[biau]

Hello,

I am currently doing a simulation (airfoil) to compare between openfoam and fluent.

The mesh was done in ICEM and converted to openfoam format using fluentmeshtofoam. However, when the mesh was converted it was expanded in the z-direction by 0.57m (It extruded from -0.28 to 0.28 in z-direction) and it kept the length as 1m.

After setting up the simulation and running it, I can't get results that are similar to Fluent results (e.g. In fluent Cl =0.507, Cd=0.027 and in OF Cl= 0.326 Cd=0.0123). I tried to use different reference areas (1m2, 0.57m2, 1.57m2). But the area that gave me the best results was 0.28m2. Is there a way to find out what is the correct reference values to use when setting up the case ? I tried to use paraview to get the surface area of the top part of the airfoil and it was 0.57m.

I am trying to run the simulation inviscid and the case and mesh files are in the link below.

https://www.dropbox.com/sh/4wbivcwswkwvgvf/PV7niWNfZ5


Any help would be appreciated.

Thank you
Julio

[biau]

Anyone please ?

[Aurelien Thinat]

Hi Julio,

Few things to start with :
- did you check the material properties in OF and Fluent ?
- did you run a checkMesh ?
- I have never used the "freestream" boundary conditions before. Have you tried to use :
velocity : fixedValue at the inlet BC and inletOutlet at outlet BC
pressure : fixedValue at the outlet BC zerogradient at the inlet BC

Aurélien

[biau]

Originally Posted by Aurelien Thinat

Hi Julio,

Few things to start with :
- did you check the material properties in OF and Fluent ?
- did you run a checkMesh ?
- I have never used the "freestream" boundary conditions before. Have you tried to use :
velocity : fixedValue at the inlet BC and inletOutlet at outlet BC
pressure : fixedValue at the outlet BC zerogradient at the inlet BC

Aurélien

Dear Aurelien,

I did check the material properties, there is a slightly difference in temperature and density, however I made adjustments to make sure they are the same.

I did run meshCheck

Create time

Create polyMesh for time = 0

Time = 0

Mesh stats
points: 33280
internal points: 0
faces: 65792
internal faces: 32512
cells: 16384
boundary patches: 4
point zones: 0
face zones: 0
cell zones: 0

Overall number of cells of each type:
hexahedra: 16384
prisms: 0
wedges: 0
pyramids: 0
tet wedges: 0
tetrahedra: 0
polyhedra: 0

Checking topology...
Boundary definition OK.
Cell to face addressing OK.
Point usage OK.
Upper triangular ordering OK.
Face vertices OK.
Number of regions: 1 (OK).

Checking patch topology for multiply connected surfaces ...
Patch Faces Points Surface topology
outlet 128 258 ok (non-closed singly connected)
inlet 256 514 ok (non-closed singly connected)
wing_profile 128 256 ok (non-closed singly connected)
frontAndBackPlanes 32768 33280 ok (non-closed singly connected)

Checking geometry...
Overall domain bounding box (-10 -10 -0.2864000698) (10.5 10 0.2864000698)
Mesh (non-empty, non-wedge) directions (1 1 0)
Mesh (non-empty) directions (1 1 0)
All edges aligned with or perpendicular to non-empty directions.
Boundary openness (-3.38429445e-18 1.524654585e-18 6.438424056e-19) OK.
Max cell openness = 6.017067417e-16 OK.
Max aspect ratio = 614.8061099 OK.
Minumum face area = 6.778323149e-06. Maximum face area = 1.570952427. Face area magnitudes OK.
Min volume = 3.882624446e-06. Max volume = 0.8998417694. Total volume = 210.0252865. Cell volumes OK.
Mesh non-orthogonality Max: 15.19406969 average: 5.646290018
Non-orthogonality check OK.
Face pyramids OK.
Max skewness = 0.6821213047 OK.
Coupled point location match (average 0) OK.

Mesh OK.

End

I have just changed my BC, but results are very similar.

Do you know if OF uses first order or second order accuracy for pressure (as standard?)

Thank you

Julio

[biau]

this is my new file if you could have a look

https://www.dropbox.com/s/3b5bjf8itswvyys/newBC.tar.gz

Thank you again

[Aurelien Thinat]

The default scheme for grad(p) is specified in the fvScheme file :

gradSchemes
{
default Gauss linear;
(...)
}

With your mesh you can use higher order schemes :
1) Gauss linear corrected schemes in laplacians instead of :

laplacianSchemes
{
default Gauss linear limited 0.333;
}

2) corrected in SnGrad instead of :

snGradSchemes
{
default limited 0.333;
}

[reza1111]

Hi forum,

I am new in OF. and for being familiar with OF I tried to simulate a flow around NACA0009 at AOA of 5 degree and Re=800000. I'm going to validte my result with data in below link:
http://library.propdesigner.co.uk/ht...teristics.html

so the NACA0009 Cl at AOA of 5 should be about 0.57
in some trial with fluent I got CL=0.49, which is good for my first attempt.
so I simuluted the flow with simpleFoam solver, I used airfoil2D test case and I changed Mesh, B.C, fluid property and initial values. I tried to set my case accurately, but I am completely confused why the CL becomes about 0.016!!!
I compared the friction and pressur force between OF and fluent result the ratio was 1:25.!!!
I am sure that I am making a mistake in OF simulation. but I can't find it.

thanks alot for attention,
Reza

[reza1111]

I checked my run setup accurately, but the result was disappointing.
any idea please?

[biau]

Originally Posted by reza1111

I checked my run setup accurately, but the result was disappointing.
any idea please?

Hi,

you are getting completely wrong results due to the vectors your are using to calculate Cl and Cd.

In your controldict you need to change:

liftDir (0 1 0);
dragDir (1 0 0);

To the angle off attack you are simulating.
E.g for aoa =4 degress

liftDir (0.05233 0.99863 0)
dragDir (0.99863 0.05233 0)

Also make sure you Area reference and Are length is the same as the one you are comparing to.

Hope it helps

Julio

[reza1111]

Hi Julio

As you see in my mesh I changed the airfoil angle to 5 degree instead of changing the inlet flow angle so I set:
liftDir (0 1 0);
dragDir (1 0 0);
In my first runs I thought the wrong result is because of changing the density of fluid, so I tried to run my case with density=1, no change happened in results.
In literature the area reference is the length of chord. so I did.
Thanks alot,
Reza.

[biau]

Originally Posted by Aurelien Thinat

The default scheme for grad(p) is specified in the fvScheme file :

With your mesh you can use higher order schemes :
1) Gauss linear corrected schemes in laplacians instead of :

2) corrected in SnGrad instead of :

Thank you for your help, but I still get the same results. Nothing major changed :-(