Distributed Stochastic Model Predictive Control for Peak Load Limiting in Networked Microgrids With Building Thermal Dynamics

Abstract

Today’s increasing in demand for renewable energy resources has attracted a lot of attention towards Microgrids (MGs). Due to their reliability, computation power and performance, a distributed network of MGs seems to be a promising approach to integrate a large number of distributed energy resources, energy storage systems and loads. However, the local privacy, the consensus process and the stochastic behaviors of renewable energy resources are some of the challenges which should be addressed before a wide deployment of such networks. In this paper, we introduce a distributed MGs integrated with buildings, focusing on their ability to provide peak load limiting. The algorithm is formalized as a two-stage stochastic problem, where the first stage variables determine the next time-step buildings’ temperatures setpoints in each microgrid, while the second-stage variables define the power exchange decisions for the purpose of limiting peak load in the network of microgrids. Leveraging by alternating direction method of multipliers (ADMM), the proposed framework can work in a real-time environment as a supervisory controller to coordinate MGs.