model.energy/future: Future German renewable power system with today's data

First Posted: 2024.01.23, Last Revised: 2024.01.23, Author: Tom Brown

Original Twitter post

🚨 new web app 🚨

future renewable power systems running on today's market data 🌬️☀️, scaled up

https://model.energy/future/

updated each day 🕰️
wind, solar, hydro, batteries, hydrogen storage 🔋
Germany as an island 🏝️
coming soon: interconnectors, new electric demands 🚗

How does it work?

We take the current day's demand, wind, solar and hydro time series from https://smard.de.

Then we divide generation by current capacities and scale them up to the future capacities.

All code and data is open (obvs).

For each day we optimise the feed-in of generation and storage (short-term batteries and long-term hydrogen) with 24 hours of foresight, mimicking the day-ahead market.

Demand here includes today's electricity demand and storage charging. Demand and supply match in each hour.

We have run this simulation one day at a time with 24 hours foresight over 9 years of data since 2015.

Here are weekly averages where you see more solar + electrolysis in summer, more wind + hydrogen-to-power in winter.

Extra renewable generation is needed to cover the storage losses (batteries have round-trip efficiency of ~90% in the model, hydrogen just 39%). This you can see by the share of each technology in the load.

Batteries are used to bridge generation gaps of a few hours, while hydrogen bridges multiple days of low wind and solar as well as managing seasonal differences.

Hydrogen storage is dispatched on a roughly seasonal basis, filling up in the summer, depleting in the winter.

Note that full power-to-hydrogen-to-power storage systems don't yet exist at scale, and more development is needed for hydrogen turbines to get the NOx emissions down.

Hydrogen storage is dispatched using a constant hydrogen value of 80 EUR/MWh (LHV). Like how water values are used to dispatch hydroelectricity systems today.

The hydrogen value influences electricity prices when hydrogen turbines supply power, and when electrolysers consume.

Prices here only drop to zero when supply is larger than all flexible demand. The mean electricity price is 82 €/MWh, which allows almost all cost recovery of the system cost 93 €/MWh. Hydrogen turbines don't cover all their investment costs (usual missing money problem).

There are many technical details and warnings on the website:

https://model.energy/future/#technical-details

In particular:

Allowing connection to Germany's neighbours would help balance variations, reducing costs and storage needs
Including new demands like heat pumps and electric vehicles will raise both total electricity demand and peak demand

We'll add these features soon!

Thanks to:

David Osmond for inspiring with simulations for Australia
#openenergytracker team (Wolf-Peter Schill et al) for brainstorming - it will be integrated there soon!
BNetzA for data via https://smard.de
Mirko Schäfer for time series rescaling

We welcome all your feedback & suggestions!

There are already many technical details on the site:

https://model.energy/future/

and planned features in the GitHub issues:

https://github.com/PyPSA/nowcast/issues