Demonstration of a Hybrid Nongray Radiative Transfer Solver Based on the Full-Spectrum Correlated K-Distribution (FSCK) Method for Combustion Applications

Combustion environments have absorbing-emitting gases such as CO2 and H2O at high temperatures, . These gases have extremely oscillatory spectral properties that necessitate repeated solution of the five-dimensional Radiative Transfer Equation (RTE). The spectral gas properties can be modeled using various nongray models that may require tens, hundreds or millions of repeated RTE solutions. Radiation modeling can, hence, incur a large computational cost in a fully coupled combustion simulation. In this study, the Full-Spectrum Correlated k-distribution (FSCK) nongray model is used in conjunction with a Hybrid RTE solver. The FSCK nongray model transforms the spectral integration of the RTE such that the solution is computed over few (8-10) quadrature points with monotonically increasing spectral absorption coefficient. The Hybrid solver switches between two popular deterministic RTE solvers: lowest order spherical harmonics approximation (P1) and the finite angle method (FAM; popularly known as the finite volume method for radiation), based on the spectral optical thickness . The approach is first validated and fine-tuned for a 2D frozen flame (specified temperature and concentration), and the optimal version is finally used for a 3D coupled flame calculation. The method is found to provide accuracy of predicting temperatures comparable to FAM, but with a speedup of 4x or more.