[skapetan]

Hello,

I want to keep it short and describe my problem.

I generated a 2D Mesh for ribbed channel where I want to analyse conjugate Heat Transfer and validate experimental Nusselt data.

Some key facts from my analysis:
K-omega SST turbulence model, Contact-IF between Solid(Rib) and Fluid, AllY+-Treatment, WallY+ 0-13 ,Air enters the channel with an Temperatur of 293,15K, I set a constant heat flux of 200 W/m² for the channel bottom and rib bottom .


The channel consists of 12 ribs. The flow can be assumed as periodic. Therefore I evaluated just the Nu-Distribution between rib 9 and 10.


I've generated a grid which had a really good accurancy but I couldn't get satisfactory results for the validation.

Based on that I decided to refine the mesh from rib 9 to rib 10 and the walls were Heat is transfered. After refining I got a Wall Y+ values of <1 ( just for the near wall area from rib 9 to 10, rest of the channel remained between 0-13)

I expected to get better results (means more equal to the experimental data as before) in the refined area.

However I got unrealistic temperature gradients in the refined area of the solid and bad accurancy.

Can someone describe why the solution is that bad ?

I put some pictures in the attachment for better illustration.

I hope this helps.Thank you.