Data-driven condition now- and forecasting of railway switches for



Published Jun 30, 2018
Daniela Narezo Guzman Edin Hadzic Robert Schuil Eric Baars Jörn Christoffer Groos


This contribution presents a data-based model that exploits
the power consumed by point engines during blades
movement of railway switches to detect relevant anomalies
in switch behavior. The model incorporates local air
temperature at the time of the measurement to account for
the significant influence of the environmental conditions on
normal switch behavior. Anomaly detection by the model is
validated against alerts triggered by the state-of-the-art
monitoring system POSS®, which is based on switchspecific and manually selected reference curves. The databased model leads to less in number and more reliable alerts
in comparison to the current version of POSS®. Especially
false alerts caused by temperature effects are significantly
reduced. Furthermore, the high sensitivity of the model
proves to be capable of detecting emerging switch failures at
an early stage of development. The detection capabilities of
switch condition (nowcast) and identification of emerging
failures at an early stage (required for failure forecast)
proves that the model is useful for traffic interference
prevention, condition-based predictive maintenance and
switch health enhancement.

How to Cite

Narezo Guzman, D., Hadzic, E., Schuil, R., Baars, E., & Groos, J. C. (2018). Data-driven condition now- and forecasting of railway switches for. PHM Society European Conference, 4(1).
Abstract 454 | PDF Downloads 433



Condition-based maintenance technologies, Data-driven and model-based prognostics, Asset health management

Technical Papers