Abstract

This chapter investigates the parallelization of discrete transfer radiation model, focusing on different solutions to the radiative transfer equation. Speed-up efficiencies of a constant absorption coefficient solution, a weighted sum of gray gases solution, and a differential total absorptivity solution are compared across a non-uniform mixture of combustion gases and soot. The discrete transfer radiation model (DTRM), has been a popular choice for many researchers in the computational fluid dynamics (CFD) combustion community. Relative to other models, it provides a relatively fast and accurate means for predicting the volumetric and surface fluxes within a domain. It is excellently suited for coupling to curvilinear (body-fitted) CFD grids and its implementation does not obscure the fundamental physics of the phenomenon. Additionally, the ray tracing approach employed is such that it provides a convenient basis for comparing different algorithms for solving the radiative transfer equation (RTE). The DTRM comprises a number of embedded loops culminating in separate integrals for the volumetric and surface fluxes. For each iteration in a sequential calculation, boundary elements are visited in turn, rays are launched from them, and the innermost loop traverses the elemental paths defined by the intersection of a ray with a control volume. Thus, within the outer iteration cycle, parallelization is achieved by splitting one of the inner loops. For the purposes of the present work, it was decided to parallelize the ray tracing loop, because this approach involves the smallest amount of communication between processors

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call