Name: Nico Burgelman (MSc)
Supervisor: Dr Zili Li (Civil Engineering and Geosciences, Road and Railway Engineering)
Subject: Train-turnout Interaction, Simulation and Validation by Axle Box Acceleration Measurements
Thesis Defence: In two years
“Everyone who has travelled by train, recognizes the sudden swinging or bouncing of a train. This could very well be caused by sagged ballast stones underneath the railway. I’m developing an automatic algorithm that shows where and how the railway is dislocated and when it needs maintenance.
In order to learn how to detect the dislocations, our department does field research. We put sensors on a train’s axle box, close to the wheels. The sensors measure accelerations to the right, left, upwards and downwards. In this way they tell us where the dislocations are. Sometimes I ride on a train equipped with our sensors. It’s fascinating to sit next to the machinist and see the enormous amount of data coming in on a laptop.
Afterwards I create an algorithm to convert the data to useful information. It should tell how many millimetres the dislocations are, and in future it will provide ProRail, the Dutch company that takes care of maintenance for the Dutch railway network, with important information regarding (future) maintenance work. It could very well be that it is better to monitor a dislocation of a couple of millimetres than to repair it straight away.
In this way the precise model I’m working on should reduce the maintenance costs of the railways, because it’s more precise and cheaper than the current method. At this moment an expensive special train measures dislocations, covering every part of railway in six months. The sensors for the method I’m developing could be placed on any train. If one train on a trajectory has our system, this will provide enough detailed and accurate information.
The method should also be used on railway switches. The maintenance of one switch is about as expensive as 500 meters of straight railway. However, it’s difficult to detect dislocations on railway switches, because of dynamic effects caused by the complicated shape of the switch. This complex geometry also requires an advanced model of wheel/rail interaction for the simulations. Nevertheless, we want to know if the movements of the train are caused by the normal shape of the switch or ballast displacements.
Over the next two years I’ll create a method that should provide this information. I’ll make use of existing wheel/rail contact algorithms and implement them in our vehicle dynamic program. Hopefully, this will result in a method that reduces maintenance costs, limits the dislocations felt by passengers and end the swinging and bouncing of trains.”