Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cyber-Resilient Fault Diagnosis Methodology in Inverter-Based Resource-Dominated Microgrids with Single-Point Measurement

Yifan Wang, Yiyao Yu, Yang Xia, Yan Xu

TL;DR

FO-MADS tackles cyber-physical threats in inverter-based microgrids with only a single VPQ sensor by combining dual fractional-order features derived from Caputo and Grünwald-Letnikov derivatives to magnify both fast transients and slow drifts. A two-stage hierarchical classifier localizes the faulty inverter and isolates the faulty IGBT switch, while Progressive Memory-Replay Adversarial Training (PMR-AT) with Online Hard Example Mining enhances robustness against diverse cyber-attacks. Validation on a four-inverter testbed shows high accuracy under bias, noise, data replacement, and replay attacks, and near-attack-free performance, demonstrating a cost-effective, deployment-ready approach to cyber-physical resilience. The work offers a practical path to resilient operation of IBR-dominated microgrids with reduced sensing requirements and integrated defense against adversarial disturbances.

Abstract

Cyber-attacks jeopardize the safe operation of inverter-based resource-dominated microgrids (IBR-dominated microgrids). At the same time, existing diagnostic methods either depend on expensive multi-point instrumentation or stringent modeling assumptions that are untenable under single-point measurement constraints. This paper proposes a Fractional-Order Memory-Enhanced Attack-Diagnosis Scheme (FO-MADS) that achieves timely fault localization and cyber-resilient fault diagnosis using only one VPQ (voltage, active power, reactive power) measurement point. FO-MADS first constructs a dual fractional-order feature library by jointly applying Caputo and Grünwald-Letnikov derivatives, thereby amplifying micro-perturbations and slow drifts in the VPQ signal. A two-stage hierarchical classifier then pinpoints the affected inverter and isolates the faulty IGBT switch, effectively alleviating class imbalance. Robustness is further strengthened through Progressive Memory-Replay Adversarial Training (PMR-AT), whose attack-aware loss is dynamically re-weighted via Online Hard Example Mining (OHEM) to prioritize the most challenging samples. Experiments on a four-inverter IBR-dominated microgrid testbed comprising 1 normal and 24 fault classes under four attack scenarios demonstrate diagnostic accuracies of 96.6% (bias), 94.0% (noise), 92.8% (data replacement), and 95.7% (replay), while sustaining 96.7% under attack-free conditions. These results establish FO-MADS as a cost-effective and readily deployable solution that markedly enhances the cyber-physical resilience of IBR-dominated microgrids.

Cyber-Resilient Fault Diagnosis Methodology in Inverter-Based Resource-Dominated Microgrids with Single-Point Measurement

TL;DR

FO-MADS tackles cyber-physical threats in inverter-based microgrids with only a single VPQ sensor by combining dual fractional-order features derived from Caputo and Grünwald-Letnikov derivatives to magnify both fast transients and slow drifts. A two-stage hierarchical classifier localizes the faulty inverter and isolates the faulty IGBT switch, while Progressive Memory-Replay Adversarial Training (PMR-AT) with Online Hard Example Mining enhances robustness against diverse cyber-attacks. Validation on a four-inverter testbed shows high accuracy under bias, noise, data replacement, and replay attacks, and near-attack-free performance, demonstrating a cost-effective, deployment-ready approach to cyber-physical resilience. The work offers a practical path to resilient operation of IBR-dominated microgrids with reduced sensing requirements and integrated defense against adversarial disturbances.

Abstract

Cyber-attacks jeopardize the safe operation of inverter-based resource-dominated microgrids (IBR-dominated microgrids). At the same time, existing diagnostic methods either depend on expensive multi-point instrumentation or stringent modeling assumptions that are untenable under single-point measurement constraints. This paper proposes a Fractional-Order Memory-Enhanced Attack-Diagnosis Scheme (FO-MADS) that achieves timely fault localization and cyber-resilient fault diagnosis using only one VPQ (voltage, active power, reactive power) measurement point. FO-MADS first constructs a dual fractional-order feature library by jointly applying Caputo and Grünwald-Letnikov derivatives, thereby amplifying micro-perturbations and slow drifts in the VPQ signal. A two-stage hierarchical classifier then pinpoints the affected inverter and isolates the faulty IGBT switch, effectively alleviating class imbalance. Robustness is further strengthened through Progressive Memory-Replay Adversarial Training (PMR-AT), whose attack-aware loss is dynamically re-weighted via Online Hard Example Mining (OHEM) to prioritize the most challenging samples. Experiments on a four-inverter IBR-dominated microgrid testbed comprising 1 normal and 24 fault classes under four attack scenarios demonstrate diagnostic accuracies of 96.6% (bias), 94.0% (noise), 92.8% (data replacement), and 95.7% (replay), while sustaining 96.7% under attack-free conditions. These results establish FO-MADS as a cost-effective and readily deployable solution that markedly enhances the cyber-physical resilience of IBR-dominated microgrids.

Paper Structure

This paper contains 10 sections, 4 equations, 10 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Proposed FO-MADS Framework
  3. Dual-Definition Fractional-Order Feature Engineering
  4. Hierarchical Diagnostic Architecture
  5. Robustness Training
  6. Progressive Memory-Replay Adversarial Training
  7. Simulation and Experimental Results
  8. Experimental Setup
  9. Performance Evaluation
  10. Conclusions

Figures (10)

  • Figure 1: Main framework of the FO-MADS
  • Figure 2: Process of Caputo derivative on VPQ signals showing enhanced detection of high-frequency transients
  • Figure 3: Process of Grünwald-Letnikov derivative on VPQ signals showing enhanced detection of slow-drift anomalies
  • Figure 4: Hyper-parameter study of FO-MADS. (a) Validation accuracy versus Caputo fractional order $\alpha$ ($\beta = 0.3$, $L = 400$). (b) Joint sensitivity heat-map of $\alpha$ and window length $L$; darker color denotes higher accuracy. The red rectangle marks the empirical sweet-spot
  • Figure 5: Illustration of the PMR-AT training procedure showing progressive attack escalation and memory replay mechanism
  • ...and 5 more figures