Numerical modelling of an experimental energy pile

Geothermal energy is a way of reducing the cost of energy. Deep geothermal energy systems extract heat from very deep soil layers where the temperatures are very high. Shallow geothermal energy systems are about 150 metres deep and they are used to store heat in the soil, to extract it later and use it for space heating. These shallow geothermal systems are generally embedded in a borehole, but they can also be cast into structures, which are called thermo-active foundations. An example of such a foundation is the energy pile, a foundation pile with a heat exchanger embedded in it, connected to a heat pump. There is no need to drill an extra hole in the ground, but the downside is that it is not well known how the bearing capacity of the pile is affected by the heating/cooling cycles. An energy pile experiment is planned to investigate the thermo-mechanical behaviours of the pile and the goal of this thesis is a numerical investigation of the pile. It serves as an estimate of the pile and soil behaviour prior to installation and the results will be used to confidently design the experiment. The model was built with DIANA FEA software, that is capable of coupling thermo-mechanical behaviour. Firstly an experiment in London clay was recreated in order to verify and validate the model and modelling approach. The experiment in Delft was modelled using site investigation that was done at the location where it will be built. Along with old Cone Penetration Test (CPT) data the subsurface was mapped and soil parameters used in the material model were chosen. The thermal cycle that was imposed on the pile was chosen on the basis of preliminary modelling done at different temperature increments. The temperatures were chosen such that the pile was affected to a significant degree of its capacity. It was cooled for three weeks to 0 °C and then heated to 24 °C for three weeks, this was repeated for 6 years. The research focussed on finding which thermo-mechanical effects can be expected and what the scale of those effects could be. The effect directly linked to an increase in temperature is thermal strain. Materials tend to expand and contract with the temperature at different rates and so do the pile and the soil. The gradient of the heat flow is also an important factor as the pile is subjected to the temperature before the soil is. The pile will expand first and this will be resisted by the soil, the strain that is resisted by the soil is called the restrained strain and that is responsible for the change in stress in the pile. A pile that is heated will have more stress than with just a mechanical load and a pile that is cooled will see a reduction in stress. The pile will expand vertically around a null-point somewhere along the pile, this is the point that does not move. In principle, an unrestrained pile will have a null-point in the centre of the pile, but because some soil layers resist the pile movement more than others the null-point is closer to the stronger layers. In the Delft experiment model the null-point was halfway down the pile at first, but as the amount of cycles progressed it moved down. This is due to a decrease in resistance from the weaker layers and an increase in resistance in the strong sand layer on which the pile is based. The amount of stress that is generated is also less in these later cycles as the resistance of the soil became less. With that reduced resistance an increase in settlement is also seen. This can lead to differential settlements of structures that are built on such a pile, possibly damaging them. The results of the modelling are used to give an advice on the experiment details, such as geometry, thermal cycle and pile layout. An advice to the layout of the sensors is included as well as a prediction of the results