[time-]

Hello foamers,
just a very simple question.

How much cell's growth rate affects turbulence residuals?

I'm having some trouble obtaining physical (and stable...) results with a simple case but with poor quality mesh. K-e model is used.

Most of the quality issue is volumetric gradation that is too high.

I have a simple cylinder with some circular patches acting as air jets. Air is blown from these jets at 20 m/s and I estimated length scale as 0.38*half_jet_width. Kappa and epsilon was calculated with this value in mind.

The grid I'm working on have huge cell size differences (jets are meshed with 1e-3 cells edges while on the opposite wall cells reach 0.2 length) and I think growth rate is totally random due to meshing software limitation.

I can't get stability as k-e just explode after less then 100 iterations and after that I get unphysical results in U and p in some cells (I'm working on a 25M grid and I didn't have time to check where this events appear on paraview, but I will for sure)

I don't have the original geometries and I actually don't know how to numerically check volume growth rate. I assume is bad as I visually inspected internal mesh with paraview...

In short, based on previous experience, how much do you think wrong growth rate can affect turbulence results?

Sorry for the lack of information, please ask whatever you need to know.

Bye