Reply To: channel3d in olb1.6 and olb1.5

[Fuxin He]

Dear Fedor

First of all thank you for your reply, and because of the time difference I want to apologize for not being able to get back to you in time.

Please allow me to ask you two more questions:
1. I found that when only adjusting “N, lx, ly, lz”, for example setting N=19 or ly=2L, the “geometry” is different from the original one. The main difference is that the “geometry 0” disappears at the x and y entrances and exits, and is also displayed in the terminal:
[setWallFunctionBoundary] Warning: Could not setWallFunctionBoundary (59, 10, 18), discreteNormal=(0,0,0,0), set to bounceBack
[BlockGeometryStatistics3D] WARNING: no discreteNormal is found
[setWallFunctionBoundary] Warning: Could not setWallFunctionBoundary (59, 11, 0), discreteNormal=(0,0,0,0), set to bounceBack
[BlockGeometryStatistics3D] WARNING: no discreteNormal is found
[setWallFunctionBoundary] Warning: Could not setWallFunctionBoundary (59, 11, 18), discreteNormal=(0,0,0,0), set to bounceBack
[BlockGeometryStatistics3D] WARNING: no discreteNormal is found
[setWallFunctionBoundary] Warning: Could not setWallFunctionBoundary (59, 12, 0), discreteNormal=(0,0,0,0), set to bounceBack
[BlockGeometryStatistics3D] WARNING: no discreteNormal is found
[setWallFunctionBoundary] Warning: Could not setWallFunctionBoundary (59, 12, 18), discreteNormal=(0,0,0,0), set to bounceBack
[BlockGeometryStatistics3D] WARNING: no discreteNormal is found

The original code, modifying only N=28 and modifying only ly = 2. * physRefL; is displayed in paraview as follows:
https://postimg.cc/gallery/Y2XKts2
After selecting the Y normal slice, I found out that “geometry 0” is the medium to realize the cycle boundary, does the presence or absence of geometry 0 in the XY direction have any effect on the simulation?

2.May I ask if you modified physRefL based on the parameters I provided?I modified physRefL based on const T charPhysU = ( util::pow((8.0/0.073), (4.0/7.0)) * util::pow((T)ReTau, (8.0/7.0)) ) * charPhysNu / (2. * physRefL); calculated physRefL = 0.0415. Since I want to study the motion of the particles in the boundary layer, set N=59, for a large eddy simulation, does this satisfy enough resolution for the first layer Y+=3.

sincerely,
Fuxin He