Ballasted Track Behaviour Induced by Absent Sleeper Support and its Detection Based on a Convolutional Neural Network Using Track Data

Abstract

With development of the heavy-haul railway, the increased axle load and traction weight bring a significant challenge for the service performance and safety maintenance of the railway track. Conducting defect recognition on concrete sleepers and ballast using big data is vital. This paper focused on the detection of absent sleeper support in a ballasted track with an emphasis on the integration of model-based and data-driven methods. To this end, a mathematical model consisting of the wagon, track and wheel–rail contact subsystems was first established to acquire the necessary raw data for the data-driven method, in which the wagon was regarded as a 47-degree-of-freedom multi-body subsystem, and the track was treated as a multi-layer discrete-elastic support beam subsystem with absent sleeper support. Then, an architectural hierarchy of a three-layer  convolutional neural network (TLCNN) was developed, which includes three convolutional layers and two pooling layers, and a method for reconstructing one-dimensional sleeper vertical displacement to a two-dimensional time–space matrix was also proposed. Thirdly, verification was carried out by comparing the simulation and experimental results to illustrate the accuracy and reliability of the mathematical model, and the dynamic behaviour of the track with absent sleeper support was investigated. Lastly, the established TLCNN was used to train the raw data of the sleeper vertical displacement and detect the existence of absent sleeper support. Results show that the integration of model-based and data-driven methods was a reliable and effective approach for the detection of absent sleeper support. The proposed TLCNN can acquire and extract robust characteristics in a noisy environment. To handle more complex recognition tasks and further improve performance, deeper CNN models and larger sample sizes should be preferentially considered in practical applications.

Keywords

Ballasted track / Absent sleeper support / Detection / Data-driven method / Model-based method

References

[1] Li C, Luo S, Cole C A signal-based fault detection and classification method for heavy haul wagons. Veh Syst Dyn, 2017, 55(12): 1807-1822

[2] Zhang D, Zhai W, Wang K. Dynamic interaction between heavy-haul train and track structure due to increasing axle load. Aust J Struct Eng, 2017, 18(3): 190-203

[3] Mei H, Leng W, Nie R Random distribution characteristics of peak dynamic stress on the subgrade surface of heavy-haul railways considering track irregularities. Soil Dyn Earthq Eng, 2019, 116: 205-214

[4] Zhang X, Zhao C, Zhai W Investigation of track settlement and ballast degradation in the high-speed railway using a full-scale laboratory test. Proc Inst Mech Eng F J Rail Rapid Transit, 2019, 233(8): 869-881

[5] Ferreira PA, Lopez-Pita A. Numerical modeling of high-speed train/track system to assess track vibrations and settlement prediction. J Transp Eng, 2013, 139(3): 330-337

[6] Li X, Palsson BA, Nielsen JCO. Simulation of track settlement in railway turnouts. Veh Syst Dyn, 2014, 52(sup1): 421-439

[7] Shi J, Chan AH, Burrow MP. Influence of unsupported sleepers on the dynamic response of a heavy haul railway embankment. Proc Inst Mech Eng F J Rail Rapid Transit, 2013, 227(6): 657-667

[8] Lam HF, Wong MT, Yang YB. A feasibility study on railway ballast damage detection utilizing measured vibration of in situ concrete sleeper. Eng Struct, 2012, 45: 284-298

[9] Kaewunruen S, Remennikov AM. Investigation of free vibrations of voided concrete sleepers in railway track system. Proc Inst Mech Eng F J Rail Rapid Transit, 2007, 221: 495-507

[10] Rezaei E, Dahlberg T. Dynamic behaviour of an in situ partially supported concrete railway sleeper. Proc Inst Mech Eng F J Rail Rapid Transit, 2011, 225: 501-508

[11] Zhu JJ, Ahmed AKW, Rakheja S Development of a vehicle-track model assembly and numerical method for simulation of wheel-rail dynamic interaction due to unsupported sleepers. Veh Syst Dyn, 2010, 48(12): 1535-1552

[12] Zhu JY, Thompson DJ, Jones CJC. On the effect of unsupported sleepers on the dynamic behaviour of a railway track. Veh Syst Dyn, 2011, 49(9): 1389-1408

[13] Recuero AM, Escalona JL, Shabana AA. Finite-element analysis of unsupported sleepers using three-dimensional wheel–rail contact formulation. Proc Inst Mech Eng K J Multi-Body Dyn, 2011, 225: 153-165

[14] Bezin Y, Iwnicki SD, Cavalletti M An investigation of sleeper voids using a flexible track model integrated with railway multi-body dynamics. Proc Inst Mech Eng F J Rail Rapid Transit, 2009, 223(6): 597-607

[15] Lundqvist A, Dahlberg T. Load impact on railway track due to unsupported sleepers. Proc Inst Mech Eng F J Rail Rapid Transit, 2005, 219: 67-77

[16] Sysyn M, Nabochenko O, Kovalchuk V. Experimental investigation of the dynamic behavior of railway track with sleeper voids. Rail Eng Sci, 2020, 28: 290-304

[17] Sysyn M, Przybylowicz M, Nabochenko O, Liu J. Mechanism of sleeper–ballast dynamic impact and residual settlements accumulation in zones with unsupported sleepers. Sustainability, 2021, 13: 7740

[18] Balouchi F, Bevan A, Formston R (2016) Detecting railway under-track voids using multi-train in-service vehicle accelerometer. In: 7th IET conference on railway condition monitoring 2016 (RCM 2016), Birmingham, UK

[19] Clark A, Kaewunruen S, Janeliukstis R, Papaelias M (2017) Damage detection in railway prestressed concrete sleepers using acoustic emission. In: 3rd International Conference on Innovative Materials, Structures and Technologies (IMST 2017), Riga, Latvia

[20] Kaewunruen S, Janeliukstis R, Freimanis A, Goto K. Normalised curvature square ratio for detection of ballast voids and pockets under rail track sleepers. J Phys: Conf Ser, 2018, 1106: 012002

[21] Lee J-H, Magno K, Joh S-H. Utilization of spectral velocity of flexural waves to detect loose sleepers. J Eng Sci Technol, 2018, 13: 1411-1419.

[22] Ankrah AA, Kimotho JK, Muvengei OM. Fusion of model-based and data driven based fault diagnostic methods for railway vehicle suspension. J Intell Learn Syst Appl, 2020, 12(3): 51-81

[23] Janssens O, Slavkovikj V, Vervisch B Convolutional neural network based fault detection for rotating machinery. J Sound Vib, 2016, 377: 331-345

[24] Wen L, Li X, Gao L A new convolutional neural network-based data-driven fault diagnosis method. IEEE Trans Industr Electron, 2018, 65(7): 5990-5998

[25] Cha YJ, Choi W, Büyüköztürk O. Deep learning-based crack damage detection using convolutional neural networks. Comput Aided Civ Infrastruct Eng, 2017, 32(5): 361-378

[26] Guo X, Chen L, Shen C. Hierarchical adaptive deep convolution neural network and its application to bearing fault diagnosis. Measurement, 2016, 93: 490-502

[27] Hu H, Tang B, Gong X Intelligent fault diagnosis of the high-speed train with big data based on deep neural networks. IEEE Trans Industr Electron, 2017, 13(4): 2106-2116

[28] Sresakoolchai J, Kaewunruen S. Detection and severity evaluation of combined rail defects using deep learning. Vibration, 2021, 4: 341-356

[29] Sresakoolchai J, Kaewunruen S. Prognostics of unsupported railway sleepers and their severity diagnostics using machine learning. Sci Rep, 2022, 12: 6064

[30] Sresakoolchai J, Kaewunruen S. Railway defect detection based on track geometry using supervised and unsupervised machine learning. Struct Health Monit, 2022, 21: 1757-1767

[31] Huang J, Kaewunruen S. Evaluation of railway passenger comfort with machine learning. IEEE Access, 2022, 10: 2372-2381

[32] Sysyn M, Przybylowicz M, Nabochenko O, Kou L. Identification of sleeper support conditions using mechanical model supported data-driven approach. Sensors, 2021, 21: 3609

[33] Kaewunruen S, Sresakoolchai J, Stittle H. Machine learning to identify dynamic properties of railway track components. Int J Str Stab Dyn, 2022, 22: 2250109

[34] Kaewunruen S, Sresakoolchai J, Thamba A. Machine learning-aided identification of train weights from railway sleeper vibration. Insight, 2021, 63: 151-159

[35] Sresakoolchai J, Kaewunruen S. Wheel flat detection and severity classification using deep learning techniques. Insight, 2021, 63: 393-402

[36] Kaewunruen S, Sresakoolchai J, Zhu G. Machine learning aided rail corrugation monitoring for railway track maintenance. Struct Monit Maint Int J, 2021, 8(2): 151-166.

[37] Zhai WM, Cai CB, Guo SZ. Coupling model of vertical and lateral vehicle/track interactions. Veh Syst Dyn, 1996, 26(1): 61-79

[38] Zhai WM, Sun X. A detailed model for investigating vertical interaction between railway vehicle and track. Veh Syst Dyn, 1994, 23: 603-615

[39] Sun YQ, Dhanasekar M. A dynamic model for the vertical interaction of the rail track and wagon system. Phys Lett B, 2002, 89(2): 169-172

[40] Xu L, Chen X, Li X Development of a railway wagon-track interaction model: case studies on excited tracks. Mech Syst Signal Process, 2018, 100: 877-898

[41] Sun W, Zhou J, Thompson D Vertical random vibration analysis of vehicle–track coupled system using Green's function method. Veh Syst Dyn, 2014, 52(3): 362-389

[42] Zhai WM, Wang KY, Lin JH. Modelling and experiment of railway ballast vibrations. J Sound Vib, 2004, 270(4): 673-683

[43] Zhang D, Wang K, Zhai W et al (2017) Track random irregularity analysis for heavy-haul railway. In: Proceedings of the 1st international conference on rail transportation. Chengdu. https://doi.org/10.1061/9780784481257.001

[44] Zhai WM. Vehicle-track coupled dynamics: theory and applications, 2020 1st English Singapore: Springer Nature

[45] Sysyn M, Kluge F, Gruen D Experimental analysis of rail contact fatigue damage on frog rail of fixed common crossing 1:12. J Fail Anal Prev, 2019, 19: 1077-1092

[46] Sysyn M, Gerber U, Nabochenko O Common crossing fault prediction with track based inertial measurements: statistical vs. mechanical approach. Pollack Periodica, 2019, 14(2): 15-26

[47] He H, Garcia EA. Learning from imbalanced data. IEEE Trans Knowl Data Eng, 2009, 21(9): 1263-1284

[48] He K, Zhang X, Ren S et al (2016) Deep residual learning for image recognition. arXiv:1512.03385

[49] Chollet F (2017) Xception: deep learning with depthwise separable convolutions. arXiv:1610.02357

[50] Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556

Funding

Open project of State Key Laboratory of Performance Monitoring and Protecting of Rail Transit Infrastructure(HJGZ2022115)

Chinese Postdoctoral Science Foundation(2021M693752)

State Key Laboratory of Traction Power(TPL2108)

National Natural Science Foundation of China(12202107)

Fundamental Research Funds for the Central Universities(300102342104)

Shannxi Science and Technology Project(2023-JC-YB-496)