Re: [ff3d-users] Inner wall definitions

[Emilio Melero García]

Thanks a lot Stephane!, i will compile it right away and if i manage maybe you want to put it in the web for other mac users?? By the way, is the cout stream to files implemented in this version?

Anyhow, i must say now that maybe i was too enthusiastic when i said that my problems were solved using a more recent version. Some of them were, but there is something a little strange i can´t seem to get around. I attached in the e-mail the ff3d and the pov file.

The problem is a time depedent diffusion one using a foward time scheme, and the geometry tries to model a box that has two capillaries inside it. The capillaries are closed on one end and the walls of the box and the walls of the capillaries are adiabatic, but diffusion should occur inside the box and inside the capillaries. To model this i use the background mesh to define the box and i construct the capillaries with two objects in the pov file that are differences of two coaxial cylinders, and outer and an inner one like

difference{
cylinder {
<0.195,0.3375,1.0>,<0.195,0.3375,6.0>,0.055
  }
cylinder {
<0.195,0.3375,1.0>,<0.195,0.3375,5.8>,0.025
  }
pigment { color rgb <1,0,0>} 
}

difference{
cylinder {
<0.195,1.0125,0.8>,<0.195,1.0125,6.0>,0.055
  }
cylinder {
<0.195,1.0125,0.8>,<0.195,1.0125,5.4>,0.025
  }
pigment { color rgb <1,1,0>}
}

By defining the background mesh in cm as (note: X and Y elements have been set so that the length of the element, i.e. distance between vertices, is 0.01cm. The z dimension needs a larger element size or i get a too large matrix)

vector a = (0,0,0);
vector b = (0.39,1.35,6.6);
vector n = (40,136,67);
mesh M = structured(n,a,b);

and the domain as 

domain Omega = domain(S,outside(<1,0,0>) and outside(<1,1,0>));

ff3d should solve the diffusion on the box with the walls of the two "capillaries" excluded from the domain, (i.e. the will be adiabatic). The strange thing comes in that, when i run the program and visualize Omega with a tetrahedrized mesh, the cylinders defined in the pov file start and end at different values from the ones written in the pov file without much correlation. Running the pov and ff files as i send them to you you should get (i get it in 1pre7 (MACOSX) and 1pre9 (WindowsXP)) that the cylinder at lower Y values does not start at 
1.0 cm, but it ends correctly at 6.0cm, with its inner cylinder ending also correctly at 5.8cm. For the case of the other "capillary" the outer cylinder begins at 0.9 instead of 0.8 as written in the pov file, and then, while the inner cylinder ends correctly at 
5.4 cm, the outer one ends at 5.9 cm instead of at 6cm. Can you help me spot what i am doing wrong?. In the definition of the background mesh i have set the vector n equal to the length of the mesh at a particular dimension (not for Z dimension) divided by 
0.01 plus one, so that there are the correct number of elements (vertices=numofelements+1).

Thank you very very much

Best regards!!

Emilio

2007/1/31, Stephane Del Pino < address@hidden>:

On Wed, Jan 31, 2007 at 12:23:26PM +0100, Emilio Melero García wrote:
> Now i need someone to help me. I have been stucked in a problem that, at the
> end, it was caused by the old version of ff3d i am using 1pre7 for mac os x.
> I have tried to download the cvs source code from savannah following the
> instructions in the ff3d web page and in the savannah page and no success!!
>
> I have tried
>
> cvs -z3 -d:pserver:address@hidden:/sources/ff3d co ff3d
>
> and no succes (connection time out)
To me, this seems that your firewall is blocking the connection. You should contact your administrator so that he allows cvs connections.

> as per instructions in the savannah page. Can anybody help me??, please send
> me a copy of the source code so that i can compile it!!
Ok. I have made a procedure that create a snapshot of the cvs version of ff3d every night. You candownload it from
         http://www.freefem.org/ff3d/sources/
it is the freefem3d-nightly-cvs.tar.gz

Best regards,
Stephane.

-- 
###############################################
Dr. Emilio Melero García
Laboratorio de Estudios Cristalográficos, IACT
Instituto de parasitología Lopez Neyra,
P.T. Ciencias de la Salud, E-18100 Armilla, 
Granada, España

Correo elec. :emilio_arroba_lec_._csic_._es
Tfno: 958181621, lab 107
###############################################

gcb.ff
Description: Binary data

gcb.pov
Description: Binary data