hi! I tried to compare the value of the wall shear stress on the walll given by fluent with three differents methodes I have done (UDF) and i have found that the value given by fluent is not correct (less than 10% with my methodes). Is there some explanation ? Thank you in advance! lagha

Dear Lagha

Here, I do not mean that the result of CFD is very good or bad. And I do not mean that your method is good or bad.

Couple of tens years, I do CFD simulation. And I have heard many many times that 'CFD results are fairly different from those of experimental ones'. Sometimes, the result of Fluent is different from that of another CFD program.

Here, I would like to say that, as far as my experience is concerned, experiment' has also its limitation. For example, there is deep discussion for the measurement of velocity and/or pressure. Some of new technologies, such as PIV(Particle Image Velocimetry), PSP(Pressure Sensitive Paint), PTV(Particle Tracking Velocimetry), which are very expensive, say that these technologies are very good. Then, in other words, conventional cheaf technology, for example, manometer measurement, is not accurate ?

So, I would like to say that CFD might be very good with high investment, including couple of tens million meshes, supercomputers, long period and so on. Who can say that CFD is good or bad, and experiment is good or bad, considering the chaotic behavior of turbulent fluctuation ?

Here I mean that most experiment has its own limitation like CFD, too. So, your another method(experiment or another analysis) might be very good or poor.

CFD is predicition. And predicition is only prediction. However, as far as my(and my team's) application experience is concerned, CFD is very powerful prediction tool. Which method can predict the detail behavior of fluid flow except CFD ? For your reference, my job is CFD consultant(that is to say, I earn my salary by only CFD projects).

As I said above, I do not mean that CFD is very beautiful. But 10% difference might not be serious problem if your CFD result has the qualitative agreement with experiment or another method.

I hope that this is helpful to you for your decision making. And, I would like to hear the opinion of CFD experts.

Sincerely, Jinwook

Dear Jinwook, thank you very much for your reponse! In fact, i know that CFD is only prediction . Bur I just want to know which method does fluent use to predict the value of the wall shear stress? That's why , I tested three method using some macro like BOUNDARY_FACE_GEOMETRY(...) which give me the norml vector to the wall face, after that I use the velocitu derivative in the cell next to this face... If you want, I will explain you in more details the other methods!

Thank you again for you response. Lagha