A novel monitoring framework for hunting motion of railway vehicles and its experimental validation on a roller rig

Abstract

While structural health monitoring has largely focused on detecting material degradation and stiffness loss, dynamic instability remains an overlooked failure mode. This study addresses railway vehicle hunting motion, a self-excited oscillation that evolves into sustained instability under adverse conditions. A novel health evaluation indicator, termed Hunting Coefficient (H _C ), is proposed, derived from measurable wheelset displacement amplitude and hunting oscillation frequency. Unlike conventional threshold-based methods on wheelset displacement or bogie acceleration, H _C enables hierarchical classification of health states (healthy, degraded, failure) and enables early warning of incipient instability. Validation was conducted through full-scale roller rig experiments, which offers a controlled and repeatable platform for reproducing hunting behavior under varying speeds and excitation conditions. The results demonstrate that H _C reliably tracks the progression of hunting instability and outperforms conventional displacement- or acceleration-based metrics in sensitivity, robustness, and real-time applicability. These findings offer a practical basis for in-service monitoring of railway vehicles, supporting condition-based maintenance strategies and enhancing the operational safety of railway systems.