Key points are not available for this paper at this time.
Fatigue assessment for floating wind turbines is commonly done through comprehensive simulation studies of integrated time-domain simulations. Procedures which incorporate simplifications of the environmental conditions in order to limit the number of simulations typically lead to more conservative designs. An alternative approach is proposed here based on response surface modeling using Latin Hypercube Sampling and artificial neural networks. The presented method takes into account the statistical characteristics of environmental parameters during the system’s life time (resulting in more realistic and accurate damage calculations) while keeping the numerical effort to a minimum. The procedure is exemplified using a hypothetical site presented linked to the H2020 project LIFES50+ site A. The considered system is a concrete semi-submersible on which a 10 MW turbine is mounted. The system is implemented in the simulation tool FAST. The effect of different numbers of samples is analyzed and a final comparison to a conventional procedure is performed. Results indicate a reduction in the predicted value of expected lifetime damage compared to conservative estimates. More research is necessary in determining the quality of the response surface, in order to be able to fully evaluate the performance of the method.
Müller et al. (Sun,) studied this question.