A Robust Method for Fault Detection and Severity Estimation in Mechanical Vibration Data
Youngjae Jeon, Eunho Heo, Jinmo Lee, Taewon Uhm, Dongjin Lee
TL;DR
The paper tackles fault detection and continuous severity estimation in multivariate vibration data by integrating Temporal Graph Convolutional Networks (T-GCN) with a robust severity index based on the mean and standard deviation of anomaly scores above a threshold. The approach leverages spatial-temporal modeling to predict normal behavior and derives a stable severity measure S_{t,m} = mu_t + m sigma_t, reducing abrupt fluctuations in anomaly scores across conditions. Empirical validation on IMS bearing and fanjet datasets demonstrates improved robustness and reliability for predictive maintenance in safety-critical mechanical systems. This work advances fault prognosis by eliminating reliance on discrete severity levels and providing a practical framework for proactive maintenance planning and risk management.
Abstract
This paper proposes a robust method for fault detection and severity estimation in multivariate time-series data to enhance predictive maintenance of mechanical systems. We use the Temporal Graph Convolutional Network (T-GCN) model to capture both spatial and temporal dependencies among variables. This enables accurate future state predictions under varying operational conditions. To address the challenge of fluctuating anomaly scores that reduce fault severity estimation accuracy, we introduce a novel fault severity index based on the mean and standard deviation of anomaly scores. This generates a continuous and reliable severity measurement. We validate the proposed method using two experimental datasets: an open IMS bearing dataset and data collected from a fanjet electric propulsion system. Results demonstrate that our method significantly reduces abrupt fluctuations and inconsistencies in anomaly scores. This provides a more dependable foundation for maintenance planning and risk management in safety-critical applications.