Large eddy simulation of reacting flow physics and combustion

The multiphysics, multiscale nature of turbulent chemically reacting flows makes these systems one of the most challenging to understand and control. A multitude of strongly coupled fluid dynamic, thermodynamic, transport, chemical, multiphase, and heat transfer processes are intrinsically coupled and must be considered simultaneously in complex domains to satisfy the governing conservation equations. The problem is compounded by the broad range of time and length scales over which interactions occur due to turbulence and differences in chemical reaction rates. The nonlinear nature of the system significantly limits the number of simplifying assumptions that can be made. Conversely, some form of modeling is always required and significant sets of assumptions must be made to derive multiscale closures that are both accurate and affordable. This combination of challenges significantly complicates the process of scientific discovery, model development, and model validation. Model development is challenging because it is nearly impossible to decouple various processes without introducing potentially significant sources of error. Model validation is problematic because detailed data over relevant parameter spaces are limited and comparisons using available data are difficult to interpret. In many cases, deviations between measured and modeled results cannot be directly linked to a particular modeling approach and the related assumptions because many other factors (e.g., boundary conditions, grid resolution, numerical errors, etc.) can have dominating effects on the results. To mitigate these challenges, there is a significant need to carefully assess the physical characteristics of the governing system and the related first-principles constitutive relations. This assessment must be done within the formalism of a generalized multiscale closure that exposes the effects of turbulence on molecular processes, within the fully coupled governing system, over 232clearly defined ranges of scales. The analysis must be performed using numerical methods that are well suited for the treatment of broadband multiscale processes and do not introduce additional extraneous errors that compete with the submodels and further complicate the analysis. The large eddy simulation (LES) approach described here is focused on addressing these needs.