Machine Learning based Solar Power Generation Forecasting with and without MPPT Controller
Debottam Mukherjee, Samrat Chakraborty, Pabitra Kumar Guchhait, Joydeep Bhunia
Abstract
The renewable resources based power generation is unpredictable since it highly depends on the conditions of climate. In India, after wind power, the second largest renewable based power generation is solar power. Therefore, forecasting for solar power generation is necessary since it depends on solar irradiance and temperature. In this paper, forecasting for solar power generation using machine learning has been done with and without using MPPT controller. The study has been done on Badabenakudi, Orissa, India. Machine learning based forecasting techniques has always been proved best than statistical based forecasting techniques. Different machine learning models have been applied on the data set taken. The result shows that Coarse Tree is the best model for solar power generating forecasting with MPPT controller having RMSE of 1.675 and Rational Quadratic Gaussian Process Regression (RQGPR) is the best model for solar power generation forecasting without MPPT controller having RMSE of 1.628.