Skip to Main content Skip to Navigation
Theses

Modèles d'âge virtuel et de risques concurrents pour la maintenance imparfaite

Yann Dijoux 1
1 SMS - Statistique et Modélisation Stochatisque
LJK - Laboratoire Jean Kuntzmann
Abstract : This work deals with stochastic modelling for the reliability of repairable systems. We consider systems subject to corrective and condition-based preventive maintenance. The main topic of this thesis is the joint modelling of the dependency between both kinds of maintenances using the competing risks framework, and the efficiency of imperfect maintenance based on the virtual age models. In this framework, we propose imperfect maintenance models for systems presenting a burn-in period. In a first part, a new usual competing risks model called alert-delay is presented. The original properties of this model allow to analyse some datasets that could not be fitted by usual models. In a second part, we propose several extensions of the usual competing risks models in a generalized competing risks framework. The assumptions proposed reflect the maintenance efficiency as well as the reactivity of the maintenance crew. Finally, we present a general modelling of the maintenance efficiency for systems presenting a burn-in period. Up to now, the classical models were only adapted for systems wearing out continuously. Two particular models are studied in details. In each part, the models proposed are applied to datasets issued from the industry.
Complete list of metadatas

Cited literature [74 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00379525
Contributor : Yann Dijoux <>
Submitted on : Tuesday, April 28, 2009 - 6:03:28 PM
Last modification on : Saturday, March 28, 2020 - 1:12:43 AM
Document(s) archivé(s) le : Thursday, June 10, 2010 - 7:05:27 PM

Identifiers

  • HAL Id : tel-00379525, version 1

Collections

Citation

Yann Dijoux. Modèles d'âge virtuel et de risques concurrents pour la maintenance imparfaite. Mathématiques [math]. Institut National Polytechnique de Grenoble - INPG, 2008. Français. ⟨tel-00379525⟩

Share

Metrics

Record views

1182

Files downloads

689