Offshore wind turbines (WTs) are crucial in offshore wind energy development. However, the dynamic responses of WTs are subject to significant uncertainties which are usually not properly considered. To the end, this paper proposes an efficient method for quantifying the uncertainties in WTs' dynamic responses based on cumulative distribution function (CDF)-manifold learning. First, a probabilistic model is developed to represent the environmental parameters and sampling for aerodynamic-hydraulic-servo-elastic simulations. Then, the CDF is obtained by statistically analyzing the simulated data. To tackle the higher dimensionality resulting from discretizing the CDF, a manifold learning-based approach is subsequently proposed to reduce its dimensionality and obtain a manifold space. Furthermore, a mapping relation is established between the environmental parameters and the low-dimensional data to efficiently obtain the response CDF under different environmental parameters, leading to the construction of a probability box (P-box) model. To demonstrate the effectiveness of the proposed method, the National Renewable Energy Laboratory (NREL) 5 MW offshore WT on an Offshore Code Comparison Collaboration (OC3) monopile is selected as a case study and analyzed accordingly. The results show P-box models of seven WT responses and validate the effectiveness of the proposed method.
Read full abstract