[inf.vish]

Hi all,
I am facing difficulty in validating openfoam for a flow in a 3D pipe. The flow is laminar and the fluid is water. Inlet velocity is 1E-4 m/s. Length of pipe = 100mm and Diameter = 10mm. I am pretty sure my boundary conditions are correct since i am able to solve the same problem using 2D axisymmetric wedge (and the results conform with the Poiseuille's law).

I am attaching both the files. Please tell me where i am going wrong.
Use fluent3dMeshToFoam for converting the .msh mesh.

2D axisymmetric - https://docs.google.com/file/d/0B8wx...it?usp=sharing
3D pipe - https://docs.google.com/file/d/0B8wx...it?usp=sharing

[romant]

could you please provide more information? maybe the problem can already be solved not having to run your cases.

1. What doesn't work?
2. What are your boundary conditions on the walls?


One thing you can do is build a pipe with blockmesh and validate the case there. It gives you good control over the mesh. I even remember someone posting a script or tool to make a simple pipe in blockmesh if you don't know how to make one yourself. Otherwise you can use swiftblock (help threads are here in the forum) http://openfoamwiki.net/index.php/Contrib/SwiftBlock to make a pipe.

[inf.vish]

Everything works. The solution even converges but the answers don't match with the theoretical values. The results i obtain from 2D axisymmetric case match the theoretical results. The results i obtain from 3D case do not. The problem is that the dimensions and transport properties are the same. so the solution shown by the two should be close.
The boundary conditions are as follow
U - inlet 10^-4 m/s, outlet zeroGradient, wall fixedValue uniform (0 0 0)
p - inlet zeroGradient, outlet fixedValue uniform 0, wall zeroGradient.

[romant]

Do you mean theoretical values for a fully developed laminar flow? Have a look at your dimensions, the pipe is only 100 mm long and 10 mm in diameter, this means that even at the end of the 100 mm you only had 10 diameters development length.

If your velocity profile is flat at the entrance, you might not have a fully developed profile at the outlet, meaning that there will be discrepancies between theory and CFD. http://www.engineeringtoolbox.com/en...low-d_615.html

[inf.vish]

Yes for a fully developed laminar flow. But the velocity is very small and if you look at the entrance region length given by the equation l_entrance = 0.06*Diameter*Re which gives me an entrance region length of 60mm which is less than 100mm and i get a fully developed flow at the outlet.
I am getting theoretically correct answers with 2D axisymmetric case if you can just open and see that file. The max velocity is 1.97e-4 and the corresponding pressure difference comes out to be around 3.2e-6 (kinematic pressure) which conforms with poiseuilles law.

Correct me if i am wrong.

[linnemann]

Hi

Please run

transformPoints -scale "(0.001 0.001 0.001)"

on the 3d case.

The dimensions are 5x5x100m which should be 0.005x0.005x0.1m

Maybe also run

refineWallLayer -overwrite WALL 0.5

a few time to get better BL resolution at the wall.

[inf.vish]

Thanks for pointing out the mistake Niels Nielsen!
The case is running properly now