HARMONIC PREDICTION FOR ELECTRIC VEHICLES IN DIFFERENT CHARGING CONDITIONS

This paper proposes a machine learning approach to predict the harmonics that occur during charging electric vehicles. Since charging of electric vehicles has a negative impact on the distribution system, this study uses a real-world dataset for harmonic estimation to ensure that all parameters during charging are taken into account and an accurate analysis is performed. An open-source dataset that consist of various charging currents and the resulting harmonics associated with these currents was used. In the study, the effective nonlinear autoregressive exogenous model approach was used. In order to make detailed prediction and to reveal the actual performance of the model, separate applications were made for each harmonic level.