Rail defects, especially transverse defects (TDs), can pose risks to safe and efficient railroad operations. Effective rail defect detection is critical for the prevention of broken rail-induced accidents and derailments. In this study, a deep autoencoder (DAE) rail defect detection framework is developed to process ultrasonic A-scan data collected by a roller search unit and to identify the presence of TDs in rail samples. An autoencoder is a semi-supervised learning algorithm that identifies observations in a dataset that significantly deviate from the remaining observations and can be used for rail defect detection. Ultrasonic A-scan signals collected from both pristine and damaged rail segments are analyzed, where the pristine dataset is used to train a DAE model. To improve the accuracy and sensitivity of defect detection, we optimize the architecture and hyperparameters of the DAE model. Moreover, we evaluate the performance of two features extracted from the DAE model through receiver operating characteristic curves and confusion matrix. The DAE features outperformed conventional knowledge-driven features in the accuracy and robustness of defect detection, especially with the presence of noise.
FRA (Federal Railroad Administration). Track Inspector Rail Defect Reference Manual, Federal Railroad Administration, Washington, DC, 2015, pp. 1–82.
4.
ClarkR.Rail Flaw Detection: Overview and Needs for Future Developments. NDT&E International, Vol. 37, No. 2, 2004, pp. 111–118.
5.
GarciaG.ZhangJ.Application of Ultrasonic Phased Arrays for Rail Flaw Inspection. FRA Report. Federal Railroad Administration, Washington, D.C., 2006.
6.
Al-NazerL.WelanderL.Rail Flaw Sizing Using Conventional and Phased Array Ultrasonic Testing, FRA Report. Federal Railroad Administration, Washington, D.C., 2012.
7.
SterniniS.LiangA. Y.Lanza di ScaleaF.Ultrasonic Synthetic Aperture Imaging with Interposed Transducer–Medium Coupling Path. Structural Health Monitoring, Vol. 18, No. 5–6, 2018, pp. 1543–1556.
8.
CawleyP.LoweM. J.AlleyneD. N.PavlakovicB.WilcoxP.Practical Long Range Guided Wave Inspection-Applications to Pipes and Rail. Materials Evaluation, Vol. 61, No. 1, 2003, pp. 66–74.
9.
BruzeliusK.MbaD.An Initial Investigation on the Potential Applicability of Acoustic Emission to Rail Track Fault Detection. NDT&E International, Vol. 37, No. 7, 2004, pp. 507–516.
10.
PohlR.ErhardA.MontagH. J.ThomasH. M.WustenbergH.NDT Techniques for Railroad Wheel and Gauge Corner Inspection. NDT&E International, Vol. 37, No. 2, 2004, pp. 89–94.
11.
FitzgeraldC. S.Inspection for Rail Defects by Magnetic Induction. Nondestructive Evaluation of Aging Railroads, Vol. 2458, 1995, pp. 40–44.
12.
BriggsC. W.Developments in Gamma Ray Radiography 1928 to 1941. II. A History of Nondestructive Testing. Materials Evaluation, Vol. 34, No. 3, 1976, p. p. vp.
13.
Lanza di ScaleaF.ZhuX.CapriottiM.LiangA. Y.MarianiS.SterniniS.Passive Extraction of Dynamic Transfer Function From Arbitrary Ambient Excitations: Application to High-Speed Rail Inspection From Wheel-Generated Waves. Journal of Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems, Vol. 1, No. 1, 2018, p. 011005.
14.
MarianiS.NguyenT.ZhuX.Lanza di ScaleaF.Field Test Performance of Noncontact Ultrasonic Rail Inspection System. Journal of Transportation Engineering, Part A: Systems, Vol. 143, No. 5, 2017, p. 04017007.
15.
WuY.ZhuX.BaillargeonJ.Deep Autoencoder for Ultrasound-Based Rail Flaw Detection. ASME/IEEE Joint Rail Conference, Vol. 85758, April2022, p. V001T09A006. American Society of Mechanical Engineers.
MahadevanV.LiW.BhalodiaV.VasconcelosN.Anomaly Detection in Crowded Scenes. 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, IEEE, San Francisco, CA, USA, 2010. pp. 1975–1981.
18.
LiW.MahadevanV.VasconcelosN.Anomaly Detection and Localization in Crowded Scenes. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 36, No. 1, 2013, pp. 18–32.
19.
XuD.RicciE.YanY.SongJ.SebeN.Learning Deep Representations of Appearance and Motion for Anomalous Event Detection, 2015. arXiv preprint arXiv:1510.01553.
20.
BeggelL.PfeifferM.BischlB.Robust Anomaly Detection in Images Using Adversarial Autoencoders. Joint European Conference on Machine Learning and Knowledge Discovery in Databases, Würzburg, Germany, 2019, pp. 206–222.
21.
BorghesiA.BartoliniA.LombardiM.MilanoM.BeniniL.Anomaly Detection Using Autoencoders in High Performance Computing Systems. Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33, No. 1, 2019, pp. 9428–9433.
22.
LuoH.ZhongS.Gas Turbine Engine Gas Path Anomaly Detection using Deep Learning with Gaussian Distribution. 2017 Prognostics and System Health Management Conference (PHM-Harbin) IEEE, Harbin, China, July 9–12, 2017, pp. 1–6.
23.
DesmetA.DeloreM.Leak Detection in Compressed Air Systems Using Unsupervised Anomaly Detection Techniques. Annual Conference of the PHM Society, St. Petersburg, Florida, Vol. 9, No. 1, 2017.
24.
MichauG.PalmT.FinkO.Deep Feature Learning Network for Fault Detection and Isolation. Annual Conference of the PHM Society, St. Petersburg, FL, Vol. 9, No. 1, 2017.
25.
MeireM.KarsmakersP.Comparison of Deep Autoencoder Architectures for Real-Time Acoustic Based Anomaly Detection in Assets. 2019 10th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS), Metz, France, 2019, pp. 786–790.
26.
XuyunF.HuiL.ZhongS.LinL.Aircraft Engine Fault Detection Based on Grouped Convolutional Denoising Autoencoders. Chinese Journal of Aeronautics, Vol. 32, No. 2, 2019, pp. 296–307.
27.
NiF.ZhangJ.NooriM. N.Deep Learning for Data Anomaly Detection and Data Compression of a Long-Span Suspension Bridge. Computer-Aided Civil and Infrastructure Engineering, Vol. 35, No. 7, 2020, pp. 685–700.
28.
LutzM. A.VogtS.BerkhoutV.FaulstichS.DienstS.SteinmetzU.OrtegaA.Evaluation of Anomaly Detection of an Autoencoder Based on Maintenance Information and SCADA-Data. Energies, Vol. 13, No. 5, 2020, p. 1063.
29.
MaoJ.WangH.Spencer Jr.B. F.Toward Data Anomaly Detection for Automated Structural Health Monitoring: Exploiting Generative Adversarial Nets and Autoencoders. Structural Health Monitoring, Vol. 20, No. 4, 2021, pp. 1609–1626.
30.
YanW.YuL.On Accurate and Reliable Anomaly Detection for Gas Turbine Combustors: A Deep Learning Approach, 2019. arXiv preprint arXiv:1908.09238.
31.
LeeH.LimH. J.SkinnerT.ChattopadhyayA.HallA.Automated Fatigue Damage Detection and Classification Technique for Composite Structures Using Lamb Waves and Deep Autoencoder. Mechanical Systems and Signal Processing, Vol. 163, 2022, p. 108148.
32.
ShinT.SunH.ZhuJ.ZhangY.Nondestructive Damage Detection of Concrete With Alkali-Silica Reactions Using Coda Wave and Anomaly Detection. IEEE Sensors Journal, Vol. 22, No. 6, 2022, pp. 6124–6135.
33.
PashoutaniS.ZhuJ.SimC.WonK.MazzeoB. A.GuthrieW. S.Multi-Sensor Data Collection and Fusion Using Autoencoders in Condition Evaluation of Concrete Bridge Decks. Journal of Infrastructure Preservation and Resilience, Vol. 2, No. 1, 2021, pp. 1–12.
34.
VincentP.LarochelleH.LajoieI.BengioY.ManzagolP. A.BottouL.Stacked Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion. Journal of Machine Learning Research, Vol. 11, No. 12, 2010, pp. 3371–3408.
35.
AnJ.ChoS.Variational Autoencoder Based Anomaly Detection Using Reconstruction Probability. Special Lecture on IE, Vol. 2, No. 1, 2015, pp. 1–18.
36.
ByrdR. H.ChinG. M.NocedalJ.WuY.Sample Size Selection in Optimization Methods for Machine Learning. Mathematical Programming, Vol. 134, No. 1, 2012, pp. ,127–155.
37.
EsterM.KriegelH. P.SanderJ.XuX.A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases With Noise. KDD-96 Proceedings, Vol. 96, No. 34, 1996, pp. 226–231.
