Reliability-Based Design of Wind Turbine Foundations – Computational Modelling

Among renewable green energy generators, wind turbines are the most technically and economically efficient. Therefore, wind power plants are experiencing a competitive increased trend in global growth. The gas and oil industry is shrouded by political conflict, not the least of which is burning of fossil fuels causing pollution, environmental degradation, and climate change, and finally mixed messages regarding declining domestic and foreign oil reserves. Therefore, the wind power industry is becoming a key player as the green energy producer in many developed countries. However, consumers demand increased cost-effectiveness in wind turbines, and an optimized design must be implemented on the expensive structural components. The traditional wind turbine foundation typically expends 25-30% of the total wind turbine budget; thus it is one of the most costly fabrication components. Therefore, a reduction in foundation cost, and optimizing foundation structural design is the best solution to cost effectiveness.

An optimized wind turbine foundation design should provide a suitable target reliability level. Unfortunately, the reliability level is not identified in most current deterministic design model methodologies due to the fact that uncertainties related to design variables in the model (e.g., resistance and/or load) are either not included or excluded due to the application of partial safety factors and quantile values. This typically results in a conservative design, leading to unnecessarily high fabrication costs. Furthermore, expensive in-situ tests to measure quantile values are usually performed at each wind turbine location in a wind farm. The test location patterns could also be optimized if the soil variability (e.g., mean value, variance, and correlation lengths) effects on the foundation reliability have already been characterized. Given that usually a conservative result has already been obtained through the current deterministic design methodologies; consequently, a reliability-based design can be suggested to quantify the uncertainties related to the design parameters, and calibrate the current deterministic methods.

The overall objective of the present Ph.D. research is to propose a reliability-based design for the traditional wind turbine foundation. For this reason, probabilistic computational models have been established to characterize the uncertainties related to the design criteria, including deflection, rotation, bearing capacity, or stiffness of a wind turbine foundation. These models were developed from frequent applied classical methods for wind turbine foundation designs, or from probabilistic numerical modelling. The models integrated soil heterogeneity as random fields, and based on the models the foundation reliability was assessed. Advanced reliability methods were introduced and employed to estimate failure probability rapidly and accurately by decreasing computational times. A primary result indicated a requirement to quantify soil uncertainty effects in the foundation design and improve current deterministic design efficiency by calibrating load and resistance safety factors proposed in relevant codes.