False Data Injection Attack Resilient Microgrid Protection Scheme Using Critical Sensor Identification and Gated Recurrent Unit Classifier

The increasing integration of information and communication technologies for enhanced microgrid monitoring has inadvertently expanded the system’s exposure to cyber threats, particularly false data injection attacks (FDIAs). As a critical function in cyber-physical microgrids, line fault protection is a prime target for attackers aiming to compromise overall grid stability. By manipulating sensor data inputs to the protection algorithm, adversaries can misrepresent fault conditions—either masking actual faults or fabricating false ones. To address this challenge, this paper presents a fault detection, classification algorithm designed with enhanced resilience to data-driven cyberattacks. The proposed methodology integrates critical sensor selection through constrained optimization with a multistage classification framework based on maximum overlap discrete wavelet transform and gated recurrent unit classifiers. This approach not only ensures accurate fault diagnosis but also identifies FDIAs effectively. Moreover, leveraging data exclusively from optimally selected sensors reduces deployment costs by 64.71% and improves the cyber-resilience of protection algorithm.