[Dragonuser]

Hello everyone, I had met some problems in modeling subcool boiling in a single tube.
Here is the brief introduction of my problems:
When I use the first set of mesh, I could get a convergence result. The result got good agreement with the experiment result.
But as I changed to the second set of mesh, I got overflow error.
The only difference between the two simulations is the mesh difference, and I really don’t know why I got overflow error by using the second set of mesh. Because in my opinion, the second set of mesh seems to have much higher quality than the first type mesh, the second type mesh's result should be better than the first mesh type.

[huisenanhai]

Did you check mesh before simulation？

[ghorrocks]

Finer meshes are less numerically stable than coarse meshes. That could explain it as well. Was the second mesh finer?

[nyyrzh]

sorrry the Internet is too slow here, i tried to upload the pics many times ,finally it got work,
the pics of two meshs are as follow:
first set mesh.png
first mesh
second set mesh.jpg
second mesh

[nyyrzh]

BTW the first new thread was posted by my friend Dragonuser, because at that time my Internet speed is two bad and i can't post a new thread, so i asked my friend for help.
And my problem is :
the first mesh has the 0 quality elements in the center point of the 1/4 circle, but it can get the reasonable result
the second mesh quality is higher than 0.5, but it will get over error.
Thank every one for your notice and help.

[ghorrocks]

The first mesh is generally coarser than the second. It could be simply your mesh density.

Have you looked at the FAQ for overflow errors: http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

[nyyrzh]

Hello ghorrocks , thank you for your reply .
But I still have trouble in understanding your suggestion after I read the FAQ.
As FAQ showed the multiphase flow simulation should have higher mesh quality than the single phase. However in my condition, the first mesh has lower quality than the second mesh, but strangely the first mesh can get convergence and reasonable result. While the second mesh would get overflow error as a result. Besides, the first mesh has 0 quality element.
And I don’t know why the 0 quality mesh could still get a convergence result, I was confused by the act.
thank you again for you help.

[ghorrocks]

I have said why I think the first mesh converged and the second did not two times on this thread already. I will try for the third time:

The first mesh is coarser than the second. Coarser meshes are often numerically more stable and therefore converge easier. In your case this appears to be a more significant effect than the element quality.

[nyyrzh]

sorry I didn't notice it before, and I finally understood.
Thank you for your suggestion.

[diamondx]

The two mesh are supposed to give approximatly same results. The second one is better of course. internal o-grid should always be present in circular geometry. That being said, it may sound stupid, but you should check the mesh's dimension, when importing your mesh, check if the dimensions are correct. The setting of your boundary condition may be perfect, but if the scaling of the mesh is not, you'll have problems. This is so important that I have post-it on my screen saying "check mesh dimension !!!" as I spent one afternoon trying to figure out the problem in one of my simulations, I felt so stupid later...