Identification of nonlinear vibrations in railway vehicles including considerations of track defects

In this work first results from an ongoing work on parameter identification for a TGV model are presented. The first step of the study is the modeling of the vehicle with a multi-body system code and the comparison of the simulation results with measurements in the time domain. The suspension characteristics are difficult to determine since they can differ from one train to another and are evolving during their life time. Besides, characterization tests on single suspension elements are expensive. The parameter identification based on in- line tests is therefore of particular interest. An adjoint state gradient method is proposed in order to minimize a misfit function describing the difference between measured and modeled response data. The objective of further work is to apply it on a simplified multi-body model of the bogie which is presented in the third part of the proceeding.