Rshunt value used to compute reverse saturation current may be inaccurate

[markcampanelli]

Rshunt can be a bad approximation to the shunt resistance at reference irradiance, as described in

pvlib/pvlib-python#1094

In particular, this affects the fit of the reverse saturation current at reference conditions

https://github.com/frivollier/pvsyst_tools/blob/master/pvsyst/module.py#L275-L278

[frivollier]

added separate parameter with computation of shunt resistance at 1000 w.m2 ('RShunt_1000') and replaced the R_sh_ref by this new computed value rather than RShunt from the PAN file.
c755d73

New notebook: https://deepnote.com/project/e3c9fe22-c006-42ba-af37-1d51ed2de5b0

[frivollier]

I will add more clarity to what each shunt resistance parameter mean and will double-check which to use with pvlib.

[markcampanelli]

@frivollier There appears to still be a small discrepancy between pvlib's implementation and the one here, see https://github.com/pvlib/pvlib-python/blob/master/pvlib/pvsystem.py#L2264.

If you concur, then I'd be happy to make a PR to address. Please let me know what you think.

[markcampanelli]

@frivollier I raised the idea of moving the code here over to pvlib-python, and the folks there seem amenable. Would you be ok with this? I'd certainly like to somehow preserve credit to you and this repo in the new code.