A Prognostic Framework for Railway Track Geometry: Tamping Detection, Settling-Aware Estimation, and Spatially Resolved RUL Prediction
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
https://orcid.org/0000-0002-1814-4278
Wolfgang Lachnit Mohammed Amin Adoul Wolfgang Birk
Abstract
Accurate estimation of railway track geometry degradation and reliable prediction of remaining useful life (RUL) are essential for cost-effective infrastructure management. This paper presents a self-contained Kalman filtering framework that integrates unsupervised tamping detection and Settling-aware post-tamping state management, jointly addressing state estimation, short-term forecasting, and RUL prediction for longitudinal level and twist parameters using only historical onboard monitoring (OBM) and measurement train (MT) data, without requiring traffic load, environmental, or maintenance metadata. The framework comprises four integrated stages: (i) source-specific outlier detection exploiting the distinct noise characteristics of OBM and MT instruments; (ii) unsupervised tamping detection through adaptive jump analysis on monthly representative values; (iii) a damped Kalman filter with adaptive noise modelling and spatial fusion of neighboring positions, augmented by a Settling-aware state management mechanism that detects the rapid post-tamping consolidation phase and injects physically informed velocity priors to prevent filter lag; and (iv) a RUL prediction module that propagates the final Kalman state forward under damped dynamics until the predicted geometry violates the Alert Limit (AL) or Immediate Action Limit (IAL) defined by EN 13848-5. The complete pipeline is evaluated on 50 consecutive track positions spanning a 50-metre segment of Sudostbahn line 870 (Switzerland), using 6 MT and 33 OBM observations per position collected over five years (2016-2021). Results demonstrate accurate estimation through both degradation and maintenance phases, with six-month forecast confidence bands. The multi-position RUL analysis classifies 98% of positions as safe beyond a 2-year horizon and identifies the remaining positions with finite RUL values, enabling spatially targeted maintenance prioritization.
How to Cite
##plugins.themes.bootstrap3.article.details##
Railway track geometry, Adaptive Kalman filter, Unsupervised tamping detection, Estimation and Forecasting, Post-Tamping Settlement, Remaining Useful Life
Bar-Shalom, Y., Li, X. R., & Kirubarajan, T. (2001). Estimation with applications to tracking and navigation. John Wiley & Sons.
Birk, W., Westerberg, J., Larsson-Kråik, P. O., & Lachnit, W. (2021). Track geometry estimation and prediction tool combining onboard monitoring and measurement vehicle data. In 2021 AREMA Annual Conference & Expo. Virtual conference.
Chen, Y., Zhang, Y., & Yang, F. (2021). Learn to predict vertical track irregularity with extremely imbalanced data. In Asian Conference on Machine Learning (pp. 1493–1504). PMLR.
Dahlberg, T. (2004). Railway track settlements: A literature review (Tech. Rep. No. 463). Report for the EU project SUPERTRACK.
Ghiasi, R., Khan, M. A., Sorrentino, D., Diaine, C., & Malekjafarian, A. (2024). An unsupervised anomaly detection framework for onboard monitoring of railway track geometrical defects using one-class support vector machine. Engineering Applications of Artificial Intelligence, 133, 108167.
Grubbs, F. E. (1969). Procedures for detecting outlying observations in samples. Technometrics, 11(1), 1–21.
Liu, J., Du, D., He, J., & Zhang, C. (2024). Prediction of remaining useful life of railway tracks based on DMG-DCCGRU hybrid model and transfer learning. IEEE Transactions on Vehicular Technology, 73(6), 7561–7575.
Sato, Y. (2018). Japanese studies on deterioration of ballasted track. In Interaction of Railway Vehicles with the Track and Its Substructure (pp. 197–208). Routledge.
Selig, E. T., & Waters, J. M. (1994). Track geotechnology and substructure management. Thomas Telford.
Simon, D. (2006). Optimal state estimation: Kalman, H infinity, and nonlinear approaches. John Wiley & Sons.
Traquinho, N., Vale, C., Ribeiro, D., Meixedo, A., Montenegro, P., Mosleh, A., & Calçada, R. (2023). Damage identification for railway tracks using onboard monitoring systems in in-service vehicles and data science. Machines, 11(10), 981.
Truong-Ba, H., Rebello, S., Cholette, M. E., Reddy, V., & Borghesani, P. (2025). Bayesian multivariate track geometry degradation modeling and its use in condition-based inspection. Railway Engineering Science, 1–25.
Yan, T. H., Hoelzl, C., Corman, F., Dertimanis, V., & Chatzi, E. (2025). Integration of on-board monitoring data into infrastructure management for effective decision-making in railway maintenance. Railway Engineering Science, 33(2), 151–168.
Züger, S., Schlatter, C., Wolter, K. U., Nerlich, I., & Hunn, S. (2020). Onboard monitoring in der Schweiz, ein Gemeinschaftswerk dreier Bahnen. ZEVrail, 144(4).
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.