Vous êtes ici : Accueil » Kiosque » Annonce

Identification

Identifiant: 
Mot de passe : 

Mot de passe oublié ?
Détails d'identification oubliés ?

Annonce

12 octobre 2017

Data integration for the management of uncertainty in a simulation process


Catégorie : Post-doctorant


Subject: Data integration for the management of uncertainty in a simulation process

Location: Institut CEA LIST, CEA Saclay – Digiteo, 91191 Gif-sur-Yvette Cedex, France

Duration: 12 months

Contacts : guillaume.damblin@cea.fr , christophe.reboud@cea.fr

 

Summary

In the field of NonDestructive Testing (NDT), Probability of detection (POD) and false alarm rate are key quantities used to quantify the performance of a particular inspection setup. Simulation can be used to estimate such quantities, through a propagation of inputs uncertainty in the physical model. The CIVA platform developed at CEA LIST is recognized as a software of reference in the community for the simulation of NDT techniques. Firstly, the project aims at enhancing the description of inputs uncertainty, which is today somewhat arbitrarily described based on experts' judgement. Secondly, it aims at designing a model calibration strategy able to correct discrepancies between theory and practice, coming from phenomena that are not taken into account by the model. Among those effects, one can cite perturbations due to the environment or human factors, for instance. To reach these two goals, the method proposed consists in integrating experimental data to the simulation process. The work will be conducted in close collaboration with the LGLS laboratory of CEA's Directon de l'Energie Nucléaire (DEN), specialized in statistical modelling and developing the URANIE platform dedicated to sensitivity analysis and uncertainty quantification.

References

F. Bachoc, G. Bois, J. Garnier and J.M. Martinez, "Calibration and improved prediction of computer models by universal Kriging", Nuclear Science and Engineering 176(1) (2014) 81-97.

Abdessalem, A. Jenson, F. Calmon, P., Improving the Reliability of POD Curves in NDI Methods Using a Bayesian Inversion Approach for Uncertainty Quantification, AIP Conference Proceedings; 2016, Vol. 1706 Issue 1, p1-7.

Damblin, G., Keller, M., Barbillon, P., Pasanisi, A., and Parent, É. (2016) Bayesian Model Selection for the Validation of Computer Codes. Qual. Reliab. Engng. Int., 32: 2043–2054. doi: 10.1002/qre.2036.

Calmon P., Jenson F., Reboud C., Simulated Probability of Detection maps in case of non- monotonic EC signal response, in 41st Annual Review of Progress in QNDE, VAIP Conf. Proc. 1650, 1993, (2014)

Dominguez N., Reboud C., Dubois A. and Jenson F., “A new approach of confidence in POD determination using simulation”, in Review of Progress in QNDE, 32, 2012.

www.extende.com

https://sourceforge.net/projects/uranie/

 

Dans cette rubrique

(c) GdR 720 ISIS - CNRS - 2011-2015.