Disable user-provided tilt angle settings in wake calculation

[paulf81]

Comment out tilt angles from calc wake function

floris_interface's calculate_wake function currently accepts tilt_angles as an input, but does not apparently apply them. Propose to comment these out (and not delete in the assumption we will want to offer this function some day)

Impacted areas of the software

floris_interface.py

[rafmudaf]

@paulf81 I see that they're used throughout the farm module: https://github.com/paulf81/floris/blob/feature/remove_tilt_from_calc_wake/floris/simulation/farm.py. Are they set in another way if it's removed from this function?

[bayc]

@paulf81 I see that they're used throughout the farm module: https://github.com/paulf81/floris/blob/feature/remove_tilt_from_calc_wake/floris/simulation/farm.py. Are they set in another way if it's removed from this function?

Hey @rafmudaf, I believe @paulf81 is right that the input right now doesn't assign the tilt angles in such a way that they get used. I believe I confirmed this with the test code below. I was able to track down that the tilt angles were not being initialized in the same was as the yaw angles when they are supplied to calculate_wake which I updated in the latest commit. The new output is shown below under Output 2.

Code:

from floris.tools import FlorisInterface
import numpy as np

fi = FlorisInterface("inputs/gch.yaml")

print(fi.floris.farm.tilt_angles)
print(fi.floris.farm.tilt_angles_sorted)

fi.calculate_wake()

print(fi.floris.farm.tilt_angles)
print(fi.floris.farm.tilt_angles_sorted)

fi.calculate_wake(tilt_angles=np.array([0.0, 0.0, 0.0]))

print(fi.floris.farm.tilt_angles)
print(fi.floris.farm.tilt_angles_sorted)

Output:

[[[5. 5. 5.]]]
[[[5. 5. 5.]]]
[[[5. 5. 5.]]]
[[[5. 5. 5.]]]
[[[5. 5. 5.]]]
[[[5. 5. 5.]]]

Output 2:

[[[5. 5. 5.]]]
[[[5. 5. 5.]]]
[[[5. 5. 5.]]]
[[[5. 5. 5.]]]
[[[0. 0. 0.]]]
[[[0. 0. 0.]]]

[misi9170]

@bayc This does seem to work, but I don't fully understand the behavior when the turbine is defined with floating_correct_cp_ct_for_tilt: True and when the definition yaml has a floating_tilt_table field.

With your changes, running

from floris.tools import FlorisInterface
import numpy as np

#fi = FlorisInterface("inputs_floating/gch_fixed.yaml")
#fi = FlorisInterface("inputs_floating/gch_floating_defined_floating.yaml")
fi = FlorisInterface("inputs_floating/gch_floating.yaml")

fi.reinitialize(layout_x = [0, 5*126, 10*126], layout_y = [0, 0, 0])

# print(fi.floris.farm.tilt_angles)
# print(fi.floris.farm.tilt_angles_sorted)

fi.calculate_wake(tilt_angles=np.array([[[5.0, 5.0, 5.0]]]))

# print(fi.floris.farm.tilt_angles)
# print(fi.floris.farm.tilt_angles_sorted)

print(fi.get_turbine_powers()/1000)

fi.calculate_wake(tilt_angles=np.array([[[0.0, 0.0, 0.0]]]))

# print(fi.floris.farm.tilt_angles)
# print(fi.floris.farm.tilt_angles_sorted)

print(fi.get_turbine_powers()/1000)

I get identical power at the first turbine (and very similar power at the other turbines). Is that expected?

[[[1688.97519743  429.74855125  477.00790152]]]
<vertical deflection warning>
[[[1688.97519743  429.74457429  476.60162974]]]

Moreover, if we are going to let calculate_wake() accept a tilt_angles argument, I think that other methods on FlorisInterface should accept them too (e.g. calculate_no_wake(), calculate_horizontal_plane(), calculate_cross_plane(), etc), which is probably a bigger task to deal with.

It seems to me the most straightforward path for v3.4 is to remove tilt_angles as an input from calculate_wake() as @paulf81 proposed, and create an issue to build out the code for allowing users to manually specify the tilt angle.

[rafmudaf]

See #653 for an ongoing discussion on handling tilt angles