Implement C1 continuous version of the Erbs diffuse-fraction/decomposition model

[adriesse]

• Closes Erbs diffuse fraction model has discontinuities at transition points #1755
• I am familiar with the contributing guidelines
• Tests added
• Updates entries in docs/sphinx/source/reference for API changes.
• Adds description and name entries in the appropriate "what's new" file in docs/sphinx/source/whatsnew for all changes. Includes link to the GitHub Issue with :issue:`num` or this Pull Request with :pull:`num`. Includes contributor name and/or GitHub username (link with :ghuser:`user`).
• New code is fully documented. Includes numpydoc compliant docstrings, examples, and comments where necessary.
• Pull request is nearly complete and ready for detailed review.
• Maintainer: Appropriate GitHub Labels (including remote-data) and Milestone are assigned to the Pull Request and linked Issue.

[adriesse]

This new model/function is using the same structure as all the other diffuse-fraction/decomposition functions with one exception: it has an optional input for dni_extra, which makes datetime_or_doy also optional.

[adriesse]

@kandersolar you have a very good point that I have also thought about. Here are a two options that are not mutually exclusive:

• delay the merge and reconsider the question just before the next release
• give the PR reviewers access to our manuscript

Here is a plot from the manuscript:

[cwhanse]

@kandersolar we added the sandia_multi inverter function before the reference paper was published. The extension of iam.physical to handle multiple interfaces was originated in pvlib discussions, if I'm not mistaken. So we've made exceptions before.

[wholmgren]

Thanks @adriesse!

I don't think we need to wait for a publication in this case.

[adriesse]

Does anyone have an idea why these test failures started to pop up?

[kandersolar]

I suspect RuntimeError: Optimal parameters not found: The maximum number of function evaluations is exceeded. errors are due to scipy/scipy#19103.

[kandersolar]

Thanks @adriesse!

[adriesse]

Ok, seems ready to merge then--or do we have to wait for scipy to sort itself out?

[kandersolar]

do we have to wait for scipy to sort itself out?

Nope, no reason to hold this up for that :)

[adriesse]

Thanks @ all!