[alireza_b]

Hi
I'm trying to solve flow over a square cylinder at Re=5,20. I used velocity inlet and outlet boundary condition. I solved the problem, But my result was wrong, i don't know why?!?????
I attached velocity and pressure contour. As you can see results is wrong.
This problem is very simple, Re=5, laminar flow and ... . But unfortunately my result is wrong.

can you help me?

[flotus1]

How did you make sure that this is a converged solution?
And could you post an image of the mesh you used?

[alireza_b]

I attached image of mesh and convergence history.

[flotus1]

Did you try running more iterations? I would run such a simple case until round-off accuracy is achieved. One thing less to worry about.
Judging from the image of the mesh, I recommend using double precision.

Did you use velocity inlet, pressure outlet and symmetry boundary conditions?

[like.no.other]

Originally Posted by alireza_b

Hi
I'm trying to solve flow over a square cylinder at Re=5,20. I used velocity inlet and outlet boundary condition. I solved the problem, But my result was wrong, i don't know why?!?????
I attached velocity and pressure contour. As you can see results is wrong.
This problem is very simple, Re=5, laminar flow and ... . But unfortunately my result is wrong.

can you help me?

Hi alireza

which parameter is wrong?
Cd?
Cp?

Please send me your case fine, I can help u

[alireza_b]

Originally Posted by flotus1

Did you try running more iterations?

Yes, I did.

Originally Posted by flotus1

Did you use velocity inlet, pressure outlet and symmetry boundary conditions?

I used velocity inlet and pressure outlet. left, top and bottom are velocity inlet and right is pressure outlet. I also used symmetry, but the answer is wrong.

[alireza_b]

Originally Posted by like.no.other

Hi alireza

which parameter is wrong?
Cd?
Cp?

Please send me your case fine, I can help u

Hi
both flow field and Cd are wrong. I check reference value, this is correct but answer is wrong.

[flotus1]

Originally Posted by alireza_b

Yes, I did.

Please run the case in double precision untill the residuals level out and show the results (residuals+contour plots).
If still not satisfied with the results, use a mesh that is more suitable for low Reynolds number flows. Its the high aspect ratio cells that cause slow convergence. You dont need such a high resolution near the wall.

One more thing: dont expect to get a value for the drag coefficient that matches any reference solution for a square cylinder in free stream.
The extent of the computational domain is not enough for low reynolds number flows.

left, top and bottom are velocity inlet

Top and botom should be symmetry boundary conditions.

[atulurade]

i think your fine mesh area has not properly defined because it will concentrate result in fine mesh area .
also one more point is that whenever their is flow over bodies then it will shows laminar region somewhat distance along the length not to the the its near by wall region..