Short-term wind speed prediction based on temporal convolutional networks

To improve the utilization efficiency of wind energy, this research proposes a hybrid model based on the Temporal Convolutional Network (TCN) and two-level wind speed decomposition. Firstly, the original wind speed data is decomposed into main and residual signals through Singular Spectrum Analysis (SSA). Then, the usage of Variational mode decomposition (VMD) decomposes the residual signals into several sub-sequences. The next step involves predicting the main wind speed signal and all the sub-sequences using TCN. Eventually, the Grey Wolf Optimizer (GWO) is employed to perform optimization on the stack of prediction results, resulting in the wind speed prediction outcomes. The research results demonstrate that the proposed hybrid SSA-VMD-TCN-GWO model outperforms the reference models. Thus, this research provides a new solution to wind speed prediction