Integration of Electric Vehicle Charging Stations into the Unit Commitment Modeling

The rapid proliferation of plug-in electric vehicles (PEVs) has resulted in a significant impact and load being placed on the nation’s power grid as a result of the charging of a number of PEVs, especially at the low and medium voltage distribution infrastructure. This creates a significant obstacle to the further mass distribution of electric vehicles (EVs). In this study, the effectiveness of the EV charging stations on the short-term power system planning and control is being studied, and the dynamic programming-genetic algorithm (DP-GA) as a hybrid optimization approach is being used to minimize the power system operational costs. IEEE 14-bus test system is being used to evaluate the suggested methodology. Long short-term memory (LSTM) as a deep learning approach is being utilized to forecast the day ahead performance of EV charging stations in order to be integrated with the power grid. The result shows that the mean square error (MSE) of LSTM is around 0.045 for EV prediction to achieve the plan of UC with minimum production costs of 514707${\$}$ for the day ahead.