PREDICTIVE MODELING OF TRACK QUALITY INDEX WITH NEURAL NETWORKS

Ivona Nedevska Trajkova; Zlatko Zafirovski

doi:10.5937/jaes0-60022

Ivona Nedevska Trajkova Ss. Cyril and Methodius University in Skopje, Faculty of Civil Engineering, North Macedonia, Chair of Railways https://orcid.org/0000-0002-0407-4846
Zlatko Zafirovski Ss. Cyril and Methodius University in Skopje, Faculty of Civil Engineering, North Macedonia, Chair of Railways https://orcid.org/0000-0003-2099-0710

DOI: https://doi.org/10.5937/jaes0-60022

Keywords: track quality index, GRNN, railway maintenance, predictive modelling, machine learning, TQI prediction, neural networks

Abstract

The Track Quality Index (TQI) was used as a primary metric to assess the structural and geometric condition of railway track infrastructure, which is directly related to traffic safety, ride comfort, and maintenance prioritization. This paper presents a predictive modeling approach aimed at forecasting future TQI values by incorporating historical measurements, geometric characteristics, and operational parameters. The analyzed dataset consisted of 821 observations of track sections collected over several measurement cycles. It included explanatory variables such as previous TQI levels, train operating speed, traffic volume, and infrastructure elements like tunnels, bridges, and records of ballast tamping. A General Regression Neural Network (GRNN) was trained using 657 data samples and validated on 164 samples, achieving a coefficient of determination (R²) of 0.699, a root mean square error (RMSE) of 4.70, and a mean absolute error (MAE) of 3.55. The results confirmed that neural networks are suitable for short-term predictions of track conditions, facilitating more effective planning and optimization of maintenance interventions. The proposed framework supports engineering decision-making related to condition-based maintenance of railways.

References

Han, Z., Han, L., Chen, C., Han, Q., Jing, G., Cheng, Z. (2024). Dynamic calibration method for track geometry measurement system: A case study in China. Measurement Science and Technology, vol. 35, no. 6, p. 066007, DOI: 10.1088/1361-6501/ad2cd8

Kite, D., Siino, G., Audley, M. (2020). Detecting embankment instability using measurable track geometry data. Infrastructures, vol. 5, no. 3, p. 29, DOI: 10.3390/infrastructures5030029

Yan, T., Corman, F. (2020). Assessing and extending track quality index for novel measurement techniques in railway systems. Transportation Research Record: Journal of the Transportation Research Board, vol. 2674, no. 8, pp. 24–36, DOI: 10.1177/0361198120923661

Amin, R., Kumar, A., Hassan, S. (2023). Predictive maintenance in rail transport: Evaluating the efficacy of the track quality index. Transport Reviews, vol. 43, no. 2, pp. 256–274.

Tsai, C. C., Lin, W. Y., Chao, J. C. (2020). Maintenance strategies for railway infrastructure: Integrating track quality indices in cost optimization. Rail Engineering and Maintenance, vol. 45, no. 2, pp. 203–219.

Li, X., Zhang, Y., Chen, H. (2016). Modeling track degradation in railway systems: Limitations of linear regression and the promise of nonlinear approaches. Journal of Rail Transportation Engineering, vol. 12, no. 3, pp. 145–159.

Branford, Q., Ellis, P., Martin, D. (2022). Evaluation of neural network architectures for predicting track condition deterioration. Journal of Infrastructure Systems, vol. 28, no. 1, p. 04022009.

Zhao, Q., Li, B., Hu, F. (2018). Integrating physical models and neural networks for enhanced track degradation forecasting. Structural Safety, vol. 66, pp. 73–81.

Zhang, Y., et al. (2020). A survey on neural networks in time series forecasting: Comprehensive framework, state of the art and future directions. Neural Networks, DOI: 10.1016/j.neunet.2019.02.013

European Committee for Standardization (CEN) (2019). EN 13848-1: Railway applications – Track geometry quality – Part 1: Characterization of track geometry. Brussels, Belgium. Available: https://standards.iteh.ai/catalog/standards/cen/19848cc1-08a5-44aa-96aa-18294c6ab12f/en-13848-1-2019

European Committee for Standardization (CEN) (2020). EN 13848-2: Railway applications – Track geometry quality – Part 2: Measuring systems – Track recording vehicles. Brussels, Belgium. Available: https://knowledge.bsigroup.com/products/railway-applications-track-track-geometry-quality-measuring-systems-track-recording-vehicles-1/standard

International Union of Railways (UIC) (2009). UIC 518: Testing and approval of railway vehicles from the point of view of their dynamic behaviour – Safety – Track fatigue – Running behaviour. UIC Code 518. Available: https://www.normadoc.com/spanish/uic-518-2009-10-4542750.html