Créé le : 09/09/2010 - Mis à jour le : 15/11/2012
In reliability analysis, the failure criterion is defined as a limit state function which depends on random inputs and response quantities. Using stochastic finite elements, the response quantity can be written as series expansion which allows an approximation of the limit state function. For computational purpose, the series must be truncated in order to retain only a finite number of terms. In the context of industrial applications, no information is available concerning the smoothness of the model. This way, the accuracy of the truncated polynomial chaos expansions cannot be directly computed. To tackle this problem, we propose a new approach coupling polynomial chaos expansions and confidence intervals. This approach, which we call RPCM (Resampling Polynomial Chaos Method), aims at exploring an existing database to find the best chaos order which allows to reach the reliability index. The proposed approach...