Scheduling of Electric Vehicle Charging in a Solar-Wind Powered Microgrid Using Machine Learning

The future of power generation lies in renewable energy, and solar and wind power are major contributors to meeting the ever-increasing energy demand brought on by the development of new technologies like electric vehicles (EVs). To address the rising energy demand, a machine learning-based approach has been explored in this research, which predicts the most practical hour for EV charging using renewable energy sources. Power generation throughout the year has been predicted using wind-solar power generation datasets. The research was conducted in two stages: applying machine learning algorithms and time series analysis. We have compared the R-squared value, correlation, and various types of errors using five machine-learning algorithms. With an R-squared value of 0.9574, the XGB Regressor model outperformed the other four machine learning algorithms in predicting solar wind power generation for EV charging demand management. On the other hand, our time series study reveals that when the sum of these two powers (solar, wind) reaches its peak between 8 AM and 8 PM over 24 hours, EV charging demand stays at its peak, suggesting the demand handling capacity of EVs by renewable energy sources.