Electric Vehicles Uncertainty Forecasting

The growing integration of electric vehicles has attracted a lot of interest. However, they are highly affected by EV charging uncertainties and are, therefore, difficult to forecast accurately. This paper presents an Artificial Neural Network (ANN) method ANN to forecast electric vehicle uncertainties. ANN was trained using historical data from a residential building, such as arrival time, departure time and initial SOC. Then, it was tested during 24 hours through different scenarios. For each one of the cases, the model’s accuracy was assessed by comparing historical data to forecast information. The associated errors were also calculated. The outcomes reveal that the suggested forecasting method is very effective in reducing EV forecasting errors and, as a result, is better at regulating EV uncertainty.