Distributed Resilient Consensus and Demand Tracking in Battery Energy Storage Systems Under Adversarial Attacks
Shiheng Zhang, Yiding Ji · 2025
Battery Energy Storage Systems (BESS) are critical in balancing power supply and demand by dynamically adjusting charging and discharging power. However, their deployment in public networks renders them vulnerable to adversarial attacks, which can disrupt system coordination and potentially lead to failures. In order to resolve these challenges, this work develops a resilient consensus method that integrates the Mean Subsequence Reduced with demand tracking, structured within a leader-follower control framework. The proposed algorithm guarantees that all non-adversarial agents achieve resilient state-of-charge (SoC) consensus and equitable power distribution, even under malicious battery storage unit conditions. Additionally, we introduce an error tracking factor for leader agents to facilitate accurate demand tracking by the BESS. We establish convergence conditions, demonstrating that the system converges to a final value determined by the communication graph, initial values, and BESS parameters. The performance of our approach is validated by a simulation under adversarial conditions, confirming its robustness and reliability.