Predge Rolling Stock – Wheel Damage Prediction (WDP)
One of our most popular analytics feature is the Wheel Damage Prediction, or WDP. Right now, this feature is in operation continuously analyzing around 10 000 wheelsets providing failure predictions and estimated life lengths to the users.
Wheel damages caused by rolling contact fatigue is an critical aspect in railway operations globally. Due to the rapid initiation and exponential growth of these damages it is crucial to have reliable and up-to-date information of these deviations to avoid the safety critical events these damages may result in.
By keeping track on each individual wheel’s condition, identifying these deviations at an early stage and having an estimated life length of each wheel, much benefits can be achieved.
WDP utilizes information from one or many wayside detectors measuring wheel impact forces. The feature synchronizes this information with estimated or measured distance information and integrates this information with secondary information such as other types of condition data and weather information.
The analytics behind the feature is based on our AI and Machine learning principles with hybrid models. The hybrid models are not purely data driven and also takes the physical aspects behind these damages into account. Still, it is continuously and automatically updated as new measurements are performed.
The users of Predge Rolling Stock having WDP as an feature will not directly access the analytics results as such.
Due to our work with continuous improvements, all our features are constantly evolving. The next step with the WDP is to improve the ability to classify types of wheel damages. By close collaboration with maintenance workshops and high quality documentation of damages, we aim at delivering high confidence type of damage, such as spalling, wheel flat and out of roundness as well as severity to our customers in the near future. As soon as these improvements shows sufficient performance, the users will have access to this capability as well.
Success story : Predge Rolling Stock WDP – Creating benefit for CargoNets operation
Transporting freight from Oslo to Narvik is a challenging task. The one-way route spans around 2000 kilometers through two countries with weather conditions varying widely even on a single trip
“The Wheel Damage Prediction (WDP) feature has given us the possibility to shunt out wagons preventively and thereby avoid stops along track. It has also allowed us to plan maintenance activities regarding wheel damages on another level considering workshop capacity and spare part availability.” – Erik Halland, Rolling Stock at CargoNet