Solar Energy Forecasting With Performance Optimization Using Machine Learning Techniques

The objective of this paper is proposed a forecasting technique system constitute the data integrity phase using Machine Learning (ML) techniques and post-processing output optimization to improve the performance. The outcomes demonstrated that the Photo Voltaic (PV) energy manufacture neural network performed more accurate by reducing solar irradiance prediction errors of linear regression characteristics. Through the adoption of both the optimized method settings and chosen function of ML methods were evaluated for predicting the performance. The accuracy of solar radiation estimation could be assessed in terms of statistical error measurement and verification measurements. The approach was tested in two climates hot and cold and the findings demonstrated using absolute error percentage with a contribution of 6.3% and 4.7%, respectively to improve the performance of the system based on data-driven ML algorithms and analytical post-processing.