A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Visual Map Navigation

View in Semantic Map View in A-Z Index

Related content in other products

International Heat Transfer Conference Digital Library International Centre for Heat and Mass Transfer Digital Library Begell House Journals Annual Review of Heat Transfer

Radiation of nonisothermal layer of scattering medium

DOI: 10.1615/thermopedia.000135

Radiation of a Nonisothermal Layer of Scattering Medium

Leonid A. Dombrovsky

Following from: P1 approximation of spherical harmonics method, The simples approximations of double spherical harmonics, Solutions for one-dimensional radiative transfer problems

Let us consider the thermal radiation transfer in a plane-parallel layer of absorbing and scattering medium between walls having temperatures Tw1, Tw2 and spectral emissivities εw1, εw2. The temperature profile in the medium layer, T(x), is assumed to be known.

According to the diffusion approximation, the spectral radiation energy density can be obtained by solving the following boundary-value problem:

(1a)

(1b)

(1c)

(1d)

Here, SB0 = 4πSB = 4αλπBλ(T) (one-temperature medium) and Swj = 4πBλ(Twj). The spectral radiation fluxes toward the walls are

(2a)

(2b)

and the profile of the heat generation in the medium can be obtained by integrating over the spectral range:

(3)

In the case of arbitrary profiles αλ(z), Dλ(z), and Sλ(z), boundary-value problem (2) cannot be solved analytically and needs a numerical solution. The usual finite-difference approximation from Eq. (1a) is

(4)

where

(5)

For a better approximation of the boundary conditions one should use the expansion of function Iλ(z) in Taylor’s series in the neighborhood of the boundary nodes of the mesh:

(6a)

(6b)

where (d2Iλ0/dz2) is determined through Iλ0 and (dIλ0/dz) is determined according to differential Eq. (1a). After transformations, the finite-difference equation and boundary conditions can be written as follows:

(7a)

(7b)

In accordance with the factorization method, the relation between the neighboring nodal values of the function to be found is

(8)

where φi and ψi are the factorization coefficients. From the finite-difference form of the boundary condition at z = 0 we find

(9)

The formulas of upward factorization

(10)

allow obtaining all of the values of φi and ψi up to i = n. After that, one can calculate the value of I0λ,n+1 from the boundary condition at z = d:

(11)

and all of the other I0λ,i are determined by Eq. (8).

A similar algorithm can also be developed with the use of the DP1 approximation. For the plane-parallel layer of a medium, boundary-value problems (1a), (1b), (2a), and (2b) from the article The Simplest Approximations of Double Spherical Harmonics have the following vector form:

(12a)

(12b)

Here,

The other designations are evident from the comparison of Eqs. (12a) and (12b) with Eqs. (1a), (1b), (2a), and (2b) from the article The Simplest Approximations of Double Spherical Harmonics. Let us rewrite Eqs. (12a) and (12b) in the form of the boundary-value problem for function :

(13a)

(13b)

(13c)

The form of these equations coincides with that of Eq. (7). Therefore, the algorithm of matrix factorization for problems (13a)-(13c) is quite similar to the above-described scalar factorization for the diffusion approximation.

Number of views: 15294 Article added: 7 September 2010 Article last modified: 18 March 2011 © Copyright 2010-2017 Back to top