Optimization of the MSMGWB model used for the calculation of infrared remote sensing signals from hot combustion gases of hydrocarbon fuel

Abstract

To address the influence of initialization on the wavenumber subinterval grouping result and calculation accuracy of the multiscale multigroup wide-band k-distribution model (MSMGWB), an optimization platform was established to determine the best combination of grouping result and Gauss quadrature scheme. An optimization objective function was established based on weighted calculation errors of the MSMGWB model in 18 1D cases, in which radiance in the 3- to 5-µm wave band emitted by hot combustion gases of hydrocarbon fuel was attenuated by the atmosphere with various thermodynamic state parameters. A new initialization method for wavenumber subinterval grouping similar to the k-means++ algorithm was established to improve the efficiency of the optimization platform. The optimization result was evaluated using additional 1D cases with and without the influences of various kinds of aerosols in the atmosphere. Finally, evaluation results are presented on the calculation of remote sensing thermal images for a supersonic aircraft exhaust system and its hot jet. The results indicate that the optimized model has a higher calculation accuracy, better compatibility and lower computational cost compared with the original one.