[Ricevind]

Hi,

I have just started my journey with Fluent . Right now I'm fiddling with simple 2d pipe flow. I want to simulate laminar flow through the pipe and determine minimal density of mesh I need to achieve less then 1% mass difference at in and out of the pipe.

My settings are:
material : water
pipe : 2d rectangle 2000mm length, 100mm height
model : laminar
bc : velocity - inlet, outflow, wall
inlet velocity: 0.005 m/s
mesh: unstructured triangles

What I'm expecting is parabolic velocity profile at exit. With my settings I found that the flow should fully develop at around 3m pipe length. () but my profiles don't change from length around 1m so I set the length to 2m (I had problems with longer pipes)

So i made a set of meshes with increasing density. But during calculation of more finer meshes I noticed that my profiles change from "cute" parabolas to more of a turbulent flow profiles. I think pictures will show what I mean.

And my question: Why it is so that with the same settings but different mesh densities I get different profiles for laminar flow ?

Much thanks in advance for Your replays

[Astrodan]

I have basically the same question, and can add that the same problem occurs for OpenFOAM, thus assuming it's a numerical problem rather than the software (which I had guessed anyway).

I'm still in the process of running simulations and generating profiles, and can attach them soon.

Rergarding your attached files: What is the difference between both pictures? Only the resolution of the profile?

[macfly]

Hi,

I don't have the answer but I think the modeling is not right:

- Modeling a 2D rectangle is like modeling a rectangular duct with infinite width. For a 2D circular pipe, you should model half of the pipe diameter, axisymmetric around pipe centerline. Bottom BC set to axis and top BC set to wall.

- you could try to impose fully developed profile at velocity-inlet (via a simple udf...)

- typical outlet BC is pressure-outlet, not outflow