Enhancing Accuracy in Wind and Photovoltaic Power Forecasting Through Neutral Network and Beluga Whale Optimization Algorithm

Abstract

As the total energy consumption continues to rise, environmental issues and the balance of energy supply and demand have become increasingly severe. With the integration of photovoltaic and wind power generation, wind and solar power generation, while reducing users' carbon emissions, also introduce additional complexity to the power system due to their inherent uncertainty. However, deep learning algorithms have shown tremendous potential in accurately predicting wind and solar power output. This paper combinate the ICEEMDAN algorithm and temporal convolutional neural network to achieve dynamic forecasting of wind and photovoltaic power outputs. In addition, an enhanced beluga whale optimization algorithm is utilized to select optimal hyperparameter combinations for the forecasting model, so that the accuracy and robustness of the model's forecasting outcomes can be enhanced. The research results show that the wind and photovoltaic power forecasting model we constructed can effectively mitigates the impact of volatility and complexity associated with these energy sources on prediction outcomes and achieve accurate forecasting.