Clarify differences and similarities to atlite

[maurerle]

This project seems to have similar use cases as https://github.com/PyPSA/atlite

It would be very cool to have a summary of the similarities and differences on the README page.

Similarities:

• model pv and wind generation for a given region/installation with a given set of weather data
• allow usage of data from ecdsapi/ERA5

Differences

• feedinlib allows using data from openFRED?
• feedinlib is more based on single installations and is more fine-grained?
  • feedinlib models using pvlib and windpowerlib while atlite models with overall heuristics
• atlite models country wide potentials without specific installation parameters

I think the comparison here is also very interesting and would be very cool to compare the performance of the feedinlib towards atlite: https://github.com/PyPSA/atlite/blob/master/examples/historic-comparison-germany.ipynb

Would be cool if others could complete the list of features/similarities/differences :)

[p-snft]

As of today, oemof in general is developed mostly focussing on local energy systems (e.g. with a granularity that allows resolving at least individual urban areas). So, your assumption that single installations play a role is indeed true. In a combined workflow, we rather try to answer the question on which of the available roofs PV installations should be installed (first). For most of the active devs, the overall potential on a global scale is not so much of interest.

As you have noticed, feedinlib has also been used in the openFRED project (and there are models using solph that optimise the national scale energy system). To my knowledge, this is currently not the focus of most of the oemof devs, so I do not really see that someone here would contribute to a more detailed comparison with atlite.