Uncertainty quantification (UQ)

Abstract

Uncertainty Quantification (UQ) is an emerging field in computational engineering that can provide certificates of fidelity in a simulation beyond the standard numerical error, and it includes uncertainty in boundary conditions, constitutive laws, materials properties and geometries. UQ is particularly impornat at microscales where geometric roughness and material properties cannot be readily quantified experimentally. Here we present a general framework for UQ based on the generalized polynomial chaos approach and various extensions that do not require modification of existing codes and are particularly effective in Microsystems with many uncertain parameters (e.g. high dimensionality).