| Résumé | The walls of combustion systems are usually assumed to be black or gray in radiative calculations, which may cause large errors. The Planck-function-weighted emissivity is usually used as the gray-wall emissivity when using the Wide-Band Correlated K-distribution method. This approach can demonstrate good accuracy when coupled with the emissivity-based optimized band interval approach proposed by Solovjov et al., 2013. To improve computational efficiency without losing accuracy, a nongray-wall emissivity model and an absorption-coefficient-based method for determining the band interval are proposed. The accuracy of nongray-wall and gray-wall emissivity models along with different approaches to determine the band intervals is evaluated in three 1D isothermal and homogeneous cases bounded by fly-ash deposit, a high-temperature alloy, and soot deposit and a 3D fuel-air flame bounded by fly-ash deposit. The results show that the nongray-wall emissivity model is much more accurate than the gray-wall one when the number of bands is greater than 1. Coupled with the absorption-coefficient-based band interval approach, the nongray-wall emissivity model becomes more accurate when the number of bands is larger than 2, especially for low-temperature walls. It is sufficient to divide the entire spectrum into 4 bands for the cases tested here. |
|---|
