Towards automated detection and diagnosis of suspension system defects in passenger railway vehicles

##plugins.themes.bootstrap3.article.main##

##plugins.themes.bootstrap3.article.sidebar##

Published Jun 28, 2018
Nikolaas N. Oosterhof
Margot F. E. Peters

Abstract

Suspension defects in passenger trains reduce safety and comfort in rail transport systems. We investigated the feasibility of detecting such defects using dynamic wheel load differences (DWLDs), which are measured from trains in operation using sensors in the track.

We found that DWLD data shows considerable variability but in a consistent manner, and developed a method that purifies the data to improve signal-to-noise. Further, we developed an algorithm that uses the purified DWLD data to detect anomalous events in suspension imbalance, which is indicative of a defect or repair event. Our algorithm further provides a diagnosis of an anomalous event as related to either the primary or the secondary suspension system. We validated our algorithm using a limited set of maintenance records and found a high detection and correct classification rate, although more extended validation is needed with more maintenance records and DWLD data.

In sum, our work indicates a promising avenue of high-frequent, automated detection and diagnosis of suspension defects, which would contribute to efficient, economical and save operation of railway vehicles.

 

How to Cite

Oosterhof, N. N., & Peters, M. F. E. (2018). Towards automated detection and diagnosis of suspension system defects in passenger railway vehicles. PHM Society European Conference, 4(1). https://doi.org/10.36001/phme.2018.v4i1.423
Abstract 384 | PDF Downloads 334

##plugins.themes.bootstrap3.article.details##

Keywords

wheel load differences, railways, safety, suspension, anomaly detection, trains

Section
Technical Papers

Similar Articles

1 2 3 4 > >> 

You may also start an advanced similarity search for this article.