Computational Tools for Battery-Electric Bus Systems: From Infrastructure Planning to Daily Operations

Abstract

Public transit operators that phase out vehicles powered by fossil fuels in favor of battery-electric buses must confront the major challenges of electric vehicles: limited driving range and significant recharging times. Buses can extend their effective range with supplemental "opportunity" charging in between passenger trips, but these operations must be managed carefully to avoid increasing delays and degrading passenger experience. Accordingly, this dissertation develops computational tools to anticipate and manage recharging needs for the complete transit system. The first tool is a public web app that processes General Transit Feed Specification data to calculate key metrics for fleet electrification planning. This processed data then serves as an input to two optimization models to design opportunity charging infrastructure and schedule its daily utilization, with an emphasis on maintaining the quality of passenger service across a broad range of operating conditions. The models' performance under real-world variations in schedule adherence and energy consumption is evaluated with a customized simulation platform based on operations data from King County Metro. The results suggest that if moderately conservative estimates of energy consumption and trip durations are used in the planning models, buses can maintain good passenger service under varying conditions.