You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We plan to merge PyPSA-Eur-Sec into PyPSA-Eur in March 2023. The PyPSA-Eur-Sec repository would then be deprecated. PyPSA-Eur would include all features from PyPSA-Eur-Sec (additional energy sectors as well as myopic and perfect foresight).
Reasons for merger
Future energy systems will have enough sector-coupling that modelling electricity separately will be hard to justify
The workflow is currently very messy, with PyPSA-Eur-Sec removing items (like storage and conventional generators) from PyPSA-Eur and then reading them in in a different way
PyPSA-Eur-Sec has features that PyPSA-Eur can benefit from (myopic and perfect foresight, new scenario management, existing wind and solar via IRENA database)
There are confusing duplications between config files
It is difficult managing the different library dependencies between the two repositories
Snakemake sub-workflows are deprecated (to be replaced by modules) and anyway hard to work with
PyPSA-Eur-Sec is more complicated, with more complicated data and library dependencies than PyPSA-Eur, which may put off newbies and casual users; we can mitigate this difficulty by turning off non-electricity sectors by default
Planned steps
introduce basic CI testing to both PyPSA-Eur and PyPSA-Eur-Sec so we can detect problems (could test inputs but also outputs e.g. for load shedding, energy mix from temporally clustered results, prices, etc.)
pull perfect foresight code into PyPSA-Eur-Sec (@lisazeyen will do this in February 2023)
merge the PyPSA-Eur and PyPSA-Eur-Sec snakemake workflows and data with minimal additional changes (perhaps as an initial test to see what goes wrong)
slowly iron out the discrepancies (adding/removing of components before/after clustering, etc.)
test it rigorously
document everything
We estimate it will take 2-3 weeks and plan it for March 2023.
Potential difficulties
PyPSA-Eur adds components at high resolution (renewables, conventional power plants, load), then clusters, whereas PyPSA-Eur-Sec removes some of these and then adds back components after the clustering; this should be harmonised so that only components that need high resolution are added before clustering (probably renewables, conventional power plants and load are needed, since they are used for clustering in some methods), then rest is added afterwards (e.g. storage that doesn't depend on high resolution and is not used for clustering)
PyPSA-Eur uses PyPSA's generator component for conventional power plants, whereas PyPSA-Eur-Sec uses links (since need to track fuel that can be fossil or green-hydrogen-based, as well as for CO2 tracking); tentative suggestion: only use links in future
PyPSA-Eur-Sec has IRENA database for existing wind and solar, whereas PyPSA-Eur has its own methods
The myopic and perfect foresight in PyPSA-Eur-Sec needs to be "back-ported" to work with electricity-only PyPSA-Eur
The deduction of heating and industry electricity demand from total electricity demand in PyPSA-Eur-Sec needs to be harmonised with PyPSA-Eur (should be fine since the deduction should only happen if these sectors are chosen in the config)
PyPSA-Eur-Sec handles CO2 with links and stores (because of CCU/S) whereas PyPSA-Eur uses a global constraint on CO2 (suggestion: use PyPSA-Eur-Sec format)
I think this makes a lot of sense, the current duplication of scripts and elements in the config file is confusing for new users. We at Aarhus University will organize to contribute or, at least, be test users of the merged model.
To add on the procedure: Given that we don't want to lose all the contributions to pypsa-eur-sec we should git merge the repositories. Therefore I suggest the following steps:
We plan to merge PyPSA-Eur-Sec into PyPSA-Eur in March 2023. The PyPSA-Eur-Sec repository would then be deprecated. PyPSA-Eur would include all features from PyPSA-Eur-Sec (additional energy sectors as well as myopic and perfect foresight).
Reasons for merger
Reasons against the merger
Planned steps
We estimate it will take 2-3 weeks and plan it for March 2023.
Potential difficulties
The text was updated successfully, but these errors were encountered: