Skip to Main content Skip to Navigation
Theses

Une stratégie d'identification robuste pour la localisation et la rupture

Abstract : The objective of this work is to propose a robust identification strategy, which allows to identify a constitutive relation for material rupture. The challenge is to be able to identify the parameters that control the rupture in a context of highly corrupted experimental data. After a first work where the strategy has been developed in an elastic framework, this thesis concentrates on the extension of the identification method to the nonlinear cases, first the viscoplasticity, then the delay damage model.
The main difficulties in these cases are the non-linearity and the unstable character of the minimization problem under nonlinear constraints which is derived from the
formulation. An extension of the LATIN method to ill-posed problems is proposed and developed in order to allow an iterative solving of this type of optimization problems. The linearized optimal control problems are solved by a robust algorithm based on optimal control techniques, mainly using the Riccati equation.
The difficulties being solved and the formulation implemented in the one-dimensional case until now, the proposed identification strategy appears to be very robust with respect to the perturbed measurements even in the case of localisation and rupture.
Complete list of metadatas

Cited literature [124 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00140895
Contributor : Ens Cachan Bibliothèque <>
Submitted on : Tuesday, April 10, 2007 - 4:12:07 PM
Last modification on : Monday, February 15, 2021 - 10:49:32 AM
Long-term archiving on: : Wednesday, April 7, 2010 - 3:15:47 AM

Identifiers

  • HAL Id : tel-00140895, version 1

Citation

Hong-Minh Nguyen. Une stratégie d'identification robuste pour la localisation et la rupture. Mécanique [physics.med-ph]. École normale supérieure de Cachan - ENS Cachan, 2006. Français. ⟨tel-00140895⟩

Share

Metrics

Record views

461

Files downloads

258