[Lym]

Hi everyone,


I am very new to CFX and my first project is the simulation a steady-state free jet out of a nozzle into the entrainment.

I'm using a 4° wedge of the geometry with only 1 cell in phi-direction, so it is basically a 2D simulation.


The problem I'm facing is the following:
While using the k-epsilon and the k-omega turbulence model I had no problems with convergence (it does not even show wall scale in the solver manager), but when I'm using the GEKO model the wall and boundary scale stays constant without convergence while the othre residuals are converging.


I'm using a hexahedral mesh with 1x1 mm cells inside the nozzle and growing cells in the entrainment.
I tried a finer resolution in the boundary layer of the nozzle wall with y+ = 50, but that also did not help.



Does anyone know the cause for this problem?

[Opaque]

can you show the plot for RMS instead of MAX?

if you include all equation residuals in your results file? Can you locate the MAX Residual? Where is it?

[Lym]

Thanks for answering.
Of course. I also uploaded the plot for Momentum and Mass

Sadly I can't find the residuals in post. They should probably be listed as variables, right?
Also what is the best way to visualize them? Do I use a contour?

[Opaque]

Residuals are only written if you select them under Output Control/Results. I am usually interested in where the maximum value is.

Your solution looks just fine. Sure it would be great if the wallscale equation was solved exactly for the provided mesh, but keep in mind the wall-distance is an approximation needed for another approximation (turbulence model).

Plot the wall distance around the boundaries, and check if the values are as expected. The Wallsale only provides a way to compute the wall-distance.

[Lym]

Hey, I hope you had a nice holiday

I just plotted the wall distance at the nozzle walls, but I am not sure what the results are supposed to show.
Is the wall distance the distance between the nearest knot and the wall?

Because the values I get are different depending of how far I am placing the line, that I am using for plotting the chart, away from the wall.

1. plot: blue line
2. plot: red line

[Opaque]

From the initial post, there is something OFF.

For all k-omega-based models, i.e. SST, k-omega, and GEKO, the wallscale is always computed to estimate the wall-normal distance.

The fact that the RMS residual is flat after many iterations indicate the equations cannot be further converged due to some mesh issue or singularity in the model.

The "entrant corner" is a natural singularity for discrete methods, so I would not be surprised if the residual is "stuck" near that corner. Have you been able to output all the equation residuals to the results file. Check your OutputControl settings.