include historical nuclear EAF for different years

[fneum]

https://github.com/PyPSA/pypsa-eur/blob/master/data/nuclear_p_max_pu.csv

based on

https://pris.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx

[fneum]

2009-2011: https://web.archive.org/web/20120613184946/https://pris.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx

2011-2013: https://web.archive.org/web/20140408215702/https://pris.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx

2014 not available

2015: https://web.archive.org/web/20181122233929/https://pris.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx

2016-2018: https://web.archive.org/web/20190822112512/https://pris.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx

2019-2021: https://web.archive.org/web/20230203100052/https://pris.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx

2022-2024: https://web.archive.org/web/20260202053410/https://pris.iaea.org/PRIS/WorldStatistics/ThreeYrsEnergyAvailabilityFactor.aspx

[euronion]

We had some similar discussions here (open-energy-transition#202).
While p_max_pu is a reasonable setting for historic modelling, future EAFs might be different, so allowing for an operating envelope instead of hard setting the value could be a reasonable option.

Unavailability can also be retrieved from ENSTOE API on a more granular time basis, not sure how far back historically.

[fneum]

Yes, see #1895