Short-term prediction of mooring tension for floating offshore wind turbines under typhoon conditions based on the VMD-MI-LSTM method

The safety of mooring systems for floating offshore wind turbine platforms is critical for their normal operation. During extreme weather events such as typhoons, ensuring the integrity of mooring lines becomes a paramount concern. With advancements in artificial intelligence technology, the integration of deep learning methods for short-term prediction of mooring line tension under typhoon conditions has introduced innovative solutions to address this safety issue. In this study, the proposed VMD-MI-LSTM neural network is employed to forecast mooring line tension under typhoon conditions over short periods. The platform model utilized in this research is the 5-MW Braceless semisubmersible platform, with the transient wind fields of Typhoon Hagibis serving as the research scenarios. Through fully coupled simulations, the tension of mooring lines under typhoon conditions is computed. Using wave height time series and typhoon wind speed as input data and mooring line tension data as output, a dataset is constructed. The optimal model parameters are determined through exploration of the hyperparameter space to develop the multi-input long short-term memory (MI-LSTM) mooring line tension prediction model. An analysis of the prediction results for mooring line #1 is conducted. Given the similarity of environmental conditions across different platform mooring lines, the model's universality is evaluated by predicting mooring line #1 and comparing it with the VMD-MI-LSTM model. This comparison highlights the optimization effect of the VMD variational mode decomposition method. This study provides short-term predictions of mooring line tension under typhoon conditions. By integrating with the mooring line adjustment system, effective adjustment of the mooring system of the floating wind turbine platform can be achieved under extreme environmental conditions, thereby enhancing the platform's safety and resilience against risks.