Optimal Infrastructure Design of a Public Electric Vehicle Charging Network

Project Type: Research
LoDI Team: Lihui Bai, Sumei Zhang, Shahab Sadri

Overview

LG-&-E and KU (Louisville Gas and Electricity and Kentucky Utilities) is a company which generates, distributes, transforms, and sells electricity generated form gas, oil, hydro energy, and coal sources. The aim of this study is finding the best location within Louisville area to establish EV charging stations.

Purpose and Need

There is an increase trend in adopting electric vehicles (EV). This trend will increase the need for an adequate and sustainable network of charging stations. Strategic decisions on location and capacity of the public charging stations in the urban/suburban as well as highways transportation systems are of great interest to transportation, energy and environment and regional planning authorities. LG-&-E and KU and LoDI conducted this study to first predict the potential EV charging demand in the Louisville area (both urban and surrounding highway systems) and then find out the best locations for new charging stations to satisfy the growing demand.

Methodology

This project designs an optimal EV charging network with a multi-disciplinary approach. Optimization and simulation models are first developed to obtain a set of locations/sites in the study area based on traffic flow analysis, EV ownership distribution and operational requirements for EV charging, among others. Secondly, urban planning analysis is used to refine the set of public EV charging sites based on socio-economic factors, spatial & population factors, real estate market forecasting, community amenities and economic development trends.  The combined approach is applied to both urban/community and highway transportation networks. Particularly, the EV charging network design is modeled as a mixed-integer linear optimization problem (MILP). The optimization model takes forecasted charging demands and returns the optimal strategic location and capacities (i.e., number of chargers) of EV charging stations. To account for the dynamic and stochastic nature of the human-physical system, an agent-based simulation (ABS) model is developed to provide feedback and recommendations on the solutions from the optimization model, based on real-time simulation of traffic and charging activities.

Results and Impacts

Using the city of Louisville, KY, and its surrounding highway system, as two case studies, the project has made recommendations consisting of top 10 locations in the city of Louisville, and 11 highways exits (8 are recommended without reservation, 3 with some reservation) as potential candidate sites for building the EV charging network infrastructure for the study region.