Multi-Agent Based Cyber Attack Detection and Mitigation for Distribution Automation System

Abstract

As electric power became an essential part of daily life, resiliency and reliability of operation became important. A distribution management system (DMS) enables real-time monitoring and dynamic controls of the power distribution networks. As such, their controls should be designed to be resilient against distribution grid disturbances and cyber events. Due to the high penetration of Information and Communication Technology (ICT) in the DMS, the reliability of power distribution grids is highly dependent on the cyber system. However, most power distribution applications, communication protocols, and devices are vulnerable to the cyber attacks since they are designed and implemented before cybersecurity became a critical issue. Recent reports and cyber attack incidents clearly indicate that cyber-attacks are increasingly likely on the power system infrastructures, e.g., control centers, nuclear power plants, and substations. These attacks may cause significant damages to the power grid. Cybersecurity research for the power distribution grid is a high priority subject in the emerging smart grid environment. This article proposes cybersecurity enhanced distribution automation system (DAS) with a multi-agent system (MAS). The proposed multi-agent based cyber attack detection and mitigation algorithms can identify the anomalies, abnormal activities, and unusual system operations of the DAS. The proposed algorithms have been applied to the existing communication protocols, protection schemes, and restoration applications. The result is validated with a testbed and proposes a new integrated tool for the detection and mitigation of cyber intrusions at a power distribution grid with multiple feeders.

Keywords

• Cybersecurity of distribution automation system
• multi-agent based system
• fault detection
• isolation and detection