Conditional source-term estimation with laminar flamelet decomposition in large eddy simulation of a turbulent nonpremixed flame

Abstract

Conditional source-term estimation (CSE) is a method to close the mean chemical reaction source term based on the conditional moment closure hypothesis. It has been shown previously to be successful in a priori tests against direct numerical simulation data and in the large eddy simulation (LES) of a nonpremixed flame using reduced chemistry. Laminar flamelet decomposition (LFD) is a method to incorporate more complex chemistry into CSE by using laminar flamelets as basis functions and inverting an integral equation for a basis function coefficient vector which describes the linear combination of flamelets that best approximates the conditional average of certain scalar fields within an ensemble of points in a flow field. This coefficient vector is used to obtain the conditional average of the chemical source term for this ensemble of points which can then be transformed into the unconditional average chemical source term in the transport equations of reactive scalars in the flow. This study focuses on the application of CSE with LFD in LES of the Sandia D-flame. The simulation results show that LFD is able to predict temperature and major species well with both steady and unsteady flamelet libraries. However, in order to predict NO well, it is necessary to use a mixed library that includes both steady and unsteady flamelets. The computational cost of this method is low because very few transport equations need to be solved (specifically, equations for the filtered continuity, momentum, mixture fraction and its variance, temperature and the mass fraction of CO are solved) while other species mass fractions can be obtained directly from the flamelet library.