Distributed asynchronous energy management for dynamic large-scale energy networks

The increasing electrification due to Distributed Energy Resources (DERs), Electric Vehicles (EVs), and heat pumps has increased grid volatility. Demand-Side Management (DSM) approaches can help in solving the grid issues. However, DSM approaches require a group of energy agents to coordinate amongst themselves to achieve a global objective conducive to grid health. This paper extends upon previous work and presents an asynchronous and distributed coordination approach amongst a group of agents consisting of event-driven and periodic optimization processes. Specifically, in our approach agents formulate and constantly update a coordination tree in a distributed manner. We use the coordination tree to disseminate steering signals for optimization which the agents use to plan and update the aggregate group power profile. Our contribution deals with unplanned events, such as EV arrivals or departures or a loss in communication with an agent. For the evaluation, we use real EV charging data from a parking lot in the Netherlands. We compare our approach with the offline planning from greedy charging and an optimization algorithm called Profile Steering. Our approach achieves a solution with a deviation of 9.61% when compared to Profile Steering and has a maximum power of 44.64% less than that from greedy charging.