[Mixo123]

Hello, dear CFD experts!

I am in a process of writing scope statement for CFD software, but I have extremely vague knowledge in CFD.

I want to ask you: what are the most important features for almost-general-purpose CFD software? This particular program will be mainly focused on nonlinear waves, turbulence and in providing high detalization.

What are important CFD features, missing in common CFD software like ANSYS? And what features are "must have"?

I would also be glad to hear some advices and thoughts: about models, recommended methods, etc.

Thank you very much!

p.s.: Sorry for overly vague question.

[cfdnewbie]

Originally Posted by Mixo123

Hello, dear CFD experts!
This particular program will be mainly focused on nonlinear waves, turbulence and in providing high detalization.

What are important CFD features, missing in common CFD software like ANSYS? And what features are "must have"?

if turbulence resolution and wave interaction is what you are looking for, one feature absent in codes like ansys and co is a high order discretization with low dispersion and dissipation. Another issue is parallel scaling, which is also sadly lacking in commercial codes, at least above 100 or 1000 cores...

[Mixo123]

Ok, I see.
May you recommend some techniques or methods for achieving this?

Some of our programmers want to use SPH, but I am not sure that is smart for our task.

[julien.decharentenay]

Hi Michail,

Sorry to be a bit blunt. But can you please state your affiliation?

There has been instances where someone started a discussion and then used the contents (and ideas) in a non-free report.

Julien

[Mixo123]

Oh, I am undergraduate student in Russia. Small team of students here has an assignment in CFD (despite we never had hydrodynamics course).
Our assigment is not of hard ones, but suddenly we became interested in CFD on our own.

[FMDenaro]

Originally Posted by Mixo123

Oh, I am undergraduate student in Russia. Small team of students here has an assignment in CFD (despite we never had hydrodynamics course).
Our assigment is not of hard ones, but suddenly we became interested in CFD on our own.

I suggest to read the introduction of some CFD books ... it should be sufficient for your aims

[Mixo123]

Originally Posted by cfdnewbie

if turbulence resolution and wave interaction is what you are looking for, one feature absent in codes like ansys and co is a high order discretization with low dispersion and dissipation. Another issue is parallel scaling, which is also sadly lacking in commercial codes, at least above 100 or 1000 cores...

By the way, what are scales, on which those effects appear?

[leflix]

Originally Posted by Mixo123

Ok, I see.
May you recommend some techniques or methods for achieving this?

Some of our programmers want to use SPH, but I am not sure that is smart for our task.

Yes I completely agree with you !! I'm not convinced at all that SPH could fulfil the requirements needed for a state-of-the-art CFD solver in 2012 !! lol
Use FVM on arbirary control volume shape, or at least hexaedral block structured grids.

[JoshS]

Originally Posted by Mixo123

Oh, I am undergraduate student in Russia. Small team of students here has an assignment in CFD (despite we never had hydrodynamics course).
Our assigment is not of hard ones, but suddenly we became interested in CFD on our own.

Mixo123 the place to start is Wilcox's Turbulence Modeling for CFD.