• No results found

Solution of Pure Scattering Radiation Transport Equation (RTE) Using Finite Difference Method (FDM)

N/A
N/A
Protected

Academic year: 2022

Share "Solution of Pure Scattering Radiation Transport Equation (RTE) Using Finite Difference Method (FDM)"

Copied!
1
0
0

Laster.... (Se fulltekst nå)

Fulltekst

(1)

Solution of Pure Scattering Radiation Transport Equation (RTE) using Finite Difference Method (FDM)

Abstract. Radiative transfer is the physical phenomenon of energy transfer in the form of electromagnetic radiation. The propagation of radiation through a medium is affected by absorption, emission, and scattering. Radiative Transfer Equation (RTE) have been applied in a many subjects including optics, astrophysics, atmospheric science, remote sensing, etc. Analytic solutions for RTE exist for simple cases, but, for more realistic media with complex multiple scattering effects, numerical methods are required. In the RTE, six different independent variables define the radiance at any spatial and temporal point. By making appropriate assumptions about the behavior of photons in a scattering medium, the number of independent variables can be reduced. These assumptions lead to the diffusion theory (or diffusion equation) for photon transport. In this work, the diffusive form of RTE is discretized, using a Forward-Time Central-Space (FTCS) Finite Difference Method (FDM). The results reveal the radiance penetration according to Beer-Lambert law.

Keywords: Radiation Transport Modelling (RTM), Radiation Transport Equation (RTE) Finite Difference Method (FDM), Forward-Time Central-Space (FTCS)

Hassan A. Khawaja

Department of Engineering and Safety (IIS-IVT),

UiT The Arctic University of Norway, 9037, Tromsø, Norway [email protected]

SCIA 2017 - SCANDINAVIAN CONFERENCE ON IMAGE ANALYSIS, 12-14 JUNE 2017, TROMSØ

Radiation Transport Equation.

𝜕𝜕𝐼𝐼𝑣𝑣 𝒓𝒓, �𝒏𝒏, 𝑡𝑡

𝑐𝑐𝜕𝜕𝑡𝑡 + �Ω. 𝛻𝛻𝐼𝐼𝑣𝑣 𝒓𝒓, �𝒏𝒏, 𝑡𝑡 + 𝑘𝑘𝑣𝑣,𝑠𝑠 + 𝑘𝑘𝑣𝑣,𝑎𝑎 𝐼𝐼𝑣𝑣 𝒓𝒓, �𝒏𝒏, 𝑡𝑡 = 𝑗𝑗𝑣𝑣 𝒓𝒓, 𝑡𝑡 + 1

4𝜋𝜋 𝑘𝑘𝑣𝑣,𝑠𝑠

Ω

𝐼𝐼𝑣𝑣 𝒓𝒓, �𝒏𝒏, 𝑡𝑡 𝑑𝑑Ω

𝐼𝐼𝑣𝑣 is spectral radiance of electromagnetic waves 𝑐𝑐 is speed of light

�Ω is the vectorial position of a solid angle

𝑘𝑘𝑣𝑣,𝑠𝑠 is the scattering opacity of the medium

𝑘𝑘𝑣𝑣,𝑎𝑎 is the absorption opacity of the medium

𝑗𝑗𝑣𝑣 is the emission coefficient of the medium 𝑡𝑡 is time variable

Two Dimensional Pure Scattering Radiation Transport Equation.

𝜕𝜕𝜙𝜙𝑣𝑣 𝒓𝒓, 𝑡𝑡

𝑐𝑐 𝜕𝜕𝑡𝑡 = 𝐷𝐷𝛻𝛻𝟐𝟐𝜙𝜙𝑣𝑣 𝒓𝒓, 𝑡𝑡 = 𝐷𝐷 𝜕𝜕𝟐𝟐𝜙𝜙𝑣𝑣 𝒓𝒓, 𝑡𝑡

𝜕𝜕𝑥𝑥𝟐𝟐 + 𝜕𝜕𝟐𝟐𝜙𝜙𝑣𝑣 𝒓𝒓, 𝑡𝑡

𝜕𝜕𝑦𝑦𝟐𝟐 𝜙𝜙𝑣𝑣(𝒓𝒓, 𝑡𝑡) is radiance intensity

𝑐𝑐 is speed of light

𝑥𝑥, 𝑦𝑦 are the space dimensions 𝐷𝐷 is diffusion coefficient

𝑡𝑡 is time variable

Forward-Time Central-Space (FTCS) Finite Difference Method.

𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖,𝑗𝑗𝑡𝑡+1

= 𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖,𝑗𝑗𝑡𝑡

+ 𝐷𝐷 𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖+1,𝑗𝑗𝑡𝑡

− 2𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖,𝑗𝑗𝑡𝑡

+ 𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖−1,𝑗𝑗𝑡𝑡

∆𝑥𝑥

2

∆𝑡𝑡 + 𝐷𝐷 𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖𝑡𝑡,𝑗𝑗+1

− 2𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖𝑡𝑡,𝑗𝑗

+ 𝜕𝜕𝜙𝜙

𝑣𝑣

𝒓𝒓, 𝑡𝑡

𝑖𝑖𝑡𝑡,𝑗𝑗−1

∆𝑦𝑦

2

∆𝑡𝑡

subscripts i and j represent the computational points in the space domain superscript t represents the transient state

Figure.

Contour plots with various radiance intensities

.

Conclusion. The Radiation Transport Equation (RTE) can be solved using a Forward-Time Central-Space (FTCS) Finite Difference Method (FDM) in special cases such as a highly diffusive mediums (as discussed above). The results show the radiance intensity in space, which reflects on the radiance penetration.

Future Work. This work can be further extended by varying the diffusion coefficient and observing its impact on the radiance penetration. It will help to find the validity limits of Radiation Transport Equation (RTE) solution using Forward-Time Central-Space (FTCS) Finite Difference Method (FDM).

Figure.

Radiance intensities in the two- dimensional discretized domain. Indices i and j refer to the nodal position of radiance intensity.

Figure.

Flow chart of the method of solution.

Loop over time and space!

No!

Start

Define total time, total space variables…

Allocate size, dimensions to radiance intensity variable…

Define boundary conditions…

Solve equation (over time and space) … Define time steps, space steps variables…

Define speed of light DQG spectral diffusion

Meet CFL criteria…

Yes!

Define Lnitial conditions…

Loop over time and space finishes!

Check Uesults…

݅ = 1 ݅ = 2 ݅ = 3 ݅ = 4 ݅ = 5 ݅ = 6 ݅ =݊

݆ = 1

݆ = 2

݆ = 3

݆ = 4

݆ = 5

݆ = 6 ڭ

݆ =݉

߶ ࢘,ݐ ଵ,ଵ ߶ ࢘,ݐ ଶ,ଵ ߶ ࢘,ݐ ଷ,ଵ ߶ ࢘,ݐ ସ,ଵ ߶ ࢘,ݐ ହ,ଵ ߶ ࢘,ݐ ଺,ଵ ߶ ࢘,ݐ ௡,ଵ

߶ ࢘,ݐ ଵ,ଶ ߶ ࢘,ݐ ଶ,ଶ ߶ ࢘,ݐ ଷ,ଶ ߶ ࢘,ݐ ସ,ଶ ߶ ࢘,ݐ ହ,ଶ ߶ ࢘,ݐ ଺,ଶ ߶ ࢘,ݐ ௡,ଶ

߶ ࢘,ݐ ଵ,ଷ ߶ ࢘,ݐ ଶ,ଷ ߶ ࢘,ݐ ଷ,ଷ ߶ ࢘,ݐ ସ,ଷ ߶ ࢘,ݐ ହ,ଷ ߶ ࢘,ݐ ଺,ଷ ߶ ࢘,ݐ ௡,ଷ

߶ ࢘,ݐ ଵ,ସ ߶ ࢘,ݐ ଶ,ସ ߶ ࢘,ݐ ଷ,ସ ߶ ࢘,ݐ ସ,ସ ߶ ࢘,ݐ ହ,ସ ߶ ࢘,ݐ ଺,ସ ߶ ࢘,ݐ ௡,ସ

߶ ࢘,ݐ ଵ,ହ ߶ ࢘,ݐ ଶ,ହ ߶ ࢘,ݐ ଷ,ହ ߶ ࢘,ݐ ସ,ହ ߶ ࢘,ݐ ହ,ହ ߶ ࢘,ݐ ଺,ହ ߶ ࢘,ݐ ௡,ହ

߶ ࢘,ݐ ଵ,଺ ߶ ࢘,ݐ ଶ,଺ ߶ ࢘,ݐ ଷ,଺ ߶ ࢘,ݐ ସ,଺ ߶ ࢘,ݐ ହ,଺ ߶ ࢘,ݐ ଺,଺ ߶ ࢘,ݐ ௡,଺

߶ ࢘,ݐ ଵ,௠߶ ࢘,ݐ ଶ,௠߶ ࢘,ݐ ଷ,௠߶ ࢘,ݐ ସ,௠߶ ࢘,ݐ ହ,௠߶ ࢘,ݐ ଺,௠ ߶ ࢘,ݐ ௡,௠

1

/11 1 *1+,1

1

1

1

1

1 1

/1 /-1

+,1 %&1

1 *1 1

1

+,1 +,1

1 1

1

1

1

/$11 1

+,1 /$1+,1

1 *1 1 *1

1 1

+,1 +,1

1 1

1

1

1

/$11 1

+,1 /$1+,1

1 *1 1 *1

1 1

+,1 +,1

1 1

1

1

1

/$11 1

+,1 /$1+,1

1 *1 1 *1

1 1

+,1 +,1

1 1

1

1

:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

UDGLDQFHLQWHQVLW\:P

\SRVLWLRQP

[SRVLWLRQP

:P

:P :P

[SRVLWLRQP

[SRVLWLRQP

:P :P

[SRVLWLRQP

[SRVLWLRQP

:P :P

[SRVLWLRQP

[SRVLWLRQP

:P :P

[SRVLWLRQP

[SRVLWLRQP [SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

\SRVLWLRQP

Referanser

RELATERTE DOKUMENTER

We present a method for constructing first integral preserving nu- merical schemes for time-dependent partial differential equations on non- uniform grids, using a finite

Kato, On the Cauchy problem for the (generalized) Korteweg–de Vries equation. Studies in Applied Mathematics, Adv. Introduction to Nonlinear Dispersive Equations. A comparison

We prove regularity in time and space by using Duhamel's formula, combining it with known regularity of the unique viscosity solution for the fractional Hamilton-Jacobi

One plot is the solution using standard Galerkin finite element method with linear elements and one plot is the solution using mixed finite element method with Arnold-Winther

The SHDB boundary condition is a general linear boundary condition that contains two scalar equations that involve both the tangential and normal components of the

stochastic partial differential equation, Monte Carlo method, random advection equation, finite difference/volume schemes, uncertainty quantification, stochastic

A key advantage of the linearized complex eikonal equations is providing possible method of solving the TTI complex eikonal equation using a finite-difference method, which can

We show that the Engquist-Osher (and hence all monotone) finite difference approximations converge to the unique entropy solution of the governing equation if, among other demands,