[blin]

Hi all,

My company is in the selection of a CFD program that is on the lower end of the spectrum cost wise (so ANSYS, and CD-Adapco are out). We will be performing external aerodynamics only, something with k-e or spalart-allmaras modeling is preferred.

We are looking for a software that will be a good supplement to the design process, preferably with a built in pre-processor. We are looking at CFDesign, FlowEFD, FloWorks, Flow3D, NUMECA, Phoenics and OpenFOAM but don't know what are the advantages/disadvantages of each. More suggestions to software options are welcomed too.

Any suggestions?

[Lysistrata]

Do you want a variety of turbulence models as well?

[blin]

I think the S-A and k-e model will be used most of the time, so other types of models isn't really necessary

[CapSizer]

Something worth considering is the question of what is available in terms of local expertise and support. For example, if you have a strong group in the immediate vicinity working with OpenFOAM, it becomes quite a strong candidate, but if you don't have that, you would be in for quite a tough struggle. Modern communications are wonderful, and one can do much remotely, but there is a lot to be said for quickly calling somebody over to come and sit with you to sort a difficult problem out.

[Lysistrata]

Originally Posted by blin

I think the S-A and k-e model will be used most of the time, so other types of models isn't really necessary

If you are going to stray into the realms of voodoo that's a good idea. Fewer chickens die pointlessly that way.

[blin]

Originally Posted by CapSizer

Something worth considering is the question of what is available in terms of local expertise and support. For example, if you have a strong group in the immediate vicinity working with OpenFOAM, it becomes quite a strong candidate, but if you don't have that, you would be in for quite a tough struggle. Modern communications are wonderful, and one can do much remotely, but there is a lot to be said for quickly calling somebody over to come and sit with you to sort a difficult problem out.

There is not much local support for any of the software, but note taken regarding OpenFOAM. Thank you.

[gocarts]

Originally Posted by blin

More suggestions to software options are welcomed too

Maybe you'd also like to take a look at the Caedium RANS Flow add-on which uses a custom version of OpenFOAM and includes the S-A and k-e turbulence models.

Caedium Professional (includes RANS Flow) provides an easy-to-use GUI environment for geometry import/creation (not embedded in a CAD system), meshing, CFD solver setup and control, and post processing "at an unbelievable price." - phrase of the moment thanks to Apple

Full disclosure: You've probably guessed already that I represent Symscape the developer of Caedium.
OpenFOAM is a trademark of OpenCFD and is unaffiliated with Symscape.

[monkeytron]

Do you have an idea of what hardware you will be solving on?
I'm imagining you'll need something reasonably meaty to do external race-car aero. OpenFOAMs meshser - snappyHexMesh - will mesh in parallel which would be an advantage. I'm not too sure about the pre-processing abilities of the packages you mention, but I dont think they mesh in parallel. CapSizer is right about access to expertise wrt openfoam though...

[blin]

Thanks for the response guys. My previous experience is with GAMBIT/Fluent, how does OpenFOAM compare to that in terms of user difficulty? Also, does OpenFOAM come with a pre-processor that is good enough for complex meshing? Say, compared to GAMBIT.....

[derrekcooper]

could be me, but seems like you really need to decide what you need in a cfd solution.. honestly, turbulence models should be the least of your worries

1. where does our geom originate?
2. how robust is the mesher?
3. what OS are supported?
4. external flow- is hpc an option?
5. how about running a what-if scenario?
6. how can you compare one design vs the other?

who cares about turb models until the above is worked out...

[vegbird]

Once I find way to find the pressure on a body in a compressible air flow, you come to see me and don't need to seek any software.

[monkeytron]

Originally Posted by derrekcooper

could be me, but seems like you really need to decide what you need in a cfd solution.. honestly, turbulence models should be the least of your worries

1. where does our geom originate?
2. how robust is the mesher?
3. what OS are supported?
4. external flow- is hpc an option?
5. how about running a what-if scenario?
6. how can you compare one design vs the other?

who cares about turb models until the above is worked out...

You have to understand your turbulence model in order to understand what your meshing requirements are and what you need from a mesher. You also need to understand your turbulence model so that you can be confident the results you get by running a what if scenario and by comparing one design to another are not affected by your choice of model. Neglect turbulence models at your peril