Blowout week 105 linked to a recently-completed study from the Solar Trade Association which reached the following conclusion:
…. solar together with storage and flexibility would cost roughly half that of (Hinkley Point Unit C) over the 35 year lifetime.
And a comment posted by robertok06 had this to say about the Solar Trade Association study:
I have hardly read more BS in one single document…
Here we take a closer look at these contrasting viewpoints.
With a capacity of 3,200MW and at the 91% capacity factor assumed by the Solar Trade Association Hinkley Point C will generate a constant 2,912 MW of power and 2,912 * 8760 = 25,509,120 MWh in a year of operation. So to compare solar directly with Hinkley we need 25,509,120 MWh of annual solar generation. What does the annual solar generation curve look like at this level?
We begin with Figure 1, which plots hourly embedded solar generation in UK during 2014, the last year for which complete records are available and one which I assume will be representative of solar output in coming years. The data are from the “UK Grid Graphed” data base and were supplied by Neil Mearns:
Figure 1: UK solar generation, 2014, hourly data
Solar in UK generated only about one-seventh as much power in 2014 as Hinkley will generate in a year, so to match solar to Hinkley output we have to factor 2014 solar generation up by seven. Then another adjustment is necessary. Figure 1 is right-skewed because installed solar capacity doubled from 2,691MW to 5,131MW between the beginning and end of 2014, as reported in solar PV statistics. To remove this skewness I normalized all the 2014 generation data to 3,911MW, the average installed solar capacity in that year. Figure 2 compares the solar generation curve after application of these adjustments with the baseload generation from Hinkley (note that all the solar peaks are separated by nighttime periods of zero generation, although this is difficult to see at this scale):
Figure 2: UK solar generation factored up to match annual generation from Hinkley Point C and adjusted for the increase in installed solar capacity during 2014, hourly data. Approximately 27GW of installed solar capacity is needed.
The question that now arises is how to compare the baseload generation from Hinkley with the highly irregular solar generation, which varies over day/night ranges approaching 20GW. The Solar Trade Association assumes that with storage and “flexibility” the solar curve can be flattened out to the point where it can be considered dispatchable , but because no specifics are given I have had to make my own estimates of what would be needed to flatten it. I did this by calculating how much energy storage would be needed to convert the solar generation into baseload generation at the same level as Hinkley, which is the only way I could see of making an apples-to-apples comparison. I ignored potential contributions from “flexibility” partly because I had no way of estimating how large they might be and partly because I doubt they would be significant.
First I estimated the amount of storage needed to remove the diurnal variations. Figure 3 plots the data for July 24, 2014, which having the largest day-night generation change can be considered the worst case. Average generation during the 24-hour period is 7,100MW, and storing the surplus power generated in the day for re-use at night to obtain 24 hours of continuous 7,100MW output requires 3.4GWh of storage capacity. This isn’t a prohibitive amount, and because demand is higher during the day than at night the actual storage requirement would probably be lower. So we can reasonably assume that diurnal variations in solar output can be smoothed out without a large cost penalty.
Figure 3: UK solar generation on July 24, 2014, showing the storage and release requirements needed to smooth out diurnal variations, half-hourly data.
But after removing the diurnal variations we are still left with the daily solar generation curve shown in Figure 4. The large variations between winter and summer generation must also be smoothed out to convert solar into year-round baseload generation, and a substantial amount of energy storage will obviously be needed to do it:
Figure 4: Average daily UK solar generation needed to match Hinkley annual generation, 2014
To estimate how much would be needed I calculated the daily solar surpluses and deficits relative to a constant 2,912 MW baseload level. These are shown in Figure 5:
Figure 5: Daily solar surpluses and deficits (GWh) relative to the constant 2,912 MW baseload level.
Then I accumulated the surpluses and deficits to calculate how much energy would have to be in storage at any time to obtain 2,912MW of constant output throughout the year. The results are plotted in Figure 6. There is a requirement for 7 terawatt-hours of storage, roughly the equivalent of eight hundred more Dinorwigs, or if you like two hundred and thirty Coire Glases.
Figure 6: Energy in storage needed to maintain a constant 2,912MW of baseload solar output throughout the year
How much will this 7 TWh of storage add to solar costs? I didn’t bother to make an estimate because the question is academic. There is no way this much additional energy storage capacity could possibly be installed in UK by the time Hinkley begins operations, if ever.
Yet the Solar Trade Association comes up with cost numbers that allow for storage. How much seasonal storage do they allow for? None. They simply assume that seasonal solar variations will be “complemented” by wind, which blows more strongly in the winter, to the point where they “more closely match electricity demand”:
“Our analysis does not include inter-seasonal storage to match Hinkley’s winter output. The storage and balancing aims to both smooth and shift the solar output to more closely match electricity demand. From a broader renewable energy perspective, solar generation can be complemented by wind power whose output peaks in the winter months.”
Now let us see what adding wind to solar does. I began by taking daily wind generation data for 2014 (again from UK Grid Graphed), factored them by 0.9 to make annual wind generation equal to annual solar generation and combined the two. Figure 7 shows the results in a stacked bar chart. Adding wind to solar indeed reduces the winter/summer range but the daily generation curve is now much more erratic than before, which will make it more, not less difficult to match generation to electricity demand:
Figure 7: Solar generation plus an equal amount of wind generation, 2014 daily averages.
And how much difference does the reduced summer range make to storage requirements? Using the Figure 7 data I accumulated the surpluses and deficits to calculate how much energy would have to be in storage at any time to maintain 2,912MW of baseload solar output (plus an equal amount of wind) throughout the year. The results are shown in Figure 8. Combining solar with wind halves the storage requirement from 7 TWh to 3.5 TWh, but 3.5TWh is still more than 100 times current installed UK energy storage capacity and the equivalent of roughly four hundred more Dinorwigs:
Figure 8: Energy in storage needed to maintain a constant 2,912MW of baseload solar output throughout the year with wind contributing.
So what to make of the Solar Trade Association’s claim that solar is cheaper than Hinkley nuclear? Well, robertok06 was right, it’s BS. Barring miraculous breakthroughs in energy storage technology within the next few years or a populace that is willing to freeze in the dark when the sun doesn’t shine it is simply not possible to replace baseload generation from Hinkley with intermittent solar power.
A final point. The UK Solar Trade Association is in the business of selling solar systems, and its claim that solar is cheaper than nuclear is clearly designed to help it sell more solar systems. This I believe makes its report a marketing document subject to UK Advertising Standards regulations, one of which is:
“You must describe your product accurately. This means if you make a claim about your product, you must be able to prove what you say.”
In this case the Solar Trade Association is unable to prove what it says. Does this put it in violation of UK advertising standards? Feedback is requested.