First Post
I recently commented on an article which perpetuated some of the myths common when Concentrated Solar Power (CSP) is discussed. Here’s what I commented, along with two replies to my post:
Counting the “thumbs-down” to my post and the “thumbs-up” (5+7+3 = 13) and the opposite reactions (1+1+0 = 2), it’s clear that the consensus is that my comment was “deeply ignorant”.
Second Post
When I read Marcan’s reply, I responded.
Full details backing my comment are supplied in my reply - but the publication disallowed the reply, as indicated by the “Removed” tag in Disqus. In short, the publication censored me.
Setting the scene with two topics
It is common, in South Africa, to celebrate the 1st of September as “the first day of Spring” (Scientifically, this day is actually about three weeks later.) Taking the South African convention as a point of reference. we can divide the year into its seasons matching the months of the year.
Taking the first month of Spring to be October would be more accurate, but that would make December the last month of Spring. I suspect just about every South African would describe December as a Summer month, so I will use the convention described in the table above.
At the beginning of 2024, South Africa had 500 MW of CSP installed and operating, as listed in this publication. Excluding Green Duba 1 because it is (at the time of writing) still under construction, but not excluding Redstone (100 MW) because it is destined to be completed in 2024, we have the 600 MW listed in the table below.
Note that most of the CSP exploits so-called trough technology. Only Khi Solar One is tower technology, the technology used by Redstone.
Justifying my rebuttal
In the last sentence of Marcan’s response to my post, note how he is careful to hedge his bets, stating that “Redstone CSP can produce power practically 24 hours a day with sunny weather” (my italics), compared to reality - which is that Redstone does not produce consistent power 24 hours a day. It’s almost comical that Marcan’s careful choice of words neatly negates the claim he’s implying. My post indicated that there is a reason why I claim CSP is fundamentally fatally flawed. The absence of the heat source for roughly 12 hours of every 24 is not a feature, it’s a bug!
It is easy to demonstrate Redstone’s limitation. In the table above the plant’s expected annual output is given as 466 GWh (466,000 MWh). The largest value I can find predicts that this plant’s output will be 480,000 MWh (NB, not “MW” as stated in the link!). (I found confirmation of 480 GWh here.) Using the last value, we can ask for how long, per day, Redstone could produce 100 MW, the installation’s rated capacity.
If Redstone produced 100 MW for 24 hours of every day of the 365 in a year, it would be producing 100×365×24 = 876,000 MWh annually. If we divide the predicted value by the maximum value, we get 480÷876 = 0.548 or 54.8%. In short, if Redstone produced 54.8 MW every hour of the day, all year long, the annual output would be 54.8×24×365 = 480,048 MWh. The apparent extra 48 MWh is because I rounded 0.7945… up to 0.8 and I can use a percentage and an output interchangeably because of the fortunate fact that Redstone is rated at 100 MW.
Of course, Redstone will not be used to output a steady 54.8 MW. Without doubt, at times of peak demand and or at times of peak sun, Redstone will deliver it’s maximum possible output. Taking this to be 100 MW, Redstone could only do this for 0.548×24 = 13.2 hours per day (on average). So, there can be no doubt, the plant’s specifications declare the fact that Redstone can only produce 100 MW for h hours per day (0≤h≤13.2) with a reduced amount in the remaining (24-h) hours. For example, if, on average, one output’s 100 MW from Redstone for 8 hours per day, the average output in the remaining 16 hours will be (54.8×24 - 8×100)÷16 = 32.2 MW, roughly a third of Redstone’s nominal output.
My point is, the likes of Redstone are set up to fail. They fail because CSP’s heat-source is absent for half the time. I use the word “fail” to describe the fact that these plants are unable to produce their rated capacity for a full 24 hour period. The company with a 49% share in Redstone makes this declaration: “Redstone with 12 hours of full-load energy storage will be able to reliably deliver a stable electricity supply to more than 200,000 South African homes during peak demand periods, even well after the sun has set.” Although vague, I think the implication is clear. On average, output can be 100 MW for 12 hours - and there should be some heat left to supply an average of (54.8×24 - 12×100)÷12 = 9.6 MW during remaining hours of the day.
What does the data say
Back-of the envelope calculations can be used to drive a point home - but reality is messier than can be displayed using such calculations.
I’ve plotted the data Eskom offers since January 2024. I’ve used these downloads, in 2024, to plot the performance of “renewables” for the most recent month and the month before that. The downloaded dataset goes back some five years so historical data is available if needed.
Here’s the output from CSP in May and June 2024. Note that Redstone was still under construction at this time so the plot is displaying the output of 450 MW trough technology and 50 MW of tower technology.
At the end of Autumn, the beginning of Winter, it is to be expected that the sun will be weaker and therefore CSP’s output less than when the sun is stronger. Compare May’s output to June’s. There’s a clear decline.
The data in this plot can be rearranged so that the hours are in sequence and the data points within an hour are for the days in the period.
We see that there is virtually no output before 9 in the morning and after 10 in the evening. We also see that output is peaking after the sun has gone down. CSP might not be able to contribute to the morning peak demand from households but it can store energy for use after the sun has gone down.
Now imagine if “the sun never went down”, that the heat source did not disappear for half the day, every day.
Redstone was synchronised to the grid in the middle of September 2024. I intend plotting the output from CSP for November and December 2024. I want to stay within 2024 but I want to view the performance of CSP in the late Autumn, early Summer, a time when the sun is at its strongest.
First, we need to establish a point of reference. Since Eskom does not publish individual plant output’s, only the total output of technologies, one cannot view “Redstone’s output”. Therefore, here are the plot's for November and December 2023.
There is clearly a daily pattern in this output from 500 MW of (mostly) trough CSP.
Now that we know what the output should look like without Redstone, we can examine the output when Redstone is included, at a similar point in time.
Frankly, I don’t see much difference - the average changed from 188.63 MW to 212.96 MW, an increase of 24.33, about half what one would expect (54.8 MW). The pattern during a day is also much the same.
Note, however, that while the 2023 peaks are at about 400 MW, the 2025 peaks are at about 500 MW, the extra 100 MW presumably coming from Redstone.
We see the same pattern with Redstone as we did without. More or less constant output once the sun is high enough with output continuing after sunset, although declining as 10 o’clock approaches, at which point CSP effectively shuts down for the night. Yes, there’s output, but roughly a tenth of the output around midday.
If only the heat source were available for all 24 hours of a day!
Conclusion
Please scroll to the top of this post and read the comment I made. Then read the two responses. The links supplied in my reply showed what I show above - just not as concentrated. So my question is: Why was my reply censored by the publication concerned?!