Development of a Machine Learning Method to Calculate Real-time 3D Deflections in a Dynamic Magnetic Field

  • Links:Ausschreibung
  • Trelleborg is developing a contactless 3D deflection measuring system. The system allows capturing static and dynamic deflections as a basis for complex condition monitoring and predictive maintenance approaches – especially in railway applications.

    The task contains:

    • Development of a Machine Learning model which is able to infer 3D deflections by combining different sensor signals 
    • Gaining knowledge of the physical system behavior as well as correlations in the sensor values
    • Validating approach on laboratory tests and benchmarks 

    Current status:

    • First prototype of sensor hardware available and tested
    • Extensive datasets for analysis available
    • Improved sensor hardware available in 08/2019
    • First approach of algorithm as base for improvements / benchmarking
    • Wide range of tools and hardware for additional tests available


    • Real-time capable prediction model / algorithm
    • Measuring accuracy of 0.1 mm
    • Development of proper calibration and commissioning processes 

    If you are interested, send a short letter of motivation, your CV and a transcript of records to dominik.martin∂kit.edu.