Prediction of Global Whipping Responses on a Large Cruise Ship Under Unknown Sea States Using an LSTM Based Encoder-Decoder Model

Global whipping responses, characterized by high-frequency vibration components on the hull girder, are typically induced by stern and bow slamming loads. These responses contribute to an increase in vertical bending moments, making their accurate prediction crucial for ship safety. In this study, we build upon our prior work to establish an encoder-decoder model based on Long Short-Term Memory (LSTM), designed to predict whipping responses under varying sea states. The model is trained on a comprehensive dataset, which includes motion data and vertical bending moment history of a large cruise ship under various sea states. This dataset is established via numerical simulation, ensuring a wide range of scenarios for the model to learn from. The efficacy of the LSTM encoder-decoder model in capturing global whipping responses is initially verified under a single sea state case. This step confirms the model’s ability to accurately predict vertical bending moments under known conditions. Subsequently, the model’s performance under unknown sea states is examined. Given that the distribution of training data significantly influences the model’s performance under unknown sea states, a data mixing strategy is employed during the training process in this scenario. The results indicate that the LSTM encoder-decoder model effectively captures whipping responses. Furthermore, the data mixing strategy significantly improves prediction accuracy under unknown sea states, demonstrating the potential of this approach in enhancing ship safety.