Hydrodynamics of an Oilwell Blowo'ut

D.R. TOPHAM Technical Report No. 33

HYDRODYNAMICS OF AN OILWELL BLOWOUT

D. R. Topham

Frozen Sea Research Group > Department of the Environment

825 Devonshire Road Victoria, B.C. V8W lY4

Technical Report

Beaufort Sea Project Dept, of the Environment

512 Federal Building 1230 Government St.

Victoria, B.C. V8W lY4

December, 1975

TABLE OF CONTENTS

1. INTRODUCTION

2. SIMULATION OF AN OILWELL BLOWOUT IN SHALLOW COASTAL WATERS

2.1 Selection of Test Conditions

2.2 Experimental Layout

2.3 Results

2.3.1 60 m depth

2.3.2 23 m depth

2 . 4 Di scuss ion

2.5 Conclusions

2.6 References

3. THE BEHAVIOUR OF CRUDE OIL/GAS MIXTURES INJECTED UNDER WATER

3.1 Introd ucti on

3.2 Test Te chn i ques

3.3 Results

3.4 D i scu ssi on

3.5 Conclusions

3.6 References

4. THE BLOWOUT SCENARIO

4.1 Refere nces

1

1. SUMMARY

Two aspects of an under sea oil well blowout have been investigated experi­mentally. Firstly a full scale blowout was simulated by pumping the equivalent of up to 27 m3/min of air at atmospheric pressure down to depths of 60 m and 23 m of seawater and measuring the induced flow patterns. These were similar in all cases, a central rising plume with a surface radial flow at the point of impingement. A striking feature of the radial flow pattern was a ring of waves concentric with the plume centre. This marks a division in the directions of the surface radial currents, outwards within, and inwards beyond the ring. This flow system would provide a certain amount of natural containment for the oil.

Secondly, an experiment was carried out to investigate the possibility of stable emulsions being formed close to the well exit. Mixtures of oil and gas were injected under water with appropriate velocities through a common pipe exit. Two specific oils were used to represent extremes of the types expected in the Beaufort Sea; one was known not to form stable emulsions and the other thought likely to form stable water-in-oil emulsions. In both cases the oil was shattered into droplets within a short distance of the pipe exit, with the major part of the oil in droplets 1 mm in diameter. A small proportion, of the order of 1%, was in droplets of 50 microns or less in diameter.

Most of the oil will reach the surface as 1 mm diameter droplets and be swept radially outwards by the induced surface currents and coalesce on the surface within the wave ring. After a critical depth is exceeded the oil will overcome the retaining flows and collect beyond the wave ring. The smaller drops may be carried down to depths of up to 10 m in the case of a 60 m plume and if released into a current would be carried several kilometers downstream.

····2. I NTRODUCTI ON

The rapid expansion of the exploration for oil and gas in the Canadian Arctic with the consequent risk to the environment has led to a large-scale environmental study - the Beaufort Sea Project. Of prime concern are the environmental risks attendant on an uncontrolled oil well blowout on the sea bed in ice covered waters. To assess this problem a 'standard Beaufort Sea blowout' was agreed upon as that expected from the geological forma­tions of the area. This is as follows: an initial oil flow rate of 2500 barrels/day (398 m3/day) decaying linearly to 1000 barrels/day (159 m3/day) after one month and then continuing at this rate indefinitely or until plugged. The oil contains dissolved gas within the reservoir in an amount equal to 22.7 m3/barrel when expanded to atmospheric pressure.

The work described in this report concerns the behaviour of the oil/gas mixture as it leaves the pipe exit, its transport to the sea surface, and the locally induced water flows. A preliminary appraisal of this problem based on existing information on rising bubble plumes revealed a lack of knowledge of the behaviour of such systems on the scale of the projected oil well blowout (Topham 1974), and the experiments reported here were undertaken to obtain data under full scale conditions.

2

Two separate experiments were undertaken, a full-scale simulation of the gas bubble plume in EO m of seawater using air compressors, and a tank experiment investigating the behaviour of gas/oil mixtures at an underwater pipe exit. As these two experiments were separate investigations, they are presented as self-contained sections and the major consequences are delineated in a separate section describing a probable blowout scenario.

3. SIMULATION OF AN OIL WELL BLOWOUT IN SHALLOW COASTAL WATERS

2.1 Selection of Test Conditions

It is anticipated that exploratory drilling will take place in water ranging between 15 m and 180 m in depth. The standard Beaufort Sea blowout detailed in the introduction yields an initial gas volume flow of 39 m3/min of gas at atmospheric pressure, and the experiment was designed to duplicate these conditions as closely as possible.

Bubble plumes have been studied by a number of workers, although more frequently for the case of a 2-dimensional line source of bubbles. Such systems find practical application as breakwaters (Bulson 1961), and as a means of mixing salt and fresh water to prevent freezing, (Carstens 1971). A comprehensive review of this literature has been given by Jones (1972) and the appli on of axisymmetric plumes to the oil well blowout discussed by the present author (Topham 1974), and ice melting by Ashton (1974). The mathematical modelling of such plumes is based on the analogy between density deficiency due to heat, and that due to the presence distributed bubbles throughout water column. None the theoretical models so far applied to le lumes have the expansion of the gas bubbles nted Turner (1973)

Experimenta res to water depths anticipated oil 1 blowout. are 4.5 m and 0.36 m3/min for and 10 m and 9.3 m3/min/m for

constant the signifi­

ume and i iterature are

gas ow rates far ow those of the Maximum depths and flow rates reported

axisymmetric plumes by Kobus (1968), a 2-dimensional plume by Bulson (1961).

The gas flow rates and depths of the projected well blowout, 39 m3/min of free gas and 180 m are an order of magnitude greater than previous experiments, a the present work was undertaken to provide informa­tion under near full-scale conditions.

Air was substituted for natural gas, as its solubility in water is close to that of methane, the major constituent of natural gas, allowing conventional high capacity air compressors to be used. It was not practical to perform the experiment under ice or to provide a solid sea cover over a large enough area, but it is believed that the broad features of the resulting flow patterns will be similar and can be derived from the open water case.

� &.

3

2. 2 Experimental Layout

Air was supplied from two 200 kw rotating screw compressors having a nominal maximum total delivery of 50 m3/min of free air at a supply pressure of 7.8 atmospheres. This supply pressure determined the max­imum working depth of the discharge nozzle as 60 m. Two sites were chosen in Saanich Inlet, Vancouver Island, one with a sea bed depth of 73 m with the discharge nozzle at 60 m and the other with the discharge nozzle on the sea bed at 23 m depth.

A 3.7 m sq. raft moored above the nozzle outlet served as a suspen­sion platform for a 12 m horizontal beam carrying 20 Savonius rotors mounted to measure flow speeds in the vertical plane. A steel cable between the raft and the sea bed served to locate the discharge nozzle and also as a guide for the instrument carrying beam. Addi­tional surface mooring lines provided the means to centre the raft over the plume, Figure 2.1. Radial profiles of the vertical flows were obtained by lowering the beam down the guide wire, recording the rotor speeds by means of electro-mechanical counters and stopwatch. The horizontal current strengths and directions away from the central plume were measured with a single current meter suspended from a small boat moored at various radjal positions along the surface moor­ing lines. Vertical profiles of the water temperature and density both before the plume was started and at various positions around the established plume were measured with a Gui1d1ine Model 8l01A CTO unit. Additional information on the plume structure was obtained from a vertically oriented side-scan sonar system.

2.3 Results

•

2.3.1 60 m depth

Measurements were made on plumes having flow discharges of 11, 22, and 26.6 m3/min of free air and of one at a slightly reduced depth of 53 m with a discharge rate of 27. 6 m3/min. The general features of the induced flow pattern were similar in all cases, consisting of a central rising plume with a surface radial flow as it impinged on the surface. A strik­ing feature of the radial flow pattern was a ring of waves concentric with the plume centre. This marked a division in the direction of the radial surface currents, outwards within, and inwards beyond the ring. This flow system forms a natural capture area for floating debris in the vicinity. The surface flows were not steady, but contained large scale fluctuations arising from eddies within the central plume After impinging on the surface, these spread rapidly as expanded vortices, coming to a halt at the wave ring radius which marks the complex interaction region where the fluid mixes downwards. This downward ,mixing entrains the inwards directed flow outside the ring. The side scan sonar records suggest that some bubbles are carried out to the wave ring and are mixed down to a depth of 12 m in an annular ring positioned below the wave ring.

4

A series of velocity profiles was obtained from each depth, each being based on a one-minute integration period of the meter revolution counters. Six consecutive counting periods were monitored, each separated by about 30 seconds required to record the readings; the curves shown are the means of all six of these one-minute integration periods. The indivi-dual velocity profiles were irregular in shape and the position of the maximum velocity relative to the mean profile centerline varied considerably. It is difficult to separate such varia­tions from those arising from lateral movement of the surface raft and its mooring system. Such motions, however, would tend to exaggerate the plume radius towards the surface and increase the apparent angular divergence of the plume. It was not practical to measure velocity profiles with more than one radial orientation of the instrument boom and the profiles do not account for any asymmetry due to cross currents although measurements of these before the plume was started showed them to be negli­gible in most cases.

The central plumes were initially conical in shape, but after rising to a height of roughly 23 m the radius remained approxi­mately constant with a further expansion approaching the surface. The centerline velocity decreases with height as the plume diameter increases, and changes little over the remainder of the plume height. The distributions of plume radius and centerline velocity with height are shown in Figures 2. 2 to 2.5, where the radius has been defined as that at which the velocity has fallen to 15% of the centerline value. This rather large value was chosen to minimize errors accumulated on the current meters from the transmission of surface wave action to the instrument mounting boom.

The mean velocity proviles for the 11.0 m3/min and 22 m3/min flows at 60 m depth are shown in Figures 2.6 and 2.7 and for. the 27.6 m3/min flow at 53 m depth in Figure 2.8.

These velocity profiles show little similarity with changing depth and the changes in plume shape from conical to cylin­drical suggest some change in plume development mechanisms. This is better illustrated in Figure 2.9 which shows the variation of total mass flow in the plume with height from the source for different flow rates. These all show marked irregularities with distinct reductions in gradient, a mea­sure of the rate of entrainment of fluid. Figure 2. 10 shows the mean profiles of radial velocity near the surface away from the plume centerline. These clearly show the reversal of surface current beyond the wave ring radius.

Density and temperature profiles taken before starting the plume revealed a cold layer of low salinity roughly 9 m deep overlying the main body of water, Figure 2.1 1. Profiles taken with the established plume showed strong mixing of the surface layer to have occurred within the wave ring, Figure 2.12, whilst the profiles outside remained unchanged.

5

2.3.2 23 m depth

Three airflow rates, 3. 6 m3, 29.2 m3 and 40 m3 of free air per minute were studied in detail at this depth, whilst visual observations were made of a flow of 6 2 m3/min, the latter pro­ducing such a violent interaction with the surface that the raft became unstable. At this large flow rate surface erup­tions of foam up to 1 metre in height were observed. Figures 2.13, 2.14, and 2.15 show these results, the lowest flow rate giving a conical plume with constant centerline velocity, whilst at the higher flows the plume expanded more rapidly, with a sudden expansion in diameter occurring at a height of 12 m.

These flows produced surface features similar to those of the 60 m case, with a reduced radius for the wave ring 20 m in the case of the 40 m3/min flow. This radius was a weak function of air flow rate, reducing to about 9 m for the 3.6 m3/min flow.

2.4 Discussion

The significant differences between these plumes and those reported in the literature are the high air injection rates and the large changes in buoyancy due to the expansion of the rising bubbles. The latter are shown in Figures 2.16 and 2.17 for the two test depths, calculated on a simple pressure-volume basis. As noted in the intro­duction, the effect of changing buoyancy can be treated in the same manner as a variable ambient density, and in principle the heated plume analogy can be extended to the deep bubble plume. However, this analogy requires that the bubbles be small, so that their natural rise velocity is small compared with local fluid velocities .

.. �. Observations of the size of the bubbles using an underwater TV camera

showed that this requirement was not met by a considerable fraction of the air volume. The larger bubbles ranged roughly from l.5 cm to 3 cm in diameter with separations of several centimeteres. Single bubbles of this size would have terminal velocities of 30 to 60 cm/ sec. (Datta et al, 1950).

This implies a relative velocity between the bubbles and the water of the same magnitude as the maximum measured water velocities, and this violates a fundamental assumption of the heated plume analogy. The slip velocity redu ces the effectiveness of the bubble as a buoy­ancy. The relative velocity between the bubbles and the rising water column gives rise to another difference from the heated plume. The bubble acts as a distributed smoothing screen and this may inhibit the development of large-scale eddies in the turbulent structure, and entrainment data derived from thermal plumes may not be appropriate.

These factors preclude a direct comparison with existing theories and make t�e extrapolation of the results beyond the range of the experi­ment uncertain. From the point of view of the oil well blowout the significant feature of these results is the self-containing effect of the surface flow. The following empirical relationship has been

6

developed for the dependence of the wave ring radius R metres, on gas flow rates and water depth.

1/3 R = 0.39Z ( Vf x 10.36\

Z + 10.36} where z is the water depth in metres and Vf if the volume flow of free gas m3/min. This gives reasonable estimates within the range of test results but may not be applicable much beyond this range, i.e. water depths 33 m to 60 m and volume flows 3.6 to 40 m3/min.

Figure 2.18 {a ) shows a view of the bubble plume operating in 60 m of water with 26.6 m3/min of free air injected, the dark line beyond the instrument raft marks the wave ring. Figure 2.18(b) shows a real underwater blowout. The water depths are not known accurately but are probably less than 30 m. Comparing the centre of the surface boil with that of the 62 m3/min flow obtained in the 23 m deep experiment suggests that the flow is considerably in excess of this value. The surface interaction region is distorted by a strong sea current. the 'arms' resulting from the shearing action between this and the flows induced by the plume. Similar distortions were seen during the experi­ment under strong wind conditions.

2.5 Conclusions

2.5.1 The rising gas bubbles entrain the surrounding water to form a rising plume which is initially conical in shape, but becomes cylindrical above a certain height. The plumes in 60 m of water became cylindrical about 23 m from the source. In shallow water of 23 m, the low flow plume, 3.6 m3/min, remained conical throughout, with a divergence of 6° half angle, this angle increasing with increasing gas flow. It is thought that the height at which the plumes become cylindrical is determined by the local ratio of bubble volume to entrained water flow volume.

2.5.2 The mean centerline velocities did not vary significantly with either depth or air flow for a given plume height, being about 60 em/sec for the 60 m plumes and 90 to 120 em/sec for the 23 m plumes.

2.5.3 The interaction of the plume with the surface produces a ring of waves concentric with the plume centre which acts to confine surface debris. This wave ring marks a division in the direc­tion of the radial surface flows, these being directed inwards at radii beyond the wave ring. In the case of an oil well blowout this feature should provide a certain amount of natural containment.

2.5.4 The behaviour of deep bubble plumes ;s not well described by present theoretical models and in the absence of such a model it is difficult to extrapolate the data far beyond the range covered by the present experiments. An appropriate theory is currently being developed by the present author.

7

2.6 References

Ashton G. D. Air bubble systems to suppress ice. Special Report 210 , Cold Regions Research & Engineering Lab., U.S. Army, Hanover, New Hampshire. 1974.

Bulson, P. S. Currents produced by an air curtain in deep water. The Dock & Harbour Authority, Vol. 42, 1951, pp 15-22.

Carstens T. Prevention of ice formation by forced mixing. Proc. 1st Conf. in Port & Ocean Engineering under Arctic Conditions. Tech. Univ. Trondheim, Norway, August 1971, pp 140-151.

Datta, R. L . , D. H. Napier and D. M. Newitt The properties and behav­iour of gas bubbles formed at a circular ori.nce. Conf. on Form­ation & Properties of Gas Bubbles. Inst . of Chemical Engineers ( London ) Feb. 14, 1950.

Jones W. T. Air barriers as oi.l spi.ll containment devices. J. Pet. Engr., April 1972, pp 126-14 2 ..

Kobus H. E. Ana lysis of tne flow' induced by a i,r-bubb1 e systems. Proc. 11 th Conf. on Coastal Enginee.ri:ng, 1.968, pp 1016-1031.

Topham D. R. Hydrodynamic aspects of an oil well blowout under sea ice. Interim Report, Beaufort Sea Project, Dec. 1974 , Dept. of the Environment, Victoria. B.C.

Turner J. S. Buoyancy effects in flui,ds. Cambridge Uniyersi,ty Press, England, 1973, pp 195.

8

.. +l . QJ ::::I

4. Go 0 3:

0 0 0 " 00 r-

0 o 0 C'lI o::l 0 C> 00 • • 0 -0

0 a Q) +l

0 0 0 o· 10 r-

0 ::::I 0 E

.. or-

t/) 4-0

I E 'e:; E +l 0 (V) ::::I 0 \0 0 (V) >,

10 -I

r-

IO � +l

C

J, Q) E

� or-

� Q) 0..

............, X W

r-

N

C1 Or-

1..1-E

0 r-

0.5 Velocity, m/sec.

Height, m

50

40

30

20

10

9

2 4 Radius, m

Fig. 2.2 Plume Shape, 1 1 m3/min of free air at 60 m depth.

6

0.5 Velocity, m/sec

10

Height, m

50

40

30

20

10

2

Fig. 2. 3 Plume Shape, 22 m3/min of free air at 60 m depth.

6 Radi us, m

0.5 Velocity, m/sec

Height, m 50

40

30

20

10

11

2 4

Fig. 2.4 Plume Shape, 26.6 m3/min of free air at 60 m depth

6

Radius, m

0.5 Velocity, m/sec

12

Height, m

50

40

30

20

10

2 4 6 Radius, m

Fig . 2.5 Plume Shape 27 .6 m3/min of free air at 53 m depth.

Vertical Velocity, m/sec

2

2 4 6 Radius, m

Fig. 2.6 Velocity Profiles for 22 m3/min of free air at 60 m depth Numbers on curves indicate depth in metres.

f-' \.oJ

Vertical Velocity, mlsec

1. 0 .-r-__

2

Fig. 2.7 Velocity Profiles Numbers on curves

4 6

for 26.6 m3/� of free air at 60 m depth, indicate � in metres.

--' �

8 Radius, m

,'/

Vertical Velocity, m/sec

2.

1.

2 4 6

.

Fig. 2.8 Velocity Profiles fo r 27.6 m3/� of free air at 53 m depth .

Numbers on curves indi cate � in metres.

8 Radius, m

i-' (J'l

16

Volume flow, m3/sec

60

50

40

30

20

10

20 40 60 Plume Height, m.

Fig. 2.9 Variation of total vertical volume flow of water with heiqht, ().ll m3/min, �.22 m3/min,�26.6 m3/min�27.6 m3/min of free air.

5

10

15

/

+0. 2 o -0. 2 +0. 2

34 m

Depth, m

Radlal veloci'ty, m/sec. o -0.2 +0.4 +0.2 D +0.6 +0.4 +0. 2

12 m

Radial Position, m.

.

�v

Fig. 2. 10 Radial Velocity Profiles of Induced Currents, 26 .6 m3/s.ec air at 60 m depth.

o

8 m

--'

'-I

o

10

20

30

40

50

6 7 T·C -.-._------------,

!/

27 28 29 30 5·' •• , i i i

Fig. 2.11 Temperature, salinity and density ,profiles, 60 m site, without p l ume.

21 22 23 24 d (s.t.p) .....-

--' 00

o

10

20

30

40

50

l

6 7 rOc 27 28 29 30 50/00 i ) i i "

Fig. 2.12 Temperature, salinity and density profiles, 60 m site, with plumes.

21 22 23 24 d (s.tp) $ i i i

� 1.0

1 Velocity, m/sec.

20

Height, m

20

15

10

5

2 4 Radius, m

Fig. 2. 13 Plume Shape, 3.6 m3/min of free air at 23 m depth.

21

Height. m

20

15

10

5

2 4 Velocity . m/sec Radi us, m

Fig. 2.14 Plume Shape, 29 m3/min of free air at 23 m depth

1

Velocity, mjsec

Height, m

20

15

10

5

22

2

I I

./' /

4 Radius, m

Fig. 2.15 Plume Shape, 40 m3jmin of free air at 23 m depth.

23

Height, m

60

50

40

30

20

10

2 3 4 5 6 VIVo

Fig. 2.16 Change of Relative Buoyancy with Height. 16 m.

24

He ight , m

20

15

10

5

2 3 VIVo

Fig. 2.17 Change of Relative Buoyancy with Height, 23 m.

25

(a)

(b)

Fig. 2. 18 (a) simulated blowout, (b) underwater gas flare

26

3. THE BEHAVIOUR OF CRUDE OIL/GAS MIXTURES INJECTED UNDER WATER

3. 1 Introduction

The work in this section was carried out to examine the possibility of stable emulsions being formed close to the well exit, in parti­cular any occurrence of the water-tn-oil emulsions encountered in some open sea oil spills - 'chocolate mousse' formations. These are known to be very stable and as their specific g ravity is close to that of water they are easily transported by sea currents.

The test conditions selected again correspond to those of the 'stand­ard Beaufort Sea Project oil well I, i.e. initial flow rates of 2 5 00 barrels/day (398 m3/day). with a gas content of 2 2 .7 m3 of free g as/ barrel with a range of seawater depths detween 1 5 m and 1 80 m. A pipe diameter of 1 5 cm was taken as the well exit conditinn. Cal­culations on the basis of this ori fice immersed in 15 m and 1 8 0 m of seawater with the above gas and oil flows yield the following flow velocities.

Depth m

1 5

1 80

TABLE 1

Superficial gas velocity m/sec

1 4 . 3

1 .92

Superficial 0;,1 velocity m/sec

0.2 5

0.25

The superfi c i a l v eloc i ty being that whi, ch each flui,d would have if occupying the pipe one.

Two types of crude oil were selected fOr testi,ng, Norma n Wells crude 1 supplied by Imperial Oil Limited and Swan Hills crude oil sup­

plied by Gu l f Oil Limited. The physical properties of the two oils a t 2 5 °C are shown below.

TABLE 2

Oil Type Specific Gravity Viscosi,ty Pour Point Surface Tension

Norman Wells 0 . 82 1 6 . 0 c . p . -50aC 2 4 . 7 dyns/cm

Swan Hi 1 1 s 0 . 81 5 6 . 5 c . p . 26.3 dyns/cm

These oils are considered to represent extremes of those types expected to be found in the Beaufort Sea and are those used in related studies of the effects of oil in sea ice (NORCOR 1975).

3 . 2 Test Techniques

It was anticipated that any mechanism forming an emulsion close to the nozzle would be complex in nature and that it was unlikely that

27

simple scaling parameters coul d be devised for model test ing. Tests we re therefore carried out under conditions as closely approaching ful l scale as was practical under the ti'me scale of t he study. By testing a number of different scale systems the important parameters controlling the processes could be identified.

Tests were conducted in an oval cross section tank of 68 litres capacity (standard domestic oil storage tank) containing approxi­mately 1.5 m depth of water, Fig . 3. 1 . Oil and gas were injected. separately into a 2. 2 cm diameter mixing chamber beneath the tank and fed into the water through a vertical feed pipe. Feed pipe dia­meters of 0. 6 4, 2. 2, 7.6 and 14.7 em were used to chang e the scale of the experiment. The l ength of the feed pipe was 70 cm for the 0 . 64 cm and 2. 2 cm diame ter pipe, giving tn excess of 20 pipe dia­meters for the oil/gas mixtures to stablize. Lengths of 90 cm were used for the 7 . 6 and 14.7 em diameter pipes, this being the longest length that could be accommodated and still leave an adequate water depth to the surface.

Flow durations of the order of 3 t o 5 seconds were sufficient to establish stabilized fl ow patterns and tests were carried out in both sea water and fresh water with gas only injected and with gas/ oil mixtures. Nitrogen was used as a test gas rather than methane to r educe explosion hazard, the ma;-n di'fference between t he two being the greater solubility of methane in water, and close to tile nozzle the time scale is too short for tht's t o be si:gnifi'cant. It was also assumed that the oi'l/gas -mixture leaving the well exit woul d be in equi 1 i brium and that gas di'd not rapi'dly come out of solution in the oil as it left the nozzle.

Both still and cine films of Ute resulting bubble systems were taken through windows covering the full depth of the tank.. In addition microscopic photography was employed to obtain an estimate of the

..•. small scale droplet distributions remain ing suspended in the tank.. after firing.

The configuration of the oil/gas mixture within the exit pipe was felt to be of some importance and this is known to be a strong fun c­t ion of the superficial oil and gas velocities in the pipe (Govier and Aziz 1972). The flow conditions given in Table 1 result tn different flow regimes for the two extremes of water depth. In t he 1 80 m case the flow is expected to be of the 'slug' type with alter­nate sections of oil and gas emerging. For the flow conditi.ons corresponding to 1 5 m of water the flow is expected to be of an annular pattern with the oil flowing up the sides of t he pipe and the gas up the centre. The flow vo 1 urnes of gas and on were there­fore adjusted to give the same superficial velocities for each pipe size corresponding to those of Table 1 .

The gas flow rates were cal ibrated by inverting a water filled container over the pipe exit and measuring the volume of water displaced over a k nown f l ow period. This gave flow calibrations to an accuracy of ± 1 0% . Oil fl ows were obtai'ned directly from measure­me nts of the injected vol umes.

......

28

3.3 Results

The results were all obtained in the form of either single flash pictures, exposure approximately 1 /1 000 sec or cine film taken at 64 frames/sec with an exposure of approximately 1 /200 sec. For each set of test conditions photographs were taken showing gas alone being injected and gas with the appropriate addition of oil. For the tests with oil injection an attempt was made to determine droplet size distributions from the photographs. These had to be confined to the edges of the plumes where individual droplets could be clearly distinguished and were limi'ted by the resolution to droplets of approximately 0.5 mm diameter. For one set of conditions the sma 11 drops 1 eft suspended after a test fi ring were photographed through a microscope focussed into the interior of the tank. Know­ing the magnif ication and depth of field the droplet density distri­bution was estimated. extending the resolution down to 1 5 m icrons (1 micron = 1 0-Sm).

The behaviour of gas bubbles formed at submerged orifices has been extensively studied and is well understood for low flow rates and sma 11 orifi ces where surf ace tension and buoyancy are the dom; nati ng forces . Single bubbles form and are released when the buoyancy force becomes equal to the h olding force of the surface tension around the pipe exit. The rati'o of bubble r adi'us to p ipe exi.t radius at the moment of rel ease becomes

R (3 (j' :\1/3 r = \2 r 2 pg)

where R is the b ubble radius at release and r the pipe radius, thus it can be seen that bubble size does not scale geometrically with pipe radius.

As gas flow rates are increased the dynamic pressure forces arising from the rapid bubble expansion come into play and considerably modify the expanding bubbl e shape. In addition, the bubbles merge into each other at the pipe exit. As the size of the orifice is increased the emergent bubbles fracture into small bubbles within a short distance. The overall picture of bubble behaviour over the range of nozzle sizes and gas fl ow rates tested is one of r ounded bubbles growing from the orifice, releasing and bursting into smaller fragments. At the h igher flow velocity the emerging bubbles have a fluted surface appearance and the bursting is more viol ent . As the nozzle exit is increased in diameter the bubbles became flattened horizontally and the breakaway is again accompanied by more violent bursting.

Typical bubble configurations for the pipe sizes tested are sketched in Figures 3. 2, 3.3, 3.4 and 3.5. The apparatus was too small in scale to allow testing the 7.6 cm and 1 4.7 cm diameter pipes at 14 m/sec gas exit velocity.

29

With simultaneous injection of oil and gas the oil appears to be drawn around the surface of the emerging bubble and shattered into small droplets during the release and breakup of the bubble. These droplets are then carried upwards with the rising bubble column. Figure 3.6 shows still photographs of nitrogen beinq injected under water through the 2. 2 cm diameter pipe at exit velocities of 1.9 m/ sec and 1 4 m/sec, and Figure 3.7 the same gas flow conditions with the addition of Norman Wells crude oil to the flow.

Photographs of the gas and oil column were analyzed for droplet size distribution for the following cases:

TABLE 3

Pi�e dia . Water T'y�e Gas Exit Vel . Oil T'yQe_

0 .64 cm fresh 1 . 9 m/sec Norman Wells

II " 14 m/sec H II

2. 2 cm II 1 .9 m/sec II Ii

II II 14 m/sec II II

II sea 1.9 m/sec Ii "

II Ii 14 m/sec Ii "

II fresh 1.9 m/sec Swan Hill s

Graphs resulting droplet in Fi 3.8, 3.9, 3.10 and 3. 1 1 . As noted ea ons are confined to the edges of the plumes 1 in u on to 0.5 mm. Each sample consisted of a count of approximately 100 drops total. This sample size is rather small and the graphs should be regarded as indications of drop sizes rather than definitive distri­bution functions. The results reflect the bubble behaviour in that the most marked differences can be seen between the 1.9 m/sec and 14 m/sec gas velocity exit conditions with a concentration towards the smaller sizes for the 14 m/sec gas flow velocity .

The apparatus was not capable of delivering sufficient oil volume flows to the 7. 6 and 1 4. 7 cm diameter pipes, and in these cases most of the droplets seemed to be 0.5 mm or less in diameter. The violent mixing caused by these large jets rapidly distributed the droplets over the tank volume, considerably reducing the clarity of the photographs to a point where droplet distribution functions could not be obtained.

Microscopic photography was employed to obtain the density distri­bution of small size droplets for the case of Swan Hills crude injected into fresh water through the 2. 2 cm diameter pipe with 1 . 9 m/sec gas exit velocity. These show a strong peak at about

30

15 microns diameter. The microscope was focussed into the interior of the tank through an air filled extension tube fixed to the inside of the window. The oil /gas mixture was injected and 1 eft to stand for five minutes. This allowed the circulation patterns set up by the jet to distribute small droplets throughout the tank and also for large scale droplets to rise to the surface, the droplet sizes of interest being those which would easily be carried with water move­ment if they escaped from the rising bubble column. Photographs were taken at random intervals with different magnifications, and k nowing the depth of focus, an approximate density distribution was calculated ( Figure 3.12). The total sample consisted of a count of 1 40 droplets. To check that the droplets were formed at the pipe exit and not by the bubbles hitting a layer of oil at the surface of the water, separate tests were made firing gas only with oil floating on the water . Drops of the order of 2-3 mm were formed with no sign of microscopic particles.

The density distribution of Fig ure 3.12 was obtained at half the tank depth and assuming that this i,s uniform over the tank volume, a figure of 1% is obtained for the proporti'on of injected oil suspended in a fine enoug h form to De easily carried by water movements.

3 . 4 Discussion

One of the motivations for this experiment was the question of whether or not stable emulsions could be formed clos e to the sea bed. The violent bursting action of the oil covered gas bubbles at the pipe exit produces a finely divided suspension of oil droplets in water, and the nature of this splittin g process is such that a con­tinuous vol ume of water-in-oil ernul si'on is unl ikely to be formed even w ith oil containing surfactants favour in g such formations.

The water volume flows entrained by the rising g as bubbles are such that the average density of 1 mm diameter oil droplets ;n the water column is 1 per cm3, and coalescence during their rise ;s unlikely. The remaining question to be answered ;s the behaviour of the oil on reaching the surface. According to Canevari, 1969, most naturally occuring oils favour the formation of water-in-o;l emulsions and an initial dispersion of oil drops of this type in water reverts to a continuous oil slick at the surface. It is only by the addition of further mixing energy to this continuous slick that water-in-oil emulsion ;s formed. Thus oil droplets rising under an ice sheet would quickly collect into a pool .

At a late stage in the investigation it came to light that a sample of Swan Hills crude used in related tests on sessile drop behaviour, Rosseneger 1975, formed under certain mixing conditions, much more stable water-in-oil emulsions than the sample used in the injecti:on tests. Supplementary tests were therefore carried out to compare the behaviour of these two samples, henceforth referred to as samples I and II respectively. Two tests were performed. In the first, air was bubbled continuously for several hours through a layer of 0;1 floating on water. Sample I rapidly formed a water-in-oil emulsion,

31

a large proportion of which remained stable for many days at room temperature. With Sample II, although water-in-oi1 emulsion was formed, a large proportion reverted to a continuous Oil slick within minutes of the cessation of bubbling.

In the second test oil was injected into water through a fine syringe to produce a suspension of droplets which was allowed to coalesce at the surface to simulate the final stages of the rise of the oil droplets produced in the blowout. Under these conditions both samples coalesced into continuous sli cks. The problem of the formation of stable emulsions is thus one of surface mixing and in this respect reduces to the case of an open water oil spill .

3 . 5 Conclusions

3 . 5 . 1 Tests were performed injecting both Norman Wells and Swan Hills crude oil concurrently with nitrogen gas at velocities of 1 . 9 and 1 4 m/sec into both sea water and fresh water to simulate an undersea oil well Blowout. Pipe exit diameters of 0 . 6 4, 2 . 2 , 7 . 6 and 1 4 . 7 cm were used. The release of the bubbles at the pipe exit breaRs the otl into fine droplets. In general the higher flow velocity produced smaller droplet sizes, as did the larger pipe diameters, this being associat­ed with greater violence in the bubBle release . The major portion of the oil was contained in droplets in the range 0 . 5 to 1 . 0 mm d iameter. These have terminal velocities between 5 and 7 cm/sec and if they escape from the main plume of bubbles and entrained water, w'ill rise to the surface in a comparatively short space of time.

.

3.5 .2 A small proportion , of the order of 1 %, of the oil injected, is in drops of less than 50 microns diamater whose terminal rise velocity is 0.5 mm/sec or less and t hese could be mixed downwards to depths of 1 0 m at t he outer edges of the plume! surface interaction region, and from there be carried away by small ocean currents.

3 . 6 References

Govier.G.W. and K . Aziz.The flow of complex mixtures in pipes. Van Norstrand Reinhold Co. 1 97 2 .

Dicken�, D, J. Overall and R . Brown , NORCOR Engineering The inter­action of crude oil with Arctic sea ice. DSS Contract number OSV4-0043 . September 1 97 5 .

Rosenegger L . W . Movement of oil under sea ice. Imperial Oil Ltd . Production Dept. Lab . , Calgary, Alberta, October 1 97 5

Canevari G. P . General dispersant theory. Proc. of Joint Conf. on Prevention & Control. of Oil Spills, Amer. Pet. Inst. , Dec. 1 5 -1 7 , 1 969, Americana Hotel, New York

152 em

32

..... --- 1 07 e�--"

-- --_._-- - - -

n II ,1 II II I I

c

Fig. 3.1 Experimental apparatus.

33

(a) (b) 1.9 m/sec exit velocity 14 m/sec exit vel ocity

Fig. 3.2 Bubble formation at 0.64 cm diameter pipe exit.

" . ' .

o 0 0 o 0

r r,. ( (I ._ _ )

( � ... , , \. - -

(a)

1.9 m/sec gas exit velocity

34

(b)

14 m/sec gas exit velocity

Fig. 3.3 Bubble formation at 2 . 2 cm diameter pipe exit.

35

(a)

Initial stages of bubble formation

u 0 0 [j {\ 0 o

(b)

Near breakaway

Fig. 3.4 Formation of bubbles at 7 . 6 cm diameter orifice with 1.9 m/sec gas exit velocity.

36

(a ) (b)

Initial stages of bubble formation Near breakaway

Fig. 3 . 5 Formation of bubbles at 14.7 cm diameter orifice with 1.9 m/sec gas exit velocity.

,:

(a) (b)

Fig. 3.6 Nitrogen injected into water, exit velocity (a) 1.9 m/sec, (b) 14m/sec.

W '-I

l

(a) (bl

Fig. 3.7 Nitrogen plus Norman Wells crude injected into water� exit velocity (a) 1.9 m/sec., (b) 14 m/sec.

w co

% of Total

60 ·

40 ·

20 ·

o o

% of Total 60 ·

40 ·

20

o o

•

1.0

1.0

39

I • •

2.0 3.0 4.0 drop dia. mm

1.9 m/sec gas exit velocity

• I

2.0 3.0

14 m/sec gas exit velocity

Fig. 3.8 Injection of Norman Wel ls crude into fresh water, 0.64 cm diameter pipe.

% of Total 60

40 .

20

o a

% of Total 60 t

40 ,

20

o

o

40

---

l •

1.0 2.0 3.0 Drop dia. mm

1.9 m/sec gas exit velocity

1""'-1 • J.

• •

1.0 2.0 3.0 drop dia. mm

14 m/sec gas exit velocity

Fig. 3.9 Injection of Norman Wells crude into fresh water, 2.2 cm diameter pipe.

% of Total 60 ·

40 ·

20 ·

o

o

% of Total

60 �

40 ·

20 ·

o

o

I

1.0

I 1 .0

41

I • I

2.0 3.0 drop dia. mm

1.9 m/sec gas exit velocity

I • I

2 .0 3.0 drop dia. mm

14 m/sec gas exit velocity

Fig. 3.10 Injection of Norman Wells crude into sea water, 2.2 cm diameter pipe.

% of Total

60 .

40 ..

20 .

o

o

% of Total

60 m

40

20

o

o

42

r-1 r

. u

1.0 2.0 3.0 drop dia. mm

Swan Hills crude

f--

I I T

1.0 2.0 3.0 drop dia. mm

Norman Wells crude

Fig. 3.11 Injection of crude oils at a gas velocity of 1.9 m/sec into fresh water, 2.2 cm diameter pipe.

\.

' " ' .

Drops/cm 3

1200 • ,..

1 000

800

600

400

200 •

-

43

I I

1 00 200 , .

3 00 drop dia. mic rons

Fig. 3 . 1 2 Swan H i l l s c rude inj ected into fre s h w ater with 1 . 9 m/sec gas exit v elocity , 2.2 cm diamete r pipe.

44

4 . THE BLOWO UT S C E NAR I O

T h e resu l ts obtained from th e exp e r i ments descr i bed i n S ection s 2 and 3 c a n b e combined to give a descrip t i on of th e l ikel y conse q u ences of an und ersea oil wel l blowout . The con d ; on s ass umed for th e oil and g as f l ow are those of the ' standa Beaufort S e a blowout ' of the i ntrod u ctio n , i . e . 2 5 00 barrel s/ day ( 398 m B/ day ) reducin g to 1 000 barrel s/ day ( 1 59 m 3 / day ) after on e month , th ereaf ter memain i n g con s tant . A constan t ga s con t ent of 2 2 . 7 m B / barrel w h e n expan ded to atmosphe ric p ressu r e is assum ed .

T h e g en er a l effect of suc h a blowout is a r i si n g mixtu re of oil dropl ets a n d g as b ub bl es w h i ch e ntrain the s ur rou n d i n g water to form a buoyan cy d r i v en p l ume . oi l ow rate i n th e presen t case i s l ess than 1 % of the g as f l ow an d its contr i bu t i on to the total buoyan cy can b e consi d ered negligib l e . Howe v e r . oi l may h av e a secondary effect on the pl ume in that a film of oi l on the g a s b u b bl e s m a y mod i fy t h e e ffecti ve surfa c e ten s i on an d in f l u e n ce the i r b e haviou r .

T h e l arge scal e b u b b l e plume e xp e r i men ts car ried out i n water d epths of 2 3 an d 60 m showed th at the se p l ume s a r e not 1 d e s cr ibed by ex i s ti ng theoreti cal mode l s and thu s the r e s ul be e xtrapol ated to d e eper water d epths w i th con f i dence . The main d e p a rtur e of th e r esu l ts from simpl e p l ume models is that are i t i al l cal i n s ha p e and then at some heig ht b ecome i n d r i ca no grea c h a n g es i n vertica l wa ter v elocity w i th e of e xpan sio n a n d the heig ht of transition be fun c t ion s of the g a s fl ow rate or , a s , of th e 0 o f b u b b l e vol ume ow g re ater d e pt h w ater con side re d is 1 80 m , the

p lume s ga

t h e 4 . 2 a rapidly ove r flow pl umes s the surface interaction the 60 m d epth .

wa ter th in Ft g u res 4 . 1 .

e vol ume i ncreases .. . son with t h e measured l ow gas

may b e mainly cyl in d rical , from a virtu al origin at

I t was su g g e sted in the intedm report , Toph am 1 97 4 . that w i th a p l ume in a stab l y strati ed • fl u i d may b e shed from the s i des of the p l ume . Hea v i e r u id is l ifted to h i g he r l e v el s by the entrainment of the r i si n g b u b bl es , and as far as the water alone is con cern ed , it is i n an u n stab l e con d ; on with resp e ct to the su r rou n d in g water . I f t h is escapes from the i n fl u e n ce of the b u b bl es it w i ll descen d to some eq u i l ­ibrium l evel . An u n p u b l ished wor k in C ambridg e , E n gl a n d (T. J . McDou g al l . p rivate communication ) with a smal l pl ume in a strati f i ed tan k con firms this view . However at this time n o c r i terion for critical d e n sity grad i e n t for this effect to be si g n i cant is av aila bl e . T h i s is a n i mportant con sid e ration where th e oil is in d ropl e ts small enou g h to be carried w i t h suc h flows , a n d th us to be rel eased from th e pl ume deep i n t h e ocean . I f any c h an g es in g as b u b b l e behaviour d u e to the pr ese nce of th e oi l a re

' " . ,

45

negl ected, t he b u oy an t pl ume act s merely to tr an sport t h e 0 ; 1 fro m the sea bed to the surf ace.

The o i l s sel ected for the i nj ecti on studi es are co n s i dered to represen t the extremes of the oi l s expected i n the Beaufort Sea, and repres ent an oi l wh i ch mi ght f orm st ab l e emu l s i on s (Sw an H i l l s crude) and one wh i ch wo ul d n ot ( Norman Wel l s crude) . The n ature of the fl ow of t he oi l and gas mi xture wi th i n t h e exi t pi pe i s largel y determi ned b y thei r relat i ve v ol umes , and for the con di t i ons ass umed here range from a ' slug ' f l ow of al tern at i n g re gi ons of gas and oil , t hro u gh a ' fr ot h ' f l ow to an an n ul ar f l ow of oi l s urroundi n g the gas flow at h i gh ex i t vel o ci t i es.

In the i n ject i on exper iments the o i l wa s br oken i nto a wi de ran ge of dr oplet s i zes , wi th di amaters from 3 mm downwards , a l im i t of 15 mi cron s bei n g set by t h e measuremen t tech n i q ues . T h e bulk of the o i l vo l ume was con t ai ned i n the ran ge 0 . 5 t o 1 mm d i ameter. T he h i gher gas i nj ecti o n vel oci t i es gave a sh i ft t ow ards t h e s mal l er si zes . T h e mai n mech an i sms of dropl et f ormati on appe ared to be cen tred i n t h e re gi on cl ose to the p i pe exi t whe re rel eased ga s b ubbl es s h attered i n to smal l frag me n ts . The h i gher the gas fl ow r ate the greater the fragment at io n . It was n o t poss i b l e to determi ne whether or not the s tate of the o i l flow w i th i n t he pi pe s tron gl y i nfl uen ced t he dropl et d is tribu ti o n s . The n atur al r i se vel oci ty of 1 rom oi l drops i s small comp ared wi, th t h e vert i cal wa ter f l ows of 60 cm/ sec measu red i n the pl ume exper imen t . I f the total o i l flow i s con s i dered t o be i n 1 mm d i ameter d ropl ets d i s tr i buted thro u ghou t the ri si n g water col umn , a mean den s ity of less tha n 1 drop/cm 3 ; s obt ai ne d , and i t i s un l i kely t h a t any coa lescen ce wou ld take pl ace wit hi n t h e ri s­

i n g co l umn . D ro ps of t h i s si z e have a natural r i se vel o city of t h e o rder of 5 cm/ sec and if they were swept by t he rad i al c u rrent to dept hs of 3m , wou ld r i se to t h e surface at a d i stan ce of 1 8 m , well w i th i n t h e wave

'

r i ng . The o i l wou l d be swept out to t h e wav e r i n g a n d a ft e r exceed i n g some cr i t ical depth would es cape and th en accumul ate i n an annu lar r i ng beyo nd t h i s .

Related s t ud i es of o i l trap ped under i ce s how t h a t t he o i l accumu l a t e s i n sessi le dro ps ap prox i matel y 0 . 8 cm t h i ck . I f i t is assumed that th e under s urf ace of the i ce i s perfe ct ly flat , t hi s 0 . 8 em m i n imum t h i ckn e s s can be u sed to pl ace upper l imi ts o n the area di'rectly con t aminated b y the oi l . Three s impl e s itu ati on s can be con s i dered, t h at the area wi t h i n the w ave r i n g i s i ce free and oi l i s con t ai n ed wi th i n t h i s vol u me , t h at i t h as a stat i o n ary i ce cover , or h as a s tearnly mov i n g i ce cover , an oi l th i ckness of 0 . 8 cm bei n g assumed i n the two l atter cas e s . F i gure 4 . 4 sh ows the ri se o f oi l w i th t i me wi t h i n a v ol u me en cl o sed by the waver i ng , F i gu re 4 . 5 , t he i n crease i n the outer r adi u s of an ann ul us o f oi l g rowi n g under a st ati o n ary sheet of flat i ce, and F i g ure 4 . 6 , th e wi dth o f the trai l of oi l l eft beh i n d on a m ov i n g cont i n uous i ce s heet . T h i s l as t curve i s deri ved from s i mp l e con s i deratinn s o f contin u i ty b ased o n the i n i t i al f l ow r ate of 2 5 00 b arrel s/ d ay .

A smal l pr oporti on of t h e oi l , of t h e o rder o f 1 % , may be i n d ro p l et s 5 0 mi crons i n di ameter or l ess , t hese have a n atur al r i se veloc i ty of 0 . 5 mm/ sec. It i s prob abl e th at dr op s of t h i s si ze woul d b e swept down to depth s of 1 0 m at the w ave ri n g and i f rel eased at t h i s dept h , wo ul d travel several ki l ometres duri n g thei r r i se t o the surf ace.

46

The c apac ity of a b ubb l e dri v e n pl u me to ra i se warm 1 0wl y in9 water t o the s urf ace and hence al ter the he at b al ance at the i ce/w ater i nterf ace i s ofte n expl o i ted to mai ntai n waterways o pen to s h i ppi ng . The operat i on of s u c h sys tems h as been s tudi ed by Ashton , 1 97 4 , the h e at transfer b etween the water and the i ce bei ng deduced fro m meas urements o n impi ng i ng heated air jets . Exper i me nts o n actua l i ce/water bu bb l er sys tems subseq uent l y confi rmed these predictions (Ashto n , 1 97 5 ) .

Combining the s e he at transfer meas urements with the s imu l ated b l owou t res u l ts e nab l e s an esti mate t o b e made o f the mel ting o f the i ce cover . Cons i'der a l ayer of warm water of th i. cknes s Llhm l y i ng o n the bottom i n wa ter of h me ters depth . The r i s i ng p l ume entr ai ns wa ter cont i nuo u s l y from i ts surrou nd i ngs , and the w ater temp er atu re wi th i n the pl u me att ai ns s ome v al ue between t h at of the ori g i nal war m l ayer and the m ai n body of the oce an. The r ate of e ntrai nment of f l u i d by the p l u me was found to v ary r ather e rrat i cal l y wi th hei gh t and for the p re s ent purpose a c o n­s tant mea n v al ue of 8 . 4 x 1 04 gm/ sec cm wi l l be t aken. Taki ng Ashton 1 s h e at tr ansfer da ta , and as s umi ng th at al l the he at trans ferred to the ice goe s towards me l ti ng, o ver a 2 0 meter di ameter ci rcl e , wi th a water impi nge ment ve l oc i ty of 0 . 7 m/sec the rate of i ce mel t e an be wri t ten as

£n = 21 LlTs em/d ay dt

where n i s the i ce th i cknes s and ATs i s the temper ature difference between the me l ti ng temperature of i ce and the imp i'ng i'nq w ater co l umn . W i t h the as s umpti o n of a co ns t an t mas s entrai.'nment rate. ,

Llh LlT s ::: h . I1T

For ex amp l e , a 10 me tre l ayer wa O . i n 1 00 metres tota l g i ves a me l t ra

4 . 1 References

C a bove the bul k temper ature of 1 . 05 em/d ay .

A s h ton , G . D . A i. r bubb l er sys tems t o s uppres s i. ce. Corps o f Eng . U . S. Army . Co l d Reg i ons Res e ar ch & Eng. L ab , H anover , New H amps h i re , USA , Speci al report 21 0 , Sept . 1 974 .

As hto n G. D . Experimen t al eva l ua t i on of bubb l e - induced heat transfer coeffi ci ents . Th i rd I nter nat i o nal Sympo s i um on I ce P ro b l e ms , H anov er, New H amp s h i re , USA

To ph am D. R. The hydrodyn ami c as pects o f an o i l we l l b l owout under sea i ce. I nteri m Re po rt, Beaufort Sea P roject , Dec . 1 974 , Dept. of E nv iro nment , V i ctor i a , B. C .

l Height , m

15

1 0

5

Rad i a l s u rface vel oc i ty :::! O . 5 m/ s ec

o o

o

0 1 0 I 0

o 0 0 0 0 0 0

o 0 0 0

o

" .J!

o 0 0

--- ----

0 0

0 0 { Vert i cal vel oc i ty :::! 1 m/ s ec 0 0

0 0 0 0 0

0 o ,/ 0 0

o 0 0 0

o 0 0 0 o 0

0

5 1 0

F i g . 4 . 1 P l ume S hape , 15 m de pth .

�

15

..j:::. �

2 0 Rad i u s , m

48

E

u

OJ

til

h

'-.

E

0

..c

<::::t

+-'

r--..

0..

OJ

0

-0

11 E

>

, +>

0

...

... \D

U

0

r-

(!)

OJ

0..

>

n:;I

..c

.-0

VI

n:;I

N

u

(J)

E

.,.

.... ::l

+-'

s.

.. (l)

CL

>-

N

r-<::::t

E

00

�

0

0

0

0

00

0

0

cO 0

0

0

•

o

•

0

0

0

CJ)

+>

00

10

0

•

0

0

o

•

00

0

0

o

•

..c

CJ)

LL

'r-OJ

0

0

0

:c

\D

<:

:::t N

Hei g ht , m

150 -

• 0 0 ') .

o o 0 o 0

0 0

o 0

0 •

0/ 0 0

· -o

• o

1 00 · � o

o o

o o

o o o

o _

• •

50- �.

•

•

49

-

F i nal Deve l o pment � > . 01 p

L ow E ntrai nment , !uJ < . 001 p

In i t i al Devel o pment � > . 01 p

•

5 0 Rad i us , m

F i g . 4 . 3 P l ume S hape , ' 80 m d e pt h .

O i l d e p t h , cm

300

200

1 00

1 0 20 30 T i me , days

F i g . 4. 4 O i l de p t h , o pen wa te r co n ta i nmen t, 6 0 m wa ter d e p t h .

O uter radius , m R

2000

U1

1 000

1 00 200 300 Ti me , days

F i g . 4 . 5 Rad i u s of con tami nated a rea a s a funct i on of t i me for ses s i l e drop th i c knes s of 0.8 cm .

Slick width

600

400

200

52

l. 0 2 . 0 k m/d ay

Fig . 4 . 6 S l ick width , moving ice cover f or oil disch arqe o f 2500 b arrels/day and 0 . 8 c m th i ckness.