Artificial Intelligence-Based Electric Vehicle Charging Station Load Forecasting Scheme for Smart Grid System

The electrification and evolvement of intelligent transportation systems (ITS) have proved to be a breakthrough paradigm for adopting the indispensable benefits of electric vehicles (EVs) in the automotive industry. This necessitates intelligent energy management during the communication between the EVs and the charging stations (CS), which is one of the critical concerns due to the huge electricity demand for EVs. Thus, many authors have adopted the smart grid as an intelligent power distribution infrastructure, which requires the EV CS load forecasting to analyze the energy consumption at CS. Therefore, we propose an artificial intelligence (AI)-based EV CS load forecasting scheme adopting the benefits of smart grid environment. Consequently, we foremost consider EV charging data to predict state-of-charge (SoC) using an AI-based sequential model based on that CS issues an energy request to the smart grid. For that, we contemplate considering CS data and predicting the energy usage of different locations based on various parameters using a sequential model. Thus, the proposed EV CS load forecasting facilitates efficient energy transfer from the smart grid to CS for optimal EV charging. The performance evaluation of the proposed scheme is analyzed considering the EV charging dataset with metrics such as EV SoC prediction comparison, error prediction with battery voltage, and mean square error (MSE (0.0007)), mean absolute error (MAE (0.019)), and error prediction with charging time for CS dataset in which Adam optimizer outperform other optimizers (RMSprop and Adadelta) attaining the efficient load forecasting.