Enhancing Solar Power Forecasting in Multi-Weather Conditions Using Deep Neural Networks

Renewable energy, with its abundant and sustainable potential, has emerged as a critical component of the global energy landscape. The need for accurate power forecasting is evident due to the demands of integrating renewable energy into existing grids, increasing load management, managing resources efficiently, and reducing environmental impact. To address this need and introduce a new era of even more accurate solar power predictions, this research paper presents a comprehensive investigation. In this pursuit, we introduce an innovative Deep Neural Network (DNN) and thoroughly assess its performance against benchmark machine learning models, including Gradient Boosting Regression (GBR), Random Forest Regression (RFR), and Support Vector Regression (SVR). To train and test the models real-world solar power generation data from the Quaid e Azam Solar Park is used. To achieve more accurate models the data is divided into diverse weather conditions sunny, cloudy, and rainy days. Each model is trained and tested separately. The results decisively demonstrate that the proposed DNN outperforms all other models, offering a promising avenue to revolutionize solar power forecasting and enhance the harnessing of renewable energy sources.