A novel hybrid surrogate model and its application on a mechanical system subjected to friction-induced vibration

A new hybrid meta-modeling approach is proposed and developed for the propagating and quantifying of uncertainties in friction-induced instabilities. According to the available knowledge about design parameters, the associated uncertainty can be described by using different models. Hence, hybrid met...

Full description

Saved in:
Bibliographic Details
Published in:Journal of sound and vibration 2018-11, Vol.434, p.456-474
Main Authors: Denimal, E., Nechak, L., Sinou, J-J., Nacivet, S.
Format: Article
Language:English
Subjects:
Complex eigenvalue analysis
Computer simulation
Design parameters
Dispersion
Engineering Sciences
Friction
Generalized polynomial chaos
Kriging
Kriging interpolation
Mechanics
Meta-modeling
Mode coupling
Monte Carlo simulation
Parameter uncertainty
Solid mechanics
Stability
Structural mechanics
Uncertainty
Vibration
Vibrations
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A new hybrid meta-modeling approach is proposed and developed for the propagating and quantifying of uncertainties in friction-induced instabilities. According to the available knowledge about design parameters, the associated uncertainty can be described by using different models. Hence, hybrid meta-models prove to be necessary to simultaneously treat different uncertainty models. So, this study presents a novel hybrid meta-model issued from the associating of the kriging formalism together with the generalized polynomial chaos for the prediction of friction-induced instabilities submitted to interval and probabilistic uncertainties. Its assessing through the considering of a friction system, reveals suitable accuracy about the estimating of the dispersion of the occurrences of instabilities. Moreover, it offers a promising alternative to the prohibitive Monte-Carlo/scanning based methods that are usually used for the same task.
ISSN:0022-460X
1095-8568
DOI:10.1016/j.jsv.2017.08.005