Resilience and operational benefits of electric vehicle and grid integration

Shared autonomous electric vehicles (SAEVs) present new opportunities to control and optimize vehicle movements. Future deployment of these vehicles may reduce the need for individuals to own a personal vehicle and can have traffic flow and environmental benefits. However, to fully realize the benefits of this technology, vehicle dispatch needs to be optimized. Fleet operators will want to own and operate as few vehicles as possible while still maintaining a reasonable level of service for passengers. SAEVs can also be used for many purposes beyond moving individual passengers across the system. They could also be used to deliver food, provide last-mile delivery for packages, and interact with the electric grid. These services must be balanced to ensure that as many people are served as possible. SAEV dispatch can be of particular interest in the aftermath of a natural disaster when there may be failures in the electric grid. In this case, vehicles can be used to transport power across broken lines to power critical facilities or reduce the number of blackouts. However, this important service must be weighed against the continued need to provide transportation to critical workers and vulnerable populations that may be reliant on SAEV service. We develop a dispatch policy that is proven to serve all demands (for both electricity and transportation service) if any policy can serve those demand. This maximum throughput policy also enables an analytical characterization of the minimum fleet size (or minimum cost fleet if it can be heterogeneous) such that queues of passengers and energy will remain bounded in the long run. Based on the stable dispatch policy we relax some assumptions and develop a policy that is more realistic for implementation. We pay particular attention to constraints on power flow in the electric grid to ensure realistic charging and discharging behavior (which is important for distribution system service restoration). The analysis and simulation also distinguishes between several potential objective functions which have important equity and stability impacts. We demonstrate how serving passengers from the longest queues first (a technique based on the 'pressure' from maximum stability dispatch) can lead to more equitable outcomes for passengers. Finally, we examine the impact of the time horizon needed for the model predictive control algorithm. A long time horizon is needed to incorporate charging and discharging as well as longer term trends in electric demand. We suggest that future research should examine heuristics to solve this problem more quickly than commercial solvers to enable real-time implementation.