Cambridge University's Geotechnical and Environmental Research Group has celebrated the award of its 150th PhD.Now it's back to business.Gareth Beazant reports.
Professor Robert Mair, head of the Geotechnical and Environmental Research Group at Cambridge University, has just witnessed the group's 150th student achieve a PhD.
Six years into his tenure, Mair's aims remain the same - to forge links with industry and boost what is already one of the leading research groups in the country. More PhD students are queuing up to join.
'A lot of research was carried out for Channel Tunnel Rail Link tunnelling projects, ' Mair says.
'We see Crossrail as the next big project. With less and less space for tunnels in the urban environment, it means tunnels come very close to heavily loaded buildings.'
Underground construction is one of seven sectors covered by the group. The others are insertion, penetration and extraction; fundamental soil behaviour; environmental engineering; earthquake engineering; numerical modelling;
and physical modelling.
There are two professors, five academic staff, five post-doctoral researchers, seven technicians and about 40 students. Each layer of the department is highly valued by Mair.
'The heart of any good research group is the technicians and we are fortunate to have some very good ones, ' he says.
Students may be sponsored by companies, research councils, EPSRC, or, in the case of overseas students, by their home countries, or the Cambridge Commonwealth Trust.
One recent area of research has seen the group working with the Massachusetts Institute of Technology to develop advanced wireless sensor technologies.
Ground Engineering takes a look at three other ongoing projects.
Influence of diaphragm wall installation on pile and soil responses Gary Choy, Jamie Standing and Robert Mair The complex 3D nature of the interaction between piling and diaphragm wall construction, means accurate predictions based on analytical and numerical approaches are often difficult.
Centrifuge modelling is being used to simulate precise field conditions. Piles were tested at three locations away from a model trench, with a guide wall installed to 20mm. During each centrifuge test, the model pile was driven to 250mm.
Results show that the location of the pile relative to the trench is very important, with significant pile movement observed during reduction of slurry level. Substantial end-bearing resistance reduction and changes in shaft friction were also recorded.
The pile has a destabilising effect on the soil mass in increasing the magnitude of soil surface settlement and a crater is formed surrounding the trench.
Seismic behaviour of layered soils Barnali Ghosh, Thusy Thusyanthan, Andrew Brennan and Gopal Madabhushi The nature of soil structure interaction effects is difficult to predict and analyse numerically.
Numerical analysis of seismic behaviour of saturated soil under earthquake shaking depends mainly on the dynamic soil model chosen.
The main aim of this work is to look at the failure of mechanisms following liquefaction of the foundation soil layer supporting a landfill subjected to earthquake loading. The dynamic response of the landfill liner system during the earthquake loading is also being studied.
Previous studies show most of the constitutive models have difficulties when layering is introduced within the soil. In most of the formulations, the soil is idealised as a linear visco-elastic medium, but most of the earthquake records indicate non-linear behaviour of the soil mass during severe earthquakes.
A series of dynamic centrifuge tests have been performed in different soil stratifications and the results are being used to validate the numerical code SWANDYNE which uses a coupled formulation for the solid and the fluid phase respectively.
In general, accelerations matched well with the centrifuge results. Pore pressures did not match the sections, especially at the interface of the loose and dense soil. It was not possible to model the migration of pore pressures from the dense zones to the unimproved zones effectively. It was also felt that the pore pressure prediction capability in dense sand is usually under-predicted.
In layered soil, the system may behave quite differently, depending on factors such as the rate of pore pressure build-up during shaking, contractive or dilative behaviour of the soil, soil permeability and the way the dynamic stiffness of the soil decreases as the excess pore pressures increase.
The overall aim is establish the validity of dynamic centrifuge modelling technique in investigating the behaviour of landfill liners during and after earthquakes.
Observation of the uplift mechanism of buried pipelines using PIV Johnny Cheuk, Dave White and Malcolm Bolton Oil pipelines are daily subjected to high temperatures and pressures with thermal expansion restricted by the side friction of the pipe and the end restraint.
This leads to the risk of upheaval buckling.
Soil cover is normally used to prevent pipelines from buckling upward. The required cover depth depends on the available uplift resistance of the overlying soil, both from weight and shear resistance.
Design models cannot accurately estimate the available resistance because of a lack of understanding of the uplift mechanism of buried objects.
As well as improving pipeline design, this project will tackle the issue of mobilisation distance scaling between full-scale and centrifuge tests when rupture bands form and the appropriate constitutive relation is stressdisplacement rather than stressstrain.
To investigate the uplift mechanism of buried pipelines, model tests have been conducted in a calibration chamber. Initially, an embedded pipe section is lifted up by an actuator.
Digital photos are taken at regular time intervals and the images are analysed by Particle Image Velocimetry (PIV) to obtain the displacement field at various stages. The development of the failure mechanism is revealed by the strain distributions at different stages.
A common model for the estimation of uplift resistance is the vertical slip model (Figure 1).
This assumes failure surfaces are vertical and start from the edge of the pipe to the ground surface.
The pull-out resistance is estimated as the sum of the soil weight above the pipe and the shear stresses mobilised along the failure surface. Shear stresses are approximated by the vertical effective stress and an earth pressure coefficient K.
Estimating K is difficult. One approach is to use the at-rest earth pressure coefficient. An alternative is to adopt a value in the range from 0.5 to 0.75, depending on the density of the soil.
Importantly, the assumed vertical failure surfaces do not represent the true failure mechanism of a buried object. The inconsistency is particularly significant for dense soil for which the failure mechanism is not kinematically admissible due to soil dilatancy. A better understanding of the failure mechanism will lead to improved confidence, and more economic design.
The calibration chamber used in this study is equipped with a sand pourer which allows uniform sand layers to be formed. By adjusting the drop height and the flow rate of the pourer, the soil density of the model can be varied.
Two glass plates are placed on the inner surface of the chamber to reduce the friction at the soil/wall interface. A PTFE material was used to reduce the friction between the pipe and the glass plates. A load cell which connects the tie rod to the actuator was used to measure the uplift force required to lift the pipe.
Pipe displacement was determined by Particle Image Velocimetry (PIV) operating on the artificial texture drawn on the pipe. Fraction A silica sand, which has a mean diameter of 2mm, was selected for the present study. A relative density of 90% was achieved at drop height of 250mm.
The 100mm diameter model pipe section was embedded at a depth of 300mm which corresponds to a cover depth ratio (H/D) of 3. During the test the pipe was pulled vertically upward at a speed of 0.003mm/s. Three digital cameras were used to take photos at 90s intervals. This corresponds to a pipe displacement of 0.27mm between images.
Results show that peak uplift resistance is mobilised at very small displacements. This is backed up by formation of a small gap underneath the pipe.
The displacement field obtained by PIV analysis shows the failure planes extend to the ground surface at peak resistance, although a progressive decrease in mobilised shear strain with distance from the pipe shoulder is present. From the shear strain contours it can be seen that the shear bands are inclined due to the dilative tendency of the soil.
Further displacement of the pipe enlarges the cavity underneath it, which soil particles then start to fall into. This gives rise to the reduction of uplift resistance; a new mechanism is operating.
At larger displacements the uplift resistance reaches a constant value, associated with a flow-round mechanism, in which soil falls into the triangular cavity below the pipe. As the slope of the cavity cannot sustain inclination higher than the soil's friction angle, soil particles keep filling the gap as if the cavity is moving with the pipe.
The images obtained will be analysed further using PIV and calibrated using close range photogrammetry. Similar experiments will be conducted on models with different densities to investigate the change in uplift mechanism with soil density.
Fraction D silica sand, which has mean particle size of 0.2mm, will be used to study the influence of particle size. The influence of relative size between the soil and the uplifting object is also under investigation.
The whole experiment will be repeated in the mini-drum centrifuge at 10g using Fraction D sand to mimic the behaviour of Fraction A sand.