Can’t get enough of the (meaningless) metric of record contribution of solar and wind to total load. This time, I will go hunting for another such record myself, using the exact same data that leads to that cheered record.
Remember from previous posts that the claim was made that Belgian solar and wind contributed 45% of total load on July 28 and that it was brought as something significant. Also remember that July 28, 2018 is a Saturday and that there was coincidentally a lot of solar PV and wind energy production, leading to a big contribution of solar and wind to total load. When it comes to intermittent power sources, that record is basically meaningless.
There was also this graph from Elia (our network manager) that showed other similar records of the last 11 months (see also this post):
As I explained in my first post on the subject, it suggests a steady increase. Until you realize that these are spot counts of coincidental high values, it is no less than cherry picking.
But if our Minister (and the network manager) are allowed to cherry pick, so am I! Solar and wind are intermittent energy sources and the record values are directly related to the failure to follow demand by those sources. Therefor, if there are records of maximum contribution (low demand at times of high production), then there are also records of MINIMUM contribution (huge demand at times of low production).
Let’s look at the data. Elia publishes a lot of Belgian electricity data, like wind, solar PV and total load.
First I wanted to see if I could reproduce the maximum records graph with this data. It was more challenging than I thought it would be. It is a huge chunk of data and not all data is in the same format. So most of the time went in reformatting the data so it could be compared. This is what the result looks like:
It looks very similar. There are some minor differences, but the trend is basically the same and all the elements point to the same thing. My guess is that Elia has more fine-grained data, I could only work with the average of quarter-hour data. The important thing however is that now I got all the data in the same format, showing the minimum records is a piece of cake. This is what Calc spitted out at the end of this process:
Ooooooops.
Where is the contribution of solar and wind in that graph?!?!
If you look very closely you will see a hint of blue glued to the x-axis on September 2017, October 2017 and June 2018. These are the only three values that are above 0.5% (but below 1%) contribution to total load. No yellow is to be found, the contribution of solar was in any of those cases a nice round figure.
The lowest contribution is from January 20, 2018 at 19:15 with a value of … wait for it … 0.016% of total load.
Where were the headlines trumpeting this record?
These minimum records are reached at weekdays and evenings. The total load values are much higher than those of the maximum values (I used the same dimension for the Y-axis). Which is not surprising. The higher the total load, the lower the contribution of the power sources. A record minimum will most probably be established when total load is high (dividing a small number by a bigger number will give a smaller ratio). And vise versa. A record maximum will be reached when high production and low total load coincide.
More importantly, it is also clear that, contrary to the maximum values, there is no real trend upwards. Minimum contribution values are crashing down in winter and crawl back up in spring/summer. If you increase the capacity of solar and wind, then at the top end you end up with huge fluctuations. Which is not really surprising. Multiplying a bigger number will give an even bigger number.
Not so at the bottom end. There, increases have much less influence. Multiplying a small number still gives a small number.
That is not hard to understand. For example, these are the extremes measured in the period October 2017 to July 2018;
- minimum production solar + wind: 1.78 MW (January 20, 2018 at 19:15)
- maximum production solar + wind: 3,800.22 MW (July 28, 2018 at 15:00)
- minimum total load: 6,818.42 MW
- maximum total load: 13,512.73 MW
If we would double the capacity of solar and wind, then we would have a production only 3.56 MW (+ 1.78) at the bottom end and 7,600.44 MW (+ 3,800.22) at the top end.
If we would increase the capacity tenfold, then we would have a meager 17.8 MW (+ 16.02) at the bottom end and 38,002.2 MW (+ 34,201.98) at the top end!
This means that wind and solar will only save 16.02 MW that other power sources don’t need to fill in (that is neglectably small) at the bottom end, but would produce almost three time the maximum total load at the top end.
The maximum record metric surely has a huge PR value, but it is a real bad metric for looking at fitting in intermittent energy sources in the current grid.
Trust, yet verify 