Modèles de Cox et réseaux de neurones pour l'évaluation de la fiabilité mécanique

Abstract : In the data reliability treatment, engineer have to choose an adequate model to do many inferences. Often exponential law is use, but this law is not good for mechanical components and there not use externs conditions, these are an major defect. We purpose to use two way, first way parametric model, and specially Cox's model and second way neural net work method. Both method are use through an application and compare.
Document type :
Conference papers
Complete list of metadatas

https://hal-ineris.archives-ouvertes.fr/ineris-00970237
Contributor : Gestionnaire Civs <>
Submitted on : Wednesday, April 2, 2014 - 3:45:01 PM
Last modification on : Thursday, February 22, 2018 - 1:06:03 PM

Identifiers

  • HAL Id : ineris-00970237, version 1
  • INERIS : EN-2007-141

Collections

Citation

Patrick Lyonnet, P. Maisonneuve, Brice Lanternier, Rosario Toscano. Modèles de Cox et réseaux de neurones pour l'évaluation de la fiabilité mécanique. 8. Colloque national en calcul des structures, May 2007, Giens, France. ⟨ineris-00970237⟩

Share

Metrics

Record views

114