Updated 12/15/2010

Nuclear Fission

Nuclear fission has been viewed by many to be the Holy Grail of power production: it’s extremely low carbon long term electricity production, and theoretically could be cheap. Moreover, its safety record is admirable.

For reasons that we explain in more detail on our “Stabilizing the Renewable Grid” page, “Wind”Fuels could benefit from a nuclear renaissance – in that the nuclear designs that are currently under consideration are baseload “always on” designs. A large-scale increase in power with poor load-following ability needs an energy storage solution

However, nothing in life is simple. In America (as well as in most other advanced democratic nations), there is little chance that new nuclear power could be competitive over the next two decades. The basic competitive problem comes down to the constitution and the political system, as we’ll explain.

Capital Costs.
The capital cost of nuclear energy is undeniably the most important nuclear issue for the next decade, so it deserves some perspective and analysis. Most advocacy groups continue to cite price data that are 5-10 years old and totally useless.

The prices being quoted for nuclear power plants in the US have doubled in the past three years, after having doubled over the preceding three years. This is due in part to the rapidly rising cost of needed materials, but more so this is due to an increased appreciation for the challenges of restarting the nuclear plant construction industry.

The DOE has recently offered an $8.3B loan guarantee for two new 1.1 GW nuclear generators in Burke Georgia. Due to the loan guarantees from the government, which significantly reduce investor risk, the $14B project may result in the first new nuclear power plants built in the U.S. in over 30 years. This would work out to $6360/kW, which is well over twice what has generally been claimed by enthusiasts. The fact that this adds two new units to an already existing campus eliminates the land costs and the cost of high capacity transmission to a remote location. It also helps bypass some of the local objections, as the citizens already live in proximity to much older and less safe nuclear power plants. (Note the older plants are still sufficiently safe, but new safety standards are far more stringent than those of 40 years ago). For comparison, a similar plant – 2 reactors, each 1.1 GW – in Levy county FL has been projected at $14B. plus $3B in transmission costs, and it has already been delayed 20 months before ground has been broken.

Due to the rapid rate of increase in the projected cost of new nuclear power plants, this is by far the most relevant price point. But this is still only a projection. Assuming this projected price is valid, assuming an 85% capacity factor, and a 7.5% discount rate; the amortized capital cost of energy for a 40 year power plant would be ~$67/MWh. This is an over-simplification, as it assumes no additional financing costs during the 6-8 year build phase, it assumes no delays or cost overruns, it assumes a lifetime capacity factor of 85% with no shut-downs or renovations, and it assumes negligible fuel, waste-storage, and decommissioning costs.

Nuclear power plants are built one at a time and require about 20,000,000 hours (10,000 person-years) of highly skilled on-site labor during construction. The reactors must be situated far from a population center, and this necessitates enormous re-location costs for thousands of construction workers, their families, and the support community. The fact that very few of these workers will have ever worked on a nuclear power plant before causes an extreme risk in delay and cost overrun.

Delays and Cost Overruns.
In 2009, MIT’s updated study showed a 15%/year increase from 2003 to 2009 in the construction cost for large-scale engineered projects. This is due in part to inflation in materials cost. Other contribution factors have been increases in the cost of labor, liabilities, licenses, insurance, and land.

Nuclear power plants are more vulnerable to this inflation than other large-scale engineering challenges due to the near certainty of delays. Local authorities and NIMBYism serve to supply ample red tape and bureaucratic obstacles that have led to a universal state of long-term delay within the industry. Finland’s Olkiluoto site was scheduled to be completed in the first quarter of 2009, but it will not become operational before 2012. As of mid-2010, the costs for that plant had already increased by 77%, and the contractor – Areva – expected the costs could rise further, depending upon the new timeline that was negotiated.

Areva may have learned its lesson from the Finnish build, and bid $7,375/kW for a new plant in Ontario. However, the Ontario government rejected that bid because Avera refused to take a sufficient share of the risk – demanding that cost overruns be paid by the Ontario Power Authority. The lowest bid that complied with Ontario’s risk sharing requirements was $10,800/kW.

Current cost overruns seem to average over 50%, though it’s hard to find concrete data, as many projects get cancelled once cost overruns exceed 100%. At $67/MWh capital cost, a 50% cost overrun adds an additional $33/MWh.
Progress Energy in Levy county Florida is more than 20 months delayed before even breaking ground on the project. Duke Energy Lee nuclear power project (Cherokee county, SC) was originally slated for 2016, but is now projected to finish in 2021. They have also seen their planned plants in Gafney (SC) and New Hill (NC) delayed. In fact, every nuclear power plant that has been proposed in America during the “nuclear renaissance” has seen some delay or has eventually been cancelled.

These delays can be extremely costly to investors. At a 7.5% discount rate on a $16B project, every month of delay costs investors over 100 million dollars, so a 20-month delay can make a very large difference in the cost of energy. For the Burke, Georgia, plant specifically, the $14B project, for a 20-month delay, would see a financing cost penalty of $92M/month. This would amortize into an additional $10/MWh.

Very likely, the spiraling costs seen for new nukes in the US over the past seven years will continue until at least the first new plant is up and running, which may be in 2017.

Political Factors Contributing to delays and Cost Overruns.
It is important to realize the distinction between the countries that have seen success (few delays) in construction verses those which have seen major delays and cost overruns. France and China are obvious examples of nuclear success, while the U.S., Canada, and most of Europe fall clearly into the latter camp. The issue is not “misguided” environmentalism – as that is a constant in all cases. The issue is the relative authority of local government verses central government, and the relative power of the citizenry verses the government.

In France, they have a unitary government. The central government issues an edict, and the local "governments" set out to obey that edict. So in France they have a single design for their nuclear power plants, and all the components for that power plant design are mass produced. When a new power plant is needed (as determined by central planners), then the site is chosen, the concrete is poured, and a new power plant is built, assembling these mass-produced components according to a standard design that has been assembled many times by experienced workers.

This is not what happens in America. The framers of the constitution chose a different division of power between federal, state, and local authorities. In America, if a power company seeks to build a nuclear power plant, then the local authorities must approve it. NIMBYism is a powerful obstacle here, as the closer one gets to the planned site of a nuclear power plant, the more tepid the support and the more vehement the opposition. The local authorities don't want to lose support from either faction, so they buy time by ordering studies: environmental impact studies, grid capacity studies, aesthetic visual impact studies, wildlife impact studies, etc... Then they invariably determine that something must be altered about the plans so that the plant can conform to local codes. This means that the designs at large must be altered – requiring custom engineering and manufacture of thousands of individual components. There has been a concerted effort to streamline siting, licensing, and certification processes in America to reduce these issues, but a single lawsuit from a grandstanding local politician can result in a new study and a re-design.

These forced customizations are more of a problem here in America than is immediately appreciated. The nuclear power-plant construction industry in America is dead. Until very recently, there hadn’t been a new nuke ordered in the US since 1973 (four years before the Three-mile Island meltdown), and the number actually operating peaked in 1991. The number of trained nuclear engineers and physicists employed in the field of nuclear energy dropped about 4% annually during the 1990’s and has been approximately stable since then at about 10,000. This serves to severely limit the ability to design custom components and adapt existing designs to accommodate them.

Perhaps it will be possible to someday again build nuclear power plants in North America at a competitive price, but it will take at least two decades and hundreds of billions of dollars to resuscitate an industry as complex as the nuclear industry. The soonest a new nuclear power plant could begin producing power in the US is 2017.

Nuclear advocates are quick to point out that some nuclear power plants may demonstrate a lifetime of 60 years, while some wind turbines have failed within a few years. However, 21% of nuclear power plants in the US have been shut down prematurely, and the average lifetime worldwide is only 25 years. Wind turbines could soon be designed for a 60-year (or more) mean lifetime in most on-shore climates at very little additional cost, as factors affecting fatigue lifetime in composites have only recently begun to be well understood. Few experts are expecting any current nuclear power plants to be approved for additional extensions beyond 60 years, as the NRC has recently noted that the cracks developing in the aging plants pose serious safety risks.

World Outlook.
Based on plants under construction as of July 2010 (see World-Nuclear-Org), 61 GW of new nuclear generating capacity will be coming online over the next six years: 2010, 6.5 GW; 2011, 8 GW; 2012, 9.8 GW; 2013, 13 GW; 2014, 13 GW; 2015, 10 GW 2016, 6 GW. None of this will be in the Western Hemisphere, and only 5 GW of it will be in Europe. According to a recent study by the German Federal Ministry of Environment, world nuclear energy production peaked in 2002 (see Figure below). However, China is clearly expanding their nuclear program very quickly. The Chinese government has authorized an expansion of over 25 GW_E of new nuclear power, and they are expected to continue to expand authorizations for as much as 130 GW_E of new capacity. In China, the government licensing process is certain, the obstacles are non-existent, the labor is inexpensive, and the liabilities and insurance are not an issue. Providers of nuclear power plant components will be much more comfortable dealing with China than with the uncertainty and red tape offered in Europe and America. India also is poised to attempt their own massive expansion within the next decade. These build-out programs will result in some key components being tight for the first world.

Global nuclear energy production is unlikely to again match the peak seen in 2002 for at least the next 15 years [16].

If the average 25 year-old world fleet does reach 40-50 years and if China and India are successful in their nuclear expansions, there will be a new problem: Uranium costs.

Future Uranium Prices.
Historically, the price of uranium has been artificially low as a direct result of state funds invested in the nuclear industry, which world-wide over the past 50 years have exceeded $300B in current dollars. An enormous amount of excess uranium inventory was built up prior to 1985, so prices and production (mining) dropped in half between 1980 and 1993. Since then, mining production has ranged from 45% to 78% of consumption. Since 1999, the US has been purchasing weapons-grade uranium and plutonium from Russia at a rate sufficient to supply about 13% of the world’s reactor requirements.

The price of natural uranium ore (U₃O₈) increased from under $10/lb in 2002 to $60/lb as of December 2010. There was a major correction in the price of uranium in 2008 for a combination of reasons: (1) the financial crisis, (2) increases in uranium reserves estimates and production from new mines, (3) it became clear that the anticipated nuclear renaissance would be much smaller than earlier thought, and (4) speculative investors began to appreciate that global uranium inventory had been 2-3 times greater than officially reported.

The last time the price of uranium peaked (thirty years ago), the ores being mined (with government subsidy) had about 50 times higher uranium concentration (over 5%) than most of the ores that will be left 30 years from now (about 0.1%) – and these ores are also generally harder and found deeper. Moreover, the CO₂ released from the mining and refining of the ores remaining after 2050 could then make the total life-cycle CO₂ release from a conventional nuclear power plant significant. The energy required to refine ore that has 0.1% U is several orders of magnitude more than what is needed to refine ore that has 5% U. Undoubtedly, technical progress will be made in the mining and refining processes, but there are limits.

Nuclear proponents (and this includes such distinguished bodies as the MITEI, in their recent update on nuclear energy) often vastly overstate the abundance of uranium by including its occurrence in ultra-low-grade hard ores (below 0.01%), in granite (4-5 ppm), and even in the oceans (0.003 ppm). Uranium is several orders of magnitude more abundant in the ocean than are most rare metals because its common oxides are much more soluble. Realistic, recent estimates put the cost of extraction of uranium from the ocean at over $20,000/kg – a few percent of the cost of small demonstrations thus far. The cost of separating uranium from granite may be higher.

A huge global glut of uranium, driven by the cold war (1953 to 1989), kept prices artificially depressed for the last 30 years, but that is changing. Total US inventories are currently about 50 kT of U₃O₈ equivalent, and current US annual consumption is about 24 kT. Global uranium mine production is currently about 78% of consumption. For the past 4 years, production has been increasing at the mean rate of about 9%/yr (~51 kT for 2009), while usage has been decreasing about 1%/yr.

Our latest estimate (based on best available data) is that the current global inventory-to-usage ratio is under 1.5 years, or about one-fifth of what it probably was at its peak in 1983. (We note that unbiased information is difficult to get because the World Nuclear Association has become a blatant advocacy organization and the International Atomic Energy Association makes limited data available for free.) By 2013 most of the HEU materials (weapons-grade, over 20% 235U) that is being reprocessed into reactor-grade will be used up (though the inventory of HEU will still be over 1000 T). If mining growth averages 9% for the next four years and usage increases 1%/yr, the global uranium inventory would be at the critical 8-month level by 2014 and the price of uranium would soar. If the more optimistic expansion rates for global nuclear power are realized, and China alone sees a global increase in online nuclear reactors of 5%/year, then this critical level will be achieved sooner.

The most dramatic growth in production over the last few years has been in Kazakhstan, which saw uranium production in the first quarter of 2010 of 4060 tons compared to 6640 tons for the year of 2007. Kazakhstan is now the world’s largest uranium producer. If strong mining growth continues there as recently projected, uranium prices will probably not increase rapidly for several more years, but the potential for Kazakstan to offset both the diminishing inventories and the rapid growth of China’s nuclear program is highly doubtful.

Official estimates of recoverable uranium resources at $130/kg were about 5.5 Mt (~80 yrs worth) in January 2007 – a 50% increase over 2002. However, those estimates apparently assumed oil will be $30/bbl. (Several other significant problems with those studies have been detailed in Part III of “The Future of Nuclear Energy: Facts and Fiction”, by Dr Michael Dittmar of the Institute of Particle Physics of ETH, Zurich.) The cost of the uranium in a free market is mostly dependent on the cost of the liquid hydrocarbon fuels (used by the mining equipment) and on the ore concentration, its depth, and its hardness. A correlation between the price of oil and the price of uranium is obvious from the market data.
Best available data on the abundance of uranium in the earth’s outer continental crust (the outer few km, that accessible to mining) put it at about half that of hafnium, which is produced in large quantities along with zirconium from abundant heavy-mineral sands. (For comparison, zinc is 30 times more abundant.
See http://en.wikipedia.org/wiki/File:Elemental_abundances.svg.)

The current cost of hafnium is about $300/kg – but it is not radioactive, and its high-grade reserves are not being quickly depleted. (Annual uranium usage relative to its natural abundance is four times that for hafnium.) Hence, it seems unrealistic to expect uranium to drop below $600/kg after the excess inventories are gone in 2015 and oil is over $180/bbl. A realistic estimate for the price of uranium in 2020 might be $700/kg – and perhaps $3000/kg in 2050 when all the high-grade and mid-grade soft ores are gone.

The analysis by Dittmar (mentioned above) of the official data in the latest “Red Book” suggests that a more realistic estimate of the Reasonably Assured Resources (RAR) of uranium at under $130/kg is probably under half the widely mentioned 5.5 Mt. One of the more alarming points of Dittmar’s analysis is that uranium production is even more concentrated than oil production. The three largest mines (McArthur River in Canada, Ranger in Australia, and Rossing in Namibia) produce 33% of the world’s uranium. (This fraction corresponds roughly to the entire OPEC share of world oil production.) The large uranium mine in Canada is probably now past its peak.

At $700/kg and assuming 2%/yr increase in enrichment and fabrication costs, the cost of the fuel alone for nuclear energy would be over $20/MWhr in 2020.

Operation and Maintenance.
O&M costs are typically higher for the nuclear power plant industry than they are for other power sources due to the safety issues involved. In 2009, the average O&M related expense for nuclear power was ~$15/MWh. However, the chance of a large-scale renovation or required re-work is a significant risk for investors. About 27% of nuclear power plants in the US have been shut down for a year or more for major re-work. As was discussed earlier with capital costs, any period of time that the plant is not producing incurs very high losses. A 27% likelihood of being shut down for a year is daunting to investors. New designs should be less likely to see these major re-works, but that cannot be assured until after the newer designs are built and performing for 30 years. This makes reasonable financing rates very difficult to obtain, and thus adds another $1/MWh to the O&M.

Waste Disposal and Decommissioning.
Much of the anti-nuclear rhetoric has focused on the supposedly “large additional costs” of waste disposal and plant decommissioning. However, these costs are less than would normally be assumed.

Waste disposal has normally added only ~$1/MWh, though that will be increasing. The US Supreme Court has ruled that a permanent waste storage site cannot be approved unless studies show little risk for 1,000,000 years into the future.

To finally decommission a plant incurs an additional ~$500 million penalty – a number that does not vary much regardless of the size of the plant being decommissioned. For a 1.1 GW plant that has a 40 year life with an 85% capacity factor, that works out to ~$1.5/MWh.

Summary of Costs.
So, for the first-world countries, an investor in a new nuclear power facility to begin operating in 2020 would be considering the costs summarized in the following table.

Levelized Cost of Energy from a new Nuclear Plant in 2020
Item	$/MWh
Initially expected capital costs	67
Cost overruns	33
Delays	10
Fuel costs	20
O&M	16
Waste disposal and decommissioning	2.5
Total	$148/MWh

That is a frightening total LCOE. If the plant is built with no cost overruns and no delays, the projected energy cost would still be $103/MWh.

The Market for Energy.
It must be stressed that every single plant that is under construction or consideration in the U.S. is baseload – a plant that maintains constant output. This means that the plant is effectively “always on”. As such, 40% of the energy produced by the plant will be produced during peak hours (7:00 am to 11:00 pm on weekdays) while the other 60% of the energy will be produced during off-peak hours (weekends and the middle of the night). As discussed in much greater detail on our “Stabilizing the Renewable Grid” page, the market price for off-peak energy is at least ~$30/MWh less than the price of peak energy. Often night-time energy is free or even has negative value in some regions.

The average U.S. peak energy value is ~$110/MWh, and the average off-peak energy value is about $80/MWh.

Since 60% of the energy produced by the nuclear power plant would be produced off-peak, even with no cost overruns or delays, a new nuclear plant bid today would lose $23/MWh 60% of the time, while gaining only $7/MWh the other 40% of the time, for an average loss of ~$11/MWh.

Thus, it’s fairly easy to see why so many government subsidies are required to get investors onboard, and it’s easy to understand why even with those government subsidies, most planned and licensed new nuclear power plants in America have been cancelled before ground has been broken.

At the very least, a uniform design is needed before nuclear plant costs can be brought under control – and then there would be significant obstacles within our current political structure.

Future Nuclear Possibilities.
It is possible that advanced fission reactor concepts with extensive reprocessing of their spent fuel will eventually become a reality. If so, this could eliminate any concerns about the cost of uranium (and CO₂ emissions from uranium refining) for several centuries, as the energy extracted per pound of natural uranium might be increased by more than an order of magnitude. However, one of the nation’s premier nuclear physicists, Frank von Hippel, has noted in “Recycling, more trouble than its worth,” (Sci Am., 2008) that recycling reduces the waste problem only minimally.

The biggest concern generally voiced about breeders and complex reprocessing is the increased security risk, as the reprocessing would invariably produce large quantities of materials that could easily be used in making high yield nuclear weapons, even by small organizations with limited resources. Perhaps this concern could be adequately addressed with suitable international agreements, but the chances of such agreements being negotiated do not appear likely in the foreseeable future.

Another major issue (that has received far too little attention) is that all prior attempts to demonstrate nuclear concepts compatible with very high utilization of fissionable resources have experienced serious technical problems, often related to stability control. These experiences raise major safety concerns that can only be satisfied after years of trouble-free testing of advanced demo reactors under strict supervision. The DOE appreciates this, and that is part of the reason they do not expect commercialization of anything other than “once-through” nuclear reactors for at least 20 years. Finally, the best available data on something close to an operational, commercial, fast-spectrum, breeder reactor (the SNR-300, Germany) suggests breeder reactors would be five times more expensive than current standard technology (pressurized water reactors).

Conclusions.
In summary, it is unrealistic to think that nuclear energy can provide a significantly increased fraction of the global energy needs within the next 15 years in first-world countries.

Is it possible that a new generation of nuclear reactors could be competitive and publicly accepted in the US 25 years from now? Yes. And well managed R&D should continue toward that goal.

Whether or not small reactors, reprocessing, and breeder reactors succeed commercially has little effect on the viability of the WindFuels RFTS plant or its development path. For now, the cheapest source of renewable energy in the US, Canada, and northern Europe is wind. If breeder reactors can produce electricity at competitive prices 25 years from now, some would eventually be used in place of wind to power new RFTS plants making alcohols and hydrocarbons. All of the essential “WindFuels” process-plant technology would still be needed to make “Nuclear-Ethanol”. Nuclear fission is a carbon-neutral energy source, but it will not likely be a major source of energy for RFTS fuels for at least three decades due to its cost.

An exeptionally good series (2009) on nuclear energy is:
“ The Future of Nuclear Energy: Facts and Fiction”, by Dr Michael Dittmar, 2009.

Part I: Nuclear Fission Energy Today
http://europe.theoildrum.com/node/5631

Part II: What is Known about Secondary Uranium Resources.
http://europe.theoildrum.com/node/5677

Part III: How (un)reliable are the Red Book Uranium Resource Data?
http://canada.theoildrum.com/node/5744

Part IV: Energy from Breeder Reactors and from Fusion?
http://www.theoildrum.com/node/5929

And here’s an exceptionally good overview of the big picture on nuclear energy possibilities from the optimistic side:
http://nucleargreen.blogspot.com/2010/04/kirk-sorensens-thorium-energy-alliance.html Apr 1, 2010

Recent nuclear plant price quotes, ~$10/W:
http://climateprogress.org/2009/07/15/nuclear-power-plant-cost-bombshell-ontario/

-------------------------------------------

References:

1. OECD NEA and IAEA, “Uranium 2007: Resources, Production and Demand”, OECD, 2008.

2. Mark Cooper, “The Economics of Nuclear Reactors: Renaissance or Relapse”, Vermont Law School, 2009. http://www.vermontlaw.edu/Documents/Cooper%20Report%20on%20Nuclear%20Economics%20FINAL%5B1%5D.pdf

3. http://www.nea.fr/html/databank/ , 2009.

4. US U₃O₈ inventories,

5. Peak uranium, http://en.wikipedia.org/wiki/Peak_uranium

6. Enriched uranium, http://en.wikipedia.org/wiki/Enriched_uranium

7. http://www.forbes.com/business/global/2007/1029/018.html

8. Oxford Research Group, http://www.oxfordresearchgroup.org.uk/work/global_security/energy.php

9. http://nucleargreen.blogspot.com/

10. http://www.world-nuclear.org/info/inf02.html

11. Joe Romm, “The Self-limiting Future of Nuclear Power”, Climate Progress, 2008, http://www.americanprogressaction.org/issues/2008/nuclear_power_report.html.

12. David Lochbaum, UCS Testimony to Congress, 2008, http://www.ucsusa.org/assets/documents/clean_energy/20040520-ucs-senate-testimony.pdf.

13. Nuclear Age Peace Foundation, Fact Sheets, 2008, http://www.wagingpeace.org/menu/issues/nuclear-energy-&-waste/start/fact-sheet_ne&w.htm

14. http://www.energyfromthorium.com/pdf/

15. Recent agreement with Russia on Uranium, http://theenergycollective.com/TheEnergyCollective/46583

16. German Federal Ministry of Environment, “Nuclear energy will continue to decline...”, http://www.bmu.de/english/nuclear_safety/downloads/doc/44832.php