Predicting Solar Irradiance Using Machine Learning Techniques

Abstract

Machine learning is applied in almost every field of knowledge in recent years. Renewable energy domain now extensively uses machine learning to provide farmers and grid operators with proper and timely information about solar irradiance prediction for future planning. Connecting grids with solar energy systems has brought unprecedented shift in the power sector. Nevertheless, the intervention poses challenges including intermittency that affects the overall energy supply and demand management. Our proposed solution intends to make prediction models using Machine Learning techniques (Linear Regression, Regression Trees and Support Vector Machine) which uses the past weather data including the parameters such as temperature, wind speed, atmospheric pressure, and humidity and predicts solar irradiance for the future. The model can help the grid operators for improved supply and demand management.