A Higher-Order Flamelet Model for Turbulent Combustion Simulations.

Abstract

Current projection of energy consumption trends has shown that combustion of fossil fuel will continue to play an important role in industrial thermal processes, power generation, and transportation for a substantial period. In order for these sectors to sustain under the finite fossil fuel reserves, improvements in existing devices and development of novel concepts that emphasize on energy efficiency are necessary. Numerical simulations can be used to address this need, in particular by complementing experiments with extensive and quick parametric studies. However, this is only viable if numerical predictions of the combustion processes are accurate, which requires adequate modeling of the multi-physics phenomena in turbulent reacting flows.

In this work, the flamelet-type combustion model, one of the most widely used approaches for turbulent reacting flow simulations, is thoroughly analyzed in terms of the validity of its underlying assumptions and limitations in its description of different combustion regimes. Diagnostic tools that account for the flamelet formulation are developed and applied to two different direct numerical simulation (DNS) results. These analyses show that the omission of the higher-order and unsteady flamelet effects by most conventional flamelet models is not valid in realistic configurations that are characterized by complex vortical structures, flame extinction and reignition, and turbulence-chemistry interactions.

Following these findings, a higher-order flamelet model that describes the conventionally omitted flamelet effects is developed for large-eddy simulations (LES) applications. This model is based on the physical interpretation of flamelets as quasi one-dimensional structures in the turbulent flow, and the consideration of the effects that the spatial-filtering in LES methodology has on these structures. The model is applied in LES of a turbulent counterflow diffusion flame configuration, demonstrating improved agreement with the reference DNS solutions of the same case than the steady flamelet/progress variable (FPV) and laminar approximation models.