ENH: model pvsyst thin-film recombination for CdTe and a:Si in pvsystem.singlediode()

[mikofski]

Problem
This request was initially made in #163 and is also related to #470. As of #504 bishop88 can accept the thin-film recombination parameters: d2mutau and vbi to output IV curve and its gradients, but there is no propagation of this ability to the proxies that use it such as pvsystem.singlediode()

Solution
Here is a proposed list of methods that should now be able to use these new terms:

• singlediode_methods.bishop88_i_from_v()
• singlediode_methods.bishop88_v_from_i()
• singlediode_methods.bishop88_mpp()
• singlediode_methods._lambertw_v_from_i()
• singlediode_methods._lambertw_i_from_v()
• singlediode_methods._lambertw()
• pvsystem.PVSystem.singlediode()
• pvsystem.PVSystem.i_from_v()
• pvsystem.singlediode()
• pvsystem.max_power_point()
• pvsystem.v_from_i()
• pvsystem.i_from_v()

Alternatives
Please suggest alternative solutions. There has been discussion of refactoring calcparams_pvsyst or even creating model specific singlediode_pvsyst methods. Is that desired?

Additional context
this is a follow on to #504

[mikofski]

@adriesse didn't #762 and #763 already address this? I think this can be closed right?

• pvsystem.max_power_point
• singlediode.bishop88_i_from_v
• singlediode.bishop88_v_from_i
• singlediode.bishop88_mpp

I think the other functions in the list above have a Lambert-W which I don't think has a way to handle the thin-film recombination terms yet.

Can I close this issue?

[cwhanse]

I think this issue should stay open, or, be closed and replaced by several smaller issues that cover the scope:

1. finish implementation in pvsystem functions (see below)
2. implementation for PVSystem methods
3. implementation for ModelChain methods

Implementation in pvlib.singlediode is complete by #763. bishop88 functions accept the arguments. The lambertw functions don't use the arguments and shouldn't; the method can't solve the single diode equation with the recombination term.

For the first item above, pvlib.pvsystem.max_power_point accepts the arguments by #763. The following functions could (and should) accept the arguments:

• pvsystem.i_from_v
• pvsystem.v_from_i
• pvsystem.single_diode

These functions have a method kwarg which defaults to lambertw, but could be set to newton or brentq and thus d2mutau could be used.