Numerical modelling of laterally loaded piles for offshore wind turbines

The offshore wind market has grown rapidly over recent years. The most widely used foundation type for offshore wind turbines is the monopile. Current design guidance for monopile foundations dates back to the 1950s and 1960s and was originally developed for the oil and gas industry, where both the pile dimensions as well as the design load conditions are different than for offshore wind. The PISA project was aimed at investigating the behaviour of monopile foundations and reducing the conservatism in design. The project used state of the art numerical modelling, which was validated through a field testing campaign. The numerical modelling was focused on capturing the monotonic response of the foundation to failure. The field testing provided valuable additional data for the rate effects and the cyclic behaviour of monopile foundations. Data analysis methods are presented to accurately interpret the foundation response. The data analysis is focused on capturing the ground level foundation response as well as the pile behaviour below ground. Additionally, the cyclic behaviour is analysed in detail. These results are used as a basis for the development of numerical models for capturing the observed behaviour in the field tests. This thesis outlines plasticity models which are integrated in a generalised Winkler model. The model development makes use of the Hyperplasticity framework. These numerical models include: (1) the kinematic hardening model to capture plastic unloading; (2) the rate effect model to capture increased foundation capacity with increasing load rate; (3) the ratcheting model to capture accumulated rotation under cyclic loading; (4) the combined rate and ratcheting model and (5) the gapping model to capture gapping on the active side of the pile. Each of the models is calibrated to the PISA eld tests illustrating the capabilities and limitations of the models for capturing the eld test response. Finally, the kinematic hardening model is integrated in software to calculate the dynamic behaviour of an operating wind turbine. Integrating more accurate soil-structure interaction models could lead to improved predictions of the turbine behaviour and reduce the cost of monopiles.