Skip to Main content Skip to Navigation
Theses

Modélisations et méthodes numériques pour l’intégration d’une solution de suivi de vieillissement d’un assemblage de puissance

Abstract : This work is part of the french Research National Agency project called "CAPTIF" (meaning"Innovative embedded sensors and software for power electronics devices"). This research project deals with power electronics health monitoring solutions in order to increase electric conversion devices reliability and decrease maintenance costs.Traditionnal reliability models limitations led to a reconsideration of life prediction methodsand mean time to failure as a reliability management tool for power electronics devices.An investigation to develop an approach more representative of failure mechanisms observed in power electronics devices was conducted. Based on ageing mechanisms of power electronics assemblies, an ageing indicator was proposed and used with a modelling methodology to characterise the residual life of a device exposed to thermal loads. The functional link between this ageing indicator and data provided by embedded sensors is approache dusing a numerical design of experiment and a response surface methodology. This modelling process is illustrated by an application to a simplified power assembly.
Document type :
Theses
Complete list of metadatas

Cited literature [103 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01830693
Contributor : Abes Star :  Contact
Submitted on : Thursday, July 5, 2018 - 11:59:05 AM
Last modification on : Wednesday, April 1, 2020 - 1:15:28 AM
Long-term archiving on: : Monday, October 1, 2018 - 4:57:38 PM

File

RENAUD_ANTOINE_2018.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-01830693, version 1

Citation

Antoine Renaud. Modélisations et méthodes numériques pour l’intégration d’une solution de suivi de vieillissement d’un assemblage de puissance. Electronique. Université de Bordeaux, 2018. Français. ⟨NNT : 2018BORD0052⟩. ⟨tel-01830693⟩

Share

Metrics

Record views

790

Files downloads

328