Prediction of 6-DOF motion response of semi-submersible floating wind turbine in extreme sea conditions using OVMD-FE-PSO-LSTM methodology

Abstract

The motion response of offshore floating wind turbines significantly influences their structural integrity, power generation efficiency, operational complexity, safety, and stability. Therefore, predicting the motion response of offshore floating wind turbines is of paramount importance. In engineering practice, especially in extreme marine environments, the motion of wind turbines becomes more complex, making accurate prediction more challenging. In this era of rapid development in deep learning technology, some solutions have emerged for this problem. In this paper, we propose a hybrid model, namely the OVMD-FE-PSO-LSTM model. We begin by conducting numerical simulations of a 5 MW-OC4 semi-submersible floating wind turbine in extreme sea conditions, obtaining motion data for the turbine’s six degrees of freedom. We then decompose the initial motion data using an optimized traditional VMD method, assess the modal complexity with the FE method, combine modal components with similar complexity to reduce computational load, and make predictions using the PSO-LSTM model. Finally, we analyze and compare the predictive results of different models. The results demonstrate that the proposed hybrid model outperforms other comparative models in terms of accuracy, providing new insights into the prediction of the motion response of offshore floating wind turbines.