An Analytical Parameter-Free Cyberattack Detection Method for Grid-Connected Converters

The control unit, a critical component of power electronic converter-based systems, interfaces with monitoring and measurement units via data communication links, a configuration that increases the risk of system penetration. Such vulnerabilities could compromise the security of the communication links between control units and physical components, like sensors, by allowing data manipulation. This article introduces an analytical-based approach to detect such cyberattacks, designing data integrity attack scenarios that emulate the actual grid events. It conducts a mathematical analysis of the control system behavior during actual events and cyberattacks, extracting essential features for attack detection. An extended state observer is developed to estimate active and reactive power, mitigating noise effects, and eliminating parameter dependency in feature calculation. Unlike model-based methods, the proposed approach is devoid of convergence issues and parameter dependence. It also surpasses data-driven methods by eliminating the need for preprocessing and model training. The effectiveness of this novel method in detecting cyberattacks is demonstrated through simulation and experimental studies.