CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Pointwise & Gridgen (https://www.cfd-online.com/Forums/pointwise/)
-   -   Non-positive volume detected cells in FLUENT, blocks orientation (https://www.cfd-online.com/Forums/pointwise/233622-non-positive-volume-detected-cells-fluent-blocks-orientation.html)

EleGiova February 4, 2021 10:21

Non-positive volume detected cells in FLUENT, blocks orientation
 
Dear all,

I have done a 3D hybrid propeller (only 1/4 of the geometry) mesh with Pointwise (the rotating part structured, outer box unstructured, 250.000.000 cells) and I am having some troubles with the hybrid inizialization with Ansys FLUENT.
I can import the mesh and then do the setup (MRF, not sliding mesh), but when I begin the hybrid inizialization it doesn't go on. I have a warning message on the console "1138109 cells with non-positive volume detected", and I think the problem is the orientation of the blocks in Pointwise.
I checked that and I have a right-hand rule for the orientation, but maybe there are other rules that I don't know of.

Can you help me ?
Thank you,
E.G

pdp.aero February 4, 2021 13:46

Quote:

Originally Posted by EleGiova (Post 795283)
Dear all,

I have done a 3D hybrid propeller (only 1/4 of the geometry) mesh with Pointwise (the rotating part structured, outer box unstructured, 250.000.000 cells) and I am having some troubles with the hybrid inizialization with Ansys FLUENT.
I can import the mesh and then do the setup (MRF, not sliding mesh), but when I begin the hybrid inizialization it doesn't go on. I have a warning message on the console "1138109 cells with non-positive volume detected", and I think the problem is the orientation of the blocks in Pointwise.
I checked that and I have a right-hand rule for the orientation, but maybe there are other rules that I don't know of.

Can you help me ?
Thank you,
E.G

Hi there,

yes, I assume it is because of the orientation.

Here is the rule I have learned back then when I was running with 6.3.26.

For structured grid you may use your right hand’s fingers and rotate them following i and j directions on boundaries then your thumb must show the k direction. Another way that I like very much is to imagine yourself walking on boundaries of the mesh following the i and j directions then your right hand must be inside the domain.

For unstructured grid you'll see the normal direction.

EleGiova February 4, 2021 13:58

But do I have to check the orientation of all domains or all blocks? Or both?



And in my case the reference system has y "vertical", in this case do I have to orient the j with y, or always k? (Talking about blocks)

pdp.aero February 4, 2021 14:59

Quote:

Originally Posted by EleGiova (Post 795306)
But do I have to check the orientation of all domains or all blocks? Or both?

In 2D you set the orient by setting the right direction for i, j and in 3D you set the orient by setting the i,j,k.

In unstructured grid you must set the normal direction.

Quote:

Originally Posted by EleGiova (Post 795306)
And in my case the reference system has y "vertical", in this case do I have to orient the j with y, or always k? (Talking about blocks)

As far as I know orient of the mesh doesn't depends on the axis direction in fluent meaning no matter how you rotate your mesh in your grid generator when you have the right orient i,j,k directions or normal everything is fine. In some other solvers like SU2 orient is fixed by default meaning is being determined based on the axis direction and you should always have the right axis directions for your mesh to get the right orient.

Good luck,
Pay

dgarlisch February 4, 2021 19:18

Select the structured blocks and use Edit, Orient... to view and change each blocks orientation. The table at the top of the panel lists the number of right hand and left hand blocks. Reorient any left hand blocks to make them right hand.

Repeat for the unstructered blocks.

The orientation of the individual domains is irrelevant in a 3D mesh.


All times are GMT -4. The time now is 01:29.