CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   comparing simulation-experimental values for Cl,Cd (http://www.cfd-online.com/Forums/fluent/28539-comparing-simulation-experimental-values-cl-cd.html)

Rodrigo Madariaga May 18, 2001 18:33

comparing simulation-experimental values for Cl,Cd
 
I am working on a NACA 0012 airfoil.

When comparing the values for the drag and lift coeficient vs. angle of attack with the experimental data from report 824 from NACA ( The 0012 is not available but 0006 and 0009 are similar), my airfoil does'nt stall (loss of lift capacity due to separation of the boundry layer)

First I thought it was because the mach number was'nt the same, I changed it so the reynolds number matched, still the same problem.

Also, I calculate the Reynolds number and after the simulation, Fluent returns Reynolds that are much smaller.

Does anyone know what I`m doing wrong?

John C. Chien May 18, 2001 19:46

Re: comparing simulation-experimental values for C
 
(1). You have two options: (a). forget about the test data, or (b). forget about the cfd solution. (2). Unless you have a paper which states that they should be the same. I strongly urge you to look for such paper or report first.

Daniel Bruno May 21, 2001 19:36

Re: comparing simulation-experimental values for C
 
Dr Chien,

You say to forget about CFD or about experimental. Why ? I am working hard on numerical simulation of airflow around a Selig 1223 airfoil with Re=171000. The results I got until now are not matching exactly the experiments, but are something ok. I tried modifying the turbulence and near-wall models. And I'll continue trying with another models and using a structured mesh instead. I have the wish to work on CFD codes implementation. Is that a distant truth or can I go trying and will reach some place ?

PS: Rodrigo, are you from Brasil ?

John C. Chien May 21, 2001 21:54

Re: comparing simulation-experimental values for C
 
(1). It is well known that standard k-epsilon two-equation turbulence model is more diffusive. For this reason, the presidcted size of a separation bubble, such as flow over a backward facing step, is always smaller than the test data. (2). So, it is consistent with your solutions which under-predict the flow separation. (3). The way to improve the solution is to improve the turbulence model. This can be done, if you think the test data is the right answer. (4). If you just run the code and ask why they don't match. Then you are creating problems. They are not suppose to match, unless you improve your turbulence model.

John C. Chien May 21, 2001 21:56

some typing errors
 
It should be " For this reason,the predicted size of a separation bubble,"

Rodrigo Madariaga May 31, 2001 04:48

Re: comparing simulation-experimental values for C
 
Daniel:

My test data comes from the NACA reports (now its NASA), they have run wind tunnel testing on all their airfoils, but they were made 50 years ago. I found other ones that, with the same Reynolds, have diferent results.

I'm a engg. Student from Chile, my University has being working on a airplane built here (Chile), the T-35 Pillan, The results that Fluent has given us are very good, specially the Spallart-Almaras model. This tells us that we are on the right track.


All times are GMT -4. The time now is 01:11.