As Ian Hore-Lacy and I have learnt from each other for over 30 years, some important points of agreement have emerged. For example, I heartily endorse the opening of the Charter of Ethics of his employer since 2001. It declares the World Nuclear Association’s “determination to promote, as a matter of ethical principle and urgent public need, an ongoing debate on energy resources that focuses citizens and governments alike on the real choices facing humankind and on the severe dangers—for the prospects of global development and for the biosphere—if decision-making on this fundamental policy is shaped by ideology and myth rather than by science and facts.” In that spirit, I wrote the NEI article ‘Mighty Mice’ (Dec. 2005, p44). Perhaps this response to Mr Hore-Lacy’s criticisms of it may usefully clarify key issues and help to cool overheated rhetoric.

Most or all of his issues with my article are resolved by its cited technical backup paper, posted for free download from www.rmi.org/sitepages/pid171.php#E05-14. That paper shows clearly the graphics that may be hard to read on NEI’s website, and it documents my entire analysis. It thus details the calculations behind Fig 3’s cost analysis for all technologies (from nuclear to efficient end-use); nuclear power’s historic and continuing subsidies (in notes 6, 13, 45, and 63); and the $0.0275/kWh delivery cost that roughly reconciles Mr Hore-Lacy’s nuclear busbar cost with Fig 3’s delivered cost. Once he’s read my underlying analysis, other NEI readers would doubtless join me in welcoming his and anyone else’s substantive critiques.

My article adduced “science and facts”—empirical evidence—that nuclear power faces formidable if not insurmountable competition from both supply- and demand-side decentralised alternatives. Mr Hore-Lacy welcomes those but “for the most part…does not see them as competing”. Yet the market does: investors have lately added upwards of ten times more electric capacity pa from these decentralised options than from nuclear power. If that’s not a competitor, what is? Since it supplies or saves kWh, why won’t it pose an increasingly serious threat to new nuclear build?

One reason might be Mr Hore-Lacy’s apparent view that big generating units are needed to do a big job—much as computing experts considered mainframe computers in the pre-PC era. The debate about power systems’ architecture and optimal unit scale is no mere abstract mindgame; on its practical outcome hangs the nuclear industry’s future. Why can’t the electric output from one 1GWe unit be produced instead by, say, a hundred 10MWe or a hundred thousand 10 kWe units? The Economist often describes the market’s rapid shift to ‘micropower’; on what grounds does the WNA base industry strategy on the supposedly manifest and ineluctable superiority of central stations? Might not industry leaders benefit from deeper enquiry into scale?

Reliable power generation too is an important issue, but proof by vigorous assertion and refutation by emphatic dismissal are no substitute for reasoned analysis. My team’s Economist book of the year (2002), Small Is Profitable: The Hidden Economic Benefits of Making Electrical Resources the Right Size, available at www.smallisprofitable.org, explains in depth why “many small units near customers are more reliable than fewer, bigger units far away.” The reasons for this are rigourously founded in engineering theory and practise. In brief, modular distributed generators have far higher and more-predictable collective availability than large non-modular units even if those have far higher unit availability (which for many technologies they don’t). And nearly all power failures in developed countries originate in the grid, so reliable power delivery requires generation at or very near the customers. This is not the fault of central stations, but it does affect their value proposition and their market.

No source of electricity is 100% available; the sent-out kWh costs of variable renewable generators already reflect their capacity factors, just as do the corresponding costs for thermal stations. Moreover, my analysis already includes, at above-market cost, enough backup and grid integration to make windpower completely dispatchable—available whether or not the wind is blowing. But it’s not “blindingly obvious” that, as Mr Hore-Lacy claims, “largish plants such as nuclear ones” are “much more reliable than wind and solar [power].” The average US nuclear unit’s latest major outage (>12 days at zero power) from whatever cause, planned or forced, lasted 36 days and occurred at a 17-month interval (15 Nov 2005 data, www.nei.org/documents/NuclearPerformanceMonthly.pdf). No wind or solar technology has such poor outage statistics. For example, modern wind turbine units are typically ~98–99% available, a long calm lasts a few days, and 34 years’ UK wind statistics show no instance where wind turbines nationwide would have been simultaneously becalmed nor shut down by high windspeeds. Indeed, windpower is highly correlated with UK electric loads, increasing their capacity factor and cost-effectiveness (www.eci.ox.ac.uk/renewables/UKWind-Report.pdf).

Of course, one must consider regional as well as single-unit reliability. Post-scram xenon and samarium poisoning and other physics and operational issues thus become relevant in case of a regional grid failure or other major common-mode interruption. After the northeast US/Canadian blackout on the afternoon of 14 Aug 2003, the nine scrammed US nuclear units achieved 0% output on the 15th, 0.3% on the 16th, 5.8% on the 17th, 38.4% on the 18th, 55.2% on the 19th, and 66.8% on the 20th (www.nrc.gov/reading-rm/doc-collections/event-status/reactor-status/2003/index.html). Those units thus lost an average of 97.5% of their rated capacity for the first 3 days, 82% for 5 days, 59% for 7 days, and 54% for 12 days—hardly a reliable resource no matter how exemplary their normal operation. This inability to restart promptly after a major grid outage makes nuclear plants least available when they’re most needed—a unique ‘anti-peaker’ attribute. Similarly, TEPCO’s 2003–4 safety shutdown closed all 17 of its nuclear units for many months. Since the grid is designed to, and does, tolerate such massive and prolonged nuclear or other central-plant outages, albeit at considerable cost for reserve margin and replacement energy, why is the occasional and predictable becalming of windfarms over a relatively small time and space, offset by higher output from statistically complementary renewable resources, such an insurmountable problem? And why does the WNA penalise windpower but not central thermal stations for requiring reserves? All resources require reserves, and their economics should properly reflect the amount that they require.

However, reserves are not a function purely of the statistics of an individual unit. Grid operators and major market actors are ultimately not so interested in the reliability of single units, nor even of regional assemblages of units. Rather they focus more on the overall delivered-service reliability of the full portfolio of technologically and geographically diverse units, integrated into a grid with diverse power sources and demand-response options, all appropriately forecasted, and optionally with storage (like the pumped-storage schemes sometimes associated with nuclear units). In this practical operational context, recent technical literature (see backup paper, note 30) does not consider wind and solar variability a serious issue even at high market shares. These are already 20–30% annually and >100% daily in some regions yet incur very modest integration costs. So again the scope for using variable electric sources becomes an empirical question. If the market does not share the nuclear industry’s belief that variability will severely constrain its renewable competitors, this collision between hope and competitive reality could dash nuclear expectations. It is thus another basic issue that the industry needs to understand based on careful analysis not slogans.

Mr Hore-Lacy assails a 1988 Energy Policy paper that I quoted to show that the climate-opportunity-cost argument isn’t new and hasn’t changed much; but he doesn’t show error in either the old or the much fuller new analysis. I agree that Vattenfall’s recent CO2 audit of the nuclear fuel cycle may be valid in its own narrow terms (backup paper, note 20). But far from “demolishing” my climate argument or rebutting Vattenfall’s own 1989 nuclear-phaseout scenario, the Vattenfall CO2 audit is irrelevant to both. The climate issue is whether buying new nuclear build releases more or less CO2 than using the same money to buy a larger amount of lower-cost low- or no-carbon alternatives.

That’s not about “funny arithmetic”; it’s about asking the right question. If new nuclear build would deliver electrical services at a higher economic cost than alternatives with similar carbon emissions, as Fig 3 shows, then those alternatives would provide more climate solution per dollar (which Mr Hore-Lacy agrees is the right goal). If so, then the nuclear industry’s climate-protection argument falls to the ground. A responsible trade association must address such a central issue. To sustain its climate-protection claim, it must show convincingly that new nuclear build can displace coal-fired generation not only cheaper and faster than other central stations but also cheaper and faster than any other supply- and demand-side option (adjusting appropriately for any carbon released). As an initial effort to test that hypothesis, my analysis offers Fig 3. If its numbers or logic are wrong, I should be grateful to know specifically how.

Why did my article assert that WNA “denies the existence” of decades of reputable scholarship documenting plausible non-nuclear futures? Because it does. WNA’s posted February 2005 ‘Sustainable Energy’ position paper states: “Recent analyses fail to come up with any 50-year scenario based on sustainable development principles which does not depend significantly on nuclear fission to provide large-scale, highly intensive energy, along with renewables to meet small-scale (and especially dispersed) low-intensity needs….Nuclear energy’s opponents have yet to credibly suggest how we should produce most of our future electricity. Certainly all the reputable energy scenarios show the main load being carried by coal, gas, and nuclear, with the balance among them depending on economic factors in the context of various levels of greenhouse constraints.” In fact, a large and compelling literature, much of it funded by EU governments and conducted by respected independent analysts, shows the opposite. I cited such a study by Vattenfall—an organization no less competent in 1989 than it is today. Mr Hore-Lacy says any such non-nuclear scenario is “extremely unpersuasive or irrelevant. We therefore ignore it. Who wouldn’t?” Anyone pursuing disinterested enquiry in the open spirit so well enunciated in WNA’s Charter of Ethics wouldn’t. Divergent views should be reconciled by detailed dialogue, not ignored as unwelcome or implausible. Hermetic belief systems immune to critical scrutiny often lead to troubled industries, strategic errors, and stranded investments.

I was aware of the proposed 10–20-MWe mini-reactors that Mr Hore-Lacy kindly mentions, but chose not to include them in Fig 3’s cost comparison for two reasons: my analysis shows only commercial technologies, and the central nuclear plants it shows have better economics than mini-reactor designs. The remote fly-in village of Galena, Alaska, is Mr Hore-Lacy’s prime example of a worthy application of a 10MWe scaledown of Toshiba’s 50MWe 4S sodium-cooled fast reactor design, unconservatively assumed to cost the same $2,500/kWe at one-fifth as at full scale. USDOE’s supporting study says that if that $25-million reactor and its licensing (for tens of millions of dollars), installation, removal, and decommissioning were free, if O&M costs were half Toshiba’s estimate for the 50MWe design, and if NRC dropped the required security staffing from 34 to 4 guards, then the ~$0.05–0.14/kWh operating cost alone may compete with today’s diesels burning costly barged-in fuel. (Paying the reactor’s capital cost would add ≥$0.09/kWh to the tariffs.) But on closer examination, the USDOE study is less encouraging (www.iser.uaa.alaska.edu/Publications/Galena_power_final.pdf), illustrating the economic challenges facing nuclear power even under these seemingly very favourable conditions:

• The study seeks to minimize average tariffs, not total costs of service, and doesn’t assess any risks. Its cursory page-and-a-half “detailed discussion” of end-use efficiency conservatively mentions potential cost-effective electric savings of one-third in homes and one-half in schools, yet assumes that none will occur and price elasticity is zero. That’s a dangerous assumption, because even modest savings would drop oil costs below minimum nuclear operating costs, as actual 2003 oil costs appear to have been.

• Galena’s electric generation is projected to rise 2.0% pa, yet sales fell faster than that during FY2001–03, whilst losses, with only 14km of lines, more than doubled to 13.8%.

• The high-oil-price case that generates large nuclear benefits escalates assumed 2039 crude-oil prices over $300/bbl (2004 $) —20x the State of Alaska’s forecast for 2015.

• Decentralised competitors are assumed away, e.g. by using windpower criteria for low-power-cost sites (Class 2–3 sites adequate to beat Galena’s high power costs are relatively nearby). Variable sources are rejected even though all options, including the nuclear plant, would require full diesel backup plus diesel replacements ~2010. Such promising non-nuclear options as fuel cells, whose hydrogen storage option and pure hot water byproduct could be valuable, are rejected as not yet commercial, yet the nuclear design isn’t commercial either, and its feasibility wouldn’t be known until at least 2010.

• A 10MWe nuclear plant could supply ~11x the town’s current winter average load and 5.6x its projected 2010 or 3.1x its 2039 peak load. Three-fifths of annual electricity has so far been sold to a military base subsequently slated for closure. With or without the base, even the one-fifth-scale plant would thus be very oversized. Part of its huge surplus of nuclear electricity and heat is assumed to be used by building large greenhouses or by cross-subsidising electric heat, which otherwise couldn’t compete with efficient oil heat. (Oil heat would need to remain ready for use in any prolonged nuclear backup, though, as backup heat by diesel electricity would quickly bankrupt the community—if not outrun capacity.) Most of the nuclear output would still remain unused—“available for hydrogen production”, which the study finds uneconomic even if the electricity costs nothing.

A cheaper solution could start with superinsulation (the average house now needs ~15kWt) and electricity-saving retrofits. My long-ago consultancy for the State of Alaska predicted ~6–7x village electricity-efficiency improvements that I later heard had been experimentally verified. A least-cost heating strategy could also retrofit home oil-heating units from the current ~75% to 95% efficiency, or use modern district heating (Galena’s current 4.3MWe of diesels discard half their heat), or adopt wood CHP (the town is now 31% wood-heated, and the similar-sized Oujé-Bougoumou Cree Nation community uses wood-and-oil-fueled district heating). Galena could also use innovative renewables to overcome the objections raised—under-ice $1000–1500/kWe Yukon River damless turbines (rejected even though cheaper than nuclear), perhaps modern PVs, and remote, balloon-mounted, or low-speed wind turbines (some work well today in harsher conditions in Alaska and Antarctica).

Illustrating efficiency potential, my own superinsulated house in the Rockies is 99% passively heated despite low temperatures only 8 C˚ above Galena’s –53˚C; its ~99% space- and water-heating savings and ~90% electric savings paid back in 10 months with 1983 technologies. For far less than the reactor’s $25 million cost—$114k for each of Galena’s 220 homes—I daresay they could be rebuilt so efficiently as to use almost no heat and electricity. In contrast, the 10MWe proposed to be supplied to ~700 people is equivalent, for each home, to ~400x what my 372-m2 household uses, or ~1,000x what it would use with 2005 efficiency techniques, both before crediting my solar power production.

In conclusion, perhaps I was unfair to imply earlier that Mr Hore-Lacy missed my article’s core message, for at one point he does precisely encapsulate its logic: that “all we need to do is keep saving energy and never again build any nuclear power plant (or by implication anything else that delivers power at similar cost).” Exactly my point: the combination of efficient end-use with the cheaper decentralised generating technologies (chiefly CHP and renewables) makes new central stations unnecessary and uneconomic. That cheaper combination is just what the market is increasing¬ly buying (Figs 1–2), uninhibited by doctrinaire rejections of smaller or more variable competitors. Providers of central stations, whether nuclear or fossil-fueled, need to understand who is right—Mr Hore-Lacy or today’s marketplace.

I appreciate his praise for my presenting evidence about the success of these decentralised alternatives, and I invite him to join in my effort to help the nuclear industry better appreciate its competitive position. An industry that does not understand who its competitors are, what its market economics are, or even that it receives substantial subsidies will not make good use of its practitioners’ talents, its customers’ goodwill, or its investors’ money. The gifted technologists and hardworking operators who have sincerely devoted their careers to our common challenge of affordable energy for sustainable development deserve better.

Amory B Lovins

CEO, Rocky Mountain Institute

Snowmass, Colorado, USA


Related Articles
Two construction milestones in China
Westinghouse and Curtiss-Wright team up
UK regulators query AP1000 ventilation system
The world’s largest canned motor pump
UK ships first nuclear power component to China