Monitoring of railway tracks maintenance needs using dynamic responses collected by an in-service train

Abstract

Railway systems are critical components of transportation networks requiring consistent maintenance. This paper proposes a novel data-driven approach to detect various maintenance needs of railway track systems using acceleration data obtained from a passenger train in operation. The framework contains four modules. Firstly, data pre-processing and cleansing are performed to extract useful data from the whole dataset. Then, condition-sensitive features are extracted from the raw data in three different domains of time, frequency, and time–frequency. In the third module, the best subset of measurement features that characterize the state of the tracks are selected using the analysis of variance (ANOVA) algorithm which eliminates irrelevant characteristics from the feature set of responses. Finally, a multilabel classification algorithm based on the cascade feed-forward neural network (CFNN) is used to classify the type of maintenance needs of the track. An open-access dataset from a field study in Pennsylvania, USA, is used in this study for validation of the proposed method. The results indicate that employing a CFNN can achieve 95% accuracy in identifying two maintenance activities, tamping and surfacing, using time-domain features. Moreover, an extensive analysis has been conducted to evaluate the influence of various feature extraction and selection methods, diverse classification algorithms, and different types of accelerometers (uni-axial and tri-axial) on the accuracy of the proposed method.

Keywords

Rail maintenance / Signal processing / Exploratory data analysis / Data-driven / Data reduction / Information and Computing Sciences / Artificial Intelligence and Image Processing

References

[1] TangR, De DonatoL, BesinovicN, et al. . A literature review of artificial intelligence applications in railway systems. Transp Res Part C Emerg Technol, 2022, 140: 103679.

[2] McBride J, Siripurapu A (2021) The state of U.S. infrastructure. https://govwhitepapers.com/whitepapers/the-state-of-u-s-infrastructure. Accessed 15 April 2024

[3] Peinado GonzaloA, HorridgeR, SteeleH, et al. . Review of data analytics for condition monitoring of railway track geometry. IEEE Trans Intell Transp Syst, 2022, 231222737-22754.

[4] AvciO, AbdeljaberO, KiranyazS, et al. . A review of vibration-based damage detection in civil structures: from traditional methods to machine learning and deep learning applications. Mech Syst Signal Process, 2021, 147: 107077.

[5] MoslehA, MeixedoA, RibeiroD, et al. . Early wheel flat detection: an automatic data-driven wavelet-based approach for railways. Veh Syst Dyn, 2023, 6161644-1673.

[6] MalekjafarianA, SarrabezollesCA, KhanMA, GolpayeganiF. A machine-learning-based approach for railway track monitoring using acceleration measured on an in-service train. Sensors, 2023, 23177568.

[7] GhiasiR, GhasemiMR, ChanTHT. Optimum feature selection for SHM of benchmark structures using efficient AI mechanism. Smart Struct Syst, 2021, 274623

[8] KeshmiryA, HassaniS, MousaviM, et al. . Effects of environmental and operational conditions on structural health monitoring and non-destructive testing: a systematic review. Buildings, 2023, 134918.

[9] EntezamiA, SarmadiH, BehkamalB. Removal of freezing effects from modal frequencies of civil structures for structural health monitoring. Eng Struct, 2024, 319: 118722.

[10] EntezamiA, SarmadiH, BehkamalB, et al. . Early warning of structural damage via manifold learning-aided data clustering and non-parametric probabilistic anomaly detection. Mech Syst Signal Process, 2025, 224: 111984.

[11] SarmadiH, KaramodinA. A novel anomaly detection method based on adaptive Mahalanobis-squared distance and one-class kNN rule for structural health monitoring under environmental effects. Mech Syst Signal Process, 2020, 140: 106495.

[12] SarmadiH, YuenK. Early damage detection by an innovative unsupervised learning method based on kernel null space and peak-over-threshold. Comput Aided Civ Infrastruct Eng, 2021, 3691150-1167.

[13] MoslehA, MeixedoA, RibeiroD, et al. . Automatic clustering-based approach for train wheels condition monitoring. Int J Rail Transp, 2023, 115639-664.

[14] MaesK, Van MeerbeeckL, ReyndersEPB, et al. . Validation of vibration-based structural health monitoring on retrofitted railway bridge KW51. Mech Syst Signal Process, 2022, 165: 108380.

[15] MalekjafarianA, CorballyR, GongW. A review of mobile sensing of bridges using moving vehicles: progress to date, challenges and future trends. Structures, 2022, 44: 1466-1489.

[16] LiuJ, ChenS, LedermanG, et al. . Dynamic responses, GPS positions and environmental conditions of two light rail vehicles in Pittsburgh. Sci Data, 2019, 6: 146.

[17] GhiasiR, KhanMA, SorrentinoD, et al. . An unsupervised anomaly detection framework for onboard monitoring of railway track geometrical defects using one-class support vector machine. Eng Appl Artif Intell, 2024, 133: 108167.

[18] LedermanG, ChenS, GarrettJH, et al. . A data fusion approach for track monitoring from multiple in-service trains. Mech Syst Signal Process, 2017, 95: 363-379.

[19] LedermanG, ChenS, GarrettJH, et al. . Track monitoring from the dynamic response of a passing train: a sparse approach. Mech Syst Signal Process, 2017, 90: 141-153.

[20] LedermanG, ChenS, GarrettJ, et al. . Track-monitoring from the dynamic response of an operational train. Mech Syst Signal Process, 2017, 87: 1-16.

[21] Liu J, Wei Y, Bergés M et al (2019) Detecting anomalies in longitudinal elevation of track geometry using train dynamic responses via a variational autoencoder. In: Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2019, March 3–7, 2019. Denver, USA. SPIE, pp. 389–401

[22] Sorrentino D, Ta QA, Latestere F et al. (2022) Improved condition monitoring of railway tracks through the analysis of on-board dynamic and geometry measurements on high-speed lines. In: World congress on railway research 2022 (WCRR2022), Birmingham, UK, 6–10 June 2022. pp. 1–6

[23] GhiasiR, LestoilleN, DiaineC, et al. . Unsupervised domain adaptation for drive-by condition monitoring of multiple railway tracks. Eng Appl Artif Intell, 2025, 139: 109516.

[24] MolodovaM, OreguiM, NúñezA, et al. . Health condition monitoring of insulated joints based on axle box acceleration measurements. Eng Struct, 2016, 123: 225-235.

[25] MalekjafarianA, OBrienE, QuirkeP, et al. . Railway track monitoring using train measurements: an experimental case study. Appl Sci, 2019, 9224859.

[26] JawalageriS, GhiasiR, JalilvandS, et al. . A data-driven approach for scour detection around monopile-supported offshore wind turbines using Naive Bayes classification. Mar Struct, 2024, 95: 103565.

[27] BuckleyT, GhoshB, PakrashiV. A feature extraction & selection benchmark for structural health monitoring. Struct Health Monit, 2023, 2232082-2127.

[28] GhiasiR, MalekjafarianA. Feature subset selection in structural health monitoring data using an advanced binary slime mould algorithm. J Struct Integr Maint, 2023, 84209-225

[29] FathnejatH, Ahmadi-NedushanB. An efficient two-stage approach for structural damage detection using meta-heuristic algorithms and group method of data handling surrogate model. Front Struct Civ Eng, 2020, 144907-929.

[30] Ghiasi R, Malekjafarian A (2023) A data-driven approach for monitoring railway tracks using dynamic responses collected by an in-service train. In: Limongelli MP, Giordano PF, Quqa S et al (eds.) Experimental vibration analysis for civil engineering structures (EVACES 2023). Springer Nature Switzerland, Cham, pp 165–174

[31] BommertA, SunX, BischlB, et al. . Benchmark for filter methods for feature selection in high-dimensional classification data. Comput Stat Data Anal, 2020, 143: 106839.

[32] DhalP, AzadC. A comprehensive survey on feature selection in the various fields of machine learning. Appl Intell, 2022, 5244543-4581.

[33] KimTK. Understanding one-way ANOVA using conceptual figures. Korean J Anesthesiol, 2017, 70122.

[34] Leon-MedinaJX, AnayaM, ParésN, et al. . Structural damage classification in a jacket-type wind-turbine foundation using principal component analysis and extreme gradient boosting. Sensors, 2021, 2182748.

[35] GhiasiR, GhasemiMR, NooriM. Comparative studies of metamodeling and AI-Based techniques in damage detection of structures. Adv Eng Softw, 2018, 125: 101-112.

[36] TooJ, AbdullahAR, Mohd SaadN. A new quadratic binary Harris hawk optimization for feature selection. Electronics, 2019, 8101130.

[37] LiaoY, HanL, WangH, et al. . Prediction models for railway track geometry degradation using machine learning methods: a review. Sensors, 2022, 22197275.

[38] HemeidaAM, HassanSA, Ali MohamedA-A, et al. . Nature-inspired algorithms for feed-forward neural network classifiers: a survey of one decade of research. Ain Shams Eng J, 2020, 113659-675.

[39] PiresAC, VianaMCA, ScaramussaLM, et al. . Measuring vertical track irregularities from instrumented heavy haul railway vehicle data using machine learning. Eng Appl Artif Intell, 2024, 127: 107191.

[40] KandiriA, GhiasiR, NogalM, et al. . Travel time prediction for an intelligent transportation system based on a data-driven feature selection method considering temporal correlation. Transp Eng, 2024, 18: 100272.

[41] Marcílio-JrWE, ElerDM. Explaining dimensionality reduction results using Shapley values. Expert Syst Appl, 2021, 178: 115020.

[42] Winter E (2002) The shapley value. In: Handbook of game theory with economic applications, vol. 3. North-Holland, pp. 2025–2054

[43] Zhao Z, Anand R, Wang M (2019) Maximum relevance and minimum redundancy feature selection methods for a marketing machine learning platform. In: 2019 IEEE international conference on data science and advanced analytics (DSAA), Washington DC, 5–8 October, 2019. IEEE, pp. 442–452

[44] ShreemSS, AbdullahS, NazriMZA, et al. . Hybridizing ReliefF, MRMR filters and GA wrapper approaches for gene selection. J Theor Appl Inf Technol, 2012, 4621034-1039

[45] GharehbaghiVR, NguyenA, Noroozinejad FarsangiE, et al. . Supervised damage and deterioration detection in building structures using an enhanced autoregressive time-series approach. J Build Eng, 2020, 30: 101292.

[46] BianchiG, FreddiF, GiulianiF, et al. . Implementation of an AI-based predictive structural health monitoring strategy for bonded insulated rail joints using digital twins under varied bolt conditions. Railw Eng Sci, 2025.

[47] LuoY, GuoX, WangLK, et al. . Unsupervised structural damage detection based on an improved generative adversarial network and cloud model. J Low Freq Noise Vib Act Contr, 2023, 4231501-1518.

[48] AlexanderW, WilliamsCMDigital signal processing: principles, algorithms and system design, 2017ElsevierAcademic Press

[49] LiuYY, JuYF, DuanCD, et al. . Structure damage diagnosis using neural network and feature fusion. Eng Appl Artif Intell, 2011, 24187-92.

[50] Lang K, Xing Z, Dong W et al (2018) A rail corrugation detection method based on wavelet packet energy entropy. In: Proceedings of the 3rd international conference on electrical and information technologies for rail transportation (EITRT) 2017. Springer Singapore, pp 205–214

[51] MohammadiM, MoslehA, ValeC, et al. . An unsupervised learning approach for wayside train wheel flat detection. Sensors, 2023, 2341910.

[52] AbdeljaberO, AvciO, KiranyazMS, et al. . 1-D CNNs for structural damage detection: verification on a structural health monitoring benchmark data. Neurocomputing, 2018, 275: 1308-1317.

[53] MousaviM, GandomiAH. Structural health monitoring under environmental and operational variations using MCD prediction error. J Sound Vib, 2021, 512: 116370.

[54] DoroudiR, LavassaniSHH, ShahrouziM. Optimal tuning of three deep learning methods with signal processing and anomaly detection for multi-class damage detection of a large-scale bridge. Struct Health Monit, 2024, 2353227-3252.

[55] MousaviZ, VarahramS, EttefaghMM, et al. . A digital twin-based framework for damage detection of a floating wind turbine structure under various loading conditions based on deep learning approach. Ocean Eng, 2024, 292: 116563.

[56] MalekjafarianA, KhanMA, OBrienEJ, et al. . Indirect monitoring of frequencies of a multiple span bridge using data collected from an instrumented train: a field case study. Sensors, 2022, 22197468.

[57] Molnar C, Freiesleben T, König G et al (2023) Relating the partial dependence plot and permutation feature importance to the data generating process. In: explainable artificial intelligence. Springer Nature Switzerland, Cham, pp. 456–479

[58] MolodovaM, LiZ, NúñezA, et al. . Automatic detection of squats in railway infrastructure. IEEE Trans Intell Transp Syst, 2014, 1551980-1990.

[59] OBrienEJ, QuirkeP, BoweC, et al. . Determination of railway track longitudinal profile using measured inertial response of an in-service railway vehicle. Struct Health Monit, 2018, 1761425-1440.

[60] Lederman G (2016) Data-driven approaches for track monitoring using in-service trains. Dissertation, Carnegie Mellon University

[61] AnastasiadisNP, SakarisCS, SchlanbuschR, et al. . Vibration-based SHM in the synthetic mooring lines of the semisubmersible OO-star wind floater under varying environmental and operational conditions. Sensors, 2024, 242543.

[62] Flores JHF, Engel PM, Pinto RC (2012) Autocorrelation and partial autocorrelation functions to improve neural networks models on univariate time series forecasting. Paper presented at the 2012 international joint conference on neural networks (IJCNN), Brisbane, QLD, Australia, 10–15 June, 2012

[63] TsunashimaH, HiroseR. Condition monitoring of railway track from car-body vibration using time–frequency analysis. Veh Syst Dyn, 2022, 6041170-1187.

[64] CorballyR, MalekjafarianA. Examining changes in bridge frequency due to damage using the contact-point response of a passing vehicle. J Struct Integr Maint, 2021, 63148-158

Funding

Science Foundation Ireland(20/FFP-P/8706)

RIGHTS & PERMISSIONS

The Author(s)