PI: Marco Pavone; Co-PI: Ram Rajagopal; Student: Justin Luke
Electrification and autonomy are driving down the total cost of ownership for vehicle fleets. Presently, autonomous electric vehicles (AEV) are being developed for fleet applications such as passenger mobility, delivery, and medium-range trucking. Autonomous fleets have the advantage of highly controllable routing and charge scheduling compared to privately owned human-operated EVs, however, the grid impact for AEV fleets is not well understood. With intelligent control, AEV fleets can not only avoid negative social externalities like exacerbating road congestion and overloading grid infrastructure, but also play an active role in increasing transportation system efficiency and providing grid services.
This project develops a tool that computes the optimal placement and charging rate of EV charging stations in an urban setting with high-penetration of autonomous EVs and PV solar installations. The optimal siting is determined jointly with the routing and charging of the fleet as a network flow problem, seeking to minimize infrastructure costs, electricity costs (including demand charges), and travel costs while observing road and power flow constraints. Following, the optimization will integrate with the cloud-enabled grid overlay project which will communicate dynamic charging bounds to the fleet operator, allowing for online operation of the fleet while obeying grid constraints without needing to compute grid power flow in real-time.
Charging infrastructure is the coupling link between power and transportation networks, thus determining station placement is necessary for any planning of power and transportation systems. This tool will be purposed to provide rigorous analysis and simulation results to inform key industry players and policymakers in both systems.