Add poa_ prefix to returned named when return_components=True in transposition functions

[kandersolar]

• Closes Add "poa_" prefix to return_components=True transposition model outputs #2529
• 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.

Note that pvfactors_timeseries (actually solarfactors under the hood) uses the old names, so the pvfactors tests are going to fail until solarfactors is updated and a new version released.

[ramaroesilva]

@kandersolar nice to see this moving! Will this be done also for isotropic, klucher, reindl, and king? I also noticed that the last two have "poa_" in the docstring but not on the actual variable within the code.

Also, I guess I should address #1553 (which is being handled here #2527) on top of this right? Once you comment on my first question (and it is addressed, if I'm right) I can later review this PR, so that its merging is sped up.