Modeling and identification of non Gaussian multivariate random fields and application to the excitation of trains by the track irregularities. - École des Ponts ParisTech Accéder directement au contenu
Communication Dans Un Congrès Année : 2013

Modeling and identification of non Gaussian multivariate random fields and application to the excitation of trains by the track irregularities.

Résumé

This presentation deals with an innovative approach to analyze complex and nonlinear systems, which are excited by non-Gaussian and non-stationary random fields, by solving of a statistical inverse problem with experimental measurements. The methodology proposed is applied to the case of a railway system: a train is a nonlinear system with many degrees-of-freedom, which is excited by the track geometry and irregularities. These irregularities are of four types (horizontal and vertical alignment irregularities on the first hand, cant and gauge irregularities on the second hand), and vary from one track to another one, from one country to another one. As the track vehicle system is very non-linear, the characterization of the train dynamics cannot be achieved from the analysis of the train response on a single track portion but has to be made on the whole set of track conditions that the train can be confronted to during its lifecycle. In reply to these expectations, the track geometry of the French railway network has been continuously measured since 2007. Based on these measurements, which can be seen as experimental realizations of the track geometry random field, we develop a two steps methodology to analyze the influence of the track geometry variability on the train dynamics. In a first step, a stochastic modeling of the track geometry is proposed. Two decompositions are therefore used to identify the statistical characteristics of this random field. At first, using the Karhunen-Loève expansion, the considered random field is approximated by its truncated projection on a particularly well adapted orthogonal basis. Then, the random vector, which gathers the projection coefficients of the random field on this spatial basis, is characterized using a polynomial chaos expansion approach.The non-Gaussian non-stationary vector-valued random field is identified using the experimental measurements following the methodology presented in and consequently, constitutes a realistic track geometry stochastic modeling. Secondly, the track geometry variability is propagated to the train dynamics by solving a nonlinear stochastic dynamical problem. The results obtained are presented and analysed.
Fichier non déposé

Dates et versions

hal-00852774 , version 1 (21-08-2013)

Identifiants

  • HAL Id : hal-00852774 , version 1

Citer

Guillaume Perrin, Christian Soize, Denis Duhamel, Christine Fünfschilling. Modeling and identification of non Gaussian multivariate random fields and application to the excitation of trains by the track irregularities.. 1ères Journées Jeunes Chercheurs en Vibrations, Apr 2013, France. ⟨hal-00852774⟩
192 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More