[s.m]

Hi Dear All,
i am working on three element airfoils. i do my mesh with snappyHexMesh, and my analysis with openFoam.
my pressure contour has a flactuation on leading edge of main element, and the leading edge of flap, why?
i change the scheme, i play with the relaxtion-factors, increase the subdivision in blockMesh to have a better mesh, change the boundary conditions, i also use the Gambit for meshing, i do all of this work but this flactuations don't remove.

what should i do, please give me a little help..
thank you very much

[s.m]

when i draw the pressureCoeffs figure, this flactuation are obvious, and it is wrong, because my experiment results donot have these flactuation.
please guide me, i really don't know what should i do else

[naveen]

dear s.m,

the problem is the mesh u hav created at the leading edge, tats why u r getting fluctuations at the leading surface of airfoil....it should be fine at the blunt edge of airfoil....

For angle of attack u hav to give in terms of UXcos(theta) in x direction and UXsin(theta) in y direction for all the three elements....dont specify for each element...

for cp have to plot only with points without line in excel sheet...


If u provide me the mesh file or snappyHexMesh file, i can hav a look and i can give some suggestion to get the correct results......

[s.m]

Originally Posted by naveen

dear s.m,

the problem is the mesh u hav created at the leading edge, tats why u r getting fluctuations at the leading surface of airfoil....it should be fine at the blunt edge of airfoil....

For angle of attack u hav to give in terms of UXcos(theta) in x direction and UXsin(theta) in y direction for all the three elements....dont specify for each element...

for cp have to plot only with points without line in excel sheet...


If u provide me the mesh file or snappyHexMesh file, i can hav a look and i can give some suggestion to get the correct results......

Hi Dear Naveen,
Thank you for answering me. i had been guessed that it may be from my mesh, so i increased the subdivision in blockMeshDict from (660 480 1) to (880 480 1) for having a better mesh, but the result for pressureCoeff figure didn't make any changes.

whould you also look at hte forceCoeffs that i define in system folder, thank you.

[s.m]

Dear Naveen,
i put my snappyHexMesh file, would you please give some advice to me?
Thank you very much.

[Artur]

I would recommend pushing this up a bit, to say 60, and seeing if it will make it better:

resolveFeatureAngle 30;

You didn't post a close-up view of your mesh so it's hard to judge but from my past endeavors I've learned that this setting can have a huge impact on the accuracy of smooth surface meshing.

If that fails try changing this as well:

tolerance 4.0;

Depending on how fine your mesh is it may be better to have it bigger or smaller. Again, hard to say without seeing the actual mesh.

Hope this helps.

[s.m]

Originally Posted by Artur

I would recommend pushing this up a bit, to say 60, and seeing if it will make it better:

resolveFeatureAngle 30;

You didn't post a close-up view of your mesh so it's hard to judge but from my past endeavors I've learned that this setting can have a huge impact on the accuracy of smooth surface meshing.

If that fails try changing this as well:

tolerance 4.0;

Depending on how fine your mesh is it may be better to have it bigger or smaller. Again, hard to say without seeing the actual mesh.

Hope this helps.

Hi Artur,
here you are.

[Artur]

Having had a closer look at picture 5 it seems that your mesh snaps to the surface of the stl quite well but it seems that the solid itself is not very smooth. Have you tried increasing the triangulation density (not sure what cad tool you're using, it's pretty straightforward to do in Rhino or Autodesk Inventor).

[s.m]

Originally Posted by Artur

Having had a closer look at picture 5 it seems that your mesh snaps to the surface of the stl quite well but it seems that the solid itself is not very smooth. Have you tried increasing the triangulation density (not sure what cad tool you're using, it's pretty straightforward to do in Rhino or Autodesk Inventor).

yes i increase the coordinates of airfoil's wall. i have used the solidWorks for drawing the airfoil, and then i use the Slome for giving the *.stl file, that is need for snappyHexMesh.
The fluctuation of pressure are better when i execute snappyHexMesh with these new *.stl file, but my results are become really inaccurate.
i don't know what should i do

[s.m]

thease are the picture of mesh with new *.stl file.

[s.m]

and also these...

[s.m]

Dear Artur and dear Naveen,

Now i got good and reasonable results, actually i reduced the domain size and also the mesh cells, and i changed the boundary condition from " velocity inlet & pressure outlet &slip " to "freestream" for all patch of domain, and i got better results for freestream boundary condition, i don't know why? do you have any idea?

p-s : i always think that it is better to increase the subdivision of the mesh, but now increasing the domain and also the mesh cells doesn't give me good result, why?

i attach the picture of result
1- " velocity inlet & pressure outlet &slip "boundary condition and also it's domain mesh
2- freestream boundary condition and also it's domain mesh
3-free stream boundary condition and also it's increased domain mesh
4- " velocity inlet & pressure outlet &slip "boundary condition with increased domain size and also domain mesh number
in the following.

[s.m]

1- " velocity inlet & pressure outlet &slip "boundary condition and also it's domain

[s.m]

2- freestream boundary condition and also it's domain mesh

[s.m]

3-free stream boundary condition and also it's increased domain mesh

[s.m]

4- " velocity inlet & pressure outlet &slip "boundary condition with increased domain size and also domain mesh number

[Artur]

It all depends on how you're treating the turbulence (k-epsilon, k-omega, S-A, LES, etc.) and how you're modeling the boundary layer (resolved or wall functions). Each combination of these will dictate different mesh resolutions and other characteristics in order to yield accurate results. Not to mention the importance of adopting appropriate boundary conditions.

I have no experience with CFD of airfoils, unfortunately, so cannot give you any more specific guidance but I'm sure there's a lot of literature out there that will explain this in much more detail.

Glad to see your results have improved and good luck.

[s.m]

Originally Posted by Artur

It all depends on how you're treating the turbulence (k-epsilon, k-omega, S-A, LES, etc.) and how you're modeling the boundary layer (resolved or wall functions). Each combination of these will dictate different mesh resolutions and other characteristics in order to yield accurate results. Not to mention the importance of adopting appropriate boundary conditions.

I have no experience with CFD of airfoils, unfortunately, so cannot give you any more specific guidance but I'm sure there's a lot of literature out there that will explain this in much more detail.

Glad to see your results have improved and good luck.

Thank you artur,
whould you please explain more the effect of increasing the "resolveFeatureAngle" ?

you recommended me to push resolveFeatureAngle up to 60, where can this angle effect?

[Artur]

I encourage you to go through these slides:

http://openfoamwiki.net/images/f/f0/...SlidesOFW7.pdf

They explain reasonably well how sHM actually works. On slide 32 you will find the explanation of what resolveFeatureAngle does exactly.

[s.m]

Originally Posted by Artur

I encourage you to go through these slides:

http://openfoamwiki.net/images/f/f0/...SlidesOFW7.pdf

They explain reasonably well how sHM actually works. On slide 32 you will find the explanation of what resolveFeatureAngle does exactly.

Thank you, i saw it before, but i couldn't understand what is the effect of increasing or decreasing this parameter!
sorry, can you explain it more for me?