Skip to content


The Cost of Sustainable Low-Carbon Electricity — Wind vs. Nuclear


United States Wind Resource Map (50 Meter Elevation)


Let’s assume that in order to reduce CO2 emissions and prepare for the day when fossil fuels run short, Congress decides to put a price on carbon, and thus coal and natural gas are no longer the lowest-cost ways to generate electricity, as they have been for many decades.

Given that assumption, we believe the public debate about renewable electricity has failed to recognize the following key points:

  1. that the public doesn’t want renewable electricity — it wants sustainable, secure, low-impact electricity at the lowest feasible price;
  2. that, although there could be six or seven new sources of sustainable carbon-free electricity over the next 10-20 years (hydro, biomass, wind, solar, nuclear, enhanced geothermal, and coal with carbon storage), the only scalable ones that we know how to build today are wind and nuclear;
  3. but wind turbines alone cost more than nuclear (on a full-time-equivalent basis);
  4. wind requires long new transmission lines which nuclear does not;
  5. and, in the absence of electricity storage, there is no such thing as wind by itself — there is only 30% wind combined with 70% natural gas, or 30% wind combined with 70% coal.
  6. Moreover, wind is not likely to benefit much from further economies of scale because it’s already a $50 billion global industry.

    Consequently, wind is not a cost-effective solution for either CO2 emissions or the fossil-fuel shortage, and subsidizing it today is not likely to lead to cost reductions in the future.  Therefore, discriminating against nuclear power (or coal with carbon storage), as 27 states have chosen to do through renewable electricity standards, has to increase the price of electricity.

    Electricity Demand and Baseload Sources

    Electricity demand typically follows a daily cycle which peaks in the afternoon and declines to a minimum overnight.  Wind and nuclear supply what is called baseload demand, or the level that’s present all the time.  Based on reports filed by the nation’s utilities with the Federal Energy Regulatory Commission, about 75% of electricity consumption is baseload and about 25% is intermediate or peak load.

    Wind has to be counted as a baseload source because it’s mostly available outside the hours of higher demand.  Nuclear power is a baseload source because the plants need to run steady-state.

    How Do We Produce Electricity Today?

    According the Energy Information Administration (EIA), in 2008, 96% of U.S. electricity came from just four sources: coal 48%, natural gas 22%, nuclear 20% and hydro 6%.  The remaining 4% came from biomass (1.4%), wind (1.3%), oil (1.1%) and geothermal (0.4%).

    Demand is Full-Time, But Wind is Part-Time;  Demand is Local, But Wind is Remote

    All previous sources of electricity, except wind, could run full-time. Thus a nuclear plant could substitute for a coal-fired plant, but a wind farm could not, unless it were combined with a backup source or with electricity storage.  But due to both its cost and environmental impact, storage has not been proposed for any major wind development.

    Likewise, all previous sources of electricity, except for some hydro and some wind, have been located near centers of demand.  But given the wind resource map shown above, we can see that any significant wind development will have to take place in remote locations on the Great Plains.

    Full-Time Equivalent (FTE) Cost

    Unlike electricity generation from fossil fuels, the economics of wind and nuclear are dominated by capital costs.  And, for baseload sources, the cost that matters is the full-time-equivalent cost, the one derived by dividing the nameplate cost by the capacity factor (the ratio of actual production to full-time production at nameplate capacity.)

    The following table compares three recent estimates for the cost of new nuclear power with recent reports for the cost of wind, natural gas and coal, all in 2007 dollars.

    [We can present a more traditional levelized cost of electricity analysis in a future post, but LCOE calculations often obscure the underlying information which we want to emphasize here.]



    Dollars per Nameplate kilowatt

    Capacity Factor

    Dollars per Full-Time kilowatt





    Combined Cycle Natural Gas

    800

    0.65

    1200

    Supercritical Coal

    2200

    0.85

    2600

    Nuclear Black &Veatch  10/07

    3200

    0.9

    3600

    Nuclear South Carolina Electric & Gas  5/08

    3400

    0.9

    3800

    Nuclear Electricite de France  12/08

    3500

    0.9

    3900


    Wind Class 3 Black & Veatch  10/07

    1650

    0.35

    4700

    Wind  Am Wind Energy Assoc / EIA 2008

    2000

    0.3

    6700

    Wind + 550 miles  Am Electric Power

    3000

    0.3

    10,000

    Wind + 1000 miles

    4000?

    0.3

    13,000?


    Description

    The first three nameplate costs and the first nameplate cost for wind are from a study by Black & Veatch, commissioned by the American Wind Energy Association (AWEA) in 2007.  The remaining nameplate costs for nuclear are from Electricite de France (EdF) and South Carolina Electric and Gas (SCE&G), and for wind are from data reported by AWEA.

    All of the capacity factors are from Black & Veatch, except the one for wind, which was calculated from data reported by AWEA and EIA.

    EdF’s and SCE&G’s cost estimates are from actual projects – Electricite de France’s most recent press release on the European Pressurized Reactor under construction in Flamanville, France; and South Carolina Electric and Gas’s engineering, procurement and construction contract for a reactor which it anticipates completing in 2016.

    The costs of inflation, financing and transmission are not included, except for the rows labeled “Wind + 550 miles” and “Wind + 1000 miles.”

    See the appendix for more information on sources and calculations.

    Observations

    The first observation is that wind’s cost and capacity factor are open to dispute, and seemingly modest differences in the assumptions can lead to large differences in the FTE cost.  While the nameplate cost increase from 2006 (the date of B&V’s data) to 2008 is probably not in dispute, the capacity factor for future projects is.  All we’re pointing out is what wind turbines actually delivered in 2008.

    A second observation is that, on an FTE basis, the cost of wind turbines alone is higher than the current and projected cost of nuclear, even if we added finance costs to nuclear or assumed a higher capacity factor for wind.  At the very least, one would have to say that on the basis of generation cost alone, wind has no obvious advantage.

    The third point to note is cost of transmission.  Even 550 miles of transmission (which is a conservative estimate for the distance from windy sites on the Great Plains back to major centers of demand) would raise wind’s cost by 50%.  1000 miles of transmission would most likely double it.  [High-voltage DC could perhaps reduce costs somewhat at the 1000-mile level.  More on that in another post.]

    Reliable estimates for new long-distance transmission are difficult to find because the most advanced technologies have not been implemented on large scale in the U.S.  We picked the 550-mile data point to use here because it comes from one of the few large-capacity transmission proposals which is based on real experience, American Electric Power’s proposal for a 765-kilovolt AC connection from West Virginia to New Jersey.

    Caveat

    Of course, the difficulty in making these comparisons is that no nuclear plants have been built from scratch in the United States for 25 years and the small amount of wind power installed to date has piggy-backed on the existing grid.  Thus for nuclear we can draw conclusions only from experience in Europe or from early-stage contracts signed by U.S. utilities, and for wind we can draw conclusions only by reviewing proposals for new transmission projects.

    But while these limitations should remind us to be cautious, they should not prevent us from making comparisons, because European nuclear regulations and labor rates are similar to those in the United States, and the proposed transmission lines are evolutionary extensions of what we have today.

    Wind’s Capacity Factor

    Although some studies have claimed onshore capacity factors of 40% or higher, production figures for 2008 from the EIA and AWEA show that the overall capacity factor for actual installations was 31%.  This was not a result of old technology.  Almost all U.S. capacity has been installed since 2001 and the majority has been installed since 2005.

    According to the EIA, 2008 wind generation equaled 52 billion kilowatt-hours.
    See table Net Generation by Other Renewables: Total (All Sectors)
    eia.doe.gov/cneaf/electricity/epm/table1_1_a.html

    According to the AWEA 2008 Annual Wind Industry Report:
    http://awea.org/publications/reports/AWEA-Annual-Wind-Report-2009.pdf
    8500 MW of wind capacity was installed in the United States in 2008 at a cost of $17 billion.  Total U.S. wind capacity was 17GW at the end of 2007, 21GW at the end of Q3 2008 and 25.3GW at the end of 2008.  From that, we can deduce that the average installed capacity in 2008 was about 19GW.

    Calculation: 52 billion kWh / (8760 hours * 19 GW) = 31%

    Why Is Wind So Expensive?

    Because wind is intermittent and the best sites are remote.  Remoteness and intermittency add three costs which no other sources of electricity have ever had: transmission, storage and backup.

    Transmission

    Contrary to popular belief, we have almost never transmitted power over long distances.  Our existing grid is primarily a local distribution network, not a transmission network.  And for good reason — transmission is expensive, and rights of ways are difficult to assemble.  In addition, the more tightly a system is connected, the larger the area that can go down in a major disruption.  Distributed generation is robust generation.

    We have typically produced electricity within 100 miles of where it was consumed.  Therefore, if we had to build a national transmission grid, its cost should be counted against the technologies which require it (wind, solar and perhaps enhanced geothermal)

    In contrast, nuclear plants can replace coal or natural gas-fired plants one for one, in nearby locations.  Thus the existing distribution grid can be re-used largely as it is.

    Storage

    Although we won’t examine electricity storage in detail here, the simple takeaway is that only one technology (pumped hydro) has ever been built at even modest scale, and environmental considerations would make it exceedingly difficult to build more.  Not to mention that suitable sites in the Appalachians are nowhere near the most desirable wind sites on the Great Plains.  Thus two different transmission systems would have to be built – one from the source to the storage facility and another from the storage facility to the load.

    The only other technology which could scale to match the requirement is compressed air energy storage (CAES.)   Unfortunately, while CAES might make sense at small scale, it has a fundamental drawback at large scale – its heavy dependence on natural gas.  CAES plus wind would consume about 60% as much natural gas as the most efficient combined cycle gas turbine alone.  But since CCGT could operate without any wind turbines, compressed-air caverns or new transmission lines at all, why build them?

    Backup

    There is no such thing as wind power by itself.  There is only 30% wind combined with 70% backup, and the only feasible choices for backup are hydro, biomass, natural gas or coal.  (Nuclear can’t be a backup because it can’t ramp easily.)  But hydro supplies only 6% of our electricity today (almost none of which is near the Great Plains) and can’t be expanded, while biomass supplies only 1.3% of our electricity today and will be costly to increase.  Natural gas and coal were the fossil fuels that we set out to eliminate in the first place.

    What’s the point of pursuing a technology that would lock us into 70% natural gas or 70% coal?

    Conclusion

    Power companies have never built remote, intermittent sources before because those sources didn’t make economic sense.

    And if we had any other choices for the future, remote intermittent sources wouldn’t make economic sense today, either.  We do have other choices.  Nuclear is a proven, scalable full-time source today, and enhanced geothermal or coal with carbon storage may be one tomorrow.

    ———————————————————————————————

    Why is the List of Scalable New Sources So Short?

    Because both our population and our per capita consumption are so large relative to the size of the continent.

    The only non-fossil sources which could plausibly join wind and nuclear include two which we use today — hydro and biomass — and three which are under development – thermal solar, enhanced geothermal (EGS), and coal with carbon capture and storage (CCS).  But all five of those are either limited in capacity or face long research and development cycles.

    [Three additional sources are not plausible.  Natural geothermal sources have mostly been developed.  Photovoltaic solar is much more expensive than its thermal solar cousin.  And nuclear fusion is too far off in the future to evaluate.]

    Hydro has been low cost, but no one believes that we could build much more.

    Biomass-fired generation could be expanded, but nowhere close to the size of U.S. electricity demand and not at low cost (or low environmental impact (the soil loses nutrients.)  Think of biomass as a low-quality form of coal, widely dispersed, and sitting outside where it collects moisture.

    Thermal solar has promise, but will face capacity factors of 20% or less outside the Southwest.

    Enhanced (deep-drilled) geothermal is years away from feasibility and its costs have yet to be determined.

    CCS coal could serve as a transition strategy, but the quantities of CO2 involved are so large that it may never be cost-effective to capture and store it, except in (relatively small) applications for enhanced oil recovery.  A one-gigawatt coal plant produces about 10 million tons of CO2 and 250,000 tons of ash per year, while the same size nuclear plant produces 20 tons of spent fuel and no ash or CO2.  Factors of 500,000 have consequences.

    ———————————————————————————————

    Sources for the Table of FTE Capital Costs:

    Black and Veatch,
    20% Wind Energy Penetration in the United States, October 2007,
    a study commissioned by the American Wind Energy Association (AWEA).
    www.20percentwind.org/Black_Veatch_20_Percent_Report.pdf

    Electricite de France, December 4, 2008.  Press release announcing that the cost of the 1.6 gigawatt Flamanville-3 European Pressurized Reactor, scheduled for completion in 2012, had risen to 4 billion euros, or 2500 euros per gigawatt (3500 dollars per gigawatt at an exchange rate of 1.4 to 1)

    http://weblog.greenpeace.org/nuclear-reaction/PR-EDF-Investor-day-nuclear-strategy-and-finance-04-12-08.pdf

    South Carolina Electric and Gas Co., April 2009.
    SCE&G Generation Strategy by Kevin Marsh, CEO.
    Available on SCANA Corp. investor relations page:
    http://www.scana.com/en/investor-relations/nuclear-financial-information/default.htm

    American Wind Energy Association, 2008 Annual Wind Industry Report
    http://awea.org/publications/reports/AWEA-Annual-Wind-Report-2009.pdf

    Transmission cost:

    American Electric Power, “Meeting America’s Future Electric Needs”, 2006
    http://www.aep.com/about/i765project/technicalpapers.aspx
    Several papers on this page refer to the following paper, for which there is no link:
    “The AEP Interstate Project Proposal – A 765 kV Transmission Line from West Virginia to New Jersey,” January 31, 2006.

    Calculation: Assume N-1 contingency — must build 3 lines in order to have 2 in operation after an outage.
    Capacity: 2 lines * 5000 MW maximum load per line = 10GW.
    Assume capacity is limited by power factor support facilities and line losses rather than by heating.
    Cost: 3 lines * $3 billion per line = $9 billion.
    Result: $1 billion per gigawatt for 550 miles.

    Notes for the Table of FTE Capital Costs:

    One could argue that inflation would affect all sources similarly, or nuclear might have an advantage because it uses 1/10th as much steel per FTE-watt as wind.

    Financing costs favor wind because of its shorter construction time.  SCE&G’s capital spending timeline indicates an average borrowing period of about 3 years, compared with 1 year or less for wind.

    Transmission costs favor nuclear.

    SCE&G’s estimate excludes inflation (estimated at a time-weighted average of 30%), financing and transmission (estimated at 10% of plant cost), but includes owner’s costs (such as site preparation) and a 10% contingency factor.

    All capacity factors represent maximum values.  When facilities are used to meet intermediate and peak loads, utilities will by necessity run them for fewer hours than their maximum capacity.

    ———————————————————————————————-

    Dis-economies Of Scale

    Unlike most industrial technologies, wind has dis-economies of scale.  That is, as more wind is added to a system, the cost per kilowatt-hour goes up.  Small amounts of wind, such as we have built to date, can piggyback on existing infrastructure.  But as more wind is added, new transmission lines must be built.  Once wind scales up to match peak demand, storage must be built.  In all cases, wind must be paired with a backup source whose capital costs must be paid for by ratepayers.

    Competing full-time sources, on the other hand, do not require long-distance transmission, storage or backup.

    ———————————————————————————————

    Cost of Fuel

    While wind’s fuel is free, that’s not much of an advantage.  Nuclear fuel costs about one half cent per kilowatt-hour, out of a total wholesale price of five to six cents per kilowatt-hour, in most parts of the country.

    Posted in Nuclear, Wind.