Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
Quantifying spatial uncertainty and inferring the stochastic wave attenuation in periodic frame structures
Submission Author:
Luiz Henrique Marra da Silva Ribeiro , SP
Co-Authors:
Luiz Henrique Marra da Silva Ribeiro, Danilo Beli, Adriano Todorovic Fabro, José Arruda
Presenter: Luiz Henrique Marra da Silva Ribeiro
doi://10.26678/ABCM.COBEM2021.COB2021-1621
Abstract
Variability is an inherent aspect of manufacturing processes such as machining and mechanical assembly. Deformations during the useful life of a structure are also a cause of variability. Statistical methods can be used to quantify variability in geometry or material properties. The variability can be in the form of random variables or stochastic processes, and the results obtained with the stochastic models will also be stochastic. The objective of the present study is to present how to use some statistical tools to simulate structural variability and to infer on the structural responses and dispersion diagrams. The simplest case, where the property of the entire structure varies according to a random variable, is presented. Then, a combination of Bayesian inference with Expansion Optimal Linear Estimation and discrete Karhunen–Loève expansion is used to model one-dimensional spatial variability. The methodology is precise and versatile to model according to spatially distributed observations. It is shown how the proposed methodology can be expanded for simulating two- and three-dimensional spatial variability. The Bayes’ factor is used to infer on the stochastic results.
Keywords
uncertainty quantification, Bayesian inference, Spatial variability, Three-dimensional frame, Phononic Crystals
