A Computationally Efficient Consensus-Based Multiagent Distributed EMS for DC Microgrids

Abstract

In this article, we present a fully distributed optimization approach for real-time energy management of dc microgrids. In the proposed approach, the energy management system (EMS) is formulated with a consensus algorithm based on incremental cost, which operates in a distributed manner for both generation and demand-side management. The proposed algorithm is privacy-preserving, scalable, plug-and-play capable, and robust against time-varying communication topologies. Furthermore, it converges very fast in terms of computation time and iterations. The scalability of the proposed algorithm was evaluated using IEEE bus systems in various sizes, number of agents, and communication topologies. A cooperative multilayered EMS optimally coordinated by multiple agents was also proposed in this article for integration. The effectiveness of the proposed system was validated through computer simulations and hardware experiments.