Development, implementation, calibration and use of practical constitutive models in finite element analysis of offshore foundations

The finite element method is routinely used for the design of offshore wind turbine foundations, with complex time-consuming three-dimensional (3D) finite element analysis (FEA) typically performed to calibrate site-specific simplified macro element or one-dimensional (1D) finite models for rapid foundation sizing design calculations. The soil constitutive model used in a FEA calculation significantly influences the predicted foundation response. Therefore, given the uptake by industry of FEA for offshore wind turbine foundation design, it is important that practical advanced constitutive models are available which can be demonstrated to work robustly and accurately as design tools. There have been significant advances in constitutive modelling of soils over the last few decades, with many advanced models presented in the literature. However, the uptake of such advanced models in industry, for foundation design FEA, is still relatively low with more basic robust models typically preferred. The principal reasons for this are considered to be: difficulties in calibrating model parameters to standard laboratory test data; robust implementation of advanced constitutive models in FEA software is time consuming and complex; excessive analysis run times and convergence issues makes the use of some advanced models on design projects unfeasible; and there are few examples that demonstrate the robustness and predictive capabilities of many advanced models for complex 3D soil-structure interaction problems. It is therefore important that practical constitutive models are also developed which can be demonstrated to work robustly and accurately as useful tools for design FEA. This project therefore considers the development and use of practical constitutive models for design, and includes a review of some existing models, with a focus on the predictive capability and parameter calibration process for foundation design FEA. A practical robust sand model formulation and implementation is presented for use in design analysis with extensive calibration of the model to laboratory tests data, at a number of different sites, followed by FEA of monopile and suction bucket foundations. In addition, the formulation and implementation of a multi-surface soil plasticity model, termed the PIMS model, which is suitable for short-term storm type loading conditions, is presented. Calibration of the PIMS model to North Sea overconsolidated clay laboratory test data is performed, followed by FEA of monopiles foundations under monotonic and cyclic loading. The suitability of a bounding surface plasticity model for sand to predict the uplift response of a suction bucket foundation under different loading rates is also investigated, with comparison of FEA predictions to previously-published centrifuge test results. Finally, numerical analysis of monopile foundations performed as part of geotechnical design progression for a large offshore wind farm development, in the North Sea is presented using a number of different constitutive models. In all cases extensive calibration to laboratory element tests (primarily from North Sea soil units) is presented, followed by the comparison of FEA predictions to foundation load tests, where possible, to demonstrate that the models are suitable for design FEA of offshore wind turbine foundations.