Demand-Side and Utility-Side Management Techniques for Increasing EV Charging Load

Abstract

Electricity authorities need capacity assessment and expansion plans for efficiently charging the growing Electric vehicle (EV) fleet. Specifically, the distribution grid needs significant capacity expansion as it faces the most impact to accommodate the high variant residential EV charging load. Utility companies employ different scheduling policies for the maintenance of their distribution transformers (hereinafter, XFR). However, they lack scenario-based plans to cope with the varying EV penetration across locations and time. The contributions of this paper are twofold. First, we propose a customer feedback-based EV charging scheduling to simultaneously minimize the peak load for the distribution XFR and satisfy the customer needs. Second, we present a deep reinforcement learning (DRL) method for XFR maintenance, which focuses on the XFR’s effective age and loading to periodically choose the best candidate XFR for replacement. Our case study for a distribution feeder shows the adaptability and success of our EV load scheduling method in reducing the peak demand to extend the XFR life. Furthermore, our DRL-based XFR replacement policy outperforms the existing rule-based policies. Together, the two approaches provide a complete capacity planning tool for efficient XFR maintenance to cope with the increasing EV charging load.