Wave Prediction Based on Long Short-Term Memory Neural Network

Wave prediction is of great importance in the field of wave energy. Machine learning provides an efficient new idea for wave prediction Due to its strong adaptability and excellent ability to deal with complex data. In this work, the long short-term memory (LSTM) model is investigated under four typical wave cases, including two extreme waves hitting the 50-year contour line of European Marine Energy Centre (EMEC) site, UK with the largest significant wave height and two different wave spectral shapes; another two wave conditions at the targeted resonance condition of a designed wave energy converter (WEC) with two different wave spectral shapes. Physical wave data recorded from wave tank are used to train a LSTM model. Validated with the physical data, the results show that the developed LSTM model has great performance for the prediction of the different waves studied in this work.