Re: [Getfem-users] solvers won't solve

[Michael . Walker]

Dear Ronan,
GOT IT!! The problem was in the region of code that builds the mesh into
getfem. Specifically, I changed myarray.transpose() into myarray =
myarray.transpose() . Presumably this is a subtle difference between
numarray and numpy.

Thank you to everyone on the list, especially Ronan, Yves and Luis, who
provided feedback on this problem. It all helped take me off wrong
directions. The final output is attached for those who would like to see.
thank you
Michael


> I had a look in 3D on your mesh and it looks really ugly...
> It is possible (I would even tell it is highly probable) that you do not
compute on the mesh you think.
> Could you send me a view of your 3D mesh?
>
> Ronan
>
> address@hidden a écrit :
>> Hi Ronan,
>> the mesh you sent me is certainly strange. Have you viewed it in three
dimensions? I viewed it using our 3D viewer and it looks fine. It is a
three dimensional object with diffusion happening over the surface. One
2-simplex is a source, another is a sink, and the rest are
>> undetermined. So the system shouldn't be singular, or do I need an initial
>> condition?
>> I'll pass this on to my colleague and see what he says.
>> thank you
>> Michael
>>> Hi Michael,
>>> I commented out the display part.
>>> It seems to work fine on my computer... However, the solution contains 0
>>> or Nan and the mesh is quite strange.
>>> Is it the good mesh (see the enclosed file)?
>>> I convert your python file to matlab and I obtained the same error as
yours (I do not know why it works in python). However the mesh is
quite
>>> strange...
>>> Regards,
>>> Ronan
>>> address@hidden a écrit :
>>>> Hi Ronan,
>>>> attached are the python file diffusion_help.py, which is a simplified
version of what I am trying to solve except that it loads the mesh
from
>>>> 'mesh.txt', which is also attached.
>>>> It runs in the same way as the actual code.
>>>> thanks again
>>>> Michael
>> Michael Walker
>> Plant Modelling Group
>> CIRAD, Montpellier
>> 04 67 61 57 27
>
>


Michael Walker

Plant Modelling Group
CIRAD, Montpellier
04 67 61 57 27

goodmeshshot.png
Description: PNG image