Thread by @bengtxyz

Have since before a simple energy/power model for Sweden, balancing the load with production + import/export (sorry in Swedish only)

https://twitter.com/bengtxyz/status/1550543370182971392?s=20&t=Lu76aFSoAWcd_kjMk7hoMw

. When duplicating the load to around 300 TWh only adding wind power we get this graph. 1/n

This is of course a very simplified picture that does not account for demand flexibility that is expected to be high for this scenario. Also import/export, water power dispatch etc are handled rather rigidly. Anyway it may be used to understand some of the dynamics. 2/n

Wind power is modelled simply by scaling the actual production data for Swedish wind power during 2020-2021 with a factor to reach the desired energy/power levels. As can be seen the produced power pattern is very fluctuating due to the high wind power content in the mix. 3/n

Also export/import and water power dispatch is fluctuating strongly between min and max values. Anyway, the purpose with this thread is to investigate the production/consumption of green hydrogen that would be used to cover for the shortage periods. 4/n

That is, when production is higher than the load, the excess energy ("curtailed") is used to produce H2 using electrolysers. And reversely when production is lower than the load ("shortage") the stored H2 is used as fuel for a gas turbine to produce the lacking power. 5/n

Using a 40% power-to-gas-to-power efficiency (which is very similar to the values used in gasturbineworld.com/gas-turbines-b…), meaning that 0.02 kt/GWh H2 is added to a store when producing H2 and 0.05 kt/GWh is removed when consuming H2. 6/n

The same article also mentions that the US fleet-wide curtailment is 3.4% of the wind power production. In this model the curtailment is around 5.3% which is not that far off also considering there is no demand flexibility in this model. 7/n

20 GW of electrolyser capacity has been used in this simulation. When just looking at the store level (store is initially empty) we can see that just using the curtailed energy is not enough to keep the store in balance, lacking almost 300 kt after the two years. 8/n

This can also be seen from the ratio of "shortage"/"curtailed" energy which is closer to unity than to 40%. Hence we need to add more production to serve the same load. We scale the wind power until the H2 store is in balance. 9/n

We have added about 10% (15 TWh) more wind energy, and the water dispatch and import/export volumes has changed slightly. Now we have the hydrogen store is in balance (shortage/curtailed ratio is 40%). The size of the needed store is approximately 270 kton H2 or 5.4 TWh. 10/n

Now there are some interesting observations here. First, the storage much take into account not only seasonal wind patterns. The relative low wind in the winter of 2021 wasn't enough to fill up the H2 store. Hence wind patterns of several years must be considered. 11/n

Secondly, the sheer amount of storage is insane for a country like Sweden. We have no salt caverns or akvifiers for large scale storage. Even for these, 5.4 TWh would be a sizeable amount. But we have probably the largest LRC in the world for natural gas called "Skallen". 12/n

The "Skallen" lined rock cavern can store about 10 GWh of H2. Hence 540 of these "Skallen" store would be needed in this scenario to hold 5.4 TWh! (OK a lot will disappear if demand flexibility was accounted for, but anyway). 13/n

And lastly, the utilization of the electrolysers. To cover the wind bursts we needed 20 GW capacity. In this model we produce 5.6 TWh worth of electric energy from H2 yearly. With 40% efficiency we need 14 TWh electricity in, or a mean power of 1.6 GW. 14/n

Thus we're utilizing the electrolyzers to 8%. With the scarcity of electrolyser resources it is probably not a good way to use them with this low utilization. It is a little bit like saying "build nuclear reactors but just run them 10% of the time..." 15/n

https://twitter.com/MLiebreich/status/1564352301640417282?s=20&t=Lu76aFSoAWcd_kjMk7hoMw

Again, this is a simple balance model that don't take account for the complex dynamic of a real system, but it is probably not completely off either (beside for the demand flexibility which clearly would impact some of the numbers, but probably not by orders of magnitude). 16/n

Hopefully it gives some insight to the challenges of large scale energy storage when having a large portion of VRE sources in the energy mix. It is probably a better idea to use the green hydrogen for other purposes before burning it in a gas turbine. 17/17

Here is same graph but with the curtailed peaks, producing H2, also removed to present the smoothed power output. Some peaks are still present (Q4 2021) where the 20 GW electrolyzer can't absorb all of the power peaks.