Assessing accuracy, reliability, and efficiency of combustion models for prediction of flame dynamics with large eddy simulation

Abstract

A variety of combustion models for large eddy simulation of premixed turbulent flames have been developed and validated over the years. Validation studies concentrate on relevant mean quantities and turbulent fluctuations; however, the prediction of flame dynamics is typically not taken into account. Furthermore, it is difficult to meaningfully compare the computational efficiency of model formulations due to different compute resources, meshes, code bases, and numerics. The present study compares turbulent combustion models on the same code base, keeping boundary conditions, meshes, and numerical settings constant. The reliability and versatility of two turbulent combustion models, i.e., the artificially thickened flame and flame surface density formulations, are assessed by applying them to a variety of operating conditions and burner configurations. In particular, for a premixed methane swirl burner, we consider three power ratings by changing the inflow velocity, which increases the demand on the sub-grid scale model due to increased sub-grid scale wrinkling. A change in swirler position modifies the interference of swirl and acoustic perturbations, with a significant impact on flame dynamics. Changes in thermal boundary condition and combustion chamber size provide insight into the consequences of quenching effects resulting from heat losses on flame anchoring and flame topology.