Decentralized multi-agent system applied to the decision making process of the microgrid restoration procedure towards sustainability

Abstract

A significant procedure to ensure the consumer supply is Power System Restoration (PSR). Due to the increase of the number of distributed generators in the grid, it is possible to shift from the conventional PSR to a new strategy involving the use of distributed energy resources (DER). In this paper, a decentralized multi-agent system (MAS) is proposed to cope with the restoration procedure in a microgrid (MG). Each agent is assigned to a specific consumer or microsource (MS), communicating with other agents at every stage of the restoration procedure so that a common decision is reached. The 0/1 knapsack problem is the problem that every agent solves to determine the best load connection sequence during the restoration of the MG. Two different case studies are used to test the MAS on a dynamically modeled benchmark MG: a total blackout and a partial blackout. Regarding the partial blackout case, demand response emergency programs are considered to manage the loads in the MG. The MAS is developed in Matlab/Simulink environment and by performing the corresponding dynamic simulations it is possible to validate this system towards sustainability.