Reduced order electrical mismatch model

[kandersolar]

Is your feature request related to a problem? Please describe.
Electrical mismatch loss due to cell-to-cell irradiance nonuniformity is relevant for bifacial system performance, but pvlib does not offer any tools to model this. PVMismatch can, but maybe I don't want to learn a new tool, maybe I don't have 2-diode model parameters, maybe I don't want another dependency, maybe I just want a quick mismatch estimate without the computational burden of a full I-V circuit solver...

Describe the solution you'd like
One of the many feature requests suggested at PVPMC (thanks @shirubana!) is an implementation of the reduced order electrical mismatch model described in Deline, Pelaez et al 2020, "Estimating and parameterizing mismatch power loss in bifacial photovoltaic systems".

The model itself is straightforward: given irradiance at several locations across a module, summarize the irradiance variation using mean absolute difference and feed that into a polynomial with known coefficients. What might be more difficult is deciding what the structure of the inputs should be, as we don't currently have a data structure convention for representing multiple irradiance values for a single module. It must be at least 2D (time, location in module), but maybe 3D would be better to allow vectorization across modules.

Describe alternatives you've considered
Of course a full I-V curve circuit simulator is still desirable in the long run, but this model is a lot simpler both to implement and to use. No reason not to have both IMHO.

Additional context

• pvfactors will give you localized/partitioned irradiance going up the row height if you ask it nicely; perhaps we should add a gallery example demonstrating how to do that. I don't think the infinite sheds code can return irradiance for small parts of the module, can it?
• bifacial_radiance already has a BSD-3 implementation we could use as a starting point:
  https://github.com/NREL/bifacial_radiance/blob/5f9de36861ec9bcffe169c9a4371806860c0e224/bifacial_radiance/mismatch.py#L166-L220

[shirubana]

There's also a pytest for half of it.

https://github.com/NREL/bifacial_radiance/blob/main/tests/test_mismatch.py#L51-L63

[shirubana]

Which module do you envision this function to live?

[cwhanse]

pvsystem.py for this simple model, I think. Code for detailed mismatch calculations would go into singlediode or ivtools.