Modeling of Local Extinction in Turbulent Flames

The Eddy Dissipation Concept (EDC), proposed by Magnussen (1985), advances the concept that the reactants are homogeneously mixed within the fine eddy structures of turbulence and that the fine structures may therefore be regarded as perfectly stirred reactors (PSRs). To understand more fully the extent to which such a subgrid scale stirred reactor concept could be applied within the context of a computational fluid dynamics (CFD) calculation to model local or global extinction phenomena: (1) Various kinetic mechanisms are investigated with respect to CPU penalty and predictive accuracy in comparisons with stirred reactor lean blowout (LBO) data and (2) a simplified time-scale comparison, extracted from the EDC model and applied locally in a fast-chemistry CFD computation, is evaluated with respect to its capabilities to predict attached and lifted flames. Comparisons of kinetic mechanisms with PSR lean blowout data indicate severe discrepancies in the predictions with the data and with each other. Possible explanations are delineated and discussed. Comparisons of the attached and lifted flame predictions with experimental data are presented for some benchscale burner cases. The model is only moderately successful in predicting lifted flames and fails completely in the attached flame case. Possible explanations and research avenues are reviewed and discussed.