The main goal of this project was a first approach of the mesher Icem, with a simple geometry and a little physical application on Fluent5.
In this report, one can notice that the meshing has been performed and the results obtained on Fluent5 are physically admittable.

However, this project did not turn out to be very successful and enriching. Indeed, we must draw conclusion from it.
Firtsly, it is obvious that Icem is too a powerful mesher for a soo simple geometry such as an air filter. It owns an impressing number of functions and options, which were completely unuseful for us.
Secondly, as a direct consequence of the first, a very long time is necessary to be able to use Icem with efficiency. Even for meshing a simple geometry like ours, many difficulties were encountered.

Finally, ICEM turns out to be a very efficient mean for meshing, and enables to generate the meshing of very complex geometries. For a first initiation, it was therefore logical to begin with a simple geometry such as ours.

As a conclusion, we can remind several results or remarks concerning the using of Icem and Fluent5 with a porous media :

- In ICEM, all the curves which are situated inside the fluid domain and which are not "wall" must be gather in a same family "intern" and declare "interior" in Fluent5 boundary conditions.

- In ICEM, the porous media needs to be considered as a fluid domain.

- In Fluent5, the porous media must be declared "porous zone" in the fluid boundary conditions. Then, several characteristics are required. The vector components, which represents the isotropic property of the porous media. The viscosity resistance which is the inverse of the permeability (1/m2). The inertial resistance (1/m) which represents a second order term in the Darcy law, when the Reynolds number is not little enough.