PSI - Issue 57

Olivier Vo Van et al. / Procedia Structural Integrity 57 (2024) 104–111

111

8

O. VoVan / Fatigue Design 2023 00 (2023) 000–000

Considering time evolution, it is of interest to examine how alterations in temperature trajectories, particularly due to climate change, a ff ect the occurrence of fatigue defects. To explore this question, a suitable approach is to employ transfer learning, specifically focusing on covariate shift. Covariate shift refers to a type of problem wherein only the distribution of variable X is modified. Namely, P ( X ) is to change due to evolution of temperature while the concept P ( Y | X ), to be understood as the link between the properties and experience of the rail with regard to the appearance of fatigue defects, remains unchanged. There are many methods for studying covariate shift based on random forests [23] or kernel-based methods [8], which will be explored in future work.

References

[1] Breiman, L., 2001. Random forests. Machine learning 45, 5–32. [2] Breiman, L., Friedman, J.H., Olshen, R.A., Stone, C.J., 1984. Classification and regression trees. brooks. Wadsworth and Brooks, Monterey, CA . [3] Chaboche, J.L., 1981. Continuous damage mechanics—a tool to describe phenomena before crack initiation. Nuclear engineering and design 64, 233–247. [4] Chapman, L., Thornes, J., Huang, Y., Cai, X., Sanderson, V., White, S., 2008. Modelling of rail surface temperatures: a preliminary study. Theoretical and Applied Climatology 92, 121–131. [5] Chapman, L., Thornes, J.E., White, S.P., 2006. Thermal imaging of railways to identify track sections prone to buckling. Proceedings of the Institution of Mechanical Engineers, Part F: Journal of Rail and Rapid Transit 220, 317–327. [6] EN 13674-1:2011, 2011. Railway applications - Track - Rail - Part 1 : vignole railway rails 46 kg / m and above. Standard. National standards and national normative documents. [7] on Fatigue, A.C.E.., Fracture, 2005. Standard practices for cycle counting in fatigue analysis. ASTM International. [8] Gretton, A., Smola, A., Huang, J., Schmittfull, M., Borgwardt, K., Scho¨lkopf, B., 2013. Covariate shift by kernel mean matching. Dataset Shift in Machine Learning , 131–160doi: 10.7551/mitpress/9780262170055.003.0008 . [9] Kaewunruen, S., Freimanis, A., Ishida, M., 2018. E ff ect of extreme climate on wheel-rail interaction over rail squats, in: The 6th European Conference on Computational Mechanics (ECCM 6): ECCM6. [10] Laurent, V., Vo Van, O., Mougeot, M., Ghidaglia, J.M., 2022. Machine learning based prediction of fatigue events in railway rails, in: 15th WCEAM Proceedings. Springer, pp. 455–466. [11] Lundberg, S., Lee, S.I., 2017. A unified approach to interpreting model predictions. arXiv:1705.07874 . [12] Lundberg, S.M., Erion, G., Chen, H., DeGrave, A., Prutkin, J.M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., Lee, S.I., 2020. From local explanations to global understanding with explainable ai for trees. Nature Machine Intelligence 2, 2522–5839. [13] Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., Galstyan, A., 2021. A survey on bias and fairness in machine learning. ACM Computing Surveys (CSUR) 54, 1–35. [14] Mingers, J., 1989. An empirical comparison of pruning methods for decision tree induction. Machine learning 4, 227–243. [15] Mirkovic´, N., Brajovic´, L., Popovic´, Z., Todorovic´, G., Lazarevic´, L., Petrovic´, M., 2021. Determination of temperature stresses in cwr based on measured rail surface temperatures. Construction and Building Materials 284, 122713. [16] Muhamedsalih, Y., Hawksbee, S., Tucker, G., Stow, J., Burstow, M., 2021. Squats on the great britain rail network: Possible root causes and research recommendations. International Journal of Fatigue 149, 106267. [19] Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V., Vanderplas, J., Passos, A., Cournapeau, D., Brucher, M., Perrot, M., Duchesnay, E., 2011. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research 12, 2825–2830. [20] Sanchis, I.V., Franco, R.I., Ferna´ndez, P.M., Zuriaga, P.S., Torres, J.B.F., 2020. Risk of increasing temperature due to climate change on high-speed rail network in spain. Transportation Research Part D: Transport and Environment 82, 102312. [21] Scornet, E., 2020. Trees, forests, and impurity-based variable importance. arXiv preprint arXiv:2001.04295 . [22] Strobl, C., Boulesteix, A.L., Zeileis, A., Hothorn, T., 2007. Bias in random forest variable importance measures: Illustrations, sources and a solution. BMC bioinformatics 8, 1–21. [23] Tsuchiya, M., Yamauchi, Y., Yamashita, T., Fujiyoshi, H., 2016. Transfer forest based on covariate shift. Proceedings - 3rd IAPR Asian Conference on Pattern Recognition, ACPR 2015 , 760–764doi: 10.1109/ACPR.2015.7486605 . [24] Ver Steeg, G., Galstyan, A., 2014. Discovering structure in high-dimensional data through correlation explanation. Advances in Neural Information Processing Systems 27. [17] Nembrini, S., Ko¨nig, I.R., Wright, M.N., 2018. The revival of the gini importance? Bioinformatics 34, 3711–3718. [18] Ngargueudedjim, K., 1997. Contribution a` l’e´tude des lois d’endommagement en fatigue. Ph.D. thesis. Lyon, INSA.