Hi,

I am looking for information about comparison among commercial CFD codes such as CFX, STARCD, FLUENT, etc in terms of strengths, weakness, models, and methods. If anybody knows the site, please let me know.

Rgds,

Jihwan

If you do not find a specific site I reccomend that you try searching this forum (this is a standard answer to a standard question). Then you have to filter the information since it is of course subjective.

Good Luck

This is problem that every time any body put such question he always gets such reply

We need the answer not the just the standard answer, so pleaseif any body knows this thing lets us know

Regards SAM

I know that 2 or 3 years ago TsAGI, Russian leading institution in aerohydrodynamics, has made such comparison and issued a report on it. The report was not open for public but from people who've seen I know the main conclusion - all those codes are nearly the same, each has its own MINOR advantages and disadvantages in some specific fields.

If you need a detailed answer to your question specify the field you're interested in and ask people/companies who has experience in this field. If they're real professionals than they've tried different codes and chose the best one for this particular field (though their reasons can be something like "I like the default colors of the interface" . But as far as I know, all major companies use at least two codes - one for "everyday" simulations and the other one for verification.

varification of results by another code?

i am interested in turbomachinery application

As far as I know for turbomachinery many people prefer CFX.

There are pros and cons with every code. The advantages or disadvantages one code has over another can be of many different kinds: accuracy, compatibility with various formats (both geometry and mesh), speed, ergonomics, price, availability on different platforms, etc. etc.

Since each user, or company, rate these topics differently it is not possible to give a generic answer to this question. If you search the forum arcives you will se many different opinions for this particular reason. An aircraft engine company has completely different demands on a code than does, say, a chemical company wishing to simulate their processes. And both of these use their code differently from an electronics company working with component cooling.

The best, and really only, way to go is to submit to the various vendors reference test cases similar to the application you are interested in. Or you could ask for test licenses to run the cases yourself. The latter approach is probably the better one, but demands more of your time and effort.

It is not possible to simply state: this and this code is the best. There are simply to many different applications, even with in the field of turbomachinery. I am in that field myself and use several codes in my daily work, not just one.

Martin

Thanks

so we can conclude (as said by one of freind in this forum) that companies use atleast two codes one for daily work and other for comparision

Thanks

One of the best comments on this issue I would say.

But this wouldn't be too expensive to have this approach

Of course it is expensive. Not many companies can afford to have two commercial codes of which one is just for benchmarking. It would also imply you have people being specialist in both codes. You only do the benchmarking once for your application and then get on working with the proper code. It's all very nice to do continues benchmarking and all but at some point you have to design something and sell it. It's better to check your code with experiments in a similar field as your application (to avoid comparison of two 'bad' commercial codes). Lots of experiments and literature out there. To add my humble view to the codes issue: All commercial codes can do simple things well. They All commercial codes are more or less user-friendly (switching from one code to the other is somewhat hard as you get used to things) When things get difficult (geometry, multiphase and such) benchmarks/experiments become rare items and you'll have to rely on your judgement on what's bogus and what's not. No code is fool-prove.

Many thanks to all the people who left messages. I can say the comments above are the best ever I've read in this forum.

Jihwan

you r most welcome. Also we learn a lot when we see such question and valuable comments from cfd experts

I hope this practice will continue on this forum

Thanks to all

SAM

test...