EWI: The default assumption assumes 2 cycles per year, whereas I assume only one (they covered this in a sensitivity), got some help with this from @benpfluger.bsky.social.

Thanks for the questions! The model only chooses H2 for backup if we restrict the options or force the gas plants to switch to H2 (like for the Ariadne scenarios). If unrestrained, it prefers barely-used gas plants (running on biomethane or fossil gas compensated by CDR to be carbon-neutral).

Read the post. I'm not advocating 1st generation biofuels, but advanced fuels without direct land usage (wastes and residues). They will also need hydrogen to boost carbon efficiency.

- in the short-term out to 2035 we need to focus 95% of our attention on electrification, clean electricity, and replumbing biomass and agriculture; grandiose hydrogen plans should be put on hold

- we need to have a rough long-term plan out to 2040-50 because industrial plant, ships and planes have lifetimes of 30-40 years; solutions need to slot into those reinvestment cycles

- we won't see much hydrogen trade, but trade in hydrogen derivatives (ammonia, direct-reduced iron, e-biofuels) could be large, since they're easy to store and move

- fuel yields can be doubled by adding green hydrogen to mop up the excess carbon in the biomass, making e-biofuels and raising the carbon efficiency (excess CO2 can also be sequestered where that's possible)

- advanced biofuels that use wastes and residues offer a path to address non-electrifiable sectors with methanol, methane and kerosene at costs of 80-120 EUR2020/MWh and abatement costs of 120-400 EUR2020/tCO2

Main points:

- reaching climate neutrality will mean managing our limited sustainable biomass resource carefully
- we need to think not just about energy efficiency, but also carbon efficiency (how well we use biogenic carbon)

Very nice, congratulations - will take a look!

Come join the community, who knows where we land in another 10 years? 🔭

PyPSA has come a long way from Jonas Hörsch and me hacking away in 2015, trying to create a tool that would meet our needs and we'd want to use. Great to see so many other folks enjoying PyPSA too, and how the development has become so professional over the years.

We would like to thank all the contributors and developers who have made PyPSA possible up to this point. This includes those who have contributed code, reported bugs, or provided feedback! We would also like to thank the @dfg.de for funding the recent developments.

#freethemodels #PyPSA

As well as additional MGA functionality, a new Xarray based backend with easier writing of custom constraints, an optional new API for components, and more!

👉 Full release notes: docs.pypsa.org/latest/relea...

🔄 A new NetworkCollection to store multiple networks in a single object for easy comparison and plotting - for example, to 𝐜𝐨𝐦𝐩𝐚𝐫𝐞 𝐝𝐢𝐟𝐟𝐞𝐫𝐞𝐧𝐭 𝐦𝐨𝐝𝐞𝐥 𝐬𝐜𝐞𝐧𝐚𝐫𝐢𝐨𝐬.

⚙️ The 𝐨𝐩𝐭𝐢𝐨𝐧𝐬 𝐦𝐨𝐝𝐮𝐥𝐞 allows you to control PyPSA behaviour without the need to pass arguments all the time.

🗺️ 𝐈𝐧𝐭𝐞𝐫𝐚𝐜𝐭𝐢𝐯𝐞 𝐦𝐚𝐩𝐬 to explore the location of all components, their attributes, and map results or other properties.

📦 Components data is now bundled together in a 𝐂𝐨𝐦𝐩𝐨𝐧𝐞𝐧𝐭𝐬 𝐜𝐥𝐚𝐬𝐬, which introduces components-specific functionality to help simplify your scripts.

✨ 𝐖𝐡𝐚𝐭'𝐬 𝐍𝐞𝐰

🎲 Introduction of 𝐭𝐰𝐨-𝐬𝐭𝐚𝐠𝐞 𝐬𝐭𝐨𝐜𝐡𝐚𝐬𝐭𝐢𝐜 𝐩𝐫𝐨𝐠𝐫𝐚𝐦𝐦𝐢𝐧𝐠 with scenario trees as well as support for changing risk preference through Conditional Value at Risk (CVaR)-based 𝐫𝐢𝐬𝐤-𝐚𝐯𝐞𝐫𝐬𝐞 𝐨𝐩𝐭𝐢𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧, allowing users to account for extreme outcomes and tail risks in their optimization.