CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Fluent-wrong cl value at high angle of attack (http://www.cfd-online.com/Forums/fluent/50620-fluent-wrong-cl-value-high-angle-attack.html)

Lucas February 19, 2009 19:39

Fluent-wrong cl value at high angle of attack
 
Hi, I trying to validate a flow around an airfoil in fluent in respect with experimental data. At low angle of attack it's ok, but at about 8 degrees I start do get cl values very diferent (lower) from the experimental ones. Well,I using SST with transitional correction (it's a flow around an airfoil similar to Selig 1223 at Reynoldy Number 300.000). The inlet condiction is I=0.0034 and viscosity ratio 10.That are good inital guesses? Could you tell me how to solver this problem? It's necessary to correcty some data to diferents values from that used at lower angle of attack?

Sorry by the lack of knoledge,I'm student and I learning to use fluent just for 2 months.

Pablo February 20, 2009 08:14

Re: Fluent-wrong cl value at high angle of attack
 
look at the reference values. Re 3e5 it's a very low Re number, you get laminar to turbulent transitional bubble. Y+?

Alagesanj February 20, 2009 09:49

Re: Fluent-wrong cl value at high angle of attack
 
Hi Pablo,

I am also simulating a flow around E387 airfoil for Re = 1e5 and 5e5. My Y+ < 1. when I use full turbulence model the drag results are very high. So I used forced transition by splitting the mesh into laminar and turbulence zone. This case the results are ok but my solutions are not convreging above alfa 8 degree.

I used another method, with full turbulence SST K -w model and by adjusting the k and omega source terms (trial and error) in fluid zone. By this I can achieve the results, able to capture the LSB. But this is not always possible because every alfa and Re i have to find a value for k and omega. I tried to write a UDF but failed,since not familiat with C++.

Is there any other method which i can try. I know i need to couple a transition method like e^n transition criterion to fluent. But I do not know how to do it. If you can provide some help it will be very helpful.

My time is running out. Next friday (27th Feb)I need to submit the report.

Any help will be appreciated.

cheers Alagesan


Pablo February 20, 2009 10:16

Re: Fluent-wrong cl value at high angle of attack
 
mmmm, now i'm feeling old.

look at the previous topics. I wanted to do the same three or four years ago.

What I found it was a turbulence model which calculates the divergence of the reynolds stress tensor instead the tensor itself. This turbulence model was developed by Pr. Perot from UMASS, I contacted him and I got the necessary files to code out the model and incorporate it into OpenFOAM. This turbulence model is able to predict the laminar to turbulent transition. Otherwise you should use the v2f turbulence model.

However, introducing the envelop method e^n into FLUENT would be quite useful to do research on low reynolds number aerodynamics, and of course is a beautifully subject of research.

may the force be with you.


Alagesanj February 20, 2009 10:40

Re: Fluent-wrong cl value at high angle of attack
 
Hi Pablo,

Thanks for your reply.

V2f turbulence model is not available. So I have to use SA, ke or kw models only. Is there any suggestion for me regarding for my problem.

I have seen your previous post about OpenFOAM and visited the site few days back. Currently i dont have time to use it. May be next week i'll check the code since I am finishing my study i have to spend my time wisely. I ve already decided to do some thing for the open source community.

cheers, Alagesan.

Carlos February 21, 2009 22:11

Re: Fluent-wrong cl value at high angle of attack
 
Hi,

There may be a few problems. Firstly, I would use the "turbulent length scale and intensity" specification method for both the inlet and outlet. Set the turbulence intensity to 0.01% (This is a typical experimental value for NACA wind tunnel) and the length scale is 0.07*chord length (derived from duct flows).

This gives the solver a better start. Make sure you are letting the solution fully converge, aerofoils typically need 3-5000 depending on angle of attack. Convergence is achieved when the residuals are flat, or oscillating about a constant value.

Ensure that you are using a higher discretization scheme than the default of first order. Pressure should be 2nd order, momentum QUICK (third order) and nut QUICK - all found under solution->controls. You MUST do this for accurate results.

Print drag and lift after the simulation then work out cl and cd that way.

Carlos.

Lucas February 22, 2009 15:37

Re: Fluent-wrong cl value at high angle of attack
 
Hi, I'm using Y+ less then 1. It's really a transitional situation, so I think that it is the hardest problem I would predict on CFD.So, the mesh of the traling edge of the airfoil may have a great contribution on this error,it is not?I will try to improve it,but I think that it will not resolve the problem


Lucas February 22, 2009 15:44

Re: Fluent-wrong cl value at high angle of attack
 
Thanks Carlos,

I just see your post now.I will try this and then post here my results

Bye

aqib December 1, 2010 07:18

hi
i am also new to fluent and using the SST transition model at high angle of attack 37.7, and controlling the separation on Cascade T106A. My solutiuon is not converging and i am running the steady case.
Please give me the solution as i am new to fluent.
Regards:
Aqib


All times are GMT -4. The time now is 02:13.