Numerical investigations of subgrid-scale mixing models for large eddy simulation of turbulent reacting flows via the filtered mass density function approach

Dinesh A Shetty, Purdue University

Abstract

Scalar mixing is an important phenomenon in many engineering application devices such as internal combustion engines, gas turbine engines, etc. The strong interaction between turbulence and non-linear chemistry makes the simulation of combustion processes highly challenging. The hybrid large eddy simulation (LES)/filtered mass density function (FMDF) method has become an important tool to analyze transient interactions between scalar and turbulence fields. The main advantage of LES/FMDF technique is that the non-linear chemistry terms appear in closed form. But on the other hand, molecular mixing terms in the formulation need to be modeled. In the current study, different mixing models are assessed by comparing the LES/FMDF results directly with experimentally measured quantities. A new mixing model has been developed and found to give better predictions. A pseudo-spectral LES/FDF code is developed to analyze the mixing models under incompressible flow formulation. This serves as an excellent platform for mixing model analysis, as the influence of boundary treatment methods is absent in this code due to the periodicity in all directions. The homogeneous, non-reacting and reacting simulations also ascertain the enhancement in LES prediction using the newly developed mixing model. A chemistry module capable of performing In-situ Adaptive Tabulation (ISAT), Adaptive Chemistry Reduction (ACR) and skeletal mechanism generation is developed to assist in the chemistry evaluation part. Accurate stiff solvers have been integrated with this module to solve the stiff equations that arise during the reacting flow simulations.

Degree

Ph.D.

Advisors

Frankel, Purdue University.

Subject Area

Mechanical engineering

COinS