BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Bayesian Uncertainty Quantification and Propagation in Structural 
 Dynamics Simulations using Vibration Measurements
DTSTART:20150327T110000
DTEND:20150327T120000
DTSTAMP:20260407T095730Z
UID:3935a911b2a07d6dd824441a831c7304c0c28b1024637f570d7710b3
CATEGORIES:Conferences - Seminars
DESCRIPTION:Costas Papadimitriou\, Professor\, Department of Mechanical E
 ngineering\, University of Thessaly\, Volos\, Greece\nBayesian inference 
 is used for quantifying and calibrating uncertainty models in structural d
 ynamics based on vibration measurements\, as well as propagating these unc
 ertainties in simulations for updating robust predictions of system perfor
 mance\, reliability and safety. The Bayesian tools for identifying system 
 and uncertainty models as well as performing robust prediction analyses ar
 e Laplace methods of asymptotic approximation and sampling algorithms. The
 se tools involve solving optimization problems\, generating samples for tr
 acing and then populating the important uncertainty region in the paramete
 r space\, as well as evaluating integrals over high-dimensional spaces of 
 the uncertain model parameters. They require a moderate to very large numb
 er of system re-analyses to be performed over the space of uncertain param
 eters. Consequently\, the computational demands depend highly on the numbe
 r of system analyses and the time required for performing a system analysi
 s.\nA computational framework for Bayesian uncertainty quantification and 
 propagation for complex models in structural dynamics will be presented. T
 he model complexity is due to the very high number (hundreds of thousands 
 or millions) of degrees of freedom arising in developing high-fidelity str
 uctural models\, the localized nonlinear actions activated during system o
 peration\, and the applied stochastic loads. High performance computing te
 chniques are integrated with Bayesian techniques to efficiently handle suc
 h complexities. Fast and accurate component mode synthesis (CMS) technique
 s are proposed\, consistent with the finite element model parameterization
 \, to achieve drastic reductions in computational effort. Surrogate models
  are also used within multi-chain sampling algorithms with annealing prope
 rties to substantially speed-up computations\, avoiding full system re-ana
 lyses. Significant computational savings are achieved for stochastic simul
 ation algorithms by adopting parallel computing algorithms to efficiently 
 distribute the computations in available GPUs and multi-core CPUs. Importa
 nt issues related to the computational efficiency of the asymptotic approx
 imations versus the stochastic simulation algorithms for serial or paralle
 l computing environments are discussed. The proposed framework is demonstr
 ated using applications in civil infrastructure and vehicle dynamics.
LOCATION:GC G1 515 http://plan.epfl.ch/?lang=fr&zoom=19&recenter_y=5864267
 .48334&recenter_x=730958.60537&layerNodes=fonds\,batiments\,labels\,inform
 ation\,parkings_publics\,arrets_metro&floor=1&q=GC_G1%20515
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR