A strategy for out-of-roundness damage wheels identification in railway vehicles based on sparse autoencoders

Abstract

Wayside monitoring is a promising cost-effective alternative to predict damage in the rolling stock. The main goal of this work is to present an unsupervised methodology to identify out-of-roundness (OOR) damage wheels, such as wheel flats and polygonal wheels. This automatic damage identification algorithm is based on the vertical acceleration evaluated on the rails using a virtual wayside monitoring system and involves the application of a two-step procedure. The first step aims to define a confidence boundary by using (healthy) measurements evaluated on the rail constituting a baseline. The second step of the procedure involves classifying damage of predefined scenarios with different levels of severities. The proposed procedure is based on a machine learning methodology and includes the following stages: (1) data collection, (2) damage-sensitive feature extraction from the acquired responses using a neural network model, i.e., the sparse autoencoder (SAE), (3) data fusion based on the Mahalanobis distance, and (4) unsupervised feature classification by implementing outlier and cluster analysis. This procedure considers baseline responses at different speeds and rail irregularities to train the SAE model. Then, the trained SAE is capable to reconstruct test responses (not trained) allowing to compute the accumulative difference between original and reconstructed signals. The results prove the efficiency of the proposed approach in identifying the two most common types of OOR in railway wheels.

References

[1.] Meixedo A, Santos J, Ribeiro D . Damage detection in railway bridges using traffic-induced dynamic responses. Eng Struct 2021, 238 112189

[2.] Meixedo A, Santos J, Ribeiro D . Online unsupervised detection of structural changes using train–induced dynamic responses. Mech Syst Signal Process 2022, 165 108268

[3.] Cury A, Ribeiro D, Ubertini F . Structural health monitoring based on data science techniques. Structural integrity 2021 Cham Springer

[4.] Wang Z, Cha YJ. Unsupervised deep learning approach using a deep auto-encoder with a one-class support vector machine to detect damage. Struct Health Monit 2021, 20 1 406-425

[5.] Sarwar MZ, Cantero D. Deep autoencoder architecture for bridge damage assessment using responses from several vehicles. Eng Struct 2021, 246 113064

[6.] Yessoufou F, Zhu J. Deep autoencoder model for direct monitoring of bridges subjected to a moving vehicle load under varying temperature conditions. Structures 2023, 52 752-767

[7.] Lee H, Lim HJ, Skinner T . Automated fatigue damage detection and classification technique for composite structures using Lamb waves and deep autoencoder. Mech Syst Signal Process 2022, 163 108148

[8.] Wang H, Li B, Gong J . Machine learning-based fatigue life prediction of metal materials: perspectives of physics-informed and data-driven hybrid methods. Eng Fract Mech 2023, 284 109242

[9.] Cha YJ, Choi W, Büyüköztürk O. Deep learning-based crack damage detection using convolutional neural networks. Comput Aided Civil Eng 2017, 32 5 361-378

[10.] Cha Y, Choi W, Suh G . Autonomous structural visual inspection using region-based deep learning for detecting multiple damage types. Comput Aided Civ Infrastruct Eng 2018, 33 9 731-747

[11.] Staśkiewicz T, Firlik B. Out-of-round tram wheels–current state and measurements. AoT 2018, 45 1 83-93

[12.] Chang C, Cai Y, Chen B . Experimental study of the wheel/rail impact caused by wheel flat within 400 km/h using full-scale roller rig. Railw Sci 2022, 1 1 76-89

[13.] Nielsen JCO, Johansson A. Out-of-round railway wheels-a literature survey. Proc Inst Mech Eng Part F J Rail Rapid Transit 2000, 214 2 79-91

[14.] Ye Y, Zhu B, Huang P . OORNet: a deep learning model for on-board condition monitoring and fault diagnosis of out-of-round wheels of high-speed trains. Measurement 2022, 199 111268

[15.] Iwnicki S, Nielsen JCO, Tao G. Out-of-round railway wheels and polygonisation. Veh Syst Dyn 2023, 61 7 1787-1830

[16.] Wu Y, Du X, Zhang HJ . Experimental analysis of the mechanism of high-order polygonal wear of wheels of a high-speed train. J Zhejiang Univ Sci A 2017, 18 8 579-592

[17.] Cai W, Chi M, Wu X . Experimental and numerical analysis of the polygonal wear of high-speed trains. Wear 2019, 440–441 203079

[18.] Vale C. Wheel flats in the dynamic behavior of ballasted and slab railway tracks. Appl Sci 2021, 11 15 7127

[19.] Amini A, Entezami M, Papaelias M. Onboard detection of railway axle bearing defects using envelope analysis of high frequency acoustic emission signals. Case Stud Nondestruct Test Eval 2016, 6 8-16

[20.] Wu Y, Wang J, Liu M . Polygonal wear mechanism of high-speed wheels based on full-size wheel-rail roller test rig. Wear 2022, 494–495 204234

[21.] Ye Y, Wei L, Li F . Multislice time-Frequency image entropy as a feature for railway wheel fault diagnosis. Measurement 2023, 216 112862

[22.] Mosleh A, Montenegro PA, Costa PA . Railway vehicle wheel flat detection with multiple records using spectral kurtosis analysis. Appl Sci 2021, 11 9 4002

[23.] Costa C, Ribeiro D, Jorge P . Calibration of the numerical model of a short-span masonry railway bridge based on experimental modal parameters. Procedia Eng 2015, 114 846-853

[24.] Meixedo A, Ribeiro D, Calçada R, et al (2014) Global and local dynamic effects on a railway viaduct with precast deck. In: The second international conference on railway technology: research, development and maintenance, Ajaccio, Corsica, France, 8–11 April 2014

[25.] Kudva JN, Munir N, Tan PW. Damage detection in smart structures using neural networks and finite-element analyses. Smart Mater Struct 1992, 1 2 108-112

[26.] Lieu QX. A deep neural network-assisted metamodel for damage detection of trusses using incomplete time-series acceleration. Expert Syst Appl 2023, 233 120967

[27.] Silva R, Guedes A, Ribeiro D . Early identification of unbalanced freight traffic loads based on wayside monitoring and artificial intelligence. Sensors 2023, 23 3 1544

[28.] Guedes A, Silva R, Ribeiro D . Detection of wheel polygonization based on wayside monitoring and artificial intelligence. Sensors 2023, 23 4 2188

[29.] Mosleh A, Meixedo A, Ribeiro D . Early wheel flat detection: an automatic data-driven wavelet-based approach for railways. Veh Syst Dyn 2023, 61 6 1644-1673

[30.] Mohammadi M, Mosleh A, Vale C . An unsupervised learning approach for wayside train wheel flat detection. Sensors 2023, 23 4 1910

[31.] Gonzalez I, Karoumi R. BWIM aided damage detection in bridges using machine learning. J Civ Struct Health Monit 2015, 5 5 715-725

[32.] Yang K, Kim S, Harley JB. Unsupervised long-term damage detection in an uncontrolled environment through optimal autoencoder. Mech Syst Signal Process 2023, 199 110473

[33.] Li L, Morgantini M, Betti R. Structural damage assessment through a new generalized autoencoder with features in the quefrency domain. Mech Syst Signal Process 2023, 184 109713

[34.] Wang Z, Cha YJ. Unsupervised machine and deep learning methods for structural damage detection: a comparative study. Eng Rep 2022, 2022 e12551

[35.] Pathirage CSN, Li J, Li L . Structural damage identification based on autoencoder neural networks and deep learning. Eng Struct 2018, 172 13-28

[36.] Finotti RP, de Souza BF, Cury AA . Numerical and experimental evaluation of structural changes using sparse auto-encoders and SVM applied to dynamic responses. Appl Sci 2021, 11 24 11965

[37.] Ye Y, Huang C, Zeng J . Shock detection of rotating machinery based on activated time-domain images and deep learning: an application to railway wheel flat detection. Mech Syst Signal Process 2023, 186 109856

[38.] Meng Q, Catchpoole D, Skillicom D, et al (2017) Relational autoencoder for feature extraction. In: 2017 international joint conference on neural networks (IJCNN). Anchorage, AK, USA. IEEE, pp 364–371

[39.] Mosleh A, Meixedo A, Ribeiro D . Automatic clustering-based approach for train wheels condition monitoring. Int J Rail Transp 2023, 11 5 639-664

[40.] Wang P, Li C, Liang R . Fault detection and calibration for building energy system using Bayesian inference and sparse autoencoder: a case study in photovoltaic thermal heat pump system. Energy Build 2023, 290 113051

[41.] Olshausen BA, Field DJ. Sparse coding with an overcomplete basis set: a strategy employed by V1?. Vis Res 1997, 37 23 3311-3325

[42.] Montenegro PA, Calçada R. Wheel–rail contact model for railway vehicle–structure interaction applications: development and validation. Railw Eng Sci 2023, 31 3 181-206

[43.] Montenegro PA, Neves SGM, Calçada R . Wheel–rail contact formulation for analyzing the lateral train–structure dynamic interaction. Comput Struct 2015, 152 200-214

[44.] Hertz H. Ueber die Berührung fester elastischer Körper. Journal für die reine und angewandte Mathematik (Crelles Journal) 1882, 1882 92 156-171

[45.] Kalker JJ. Book of tables for the Herzian creep-force law 1996 Delft Faculty of Technical Mathematics and Informatics, Delft University of Technology

[46.] The MathWorks Inc (2018) MATLAB^®, R2018a. Natick, Massachusetts

[47.] Ansys Inc (2018) ANSYS^®, Release 19.2. Academic Research, Canonsburg P

[48.] Ribeiro D, Calçada R, Brehm M . Calibration of the numerical model of a track section over a railway bridge based on dynamic tests. Structures 2021, 34 4124-4141

[49.] Ribeiro D, Calçada R, Delgado R . Finite-element model calibration of a railway vehicle based on experimental modal parameters. Veh Syst Dyn 2013, 51 6 821-856

[50.] Mosleh A, Montenegro P, Alves Costa P . An approach for wheel flat detection of railway train wheels using envelope spectrum analysis. Struct Infrastruct Eng 2021, 17 12 1710-1729

[51.] Vale C, Calçada R (2010) Dynamic response of a coupled vehicle–track system to real longitudinal rail profiles. In: Proceedings of the tenth international conference on computational structures technology, Valencia, Spain, 14–17 Sept. 2010

[52.] European Committee for Standardization (2006) Railway applications—track—track geometry quality—part 2: measuring systems—track recording vehicles (EN 13848-2)

[53.] Mosleh A, Costa PA, Calçada R. A new strategy to estimate static loads for the dynamic weighing in motion of railway vehicles. Proc Inst Mech Eng Part F J Rail Rapid Transit 2020, 234 2 183-200

[54.] Casas JR, Moughty JJ. Bridge damage detection based on vibration data: past and new developments. Front Built Environ 2017, 3 4

[55.] Mu J, Zeng J, Huang C . Experimental and numerical investigation into development mechanism of wheel polygonalization. Eng Fail Anal 2022, 136 106152

[56.] Tao G, Xie C, Wang H . An investigation into the mechanism of high-order polygonal wear of metro train wheels and its mitigation measures. Veh Syst Dyn 2021, 59 10 1557-1572

[57.] Zhang J, Han GX, Xiao XB . Influence of wheel polygonal wear on interior noise of high-speed trains. J Zhejiang Univ Sci 2018, A15 1002-1018

[58.] Peng B (2020) Mechanisms of railway wheel polygonization. Dissertaion, University of Huddersfield

[59.] Hu H, Liu J, Zhang X . An effective and adaptable K-means algorithm for big data cluster analysis. Pattern Recognit 2023, 139 109404

[60.] Andrade Nunes L, Piazzaroli Finotti Amaral R, de Souza BF . A hybrid learning strategy for structural damage detection. Struct Health Monit 2021, 20 4 2143-2160

[61.] Alves V, Cury A, Cremona C. On the use of symbolic vibration data for robust structural health monitoring. Proc Inst Civ Eng Struct Build 2016, 169 9 715-723

[62.] Barile C, Casavola C, Pappalettera G . Laplacian score and K-means data clustering for damage characterization of adhesively bonded CFRP composites by means of acoustic emission technique. Appl Acoust 2022, 185 108425

[63.] Meixedo A, Ribeiro D, Santos J . Real-time unsupervised detection of early damage in railway bridges using traffic-induced responses. Structural health monitoring based on data science techniques 2022 Cham Springer 117-142

[64.] Calinski T, Harabasz J. A dendrite method for cluster analysis. Commun Stat Theory Meth 1974, 3 1 780133830

[65.] Solorio-Fernández S, Carrasco-Ochoa JA, Martínez-Trinidad JF. A new hybrid filter–wrapper feature selection method for clustering based on ranking. Neurocomputing 2016, 214 866-880

Funding

Agência Nacional de Inova??o(NORTE-01-0247-FEDER-069595); Funda??o para a Ciência e a Tecnologia(04708/2020)