Investigation of turbulence models capability in predicting mixing in the near field region of hydrogen–hydrocarbon turbulent non-premixed flame

• Published in: Heat and mass transfer 2011-04, Vol.47 (4), p.397-406
• Main Authors: Tabet, F., Sarh, B., Gökalp, I.
• Format: Article
• Language: English
• Subjects: Applied sciences; Combustion of gaseous fuels; Combustion. Flame; Energy; Energy. Thermal use of fuels; Engineering; Engineering Thermodynamics; Exact sciences and technology; Fluid dynamics; Fundamental areas of phenomenology (including applications); Heat and Mass Transfer; Industrial Chemistry/Chemical Engineering; Original; Physics; Theoretical studies. Data and constants. Metering; Thermodynamics; Turbulence simulation and modeling; Turbulent flows, convection, and heat transfer
• ISSN: 0947-7411; 1432-1181
• DOI: 10.1007/s00231-010-0725-0

The aim of the present study is to assess numerically the capability of two turbulence approaches, Reynolds averaged Navier-Stokes equations (RANS) and large eddy simulation (LES), in predicting mixing in hydrogen–hydrocarbon turbulent non-premixed flame. The near field region of this flame is characterized by a high density ratio and a high velocity ratio between the air coflow and the fuel which makes difficult mixing modelling in this zone. This is important as flame stabilization and pollutant formation occurs in this region. In addition, differential diffusion effects may exist. The turbulence RANS models selected are Realizable κ–epsilon and Reynolds stress model, respectively. LES is performed with Smagorinsky sub-grid model. Mixing is described using mixture fraction and its variance. Mixture fraction characterizes the mixing state between fuel and oxidizer while the scalar fluctuation intensity reflects the effect of unmixedness. The effect of differential diffusion on mixing is not taken into account as the target here is mixing modelling. The predictions indicate that the RANS turbulence models applied fail to predict the initial stages of mixing while reasonable description is achieved with LES.