Bayesian Uncertainty Quantification and Propagation in Nonlinear Structural Dynamics

Nonlinear modelling and parametric identification of an experimental vehicle model , are employed in this paper. The composite structure of the vehicle model is split into a frame substructure and to four support substructures. The frame substructure possesses linear properties determined through application of a finite element analysis and designed to exhibit a relatively large modal density. A method for modal identification and structural model updating are employed in order to develop a high fidelity finite element model of the vehicle substructure. On the other hand, the four support substructures including the wheels and suspensions components, possesses strongly nonlinear characteristics, accounting mainly for absorber damping nonlinearities. Then, a Bayesian uncertainty quantification and propagation framework is adopted in order to estimate the optimal values of the four support substructures model parameters. Uncertainty models of the nonlinear wheel and suspension components are identified using the experimentally obtained response spectra for each of the components tested separately. These uncertainties, integrated with uncertainties in the body of the experimental vehicle, are propagated to estimate the uncertainties of output quantities of interest for the combined wheelsuspension-frame system. The computational challenges are outlined and the effectiveness of the Bayesian UQ&P framework on the specific example structure is demonstrated. Finally the experimental results w ere compared to those from the numerical model for verification of the numerical procedure and for the improvement of the numerical modelling of the vehicle substructuring components .