Distributed multi-agent microgrids: a decentralized approach to resilient power system self-healing

Abstract

The predominance of recent self-healing power system research has been directed towards centralized command and control functions. In this paper, a decentralized multi-agent control method for distributed microgrids is introduced. Given the complexity of a large power system spanning hundreds of miles and comprised of numerous microgrids, it is potentially unrealistic to expect that centralizing total system control functions is feasible. Therefore, the authors are particularly interested in dispersing decision-making by utilizing smart microgrid control agents that cooperate during normal and emergency situations. The combination of microgrids and agent-based control can improve power system resiliency. The method described herein lays the groundwork for a comprehensive microgrid control architecture that strikes a balance between the multiple intra-microgrid objectives defined by local operator and the situational demands of the microgrid collective as part of the power system. In this way, both self-interest and cooperation can arise, allowing microgrid agents to successfully transition from normal operations to an emergency condition and back again when conditions have resolved, independent of a central supervisor. The decentralized multi-agent methods for microgrids explored in this paper help to support what may be an enabling technology of future smart grids.