[kcoastal]

Hi all,

I'm a new user to REEF3D and have been very impressed thus far having worked through the tutorials (thanks to the development team).

I was recently tasked with assessing wave attenuation through a "wave screen" on the crest of breakwater structure. Specifically, I need to determine if a structure comprised of vertical members provides better/equivalent/worse attenuation than horizontally placed members.

The wave screen is comprised of 600mm x 300mm x 2000mm structures with a 200mm gap between each. This can be rotated by 90deg to represent vertical / horizontal.

When I run my simulation with the vertical members, water appears to flow in-between the gaps as would be expected. However, when I run the horizontal simulation, the water appears to get "stuck" between the gaps, almost as if the members are "sticky". I found this surprising given everything about the simulations are the same re. domain size and grid spacing (0.006m).

My question: Is there a minimum number of cells that should be used to represent gaps and do the structures have a friction factor? Or could this issue be related to wet/drying of cells?

I have attached a few comparison figures which hopefully explains the issue along with the control and ctrl.txt files.

[Felix_Sp]

Hey,

this is a problem with the reinitialization of the level-set in REEF3D. It inhibits the propagation on "dry" cells. Due to this problem, the calculation of wave run-up with REEF3D is computationally very costly, as the only effective and physically accurate solution is to increase the grid resolution.

You might try F 40 23 or F 49 0. While F 40 does not help a lot in my experience, F 49 does change free surface propagation quite a lot. However, the overall simulation gets unphysical as the overall mean water level does increase. F 49 0 only worked for me, while staying physical, with very long waves (wave period ~ 60 seconds).

In regard to your second question. If you mean roughness when you say friction factor, then yeah, your Nikuradse roughness is B 50 0.000001 on the solid.

So in conclusion, you might try local or global (if you have the computational power) grid refinement. Also, is your setup/ your wave screen symmetrical? You might save a lot of computational power by having a setup with a symmetry boundary conditions in the mid of the flume. So B 10 0.0 47.0 0.0 2.5 0.0 7.0 and C 12 3 in the control.txt for your setup.

Kind regards
Felix