The Electric Vehicle Requested Energy Predictions Using Machine Learning Algorithms for the Demand Side Management

The concept of demand side management involves the management and prediction of coming energy consumption. There are several factors that can affect EV charging demand, including the number of EVs in the area, the availability of public charging stations, the time of day, and local driving patterns. To make accurate predictions, it is important to consider as many of these factors as possible. Some techniques that can be used to predict EV charging demand include machine learning algorithms, statistical modeling, and simulation. In this paper, we start by addressing the problem of demand side management, focusing on the prediction of requested energy by EV users to better understand coming loads. To provide our unique contribution, we then work on predicting the requested energy of EV users, using new algorithms to predict EV charging behavior including LR, RF, Ridge Regressor, and SGDRegressor. We use the programming language Python and the scikitLearn library. Our prediction work aims to make it simpler for distribution operators to manage energy generation and consumption to avoid imbalances such as overloads or over-generations that can damage grid infrastructure. Our results indicate that SGDRegressor outperforms all other algorithms by showing the best metric values compared to other AI prediction algorithms, making it more accurate than other algorithms in these applications.