Short-Term EV Charging Demand Forecast with Feedforward Artificial Neural Network

The global increase in greenhouse gas emissions from automobiles has brought about the manufacture and usage of large quantities electric vehicles (EVs). However, to ensure proper integration EVs into grid, there is a need forecast charging demand accurately. This paper presents short-term vehicle using feedforward artificial neural network optimized with modified local leader phase spider monkey optimization (MLLP-SMO) algorithm, proposed variant optimization. A proportionate fitness selection employed improve update process algorithm trains demand. effectiveness forecasting model was tested validated public data United Kingdom Power Networks Low Carbon London Project. model's performance compared trained particle swarm optimization, genetic classical two conventional models, multi-linear regression Monte Carlo simulation. assessed mean absolute percentage error accuracy. produced accuracy 99.88% 3.384%, respectively. results show that MLLP-SMO as trainer predicted better than other models met industry standard