Decreasing demand for natural gas power thanks to wind and solar?

At the beginning of this month, I found an article in a Flemish newspaper that seemed to be inspired by the gas deficit alarm from the UK National Grid. It was titled “Renewable energy gets us though the winter cold”. This is how the story goes: the very cold February of this year did not result in record natural gas consumption. We used less natural gas in February 2018 than six years ago in February 2012. The demand for natural gas from households and industry stayed more or less the same. The difference was the demand for natural gas for electricity production, which was lower in 2018 than back in 2012.

Conclusion of the article: there was no record natural gas consumption in February because renewable energy grew in the last six years and this increased share lowered the demand for natural gas power. Because of this, we currently don’t need as much natural gas than we would only six years ago and there was no shortage in our country.

This is the part of the article that explained it (translated from Dutch):

Where does that lower gas consumption from yesterday compared to the extreme cold of 2012 comes from? ‘The current natural gas consumption by individuals and the industry stayed about the same. The main difference is in the consumption of power plants that use natural gas to generate power’, continues Remy [Fluxys spokesman Laurent Remy]. “Those power stations required 260 gigawatt-hours of gas per day in February 2012, yesterday that was only 158 gigawatt-hours.’

No worries

And this thanks to the rise of renewable energy, which is much more widespread today than six years ago. Much more power is generated by windmills and solar panels, therefor the demand for power from gas-fired power plants declined. Figures from grid manager Elia even show that January was a record month for wind energy. In February there was a lot of sunshine.

Hurrah for wind and solar! Even in winter these are decreasing demand for power from gas-fired power plants…

My first thought reading this article was: “No way!”. Solar energy will not exactly make much of a dent in energy capacity in February. Days are short (yet lengthening) in this time of the year, so solar energy will be rather limited and only a fraction of what the capacity in summer months. This leaves us with wind energy that must be responsible for most of this achievement.

Also, it seems nonsensical because natural gas is a power source that fills in the gaps. ANY gap from ANY other power source(s). It seemed a very bad metric for measuring how much the demand of gas power decreased by wind and solar. It also depends on so many variables, not just on how much gas was necessary to produce electricity on a certain day or during a certain month.

Firstly, is the comparison between February 2012 and 2018 actually a good one? I went to the KMI website and looked at the summary pages for February 2012 and February 2018. The two months were quite similar when looking at the temperatures and sunshine:

  February 2012 February 2018
Average temperature (°C) 0.7 0.8
Average maximum temperature (°C) 3.8 4.4
Average minimum temperature (°C) -3.2 -1.9
Sunshine hours (hours:minutes) 95:25 149:28

It was colder and darker in 2012 compared to 2018. Solar would be expected to do relatively better in 2018 compared to 2012.

The minimum temperature was -13°C in 2012 compared to -8.7°C in 2018. The last sentence in that article reassure us that:

… we can easily cope with temperatures until minus ten [°C].

Yet we had four days below -10°C in February 2012 and none in February 2018. If we look at temperatures, then we would expect somewhat more demand for power in 2012.

Okay, those two months are not wildly different, but 2012 will surely be at a disadvantage.

Secondly, how much is the difference in gas consumption for gas and wind & solar? The electricity production information of those two months could be downloaded from the Elia website. In the following back-of-the-envelope calculation I will use the units that Elia is providing (MW). They report the capacity that was needed to produce the needed energy at a certain time. It is a bit weird though to look at production in this way, but it has some advantages (for example, the relation with installed capacity becomes much clearer).

That being said, this is the comparison between the two months:

Energy source February 2012
February 2018
Coal 735 0 -735
Fuel 10 0 -10
Gas 3,800 2,372 -1,428
Nuclear 5,452 4,967 -485
Water 199 125 -74
Wind 114 550 +436
Other 640 764 +124
Total 10,950 8,778 -2,172

The average capacity of natural gas power was 3,800 MW in February 2012 and 2,372 MW in February 2018. That is quite a large decrease of 1,428 MW.

Wind power had an average capacity of 114 MW in February 2012 and 550 MW in February 2018. That is an increase of 436 MW.

Solar data is a bit more complicated because there is no data available for February 2012 (also not in the Solar energy section). Solar data is however conveniently placed in the “Other” category. The values for that category are not that big, so solar probably wouldn’t make much of a difference (what is expected anyway). The current capacity data for solar could be found separately on the Elia website and is on average about 290 MW for February 2018 (this for a “good” year).

The “Other” category had an average capacity of 640 MW in February 2012 and 764 MW in February 2018. If the rest of the “Other” category stayed approximately the same, then we could assume that solar had an average increase of 124 MW. This will probably not far of reality, knowing that the 290 MW capacity of February 2018 was a good month for solar and the maximum difference could only be 290 MW (in the case that there was no solar energy in 2012, which is very unlikely).

If I add the differences for wind and solar, then the result is 560. That is somewhat more than one third of 1,428. So, if this is true, then the replacement by wind and solar could never be the “main” difference between 2012 and 2018. Even if we assume that there was no solar power at all in 2012 (which is unlikely), then it only amounts to half that value. Between half and two thirds of the decrease had a different reason than wind and solar.

Remarkable is that the share of wind and the “Other” category went up, all the others, including the total, went down. The total capacity went even down by 2,170 MW. I understand that our energy consumption decreases over time, but this seems quite a lot.

But are these numbers a fair comparison between the two years? These are only the central energy production data, that is not the complete picture. Belgium is connected with its neighbor countries and there is import as well as export of energy. It is not clear where to find that data on the Elia website, but I found the grid load data (which is all electricity injected by power plants minus energy needed for pumped storage and balanced with import/export, excluding decentralized generation). So theoretically, when I subtract the central production of the grid load data, that should give me at least a rough indication of the import/export balance. If I do that, I get an average export of 138 MW in February 2012 and an average import of 955 MW in February 2018, resulting in an average difference of 1,077 in stead of 2,170. So we were self-sufficient in 2012, but not in 2018.

If my assumption is true, then it gives roughly a capacity of 1,000 MW that is not accounted for in the comparison. Where would that come from in the new situation if we would have produced it ourselves as we did in 2012? Coal is no option, we stopped using it for power generation in 2016. The fuel and water share are too small to compensate for that. Also no new nuclear plants will be build. That leaves us with gas that would close the gap, so if we want to do a correct comparison with 2012, then we would probably have 2372 + 1000 MW in February 2018 and then the neat theory that wind and solar prevented the use of gas in power plants becomes a lot less strong.

Finally, 2012 is a leap year while 2018 is not, so the figures for 2012 are for 29 days and 2018 only 28. Since this is just a back-of-the-envelope calculation, I subtract 1/28 and then I get a difference of only 693 MW between February 2012 and February 2018. The difference is suddenly a lot less than what we started with. Knowing that the temperatures in February 2012 were somewhat lower than in February 2018, it doesn’t look that good for the theory.

It is much more complicated than the simple claim that wind and solar diminish the demand for electricity from natural gas power plants. There are power sources that disappeared in the meanwhile (coal), some were less used (liquid fuel, gas, nuclear and water), there were differences in import/export and also a lower consumption of energy over time complicates things.

The big question however is whether some kind of relation could be found between wind/solar and gas power? The article suggests that it was the combo wind and solar that somehow made the difference. So how relevant is an increase of wind and solar capacity for the demand of natural gas power? If wind and solar increase, can we see the effect on demand for natural gas power? Looking at the February data for all years between 2012 and 2016, this is the result:

Wind energy gradually goes up, strong until 2014, later less steep. Gas power is strongly going down in an erratic way. It is probably correlated with something else. My guess would be this:

Nuclear power is an almost mirror image of gas power. When more nuclear capacity is available, less gas power capacity is needed. When less nuclear power capacity is available, then more gas power capacity is needed. Which is not exactly surprising. Gas fills in the gaps. If the gaps are big, then much more gas power will be needed. If the gap is small, then less gas power will be needed.

Building on that, wind and solar energy are not exactly reliable. If February 2018 was a good month for solar, then it is possible that there are also bad months, meaning less replacement of gas power by solar. It is not a given that February is sunny or windy. January was according to the article a good month for wind, which means that February was not as good. It is also perfectly possible that it could be a bad month for solar as well as wind and then more gas would be needed for electricity production and less reason to cheer for solar energy.

Concluding: did the capacity of wind and solar increased between 2012 and 2018?


Did the capacity of gas power decreased in that same period?

It sure did.

Were wind and solar the main reason for this decrease?

Not very likely.

Assuming for the sake of the argument that the impact of wind and solar was as claimed, could it be relied upon?


In fact, they were lucky that the nuclear phase-out didn’t happen on schedule and solar had a good month, otherwise the share of gas power relative to 2012 would be much higher, resulting into a potentially very different headline…

7 thoughts on “Decreasing demand for natural gas power thanks to wind and solar?

    1. trustyetverify Post author

      Sure is. That why the post took so long. There was a lot to unravel and the information was shattered over different pages on different sites. In the end, I just decided to stop working on it and put it online as it was. I learned a lot in the process though.

      Liked by 1 person

  1. chrism56

    How much power came across the border? Electricity imports can make the difference. The other factor is have any energy intensive industries closed? Like vuurklip, says, the devil is in the detail.


    1. trustyetverify Post author

      The import and export was already mentioned in the post. These numbers were rather hard to find.
      I have however no data on closure or moving away of heavy industry, so I don’t know whether this had an effect. Even without that data, I am sure that the effect was much less than was suggested in the article. I reduced the difference to a much smaller number and didn’t even account yet for the difference in temperature and gradually lowering of the demand, which would make the difference even smaller. Maybe even less than the difference of what was produced by wind and solar.


      1. chrism56

        Sorry, you are correct about mentioning interconnectors. I missed that.
        When I went to the Elia site and looked at Interconnection for February 2018, the data is just Belgium French border, Belgium Luxemburg ,and Belgium Netherlands. These seem to show the average flow was about neutral. I can’t find what the flow across the German border was.


        1. trustyetverify Post author

          Not sure what you were looking at. When I look at the graphs of the import/export balance on the page of the cross-border physical flow, then the days I looked at were mostly or totally in negative. There seemed to be a lot of import going on. I only looked at the first week and a half or so from February 2018, the graphs were building extremely slow and sometimes there were some visualization issues (didn’t see the graph lines, but could see the values of the data point when hovering over them). That gave me an idea. When I have a lot of spare time I will try to transcribe the data points of some of those graphs and compare them with load/central production data of the same period to see if my theory of “load – central production = in/export balance” is correct.

          You will not find cross-border flow with Germany, because there is no interconnection with Germany yet (it is anticipated to be ready at the end of 2019).


        2. trustyetverify Post author

          I had a quick look at the import/export balance graphs of every day in February 2018. I noticed that during 14 days there was only import and no export at all. The other 14 days had some export in it, but in 4 days just for a very short time. Looking at all the graphs learns me that, indeed, there was a lot of import going on in February 2018 and certainly not being “neutral”. The average 955 MW import seems rather plausible, it will probably not be far off.

          I took some screenshots of the import/export balance of every day, then removed the vertical red line representing the current time. Then added a thick horizontal red line at the zero point, meaning everything below that is import and everything above that is export. Graphs without a thick horizontal red line are days with a lot of import (graph line is so low that the zero-line doesn’t even come into view).


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s