A stochastic approach for the evaluation of the reliability of wood furniture in an industrial context: managing virtual standardization tests

¹ UniversitéParis-Est, Laboratoire Modelisation et Simulation Multi Echelle, MSME UMR 8208 CNRS, 5 bd Descartes, 77454 Marne-la-Vallée, France
² Institut technologique Forêt Cellulose Bois-Construction Ameublement, FCBA, 10 avenue de Saint-Mandé, 75012 Paris, France

^a Corresponding author: luc.chevalier@univ-paris-est.fr

Received: 9 April 2015
Accepted: 29 September 2015

Abstract

In this paper, a procedure is developed in order to manage virtual tests using numerical simulation in order to reduce time and to optimize design before the real validation tests on wood products. This procedure in 3 steps can be used in an industrial context: (i) characterization of the material properties, (ii) construction of probabilistic density function for each parameter, (iii) Monte Carlo simulation of the structural behavior of the wood products furniture. For this purpose, experimental tests associated with digital image correlation technique were managed using 3 point bending tests easy to develop in an industrial environment. A probabilistic density function was constructed for each mechanical parameter based on the test results using the principle of maximum entropy. The simulation is managed by a FEA developed in the Matlab environment and allows a Monte-Carlo simulation in a reasonable CPU time. We can then evaluate the effect of the input data dispersions on the overall reaction of the structure. A real test is performed on a prototype product in order to confirm the simulation results and to validate the approach.