Assessment of the progress variable variance modelling on large-eddy simulation of turbulent premixed flames using flamelet-generated manifold model

The objective of this paper is two-fold: a comprehensive literature review of the combustion model based on the laminar flamelet assumption; and implementation, application and sensitivity analysis of one of these flamelet models for simulation of turbulent premixed flames. Large-eddy simulation is one of the most reliable approaches in turbulence modelling. Since the computational cost of this approach is substantially more significant than the Reynolds-averaged Navier-Stokes models, the most economical and thus widely-used combustion models in the context of large-eddy simulation are models based on the flamelet assumption. Nevertheless, flamelet models have known shortcomings presented and discussed in detail in this work. The Flamelet-Generated Manifold (FGM) model is one of these models utilized in this work for a large-eddy simulation of a turbulent flame stabilized behind a bluff body. The results show that the accuracy of this model depends on the sub-grid scale variance sub-model its parameters. An algebraic model was used to approximate the variance, but the results were not accurate enough; so that the flame height was under-estimated by approximately 30%, and the error in the mean axial velocity was more than 60% at some points in the domain. However, solving a transport equation for this quantity improves the accuracy of the predictions.