APPLICATION OF MONTE CARLO FILTER FOR COMPUTER VISION-BASED BAYESIAN UPDATING OF FINITE ELEMENT MODEL

In this paper we describe Bayesian inference-based approach to the solution of parametric identification problem in the context of updating of a finite element model of a structure. The proposed inverse solution is based on Monte Carlo filter and on the comparison of structure displacements extracted using digital image correlation method during a quasi-static loading and the corresponding displacements predicted by finite element method program. Our approach is applied to the problem of material model parameter identification of an aluminum laboratory-scale frame. The results are also verified by comparing the Monte Carlo filter-based solution with the analytical solution obtained using Kalman filter.