Using native boundary files with FreeStream physics

[manosser]

Hi all,

I have been using some custom boundary files made from previous SOWFA simulations using the "native_boundary_plane.py" tool.
These inflow boundary files (at xlo) include just the velocity u v w field for each time-step, and seem to be ingested properly into ABL simulations.

However, this inflow data is for a wind tunnel, and I would prefer using just the FreeStream physics instead of ABL.
Is there a way to use custom boundary files in the native format with the FreeStream physics? (or another physics method that could be better suited to a large wind tunnel simulation?).

Looking though the source code a bit it seems like the following functions are not available for physics methods other than ABL.
ABL.bndry_file = bndry_file.native
ABL.bndry_io_mode = 1
ABL.bndry_planes = xlo
ABL.bndry_var_names = velocity

Any thoughts or ideas are appreciated, thanks in advance.

[marchdf]

Hi, you are right in that the boundary planes were specifically designed for ABLs. It should be possible to adapt it to other physics cases but it would require source code modification. What you would want to do is look at ABL.H and ABL.cpp and see how ABLBoundaryPlane is being used there and mimic it's usage.

In general, I can see this capability might be useful outside the ABL physics and it might make sense to make it more general. I might look into this...

You can also see if you want to still use the "abl" physics but change the boundary conditions/initial conditions from the input file to do what you are aiming for. That is probably the easiest route for you right now (no coding required). Let us know if you want help with that.

[manosser]

Hi Marc,

Thanks very much for the help!

Sounds like from what you say, modifying the source code could be my best course of action long term. I will look into the files you mentioned.

As you also said, I have been trying to manipulate the ABL parameters to recreate our wind tunnel flow. The results seem to be moving in the right direction, but the hub-height velocity is still lower than expected, and there is some strange negative u velocity phenomenon (which could also be a problem with my bndry files).

If you or one of your colleagues has time, I would gladly take up the offer for some help refining/improving my input files. I have attached one of the input files here: abl_freestream_wind_tunnel.txt
Any thoughts or recommendations you have for initializing a homogenous flow field and for correct injection of the turbulent inflow are much appreciated!

The image below shows the custom turbulent inflow entering the domain for 1000 steps @ dt=0.001s (generated using the above input file).

Here are also a couple research papers from our group at the TUM Wind Energy Institute in case you want more information about our wind-tunnel testing.
https://wes.copernicus.org/articles/4/71/2019/
https://iopscience.iop.org/article/10.1088/1742-6596/1037/6/062007

Thanks again,
Manu