While cleaning up some old files, I found a link to a tweet that I used one year ago in a post on windmills “balancing” the grid. To recapitulate, during the first lockdown in Belgium, there was a lot of solar and wind power, but there was low demand. However, nuclear produced flat out, as usual. Not wanting to produce at negative prices (because of high production and low demand), windmill owners started to shut down some of their windmills. This curtailment was then framed as wind “balancing” the grid.
Below this tweet was another one written by the same author. It is his answer to the question whether there are quantified estimates on what would eventually be realistically possible with dynamic demand response:
I apparently glossed over this tweet at the time I made the post, but that is an interesting question. So, in what way is that left graph easier/better than the right one when the aim is a balanced grid at all times?
The tweet shows two graphs. One with a constant level of generation (nuclear) and the other one with two peaks (wind or more generally intermittent power). None of the two fits demand. The left graph because it overproduces most of the time and the need for support at the very peak, the right graph because generation is not in sync with demand. Both need a balancing mechanism in order for the system to be balanced.
It is presented as a binary choice. Or generate loads of power, store some of it and curtail/lose the rest, risking frying the grid when demand is low. Or shift the generated power to when it is needed, risking frying the grid when generation is plentiful.
I don’t think this are the only two choices to chose from (and the first is not even an option for our politicians anyway), but let’s assume for the sake of the argument that these are our only two options. Which one is the easiest/better choice?
Looking at his other tweets, it seems he is not that fond on nuclear, so I guess he is in favor the the graph on the right. I also think that he doesn’t want filling in with gas or another fossil fuel, but with low emission power sources like pumped hydro or batteries.
The first option (with nuclear power, see the graph on the left) reminds me of what we currently have, but then on steroids. Currently, about 50% of our demand is provided by nuclear power and that would put the generation curve somewhere below the lowest point on the demand curve. The rest is filled in by primarily gas and pumped hydro at the peaks. This pumped hydro (Coo-trois-ponts) is specially built for this purpose. When there is excess generation (at night when the combo nuclear/gas produces more than the demand), water is pumped to a higher elevated reservoir and the kinetic power of the water in this reservoir is used to fill in the peaks.
In this option, not only nuclear power needs to be overdimensioned (to provide enough overproduction so the excess could be stored and used to fill in the top of the peaks), also pumped hydro (or another backup technology) needs to be jacked up tremendously. Knowing how little current pumped hydro contributes to the grid, there will be a need for huge amount of backup capacity.
Dynamic demand response will not be that problematic. Since there is base power generation, dynamic demand response will probably be limited to the top of the peaks.
The second option (with intermittent power sources, see the graph on the right) is something that is not tried before. Now not only the top of the peak needs to be filled in, but potentially almost the entire peak. Looking at the example graph at the right, the complete production during the period of low demand will have to be stored in order to fill in almost the entire peak demand when production is low.
An additional problem will be that production and demand will not always that symmetric. This scenario is also possible:
Now the entire deficit needs to be available in backup in order to be able to fill in almost the entire demand for that time frame. This means the need for longer term backup, which will be pretty expensive to build and to operate. If this lull continues for a couple days, then the deficit of all those days will need to be available beforehand.
The reverse is also possible:
This excess then need to be absorbed or curtailed. In the former case, an insane amount of backup will be required. In the latter case, curtailment is surely an option. Remember, curtailment was framed as wind energy “balancing” the system (although that sounds pretty neat, this was not really appreciated by the windmill owners who demanded extra financial compensation for their, ahem, “balancing”).
Dynamic demand response will probably have less impact, just compared to the energy that potentially needs to be displaced.
Concluding, both options presented in the tweet will have their share of problems, certainly when the infilling is done by for example pumped hydro or batteries instead of by dispatchable power sources like gas. Then just looking at the amount of energy that potentially needs to be stored, I think it is the first option that is the easiest/better one when the aim is to balance the grid at all times.