The prediction technologies of the power generation from the wave energy converters (WEC) are an urgent and crucial problem in the renewable energy planning, the power grid dispatching and the economic operation. Besides the statistical modelling, this paper presents a novel hybrid DDM for very short term (15 min-4 h) and short term (0-72 h) predictions of the wave energy power, based on the long-short term memory (LSTM) network and the results are compared with those obtained by the Artificial neural networks (ANN) and the support vector machine. The experimental results indicate that the proposed deep learning models enjoy a better performance with a high accuracy in the WEC power prediction than other related models. Furthermore, the proposed DDM methods are shown to be robust and timesaving in training and deployment, with advantages over the statistical methods in very short term and short term WEC power predictions.

short-term prediction  /  wave energy converter  /  data-driven modelling  /  long-short term memory