atmosphere.py bug fix on get_relative_airmass airmass

[jramirezpr]

Describe the bug
A clear and concise description of what the bug is.
on conda download, line 222 problem with cos
fix
comment out line 216
z = np.where(zenith > 90, np.nan, zenith)
and add an absolute valie on line 223
... +np.absolute (90 - z)) ** - 1.6364))))

[mikofski]

Hi @jramirezpr , thanks for reporting your issue. 😄
Can you please execute

import pvlib
pvlib.atmosphere.get_relative_airmass(zenith=60.0, model='kastenyoung1989')

and post here either the output or the traceback ? Also please output pvlib.__version__ and your version of Python.

I'm using conda (Python-3.7) and the latest version of pvlib (0.6.0+27.g3077923) and I get 1.9942928525292494 and no traceback.

[jramirezpr]

The error seems to be when zenith>90, since it filters the entry to NaN and you get a problem with np.cos

[jramirezpr]

Hi @mikofski Im using conda python 3.7 And version of pvlib is 0.6.0-py_1; updated the potential fix on my original comment

[cwhanse]

@jramirezpr why do you want to calculate air mass when zenith>90?

[jramirezpr]

@cwhanse I'm running the example on the documentation https://wholmgren-pvlib-python-new.readthedocs.io/en/extras/introexamples.html and some of the zeniths in solpos['apparent_zenith'] are greater than 90, I realize that the model is for daytimes but i want the code to run without warnings

[cwhanse]

@jramirezpr I'm having trouble duplicating your issue. Is this the warning you get?

RuntimeWarning: invalid value encountered in maximum
spectral_loss = np.maximum(0, np.polyval(am_coeff, airmass_absolute))

[jramirezpr]

@cwhanse no, I get
RuntimeWarning: invalid value encountered in cos
am = (1.0 / (np.cos(zenith_rad) +

[cwhanse]

@jramirezpr if that is the warning you see, then the cause is passing np.nan in the airmass_absolute argument on line 1917 of pvsystem.py.

[jramirezpr]

@cwhanse it is basically the same reason, (passing a nan) except on line 222.

[jramirezpr]

@cwhanse (it is in the get_relative_airmass function,after the zeniths greater than 90 have been filtered with nan)

[cwhanse]

What version of pvlib?

…

Sent from my iPhone On Oct 18, 2018, at 3:36 PM, Juan Carlos Ramirez <notifications@github.com<mailto:notifications@github.com>> wrote: @cwhanse<https://github.com/cwhanse> (it is in the get_relative_airmass function,after the zeniths greater than 90 have been filtered with nan) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#607 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AFJNL_t4rJme1SIgkt3cg36_mw3oezlWks5umPRcgaJpZM4XuZpI>.

[jramirezpr]

@cwhanse 0.6.0-py_1

[cwhanse]

numpy version?

[jramirezpr]

@cwhanse numpy 1.15.1

[cwhanse]

Could you post your code? I can't reproduce that warning.