Sparse MMC-LES of a Sydney swirl flame

Abstract Sparse multiple mapping conditioning-large eddy simulation (MMC-LES) is applied for the first time to a flame involving swirl, thus extending the model to complex flow conditions that are more indicative of practical combustors than is the case for the jet flames simulated previously. Two cases from the University of Sydney swirl flame series are chosen: (i) non-reacting case N16S159 is simulated first to test the resolution of the unstructured mesh and boundary conditions; followed by (ii) methane–air swirl flame SMA2 which has the same flow boundary conditions as N16S159. MMC-LES is a filtered density function (FDF) approach and accurate emulation of the subfilter conditional fluctuations is sensitive to the modelling of the mixing time scale. Two models are tested: (i) the original mixing time scale model (C&K model) proposed by Cleary and Klimenko (2011) and used in all past MMC-LES of laboratory flame cases, and (ii) a new, consistently anisotropic model (a-ISO model) proposed recently by Vo et al. (2017) and validated against direct numerical simulation (DNS) only. Overall, the swirl flow predictions are in very good agreement with the experimental data and both mixing time scale models produce similar statistical results for axial and azimuthal velocity, mixture fraction and unconditional reacting scalars including temperature and CO and OH mass fraction. While the predictions of conditional mean temperature show only a minor sensitivity to the mixing time scale selection, the a-ISO model produces significantly better results for the conditional fluctuations of temperature.