Descriptionhe main task of asset management is deciding whether an asset, track or turnout, requires maintenance or already needs to be re-invested. This can only be answered by forecasting the behaviour of the components and the entire turnout, which is in the same time the way for implementing predictive maintenance.
The Institute of Railway Engineering and Transport Economy at Graz University of Technology developed a methodology for project evaluation based on forecasting the maintenance demand for track. This methodology is already used by the Austrian Federal Railways since 2011 for deciding major track re‑investment projects. It is mainly based on recording car data. The different information can be used for characterising the behaviour of the entire track. Furthermore, specific signals and analysis provide information for specific track components as well.
However, this methodology up to now cannot be used for turnouts, as they are more complex with more maintenance options possible. The main limitations for transferring the prediction methodology of tracks towards turnouts are different data sources, the ever-changing stiffness of turnouts, and thus higher demand of data positioning. Furthermore, in difference to track the economic service life of turnouts differs in a wider range from the technical one.
As a first step, the reasons for re-investment of turnouts have been identified. The main source of data for this decision is the manual inspection, which is a non-loaded measurement. The existing data of turnouts are analysed in their time sequences to see whether they can be used for forecasting in order to build up a similar prediction methodology for turnouts. Unfortunately, trend analyses cannot be based on the manual inspection data, they require loaded measurements. Sensors in turnouts are already able to deliver 50% of all data derived by manual inspection. It is already possible to cover more than 95% of the required information by combining sensors in the turnout and measurements by trains and/or recording cars.
Combing these two sources will enable the understanding of the turnout behaviour and consequently a predictive maintenance will be possible for turnouts as it is already possible for track.
|Period||11 Oct 2017|
|Event title||InnoRail 2017: International Conference on the Single European Railway Area|