TrackScan – detecting of faults in track superstructure for condition-oriented maintenance
Where are there defects in a track? Should a defect be repaired immediately to avoid failures and thus delays? Or could a maintenance team carry out the repairs later when other work is due? Or even better: Where could preventative work be performed before the damage occurs? The DLR TrackScan project is looking at track superstructure condition monitoring to derive processes for condition-oriented maintenance.
Monitoring track infrastructure accounts for half of the lifecycle costs in the rail system. This comprises inspection, maintenance, fault elimination and repairs. Maintenance can be corrective or preventative. In many cases so far, corrective track maintenance has been common: action is only taken when a defined limit is exceeded. Condition-oriented preventive maintenance can, however, significantly reduce costs and avoid serious damage and failures. This requires good knowledge of the current condition of the track. It is also important to have a reliable forecast of how the condition of the track will change over time. The typical way of achieving this has normally been track inspections in intervals of two to 24 months. This does not provide sufficient data to form a basis for a reliable forecast.
The aim of TrackScan is to gather more comprehensive information on the condition of a track to form a basis for condition forecasts and preventive maintenance. One core element is the development and testing of measurement system prototypes which collect data on regular and frequent rail vehicles. This enables a semi-continuous monitoring of the track condition. The DLR is working in cooperation with Schweizerische Bundesbahnen (the Swiss railway service), Osthannoversche Eisenbahnen (East Hanover rail services) and Braunschweiger Hafenbetriebsgesellschaft (a harbour operations company in the city of Braunschweig) to enable testing on regular trains. The second part of the project is the development of algorithms and processes for geo-referencing the data that is collected and deriving information on the track condition from this data. The road-rail vehicle RailDriVE® serves as a rolling laboratory for capturing sensor data. Innovative data analysis algorithms are developed on the basis of this data. These are then tested in real track conditions. The testing comprises the automatic detection and time/space recording of potential defects and then a diagnosis (defect type / defect cause) by connecting the measurement data with parameters which influence the condition (e.g. track routing, track usage, weather).
Duration:
01/2014 to 12/2017
This project is managed by the department:
