AK-MDAmax: Maximum fatigue damage assessment of wind turbine towers considering multi-location with an active learning approach
Chao Ren, Yihan Xing
Abstract
Lifetime fatigue damage prediction plays a key factor in wind turbine structure’s reliability assessment. However, the damage estimation of wind turbines requires thousands of simulations and significant computational costs. To address this problem, this paper proposes an efficient active learning Kriging named AK-MDAmax for estimating the maximum fatigue damage of wind turbine towers with less computational cost. The proposed AK-MDAmax approach is based on the previous AK-DA approach. Kriging models are used to estimate the fatigue damage of wind turbine towers at different wind-wave conditions. An efficient active learning approach is developed to assess multi-location maximum cumulative fatigue damage. One 15MW Semi-submersible floating wind turbine model from the IEA project is used to demonstrate the efficiency of the proposed approach. Results indicate the proposed approach can efficiently and accurately estimate wind turbine towers’ maximum cumulative fatigue damage. The AK-MDAmax approach requires less than 3% of the computational effort compared with the typical simulation approach, and the related absolute error is less than 1%. The AK-MDAmax approach could be useful for designers to optimize wind turbine structures and reduce design time and costs.