CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   packed-bed geometry creation (

sara1987 May 14, 2012 08:38

packed-bed geometry creation
dear all
I'm trying to do some analysis on latent heat thermal energy storage using a packed bed of spheres, i need to run the simulation in 2d but the problem is that the direct treatment as a 2-D model will lead to no flow passage for the fluid due to the complex packing of the spheres and the contact between them.
the bed have following dimensions:
360mm diameter
460mm height
264 spheres with outer diameter of 55mm uniformly packed in eight layer

anyone can help with this?

mettler May 14, 2012 12:49

if you are using Fluent (I just know Fluent has a porous media model) you can treat it as porous media. You'll have to figure out some of the variables, but it can be done. There are journal articles written modeling it as porous media too.

sara1987 May 15, 2012 06:09


Originally Posted by mettler (Post 361009)
if you are using Fluent (I just know Fluent has a porous media model) you can treat it as porous media. You'll have to figure out some of the variables, but it can be done. There are journal articles written modeling it as porous media too.

thnx for the response mettler
do you have any specific article in mind to help me with this?
i also wanna treat it as a porous media but i have problem with the geometry, i don't know how i should put the spheres so that have acceptable flow passage. by the way the void fraction or the bed porosity is 0.48.

mettler May 15, 2012 08:47

If you are using Fluent you will not have to draw every sphere etc. You will just have to draw the dimensions of the pipe/bed/whatever that is holding the spheres and then define certain variables that will let Fluent know that the section is a packed bed. The porosity will be one of them. When I did this a few years ago I had to do some measurements in the lab to get some of the other variables. Fluent's manual has a good section on porous media that will help you understand it better.

I'll try to remember to get some author's names when I get home. Jiang comes to mind, but I am not sure if it was packed beds he was doing.

sara1987 May 15, 2012 15:20

I'm using ansys fluent 13
I need the liquid fraction of the pcm (phase change material which is paraffin wax ) and the outlet temperature after the system is being totally charged...

so you sure that i don't have to draw every sphere?

mettler May 15, 2012 15:31

get out the fluent PDF and read the porous media section.. yes, you don not have to draw all the spheres..:)

You will separate you section into an entrance length, and then the fluidized bed, and an exit section. The bed section will be defined as porous media and you will have to fill in all the necessary variables. That is what I had to do anyway.

saramito May 15, 2012 15:56

Rheolef NEW version 6.1: an efficient FEM C++ finite element library for solving PDE
Rheolef: an efficient FEM C++ finite element library for solving PDE

Version : 6.1

Distibution: sources and binaries as debian packages

Keywords: finite elements, numerical simulation, partial derivative equations,
C++, meshes, graphics

* equations on a surface: implements three diferent FEM methods
* improves the high order Pk Lagrange interpolation implementation
* ports on intel c++ 12.0 and gnu c++ 4.7 new compiler versions

Previous features:
Rheolef is a programming environment that serves as a convenient laboratory for
computations involving finite element methods (FEM) for solving partial
differential equations (PDE). Rheolef is both a C++ library and a set of
commands for unix shell programming, providing algorithms and data structures.

* Algorithms refer to the most up-to-date ones: preconditioned sparse solvers
for linear systems, incompressible elasticity, Stokes and Navier-Stokes flows,
characteristic method for convection dominated heat problems, etc. Also
nonlinear generic algorithms such as fixed point and damped Newton methods.

* Data structures fit the standard variational formulation concept: spaces,
discrete fields, bilinear forms are C++ types for variables, that can be
combined in any expressions, as you write it on the paper.

Combined together, as a Lego game, these bricks allows the user to solve most
complex nonlinear problems. The concision and readability of codes written
with Rheolef is certainly a major keypoint of this environment.

Main features

* [NEW] Massively distributed memory finite element environment, based on MPI.
* [NEW] High-order polynomial approximation.
* Poisson problems in dimension d=1,2,3.
* Stokes problems (d=2,3), with Taylor-Hood or stabilized P1 bubble-P1 elements.
* linear elasticity (d=1,2,3), including the incompressible case.
* characteristic method for time-dependent problems:
transport, convection-difusion, and Navier-Stokes equations.
* input and output in various file format for meshes generators and numerical
data visualization systems.

Advanced features

* auto-adaptive mesh algorithms.
* axisymetric problems.
* multi-regions and non-constant coefficients.
* nonlinear problems with either fixed-point algorithms or a provided generic
damped Newton solver.
* 3d stereo visualization

Both reference manual and users guide are available.

The license is GPL.

Pierre Saramito
Directeur de Recherche CNRS
Laboratoire Jean Kuntzmann, Grenoble, France

sara1987 May 16, 2012 23:53

thnx mettler it was helpfull :)

jchawner May 20, 2012 18:24


I'm aware of 2 similar things that we've done. The first is written about on our website:

In another case we scripted the meshing of the flow around an automatically generated array of a variety of sphere sizes. We can't show pictures of that because it was proprietary to the client. But we could describe how it was meshed.

If interested in the details, send me an email at and I'll forward it to the right people who can give you the answer.

Best Regards

All times are GMT -4. The time now is 21:57.