SC²S Colloquium - November 11, 2015
|Date:||November 11, 2015|
|Time:||3:00 pm, s.t.|
Ionut Farcas: High-Dimensional Uncertainty Quantification of Fluid-Structure Interaction
This work discusses uncertainty quantification problems in fluid-structure interaction, using the pseudo-spectral approach. The focus is on multidimensional stochastic settings, but one dimensional scenarios are also simulated, for obtaining a comprehensive understanding of the effects of uncertainty. Additionally, the Monte Carlo algorithm is employed for validating the results obtained via the pseudo-spectral approach. The testing is performed using two fluid-structure interaction scenarios. The first one is a two dimensional wall-mounted flap, whereas the second scenario represents an established fluid-structure interaction benchmark. Addressing the complexity posed by high dimensional uncertainty quantification is a challenging task, which is resolved using an approach comprising sparse grids interpolation and numerical quadrature. In post-processing, besides estimating the statistics of the selected quantities of interest, the estimation of the probability density function together with the sensitivity analysis based on Sobol's indices are also performed at no additional computational cost, using the already computed coefficients of the spectral expansion approximation.