The big news in energy this week is large solar. It’s exciting to see such developments, but just what does this mean for our energy challenge in the 21st century? Large solar provides appreciable quantities of energy, hamstrung by fundamental limitations in scalability and the potential for very real environmental impacts. A future with more solar is assured. But using this to argue an environmental case for a future without more nuclear too is simply deluded.

With the opening of the world’s largest solar power station I have been struck by the convergence in generating size between large solar and small nuclear. Solar is trying to get bigger. Nuclear is trying to get smaller. Both are now converging on around 300 MW- 400 MW in size.

There, the similarities end. Those who are familiar with my Zero Carbon Options project will know that I favour comparisons as a useful way to shed light on our collective energy choices, so that we might truncate the ideological hand-waving that accompanies energy discussions at large. Comparison of large solar and small nuclear holds some very important lessons for constructing a future that is both energy-rich and decarbonised for around 10 billion people.

The 392 MW, $2.2 bn Ivanpah Solar Electric Generating System (SEGS) has just started delivering electricity. For the record, SEGS was developed with a $1.3 billion loan guarantee from the US Government. Contrary to popular anti-nuclear belief, loan guarantees are not a subsidy concocted by and for the nuclear industry. SEGS also does not include any component of energy storage, however it delivers a high (for solar anyway) capacity factor of 31.4%. That’s thanks to a “remarkably intense solar resource of 2,717 kWh/m2/yr” combined with the dual-axis tracking.

So from the outset, it’s clear that the performance of a solar power station has a lot to do with the location, a critical constraint. Move away from areas of “remarkably intense” solar resource and the performance will dwindle in kind. So as far as scaling up this solution goes, how are we placed for raw resource? Figure 1 gives us the answer.

Figure 1: Map of global irradiance, units of kWh/m2

Pretty well, as it turns out. The areas of deepest red shading are blessed with this remarkably intense level of solar radiation. That’s a huge area of land, and the available energy must be mind-bendingly large. So, one of the next questions for intelligent siting of large electricity generation is proximity to large loads. Unless we can use it close to where we generate it, we are stuck with big losses through the transmission lines. This is a traditional criticism of the model of large centralised electricity generation dominated by fossil fuels and nuclear. The impact can perhaps be mitigated through the construction of new, very large direct current transmission lines. These are typically neither cheap, nor popular. Here, the news for solar is less rosy. Compare the above figure to a map of global population density in Figure 2.

With the partial exception of the Indian sub-continent, the areas with this “remarkable resource” are nigh-on uninhabited, for reasons that should be obvious. So for this to scale as an energy solution, we would have no choice but to embrace a future of massive transmission lines connecting the resource to the people.

So, it turns out Ivanpah was a canny location for SEGS indeed. Connection did not require a brand new line, just the much (politically, socially and environmentally) simpler upgrade of a 115 kilovolt line to 220 kilovolts, substation upgrades and construction of one new substation, at a cost of $446 million to $484 million. The cost was, oddly, somewhat better hidden than the bragging pamphlet… Let’s not be hasty and slap all that cost on this project. This line is capable of bringing another 1,000 MW of generation in addition to SEGS online. Sensible network planning requires investments of this kind to avoid punishing first movers (assuming, of course, we really need these remote connections in the first place).  Ivanpah also provided access to natural gas connection for the gas back up which will help it warm up in the morning and maintain operations during transient cloudy periods. Connection was just 0.5 miles away.  So that’s three important constraints Ivanpah overcame: a convergence of excellent resource, electricity network connection and gas connection. Add to this the proximity to the large loads of California and Ivanpah is a dead-set winner for location of large solar.

Then it’s worth looking at the sheer scale of this thing. At 5 square miles, around 3,500 acres, it’s just absolutely enormous. Check it out.

Figure 2: Ivanpah solar from the air

Figure 3: Ivanpah solar from the air

I’m sure most will agree, this is a staggering area. I think it’s safe to say environmentalism has comprehensively rejected “small is beautiful”.

What those less acquainted with energy may not realise is that 392 MW at 31.4% capacity factor, while hardly small, is frankly modest in terms of what we actually consume. My small state of South Australia alone, population 1.6 million, has about 4,000 MW of installed capacity. In the context of California? Let’s just say we better keep things in perspective.

Which got me thinking: what would the land requirements be for the equivalent level of electricity generation for small modular reactors (SMR)? To test this, I considered a “twin pack” of Babcock and Wilcox Generation mPower 180 MWe SMRs. This front-running SMR design is currently working through the licensing process in the US. An example design is shown below and discussed in more detail here.

The site footprint for this “Twin Pack” is just 38 acres. That, combined with the far higher capacity factor, leads to a remarkable cascade of numbers, shown in the table below.

SEGS Generation mPower Twin Pack
MW Capacity 392 360
Capacity Factor (%) 31.4 90
GWh/yr 1,078 2,838
Acres 3,500 38
GWh/acre 0.31 75
Land productivity rating (SEGS as base unit) 1 242

The small nuclear is capable of producing 242 times the electricity per unit land. Not 242% mind you. Two hundred and forty two times more. Small nuclear is as staggering in scale as large solar, just for the opposite reason.

It all comes down to energy density. With large solar we are trying to scoop up and make use of very dilute energy. For SEGS, we know this to have a value of 2,717 kWh/m2/yr. That’s a year’s worth of sunshine on a square metre of turf in the Mojave Desert. To produce 2,717 kWh from uranium dioxide fuel in the reactors shown above would require a mere 7-8 g of UO2. With a density over ten times greater than water, we are talking about a mere 0.7 cm3… barely a pebble on our square meter.

Site insolation (kWh/m2/yr) 2,717
UO2 reactor fuel kWh/kg 360,000
UO2 kg/kWh 2.8*10-6
UO2 for m2 insolation (g) 7.6
UO2 density (g/cm3) 11
UO2 volume for m2 insolation (cm3) 0.69

Of course, SEGS can’t convert all of this insolation. At a forecast output of 1,079 GWh per year, SEGS gives us around 76 kWh/m2/yr. The UO2 required to deliver this much electricity is just 0.2 g, or a volume of 0.02 cm3 …just a flint from our pebble on our square meter.

My head is already hurting with the tiny-ness of it all and yet it doesn’t even stop there. If our SMR is a GE-Hitachi PRISM (311 MWe) with Advanced Recycling Centre (put together, this is the commercial incarnation of the Integral Fast Reactor), then we have to divide our flint from our pebble from our square meter by about 100 times again.

This just goes to show that when we further developments in both the most dense and most diffuse sources of energy available to us, extraordinary comparisons will arise. Two hundred and forty two times the land consumption. Specks of mineral as a trade-off for square metres of wild land. These are profoundly different directions for humanity in the 21st century.

Some people might say, well, who cares? We have the land, let’s use it.

You might not think that if you were a desert tortoise. There is another key criteria (along with raw resource, proximity to load and ease of connection) that comes into play when we select sites for large solar: alternative uses of the land, including biodiversity.

In developing SEGS, the project proponent discovered the place was, literally, crawling with these iconic, endangered reptiles, the desert tortoise. There were up to 160 adults living on site, with likely over 600 juveniles. The project proponent has spent more than $56 million relocating the reptiles for their protection with mixed results. In a weak concession, the Sierra Club gave Ivanpah the nod despite this indisputable impact. The company version of proceedings reads like exactly what I know it to be: genuinely good people working for a very large corporation that can only ever put on a positive spin when everyone knows the simple truth: the fauna would have been best served without the disturbance. In this, the line between the “goodies” (companies developing renewables) and the “baddies” (companies developing fossil fuel and nuclear resources) is not just blurry, it has completely disappeared.

Figure 4: Sadly, the tortoises did not actually squash the heliostats. It was the other way around.

As a sustainability professional, this is starting to feel like falling through the looking glass. An iconic green energy project is disturbing a maximal area of land to give us what we need, in an environmentally sensitive area that is home to iconic fauna, which are now the unwilling recipients of tens of millions of dollars from a large corporation in an attempt to make good the unwelcome intrusion. This, surely, cannot be the direction for sustainable energy this century.

Asking whether we want large solar is an entirely moot question. We have it, and we are going to get more of it. I have seen some exquisitely degraded Australian land that would do very nicely. However, the popular supposition that solar developments rule out the need for nuclear power in a rapidly developing world is insupportable. It is every bit the “hallucinatory delusion” that Shellenberger says it is in Robert Stone’s Pandora’s Promise.

As I recently established, even if we rich folk, by some miracle, were to meet in the middle on our electricity consumption with the rest of the world in the year 2050, that world would require twice the electricity that it consumes now. We need to decarbonise our current supply, then double it with new clean supply.

If we environmentalists presume to advocate treading lightly on our earth, then it is dense energy, not dilute energy, that really shows us a way this might happen.  Maybe small is beautiful after all.

Like what you see here? Please subscribe to the blog, Like Decarbonise SA on Facebook and follow @BenThinkClimate on Twitter. Read more about the potential for nuclear power in Australia at Zero Carbon Options


  1. Hi Ben. Good article, but a minor correction. Ivanpah SEGS is a solar thermal generating station, storing energy as heat from 4 to 7 hours depending on load. Storage is how they boost solar capacity factor from nominal 22.5% (based on average insolation in that area, about 225 W/sqm divided by 1kW noon time peak ) to 32%: the steam turbines can never convert the full collected noontime solar power, and don’t need to. Peak output power is reduced, but useable energy is increased so CF = average_output / peak_output goes up. (Of course thermal storage losses also decrease average_output, but that’s engineering.)

    This amount of storage may seem trivial, but it can shift the load delivery from noonish to late afternoon-evening, when demand peaks. Particularly on hot summer days.

    So in that sense its a step in the right direction. And far be it for me to suggest the same thermal storage could be used to balance load for SMRs and increase their capacity factors..

    That shouldn’t be a problem for a few decades yet. As long as there is baseload to supply, nuclear is certainly the least-cost least-emissions way to supply it. Its when we start asking nuclear to supply variable load that we begin to run into (economic) problems. I’m thinking a combination of thermal storage and chemical processes that can economically live with variable process heat might help, such hydrogen co-generation via thermal pyrolysis. Hydro and pumped hydro were you can get it.There will be trade-offs.

    At this point it looks as if the climate and environment are still the easiest things to trade off.

  2. Not sure why people think nukes can’t be designed for load following. Think of the power plants on nuclear subs or aircraft carriers. They would be in a bit of a pickle if they took an hour to throttle up or down!
    Molten salt reactors like David LeBlanc’s DMSR are excellent load followers. Most of constraints are in the design of turbo electric equipment not the reactor. Even our venerable old Candu rectors at Bruce can load follow at reasonable rates. There is no need to store significant amounts of energy in a nuke designed for load following.

    1. You are quite right, and all new Gen III and Gen IV nuclear follow load at least as well as fossil plants, as do many “older” Gen II. As I tried to allude, the problem is economic, not technical. Nuclear plants are very capital intensive, to pay them off they have to generate power. As long as fossil plants are still providing baseload, new nuclear plants will be used to replace them in that capacity. For emissions reasons nuclear must eventually be used to follow load, but we (US) are sooooo short on nuclear right now it makes no ecological or economic sense to use them that way.

      There’s load following, and then there’s load following. We might not be thinking of the same degree. I don’t know at what capacity you run your Canadian plants. France runs theirs at about 77%. Here in U.S. its ~93% — all baseload. I once did a quick estimation on what it would take to replace coal and gas generation with nuclear, where the coal was running near baseload at 84% and NG was load-following at a capacity factor of 20%. At 84% CF nuclear economics looked quite good. At 20%, not so much.

      Marine propulsion is a different story. As is everything else if we can just get the cost down and production up.

  3. What troubles me about Ivanpah apart from cost and dust storms is the dependence on natural gas. That is apparently for both cold starting and output levelling in cloudy weather. The towers are the show ponies in the front yard and gas is the old draft horse toiling away in the back paddock. Since some are saying eastern Australia only has 22 years of proved and probable gas let’s come up with something that doesn’t need gas.

    Without doing detailed calculations it looks as if a moderate improvement in battery costs would see PV + batteries beat solar thermal every time. Indicative figures could be 15c generation + 5c retrieval costs to get 20c per kwh. The AETA figures had solar thermal with molten salt going as high as $400 per Mwh or 40c per kwh. Solar thermal is the Betamax videotape of generation technologies.

    1. This one includes “a partial-load natural gas-fired steam boiler which will be used for thermal input to the turbine during the morning start-up cycle and will also be operated during transient cloudy conditions in order to maintain turbine on-line status”. Absent data on the consumption I would give the benefit of the doubt that it is likely to be pretty minor. However, it’s yet another siting constraint.

  4. Desalination and long distance water pumping could be intermittent forms of load shifting to keep electrical output within a preferred band. From the Wiki article on desal a mid range energy requirement for reverse osmosis is 4 kwh per m3 or kL. The mothballed Whyalla desal for Olympic Dam was to produce a bit over 250 ML/d so I make the average power requirement about 42 MWe not including long distance pumping. The link says 35 MW for 200 ML/d but the water requirement was later revised. The power was to come via the NEM which is basically coal and gas.

    Large water tanks en route or at the customer end would overcome the irregular flow problem. Thus a 360 MWe SMR pair both working at full capacity could assign 0-40 MW to RO desal and 320-360 MW to other loads.

  5. Ben

    Is it not naive to say that nuclear is converging on 300-400 MW? If I look around the world I see a rather large number of old style 1 GW type plants being built, but SMRs only existing on paper. It’s jumping the gun more than a little to simply assume that SMRs will even be economical, let alone take over from traditional 1 GW plants.

    1. Well, it may be jumping the gun, but on paper SMRs have several advantages. Lower power output allows — on paper — some of them complete passive cooling, with natural air convective cooling after shutdown. The big LWRs still require water. (I know, I know). Second, at some point nuclear will be used for process heat as well as electricity, SMRs might be easier to site.

      In addition, Gen IV nukes run at near ambient pressure so their containment costs are lower. That said, the only Gen IV for commercial power generation is Russia’s single BN-600, (600 MWe) with their first BN-800 currently being fueled and scheduled to go online at end of the year. Two more BN-800’s to soon start construction in China, who is working with Russia on the BN-1200 follow-on. China is also working with Westinghouse to scale AP1000 (1120 MWe) up to CAP1400 and CAP1700. EPR is a bit over 1600, ESBWR a bit less. So yeah, for large-scale commercial power generation there appears to be a trend.

      But even in Australia some are seriously concerned with the climate catastrophe and that their country do its part in its avertment. Australia currently has some antipathy toward (and legal restrictions upon) nuclear power, and little operational or regulatory experience. Perhaps Ben thinks a well-presented walk-away safe SMR project might gain public acceptance where larger units would not. Have to ask him sometime… 🙂

    2. To a point Robert, yes it is, and it was something of a rhetorical device to frame the analysis. It was not intended to speak for all nuclear, more so the area of most development. I was struck by the curious reality that the size ratings are about the same.

  6. DSA must be making headway with 58% of respondents to an Advertiser survey wanting nuclear power for SA
    However that seems to be with the proviso it must be low cost. In my experience 100% of people directly or indirectly associated with SA defence contracting support nuclear power.

    The article outs premier Jay Weatherill as a nuclear opponent, something long suspected. There’s been talk of extending gas pipes to SA from Alice Springs with the gas new source in the Timor Sea. We’ve not only stiffed the Timorese right next to the gas field but the idea seems to be if uranium has to be mined it must be done with fossil fuel energy. Weatherill’s view is apparently that foreign people can use uranium if they must but SA will use everything including coal and remote gas in preference. SA then becomes Germany-lite. I suspect Weatherill will hang on in the March 15 election so this is what we’re up against.

  7. Ben, This was written a little while ago, but you refer the the B&W mPower as the front runner SMR. B&W have now stopped development of mPower, citing the cost of regulation and the money already spent commensurate with the lack of progress toward regulation. The DOE seems to be supporting NuScale via some research funds which has also caused Westinghouse to announce a “go slow” on their SMR development.

    The cost and bureaucracy resulting in very slow progress through the system seems to be a common story from users of the NRC, despite the view externally that it is the “gold standard” in terms of regulation.

    Perhaps something that we should not be trying to emulate here.

    Despite the development (or lack of) for mPower, your comparison of power density remains valid.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: