FLORIS wakes

[ewquon]

@lawrenceccheung, as we discussed, I've put together a demo, tested with FLORIS v3.2, to produce the following (note the turbines have different yaw misalignments relative to the mean wind):

This is the test domain / turbine config: floris_test_setup.inp.txt. The floris.yaml input file is generated under the "Farm" tab.

Todo:

• Set up FLORIS for arbitrary turbine models, based on actuator disk Cp / Ct (currently NREL5MW is hard-coded)
• Set up slicing in vertical/cross-stream planes
• ...?

[lawrenceccheung]

Hi @ewquon, this is great! I'm surprised that FLORIS was this easy to integrate. And I'm also glad that you were able to figure out how things worked in the amrwind-frontend code, I was afraid things might have gotten too confusing ;-).

This is a great demonstration and we can definitely build on it. I'm thinking that longer term, we can include more FLORIS options and better integration of FLORIS results with the frontend:

• the turbine Cp/Ct information can definitely be pulled from the turbinetype structures.
• I'm thinking I can create a popupwindow to store FLORIS parameters, in case people want to adjust the wake behaviors, etc.
• I'm also thinking that we can have an option to compute FLORIS solutions from the farm layout window, so people can play with turbine layouts and see what the impact is on power, etc.

(Those are just random ideas at the moment, we can keep adding to the list.) If there's a lot of FLORIS features, we can also split things off into a separate floris.yaml config file and maybe a separate florisfunctions.py so the files don't get too long.

Lawrence

[ewquon]

That sounds like a good plan, @lawrenceccheung. FLORIS has a few quirks (e.g., when reinitializing and then calling the wake calculation) and how you access the wake data (if you're not using their built-in visualization routines, but I've messed with it before). Happy to continue helping out, and I think we're most of the way there.