AI-Based Measurement Integrity Validation for Cyber-Resilient Microgrid Protection
A new research paper proposes an AI-based supervisory measurement integrity validation layer to protect line current differential relays (LCDRs) in inverter-based microgrids from false-data injection attacks (FDIAs). The scheme analyzes short windows of synchronized instantaneous current measurements to distinguish genuine fault-induced trajectories from cyber-manipulated streams, enhancing cyber-resilience of AC/DC protection systems.
Key facts
- arXiv:2604.23666v1
- Line current differential relays (LCDRs) are measurement-driven relays that use time-synchronized multi-phase current waveforms to infer internal faults in AC and DC power networks.
- Inverter-based microgrids rely on digitally communicated measurements, exposing LCDRs to false-data injection attacks (FDIAs).
- FDIAs manipulate remote measurement streams to create protection-triggering but physically inconsistent current trajectories.
- The proposed scheme operates as a supervisory instrumentation layer for modern LCDRs.
- It interprets short windows of synchronized instantaneous current measurements recorded during relay operation.
- The scheme assesses physical consistency to distinguish genuine fault-induced trajectories from cyber-manipulated measurement streams.
- A recurrent neural network is used for the analysis.
Entities
Institutions
- arXiv