL INJECTED UPSTREAM
OF FEED PUMP
OIL INJECTED THROUGH
A AIR BLEED PORT.
100
200
300
400
500
EMULSION INLET FLOW RATE, GPM
S-61879
Figure A6. Crossplot of Original Performance Measurements
72
-------
Based on the preceding data, it was concluded that one or more of the
following decreased the separation efficiency of this unit:
(a) A turbulent flow pattern existed internally within the discharge
section of the centrifuge, which disrupted the oil/water interface
and carried oil over the weir
(b) The average oil droplet size in the feed emulsion was decreased by
overspeeding the centrifugal feed pump, which created an emulsion
that was more difficult to separate
(c) The throttling valve downstream of the feed pump becomes an
emulsifying device
(d) High water flow rates create a flow pattern disturbance that may
cause water to be carried with the oil down the I/2-in.-diameter
discharge tubes
DISASSEMBLY AND INSPECTION
The centrifuge was disassembled and inspected for obvious mechanical
^ilures, and two separate problems were found. First, the carbon face seal
at the inlet end was found filled with fine sand, which rendered it mechanic-
a'ly inoperable and caused it to leak. The same type seal was used on the
°Utlet and although it had been leaking during the performance tests it was
n°t filled with sand. The distortion of the mating surfaces caused by the
d^ive torque on the outlet end was believed to have caused the leakage
Witnessed during operation- The carbon face seals on the inlet and outlet
should be changed, or their installation improved, to stop the leakage flow.
The second problem area was corrosion, which also affected the seals as
1 as the pump vanes, end caps, turbine vanes, and the inlet and outlet
sings- Figure 47 shows a corrosion pattern developing in the end cap.
F'gure 48, a photograph taken of this same part just prior to the design
Codification, shows that all surfaces exposed to the moving seawater were
P°rroding more rapidly than the stagnant areas found between the turbine vanes-
figure 49 shows a closeup view of a turbine vane in which all surfaces were
e9inning to corrode, and Figure 50, an inside view of the discharge housing,
feveals the blister type corrosion that had developed during exposure to
seawater.
As a result of the conclusions reached after testing and disassembly,
following recommendations were made:
(a) The internal weir should be redesigned to allow a 360-deg flow on the
oil discharge side. The original configuration involved eight, sepa-
rate I/2-in tubes. The preliminary tests indicated that the pressure
drop and turbulence associated with this design is excessive- Also,
the downstream oil discharge plumbing was enlarged to decrease the
pressure drop and reduce the possibility of plugging.
73
-------
Figure 47. End Cap Seawater Corrosion (March 6, 1970)
Figure 48. End Cap Seawater Corrosion (March 21, 1970)
74
-------
^§0
Figure 49. Corroded Discharge Vane, PN 585014
Figure 50. Corroded Discharge Housing, PN 585024
75
-------
(b) The feed pump should be disengaged from the diesel drive and be
driven by a varidrive motor- This would allow the centrifuge
pump to run at optimum speed, so that the degrees of emulsification
of inlet oil and water can be kept to a minimum-
(c) The throttling valve downstream of the feed pump should be removed-
(d) The face seals should be protected from particulate contamination-
The inlet face seal should incorporate a flow diverter that would pre-
vent particulate matter from backing up into the seal- Both the water
outlet face seal and the oil outlet face seal should be replaced by
the dual lip seals with a bleed line located between them-
(e) An automatic pressure-controlled flow regulator should be installed
on the oil discharge line- This minimizes the water content in
the oil outflow.
The centrifuge was modified to incorporate recommendations made in (a) and
(e); the other recommendations required more extensive design changes and were
not incorporated at this time-
DESCRIPTION OF THE DESIGN MODIFICATIONS
Cross sections of the original and modified designs are shown in Figure
51, with all modified parts shaded in the lower view; individual components
are shown in Figures 52 through 55- These components were painted with one
of three epoxy type corrosion-preventive compounds specifically formulated
for seawater environments-
Figure 52 shows an overall and closeup view of the inlet pump vanes as
they were installed onto the internal drum. These vanes were coated with a
corrosion-preventive epoxy type paint manufactured by Amercoat Corporation
and designated as No. 93/84.
The end cap surfaces shown in Figure 53 have been cleaned and painted
with Mogna Coatings protective primer and topcoat No. 4-G-I4-4 and 4-W-I-4.
Figures 54 and 55 show two views of the redesigned flow splitter-
Figure 54 is an overall view of the flow splitter, which has been painted
with Product Techniques PT750 corrosion-preventive compound. The stainless
steel oil discharge tube was assembled into the flow splitter and the final
configuration is shown in Figure 54. Figure 55 shows the flow splitter
without the oil discharge tube mounted onto the internal drum. Close in-
spection of the center port shown in this view reveals the l/4-in.-wide oil
flow passages between the flow splitter and Internal drum.
Figure 56 is a photograph of the automatic oil discharge control system.
A bypass line was included in this system for testing purpose and is shown
bridging the pressure regulator- Three ball valves in this system provide
independent selection of either regulated or bypassed flow control. The
sampling port at the base of the regulator was used to obtain some of the
oil samples.
76
-------
EMULSION
INLET
FLOW. SPLITTER
OIL PASSAGE
WATER PASSAGE
OIL OUTLET
ORIGINAL CONFIGURATION
FLOW SPLITTER
OIL PASSAGE
WATER PASSAGE
.OIL OUTLET
MODIFIED CONFIGURATION
S-60940 -A
Figure 51. Cross Section of Original and Modified Design
77
-------
Figure 52. Inlet Configuration of Modified Centrifuge
78
-------
f »
OUTLET END CAP
Figure 53. Corrosion Protection for End Cap Surfaces
79
-------
00
c
Figure 54. Outlet Flow Splitter and Oi
Discharge Tube Assembly
Figure 55. Outlet Flow Splitter and
Drum Assembly
-------
Figure 56. Oil Discharge Pressure Regulator
81
-------
PERFORMANCE TEST OF MODIFIED CENTRIFUGE
A new test program with the following objectives was performed with the
reworked centrifuge during May and June of 1970.
(a) Define the separator performance characteristics
(b) Check the mechanical and functional centrifuge characteristics
during operation with warm emulsions (I20-I60°F)
(c) Investigate the effect of solid particles on separation ability
and determine their distribution within the centrifugation
system
(d) Determine the feasibility of increasing the separation efficiency
by injecting additives upstream of the centrifuge
(e) Determine whether a multiple-stage centrifuge would be more
efficient
The tests were conducted at the Huntington Beach field laboratory and
the test setup was similar to that used previously (see Figure 41). A
particle counter was used to measure particle size and distribution at points
upstream and downstream of the centrifuge. (See Appendix 3.)
Test Results
The improved separation with the modified unit is graphically presented
in Figure 57. This curve shows that the separation capability was improved
by over 50 percent at the higher inlet emulsion concentrations. It was
necessary to sacrifice a portion of the flow capacity in order to install
the modified parts. The reduced flow capacity of the modified unit is caused
by the increased pressure drop show in Figure 58. The increased pressure
drop was caused by the unbalanced inlet and outlet pump-turbine configuration.
As originally designed, the pump-turbine relationship was hydrodynamically
balanced and functioned as an effective power conservation mechanism. The
power consumption for the modified unit remained low, however, as shown in
Figure 59.
Figure 60 confirms increased separation with increase in centrifuge
speed, as would be expected from centrifugal theory. This curve is a plot
of the oil concentration in the discharge water (mg/l) vs centrifuge speed
for a constant inlet flow rate of 200 gpm. The inlet emulsion contained
10,000 mg/l oil of 21 to 25° API gravity. The results showed a vast
improvement over the original unit for which the separation ability decreased
with speed, as shown in Figure 60. This phenomenon was due primarily to the
emulsifying effect of the feed pump on the original unit, since that pump
was directly connected to the engine driving the centrifuge. The resultant
excess flow required a throttle valve between the feed pump and the centrifuge-
82
-------
100
90
80
£ 70
60
50
£
UJ
vt
£ 30
ui
20
10
0
0 0.5 1.0 1.5 2.0 2.5
INLET EMULSION OIL CONCENTRATION, PERCENT
PERCENT SEPARATION IMPROVEMENT = 100 (I - §)
D
A = mg/1 OIL IN DISCHARGE WATER FOR MODIFIED CONFIGURATION
(PN 585010-1-3)
B = mg/I OIL IN DISCHARGE WATER FOR ORIGINAL CONFIGURATION
(PN 585010-1-1)
S-618/8
Figure 57. Performance Improvement
83
-------
PSIG
oo
'. >
LU
Q
Q
UJ
_J
z
LU
CD
U_
t-l
cc
t-
LU
O
50
o
-^i
m
rvtrooj^uv
omooo
10
: I : i |1|
i ,. 1. ,
:: :::=rit
. • .1
tEEiiHi
sr
. : j L+...
.. i !.j : i ,
t.L.
i. IE
Hrl; • • ! •'
•
: ::
,
...... i ,(..»!
:i
t:j
' i r '
£±
:;:::':
: ^-!f
. ..L.
.
tg
!
.:....
i 11 11
1 . ,
il.,Iill
.
.,;..,.,
-* -|—*-r- -j—
'
M
:
-j-1
±rr:-
uf,
;
-
£[!
1 I.I
t
•W
!
V
, ,
I
•
,
1.5 2 2.5 3
CENTRIFUGE SPEED, RPM X 10"
S-61876
Figure 58. Oil/Water Separator Inlet Pressures
-------
100
Q.
CQ
Q.
to
o
o
UJ
CD
U.
h-l
o:
UJ
O
80
60
40
20 •
TOTAL SYSTEM
LESS FEED PUMP
2000 2500
CENTRIFUGE RPM
3000
3500
Figure 59. Preliminary Centrifuge Power
Requirements (Steady-State)
85
-------
500
400
o
to
S 300
200
o
o
u
100
MODIFIED DESIGN
21-25 API OIL
10,000 mg/1 EMULSION
200 GPM FLOW RATE
ORIGINAL DESIGN
1500
2000 2500 3000
CENTRIFUGE SPEED, RPM
3500
S-61877 -A
Figure 60. Variable Speed Performance
86
-------
The performance plot of the modified centrifuge is shown in Figure 61.
This figure shows the oil concentration in the discharge water as a function
°f inlet emulsion concentration for inlet flow rates of 100, 200, and 300 gpm
at a centrifuge speed of 2750 rpm. The oil concentration in the discharge
should be and was reduced at higher speed, as shown in Figure 60.
The effect of improved separation with emulsions containing oil of higher
gravity ,(lov/e,r specific gravity), is shown in Figure 62. The unit was
operated at|2750 rpm with constant inlet flow rates of 200 gpm. Two oils
°f different API gravity were used to derive these curves- This plot con-
firms that the unit will separate lower density oils more easily than higher
Density oils, as would be expected from theory.
Two mefhods of increasing the separation efficiency were demonstrated.
The first m|thod was to; inject an emulsion breaker (coagulant) into the feed
eni
-------
500
400
300
I1 250
200
150
to
i—i
o
100
75
iu 50
40
30
20
INLET
FLOW RATES
MODIFIED DESIGN
2I-25°API OIL
2750 CENTRIFUGE RPM
3 456 810 15 20
INLET EMULSION CONCENTRATION, gm/1
30
40 50
S-6I875-A
Figure 6,1. Detailed Performance Plot at Reduced Centrifuge Speed
88
-------
300
200 GPM INLET FLOW RATE
2750 CENTRIFUGE RPM
457 10 20 30 40 50
INLET EMULSION CONCENTRATION, gm/1
S-61874 -A
Figure 62. Separation as a Function of Oil API Gravity
89
-------
O
00
I—I
Q
CD
z E
o
»-( «\
< UJ
LU
O
O
O
300
200
100
50
MODIFIED UNIT
2I-25°API OIL
200 GPM EMULSION FLOW
2750 RPM CENTRIFUGE SPEED
I I I i
6 78 9 10
INLET EMULSION CONCENTRATION, gm/
20
30
40 50
Figure 63. Emulsion Breaker Evaluation, Tretolite JN9045
INLET EMULSION
TEMPERATURE
5 6 7 8 9 10
INLET EMULSION CONCENTRATION, gm/1
S-61873'"
Figure 64. Influence on Separation by Increased Temperatures
90
-------
INLET SOLIDS = 9.5 mg/1 CONSTANT
FLOW RATE = 100 GPM SOURCE WATER
OIL FREE SOLIDS
ASHED SOLIDS
MODIFIED UNIT
1500 2000 2500 3000
CENTRIFUGE SPEED, RPM
3500
4000
S-61872 -A
Figure 65. Solids Distribution within CentrifugatIon System
-------
rv>
1000
500
S 200
i
LiJ
S 100
o
CO
LU
LJ
O
(J
50
20
10
2I-3I°API DIL
2750-3350 RPM
500 1000 5000 10,000
INLET OIL CONCENTRATION, mg/1
100,000
S-61883
Figure 66. General Map Representing All of the Modified Unit Performance Data
-------
This plot also provides the necessary information required to deduce
the performance of two centrifuge units in series or a two-stage centrifuge.
°V selecting a starting emulsion of 10,000-ppm oil and using the 500-gpm data
curve, the outlet from the first unit would contai'n approximately 390 mg/1
°'I. When this concentration is used for the inlet of the second unit, its
Oljtlet concentration would be 58 mg/1 oil. This indicates that series units
°r a multiple-stage unit would be more efficient than a single unit. Data
contained within Appendix 3 also confirms this analysis of Figure 66.
In order to determine the effect of multistaging the centrifuge, or
Derating two centrifuges in series, the effluent from a series of test runs
^as collected in a vacuum truck. The data from these runs is shown as the
|jrst nine runs of data page 5 of 5, Data Group 6, of Appendix 3. This
''quid was then withdrawn from the vacuum truck and rerun through the cen-
trifuge, with the results shown in runs 10 through 13 on the same data sheet.
93
-------
SECTION VIII
SYSTEM TESTING
Skimmer barge testing at sea was conducted off the coast near Santa Barbara
California by General Marine Transport of Santa Barbara, Inc. under the direct
supervision of AiResearch. (The General Marine Transport Company was intimately
'nvolved in the cleanup of the notorious oil leak at Santa Barbara in February
0[ 1969.) These tests were part of the Sea Dragon test program conducted by
^Research under contract for the American Petroleum Institute. The EPA per-
^'tted the skimmer barge and centrifuge to be part of the Sea Dragon test
ecluipment, which is a total containment and oil spill recovery system. In
Edition to the skimmer barge and centrifuge, Sea Dragon test equipment in-
cluded oil containment/sweeping booms used to funnel the surface oil into the
Dimmer barge and an inflatable oil storage receiver for the recovered oil.
^'Research Report 70-6787 describes the Sea Dragon tests and was concerned
^ainly with the evaluation of the booms. The skimmer barge was tested for the
c°l lection of oil with and without the use of the boom.
The objectives for the skimmer tests were:
(a) To check seaworthiness under various sea conditions
(b) To determine towing forces
(c) To determine the skimmer efficiency
(d) To evaluate the paddle wheel for skimming improvements
DESCRIPTION OF EQUIPMENT USED FOR SEA TESTS
^Hjts Under Test
(a) Skimmer barge (see Section VI for description)
(b) Centrifuge installed on the skimmer barge (see Section VII for
description)
s Employed
The following General Marine Vessels were used for towing, boom deploy-
016 nt, and transferring personnel.
(a) Sea Truck — 65- by 23-ft supply boat powered by two 400-hp diesel
engines. This was the the principal boat used in the program for
moving the skimmer to the test site and for towing the skimmer,
with or without booms.
95
-------
(b) Packer—65- by 21-ft supply boat, powered by two 300-hp diesel
engines. This vessel was used for boom deployment and boom towing
when two boats were required.
Other smaller boats were used for boom deployment, personnel transfer,
lighter towing jobs, and other tasks.
It should be noted that only approximately 100 hp is needed for towing
while skimming oil, but 600- to 800-hp vessels are required to move the skimmer
rapidly from one site to another.
Booms
The description of the booms follows:
(a) Rapidly Deployable Boom—The rapidly deployable boom, Figure 67, was
designed by Headrick Industries of Glendale, California. It con-
sisted of four 10-in.-dIameter tubes each 250 ft long, arranged in
the manner shown in Figure 68. The three top tubes were air-filled
and the lower tube was water-filled, with bulkheads every 25 ft. A
major feature of the Headrick boom was the cable harness, Figure 69,
which kept the boom in a true V-shape while being towed.
(b) Heavy Duty Boom—The heavy duty boom, Figure 70, was designed and
manufactured by Gates Rubber Company of Denver, Colorado. It con-
sisted of a series of hose sections measuring 25 in. OD and 25 ft
in length. The hose was bias-cut nylon tire cord carcass spirally
wrapped onto a helix of I/A-in.-diameter wire and bonded with neo-
prene rubber. A neoprene-coated nylon skirt was vulcanized to the
bottom of the boom hose. The hose sections were terminated with
aluminum heads, and sections could be connected by Marman (V-band)
clamps, as shown in Figure 71.
(c) Water Spray Boom—The water spray boom, Figure 72, was designed and
fabricated by Pacific Northwest Laboratories Division of Battelle
Memorial Institute, Richland, Washington. It consisted of water
spray nozzles mounted on five sponsons, as shown in Figure 73. The
spray nozzles were mounted in a 6-in.-diameter pipe and sprayed water
ahead of the sponsons at about a 20-deg angle from the horizontal.
SEA TESTS
Following is a description of the sea tests.
Test Location
The tests were conducted in three separate areas off the coast near Santa
Barbara, as shown in Figure 74. Area A, which is immediately adjacent to the
harbor, was used for general shake-down tests and closed-circuit skimmer oil
tests. Since any oil spillage in this area would be noticed immediately,
testing that would involve placing oil on the open ocean was not permitted.
96
-------
Figure 67. Sweeping Oil with Single Headrick Boom
97
-------
3-1/2 IN,
(TYP)
XD
00
S-60528
•TOWING EYE
Figure -68. HeadrvcV. Irvf\atabAe Boom
-------
125 FT
FIVE ATTACHMENT
POINTS TO SKIMMER
WIRE HARNESS
5/16-IN. DIA-
TOW LINE
ATTACHMENT.
S-60597
Figure 69. Headrick Boom Harness [Original Configuration)
-------
Figure 70. Skimmer with Single 500-ft Section of Gates Boom
00
-------
BOOM SIZE:
BOOM WEIGHT:
BOOM MATERIAL:
LENGTH, 25 FT; DIAMETER, 25 IN.
APPROX. 700 LB
NEOPRENE COVERED NYLON (REINFORCED
WITH WIRE AND FIBERGLAS)
ALUMINUM FLANGES AND END PLATES
INTERNAL PRESSURE: AMBIENT
PLEATS
CRES BOLTS
MARMAN
CLAMP
TOWING
PLATE
MARMAN CLAMP
ALIGNMENT RO
-*>>,s>v-»»yYv..
SX5>^ Z2223SZZ222
'"""""V^KK««
-------
Figure 72. Battelle-Northwest Water Spray Boom
02
-------
TOWBOAT
.BOOM DIMENSIONS
TOTAL LENGTH
PIPE LENGTH PER SECTION
OUTSIDE DIMENSION OF SPONSON
ASSEMBLED HEIGHT
SPRAY NOZZLE HEIGHT ABOVE
WATER SURFACE
TOTAL WEIGHT
WATER SUPPLY
75 FT
15 FT
10 BY 8 BY 6 FT
6 FT
3 TO 4 FT
7000 LB
3500 GPM AND
100 PSIG
S-60916
Figure 73. Battelle Spray Boom during Skimming
103
-------
w
-ELLWOOD TANKER
TERMINAL
SANTA BARBARA
COAL OIL POINT
NATURAL OIL SEEP
PLATFORM
HOLLEY
SANTA BARBARA CHANNEL
SCALE IN MILES
01 2 3 A 5 6
• OFFSHORE OIL DRILLING PLATFORM
TEST AREA C
50 MILES OFFSHORE
33° 40' N
I 19° 30' W
S-60503
Figure 14. Test. Areas
-------
Area B is the site of a natural oil seep approximately 10 miles west of Santa
Barbara. In this area a light weathered slick of several square miles is con-
tinuously present. The majority of the boom testing was done at this site,
and some 26 and 35° API oil was occasionally dumped to test its effect.
The third location, Area C, was located more than 50 miles offshore.
This area was used for the final system tests wherein oil was spilled on the
ocean surface and then swept by the booms and recovered by the skimmer.
Testing was done with the technical assistance of representatives from
the various subcontractors whose equipment was involved (Gates, Headrick,
USCG). General Marine Transport supplied the boats, crews, and miscellaneous
services required to support the operation.
lest Results
A summary of all of the testing at sea is presented in chronological
order in Table 3. The skimming system, and usually the centrifuge, were in
operation during many of the boom tests, so additional data under various
operating conditions was obtained.
(a) Towing Tests
Sea testing of the skimmer commenced in June 1970. Early tests included
towing tests at various speeds (forward and backwards) and at various depths
(drafts). Figure 75 depicts the results of these tests against a background
°f the previously calculated predicted forces. The forces were measured with
a dynamometer installed on the towing vessel in the manner shown in Figure 76.
uf particular interest is the extremely small force involved in towing the
skimmer at a reasonable boom-towing speed of one knot. This force is so low
't would have no effect upon the tested booms, and the skimmer can be readily
towed, at sweeping speed, through the booms. A separate towing line is not
•"equired for the skimmer.
(b) Skimmer Operation Tests
In order to check out the skimmer components a second series of tests,
conducted in mid-June, included closed-circuit operation of the oil/water
circui,ts. In these tests 15 or 26° API oil was introduced in the center of
the spillway and circulated through the quiet pond, weirs, pumps, surge tank,
snd oil transfer pump back to the top of the spillway. It was observed that
a'l components functioned.
(c) Centrifuge Tests
On June 16 the oil/water centrifuge was installed on the skimmer.
Since the centrifuge performance was mapped prior to installation on the
skimmer, only sufficient data was taken to compare general adherence to
Previous data. For example, two test points taken during the at-sea test on
18 using 26° API oil are shown on the performance map of Figure 66. A
105
-------
TABLE 3
CHRONOLOGY OF TESTS
Date
(1970)
LocatIon*
Test
Test Objectives and Results
Jun 3
Jun 9
Jun I I
Jun 18
Jun 19
Jun 24
Jul 23
Jul 27
Jul 29
Sept 23
Sept 24
Area A
Area A
Area A
Area A
Area B
Area B
(Coal Oil
Point)
Area B
Area C
(Off Santa
Cruz
Island)
Area C
Area B
Area A
• Skimmer tow tests
• Skimmer closed-circuit oil
tests
• 26° API oil
• Skimmer closed-circuit oil
tests
• 16° API oil
• Skimmer closed-clrcuIt oil
test with centrifuge using
26° API oil
• Skimmer plus centrifuge and
Gates boom with V harness
• Skimmer and Gates boom with
V harness
• Single Gates boom and
skinnier; no harness
• Skimmer efficiency tests
• Skimmer and single and double
Gates boom efficiency tests
• Skimmer and Head rick boom
tests
• Single 250-ft section Head rick
boom and skimmer with paddle
wheel
• Single 250-ft section Head rick
boom and skimmer with paddle
wheel
• Skimmer towing forces.
• Skimmer ballasting times.
• Weir operation in Sea States 0 through 3.
• Quiet pond holds oil.
• Floating weirs skim oil
• Pumps pump weirs.
• Surge tank separates oil and water to less than 25 ppm
oil at gross flow rate of SO gpm.
• System pumps 500 gpm (but surge tank does not separate;
centrifuge required).
• Lower API gravity oil much easier to process (will not
mix with water as readily).
• Sea States 0-2.
* Centrifuge operated properly under bc.ge pitch and roll
env i ronment.
• Centrifuge separated oil from surge tank water
discharge.
• Sea States 0-3.
• Towing speed to test site was 5 to 6 knots.
• Used natural oil seepage at Coal Oil Point.
• Sea States 0-1.
• Sea States 0-1.
• Two barrels of tar collected.
• Weathered oil coagulates into large lumps of tar.
Will not flow through weirs.
• Sea State 0.
• Skimmer overall efficiency obtained with 35° API oil
in Sea States 0 to 2.
• Headrick boom, with harness, swept oil at
speed of 2.5 knots in Sea States I through 3.
Made 75- to 100-ft wide sweeps at 2 knots.
Collected 18 barrels of tar from weathered natural seep-
Weirs and centrifuge ineffective with tar lumps.
Sea State 0.
Hade 75- to 100-ft wide sweeps at 2 knots.
Collected 5 barrels of weathered oil (tar) from
accumulated leakages in this area.
• Weirs and centrifuge ineffective in tar lumps formed
from weathered oil.
• Sea State 0.
*See Figure 74.
106
-------
100
3 4
SPEED, KNOTS
Figure 75. Barge Towing Forces
107
-------
M = 0 TO 10,000 LB DYNAMOMETER
D = DYNAMOMETER READING
T = TOWING LOAD
SET FOR TOWING LOADS UNDER 10,000 LB
T = D
SETUP FOR TOWING LOADS GREATER THAN 10,000 LB
X = Y; | - TAN 9
D
T =
2 SIN 6
S-60502
Figure 76. Dynamometer Setup for Measuring Towing Forces
108
-------
typical result of these tests is shown in Figure 77. The jar on the right
contains the centrifuge inlet mixture (approximately 2 percent oil), the
center jar contains the oil discharge, and the left-hand jar water discharged
by the centrifuge. During this time period, skimmer barge equipment was
further developed to improve the handling of the oi1/water/trash mixtures.
After these tests were completed the skimmer was assigned to the API
boom testing program.
(d) Skimmer Tests with Boom
The skimmer was tested with the Gates, Headrick, and Battelle booms already
described. Use of the Gates and Headricks booms enabled a wide expanse (greater
than 100 ft) of ocean surface to be swept into the skimmer. It was observed
that oil spread on the ocean surface would weather in just a few hours to a
tar-like product. Although the oil on the ocean surface appeared to be very
light and thin (with a silvery sheen, color bands, and some "elephant skin"
Windrows), it was actually made up of the weathered heavy ends. As it col-
lected in the quiet pond it coagulated into a heavy black sponge-like mass
several inches thick. Some of this was skimmed off by the floating weirs
and pumped to the surge tank. However, the weirs quickly became clogged with
the thick masses of tar and the material had to be manually broken up and
lifted from the quiet pond like pieces of asphalt paving.
Oil that was pumped to the surge tank rapidly coagulated on the surface
afid would not drain off through the 3-in.-diameter oil drain pipe. This
Material had to be shoveled into open oil drums. A quart jar sample of the
°il could be inverted without the oil running out. A total of 18 barrels of
this material was collected during a 4-hr period on September 23, 1970. Two
Samples taken during this test period had these characteristics:
Sample No. 4 Sample No. 9**
Sample location Quiet pond Collecting drum
Date Sept. 23, 1970 Sept. 25, 1970
Time 10:00 a.m. 2:00 p.m.
Percent water 38 percent 26 percent
(including II (including 4
percent sand) percent sand)
Specific gravity 1.0557 1.0507
API gravity 2.5°** A.5°**
Appearance Wet mass of tar Damp mass of tar
^Material collected at Coal Oil Point Sept. 23, 1970 and placed in drums.
0il of these gravities, is denser than sea water and should sink. Samples
taken at Coal Oil Point and stored in l-pt Mason jars usually sink within a
day or two. Apparently entrapped gas keeps this material afloat.
109
-------
DISCHARGED
WATER
DISCHARGED
OIL
CENTRIFUGE
INLET
Figure 77. Centrifuge Inlet and Discharge Samples
\ 10
-------
(e) Skimmer Efficiency Tests
During the sea tests conducted July 27 to 30, 1970, the overall collection
efficiency of the skimmer was determined for a particular set of conditions.
The skimmer was deployed off the stern of the supply boat Sea Tender in the
manner shown in Figure 78, and oil was dumped in front of the skimmer by means
of a 3/4-in. hose attached to the towline. The oil circuit is shown in
Figure 79.
Overall skimmer efficiency is dependent upon all of these factors:
(a) Skimmer speed
(b) Sea state
(c) Direction with respect to wind and waves
(d) Depth of spillway
(e) Condition of inlet as affected by booms, etc., in front of the
s k i mme r
(f) Oil fi1m thickness
(g) 0 i1 gravi ty
During the test period between.I 3:30 and 15:30 the skimmer was progress-
ing at 1.2 knots starting under Sea State 0 as shown in Figure 80. As the test
Proceeded, the sea state condition gradually increased until at the termination
°f the test at 16:30, the sea state was between I and 2. Results of the
efficiency testing are shown in Figure 81.
The rate of dump was determined by recording the total quantity (barrels)
Passed through a flowmeter on the dump hoss and applying a meter correction as
Determined by a calibration of the flowmeter at the Signal Oil Company Labora-
tory in Long Beach, California.
Meter calibrat ion:
10 bbl actual = 10.45 bbl indicated
Meter correction:
ng = actual bbl
These points are plotted in Figure 81.
I I I
-------
UN I ROYAL TANK
SKIMMER
6-IN.-DIAMETER
OIL TRANSFER HOSE
OIL SPREADING BOARD
OIL DUMP HOSE
DOWN TOW LINE
TO SKIMMER
SEA TENDER
WITH TWO 500 BBL TANKS
S-60330
Figure 78. Skimmer Efficiency Test Setup
-------
BAKER TANK
OIL PUMP /
- FLOW METER NO. 1 SURGE TANK
ucIP »,„,,_ CENTRIFUC
—" -.•*». ^^.— . — —
8rr-i
0=3 .1 for-
V. J\
V
SEA TENDER
h __fL
^ f — -*-^^**~ — - -», i — i <-._/^..., ^ ..^-. -»-—
^v
V
OCEAN
FLOW METE RM NO.. _2 mn ., _ FFFirrFwrv
/ U 1 L. r\L. V/ L. 1 V L. I\
/ r-FLOW METER
/ I NO. 2
iE / PUMP f
-m — — V
1 1
WATER OIL
OVERBOARD STORAGE
J
V
SKIMMER
FLOW METER NO. I
K = CORRECTION FACTOR FOR WATER CONTENT AND LIGHT END LOSS,
S-60335
Figure 79. Setup for Determination of System Efficiency
-------
Figure 80. Skimmer during Efficiency Test
I 14
-------
DUMP
2.53 BBL/HR
RECOVER
.90 BBL/HR
EFF.)
RECOVER
2.28 BBL/HR
(90% EFF.)
ELAPSED TIME, HR
1300
1500
CLOCK TIME
Figure 81. Skimmer Efficiency Test
I 15
-------
(f) 0?1 Recovery Rate
Recovered oil was collected in the transfer barrel, and at suitable inter-
vals when the barrel was nearly full, it was pumped through the skimmer flow-
meter into the 6-in. hose to the Uniroyal tank. The time at which the pumping
operation was completed was recorded. The flowmeter readings were corrected by
the meter calibration factor:
Meter calibrat ion:
10 bbl actual = 10.8 bbl indicated
Meter correction:
Meter read ing . , ,, ,
1-7—5 a = actual bbl
I u. o
During the test the water level in the surge tank was maintained at a level
well below the inlet pipe and a little above the lower outlet pipe to the
centrifuge. This kept the oil mixed into the water so it would not accumulate
on the surface. At approximately 1600 hr, the crew member assigned to main-
taining this level had to leave his station to assist in ballasting the barge
to a somewhat deeper waterline in an attempt to reduce underflow at the spill-
way. During this time the water level in the surge tank rose above the inlet
pipe and, as a result, oil accumulated on the surface of the water in the
surge tank. The test was terminated at 1620 hr when the barge towline parted
from chafing at the towboat. At this time there was 1-1/2 in. of oil on the
surface of the surge tank. This quantity, which is 5/8 (0.63) bbl, was there-
fore added to the last reading at 1620 hr.
Samples of oil were taken from the centrifuge discharge intermittently
during the test. These were analyzed and were found to have an average oil
content of 78.5 percent. The recovered oil/water emulsion data was then cor-
rected for water content by multiplying by 0.785. This produced a figure for
the dry oil collected.
On July 29, while operating with the same type of oil under similar cir-
cumstances except that a boom was being used, oil samples were taken of the
original and recovered oil. These were analyzed for percentage of major con-
stituents with these results:
Ori gi nal Oil Recovered 0 i1
Sulfur, weight percent 0.15 0.25
Gasoline, percent 33.9 15.9
Kerosine, percent 22.7 26.5
Gas-oil, percent 22.0 28.8
Still residue, percent 21.7 27.6
Loss, percent 1.3 1.2
(See Tables 4 and 5 for complete data)
I 16
-------
TABLE 4
GENERAL CRUDE OIL TESTS
ON OIL SAMPLES SUBMITTED BY AIRESEARCH MANUFACTURING COMPANY
Original Oil Recovered Oil 1842
1842 7/29/70 V
API gravity, 60 F 33.5* 28.5
Specific gravity, 60/60 F 0.8576 0.8844
Pour point, F +35 4-55
Sulfur, weight percent 0.15 0.25
Viscosity, SUS
70 F 55.0 105.9
100 F 44.I 60.3
Initial boiling point 126 213
The corrected API gravity of the tank sample of the oil delivered on July 24,
'970 was 34.9° API and has been regarded as 35° API throughout this report.
This particular sample, taken after five days at sea in a vented 250-barrel
tank, indicated 33.5° API.
Assuming that no evaporation or solution of the gas-oil constituent
°ccurred, the loss of the other constituents and the quantity of original oil
Pe'~ unit volume of the recovered sample can be determined. -With this assump-
tion, 23.7 percent of the original oil spilled was lost through evaporation
°r solution. Assuming that 5 percent of the gas-oil constituent was also
lost, the total loss would increase to 27.5 percent. This analysis is shown
below:
(I) No Gas-Oil Loss
The original oil sample was 22.0 percent gas-oil and the recovered sample
was 28.8 percent gas-oil. If it were assumed that no gas-oil was lost/ the
0|"iginal volume of crude represented by the recovered sample would have been
'31 ml/100 ml recovered:
- 1.31
I 17
-------
TABLE 5
MODIFIED HEMPEL DISTILLATION OF OILS SUBMITTED
BY AIRESEARCH MANUFACTURING COMPANY
Atmospheric Distillation
Temp . , F
IBP- 122
122-167
167-212
212-257
257-302
302-347
347-392
392-437
437-482
482-527
Vacuum Distillation,
40 mm Hg . , F
Up to 392
392-437
437-482
482-527
527-572
Distillation Summary
Gasol ine
Kerosine
Gas Oil
Still Residue
Error or Loss
Original
Percent
Distilled
-
1.0
3.4
7.8
7.9
7.6
6.2
6. 1
8. 1
8.5
2.5
5. 1
5.3
4.3
4.8
-
-
-
-
-
Total
Percent
Disti 1 led
-
1.0
4.4
12.2
20. 1
27.7
33.9
40.0
48. 1
56.6
59. 1
64.2
69.5
73.8
78.6
33.9
22.7
22.0
21.7
-0.3
API
Gravity
-
76.4
62.2
55.8
51.9
48.5
45. 1
41.8
34.8
35.6
32.7
31.5.
29.3
26.9
23.5
52.3
38.2
28.6
1 1.7
-
Recovered
Percent
Distil led
_
-
-
1.8
3.0
4.8
6.3
8.8
10. 1
7.6
3. 1
7.4
6.5
5.6
6.2
-
-
-
-
-
Total
Percent
Distil led
-
-
-
1.8
4.8
9.6
15.9
24.7
34.8
42.4
45.5
52.9
59.4
65.0
71.2
15.9
26.5
28.8
27.6
1.2
API
Gravity
-
-
-
44.6
46.7
45.7
45.3
41.6
37.8
35.3
32.8
31.6
29. 1
27.1
24.8
45.7
38.7
28.8
9.4
_
18
-------
A 131-ml sample of original crude would contain the quantities of constituents
as shown in line 2 of Table 6. Actuany the amounts recovered were as shown
in line 3. Line 4 shows the percentage of each of the original constituents
recovered^ and line 5 the percentage loss.
(2) Five Percent Gas-Oil Loss
If it were assumed that some of the gas-oil was also lost, these loss
figures would be even higher. As an example, lines 6 through 8 of Table 6
illustrate the case in which a loss of 5 percent of the gas-oil is assumed
and the original gas-oil amount would be 30.3 ml for each 28.8 ml recovered.
Since the oil placed on the water for the efficiency tests would be sub-
ject to the same loss phenomena, although for varying exposure times, it was
assumed that an average evaporation and solution loss of 20 percent had
occurred. Therefore, the dry oil figures were corrected by dividing these
Values by 0.80 to obtain a value for the equivalent original oil collected.
This is plotted as the "recover" curve of Figure 81. A summary of the test
data and the calculations from which these curves were plotted is included
as Table 7.
From observations of these tests in comparison to those at Coal Oil Point
't appeared that the 35° API oil mixes very easily with water forming small
droplets that were readily carried below the surface. The heavier Coal Oil
Point oil, or weathered oil, tended to form larger globules, which tended to
stay on the surface of the water and not be carried under the skimmer.
(g) Paddle Wheel Tests
On September 16, 1970 a paddle wheel was installed as shown in Figures
35 and 36. It was driven by a V-belt gear-reducer chain-drive assembly by
means of a small diesel engine removed from one of the weir pumps. The over-
all drive ratio was 119:1 in three steps as follows.
V-belt 5/8 = 0.625:1
Worm gear reducer = 30:I
Chain 70/11 - 6.36:1
Overall = 0.625 x 30 x 6.36 = 119:1
The diesel engine could be run at speeds between 900 and 3600 rpm, but
a'l of the testing was done at a constant engine speed of 1460 rpm. This pro-
duced a peripheral speed on the 36-in.-diameter paddle wheel of 1.2 knots.
This speed appeared to be most compatible with the 2.0-knot skimmer velocity
used during the tests.
I 19
-------
TABLE 6
SUMMARY OF EVAPORATION LOSS CALCULATION
Line
Gasol ine
Keros ine
Gas-Oi 1
Res idual
Error
Total
100 percent of gas-oil recovered
1
2
3
4
5
New oil sample percent
ml
Recovered oil sample
percent
Percent recovered
Percent lost
33.9
44.5
15.9
35.7
64.3
95 percent of gas-oil recovered
6
7
8
Size of original sample
per 100 ml recovered
Percent recovered
Percent Lost
46.6
34. 1
65.9
22.7
29.8
26.5
89.0
1 1.0
22.0
28.8
28.8
100.0
0.0
21.7
28.4
27.6
97.3
2.7
-0.3
-0.4
1.2
-
-
100.0
131
100.0
76.3
23.7
31.2
85.0
15.0
30.3
95.0
5.0
29.8
92.7
7.3
-0.4
-
-
137.5
72.7
27.3
-------
TABLE 7
SUMMARY OF RESULTS OF SKIMMER EFFICIENCY TEST JULY 21, 1970
Test Time
(Hr)
1340
1345
1353
1404
U06
1407
U09
1417
1436
1437
1448
1450
1451
1500
1502
1504
1519
1535
1539
1541
1542
1545
1556
1558
1559
1606
1613
1616
1620
1630
Sea Tender
Flowmeter
(bbls)
22.3
22.4
22.7
23.2
24.6
25.5
26.5
27.2
27.6
28.6
28.9
29.2
Oil Dumped
(Unconnected)
(bbls)
-
O.I
0.4
0.9
2.3
3.2
4.2
4.9
5.3
6.3
6.6
6.9
0 i 1 Dumped
(Meter Corrected)
(bbls)
_
0.10
0.38
0.86
2.20
3.06
4.02
4.68
5.06
6.02
6.30
6.58
Oil
Transfer
Pump
Start
Stop
Start
Stop
Start
Stop
Start
Stop
Start
Stop
Start
Stop
St^rt
Stop
Skinnier
Flowmeter
(bbls)
3.50
3.90
3.90
4.95
5.64
5.64
6.22
7.68
7.68
8.26
8.26
8.43
Flowmeter
bbls
Col lected
-0-
0.40
1.45
2. 14
2.72
4.18
4.76
4.93
Flowmeter
Collected
(bbls)
0.37
1.34
1.98
2.52
3.87
4.41
4.56
Plus 5/8 bbl
in Surge Tank
At end of Test
-
-
-
-
-
-
-
5.19
Corrected
For Water
Content
-0-
0.29
1.05
1.55
1.98
3.04
3.46
4.07
Corrected
for
Evaporat ion
-0-
0.36
1.31
1.94
2.47
3.80
4.33
5.08
-------
It was observed that the paddle wheel was able to maintain a flow into
the skimmer inlet even when the skimmer was standing still in the water.
After the checkout runs were made alongside the pier in Santa Barbara, it was
noted that another boat in the harbor had discharged several gallons of black
oil while pumping tanks. The skimmer, which lay immediately downwind of the
offending vessel, was submerged to operational depth and the paddle wheel
started. Garden hoses and spray nozzles fed by the skimmer's high pressure
(60 psig) water supply were employed to direct the spill to the front of the
skimmer.' The paddle wheel maintained a flow into the quiet pond and the
entire spill was collected without moving the skimmer from its mooring.
During the at-sea tests on September 23, and 24, the paddle wheel main-
tained a constant flow across the spillway lip, appreciably increasing the
flow of oil into the skimmer. Furthermore, the paddle wheel did not cause
the severe wave reflections as did the wave gate previously used. As the
skimmer became increasingly loaded with the collected weathered oil, the
paddle wheel became an effective check valve to prevent the oil from passing
back out through the entrance when the skimmer stopped.
122
-------
SECTION IX
DISCUSSION
The skimmer barge and centrifuge recovered and separated oil in Sea
States 0 through 3. The effectiveness of the skimming system was reduced as
the sea state increased and although no efficiency data was taken, the system
Was shown operable in Sea State 3. In rougher seas the floating weirs had
to be secured to prevent damage from sloshing, but under these conditions,
they become quite inefficient anyway. If the floating weirs were replaced
by a fixed weir, the skimming could continue, although at reduced efficiency,
until the barge began to take water over the gunwales in large amounts
(estimated Sea State 7).
The quality of the oil being recovered had a marked effect on the opera-
tion. Light oils of 26 to 35° API were readily passed through the system.
However, they appeared to readily emulsify. Heavier oils separated readily
'n the surge tank so that the centrifuge inlet emulsion appeared to contain a
'ow percentage of oil. The weathered oils, which had the consistency of tar,
Were very difficult to handle within the skimmer, but were so cohesive the
collection efficiency appeared to be perfect. This material required manual
recovery, as it formed into lumps too large to pass through the weirs- Addi-
tional equipment would have to be installed on the barge to handle this
"eterlal.
SKIMMER BARGE
£f_fect of Wave Action
The skimmer barge was operated in seas from flat calm to Sea State 3
(with 4-ft waves), and the waves had these effects on the skimming effective-
ness:
(a) Increasing wave heights would increase the surface disturbance at
the entrance to the skimmer, thereby causing the oil to be mixed
into the surface water prior to entrance into the skimmer. This
phenomenon was aggravated by the higher gravity, noncoheslve oils.
(b) Waves entering the skimmer were observed to cause disturbances in
the quiet pond, which would tend to prevent complete settling.
This was particularly noticeable at the forward end of the quiet
pond, especially when the wave gate was being used. These dis-
turbances were observed to extend all the way to the bottom of the
quiet pond and would cause some oil to escape through the louvered
bottom.
(c) Visually, the pitching, rolling, and heaving of the barge due to
wave action had minor effects- These motions caused sloshing in
the quiet pond, which was partially attentuated by the wave fences.
123
-------
(d) The paddle wheel was not tested in sea states high enough to deter-
mine its relative effectiveness in reducing wave disturbances at
the skimmer entrance. Results of testing with the paddle wheel in
lower sea states, however, indicate that it might reduce the verti-
cal currents generated immediately aft of the spillway more effectively
than does the wave gate. The paddle wheel, on the other hand, would
not offer the full protection available with a completely closed
wave gate under very high sea states.
Loss of Oil
The principal phenomenon that affected skimmer efficiency was loss of oil
at the front of the quiet pond- This was due to the extreme turbulence and
vertical circulation immediately aft of the spillway and in the region of the
first wave fence, as shown in Figure 82. Oil droplets would beat into the
water while passing the disturbed areas of the wave gate, the aft side of the
spillway, and the wave fences. Vertical circulation, caused by the aft side
of the spillway, carried these oil droplets to the bottom of the quiet pond
where they escaped through the forward louvers. This phenomenon was observed
through the underwater viewing ports in the sides of the quiet pond.
Oil was observed to be lost through the bottom of the quiet pond in
quantities sufficient to leave an easily discernible slick behind the barge
when processing oil in the 26 to 35° API gravity numbers. Observation through
the underwater viewing ports showed that this is a phenomenon restricted to
the forward end of the pond.
It is anticipated that this shortcoming can be eliminated by some minor
modification to the forward end of the quiet pond. Changes that could be con-
sidered, as shown in Figure 83, are:
(a) Closing the louvers for the first several feet of the quiet pond
(b) Installing horizontal baffles in the forward section of the quiet
pond to reduce vertical circulation
(c) Improving adjustment of the wave gate to reduce water disturbance
(d) Extending the spillway slope to better diffuse the incoming stream
and to reduce the effect of the vertical aft bulkhead of the
spllIway
Seaworthiness of Barge
No problems were encountered with respect to the stability or seaworthi-
ness of the skimmer barge during the tests at sea.
124
-------
ro
m
WAVE FENCES
LOUVERED
BOTTOM
UNDESIRABLE
CIRCULATION
VIEWING
PORT
S-60608 -A
Figure 82. Turbulence at Forward End of Quiet Pond
-------
a. ELIMINATION OF LOUVERS IN FORWARD SECTION OF QUIET POND
b. HORIZONTAL BAFFLES
c. EXTENSION OF SPILLWAY SLOPE
S-60616
Figure 83. Possible Solutions to Quiet Pond Turbulence
126
-------
CENTRIFUGE
.Analysis of Separable Oil Drop Size
A small oil drop entering the annulus of the oil/water centrifuge separa-
tor is envisioned to be the same as if it were entering a straight nonrotating
channel and under the influence of a high gravitational field equivalent to
the centrifugal force produced by rotation of the centrifuge. The following
assumptions are made in the analysis.
'' Assumpt ions
(a) The drops are spherical in shape.
(b) The distance between drops is large enough to assume no interaction
between drops (this allows the use of analysis on a single drop).
(c) The drops are not near walls or other boundaries.
(d) The forces acting on the drops are:
Buoyant force due to density difference between the drops and
the seawater
The external force which is equivalent to the centrifugal force
produced by rotation of the centrifuge
Viscous drag (represented by Stokes' drag law, which requires
a Reynolds number < I) (Figure 84)
(e) All other forces are assumed to be negligible.
2. Initial Conditions
The initial operating conditions considered for this analysis are the
low!ng:
(a) The drop enters the centrifuge annulus at time t = 0 and has zero
radial velocity.
(b) The drop enters the centrifuge annulus at a distance R from the
axis of rotation.
127
-------
t/J
o
on
o
on
LU
t
o
LU
_J
O
l~i
fe
100
80
60
50
40
30
20
10
5
4
PARTICLES ARE SPHERICAL
OIL DROPS (21° API) SUSPENDED
IN SEA WATER
DENSITY OF PARTICLE, LB/IN.
3
'P
U) =
r =
DENSITY OF FLUID, LB/IN.
DIAMETER OF PARTICLE, IN.
CENTRIFUGE ROTATION, RAD/SEC
DISTANCE OF PARTICLE FROM
CENTRIFUGE AXIS, IN.
= FLUID VISCOSITY, LB/IN-SEC
I I I I I I I I
= (pp-PF)
R =
PF
i i
Reynolds number
0.
0.2
0.3 0.4 0.5 0.7
1 2 3456
-V , PARTICLE TERMINAL VELOCITY (INCHES PER SECOND)
Figure 84. Oil Drop Terminal Velocity as a Function of
Drop Diameter and Centrifugal Force
S-6188*
128
-------
3. Analysi s
The equation of motion is:
? DP3 <>P + I "F> § ' 2 °P
r = radius vector from the centrifuge axis of rotation to the oil
drop (the origin moves along the centrifuge axis with the
velocity of the suspending fluid)
t = t i me
D = diameter
p = densi ty
u> = angular velocity
u, = vi scosi ty
P = particle (oil drop)
subscripts p = fluid (sea water)
^he term on the left-hand side of the equal sign is the inertia force. It is
composed of the ordinary inertia plus an additional apparent mass. This
additional apparent mass arises because we have applied a Galilean transforma-
t!on to an unsteady flow problem. The first term on the right-hand side is
the buoyant force caused by the fluid density difference in the centrifugal
force field. The last term on the right is the Stokes1 drag force.
The solution of the equation is:
« z at + ""
Vi
(i\ - e-at fcosh f/l + ((»/«) z at) + "" h ' •+ <>/«2t
w I A '
2Vl
Vr/ [l H- Vl + (P/a)2 coth VI + (0/a)2 at 1
L J
v =-v- = radial velocity
r dt
f \ r. 2 2
(pp-pF) Dp to r
V = terminal radial velocity = r=
r IS
= initial radial distance (r = R at ,t = o)
129
-------
n ( -\- — ~\
A'pP-pF)
p =//~^~r
y(pp+ 2 PF)
The solutions to the equation of motion for 21° API oil drops suspended
in seawater are presented graphically in Figures 84 through 86. The oil drop
residence time in the oil/water centrifuge as a function of oil/water mixture
flow rate is shown in Figure 87.
As an example, consider the case where the oil/water mixture flow rate
is 500 gpm and the rotational speed is 3600 rpm. With a centrifuge annul us
having a 24-in. OD and 20-in. ID and a length of 60 in., the residence time
in the centrifuge of an oil drop would be 4.2 sec (from Figure 87).
The centrifugal force at the mean radius of the annulus is 4000 g (Figure
84). Also, it can be seen that Stokes1 drag law is not valid for particles
larger than approximately 18p, (microns). For an 18u,- part ic le, the terminal
velocity is 2.5 in./sec.
Figure 85 shows that it takes an 18p,-part ic le approximately 10~* sec to
reach 99 percent of its terminal velocity. Smaller particles reach terminal
velocity even faster. Thus, for the oil drops being considered here (< 18pJ;.
terminal velocity is attained instantaneously, for all practical purposes.
Figure 86 indicates the particle size that can be separated. For the oi'/
water separator the parameter C(R-r)/R] is equal to 1/6 and the residence time
in the centrifuge is 4.2 sec. Ideally, at 3600 rpm all oil drops larger than
8u, can be separated. In actual practice, this drop size will be somewhat
larger because of the influence of the adjacent particles and the walls.
Droplet Size Measurement
I. The Counter
The theoretical predictions of separator efficiency can be utilized only
if the actual size of the oil droplets present in the emulsion is known. There
are several well-established methods for measuring drop size but the methods
that afford the highest accuracy and the highest degree of automation are
those based on photoelectric optical systems.
An HIAC-SS Automatic Particle Counter, made by High Accuracy Products,
Claremont, California, was used during land test. No droplet measurements
made at sea or on samples recovered at sea. The counter schematic diagram is
shown in Figure 88. A sample of the oil/water emulsion flows through a small
rectangular fluid passage and past a window. Oil particles in the fluid pass
130
-------
IOC
8C
6C
5C
PARTICLES ARE SPHERICAL.
OIL DROPS (21° API) IN
SEA WATER
3C
I
-------
in
o
o
i—i
5
•*
a:
LU
b
1
a
to
LU
100
70
60
50
40
30
20
10
8
6
5
4
PARTICLES ARE SPHERICAL OIL
DROPS IN SEA WATER
R = PARTICLE DISTANCE FROM AXIS
OF ROTATION OF CENTRIFUGE
AT TIME t = o, in.
r » PARTICLE DISTANCE FROM AXIS
OF ROTATION OF CENTRIFUGE
AT TIME t, in.
t, TIME (SECONDS' TO TRAVEL THE DISTANCE (R-
STARTING FROM THE DISTANCE R
S-6I8B'
Figure 86. Travel Time as a Function of Drop
Diameter and Rotational Speed
132
-------
£
o
Ul
I
s
o>
700
500
400
300
200
100
80
70
60
50
40
30
20
10
D = Out
o
D. = Inn
L =» Len
\
X
\
\
\
\
V
\
\J
s
s
er diameter of centrifuge
er diameter of centrifuge
gth of centrifuge annuius
s
\
T =
\y
>y
\
annuius = 2.00 ft
annuius =• 1.6'
» 5.00 ft
1 ft
= 351
*\
S
1
\.
\
Q
\
2)
\
L
X
\
S
2 345 7 10 20 30 40 50 70 100
T, PARTICLE RESIDENCE TIME IN CENTRIFUGE (SECONDS)
S-61886
Figure 87. Oil/Water Centrifuge Oil Drop
Residence Time as a Function of Flow
133
-------
WINDOW
-SAMPLE FLUID
-FLUID PASSAGE
PANEL BASE
OUTPUT METER
LIGHT INTENSITY
ADJUST
FIANGE
IMENTS -v
X i
— — ^ —
F
O O O O O
o o o o o
i
i
o
1 \
' V \
1
20 | | 40 | | 80 | 150
/
O OPERATE
CALIBRATION
PULSE GENERATOR
^READOUT
CHANNELS
S-59673
Figure 89. Photoelectric Particle Counter Schematic
134
-------
the window one by one. Light from a tungsten lamp is formed by the window to
a parallel beam of an exact size, and directed onto a photodetector. Using
the Light Intensity Adjust the operator establishes the proper base voltage
from the photodetector (as indicated on the Panel Base Output Meter). Each
particle,, as it passes the window, interrupts a portion of the light beam
according to its size. This causes a specific reduction (or pulse) in the
voltage that is proportional to the size of the particle as long as specified
limits of particle concentration are not exceeded. Five counting circuits
(channels) with preset thresholds tally the particles by size. A Size Range
Adjustment is provided for each channel to permit the operator to select any
desired size ranges. A built-in Calibration Pulse Generator provides the
operator with reference pulses to simulate any particle size for adjusting
and verifying the size ranges.
Accurate sizing is obtained regardless of the color or shape characteris-
tics of the particles. Adjustment of light intensity by the operator immedi-
ately corrects for the optical density of any fluid and also for any deposits
on the window (with no change in calibration settings required). If two
Particles are in the sensing zone at the same time, their areas are summed
and reported as one large particle. This will not occur if concentration
limits recommended by the manufacturer are followed. Initial tests on oil-
Water samples at the manufacturer's laboratory also showed that the counter
becomes saturated if the recommended particle concentration is markedly ex-
ceeded. Under these conditions it is necessary to dilute the sample until
the particle concentration is below the recommended limits.
2. Initial Test i ng
The oil droplet measurement program was conducted at the Signal Oil
Company lease in Huntington Beach, California. In order to simulate the Sea
Dragon operating conditions, the centrifuge was run on salt water mixed with
Various proportions of various oils. This water is obtained from a well con-
taining seawater that has percolated through the ground from the nearby sea.
The well water contains particulate matter and it is essential to determine
its size distribution since this forms a background count that, if significant,
must be subtracted from the overall count to give the oil size distribution.
Samples of the well water were taken in bottles at the well outlet. It was
found that the suspended particulates (an iron salt complex) were photo-
Sensitive and broke down in the sunlight to give a black coloration to the
water. Table 8 gives the results of these tests-
An examination of Table 8 shows that the number of particles in a well
Water sample decreases the longer the water is allowed to stand before count-
ing. This is quite consistent with the sedimentation of the particles since
a 5^-diameter particle with a specific gravity of 3.0 has a terminal velocity,
°f about 0.0001 fps in water, and so would fall about I in. in 15 min. A
correction, therefore, should be made for this background count. If the well
Water has passed through either a centrifugal pump or the centrifuge, however,
then the majority of these particles will have been precipitated, since the
centrifugal force in water on a particle with a specific gravity of 3.0 is 20
times that on a similar sized particle with a specific gravity of 0.9 (typical
value of oi1).
135
-------
TABLE 8
EFFECT OF TIME ON PARTICLE COUNTS IN WELL WATER, MAY 14, 1970
Test
No.
Al
A2
Bl
B2
B3
Cl
•C2
Vol ume
Sampled
10 cc
10 cc
10 cc
10 cc
10 cc
10 cc
10 cc
Number Particles
5 M-
5801
5819
4583
4852
4371
3849
4483
10 M-
2051
2392
1096
1 105
1046
445
416
20 M-
748
763
269
202
202
80
52
40 U
150
166
29
27
40
8
1
80 M.
17
21
2
2
8
0
0
Comments
Counted at once,
water clear.
Counted after 5
min, water clear
to grey
Counted after 15
min, water black
Trial tests using samples of oil-water emulsion showed that the recom-
mended number of particles per unit volume sampled was usually exceeded except
for the lowest oil concentrations. It was necessary, therefore, to arrange
for a method of diluting the sample in order to reduce the particle concentration-
^' Te s t A r r a n g e me n t s
The arrangement finally used is shown schematically in Figure 89. Samples,
of the oil-water emulsion may be taken at one of three stations by pitot-type
probes inserted into pipe lines at entry to the centrifugal pump, at entry to
the centrifuge, and at exit from the centrifuge. The emulsion passes through
a flowmeter to a mixing section where filtered seawater is added and the mix-
ture then passes to the particle counter. The seawater is supplied from a
pressurized storage tank, and passes through a 5u, absolute filter and a flow-
meter before mixing with the emulsion. Thus by suitable adjustment of the
control valves the emulsion can be diluted in any required ratio and control
gained over the particle concentration as seen by the counter. The possibility
that the absolute filter was not absolute was checked by passing the dilutant
alone through the counter. The results are shown in Table 9.
Background count due to the dilutant is negligible and the steady decline
in number or particles with time suggests that the particles recorded were
actually those already in the lines downstream of the filter and dislodged at
startup.
In order to determine the effect of centrifuge rotation on the iron salt
particles in the well water, samples were taken downstream of the centrifuge
pump (i.e. at entry to the centrifuge) and at"exit from the centrifuge. The
results are shown in Table 10.
136
-------
FEED
EMULSION
CENTRIFUGE
T
OIL
DISCHARGE
SAMPLE
LINES
WASTE
t
SEAWATER
DILUTANT
STORAGE
ELECTRONIC
PULSES
TO PARTICLE
COUNTER
AIR
FLOWMETERS
DILUTANT-
0-100 CC/MIN
ABSOLUTE
FILTER
VALVE
AIR
SAMPLE-
0-100 CC/MIN
0 FLUSHING
SOLVENT
AIR
70°F,
20 PSIG
S-6IOOI
SEAWATER
DISCHARGE
Figure 89. Schematic of the Particle Sampling System
-------
TABLE 9
DILUTANT PARTICLE COUNTS, MAY 25, 1970
Test
No.
Dl
D2
D3
D4
D5
T i me ,
Hrs
13.55
13.56
13.57
13.58
13.59
Vol ume
Samp 1 ed
44 cc
44 cc
44 cc
44 cc
44 cc
Number Particles
5 V-
226
\ 12
51
46
36
10 U
17
8
3
4
3
20 y.
2
2
0
1
0
40 M*
0
0
0
0
0
80 U
0
0
0
0
1
TABLE 10
WELL WATER PARTICLE COUNTS, MAY 26, 1970
Test
No.
El
E2
E3
Fl
F2
F3
Gl
G2
G3
Ml
Vol ume
Samp 1 ed
100 cc
100 cc
100 cc
100 cc
100 cc
100 cc
50 cc
50 cc
50 cc
0
Number Particles
5 P-
30223
30927
32766
28316
27953
28482
1725
1910
1757
1520
10 M-
26023
25012
26197
27829
27648
27680
4820
5273
5002
0
20 [i
15476
13652
14837
14830
15582
15470
8189
8934
8819
0
40 (J.
760
574
662
580
674
713
1 1 1 86 .
12607
1 1833
0
80 M-
8
3
2
12
4
6
8773
8680
8569
0
Comments
Downstream of
pump; flow rate
1 07 gpm
Downstream of
pump; flow rate
225 gpm
Downstream of
cent r i f uge ; f I c
rate 225 gpm;
centrifuge spee
1 750 rpm.
138
-------
Tests E and F were made at entry to the centrifuge. The effect of water
flow rate is very small, as the number of particles in suspension remains vir-
tually unchanged as the flow rate is more than doubled. Particle deposition
in the centrifugal pump is small when compared with the results in Table 8.
(The number of particles in tests E and F should be divided by ten for a
direct comparison with Table 8, since the volume sampled is ten times
greater.)
The particle counts at exit from the centrifuge (test G) show an extra-
ordinary increase in the number of large particles instead of the expected
decrease. This was eventually found to be caused by gas bubbles whose origin
Was a pressurized nitrogen blanket on the source well that prevented entrance
of oxygen into the injection system. By extending the length of line between
the sample collection point and the counter, coalescence of these bubbles was
encouraged so that they collected on the pipe walls and did not pass through
the counter.
Test H revealed another cause of spurious readings. The vibrations trans-
mitted from the centrifuge to the counter caused the illuminating lamp filament
to vibrate; these intensity fluctuations were seen by the photodetector as
small particles. This effect is masked when there is flow through the counter
but under no-flow conditions the effect is evident. The lengthening of the
connecting line between the collection pitot and the counter (in order to
allow the bubbles to coalescence) also served to isolate the counter from these
vibrations. Further check tests with this modification showed that the back-
ground count due to well water particulates at the centrifuge exit was essen-
tial ly negli gi ble.
4. Data Reduction
It is simple but tedious to calculate the mg/l of oil present in a given
volume of emulsion if the number and size of all the oil particles are known.
The particle counter, however, does not give the exact size of each particle
but instead gives the number of particles in a given size range. Thus, an
average value of the size range has to be chosen in order to evaluate the
Volume of oil. The usual average is the arithmetic mean diameter, which is
9iven by (D. + D-)/2 where D. and D? are the upper and lower diameters in the
range. For a uniformly distributed population of droplets, the use of this
average is in error since larger droplets make a much greater contribution to
the total volume than do an equal number of smaller ones. The correct average
r 3 3 il/3
to use is the volumetric mean diameter, which is given by (D. + D?)/2
since this biases the average toward the upper limit. L J
A small program was written in FORTRAN A to evaluate the concentration
°f oil in the emulsion from an input of the size ranges, the number of drops
•n each range, and the total volume sampled. The program is listed in Figure
90.
139
-------
C PROGRAM FOR ANALYSIS OF OIL-WATER COMPOSITION DATA
C
C WRITTEN BY HERBERT N. ROSENBERG / AIRE SEARCH-LOS ANGFLES
C .^_
0001
0302
0003
0004
0005
0006
0007
0008
0009
001U
0011
0012
0013
0014
0015
0016
0017
0018
0019
0020
0021
OC22
0023
0024
0025
0026
0027
0028
0029
0030
0031
0032
0033
0034
0035
0036
0037
0038
0039
0040
0041
0042
0043
0044
0045
0046
0047
0048
0049
0050
0051
0052
0053
0054
0055
0056
C057
0058
0059
0060
0061
0062
0063
OC64
0065
0046
REAL N(5)
DIMENSION TlTLE<20),DAM<5),OVM(5),VAMI5»,VVMI5»,FAMI5),FVMm
1 NI(5),DL(5),OH(5)
1000 READ(5,5000,END=9999) TITLE
5000 FORMAT! 20A4)
RfcAU(5,500l) fl, VTUT
5001 FORMAT) 6F10.0)
REALMS, 5002) t OLl I ) ,DH{ I ) , 1-1,5)
5002 FORMATI8F10.0)
PI-J. 14159
UU IOC 1-1,5
DAM(I)-(DL(I) + OHU 11/2.0
DVMUI-ttDLl t)**3 * DHtI)**3>/2.0)**0.33333
VAM(II«PI*DAMtI )**3*NtI)/6.0
100 VVMt I)-PI*OVMtI)**3*NtI)/6.0
VSAM-O.C
VSVM-0.0
DO 110 J-1.5
VSAM-VSAM * DAM(J}*»3*NU)*PI/fe.O
110 VSVM-VSVM * DVM( J)»«3*NUI*PI/6.0
DC I?-? 1-1,5
FAh( . « (FVMtl). 1-1,5)
60C5 FORMATUHO, 'PCT VOLUMEIVOL. MEAN! ' ,5«F3. 2.8XJ 1
WRITEI6.6006) CAM
6006 FORMATt///,iHO,20X,'OIL-^ATeR PPMIAPITH. MtAN) '.F8.1
WRITE(6,600T) CVM
6007 FORMATI1HC.20X, 'GIL-WATER PPMtVOL. KLAf,) ',F8.
WRITEI&.6008) UAAM
600B FOkMATt 1H0.20X, «MCAf< DIAIAP.ITH. MLAM , hKCJNS ',F8.
6009 FORMAT(1HC,20X, 'KEAN CIrttVOL. MF.Af4), MICFUNi. ',F8.
6010 FOPMATt lrt0.20X, 'VCL. D1AIARITH. MEAN), MICRONS *,F8.
6011 FURMATUH0.20X, 'VOL. DIAJVOL. MEA'N), MICPONS ',FB.
WRITEtt, 60091 DAVM
rtRITElb.oaiO) DVAM
WRITt«6,601l) DVVM
GO TO 10CO
9999 STOP
sx)»
*• ^*"
'_--
1)
1)
I)
1)
^.
0067
END
Figure 90. Data Reduction Program
»40
-------
The program essentially calculates the oil-in-water concentration in mg/1
and is based on the volumetric mean diameter of each size range. It also
calculates what percentage the drops in each size band contribute to the total
volume, as well as two overall mean diameters that can be used to typify the
distributions. For comparison purposes the program also calculates these
values based on the arithmetic mean diameter of each size range.
Typical program outputs are shown in Figure 91.
5.
The particle counter was used on some of the centrifuge tests in order
to try to understand the basic mechanisms governing centrifuge performance.
Table II gives the recorded data and the computed results for oil concentra-
tion and volumetric mean diameter for a series of tests on the oil-containing
waste water discharge from the settling tanks used on the Signal Oil Company
Lease.
Runs I to 5 are taken at the outlet from the centrifuge under conditions
as steady as it was possible to maintain. The waste water system was not
under direct control and certain results should be discarded because of sus-
pected system transients; for example, run 2.
Runs 7 to 12 are taken at the inlet to the centrifuge. The effectiveness
of the centrifuge in reducing the oil concentration and also in mean particle
diameter can be seen. Morever what is really significant is the fact that
there are considerably more particles per cc in the 5- to I0u,-range at exit
from the centrifuge than at entry. This indicates the existence of some
Powerful droplet-shearing mechanism in the centrifuge. In runs I to 12, the
centrifugal pump was bypassed, as the waste water inlet pressure was suffi-
cient. In runs 15 to 18, although the pressure was the same, the centrifugal
Pump was run and throttled to give the same flow rate. The effect was a con-
siderable reduction of the number of large particles, indicating that the
Pump exerts a shearing action on the oil resulting in smaller particles that
are more difficult to separate.
The next series of tests, still using the oil-bearing waste water dis-
charge, was designed to determine the effect of emulsion flow rate on the
Performance of the centrifuge since the particle residence time in the centri-
fuge is reduced directly as the flow rate increases. The results are given in
Table 12.
There is a small but discernible increase in the exit oil concentation
as the throughput emulsion flow rate increases.
A significant result i s the presence of large oil particles at exit from
the centrifuge. The calculations outlined in previous paragraphs indicate
that no 40-n particles should be able to pass through the centrifuge at these
flow rates (residence times) and rotational speeds, especially when the inlet
emulsion has such a relatively low oil concentration. The inlet emulsion oil
concentration was less than 5000 mg/1, and this amount should not be sufficient
to induce hindered settling.
141
-------
Test Identification
Test June 2, 1970 Run 2
Outlet
Size range, microns
Number of drops
Pet volume (arith. mean)
Pet volume (vol. mean)
Sample Volume is 20.00 Cubic Centimeters
5.0 to 10.
79,847
i) 6.36
6.4-4
Oil-water mg/1
Oil-water mg/1
Mean dia (arith
Mean dia (vol.
0 10.0 to 20.0
35,345
22.51
22.80
(arith. mean)
(vol. mean)
. mean) microns
mean) microns
20.0 to 40.0
3,534
18.01
18.24
^
10.7
1 1.8
40.0 to 80.0 80.0 to 150.0
543 108
22.13 31.00
22.42 30.09
Measured Values
- Discharge oi 1 16 mg/1
Inlet oil 47 mg/1
Test Identification Test June 2, 1970 Run II inlet
Size range, microns
Number of drops
Pet volume (arith. mean)
Pet volume (vol. mean)
Sample Volume is 20.30 Cubic Centimeters
5.0 to 10.0 10.0 to 20.0 20.0 to 40.0 40.0 to 80.0
25,073 31,333 15,475 3,952
0.63 6.34 25.03 51.14
0.64 6.38 25.21 51.50
80.0 to 150.0
185
16.86
16.27
Oil-water mg/1 (arith. mean)
Oil-water mg/1 (vol. mean)
Mean dia (arith. mean) microns
Mean dia (vol. mean) microns
43.1 )
57.0 )
18.2
20.0
Measured Values
Inlet oi1 42 mg/l
Discharge oi1 19 mg/1
Figure 9\ . Sample Performance Test Data
-------
TABLE
WASTE WATER DROPLET COUNTS, JUNE 2, 1970
Run
1
2
3
4
5
7
8
9
10
1 1
12
15
16
17
18
Vol ume
Samp led
20.0
20.0
21 .0
21 .0
17.3
24.0
24.1
24.0
24.0
20.3
20.0
23.3
20.6
20.6
21 .0
Number of Particles
5 u.
75210
79487
50826
47775
43516
27959
20063
22767
23107
25073
24709
30193
28993
27877
27523
10 u,
14390
35345
42208
41212
37856
36447
28860
31 123
35157
31333
30265
40108
38889
36999
36482
20 u,
1 1 18
3534
4199
3122
2892
1 7732
19284
18137
1 7472
15475
14249
19645
18804
18579
18660
40 M.
186
543
265
1 16
121
5004
8329
5342
4866
3952
3360
2400
2284
2308
2021
80 |l
65
108
5
15
3
130
370
485
320
185
168
1 1
8
10
28
PPM
8.6
18.3
1 1 .4
9.7
10.3
54.8
85.9
71 .7
61 .6
57.0
51 .2
36.7
39.2
39.0
41 .4
i
Mean
Di ameter
15.3
18.0
16.7
16. 1
15.9
30.6
37.2
31.8
32.7
30.7
30.0
26.1
25.9
26.2
26.9
Comments
Samp 1 i ng at
cent r i f uge
out 1 et ,
cent r i f uge
speed i s
1770 rpm,
flow rate
is 194 gpm.
Centr i fugal
pump off.
Sampl i ng at
centr i fuge
inlet; con-
di t ions as
above. Al 1
test i ng
done on
waste water
di scharge
Sampl i ng at
centr I fuge
inlet; con-
di t i ons as
above
except pump
turned on
950 rpm
143
-------
TABLE 12
EFFECT OF FLOW RATE ON EXIT OIL CONCENTRATION, JUNE 2, 1970
Run
28
29
30
31
32
33
45
46
47
49
Diameter
27.0
27.0
26.6
30.0
29.0
29.0
29.0
29.0
29.0
29.3
Number of Part ic 1 es
5 U
55906
61 1 18
52654
57074
73829
71914
67143
66347
65352
66620
10 u,
31 76
2202
3551
46670
24460
1 7493
19614
26034
26989
30190
20 u,
1 123
672
1686
2402
516
1 146
975
1586
1676
1296
40 u
208
98
410
74
23
144
120
216
368
175
80 M-
1 1
4
14
1
1
26
1
5
9
1
mg/1
3.2
2.0
5.0
(3.4)*
6. 1
3.2
4.6
(4.6)*
3.6
5.2
6.2
4.9
(5.0)*
Mean
Di ameter
14.0
1 1 .8
16.3
14.9
12.2
14. I
13. 1
14.4
15.4
14.0
Comment
Cent r i f uge
outl et ,
2750 rpm,
flow rate
1 1 0 gpm
As above
but flow
rate 194
gpm
As above
but flow
rate 320
gpm
^Average exit oil concentration
Conclusions and Recommendations
More tests were made using the particle counter; they merely substantiated
the two main observations, namely (l) the centrifuge failed to remove all the
large oil particles and (2) there were more small particles leaving the centri-
fuge than entering it.
In order to explain these results, the
examined more closely. Stokes' Law assumes
initial design assumptions were
that the particle is alone and
falls through a fluid of infinite extent; that is, the flow field around the
particle is not affected by neighboring particles or containing walls. In
reality when there are many particles close together, the cloud of particles
falls as though it were a solid body of the same diameter as the cloud with
the average density of the oil and water. It therefore falls far faster than
would the individual particles that comprise it. If, however, the cloud ex-
tends close to the containing walls, the liquid that is displaced as the cloud
falls is prevented from freely circulating and has to percolate back through
144
-------
the cloud. The closer the particles are together, the more significant this
effect becomes. Thus, if 5-percent oil is present and distributed uniformly
as equally sized droplets, the distance between particles is only 1.3 particle
diameters, and the whole cloud falls at 53 percent of the velocity of a single
isolated particle. This effect therefore impairs the operation of the centri-
fuge, but it is not enough to explain the presence of the largest droplets in
the outlet.
It is also assumed that the liquid flowed through the centrifuge without
rotational slip. This is true in bowl centrifuges, but here the absence of
liquid prerotation, together with high flow rates, combine to produce a secon-
dary flow vortex. These coriolis force-induced vortexes exist in each flow
passage and rotate with respect to the guide vanes in the opposite direction
to the main rotation. As the fluid flows axially through each flow passage,
it rotates about its own axis as well as about the centrifuge axis. The cen-
trifugal force acting on the oil particles is different from that if the
liquid rotated as a solid body. The net result is to reduce the effective
centrifugal force. Calculations for this particular centrifuge geometry
showed that this centrifugal force reduction only amounts to about 10 percent
and thus is not sufficient to explain the oil carryover.
The mean axial velocity of the water through the centrifuge is not large
enough to re-entrain any oil from the surface of the inner barrel nor has the
velocity any radial component. The tangential velocity of the induced vortex,
however, is as high as 30 fps and it is directed radially outwards, with
respect to the centrifuge axis, for part of each revolution made by the vor-
tex. The re-entrainment of oil already deposited on the inner barrel is par-
ticularly likely where the oil film is thickest. Any devices that reduce
this thickness may be expected to benefit the operation of the centrifuge.
This reduction may be accomplished either by increasing its velocity or by
removing some (or all) of the oil.
A conical inner barrel would cause the oil to flow very rapidly towards
the exit weir under the influence of the centrifugal force component directed
along its surface. This rapidly flowing oil film is much thinner than the
cylindrical barrel film and resists re-entrainment. This probably explains
why the separation efficiency using hot emulsion is so high. The reduced
viscosity of the oil film allows faster and hence thinner flow.
The alternative is to remove the oil as soon as it is deposited. This
rciay be achieved by using more than one removal weir. This concept of oil
removal may also be achieved using a porous shield positioned around the
inner barrel. This shield is effectively an infinite number of stages. The
oil passes through the shield under the influence of the centrifugal force
and is shielded from the re-entraining effects of the vortex swirl. The oil
now flows axially along the annulus between the inner barrel and the shield
towards the exit weir. The final configuration must await the results of
further analysis and tests.
145
-------
ECONOMICS OF OPERATION
Ocean Oil Spill Recovery with a Skimming System
In an oil skimming system of large capacity, the cost of recovery per
barrel of oil is quite dependent upon the recovery rate. This is as opposed
to sorbent, sinking, or dispersant systems in which the total cost is more
dependent upon the quantity of oil recovered due to purchase and handling of
the active materials used. These costs have been estimated as follows for a
200 bbl/hr spill:
Sorbent (straw) $!8/bbl recovered
Sinking (chalk) $44/bbI recovered
Dispersant $40/bbl recovered
The cost for recovery of oil by means of the skimming system described
in this report is shown in Figure 92. The lowest cost curve is for the
recovery of oil from the water surface on a 24-hr/day basis considering
skimmer operation only. Such an operation might occur where the oil spill
is contained by booms or other restraints and the oil is being carried to
the skimmer by the wind or water currents. The second curve shows the cost
if the system also involves a towboat and an oil boom. The third line shows
cost of the entire operation supported at sea by an additional boat that
would provide accommodations and meals for the alternate crews, and storage
facilities for oil recovered (up to 2000 bbl). The uppermost curve includes
the cost of disposing of the recovered oil assuming that it must be trans-
ported to shore by barge, offloaded into vacuum or tank trucks, transported
to a disposal site, and disposed of by burial. These costs per barrel were
estimated as follows:
Barge transport to shore and unloading $1.00
Truck transport 2.00
Disposal 1.00
$4.00
Thus, for any significant recovery rate, the cost of transport and
disposal of the oil is the principal cost-
Also plotted on Figure 92 are the costs for sorbent, sinking, and
dispersal methods as estimated by the Dilltngham Corporation and reported
in their Final Report, Contract OS-I, with the American Petroleum Institute-
146
-------
$1000.00
$100.00
S $10.00
CQ
CH
LLJ
O-
to
01
o
o
2 $ I .00
to
o
o
0.10
0.01
\
X
X
X
\
N
x
k.
\
\
\
^
s
\
\
^s
V
S
^
>
\
S
<:
S
V
>
S
S
\
s
^
V
S
v
^
1
\>
— v-
X
v
\
X
s s
V
\
X
N^
s
s
v
\
\
v_
^
\
^
V
v
1 DATA FROM DILLINGHAM CORPORATION
FINAL REPORT CONTRACT OS-I
AMERICAN PETROLEUM INSTITUTE
V
4-
\^
'
^
\
SIN
y
\
<\ . ,.,.
^
N
54 S
^WHB
\
\
s,
S
k
_\
^/
™^
\
^>
\
S
bbl/hr -
bb
mm
\
\
V
>
/
••
s,
V
y
>•
S
1 •
10 100
REMOVAL RATE IN BARRELS PER HOUR
1000
S-65059
Figure 92. Cost of Removing Oil from Ocean Surface
147
-------
Tanker Ballast Water Treatment
A centrifuge of the type developed during this program may be economically
useful for the treatment of tanker ballast water- Savings might accrue by
eliminating unwanted ship operation required to wash down tanks; reduced
wear and tear on equipment; recovery of oil product and valuable cargo space;
and elimination of slop tanks, heating requirements, and harbor pollution.
In addition, there are such indirect cost savings as elimination of equipment
for tank wash-down plus appreciable reduction in steel replacement costs due
to less corrosion on tank walls.
The operational and economic data shown tabulated in this section are
based on a tanker of approximately 200,000 dwt. The various ballast treatment
techniques presented are not meant to indicate which procedure must be used,
but are given to show magnitude of savings for each method. When added to the
many items not included in the savings column such as heating costs, increased
transfer pump life, etc-, the savings projected represent exceptional values-
This economic analysis is based upon the use of a fully qualified 2000-gpm
unit; if a 500-gpm unit were used, the centrifuge costs would be approximately
doubled.
(a) Method I - Butterworth Method with Slop Tanks
The following example is given for a "load-on-top" practice
with a normal routine of cleaning tanks for clean ballast using a
2000-gpm centrifuge for final slop tank clean-up only.
Step No. I
Clean in rotation approximately one-half the cargo compartments
of the ship using Butterworth method. This puts approximately 3000 tons
of dirty wash water into the slop tanks-
Step No. 2
Fill appropriate clean tanks with clean ballast water while at
the same time draining off the clean lower part of the original
dirty ballast (dirty ballast will have settled out so that the oil
will be in the top few feet of the water in the ballast compartment).
Step No- 5
Pump dirty portion of original ballast plus wash water (collected
from cleaning remaining tanks in rotation) into a slop tank (total
capacity is approximately 7000 tons). This waste water contains
from 300 to 800 tons of oil or an average of 550 tons (4000 barrels).
148
-------
Step No. 4
Two 2000-gpm centrifuge separators can process 4000 gpm or
approximately 1000 tons per hour- This allows the dirty water to
be processed in 6 to 8 hours. Only the clean water-free oil will
be retained onboard for load-on-top or fuel as the purified water
will be pumped overboard.
A cost summary for this method is shown in Table 13.
(b) Method II - Straight Separation of Settled Ballast Water
An alternate method of using a 2000-gpm centrifuge is presented
be low-
Load ballast at discharge port and proceed to loading port-
During the voyage, the dirty ballast will settle out and all but
the top few feet can be discharged while loading. The top few
feet can be discharged through the centrifuge separator at a rate
of 2000 gpm per separator or approximately 500 tons per hour per
separator- In a 200,000-dwt tanker this last dirty ballast portion
will be approximately 10 percent of the total ballast water or
10,000 tons- This can be discharged in 6 to 8 hr using 3 centrifuge
separators.
The savings and costs between direct and indirect are somewhat
different in this case than in the normal load-on-top case. In-
direct costs in steel replacement, savings-in-washing costs,
including use of equipment, labor, etc., plus the savings in product,
etc-, are a few samples. Although less oil is reclaimed because
only the oil in the tanks filled with ballast (100,000 tons) is
processed, large savings are realized by eliminating the slop tanks.
A cost summary for this method is shown in Table 14.
(c) Method III - Straight-Through Separation All Ballast Water
The ability to directly process ballast water through the centri-
fuge while simultaneously loading crude oil in the harbor eliminates
the necessity of washing down the tanks during the voyage and there-
fore the need for slop tanks. Method II, however, assumes that
settling between the oil emulsion and seawater takes place In the
ballast tanks*
This may not be a good assumption, especially for short duration
runs. It is possible that all the ballast water (100,000 tons) must
be processed. If such is the case, a longer period of time and more
centrifuge units would be necessary. Assuming a 25-hour processing
period, eight 2000-gpm centrifuge units (500 tons per hour each)
would be necessary to process 100,000 tons of ballast water. A
cost summary for Method III is given in Table 15-
149
-------
TABLE 13
COST SUMMARY - METHOD I
Costs
Instal lat ion Costs
2 units at $100,000 = $200,000
(approximate) *
$200,000 at 9 percent
interest for 5-yr payoff = $5lJ)300/yr
Fuel Costs
Each unit requires approxi-
mately 100 hp; using an
SFC of 0.4 Ib/hp-hr, each
unit uses 40 lb (^ gal)/hr
of operation.
At $0. 15/gal, cost per unit
is $0.90/hr
Total operation per trip =
7 hr per unit
Cost per trip = $12.50
Based on a round trip of
10 days or 30 trips/yr,
Cost of fuel = $ 375/yr
Maintenance Costs
Includes periodic in-
spection and spare parts = $ 3,000/yr
Total cost per year = $54,675
Savings
Reclaimed oil
550 tons = 4000 bbl per trip at
$l.50/bbl
Savings per round
trip = $6,000
At 30 trips/yr, savings = $180, 000/yr
Total savings per year = $180,000
^Estimated procurement and installation cost based on AiResearch analysis.
150
-------
TABLE 14
COST SUMMARY. - METHOD
Costs
Savings
Installation Costs
3 units (1.5 x Method l) = $77,000/yr
Fuel Costs
(I.5 x Method I)
Maintenance Cost
$ 550/yr
$ 4,000/yr
Total costs per year
11,550
Reclaimed Oil
100,000 tons of ballast -con-
tains approximately 0.4 per-
cent oil or 400 tons per
trip, or about $4,500 per
round trip x 30 round trips
Savings = $135,000/yr
Reclaimed Cargo Space
Space of slop
tanks = 7,000 tons
Value per ton per round
trip $2.50
Savings per round trip =
$2.50,x 700 = $17,500
x 30 trips
Savings =
Reduced Repair Costs
Based on normal (using
Butterworth) repair costs
for 200,000-dwt tanker of
$3,000,000 over 20 yr
($150, 000/yr average) and
the fact that corrosion
rate can be reduced 15-fold
(API data), assume 2/3
reduction in steel replace-
ment
Savings =
$525,000/yr
$100,000/yr
Total savings per yr = $760,000
151
-------
TABLE 15
COST SUMMARY - METHOD III
Costs
Instal lat ion Costs
8 units (4 x Method I) = $205,000
Fuel Costs
8 units for 25 hr
per round trip = $5,400
(30 trips/yr)
Maintenance Costs = $ 8,000
Total costs per year = $218,400
Savings
Total savings per year = $760,000
(same as Method II)
The discussion above has attempted to give an order of magnitude in the
savings realized by using the 2000-gpm centrifugal oil/water separator.
Not only is treatment of oily ballast water a legal requirement, it is also
economically attractive. The basic improvement of the centrifuge unit pro-
posed herein over static settling type coalescers is that the unit occupies
only one-fortieth the volume required to handle similar rates of throughflow.
Based upon comparison with a settling tank/coalescer type separator of
equivalent throughflow (500 ton/hr), the savings in displacement tonnage is
80 tons per unit. This equivalent cargo space would amount to a dollar
savings of over $30,000 per unit over a 5-yr period. The design of the unit
reduces the power requirements to one-tenth of a conventional centrifuge.
In addition., the speed of the unit can be adjusted to handle various kinds of
gravity oil emulsions. Therefore, resulting separation performance (water
quality) can be varied as necessary to meet harbor or inland requirements.
The capability of the centrifuge in providing an efficient, compact, long-1ife>
separation system for a variety of inlet conditions makes it an ideal candidate
for any oil tanker requiring treatment of oily ballast water.
152
-------
Oil Field Waste Water Treatment
The following is an example from a study made by a major oil company on
the economics of a centrifuge system in the treatment of waste water in an oi
field on the Southern California coast. The requirements involved handling
500,000 bbl/day of skim tank water that averaged approximately 200 mg/1 oil,
and reducing the oil content to less than 15 mg/1.
I. Conventional Waste Water Treatment with Filtration
Cap i tal:
$8.8/bbl/day x 500,000 bbl/day $4,400,000
Future operating expense for 10 years:
$0.006/bbl/day x 500,000 bbl/day x
365 days/yr x 10 yr = $11,000,000
Present worth of future operating expense
at 10 percent discount factor:
0.52 x $11,000,000 = 5,700.000
Approximate present value of total future costs $10,100,000
2. Centrifuge Waste Water Cleanup- System
Capital:
20,000 bbl/day unit* $50,000
Miscellaneous hookup 20,000
Total $70,000
500,000 bbl/day capacity $ 1,700,000
Future operating expense for 10 years:
$30/bbl/day x 500,000 bbl/day x
365 days/yr x 10 yr = $2,700,000
Present worth of future operating
expense at 10 percent discount factor: $ 1,400,000
Approximate present value of total
future costs $ 5.100,000
3. Savings Using Centrifuge:
Approximate present value of potential
savings in total future costs:
^Estimated centrifuge cost based on AiResearch analysis
153
-------
SECTION X
ACKNOWLEDGMENTS
The design, development., and test supervision of the prototype skimmer/
centrifuge was performed by a team from AiResearch Manufacturing Company, a
division of The Garrett Corporation, under the direction of Mr. J,, S. Tyler,
Program Manager. Early system concepts and centrifuge design were completed
by Mr. J. W. Abbott. Mr. D. S. Wimpress was responsibl'e for design and develop-
ment of the skimmer and the at-sea test program. Centrifuge development was
supervised by Mr. R. W. Lewis, and Or. John Fallen made the analyses of the
size and distribution of particles throughout the centrifuge.
Mr. J. Malberti, of California Shipbuilding an'd Dry Dock Company, made
the structural design of the skimmer barge.
Mr. C. Lang, Ocean Science and Engineering, assisted in- the marine
engineering and oceanographic aspects of the project.
The practical knowledge and background experience with oil spills pro-
vided by Mr. K. S. Elmes, President, General Marine Transport of Santa Barbara,
and by the boat crews, is gratefully acknowledged.
Mr. E. 0. Kartinen, Chief Engineer, Technical Services, Signal and Gas
Company, provided valuable assistance in the centrifuge design and technical
liaison with the oil industry in obtaining test facilities and materials.
The support of the project by the Federal Water Quality Administration
(now a part of the Environmental Protection Agency) and the help provided by
Mr. Allen Cywin, Director, Division of Applied Science and Technolo-gy; Mr.
Harold Bernard, Chief, Agriculture and Marine Pollution Control Branch; Mr.
R. T. Dewling, Director of Edison Laboratories; and Mr. Gerald Stern, Project
Officer, is acknowledged with sincere thanks.
155
-------
SECTION XI
SELECTED BIBLIOGRAPHY
I. Rules for Building and Classing Steel Barges for Offshore Services,
1967 Edition, American Bureau of Shipping, New York.
2. Principles of Naval Architecture, John B. Comstock, ed., Society of
Naval Architects and Marine Engineers, 1967.
3. Chemical Engineers' Handbook, John H. Perry, ed., Fourth ed., McGraw-Hill,
New York, 1950
4. Handbook of Ocaan and Underwater_Engi_neerijig, John J. Meyers, ed.,
McGraw-Hill, New York, 1969.
157
-------
SECTION XII
PUBLICATIONS
I. AiResearch Staff, "Final Report, Sea Dragon Oil Spill Containment and
Removal System/' AiResearch Report 70-6787, December 1970.
2. Wimpress, D. S., "FWPCA Oil Skimmer Model Test/' AiResearch Report
70-6012, January 20, 1970.
3. Lewis, R. W., "Oil/Water Separator PN 585010-1-1 Performance Testing
from March 3 through March 26, 1970," AiResearch Report 70-6406,
April 4, 1970.
159
-------
SECTION XIII
GLOSSARY AND ABBREVIATIONS
API Gravity - An arbitrary scale used by the petroleum industry in the United
States to describe the specific gravity of oils. The relation between degrees
API and specific gravity is expressed by the following equation:
De9rees API B sp gr460V60"F ' l31'5 = ° API
.Boom - A device of considerable length that will float on the surface of a
body of water, and form a fence or dam to restrain the surface movement of
oil. Booms are usually flexible to conform to the water surface contours
under various sea conditions.
.Centrifugation - The process of centrifuging; the process of separating oil
and water by subjecting it to centrifugal force with a centrifuge drum.
.Emulsion - A mixture of oil and water in which the oil is distributed through-
out the water in small droplets.
How Spl itter - A device at the downstream end of the centrifuge drum to prevent
the oil layer from passing into the water discharge.
.Hertz - A unit of cycle frequency measurement. One Hertz equals one cycle
Per second.
OjJiet Pond - The central portion of the skimmer barge, bounded by the aft
bulkhead, the twin hulls, the spillway and the louvered bottom.
.Skimmer. Skimmer Barge - The sea-going vessel that supported the centrifuge
and other equipment involved in the recovery of oil slicks. The vessel was
specially designed and constructed to skim the surface of a body of water.
Ipecific Gravity - The ratio of the density of a substance to the density of
Pure water at 4"C.
Ipi1Iway - The underwater structure that extends across the entrance to the
skimmer barge. The spillway limits the amount of water that enters the quiet
Pond.
Sjjave Fence - A device that contains a series of spaced vertical elements, and
extends across the quiet pond to reduce the wave action.
jjjaye Gate - A horizontally-hinged door at the entrance to the skimmer quiet
Pond.
161
-------
°API - Degrees API (See API Gravity)
bbl - Barrel or barrels (a unit of measurement equal to 42 U.S. gallons)
F - Fahrenheit
gm - Gram
gm/1 - Grams per liter
GPM - Gallons per minute
mg/1 - Milligrams per liter
ml - Milliliter
mm - Mi 11imeter
RPM - RevolutTons 'per minute
162
-------
SECTION XIV
APPENDIXES
Page
I. Instrumentation 163
Table 16: Sea Dragon Instrumentation 164
Figure 93: Wave Height Measuring Device 166
2. Oil/Water Test, Sample Analysis Method 169
3. Centrifuge Test Data 173
163
-------
APPENDIX I
INSTRUMENTATION
The skimmer barge was instrumented to provide data as to the environmental
conditions and performance of the various equipment. A summary of this instru-
mentation is given in Table 16.
PROCESSED FLUIDS FLOW MEASUREMENT
Weir pump flow rate, overflow rate, and centrifuge inlet flow were measured
by timing the rise (or fall) of the surface level in the surge tank when selected
valves were opened or closed. The centrifuge oil flow rate was measured by
three different methods at various times during the tests.
(a) Time to fill a quart jar
(b) Time to fill oil receiver between two marks placed l/2-barrel
(21 gal) apart
(c) Positive displacement flowmeter
The oil flow rate from the surge tank oil overflow was measured either
by (l) measuring barrels of oil pumped from the oil transfer barrel and then
subtracting the centrifuge oil flow rate or (2) measuring l/2-barrels drawn
off the oil overflow weir.
WAVE HEIGHT MEASUREMENT
The ocean wave heights were measured by means of a spar buoy as shown in
Figure 93. The buoy consists of a calibrated 24-ft pole that floats vertically
in the water. A IA-in.-diameter saucer at the lower end, which is nominally
some 16 ft underwater, damps vertical motion of the pole. A small flotation
ring located 8 ft above the saucer brings the center of flotation above the
center of gravity so that the pole remains upright. As waves pass the pole,
there is very little change in buoyancy because of the small cross section of
the pole. Therefore, the pole stands still in the water as the surface waves
Pass by. Wave height and frequency are determined visually by observing the
rise and fall of the water surface on the calibrated portion of the pole.
WIND MEASUREMENT
Wind velocity was determined by means of a Danforth Model M50SB wind
velocity indicator. The 3-cup generator was mounted approximately 10 ft
above the deck of the barge. Two scales were available through a selector
switch: 0 to 25 knots and 0 to 125 knots.
Wind direction was obtained with a Newport Supply Company Model NI98
Windetector wind vane located adjacent to the anemometer generator.
165
-------
TABLE 16
SEA DRAGON INSTRUMENTATION
Sea
Air
Barge
Quiet pond
Var i able
Wave height
Wave frequency
Wave direction
Type
Water temperature
Veloc i ty
Di rection
Temperature
Draft
Velocity
Pitch magnitude
Pitch period
Rol 1 magnitude
Rol 1 period
Di rection
Inlet f 1 ow
Wave gate opening
Underwater condition
Underwater currents
Instrument
Spar buoy
Stopwatch
Compass
Visual
Thermometer
Anemometer
Wi nd vane
Thermometer
Draft 1 i nes
Sailboat speedometer
Incl tnometer
Stopwatch*
Incl inometer
Stopwatch*
Compass
Calculate (draft
times speed)
Marks on gate bracket
Port holes
Ribbons
Range
±6.0 ft
0 to 360 deg
--
0 to I20°F
0 to 25, 0 to 125 knots
No scale
0 to I20°F
0 to 6 ft
0 to 5, 0 to 10 knots
--
0 to 360 deg
—
0 to 36 in.
Not appl icabl e
Not appl i cable
Least
Count
i ft
0.2 sec
5 deg
--
I°F
1 , 5 knots
Vi sual
3 in.
5 deg
—
1 in.
--
--
Source
Spec i al bu i 1 d
A i r g u i de Mo de i 87
Danforth Model MSB
Wi ndetector
Pai nted
Kenyon Model KSI
Ai rgui de Model 87
--
Spec! al build
Spec i al bui 1 d
Spec! al bui 1 d
o
o
e\sevihere
-------
TABLE 16 (Continued)
Processed
fluids
Equipment
Centrifuge
Variable
Weir pumps flow rate
Centrifuge inlet flow rate
Tank overflow flow rate
Centrifuge oil flow rate
Tank oil flow rate
Oil vi scosi ty
Oil in water
Water in oil
Oi 1 gravity
Centrifuge speed
Wei r pump
Bal last tanks
Towing force
Speed and chronometer
Pump inlet pressure
Pump discharge pressure
Oil discharge pressure
Diesel cooling water temp
Diesel oil pressure
Diesel torque converter
pressure
Diesel fuel inlet
pressure
Oi 1 mist manifold
pressure
Bearing temperature
I nstrument
>
•
'
j
»• See text
Samples col lected
and sent to
1 aboratory
Tach on centrifuge
Hand-held tachometer
Cal ibrated rod
Dynometer
Tachometer
Gauge
Gauge
Gauge
Gauge
Gauge
Gauge
Vacuum gauge
Gauge
Battery powered bridge
Range
0 to 4000 rpm
0 to 8 ft
0 to 10,000 Ib
0 to 3500 rpm
0 to 50 psi
0 to 100 psi
0 to 50 psi
0 to 200°F
0 to 80 psi
0 to 300 psi
0 to 30 in. Hg
0 to 100 in. H20
0 to 50 mv
Least
Count
1 in.
100 Ib
—
Source
Stewart-Warner 757-W
Special build
Di 1 Ion Type AN
Part of centrifuge
Part of centrifuge
Part of centrifuge
Part of centrifuge
Part of centrifuge
Part of centrifuge
Part of centrifuge
Part of centrifuge
J
Part of centrifuge
"~ g^-
Lab supply
-------
RESERVE FLOTATION
^24 FT, I-I/4-IN. DIA
-ALUMINUM TUBING
^ FLOTATION -_-_-_-_-_
wr= WEIGHTED DISK
S-60500
Figure 93. Wave Height Measuring Device
168
-------
SKIMMER SPEED
A Kenyon Model KSI sailboat speedometer, with selectable ranges of 0 to
5 knots and 0 to 10 knots, was installed to obtain skimmer speed. This equip-
ment, however, was subject to malfunctions and speeds were usually determined
by throwing small floating objects onto the water and timing their passage
the length of the barge.
WATER TEMPERATURE
A dial thermometer was installed in the surge tank to indicate the tem-
perature of the water being processed, which was the same as the seawater
temperature.
TOWING FORCES
Towing forces were measured with a Dillon Model AN 0 to 10,000 Ib (±2 per-
cent) dynamometer. When higher than 10,000-lb peak loads were expected, the
dynamometer was installed as is shown in Figure 76. Towing forces varied
considerably as the towboat and skimmer passed over the ocean swells. The
maximum towing force was indicated by a needle follower; the minimum and
average were determined by observation.
169
-------
APPENDIX 2
OIL/WATER TEST, SAMPLE ANALYSIS METHOD
METHOD USED FOR DETERMINING CONCENTRATION OF OIL IN WATER
The trichloroethylene method was used in determining the amount of oil
in the water during the centrifuge and sea skimmer test. This method was used
rather than the Soxhlet extraction method as discussed in the FWQ.A manual
entitled "FWPCA Methods for Chemical Analysis of Water and Wastes/1 dated
November 1969. The following comments are given to justify the method used.
(a) The asphaltene content of the oils tested varied from 4 to 10 percent.
Asphaltenes are soluble in trichloroethylene; however, this material
is only slightly soluble in hexane. This would reduce the accuracy
of the hexane method.
(b) The oils used contained varying amounts of volatile hydrocarbons
that would be lost in the hexane method as outlined in page 205 of
the FWQA manual.
(c) Calibration factors were determined for each oil used in the program.
This resulted in an accuracy of ±10 percent for the colorimetric
method. Data for verification of the hexane Soxhlet extraction
method accuracy are not available per page 209 of the FWQA manual.
(d) The large number of samples obtained during the test program required
rapid method of determining oil content. The colorimetric requires
about 10 min/sample; the hexane method takes 6 to 8 hr.
71
-------
SIGNAL OIL AND GAS COMPANY
Technical Service Laboratory
Long Beach, California
Colorimetric Determination Of Oil In Waste Waters
Introduct i on
The following colorlmetric procedure was devised for rapid determinations
of small amounts of oil in waste waters. Both emulsified and floatable oils
are measured. The oil is extracted from the water by intimate contact with
an organic solvent. Although any good colorless organic solvent may be used,
trichloroethylene was selected because it has a slightly higher boiling point
(87°C) than similar chlorinated solvents. Since it has a solubility in water
of less than O.I percent and a specific gravity of 1.466 at 20°C, a good separa-
tion of the two liquids is obtained.
Apparatus
Balance
Colorimeter
Cuvettes
1000 ml separatory funnel with no-lubrication stopcock or with all
lubricant removed from ground glass surfaces
I 00 ml graduate
500 ml graduate
Glass funnel
Pint-sized sample bottle with caps
Weighing flask
China marking pencil
Ri ngstand
Alumi num foi1
Reagents
Trichloroethylene
172
-------
Procedure
Sampling Procedure. Obtaining a representative sample is of the
utmost importance. It is at this point where the greatest error
may be introduced in any residual oil determination. Since every
waste water sampling has its own peculiar difficulties, no definite
procedure can be given which would be applicable in all cases. In
general, samples from open discharge lines should be taken in wide
mouth pint bottles, preferably, the Mason type. Samples from lines
with sample cocks may be obtained in any suitable pint-size container.
During sampling the bottle should not be overflowed. Before turning
on the cap, place a square of aluminum foil over the jar mouth.
Laboratory Procedure. Mark the water level on the outside of the
sample jar with a china marking pencil. Remove the cap with the
foil and using a funnel pour the contents of the jar into the
separatory funnel. Measure a volume of 100 ml trichloroethylene
in a 100 ml graduate. Transfer approximately 50 ml of this volume
into the sample jar, recap, and shake vigorously. Remove the cap
and pour the solvent into the separatory funnel. Repeat using
approximately 25 ml of the remaining solvent to rinse off oil ad-
hering to walls of the bottle. Before removing the funnel, rinse
it with the remainder of the solvent.
Stopper the separatory funnel and shake. Set aside for a few minutes
to allow water and solvent separation. During this time fill the
sample jar to the water level mark with tap water. Transfer this
water into the 500 ml graduate and record its volume. When the sol-
vent in the separatory funnel has settled out, open the stopcock and
obtain a sample in the colorimeter cuvette. Occasionally it may be
necessary to filter the solvent containing the extracted oil to
remove small suspended water droplets. If this is necessary, moisten
the filter paper with a clean solvent before filtering. This will
aid in retaining the water and allowing the solvent phase only to
pass through the filter paper.
Zero the colorimeter with a solvent blank and read the sample at
400-465 m (j,.
Calculate as follows:
OM (mq/|)= Colorimeter Reading so,vent volume x calibration factor.
Ul ' \m9' '' Sample Volume
Standardization. Since crude oils exhibit a great range in chemical
and physical properties, it is impossible to measure minute quanti-
ties of oil with a great degree of accuracy and reproducibi1ity in
oil field waste waters by any method. To reduce errors caused by
these variables, it is suggested that calibration factors or curves
be determined for each field from which samples for oil determinations
might be encountered.
73
-------
Samples of representative crude are weighed to the nearest mg. If
the volatility of the crude is high enough to cause weight changes
when using an open container, it is suggested that a Lunge-type
weighing flask be used. For most crudes, a watch glass is satis-
factory. Samples of 50 mg or less are preferred for standardization.
The addition of 100 mis. of solvent to 50 mg of oil gives a concen-
tration of 500 gm/1. Aliquots of this stock solution are then
diluted with solvent to obtain lower concentrations. Depending upon
the type of colorimeter used, the standard containing 500 gm/1 will
probably exceed the upper limit of the colorimeter scale. With the
Klett-Summersen colorimeter, it was noted that above 250 gm/1 the
calibration factors became increasingly larger compared to those
obtained at lower concentrations.
i
Readings taken at various concentrations are recorded and the factor
is determined as follows:
. ... . gm/1 oil in standard
Calibration Factor =7—;—"-.— TT-
Colorimeter reading
Usually there is a variation of less than 5 percent between calibra-
tion factors, so an average calibration factor is used.
Limi tat ions of Test
This method was devised in an attempt to obtain a rapid colorimetric
procedure for determining oil in waste water. As such, the results may be
affected by any of the limitations and errors inherent in colorimetric
determinations.
In practice, the accuracy and reproducibi1ity of results compares favor-
ably with those obtained by the use of, ttme-consuming flocculation and extrac-
tion methods.
174
-------
APPENDIX 3
CENTRIFUGE TEST DATA
Presented in this section are data that describe the performance of the
centr i fuge.
Data in Group I were taken during one of the first runs on the centrifuge
wherein the general characteristics (pressures, flow rates, and bearing tem-
peratures) were being investigated.
Data in Group 2 were taken the following day and include further investi-
gations into the oil separation characteristics.
Data in Group 3 were taken at essentially constant centrifuge speed but
with the flow rate varied by means of the throttling valve downstream of the
pump. The oil in and out, and the water flow rates, were measured with flow-
meters. From this was obtained the flow emulsion percentage (last column).
The lab emulsion percentage was obtained by analysis of samples taken of the
inlet emulsion. The samples columns show the mg/1 of oil in the discharge
water and the percent water in the discharge oil.
Data in Group 4 were taken at a slightly lower centrifuge speed.
Group 5 includes data taken at a number of different conditions to pro-
vide a broader background for the analysis of the preceding data. To deter-
mine some of the mechanical problems, some miscellaneous tests were also made,
such as the backflow oil test to test the seals.
These above-described data were all taken on the original model of the
centrifuge, which was then modified and the following data were taken.
Group 6 data are summaries excerpted from raw data sheets similar to
sheets I through 5, to arrange the data with certain parameters constant.
Group 7 data were taken to obtain the fuel consumption rates for the
centrifuge under several operating conditions.
175
-------
Speed, rpm Pressures, psig Flow Rates, gpm 7rg Terip.,°F
Time
1315
1326
1334
1338
1340
1345
1350
1355
1358
1403
1408
1419
1420
1421
1426
1430
1433
1435
1440
1440+
1442
1450
1451
1458
1500
1507
1515
2 hr 1516
1625
1635
1640
Chron
1.24
1.30
1.35
1.38
1.39
1.43
1.46
1.48
1.51
1.55
1.58
1.66
1.69
1.73
1.78
1.84
1.90
1.94
1.96
2.00
2.12
2.17
2.31
2.37
2,76
Engine
Start
Decrease
StOD
Started
Stop
Cent.
1000
1000
1000
~IOOO
~IOOO
-1000
-1000
-IOOO
1100
1100
1050
1525
2075
2500
3125
3650
3300
3470
3325
3150
3150
3500
>3500
3575
Putrp IN
r"uirp Out
20
20
20
20
20
19
20
20
20
80
80
95
on out
i.
2.
2.8
50.
27
37
48
47
35
2.8
6.6
10.5
22.
St.
41.
57.
46
45
2
Water Ou
Water I
108
168
217
280
375
435
364
297
253
130
*KOtft»
MMm
R.U.I
3/I2/:
si Oil 0
Pressure
Pressure
Pressure
Pressure
1 Ions of
ow Rate
emperatu
to Enqin
later Eff
a! Leaki
n
it let Pr
'• taken <,
Taken 1
Taken !
Taken 'i
Oil Inss
deduced t
e >I95;
!, Strair
g Sadlv
WEIR PR^
3 sec Aft
3 sec Aft
) sec Aft
3 sec AfJ
3 Sec Aft
rted intt
J < 100 (
Drop Spe<
er Must b
r Seoara
Hfter Wat
5SURE DR( P TE
Ids Slow
"• Closi
sr Closi
er Closi
Centrif
t; Incre«
e oettin^
ion}
STS
Bleeds Fi
a I ts Va
a Its Va
g Its Va
9 It
.
s Va
se Cool ir
dirty
Imrf
CE^^TRIFUGE PERFORMANCE DATA
AIRESEARCH MANUFACTURING CO.
LOS ANGELES. CALIFORNIA
st
ve
g
Group 1
-------
Speed, rpoi Pressures, p£ig Flow Rates, gpm Brg. Ts:p.,°F
Time
1740
1750
1755
1800
Chron
2.97
ing me
Start
StOD
Cent.
2300
3300
3ump Zn
Pump Out
Oil Out
'ater Oul
Water Ii
325
295
Oil In
5.75
4.20
Oil Out
1.5
Inlet
7B,7B
78.78
Samples, ppin
Outlet
78
7S
M.TITUDC numm
OLNUTM
•Kjummi
HUM
R.U.L
Th<-
the
Cent
Foil
Low
Oil
MVHKM
3/1 Z/7
)
rifugal 1
Concent r,
nn of th
tion in
Day (Go
i/ater Sai?
>le. Thi
ion)
rhf
CENTRIFUGE PERFORMANCE DATA
Original Data in Notebook
AfRESEARCH MANUFACTURING CO.
LOS ANSELES. CALIFORNIA
Best
Group. 1
•vl
•vl
-------
R P IV\ 'S
FL0U3
times TIME CHR.OM. EM&mC £EMT. PUMP IM PBMPmir aii-aur
-RR.G. TT=M?.
-^ PPM
U// i/j % a/iiJ.
3. If
75-
75-
77
v/
3.OOO
3340
Tt.Tf
JCf
4.0
•3,85-
PUMP
32-
rd
9-7 5V
AfZ.2.
/a. 3
92. 99
9Z
32.STO
/(3.7
13. 0
92. 99
v/
F-UMP
3-/7
32.75"
7°
10
/r-7
?9
-llf-fP
43
_ZfiL
/•s-.o
9Z 9?
V/
_£0_
10
^2^
^2J_
Pu/^P
2,8*
^^?
/^•g
/J"-o
_&&
PUM?
/^gtf
•^.7/
_42_
/0.7
22^
^_
PUM?
g.l^
97 9?
z.-s/
/LOO
/.TO
-------
r.EM-r
ilM? IKf
TART
I 1 14,
STAAT
SO
10O
2000
32
So
300
Lfli
Sf K
IPS'
_30.
-29
2OCO
g^y
_LS2
700
2.0OO
?9o
20OO
o
_Q_
A22.S-
smscr
81
aaso
7, a*
33 SO
0
90
O
9O
J5L
-------
on. OUT* \tLjfff*=*/vw w*n& //u OIL, lu /Mt- 007* QtiruET" JAJL&T*
*ZA**PL£S -**}?J4 LXB . Sf
_Vf&D£R^ wA I fjf our* atrrj*
rauf »»i
IO3S
•9«so
11
II OY
9.88
/7SO
/.ad
1 .33
S2
;of
II IO
I&OO
27SO
298*
57
¥0
1133
10. jy
IKOO
70
0
Zoo tet
Z.OO
.57
1-2.
11*0
I7JTO
27SO
10
417
2.SO
»/ #7
2. SO
'f?*
AO-S?
reoo
SttP
-77
777
a
Z77O
7S"
318
O
8L
41 48
1-750
O
-395"
(3.S*
^1
^£2
/.s-V
SToT
GO
O
] - F*
-------
CMS*.*
CEttr
ill
•OM? OUT rruT. ikl OIL, mrr vtna M OIL. itJ «n. mi-r n.mer IULJTT urttBL, ,,MiBt* au.
A^g. F6HHK*. •/,
at
s^tar
•70
707&
smer
J//V
(/,»
800
is-
13
*LO
O
JZ2_
0£.
800
IOKO
is-
AS
II. TO
zoo
372.
50
3.0
406
2.0
/o
-0
235T
/ira.
it.-fl
to
JQ.
Voo
Ji.O
-ttL-
100
70
M11L
J!°.
tf°
t*° I3S
P-1h
/7C<
_Q_
rro
.22*.
3£L
•TOO
3.0
/zo/oj.
2.0
_D£ iZl
JfifO
/?.?»
0
J00_
O
70 -70
CO
1.ZS
0
-gas
o
0
-7Ci
^i.
72- 7?
_u_
27QQ
7O
-*^Q
9^/aa
^?i HM
-UL.
Lg L/ILJJ QC
«Limiot caumon otrnxM
TOTAL
f.U.L.
AIRESEARCH MANUFACTURING CO.
L.OS ANGELES. CALIFORNIA
Group 5
-------
ooTuGT IK>U£T IKIUET
a: /rt
_*£
±A.
42
-2-
.62,
.!£
_ZA
Jx£
^1
2
/2o
2?
S.9
J&2,
Z-7
46
fo
42.
2.0
^2Z
^1
/a
3Z7
g/^r
ji.1
33.000
C.2,
403.
Jl
C.8.
62
7} 7
/ /. 1
43
J
7,2.
V/
4?
J.9
/a
3.7
^a
n. ANOKLE*. CALirOMNIA
Group
7/e/y
-------
^.owc. gx^-ass
-40000
/8 Coo
•^5— d&O
1
O T^ AT
CALCULATIB
RECORDED
DRAWN
CHECKED
APPROVED
Tl
£(
,%%
\ f^/eo
/07
2,75-
/0&
vs 57^
— A
^5-
^^
32^
6, /? o
J>
OAJ1 V
^"6^
J6>
^l^f-
z^3
fX^2> o
ft! 22,
To -3/
flP£-
Data Summary - Oil Water Separator rroun ft
at 1750 + 150 RPM p
(RESEARCH MANUFACTURING O
UX ANOKLCS. CALirOMNIA
it* '/to. no
9.
*T> ^ .^ » JJ**
183
-------
J03I
IZ8
Data Summary - Oil Water Separator
at 2750 4 ISO RPH 21 to 26 API -
Sea Water
AIRESEARCH MANUFACTURING CO
LOS ANOELES. CALIFORNIA
8A
-------
HtCOIIOID
OftAWM
rnrr
7/7
Data Summary - 0)1 Water Separator
«t 2750 * 450 *PM 21 to 26 API -
" AM Sea Water
CMICKCO
AIME8CARCH MANUFACTURtNO CO.
CAUf
185
-------
GIL. &5AJC.
><9<3
/33'. HUT
j_Q£
J-L
/by
222,
JOO
£*£.
m_
l&L
/.asr
^Z
^2-
LJL f k.
CALCULATID
RECORDED
DRAWN
CHECKED
APPROVED
PATA
Oil,
AIRESEARCH MANUFACTURING CO.
UO« ANOELCS. CALIFORNIA
Group 6
>P^. 7/Z/^
186
-------
00
•vl
/ jf
/ooo
WOO
/ VOG
J 7
/
PSC6,
3.0
3o
/
o i
•
70
70
tt-ri.ists+4
0:15~
0:07,5-
O: &7
0:32,5
o:3f
o>. 11 .i
3.10V.
Oil?
O:3f
1107
/ o
1:39
7: S~o
, 5"
r.ji
0:V7
/.•/y
6:20
£:<
: 69.5
:/
Jf
ir W/fc
/
+1*
•« x n
> o >
50 O P
-i m n
z
o
z
0
m
rn
>
O
I
P
O
O
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
/. Report No.
3. Accession No.
w
4. Title FINAL REPORT
OIL/WATER SEPARATION SYSTEM WITH SEA SKIMMER,
7.
Wimpress, D.S., Goodsell, R.A.,
Lewis, R. W., and Fallen, J.
9. Organization The Qarrett Corporation
AiResearch Manufacturing Company
Los Angeles Division
Los Angeles, California
12. Sponsoring Organization
IS. Supplementary Notes
5. Report Date
6.
8. Performing Organization
Report No.
10. Project No.
15080 DJP
11. Contract I Grant No.
14-12-524
13. Type of Report and
Period Covered
16. Abstract
An oil skimming and separation system capable of processing 30,000 gallons per hour
and operating on the open ocean under Sea State 3 conditions, was designed, constructed,
and tested by the AIResearch Manufacturing Company. A 45 x 26-foot twin-hulled barge,
which contained an entrance paddle wheel and self-adjusting skimming weirs, was built
to support the skimming and separation equipment. A 500-gpm centrifuge developed during
the program was used to reduce the oil content of the discharge water to less than
100 ppm. The recovered oil contained less than 5 percent water. The oil content of the
discharge water could be reduced to Jess than 20 ppm by recycling it through the centri-
fuge a second time.
The equipment was tested in natural and simulated oil slicks off the coast of
Southern California under environmental conditions up to and including Sea State 3.
Intentionally spilled oils of API J5, ?$, and 35 gravities were recovered and separated,
as were weathered oil slicks that resulted from natural underwater seeps in the test
area. The centrifuge was particularly useful in separating the mixtures of water and
the higher API gravity oils.
This report was submitted in fulfillment of Project No. 15080 DJP Contract No.
14-12-524 under the sponsorship of the Water Quality Office, Environmental Protection
Agency.
Ha. Descriptors
»Centrifugation, *0il Wastes, *0ily Water, ^Secondary Recovery of Oil, Oil-
Water Interface, Water Pollution Treatment
ITb. Identifiers
*Sea Skimmer, *0iI/Water Separation, Oil Booms, Oil Spills, Oil Skimmer,
Oil Pollution, Skimmer
17c. CO WRR Field & Group 05E, 08C
18. Availability
19. Security Class.
(Report)
tO. Security Class.
(P»ge)
Abstractor p. 5. W Impress
21. No. of
Pages
Send To :
22. Price
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OFTHE INTERIOR
WASHINGTON, D. C. 20240
Institution
AIResearch Manufacturing Company
WRSIC 102 (REV. JUNE 1971)
GPO 913.261
-------
|