|
Abstract:
|
From both mathematical and statistical perspectives, the fundamental goal of Uncertainty Quantification (UQ) is to ascertain uncertainties inherent to parameters, initial and boundary conditions, experimental data, and models themselves to make predictions with improved and quantified accuracy. The central role of uncertainty quantification for the Mars Sample Return Mission is compounded by the fact that models are critical to complement often sparse experimental data for the extreme conditions encountered at various points in the mission. This includes high-fidelity simulation models, which incorporate the full physics, and highly efficient surrogate models, which provide the computational efficiency required for design and uncertainty analysis. This presentation will detail the use of Bayesian inference to quantify uncertainties inherent to model parameters and experimental data. We will subsequently discuss the construction of prediction intervals for responses and detail how this can guide the quantification of model discrepancies and be employed for rigorous model validation.
|
