Model form uncertainty quantification in turbulent combustion simulations

Uncertainties in turbulent combustion simulations come from three distinct sources: uncertainties in operating parameters such as boundary conditions, uncertainties in model parameters such as chemical kinetic rates, and structural uncertainties associated with the form of the various component models. While the quantification of parametric uncertainty is relatively straightforward conceptually, the quantification of model form uncertainty requires the translation of physical assumptions into mathematical statements. In this work, two methods are proposed for quantifying such model form uncertainty directly from physics: hierarchical models and peer models. Hierarchical models use a higher-fidelity model to estimate the uncertainty in a lower-fidelity model, and peer models use nominally equivalent models with differing assumptions to derive an uncertainty estimate. These notions will be utilized to quantify model form error in a nonpremixed combustion model and in a subfilter dissipation rate model, respectively.