Brian’s post has cast his financial eye over the WorleyParsons proposal to build solar thermal power stations across Australia. Now, some technological perspective…but with a financial sting in the tail!
Essentially, WorleyParsons is looking into a parabolic trough solar thermal design. The idea is straightforward; the radiant heat from the sun is focused by the curved mirror on to a tube containing a fluid, usually an oil. The heated fluid circulates into a heat exchanger, where the heat boils water to make steam. The steam is used to turn a conventional steam turbine, just as in a coal-fired power station. This is by no means new. Most of the solar thermal power stations built so far have used this approach, including the SEGS stations at Kremer Junction in California’s inland desert (which look rather pretty in street view).
What is new - at least in terms of industrial implementation rather than prototyping - about their proposed plant is the used of thermal energy storage. As I’ve tried to point out on many occasions, the problem with intermittent renewables isn’t simply their cost; it’s that their power production occurs only when the renewable resource is available, which means you need some kind of backup when it’s not. The proposed plant would reduce this problem quite a lot. Their solar collectors are actually considerably bigger than what would be required to run their turbine at full capacity. The surplus energy is used to heat a type of salt to the point that it melts and liquifies. The molten salt stores heat until sundown, where it can keep the plant running for several hours after dark. This is, quite conveniently, the peak period of demand for electricity. It also means that the plant doesn’t just immediately stop working if a cloud blows over the collectors (solar thermal plants, unlike solar PV, don’t usually work at all unless there’s direct sunlight).
You may remember Ausra, an American startup company with with Australian origins, who’ve gotten a good deal of publicity for their version of solar thermal technology. There’s a couple of major differences between this and what Ausra is proposing to do. As their fact sheet (PDF) shows, their collectors are flat glass, rather than the funky parabolic curves of the trough system. Furthermore, water is boiled directly in the steel pipes of the Ausra design, rather than using a separate hot fluid to absorb the heat from the sun, and then pumping that back to a boiler. These two features should make Ausra’s collectors cheaper to make. However - and this is something nobody except Ausra themselves know - the simpler design is probably less efficient, both in terms of the collector area required to collect a certain amount of energy, and the temperature the steam reaches (essentially, the hotter the steam, the the better) as it goes into the turbine. Until Ausra has actually built a plant, their projected cost-competitiveness with gas-fired power - which substantially undercuts the likely costs of power from the WorleyParsons proposal - is going to be taken with a substantial amount of salt by rather conservative companies like WP, I suspect.
The first thing to note about this study is that the technology is both fundamentally simple, and, with the exception of the energy storage system, industrially proven. There’s no “and then something amazing happens” relied upon to make this work. And in the sunnier parts of Australia, it’s going to provide energy on a far more reliable basis than wind, making it a lot easier to integrate into the broader grid. So I’ve no doubt that WorleyParsons could make this happen on a large scale, and that their goals of providing half of Australia’s renewable energy target - in other words, 10% of Australia’s electricity - by 2020 is doable. But the big question - and one that will be keeping WP and their partners busy crunching the numbers over the next few months - is whether it’s going to be financially viable.
Various media reports have described the proposed plant as a “1 billion dollar” plant. I checked with WP’s media people, and apparently that figure was based on their previous solar project experience, but was a very rough estimate provided in response to media queries - a more accurate costing is a major part of their study. But taking the $1 billion figure for the moment - a 50MW solar plant of similar design currently being built in Spain is projected to cost 310 million Euros - that equates to a construction cost of $4,000 per kilowatt of nameplate capacity. That’s eye-wateringly expensive. It’s roughly what the now-cancelled Kwinana clean coal project would have cost, and it would have produced power for a lot more hours in a day than a solar thermal plant does. While the minute-to-minute fluctuations of the power market, and the strong correlation between power production and power demand probably helps tilt the balance in favour of this technology, those costs are pretty scarily high.
With the mandatory renewable energy target, of course, this proposal isn’t competing against clean coal (or gas, or nuclear). It’s competing against wind, solar PV, geothermal, and other solar thermal projects. And, as far as I know, there hasn’t been a great deal of work on geothermal energy in Western Australia, so it’s possible that there it’s really only competing against wind, whose irregularity will start to become a major nuisance as the MRET approaches 20%. But unless they can get the costs down substantially - and given the simplicity of the technology, there doesn’t seem to be that much room for paradigm-busting breakthroughs - it’s going to be an awfully expensive way to get clean electricity.
NOTE: Story corrected.
Hi Robert,
Sorry if this is a stupid question (couldn’t find out) - what kind of salt are they using?
Thanks,
Steve H
They haven’t said.
By the way, I should make clear that one of the major goals of the WP study is to provide a more accurate costing.
You used to be able to buy off-peak electric house heating which stored heat in a big block of common salt (whatever happened to that anyway?). I’d imagine common salt would be fine for this too, as long as the troughs concentrated the heat enough to melt it.
DD: Natural gas and reverse-cycle air conditioning is what happened to it. Burning gas to produce heat, turning it into electricity, sending through the grid, only to turn it back into heat is a rather inefficient process.
Not strictly relevant, but did you know that heating salt with oil until it melts in the bottom of stainless steel cookware every so often makes a virtually non-stick pan?
Interesting post, Robert, particularly in the context of the clean coal guys’ latest plea for funding. I saw Peter Cook of the CO2CRC speak recently and he said they need $300m stat just to keep things going, and offered no guarantee of any commerical projects at any time soon.
Anyone who thinks there are cheap solutions are either kidding themselves are ignoring the externalities.
But, but, I thought decarbonising the energy sector was going to be cheap and easy!
So, if solar thermal is “eye-wateringly expensive”, and wind is hopelessly intermittent, where exactly is the miraculous technolgy that will deliver clean energy cheaply? I don’t believe its nukes, just the decommissioning costs are horrendous (ask the Brits!) not to mention the costs of waste disposal, planning and construction.
two points:
“cheap” is a relative term. Australia’s GDP is a trillion dollars. In that context, spending a couple of hundred billion dollars over the course of a decade is not a lot of money.
That said, I still wonder just how high a price Australians are prepared to pay to avoid dealing with their distaste for nuclear power. It’s expensive, but it ain’t as expensive as this proposal by a long shot.
Ummmm … where did my comment go?
Well, Robert Merkel does for one, and so does John Quiggin.
You see, the energy sector is a very small fraction of the world economy, therefore decarbonising the energy sector will be cheap. Make sense? No, it didn’t make sense to me either.
But today we’re told that solar thermal is “eye wateringly expensive”, wind is hopelessly intermittent, and nukes are expensive (but not as expensive as solar thermal).
Who here is buying the argument that decarbonising the energy sector will be “cheap”, in any sense of the word, relative or otherwise?
Oh come on Rob, putting aside the costs of nukes (which I think you underestimate considerably) the reality is nuclear power is politically impossible in Australia for the foreseeable future. Australians would much rather put aside their concerns about climate change than have nuclear power stations up and down the east coast.
Carbonsink: we’ve been over this a dozen times. Energy will get more expensive than it is now. But the fact is - and JQ has pointed this out many times to you - even if it doubles or triples in price, it will still be affordable, particularly if you use it more efficiently.
Your second claim - that Australians prefer global warming to nuclear waste - is true in the short term. We’ll see how long that attitude lasts.
Robert I presume the critical factor in a comparison of costs and output for such a facility in Oz as compared to Spain is this number “insolation : 2000kWh/m2y).” which is the figure given for the Spanish facility in your link?
What would a comparable figure be for possible sites in Oz and how would that impact on costs and output?
Or am I asking irrelevancies?
Incidentally I’ll be in the area in Spain in a coupla weeks time, I might check the site out.
I managed to find this:
http://en.wikipedia.org/wiki/Image:Solar_land_area.png
A map of world insolation rates.
It appears from this map that most of Oz has higher rates of insolation than the sunniest areas of Spain and northern Oz is even higher.
So, presumably, a facility in Oz similar to the Spanish site, would produce more ‘bang for the buck’?
That is indeed a key figure. here is an insolation map of Australia, which suggests much higher numbers. That would of course knock the capital cost down substantially. Obviously, only some of the cost relates to collector area; the steam turbine and the energy storage system would cost the same regardless.
It’s not that new. Rocketdyne had a demonstration plant using molten salt running in the 90s. That technology is now owned by United Technologies who have been “developing power systems and molten salt technology for 30 years in both terrestrial and space applications”. They’re commercializing the technology through a company called Solar Reserve (http://www.solar-reserve.com/technology.html with a little video) The US National Solar Test Facilty at Sandia Labs thinks the techology has legs, observing that it’s used on an industrial scale for heat transfer in the Chemical and Metal industries. They use a 60-40 mix of Sodium and Potassium Nitrate and think they can get 99% efficiency in the heat transfer in and out of the salt.
There’s also work being done in second tier heat storage using molten graphite, which works better at larger scales, but that’s pretty untried, I think.
d
Well, that Wikipedia page suggests “insolation : 2000kWh/m2y” for Spain - I assume that means 2000 kWh per square meter per year. That’s 5.5 kWh per square meter per day - so, based on that insolation map posted by Robert above, for most of Australia, the insolation is worse than or equal to the Spanish performance.
Even in that band of remote Australian outback where we’ve got 6 kWh per square meter per day, that’s only 2190 kWh per year - hardly more than the 2000 kWh in Spain.
To answer Steve’s question - the actual salt in molten salt energy storage is potassium nitrate, sodium nitrate, or a mixture of the two, usually 60% sodium nitrate and 40% potassium nitrate, which will melt at around 300 degrees C. Sodium chloride isn’t used, because its melting point is far higher.
Thanks for the correction on the insolation. However, I would be very surprised if Spain really got that much more sunlight than the Aussie outback. Sure, it’s sunny, but it’s also a fair bit further away from the equator.
http://en.wikipedia.org/wiki/Photovoltaics
Here, try this world map of insolation.
It shows the sunniest part of Spain to be less sunny than than most of Oz.
I’ve seen a hard copy global version of the map Robert posted. Of course I can’t speak for its accuracy, but I know at least one solar power company relies on it. There’s pretty much nowhere in the world higher than that strip running through WA and NT. A small patch of the US and quite a large area of the Sahara, and some tiny spots elsewhere are the only places with higher insolation if the map can be believed.
The Southern tip of Spain is similar to Perth from memory. Aside from Tassie and Victoria south of the Great Divide there is nowhere in Australia with lower insolation than southern Spain, and half the country is quite a bit higher, although of course the transport costs for getting the energy to the cities would be high.
It seems to me that being closer to the equator would not give you more hours of sunlight in a year per se, that would depend on cloud cover. But there would be less variability in terms of the hours of daily sunlight throughout the year and less variability also in heat intensity.
Robert, the AFR article cited the proposed plant as 250MW and that checks out with the literature from WP (see fact sheet) where they also say that 50-300MW gives the best efficiency. This would make the economics more reasonable. I’d find it hard to believe that a competent engineering firm would be chasing a project so far off the pace.
I stuffed up in my post by thinking that they were after 40% of the entire electricity market. It should be the renewable market with a 20% MRET, which leaves space for other players and other technologies. (In one place they say half, in another they say 40%.)
My memory is that Geodynamics (in which I have a few shares) saw themselves as being competitive with fossil fuel energy, given what they anticipated the cost of carbon would be in 2015.
I was wondering whether the WP design, which would have a closed circuit of oil passing through a heat exchanger, would be easier to manage in terms of a steady power output.
I agree that we’d expect higher irradiation in Australia than in Spain. Keep in mind that our only source for the Spain data is a brief snippet from a badly written Wikipedia page - a little dubious, to put it lightly.
BOM puts the average annual solar irradiation for most of Australia at 21 MJ per square meter per day, or 5.8 kWh, which is fairly consistent with the other map posted for Australia - so, I’d say the claimed figure for spain is dodgy.
http://www.bom.gov.au/cgi-bin/climate/cgi_bin_scripts/solar-radiation.cgi
Brian: Geodynamics will also be eligible for participation in MRET.
If they can pull it off in a reasonable timescale and something close to their cost estimates, they’re sitting on a goldmine.
A few points here:
- the solar insolation figures in Australia are much above Spain. Solar insolation can be quoted in either global or direct terms, and also relates to the angle of incidence ie a fixed horizonal plane < fixed plane angled at the location’s latitude < tracking ie always normal to the sun. From memory Direct Solar Radiation or DSR (the last one mentioned) in Alice Springs is 7.7 kWh/a.sqm and tops out at nearly 8 at Pt Hedland. Mildura is about 5.5 and it drops to 4 in Melbourne.
- the intermittency factor is bit of a furphy, as long as there are sufficient plants across a wide enough area. Even with wind (and a well-connected grid) it’s blowing somewhere. The South Austalian electricity market has been studied by NEMMCO and is demonstrably cheaper with wind in the mix. If you divide the total annual electricity output of Australia by the total installed capacity the average is just over 50% ie that’s the load factor of the fossil fuel installed base. Load growth is not so much baseload (well catered for by coal) but in the intermediate to peak loads (ie hot sunny days where people run their airconditioners). In this scenario, solar makes sense - see the work done by Muriel Watt at UNSW.
- one very big problem for renewables is that the energy source is usually not close to the population centres. It costs nearly as much to connect to the grid as it does to build the wind/solar/hydro plant. Coal generators have 60 years of infrastructure set up to deliver electricity from LaTrobe Valley etc to the far reaches of the land, a benefit not shared by wind or geothermal. Solar is far more distributed - a big solar plant could be placed almost anywhere in an arc from Mildura up to Mt Isa and have a very well defined resource base. For fossil fuel to deliver electricity to Broken Hill, for example, the transmission costs are more than the generating cost, and a renewable generator needs only to be less than the total delivered cost.
Wind’s intermittancy is OK if it makes up bugger-all of your grid and/or you’ve got lots of hydro in the mix.
Brian - You mentioned some concerns with Geodynamics’ share price in the other story. I’ve got some shares in Petratherm, and their share price has dropped about 25% since I bought them a few months ago. I reckon someone’s playing funny buggers with the market.
So it’s expensive to set up initially; but what about ongoing costs? Does it become more competitive in a longer time frame as there are no ongoing costs (for coal and gas) to fire the turbines day after day?
rf: the difference in construction costs is such that it’s going to be hard to compete with coal and gas on a pure cost basis.
The price of gas over the longer term may well be heading up, making such calculations look more favourable to the solar option.
David, the market as a whole is crap anyway at present and these small single product startups can be volatile at the best of times.
There are laws against insider information, but there will always be people who know more than I do about what’s going on. Also brokers don’t follow the tiddlers, so there’s usually no advice there except perhaps from a firm involved in the float, and that’s inherently unreliable.
It makes it hard for ordinary people.
“There’s also work being done in second tier heat storage using molten graphite …”
“Sodium chloride isn’t used, because its melting point is far higher [than nitrate salts].”
I’m sure you’re both right, but there’s gotta be more to it than that. Graphite melts at 3700 degrees while sodium chloride melts at 800 degrees.
Thanks everyone for the salt details - basically the biggest challenge in such a closed-loop system would probably be the startup/shutdown process. That’s where you’ll get some major efficiency losses.
Derrida, it’s not just a simple matter of temperature - there are a whole host of thermal engineering issues not to mention purity and storage requirements. My knowledge of this stuff is a bit sketchy but the Sodium/Potassium mix is a hell of a lot easier to handle which would also be a factor.
I’d say they’ll have to stick with single-medium storage solutions (pun intended!) for the moment - if nothing else to keep costs under control.
Brian, I invested in Petratherm from belief in what they’re doing more than an expectation of profit. I just thought it was a better place for my money than the Telstra shares I inherited from my late mother.
The thermal storage in oil based systems is usually a Eutectic salt mixture, the problem with these systems so far is that the mixture tends to stratify with thermal cycling and goes off eutectic. If the eutectic point goes too high the salt will not melt or solidify at the right temperature. Lot’s of people are working on this, have been for 20+ years and have yet to really solve it.
The major point about solar energy in OZ is that the majority of the population lives on the East coast in cities strung out on a North South axis -so we tend to track each others power usage. (We are talking solar time here - eg it gets dark along the east coast at the same instant- more or less-for every-body) If we all lived along the East West axis we would have two hours of “spread” of solar energy.
Parabolic trough systems that use oil as the transfer mechanism have actually caught fire in the US;a huge fire, so care will be required. I think that is why others have elected to use water as the heat transfer medium.
The upside of solar thermal is that its conversion efficiency is about 20-35% Photo-Voltaic is sort of stuck in the 10-20% well.
One way to solve the intermittency problem is to add in a gas fired front cycle using a gas turbine and to direct the turbine exhaust to steam raising;
I like the systems where you build a boiler/collector on a tower and bombard it with reflections from a whole bunch of heliostats. The advantage of this is that you can get the temperature high enough to use liquid sodium as a thermal storage medium. Liquid Sodium has a high specific heat, high melting point and is used in nukes as a thermal transfer medium - so it is readily available and well understood.
Huggy
Robert Merkel @ 10:
A doubling or tripling of energy prices is a near certainty. I mean, oil doubled in less than 12 months, and coal prices tripled overnight recently. Far more likely we’ll see price rises of an order of magnitude or more over the next 5-10 years. You can’t conserve your way out of that!
The price of oil rose 15-fold in nominial terms between 1998-2008. Who’s to say that won’t happen again over the next 5 years, especially if we do hit the “supply crunch” in 2010 that the IEA is forecasting? If that happens it will put enormous pressure on energy prices across-the-board, both fossil and renewable. How much will it cost to rebuild our entire energy infrastructure with oil at $500/barrel? Sounds absurd now, but $120/barrel would have seemed impossible in the late 90s.
Seems to me both you and JQ assume we will be rebuilding our energy infrastructure in a benign economic environment, and in a world without energy constraints. It seems unlikely this will be the case.
Rusty @ 22:
Important point. Chapter 17 of the Garnaut Review Draft Report (network infrastructure market failures) discusses this issue. It is also worth checking out the “TREC” proposal to use high voltage DC transmission lines to supply power for Europe from renewable infrastructure.
On the subject of
crowdsmarkets, as people may have noticed, there has been quite a bit of bubbles, busts and noise lately. I bought a small parcel of shares in Patratherm just before Garrett announced $50 million for geothermal development. This sparked a bubble in the Geothermal industry for the second quarter of 2007, which saw my Petratherm shares go up by something like 110%, and they are now down by about 10%. I’m not too concerned about the short term movements - If I had cash to spare I would be tempted to buy more of some of these firms at the current bargain prices. I’m also not too concerned about shorters, they will have to cover their positions sooner or later. I once heard that Warren Buffet only buys shares in a company when they are trading at at least 50% less than he thinks that they are worth.One thing that is a bit odd is that Geodynamics 2009 options (GDYOA) are trading at 20c, which would suggest that the market thinks that there is a good chance that Geodynamics will trade at $2.20 next year, a bit higher than their current price of $1.31.
Gosh, Robert to write this…
“What is new - at least in terms of industrial implementation rather than prototyping - about their proposed plant is the used of thermal energy storage. As I’ve tried to point out on many occasions, the problem with intermittent renewables isn’t simply their cost; it’s that their power production occurs only when the renewable resource is available, which means you need some kind of backup when it’s not.”
…..now, says that you really haven’t been following the information on CSP at all well. If you care to read this…..
http://www.solar-thermie.org/hintergruende/documents/cspnow.pdf
…..,information that has been around for a few years now, you will see that the storage technology has been installed an running for years and is giving up to 8 hours storage. Further if you care to follow the implementation steps (in the publication) you will see that at step 4 there is a reasonably clear costing formula.
The Spanish installation is expensive because it is just 1 fifth of the economic optimal 250 megawatt size. The optimum size is determined by the heat medium pumping distance to the turbine house. And it is the turbine house which is the most expensive cost item in the system, an item that must be present regardless of the size of the installation. The other limiting factor is the mirror production environment. For a small plant these mirrors must be produced elsewhere and transported, whereas for 250 megawatt and upwards the mirror plant will be built on site with all of the cost efficiencies that that brings. And on top of all of that is the cost of 1 square kilometre or European landscape.
No level playing field there.
If you would prefer to read all of that from the glossey brochure, I have a few of them here, and would be happy to send you one.
And the difference between the Ausra and SEGS systems in efficiency is also explained in the PDF file. Simply put the Ausra (fresnel reflector system) allows for over 400,000 square metres of reflector area per square kilometre but achieves a lower temperature against the SEGS 350,000 square metres per square kilometre at 400deg C. The area differences are in the casting of shadows and mirror efficiencies.
All up the Worley Parsons project proposal is double what it should be 2 billion dollars per gigawatt is a fair thumb guide for multimegawatt installations), but then as a first plant with the extra costs of transmission cables, setting up the process in a new country, building a mirror plant from scratch, building a whole new town for the staff, maybe it is reasonable.
Or maybe it is a political scare off proposal!
David, I had a bit of a look at Petratherm. It’s a small company with a market cap of about $35m. Looking at the Aspect-Huntley consensus forecasts it appears that none of the main brokers follow it.
Many traders don’t like buying stocks unless they have significant volumes of trades. I suspect Petratherm is a bit under the radar and would need some exciting announcement to stir up some activity. The whole sector of geothermal energy is a bit under the radar. Petratherm could nevertheless be long term investment. I’d say it depends mainly on the quality of the management, which is the hardest thing to know anything about. So in short I wouldn’t presume to advise.
I checked out Geodynamics and they are only followed by ABN AMRO who have a “strong buy” on it. I have access to their research and it seems that they concentrate on companies with a market cap of $300m-$1b in the small cap sector.
Hopefully big things from little things grow!
Bill: I know there’s been a pile of research into TES. However, it’s not been deployed on an industrial scale yet.
Re 26 (Robert) rf: the difference in construction costs is such that it’s going to be hard to compete with coal and gas on a pure cost basis. The price of gas over the longer term may well be heading up, making such calculations look more favourable to the solar option.
Solar can’t compete with coal (yet) without carbon pricing, but when considering WA power prices you need to consider 2 things:
1. WA gas prices are rapidly heading towards parity with international LNG prices (which is what the gas companies want) and they have a shortage of gas for the domestic market.
2. International LNG prices are only heading one way.
3. 2 major gas supply disruptions to WA over the past 12 months have cost local industry billions of dollars in lost income - diversifying away from gas has a lot of other benefits besides just cost comparisons.
4. Solar thermal prices will drop as volumes expand and the components start to achieve economies of scale - comparisons to bespoke plants built in Spain that led this new wave of construction aren’t really fair.
North west WA is probably the best place on the planet for building large scale solar thermal, with the possible exception of some parts of the Mojave near LA.
I think pretty much every stakeholder in WA has something to gain from this development, so it may well be reality before too long.
Distance is generally thought not to be a problem, as the extra kwh of energy generated more than makes up for the loss of power in transmission. Thus they could transmit solar energy over DC lines several thousand kilometres and it would still be cheaper than building the solar power stations close to cities near costal areas, that get less solar insolation.
Cost is a factor, but with carbon cap and trade, existing fossil fuel based electricity should increase in cost to the consumer. Clean coal technology will add further to that cost. Also the 2020 / 20% renewable mandate will play into the hands of renewables like solar and wind. Also we have to reduce greenhouse gasses anyway and as was said, nuclear seems unlikely to be accepted in Australia in the short term. Thus that leaves mainly solar and wind to do the job. I suspect in this context, cost will take a bit of a back seat. The job has to be done. It was never going to be cheap in the short term, but the nor were the consequences of inaction in the longer term. Solar thermal with thermal storage and large scale wind are best placed for utility scale power generation. Solar PV and solar hot water best suited for domestic use.
There was an interesting segment about solar on Saturday Extra on the weekend. Peter Meurs of WorleyParsons made it clear that they were playing a niche strategy to provide peak power that sits on top of base-load provision.
So solar would be available during the day, and by storing some, in the early evening and the following morning for heating, getting breakfast etc. So the strategy was to provide 1.35 times the peak power needed during the day and store the 0.35 for use in the evening and early morning. Storing more than that would blow out the costs too much.
So it is not intended to be a major source of power, but fit into a smorgasbord of offerings.
Another interesting aspect is that because of their global reach WorleyParsons is already involved in a range of solar projects around the world, including California. They did see a role for manufacturing some components locally in the Oz program.