WP2

IEA Wind Task 26

The Past and Future Cost of Wind Energy
Leading Authors
Eric Lantz: National Renewable Energy Laboratory
Ryan Wiser: Lawrence Berkeley National Laboratory
Maureen Hand: National Renewable Energy Laboratory

IEA Wind Task 26:

The Past And Future Cost
Of Wind Energy
Work Package 2
Lead Authors:

Eric Lantz: National Renewable Energy Laboratory

Ryan Wiser: Lawrence Berkeley National Laboratory
Maureen Hand: National Renewable Energy Laboratory

Contributing Authors:

Athanasia Arapogianni: European Wind Energy Association

Alberto Cea: Spanish Wind Energy Association
Emilien Simonot: Spanish Wind Energy Association
Edward James-Smith: Ea Energy Analyses

The IEA Wind agreement, also known as the Implementing Agreement

for cooperation in the Research, Development, and Deployment of
Wind Energy Systems, functions within a framework created by the
International Energy Agency (IEA). Views, findings, and publications of
IEA Wind do not necessarily represent the views or policies of the IEA
Secretariat or of all its individual member countries.

NREL is a national laboratory of the U.S. Department of Energy, Office of Energy

Efficiency & Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.

National Renewable Energy Laboratory

15013 Denver West Parkway
Golden, Colorado 80401
303-275-3000 [Link]

Technical Report
NREL/TP-6A20-53510
May 2012
Contract No. DE-AC36-08GO28308

NOTICE
This report was prepared as an account of work sponsored by an agency of the United States government.
Neither the United States government nor any agency thereof, nor any of their employees, makes any warranty,
express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of
any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation,
or favoring by the United States government or any agency thereof. The views and opinions of authors
expressed herein do not necessarily state or reflect those of the United States government or any agency thereof.
Available electronically at [Link]
Available for a processing fee to U.S. Department of Energy
and its contractors, in paper, from:
U.S. Department of Energy
Office of Scientific and Technical Information
P.O. Box 62
Oak Ridge, TN 37831-0062
phone: 865.576.8401
fax: 865.576.5728
email: [Link]
Available for sale to the public, in paper, from:
U.S. Department of Commerce
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
phone: 800.553.6847
fax: 703.605.6900
email: orders@[Link]
online ordering: [Link]

Printed on paper containing at least 50% wastepaper, including 10% post consumer waste.

Acknowledgments
This report has been sponsored by the International Energy Agency (IEA) Wind Implementing
Agreement for Co-operation in the Research, Development, and Deployment of Wind Energy
Systems, and funded by the respective entities in the participating countries of Task 26, The Cost
of Wind Energy, including Denmark, Germany, Netherlands, Spain, Sweden, Switzerland, and
the United States. The authors of this report would like to thank each of the participating
countries and their representatives for supporting the work of this task. United States
participation in Task 26 is supported by the U.S. Department of Energy under Contract No. DEAC36-08GO28308 with the National Renewable Energy Laboratory.
Special thanks to Graham Sinden (Climate Strategies), Paul Schwabe (National Renewable
Energy Laboratory), Roberto Lacal Arantegui (European Commission), and Sander Lensink
(Energy Research Centre of the Netherlands) for detailed comments and input on earlier versions
of this report. Thanks also to Mark Bolinger (Lawrence Berkeley National Laboratory) for
reviewing portions of the report and for providing valuable insights and guidance at various
points in the development of this paper. Finally thanks to Scott Gossett and Linda Huff (National
Renewable Energy Laboratory) for editing assistance and Lynn Billman, David Kline, Robin
Newmark, and Brian Smith (National Renewable Energy Laboratory) for additional input and
direction in completing this work. Of course, any remaining errors or omissions are the sole
responsibility of the authors.

iii

Executive Summary
Over the past 30 years, wind power has become a mainstream source of electricity generation
around the world. However, the future of wind power will depend a great deal on the ability of
the industry to continue to achieve cost of energy reductions. This summary report, developed as
part of the International Energy Agency (IEA) Wind Implementing Agreement Task 26, The
Cost of Wind Energy, provides a review of historical costs, evaluates near-term market trends,
reviews the methods used to estimate long-term cost trajectories, and summarizes the range of
costs projected for onshore wind energy across an array of forward-looking studies and
scenarios. It also highlights high-level market variables that have influenced wind energy costs in
the past and are expected to do so into the future.
Historical and Near-Term Trends in the Levelized Cost of Wind Energy
Between 1980 and the early 2000s, significant reductions in capital cost and increases in
performance had the combined effect of dramatically reducing the levelized cost of energy
(LCOE) for onshore wind energy. Data from three different historical evaluations, including
internal analysis by the Lawrence Berkley National Laboratory (LBNL) and the National
Renewable Energy Laboratory (NREL) as well as published estimates from Lemming et al.
(2009) and the Danish Energy Agency (DEA) (1999), illustrate that the LCOE of wind power
declined by a factor of more than three, from more than $150/MWh to approximately $50/MWh
between 1980s and the early 2000s (Figure ES-1). However, beginning in about 2003 and
continuing through the latter half of the past decade, wind power capital costs increaseddriven
by rising commodity and raw materials prices, increased labor costs, improved manufacturer
profitability, and turbine upscalingthus pushing winds LCOE upward in spite of continued
performance improvements (Figure ES-1).

Levelized Cost of Energy

(2010 USD/MWh)

$300
LBNL/NREL Internal Analysis
DEA 1999
Lemming et al. 2009 (Coastal European Sites)

$250
$200
$150
$100
$50
$0
1980

1985

1990

1995

2000

2005

2010

Figure ES-1. Estimated LCOE for wind energy between 1980 and 2009 for the United States and
Europe (excluding incentives)

Sources: LBNL/NREL (internal analysis), Lemming et al. 2009, and DEA 1999

iv

More recently, turbine prices and therefore project capital costs have declined, but still have not
returned to the historical lows observed earlier in the 2000s. At the same time, however,
performance improvements have continued. As a result, modeling based on capital cost and
performance data from the United States and Denmark for projects expected to be built in 2012
2013 suggests that the LCOE of onshore wind energy is now at an all-time low within fixed wind
resource classes, and particularly in low and medium wind speed areas (Figure ES-2). Moreover,
the fact that capital costs remain higher than in the early 2000s but that those increased costs are
rewarded by improved performance and a lower LCOE demonstrates the fundamental
interdependence of capital cost and performance in wind turbine and project design.
$170

$150
Estimated LCOE:
2009-10

$130

Estimated
LCOE:
2002-03

$110
$90

Estimated
LCOE:
2012-13

$70
$50
5.5

Class 2

6.5

Class 3

7.5

Class 4

Class 5

Levelized Cost of Energy

(2010 USD/MWh)

Levelized Cost of Energy

(2010 USD/MWh)

$170

$150
Estimated LCOE:
2009

$130

Estimated
LCOE:
2002

$110
$90

Estimated
LCOE:
2012

$70
$50

8.5

5.5

Class 2

Class 6

50 Meter Wind Speed (m/s)

6.5

Class 3

7.5

Class 4

8.5

Class 5

Class 6

50 Meter Wind Speed (m/s)

air density = 1.225 kg/m 3

air density = 1.225 kg/m 3

Figure ES-2. LCOE for wind energy over time in the United States (left) and Denmark (right)

Sources: Wiser et al. 2012, James-Smith 2011b

Long-term Trends in Wind Energy LCOE
Further into the future, the LCOE of wind energy is expected to continue to fall, at least on a
global basis and within fixed wind resource classes. Performance improvements associated with
continued turbine upscaling and design advancements are anticipated, and lower capital costs
may also be achievable. The magnitude of future cost reductions, however, is highly uncertain.
Estimates of the future cost of onshore wind energy conducted to date have often been the result
of an iterative process that incorporates some combination of historical trends, learning curve
analysis, expert elicitation, and engineering modeling. Figure ES-3 presents the estimated cost
reductions anticipated by 13 recent analyses covering 18 cost scenarios. Many of these studies
utilize learning curves in combination with expert elicitation, engineering models, and near-term
market analysis (e.g., EWEA 2009, U.S. DOE 2008, GWEC/GPI 2010, and Lemming et al.
2009).

Trends In Wind Power LCOE

2011 = 100%

100%
90%
80%
70%
60%
50%

Note:
-Shaded area represents the full
range of expectations in the
literature
-Each individual line details the
expected cost of energy pathway
for a given study

2010

20th to 80th

2015

2020

2025

2030

Figure ES-3. Estimated range of wind LCOE projections across 18 scenarios

Sources: EREC/GPI 2010, Tidball et al. 2010 (includes modeling scenarios from multiple other
sources), U.S. DOE 2008, EIA 2011, Lemming et al. 2009, EWEA 2011, EPRI 2010, Peter and
Lehmann 2008, GWEC/GPI 2010, IEA 2009, and European Commission 2007
The data presented in Figure ES-3 suggest an approximate 0%40% reduction in LCOE through
2030. The single scenario anticipating no further cost reductions assumes that the upward price
pressures observed between 2004 and 2009 are moderated but remain significant enough to
prevent future reductions in LCOE. The three studies anticipating a 35%40% reduction in
LCOE by 2030 represent ambitious scenarios requiring concentrated efforts to reduce the cost
wind energy, relatively high rates of global deployment, and levels of investment that exceed
business as usual. By focusing on the results that fall between the 20th and 80th percentiles of
scenarios, the range is narrowed to roughly a 20%30% reduction in LCOE. Cost of energy
reductions are generally expected to be greater in the early years and then slow over time. Initial
cost reductions range from 1%6% per year. By 2030, all but one scenario envisions cost
reductions falling below 1% per year.
A large number of technological and market-based drivers are expected to determine whether
these projections are ultimately realized. Possible technical drivers are summarized in Table ES1 and include reduced component loads and increased reliability. At the same time, a resurgence
in turbine demand, resulting in supply chain pressures similar to those observed between 2004
and 2009 could counter cost reductions resulting from continued technical advancements.
Continued movement toward lower wind speed sites may also increase fleet-wide LCOE, despite
technological improvements that would otherwise yield a lower LCOE. On the other hand,
increasing competition among manufacturers and developers could drive down the LCOE of
onshore wind energy to a greater extent than otherwise envisioned.

vi

Table ES-1. Potential Sources of Future Wind Energy Cost Reductions

R&D/Learning Area

Potential Changes

Expected Impact

(For more detail on technology changes and

expected impacts, see references below)

Advanced drivetrain designs, reduced

loads via improved controls, and condition
monitoring (Bywaters et al. 2005)

Enhanced drivetrain reliability and

reduced drivetrain costs

Manufacturing
Efficiency

Higher production volumes, increased

automation (Cohen et al. 2008), and
onsite production facilities

Enhanced economies of scale, reduced

logistics costs, and increased component
consistency (allowing tighter design
standards and reduced weights)

O&M Strategy

Enhanced condition monitoring

technology and design-specific
improvements and improved operations
strategies (Wiggelinkhuizen et al. 2008)

Real-time condition monitoring of turbine

operating characteristics, increased
availability, and more efficient O&M
planning

Power
Electronics/Power
Conversion

Enhanced frequency and voltage control,

fault ride-through capacity, and broader
operative ranges (UpWind 2011)

Improved wind farm power quality and

grid service capacity, reduced power
electronics costs, and improved turbine
reliability

Resource Assessment

Turbine-mounted real-time assessment

technology (e.g., LIDAR) linked to
advanced controls systems, enhanced
array impacts modeling, and turbine siting
capacity (UpWind 2011)

Increased energy capture while reducing

fatigue loads, allowing for slimmer design
margins and reduced component masses;
increased plant performance

Rotor Concepts

Larger rotors with reduced turbine loads

allowed by advanced controls (Malcolm
and Hansen 2002) and application of lightweight advanced materials

Increased energy capture with higher

reliability and less rotor mass; reduced
costs in other turbine support structures

Tower Concepts

Taller towers facilitated by use of new

design architectures and advanced
materials (Cohen et al. 2008, LaNier 2005,
Malcolm 2004)

Reduced costs to access stronger, less

turbulent winds at higher above-ground
levels

Drivetrain Technology

vii

Conclusions and Future Work

Following a long period of historical declines, wind energy costs were increasing for much of the
past decade. However, today, the cost of onshore wind energy once again appears to be falling
and is expected to reach a historic low in the near future within fixed wind resource areas.
Continued cost reductions are expected through 2030, but the anticipated magnitude of those
reductions varies widely and will ultimately be determined by an array of technical and nontechnical variables.
Recent capital cost and performance trends have underscored the need for a view of the cost of
wind energy that equally weighs both trends in capital cost and performance, particularly when
trying to understand the future cost of wind energy. The technology is now at a point where an
optimal cost of onshore wind energy may result from little or no further capital cost reductions
(and perhaps even modest capital cost increases), but continued performance improvements. In
this environment, it is possible to see capital costs remain relatively flatwith possible modest
reductions or increases, depending on local market conditionsinto the future and to see
performance increases as the primary target of original equipment manufacturers (OEMs).
Further improving our understanding of possible future cost trends will require additional data
gathering and improved modeling capability. Robust data collection is needed across the array of
variables that must be factored into estimating LCOE (e.g., capital cost, capacity factor, O&M
costs, component replacement rates and costs, and financing costs) and in each of the wind
energy markets around the globe. Also needed are data on the many contextual factors that
impact the overall cost of wind energy and that may also vary with time, such as interconnection
costs, permitting costs, and the average wind speed of installed wind projects. Such data would
allow historical LCOE trends to be more closely analyzed, with insights gleaned both through
more sophisticated learning curve analysis as well as bottom-up assessments of historical cost
drivers. Additional data could also assist in better distinguishing those cost reductions that result
from technological improvements from those changes in cost that result from external supply and
demand market variables or changes in raw material and commodity prices. It is only with this
improved historical understanding that future possible cost trajectories can be fully understood
(Dinica 2011). An enhanced capacity to model the cost and performance impacts of new
technological innovation opportunities, taking into account the full system dynamics that result
from a given technological advancement, is also essential. Component, turbine, and project-level
design and cost tools of this nature would allow for more sophisticated cost modeling and
provide greater insights into possible future costs based on changes in material use and design
architectures. Together these efforts would enhance our ability to understand future costs,
facilitate prioritization of research and development (R&D) efforts, and help to understand the
role and required magnitude of deployment incentives into the future.

viii

Table of Contents
1
2

Introduction ........................................................................................................................................... 1
Trends in Wind Energy Capital Costs and Performance .................................................................. 3

Impact of Capital Cost and Performance Trends on Levelized Cost of Energy .......................... 14

2.1 Capital Cost Reductions: 19802003...................................................................................3

2.2 Capital Cost Increases: 20042009 ......................................................................................5
2.3 Performance Increases: 19802010 .....................................................................................8
2.4 Recent and Near-term Trends in Capital Cost and Performance .......................................12
3.1 Historical LCOE Trends: 19802009 ................................................................................14
3.2 Recent and Near-term Trends in LCOE: 20102013 ........................................................15
Analytical Methods to Project the Future Cost of Wind Energy .................................................... 18

4.1 Learning Curve ..................................................................................................................18

4.2 Expert Elicitation ...............................................................................................................20
4.3 Engineering Model.............................................................................................................21
4.4 Sources of Cost Reduction Identified by Expert Elicitation and Engineering Modeling
Assessments .......................................................................................................................22
4.5 Quantitative Summary of Future Cost Estimates...............................................................25

5 Conclusions ........................................................................................................................................ 28
6 References .......................................................................................................................................... 30
Appendix A ................................................................................................................................................ 36

Data Inputs for U.S. and Denmark Levelized Cost of Energy Modeling ................................36

Appendix B ................................................................................................................................................ 38

ix

List of Figures
Figure ES-1. Estimated LCOE for wind energy between 1980 and 2009 for the United States and
Europe (excluding incentives) ........................................................................................... iv
Figure ES-2. LCOE for wind energy over time in the United States (left) and Denmark (right)... v
Figure ES-3. Estimated range of wind LCOE projections across 18 scenarios ............................. vi
Figure 1. Capital cost trends in the United States and Denmark between 1980 and 2003 ............. 4
Figure 2. Capital cost trends in the United States, Denmark, Spain, and Europe from 2003 to
2009..................................................................................................................................... 5
Figure 3. Wind turbine prices in the United States ......................................................................... 6
Figure 4. Representative turbine architectures from 1980 to 2010 ................................................. 9
Figure 5. Fleet-wide capacity factor data for the United States, Denmark, and Spain from 1999 to
2010................................................................................................................................... 10
Figure 6. U.S. project capacity factors by vintage and wind power class .................................... 11
Figure 7. Modeled capacity factors for current turbine models relative to historical technology 13
Figure 8. Estimated LCOE for wind energy from 1980 to 2009 for the United States and Europe
(excluding incentives) ....................................................................................................... 15
Figure 9. LCOE for wind energy over time in the United States (left) and Denmark (right)....... 16
Figure 10. Estimated change in the LCOE between low and high wind speed sites resulting from
technological advancement ............................................................................................... 17
Figure 11. Estimated range of wind LCOE projections across 18 scenarios ................................ 26

List of Tables
Table ES-1. Potential Sources of Future Wind Energy Cost Reductions ..................................... vii
Table 1. Potential Sources of Future Wind Energy Cost Reductions ........................................... 23
Table 1A. Inputs in Modeling of U.S. LCOE Estimates 20022003 through 20122013 ........... 36
Table 2A. Inputs in Modeling of Danish LCOE Estimates 2002 through 2012 ........................... 37

1 Introduction
Wind energy has been utilized by human society for millennia. The first experiments with windgenerated electricity, however, date to the late 19th century, and it was not until the 1970s that
wind energy began to penetrate commercial power marketsfirst in Denmark, then in California
in the 1980s, and in Germany and Spain in the 1990s (Wiser et al. 2011b). Since the mid-1990s,
wind energy has evolved into a mainstream source of power generation, with more than 200
gigawatts (GW) installed around the world (GWEC 2011, BNEF 2011b). A substantial fraction
of the newly installed total global electric generation capacity each year now comes from wind.
Wind energy is estimated to deliver the equivalent of just over 2% of global electricity
consumption (Wiser and Bolinger 2011), though that contribution varies considerably by region
(GWEC 2011).
The emergence of wind energy as a major source of power around the world correlates with
policy support for renewable energy technologies. Policy support has been driven by concerns
around energy security, economic development, and environmental protection. Without policy
support, much of the deployment occurring today would not be economically feasible (Wiser et
al. 2011). 1 Deployment of wind power is also correlated with significant reductions in the cost of
wind-generated electricity. Between 1980 and the early 2000s, wind power installation costs fell
by more than 65% in the United States (Wiser and Bolinger 2011) and 55% in Denmark (Nielsen
et al. 2010). As a result of these dramatic cost reductions, wind energy, in some parts of the
world, has achieved costs that are competitive with prevailing market prices without policy
support (Berry 2009, IEA 2009, and IEA and OECD 2010).
The future of wind energy will depend a great deal on the ability of the industry to continue to
achieve cost reductions and, ultimately, to achieve cost parity with conventional sources of
generation (i.e., to compete without direct policy support) across a broad array of contexts and
locations. The importance of future cost competitiveness has been reinforced by the difficulty in
developing international consensus around climate change policy (Pielke 2010). Moreover, with
increasing pressure on governments to reduce spending and debt, long-term policy support for
wind energy remains uncertain. Estimates by the Global Wind Energy Council (GWEC) and
Green Peace International (GPI) (GWEC/GPI 2010) suggest that supplying 20% of global
electricity demand with wind energy is possible; however, achieving such penetration would be
greatly facilitated by significant future cost reductions.
At the highest level, this summary report provides a review of historical cost, evaluates near-term
market trends, reviews the methods used to estimate long-term costs trajectories, and summarizes
the range of costs projected for onshore wind energy across an array of forward-looking studies
and scenarios. It also highlights high-level market variables that have influenced wind energy
1

Arguably, the need for policy support of wind energy to drive deployment in much of the world is, in part, due to
the tendency of electricity markets to inherently favor incumbent generation technologies. For example, grid
systems were not historically designed to integrate variable generation resources, and much of the existing grid
system was not designed to access geographically constrained renewable energy resources. Moreover, incumbent
technologies have, in many cases, continued to benefit from direct and indirect policy support in the form of fuel
production incentives, rail and pipeline development incentives, and socialization of risk (e.g., nuclear power). For
these reasonsand because wind energy is believed to provide public benefits (e.g., environmental gains)
advocates argue that policy support for wind energy is often justified.

costs in the past and are expected to do so into the future. More specifically, Section 2 begins by
summarizing historical capital costs and performance data in the United States and in various
European countries (for which data were available). Section 3 considers these trends from the
perspective of the cost of delivered energy and estimates the near-term cost of energy for
projects in late-stage development in the United States and Denmark today. Section 4 provides
an in-depth look at the strengths and weaknesses of the various methods used to estimate the
future cost of wind energy over the long term, summarizes a sample of projections made by a
variety of institutions under both conservative and optimistic assumptions, and discusses
potential sources of future cost reductions.
This report has been developed as part of the International Energy Agency (IEA) Wind
Implementing Agreement Task 26, The Cost of Wind Energy, and builds on the prior work of this
task to estimate the 2008 cost of wind energy among the participating countries (Schwabe et al.
2011). As in the prior report, analysis estimating the levelized cost of energy (LCOE) for wind
included here utilized the discounted cash-flow model developed by the Energy Research Centre
of the Netherlands (Schwabe et al. 2011); however, data presented here also rely on LCOE
estimates generated from other sources, as noted in the text. Additional input to this report comes
from published data as well as data provided to the Task 26 Working Group by participating
members. Specifically, Task 26 participant presentations, which provided key data and analysis
insights to the working group, are included in Appendix B and formed much of the basis for this
summary. 2
To be clear, this report does not make new, long-term cost-of-energy forecasts. Instead, it is
intended to inform the reader of future possibilities. The focus of the report is on onshore wind
energy, as the vast majority of historical wind power investments have occurred on land.
Offshore wind has yet to penetrate many of the worlds commercial power markets and has not
yet achieved the status of a mainstream source of power generation in much of the world.
Nonetheless, useful analyses of past and possible future costs for offshore wind are available in
the literature (e.g., Carbon Trust 2008, Ernst & Young 2009, UKERC 2010, Wiser et al. 2011,
Levitt et al. 2011, ARUP 2011, and Doyle et al. 2011). 3

Individual presentations are referenced by author in the body of this report and included in the reference list.
As a less mature technology and industry, offshore wind energy is at a different point on the technology
development and deployment cost curve. Looking forward, offshore wind costs are generally expected to follow a
steeper downward trajectory than costs for onshore wind energy (Neij 2008, Wiser et al. 2011, Doyle et al. 2011,
and ARUP 2011).
3

2 Trends in Wind Energy Capital Costs and Performance

The cost of wind energy is a function of the cost to build and operate a wind energy project and
the amount of energy produced by the facility over its lifetime. For the purposes of this report,
the cost of wind energy is represented by the levelized cost of energy (LCOE). The LCOE is a
simple way of comparing the unit costs of different generation technologies and is widely used in
policy decision-making. 4 Additional detail on LCOE and methods for estimating it can be found
in Schwabe et al. (2011). Precise estimates of the LCOE require detailed data on capital costs,
financing costs, operation and maintenance (O&M) expenditures, and plant production data.
However, detailed data across countries on each of these variables are often not in the public
domain. The most detailed publicly available data sets often capture only capital cost and, in
some cases, project performance data. Moreover, data on one or both of these two variables are
available, over an extended time period, for only a handful of countries.
As a result of the aforementioned data limitations, independent capital cost and performance data
have often been used historically as a proxy for the cost of wind energy. Reliance on these two
variables as an indicator of the historical cost of wind is justified, in part, because they are two of
the more significant contributors to the LCOE. However, because capital cost and performance
are interdependent (e.g., newer, better performing technology sometimes costs more), reliance on
only one of these variables may fail to fully explain historical (or forecast future) LCOE trends
(e.g., EWEA 2009, Ferioli et al. 2009, Berry 2009, and Dinica 2011). This report includes data
and analysis of both historical capital costs and project performance to inform our understanding
of long-term cost trends and the high-level factors that influence the cost of wind energy.
Moreover, the report does so with greater consideration for the interdependency between capital
costs and project performance than has been typical in the past.
Here we examine capital cost and performance trends from the 1980s through present day.
Capital cost data are broken into three different epochs, from 1980 to 2003, from 2004 to 2009,
and from 2009 to projects planned for commissioning in 2012 and 2013. These time periods
cover both significant cost reductions as well as the period of cost increases observed during the
last decade. Included is a discussion of the drivers of both capital cost reductions and increases.
Developments in turbine performance since the 1980s are also discussed; however, performance
data are focused primarily on the more recent period, from the turn of the century to todays
current turbine offerings.
2.1 Capital Cost Reductions: 19802003
From the 1980s to the early 2000s, average capital costs for wind energy projects declined
markedly. In the United States, capital costs achieved their lowest level from roughly 2001 to
2004, approximately 65% below costs from the early 1980s (Wiser and Bolinger 2011). In
Denmark, capital costs followed a similar trend, achieving their lowest level in 2003, more than
55% below the levels seen in the early 1980s (Nielsen et al. 2010) (Figure 1). Over the same time
period, global installed wind power capacity grew from a negligible quantity to nearly 40,000
4

It can be seen to represent a break-even cost per unit of energy produced for a generating facility and is typically
based on the cost of capacity, operation and maintenance costs, the expected level of production, and financing costs
as represented by a discount rate. LCOE typically does not capture or represent all societal costs or benefits resulting
from wind energy deployment (e.g., grid integration costs and environmental externalities).

megawatts (MW) (GWEC 2006), with the bulk of this growth (>85%) occurring between 1995
and the early 2000s. The primary markets for wind energy during this time were Europe and the
United States.

Installed Project Captial Cost

(2010 USD/kW)

$5,000
US Capacity-Weighted
Average
DK Capacity-Weighted
Average

$4,000
$3,000
$2,000
$1,000
$0
1980

1985

1990

1995

2000

2005

Figure 1. Capital cost trends in the United States and Denmark between 1980 and 2003

Sources: Wiser and Bolinger 2011, Nielsen et al. 2010

Over this time frame, technological innovations allowed for the development of larger turbines at
lower costs. Economies of scale resulting from increased turbine size were followed by
economies of scale in project size and manufacturing (EWEA 2009). More specifically,
innovations in design, materials, process, and logistics helped to drive down system and
component costs while facilitating turbine upscaling (EWEA 2009). Between 1980 and the early
2000s, commercial turbines grew from less than 100 kilowatts (kW) to more than 1 MW in
capacity, rotor diameters grew from roughly 15 meters to more than 70 meters in some cases,
and tower heights grew from 20 meters to more than 65 meters in some cases (Wiser et al. 2011).
Larger turbines provided access to better wind resources while lowering the plant-wide parts
count and generating turbine-level economies of scale for many components for which costs do
not vary proportionally with turbine size (e.g., controls). For those components where costs tend
to increase significantly with size, such as rotors, towers, and generators, engineering design
innovations mitigated the otherwise-expected cost increases. Turbine upscaling also facilitated
reductions in required project infrastructure, including roads, total project foundation costs, and
underground cabling, helping to drive down balance of plant (BOP) costs on a dollars per
kilowatt ($/kW) basis (EWEA 2009).
4

As costs declined and the technology matured, larger projects also began to emerge. These
projects offered further potential economies of scale, this time in development costs, substation
and interconnection infrastructure, transmission tie lines, and O&M facilities. Again, efficiencies
were gained because these types of costs do not vary proportionately with project size (EWEA
2009). Finally, as the market expanded, greater production volumes allowed for investment in
manufacturing facilities and created opportunities to increase production efficiencies, thus
offering an additional source for capital cost reduction.
2.2 Capital Cost Increases: 20042009
The initial period of capital cost reductions came to an end in the early-to-mid 2000s. Data from
the United States, Denmark, Spain, and Europe show capital cost increases beginning around
2004 and continuing through at least 20072009 (Figure 2). Capital costs in the United States
were maintained at peak levels through 2010, although preliminary data indicate that capital cost
reductions are likely in 2011, (see also Section 2.4) (Wiser and Bolinger 2011). In Denmark and
Europe capital costs peaked earlier, in 2008 and 2007, respectively, and have declined modestly
since then (see Figure 2 and Section 2.4).
An important exception to this general trend of substantially rising costs from 2004 to 2009 was
China. Specifically, the emergence of a handful of strong domestic original equipment
manufacturers (OEMs) has resulted in significantly lower capital costs in China (i.e.,
$1,100/kW$1,500/kW [2010 U.S. dollars] in 20082009) than witnessed in Europe or the
United States (Wiser et al. 2011b).

Installed Project Captial Cost

(2010 USD/kW)

$2,500
$2,000
$1,500
$1,000

EWEA Capacity-Weighted Average

Spain Capacity-Weighted Average
US Capacity-Weighted Average
DK Capacity-Weighted Average

$500

$0
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Figure 2. Capital cost trends in the United States, Denmark, Spain, and Europe from 2003 to 2009

Sources: Wiser and Bolinger 2011, Nielsen et al. 2010, Cea and Simonot 2011, and EWEA
2011
Although balance-of-plant costs have played a role (e.g., Fowler 2008), the increase in capital
costs observed between roughly 2004 and 2009 has been largely tied to increases in the price of
wind turbines (Wiser and Bolinger 2011, Cea and Simonot 2011). For example, Figure 3
depicts reported wind turbine transaction prices in the United States. Because visibility
5

surrounding wind turbine transactions has declined in recent years, the figure also presents a
range of reported pricing for U.S. transactions signed in 2010 and early 2011 and includes
average global turbine prices reported by Vestas for the years 2005 through 2010. Leaving
discussion of the sizable drop in turbine prices from 2009 to 2011 for Section 2.4, the figure
clearly depicts the sharp rise in turbine pricing experienced during the mid-2000s. 5

Figure 3. Wind turbine prices in the United States

Sources: Wiser and Bolinger 2011, Vestas (2011a, 2011, 2011c), and BNEF (2011a)
An array of factors has contributed to this increase in turbine prices. Raw material commodity
prices and energy prices have been among the exogenous factors influencing capital costs over
this time period. Both materials and energy prices increased substantially up to the time of the
financial crisis in late 2008. Bolinger and Wiser (2011) estimate that materials price increases,
including steel, iron, copper, aluminum, and fiberglass, resulted in an approximately $71/kW
price increase for wind turbines in the United States from 2003 through 2008. Increases in
energy prices were estimated to be responsible for an additional increase of about $12/kW
(Bolinger and Wiser 2011). A similar Danish study (Nielsen et al. 2010) found that steel prices
accounted for about $80/kW of the difference in Danish turbine prices over this same time
period. Excluding steel, Nielsen et al. (2010) found that other raw material prices accounted for
about $20/kW of the difference in Danish turbine prices from 2003 through 2008. Analysis of
changing commodity prices in the Spanish wind industry suggest that changes in copper and
steel pricing accounted for 46% and 34%, respectively, of the overall raw material cost increase
of wind turbine nacelles from 2006 through 2010, while increases in the price of steel have been
responsible for about 38% of the increase in the raw material cost of a typical wind turbine tower
over this same time period (Cea and Simonot 2011).
5

Other forms of electricity production equipment also experienced price increases over this same time frame
(Chupka and Basheda 2007).

Growth from approximately 40,000 MW to nearly 200,000 MW of installed wind power

capacity (GWEC 2011) in a period of about 6 years also resulted in significant supply
constraints. Coupled with increasing OEM profitability and increasing labor costs, these
constraints resulted in significant upward pressure on wind turbine prices. Analysis of the
financial reports from the worlds largest turbine manufacturer, Vestas, shows that manufacturer
profitability grew notably from 2003 through 2008 and accounted for an estimated $106/kW of
the increase in U.S. turbine prices during this time (Bolinger and Wiser 2011). Nielsen et al.
(2010), using a slightly different methodology also based on Vestas financial reports, estimate an
impact to turbine prices of about $125/kW over this time period. While these two particular
quantitative estimates are based solely on data from Vestas, there is solid evidence that the
profitability of other manufacturers also increased during this time (Bolinger and Wiser 2011).
According to Nielsen et al. (2010), these higher earnings followed substantial price reductions
associated with excess manufacturing capacity in the early 2000s, as well as financial losses that
were accrued during the first half of 2000s.
In addition, Bolinger and Wiser (2011) found that growth in labor costs accounted for roughly
$60/kW of the observed turbine price increase from 2003 through 2008 for Vestas, with similar
results for Suzlon (Bolinger and Wiser 2011). Nielsen et al. (2010) estimate that the difference in
labor costs from 2003 through 2008 was comparable, at about $50/kW. Though not quantified,
additional upward pressure on turbine prices presumably came from the increased profitability
and labor costs of component manufacturers. Supply chain bottlenecks also resulted in extended
development lead times and placed additional upward price pressure on (non-turbine) BOP costs.
Nielsen et al. (2010) observed that the rapid increase in demand for turbines from 2005 resulted
in supply side constraints for key components such as bearings and gear boxes, and that these
constraints may have been compounded by sub-contractors who optimized their production to
smaller wind turbine models that dominated the market up until 2005.
Continued turbine upscaling and increasingly sophisticated designs have also resulted in turbine
price increases due to increased material use, more complex control and sensing systems, and
enhanced power conversion capabilities (Wiser and Bolinger 2011). Using an NREL scaling
model described in Fingersh et al. (2006), turbine upscaling alonenot considering other design
advancements, which have also improved the performance of modern turbines and resulted in
improved grid interactionshas been estimated to be responsible for $114/kW of the turbine
price increases observed from 2003 to 2008 in the United States, representing the single largest
factor driving prices higher over this time period (Bolinger and Wiser 2011). Over the time
period from 2006 through 2010, Cea and Simonot (2011) attribute upwards of 50%70% of
individual nacelle, blade, and tower raw material cost increases to the impacts of increased
material use (weight) in the larger turbines of today. Focusing solely on steel costs, Nielsen et al.
(2010) estimated that the structural strengthening required to accommodate larger rotors and
higher towers alone resulted in additional cost for wind turbines of approximately $60/kW when
comparing turbines built in 2003 with those in 2008. As shown in Bolinger and Wiser (2011),
however, the increased capital costs due to upscaling have been rewarded by higher energy
yields. Excluding other drivers of turbine prices, upscaling was estimated to yield a lower
turbine-level LCOE, despite the increase in capital cost associated specifically with upscaling.
This highlights the interdependence of capital cost and performance.

Manufacturer warranty provisions also increased over this time period. Estimates by Bolinger
and Wiser (2011) indicate that Vestas increases in warranty provisions were responsible for
$31/kW of wind turbine price increases from 2003 through 2008, with strong evidence that other
manufacturer warranty provisions were also increasing over this time period. For wind power
markets outside of Europe, exchange rate fluctuations may have also contributed to increases in
observed turbine prices. In the United States, for example, the declining value of the dollar
relative to the Euro was estimated to have been responsible for $87/kW of the total observed
increase in turbine prices from 2003 through 2008 (Bolinger and Wiser 2011). As markets
outside of Europe have begun to build up their own domestic manufacturing capacity, however,
fluctuations in exchange rates will likely have a less significant impact on turbine prices and
subsequently capital costs moving forward.
Further discussion and analysis of the role of some of these factors in driving historical onshore
wind energy costs is included in Milborrow (2008), Blanco (2009), and Dinica (2011).
Moreover, other authors note the importance of many of these same factors in driving up the cost
of offshore wind energy (e.g., Carbon Trust 2008, Greenacre et al. 2010), as well as other forms
of electricity generation equipment (e.g., Chupka and Basheda 2007, Winters 2008) over a
similar time frame.
2.3 Performance Increases: 19802010
To maximize turbine performance, manufacturers have sought to develop more advanced turbine
components and larger turbines. More advanced components promise greater efficiency,
improved availability, and reduced generation losses (EWEA 2009, Wiser et al. 2012). Scaling to
taller towers allows wind turbines to capture less turbulent and often stronger wind resources.
Meanwhile, larger turbine rotor diameters allow a turbine to generate more electricity than would
otherwise be the case. In Figure 4, the growth of turbine nameplate capacity, hub height, and
rotor diameter, over the past 30 years is illustrated. 6

The development of larger turbines may not result in an increased ability to extract energy from a fixed amount of
wind energy, and as such, scaling does not improve turbine efficiency. However, by increasing the energy
production of a given turbine, which requires innovations in rotor design, tower design, the drivetrain, and transport
logistics, turbine scaling and the development of larger turbines is a technological change that has advanced the state
of the art.

Figure 4. Representative turbine architectures from 1980 to 2010

Source: NREL
A review of annual fleet-wide capacity factor data for the United States, Denmark, and Spain,
spanning 1999 through 2010, demonstrates the resulting performance improvements, to some
degree (Figure 5). Specifically, data for the United States and Denmark demonstrate overall
increases in average fleet-wide capacity factors on the order of 20% or more over this period.
However, such data are often confounded by inter-annual wind resource variability, dispatch
curtailment due in part to transmission congestion, 7 and long-term trends toward siting projects
in lower wind resource areas as the best resource sites are developed. The latter variable has been
especially significant in Spain, where fleet-wide capacity factor data show relatively flatand to
some extent, decliningcapacity factors as a result of new developments being pushed to lower
quality wind resource areas, simply because they are the only readily developable sites that
remain.

Although not shown in the data in Figure 5, this latter issue has emerged in the United States in certain regions
(Wiser and Bolinger 2011).

Capacity Factor

40%

DK Capacity Factor (Years for which data were provided)

Spanish Capacity Factor
US Capacity Factor

35%

30%

25%

20%
1998

2000

2004

2002

2006

2008

2010

Figure 5. Fleet-wide capacity factor data for the United States, Denmark, and Spain from 1999 to
2010

Source: Wiser and Bolinger 2011, James-Smith 2011a, and Mesa 2011
Note: U.S. data for 2007, 2008, 2009, and 2010 are estimated based on actual production plus
estimated curtailment (where curtailment data are available).
As a result of the limitations noted above, fleet-wide capacity factor data are incapable of
demonstrating the true level of performance improvement achieved over the past three decades.
However, by evaluating overall multi-year average capacity factor changes within specific wind
power classes and for specific project vintages, greater insights into the overall magnitude of
technical improvement can be gained. Such an exercise is particularly important in places like
Spain (and many other parts of the world) where projects have increasingly been installed in
lower wind power class sites, and the actual degree of capacity factor improvement over time
within individual wind resource classes is less apparent (Wiser 2011, Dinica 2011). 8 Figure 6
illustrates the substantial improvements over time in multi-year average capacity factors that
have been observed from projects installed in the United States when sorted by wind power class
and project vintage. These improvements have been directly linked to the development of taller
towers and larger rotors (Wiser 2010). In addition, by relying on modeled performance of
representative turbines installed between the early 2000s to today, Wiser et al. (2012) show
improvements in capacity factor on the order of 20% for projects in Class 5 and Class 6 wind
regimes and as much as 50% for projects in mid-Class 3 wind regimes (see Figure 7).

Aside from resource exhaustion as has been observed in Spain, drivers of project siting in lower wind resource
class areas could also include limited transmission availability, environmental and wildlife exclusions, and
potentially social opposition, among other siting limitations.

10

Capacity Factor

50%
40%
30%
20%

Projects installed: 2005 - 2007

10%

Projects installed: 1998 - 2001

Projects installed: 2002 - 2004

0%
2

4
Wind Power Class

Figure 6. U.S. project capacity factors by vintage and wind power class

Source: Wiser 2010

Using a different but related metric for wind power plant performance, annual energy production
per square meter of swept rotor area (kilowatt-hour [kWh]/m2) for a given wind resource site,
improvements of 2%3% per year over the last 15 years have been documented (IEA 2008,
EWEA 2009). 9

These data suggest that turbines are not only capturing more energy as a result of scaling but are also becoming
more efficient.

11

2.4 Recent and Near-term Trends in Capital Cost and Performance

Since the financial crisis of late 2008, global turbine production capacity has increased while
demand in some markets has softened. In combination with a shift in the directional impact of
many of the turbine price drivers discussed in Section 2.2 (e.g., commodity or raw materials
prices and manufacturer profitability), the result has been significant turbine price declines (see
Figure 3). Declining turbine prices (as well as BOP costs) began to reduce average capital costs
in some markets as early as 2008 (e.g., Denmark), but other markets took more time to work
through higher cost inventory that was contracted before prices began to fall (e.g., the United
States). 10 Data from Denmark show that capital costs fell about 2% from 2008 to 2009, and then
by more than 22% between 2009 and 2010 (Nielsen et al. 2010); EWEA estimates an 8%
reduction from 2007 through 2010 in the whole of Europe (EWEA 2011) (see Figure 2).
Analysts expect capital cost reductions of this magnitude to follow in the near term for the
United States (Wiser et al. 2012, Wiser and Bolinger 2011) 11 and elsewhere, perhaps with some
exceptions and variability in the overall magnitude of capital cost reduction. Capital cost
expectations farther into the future suggest continued declines (See Section 4).
Despite the recent decline in turbine prices, hub heights and rotor diameters have continued to
trend toward larger machines, suggesting that turbine performance improvements will also
continue. Preliminary analysis conducted by Wiser et al. (2012) suggests that capacity factors for
projects to be installed with current state-of-the-art technology in the United States (Current:
2012-13 in Figure 7) will improve significantly within a given wind power class, relative to
older technology (Figure 7). Moreover, also shown in the figure, the most significant
performance improvements are occurring in equipment designed for low wind speed sites
(typical average hub-height wind speeds of 7.5 m/s). 12 As a result of these technical and design
advancements, Wiser et al. (2012) find that the amount of U.S. land area that could achieve 35%
or higher wind project capacity factors has increased by as much as 270% when going from
turbines commonly installed in the 20022003 time frame to current low wind speed turbine
offerings.

10

In the United States, turbine lead times approached 2 years during the peak demand period in the first half of
2008. Market fundamentals have since changed, and lead times have dropped significantly. Nevertheless, there is a
natural lag between turbine contract and power purchase agreement signing and project commissioning such that
turbines ordered in early 2008 were still working their way through projects that were completed in 2010.
11
Initial 2011 data for the United States indicate that average project capital costs will likely fall relative to 2010,
while greater cost reductions are expected for projects to be completed in 20122013, based on current turbine
orders.
12
As previously noted, performance improvements may not actually result from increased turbine or rotor efficiency
but rather from new equipment that can access more valuable wind regimes as a result of higher hub heights and
capture more energy as a result of larger rotors.

12

50%
Current: 2012-13
Low Wind-speed (100m Tower)

45%

Current: 2012-13
Low Wind-speed
(80m Tower)

Current: 2012-13
Standard
Technology
2009-10
Standard
Technology

Capacity Factor

40%

2002-03
Standard
Technology

35%
30%
25%
20%
15%
5.5

6
Class 2

6.5

7
Class 3

7.5
Class 4

8
Class 5

8.5
Class 6

50m Wind-speed (m/s)

air density = 1.225 kg/m3
Figure 7. Modeled capacity factors for current turbine models relative to historical technology

Source: Wiser et al. 2012

Note: Low wind speed turbines are generally certified for IEC Class III wind conditions. IEC
Class definitions are distinct from and independent of the Wind Power Classes noted in the
figure. IEC standards define the reference average annual hub height wind speed for Class III
machines as 7.5 m/s at an air density of 1.225 kg/m3. Although site-specific factors (e.g.,
turbulence, gust conditions, air density) ultimately determine the viability of a specific turbine at
a particular site, Wiser et al. use the 7.5 m/s (at an air density of 1.225 kg/m3) reference as the
maximum average annual wind speed at hub height for these machines.

13

3 Impact of Capital Cost and Performance Trends on Levelized

Cost of Energy
As noted earlier, attempting to elicit LCOE trends from either capital cost or performance
independently is tenuous. Both variables are important components of the overall LCOE
equation. Historical and near-term LCOE trends are highlighted in this section to demonstrate the
value of focusing on LCOE and the shortcomings associated with having an exclusive focus on
either capital cost or performance. Data shown here generally reflect the impact on LCOE from
capital cost and performance changes. More precise estimates would also factor in the effects of
changes in O&M costs, major equipment replacement costs, financing costs, and actual turbine
lifetime. However, due to limited data availability, the latter variables were not considered in the
analysis presented in Section 3.2. Also note that the LCOE figures presented in this section
exclude available incentives that might affect the price of wind energy in wholesale markets. 13
3.1 Historical LCOE Trends: 19802009
Significant reductions in capital cost and increases in performance between 1980 and 2003 had
the combined effect of dramatically reducing the LCOE of wind energy. Data from three
different historical evaluations, including internal analysis by the Lawrence Berkley National
Laboratory (LBNL) and the National Renewable Energy Laboratory (NREL) as well as
published estimates from Lemming et al. (2009) and DEA (1999), illustrate that the LCOE of
wind power declined by a factor of more than three from more than $150/MWh to approximately
$50/MWh over this period (Figure 8). During this time, capital cost and performance trends were
both generally aligned with substantial capital cost declines and performance improvements. As
a result, there was little risk associated with an exclusive focus on one or the other when
attempting to understand broad trends in the LCOE of wind. 14
The increase in capital cost observed between 2004 and 2009, however, pushed the LCOE
upward despite the continued performance increases noted during this period (see Section 2).
The observed increase in LCOE was mitigated to some extent by performance increases,
however, and the overall LCOE trend is only modestly upward within the broader historical
timeline depicted in the figure. In fact, as discussed earlier, the single largest (but certainly not
the only) driver for the increased capital costs over this period was turbine upscaling (Bolinger
and Wiser 2011), which was rewarded with performance improvements. Clearly, from 2004 to
2009, an exclusive emphasis on either performance or capital costs to understand the trends in
the cost of wind energy would have been misleading. In the former case, one would have
assumed costs would have continued to fall due to performance increases, while in the latter
instance, one might have assumed that the LCOE increases would have been more dramatic than
they ultimately were.

13

In the United States, this is primarily the production tax credit, which is not accounted for in the data below. To
the extent that other industry support schemes (e.g., feed-in-tariffs and renewables obligations certificates) might
also affect the price of wind energy, they are not captured here.
14
Of course, an accurate estimate of actual LCOE trends would have required an assessment of both factors, as both
were improving over this period.

14

Levelized Cost of Energy

(2010 USD/MWh)

$300
LBNL/NREL Internal Analysis
DEA 1999
Lemming et al. 2009 (Coastal European Sites)

$250
$200
$150
$100
$50
$0
1980

1985

1990

1995

2000

2005

2010

Figure 8. Estimated LCOE for wind energy from 1980 to 2009 for the United States and Europe
(excluding incentives)

Sources: LBNL/NREL (internal analysis), Lemming et al. 2009, and DEA 1999
3.2 Recent and Near-term Trends in LCOE: 20102013
Turbine prices and therefore project capital costs have recently declined since their peak in the
late 2000s, but turbine pricing and capital costs have not returned to the historical lows observed
earlier in the 2000s. Despite this, continual improvements in turbine technology are expected to
result in the industry achieving an apparent historic low in the LCOE of wind, particularly in low
and medium wind speed sites (i.e., 6.0 m/s to 8.5 m/s average annual hub-height wind speed)
(Wiser et al. 2012, James-Smith 2011b).
Applying capital cost and performance data from the United States and Denmark, along with
standard industry assumptions derived from historical data and industry inquiries for O&M and
replacement costs, financing costs, and project availability, Figure 9 illustrates the results of two
analyses completed by the IEA Task 26 working group. 15 Both these LCOE estimates rely on the
same cash flow model used previously by the IEA Task 26 working group model and described
in detail in Schwabe et al. (2011). The results show that the LCOE of wind increased from 2002
to 2009 as a result of capital cost increases that outweighed the performance improvements
otherwise experienced over that time frame. In comparison, current turbine offerings (i.e., those
that might be installed in the 20122013 time frame) also have higher project-level installed
15

The turbine technologies considered in these two analyses include the GE 1.5/1.6 series machines in the United
States analysis and the NEG Micon NM52/90, Siemens 2.3-93, and Vestas 3.0 V112 turbines in the Danish analysis.
Capital costs range from $1,300/kW to $2,150/kW depending on the commercial operation date, local market
conditions, and whether low wind speed turbines are considered (low wind speed turbines are only considered in
20122013 for those wind speed classes where they are IEC certified and are somewhat higher in cost than their
standard technology analogs). Turbine production varies based on the technology being considered and, of course,
the average annual wind speed. A complete description of technology, cost, and performance modeling inputs as
well as the constants utilized for annual operations expenditures and financing costs is included in Appendix A.

15

costs than those installed in 20022003, but those costs are lower than in 20092010 and are
more than offset by the sizable expected performance improvements, yielding a lower LCOE
than in 20022003, despite the higher capital costs.

Figure 9. LCOE for wind energy over time in the United States (left) and Denmark (right)

Sources: Wiser et al. 2012, James-Smith 2011b

Note: Estimates assume low wind speed turbines are utilized at sites with 7.5m/s equivalent sea
level annual average hub height wind speed for 20122013 only. All other estimates are based
on modeled performance for standard industry technology during each time period. Estimates
also exclude available incentives and policies that might reduce the price of wind energy in
wholesale power markets. Finally, these estimates assume constant standardized industry values
for O&M, financing costs, and losses.
Looking ahead, because LCOE is the primary cost metric that impacts the bottom line of project
owners, utility purchasers, and to some extent even society at large (although, of course, not all
societal costs are captured in the LCOE), it is justified that analysis of future wind energy costs
be based on estimates of LCOE rather than capital cost alone. Moreover, unlike between 1980
and the early 2000s when both performance and capital cost trends suggested significant
reductions in the cost of wind energy, the technology is now at a point where an optimal cost of
onshore wind energy may result from little or no further capital cost reductions (and perhaps
even modest capital cost increases), but continued performance improvements. 16 Continued
incremental optimization on a cost or performance basis, as is likely into the future, is expected
to require analysis of future cost based on LCOE.
Aside from simple technological maturation, the potential for limited capital cost reduction is
further justified by the current trend of OEMs to offer a suite of turbine designs, including an
increasing number of offerings that are specifically designed for lower wind speed sites. Low
wind speed turbines typically have higher hub heights and larger rotors (in comparison to their
capacity rating), necessarily resulting in a higher capital cost product. But those capital cost
increases are rewarded by improved performance and, in many instances, a lower LCOE than
standard turbine offerings that feature smaller towers and rotors. In fact, as shown in Wiser et al.
(2012), by generating more significant performance increases in the emerging fleet of low wind
16

Simultaneous reductions of capital cost and performance improvements do remain a possibility, albeit
increasingly unlikely.

16

speed turbines, the difference in the LCOE of wind between low to medium wind speed sites and
high wind speed sites has declined over time (Figure 10). As a result of the improved economics
of wind and this narrowing in the LCOE between lower and higher wind speed sites, an
increasing amount of land area has become economically viable for wind energy development. In
the United States, for example, the available land area capable of generating wind energy at an
unsubsidized cost of $62 per megawatt-hour (MWh) (2010 U.S. dollars) is estimated to have
increased by 42% relative to the land area capable of producing power from wind at this cost in
the early 2000s (Wiser et. al. 2012).

Levelized Cost of Energy ($/MWh)

No Incentives

$120
$100
6 m/s

$80

$44/MWh
7 m/s

$25/MWh

8 m/s

$60
$40
$20
$0

Current, 2012-13

2002-03
Standard Technology

Technology Choice

Figure 10. Estimated change in the LCOE between low and high wind speed sites resulting from
technological advancement

Source: Wiser et al. 2012

Note: Todays turbine offerings include typical standard technology with an IEC Class I/II
rating as well as low wind speed technology designed for IEC Class III sites. Wind speeds noted
here are average annual sea level equivalent wind speeds at 50 m. Technology Choice
signifies the lowest cost turbine for a given wind regime. At average annual hub height wind
speeds below the sea level equivalent of 7.5 m/s, Technology Choice results in the use of low
wind speed technology.

17

4 Analytical Methods to Project the Future Cost of Wind Energy

A historical perspective provides a great deal of insight into the trajectory of the industry and
likely near-term trends (25 years), as well as an understanding of the driving forces behind past
and therefore potential future cost trends. However, the historical perspective alone fails to
reflect the specific technologies that remain in the research and development (R&D) pipeline and
may not adequately reflect the temporal nature of future cost reductions, which are likely to
depend on the rate of deployment and the amount of public and private sector R&D invested in
technology development. To complement an analysis of historical trends, three analytical
methods, the learning curve, expert elicitation, and the engineering model, have been developed
to provide additional insights into the future cost of wind energy over periods beyond the
immediate future.
This section examines the strengths and weaknesses of each of these three methods, highlights
examples of each as applied today, and summarizes the results of an array of studies utilizing one
or more of these methods to project the cost of wind energy through 2030. It also discusses
potential sources of future cost reductions, primarily focusing on changes in wind turbine
technology anticipated to reduce initial costs and operations expenditures and/or increase turbine
performance.
4.1 Learning Curve
The learning curve methodology has been the most common approach to forecasting future wind
energy costs. This approach assumes an explicit cost reduction, typically calculated from
historical trends, as a function of technology deployment or, in some cases, equipment
production. By using a learning curve, a specific learning rate, or percentage reduction associated
with every doubling in the capacity of installed wind energy projects, can be calculated. The
estimation of a historical learning rate is not in itself a projection, but rather a means of
evaluating historical trends (Lemming et al. 2009). Nonetheless, learning rates derived from
historical trends are commonly used to forecast future costs. A recent review of 18 estimates of
learning rates for onshore wind energy, completed for the Intergovernmental Panel on Climate
Change Special Report on Renewable Energy (Wiser et al. 2011), found that the overall range of
learning rates available in the published literature varied from 4%32%. However, by focusing
on a more limited sample of studies conducted since 2004 and using common variables (e.g.,
global installations and total investment costs) the range was narrowed to 9%19%.
The broad range of learning rates derived from past analysis suggests that a few caveats are
necessary when interpreting learning rates and using them to forecast future wind energy costs.
With respect to assessing previously calculated learning rates, the boundaries of a given study
are important. As the wind industry has only recently become truly global, many past studies
focused on deployment within a given country or regional market (e.g., Neij 1997, Mackay and
Probert 1998, Wene 2000, and Sderholm and Sundqvist 2007). Today, however, the wind
industry serves a multi-billion dollar global market and provides manufacturers and installers
with opportunities to learn from technology improvements and installation efficiencies from
around the world. Cost reductions used to calculate learning rates should therefore ideally be
associated with global installationsnot simply those in a specific local market (Ek and

18

Sderholm 2010). 17 Boundaries around the time period of historical data used are also critical.
For example, including data from 2002 to 2009 (which observed price increases due to a range of
factors discussed earlier) will produce a lower learning rate than if the rate were only calculated
from data that extends to the early 2000s; whether or not using 20022009 data is appropriate
depends on the nature of the drivers that impacted wind energy costs during that period.
The dependent variable that is studied will also impact the results of a learning curve study. Data
limitations have resulted in learning curve studies that frequently used project capital costs or
turbine prices as the dependent variable and as a proxy for wind energy costs, as opposed to the
LCOE. However, as discussed in Section 3, if capital costs are flat or increasing, but
performance is also increasing, reductions in LCOE could be occurring that are not reflected in
capital cost trends. Use of a dependent variable other than LCOE could potentially distort the
true learning rate for the technology (EWEA 2009, Ferioli et al. 2009, and Dinica 2011).
Finally, learning rates used to estimate future costs would ideally also be adjusted to account for
cost reductions resulting from drivers that are unrelated to cumulative installed capacity. For
example, cost reductions may result from a targeted R&D investment, economies of scale,
market supply and demand forces, or changes in commodity prices, all factors that are not
necessarily affected by learning associated with industry maturation and increased deployment.
Unless such variables are accounted for when estimating a learning rate (which results in a lower
learning rate) or can be assumed to provide equivalent cost reductions into the future, future
projections may overestimate the level of cost reduction associated with a specific increase in
total installed capacity (Wiser et al. 2011). Even when controlling for such variables, however,
declining returns are possible (Ferioli et al. 2009, Nemet 2009), and an assumption that past
learning rates can be applied to forecast future costs may be suspect.
4.1.1 Future Onshore Cost Estimates from Learning Curves
Despite their limitations, learning curve techniques can be informative with respect to the general
magnitude of possible future cost reductions. Learning curves are likely to remain in use due to
their relatively basic analytical underpinnings, reliance on real historical data, limited data
requirements, and the lack of obvious alternatives, particularly when looking more than a decade
into the future. As a result, the European Wind Energy Association (EWEA) and many others,
including governments, have relied extensively on learning curves to project the cost of wind
energy into the future. Often studies that apply learning curves and rates include external
assumptions that adjust for some of the limitations noted above by, for example, assuming a
reduction in the learning rate over time or only applying the learning curve for a short period of
time (e.g., EWEA 2009). Others focus the use of learning exclusively on capital costs and
assume little or no improvement in wind project performance (e.g., IEA 2009, Lemming et al.
2009), while still others apply the assumed learning rate to LCOE estimates (EWEA 2009).
Based on a range of learning curve estimates for LCOE of 9%17%, EWEA (2009) used a
learning rate of 10% to estimate that LCOE would drop by roughly 12 euro cents/kWh between
2010 and 2015. Using three different learning rates that are assumed to fall with deployment and
17

If a given cost reduction that is the result of learning by a global industry is attributed to only a fraction of global
installations, and represents a national or regional market, the degree of cost reduction resulting from a given set of
installations can be over or underestimated. As a result, errors can be introduced into both historical comprehension
and projections.

19

industry maturation from 10% to 6%, and ultimately to 3%, in order to represent the diminishing
returns associated with increased deployment, Lemming et al. (2009) estimate that wind energy
capital costs will fall by approximately 20% by 2020 and 30% by 2050 while capacity factors
remain at current levels. The Global Wind Energy Council (GWEC) and Green Peace
International (GPI) (GWEC/GPI 2010) utilize a learning rate and variable deployment rates to
estimate that capital costs will fall by 10% to 20% between 2010 and 2030. IEA (2009) applied a
7% learning curve to land-based wind energy to determine that capital costs will fall by nearly
20% by 2030 and 25% by 2050.
By and large, institutions using learning rates to project future costs have often tended toward
relatively conservative assumptions and largely arrived at capital cost reductions on the order of
20%30% by 2030 with somewhat smaller, incremental cost reductions accruing through to
2050. A review of learning rates in the literature conducted by the IEA Task 26 working group
suggests that the most aggressive use of learning rates in the literature results in a reduction in
the LCOE from wind power on the order of 30%40% by 2030 (see Section 4.5).
4.2 Expert Elicitation
To bolster the reliability of learning curve estimates and to garner a more detailed understanding
of how future cost reductions may actually be realized, analysis of the future cost of wind energy
also sometimes includes expert elicitation. This approach is based on surveying or interviewing
industry executives and technology design experts. Interviews are typically focused at the turbine
component and system level and may also attempt to capture trends in various aspects of
installation costs (e.g., underground cabling, erection costs, and required on-site monitoring
infrastructure). By evaluating the potential for cost reductions or performance improvements at
the component or system level and combining the estimated potential from an array of concrete
possible technological advancements, this approach constitutes a simple but technology-rich,
bottom-up analysis. In fact, expert elicitation is unique in that it allows for a relatively simple
bottom-up analysis and for a diverse set of variables (i.e., market pressures or system-level
turbine interactions) to be considered. However, it also introduces a relatively high level of
subjectivity into the analysis, as the responses to the elicitation may be affected by the design of
the data collection instrument and by the individuals selected to submit their views through that
instrument.
4.2.1 Examples of Expert Elicitation
Expert elicitation is sometimes combined with learning curve analysis and may be used to inform
public sector R&D investments. In the latter case, explicit quantitative projections may or may
not be included, but expert interviews are utilized to identify specific areas where focused R&D
is likely to have the greatest impact on the future cost of energy. The European wind industry,
with funding from the European Union, has used the expert elicitation approach in its European
Wind Energy Technology Platform (TP Wind) to identify specific research priorities. The U.S.
Department of Energys study 20% Wind Energy by 2030 (U.S. DOE 2008) utilized the expert
elicitation derived from insights in the Wind Partnership for Advanced Technology Components
(WindPACT) design studies (e.g., Cohen et al. 2008) to estimate an approximately 10%
reduction in wind energy capital costs and a nearly 20% increase in capacity factors (36%43%
in a Class 4 wind resource) between 2005 and 2030. Neij (2008), in work conducted for the
European Unions New Energy Externalities Developments for Sustainability (NEEDS) project,
combined expert elicitation with learning curve analysis to estimate that future turbine costs
20

could be approximated with a learning rate of about 10% and that reductions in the LCOE could
be approximated with a 15% learning rate when accounting for increased performance. Although
applied to offshore wind, Junginger et al. (2004) combine a sophisticated learning curve
approach (i.e., estimating learning curves for individual turbine components) and expert
elicitation to estimate future offshore wind energy cost reductions.
4.3 Engineering Model
Similar to expert elicitation, engineering modeling analysis provides a bottom-up alternative or
complement to the learning curve. Rather than relying on high-level data or expert opinion, this
approach utilizes detailed modeling of specific possible technology advancements that are
expected to result in cost reductions or performance increases. Because this approach typically
models both cost and performance, it inherently emphasizes expected reductions in LCOE. It
also requires a relatively robust understanding of possible technology advancements and, as a
result, the opportunities captured by engineering studies are often incremental and generally
realizable in the near term (510 years).
In addition to primarily being focused on the near to medium term, the main limitation of the
engineering model approach is that it requires highly sophisticated design and cost models to
capture the full array of component- and system-level interactions. Often the level of
sophistication achieved with todays modeling tools is insufficient to truly capture the systemlevel interactions that are common in wind turbine design. Cost models are also unable to make
projections about future commodity prices or supply and demand pressures throughout the
supply chain (of course, learning curves and expert elicitation face similar challenges). 18
Accordingly, the projected costs are generally based on the impact of a particular technical
innovation, all else being constant.
4.3.1 Engineering Model Examples
One of the prime examples of the engineering modeling approach comes from the U.S.
Department of Energys WindPACT project (e.g., Bywaters et al. 2005, Malcolm and Hansen
2002). Under that project, an array of system design studies was used to understand how various
innovation opportunities might affect turbine performance into the future. These results were
ultimately tied to cost functions to quantify their impact on turbine and project costs (Fingersh et
al. 2006). More recent NREL modeling work that builds upon these studies suggests that
performance increases on the order of 20% and cost reductions on the order of 10% over the next
one to two decades are possible but may require additional technological advancements not
captured by the WindPACT studies (e.g., Lantz and Hand 2011). Another example of the
engineering modeling approach being applied to future costs is in the European Commissions
UpWind project. In this effort, technical experts identified and analyzed an array of actions
required to achieve a functional 20-MW turbine, and cost modeling was used to estimate the
potential cost of this machine and the impact to costs from the various technological
enhancements (UpWind 2011).

18

Of course cost models are able to analyze the impacts of changes in commodity prices and labor costs, they
simply cannot forecast or anticipate such changes.

21

4.4

Sources of Cost Reduction Identified by Expert Elicitation and Engineering

Modeling Assessments
Incremental innovations in a variety of aspects of wind turbine design as well as in turbine
erection, O&M strategies, and manufacturing are expected to contribute to reduced wind energy
costs in the future. Table 1 summarizes the broad categories of opportunities envisioned to apply
to onshore wind energy projects, based on engineering studies and expert elicitation. Much of the
opportunity to drive down costs is perceived to be in the design and performance of wind
turbines because of their critical role in calculating wind energys LCOE; initial turbine
expenditures alone account for roughly 70%75% of project capital costs (Wiser and Bolinger
2011, Blanco 2009) and about 60% of lifetime project costs (Blanco 2009). However, O&M
strategies and manufacturing efficiencies are also anticipated to help reduce the cost of wind
energy in the future.

22

Table 1. Potential Sources of Future Wind Energy Cost Reductions

R&D/Learning Area

Potential Changes

Expected Impact

(For more detail on technology changes and

expected impacts, see references below)

Advanced drivetrain designs, reduced

loads via improved controls, and condition
monitoring (Bywaters et al. 2005)

Enhanced drivetrain reliability and

reduced drivetrain costs

Manufacturing
Efficiency

Higher production volumes, increased

automation (Cohen et al. 2008), and
onsite production facilities

Enhanced economies of scale, reduced

logistics costs, and increased component
consistency (allowing tighter design
standards and reduced weights)

O&M Strategy

Enhanced condition monitoring

technology, design-specific
improvements, and improved operations
strategies (Wiggelinkhuizen et al. 2008)

Real-time condition monitoring of turbine

operating characteristics, increased
availability, and more efficient O&M
maintenance planning

Power
Electronics/Power
Conversion

Enhanced frequency and voltage control,

fault ride-through capacity, and broader
operative ranges (UpWind 2011)

Improved wind farm power quality and

grid service capacity, reduced power
electronics costs, and improved turbine
reliability

Resource Assessment

Turbine-mounted real-time assessment

technology (e.g., LIDAR) linked to
advanced controls systems, enhanced
array impacts modeling, and turbine siting
capacity (UpWind 2011)

Increased energy capture while reducing

fatigue loads, allowing for slimmer design
margins and reduced component masses;
increased plant performance

Rotor Concepts

Larger rotors with reduced turbine loads

allowed by advanced controls (Malcolm
and Hansen 2002) and application of lightweight advanced materials

Increased energy capture with higher

reliability and less rotor mass; reduced
costs in other turbine support structures

Tower Concepts

Taller towers facilitated by use of new

design architectures and advanced
materials (Cohen et al. 2008, LaNier 2005,
and Malcolm 2004)

Reduced costs to access stronger, less

turbulent winds at higher above-ground
levels

Drivetrain Technology

23

As suggested by Table 1, fundamental elements of wind turbine innovations include advanced

control systems to assist in shedding loads, advanced condition monitoring to minimize major
component failures and unplanned turbine downtime, and forward-looking resource evaluation
(i.e., seeing the wind that is approaching the turbine) in order to better position turbines to
maximize production or minimize loads.
Cohen et al. (2008) estimate that taller towers could increase annual energy production by 11%.
In the absence of new design innovations, however, the increased costs associated with
extending towers beyond current industry norms may not be cost effective except in those sites
with higher-than-average wind shear (Wiser et al. 2012). Advanced design concepts considered
by Fingersh et al. (2006) suggest that higher tower heights might be achieved in the future with
only incremental cost increases. Advanced control systems and integrated system design are
considered critical for reducing tower loading and allowing for the development and application
of lower-cost tower and foundation designs (UpWind 2011, U.S. DOE 2008). Innovation
opportunities include hybrid, or concrete and steel towers, application of advanced materials, and
new structural designs. Cea and Simonot (2011) estimate that a 125-m hybrid steel and concrete
tower could result in a total tower materials cost reduction on the order of 40%50%, although
they expect this cost savings to be slightly less for an installed tower due to greater labor
requirements associated with hybrid and concrete towers. Future tower innovations may also be
targeted at easing crane height requirements, for example, with self-erecting or partially selferecting towers, and minimizing logistics constraints, the latter of which could support reduced
transportation costs.
Moving to larger rotor diameters may increase annual energy output from a given wind turbine
by as much as 10%30% (Cohen et al. 2008). To achieve larger rotor diameters without
significant cost increases or efficiency losses, however, new innovations that reduce blade
weight and loads are necessary (Griffin 2001). Some weight reductions may be achieved by
incremental design refinements and optimizations or the application of new materials with lower
mass-to-strength ratios. More significant weight reductions might be achieved with designs that
passively shed loads by twisting (Ashwill 2009) or that include partial blade span actuation (i.e.,
the ability to control different sections of a single turbine blade) coupled with sensing capacities
that allow the rotor to adapt to variability in wind conditions and turbulence in different parts of
the rotor disk (Buhl et al. 2005, Lackner and van Kuik 2009, and UpWind 2011). Trailing edge
flaps that react to the wind as it is moving towards the rotor rather than the wind that has already
passed by could also assist in load shedding and ultimately weight reductions (Andersen et al.
2006, Berg et al. 2009).
Optimization of current drivetrain designs to increase reliability and reduce weight is expected to
reduce drivetrain cost and increase performance (e.g., Peeters et al. 2006, Heege et al. 2007).
However, wholly new architectures are also possible and are in development and production,
including direct drive turbines, single-stage medium-speed geared designs, and multi-generator
architectures (Cohen et al. 2008). Permanent magnet generators have become increasingly
common in wind turbine design and may resolve some of the longstanding challenges of weight
and size typically associated with direct drive machines. UpWind (2011) specifically identified
permanent magnet transversal flux generators from among 10 specific drivetrain configurations
to be particularly promising in terms of drivetrain weight reduction. Drivetrain advancements

24

were estimated by Cohen et al. (2008) to result in increases in energy production up to 8% while
also potentially reducing costs by as much as 11%.
Advanced manufacturing strategies are expected to result in tighter design tolerances, driving
down weight and increasing product reliability as well as lowering overall turbine costs as a
result of process improvements and greater economies of scale (Cohen et al. 2008, UpWind
2011). New manufacturing strategies that allow for on-site production could also be employed to
lower logistics and transportation costs. Broader use of condition monitoring and increasingly
refined operational strategies (e.g., use of preventive maintenance strategies and enhanced
planning for turbine downtime) are expected to facilitate reductions in operations and
replacement costs. Finally, improvements in power electronics may reduce costs while
increasing the ability of wind turbines to provide grid services.
4.5 Quantitative Summary of Future Cost Estimates
Policy analysts, researchers, trade groups, and others have commonly utilized one or more of the
approaches described above to estimate the future cost of wind energy. Cost projections applying
these approaches vary, although when considered against at least the higher end of the range of
historical learning rates observed in the literature and suggested for every doubling of global
capacity (see Section 4.1 or Wiser et al. 2011), the range is somewhat less dramatic. 19
Figure 11 compiles and normalizes data from 13 relatively recent analyses (including both
research studies and policy analysis modeling inputs) and 18 scenarios to illustrate the expected
range of the future costs for onshore wind energy anticipated in the literature. The LCOE
reduction estimates shown in Figure 11 were developed by extracting data from individual
studies and scenarios. Raw cost data were converted to common values (i.e., 2009 dollars), and a
minimum performance starting point. From these normalized data, an estimated LCOE for each
scenario was calculated. LCOE values for this portion of the analysis were estimated with the
simplified LCOE methodology also employed in the IPCC Special Report on Renewable Energy
Sources (Moomaw et al. 2011). Where there were incomplete data for a given scenario (e.g.,
only capital costs available) representative industry data were utilized for other assumptions.
After calculating the annual percent reduction in LCOE implied by each scenario, scenarios were
aligned to a common starting point to estimate the percentage reduction in LCOE from 2010 to
2030. Studies analyzed were dated between 2007 and 2011. Based on this relatively narrow and
recent time frame, it was assumed that the drivers of specific cost reduction trajectories remain
generally fully available to be integrated into the average fleet turbine (i.e., if a study was
completed in 2007, the technological advancements envisioned would not be fully tapped by the

19

Learning rates are a function of cumulative installed capacity, and it is not unreasonable to expect that global
installed wind power capacity will double more than once over the next 20 years. Achieving 20% wind energy by
2030 in the United States has been estimated to require 305 gigawatts (GW), or an additional 260 GW from the endof-year 2010 figure (DOE 2008). China expects to have perhaps 200 GW installed by 2020, roughly a 150-GW
increase (GWEC 2011). Global capacity was 197 GW at year-end 2010 and thus Chinas activities alone could
nearly double global capacity by 2020. The Global Wind Energy Council deployment forecast based on data through
2010 suggests an average 18% growth annual rate over the next 5 years, which would result in a doubling of global
capacity by the end of 2014 (GWEC 2011). The higher end of the historical learning rates presented in the literature
for onshore wind energy, if applied to the anticipated number of doublings of installed wind power capacity by 2020
or 2030, would, in many cases, yield cost reductions that are greater than those presented in Figure 11.

25

2010 starting point in Figure 11). 20 Many of these studies utilize learning curves in combination
with expert elicitation, engineering models, and near-term market analysis (e.g., EWEA 2009,
U.S. DOE 2008, GWEC/GPI 2010, and Lemming et al. 2009). Some of the more extreme results
are generated from comparably conservative assumptions (e.g., Tidball et al. 2010) or from
advanced scenarios with the most optimistic assumptions (e.g., EREC/GPI 2010, GWEC/GPI
2010, and Peter and Lehmann 2008).
The normalized data suggest an absolute range of roughly a 0%40% reduction in LCOE
through 2030 (Figure 11). The single scenario anticipating no further cost reductions assumes
that the upward price pressures observed between 2004 and 2009 are moderated but remain
significant enough to prevent future reductions in LCOE. The three studies anticipating a 35%
40% reduction in LCOE by 2030 represent ambitious scenarios requiring concentrated efforts to
reduce the cost of wind energy and levels of investment that exceed business as usual. In
addition as virtually all the studies reviewed here incorporate learning curve concepts at some
level, differences in the respective scenarios also depend on the assumed levels of deployment.
Nevertheless, by focusing on the results that fall between the 20th and 80th percentiles of
scenarios, the range is narrowed to roughly a 20%30% reduction in LCOE.

Trends In Wind Power LCOE

2011 = 100%

100%
90%
80%
70%
60%

Note:
-Shaded area represents the full range of
expectations in the literature

20th to 80th

-Each individual line details the expected

cost of energy pathway for a given study

50%
2010

2015

2020

2025

2030

Figure 11. Estimated range of wind LCOE projections across 18 scenarios

Sources: EREC/GPI 2010, Tidball et al. 2010 (includes modeling scenarios from multiple other
sources), U.S. DOE 2008, EIA 2011, Lemming et al. 2009, EWEA 2011, EPRI 2010, Peter and
Lehmann 2008, GWEC/GPI 2010, IEA 2009, and European Commission 2007
LCOE reductions are generally expected to be greater in the early years and then slow over time.
Initial cost reductions range from 1%6% per year, fall to 1%4% by 2020, and decline further
20

In some cases, new turbine models or existing prototypes may already include technology improvements that have
been captured in various studies, suggesting that some portion of the cost reduction projection has already been
realized. However, when thinking of the industry not in terms of cutting edge, latest technology, but rather in terms
of the industry standard or average turbine among the existing fleet today, it is reasonable to assume that the LCOE
represented by the existing fleet average turbine has not incorporated the vast majority of the technology
improvement opportunities that are assumed to be realized in individual studies or modeling scenarios.

26

to 0%1.5% by 2025. By 2030, all but one scenario envisions cost reductions falling below 1%
per year.
A variety of factors are expected to influence whether or not these estimates are realized.
Resurgence in turbine demand resulting in supply chain pressures similar to those observed
between 2004 and 2009 could, again, drive wind energy LCOE higher. Renewed upward
pressure on commodity prices that might be associated with a strong economic recovery could
also drive prices higher.
At the same time, factors that may not be directly captured in the studies highlighted here,
including the impact of increasing competition among manufacturers in general, could drive
down costs further.

27

5 Conclusions
Over the past 30 years, the wind industry has become a mainstream source of electricity
generation around the world. The industry has observed significant cost of energy reductions.
However, from roughly 2004 to 2009, the cost of wind energy increased. Historically, cost
reductions have resulted from both capital cost reductions and increased performance. From
2004 to 2009, however, continued performance increases were not enough to offset the sizable
increase in capital costs that were driven by turbine upscaling, increases in materials prices,
energy prices, labor costs, manufacturer profitability, andin some marketsexchange rate
movements. Nevertheless, as capital costs have moderated from their 20092010 levels, the cost
of wind energy has fallen and is now at an all-time low within fixed wind resource classes.
Looking forward, the LCOE of wind energy is expected to continue to fall, at least on a longterm global basis and within fixed wind resource classes. Performance improvements associated
with continued turbine upscaling and design advancements are anticipated, and lower capital
costs may also be achievable. The magnitude of future cost reductions, however, remains highly
uncertain, although most recent estimates project that the LCOE of onshore wind could fall by
20%30% over the next two decades.
As the industry continues to mature and future technology advancement opportunities become
increasingly incremental, however, LCOE reductions can be anticipated to slow. Moreover,
continued movement towards lower wind speed sites may invariably increase industry-wide
LCOE, despite technological improvements that would otherwise yield a lower LCOE. Other
local factors such as transmission needs may also push towards higher costs. With these factors
in mind, it is of important to consider the interdependence of capital costs and performance, and
to evaluate the future cost of wind energy on an LCOE basis. Moreover, such evaluations must
consider trends in the quality of the wind resource in which projects are located, as well as
development, transmission, integration, and other cost elements that may also change (and
increase) with time and deployment levels, but are sometimes ignored in traditional LCOE
analyses (e.g., Dinica 2011).
Estimates of the future cost of wind energy conducted to date have often been the result of an
iterative process that incorporates some combination of historical trends, learning curve analysis,
expert elicitation, and engineering modeling. Theoretically, each of these approaches could
independently provide an estimate of the future cost of energy; however, it has often been
recognized that it is a combination of these different methodologies that is likely to yield the
most accurate results. The individual strengths and weaknesses of each approach are in some
ways complementary, so future projections are also expected to employ various combinations of
these methods.
Further improving our understanding of possible future cost trends will require additional data
gathering and improved modeling capability. Robust data collection is needed across the array of
variables that must be factored into estimating LCOE (e.g., capital cost, capacity factor, O&M
costs, component replacement rates and costs, and financing costs) and in each of the wind
energy markets around the globe. Also needed are data on the many contextual factors that
impact the overall cost of wind energy and that may also vary with time, such as interconnection
costs, permitting costs, and the average wind speed of installed wind projects. Such data would
28

allow historical LCOE trends to be more closely analyzed, with insights gleaned both through
more-sophisticated learning curve analysis as well as bottom-up assessments of historical cost
drivers. Additional data could also assist in better distinguishing those cost reductions that result
from technological improvements from those changes in cost that result from external supply and
demand market variables or changes in raw material and commodity prices. It is only with this
improved historical understanding that future possible cost trajectories can be fully understood
(Dinica 2011). An enhanced capacity to model the cost and performance impacts of new
technological innovation opportunities, taking into account the full system dynamics that result
from a given technological advancement, is also essential. Component, turbine, and project-level
design and cost tools of this nature would allow for more sophisticated cost modeling and
provide greater insights into possible future costs based on changes in material use and design
architectures. Together these efforts would enhance our ability to understand future costs,
facilitate prioritization of R&D efforts, and help to understand the role and required magnitude
of deployment incentives into the future.

29

6 References
Andersen, P.B.; Mac Gaunaa, C.B.; Buhl, T. (2006). Load Alleviation on Wind Turbine Blades
Using Variable Airfoil Geometry. In proceedings of the 2006 European Wind Energy
Conference and Exhibition, Athens, Greece, February 27March 2, 2006.
Ove Arup & Partners Ltd. (ARUP). (2011). Review of the Generation Costs and Deployment
Potential of Renewable Electricity Technologies in the UK. Department of Energy and Climate
REP001, Prepared by Ove Arup & Partners Ltd. London, UK.
Ashwill, T.D. (2009). Materials and Innovations for Large Blade Structures: Research
Opportunities in Wind Energy Technology. 50th AIAA/ASME/ASCE/AHS/ASC Structures,
Structural Dynamics, and Materials Conference, Palm Springs, CA, May 2009. Reston, VA:
American Institute of Aeronautics and Astronautics (AIAA).
Berg, D.E.; Wilson, D.G.; Barone, M.F.; Resor, B.R.; Berg, J.C.; Paquette, J.A.; Zayas, J.R.;
Kota, S.; Ervin, G.; Maric, D. (2009). The Impact of Active Aerodynamic Load Control on
Fatigue and Energy Capture at Low Wind Speed Sites. European Wind Energy Association.
Marseille, France, March 1619, 2009.
Berry, D. (2009). Innovation and the price of wind energy in the US. Energy Policy, 37: pp.
44934499.
Blanco, M.I. (2009). The Economics of Wind Energy. Renewable and Sustainable Energy
Reviews (13:67); pp. 13721382.
Bloomberg New Energy Finance (BNEF). (2011a). Wind-Research Note: Wind Turbine Price
Index. July 27, 2011.
BNEF. (2011b). Wind Market Outlook: Q3 2011. November 2, 2011.
Bolinger, M.; Wiser, R. (2011). Understanding Trends in Wind Turbine Prices Over the Past
Decade. LBNL-5119E. Berkeley, CA: Lawrence Berkeley National Laboratory.
Buhl, T.; Gaunaa, M.; Bak, C. (2005). Potential Load Reduction Using Airfoils with Variable
Trailing Edge Geometry. Journal of Solar Energy Engineering (127:4); pp. 503516.
Bywaters, G.; John, V.; Lynch, J.; Mattila, P.; Norton, G.; Stowell, J.; Salata, M.; Labath, O.;
Chertok, A.; Hablanian, D. (2005). Northern Power Systems WindPACT Drive Train Alternative
Design Study Report. NREL/SR-500-35524. Golden, CO: National Renewable Energy
Laboratory.
Carbon Trust. (2008). Offshore Wind Power: Big Challenge, Big Opportunity: Maximizing the
Environmental, Economic, and Security Benefits. London, UK: The Carbon Trust.
Cea, A; Simonot, E. (2011). The Cost of Wind Energy. Spanish Wind Energy Association
(AEE) contribution to IEA Task 26.

30

Chupka, M.W.; Basheda, G. (2007). Rising Utility Construction Costs: Sources and Impacts.
Prepared by The Brattle Group for The Edison Foundation.
Cohen, J.; Schweizer, T.; Laxson, A.; Butterfield, S.; Schreck, S.; Fingersh, L.; Veers, P.;
Ashwill, T. (2008). Technology Improvement Opportunities for Low Wind Speed Turbines and
Implications for Cost of Energy Reduction. NREL/TP-500-41036. Golden, CO: National
Renewable Energy Laboratory.
Danish Energy Agency (DEA). (1999). Wind Power in Denmark: Technologies, Policies, and
Results.
Dinica, V. (2011). Renewable electricity production costsA framework to assist policymakers decisions on price support. Energy Policy, 39: pp. 41534167.
U.S. Department of Energy (U.S. DOE). (2008). 20% Wind Energy by 2030: Increasing Wind
Energy's Contribution to U.S. Electricity Supply. DOE/GO-102008-2567. Washington, DC: U.S.
DOE.
Doyle, G.; Borkowski, G.; Vansiotis, G.; Dodds, J.; Critten, S. (2011). Costs of Low-Carbon
Generation Technologies. Bristol, UK: Mott McDonald and the Committee on Climate Change.
Ernst & Young. (2009). Cost of and Financial Support for Offshore Wind. The UK Department
of Energy and Climate Change.
U.S. Energy Information Administration (EIA). (2011). Assumptions to the Annual Energy
Outlook 2011. DOE/EIA-0554. Washington, DC: U.S. Energy Information Administration.
Ek, K.; Sderholm, P. (2010). Technology learning in the presence of public R&D: The case of
European wind power. Ecological Economics, 69: pp. 23562362.
Electric Power Research Institute (EPRI). 2010. NESSIE Modeling Base Case 2009. Personal
communication.
European Renewable Energy Council (EREC); Greenpeace International (GPI). (2010). Energy
[R]evolution: A Sustainable World Energy Outlook. Brussels, Belgium: Greenpeace
International, European Renewable Energy Council.
European Commission. (2007). Renewable Energy Roadmap Impact Assessment. Commission
staff working document.
European Wind Energy Association (EWEA). (2009). Economics of Wind Energy. Brussels,
Belgium: European Wind Energy Association.
EWEA. (2011). Pure Power Wind Energy Targets for 2020 and 2030. Brussels, Belgium:
European Wind Energy Association.

31

Ferioli, F.; Schoots, K; Van der Zwaan, B.C.C. (2009). Use and limitations of learning curves
for energy technology policy: A component-learning hypothesis. Energy Policy, 37: pp. 2525
2535.
Fingersh, L.; Hand, M.; Laxson, A. (2006). Wind Turbine Design Cost and Scaling Model.
NREL/TP-500-40566. Golden, CO: National Renewable Energy Laboratory.
Fowler, A. (2008). Trends in Wind Power Prices, B.O.P., and Turbine Costs. WINDPOWER
2008. Houston, Texas, June 2008.
Global Wind Energy Council (GWEC). (2006). Global Wind Report: Annual Market Update
2006. Brussels, Belgium: Global Wind Energy Council.
GWEC. (2011). Global Wind Report: Annual Market Update 2010 2nd Edition. Brussels,
Belgium: Global Wind Energy Council.
GWEC; Greenpeace International (GPI). (2010). Global Wind Energy Outlook 2010. Brussels,
Belgium: Global Wind Energy Council, Greenpeace International.
Greenacre, P.; Gross, R.; Heptonstall, P. (2010). Great expectations: The cost of offshore wind in
UK waters Understanding the past and projecting the future. London, UK: UK Energy
Research Centre.
Griffin, D.A. (2001). WindPACT Turbine Design Scaling Studies Technical Area 1Composite
Blades for 80- to 120-Meter Rotor; 21 March 2000 15 March 2001. NREL/SR-500-29492.
Work performed by Global Energy Concepts, Kirkland, WA. Golden, CO: National Renewable
Energy Laboratory.
Heege, A.; Betran, J.; Radovcic, Y. (2007). Fatigue Load Computation of Wind Turbine
Gearboxes by Coupled Finite Element, Multi-Body System and Aerodynamic Analysis. Wind
Energy (10:5); pp. 395413.
International Energy Agency (IEA). (2008). Energy Technology Perspectives 2008. Scenarios
and Strategies to 2050. Paris, France: International Energy Agency.
IEA. (2009). Technology Roadmap Wind Energy. Paris, France: International Energy Agency.
IEA; Organisation for Economic Co-operation and Development (OECD). (2010). Projected
Costs of Generating Electricity, 2010 Edition. Paris, France: International Energy Agency and
the Organisation for Economic Co-operation and Development. 218 pp.
James-Smith, E. (2011a). Danish Wind Project Performance and Capital Price Trends.
Presentation to IEA Wind Task 26.
James-Smith, E. (2011b). Development in LCOE for Wind Turbines in Denmark. Presentation to
IEA Wind Task 26.

32

Junginger, M.; Faaij, A. Turkenburg, W.C. (2004). Cost Reduction Prospects for Offshore Wind
Farms. Wind Engineering, 28: pp. 97118.
Lackner, M.; van Kuik, G. (2009). A Comparison of Smart Rotor Control Approaches Using
Trailing Edge Flaps and Individual Pitch Control. 47th AIAA Aerospace Sciences Meeting and
Exhibit, Orlando, Florida, January 58, 2009. Miami, FL: American Institute of Aeronautics and
Astronautics (AIAA).
LaNier, M.W. (2005). LWST Phase I Project Conceptual Design Study: Evaluation of Design
and Construction Approaches for Economical Hybrid Steel/Concrete Wind Turbine Towers;
June 28, 2002 July 31, 2004. NREL/SR-500-36777. Golden, CO: National Renewable Energy
Laboratory.
Lantz, E.; Hand, M. (2011). Engineering Model Based Technology Projections. Presentation to
IEA Task 26.
Lemming, J.K.; Morthorst, P.E.; Clausen, N.E.; Hjuler Jensen, P. (2009). Contribution to the
Chapter on Wind Power in Energy Technology Perspectives 2008. Ris-R-1674(EN). Roskilde,
Denmark: Ris National Laboratory for Sustainable Energy.
Levitt, A.; Kempton, W.; Smith, A.; Musial, W.; Firestone, J. (2011). Pricing Offshore Wind
Power. Energy Policy, 39: pp 64086421.
Mackay, R.M.; Probert, S.D. (1998). Likely market-penetrations of renewable-energy
Technologies. Applied Energy, 59: pp. 138.
Malcolm, D. J. (2004). WindPACT Rotor Design Study: Hybrid Tower Design; Period of
Performance: 29 June 2000 28 February 2004. NREL/SR-500-35546. Golden, CO: National
Renewable Energy Laboratory.
Malcolm, D.J.; Hansen, A.C. (2002). WindPACT Turbine Rotor Design Study. NREL/SR-50032495. Golden, CO: National Renewable Energy Laboratory.
Mesa, H.W. (2011). Onshore Wind in Spain. Spanish Wind Energy Association (AEE)
contribution to IEA Task 26.
Milborrow, D. (2008). Dissecting Wind Turbine Costs. WindStats Newsletter (21:4); pp. 35.
Moomaw, W.; Burgherr, P.; Heath, G.; Lenzen, M.; Nyboer, J.; Verbruggen, A. (2011) Annex II:
Methodology. In IPCC Special Report on Renewable Energy Sources and Climate Change
Mitigation [O. Edenhofer, R. Pichs-Madruga, Y. Sokona, K. Seyboth, P. Matschoss, S. Kadner,
T. Zwickel, P. Eickemeier, G. Hansen, S. Schlmer, C. von Stechow (eds)]. Cambridge, UK and
New York, NY, USA: Cambridge University Press.
Nemet, G. (2009). Interim monitoring of cost dynamics for publicly supported energy
technologies. Energy Policy, 37: 825835.

33

Neij, L. (1997). Use of experience curves to analyse the prospects for diffusion and adoption of
renewable energy technology. Energy Policy, 25: pp. 10991107.
Neij, L. (2008). Cost development of future technologies for power generation A study based
on experience curves and complementary bottom-up assessments. Energy Policy, 36: pp. 2200
2211.
Nielsen, P.; Lemming, J.; Morthorst, P.E.; Clausen, N.E.; Lawetz, H.; Lindboe, H.H.; JamesSmith, E.; Bang, N.C.; Strm, S.; Larsen, J. (2010). Vindmllers konomi (The Economy of
Wind Power). EUDP 33033-0196. Prepared by EMD International, Aalborg, Denmark.
Peeters, J.L.M.; Vandepitte, D.; Sas, P. (2006). Analysis of Internal Drive Train Dynamics in a
Wind Turbine. Wind Energy (9:12); pp. 141161.
Peter, S.; Lehmann, H. (2008). Renewable Energy Outlook 2030 Energy Watch Group Global
Renewable Energy Scenarios. Berlin, Germany: Energy Watch Group.
Pielke, R. (2010). The Climate Fix: What Scientists and Politicians Wont Tell You About Global
Warming. New York, NY: Basic Books.
Schwabe, P.; Lensink, S.; Hand, M.; Arapogianni, A.; Cea, A.; Cory, K.; Geissmann, M.;
Hammes, K.; Luxembourg, S.; Sanchez, A. M.; James-Smith, E.; Wallasch, J.; Wilczek, P.
(2011). Multi-national Case Study of the Financial Cost of Wind Energy. Paris, France:
International Energy Agency Wind Implementing Agreement (IEA Wind).
Sderholm, P.; Sundqvist, T. (2007). Empirical challenges in the use of learning curves for
assessing the economic prospects of renewable energy technologies. Renewable Energy, 32: pp.
25592578.
Tidball, R.; Bluestein, J.; Rodriguez, N.; Knoke, S. (2010). Cost and Performance Assumptions
for Modeling Electricity Generation Technologies. NREL/SR-6A20-48595. Work performed by
ICF International, Fairfax, VA. Golden, CO: National Renewable Energy Laboratory.
UK Energy Research Centre (UKERC). (2010). Great Expectations: The cost of offshore wind in
UK waters understanding the past and projecting the future. London, UK: UK Energy
Research Centre. ISBN 1 903144 0 9 4.
UpWind. (2011). Design Limits and Solutions for Very Large Turbines. European Commission,
Sixth Framework Programme.
Vestas. (2011a). Annual Report 2010.
Vestas. (2011b). Full year 2010 and guidance for 2011: Operating in a market in recovery.
Presentation accompanying the Vestas Annual Report 2010.
Vestas. (2011c). Shareholder information: 1/2011.

34

Wene, C.O. (2000). Experience curves for energy technology policy. Paris, France: International
Energy Agency and Organisation for Economic Co-operation and Development. 127 pp.
Wiggelinkhuizen, E.; Verbruggen, T.; Braam, H.; Rademakers, L.; Xiang, J.; Watson, S. (2008).
Assessment of Condition Monitoring Techniques for Offshore Wind Farms. Journal of Solar
Energy Engineering (130:3); pp. 031004031009.
Winters, T. (2008). The Rising Cost of Electricity Generation. The Electricity Journal (21:5);
pp. 5763.
Wiser, R. (2010). Understanding Trends in Overall U.S. Wind Project Performance.
WINDPOWER 2010. Dallas, Texas, May 2008.
Wiser, R. (2011). Unpacking the Drivers Behind Recent U.S. Wind Project Installed Cost and
Performance Trends. Presentation submitted to IEA Task 26.
Wiser, R.; Bolinger, M. (2011). 2010 Wind Technologies Market Report. DOE/GO-1020113322. Washington, DC: U.S. Department of Energy Office of Energy Efficiency and Renewable
Energy.
Wiser, R.; Lantz, E.; Bolinger, M.; Hand, M. (2012). Recent Developments in the Levelized Cost
of Energy From U.S. Wind Power Projects. Presentation submitted to IEA Task 26.
Wiser, R.; Yang, Z.; Hand, M.; Hohmeyer, O.; Infield, D.; Jensen, P.H.; Nikolaev, V.; OMalley,
M.; Sinden, G.; Zervos, A. (2011). Wind Energy. In IPCC Special Report on Renewable Energy
Sources and Climate Change Mitigation [O. Edenhofer, R. PichsMadruga, Y. Sokona, K.
Seyboth, P. Matschoss, S. Kadner, T. Zwickel, P. Eickemeier, G. Hansen, S. Schlmer, C. von
Stechow (eds)]. Cambridge, UK and New York, NY, USA: Cambridge University Press.

35

Appendix A
Data Inputs for U.S. and Denmark Levelized Cost of Energy Modeling
Modeling inputs from the independent U.S. and Danish levelized cost of energy (LCOE)
analyses are shown in Table 1A and Table 2A, respectively. In both cases, assumptions are
intended to represent industry-wide average estimates; actual performance, capital costs,
operating costs, and other terms may vary widely from one project to the next. All dollar values
are in real 2010 U.S. dollars, and each scenario assumes a 20-year project or economic life. For
the U.S. analysis, performance is modeled and based on a 50 meter (m) annual average wind
speed. As such, the U.S. analysis applied a Weibul k Factor of 2 across all scenarios and utilized
the 1/7th power law to estimate hub height wind speeds, again across all scenarios. U.S.
aggregate income taxes are assumed to be 38.9%. In Denmark, modeling relies on estimates of
the typical full load hours for a wind regime in Eastern Denmark as acquired from industry data
and sources. A Weibul k factor of 2 is assumed in the Danish analysis, and Danish corporate
taxes are 25%.
Table 1A. Inputs in Modeling of U.S. LCOE Estimates 20022003 through 20122013
Characteristics

20022003

20092010

1.5 megawatts
(MW)

1.5 MW

1.62 MW

1.62 MW

1.62 MW

65 m

80 m

80 m

80 m

100 m

70.5 m

77 m

82.5 m

100 m

100 m

$1,300/kilowatt
(kW)

$2,150/kW

$1,600/kW

$1,850/kW

$2,025/kW

$60/kW-year

$60/kW-year

$60/kW-year

$60/kW-year

$60/kW-year

Losses (availability,
array, other)

15%

15%

15%

15%

15%

Financing cost/
discount rate (nominal)

9%

9%

9%

9%

9%

Nameplate capacity
Hub height
Rotor diameter
Installed capital cost
Operating costs

36

Current Turbine Pricing: ~20122013

Table 2A. Inputs in Modeling of Danish LCOE Estimates 2002 through 2012
Characteristics

2002

2009

2012

0.9 MW

2.3 MW

3.0 MW

Hub height

49 m

80 m

80 m

Rotor diameter

52 m

93 m

112 m

$1,465/kW

$1,908/kW

$1,857/kW

Operating costs

$16/megawatthour (MWh)

$16/MWh

Full Load Hours

2209

3102

3602

8%

8%

8%

Nameplate capacity

Installed capital cost

Financing cost/
discount rate (nominal)

37

$16/MWh

Appendix B. The Cost of Wind Energy Presentations

The Cost of Wind Energy

THE COST OF WIND ENERGY

AEE contribution to
IEA task 26, WP#2

Credit: AEE

Alberto Cea & Emilien Simonot

esimonot@[Link]
Technical department
Spanish Wind Energy Association

July 2011

OBJECTIVES OF THE STUDY

Identify the factors influencing the evolution of wind

turbine generator (WTG) design and technical
characteristics
Understand the influence of design evolution and
external factors on the cost of WTG
Highlight critical issues that could influence the cost of
WTG in the near future.

38

APPROACH OF THE STUDY

Study challenges:
Assess how WTG costs vary over time
Determine the factors that influence WTG cost variations,
including:
Needed technology changes (e.g., size and weight increase,
and new technology standards)
External factors (e.g., raw material costs, energy, and
manpower)
Try to identify the effect of each factor
Data inputs:
All sources available, including AEEs statistics, WTG
manufacturers portfolios, and theses
Study criteria:
WTG broken down into three main types studied separately:
Rotor, Nacelle, and Tower.

WIND ENERGY
DEPLOYMENT IN
SPAIN

39

HISTORICAL AND GEOGRAPHIC TRENDS

Most of the best spots
occupied during the early
development of wind energy
were in place before 2004.

6.000

En 2010
En 2009

5.000

En 2008
En 2007

4.000

MW

En 2006

Autonomous
community

Accumulated
capacity end
2010 (megawatts
[MW])

Castilla y Len

4803,815

Total a 01/01/2004

1.000

3709,19
3289,325
2979,33
1764,01

Comunidad
Valenciana

986,99

Navarra
Catalua
La Rioja
Asturias

968,37
851,41
446,62
355,95

Pas Vasco

153,25

Murcia
Canarias
Cantabria
Baleares

189,91
138,92
35,3
3,65

En 2004

2.000

Fuente: AEE

Today, leading regions (in terms of installed capacity) are not the best
wind resource regions. As such:
- Developers have to deal with medium and low wind conditions
- There is a need to adapt technology to these sites.

TREND IS TO INSTALL BIGGER AND BIGGER WTG

Installed WTG average capacity (Spain)

2000
1500
1000
500

40

2010

2009

2008

2007

2005

2006

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

(kW)
Average capactiy (MW)

2500

1990

Castilla-La
Mancha
Galicia
Andaluca
Aragn

En 2005
3.000

THE OPERATING HOURS OF WIND FARMS DECREASED SLOWLY

DURING THE LAST DECADEFROM 2,500 TO 2,000 HOURS AND
STABILIZED AROUND 2,000 HOURS
Average operating hours for projects, by year of instalation.
3.000

2.500

Average operating hours

2.000

1.500

1.000

500

Year of instalation

Ant98

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2.440

2.496

2.544

2.366

2.328

2.181

2.201

2.077

2.148

2.147

1.973

1.931

2.170

1.993

Source: CNE

GLOBAL CAPACITY FACTOR EVOLUTION

29%
27%

CAPACITY FACTOR

25%
23%
21%
19%
17%
15%
1998

1999

2000

Early development of
wind power: best wind
condition sites

2001

2002

2003

2004

2005

Worse wind conditions

2006

2007

2008

2009

2010

Worse wind conditions but

technology adaptation

The current average capacity factor is lower than in early 2000 and is oscillating from
23% to 25%.
9

41

FACTORS
INFLUENCING WIND
POWER COSTS
10

WTG AND WIND FARM COSTS

Otros costes de
inversin
8,78%
Conexin
10,84%
Obra Civil
7,36%

Aerogeneradores
73,02%
Fuente: Estudio INTERMONEY y AEE

Onshore wind farm investment costs are closely linked to WTG

costs which represent 73% of the total investment costs.

11

42

WTG AND WIND FARM COSTSNO APPLICATION OF THE

LEARNING CURVE
Wind farms costs 2001-2009

Wind turbines costs 2001-2009

Source: CNE

Source: CNE

Average installed project costs have increased 100% in

the period 2001-2009.

Average WTG costs have increased 50% in the period

2001-2009.

Before 2006: investment costs increase and standard deviation

remains low. Low cost period.

Before 2004: demand meets offer.

From 2004 to 2006: high demand and strong cost increases.

2007: 3,520 MW installed. The high demand pushes costs up. Wide
range of costs. Transition period.

2007: Wide range of costs due to very high demand. The variation of
raw material costs starts.

Since 2008: Stabilization of costs at higher levels. The standard

deviation decreases but the average costs remain high.
High cost period.

Since 2008: Costs remain high as the result of:

- Lower demand due to the end of the regulatory framework
- Increase of raw material prices (i.e., steel, copper, and fiberglass)

Future trends: - Prices may decrease as a result of more offers at lower prices (new manufacturers especially from Asia)
- Technology evolution and costs of material have to be taken into account.

LEARNING CURVE AND FUTURE TRENDS

In Spain, it appears that no learning curve exists, from the early
industrial stage of wind power to the present, no cost reductions have
been observed.
Reduction of production costs must have happenedbut were
hidden by a high demand marketwhich is not an appropriate
scenario to incentivize cost reduction.
Recent changes introduce new industry challenges and costs:
The regulatory framework: needs for local production in the new
regional tenders mean a rise in investment costs
Less and less good wind condition sites: bigger diameter and higher
towers
Market volatility and supply insecurity for raw materials (i.e., steel,
copper, and rare earth metals)
Decrease of Spanish market installation rhythm means WTG
manufacturer strategy to look abroad new investments.
13

43

COST OF MATERIALS

14

WEIGHT AND RAW MATERIAL INFLUENCE ON WTG COST

VARIATION
Formula used to determine how weight increase and raw material costs
independently influence WTG cost:
Determination of influence of weight increase W for each material between year
n-x and year n:

W=

(Weightn Weightn x ) Cost n x

(Cost n Weightn ) (Cost n x Weightn x )

Determination of influence of raw material cost variation C for each material

between year n-x and year n:

C=

(Cost n Cost n x ) Weightn

(Cost n Weightn ) (Cost n x Weightn x )

15

44

RAW MATERIALS COSTS

MATERIAL

2006 PRICE

CURRENT PRICE

Iron (Cast Steel) (1)

370 /t

500 /t

Steel (2)

460 /t

620 /t

Copper

~ 5.500 $/t = ~ 4.500/t (2)

~ 9.000 $/t = ~ 7.000 /t (3)

Glass fiber

~ 450 /t (4)
Unknown composition

Resin

Values to be considered in the next slides.

Iron price: 80% of steel price
Flat Products/HRC/S. Europe domestic Ex-Works /t
IMF Commodity Price Survey
Personal communication
16

STEEL COST EVOLUTION SINCE 2006

Cost of Steel Flat Products (South Europe domestic)

900
800

85.000 /MW

700

70.000 /MW

600
500

/t

Monthly min
400

45.000 /MW

300
200

Monthly max

42.000 /MW

+ 40 /kW

- 43 /kW

Monthly Avg

+ 28 /kW

100

Flat Products/HRC/S. Europe domestic Ex-Works /t

45

ene 11

jul 10

oct 10

abr 10

ene 10

jul 09

oct 09

abr 09

ene 09

jul 08

oct 08

abr 08

ene 08

jul 07

oct 07

abr 07

ene 07

jul 06

oct 06

abr 06

ene 06

(1)
(2)
(3)
(4)

STAINLESS STEEL COST EVOLUTION SINCE 2006

Stainless Steel /t (South Europe

domestic)
5000
4500
4000
3500
3000
2500
2000
1500
1000
500

ene 11

jul 10

oct 10

abr 10

ene 10

jul 09

oct 09

abr 09

ene 09

jul 08

oct 08

abr 08

ene 08

jul 07

oct 07

abr 07

ene 07

jul 06

oct 06

abr 06

ene 06

Stainless Steel/CR 304 2B 2mm coil transaction/S. Europe domestic delivered /t

1- NACELLE

19

46

NACELLE WEIGHT: INFLUENCE OF

ELECTRICAL CONVERSION TECHNOLOGY
TYPE OF ELECTRICAL CONVERSION - WEIGHT
500
450
Asncrono

NACELLE WEIGHT (t)

400
350

DFIG

300

Multipole Wound rotor

250

PMG

200

Squirrel Cage

150

Polinmica (DFIG)

100

Polinmica (Multipole Wound

rotor)

50

Lineal (PMG)

0
0

1000

2000

3000

4000

5000

6000

RATED POWER (MW)

Typology with permanent magnet generator could be the lightest one,

whereas multipole wound generator technology is the heaviest one.

20

EVOLVING GRID CODES WILL ENCOURAGE THE USE OF NEW

CONVERSION SYSTEMS
Breakdown of electrical conversion tecnology in Spain.
25000

Cumulative installed capacity

20000

15000
Full Converter

For installed WTG,

retrofits are needed to
comply with new grid
codes. As a result,
unplanned costs can
arise during the
operation phase.

DFIG
10000

Squirel cage

5000

0
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Due to new grid codes requirements, the technological trend is to use

full converter in new WTG designs.
21

47

ISSUES: GENERATORS AND POWER ELECTRONICS

Trends to build WTG equipped with permanent magnet

generators or multipole technology include:
No gearbox = may reduce operation and maintenance (O&M) costs
Full converter = easier grid integration but higher cost
Larger generators = heavier

Permanent magnet generators use of rare earth:

Demand of rare earth is high because of lots of other applications
China generates 97% of world production and risks exist regarding
future delivery.
Symbol

Name

Pr

Praseodymium

Nd

Neodymium

Sm

Samarium

Gd

Gadolinium

Dy

Dysprosium

Selected Usages
Rare-earth magnets, lasers, core material for carbon arc lighting,
colourant in glasses and enamels, additive in Didymium glass used in
welding goggles, ferrocerium firesteel (flint) products
Rare-earth magnets, lasers, violet colors in glass and ceramics, ceramic
capacitors
Rare-earth magnets, lasers, neutron capture, masers
Rare-earth magnets, high refractive index glass or garnets, lasers, x-ray
tubes, computer memories, neutron capture
Rare-earth magnets, lasers

Rare earth used in the construction of permanent magnets.

22

ISSUES RELATED TO NACELLE MATERIALS

The weight of nacelle rises from 30 tons in 1 MW WTG up to

more than 120 tons in multi MW WTG
Need to consider specific logistic solutions: higher load trucks
and heavier lifting capacities
Nacelle raw material structure:
2% 3% 1%
9%

Raw material
28%

Iron
Steel
Copper
Glassfiber
Resin
Quartz

57%

Tons

Iron

10, 5

Steel

21, 69

Copper

3, 5

Glass fiber

0, 8

Resin

1, 2

Quartz

0, 35

Gamesa G8X Nacelle (real data)

WTG nacelle raw material breakdown (GAMESA G8X)

Source: Life cycle assessment of a multi-megawatt wind turbine - E. Martnez, F. Sanz - Renewable Energy 34 (2009) 667673

48

23

EVOLUTION OF COST OF MATERIALS IN NACELLE

Use of raw material in WTG nacelle
Period

2004 2010

Present and
Future

WTG type

2 MW doubly

fed induction
generator (DFIG)

Cost

Estimation for a
3 MW DFIG or
similar

Cost

Variation
Due to raw
material cost
changes

Due to WTG
weight increase

34%

66%

Iron (cast
steel)

~10.5 ton

3.885

~33 tons

16.500

Steel

~22 ton

10.120

~68 ton

42.160

34%

66%

Copper

~3.5 ton

15.750

~10 ton

70 000

46%

54%

Glass fiber

~0.8 ton

~2.3 ton

1100

Resin

~1.2 ton

~3.5 ton

Unknown composition

Quartz

~0.3 ton

~1.2 ton

No data available

No historical data available

The cost increase of raw material in WTG nacelles from steel:

Approximately 1/3, from the cost increase of raw material.
Approximately 2/3, to the size and weight increase of the WTG.
For copper, these two factors each share 50% of the variation.

24

2- BLADES

25

49

ROTOR SIZE: BIGGER AND BIGGER

Rotor diameter of installed WTG (Spain)
140

120

Rotor diameter (m)

100

80
Mean Diameter
Min Diameter

60

Mx Diameter
40

20

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

The increase of diameter has been constant since the earliest stages of wind
power development.
Latest WTG designs with rotor diameter of more than 100 meters (m) show that
this tendency should continue.

BLADES ARE ALSO GETTING HEAVIER AND HEAVIER

BLADE

5000

18000
16000

Resin

5540

14000

4200

3000

12000
10000
8000

2000

1250

4000
2000

300

0
0

10

1650

850

600
500

3000
2500 2300

2500

1500

6000

2000

1500

From 1 to 18 tons

750 1000
660

20

30

3000

40

50

60

70

3.9 tons

Fiberglass

2.6 tons

Total

6.5 tons

Continuous growth is
observed, however,
length seems to grow
quicker than weight.

BLADE LENGTH (m)

7000

7000

6000

6000

WTG rated power (kW)

WTG rated power (kW)

BLADE WEIGHT (Kg)

G8X (~40m) Blade material

breakdown

RATED POWER (kW)

20000

5000
4000
3000
2000
1000

5000
4000
3000
2000
1000
0

0
0

10

20

30

40

50

60

70

5000

10000
Blade weight (kg)

Blade length (m)

50

15000

20000
27

ISSUES RELATED TO BLADE MATERIALS

Wind turbine blades are made of a main frame and several different layers:
Frame and internal layers are designed to be as light as possible
External layers responsible for protection and wind harnessing are
composed of glass fiber and resins
Chinese fiberglass is gaining market share due to a low price policy.

Technological trends:
Segmented blades, optimized logistics
New materials, including use of carbon fiber, resins improvement, and
special coatings (e.g., anti biofouling and stealth)

WTG blades have not experienced very important changes in their

structural designs since the 90s. The possibility to reach a maximum
length/weight ratio has to be considered and R&D efforts are planned to
address this issue (including research on new materials and designs).

EVOLUTION OF COST OF MATERIAL FOR BLADES

Use of raw material in WTG blades
Period

2004 2010

Blade type

40 m long/
7, 5 ton

Cost

~ 2.6 ton

No historical
data available

Glass fiber
Resin

~1.2 ton

Present and
Future
60 m long/
17, 7 ton

Cost

~6.2 ton

2790

~3.5 ton

Variation
Due to raw
material cost
changes

Due to WTG
weight increase

No historical data available

Unknown Composition

29

51

3-TOWERS

30

WEIGHT INCREASE IN WTG TOWERS

TOWERS

RATED POWER (kW)

1200

CONCRETE TOWER

3000

1000

3000

TOWER WEIGHT (tons)

800

4500

Enercon E-112 1

600

STEEL TOWER

OFFSHORE SPECIFIC
DESIGN2

400

5540
1800

220 tons

200

2500

2000

900

1000
1000

1300
1000
1000

45 tons

1000

0
0

(1)
(2)

5000

20

40

[Link] for E-112 reference

Lower hub heights

1250

60

80
HUB HEIGHT (m)

52

1250

2000
2000
1500
100

120

140

31

CORRELATION BETWEEN DIAMETER/TOWER HEIGHT AND WTG

IEC CLASS
IEC WTG Class: Rotor Diameter and Hub Height
140
140

120
120

Class III
Rotor Diameter (m)

100
100

80
80

Class II

Clase I I
Clase IIII

60
60

Clase IIIIII a
40
40

Class I

20
20

00

20
20

40
40

60
60

8080

100100

120120

140 140

160 160

Hub Height (m)

(Difficult interpretation: WTG Class I/II and WTG Class II/III mixed)

Higher class (lower wind speeds) means larger rotor diameters and higher
towers.

32

EVOLUTION OF COST OF MATERIAL FOR TOWERS

Use of raw material in towers

Period
WTG type

2006

2010

2 MW DFIG
Hub height: 80 m
(steel tower)

Cost
STEEL
TOWER

~ 180 ton

Variation

Estimation for a 3
MW DFIG or similar
Hub Height: 125 m
(steel or hybrid
tower)

Cost

82.800

~ 425 ton

~ 263.000

Steel

+ 38%

HYBRID TOWER

Steel
Concrete

Due to raw Due to WTG

material cost
weight
changes
increase

162 ton (including steel

segment and steel
reinforcement in the
concrete segment)

100.400

740 ton

46 000

Total Hybrid Design

146.400

+ 62%

- 62% (compared to full steel

structure)

- 44% (compared to full steel

structure)

The cost increase of raw material in WTG towers from steel:

Approximately 1/3, to the cost increment of raw material
Approximately 2/3, to the size and weight increment of the WTG.
When using hybrid designs, the raw material cost reduction resulting from the use of
concrete is important, however, it must also consider an increase in manpower costs (not
counted here).

53

33

SUMMARY

34

Summary: Use of raw material in WTG

Period
WTG type

2006

2010
Variation

NACELLE

Cost

Estimation for a 3
MW DFIG or similar
Blade: 60 m long/
17, 7 ton
Hub Height: 125 m
(steel or hybrid
tower)

Iron (Steel Casts)

~10.5 ton

3.885

~33 tons

16.500

+ 34%

+ 66%

Steel

~22 ton

10.120

~68 ton

42.160

+ 34%

+ 66%

Copper

~3.5 ton

15.750

~10 ton

70 000

+ 46%

+ 54%

~0.8 ton

No historical
data available

~2.3 ton

1100

Glass fiber
Resin

BLADE

Glass fiber
Resin

2 MW DFIG
Blade 40 m long/
7, 5 ton
Hub height: 80 m
(steel tower)

~1.2 ton
~ 2.6 ton

~3.5 ton

STEEL HYBRID
TOWER TOWER1

No historical
data available

~6.2 ton

82.800

~ 425 ton

~1.2 ton
~ 180 ton

Cost

Concrete

No historical data available

Unknown composition

2790

~3.5 ton

No historical data available

Unknown composition

~ 263.000

Steel

Steel

Due to raw Due to WTG

material cost
weight
changes
increase

+ 38%
-

162 ton

100.400

740 ton

46 000

Total Hybrid Design

146.400

+ 62%

- 62% (compared to full steel

structure)
- 44% (compared to full steel
structure) 35

(1) Data for hybrid tower extracted from tall tower for large wind turbines, report from Vindfork project V-324 Hga torn fr vindkraftverk Elforsk rapport 10:48

54

THANK YOU FOR YOUR

ATTENTION

Credit: AEE

Alberto Cea & Emilien Simonot

esimonot@[Link]
Technical department
Spanish Wind Energy Association

55

Danish Wind Project Performance, Capital Cost, and Price Trends

Danish Wind Project Performance,

Capital Cost, and Price Trends
IEA Wind Task 26: Cost of Wind Power
Edward James-Smith
Golden, March 29, 2011

Explanatory Notes
Price data is from actual projects in Denmark and is published
in the report Vindmllers konomi, February 2010
Updates for 2010 from projects and turbine manufacturer
Data for 2004 to 2007 is from a very small pool of turbines
due to the low deployment in Denmark during this period
Production data is from the Danish Energy Agency wind
turbine register
Wind data is from the Danish Wind Turbine Owners
Association
Performance data from 2004 is high as most turbines erected
that year were demonstration turbines with large rotors and
high hubs.
2

56

PRICE TRENDS IN DENMARK

General Trends in Denmark - Price

Capital costs have fallen sharply since 2008
Some turbines have been available at lower
prices due to orders being cancelled or defaulted
can mask true price of turbines
Overcapacity in manufacturing downward
pressure
Turbines are getting higher and rotor diameters
larger
Manufacturing base is moving away from
Denmark.
4

57

Breakdown of costs
Project
development
4%
Roads
2.1%

Other
20%

Turbine
80%

Financing
costs
3%
Land
3.7%

Internal grid
1.3%
Foundations
6.4%

Data from Vindmllers konomi (2010)

On average, wind projects spend 9% of the total budget on

purchasing old turbines for repowering. These costs are
considered to be covered by the repowering subsidy and are
therefore not included.
5

Turbine Prices in Denmark

2008 Euro/kilowatt (kW)
2500

2000

1500

1000

500

Data From: Vindmllers konomi (2010)

6

58

Diverse
0.4%

Learning Rates and Prices

16000
14000

DKK/kW installed capacity

12000
10000
8000
6000
4000
2000
0
1990

1995

2000

2010

2005

LR 6,7 % 1990 - 2009

2015

Faktiske priser/prisprognose

2020

2025

2030

LR 9 % 1990 - 2001

Source: Vindmllers konomi (2010)

7

Prices in the European Union Using European

Wind Energy Association Targets and
Learning Rates
10,000
9,000

DKK/kW installed capacity

8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0
2010

2015
Hj estimat

2020
Mellem estimat

Source: Vindmllers konomi (2010)

8

59

2025
Lav estimat

2030

Price Prognosis: Total Investment (Danish Krone per

Megawatt-hour [MWh]) in Year One Equivalent of
One Share in Danish Turbine Cooperative
3,500

Investment per 1000 kWh production in year 1 (2008-kr.)

3,228
2,990

3,000

2,808

2,711

2,688

2025

2030

2,500

2,000

1,500

1,000

500

0
2010

2015

2020

Source: Vindmllers konomi (2010)

9

PERFORMANCE OF DANISH
TURBINES
10

60

General Trends in Performance

Turbines have higher hub heights and larger
rotors
Resulted in marked increase in capacity factor
since 2000 average capacity factor for
onshore turbines has increased by ~ 50%
The increase in capacity factor appears to
have resulted in lower prices per kilowatthour (kWh).
11

Hub Heights
90

Average hub height for new capacity (m)

80
70
60
50
40
30
20
10
0
2000

2001

2002

2003

2004

2005

Source: Vindmllers konomi (2010)

12

61

2006

2007

2008

2009

2010

Rotor Diameters
120

Average rotor diameter for new turbines (m)

100

80

60

40

20

0
2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

Source: Vindmllers konomi (2010)

13

Annual Wind Energy in Denmark

Source: Danish Energy Agency Wind Turbine Register

Production data from January 1 to September 30, 2010 was used for all
onshore wind turbines commissioned from 2000 to September 2010 and
adjusted to normal wind year to determine average performance per
commissioning year.

14

62

Performance of Danish Onshore Turbines From

January to September 2010
(Adjusted to Normal Wind Year)
110
100
90
80
70
60
Rotor diameter

50

Hub height
Capacity factor

40
30
20
10
0
2000

2001

2002

2003

2004

2005

2006

2007

2009

2008

2010

Year of commissioning

15

Price Equivalent of One Share in Danish Turbine

Cooperative, Total Investment Euro/MWh in
Year One
800

Investment in EURO/MWh of first year production

700

600

500

400

300

200

100

0
2000

2001

2002

2003

2004

2005

16

63

2006

2007

2008

2009

2010

ADDITIONAL SLIDES ON COSTS AND

PRODUCTION
17

Change in Energy Production and Price Per Megawatt for SWT

2.3 MW with 82, 93, and 101 Meters at Different Wind Speeds

18

Source: Vindmllers konomi (2010)

64

Production in Relation to Weight

Source: Vindmllers konomi (2010)

19

Cost of Steel

20

Source: Vindmllers konomi (2010)

65

Labor Costs

Source: Vindmllers konomi (2010)

21

Summary of Costs

22

Source: Vindmllers konomi (2010)

66

Development of LCOE for Wind Turbines in Denmark

Development in LCOE for Wind

Turbines in Denmark
Edward James-Smith

Methodology
Input

Power curves for typical turbines built in 2002 and 2009 and an
example of a high-performance turbine available for delivery in 2012
Capacity factor calculated for turbines at average wind speeds of 7.5
meters (m) per second (s), 7.2 m/s, 6.9 m/s, and 5.9 m/s and Weibull
shape parameter of 2.3
Cost data from IEA Wind Task 26 and turbine manufacturer
Historical generation data for 2002 and 2009 turbines and first year
generation data for two turbines of type available for delivery in 2012
Historical data used to validate calculated data.

Output

Levelized cost of energy (LCOE) determined in 2008 Euro using IEA

Wind Task 26 LCOE calculator
Comparison of relative price and relative value in Denmark using
current subsidy program.

67

Capacity Factor for Turbines

2002

2009

2012

7.5 m/s

34%

37%

48%

7.2 m/s

31%

34%

45%

6.9 m/s

28%

32%

43%

5.9 m/s

20%

24%

33%

Actual Danish
Average

25%

35%

41%

LCOE 2002
90
80
70
60
50
40
30
20
10
0
7.5 m/s

7.2 m/s

6.9 m/s
LCOE (2008 Euro/MWh)

68

DK avg
Capacity factor (%)

5.9 m/s

LCOE 2009
90
80
70
60
50
40
30
20
10
0
7.5 m/s

DK avg

7.2 m/s
LCOE (2008 Euro/MWh)

6.9 m/s

5.9 m/s

Capacity factor (%)

LCOE 2012
70

60

50

40

30

20

10

0
7.5 m/s

7.2 m/s

6.9 m/s
LCOE (2008 Euro/MWh)

69

DK avg
Capacity factor (%)

5.9 m/s

Remarks
The average LCOE for turbines commissioned
or ordered for commissionining fell from 65
in 2002 to 60 in 2009 and 53 in 2012.
A comparison of calculated LCOE for 2002,
2009, and 2012 indicates that higher wind
sites were required in 2009 to bear the higher
costs of turbines despite higher performance
than in 2002.
7

Comparison of Turbines with Danish

Support Program
Price (2008
Euro)/
Kilowatt
[kW])

Turbine

Megawatthour
(MWh)/
Megawatt
(MW)

EUR/MWh

Relative
price

Income:
subsidy
(M)

Income:
market
(M)

Total
(M)

Income/
project
price

Relative
value

2002

1104

0.9 kW

2209

500

0.66

2.12

2.78

2.79

0.85

2009

1438

3 MW

3102

464

0.93

2.2

9.9

12.13

2.81

0.86

2012

1400

3 MW

3602

389

0.78

2.2

11.52

13.73

3.27

Average market price used: EUR53/MWh

Subsidy: EUR33/MWh for 22,000 full load hours
Balancing subsidy and balancing costs not included
Current subsidy regime used for 2002 turbines. Another subsidy regime was in place
at the time.
8

70

Engineering Model-Based Technology Projections

Innovation for Our Energy Future

Engineering Model-Based Technology

Projections
IEA Wind Task 26 Working
Group Meeting
E. Lantz (NREL) and M. Hand
(NREL)

March 29, 2011

NREL PIX 16813

NREL is a national laboratory of the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.

Historical Costs and Future Projections Appear at Odds

$3,000

Capital Cost (2009$/kW)

$2,500

EPA 2009

$2,000

AEO 2010
DOE 2008

$1,500

$1,000

$500

EPRI 2009

EREC/GPI 2008
Lemming et al. 2009
Wiser and
Bolinger 2010

$0
2000

2005

2010

2015

2020

2025

71

2030

2035

2040

2045

2050

Innovation for Our Energy Future

Attempting to Reconcile Recent Trends and Cost

Projections Generates Critical Questions
Working Premises
Cost of energy (not installed cost) is the design driver
Turbine manufacturers must balance lower installed costs
against increased performance
Applicable Questions
Are projections of lower installed costs and increased
performance realistic?
What level, and in what turbine system(s), of innovation is
required to achieve installed cost and performance
projections?

Innovation for Our Energy Future

Learning Curve Alternatives: Expert Elicitation

TIOs Potential for Improvement

(% change from reference turbine)

Annual Energy Production

Survey industry experts to

gather a range of possible
technology outcomes

Turbine Capital Cost

Balance of Station Cost
Levelized Replacement Cost
O&M Costs

% Change

-30 -20 -10

Develop probability
distributions associated with
various technical outcomes

+10 +20 +30 +40

Advanced (Enlarged) Rotor TIOs

Manufacturing TIOs

Reduced Energy Losses and

Increased Availability TIOs

Advanced Tower TIOs

Example

Site-Specific Design/Reduced
Design Margin TIOs

U.S. Department of Energy

(DOE) Risk Analysis
project conducted in
association with
WindPACT analytic studies

New Drive Train Concept TIOs

Advanced Power Electronics TIOs

Learning Curve Effects

Note: TIO =Technology Improvement Opportunity;

O&M = Operations and Maintenance
Source: Cohen et al., 2008
4

72

Innovation for Our Energy Future

Learning Curve Alternatives: Engineering Model

Bottom-up, component-level, system analysis

1. Evaluates tangible technology advancements: proposed and
anticipated technology advancements, with a focus on realizable
opportunities
2. Measures the potential value of individual innovations: specific
opportunities are quantified independently, and results are combined
to arrive at a cost estimate
3. Considers installed cost and technology performance: allows
weighting of tradeoffs between cost increases and improvements in
energy capture
4. Offers the opportunity to substantiate a learning curve
projection: creates a pathway to a future point

Requires simplification of complex engineering problems

Does not typically represent economy of scale or volume-based
cost improvements explicitly
5

Innovation for Our Energy Future

NRELs Wind Turbine Design and Cost Scaling Model

Flexible, modular, spreadsheet
model:
Perform trade-off studies of turbine
technology options
o Determine technology changes with
the greatest potential to reduce cost of
energy (COE)

Generate wind technology cost

and performance trajectories
o Used in generation capacity
expansion modeling

Innovation for Our Energy Future

73

Cost Model Features

Permits scaling of components to analyze turbine
configurations
Costs based on DOE WindPACT analysis (and
development subcontracts)
Includes simple and advanced scaling curves

Illustrates some technology pathways in relation to

industry
Pathways based on industry trends and WindPACT
analysis

Validated using industry data where possible

Innovation for Our Energy Future

Cost Model Functional Overview

Spreadsheet tool consists of two primary elements:
1. Cost and scaling calculator(s)

Series of models for individual turbine components (e.g., blades,

hub, low-speed shaft, pitch mechanisms, and mainframe)
Most relationships are simple functions of rating, rotor diameter,
and hub height
Does not directly consider variable cost of capital or supply and
demand market price pressures

2. Simplified energy production calculator

Assumes equivalent blade geometry and airfoil performance

Only utilizes the peak coefficient of power (Cp)
Operational range divided into meter (m) per second (s) bins,
for each bin hub power, drive-train efficiency, and total energy are
calculated, energy capture and capacity factor are computed from
totals
8

74

Innovation for Our Energy Future

Wind Turbine Blade Innovation Pathway

30000
WindPACT Baseline Design

Commercial Blade Data

25000

Mass (kg)

20000
15000
10000
5000
0
0

10

20

30

40

50

60

70

Rotor Radius (m)

Source: Based on Fingersh et al., 2006
Innovation for Our Energy Future

Wind Turbine Blade Innovation Pathway

30000
LM Glasfiber Blades

25000

WindPACT Baseline Design

Commercial Blade Data

Mass (kg)

20000
LM Glasfiber Blades

15000
10000
5000
0
0

10

20

30

40

50

60

70

Rotor Radius (m)

Source: Based on Fingersh et al., 2006
Innovation for Our Energy Future

75

Wind Turbine Blade Innovation Pathway

30000

LM Glasfiber Blades

25000

TPI Innovative Blade

Concepts

20000

Mass (kg)

WindPACT Baseline Design

Commercial Blade Data

LM Glasfiber Blades

15000
WindPACT Final Design

10000
5000
0
0

10

20

30

40

50

60

70

Rotor Radius (m)

Source: Based on Fingersh et al., 2006
Innovation for Our Energy Future

NREL Engineering Model Analysis Scope

Objective

Quantify the value of potential technological advancements on larger turbines

Evaluate the feasibility of achieving the U.S. 20% wind cost and performance
trajectory while simultaneously moving to 5-megawatt (MW) turbines

Approach

Develop hypothetical turbines that are in line with the 20% wind study, to
achieve 20% wind performance trajectory, and scale gradually to 5-MW
Apply individual technological improvements to those turbines to evaluate their
impact on installed cost and turbine capacity factors

Caveats

Emphasis on incremental technology and marketplace changes

Innovations must be adopted widely
Continued analysis of how balance-of-plant (BOP) costs are affected by larger
turbines and consideration of emerging innovation strategies would further
enhance our understanding

12

76

Innovation for Our Energy Future

Land-based Wind Technology Development

Pathway to Achieve 20% Wind Targets
No R&D Basecase (simple scaling)
Best Technology

Meters

160

Expected Technology
Estimated Capacity Factor (Right Axis)

49%

140

47%

120

45%

100

43%

80

41%

60

Capacity Factor

180

39%

40

37%

20
0

35%
2.5 MW

3.5 MW

5.0 MW

2.5 MW

Hub Height

3.5 MW

5.0 MW

Rotor Diameter

Drivetrain Design
2.5 MW
No Research and Development (R&D)
Basecase (Simple Scaling)

Expected Technology

3.5 MW

5.0 MW

3-stage geared
multi-generator

single-stage geared

direct drive

Best Technology

Notes: Technology scenario labels are generally defined by WindPACT risk analysis and summarized in Cohen et al.
2008. Individual turbine designs are optimized to reach capacity factor targets of the 20% wind study; capacity factor
targets assume continued scaling of turbines, hence, higher capacity factor targets are expected for larger machines.
13

Innovation for Our Energy Future

Technology Captured by Inputs Range

Turbine Cost Reduction
Blade Technology
Tower Technology

Drivetrain Technology

Manufacturing Efficiency

Power Electronics
Annual Energy Production
(AEP) Increase
Reduced Losses
Power Electronics (Higher
Efficiency)

Innovations Reflected in the Advanced

Technology Parameters
Enhanced structural design, targeted reinforcement
High-tech composites possibly including carbon fiber
Tower feedback to blade pitch controls
Flap-twist coupling in blade design
Reduced blade chord with tip speed increase
Multi-generator drivepath
Single-stage, medium-speed gearbox
Direct drive
Increased automation
Improved resins with greater ease of use
Reduced design margins resulting from more consistent
fabrication
Reduced profit margins as a result of increased volume
High-voltage circuitry
Multi-switch capacity
Semi-conductor devices

Improved micro-siting to reduce array losses

Real-time monitoring and operational modifications
Low soil blades
High-voltage circuit topologies
Multi-switch capacity
14

77

Input Data Source

Cost and scaling model

advanced technology
scaling curves

Cohen et al. 2008 TIOs

(WindPACT Risk
Analysis)

Industry estimates
Cohen et al. 2008 TIOs
Innovation for Our Energy Future

Modeling Changes to Cost and Performance

Note: One additional scenario considers best technology improvements combined with scaling at todays blade
and nacelle mass ratios (i.e., scaling with proportionally comparable masses for blades and nacelles to those
observed in the industry today).

15

Innovation for Our Energy Future

Advanced Technology Permits Scaling to 3.5-MW with

Reduced Costs and Increased Performance

16

78

Innovation for Our Energy Future

Incremental Advancements Across All Turbine

Components/Systems Are Necessary
$2,149

$2,000

$220 $57

$1,750

$186

$1,500

$86

$1,250

$80

$1,236
$171

$1,000

Impact from Expected Technology Advancments

in Specific Turbine Systems

$750

$143

$29

$500
$250
Efficiency (Siting &
Turbine)

Manufacturing

Power Electronics

Towers

Blades

Drivetrain

No R&D Base case

$0

17

20% Wind 2030

Estimate

Reduction
from Other
Technology
Scenarios

Current Mass
Ratios

$2,250

Best Technology

Turbine Cost ($/kW; 3.5 MW)

$2,500

Innovation for Our Energy Future

Additional Technology Advancement is Required to

Achieve Cost and Performance Targets at 5.0-MW

$2,481

$2,500

$91

$2,000

$59
$229

$99

$100

$1,500

$261
Impact from Expected Technology Advancments
in Specific Turbine Systems

$1,000

$1,236

$179
$238

18

79

20% Wind 2030

Estimate

Efficiency (Siting &

Turbine)

Manufacturing

Power Electronics

Towers

Blades

Drivetrain

$0

Current Mass
Ratios

$500

Best Technology

Reduction
from Other
Technology
Scenarios

No R&D Base case

Turbine Cost ($/kW; 5 MW)

$3,000

Innovation for Our Energy Future

Conclusions
1. It is not unreasonable to expect significant turbine cost reductions
(20%) while increasing turbine performance, when excluding shortterm market dynamics.
Based on technological change, we can generally reconcile future cost
reductions with recent trends.

2. Modest cost/performance projections (i.e., 20% wind) can be

achieved, but 5.0-MW machines will require all of the advanced
technology captured in this analysis and then some.
3. All turbine systems can contribute to future cost reductions.

Advanced tower designs and reduced losses appear to offer the greatest
potential to minimize costs while maintaining performance.

4. If existing prototypes and recently commercialized equipment prove

viable, these cost reductions may be closer than we think.

19

Innovation for Our Energy Future

Questions
Eric Lantz
Research Analyst

Strategic Energy Analysis Center

National Renewable Energy Laboratory
[Link]
[Link]
1617 Cole Blvd.
Golden, CO 80401-3393
P: (303) 384-7418
email: [Link]@[Link]
20

80

NREL PIX 05593

Innovation for Our Energy Future

On-Shore Wind in Spain

ON-SHORE WIND IN SPAIN

Heikki Willstedt Mesa

hwillstedt@[Link]

Direccin Polticas Energticas

Asociacin Empresarial Elica

Credit: AEE

Web2Present Webinar Series- Wind Power 2011

MAIN ISSUES

WHY A SUCCESS STORY?

Early mover (90s)
Feed-in tariff (no system changes)
Utilities among the main developers
Own industry with research and development (R&D) and investment

CURRENT SITUATION 2010

Fourth country in installed power in the world (second in the European
Union [EU]) with 20,676 megawatts (MW)
First country in the EU with a wind power generation of 42.7 terawatt-hours
(TWh)
Second country in the world with a wind power penetration of 16.4%
Spanish companies had 9,200 MW installed outside the country

CHALLENGES AHEAD
Achievement of the National Renewable Energy Action Plan (NREAP)
objective of 35,000 MW on-shore and 3,000 MW off-shore
Grid and storage integration (e.g., pumped storage, electrical vehicles, and
other storage)

81

1.
Wind energy: Context and Situation

ENERGY AND ENVIRONMENTAL CONTEXT

The external energy dependence of the European Union is over 50%
and it is forecasted to be above 70% by 2020. Energy diversification is a
priority in the European energy policy and reducing energy dependence
is a geostrategic necessity.
In 2007, the rate of external energy dependence of Spain reached 85.1%
which is a strategic weakness and a burden on the trade balance.
The current rate is 79%.
According to the commitments made within the Kyoto Protocol, Spain
aims to limit carbon dioxide (CO2) emissions by more than 15%
(reference 1990), but it closed 2007 with a 52% increase. Currently
(2010), the rate is at +24%28%.
The long construction period for conventional power stations and the
uncertainty of fossil fuel prices prices triggers the need to use wind
energy as a source of electrical generation.

82

IMPORTANT AND CONTINUOUS GROWTH OF

WIND POWER IN SPAIN
MW

20.676
19.160

20.000
16.699

ANUAL

ACUMULADA

15.115

15.000
11.595
10.013

10.000

8.462
6.185
5.033

5.000

3.522
723

1.408

2.358

0
1998

1999

2000

2001

2002

2003

Fuente: AEE

2004

2005

2006

2007

2008

2009

2010

AOS

Renewable Energies Plan (20052010): 20,155 MW

Official network planning for 2016: 29,000 MW
Further increase expected for compliance with the
goal of 20% of energy demand with renewable
technologies 40% -> 38,000 MW.
5

GWh

WIND POWER PRODUCTION AND CAPACITY FACTOR

45.000
40.000
35.000
30.000
25.000
20.000
15.000
10.000
5.000
0

40%
35%
30%
25%
20%
15%
10%
5%
2004

2005

2006

2007

2008

2009

2010

Generacin elica

15.744

20.520

22.684

27.169

31.136

36.188

42.702

Tasa de variacin elica (%)

34,3%

30,3%

10,5%

19,8%

14,6%

16,2%

18,0%

Cobertura de la demanda (%)

6,7%

8,3%

8,9%

10,3%

11,8%

14,4%

16,4%

0%

Fuente: REE

Wind production increases due to the

growth of installed power, but individual
wind farms have reduced their capacity
factor up until 2008. In 2009 and 2010,
the capacity factor has increased.
Wind production remains constant for
long periods of time.
The maximum coverage was 53% of the
total demand.

40,00%

Factor de capacidad mensual promedio, mnimo y mximo

1998-2010

35,00%
Factor de Capacidad

30,00%
25,00%
20,00%
15,00%
10,00%

MXIMO
PROMEDIO
MNIMO
2009
2010

5,00%
0,00%

Fuente: Elaboracin AEE

83

DEMAND COVERAGE HISTORIC HIGH 9/11/2010

Source: REE
7

2.
Main Technical Challenge: Integration of Wind
Energy into the Grid
Spain is an Electrical Island

84

THE INTEGRATION OF WIND ENERGY RAISES SOME

CHALLENGES

Technical challenges:
Reaction of wind farms to voltage drops: contribution to grid
stability
Voltage control
Grid security and safety.

System operation challenges:

Wind production program: affects the deviations and the use of
ancillary services
Coordinated operation of generation units.

THE ANSWER TO THOSE CHALLENGES REQUIRES THE

DEFINITION OF THE TECHNICAL REQUIREMENTS
Universities and
technological
centres

Power
system
needs: TSO

Turbine
capabilities:
Manufacturers

TECHNICAL
REQUIREMENTS
Laboratories,
testing and
certification
entities

Cost: Promoters

10

85

SINCE 2004, THE WIND SECTOR AND THE TSO HAVE CLOSELY
COLLABORATED TO DEFINE THOSE REQUIREMENTS IN A REALISTIC AND
PRACTICAL WAY
STUDY OF WIND ENERGY INTEGRATION

2004

COORDINATION AND FOLLOW

UP

WORKING GROUP FOR THE

COORDINATION BETWEEN
THE WIND SECTOR (6
REPRESENTATIVES) AND REE

18 MEETINGS
UP TO MARCH 2010

(WITH THE PORTUGUESE TSO

INVOLVEMENT)

LVRT REQUIREMENTS

COORDINATED SYSTEM
OPERATION:
TSO: CECRE
CONTROL CENTRES (18)
WIND FARMS (AROUND 700)

DEFINITION OF THE VERIFICATION

AND CERTIFICATION PROCEDURE
(PVVC)

CREATION OF THE
TECHNICAL VERIFICATION
COMMITTE (CTV):

FOLLOW UP AND UPDATE OF THE

PVVC

STUDY OF WIND ENERGY

INTEGRATION

PROMOTERS
MANUFACTURERS
Wind turbine
generator (WTG)
FACTS
LABORATORIES
AEE
REE

95% OF POWER ADAPTED TO

LVRT
11

WIND POWER CUTS > 100 MW

100

16000

90

14000

80
Number of Power Losses

12000
70
10000

60
50

8000

40

6000

30
4000
20
2000

10
0

2005
2006
2007
Number of Wind Power Losses Bigger than 100 MW

0
2008
2009
Wind Power NO Fulfilling PO 12.3

12

86

A COORDINATED OPERATION IS IMPORTANT TO GUARANTEE

WIND ENERGY INTEGRATION
Parques
R.E. 1

CECRE

Parques
R.E. 2

TSO: COORDINATION CENTRE OF

RENEWABLE ENERGIES
Parques
R.E. n

Solucin Provisional

CC1

CC2

Enlace

CCn
Enlace

Enlace

CECRE: REE Control Centre

of Special Regime
CC: Control Centre

CECRE: OBJECTIVES AND FUNCTIONS

Main function: To achieve a greater level

of integration for renewable energy
sources without compromising the
system s security.

Improves security and effectiveness in

the system s operation
Allows substitution of permanent
preventive criteria for real-time
production control.

GEMAS

Real-time risk assessment due to

voltage dip wind generation losses
Calculation of wind production
limitations
Filter limits for stable solutions in
accordance with legislation.

87

14

AN UNDESIRED SITUATION: WIND POWER REDUCTION

AFTER CECRE ORDER (24/2/2010)

15

3.
The Spanish Market

88

MAIN PLAYERS IN THE SUPPLY CHAIN OF

WIND ENERGY IN SPAIN
Main players in major activity areas include:
Raw materials

Manufacture of
key components

Energy
Shipping

Project
development,
construction,
financing, and
so on

Manufacture of
wind turbines

Etc.
HANSEN
WINERGY
MOVENTAS
ECHESA
SKF
FAG / INA
TIMKEN
LM GLASSFIBER
FIBERBLADE

GES

IBERDROLA

VESTAS
GE Wind
GAMESA
ENERCON
SIEMENS
NORDEX
ALSTOM
ACCIONA
M TORRES
REPOWER
SUZLON

Service
operation,
maintenance,
and sale of
electricity

ACCIONA

IMFUTURE

NEO ENERGA (HC)

GUASCOR SERVICIOS

ECYR (ENDESA)

INGETEAM

EUFER (UN. FENOSA)

REETEC

GAS NATURAL

EFACEC

ENERFN

SBS

EYRA (ACS)
EON
EOLIA

MOLINOS DEL EBRO

..

BREAKDOWN BY MANUFACTURERS
BREAKDOWN BY MANUFACTURERS
OF INSTALLED WIND POWER IN 2010

Fabricantes. Potencia 2010

GE Fuhrlander
6,23% 0,79%

M-TORRES
0,44%

ALSTOM-WIND
9,35%

BREAKDOWN BY MANUFACTURERS
OF ACCUMULATED WIND POWER AT
THE END OF 2010

GAMESA
50,18%
VESTAS
33,01%

DESA
0,49%

SUZLON
1,05%

Fuente: AEE

NORDEX
0,65%

ENERCON
2,34%
SIEMENS
3,52% GE

Fabricantes. Potencia Acumulada

LAGERWEY M-TORRES
0,18%
0,18%
KENETECH
0,18%
REPOWER
0,12%

FUHRLANDER
0,06%
OTROS
0,03%

5,82%

More than 700 companies

from manufacturers to
financial serviceshave
some kind of involvement in
the wind energy sector.

ACCIONA
WIND POWER
7,04%
ALSTOM-WIND
7,54%

GAMESA
53,72%
VESTAS
17,07%

Fuente: AEE

18

89

A SOLID INDUSTRIAL NETWORK TO SUPPORT THE NATIONAL

AND EXTERNAL MARKETS

Spanish wind sector industrial centres

PROMOTERS: ACCUMULATED INSTALLED POWER IN 2010

Promotores. Potencia Acumulada
EOLIA
RENOVABLES
1,78%
GAS NATURAL
1,84%

E. ON Renovables
1,55%

AGRUPACI DE ENERGAS
RENOVABLES, S.A. (AERSA)
1,89%
ENERFN
1,96%
OLIVENTO, S.L.
2,04% EyRA

MOLINOS DEL
EBRO
1,14%
GOVADE
ELICA DE
1,12%
NAVARRA
0,72%
IBERELICA
0,77%
GECAL,
FERSA
S.A.
0,69%
MEDWIND
1,12%
1,19%

GAMESA
ENERGA
1,36%

ELECDEY
0,68%
RENOVALIA
0,51%

3,74%
ENEL GREEN POWER
ESPAA
3,91%
EUFER
4,42%

OTROS
14,06%

EDPR
9,01%
IBERDROLA RENOVABLES
25,00%

Fuente: AEE
ACCIONA ENERGA
19,52%

20

90

4.
An Adequate Legal Framework Until 2009

A STABLE LEGAL FRAMEWORK IS ESSENTIAL FOR THE

DEVELOPMENT OF THE WIND SECTOR

The 2007 Royal Decree, in force from 06/01/07, allows two

remuneration options:
Regulated tariff
Market price and premium: with a cap and floor for the
premium

Wind farm installations governed by the previous regulation

(RD 436/2004) had until 01/01/2009 to decide whether they will
continue with it or choose the new RD

Remuneration updated using the RPI, less an adjustment factor

(0,25% until 2012, 0,5% from then on).

22

91

WIND POWER HELPS KEEP POWER MARKET PRICES LOWER

THAN DURING THE PREVIOUS OIL SHOCK IN 2008
% of demand covered by wind versus market and gas
prices in the Spanish power sector
30.00
25.00
20.00
15.00
10.00

0.00

Jan-05
Mar-05
May-05
Jul-05
Sep-05
Nov-05
Jan-06
Mar-06
May-06
Jul-06
Sep-06
Nov-06
Jan-07
Mar-07
May-07
Jul-07
Sep-07
Nov-07
Jan-08
Mar-08
May-08
Jul-08
Sep-08
Nov-08
Jan-09
Mar-09
May-09
Jul-09
Sep-09
Nov-09
Jan-10
Mar-10
May-10
Jul-10
Sep-10
Nov-10
Jan-11

5.00

% of demand covered by wind

Sources: CNE, OMEL, and REE

Pool price (c/KWh)

Natural gas price (/MWh)

In 2008, the price of the market was so

high that during 3% of the hours, wind
power did not receive premiums.

23

WIND FARMS REMUNERATION

Pool + prima DT 1 RD 661/2007

120

Pool + prima RD 661/2007

102,723
100

Tarifa regulada RD 661/2007

85,941
75,681

/MWh

80

74,073

77,048

78,183

76,896

77,471

72,420

60

40

20

0
2008

2009

2010

Fuente: AEE

24

92

THE SPANISH FEED-IN TARIFF HAS MADE POSSIBLE THE ACHIEVEMENT OF THE 2010
RES ELECTRICAL OBJECTIVE, AND AS A BONUS HAS ENABLED THE SPANISH WIND
INDUSTRY TO BECOME A WORLD PLAYER
For each Euro spent in wind power the Spanish Economy
has gained three in return

=
3.705 M
Premiums
(Tot. 2007-2009)

Power

14,4 % of national
demand
(2009)

Exports +7.554 M

Monthly
contribution to
wind power by
each Spanish
household

PIB Direct
+ 6.192 M

PIB Indirect
4.087 M

Jobs 37.900
Average (07-09)

Fuel
Imports
+4.463 M
CO2 Emissions
+ 1.035 M
Fiscal balance
648 M

R&D&I
556 M

Other social benefits

Datos 2007-09

5.
Challenges Ahead

93

2
5

WIND POWER AND OTHER TECHNOLOGIES: INTEGRATION

Wind and CCGT generation (Jan-Feb/2011)
900,000
800,000
700,000

MWh

600,000
500,000
400,000
300,000
200,000
100,000
0

Wind

CCGT

Demand

Source: REE
27

WIND POWER AND STORAGE: ARE THERE INCENTIVES TO

MAKE IT WORK TO ITS FULL POTENTIAL?
Wind power generation and pumped storage (Jan-Feb/2011)
300,000
250,000
200,000
150,000
100,000
50,000
0
-50,000
-100,000

Source: REE

Wind

Pumped storage

94

The electric car:

a reality after 2015?

28

THE SPANISH WIND POWER MARKET IS SLOWING DOWN ITS GROWTH DUE TO
EXCESSIVE REGULATION AND LACK OF LONG-TERM STABILITY
RD 1565/2010
RD 1624/2010
RDL 14/2010

Ley 17/2007
RD 661/2007
Tarifa regulada
prima variable

4.000
Incremento anual (MW)

100,0%

3.520

3.500

90,0%

RD-L 7/2006

RD-L 6/2009
RES SEE 19/11

TASA DE VARIACIN %
3.000

RD 436/2004

2.276

60,0%
50,0%

42,9%

1.511

1.500

Ley 54/1997

1.165

36,8%

1.551

1.584

1.582

1.152

10,5%

15,8%

500

40,0%
30,0%

18,3%

22,9%

692

1.516

30,4%

950

1.000

70,0%

2.461

49,4%

RD 2818/1998

2.000

20,0%

14,7%
7,9%

287

10,0%
0,0%

1998

1999

2000

2001

2002

2003

2004

RD-L 6/2000

2005

2006

2007

2008

2009

2010
29

THE OBJECTIVE FOR WIND POWER OF 38,000 MW IN 2020 WILL BE

DIFFICULT TO ACHIEVE WITHOUT THE RIGHT POLICIES
Noncompliance
with NREAP
obj.= Gap
5.0009.000
MW

40000
35000

30000

Frontrunner

25000
20000

Uncertain
future

15000
10000
5000

Potencia instalada 2011-2020 (senda paralizacin sector 2010-15)

95

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

0
1998

MW

Tarifa regulada
Prima fija

67,4%

65,8%

2.500

80,0%

Fuente: AEE

94,6%

CONCLUSIONS
The on-shore wind energy sector has become a world player due to the
extraordinary development of the Spanish market, spurred by ambitious
objectives, an efficient feed-in tariff system, and the involvement of the
main utilities.
The collaboration between the Spanish wind sector and the TSO has been
very fruitful and essential to guarantee the integration of wind energy in
the electrical system.
The development of wind energy in Spain has also been backed by an
important industrial capacity and now the main challenge is to maintain a
competitive position in the global market.
To achieve the 2020 NREAP objectives for wind power, a long-term policy
framework is necessary to attract investments in an increasingly
competitive European wind market.

31

THANK YOU FOR YOUR

ATTENTION!

Credit: AEE

Heikki Willstedt Mesa

hwillstedt@[Link]

Direccin Polticas Energticas

Asociacin Empresarial Elica

Recent Developments in the Levelized Cost of Energy from U.S. Wind Power
Projects

96

Recent Developments in the

Levelized Cost of Energy from
U.S. Wind Power Projects
Ryan Wiser, Lawrence Berkeley National Laboratory
Eric Lantz, National Renewable Energy Laboratory
Mark Bolinger, Lawrence Berkeley National Laboratory
Maureen Hand, National Renewable Energy Laboratory
February 2012
This analysis was funded by the Wind & Water Power Program, Office of
Energy Efficiency and Renewable Energy of the U.S. Department of
Energy under Contract No. DE-AC02-05CH11231
1

Background and Motivation

Some recent focus has been placed on the increasing capital cost
of wind projects observed from the early 2000s to 2010, and the
apparent flattening of fleet-wide capacity factors in recent years
These trends are important, but ignore other developments:
Continued improvements in capacity factors within individual wind resource
classes (if not fleet-wide) due to hub height and rotor diameter scaling
Significant recent improvements in low-wind-speed technology, resulting in
increased capacity factors and increased land area open to development
Steep reductions in turbine prices negotiated in last 2 years, and some
evidence of reductions in balance of plant costs expected (with some lag) to
lead to sizable near-term reductions in project-level capital costs
Possible longer-term (i.e., since early 2000s) reductions in the cost of operating
and maintaining as well as financing wind projects, as well as potentially
improved turbine/project availability
Incentive choice with respect to 30% ITC/1603 Treasury grant in lieu of PTC
2

97

Background and Motivation

Exclusive focus on historical capital cost and fleet-wide capacity
factor trends fails to convey recent improvements in the levelized
cost of energy (LCOE) from wind projects, the opening of new
lower-wind-speed areas for development, and the fundamental
interdependency of capital costs and capacity factors
Wind turbine manufacturers, wind project developers, and wind
power purchasers are not focused solely on optimizing capital costs
and capacity factors, individually, but are more interested in:
Levelized cost of wind energy across all wind resource classes
Lower LCOE lower power sales prices greater demand for wind energy

Amount of land area that might be reasonably developed

Transmission/siting/policy influences can constrain development in high-windresource regimes opening new lower-wind-resource regimes for development
can significantly increase potential development opportunities
3

Objectives of Present Work

(1) Develop consistent levelized cost of energy estimates for wind in
the U.S. in various wind resource regimes for:
Projects installed in 2002-2003: The low-point of wind capital costs
Projects installed in 2009-2010: The likely peak of wind capital costs
Projects to be installed in 2012-2013: When current turbine pricing is likely to
more-fully make its way into observed capital costs

Focus on direct costs, accounting primarily for capital cost trends and
trends in estimated capacity factors; conduct analysis with/without
PTC/MACRS; emphasize, only as an example, GE turbines

(2) Estimate the amount of available land area that would exceed
certain capacity factor and LCOE thresholds using the same
assumed technology, assumptions, and time periods as above
(3) Conduct two side-case analyses: (1) impact of incentive choice
between PTC and ITC/Section 1603; and (2) impact of possible
O&M, financing, and availability trends
4

98

Caveats
This is a preliminary assessment of the impact of various trends on
LCOE; the analysis does not consider all factors, and the results have
not undergone rigorous peer-review or been published
The analysis uses GE turbine technology only as an example to
facilitate assessment
This work only seeks to understand and estimate recent and near-term
developments
LCOE estimates for 2012-2013 are based on current turbine pricing,
but are nonetheless speculative
This work has not attempted to track developments over a longer
historical record or to forecast longer-term future trends
The present analysis is focused on the U.S., though the basic findings
should hold for other regions of the world as well
5

Presentation Outline

Recent
Trends

Analysis
Assumptions

99

Analytical
Results

Datasets Used To Explore Recent Capital

Cost and Performance Trends
Project- and
TurbineLevel Capital
Costs

488 projects built from 1983-2011, 34.6 GW

81 turbine transactions from 1997-2011, 23.9 GW

ProjectLevel
Performance
/ Capacity
Factors

338 projects built from 1983-2009, 32.0 GW

Data shown here are primarily from the U.S. DOEs 2010 Wind
Technologies Market Report (U.S. DOE 2011)
7

Installed Wind Project Capital Costs

Increased from Early 2000s through 2010
5,000

Installed Project Cost (2010 $/kW)

4,500

Individual Project Cost (488 projects totaling 34,616 MW)

Capacity-Weighted Average Project Cost
Polynomial Trend Line

4,000
3,500
3,000
2,500
2,000
1,500
1,000
500

1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011

Source: U.S. DOE 2011

Project costs bottomed out in 2001-2004; rose by $850/kW

on average through 2009; held steady in 2010 ($2,155/kW);
based on limited available data, may have dropped in 2011
8

100

Fleet-Wide Capacity Factors (CF) Have

(Generally) Increased Over Time
35%

Some
leveling off in
fleet-wide
capacity
factors in
recent years
is also
apparent

25%
20%
15%
10%
5%
0%

Year: 1999

Projects:

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

12

41

85

98

118

144

169

212

256

358

338

MW: 549

1,005 1,545 3,285 3,826 5,182 5,894 8,726 10,712 15,686 24,403 31,986

50%
2010 Capacity Factor (by project vintage)

BUT:

Capacity Factor

30%

45%
40%

Sample includes 338

projects totaling 32 GW

35%
30%
25%
20%
15%
10%

Capacity-Weighted Average 2010 Capacity Factor (by project vintage)

5%
0%

Source: U.S. DOE 2011

Individual Project 2010 Capacity Factor (by project vintage)

Pre-1998

1998-99

2000-01

2002-03

2004-05

2006

2007

2008

2009

9 projects

22 projects

25 projects

35 projects

31 projects

19 projects

34 projects

76 projects

87 projects

1,553 MW

1,906 MW

3,414 MW

1,591 MW

4,718 MW

8,409 MW

8,967 MW

647 MW

9
780 MW

Historical Trends Important, But Ignore

Other Notable Developments
Project
Performance

Sizable historical/continued increases in hub heights

and rotor swept area (in proportion to nameplate
capacity), leading to improvements in capacity factors
within individual wind resource classes, especially in
lower-wind-speed sites

Installed
Capital Costs

Steep reductions in wind turbine prices negotiated over

the last 2 years, with some evidence of a simultaneous
reduction in balance of plant costs, which are expected
to lead to sizable near-term capital cost reductions

Other Possible
Advancements

Operation and maintenance: Potential reduction in the

cost of O&M for newer wind power projects
Financing: Notwithstanding the impact of the financial
crisis, generally improved financing terms over the
longer-term as the sector and technology have matured
Availability: Potential reduction in losses due to
improved turbine/project availability

(since early 2000s)

See following slides for more details

10

101

Recent Moderation of CF Increase Driven In Part

By Move Towards Lower Wind Speed Sites
Projects increasingly sited in poorer wind regimes at 50m:
1998-2001: Class 4-5 common
2006-2009: Class 3-4 common
Trend likely driven by transmission/siting limitations and policy
influences, as well as improvement in low wind speed technology
Class
7

Wind Power Density at Project Level

900

Wind Power Density

(Average, W/m2 @ 50 m)

800
Class
6

700
600

Class
5

500

Class
4

400

Class
3

300
200
100

Class
2

Note:
Box plots identif y 10th/25th/75th/90th
percentile values
Percentages in labels represent % of
total in LBNL database

Project Level Capacity Weighted Mean

Class
1

Project Level Median

0
1998-99
48 projects
988 MW
(97%)

2000-01
43 projects
1725 MW
(98%)

2002-03
50 projects
2067 MW
(97%)

2004-05
54 projects
2611 MW
(93%)

2006
38 projects
2455 MW
(100%)

2007
53 projects
5250 MW
(100%)

2008
108 projects
8169 MW
(98%)

2009
88 projects
9479 MW
(95%)

Source: Lawrence Berkeley National Laboratory

11

and Growing Amounts of Wind Energy

Curtailment in Some Regions
Wind Curtailment, GWh
(percent potential generation)
Electricity Reliability Council of Texas

2007

2008

2009

2010

109
(1.2%)

1,417
(8.4%)
0
(0.0%)
2.5
(0.1%)
25.4
(0.8%)

3,872
(17.1%)
0
(0.0%)
19.0
(0.6%)
42.4
(1.2%)
250
(2.2%)

2,067
(7.7%)
0.9
(0.0%)
81.5
(2.2%)
42.6
(1.2%)
781
(4.4%)
4.6
(0.1%)
2,978
(5.1%)

Southwestern Public Service Co.

N/A

Public Service Company of Colorado

N/A

Northern States Power Co.

N/A

Midwest ISO, less NSP

N/A

N/A

Bonneville Power Administration

N/A

N/A

N/A

109
(1.2%)

1,445
(6.4%)

4,183
(10.4%)

Total Across These 6 Areas:

Source: U.S. DOE 2011

U.S. fleet-wide capacity factors from 2008-2010 would have

been 1-2 percentage points higher absent curtailment

(curtailment largely caused by inadequate transmission and minimum load)

12

102

Move to Lower Wind Speed Sites and Increased

Curtailment Hide the Very Real Increases in CFs
Witnessed in Individual Wind Resource Classes
80
70
60

Rotor Diameter

50

Hub Height

Historical increases in avg.

hub height and rotor diameter
have been substantial

40
30
20
10
0

COD:
Turbines:
Capacity:

1998-99

2000-01

2002-03

2004-05

2006

2007

2008

2009

2010

1,418

1,982

1,686

1,942

1,515

3,190

5,004

5,733

2,855

1,014 MW 1,756 MW 2,080 MW 2,779 MW 2,440 MW 5,249 MW 8,348 MW 9,993 MW 5,113 MW

Source: U.S. DOE 2011

Leading to sizable increases

in avg. capacity factors within
specific resource classes

50%
Capacity Factor (Weighted Average)

Average Rotor Diameter and

Hub Height (m)

90

45%
40%
35%
30%
25%
20%

2005-2007 Projects

15%

2002-2004 Projects

10%

1998-2001 Projects

5%
0%
1

4
5
6
Wind Resource Class @ 50m

Source: Lawrence Berkeley National Laboratory

13

Wind Turbine Prices Have Softened

Since Their Highs in 2008
Lag between
turbine prices
and project costs
should lead to
substantial
project-level
installed capital
cost reductions
by 2012-2013

Source: Bolinger and Wiser 2011

Turbine price quotes in 2011 for standard technology are reportedly as low
as $900/kW (Tier 1: ~$1,100-1,250/kW, with average at ~$1,100/kW); higher
costs typical for smaller orders, larger rotors/towers, etc.
(also more-favorable terms for buyers and improved technology; balance-ofplant costs also reportedly lower than in recent past)
14

103

Some Evidence of Lower O&M Costs,

Higher Availability, Improved Financing
O&M costs from
limited sample of US
wind power projects
(U.S. DOE 2011)

Strength of evidence
supporting specific
improvements in
O&M, availability, and
financing somewhat
lower than trends in
capacity factor and
capital cost

GEs reported median

availability by model year
(Mesh et al. 2011, GE
Wind Turbine Availability)

15

Recent Chatter Suggests that, As a Result of

These Trends, Delivered Wind Energy Costs
Have Declined Substantially
BNEF 2011: "The cost of wind generation has been driven to record lows by declines in turbine
prices and the cash grant, which eliminates the cost of securing tax equity financing.
Austin Energy officials say those wind contracts are among the cheapest deals available, when
the cost of building power plants is taken into account, and comparable to what the historically
volatile natural gas market has been offering recently. ([Link] article)
Our contract with NextEra Energy Resources is one of the lowest weve ever seen and results
in a savings of nearly 40 percent for our customers, said David Eves, president and CEO of
Public Service Company of Colorado. The addition of this 200-megawatt wind farm
demonstrates that renewable energy can compete on an economic basis with more traditional
forms of generation fuel, like natural gas, and allows us to meet the states Renewable Energy
Standard at a very reasonable cost to our customers. (Reuters article)
Consumers Energy, Michigan: "Lower wind power costs mean $54m saving for Consumers
Energy. (newspaper article)
Westar, Kansas: Signed more wind contract than needed "...because pricing is so attractive now
and to minimize tax risk to our investors" (Westar Q4 earnings call)

16

104

Analysis Objective
Estimate Capacity Factors, LCOE, and Developable
Land Area By Selecting Representative Assumptions
for Turbines Used in U.S. Wind Projects
Installed in
2002-2003

The low-point of wind project capital costs

Installed in
2009-2010

The likely peak of wind project capital costs

(To be)
Installed in
~2012-13

When current turbine pricing is likely to morefully make its way into observed capital costs

18

Basic Approach
In each period, account for common actual/expected trends in:
(1) installed capital costs (based on actual/estimated cost)
(2) capacity factors in different wind resource classes (estimated based on
available power curves, assuming sea level air density of 1.225 kg/m 3)

For simplicity, in most cases hold constant: O&M costs; financing

structure and costs; project/turbine availability and other losses;
project life, income taxes, decommissioning, etc.
Estimate LCOE: (1) without PTC or 5-yr MACRS (depreciation
assumed to be 12-yr straight line); and (2) with PTC and 5-yr MACRS
Conduct two side analyses: (1) impact of incentive choice between
PTC and ITC/Treasury Grant; and (2) impact of possible O&M,
financing, and availability trends
Use IEA Task 26 Work-package 1 LCOE model
19

105

For Analytical Simplicity, Focus on

GE Turbine Evolution
Typical GE Turbine Offerings Over Time
(does not include 2.5-MW turbine platform)

Credit: Al Hicks, NREL

GE has been the dominant supplier of turbines to the U.S. market over this
timeframe, ensuring that a focus on GE as an example of the evolution of
cost, performance, and LCOE trends is appropriate
20

Summary of Core Input Assumptions

2002-2003

2009-2010

Technology type

Characteristics

Standard

Standard

Standard*

Low Wind*

Low Wind*

Nameplate capacity

1.5 MW

1.5 MW

1.62 MW

1.62 MW

1.62 MW

65 m

80 m

80 m

80 m

100 m

Hub height (HH)

Current Turbine Pricing: ~2012-2013

Rotor diameter (RD)

70.5 m

77 m

82.5 m

100 m

100 m

Installed capital cost

$1,300/kW

$2,150/kW

$1,600/kW

$1,850/kW

$2,025/kW

Operating costs

$60/kW-yr

$60/kW-yr

$60/kW-yr

$60/kW-yr

$60/kW-yr

Losses (availability,
array, other)

15%

15%

15%

15%

15%

Financing (nominal)

9%

9%

9%

9%

9%

Dollar values are all real 2010$

Financing cost / discount rate reported in nominal terms
Air density = 1.225 kg/m3 (sea level wind speed)
Weibul K Factor = 2 in all scenarios
1/7th power law scaling to estimate hub height wind speed
20-year assumed project/economic life in all scenarios
Aggregate income taxes assumed to equal 38.9%

All assumptions intended to reflect

representative actual/common
conditions; installed costs, operating
costs, losses and other assumptions
can vary considerably from one
project to the next

*These turbines are assumed viable in sites up to the respective IEC Class II and Class III reference average annual
wind speed. Depending on site specific gust, turbulence, and air densities these turbines in actuality may be
21 annual wind speeds than applied in this analysis.
reasonably placed in sites with higher average

106

Basis For Core Assumptions

Turbine Characteristics (capacity, hub height, rotor diameter)
Common GE designs for relevant installation years

Installed Capital Cost

2002-2003 and 2009-2010: based on actual average costs for installed projects
2012-2013: assumes $550/kW drop in turbine/BOP costs since high-point for 80 m HH /
82.5 m RD turbine based on earlier data on turbine cost trends, BNEF (2011), and
discussions with wind developers/manufacturers; $250/kW assumed increase for 100 m RD
upgrade and additional $175/kW increase for 100 m HH upgrade based on discussions with
wind developers/manufacturers; result is an estimate of the average installed cost of
projects based on current turbine orders; actual project costs will have a large spread
around the average, with both lower- and higher-cost projects anticipated

Other Costs and Characteristics

Operating costs: Assumed to be exclusively denominated in $/kW-yr, though, in practice,

some costs may be more-appropriately denominated in $/MWh; all-in operating costs
assumed to equal $60/kW-yr in core analysis based on review of available data/literature
Losses: 15% losses assumed in core analysis based on review of available data/literature,
as well as matching actual and estimated project-level capacity factors
Financing: 9% assumed in core analysis based on review of available data/literature
22

Summary of Side-Case Input

Assumptions
Incentive
Choice

Assumes that project owners choose between the PTC

and ITC/Treasury Grant based only on the face-value
of those incentives (i.e., the incentive that minimizes
the LCOE with no additional changes and ignoring any
ancillary benefits of the ITC/Treasury Grant)

O&M,
Financing,
Availability

O&M: Cost reduction from $64/kW-yr in 2002-2003

$57/kW-yr in 2012-2013 based on review of available
data/literature, including 2010 Wind Technology
Market Report, BNEF (2011), etc.
Financing: All-in cost of finance decreases as
technology matures from 10.5% in 2002-2003 9%
in 2012-2013 based on review of available
data/literature and discussions with wind developers
Availability: Reduction in average project-level
losses from 16.5% in 2002-2003 14% in 20122013 based only on assumed improvement in
availability, itself based on review of limited available
data/literature and discussions with wind consultants
and developers
23

107

Presentation Outline

Recent
Trends

Analysis
Assumptions

Analytical
Results

24

Outline of Key Results

Estimated Capacity Factors in Varying Resource Classes
Levelized Cost of Energy in Varying Resource Classes
Side Analysis: Impact of Improvements in O&M, Financing, and Availability
Designing Turbines for Low Wind-Speed Sites: Narrowing the Gap in LCOE
Side Analysis: Impact of ITC/Treasury Grant Option In Lieu of PTC
Increased Land Area Exceeding Capacity Factor Thresholds
Increased Land Area Exceeding LCOE Thresholds
25

108

Turbine Design Advancement Leads To

Enormous Increase in Capacity Factors
Estimated capacity factor improvement driven by larger rotor swept area in
proportion to nameplate capacity, as well as higher hub heights; increase is
especially apparent with newest batch of low-wind-speed turbines
50%
Current: 2012-13
Low Wind-speed (100m Tower)

45%

Current: 2012-13
Standard Technology

Current: 2012-13
Low Wind-speed (80m Tower)

2009-10
Standard Technology

Capacity Factor

40%

2002-03
Standard Technology

35%
30%

Note: Low wind-speed turbine performance is modeled up

to a 7.5 m/s hub height wind speed, per IEC standards;
site specific conditions may allow these machines to be
installed in wind regimes that exceed the IEC average
annual wind speed, leading to higher capacity factors (at
higher average 50m wind speeds) than shown here.

25%
20%
15%
5.5

6.0

6.5

Class 2

7.0
Class 3

7.5
Class 4

8.0
Class 5

8.5
Class 6

50m26
Wind-speed (m/s)
air density = 1.225 kg/m3

Levelized Cost of Wind Energy Is Estimated

To Be at An All-Time Low for 2012-2013
Accounting only for assumed capacity factor and capital cost trends, LCOE increased
substantially from 2002-2003 to 2009-2010 because capital cost increases were not
fully compensated by CF improvements; estimated LCOE has since dropped below its
earlier lows within individual wind resource classes because capital cost increases from
2002-2003 to the present have been more-than offset by CF improvements, yielding a
lower LCOE than at any time in the past within individual resource classes

Without Federal Incentives

With Federal PTC/MACRS

$120

$120

Mid-Class 5

$80
$60
$40
$20

Levelized Cost of Energy ($/MWh)

Includes Federal PTC & MACRS

Levelized Cost of Energy ($/MWh)

No Incentives

Mid-Class 3
$100

$0

$100
$80

Mid-Class 3

$60

Mid-Class 5

$40
$20
$0

2002-03:
Standard Tech.

2009-10:
Standard Tech.

Current, 2012-13
Tech. Choice

2002-03:
Standard Tech.

2009-10:
Standard Tech.

Current, 2012-13
Tech. Choice

Note: Tech. Choice assumes that IEC Class III machines are only available for sites up to 7.5 m/s average
wind speed at hub height (sea level air density)
27

109

Levelized Cost of Energy in Varying

Resource Classes (without PTC/MACRS)
Based on current pricing and assumptions: 100m rotor diameter is found to be
economically attractive in comparison to 2012-2013 Standard Technology where it
can be deployed; a wind sheer higher than 1/7th is found to be needed for the
100m tower to be least cost in comparison to the 80m option (with the 100m rotor)
Levelized Cost of Energy ($/MWh)
No Incentives

$160
$140

$120

2009-10: Standard
Technology

$100
$80
$60

2002-03: Standard
Technology

Current, 2012-13:
Low Wind-speed
(100m Tower)

$40

Current, 2012-13:
Standard
Technology

Current, 2012-13:
Low Wind-speed
(80m Tower)

$20
$0
5.5

6.0
Class 2

6.5

7.0

7.5

Class 3
Class 4
50m Wind-speed (m/s)

8.0
Class 5

8.5
Class 6

(air density = 1.225 kg/m3 )

28

Levelized Cost of Energy in Varying

Resource Classes (with PTC/MACRS)
With
PTC/MACRS
and
with02-03
currenttoturbine
andsince
otherdropped
specific assumptions,
LCOE
increased
from
09-10,pricing
but has
below its
the highest
wind speed
evaluated
belowcost
can from
support
LCOEs
as low
as
earlier
lows despite
the sites
increase
in capital
02-03
to the
present
~$33/MWh (real$), while the lower wind speed sites have LCOEs of ~$65/MWh
Levelized Cost of Energy ($/MWh)
Includes Federal PTC & MACRS

$120
1

$100

$80

2009-10: Standard
Technology

$60
$40

2002-03: Standard
Technology

Current, 2012-13
Low Wind-speed
(100m Tower)

$20

Current, 2012-13:
Low Wind-speed
(80m Tower)

Current, 2012-13:
Standard Technology

$0
5.5

6.0
Class 2

6.5

7.0

Class 3
Class 4
50m Wind-speed (m/s)
air density = 1.225kg/m3

29

110

8.0

7.5
Class 5

8.5
Class 6

All Else Being Equal, Capital Cost Increases

Would Have Led To Much Higher LCOEs
(Mid Class 3 Wind Resource, with PTC/MACRS)
Capital Cost Impacts on LCOE
Assumes Fixed Performance circa 2002-03 and Class 3 Wind Resource

$100

Levelized Cost of Energy ($/MWh)

Includes Federal PTC & MACRS

$90
$80
$70
$60
$50
$40
$30
$20
$10

Current Pricing: Variable Turbine Designs

$0
2002-03:
Standard Tech.

2009-10:
Standard Tech.

2012-13:
Standard Tech.

2012-13: Low Wind

(80m Tower)

2012-13: Low Wind

(100m Tower)

30

All Else Being Equal, Performance Improvements

Would Have Led To Much Lower LCOEs
(Mid Class 3 Wind Resource, with PTC/MACRS)
Performance Impacts on LCOE from Advanced Turbine Designs
Assumes Fixed Capital Cost circa 2002-03 and Class 3 Wind Resource

Levelized Cost of Energy ($/MWh)

Includes Federal PTC & MACRS

$60
$50
$40
$30
$20
$10

Current Pricing: Variable Turbine Designs

$0
2002-03:
Standard Tech.

2009-10:
Standard Tech.

2012-13:
Standard Tech.

31

111

2012-13: Low Wind

(80m Tower)

2012-13: Low Wind

(100m Tower)

Same Tradeoff Between Capital Cost and

Performance Exists in Other Wind Resource Classes
(Mid Class 5 Wind Resource, with PTC/MACRS)
Capital Cost Impacts on LCOE

Performance Impacts on LCOE from Advanced

Turbine Designs
Assumes Fixed Capital Cost circa 2002-03 and Class 5 Wind
Resource

$70
Levelized Cost of Energy ($/MWh)
Includes Federal PTC & MACRS

Levelized Cost of Energy ($/MWh)

Includes Federal PTC & MACRS

Assumes Fixed Performance circa 2002-03 and Class 5 Wind

Resource

$60
$50
$40
$30
$20
$10
$0
2002-03:
Standard Tech.

2009-10:
Standard Tech.

$40
$35
$30
$25
$20

$15
$10
$5
$0
2002-03:
Standard Tech.

2012-13:
Standard Tech.

2009-10:
Standard Tech.

2012-13:
Standard Tech.

32

Same Tradeoff Between Capital Cost and

Performance Exists without PTC/MACRS
Class 3

Capital Cost Impacts on LCOE

Assumes Fixed Performance circa 2002-03 and Class 3 Wind Resource

Performance Impacts on LCOE from Advanced Turbine Designs

Assumes Fixed Capital Cost circa 2002-03 and Class 3 Wind Resource

$90
$80
Levelized Cost of Energy ($/MWh)
No Incentives

$120
$100
$80
$60
$40
$20

Current Pricing: Variable Turbine Designs

$70
$60
$50
$40
$30
$20
$10

Current Pricing: Variable Turbine Designs

$0

$0
2002-03:
Standard Tech.

2009-10:
Standard Tech.

2012-13:
Standard Tech.

2012-13: Low Wind

(80m Tower)

2002-03:
Standard Tech.

2012-13: Low Wind

(100m Tower)

Capital Cost Impacts on LCOE

Assumes Fixed Performance circa 2002-03 and Class 5 Wind
Resource

Class 5

$100
$90
$80
$70
$60
$50
$40
$30
$20
$10
$0

$70

2009-10:
Standard Tech.

2012-13:
Standard Tech.

2012-13: Low Wind 2012-13: Low Wind

(80m Tower)
(100m Tower)

Performance Impacts on LCOE from Advanced

Turbine Designs
Assumes Fixed Capital Cost circa 2002-03 and Class 5 Wind
Resource

Levelized Cost of Energy ($/MWh)

No Incentives

Levelized Cost of Energy ($/MWh)

No Incentives

Levelized Cost of Energy ($/MWh)

No Incentives

$140

$60
$50
$40
$30
$20
$10
$0

2002-03:
Standard Tech.

2009-10:
Standard Tech.

2012-13:
Standard Tech.

2002-03:
Standard Tech.

33

112

2009-10:
Standard Tech.

2012-13:
Standard Tech.

Side Analysis: Impact of Improvements in O&M,

Financing, and Availability (with PTC/MACRS)
Assumed improvements in O&M costs, financing rates, and availability lead to
substantial additional estimated LCOE reductions from 2002-2003 to 2012-2013 in
comparison to core analysis that only varies capital cost and capacity factor

Core Assumptions

Side Analysis Assumptions

only varies capital cost and capacity factor

also varies O&M, availability, and financing

$120

Levelized Cost of Energy ($/MWh)

Includes Federal PTC & MACRS

Levelized Cost of Energy ($/MWh)

Includes Federal PTC & MACRS

$120
$100
$80

26% Cost Reduction

6 m/s

$60
7 m/s

$40

8 m/s
5% Cost Reduction

$20
$0

$100
$80

39% Cost Reduction

6 m/s

$60
7 m/s

$40

8 m/s
24% Cost Reduction

$20
$0

2002-03

Current, 2012-13

2002-03

Current, 2012-13

Standard Technology

Technology Choice

Standard Technology

Technology Choice

Note: Technology Choice assumes that IEC Class III machines are only available for sites up to 7.5 m/s
average wind speed at hub height (sea level air density)
34

Side Analysis: Impact of Improvements in O&M,

Financing, and Availability (without PTC/MACRS)
Assumed improvements in O&M costs, financing rates, and availability lead to
substantial additional estimated LCOE reductions from 2002-2003 to 2012-2013 in
comparison to core analysis that only varies capital cost and capacity factor

Core Assumptions

only varies capital cost and capacity factor

Side Analysis Assumptions

also varies O&M, availability, and financing

$140

Levelized Cost of Energy ($/MWh)

No Incentives

Levelized Cost of Energy ($/MWh)

No Incentives

$120
20% Cost Reduction
6 m/s

$100
$80

7 m/s

$60

8 m/s
3% Cost Reduction

$40
$20

$120
31% Cost Reduction
6 m/s

$100
$80

7 m/s

$60

8 m/s
17% Cost Reduction

$40
$20
$0

$0
2002-03

Current, 2012-13

2002-03

Current, 2012-13

Standard Technology

Technology Choice

Standard Technology

Technology Choice

Note: Technology Choice assumes that IEC Class III machines are only available for sites up to 7.5 m/s
average wind speed at hub height (sea level air density)
35

113

Designing Turbines for Low Wind-Speed

Sites: Narrowing the Gap in LCOE
The proliferation of turbines designed for lower wind speeds has narrowed the
gap between the LCOE of high- and low- wind speed sites, increasing the
economic attractiveness of developing wind projects in lower wind speed areas
Without Federal Incentives

With Federal PTC/MACRS

$100

$100

$80

$90

6 m/s
$32/MWh

$70

7 m/s

$60

8 m/s

$25/MWh

$50
$40
$30
$20

Levelized Cost of Energy ($/MWh)

Includes Federal PTC & MACRS

Levelized Cost of Energy ($/MWh)

No Incentives

$90

$10

$80
$70
$60
$50

6 m/s
$31/MWh
$23/MWh

7 m/s

$40

8 m/s

$30
$20
$10

$0

$0
Current, 2012-13:
Standard Technology

Current, 2012-13:
Tech. Choice Including Low Wind-speed

Current, 2012-13:
Standard Technology

Current, 2012-13:
Tech. Choice Including Low Wind-speed

Notes: Does not consider Treasury Grant program / 30% ITC (see later results); Tech. Choice assumes that IEC Class III
machines are only available for sites up to 7.5 m/s average wind speed at hub height (sea level air density)
36

Side Analysis: Incentive Choice Has Also Impacted

the Economics of Low Wind Speed Sites
30% ITC/Grant applied to capital cost is relatively more lucrative than PTC on facevalue basis when a project has a low capacity factor and high costs see results
for 2009-2010 Standard Technology in chart on left
Current turbines / pricing have higher assumed capacity factors and lower costs,
making PTC the better choice in virtually all developable wind regimes on face
value basis see results for 2012-2013 turbines in chart on right
$70

$120

$100

Levelized Cost of Energy ($/MWh)

Levelized Cost of Energy ($/MWh)

$110
2009-10:
Standard
Technology (PTC)

$90
$80

2009-10:
Standard
Technology
(ITC/Cash Grant)

$70
$60
$50
$40

$65
2012-13:
Technology
Choice
(ITC/Cash Grant)

$60
$55
$50
2012-13:
Technology
Choice
(PTC)

$45
$40
$35
$30

$30
5.5

6.0
Class 2

6.5

7.0

7.5

Class 3
Class 4
Class 5
50m Wind-speed (m/s)

8.0

8.5
Class 6

5.5

6.0
Class 2

6.5

7.0

7.5

Class 3
Class 4
Class 5
50m Wind-speed (m/s)

8.0

8.5
Class 6

air density = 1.225 kg/m3

air density = 1.225 kg/m3

Note: Results ignore benefits of ITC/Treasury grant beyond direct face value;
Technology Choice assumes that IEC Class III machines are only available for sites up to 7.5 m/s average wind
speed at hub height (sea level air density)
37

114

Result Was Narrowing in LCOE Between High &

Low Wind Sites in 2009-2010, Not in 2012-2013
ITC/Grant improved the economics of low wind speed sites in 2009-2010;
estimated lower capital costs and improved capacity factors result in face
value of PTC > ITC/Grant in 2012-2013 even in low wind speed sites

(ancillary benefits of ITC/Grant may still outweigh loss in face value in such sites)
$100

$100
6 m/s

$90

$80
$70

$45/MWh

$38/MWh

7 m/s

$60
8 m/s

$50
$40
$30
$20

Levelized Cost of Energy ($/MWh)

Includes Federal Incentives

Levelized Cost of Energy ($/MWh)

Includes Federal Incentives

$90

$80
$70
$60
6 m/s

$50
$40

$23/MWh

$23/MWh

7 m/s
8 m/s

$30
$20
$10

$10

$0

$0
2009-10:
Standard Technology
PTC & MACRS

2012-13:
Tech. Choice Incl. Low Wind-speed
PTC & MACRS

2009-10:
Standard Technology
Incentive Choice

2012-13:
Tech. Choice Incl. Low Wind-speed
Incentive Choice

Note: Tech. Choice assumes that IEC Class III machines are only available for sites up to 7.5 m/s average
wind speed at hub height (sea level air density)
38

Land Area Exceeding Capacity Factor

Thresholds Has Increased Dramatically
Percentage Increase Since 2002-2003:
> 30% CF: 80% 140%
> 35% CF: 130% 270%
> 40% CF: 600% 1400%

3,000,000

Available Land Area (km2)

2,500,000

Current, 2012-13:
Low Wind-speed (100m Tower)
[+ other 2012-13 Turbines]

2,000,000

Current, 2012-13:
Low Wind-speed (80m Tower)
[+ 2012-13 Standard Tech]

1,500,000
Current, 2012-13: Standard
Technology

1,000,000

2009-10: Standard
Technology

500,000

2002-03: Standard
Technology

0
28%

30%

32%

34%

36%
38%
Capacity Factor

40%

42%

44%

Notes: Wind speed data come from the 50 m long-term assessments produced by AWS Truepower, MN Dept of Commerce, Iowa State Energy Center,
Alternative Energy Institute (Texas), and NREL. Alabama, Louisiana, Mississippi, and Florida were not covered by any of these datasets. Standard wind
resource exclusions were applied, as documented on the Wind Powering America website. Low wind-speed turbines are assumed to be utilized in sites
up to 7.5 m/s sea level equivalent average annual wind-speed, per IEC standards. Site specific conditions may allow these machines to be placed in
higher average annual wind-speed sites, which would further increase the percentage increase in available land area beyond what is estimated here.

39

115

Available Land Area (km2)

Land Area Exceeding LCOE Thresholds

Has Also Increased Substantially
3,000,000

2012-13:
Lowest Cost
Tech Choice

2,500,000

2002-03:
Standard Technology

2,000,000
1,500,000

Percentage Increase Since 2002-2003:

< $35/MWh w/ PTC & MACRS: 42%
< $55/MWh w/ PTC & MACRS: 48%
< $75/MWh w/ PTC & MACRS: 25%

1,000,000
500,000
0

w/: PTC/MACRS
w/o: PTC/MACRS

$30

$40

$50

$60

$70

$80

$90

$57

$67

$77

$89

$99

$110

$121

Levelized Cost of Energy ($/MWh)

Notes: Increase in land area meeting a LCOE threshold is lower than the increase from
a CF threshold because increased capital cost trends impact LCOE estimates, but not
CF; Tech. Choice assumes that IEC Class III machines are only available for sites up
to 7.5 m/s average wind speed at hub height (sea level air density)
40

Conclusions
Economic attractiveness of wind projects in recent past was reduced due
to increased capital cost, move toward lower wind speed sites, and lower
electricity prices
Examination of historical trends in capital costs and capacity factors,
individually, gives an incomplete picture of technology advancement as
well as historical & current developments
Recent declines in turbine prices & improved technology have reduced
the estimated LCOE of wind; LCOE for projects being planned today in
fixed resource areas is estimated to be at an all-time low
Considering plausible assumptions for not only capital cost and capacity
factor, but also O&M, financing & availability, the LCOE for 2012-2013
projects is estimated to be as much as ~24% and ~39% lower than
the previous low in 2002-2003 in 8 m/s and 6 m/s (at 50 m) resource
areas, respectively (with the PTC/MACRS); when only considering
capital cost and capacity factor, the reduction is ~5% and ~26%
41

116

Conclusions
Technology advancement for lower wind speeds has narrowed the gap in
LCOE between lower and higher wind speed sites; choice of 30%
ITC/Treasury Grant may have further encouraged development in lower wind
speed sites, especially in 2009-2010
The amount of land area meeting or exceeding certain capacity factor and
LCOE thresholds has substantially increased as a result of these technology
improvements helps alleviate to a degree transmission and siting barriers
Technology advancement & learning still applies to onshore wind, despite its
relative maturity, but all modes of technical advancement must be
considered rather than emphasizing individual parameters
Despite these recent and impressive technological advancements, three
counter-veiling factors may intervene to raise LCOE:
potential for increased pricing if demand for wind turbines begins to catch up with supply, or
if other exogenous influences are triggered (e.g., higher commodities and/or labor costs)
potential continued move towards lower wind speed sites as a result of severe
transmission/siting limitations
potential near-term loss of federal PTC/ITC/Treasury Grant
42

Unpacking the Drivers Behind Recent U.S. Wind Project Installed Cost and Performance Trends

Unpacking the Drivers Behind

Recent U.S. Wind Project Installed
Cost and Performance Trends
Ryan H. Wiser

Lawrence Berkeley National Laboratory

IEA Cost of Wind Energy Workshop

Denver, Colorado
March 29, 2011
This analysis was funded by the U.S. Department of Energys (DOE) Office of Energy
Efficiency and Renewable Energy, Wind and Hydropower Technologies Program
Environmental Energy Technologies Division Energy Analysis Department

117

Understanding Trends in
Wind Project Costs
This portion of the presentation has been
published. Please see the final report and
presentation at:
[Link]
2

Electricity Markets and Policy Group Energy Analysis Department

Understanding Trends in
Wind Project Performance
- Preliminary Analysis -

Electricity Markets and Policy Group Energy Analysis Department

118

Wind Project Capacity Factors (CF) in the

United States Vary Substantially By Project
Focusing on 14,722 megawatts (MW) installed from 19982007, capacityweighted fleet-average CFs from 19992008 equaled 33.7%
Installed Capacity (MW)

5,000
4,500
4,000
3,500
3,000

14,722 MW
equals 95%
of all wind
capacity
added over
this time
frame

2,500
2,000
1,500
1,000
500
0
<15% 15-20% 20-24% 25-29% 30-34% 35-39% 40-44% 45-49% 50-54%

Capacity Factor (Project Average)

Electricity Markets and Policy Group Energy Analysis Department

Average Capacity Factors Have Increased

Over Time, by Project Vintage
2% of projects installed from 19982003 had average CFs > 40%
25% of projects installed from 20042007 had average CFs > 40%
Installed Capacity (MW)

3,500
3,000

1998-2003
2004-2007

2,500
2,000
1,500
1,000
500
0

<15% 15-20% 20-24% 25-29% 30-34% 35-39% 40-44% 45-49% 50-54%

Capacity Factor (Project Average)

5

Electricity Markets and Policy Group Energy Analysis Department

119

Increase In Capacity Factors Is Widespread

(Not Limited to Certain Regions of the Country)
30%
40%
30%
20%
10%
0%

20%
10%
0%
98-03

40%
30%
20%
10%
0%
98-03

04-07

40%
30%
20%
10%
0%

40%
30%
20%
10%
0%

04-07

40%
30%
20%
10%
0%

98-03

30%
98-03

98-03

04-07

04-07

20%
10%
0%
98-03

98-03

04-07

04-07

Note: Years represent

project vintage (i.e., COD);
capacity factors reflect
multi-year average

40%
30%
20%
10%
0%

40%
30%
20%
10%
0%

n/a
98-03

98-03

04-07

Northwest
California
Mountain
Texas
Heartland
Great Lakes
East
New England
Southeast

04-07

04-07

Electricity Markets and Policy Group Energy Analysis Department

Average CFs Have Improved with Project

Vintage, but Flattened from 20042007
50%

Capacity Factor (Average)

45%
40%
35%
30%
25%
20%
15%
10%
Note: Box plots identif y

5%

Capacity Factor (Weighted Average)

0%
1998-99
32 projects
935 MW

2000-01
33 Projects
1,446 MW

2002-03
34 Projects
2,227 MW

2004-05
37 Projects
2,680 MW

10th/25th/75th/90th
percentile CF values

2006
31 Projects
2,394 MW

Electricity Markets and Policy Group Energy Analysis Department

120

2007
42 Projects
5,001 MW

Possible Drivers of Project Performance

Project siting: Projects located in stronger wind regimes, as well as
better micro-siting, may increase capacity factors
Higher hub heights: Given wind shear, taller towers will access
better wind speeds
Larger rotor diameters (relative to rated capacity): Longer
blades will increase energy capture
More-efficient turbine designs: Evolutionary gains in turbine and
blade design/electronics may increase energy capture
Improved availability: Possible improvement in project-level
availabilities over time
Annual wind resource variation: Wind resource quality can vary
from one year to the next, nationally and regionally.
Output curtailment: Physical or financial curtailment of output due
to transmission limitations and minimum generation conditions
8

Electricity Markets and Policy Group Energy Analysis Department

Objective: Evaluate a Subset of These

Drivers to Better Understand Overall Trends
For now, focus only on projects built from 19982007 (older
projects not representative, and have data issues)
Determine turbine coordinates, hub height, and rotor diameter
using Federal Aviation Administration (FAA) database and other
sources
Correlate turbine/project coordinates to respective average
estimated wind resource using available wind maps
Collect and organize project performance data from the Federal
Energy Regulatory Commission (FERC), Energy Information
Administration (EIA), and other sources for data years 19992008
Document trends in project performance nationally, by region, by
turbine type, by wind resource class, etc.
Analyze resulting data (i.e., average wind resource, hub height,
and rotor diameter) to discern relative impact on capacity factors
9

Electricity Markets and Policy Group Energy Analysis Department

121

Current Data Sample: Projects Installed from

19982007, with Turbines >200 kilowatts (kW)
Project Sample Description
Total sample of possible projects
Projects for which we have rotor diameter,
hub height, and location/resource
Projects for which we have above data
and project-level capacity factor (for at
least one year)

Nameplate
Capacity
15,537 megawatts
(MW)

15,268 MW
14,722 MW
95% of possible
sample

Rotor diameter, hub height, and location/modeled resource data

have also been collected for 8,376 MW built in 2008; currently
collecting data for 20092010
10

Electricity Markets and Policy Group Energy Analysis Department

Results: CFs Increased Despite Significant Decrease

in Average Wind Resource at 50 Meters (m)

High Class 5
in 19981999
High Class 3
in 2008

900

Note: Box plots identify

10th/25th/75th/90th
percentile values

800

Wind Power Density

(Average, W/m2 @ 50 m)

Projects
increasingly
sited in
poorer wind
regimes at
50 m:

Class
6

700
600

Class
5

500

Class
4

400

Class
3

300

Class
2

200
100
0

Wind Power Density (Weighted Average)

1998-99
2000-01
2002-03
2004-05
2006
2007
2008
32 projects 33 Projects 34 Projects 37 Projects 31 Projects 42 Projects 110 Projects
935 MW 1,446 MW 2,227 MW 2,680 MW 2,394 MW 5,001 MW 8,376 MW

11

Class
7

Electricity Markets and Policy Group Energy Analysis Department

122

Class
1

This Must Be a Story of Technical Change: Hub

Height, Rotor Diameter, Efficiency, Etc.
Capacity factors have increased with project vintage within
fixed estimated wind resource regimes at 50 m
Capacity Factor (Weighted Average)

50%
45%
40%
35%
30%
25%
20%

2005-2007 Projects

15%

2002-2004 Projects

10%

1998-2001 Projects

5%
0%
1

4
5
6
Wind Resource Class @ 50m

12

Average Hub Heights (HH) Have

Experienced a Dramatic Increase
90

Turbine Hub Height (Average, m)

80
70
60
50
40
30
20
10

Hub Height (weighted average)

0
1998-99
32 projects
935 MW

13

2000-01
33 Projects
1,446 MW

2002-03
34 Projects
2,227 MW

2004-05
37 Projects
2,680 MW

2006
31 Projects
2,394 MW

Note: Box plots identif y

10th/25th/75th/90th
percentile project hub
height values

2007
42 Projects
5,001 MW

2008
110 Projects
8,376 MW

Electricity Markets and Policy Group Energy Analysis Department

123

ButAssuming 1/7th Power Law, Average

Wind Resource at HH Has Still Declined
Wind Power Density (Average, W/m2)

900
800
700
600
500
400
300
200
Wind Power Density @ HH (weighted average)

100

Wind Power Density @ 50 m (weighted average)

0
1998-99
32 projects
935 MW

2000-01
33 Projects
1,446 MW

2002-03
34 Projects
2,227 MW

2004-05
37 Projects
2,680 MW

2006
31 Projects
2,394 MW

2007
2008
42 Projects 110 Projects
5,001 MW
8,376 MW

Electricity Markets and Policy Group Energy Analysis Department

14

Specific Power (Average, W/m2)

15

100
90
80
70
60
50
40
30
20
10
0
500

Rotor Diameter (weighted average)

Turbine Capacity (weighted average)
98-99 00-01 02-03 04-05

06

07

2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0

08

Turbine Capacity (Average, MW)

Turbine Rotor Diameter (Average, m)

Rotor Diameter Has Increased; Specific

Power (W/m2 Swept Area) Has Decreased
Larger rotors optimized
for lower wind speeds
Rotor swept area has
increased more rapidly
than turbine capacity,
leading to reductions in
average specific power
(W/m2)

400

Turbines with lower

specific power will tend
to have higher CFs (not
greater kWh/m2)

300
200
100
0

Specific Power (weighted average)

98-99 00-01 02-03 04-05

06

07

08

124

More Efficient Turbine and Blade Designs,

Electronics, and Project Availabilities
In combination, the following factors (among
others) may have contributed to an increase in
CFs over time:
Variable speed technology
Improved blade design
More aerodynamically efficient airfoils
Lower soiling losses

16

Better control systems and algorithms

Better understanding of wake effects
Drive-train efficiency improvements
Higher project-level availabilities
Electricity Markets and Policy Group Energy Analysis Department

Curtailment Needs to Be Considered As

One Factor That Might Drive CFs Lower
Curtailment has been greatest in Texas, and increased
substantially in 2009. Our performance data extends through
2008, so curtailment is less of an issue in the present analysis.
3,000,000
2,500,000

local curtailment
wind production
8% of Potential Wind
Generation Curtailed

2,000,000
1,500,000

17% of Potential Wind

Generation Curtailed

zonal curtailment

1% of Potential Wind
Generation Curtailed

1,000,000
500,000
0
1 2 3 4 5 6 7 8 9 101112 1 2 3 4 5 6 7 8 9 101112 1 2 3 4 5 6 7 8 9 101112

17

2007

2008

125

2009

Regression Analysis Helps To

Piece Together the Story
Regression analysis shows relative influence of: (1) wind power density
at 50 m; (2) hub height; (3) specific power; (4) installation year (COD) to
account for efficiency/availability; (4) Texas/state dummy variables to
account for curtailment, errors in state wind maps, etc.
HH 3 7
1
ln (CFavg ) = 0 + 1 ln (WPD ) + 2 ln
+ 4 COD + i + 4 Statei +
+ 3 ln
50
SP

ln(WPD )

ln HH 7
50

18

ln 1
SP

COD

Texas

Beta

0.527

0.989

0.584

0.014

0.075

Stand.
Error

0.066

0.324

0.194

0.008

0.033

P-value

0.000

0.003

0.003

0.153

0.023

Electricity Markets and Policy Group Energy Analysis Department

Decoding the Influence of the Variables:

Graphic Presentation of Regression Results
Predicted Absolute Change in CF
from 1998 to 2007 (modeled)

6%
4%
2%
0%
-2%
-4%
-6%
Hub Height

Specific
Power

Installation Wind Density

Year

ALL

Actual data show absolute increase in CF of 5.8% from

19981999 to 2007 projects; regression analysis estimates 4.6%
19

126