This is a talk I gave at the Detroit Edison Company on April 17, 2007. It's the most complete version summarizing our work on the historical costs of nuclear power. That worked appeared in three refereed journal articles:
Hultman, Nathan E., and Jonathan G. Koomey. 2007. "The risk of surprise in energy technology costs." Environmental Research Letters. vol. 2, no. 034002. July. <http: />
Hultman, Nathan E., Jonathan G. Koomey, and Daniel M. Kammen. 2007. "What history can teach us about the future costs of U.S. nuclear power." Environmental Science & Technology. vol. 41, no. 7. April 1. pp. 2088-2093.
Koomey, Jonathan G., and Nathan E. Hultman. 2007. "A reactor-level analysis of busbar costs for U.S. nuclear plants, 1970-2005." Energy Policy. vol. 35, no. 11. November. pp. 5630-5642. <http: />
The last article is the primary source--the other two derive policy insights from the data developed in that last article.
2007 Koomey talk on historical costs of nuclear power in the US
1. A Reactor-Level Analysis of
Busbar Costs for U.S. Nuclear
Plants, 1970-2005
Jonathan G. Koomey* and Nathan Hultman**
*Staff Scientist, Lawrence Berkeley National Laboratory
and Consulting Professor, Stanford University
**Assistant Professor, Georgetown University
Presented by Koomey at DTE, in Detroit, MI
April 17, 2007
JGKoomey@stanford.edu
http://www.koomey.com
Copyright Jonathan Koomey 2007 1
2. Recent interest in nuclear power
• Climate change mitigation issues (e.g.
Blair announcement)
• EPACT 2005
– Subsidies
– Loan guarantees
• International investments (e.g. China)
• Expressions of interest by U.S. utilities
Copyright Jonathan Koomey 2007 2
3. New plant status from NEI
Updated 4/2007
FY = Federal fiscal year, CY = calendar year Copyright Jonathan Koomey 2007 3
4. Goals of this work
• Creating fair, accurate, and comparable
historical cost estimates
• Explaining history using limiting cases and
statistics
• Comparing historical cost estimates to recent
projections of nuclear costs
• Characterizing mistakes to avoid if nuclear
power is to undergo a rebirth in the U.S.
Copyright Jonathan Koomey 2007 4
5. Status of work
• Hultman, Nathan E., Jonathan Koomey, and Dan
Kammen. 2007. "What history can teach us about the
future costs of nuclear power." Environmental Science &
Technology. vol. 41, no. 7. April 1. pp. 2088-2093.
• Koomey, Jonathan, and Nate Hultman. 2007. "A reactor-
level analysis of busbar costs for U.S. nuclear plants,
1970-2005." Energy Policy (accepted, under revision).
January.
• Hultman, Nathan E., and Jonathan G. Koomey. 2007.
"Energy technology costs and public support for R&D:
Accounting for surprise by disentangling uncertainties."
Submitted to Environmental Research Letters. April 15.
Copyright Jonathan Koomey 2007 5
6. The Generations of Nuclear Energy
Source: DOE Generation IV Project
Copyright Jonathan Koomey 2007 6
7. Status of new reactor designs
Source: Energy Information Administration, “New Reactor Designs”.
http://www.eia.doe.gov/cneaf/nuclear/page/analysis/nucenviss2.html
Copyright Jonathan Koomey 2007 7
9. Focus on direct costs
• Capital costs
• Construction duration
• Interest
• Capacity factors
• Heat rates
• O&M costs
• Fuel costs
• Incremental capital additions
• Waste disposal
• Decommissioning
Copyright Jonathan Koomey 2007 9
10. Some costs not included
• Taxes (these are transfer payments)
• Waste disposal costs beyond the 0.1
¢/kWh current fee (uncertainties on
Yucca Mountain)
• Externalities (e.g. accident risks beyond
those covered by insurance), routine
radiation emissions (generally small)
• Subsidies
Copyright Jonathan Koomey 2007 10
11. Why levelized costs?
• ¢/kWh a common metric for comparing
supply and demand-side technologies
• Just analyzing capital costs ignores
other big effects (eg. ∆ capacity factors)
• Collapsing the time dimension can yield
useful insights
Copyright Jonathan Koomey 2007 11
12. Levelization method
• Calculate present value of some future
expenditure or stream of expenditures
using the real discount rate (e.g.
decommissioning 40 years hence)
• Annualize that present value using the
capital recovery factor over the reactor
lifetime (generally 40 years) to get the
levelized cost
Copyright Jonathan Koomey 2007 12
13. Weaknesses of levelized costs
• Method depends somewhat on forecasts
– less accurate for reactors completed most recently
• Method submerges discontinuous events
– long periods of a reactor being off-line
– rapid changes in O&M costs or capacity factors
• Levelized cost figures often are compared
inconsistently to other ¢/kWh numbers
Copyright Jonathan Koomey 2007 13
14. Methods
• Calculate costs at the reactor level
– Discount rate = 6% real
– Use reactor specific data when possible
• Capital costs
• Construction duration
• Capacity factors
• O&M (needs improvement)
• Decommissioning (split by PWR vs. BWR)
– Use averages when necessary
• Fuel costs
• Incremental capital additions (better if by reactor)
• Waste disposal
Copyright Jonathan Koomey 2007 14
15. Key data sources
• Capital costs: Komanoff (except 4 reactors)
• Construction duration: Komanoff
• Interest: Avg US electric utility ROI 1971-2001
• Capacity factors: Komanoff, IAEA
• Heat rates: NEI
• O&M (fixed and variable): NEI, EIA
• Fuel: EIA
• Incremental capital additions: EIA
• Waste disposal: MIT
• Decommissioning: OECD
Copyright Jonathan Koomey 2007 15
16. Detective work on four reactors
Koomey with Joe Roy of MMWEC in August 2006. Roy
supplied monthly construction cost data for Seabrook. Updated
capital costs also compiled for Comanche Peak 1&2 and Watts Bar.
Copyright Jonathan Koomey 2007 16
17. Results
• Key components of total levelized costs
– Installed costs
– Capacity factors
– Total O&M costs
– Construction duration
– Capacity
• Total levelized busbar costs
– Over time and rank ordered
• Comparison with current projections
Copyright Jonathan Koomey 2007 17
18. Large increases in installed capital
costs for plants completed >1982
Installed costs include 6%/yr real interest during construction
Copyright Jonathan Koomey 2007 18
19. Projected installed costs are at the
mid-to low end of historical costs
Copyright Jonathan Koomey 2007 19
20. Projected and historical capacity
factors match up well
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 20
22. Projected construction durations
generally lower than historical data
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 22
23. Projected capacity similar to history
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 23
24. Projected O&M costs generally
lower than historical experience
Copyright Jonathan Koomey 2007 24
27. Comparison of best case historical
costs with current estimates
Copyright Jonathan Koomey 2007 27
28. Implications for new reactors?
• Advocates for new plants say “So what? New
plants are totally different from the old ones”
– Need to prove this assertion by actually
building plants cheaply
• Some lessons have already been learned
– Streamline licensing
– Standardize designs
– Operate plants more effectively
• IT for construction management and operations
much more sophisticated now
• International experience accumulating
Copyright Jonathan Koomey 2007 28
29. Opportunities for a nuclear revival
• Interest driven by
– Tax credits/R&D
– Need for reductions in
• Greenhouse gas reductions
• Imported oil (need electric vehicles)
– International markets
• Much improved
– Technology
– Design
– Management
– Construction processes
– Licensing
Copyright Jonathan Koomey 2007 29
30. New nuclear infrastructure will be
more highly optimized
1978: Plastic models on roll-around carts 2000: 4-D computer aided design
and virtual walk-throughs
McGuire Nuclear Station Reactor Building Models.
2002 NRC processing time for 20-year
license renewal: ~18 months
1000 MW Reactor (Lianyungang Unit 1)
Source: Per Peterson, UC Berkeley Copyright Jonathan Koomey 2007 30
31. Uncertainties for a nuclear revival
• Plants still mostly site built
• Are there near-term construction bottlenecks?
• U.S. political system still decentralized
• Fuel prices up a lot recently
• Risk of accident anywhere in the world
• Link to proliferation (N. Korea, Iran)
• Need new repository?
• Competition more intense now
– More electricity generation options
– Deregulated markets
• Capital markets may be skeptical
Copyright Jonathan Koomey 2007 31
32. Spot prices for Uranium Up!
$113/lb in
early
April 2007
Copyright Jonathan Koomey 2007 32
33. Advice on evaluating projected
costs for new reactors
• Compare apples to apples ($/kW)
– Inflation adjustment
– 1st of a kind vs. Nth of a kind
– One vs. two reactors on a site
– Full reactor costs vs. power block
• Beware of
– Possible certification delays
– Lack of specificity of possible bottlenecks
– Lack of real world experience in construction
• Check modularization and parts counts
• DTE: First mover or fast follower?
Copyright Jonathan Koomey 2007 33
34. Questions to consider on
projected costs of new reactors
• Can DTE afford to wait?
– Risks of early adoption
– Risks of delay
• Can contract be structured so vendor
bears some risk for cost overruns?
• What if uranium prices stay high?
• Has DTE considered a scenario exercise?
Copyright Jonathan Koomey 2007 34
35. Comparing projected costs for other
new reactors (EPR, ABWR, ESBWR)
ABWR
ABWR
ESBWR
Copyright Jonathan Koomey 2007 35
36. Future work
• Improved reactor by reactor O&M costs,
forced outage rates, and incremental capital
additions (sources?)
• Statistical analysis and development of
limiting cases (e.g., pairing of example
reactors).
• Compare cost estimates for site-built and
mass produced technologies
– is more cost variability guaranteed for site-built
technologies?
– How can cost variability be minimized?
• International comparisons
Copyright Jonathan Koomey 2007 36
37. Conclusions
• Just examining capital costs gives an
incomplete picture of historical developments
• Reactor size interacted with regulatory delays
and slowing electricity demand growth,
affecting capital costs, financing costs, O&M
costs (and perhaps capacity factors)
• Projected capital costs, O&M costs, and
construction durations are low compared to
historical experience (but may still be correct)
Copyright Jonathan Koomey 2007 37
38. Conclusions (continued)
• Larger, more sophisticated operators
have been able to increase capacity
factors and reduce O&M costs
• Key uncertainty is whether the
regulatory and technical changes
implemented thus far are enough to
result in real reductions in installed
costs and construction duration
Copyright Jonathan Koomey 2007 38
40. Another look at total busbar costs
PWR=Pressurized water reactor,
BWR=Boiling water reactor. Cohort
indicates one of eight predictive cost
categories described by Rothwell :
Category 1 denotes Westinghouse PWR
< 700 MW; Category 2 = Westinghouse
PWR 700-1000 MW; Category 3 =
Westinghouse PWR>1000 MW;
Category 4 = Babcock & Wilcox PWR;
Category 5 = Combustion Engineering
PWR; Category 6 = General Electric
BWR < 700 MW; Category 7 = GE
BWR 700-1000 MW; Category 8 = GE
BWR > 1000 MW. Levelized costs
(which exclude subsidies and
externalities) are calculated using a real
discount rate of 6% as described in
online supplemental material.
Copyright Jonathan Koomey 2007 40
42. For plants started > 1968 (with one
exception), capacity was ≥ 800 MW
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 42
43. For plants finished > 1975,
capacity was ≥ 800 MW
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 43
45. Construction duration increased
for reactors started >1972
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 45
46. As reactor capacity became > 800 MW,
construction duration varied more
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 46
47. U.S. average nuclear capacity
factors improved substantially after
1985
Copyright Jonathan Koomey 2007 47
48. Lifetime capacity factors have also
improved, partly due to retirements
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 48
49. Reactor efficiencies don’t vary much
Early reactors for which we don’t have cost data
Main sample for which we do have cost data
Copyright Jonathan Koomey 2007 49
50. Limited data on incremental
additional capital expenditures
Copyright Jonathan Koomey 2007 50
51. Avg costs for O&M, fuel, and incremental
capital additions vary over time
Copyright Jonathan Koomey 2007 51
52. Improvements in capacity factors affect
total O&M costs after the mid 1980s
Copyright Jonathan Koomey 2007 52
53. Total O&M costs vary widely
Copyright Jonathan Koomey 2007 53
55. RECENT CONSTRUCTION COST
EXPERIENCE ($2002)
Genkai 3 $2,818/kW (overnight)
Genkai 4 $2,288/kW (overnight)
Onagawa $2,409/kW (overnight)
KK6 $2,020/kW (overnight)
KK7 $1,790/kW (overnight)
Yonggwang 5&6 $1,800/kW (overnight)
Browns Ferry RESTART $1,280/kW (overnight estimate)
Finland EPR (AREVA-Seimens contract only)
$2,350/kW (nominal estimate 2005)
Bruce RESTART $1,425/kW (nominal estimate 2005)
Source: Paul Joskow, MIT Copyright Jonathan Koomey 2007 55
56. New U.S. Reactor Licensing Process
Old Process: The two-step licensing process (10 CFR 50)
Construction Build Operating
License Plant License
New Process: Combined licensing process (10 CFR 52)
Early Verification
Site of
Combined Inspections,
Permit
Construction Tests,
and Build
Plant Analysis,
Operating and
Standard License
Design Acceptance
Specification Criteria
Copyright Jonathan Koomey 2007 56
Source: Berger and Parsons (MIT CEEPR 2005), via Joskow (MIT)
57. Energy Policy Act of 2005
• Loan guarantees for up to 80% of project cost
– Valid for all GHG-free technologies
– Higher leverage, lower debt cost reduces overall project cost
• Production tax credit of $18 per MWh for new nuclear capacity
through 2021, subject to 2 limitations:
– $125 million per 1,000-MW per year
– 6,000-MW eligible, allocated among available capacity
• Insurance protection against delays during construction and until
commercial operation caused by factors beyond private sector’s
control
– Coverage: $500 million apiece for first two plants, $250 million for
next four
– Covered delays: NRC licensing delays, litigation delays
Source: Joskow (MIT)
Copyright Jonathan Koomey 2007 57
58. Energy Policy Act of 2005
• Renewal of the Price-Anderson Act of 1957
– Liability protection extended until 2025
• Legislation updates tax treatment of nuclear decommissioning
trust funds to reflect competitive electricity markets
– All decommissioning trust funds will qualify for tax
deductibility (not only those of regulated utilities)
• Federal commitment on R&D portfolio ($2.95 billion authorized)
• Creates Assistant Secretary for Nuclear Energy at DOE
Source: Berger and Parsons (CEEPR, 2005), via Joskow (MIT)
Copyright Jonathan Koomey 2007 58