Application of BDI-based intelligent multi-agent systems for distribution system service restoration planning



By changing the status of switches on distribution feeders, service to de-energized customers can be restored after an occurrence of a fault. In this paper, multi-agent system (MAS) is applied for the power distribution system service restoration task using local information by utilizing the characteristics of information sharing. Each software agent is running autonomously and is able to exchange information within limited distance. Every agent is capable of performing service restoration such that service restoration can be performed in parallel when a multiple fault scenario occurs. As a result, the outage period can be reduced. BDI (believe–desire–intention) agent architecture is implemented in this paper for group and zone service restoration for power distribution systems. Zone/switch relation table is built and used by agents in order to identify adequate switching actions during the service restoration planning stage. The proposed MAS is implemented on JADE (Java Agent Development) platform. The simulation results show that the proposed BDI-based service restoration agent system is capable of providing adequate restoration switching plan in several scenarios. Copyright © 2010 John Wiley & Sons, Ltd.