Learning Forum

[pablobhs]

I am trying to model a hot jet impinging at an angle on an effusion cooling plate. While comparing my simulation with a colleague's results, I saw severe underprediction of TKE generation in my shear regions. I plotted production of k which seems to be fine, however the dissipation is probably too high. Is the fix more mesh-related or a turbulence model issue? Im using k-omega sst with default settings except production limiter is turned off and viscous heating and corner flow correction are turned on. 

 

 

[Essence]

Hello,

Mesh is a bit slender at inflation layers. This can create the cell jump issues. Can you try using polyhedra mesh?

[pablobhs]

Hey, thanks for the reply.

I did try a polyhedra mesh but the solution was diverging, maybe i need to resolve it better.

What about the free shear region? That has structured cells everywhere still the TKE generation is poor.

[Essence]

The issue may be related to sudden cell jumps in shear region. Since the velocity is quite high, these cell jumps can be detrimental to the solution.

Is the cross-section a square or a circular? Two-equation models are known to overpredict dissipation, which can result in lower TKE. Please use higher order discretization schemes. Ensure the y+ values are adequate.

[pablobhs]

The cross section is essentially a "sliver" of two rows of cooling holes, 2.18mm wide. 

y+ is <0.3 throughout the domain, except on the free slip "side walls" where its 0.5.

I tried running a transition SST with QUICK discretization schemes for everything. No noticable changes. 

 

[Essence]

I see. Thanks for the information. What about the mesh? For the given velocity field, the mesh seems a bit coarse. Can you try to run the case once using Reynolds Stress Model (RSM)? This will help us to narrow down the issue.

[pablobhs]

Here's how the global mesh looks. Max. cell length is 3e-4m. 

Ran the RSM model. It gave a floating point exception with QUICK schemes and diverged with 1st order upwind schemes.