An integrated decomposition algorithm based bidirectional LSTM neural network approach for predicting ocean wave height and ocean wave energy

Abstract

A potential and limitless source of renewable energy is the ocean. Yet, the intermittent and irregular characteristic of wave energy is a cause of concern for the stability of the power system. Thus, wave energy must be anticipated in order to be integrated into power systems. For harnessing energy from ocean waves, ocean wave height is a crucial parameter and due to this accurate and trustworthy forecasts of ocean wave height have gained increased attention recently. This makes it crucial to come up with a forecasting algorithm that is trustworthy, precise and can be used in varied situations. Therefore, in the proposed framework, well known data decomposition algorithm i.e. Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) is combined with Bidirectional Long Short-Term Memory (BiDLSTM) deep learning algorithm to accurately forecast ocean significant wave height and ocean wave energy from the standard datasets acquired from National Renewable Energy Laboratory (NREL) data site. Empirical outcomes produced using the proposed CEEMDAN-BiDLSTM method is found to be highly accurate as compared to the recently established forecasting algorithms in literature.