Selecting Layover Charging Locations for Electric Buses: Mixed-Integer Linear Programming Models

Abstract

Public transit agencies across the United States are rapidly converting their bus fleets from diesel or hybrid powertrains to battery-electric propulsion systems. To realize the benefits of this transition while retaining acceptable quality of service and limiting capital costs, agencies must intelligently decide where to locate recharging infrastructure. While most agencies electrifying their fleets plan to install chargers at bases where buses are kept overnight, a question faced by many fleet operators is where to install layover chargers that provide additional energy while buses are in operation during the day. To address this challenge, this thesis presents two mixed-integer linear programming models that optimize the tradeoff between upfront charging infrastructure costs and operational performance in the form of trip delays and recovery times. A discrete-event simulation model is also developed to accurately quantify queue delays at heavily used chargers and better evaluate system performance under real-world variations in key parameters such as bus energy consumption per mile. The models are applied to a case study of South King County, WA, where an electric bus deployment is planned in the near future. The results show that the models are effective at identifying sensible locations and ensuring that buses can charge without incurring additional delays.