Proceedings of The Twelfth (2002) International Offshore and Polar Engineering Conference Kitakyushu, Japan, May 26 31, 2002

Copyright 2002 by The International Society of Offshore and Polar Engineers ISBN 1-880653-58-3 (Set); ISSN 1098-6189 (Set)

Application of a Small Beam Centrifuge in Offshore Foundation Engineering

H.G.B. Allersma
Delft University of Technology Delft, The Netherlands

ABSTRACT Since geotechnical centrifuges have become a widely accepted tool for investigating soil structures in scale model tests, there has been a trend to increase the size of the devices. The reasons to increase the size of the centrifuges are that more details can be modeled in the samples and that more space is available to install sensors and intelligent actuators. However, as a result of miniaturization of electronics and the development of new measuring techniques, small centrifuges have become also a serious tool for performing model tests. The advantages of a small centrifuge are the lower costs of operation, the much shorter time between conceptualization and test results and the relatively large number of tests which can be performed in a project. A small beam centrifuge of 12 g-ton has been developed at the Geotechnical Laboratory of the University of Delft. The dimensions have been chosen in such a way that it is believed that an optimal ratio is obtained between capabilities and convenience of operation (soil samples and tools can be carried by one person). Several topics can been tested under which several problems related to offshore engineering such as; sliding behavior of spudcan footings, buckling of large diameter piles during driving, gas blowouts, installation and bearing capacity of suction piles and cyclic loading buried pipes.

a small centrifuge is quite adequate. By making an optimal choice between size and facilities and using up-to-date electronics and computer control, advanced tests can also be performed in a small centrifuge. A small device is inexpensive to operate, and the development of the equipment was much faster compared to a large centrifuge. A geotechnical centrifuge with a diameter of 2.5m has been built at the Geotechnical Laboratory of the University of Delft. To enable the performance of advanced tests in flight, the swing up platforms of the centrifuge were made large enough to contain computer-controlled devices. The test containers and actuators are so small that they can be handled by one person. This is convenient during the preparation of the tests and leads to good reproducibility of the soil samples. Due to the low weight modification of the centrifuge for different tests is simple, so that a flexible operation is obtained. A disadvantage of a small centrifuge is the limitation in the use of sensors during a test. This restriction, however, can be compensated by digital processing of images taken with the on-board video camera. Miniature devices have been developed for performing advanced tests in flight, such as: loading, displacement and controlling the flow of sand, water and air. The devices operate under the control of software, which runs on a PC compatible computer located in the rotating part of the centrifuge. To improve the reproducibility, sample preparation is automated as much as possible. A special centrifuge has been built to consolidate clay slurry, in order to obtain a very soft normally consolidated clay. Several research projects related to offshore engineering have been carried out in the centrifuge, e.g.: sliding behavior of spudcan footings, buckling of large diameter piles during driving, gas blowouts, installation and bearing capacity of suction piles and behavior of cyclic loaded buried pipes. Some related areas are; flow induced slope instability, trapdoor problem and pollution transport. The flexibility of the centrifuge is demonstrated by the fact that three quite different model tests can be performed on the same day. THE SMALL CENTRIFUGE The geotechnical centrifuge at the University of Delft (Allersma, 1994a)

KEY WORDS:
engineering. INTRODUCTION

geotechnical centrifuge, offshore foundation

Centrifuge research (e.g. Kimura et al., 1998) has proven to be effective in investigating the behavior of soil and other granular materials. With this technique it is possible to use scale models to examine the behavior of large scale problems. Clay is a typical example of a material with a strong stress dependent behavior but also materials like sand behave differently under different stress levels. In most practical problems the stress dependent behavior has to be taken into account if one is to make reliable predictions. The tendency has been to increase the size of the centrifuges in order to model more details. For several geotechnical problems, however, the use of

699

Electronic Facilities The system electronics enable the performance of computer-controlled tests in flight (Fig.2). To minimize electrical disturbances, the control unit is placed in the spinning part of the centrifuge. The unit contains a small single board IBM-PC compatible computer (180x120x25mm; 486CPU; 66Mhz; 16Mbyte RAM; 32Mbyte ROM disk; 1 Gbyte hard disk; 1.44Mbyte floppy disk), a 12-bit analog to digital converter with a 16-channel multiplexer, two voltage controlled outputs of 8 Ampere each; two 16-bit counters and several digital inputs and outputs. The 1Gbyte hard disk is placed just in the center of the centrifuge and operates correctly up to at least 160g at 1 meter. The signals from the sensors can be read by a computer program via a analog to digital converter. The voltage controlled outputs are used for proportional control of small DC-motors, where the digital inputs are used to detect the rotation of pulse wheels in order to deduce displacements. The digital outputs can be used for on/off control of several devices, such as DC-motors, electro/pneumatic valves, etc. Eight power slip rings are available to feed the electronics and the actuators. 24 high quality slip rings are used to transmit the more sensitive signals, such as, for example, two video lines and the connection between the on-board computer and the key-board and monitor in the control room (10 lines). Thanks to commercial available line driver units the on-board computer is accessible just like a normal PC. During a test the relevant parameters are displayed in graphical form and stored on the solid state disk unit or the hard disk. A special feature is that several phenomena can be measured using the video images. In this technique the video images of the in flight test are captured by the frame grabber in the PC and processed until the relevant parameters are isolated and digitized. Image processing can be used to visualize and digitize the surface deformation of clay and sand samples, to digitize the consolidation of a clay layer or to digitize the displacement of objects (Allersma, 1990). Since light is used as the measuring medium the measurement did not influence the test in any way. DEVICES FOR IN FLIGHT TESTING Several devices have been developed (Allersma, 1994b) in order to perform advanced tests in flight. Devices used for offshore projects are: a two dimensional loading system, pile driving hammer and air supply system.

Fig 1 Geotechnical centrifuge of the University of Delft. was designed by the Geotechnical Laboratory of the Department of Civil Engineering and was built by the mechanical workshop of the University. The electronic systems were designed and built by the Geotechnical Laboratory. The advantage of an in-house design is that the system can be expanded and modified under its own supervision, thus guaranteeing a good interaction between the facilities of the device and the tests. Mechanical Construction The centrifuge is a beam type centrifuge with two swinging platforms (Fig.1). A beam with a length of 1500 mm is connected to the axis, so that it can be rotated in the horizontal plane. Two swinging platforms are connected to the beam by means of brackets. The platforms are formed by two plates at a distance of 410 mm apart, which are connected to each other by four cylindrical steel beams. The surface of the plates is 400 x 300 mm. Samples with a weight of 400N can be accelerated up to 300 g (12g-tons). The centrifuge is driven by an 18 kW electric motor via a hydraulic speed control unit. The hydraulic speed controller is manipulated by a step motor, which is interfaced to a PC. A computer program has been developed to adjust the speed of the centrifuge using the signal of a tachometer. Several options are available to control the speed. It is, for example possible to make the acceleration dependent on time or on test parameters, such as the pore water pressure in a clay sample.

Fig.3 Diagram of the two dimensional loading system. Fig.2 Diagram of the electronic facilities. Two Dimensional Loading System The two dimensional loading system (Fig. 3) can be considered as a universal tool, which can be used for several tests. The system is driven by two miniature DC-motors. The displacement is measured by means of small

700

pulse generators. A vertical and horizontal displacement of 100mm and 200mm can be adjusted, respectively. The accuracy is better than 0.1mm. The loads in the two perpendicular directions are measured. The measured loads are used by a computer program to control the device. Loads of more than 5 kN can be applied by the system. Except for loading tests the device can be used as a simple robot to manipulate tests in flight or to take samples. To date the device has been used at gravitation levels of more than 150 g. Air Supply system. In some tests it is required that gas be supplied to a soil sample. Since the small centrifuge is not equipped with fluid slip rings the gas has to be stored in the spinning section of the centrifuge. To make the storage as compact as possible, two high pressure (200 bar) cylinders of 5 liters each are mounted on the beam of the centrifuge. A computer controlled air supply system has been developed in order to regulate the pressure and the gas flow. Flow rates of 10 l/s can be achieved. The gas in the high pressure cylinders represents a considerable amount of power, which can be used, in principle, for tests in which large loads or energy are needed. The gas supply system is used for several tests, such as: pile driving, water circulation, controlling a trapdoor test, simulating gas blowouts in soil layers and simulation of suction pile installation.

Fig.5 Computer controlled sand preparation device. prepare well defined sand layers in the test containers (Fig.5). The falling height of the sand can be adjusted in order to control the density. The height is kept constant during raining by means of an optical sensor and an actuator. Sand samples with a surface area of 300x300 mm and a maximum thickness of 150 mm can be made. The porosity can be varied between 35% and 39%. The sand samples can be reproduced with a standard deviation of less than 0.2 percent. The preparation of the sample with a thickness of 100mm takes about 20 minutes. APPLICATIONS IN OFFSHORE ENGINEERING Tests have been carried out in order to analyze several different geotechnical phenomena in offshore engineering. For most geotechnical simulations, many tests are required to get a better insight into the problem. The special feature of a small centrifuge is that the tests can be performed in a relatively short period.

Fig.4 Diagram of the air driven pile driving hammer. Pile Driving Hammer A simple in flight hammer has been developed to enable the simulation of pile driving. The hammer is driven by compressed air. A cylinder with a piston is mounted on the pile head (Fig.4). During flight (150g) the piston is lifted up by air burst. The supplied air escapes via holes in the cylinder, so that the piston falls down abruptly. The bounce with the bottom plate of the cylinder gives sufficient impact to penetrate a pile over some distance. To allow settlement of the hammer the air is supplied by means of a telescopic tube through the center of the piston. A rate of about 5 blows per second can be reached, where hollow cylinder piles with a prototype diameter of at least 2.5 meters can be simulated. SAMPLE PREPARATION An important aspect of centrifuge research is sample preparation. Samples of different soil types have to be prepared, in which the density can be varied. To enable comparing the results of different tests good reproducibility of the samples is required. An automated computer controlled curtain rainer has been developed to Fig.6 Load path applied to a spudcan footing, simulating storm conditions. Sliding Behavior of Shallow Footings Shallow footing (spudcans) are used for the foundation of offshore mobile drilling platforms. If the spudcans are installed on clay the penetration depth of the footing due to the pre-load is approximately one diameter. In the case of sand, however, the penetration depth is much less, so that the foundation elements are more sensitive to sliding during storms. To provide more insight into the sliding behavior several research projects have been carried out in centrifuges (e.g. Murff, 1996). More information, however, was needed about the sliding behavior of spudcans at low vertical loads, with shallow embedment on sand and sand over clay. Furthermore methods needed to be tested to improve the sliding capacity. The small centrifuge appeared to be very suitable to perform the required large number of tests on well reproduced sand beds. By means of the two dimensional loading system the loading path which can occur during a storm is simulated (Fig.6). After a pre-load (applied in practice during installation) has been applied, the vertical load is reduced to

701

e.g. half the maximum value. From this point the combined loading program is started (Fig6c). The path of the horizontal and vertical load is shown separately in Fig.6a and b, respectively. By performing tests with different loading patterns a curve is obtained representing critical combinations of the horizontal and vertical load. At low vertical loads the critical points form a more or less straight line. The slope of this line is an important parameter for design practise. Most tests are carried out at 150g. Since dry sand was used, the stress gradient agrees with a higher acceleration if a prototype of saturated sand layer is simulated. A footing with a diameter of 6 cm is therefore equivalent to a prototype footing with a diameter of 14.4m. This is a realistic prototype dimension. The test results showed that the first sliding check (based on the slope of the failure curve) as advised in the Recommended Practice for Site Specific Assessment of Mobile Units (1994) is applicable to restricted loading conditions and may therefore be too optimistic (Allersma et al., 1997b). In a next test program the (sliding) behavior of spudcans on sand overlaying a stiff clay (Cu=95kPa) was examined (Allersma et al., 1998). It was found that the clay layer has a significant influence on the bearing capacity, even with sand layers with a height of more than the spudcan diameter (Fig.7).

A B Fig.8 Optically measured principal stress trajectories in a centrifuge test; a) vertical load, b) sliding. sliding capacity (Allersma, 1997a). Three pins and a skirt appeared to be most effective, but also the roughness contribute significantly to the sliding capacity. In Table 1 the different modifications are compared with a smooth spudcan without a tip. Due to the good reproducibility of the sand samples the effect of small changes in the shape of the footing and installation procedure could be determined. A missing element in experimental geotechnics is that the stresses cannot be measured in the interior of a sample of granular material. The only available method of obtaining continuous information about the stress distribution is the use of optically sensitive granular material, such as crushed glass (Allersma, 1987). An assembly of crushed glass can be made transparent by submerging the pores with a liquid with a matching refractive index. By means of polarized light, the stresses, for example, during a cone penetration test can be made visible. A first attempt to apply this test technique in a centrifuge is shown in Fig.8, where the major principal stress directions underneath a footing before and after sliding are plotted.

Fig.7 Load displacement diagrams of spudcans founded on sand overlaying clay Probably because the clay was rather stiff the so called punch through mechanism could not be observed. Since the penetration depth in layered soil is more than in pure sand the spudcans are less sensitive to sliding in layered soils. Table 1 Comparison of different modifications to improve the sliding capacity of spudcan footings.

Fig.9 Diagram of the test setup to simulate suction pile installation. Suction Pile Installation Suction pile foundations have been applied increasingly in offshore engineering. Suction piles are attractive because of the convenient method of installation. A pile with a diameter of 9 m and a height of 10 m can be installed in 1 hour, by using a pump only. Several research programs in centrifuges have been carried out to investigate the bearing capacity of suction piles. However, little research has been performed to examine the behavior of suction piles during installation. Centrifuge tests on the installation of suction piles in clay are performed by Renzi et al. 1991. Since the resistance of clay is relatively small the tests could be performed at atmospheric conditions. The installation of suction piles in clay or sand generally poses no problems, so that there is not much interest in performing centrifuge tests on this subject. There are some special cases, however, where testing can yield interesting information. For example, the behavior of suction piles in layered soil (sand, clay, sand) was not well known. It was not clear what happen when the pile enters the sand after passing the clay layer. There were doubts if the clay layer in the pile would allow sufficient water flow. Furthermore, for engineering practice it is interesting to know what

To achieve a more save foundation the sliding capacity of the spudcans can be improved by modifications of the shape. Examples are: a spicular tip, pins, ribs, skirt or a rough surface. Centrifuge research has proven to be extremely suitable to examine the influence of these modifications on the

702

limits apply with respect to the height and diameter of the piles. A suction pile is a huge can which is placed upside down on the sea bed. Installation occurs by pumping water out of the pile. The pressure difference causes a driving force, where the boundary and tip resistance is reduced by local fluidization of the sand as a result of the induced ground water flow. A problem in centrifuge tests is that the maximum pressure difference is dependent on the water depth. In particular, in a small centrifuge it is not easy to simulate a water depth of 40m by means of water. However, by performing the test in a closed container and by connecting the inside of the suction pile to atmospheric pressure, a pressure difference can easily be achieved, by pressurizing the container. No pumps are required in this case and a large water depth can be simulated. The limit of the pump capacity can be simulated simply by means of a nozzle. A diagram of the test set-up is shown in Fig.9. At 150g prototype piles can be simulated with models with a diameter of 60mm and a height of 66mm. Parameters that can be measured are: the pressure difference, the penetration depth and the water flow. A typical output of a test is shown in Fig.10. In homogeneous sand it was found that there was almost a linear relationship between the dimensions of the pile and the required pressure (Allersma et al., 1997c). The load required for

pulses. Thanks to this technique the maximum pressure difference is available. In the centrifuge tests, situations were created where the suction pile did not enter the soil under continuous pumping conditions. Hence the new technique could then be tested to observe any improvement. Indeed it could be shown that percussion method yields beter results in the case of coarse material, uneven seabed and in layered soils. Also there was an indication that piles with smaller diameter could be used. Buckling Behavior Large Diameter Piles The need for a pile driving device originated in the investigation of the buckling behavior of open ended hollow cylinder steel piles. This problem was related to an offshore project at the West coast of Australia . A production platform was founded in soft soil using piles with a diameter of 2.7m, wall thickness of 45mm and a length of 100m. The piles were installed down to a sintered calcareous soil layer by hammering. Since the soft soil did not yield enough shaft friction, the piles were not supposed to be capable of resisting the pullout forces. Therefore it was planned to drill through the open piles into the calcareous layer (50m), so enabling the installation of an insert pile. The whole system was to be fixed by grouting. However, a problem arose when an attempt was made to enter the piles with the drilling machine. It appeared that almost all piles had progressively collapsed over a distance of 30m from the tip. Looking at the tip a peanut shaped cross section could be observed.

Fig.10 Pressure and displacement during a tests on a suction pile. installation by suction is approximately eight times smaller than needed for mechanical penetration. It could be observed in transparent piles that fluidization only occurs in a restricted zone close at the boundary. If the diameter becomes too small, the whole soil plug starts to fluidize, which causes considerable upheave. It was found, however, that piles with a diameter over height ratio of 1 to 4 could be installed without any problem. An initially inclined pile tends to become more vertical during installation. Centrifuge tests have shown that a clay layer in layered soil (sand, clay, sand) did not hamper the installation of a suction pile. In the meantime, the feasibility of installing suction piles in layered soil has also been demonstrated in practice. A new installation technique (Allersma et al., 2001) has been tested in the small centrifuge, in order to investigate if installation can be extended to more difficult circumstances, such as: coarse material, an irregular sea floor or clay over sand. In this method the driving force is not obtained by continuous pumping, but by creating short pressure

Fig.11 Examples of deformed large diameter piles during driving. In order to gain more insight into this problem, tests have been performed in the centrifuge. Piles with a diameter of D=12mm and a wall thickness of d=.2mm (D/d=60) were inserted into dry sand at a gravity of 150g. To prevent plugging of the sand during installation, it appeared to be necessary to install the pile by hammering. For this purpose it was not necessary to know the energy of each blow exactly. The important point was to ensure similarity of the installation with practice in the field. Therefore the simple pneumatic pile driving hammer could be used to drive the piles in flight. About 200 blows were required for the pile to penetrate medium dense sand by 20m. The tests on the large diameter pile have shown clearly that a small initial physical damage is required at the tip in order to simulate collapse of piles during driving. Some shapes of pile tips are shown in Fig.11.

703

peak load of 15% below the static pullout load. After performing several cycles the peak load was increased automatically by 5%. A typical test result is presented in Fig.13. It appeared that cyclic loading did not decrease the pullout resistance significantly. This test program was an example of a direct application of centrifuge tests for the validation of a design. Since the tests could be performed in a short period and the relative low cost of operation made the small centrifuge very attractive for this purpose. Simulation of Cratering Due to Gas/Liquid Flow A man made crater can occur during drilling for gas/oil reservoirs. If the control over the pressure is lost at the moment that the cover of the reservoir is penetrated an internal blow-out occurs. Because of accidental canals the escaped fluid and gas is able to reach the surface. If the flow is strong enough, a crater with fluidized soil can be formed, which can, for example cause the drilling installation to sink down. A chain reaction can be initiated in the case that neighboring production pipes enter the crater zone. Therefore safety valves are installed in many cases. A prediction has to be made about the installation depth. The formation of a crater is a complicated process, which cannot be described simply by analytical or numerical models. Therefore tests are the only method for getting insight into this phenomenon. Since it is almost impossible to perform field tests small scale tests are the only way to study the mechanism. Small scale tests at 1g are not realistic, because the capillary cohesion influences the results significantly. In a centrifuge, the self-weight stresses are increased, while the capillary cohesion remains the same. To enable visualization of the process several tests are performed under plane strain conditions. Sand layers with a maximum height of 40 m are simulated. The gas pressure and flow rate can be adjusted in flight. In Fig.14 a typical test result is presented in a sand layer. The most important results are (Allersma et al., 1994c):

Fig.12 Influence of attachment height on the horizontal bearing capacity of suction piles. Horizontal Bearing Capacity Suction Piles A test program has been carried out to examine the influence of several parameters on the horizontal bearing capacity of suction piles (Allersma et al., 1999a, 1999b). In these tests the sand was poured around and in the piles, in order to achieve a good reproducibility (better than 0.2% in density) of the sand sample. Some of the variables were: attachment angle (see Fig.3), attachment height, H./D ratio and soil cover. In Fig.12 the results are shown graphically of a test series in which the attachment height is varied. It appeared that there was an optimum height at 2/5 from the bottom of the pile. The experimental results are compared with a three dimensional finite element calculation. It was found that the calculation show the same tendency as the tests, however the calculated bearing capacities were about 10% higher than the measured bearing capacities. The attachment angle was influencing the bearing capacity significantly. This indicated that it has to be considered to bury the cable in order to keep the cable as horizontal as possible. An interesting observation was that the horizontal bearing capacity increases by 30% if 1/3 of the upper part of the pile was removed after installation.

Fig.14 Simulation of a gas blowout underneath a footing in sand at 150g. - A crater is formed mainly when the gas supply is started abruptly. - A clay layer also initiates a crater at a gradual gas supply, because the gas accumulates in the first instance under the clay layer, where the sudden escape simulates an abrupt supply. Also a foundation element at the sand surface causes a crater at a gradual gas supply. - By interpolation it could be deduced that the maximum crater depth was between 60m and 100m. CONCLUSIONS The small geotechnical centrifuge at the University of Delft has proven to be a successful tool in testing designs of offshore foundations. The small size of the samples means that the machine is very flexible in operation and tests can be performed in a short period after an idea has been formulated. Due to the application of state-of-the-art electronics, measuring techniques

Fig.13 Behavior of a buried pipe segment during cyclic vertical loading. Cyclic Behavior Buried Pipes In some cases buried pipes at the sea floor tend to heave in the soil. This can cause a dangerous situation when fishing lines are hooked by the pipes. A possible reason for upheave is that the pipes buckle somewhat due to cyclic temperature changes in the transported fluid. The buckling forces can be estimated by theoretically, so that it is possible to test this in the centrifuge to see if the cyclic buckling loads can cause upheave. To meet the practical conditions as close as possible soil was used from the actual site (only 10 litres was needed). In the first instance the static pullout capacity was determined and next a pipe segment was subjected to cyclic loading with a

704

and special tools, advanced tests can be performed in flight. Since the computer is located on the spinning part of the centrifuge, only a few slip rings are required to manipulate on board equipment, and the electrical noise is minimal. The small sand samples could be reproduced accurately with the automated preparation device. It appeared that several offshore problems could be examined effectively in a small centrifuge. With small models the behavior of structures could be examined, which could not easily be predicted by calculations or real scale tests. Modifications can be made very easily and a large number of tests can be performed in a project to examine the influence of several parameters. The test have led to a better understanding of several problems and the results have proven to be valuable for the design practice. ACKNOWLEDGEMENT The centrifuges, electronics, soil preparation devices and testing equipment were all designed by the Geotechnical Laboratory of the Department of Civil Engineering at the University of Delft. Many thanks are given to several graduate students and the technicians of the laboratory for their contributions to this research. Several research projects are supported by Shell International Exploration and Production (SIEP), Shell Expro UK, Fugro Engineers b.v. and Suction Pile Technology.b.v. REFERENCES Allersma, H.G.B. 1987: Optical analysis of stress and strain in photoelastic particle assemblies, Thesis, University of Delft, Netherlands. Allersma, H.G.B. 1990: On line measurement of soil deformation in centrifuge tests by image processing. Proc. Int. Conf. on Experimental Mechanics, Copenhagen, pp. 1739-1748. Allersma, H.G.B. 1994a: The University of Delft geotechnical centrifuge. Int. Conf. Centrifuge94, Singapore, Balkema Rotterdam, pp. 47-52. Allersma, H.G.B. 1994b: Development of miniature equipment for a small geotechnical centrifuge. Transportation Research Record No. 1432. Nat. Academy Press, Washington D.C., pp. 99-105. Allersma, H.G.B. 1997a: Centrifuge tests on the influence of shape on the sliding behaviour of spudcans. Int. Conf. Behaviour of Offshore structures, Delft, the Netherlands, pp.225-233. Allersma, H.G.B., W.J. van Niekerk, A.P. Kooijman, 1994c: Simulation of cratering in a small geotechnical centrifuge. Int. Conf. Centrifuge94, Singapore, Balkema Rotterdam, pp. 325-330. Allersma, H.G.B., B. Hospers, J.G. den Braber, 1997b: Centrifuge tests on the sliding behaviour of spudcans. Canadian Geot. J., 34(5), pp.658-663. Allersma, H.G.B., H.G. Stuit, P. Holscher, 1994d: Using image processing in soil mechanics. XIII Int. Conf. on Soil Mechanics and Foundation Eng., New Delhi, pp. 1341-1344. Allersma, H.G.B., F.J.A. Plenevaux and J.F. Wintgens 1997c: Simulation of suction pile installation in sand in a geocentrifuge. ISOPE97., Hawaii, May, Vol. 1, pp. 761-766. Allersma, H.G.B. and F.M. van Woensel, 1998: Centrifuge research on the sliding behaviour of spudcans in layered soils, ISOPE98, Montreal, pp. 621-627. Allersma, H.G.B., A.A. Kirstein, R.B.J. Brinkgreve, T. Simon 1999a: Centrifuge and numerical modelling of horizontally loaded suction piles. ISOPE99, Brest, pp. 711-717. Allersma, H.G.B., A.A. Kirstein, R.B.J. Brinkgreve 1999b: Centrifuge and numerical modelling of methods to optimize the horizontal bearing capacity of suction piles.. OMAE99, St. Johns, pp.621-627. Allersma, H.G.B., J.R. Hogervorst, M. Pimoulle, 2001: Centrifuge modelling of suction pile installation using a percussion technique. ISOPE01, Stavanger, Vol.2. pp. 620-625. Kimura, T., O. Kusakabe, J. Takemura 1998: Editors Proc. of the Int. Conf.

Centrifuge98, Tokyo, Balkema, Rotterdam. Murff, J.D. 1996: The geotechnical centrifuge in offshore engineering. Int. Offshore Technology Conf., Houston, 6-9 May, pp. 675-689. Recommended practice for site specific assessment of mobile jack-up units; 1st edition, 1994. The Society of Naval and Marine Engineers, Jersey City. Renzi, R., W. Maggioni, F. Smits (1991), A centrifugal study on the behaviour of suction piles. Int. Conf. Centrifuge91, pp.169-176.

705