[BUGFIX] AWC wake mixing factor with mixed awc_modes inputs

[gyalla]

This PR addresses a bug in the mixing factor when using AWC that was identified in issue #1089.

[gyalla]

@misi9170 These changes fix the issue with AWC in the cases I'm testing. However, I'm not sure I fully understand the shape of the numpy arrays awc_mode_i and awc_amplitude_i in this part of the code, and particularly why the [:,:,0,0] slicing is used. Can you double check the changes? Thank you.

[misi9170]

@gyalla , totally fair, yes I'll take a look now

[misi9170]

@gyalla , your proposal looks good to me. I've reformatted it a little bit, but all works fine. I was going to add another test to ensure that checks this functionality, but I realized we don't yet have a place where we test how the operation models affect flows, so to speak---I'll open an issue for that (#1091 ), but likely that'll happen at a later date.

Thanks again for providing an example runscript in #1089---that's really helpful.

[gyalla]

Thanks for taking a look @misi9170, and for reformatting the code. All looks good to me, so feel free to merge whenever makes sense.

[misi9170]

Making a note here that this change means that if any entries in awc_mode_i are something other than "helix", zeros will be returned. This works as intended (for instance, if the entry is "baseline"), but will also return zeros if the awc_mode_i contains anything other than "helix" or "baseline". Previously, passing anything else would have raised a NotImplementedError. Still, I think that's fine, since there is already a warning raised in the AWCTurbine model if this occurs. Moreover, I'd like to incorporate this kind of "checking" into a future class that is dedicated to handling control setpoints #1032.

[gyalla]

@misi9170, we need to specify the mixing factor as

awc_mixing_factor = np.zeros_like(awc_amplitude_i,dtype=float)

Otherwise, if awc_amplitudes are specified with integers, which they commonly are, then awc_mixing_factor will be rounded down and likely zeroed out. Thanks to @moprak-nrel for helping me track this one down! I'll push the fix.

[misi9170]

Good catch @gyalla ! Actually, I was just talking to @rafmudaf about exactly this type of thing, and what we should really be using is the floris_float_type---I'm going to quickly change it to that, but you're right about specifying the type!