Oh, I see. Obviously, that won't work. If I do the mesh again and have all tetrahedrals, I get the previous error message "Index 0 out of bounds for axis 0 with size 0". The inlet planes of the mixer are at x=0. Would it help to shift all coordinates to positive values?!

h.

On Friday, August 2, 2013 10:35:32 AM UTC+2, Robert Cimrman wrote:

On 08/02/2013 09:48 AM, Hendrik Mohrmann wrote:
> Here they are. Thanks again for your help.
>
> Regards, Hendrik

The mesh is 3D, but contains only a (partial?) surface made of triangles - no
3D elements like tetrahedrons.

You can see that yourself by running

$ ./postproc.py micromixer.mesh --wireframe

SfePy should definitely die more gracefully in this case. I will add some
checks and hopefully helpful error messages.

r.

>
>
>
>
> On Thursday, August 1, 2013 5:31:46 PM UTC+2, Robert Cimrman wrote:
>>
>> Could you send me a complete problem description file + mesh so that I can
>> try it?
>>
>> r.
>>
>> On 08/01/2013 05:18 PM, Hendrik Mohrmann wrote:
>>>
>>>>
>>>> Great, that helped for that issue! Now I got the next error. Sfepy is
>>>> giving up on my stupid attempts :)
>>>>
>>>
>>>
>>> ...
>>> sfepy: fluid
>>> sfepy: ...done in 0.00 s
>>> warp violation 0.000000e+00 at (iel: 0, iqp: 0)!
>>> warp violation 0.000000e+00 at (iel: 0, iqp: 1)!
>>> warp violation 0.000000e+00 at (iel: 0, iqp: 2)!
>>> warp violation -2.282583e-320 at (iel: 0, iqp: 3)!
>>> mem_free_mem(): error exit!
>>> mem_free_mem(): error exit!
>>> mem_free_mem(): error exit!
>>> mem_free_mem(): error exit!
>>> mem_free_mem(): error exit!
>>> mem_free_mem(): error exit!
>>> sfepy: giving up!
>>> Traceback (most recent call last):
>>> File "./simple.py", line 146, in <module>
>>> main()
>>> File "./simple.py", line 143, in main
>>> app()
>>> File "/home/rasputin/Uni/sfepy/sfepy/applications/application.py",
>> line
>>> 29, in call_basic
>>> return self.call(**kwargs)
>>> File "/home/rasputin/Uni/sfepy/sfepy/applications/pde_solver_app.py",
>>> line 213, in call
>>> nls_status=nls_status)
>>> File "/home/rasputin/Uni/sfepy/sfepy/solvers/ts_solvers.py", line 29,
>> in
>>> __call__
>>> state = problem.solve(state0=state0, nls_status=nls_status)
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/problemDef.py", line 933, in
>>> solve
>>> vec = solvers.nls(vec0)
>>> File "/home/rasputin/Uni/sfepy/sfepy/solvers/nls.py", line 278, in
>>> __call__
>>> vec_r = fun( vec_x )
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/evaluate.py", line 52, in
>>> eval_residual
>>> vec_r = self.problem.equations.eval_residuals(vec)
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/equations.py", line 583, in
>>> eval_residuals
>>> self.evaluate(mode='weak', dw_mode='vector', asm_obj=out)
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/equations.py", line 526, in
>>> evaluate
>>> asm_obj=asm_obj)
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/equations.py", line 753, in
>>> evaluate
>>> ret_status=True)
>>> File "/home/rasputin/Uni/sfepy/sfepy/terms/terms.py", line 1457, in
>>> evaluate
>>> fargs = self.call_get_fargs(_args, kwargs)
>>> File "/home/rasputin/Uni/sfepy/sfepy/terms/terms.py", line 1278, in
>>> call_get_fargs
>>> fargs = self.get_fargs(*args, **kwargs)
>>> File "/home/rasputin/Uni/sfepy/sfepy/terms/termsNavierStokes.py",
>> line
>>> 54, in get_fargs
>>> vg, _ = self.get_mapping(state)
>>> File "/home/rasputin/Uni/sfepy/sfepy/terms/terms.py", line 1205, in
>>> get_mapping
>>> return_key=return_key)
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/fields_base.py", line 1036,
>> in
>>> get_mapping
>>> return_mapping=True)
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/fea.py", line 349, in
>>> describe_geometry
>>> ori=self.ori)
>>> File "/home/rasputin/Uni/sfepy/sfepy/fem/mappings.py", line 270, in
>>> get_mapping
>>> cmap.describe(self.coors, self.conn, bf_g, ebf_g, weights)
>>> File "mappings.pyx", line 101, in
>>> sfepy.fem.extmods.mappings.CMapping.describe
>>> (sfepy/fem/extmods/mappings.c:2778)
>>> ValueError: ccore error (see above)
>>>
>>>
>>> Another question: Wouldn't it be helpful to post this in a public forum,
>> so
>>> everyone would have these helpful information if they google the
>> problem?
>>>
>>
>>
>