Evaluation of emissivity correlations for H2O-CO2-N-2/air mixtures and coupling with solution methods of the radiative transfer equation

This study is directed towards the limitations of applying total emissivity correlations in computational fluid dynamics (CFD) computer codes for flame modeling. The predictions of nine widely applied total emissivity models for H2O-CO2 homogeneous mixtures are compared with the exponential wide band model (EWBM) calculations. The comparison covers a range of total pressures, temperatures and path lengths which are suitable for the use of fine numerical grids in CFD simulations of atmospheric and high pressure combustors. Attention is paid to coupling of the property models with the radiative transfer equation (RTE) and their performance in non-homogeneous applications. In this respect both the total transmittance non-homogeneous (TTNH) model and the spectral group model (SGM) are used. The latter model is combined with five weighted sum of gray gases models (WSGGM), the single line based sum of gray gases model (SLW) and the k-distribution model. The non-homogeneous validation tests used in situ total radiance measurements in two non-luminous natural gas flames representing two industrial situations, a water cooled furnace and a refractory lined furnace. The main conclusions are as follows. The spectral group model provides an elegant and accurate method of coupling WSGGM, k-distribution and SLW property models to the equation of radiative transfer. Both homogeneous and non-homogeneous tests indicate the advantage of using the Smith, Shen and Friedman weighted sum of gray gases model over polynomial correlations and the SLW model. It has been shown that in the near burner region of a natural gas diffusion flame, the water vapor to carbon dioxide partial pressure ratio departs significantly from the value expected for the complete combustion of methane in air. This finding emphasizes the limitation of existing WSGGM to H2O to CO2 partial pressure ratios of one and two only. (C) 1997 Elsevier Science Ltd.