201517th International Congress of Metrology
|Number of page(s)||6|
|Section||Maitrise de l’incertitude / Uncertainty management|
|Published online||21 September 2015|
L’approche Monte Carlo à plusieurs niveaux pour évaluer le processus au plus proche de la physique
1 LE2I UMR6306, CNRS, Arts et Métiers, Univ. Bourgogne Franche-Comté, F-21000 Dijon, France
2 Laboratoire MSMP, Arts et Métiers Paris Tech - 8 Boulevard Louis XIV, 59046 LILLE Cedex
3 Centre Technique des Industries Mécaniques (CETIM) – 52, avenue Félix-Louat – BP 80067 – 60304 Senlis Cedex2
a Email de correspondance de l’auteur : firstname.lastname@example.org
Monte Carlo method have been introduced in metrology in the 1990’s and integrated in the GUM (Supplement 1) in 2008. This method is more and more used. Typically the users of this method realize a complete simulation in one step, like the GUM, one step for one model. This is unfortunate, the simulation loses its physical sense. The study aim is to present a multi-level Monte Carlo approach which allows being near of the reality of the measurement process. Two applications are developed: evaluation of the uncertainties on CMM and on AACMM. This principle has been developed with CETIM for COFRAC accreditation on CMM for gear measurands. The simulation is divided into two principle stages, namely the first is the comprehensive evaluation of possible changes in the geometry of the CMM and the second step, directly related to the measure of the piece, is the evaluation of the analyzed measurand. For AACMM, same principle is realized but the first level is divided into three sub-levels. The division into several levels has many advantages. Indeed, it makes it easier to understand the key sources of uncertainty and thus optimize processes.
© Owned by the authors, published by EDP Sciences, 2015
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.