Defects detection for railway catenary system with encoder-decoder architecture

Abstract

In this research, a robust and efficient damage detection methodology for identifying defects in railway catenary systems is presented. An encoder-decoder architecture supplemented by residual analysis is employed for this purpose. A novel signal segmentation strategy based on the structural features of catenaries is introduced, coupled with a quasi-Welch method designed to mitigate edge effects. The potential impact of GPS inaccuracies on detection precision is examined. Additionally, a comprehensive analysis of various normalization techniques and their significant effects on defect identification outcomes is conducted. Two primary types of defects are considered: hard points in the contact wire and periodic short-wavelength irregularities (PSWI) of the contact wire, with variations in train speeds and defect magnitudes. A defect detection criterion has been developed, facilitating rapid and automatic identification of catenary defects. This integrated approach enables effective detection of defects and accurate determination of their location and can overcome the limitations of previous approaches, such as the requirement for high sampling frequency. This work not only advances the methodology for catenary inspection but also contributes to enhancing the safety and reliability of railway operations. The innovation of this work lies in the integration of the reconstruction capabilities of the encoder-decoder architecture with a residual-based defect detection method. This synergy allows the respective features of each to complement the other effectively.

Keywords

Encoder-decoder architecture / Long-short time memory (LSTM) / Railway catenary system / Residual-based defects detection

References

[1] Chen S, Frøseth GT, Derosa S, et al.. Railway catenary condition monitoring: a systematic mapping of recent research. Sensors (Basel), 2024, 24(3 1023

[2] Frøseth GT, Nåvik P, Rønnquist A. Operational displacement estimations of railway catenary systems by photogrammetry and the integration of acceleration time series. Int J Railw Tech, 2017, 6(3): 71-92

[3] Derosa S, Nåvik P, Collina A, et al.. Railway catenary tension force monitoring via the analysis of wave propagation in cables. Proc Inst Mech Eng Part F J Rail Rapid Transit, 2021, 235(4): 494-504

[4] Carnevale M, Collina A. Processing of collector acceleration data for condition-based monitoring of overhead lines. Proc Inst Mech Eng Part F J Rail Rapid Transit, 2016, 230(2): 472-485

[5] Nåvik P, Derosa S, Rønnquist A. On the use of experimental modal analysis for system identification of a railway pantograph. Int J Rail Transp, 2021, 9(2): 132-143

[6] Jiang T, Frøseth GT, Rønnquist A, et al.. A robust line-tracking photogrammetry method for uplift measurements of railway catenary systems in noisy backgrounds. Mech Syst Signal Process, 2020, 144 106888

[7] Jiang T (2021) Development and application of a vision-based system for structural monitoring of railway catenary system. Dissertation, Norwegian University of Science and Technology

[8] Ambrósio J, Pombo J, Pereira M, et al.. Recent developments in pantograph-catenary interaction modelling and analysis. Int J Railw Tech, 2012, 1(1): 249-278

[9] Antunes P, Ambrósio J, Pombo J, et al.. A new methodology to study the pantograph–catenary dynamics in curved railway tracks. Veh Syst Dyn, 2020, 58(3): 425-452

[10] Bruni S, Bucca G, Carnevale M, et al.. Pantograph–catenary interaction: recent achievements and future research challenges. Int J Rail Transp, 2018, 6(2): 57-82

[11] Liu Z, Liu K, Zhong J, et al.. A high-precision positioning approach for catenary support components with multiscale difference. IEEE Trans Instrum Meas, 2020, 69(3): 700-711

[12] He K, Zhang X, Ren S et al (2016) Deep residual learning for image recognition. In: 2016 IEEE conference on computer vision and pattern recognition (CVPR). Las Vegas, pp 770–778

[13] Deng J, Dong W, Socher R et al (2009) ImageNet: a large-scale hierarchical image database. In: 2009 IEEE Conference on computer vision and pattern recognition. Miami, pp 248–255

[14] Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Commun ACM, 2017, 60(6): 84-90

[15] Zhong J, Liu Z, Wang H et al (2020) Action-driven Reinforcement Learning for Improving Localization of Brace Sleeve in Railway Catenary2020 International Conference on Sensing, Measurement & Data Analytics in the era of Artificial Intelligence (ICSMD). Xi’an, pp 100–105

[16] Ye Y, Zhu B, Huang P, et al.. 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

[17] Ye Y, Gao H, Huang C, et al.. Computer vision for hunting stability inspection of high-speed trains. Measurement, 2023, 220 113361

[18] Song Y, Rønnquist A, Jiang T, et al.. Identification of short-wavelength contact wire irregularities in electrified railway pantograph–catenary system. Mech Mach Theory, 2021, 162 104338

[19] Wang H, Liu Z, Núñez A, et al.. Entropy-based local irregularity detection for high-speed railway catenaries with frequent inspections. IEEE Trans Instrum Meas, 2019, 68(10): 3536-3547

[20] Wang H, Núñez A, Liu Z, et al.. A Bayesian network approach for condition monitoring of high-speed railway catenaries. IEEE Trans Intell Transp Syst, 2020, 21(10): 4037-4051

[21] European Committee for Electrotechnical Standardization (2012) Railway applications—Current collection systems—Requirements for and validation of measurements of the dynamic interaction between pantograph and overhead contact line. EN 50317:2012

[22] Wang Y, Vinogradov A. Simple is good: investigation of history-state ensemble deep neural networks and their validation on rotating machinery fault diagnosis. Neurocomputing, 2023, 548 126353

[23] Cui B, Weng Y, Zhang N. A feature extraction and machine learning framework for bearing fault diagnosis. Renew Energy, 2022, 191: 987-997

[24] Chen X, Zhang B, Gao D. Bearing fault diagnosis base on multi-scale CNN and LSTM model. J Intell Manuf, 2021, 32(4): 971-987

[25] Bharti R, Khamparia A, Shabaz M, et al.. Prediction of heart disease using a combination of machine learning and deep learning. Comput Intell Neurosci, 2021, 2021: 8387680

[26] Maragatham G, Devi S. Retracted article: LSTM model for prediction of heart failure in big data. J Med Syst, 2019, 43(5): 111

[27] Li S, Yang J, Teng P. Data-driven prediction of cylinder-induced unsteady wake flow. Appl Ocean Res, 2024, 150 104114

[28] Zhang Q, Zhu S, Yuan Z et al (2022) Compound damage discrimination for steel-spring floating-slab track using 1D convolutional neural network. In: Proceedings of the second international conference on rail transportation. Chengdu, pp 33–40

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

[30] Sarwar MZ, Cantero D. Probabilistic autoencoder-based bridge damage assessment using train-induced responses. Mech Syst Signal Process, 2024, 208 111046

[31] Zhang S, Huang P, Yan W. A data-driven approach for railway in-train forces monitoring. Adv Eng Inform, 2024, 59 102258

[32] Kulkarni R, Qazizadeh A, Berg M. Unsupervised rail vehicle running instability detection algorithm for passenger trains (iVRIDA). Measurement, 2023, 216 112894

[33] Fink O, Wang Q, Svensén M, et al.. Potential, challenges and future directions for deep learning in prognostics and health management applications. Eng Appl Artif Intell, 2020, 92 103678

[34] Hochreiter S, Schmidhuber J. Long short-term memory. Neural Comput, 1997, 9(81735-1780

[35] Cho K, van Merrienboer B, Gulcehre C et al (2014) Learning phrase representations using RNN encoder-decoder for statistical machine translation. In: Proceedings of the 2014 conference on empirical methods in natural language processing. Doha, pp 1724–1734

[36] Chung J, Gulcehre C, Cho K et al (2014) Empirical evaluation of gated recurrent neural networks on sequence modeling. https://arxiv.org/abs/1412.3555v1. Accessed 4 July 2024

[37] Song Y, Lu X, Yin Y, et al.. Optimization of railway pantograph-catenary systems for over 350 km/h based on an experimentally validated model. IEEE Trans Ind Inform, 2024, 20(57654-7664

[38] Song Y, Jiang T, Nåvik P, et al.. Geometry deviation effects of railway catenaries on pantograph–catenary interaction: a case study in Norwegian Railway System. Railw Eng Sci, 2021, 29(4): 350-361

[39] Standard B. Railway applications—Fixed installations and rolling stock—Criteria to achieve technical compatibility between pantographs and overhead contact line. NEK EN, 2020, 50367: 2020

[40] U.S. Government (2020) GPS SPS performance standard. https://www.gps.gov/technical/ps/2020-SPS-performance-standard.pdf. Accessed 4 July 2024

[41] Welch P. The use of fast Fourier transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Trans Audio Electroacoust, 1967, 15(270-73

Funding

Norwegian railway directorate

RIGHTS & PERMISSIONS

The Author(s)