Stochastic dynamic simulation of railway vehicles collision using data-driven modelling approach

Abstract

Using stochastic dynamic simulation for railway vehicle collision still faces many challenges, such as high modelling complexity and time-consuming. To address the challenges, we introduce a novel data-driven stochastic process modelling (DSPM) approach into dynamic simulation of the railway vehicle collision. This DSPM approach consists of two steps: (i) process description, four kinds of kernels are used to describe the uncertainty inherent in collision processes; (ii) solving, stochastic variational inferences and mini-batch algorithms can then be used to accelerate computations of stochastic processes. By applying DSPM, Gaussian process regression (GPR) and finite element (FE) methods to two collision scenarios (i.e. lead car colliding with a rigid wall, and the lead car colliding with another lead car), we are able to achieve a comprehensive analysis. The comparison between the DSPM approach and the FE method revealed that the DSPM approach is capable of calculating the corresponding confidence interval, simultaneously improving the overall computational efficiency. Comparing the DSPM approach with the GPR method indicates that the DSPM approach has the ability to accurately describe the dynamic response under unknown conditions. Overall, this research demonstrates the feasibility and usability of the proposed DSPM approach for stochastic dynamics simulation of the railway vehicle collision.

Keywords

Dynamic simulation / Railway vehicle collision / Stochastic process / Data-driven stochastic process modelling

References

[1.] Zhu T Xiao S Lei C . Rail vehicle crashworthiness based on collision energy management: an overview. Int J Rail Transp, 2021 9 2 101-131

[2.] Zhu T Xiao SN Hu GZ . Crashworthiness analysis of the structure of metro vehicles constructed from typical materials and the lumped parameter model of frontal impact. Transport, 2019 34 1 75-88

[3.] Li R Xu P Peng Y . Scaled tests and numerical simulations of rail vehicle collisions for various train sets. Proc Inst Mech Eng Part F J Rail Rapid Transit, 2016 230 6 1590-1600

[4.] Lu S Xu P Yan K . A force/stiffness equivalence method for the scaled modelling of a high-speed train head car. Thin-Walled Struct, 2019 137 129-142

[5.] Cho HJ Koo JS. A numerical study of the derailment caused by collision of a rail vehicle using a virtual testing model. Veh Syst Dyn, 2012 50 1 79-108

[6.] Scholes A Lewis JH. Development of crashworthiness for railway vehicle structures. Proc Inst Mech Eng Part F J Rail Rapid Transit, 1993 207 1 1-16

[7.] Montgomery DC Runger GC Hubele NF. Engineering statistics, 2009 New York Wiley

[8.] Tang Z Zhu Y Nie Y . Data-driven train set crash dynamics simulation. Veh Syst Dyn, 2017 55 2 149-167

[9.] Dong S Tang Z Yang X . Nonlinear spring-mass-damper modeling and parameter estimation of train frontal crash using CLGAN model. Shock Vib, 2020 2020 9536915

[10.] Nie Y Tang Z Liu F . A data-driven dynamics simulation framework for railway vehicles. Veh Syst Dyn, 2018 56 3 406-427

[11.] Müller M Botsch M Böhmländer D . Machine learning based prediction of crash severity distributions for mitigation strategies. J Adv Inf Technol, 2018 9 1 15-24

[12.] Li YR Zhu T Xiao SN . Application of the collision mathematical model based on a BP neural network in railway vehicles. Proc Inst Mech Eng Part F J Rail Rapid Transit, 2021 235 6 713-725

[13.] Luo R Shi H Teng W . Prediction of wheel profile wear and vehicle dynamics evolution considering stochastic parameters for high-speed train. Wear, 2017 392 126-138

[14.] Lu Y Yang S Li S . Numerical and experimental investigation on stochastic dynamics load of a heavy duty vehicle. Appl Math Model, 2010 34 10 2698-2710

[15.] Souffran G Miègeville L Guérin P. Simulation of real-world vehicle missions using a stochastic Markov model for optimal powertrain sizing. IEEE Trans Veh Technol, 2012 61 8 3454-3465

[16.] Xu L Zhai W. A new model for temporal–spatial stochastic analysis of vehicle–track coupled systems. Veh Syst Dyn, 2017 55 3 427-448

[17.] Hao P Boriboonsomsin K Wu G . Modal activity-based stochastic model for estimating vehicle trajectories from sparse mobile sensor data. IEEE Trans Intell Transp Syst, 2017 18 3 701-711

[18.] Hoffman MD Blei DM Wang C . Stochastic variational inference. J Mach Learn Res, 2013 14 5 1303-1347

[19.] Urtasun R, Darrell T (2008) Sparse probabilistic regression for activity-independent human pose inference. In: 2008 IEEE conference on computer vision and pattern recognition. Anchorage, AK, USA. IEEE, pp 1–8

[20.] Hensman J, Rattray M, Lawrence N D (2012) Fast variational inference in the conjugate exponential family. arXiv preprint arXiv:1206.5162

[21.] Hensman J, Fusi N, Lawrence ND (2013) Gaussian processes for big data. arXiv preprint arXiv:1309.6835

[22.] Raissi M Babaee H Karniadakis GE. Parametric Gaussian process regression for big data. Comput Mech, 2019 64 2 409-416

[23.] Shokravi H Shokravi H Bakhary N . A comparative study of the data-driven stochastic subspace methods for health monitoring of structures: a bridge case study. Appl Sci, 2020 10 9 3132

[24.] Prudencio EE Bauman PT Williams SV . A dynamic data driven application system for real-time monitoring of stochastic damage. Procedia Comput Sci, 2013 18 2056-2065

[25.] Jiang B Fei Y. Vehicle speed prediction by two-level data driven models in vehicular networks. IEEE Trans Intell Transp Syst, 2017 18 7 1793-1801

[26.] Maraun D Huth R Gutiérrez JM . The VALUE perfect predictor experiment: evaluation of temporal variability. Int J Climatol, 2019 39 9 3786-3818

[27.] Fortunato AB Bertin X Oliveira A. Space and time variability of uncertainty in morphodynamic simulations. Coast Eng, 2009 56 8 886-894

[28.] Haigermoser A Luber B Rauh J . Road and track irregularities: measurement, assessment and simulation. Veh Syst Dyn, 2015 53 7 878-957

[29.] Oprea RA. A constrained motion perspective of railway vehicles collision. Multibody SysDyn, 2013 30 1 101-116

[30.] Rasmussen CE (2004) Gaussian processes in machine learning. In: Advanced lectures on machine learning. Springer, Berlin, pp 63–71

[31.] Shimodaira H. Improving predictive inference under covariate shift by weighting the log-likelihood function. J Stat Plann Inference, 2000 90 2 227-244

[32.] Zhou Q Jiang P Shao X . A variable fidelity information fusion method based on radial basis function. Adv Eng Inf, 2017 32 26-39

[33.] Oloyede I. Bayesian classification of high dimensional data with gaussian process using different kernels. Anale Seria Informatică, 2018 16 1 164-167

[34.] Longman FS, Mihaylova L, Yang L et al. (2019) Multi-band image fusion using Gaussian process regression with sparse rational quadratic kernel. In: 22th international conference on information fusion (FUSION). Ottawa, ON, Canada. IEEE, pp 1–8

[35.] Hida T Kuo HH Potthoff J . White noise: an infinite dimensional calculus, 2013 New York Springer

[36.] Sonnenburg S Rätsch G Henschel S . The SHOGUN machine learning toolbox. J Mach Learn Res, 2010 11 1799-1802

[37.] Fushiki T. Estimation of prediction error by using K-fold cross-validation. Stat Comput, 2011 21 2 137-146

[38.] Alpaydin E. Introduction to machine learning, 2020 London MIT Press

[39.] Yang RJ Wang N Tho CH . Metamodeling development for vehicle frontal impact simulation. J Mech Des, 2005 127 5 1014-1020

[40.] Chen Q, Song X, Yamada H et al. (2016) Learning deep representation from big and heterogeneous data for traffic accident inference. In: Proceedings of the thirtieth AAAI conference on artificial intelligence. Phoenix, Arizona, USA, February 12–17, 2016

[41.] Besbeas P Morgan BJT. Goodness-of-fit of integrated population models using calibrated simulation. Methods Ecol Evol, 2014 5 12 1373-1382

[42.] Gao G Wang S. Crashworthiness of passenger rail vehicles: a review. Int J Crashworthiness, 2019 24 6 664-676

[43.] Li SY Zheng ZJ. Energy-absorbing structure design and crashworthiness analysis of high-speed trains. Explos Shock Waves, 2015 35 2 164-170

Funding

National Key Scientific Instrument and Equipment Development Projects of China(2019YFB1405401); National Natural Science Foundation of China(5217120056)