This study proposes an integrated methodology for modelling and predicting ballast degradation in turnouts by exploiting loaded track geometry data. The rate of increase in track irregularities with respect to the cumulative loading of the track is the index of ballast degradation to be predicted. The methodology combines the fractal analysis method with probabilistic modeling and Bayesian inference to design a prognostic tool that forecasts the expected ballast degradation rate over a period of three months. The foundation of the proposed method is a universal degradation model for the ballast in turnouts, which fuses technological, operational, and environmental information to prognose the rate of ballast deterioration. The methodology has been verified on 15 different turnouts of the Danish railway infrastructure.
|Journal||Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit|
|Publication status||Published - 2020|
- Ballast degradation
- Vertical track geometry
- Bayesian estimation
- Fractal analysis
- Railway turnouts