contract research report 405/2002 · the experiments studied the flow of liquid across a bund floor...

HSEHealth & Safety

Executive

A series of experiments to studythe spreading of liquid pools with

different bund arrangements

Prepared byAdvantica Technologies Limited

for the Health and Safety Executive

CONTRACT RESEARCH REPORT

405/2002

HSEHealth & Safety

Executive

A series of experiments to studythe spreading of liquid pools with

different bund arrangements

P S Cronin and J A EvansAdvantica Technologies Limited

Ashby RoadLoughboroughLeicestershire

LE11 3GRUnited Kingdom

Assessment of the hazards posed by the storage of flammable or toxic liquids in large tanks can beassisted by the use of mathematical models to calculate the consequences of leakages. Theseconsequences may include fires or explosions from dispersion of flammable vapours, or harm topersons from inhalation of toxic vapours. One component of such models is a mathematicalrepresentation of the spreading of a liquid pool, and in recent years a number of pool spread modelshave been proposed and implemented.

However, there are a number of issues in the formulation of the spreading models that have yet to beresolved. In particular, one area of uncertainty is the boundary condition applied at the front of thespreading pool. Boundary conditions which are generally accepted as applicable to the spread of oil onwater and to the dispersion of a cloud of dense gas in air may not be applicable to the spread of aliquid on land, as the balance of competing physical phenomena at the spread front is quite different.Resolution of the uncertainties this creates has been hampered by a lack of reliable experimental dataat a large scale. As a result, the Health and Safety Executive have contracted Advantica to carry out aseries of ‘liquid spread on land’ experiments at their Spadeadam Test Site to provide a sufficientlydetailed database to help to resolve this issue. This report describes the resulting programme of58 experiments.

The experiments studied the flow of liquid across a bund floor and measured the amount of liquid thatescaped the bund for a wide range of bund geometries.

This report and the work it describes was funded by HSE. Its contents, including any opinions and/orconclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy.

HSE BOOKS

ii

© Crown copyright 2002Applications for reproduction should be made in writing to:Copyright Unit, Her Majesty’s Stationery Office,St Clements House, 2-16 Colegate, Norwich NR3 1BQ

First published 2002

ISBN 0 7176 2255 X

All rights reserved. No part of this publication may bereproduced, stored in a retrieval system, or transmittedin any form or by any means (electronic, mechanical,photocopying, recording or otherwise) without the priorwritten permission of the copyright owner.

Report Number: R 4418 Issue:01

COMMERCIAL IN CONFIDENCE Page iii

Executive Summary Assessment of the hazards posed by the storage of flammable or toxic liquids in large tanks can be assisted by the use of mathematical models to calculate the consequences of leakages. These consequences may include fires or explosions from dispersion of flammable vapours, or harm to persons from inhalation of toxic vapours. One component of such models is a mathematical representation of the spreading of a liquid pool, and in recent years a number of pool spread models have been proposed and implemented. However, there are a number of issues in the formulation of the spreading models that have yet to be resolved. In particular, one area of uncertainty is the boundary condition applied at the front of the spreading pool. Boundary conditions which are generally accepted as applicable to the spread of oil on water and to the dispersion of a cloud of dense gas in air may not be applicable to the spread of a liquid on land, as the balance of competing physical phenomena at the spread front is quite different. Resolution of the uncertainties this creates has been hampered by a lack of reliable experimental data at a large scale. As a result, the Health and Safety Executive have contracted Advantica to carry out a series of ‘liquid spread on land’ experiments at their Spadeadam Test Site to provide a sufficiently detailed database to help to resolve this issue. This report describes the resulting programme of 58 experiments. The experiments studied the flow of liquid across a bund floor and measured the amount of liquid that escaped the bund for a wide range of bund geometries. From a large tank with an initial fill height of up to 1.8 m, water was released through a slot at the base and was free to spread over distances of up to 10m on a specially constructed horizontal concrete surface. The rate of release was such that the vessel emptied in about 30 seconds. The experiments investigated the flow rate of water from the vessel, the initial rate at which the pool spread and the quantity of water which overtopped the bund. The data collected are summarised in this report. The large scale of the test rig means that the experiments also have a direct bearing on issues of a practical nature. Most tanks used for storage of flammable or toxic liquids in the United Kingdom have capacities less than 50,000 cubic metres, although several tanks do exist with capacities in excess of 100,000 cubic metres. The test tank constructed at Spadeadam replicates a tank of about 150,000 cubic metres capacity at a linear scale of one-twentieth, or alternatively a tank of about 20,000 cubic metres capacity at a linear scale of one-tenth. Hence, the experiments give information that can be related to behaviour on onshore sites.


COMMERCIAL IN CONFIDENCE Page iv


COMMERCIAL IN CONFIDENCE Page v

Contents 1 Introduction .......................................................................................................1 2 EXPERIMENTAL DETAILS ................................................................................2

2.1 THE TEST RIG ................................................................................................2 2.2 INSTRUMENTATION..........................................................................................2

3 Experimental Programme.................................................................................3 3.1 PHASE 1 ........................................................................................................3 3.2 PHASE 2 ........................................................................................................4 3.3 EXPERIMENTS TO VIDEO THE WATER FRONT ....................................................4

4 Experimental Results........................................................................................5 4.1 WATER FLOW.................................................................................................5 4.2 TIME OF ARRIVAL OF THE LIQUID FRONT ...........................................................6 4.3 WATER OVERTOPPING ....................................................................................7 4.4 IMAGES OF THE WATER FRONT ........................................................................8

5 Discussion .........................................................................................................8 5.1 PHASE 1 ........................................................................................................8 5.2 PHASE 2 ........................................................................................................9

6 Summary..........................................................................................................10 7 References.......................................................................................................10


COMMERCIAL IN CONFIDENCE Page vi


COMMERCIAL IN CONFIDENCE Page 1

1 INTRODUCTION Assessment of the hazards posed by the storage of flammable or toxic liquids in large tanks can be assisted by the use of mathematical models to calculate the consequences of leakages. These consequences may include fires or explosions from dispersion of flammable vapours, or harm to persons from inhalation of toxic vapours. One component of such models is a mathematical representation of the spreading of a liquid pool, and in recent years a number of pool spread models have been proposed and implemented [1,2].

However, there are a number of issues in the formulation of the spreading models that have yet to be resolved. In particular, one area of uncertainty is the boundary condition applied at the front of the spreading pool; in this respect the modelling described in [1] is fundamentally different to that described in [2], for example, and in a number of other software models known to be in use. Boundary conditions which are generally accepted as applicable to the spread of oil on water and to the dispersion of a cloud of dense gas in air may not be applicable to the spread of a liquid on land, as the balance of competing physical phenomena at the spread front is quite different. Although information is available from a variety of sources ([3,4,5], for example), resolution of the uncertainties this creates has been hampered by a lack of reliable experimental data at a large scale. As a result, the Health and Safety Executive have contracted Advantica to carry out a series of “liquid spread on land” experiments at their Spadeadam Test Site to provide a sufficiently detailed database to help to resolve this issue, as well as to allow more general validation. This report describes the resulting programme of 58 experiments.

The experiments studied the flow of liquid across a bund floor and measured the amount of liquid that escaped the bund for a wide range of bund geometries. From a large tank with an initial fill height of up to 1.8 m, water was released through a slot at the base and was free to spread over distances of up to 10m on a specially constructed horizontal concrete surface. The size of the slot was such that the tank emptied completely in about thirty seconds. More rapid releases can be imagined, and indeed have occurred in practice when storage tanks have failed catastrophically. However, the rate of release in these tests was judged adequate for the intended purpose of the tests. The experiments investigated the flow rate of water from the vessel, the initial rate at which the pool spread and the quantity of water which overtopped the bund. The data collected are summarised in this report and full copies of the data are provided with the report in electronic format.

The large scale of the test rig means that the experiments also have a direct bearing on issues of a practical nature. Most tanks used for storage of flammable or toxic liquids in the United Kingdom have capacities less than 50,000 cubic metres, although several tanks exist with capacities in excess of 100,000 cubic metres. The test tank constructed at Spadeadam replicates a tank of about 150,000 cubic metres capacity at a linear scale of one-twentieth, or alternatively a tank of about 20,000 cubic metres capacity at a linear scale of one-tenth. Hence, the experiments give information that can be related to behaviour on onshore sites.



2 EXPERIMENTAL DETAILS

2.1 The Test Rig The release vessel (Figures 1 and 2) was designed to represent at 1/20th scale a quadrant of a 70m diameter storage tank. The lower section of the vessel was constructed from a section of curved mild steel with a 3.5m diameter. Elsewhere the cross-sectional shape was modified slightly in order to strengthen the tank whilst maintaining the same cross-sectional area, as shown on Figures 1 and 2. For all of the tests, the vessel was filled with ambient temperature water, fed through a 3” hose, from the site fire water main. The tank was also fitted with a 2” nominal bore gate-valve, which could be used to drain it.

The liquid release mechanism is shown in Figure 3. A slot was cut around the full circumference of the lower section of the tank wall, and into it was fastened a quarter-circular strip of mild steel containing a slot 25 mm high through which the water could be released. The bottom of the slot was 20mm above the concrete surface. The water was initially held in the tank by an array of five closed flaps. On the outside of the tank was a mechanism whose activation opened the flaps and allowed the water to flow freely out of the tank. The release mechanism was controlled by a PC based SCADA system connected to a sequence timer system which co-ordinated the operation of the release mechanism and the instrumentation.

Bund walls were fabricated from mild steel and set up on a flat 15m square concrete pad with the release vessel at one corner. Thirteen different bund configurations were used, 9 circular and 4 square, see Figures 4 to 10. All configurations were sized to give nominally the same containment volume. The two sides of the 90o quadrant were enclosed with 250mm deep flat plate, and were assumed to act as walls of symmetry. Perspex sections were fitted into one of the side walls to allow video records to be made of the water flow.

For some tests, 1 circular and 2 semi circular sections were installed within the bund to represent a regular array of additional tanks, see Figures 10 and 11. The sections were fabricated from 200mm wide flat steel strip rolled to a 3.5m diameter.

2.2 Instrumentation A scale was marked on the inside of the tank to show the level of water within it above the bottom of the slit. In addition a pressure transducer was fitted on the back of the tank at a height level with the slit. During the tests the signal from the pressure transducer was recorded on a transient recorder sampling at 200Hz.

The movement of water across the bund floor was monitored using up to 60 resistance probes, fixed in position on the concrete bund floor using terminal block. The positions of the resistance probes used for these tests are shown in Figures 12 and 13 and the polar co-ordinates of the probes are listed in Table 1. The resistance probes consist of two electrodes separated by a gap (typically 5-10 mm), which provides a high electrical resistance. The arrival of the water lowers the resistance across the gap and triggers a TTL voltage step output from a purpose-built electronic circuit, which acts to terminate a computer based counting register on a counter



board. Counting was started when the release was initiated, and thus, using the known count frequency, a measure of the time of arrival of the water front at a specific probe location was obtained. The accuracy of the system has been checked by comparison with cine records and has been shown to indicate arrival times with an accuracy of better than 1 millisecond. As there may be a slight variation in the time at which the release valves opened after initiation of the release, time zero for the resistance probe records has been taken as the time when the first resistance probe was triggered.

Water that overtopped the bund was caught in a polythene sheet attached to the (outside) edge of the bund, then pumped into calibrated containers to measure its volume.

Up to four video cameras were used in the experiments; the images being recorded on MiniDV video tape. Video timers provided a timescale for the images on the video tape. One camera was used to provide general images of the release, another provided an overhead view of the floor of the bund whilst the other two cameras were used to monitor items of specific interest, for example, the time of arrival of the water front at the bund wall, or the release of water in the vicinity of the tank.

3 EXPERIMENTAL PROGRAMME The Test Programme was made up of two Test Phases. The experiments in Phase 1 were conducted using circular bunds and the experiments in Phase 2 were conducted using square bunds. The experiments conducted in each phase are detailed in Tables 2 and 3.

In both Test Phases, three nominal fill heights were used for the experiments: 1.45m, 1.60m and 1.75m, corresponding to water inventories of approximately 3.9 cubic metres plus and minus 10%. The actual fill height in each experiment is given in Tables 2 and 3. The water inventory of the release vessel at the three nominal fill heights is given in Table 4.

3.1 Phase 1 In the first of the two Test Phases, 37 experiments were carried out to examine liquid spread over flat uninterrupted terrain and interaction with a single, circular bund wall. Different combinations of bund location and height were investigated and bunds with face angles, relative to the oncoming water front, of either 30 degrees, 45 degrees or 90 degrees were used. Figures 4 to 6 show the different bund arrangements used in the experiments in Phase 1. They were chosen so that the capacity of the bund was close to 3.9 cubic metres in each case. Table 5a gives the actual volume of the bund for the different experimental arrangements.

The resistance probes were positioned as shown in Figure 12. Probes deployed along two radial lines monitored general progress of the front, and probes located along circular arcs gave information about the radial symmetry.



3.2 Phase 2 In the Phase 2, 21 experiments investigated releases into square bunded areas, in some cases with obstructions present. Figures 7 to 10 illustrate the different configurations. In all of these cases, the bund walls presented a vertical face to the oncoming water. Again each configuration was chosen so that the capacity of the bund was close to 3.9 cubic metres. Table 5b gives the actual volume of the bund for the different experimental arrangements

Configurations SQ1 and SQ2 each represented one-quadrant of the case of a square bund surrounding a concentric circular tank, a situation that is often encountered in practice; the two configurations differed because they made use of the different lines of symmetry in that full-plan arrangement. If the edges of the test pad had no impact on the water flows then bunds SQ1 and SQ2 would give identical overtopping results. Significant differences between results would show that edge effects were present, suggesting that a degree of caution would be necessary in extrapolating these quadrant results to the full-plan situation.

In addition, because each of configurations SQ1 and SQ2 had the same plan area and the same height as the highest of the circular bunds, appropriate comparison of results would address the question of how, for a given release into a bund of given capacity, the retention capability is affected by the shape of the bund. Similarly, square configuration SQ3 was chosen to have the same plan area and the same height as the lowest of the circular bunds, allowing the same question to be addressed in these different circumstances.

Finally, square arrangement SQ4 was defined to have the same (free) area and the same height as square arrangement SQ3, the only difference being the introduction of obstacles. Appropriate comparison would allow conclusions on the effect of obstacles on the flow and overtopping behaviour.

The resistance probes were positioned as shown in Figure 12 for the experiments with bund arrangements without obstacles (SQ1, SQ2 and SQ3) and as shown in Figure 13 for the experiments with obstacles present (bund arrangement SQ4).

3.3 Experiments to Video the Water Front Three further experiments were carried out to film the shape of the profile of the water. The water front was filmed using a Kodak Motioncorder 500 high speed camera system operating at 125f.p.s., set-up as shown in Figure 14. For all three releases an initial water head of 1.60m was used and no bund walls were used so the water spread unrestricted across the concrete pad. Video images of the water front were taken at 5m, 7.1m and 10m radii.



4 EXPERIMENTAL RESULTS

4.1 Water Flow The water pressure (and hence water depth) was measured by the pressure transducer located at the base of the release vessel in each experiment. The liquid head data for all the tests carried out is included with this report in electronic format.

Figure 15 shows the variation of water depth in the release vessel with time, for three typical tests with the three initial fill heights. Also shown on this figure are predicted values obtained using the Bernoulli equation for the outflow velocity and the measured value for the area available for outflow, assuming a discharge coefficient (Cd) with a value of 0.64:

( )2

0

221

��

�

�

��

�

� ×××××−=

cross

outd

A

tgACHtH

where

H0 is the initial height of the water, nominally 1.75m, 1.60m or 1.45m

Aout is the cross-sectional area of the hole = (0.025 * 2 * 1.75 * π)/ 4 m2

Across is the cross sectional area of tank = 2.411m2 (from tank capacity),

g is acceleration due to gravity = 9.81ms-2, and

t is the time, in seconds.

The agreement between the values inferred from the measurements and the predictions is very close. This gives some confidence that the measurements were made correctly and that the release mechanism worked as intended.

The measurement of the liquid head can be used to determine the water flow rate from the vessel. Figure 16 shows the variation in water flow rate with time inferred from the measurements of pressure for the same three experiments.

The data recorded by the pressure transducer inevitably contain some random, high frequency noise and, possibly, other more physically-based, oscillations, superposed on the signal recording the changing head of water as the tank empties. Such oscillations are particularly apparent in the first few seconds immediately following the opening of the flaps to release the water. As a result, some form of time averaging is required to remove this high frequency component from the data if they are to be differentiated numerically in order to infer a representative flow rate from the vessel. Figure 15 and Figure 16 show the liquid level and the flow rate from the vessel inferred from the numerical differentiation of 'smoothed' pressure data for three specific experiments. The curves shown in this figure were produced using 0.5 second time averaging to remove the high frequency oscillations. As a check, it was found that integrating the resulting flow rate over time gave agreement to within 1% of the original volume of water recorded as being in the tank at the start of the experiment.



4.2 Time of Arrival of the Liquid Front Resistance probes data were collected for thirteen of the experiments conducted with circular bunds and sixteen of the experiments conducted with square bunds. Tables 2 and 3 show the experiments in which the resistance probes were used. The time of water arrival at each resistance probe in these experiments is given in Tables 6 to 9. As there may be a slight variation in the time at which the release valves opened after initiation of the release, time zero for the resistance probe records has been taken as the time when the first resistance probe was triggered. The complete resistance probe data are included with this report in electronic format.

Figure 17 shows the time of arrival of the liquid front against distance from the release vessel for representative tests with a circular bund with the three initial fill levels. These data were collected in experiments where the bund wall was located 10m from the tank centre.

The time of detection of water at the resistance probes was used to infer the velocity of the front. Using the time difference between arrival of water at adjacent resistance probes along a radial line was found to give a local velocity with a significant amount of variability. Consistent with this, small local variations in progress can be observed on the video records. However, a smooth polynomial curve was fitted to all of the time of arrival data collected in experiments with the same initial fill height in the vessel and this curve was differentiated to obtain an ‘averaged’ front velocity. A comparison of this inferred ‘average’ velocity with one particular set of ‘local’ velocity data is shown in Figure 18.

It should be noted that an examination of the video records and time of arrival records of the experiments suggests that the disruption to the flow caused by the presence of the resistance probes may have slowed the progress of the front slightly along the line of the probes. The size of the terminal block used to hold the resistance probe wires close to ground level was approximately 14mm and it was observed that there was a small wake-like region created in the lee of each probe. The video records suggest that these wake-like regions did not spread to affect the radial outflow at other locations. The maximum effect this produced appears to have occurred in the tests with the greatest initial fill level in the tank. In this case, the maximum difference in the time of arrival of the flow along the line of the probes and elsewhere at a distance of 10m (i.e arrival at the bund wall in the relevant tests) is about 0.5 secs. This suggests that the average speed of progress inferred from the resistance probe measurements systematically underestimates the true value by 0.05 m/s at most (c.f. observed value of about 0.37 m/s – corresponding to a discrepancy of about 13% in the values obtained from the resistance probe data along the 45 degree radial line). Whilst such differences should be born in mind if the data are used to compare with the predictions of mathematical models, they are of a smaller magnitude than differences in the predictions from the models referred to in [1] and [2].

Figure 19 shows the time of arrival of the liquid front along the radial line from the release vessel at 45 degrees for bund geometries SQ3 and SQ4 for the three initial fill heights. Examination of the video records of these experiments reveals that the flow was channelled between the adjacent obstacles, resulting in a faster, deeper



flow between these ‘model’ tanks. The radial symmetry was disrupted and the flow appeared to separate from the obstacle tank sides, producing two tongues of fluid from between the model tank obstacles. This flow hit the retaining walls and spread out along them, meeting at the sides and in the corner and then moving back towards the rear of the tank obstacles. The data show that the effect of the obstacle is similar in all of the experiments, with the water reaching the rear face of the obstacle (in its wake) after it has reached the far bund wall.

The time of arrival of the liquid front along the 22.5 degree radius against distance from the release vessel for bund geometries SQ3 and SQ4 for the three initial fill heights is shown in Figure 20.

4.3 Water Overtopping The volume of water that overtopped the bund in each experiment was found to vary with the initial fill height, the profile of the bund wall and its distance from the vessel. The volume of water that overtopped the bund in each test is given in Tables 10 and 11. In order to compare the information from different experiments in graphical form, the volume of water that overtops has been normalised by reference to the bund capacity for a 90 degrees bund wall (3.9 m3). Figure 21 shows the results for a number of the Phase 1 experiments in which the tank was filled to a height of approximately 1.45m. The volume of water released in these experiments corresponds to approximately 90% of the capacity for a 90 degrees bund wall. In all cases, even allowing for the reduced capacity as a result of the sloping bund walls, the volume was less than the capacity of the bund. The figure shows that the amount of water overtopping the bund varies with the wall angle and distance of the bund from the vessel, with most overtopping occurring for a 30 degrees angle bund wall situated at 5m.

In the experiments carried out in configuration SQ4, with the obstacles present, it was observed that water overtopped the cylindrical sheeting that formed the obstacles and flowed into the obstacles. The depth of water retained in the cylindrical obstacles could not be measured and was less than 1 mm, corresponding to a loss of less than 19 litres. Hence, the amount of overtopping recorded in the later interaction with the bund walls could have been underestimated by up to this amount. This is noted subsequently on the relevant Tables of data and Figures.

For the experiments carried out with initial water heads of approximately 1.75mm (nominally 110%), the volume of water that was released was greater than the capacity of the bund. For these experiments, two values were determined:

• Initial: This is the amount of water collected from the initial surge of water, the release being shut-off before the bund was full.

• Total: This is the total amount of water collected after the release vessel was allowed to empty without interruption.

Figure 22 shows the percentage of water overtopping the bund for the different square bund arrangements and the different nominal fill heights. The figure shows that there is more water overtopping the bund when the bund is further away from



the vessel. Also the presence of obstacles within the bunded area has little effect on the volume of water overtopping, although it should be noted that up to an additional 19 litres, or 5% of the bund capacity, could have been lost through water overtopping the obstacles in bund arrangement SQ4.

Plotting the percentage of water overtopping the bund for the circular and square bunds with the same bunded area (5.0m circular bund, SQ1 and SQ2 as one group and 10.0m circular bund and SQ3 and SQ4 as another) gives an indication of the effect of the shape of the bund. Figure 23 shows the percentage of water overtopping the bund for arrangements SQ1, SQ2 and a 5.0m circular bund, with a 90 degrees bund wall, for the different nominal fill heights. The figure shows that the percentage of water overtopping the bund is approximately the same for the different bund shapes.

Figure 24 shows the percentage of water overtopping the bund for bund arrangements SQ3, SQ4 and a 10.0m circular bund, with a 90 degrees bund wall, for the different nominal fill heights. This figure shows that for fill heights of 90% and 100% less water overtopped the circular bund than the square bunds.

4.4 Images of the Water Front Figures 25 to 27 show still images of the water front recorded at 5.0m, 7.1m and 10.0m radii. These images were downloaded from the high-speed video system that was operated at 125f.p.s with the shutter set at 1/1000th s.

5 DISCUSSION

5.1 Phase 1 Phase 1 has provided data on the spreading of water released from the base of a tank, that at a scale of 1:20, represents a 70m diameter storage tank. The release and the surrounding geometrical arrangement are of an idealized nature. This creates a flow that can be simulated in a straightforward manner by mathematical models for liquid spread over land. Hence, the results from the experiments provide an ideal dataset for the development or validation of mathematical models.

The interaction of the spreading front with a number of different bunds has been examined and the amount of water, if any, overtopping the bunds has been determined. A number of observations can be made on the data, as follows.

Firstly, as can be seen from Figure 15, the outflow from the vessel is in close agreement with expectations from theory. Also, the initial velocity of the water flowing out of the vessel is consistent as a starting value for the inferred spreading velocity data, such as that shown in Figure 17.

Within the programme, a number of experiments were carried out with the same initial head in the tank. A comparison of the spreading behaviour in these experiments gives some idea of the repeatability of the observed behaviour. It is found that there is some radial variation in the time of arrival at a given distance from the tank (of the order of 10 to 20%) in any one experiment. However, the results from



all of the experiments, taken before interaction with the bund wall, show similar behaviour. Even taking into account the observed slower rate of progress along the line of the probes, if the time of arrival is plotted against distance from the tank centre, all of the measurements lie in a band of relatively narrow width, indicating a good degree of reproducibility in flow behaviour. Figures 17 and 18 illustrate this for the results obtained with one particular initial fill level in the vessel. Results such as those shown in Figures 17 and 18 provide the data that can be used to help investigate a number of outstanding issues, such as those concerning the boundary conditions to be used at the front of the spreading pool on land.

The measurements of the amount of water overtopping the bund have also been examined. As Figure 21 demonstrates, if a large release occurs from around the base of the tank, the water is capable of overtopping the bund walls, even if the bunded area has the capacity to hold all of the water that is released. Based on these results, it appears that generally the amount that overtops the bund decreases as the distance of the bund wall from the vessel increases (note: the bund heights were chosen to ensure that the bunded area had nominally the same capacity in each case). The amount also appears to be sensitive to the bund wall angle in this case. Even allowing for the differences in capacity of the bunds for the different slopes of the bund wall, it appears that a vertical wall is more effective than a sloping bund wall at retaining the water within the bund.

5.2 Phase 2 The experiments in Phase 2 have provided data on the interaction of the spreading front on square bunds and allow a comparison of different geometry square bunds and circular bunds. Experiments were also conducted to study the effect of additional tanks within the bunded area.

As expected, for the experiments without additional obstacles, the spreading of the liquid front prior to interaction with the bund wall was the same for square and circular bunds allowing for a radial variation of 10 to 20%, see Section 5.1.

The presence of additional tanks within the bunded area interrupts the spreading of the liquid front towards the bund wall. The results show that as the initial fill height increases the liquid front reaches the region between the additional tank and the bund wall at an earlier time, see Figure 19. The presence of the additional tanks appears to channel the water between the tanks at a greater velocity as shown in Figure 20 where the time of arrival of the liquid front occurs earlier at certain parts of the bund wall than in the immediate wake of the additional tanks.

The measurements of the amount of water overtopping the bund show that with additional tanks in the bunded area the amount of water overtopping the bund is similar to when there are no additional tanks present (although there is an uncertainty of up to approximately 5% of the bund capacity introduced because some of the flow overtopped the model tanks themselves in the experiments in which the obstacles were present). It should be noted that although the bunded area is the same in the two experimental set-ups, when the obstacles are present, the bund wall is further away from the release vessel. This is to allow for the area of the additional tanks within the bund and to maintain the same area available for liquid spreading.



Phase 1 experiments have shown that the amount of water overtopping the bund reduces as the distance from the release vessel increases, so this may explain part of this result.

The measurements show that for a bund area nominally equivalent to a circle of radius 5m, the amount of water overtopping a bund is similar, independent of the bund geometry (Figure 23). However, it appears (Figure 24) that for a bund area equivalent to a circle of radius 10m, less water overtopped a circular bund compared to a square bund.

6 SUMMARY A series of 58 experiments has been carried out to study the flow of water released from a slit at the base of a cylindrical tank into a bund. During these experiments the water pressure (liquid head) in the vessel, the initial flow of water across the bund floor and the quantity of water which overtopped the bund were measured. The data collected have been summarised in this report and full copies of the data are provided with the report in electronic format.

7 REFERENCES 1. Webber, D.M., A Model for Pool Spreading and Vaporisation and its

Implementation in the Computer Code G*A*S*P”, Report SRD/HSE/R507, September 1990.

2. Linden, P., Daish, N., Dalziel, S., Halford, A., Jackson, M., Hirst, I., Perroux, J., Wiersma, S. 'LSMS: A New Model For Spills Of LNG And Other Hazardous Liquids', Proceedings of 1998 International Gas Research Conference, San Diego, November 8th - 11th.

3. Greenspan, H.P. and Johansson, A.V. An experimental study of flow over an impounding dyke. Studies in Applied Mathematics, 64, 211-233, 1981.

4. Moorhouse, J and Carpenter, R.J. Factors affecting vapour evolution rates from liquefied gas spills. I Chem E North Western Branch, Hazards Symposium, 1986.

5. Sharifi, T. An experimental study of the catastrophic failure of storage tanks. Ph D Thesis. University of London, Imperial College of Science and Technology, Dept. of Chem Eng. And Chem. Technology, 1987.



Table 1: Coordinates for Resistance Probes Using a Polar Coordinate System, see Figures 12 and 13

Coordinates Instrument Number θ (θ (θ (θ (οοοο)))) R (m)*

RP01 45 1.88 RP02 45 2.16 RP03 45 2.44 RP04 45 2.72 RP05 45 3.00 RP06 45 3.28 RP07 45 3.56 RP08 45 3.84 RP09 45 4.12 RP10 45 4.40 RP11 45 4.68 RP12 45 4.96 RP13 45 5.24 RP14 45 5.52 RP15 45 5.80 (13.06) RP16 45 6.08 (13.32) RP17 45 6.36 (10.20) RP18 45 6.64 (10.46) RP19 45 6.92 (10.72) RP20 45 7.20 (10.98) RP21 45 7.48 (11.24) RP22 45 7.76 (11.50) RP23 45 8.04 (11.76) RP24 45 8.32 (12.02) RP25 45 8.60 (12.28) RP26 45 8.88 (12.54) RP27 45 9.16 (12.80) RP28 45 9.44 RP29 45 9.72 RP30 45 9.93 RP31 22.5 1.88 RP32 22.5 2.50 RP33 22.5 3.12 RP34 22.5 3.75 RP35 22.5 4.38 RP36 22.5 5.00 RP37 22.5 5.53 RP38 22.5 6.05 RP39 22.5 6.57 RP40 22.5 7.10 RP41 22.5 7.83



RP42 22.5 8.55 RP43 22.5 9.27 RP44 22.5 10.00 RP45 11 5.00 RP46 34 5.00 RP47 56 5.00 RP48 67.5 5.00 RP49 79 5.00 RP50 7 7.10 RP51 15 7.10 RP52 30 7.10 RP53 37 7.10 RP54 52 7.10 RP55 60 7.10 RP56 67.5 7.10 RP57 75 7.10 RP58 83 7.10 RP59 11 10.00 RP60 34 10.00

* Radii in brackets are for tests in bund configuration SQ4



Table 2: Test Conditions for the Experiments Conducted in the Circular Bunds

Test Bund

Height (mm)

Bund Radius

(m)

Bund Angle

(o)

Nominal Fill (%)

Actual Fill Height

(m) Resistance Probe Data

H1 50 10.0 90 90 1.449 Yes H2 50 10.0 90 100 1.588 Yes G2 50 10.0 90 110 No data Yes

G2_1 50 10.0 90 110 1.808 No H3 50 10.0 45 90 1.428 No H4 50 10.0 45 100 1.579 No G3 50 10.0 45 110 1.797 Yes

G3_1 50 10.0 45 110 1.802 Yes H5 50 10.0 30 90 1.438 No H6 50 10.0 30 100 1.586 No G4 50 10.0 30 110 1.804 Yes

G4_1 50 10.0 30 110 1.806 No

H7 100 7.1 90 90 1.428 No H8 100 7.1 90 100 1.577 No G5 100 7.1 90 110 1.817 Yes

G5_1 100 7.1 90 110 1.813 No G5_2 100 7.1 90 110 1.814 No

H9 100 7.1 45 90 1.431 No H10 100 7.1 45 100 1.593 No G6 100 7.1 45 110 1.811 Yes

G6_1 100 7.1 45 110 1.806 No H11 100 7.1 30 90 1.437 No H12 100 7.1 30 100 1.592 No G7 100 7.1 30 110 1.808 Yes

G7_1 100 7.1 30 110 1.806 No

H13 200 5.0 90 90 1.442 No H14 200 5.0 90 100 1.593 No G8 200 5.0 90 110 1.758 Yes

G8_1 200 5.0 90 110 1.808 No H15 200 5.0 45 90 1.439 No H16 200 5.0 45 100 1.585 No G9 200 5.0 45 110 1.751 Yes

G9_1 200 5.0 45 110 1.788 Yes H17 200 5.0 30 90 1.441 No H18 200 5.0 30 100 1.588 No G10 200 5.0 30 110 1.819 Yes

G10_1 200 5.0 30 110 1.753 No



Table 3: Test Conditions for the Experiments Conducted in the Square Bunds

Test Bund

Height (mm)

Bund Configuration

Bund Angle

(o)

Nominal Fill (%)

Actual Fill Height

(m) Resistance Probe Data

J1 200 SQ1 90 90 1.430 Yes J2 200 SQ1 90 100 1.585 Yes J3 200 SQ1 90 110 1.733 Yes

J3_1 200 SQ1 90 110 1.732 No

J4 200 SQ2 90 90 1.437 Yes J5 200 SQ2 90 100 1.586 Yes J6 200 SQ2 90 110 1.730 No

J6_1 200 SQ2 90 110 1.734 Yes

J7 50 SQ3 90 90 1.439 Yes J8 50 SQ3 90 100 1.588 Yes J9 50 SQ3 90 110 1.731 Yes

J9_1 50 SQ3 90 110 1.730 No

J10 50 SQ4 90 90 1.436 Yes J10_1 50 SQ4 90 90 1.447 Yes J10_2 50 SQ4 90 90 1.438 Yes

J11 50 SQ4 90 100 1.593 Yes J11_1 50 SQ4 90 100 1.587 Yes J11_2 50 SQ4 90 100 1.638 Yes

J12 50 SQ4 90 110 1.716 No J12_1 50 SQ4 90 110 1.740 Yes J12_2 50 SQ4 90 110 1.738 No



Table 4: Release Vessel Water Inventory

Fill Height/Initial Head Water Inventory

1.45m 3.50m3

1.60m 3.86m3

1.75m 4.22m3

Table 5a: Volume of the Circular Bunds in Phase 1

Volume of Bund (m3) Nominal Radius (m) 30o bund angle 45o bund angle 90o bund angle

5.0 3.66 3.77 3.93

7.1 3.86 3.91 3.96

10.0 3.89 3.93 3.93

Table 5b: Volume of the Square Bunds in Phase 2

Bund Configuration Volume of Bund (m3)

SQ1 3.88

SQ2 3.88

SQ3 3.90

SQ4 3.88



Table 6: Time of Water Arrival for Experiments with a 10.0m Circular Bund Probe Test H1

(ms) Test H2

(ms) Test G2

(ms) Test G3

(ms) Test G4

(ms) RP01 0 0 10.8 3.1 - RP02 69.7 62.5 108.9 96 79.4 RP03 180.8 138.7 190.6 160.6 146.1 RP04 248.9 257.7 242.2 221.2 216.7 RP05 256.9 286.9 283.6 288.9 282.4 RP06 404.7 373.3 340.4 343.4 331.6 RP07 541.3 480.8 406.9 413.1 412.9 RP08 643.7 592.8 459.5 536.7 523.8 RP09 728.3 696.6 596.4 670.1 597.9 RP10 829.8 780.3 627.4 735.6 725.4 RP11 931.4 868.9 812.5 803.4 786.3 RP12 1033.1 997.1 887.6 867.9 898.4 RP13 1159.8 1126.6 972 979.6 971.8 RP14 1284.7 1218.7 1076.4 1056.7 1056 RP15 1422.3 1339.4 1183.3 1181.6 1150.5 RP16 1559 1475.9 1270.1 1276.8 1263 RP17 1704 1606.2 1337.3 1345.4 1339.1 RP18 1863.5 1778.7 1429.8 1430.2 1431.8 RP19 2030.5 1923.9 1548.5 1542.2 1552.9 RP20 2428.3 2271.5 1824.7 1884.9 1832.7 RP21 2652.8 2483.5 1998.6 2064.1 2032.7 RP22 2874.4 2653.6 2180.2 2223.9 2200.3 RP23 3081.7 2842.4 2360.7 2376.3 2367.4 RP24 3265.5 3091.4 2550.8 2579.4 2543.5 RP25 3474.2 3339.4 - - 2476.2 RP26 3675.5 3491.5 2960.3 3004.4 2939.7 RP27 4070.1 3816.2 3157.1 3183.1 3142 RP28 4384.3 4079 3385.7 3400.2 3350.4 RP29 5050.7 4439 3580.1 3574.2 3558.8 RP30 5029.6 4641.1 3734 3754.7 3721.9 RP31 23.4 7.7 0 0 0 RP32 144.3 145.5 174.9 164 156.9 RP33 309.5 306.4 284.5 273.9 275.9 RP34 495.7 496.1 417.1 435.6 416.6 RP35 725.2 780.7 615.4 609.6 619.2 RP36 909.2 850.2 792.2 796.2 819.8 RP37 1101.2 1040.8 1003.5 990 968.7 RP38 1351.5 1276.4 1152.1 1150 1127.7 RP39 1565.2 1509.7 1325.4 1331.9 1321.5 RP40 1934.9 1853.1 1526.1 1552.2 1532.2 RP41 2720.4 2519.6 2090.5 2114 2053.9 RP42 3325.1 3092.2 2477.7 2488.2 2425.5 RP43 4387.9 3670.2 2940.9 2938.1 2900.5 RP44 4670 4336.3 3217.2 3226.7 3167.1 RP45 908 849.1 731.9 759.1 778.6 RP46 951.6 875.8 873.8 857.9 863 RP47 977.3 917.9 870.9 843.4 845.3 RP48 951.3 862.8 875.2 889.7 840.5 RP49 866.2 875.5 686.9 784.5 730.2 RP50 2019.2 1859.8 1512.5 1499.8 1499.4 RP51 2038.7 1889.3 1546.2 1525.7 1507.3 RP52 1922.4 1883.5 1584.4 1591.5 1535 RP53 2029.1 1915.8 1580.6 1578 1568.8 RP54 1884.1 1780.3 1586.5 1571.4 1531.8 RP55 2040.6 1864.6 1632.1 1621.3 1641.1 RP56 2073.1 1909.6 1651 1667.6 1621.4 RP57 1887.2 1714.3 1590.7 1587.2 1559.6 RP58 1836.2 1664.3 1542.3 1547.4 1512 RP59 4983.7 4509.6 3378.4 3367.9 3380.5 RP60 4895.1 4222.5 3350.6 3272.9 3326.4

‘-‘ means no data were collected by that resistance probe.



Table 7 Time of Water Arrival for Experiments with a 7.1m Circular Bund Probe Test G5

(ms) Test G6

(ms) Test G7

(ms) RP01 3.8 0 0 RP02 87.8 107.7 87.7 RP03 180.5 197 159.4 RP04 234.6 261.1 214.4 RP05 286.6 310.5 270.9 RP06 351.4 395.8 353.7 RP07 460.7 468.1 451.1 RP08 490.2 613.3 539.8 RP09 667.3 707.4 658.7 RP10 724 774.1 729.5 RP11 788.8 839.3 817.8 RP12 822.1 933.8 851 RP13 961.1 1023.3 988.1 RP14 - 1112.7 1069 RP15 1177.5 1213.2 1162.9 RP16 1260.4 1279.7 1236.1 RP17 1354.8 1353.8 1327.1 RP18 1451.4 1451.1 1440.9 RP19 1566.9 1566.1 1542.3 RP20 - - - RP21 - - - RP22 - - - RP23 - - - RP24 - - - RP25 - - - RP26 - - - RP27 - - - RP28 - - - RP29 - - - RP30 - - - RP31 0 42 0 RP32 163.1 203.4 160.7 RP33 267.1 312.8 268.6 RP34 418.6 462.4 430.3 RP35 630.3 649.7 637.2 RP36 844.8 874.7 808.8 RP37 994.1 1045.2 994.7 RP38 1157.5 1209.1 1143.1 RP39 1319.7 1366.1 1329 RP40 1515.1 1579 1513.1 RP41 - - - RP42 - - - RP43 - - - RP44 - - - RP45 776.9 670.2 778.8 RP46 855.3 903.2 850.9 RP47 859 895.9 860.3 RP48 834.1 890.8 869.7 RP49 692.1 806.8 652.3 RP50 1501.8 1538.3 1495.1 RP51 1530.9 1539.1 1504.3 RP52 1561.1 1613.6 1571.1 RP53 1577.7 1605.2 1591.4 RP54 1565.7 1566.2 1532.4 RP55 1599.2 - 1597.2 RP56 1637.2 - 1613.7 RP57 1521.8 1616.6 1490.7 RP58 1494 1568.9 1498.4 RP59 - - - RP60 - - -




Table 8: Time of Water Arrival for Experiments with a 5.0m Circular Bund

Probe Test G8 (ms)

Test G9 (ms)

Test G9_1 (ms)

Test G10 (ms)

RP01 6.5 0 0 0 RP02 84.6 60.1 74.3 79.7 RP03 162 149.2 154.1 193.1 RP04 223.3 233.8 211.1 265.3 RP05 269.6 256.6 283.2 309 RP06 332.6 323.7 347.4 362 RP07 415.1 419.7 462.6 481.9 RP08 486.3 527.4 494.6 600.1 RP09 518.2 669.8 561.9 699.6 RP10 584.8 731.5 738.5 773.5 RP11 799.2 790.3 801.6 824.3 RP12 920.9 954.1 935.3 929.8 RP13 - - - - RP14 - - - - RP15 - - - - RP16 - - - - RP17 - - - - RP18 - - - - RP19 - - - - RP20 - - - - RP21 - - - - RP22 - - - - RP23 - - - - RP24 - - - - RP25 - - - - RP26 - - - - RP27 - - - - RP28 - - - - RP29 - - - - RP30 - - - - RP31 0 0 0 24.6 RP32 169 150.2 158.9 191 RP33 275.7 257.7 270.8 304.4 RP34 427.5 415.8 416.5 437.5 RP35 606.8 615.2 549.4 RP36 822.2 806 807.8 839.6 RP37 - - - - RP38 - - - - RP39 - - - - RP40 - - - - RP41 - - - - RP42 - - - - RP43 - - - - RP44 - - - - RP45 783.7 772.1 746.9 795.8 RP46 - 840.4 843.3 868.2 RP47 845.4 851.5 829.6 868.8 RP48 846.6 844.3 851.4 873.8 RP49 698.4 735.9 - - RP50 - - - - RP51 - - - - RP52 - - - - RP53 - - - - RP54 - - - - RP55 - - - - RP56 - - - - RP57 - - - - RP58 - - - - RP59 - - - - RP60 - - - -


R

epor

t Num

ber:

R 4

418

Issu

e:01

C

OM

MER

CIA

L IN

CO

NFI

DEN

CE

Pa

ge 1

9 Ta

ble

9: T

ime

of W

ater

Arr

ival

for E

xper

imen

ts w

ith S

quar

e B

unds

Test

s Pr

obe

J1

(ms)

J2

(m

s)

J3

(ms)

J4

(m

s)

J5

(ms)

J6

_1

(ms)

J7

(m

s)

J8

(ms)

J9

(m

s)

J10

(ms)

J1

0_1

(ms)

J1

0_2

(ms)

J1

1 (m

s)

J11_

1 (m

s)

J11_

2 (m

s)

J12_

1 (m

s)

RP0

1 0

- 0

0 0

0 0

0 0

0 -

0 0

0 0

0 R

P02

59

118.

9 72

.7

70

87.6

90

.9

74.7

95

97

.2

69.1

90

10

2 80

.6

88

96

61

RP0

3 15

2.5

189.

6 13

4.7

137.

9 16

9.4

235.

2 17

5.4

215.

9 20

8.3

170.

9 18

4 17

1 17

2.3

178

208

163

RP0

4 18

8.1

240.

9 20

7.5

188.

3 19

4.2

220.

4 32

0.7

237.

1 26

9.2

293.

5 28

0 24

1 28

3.1

258

262

221

RP0

5 27

3.4

327.

6 28

6.6

326.

6 31

1.3

304

364.

5 31

4.2

323.

9 33

2.5

333

397

312.

3 30

7 31

8 28

4 R

P06

397.

5 38

5.3

369.

2 39

8.3

377.

8 35

9.8

470.

6 38

6.8

384

403.

1 40

4 45

9 36

6.1

355

443

336

RP0

7 55

3.3

464.

8 46

5.6

541.

1 48

2.8

445.

2 60

2.4

558.

6 47

9.7

561

503

572

520.

9 46

6 59

8 44

8 R

P08

626.

2 55

6.5

643

646.

9 62

2.7

473.

9 68

6.5

663.

2 59

7.6

653.

2 55

7 69

7 63

1.9

605

662

555

RP0

9 72

3.1

704.

1 70

9.6

730.

1 70

5.6

577.

1 78

0.7

723.

4 70

0.1

736.

7 73

8 78

2 69

8.1

714

744

651

RP1

0 -

- -

844.

5 80

4.3

778.

3 91

5.7

814.

6 78

5.3

830.

5 83

7 88

0 78

3.7

791

865

778

RP1

1 -

- -

1091

.4

899.

1 85

0.3

1009

.2

934.

6 90

8.3

944.

8 96

0 99

7 91

3.4

924

955

845

RP1

2 -

- -

1032

.1

1020

.4

963.

9 10

67.9

99

2 99

3.1

1031

.5

1068

11

30

1067

.2

985

1022

92

9 R

P13

- -

- 11

74.7

11

15.3

11

02.1

12

31.4

11

54.5

11

32.4

11

82.1

12

05

1233

11

83.7

11

34

1174

10

92

RP1

4 -

- -

1308

.2

1230

.2

1112

.5

1364

.5

1282

.1

1288

.7

1284

.7

13`3

3 13

43

1296

.7

1269

12

76

1174

R

P15

- -

- 14

36.1

13

31.8

13

23.7

14

35.9

13

75.5

13

80.1

-

- 10

567

- -

1005

8 97

73

RP1

6 -

- -

1553

.6

1445

.7

1429

.9

1573

.6

1534

14

89.3

-

- 10

820

- -

1034

6 10

029

RP1

7 -

- -

- -

- 17

44.3

16

59.2

15

81.2

-

- 11

774

- -

1135

5 10

967

RP1

8 -

- -

- -

- 19

12.8

18

09.8

17

83.5

-

- 12

725

- -

1137

5 11

125

RP1

9 -

- -

- -

- 20

66.4

20

08.7

19

49

- -

1132

7 -

- 10

830

1075

8 R

P20

- -

- -

- -

2503

.1

2338

.3

2273

.7

- -

1104

3 -

- 10

659

1046

2 R

P21

- -

- -

- -

2746

.2

2591

.9

2500

.4

- -

1063

1 -

- 10

001

1060

7 R

P22

- -

- -

- -

2958

.1

2818

.8

2700

.2

- -

1038

2 -

- 96

80

9552

R

P23

- -

- -

- -

3175

.5

3013

.4

2895

.5

- -

1028

8 -

- 97

56

9437

R

P24

- -

- -

- -

3377

.6

3180

.8

3082

.3

- -

1046

9 -

- 99

13

- R

P25

- -

- -

- -

3603

.3

3403

.3

3274

.9

- -

1043

0 -

- 99

06

9679

R

P26

- -

- -

- -

3838

.4

3632

.9

3508

-

- 10

276

- -

9834

96

18

RP2

7 -

- -

- -

- 40

41.6

39

13.2

37

30.9

-

- 10

472

- -

1003

6 95

16

RP2

8 -

- -

- -

- 40

41.6

42

08.5

40

31

- -

1340

8 -

- 13

467

1289

1 R

P29

- -

- -

- -

- 45

65.1

43

34.2

-

- 12

287

- -

1218

9 11

784

RP3

0 -

- -

- -

- -

4848

.8

4564

.1

- -

1195

5 -

- 11

427

1144

1 R

P31

- -

- 22

.1

19.7

13

12

.4

31.4

38

.3

13.9

37

55

22

.1

39

57

- R

P32

- -

- 15

5.1

152.

1 15

4.7

178.

8 17

2.2

162.

3 14

6.3

168

178

149

168

193

- R

P33

- -

- 32

6 31

3.5

278.

6 33

5.9

335.

8 30

8.7

319.

1 32

0 34

8 30

3.9

301

317

- R

P34

- -

- 53

4 49

6.3

462.

4 52

4.2

490.

2 45

5.1

483.

7 49

5 51

6 47

9.3

474

514

- R

P35

- -

- 71

7.1

663.

3 59

8.9

- 71

6.5

651.

5 69

6.9

740

744

650.

5 65

2 75

5 -

RP3

6 -

- -

- -

- 91

2.1

894.

7 83

2.1

887.

5 90

7 95

1 88

5.3

891

939

- R

P37

- -

- -

- -

1100

.9

1055

10

29.4

10

72

1115

11

74

1081

.5

1048

11

33

-

R

epor

t Num

ber:

R 4

418

Issu

e:01

C

OM

MER

CIA

L IN

CO

NFI

DEN

CE

Pa

ge 2

0

RP3

8 -

- -

- -

- 13

39.4

18

99

1225

.2

1335

.4

1365

15

87

1312

.9

1339

13

88

- R

P39

- -

- -

- -

1640

.3

1585

.8

1479

.4

1629

.4

1643

17

33

1551

.3

1499

16

90

- R

P40

- -

- -

- -

- 19

56.8

17

62.7

19

50.5

19

67

2088

18

59.8

17

98

1923

-

RP4

1 -

- -

- -

- 28

28.7

27

89.9

26

11.3

28

01.4

28

12

2906

26

51.8

25

84

2799

23

97

RP4

2 -

- -

- -

- 33

19.9

32

14.2

30

44

2998

.8

2997

31

08

2862

.2

2836

30

47

2686

R

P43

- -

- -

- -

- 37

91

3570

.2

3329

.2

3391

34

84

3216

.6

3150

33

91

3015

R

P44

- -

- -

- -

4731

.9

4571

47

39.1

37

48.5

37

77

3883

35

99.3

35

08

3756

33

19

RP4

5 -

- -

665.

1 70

0.7

635.

1 -

884.

6 84

6.2

2282

90

03

1369

7 -

9124

14

261

1024

2 R

P46

- -

- 79

2 75

9.3

755.

6 10

42.9

95

4.5

932.

9 10

08.9

99

0 10

95

930.

9 91

1 10

26

852

RP4

7 -

- -

808.

9 63

8.8

735.

8 10

03.2

94

8.7

906.

6 97

2.7

994

982

914.

6 90

7 10

02

844

RP4

8 -

- -

743.

3 73

7.8

720.

1 98

6.4

955

870.

9 94

3.9

981

958

896.

9 89

2 96

2 83

1 R

P49

- -

- 69

2.6

614.

7 56

5.5

916

863.

2 85

9.5

2282

.1

9003

13

697

- 91

24

1426

0 10

277

RP5

0 -

- -

- -

- 20

42.5

20

05.3

18

38.5

12

4 63

66

6709

-

6916

72

42

6809

R

P51

- -

- -

- -

- 19

75.4

19

27.5

45

58.7

46

80

4780

47

87.7

45

67

4562

44

82

RP5

2 -

- -

- -

- 20

59.5

20

11.9

18

93

1996

.5

2008

21

54

1917

19

25

2022

18

07

RP5

3 -

- -

- -

- 20

57.8

20

00.6

19

09.3

22

84.3

90

05

1370

0 -

9127

14

264

1031

0 R

P54

- -

- -

- -

1945

18

73.6

17

77.9

-

9005

13

700

- 91

27

1426

4 10

310

RP5

5 -

- -

- -

- 20

67.5

19

54.6

18

78.5

19

02.5

19

14

1930

18

33.8

18

35

1922

17

56

RP5

6 -

- -

- -

- 20

30.3

19

41

1893

.6

1884

.8

2060

19

86

1777

.9

1817

18

64

1681

R

P57

- -

- -

- -

1885

.2

1781

.1

1713

.2

5150

.3

4909

51

37

2955

.6

3104

37

84

2705

R

P58

- -

- -

- -

1878

.5

1756

.6

1688

.9

322.

1 65

67

7689

-

6829

74

22

7389

R

P59

- -

- -

- -

- 35

88.8

-

- 55

16

5758

-

5138

54

69

5126

R

P60

- -

- -

- -

- 44

39.2

42

38.5

93

1.4

8681

80

35

- 65

90

7866

84

26

‘-‘ m

eans

no

data

wer

e co

llect

ed b

y th

at re

sist

ance

pro

be.



Table 10: Results of the Volume of Water Overtopping the Bund in the Experiments Conducted in the Circular Bunds

* Nominal inventory quoted, as no fill height data available for this test

Test Bund

Height (mm)

Bund Radius

(m)

Bund Angle

(o)

Tank Inventory

(cu.m) (from Tables

2 and 4)

Bund Volume (cu.m) (from

Table 5a)

Water Overtopping

(litres)

Initial Wave or

Total Amount

H1 50 10.0 90 3.50 3.93 None Total H2 50 10.0 90 3.83 3.93 55 Total G2 50 10.0 90 4.22* 3.93 443 Total

G2_1 50 10.0 90 4.36 3.93 476 Initial H3 50 10.0 45 3.45 3.93 None Total H4 50 10.0 45 3.81 3.93 39 Total G3 50 10.0 45 4.33 3.93 323 Total

G3_1 50 10.0 45 4.34 3.93 323 Initial H5 50 10.0 30 3.47 3.89 7 Total H6 50 10.0 30 3.83 3.89 67 Total G4 50 10.0 30 4.35 3.89 468 Total

G4_1 50 10.0 30 4.35 3.89 419 Initial

H7 100 7.1 90 3.45 3.96 37 Total H8 100 7.1 90 3.80 3.96 160 Total G5 100 7.1 90 4.38 3.96 545 Total

G5_1 100 7.1 90 4.37 3.96 654 Total G5_2 100 7.1 90 4.37 3.96 272 Initial



G7_1 100 7.1 30 4.35 3.86 511 Initial




G10_1 200 5.0 30 4.23 3.66 827 Initial



Table 11: Results of the Volume of Water Overtopping the Bund in the Experiments Conducted in the Square Bund Configurations

Test Bund

Configuration

Bund Height (mm)

Bund Angle

(o)

Tank Inventory

(cu.m) (from Tables

2 and 4)

Bund Volume (cu.m)

(from Table 5b)

Water Overtopping

(litres)

Initial Wave or Total Amount

J1 SQ1 200 90 3.45 3.88 19 Total J2 SQ1 200 90 3.82 3.88 55 Total J3 SQ1 200 90 4.18 3.88 170 Total

J3_1 SQ1 200 90 4.18 3.88 21 Initial

J4 SQ2 200 90 3.47 3.88 25 Total J5 SQ2 200 90 3.83 3.88 59 Total J6 SQ2 200 90 4.17 3.88 38 Initial

J6_1 SQ2 200 90 4.18 3.88 249 Total

J7 SQ3 50 90 3.47 3.90 175 Total J8 SQ3 50 90 3.83 3.90 286 Total J9 SQ3 50 90 4.17 3.90 235 Initial

J9_1 SQ3 50 90 4.17 3.90 485 Total

J10 SQ4* 50 90 3.47 3.88 160 Total J10_1 SQ4* 50 90 3.49 3.88 120 Total J10_2 SQ4* 50 90 3.47 3.88 125 Total J11 SQ4* 50 90 3.84 3.88 225 Total

J11_1 SQ4* 50 90 3.83 3.88 294 Total J11_2 SQ4* 50 90 3.95 3.88 292 Total J12 SQ4* 50 90 4.14 3.88 443 Total

J12_1 SQ4* 50 90 4.20 3.88 445 Total J12_2 SQ4* 50 90 4.19 3.88 63 Initial * In experiments carried out in configuration SQ4, up to approximately 19 litres of water may have overtopped and flowed into the model obstacles, and hence have not been available to overtop the constraining bund walls.



Figure 1: Photograph of the Release Vessel



Figure 2: Schematic Diagram of the Release Vessel

Flap valve opening mechanism

Water tank

Water tank

2.6m

1.1m

1.8m

Steel sheet

Elevation

Plan

Concrete pad

Concrete pad

1.75m



Figure 3: Photograph of the Release Mechanism

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

26

Fi

gure

4: B

und

Arra

ngem

ent f

or a

Circ

ular

Bun

d w

ith a

Dia

met

er o

f 5.0

m

5.0m

Rel

ease

Ve

ssel

Wal

l of

sym

met

ry

3.5

Dia

200m

m

200m

m

45 d

egre

es

200m

m

30 d

egre

es

Bund

Wal

l Geo

met

ries

Bund

wal

l

Plan

of B

und

Layo

ut

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

27

Fi

gure

5: B

und

Arra

ngem

ent f

or a

Circ

ular

Bun

d w

ith a

Dia

met

er o

f 7.1

m

7.1m

Rel

ease

Ve

ssel

Wal

l of

sym

met

ry

3.5

Dia

100m

m

100m

m

45 d

egre

es

100m

m

30 d

egre

es

Bund

Wal

l Geo

met

ries

Bund

wal

l

Plan

of B

und

Layo

ut

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

28

Fi

gure

6: B

und

Arra

ngem

ent f

or a

Circ

ular

Bun

d w

ith a

Dia

met

er o

f 10.

0m

10.0

mR

elea

se

Ves

sel

Wal

l of

sym

met

ry

3.5m

Dia

50m

m

45 d

egre

es

50m

m

30 d

egre

es

Bun

d W

all G

eom

etrie

s

Bund

wal

l

Plan

of B

und

Layo

ut

50m

m

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

29

Fi

gure

7: B

und

Arra

ngem

ent f

or S

quar

e B

und

Arra

ngem

ent S

Q1

6.27

m

200m

m v

ertic

albu

nd w

all

Rel

ease

Ve

ssel

Wal

l of

sym

met

ry

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

30

Fi

gure

8: B

und

Arra

ngem

ent f

or S

quar

e B

und

Arra

ngem

ent S

Q2

4.43

m

200m

m v

ertic

albu

nd w

all

Rel

ease

Ve

ssel

Wal

l of

sym

met

ry

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

31

Fi

gure

9: B

und

Arra

ngem

ent f

or S

quar

e B

und

Arra

ngem

ent S

Q3

8.89

m

50m

m v

ertic

albu

nd w

all

Rel

ease

Ve

ssel

Wal

l of

sym

met

ry

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

32

Fi

gure

10:

Bun

d Ar

rang

emen

t for

Squ

are

Bun

d Ar

rang

emen

t SQ

4

3.5m

9.89

m

5.3m

50m

m v

ertic

albu

nd w

all

Rel

ease

Ve

ssel

Wal

l of

sym

met

ry

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

33

Fi

gure

11:

Pho

togr

aph

of th

e Te

st R

ig fo

r Bun

d Ar

rang

emen

t SQ

4

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

34

Fi

gure

12:

Loc

atio

n of

Res

ista

nce

Prob

es fo

r All

the

Bun

d Ar

rang

emen

ts W

ithou

t Obs

tacl

es W

ithin

the

Bun

ded

Area

wate

r tan

k

10m

7.1m

5m

orig

in

O

rCoor

dina

te S

yste

m θ=

0θ

5857

5655

54

4948

47

46

53

52

3132

3334

3536

3738

3940

4142

43

44

45

51 50

59

60

Probes

1-30 A

long G

antry

Not

e: S

ee T

able

1 fo

r pr

obe

pola

r co-

ordi

nate

s

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

35

Fi

gure

13:

Loc

atio

n of

Res

ista

nce

Prob

es fo

r the

Bun

d Ar

rang

emen

t With

Obs

tacl

es in

the

Bun

ded

Area

w

ater

tank

9.93

m7.

1m5m

orig

in

O

r

Coor

dinat

e Sy

stem θ

= 0

θ58

5756

55

54

4948

47

46

53

52

3132

3334

3536

3738

3940

4142

43

44

45

51 50

59

60

Probe

s 1-30

Alon

g Gan

try

Not

e: S

ee T

able

1 fo

r pr

obe

pola

r co-

ordi

nate

s

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

36

Fi

gure

14:

Dia

gram

Sho

win

g th

e Se

t-up

of th

e H

i Spe

ed V

ideo

Cam

era

Gra

duat

ed s

teel

rule

rH

igh

spee

d vi

deo

cam

era

Gan

try

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

37

Figu

re 1

5: V

aria

tion

of M

easu

red

and

Cal

cula

ted

Hei

ght o

f Wat

er R

emai

ning

in th

e Ve

ssel

with

Tim

e fo

r Diff

eren

t Nom

inal

Fi

ll H

eigh

ts

0.00

0.25

0.50

0.75

1.00

1.25

1.50

1.75

2.00

05

1015

2025

Tim

e (s

)

Head (m)

Fill

Hei

ght 1

.45m

Fill

Hei

ght 1

.60m

Fill

Hei

ght 1

.75m

Fill

Hei

ght 1

.45m

cal

cula

ted

Fill

Hei

ght 1

.60m

cal

cula

ted

Fill

Hei

ght 1

.75m

cal

cula

ted

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

38

Fi

gure

16:

Var

iatio

n of

Wat

er F

low

Rat

e In

ferr

ed fr

om H

eigh

t of W

ater

in th

e Ve

ssel

with

Tim

e fo

r Diff

eren

t Nom

inal

Fill

H

eigh

ts

050100

150

200

250

300

350

05

1015

2025

3

Tim

e (s

)

Water Flow Rate (litres/s)

Fill H

eigh

t 1.4

5m

Fill H

eigh

t 1.6

0mFi

ll Hei

ght 1

.75m

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

39

Figu

re 1

7: P

lot o

f Tim

e of

Wat

er A

rriv

al a

t the

Res

ista

nce

Prob

es in

a 1

0m C

ircul

ar B

und

for

Expe

rimen

ts w

ith D

iffer

ent

Nom

inal

Fill

Hei

ghts

(Not

e: P

robe

s w

ere

loca

ted

at 5

m, 7

.1m

and

10m

to te

st fo

r rad

ial s

ymm

etry

of t

he fl

ow.

The

spre

ad

of th

e da

ta re

cord

ed a

t the

se p

oint

s gi

ves

som

e in

dica

tion

of th

e va

riabi

lity

and

effe

ct o

f the

pro

bes

them

selv

es.)

0

1000

2000

3000

4000

5000

6000

02

46

810

12

Radi

us (m

)

Time (ms)

Fill

Hei

ght 1

.45m

Fill

Hei

ght 1

.60m

Fill

Hei

ght 1

.75m

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

40

Figu

re 1

8:

Com

paris

on o

f Fr

onta

l Spe

ed In

ferr

ed f

rom

Res

ista

nce

Prob

e D

ata

Alon

g a

Rad

ial L

ine

in T

est

G3

with

the

Va

lues

Infe

rred

from

a C

urve

Fit

to a

ll th

e D

ata

Col

lect

ed in

Sim

ilar E

xper

imen

ts.

Spee

d of

Pro

paga

tion

of F

ront

as a

Fun

ctio

n of

Dis

tanc

e

012345678

02

46

810

12

Dis

tanc

e (m

)

Front speed, m/s

Test

G3

Infe

rred

from

cur

vefit

tosp

read

ing

data

from

sim

ilar t

ests

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

41

Fi

gure

19:

Plo

t of

Tim

e of

Wat

er A

rriv

al a

t th

e R

esis

tanc

e Pr

obes

Alo

ng t

he 4

5 D

egre

e R

adiu

s in

a B

und

Arra

ngem

ents

SQ

3 an

d SQ

4 fo

r Exp

erim

ents

with

Diff

eren

t Nom

inal

Fill

Hei

ghts

0

2000

4000

6000

8000

1000

0

1200

0

1400

0

1600

0 0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

Rad

ius

(m)

Time (ms)

SQ 4

, Fill

Hei

ght 1

.45m

SQ 4

, Fill

Hei

ght 1

.60m

SQ 4

, Fill

Hei

ght 1

.75m

SQ 3

, Fill

Hei

ght 1

.45m

SQ 3

, Fill

Hei

ght 1

.60m

SQ 3

, Fill

Hei

ght 1

.75m

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

42

Figu

re 2

0: P

lot o

f Tim

e of

Wat

er A

rriv

al a

t the

Res

ista

nce

Prob

es A

long

the

22.5

Deg

ree

Rad

ius

in a

Bun

d Ar

rang

emen

ts

SQ3

and

SQ4

for E

xper

imen

ts w

ith D

iffer

ent N

omin

al F

ill H

eigh

ts

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

02

46

810

1

Rad

ius

(m)

Time (ms)

SQ

4 Fi

ll H

eigh

t 1.4

5m S

Q4

Fill

Hei

ght 1

.60m

SQ

4 Fi

ll H

eigh

t 1.7

5m S

Q3

Fill

Hei

ght 1

.45m

SQ

3 Fi

ll H

eigh

t 1.6

0m S

Q3

Fill

Hei

ght 1

.75m

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

43

Figu

re 2

1: C

ompa

rison

of V

olum

e of

Wat

er O

vert

oppi

ng th

e B

und

for C

ircul

ar B

unds

for D

iffer

ent A

ngle

s of

the

Bun

d W

all

(Cas

es s

how

n ha

ve in

itial

nom

inal

fill

heig

ht o

f app

roxi

mat

ely

90%

of t

he c

apac

ity o

f the

90

degr

ees

bund

wal

l cas

e).

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

010

2030

4050

6070

8090

100

Faci

ng a

ngle

of b

und

wal

l (de

gree

s)

Percentage overtopping (%)

bund

radi

us 1

0.0m

bund

radi

us 7

.1m

bund

radi

us 5

.0m

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

44

Fi

gure

22:

Com

paris

on o

f Vo

lum

e of

Wat

er O

vert

oppi

ng t

he B

und

for

Squa

re B

unds

for

Diff

eren

t N

omin

al In

itial

Wat

er

Dep

ths

in th

e Ve

ssel

0.0

2.0

4.0

6.0

8.0

10.0

12.0

1.4

1.45

1.5

1.55

1.6

1.65

1.7

1.75

1.8

Vess

el F

ill H

eigh

t (m

)


Arra

ngem

ent S

Q1

Arra

ngem

ent S

Q2

Arra

ngem

ent S

Q3

Arra

ngem

ent S

Q4

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

45

Fi

gure

23:

Per

cent

age

of W

ater

Ove

rtop

ping

the

Bun

d fo

r th

e Sa

me

Size

Bun

ded

Area

s fo

r D

iffer

ent S

hape

d B

unds

and

D

iffer

ent N

omin

al F

ill H

eigh

ts

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

1.4

1.45

1.5

1.55

1.6

1.65

1.7

1.75

1.8

Vess

el F

ill H

eigh

t (m

)


Arra

ngem

ent S

Q1

Arra

ngem

ent S

Q2

5.0m

circ

ular b

und

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

46

Figu

re 2

4: P

erce

ntag

e of

Wat

er O

vert

oppi

ng th

e B

und

for

the

Sam

e Si

ze B

unde

d Ar

eas

for

Diff

eren

t Sha

ped

Bun

ds a

nd

Diff

eren

t Nom

inal

Fill

Hei

ghts

0.0

2.0

4.0

6.0

8.0

10.0

12.0

1.4

1.45

1.5

1.55

1.6

1.65

1.7

1.75

1.8

Vess

el F

ill H

eigh

t (m

)

Arra

ngem

ent S

Q3

Arra

ngem

ent S

Q4

10.0

m c

ircul

ar b

und

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

47

Figu

re 2

5: V

ideo

Imag

e of

the

Wat

er F

ront

5.0

m fr

om th

e R

elea

se V

esse

l

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

48

Fi

gure

26:

Vid

eo Im

age

of th

e W

ater

Fro

nt 7

.1m

from

the

Rel

ease

Ves

sel

R

epor

t Num

ber:

R 4

418

Issu

e: 0

1

CO

MM

ERC

IAL

IN C

ON

FID

ENC

E

Page

49

Fi

gure

27:

Vid

eo Im

age

of th

e W

ater

Fro

nt 1

0.0m

from

the

Rel

ease

Ves

sel

Printed and published by the Health and Safety ExecutiveC30 1/98

Printed and published by the Health and Safety ExecutiveC1.25 01/02

CRR 405

£20.00 9 780717 622559

ISBN 0-7176-2255-X

contract research report 405/2002 · the experiments studied the flow of liquid across a bund floor...

Documents