Nitric oxide formation in dilute hydrogen jet flames: isolation of the effects of radiation and turbulence-chemistry submodels

Abstract

Nitric oxide (NO) levels in turbulent flames are sensitive to temperature, O atom concentrations, and residence times at the NO forming conditions. Consequently, accurate prediction of NO formation is a challenge for turbulent combustion models, and NO is a useful scalar to consider when evaluating turbulence-chemistry submodels. Predictions of NO formation in a series of hydrogen flames with varying levels of helium dilution are compared with experimental results. Dilution reduces flame radiation to the point where uncertainties in the radiation calculation have only a small influence on NO emission, and this allows different models for the coupling of turbulence and chemistry to be compared in relative isolation from the effects of radiation. Calculations using the Probability Density Function (PDF) method and the Conditional Moment Closure (CMC) method are carried out using the same turbulence model, radiation model, reduced chemical mechanism, and boundary conditions, so that similarities and differences between these two approaches for the coupling of turbulence and chemistry may be isolated and more fully understood. To accomplish this, flow model constants are adjusted to match measured axial profiles of mean velocity and mixture fraction in the undiluted flame, before detailed scalar comparisons are carried out. The optically thin radiation assumption is shown to be appropriate for these hydrogen flames, and the effects of flame radiation on NO emission are examined. The sensitivities of the calculated results to certain boundary conditions and submodels are also quantified. Both turbulence-chemistry models yield good results for conditional mean temperature and H 2O mole fraction. Results also show that both models yield quantitatively useful predictions for the NO formation and emission. With all other submodels being the same, the PDF method predicts lower NO levels than does the CMC method. This difference is associated with lower O atom concentrations predicted by the PDF method. Both methods underpredict NO formation in the lower portion of the jet flames. This deficiency is attributed to the effects of differential diffusion, which are not included in the models, and to the influence of heat release on turbulence structure near the flame base.