Publications

Abstract : As solar energy is one of the most abundant sources of renewable energy, there has been substantial fluctuations in the energy supply from PV systems when used with power grids. Solar power output depends on various factors such as cloud cover, temperature & relative humidity. By using the meteorological data to predict solar power generation, solar power becomes a reliable energy resource. This study investigates the use of Support Vector Regression for solar power prediction, which performs better than other machine learning algorithms. A variety of parameter tuning methods, such as Random search, Grid search and Tree based optimization tools are implemented to obtain a robust model that can identify the best model that will yield the least error when predicting the solar power generation.