Predictive Alarm Detection
SBB ZKE measures the bearing and wheel temperatures of passing trains on the open track. SCS has trained an AI model that can predict applied brakes based on measurements from several consecutive measurement systems. This makes it possible to predict an alarm before it occurs.
-
Initial situation
SBB operates measuring systems that detect brakes that have not been released (fixed brakes). In the Gotthard Base Tunnel, it is advantageous if an imminent fixed brake can be predicted before travelling through the tunnel, as a malfunction in the tunnel would be very inconvenient from an operational point of view.
-
SCS solution
SCS has trained a machine learning model that can predict fixed brakes. Using measurements from several consecutive measuring systems, the beginnings of a fault can be recognised at an early stage.
-
Added value
The study has shown that the probability of successful early detection of fixed brakes is high. Affected trains could thus be withdrawn from traffic before entering critical sections of track. This prevents the section of track from becoming blocked.
Project insights
As part of the ZKE (train control equipment), the bearing and wheel temperatures of passing trains are measured on the open track with the help of HFO systems (hot axle box and fixed brake detection). If excessive temperatures are measured, an alarm is triggered and, if necessary, an intervention is initiated (speed reduction, stop at an intervention station, inspection of the vehicle by the locomotive driver, rectification of the problem). Applied brakes (fixed brakes) or wheel bearings that are too hot (hot wheels) cause damage to the rolling stock and can lead to axle or wheel breakage and thus derailment.

In certain situations, it is advantageous if it is known that a critical situation is imminent before limit values are exceeded. This is particularly important on sections of the railway network where it is not possible to stop the train for a check for a long period of time due to the network topology (e.g. in the Gotthard Base Tunnel). If a limit value is detected to have been exceeded on such a section, the train may only travel at a reduced speed, thereby blocking the route for subsequent trains.
When predicting an alarm, one challenge that should not be underestimated is that every axis is analysed by the algorithm. The algorithm must therefore be very certain of an alarm in order to minimise possible false alarms. If an alarm is predicted for an axis without a problem, the resulting unnecessary intervention is associated with high costs. This means that the proportion of axes that lead to an alarm, even though there is no alarm, must be around one millionth. Nevertheless, the real alarms should be detected.

A study carried out by SCS has delivered promising results for the prediction of fixed brakes. With the help of an AI model(machine learning), the measurements of previously travelled HFO systems are used to predict whether an alarm will occur on the next HFO system due to a fixed brakeman. Such an approach is made possible by the fact that SBB stores the measurements of the HFO systems and their assessment. Over the years, this has created a database that is suitable for training machine learning approaches.
The close collaboration between SBB and SCS made it possible to incorporate the knowledge of SBB’s technical experts into the design of the early detection system. This is because a good understanding of the system helps to develop meaningful parameters that are ultimately fundamental to the success of a machine learning approach.
Do you have any questions or would you also like to look into the future through your data? Get in touch with us!
