We have 104 licensed commercial nuclear reactors—generating about 20% of our electricity and more than 70% of all carbon-free electricity. My company, the North Carolina-based Duke Energy, has seven reactors and we are planning three more. France operates 58 reactors and China has 11, but it is currently building 24 more.
These are numbers for the US. In France an even larger percentage of carbon free energy comes from nukes since 80% of France's electricity comes from nukes.
Out of the remaining 30% most of that is hydroelectric. Wind and solar are still pretty small sources of electric power.
China has become the big test bed for faster and therefore lower cost construction of nuclear power plants. If China succeeds (with help of Western companies who will go up the learning curve too) then that will change how nukes get built in Western countries.
For those who want to see a reduction in carbon dioxide emissions the role of nuclear energy as a low carbon power source should not be ignored. The Electric Power Research Institute (EPRI) says that an effort to cut CO2 emissions 41% by 2030 as compared to a 2005 baseline would require 45 new nuclear power plants along with other measures.
PALO ALTO, Calif. (August 3, 2009) – The Electric Power Research Institute (EPRI) today released updated “Prism and Merge” analyses that show a full portfolio of electricity sector technologies could simultaneously address the challenge of growing load demand while meeting carbon constraints and limiting increases in the cost of electricity.
The research shows that the sector could potentially reduce annual CO2 emissions in 2030 by 41 percent relative to 2005 emissions levels, but that it will require sustained research, development and demonstration and aggressive deployment of the full technology portfolio.
They include coal with carbon capture and storage (CCS). But every report I come across on comparative electric power costs shows coal with CCS as costing more than nuclear power. So why bother with the more expensive choice?
The full portfolio includes coal-fired generation with carbon capture and storage, renewable resources, and nuclear generation, as well as significant efficiency improvements throughout the electricity production and delivery system and reduced consumption through end-use efficiency.
The full portfolio requires deployment of advanced technologies by 2030 comparable to those assumed in the Prism analysis; an 8 percent reduction in electricity consumption through improved end-use efficiency; 45 new nuclear units; new renewables generation equivalent to four-fold increase in current wind and solar generation capacity; and, 100 million plug-in electric vehicles.
Here's an example of cost estimates for coal with CCS versus nuclear power. You can see this chart for a succinct summary comparison of expected electric power costs for a variety of sources in 2016. Coal with CCS doesn't compete. Of course, neither do wind or solar without subsidies. I'm more optimistic about solar in the afternoon due to its peak time falling closer to the time of peak demand. But nuclear power makes more sense for baseload.
A more electrified and cleaner economy is possible. Much more space heating could be done with air sink and ground sink heat pumps. Trains could be electrified. The hardest things to electrify? Airplanes. They need high energy density chemical fuel. In a world with high and rising oil prices airlines take the biggest hit.
France already produces low carbon electricity using nuclear reactors.
Each megawatt-hour of electricity in China requires 868 kilograms of carbon, compared with 611 kilograms in the United States and just 88 kilograms in France, which produces 80 percent of its electricity from clean nuclear energy.
The May 2009 update of MIT's cost study of nuclear power costs shows on page 7 that coal with a $25 tax per ton of carbon costs about the same as nuclear power (8.3 cents per kwh versus 8.4 cents per kwh for nuclear). Given that nuclear doesn't emit any CO2 why bother with coal? The main factor making nuclear power expensive? The cost of capital. If you equalize the interest cost on a nuclear plant as compared to a coal plant the nuclear plant's cost per kwh goes way below coal electric's cost.
For this reason, the 2003 report applied a higher weighted cost of capital to the construction of a new nuclear plant (10%) than to the construction of a new coal or new natural gas plant (7.8%). Lowering or eliminating this risk-premium makes a significant contribution to making nuclear competitive. With the risk premium and without a carbon emission charge, nuclear is more expensive than either coal (without sequestration) or natural gas (at 7$/MBTU). If this risk premium can be eliminated, nuclear life cycle cost decreases from 8.4˘ /kWe-h to 6.6 ˘/kWe-h and becomes competitive with coal and natural gas, even in the absence of carbon emission charge.
The MIT study also assumes a slow rate of increase in coal prices. If the pessimists on remaining coal reserves are correct then coal will become much more expensive.
If the efforts of Chinese and Western companies to build Westinghouse AP1000 reactors in 36 months succeed then I'm expecting a decline in the cost of nuclear reactors. Fast build times using module construction and passive designs ought to reduce the perceived risk and therefore lower interest rates on capital for nuclear plant construction.
By Randall Parker at 2009 August 08 10:51 PM Energy NuclearThe most likely cause for slow construction of US reactors will be red tape and obstructionism by enviornmentalists and NIMBY's. Over the years, all sorts of laws and regulations have been put in place that can be used to delay almost indefinitely the construction. Environmentalists showed how well this could be done in their obstruction of the Mt. Graham telescopes, and the barriers there were trivial compared to that of nuclear plants.
Right now, in California, a union is using threats of tying companies up in these red tape tangles to extort union wages from workers - on solar power plants!
Engineering is no longer a significant issue with politically incorrect power - it's the politics that controls.
John Moore,
While several states ban nuclear reactors outright the regulatory landscape for nukes has gotten much more favorable. Some federal legislation in 2005 and changes made by the NRC has reduced the number of regulatory approval steps and greatly reduced the regulatory risk.
In some parts of the US nuclear does not elicit the sort of opposition we see in California. The southeast is especially favorable - sensibly since wind power there is so weak.
Nuclear faces subsidized wind power and construction time risks. Areva's mistakes in Finland have hurt nuclear's prospects. But fast construction times in East Asia will help nuclear in the US.
I'm struck by several things in the EPRI study.
First, their low-cost scenario includes a lot of PHEVs - 40% market share by 2020.
2nd, they use as a baseline the highly unrealistic EIA base-case, which assumes that nothing changes from Bush-era policies.
3rd, they assume a very low level of renewables: 135GW by 2030, which is only about 5GW per year, when 8.5GW of wind alone was installed in 2008.
Re: France. They don't get 80% of KWHs from nuclear. If you divide nuclear production by French consumption you get about 80%, but that's dishonest: much of the nuclear production gets sold at night to neighbors (at very low prices).
Nick,
Was that 8.5 GW number for wind a nameplate number? If so, the real amount was less than 3 GW. For baseload it is much lower still.
France: You got any numbers behind this (admittedly plausible) assertion? I wonder what their demand curve looks like. Both Germany and Britain are winter peaking. My guess is that their day peaks over night aren't as steep as ours either.
Those US environmentalist red-tape NIMBY types are more powerful than anyone could have imagined.
They drove up the cost of Florida and then Turkish and Finnish reactors into the almost unaffordable range.
The fact is, nuclear plants require very high standards in construction, inspection and design.
American construction firms that once supported the industry have completely forgotten what they knew.
The familiar slow learning curve will drive nuclear construction prices up and delay plant openings. Yet again.
This isn't helped by the US approving several competing designs rather than insisting on one.
Was that 8.5 GW number for wind a nameplate number?
Yes.
If so, the real amount was less than 3 GW.
Sure. Keep in mind that natural gas generation commonly has capacity factors even lower than wind, and that the overall US capacity factor is only about 41%.
For baseload it is much lower still.
You're thinking of peak-load capacity credits. For baseload, which is at night, wind produces a bit more than average.
With the 100M PHEVs projected by EPRI, it will be much, much easier to deal with wind intermittency.
Well, I went back to the EPRI slideshow. As best I can tell, they're projecting about 90GW average output from wind in 2030 for their high case (Full Portfolio), so that's about 300 nameplate, or about 270 more than now, or about 13GW per year. That's not as unrealistic as it looked at first blush, but it's clearly a lot less than we can do, should we choose to do so.
Randell,
What have you hear about thorium cycle nuclear power?
Trent,
Thorium: The Indians are hard at work on it because they have a lot of thorium. There's also a company in Virginia or DC, Thorium Power, that is testing a thorium-uranium fuel package in Russian reactors that will let reactors go longer between refuelings (increase profits and output) and lower costs of fuel and waste. But so far thorium isn't ready for mass deployment. Looks promising though. Here's a post I did on thorium power.
My expectation is that thorium and ways to burn U-238 will both come along before U-235 supplies become an issue.
I'd especially like U-238 to come along faster because some of the proposed designs for burning it will hugely cut waste to dispose and run for a long time (decades and longer) between refuelings. Plus, the amount of U-238 we have is hugely greater than the amount of thorium or U-235.
Nick,
A power source that produces at night is not baseload. To be baseload you have to do 24x7.
Natural gas's capacity factors are due to its dispatchable nature. It can supply peak demand. Wind's capacity factors tell a far worse tale. You mention France selling nuclear at night at lower prices. Well, wind at night sells for lower prices too. At least nuclear also sells in afternoons when prices are much higher.
Randall,
Wind's night production is only a bit higher at night. It does produce 24x7, and in some places peaks in late afternoon.
No question that it has more variance, and is less predictable than nuclear. OTOH, nuclear has more variance than nuclear advocates like to admit. Most US plants have gotten very, very reliable, but that's after decades of service, and US nuclear production works so well in large part because the US has far more plants than anyone else, and a far larger grid. For instance, grid engineers in Ireland toy with the idea of nuclear, but they don't take it seriously, because nuclear plants can always trip, with seconds notice, and take days (and, occasionally, a year) to restart (1GW lost in a few seconds is a lot of variance) - the loss would be too large for Ireland's small grid. If wind had the equivalent size (325GW) it would be far more reliable.
Neither nuclear nor wind are dispatchable: their production will always be maximized, and other sources will be used mitigate unexpected variance.
The France electricity sold to Italy during the night time is used to pump water up in hydroelectric dams. Then, during daylight the dam let the water go down and produce electricity when it is more needed.
A few years ago, a tree broken a line bringing electricity in Italy, at night, and resulted in a nation wide black-out as the lines to/from France and Switzerland failed one after the other.
Nick,
I find "a bit higher" as pretty imprecise. Evidence please.
Nuclear has more variance than nuclear advocates like to admit? I think you are attacking a straw man here. I do not see the NEI exaggerating up-time. They point to up-time stats. But the stats look accurate.
Why have up-times improved? I suspect economic incentives and consolidation of nukes under a much smaller number of owners.
Ireland: Yes, and the relevance? China isn't Ireland. Continental Europe isn't China. The US and Canada on a big grid aren't Ireland. These are all big grids that can easily handle lots of nukes.
Nuclear and wind: But nuclear's output will be more consistent and most of its down time is scheduled.
I find "a bit higher" as pretty imprecise.
True. I've looked at a number of wind diurnal production curves: some were higher during the day, my impression is that the majority were slightly higher at night. I downloaded some data from Ontario, recently ( http://www.theimo.com/imoweb/marketData/marketData.asp ). I took a look: 9AM-9PM wind output was 1.8% lower than the daily average.
more later...
I do not see the NEI exaggerating up-time. They point to up-time stats. But the stats look accurate.
It's a question of the context, and the impressions and perceptions created in the minds of readers. I see a lot people attacking wind as being "unreliable", and it seems to me that people are getting the idea that there's a clear, black and white line between wind and nuclear. I think they don't realize that all power generators have an element of unreliability, that ISO's evaluate all sources that way, and that nuclear is not 100% reliable: as I noted, any nuclear unit can go out at a moment's notice, and stay out for days or years.
Why have up-times improved? I suspect economic incentives and consolidation of nukes under a much smaller number of owners.
Hmmm. What economic incentives are you thinking of? Early in the history of the US nuclear industry it wasn't uncommon for plants to have 70 or 80% uptime. Raising that number has a very direct payoff. It took a long time to do it. One of the problems of the industry is that plants are so large that there's too much time between plants. In effect, they're usually new designs, being handmade for the first time. Operating experience doesn't necessary apply fully to new plants. If we could design small, modular, manufactured reactors that might change, but for the moment....
Ireland: Yes, and the relevance?
The point is that nuclear has variance, and that it's variance is solved by scale. The same applies fairly strongly to wind.
Don't get me wrong. I think nuclear is ok for OECD countries, and that it will be a very useful part of the generation mix. But....it's too slow to do everything, and I'd prefer to see us maximize our investment in renewables, because...I'd be happy to see Iran (and many other similar countries) installing wind and solar, and nuclear - not so much.
edit:
"If we could design small, modular, manufactured reactors" should be
"If we could produce small, modular, manufactured reactors"
I'd especially like U-238 to come along faster because some of the proposed designs for burning it will hugely cut waste to dispose and run for a long time (decades and longer) between refuelings.U-238 is somewhat of a problem, because getting a breeding ratio > 1 requires a fast-neutron reactor. This requires funky coolants and a lot of other stuff that we don't have the experience to handle well.
One advantage of the Th-U fuel cycle is that it runs just fine with thermal neutrons; U-233 produces > 2 neutrons per fission even with a thermal spectrum. There is lots of experience with both water coolants and molten salts, and the inherent safety of MS (it load-follows naturally and the fuel drains with the coolant, leaving nothing to melt down) requires less fussiness overall.
Plus, the amount of U-238 we have is hugely greater than the amount of thorium or U-235.Actually, it's the opposite; the world is about 2.5 ppm U and about 10 ppm Th.
E-P,
The current designs that use enriched Uranium seem to me to be too dangerous, given that they can be exploited the way Iran has.
What combination of nuclear design elements would have the least propensity to support nuclear weapon proliferation? Would you feel comfortable with that design in the hands of a malignant government?
Based on the data and claims I've seen, the ones least liable to proliferation would appear to be solid-fuel thorium reactors (PWRs, ala Shippingport) and denatured (U-233 + U-238) molten salt reactors. Both of them would have piles of U-232 in everything they produced (nasty gamma-emitter Tl-208 in the decay chain, deadly to weapons electronics and easy to detect), and the DMSR requires isotope separation to get material that can go prompt-supercritical. The breeding ratio can be kept such that they can't be used to irradiate much in the way of other fertile material without running themselves down, and PWRs are inherently hard to tamper with without showing obvious signs.
hmm. So nothing in the fuel cycle lends itself to hiding a weapons-intent enrichment program?
Would you feel fully comfortable with either design in the hands of a malignant government?
The fuel cycle is another matter, because the initial fuel loads of thorium-based PWRs and the additional fuel required by DMSR's require enriched uranium. But troublesome nations don't have to have the fuel cycle, just the fuel at guaranteed prices. Foregoing any part of the fuel cycle can be a treaty obligation as a condition of getting the reactors.
the initial fuel loads of thorium-based PWRs...require enriched uranium.
hmmm. So once it's up and running, a nation using a thorium-based PWR would no longer depend on enriched uranium? Would that be a way to solve fears of dependency on other nations for supplies?
Foregoing any part of the fuel cycle can be a treaty obligation as a condition of getting the reactors.
That didn't work with Iran. It's not a part of international law at the moment. Building that in sounds very, very hard.
So, maybe a thorium-based PWR is the answer?
Yes and no. Thorium-based reactors can run a really long time on a fuel load (Shippingport ran 5 years on its Th-U load without refuelling and ended with more fissionables than when the run started), but I don't think that the proliferation-proof technologies are compatible with complete fuel cycles. I'd love someone to prove me wrong, though.
LeBlanc suggests denatured molten-salt reactors (first reference on page 15, video of presentation here) which would not achieve breakeven breeding ratios (they would be burners), but the amount of enriched uranium required (denatured with U-238) would be very small and the salt would require no removal of fission products for as much as 20 years. If a nation got the reactors as part of a treaty agreement (separate from the NPT) and all nations with enriched fuel for sale agreed to supply it only for direct installation in reactors under international supervision, that should deal with most of the dependency problems (many suppliers) and proliferation risks (minimal or no threat of diversion).
