Improving the Resilience of Modular Multilevel Converters using Concurrent Estimators
Due to the widespread employment of power electronic devices with complex and occasionally remote/cloud control or monitoring in the future power systems, their secure operation would play a significant role in the resilient supply of critical loads. The modular multilevel converter (MMC) is a well-established topology in high and medium voltage levels due to its inherent redundancy and numerous technical advantages. This study investigates the operation of MMCs under false data injection attacks (FDIA) and proposes a practical solution for detecting and mitigating them. We propose to employ two concurrent estimators, module level and arm level using Kalman filter and neural networks, in order to improve the resilience of the MMC and compensate for the shortcomings of the detection methods presented in the literature against sophisticated attacks. The simulation study clearly proves the feasibility of the proposed dual estimator technique, and thus its contribution to improving the cyber resilience of MMCs.
