Remaining Useful Life Prediction for Railway Switch Engines Using Classification Techniques

##plugins.themes.bootstrap3.article.main##

##plugins.themes.bootstrap3.article.sidebar##

Published Nov 17, 2020
Thomas B¨ohm

Abstract

A highly available infrastructure is a premise for capable railway operation of high quality. Therefore, maintenance is necessary to keep railway infrastructure elements available. Railway switches, especially, are critical because they connect different tracks and allow a train to change its moving direction without stopping. Their inspection, maintenance and repair have long been identified as a cost driver. Switch failures, particularly, are responsible for a comparable high number of failures and delay minutes. The reduction of failures would not only save maintenance costs, but also let more trains arrive on time and hence increase the attractiveness of the railway transport. Therefore, upcoming failures need to be revealed early enough to allow an effective planning and execution of failure preventing maintenance activities. Research is exploring ways to predict the remaining useful life of switches.
This paper presents an approach to predict the remaining useful life (RUL) of railway switch engine failures. The development is based on measurement data of the electrical power consumption of switch engines. The two year time series of 29 switches of Deutsche Bahn was recorded by a commercial switch diagnostic system leading to roughly 250 000
measurement tuples. Since earlier researched showed that the electrical data alone is not sufficient enough, additional data is integrated. It takes into account the dependency of the switch condition data from climatic conditions and certain properties of the switch construction type.
Predicting a RUL is quite challenging in many PHM applications. To avoid common problems with uncertainty in measurement data, a long prediction horizon (month) of small time units (hours) and to stabilise end user acceptance the approach transforms the RUL prediction problem into a classification problem of multiple classes. It, then, uses two different supervised classification techniques, Artificial Neural Networks (aNN) and Support Vector Machines (SVM), to predict the RUL in the form of classes. However, as known from the no free lunch-theorem of classification, there is no ultimately best performing technique. The success depends on the problem and data structure as well as on the parametrisation of the technique or the selected algorithm respectively. Especially aNN and SVM have a high number of possible parametrisations. They can fail the task or result in a very good performance under the heavy influence of their parametrisation. Hence, it is an important aspect of this paper to share how the different parameters effect the RUL prediction and which parameters result in maximum performance. In order to compare the performance, two metrics are chosen, the Matthews Correlation Coefficient (MCC) as single value metric and a visualisation of the confusion matrix as more comprehensible metric. Finally, deriving those parameters maximising the RUL prediction results enables one of the two classification techniques to reveal upcoming failures of the switch engine early enough to prevent them.

Abstract 531 | PDF Downloads 361

##plugins.themes.bootstrap3.article.details##

Keywords

Condition monitoring, anomaly detection, time series prediction, railway infrastructure

References
Asada, T., Roberts, C., & Koseki, T. (2013). An algorithm for improved performance of railway condition monitoring equipment: Alternating-current point machine case study. Transportation Research Part C: Emerging Technologies, 30(0), 81–92.
Atamuradov, V., Camci, F., Baskan, S., & Sevkli, M. (2009). Failure diagnostics for railway point machines using expert systems. In Ieee international symposium on diagnostics for electric machines, power electronics and drives, 2009 (pp. 1–5). Piscataway (NJ USA): IEEE.
Bai, H. (2010). A generic fault detection and diagnosis approach for pneumatic and electric driven railway assets (Unpublished doctoral dissertation). University of Birmingham, Birmingham.
Ben-Hur, A., & Weston, J. (2010). A user’s guide to support vector machines. In O. Carugo & F. Eisenhaber (Eds.), Data mining techniques for the life sciences (Vol. 609, pp. 223–239). New York (NY USA): Humana Press. Retrieved from http://dx.doi.org/10.1007/978-1-60327 -241-4n 13 doi: 10.1007/978-1-60327-241-4n 13
B¨ohm, T. (2012). Accuracy improvement of condition diagnosis of railway switches via external data integration. In C. Boller (Ed.), Structural health monitoring 2012 (pp. 1550–1558). Germany.
B¨ohm, T. (2015). Pr¨azise vorhersage von weichenst¨orungen. EI - Eisenbahningenieur, 66.(10), 50–53.
B¨ohm, T., & Doegen, C. (2010). Diagnosis without sensors - integration of external data for condition monitoring of railway switches. In S. Okumura, T. Kawai, P. Chen, & R. B. Rao (Eds.), Comadem 2010 - advances in maintenance and condition diagnosis technologies towards sustainable society (pp. 619–622). Hikone (Japan): Sunrise Publishing.
Burges, C. J. C. (1998). A tutorial on support vector machines for pattern recognition. Data Mining and Knowledge Discovery, 2, 121–167.
Chamroukhi, F., Samé, A., Aknin, P., & Antoni, M. (2008). Switch mechanism diagnosis using a pattern recognition approach. In The 4th iet international conference on railway condition monitoring (pp. 1–4). Derby (UK).
Cocciaglia, D. (2012). Case study: Switch&crossing diagnostics. In London Business Conferences (Ed.), Cost optimisation track maintenance and renewal congress 2012.
DB Netz AG. (April 2016). Infrastrukturzustandsund-entwicklungsbericht 2015: Internetversion. Frankfurt am Main. Retrieved 15.10.2016, from http://www.eba.bund.de/SharedDocs/Publikationen/DE/Finanzierung/IZB/IZBn 2015.html;jsessionid=35B3BE62A8AA227E32055FF48043F670.live11292?nn=491736
Duda, R. O., Hart, P. E., & Stork, D. G. (2012). Pattern classification (2. Aufl. ed.). s.l.: Wiley-Interscience.
Eker, O. F., Camci, F., & Kumar, U. (2010). Failure diagnostics on railway turnout systems using support vector machines. In U. Kumar (Ed.), emaintenance 2010 (pp. 248–251). Lulea (Sweden): Univ.
Ferri, C., Hernández-Orallo, J., & Modroiu, R. (2009). An experimental comparison of performance measures for classification. Pattern Recognition Letters, 30(1), 27–38. Retrieved from http://www.sciencedirect.com/science/article/pii/S0167865508002687
García Márquez, F. P., Weston, P., & Roberts, C. (2007). Failure analysis and diagnostics for railway trackside equipment. Engineering Failure Analysis, 14(8), 1411–1426. Retrieved from http://www.sciencedirect.com/science/article/pii/S1350630707000556
Hand, D. J., Mannila, H., & Smyth, P. (2001). Principles of data mining. Cambridge (Mass. USA): MIT Press. Retrieved from http://www.gbv.de/dms/bowker/toc/9780262082907.pdf
Hsu, C.-W., Chang, C.-C., & Lin, C.-J. (2003). A practical guide to support vector classification. Retrieved from http://www.csie.ntu.edu.tw/cjlin/papers.html
Jurman, G., & Furlanello, C. (2010). A unifying view for performance measures in multi-class prediction. ArXiv e-prints(Aug), 1–5.
Keerthi, S. S., Shevade, S. K., Bhattacharyya, C., & Murthy, K. R. K. (2001). Improvements to platt’s smo algorithm for svm classifier design. Neural Computation, 13(3), 637–649. doi: 10.1162/089976601300014493
Leao, B. P., Yoneyama, T., Rocha, G. C., & Fitzgibbon, K. T. (2008). Prognostics performance metrics and their relation to requirements, design, verification and costbenefit. In International conference on prognostics and health management, 2008 (pp. 1–8). Piscataway (NJ USA): IEEE.
Lee, J. (2010). Design of selfmaintenance and engineering immune systems for smarter machines and manufacturing systems. In S. Okumura, T. Kawai, P. Chen, & R. B. Rao (Eds.), Comadem 2010 - advances in maintenance and condition diagnosis technologies towards sustainable society (pp. 1–13). Hikone (Japan): Sunrise Publishing.
Matthews, B. W. (1975). Comparison of the predicted and observed secondary structure of t4 phage lysozyme. Biochimica et biophysica acta, 405(2), 442–451. Retrieved from http://view.ncbi.nlm.nih.gov/pubmed/1180967
Min, J. H., & Lee, Y.-C. (2005). Bankruptcy prediction using support vector machine with optimal choice of kernel function parameters. Expert Systems with Applications, 28(4), 603–614. doi: 10.1016/j.eswa.2004.12.008
Pedregal, D. J., García Márquez, F. P., & Roberts, C. (2009). An algorithmic approach for maintenance management based on advanced state space systems and harmonic regressions. Annals of Operations Research, 166(1), 109–124.
Platt, J. C. (1999). Fast training of support vector machines using sequential minimal optimization: Advances in kernel methods. In B. Sch¨olkopf, C. J. C. Burges, & A. J. Smola (Eds.), Advances in kernel methods (pp. 185–208). Cambridge (Mass. USA): MIT Press. Retrieved from http://dl.acm.org/citation.cfm?id=299094.299105
Pool, C., & Vlek, J. (2016). Making the switch: Predictive maintenance on railway switches. In databricks (Ed.), Spark summit europe 2016. Retrieved 23.04.2017, from https://youtu.be/HEDMET3K6nA
Rama, D., & Andrews, J. D. (2013). A reliability analysis of railway switches. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit, 227(4), 344–363. doi: 10.1177/0954409713481725
Rausch, W. (2010). Diagnosesysteme f¨ur weichen als grundlage f¨ur eine optimierte instandhaltungsplanung. In VDEI (Ed.), Symposium zum thema: ”moderne instandhaltungsverfahren f¨ur weichen – qualit¨atsanspr¨uche – wirtschaftlichkeit“. Brandenburg-Kirchm¨oser.
Riedmiller, M., & Braun, H. (1993). A direct adaptive method for faster backpropagation learning: therprop algorithm. In Proceedings of the ieee international conference on neural networks (pp. 586–591). Piscataway (NJ USA).
Sebastiani, P., Abad, M. M., & Ramoni, M. F. (2010). Bayesian networks. In O. Maimon & L. Rokach (Eds.), Data mining and knowledge discovery handbook (pp. 175–208). Boston (MA USA): Springer Science+ Business Media LLC.
Shmilovici, A. (2010). Support vector machines. In O. Maimon & L. Rokach (Eds.), Data mining and knowledge discovery handbook (pp. 231–247). Boston (MA USA): Springer Science+Business Media LLC.
Silmon, J. A. (2009). Operational industrial fault detection and diagnosis: Railway actuator case studies (Unpublished doctoral dissertation). University of Birmingham, Birmingham.
Sokolova, M., & Lapalme, G. (2009). A systematic analysis of performance measures for classification tasks. Information Processing and Management, 45(4), 427–437.
Stoll, H., & Bollrath, B. (2002). Weichendiagnosesystem sidis w. Signal + Draht, 94(4), 26–29.
University of Birmingham. (2008). Innotrack (fp6 eu project): Deliverable d3.3.1 – list of key parameters for switch and crossing monitoring. Br¨ussel.
Vapnik, V. N. (2008). The nature of statistical learning theory (2. ed., 6. print ed.). New York: Springer. Retrieved from http://www.loc.gov/catdir/enhancements/fy0816/99039803-d.html/http://www.loc.gov/catdir/enhancements/fy0816/99039803-t.html/http://www.gbv.de/dms/hbz/toc/ht012773848.pdf
Zhang, P. G. (2010). Neural networks for data mining. In O. Maimon & L. Rokach (Eds.), Data mining and knowledge discovery handbook (pp. 419–444). Boston (MA USA): Springer Science+Business Media LLC.
Zwanenburg, W.-J. (2009). Degradation processes of railway switches & crossings: To improve maintenance & renewal planning on the swiss railway network. Saarbr¨ucken: S¨udwestdeutscher Verlag f¨ur Hochschulschriften.
Section
Technical Papers