Messed up problem...IGES


I'm just new to this CFD software; and I have a problem.

From my drawing program I import a block with my model extracted,
as an .igs file. I simplified the model to make the calculation as simple as possible.

All conditions are set; walls, inlet, outlet, medium, and speed.

Initial arrows are displayed well, but when I run the solver I get the following Warning:
(earlier models worked fine)

volume [www.1341.5] RANS Flow: Failed
Solver singularity: smoothSolver: Solving for epsilon, Initial residual = -1.#IND, Final residual = 1.#QNAN, No Iterations 100

I have red that about 90% of the warnings are messing problems;
I doesn't seem to be my model (it's a converted solid, must be ok).

Has anyone a clue to solve this?.

Another question; is ther a simple way to select more faces in one selection?;
I wan to make a group op selected faces, but I don't want to select each face seperatly.

Kind regards, fhoos

Still Likely a Geometry/Mesh Problem

Imported IGES (.igs) files only provide faces so I guess you successfully used the Volumes->Stitch tool on the Geometry Tool Palette to form a volume, right?

If you have the option STEP (.stp) is a much better geometry exchange format - it preserves topology (volumes) and accuracy better than IGES.

When you say you successfully set the speed, I presume that you mean you set the Substance->Properties->Reference->U, right?

If you are confident that your boundary conditions are correct, then it is still likely a meshing problem you are facing. You need to check the quality of your face and volume cells as outlined in the FAQ "How do I fix mesh and solver failures?" and apply the Accuracy tool from the Physics Tool Palette as needed.

Yes you can make multiple selections, e.g., all faces, for details review, or better still try, the tutorial "Flow Over a Rotating Wheel with Moving Ground" and see "Camera View Control and Entity Selection"

Even more messed up...


I just looked up my file and solid, to make it better I deleted, all thin and difficult faces,
the model is now as simple as can be.The software gives the following warning:

volume [www.1341.5] RANS Flow: Failed
Unexpected process exit.

Illegal cell label -1 in neighbour addressing for face 347362

From function polyMesh::initMesh()
in file meshes/polyMesh/polyMeshInitMesh.C at line 65.

FOAM exiting

Is it hard to trace this errror?.

Kind regards, fhoos

Check mesh metrics

I'd guess the meshing failed. Was there an earlier warning concerning meshing?

Have you tried looking at face metrics and then volume metrics as outlined in "How do I fix mesh and solver failures?"

Still a mess...


Since my last post I threw away my model and built a complete new one.

The model was as simple as possible, al strange corners etc. were removed.
All doubtfull geometrie was simplified, to simple as a rock solid features.

I have many problems making the model work. at first it wouldn't solve at all.
(I got all the time warnings about settings as Epsilon, p, and others).

Yesterday evening it solved for one time, and the results were perfect.

Today I had the same mess problems all over, and now it solves but the outcome
doesn't make any sence, even after re-making the complete file. I also checked the in-and outlet settings and if the speed was determed in the correct direction.

(I got a turbulence through about my whole model, in all directions, that can't be the case with a medium as water with a speed of 2 m*s, a max speed of 53 m*;s when you have an very open structure and a one way flow).

I know I'am still a beginner and shouldn't expect wonders in this stage, but I wonder if someone know's what could be the problem?. Is it maybe my PC, or has it got to do that this is the trial version?.

I hope somebody can help, when it was working yesterday it really looked brilliant fur such an suffisticated and easy to use program. Regards, fhoos

Investigate your mesh

Have you tried examining the mesh metrics as described in "How do I fix mesh and solver failures?"? Have you tried using the Accuracy tool to improve the mesh?

I doubt there is a problem with your PC. The Caedium trial version is identical to the paid version.

If you are still having problems use our contact form to share your simulation (.sym) via a file sharing service, such as Dropbox.

Getting further to some results...


Many thanks for all your/these replies;

To help other users here is wat I found so far; It seems that my structure is placed with a narrow gab to a wall (this was intended to be so, because it is). The best result is to make this gab a bit wider, or remove the gab (you have to think wich situation is more representative to the real world situation).

Still sometimes there occur some smallproblems but the main thing I get it to work.

Kind regards, fhoos

Use the Accuracy tool

If the small gap is reflective of the actual physics you want to model then you should keep it as is. Then apply the Accuracy tool to nearby faces or edges to improve the gap mesh (i.e., better resolve it) as demonstrated in the section "Specify the Accuracy" for the tutorial "Flow Over a CO2 Dragster".