Power from the midnight sun

A solar power station in Spain runs for 24 hours, and the overhyped intermittency problem.

A news item to cheer up any of our readers who worry about the intermittency of wind and solar energy. (Tidal and geothermal, not so much).

The Gemasolar thermal solar tower plant near Seville achieved 24-hour operation on a day in July, using molten salt heat storage. At 20MW, it’s a real pocket-sized power station, not a tiny pilot. It achieved this  only one month after commissioning, so the design clearly works. The average capacity factor will not of course be 100%, especially in winter. It doesn’t have to be, and every power plant is offline part of the time for maintenance. The molten salt technology gets solar thermal close to the availability of the conventional or nuclear pack, so problem apparently solved, technically.

Costs? They are not telling. and solar energy in Spain still relies on public subsidy through a generous feed-in tariff.

This was reduced in 2008 causing an investment stall, but things seem to be picking up again. Since CSP is a young technology, the current costs don’t mean very much looking forward. Economies of scale – mass-produced heliostats (steerable mirrors) and standardised receivers and turbines – will bring them down. The long-term problem is the land take; it’s not like wind where cows graze up to the pylon base. I don’t myself see why you couldn’t raise the heliostat supports and intercrop with vegetables or parked cars – the partial shade isn’t a big drawback in climates good for CSP.

Molten salt isn’t the only scheme around. Gemasolar’s steam plant needs water, a constraint in hot countries. If you use air as the working fluid instead, you can today get it up to 900ºC or more in the receiver. This is hot enough to drive a gas turbine, aka known as a Brayton cycle. Even higher temperatures of 1200ºC or so are more efficient thermodynamically. This is not directly relevant to a free resource, but it’s desirable to reduce the number and area of mirrors. Pilot towers are being built in Spain, Israel and Australia that combine hot air with gas or biofuel. You can drive up the temperature, and above all have a 24-hour combined cycle that doesn’t need water, though it’s higher carbon than the hot-salt system. There are ideas for heat storage too but 1000ºC is a hard target. One plus here is modularity – you can build small as well as well as large units for remote mines and city rooftops.

Go, go, go. We need all of this, urgently.


How serious is the intermittency worry anyway? It’s not the killer objection some dinner-table conservatives would make it. For 15 years, electricity professionals have been saying that integration of intermittent sources up to 20% of supply is manageable without reducing service standards. The integration has costs, but also benefits – a more diverse and fine-grained generating park, linked to the required better grid, can be more reliable and flexible overall. Multiple sources cited here, including 2006 UK estimates for grid integration costs. Denmark is now close to 20% wind supply, and SFIK manages fine.

This chart for SE Australia in 2010, taken from here, tells the story.

(Sorry about the poor resolution, better pdf here.)

The variation in wind supply (about 4% of mean demand) is dwarfed by the 20% daily variation in demand, so it’s lost in the overall balancing. But what would happen if you scaled up? This chart from the same source plots total demand against wind’s capacity utilisation factor, on incommensurate scales.


If you scaled wind up to match 100% of demand at the average utilisation factor of 30%, superimposing the blue and red, you would need almost as much conventional capacity in reserve. Not a tragedy since you have it already.

My calculations for a sample month (April 2011) suggest that the balancing capacity (coal, oil, gas, hydro, imports, or storage) on this thought experiment would need to be 98% of mean demand, or 30% of nominal wind capacity, and it would supply about 40% of total power summed over the month. The true carbon intensity of a wind-heavy power system includes this smoothing overhead, so it’s low but far from zero. If you look at marginal impact, the story is different: you are starting, unless you are France, with a mainly fossil mix, so the marginal impact of wind is a near-total displacement of fossil carbon emissions.


Note: My calculations are from the data set supplied here, see my spreadsheet. (Warning: 5-minute sampling of 24 wind farms, so big file.) I can’t be fagged to repeat them for all 12 months, but anyway the numbers won’t generalise to other countries or continents. It’s only an approximate simulation anyway – a real expansion would be more dispersed and smoother in output than the grainy current installed base. A lot depends on how wide the reference area is; on the scale of a continent with a good long-distance grid, there’s natural smoothing from the cycle of weather systems. So my 42% balancing is an illustrative upper bound, not a central estimate.

Author: James Wimberley

James Wimberley (b. 1946, an Englishman raised in the Channel Islands. three adult children) is a former career international bureaucrat with the Council of Europe in Strasbourg. His main achievements there were the Lisbon Convention on recognition of qualifications and the Kosovo law on school education. He retired in 2006 to a little white house in Andalucia, His first wife Patricia Morris died in 2009 after a long illness. He remarried in 2011. to the former Brazilian TV actress Lu Mendonça. The cat overlords are now three. I suppose I've been invited to join real scholars on the list because my skills, acquired in a decade of technical assistance work in eastern Europe, include being able to ask faux-naïf questions like the exotic Persians and Chinese of eighteenth-century philosophical fiction. So I'm quite comfortable in the role of country-cousin blogger with a European perspective. The other specialised skill I learnt was making toasts with a moral in the course of drunken Caucasian banquets. I'm open to expenses-paid offers to retell Noah the great Armenian and Columbus, the orange, and university reform in Georgia. James Wimberley's occasional publications on the web

17 thoughts on “Power from the midnight sun”

  1. What the world needs is a good battery. “Battery” is not meant literally; it means any system or component in the energy storage trade. If I could store sunlight for 12 hours and heat for 6 months, the utility bills for my little house in MA would be mighty small.

    Way back (around the time of Enron’s collapse, I think) I read a story somewhere about a software company in Bangalore which got around the problem of intermittent electric service by filling a warehouse next door with standard automobile batteries. The intermittency was not due to variable sun or wind, of course, but the solution would have worked equally well if it was.

    Every time I see a “noisy” signal, I keep thinking “There ought to be a capacitor across that sucker”. A big capacitor at the power source is one approach. Lots of little capacitors out at the tendrils of the grid would be another. One advantage of the second approach might be that the distribution grid would not have to be sized for peak demand, only for average demand. The obvious disadvantage is cost — for now.


  2. Thanks for posting this, James. I also like your point that partial shade can be good for crops.

    The obvious place to start in the USA is hotter areas, especially as (IIRC), the peak electricity loads are on hot, sunny, windless afternoons.

  3. “Costs? They are not telling. and solar energy in Spain still relies on public subsidy through a generous feed-in tariff.”

    I’d like to know if *any* large-scale energy production in the developed world is not subsidized.

  4. Seville often registers the highest temperatures in Spain, a hot country to start with. They exceed 100F and frequently reach 110 in the summer. Would this 24h cycle possible in more temperate zones?

    Whimsical thought: So the hotter the planet is, the more solar energy we get? So solar plants act as negative feedback loops for world climate?

  5. One sympathetic article describes the capital requirement as “$33 per average Watt delivered, several times higher than the cost of wind, geothermal, or nuclear”.


    In other words this is a massively expensive approach, way more so even than other heavily subsidized generation methods. So, not at this point either an economically viable or relevant technology, or a contribution to the general welfare.

  6. “The molten salt technology gets solar thermal close to the availability of the conventional or nuclear pack, so problem apparently solved, technically”

    You think this is “close” to the availability of a nuclear plant? Many nuclear plants operate continuously, at full power, for over _2 years_ at a time, between refuelings. Overall industry-average capacity factors run over 85% over the entire year. Coal plants essentially run till they break down. Which is not very often at all.

    There are standby emergency diesel generators at some REAL power plants that produce more power than this boondoggle produces.

  7. Neville: no new technology starts cheap. The argument (textbook orthodoxy) for subsidy is economies of scale. Look at the cost curve for solar PV. Give me some reasons why this won’t apply to CSP. $33 per kw/hr is not bad for – remember – the very first industrial implementation of the technology. The company estimate a 20% reduction for their next two plants. I know they laughed at Bozo the clown as well as Columbus, but does this look to you like Bozo?
    Freddy Hill: SFIK nobody is talking about solar CSP anywhere north of Madrid. The industry is indeed looking at Africa: the Sahara is enormous, empty, continuously sunny, and close to Europe. There’s already a substantial interconnector across the Straits of Gibraltar. The absence of water points to the Brayton cycle.
    The concentration in Spain of CSP plants near Seville, on high-quality farmland, is strange. Why not put them in empty Estremadura? The answer must be either water or pork-barrel politics: Andalucia is a richer region, and mayors nearer the capital have more clout with the Junta.

  8. rxc: The availability of a salt-storage CSP plant at rated output is mainly the result of economic decisions how much salt storage and mirror area to install in relation to the power train capacity, not a limitation of the technology. Gemasolar decided on 15 hours storage and an very large heliostat field, so they get high availability and their 63% capacity factor over the year. This can only be an estimate as the plant was commissioned in June, and they don’t have a year’s data on pump reliability and so on.

    The capacity factor (actual output over time as a percent of theoretical maximum) is not the same thing as availability (percent of time potentially operating). As with all renewables, you are getting partial output under most weather conditions, so the old yes/no availability metric is not quite right. It might be better to think of a CSP plant as the sum of a sheaf of seasonal plants with different peak ratings, varying with theoretical clear-sky insolation, and actual availability varying with cloudiness offset by storage.

    It looks to my amateur eye as if Gemasolar may have over-engineered the plant to make their headline point. If the grid has a big air-conditioning load in summer, and a large wind park with maximum output in winter, the economics might dictate tuning your CSP plants for summer output and sacrificing some winter capacity. These are so-what optimisation issues we can safely leave to the professionals.

  9. As James Wimberley points out “Neville: no new technology starts cheap.”, I’d like to add that no “old” technology is cheap either. We do not lump the costs of oil, gas, coal, nuclear production into a clearly visible form, but surely none here would argue that wars for oil, subsidies for fossil fuel extracting, refining, production, delivery are COSTS which are subsidized by the Public. Nor do we calculate and compare the costs to the public of negative impacts on health costs, water resources, air quality, ozone, wildlife, etc.

    I was shocked to hear a report that after a century of dependence on oil and gas, there is still only a capture and transmission of 35% of energy from those sources. We spend an awful lot in blood, treasure, health, environmental contamination for such a small return, in my opinion. As companies claim record profits, one would think they’d invest in their own processes in order to effect at least 90% of potential energy from their product.

    I’d like to see the comparisons in negative environmental and health impacts of solar vs fossil fuels and nuclear from start to finish along with a comparative table of public subsidization. I celebrate the great success and creative materials application by Gemasolar!

  10. Any real environmentalist must be appalled at this report. How much land has been destroyed by this boondoggle? Judging from the photo, it appears to be quite extensive. This carpet of mirrors will eliminate most forms of life from the afflicted space. The dense shade of these closely spaced mirrored roofs precludes most plant life, even if the mirrors are elevated above the surface, as the author suggests. With little or no primary productivity from plants, the land will harbor few mammals, birds, reptiles or even insects. Ask yourself this: if an equivalent amount of open land were being converted to a paved-over industrial site for any other purpose, would you support it enthusiastically? I suspect not.

    Belief in CO2-mediated anthropogenic global warming catastrophism is a religious conviction, not a scientific conclusion. Elimination of CO2 has become a compulsion for true believers, just as the avoidance of ritual pollution consumes the imaginations of the acolytes of several other more traditional religions. True believers seem willing to sacrifice the very environment they purport to protect if it might displace a little fossil fuel consumption. This way lies madness.

    Wind power is even more absurd than solar schemes. Over time most wind installations generate about 15% of rated output. Most of that generation occurs when it is least needed, during temperate spring and fall. Hot summers and frigid winter depths are often accompanied by high pressures and low winds. In Britain’s recent January cold spells their extensive wind networks generated less than 1% of capacity. In this blistering summer, Texas wind farms worked at 5% capacity and contributed to electricity shortages. If power grids are to function, they must have conventional (usually fossil fueled) backups equal to the unreliable renewable input. In consequence, even extensive and expensive renewable facilities actually save very little in CO2 emissions.

    Why waste so much money? As economic investment it is nonsense; as religious sacrifice, it is apparently a great consolation to true believers.

  11. It seems to me that someone is making a big thing out of a technology that is thirty years old. This plant looks and sounds very much like Solar One that was tested in California in the the 1980s. Even the molten salt method for storing heat energy for use during cloudy days or night was proven by these tests. To me this report is more a cheer-leading/PR action to make up for the loss of the Solar Panel maker in Fremont.

  12. Interesting article and it is true that molten salt technologies may be helpful for real-world storage. However, there are a couple of missing pieces to your article. First, CSP is not a new technology, it has been around since ancient Egypt. Solar PV is likely to be the solar technology of the future. Second, you are omitting the cost of reserve operations in your calculations. Finally, it’s intermittency over a fine time-scale that matters, not the ability to produce throughout the night. Having said that, I analyze the intermittency effect of solar PV in Arizona and find them to be about $11 / MWh, i.e. not insurmountable. Feel free to check out my paper (which will require some mathematical and economics training to fully understand) http://www.u.arizona.edu/~gowrisan/pdf_papers/renewable_intermittency.pdf .

  13. “Denmark is now close to 20% wind supply, and SFIK manages fine.”

    That’s not really an accurate summary. Denmark is connected to several other countries — notably Norway and Sweden, which have massive amounts of hydropower, which can be switched on and off fast enough to match the vagaries of Denmark’s wind production. So Denmark’s wind farms should be considered a small fraction of the total north European power supply. It can’t be considered in isolation any more than can Nolan County, TX — which has about as much wind capacity as Denmark.

Comments are closed.