CoMPAcT-Data Based Condition Monitoring and Prediction Analytics for Turnouts

Michael Fellinger*, Petra A. Wilfling, Stefan Marschnig

*Korrespondierende/r Autor/-in für diese Arbeit

Publikation: Beitrag in Buch/Bericht/KonferenzbandBeitrag in Buch/BerichtBegutachtung


Unloaded measurements of turnouts guarantee the adherence of safety critical geometric limits but do not allow behaviour forecasting. To push maintenance strategies for turnouts up to a preventive level, performance prediction is required. Therefore, it was necessary to identify data, being suitable to describe the condition of a turnout, based on loaded measurements. The measurement data from the standard track recording car EM250 of the Austrian Federal Railways are best suited for this purpose. Due to the limited length and the varying stiffness of a turnout, the measurement data must be positioned much more precisely. In this context, it was necessary to develop a methodology for a defined and comprehensible measurement data positioning. This method enables the positioning of all available measurement signals with a maximum deviation of one measurement point (25 cm) and ensures the synchronicity between each single measurement signal. Applying this positioning algorithm to the measurement data, it is possible to describe the behaviour of turnouts quite well. Different methods for describing the necessary tamping actions as well as the actual condition of the ballast are presented. These methodologies enable describing the present condition of turnout components and deliver statistical values, required for prediction algorithms.

TitelSpringer Series in Reliability Engineering
Herausgeber (Verlag)Springer Science and Business Media Deutschland GmbH
PublikationsstatusVeröffentlicht - 2021


NameSpringer Series in Reliability Engineering
ISSN (Print)1614-7839
ISSN (elektronisch)2196-999X

ASJC Scopus subject areas

  • Sicherheit, Risiko, Zuverlässigkeit und Qualität


Untersuchen Sie die Forschungsthemen von „CoMPAcT-Data Based Condition Monitoring and Prediction Analytics for Turnouts“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren