Short-term photovoltaic power forecasting of improved humpback whale migration algorithm

To achieve accurate prediction of short-term photovoltaic power generation, based on Variational Mode Decomposition (VMD), Improved Whale Migration Algorithm (IWMA), Convolutional Neural Network (CNN), and Kernel Extreme Learning Machine (KELM), a novel photovoltaic power generation power prediction model (VMD-CNN-IWMA-KELM) is proposed. Firstly, VMD is utilized to decompose the original power sequence to overcome the influence of noise and randomness. Subsequently, a CNN-KELM hybrid model is constructed, and the IWMA algorithm is introduced to optimize its hyperparameters, thereby enhancing the predictive accuracy of the model. Finally, individual VMD-CNN-IWMA-KELM models are constructed for each stationary component to conduct training and prediction. The predictive results of all components are then reconstructed and superimposed to obtain the final photovoltaic power generation forecast. Experimental results demonstrate that, compared to existing models, VMD-CNN-IWMA-KELM achieves superior prediction accuracy and more accurately characterizes the variation trend of photovoltaic power output.