Modeling thermal radiation in combustion environments can be extremely challenging for two main reasons. First, the radiative transfer equation (RTE), which is the cornerstone of modeling radiation in such environments, is a five-dimensional integro-differential equation. Second, the absorption and scattering coefficients of molecular gases and particulates prevalent in combustion environments oscillate strongly with the wavenumber (or wavelength), i.e., the medium is strongly nongray, requiring the solution of the RTE for a large number of wavenumbers. This article reviews the progress that has been made in this area to date with an emphasis on the work performed over the past three decades. Progress in both deterministic and stochastic (Monte Carlo) solutions of the RTE is reviewed, in addition to the review of the treatment of the spectral properties of gases, soot, and fuel droplets that dominate combustion environments, i.e., spectral or nongray models. The application of the various state-of-the-art nongray models and RTE solution methods to flames (particularly turbulent), fires, combustors, and other combustion systems are summarized along with a critical discussion of the pros and cons of the models and methods. Finally, the challenges that remain in modeling thermal radiation in combustion systems are highlighted and future outlooks are shared.
The semiempirical model developed by Springer et al. [ J. Electrochem. Soc. , 138 , 2334 (1991) ] continues to be popular for water transport in polymer membranes. The implementation of this model into a computational fluid dynamic (CFD) code poses some problems. First, the dependent variable in the membrane is the water content, λ, while that in the diffusion layers and channels is the species mass fractions. Thus, the membrane has to be designated a separate domain, wherein a transport equation for λ is solved and iteratively coupled to the mass fractions. This is computationally expensive and sometimes unstable. The second problem is associated with the notion that because the membrane is “thin” it is sufficient to have a few computational nodes across its thickness. A model that extends Springer’s model to include convective transport is presented. It is found that accurate calculation of water transport through the membrane requires few hundred nodes. A database that provides water flux and average membrane electrical conductivity for relevant mass-transport Peclet numbers and interfacial conditions is created. When used with CFD calculations, the database will allow accurate and efficient calculation of water and current transport in the membrane without solution of the transport equation for λ.