Updating an existing pv model from pvib 0.8.x to 0.9.1

[Ludee]

I am currently updating an existing pvlib model and having some problems with different errors.
In addition it is not clear what parts need to updated to run with the latest version.
The topics I found are:

• PVSystem module_type
• PVSystem racking_model
• TEMPERATURE_MODEL_PARAMETERS

Any help is appreciated. The result would be a step-by-step instruction to update your existing model.

[mikofski]

Can you please share a minimum complete verifiable example? Or a little more detail about each of the above specifically that you are having issues with?

[kandersolar]

@Ludee in case you aren't aware, we do log changes for each version in our "what's new?" documentation page: https://pvlib-python.readthedocs.io/en/stable/whatsnew.html

But if that doesn't answer all your questions we can help with specifics like @mikofski says.

[Ludee]

Thanks for the fast reply. Indeed my question is lacking a good example. sorry!
I am developing a model with different setups and weather data, thus I am not able to create a MWE that reproduces the errors.
The model is running with pvlib 0.8.1 without any problems and with pvlib 0.9.1:

A:

lib\site-packages\pvlib\tools.py:340: RuntimeWarning: divide by zero encountered in true_divide
np.trunc(np.log(atol / (df['VH'] - df['VL'])) / np.log(phim1)))
pvlib\tools.py:340: RuntimeWarning: invalid value encountered in log
np.trunc(np.log(atol / (df['VH'] - df['VL'])) / np.log(phim1)))

B:

lib\site-packages\pvlib\modelchain.py", line 1770, in run_model
self._run_from_effective_irrad(weather)
...
lib\site-packages\pvlib\tools.py", line 364, in _golden_sect_DataFrame
raise Exception("Iterations exceeded maximum. Check that func",
Exception: ('Iterations exceeded maximum. Check that func', ' is not NaN in (lower, upper)')

Any further help is appreciated.

[cwhanse]

@Ludee between v0.8.1 and v0.9.1 there were some improvement to the code that computed the maximum power point on IV curves. The warnings and error you are encountering relate to that code. For the warnings (A), I would guess that at least one IV curve was requested with POA irradiance of 0, so photocurrent is 0, and Voc is zero. For the error, I'd guess something upstream produced a NaN for the IV curve Voc.

Without a reproducible example I couldn't diagnose further. I'm interested to know the cause of the error so we can improve the function if warranted.