Short-Term Forecasting of Total Aggregate Demand in Uncontrolled Residential Charging with Electric Vehicles Using Artificial Neural Networks

Electric vehicles are gaining focus and adoption, with new users every day. Their widespread use introduces a new scenario and challenge for the electrical power system due to the high energy storage demands they create. Predicting these loads using artificial neural networks has proven to be an efficient approach for solving time series problems. This work utilizes a multilayer perceptron network through supervised backpropagation training with Bayesian regularization to enhance generalization, reducing overfitting errors. The research aggregates actual consumption data from 200 households and 348 plug-in electric vehicles in the training and forecasting process. To validate the developed method, MAPE was used. Short-term forecasts were made across the four seasons, predicting the total aggregate demand from homes and vehicles for the next 24 hours. The methodology demonstrated significant and relevant results using hybrid training for this problem, with the potential for real-world application.