Thermal Vision Analysis of Wheel-Rail Interaction: Application of Convolutional Neural Networks
[ 1 ] Instytut Transportu, Wydział Inżynierii Lądowej i Transportu, Politechnika Poznańska | [ 2 ] Instytut Technologii Mechanicznej, Wydział Inżynierii Mechanicznej, Politechnika Poznańska | [ 3 ] Instytut Mechaniki Stosowanej, Wydział Inżynierii Mechanicznej, Politechnika Poznańska | [ SzD ] doctoral school student | [ P ] employee
[2.7] Civil engineering and transport[2.7] Civil engineering, geodesy and transport[2.9] Mechanical engineering
2024
chapter in monograph / paper
english
- wheel-rail interface
- thermal imaging
- convolutional neural networks
- image classification
- frictional heating
- creepages
EN This paper focused on the application of a Convolutional Neural Network in the analysis of the thermal images of the wheel-rail interface. The first part is centered on the primary motivations behind conducting thermal imaging measurements in the wheel-rail contact area and the analysis of the thermograms using Convolutional Neural Networks. It was emphasized that automatic classification of the wheel-rail contact types can provide valuable insights, especially considering the spatial data and statistical metrics, which can be useful in the case of wear intensity analysis. Subsequently, the methodology of thermal imaging measurement and the design of a classifier utilizing Convolutional Neural Network technology were presented. In the context of result analysis, the very promising capabilities of detecting various contact types using Convolutional Neural Network s were highlighted. The paper concluded by summarizing the key benefits arising from the proposed technology and its potential impact on wheel-rail interaction studies.
9.7-1 - 9.7-5
Paper 9.7
copyright
publisher's website
final published version
5