ADMM-based Coordinated Control of Electric Vehicles in an Unbalanced Distribution Grid: Assessing Tolerance to Communication Failures

In this paper, we propose an Alternating Direction Method of Multipliers (ADMM) approach combined with consensus optimization to coordinate electric vehicle (EV) charging and discharging in an unbalanced distribution grid. The underlying optimization problem has an objective function to minimize the customer energy costs, subject to a globally coupled network constraint in the form of a linear inequality (representing grid voltage and transformer core temperature limits) and local polyhedron constraints (representing customer charging requirements and limitations in charging infrastructure capacity). EV coordination is achieved by means of an iterative routine, wherein EVs (agents) exchange limited information with their neighboring EVs through a connected peer-to-peer communication network. Our proposed approach is also designed to accommodate time-varying communication networks arising from communication link faults and random disconnection of agents from communication channels. Numerical simulations verify the performance of our network-aware EV coordination approach to communication-based failures.