Predicting Solar Irradiance in Saudi Arabia via Machine Learning & Artificial Neural Networks for Efficient Grid Integration

Abstract

Solar energy is becoming an essential part of the energy mix of several counties. However, there are challenges with implementing such renewable sources including intermittency, difficulty in managing energy flux, integrating & operating the power grid, and the implicit relationship between weather parameters and solar irradiance requiring data analysis techniques to be implemented. These techniques would uncover the hidden patterns and correlations between weather features (e.g., humidity, temperature and solar irradiance) to enhance solar irradiance prediction accuracies for efficient planning of electricity production of solar panels connected to the grid. Saudi Arabian weather features will be employed for prediction assessments of machine learning & artificial neural network algorithms with multiple data sources & public domain entities including Photovoltaic Geographical Information System (PVGIS), King Abdullah City for Atomic and Renewable Energy (KACARE), and the National Renewable Energy Laboratory (NREL).