Nonparametric stochastic modeling of structural dynamic systems with uncertain boundary conditions

M. P. Mignolet (1) , Christian Soize (2)

1 Fac Mech & Aerosp Engn
2 LaM - Laboratoire de Mécanique

A notorious source of uncertainty are physical clamps which are often used and traditionally modeled by clamped boundary conditions. It is however well recognized that the mathematical clamp boundary condition is never achieved and that significant variations do exist in the boundary conditions physically realized. This issue belongs in fact to model uncertainties and may be directly addressed by the nonparametric stochastic modeling approach. A different, more detailed approach is sought here and carried out on the system with an "incomplete" boundary condition in which motion at the boundary is restrained by springs. The application of the nonparametric modeling approach to such systems will be performed and discussed. Comparisons with the direct application of the nonparametric stochastic modeling approach will also be presented. Finally, these discussions and results will be exemplified on a plate clamped on all sides.

Mots clés: uncertainty quantification; computational mechanics; robust updating

Domaines: Mécanique [physics.med-ph]; Probabilités [math.PR]

M. P. Mignolet, Christian Soize. Nonparametric stochastic modeling of structural dynamic systems with uncertain boundary conditions. USNCCM IX 2007, Ninth U. S. National Congress on Computational Mechanics, Jul 2007, San Francisco, United States. pp.2. ⟨hal-00698959⟩