CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   GridPro (https://www.cfd-online.com/Forums/gridpro/)
-   -   GridPro (https://www.cfd-online.com/Forums/gridpro/12510-gridpro.html)

Vinod Dhiman November 6, 2006 03:12

GridPro
 
Hi All

Has anyone ever used GridPro? If anyone has used I'd like to have a review regarding this. We are using ICEM CFD and Gambit right now. GridPro people came to us and we'll have a Demo Version for a month. However, if someone can give a review that'll be very helpful.

With Thanks Vinod Dhiman

myron November 6, 2006 08:48

Re: GridPro
 
I used GridPro several years ago - and things have surely changed some since then.

I can speak only on Hex meshing.

Both GridPro and ICEM require a "blocking topology" to describe the mesh. Gambit uses a "sweeping" technology with their Cooper tool. My opinion is you can get much more complex grids completed with the blocking topology.

ICEM topology is typically easier to build - but the mesh will be much smoother from GridPro. In most cases the "smoothness" of the grids does not really impact the solution that much - but it does look nice. I'm sure there are some cases where orthogonality to the wall is required, etc - and ICEM does have some smoothing, but it's a pain to use.

With ICEM you also have the Ansys brand - which could be good or bad depending on whom you ask. But you also get some very powerful interfaces to CAD packages with ICEM. Integration with Fluent, CFX.

For grids themselves comes down to preference - trade ease of the topology creation for smoothness of the grids...

Simply my opinion....

Mani November 6, 2006 09:26

Re: GridPro
 
One advice on testing Gridpro: Make sure to try it out on one of your own problems. GridPro gets really impressive (very smooth looking) grids on all their published examples and tutorials. However, even if you have a relatively simple topology (in my case a fan), and you can apply a block-topology similar to one of their examples, it may take a lot of tricks to get it to converge on your geometry. In general I encountered problems when the geometry had a large range of scales, e.g. a small tip clearance region on a high aspect-ratio fan, or even just a very thin airfoil with large chord and sharp leading and trailing edges. It's not so much the grid generation that's giving you trouble than the iterative process in which Gridpro is trying to map block corners onto boundaries. If all works well, this process is very effective, but often enough it does not work as in the tutorials. My advice: Test in on your geometry.

QCFD March 24, 2015 13:44

GridPro has a steeper learning curve than a software like ICEM for block structured meshing, but the power of the smoother is well beyond what I have been able to accomplish in ICEM.

As mentioned above, for a lot of general purpose CFD, the buttery smoothness of a GridPro mesh might not impact your final results a lot (If you follow good practice with the unsmoothed mesh, volume transitions, skew...) However, if you run LES it can really help. Additionally, I tend to get lower residuals and convergence in fewer iterations when using a high quality smoothed GridPro mesh than an unsmoothed mesh (maybe out of ICEM) for simple RANS analysis as well.

If you only expect to mesh occasionally and need quick turnaround, GridPro may not be all that practical, but for me it a valuable tool... and the meshes just look beautiful


All times are GMT -4. The time now is 20:28.