Resilience assessment framework for cyber-physical distribution power system based on coordinated cyber-physical attacks under dynamic game

TL;DR

This paper proposes a dynamic game-based resilience assessment framework for cyber-physical distribution power systems under coordinated cyber-physical attacks, aiming to optimize defense strategies and evaluate system robustness.

Contribution

It introduces a three-stage defender-attacker-defender game model considering fake base station and physical attacks, with a novel solution method for resilience evaluation.

Findings

Effective defense resource deployment strategies identified.

The framework accurately assesses load loss and service restoration.

Simulation results validate the model's applicability to real-world systems.

Abstract

Owing to the advanced communication networks and intelligent electronic devices, the cyber-physical distribution systems (CPDSs) possess the capability to perform flexible economic dispatch and achieve rapid self-healing from extreme events. Meanwhile, the deep integration of cyber and physical systems makes CPDS vulnerable to coordinated cyber-physical attacks. In this paper, a resilience assessment framework for the CPDS under coordinated cyber-physical attacks is proposed to investigate the impact of the coordinated attacks on load loss and service restoration in CPDS. First, a three-stage defender-attacker-defender dynamic game model considering fake base station (FBS) and physical attacks for CPDS is established, aiming at seeking the optimal defense resource deployment strategy to enhance the resilience of the CPDS. The physical attack is launched to cause faults on the power lines, and the FBS attack is employed to interrupt the service of wireless cellular network to hinder the self-healing process of the CPDS. The lognormal shadowing model and search theory are applied to quantitatively describe the process of the coordinated cyber-physical attacks. Further, the constructed three-stage dynamic game model is equivalently recast as a tri-level max-min-max optimization model, which is solved using column-and-constraint generation combined with enumeration method. Finally, the effectiveness of the proposed resilience assessment framework and solution strategy is demonstrated by conducting simulation analysis on the modified IEEE 33-node CPDS and a real-world 47-node CPDS in China.