Abstract
Ocean wave energy is expected to play a significant role as a clean and stable renewable energy source in power generation. However, the intermittency and irregularity of wave energy pose challenges for the safe operation of power grids. Thus, it is essential to forecast wave energy accurately before it is integrated into the power grid. In this study, a novel wave energy forecasting model composed of a two-layer decomposition technique and a long short-term memory network with an attention mechanism was proposed. The two-layer decomposition technique can decompose and reconstruct a wave parameter series into multiple subseries that are more accurately forecastable. This technique incorporates an improved complete ensemble empirical mode decomposition with adaptive noise, variational mode decomposition, and sample entropy algorithms. In addition, a sparrow search algorithm was used to adaptively adjust the parameters of the two-layer decomposition technique to ensure optimal decomposition. The forecasting performance were assessed using four datasets from the US National Data Buoy Center. The results show that the proposed model is superior to seven other well-known forecasting methods and compared with the long short-term memory network; the correlation coefficient of the proposed model increases by 58.72% on average, and the mean absolute percentage error decreases by 75.71% on average.