In previous post, I assumed that a grid with a base load might be easier/better than a grid based on intermittent power sources when the aim is a balanced grid. I based this on the hypothesis that in the former system much less energy needs to be displaced than in the latter, therefor it would be easier to balance.

That is something that I can check. I could put some grid data in a model and then change a parameter in order to see which one of the two is easier/better and to what extent.

Without further ado, this is what I did:

• I downloaded 2020 data of solarwind power from the Elia website
• I took a constant level of 5,650 MW (the level at the time when I started to write this post) as the constant value for nuclear
• Read this data in the model and removed all rows that had missing data (whether from solar, wind or total load), ending up with values in all rows
• Assumed storage with a charge/discharge capacity of 95% and a capacity of 130 GWh (I used that scenario before, at first it doesn’t seem much, but be aware that it is about 1,000 times the capacity of the old Hornsdale power reserve (or about 700 times of the Hornsdale power reserve after increasing capacity)
• Both the nuclear and the solar + wind scenarios will run through the exact same routine, so the difference in result will originate from their intrinsic differences.

I ran the model several times iterating over different multipliers in order to figure out the point of optimal balance between deficit and surplus (as I did earlier here) and also the point where there would not be any deficit anymore in this test system.

These are the results for the base load system (nuclear and 130 GWh storage):

At 1.4 times the current nuclear output, the combo nuclear/storage can not supply enough energy to cover demand and there is a need for an additional 5,434 MW of some other dispatchable power source. At higher iterations, this need for other dispatchable power sources decreases and gets zero at about 1.9 times the current output. Which is something that was expected, about half of our demand is currently filled in by nuclear.

The balance point between surplus and deficit is situated somewhere between 1.6 and 1.7 times the current nuclear output (in reality it is not exactly a point, but more likely a range):

After this point more excess power is produced that can not be absorbed by demand or by storage. Adding extra capacity will bring the need for more curtailment.

The deficit curve hits the x-axis at around 1.9. That is as expected. Currently nuclear provides about half our demand, so it is expected that it would be able to cover demand at around two times the current output.

Solar and wind are a different story. These are the results of them in combination with the same 130 GWh:

The optimal point is situated somewhere between 5.4 and 5.5 of the current solar and wind output:

Contrary to the scenario of the nuclear output, the point where there is no deficit anymore is not in sight. Not even close.

The deficit curve eventually hits the x-axis at around 94 times the current output. That will be a pretty long stretched tail towards zero (for those who think that the deficit curve is already pretty close to the x-axis, look a bit closer to the corresponding values on the y-axis).

The deficit curve also sits a lot higher than the deficit curve of the nuclear curve. The maximum deficit (meaning the part that still needs to be filled in by dispatchable power sources) sits at somewhat north 13,000 MW at the optimum point and, because of the properties of intermittent power sources, will only very slowly decrease. This means that dispatchable output will be needed for a long time, even at high shares of intermittent output. Leading to a lot of energy that needs to be curtailed at peak production, while still experiencing shortages at low production.

Unless I misunderstand that easier/better-when-the-aim-is-a-balanced-grid part of the claim, the scenario with base load is obviously easier/better to balance than the scenario with intermittent energy. The rough drawing discussed in previous post seems to suggest that there might not be that much of a difference between the two, but looking at the actual intermittency tells a different story.

via Trust, yet verify

https://ift.tt/3hNDp9B

May 23, 2021