CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Phoenics (
-   -   First steps - laminar flow in a pipe (

Maria November 20, 2001 15:54

First steps - laminar flow in a pipe
I'm trying to begin a study of fluid flow in pipes. I intend to have in the end a 3D pipe with bends, so I am not going to use cylindrical coordinates. As a first step, I built a model consisting of a solid block with a vertical pipe of air (I used cylinder.dat and set porosity=1 to the volume), an inlet at the base and an exit at the top. I'm using laminar model, with a constant velocity at inlet, and I expected to see the development of the boundary layer, but all I get is a constant velocity all over the pipe. I must be forgetting something very simple, but even reading the example from Simuserve of a flow in a 3D pipe, I don't see where my mistake is.

leon November 20, 2001 19:25

Re: First steps - laminar flow in a pipe
What have you set the surface roughness of the 'pipe' to ? - it is the fiction on the surface of the pipe that generates the boundary layer and needs to be set.


Tim November 21, 2001 11:38

Re: First steps - laminar flow in a pipe
What is the material properties of the block that you inserted? I think (though I am not sure) that the default material is considered not to have a wall rougness at all (ie, the no slip condition does not apply to it). Changing the block material to something "real" may aid your problem. However, I am not 100% sure about the default properties I mentioned. Give it a try though.

Maria November 22, 2001 19:40

Re: First steps - laminar flow in a pipe
I have set a global roughness of 5E-5 m under the Sources menu. I don't see where else to set it.

Regards, Maria

Maria November 22, 2001 19:47

Re: First steps - laminar flow in a pipe
I've set the domain as steel under Properties, and a global roughness of 5E-5 m under Sources. I don't see where else to change it. Any more ideas?

Regards, Maria

leon November 22, 2001 19:49

Re: First steps - laminar flow in a pipe

What material is the object set to ?

The default is set to a smooth wall with no friction. This needs to be changed to you pipe material.

The Sources menu is the correct place to start to set up global roughness factors. Have a look under POLIS for further information.


Tim November 23, 2001 14:50

Re: First steps - laminar flow in a pipe
If you are running a laminar flow situation then I do not think that the roughness will be a factor in the boundary layer development. The development will be a function of the viscosity of the fluid. I think (not 100% sure) that the roughness coefficient will only be used for turbulent flows. (which would make sense, check any moody chart, laminar flow is independent of surface roughness). One of the problems you may be facing is the fact that it is air that you are simulating. Air has a very low viscosity, and hence the boundary layer development should be relatively slow. You may not be able to see the development if your flow situation is too short or if the grid size is not adequate. What is the cylinder diameter you are using and how long is the cylinder? You may want to try simulating one of the viscous fluids. Water may show a faster development because of the difference in kinematic viscosity (by about a factor of 10). Don't know if any of this will help. Hope so.

Maria November 24, 2001 19:31

Re: First steps - laminar flow in a pipe
I thought that maybe an overall roughness for the domain would impose a no-slip condition at the surface. I don't see any specific setting for no-slip.

I will try water, just to see if something changes. I am using 0.2m diameter and 10m length and an average speed of 0.1m/s. I believe that for air this means the entrance length is roughly 13m, so I should be able to see something. My mesh has 18 cells on each of the transverse directions X and Y (using power law, coarser at the center) and 40 along the tube.

Tim November 27, 2001 12:26

Re: First steps - laminar flow in a pipe
Try your domain fluid as 198 Solid with Smooth-Wall Friction. That should result in the desired parabolic flow. I do not know why the steel will not work.

All times are GMT -4. The time now is 04:16.