Neural Network Based Energy Management in Solar PV Powered EV Charging Station
This work presents a new system architecture for a low-cost photovoltaic (PV) battery charging station that can balance: 1) the charging time of each individual battery and 2) the total charging time of all batteries in the system. The control strategy for the new system first charges each individual battery to either the same voltage or same state of charge (SOC) level and then charges multiple batteries in parallel simultaneously. Solar PV system employed with ANFIS MPPT to extract the maximum power from the PV array. Stationary battery storage and EV battery are controlled by a feedback voltage control scheme to maintain dc bus voltage constant. The neural network-based energy management system is proposed for the solar-powered EV charging stations. NN receives two inputs i.e., PV power and soc of the stationary battery, and generates the current reference signal for the grid inverter control. the simulation results are analyzed for different operating modes such as a change in irradiance conditions and different soc conditions of the stationary battery and EV battery. from the test results, energy has been managed effectively using a neural network in a solar PV-powered EV charging station.