Abstract
Despite many advantages, wide-spread integration of electric vehicles (EVs) in the power systems is challenging. Large-scale uncontrolled EV charging load may lead to under-voltage violations, higher power losses, overloading of transformers and transmission lines. In this research work, a machine learning-based communication-free EV charge control strategy is developed to mitigate the issues caused by the uncontrolled EV charging. Another notable feature of the proposed controller is to ensure the fairness among the EVs available at different locations in the distribution system. Moreover, two indices are proposed, that is system utilisation index (SUI) and system average voltage violation duration index (SAVVDI), to assess the performance of any EV charge controller. In the proposed control structure, a nodal voltage and the voltage-to-load sensitivity are measured at each load node which are fed to the EV charge controller. In fact, an upstream node is generally less sensitive to the changes in load, since it is closer to the feeding point, as compared to a downstream node. The output of the charge controller is the charging rate of an EV. In order to validate the robustness of the proposed controller, light and heavy loading conditions are considered which mimic the daily, monthly, yearly and seasonal load variations. Simulation results prove that the proposed controller effectively improves the voltage profiles and ensures fairness among the EVs connected at various charging points in the distribution system. Moreover, it effectively utilises the power system resources.