[tomding0107]

Hi everyone,

I have a question about the two schemes available in CONVERGE.

The diffrence between these two schemes is simply that, one computes the wall distance as y=0.5V^(1/3), and the other compute like y=0.3V^(1/3).

So I'm wondering which one we should use while simulating with LES model and RANS models.

And also, if the law of the wall is applied, does the choice of the scheme will lead to different results?

Thank you !

[ksrivast]

Hello Zhihao,

To expand on the differences between the two calculation schemes for effective wall distance for the wall model:

With the Non-constant method (wall_dist_flag=0), y=0.5V^(1/3) is only used for MOVING walls. Note that V is the volume of the full-sized cell at the same embedding level. For STATIONARY walls, the distance equals to the distance from effective boundary cell center (cut cell or paired cell) to the boundary itself.

With the Constant method (wall_dist_flag=1), y=0.3V^(1/3) for ALL walls. This value of 0.3 is empirical.

The recommendation is that if you want smoother (3D) results for heat transfer (and y+ plus) along a boundary, especially one that has sharp features leading to weirdly-shaped cut cells, turn on the wall_dist_flag in turbulence.in. This is also seen to control small oscillations in y+ values reported in bound*wall.out. However, the choice in value should have negligible effects in your overall results.

Hope this helps.

Sincerely,