Combined Forecasting Model of Tidal Energy Based on Improved Variational Mode Decomposition and Bidirectional Long-Term and Short-Term Memory Network
Junsen Huang, Guohui Li · 2025
Accurate and reliable tidal energy prediction is very important for the study of open ocean activities. Aiming at the nonlinear characteristics of tidal energy data, combined forecasting model of tidal energy based on improved variational mode decomposition and bidirectional long-term and short-term memory network is proposed. Aiming at the problem that the penalty coefficient and decomposition level need to be set manually in variational mode decomposition (VMD), VMD based on lotus effect optimization algorithm (LEA) is proposed, named LEAVMD. Firstly, the tidal energy data are decomposed by LEAVMD, and several Intrinsic Mode Functions (IMFs) are obtained. Then, BiLSTM is used to predict IMFs and residuals. Finally, the predicted values of different components and residuals are superimposed to obtain the final prediction result. The data of the Caribbean Sea are selected from the National Oceanic and Atmospheric Administration of the United States, and the time is from January 2023 to January 2024. The sampling interval is one day, and 2 to 4 data are sampled at a time, including tidal peaks and troughs. The results show that the root mean square error, mean absolute error, mean absolute percentage error and determining coefficient are 0.1965,0.1399, 0.3589 and 0.9721, respectively. The indexes of the proposed model are all better than those of the comparison model, which proves that the proposed model has higher prediction accuracy and provides an effective method for tidal energy prediction.