Large-eddy simulation of the Darmstadt multi-regime turbulent flame using flamelet-generated manifolds

The Darmstadt multi-regime turbulent flame MRB26b is modelled using the Flamelet-Generated Manifold (FGM) in the context of Large Eddy Simulation (LES). Laminar free-propagating premixed flames are used to build a premixed chemical database, and laminar counter-flow flames are adopted to build a non-premixed and a partially premixed database. A detailed comparison of the LES-FGM results with the experimental data using the three databases is conducted, combining two sub-filter closure models, i.e., the PPDF (Presumed Probability Density Function) and DTF (Dynamically Thickened Flame). It is found that different databases lead to similar modelling of velocity, temperature, mixture fraction and major species (e.g., CH4, O2, CO2 and H2O). However, the intermediate minor species CO and H2 are better predicted using the two databases based on counter-flow flames due to a better inclusion of the fuel/air mixing and straining effects. Similar to previous studies, using the DTF model is observed to over-predict CO and H2. This is revealed to be a combined effect of the artificial flame front thickening and flame wrinkling loss, which is characterized by a lower variance of the progress variable. Comparatively, better minor species predictions are obtained using the PPDF without flame thickening. A correction method is proposed and validated to improve the minor species prediction for the DTF model. The combustion regimes and flame propagation directions are also examined. The results are significant for high-fidelity simulation of multi-regime turbulent flames using the flamelet-tabulated combustion model.& COPY; 2023 Published by Elsevier Inc. on behalf of The Combustion Institute.

Automated summary

This study models the Darmstadt multi-regime turbulent flame MRB26b using the Flamelet-Generated Manifold (FGM) in the context of Large Eddy Simulation (LES). Different databases are used to build premixed, non-premixed, and partially premixed chemical databases, and the LES-FGM results are compared with experimental data using two sub-filter closure models. The results show that the databases lead to similar modelling of major species but better predict intermediate minor species using counter-flow flame databases. The study also reveals the over-prediction of minor species using the DTF model and proposes a correction method for improvement.