FaultGuard: A Generative Approach to Resilient Fault Prediction in Smart Electrical Grids

TL;DR

FaultGuard tackles the vulnerability of ML-based fault prediction in smart grids to adversarial attacks by coupling a GAN-based Anomaly Detection System with a low-complexity, online adversarially trained fault predictor. The ADS detects and rejects malicious inputs, while online adversarial training strengthens the predictor against perturbations, achieving up to $0.958$ fault-type and $0.958$ fault-zone accuracy under attack and ADS detection up to $1.000$. Evaluations on the IEEE-13 AdvAttack dataset show significant robustness gains and reveal the persistent challenge posed by sophisticated attacks like CW, motivating multi-layer defenses. The work demonstrates substantial practical benefits for secure fault management in smart grids, providing a framework that is both effective and extensible to evolving threat models. The authors also discuss computational trade-offs and outline avenues for richer datasets and broader threat coverage to further enhance resilience.

Abstract

Predicting and classifying faults in electricity networks is crucial for uninterrupted provision and keeping maintenance costs at a minimum. Thanks to the advancements in the field provided by the smart grid, several data-driven approaches have been proposed in the literature to tackle fault prediction tasks. Implementing these systems brought several improvements, such as optimal energy consumption and quick restoration. Thus, they have become an essential component of the smart grid. However, the robustness and security of these systems against adversarial attacks have not yet been extensively investigated. These attacks can impair the whole grid and cause additional damage to the infrastructure, deceiving fault detection systems and disrupting restoration. In this paper, we present FaultGuard, the first framework for fault type and zone classification resilient to adversarial attacks. To ensure the security of our system, we employ an Anomaly Detection System (ADS) leveraging a novel Generative Adversarial Network training layer to identify attacks. Furthermore, we propose a low-complexity fault prediction model and an online adversarial training technique to enhance robustness. We comprehensively evaluate the framework's performance against various adversarial attacks using the IEEE13-AdvAttack dataset, which constitutes the state-of-the-art for resilient fault prediction benchmarking. Our model outclasses the state-of-the-art even without considering adversaries, with an accuracy of up to 0.958. Furthermore, our ADS shows attack detection capabilities with an accuracy of up to 1.000. Finally, we demonstrate how our novel training layers drastically increase performances across the whole framework, with a mean increase of 154% in ADS accuracy and 118% in model accuracy.

Figures (5)

• Figure 1: FaultGuard framework.
• Figure 2: Schema of our gan-based ads training process for each epoch.
• Figure 3: Combinatorial accuracy of fault prediction tasks when delaying notification by the number of consecutive faulty batches.
• Figure 4: Model's accuracy at varying epsilon values on the white-box attacks.
• Figure 5: Model's accuracy at varying epsilon values on the white-box attacks when equipped with the defense mechanism.