Research on the Application of Electric Power Big Data in Renewable Energy Power Generation Forecasting

This paper proposes an innovative adaptive integrated neural network model, which combines electric power big data (such as meteorological data, historical power generation data and power grid load data) and uses deep learning technology to accurately predict the power generation of renewable energy. The AI-Net model integrates multiple neural networks structures and adopts an adaptive weighting mechanism to automatically adjust the prediction strategy according to different input data and real-time environmental conditions, thereby improving the adaptability and prediction accuracy of the model. Experimental results show that the AI-Net model performs well in multiple experimental scenarios. Compared with traditional time series models (such as ARIMA) and deep learning models (such as LSTM), the mean square error (MSE) is reduced by about 15% and the mean absolute error (MAE) is reduced by 12%. In addition, the AI-Net model shows higher computational efficiency in real-time prediction tasks and can effectively cope with the processing requirements of large-scale electric power big data. Experiments also show that this method can maintain high prediction accuracy in different renewable energy power generation scenarios such as wind power and solar power, providing strong support for power system scheduling, energy optimization and power grid stability.