Model-based condition monitoring of crossing panels with on-board sensors

Abstract

Excessive wear and failure of railway crossings pose significant safety and economic risks, necessitating effective condition monitoring strategies. Current monitoring approaches rely primarily on manual inspections that provide only intermittent snapshots of the crossing condition. This paper presents a condition estimation method that integrates multibody dynamics (MBD) simulations with on-board sensor data to evaluate the crossing panel condition through vehicle-track interaction analysis. The method employs an optimisation algorithm to identify unknown condition parameters within the simulation model, enabling the assessment of the crossing geometry based on historical measurements and current axlebox acceleration data. The approach provides automated condition assessment and enables predictive capabilities, facilitating proactive maintenance scheduling before failure occurrence.