Bug: Axial induction calculation called from the floris interface crashes

[MarcusBecker-GitHub]

Axial induction calculation called from the floris interface crashes

The function get_turbine_ais() in the floris interface crashes, since it is called with ref_tilt_cp_cts instead of ref_tilt_cp_cts_sorted. ref_tilt_cp_cts is a n-turbine array, ref_tilt_cp_cts_sorted is a 1 x 1 x n-turbine array.
As a result, line 432 in /simulation/turbine.py throws an error:

    if ix_filter is not None:
        yaw_angle = yaw_angle[:, :, ix_filter]
        tilt_angle = tilt_angle[:, :, ix_filter]
        ref_tilt_cp_ct = ref_tilt_cp_ct[:, :, ix_filter] # <---- ERROR

As I see it, the bug could be fixed by either changing the function get_turbine_ais() in the floris interface:

    def get_turbine_ais(self) -> NDArrayFloat:
        turbine_ais = axial_induction(
            velocities=self.floris.flow_field.u,
            yaw_angle=self.floris.farm.yaw_angles,
            tilt_angle=self.floris.farm.tilt_angles,
            ref_tilt_cp_ct=self.floris.farm.ref_tilt_cp_cts_sorted, # <---- CHANGE
            fCt=self.floris.farm.turbine_fCts,
            tilt_interp=self.floris.farm.turbine_fTilts,
            correct_cp_ct_for_tilt=self.floris.farm.correct_cp_ct_for_tilt,
            turbine_type_map=self.floris.farm.turbine_type_map,
            average_method=self.floris.grid.average_method,
            cubature_weights=self.floris.grid.cubature_weights,
        )
        return turbine_ais

or by modifying the way ix_filter is used in the axial_induction() function in turbine.py, but I don't know what would be the favoured approach. Additionally, it might be good to add a comment to the axial_induction() function in turbine.py to define the input structure, as it is done with the other inputs.

How to reproduce

from floris.tools import FlorisInterface
fi = FlorisInterface('*Path*/examples/inputs/gch.yaml')
fi.calculate_wake()
fi.get_turbine_ais()

Floris version

FLORIS v3.4

[rafmudaf]

Thanks for documenting this @MarcusBecker-GitHub. I've created a pull request that I believe fixes this, see #665. @bayc is more familiar with the architecture of the Farm attributes and what the correct fix should be, so I'll wait until he reviews it to determine whether this is closed.

[rafmudaf]

@MarcusBecker-GitHub I've verified your example script in the original post runs after #665. That pull request was merged into develop and will be included in the next release. Thanks for reporting this.