Prediction of Electric Vehicle Penetration and Its Impacts on Distribution Systems: A Real-World Case Study in Maryland

With the growing penetration of electric vehicles (EVs) in power distribution networks, electric utilities are facing many challenges. For example, the growing EV charging load may cause degradation in voltage quality, higher network losses, and overloading of equipment. The distribution grid infrastructure needs to be upgraded to handle these problems. In this study, we develop an integrated data-driven planning framework for electric utilities to predict EV adoption and analyze their impacts on distribution feeders. This planning framework consists of two modules. In the first module, we design a generalized Bass diffusion model (GBM), which utilizes historical adoption data, EV availability, incentives, cost, and demographic information to predict EV adoption at the zip code or feeder level. Subsequently, in the second module, we combine the EV adoption prediction and the representative EV charging load profiles to analyze the impacts of EVs on distribution feeders such as voltage violations and equipment overloading. The proposed solution framework can be applied universally to different locations and population groups and was tested in a case study in Maryland, U.S. using real-world data and distribution circuit models. The results feature accurate predictions of EV adoption and reveal when, where, and how severe the voltage violations and overloading issues will be with the growing EV penetration. The proposed framework serves as a valuable tool for system planners to determine distribution system upgrade plans.