GEOTECHNICAL RESEARCH

by

Andrew Schofield

The GCC is devoted to physical modelling, and especially to the proposition that soil stresses must be replicated if similitude is to be obtained. Since geostatic stresses in the field increase with depth, due to the self-weight of soils, it is necessary to enhance body forces in small-scale models tested in the laboratory. If boundary conditions are sufficiently representative of the field, if the geometry and composition of soil layers or slopes is also representative of the full-scale, and if there is a technique by which the strain history of the field problem can be replicated prior to the creation at model scale of the intervention of interest, the centrifuge model will recreate in each elementary volume the stress rotations and anisotropy seen at corresponding points in the field-scale problem.

 Triaxial tests and standard soil constitutive models, such as Cam Clay, fall far short of being able to induce or describe these more complex facets of soil behaviour. Not only are such phenomena automatically produced in a well-designed centrifuge test, the interactions between the various soil elements and structures are equally accounted for. Very often, centrifuge testing provide the only practical means of validating numerical simulations or design methods; this is especially so when the events of interest are catastrophic.

 As with other facilities in the Group, the GCC can be made available to all research students whose supervisors can fund appropriate experiments. Current investigations number about fifteen, and they include:

* soil loading/unloading changing skin friction in piles (with Dr Al Tabbaa),

* stability of landfill repositories in earthquakes (with Dr Madabhushi),

* spillage of DNAPL contaminant leading to pollution transport (with Dr Soga),

* sorption of aqueous pollutants onto soil minerals (with Dr Lynch),

* injection of grout near tunnel linings (with Dr Bolton and Professor Mair),

* the burial of submarine pipelines (with Dr Bolton and Professor Palmer).

 Each element of soil in the model should respond to the complex intervention of the research worker in a way which makes it possible to create and validate mechanisms and analyses which could later be used in the field. The precise "scaling laws" governing similitude between centrifuge models and full-scale structures is a matter for experiment, analysis and comparison. The most obvious distortion is the much larger particle / structure size ratio in small physical models compared with the field. Particle size and strength effects are, accordingly, a key research interest of Dr Malcolm Bolton. One obvious fact is that the particle / model size ratios in a typical soil model in a centrifuge are no larger than typical triaxial tests, and may be much smaller.

 The Centre houses three geotechnical centrifuges, which are used for different classes of experiment. A BALANCED BEAM CENTRIFUGE , capable of creating up to 150 "gravities" of centrifugal body force on a 900 kg test package mounted with a typical working radius of 4.0 m, has been in use for 24 years. The test container needs to be capable of withstanding full-scale lateral earth pressures. A 300 mm depth of soil at 100 g is equivalent to 30 m at field scale, giving an equivalent vertical total stress of about 600 kPa. If the soil became fluidised during a test, a similar lateral pressure would have to be carried by the container, which may alternatively and justifiably be called a "strong box". The soil in its container, together with whatever actuators, standpipes etc are necessary to the task, form a test PACKAGE which is placed on a SWINGING ARM which swings outwards as the centrifuge achieves about 10 g, landing the package safely on the strong back plate of the rotor. From there, as the rotor accelerates up to scale speed, the centrifugal force acting in the radial direction replaces earth's gravity as the principal body force on each component of the model.

 Pore pressures can be measured using buried transducers, cross-sectional displacements can be measured using PHOTOGRAPHS taken through a thick window, and lasers can be used to scan surfaces for settlement. Complex soil-structure interactions, such as simulated storm loading on a lattice leg "spud-can" foundation for an off-shore JACK-UP RIG can be created by actuators and monitored via load cells. New work this year includes a study of retro-fit strengthening for clay embankments, in a collaboration with Mott Macdonald for LUL.

 The power and sophistication of actuators extends to a shaking table, flying on the beam centrifuge, which is capable of simulating severe earthquakes. This development is in the hands of Dr Gopal Madabhushi who is the Assistant Director of the Centrifuge Centre. Powerful earthquake actuators have been available on the beam centrifuge for nearly 20 years, but the new Stored Angular Momentum (SAM) actuator extends the range of frequencies and intensities available to the research worker. Soil models can be constructed in a laminated container whose end walls flex to simulate free-field soil displacements. Investigations of slopes and bridge foundations in earthquakes are under way, together with an on-going study of the engineering implications of soil liquefaction.

 The GCC also houses two drum centrifuges, one 2 m diameter and one 0.8 m diameter - both capable of running to 500 g. Soil can be sprayed or poured into the ring channel of a drum centrifuge as it rotates, to form a uniform carpet around the circumference. At top speed, our 2 m DRUM models a strip of soil in the field 3km long, 500 m wide, and 75 m deep. It is therefore particularly well-suited to research programmes in which many tests should be performed on a single, uniform soil bed. For example, different combinations of vertical, horizontal and moment loading on a 3-LEG JACK-UP were investigated for EXXON. The 0.8 m mini-DRUM is particularly useful for outlining mechanisms, getting a quick start on new problems, by handling a relatively small and manageable sample of soil.

 A new application in the 0.9 m drum this year is the fast measurement of uplift resistance of buried pipelines, especially of subsea pipelines with clay back-fill, which has been funded by industrial research contracts via Cambridge University Technical Services Ltd with companies including Coflexip Stena Offshore, Andrew Palmer and Associates, and Brown and Root.

 A particular feature of the Cambridge drum centrifuges, pioneered by Prof Andrew Schofield , is their twin coaxial shafts - they are, effectively, one centrifuge (the central turntable, carrying actuators or tools) rotating independently of the other (the ring channel, carrying the soil). This makes it possible, in principle, to keep the soil spinning correctly while the various stages of soil pouring and actuator interventions - such as a hopper for embankment construction IN FLIGHT are conducted in sequence by stopping and starting the central turntable.