A multi-agent privacy-preserving energy management framework for renewable networked microgrids

Abstract

This paper proposes a fully distributed scheme to solve the day-ahead optimal power scheduling of networked microgrids in the presence of different renewable energy resources, such as photovoltaics and wind turbines, considering energy storage systems. The proposed method enables the optimization of the power scheduling problem through local computation of agents in the system and private communication between existing agents, without any centralized scheduling unit. In this paper, a cloud-fog-based framework is also introduced as a fast and economical infrastructure for the proposed distributed method. The suggested optimized energy framework proposes an area to regulate and update policies, detect misbehaving elements, and execute punishments centrally, while the general power scheduling problem is optimized in a distributed manner using the proposed method. The suggested cloud-fog-based method eliminates the need to invest in local databases and computing systems. The proposed scheme is examined on a small-scale microgrid and also a larger test networked microgrid, including 4 microgrids and 15 areas in a 24-h time period, to illustrate the scalability, convergence, and accuracy of the framework. The simulation results substantiate the fast and precise performance of the proposed framework for networked microgrids compared with other existing centralized and distributed methods.