[rajnarayang]

Hi Everyone,
I am facing issues with a case for laminar flow over a profile. This specific shape of the section is being experimentally studied in the laboratory.
My geometry and mesh details are as shown below. I have to run a 3D case, but due to convergence issues; shifted to 2D. Even in 2D, I am facing issues.
I am using laminar model, with energy considered (the top surface has a temperature of 70c) with Simple scheme and second order upwind scheme for momentum.
My fluid is air and my inlet velocity is at 2m/s thereby making Rex approximately 25000 on the trailing flat plate. I am using Ansys fluent to solve the problem.
Now, I figured initially that my mesh at the leading edge may be an issue, hence I refined it couple of times, but never attained basic residual convergence. I tried tet mesh and polyhedral mesh also. I also tried solving it in CFX, but to no avail.
I have tried increasing the length of the domain beyond the plate to 1.3m (its currently 0.22m) to make sure that there was no unresolved flow physics that causing the problem. Convergence, Nyet. And same goes for increasing the height of the domain.
Tried Inviscid. No convergence.
To cases that converge:
K epsilon model converges in 2D, but not in 3D.
Low velocity (0.002m/s) converges. But not higher.
Very coarse mesh converges. But that doesn’t capture the boundary layer.
I believe I have tried all the option that I can think off. I am not able to figure out any further issues that I can amend. Your valuable suggestions can help.
Dimensions.jpg

Convergence.jpg

Mesh.jpg

Mesh Zoom.jpg

Velocity Profile.jpg

[rajnarayang]

Zoomed in version of Velocity profile. The instability is visible in the trailing edge as "flapping".

[Shamoon Jamshed]

the thickness seems to be unecessarily high, did you just make it on your own or found it in some literature. Try to keep it to a line and just run 2D

[rajnarayang]

Originally Posted by Shamoon Jamshed

the thickness seems to be unecessarily high, did you just make it on your own or found it in some literature. Try to keep it to a line and just run 2D

The peculiar shape is being studied experimentally in our lab. I will be getting my data from them, and hence decided to use the same shape.

Sent from my Nexus 5 using CFD Online Forum mobile app