Wind Power Forecast Based on Multi-Source Data and RNN: A Case Study of Jiangsu

The unpredictability and intermittency of reneable energy sources have made it increasingly difficult to balance the power system, threatening its stability and sustainability. This study developed a multi-source Recurrent Neural Network (RNN) model aimed at improving the accuracy of wind power generation forecasts while mitigating the risks of overfitting in small datasets. Leveraging historical data, meteorological conditions, and electricity load data, we constructed a large-scale wind power generation dataset for Jiangsu Province, encompassing information from 116 district-level cities. Experimental results demonstrate the superior performance of the multi-source RNN model in wind power generation forecasting compared to traditional models. Our research underscores the potential of deep learning techniques in addressing the challenges of renewable energy and offers a practical pathway toward building a more resilient and efficient power system.