Fluent Capability for solving Polyhedral Mesh

jsm · September 16, 2015, 11:43

Hello All,

I found an beta feature for polyhedral mesh generation in fluent-meshing 16.2 version. It looks like, we can get good quality than tetra mesh for complex geometries.

My question is, Is fluent capable to solve with polyhedral elements robustly? Anyone studied about results difference for tetra and polyhedral mesh.

Your comments are appreciated!

Antanas · September 16, 2015, 12:57

Quote:

Originally Posted by jsm

Hello All,

I found an beta feature for polyhedral mesh generation in fluent-meshing 16.2 version. It looks like, we can get good quality than tetra mesh for complex geometries.

My question is, Is fluent capable to solve with polyhedral elements robustly? Anyone studied about results difference for tetra and polyhedral mesh.

Your comments are appreciated!

If I am not mistaken, Fluent can convert tetra meshes to polyhedral and solve on 'em for a long time.

jsm · September 16, 2015, 13:32

Yes. It is available for long time. But I am not sure, solver and discretization equations can handle polyhedral elements and produce accurate results. Because of this, I hesitate to use polyhedral elements til now.

If someone did correlation study with tetra and polyhedral mesh, then it will be interesting to discuss.

arjun · September 16, 2015, 17:40

Quote:

Originally Posted by jsm

Yes. It is available for long time. But I am not sure, solver and discretization equations can handle polyhedral elements and produce accurate results. Because of this, I hesitate to use polyhedral elements til now.

If someone did correlation study with tetra and polyhedral mesh, then it will be interesting to discuss.

Do yo u know any code that can do accurate polyhedral treatment?

jsm · September 16, 2015, 22:12

I think, Star CCM+ has polyhedral element treatment. But I don't know about Fluent. That is what I would like to know.

Antanas · September 16, 2015, 22:18

Quote:

Originally Posted by jsm

I think, Star CCM+ has polyhedral element treatment. But I don't know about Fluent. That is what I would like to know.

From Fluent Manual:

Quote:

.
.
.
When in solution mode, Fluent allows you to simulate the following:

- 2D planar, 2D axisymmetric, 2D axisymmetric with swirl (rotationally symmetric), and 3D flows

- Flows on quadrilateral, triangular, hexahedral (brick), tetrahedral, wedge, pyramid, polyhedral, and mixed element meshes

- Steady-state or transient flows
.
.
.

Dou you have doubt that this is true?

LuckyTran · September 16, 2015, 23:45

As others have mentioned, Fluent has supported the ability to solve polyhedral meshes for a long time and even has built-in support to convert tetrahedral mesh to a polyhedral one.

One of the niceties of having an unstructured solver is that the solver doesn't care what the grid geometry is. Unfortunately, for the same reason, it's difficult to implement higher order discretization schemes and that's why you're typically limited to 1st or 2nd order upwind schemes. The only things that would affect the robustness are things like cell skewness or area-to-volume ratio, but those problems are general and apply to all grid types. In other words, it really doesn't matter whether your cell is hexahedral with 6 faces or 5 faces or 7 faces or 12 faces.

jsm · September 17, 2015, 09:26

Thank you for comments - Antanas & LuckyTran. I understood your points.

Have you used polyhedral elements and compared the results?

Antanas · September 17, 2015, 10:08

Quote:

Originally Posted by jsm

Have you used polyhedral elements ...?

Only in CFX

arjun · September 17, 2015, 11:22

Quote:

Originally Posted by jsm

I think, Star CCM+ has polyhedral element treatment. But I don't know about Fluent. That is what I would like to know.

Fluent can also treat polyhderal and results will be no worse than ccm. There is nothing special in ccm that fluent does not have for polyhedrals.

jsm · September 17, 2015, 11:23

Quote:

Originally Posted by Antanas

Only in CFX

How it is worked for you? Do you got accurate results than tetra mesh? And do you compared with test?

jsm · September 17, 2015, 11:29

Quote:

Originally Posted by arjun

Fluent can also treat polyhderal and results will be no worse than ccm. There is nothing special in ccm that fluent does not have for polyhedrals.

I never had chance to compare Fluent vs ccm+ with polyhedral mesh. But I should try polyhedral mesh in fluent and compare with test data.

Bharatt · September 17, 2015, 11:55

Quote:

Originally Posted by jsm

I think, Star CCM+ has polyhedral element treatment. But I don't know about Fluent. That is what I would like to know.

Yes, Star CCM+ has polyhederal element treatment in it. in ccm the mesh elements it have are trimmer and polyhederal and the convergence time for poly mesh is less when compared to trimmer and with somewhat better flow results. but its still a debatable topic, with type of CFD problems one is dealing.

Regards,
Bharat

September 16, 2015, 11:43	Fluent Capability for solving Polyhedral Mesh	#1
jsm Senior Member JSM Join Date: Mar 2009 Location: India Posts: 192 Rep Power: 20	Hello All, I found an beta feature for polyhedral mesh generation in fluent-meshing 16.2 version. It looks like, we can get good quality than tetra mesh for complex geometries. My question is, Is fluent capable to solve with polyhedral elements robustly? Anyone studied about results difference for tetra and polyhedral mesh. Your comments are appreciated! __________________ With regards, JSM

September 16, 2015, 13:32		#3
jsm Senior Member JSM Join Date: Mar 2009 Location: India Posts: 192 Rep Power: 20	Yes. It is available for long time. But I am not sure, solver and discretization equations can handle polyhedral elements and produce accurate results. Because of this, I hesitate to use polyhedral elements til now. If someone did correlation study with tetra and polyhedral mesh, then it will be interesting to discuss. __________________ With regards, JSM

September 16, 2015, 22:12		#5
jsm Senior Member JSM Join Date: Mar 2009 Location: India Posts: 192 Rep Power: 20	I think, Star CCM+ has polyhedral element treatment. But I don't know about Fluent. That is what I would like to know. __________________ With regards, JSM

September 17, 2015, 09:26		#8
jsm Senior Member JSM Join Date: Mar 2009 Location: India Posts: 192 Rep Power: 20	Thank you for comments - Antanas & LuckyTran. I understood your points. Have you used polyhedral elements and compared the results? __________________ With regards, JSM

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
how to set periodic boundary conditions	Ganesh	FLUENT	15	November 18, 2020 06:09
Star CCM Overset Mesh Error (Rotating Turbine)	thezack	Siemens	7	October 12, 2016 11:14
[snappyHexMesh] No layers in a small gap	bobburnquist	OpenFOAM Meshing & Mesh Conversion	6	August 26, 2015 09:38
Polyhedral mesh has many more celle than standard mesh	klausb	enGrid	2	March 16, 2015 08:01
Polyhedral mesh in Fluent	Shri	FLUENT	3	August 22, 2007 04:31

September 16, 2015, 23:45		#7
LuckyTran Senior Member Lucky Join Date: Apr 2011 Location: Orlando, FL USA Posts: 5,674 Rep Power: 65	As others have mentioned, Fluent has supported the ability to solve polyhedral meshes for a long time and even has built-in support to convert tetrahedral mesh to a polyhedral one. One of the niceties of having an unstructured solver is that the solver doesn't care what the grid geometry is. Unfortunately, for the same reason, it's difficult to implement higher order discretization schemes and that's why you're typically limited to 1st or 2nd order upwind schemes. The only things that would affect the robustness are things like cell skewness or area-to-volume ratio, but those problems are general and apply to all grid types. In other words, it really doesn't matter whether your cell is hexahedral with 6 faces or 5 faces or 7 faces or 12 faces.