On-board wheel flat detection for heavy haul wagons using ultra-low power sensor nodes

Condition monitoring of general freight and heavy haul railway vehicles is typically performed by wayside equipment scattered along the network. Hence, these systems are not able to provide instantaneous alarms and real-time diagnostics of the vehicles and their components. Although wayside systems deliver important data for monitoring of condition trends, allow planning of vehicle maintenance, and provide alarms that mitigate damage to infrastructure caused by rolling stock failures, implementing on-board instrumentation for monitoring vehicle components would further improve safety by allowing instant fault detection. On-board instrumentation could also further increase efficiency by enabling more predictive maintenance techniques and enhance operation and revenue by delivering insights that permit faster, longer and heavier trains. To study the feasibility of on-board fault detection applications, a multibody dynamic model of a railway freight wagon, operating with a variety of wheel flat defects was developed. The model enabled the study of the vehicle dynamic behaviour and the development of a wheel flat detection technique based on axle adaptor acceleration signals. By mapping the signal processing algorithm to an analogue electronic circuit, it was possible to optimize the sensor node hardware architecture. The proposed technique successfully detected the wheel defect in a variety of scenarios. The optimised sensor node was able to deliver the detection data with considerably reduced power and hardware requirements compared to traditional all-digital signal processing and data acquisition. The minimal computational and data storing requirements, open the possibilities to use simpler microcontrollers in advanced signal processing applications. The paper will present details of the design, the processing methodology and introduce a version of the optimised sensor node hardware architecture. The innovation presents the opportunity of developing ultra-low power, low-cost onboard condition monitoring applications for heavy haul wagons, suitable for deployment on large rolling stock fleets.