Help! Validity of unstructured grid on S809 airfoil

 Register Blogs Members List Search Today's Posts Mark Forums Read

 March 28, 2011, 08:33 Help! Validity of unstructured grid on S809 airfoil #1 Member   Felix Join Date: Mar 2011 Posts: 45 Rep Power: 6 Hello, everyone~ I want to simulate the influence of motion and deformation of an airfoil on the flow field using dynamic mesh, so the unstructured grid was adopted. Before the simulation, I try to simulate the static airfoil using unstructured grid. Unfortunately, the lift coefficients are very different from the experiment data. BUT results using the structured grid are almost same with experiment data. I am very confused. I have built a model with all unstructured grids and another model with boundary layer together with unstructured grids. When the angle of attack locates between 0degree and 10degree, results of the two models are similar, but quit different from those of the structured grids and experiment. Why does such kind of thing happen? Would anyone be so kind to explain for me? Or, has anyone ever simulated the airfoil using unstructured grid and got good results? My English is not good, hope I have described my question clearly and any comment is welcome! Thank you!

 March 29, 2011, 12:30 Would anybody be so kind to help me? #2 Member   Felix Join Date: Mar 2011 Posts: 45 Rep Power: 6 It seems that arranging the grid of an 2D airfoil using structured boundary grid (nearly 2% chord high ) together with unstructured grid outside cannot simulate the lift coefficients well. Does anyone know the reason? Pleeease help me, it is driving me mad because I can make no progress without solve this problem.

March 30, 2011, 02:00
#3
Senior Member

Raashid Baig
Join Date: Mar 2010
Location: Bangalore, India
Posts: 136
Rep Power: 7
Quote:
 Originally Posted by didiean It seems that arranging the grid of an 2D airfoil using structured boundary grid (nearly 2% chord high ) together with unstructured grid outside cannot simulate the lift coefficients well. Does anyone know the reason? Pleeease help me, it is driving me mad because I can make no progress without solve this problem.
Hi Joshua,
Take a look at the following paper -
"Predicting 2D Airfoil and 3D Wind Turbine Rotor Performance using a Transition Model for General CFD Codes",
R. Langtry, J. Gola and F. Menter, ANSYS CFX, Otterfing, Germany, AIAA-2006-0395 44th AIAA Aerospace Sciences Meeting and Exhibit

and see the forum thread - Two-equation turbulent models: low re airfoils,

Raashid

March 30, 2011, 02:50
#4
Member

Felix
Join Date: Mar 2011
Posts: 45
Rep Power: 6
Quote:
 Originally Posted by cfd_newbie Hi Joshua, Take a look at the following paper - "Predicting 2D Airfoil and 3D Wind Turbine Rotor Performance using a Transition Model for General CFD Codes", R. Langtry, J. Gola and F. Menter, ANSYS CFX, Otterfing, Germany, AIAA-2006-0395 44th AIAA Aerospace Sciences Meeting and Exhibit and see the forum thread - Two-equation turbulent models: low re airfoils, these 2 resources will help you understand the problem more. If you want I am willing to help you with this. Raashid
Hi, Raashid
What is troubling me most is that the structured grid can simulate the lift coefficients between 0~8degree quite well, but the results of unstructured grid, with structured boundary layer grid near the airfoil,are very different, as more as 20% different to those of the experiment. But I don't know why this happens. Does it mean that the CFD using unstructured grid cannot predict the lift coefficients of the 2D airfoil?

March 30, 2011, 03:05
#5
Senior Member

Raashid Baig
Join Date: Mar 2010
Location: Bangalore, India
Posts: 136
Rep Power: 7
Quote:
 Originally Posted by didiean Hi, Raashid Thanks a lot for your reply! What is troubling me most is that the structured grid can simulate the lift coefficients between 0~8degree quite well, but the results of unstructured grid, with structured boundary layer grid near the airfoil,are very different, as more as 20% different to those of the experiment. But I don't know why this happens. Does it mean that the CFD using unstructured grid cannot predict the lift coefficients of the 2D airfoil?
Hi,
Good to know that you are getting correct result with structured grid.
If you are getting incorrect results with unstructured grid that your grids are not properly made. People have used unstructured grids in much more difficult problems than this so you cannot say that unstructured grids are bad.

See the following thread for grid generation discussion - Meshing a 2D airfoil properly

Also look at the following paper for a S809 grid generation study -
http://wwweng.uwyo.edu/mechanical/fa...-2009-1221-908

March 30, 2011, 04:01
#6
Member

Felix
Join Date: Mar 2011
Posts: 45
Rep Power: 6
Quote:
 Originally Posted by cfd_newbie Hi, Good to know that you are getting correct result with structured grid. If you are getting incorrect results with unstructured grid that your grids are not properly made. People have used unstructured grids in much more difficult problems than this so you cannot say that unstructured grids are bad. See the following thread for grid generation discussion - Meshing a 2D airfoil properly Also look at the following paper for a S809 grid generation study - http://wwweng.uwyo.edu/mechanical/fa...-2009-1221-908
The reason I choose unstructured grid outside the structured boundary layer grids is that I want to simulate influence of flexible airfoil. As you know, using fluent to simulate an object having large deflection and deformation, only the dynamic mesh method is available. AND remeshing method should be used which are only available for triangular grid.
Before the unsteady calculation, I want to test the availability of the grids, so I simulate the original S809 under Re=1X10^6, but the results are beyond my hope. And I don't know whether I should trust it.

 March 31, 2011, 10:35 #7 Member   L.Y LIN Join Date: Mar 2010 Posts: 47 Rep Power: 7 are you sure your code is right? did you do other benchmark tests?

April 1, 2011, 00:22
#8
Member

Felix
Join Date: Mar 2011
Posts: 45
Rep Power: 6
I use fluent to do the simulation, and SIMPLE method is adopted by default.
I try to simulate flow field around an airfoil, and the mixture of structured grids around the airfoil (y+<1) and the unstructured grids outside is used. But the lift coefficients are 20% lower than those of the experiment.
If I use only structured grid, even just part of the mixture grid mentioned above, the results are better.

Quote:
 Originally Posted by iilw1314 are you sure your code is right? did you do other benchmark tests?

April 1, 2011, 08:37
#9
Member

L.Y LIN
Join Date: Mar 2010
Posts: 47
Rep Power: 7
Quote:
 Originally Posted by didiean Hello, everyone~ I want to simulate the influence of motion and deformation of an airfoil on the flow field using dynamic mesh, so the unstructured grid was adopted. Before the simulation, I try to simulate the static airfoil using unstructured grid. Unfortunately, the lift coefficients are very different from the experiment data. BUT results using the structured grid are almost same with experiment data. I am very confused. I have built a model with all unstructured grids and another model with boundary layer together with unstructured grids. When the angle of attack locates between 0degree and 10degree, results of the two models are similar, but quit different from those of the structured grids and experiment. Why does such kind of thing happen? Would anyone be so kind to explain for me? Or, has anyone ever simulated the airfoil using unstructured grid and got good results? My English is not good, hope I have described my question clearly and any comment is welcome! Thank you!
well,i have little experiences with fluent.But i am sure you must make some mistakes somewhere since the technique with unstructured grids for 2D airfoil is very mature and it would not produce such a wrong result ,as you'v said it is 20% away from the experiment .Maybe you should check your actions on every step carefully ,you must have made some mistakes such as interpolation to the points on the airfoil surface.
bless.

April 2, 2011, 12:26
#10
Member

Felix
Join Date: Mar 2011
Posts: 45
Rep Power: 6
Quote:
 Originally Posted by iilw1314 well,i have little experiences with fluent.But i am sure you must make some mistakes somewhere since the technique with unstructured grids for 2D airfoil is very mature and it would not produce such a wrong result ,as you'v said it is 20% away from the experiment .Maybe you should check your actions on every step carefully ,you must have made some mistakes such as interpolation to the points on the airfoil surface. bless.
Well, for the structured grid and the mixture grid (structured + unstructured), I use the same journal file of fluent, so all their sets are the same. Maybe, for the triangular grids, high order discretized method should be used in order to reduce the numerical viscosity. I will try to test it! Do you think my guess is possible?

April 3, 2011, 01:06
#11
Member

L.Y LIN
Join Date: Mar 2010
Posts: 47
Rep Power: 7
Quote:
 Originally Posted by didiean Thanks for your reply~ Well, for the structured grid and the mixture grid (structured + unstructured), I use the same journal file of fluent, so all their sets are the same. Maybe, for the triangular grids, high order discretized method should be used in order to reduce the numerical viscosity. I will try to test it! Do you think my guess is possible?
you mentioned "high order" here , i know little about fluent ,but i know that it is usually 2 order for FVM and if you would use higher order ,you may choose FEM.But i think 2order FVM is sufficient enough for this type of problem.you may try but i think that is not the point here .i can just afford this information to you,you may ask the guys around you who know fluent for help,it will be more helpful.bless!

April 3, 2011, 02:10
#12
Member

Felix
Join Date: Mar 2011
Posts: 45
Rep Power: 6
Quote:
 Originally Posted by iilw1314 you mentioned "high order" here , i know little about fluent ,but i know that it is usually 2 order for FVM and if you would use higher order ,you may choose FEM.But i think 2order FVM is sufficient enough for this type of problem.you may try but i think that is not the point here .i can just afford this information to you,you may ask the guys around you who know fluent for help,it will be more helpful.bless!
I tried to use QUICK for discretization of momentum and turbulent viscosity, but it didn't work!
Usually, the structured grids are adopted, but I want to simulate it using dynamic mesh, so I use triangular grids.
It is driving me crazy！ Maybe, I should change my thesis!

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post user Main CFD Forum 6 November 25, 2010 02:14 louis.langouche@gmail.com FLUENT 0 March 3, 2010 09:19 David FLUENT 4 June 8, 2005 16:20 Art Stretton Phoenics 5 April 2, 2002 05:59 Chuck Leakeas Main CFD Forum 2 May 26, 2000 11:18

All times are GMT -4. The time now is 09:09.