Hi, I'm new learner in Gambit & Fluent.

Can anyone tell me where can i get Journal file examples of Gambit & Fluent for my practise especially that Online.

Thanks in advance

Why don't you use the gambit and fluent gui's/tui's to build models and then look at the journal files produced. This way you can see what gambit/fluent produce depending upon your actions.

In gambit they are .jou files. In fluent you need to record your actions from the file menu. I find its better to use the text interface rather than the gui in fluent for recording journal commands.

Greg

I had done all tutorials in Gambit(tut's 1~6).

Actually, what I want to learn is how to generate mesh for complex/shopiscated geometry using Hexahedral cell.

Therefore, can anyone share their Jou files with me especially those who had succeed to generate hexahedral mesh for complex geometry by Gambit.

I'll appreciate a lot.

If you want to make hex meshes in complex geometries, you have a couple of options which are more viable than the tools in Gambit:

1. Use a block structured mesher like Gridpro or Gridgen. You will need to invest more time up front in decomposing your geometry into blocks, but you will have much better control on the meshing process and a better chance of getting a mesh at all on some geometries.

2. Use a volumetric unstructured mesher like samm/pro*am from adapco. You will get some N-faced cells where the algorithm "trims" the otherwise perfect hex cells to fit the geometry, but you can easily add boundary layers to get a very high quality mesh in much less time than the structured approach.

Good luck,

Calvin

Calvin do you have any comments on Tgrid.

I'm thinking of trying it for complex 3d geometries where I want to have a hex boundary layer and can live with a tet mesh far from the boundaries.

Greg

Greg,

I have used TGrid, and it's a very nice tetrahedral mesher with the option to create prismatic boundary layers. TGrid builds volume meshes based on the surface mesh you provide, so if you want hexas in the boundary layer, you will need to provide a surface mesh consisting of quadrilaterals. (This is not so easy to do, and can't be done at all within TGrid.) If you can live with triangular based prisms, TGrid will be able to do this, with a couple of caveats:

1. The boundary layers are generated by an advancing front method. There is no recognition of proximity, so if your surfaces get closer to each other than the combined thickness of the boundary layers you are trying to build, the volume mesh generation will fail.

2. The boundary layers you generate from a triangulated surface will be prisms, and given that all of the important gradients are normal to the wall, you will end up putting a lot more total cells near the wall than you would have to if you used hexas, which have the additional advantage of being able to be stretched much more along the wall than prisms/tets can be.

I know this is the Fluent forum, but if what you really want to do is create unstructured meshes with mainly hexahedra in the boundary layer, you really ought to have a look at pro*am from CD/adapco. I think the agent for them in Oz is Peter Ewing at Orbital Engine Co. in Perth, WA.

Good luck,

Calvin

Thanks Calvin,

I read about your points in the manual. I started in gambit and then realised that its not really that good for complex geometries. But I am thinking of using it to generate the surface meshes with a pave hex algorithm.

My surfaces are complex, since they move as a function of the solution. So I've been writing some C programs to generate gambit journal files.