Electricity Markets and Renewable Generation

2014
Energy Risk Professional

ERP Exam Course Pack
READINGS THAT ARE FREELY AVAILABLE ON THE GARP WEBSITE

In addition to the published readings listed, the Electricity Markets and Renewable Generation
section of the 2014 ERP Study Guide includes several additional readings from online sources that
are freely available on the GARP website (link to 2014 Online Readings). These readings include
learning objectives that cover specialized topics or current trends in the electricity markets that
are unavailable in traditional text books.
The 2014 ERP Examination will include questions drawn from the following AIMs for each reading:
Readings for Electricity Markets and Renewable Generation

System Reliability and Demand Response
1.
Bo Shen, Girish Ghatikhar, Chun Chun Ni, and Junqiao Dudley. Addressing Energy
Demand Through Demand Response. (Berkeley National Laboratory, June 2012).
(Sections 1 to 4 only)
Define demand response (DR) and understand how DR works to curtail shortages
on a power grid.
Understand how government policy and market deregulation have been instrumental
Compare and contrast the DR programs for various RTOs.
Understand how bilateral DR programs like cost recovery and demand-side manage-
Describe the various methods used to encourage end-user participation in DR programs.
in the creation of DR programs.
ment (DSM) operate.
Global Electricity Market Applications and Current Trends

2.
Johannes P. Pfeifenberger and Kathleen Spees. Evaluation of Market Fundamentals and

Challenges to Long-Term System Adequacy in Alberta's Electricity Market. The Brattle
Group, April 2011. (Sections I, II and III only)
Understand the potential impact that regulatory changes, renewable generation, and
Identify available features of electricity market design that address resource adequacy.
Understand how price spikes can impact energy-only power markets like Alberta.
Summarize operational decisions associated with the expiration of Power Purchase
fuel-on-fuel competition may have on the electricity market in Alberta.
Agreements (PPAs) and understand the impact that expiring PPAs have had on
generating capacity in the Alberta electricity market.
Explain the effect that greenhouse gas (GHG) reduction legislation will have on the
power market in Alberta and the methods currently employed to meet GHG target levels.
Describe the challenges associated with integrating wind generation into the power
grid, including the challenges of long-term resource adequacy and the impact wind
power can have on price curves.
2014 Global Association of Risk Professionals. All rights reserved.
2014

3.
Australian Energy Market Operator. An Introduction to Australias National

Electricity Market.
Understand the roles of AEMO and NEM in the Australian market.
Identify the Market Price Cap and Market Floor Price and understand their use in
electricity transactions.
Describe the scheduling of generators for a given hour under NEM rules and calculate
the market price.
Understand the difference between regulated and unregulated interconnectors and

the auction of inter-region settlement residues.
4.
Describe how power forecasts affect market operation.
Explain the use of hedge contracts and how these transactions are settled.
Nord Pool Spot. The Nordic Electricity Exchange and Model for a Liberalized
Electricity Market.
Identify the stakeholders in a liberalized electricity market.
Describe the role and duties of the Transmission System Operator (TSO) in a deregulated market.
Summarize the process a TSO uses to manage supply and demand on the electric grid.
Calculate the settlement (payment) for a quantity of dispatched/consumed power

under NORD Pool rules.
Understand how NORD Pool up-regulation and down-regulation pricing ensures

efficient market operation.
5.
Hogan, Lovells, Lee & Lee. Singapore Energy Market (Schedule 1: Summary of Singapore
Electricity Market Deregulation and Wholesale Market Operations).
Understand Singapores creation of artificial LNG prices and the rationale for
Describe the deregulation of Singapores electricity market.
Describe the financial flows and contract settlement in both the wholesale and
Understand the types of offers made by a generator, how these offers are accepted,
this policy.
retail markets.
market prices established and offers cleared by the market.
Describe vesting contracts, their use and their settlement process.
2014

6.
P. E. Baker, Prof. C. Mitchell and Dr. B. Woodman. Electricity Market Design for a
Low-carbon Future. (Sections 1 to 6 only)
Explain the impact of wind power on electricity price cost curves, including the
Discuss how investments in renewable power generation will be incentivized under
Define the balancing market and understand the relationship between wind power
Define feed-in tariffs and understand their use and the effect on the market.
Discuss the need to provide market pricing signals to address grid congestion.
impact of negative prices.

the new market design, including capacity-based mechanisms.
production and balancing costs, and explain how balancing costs are reduced.
Global Renewable Generation, Carbon Emissions and Project Finance

7.
Nuclear Energy Agency. Nuclear Energy Today, Second Edition (2012).

(Sections 2, 4, 6 and 8 only)
Compare and contrast different nuclear reactor technologies, including pressurized

water reactors, boiling water reactors, pressurized heavy water reactors, fast breeder
reactors, Generation IV reactors, small modular reactors (SMRs), and fusion reactors.
Identify the safety risks associated with nuclear power generation, and describe the
methodologies and features used to assess and mitigate these risks.
Understand the causes, impact, and safety implications of the Three Mile Island,
Chernobyl, and Fukushima Daiichi incidents.
Compare low-level, intermediate-level, and high-level radioactive waste and identify

common methodologies for storage, transport, and disposal of each type of waste.
Understand the economics of nuclear power generation and compare the economics
to other types of electricity generation.
8.
Intergovernmental Panel on Climate Change (IPCC). IPCC Special Report on Renewable

Energy Sources and Climate Change Mitigation.
Chapter 3.........................Direct Solar Energy (Sections 3.33.5 only)
Compare and contrast various solar generation technologies, including photovoltaic

(PV) and concentrating solar power (CSP).
Understand how PV systems are integrated into the power grid.
Understand technologies used to transport and store solar generated electricity.
Explain the smoothing effect with respect to multiple PV solar systems.
Chapter 5 ........................Hydropower (Sections 5.35.5 only)
Compare and contrast run-of-river, storage hydropower, and pumped storage

technologies.
Understand factors that impact the efficiency of a hydropower plant, and compare
the efficiency of hydropower turbines over a range of discharge levels.
Describe the characteristics of hydropower generation and understand how these

characteristics can contribute to the reliable operation of a power grid.
2014

Chapter 7.........................Wind Energy (Sections 7.37.5 and 7.8 only)
Understand the dynamics of the power curve for a wind turbine.

Compare and contrast the generation technology and output associated with onshore
and offshore wind facilities.
Understand the challenges associated with planning electric power systems that
include wind energy, including the integration of wind power to the grid.
Define capacity credits and compare the capacity credit of wind power to hydrocarbon-based power.
Explain how incremental wind power capacity impacts price formation, volatility, and
balancing costs in an electricity market.
Understand the factors that impact the economics of wind power generation, including
the capacity factor, the levelized cost of energy, and the effect of government policies.
9.
Larry Parker (U.S. Congressional Research). Climate Change and the EU-Emissions
Trading Scheme (ETS): Looking to 2020.
Understand the ETS system; assess the greenhouse gas reduction commitment under
Understand how an auction system can address the issue of windfall profits that often
the ETS and identify the industries covered.

accrue to power producers under a cap and trade program like the ETS.
Identify key changes in Phase III of the ETS and understand how carbon allowances
are phased out for non-power producing industries under Phase III.
Describe the EU ETS provisions that will support industries in energy intensive,
trade-exposed areas.
10. Chris Groobey, John Pierce, Michael Faber and Greg Broome. Project Finance Primer for
Renewable Energy and Clean Tech Projects.
Describe project finance, and explain the structure of a typical project finance agreement.
Understand the importance of power purchase agreements (PPAs) in securing
project finance.
Compare different loan structures which can be used to raise project debt for a
renewable project.
Explain how project revenues are distributed to stakeholders (i.e. the project "waterfall").
Describe key U.S. government incentive structures for renewable energy projects,
including production tax credits (PTCs), investment tax credits (ITCs), and accelerated
depreciation.
LBNL-5580E
Addressing Energy Demand

through Demand Response:
International Experiences and
Practices
Bo Shen, Girish Ghatikar, Chun Chun Ni, and
Junqiao Dudley
Environmental Energy Technologies Division
Lawrence Berkeley National Laboratory
Phil Martin and Greg Wikler
ENERNOC, INC.
Reprint version of the report prepared for AZURE

INTERNATIONAL , December 1, 2011 .
June 2012
This work was supported by AZURE INTERNATIONAL, CESP and Energy Foundation
through the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.
Disclaimer
This document was prepared as an account of work sponsored by the United States Government. While
this document is believed to contain correct information, neither the United States Government nor any
agency thereof, nor The Regents of the University of California, nor any of their employees, makes any
warranty, express or implied, or assumes any legal 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 its 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, or The Regents of the University of California. The views and opinions of authors expressed
herein do not necessarily state or reflect those of the United States Government or any agency thereof,
or The Regents of the University of California.
Ernest Orlando Lawrence Berkeley National Laboratory is an equal opportunity employer.
Table of Contents
1.
2.
3.
INTRODUCTION ..........................................................................................................................1
1.1.
Definition of Demand Response ..................................................................................................................1
1.2.
Benefits Brought by Demand Response .......................................................................................................1
DEMAND-SIDE MANAGEMENT AND THE ROLE OF DR ..................................................................4

2.1.
Overview of Demand Side Management .....................................................................................................4
2.2.
Role of DR in Demand-Side Management ....................................................................................................5
2.3.
Coordination of Demand Response and Energy Efficiency ..........................................................................6
REGULATORY AND POLICY FRAMEWORKS THAT PROMOTE DEMAND RESPONSE .........................7

3.1.
3.1.1
Wholesale Market Access ......................................................................................................................................... 7
3.1.2
Bilateral Programs with Vertically Integrated Utilities and Network Operators..................................................... 12
3.2.
5.
Lack of Sufficient Incentives from Standard and TOU Pricing: Experience with Interruptible Tariffs ...................... 19
3.2.2
Cost and RisksHow Load Aggregators have Removed Traditional Barriers to DR Participation ........................... 20
3.2.3
Avoid Energy Costs vs. Resource Payments ............................................................................................................ 21
3.2.4
Emerging Trends: Dynamic Pricing ......................................................................................................................... 21
Summary/Comparative Analysis of Policy and Regulatory Frameworks ...................................................22
ENABLING TECHNOLOGY SOLUTIONS FOR DEMAND RESPONSE ................................................. 25

4.1.
Metering and Control Solutions .................................................................................................................25
4.2.
Auto-DR and OpenADR (with the AMI linkage) ..........................................................................................27
4.3.
Smart Meter and Advanced Metering Infrastructure (AMI) and OpenADR...............................................29
BEST PRACTICES AND RESULTS OF DR IMPLEMENTATION .......................................................... 31

5.1.
6.
Encouraging End-User Participation: The Role of Incentives .....................................................................16
3.2.1
3.3.
4.
Demand Response Enabling Policies ............................................................................................................7
DR Strategies in Commercial and Industrial Buildings ...............................................................................32
5.1.1
DR Strategies for Commercial Buildings ................................................................................................................. 33
5.1.2
DR Strategies for Industrial Facilities ...................................................................................................................... 33
5.1.3
DR Strategies in Water or Wastewater Facilities .................................................................................................... 33
5.1.4
DR Strategies in Refrigerated Warehouse Facilities ................................................................................................ 34
5.1.5
DR Strategies in Food Processing Facilities ............................................................................................................. 34
5.1.6
DR Strategies in Data Centers ................................................................................................................................. 35
5.1.7
DR Strategies in Heavy Industry .............................................................................................................................. 35
RECOMMENDATIONS AND KEY PRINCIPLES FOR DESIGNING AND IMPLEMENTING DR IN CHINA. 36
List of Acronyms
AESO
Alberta Energy System Operator
AMI
AMP
Auto-DR
BRA
advanced metering infrastructure
C&I
CAP
CEC
CPP
CSPs
DECC
aggregator managed portfolio

automated demand response
base residual auction
commercial and industrial
Climate Action Plan
California Energy Commission

critical peak pricing
curtailment service providers
U.K. Department of Energy and Climate Change
DLC
DOE
DPCR5
DR
direct load control

U.S. Department of Energy
DRAS
DRRC
demand response automation server

Demand Response Research Center
DSM
EE
EEPS
EILS
EISA 2007
EMCS
EPACT
ERCOT
FCM
FERC
FRCC
demand side management

energy efficiency
HVAC
heating ventilation and air-conditioning
I/C
IA
IESO
IOUs
Interruptible/curtailable
IT
information technology
LMP
locational marginal price
M&V
measurement and verification
MRO
Midwest Reliability Organization
NOCs
network operation centers
distribution price control review
demand response
Energy Efficiency Portfolio Standard

emergency interruptible load service
Energy Independence and Security Act of 2007

energy management control systems
The Energy Policy Act of 2005

Electric Reliability Council of Texas
forward capacity market
U.S. Federal Energy Regulatory Commission

Florida Reliability Coordinating Council
incremental auctions
independent electricity system operator
investor-owned utilities
NPCC
OPA
PG&E
PPA
PTR
RFC
RPM
RTDR
RTEG
RTO
RTP
SCE
SERC
SPP
SRM
STOR
TOU
TVA
WECC
WEM
Northeast Power Coordinating Council, Inc.

Ontario Power Authority
California Pacific Gas and Electric
power purchase agreement
peak time rebates
ReliabilityFirst Corporation
reliability pricing model
real time demand response
real time emergency generation
regional transmission organization
real-time pricing
Southern California Edison
Southeastern Electric Reliability Council

Southwest Power Pool
sycnhronized reserve market
short term operating reserves
time-of-use
Tennessee Valley Authority
Western Electricity Coordinating Council

whosesale electricity market
1. INTRODUCTION
1.1. Definition of Demand Response
Demand response (DR) is a load management tool which provides a cost-effective alternative to
traditional supply-side solutions to address the growing demand during times of peak electrical load.
According to the US Department of Energy (DOE), demand response reflects changes in electric usage
by end-use customers from their normal consumption patterns in response to changes in the price of
electricity over time, or to incentive payments designed to induce lower electricity use at times of high
wholesale market prices or when system reliability is jeopardized. 1 The California Energy Commission
(CEC) defines DR as a reduction in customers electricity consumption over a given time interval relative
to what would otherwise occur in response to a price signal, other financial incentives, or a reliability
signal. 2 This latter definition is perhaps most reflective of how DR is understood and implemented
today in countries such as the US, Canada, and Australia where DR is primarily a dispatchable resource
responding to signals from utilities, grid operators, and/or load aggregators (or DR providers).
1.2. Benefits Brought by Demand Response

There are a variety of benefits brought by DR, ranging from the environmental to the economic.
Environmental benefits
By reducing electric demand to ensure the sufficiency of existing supply, rather than increasing supply to
meet rising demand, DR avoids power plant operation and its associated emissions. Moreover, because
DR capacity is distributed, there are added benefits due to the avoidance of electrical losses in the
transmission and distribution lines typically experienced from centrally-generated utility power. The USbased energy consultancy Synapse Energy Economics addressed this issue in its study of DR and air
emissions in the US market of ISO New England:
when DR operates it reduces system line losses relative to reference case system operation.
This is because when energy is provided to customers from the grid it often comes from power
plants a considerable distance from the point of end use, and energy is lost in transmission.
Usually line losses are in the range of 5 to 10 percent, but they can be higher during periods
when transmission lines are heavily loaded. In contrast, the DR resource be it a load reduction
or a generator is located at the site of energy use, so no energy is lost in transmission.
Because DR avoids line losses, a DR resources of five MW is comparable to a grid-connected
asset of slightly larger than five MWFor example, a five-MW DR resource might be credited as
U.S. Department of Energy (February 2006), Benefits of Demand Response in Electricity Markets and Recommendations for
Achieving Them: A Report to the United States Congress Pursuant to Section 1252 of The Energy Policy Act of 2005, pp. 11-12.
(http://eetd.lbl.gov/ea/ems/reports/congress-1252d.pdf).
2
http://www.energy.ca.gov/2011_energypolicy/documents/2011-07-06_workshop/background/Metrics_July_IEPR_DR_v1.pdf.
providing 5.5 MW of reserve capacity is average system line losses during DR events were
determined to be roughly 10 percent.3
In addition, because DR is often procured on a forward basis, it may not only offset the operation of
power plants but also their very construction. In this manner, the environmental benefits of DR extend
to the avoided emissions associated with the construction of the materials for the power plant itself (i.e.
cement, steel, etc.), as well as the potential ecological impact that may have resulted should the unit
have been constructed.
The use of DR for non- peak-shaving purposes such as for ancillary services, also comes with significant
environmental benefits, despite the very short duration dispatches of such resources. In many systems,
ancillary services (also known as reserves), are primarily provided by plants in running operating mode,
as there may be an insufficient number of quick-start generating units able to start, synchronize, and
export power to the grid in the requisite period of time. These plants tend to be fueled by diesel or oil,
which add to local and regional pollution. Increased use of quick-response DR can reduce the need for
power plants to run in operating mode, as well as potentially lead to a more efficient overall use of
resources within the system.
Economic benefits
The economic benefits of DR oftentimes may be more significant than the environmental benefits.
While there is a clear environmental benefit to avoiding or reducing power plant operation, the targeted
usage of DR will not save the same amount of energy as permanent load reductions that come from
energy efficiency measures. As peak periods are relatively infrequent, so too tends to be the use of DR.
Yet, the infrequent spikes in demand have a significant economic impact: in many systems, 10% (or
more) of costs are incurred to meet demands which occur less than 1% of the time.4 Reducing this peak
demand through DR programs means that the capacity requirements which drive investments in
generation, transmission, and distribution assets can also be proportionally reduced. The US-based
energy consultancy the Brattle Group, in its 2007 paper The Power of Five Percent, found that a 5%
reduction in peak demand would have resulted in avoided generation and T&D capacity costs of $2.7
billion per year.5
In addition, the use of DR during peak periods can result in significant savings in terms of energy
expenditure. In wholesale markets, spot energy prices during peak periods can skyrocket due to
increased demand. Similarly, energy prices in vertically integrated, non-wholesale market systems can
also increase during peak periods as less efficient units (i.e. with a higher heat rate) are utilized in order
to meet the rising demand. As retail energy rates tend to not reflect the true cost of energy during peak
periods, the expensive utilization of generation during these times is socialized among all customers. By
reducing the need to purchase high-priced power, all customers in a system are positively impacted. The
3
Synapse Energy Economics, Modeling Demand Response and Air Emissions in New England, September 4, 2003. Page 16.
EnerNOC, Inc. Analysis of US and Australian Electricity System Data.; Brattle Group, The Power of Five Percent, May 16, 2007
5
Brattle Group, The Power of Five Percent, May 16, 2007.
4
aforementioned Brattle report also identified energy savings on the order of $300 million per year from
the same 5% reduction in the peak demand of the US as a whole. Figure 1 further illustrates the point
that the avoided capacity costs far outweigh the avoided energy and avoided T&D costs.
Figure 1: Annual Benefits of 5% Remand Response in the US6

Indeed, it is important to recognize the financial benefits participants in these programs receive, which
are the sum of both the avoided energy costs (and demand charges) as well as the direct incentive
payments for participation and successful performance.
Id.
2. DEMAND-SIDE MANAGEMENT AND THE ROLE OF DR

2.1. Overview of Demand Side Management
Demand-side management (DSM) consists of a broad range of planning, implementing and monitoring
of activities designed to encourage end-users to modify their levels and patterns of electricity
consumption. DSM programs and initiatives are typically implemented to achieve two basic objectives:
energy efficiency (EE) and load management. EE is primarily achieved through programs that reduce
overall energy consumption of specific end-use devices and systems by promoting high-efficiency
equipment use and building design. Conversely, load management programs are designed to achieve
reductions in consumption primarily during times of peak demand, rather than on a permanent or
ongoing basis. Load management programs can include permanent-load shifting and peak-shaving
activities traditionally associated with demand response. With improvements in technology, load
management programs are also increasingly dispatched on a level playing field with supply-side
resources.
Figure 2 presents the total peak load reduction through DSM from 1998 to 2009 in the US. Peak load
reduction through energy efficiency programs increased from 13,591 MW in 1998 to 19,766 MW in 2009,
but decreased from 13,640 MW to 11,916 MW during the same period by load management programs.7
35,000
25.0%
Peak Load Reduction (MW)
15.0%
25,000
10.0%
5.0%
20,000
0.0%
15,000
-5.0%
-10.0%
10,000
Annual Growth Rate (%)
20.0%
30,000
-15.0%
5,000
-20.0%
-25.0%
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Energy Efficiency
Load Management
Annual Growth Rate of Energy Efficiency
Annual Growth Rate of Load Management
Figure 2: Peak Load Reductions from DSM Programs by Program Category8
U.S. Energy Information Administration (November 23, 2010), Demand-Side Management Actual Peak Load Reduction by
Program Category. (http://205.254.135.24/cneaf/electricity/epa/epat9p1.html)
8
US Energy Information Agency. 2009.
2.2. Role of DR in Demand-Side Management

Demand Response is normally included as part of utility DSM program or a potential DSM program
solution which helps make the electric grid much more efficient and balanced by assisting the electric
grid's commercial and industrial customers in reducing their electric peak demands, and/or shifting the
time period when they use their electricity, and/or prioritizes the way they use electricity, and in return
reduces their overall energy costs.
A key difference between DR and EE is the energy reductions for DR are time-dependent, whereas
reductions for EE are not. Demand response programs yield reductions in demand at critical times,
which typically corresponds to time of peak power demand, while EE programs yield permanent energy
savings. However, the two programs have overlapping effects: EE can permanently reduce demand
including those occurred during the peak time while demand response with well-targeted control
strategies can also produce energy savings.9
Up to 2003, EE programs in the US contributed to more than 60% of actual peak load reduction;10
however, its share dropped by almost 10% from 59.3% in 2004 to 49.9% in 2009. Meanwhile,
contributions from load management had increased by about 12% from 37.6% to 50.1% during the same
period (see Figure 3).
100.0%
90.0%
80.0%
(Unit: %)
70.0%
60.0%
50.0%
40.0%
30.0%
20.0%
10.0%
0.0%
1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Energy Efficiency
Load Management
Goldman, Charles, M. Reid, R. Levy, and A. Silverstein (January 2010), Coordination of Energy Efficiency and Demand Response,
Lawrence Berkeley National Laboratory, LBNL-3044E.
(http://eetd.lbl.gov/ea/ems/reports/lbnl-3044e.pdf)
10
Peak load reductions are categorized as potential or actual. Potential peak load reductions are the amount of load available
for curtailment through load control programs such as direct load control, interruptible load control, other load management,
or other DSM programs. Actual peak load reductions are the amount of reduction that is achieved from load control programs
that are put into force at the same time as peak load and the amount of reductions that result from energy efficiency programs
at the time of peak load.
Figure 3: Share of Total Actual Peak Load Reduction by Program Category11
2.3. Coordination of Demand Response and Energy Efficiency

Demand response and energy efficiency programs could be coordinated at the customer level at least in
the following four ways:12
Offering combined programs: Although separating energy efficiency and demand response
programs are quite common, customers could be presented with both opportunities at the
same time. Furthermore, technologies that are commonly used for DR such as energy
monitoring, building automation systems and load control equipment can also be leveraged to
help inform about opportunities that would lead to energy efficiency improvements.
Coordinating program marketing and education: Program sponsors could package and promote
demand response and energy efficiency in a closely coordinated way. Because the two programs
are quite complicated to customers, program sponsors could help customers addresses both
topics under a broad DSM and management theme.
Market-driven coordination services: Effective coordination can be done not only by utilities
and Independent System Operators (a.k.a. ISO), but also by the initiative of private firms that
find a market among customers who are interested in reducing their energy costs or receiving
incentives.
Incorporating Building codes and appliance standards: Building codes and appliance efficiency
standards can incorporate demand response and energy efficiency functions into the design of
buildings, infrastructure, and power-consuming appliances/equipments. Integrating those codes
and standards can lead to significant reduction in the costs to customers of integrating demand
response and energy efficiency strategies and measures.
11
Id.
Cappers, Peter, C. Goldman, and D. Kathan (2009), Demand Response in U.S. Electricity Markets: Empirical Evidence,
Lawrence Berkeley National Laboratory, LBNL-2124E.
(http://eetd.lbl.gov/ea/ems/reports/lbnl-2124e.pdf).
12
3. REGULATORY AND POLICY FRAMEWORKS THAT PROMOTE DEMAND

RESPONSE
3.1. Demand Response Enabling Policies
DR, at least in a basic form, has been around for decades. In the US, load management and
interruptible/curtailable tariffs were first introduced in the early 1970s. The primary interest in load
management was driven in part by the increasing penetration of air conditioning which resulted in
needle peaks and reduced load factor. These programs were effectively limited to the largest industrial
customers in a given system, and in many cases never used. Deployed before the advent of the internet
or the load aggregator business model, these programs were very manual and typically featured slow
response times. With such limited capabilities, interruptible programs served less as an alternative to
generation investments, and more as a load management tool that could theoretically be used in
emergencies in reality though, they were more often than not a customer retention tool allowing
utilities to offer discounted service rates to customers large enough to fund the installation of their own
generation assets.
This base of demand response was then further spurred by two important developments: Within
traditionally-regulated, vertically-integrated utilities, the advent of integrated resource planning in the
late 1970s and 1980s made utilities increasingly aware of the system cost impacts of meeting peak loads,
and load management began to be viewed as a reliability resource. The results of this perspective were
first evident in the rise of Direct Load Control (DLC) programs that cycled residential air conditioning
units during peak periods. Even more significant, in the mid 1990s, policymakers and utilities interested
in facilitating the development of regional, competitive wholesale markets primarily based on re-design
and re-structure markets.
3.1.1 Wholesale Market Access

UNITED STATES
It is well-known to industry observers that the growth of the demand response industry in the United
States can in many ways be traced to the opportunities in these wholesale power markets, particularly
in the systems of ISO-New England13 and the PJM Interconnection14. According to the most recent
government statistics, more than 31 GW of demand response was active in the US RTO/ISO markets in
201115. While the opportunities for DR in California that emerged after the states energy crisis in 2001
certainly contributed to the growth of the industry as well, the scale of the opportunities (and the
realization of them) in the aforementioned wholesale markets has proven to be a stronger influence on
the growth of the industry in the United States.
13
The current size of the ISO-New England system is approximately 26 GW.

The current size of the PJM system is approximately 165 GW.
15
Federal Energy Regulatory Commission (2011). Assessment of Demand Response & Advanced Metering Staff Report.
November 2011. Table 2.
14
The role of government policy in the establishment of these opportunities has been an essential driver
to the growth of the DR industry in the US. The foundation of competitive power markets in the US can
be traced to the Energy Policy Act of 1992 (EPAct) and Order 888 from the Federal Energy Regulatory
Commission (FERC). EPAct began the process of electric industry deregulation and opened up the
opportunity for independent power generators to participate in wholesale markets, which FERC Order
888 furthered by requiring fair access and market treatment to transmission systems. While the
aforementioned legislation and Order were primarily focused on increasing competition among
generators, the concepts laid the groundwork for demand response to enter wholesale markets when
such resources could meet the same technical requirements as their supply-side counterparts. The
Energy Policy Act of 2005 (EPACT) further codified that a key objective of US national energy policy was
to eliminate unnecessary barriers to wholesale market demand response participation in energy,
capacity, and ancillary services markets by customers and load aggregators,16 at either the retail or
wholesale level.17
While demand response began participating at scale in wholesale power markets in the early 2000s
particularly in emergency capacity programs many market barriers remained. Fortunately, in October
2008, FERC issued Order 719, which focused on the operation of the countrys wholesale electric
markets. A major component of Order 719 was eliminating barriers to the participation of demand
response in wholesale markets operated by wholesale market operators. Order 719 permitted load
aggregators to bid demand response directly into organized markets, unless the relevant laws of the
local electric retail regulatory authority prohibit such activity.
Demand response integration into US wholesale power markets was further bolstered with the March
2011 issuance of FERC Order 745. Order 745 requires that demand response resources are paid the
Locational Marginal Price (LMP), or the wholesale market price for energy. By codifying the ability for DR
to be compensated in the same fashion as generation resources for services provided to the energy
markets, Order 745 advances the cause of equal treatment between generation and demand side
resources.
In the US, DR is primarily seen in the wholesale capacity markets, most notably in the PJM
Interconnection and ISO-New England. DR in these markets is procured in a competitive process that
places demand side resources on equal footing with generation, creating an opportunity for costeffective DR that can easily enter the market (should technical requirements be able to be met). In
addition, in both markets, capacity DR is dispatched only during the very critical peak or emergency
periods, making end-user participation relatively simple (compared to other markets to be profiled in
this paper that are solely for balancing resources). Examples of both markets are provided below.
The PJM Interconnection
16
Load aggregation is the process by which individual energy users band together in an alliance to secure more competitive
prices than they might otherwise receive working independently. Oftentimes, load aggregator companies are formed to
represent the interests of these groups of customers.
17
Cappers, Peter, C. Goldman, and D. Kathan (2009).
PJM (Pennsylvania-New Jersey-Maryland) Interconnection is a regional transmission organization (RTO)

that coordinates the movement of wholesale electricity in all or parts of Delaware, Illinois, Indiana,
Kentucky, Maryland, Michigan, New Jersey, North Carolina, Ohio, Pennsylvania, Tennessee, Virginia,
West Virginia and the District of Columbia. PJM is the largest market in the US and allows DR to
participate in all of its markets types capacity, energy, and ancillary services. Today, more than 60
entities serve as Curtailment Service Providers (CSPs), or load aggregators, in the PJM system.
Like in most systems, the bulk of the DR in PJM participates in the capacity market. Capacity markets are
particularly well-suited to peaking resources like DR which operate for relatively few hours a year and
may have trouble accessing the proper price signals from an energy-only market. PJMs current capacity
market, the Reliability Pricing Model (RPM), was instituted in 2007. In the RPM, those resources include
not only generating stations, but also demand response actions and energy efficiency measures by
consumers to reduce their demand for electricity. In this manner, demand side management is directly
integrated into the wholesale capacity market structure.
Every year PJM conducts a Base Residual Auction (BRA) for delivery of capacity three years in the future.
The BRA is held in May and the delivery year begins 3 years later on June 1st and ends on May 31st of
the following year. In addition to the BRA, PJM conducts three Incremental Auctions (IA) that are held in
advance of each corresponding delivery year. The purpose of the IA is to balance any changes in the
load forecast and to allow suppliers of capacity resources to adjust their positions. In PJMs most recent
Base Residual Auction in May 2011, the market procured 149,974 MW of capacity for the 2014/2015
delivery year. Of note, 14,118 MW of this capacity or 9.4% of the total came from demand response
resources. Once cleared through the capacity market, these DR resources become participants in PJMs
Emergency Load Response Program.
In PJM, qualifying DR resources can also participate in the wholesale energy market (both day-ahead
and real-time) as well as various ancillary service markets (primarily, the synchronized reserve market).
However, these markets are not the same drivers of DR growth that the capacity market is. The energy
market does not feature capacity incentives, and therefore requires significantly more participation to
garner the same financial opportunity. The Synchronized Reserve Market, on the other hand, requires
full response within 10-minutes of a dispatch signal and generation-grade telemetry, limiting the pool of
potential participants.
ISO New England (ISO-NE)
PJMs counterpart to the north, ISO New England, also operates a forward capacity market (FCM) in
which DR can participate alongside generation, and which accounts for the majority of demand side
participation within the New England system. Similar to PJMs BRA, the ISO-NE FCM also allows both
dispatchable demand response and energy efficiency measures to participate in the market.
ISO-NEs use of demand response may be the clearest example of a resource designed specifically for
reliability and/or emergency prevention purposes. While the ELRP in PJM is also designed for similar
9
purposes, the trigger for usage in ISO-NE is even more defined. ISO-NE treats DR provided by
curtailment and on-site generation as distinct resources, labeling the former Real Time Demand
Response (RTDR) and the latter Real Time Emergency Generation (RTEG). RTDR may be called by ISO-NE
only when the system reaches an emergency level known as Operating Procedure 4 Action 9 (OP4
Action 9). RTEG, on the other hand, cannot be dispatched until a further level of emergency has been
reached, OP4 Action 12. For customers that utilize both load curtailment and on-site generation to
provide DR capacity, they (or their DR provider), must be able to call those distinct loads separately in
order to comply with ISO-NE requirements. Today, approximately 2,000 MW, or 8% of the resources in
the capacity market, are dispatchable demand response. This figure grows to 3,400 MW, or 10% of the
ISO-NE system, in 2014/15.
Demand response resources can also provide energy to the ISO-NE market through the Real-Time Price
Response and Day-Ahead Load Response Programs. As with energy market participation in PJM, these
programs are relatively unpopular compared to the capacity market, as they require much more
frequent participation and have comparatively lower economic benefit. In both markets, sites that
participate in the energy programs tend to be among the most flexible participants in the capacity
markets who are looking for an additional economic opportunity, rather than the energy program
serving as the sole method of DR participation in the market. While ISO-NE formerly had a pilot program
testing the ability for DR to provide ancillary services the Demand Response Reserve Pilot (DRRP) it
no longer has an active mechanism for DR to provide operating or spinning reserves. DR participation in
these markets is now under active consideration, in part due to the aforementioned FERC Order 719.
UNITED KINGDOM
National Grid Short Term Operating Reserves (STOR) Market
Demand response resources also enjoy wholesale market access in the United Kingdom, albeit in a much
more limited context. Market-based opportunities for demand-side resources in the UK are currently
restricted to ancillary service markets, primarily the Short Term Operating Reserves Market. While
others exist, the parameters result in low levels of participation and or a small addressable market.18 DR
cannot access the nations wholesale energy market and unlike PJM and ISO-NE, there is no capacity
market in the UK. That said, the government, spearheaded by the Department of Energy and Climate
Change (DECC), is pushing forward legislation to launch one in the coming years and which will also
allow for demand side participation.
STOR is essentially a supply and demand balancing service that meets the need of the grid as demand
changes and as traditional power plants come online and ramp up and down, similar in many ways to
the Sycnhronized Reserve Market (SRM) in the PJM Interconnection. While not a capacity market per se,
cleared resource receive an availability payment for each hour they are in the market and available to be
dispatched. Utilization (energy) payments are also given for the actual load reduction provided. Both
features are also present in the aforementioned SRM. As a balancing market, STOR is called much more
18
For example, the Fast Reserves (FR) program has a 50 MW minimum requirement for participation.
10
frequently than the capacity programs in PJM and ISO-NE which are used primarily to address
emergency conditions. STOR participants, on average, must be prepared to respond to a dispatch every
week. Such frequent participation requires the employment of different curtailment strategies than
those that are found in capacity programs designed to shave consumption only during infrequent, peak
periods.
AUSTRALIA
Independent Market Operator Wholesale Electricity Market (WEM)
Another successful example of DR participation in wholesale markets can be found in the South West
Interconnected System of Western Australia, run by the Independent Market Operator (IMO). There are
two wholesale markets in Australia; the Whosesale Electricity Market (WEM) in Western Australia, and
the National Electricity Market (NEM) in the eastern states (except for the Northern Territory). The NEM
is an energy-only market and has very low levels of DR participation for the reasons mentioned earlier,
whereas the WEM is a capacity market similar in many ways to ISO-NE and PJM, and with a significant
penetration of DR. In the most recent Reserve Capacity Cycle, more than 8% of the capacity procured
came from demand-side resources.19
The IMO-administered WEM procures system resources through the Reserve Capacity Mechanism, in
which capacity can be traded bilaterally to the IMO directly, or to retailers. Unlike in PJM and ISO-NE
where capacity prices are the result of competitive offers, the IMO sets a price for all capacity based on
the avoided cost of a marginal new peaking unit, specifically a 160 MW open cycle gas turbine. Auctions
are only triggered if the bilateral trading mechanism secures insufficient capacity.20 As in the previously
discussed capacity markets of the US, DR and generation receives the same exact market payment.
As with PJM and ISO-NE, DR is assumed to have different levels of dispatch capability than traditional
supply-side resources. The RCM has 4 Availability Classes; Generation must all list itself as Class 1, or
available for more than 96 hours a year; DR meanwhile can offer at between 24-96 hours of dispatch.
One important distinction about DR in WA is that unlike PJM and ISO-NE; DR in the WEM can be
dispatched when it is deemed to be economic and is not dependent on emergency conditions. The
system operator is required to first utilize the plants of the former state-owned generation company,
but afterwards all dispatch is determined by the market energy price offered by the resource.
Unlike its counterparts in the US, DR in WA is limited to participation in the wholesale capacity market.
While a wholesale energy market also exists in WA, the STEM, DR resources do not have access to the
market. Meanwhile, a competitive balancing market is only just now being designed, and access for DR
is expected once the market is fully operational in the coming years.
19
The Reserve Capacity Cycle (RCC) is the process that is used in Australia to procure DR resources as part of the Reserve
Capacity Mechanism.
20
Note that to date this situation has never been experienced so no auctions have been called.
11
Capacity-based Programs in Energy-only Markets

Some wholesale market operators have taken a slightly different approach, creating DR-specific
opportunities outside of the standard wholesale markets themselves. For the most part, these are
energy-only markets, where the underlying structure is not as conducive to peaking resources like
demand response, which operate for relatively few hours per year.
In the Canadian Province of Ontario, the Independent Electricity System Operator (IESO) and the Ontario
Power Authority (OPA) launched a large scale DR program (DR3) in 2007 to provide additional capacity
to the market due to planned retirement of coal-fired power plants in the province. DR3s inclusion of a
capacity payment represented a departure from previous DR programs in Ontario that failed to gain
traction, primarily due to incentives being limited to energy payments.
Another example is the Emergency Interruptible Load Service (EILS) program in the Texas market of
ERCOT. EILS is essentially a standalone markets that exists alongside the wholesale markets open to
generation resources in ERCOT. EILS is designed to provide reserve capacity to the energy-only ERCOT
market, and is procured during four separate markets spaced evenly throughout the year. Unlike DR3,
pricing in EILS is the result of offers made by DR providers, similar to how the capacity markets in PJM
and ISO-NE work. A further similarity with PJM is that ERCOT allows EILS participating loads to provide
operating reserves in the ERCOT ancillary service markets and receive the same payment as generation
resources, similar to the SRM.
3.1.2 Bilateral Programs with Vertically Integrated Utilities and Network Operators
Access to existing wholesale markets are just one mechanism for creating and leveraging demand
response resources. In recent years, much growth in the industry has been found in bilateral programs
with vertically integrated utilities in traditionally regulated environments, and with network (T&D)
operators located within a liberalized market structure. These bilateral programs are most often used as
a way to avoid or defer investments in generation and/or T&D infrastructure, and tend to look similar in
structure to a power purchase agreement (PPA) that a utility might sign with an independent power
producer. These utility programs are likely better proxies for how the implementation of nextgeneration demand response could manifest itself in China, given the lack of a wholesale market.
There are a number of enabling policies that have encouraged the development of bilateral DR
programs throughout North America, the UK and Australia. These policies include:
Cost recovery and DSM funds
Loading orders and similar regulations
Peak demand mandates and energy efficiency portfolio standards
Cost Recovery and DSM Funds
Whether in the US, the UK, or Australia, vertically integrated utilities and distribution network operators
are regulated monopolies whose revenues are dependent on government policy and regulation. As such,
12
it is essential to understand the regulatory environments in which these utilities operate in order to
understand how regulatory policies have both contributed to, and hindered, the growth of demand
response.
Perhaps the most basic and essential enabling policy is a cost-recovery mechanisms. Under a costrecovery mechanism, a utility can recover prudently-incurred costs of DR and EE investments on a
dollar-for-dollar basis, typically through a rider or customer surcharge. Cost recovery is designed to
make a utility whole on its DR and EE investments. However, there are challenges with this approach.
First, cost recovery alone will not address the lost margin revenue the utility will face due to reduced
energy sales from DR and EE programs. Second, cost recovery does not factor in opportunity costs: DR
and EE investments displace supply-side investments for which the utility can earn a profit. Given these
opportunity costs, absent a statutory or regulatory mandate, program cost recovery alone will generally
not attract utility interest in DR and EE programs. However, in some jurisdictions, utilities are
authorized to recover additional costs associated with the lost revenue due to the energy efficiency
measures. There are also provisions for earning a fair rate of return on the DSM investment, typically at
levels that are equivalent to allowable returns on power generation assets.
Loading Orders and Similar Regulations
Loading orders are governmental proclamations that define the priority order in which resources are to
be developed. To underscore the importance of energy efficiency and demand response in Californias
future energy picture, the state government developed the Energy Action Plan established a loading
order of preferred resources, placing energy efficiency and demand response as the states highestpriority procurement resource, and set aggressive long-term goals for energy efficiency and demand
response resources. In addition, energy efficiency and demand response strategies were implemented
to address greenhouse gas emission reduction targets specified by AB32, a law adopted in California to
create regulatory policy mechanisms to combat global warming. As a result of these policies,
Californias energy efficiency and demand response efforts have proven to be very successful. California
leads the nation in term of energy saved. The state invests nearly $3 billion per year in energy efficiency
and demand response programs that target electricity and natural gas customers to install high
efficiency equipment, take measures to reduce their peak demands, and establish time-sensitive price
structures that are more in line with the actual cost of providing the electricity. Resources such as
renewable generation, distributed generation, and traditional generation are considered as the second
and third priorities, respectively in the loading order, and should only be considered once all energy
efficiency and demand response resources are exhausted.
In Massachusetts, a law known as the Green Communities Act was passed in 2008 and implemented
shortly thereafter. The law requires the states utilities to procure all available energy efficiency
resources that cost less than traditional energy sources do. The law in effect prioritizes energy efficiency
as being at the top of the loading order, ahead of renewable energy, and more traditional forms of
generation. Among the major provisions is a requirement for utilities to invest in energy efficiency when
it is less expensive than buying power. Previously companies purchased more power when demand
13
increased. The effect of the law is that the state is seeing significant investments in energy efficiency,
leading toward the ultimate goal of reducing the states use of fossil fuels in buildings by 10% and overall
greenhouse gas emissions by 20% in the year 2020.
Peak Demand Mandates, Energy Efficiency Portfolio Standards
Peak demand mandates and energy efficiency portfolio standards have recently emerged as another
mechanism to encourage DR outside of market-based opportunities. Perhaps most well known is a
mandate in the state of Pennsylvania, the so-called Act 129 legislation, signed into law in October 2008,
which requires all electric distribution companies to achieve peak demand reduction targets of 4.5% and
energy efficiency reductions of 4% by 2015. While the legislation does not expressly encourage DR over
other types of peak reduction such as energy efficiency and or solar PV, Pennsylvania utilities appear to
have determined C&I DR was the most cost effective way to reach compliance and several large deals
with aggregators have already been publicly announced.
Other states with peak demand mandates that are similar to Pennsylvania include New York, Colorado,
Michigan and Ohio. In New York, the Public Service Commission established an Energy Efficiency
Portfolio Standard (EEPS) which ordered the states utilities to achieve a 15% reduction in forecast
electricity usage by the year 2015. The states utilities are implementing aggressive EE and DR programs
in order to meet that goal, which specifies that each of the states utilities realize specific MWh and peak
MW reduction amounts by 2015. In Colorado, the Climate Action Plan (CAP) sets carbon reduction goals
for the state and proclaims that energy efficiency programs are the most important responses to the
carbon-reduction challenge. In response, the Colorado Public Utilities Commission has ordered the
states utilities to implement EE and DR programs to meet that goal. Michigan and Ohio have similar
statutory mandates to lower energy usage and peak demand.
Parity of Treatment
Traditional utility regulation favors supply-side resources over DR and EE resources. First, utilities earn a
rate of return on investments in generation, transmission and distribution infrastructure. The absence of
a parallel incentive for DR and EE investments creates a bias against demand-side resources. This has
been described in the economic literature as the Averch-Johnson Effect. That is, where a firms profits
are linked to its capital investment, as is the case with utilities under traditional regulatory structures,
there is an embedded incentive for the firm to increase its capital outlay in a manner that does not
necessarily maximize producer and consumer surplus. Stated another way, traditional regulatory
frameworks create a disincentive for utilities to meet resource needs using approaches that are less
capital intensive. Thus, faced with otherwise equivalent alternatives of building a power plant that
contributes to profitability or making investments in DR and EE that allow for cost-recovery only, a utility
would generally prefer to build a power plant (or T&D).
The government of the United Kingdom recently recognized and addressed this very challenge. In the
2010-2015 Distribution Price Control Review 5 (DPCR5), Ofgem the national electricity and gas
regulator instituted the so-called Equalisation Incentive which establishes parity in the treatment of
14
capital and operating expenditures by distribution utilities. Thus, any utility acting in its own rational
economic interest will clearly pursue the most cost-effective way to meet network needs and reliability
requirements, whether that is through traditional investments in infrastructure or through non-network
alternatives like DSR. As a result of this new regulation, one local distribution network operator
Electricity North West has already deployed a commercial scale DR program in which an aggregator is
deploying DR on specific circuits in order to defer investments in substations. Other distribution network
operators, such as UK Power Networks, are also conducting pilot projects using DR for distribution relief
as they hope to prepare themselves to launch commercial-scale programs under this new regulatory
framework.
Example Utility DR Programs
There are several examples of bilateral DR programs. In California, Pacific Gas and Electric (PG&E)
implements the Aggregator Managed Portfolio (AMP) program. AMP is a non-tariff program that
consists of bilateral contracts with aggregators to provide PG&E with price-responsive demand response.
The program can be called at PG&Es discretion. Each aggregator is responsible for designing and
implementing their own demand response program, including customer acquisition, marketing, sales,
retention, support, event notification and payments. To participate, customers must enroll through a
load aggregator. The customer in turn authorizes the aggregator to act on their behalf with respect to
all aspects of AMP, including receipt of notification of an event, receipt of incentive payments and/or
penalties. Southern California Edison (SCE) operates the Demand Response Contracts (DRC) program.
SCE has contracted with several aggregator companies to provide SCE with price-responsive and/or
demand response events that SCE may call at its discretion. Each aggregator designs their own programs,
and offers demand response program structures and options that may not be directly available through
SCE. Customers may select an aggregator with services that best meet their business needs.
More common are arrangements where a utility contracts with a single DR load aggregator for a
program in their territory (or a single provider per customer class). For example, EnerNOC, a Bostonbased load aggregator has a program in place with the Tennessee Valley Authority (TVA) in the
southeastern US, the largest public power company in the country. TVA procured a long-term, 560 MW
resource from EnerNOC which it is required to deliver in line with contract requirements over the 10year contract length. There are many other load aggregator companies operating in the various
electricity markets throughout North America. As with the aforementioned DR programs in California,
the load aggregator is responsible for all roles from customer acquisition through resource dispatch and
settlement.
As with similar DR programs, TVA has purchased a guaranteed firm resource. In addition to identifying
and enabling DR capacity in line with contract milestones, the load aggregator must also meet
performance standards when dispatched by TVA. Should the load aggregator fail to do either, financial
penalties against the aggregator may be assessed. In this manner, TVA can depend on its DR-based
virtual power plant in the same way its system planners and operators can trust a traditional
generation resource. Figure 4 provides a summary of the TVA bi-lateral program parameters.
15
Program Size
Advanced Notification
Dispatch Trigger
Availability Window
Maximum Cumulative Dispatches
Term Length
Up to 560 MW
30 minutes
TVAs discretion
April October: 12:00-20:00, Mon-Fri
November March: 5:00-13:00, Mon-Fri
40 hours per annum
10 years
Figure 4: TVA Bi-lateral DR Program Parameters

Other vertically-integrated utilities in the US that have implemented similar programs include: Arizona
Public Service, Idaho Power, NV Energy, Public Service Company of New Mexico, Puget Sound Energy,
Salt River Project, San Diego Gas & Electric, Tampa Electric, Tucson Electric Power, and Xcel Energy.
Per the aforementioned equalisation incentive now in effect in the UK, distribution network operators
(DNOs) in the country are also now deploying demand response programs to defer or avoid investments
in network infrastructure. Electricity North West (ENW), one of the 14 regulated DNOs in the UK with a
network that includes the Greater Manchester and Cumbria areas, has recently launched a DR program
along with a third-party load aggregator. Under this program, DR is deployed within specified circuits in
the network, allowing demand to be controlled on a geographically-targeted basis that will prevent the
need to upgrade the substations on those portions of the network. The program, announced in May
2011, is set to last for five years. Such network support contracts are also commonly found in Australia,
particularly in New South Wales.
The same Distribution Price Control Review that launched the equalisation incentive, also included funds
from Ofgem the UK electric regulator for Low Carbon Network (LCN) projects that will pilot new
technologies and facilitate the development of an environmentally-friendly electricity system in the
country. Many DNOs throughout the UK have successfully applied for LCN funding to pilot the use of DR
in their networks, including UK Power Networks (UKPN) and Northern Powergrid (formerly CE Electric).
ENW has also recently been awarded LCN funding from Ofgem to pilot the use of DR in new ways within
their system that, if successful, would reduce the amount of network capacity DNOs would need to have
in order to comply with reliability standards.
3.2. Encouraging End-User Participation: The Role of Incentives

The U.S. Department of Energy classifies demand response into two categories, i.e. price-based demand
response and incentive-based demand response.21 Each category has its own subcategories. Pricing
mechanisms vary on each subcategory as shown in Table 1.
Price-based demand response refers to changes in usage by customers in response to changes in
the prices they pay and include real-time pricing, critical-peak pricing, and time-of-use rates. If
the price differentials between hours or time periods are significant, customers can respond to
21
U.S. Department of Energy (February 2006).
16
the price structure with significant changes in energy use, reducing their electricity bills if they
adjust the timing of their electricity usage to take advantage of lower-priced periods and/or
avoid consuming when prices are higher. Customers load use modifications are entirely
voluntary (Table 1)
Incentive-based demand response programs are established by utilities, load-serving entities, or
a regional grid operator. These programs give customers load-reduction incentives that are
separate from, or additional to, their retail electricity rate, which may be fixed (based on
average costs) or time-varying. The load reductions are needed and requested either when the
grid operator thinks reliability conditions are compromised or when prices are too high. Most
demand response programs specify a method for establishing customers baseline energy
consumption level, so observers can measure and verify the magnitude of their load response.
Some demand response programs penalize customers that enroll but fail to respond or fulfill
their contractual commitments when events are declared (Table 1).
17
Table 1: Demand Response Options and Related Pricing Mechanisms

Price-Based (Voluntary)
Time-of-use (TOU): a rate with different unit
price for usage during different blocks of time,
usually defined for a 24 hour day. TOU rates
reflect the average cost of generating and
delivering power during those time periods.
Real-time pricing (RTP): a rate in which the
price for electricity typically fluctuates hourly
reflecting changes in the wholesale price of
electricity. Customers are typically notified of
RTP prices on a day-ahead or hour-ahead basis.
Critical Peak Pricing (CPP): CPP rates are a
hybrid of the TOU and RTP design. The basic
rate structure is TOU. However, provision is
make for replacing the normal peak price with
a much higher CPP event price under specified
trigger conditions (e.g., when system reliability
is compromised or supply prices are very high).
Incentive-Based
(Contractually Mandatory)
Direct load control: a program by which the
program operator remotely shuts down or cycles
a customers electrical equipment (e.g., air
conditioner, water heater) on short notice. Direct
load control programs are primary offered to
residential or small commercial customers.
Interruptible/curtailable (I/C) service: curtailment
options integrated into retail tariffs that provide a
rate discount or bill credit for agreeing to reduce
load during system contingencies. Penalties
maybe assessed for failure to curtail. Interruptible
programs have traditionally been offered only to
the largest industrial (or commercial) customers.
Demand Bidding/Buyback Program: customers
offer bids to curtail based on wholesale electricity
market prices or an equivalent. Mainly offered to
large customers (e.g., one megawatt [MW] and
over).
Emergency Demand Response Programs: programs
that provide incentive payments to customers for
load reductions during periods when reserve
shortfall arise. (e.g. ERCOT EILS)
Capacity Market Programs: customers offer load
curtailments as system capacity to replace
conventional generation or delivery resources.
Customers typically receive day-of notice of events.
Incentives usually consist of up-front reservation
payments, and face penalties for failure to curtail
when called upon to do so. (e.g. PJM ELRP, IMO WA)
Ancillary Services Market Program: customers bid
load curtailments in ISO/RTO markets as operating
reserves. If their bids are accepted, they paid the
market price for committing to be on standby. If
their load curtailments are needed, they are called
by the ISO/RTO, and may be paid the spot market
energy price. (e.g. PJM SRM, UK STOR)
Source: DOE (2006), p.12.
In addition to federal regulation as described in Section 3.1 and economic benefits described in Section
3.2, numbers of the U.S. utilities have taken action to expand their retail demand response programs.
One incentive factor for many of them has been concern about peak load growth and rising energy
prices.22
22
U.S. Federal Energy Regulatory Commission (December 2008), Assessment of Demand Response and Advanced Metering,
Washington D.C.
(http://www.ferc.gov/legal/staff-reports/12-08-demand-response.pdf).
18
3.2.1. Lack of Sufficient Incentives from Standard and TOU Pricing: Experience with
Interruptible Tariffs
Many utilities have offered a variety of traditional DR programs for many years. These legacy programs
are typically referred to as load management programs. There are three types of legacy load
management programs: direct load control (DLC), time-of-use (TOU) rates, and interruptible contracts.
Each of these programs use some form of incentive to encourage customers to participate. However,
the amount of the incentives or the nature of the incentives has not been sufficient to bring about
meaningful levels of demand reductions.
DLC programs allow the utility to directly control customer end-uses during certain periods when the
electrical system is under strain. The customer end-uses are directly controlled by the utility and when
events are called, those loads are either shut down, cycled on and off, or moved to a lower consumption
periods. Residential DLC programs often target air conditioners or electrical water heaters. Nonresidential DLC programs include air conditioner systems, lighting and in some regions irrigation control.
There are a number of challenges with DLC programs. First, customers tend to become frustrated with
effects of the service interruptions and oftentimes will leave the program if they are called too
frequently. Second, the incentives offered by the utilities have been insufficient to encourage their
sustained participation.
TOU rates are tariff schedules that are typically offered to residential and small business customers on a
voluntary basis and are mandatory for the largest commercial and industrial customers. The TOU rates
are structured to charge lower rates during a utilitys off-peak and partial-peak periods and higher rates
during seasonal and daily peak demand periods. By charging more during the peak period, when
incremental costs are highest, TOU rates send accurate marginal-cost price signals to customers. TOU
rates encourage customers to shift energy use away from peak periods to partial-peak or off-peak
periods and enable customers to lower their electricity bills. There are two common challenges with
TOU rates. First, the utilities have often set the TOU peak periods to be for long periods at a time, thus
limiting customers abilities to shift their loads to the lower price off-peak periods. Second, the TOU rate
programs tend to be static in nature in that the peak and off-peak prices do not change regardless of
system conditions and the true costs required to deliver electricity to customers. Because of the static
nature of the TOU rates, they cannot be counted on for meeting the peak demand needs of the utility.
In addition, the utilities often design these tariffs to be revenue neutral. That is, the price differentials
between on-peak and off-peak are intended to not change the utilitys overall revenue. This goal
oftentimes is inconsistent with a goal of maximizing customer participation in order to have meaningful
peak demand reductions as a result of the TOU tariff.
Interruptible tariffs are contractual arrangements set up between the utility and large non-residential
customers. Customers agree to reduce their electrical consumption to a pre-specified level, or by a prespecified amount, during system reliability problems in return for an incentive payment or a similar rate
19
discount. Customers are given the incentive regardless of whether reliability events are called. In the
past, these programs were developed mostly for customer retention as the utilities assured customers
that reliability events were so rare and would never be called. However, as reliability problems are
becoming more acute, utilities are calling more interruptible events. As a result, many customers are
opting to negotiate an exit to their contractual obligations for these programs as they cannot tolerate
the volume interruptions to their businesses.
3.2.2. Cost and RisksHow Load Aggregators have Removed Traditional Barriers to DR
Participation
Complicated tariff structures and insufficient incentives are just a few of the challenges utilities face
when trying to garner customer interest in traditional, non-aggregator-based DR programs. Equally
important are the costs and risks customers must bear in order to participate.
While the costs for metering and load control equipment may not always be borne by the customer in
these situations, the exposure to performance penalties remains essentially a constant. Without an
aggregator to guarantee the load response, utilities have no choice but to penalize customers if they
dont fully comply with a dispatch in order to ensure proper response. However, C&I loads are
inherently volatile and customers may not always be able to participate and consequently customers
may need to be willing to face a strong likelihood of penalties if they seek to participate.
Using load aggregators is one proven approach to removing many of these traditional barriers to DR
participation. It is typical that the load aggregator pays all costs for the installation of metering and load
control equipment, making participation for the customer a no-cost proposition. More importantly,
because load aggregators are measured on the total load reduction their entire portfolio of sites
provides, and not on a site-by-site basis, they are able to pool resources in a way that ensures that
contract performance requirements can be met. In the event that performance penalties are assessed
on the aggregator, many will still refrain from passing these onto the customer. Figure 5 illustrates this
concept.
20
Load
Aggregator
Utility/TSO
Figure 5: Aggregated Performance and Risk-Shielding
3.2.3. Avoid Energy Costs vs. Resource Payments

As is evident from the relative lack of uptake in energy-based demand response opportunities in
wholesale markets (where prices are higher and more volatile than what customers face at a retail level),
the economic benefit of avoided energy costs alone is likely to be an insufficient driver of customer
participation. While dynamic pricing may impact this trend in some ways (as discussed in the next
section), currently it is the ability to receive resource payments from DR participation that are driving
customer involvement. In both wholesale market and bilateral programs, aggregators or very large
customers that qualify for direct participation receive a payment from the entity purchasing the DR
resources, either the system operator or the utility.
Whether determined through market pressures or a utility decision, these payments are almost always
based on the avoided costs of providing the same functional service through a traditional supply-side
resource. Aggregators then use a portion of this payment stream to cover their costs of customer
acquisition, site enablement, dispatch and settlement; the remainder is used to pay the customer an
incentive payment for their participation. These payments tend to exist in the same form as those the
aggregator receives, namely energy and capacity. In wholesale markets where very large customers can
participate directly, the customer would individually receive the full payment stream, but would be
responsible for adhering to technical requirements and managing performance risk. For these reasons,
many large customers continue to work with aggregators even when the market requirements dont
make it a necessity.
3.2.4. Emerging Trends: Dynamic Pricing

Dynamic pricing refers to a category of rates that offer customers time-varying electricity prices on a
day-ahead or real-time basis. Prices are higher during peak periods to reflect higher-than-average cost
21
of providing electricity during those times, and lower during off-peak periods, when it is cheaper to
provide electricity. Dynamic pricing incentivizes customers to lower their usage during peak times,
particularly during the most critical hours of the year when peak demands spike and the cost of
acquiring electricity tends to be the highest. Dynamic pricing can take many forms. The most
sophisticated form of dynamic pricing is real-time pricing (RTP). RTP programs are where prices are set
by the utility in near real-time to match the market conditions for available power. Customers must be
able to accommodate whatever price is given, which means that they take a significant risk that if prices
spike they will either accept the higher price or be capable to rapidly reduce their consumption levels to
avoid the high prices. Because of the complexities of RTP programs, most of the examples are in the
pilot stages. Once sophisticated metering infrastructures are put into place and customers have the
necessary building automation systems, it is likely that there will be more RTP programs coming on line
in the future.
Critical peak pricing (CPP) is a less complex form of dynamic pricing. CPP programs are designed such
that the prices for the top 60 to 100 hours are defined ahead of time, but the actual times in which
these prices are in effect is not known until the day before the DR event or sometimes on the same day
as the DR event. The price differentials are intended to be quite steep (oftentimes set at three to fivetimes the peak price) to encourage the customer to reduce or shift their loads during the critical peak
times. CPP programs are offered to all customer types from residential to large commercial and
industrial. A variant of CPP is peak time rebates (PTR). In PTR programs, a standard rate is applied
during all hours but customers can earn a rebate if they reduce their consumption during the critical
peak hours. PTR programs are most applicable to residential customers.
3.3. Summary/Comparative Analysis of Policy and Regulatory Frameworks

The global survey of demand response programs in this report illustrates a variety of regulatory
constructs under which DR can thrive. Fundamentally, all of these regulations and policies in one way or
another attempt to change the traditional paradigm that has historically lead to investments in
additional supply-side infrastructure rather than load management.
Clearly, one method that has been incredibly successful in this regard is the wholesale capacity market.
By removing the type of resource from the decision-making equation altogether and rather basing
procurement decision on price alone, any resource that can meet the necessary market requirements
can be purchased. With more 8-10%, or more, of system capacity met by DR in the PJM Interconnection,
ISO New England, and the Western Australia Independent Market Operator, these markets have shown
a clear ability to drive significant penetration of demand response.
Outside of the established liberalized markets where DR is present, there tend to be more fragmented
regulatory efforts to mitigate but not eliminate the disparity in incentives between supply-side and
demand-side investments by utilities. In fact, one could argue that these multitude of policies and
initiatives are required because in most areas, the underlying financial drivers that encourage a supply-
22
side-focused perspective have not been modified: utility revenue is still tied to the amount of kWh sold,
and the amount of capital they invest in generation and/or network infrastructure.
In many regions, utilities are only allowed to recover their DSM expenditures, but cannot earn a rate of
return in the same manner as they would for supply-side investments. Because of this unequal
treatment, some jurisdictions require their utilities to first pursue DSM programs before they can build
generation assets to ensure solutions that may be cost-effective, but not financially beneficial, are
considered. In other areas, utilities are mandated to reduce the peak demands (and energy
consumption) or face penalties such as in Pennsylvania where there is no financial driver for the
utility to do anything other than build more and more infrastructure.
It is within this environment that the UKs equalisation incentive, is significant as it demonstrates a
way to create true parity of treatment outside of a wholesale market context. While wholesale
generation is competitive in the UK, distribution network operation is not they are regulated
monopolies in the same manner as vertically-integrated utilities in traditionally-regulated markets.
Moreover, in traditionally regulated areas, such a mechanism could be applied to all investments so that
generation (or alternatives to it) were also covered. In many ways, it is the concept of the equalisation
incentive that is most important, and not its exact methodology. A multitude of regulatory mechanisms
could likely be developed that would result in equal financial treatment between supply-side and
demand-side investments, and it is important to not prescribe specific methodologies that may be
better suited for one system than another.
This global survey demonstrates that good program designs are crucial to the success of demand
response, perhaps more so than the existence of a formal market structures. Regardless of how DR
programs or opportunities are engendered, programs must have the essential elements outlined in this
paper in order to be sustainable, whether they are in liberalized markets or operated by verticallyintegrated utilities.
In the wholesale capacity markets profiled in this paper, the programs found in the investor-owned
utilities of California, as well as the program for the public utility TVA, clear similarities are evident. All
such programs and markets are capacity-based, in which demand response resources are paid an
ongoing payment for being available to provide capacity. In addition, all these examples are mainly
targeted at the infrequent, yet expensive, top peak hours of the year. While there is indeed the ability
for DR to provide more frequent response, such as in ancillary service markets, these general peakshaving or emergency-prevention programs are suitable for the widest number of participants and can
therefore lead to the highest levels of customer penetration.
Lastly, the inclusion of demand response load aggregators is another key recipe for success. In wholesale
markets, often only the largest industrial customers can participate directly and aggregators are a
mechanism for small and medium sized C&I customers to participate as well. Yet, even in such
conditions, it is common for customers that could otherwise directly access the market do so instead
through aggregators for the risk-mitigation benefits discussed in this paper. And in both the wholesale
markets and among the regulated utility environments that are indeed more similar to the landscape in
23
China, we see aggregators play two other key roles that contribute to the success of DR. From the utility
or system operator perspective is the ability to provide guaranteed capacity. Once reliability can be
ensured, system planners and operators are subsequently able to depend on the DR resource and
reduce the usage of, or construction of, supply-side infrastructure. Put another way, without these
guarantees, there would be limited ability for investments in demand response that lead to
opportunities for participation among end-users. Equally if not more important is the behind-the-meter
expertise that aggregators offer. With specialized staff and technology able to implement repeatable
curtailment strategies that do not negatively impact commercial business operations, aggregators can
both identify and leverage more capacity, and achieve higher levels of customer participation.
24
Enabling Technology Solutions for Demand Response
4.
Demand Response enabling technology solutions are dependent on the level of automation a particular
facility participating in DR program is capable of. Understanding the functional capabilities of building
control systems, including the underlying technologies and software capabilities as installed, is essential
to identify and quantify a specific facilitys potential to participate in Automated Demand Response
(Auto-DR) and to maximize load reduction savings without affecting day-to-day business or operations.
The three key ways a DR program can be implemented are:
1)
2)
3)
Manual DR: This involves manually turning off or changing comfort set points, lights, or
processes or each equipment, switch, or controller.
Semi-Automated DR: This involves automation of HVAC or one or several processes or systems
within a facility using Energy Management Control Systems (EMCS) or centralized control
system, with the remainder of the facility under manual operations.
Fully Automated DR: This involves automation of an entire facility, with integration of end use
loads into an EMCS and centrally managed with no human intervention.
Regardless of the type, technology plays an important role in the reliable operation of demand response.
4.1. Metering and Control Solutions

Metering
Granular meter data is essential to the successful operation of a demand response program. First and
foremost, it is the foundation of accurate measurement and verification (M&V), which is necessary for
both proper measurement of the performance of the DR resource as well as financial settlement.
Ensuring that DR performs as expected requires real-time data, so that the actual consumption of
participating facilities can be compared to accurate forecasts of what their consumption would have
been should a dispatch not have occurred. Furthermore, real-time metering and data presentment
allows for performance monitoring during a dispatch. For aggregators, this enables them to ensure their
entire portfolio is cumulatively delivering the load reduction required, and if not, allows them to utilize
other resources to provide the proper level of curtailment. From an end-user perspective, particularly
among larger sites responding with some level of manual action, real-time data also allows for them to
ensure they have taken the proper steps necessary to comply with their intended response.
That said, it is important to differentiate between real-time meter data for DR and typically advanced
metering infrastructure (AMI). First, while AMI can facilitate DR it is by no means a prerequisite for
successful deployments. In fact, because AMI deployments are in their early stages, and often focused
on the residential customer classes, most of the technology-enabled DR present today utilizes real-time
meter data by the installation of additional technology. This typically includes directly accessing a utility
meter through analog pulse or digital serial outputs, as well as metering/sub-metering specific loads
such as a generator, and then transferring this information back to the DR aggregator using existing
broadband and wireless infrastructure.
Even when and where AMI is present, it may be insufficient. Most smart meters and their supporting
infrastructure are designed primarily with automated meter reading, and not DR in mind. As such, it is
25
common for these new smart meters to read data every half-hour or hour, and then backhaul the
consumption data once a day. Such infrequent and delayed measurements, while appropriate for AMR
purposes, do not provide the needed functionality for DR aggregators whom need to ensure delivery
standards are met in real time. In this manner, the installation of additional or specialized metering
equipment is likely required even where AMI is present.
Load Control
Load control hardware is another essential component of modern-day, technology-enabled DR
deployments and is often part of the same advanced metering kit that is installed on customer premises.
Many customer types require some level of automation in order to be able to respond to a dispatch
signal. A grocery store, for example, will typically not have an energy or facilities manager on staff able
to initiate curtailment measures. Even if personnel was present, without automation, they would likely
be unable to manually enact common strategies for this customer segment, including HVAC cycling,
partial lighting curtailment, and anti-sweat heater (condensation) control. In other situations, it is the
program requirements that require load control in order to comply with the response time. Ancillary
service programs, and some bilateral utility programs, can have response times of ten minutes, or less.
In fact, frequency responsive DR programs can have even shorter response times. For example, the
Alberta Energy System Operator (AESO) just launched a DR program with a 200-millisecond response
time. With such requirements, automation and load control is an absolute necessity. Yet even in
traditional peak management programs, remote load control is increasingly being utilized for customer
convenience and enhanced resource reliability.
The aforementioned metering/gateway devices installed are often the foundation for initiating load
control as they feature two-way communication. Such devices may toggle relays attached to specific
circuits, send scripts to Building Energy Management Systems (BEMS) to begin pre-defined curtailment
actions, or attach directly to industrial control equipment.
Dispatch, Monitoring and Management
In order to successfully leverage the metering and load control hardware described above, DR providers
commonly deploy Network Operation Centers (NOCs) to utilize the aforementioned foundation
technologies. It is from these NOCs that load aggregators can initiate automatic dispatch notifications to
participating customers, remotely control customer loads and generation, monitor performance in order
to ensure performance compliance, and coordinate technicians in the field.
Centralized control centers also allow DR to comply with telemetry requirements in a cost-effective way.
Some grid operators require resource in some of their markets (e.g. PJM Synchronized Reserves,
National Grid STOR) to be directly integrated into their respective control rooms with remote terminal
units, or other similar equipment. Such generation-grade hardware is expensive, and would be costprohibitive to deploy at individual customer sites.
26
4.2. Auto-DR and OpenADR (with the AMI linkage)

Increasingly, Auto-DR activities in California and in pilots across the U.S. are carried out through Open
communication technologies, namely the Open ADR technology developed by LBNL. Since 2010,
OpenADR is being formally standardized within standard organizations and it is selected by the U.S.
national Smart Grid activity coordinated by the National Institute of Standards and Technology (NIST) as
the only standard to communicate price and reliability-based information.23
In the Open Automated Demand Response Communications Specification (Version 1.0),24 OpenADR is
defined as a communications data model designed to facilitate sending and receiving DR signals from a
utility or independent system operator to electric customers. The intention of the data model is to
interact with building and industrial control systems that are pre-programmed to take action based on a
DR signal, enabling a demand response event to be fully automated, with no manual intervention. The
OpenADR specification is a highly flexible infrastructure design to facilitate common information
exchange between a utility or regional transmission organization (RTO)/Independent System Operator
(ISO) and their end-use participants. The concept of an open specification is intended to allow anyone
to implement the signaling systems, providing the automation server or the automation clients.25
The specification also describes the scope of the OpenADR standard: The Open Automated Demand
Response Communications Specification defines the interface to the functions and features of a Demand
Response Automation Server (DRAS) that is used to facilitate the automation of customer response to
various Demand Response programs and dynamic pricing through a communicating client. This
specification, referred to as OpenADR, also addresses how third parties such as utilities, ISOs, energy
and facility managers, aggregators, and hardware and software manufacturers will interface to and
utilize the functions of the DRAS in order to automate various aspects of demand response (DR)
programs and dynamic pricing. The OpenADR structure is illustrated in Figure 6, with the key features
defined in Box 1.
23
http://collaborate.nist.gov/twiki-sggrid/bin/view/SmartGrid/OpenADR.
Piette, M.A., G. Ghatikar, S. Kiliccote, E. Koch, D. Hennage, P. Palensky, and C. McParland. 2009. Open Automated Demand
Response Communications Specification (Version 1.0). California Energy Commission, PIER Program. CEC-500-2009-063 and
LBNL-1779E.
25
The OpenADR Primer, White paper by the OpenADR Alliance (http://www.openadr.org/).
24
27
Figure 6: OpenADR Structure

Box 1: OpenADR Features
Continuous, Secure and Reliable Provides continuous, secure, and reliable two-way communications
infrastructures where the clients at the end-use site receive and acknowledge to the DR automation sever
upon receiving the DR event signals.
Translation Translates DR event information to continuous Internet signals to facilitate DR automation.
These signals are designed to interoperate with Energy Management and Control Systems, lighting, or other
end-use controls.
Automation Receipt of the external signal is designed to initiate automation through the use of preprogrammed demand response strategies determined and controlled by the end-use participant.
Opt-Out Provides opt-out or override function to participants for a DR event if the event comes at a time
when reduction in end-use services is not desirable.
Complete Data Model Describes a rich data model and architecture to communicate price, reliability, and
other DR activation signals.
Scalable Architecture Provides scalable communications architecture to different forms of DR programs,
end-use buildings, and dynamic pricing.
Open Standards Open standards-based technology such as Simple Object Access Protocol (SOAP) and Web
services form the basis of the communications model.
During a Demand Response event, the utility or RTO/ISO provides information to the DRAS about what
has changed and on what schedule, such as start and stop times. A typical change would specify one or
more of the following:
Price signals: This would include a price multipler, a price relative, or an absoulte price
Reliability signals: This would include the load amount to be shed (difference, load level, or setpoint that a load should go to).
Levels: These are simple representations of the price and reliability signals such as NORMAL,
MODERATE, and HIGH.
28
The standard also specifies considerable additional information that can be exchanged related to DR and
Distributed Energy Resources (DER) events, including event name and identification, event status,
operating mode, various enumerations (a fixed set of values characterizing the event), reliability and
emergency signals, renewable generation status, market participation.
Widespread adoption of OpenADR will accelerate the successful implementation of DR programs

andDER, thereby providing the following four major benefits for all stakeholders:
Lower Costs Standardization lowers development and support costs for vendors and,
ultimately, their utility customers. Standardization also fosters technology innovation and
competition, which expands product choices for both utilities and end users.
Assured Interoperability Electricity providers and consumers alike benefit from being able to
choose from among a wide range of different products and services without concern for any
incompatibility or inevitable obsolescence.
Greater Reliability Products based on robust standards function dependably under normal
circumstances and are able to recover from any anticipated error conditions to deliver
dependable operation.
Enhanced Flexibility OpenADR has been designed to work with existing DR equipment (socalled backwards compatibility), as well as with newer, more sophisticated systems offering
advanced feature sets.
Commercial, industrial and residential customers, and energy aggregators, will all be able to reduce
costs, time and risk in the selection and deployment of products and systems based on the OpenADR
standard. Work being performed by the OpenADR Alliance26will educate these stakeholders about the
benefits of DR, and will increase their confidence in the available solutions with rigorous testing and
certification programs.
As a result, equipment vendors and systems integrators will be able to accelerate the time-to-market for,
and lower the development costs of, innovative products and services, while electric utilities, ISOs and
RTOs will gain faster access to the market, experience lower capital and operational expenditures, and
achieve greater success with DR programs. Even regulatory agencies will benefit from knowing that the
introduction of new pricing policies will not be undermined by incompatibilities or other end-to-end
impediments in the marketplace.
4.3. Smart Meter and Advanced Metering Infrastructure (AMI) and OpenADR
As the use of OpenADR for commercial and industrial facilities has gained significant traction in
California and other parts of the U.S., the Advanced Metering Infrastructure (AMI) and smart home
technologies are currently being implemented on a large-scale basis in residences. The AMI system wide
implementation in California residences by the utilities together with development of the supporting
26
http://www.openadr.org/.
29
technologies has provided opportunity for wide range of system operation and customer management
applications, including communicating DR information through the AMI communication channels. The
AMI communication is not open and accessible outside the utility network. AMI infrastructure can
include smart meter, which is a revenue-qualified device from which charges can be derived. Other
means of measuring power may be used, but they would generally not be qualified for revenue use from
the residence point-of-view, the advanced meter contains valuable information about current and past
power usage. This advanced infrastructure is however, not needed in most DR programs. While the
traditional electrical meters only measure total consumption and as such provide no information of
when the energy is consumed, an interval meter can usually record consumption of electricity in
intervals of an hour or less and communicate that information at least daily back to the utility for
monitoring and billing purposes. In some DR programs, an interval meter is all that is needed for a
customer to be qualified to participate.
LBNL is working with the utilities and other stakeholders to provide an external non-AMI based
OpenADR interface to the residential technologies and home automation networks (HAN). These
interfaces coexist with the AMI infrastructure that the utilities plan to use for their metering and billing
purposes. Figure 7 shows these interfaces where OpenADR can be used as a means of communication
directly with the residential gateway or the end-use devices such as the appliances:27
Figure 7: AMI-HAN Interface

Further details on the home automation technologies and its use within the DR context are available
from previous LBNL studies.28
27
28
Figure courtesy: Ron Hoffman, California Energy Commission.

McParland, Charles. Home Network Technologies and Automating Demand Response. LBNL, 2008. LBNL-3093E.
30
Evaluation of Market Fundamentals and

Challenges to Long-Term System
Adequacy in Albertas Electricity Market
April 2011
Johannes P. Pfeifenberger
Kathleen Spees
Prepared for
Copyright 2011 The Brattle Group, Inc. This material may be cited subject to inclusion of this copyright notice.
Reproduction or modification of materials is prohibited without written permission from the authors.
Acknowledgements
The authors would like to thank the AESO staff for their cooperation and responsiveness to our
many questions and requests. We would also like to acknowledge the research and analytical
contributions of Lucas Bressan and Robert Carlton. Opinions expressed in this report, as well as
any errors or omissions, are the authors alone.
TABLE OF CONTENTS
I.
Executive Summary ...........................................................................................................1
II.
Background ........................................................................................................................4
A. Market Designs to Address Resource Adequacy ........................................................ 4
1. Energy-Only Markets............................................................................................ 5
2. Market Designs Based on Administrative Capacity Payments............................. 7
3. Market Designs with Resource Adequacy Requirements ..................................... 9
B. AESOs Energy-Only Market Design ........................................................................ 9
III.
Long-Term System Adequacy Challenges Faced in Alberta .......................................11

A. Low Natural Gas and Electric Prices ........................................................................ 11
B. Expiration of Power Purchase Arrangements ........................................................... 13
C. Alberta and Federal Carbon Legislation ................................................................... 16
1. Alberta Carbon Policy......................................................................................... 16
2. Federal Carbon Policy......................................................................................... 24
D. Air Quality Emissions Regulation ............................................................................ 25
E. Wind Integration and Ancillary Services .................................................................. 27
F. Expanded Interconnections with Neighboring Markets............................................ 32
IV.
Supply-Demand Outlook .................................................................................................36

A. Supply Outlook Under Various Retirement Scenarios ............................................. 36
1. Reserve Margin and Supply Outlook without Retirements or Additions ........... 36
2. Supply Outlook with PPA Retirements .............................................................. 38
3. Supply Outlook with Federal Coal CO2 Emissions Standard ............................. 40
B. AESO Projected Future Generation Additions ......................................................... 41
V.
Analysis of Generator Economics...................................................................................43

A. Historic Trends in Generator Economics .................................................................. 43
1. Energy and Operating Reserves Prices ............................................................... 43
2. Historic Scarcity Pricing Levels ......................................................................... 45
3. Impact of Price Cap and Administrative Scarcity Pricing .................................. 47
4. Price Impact of Declining Reserve Margins ....................................................... 49
5. Costs and Operating Parameters of New Generating Plants ............................... 50
6. Historic Generator Operating Margins vs. Fixed Costs ...................................... 52
B. Outlook for Generator Economics ............................................................................ 55
1. Projecting Future Market Prices ......................................................................... 55
2. Projection of Generator Operating Margins vs. Investment Costs ..................... 61
3. Breakeven Future Prices by Technology Type ................................................... 64
VI.
Findings and Recommendations.....................................................................................66
Bibliography .................................................................................................................................68
List of Acronyms ..........................................................................................................................75
Appendices ....................................................................................................................................77
A. Generator Operating Margins versus Fixed Costs .................................................. A-1
B. Method For Projecting Operating Margins ............................................................. B-1
C. Projection of Generator Operating Margins versus Fixed Costs ............................ C-1
I.
EXECUTIVE SUMMARY
The Alberta Electric System Operator (AESO) asked The Brattle Group to review long-term
challenges to resource adequacy in Albertas electricity market and assess the following four
questions:
1. Is the market design sustainable in its current state?
2. Is the energy-only market design sustainable with minor changes?
3. Are major changes required to maintain resource adequacy?
4. What long-term adequacy metrics can be used as milestones for change?
The challenges that Alberta will face over the coming decade include: (1) the potential
introduction of new environmental regulations that could force aging plants to retire or incur
significant capital expenditures; (2) the expiration of power purchase arrangements, which may
trigger accelerated retirement partly due to decommission cost recovery regulations; (3) the
addition of wind generation capacity, which suppresses energy prices and increases price
volatility; (4) expanded interconnections with neighboring markets, which has the potential to
reduce reliability in Alberta if they lead to the province becoming dependent on interties for
resource adequacy, and (5) the continued long-term outlook of low natural gas and power prices,
which result in low operating margins and limits investment cost recovery particularly for coal
and hydro plants.
Individually, each of these challenges may impose a manageable downward pressure on reserve
margins and consequently upward pressure on market prices, ultimately resulting in relatively
stable levels of market prices and reliability. However, the combined impact of these factors
might create a resource adequacy challenge for the Alberta electricity market large enough to
result in unacceptably low levels of reliability or higher, more volatile power prices. The overall
challenge is amplified to the extent that the market will be exposed to all of these pressures
simultaneously over a relatively short period of time.
Most electricity markets around the world face a similar set of challenges, although some of
these challenges are unique to Alberta. For example, most US electricity markets do not rely on
market mechanisms to determine the desired level of reliability, but instead impose resource
adequacy standards that ensure a specific reserve margin. By doing so, reserve capacity becomes
valuable and power plants can earn revenues through bilateral or centralized capacity markets.
Other power markets offer regulated capacity payments to encourage investment. In contrast, no
similar capacity-related revenue sources are available to power plants or demand-side resources
in Alberta. Rather, investment costs need to be recovered solely through revenue earned in
Albertas energy and ancillary service markets. Note that these energy revenues may also be
hedged through short-term and long-term bilaterally contracted sales, although the prices agreed
upon in these contracts will be ultimately informed and driven by energy spot prices from the
centralized wholesale market. This energy-only market design creates significant uncertainties
about whether the market will maintain resource adequacy in the presence of the identified
challenges. In fact, some other energy-only markets, such as in Great Britain, are in the midst of
significant market redesign efforts to address these challenges.
We find that the identified challenges will come about gradually and increase the rate of plant
retirements and investment needs. However, with the possible exception of accelerated
1
retirements related to decommissioning cost recovery, the identified challenges should not result
in substantial simultaneous retirements of existing plants. The rate of plant retirement will most
likely average 220 MW per year over the next two decades, which is 1.5 times the 150 MW of
annual retirements experienced during the last decade. Considering both these retirements as
well as the anticipated load growth of 3.2% per year and an associated reserve margin
requirement increase, this would require the addition of 740 MW per year over the next 20 years.
This is almost twice the rate of historic generation additions, which averaged 380 MW over the
past decade.
We conclude that the current market design should be able to support this higher and
consequently more challenging rate of generation additions. Our analysis shows that the Alberta
market design is generally well-functioning, with energy and ancillary service prices that have
been relatively low when reserve margins were high, but that have increased enough to attract
new plant additions when system-wide reserve margins declined.
We also find that the Alberta market design will likely be able to retain existing resources and
attract new entry without dramatic price increases or a significant reduction in resource
adequacy. Our projections of future energy and ancillary service prices based on recentlyexperienced market conditions show that only modest increases in market prices, consistent with
projected increases in natural gas and carbon emission costs, should be sufficient to avoid
premature retirement of existing resources and, importantly, support investments in new
generation. We find that projected future market prices based on current fundamentals strongly
favor a shift in the resource mix from coal generation to natural-gas-fired power plants, which
are more flexible and have lower capital costs. The entry of additional wind turbines and coal
plants with carbon capture and storage may be supported by government policies and through the
value of green attributes.
As a result, and perhaps contrary to our initial expectations, we currently see no compelling need
for major changes in Albertas electricity market design. However, the outlook for resource
adequacy remains uncertain and sensitive to changes in market fundamentals and continued
evolution of the identified challenges, which must not be underestimated. It also needs to be
recognized that an energy-only market design will not be able to guarantee that a certain
reserve margin will be maintained. In fact, in a small system such as Albertas, the lack of
coordination between the retirement and online dates of individual units can cause transitional
reliability concerns and price spikes, as has been highlighted by the recently announced,
unexpected potential early retirements of Sundance 1 and 2.
Overall, we offer the following recommendations.
The AESO should carefully monitor market fundamentals in light of the identified
challenges. In addition to the already ongoing monitoring of resource adequacy metrics
based on a 24-month outlook, we recommend monitoring: (1) trends in market heat rates
and the long-term outlook for technology-specific operating margins; (2) retirement
schedules and associated system reserve margins; (3) market price impacts of wind
generation as more wind power plants come on line; and (4) the impact of interties as
they are expanded and market rules related to the use of these interties evolve.
Alberta policy makers should consider relaxing or revising the existing decommissioning
cost recovery rule to reduce the risk of large simultaneous plant retirements in 2020 when
most of the existing purchase power arrangements expire. More generally, policy makers
2
should avoid introducing regulations that could result in large simultaneous retirements,
which are difficult to manage in any market or regulated environment.
We recommend that the AESO consider increasing the current price cap from
$1,000/MWh to the lower end of estimates for the value of lost load, which tend to be
in the range of approximately $3,000/MWh. We also recommend reducing the price
floor below zero to a level where generators, including wind plants, would have an
incentive to shut down when it is economic to do so. These adjustments would also
allow for economically efficient prices during reliability events, stimulate demandresponse, facilitate entry of resources at lower average annual market prices, and make
the level of the price cap more consistent with those in other energy-only markets, such
as Texas and Australia.
Coincidentally with increasing its price cap, the AESO should consider revising its
mechanism for setting administrative prices under emergency conditions when out-ofmarket reliability actions become necessary. Under these conditions, prices should be set
to reflect the marginal cost of any out-of-market actions.
The AESO should carefully consider the long-term resource adequacy implications of its
efforts to refine the Alberta market design, which include: (1) the integration of
additional wind generation; (2) refining ancillary service markets and market designs for
demand response; and (3) the expansion of interconnections with neighboring systems.
Overall we conclude that Albertas energy-only market is generally well-functioning and
sustainable, although its efficiency and effectiveness can be improved with some design changes.
However, we caution that the current positive outlook cannot guarantee resource adequacy longterm for the simple reason that Albertas market design, like other energy-only markets, does not
include a resource adequacy requirement. For this reason the AESO must continue to monitor
potential challenges to resource adequacy over time.
II.
BACKGROUND
The Alberta Electric System Operator (AESO) asked The Brattle Group to review long-term
challenges to resource adequacy in Albertas electricity market and assess the sustainability of
the current energy-only market design from a long-term resource adequacy perspective. This
report assesses the possible impact of these challenges to the long-term sustainability of
Albertas energy-only market and explores options that may help reduce the risk of highly
undesirable outcomes. In this context, our report explores four questions:
1. Is the market design sustainable in its current state?
2. Is the energy-only market design sustainable with minor changes?
3. Are major changes required to maintain resource adequacy?
4. What long-term adequacy metrics can be used as milestones for change?
Our evaluation defines a sustainable market design as one that will provide long-term resource
adequacy through pricing signals that are sufficient to attract and retain capacity when needed.
A sustainable design can provide an efficient level of reliability without reliance on out-ofmarket or backstop mechanisms. The scope of our analysis does not include challenges
related to transmission planning, system operations, and short-term market design initiatives.
A. M ARKET D ESIGNS TO A DDRESS R ESOURCE A DEQUACY
Albertas energy-only market design lies within a spectrum of resource adequacy constructs that
have been implemented in North America and around the world, as summarized in Table 1.
Table 1 describes four different electricity market design approaches: (1) energy-only markets,
which are usually accompanied by a set of ancillary services markets, but without an explicit
resource adequacy requirement; (2) markets in which resource adequacy is ensured through
administratively determined capacity payments made directly to suppliers; (3) markets with
explicit resource adequacy requirements that mandate the procurement of reserve capacity by
retail suppliers on a short-term basis (e.g., for the next peak season); and (4) market designs that
mandate procurement of reserve capacity by retail suppliers on a forward basis (e.g., one to
several years prior to the year when the capacity is needed).
Table 1
Spectrum of Approaches to Resource Adequacy1
Type of
Centralized
Capacity
Market
None
Designs without Explicit Resource

Adequacy Requirement
Designs With Explicit Resource

Adequacy Requirement
Energy-Only
Markets
With Capacity
Payments or
PPAs
Short-Term
Forward
AESO, Australias
NEM, ERCOT,
Great Britain,
NordPool
Argentina, Chile,
Colombia, Peru,
Spain, South
Korea, Ontario
SPP, former
power pools
(NYPP, PJM,
NEPOOL)
CAISO
Voluntary
Midwest ISO
Mandatory
NYISO, former
PJM, Australias
SWIS
PJM, ISONE, Brazil
The three rows of Table 1 show that in market designs with a resource adequacy requirement for
retail suppliers, the procurement of reserve capacity may be based on bilateral contracting or
self-supply without a centralized capacity market administered by Independent System Operators
(ISO) (row 1), or they may include ISO-administered capacity markets that are either
voluntary (row 2) or mandatory (row 3).
1. Energy-Only Markets
In an energy-only market like Alberta, there is no mandated and no guaranteed level of resource
adequacy. Instead, the amount of capacity in the system is determined by the aggregate effect of
market-based private investment decisions, which are made in response to the prices and
revenues available from the energy and ancillary services markets or through bilateral
contracting with retail suppliers.2,3 Energy-only markets are usually characterized by moderate
2
3
Table 1 is based on a report prepared by The Brattle Group for PJM (see Pfeifenberger, et al. (2009)). The
table refers to the following markets according to their short names: California ISO (CAISO), Southwest
Power Pool (SPP), Electric Reliability Council of Texas (ERCOT), Alberta Electric System Operator
(AESO), Australias National Electricity Market (NEM), Australias South West Interconnected
System (SWIS), PJM Interconnection (PJM), ISO New England (ISO-NE), New York Power Pool
(NYPP), New England Power Pool (NEPOOL), and New York Independent System Operator
(NYISO).
For a full discussion of the theoretical basis for pure energy-only markets, see Hogan (2005) and Joskow
and Tirole (2004).
In many energy-only markets, there often are market interventions through the system operator or
government entities in the case of insufficient resources. Such out-of-market interventions can take the
form of backstop procurement mechanisms, government-built generation, or out-of-market cost recovery
such as government-supported long-term power purchase arrangements.
These out-of-market
interventions damage the function of the energy-only market by artificially suppressing energy-market
levels of energy prices punctuated by occasional severe price spikes. This is because sufficient
resources are available most of the time, and competitive market forces depress prices towards
the production cost of the most expensive unit dispatched. These prices near marginal
production costs are below the price levels needed for full investment cost recovery for marginal
resources. However, there will also be occasional conditions in which supplies become scarce
and energy prices increase (or even spike) to include a scarcity premium that provides
generators with the operating margins needed to recover their investment and other fixed costs.
These occasional price spikes must be large enough and frequent enough to allow the full
recovery of fixed operations and maintenance and investment costs if capacity resources are to
be attracted to and retained in the market. Revenues received from the ancillary services
markets, which tend to track with prices in the energy market, also help determine when and
which types of new capacity investments are attractive.
While such scarcity-based price spikes are inherent to the design of energy-only markets, they
can impose economic impacts on retail customers that create political challenges to maintaining
the market design. However, retail suppliers have the option to hedge against the economic
impact of this price volatility, a practice that is widespread in some energy-only markets, such as
Australias National Electricity Market (NEM).4 For buyers and sellers that are fully hedged
with long-term contracts for power, the hourly energy price has no effect other than as a
settlement tool, or as a benchmark helping to determine a reasonable price for a new long-term
contract.
Occasional high scarcity prices also motivate demand reductions through price-responsive
demand (PRD) and interruptible retail services. The price during a scarcity event must rise
until supply and demand are balanced. If that happens, the scarcity price represents an
economically efficient and accurate representation of the value customers place on consuming
peak power and avoiding interruptions in service. Energy suppliers, likewise, have an efficient
price signal indicating whether or not to invest in capacity without any administrativelydetermined resource adequacy standard. The ability to rely on customers to choose their own
desired level of reliability through the marketplace, rather than relying on administrative
determinations, is one of the (at least theoretical) advantages of energy-only markets.
Demand can adequately adjust to balance the system during supply shortages only if: (1) a large
enough fraction of the load is exposed to and is responsive to market prices; and (2) prices are
allowed to rise to the high value that customers place on reliability. In most real-world energyonly markets, there is not yet sufficient price response or interruptible load to realize the
theoretical model of how the market should behave under scarcity conditions. Instead, during a
scarcity event, the system administrator may have to rely on out-of-market actions such as
expensive off-system power purchases, voluntary emergency load shedding contracts, or resort
to involuntary load curtailments. In some markets, the actions of the system operator to increase
prices and tend to be self-perpetuating. A well-functioning energy-only market should not require such
interventions. See Pfeifenberger, et al. (2009), pp. 19-38. Albertas backstop reliability mechanism,
instituted as part of its Long-Term Adequacy Rules, allows the market operator to intervene to procure
sufficient capacity when the 2-year supply outlook is insufficient to maintain a reliability threshold, see
AESO (2008).
See AER (2007), Ch. 3.
supply through out-of-market actions during emergency events can actually have the undesirable
effect of artificially suppressing the market price. Finally, in the extreme event of firm load
shed, the market price has to be set to an administratively-determined level because the market
clearing price is an undefined quantity during a rationing event. In Alberta the price is set at the
price cap under such conditions.
The theoretically efficient price during emergency operations is the marginal cost of the next
emergency procedure. For example, if voluntary curtailments are required, the pool price should
be set equal to the per-MWh cost of the most expensive load-shed contract called upon during
the emergency.5 If involuntary curtailments of firm load are required, the most efficient price
during the rationing event is the estimated price that the average interrupted customer would
have been willing to pay to avoid interruption. This price level is referred to as the Value of Lost
Load (VOLL).6 Estimates of VOLL vary widely depending partly on the makeup of the
customer base and partly on uncertainty in estimation methods, but usually are at least in the
range of $3,000-$10,000/MWh.7,8 Administrative scarcity pricing at the VOLL crudely
approximates a demand curve for energy.9 More advanced administrative scarcity pricing
schemes, as used by the Midwest ISO for example, gradually increase the price toward the
VOLL as the necessity of involuntary curtailments becomes more likely.10
When the potential for exercise of generator market power is a concern, administrative scarcity
pricing can also allow the system operator to maintain a generator bid cap below the VOLLbased price cap, without undermining efficiently high prices during scarcity events. For
example, prices can increase to the generator bid cap as the supply stack runs out. At even
higher levels of scarcity, a combination of high-priced demand bids (which can be higher than
the generator bid cap) and administrative scarcity pricing can tie the prevailing market price
directly to the marginal cost of demand interruptions or the marginal cost of out-of-market
emergency operations.
2. Market Designs Based on Administrative Capacity Payments
Some energy market designs mitigate or otherwise suppress market prices to levels far below the
VOLL, such that they do not include a sufficient scarcity premium. As a result, suppliers are
generally unable to recover their fixed costs solely through energy and ancillary services
6
7
8
10
For example, if the load-shed contract for a 1 MW reduction costs $10,000/year and stipulates an expected
5 hours of curtailment per year, then the hourly system-wide price for any hour when the contract is
enacted should be $2,000/MWh.
See Joskow and Tirole (2004) and Hogan (2005).
See Centolella and Ott (2009) and Midwest ISO (2006).
Note that some sectors of industry, such as mining, place an extremely high value on lost load, exceeding
$50,000/MWh, see Midwest ISO (2006). However, a system-wide estimate of average VOLL does not
need to include the full VOLL of these customers if they exceed the cost of private investments in back-up
generation.
Note that if there actually were sufficient levels of demand response and interruptibility in the market, the
outcome during a scarcity event would be much more efficient because customers would self-select
reductions from low-value uses of power. Under involuntary curtailments, high and low value
applications for power are indiscriminately interrupted.
See Hogan (2005) and Newell, et al. (2010), Section IV.A.4.
markets, resulting in missing money relative to what is needed to attract and retain sufficient
capacity to meet reliability targets.11 In a market design with administrative capacity payments
as shown in Table 1, the system operator makes direct payments to suppliers or signs PPAs with
suppliers of capacity. The system administrator then recovers the costs associated with these
capacity payments via an uplift charge assessed to customers.12
There has been great variation in the determination of administrative capacity payments and the
designation of eligible suppliers. The most widely-used capacity payment design is similar to
the one first implemented in Chile in 1982.13 This was an availability-based compensation
mechanism under which any supplier bidding into the energy market would receive a capacity
payment whether or not the unit was dispatched. These capacity payments would be set such
that, over the course of the year, they would cover the annual investment costs of a peaking unit
as long as the plant demonstrated sufficient availability during months of peak demand or
capacity shortage.14
The major criticism of capacity payment systems is that they rely on administrative judgment
rather than market forces.15 In a capacity payment system, the system administrator is
extensively involved in determining the size of the payments that will be made and the type of
capacity resources that would be eligible. However, the quantity that will be supplied in
response of such payments can remain uncertain, which can lead to excess capacity or reliability
levels that remain below targets despite the administrative payments.
Maintaining target levels of resource adequacy by making administratively-set capacity
payments available only to new resources is sometimes viewed as a more cost-effective
solution than providing capacity payments to all resources. However, attempts to limit payments
only to new resources, while implemented in some places such as Spain, will not likely result in
lower costs in the long run, particularly in cases where re-investing in existing facilities would
have been lower in cost than building new facilities. Such an approach also generally risks
higher long-term costs because capacity payments are generally not made available to low-cost
capacity supply from demand-side resources, capacity uprates, or postponed retirements.
Finally, the cost of these payments is not generally reflected in market prices during peak load
conditions, which means that efficient levels of demand-response cannot be achieved even in the
absence of other barriers to demand-response.
11
12
13
14
15
For additional discussion and explanation of the meaning of missing money and how baseload,
intermediate, and peaking capacity is affected, see Hogan (2005), pp. 2-7.
See Adib, et al. (2008), pp. 336-337.
See Batlle (2007), p. 4547; Larsen (2004); Rudnick (2002).
In Chile, the peak demand months are May-September; in Colombia, the payments are made during the
dry season of December-April when hydro capacity is limited, see p. 161, Rudnick (2002). Sometimes the
capacity payments are differentiated depending on the type of resource, for example, in order to incent
investments in thermal capacity after a period of drought and associated electric shortages, Colombia
introduced increased capacity payments for thermal units. However, the units would have to make at least
some energy margins to be profitable overall, see Larsen, (2004).
For example, both the South Korean and Colombian systems have been criticized for lack of transparency
and predictability. See Park (2007), pp. 5821-22; Larsen, et al. (2004), p. 1772.
3. Market Designs with Resource Adequacy Requirements

The approach to ensuring resource adequacy used in most of the United States is based on
reserve margin requirements imposed on retail suppliers. Under this market design, the regulator
or system administrator determines the amount of capacity each retail supplier must procure to
ensure resource adequacy. For example, to ensure a system-wide reserve margin of 15%, each
retail supplier would be required to procure capacity amounting to 115% of its projected
coincident peak load.
These reserve margin requirements can be imposed on a current or forward basis. As shown in
Table 1, the Midwest ISO, SPP, and NYISO require that sufficient capacity commitments are
demonstrated immediately prior to each delivery period (e.g., prior to each delivery month). In
contrast, PJM, ISO-NE, and CAISO all require capacity procurement on an annual or 3-year
forward basis.
Another key difference among these markets is whether the design relies exclusively on bilateral
contracting and self-supply, or whether the system operator facilitates procurement through a
centralized capacity market. While the creation of a resource adequacy requirement always
creates a bilateral market for capacity, centralized capacity markets are not a necessary design
element. For example, in SPP, retail suppliers procure capacity through self-supply or bilateral
contracting.16 The Midwest ISO operates in largely the same way, but it also administers a
Voluntary Capacity Auction (VCA) through which market participants can buy or sell capacity
on a voluntary basis.17 In both Midwest ISO and SPP, retail suppliers are solely responsible for
procuring capacity, and the system operator would not intervene to fill deficiencies if any
existed. This is unlike California, where the CAISO will bilaterally procure capacity when
needed to fill any deficiencies that remain beyond what LSEs have already procured and
submitted in their capacity procurement plans. In PJM, NYISO, and ISO-NE, the ISOs procure
capacity deficits through their centralized capacity auctions.
Participation in the centralized market for procuring residual capacity is mandatory in PJM,
NYISO, and ISO-NE.18 Under these designs, retail suppliers have the option to self-supply or
contract bilaterally, and the RTO will procure any residual capacity requirements through the
mandatory centralized auction and assign responsibility for payment to retail providers.
Participation in the capacity auction is also mandatory for all existing capacity with competitive
bid levels overseen by the market monitor.
B. AESO S E NERGY-O NLY M ARKET D ESIGN
Prior to deregulating, Albertas electric sector consisted of vertically-integrated regulated
investor-owned utilities as well as municipalities and cooperatives. Under this market structure,
supply adequacy was ensured by regulator-approved cost recovery for assets needed for
16
17
18
Member utilities in SPP are mandated to fulfill the 12% capacity margin. The RTO oversees but does not
enforce this provision, with overall resource adequacy and enforcement handled by state regulators. See
NERC (2008), p. 222; SPP (2009), pp. 2.2-2.4.
See Midwest ISO (2009).
See PJM (2009); NYISO (2009).
reliability. In the 1990s, Alberta began a deregulation initiative to create competition in the
electric sector. In 1996, Alberta introduced a power pool, creating the wholesale energy market,
and over 1998-2001 Alberta deregulated its electric generation fleet.19 With these reforms,
Alberta transitioned to an energy-only market in which new generation investments would not be
mandated by regulators but rather would be attracted by market incentives. After the first few
years of experience with the energy-only market, the Alberta Department of Energy initiated a
market design review to determine whether major market modifications were required for longterm adequacy, including the option of imposing adequacy obligations on retail suppliers. The
review concluded that such a redesign was not necessary at the time, noting that the market had
attracted more than 3,500 MW of competitive new generation between 1998 and 2005.20
However, the review did recommend an initiative toward the long-term adequacy (LTA) rules.
Albertas energy-only market design is implemented along with a set of ancillary services
markets including operating reserves to ensure sufficient operating flexibility. The energy and
ancillary markets are also accompanied by a dispatch down service (DDS) settlement
mechanism to mitigate against energy price distortions from out-of-market transmission must run
(TMR) dispatch. Like other energy-only markets such as those in Great Britain, Scandinavia,
Texas, and Australia, Albertas electricity market design does not offer capacity payments and
does not have a mandated resource adequacy requirement.
Also similarly to other market designs, Alberta has out-of-market backstop mechanisms for
providing reliability when in-market signals have failed to provide sufficient supply for
reliability. One backstop mechanism is the option to sign TMR contracts with generation units
that are needed for locational resource adequacy or voltage stability although they are
uneconomic to operate as market-based assets. The LTA rule sets out another set of backstop
mechanisms that may be implemented if the two-year supply outlook appears insufficient to
maintain a reliability threshold. In this case AESO can engage in out-of-market reliability
contracts for load shedding, back-up generation, or the temporary installation of emergency
portable generation.21 The need to rely on out-of-market backstop reliability contracts such as
these could be a strong indicator of problems in the market design, which may not be providing
sufficient price signals for supply investments. Out-of-market reliability contracts can also add
to the reliability problem by suppressing market prices during periods of scarce supply, unless
they are managed carefully.
Like other energy-only markets, Alberta takes a carefully restrained approach to mitigation of
market power, allowing energy prices to spike sufficiently in response scarcity events to attract
and retain generation and demand response investments.22 This differs from the more heavyhanded price mitigation in U.S. and other electricity markets with resource adequacy standards,
which mitigate energy market prices to much lower levels but supplement suppliers cost
recovery with a capacity market or capacity payments.
19
20
21
22
See AESO (2006), pp. 4, 7-9.

See DOE (2005), pp.3-4, 26-35.
See AESO (2006), pp. 10-11.
For example, see MSA (2010a) and (2010b).
10
However, Albertas energy-only market design and market fundamentals also differ from other
energy-only markets in a number of respects. Albertas market price limits are more restrictive
than in other markets, with a price floor at zero and a $1,000/MWh price cap that is far below
reasonable estimates of VOLL.23 The relatively low price cap along with a high load factor and
other features likely combine to limit the potential for demand response (DR), which may only
choose to respond at much higher energy prices.24 Albertas centralized market is an ex-post real
time market, with no day-ahead market, hour-ahead market, or centralized generation unit
commitment.
Importantly, Alberta is a relatively small market which naturally limits the number of market
participants and the extent to which competitive locational submarkets could be maintained. The
Alberta energy-only market is also surrounded by non-market-based regions with resource
adequacy requirements, which creates some unusual challenges at the market seams. Finally,
provincial regulations, the upcoming expiration of power purchase arrangements (PPA), and
high dependence on coal under a potential federal coal retirement mandate all create challenges
unique to Alberta. These unique factors also limit the extent to which experience in other
markets can be directly applied in our analysis.
III. LONG-TERM SYSTEM ADEQUACY CHALLENGES FACED IN ALBERTA
Like other electric markets around the world, Alberta faces a series of challenges to resource
adequacy over the coming decades. Existing generators will face retirement pressures from a
number of directions, including the potential federal coal retirement mandate, Albertas carbon
and air quality emissions standards, the expiration of PPAs for most of the coal generation fleet,
and reduced operating margins caused by low electric prices. Low electric prices are driven by
the economic turndown, low natural gas prices, the growth of wind power, and the potential
expansion of interties with neighboring power markets where generators do not need to rely only
on energy market revenues to recover investment costs. In particular, the growth of wind power
and increased intertie capacity may reduce energy prices without substantially contributing to
dependable capacity available for resource adequacy. We describe each of these challenges here,
document the scale of impact that these challenges may have on the Alberta market, and discuss
the potential resource adequacy implications.
A. L OW N ATURAL G AS AND E LECTRIC P RICES
The price of natural gas directly impacts the production cost and offer prices of gas generators in
the wholesale electricity market. Because natural gas generators are the price-setting suppliers in
many hours, the price of natural gas also has a strong impact on the market clearing price for
23
24
For example, Australias NEM currently has a price floor of -$1001/MWh (-$1000 AUD/MWh) and a
price cap at their estimated VOLL of $12,512/MWh ($12,500 AUD/MWh). See AEMC (2010).
Exchange rate of 1.0009 CAD/AUD is the December, 2010 monthly average exchange rate from FRB
(2011).
For a comprehensive review of these issues, see our review of market design for DR in AESO,
Pfeifenberger and Hajos (2011).
11
electric energy.25 The close relationship between natural gas and electricity prices can be seen in
Figure 1. The figure shows historic spot and futures prices for gas at Alberta Energy Company
(AECO) C Hub, along with the AESO electricity prices over the same period. While the
relationship is not one-for-one, the impact of natural gas prices on electric energy prices can be
seen clearly during several periods of high gas prices, including early 2003, late 2005, and early
2008.26
Figure 1
Monthly Electric and Gas Prices in Alberta
$200
Historic Future
Gas
$12
$180
Electric
(Left Axis)
(Right Axis)
$160
Gas Price ($/GJ)
$10
$140
$120
$8
$100
$6
$80
$60
$4
Electric Price ($/MWh)
$14
$40
$2
$20
$0
Jan-16
Jan-15
Jan-14
Jan-13
Jan-12
Jan-11
Jan-10
Jan-09
Jan-08
Jan-07
Jan-06
Jan-05
Jan-04
Jan-03
Jan-02
Jan-01
Jan-00
$0
Sources and Notes:

Historic and Future gas prices from Bloomberg (2010); AESO electric prices from Ventyx (2010).
AECO C basis futures shown through 2010; AECO price in out years estimated by AESO (2010c) based on Henry Hub futures.
More recently, the combination of economic downturn and rapid increases in shale gas
production have resulted in lower prices for natural gas and electricity.27 AECO C Hub prices
dropped to $3.82/GJ in 2009-10 from an average of $6.14/GJ for 2003 through 2008. Coincident
with this drop in natural gas prices, electric prices have also dropped to $50.86/MWh for 20092010 from an average of $70.83/MWh for 2003 through 2008. Given the changed fundamentals
of the natural gas industry due to shale gas developments, these low gas prices are expected to
25
26
27
For example, in 2009, gas and cogen units submitted price-setting bids in 40% of hours while coal units
submitted price-setting bids in 60% of all hours. Note that this means that gas-based generators have a
disproportionately large impact on the average price. See AESO (2010e), p. 6.
Note however, that gas prices are not the only or necessarily the dominant reason for many of the observed
variations in electric prices. For example, the Alberta energy price spike in May 2010 was caused by
transmission outages. Note also that while monthly energy and gas prices have a relatively strong
correlation, hourly energy prices have a relatively weaker relationship to gas prices, with most of the
volatility explained by short-term fluctuations in supply and demand.
See, for example, Saur and Wallace (2011).
12
continue for the foreseeable future, as also indicated by the futures market for gas for the next
several years. AECO C gas prices will make only a modest recovery to approximately $4.90/GJ
by 2015 as shown in Figure 1.
These low gas and electric prices have already greatly reduced the operating margins of existing
and potential new generators as discussed in Section V.A.6. The impact is particularly
pronounced for baseload coal generators that have low operating costs but high fixed costs.
These generators require higher operating margins to cover the capital costs of new coal units
and fixed costs of existing baseload coal units. Given the additional environmental challenges
that coal generators face, and associated environmental upgrades that may be required to keep
existing units operating, these low energy margins may not only deter new entry but also force
some existing units to retire early. The likely impact that these low gas and electric prices have
had on existing coal generators in Alberta are discussed further in Section V.A.6.
B. E XPIRATION OF P OWER P URCHASE A RRANGEMENTS
As part of the transition to a competitive wholesale electricity market, 7,600 MW or
approximately 78% of the Alberta electric generating fleet were placed under PPAs in 2001.28
These PPAs were introduced to assure that generation assets built under the previous regulated,
rate-of-return regime would be able to recover their costs, while still allowing for a transition to a
competitive wholesale market. Under the terms of these PPAs, the original generation suppliers
retained ownership of the facilities but were provided with PPAs that ensured full cost recovery
through the remainder of the assets lifetime. The buyer of the PPA was obligated to make the
agreed-upon payments to the asset owner and, in return, gained the right to schedule sales and
collect revenues from the wholesale market.29 These PPAs expire over the 2003-2020 period.
Rights to 4,460 MW of PPAs covering thermal capacity were sold at a competitive auction in
August 2000, with auction proceeds returned to retail customers. An additional 2,350 MW of
thermal capacity failed to sell in the auction and 790 MW of hydroelectric capacity were not
placed in the auction.30 These unsold PPAs were transferred to the Balancing Pool, an entity
created by the Alberta government in 1998, which manages these assets as the PPA buyer and
returns any net revenues to retail customers in Alberta.31
A potential resource adequacy challenge is created by the possibility that a substantial proportion
of the units currently operating under PPAs may retire after the PPAs expire. For example, some
asset owners may be operating facilities that are recovering their fixed costs under the terms of
the PPA even though those units would not be economically viable without the PPA payments.
Additionally, asset owners need to make continuous investments into their facilities over time to
28
29
30
31
List of units originally under PPA and their MW ratings are from Appendix D, AESO (2006), AESO
(2010c), and Balancing Pool (2004), p. 9. Percentage is based on a total installed fleet of 9,400 MW in
2000 and a hydro derate to 67% of installed capacity value. List of units online in 2000 and MW ratings
from Ventyx (2010) and AESO (2010c).
See AESO (2006), pp. 12-13, 58-60.
List of units sold at auction and their MW ratings are from AESO (2006), Appendix D; AESO (2010c);
and Balancing Pool (2004), p. 9.
See Balancing Pool (2009), p. i.
13
maintain the assets and extend the operating lives of the plants beyond PPA termination. They
may, however, choose not to make these investments if they do not expect to be able to recoup
the costs once the PPAs expire. Finally, by December 31, 2018 asset owners need to determine
whether or not to decommission the facility within one year of PPA expiration to be eligible for
payment of decommissioning costs.32 The payment of these decommissioning costs may be a
major factor in the retirement decision for units with large environmental liabilities such as ash
or asbestos cleanup, particularly if these units would expect to operate only a few years beyond
PPA expiration in any case. For these reasons, the expiration of PPAs is an important factor to
consider when assessing long-term resource adequacy in the Alberta electricity market.
Figure 2 shows the historic and future PPA expiration dates and generating capacities by unit
type. The figure also shows the PPA capacity that has already retired. The experience to date
shows that generation retirement after PPA expiration is a possibility. In fact, among coal units
with already expired PPAs, 540 MW out of 680 MW have since retired, and among natural gasfired steam turbines (STs) all 840 MW have since retired.33 The figure shows some delayed
retirement dates for some of the capacity with expired PPAs in light green and light purple for
past years. However, several of these retirements may have been delayed not because they were
economically viable, but rather because they were awarded temporary (non-market-based) TMR
contracts by the AESO to avoid local reliability problems that would have been introduced by
their retirement.34 Finally, the figure also shows the potential early retirement and PPA
termination of the Sundance 1 and 2 units, which reportedly developed mechanical problems so
substantial that they may not be resolved to fulfill the PPA term.35
A key problem introduced by the scheduled PPA expirations is the fact that a large proportion of
them occur at the same time. Of the 5,400 MW of capacity currently still operating under PPAs,
4,300 MW of coal and 780 MW of hydro PPAs will expire on December 31, 2020.36 This large
quantity of simultaneous PPA expirations represents 41% of the currently-available generation
fleet, and may represent 28% of the fleet in 2020.37
Fortunately, the simultaneous retirement of all of these units after their PPAs expire in 2020 is
unlikely. As discussed further in Section V.A.6, these units generally earn sufficient returns in
32
33
34
35
36
37
Units that apply to the Alberta Utilities Commission (AUC) for retirement within one year of PPA
expiration are entitled to receive payments for any decommissioning costs unrecovered from the PPA or
from consumers prior to PPA commencement. See Balancing Pool (2009), pp. 39-40; Alberta
Government (2007b), Section 7; and Alberta Government (2003), Section 5.
Retirement dates from Ventyx (2010) and AESO (2010c).
For example, the Rainbow gas CTs and Rossdale gas STs received substantial non-market TMR contract
payments that may have contributed to their delayed retirement dates, AESO (2010c).
See TransAlta (2011).
Future PPA expiration dates from AESO (2010c). Current PPA capacity number is after the potential
early Sundance 1 and 2 retirement and PPA termination.
Dependable capacity value of hydro and wind units derated to 67% and 0% of installed capacity
respectively. Calculation is based on a current effective installed capacity of 11,730 MW and an estimated
2020 installed capacity of 17,440 assuming that future capacity will be large enough to meet projected
peak load and a 15% reserve margin. List of units online in 2010 and MW ratings from Ventyx (2010)
and AESO (2010c). The AESO projection of 2020/21 winter peak Alberta Internal Load is 15,162, see
AESO (2010a).
14
the energy market to cover their ongoing fixed costs even at relatively low market prices.
Therefore, unless faced with significant investment needs or near-term decommissioning costs,
most units will have sufficient economic incentive to continue operating beyond the PPA
expiration for the remainder of the economic life of the plant.
Figure 2
Historic and Future PPA Expirations by Unit Type
8
Historic
Future
Hydro
Gas CT
Capacity (GW) &
6
5
4
3
Coal
2
Sundance 1 & 2
Retire Prior to PPA Expiration
Coal (Now Retired)
Gas ST
Gas ST (Now Retired)
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
Sources and Notes:

Dependable capacity rating reported above for hydro is 67% of installed capacity.
Unit online and retirement dates, MW rating, and future PPA expiration dates from Ventyx (2010) and AESO (2010c).
Historic PPA expiration dates from Appendix D, AESO (2006).
Clover Bar and HR Milner PPAs originally expired in 2010 and 2012, but are reported above at their early termination dates of 2005 and 2001 respectively,
CRA (1999); p.3, Balancing Pool (2001), p.9 (2004), p. 3 (2005).
An additional challenge is that a substantial portion of these coal units likely will be forced into
retirement within a few years after 2020 regardless of the PPA expiration. These retirements will
be driven by a combination of factors discussed below, chiefly the pending federal coal
retirement mandate, large capital expenditures that might be required to life-extend an aging unit,
or capital investments that may be required for environmental upgrades. If these units would be
forced into retirement within a few years of PPA expiration, there is an increased risk that the
owners will opt to accelerate retirement by a few years in order to recover decommissioning
costs.38 For example, the federal coal retirement mandate would force 1,170 MW into retirement
over the 2020-25 period if enacted, which could lead to large simultaneous retirements if these
facilities were to accelerate retirement to recover decommissioning costs.
38
See Balancing Pool (2009), pp. 39-40; Alberta Government (2007b), Section 7.
15
Overall, these factors combine to introduce a substantial risk of a step change of an unusually
large number of retirements in 2020 and 2021. If these retirements were phased in on a more
gradual basis, it would be less challenging for market-based investments to replace the units
without introducing temporary reliability problems. A large step change in the number of
retirements may be too abrupt for the market to absorb without administrative intervention. The
number of retirements in 2020 thus should be monitored and the decommissioning cost rule
stipulated in the Power Purchase Arrangements Regulation may have to be reexamined and
relaxed to spread retirements over several years.39 The AESOs forward-looking supply
adequacy review, which summarizes suppliers announced retirement and online dates, will also
be a helpful mitigation factor. However, the AESO cannot modify announced retirements
without a resource adequacy requirement or market interventions.
C. A LBERTA AND F EDERAL C ARBON L EGISLATION
Both Alberta and the Canadian federal government have greenhouse gas (GHG) reduction
goals that could substantially affect plant retirement, resource adequacy, and the operation of the
energy-only market over the next 20 years. The Federal GHG reduction target is a 17%
reduction below 2005 levels by 2020. This compares to a less ambitious Alberta reduction target
of 21% above 2005 levels. Albertas major carbon policy initiatives are $2 billion in investments
in carbon capture and storage (CCS) technology and the Specified Gas Emitters Regulation.
Both of these efforts as well as the implications of the overall GHG strategy are discussed below.
Federal policy on GHG has yet to be codified, although the recently proposed strict carbon
emissions standard for coal would effectively require either a CCS retrofit or retirement, and
could significantly impact resource adequacy in Alberta as discussed in Subsection III.C.2.
1. Alberta Carbon Policy
The government of Alberta has a Climate Change Strategy for reducing GHG emissions in the
province as a whole, which will require large contributions from the electricity sector. The two
current initiatives that may have the largest impact on the wholesale electricity market are the
carbon capture and sequestration objectives and the Specified Gas Emitters Regulation
a. Alberta Climate Change Strategy
In January 2008, Alberta Environment published its climate change strategy, laying out a policy
framework for reducing GHG emissions in the province.40 The strategy sets a GHG reduction
target of 15% below a business-as-usual (BAU) case by 2020, and 50% below BAU by 2050.
This is equivalent to 21% above 2005 CO2-equivalent (CO2e) output levels in 2020 and 14%
below 2005 levels by 2050.41 Alberta Environments strategy includes a 139 MT CO2e
39
40
41
See Alberta Government (2007b), Section 7. Note that AESO does not have authority to revise the
decommissioning cost recovery rule, which may need to be reviewed by the Department of Energy and the
Alberta Utilities Commission.
See Alberta Environment (2008).
The unit for measuring GHG emissions used by the Alberta government and in this report is tonnes of
CO2e. Under this unit non-CO2 greenhouse gases are converted into the equivalent global warming
potential of CO2. For the electric sector, the non-CO2 emissions covered are methane (CH4) and nitrous
oxide (N2O), which typically contribute approximately 0.6% of the total CO2e emissions for coal
16
reduction (70% of total reductions) through CCS, a 37 MT reduction (19%) through greening
energy production, and a 24 MT reduction (12%) through conservation and energy efficiency as
shown in Figure 3.42,43
Figure 3
Alberta Climate Strategy GHG Reductions Plan
450
MegaTonnes of CO 2 e &
400
Carbon Capture and Sequestration

Greening Energy Production
Conservation and Energy Efficiency
450
Business as Usual
400
350
350
50 MT
300
250
200
139 MT
200 MT
300
37 MT
250
24 MT
2005 GHG Output
200
Alberta Plan
150
150
100
100
50
Historic Electric Emissions
50
0
2005 2010
0
2020
2030
2040
2050
Sources and Notes:

Alberta Environment (2008), pp. 23-24. Historic calculated from AESO (2010c)
A large fraction of these emissions reductions will be achieved within the electric sector, which
accounted for 44.1% of Albertas registered GHG emissions as of 2008 as shown in Table 2,
although only approximately 21% of total emissions as shown in Figure 3.44 Table 2 shows that
the utilities sector contributes more registered emissions than any other sector.45 The high
42
43
44
45
generators and 1.0% for gas generators. Calculation based on emissions rates from Alberta Environment
(2010a).
Id., pp. 23-24. Year 2020 50 MT reduction was explicitly reported, but percentage numbers are estimated
from a graphic representation.
One MT is equivalent to one million tonnes or one megatonne.
Total Alberta 2008 emissions of approximately 243 MT from visual inspection of figure in Alberta
Environment (2008), pp. 23-24. Total reported and unreported 2008 electric sector emissions were
approximately 51.4 MT as explained in footnote 45.
While the utilities sector is almost totally comprised of electric generation plants, emissions from each
sector are not strictly separated in all cases. In particular for cogeneration units, the electric-related
emissions and industrial process emissions are generally reported together and may be included under
either utilities or under another industry such as oil sands mining or petroleum refining. See Alberta
Environment (2010a). In an independent calculation of the sector GHG emissions based on AESO data
and separating out the cogen emissions attributable to electricity generation, the electric sector emissions
17
registered proportion from electricity is partly because more than 99% of the GHG emissions in
the electricity sector are from large point sources emitting more than 100 kT/yr, while sources
from some other sectors, such as transportation, are from diffuse sources and therefore are not
covered under the current reporting rules.46 Including unregistered emissions, the electric sector
accounted for only approximately 21% of Albertas total GHG output in 2008 as shown in
Figure 3.
Table 2
Alberta Registered GHG Emissions by Sector, 2008
Sector
Number of
Total Sector Percent of
Reporting
Total
Emissions, kT
Facilities
Chemical Manufacturing
Coal Mining
Conventional Oil and Gas Extraction
Mineral Manufacturing
Oil Sands In Situ Extraction
Oil Sands Mining and Upgrading
Paper Manufacturing
Petroleum Refineries
Pipeline Transportation
Utilities
Waste Management
Total
15
3
29
6
13
5
4
3
4
26
1
10,270
497
6,845
2,403
10,927
23,848
478
3,862
2,797
48,903
90
9.3%
0.4%
6.2%
2.2%
9.9%
21.5%
0.4%
3.5%
2.5%
44.1%
0.1%
109
110,921
100%
Sources and Notes:

Alberta Environment (2010a), p. 7.
Total Alberta GHG emissions are not represented in this table; only facilities
outputting more than 100 kT of CO2e annually must report their emissions.
Meeting these targets of 15% CO2e reductions below BAU by 2020 and 50% below BAU by
2050 will have a large impact on Albertas generation fleet and its energy-only market. Many of
the impacts can only be inferred, however, because the measures that will be enacted to meet
these goals have not yet been specified. Nevertheless, the most immediate impacts on Alberta
wholesale electricity prices and resource adequacy will come from the 2020 goals, toward which
the electric sector may have to contribute approximately 15 MT of reductions from CCS, 4 MT
from greening production, and 3 MT from efficiency.47 These reduction targets may have the
following impacts:
46
47
were approximately 51.4 MT in 2008, 51.0 MT of which were from units emitting more than 100 kT/yr.
This calculation would put electric sector emissions at 45.9% of all registered emissions in Alberta, AESO
(2010c).
One kT is equivalent to one thousand tonnes or one kilotonne. One MT or megatonne is equal to 1,000 kT
or one million tones.
Based on approximate BAU emissions estimate, see Footnote 42. Assumes that contributions toward CO2e
reductions by category are the same in 2020 as in 2050, or 35 MT CCS, 9 MT greening production, and 6
MT efficiency over all of Alberta. Electric sector reductions are assumed to be achieved in proportion to
18
Carbon Capture and Sequestration Achieving a 15 MT reduction of CO2e in the electricity

sector by 2020 may require approximately 3,620 MW of CCS-enabled coal generation,
compared to a coal fleet of 5,780 MW in 2010. The scale and implications of this goal
and current large scale CCS projects are discussed in the next subsection.
Greening Energy Production The largest initiative enacted to date toward achieving 4 MT
of CO2e reductions through greening energy production is Albertas Specified Gas
Emitters Regulation. This regulation requires GHG reductions below a per-unit historic
baseline or else requires payments on excess emissions as discussed further below.
Alberta is investing revenue collected under this regulation into carbon-reducing
programs and renewable energy sources, the potential AESO impacts of which are
discussed further in Section III.E.
Energy Efficiency and Conservation The Climate Change Strategys efficiency goal
amounts to approximately 4% consumption reductions below BAU by 2020.48 This goal
could reduce energy load growth from the forecasted 4.6% to 4.1% annually between
2010 and 2020.49 While large efficiency gains would tend to reduce wholesale energy
prices and relieve resource adequacy concerns if implemented quickly on a large scale,
the gradual introduction planned is unlikely to substantially impact the Alberta energyonly market either in terms of prices or resource adequacy.
Overall, however, Albertas climate change strategy may require significant changes to the
makeup of the generation fleet, impacting the wholesale electricity prices and resource adequacy.
b. Carbon Capture and Sequestration
The Government of Alberta has awarded $2 billion in financial commitments to developing four
large CCS projects in Alberta, along with $526 million in federal investments as shown in Table
3. Together, these four projects are expected to achieve 5 MT of annual CO2 sequestration by
2015.
48
49
the electric sectors share of registered GHG emissions, or 44.1% of the total as shown in Table 2. In
reality, the share of reductions required from the electric sector may be greater than from other sectors.
Percentages assume that efficiency gains in the electric sector will be proportional to its share of the
registered Alberta GHG emissions or 3 MT by 2020. From Table 2, the registered CO2e rate in the electric
sector was 48.9 MT in 2008 over 69,947 GWh of AIL from AESO (2010a). At this same emissions rate of
0.70 kT/GWh, a 3 MT reduction by 2020 would require a 4,291 GWh reduction in AIL by 2020 or 3.9%
of the current projection of 108,638 GWh.
Compound annual growth rates calculated from AESO projected Alberta Internal Load (AIL) energy of
72,459 GWh in 2010 to 113,652 GWh in 2020 and an alternative 2020 load reduced by 2%. AESO
(2010a).
19
Table 3
Large-Scale CCS Projects under Development in Alberta
Project
Description
Online Date
Generation Government
Capacity Awards, $M
2015
300 MW
Alberta Carbon Trunk Line
[1] Underground coal gasification.

Above ground separation of syngas
from CO2 and impurities. Power
production from syngas in separate
combined cycle facility.
[2] 240 km CO2 pipeline from near Fort
Saskatchewan south to Clive, to be
used for EOR. Initial CO2 will
come from Agrium Redwater
Complex and the North West
Upgrading facility once completed.
2012
n/a
$495 Alberta 14.6 MT Capacity

$63 Federal 1.8 MT Initially
Shell Quest Project
[3] Scotford bitumen upgrader facility.
2015
n/a
$745 Alberta
$120 Federal
$436 Alberta
$343 Federal
Swan Hills Synfuels
Project Pioneer on Keephills 3 [4] Carbon capture retrofit to coal plant. 2011 Power Output
2015 Carbon Capture
450 MW
$285 Alberta
Annual CO2
Sequestration
1.3 MT
1.2 MT
1 MT
Sources and Notes:

[1 - 4] Alberta Government (2010a); Natural Resources Canada (2010).
[1] Alberta Government (2009).
[2] Enhance Energy, (2010).
Agrium fertilizer plant retrofit is planned to provide 0.25-0.55 MT of CO2 annually, Agrium (2010).
North West Upgrading will produce 1.3 MT annually from each of 3 identical phases with Phases I and II online 2013 and 2018,
North West Upgrading (2010a-b).
1.8 MT initial output assumes both Agrium and North West Upgrading Phase 1 are operational.
[4] TransAlta (2010).
Two of these planned CCS projects are planned for new coal generation facilities with a total
capacity of 750 MW and an expected 2.3 MT of total annual CO2 sequestration. Once energy
consumption of the CCS equipment is accounted for, these projects may contribute
approximately 1.7 MT of net avoided CO2 or 11% of the 2020 GHG reduction target for the
electric sector.50 If the rate of avoided CO2 emissions can be improved on future projects to 81%
below the emissions rate of a new coal plant without CCS, then an additional 2,880 MW of CCSenabled coal generation may have to be built or retrofitted by 2020.51 Combined with the
projects currently under way, the potential 3,630 MW of CCS-enabled coal generation by 2020
compares to an existing coal fleet of approximately 5,780 MW as of 2010. This ambitious CCS
goal represents a massive build-out of capacity that will likely be too aggressive to achieve.
However, if this target is met, CCS-enabled coal will represent two thirds of the current coal fleet
50
51
A fraction of the CO2 sequestered is not counted toward net avoided CO2 emissions. Net avoided CO2
emissions are approximately 72%-76% of captured CO2 emissions for pulverized coal plants because CCS
technology consumes power itself, and therefore decreases the net plant capacity rating for power
deliverable to the grid. See IPCC (2005), Table 8.3a.
Calculation assumes that 15 MT of the 2020 CCS coal must be met in the electric sector, or a fraction
proportional to the currently registered emissions from the utilities sector, of which 1.7 MT will be met by
the two electric projects already funded as described in Table 3. Also assumes that 620 kg CO2/MWh can
be avoided and units would operate at 85% capacity factor. See IPCC (2005), Table 8.3a.
20
and about 21% of the entire Alberta generation fleet by 2020. For comparison, coal currently
accounts for 49% of the generation fleet.52
While this estimate of the required CCS-enabled coal generation is only a rough approximation
of the investment required to meet the Climate Strategy targets, it can be used to infer the scale
of impacts on the AESO electricity market. Large governmental investments in CCS-enabled
coal over the coming decade could boost resource adequacy in Alberta by supporting new
generation additions or possibly enabling the retrofit and refurbishment of coal units that
otherwise would be retired.
These CCS-enabled coal plants may also impact wholesale electricity market prices by operating
as must-run units to achieve high levels of CO2 sequestration. If operating as must-run
generation, they are likely to bid into the wholesale energy market at or near zero, thereby
tending to suppress market prices. During peak hours, this price suppression may not be a
problem, especially if peak prices are allowed to rise to levels that can support new entry.
During off-peak hours, however, this addition of must-run units could potentially increase the
frequency of surplus supply conditions. At these times, wholesale electricity prices can drop to
zero and must-run units will operate at a loss because they are unable to reduce output without
incurring even larger shutdown-related costs. During some low-load conditions, the AESO must
force these units to ramp down or shut down to maintain system stability, regardless of additional
costs. Note that the efficient market price during such events would be negative because
generators would rather pay some amount (up to their shut-down-related costs) than be forced to
reduce output further, as discussed further in Section III.D. The economics of must-run coal,
cogeneration and, increasingly, wind generation already result in occasional surplus supply
conditions. Due to the low dispatch flexibility of CCS plants, the frequency and severity of these
conditions could increase as CCS generation expands.53
c. Specified Gas Emitters Regulation
Albertas Specified Gas Emitters Regulation went into effect July 1, 2007, requiring emissions
reductions from all Alberta facilities outputting more than 100 kT of CO2e annually.54 These
facilities were assigned a GHG emissions intensity reduction target of 12% below their baseline
output established over 2003-2005.55 For electric generators, this target is a requirement to
reduce the quantity of CO2e emitted per MWh produced. In order to comply with the regulation,
facilities have four options:
Improve the efficiency of operations to reduce per-unit output by 12%,
52
53
54
55
Year 2020 percentage assumes installed capacity will be equal to projected Alberta Internal Load of
15,160 MW plus a 15% reserve margin; current effective installed capacity is 11,730 MW assuming that
hydro dependable capacity is 67% of installed capacity and wind dependable capacity is 0%. From AESO
(2010a), AESO (2010c).
For a full analysis of minimum generation conditions in AESO, see AESO (2010b).
See Alberta Government (2007a), pp. 5-7.
A new units baseline is determined from the 3rd year of operations, with the efficiency requirement
ramped up to the full 12% by the 9th year of operations. See Alberta Government (2007a), pp. 7, 17.
21
Contribute $15/tonne of CO2e to the Climate Change and Emissions Management

(CCEM) Fund, which invests these funds in projects to reduce emissions
elsewhere,
Purchase offset credits for CO2e emissions from Alberta-based projects that reduced
output but are not covered by the regulation, or
Purchase performance credits from other Alberta GHG emitters that exceeded their
12% GHG reduction target.56
Year 2009 program results show that Alberta-wide reductions targets were approximately 11 MT
in total and approximately 4.9 MT for the electricity generation sector. Of the total from all
sectors, 38% or 4.2 MT of the reductions target were met via payments to the CCEM, 28% or 3.1
MT were met by either operational improvements or performance credits to facilities covered by
the regulation, and the remaining 34% or 3.8 MT were supplied by Alberta-based CO2e offsets
as shown in Figure 4.
CO2 e Reductions or Payments, MT&
Figure 4
Alberta GHG Specified Gas Emitters Regulation Compliance by Category
12
10
CCEM Fund Payments
8
6
Alberta Offsets
4
Performance Credits
Operational Improvements
0
2007
(half year)
2008
2009
Sources and Notes:

Alberta Government (2010c).
These compliance requirements put a cost burden on large GHG emitters participating in the
Alberta market. As of 2009, there were approximately 4,090 MW of natural gas and 6,060 MW
of coal plants subject to this regulation, representing 85% of Albertas generation fleet.57
56
57
See Alberta Government (2010b), pp. 2-3.

Based on 2009 effective installed capacity of 11,920 after accounting for a 67% net dependable capacity
rating for hydro and 0% dependable for wind. Determination of units covered by the regulation is based
on a calculation of estimated GHG output in 2009 for each unit from AESO internal generation data and
estimated heat rates, AESO (2010c), Ventyx (2010).
22
Covered suppliers are likely to pass these increased production costs through to the wholesale
electricity market in the form of increased offer prices to the extent that their offer prices are
based on their marginal production cost rather than strategic bidding.
To scale the total impact that this regulatory requirement may have on the market, we can
examine a case in which suppliers meet their entire regulated efficiency reduction through
$15/tonne CO2e payments. Table 4 shows the approximate impact that paying full price for
these emissions would have on the production cost for gas and coal units. As the table shows, if
the full $15/tonne compliance cost is paid, then this regulation increases production costs by
approximately 2% for natural gas-fired combustion turbine (CT) and combined cycle (CC)
plants, and by approximately 13% for coal plants.
Table 4
Approximate Production Cost Impact of Alberta CO2 Emissions Regulation
Gas CC
Gas CT
Coal
Assumptions
Fuel Cost, $/GJ
Heat Rate, GJ/MWh
GHG Rate, kg/GJ
CO2e Cost, $/tonne
Fraction of CO2e Output Charged
[1]
[2]
[3]
[4]
[5]
6
7.7
56.6
$15
12%
6
13.2
56.6
$15
12%
1
10.4
103.1
$15
12%
Costs, $/MWh
Fuel
VOM
CO2e
[6]
[7]
[8]
$46.29
$2.22
$0.79
$79.43
$3.84
$1.35
$10.37
$4.93
$1.92
Total Cost w/o CO2e Charges
[9]
$48.50
$83.26
$15.30
Total Cost w/ CO2e Charges
[10]
$49.29
$84.61
$17.22
% Cost Increase w/ CO2e Charges
[11]
1.6%
1.6%
12.6%
Sources and Notes

[1] Mine-mouth coal cost assumption provided by AESO.
Approximate average AECO C price over 2008-09, Bloomberg (2010).
[2] AESO class average fully loaded heat rate from Ventyx (2010).
[3] CO2 rate from EIA (2010b).
Ratio of CO2e to CO2 calculated from Alberta Government (2010b).
[6] = [1] * [2]
[7] See Table 8.2, EIA (2010a). Converted to 2010 CAD.
[8] = [2] * [3] * [4] * [5] / 1000
[9] = [6] + [7]
[10] = [6] + [7] + [8]
[11] = ( [10] - [9] ) / [11]
Because the payments apply to only 12% of total plant output, the increase in production costs
associated with this regulatory requirement is quite small, amounting to approximately
$0.80/MWh to $1.90/MWh. The cost impact on natural gas plants is comparable to the impact
of a 2% increase in natural gas prices, a minor impact compared to the daily and monthly
23
volatility of gas prices.58 Given the small scale of this impact, it appears that the Specified Gas
Emitters Regulation is unlikely to substantially impact retirements or resource adequacy. The
cost of emitting CO2e and the portion of the sectors output that is covered by the regulation
would have to be increased substantially before the regulation would have a material impact on
the economics of existing units or the wholesale electricity market.
2. Federal Carbon Policy
Federal GHG reduction goals are substantially more ambitious than Albertas. The Canadian
federal government has committed to reducing GHG output to 17% below 2005 levels by 2020,
compared to the Alberta goal of 21% above 2005 levels by 2020.59 The federal government
announced this regulatory framework for GHG emissions reductions in 2007, but the framework
has not been translated into binding regulation in time to meet the original 2010 reductions goals.
Although the Alberta regulation discussed in Section III.C.1.c remains the only binding GHG
regulation currently affecting Alberta, the more ambitious federal commitment highlights the
possibility of substantial federal mandates.
A federal policy that would have a large impact on Alberta is the coal generation performance
standard proposed by the Minister of the Environment on June 23, 2010, for which draft
regulations may be published in spring 2011.60 The proposed regulation would effectively phase
out all coal generation in Canada without CCS. The proposal would require coal plants to meet a
strict performance standard based on CCS technology, or else retire after the later of its 45th
operating year or PPA expiration. In order to build new coal plants or extend the lives of
existing coal plants, operators would have to meet a GHG emissions performance standard of
approximately 360 to 420 kg of CO2e/MWh, putting it in the range of the emissions rate of a
natural gas-fired CC or a CCS-enabled coal unit with an overall 50% rate of net avoided CO2e.61
The entire Alberta coal fleet would be affected by this regulation, but the impact would be
phased in over the next twenty years, as shown in Figure 5. The figure shows Alberta coal
capacity that would be subject to retirement under the regulation. These retirements add up to an
overall retirement rate of approximately 210 MW per year starting in 2015. This is 1.5 times the
retirement rate observed in Alberta over the past decade.62 The retirements driven by this
58
59
60
61
62
For example, in 2009 daily AECO C gas prices had a standard deviation of 26% of the average annual
value, while monthly gas prices had a standard deviation of 25% of the average. Bloomberg (2010).
See Environment Canada (2010a).
See Environment Canada (2010a-b).
Note that the total rate of avoided CO2e is lower than the rate of captured CO2e because of the efficiency
losses associated with CCS. For comparison, the emissions rate of a typical gas CCs is approximately 344
to 379 kg/MWh and the emission rate for a new coal unit without CCS is approximately 736 to 811
kg/MWh. See IPCC (2005), Table 8.1.
Over 2001 through 2010, the annual retirement rate in AESO has been approximately 150 MW per year.
AESO (2010c). Note that these numbers are different from those reported on page 2. Page 2 reports the
total retirements over 2011-29 projected by AESO in their long-term planning activities, including
Sundance 1 and 2 and some adjustments to assumed retirements timing as informed by factors other than
just the federal coal mandate. The numbers here cover a shorter time span, exclude Sundance 1 and 2, and
include only retirements that would be driven by the federal coal mandate over 2015-29.
24
potential federal mandate, along with retirements that may occur for other reasons, would
noticeably increase the rate of new plant additions required to maintain resource adequacy.
Figure 5
Alberta Coal Units Subject to Proposed Federal Coal CO2 Emissions Standard
4.0
3.5
Capacity (GW) &
3.0
2.5
Units Subject to Retirement Mandate
2.0
1.5
1.0
Sundance 1 & 2
Retire Prior to Mandate
0.5
2029
2028
2027
2026
2025
2024
2023
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
Sources and Notes:

Sundance 1 & 2 retirement mandates would have been in 2017 and 2018 respectively. No units will be under PPA past their 45th year.
Terms of retirement mandate from Environment Canada (2010a-b).
Unit online date and MW rating from Ventyx (2010) and AESO (2010c).
D. A IR Q UALITY E MISSIONS R EGULATION

In Alberta, air quality emissions from the electric sector are regulated under the Emissions
Management Framework for the Alberta Electricity Sector. The framework was developed by
the Clean Air Strategic Alliance (CASA) stakeholder group and proposed to Alberta
Environment in 2003.63 The original framework standards were adopted in 2006, as was the
recommendation that the group reconvene each five years to determine whether new substances
need to be covered or whether standards need to be tightened.64 Currently, the framework is
implemented under two sets of emissions standards, one standard covering SO2, NOx, and
particulate matter (PM), and another standard covering mercury.
Emissions of SO2, NOx, and PM are covered under the Alberta Air Emissions Standards for
Electricity Generation. These standards set out maximum emissions rates for new units, after the
50th operating year for coal units, after the 40th operating year for most natural gas units, and
63
64
See CASA (2003).

See CASA (2010), p. 1; Alberta Environment (2010c).
25
after the 60th operating year for gas peaking units.65 The allowed emissions rates are set based
on a determination of the best available technology economically achievable (BATEA), which
may improve over time and will therefore result in more strict emissions standards over time.
These emissions standards somewhat increase the costs of building new units by requiring that
new generators have pollution controls. For aging units past their design life, these standards are
likely to require a retrofit installation of emissions controls for the unit to continue operating.
For many units, the costs associated with these upgrades may be too high relative to potential
going-forward operating margins to remain viable. For this reason, these control standards could
force some aging units to retire before these controls would need to be installed.
Figure 6 shows the timeline over which the existing AESO coal and gas fleet will be subjected to
these SO2, NOx, and PM standards. This 50-year coal retirement timeline corresponds to a 5year delay relative to when the federal coal retirement mandate would force coal plants to retire.
For this reason, if the federal coal mandate is enacted, the Alberta air quality standard will have
no incremental effect on coal retirements or resource adequacy. In either case, the standard will
have an incremental impact on natural gas units past their 40th operating year or past the 60th
operating year for peaking units. Overall, these standards could force 1,630 MW of coal and 460
MW of natural gas plant retirements by 2029.66 While the quantity of capacity affected is large,
the resource adequacy impact is likely to be very limited because of the gradual timeline and the
imposition of standards only on older units that are already past their design life.
Albertas Mercury Emissions from Coal-Fired Power Plants Regulation is an additional
emissions standard based on output levels under BATEA.67 However, the mercury standard is
imposed on all existing generators at the same time regardless of the unit age. In order to
comply with the mercury standard, coal generators had to meet the following requirements by
the first of the year:
Continuous monitoring equipment installed by January 1, 2010
70% mercury capture by January 1, 2011
80% mercury capture by January 1, 2013
The mercury standard was implemented with some flexibility that allowed some older units to
avoid installing mercury controls as long as they committed to retiring by unit-specific deadlines
over the 2012-17 timeframe.68 Of these, Sundance 1 and 2 may already be considered retired for
unrelated reasons, but all other coal facilities in Alberta will meet the mercury standard.69 The
recent CASA review contained recommendations for increasing flexibility in meeting the
requirement by allowing credits for early reductions that could later be used at the same facility
65
66
67
68
69
Peaking units are regulated based on an annual total emissions limit that assumes approximately 1500 MW
of operation per year, see CASA (2010), Sections 3 and 6; Alberta Environment (2005).
Note that Sundance 1 and 2 are excluded from this total as they may have already retired, pending
determination of whether repowering will occur.
See Alberta Government (2006).
Specifically, HR Milner would have had to retire by 2012, Battle River 3 and 4 by 2015, and Sundance 1
and 2 by 2017. See Alberta Government (2006), p. 4.
Confirmed via personal communication with Alberta Environment staff director of the mercury program.
Note that Sundance 1 and 2 will retire early in advance of the air quality mandate because of unrelated
large investment costs that would be required for continued operation. See TransAlta (2011).
26
if it had equipment problems.70 Overall, it appears that the mercury standard has been
successfully implemented without imposing any resource adequacy concerns and even without
imposing any incremental retirements.
Figure 6
Capacity Subject to Provincial Air Quality Emissions Standards
3.0
Gas Peakers
Gas
2.5
Capacity (GW) &
(Non-Peakers)
2.0
Coal
1.5
1.0
0.5
Sundance 1 & 2
Retire Prior to Air
Quality Mandate
2029
2028
2027
2026
2025
2024
2023
2022
2021
2020
2019
2018
2017
2016
2015
2014
2013
2012
2011
2010
Sources and Notes:

Peaking units identified as units operating at less than 17% capacity factor, see p. 4, Alberta Environment (2005) and AESO (2010c).
Unit online date and MW rating from Ventyx (2010), AESO (2010c), and Alberta Environment (2005).
E. W IND I NTEGRATION AND A NCILLARY S ERVICES

Wind generation capacity in Alberta has increased quickly over the past decade, from 30 MW in
2000 to 630 MW in 2010 as shown in Figure 7. These increases in wind capacity can be
expected to continue, with 240 MW of wind currently under construction and another 1300 MW
either permitted or proposed. While many of the proposed or even permitted projects may never
get built, they do indicate a high level of continued investor interest. The large increases in wind
penetration have been driven by a variety of policies subsidizing wind. For example, wind
suppliers are eligible to create and sell Alberta-based GHG offsets under the Specified Gas
Emitters Regulation discussed in Section III.C.1.c and are now able to sell renewable energy
70
Assuming the recommendations are adopted, 50% of the early reductions above 75% capture would earn a
credit starting 2011, while 50% of reductions above 80% would earn a credit starting 2013. These credits
could not be used to delay controls upgrades or transferred to other facilities, but they could be used to
offset excess emissions caused by maintenance or operational issues. All credits would expire by the end
of 2015. See CASA (2010), p. 4.
27
certificates to allow utilities in California to meet the states ambitious renewable energy
standards.71
Figure 7
Historic and Potential Future Wind Capacity Growth
2.5
Wind Nameplate Capacity (GW) &
Historic Projected
2.0
Proposed Additions
1.5
1.0
Permitted Additions
Under Construction
0.5
Historic Additions
Winter 2000/01 Capacity
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
0.0
Sources and Notes:

Unit online and retirement dates and MW rating from Ventyx (2010) and AESO (2010c).
Units under construction, permitted, and proposed from AESO (2010d).
Large wind penetration levels can introduce a variety of operational challenges as the system
operator must develop wind forecasting capability and operate the power grid with a highly
intermittent generation resource. The risk of sudden drop-off in wind output increases the need
for additional operating reserves. Unexpectedly high wind output during low load periods can
also create operational challenges by creating minimum generation conditions in which market
prices are zero, baseload generators are operating at minimum output, and the system operator
must order further involuntary generation reductions or shutdowns. These operational challenges
are the subject of ongoing market design effort by AESO and stakeholders to address increasing
wind penetration in the near term and longer term.72
High levels of wind generation can also introduce long-term resource adequacy challenges. Due
to intermittent output levels, wind resources have very little capacity value during peak load
conditions. Albertas capacity factor during peak times is higher than in many other systems,
simply because Albertas peak load and highest wind season both occur in winter. In fact, the
wind capacity factor is about 41% over November-January, which is much higher than the
71
72
See Herndon (2011).

See AESO (2010b); AESO (2010h).
28
approximate 29% annual capacity factor..73 However, this capacity factor substantially
overstates the capacity value of wind, because wind is not firm supply and will be unavailable
periodically despite relatively high average monthly output. For example, Midwest ISO studies
have shown that only 8% of a wind turbines nameplate capacity can be reliably counted toward
the overall system installed capacity although the wind fleet has a 27% average capacity factor.74
While not contributing substantially to system adequacy, wind generation does have a large
impact on the energy market because it enters the supply stack at zero (or even negative)
marginal cost. These negative marginal costs can arise if suppressing power output during high
wind conditions causes lost revenues from renewable energy credits (RECs) or if it imposes
additional O&M costs to slow turbine speeds. Wind generation consequently tends to depress
average energy prices and reduce the net revenues received by other generators, making them
more likely to retire and potentially making it less likely that new resources are built. Figure 8
shows the short-term price impact of wind output fluctuations, by separately showing the price
duration curves for high-wind and low-wind hours during 2008-10. The figure shows that lowwind hours with less than 100 MW of wind output had an average price of $77/MWh, while
high-wind hours with more than 400 MW of wind had an average price of $42/MWh. However,
this does not mean that 300 MW of wind can suppress average prices by more than $30/MWh,
because the analysis does not control for factors such as natural gas price changes, time of day,
or the difference between forecasted and realized wind output. Nevertheless, the figure
highlights the importance of monitoring and further analyzing the potential price-suppressing
effects of additional wind investments.
The lower energy prices during high wind events do not mean that energy market prices will
need to be artificially propped up or otherwise revised. In fact, low or even negative hourly
prices during high wind hours correctly represent the short-run marginal cost of supply at those
times, which is the efficient energy price signal at these specific instances in time. In fact,
negative prices would enhance market efficiency by creating an additional incentive for wind
and other suppliers to ramp down or for load to ramp up during high wind events, making
flexibility more valuable. While some market participants may fear that negative prices will
undermine overall incentives for conventional generation investment, we believe that this will
not be the case as long as ancillary services requirements and operational requirements on wind
suppliers are carefully designed. The overall market impact of increased wind integration should
be to increase the value of flexible generation and demand response relative to inflexible
generation.
The immediate-term effect of wind generation-related price suppression may be to replace more
traditional resources that have high capacity value with wind resources that have very little
capacity value, reducing the system reserve margin. The lower system reserve margin, however,
would increase price spikes in response to low-wind conditions. This will tend to increase
average prices and price volatility, but will also prevent further deterioration of reserve margins
by making it more attractive to build flexible generating resources that can take advantage of the
higher prices and price volatility.
73
74
Calculated from hourly wind data and wind installed capacity data from AESO (2010c).
See Midwest ISO (2009), pp. 3, 13.
29
Figure 8
Price Duration Curve at Different Levels of Wind Output
$500
$450
$400
AESO Price ($/MWh)
$350
$300
$250
$200
Higher Prices
Wind < 100 MW
$150
$100
All Hours
$50
Lower Prices
Wind > 400 MW
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
$0
Percent of Hours with Each Wind Level

Sources and Notes:
AESO analysis of hourly wind and price data over Jan. 2008 through Nov. 2010.
Finally, increased wind generation will also increase the need for operating reserves. If
additional reserves requirements are instituted, flexible resources will become, again, more
valuable because of their ability to provide operating reserves. The Alberta generation fleet may
also have some additions from less flexible, new baseload generation sources, such as new
cogeneration for the oil sands industry and coal plants fitted with CCS. This means even if
resource adequacy can be maintained, the added wind generation may create system operations
challenges. This added challenge will require continued close attention to current market design
efforts to facilitate the integration of additional wind resources.75
The quantity of reserves that are currently required are somewhat variable, but the average level
of operating reserves scheduled over 2009 is shown in Table 5. The table also shows the total
fleet capability for supplying operating reserves of each type, based on AESOs qualified
provider list. Note that the total capability for reserves is less than the sum of the capability for
each individual type of reserves because suppliers cannot supply their maximum capability for
more than one type of reserves at a time. Overall, the fleet capability is 6 times the currently
75
See AESO (2010b); AESO (2010h).
30
required level of regulating reserves, 9 times the required spinning reserves, 12 times the
required supplemental reserves, and 4 times the total simultaneously required reserves.
Table 5
Operating Reserves Need and Fleet-Wide Capability
Average Scheduled in 2009
Active
Standby
Total
MW
MW
MW
Fleet-Wide
Capability
MW
Regulating
Spinning
Supplemental
160
243
243
131
106
37
290
350
280
1,785
3,148
3,502
Total
646
274
920
3,780
Sources and Notes:

AESO (2010c).
AESO has examined the potential for mitigating wind variability by increasing regulating
reserves capability, among other options. In a year 2020 scenario with 4,000 MW of installed
wind capacity, the AESO analysis found that an additional 300 MW of regulating reserves could
mitigate approximately half of the Area Control Error (ACE) events, although 2,000 MW of
regulating reserves would be required to resolve 98% of the events.76 An increase in regulating
reserves of this magnitude would be difficult to achieve, since it is higher than the currently
installed capability. However, it is likely within a potential feasible range by the time overall
growth in the generation fleet by 2020 is accounted for. Further, it is likely that much of the
difficulty with wind variability can be mitigated with other options that AESO is considering
including acquiring additional operating reserves from demand response and placing additional
requirements on wind generators.
Finally, if AESO increases reserves requirements, market prices for reserves are also likely to
increase and help attract incremental reserves. If resources that can provide operating reserves
become scarce relative to increasing demand for such reserves, then prices for reserves will tend
to rise relative to the price of energy. This should support the entry by resources that can provide
operating reserves beyond what is reported in Table 5 High reserves prices could even attract
additional reserves supply from the existing fleet as some suppliers may choose to invest in
upgrading their reserves capability, for example, by adding active generation controls (AGC) that
would allow them to supply regulation.
The total system capability for supplying operating reserves can also be increased through the
integration of demand-side resources. This could be achieved through market designs that
reduce barriers that limit the participation by demand-side resource in energy markets and
operating reserves.77
76
77
See AESO (2010h), pp. 15-16.

This option is discussed in a forthcoming Brattle study for AESO, Demand Response Review, by Johannes
Pfeifenberger and Attila Hajos.
31
F. E XPANDED I NTERCONNECTIONS WITH N EIGHBORING M ARKETS

Alberta currently is only weakly interconnected with neighboring systems, but the Alberta
government, in its Provincial Energy Strategy, has set out a policy objective of expanding
interties with neighboring markets. By expanding these interconnections, the government aims
to increase reliability, supply adequacy, market competitiveness, and access to wind
generation.78 However, expanding interconnections to neighboring markets, all of which have
resource adequacy requirements, also introduces risks which must be monitored carefully. This
includes the possibility that the interaction with external markets could depress Alberta market
prices and deter needed investment in new resources, thereby decreasing long-term supply
adequacy and reliability.
The Alberta electric system currently has two major interties, one with BC Hydro and the other
with Saskatchewan. The BC Hydro intertie is currently operating with available transfer
capability (ATC) less than its design capacity due to Alberta-internal transmission constraints
and other operational restrictions.79 The BC Hydro intertie has a design rating of 1,200 MW for
imports and 1,000 MW for exports, but currently has a maximum ATC value of only 650 MW
for imports and 735 MW for exports.80 The Saskatchewan intertie design rating is 150 MW for
imports and exports and its ATC has recently been restored to its design rating.81 The ATC on
the BC Hydro intertie is also anticipated to be restored to its original design ratings after the
creation of intertie restoration products including load shed service and other system
enhancements.82
In addition to restoring intertie ATC to design rating, there is one intertie project that will further
expand Albertas interconnections with neighboring markets, although it will not necessarily
increase ATC. This new intertie is the 300 MW Montana-Alberta Tie Limited (MATL) line
that is currently under construction with an estimated online date in late 2011. In addition, the
AESO has begun considering several other potential interconnection options, although specific
projects have not yet been determined.83
Expanded interconnections increase market efficiency by allowing more power to flow from
locations with lower-priced supplies to locations with higher-priced supplies. These increased
transmission flows tend to reduce price differentials between regions by increasing prices and
supplier profits in the lower-priced locations while decreasing prices and supplier profits in high-
78
79
80
81
82
83
See AESO (2009a), pp. 10, 14, 26-27.

The north-south transmission constraint between Calgary and Edmonton is the primary constraint reducing
intertie ATC values to below design ratings. Both the BC Hydro and Saskatchewan interties are connected
south of AESOs north-south transmission cut plane, which places a limit on the total imports and exports
that can be scheduled. The loss of an intertie would represent too large a contingency for the system south
of the cut plane to absorb by itself. From communication with AESO staff and AESO (2009a), Section
4.3.
Current value from AESO staff; design rating from AESO (2009a).
Current value from AESO staff; design rating from AESO (2009a), p. 303.
Based on communication with AESO staff and AESO (2009a), p. 39.
The expansion paths being considered include: southern Alberta to Saskatchewan and Manitoba, southern
Alberta to the US Pacific Northwest, northern Alberta to northern BC, and northern Alberta to northern
Saskatchewan. See AESO (2009a), p. 36 and Section 4.9.3.
32
priced locations. Because Alberta is typically an exporter at night and an importer during the
day, expanded interconnections may increase off-peak prices while decreasing on-peak prices.84
The combined impact from both effects is likely to be a suppression of Alberta prices overall,
given that Alberta prices are higher on average than are those of its neighbors as shown in Figure
9. Part of the reason for the lower energy prices in neighboring markets is that these other
markets are cost-of-service regulated, meaning that suppliers recover the capacity costs of their
generation through regulated retail rates or through public ownership as in British Columbia
rather than solely through wholesale market prices for energy. Other more distant markets, such
as California, have resource adequacy requirements that allow suppliers to earn capacity
payments through a bilateral market to supplement their energy market revenues.85 As a
consequence, the energy prices of neighboring regions typically reflect only short-term
generation dispatch costs without sufficient contributions to recover investment costs. This is in
contrast to AESO, where suppliers need to recover their investment costs through the wholesale
energy and ancillary service markets.86
Price suppression in Albertas energy-only market through expanded interties is likely to be
magnified by an increase in imports from zero-marginal-cost technologies, such as new wind
generation. For example, the MATL developer has predicted that adding the new transmission
line will allow for the development of a large wind farm at its source in Montana.87 Similarly,
increased intertie capacity with BC Hydro will interconnect Alberta more heavily with a market
that is planned to become a large exporter of green power, likely from wind and hydro, which
could further depress Alberta energy prices because of lower energy prices in British
Columbia.88
These low-marginal-cost imports would directly benefit Alberta customers in the near term.
However, in the long term, the price suppression would reduce the profitability of generation
resources in Alberta, which would make it less likely that new resources would be built while
increasing the likelihood that existing generators would retire prematurely. These impacts could
tend to reduce the reserve margin within Alberta and make the system more dependent on the
interties for resource adequacy.
84
85
86
87
88
See AESO (2009a), p. 50.

See CPUC (2006); Blakes (2008), Section VI.
See Section II.A for more discussion of the difference between energy-only markets and markets with
resource adequacy standards.
See Puckett (2009).
See BC Government (2010), Section 2.n.
33
Figure 9
Monthly On-Peak Energy Prices in Alberta, Northern California, and Mid-Columbia
Sources and Notes:

Bloomberg (2010); Ventyx (2010); and UCEI (2003).
Prices for North Path 15 (NP15) are zonal prices until April 2009, after which NP15 Gen Hub is shown.
NP15 and Mid C prices converted to CAD from USD, Bloomberg (2010).
While expanded interconnections increase the capability for importing power, they do not
guarantee that external supplies will be available for import when they are needed most. This is
because generators in neighboring markets tend to be obligated to supply their local customers
during peak load conditions. If such peak load or emergency conditions occur simultaneously in
Alberta and neighboring markets, Alberta will not be able to import the needed supplies no
matter how much intertie capacity is available and no matter how high the AESO price. It also
means that the resource adequacy value of the increased interties is limited to a probabilistic
value that depends on the extent to which neighboring markets are over-built beyond their own
resource adequacy requirements and the extent to which those markets experience system peak
conditions at times different from Alberta.
In addition, shortage conditions in neighboring markets can introduce supply shortages in
Alberta. Because Alberta generators do not have the obligation to serve Alberta load, they have
the option to export power to neighboring systems even during peak conditions and would
rationally choose to do so any time external power prices are higher.89 The converse is not true,
however. Because generators in neighboring markets will typically not be able to sell power into
89
This impact applies to peak conditions only prior to the initiation of emergency procedures, as AESO will
intervene to curtail exports to zero during peak load conditions. See AESO (2010f).
34
Alberta during emergencies, this means that neighboring markets are largely insulated from any
resource adequacy challenges in Alberta.90
The scale of the impact that shortages in external markets may have on Alberta can be gauged to
some extent by examining the level of correlation that already exists between prices in Alberta
and those in neighboring markets. Figure 9 shows electricity prices in Alberta compared with
electricity prices in Northern California and Mid-Columbia in Washington, the two closest liquid
power trading hubs with transparent market prices. While the figure shows a strong correlation
among the electric prices in the three locations, this is somewhat misleading since a significant
portion of the common price movements starting in 2001 have been driven by changes in the
regional price for natural gas.91
The extreme high prices in the year 2000 affecting all three locations were caused by the
California power crisis.92 These extreme price conditions during 2000 were driven partly by
tight supply conditions and partly by market power abuses. The market power abuses have been
the subject of substantial litigation, investigation, and damages settlements before the Federal
Energy Regulatory Commission (FERC) in the United States, and also have been investigated
for their impact on Alberta by the Market Surveillance Administrator (MSA).93 The
experience does highlight the point that real and even manipulated shortages in neighboring
markets can substantially impact Alberta. Expanding interconnections will further increase
AESOs exposure to the market fundamentals and potential shortages in neighboring markets in
the future. At the same time, as Figure 9 also shows, the price spikes that occurred in the Alberta
energy market since 2006, with proximate causes often related to short-term supply adequacy
problems, had virtually no impact on the market prices in these neighboring regions.
These factors mean that the expansion of interconnections with neighboring markets will require
the AESO to increase the extent to which it monitors for potential shortage conditions, including
assigning a realistically low capacity value to total import capability. It also means that the
AESO should maintain its procedures for limiting exports during scarcity conditions, and not
introduce firm export transmission service without careful consideration of the potential resource
adequacy consequences.94
90
91
92
93
94
If Alberta had the potential to become a large net exporter of power, then increased interties would have
the potential to increase reliability in Alberta by incenting generation build-out in excess of supply needs
to meet export demand. This scenario would only materialize if Alberta had structural potential for lowercost energy market prices over the long run, even after accounting for the capital cost recovery required
through Albertas energy market (compared to Albertas neighbors which award cost recovery outside the
energy market).
Excluding the year 2000, the R2 values of predicting AESO prices from Mid-Columbia and Northern
California power prices are 0.122 and 0.221 respectively, while AESO, Mid-Columbia, and Northern
California power prices have R2 values of 0.227, 0.119, and 0.302 when predicted by gas prices.
Natural gas prices were also high during the year 2000, but no higher than in other years with moderately
high electric prices such as 2005.
While the investigation into the bidding strategies and intertie conduct of Enron Canada Corporation and
Powerex Corporation did not uncover prohibited behaviors, it did result in the revision of rules governing
intertie conduct that would prevent those behaviors in the future. See MSA (2005); FERC (2010).
No similar concern would be introduced by allowing for firm import capability, which would allow for the
possibility that suppliers would have the firm option to sell into Alberta. However, only external suppliers
35
LIST OF ACRONYMS
ACE
Area Control Error
AECO
Alberta Energy Company
AESO
Alberta Electric System Operator
AEMC
Australian Energy Market Commission
AIES
Alberta Internal Electric System
AIL
Alberta Internal Load
ATC
Available Transfer Capability
AUC
Alberta Utilities Commission
AUD
Australian Dollars
BATEA
Best Available Technology Economically Achievable
BAU
Business as Usual
CAD
Canadian Dollars
CAISO
California Independent System Operator
CASA
Clean Air Strategic Alliance
CC
Combined Cycle
CCEM
Climate Change Emissions Management
CCS
Carbon Capture and Sequestration
CH4
Methane
CO2
Carbon Dioxide
CO2e
Carbon Dioxide Equivalent
CONE
Cost of New Entry
CPUC
California Public Utilities Commission
CT
Combustion Turbine
DDS
Dispatch Down Service
DR
Demand Response
EDC
EDC Associates, Ltd.
EIA
Energy Information Administration
ERCOT
Electric Reliability Council of Texas
FERC
Federal Energy Regulatory Commission
FOM
Fixed Operations and Maintenance
FRB
Federal Reserve Board
GHG
Greenhouse Gas
75
GJ
Gigajoules
GW
Gigawatts
GWh
Gigawatt-hours
IPCC
Intergovernmental Panel on Climate Change
ISO
Independent System Operator
ISO-NE
ISO New England
kT
Kilotonnes
MATL
Montana-Alberta Tie Limited
MT
Megatonnes
MW
Megawatts
MWh
Megawatt-hours
MSA
Market Surveillance Administrator
N2O
Nitrous Oxide
NOX
Mono-nitrogen Oxides
NEM
National Electricity Market
NEPOOL
New England Power Pool
NERC
North American Electric Reliability Corporation
NP15
North Path 15
NYPP
New York Power Pool
O&M
Operations and Maintenance
PJM
PJM Interconnection, LLC
PM
Particulate Matter
PPA
Power Purchase Arrangement
PRD
Price-Responsive Demand
SO2
Sulfur Dioxide
SPP
Southwest Power Pool
SWIS
South West Interconnected System
TMR
Transmission Must Run
VCA
Voluntary Capacity Auction
VOLL
Value of Lost Load
76
APPENDICES
77
A. G ENERATOR O PERATING M ARGINS VERSUS F IXED C OSTS

Figure 32 through Figure 36 show the estimated generator operating margins and fixed costs
over the past decade for natural gas CCs, natural gas cogen, hydro, wind, and natural gas CTs
including TMR units. A discussion of the implications of this information is in Section V.A.6,
along with similar figures for coal units and gas CTs without TMR units.
Figure 32
Historic Gas CC Operating Margins vs. Fixed Costs
Operating Margins and Fixed Costs ($/kW-y) &
$250
DDS
Operating Reserves
$200
Energy Margins
$150
Cost of New Plant

$100
$50
Fixed O&M
$0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Sources and Notes:
Energy margins represent revenues minus estimated operating costs in energy market. Cost of New Plant includes capital costs and FOM.
Unit-specific volumes and revenues as well as 2010 VOM, CONE, and FOM by unit type are from AESO (2010c).
Gas prices and exchange rates from Bloomberg (2010). Heat rates estimated from Ventyx (2010), AESO (2010c), Alberta Environment (2010a-b).
Historic CONE and FOM numbers are inflated according to the Handy-Whitman Index (converted from USD to CAD) between 2000 and 2009 from
Whitman, et al. (2008) andPJM (2009); and by inflation between 2009 and 2010 from Bank of Canada (2010).
A-1
Figure 33
Historic Gas Cogen Operating Margins vs. Fixed Costs
Operating Margins and Fixed Costs ($/kW-y)
$250
Operating Reserves
Energy Margins
$200
$150
Cost of New Plant

$100
$50
Fixed O&M
$0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Sources and Notes:
Figure 34
Historic Hydro Operating Margins vs. Fixed Costs
$400
Other AS
$350
TMR
$300
Supplemental
Spinning
Cost of New Plant
$250
Regulating
Energy Margins
$200
$150
$100
$50
Fixed O&M
$0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Sources and Notes:
Gas prices and exchange rates from Bloomberg (2010).
A-2
Figure 35
Historic Wind Operating Margins vs. Fixed Costs
Cost of New Plant
$250
$200
Energy Margins
$150
$100
$50
Fixed O&M
$0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Sources and Notes:
Gas prices and exchange rates from Bloomberg (2010).
Figure 36
Historic Gas CT Operating Margins vs. Fixed Costs (Including TMR Units)
Other AS
$200
$150
TMR
Cost of New Plant
$100
Operating Reserves
$50
Energy Margins
Fixed O&M
$0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Sources and Notes:
Whitman, et al. (2008) and PJM (2009); and by inflation between 2009 and 2010 from Bank of Canada (2010).
A-3
B. M ETHOD F OR P ROJECTING O PERATING M ARGINS

Section V.A.6 describes our approach to projecting future operating reserves revenues and
energy margins as a linear function of perfect dispatch margins that could be achieved by a
plant with no startup costs, outages, or dispatch constraints. The parameters of these linear
relationships are shown in Table 8.
Table 8
Energy Margins and Operating Reserves Revenue versus Perfect Dispatch Margins
(Linear Relationships Based on Historic Monthly Data)
Energy Margins vs. Perfect Dispatch Margins
a * (perfect energy margins) + b = (actual energy margins)
Coal
Gas Cogen
Gas CC
Gas CT
Hydro
Wind
b
($/kW-yr)
R2
0.726
0.667
0.748
0.484
0.295
0.287
19.86
-21.68
-42.91
-26.91
-10.81
n/a
0.710
0.817
0.918
0.655
0.440
n/a
Reserves Revenues vs. Perfect Dispatch Margins

a * (perfect energy margins) + b = (OR revenues)
Coal
Gas Cogen
Gas CC
Gas CT
Hydro
Wind
b
($/kW-yr)
R2
0.001
0.033
0.031
0.209
0.290
0.000
0.219
-0.076
2.877
2.500
-60.828
0.000
0.030
0.081
0.104
0.167
0.628
n/a
Sources and Notes:

Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c).
Wind is calculated as a simple percentage.
These relationships were developed based on the historic relationship between historic energy
margins and operating reserves revenue calculated from AESO internal data as described in
Section V.A.6 and a theoretical back-cast of perfect dispatch margins. These data are
represented at the unit level for each month from January 2008 through October 2010. Figure 37
through Figure 42 are scatter plots of the data used to determine these linear relationships. Note
that some data points show zero historic energy margins, which is an indication that the unit was
on outage during that month.
B-1
Figure 37
Historic Gas CT Operating Margins vs. Perfect Dispatch Margins
Energy
Actual Energy Margins ($/kW-yr) &
$800
y = 0.484x
+ -26.91
Energy
R-Squared: 0.7095
$700
100%
100%
$600
$500
$400
$300
$200
$100
Opeating Reserves
$0
OR Revenue ($/kw-yr)
&&
$0
$800
$100
$200
$300
$400
$500
$600
$700
$800
Operating Reserves
$600
y = 0.209x + 2.5
R-Squared: 0.1666
$400
100%
$200
$0
$0
$100
$200
$300
$400
$500
$600
$700
$800
Perfect Dispatch Energy Margins ($/kW-yr)

Sources and Notes:
Figure 38
Historic Gas CC Operating Margins vs. Perfect Dispatch Margins
$1,000
Energy
0.484x + 42.914
-26.91
yy==Energy
0.748x
R-Squared: 0.7095
R-Squared:
0.9177
$900
$800
100%
100%
$700
$600
$500
$400
$300
$200
$100
Opeating Reserves
$0
&&
$0
$1,000
$800
$600
$400
$200
$0
$100
$200
$300
$400
$500
$600
$700
$800
$900
$1,000
Operating Reserves
y = 0.031x + 2.877
R-Squared: 0.104
$0
$100
100%
$200
$300
$400
$500
$600
$700
$800
$900

Sources and Notes:
B-2
$1,000
$1,000
Figure 39
Historic Gas Cogen Operating Margins vs. Perfect Dispatch Margins
Energy
y =Energy
+ -26.91
y 0.484x
= 0.667x
-21.684
$900
100%
100%
R-Squared:
0.7095
R-Squared:
0.8174
$800
$700
$600
$500
$400
$300
$200
$100
Opeating Reserves
$0
&&
$0
$1,000
$800
$600
$400
$200
$0
$100
$200
$300
$400
$500
$600
$700
$800
$900
$1,000
Operating Reserves
y = 0.033x + -0.076
R-Squared: 0.0805
$0
$100
100%
$200
$300
$400
$500
$600
$700
$800
$900

Sources and Notes:
Figure 40
Historic Coal Operating Margins vs. Perfect Dispatch Margins
Energy
y =Energy
+ -26.91
y0.484x
= 0.726x
+ 19.865
$1,000
100%
100%
R-Squared:
0.7095
R-Squared:
0.7095
$800
$600
$400
$200
Opeating Reserves
$0
&&
$0
$200
$400
$600
$800
$1,000
Operating Reserves
$1,000
$800
$600
$400
$200
$0
y = 0.001x + 0.219
R-Squared: 0.0302
$0
$200
100%
$400
$600
$800
$1,000

Sources and Notes:
B-3
$1,000
Figure 41
Historic Hydro Operating Margins vs. Perfect Dispatch Margins
$1,200
Energy
0.484x + -26.91
yy==Energy
0.295x
10.806
100%
100%
R-Squared: 0.7095
$1,000
R-Squared: 0.9177
$800
$600
$400
$200
Opeating Reserves
$0
&&
$0
$1,200
$1,000
$800
$600
$400
$200
$0
$200
$400
$600
$800
$1,000
$1,200
Operating Reserves
y = 0.29x + -60.828
R-Squared: 0.6277
$0
$200
100%
$400
$600
$800
$1,000
$1,200

Sources and Notes:
Figure 42
Historic Wind Operating Margins vs. Perfect Dispatch Margins
$1,000
Energy
yy ==
0.484x
0.287x++-26.91
0
Energy
$900
100%
100%
R-Squared:
R-Squared:0.7095
n/a
$800
$700
$600
$500
$400
$300
$200
$100
Opeating Reserves
$0
&&
$0
$1,000
$800
$600
$400
$200
$0
$100
$200
$300
$400
$500
$600
$700
$800
$900
$1,000
Operating Reserves
y = 0x + 0
R-Squared: n/a
$0
$100
100%
$200
$300
$400
$500
$600
$700
$800
$900
$1,000

Sources and Notes:
B-4
C. P ROJECTION OF G ENERATOR O PERATING M ARGINS VERSUS F IXED C OSTS

Figure 43 through Figure 46 show a projection of future generator operating margins and fixed
costs under baseline CO2e and gas assumptions using historic heat rates from 2006-10. The
figures show results for natural gas CCs, natural gas cogen, hydro, and wind power plants. A
discussion of the implications of this information is in Section V.B.2, along with similar figures
for coal units and gas CTs.
Figure 43
Projected Gas CC Operating Margins vs. Fixed Costs
Sources and Notes:

Future price duration curve is calculated based on heat rates from 2006-10 and baseline gas and CO2e price forecasts.
Cost of new plant includes FOM and real levelized capital costs from Section V.A.5, future escalation at 2.4% annually.
C-1
Figure 44
Projected Gas Cogen Operating Margins vs. Fixed Costs
Sources and Notes:

Figure 45
Projected Hydro Operating Margins vs. Fixed Costs
Sources and Notes:

C-2
Figure 46
Projected Wind Operating Margins vs. Fixed Costs
Sources and Notes:

C-3
AN INTRODUCTION TO AUSTRALIAS NATIONAL ELECTRICITY MARKET

JULY 2010
Disclaimer
This document is made available to you on the following basis:
(a) Purpose - This document is provided by the Australian Energy Market Operator Limited
(AEMO) to you for information purposes only. You are not permitted to commercialise it or
any information contained in it.
(b) No Reliance or warranty - This document may be subsequently amended. AEMO does
not warrant or represent that the data or information in this document is accurate, reliable,
complete or current or that it is suitable for particular purposes. You should verify and check
the accuracy, completeness, reliability and suitability of this document for any use to which
you intend to put it and seek independent expert advice before using it, or any information
contained in it.
(c) Limitation of liability - To the extent permitted by law, AEMO and its advisers, consultants
and other contributors to this document (or their respective associated companies,
businesses, partners, directors, officers or employees) shall not be liable for any errors,
omissions, defects or misrepresentations in the information contained in this document, or
for any loss or damage suffered by persons who use or rely on such information (including
by reason of negligence, negligent misstatement or otherwise). If any law prohibits the
exclusion of such liability, AEMOs liability is limited, at AEMOs option, to the re-supply of
the information, provided that this limitation is permitted by law and is fair and reasonable.
2010 - All rights reserved.
ELECTRICITY
3
The National Electricity Market
The Australian Energy Market Operator
National Electricity Law and Rules
The Spot Market
Key Parameters for NEM Operation
Operating the NEM
Ancillary Services
14
Inter-regional Trade
15
Market Forecasts
17
Full Retail Competition
19
Registered Participants
19
Financial Contracts for Electricity
20
Alternative Generation Technologies
22
AEMO and the Environment
22
Regulatory Arrangements
23
Glossary
24
NATIONAL ELECTRICITY MARKET
The Electricity Supply Industry
CONTENTS
1
THE ELECTRICITY SUPPLY INDUSTRY

Sectors of the electricity supply industry are involved
with the generation, transmission, distribution and
retail sale of electricity. Australias social, industrial and
commercial success depends on the reliability of the
electricity supply. In this way, the industry contributes
significantly to the national economy.
What is Electricity?
Electricity is a form of energy produced by the flow of
electrons in a substance known as a conductor. The best
conductors are metals such as copper and aluminium,
and are commonly used in electrical wiring.
Energy exists in many forms. Electricity is a secondary
energy source as it is produced by the conversion of other
energy sources like the chemical energy in coal, natural
gas and oil. Other primary sources of energy, like the
sun and wind, are increasingly being used to produce
electricity. A quantity of energy can be changed or
converted, but can never be created or destroyed.
Electricity can be converted readily to heat and light

and used to power machines. It can also be transported
with relative ease. These characteristics make electricity
a convenient and manageable form of energy, and
contribute both to its value as a commodity and its
versatility as a source of power.
A unit of power is referred to as a watt. The number of
watts, or wattage, of an electrical appliance indicates the
rate at which the appliance converts electrical energy to
another form of energy such as heat or light. One watt
is equivalent to one joule of work per second. Both the
electrical pressure (voltage) and the number of electrons
flowing (current) determine the electrical power or rate
of energy conversion. A 60-watt light globe uses 60 watts
of electricity to produce light, and a typical electric kettle
uses 2400 watts to produce heat.
How is Electricity Produced?

Electricity can be produced by either chemical means or
mechanical action. Electricity produced by chemical means
relies on a flow of charged particles from cells in a battery.
While this type of electricity has some very important
applications in modern society, it is an expensive
production process and can meet only limited, specific
requirements for electricity. The generators in modern
power stations produce electricity by the mechanical
action of large, powerful magnets that spin rapidly inside
the huge coils of conducting wire driven by steam, gas
or water turbines.
More than 90 per cent of Australias electricity production
relies on the burning of fossil fuels - coal, gas and oil.
The chemical energy stored in these fuels is used to heat
water and produce steam. The steam is then forced under
great pressure through a turbine that drives a generator
to produce electricity. The complete process involves
the conversion of chemical energy to kinetic energy to
electrical energy. In a similar way, the kinetic energy of
falling water drives turbine blades to produce electrical
energy at a hydro-electricity plant, and the kinetic energy
of wind drives the blades of a wind-power turbine to
produce electricity.
UNITS EXPLAINED
One megawatt (MW)
is equal to one million
watts (W).
One gigawatt (GW)

is equivalent to one
thousand megawatts.
One megawatt hour

(MWh) is the energy
required to power ten
thousand 100 W light
globes for one hour.
A 100 MW generator
will power one million
100 W light globes
simultaneously.
A 600 MW generator
has sufficient capacity
to service 200,000
domestic customers.
THE ELECTRICITY SUPPLY INDUSTRY CONTINUED
How is Electricity Transported?

Transmission lines
carry electricity
long distances
Electricity travels along a conductor at close to the

speed of light. When an appliance is switched on, power
is instantly transmitted from a power station to the
appliance. Although this occurs instantaneously, a specific
sequence of events takes place to ensure the delivery
of the required electricity.
A transformer converts the electricity produced at a
generation plant from low to high voltage to enable its
efficient transport on the transmission system. When the
electricity arrives at the location where it is required, a
substation transformer changes the high voltage electricity
to low voltage for distribution. Distribution lines then carry
low voltage electricity to consumers who access it through
the power outlets in homes, offices and factories.
Power plant
generates
electricity
Transformer
converts low voltage
electricity to high
voltage for efficient
transport
Distribution lines
carry low voltage
electricity to
consumers
Substation transformer
converts high voltage
electricity to low voltage
for distribution
Homes, offices and

factories use electricity
for lighting and
heating and to power
appliances
TRANSPORT OF ELECTRICITY
Energy exists in many forms. Electricity is a secondary energy

source as it is produced by the conversion of other energy
sources like the chemical energy in coal, natural gas and oil.
Other primary sources of energy, like the sun and wind,
are increasingly being used to produce electricity.
3
THE NATIONAL ELECTRICITY MARKET

The National Electricity Market (NEM) began operating as
a wholesale market for the supply of electricity to retailers
and end-users in Queensland, New South Wales, the
Australian Capital Territory, Victoria and South Australia
in December 1998. Tasmania joined the NEM in 2005 and
operations today are based in five interconnected regions
that largely follow state boundaries.
The NEM operates on the worlds longest interconnected
power system from Port Douglas in Queensland to Port
Lincoln in South Australia a distance of around 5,000
kilometres. More than $10 billion of electricity is traded
annually in the NEM to meet the demand of more than
eight million end-use consumers.
GENERATION BY FUEL TYPE1

OIL AND OTHER: 0.2%
WIND2: 1.5%
HYDRO: 5.0%
NATURAL GAS3: 12.2 %
BROWN COAL: 24.8 %
BLACK COAL: 56.3%
1 Excludes embedded and
non-grid private generation
2 Includes generation from
semi-scheduled and large
non-scheduled intermittent
generators
3 Includes generation from
coal seam methane
Some assets that comprise the NEMs infrastructure

are owned and operated by state governments, and
some are owned and operated under private business
arrangements.
Exchange between electricity producers and electricity
consumers is facilitated through a pool where the output
from all generators is aggregated and scheduled to meet
demand. The electricity pool is not a physical location;
rather it is a set of procedures that AEMO manages
according to the provisions of National Electricity Law and
Statutory Rules (the Rules) and in conjunction with market
participants and regulatory agencies.
Electricity is an ideal commodity to be traded using pool

arrangements because of two of its unique characteristics.
Electricity cannot be stored for future use, so supply must
vary dynamically with changing demand. And because one
unit of electricity is indistinguishable from all other units,
it is impossible to determine which generator produced
which electricity.
Sophisticated information technology systems underpin
the operation of the NEM. The systems balance supply
with demand, maintain reserve requirements, select which
components of the power system operate at any one
time, determine the spot price, and thereby facilitate the
financial settlement of the physical market.
ELECTRICITY CONSUMPTION BY SECTOR

TRANSPORT AND
STORAGE: 1.0%
MINING: 9.4%
MANUFACTURING: 9.1%
ALUMINIUM
SMELTING: 11.0%
NUMBER OF CUSTOMERS BY SECTOR
AGRICULTURE: 0.8%
RESIDENTIAL: 27.7%
BUSINESS: 12%
DOMESTIC: 88%
COMMERCIAL: 22.8%
METALS: 18.3%
Source: Electricity Gas Australia 2010 ESAA
The Australian Energy Market Operator (AEMO) was

established to manage the NEM and gas markets from
1 July 2009.
AEMOs core functions can be grouped into the
following areas:

Electricity Market - Power System and Market Operator

Gas Markets Operator
National Transmission Planner
Transmission Services
Energy Market Development
Created by the Council of Australian Governments

(COAG) and developed under the guidance of the
Ministerial Council on Energy (MCE), AEMO strengthens
the national character of energy market governance by
drawing together under the one operational framework
responsibility for electricity and gas market functions,
NEM system operations, management of Victorias gas
transmission network and national transmission planning.
AEMO carries out the electricity functions previously
undertaken by the National Electricity Market
Management Company (NEMMCO) with respect to the
NEM and the planning responsibilities of the Electricity
Supply Industry Planning Council (ESIPC, South Australia).
Additionally, AEMO assumed the retail and wholesale gas
market responsibilities of the Victorian Energy Networks
Corporation (VENCorp), Retail Energy Market Company
(REMCO), Gas Market Company (GMC) and Gas Retail
Market Operator (GRMO).
As part of its gas market functions, AEMO is responsible
for the establishment of a Short Term Trading Market, due
to commence in 2010 (initially in the New South Wales
and South Australia hubs), which sets a daily wholesale
price for natural gas.
AEMO operates on a cost recovery basis as a corporate

entity limited by guarantee under the Corporations Law.
Its membership structure is split between government
and industry, respectively 60 and 40 percent, with
this arrangement to be reviewed after three years of
operation. Government members of AEMO include
the Queensland, New South Wales, Victorian, South
Australian and Tasmanian state governments, the
Commonwealth and the Australian Capital Territory.
AEMO and the NEM
A key aim of AEMO is to provide an effective
infrastructure for the efficient operation of the wholesale
electricity market, to develop the market and improve
its efficiency and to coordinate planning of the
interconnected power system.
AEMOs primary responsibility is to balance the demand
and supply of electricity by dispatching the generation
necessary to meet demand.
AEMOs key financial objective of being self-funding is
achieved through the full recovery of its operating costs
from fees paid by market participants.
The National Electricity Law and the Rules were amended
to replace NEMMCO with AEMO as the national
electricity market and system operator.
AEMOs functions are prescribed in the National
Electricity Law while procedures and processes for market
operations, power system security, network connection
and access, pricing for network services in the NEM and
national transmission planning are all prescribed in
the Rules.
With respect to the electricity market AEMO has two

core roles:
Power System Operator
Market Operator
The market requirements determine how the power
system is operated.
The Australian Energy Market Operator
AEMOs electricity market and system operation

responsibilities include:

anagement of the NEM

M
Overseeing reliability and security of the NEM
Ensuring supply reserve to meet reliability standards
Directing generators to increase production during
periods of supply shortfall
Instruction of load shedding to rebalance supply and
demand and protect power system operations
Implementation of reserve trading to maintain supply
and reliability levels through demand-side response
National transmission planning for the electricity
transmission grid and production of a National
Transmission Network Development Plan
Publication of the Electricity Statement
of Opportunities
Electricity emergency management
Facilitation of Full Retail Competition.
AEMO manages the market and power system from two

control centres in different states. Both centres operate
around the clock, and are equipped with identical
communication and information technology systems.
The entire NEM, or individual regions within it, can be
operated from either or both centres. This arrangement
ensures continuous supply despite the risks posed by
natural disasters or other critical events, and provides
AEMO with the flexibility to respond quickly to dramatic
changes in the market or the power system.
5
NATIONAL ELECTRICITY LAW

AND RULES
THE SPOT MARKET
When the NEM commenced, a National Electricity

Code provided guidelines for how the market was to
operate. These guidelines were developed following
comprehensive consultation and extensive trials
conducted between governments, the electricity
supply industry and electricity users as part of a
government-driven deregulation and reform agenda.
Wholesale trading in electricity is conducted as a spot

market where supply and demand are instantaneously
matched in real-time through a centrally-coordinated
dispatch process. Generators offer to supply the
market with specific amounts of electricity at particular
prices. Offers are submitted every five minutes
of every day. From all offers submitted, AEMO
determines the generators required to produce
electricity based on the principle of meeting prevailing
demand in the most cost-efficient way. AEMO then
dispatches these generators into production.
In June 2005, the National Electricity Code was replaced

by the National Electricity Law and Rules. The Law and
Rules were recently amended to replace NEMMCO
with AEMO as the national electricity market and
system operator.
AEMOs functions are prescribed in the National
Electricity Law while procedures and processes for market
operations, power system security, network connection
and access, pricing for network services in the NEM and
national transmission planning are all prescribed
in the Rules.
A dispatch price is determined every five minutes, and six

dispatch prices are averaged every half-hour to determine
the spot price for each trading interval for each of the
regions of the NEM. AEMO uses the spot price as the
basis for the settlement of financial transactions for all
energy traded in the NEM.
The Rules set a maximum spot price, also known as
a Market Price Cap, of $12,500 per megawatt hour (MWh).
This is the maximum price at which generators can bid
into the market and is the price automatically triggered
when AEMO directs network service providers to interrupt
customer supply in order to keep supply and demand in
the system in balance.
Market Price Cap

The Rules place a limit on the maximum spot price at any
regional reference node. This limit is called the Market
Price Cap and is $12,500 per MWh. The spot price may be
set to the cap value when it is necessary to involuntarily
interrupt electricity supplies (ie load shedding) to Market
Customers in order to balance the overall electricity
supply and demand.
Market Floor Price
The Rules place a limit on the minimum spot price. This
limit is called the Market Floor Price and is currently set at
-$1,000 per MWh.
Trends in spot price movement provide signals for future

investment in generation and transmission infrastructure
in the NEM. As the capacity of available generation to
meet demand diminishes, relative scarcity will lead to an
increase in the spot price, and new generation or network
capacity will be attracted into the market. High spot
prices during periods of supply scarcity may also act as an
incentive for consumers to reduce their demand.
The NEM is a wholesale market. Up to 50 percent of
the price paid by domestic and business consumers for
electricity supply is accounted for by the direct cost of
the energy. Additional charges are added to retail
accounts for network usage, service fees, market charges,
retail charges and GST.
The Reliability Panel reviews the level of the Market Floor

Price and Market Price Cap every two years.
Two aspects of the transmission network contribute
to varying costs of electricity supply within different
areas of the NEM. Firstly, losses are incurred as power
is transported from where it is produced to where it is
consumed through electrical resistance and the heating
up of conductors. Secondly, electricity being transported
along certain elements of the network may encounter
technical constraints on capacity or bottlenecks.
AEMO
THE SPOT MARKET CONTINUED
GENERATOR
TRANSMISSION NETWORK
SERVICE PROVIDER
TOTAL ENERGY SENT OUT 2008/09

TAS 4.9%
DISTRIBUTION NETWORK
SERVICE PROVIDER
SA 6.5%
NSW 38%
VIC 25.1%
QLD 25.4%
Source: ESAA
MARKET CUSTOMER
DISPATCH INSTRUCTIONS
PHYSICAL ELECTRICITY FLOW
FINANCIAL FLOWS
ENERGY AND FINANCIAL FLOWS
KEY PARAMETERS FOR NEM OPERATION

AEMO is required to operate the power system
efficiently and ensure agreed standards of security and
reliability are maintained.
The minimum reserve levels across the different

NEM regions are listed in the Electricity Statement of
Opportunities on the AEMO website.
Security of Supply
Managing Security and Reliability
AEMOs highest priority as power system and market

operator of the NEM is the management of power system
security. Security of electricity supply is a measure of the
power systems capacity to continue operating within
defined technical limits despite the disconnection of a
major power system element, such as a generator or
interconnector.
In all but extraordinary circumstances, market forces keep

supply and demand in the NEM in balance. However,
during periods of supply shortfall when system security or
reliability of supply is threatened, the Rules endow AEMO
with authority to use a variety of tools to restore supply
and demand balance. The tools include demand side
management, the power of direction, load shedding and
reserve trading.
The maintenance of power system security ensures the

power system is operated in a way that does not overload
or damage any part of it or risk overload or damage after
a credible event.
Power System Reliability

Reliability is a measure of the power systems capacity to
continue to supply sufficient power to satisfy customer
demand, allowing for the loss of generation capacity.
The shortfall of supply against demand is referred to as
unserved energy. Reliability standards are established in
the NEM that determine that unserved energy per year
for each region must not exceed 0.002 percent of the total
energy consumed in that region that year.
Supply Reserve
The power system is required to be operated at all
times with a certain level of reserve in order to meet
the required standard of supply reliability across the
NEM. Calculation of the minimum reserve requirements
recognises reserve sharing in a national context.
Security and Reliability Directions

AEMO has the power to direct registered generators into
production when a supply shortfall is expected and some
generators are known to have withheld some of their total
capacity from the market. AEMO only uses this power
of direction to protect power system security or supply
reliability.
Load Shedding
In the event that demand in a region exceeds supply
and all other means to satisfy demand have been
implemented, AEMO can instruct network service
providers to shed some customer load. This action is only
taken when there is an urgent need to protect the power
system by reducing demand and returning the system to
balance. Load shedding involves a temporary suspension
of supply to customers in a specific part or region of the
NEM where system security is at risk.
During a period of load shedding, supply is withdrawn

from those NEM regions affected by the shortfall in
proportion to the demand levels at the time the shortfall
began. The proportioning process determines the amount
of load shedding for each affected region up to the point
where interconnectors are operating to their maximum
transfer capacity. Once the interconnectors reach their
maximum transfer capacity, the importing region must
bear any additional load shedding locally.
By implementing load shedding, AEMO protects the
integrity of power system operation so that widespread
and long-lasting blackouts are avoided. It also ensures
that the hardship caused by a sustained supply shortfall is
shared in an equitable fashion.
Reserve Trading
When there is sufficient notice of an upcoming shortfall
of supply that threatens to compromise minimum reserve
margins, AEMO may tender for contracts for electricity
supply from sources beyond those factored into AEMOs
usual forecasting processes. At these times, emergency
generators and other generators connected directly to
the distribution network who submit tenders may enter
contracts to boost supply in the NEM so the widespread
supply interruptions that may otherwise have occurred can
be avoided. In the same way, some electricity consumers
may offer for a financial consideration to decrease their
demand at times of supply shortfall so that demand and
supply are brought into balance.
8,984
5,935
6,004
1,158
1,542
3000
TAS
SA
Demand in the Victorian and South Australian regions of

the NEM is characterised by short-term demand peaks
during the summer months. It makes economic and
market sense that these extreme peaks of demand be
met by special arrangements rather than having excess
base-load generation capacity in the system at all times.
The peaks are currently being met by a combination of
generators that have been specifically built to service
extreme demand periods (peak generators), and demand
side participation, where consumers voluntarily and
temporarily withdraw from the market when the spot price
reaches a threshold level.
6000
VIC
AEMO conducts forecasts of expected electricity demand

in order to operate the NEM. Demand varies from region
to region depending on population, temperature, and the
industrial and commercial needs. It also varies throughout
the day, with daily demand peaks (driven by domestic
activity) typically occurring between 7:00 am and 9:00 am
and between 4:00 pm and 7:00 pm.
9000
NSW
Demand
A typical level of demand for electricity across the NEM

is approximately 25,000 megawatts on a business day of
average temperatures. There is ample supply available in
the system to meet this level of demand. In fact, supply
only comes under extreme pressure for a few hours on
just a few days of extreme high temperature that occur
each year. Further, because peak demand does not occur
simultaneously in all regions, total supply can be shared
between regions using the interconnected power network.
QLD
Operating the NEM involves conducting a sequence

of activities to facilitate trade between the producers
and wholesale consumers of electricity. These activities
include establishing demand levels, receiving offers
to supply from generators, scheduling generators,
dispatching generators into production, calculating
the spot price, measuring electricity use and financially
settling the market.
OPERATING THE NEM
AVERAGE DEMAND
AVERAGE DEMAND (MW) 2008/09

Source: ESAA
OPERATING THE NEM CONTINUED
Supply
Scheduling and Dispatching Generators
The delivery of electricity to market customers comprises

a sequence of distinct processes that AEMO manages
according to strict timetables.
From the bids submitted, AEMOs systems determine

which generators are required to satisfy demand, at
what time, and their production levels in a process called
scheduling. Offers to generate are stacked in order of
rising price, and are then scheduled and dispatched into
production. The use of the rising-price stack means that
more expensive generators are scheduled into production
as total demand for electricity increases.
Submitting Offers to Supply

To enable AEMOs systems to facilitate supply, scheduled
NEM generators are required to submit to AEMO offers
indicating the volume of electricity they are prepared to
produce for a specified price.
There are three types of bids or offers to supply daily
bids, re-bids and default bids. Daily bids are submitted
before 12:30 pm on the day before supply is required,
and are reflected in pre-dispatch forecasts. Generators
may submit re-bids up until approximately five minutes
prior to dispatch. In doing so, they can change the volume
of electricity from what it was in the original offer, but they
cannot change the offer price.
Default bids are standing bids that apply where no daily
bid has been made. These bids are of a commercialin-confidence nature and, in general, reflect the base
operating levels for generators.
10
At times, the technical capacity of the transmission

network may determine which generators are scheduled
to meet demand. In such a situation, generators may be
scheduled out of price order so that demand in
a particular area supplied through the network may
be satisfied.
From the bids

submitted, AEMOs
systems determine
which generators are
required to satisfy
demand.
TOTAL DEMAND OF ELECTRICITY FROM

THE POOL (MW)
500
400
C
B
300
$38
$37
$35
200
$28
100
$20
0
4:05
4:10
4:15
4:20
4:25
4:30
5-MINUTE PERIODS THROUGHOUT A HALF HOUR TRADING PERIOD
GENERATOR:
ONE
TWO
THREE
FOUR
Bids to produce electricity received by AEMO

are stacked in ascending price order for
each dispatch period. Generators are then
progressively scheduled into production to
meet prevailing demand, starting with the
least-cost generation option.
A. In order to supply demand for power at 4:05
pm, Generators 1 and 2 are dispatched to their
full bid capacity, and Generator 3 is only partly
dispatched. The price is $35 per MWh.
B. At 4:10 pm, demand has increased:
Generators 1, 2 and 3 are fully dispatched,
and Generator 4 is partly dispatched.
The price is $37 MWh.
C. At point C (4:15 pm) demand has increased
a further 30 MW. Generators 1, 2, 3 and 4
continue producing power and the price remains
at $37 MWh.
D. By 4:20pm, demand has increased to the

point that Generator 5 is just required to meet
demand, and the price increases to $38 per MWh.
E. At 4.25 pm, Generators 1-4 are fully
dispatched and Generator 5 partly dispatched.
The price remains at $38 per MWh.
F. By 4:30 pm, demand has fallen. Generator
5 (the most expensive generator) is no longer
required, and Generator 4 is only partly
dispatched. The price returns to $37 per MWh.
The spot price for the trading period is

calculated as the average of the six dispatch
prices. That is, $(35+37+37+38+38+37) per MWh
divided by six, or $37 per MWh. This is the price
all generators receive for production during this
period, and the price market customers pay for
electricity they consume from the pool during
this period.
FIVE
SCHEDULING OF NEM GENERATORS

Characteristic
Type
Gas and Coal-fired Boilers
Gas Turbine
Water (Hydro)
Renewable (Wind/Solar)
Time to fire-up generator from cold
8-48 hours
20 minutes
1 minute
dependent on prevailing
weather
Degree of operator control over energy source
high
high
medium
low
Use of non-renewable resources
high
high
nil
nil
Production of greenhouse gases
high
medium-high
nil
nil
Other characteristcs
medium-low operating cost
medium-high
operating cost
low fuel cost with plentiful

water supply; production
severely affected by drought
suitable for remote and

stand-alone applications;
batteries may be used to
store power
CHARACTERISTICS OF GENERATORS
11
Setting the Spot Price

AEMO issues dispatch instructions to generators at
five-minute intervals throughout each day based on the
offers generators have submitted in the bidding process.
In this way, there are 288 dispatch intervals every day.
The dispatch price represents the cost to supply the last
megawatt of electricity to meet demand, and applies to
all generators scheduled into production regardless of
the level of their original offer.
A trading interval in the NEM is a half-hour period. Hence,
there are 48 trading intervals in the market each day. The
spot price of electricity for all 30-minute trading intervals
each day is the average of the six dispatch prices during
the preceding half-hour. There is a separate spot price for
each trading interval in each of the NEMs five regions.
Factors that contribute to variations in the spot price

in different regions of the NEM include limits on
interconnector capacity and reliance on differing fuel
sources for local supply in different NEM regions. Because
gas is a more expensive fuel than coal or water, electricity
produced using gas will generally cost more than
electricity produced by the other means. Other factors
including total system load, plant outages, frequency
control, voltage control, testing and transmission outages
also affect the dispatch and spot prices. During 2007-08,
the average daily spot price across all regions of the NEM
was $52 per MWh.
Measuring Electricity Use

All market customers are required to install equipment
to record their electricity consumption. AEMO registers,
accredits and audits a range of metering services provided
THE GENERATOR DISPATCH CYCLE

Scheduling
Ranking bids
Identifying the dispatch levels
of generating units
Data input
Establishing current
operational status of
generating units
Assessing demand
forecasts
Applying loss factors
Determining system
conditions
12
Dispatch
Issuing dispatch instructions
to generators
by local network service providers. These service providers

are responsible for measuring the volume of electricity
supplied, validating the data from the meters, applying
distribution loss factors, and forwarding the information to
AEMO for use in calculating and preparing accounts for
financial settlement.
Settling the Market

AEMO calculates the financial liability of all market
participants on a daily basis and settles transactions for all
trade in the NEM weekly. This involves AEMO collecting
all money due for electricity purchased from the pool
from market customers, and paying generators for the
electricity they have produced. The spot price is the basis
for all these financial transactions.
NEM financial settlement operates four weeks in arrears
and generally includes millions of dollars of trading
funds. In order to ensure that generators are paid for
their electricity production, AEMO has strict prudential
arrangements and a robust risk management program in
place. As part of this, AEMO requires the deposit of bank
guarantees and security deposits against an established
maximum credit limit for each market customer. AEMO
closely monitors the activities of all participants in the
market and has a firm timetable in place for the entire
settlement process.
The settlement process involves determining the financial
liabilities, issuing accounts, and settling amounts payable
and receivable for electricity sold to and purchased
from the pool. The settlement price for both generators
and market customers is equal to the amount of energy
produced or consumed multiplied by both the spot price
that applies in the region of their operation and any loss
factors that apply.
Demand Side Participation

Demand side participation refers to the situation where
market customers reduce their consumption of electricity
in response to a change in market conditions, such as
high spot prices. This is a deliberate action taken when
demand for power drives spot prices high.
Under similar arrangements scheduled loads, such as
smelters, may elect to withdraw from the market when
the spot price reaches a particular threshold, and resume
trading when the price falls to the level of their bids again.
This strategy is beneficial to both the customer and the
market in that it allows the smelter to avoid the peaks
of high spot prices without damaging their production
processes, and provides a short-term response to a supply
shortfall in the market. A similar strategy, called load
shifting, describes a process where specific demand is
intentionally moved to a time when there is lower overall
demand and consequent lower spot prices. Off-peak hot
water arrangements are an example of the deliberate
shifting of demand for electricity to a low-demand period.
All market customers are

required to install equipment
to record their electricity
consumption. AEMO registers,
accredits and audits a range
of metering services provided
by local network service
providers.
If a market participant breaches their maximum credit limit

on any one day of trading, a call notice for rectification of
the situation and then a default notice may be issued to
ensure that AEMO is able to settle the market according
to its fixed timetable. AEMO has the authority to suspend
a market participant who fails to respond adequately
to a default notice, and to reinstate that market participant
only when their required financial position
is re-established.
13
ANCILLARY SERVICES
Ancillary services are those services used by AEMO
to manage the power system safely, securely and
reliably. Ancillary services maintain key technical
characteristics of the system, including standards
for frequency, voltage, network loading and system
re-start processes.
NCAS are primarily used to:
AEMO operates eight separate markets for the delivery

of Frequency Control Ancillary Services (FCAS), and
purchases Network Control Ancillary Services (NCAS)
and System Restart Ancillary Services (SRAS) under
agreements with service providers.
SRAS are reserved for contingency situations in which

there has been a major supply disruption or where the
electrical system must be restarted.
FCAS providers bid their services into the FCAS markets

in a similar way to how generators bid into the energy
market. The FCAS markets were introduced to the NEM
in September 2001 and provide simpler, more dynamic
and transparent arrangements that have further increased
competition and contributed to improved overall market
efficiency.
Payments for ancillary services include payments for
availability and for the delivery of the services. The market
participant or participants responsible for a situation
that requires ancillary services pay for individual services
whenever regulation FCAS are needed to automatically
raise or lower frequency to within the normal operating
band of 49.9 Hertz to 50.1 Hertz.
14
C
ontrol the voltage at different points of the electrical
network to within the prescribed standards; or
Control the power flow on network elements to within
the physical limitations of those elements
Ancillary service costs are dependant upon the amount

of service required at any particular time and, as these
amounts can vary significantly from period to period, costs
will also vary.
The NEM comprises five interconnected electrical

regions. There is a designated region reference node in
each region where the regional spot price of electricity
is set. The Queensland, New South Wales, Victoria,
Tasmania and South Australia regions all contain both
major generation and demand centres.
Interconnectors
The high-voltage transmission lines that transport
electricity between adjacent NEM regions are called
interconnectors. Interconnectors are used to import
electricity into a region when demand is higher than can
be met by local generators, or when the price of electricity
in an adjoining region is low enough to displace the
local supply.
AEMOs ability to schedule generators to meet demand
using an interconnector to facilitate importing electricity
is sometimes limited by the physical transfer capacity
of the interconnector. When the technical limit of an
interconnectors capacity is reached, the interconnector
is said to be constrained. For example, if prices are
very low in one region and high in an adjacent region,
electricity can be sent from the first to the second region
across an interconnector up to the maximum technical
capacity of the interconnector. AEMOs systems will then
dispatch local generators with the lowest price offers
from within the second region to meet the outstanding
consumer demand.
Regulated Interconnectors
A regulated interconnector is an interconnector that has
passed the ACCC-devised regulatory test and has been
deemed to add net market value to the NEM. Having
passed the test, a regulated interconnector becomes
eligible to receive a fixed annual revenue set by the ACCC
and based on the value of the asset, regardless of actual
usage. The revenue is collected as part of the network
charges included in the accounts of electricity end-users.
At present, regulated interconnectors operate between
all adjacent regions of the NEM, except Tasmania.
INTER-REGIONAL TRADE
INTERCONNECTORS IN THE NEM
REGIONAL REFERENCE NODE

REGULATED INTERCONNECTOR
MARKET NETWORK SERVICE
PROVIDER
QLD
SOUTH PINE
NSW-QLD
(QNI)
SA
NSW-QLD
TERRANORA
TORRENS ISLAND
VIC-SA NSW WEST SYDNEY

(MURRAYLINK)
VIC
VIC-SA
(HEYWOOD)
VIC-NSW
THOMASTOWN
TAS-VIC
(BASSLINK)
GEORGE TOWN
TAS
15
INTER-REGIONAL TRADE CONTINUED
Unregulated Interconnectors
Loss of Energy in the System
Unregulated (or market) interconnectors derive revenue

by trading in the spot market. They do this by purchasing
energy in a lower price region and selling it to a higher
price region, or by selling the rights to revenue generated
by trading across the interconnector. Unregulated
interconnectors are not required to undergo regulatory
test evaluation.
As electricity flows through the transmission and

distribution networks, energy is lost due to electrical
resistance, and the heating of conductors. The losses
are equivalent to approximately 10 per cent of the total
electricity transported between power stations and
market customers.
An unregulated interconnector Basslink operates

between the Tasmanian and Victorian regions of the
NEM. Murraylink and Directlink were built as unregulated
interconnectors between Victoria and South Australia,
and New South Wales and Queensland respectively.
They successfully applied to the ACCC for conversion
to regulated status.
Energy losses on the network must be factored in at all

stages of electricity production and transport to ensure
the delivery of adequate supply to meet prevailing
demand and maintain the power system in balance.
In practical terms, this means that more electricity must
be generated than indicated in demand forecasts in order
to allow for this loss during transportation.
The impact of network losses on spot prices is
mathematically represented as transmission and
distribution loss factors. Loss factors within each region
of the NEM are calculated based on forecast demand,
and fixed for a period of 12 months to facilitate efficient
scheduling and settlement processes in the NEM. Loss
factors between regions of the NEM are dynamically
calculated and reflect the operating conditions at the time
of the transmission of the electricity.
As electricity flows through the

transmission and distribution networks,
energy is lost due to electrical resistance,
and the heating of conductors.
16
REGION A
REGION A
REFERENCE
NODE
REGION
BOUNDARY
REGION B
REFERENCE
NODE
CUSTOMER 2
CUSTOMER 1
REGION B
INTER-REGIONAL LOSS FACTOR APPLIES
INTRA-REGIONAL LOSS FACTOR APPLIES
Electricity losses occur between regions and within

regions. Losses between regions are of the order of
10 per cent of electricity transported. Therefore, to ensure
that 100 MW of energy committed to be supplied to
Region B (in the diagram) from generators within Region
A, 110 MW of electricity must be exported from Region A.
Intra-regional losses occur between the region reference
node, where the region spot price is set, and the
customers connection point to the grid. In the diagram,
customer C1 would require more energy to be imported
to receive the same amount of supply as customer C2,
because C2 is closer to the regional reference node.
LOSS OF ENERGY IN THE POWER SYSTEM
AEMO uses a variety of forecasting processes to

determine the level of demand for every dispatch
interval in the NEM. Then using the submitted offers
to generate electricity, AEMO produces a schedule or
timetable of generation to ensure that the forecast
demand will be met based on the requirements that
the least expensive generators are dispatched into
production and the power system remains in a secure
operating state.
As a prerequisite for maintaining supply and demand in
balance, it is important for AEMOs planning processes
to be informed in advance of any limits on the capacity
of generators to supply electricity or networks to
transport electricity. This enables the remainder of market
participants to respond to potential supply shortfalls by
increasing their generation or network capacity to the
market. Market participants are able to signal upcoming
limitations on supply by means of a variety of forecasting
tools designed to improve the overall efficiency of
the market.
Pre-dispatch Forecasting
Projected Assessment of System Adequacy
Pre-dispatch is a short-term forecast of supply and

demand in the market. It is used to estimate the price
and demand for the upcoming trading day, and the
volume of electricity expected to be supplied through the
interconnectors between regions.
AEMO monitors the future adequacy of generating

capacity based on the predicted availability of generating
units at power plants. AEMO produces both seven-day
and two-year forecasts because of the variability
of demand for electricity. These forecasts are called the
short-term and medium-term Projected Assessments of
System Adequacy, or PASA, respectively. They are used
by AEMO to ensure that adequate levels of reserve are
in the system at all times, and by generators and network
operators to plan augmentation, maintenance and
other outages.
Generators and network operators are required to notify

AEMO of their maximum supply capacity and availability,
and this information is matched against regional demand
forecasts. All offers to supply are then collated so that
potential shortfalls of supply against demand can be
identified and published. Participants in the market use
this information as the basis for any re-bids of the capacity
they wish to bring to the market.
Five-minute Matching of Supply and Demand

Generators are scheduled and dispatched into production
to match supply with prevailing demand every five minutes
of every day. This process, in turn, produces dynamic
price signals that guide market participants as they bid to
supply electricity to the market.
MARKET FORECASTS
AEMO PRODUCES TWO PASA FORECASTS

Forecast
Forecast
Period
Updated/
Published
Short-term
PASA
7 days
2-hourly from
4:00am
Medium-term
PASA
2 years
2:00pm every
Tuesday
A DAY IN THE NEM

12.00 MIDNIGHT START/
END OF SETTLEMENT DAY
24
18
24 HOUR
CLOCK
4.00AM (EST) TRADING DAY

STARTS AND ENDS
30 MINUTE TRADING
INTERVAL (48 PER DAY)
1 HOUR
12:30PM DEADLINE FOR DAILY BIDS FOR
NEXT TRADING DAY (RE-BIDS CAN BE UP
TO 5 MINUTES PRIOR TO DISPATCH
PRE-DISPATCH FORECAST
PUBLISHED
60
45
1 HOUR
CLOCK
15
5 MINUTE DISPATCH INTERVAL
(288 PER DAY)
PRE-DISPATCH FORECAST
PUBLISHED
17
MARKET FORECASTS CONTINUED
Electricity Statement of Opportunities

AEMO publishes a 10 year forecast called the electricity
Statement of Opportunities (SOO) each year. This
publication provides information to assist market
participants assess the future need for electricity
generating capacity, demand side capacity and
augmentation of the network to support the operation
of the NEM.
It also contains forecasts of ancillary service requirements,
minimum reserve levels, and economic and operational
data to assist potential investors gain a full understanding
of the NEM.
The Electricity SOO brings together information supplied
to AEMO by the planning bodies in each jurisdiction of
the NEM. A year-by-year annual supply-demand balance
is presented for each region in the SOO as a snapshot
forecast of the capacity of generation and distribution to
satisfy demand for electricity into the future.
National Transmission Network

Development Plan
AEMO is the National Transmission Planner for the
electricity transmission grid. A core component of this
transmission planning responsibility involves preparing
annual network development plans to guide investment in
the power system.
In 2009 an interim National Transmission Statement (NTS)
replaced the previous Annual National Transmission
Statement produced by NEMMCO. This document will
be superseded by the National Transmission Network
Development Plan (NTNDP) from 2010.
The NTNDP will:
p
rovide historical data and projections of network
utilisation and congestion;
summarise emerging reliability issues and potential
network solutions identified by the Jurisdictional
Planning Bodies; and
present information on potential network
augmentations and non-network alternatives and their
ability to address the projected congestion.
Transmission planning documents rely heavily on market
simulations. Consultation is conducted with interested
parties to comment on the input data and assumptions
that are used in market simulations.
18
Since the commencement of the NEM, electricity

consumers have progressively gained the right to
choose their own supplier. This has meant that AEMOs
responsibilities have extended from managing the
wholesale market to providing the systems and processes
to support competition and choice for all end-users in the
retail electricity market. Delivering full retail competition
(FRC), or contestability, has required new information
technology systems to process transfers of customers
between registered retailers in the NEM. The systems that
facilitate this function contain one of the largest metering
databases in the world. They accept data from a variety of
electricity meter types and have the capacity to process
information from up to 10 million meters.
One of AEMOs responsibilities under the Rules is to

register participants in the NEM. There are six main
categories of registered participant, including those
who participate directly in trading activities, and other
participants who provide services essential for the
operation of the market. The categories are generator,
market customer, intending participant, network service
provider, trader, reallocator and special participant.
AEMOs systems are set up to provide key meter

installation details to support a simple and rapid
information transfer process. Different metering processes
are required for different types of meters used in the
NEM, to support consumer transfer and core settlement
procedures and to calculate load profiles. The cost of
electricity consumed is then calculated according to
a user profile that approximates the pattern of use in
a typical situation.
REGISTERED PARTICIPANTS
By June 2009, approximately 6.3 million customer transfers

from one retailer to another had taken place. As a result
of the introduction of full retail competition, electricity
retailers are increasingly competing, and creating new
and unique products as a means to increase their
customer bases.
Scheduled: aggregate
generation capacity of more than
or equal to 30 megawatts.
Semi-scheduled: aggregate
generation capacity of more than
or equal to 30 megawatts where
output is intermittent.
Non-scheduled: aggregate
generation capacity of less than
30 megawatts.
Market participants include market generators, market

network service providers and market customers. A market
participant must be separately registered in each category
of the market in which it participates. For example, a
business that participates as a generator (a peaking
generating plant for instance) and as a market customer
(retailer of electricity to end-use customers) would be
required to be registered as both a generator and a
market customer.
FULL RETAIL COMPETITION
The registration of participants is a formal process, strictly

defined in the Rules. Registered participants are required
to pay participant fees that are levied to recover the costs
associated with managing the market.
Market Participants
Other Registered Participants
Registered to participate in the National Electricity Market

Market Generators
Sell entire electricity output
through the spot market
and receive the spot price at
settlement.
Market Network Service

Providers
Own and operate a network
linked to the national grid at
two terminals in different NEM
regions. Pay market participant
fees and obtain revenue from
trading in the NEM.
Market Customers
Transmission Network Service Provider

Owner and operator of the high-voltage transmission towers
and wires that transport electricity.
Distribution Network Service Provider
Owner and operator of substations and the wires that transport
from distribution centres to end-use consumers. Also provider
of technical services, including construction of power lines,
inspection of equipment, maintenance and street lighting.
Reallocator
Registered with AEMO to participate in reallocation transactions
under clause 3.15.11 of the National Electricity Rules.
Purchase electricity supplied to

a connection point on a NEM
transmission or distribution
system for the spot price.
Special Participant
System operators or agents appointed to perform power security
functions. Distribution system operators and controllers or
operators of any portion of the distribution system.
Electricity Retailers: buy

electricity at spot price and
on-sell it to end-use customers.
End-use Customers: buy directly
from the market for own use.
Intending Participant
Must reasonably satisfy AEMO of intention to perform activity that
would entitle it to be a registered participant.
Trader
Party registered to participate in the settlement residue auction.
19
FINANCIAL CONTRACTS FOR ELECTRICITY
Hedge Contracts
Hedge contracts are typically agreements between
generators and customers that operate independently of
both the market and AEMOs administration. The details
of hedge contracts are not factored into the balancing
of supply and demand, and are not regulated under the
Rules. These contracts can be entered into under either
long-term or short-term arrangements that set an agreed,
or strike, price for electricity traded through the pool. In
this way, hedge contracts are financial instruments that
participants can use to manage the financial risk that
results from potential volatility of the spot price.
The basic form of a hedge contract exists where two
parties agree to exchange cash so that a defined quantity
of electricity over a nominated period is effectively valued
at an agreed strike price. Under such an agreement,
generators pay customers the difference when the spot
price is above the strike price. When the spot price is
below the strike price, customers pay generators the
difference between the spot price and the strike price.
20
$160
SPOT PRICE
$140
$120
$100
PRICE
Participants in the NEM require a means of managing

the financial risks associated with the significant
degree of spot price volatility that occurs during
trading periods. They typically achieve this by
using financial contracts that lock in a firm price for
electricity that will be produced or consumed at a
given time in the future. These contracts serve to
substantially reduce the financial exposure of market
participants and contribute to spot market stability.
They are known as derivatives, and include swaps
or hedges, options and futures contracts.
$80
SELLER PAYS BUYER

DIFFERENCE BETWEEN
AGREED STRIKE PRICE
AND SPOT PRICE
BUYER PAYS SELLER

DIFFERENCE BETWEEN
AGREED STRIKE PRICE
AND SPOT PRICE
$60
$40
STRIKE PRICE
(AGREED CONTRACT PRICE)
$20
In this example of a hedge contract,

the two parties have agreed to set a
price (the strike price) of $40/MWh.
The graph shows the strike price
and the actual spot price over a
hypothetical days trading.
When the spot price is below the
strike price, the market customer pays
the difference in these prices to the
generator. In this case, when the spot
price is $17, the market customer pays
the generator $23/MWh.
When the spot price exceeds the
strike price, the generator pays the
market customer the extra required
to purchase electricity from the pool.
In this case, when the spot price rises
to $145/MWh, the generator pays
the market customer the $105/MWh
difference.
$0
0400
0800
1200
1600
TIME
2000
0000
HEDGE CONTRACTS IN THE NEM
Hedge contracts are typically

agreements between generators
and customers that operate
independently of both the market
and AEMOs administration.
FINANCIAL CONTRACTS FOR ELECTRICITY CONTINUED
Auctions of Inter-region Settlement Residues

REGION A
The spot price for electricity in each region of the NEM is

determined by a number of factors, including supply and
demand, the physical limitations of interconnectors, and
the loss factors for both the transmission and distribution
networks. This means that there may be significant
differences in the spot price for any trading interval across
NEM regions.
The difference between the value of electricity in the
region where it is generated and its value if sold in another
region is called the inter-regional settlement residue.
The settlement residue that accumulates is made available
to the market by the conduct of an auction. The auction
process establishes the market value of the residue,
and contributes to inter-regional trade by providing
registered generators, market customers and traders with
a mechanism to manage the risk associated with different
price outcomes between trading regions.
Registered participants who purchase auction units obtain
access to a share of the residue. In this way, the premium
paid for the auction units provide protection against
high price differences between regions in the
wholesale market.
SPOT PRICE
$100/MWH
GENERATORS
ARE PAID AT
$100/MWH
REGION BOUNDARY
SETTLEMENT RESIDUE
ACCRUED IN REGION B:
$120-$100=$20
(EXCLUDING LOSS
FACTORS)
CUSTOMERS
BUY AT
$120/MWH
SPOT PRICE
$120/MWH
REGION B
POWER FLOW
INTER-REGIONAL SETTLEMENTS RESIDUE
21
ALTERNATIVE GENERATION TECHNOLOGIES
AEMO AND THE ENVIRONMENT
The range of technologies for the generation of

electricity is expanding to accommodate alternative
energy sources such as wind energy.
AEMOs role as the manager of both the power system

and the electricity market means that it is sometimes
asked about issues of environmental management, the
sustainability of the market and the electricity supply
industry in general.
Large wind generators are typically registered as semischeduled generators (rather than scheduled) because
their energy source is intermittent and their generation
cannot increase on demand. The market is designed to
allow intermittent generators to participate and share the
same power system and the same consumers.
The NEMs base-load generators are scheduled
according to bids, and production from each generating
unit is controlled by operators. The changeability and
unpredictability of wind means that wind generators
cannot be scheduled in the usual way.
In late 2008, the Australian Wind Energy Forecasting

System (AWEFS) was implemented to forecast the
energy likely to be produced by major wind farms in the
NEM. This system enabled the generator classification
semi-scheduled to be introduced in the NEM, allowing
intermittent generators (including wind farms) to compete
in the NEM through the bidding process.
The success of AWEFS and the semi-scheduled category
will help the energy market respond to an ongoing
expansion in renewable generation while maintaining
the security of supply that has been the hallmark
of the industry.
The integration of wind and other intermittent generators

to the NEM must take account of AEMOs responsibility to
maintain power system security, and be managed during
each five-minute dispatch interval. The variation of output
associated with wind generators (with individual output
that can change by as much as 50 per cent in a five-minute
dispatch interval) may require interconnectors to operate
at lower limits to avoid overloads, and hence reduce the
total supply capacity available to the market.
Under the Rules, AEMOs charter focuses specifically

on efficiency, security and reliability of power supply,
and excludes favouring one fuel source over any other.
Consequently, AEMO has neither the power nor the
authority to make decisions based on considerations of
sustainability and balance in resource management.
The various state regulators ensure that environmental
impact assessments are conducted as part of any power
industry planning initiatives. The regulators also monitor
operations at industry sites within their jurisdictions, and
the industry itself operates and audits waste reduction and
recycling programs.
Renewable Energy Target

The Federal Governments expanded Renewable Energy
Target (RET) will result in changes to the generation mix in
the NEM over the next decade.
AEMO will continue to provide advice as requested by
government and the Ministerial Council on Energy with
regard to these changes in the NEMs supply mix and the
ongoing management of power system reliability.
Large wind generators are typically registered as

semi-scheduled generators because their energy
source is intermittent and their generation cannot
increase on demand.
22
AEMO is not responsible for market regulation. Since

mid-2005, the Australian Energy Market Commission
(AEMC) and the Australian Energy Regulator (AER)
have had responsibility for oversight and regulation
of the National Electricity Market.
The AEMC is responsible for rule making and market
development. The rule-making role does not involve
initiating changes to the Rules other than where the
change involves correcting minor errors or where
the change is of a non-material nature. Rather, the
role involves managing the rule change process, and
consulting and deciding on rule changes proposed by
others. In regard to its market development function, the
AEMC conducts reviews at the request of the Ministerial
Council on Energy or at its own volition on the operation
and effectiveness of the Rules or any matter relating to
them. In doing this, the AEMC relies on the assistance
and cooperation of industry relationships and interested
parties in its decision making.
The AER has responsibility for the enforcement of

and monitoring compliance with the Rules, as well
as responsibility for economic regulation of electricity
transmission. The AER issues infringement notices for
certain breaches of the National Electricity Law and
Rules, and is the body responsible for bringing court
proceedings in respect of breaches.
REGULATORY ARRANGEMENTS
A Memorandum of Understanding between the ACCC,

the AER and the AEMC guides interaction between these
three bodies and their function in the Australian energy
industry. The regulatory bodies have been created under
the auspices of the Ministerial Council on Energy and take
over many of the electricity regulatory arrangements
that were previously the responsibility of state government
authorities.
23
GLOSSARY
Renewable generation
Energy conversion techniques including wind, solar, hydro
and geothermal. The primary commercial sources at this
time are hydro and wind power.
Networks (Transmission)
Transmission lines carry high voltage electricity to
substation transformers where it is changed to low voltage
for distribution.
Hedge contracts
Long term or short term arrangements that set a strike
(agreed) price for electricity traded through the pool.
Networks (Distribution)
Distribution lines carry low voltage electricity to
consumers who access it through the power outlets
in homes, offices and factories.
Interconnectors
The high-voltage transmission lines that transport
electricity between adjacent NEM regions.
Intermittent generators
Where the changeability and unpredictability of the source
(such as wind) means the generators cannot be scheduled
to operate in the same way as conventional generators
(coal, gas or oil).
Load shedding
AEMO can request network service providers to
disconnect some customers when demand in a region
exceeds supply. This action is only taken when there is
an urgent need to reduce demand and return the system
to balance.
Market Price Cap
The maximum price at which generators can bid into
the market.
24
Transformers
Convert the electricity produced at a generation plant
from low to high voltage to enable its efficient transport
on the transmission system. When the electricity arrives at
the location where it is required, a substation transformer
changes the high voltage electricity to low voltage for
distribution to end users.
REGIONS AND NETWORKS IN AUSTRALIAS

2
2
TRANSMISSION INFRASTRUCTURE
2
POWER STATION
SUBSTATION
WINDFARM
500 KV TRANSMISSION LINE
DC
3
2
2 2
2
2
2
2
2
2
2

132 / 110 KV LINE
66 KV LINE
DC LINK
MULTIPLE CIRCUIT LINES
2
2
REGIONAL BOUNDARIES
REGIONAL REFERENCE NODE
QUEENSLAND
NEW SOUTH WALES
VICTORIA
SOUTH AUSTRALIA
TASMANIA
2
2
2
2
2
2
2
2
2
2 2
2 3
SOUTH PINE
DC
2
2
2
2
DC
2
2
2
TORRENS ISLAND
2
2 2
2
2
2
2
WEST SYDNEY
THOMASTOWN
DC
SYD WEST
2
2
GEORGE TOWN
2
2
25
ELECTRICITY
AEMO GPO Box 2008 Melbourne VIC 3001 Website: www.aemo.com.au
INFORMATION CENTRE Telephone: 1300 361 011
ISBN 0-646-41233-7
The Nordic Electricity

Exchange and
The Nordic Model for a
Liberalized Electricity
Market
1 The market
When the electricity market is liberalized, electricity becomes a commodity like, for
instance, grain or oil. At the outset, there is as in all other markets a wholesale
market and a retail market and there are the three usual players: the producers, the
retailers and the end users.
However, for electricity, a more advanced trading pattern quickly develops. New
players enter the scene: the traders and the brokers (Figure 1).
Figure 1: The commercial players and the electricity exchange
A trader is a player who owns the electricity during the trading process. For example,
the trader may buy electricity from a producer and subsequently sell it to a retailer.
The trader may also choose to buy electricity from one retailer and sell it to another
retailer and so forth: there are many routes from the producer to the end user.
The brokers play the same part in the electricity market as the estate agent in the
property market. The broker does not own the commodity he acts as an
intermediary.
A retailer may, for example, ask the broker to find a producer who will sell a given
amount of electricity at a given time.
The Nordic electricity exchange Nord Pool Spot covers Denmark, Finland, Sweden,
Norway , Estonia and Lithuania. Nord Pool Spot is an exchange primarily servicing the
players at the wholesale market for electricity. The customers on Nord Pool Spot are
Nord Pool Spot AS, Tel +47 6710 9100, Fax +47 6710 9101, PO Box 121, NO-1325 Lysaker, Norway,
info@nordpoolspot.com, org nr. NO 984 058 098 MVA, www.nordpoolspot.com
2
the producers, retailers, and traders who choose to trade on the electricity
exchange. In addition, large end users trade on the electricity exchange. In this article,
the term the Nord Pool Spot exchange area denotes Denmark, Finland, Sweden,
Norway, Estonia and Lithuania.
2 The Point Tariff System
In Figure 2, the water illustrates the electrical power and the walls of the tanks
illustrate the transmission grid.
Figure 2: Illustration of the Nordic Point tariff system
The idea of the system of point tariff is that the producers pay a fee to the grid owner
for each kWh they pour into the grid. Correspondingly, the end users pay a fee for
each kWh they draw from grid.
This means for example, that a retailer in Southern Sweden may buy electricity from a
producer in Northern Sweden. Of course, such a deal does not cause the producers
electricity to travel all the way from Northern Sweden to Southern Sweden. The
principle is simply that for each hour somewhere a producer has to pour an amount
of electricity to the grid which corresponds to the amount the retailers customers
have tapped from the grid.
3 The non-commercial players
The roads in the Nordic countries are operated by monopolies: The municipalities,
the counties and the state. For electricity, the grid functions like the roads
transporting the power. Correspondingly, the grid is operated by non-commercial
monopolies (Figure 3). For each local area, there is a local grid operator who handles
the local low-voltage grid (cf. the municipalities and counties operating the local
roads). The high-voltage grid is operated by the transmission system operator (TSO)
just as the motorways are operated by the state.
In addition to owning and operating the high-voltage grid, the TSO is responsible for
the security of supply in its country. Consequently, the TSO rules and controls the
electricity system in his country. Basically, the physical control and maintenance of
3
the electricity system is done in the same way, whether you have market
economy or planned economy.
Figure 3: The grid connection between producers and End-users
Only the financial organization is changed when we shift from planning economy to
market economy. This is because the laws of nature are the same whether we have
planned economy or market economy.
This also holds for corn flakes: the machine filling the corn flakes into cartons does
not care whether there is market economy or planned economy. It makes no
difference to the physics whether there is planned economy or market economy.
The commercial players are not and cannot be responsible for the security of supply.
If a South Swedish retailer, for example, has bought electricity from a North Swedish
producer, the North Swedish producer cannot guarantee that there will be electricity
in the plug at the retailers customers.
What the commercial players deliver to each other and the end users are only the
prices (and the bills). Hence, the commercial players deliver financial services only.
The commercial players work in the domain which is changed when the electricity
market is liberalized: the financial domain.
4
4 The transmission system operator (TSO)

The TSO is responsible for keeping the respective area electrically stable. Technically,
this means that the frequency must be kept at 50 Hz. In other words, the TSO is
responsible for the commodity (electricity) arriving at the end users sites.
The TSO must be a non-commercial organization, neutral and independent of
commercial players. The TSOs in the exchange areas thus have the responsibility for
both the high-voltage grid and the security of supply. In Norway, the TSO is the stateowned grid company Statnett. In Sweden, the TSO is the state-owned grid company
Svenska Kraftnt. The TSO in Finland is the grid company Fingrid and is owned partly
by the Finnish State and partly by Finnish insurance companies. In Denmark, the TSO
is the state-owned grid company Energinet.dk. Energinet.dk is TSO for both electricity
and gas. The TSO in Estonia is Elering and is fully owned by the Estonian state. In
Lithuania the TSO, Litgrid, is also stately owned.
5 Regulating power market
The regulating power market is managed by the TSO in order to obtain stable
frequency in the transmission grid. It may happen that the consumption exceeds the
generation. In this case, the frequency of the alternating current will fall to a value
below 50 Hz. When this happens, the TSO must ensure that one or more producers
deliver(s) more electricity to the grid (Figure 2). In this case, the TSO buys more
electrical power from producer(s) who has proclaimed excess generation capacity.
We say that the TSO is procuring up regulation.
The generation of electricity may also be too big exceeding the consumption. In this
case, the frequency will rise to a value above 50 Hz. Now, the TSO must ensure that
one or more producers reduce(s) the generation of electricity. In this case, the TSO is
selling electrical power to the producers thereby causing the producers to reduce
their generation. We say that the TSO is procuring down regulation.
The electricity, which the TSO in this way trades with selected market players, is
called regulating power. Hence, the regulating power is traded by the TSO in order to
regulate the frequency to keep it at 50 Hz.
To illustrate the setting of prices in the regulating power market an example is
presented. The green rectangles in Figure 6 illustrate up-regulation orders, e.g.
producers with available generation capacity. The orange rectangles illustrate downregulation orders, e.g. consumers being able to reduce consumption. All regulating
power orders submitted to the TSOs are ranked with increasing price (merit-order).
Assume there is a need for 400 MW up regulations. All the up-regulation orders with
lowest prices are activated until 400 MW is reached. The price of the last up regulated
MW sets the up-regulation price. The orders with prices below the up-regulation
price have a profit, equal to the difference between final regulation price and the
offered price. The same procedure is used to find the down-regulation price.
5
Figure 4: Price setting in the regulating power market
6 Balancing Power
In the wholesale market electricity is bought and sold hourly. Figure 4 illustrates an
example where a retailer buys electricity for one particular hour at one specific date.
The hour during which the power is delivered and consumed is called the hour of
operation.
In the example, the retailer has two contracts of 30 MWh and 70 MWh, respectively:
the retailer expects that his customers will consume 100 MWh during this hour of
operation (1 MWh is 1,000 kWh).
6
Figure 5: The retailers purchase of electricity for the hour 1pm 2pm on September 23th 2010
Before the hour of operation, the purchases must be made. After the hour of
operation, the settlement is done (Figure 5). The retailer pays the suppliers for the 30
MWh and the 70 MWh.
Assume that the retailers customers have only used 85 MWh during this hour of
operation. In this case, the retailer has per definition sold the 15 MWh to the TSO. The
TSO pays the retailer for the 15 MWh.
Figure 6: Settling the consumption of electricity
This trade with the TSO creates a balance between the retailers total trading and the
retailers customers consumption. The electricity, which the retailer trades with the
TSO, is therefore called balancing power, or often referred to as regulating power.
If the TSO had to procure up-regulation during this hour, the TSO will pay the retailer
the up-regulating price for the balancing power (i.e. the retailer will get the same
price as the producers, who sold up-regulating power to the TSO during this hour).
Normally, the up-regulating price will be higher than the market price (in this article,
the market price is the day-ahead exchange price for this hour).
If the TSO had to procure down-regulation during this hour, the TSO will pay the
retailer the down-regulating price for the balancing power (i.e. the retailer will get
the same price as the producers, who bought down-regulating power from the TSO
7
during this hour). Normally, the down-regulating price will be lower than
the market price.
In a different case, assume the retailers customers have used 110 MWh during this
hour of operation. This is 10 MWh more than the retailer bought before the hour of
operation. In this case, the retailer has to buy the additional 10 MWh from the TSO. In
this situation, the TSO will invoice the retailer for the 10 MWh.
7 Settlement in the balancing market

When the TSO sells regulating power, the price is set the same way as when the TSO
buys regulating power : if there was up-regulation during this hour, the TSO will
invoice the up-regulating price (normally higher than the market price). If there was
down-regulation, the TSO will invoice the down-regulating price (normally lower
than the market price).
Suppose one of the retailers suppliers is a producer whose plant breaks down just
before the hour of operation starts. As the market closes one hour before the hour of
operation, the producer cannot buy electricity from another supplier if his power
station breaks down 10 minutes before the hour of operation starts.
The retailer has to pay the producer, even though the producer has not produced
anything. In this case, the TSO sells balancing power to the producer, and the
producer resells the power to the retailer.
Hence, if a producer fails to produce according to his plan, the producer must also
settle balancing power with the TSO. However, for the producers the price is set a bit
differently: during an hour with up-regulation, producers producing too much will
only get paid the market price (not the up-regulating price). During an hour with upregulation, producers producing too little will be invoiced the up-regulating price
(normally higher than the market price).
During hours with down-regulation, producers producing too much will get paid the
down-regulating price (normally lower than the market price). Producers producing
too little will be invoiced the market price (not the down-regulating price).
That a trader owns electricity means in practice that the trader must settle balancing
power with the TSO, if his purchase and sale are imbalanced. Hence, in order to avoid
settling balancing power with the TSO, the trader must ensure that he is buying and
selling the same amount of power produced or consumed during each hour.
8
Elspot Nord Pool Spots Day-ahead Auction Market

8
Elspot is Nord Pool Spots day-ahead auction market, where electrical power
is traded.
Players, who want to trade power on the Elspot market, must send their purchase
orders to Nord Pool Spot at the latest at noon the day before the power is delivered to
the grid.
Correspondingly, participants who want to sell power to Elspot must send their sale
offers to Nord Pool Spot at the latest at noon the day before the power is delivered to
the grid (i.e. gate closure is 12.00).
Figure 7: Bid/Offer from one player for the hour 1pm 2pm of tomorrow.
The orders and offers are sent electronically to Nord Pool Spot in Oslo: the
participants send the orders to Nord Pool Spot via the Internet.
Figure 7 shows an example of orders submitted by a retailer for one hour of the
following day. The retailer expects that his customers will consume 50 MWh during
this hour.
This retailer has his own generation facility. Hence, he can choose whether he will
either:
- buy the 50 MWh from the exchange and therefore not produce anything himself.
- buy some of the electricity from the exchange and produce the rest himself.
- produce precisely 50 MWh.
9
- or sell electricity to the exchange and consequently produce more than

50 MWh.
The retailer in the example has informed the electricity exchange that he will buy 50
MWh from Elspot, if the exchange price for this hour turns out to be 20 EUR/MWh or
less.
If the exchange price for this hour turns out to be 40 EUR/MWh, the retailer will buy
10 MWh. In this case, the retailer will produce the remaining 40 MWh at his own
generation facility. The retailer will sell 10 MWh if the price turns out to be 50
EUR/MWh. If the price is between 50 and 60 EUR/MWh, the retailer will sell an
amount corresponding to the sloping curve. If the price is 60 EUR/MWh or more the
retailer will sell 30 MWh.
At Nord Pool Spot, the purchase orders are aggregated to a demand curve. The sale
offers are aggregated to a supply curve (Figure 8). The intersection of the two curves
gives the market price for one specific hour.
Figure 8: Aggregated supply and demand curves
Nord Pool Spot calculates a price for each hour. Elspot is a day-ahead market, as this
is trading for the following day.
This way of calculating the price is called a double auction, as both the buyers and the
sellers have submitted orders (for many other auction types, only the buyers submit
orders). Hence, Elspot is called a day-ahead auction market (as the word double is
cut out from the type description).
Figure 9 shows the prices during one specific day, July 14th 2006..
10
At noon, Nord Pool Spots computer in Oslo starts calculating the day-ahead prices.
Having finished the calculation, Nord Pool Spot publishes the prices. At the same time,
Nord Pool Spot reports to the participants how much electricity they have bought or
sold for each hour of the following day. These reports on buying and selling are also
sent to the TSOs in the Nord Pool Spot area. The TSOs use this information, when they
later calculate the balancing power for each player.
Figure 9: System price Thursday September 22nd 2011
There is a standard Elspot trading fee in EUR/MWh which is paid by both buyers and
sellers.
9 Bidding areas
Actually, chapter 6 describes how the so-called System Price is calculated. The System
Price is the theoretical, common price we would have in the Nordic area if there were
no grid bottlenecks.
Due to the bottlenecks, the Nord Pool Spot exchange area is divided into a number of
bidding areas. For example, when a producer in Eastern Denmark sends his orders to
Nord Pool Spot, he must specify that these orders are submitted for delivery in the
bidding area Eastern Denmark.
The TSOs decides the number of bidding areas its boundaries. Eastern Denmark and
Western Denmark are always treated as two different bidding areas. Sweden
constitutes one bidding area until November 2011 when it is to be divided into four
bidding areas. Also, Finland, Estonia and Lithuania constitutes one bidding area while
Norway currently (August 2011) has five bidding areas.
11
Nord Pool Spot calculates a price for each bidding area for each hour of the
following day.
Naturally, there are often hours, where neighboring bidding areas have the same
price. Likewise there may also be hours, where the whole exchange area has the same
price: for example, during 2010, the whole exchange area had the same day-ahead
price during 10 % of the hours.
10 Day-ahead Congestion Management: Implicit Auction

Apart from calculating day-ahead prices, the Elspot market is also used to carry out
day-ahead congestion management in the Nord Pool Spot exchange area through an
implicit auction.
In the price calculation supply and demand orders are aggregated. The intersection of
the curves gives the market price and turnover. Depending on available transmission
capacity in the transmission grid, the spot markets in the different bidding areas are
integrated to maximize the overall social welfare in both (or more) markets.
Some areas have surplus of power while others have deficit of power. The area in
deficit is dependent on import from surplus areas. If there is insufficient transmission
capacity between the two areas bottlenecks occur and price differences arise. The
surplus area will have a lower price than the deficit area as more power is available
compared to consumption.
The export of power from surplus area to deficit area is reflected as an additional
purchase in the surplus area, and additional sale in the deficit area. An example with
Norway as surplus area and Sweden as deficit area is used to illustrate the principles.
The demand supply curves are chosen randomly. If no transmission capacity were
available between the two areas they would have different prices. Norway would
have a price of 200 NOK/MWh, while Sweden would have a price of 300 NOK/MWh.
12
Figure 10: Export of 50 MW from Norway to Sweden
Figure 11: Import of 50 MW to Sweden from Norway
Assume there is 50 MW available transmission capacity between Norway and

Sweden. The price in Sweden would be lowered to 233.33 NOK/MWh due to
additional available production. The price in Norway would increase to 283.33
NOK/MWh due to higher consumption.
13
Available transmission capacity
0 MW exchange
50 MW exchange
CONGESTION
[NOK/MWh]
200
300
[NOK/MWh]
233.33
283.33
Bidding area
Norway
Sweden
In the implicit auction the available transmission capacity is used to level out price
differences as much as possible.
Nord Pool Spot carries out the day-ahead congestion management on both external
and internal transmission lines between and within Denmark, Norway, Sweden,
Finland, Estonia and Lithuania.
11 Day-ahead Congestion Management: Market Coupling

When a bottlenecks occur on the cross-border between two exchange areas the flow
on this needs to be determined in cooperation between the two involved power
exchanges. This is called market coupling and exists today between the Nordic,
German and Central Western European exchange areas, which is shown in Figure 12.
Each single power exchange gives their orders to a central organ, European market
coupling company (EMCC). EMCC performs a price calculation with implicit auction
which determines the power flow on the cross-border lines, shown as green in Figure
12. This flow is used as input in the local price calculation performed by each single
power exchange.
14
Figure 12: Market coupling between the Nordic, German and Central Western European exchange
areas.
12 Cross-border trading
Inside the Nord Pool Spot exchange area, all the transmission capacity on the external
transmission lines is handled by Nord Pool Spot through implicit auction during price
calculation.
Two Nordic commercial players situated in different bidding areas cannot trade
electricity with each other. This is because Nord Pool Spot handles all the trading
capacity on the cross-border links, on behalf of the Nordic TSOs.
In order to trade with each other, Nordic players in different bidding areas can use
the financial electricity market (Figure 13). The two players can trade the power on
Nord Pool Spot or with a player situated in their own bidding area (i.e. the power is
traded locally). In addition the two players have a settlement in accordance with the
financial contract.
15
Figure 13: The capacity on the Nordic bottlenecks is given to E.E. (Electricity Exchange). How can a
producer P and a retailer R trade, if they are separated by one or more bottleneck(s)? Answer: They
trade the power with E.E. or with another local counterpart. Furthermore, they have a financial
contract.
The idea of this principle is the following: you can always buy or sell electrical power.
For example, you can trade on the electricity exchange. Hence, what is interesting for
the commercial players is only the price. However by means of a financial contract,
the players can lock the price.
13 The financial electricity market
At the financial electricity market you cannot trade one single kWh. As mentioned
above, the financial market is used for price hedging and risk management.
Figure 14 illustrates how a financial contract works. The example illustrates a
financial contract of the type called a futures contract.
In the example, a retailer and a supplier have entered into a futures contract with a
volume of 4 MWh and a hedge price of 65 EUR/kWh. In the example, the contracts
so-called delivery period is a specific month (for instance, it may be June 2014).
16
Figure 14: Producer and retailer sign a future contract with hedge price 65 EUR/MWh; If, for instance,
the average system price in the month concerned turns out to be 66 EUR/MWh; The producer pays the
retailer 1 EUR/MWh * 4 MWh. If, for instance, the average system price in the month concerned turns
out to be 63 EUR/MWh; the retailer pays the producer 2 EUR/MWh * 4 MWh. In the example, the
parties have cleared the contract. Hence, the settlement runs via clearing house.
The parties have a mutual insurance (and a mutual obligation). Suppose the average
system price for the month in question turns out to be 66 EUR/MWh. A high price on
the wholesale market is obviously disadvantageous for the retailer. However in this
situation, the supplier will compensate the retailer. The supplier pays the retailer
1 EUR/MWh * 4 MWh = 4 EUR.
Suppose instead the average system price for the month in question turns out to be
63 EUR/MWh. A low price on the wholesale market is obviously disadvantageous for
the supplier. In this situation, the retailer will compensate the supplier. The retailer
pays the supplier
2 EUR/MWh * 4 MWh = 8 EUR.
The contract is therefore settled by comparing the hedge price of the contract with
the average system price for the period in question. The difference in price is
multiplied by the contracts volume. Eventually, this amount of money is transferred
between the parties.
It is important to note that the parties of a financial contract are not delivering
physical power with each other. Only money is exchanged between them (therefore,
the name financial electricity market). However, in addition, the retailer may
submit a purchase order with an unspecified price to Elspot. The retailer can notify
Elspot that he will buy 5 MWh each hour during the month irrespective of the price.
With a purchase of 5 MWh each hour during the whole month, the retailer will in total
have bought 3600 MWh by the end of a 30-days month:
5 MWh/h * 24 h * 30 days = 3600 MWh.
17
The retailer does not need to worry about the price. If it is higher than 65 EUR/MWh,
he will be compensated. On the other hand, if the price is lower than 65 EUR/MWh,
he has to compensate the opposite party of the futures contract.
The retailer, therefore, has two trade arrangements: a purchase on Elspot and a
futures contract. In total, the two trade arrangements guarantee his price for the
3600 MWh will be 65 EUR/MWh.
14 Clearing of financial contracts
The two parties of a financial contract can choose to clear the contract using a
clearing house. In this case, the clearing house takes care of the settlement of the
contract (Figure 14). Furthermore, the clearing house guarantees the settlement: the
clearing house will ensure that the settlement is carried out, even if one of the parties
cannot fulfill his obligations.
If the parties have entered the contract via a financial electricity exchange, clearing is
mandatory. This is because the trading at the financial exchange is anonymous: the
parties do not know each others identity. Hence, the contract must be cleared, so the
clearing house sits between the parties.
15 Long-term contracts
At Elspot, the commercial players can trade power day-ahead. Now, let us take a look
at the market for long-term contracts.
For example, let us consider a retailer who has sold 100 MWh to an end user at a
price of 67 EUR/MWh for the following year. The retailer now has to make a
corresponding purchase on the wholesale market.
However, the retailer does not need to buy the power immediately. In order to hedge
his position, all the retailer needs now is a futures contract. For example, the retailer
has earned 2 EUR/MWh if he enters into a futures contract with a hedge price of 65
EUR/MWh.
Next year, the retailer can simply buy the power from Elspot or from a local supplier.
Therefore, the financial market is also the market for long-term contracts.
16 The day-ahead price must be reliable
As it appears, the Elspot day-ahead price is used, when the financial contracts are
settled. We say that the day-ahead price is the underlying reference for the financial
contracts.
18
A reliable day-ahead power price is an absolutely essential basis for a financial

market. It is imperative that all the players regard the day-ahead price as the true
market price. For obvious reasons, only in this case the players will be interested in
making financial contracts, with the day-ahead price as the underlying reference.
Through Nord Pool Spots Elspot market such a reliable day-ahead price in the Nord
Pool Spot exchange area has been created.
17 Why an electricity exchange?
For society, the Elspot market provides price transparency. For example at
www.nordpoolspot.com, everybody can see the wholesale markets day-ahead price.
In addition, the day-ahead price is used as the underlying reference for financial
electricity exchanges. Via their quotation of financial contracts, price transparency is
also provided for long-term contracts. For example, via the financial electricity
exchanges, you can see the market players estimate of next years electricity prices.
The electricity exchange also provides another service to society: the electricity
exchange handles transmission capacity in a market-oriented way. With this, there is
a neutral and fair day-ahead congestion management. The system secures that the
day-ahead plans send the commodity in the right direction: from low-price areas
towards the high price areas.
19
Singapore energy market

September 2012
Further information
If you would like further information on any aspect of this client
note, please contact a person mentioned below.
Contact
Alex Wong
Partner
T +65 6302 2557
alex.wong@hllnl.com
Amy Lee
CEO
T +65 6302 2558
amy.lee@hllnl.com
Ming Hui Chock
Senior Associate
T +65 6302 2560
minghui.chock@hllnl.com
Adrian Wong
Senior Associate
T +65 6302 2568
adrian.wong@hllnl.com
This note is written as a general guide only. It should not be

relied upon as a substitute for specific legal advice
Contents
Introduction Recent developments in the Singapore energy industry
Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations
Schedule 2 Summary of Singapore gas market deregulation and the Gas Network Code
15
With the merchant power market celebrating its first decade of

operations and the gas market liberalisation recently
completed, Singapore is now looking to finalise its long term
energy market strategy. This note serves as an update on
some of the recent developments in the Singapore Energy
Industry as it enters a crucial phase of development. The
attached schedules also present a summary of the functions
of the liberalised Singapore Electricity and Gas Markets.
POWER MARKET UPDATE
The Singapore power market has witnessed, over the last few
years, both expansion and consolidation as it looks forward to
the coming years of energy market development. From a
consolidation perspective, almost all of the large incumbent
power generators have undergone repowering exercises and
moved almost exclusively towards CCGT plants for power
generation as they reposition themselves in a more
competitive wholesale market for power generation. Today,
more than 80% of Singapores electricity is fuelled by piped
natural gas.
Singapore also continues to encourage energy efficiency as it

moves towards a greener future.
The most direct
manifestations of this include the continued steady growth of
the National Environment Agencys waste to energy plants
and the likely increase in the number of buildings taking
advantage of rooftop solar to reduce their carbon footprints.
This is supported by the recently passed Energy Conservation
Act 2012 which aims to mandate good energy management
practices for large energy users and the support provided by
the Economic Development Board (EDB) for the
establishment of an energy efficiency fund in Singapore to
grow businesses that focus on energy efficiency. All this
means that there will continue to be overcapacity in the
generation market for the foreseeable future and, at least in
2
theory, the consumers should benefit from this competition .
Expansion of the power generation sector has mainly come

from new market entrants GMR Energy who will bring 800MW
of gas fired units online over the next two years. After an
almost decade long fight for access to gas, the long awaited
project achieved financial close in 2011. GMR Energy is not
alone though in the contest for space in the generation
segment of the market. Onsite and captive power generation
1
remain permitted activities subject to certain restrictions and
the decision to allow Hyflux (through Tuaspring Pte Ltd) to join
in the market for power generation as part of the second Tuas
Desalination project also caught market watchers by surprise.
2
1
Policy on Self-Supply of Electricity Information Paper, Energy Market

Authority of Singapore (EMA), 21 April 2008
Full retail contestability has still not been achieved so household and small
consumers retain the Market Support Services Licensee as their default
power retailer.

(contd)
Increased competition in the generation industry however

almost immediately translates into a battle for the cheapest
fuel source for power generation. For a large part of the
foreseeable past, this has been piped natural gas from
Indonesia and Malaysia. This dominance (which not only
fuels power generation but also major industries in the
petrochemicals, electronics and biomedical sectors) has been
deliberately challenged in Singapore in the recent past by the
importance
the
Singapore
government
places
in
diversification of fuel sources and by the global fall in LNG
prices (particularly long-term prices). It is Singapores longterm strategy for fuel mix diversification for power generation
and industry that will dominate this discussion in the coming
months.
LNG AND ITS IMPACT ON THE SINGAPORE GAS
MARKET
Singapores current gas import consists almost exclusively of
piped natural gas imported through four pipelines from South
Sumatra and West Natuna gas fields in Indonesia and the gas
imports from Malaysia.
Significant focus has been placed by the Singapore

government in reducing the reliance on these piped natural
gas imports and increasing gas imports from other sources.
This led to the Singapore government moving ahead with the
development of its first LNG import and regasification terminal
and the appointment of BG Asia Pacific Pte Limited (BG) as
Singapores first LNG aggregator.
At the time this decision was made, piped natural gas into
Singapore was significantly less expensive than the
comparable LNG price but the decision to proceed was
nonetheless made on strategic grounds and ignoring the
immediate economics of the decision. Singapore had to
create an artificial demand for LNG in order to wean the
market off its reliance on piped natural gas. This artificial
market was created by simultaneously requiring power
generators to accept LNG as part of their fuel portfolio as well
as imposing a moratorium on future piped natural gas imports.
In hindsight, the strategy appears to have been prescient.
With new sources of gas flooding the global market (both
traditional gas fields in Australia (which is set to overtake
Qatar as the largest LNG exporter by the end of the decade)
and East Africa and shale gas in North America), LNG prices
are set to fall in the medium to long term in spite of strong
demand from China and Japan. At the same time, the West
Natuna and South Sumatra gas fields in Indonesia appear to
be depleting at a higher rate than originally envisaged and
this, coupled with pressure on the Indonesia government to
reserve more of its precious resources for domestic
consumption, would likely mean a far less reliable future for
piped natural gas imports.
This has led to BGs 3 Mtpa franchise for aggregating LNG for
Singapore having a strong uptake. With 2.65 Mtpa of LNG
having been taken up by February 2012, the EMA expects the
entire franchise to be fulfilled by 2013 at the latest. This begs
the immediate question of Singapores LNG import policy
beyond BGs franchise. The EMA has sought feedback
3
through a consultation exercise that will aim to determine
whether Singapores future LNG import framework will be
carried out through a regulated sole importer framework
(BG+1) or a multiple aggregator framework (BG+3).
Under the BG+1 framework, the EMA will appoint a Regulated
Sole Importer (RSI) who will import all incremental LNG
beyond BGs 3 Mtpa supply. The EMA will regulate the RSIs
returns as well as its LNG procurement, gas sales prices and
contract terms. A variant of this framework is adopted by
Asian importers like South Korea, Taiwan and Thailand. It is
expected that the appointment will be conducted through a
similar RFP exercise that resulted in BGs appointment.
The BG+3 framework would see up to 4 large LNG importers
serving up to 15 Mtpa of LNG capacity (which is the projected
LNG demand by 2024) in Singapore. Under the government
mandated aggregation variant of this model, importers will be
selected by EMA (either sequentially over time or
concurrently) and awarded LNG import licences through a
competitive RFP process. Each importer would be awarded a
franchise to import a specific amount of LNG. Under the
market-driven aggregation model, competition between
players in the LNG import sector would give rise to natural
aggregation of demand into a few dominant players (e.g. the
formation of a few buyer groups). EMA would set entry
criteria that importers must fulfil to qualify for access to the
LNG terminal.
With the LNG terminal able to support 7 LNG storage tanks
and up to 15Mtpa of LNG demand, the Singapore government
is also open to seeing the LNG terminal being used as a
trans-shipment and trading hub. As of today, a number of
international LNG trading companies (such as Shell, GDF
Suez, ConocoPhillips and BP) have set up LNG trading
offices in Singapore and they will be looking to tap the strong
regional growth of spot and short-term LNG contracts. There
should be no doubt though that the government sees the LNG
3
LNG Procurement Framework Consultation Paper, EMA, 30 March 2012

(contd)
terminals first priority as that of helping to ensure security of

fuel supply for Singapore.
selection regime, any electricity import selection will likely be

subject to a similar tender process.
With decisions on the LNG import strategy likely to be made

in the coming months, the EMA will likely also have to resolve
the wider question on the long term fuel mix strategy and the
on-going question of how LNG imports will affect the import of
piped natural gas. Given the long term investments and
decisions required with respect to new piped natural gas
import contracts, the EMA is aware that it will not be able to
make sharp changes in direction once its long term policy and
strategy have been settled. Amongst other things, the EMA
will have to make challenging decisions on issues such as
LNG imports into the nearby Malaysian LNG import terminals.
Where regasified LNG is piped into Singapore from these
terminals, they will likely be considered to be piped natural
gas and thus subject to EMAs current moratorium. The
extent to which the moratorium may be lifted for such future
imports (and thus a potential direct challenge to the business
of the Singapore LNG terminal) will be a difficult decision that
the EMA will have to make in terms of striking a balance
between market forces and price competitiveness and
security of supply.
In summary, the Singapore energy market is best described

as a managed privatised market. Singapore is adamant not
to fall into the cycle of underinvestment and high energy
prices that befalls many other privatised energy markets
around the world but yet wants to take full advantage of
deregulated market practices. The strategy appears to have
worked so far but its continued success will depend on the
ability of the EMA and the Singapore government as a whole
to balance the supply and demand pressures of the merchant
market with the strategic priorities of the city state.
SINGAPORE ENERGY MARKET INVESTMENTS IN THE

NEAR FUTURE
The business of the Singapore LNG terminal is, of course, of
significant interest to the EMA and the Singapore government
as a whole. The initial intention was for the LNG terminal to
be a self-sustaining business model to be run independent of
government subsidy (but albeit subject to regulation by the
EMA).
The decision for EMA to take over terminal
development reflected the urgency and importance with which
the government saw the terminal as a cornerstone of its
energy strategy. That said, it remains the intention of the
government to, at an appropriate juncture, divest the
ownership of the LNG terminal to the private sector (perhaps
similar to the way in which the previous Temasek owned
generating companies were divested in 2008) a decision
that will no doubt attract significant market interest.
Whilst natural gas dominates the discussions on energy policy
and strategy, there remain other elements of the
governments overall energy diversification strategy that
deserve a mention. The EDB has released an RFP for a
synthetic gas facility (through coal gasification) to be
developed on Jurong Island to provide feedstock for industry
and potentially also for power generation. The ability of the
syngas facility to provide fuel for power generation is
nonetheless subject to EMAs overall study and policy
finalisation on gas import control. Electricity imports also
remain on the horizon (with a possible 4-5 year timeline from
2012). Subject to a restriction of 600MW per country (under
the ASEAN Grid masterplan) any such import will have to be
licensed by the EMA and will likely involve a contracts-fordifferences pricing structure.
Similar to the aggregator
Schedule 1 Summary of Singapore electricity market deregulation and

wholesale market operations4
HISTORICAL BACKGROUND
REGULATORY FRAMEWORK
The electricity and piped gas industries in Singapore have

traditionally been vertically integrated and government-owned.
The Public Utilities Board (PUB) was formed in May 1963 to
undertake the supply of water, electricity and piped gas to the
population of Singapore.
Electricity Act
In 1995, the Government began to implement a number of

changes to deregulate the electricity industry. On 1 October
1995, the PUB transferred its electricity and gas activities to
Temasek Holdings. Within Temasek Holdings, Singapore
Power was created as the holding company for several other
new companies including the generation companies,
PowerSenoko (now known as Senoko Power) and
PowerSeraya; the transmission company, PowerGrid; and SP
Services Ltd, the electricity supply and utilities support
services company. A further generator, Tuas Power, was set
up as an independent company directly under Temasek
Holdings.
THE NATIONAL ELECTRICITY MARKET OF SINGAPORE
4
(NEMS)
On 1 April 2001, the Government established a body
corporate, the EMA, under the Ministry of Trade and Industry
(MTI), to regulate, among others, the electricity industry. In
that same year, PowerGrid transferred its system operator
function to the Power System Operator (PSO), and the
market operator function and pooling and settlement
responsibilities to the Energy Market Company Pte Ltd
(EMC), which was formed as a subsidiary of the EMA to
operate the Singapore Electricity Pool (SEP) and
subsequently the wholesale electricity market in the NEMS.
The NEMS consists of a wholesale electricity market and a
retail electricity market. The wholesale market consists of two
markets: the real-time (or the spot market) for energy,
reserve and regulation; and the procurement market for
other ancillary services. The sale of energy, reserve and
regulation are done through price/offer quantities submitted by
generation companies every half hour.
The Electricity Act is the principal legislation governing the

electricity industry and the NEMS. The principal rights and
obligations of the participants in the wholesale and retail
electricity markets are set out in the Singapore Electricity
Market Rules (the Market Rules), the electricity licences
and in the codes of practice (the Codes of Practice) issued
by the EMA.
The Singapore Wholesale Market Rules
The Market Rules are effectively contracts between each
market participant and the EMC under section 49 of the
Electricity Act. This ensures that market participants have the
option to take legal action against the EMC for damages
sustained as a result of the non-observance of the Market
Rules by the EMC and vice versa. The Market Rules also
contain dispute resolution procedures.
The objectives of the Market Rules are:
to provide market participants and the Market Support

5
Services Licensee (the MSSL) with non-discriminatory
access to the transmission system; and
to facilitate competition in the generation of electricity.
Under the Electricity Act, an entity may not engage in certain

electricity-related activities unless it has been issued with an
electricity licence by the EMA (or it has been exempted from
holding one). The electricity-related activities that require an
electricity licence are:
EMA Introduction to the National Electricity Market of Singapore (version

6) October 2010.
to establish and govern efficient, competitive and reliable

markets for the wholesale selling and buying of electricity
and ancillary services in Singapore;
Electricity Licenses
In addition, as part of the governments policy of separating

ownership of electricity generation assets from ownership of
the Transmission and Distribution Systems, Singapore Power
divested its ownership interests in Senoko Power and
PowerSeraya to Temasek Holdings. Temasek Holdings
subsequently divested Tuas Power, Senoko Power and
PowerSeraya to private sector investors.
operation of any wholesale electricity market;
generation of electricity;
transmission of electricity;
provision of market support services (such as meter

reading and meter data management);
retail of electricity;
It should be noted that although the MSSL is obtaining supply from the
wholesale market, it is not technically a market participant. However, the
Market Rules provide for MSSL to be treated, for the most part, similarly to
the manner in which market participants are treated. Thus, a MSSL is
subject to most of the same obligations as market participants are under
the Market Rules.

wholesale market operations (contd)
trading in the wholesale electricity market; and
importing or exporting electricity.
Codes of Practice
The electricity licences require that licensees comply with
relevant Codes of Practice and other standards of
performance that govern their activities. The Codes of
Practice contain detailed rules that govern the electricity
licensees in conducting their activities. The Codes of Practice
developed to date include:
The Transmission Code
The Transmission Code is binding on the Transmission
Licensee, which is SP PowerAssets. It sets out the minimum
conditions that SP PowerAssets must meet in carrying out its
obligations as owner of the Transmission System and to
facilitate non-discriminatory access to the Transmission
System.
Regulated Supply Service Code
The Regulated Supply Service Code is binding on MSSLs,
which is currently SP Services only. It sets out the minimum
conditions that a MSSL must meet in carrying out its
obligations to procure the supply of electricity and provide
market support services to non-contestable consumers under
section 21 of the Electricity Act.
Market Support Services Code
The Market Support Services Code (the MSS Code) is
binding on MSSLs. It sets out the minimum conditions that a
MSSL must meet in carrying out its obligations to provide
market support services to Retail Electricity Licensees
(RELs) and contestable consumers, and facilitate their
access to the wholesale electricity market.
Metering Code
The Metering Code is binding on the Transmission Licensee,
generation licensees and MSSLs, and sets out the minimum
conditions that a metering equipment service provider must
meet in carrying out its obligations to install and maintain
meters. It also sets out the roles and obligations of the meter
reader and meter data manager.
Codes of Practice for RELs
The Codes of Practice for RELs sets out the minimum
standards of behaviour that a REL must observe in retailing to
consumers.
Market Agreements and Contracts
Most market participants in the NEMS are required to enter
into a number of agreements and contracts. These are
generally a consequence of their respective licence
conditions, the Market Rules and Codes of Practice. The

table in page 6 sets out the various market agreements and
contracts in the NEMS.

Agreements and Contract Parties
Purpose
Operating Agreement
PSO and SP Power Assets
The Operating Agreement gives the PSO the authority to direct

the operations of the Transmission System subject to certain
limitations on the manner and extent of those operations.
MSSL EMC Agreement
EMC and MSSLs
This agreement establishes a contractual relationship between

the EMC and the MSSL and provides that the Market Rules will
have the effect of contract as between the EMC and the MSSL
in so far as it applies to them.
MSSL Market Participant

Agreement
MSSL, generation licensees

and direct market participants
(DMP)
This agreement provides for meter reading services for

wholesale settlement. The generation licensees pay directly for
the services with fees set by the MSSL.
PSO Market Participant

Agreement
PSO, generation licensees,

RELs and DMP
This agreement establishes a contractual relationship between

the PSO and the market participant and provides that the
Market Rules will have the effect of contract as between the
PSO and the market participant in so far as it applies to them.
MSSL Agreement
MSSL and RELs
This agreement is a services agreement between the MSSL as

provider and the RELs as procurers in relation to the various
customer support services as defined in the MSS Code.
Connection Agreement
SP PowerAssets, generation
licensees, party wanting DMP
and electricity consumers
This agreement gives effect to the obligations that must exist

between SP PowerAssets and the party wanting connection
service.
Retailer Use of System

Agreement
SP PowerAssets and RELs
This agreement is for the collection of Use of System charges

(UoS Charges), that is, the transmission and distribution
tariffs, when a REL opts for consolidated billing, that is, when a
REL assumes the payment responsibility of its customers for
the transmission charges.
Agency Agreement
MSSL and SP PowerAssets
This agreement is an agency agreement for the provision of

UoS Charges collection services for SP PowerAssets.
Ancillary Services Agreement
EMC on behalf of the PSO

and the generation licensees
This agreement is a contract between the EMC and a market

participant (usually a generation licensee) supplying ancillary
services.
Source: EMA

ELECTRICITY INDUSTRY STRUCTURE
facilitate the planning

Transmission System;
provide information and other services to facilitate

decisions for investment and the use of resources in the
electricity industry; and
exercise and perform the functions, powers and duties

assigned to the EMC under the EMA Act, its electricity
licence, the Market Rules and applicable Codes of
Practice.
The EMA
The EMA was established in April 2001 pursuant to the EMA
Act as an independent regulator overseeing the electricity and
gas industries in Singapore. Under section 3 of the EMA Act,
the EMA is charged with the general administration of the
EMA Act, and its functions and duties include:
to perform the interests of consumers with regard to

prices, reliability and quality of services;
to perform the functions of economic and technical

regulator;
to ensure that electricity licensees provide an efficient

service;
to ensure security of supply of electricity to consumers

and to arrange for the secure operation of the
transmission system;
to protect the public from dangers arising from electricityrelated activities;
to create an economic and regulatory framework for the

electricity sector that promotes competitive, fair and
efficient market conduct and prevents the misuse of
monopoly or market power;
to advise the Government on matters relating to the

electricity system; and
In fulfilling these functions, the EMA has at its disposal a
number of regulatory tools and powers. These include the
authority to issue, suspend, revoke or modify an
electricity licence; the power to issue and modify codes of
practice and other standards of performance; the power
to issue directions to electricity licensees; the power to
fine electricity licensees; and the authority to investigate
and sanction anti-competitive conduct.
The EMC
The EMC is licensed to operate the wholesale electricity
market in the NEMS. The EMCs functions are to:
operate and administer the wholesale electricity market in

the NEMS;
prepare schedules for generation facilities, loads (that is,

the withdrawal of electricity from the Transmission
System) and the Transmission System;
settle accounts of market participants;
and
augmentation
of
the
The EMC is a 51:49 joint venture between the EMA and M-Co
(The Marketplace Company) Pte Limited (M-Co Singapore).
M-Co Singapore is a related company of The Marketplace
Company Limited, which developed, implemented and
currently operates the wholesale electricity market in New
Zealand.
PSO
The role of the PSO (a division of EMA) is to ensure the
security of supply of electricity to consumers and to arrange
for the secure operation of the electricity system.
The functions of the PSO include:
maintaining the reliability of the electricity system;
forecasting and reporting

Transmission System;
coordinating the outages of generation facilities;
providing Transmission System status and load

forecasting to the EMC for the purposes of market
clearing;
coordinating the actions of the EMC and market

participants during emergencies; and
dispatching generation facilities.
on
conditions
on
the
Market Participant
A Market Participant in the NEMS is defined as a person (that
is, an entity or organisation, as well as people) that:
has an electricity licence issued by the EMA; and
has been registered with the EMC as a market

participant.
The wholesale electricity market is a mandatory market in the

sense that any person who wishes to convey electricity over
the Transmission System must be registered as a market
participant with the EMC. Market participants may be:

the Transmission Licensee;
generation licensees;
RELs;
persons, other than the generation licensees and RELs,

who have been licensed to trade in the wholesale
electricity market; and
any department of the Government that generates

electricity before 1 April 2001.
A MSSL is not a market participant.

In the NEMS, it is mandatory for all generation facilities above
to be licensed by the EMA. It is also mandatory for
generation facilities above 10MW to be registered for dispatch
by the PSO.
Mandatory registration ensures that all
generation facilities of any significant size are subject to the
Market Rules.
Transmission Licensee
SP PowerAssets is currently the sole Transmission Licensee
in Singapore.
The responsibilities of SP PowerAssets and of the persons
whose facilities are connected to the Transmission System
are set out in the Transmission Code and the connection
agreements.
The Market Rules also contain specific
provisions for SP PowerAssets and the PSOs obligations in
respect of the reliability and security of the Transmission
System.
RELs
The retail electricity market does not come under the
jurisdiction of the Market Rules and the EMC. It is created
and regulated under the EMA Act, the electricity licenses and
the Codes of Practice issued by the EMA.
RELs may be market participants who purchase electricity
directly from the wholesale electricity market or purchase
through the MSSL. Since the RELs are permitted to trade in
electricity and are not subject to the same degree of
regulation as the MSSL, they may offer contestable
consumers contracts different from those available from the
MSSL. RELs can bundle energy and other charges into a
single invoice, charge a price other than the Uniform
6
Singapore Energy Price (the USEP) for energy, and offer
additional services to consumers.
Consumers
Consumers are classified as either contestable or noncontestable, depending on their electricity usage.
Contestable consumers are entitled to purchase electricity
from a REL, or directly from the wholesale electricity market,
or indirectly through MSSLs. Non-contestable consumers are
supplied by MSSLs.
Currently, consumers with a monthly usage of 10,000kWh
and above are contestable. The EMA continues to study
when full contestability of all retail consumers will be allowed.
Relationships between the Market Participants
The figure on page 9 shows the financial flows between the
Market Participants in the NEMS.
The transmission network transports electricity at high voltage

from generators to the low voltage distribution network (or, in
a small number of cases, directly to large industrial
consumers).
SP PowerAssets, being the monopoly provider of
transmission services, is not permitted to compete in the
energy market, whether as a generator, retailer or trader
(either directly or indirectly by ownership of companies
engaged in such activities), because opportunities exist for it
to afford a preference to its competitive activities or its
competitive affiliates.
Market Support Services Licensee
SP Services is currently the sole MSSL and provides market
support services to the majority of electricity consumers in
Singapore. SP Services charges regulated fees, as approved
by the EMA, for market support services provided.
6
Please refer to page 13 for further discussions on the USEP.

Financial Flows between the Market Participants in the NEMS
Competition
OVERVIEW OF WHOLESALE MARKET OPERATIONS
An important role for the EMA is the monitoring of market

behaviour to guard against the concentration of market power,
abuse of dominance and exercise of anticompetitive
behaviours such as collusion and capacity withdrawal to drive
up prices. The EMA is also empowered to prohibit anticompetitive practices including restrictions on mergers and
acquisitions (that is, there will be cross ownership limitations).
In the NEMS, the real-time dispatch of electricity (scheduling

generators to supply energy, reserve and regulation) is
determined by the operation of a wholesale spot market run
every half-hour. Generators offer their capacity (specifying
price/quantity pairs) to the market and the PSO provides a
prediction of the expected load along with any system
constraints for that half-hour. The market then determines the
least-cost dispatch quantities and the corresponding market
clearing prices based on the offers made by generators.
Significantly, the anti-competition provisions in the Electricity

Act have retrospective effect. Where the EMA finds that a
person has acted in an anti-competitive manner, the EMA can
give a direction for the cessation of such conduct and impose
a financial penalty on the person.
The electricity industry is excluded from the Competition Act
2004 as the EMA will continue to regulate anti-competitive
practices and other competition issues in the electricity
industry.
The wholesale electricity market consists of two markets:
the spot market for energy, reserve and regulation; and
the procurement market for other ancillary services.
Every half-hour, the Market Clearing Engine (MCE), a linear

programming computer model, determines the spot market
outcomes for:
EMA Introduction to the National Electricity Market of Singapore (version

6) October 2010.
10

the dispatch quantity produced by each generation

facility;
the reserve and regulation capacity required to be

maintained by each generation facility; and
the corresponding wholesale spot market prices for

energy, reserve and regulation.
Quantities and prices are based on price/quantity offers made

by the generators and load forecasts prepared by the EMC
based on demand forecast information received from the
PSO.
RESERVE AND REGULATION
Reserve is unused capacity that has to be made available to
the electricity system quickly to correct any imbalance and
maintain reliable supply in case of an unexpected outage of a
scheduled generation facility. This capacity must be able to
be in production within a short timeframe, depending on the
arrangement. There are three reserve classes: primary
reserve (eight seconds response), secondary reserve (30
seconds response) and contingency reserve (10 minutes
response).
Regulation, or load-following, is a normal operational
requirement to cover second-to-second variations in load
away from estimated load.
DISPATCH SCHEDULING
Dispatch scheduling is the process of matching the generation
capacity needed to meet forecast demand. It is at the heart of
running an electricity system. To enable the MCE to generate
the dispatch schedule, the PSO and the generation facilities
dispatch coordinators need to know in advance when each
generation facility will be operating and how much output is
expected from each. The dispatch schedule comes from the
MCE.
electricity. There is no overall cheaper dispatch available, in

terms of the offers that have been made to the market, by
providing energy, reserve and regulation from different
sources, or in different quantities from the same sources.
This is the minimum cost market dispatch.
The supply of energy, reserve and regulation is specified by
means of offers that contain price/quantity tranches indicating
the quantity of energy, reserve or regulation that each
dispatchable generation facility is willing to supply at the
corresponding energy, reserve or regulation prices.
With the dispatch based on estimates of the demand for the
coming dispatch period, the market prices are similarly the
prices set according to those estimates. This form of price
setting, called ex ante pricing (pricing before the event), gives
the market participants certainty about market prices even if
8
dispatch quantities differ from those scheduled .
The MCE does not produce a single market energy price
because of the effects of losses and congestion on the
transmission system.
Different energy prices apply to
different nodes on the transmission system.
THE OFFER PROCESS
Energy, Reserve and Regulation Offers
Generators make offers to supply energy, reserve, and
regulation for each of their units in each half-hourly dispatch
period in which they want to operate. They are similarly
permitted to offer interruptible load to supply reserve. Offers
can vary for each half-hour, and are assumed to stand, unless
modified, from the time they are made through to dispatch.
The market does not distinguish between offers used for the
market outlook, pre-dispatch and real-time processes. It
simply uses the most recent offer made for each half-hour.
Key features of the generator offer process are:
standing offers are required. Generators are required to

make standing offers into the market. The standing
offers form a pattern for a week. The use of standing
offers is particularly valuable for smaller generators, since
it eases the administrative burden of participating in the
market;
continuous adjustment of offers. Market participants are

allowed to continually adjust their offers up to gate
9
closure ;
The PSO instructs the generation facilities to conform to the

dispatch schedule. Any deviations from the estimated load
and corresponding schedule are handled by the PSO using
ancillary services.
THE MARKET CLEARING ENGINE
Every half-hour, the MCE is run to determine the dispatch
schedule and the associated energy market prices for the
upcoming dispatch period. The MCE also determines which
generation facility is on reserve and regulation duty along with
the market prices for reserve and regulation.
The objective of the MCE is to find a set of dispatch
instructions that minimises the cost of supplying load at all
nodes (injection or exit points) of the Transmission System, as
well as meeting the reserve and regulation requirements for
For all but plants providing regulation, energy actually injected should not
differ greatly from scheduled energy, under normal circumstances.
Although offers modified within the last hour may be subject to scrutiny
from the market surveillance panel.
11

up to 10 price/energy quantity bands. Generators may

make energy offers consisting of up to 10 price/quantity
bands (tranches) for each facility for each half-hour;
up to five price/reserve and regulation quantity bands.

Generators and interruptible load may make reserve
offers (of different classes) and generators may make
regulation offers if they are registered to do so;
combined offers. Energy, reserve and regulation are all

offered simultaneously, and are co-optimised by the
market clearing model. The model respects the trade-
offs between the commodities so that a facility will not be

scheduled to produce more energy, reserve and
regulation than it can simultaneously manage; and
offers at a node. Energy offers for each generation

facility are made at the node where that facility is located.
Market Clearing Process in the Spot Market

Prices in S$ per megawatt hour (MWh)
12

MARKET CLEARING PROCESS

The figure on page 11 shows a simplified example of the
market clearing process in the spot market.
In this example, there are three generators (A, B and C)
whose offers consist of four price/quantity tranches each. The
tranches are arranged in ascending price order. The marketclearing price is found at the point where total demand from
consumers is met by the offer tranches. In this example, the
third tranche from Company C sets the market-clearing price
with its offer price of S$85/MWh. The total demand is a
forecast of the load for that period.
Offers below the Market Clearing Price are accepted and
those generation facilities are dispatched, in full, to the offers.
Offers above the Market Clearing Price are not accepted, and
so the generation capacity represented by those offers is not
taken at all. At the margin, the offer that sets the price is
usually only partially dispatched. The generation facility at the
margin is called the marginal unit (the Marginal Unit).
UNIT COMMITMENT
The NEMS is a self-commitment market. This means that unit
commitment is the responsibility of each generation company,
and no start-up or shutdown payments are made (generation
companies are expected to factor these payments in to their
offering strategy). This fact is important because some
generation units require a significant period of time to warm
up before they can produce electricity and hence need to be
committed some time in advance. The PSO needs to know
and account for the ability of the generation unit to ramp up or
down. This information is part of the standing capability data
required from each generation unit.
PRICES AND CHARGES
Energy Price
In common with many modern electricity markets, the NEMS
uses a form of energy pricing referred to as nodal pricing,
meaning that prices at each node in the network will be
influenced by the physical properties and constraints of the
transmission system. This results in the price of energy
differing at different physical locations on the network.
The MCE automatically produces a different price at each
node on the network. Dispatchable generators are paid the
nodal price at their point of injection.
Reserve and Regulation Price
Reserve is generation capacity that is required in case of an
unexpected outage of scheduled plant. Because generating
units may fail without warning, some reserve capacity has to
be made available to the system to correct any imbalance
quickly and maintain reliable supply. Reserve is a significant
factor in the Singapore system since some generating units
(600MW thermal units, for example) are large relative to the

total load.
A generator would wish to receive payment for the reserve
and regulation it provides because it forgoes the opportunity
of being dispatched fully, by being partially available for PSO
to call on it for reserve and regulation, as and when required.
Reserve in Singapore can be provided by generation facilities
10
and load . For a facility to provide reserve as quickly as in
eight seconds or even in 30 seconds, it needs to be already
spinning and synchronised. In most instances, that requires
the unit also to be supplying energy, with reserve capability
coming from its ability to ramp up its scheduled output very
quickly. Since not all plants have this technical capability, a
plant has to be certified as meeting the requirements for
registration to provide reserve before it can be offered in the
reserve market.
There are also different reserve provider groups for each
class of reserve. These groups represent the reliability of
different reserve sources in providing reserve, and their
effectiveness in curtailing falls in system frequency.
Since a facilitys capacity may be available for both energy
and reserve/regulation, the MCE must consider the optimal
trade-off between the offers for reserve, regulation and
energy. In solving the markets for each class of reserve and
regulation, the MCE simultaneously finds the lowest cost
solution (in terms of the offers made) that trades off between
these products for the various facilities. Within the MCE,
optimisation of the supply of energy must account for the
minimum running level of facilities that provide reserve. The
overall optimal solution may result in a unit being run out of
11
merit for energy so that the unit is available for reserve .
Offers for reserve from a generator can only be made in
association with a corresponding offer for energy. Part of the
standing capability date for the plant is a function relating its
reserve capability to its energy capability. This relationship is
entered into the MCE.
The cost of regulation is recovered from consumers and
generation facilities. The cost is allocated on a S$ per MW
basis across all MW of consumption in a dispatch period plus
the first 10MW of electricity dispatched by each generation
facility in that dispatch period.
10
An Interruptible Load scheme was introduced into the Singapore market in

2004.
11
Since often a facility must be running in order to be available for reserve, it

may be dispatched for energy even though its energy offer is higher than
that of the marginal plant for energy. This is acceptable because there is
no cheaper energy and reserve solution for the system as a whole. The
reserve price received by such a plant will compensate it for the shortfall
between its energy offer price and the energy spot price.
13

USEP
While the generation facilities are paid their nodal price,
buyers from the wholesale electricity market pay a uniform
overall average price so that no customers are disadvantaged
by location. The USEP is calculated from the weighted
average of the nodal prices at all of the exit nodes on the
Transmission System. The nodal energy price at each node
is weighted by the energy withdrawn from that node.
VESTING CONTRACTS
In the transition to the NEMS, the EMA had concerns with the
degree of market power that will exist in the wholesale
electricity market. The EMA has addressed these concerns
using Vesting Contracts without interfering with the structure
of the wholesale electricity market. Vesting Contracts are
contracts for differences (CfDs) vested on the large
incumbent electricity generation companies, for a transitional
period.
In Singapore, the Vesting Contracts take the following form:
generators will be required to enter into Vesting Contracts

with the MSSL, who is the counterparty to all of the
Vesting Contracts. The MSSL will then distribute debits
and credits associated with the contracts to consumers,
both contestable and non-contestable;
Vesting Contracts have a contract price (or strike price)

set at about the economic or long-run marginal cost
(LRMC) of a new entry generator (i.e. the electricity price
that a new investor in base-load capacity would require in
order to cover its fixed and variable costs at a reasonable
return to shareholders over the life of the plant). The
same strike price applies to all generators;
the contract quantity will be set to keep the market power

of large generators at an acceptable level. During peak
load times, the contract quantity will be a larger
proportion of total load, while in off-peak times it will be a
smaller proportion. The average contract quantity will
reduce over time as new capacity is built to mitigate the
market power of incumbents; and
the contract quantities for each generator are based on

the generation capacity for each company.
Vesting Contracts are settled by the EMC in the wholesale

electricity market and by the MSSL in the retail electricity
market. The figure on page 14 illustrates the settlement
process for Vesting Contracts.
14

BILATERAL CONTRACTS
The wholesale market in Singapore is not designed to
eliminate or be immune to price volatility; rather, it is important
to the market that prices move freely. As a result, the design
also recognises the need to allow participants to manage
price risk.
The generation and electricity retail companies can enter into
bilateral contracts, at their discretion, to reduce price
fluctuations.
These contracts are purely financial
arrangements, the most common of which are CfDs.
Under such an arrangement, the contracting parties agree to
a strike price for a given volume of energy. They continue to
buy and sell on the spot market but settle between
themselves any financial difference between the spot and CfD
strike price. When the strike price is higher than the spot
price, the electricity retail companies make a payment to the
electricity generation companies for the difference, and vice
versa.
Bilateral contracts create price certainty for the parties and
limit their exposure to potential volatility in the spot market.
Bilateral contracts are outside the wholesale electricity market
and are not taken into account in the physical dispatch
process, and are not in any way regulated by the Market
Rules. However, the facility exists for the parties to the
bilateral contracts to settle their contracts through the EMCs
settlement system.
Electricity Market Design for a Low-carbon Future

P E Baker, Prof: C Mitchell and Dr B Woodman
October 2010
THE UK ENERGY RESEARCH CENTRE
The UK Energy Research Centre carries out world-class research into sustainable future
energy systems.
It is the hub of UK energy research and the gateway between the UK and the
international energy research communities. Our interdisciplinary, whole systems
research informs UK policy development and research strategy.
www.ukerc.ac.uk
The Energy Supply Theme of UKERC

UKERCs energy supply research activities are being undertaken by the University of
Cardiff, Imperial College and the University of Exeter.
Energy Research Centre
UKERC/WP/ESM/2005/004
This paper considers GB electricity market and network regulatory arrangements in the
context of transitioning to a low carbon electricity system. By considering some of the
primary features of a low carbon electricity system and building on themes raised by a
previous UKERC Supply Theme paper (Baker, 2009), the paper attempts to identify what
characteristics an appropriate market and regulatory framework would need to posses.
The paper goes on to consider how existing market arrangements perform in these
areas and the possible need for change.
The aim of the paper is to contribute to the debate on energy market reform that is now
underway. Currently, discussion seems to be focussing primarily on how to ensure
adequate investment in low carbon and, in the medium term, conventional generation to
meet the UKs climate change and security of supply goals. Delivering the necessary
generation capacity is clearly crucial and by reviewing some of the mechanisms that
could be used to encourage investment, this paper attempts to contribute in this area.
However, the paper also addresses other areas where reform may be required but that
have, to date, received less attention; issues such as arrangements to ensure efficient
dispatch and energy balancing, efficient mechanisms to deal with network congestion
and measures necessary to facilitate demand side participation.
The approach taken by the paper is incremental in nature, focussing on how current
market arrangements may need to develop in the coming years, rather than proposing
radical change. It is likely that successfully decarbonising the electricity sector may
ultimately require a fundamentally different market design and that change, particularly
in relation to low-carbon investment, may be required sooner rather than later.
However, the transition to a low carbon electricity system will be gradual and arguably
best served by incremental change in response to demonstrated need.
Contents
1. SUMMARY .......................................................................................................... 3
2. A LOW CARBON ELECTRICITY SECTOR AND ITS IMPLICATIONS FOR MARKET DESIGN .. 4
3. GENERATION INVESTMENT ................................................................................... 5
3.1 THE GENERATION INVESTMENT CHALLENGE ......................................................................5
3.2 IMPACT OF WIND ON ENERGY PRICES ..............................................................................6
3.3 ENERGY-ONLY MARKETS ............................................................................................8
3.4 ENCOURAGING GENERATION INVESTMENT ......................................................................10
3.4.1 Output-based mechanisms. ........................................................................11
3.4.2 Capacity-based mechanisms ......................................................................12
3.5 A SINGLE BUYER .....................................................................................................16
4. ENERGY DISPATCH AND BALANCING IN A LOW CARBON ELECTRICITY SYSTEM ......... 17
4.1 PROBLEMS WITH EXISTING MARKET ARRANGEMENTS. .........................................................17
4.2 SYSTEM RESERVES ...................................................................................................19
4.3 MARKET LIQUIDITY .................................................................................................19
4.4 A SEPARATE MARKET FOR INTERMITTENT GENERATION? .....................................................20
4.5 THE CASE FOR A MORE INTEGRATED APPROACH TO MARKET DESIGN ......................................21
4.6 INTEGRATED MARKETS AND THE NEED FOR PRIORITY DISPATCH ............................................21
4.7 BALANCING MARKET SIGNALS WITH DEPLOYMENT RISKS FOR WIND. .......................................22
5.NETWORK CONGESTION AND APPROPRIATE NETWORK INVESTMENT SIGNALS........... 23
5.1 CONGESTION VOLUME .............................................................................................23
5.2 MINIMISING CONGESTION COSTS .................................................................................25
5.3 TRANSMISSION INVESTMENT SIGNALS ...........................................................................26
6. ENCOURAGING DEMAND-SIDE PARTICIPATION..................................................... 26
6.1 REDUCING CAPACITY AND RESERVE ..............................................................................27
6.2 SETTLEMENT IMPACTS ..............................................................................................27
7. NETWORK REGULATION .................................................................................... 29
7.1 ENSURING EFFICIENT NETWORK UTILISATION...................................................................30
7.2 DELIVERING NETWORK INVESTMENT IN A TIMELY FASHION ..................................................31
7.3 FUNDING NECESSARY NETWORK INVESTMENT ..................................................................32
8. CONCLUSIONS .................................................................................................. 33
8.1 ENCOURAGING GENERATION INVESTMENT ......................................................................34
8.2 ENERGY DISPATCH AND BALANCING .............................................................................36
8.3 DEMAND RESPONSE ................................................................................................38
8.4 NETWORK REGULATION............................................................................................39
1. Summary
The electricity system will have a pivotal role in delivering the UKs climate change
obligations and longer term aspirations. The need to accommodate large amounts of
renewable, mainly intermittent, generation by 2020, replace generation expected to
decommission in the same timescales and the need to effectively decarbonise the
electricity system by 2030 through the introduction of new low-carbon technologies,
represent huge challenges to be overcome. In addition, the need to partially electrify
the heat and transport sectors with the introduction of heat pumps and electric vehicles
will require further investment in generation capacity and could, if not adequately
managed, place additional strains on the electricity infrastructure.
If these challenges are to be met, some aspects of electricity network regulation and
market arrangements will need to change. While the current arrangements have
arguably served us well, delivering secure electricity supplies and driving out
unnecessary cost, they are designed around controllable conventional generation
capacity serving demand that varies in a predictable fashion. Tomorrows electricity
sector will, however, look very different with a more flexible demand base required to
accommodate a low carbon generation fleet that contains large amounts of high capital
cost, intermittent and inflexible capacity.
By considering some of the primary characteristics of a low carbon electricity sector, this
paper attempts to identify things that an appropriate market and regulatory regime will
need to do well. The analysis suggests that the highly disaggregated, energy-only and
illiquid nature of the current market, reinforced by asymmetrical and non cost-reflective
imbalance charges, may not be the most appropriate arrangement for dealing with
intermittent renewable generation. A return to a more integrated market design is
proposed, operating seamlessly down to real time in order to provide the liquidity and
near real time balancing opportunities necessary to accommodate intermittent
generation technologies such as wind. A more integrated electricity market would allow
reserves, energy, and potentially, network requirements, to be optimised
simultaneously.
The introduction of wind and other intermittent generation technologies will cause
energy prices to fall on average, but become far more volatile. This will make financing
both capital intensive low carbon and peaking generation more difficult and, given the
general scepticism over the ability of emissions trading to fully internalise the costs of
carbon, there would seem to be a need for additional measures to support investment.
Options include the extension of existing supplier-based obligations, Feed in Tariffs
(FiTs), capacity obligations or capacity payments. In addition, the more radical option of
creating a central entity to procure both capacity and energy has been suggested.
The need to retain substantial amounts of conventional plant to back-up intermittent
generation can be expected to significantly increase network congestion. The current
market arrangements, where network requirements are only considered one hour before
real time, arguably encourage practices that increase congestion volumes and make the
resolution of that congestion unnecessarily expensive. In addition to incurring
unnecessary costs, which are ultimately borne by electricity customers, the current
arrangements also cause network investment to appear overly attractive. The adoption
of more integrated market arrangements would allow earlier consideration of network
requirements and a more cost effective resolution of network congestion.
The development of a more responsive demand side to accommodate a partially
intermittent and inflexible generation fleet will be facilitated by the introduction of
advanced or smart metering, where domestic and small commercial customers are
metered on a half-hourly basis. This will require fundamental changes to the settlement
processes and it will be necessary to ensure that increased data retrieval, handling and
aggregation requirements do not impose unnecessary burdens on small customers or
impede the development of a more responsive demand base.
Finally, the paper briefly addresses some regulatory issues and makes the case for a
regulatory environment that more effectively supports innovation and equalizes
incentives for network investment and operational alternatives. Regulation will also
need to ensure that network investments necessary to accommodate renewable and low
carbon generation can be delivered in a timely fashion and that those investments can
be adequately financed.
2. A Low carbon electricity sector and its implications for

market design
If the UK is to achieve the target of an 80% reduction in CO2 emissions by 2050, the
electricity sector will need to be effectively decarbonised by 2030 (Committee on
Climate Change, 2008). Furthermore, as the sector is decarbonised, energy
consumption is likely to increase as low carbon electricity replaces fossil fuels in the
surface transport and heating sectors. It is difficult to be precise about the makeup of
the future generation portfolio given the wide range of possible outcomes (UKERC,
2010). However, a plausible scenario (Electricity Networks Strategy Group (ENSG), 2009)
is that around 45 GW of wind and other intermittent renewables, 10 GW of new nuclear,
together with 12 GW of supercritical coal and gas-fired plant equipped with carbon
sequestration technology will be required by 2030 in order to decarbonise the

electricity sector.
Wind, nuclear and CCS technologies all have high capital costs and, as such, may not be
plant that investors would necessarily choose to support. An appropriate electricity
market design will, therefore, need to ensure that sufficient low-carbon capacity is
brought forward. Furthermore, as much of this low-carbon capacity will be intermittent
in nature, with output difficult to predict with any accuracy until close to real time, the
electricity market will need to accommodate the increased short-term trading and
balancing activity necessary to maintain security of supply. In fact it seems likely that
these two issues, i.e. the need to ensure adequate investment in low-carbon
technologies and accommodate high levels of short-term trading and balancing activity,
will be the principle determinants in designing an electricity market for the future.
Other issues influencing market design will be the need to generally minimise
emissions, manage network congestion efficiently and accommodate a flexible and
price-sensitive demand base. Minimising emissions will require that low-carbon
generation has priority of use over conventional technologies and that overall dispatch
efficiency is maximised. The need for an effective means of managing network
congestion stems from the intermittent nature of technologies such as wind and the
consequent need for conventional generation back-up, which would make the provision
of sufficient network capacity to accommodate the simultaneous operation of all
generation capacity unnecessary and prohibitively expensive. The requirement for
market arrangements to facilitate the development of a flexible demand base is also
associated with need to minimise the impacts of intermittency, both in terms of
generation investment and energy balancing, and to allow the partial electrification of
the heat and surface transport sectors to be accomplished in a cost effective fashion.
3. Generation investment
3.1 The generation investment challenge
The UK will need to invest heavily in generation capacity over the coming years.
Deploying sufficient renewable and low-carbon generation to meet our climate change
obligations while replacing plant expected to close as a result of E U Large Combustion
Plant and Industrial Emissions Directives, is likely to require some 140 billion of
investment by 2025 (Ernst & Young, 2009). Delivering the necessary investment will be
all the more challenging given that many other countries will be embaking on similar
programmes. It is estimated that global investment in generation could run at around
$550 billion/year until 2030, with investment in Europe running at some 60
billion/year over the same period (E.on, 2009). This international dimension is
particularly relevant given that the UK will be heavily dependent on large European-
based energy companies, operating away from their home markets, to deliver the
investment in generation capacity required. The UK will, therefore, need to maintain a
regulatory and market environment that is attractive to these companies, who clearly
have choices in terms of where they invest.
The following paragraphs in this section consider how the introduction of intermittent
generation technologies such as wind make the investment challenge more difficult and
why the current GB energy only electricity market may not be the most appropriate
design to deliver the investment required to achieve our climate change goals. The
section then moves on to consider the various options available for encouraging
necessary investment, drawing on experience from the UK and overseas.
3.2 Impact of wind on energy prices

Meeting our climate change obligations implies that, by 2020, some 30GW of
predominately intermittent renewable generation will be connected to the electricity grid
supplying around 30% of our electrical energy, while around 45GW could be required by
2030. As can be deduced from figure 1, injecting such large amounts of zero-marginal
cost energy into the electricity market is likely to have a significant impact on wholesale
electricity prices, with the displacement of expensive and polluting fossil fuels and the
reduced utilisation of high variable cost, low efficiency, plant.
High wind
/Mwh
Night
Peak
Low wind
Price l (low wind

)
OCGT
CCGT & Coal

Price h (high wind)
Wind & nuclear
MWh
Figure 1. Impact of wind on energy prices
Indeed, as wind penetration increases, spot electricity prices may fall to zero and even
go negative on those occasions when windy conditions coincide with periods of low
demand and wind generators attempt to retain access to operational subsidies 1. The
negative impact of wind generation on wholesale electricity prices has been observed in
countries such as Denmark, Germany and Spain, which have installed large amounts of
wind generation as a proportion of their peak electrical demand (Poyry, 2010). In fact,
the impact of wind energy on electricity prices could be even more pronounced in GB,
due the island nature of the electricity system with little interconnection currently
available to smooth variations in supply.
While the injection of large amounts of zero-marginal cost energy will tend to reduce
average wholesale electricity prices, the intermittent nature of that energy will introduce
some additional, offsetting, costs. Intermittent generation such as wind cannot be relied
upon to be available at any particular point in time and contributes little to security of
supply. Back up resources in the form of flexible conventional generation or
alternatives such as demand response, storage or support from adjacent systems via
interconnection capacity, therefore need to be retained on almost a MW for MW basis 2 in
order to operate when wind output is low. Conventional generation, typically CCGTs,
operating in this role will experience decreasing utilisation as wind capacity builds, but
be expected to operate more flexibly, i.e. starting and stopping more frequently and
being part-loaded in order to provide both upwards and downwards reserve. These
modes of operation will introduce operational inefficiencies and associated costs, which
will need to be spread over a reducing number of running hours.
In addition to the impact of these operational costs, consumers will also need to bear
the costs of retaining conventional back up plant in service and of eventually funding
its replacement. A sustainable electricity system with high levels of intermittent
renewable generation will require far more generating capacity than there is peak
demand to be supplied and the fixed costs of this additional capacity will need to be
supported through more volatile energy prices or, alternatively, mechanisms that reward
capacity explicitly.
Renewable generation receives Renewable Obligation Certificates (ROCs) for each MW of energy
generated. These can be sold on to suppliers to help meet their obligation to purchase energy from
renewable sources, thereby creating an income stream. During periods when the combination of
renewable output and that of inflexible sources such as nuclear exceed demand, it is worth
renewable generation paying suppliers to take energy in order to retain access to the ROC income
stream. In these circumstances the spot price of energy would enter negative territory.
1
Conventional generation is generally held to a have a 95% availability forecast error over peak
demand periods. For wind to have a similar firmness, its capacity would need to be factored down
to approximately 4% of installed capacity.
2
3.3 Energy-only markets

In common with many other jurisdictions both in Europe and elsewhere, GB operates an
energy-only electricity market where non-subsidised, conventional generation relies
on the difference between energy prices and variable cost to service its investment and
other fixed costs3. Energy-only markets rely on the theory of peak load pricing
(Kleindorfer), which describes how generation investment can be optimised through
efficient pricing signals. For the majority of time, available generation capacity will
exceed demand and wholesale energy prices will reflect the variable costs of the
marginal plant. Low variable cost generation such as nuclear or wind will receive excess
income when prices are set by higher-variable cost plant such as CCGTs or coal,
contributing to their fixed costs. However, marginal and peaking generation will need
to rely on periods of high energy prices associated with tight capacity margins, which
may only occur for just a few hours per year.
To work effectively, energy-only markets require demand to be sensitive to price. Spikes
in energy prices caused by plant scarcity will be attenuated by price sensitivity and the
value that different classes of demand place on an additional MWh of supply will be
exposed. In this way, the market effectively determines how much generation capacity
is required, rather than compliance with some arbitrary generation adequacy standard.
In the absence of demand price sensitivity, as is the case in GB, the energy market may
become distorted with the GBSO having to impose voltage reductions or physical
disconnection to curtail demand in the face of generation shortages. The need to
impose operational measures to curtail demand in the absence of any natural response
to increasing price may result in some customers loosing access to supplies before
prices have reached the level where they would restrict consumption voluntarily.
Despite these general concerns, the GB energy-only market in the form of NETA and
BETTA4 has been relatively effective in bringing forward new capacity. While there was a
sharp drop in generation commissioning following the introduction of NETA in 2000,
this was probably in part a reaction to the increased plant margins that applied in the
latter years of the England & Wales Electricity Pool. The 16 GW of new, non renewable,
capacity that is forecast5 to commission by 2015/16, suggests that the current
arrangements are capable of dealing with the immediate requirements for generation
Notable examples of energy only markets are Australia (NEM), ERCPT, Nordpool & Ontario
New Electricity Trading Arrangements (NETA). Introduced to replace the E&W Electricity Pool in
April 2000. The bilateral trading arrangements introduced by NETA were extended to Scotland in
April 2005 with the introduction of the British Electricity Trading & Transmission Arrangements
(BETTA).
4
National Grid Seven Year Statement, table 3.8.
investment. However, as wind and nuclear capacity builds, conventional capacity will
experience reducing utilisation and marginal prices will increasingly be set by lower
variable cost plant. Conventional plant will therefore require ever higher energy prices
during non-windy periods in order to recover investment costs. Low carbon
technologies such as nuclear can expect to see high load factors, however they will also
be disadvantaged as average energy prices decline but become more volatile.
In a recent study to examine how the GB and All-Ireland electricity markets may might
perform as the capacity of wind and low-carbon plant grows, (Poyry, 2009) suggest that
energy price volatility can be expected to increase dramatically, with pricing peaks of
almost 8000/MWh necessary by 2030 to support the continued availability of peaking
plant. Poyry also conclude that the incidence of extremely high prices will vary
significantly from year to year due to normal variations in weather, introducing
additional uncertainties for potential investors. It is worth noting, however, that the
Poyry studies assumed demand to be insensitive to price. If demand becomes more
price sensitive through the introduction of smart metering however, the future pricing
peaks predicted by Poyry, which exceed by some margin the accepted value that
customers place on maintaining access to supply6, would be considerably reduced.
Furthermore the Poyry studies take no account of the partial electrification of the heat
and transport sectors, a requirement of achieving the UKs climate change goals, which
would inject a large amount of controllable demand and allow further demand
smoothing. Increased interconnection and storage would have a similar effect.
Notwithstanding this mitigation, the increase in zero and low marginal cost generation
will undoubtedly challenge the ability of an energy-only market to deliver adequate
levels of generation investment. There must be a limit to the extent to which price
sensitivity, particularly during periods of cold weather that often coincide with calm
conditions, can be expected to limit electrical demand. Moreover, energy price spikes
will still be necessary to adequately reward low merit and peaking plant and there is a
concern that periods of extreme, if temporary, energy prices may prove to be
unacceptable from a political or regulatory point of view.
Consumers exposed to real time electricity prices seem likely to press for prices to be
capped and, given the year to year variability suggested by Poyry, it might be difficult to
distinguish between justified price spikes and those resulting from an abuse of market
power (Poyry, 2009). There is a danger, therefore, that regulatory or political
interventions may result in measures that prevent energy prices from rising to the levels
necessary to justify investment in new capacity. Indeed, regulatory and political
pressures have resulted in the application of measures to contain wholesale prices in
Defined as the Value of Lost Load or VOLL, currently assumed to be around 4000/MWh.
many electricity markets and, while not currently applied in GB, price caps have been
applied in the past. It is also worth noting that there are other mechanisms at work
within BETTA that tend to attenuate spot prices, for example the use of contracted
reserve contracts by the GBSO as an alternative to accepting more expensive Balancing
Mechanism7 offers to adjust output in real time.
In conclusion therefore, energy only markets can claim to have the virtue of relative
simplicity, with reliability and generation investment set by market participants, rather
than arbitrary rules. However, reliance on scarcity pricing to recover fixed costs
increases investment uncertainties and finance risk. Furthermore, the ability of scarcity
pricing to stimulate adequate investment will be tested to the extreme by the
introduction of zero-marginal cost intermittent generation technologies, with the
consequent decline in average energy prices and increased price uncertainty and
volatility. While studies such as that carried out by Poyry do not, of themselves, make
the case against energy-only markets and the need to reward for generation capacity
explicitly, they do clearly demonstrate the challenges to be faced.
3.4 Encouraging generation investment

There appears to be an emerging consensus that existing market arrangements are
unlikely to deliver the low carbon investment necessary to satisfy the UKs climate
change ambitions (Ofgem, 2010), (Committee on Climate Change, 2010), (HM Treasury,
2010). Whereas none of this analysis currently goes beyond presenting options, a
common theme is the need to introduce some form of mechanism, external to the main
energy market, to encourage investment in low carbon generation capacity.
In their Energy Market Assessment (EMA), HM Treasury/DECC set out five possible
models for market reform, shown in figure 2. These models escalate in terms of
intervention from simply adding a carbon floor price to existing market arrangements,
the provision of additional low carbon incentives, regulation to limit investment in highcarbon technologies, the provision of long-term low carbon payments, and finally the
creation of a central buyer for all generation capacity and output. In discussing the
relative merits of these five options, the EMA concludes that Option A, which would
introduce a floor price for carbon while leaving other market arrangements as they are,
is unlikely to drive the pace and scale of investment required. At the other end of the
intervention spectrum, the single buyer proposed by EMA Option E is discounted as not
having sufficient benefits over less interventionist options that retain a competitive
Balancing Mechanism (BM). The BM commences at market closure, one hour before real time.
Generators (or demand) submit bids and offers to vary output (or demand) and these may be
accepted by the GBSO to ensure final energy balancing and that network congestion is resolved.
7
10
approach involving incentives, payments or restrictions in investing in conventional

technologies.
Increasing centralisation
Option A
Option B
Option C
Option D
Option E
Greater carbon
price certainty
alone
Support low
carbon in the
current market
Regulate to limit
high carbon
generation
Separate low
carbon market
Single buyer
agency
Minimum carbon
price guarantee at
currently expected
level
Additional
incentives for low
carbon generation
price above carbon
price
Regulate to drive
decarbonisation
Long term
payments to low
carbon generators
to provide revenue
certainty
Conventional plant
trades in
competitive
market framework
Single agency is
the only purchaser
of electricity
generation
existing and new
low and high
carbon- and only
seller of this on to
suppliers
Competitive
market framework
as today
Competitive
market framework
as today
Competitive
market framework
as today
Figure 2. EMA market reform options

The following paragraphs consider some of the alternatives for encouraging sufficient
low-carbon and conventional generation capacity in the context of the Energy Markets
Assessment Options B & D. The mechanisms considered include energy or output based
obligations, such as the GB Renewable Obligation (RO), capacity obligations and capacity
payments. In addition, and despite DECCs rather summary dismissal of their Option E,
the possible merits of a single-buyer concept are considered.
3.4.1 Output-based mechanisms.

Since the demise of the Non Fossil Fuel Obligation, support for the deployment of
renewable generation in the UK has been via the Renewable Obligation (RO) and, from
April this year, via a feed in tariff (FiT) for smaller generation. Both are output-based
support mechanisms that reward generators for producing renewable energy, and EMA
Option B appears to propose extending this type of arrangement to other forms of low
carbon generation. The Committee on Climate Change (CCC, 2010) also recommends
an extension of FiTs and the introduction of a low carbon obligation to ease
uncertainties over cost recovery, thereby reducing investment costs.
While output based mechanisms have been effective in bringing forward low carbon
investment worldwide and provide strong delivery incentives, they are not without
problems. Quantity based output mechanisms such as the RO, for example, suffer
from uncertainties over future ROC prices due to headroom issues (i.e. prices decline
as the specified quantity is achieved) that increase investment risk and can also over-
11
reward the cheapest low carbon technologies (BERR, 2008). Price based mechanisms,
such as FiTs, suffer from uncertainty in terms of actual response to the guaranteed price
and, if that price is incorrectly set, can result in over or under supply (Cory, 2009).
Output based mechanisms also have the potential to distort the energy market. As
indicated in 3.2, increasing wind and nuclear capacity will give rise to the possibility of
wind becoming the marginal plant when periods of low demand coincide with high wind
output. During these periods, wind generation will seek to retain access to ROC income,
driving energy prices into negative territory. Some analysis (Strbac, 2008) suggests
that, taking into account the need to carry addition reserves on part-load thermal plant,
up to 25% of wind energy may need to be rejected when wind capacity exceeds 30 GW.
Clearly, this could have a serious impact of the financial viability of wind as well as other
low carbon technologies such as nuclear, which will be dependent on high energy prices
to recover investment costs.
In order to avoid or reduce the prospect of negative prices, measures such as curtailing
ROC or FiT payments during periods of excess low carbon output could be considered.
Some possibility of low and damaging energy prices would remain however and
attention is likely to turn to the deployment of additional storage or interconnection
capacity in order to artificially boost demand. An alternative approach, albeit involving a
degree of central planning, would be to take a more strategic view of the interactions
between nuclear and wind generation and attempt to optimise the capacity mix.
The premium FiTs discussed in EMA Option B, which top up revenues from the
energy market and provide investors with a guaranteed income, have the advantage over
standard FiT designs of keeping generation involved and interested in the energy
market. Dispatchable renewable generation would be able to respond to energy price
signals and therefore be less likely to contribute to negative price problems. However,
intermittent renewable technologies such as wind are not dispatchable in any
meaningful way and are less able to respond to price signals (Poyry, Elementenergy,
2009). By supplementing energy market revenues, premium FiTs therefore could still act
in a similar fashion to the RO, particularly in the case of intermittent technologies, in
encouraging negative biding during periods of excess low carbon output.
3.4.2 Capacity-based mechanisms

An alternative approach to encouraging investment in low carbon technologies, which
would fit comfortably in EMA options B or D, would be to focus on low-carbon capacity
rather than output. There are numerous examples world-wide where obligations are
placed on suppliers to procure sufficient generation capacity to satisfy demand to some
standard of supply security. To date, the focus of such mechanisms has been security
of supply alone and they have not therefore been technology specific. However, there
seems no reason in principle why such obligations could not be broadened to deliver
12
both security of supply and investment in low carbon technologies, albeit at the cost of
some complexity in design.
3.4.2.1 Capacity obligations on suppliers
There are numerous examples of supplier-based obligations in Europe (often referred to
as Public Service Obligations), the US and elsewhere. In the US, capacity obligations are
a carryover from the old regional power pool structures in which all participating
suppliers, referred to as Load serving Entities or LSEs, were required to acquire
sufficient capacity to serve their peak demand plus a reliability margin set by the pool.
With the restructuring of the US electricity system in the mid-1990s, many of these
arrangements developed into organized capacity markets (see 3.4.2.2); however, some
of the original pooling arrangements remain8, with capacity being traded bilaterally to
meet obligations in response to demand movements or diversity in demand peaks.
Penalties usually apply in the event of an LSE failing to acquire sufficient generation
capacity to satisfy the obligation, although there is concern that there may not
necessarily be time for the Independent System Operator (ISO) 9, to access capacity in the
event of shortages. Capacity obligations can and often do allow demand-side
participation however, allowing relief in shorter timescales. Some jurisdictions, i.e.
California, have addressed this problem by imposing forward obligations to ensure
potential capacity shortages are identified in good time.
As with all capacitybased obligations that flow from an administered reliability
standard, there are concerns that the value of reliability may not be adequately balanced
against the cost of providing that reliability. Concerns have also been raised in the US
about a lack of market liquidity and that the prices paid for capacity are not always
transparent (Brattle Group, 2009).
Addressing the issue of how capacity obligations might be applied in GB, and taking a
cue from the Renewable Obligation, suppliers could be required to purchase capacity
certificates in proportion to their demand or pay a buyout price, with the proceeds
distributed to certificate holders. As the object would be to ensure both security of
supply and decarbonisation, the obligation would need to recognise the carbonintensity of different technologies, possibly through premium payments for low carbon
technologies or selecting successful bids on the basis of low-carbon emissions as well
as bid price (Gottstein, 2010).
For example, Southwest & Southwest Power Pool covering most of the Southern & South-western
states except California and Texas.
8
Independent System Operator (ISO). A not for profit entity, charged with the operation of the
electricity network and who may also administer the electricity market.
9
13
3.4.2.2 Capacity markets

Placing obligations on suppliers will naturally lead to capacity trading, as individual
parties seek to satisfy specific capacity requirements. However, where a more organised
approach is required, or where doubts exist as to the effectiveness of supplier
obligations in bringing forward sufficient generation investment, a more reliable option
may be to place an obligation on the System Operator. There are a number of examples
in Europe, the US and elsewhere of such obligations, which typically involve the System
Operator establishing a generation capacity requirement sufficiently far ahead to
encourage new investment and procuring the capacity necessary to meet that
requirement.
PJM, ISO10 New England, ISO New York and ISO Midwest are examples in the US of where
trading around the original regional pool-based LSE capacity obligations developed into
centralised capacity markets, administered by the ISO. Early market designs delivered
mixed results for a number of reasons, including an initial focus on short-term supply
reliability at the expense of signalling the need for investment in new capacity and the
distorting impact of price caps. This short term focus resulted in bipolar pricing
(Gottstein, 2010), with prices collapsing when a surplus of capacity existed but rising to
high levels in the event of capacity shortfalls.
Market designs developed to overcome these initial problems and now include forward
capacity auctions, typically three years before the year of delivery. LSEs retain their
supply obligation and can choose to contract bilaterally for capacity. However, where
participation in the market is mandated, this capacity must be input to the auction with
the ISO effectively procuring any residual capacity required. Both existing and new
capacity, and in the case of PJM & ISO New York demand response, bid into the auction
and the clearing price is paid to all successful bidders. The costs of procuring the
required capacity are allocated to LSEs on a pro-rata basis. There is also a locational
element to the auctions to ensure that transmission constraints are respected.
Capacity markets operating in the US and elsewhere are essentially non-technology
specific in nature, focussing on security of supply alone. However, as is the case with
simple capacity obligations, there seems to be no reason why capacity markets could
not be designed to take into account the carbon-intensity of generation in order to
deliver both security of supply and low carbon objectives
It can be argued that the System Operator or ISO is better placed than individual
suppliers to anticipate future system demand and optimize the shape of the generation
10
Independent System Operator (ISO).
14
portfolio. Individual suppliers may, for example, be prepared to shed market share
rather than commit to new capacity in an uncertain world and might favour one
technology over another. An independent System Operator may well take a more holistic
view but would be sourcing capacity on the basis of generation adequacy rather than
market signals. In the context of the current GB market arrangements, it is difficult to
see how investment to satisfy some non-market based adequacy requirement could coexist with investment on a commercial basis. There is a danger therefore, that placing
an obligation on the System Operator to procure capacity, even as provider of last
resort, could deter normal commercial investment.
3.4.2.3 Capacity payments
A limitation of some early capacity-based obligations and markets was a lack of
incentives to ensure real time plant availability. Although financial penalties for nondelivery are now common, these penalties do not always reflect the real value of
availability during times of system stress. Capacity payment designs are considered
more effective in providing real time availability incentives (Oren, 2000).
Capacity payment mechanisms normally involve a payment being made to every
generator that is available to meet demand for each trading period irrespective of
whether or not the generator is actually required to run. Payments, which are normally
funded via an uplift on suppliers, are usually linked to the value of capacity, i.e. when
the supply position is tight payments will be higher. Capacity payment mechanisms are
invariably associated with more integrated scheduling and dispatch arrangements such
as the England & Wales Pool, which operated from privatisation of the GB electricity
sector in 1990 to the introduction of NETA in 2001.
In the case of the England & Wales Electricity Pool, capacity payments were a function of
the loss of load probability (LOLP)11 and the value of lost load (VOLL)12. If the supply
situation became very tight and LOLP approached unity, then capacity payments could
reach very high levels, capped only by the value of VOLL (currently assumed to be
around 4000/MWh). The all-Ireland Single Electricity Market introduced in 2007 also
incorporates a capacity payment mechanism but utilising a softer link to the fluctuating
value of capacity based on the cost of building efficient open cycle gas fired plant.
LOLP is a measure of the likelihood of insufficient generation capacity being available to meet peak
demand, varying from zero when there is no risk to unity when there is certainty.
11
VOLL is an estimate of the maximum price a customer is prepared to pay to maintain access to
electricity supplies. In practice, VOLL will vary between customer groups and on other
circumstances.
12
15
While linking payments to the scarcity value of capacity incentivises additional provision
during periods of system stress, perversely it also provides an incentive on portfolio
generators to withdraw capacity in order to increase those payments. To some extent
this is a criticism that can be levelled at all capacity mechanisms, however the more
granular nature of capacity payments compared with capacity obligations provides
rather more scope for abuse. Problems associated with the withholding of capacity were
a principal driver behind the abandonment of the E&W Electricity Pool and the
development of a bilateral market with incentives to contract forward (Patrick, 2001).
Other criticisms of capacity payments relate the value of VOLL, which is estimated rather
than set by the market, and the usually simplistic methods used for calculating of the
value of LOLP. In combination, these issues are almost certain to result in a mismatch
between the value of capacity payments and that which would arise from a capacity
market.
As is the case with capacity obligations and markets, the application of capacity
payment mechanisms to date has been linked exclusively to maintaining security of
supply. There seems to be no reason however why a capacity payment mechanism
could not be designed to recognise the carbon intensity of generation in order to
address both decarbonisation and security of supply objectives.
3.4.3 A single buyer

Although dismissed rather abruptly by the EMA as unnecessarily interventionist and
lacking the disciplines to drive efficiency, the single buyer model would seem to have
some relevance to a low carbon world. A central agency would identify the need for and
procure low-carbon, and possibly conventional, generation capacity via a tender
process. Successful bids up to a defined capacity requirement would be awarded a
fixed annual income over the lifetime of the project on a /MW basis, or to reflect
levelised costs. If capacity was to be rewarded on a /MW basis, the arrangement would
look rather similar to the centralised electricity markets that have developed in the US
and elsewhere, and which are discussed in 3.4.2. Energy would be sold into and bought
through a spot market, with parallel contracting between generators and suppliers via
contracts for differences (cfds), where price certainty was required or to meet any lowcarbon obligations that might be imposed. Dispatch priority may need to be given to
low-carbon generators in order to avoid any risk of a price-based dispatch process
preventing compliance with those obligations.
If, however, generation capacity was procured on the basis of, say, levelised costs, the
single buyer model could transform the nature of the electricity market. As revenues
would be agreed during the tender process, indexed to cover fuel price variation in the
case of nuclear or carbon sequestered plant, there would be little point attempting to
dispatch plant on the basis of submitted bids via a spot market. Generation could be
dispatched to meet demand on the basis of a carbon-emissions hierarchy and to resolve
16
network congestion, with differentiation within plant technologies on the basis of

marginal cost. The need for a plethora of confusing renewable and low carbon
obligations would be removed.
The value of a single buyer approach would be in substantially improving the investment
climate. Continuing with the existing market arrangements implies increasing energy
price uncertainty and volatility, increasing the cost of project capital. However, the
guaranteed income stream associated a single buyer model would improve investor
confidence and reduce the cost of capital. Low-carbon generation projects, which have
high capital costs, would particularly benefit from a more benign investment climate and
the overall costs of decarbonising the electricity sector would be reduced.
The EMAs concern that a single buyer model would lack incentives to drive efficiency
seems to some extent misplaced. In addition to the advantages flowing from an
improved investment climate, a single buyer approach would bear down on project costs
through the tender process. While fuel price risk and risk of generation assets become
stranded would ultimately lie with the customer, construction and operational risks
would remain with the generator. Overall, the single buyer model would seem to
provide a competitive and less complex environment for the delivery and operation of a
low carbon electricity sector.
4. Energy dispatch and balancing in a low carbon electricity

system
4.1 Problems with existing market arrangements.
Currently, the GB market arrangements make no organised attempt to optimise
generation dispatch. Generators and suppliers trade energy in advance on a bilateral
basis and, at gate closure13, present the GBSO with generation or demand schedules
necessary to deliver contractual commitments a process referred to as self dispatch.
Furthermore, as the majority of energy is produced by vertically integrated utilities with
both generation and supply businesses, much of this trading is internal - i.e. these
utilities self supply to a significant extent. Self supply limits the competitive pressures
on which bilateral markets depend to ensure efficiency (Sioshansi, 2009) and the
combination of self dispatch and self supply creates the potential for non-optimised
dispatch outcomes.
13Gate
closure 1 hour before real time, the energy markets close and the Balancing Mechanism
commences. Contractual positions at gate closure are compared with outturn in order to determine
imbalance.
17
An additional issue to be considered in the context of dispatch efficiency is the tendency

for generators to self insure. The asymmetrical and non cost-reflective cash out
prices14 applied to residual imbalances resolved by the GBSO in the Balancing
Mechanism, which operates from gate closure, encourages generators to carry reserves
in order to minimise imbalance. These reserves are additional to those specified by the
GBSO to cover demand or generation uncertainties and result in an energy market that is
predominately long, with connected generation capacity exceeding the demand to be
supplied. Consequently, more generation is part-loaded than is actually required,
reducing overall efficiency and causing unnecessary emissions. An indication of the
extent to which generation companies self insure is given by the monthly Trading
Operation Report15 published by Elexon, which suggests that there is typically between
1000 and 2000MW of unused reserve available on part-loaded plant over demand peaks
and considerably more during other periods.
The extent to which the combination of self dispatch, self supply and self insurance
reduces the overall efficiency of dispatch is unclear. However, there is evidence from
both the US (Sioshansi, 2009) and GB (ILEX, March 2002) to suggest that fuel inputs
could be around 3 or 4% higher than would be the case if an organised attempt were
made to fully optimise generation dispatch. Further anecdotal evidence that the current
GB market arrangements may produce generation dispatch outcomes that differ from
the optimum is given by analysis undertaken for Elexon in developing a mechanism to
account for transmission losses (Siemens, 2009). This analysis demonstrated that, in
some cases, transmission line loss factors calculated from actual line flows differed from
those produced using a load flow model that dispatched plant on the basis of marginal
cost. In other words, the disposition of generation resulting from existing market
arrangements appears to differ to some extent from that which might be delivered by a
truly optimised dispatch process based on actual marginal costs.
Although the inefficiencies introduced by self dispatch might currently be of a low order,
they do result in unnecessary cost and carbon emissions. Dispatch inefficiency is also
likely to increase with the growth of intermittent generation. With relatively little wind
capacity connected, portfolio generating companies can hide intermittency within their
14Imbalances
that add to the net system imbalance are treated differently than those that reduce net
imbalance. For example, a generator whose imbalance adds to system imbalance is exposed to the
balancing costs incurred by the GBSO. Generators whose imbalances reduce net imbalance
pay/receive prices related to the short term energy prices. This asymmetry encourages parties to self
balance and penalizes inflexible or intermittent generators.
Operational Trading Reports are available at
http://www.elexon.co.uk/search/default.aspx?qs=operational%report
15
18
settlement production account16 relatively easily. However, as wind capacity builds,

internalising the impact of intermittency within a generation portfolio will become
more difficult. Generators will attempt to trade out intermittency close to real time as
wind forecasts become more accurate, however to limit imbalance risk and exposure to
cash out prices, generators will need to carry more reserve as the intermittent capacity
within their portfolio increases.
4.2 System reserves

In addition to any reserve held by individual portfolio generators as insurance against
exposure to imbalance charges, the growth in intermittent capacity will also require the
GBSO to procure additional reserves. Currently, system reserve levels are relatively
modest at around 4GW (4 hours ahead of real time) and predicable, varying only slightly
with demand level and time of day. However, as intermittent wind capacity builds,
reserve levels will increase and are predicted to exceed 9GW by the middle of the next
decade (National Grid, 2009). The requirement to carry reserves will also become
considerably more unpredictable and volatile, increasing when high wind output is
forecast and decreasing during periods of relative calm.
Currently, the GBSO procures reserve through a combination of periodic tenders 17, some
intra-day power exchange trading together and Balancing Mechanism bid/offer
acceptances close to real time. While these arrangements are effective in procuring
sufficient reserves to meet current requirements, they are unlikely to produce an
optimised outcome or reveal the true real time value of reserve. If the GB market is to
deal effectively with an increased and more volatile requirement for system reserves in
the future, some means of more formally integrating energy and reserve requirements in
the short term and intraday markets will be required.
4.3 Market liquidity

As suggested in 4.1, the growth in wind capacity will significantly increase short term
trading as generators attempt to match commitments to updated and more accurate
wind output forecasts. This increased trading close to real time will require efficient,
liquid short term markets and there are factors which suggest that the current market
structures may not be best placed to provide that liquidity or deal with the challenges
associated with large amounts of wind generation. The GB market is the least liquid of
all comparable European markets (Weber, 2009) due primarily to the vertically
For the purposes of settlement, a generating company has a single production account. In the case
of a portfolio generator, imbalance prices are applied to the aggregated production account
imbalance, rather than the imbalance of each individual generator in that account.
16
The GBSO contracts for Short Term Operational Reserve (STORR) via auctions held three times a
year.
17
19
integrated nature of the sector and the high level of internal trading. Liquidity is also
reduced by the continuous nature of trading and the existence of alternative trading
platforms, which tend to disperse trading activity.
This lack of short term market liquidity acts against the interests of both intermittent
generation such as wind and also small independent players, who have a greater need
for balancing in the shorter term. Prompted by these and more general concerns about
market efficiency, Ofgem has consulted on measures to improve market liquidity and
has threatened action by the end of 2010 if the situation has not sufficiently improved
(Ofgem, 2010).
4.4 A separate market for intermittent generation?

In the context of the existing, disaggregated, bilateral trading arrangements, there may
be some value in creating a separate market for wind and other intermittent
technologies. As suggested in 4.1, portfolio generators will find it increasingly difficult
to internalise the impacts of intermittency as capacity increases and will need to resort
to short term trading. However, internalising or trading out intermittency on an
individual generator or portfolio basis is unlikely to take full advantage of the
geographic diversity of wind output, which can significantly reduce wind output
uncertainty18.
As wind capacity grows, there may be a case for creating separate market arrangements
for wind in order to capture the value of geographic diversity. Wind output could be
aggregated across the whole of GB and auctioned into the electricity market, reducing
forecast error and overall imbalance. Charging for imbalance on aggregated basis
rather than against individual or portfolio generator output would arguably be more
cost-reflective, as balancing costs incurred by the GBSO reflect net generation-demand
imbalance rather than the imbalance of any particular generator.
Carving out a separate market for wind would be a radical departure from current
practice and might, therefore, encounter opposition from portfolio generators.
However, the increasing difficulty and inefficiency associated with attempting to manage
intermittency on an individual company basis may cause support for a separate market
for wind to grow with time. A separate market for wind would be particularly helpful for
independent wind operators who, unlike portfolio generators, currently have little
opportunity to mitigate the impact of intermittency and reduce imbalance charges.
18
Aggregating wind output over a wide geographic area significantly reduces wind output forecast error
together with associated reserve and capacity requirements. See for example
www.nationalgrid.com/.../GBSQSSIntegratedReliabilityAndEconomicsAssessment.pdf
20
4.5 The case for a more integrated approach to market design

Whereas creating a separate market for wind might be appropriate given a continuation
of bilateral energy trading, issues of dispatch efficiency, market liquidity and the need to
deal with increasing and more volatile reserve requirements suggest that the existing
arrangements may not be appropriate for a low-carbon electricity system. The need for
market participants to adjust their contractual positions in response to more accurate
short term forecasts needs to be recognised and facilitated, while increased and volatile
reserve requirements need to be coordinated more effectively with energy procurement
in order to ensure efficient dispatch. The more integrated electricity market designs, as
adopted by PJM, New England, New York and, to a lesser extent Spain, seem more
effective in dealing with these issues and therefore more appropriate in terms of
transitioning to a low carbon electricity sector.
In the integrated US markets, energy is traded via day ahead and near real time auctions
run by the ISO, based on bids submitted by generators. The timed nature of the auctions
maximise liquidity, contrasting with the situation in GB where liquidity is reduced by the
disaggregated and continuous nature of trading. The simultaneous procurement of
energy and reserve requirements based on production costs ensures that generation
dispatch is optimised and the real time value of plant flexibility is revealed.
With around 17GW of wind capacity currently installed, the Spanish electricity market
has evolved to deal with the impacts of intermittency. Market arrangements lie
somewhere between the fully integrated designs seen in the US and the disaggregated
approach adopted by GB. The majority of energy is traded via timed day-ahead and
intra-day auctions in a similar fashion to PJM and other US markets, ensuring high levels
of liquidity. Unlike the US however, the auctions are administered by a Market Operator.
The System Operator inputs reserve requirements to the multi intra-day auction
process, ensuring that energy and reserve requirements are optimised simultaneously
and an efficient generation dispatch outcome is achieved.
4.6 Integrated markets and the need for priority dispatch

If the costs of carbon emissions are fully internalised, an integrated dispatch process
that attempts to minimise the overall cost of meeting demand securely should also
minimise carbon emissions. However, if the cost of carbon remains low, this will not
necessarily be the case. While intermittent wind and nuclear generation plant have zero
or low marginal costs and will always be dispatched before carbon emitting generation,
renewable technologies such as biomass 19 have non-trivial marginal costs and carbon
EU Directive 2009/28/EC requires that member states introduce regulations to ensure that
renewable generation is given priority in dispatch over other forms of generation. The UK has not
introduced regulations to give effect to priority dispatch as, in the GB electricity market, all
generation can self dispatch and therefore achieve priority unilaterally.
19
21
sequestrated generation is likely to have higher high marginal costs than nonsequestrated plant due to the associated efficiency penalty.
There is a possibility therefore, that an integrated dispatch process may not minimise
carbon emissions if carbon is incorrectly priced. In the transition to a low-carbon
electricity system, the introduction of a more integrated dispatch process would need to
be accompanied by some means of prioritising low carbon generation. Rather than
dispatching generation on the basis of cost, low carbon generation would need to be
dispatched on the basis of carbon emissions, or some function of marginal cost and
carbon emissions, to ensure that overall emissions were minimised. While low carbon
capacity remained at modest levels, there would be little need to differentiate between
individual generators or technology for the purpose of dispatch. However, as capacity
increased, network or energy-related constraints would become more frequent, and
some means of differentiation would be required to ensure that carbon emissions were
minimised at the lowest possible cost. Differentiation between technologies could be
achieved on the basis of emissions, while differentiation within technologies could be
achieved when necessary on the basis of marginal cost or some other measure, such as
transmission losses.
It is interesting to note that there is some experience of non-marginal cost related
generation dispatch in the UK, albeit in a rather different context. The Central Electricity
Generating Board (CEGB ), which operated a highly detailed centralised dispatch
optimisation process, was able to move seamlessly from dispatching on the basis of
marginal cost to a heat rate based dispatch during the frequent fuel emergencies of
the 19770s & 1980s. Dispatching fossil fired generation on the basis of heat rate
rather than marginal cost resulted in a significant reduction in fuel inputs and, as a
consequence, would have reduced carbon emissions/MWh of electrical energy
generated.
4.7 Market signals v deployment risks for wind.

While a more integrated market design, coupled with priority in dispatch, would create a
more benign environment for intermittent technologies such as wind, additional
measures may be required given the characteristics of wind generation and the scale of
deployment required. The increasingly volatility of energy prices, with wind always
likely to be on the wrong side of the balancing argument - attempting to sell energy
when wind output is high and energy prices low (or even negative) - will decrease the
value that wind can extract from the energy market (Redpoint, 2009) over time. To this
erosion of value can be added system integration costs that will also rise steadily as
wind capacity builds, further impacting on the viability of future wind projects.
22
While these market signals may simply reflect the economic consequences of
intermittency, they could have a negative impact on deployment or at least on the need
for subsidy to maintain the level of deployment required to deliver climate change goals.
The extent to which wind generation is exposed to market signals varies across Europe.
For example GB chooses to draw no distinction on the basis of generation technology,
exposing the full costs of balancing and imposing technical requirements that require
wind to behave as any other generation even though a system approach may result in
lower overall cost. Germany, on the other hand, protects wind generation from the full
rigour of market and technical signals with the costs of integration falling mainly on the
System Operator. A question to be addressed by the UK and indeed all jurisdictions that
intend to connect large amounts of wind or other intermittent renewable generation is,
therefore, how to balance the exposure of that generation to market signals with the
risks to deployment inherent in those signals particularly given the limited ability of
wind to respond.
5. Network congestion and the need for appropriate network

investment signals.
Commissioning large amounts of wind or other intermittent generation together with
the associated need to retain back-up generation will result in far more generation
capacity being connected to the electricity grid than there is peak demand to be
supplied20. This will result in a significant rise in potential network congestion 21, indeed
that process has already begun and the GBSO is forecasting that the cost of resolving
network congestion will approach 600 million by 2011 with the prospect of substantial
rises after that time (Redpoint, 2010). A future electricity market will therefore need to
be capable of dealing with congestion in a cost-effective fashion. Unfortunately, the
current GB market arrangements are not particularly effective in controlling the volume
of congestion or minimising the costs of resolving that congestion.
5.1 Congestion volume

In terms of controlling the magnitude of network congestion, the current market
arrangements are deficient in two respects. Firstly, market participants can trade energy
bilaterally in forward markets without the need to consider the costs that those trades
will impose on the electricity grid. The implications of this unconstrained trading are
The margin of generation capacity over demand is expected to rise from historic levels of around
24% to nearer 90% by 2020
20
Congestion arises when potential power flows exceed the capability of network circuits or
boundaries. Congestion is resolved either by adjusting generation patterns to reduce power flows, or
by increasing network capacity either by investment in primary assets or by operational means.
21
23
presented to the GBSO in the form of individual generator dispatch schedules one hour
before real time, at which point the GBSO is required to establish counter-flows via the
Balancing Mechanism (BM) to ensure that actual power flows do not exceed network
capacity. Secondly, the cost of resolving this congestion is recovered via Balancing Use
of System (BUSoS) charges paid by all trading parties on a per kWh basis and there is
therefore no incentive on parties causing that congestion to modify their behaviour.
In fact, it can be argued that market arrangements currently encourage behaviour that
leads increased network congestion. The separation of energy trading and congestion
management into distinct markets prompts generators to consider how best they can
maximise returns. Consider for example a portfolio generating company with a large
installed wind capacity in Scotland and conventional generation assets on both sides of
the cheviot22 network boundary. The company would contract ahead to supply energy
assuming, due to its intermittent nature, a modest contribution from wind. If,
approaching gate closure, it appeared that wind output would be high, the generator
would need to decide what conventional generation to stand down plant on the export
side of the boundary or plant on the import side. If the company stands plant down on
the export side, potential congestion across the boundary is eased but that plant earns
no income. If however, the company stands down conventional plant in E&W, then
congestion across the boundary is increased and the GBSO is likely to accept bids from
Scottish conventional plant to reduce output. As these bids are invariably less than
variable cost of generation, the Scottish plant earns income by not producing or by
producing less. Furthermore, the conventional plant in E&W that was stood down is now
free to offer replacement energy at a significant premium to market prices.
Market rules therefore allow, in fact encourage, companies to maximise income by
acting in a fashion which is detrimental to the efficient operation of the system (LECG
Consulting, 2010). Ofgem appears to consider that such behaviour amounts to an abuse
of market power and a Market Power License Condition (MPLC)
23was
passed into law by
the 2010 Energy Act. The new Condition gives Ofgem the power to penalise the
withholding or manipulation of output however, given the short term market volatility
that the deployment of wind at scale will bring, deliberate manipulation of output to
exploit network constraints will become more difficult to demonstrate. Furthermore,
even if demonstrated, such behaviour is arguably no more than the expected
commercial response to a particular set of flawed market arrangements. The MPLC
therefore addresses the symptoms of the problem, rather than the problem itself.
The Cheviot boundary is that which cuts the four transmission circuits connecting Scotland with
England, currently having a capacity of around 2.4 GVA.
22
23
http://www.opsi.gov.uk/acts/acts2010/pdf/ukpga_20100027_en.pdf
24
5.2 Minimising congestion costs

Not only are existing GB market arrangements ineffective in managing congestion
volume, they make dealing with congestion particularly expensive. This stems from the
energy only nature of the electricity market and the need for mid-merit plant to recoup
a proportion of their fixed costs by extracting a discount or premium on forward market
prices through Balancing Mechanism bids and offers. The need for mid merit plant to
attempt to recover fixed costs through the BM in this fashion is reinforced by the fact
that peaking plant is able to partially recover investment and other fixed costs through
pre-gate closure energy contracts. By utilising contracted plant over demand peaks, the
GBSO is able to reduce spikes in energy prices, therefore reducing the income available
to non-contracted mid merit plant (SEDG, 2009).
The impact of fixed cost recovery through the BM can be seen in figure 3, which
illustrates the relationship between accepted offers and bids to the market index price
(MIP)24. It can be seen that accepted BM offers are invariable at a significant premium to
MIP, while accepted bids are invariably discounted. As the cost of resolving congestion
is the difference between the associated bids and offers, these costs can on occasion
exceed 150/MWh.
250
200
/MWh
150
100
50
0
1
10
Month (08/09)
Of f ers
Bids
Energy price
Figure 3. Balancing Mechanism bids & offers compared with market Index price,
2008/09 (National Grid)
24
Market Index price (MIP) is indicative of intra-day market energy prices.
25
It is instructive to compare the costs of resolving congestion under current market rules
with those observed under previous market regimes, for example the England & Wales
Electricity Pool, which precede the introduction of NETA/BETTA. Under the old Pool
rules, the cost of resolving congestion was essentially the difference between the offers
made by generation that was ultimately constrained and offers made by replacement
plant at the day-ahead schedule stage. For similar technologies, i.e. where coal plant
was displaced by other coal plant in order to resolve a network constraint, the difference
in day ahead offers may only have been a few pounds and maybe around 15/MWh
where CCGT plant was replaced by coal. Resolving network congestion under BETTA is
therefore around ten times as expensive as was the case under the E&W Electricity Pool
rules or the old CEGB merit order arrangements which existed before the industry was
privatised.
5.3 Transmission Investment signals

Higher than necessary costs of resolving congestion make investment in network assets
to remove that congestion appears overly attractive. Transmission reinforcement can be
justified up to the point where the marginal cost of reinforcement equals the marginal
reduction in congestion costs brought about by that reinforcement. Clearly, if the costs
of resolving congestion are at least 10 times higher than necessary, much more
transmission reinforcement can be justified than is actually required. This together with
network design rules that tend to provide sufficient network capacity to allow the
simultaneous contribution of all generation to system peak demands ( inappropriate as
there will be far more generation connected to the network than there is demand to
supply), suggests that rather more network capacity is likely to be built than is actually
required. While there is considerable uncertainty around just what network investment
may be required to deliver a decarbonised electricity system and the consequences of
having too little network capacity are likely to outweigh those of too having much, overinvesting due to inappropriate market signals or design rules would impose unnecessary
costs on customers and could ultimately undermine the case for connecting renewable
generation.
6. Encouraging demand-side participation

Mature electricity systems around the world can be described as having a flexible
generation portfolio able to respond to a variable, relatively price-insensitive but
predictable demand base. However, with the introduction of large amounts of
intermittent renewable generation, this model is likely to be reversed, with the demand
side needing to become more flexible in order to accommodate a variable supply.
26
6.1 Reducing capacity and reserve

Effective demand side participation can facilitate the development of a low-carbon
electricity system in both investment and operational timescales. By competing with
generation in capacity auctions, the overall requirement for generation capacity will be
reduced. Similarly, the impact of partially electrifying the heat and surface transport
sectors on network investment could be minimised by effectively managing that demand
and utilising its inherent storage capacity. In operational timescales, demand response
has the potential to reduce the requirement for reserve to be held on part-loaded
generation, while generally reducing the impact of intermittency and energy price
volatility which could otherwise reach unacceptable levels.
Currently, demand response in GB is limited to relatively large industrial demand,
usually contracting with the GBSO ex-anti to supply load reduction when required, for
example in the event of a low unexpected generation loss. To date, response from the
domestic or small commercial demand has mostly been limited to shifting demand
from peak to off-peak periods through fixed time of use (ToU) tariffs such as
Economy 7, although more flexible demand shifting via tele-switching has been
utilised to some extent.
The introduction of smart metering, incorporating communication capabilities and the
availability of smart appliances that can respond to price or other signals, will make
domestic and small commercial customers more aware of their consumption and
become more active providers of demand response. This could be achieved through
suppliers offering interruptible tariffs, with domestic appliances or heating being
switched automatically and allowing suppliers to offer aggregated demand response in
both investment and operational timescales. Alternatively, dynamic ToU tariffs could be
offered, with pricing being set to reflect short term wholesale market conditions and
consumers responding to price signals either manually or, more conveniently, by relying
on smart appliances.
6.2 Settlement impacts

The delivery of domestic and small commercial sector demand response will have
implications for the electricity market settlement process. As demand less than 100kW
is currently metered on a summation basis, it is input to the electricity market via a
profiling process, where customers are allocated to one of eight demand profiles for
the purposes of settlement. Rather than being charged for the actual half-hourly
consumption of their smaller customers, suppliers are charged on the deemed
consumption given by these profiles.
27
The settlement process has the innate ability to allocate actual energy consumption to
the appropriate settlement periods and profiling could probably be extended to
accommodate interruptible demand and ToU tariffs, provided that the shape and timing
of those tariffs were known in advance. Individual profiles could be constructed around
each ToU tariff, once customer response to those tariffs had been demonstrated by
experience. However, while fixed ToU tariffs are an appropriate response to predicable
demand characteristics where the timing of demand and price peaks can readily be
forecast, they will be less so with the growth of intermittent generation. Dynamic ToU
tariffs will be required to respond to variations in energy supply energy prices that can
only be forecast with any accuracy in short timescales.
Truly dynamic ToU tariffs will therefore require energy consumption to be settled on a
half-hourly, rather than a profiled basis. Profiling will clearly need to be retained
though the smart meter rollout process, however it seems likely that there will be a
gradual migration of non-half hourly metered demand to half-hourly settlement over
time. This will have implications for the settlement process. Firstly, profiling would
need to ensure the appropriate half-hourly allocation of energy consumption as the
number of customers being profiled diminished and, secondly, that differences in actual
and estimated consumption continued to be dealt with appropriately. Differences
between actual and estimated energy consumption are allocated to suppliers on the
basis of their market share of non-half hourly metered demand and this may no longer
be appropriate with ever diminishing customer numbers (Elexon, 2008).
A transition to full half hourly settlement will involve a very substantial increase in the
volume of metering data to be processed and the costs of data retrieval, handling and
aggregation will clearly increase. While the central settlement systems may not be
significantly affected as data will be received in an aggregated form, the need to process
half hourly, rather than summated, customer energy consumption will substantially
increase the data volumes to be handled by suppliers. Although modifying existing half
hourly settlement processes, e.g. by extending the period over which half hourly data
may be entered into the settlement system (Frontier Economics, 2007), could mitigate
cost increases to some extent, substantial increases in cost seem unavoidable. An
indication of scale of these additional costs can be inferred from those incurred by
customers who currently elect to be metered on a half-hourly basis. In 2007, the
additional costs associated with data aggregation and collection was estimated at
around 250 (Elexon, 2010). Although there is evidence to suggest that these costs
have reduced, it will be necessary to ensure that overheads are commensurate with the
relatively low energy requirements of individual and do not become a barrier to smart
metering delivering small customer demand response.
A further settlement-related issue is the extent to which current arrangements will
encourage dynamic customer demand response, i.e. response to real time situations.
28
References
Abbad, J. R. (2010). Electricity Market Participation of wind Farms:the success story of
the Spannish pragmitism. Energy Policy.

Baker, Mitchell & Woodman. (2009). The Extent to Which Economic Regulation Enables
the Transition to a Sustainable Electricity System.

Brattle Group. (2009). A Comparison of PJM's RPM with Alternative Energy and Capacity
Market Designs.
Committee on Climate Change. (2010). Meeting Carbon Budgets - ensuring a low
carbon recovery.
E.on. (2009). http://www.eon.com/en/downloads/E.ON_Capital_Market_Day_09.pdf.
ENA. (2010). ENA Response to Ofgems Embedding Financeability in a New Regulation"
Consultation.
Ernst & Young. (2009). Securing the UK's Energy Future.
Geen Investment Bank Commission. (June, 2010). Unlocking Investment to Deliver
Britain's low carbon future.

Gottstein, S. (2010). the Rol of Forward Capacity Markets in Increasing Demand-Side and
Other Low Carbon Resources: Experiences and Prospects.

HM Treasury. (2010). Energy Market Assessment.
House of Commons Business and Enterprise Select Committee. (2009). Energy Policy -
Future challenges.
IEA. (2008). Design of Power Systems with Large Amounts of Wind Power. Wind Task 25
Phase 1 Final Report.

ILEX. (March 2002). NETA- the Next Phase.
Investco Perpetual. (2010). Open Letter to Lord Mogg from Neil Woodford.
Kleindorfer, C. &. The Economics of ublic tility egulation. 1986: Macmillan Press.
LECG Consulting. (2010). The UK Transmission Congestion Problem.
41
Littlechild. (2007). A Proposal for a Balancing Market to Determine Cash Out Prices.
Natioanl Grid. (2009). Grid Code Working Group - Intermittent Generation Data.
National Grid. (2010). Letter from Paul Whittaker to Scott Phillips, Ofgem; RPI-X@20
Current Thinking Working Paper - Financeability.

National Grid. (2009). Operating the Electricity Transmission Networks in 2020.
Ofgem. (2010). Liquidity Proposals for the GB wholesale Electricity Market. Ref 22/10.
Ofgem. (2010). Project Discovery.
Ofgem. (2009). Project Discovery: Energy Market Scenarios consultation, apendix 2.
Ofgem. (2010). Regulation energy networks for the future: RPI-X@20 Recomendations:
Ref 91/10.
Ofgem. (2008). Transmission Access Review - Initial Consultation on Enhanced
Investment Incentives. Document 175/08.

Poyry. (2009). Impact of Intermittency - How Wind Variability could Change the Nature
of the British & Irish Electricity Markets.

Poyry. (2010). Wind Energy and Electricity Prices - Exploring the merit order effect.
Areport to the EWEA.

Redpoint. (2009). Decarbonising the GB Power Sector, evaluating investment pathways,
generation patterns and emissions through to 2030. A report to the Committee on

Climate Change.
Redpoint Energy. (2006). Dynamics of GB Electricity Generation Investment.
Redpoint. (2010). Improving Grid Access: Modeling the Impacts of the consultation
Options.
SEDG. (2009). Probabalistic Network Operation and Design standards to Support the
Development of the UK Low carbon Electricity System.

Siemens. (2009). MP229 Load Flow Modelling Service. Report to Elexon.
Sioshansi, O. O. (2009). The cost of Anarchy in Self-commitment Based Electricity
Markets.
42
Strbac. (2008). Transmission Systems with Wind & Nuclear.

Weber. (2009). Adequate Intraday market desig to Allow the integration of wind energy
into the European power systems. Elssvier.
43
Chapter 3
Direct Solar Energy

Coordinating Lead Authors:
Dan Arvizu (USA) and Palani Balaya (Singapore/India)
Lead Authors:
Luisa F. Cabeza (Spain), K.G. Terry Hollands (Canada), Arnulf Jger-Waldau (Italy/Germany),
Michio Kondo (Japan), Charles Konseibo (Burkina Faso), Valentin Meleshko (Russia),
Wesley Stein (Australia), Yutaka Tamaura (Japan), Honghua Xu (China),
Roberto Zilles (Brazil)
Contributing Authors:
Armin Aberle (Singapore/Germany), Andreas Athienitis (Canada), Shannon Cowlin (USA),
Don Gwinner (USA), Garvin Heath (USA), Thomas Huld (Italy/Denmark), Ted James (USA),
Lawrence Kazmerski (USA), Margaret Mann (USA), Koji Matsubara (Japan),
Anton Meier (Switzerland), Arun Mujumdar (Singapore), Takashi Oozeki (Japan),
Oumar Sanogo (Burkina Faso), Matheos Santamouris (Greece), Michael Sterner (Germany),
Paul Weyers (Netherlands)
Review Editors:
Eduardo Calvo (Peru) and Jrgen Schmid (Germany)
This chapter should be cited as:

Arvizu, D., P. Balaya, L. Cabeza, T. Hollands, A. Jger-Waldau, M. Kondo, C. Konseibo, V. Meleshko,
W. Stein, Y. Tamaura, H. Xu, R. Zilles, 2011: Direct Solar Energy. 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 University Press, Cambridge, United Kingdom and New York, NY, USA.
333
Chapter 3
the primary source of information, but their accuracy is inherently lower

than that of a well-maintained and calibrated ground measurement.
Therefore, satellite radiation products require validation with accurate
ground-based measurements (e.g., the Baseline Surface Radiation
Network). Presently, the solar irradiance at the Earths surface is estimated with an accuracy of about 15 W/m2 on a regional scale (ISCCP
Data Products, 2006). The Satellite Application Facility on Climate
Monitoring project, under the leadership of the German Meteorological
Service and in partnership with the Finnish, Belgian, Dutch, Swedish and
Swiss National Meteorological Services, has developed methodologies
for irradiance data from satellite measurements.
Various international and national institutions provide information
on the solar resource, including the World Radiation Data Centre
(Russia), the National Renewable Energy Laboratory (USA), the National
Aeronautics and Space Administration (NASA, USA), the Brasilian
Spatial Institute (Brazil), the German Aerospace Center (Germany), the
Bureau of Meteorology Research Centre (Australia), and the Centro de
Investigaciones Energticas, Medioambientales y Tecnolgicas (Spain),
National Meteorological Services, and certain commercial companies.
Table 3.2 gives references to some international and national projects
that are collecting, processing and archiving information on solar irradiance resources at the Earths surface and subsequently distributing it in
easily accessible formats with understandable quality metrics.
Direct Solar Energy
3.2.4
Possible impact of climate change on resource

potential
Climate change due to an increase of greenhouse gases (GHGs) in the

atmosphere may inuence atmospheric water vapour content, cloud
cover, rainfall and turbidity, and this can impact the resource potential
of solar energy in different regions of the globe. Changes in major climate variables, including cloud cover and solar irradiance at the Earths
surface, have been evaluated using climate models and considering
anthropogenic forcing for the 21st century (Meehl et al., 2007; Meleshko
et al., 2008). These studies found that the pattern of variation of monthly
mean global solar irradiance does not exceed 1% over some regions of
the globe, and it varies from model to model. Currently, there is no other
evidence indicating a substantial impact of global warming on regional
solar resources. Although some research on global dimming and global
brightening indicates a probable impact on irradiance, no current evidence is available. Uncertainty in pattern changes seems to be rather
large, even for large-scale areas of the Earth.
3.3
Technology and applications
This section discusses technical issues for a range of solar technologies,

organized under the following categories: passive solar and daylighting,
Table 3.2 | International and national projects that collect, process and archive information on solar irradiance resources at the Earths surface.
Available Data Sets
Responsible Institution/Agency
Ground-based solar irradiance from 1,280 sites for 1964 to 2009 provided by national meteorological services around the
world.
World Radiation Data Centre, Saint Petersburg, Russian

Federation (wrdc.mgo.rssi.ru)
National Solar Radiation Database that includes 1,454 ground locations for 1991 to 2005. The satellite-modelled solar
data for 1998 to 2005 provided on 10-km grid. The hourly values of solar data can be used to determine solar resources for
collectors.
National Renewable Energy Laboratory, USA (www.nrel.gov)
European Solar Radiation Database that includes measured solar radiation complemented with other meteorological data
necessary for solar engineering. Satellite images from METEOSAT help in improving accuracy in spatial interpolation. Test
Reference Years were also included.
Supported by Commission of the European Communities,

National Weather Services and scientic institutions of the
European countries
The Solar Radiation Atlas of Africa contains information on surface radiation over Europe, Asia Minor and Africa. Data
covering 1985 to 1986 were derived from measurements by METEOSAT 2.
Supported by the Commission of the European Communities
The solar data set for Africa based on images from METEOSAT processed with the Heliosat-2 method covers the period 1985
to 2004 and is supplemented with ground-based solar irradiance.
Ecole des Mines de Paris, France
Typical Meteorological Year (Test Reference Year) data sets of hourly values of solar radiation and meteorological parameters
derived from individual weather observations in long-term (up to 30 years) data sets to establish a typical year of hourly data.
Used by designers of heating and cooling systems and large-scale solar thermal power plants.
National Renewable Energy Laboratory, USA.

National Climatic Data Center, National Oceanic and
Atmospheric Administration, USA. (www.ncdc.noaa.gov)
The solar radiation data for solar energy applications. IEA/SHC Task36 provides a wide range of users with information on
solar radiation resources at Earths surface in easily accessible formats with understandable quality metrics. The task focuses
on development, validation and access to solar resource information derived from surface- and satellite-based platforms.
International Energy Agency (IEA) Solar Heating and Cooling

Programme (SHC). (swera.unep.net)
Solar and Wind Energy Resource Assessment (SWERA) project aimed at developing information tools to simulate RE
development. SWERA provides easy access to high-quality RE resource information and data for users. Covered major areas
of 13 developing countries in Latin America, the Caribbean, Africa and Asia. SWERA produced a range of solar data sets and
maps at better spatial scales of resolution than previously available using satellite- and ground-based observations.
Global Environment Facility-sponsored project. United Nations

Environment Programme (swera.unep.net)
343
Direct Solar Energy
active heating and cooling, PV electricity generation, CSP electricity

generation and solar fuel production. Each section also describes applications of these technologies.
3.3.1
Passive solar and daylighting technologies
Passive solar energy technologies absorb solar energy, store and distribute it in a natural manner (e.g., natural ventilation), without using
mechanical elements (e.g., fans) (Hernandez Gonzalvez, 1996). The term
passive solar building is a qualitative term describing a building that
makes signicant use of solar gain to reduce heating energy consumption based on the natural energy ows of radiation, conduction and
convection. The term passive building is often employed to emphasize
use of passive energy ows in both heating and cooling, including redistribution of absorbed direct solar gains and night cooling (Athienitis and
Santamouris, 2002).
Daylighting technologies are primarily passive, including windows, skylights and shading and reecting devices. A worldwide trend, particularly
in technologically advanced regions, is for an increased mix of passive
and active systems, such as a forced-air system that redistributes passive solar gains in a solar house or automatically controlled shades that
optimize daylight utilization in an ofce building (Tzempelikos et al.,
2010).
The basic elements of passive solar design are windows, conservatories
and other glazed spaces (for solar gain and daylighting), thermal mass,
protection elements, and reectors (Ralegaonkar and Gupta, 2010). With
the combination of these basic elements, different systems are obtained:
direct-gain systems (e.g., the use of windows in combination with walls
able to store energy, solar chimneys, and wind catchers), indirect-gain
systems (e.g., Trombe walls), mixed-gain systems (a combination of
direct-gain and indirect-gain systems, such as conservatories, sunspaces
and greenhouses), and isolated-gain systems. Passive technologies are
integrated with the building and may include the following components:
Windows with high solar transmittance and a high thermal resistance facing towards the Equator as nearly as possible can be
employed to maximize the amount of direct solar gains into the living space while reducing heat losses through the windows in the
heating season and heat gains in the cooling season. Skylights are
also often used for daylighting in ofce buildings and in solaria/
sunspaces.
Building-integrated thermal storage, commonly referred to as thermal mass, may be sensible thermal storage using concrete or brick
materials, or latent thermal storage using phase-change materials
(Mehling and Cabeza, 2008). The most common type of thermal storage is the direct-gain system in which thermal mass is adequately
distributed in the living space, absorbing the direct solar gains.
Storage is particularly important because it performs two essential
functions: storing much of the absorbed direct solar energy for slow
344
Chapter 3
release, and maintaining satisfactory thermal comfort conditions by

limiting the maximum rise in operative (effective) room temperature
(ASHRAE, 2009). Alternatively, a collector-storage wall, known as
a Trombe wall, may be used, in which the thermal mass is placed
directly next to the glazing, with possible air circulation between
the cavity of the wall system and the room. However, this system has
not gained much acceptance because it limits views to the outdoor
environment through the fenestration. Hybrid thermal storage with
active charging and passive heat release can also be employed in
part of a solar building while direct-gain mass is also used (see, e.g.,
the EcoTerra demonstration house (Figure 3.2, left panel), which
uses solar-heated air from a building-integrated photovoltaic/thermal system to heat a ventilated concrete slab). Isolated thermal
storage passively coupled to a fenestration system or solarium/sunspace is another option in passive design.
Well-insulated opaque envelope appropriate for the climatic conditions can be used to reduce heat transfer to and from the outdoor
environment. In most climates, this energy efciency aspect must be
integrated with the passive design. A solar technology that may be
used with opaque envelopes is transparent insulation (Hollands et
al., 2001) combined with thermal mass to store solar gains in a wall,
turning it into an energy-positive element.
Daylighting technologies and advanced solar control systems, such
as automatically controlled shading (internal, external) and xed
shading devices, are particularly suited for daylighting applications in the workplace (Figure 3.2, right panel). These technologies
include electrochromic and thermochromic coatings and newer
technologies such as transparent photovoltaics, which, in addition
to a passive daylight transmission function, also generate electricity. Daylighting is a combination of energy conservation and passive
solar design. It aims to make the most of the natural daylight that
is available. Traditional techniques include: shallow-plan design,
allowing daylight to penetrate all rooms and corridors; light wells in
the centre of buildings; roof lights; tall windows, which allow light
to penetrate deep inside rooms; task lighting directly over the workplace, rather than lighting the whole building interior; and deep
windows that reveal and light room surfaces to cut the risk of glare
(Everett, 1996).
Solariums, also called sunspaces, are a particular case of the directgain passive solar system, but with most surfaces transparent, that
is, made up of fenestration. Solariums are becoming increasingly
attractive both as a retrot option for existing houses and as an
integral part of new buildings (Athienitis and Santamouris, 2002).
The major driving force for this growth is the development of new
advanced energy-efcient glazing.
Some basic rules for optimizing the use of passive solar heating in buildings are the following: buildings should be well insulated to reduce
overall heat losses; they should have a responsive, efcient heating system; they should face towards the Equator, that is, the glazing should
Chapter 3
Direct Solar Energy
BIPV/T
Roof
Exhaust
Fan
q
External
Rolling
Shutter
Variable
Speed Fan
Solar
Tilted
Slats
Dryer
Air
Inlet
Geothermal
Pump
HRV
Passive
Slab
Light
Shelf
DHW
Ventilated
Slab
Internal
Rolling
Shutter
Blinds
Side-Fin
Figure 3.2 | Left: Schematic of thermal mass placement and passive-active systems in a house; solar-heated air from building-integrated photovoltaic/thermal (BIPV/T) roof heats
ventilated slab or domestic hot water (DHW) through heat exchanger; HRV is heat recovery ventilator. Right: Schematic of several daylighting concepts designed to redistribute daylight
into the ofce interior space (Athienitis, 2008).
be concentrated on the equatorial side, as should the main living rooms,

with rooms such as bathrooms on the opposite side; they should avoid
shading by other buildings to benet from the essential mid-winter sun;
and they should be thermally massive to avoid overheating in the summer and on certain sunny days in winter (Everett, 1996).
Clearly, passive technologies cannot be separated from the building itself.
Thus, when estimating the contribution of passive solar gains, the following must be distinguished: 1) buildings specically designed to harness
direct solar gains using passive systems, dened here as solar buildings,
and 2) buildings that harness solar gains through near-equatorial facing
windows; this orientation is more by chance than by design. Few reliable
statistics are available on the adoption of passive design in residential
buildings. Furthermore, the contribution of passive solar gains is missing in existing national statistics. Passive solar is reducing the demand
and is not part of the supply chain, which is what is considered by the
energy statistics.
The passive solar design process itself is in a period of rapid change,
driven by the new technologies becoming affordable, such as the recently
available highly efcient fenestration at the same prices as ordinary glazing. For example, in Canada, double-glazed low-emissivity argon-lled
windows are presently the main glazing technology used; but until a
few years ago, this glazing was about 20 to 40% more expensive than
regular double glazing. These windows are now being used in retrots
of existing homes as well. Many homes also add a solarium during

retrot. The new glazing technologies and solar control systems allow
the design of a larger window area than in the recent past.
In most climates, unless effective solar gain control is employed, there
may be a need to cool the space during the summer. However, the need
for mechanical cooling may often be eliminated by designing for passive cooling. Passive cooling techniques are based on the use of heat
and solar protection techniques, heat storage in thermal mass and heat
dissipation techniques. The specic contribution of passive solar and
energy conservation techniques depends strongly on the climate (UNEP,
2007). Solar-gain control is particularly important during the shoulder seasons when some heating may be required. In adopting larger
window areasenabled by their high thermal resistanceactive solargain control becomes important in solar buildings for both thermal and
visual considerations.
The potential of passive solar cooling in reducing CO2 emissions
has been shown recently (Cabeza et al., 2010; Castell et al., 2010).
Experimental work demonstrates that adequate insulation can reduce
by up to 50% the cooling energy demand of a building during the hot
season. Moreover, including phase-change materials in the alreadyinsulated building envelope can reduce the cooling energy demand in
such buildings further by up to 15%about 1 to 1.5 kg/yr/m2 of CO2
emissions would be saved in these buildings due to reducing the energy
345
Direct Solar Energy
consumption compared to the insulated building without phase-change

material.
Passive solar system applications are mainly of the direct-gain type,
but they can be further subdivided into the following main application
categories: multi-story residential buildings and two-story detached or
semi-detached solar homes (see Figure 3.2, left panel), designed to have
a large equatorial-facing faade to provide the potential for a large solar
capture area (Athienitis, 2008). Perimeter zones and their fenestration
systems in ofce buildings are designed primarily based on daylighting
performance. In this application, the emphasis is usually on reducing
cooling loads, but passive heat gains may be desirable as well during
the heating season (see Figure 3.2, right panel, for a schematic of shading devices).
In addition, residential or commercial buildings may be designed to use
natural or hybrid ventilation systems and techniques for cooling or fresh
air supply, in conjunction with designs for using daylight throughout
the year and direct solar gains during the heating season. These buildings may prot from low summer night temperatures by using night
hybrid ventilation techniques that utilize both mechanical and natural
ventilation processes (Santamouris and Asimakopoulos, 1996; Voss et
al., 2007).
In 2010, passive technologies played a prominent role in the design
of net-zero-energy solar homeshomes that produce as much electrical and thermal energy as they consume in an average year. These
houses are primarily demonstration projects in several countries currently collaborating in the International Energy Agency (IEA) Task 40 of
the Solar Heating and Cooling (SHC) Programme (IEA, 2009b)Energy
Conservation in Buildings and Community Systems Annex 52which
focuses on net-zero-energy solar buildings. Passive technologies are
essential in developing affordable net-zero-energy homes. Passive solar
gains in homes based on the Passive House Standard are expected to
reduce the heating load by about 40%. By extension, systematic passive solar design of highly insulated buildings at a community scale,
with optimal orientation and form of housing, should easily result in
a similar energy saving of 40%. In Europe, according to the Energy
Performance of Buildings Directive recast, Directive 2010/31/EC (The
European Parliament and the Council of the European Union, 2010), all
new buildings must be nearly zero-energy buildings by 31 December
2020, while EU member states should set intermediate targets for 2015.
New buildings occupied and owned by public authorities have to be
nearly zero-energy buildings after 31 December 2018. The nearly zero
or very low amount of energy required should to a very signicant level
be covered by RE sources, including onsite energy production using
combined heat and power generation or district heating and cooling, to
satisfy most of their demand. Measures should also be taken to stimulate building refurbishments into nearly zero-energy buildings.
Low-energy buildings are known under different names. A survey carried out by Concerted Action Energy Performance of Buildings (EPBD)
identied 17 different terms to describe such buildings across Europe,
346
Chapter 3
including: low-energy house, high-performance house, passive house

(Passivhaus), zero-carbon house, zero-energy house, energy-savings
house, energy-positive house and 3-litre house. Concepts that take into
account more parameters than energy demand again use special terms
such as eco-building or green building.
Another IEA AnnexEnergy Conservation through Energy Storage
Implementing Agreement (ECES IA) Annex 23was initiated in
November 2009 (IEA ECES, 2004). The general objective of the Annex is
to ensure that energy storage techniques are properly applied in ultralow-energy buildings and communities. The proper application of energy
storage is expected to increase the likelihood of sustainable building
technologies.
Another passive solar application is natural drying. Grains and many
other agricultural products have to be dried before being stored so that
insects and fungi do not render them unusable. Examples include wheat,
rice, coffee, copra (coconut esh), certain fruits and timber (Twidell and
Weir, 2006). Solar energy dryers vary mainly as to the use of the solar
heat and the arrangement of their major components. Solar dryers
constructed from wood, metal and glass sheets have been evaluated
extensively and used quite widely to dry a full range of tropical crops
(Imre, 2007).
3.3.2
Active solar heating and cooling
Active solar heating and cooling technologies use the Sun and mechanical elements to provide either heating or cooling; various technologies
are discussed here, as well as thermal storage.
3.3.2.1
Solar heating
In a solar heating system, the solar collector transforms solar irradiance into heat and uses a carrier uid (e.g., water, air) to transfer
that heat to a well-insulated storage tank, where it can be used when
needed. The two most important factors in choosing the correct type
of collector are the following: 1) the service to be provided by the
solar collector, and 2) the related desired range of temperature of the
heat-carrier uid. An uncovered absorber, also known as an unglazed
collector, is likely to be limited to low-temperature heat production
(Dufe and Beckman, 2006).
A solar collector can incorporate many different materials and be manufactured using a variety of techniques. Its design is inuenced by the
system in which it will operate and by the climatic conditions of the
installation location.
Flat-plate collectors are the most widely used solar thermal collectors
for residential solar water- and space-heating systems. They are also
used in air-heating systems. A typical at-plate collector consists of an
absorber, a header and riser tube arrangement or a single serpentine
Chapter 3
Direct Solar Energy
tube, a transparent cover, a frame and insulation (Figure 3.3a). For

low-temperature applications, such as the heating of swimming pools,
only a single plate is used as an absorber (Figure 3.3b). Flat-plate collectors demonstrate a good price/performance ratio, as well as a broad
range of mounting possibilities (e.g., on the roof, in the roof itself, or
unattached).
Flat Plate Collectors
Unglazed Solar Collectors

Tube-on-Sheet Collector
Single or
Double
Glazing
Pump
Flow
Glazing Frame
Metal Deck
Rubber
Grommet
Serpentine Plastic
Pipe Collector
Box
Flow Passages
Absorber
Plate
Insulation
Backing
Typically
1 1/2 ABS Pipe
Figure 3.3a | Schematic diagram of thermal solar collectors: Glazed at-plate.
Pump Flow
Evacuated-tube collectors are usually made of parallel rows of transparent glass tubes, in which the absorbers are enclosed, connected to
a header pipe (Figure 3.3c). To reduce heat loss within the frame by
convection, the air is pumped out of the collector tubes to generate
a vacuum. This makes it possible to achieve high temperatures, useful
Figure 3.3b | Schematic diagram of thermal solar collectors: Unglazed tube-on-sheet

and serpentine plastic pipe.
Evacuated-Tube Collectors
ater
ed W
Heat
Insulation
Solar
Energ
Heat
Heat
es to
r Ris
Vapo
Top
ns to
ensed
Cond
Vacuum Indicator
etur
uid R
Botto
fe
Trans
fe
Trans
Copper
Sleeve in
Manifold
Aluminium
Header
Casing
Liq
Evacuated Glass Tube
Copper
Manifold
Evacuated Heat Pipe
Figure 3.3c | Schematic diagram of thermal solar collectors: Evacuated-tube collectors.
347
Direct Solar Energy
Chapter 3
for cooling (see below) or industrial applications. Most vacuum tube

collectors use heat pipes for their core instead of passing liquid directly
through them. Evacuated heat-pipe tubes are composed of multiple
evacuated glass tubes, each containing an absorber plate fused to a
heat pipe. The heat from the hot end of the heat pipes is transferred
to the transfer uid of a domestic hot water or hydronic space-heating
system.
Solar water-heating systems used to produce hot water can be classied
as passive or active solar water heaters (Dufe and Beckman, 2006).
Also of interest are active solar cooling systems, which transform the hot
water produced by solar energy into cold water.
Passive solar water heaters are of two types (Figure 3.4). Integral collector-storage (ICS) or batch systems include black tanks or tubes in
an insulated glazed box. Cold water is preheated as it passes through
the solar collector, with the heated water owing to a standard backup
water heater. The heated water is stored inside the collector itself. In
thermosyphon (TS) systems, a separate storage tank is directly above
the collector. In direct (open-loop) TS systems, the heated water rises from
the collector to the tank and cool water from the tank sinks back into the
collector. In indirect (closed-loop) TS systems (Figure 3.4, left), heated uid
(usually a glycol-water mixture) rises from the collector to an outer tank
that surrounds the water storage tank and acts as a heat exchanger
(double-wall heat exchangers) for separation from potable water. In climates where freezing temperatures are unlikely, many collectors include
an integrated storage tank at the top of the collector. This design has
many cost and user-friendly advantages compared to a system that uses
a separate standalone heat-exchanger tank. It is also appropriate in
households with signicant daytime and evening hot water needs; but
it does not work well in households with predominantly morning draws
because sometimes the tanks can lose most of the collected energy
overnight.
Active solar water heaters rely on electric pumps and controllers to circulate the carrier uid through the collectors. Three types of active solar
water-heating systems are available. Direct circulation systems use pumps
to circulate pressurized potable water directly through the collectors.
These systems are appropriate in areas that do not freeze for long periods
and do not have hard or acidic water. Antifreeze indirect-circulation systems pump heat-transfer uid, which is usually a glycol-water mixture,
through collectors. Heat exchangers transfer the heat from the uid to
the water for use (Figure 3.4, right). Drainback indirect-circulation systems
use pumps to circulate water through the collectors. The water in the
collector and the piping system drains into a reservoir tank when the
pumps stop, eliminating the risk of freezing in cold climates. This system should be carefully designed and installed to ensure that the piping
always slopes downward to the reservoir tank. Also, stratication should
be carefully considered in the design of the water tank (Hadorn, 2005).
A solar combisystem provides both solar space heating and cooling as
well as hot water from a common array of solar thermal collectors, usually backed up by an auxiliary non-solar heat source (Weiss, 2003). Solar
combisystems may range in size from those installed in individual properties to those serving several in a block heating scheme. A large number
of different types of solar combisystems are produced. The systems on
the market in a particular country may be more restricted, however,
because different systems have tended to evolve in different countries.
Solar
Collector
A Close-Coupled
Solar
Energy
Solar Water
Heater
Solar Energy
To Taps
Tank
Controller
Boiler
Arrows Show
Direction of Water
Flow Through Copper
Cold
Water
Feed
Pipes When the Sun

Heats the Collector Panels.
Figure 3.4 | Generic schematics of thermal solar systems. Left: Passive (thermosyphon). Right: Active system.
348
Pump
Chapter 3
Depending on the size of the combisystem installed, the annual space

heating contribution can range from 10 to 60% or more in ultra-low
energy Passivhaus-type buildings, and even up to 100% where a large
seasonal thermal store or concentrating solar thermal heat is used.
Direct Solar Energy
include the heat recovery units, heat exchangers and humidiers. Liquid
sorption techniques have been demonstrated successfully.
3.3.2.3
3.3.2.2
Solar cooling
Solar cooling can be broadly categorized into solar electric refrigeration, solar thermal refrigeration, and solar thermal air-conditioning.
In the rst category, the solar electric compression refrigeration uses
PV panels to power a conventional refrigeration machine (Fong et al.,
2010). In the second category, the refrigeration effect can be produced
through solar thermal gain; solar mechanical compression refrigeration,
solar absorption refrigeration, and solar adsorption refrigeration are the
three common options. In the third category, the conditioned air can be
directly provided through the solar thermal gain by means of desiccant
cooling. Both solid and liquid sorbents are available, such as silica gel
and lithium chloride, respectively.
Solar electrical air-conditioning, powered by PV panels, is of minor interest from a systems perspective, unless there is an off-grid application
(Henning, 2007). This is because in industrialized countries, which have
a well-developed electricity grid, the maximum use of photovoltaics is
achieved by feeding the produced electricity into the public grid.
Solar thermal air-conditioning consists of solar heat powering an absorption chiller and it can be used in buildings (Henning, 2007). Deploying
such a technology depends heavily on the industrial deployment of lowcost small-power absorption chillers. This technology is being studied
within the IEA Task 25 on solar-assisted air-conditioning of buildings,
SHC program and IEA Task 38 on solar air-conditioning and refrigeration, SHC program.
Closed heat-driven cooling systems using these cycles have been known
for many years and are usually used for large capacities of 100 kW
and greater. The physical principle used in most systems is based on
the sorption phenomenon. Two technologies are established to produce
thermally driven low- and medium-temperature refrigeration: absorption and adsorption.
Open cooling cycle (or desiccant cooling) systems are mainly of interest
for the air conditioning of buildings. They can use solid or liquid sorption. The central component of any open solar-assisted cooling system
is the dehumidication unit. In most systems using solid sorption, this
unit is a desiccant wheel. Various sorption materials can be used, such
as silica gel or lithium chloride. All other system components are found
in standard air-conditioning applications with an air-handling unit and
Thermal storage
Thermal storage within thermal solar systems is a key component to

ensure reliability and efciency. Four main types of thermal energy storage technologies can be distinguished: sensible, latent, sorption and
thermochemical heat storage (Hadorn, 2005; Paksoy, 2007; Mehling and
Cabeza, 2008; Dincer and Rosen, 2010).
Sensible heat storage systems use the heat capacity of a material. The
vast majority of systems on the market use water for heat storage. Water
heat storage covers a broad range of capacities, from several hundred
litres to tens of thousands of cubic metres.
Latent heat storage systems store thermal energy during the phase
change, either melting or evaporation, of a material. Depending on the
temperature range, this type of storage is more compact than heat storage in water. Melting processes have energy densities of the order of
100 kWh/m3 (360 MJ/m3), compared to 25 kWh/m3 (90 MJ/m3) for sensible heat storage. Most of the current latent heat storage technologies
for low temperatures store heat in building structures to improve thermal performance, or in cold storage systems. For medium-temperature
storage, the storage materials are nitrate salts. Pilot storage units in the
100-kW range currently operate using solar-produced steam.
Sorption heat storage systems store heat in materials using water
vapour taken up by a sorption material. The material can either be a solid
(adsorption) or a liquid (absorption). These technologies are still largely
in the development phase, but some are on the market. In principle,
sorption heat storage densities can be more than four times higher than
sensible heat storage in water.
Thermochemical heat storage systems store heat in an endothermic
chemical reaction. Some chemicals store heat 20 times more densely
than water (at a T100C); but more typically, the storage densities
are 8 to 10 times higher. Few thermochemical storage systems have
been demonstrated. The materials currently being studied are the salts
that can exist in anhydrous and hydrated form. Thermochemical systems
can compactly store low- and medium-temperature heat. Thermal storage is discussed with specic reference to higher-temperature CSP in
Section 3.3.4.
Underground thermal energy storage is used for seasonal storage and
includes the various technologies described below. The most frequently
used storage technology that makes use of the underground is aquifer
349
Direct Solar Energy
thermal energy storage. This technology uses a natural underground layer

(e.g., sand, sandstone or chalk) as a storage medium for the temporary
storage of heat or cold. The transfer of thermal energy is realized by
extracting groundwater from the layer and by re-injecting it at the modied temperature level at a separate location nearby. Most applications
are for the storage of winter cold to be used for the cooling of large
ofce buildings and industrial processes. Aquifer cold storage is gaining interest because savings on electricity bills for chillers are about
75%, and in many cases, the payback time for additional investments
is shorter than ve years. A major condition for the application of this
technology is the availability of a suitable geologic formation.
3.3.2.4
Active solar heating and cooling applications
For active solar heating and cooling applications, the amount of hot
water produced depends on the type and size of the system, amount of
sun available at the site, seasonal hot-water demand pattern, and installation characteristics of the system (Norton, 2001).
Solar heating for industrial processes is at a very early stage of development in 2010 (POSHIP, 2001). Worldwide, less than 100 operating solar
thermal systems for process heat are reported, with a total capacity of
about 24 MWth (34,000 m collector area). Most systems are at an experimental stage and relatively small scale. However, signicant potential
exists for market and technological developments, because 28% of the
overall energy demand in the EU27 countries originates in the industrial
sector, and much of this demand is for heat below 250C. Education and
knowledge dissemination are needed to deploy this technology.
In the short term, solar heating for industrial processes will mainly be used
for low-temperature processes, ranging from 20C to 100C. With technological development, an increasing number of medium-temperature
applicationsup to 250Cwill become feasible within the market.
According to Werner (2006), about 30% of the total industrial heat
demand is required at temperatures below 100C, which could theoretically be met with solar heating using current technologies. About 57%
of this demand is required at temperatures below 400C, which could
largely be supplied by solar in the foreseeable future.
In several specic industry sectorssuch as food, wine and beverages,
transport equipment, machinery, textiles, and pulp and paperthe
share of heat demand at low and medium temperatures (below 250C)
is around 60% (POSHIP, 2001). Tapping into this low- and mediumtemperature heat demand with solar heat could provide a signicant
opportunity for solar contribution to industrial energy requirements. A
substantial opportunity for solar thermal systems also exists in chemical industries and in washing processes.
Among the industrial processes, desalination and water treatment
(e.g., sterilization) are particularly promising applications for solar
thermal energy, because these processes require large amounts of
350
Chapter 3
medium-temperature heat and are often necessary in areas with high

solar irradiance and high energy costs.
Some process heat applications can be met with temperatures delivered by ordinary low-temperature collectors, namely, from 30C to
80C. However, the bulk of the demand for industrial process heat
requires temperatures from 80C to 250C.
Process heat collectors are another potential application for solar
thermal heat collectors. Typically, these systems require a large capacity (hence, large collector areas), low costs, and high reliability and
quality. Although low- and high-temperature collectors are offered
in a dynamically growing market, process heat collectors are at a
very early stage of development and no products are available on an
industrial scale. In addition to concentrating collectors, improved at
collectors with double and triple glazing are currently being developed, which could meet needs for process heat in the range of up
to 120C. Concentrating-type solar collectors are described in Section
3.3.4.
Solar refrigeration is used, for example, to cool stored vaccines. The
need for such systems is greatest in peripheral health centres in rural
communities in the developing world, where no electrical grid is
available.
Solar cooling is a specic area of application for solar thermal technology. High-efciency at plates, evacuated tubes or parabolic
troughs can be used to drive absorption cycles to provide cooling. For
a greater coefcient of performance (COP), collectors with low concentration levels can provide the temperatures (up to around 250C)
needed for double-effect absorption cycles. There is a natural match
between solar energy and the need for cooling.
A number of closed heat-driven cooling systems have been built,
using solar thermal energy as the main source of heat. These systems
often have large cooling capacities of up to several hundred kW. Since
the early 2000s, a number of systems have been developed in the
small-capacity range, below 100 kW, and, in particular, below 20 kW
and down to 4.5 kW. These small systems are single-effect machines
of different types, used mainly for residential buildings and small commercial applications.
Although open-cooling cycles are generally used for air conditioning
in buildings, closed heat-driven cooling cycles can be used for both air
conditioning and industrial refrigeration.
Other solar applications are listed below. The production of potable
water using solar energy has been readily adopted in remote or
isolated regions (Narayan et al., 2010). Solar stills are widely used
in some parts of the world (e.g., Puerto Rico) to supply water to
households of up to 10 people (Khanna et al., 2008). In appropriate
isolation conditions, solar detoxication can be an effective low-cost
Chapter 3
treatment for low-contaminant waste (Gumy et al., 2006). Multipleeffect humidication (MEH) desalination units indirectly use heat
from highly efcient solar thermal collectors to induce evaporation
and condensation inside a thermally isolated, steam-tight container.
These MEH systems are now beginning to appear in the market. Also
see the report on water desalination by CSP (DLR, 2007) and the discussion of SolarPACES Task VI (SolarPACES, 2009b).
In solar drying, solar energy is used either as the sole source of the
required heat or as a supplemental source, and the air ow can be
generated by either forced or free (natural) convection (Fudholi et al.,
2010). Solar cooking is one of the most widely used solar applications
in developing countries (Lahkar and Samdarshi, 2010) though might still
be considered an early stage commercial product due to limited overall
deployment in comparison to other cooking methods. A solar cooker
uses sunlight as its energy source, so no fuel is needed and operating
costs are zero. Also, a reliable solar cooker can be constructed easily and
quickly from common materials.
Direct Solar Energy
Anti-Reection Coating
n-Type Semiconductor
Front Contact
Electron (-)
Hole (+)
Recombination
p-Type Semiconductor
Back Contact
Figure 3.5 | Generic schematic cross-section illustrating the operation of an illuminated

solar cell.
3.3.3
Photovoltaic electricity generation
Photovoltaic (PV) solar technologies generate electricity by exploiting

the photovoltaic effect. Light shining on a semiconductor such as silicon (Si) generates electron-hole pairs that are separated spatially by an
internal electric eld created by introducing special impurities into the
semiconductor on either side of an interface known as a p-n junction.
This creates negative charges on one side of the interface and positive
charges are on the other side (Figure 3.5). This resulting charge separation creates a voltage. When the two sides of the illuminated cell are
connected to a load, current ows from one side of the device via the
load to the other side of the cell. The conversion efciency of a solar cell
is dened as a ratio of output power from the solar cell with unit area
(W/cm2) to the incident solar irradiance. The maximum potential efciency of a solar cell depends on the absorber material properties and
device design. One technique for increasing solar cell efciency is with a
multijunction approach that stacks specially selected absorber materials
that can collect more of the solar spectrum since each different material
can collect solar photons of different wavelengths.
PV cells consist of organic or inorganic matter. Inorganic cells are based
on silicon or non-silicon materials; they are classied as wafer-based cells
or thin-lm cells. Wafer-based silicon is divided into two different types:
monocrystalline and multicrystalline (sometimes called polycrystalline).
3.3.3.1
Existing photovoltaic technologies
Existing PV technologies include wafer-based crystalline silicon (c-Si)

cells, as well as thin-lm cells based on copper indium/gallium disulde/diselenide (CuInGaSe2; CIGS), cadmium telluride (CdTe), and
thin-lm silicon (amorphous and microcrystalline silicon). Mono- and
multicrystalline silicon wafer PV (including ribbon technologies) are the

dominant technologies on the PV market, with a 2009 market share
of about 80%; thin-lm PV (primarily CdTe and thin-lm Si) has the
remaining 20% share. Organic PV (OPV) consists of organic absorber
materials and is an emerging class of solar cells.
Wafer-based silicon technology includes solar cells made of monocrystalline or multicrystalline wafers with a current thickness of around 200
m, while the thickness is decreasing down to 150 m. Single-junction
wafer-based c-Si cells have been independently veried to have record
energy conversion efciencies of 25.0% for monocrystalline silicon
cells and 20.3% for multicrystalline cells (Green et al., 2010b) under
standard test conditions (i.e., irradiance of 1,000 W/m2, air-mass 1.5,
25C). The theoretical Shockley-Queisser limit of a single-junction cell
with an energy bandgap of crystalline silicon is 31% energy conversion
efciency (Shockley and Queisser, 1961).
Several variations of wafer-based c-Si PV for higher efciency have
been developed, for example, heterojunction solar cells and interdigitated back-contact (IBC) solar cells. Heterojunction solar cells consist
of a crystalline silicon wafer base sandwiched by very thin (~5 nm)
amorphous silicon layers for passivation and emitter. The highest-efciency heterojunction solar cell is 23.0% for a 100.4-cm2 cell (Taguchi
et al., 2009). Another advantage is a lower temperature coefcient. The
efciency of conventional c-Si solar cells declines with elevating ambient temperature at a rate of -0.45%/C, while the heterojunction cells
show a lower rate of -0.25%/C (Taguchi et al., 2009). An IBC solar
cell, where both the base and emitter are contacted at the back of the
cell, has the advantage of no shading of the front of the cell by a top
electrode. The highest efciency of such a back-contact silicon wafer
351
Direct Solar Energy
cell is 24.2% for 155.1 cm2 (Bunea et al., 2010). Commercial module
efciencies for wafer-based silicon PV range from 12 to 14% for multicrystalline Si and from 14 to 20% for monocrystalline Si.
Commercial thin-lm PV technologies include a range of absorber
material systems: amorphous silicon (a-Si), amorphous silicon-germanium, microcrystalline silicon, CdTe and CIGS. These thin-lm cells have
an absorber layer thickness of a few m or less and are deposited on
glass, metal or plastic substrates with areas of up to 5.7 m2 (Stein et al.,
2009).
The a-Si solar cell, introduced in 1976 (Carlson and Wronski, 1976) with
initial efciencies of 1 to 2%, has been the rst commercially successful
thin-lm PV technology. Because a-Si has a higher light absorption coefcient than c-Si, the thickness of an a-Si cell can be less than 1 mthat
is, more than 100 times thinner than a c-Si cell. Developing higher efciencies for a-Si cells has been limited by inherent material quality and
by light-induced degradation identied as the Staebler-Wronski effect
(Staebler and Wronski, 1977). However, research efforts have successfully lowered the impact of the Staebler-Wronski effect to around 10%
or less by controlling the microstructure of the lm. The highest stabilized efciencythe efciency after the light-induced degradationis
reported as 10.1% (Benagli et al., 2009).
Higher efciency has been achieved by using multijunction technologies
with alloy materials, e.g., germanium and carbon or with microcrystalline silicon, to form semiconductors with lower or higher bandgaps,
respectively, to cover a wider range of the solar spectrum (Yang and
Guha, 1992; Yamamoto et al., 1994; Meier et al., 1997). Stabilized
efciencies of 12 to 13% have been measured for various laboratory
devices (Green et al., 2010b).
CdTe solar cells using a heterojunction with cadmium sulphide (CdS)
have a suitable energy bandgap of 1.45 electron-volt (eV) (0.232 aJ)
with a high coefcient of light absorption. The best efciency of this
cell is 16.7% (Green et al., 2010b) and the best commercially available
modules have an efciency of about 10 to 11%.
The toxicity of metallic cadmium and the relative scarcity of tellurium
are issues commonly associated with this technology. Although several
assessments of the risk (Fthenakis and Kim, 2009; Zayed and Philippe,
2009) and scarcity (Green et al., 2009; Wadia et al., 2009) are available,
no consensus exists on these issues. It has been reported that this potential hazard can be mitigated by using a glass-sandwiched module design
and by recycling the entire module and any industrial waste (Sinha et
al., 2008).
The CIGS material family is the basis of the highest-efciency thin-lm
solar cells to date. The copper indium diselenide (CuInSe2)/CdS solar
cell was invented in the early 1970s at AT&T Bell Labs (Wagner et al.,
1974). Incorporating Ga and/or S to produce CuInGa(Se,S)2 results in the
benet of a widened bandgap depending on the composition (Dimmler
and Schock, 1996). CIGS-based solar cells have been validated at an
352
Chapter 3
efciency of 20.1% (Green et al., 2010b). Due to higher efciencies and

lower manufacturing energy consumptions, CIGS cells are currently in
the industrialization phase, with best commercial module efciencies
of up to 13.1% (Kushiya, 2009) for CuInGaSe2 and 8.6% for CuInS2
(Meeder et al., 2007). Although it is acknowledged that the scarcity of
In might be an issue, Wadia et al. (2009) found that the current known
economic indium reserves would allow the installation of more than 10
TW of CIGS-based PV systems.
High-efciency solar cells based on a multijunction technology using
III-V semiconductors (i.e., based on elements from the III and V columns
of the periodic chart), for example, gallium arsenide (GaAs) and gallium
indium phosphide (GaInP) , can have superior efciencies. These cells
were originally developed for space use and are already commercialized. An economically feasible terrestrial application is the use of these
cells in concentrating PV (CPV) systems, where concentrating optics are
used to focus sunlight onto high efciency solar cells (Bosi and Pelosi,
2007). The most commonly used cell is a triple-junction device based on
GaInP/GaAs/germanium (Ge), with a record efciency of 41.6% for a
lattice-matched cell (Green et al., 2010b) and 41.1% for a metamorphic
or lattice-mismatched device (Bett et al., 2009). Sub-module efciencies have reached 36.1% (Green et al., 2010b). Another advantage of
the concentrator system is that cell efciencies increase under higher
irradiance (Bosi and Pelosi, 2007), and the cell area can be decreased in
proportion to the concentration level. Concentrator applications, however, require direct-normal irradiation, and are thus suited for specic
climate conditions with low cloud coverage.
3.3.3.2
Emerging photovoltaic technologies
Emerging PV technologies are still under development and in laboratory

or (pre-) pilot stage, but could become commercially viable within the
next decade. They are based on very low-cost materials and/or processes
and include technologies such as dye-sensitized solar cells, organic solar
cells and low-cost (printed) versions of existing inorganic thin-lm
technologies.
Electricity generation by dye-sensitized solar cells (DSSCs) is based on
light absorption in dye molecules (the sensitizers) attached to the very
large surface area of a nanoporous oxide semiconductor electrode (usually titanium dioxide), followed by injection of excited electrons from the
dye into the oxide. The dye/oxide interface thus serves as the separator
of negative and positive charges, like the p-n junction in other devices.
The negatively charged electrons are then transported through the semiconductor electrode and reach the counter electrode through the load,
thus generating electricity. The injected electrons from the dye molecules
are replenished by electrons supplied through a liquid electrolyte that
penetrates the pores of the semiconductor electrode, providing the electrical path from the counter electrode (Graetzel, 2001). State-of-the-art
DSSCs have achieved a top conversion efciency of 10.4% (Chiba et
al., 2005). Despite the gradual improvements since its discovery in 1991
(ORegan and Graetzel, 1991), long-term stability against ultraviolet light
Chapter 3
irradiation, electrolyte leakage and high ambient temperatures continue

to be key issues in commercializing these PV cells.
Organic PV (OPV) cells use stacked solid organic semiconductors, either
polymers or small organic molecules. A typical structure of a smallmolecule OPV cell consists of a stack of p-type and n-type organic
semiconductors forming a planar heterojunction. The short-lived nature
of the tightly bound electron-hole pairs (excitons) formed upon light
absorption limits the thickness of the semiconductor layers that can be
usedand therefore, the efciency of such devices. Note that excitons
need to move to the interface where positive and negative charges can
be separated before they recombine. If the travel distance is short, the
active thickness of material is small and not all light can be absorbed
within that thickness.
The efciency achieved with single-junction OPV cells is about 5% (Li et
al., 2005), although predictions indicate about twice that value or higher
can be achieved (Forrest, 2005; Koster et al., 2006). To decouple exciton
transport distances from optical thickness (light absorption), so-called
bulk-heterojunction devices have been developed. In these devices,
the absorption layer is made of a nanoscale mixture of p- and n-type
materials to allow excitons to reach the interface within their lifetime,
while also enabling a sufcient macroscopic layer thickness. This bulkheterojunction structure plays a key role in improving the efciency, to
a record value of 7.9% in 2009 (Green et al., 2010a). The developments
in cost and processing (Brabec, 2004; Krebs, 2005) of materials have
caused OPV research to advance further. Also, the main development
challenge is to achieve a sufciently high stability in combination with
a reasonable efciency.
Direct Solar Energy
Committee, 2001; Navigant Consulting Inc., 2006; EU PV European

Photovoltaic Technology Platform, 2007; Kroposki et al., 2008; NEDO,
2009).
At the component level, BOS components for grid-connected applications
are not yet sufciently developed to match the lifetime of PV modules.
Additionally, BOS component and installation costs need to be reduced.
Moreover, devices for storing large amounts of electricity (over 1 MWh
or 3,600 MJ) will be adapted to large PV systems in the new energy
network. As new module technologies emerge in the future, some of the
ideas relating to BOS may need to be revised. Furthermore, the quality
of the system needs to be assured and adequately maintained according
to dened standards, guidelines and procedures. To ensure system quality, assessing performance is important, including on-line analysis (e.g.,
early fault detection) and off-line analysis of PV systems. The knowledge
gathered can help to validate software for predicting the energy yield of
future module and system technology designs.
To increasingly penetrate the energy network, PV systems must use
technology that is compatible with the electric grid and energy supply
and demand. System designs and operation technologies must also be
developed in response to demand patterns by developing technology to
forecast the power generation volume and to optimize the storage function. Moreover, inverters must improve the quality of grid electricity by
controlling reactive power or ltering harmonics with communication in
a new energy network that uses a mixture of inexpensive and effective
communications systems and technologies, as well as smart meters (see
Section 8.2.1).
3.3.3.5
3.3.3.3
Novel technologies are potentially disruptive (high-risk, high-potential)

approaches based on new materials, devices and conversion concepts.
Generally, their practically achievable conversion efciencies and cost
structure are still unclear. Examples of these approaches include intermediate-band semiconductors, hot-carrier devices, spectrum converters,
plasmonic solar cells, and various applications of quantum dots (Section
3.7.3). The emerging technologies described in the previous section primarily aim at very low cost, while achieving a sufciently high efciency
and stability. However, most of the novel technologies aim at reaching
very high efciencies by making better use of the entire solar spectrum
from infrared to ultraviolet.
3.3.3.4
Photovoltaic applications
Novel photovoltaic technologies
Photovoltaic systems
A photovoltaic system is composed of the PV module, as well as the

balance of system (BOS) components, which include an inverter, storage
devices, charge controller, system structure, and the energy network. The
system must be reliable, cost effective, attractive and match with the
electric grid in the future (US Photovoltaic Industry Roadmap Steering
Photovoltaic applications include PV power systems classied into two

major types: those not connected to the traditional power grid (i.e., off-grid
applications) and those that are connected (i.e., grid-connected applications). In addition, there is a much smaller, but stable, market segment
for consumer applications.
Off-grid PV systems have a signicant opportunity for economic application in the un-electried areas of developing countries. Figure 3.6
shows the ratio of various off-grid and grid-connected systems in the
Photovoltaic Power Systems (PVPS) Programme countries. Of the total
capacity installed in these countries during 2009, only about 1.2% was
installed in off-grid systems that now make up 4.2% of the cumulative
installed PV capacity of the IEA PVPS countries (IEA, 2010e).
Off-grid centralized PV mini-grid systems have become a reliable alternative for village electrication over the last few years. In a PV mini-grid
system, energy allocation is possible. For a village located in an isolated
area and with houses not separated by too great a distance, the power
may ow in the mini-grid without considerable losses. Centralized
systems for local power supply have different technical advantages concerning electrical performance, reduction of storage needs, availability
353
Installed PV Power [MWp]
Direct Solar Energy
25,000
Cumulative Grid-Connected
20,000
Cumulative Off-Grid
an existing structure; and the PV array itself can be used as a cladding

or roong material, as in building-integrated PV (Eiffert, 2002; Ecofys
Netherlands BV, 2007; Elzinga, 2008).
An often-cited disadvantage is the greater sensitivity to grid interconnection issues, such as overvoltage and unintended islanding (Kobayashi
and Takasaki, 2006; Cobben et al., 2008; Ropp et al., 2008). However,
much progress has been made to mitigate these effects, and today, by
Institute of Electrical and Electronics Engineers (IEEE) and Underwriter
Laboratories standards (IEEE 1547 (2008), UL 1741), all inverters must
have the function of the anti-islanding effect.
15,000
10,000
5,000
0
92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09
Figure 3.6 | Historical trends in cumulative installed PV power of off-grid and gridconnected systems in the OECD countries (IEA, 2010e). Vertical axis is in peak megawatts.
of energy, and dynamic behaviour. Centralized PV mini-grid systems

could be the least-cost options for a given level of service, and they may
have a diesel generator set as an optional balancing system or operate
as a hybrid PV-wind-diesel system. These kinds of systems are relevant
for reducing and avoiding diesel generator use in remote areas (Munoz
et al., 2007; Sreeraj et al., 2010).
Grid-connected PV systems use an inverter to convert electricity from
direct current (DC)as produced by the PV arrayto alternating current (AC), and then supply the generated electricity to the electricity
network. Compared to an off-grid installation, system costs are lower
because energy storage is not generally required, since the grid is used
as a buffer. The annual output yield ranges from 300 to 2,000 kWh/
kW (Clavadetscher and Nordmann, 2007; Gaiddon and Jedliczka, 2007;
Kurokawa et al., 2007; Photovoltaic Geographic Information System,
2008) for several installation conditions in the world. The average annual
performance ratiothe ratio between average AC system efciency and
standard DC module efciencyranges from 0.7 to 0.8 (Clavadetscher
and Nordmann, 2007) and gradually increases further to about 0.9 for
specic technologies and applications.
Grid-connected PV systems are classied into two types of applications:
distributed and centralized. Grid-connected distributed PV systems are
installed to provide power to a grid-connected customer or directly to
the electricity network. Such systems may be: 1) on or integrated into
the customers premises, often on the demand side of the electricity
meter; 2) on public and commercial buildings; or 3) simply in the built
environment such as on motorway sound barriers. Typical sizes are 1 to
4 kW for residential systems, and 10 kW to several MW for rooftops on
public and industrial buildings.
These systems have a number of advantages: distribution losses in the
electricity network are reduced because the system is installed at the
point of use; extra land is not required for the PV system, and costs
for mounting the systems can be reduced if the system is mounted on
354
Chapter 3
Grid-connected centralized PV systems perform the functions of centralized power stations. The power supplied by such a system is not
associated with a particular electricity customer, and the system is
not located to specically perform functions on the electricity network
other than the supply of bulk power. Typically, centralized systems are
mounted on the ground, and they are larger than 1 MW.
The economical advantage of these systems is the optimization of installation and operating cost by bulk buying and the cost effectiveness of
the PV components and balance of systems at a large scale. In addition,
the reliability of centralized PV systems can be greater than distributed
PV systems because they can have maintenance systems with monitoring equipment, which can be a smaller part of the total system cost.
Multi-functional PV, daylighting and solar thermal components involving PV or solar thermal that have already been introduced into the built
environment include the following: shading systems made from PV
and/or solar thermal collectors; hybrid PV/thermal (PV/T) systems that
generate electricity and heat from the same panel/collector area; semitransparent PV windows that generate electricity and transmit daylight
from the same surface; faade collectors; PV roofs; thermal energy roof
systems; and solar thermal roof-ridge collectors. Currently, fundamental and applied R&D activities are also underway related to developing
other products, such as transparent solar thermal window collectors, as
well as faade elements that consist of vacuum-insulation panels, PV
panels, heat pump, and a heat-recovery system connected to localized
ventilation.
Solar energy can be integrated within the building envelope and with
energy conservation methods and smart-building operating strategies.
Much work over the last decade or so has gone into this integration,
culminating in the net-zero energy building.
Much of the early emphasis was on integrating PV systems with thermal
and daylighting systems. Bazilian et al. (2001) and Tripanagnostopoulos
(2007) listed methods for doing this and reviewed case studies where
the methods had been applied. For example, PV cells can be laid on
the absorber plate of a at-plate solar collector. About 6 to 20% of the
solar energy absorbed on the cells is converted to electricity; the remaining roughly 80% is available as low-temperature heat to be transferred
to the uid being heated. The resulting unit produces both heat and
Chapter 3
Direct Solar Energy
electricity and requires only slightly more than half the area used if the
two conversion devices had been mounted side by side and worked
independently. PV cells have also been developed to be applied to windows to allow daylighting and passive solar gain. Reviews of recent
work in this area are provided by Chow (2010) and Arif Hasan and
Sumathy (2010).
gas, nuclear, oil or biomasscomes from creating a hot uid. CSP simply provides an alternative heat source. Therefore, an attraction of this
technology is that it builds on much of the current know-how on power
generation in the world today. And it will benet not only from ongoing
advances in solar concentrator technology, but also as improvements
continue to be made in steam and gas turbine cycles.
Considerable work has also been done on architecturally integrating the

solar components into the building. Any new solar building should be
very well insulated, well sealed, and have highly efcient windows and
heat recovery systems. Probst and Roecker (2007), surveying the opinions of more than 170 architects and engineers who examined numerous
existing solar buildings, concluded the following: 1) best integration is
achieved when the solar component is integrated as a construction element, and 2) appearanceincluding collector colour, orientation and
jointingmust sometimes take precedence over performance in the
overall design. In describing 16 case studies of building-integrated photovoltaics, Eiffert and Kiss (2000) identied two main products available
on the architectural market: faade systems and roof systems. Faade
systems include curtain wall products, spandrel panels and glazings;
roong products include tiles, shingles, standing-seam products and
skylights. These can be integrated as components or constitute the
entire structure (as in the case of a bus shelter).
Any concentrating solar system depends on direct-beam irradiation

as opposed to global horizontal irradiation as for at-plate systems.
Thus, sites must be chosen accordingly, and the best sites for CSP are
in near-equatorial cloud-free regions such as the North African desert.
The average capacity factor of a solar plant will depend on the quality
of the solar resource.
The idea of the net-zero-energy solar building has sparked recent interest. Such buildings send as much excess PV-generated electrical energy
to the grid as the energy they draw over the year. An IEA Task is considering how to achieve this goal (IEA NZEB, 2009). Recent examples for
the Canadian climate are provided by Athienitis (2008). Starting from a
building that meets the highest levels of conservation, these homes use
hybrid air-heating/PV panels on the roof; the heated air is used for space
heating or as a source for a heat pump. Solar water-heating collectors
are included, as is fenestration permitting a large passive gain through
equatorial-facing windows. A key feature is a ground-source heat pump,
which provides a small amount of residual heating in the winter and
cooling in the summer.
Smart solar-building control strategies may be used to manage the collection, storage and distribution of locally produced solar electricity
and heat to reduce and shift peak electricity demand from the grid. An
example of a smart solar-building design is given by Candanedo and
Athienitis (2010), where predictive control based on weather forecasts
one day ahead and real-time prediction of building response are used to
optimize energy performance while reducing peak electricity demand.
3.3.4
Concentrating solar power electricity generation
Concentrating solar power (CSP) technologies produce electricity by

concentrating direct-beam solar irradiance to heat a liquid, solid or gas
that is then used in a downstream process for electricity generation. The
majority of the worlds electricity todaywhether generated by coal,
Some of the key advantages of CSP include the following: 1) it can be

installed in a range of capacities to suit varying applications and conditions, from tens of kW (dish/Stirling systems) to multiple MWs (tower
and trough systems); 2) it can integrate thermal storage for peaking
loads (less than one hour) and intermediate loads (three to six hours);
3) it has modular and scalable components; and 4) it does not require
exotic materials. This section discusses various types of CSP systems and
thermal storage for these systems.
Large-scale CSP plants most commonly concentrate sunlight by reection, as opposed to refraction with lenses. Concentration is either to a
line (linear focus) as in trough or linear Fresnel systems or to a point
(point focus) as in central-receiver or dish systems. The major features of
each type of CSP system are illustrated in Figure 3.7 and are described
below.
In trough concentrators, long rows of parabolic reectors concentrate
the solar irradiance by the order of 70 to 100 times onto a heat collection element (HCE) mounted along the reectors focal line. The troughs
track the Sun around one axis, with the axis typically being oriented
north-south. The HCE comprises a steel inner pipe (coated with a solarselective surface) and a glass outer tube, with an evacuated space in
between. Heat-transfer oil is circulated through the steel pipe and heated
to about 390C. The hot oil from numerous rows of troughs is passed
through a heat exchanger to generate steam for a conventional steam
turbine generator (Rankine cycle). Land requirements are of the order of
2 km2 for a 100-MWe plant, depending on the collector technology and
assuming no storage. Alternative heat transfer uids to the synthetic oil
commonly used in trough receivers, such as steam and molten salt, are
being developed to enable higher temperatures and overall efciencies,
as well as integrated thermal storage in the case of molten salt.
Linear Fresnel reectors use long lines of at or nearly at mirrors, which
allow the moving parts to be mounted closer to the ground, thus reducing structural costs. (In contrast, large trough reectors presently use
thermal bending to achieve the curve required in the glass surface.) The
receiver is a xed inverted cavity that can have a simpler construction
than evacuated tubes and be more exible in sizing. The attraction of
355
Direct Solar Energy
Chapter 3
(a)
(b)
Reector
Absorber Tube
Curved Mirrors
Curved Mirrors
Solar Field Piping
Absorber Tube and

Reconcentrator
(c)
(d)
Central Receiver
Reector
Receiver/Engine
Heliostats
Figure 3.7 | Schematic diagrams showing the underlying principles of four basic CSP congurations: (a) parabolic trough, (b) linear Fresnel reector, (c) central receiver/power tower,
and (d) dish systems (Richter et al., 2009).
linear Fresnel reectors is that the installed costs on a per square metre
basis can be lower than for trough systems. However, the annual optical
performance is less than that for a trough.
Central receivers (or power towers), which are one type of point-focus
collector, are able to generate much higher temperatures than troughs
and linear Fresnel reectors, although requiring two-axis tracking as
the Sun moves through solar azimuth and solar elevation. This higher
356
temperature is a benet because higher-temperature thermodynamic

cycles used for generating electricity are more efcient. This technology
uses an array of mirrors (heliostats), with each mirror tracking the Sun
and reecting the light onto a xed receiver atop a tower. Temperatures
of more than 1,000C can be reached. Central receivers can easily generate the maximum temperatures of advanced steam turbines, can use
high-temperature molten salt as the heat transfer uid, and can be used
to power gas turbine (Brayton) cycles.
Chapter 3
Dish systems include an ideal optical reector and therefore are suitable
for applications requiring high temperatures. Dish reectors are paraboloid
and concentrate the solar irradiation onto a receiver mounted at the
focal point, with the receiver moving with the dish. Dishes have been
used to power Stirling engines at 900C, and also for steam generation. There is now signicant operational experience with dish/Stirling
engine systems, and commercial rollout is planned. In 2010, the capacity of each Stirling engine is smallon the order of 10 to 25 kWelectric.
The largest solar dishes have a 485-m2 aperture and are in research
facilities or demonstration plants.
In thermal storage, the heat from the solar eld is stored prior to
reaching the turbine. Thermal storage takes the form of sensible or
latent heat storage (Gil et al., 2010; Medrano et al., 2010). The solar
eld needs to be oversized so that enough heat can be supplied to
both operate the turbine during the day and, in parallel, charge the
thermal storage. The term solar multiple refers to the total solar eld
area installed divided by the solar eld area needed to operate the turbine at design point without storage. Thermal storage for CSP systems
needs to be at a temperature higher than that needed for the working uid of the turbine. As such, system temperatures are generally
between 400C and 600C, with the lower end for troughs and the
higher end for towers. Allowable temperatures are also dictated by
the limits of the media available. Examples of storage media include
molten salt (presently comprising separate hot and cold tanks), steam
accumulators (for short-term storage only), solid ceramic particles,
high-temperature phase-change materials, graphite, and high-temperature concrete. The heat can then be drawn from the storage to
generate steam for a turbine, as and when needed. Another type of
storage associated with high-temperature CSP is thermochemical storage, where solar energy is stored chemically. This is discussed more
fully in Sections 3.3.5 and 3.7.5.
Thermal energy storage integrated into a system is an important attribute of CSP. Until recently, this has been primarily for operational
purposes, providing 30 minutes to 1 hour of full-load storage. This
eases the impact of thermal transients such as clouds on the plant,
assists start-up and shut-down, and provides benets to the grid.
Trough plants are now designed for 6 to 7.5 hours of storage, which is
enough to allow operation well into the evening when peak demand
can occur and tariffs are high. Trough plants in Spain are now operating with molten-salt storage. In the USA, Abengoa Solars 280-MW
Solana trough project, planned to be operational by 2013, intends
to integrate six hours of thermal storage. Towers, with their higher
temperatures, can charge and store molten salt more efciently.
Gemasolar, a 17-MWe solar tower project under construction in Spain,
is designed for 15 hours of storage, giving a 75% annual capacity factor (Arce et al., 2011).
Thermal storage is a means of providing dispatchability. Hybridization
with non-renewable fuels is another way in which CSP can be
designed to be dispatchable. Although the back-up fuel itself may
Direct Solar Energy
not be renewable (unless it is biomass-derived), it provides signicant

operational benets for the turbine and improves solar yield.
CSP applications range from small distributed systems of tens of kW to
large centralized power stations of hundreds of MW.
Stirling and Brayton cycle generation in CSP can be installed in a wide
range from small distributed systems to clusters forming medium- to
large-capacity power stations. The dish/Stirling technology has been
under development for many years, with advances in dish structures, high-temperature receivers, use of hydrogen as the circulating
working uid, as well as some experiments with liquid metals and
improvements in Stirling enginesall bringing the technology closer
to commercial deployment. Although the individual unit size may only
be of the order of tens of kWe, power stations having a large capacity
of up to 800 MWe have been proposed by aggregating many modules.
Because each dish represents a stand-alone electricity generator, from
the perspective of distributed generation there is great exibility in
the capacity and rate at which units are installed. However, the dish
technology is less likely to integrate thermal storage.
An alternative to the Stirling engine is the Brayton cycle, as used by
gas turbines. The attraction of these engines for CSP is that they are
already in signicant production, being used for distributed generation
red with landll gas or natural gas. In the solarized version, the air is
instead heated by concentrated solar irradiance from a tower or dish
reector. It is also possible to integrate with a biogas or natural gas
combustor to back up the solar. Several developments are currently
underway based on solar tower and micro-turbine combinations.
Centralized CSP benets from the economies of scale offered by largescale plants. Based on conventional steam and gas turbine cycles,
much of the technological know-how of large power station design
and practice is already in place. However, although larger capacity has
signicant cost benets, it has also tended to be an inhibitor until
recently because of the much larger investment commitment required
from investors. In addition, larger power stations require strong infrastructural support, and new or augmented transmission capacity may
be needed.
The earliest commercial CSP plants were the 354 MW of Solar Electric
Generating Stations in Californiadeployed between 1985 and
1991that continue to operate commercially today. As a result of the
positive experiences and lessons learned from these early plants, the
trough systems tend to be the technology most often applied today as
the CSP industry grows. In Spain, regulations to date have mandated
that the largest capacity unit that can be installed is 50 MWe to help
stimulate industry competition. In the USA, this limitation does not
exist, and proposals are in place for much larger plants280 MWe in
the case of troughs and 400-MWe plants (made up of four modules)
based on towers. There are presently two operational solar towers of
10 and 20 MWe, and all tower developers plan to increase capacity in
357
Direct Solar Energy
line with technology development, regulations and investment capital.

Multiple dishes have also been proposed as a source of aggregated
heat, rather than distributed-generation Stirling or Brayton units.
CSP or PV electricity can also be used to power reverse-osmosis plants
for desalination. Dedicated CSP desalination cycles based on pressure and temperature are also being developed for desalination (see
Section 3.3.2).
3.3.5
Solar fuel production
Solar fuel technologies convert solar energy into chemical fuels, which
can be a desirable method of storing and transporting solar energy.They
can be used in a much wider variety of higher-efciency applications
than just electricity generation cycles. Solar fuels can be processed into
liquid transportation fuels or used directly to generate electricity in
fuel cells; they can be employed as fuels for high-efciency gas-turbine
cycles or internal combustion engines; and they can serve for upgrading
fossil fuels, CO2 synthesis, or for producing industrial or domestic heat.
The challenge is to produce large amounts of chemical fuels directly
from sunlight in cost-effective ways and to minimize adverse effects on
the environment (Steinfeld and Meier, 2004).
Solar fuels that can be produced include synthesis gas (syngas, i.e.,
mixed gases of carbon monoxide and hydrogen), pure hydrogen (H2)
gas, dimethyl ether (DME) and liquids such as methanol and diesel. The
high energy density of H2 (on a mass basis) and clean conversion give it
attractive properties as a future fuel and it is also used as a feedstock for
many industrial processes. H2 has a higher energy density than batteries,
although batteries have a higher round-trip efciency. However, its very
low energy density on a volumetric basis poses economic challenges
associated with its storage and transport. It will require signicant new
distribution infrastructure and either new designs of internal combustion
engine or a move to fuel cells. Additionally, the synthesis of hydrogen
with CO2 can produce hydrocarbon fuels that are compatible with existing infrastructures. DME gas is similar to liqueed petroleum gas (LPG)
and easily stored. Methanol is liquid and can replace gasoline without
signicant changes to the engine or the fuel distribution infrastructure.
Methanol and DME can be used for fuel cells after reforming, and DME
can also be used in place of LPG. Fischer-Tropsch processes can produce
hydrocarbon fuels and electricity (see Sections 2.6 and 8.2.4).
There are three basic routes, alone or in combination, for producing
storable and transportable fuels from solar energy: 1) the electrochemical route uses solar electricity from PV or CSP systems followed by an
electrolytic process; 2) the photochemical/photobiological route makes
direct use of solar photon energy for photochemical and photobiological
processes; and 3) the thermochemical route uses solar heat at moderate
and/or high temperatures followed by an endothermic thermochemical
process (Steinfeld and Meier, 2004). Note that the electrochemical and
thermochemical routes apply to any RE technology, not exclusively to
solar technologies.
358
Chapter 3
Figure 3.8 illustrates possible pathways to produce H2 or syngas from

water and/or fossil fuels using concentrated solar energy as the source
of high-temperature process heat. Feedstocks include inorganic compounds such as water and CO2, and organic sources such as coal,
biomass and natural gas (NG). See Chapter 2 for parallels with biomass-derived syngas.
Electrolysis of water can use solar electricity generated by PV or CSP
technology in a conventional (alkaline) electrolyzer, considered a
benchmark for producing solar hydrogen. With current technologies,
the overall solar-to-hydrogen energy conversion efciency ranges
between 10 and 14%, assuming electrolyzers working at 70% efciency and solar electricity being produced at 15% (PV) and 20%
(CSP) annual efciency. The electricity demand for electrolysis can be
signicantly reduced if the electrolysis of water proceeds at higher
temperatures (800 to 1,000C) via solid-oxide electrolyzer cells
(Jensen et al., 2007). In this case, concentrated solar energy can be
applied to provide both the high-temperature process heat and the
electricity needed for the high-temperature electrolysis.
Thermolysis and thermochemical cycles are a long-term sustainable
and carbon-neutral approach for hydrogen production from water. This
route involves energy-consuming (endothermic) reactions that make
use of concentrated solar irradiance as the energy source for hightemperature process heat (Abanades et al., 2006). Solar thermolysis
requires temperatures above 2,200C and raises difcult challenges
for reactor materials and gas separation. Water-splitting thermochemical cycles allow operation at lower temperature, but require several
chemical reaction steps and also raise challenges because of inefciencies associated with heat transfer and product separation at each
step.
Decarbonization of fossil fuels is a near- to mid-term transition pathway to solar hydrogen that encompasses the carbothermal reduction
of metal oxides (Epstein et al., 2008) and the decarbonization of fossil
fuels via solar cracking (Spath and Amos, 2003; Rodat et al., 2009),
reforming (Mller et al., 2006) and gasication (ZGraggen and
Steinfeld, 2008; Piatkowski et al., 2009). These routes are being pursued by European, Australian and US academic and industrial research
consortia. Their technical feasibility has been demonstrated in concentrating solar chemical pilot plants at the power level of 100 to 500
kWth. Solar hybrid fuel can be produced by supplying concentrated
solar thermal energy to the endothermic processes of methane and
biomass reformingthat is, solar heat is used for process energy only,
and fossil fuels are still a required input. Some countries having vast
solar and natural gas resources, but a relatively small domestic energy
market (e.g., the Middle East and Australia) are in a position to produce and export solar energy in the form of liquid fuels.
Solar fuel synthesis from solar hydrogen and CO2 produces hydrocarbons that are compatible with existing energy infrastructures such as
the natural gas network or existing fuel supply structures. The renewable methane process combines solar hydrogen with CO2 from the
Chapter 3
Direct Solar Energy
Concentrated
Solar Energy
H2O
Solar Thermolysis
Decarbonization
H2O Splitting
Solar Thermochemical
Cycle
Solar Electricity &

Electrolysis
Solar Reforming
Solar Cracking
Fossil Fuels
(NG, Oil, Coal)
Solar
Gasication
Optional CO2/C
Sequestration
Solar Fuels (Hydrogen, Syngas)
Figure 3.8 | Thermochemical routes for solar fuels production, indicating the chemical source of H2: water (H2O) for solar thermolysis and solar thermochemical cycles to produce H2
only; fossil or biomass fuels as feedstock for solar cracking to produce H2 and carbon (C); or a combination of fossil/biomass fuels and H2O/CO2 for solar reforming and gasication to
produce syngas, H2 and carbon monoxide (CO). For the solar decarbonization processes, sequestration of the CO2/C may be considered (from Steinfeld and Meier, 2004; Steinfeld, 2005).
atmosphere or other sources in a synthesis reactor with a nickel catalyst. In this way, a substitute for natural gas is produced that can be
stored, transported and used in gas power plants, heating systems
and gas vehicles (Sterner, 2009).
Solar methane can be produced using water, air, solar energy and a
source of CO2. Possible CO2 sources are biomass, industry processes
or the atmosphere. CO2 is regarded as the carrier for hydrogen in this
energy system. By separating CO2 from the combustion process of
solar methane, CO2 can be recycled in the energy system or stored
permanently. Thus, carbon sink energy systems powered by RE can
be created (Sterner, 2009). The rst pilot plants at the kW scale with
atmospheric CO2 absorption have been set up in Germany, proving the
technical feasibility. Scaling up to the utility MW scale is planned in
the next few years (Specht et al., 2010).
In an alternative conversion step, liquid fuels such as Fischer-Tropsch
diesel, DME, methanol or solar kerosene (jet fuel) can be produced
from solar energy and CO2/water (H2O) for long-distance transportation. The main advantages of these solar fuels are the same range
as fossil fuels (compared to the generally reduced range of electric
vehicles), less competition for land use, and higher per-hectare yields
compared to biofuels. Solar energy can be harvested via natural photosynthesis in biofuels with an efciency of 0.5%, via PV power and
solar fuel conversion (technical photosynthesis) with an efciency of

10% (Sterner, 2009) and via solar-driven thermochemical dissociation
of CO2 and H2O using metal oxide redox reactions, yielding a syngas
mixture of carbon monoxide (CO) and H2, with a solar-to-fuel efciency approaching 20% (Chueh et al., 2010). This approach would
provide a solution to the issues and controversy surrounding existing
biofuels, although the cost of this technology is a possible constraint.
3.4
Global and regional status of market and

industry development
This section looks at the ve key solar technologies, rst focusing on

installed capacity and generated energy, then on industry capacity
and supply chains, and nally on the impact of policies specic to
these technologies.
3.4.1
Installed capacity and generated energy
This subsection discusses the installed capacity and generated energy

within the ve technology areas of passive solar, active solar heating
and cooling, PV electricity generation, CSP electricity generation, and
solar fuel production.
359
Direct Solar Energy
For passive solar technologies, no estimates are available at this time for
the installed capacity of passive solar or the energy generated or saved
through this technology.
For active solar heating, the total installed capacity worldwide was
about 149 GWth in 2008 and 180 GWth in 2009 (Weiss and Mauthner,
2010; REN21, 2010).
In 2008, new capacity of 29.1 GWth, corresponding to 41.5 million m2 of
solar collectors, was installed worldwide (Weiss and Mauthner, 2010).
In 2008, China accounted for about 79% of the installations of glazed
collectors, followed by the EU with 14.5%.
The overall new installations grew by 34.9% compared to 2007. The
growth rate in 2006/2007 was 18.8%. The main reasons for this growth
were the high growth rates of glazed water collectors in China, Europe
and the USA.
In 2008, the global market had high growth rates for evacuated-tube
collectors and at-plate collectors, compared to 2007. The market for
unglazed air collectors also increased signicantly, mainly due to the
installation of 23.9 MWth of new systems in Canada.
Compared to 2007, the 2008 installation rates for new unglazed, glazed
at-plate, and evacuated-tube collectors were signicantly up in Jordan,
Cyprus, Canada, Ireland, Germany, Slovenia, Macedonia (FYROM),
Tunisia, Poland, Belgium and South Africa.
New installations in China, the worlds largest market, again increased
signicantly in 2008 compared to 2007, reaching 21.7 GWth. After a
market decline in Japan in 2007, the growth rate was once again positive in 2008.
Market decreases compared to 2007 were reported for Israel, the Slovak
Republic and the Chinese province of Taiwan.
The main markets for unglazed water collectors are still found in the
USA (0.8 GWth), Australia (0.4 GWth), and Brazil (0.08 GWth). Notable
markets are also in Austria, Canada, Mexico, The Netherlands, South
Africa, Spain, Sweden and Switzerland, with values between 0.07 and
0.01 GWth of new installed unglazed water collectors in 2008.
Comparison of markets in different countries is difcult due to the
wide range of designs used for different climates and different demand
requirements. In Scandinavia and Germany, a solar heating system
will typically be a combined water-heating and space-heating system,
known as a solar combisystem, with a collector area of 10 to 20 m2. In
Japan, the number of solar domestic water-heating systems is large, but
most installations are simple integral preheating systems. The market in
Israel is large due to a favourable climate, as well as regulations mandating installation of solar water heaters. The largest market is in China,
where there is widespread adoption of advanced evacuated-tube solar
360
Chapter 3
collectors. In terms of per capita use, Cyprus is the leading country in

the world, with an installed capacity of 527 kWth per 1,000 inhabitants.
The type of application of solar thermal energy varies greatly in different countries (Weiss and Mauthner, 2010). In China (88.7 GWth), Europe
(20.9 GWth) and Japan (4.4 GWth), at-plate and evacuated-tube collectors mainly prepare hot water and provide space heating. However,
in the USA and Canada, swimming pool heating is still the dominant
application, with an installed capacity of 12.9 GWth of unglazed plastic
collectors.
The biggest reported solar thermal system for industrial process heat
was installed in China in 2007. The 9 MWth plant produces heat for a textile company. About 150 large-scale plants (>500 m2; 350 kWth)1 with a
total capacity of 160MWth are in operation in Europe. The largest plants
for solar-assisted district heating are located in Denmark (13 MWth) and
Sweden (7 MWth).
In Europe, the market size more than tripled between 2002 and 2008.
However, even in the leading European solar thermal markets of Austria,
Greece, and Germany, only a minor portion of residential homes use
solar thermal. For example, in Germany, only about 5% of one- and twofamily homes are using solar thermal energy.
The European market has the largest variety of different solar thermal
applications, including systems for hot-water preparation, plants for
space heating of single- and multi-family houses and hotels, large-scale
plants for district heating, and a growing number of systems for airconditioning, cooling and industrial applications.
Advanced applications such as solar cooling and air conditioning
(Henning, 2004, 2007), industrial applications (POSHIP, 2001) and desalination/water treatment are in the early stages of development. Only a
few hundred rst-generation systems are in operation.
For PV electricity generation, newly installed capacity in 2009 was
about 7.5 GW, with shipments to rst point in the market at 7.9 GW
(Jger-Waldau, 2010a; Mints, 2010). This addition brought the cumulative installed PV capacity worldwide to about 22 GWa capacity
able to generate up to 26 TWh (93,600 TJ) per year. More than 90%
of this capacity is installed in three leading markets: the EU27 with 16
GW (73%), Japan with 2.6 GW (12%), and the USA with 1.7 GW (8%)
(Jger-Waldau, 2010b). These markets are dominated by grid-connected
PV systems, and growth within PV markets has been stimulated by
various government programmes around the world. Examples of such
programmes include feed-in tariffs in Germany and Spain, and various
mechanisms in the USA, such as buy-down incentives, investment tax
credits, performance-based incentives and RE quota systems. For 2010,
1
To enable comparison, the IEAs Solar Heating and Cooling Programme, together
with the European Solar Thermal Industry Federation and other major solar thermal
trade associations, publish statistics in kWth (kilowatt thermal) and use a factor of
0.7 kWth/m2 to convert square metres of collector area into installed thermal capacity
(kWth).
Chapter 3
Direct Solar Energy
Cumulative Installed Capacity [MW]
Figure 3.9 illustrates the cumulative installed capacity for the top eight
PV markets through 2009, including Germany (9,800 MW), Spain (3,500
MW), Japan (2,630 MW), the USA (1,650 MW), Italy (1,140 MW), Korea
(460 MW), France (370MW) and the Peoples Republic of China (300
MW). By far, Spain and Germany have seen the largest amounts of
growth in installed PV capacity in recent years, with Spain seeing a huge
surge in 2008 and Germany having experienced steady growth over the
last ve years.
10,000
9,000
8,000
Germany
Italy
7,000
Spain
Korea
6,000
Japan
France
5,000
USA
China
4,000
3,000
2,000
1,000
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Figure 3.9 | Installed PV capacity in eight markets. Data sources: EurObservER (2009);
IEA (2009c); REN21 (2009); and Jger-Waldau (2010b).
Concentrating photovoltaics (CPV) is an emerging market with about 17

MW of cumulative installed capacity at the end of 2008. The two main
tracks are high-concentration PV (>300 times or 300 suns) and lowto medium-concentration PV with a concentration factor of 2 to about
300 (2 to ~300 suns). To maximize the benets of CPV, the technology
requires high direct-beam irradiance, and these areas have a limited
geographical rangethe Sun Belt of the Earth. The market share of
CPV is still small, but an increasing number of companies are focusing
on CPV. In 2008, about 10 MW of CPV were installed, and market estimates for 2009 are in the 20 to 30MW range; for 2010, about 100MW
are expected.
661/2007 has been a major driving force for CSP plant construction and
expansion plans. As of November 2009, 2,340 MWe of CSP projects had
been preregistered for the tariff provisions of the Royal Decree. In the
USA, more than 4,500 MWe of CSP are currently under power purchase
agreement contracts. The different contracts specify when the projects
must start delivering electricity between 2010 and 2015 (Bloem et al.,
2010). More than 10,000 MWe of new CSP plants have been proposed in
the USA. More than 50 CSP electricity projects are currently in the planning phase, mainly in North Africa, Spain and the USA. In Australia, the
federal government has called for 1,000 MWe of new solar plants, covering both CSP and PV, under the Solar Flagships programme. Figure 3.10
shows the current and planned deployment to add more CSP capacity
in the near future.
Hybrid solar/fossil plants have received increasing attention in recent
years, and several integrated solar combined-cycle (ISCC) projects
have been either commissioned or are under construction in the
Mediterranean region and the USA. The rst plant in Morocco (Ain
Beni Mathar: 470 MW total, 22 MW solar) began operating in June
2010, and two additional plants in Algeria (Hassi RMel: 150 MW total,
30 MW solar) and Egypt (Al Kuraymat: 140 MW total, 20 MW solar)
are under construction. In Italy, another example of an ISCC project is
Archimede; however, the plants 31,000-m2 parabolic trough solar eld
will be the rst to use molten salt as the heat transfer uid (SolarPACES,
2009a).
Solar fuel production technologies are in an earlier stage of development. The high-temperature solar reactor technology is typically being
developed at a laboratory scale of 1 to 10 kWth solar power input.
Installed Capacity [MW]
the market is estimated between 9 and 24 GW of additional installed

PV systems, with a consensus value in the 13GW range (Jger-Waldau,
2010a).
12,000
South Africa
China
10,000
Israel
Jordan
Egypt
8,000
Algeria
Morocco
Tunesia
Regarding CSP electricity generation, at the beginning of 2009, more

than 700 MWe of grid-connected CSP plants were installed worldwide,
with another 1,500 MWe under construction (Torres et al., 2010). The
majority of installed plants use parabolic trough technology. Centralreceiver technology comprises a growing share of plants under
construction and those announced. The bulk of the operating capacity is
installed in Spain and the south-western United States.
In 2007, after a hiatus of more than 15 years, the rst major CSP plants
came on line with Nevada Solar One (64 MWe, USA) and PS10 (11 MWe,
Spain). In Spain, successive Royal Decrees have been in place since 2004
and have stimulated the CSP industry in that country. Royal Decree
6,000
Abu Dhabi
Australia
Spain
4,000
USA
2,000
0
1990
2000
2006
2007
2008
2009
2010
2012
2015
Figure 3.10 | Installed and planned concentrated solar power plants by country (Bloem
et al., 2010).
361
Direct Solar Energy
Scaling up thermochemical processes for hydrogen production to the

100-kWth power level is reported for a medium-temperature mixed
iron oxide cycle (800C to 1,200C) (Roeb et al., 2006, 2009) and for
the high-temperature zinc oxide (ZnO) dissociation reaction at above
1,700C (Schunk et al., 2008, 2009). Pilot plants in the power range of
300 to 500 kWth have been built for the carbothermic reduction of ZnO
(Epstein et al., 2008), the steam reforming of methane (Mller et al.,
2006), and the steam gasication of petcoke (ZGraggen and Steinfeld,
2008). Solar-to-gas has been demonstrated at a 30-kW scale to drive
a commercial natural gas vehicle, applying a nickel catalyst (Specht et
al., 2010). Demonstration at the MW scale should be warranted before
erecting commercial solar chemical plants for fuels production, which
are expected to be available only after 2020 (Pregger et al., 2009).
Direct conversion of solar energy to fuel is not yet widely demonstrated
or commercialized. But two options appear commercially feasible in the
near to medium term: 1) the solar hybrid fuel production system (including solar methane reforming and solar biomass reforming), and 2) solar
PV or CSP electrolysis.
Australias Commonwealth Scientic and Industrial Research
Organisation is running a 250-kWth reactor and plans to build a
MW-scale demonstration plant using solar steam-reforming technology,
with an eventual move to CO2 reforming for higher performance and
less water usage. With such a system, liquid solar fuels can be produced
in sunbelts such as Australia and solar energy shipped on a commercial
basis to Asia and beyond.
Oxygen gas produced by solar (PV or CSP) electrolysis can be used for
coal gasication and partial oxidation of natural gas. With the combined
process of solar electrolysis and partial oxidation of coal or methane,
theoretically 10 to 15% of solar energy is incorporated into the methanol or DME. Also, the production cost of the solar hybrid fuel can be
lower than the solar hydrogen produced by the solar electrolysis process
only.
3.4.2
Industry capacity and supply chain
This subsection discusses the industry capacity and supply chain within
the ve technology areas of passive solar, active solar heating and cooling, PV electricity generation, CSP electricity generation and solar fuel
production.
In passive solar technologies, people make up part of the industry
capacity and the supply chain: namely, the engineers and architects
who collaborate to produce passively heated buildings. Close collaboration between the two disciplines has often been missing in the past,
but the dissemination of systematic design methodologies issued by
362
Chapter 3
different countries has improved the design capabilities (Athienitis and

Santamouris, 2002).
The integration of passive solar systems with the active heating/cooling air-conditioning systems both in the design and operation stages
of the building is essential to achieve good comfort conditions while
saving energy. However, this is often overlooked because of inadequate
collaboration for integrating building design between architects and
engineers. Thus, the architect often designs the building envelope based
solely on qualitative passive solar design principles, and the engineer
often designs the heating-ventilation-air-conditioning system based
on extreme design conditions without factoring in the benets due to
solar gains and natural cooling. The result may be an oversized system
and inappropriate controls incompatible with the passive system and
that can cause overheating and discomfort (Athienitis and Santamouris,
2002). Collaboration between the disciplines involved in building design
is now improving with the adoption of computer tools for integrated
analysis and design.
The design of high-mass buildings with signicant near-equatorial-facing
window areas is common in some areas of the world such as Southern
Europe. However, a systematic approach to designing such buildings is
still not widely employed. This is changing with the introduction of the
passive house standard in Germany and other countries (PHPP, 2004),
the deployment of the European Directives, and new national laws such
as Chinas standard based on the German one.
Glazing and window technologies have made substantial progress in
the last 20 years (Hollands et al., 2001). New-generation windows result
in low energy losses, high daylight efciency, solar shading, and noise
reduction. New technologies such as transparent PV and electrochromic
and thermochromic windows provide many possibilities for designing
solar houses and ofces with abundant daylight. The change from regular double-glazed to double-glazed low-emissivity argon windows is
presently occurring in Canada and is accelerated by the rapid drop in
prices of these windows.
The primary materials for low-temperature thermal storage in passive
solar systems are concrete, bricks and water. A review of thermal storage materials is given by Hadorn (2008) under IEA SHC Task 32, focusing
on a comparison of the different technologies. Phase-change material
(PCM) thermal storage (Mehling and Cabeza, 2008) is particularly
promising in the design, control and load management of solar buildings because it reduces the need for structural reinforcement required
for heavier traditional sensible storage in concrete-type construction.
Recent developments facilitating integration include microencapsulated
PCM that can be mixed with plaster and applied to interior surfaces
(Schossig et al., 2005). PCM in microencapsulated polymers is now on
the market and can be added to plaster, gypsum or concrete to enhance
Chapter 3
In spite of the advances in PCM, concrete has certain advantages for

thermal storage when a massive building design approach is used, as
in many of the Mediterranean countries. In this approach, the concrete
also serves as the structure of the building and is thus likely more cost
effective than thermal storage without this added function.
For active solar heating and cooling, a number of different collector
technologies and system approaches have been developed due to different applicationsincluding domestic hot water, heating, preheating
and combined systemsand varying climatic conditions.
Figure 3.11 plots the increase in production from 2000 through 2009,
showing regional contributions (Jger-Waldau, 2010a). The compound
annual growth rate in production from 2003 to 2009 was more than
50%.
Annual PV Production [MW]
the thermal capacity of a room. For renovation, this provides a good

alternative to new heavy walls, which would require additional structural support (Hadorn, 2008).
Direct Solar Energy
12,000
Rest of World
10,000
United States
China
8,000
Europe
Japan
6,000
4,000
The largest exporters of solar water-heating systems are Australia,

Greece and the USA. The majority of exports from Greece are to Cyprus
and the near-Mediterranean area. France also sends a substantial
number of systems to its overseas territories. The majority of US exports
are to the Caribbean region. Australian companies export about 50%
of production (mainly thermosyphon systems with external horizontal
tanks) to most of the areas of the world that do not have hard-freeze
conditions.
PV electricity generation is discussed under the areas of overall solar
cell production, thin-lm module production and polysilicon production.
The development characteristic of the PV sector is much different than
the traditional power sector, more closely resembling the semiconductor market, with annual growth rates between 40 to 50% and a
high learning rate. Therefore, scientic and peer-reviewed papers can
be several years behind the actual market developments due to the
nature of statistical time delays and data consolidation. The only way
to keep track of such a dynamic market is to use commercial market
data. Global PV cell production2 reached more than 11.5 GW in 2009.
2
Solar cell production capacities mean the following: for wafer-silicon-based solar
cells, only the cells; for thin lms, the complete integrated module. Only those companies that actually produce the active circuit (solar cell) are counted; companies
that purchase these circuits and then make modules are not counted.
2,000
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Figure 3.11 | Worldwide PV production from 2000 to 2009 (Jger-Waldau, 2010b).
The announced production capacitiesbased on a survey of more

than 300 companies worldwideincreased despite very difcult economic conditions in 2009 (Figure 3.12) (Jger-Waldau, 2010b). Only
published announcements from the respective companies, not thirdparty information, were used. April 2010 was the cut-off date for the
information included. This method has the drawback that not all companies announce their capacity increases in advance; also, in times of
nancial tightening, announcements of scale-backs in expansion plans
are often delayed to prevent upsetting nancial markets. Therefore, the
capacity gures provide a trend, but do not represent nal numbers.
In 2008 and 2009, Chinese production capacity increased overproportionally. In actual production, China surpassed all other countries,
Annual Production/Production Capacity [MW]
In some parts of the production process, such as selective coatings,

large-scale industrial production levels have been attained. A number of
different materials, including copper, aluminium and stainless steel, are
applied and combined with different welding technologies to achieve
a highly efcient heat-exchange process in the collector. The materials used for the cover glass are structured or at, low-iron glass. The
rst antireection coatings are coming onto the market on an industrial
scale, leading to efciency improvements of about 5%.
In general, vacuum-tube collectors are well-suited for higher-temperature
applications. The production of vacuum-tube collectors is currently dominated by the Chinese Dewar tubes, where a metallic heat exchanger is
integrated to connect them with the conventional hot-water systems.
In addition, some standard vacuum-tube collectors, with metallic heat
absorbers, are on the market.
70,000
60,000
ROW
India
50,000
South Korea
USA
40,000
30,000
China
Europe
Japan
20,000
10,000
0
Estimated
Production
2009
Planned
Capacity
2009
Planned
Capacity
2010
Planned
Capacity
2012
Planned
Capacity
2015
Figure 3.12 | Worldwide annual PV production in 2009 compared to the announced

production capacities (Jger-Waldau, 2010a).
363
Direct Solar Energy
Chapter 3
estimated in 2009 at between 5.4 and 6.1GW (including 1.5 to 1.7GW

production in the Chinese province of Taiwan), Europe had 2.0 to 2.2GW,
and was followed by Japan, with 1.5 to 1.7GW (Jger-Waldau, 2010b).
In terms of production, First Solar (USA/Germany/France/Malaysia) was
number one (1,082 MW), followed by Suntech (China) estimated at
750MW and Sharp (Japan) estimated at 580MW.
If all these ambitious plans can be realized by 2015, then China will
have about 51% (including 16% in the Chinese province of Taiwan) of
the worldwide production capacity of 70GW, followed by Europe (15%)
and Japan (13%).
Production Capacity [MW/yr]
Worldwide, more than 300 companies produce solar cells. In 2009,

silicon-based solar cells and modules represented about 80% of the
worldwide market (Figure 3.13). In addition to a massive increase in production capacities, the current development predicts that thin-lm-based
solar cells will increase their market share to over 30% by 2012.
70,000
60,000
Crystalline Wafer Silicon

Thin Films
50,000
40,000
30,000
20,000
10,000
0
2006
2009
2010
2012
2015
Figure 3.13 | Actual (2006) and announced (2009 to 2015) production capacities of
thin-lm and crystalline silicon-based solar modules (Jger-Waldau, 2010b).
In 2005, production of thin-lm PV modules grew to more than 100 MW

per year. Since then, the compound annual growth rate of thin-lm PV
module production was higher than that of the industrythus increasing the market share of thin-lm products from 6% in 2005 to about
20% in 2009. Most of this thin-lm share comes from the largest PV
company.
More than 150 companies are involved in the thin-lm solar cell production process, ranging from R&D activities to major manufacturing plants.
The rst 100-MW thin-lm factories became operational in 2007, and
the announcements of new production capacities accelerated again in
2008. If all expansion plans are realized in time, thin-lm production
capacity could be 20.0GW, or 35% of the total 56.7GW in 2012, and
23.5 GW, or 34% of a total of 70 GW in 2015 (Jger-Waldau, 2009,
364
2010b). The rst thin-lm factories with GW production capacity are

already under construction for various thin-lm technologies.
The rapid growth of the PV industry since 2000 led to the situation
between 2004 and early 2008 where the demand for polysilicon outstripped the supply from the semiconductor industry. This led to a silicon
shortage, which resulted in silicon spot-market prices as high as USD2005
450/kg (USD2005, assumed 2008 base) in 2008 compared to USD2005 25.5/
kg in 2003 and consequently higher prices for PV modules. This extreme
price hike triggered the massive capacity expansion, not only of established companies, but of many new entrants as well.
The six companies that reported shipment gures delivered together
about 43,900 tonnes of polysilicon in 2008, as reported by Semiconductor
Equipment and Materials International (SEMI, 2009a). In 2008, these
companies had a production capacity of 48,200 tonnes of polysilicon (Service, 2009). However, all polysilicon producers, including new
entrants with current and alternative technologies, had a production
capacity of more than 90,000 tonnes of polysilicon in 2008. Considering
that not all new capacity actually produced polysilicon at nameplate
capacity in 2008, it was estimated that 62,000 tonnes of polysilicon
could be produced. Subtracting the needs of the semiconductor industry
and adding recycling and excess production, the available amount of
silicon for the PV industry was estimated at 46,000 tonnes of polysilicon. With an average material need of 8.7 g/Wp (p = peak), this would
have been sufcient for the production of 5.3 GW of crystalline silicon
PV cells.
The drive to reduce costs and secure key markets has led to the emergence of two interesting trends. One is the move to large original design
manufacturing units, similar to the developments in the semiconductor
industry. A second is that an increasing number of solar manufacturers
move part of their module production close to the nal market to demonstrate the local job creation potential and ensure the current policy
support. This may also be a move to manufacture in low-cost or subsidized markets.
The regional distribution of polysilicon production capacities is as follows: China 20,000 tonnes, Europe 17,500 tonnes, Japan 12,000 tonnes,
and USA 37,000 tonnes (Service, 2009).
In 2009, solar-grade silicon production of about 88,000 tonnes was
reported, sufcient for about 11GW of PV assuming an average materials need of 8g/Wp (Displaybank, 2010). China produced about 18,000
tonnes or 20% of world demand, fullling about half of its domestic
demand (Baoshan, 2010).
Projections of silicon production capacities for solar applications in 2012
span a range between 140,000tonnes from established polysilicon producers, up to 250,000tonnes including new producers (e.g., Bernreuther
Chapter 3
and Haugwitz, 2010; Ruhl et al., 2010). The possible solar cell production will also depend on the material use per Wp. Material consumption
could decrease from the current 8g/Wp to 7g/Wp or even 6g/Wp (which
could increase delivered PV capacity from 31 to 36 to 42 GW, respectively), but this may not be achieved by all manufacturers.
Forecasts of the future costs of vital materials have a high-prole history,
and there is ongoing public debate about possible material shortages
and competition regarding some (semi-)metals (e.g., In and Te) used in
thin-lm cell production. In a recent study, Wadia et al. (2009) explored
material limits for PV expansion by examining the dual constraints of
material supply and least cost per watt for the most promising semiconductors as active photo-generating materials. Contrary to the commonly
assumed scarcity of indium and tellurium, the study concluded that
the currently known economic reserves of these materials would allow
about 10TW of CdTe or CuInS2 solar cells to be installed.
In CSP electricity generation, the solar collector eld is readily scalable,
and the power block is based on adapted knowledge from the existing
power industry such as steam and gas turbines. The collectors themselves
benet from a range of existing skill sets such as mechanical, structural
and control engineers, and metallurgists. Often, the materials or components used in the collectors are already mass-produced, such as glass
mirrors.
By the end of 2010, strong competition had emerged and an increasing number of companies had developed industry-level capability to
supply materials such as high-reectivity glass mirrors and manufactured components. Nonetheless, the large evacuated tubes designed
specically for use in trough/oil systems for power generation remain
a specialized component, and only two companies (Schott and Solel)
have been capable of supplying large orders of tubes, with a third
company (Archimedes) now emerging. The trough concentrator itself
comprises know-how in both structures and thermally sagged glass mirrors. Although more companies are now offering new trough designs
and considering alternatives to conventional rear-silvered glass (e.g.,
polymer-based reective lms), the essential technology of concentration remains unchanged. Direct steam generation in troughs is under
demonstration, as is direct heating of molten salt, but these designs are
not yet commercially available. As a result of its successful operational
history, the trough/oil technology comprised most of the CSP installed
capacity in 2010.
Linear Fresnel and central-receiver systems comprise a high level of
know-how, but the essential technology is such that there is the potential for a greater variety of new industry participants. Although only a
couple of companies have historically been involved with central receivers, new players have entered the market over the last few years. There
are also technology developers and projects at the demonstration level
(China, USA, Israel, Australia, Spain). Central-receiver developers are
aiming for higher temperatures, and, in some cases, alternative heat
Direct Solar Energy
transfer uids such as molten salts. The accepted standard to date has
been to use large heliostats, but many of the new entrants are pursuing
much smaller heliostats to gain potential cost reductions through highvolume mass production. The companies now interested in heliostat
development range from optics companies to the automotive industry
looking to diversify. High-temperature steam receivers will benet from
existing knowledge in the boiler industry. Similarly, with linear Fresnel,
a range of new developments are occurring, although not yet as developed as the central-receiver technology.
Dish technology is much more specialized, and most effort presently
has been towards developing the dish/Stirling concept as a commercial
product. Again, the technology can be developed as specialized components through specic industry know-how such as the Stirling engine
mass-produced through the automotive industry.
Within less than 10 years prior to 2010, the CSP industry has gone from
negligible activity to over 2,400 MWe either commissioned or under
construction. A list of new CSP plants and their characteristics can be
found at the IEA SolarPACES web site.3 More than ten different companies are now active in building or preparing for commercial-scale
plants, compared to perhaps only two or three who were in a position to
build a commercial-scale plant three years ago. These companies range
from large organizations with international construction and project
management expertise who have acquired rights to specic technologies, to start-ups based on their own technology developed in-house. In
addition, major independent power producers and energy utilities are
playing a role in the CSP market.
The supply chain does not tend to be limited by raw materials, because
the majority of required materials are bulk commodities such as glass,
steel/aluminium, and concrete. The sudden new demand for the specic
solar salt mixture material for molten-salt storage is claimed to have
impacted supply. At present, evacuated tubes for trough plants can be
produced at a sufcient rate to service several hundred MW per year.
However, expanded capacity can be introduced readily through new factories with an 18-month lead time.
Solar fuel technology is still at an emerging stagethus, there is no
supply chain in place at present for commercial applications. However,
solar fuels will comprise much of the same solar-eld technology being
deployed for other high-temperature CSP systems, with solar fuels
requiring a different receiver/reactor at the focus and different downstream processing and control. Much of the downstream technology,
such as Fischer-Tropsch liquid fuel plants, would come from existing
expertise in the petrochemical industry. The scale of solar fuel demonstration plants is being ramped up to build condence for industry,
which will eventually expand operations.
3
See: www.solarpaces.org.
365
Direct Solar Energy
Chapter 3
Hydrogen has been touted as a future transportation fuel due to its

versatility, pollutant-free end use and storage capability. The key is a
sustainable, CO2-free source of hydrogen such as solar, cost-effective
storage and appropriate distribution infrastructure. The production of
solar hydrogen, in and of itself, does not produce a hydrogen economy
because many factors are needed in the chain. The suggested path to
solar hydrogen is to begin with solar enhancement of existing steam
reforming processes, with a second generation involving solar electricity
and advanced electrolysis, and a third generation using thermolysis or
advanced thermochemical cycles, with many researchers aiming for the
production of fuels from concentrated solar energy, water, and CO2. In
terms of making a transition, solar hydrogen can be mixed with natural gas and transported together in existing pipelines and distribution
networks to customers, thus enhancing the solar portion of the global
energy mix.
Chapter 1. Solar technologies differ in levels of maturity, and although

some applications are already competitive in localized markets, they
generally face one common barrier: the need to achieve cost reductions
(see Section 3.8). Utility-scale CSP and PV systems face different barriers than distributed PV and solar heating and cooling technologies.
Important barriers include: 1) siting, permitting and nancing challenges
to develop land with favourable solar resources for utility-scale projects;
2) lack of access to transmission lines for large projects far from electric
load centres; 3) complex access laws, permitting procedures and fees for
smaller-scale projects; 4) lack of consistent interconnection standards
and time-varying utility rate structures that capture the value of distributed generated electricity; 5) inconsistent standards and certications
and enforcement of these issues; and 6) lack of regulatory structures
that capture environmental and risk mitigation benets across technologies (Denholm et al., 2009).
Steam reforming of natural gas for hydrogen production is a conventional industrial-scale process that produces most of the worlds
hydrogen today, with the heat for the process derived from burning a
signicant proportion of the fossil fuel feedstock. Using concentrated
solar power, instead, as the source of the heat embodies solar energy in
the fuel. The solar steam-reforming of natural gas and other hydrocarbons, and the solar steam-gasication of coal and other carbonaceous
materials yields a high-quality syngas, which is the building block for a
wide variety of synthetic fuels including Fischer-Tropsch-type chemicals,
hydrogen, ammonia and methanol (Steinfeld and Meier, 2004).
Through appropriate policy designs (see Chapter 11), governments have

shown that they can support solar technologies by funding R&D and by
providing incentives to overcome economic barriers. Price-driven instruments (see Section 11.5.2), for example, were popularized after feed-in
tariff (FIT) policies boosted levels of PV deployment in Germany and
Spain. In 2009, various forms of FIT policies were implemented in more
than 50 countries (REN21, 2010) and some designs offer premiums for
building-integrated PV. Quota-driven frameworks such as renewable
portfolio standards (RPS) and government bidding are common in the
USA and China, respectively (IEA, 2009a). Traditional RPS frameworks
are designed to be technology-neutral, and this puts at a disadvantage
many solar applications that are more costly than alternatives such as
wind power. In response, features of RPS frameworks (set-asides and
credits) increasingly are including solar-specic policies, and such programs have led to increasing levels of solar installations (Wiser et al.,
2010). In addition to these regulatory frameworks, scal policies and
nancing mechanisms (e.g., tax credits, soft loans and grants) are often
employed to support the manufacturing of solar goods and to increase
consumer demand (Rickerson et al., 2009). The challenge for solar projects to secure nancing is a critical barrier, especially for developing
technologies in market structures dominated by short-term transactions
and planning.
The solar cracking route refers to the thermal decomposition of natural

gas and other hydrocarbons. Besides H2 and carbon, other compounds
may also be formed, depending on the reaction kinetics and on the
presence of impurities in the raw materials. The thermal decomposition
yields a carbon-rich condensed phase and a hydrogen-rich gas phase.
The carbonaceous solid product can either be sequestered without CO2
release or used as material commodity (carbon black) under less severe
CO2 restraints. It can also be applied as reducing agent in metallurgical
processes. The hydrogen-rich gas mixture can be further processed to
high-purity hydrogen that is not contaminated with oxides of carbon;
thus, it can be used in proton-exchange-membrane fuel cells without
inhibiting platinum electrodes. From the perspective of carbon sequestration, it is easier to separate, handle, transport and store solid carbon
than gaseous CO2. Further, thermal cracking removes and separates
carbon in a single step. The major drawback of thermal cracking is the
energy loss associated with the sequestration of carbon. Thus, solar
cracking may be the preferred option for natural gas and other hydrocarbons with a high H2/C ratio (Steinfeld and Meier, 2004).
3.4.3
Impact of policies4
Direct solar energy technologies support a broad range of applications,

and their deployment is confronted by many of the barriers outlined in
4
Non-technology-specic policy issues are covered in Chapter 11 of this report.
366
Most successful solar policies are tailored to the barriers posed by specic applications. Across technologies, there is a need to offset relatively
high upfront investment costs (Denholm et al., 2009). Yet, in the case
of utility-scale CSP and PV projects, substantial and long-term investments are required at levels that exceed solar applications in distributed
markets. Solar heating and cooling technologies are included in many
policies, yet the characteristics of their applications differ from electricity-generating technologies. Policies based on energy yield rather than
collector surface area are generally preferred for various types of solar
thermal collectors (IEA, 2007). See Section 1.5 for further discussion.
Similar to other renewable sources, there is ongoing discussion about
the merits of existing solar policies to spur innovation and accelerate
deployment using cost-effective measures. Generallyand as discussed
Chapter 3
in Chapter 11the most successful policies are those that send clear,
long-term and consistent signals to the market. In addition to targeted
economic policies, government action through educationally based
schemes (e.g., workshops, workforce training programs and seminars)
and engagement of regulatory organizations are helping to overcome
many of the barriers listed in this section.
3.5
Integration into the broader energy

system5
This section discusses how direct solar energy technologies are part of
the broader energy framework, focusing specically on the following:
low-capacity energy demand; district heating and other thermal loads;
PV generation characteristics and the smoothing effect; and CSP generation characteristics and grid stabilization. Chapter 8 addresses the
broader technical and institutional options for managing the unique
characteristics, production variability, limited predictability and locational dependence of some RE technologies, including solar, as well as
existing experience with and studies associated with the costs of that
integration.
3.5.1
Low-capacity electricity demand
There can be comparative advantages for using solar energy rather than
non-renewable fuels in many developing countries. Within a country, the
advantages can be higher in un-electried rural areas compared to urban
areas. Indeed, solar energy has the advantage, due to being modular, of
being able to provide small and decentralized supplies, as well as large
centralized ones. For more on integrated buildings and households, see
Section 8.3.2.
In a wide range of countries, particularly those that are not oil producers,
solar energy and other forms of RE can be the most appropriate energy
source. If electricity demand exceeds supply, the lack of electricity can
prevent development of many economic sectors. Even in countries with
high solar energy sustainable development potential, RE is often only considered to satisfy high-power requirements such as the industrial sector.
However, large-scale technologies such as CSP are often not available to
them due, for example, to resource conditions or suitable land area availability. In such cases, it is reasonable to keep the electricity generated near
the source to provide high amounts of power to cover industrial needs.
Applications that have low power consumption, such as lighting in rural
areas, can primarily be satised using onsite PVeven if the business plan
for electrication of the area indicates that a grid connection would be
more protable. Furthermore, the criteria to determine the most suitable
technological option for electrifying a rural area should include benets
such as local economic development, exploiting natural resources, creating jobs, reducing the countrys dependence on imports, and protecting
the environment.
5
Non-technology-specic issues related to integration of RE sources in current and

future energy systems are covered in Chapter 8 of this report.
Direct Solar Energy
3.5.2
District heating and other thermal loads
Highly insulated buildings can be heated easily with relatively lowtemperature district-heating systems, where solar energy is ideal, or
quite small quantities of renewable-generated electricity (Boyle, 1996).
A district cooling and heating system (DCS) can provide both cooling
and heating for blocks of buildings. Since the district heating system
already makes the outdoor pipe network available, a district cooling system becomes a viable solution to the cooling demand of buildings. There
are already many DCS installations in the USA, Europe, Japan and other
Asian countries because this system has many advantages compared to
a decentralized cooling system. For example, it takes full advantage of
economy of scale and diversity of cooling demand of different buildings,
reduces noise and structure load, and saves considerable equipment area.
It also allows greater exibility in designing the building by removing the
cooling tower on the roof and chiller plant in the building or on the roof,
and it can provide more reliable and exible services through a specialized professional team in cold-climate areas (Shu et al., 2010). For more
on RE integration in district heating and cooling networks, see Section
8.2.2.3.
In China, Greece, Cyprus and Israel, solar water heaters make a signicant
contribution to supplying residential energy demand. In addition, solar
water heating is widely used for pool heating in Australia and the USA.
In countries where electricity is a major resource for water heating (e.g.,
Australia, Canada and the USA), the impact of numerous solar domestic
water heaters on the operation of the power grid depends on the utilitys load management strategy. For a utility that uses centralized load
switching to manage electric water heater load, the impact is limited to
fuel savings. Without load switching, the installation of many solar water
heaters may have the additional benet of reducing peak demand on the
grid. For a utility that has a summer peak, the time of maximum solar
water heater output corresponds with peak electrical demand, and there is
a capacity benet from load displacement of electric water heaters. Largescale deployment of solar water heating can benet both the customer
and the utility. Another benet to utilities is emissions reduction, because
solar water heating can displace the marginal and polluting generating
plant used to produce peak-load power.
Combining biomass and low-temperature solar thermal energy could provide zero emissions and high capacity factors to areas with less frequent
direct-beam solar irradiance. In the short term, local tradeoffs exist for
areas that have high biomass availability due to increased cloud cover
and rainfall. However, solar technology is more land-efcient for energy
production and greatly reduces the need for biomass growing area and
biomass transport cost. Some optimum ratio of CSP and biomass supply
is likely to exist at each site. Research is being conducted on tower and
dish systems to develop technologiessuch as solar-driven gasication of
biomassthat optimally combine both these renewable resources. In the
longer term, greater interconnectedness across different climate regimes
may provide more stability of supply as a total grid system; this situation
could reduce the need for occasional fuel supply for each individual CSP
system.
367
Direct Solar Energy
3.5.3
Photovoltaic generation characteristics and the

smoothing effect
At a specic location, the generation of electricity by a PV system varies

systematically during a day and a year, but also randomly according to
weather conditions. The variation of PV generation can, in some instances,
have a large impact on voltage and power ow of the local transmission/
distribution system from the early penetration stage, and on supplydemand balance in a total power system operation in the high-penetration
stage (see also Section 8.2.1 for a further discussion of solar electricity
characteristics, and the implications of those characteristics for electricity
market planning, operations, and infrastructure).
Various studies have been published on the impact of supply-demand
balance for a power system with a critical constraint of PV systems integration (Lee and Yamayee, 1981; Chalmers et al., 1985; Chowdhury and
Rahman, 1988; Jewell and Unruh, 1990; Bouzguenda and Rahman, 1993;
Asano et al., 1996). These studies generally conclude that the economic
value of PV systems is signicantly reduced at increasing levels of system
penetration due to the high variability of PV. Todays base-load generation
has a limited ramp ratethe rate at which a generator can change its outputwhich limits the feasible penetration of PV systems. However, these
studies generally lack high-time-resolution PV system output data from
multiple sites. The total electricity generation of numerous PV systems in
a broad area should have less random and fast variationbecause the
generation output variations of numerous PV systems have low correlation and cancel each other in a smoothing effect. The critical impact on
supply-demand balance of power comes from the total generation of the
PV systems within a power system (Piwko et al., 2007, 2010; Ogimoto et
al., 2010).
Some approaches for analyzing the smoothing effect use modelling
and measured data from around the world. Cloud models have been
developed to estimate the smoothing effect of geographic diversity
by considering regions ranging in size from 10 to 100,000 km2 (Jewell
and Ramakumar, 1987) and down to 0.2 km2 (Kern and Russell, 1988).
Using measured data, Kitamura (1999) proposed a set of specications
for describing uctuations, considering three parameters: magnitude,
duration of a transition between clear and cloudy, and speed of the
transition, dened as the ratio of magnitude and duration; he evaluated the smoothing effect in a small area (0.1 km by 0.1 km). A similar
approach, ramp analysis, was proposed by Beyer et al. (1991) and
Schefer (2002).
In a statistical approach, Otani et al. (1997) characterized irradiance
data by the uctuation factor using a high-pass ltered time series of
solar irradiance. Woyte et al. (2001, 2007) analyzed the uctuations of
the instantaneous clearness index by means of a wavelet transform. To
demonstrate the smoothing effect, Otani et al. (1998) demonstrated that
the variability of sub-hourly irradiance even within a small area of 4
km by 4 km can be reduced due to geographic diversity. They analyzed
the non-correlational irradiation/generation characteristics of several PV
systems/sites that are dispersed spatially.
368
Chapter 3
Wiemken et al. (2001) used data from actual PV systems in Germany

to demonstrate that ve-minute ramps in normalized PV power output
at one site may exceed 50%, but that ve-minute ramps in the normalized PV power output from 100 PV systems spread throughout the
country never exceed 5%. Ramachandran et al. (2004) analyzed the
reduction in power output uctuation for spatially dispersed PV systems
and for different time periods, and they proposed a cluster model to
represent very large numbers of small, geographically dispersed PV systems. Results from Curtright and Apt (2008) based on three PV systems
in Arizona indicate that 10-minute step changes in output can exceed
60% of PV capacity at individual sites, but that the maximum of the
aggregate of three sites is reduced. Kawasaki et al. (2006) similarly
analyzed the smoothing effect within a small (4 km by 4 km) network
of irradiance sensors and concluded that the smoothing effect is most
effective during times when the irradiance variability is most severe
particularly days characterized as partly cloudy.
Murata et al. (2009) developed and validated a method for estimating
the variability of power output from PV plants dispersed over a wide
area that is very similar to the methods used for wind by Ilex Energy
Consulting Ltd et al. (2004) and Holttinen (2005). Mills and Wiser (2010)
measured one-minute solar insolation for 23 sites in the USA and characterized the variability of PV with different degrees of geographic
diversity, comparing the variability of PV to the variability of similarly
sited wind. They determined that the relative aggregate variability of PV
plants sited in a dense ten by ten array with 20-km spacing is six times
less than the variability of a single site for variability on time scales
of less than 15 minutes. They also found that for PV and wind plants
similarly sited in a ve by ve grid with 50-km spacing, the variability
of PV is only slightly more than the variability of wind on time scales of
5 to 15 minutes.
Oozeki et al. (2010) quantitatively evaluated the smoothing effect in a
load-dispatch control area in Japan to determine the importance of data
accumulation and analysis. The study also proposed a methodology to
calculate the total PV output from a limited number of measurement
data using Voronoi Tessellation. Marcos et al. (2010) analyzed onesecond data collected throughout a year from six PV systems in Spain,
ranging from 1 to 9.5 MWp, totalling 18 MW. These studies concluded
that over shorter and longer time scales, the level of variability is nearly
identical because the aggregate uctuation of PV systems spread over
the large area depends on the correlation of the uctuation between
PV systems. The correlation of uctuation, in turn, is a function both
of the time scale and distance between PV systems. Variability is less
correlated for PV systems that are further apart and for variability over
shorter time scales.
Currently, however, not enough data on generation characteristics exist
to evaluate the smoothing effect. Data collection from a sufciently
large number of sites (more than 1,000 sites and at distances of 2 to 200
km), periods and time resolution (one minute or less) had just begun
in mid-2010 in several areas in the world. The smoothed generation
characteristics of PV penetration considering area and multiple sites will
Chapter 3
be analyzed precisely after collecting reliable measurement data with

sufcient time resolution and time synchronization. The results will contribute to the economic and reliable integration of PV into the energy
system.
3.5.4
Concentrating solar power generation

characteristics and grid stabilization
In a CSP plant, even without integrated storage, the inherent thermal

mass in the collector system and spinning mass in the turbine tend to
signicantly reduce the impact of rapid solar transients on electrical output, and thus, lead to less impact on the grid (also see Section 8.2.1). By
including integrated thermal storage systems, base-load capacity factors
can be achieved (IEA, 2010b). This and the ability to dispatch power on
demand during peak periods are key characteristics that have motivated
regulators in the Mediterranean region, starting with Spain, to support
large-scale deployment of this technology with tailored FITs. CSP is suitable for large-scale 10- to 300-MWe plants replacing non-renewable
thermal power capacity. With thermal storage or onsite thermal backup
(e.g., fossil or biogas), CSP plants can also produce power at night or
when irradiation is low. CSP plants can reliably deliver rm, scheduled
power while the grid remains stable.
CSP plants may also be integrated with fossil fuel-red plants such as
displacing coal in a coal-red power station or contributing to gasred integrated solar combined-cycle (ISCC) systems. In ISCC power
plants, a solar parabolic trough eld is integrated in a modern gas and
steam power plant; the waste heat boiler is modied and the steam
turbine is oversized to provide additional steam from a solar steam
generator. Better fuel efciency and extended operating hours make
combined solar/fossil power generation much more cost-effective than
separate CSP and combined-cycle plants. However, without including
thermal storage, solar steam could only be supplied for some 2,000 of
the 6,000 to 8,000 combined-cycle operating hours of a plant in a year.
Furthermore, because the solar steam is only feeding the combined-cycle
turbinewhich supplies only one-third of its powerthe maximum
solar share obtainable is under 10%. Nonetheless, this concept is of
special interest for oil- and gas-producing sunbelt countries, where solar
power technologies can be introduced to their fossil-based power market (SolarPACES, 2008).
3.6
Environmental and social impacts6
This section rst discusses the environmental impacts of direct solar

technologies, and then describes potential social impacts. However, an
overall issue identied at the start is the small number of peer-reviewed
studies on impacts, indicating the need for much more work in this area.
6
Direct Solar Energy
3.6.1
Environmental impacts
No consensus exists on the premium, if any, that society should pay for
cleaner energy. However, in recent years, there has been progress in
analyzing environmental damage costs, thanks to several major projects
to evaluate the externalities of energy in the USA and Europe (Gordon,
2001; Bickel and Friedrich, 2005; NEEDS, 2009; NRC, 2010). Solar energy
has been considered desirable because it poses a much smaller environmental burden than non-renewable sources of energy. This argument
has almost always been justied by qualitative appeals, although this
is changing.
Results for damage costs per kilogram of pollutant and per kWh were
presented by the International Solar Energy Society in Gordon (2001).
The results of studies such as NEEDS (2009), summarized in Table 3.3
for PV and in Table 3.4 for CSP, conrm that RE is usually comparatively
benecial, though impacts still exist. In comparison to the gures presented for PV and CSP here, the external costs associated with fossil
generation options, as summarized in Chapter 10.6, are considerably
higher, especially for coal-red generation.
Considering passive solar technology, higher insulation levels provide
many benets, in addition to reducing heating loads and associated
costs (Harvey, 2006). The small rate of heat loss associated with high
levels of insulation, combined with large internal thermal mass, creates
a more comfortable dwelling because temperatures are more uniform.
This can indirectly lead to higher efciency in the equipment supplying the heat. It also permits alternative heating systems that would not
Table 3.3 | Quantiable external costs for photovoltaic, tilted-roof, single-crystalline silicon, retrot, average European conditions; in US2005 cents/kWh (NEEDS, 2009).
2005
2025
2050
Health Impacts
0.17
0.14
0.10
Biodiversity
0.01
0.01
0.01
Crop Yield Losses
0.00
0.00
0.00
Material Damage
0.00
0.00
0.00
Land Use
Total
N/A
0.01
0.01
0.18
0.17
0.12
Table 3.4 | Quantiable external costs for concentrating solar power; in US2005 cents/
kWh (NEEDS, 2009).
2005
2025
2050
Health Impacts
0.65
0.10
0.06
Biodiversity
0.03
0.00
0.00
Crop Yield Losses
0.00
0.00
0.00
Material Damage
0.01
0.00
0.00
Land Use
N/A
N/A
N/A
0.69
0.10
0.06
Total
A comprehensive assessment of social and environmental impacts of all RE sources

covered in this report can be found in Chapter 9.
369
Chapter 5
Hydropower
Arun Kumar (India) and Tormod Schei (Norway)
Lead Authors:
Alfred Ahenkorah (Ghana), Rodolfo Caceres Rodriguez (El Salvador),
Jean-Michel Devernay (France), Marcos Freitas (Brazil), Douglas Hall (USA),
nund Killingtveit (Norway), Zhiyu Liu (China)
Emmanuel Branche (France), John Burkhardt (USA), Stephan Descloux (France),
Garvin Heath (USA), Karin Seelos (Norway)
Review Editors:
Cristobal Diaz Morejon (Cuba) and Thelma Krug (Brazil)

Kumar, A., T. Schei, A. Ahenkorah, R. Caceres Rodriguez, J.-M. Devernay, M. Freitas, D. Hall, . Killingtveit,
Z. Liu, 2011: Hydropower. 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 University Press, Cambridge,
United Kingdom and New York, NY, USA.
437
Chapter 5
signicant effect on the power output. Increased risks of landslides

and glacial lake outbursts, and impacts of increased variability, are
of particular concern to Himalayan countries (Agrawala et al., 2003).
The possibility of accommodating increased intensity of seasonal precipitation by increasing storage capacities may become of particular
importance (Iimi, 2007).
In Europe, by the 2070s, hydropower potential for the whole of Europe
has been estimated to potentially decline by 6%, translated into a 20
to 50% decrease around the Mediterranean, a 15 to 30% increase in
northern and Eastern Europe, and a stable hydropower pattern for western and central Europe (Lehner et al., 2005).
In New Zealand, increased westerly wind speed is very likely to enhance
wind generation and spill over precipitation into major South Island
watersheds, and to increase winter rain in the Waikato catchment.
Warming is virtually certain to increase melting of snow, the ratio of
rainfall to snowfall, and to increase river ows in winter and early
spring. This is very likely to increase hydroelectric generation during the
winter peak demand period, and to reduce demand for storage.
In Latin America, hydropower is the main electrical energy source for
most countries, and the region is vulnerable to large-scale and persistent
rainfall anomalies due to El Nio and La Nia, as observed in Argentina,
Colombia, Brazil, Chile, Peru, Uruguay and Venezuela. A combination of
increased energy demand and droughts caused a virtual breakdown of
hydroelectricity in most of Brazil in 2001 and contributed to a reduction
in gross domestic product (GDP). Glacier retreat is also affecting hydropower generation, as observed in the cities of La Paz and Lima.
In North America, hydropower production is known to be sensitive to
total runoff, to its timing, and to reservoir levels. During the 1990s, for
example, Great Lakes levels fell as a result of a lengthy drought, and in
1999, hydropower production was down signicantly both at Niagara
and Sault St. Marie. For a 2C to 3C warming in the Columbia River
Basin and BC Hydro service areas, the hydroelectric supply under worstcase water conditions for winter peak demand is likely to increase (high
condence). Similarly, Colorado River hydropower yields are likely to
decrease signicantly, as will Great Lakes hydropower. Northern Qubec
hydropower production would be likely to benet from greater precipitation and more open-water conditions, but hydropower plants in
southern Qubec would be likely to be affected by lower water levels.
Consequences of changes in the seasonal distribution of ows and in
the timing of ice formation are uncertain.
Hydropower
Table 5.2 | Power generation capacity in GW and TWh/yr (2005) and estimated changes
(TWh/yr) due to climate change by 2050. Results are based on an analysis using the SRES
A1B scenario in 12 different climate models (Milly et al., 2008), UNEP world regions and
data for the hydropower system in 2005 (US DOE, 2009) as presented in Hamududu and
Killingtveit (2010).
TWh/yr (PJ/yr)
Change by 2050
TWh/yr (PJ/yr)
Africa
22
90 (324)
0.0 (0)
Asia
246
996 (3,586)
2.7 (9.7)
Europe
177
517 (1,861)
-0.8 (-2.9)
North America
161
655 (2,358)
0.3 (1)
South America
119
661 (2,380)
0.3 (1)
Oceania
13
40 (144)
0.0 (0)
TOTAL
737
2931 (10,552)
2.5 (9)
In general the results given in Table 5.2 are consistent with the (mostly
qualitative) results given in previous studies (IPCC, 2007b; Bates et al.,
2008). For Europe, the computed reduction (-0.2%) has the same sign,
but is less than the -6% found by Lehner et al. (2005). One reason could
be that Table 5.2 shows changes by 2050 while Lehner et al. (2005) give
changes by 2070, so a direct comparison is difcult.
It can be concluded that the overall impacts of climate change on the
existing global hydropower generation may be expected to be small, or
even slightly positive. However, results also indicated substantial variations in changes in energy production across regions and even within
countries (Hamududu and Killingtveit, 2010).
Insofar as a future expansion of the hydropower system will occur incrementally in the same general areas/watersheds as the existing system,
these results indicate that climate change impacts globally and averaged across regions may also be small and slightly positive.
Still, uncertainty about future impacts as well as increasing difculty of
future systems operations may pose a challenge that must be addressed
in the planning and development of future HPP (Hamududu et al., 2010).
Indirect effects on water availability for energy purposes may occur if
water demand for other uses such as irrigation and water supply for
households and industry rises due to the climate change. This effect is
difcult to quantify, and it is further discussed in Section 5.10.
5.3
In a recent study (Hamududu and Killingtveit, 2010), the regional and
global changes in hydropower generation for the existing hydropower
system were computed, based on a global assessment of changes in
river ow by 2050 (Milly et al., 2005, 2008) for the SRES A1B scenario
using 12 different climate models. The computation was done at the
country or political region (USA, Canada, Brazil, India, China, Australia)
level, and summed up to regional and global values (see Table 5.2).
Power Generation Capacity (2005)

GW
REGION
Head and also installed capacity (size) are often presented as criteria for
the classication of hydropower plants. The main types of hydropower,
however, are run-of-river, reservoir (storage hydro), pumped storage,
and in-stream technology. Classication by head and classication by
size are discussed in Section 5.3.1. The main types of hydropower are
presented in Section 5.3.2. Maturity of the technology, status and
449
Hydropower
Chapter 5
current trends in technology development, and trends in renovation

and modernization follow in Sections 5.3.3 and 5.3.4 respectively.
5.3.1
Classication by head and size
A classication by head refers to the difference between the upstream

and the downstream water levels. Head determines the water pressure
on the turbines that together with discharge are the most important
parameters for deciding the type of hydraulic turbine to be used.
Generally, for high heads, Pelton turbines are used, whereas Francis
turbines are used to exploit medium heads. For low heads, Kaplan and
Bulb turbines are applied. The classication of what high head and
low head are varies widely from country to country, and no generally
accepted scales are found.
Classication according to size has led to concepts such as small hydro
and large hydro, based on installed capacity measured in MW as the
dening criterion. Small-scale hydropower plants (SHP) are more likely
to be run-of-river facilities than are larger hydropower plants, but reservoir (storage) hydropower stations of all sizes will utilize the same
basic components and technologies. Compared to large-scale hydropower, however, it typically takes less time and effort to construct and
integrate small hydropower schemes into local environments (Egr and
Milewski, 2002). For this reason, the deployment of SHPs is increasing in
many parts of the world, especially in remote areas where other energy
sources are not viable or are not economically attractive.
Nevertheless, there is no worldwide consensus on denitions regarding
size categories (Egr and Milewski, 2002). Various countries or groups
of countries dene small hydro differently. Some examples are given
in Table 5.3. From this it can be inferred that what presently is named
large hydro spans a very wide range of HPPs. IJHD (2010) lists several
more examples of national denitions based on installed capacity.
This broad spectrum in denitions of size categories for hydropower may

be motivated in some cases by national licensing rules (e.g., Norway9)
to determine which authority is responsible for the process or in other
cases by the need to dene eligibility for specic support schemes (e.g.,
US Renewable Portfolio Standards). It clearly illustrates that different
countries have different legal denitions of size categories that match
their local energy and resource management needs.
Regardless, there is no immediate, direct link between installed capacity as a classication criterion and general properties common to all
HPPs above or below that MW limit. Hydropower comes in manifold
project types and is a highly site-specic technology, where each project
is a tailor-made outcome for a particular location within a given river
basin to meet specic needs for energy and water management services.
While run-of-river facilities may tend to be smaller in size, for example,
large numbers of small-scale storage hydropower stations are also in
operation worldwide. Similarly, while larger facilities will tend to have
lower costs on a USD/kW basis due to economies of scale, that tendency will only hold on average. Moreover, one large-scale hydropower
project of 2,000 MW located in a remote area of one river basin might
have fewer negative impacts than the cumulative impacts of 400 5-MW
hydropower projects in many river basins (Egr and Milewski, 2002).
For that reason, even the cumulative relative environmental and social
impacts of large versus small hydropower development remain unclear,
and context dependent.
All in all, classication according to size, while both common and administratively simple, isto a degreearbitrary: general concepts like
small or large hydro are not technically or scientically rigorous indicators of impacts, economics or characteristics (IEA, 2000c). Hydropower
projects cover a continuum in scale, and it may be more useful to evaluate a hydropower project on its sustainability or economic performance
(see Section 5.6 for a discussion of sustainability), thus setting out more
realistic indicators.
Table 5.3 | Small-scale hydropower by installed capacity (MW) as dened by various countries
Country
Small-scale hydro as dened by

installed capacity (MW)
Reference Declaration
Brazil
30
Brazil Government Law No. 9648, of May 27, 1998
Canada
<50
Natural Resources Canada, 2009: canmetenergy-canmetenergie.nrcan-rncan.gc.ca/eng/renewables/

small_hydropower.html
China
50
Jinghe (2005); Wang (2010)
EU Linking Directive
20
EU Linking directive, Directive 2004/101/EC, article 11a, (6)
India
25
Ministry of New and Renewable Energy, 2010: www.mnre.gov.in/
Norway
10
Norwegian Ministry of Petroleum and Energy. Facts 2008. Energy and Water Resources in Norway; p.27
Sweden
1.5
European Small Hydro Association, 2010: www.esha.be/index.php?id=13
USA
5100
US National Hydropower Association. 2010 Report of State Renewable Portfolio Standard Programs (US
RPS)
450
Norwegian Water Resources and Energy Directorate, Water resource act and
regulations, 2001.
Chapter 5
5.3.2
Hydropower
Classication by facility type
HPPs is relatively inexpensive and such facilities have, in general, lower

environmental impacts than similar-sized storage hydropower plants.
Hydropower plants are often classied in three main categories according to operation and type of ow. Run-of-river (RoR), storage (reservoir)
and pumped storage HPPs all vary from the very small to the very large
scale, depending on the hydrology and topography of the watershed. In
addition, there is a fourth category called in-stream technology, which is
a young and less-developed technology.
5.3.2.1
Run-of-River
A RoR HPP draws the energy for electricity production mainly from the
available ow of the river. Such a hydropower plant may include some
short-term storage (hourly, daily), allowing for some adaptations to the
demand prole, but the generation prole will to varying degrees be
dictated by local river ow conditions. As a result, generation depends
on precipitation and runoff and may have substantial daily, monthly or
seasonal variations. When even short-term storage is not included, RoR
HPPs will have generation proles that are even more variable, especially when situated in small rivers or streams that experience widely
varying ows.
5.3.2.2
Storage Hydropower
Hydropower projects with a reservoir are also called storage hydropower since they store water for later consumption. The reservoir
reduces the dependence on the variability of inow. The generating
stations are located at the dam toe or further downstream, connected
to the reservoir through tunnels or pipelines. (Figure 5.6). The type and
design of reservoirs are decided by the landscape and in many parts of
the world are inundated river valleys where the reservoir is an articial
lake. In geographies with mountain plateaus, high-altitude lakes make
up another kind of reservoir that often will retain many of the properties
Dam
Penstock
In a RoR HPP, a portion of the river water might be diverted to a channel

or pipeline (penstock) to convey the water to a hydraulic turbine, which
is connected to an electricity generator (see Figure 5.5). RoR projects
Powerhouse
Switch Yard
Tailrace
Desilting
Tank
Diversion
Weir
Headrace
Forebay
Penstock
Powerhouse
Tailrace
Figure 5.6 | Typical hydropower plant with reservoir.
Intake
Stream
of the original lake. In these types of settings, the generating station is

often connected to the lake serving as reservoir via tunnels coming up
beneath the lake (lake tapping). For example, in Scandinavia, natural
high-altitude lakes are the basis for high pressure systems where the
heads may reach over 1,000 m. One power plant may have tunnels coming from several reservoirs and may also, where opportunities exist, be
connected to neighbouring watersheds or rivers. The design of the HPP
and type of reservoir that can be built is very much dependent on opportunities offered by the topography.
5.3.2.3
Pumped storage
Figure 5.5 | Run-of-river hydropower plant.
may form cascades along a river valley, often with a reservoir-type HPP
in the upper reaches of the valley that allows both to benet from the
cumulative capacity of the various power stations. Installation of RoR
Pumped storage plants are not energy sources, but are instead storage
devices. In such a system, water is pumped from a lower reservoir into
an upper reservoir (Figure 5.7), usually during off-peak hours, while ow
is reversed to generate electricity during the daily peak load period or at
451
Hydropower
Chapter 5
5.3.3
Status and current trends in technology

development
Upper Reservoir
Pumping
Pumped Storage
Power Plant
Generating
Lower
Reservoir
Hydropower is a proven and well-advanced technology based on more

than a century of experiencewith many examples of hydropower
plants built in the 19th century still in operation today. Hydropower
today is an extremely exible power technology with among the best
conversion efciencies of all energy sources (~90%, water to wire) due
to its direct transformation of hydraulic energy to electricity (IEA, 2004).
Still, there is room for further improvements, for example, by improving operation, reducing environmental impacts, adapting to new social
and environmental requirements and by developing more robust and
cost-effective technological solutions. The status and current trends are
presented below, and options and prospects for future technology innovations are discussed in Section 5.7.
5.3.3.1
Figure 5.7 | Typical pumped storage project.
other times of need. Although the losses of the pumping process make
such a plant a net energy consumer overall, the plant is able to provide
large-scale energy storage system benets. In fact, pumped storage is
the largest-capacity form of grid energy storage now readily available
worldwide (see Section 5.5.5).
Efciency
The potential for energy production in a hydropower plant is determined

by the following parameters, which are dependent on the hydrology,
topography and design of the power plant:
The amount of water available;
Water loss due to ood spill, bypass requirements or leakage;
5.3.2.4
In-stream technology using existing facilities
To optimize existing facilities like weirs, barrages, canals or falls, small

turbines or hydrokinetic turbines can be installed for electricity generation. These basically function like a run-of-river scheme, as shown in
Figure 5.8. Hydrokinetic devices being developed to capture energy from
tides and currents may also be deployed inland in both free-owing
rivers and in engineered waterways (see Section 5.7.4).
Spillway
Diversion Canal
Irrigation Canal
Switch Yard
Powerhouse
Tailrace Channel
Figure 5.8 | Typical in-stream hydropower plant using existing facilities.
452
The difference in head between upstream intake and downstream

outlet;
Hydraulic losses in water transport due to friction and velocity
change; and
The efciency in energy conversion of electromechanical equipment.
The total amount of water available at the intake will usually not be
possible to utilize in the turbines because some of the water will be lost
or will not be withdrawn. This loss occurs because of water spill during
high ows when inow exceeds the turbine capacity, because of bypass
releases for environmental ows, and because of leakage.
In the hydropower plant the potential (gravitational) energy in water
is transformed into kinetic energy and then mechanical energy in the
turbine and further to electrical energy in the generator. The energy
transformation process in modern hydropower plants is highly efcient,
usually with well over 90% mechanical efciency in turbines and over
99% in the generator. The inefciency is due to hydraulic loss in the
water circuit (intake, turbine and tailrace), mechanical loss in the turbogenerator group and electrical loss in the generator. Old turbines can
Chapter 5
Hydropower
have lower efciency, and efciency can also be reduced due to wear
and abrasion caused by sediments in the water. The rest of the potential
energy is lost as heat in the water and in the generator.
In addition, some energy losses occur in the headrace section where
water ows from the intake to the turbines, and in the tailrace section
taking water from the turbine back to the river downstream. These losses,
called head loss, reduce the head and hence the energy potential for the
power plant. These losses can be classied either as friction losses or
singular losses. Friction losses depend mainly on water velocity and the
roughness in tunnels, pipelines and penstocks.
Efciency () [%]
The total efciency of a hydropower plant is determined by the sum of

these three loss components. Hydraulic losses can be reduced by increasing the turbine capacity or by increasing the reservoir capacity to get
better regulation of the ow. Head losses can be reduced by increasing the area of headrace and tailrace, by decreasing the roughness in
these and by avoiding too many changes in ow velocity and direction.

The efciency of electromechanical equipment, especially turbines, can
be improved by better design and also by selecting a turbine type with
an efciency prole that is best adapted to the duration curve of the
inow. Different turbine types have quite different efciency proles when
the turbine discharge deviates from the optimal value (see Figure 5.9).
Improvements in turbine design by computational uid dynamics software
and other innovations are discussed in Section 5.7.
5.3.3.2
Tunnelling capacity
In hydropower projects, tunnels in hard and soft rock are often used for
transporting water from the intake to the turbines (headrace), and from the
turbine back to the river, lake or fjord downstream (tailrace). In addition,
tunnels are used for a number of other purposes when the power station
is placed underground, for example for access, power cables, surge shafts
100
Francis
Kaplan
90
80
Pelton
70
Crossow
60
50
Propeller
40
30
20
10
10
20
30
40
50
60
70
80
90
100
110
120
Relative Discharge [%]

Figure 5.9 | Typical efciency curves for different types of hydropower turbines (Vinogg and Elstad, 2003).
453
Hydropower
Chapter 5
and ventilation. Tunnels are increasingly favoured for hydropower construction as a replacement for surface structures like canals and penstocks.
Tunnelling technology has improved greatly due to the introduction
of increasingly efcient equipment, as illustrated by Figure 5.10 (Zare
and Bruland, 2007). Today, the two most important technologies for
hydropower tunnelling are the drill and blast method and the use of
tunnel-boring machines (TBM).
The drill and blast method is the conventional method for tunnel excavation in hard rock. Thanks to the development in tunnelling technology,
excavation costs have been reduced by 25%, or 0.8%/yr, over the past 30
years (see Figure 5.10).
Tunnel Excavation Costs [USD/m]
TBMs excavate the entire cross section in one operation without the use
of explosives. TBMs carry out several successive operations: drilling, support of the ground traversed and construction of the tunnel. The diameter
of tunnels constructed can be from <1 m (micro tunnelling) up to 15 m.
The excavation progress of the tunnel is typically from 30 up to 60 m/day.
3,000
2,500
In addition, for HPPs with reservoirs, their storage capacity can be lled
up by sediments, which requires special technical mitigation measures
or plant design.
Lysne et al. (2003) reported that the effects of sediment-induced wear of
turbines in power plants can be, among others:
Generation loss due to reduction in turbine efciency;

Increase in frequency of repair and maintenance;
Increase in generation losses due to downtime;
Reduction in lifetime of the turbine; and
Reduction in regularity of power generation.
All of these effects are associated with revenue losses and increased
maintenance costs. Several promising concepts for sediment control at
intakes and mechanical removal of sediment from reservoirs and for
settling basins have been developed and practised. A number of authors
(Mahmood, 1987; Morris and Fan, 1997; ICOLD, 1999; Palmieri et al.,
2003; White, 2005) have reported measures to mitigate the sedimentation problems by better management of land use practices in upstream
watersheds to reduce erosion and sediment loading, mechanical removal
of sediment from reservoirs and design of hydraulic machineries aiming
to resist the effect of sediment passing through them.
2,000
5.3.4
Renovation, modernization and upgrading
1,500
1,000
500
0
1975
1979
1983
1988
1995
2005
Figure 5.10 | Developments in tunnelling technology: the trend in excavation costs for a
60 m2 tunnel, in USD2005 per metre (adapted from Zare and Bruland, 2007).
5.3.3.3
Technical challenges related to

sedimentation management
Although sedimentation problems are not found in all rivers (see Section
5.6.1.4), operating a hydropower project in a river with a large sediment
load comes with serious technical challenges.
Specically, increased sediment load in the river water induces wear
on hydraulic machinery and other structures of the hydropower plant.
Deposition of sediments can obstruct intakes, block the ow of water
through the system and also impact the turbines. The sediment-induced
wear of the hydraulic machinery is more serious when there is no room
for storage of sediments.
Renovation, modernization and upgrading (RM&U) of old power stations is often less costly than developing a new power plant, often has
relatively smaller environment and social impacts, and requires less time
for implementation. Capacity additions through RM&U of old power
stations can therefore be attractive. Selective replacement or repair of
identied hydro powerhouse components like turbine runners, generator
windings, excitation systems, governors, control panels or trash cleaning
devices can reduce costs and save time. It can also lead to increased
efciency, peak power and energy availability of the plant (Prabhakar
and Pathariya, 2007). RM&U may allow for restoring or improving
environmental conditions in already-regulated areas. Several national
programmes for RM&U are available. For example, the Research Council
of Norway recently initiated a program with the aim to increase power
production in existing hydropower plants and at the same time improve
environmental conditions.10 The US Department of Energy has been
using a similar approach to new technology development since 1994
when it started the Advanced Hydropower Turbine Systems Program
that emphasized simultaneous improvements in energy and environmental performance (Odeh, 1999; Cada, 2001; Sale et al., 2006a).
Normally the life of hydroelectric power plants is 40 to 80 years.
Electromechanical equipment may need to be upgraded or replaced
after 30 to 40 years, however, while civil structures like dams, tunnels
10 Centre for Environmental Design of Renewable Energy: www.cedren.no/.
454
Chapter 5
Hydropower
etc. usually function longer before they requires renovation. The lifespan of properly maintained hydropower plants can exceed 100 years.
Using modern control and regulatory equipment leads to increased reliability (Prabhakar and Pathariya, 2007). Upgrading hydropower plants
calls for a systematic approach, as a number of hydraulic, mechanical,
electrical and economic factors play a vital role in deciding the course
of action. For techno-economic reasons, it can also be desirable to
consider up-rating (i.e., increasing the size of the hydropower plant)
along with RM&U/life extension. Hydropower generating equipment
with improved performance can also be retrotted, often to accommodate market demands for more exible, peaking modes of operation.
Most of the existing worldwide hydropower equipment in operation
will need to be modernized to some degree by 2030 (SER, 2007).
Refurbished or up-rated hydropower plants also result in incremental
increases in hydropower generation due to more efcient turbines and
generators.
In addition, existing infrastructure without hydropower plants (like
existing barrages, weirs, dams, canal fall structures, water supply
schemes) can also be reworked by adding new hydropower facilities. The majority of the worlds 45,000 large dams were not built
for hydropower purposes, but for irrigation, ood control, navigation
and urban water supply schemes (WCD, 2000). Retrotting these
with turbines may represent a substantial potential, because only

about 25% of large reservoirs are currently used for hydropower
production. For example, from 1997 to 2008 in India, about 500 MW
have been developed on existing facilities. A recent study in the USA
indicated some 20 GW could be installed by adding hydropower
capacity to 2,500 dams that currently have none (UNWWAP, 2006).
5.4
Global and regional status of market and

industry development
5.4.1
Existing generation
In 2008, the generation of electricity from hydroelectric plants was

3,288 TWh (11.8 EJ)11 compared to 1,295 TWh (4.7 EJ) in 1973 (IEA,
2010a), which represented an increase of roughly 25% in this period,
and was mainly a result of increased production in China and Latin
America, which reached 585 TWh (2.1 EJ) and 674 TWh (2.5 EJ),
respectively (Figures 5.11 and 5.12).
Hydropower provides some level of power generation in 159
countries. Five countries make up more than half of the worlds hydropower production: China, Canada, Brazil, the USA and Russia. The
1973
2008
Africa 2.2%
Asia 7.2%
Non-OECD
Europe
2.1%
Middle East
0.3%
Latin
America
7.2%
Asia 25.5%
Non-OECD
Europe
1.5%
Former
USSR 9.4%
Latin America
20.5%
Africa 3%
Former
USSR 7.3%
Middle East
0.3%
OECD 72%
OECD 41.9%
1,295 TWh (4,662 EJ)
3,288 TWh (11,836 EJ)
Figure 5.11 | 1973 and 2008 regional shares of hydropower production (IEA, 2010a).
11 These gures differ slightly from those presented in Chapter 1.
455
Hydropower
Chapter 5
possibility of drawing on the hydroelectric resource was important

for the introduction and consolidation of the main electro-intensive
sectors on which the industrialization process in these countries was
based during a considerable part of the 20th century.
Rest of World;
1007; 31%
Sweden; 69; 2%
Venezuela; 87; 3%
PR of China;
585; 18%
Japan; 83; 2%
India; 114; 3%
Despite the signicant growth in hydroelectric production, the percentage share of hydroelectricity on a global basis has dropped during the
last three decades (1973 to 2008), from 21 to 16%. This is because
electricity demand and the deployment of other energy technologies
have increased more rapidly than hydropower generating capacity.
Norway; 141; 4%
Canada;
383; 12%
Brasil;
370; 11%
United
States;
282; 9%
Russia; 167; 5%
Figure 5.12 | Hydropower generation in 2008 by country, indicating total generation

(TWh) and respective global share (IEA, 2010a).
importance of hydroelectricity in the electricity mix of these countries

is, however, different (Table 5.4). On the one hand, Brazil and Canada
are heavily dependent on this source, with a percentage share of
total domestic electricity generation of 83.9% and 59%, respectively,
whereas in Russia the share is 19.0% and in China 15.5%.
China, Canada, Brazil and the USA together account for over 46%
of the production (TWh/EJ) of hydroelectricity in the world and are
also the four largest in terms of installed capacity (GW) (IEA, 2010a).
Figure 5.12 shows hydropower generation by country. It is noteworthy that 5 out of the 10 major producers of hydroelectricity are
among the worlds most industrialized countries: Canada, the USA,
Norway, Japan and Sweden. This is no coincidence, given that the
5.4.2
The hydropower industry
In developed markets such as the Europe, the USA, Canada, Norway

and Japan, where many hydropower plants were built 30 to 60 years
ago, the hydropower industry is focused on re-licensing and renovationas well as on adding new hydropower generation to existing dams.
In emerging markets such as China, Brazil, Ethiopia, India, Malaysia,
Iran, Laos, Turkey, Venezuela, Ecuador and Vietnam, utilities and private
developers are pursuing large-scalenew hydropower construction (116
GW of capacity is under construction; IJHD, 2010). Canada is still on the
list of the topve hydropower markets for new installations worldwide.
Orders for hydropower equipment werelower in 2009 and 2010 compared tothe peaks in 2007 and 2008, thoughthe general high level after
2006, when the hydropower market almost doubled, is anticipated to
continue for the near future.With increasing policy support of governments for new hydropower (see Sections 5.4.3 and 5.10.3) construction,
hydropowerindustrial activity is expected to be higher in the coming
years compared to theaverage since 2000 (IJHD, 2010).As hydropower
and its industry are mature, it is expected that the industry will be able to
meet the demand that materializes (see Section 5.9). In 2008, the hydropower industry installed more than 40 GW of new capacity worldwide
(IJHD, 2010), with 31 GW added in 2009 (REN21, 2010; see Chapter 1).
Table 5.4 | Major hydroelectricity producer countries with total installed capacity and percentage of hydropower generation in the electricity mix. Source: IJHD (2010).
Country
Installed Capacity (GW)
China
Country Based on Top 10 Producers
Percent of Hydropower in Total

Domestic Electricity Generation (%)
200
Norway
Brazil
84
Brazil
USA
78.2
Venezuela
73.4
Canada
74.4
Canada
59.0
Russia
49.5
Sweden
48.8
India
38
Russia
19.0
Norway
29.6
India
17.5
Japan
27.5
China
15.5
France
21
Italy
14.0
Italy
20
Rest of the world
301.6
World
926.1
Note: 1. Excluding countries with no hydropower production.
456
99
83.9
France
8.0
Rest of the world1
14.3
World
15.9
Chapter 5
5.4.3
Impact of policies12
Hydropower infrastructure development is closely linked to national,

regional and global development policies. Beyond its role in contributing
to a secure energy supplysecurity and reducing a countrys dependence
on fossil fuels, hydropower offers opportunities for poverty alleviation and sustainable development. Hydropower also can contribute to
regional cooperation, as good practice in managing water resources
requires a river basin approach regardless of national borders (see also
Section 5.10). In addition, multipurpose hydropower can strengthen a
countrys ability to adapt to climate change-induced hydrological variability (World Bank, 2009).
The main challenges for hydropower development are linked to a number
of associated risks such as poor identication and management of environmental and social impacts, insufcient hydrological data, unexpected
adverse geological conditions, lack of comprehensive river basin planning, shortage of nancing, scarcity of local skilled human resources and
lack of regional collaboration. These challenges can be and are being
addressed to varying degrees at the policy level by a number of governments, international nancing institutions, professional associations and
nongovernmental organizations (NGOs). Examples of policy initiatives
dealing with the various challenges can be found in Sections 5.6.2 and
5.10.
Challenges posed by various barriers can be addressed and met by
public policies, bearing in mind the need for an appropriate environment for investment, a stable regulatory framework and incentives for
research and technological development (Freitas and Soito, 2009; see
Chapter 11). A variety of policies have been enacted in individual countries to support certain forms and types of hydropower, as highlighted
generally in Chapter 11. More broadly, in addition to country-specic
policies, several larger policy issues have been identied as particularly
important for the development of hydropower, including carbon markets, nancing, administration and licensing procedures, and size-based
classication schemes.
5.4.3.1
International carbon markets
As with other carbon reduction technologies, carbon credits can benet

hydropower projects by bringing additional funding and thus helping
to reduce project risk and thereby secure nancing. Though the Clean
Development Mechanism (CDM) is not unique to hydropower, hydropower projects are one of the largest contributors to the CDM and Joint
Implementation (JI) mechanisms and therefore to existing carbon credit
markets. In part, this is due to the fact that new hydropower development is targeted towards developing countries that are in need of
Hydropower
investment capital, and international carbon markets offer one possible route to that capital. Out of the 2,062 projects registered by the
CDM Executive Board (EB) by 1 March 2010, 562 were hydropower
projects. When considering the predicted volumes of Certied Emission
Reductions to be delivered, registered hydropower projects are expected
to avoid more than 50 Mt of carbon dioxide (CO2) emissions per year
by 2012. China, India, Brazil and Mexico represent roughly 75% of the
hosted projects.
5.4.3.2
Project nancing
Hydropower projects can often deliver electricity at comparatively

low costs relative to existing market energy prices (see Section 5.8).
Nonetheless, many otherwise economically feasible hydropower projects
are nancially challenging because high upfront costs are often a deterrent to investment. Related to this, hydropower projects tend to have
lengthy lead times for planning, permitting and construction, increasing development risk and delaying revenue generation. A key challenge,
then, is to create sufcient private sector condence in hydropower
investment, especially prior to project permitting. Deployment policies
of the types described in Chapter 11 are being used in some countries
to encourage investment. Also, in developing regions such as Africa,
interconnection between countries and the formation of larger energy
markets is helping to build investor condence by reducing the risk of a
monopsony buyer. Feasibility and impact assessments carried out by the
public sector, prior to developer tendering, can also help ensure greater
private sector interest in hydropower development (WEC, 2007; Taylor,
2008). Nonetheless, the development of appropriate nancing models
that consider the uncertainty imposed by long planning and regulatory
processes, and nding the optimum roles for the public and private sectors, remain key challenges for hydropower development.
5.4.3.3
Administrative and licensing process
Hydropower is often regarded as a public resource (Sternberg, 2008),

emphasized by the operating life of a reservoir that may be more
than 100 years. Legal frameworks vary from country to country, however, including practices in the award and structuring of concessions,
for instance, regarding concession periods, royalties, water rights etc.
Environmental licensing procedures also vary greatly. With growing
involvement of the private sector in what was previously managed by
public sector, contractual arrangements surrounding hydropower have
become increasingly complex. There are now more parties involved and
much greater commercial accountability, with a strong awareness of
environmental and social indicators and licensing processes. Clearly, the
policies and procedures established by governments in granting licenses
and concessions will impact hydropower development outcomes.
12 Non-technology-specic policy issues are covered in Chapter 11 of this report.
457
Hydropower
5.4.3.4
Classication by size
Finally, many governments and international bodies have relied upon

various distinctions between small and large hydro, as dened by
installed capacity (MW), in establishing the eligibility of hydropower
plants for certain programs. While it is well known that large-scale
HPPs can create conicts and concerns (WCD, 2000), the environmental
and social impacts of a HPP cannot be deduced by size in itself, even if
increasing the physical size may increase the overall impacts of a specic HPP (Egr and Milewski, 2002; Sternberg, 2008). Despite their lack
of robustness (see Section 5.3.1), these classications have had signicant policy and nancing consequences (Egr and Milewski, 2002).
In the UK Renewables Obligation,13 eligible hydropower plants must
be below 20 MW in size. Likewise, in several countries, feed-in tariffs
are targeted only towards smaller projects. For example, in France, only
projects with an installed capacity not exceeding 12MW are eligible,14
and in Germany, a 5 MW maximum capacity has been established.15
In India, projects below 5 and 25 MW in capacity obtain promotional
support that is unavailable to projects of larger sizes. Similar approaches
exist in many developed and developing countries around the world, for
example, in Indonesia.16 Because project size is neither a perfect indicator of environmental and social impact nor of the nancial need of a
project for addition policy support, these categorizations may, at times,
impede the development of socially benecial projects.
Similar concerns have been raised with respect to international and
regional climate policy. Though hydropower is recognized as a contributor to reducing GHG emissions and is included in the Kyoto Protocols
exible mechanisms, those mechanisms differentiate HPPs depending
on size and type. The United Nations Framework Convention on Climate
Change (UNFCCC) CDM EB, for example, has established that storage
hydropower projects are to follow the power density indicator (PDI),
W/m2 (installed capacity/reservoir area), to be eligible for CDM credits.
The PDI indicates tentative GHG emissions from reservoirs. The CDM
Executive Board stated (February 2006) that Hydroelectric power plants
with power densities greater than 4 W/m2 but less than or equal to 10
W/m2 can use the currently approved methodologies, with an emission
factor of 90 g CO2eq/kWh for projects with reservoir emissions, while
less than or equal to 4 W/m2 cannot use current methodologies.
There is little link, however, between installed capacity, the area of a
reservoir and the various biogeochemical processes active in a reservoir.
Hypothetically, two identical storage HPPs would, according to the PDI,
have the same emissions independent of climate zones or of inundated
Chapter 5
biomass and carbon uxes (see Section 5.6.3). As such, the PDI rule may
inadvertently impede the development of socially benecial hydropower
projects, while at the same time supporting less benecial projects. The
European Emission Trading Scheme and related trading markets similarly treat small- and large-scale hydropower stations differently.17
5.5
Hydropowers large capacity range, exibility, storage capability when

coupled with a reservoir, and ability to operate in a stand-alone mode or
in grids of all sizes enables hydropower to deliver a broad range of services. Hydropowers various roles in and services to the energy system
are discussed below (see also Chapter 8).
5.5.1
14 Dcret n2000-1196, Decree on capacity limits for different categories of systems for
the generation of electricity from renewable sources that are eligible for the feed-in
tariff: www.legifrance.gouv.fr.
15 EEG, 2009 - Act on Granting Priority to Renewable Energy and Mineral Sources:
bundesrecht.juris.de/eeg_2009/.
Grid-independent applications
Hydropower can be delivered through national and regional interconnected electric grids, through local mini-grids and isolated grids, and can
also serve individual customers through captive plants. Water mills in
England, Nepal, India and elsewhere, which are used for grinding cereals, for lifting water and for powering machinery, are early testimonies
of hydropower being used as captive power in mechanical and electrical
form. The tea and coffee plantation industries as well as small islands
and developing states have used and still make use of hydropower to
meet energy needs in isolated areas.
Captive power plants (CPPs) are dened here as plants set up by any
person or group of persons to generate electricity primarily for the
person or the groups members (Indian Electricity Act, 2003). CPPs are
often found in decentralized isolated systems and are generally built by
private interests for their own electricity needs. In deregulated electricity
markets that allow open access to the grid, hydropower plants are also
sometimes installedfor captive purposes by energy-intensive industries
such as aluminium smelters, pulp and paper mills, mines and cement factories in order to weather short-term market uncertainties and volatility
(Shukla et al., 2004). For governments of emerging economies such as
India facing shortages of electricity, CPPs are also a means to cope with
unreliable power supply systems and higher industrial tariffs by encouraging decentralized generation and private participation (Shukla et al.,
2004).
5.5.2
13 The Renewables Obligation Order 2006, No. 1004 (ROO 2006): www.statutelaw.
gov.uk.
Integration into broader energy systems
Rural electrication
According to the International Energy Agency (IEA, 2010c), 1.4 billion

people have no access to electricity (see Section 9.3.2). Related to the
discussion in Section 5.5.1, small-scale hydropower (SHP) can sometimes be an economically viable supply source in these circumstances,
as SHP can provide a decentralized electricity supply in those rural areas
17 Directive 2004/101/E, C article 11a(6), www.eur-lex.europa.eu.
16 Regulation of the Minister of Energy and Mineral Resources, No.31, 2009.
458
Chapter 5
that have adequate hydropower technical potential (Egr and Milewski,

2002). In fact, SHPs already play an important role in the economic
development of some remote rural areas. Small-scale hydropower-based
rural electrication in China has been one of the most successful examples,
where over 45,000 small hydropower plants totalling 55 GW have been
built that are producing 160 TWh (0.58 EJ)annually. Though many of
these plants are used in centralized electricity networks, SHPs constitute one-third of Chinas total hydropower capacity and are providing
services to over 300 million people (Liu and Hu, 2010). More generally,
SHP is found in isolated grids as well as in off-grid and central-grid settings. As 75% of costs are site-specic, proper site selection is a key
challenge. Additionally, in isolated grid systems, natural seasonal ow
variations might require that hydropower plants be combined with other
generation sources in order to ensure continuous supply during dry periods (World Bank, 2008) and may have excess production during wet
seasons; such factors need to be considered in the planning process
(Sundqvist and Wrlind, 2006).
In general, SHPs
Are often but certainly not always RoR schemes;
Can use existing infrastructure such as dams or irrigation channels;
Are located close to villages to avoid expensive high-voltage distribution equipment;
Can use pumps as turbines and motors as generators for a turbine/
generator set; and
Have a high level of local content both in terms of materials and
work force during the construction period and local materials for the
civil works.
A recent example from western Canada18 shows that SHP might also be
a solution for remote communities in developed countries by replacing
fossil-red diesel generation with hydropower generation.
All in all, the development of SHP for rural areas involves environmental, social, technical and economic considerations. Local management,
ownership and community participation, technology transfer and
capacity building are basic issues for sustainable SHP plants in such
circumstances.
5.5.3
Power system services provided by hydropower
Hydroelectric generation differs from thermal generation in that the

quantity of fuel (i.e., water) that is available at any given time is determined by river ows leading to the hydroelectric plant. Run-of-river
HPPs lack a reservoir to store large quantities of water, though large
RoR HPPs may have some limited ability to regulate river ow. Storage
18 Natural Resources Canada. 2009. Isolated-grid case study: the Hluey Lake project
in British Columbia: www.retscreen.net/ang/case_studies_2900kw_isolated_grid_
internal_load_canada.php.
Hydropower
hydropower, on the other hand, can largely decouple the timing of

hydropower generation and variable river ows. For large storage reservoirs, the storage may be sufcient to buffer seasonal or multi-seasonal
changes in river ows, whereas for smaller reservoirs the storage may
buffer river ows on a daily or weekly basis.
With a very large reservoir relative to the size of the hydropower plant
(or very consistent river ows), HPPs can generate power at a nearconstant level throughout the year (i.e., operate as a base-load plant).
Alternatively, in the case that the hydropower capacity far exceeds
the amount of reservoir storage, the hydropower plant is sometimes
referred to as energy-limited. An energy-limited hydropower plant would
exhaust its fuel supply by consistently operating at its rated capacity
throughout the year. In this case, the use of reservoir storage allows
hydropower generation to occur at times that are most valuable from
the perspective of the power system rather than at times dictated solely
by river ows. Since electrical demand varies during the day and night,
during the week and seasonally, storage hydropower generation can
be timed to coincide with times where the power system needs are the
greatest. In part, these times will occur during periods of peak electrical
demand. Operating hydropower plants in a way that generates power
during times of high demand is referred to as peaking operation (in contrast to base-load). Even with storage, however, hydropower generation
will still be limited by the size of the storage, the rated electrical capacity
of the hydropower plant, and downstream ow constraints for irrigation, recreation or environmental uses of the river ows. Hydropower
peaking may, if the outlet is directed to a river, lead to rapid uctuations
in river ow, water-covered area, depth and velocity. In turn this may,
depending on local conditions, lead to negative impacts in the river (see
Section 5.6.1.5) unless properly managed.
Hydropower generation that consistently occurs during periods with
high system demand can offset the need for thermal generation to
meet that same demand. The ratio of the amount of demand that can
be reliably met by adding hydropower to the nameplate capacity of the
hydropower plant is called the capacity credit. Even RoR hydropower
that consistently has river ows during periods of high demand can
earn a high capacity credit, while adding reservoir storage can increase
the capacity credit to levels comparable to thermal power plants (see
Section 8.2.1.2).
In addition to providing energy and capacity to meet electrical demand,
hydropower generation often has several characteristics that enable it
to provide other services to reliably operate power systems. Because
hydropower plants utilize gravity instead of combustion to generate
electricity, hydropower plants are often less susceptible to the sudden
loss of generation than is thermal generation. Hydropower plants also
offer operating exibility in that they can start generating electricity with
very short notice and low start-up costs, provide rapid changes in generation, and have a wide range of generation levels over which power
can be generated efciently (i.e., high part-load efciency) (Haldane and
Blackstone, 1955; Altinbilek et al., 2007). The ability to rapidly change
459
Hydropower
output in response to system needs without suffering large decreases in

efciency makes hydropower plants well suited to providing the balancing services called regulation and load-following. RoR HPPs operated
in cascades in unison with storage hydropower in upstream reaches
may similarly contribute to the overall regulating and balancing ability
of a eet of HPPs. With the right equipment and operating procedures,
hydropower can also provide the ability to restore a power station to
operation without relying on the electric power transmission network
(i.e., black start capability) (Knight, 2001).
Overall, with its important load-following and balancing capabilities,
peaking capacity and power quality attributes, hydropower can play a
signicant role in ensuring reliable electricity service (US Department of
the Interior, 2005).
5.5.4
Hydropower support of other generation

including renewable energy
Electricity systems worldwide rely upon widely varying amounts of hydropower today. In this range of hydropower capabilities, electric system
operators have developed economic dispatch methodologies that take
into account the unique role of hydropower, including coordinating the
operation of hydropower plants with other types of generating units. In
particular, many thermal power plants (coal, gas or liquid fuel, or nuclear
energy) require considerable lead times (often four hours for gas turbines
and over eight hours for steam turbines) before they attain an optimum
thermal efciency at which point fuel consumption and emissions per
unit output are minimum. In an integrated system, the considerable exibility provided by storage HPPs can be used to reduce the frequency
of start ups and shut downs of thermal plants; to maintain a balance
between supply and demand under changing demand or supply patterns and thereby reduce the load-following burden on thermal plants;
and to increase the amount of time that thermal units are operated at
their maximum thermal efciency. In some regions, for instance, hydroelectric power plants are used to follow varying peak load demands
while nuclear or fossil fuel power plants are operated as base-load units.
Chapter 5
hydropower being a net energy consumer. A traditional storage HPP may

also be retrotted with pump technologies to combine the properties
of storage and pump storage HPPs (SRU, 2010). The use and benet
of pumped storage hydropower in the power system will depend on
the overall mix of existing generating plants and the architecture of
the transmission system. Pumped storage represents about 2.2% of all
generation capacity in the USA, 10.2 % in Japan and 18.7 % in Austria
(Deane et al., 2010). Various technologies for storing electricity in the
grid are compared by Vennemann et al. (2010) in Figure 5.13 for selected
large storage sites in different parts of the world.
In addition to hydropower supporting fossil and nuclear generation
technologies, hydropower can also help reduce the challenges of integrating variable renewable resources. In Denmark, for example, the high
level of variable wind (>20% of the annual energy demand) is managed
in part through strong interconnections (1 GW) to Norway, where there
is substantial storage hydropower (Nordel, 2008). More interconnectors
to Europe may further support increasing the share of wind power in
Denmark and Germany (SRU, 2010; see also Section 11.6.5). From a
technical viewpoint, Norway alone has a long-term potential to establish pumped storage facilities in the 10 to 25 GW range, enabling energy
storage over periods from hours to several weeks in existing reservoirs,
and more or less doubling the present installed capacity of 29 GW (IEAENARD, 2010).
Increasing variable generation will also increase the amount of balancing services, including regulation and load following, required by the
power system (e.g., Holttinen et al., 2009). In regions with new and
existing hydropower facilities, providing these services with hydropower
may avoid the need to rely on increased part-load and cycling of thermal
plants to provide these services. Similarly, in systems with high shares
of variable renewable resources that provide substantial amounts of
energy but limited capacity, the potential for a high capacity credit of
hydropower can be used to meet peak demand rather than requiring
peaking thermal plants.
5.5.5
Pumped hydropower storage can further increase the support of other
resources. In cases with pumped hydropower storage, pumps can use
the output from thermal plants during times that they would otherwise operate less efciently at part load or be shut down (i.e., low load
periods). The pumped storage plant then keeps water in reserve for generating power during peak period demands. Pumped storage has much
the same ability as storage HPPs to provide balancing and regulation
services.
Pumped storage hydropower is usually not a source for energy, however.
The hydraulic, mechanical and electrical efciencies of pumped storage
determine the overall cycle efciency, ranging from 65 to 80% (Egr
and Milewski, 2002). If the upstream pumping reservoir is also used as
a traditional reservoir the inow from the watershed may balance out
the energy loss caused by pumping. If not, net losses lead to pumped
460
Reliability and interconnection needs

for hydropower
Though hydropower has the potential to offer signicant power system

services in addition to energy and capacity, interconnecting and reliably
utilizing hydropower plants may also require changes to power systems.
The interconnection of hydropower to the power system requires adequate transmission capacity from hydropower plants to demand centres.
Adding new hydropower plants has in the past required network investments to extend the transmission network (see Section 8.2.1.3). Without
adequate transmission capacity, hydropower plant operation can be
constrained such that the services offered by the hydropower plant are
less than what it could offer in an unconstrained system.
Aside from network expansion, changes in the river ow between
a dry year and a wet year can be a signicant concern for ensuring
Hydropower
1,000,000
Pumped Storage Plants (PSP)

100,000
PSP, Limberg I, II, III, AT, 2018
PSP, Linth Limmern, CH, 2015
PSP, Bath County, US, 1985
Compressed Air/Thermal
10,000
PSP, Goldisthal, DE, 2003
1,000
Molten Salt, Andasol I,

ES, 2008
Batteries
PSP, Dinorwig, GB, 1984

PSP, Vianden, LU, 1964
CAES, McIntosh, US, 1991
1 month
PSP, Atdorf, DE, 2018
AA-CAES, RWE/GE,
Project
CAES, Huntorf, D, 1978

NaS, Rokkasho, JP, 2008
3d
100
32 h
In Operation
7d
Lead Acid, Chino, US, 1988
16 h
Pumped Storage
Compressed Air
8h
10
4h
Under Construction/
Project
Storage Capacity per Cycle [MWh]
Chapter 5
Batteries
VRB, Tomamae, JP, 2000
Thermal Storage
2h
NiCd, Fairbanks, US, 2003
1h
1
1
10
100
1,000
10,000
Generation Power [MW]

Figure 5.13 | Storage and installed capacity of selected large electricity storage sites (Vennemann et al., 2010).
Note: PSP = Pumped storage plants; CAES = compressed air energy storage, AA-CAES = advanced adiabatic compressed air energy storage; Batteries: NaS = sodium-sulphur, NiCd =
nickel cadmium, VRB = vanadium redox battery.
that adequate total annual energy demand can be met. Strong interconnections between diverse hydropower resources or between
hydro-dominated and thermal-dominated power systems have been
used in existing systems to ensure adequate energy generation (see
Section 8.2.1.3). In the future, interconnection to other renewable
resources could also ensure adequate energy. Wind and direct solar
power, for instance, can be used to reduce demands on hydropower,
either by allowing dams to save their water for later release in peak periods or letting storage or pumped storage HPPs consume excess energy
produced in off-peak hours.
5.6
Like all energy and water management options, hydropower projects

have negative and positive environmental and social impacts. On the
19 A comprehensive assessment of social and environmental impacts of all RE sources
environmental side, hydropower may have a signicant environmental

footprint at local and regional levels but offers advantages at the macroecological level. With respect to social impacts, hydropower projects may
entail the relocation of communities living within or nearby the reservoir
or the construction sites, compensation for downstream communities,
public health issues etc. A properly designed hydropower project may,
however, be a driving force for socioeconomic development (see Box
5.1), though a critical question remains about how these benets are
shared.
Because each hydropower plant is uniquely designed to t the sitespecic characteristics of a given geographical site and the surrounding
society and environment, the magnitude of environmental and social
impacts as well as the extent of their positive and negative effects is
highly site dependent. Though the size of a HPP is not, alone, a relevant criterion to predict environmental performance, many impacts are
related to the impoundment and existence of a reservoir, and therefore
do not apply to all HPP types (see Table 5.5). Section 5.6.1 summarizes
461
Chapter 7
Wind Energy
Ryan Wiser (USA), Zhenbin Yang (China)
Lead Authors:
Maureen Hand (USA), Olav Hohmeyer (Germany), David Ineld (United Kingdom),
Peter H. Jensen (Denmark), Vladimir Nikolaev (Russia), Mark OMalley (Ireland),
Graham Sinden (United Kingdom/Australia), Arthouros Zervos (Greece)
Nam Darghouth (USA), Dennis Elliott (USA), Garvin Heath (USA), Ben Hoen (USA),
Hannele Holttinen (Finland), Jason Jonkman (USA), Andrew Mills (USA),
Patrick Moriarty (USA), Sara Pryor (USA), Scott Schreck (USA), Charles Smith (USA)
Review Editors:
Christian Kjaer (Belgium/Denmark) and Fatemeh Rahimzadeh (Iran)

Wiser, R., Z. Yang, M. Hand, O. Hohmeyer, D. Ineld, P. H. Jensen, V. Nikolaev, M. OMalley, G. Sinden,
A. Zervos, 2011: Wind Energy. 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 University Press, Cambridge, United
Kingdom and New York, NY, USA.
535
Wind Energy
Chapter 7
loading for offshore plants, and changes in sea ice and/or permafrost
conditions may also inuence access for performing wind power plant
O&M (Laakso et al., 2003). One study focusing on northern Europe
found substantial declines in sea ice under reasonable climate change
scenarios (Claussen et al., 2007). Other meteorological drivers of turbine
loading may also be inuenced by climate change but are likely to be
secondary in comparison to changes in resource magnitude, weather
extremes, and icing issues (Pryor and Barthelmie, 2010).
Power in
Wind
Rated Power
Power Captured
7.3
Modern, commercial grid-connected wind turbines have evolved from

small, simple machines to large, highly sophisticated devices. Scientic
and engineering expertise and advances, as well as improved computational tools, design standards, manufacturing methods, and O&M
procedures, have all supported these technology developments. As a
result, typical wind turbine nameplate capacity ratings have increased
dramatically since the 1980s (from roughly 75 kW to 1.5 MW and larger),
while the cost of wind energy has substantially declined. Onshore wind
energy technology is already being manufactured and deployed on a
commercial basis. Nonetheless, additional R&D advances are anticipated, and are expected to further reduce the cost of wind energy while
enhancing system and component performance and reliability. Offshore
wind energy technology is still developing, with greater opportunities
for additional advancement.
This section summarizes the historical development and current technology status of large grid-connected on- and offshore wind turbines (7.3.1),
discusses international wind energy technology standards (7.3.2), and
reviews power conversion and related grid connection issues (7.3.3); a
later section (7.7) describes opportunities for further technical advances.
7.3.1
Technology development and status
7.3.1.1
Basic design principles
Generating electricity from the wind requires that the kinetic energy of
moving air be converted to mechanical and then electrical energy, thus
the engineering challenge for the wind energy industry is to design costeffective wind turbines and power plants to perform this conversion.
The amount of kinetic energy in the wind that is theoretically available
for extraction increases with the cube of wind speed. However, a turbine
only captures a portion of that available energy (see Figure 7.3).
550
Rotor RPM
Power
Additional research on the possible impact of climate change on the

size, geographic distribution and variability of the wind resource is
warranted, as is research on the possible impact of climate change on
extreme weather events and therefore wind turbine operating environments. Overall, however, research to date suggests that these impacts
are unlikely to be of a magnitude that will greatly impact the global
potential of wind energy deployment.
Wind Speed
Cut-In Speed
Region I
Region II
Rated Speed
Cut-Out Speed
Region III
Figure 7.3 | Conceptual power curve for a modern variable-speed wind turbine (US
DOE, 2008).
Specically, modern large wind turbines typically employ rotors that

start extracting energy from the wind at speeds of roughly 3 to 4 m/s
(cut-in speed). The Lanchester-Betz limit provides a theoretical upper
limit (59.3%) on the amount of energy that can be extracted (Burton
et al., 2001). A wind turbine increases power production with wind
speed until it reaches its rated power level, often corresponding to a
wind speed of 11 to 15 m/s. At still-higher wind speeds, control systems limit power output to prevent overloading the wind turbine, either
through stall control, pitching the blades, or a combination of both
(Burton et al., 2001). Most turbines then stop producing energy at wind
speeds of approximately 20 to 25 m/s (cut-out speed) to limit loads on
the rotor and prevent damage to the turbines structural components.
Wind turbine design has centred on maximizing energy capture over
the range of wind speeds experienced by wind turbines, while seeking
to minimize the cost of wind energy. As described generally in Burton et
al. (2001), increased generator capacity leads to greater energy capture
when the turbine is operating at rated power (Region III). Larger rotor
diameters for a given generator capacity, meanwhile, as well as aerodynamic design improvements, yield greater energy capture at lower
wind speeds (Region II), reducing the wind speed at which rated power
is achieved. Variable speed operation allows energy extraction at peak
efciency over a wider range of wind speeds (Region II). Finally, because
the average wind speed at a given location varies with the height above
ground level, taller towers typically lead to increased energy capture.
To minimize cost, wind turbine design is also motivated by a desire
to reduce materials usage while continuing to increase turbine size,
increase component and system reliability, and improve wind power
plant operations. A system-level design and analysis approach is necessary to optimize wind turbine technology, power plant installation and
O&M procedures for individual turbines and entire wind power plants.
Moreover, optimizing turbine and power plant design for specic site
Chapter 7
Wind Energy
conditions has become common as wind turbines, wind power plants

and the wind energy market have all increased in size; site-specic conditions that can impact turbine and plant design include geographic
and temporal variations in wind speed, site topography and access,
interactions among individual wind turbines due to wake effects, and
integration into the larger electricity system (Burton et al., 2001). Wind
turbine and power plant design also impacts and is impacted by noise,
visual, environmental and public acceptance issues (see Section 7.6).
7.3.1.2
Onshore wind energy technology
In the 1970s and 1980s, a variety of onshore wind turbine congurations

were investigated, including both horizontal and vertical axis designs
(see Figure 7.4). Gradually, the horizontal axis design came to dominate,
although congurations varied, in particular the number of blades and
whether those blades were oriented upwind or downwind of the tower
(EWEA, 2009). After a period of further consolidation, turbine designs
largely centred (with some notable exceptions) around the three-blade,
upwind rotor; locating the turbine blades upwind of the tower prevents
the tower from blocking wind ow onto the blades and producing extra
aerodynamic noise and loading, while three-bladed machines typically
have lower noise emissions than two-bladed machines. The three blades
are attached to a hub and main shaft, from which power is transferred
(sometimes through a gearbox, depending on design) to a generator. The
main shaft and main bearings, gearbox, generator and control system
are contained within a housing called the nacelle. Figure 7.5 shows the
components in a modern wind turbine with a gearbox; in wind turbines
without a gearbox, the rotor is mounted directly on the generator shaft.
In the 1980s, larger machines were rated at around 100 kW and primarily
relied on aerodynamic blade stall to control power production from the
xed blades. These turbines generally operated at one or two rotational
speeds. As turbine size increased over time, development went from stall
control to full-span pitch control in which turbine output is controlled by
pitching (i.e., rotating) the blades along their long axis (EWEA, 2009). In
addition, a reduction in the cost of power electronics allowed variable
speed wind turbine operation. Initially, variable speeds were used to
smooth out the torque uctuations in the drive train caused by wind
turbulence and to allow more efcient operation in variable and gusty
winds. More recently, almost all electric system operators require the
continued operation of large wind power plants during electrical faults,
together with being able to provide reactive power: these requirements
have accelerated the adoption of variable-speed operation with power
electronic conversion (see Section 7.3.3 for a summary of power conversion technologies, Section 7.5 for a fuller discussion of electric system
integration issues, and Chapter 8 for a discussion of reactive power and
broader issues with respect to the integration of RE into electricity systems). Modern wind turbines typically operate at variable speeds using
full-span blade pitch control. Blades are commonly constructed with
composite materials, and towers are usually tubular steel structures that
taper from the base to the nacelle at the top (EWEA, 2009).
Over the past 30 years, average wind turbine size has grown signicantly (Figure 7.6), with the largest fraction of onshore wind turbines
installed globally in 2009 having a rated capacity of 1.5 to 2.5 MW; the
average size of turbines installed in 2009 was 1.6 MW (BTM, 2010). As
of 2010, wind turbines used onshore typically stand on 50- to 100-m
towers, with rotors that are often 50 to 100 m in diameter; commercial
Horizontal-Axis Turbines
Vertical-Axis Turbines
Figure 7.4 | Early wind turbine designs, including horizontal and vertical axis turbines (South et al., 1983).
551
Wind Energy
Chapter 7
Rotor Blade
Wind
Direction
D
ireectio
io
on
on
FFor
or aan
n
Upwind
Rotor
Wind
Direction
Nacelle with
Gearbox and
Generator
Low-Speed
Shaft
Rotor
Diameter
Swept Area
of Blades
Gear Box
Pitch
Generator
Rotor
Hub
Controller
Anemometer
Wind
Direction
Brake
Hub
Height
Yaw
Drive
Blades
Yaw
Motor
Tower
Tower
High-Speed
Shaft
Nacelle
Wind Vane
Transformer
Figure 7.5 | Basic components of a modern, horizontal-axis wind turbine with a gearbox (Design by the National Renewable Energy Laboratory (NREL)).
machines with rotor diameters and tower heights in excess of 125 m are
operating, and even larger machines are under development. Modern
turbines operate with rotational speeds ranging from 12 to 20 revolutions per minute (RPM), which compares to the faster and potentially
more visually disruptive speeds exceeding 60 RPM common of the
smaller turbines installed during the 1980s.13 Onshore wind turbines
are typically grouped together into wind power plants, sometimes also
called wind projects or wind farms. These wind power plants are often 5
to 300 MW in size, though smaller and larger plants do exist.
The main reason for the continual increase in turbine size to this point
has been to minimize the levelized generation cost of wind energy
13 Rotational speed decreases with larger rotor diameters. The acoustic noise resulting
from tip speeds greater than 70 to 80 m/s is the primary design criterion that governs
rotor speed.
552
by: increasing electricity production (taller towers provide access to a

higher-quality wind resource, and larger rotors allow a greater exploitation of those winds as well as more cost-effective exploitation of
lower-quality wind resource sites); reducing investment costs per unit
of capacity (installation of a fewer number of larger turbines can, to a
point, reduce overall investment costs); and reducing O&M costs (larger
turbines can reduce maintenance costs per unit of capacity) (EWEA,
2009). For onshore turbines, however, additional growth in turbine size
may ultimately be limited by not only engineering and materials usage
constraints (discussed in Section 7.7), but also by the logistical constraints (or cost of resolving those constraints) of transporting the very
large blades, tower, and nacelle components by road, as well as the cost
of and difculty in obtaining large cranes to lift the components into
place. These same constraints are not as binding for offshore turbines,
so future turbine scaling to the sizes shown in Figure 7.6 are more likely
to be driven by offshore wind turbine design considerations.
Hub Height (m)
Chapter 7
Wind Energy
320
250m
20,000kW
300
280
Past and Present

Wind Turbines
260
Future Wind
Turbines
150m
10,000kW
240
220
125m
5,000kW
200
180
100m
3,000kW
160
140
120
100
Rotor Diameter (m)

Rating (kW)
80m
1,800kW
50m
750kW
80
60
40
70m
1,500kW
17m
75kW
30m
300kW
20
0
1980- 19901990 1995
19952000
20002005
20052010
2010-?
2010-?
Future
Future
Figure 7.6 | Growth in size of typical commercial wind turbines (Design by NREL).
As a result of these and other developments, onshore wind energy technology is already being commercially manufactured and deployed on a
large scale. Moreover, modern wind turbines have nearly reached the
theoretical maximum of aerodynamic efciency, with the coefcient of
performance rising from 0.44 in the 1980s to about 0.50 by the mid
2000s.14 The value of 0.50 is near the practical limit dictated by the
drag of aerofoils and compares with the Lanchester-Betz theoretical
limit of 0.593 (see Section 7.3.1.1). The design requirement for wind
turbines is normally 20 years with 4,000 to 7,000 hours of operation
(at and below rated power) each year depending on the characteristics
of the local wind resource. Given the challenges of reliably meeting this
design requirement, O&M teams work to maintain high plant availability despite component failure rates that have, in some instances, been
higher than expected (Echavarria et al., 2008). Though wind turbines are
reportedly under-performing in some contexts (Li, 2010), data collected
through 2008 show that modern onshore wind turbines in mature markets can achieve an availability of 97% or more (Blanco, 2009; EWEA,
2009; IEA, 2009).
These results demonstrate that the technology has reached sufcient
commercial maturity to allow large-scale manufacturing and deployment. Nonetheless, additional advances to improve reliability, increase
electricity production and reduce costs are anticipated, and are discussed
in Section 7.7. Additionally, most of the historical technology advances
have occurred in developed countries. Increasingly, however, developing
countries are investigating the use of wind energy, and opportunities for
14 Wind turbines achieve maximum aerodynamic efciency when operating at wind
speeds corresponding to power levels below the rated power level (see Region II in
Figure 7.3). Aerodynamic efciency is limited by the control system when operating
at speeds above rated power (see Region III in Figure 7.3).
technology transfer in wind turbine design, component manufacturing

and wind power plant siting exist. Extreme environmental conditions,
such as icing or typhoons, may be more prominent in some of these markets, providing impetus for continuing research. Other aspects unique to
less-developed countries, such as minimal transportation infrastructure,
could also inuence wind turbine designs if and as these markets grow.
7.3.1.3
Offshore wind energy technology
The rst offshore wind power plant was built in 1991 in Denmark,
consisting of eleven 450 kW wind turbines. Offshore wind energy technology is less mature than onshore, and has higher investment and
O&M costs (see Section 7.8). By the end of 2009, just 1.3% of global
installed wind power capacity was installed offshore, totalling 2,100
MW (GWEC, 2010a).
The primary motivation to develop offshore wind energy is to provide access
to additional wind resources in areas where onshore wind energy development is constrained by limited technical potential and/or by planning
and siting conicts with other land uses. Other motivations for developing
offshore wind energy include: the higher-quality wind resources located
at sea (e.g., higher average wind speeds and lower shear near hub height;
wind shear refers to the general increase in wind speed with height);
the ability to use even larger wind turbines due to avoidance of certain
land-based transportation constraints and the potential to thereby gain
additional economies of scale; the ability to build larger power plants than
onshore, gaining plant-level economies of scale; and a potential reduction
in the need for new, long-distance, land-based transmission infrastructure
553
Wind Energy
to access distant onshore wind energy15 (Carbon Trust, 2008b; Snyder

and Kaiser, 2009b; Twidell and Gaudiosi, 2009). These factors, combined
with a signicant offshore wind resource potential, have created considerable interest in offshore wind energy technology in the EU and,
increasingly, in other regions, despite the typically higher costs relative
to onshore wind energy.
Offshore wind turbines are typically larger than onshore, with nameplate capacity ratings of 2 to 5 MW being common for offshore wind
power plants built from 2007 to 2009, and even larger turbines are
under development. Offshore wind power plants installed from 2007 to
2009 were typically 20 to 120 MW in size, with a clear trend towards
larger turbines and power plants over time. Water depths for most offshore wind turbines installed through 2005 were less than 10 m, but
from 2006 to 2009, water depths from 10 to more than 20 m were
common. Distance to shore has most often been below 20 km, but average distance has increased over time (EWEA, 2010a). As experience is
gained, water depths are expected to increase further and more exposed
locations with higher winds will be utilized. These trends will impact the
wind resource characteristics faced by offshore wind power plants, as
well as support structure design and the cost of offshore wind energy. A
continued transition towards larger wind turbines (5 to 10 MW, or even
larger) and wind power plants is also anticipated as a way of reducing the cost of offshore wind energy through turbine- and plant-level
economies of scale.
To date, offshore turbine technology has been very similar to onshore
designs, with some modications and with special foundations (Musial,
2007; Carbon Trust, 2008b). The mono-pile foundation is the most common, though concrete gravity-based foundations have also been used
with some frequency; a variety of other foundation designs (including oating designs) are being considered and in some instances
used (Breton and Moe, 2009), especially as water depths increase,
as discussed in Section 7.7. In addition to differences in foundations,
modication to offshore turbines (relative to onshore) include structural upgrades to the tower to address wave loading; air conditioned
and pressurized nacelles and other controls to prevent the effects of
corrosive sea air from degrading turbine equipment; and personnel
access platforms to facilitate maintenance. Additional design changes
for marine navigational safety (e.g., warning lights, fog signals) and to
minimize expensive servicing (e.g., more extensive condition monitoring, onboard service cranes) are common. Wind turbine tip speed could
be chosen to be greater than for onshore turbines because concerns
about noise are reduced for offshore power plantshigher tip speeds
can sometimes lead to lower torque and lighter drive train components
for the same power output. In addition, tower heights are sometimes
15 Of course, transmission infrastructure is needed to connect offshore wind power
plants with electricity demand centres, and the per-kilometre cost of offshore transmission typically exceeds that for onshore lines. Whether offshore transmission needs
are more or less extensive than those needed to access onshore wind energy varies
by location.
554
Chapter 7
lower than used for onshore wind power plants due to reduced wind
shear offshore relative to onshore.
Lower power plant availabilities and higher O&M costs have been common for offshore wind energy relative to onshore wind both because of
the comparatively less mature state of offshore wind energy technology
and because of the inherently greater logistical challenges of maintaining and servicing offshore turbines (Carbon Trust, 2008b; UKERC, 2010).
Wind energy technology specically tailored for offshore applications
will become more prevalent as the offshore market expands, and it is
expected that larger turbines in the 5 to 10 MW range may come to
dominate this market segment (EU, 2008). Future technical advancement possibilities for offshore wind energy are described in Section 7.7.
7.3.2
International wind energy technology standards
Wind turbines in the 1970s and 1980s were designed using simplied
design models, which in some cases led to machine failures and in other
cases resulted in design conservatism. The need to address both of these
issues, combined with advances in computer processing power, motivated designers to improve their calculations during the 1990s (Quarton,
1998; Rasmussen et al., 2003). Improved design and testing methods
have been codied in International Electrotechnical Commission (IEC)
standards, and the rules and procedures for Conformity Testing and
Certication of Wind Turbines (IEC, 2010) relies upon these standards.
Certication agencies rely on accredited design and testing bodies to
provide traceable documentation of the execution of rules and specications outlined in the standards in order to certify turbines, components
or entire wind power plants. The certication system assures that a wind
turbine design or wind turbines installed in a given location meet common guidelines relating to safety, reliability, performance and testing.
Figure 7.7(a) illustrates the design and testing procedures required to
obtain a wind-turbine type certication. Plant certication, shown in
Figure 7.7(b), requires a type certicate for the turbine and includes procedures for evaluating site conditions and turbine design parameters
associated with that specic site, as well as other site-specic conditions
including soil properties, installation and plant commissioning.
Insurance companies, nancing institutions and power plant owners
normally require some form of certication for plants to proceed, and
the IEC standards therefore provide a common basis for certication to
reduce uncertainty and increase the quality of wind turbine products
available in the market (EWEA, 2009). In emerging markets, the lack of
highly qualied testing laboratories and certication bodies limits the
opportunities for manufacturers to obtain certication according to IEC
standards and may lead to lower-quality products. As markets mature
and design margins are compressed to reduce costs, reliance on internationally recognized standards is likely to become even more widespread
to assure consistent performance, safety and reliability of wind turbines.
Chapter 7
Wind Energy
(a) Wind Turbine Type Certication Procedure
(b) Wind Project Certication Procedure
Site Conditions
Assessment
Design Basis
Evaluation
Type Certicate
Design Basis
Evaluation
Design
Evaluation
Foundation Design
Evaluation
Manufacturing
Evaluation
Foundation
Manufacturing
Evaluation
Integrated Load
Analysis
Wind Turbine/RNA
Design Evaluation
Support Structure
Design Evaluation
Other Installations
Design Evaluation
Wind Turbine/RNA
Manuf. Surveillance
Support Structure
Manuf. Surveillance
Other Installations
Manuf. Surveillance
Type Testing
Transportation &
Install Surveillance
Type Characteristics
Measurements
Commissioning
Surveillance
Project Characteristics
Measurements
Final Evaluation
Optional Module
Final Evaluation
Optional Module
Type Certicate
Project Certicate
Operational &
Maintainance Surveillance
Figure 7.7 | Modules for (a) turbine type certication and (b) wind power plant certication (IEC, 2010).
Notes: RNA refers to Rotor Nacelle Assembly. The authors thank the IEC for permission to reproduce information from its International Standard IEC 61400-22 ed.1.0 (2010). All such
extracts are copyright of IEC, Geneva, Switzerland. All rights reserved. Further information on the IEC is available from www.iec.ch. IEC has no responsibility for the placement and
context in which the extracts and contents are reproduced by the authors, nor is IEC in any way responsible for the other content or accuracy therein. Copyright 2010 IEC Geneva,
Switzerland, www.iec.ch.
7.3.3
Power conversion and related grid

connection issues
From an electric system reliability perspective, an important part of the

wind turbine is the electrical conversion system. For large grid-connected
turbines, electrical conversion systems come in three broad forms.
Fixed-speed induction generators were popular in earlier years for both
stall-regulated and pitch-controlled turbines; in these arrangements, wind
turbines were net consumers of reactive power that had to be supplied
by the electric system (see Ackermann, 2005). For modern turbines, these
designs have now been largely replaced with variable-speed machines.
Two arrangements are common, doubly-fed induction generators and
synchronous generators with a full power electronic converter, both of

which are almost always coupled with pitch-controlled rotors. These
variable-speed designs essentially decouple the rotating masses of the
turbine from the electric system, thereby offering a number of power
quality advantages over earlier turbine designs (Ackermann, 2005; EWEA,
2009). For example, these turbines can provide real and reactive power as
well as some fault ride-through capability, which are increasingly being
required by electric system operators (these requirements and the institutional elements of wind energy integration are addressed in Section 7.5).
These designs differ from the synchronous generators found in most largescale fossil fuel-powered plants, however, in that they result in no intrinsic
inertial response capability, that is, they do not increase (decrease) power
555
Wind Energy
Chapter 7
7.4
Global and regional status of market

and industry development
The wind energy market expanded substantially in the 2000s, demonstrating the commercial and economic viability of the technology and
industry, and the importance placed on wind energy development by
a number of countries through policy support measures. Wind energy
expansion has been concentrated in a limited number of regions, however, and wind energy remains a relatively small fraction of global
electricity supply. Further expansion of wind energy, especially in regions
of the world with little wind energy deployment to date and in offshore
locations, is likely to require additional policy measures.
This section summarizes the global (Section 7.4.1) and regional (Section
7.4.2) status of wind energy deployment, discusses trends in the wind
energy industry (Section 7.4.3) and highlights the importance of policy
actions for the wind energy market (Section 7.4.4).
7.4.1
Global status and trends
Annual Capacity Additions [GW]
Wind energy has quickly established itself as part of the mainstream

electricity industry. From a cumulative capacity of 14 GW at the end
of 1999, global installed wind power capacity increased 12-fold in 10
years to reach almost 160 GW by the end of 2009, an average annual
increase in cumulative capacity of 28% (see Figure 7.8). Global annual
wind power capacity additions equalled more than 38 GW in 2009, up
from 26 GW in 2008 and 20 GW in 2007 (GWEC, 2010a).
The majority of the capacity has been installed onshore, with offshore
installations constituting a small proportion of the total market. About 2.1
GW of offshore wind turbines were installed by the end of 2009; 0.6 GW
were installed in 2009, including the rst commercial offshore wind power
plant outside of Europe, in China (GWEC, 2010a). Many of these offshore
installations have taken place in the UK and Denmark. Signicant offshore
wind power plant development activity, however, also exists in, at a minimum, other EU countries, the USA, Canada and China (e.g., Mostafaeipour,
2010). Offshore wind energy is expected to develop in a more signicant
way in the years ahead as the technology advances and as onshore wind
energy sites become constrained by local resource availability and/or siting
challenges in some regions (BTM, 2010; GWEC, 2010a).
The total investment cost of new wind power plants installed in 2009 was
USD2005 57 billion (GWEC, 2010a). Direct employment in the wind energy
sector in 2009 has been estimated at roughly 190,000 in the EU and
85,000 in the USA. Worldwide, direct employment has been estimated at
approximately 500,000 (GWEC, 2010a; REN21, 2010).
Despite these trends, wind energy remains a relatively small fraction of
worldwide electricity supply. The total wind power capacity installed by the
end of 2009 would, in an average year, meet roughly 1.8% of worldwide
electricity demand, up from 1.5% by the end of 2008, 1.2% by the end of
2007, and 0.9% by the end of 2006 (Wiser and Bolinger, 2010).
7.4.2
Regional and national status and trends
The countries with the highest total installed wind power capacity by the
end of 2009 were the USA (35 GW), China (26 GW), Germany (26 GW),
Spain (19 GW) and India (11 GW). After its initial start in the USA in the
1980s, wind energy growth centred on countries in the EU and India during the 1990s and the early 2000s. In the late 2000s, however, the USA and
then China became the locations for the greatest annual capacity additions (Figure 7.9).
Regionally, Europe continues to lead the market with 76 GW of cumulative installed wind power capacity by the end of 2009, representing 48%
of the global total (Asia represented 25%, whereas North America
180
60
50
150
Annual Wind Power Installations (Left Axis)

Cumulative Wind Power Capacity (Right Axis)
40
120
30
90
20
60
10
30
0
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
Figure 7.8 | Global annual wind power capacity additions and cumulative capacity (Data sources: GWEC, 2010a; Wiser and Bolinger, 2010).
556
2007
2008
2009
Cumulative Capacity [GW]
output in synchronism with system power imbalances. This lack of inertial

response is an important consideration for electric system planners because
less overall inertia in the electric system makes the maintenance of stable
system operation more challenging (Gautam et al., 2009). Wind turbine
manufacturers have recognized this lack of intrinsic inertial response as
a possible long-term impediment to wind energy and are actively pursuing a variety of solutions; for example, additional turbine controls can be
added to provide inertial response (Mullane and OMalley, 2005; Morren
et al., 2006).
Cumulative Wind Power Capacity [GW]
Chapter 7
Wind Energy
40
35
2006
2007
30
2008
2009
25
20
15
10
5
0
US
China
Germany
Spain
India
Italy
France
UK
Portugal
Denmark
Figure 7.9 | Top-10 countries in cumulative wind power capacity (Date source: GWEC, 2010a).
represented 24%). Notwithstanding the continuing growth in Europe,

the trend over time has been for the wind energy industry to become
less reliant on a few key markets, and other regions of the world have
increasingly become the dominant markets for wind energy growth. The
annual growth in the European wind energy market in 2009, for example, accounted for just 28% of the total new wind power additions in
that year, down from over 60% in the early 2000s (GWEC, 2010a). More
than 70% of the annual wind power capacity additions in 2009 occurred
outside of Europe, with particularly signicant growth in Asia (40%) and
North America (29%) (Figure 7.10). Even in Europe, though Germany
and Spain have been the strongest markets during the 2000s, there is a
trend towards less reliance on these two countries.
Annual Capacity Additions, by Region [GW]
Despite the increased globalization of wind power capacity additions, the

market remains concentrated regionally. As shown in Figure 7.10, Latin
America, Africa and the Middle East, and the Pacic regions have installed
relatively little wind power capacity despite signicant technical potential

in each region, as presented earlier in Section 7.2. And, even in the regions
of signicant growth, most of that growth has occurred in a limited
number of countries. In 2009, for example, 90% of wind power capacity
additions occurred in the 10 largest markets, and 62% was concentrated
in just two countries: China (14 GW, 36%) and the USA (10 GW, 26%).
In both Europe and the USA, wind energy represents a major new source
of electric capacity additions. From 2000 through 2009, wind energy
was the second-largest new resource added in the USA (10% of all gross
capacity additions) and EU (33% of all gross capacity additions) in terms
of nameplate capacity, behind natural gas but ahead of coal. In 2009,
39% of all capacity additions in the USA and 39% of all additions in the
EU came from wind energy (Figure 7.11). In China, 5% of the net capacity additions from 2000 to 2009 and 16% of the net additions in 2009
came from wind energy. On a global basis, from 2000 through 2009,
16
14
2006
2007
12
2008
2009
10
8
6
4
2
0
Europe
North America
Asia
Latin America
Africa &
Middle East
Pacic
Figure 7.10 | Annual wind power capacity additions by region (Data source: GWEC, 2010a).
Note: Regions shown in the gure are dened by the study.
557
Wind Energy
Chapter 7
Percent of Nameplate Electric Capacity Additions [%]
201 GW
324 GW
26 GW
Total Additions
25 GW
100
90
Other
80
Coal
Natural Gas
70
Wind
60
50
40
A number of countries are beginning to achieve relatively high levels of

annual wind electricity penetration in their respective electric systems.
Figure 7.12 presents data for the end of 2009 (and the end of 2006, 2007
and 2008) on installed wind power capacity, translated into projected
annual electricity supply, and divided by electricity consumption. On this
basis, and focusing only on the 20 countries with the greatest cumulative wind power capacity, at the end of 2009, wind power capacity
was capable of supplying electricity equal to roughly 20% of Denmarks
annual electricity demand, 14% of Portugals, 14% of Spains, 11% of
Irelands and 8% of Germanys (Wiser and Bolinger, 2010).17
30
20
7.4.3
10
0
EU
US
EU
2000-2009
US
2009
Figure 7.11 | Relative contribution of electricity supply types to gross capacity additions
in the EU and the USA (Data sources: EWEA, 2010b; Wiser and Bolinger, 2010).
Projected Wind Electricity as a

Proportion of Electricity Consumption [%]
Note: The other category includes other forms of renewable energy, nuclear energy,
and fuel oil.
Industry development
The growing maturity of the wind energy sector is illustrated not only
by wind power capacity additions, but also by trends in the wind energy
industry. In particular, major established companies from outside the
traditional wind energy industry have become increasingly involved in
the sector. For example, there has been a shift in the type of companies
developing, owning and operating wind power plants, from relatively
small independent power plant developers to large power generation
22
20
Approximate Wind Penetration, End of 2009
18
16
14
12
10
8
6
4
TOTAL
Japan
Brazil
China
Turkey
Canada
Australia
France
Sweden
US
Austria
India
Italy
UK
Netherlands
Greece
Germany
Ireland
Spain
Portugal
Denmark
Figure 7.12 | Approximate annual average wind electricity penetration in the twenty countries with the greatest installed wind power capacity (Wiser and Bolinger, 2010).
roughly 11% of all newly installed net electric capacity additions came
from new wind power plants; in 2009 alone, that gure was probably
more than 20%.16
16 Worldwide capacity additions from 2000 through 2007 come from historical data
from the US Energy Information Administration. Capacity additions for 2008 and
2009 are estimated based on historical capacity growth from 2000 to 2007. The focus here is on capacity additions in GW terms, though it is recognized that electricity
generation technologies often have widely divergent average capacity factors, and
that the contribution of wind energy to new electricity demand (in GWh terms) may
differ from what is presented here.
558
companies (including electric utilities) and large independent power

plant developers. With respect to wind turbine and component manufacturing, the increase in the size and geographic spread of the wind
energy market, along with manufacturing localization requirements in
some countries, has brought in new players. The involvement of these
new players has, in turn, encouraged a greater globalization of the
industry. Manufacturer product strategies are shifting to address larger
17 Because of interconnections among electricity grids, these percentages do not necessarily equate to the amount of wind electricity consumed within each country.
Chapter 7
scale power plants, higher capacity and offshore turbines, and lower
wind speeds. More generally, the signicant contribution of wind energy
to new electric capacity investment in several regions of the world has
attracted a broad range of players across the industry supply chain,
from local site-focused engineering rms to global vertically integrated
utilities. The industrys supply chain has also become increasingly competitive as a multitude of rms seek the most protable balance between
vertical integration and specialization (BTM, 2010; GWEC, 2010a).
Despite these trends, the global wind turbine market remains somewhat
regionally segmented, with just six countries hosting the majority of
wind turbine manufacturing (China, Denmark, India, Germany, Spain
and the USA). With markets developing differently, market share for turbine supply has been marked by the emergence of national industrial
champions, the entry of highly focused technology innovators and the
arrival of new start-ups licensing proven technology from other regions
(Lewis and Wiser, 2007). Regardless, the industry continues to globalize:
Europes turbine and component manufacturers have penetrated the
North American and Asian markets, and the growing presence of Asian
manufacturers in Europe and North America is expected to become more
pronounced in the years ahead. Chinese wind turbine manufacturers, in
particular, are dominating their home market, and will increasingly seek
export opportunities. Wind turbine sales and supply chain strategies are
therefore expected to continue to take on a more international dimension as volumes increase.
Amidst the growth in the wind energy industry also come challenges. As
discussed further in Section 7.8, from 2005 through 2008, supply chain difculties caused by growing demand for wind energy strained the industry,
and prices for wind turbines and turbine components increased to compensate for this imbalance. Commodity price increases, the availability of skilled
labour and other factors also played a role in pushing wind turbine prices
higher, while the underdeveloped supply chain for offshore wind power
plants strained that portion of the industry. Overcoming supply chain difculties is not simply a matter of ramping up the production of wind turbine
components to meet the increased levels of demand. Large-scale investment decisions are more easily made based on a sound long-term outlook
for the industry. In most markets, however, both the projections and actual
demand for wind energy depend on a number of factors, some of which
are outside of the control of the industry, such as political frameworks and
policy measures.
7.4.4
Impact of policies18
The deployment of wind energy must overcome a number of challenges

that vary in type and magnitude depending on the wind energy application and region.19 The most signicant challenges to wind energy
deployment are summarized here. Perhaps most importantly, in many
18 Non-technology-specic policy issues are covered in Chapter 11 of this report.
19 For a broader discussion of barriers and market failures associated with renewable
energy, see Sections 1.4 and 11.1, respectively.
Wind Energy
(though not all) regions of the world, wind energy is more expensive
than current energy market prices, at least if environmental impacts are
not internalized and monetized (NRC, 2010a). Wind energy also faces
a number of other challenges, some of which are somewhat unique to
wind energy or are at least particularly relevant to this sector. Some
of the most critical challenges include: (1) concerns about the impact
of wind energys variability on electricity reliability; (2) challenges to
building the new transmission infrastructure both on- and offshore (and
within country and cross-border) needed to enable access to the most
attractive wind resource areas; (3) cumbersome and slow planning, siting and permitting procedures that impede wind energy deployment;
(4) the technical advancement needs and higher cost of offshore wind
energy technology; and (5) lack of institutional and technical knowledge
in regions that have not experienced substantial wind energy deployment to this point.
As a result of these challenges, growth in the wind energy sector is
affected by and responsive to political frameworks and a wide range
of government policies. During the past two decades, a signicant
number of developed countries and, more recently, a growing number
of developing nations have laid out RE policy frameworks that have
played a major role in the expansion of the wind energy market. These
efforts have been motivated by the environmental, fuel diversity, and
economic development impacts of wind energy deployment, as well as
the potential for reducing the cost of wind energy over time. An early
signicant effort to deploy wind energy at a commercial scale occurred
in California, with a feed-in tariff and aggressive tax incentives spurring
growth in the 1980s (Bird et al., 2005). In the 1990s, wind energy deployment moved to Europe, with feed-in tariff policies initially established
in Denmark and Germany, and later expanding to Spain and then a
number of other countries (Meyer, 2007); renewable portfolio standards
have been implemented in other European countries and, more recently,
European renewable energy policies have been motivated in part by the
EUs binding 20%-by-2020 target for renewable energy. In the 2000s,
growth in the USA (Bird et al., 2005; Wiser and Bolinger, 2010), China (Li
et al., 2007; Li, 2010; Liu and Kokko, 2010), and India (Goyal, 2010) was
based on varied policy frameworks, including renewable portfolio standards, tax incentives, feed-in tariffs and government-overseen bidding.
Still other policies have been used in a number of countries to directly
encourage the localization of wind turbine and component manufacturing (Lewis and Wiser, 2007).
Though economic support policies differ, and a healthy debate exists
over the relative merits of different approaches, a key nding is that
both policy transparency and predictability are important (see Chapter
11). Moreover, though it is not uncommon to focus on economic policies for wind energy, as noted above and as discussed elsewhere in this
chapter and in Chapter 11, experience shows that wind energy markets
are also dependent on a variety of other factors (e.g., Valentine, 2010).
These include local resource availability, site planning and approval procedures, operational integration into electric systems, transmission grid
expansion, wind energy technology improvements, and the availability
of institutional and technical knowledge in markets unfamiliar with
559
Wind Energy
wind energy (e.g., IEA, 2009). For the wind energy industry, these issues
have been critical in dening both the size of the market opportunity
in each country and the rules for participation in those opportunities;
many countries with sizable wind resources have not deployed signicant amounts of wind energy as a result of these factors. Given the
challenges to wind energy listed earlier, successful frameworks for
wind energy deployment might consider the following elements: support systems that offer adequate protability and that ensure investor
condence; appropriate administrative procedures for wind energy
planning, siting and permitting; a degree of public acceptance of wind
power plants to ease implementation; access to the existing transmission system and strategic transmission planning and new investment
for wind energy; and proactive efforts to manage wind energys inherent output variability and uncertainty. In addition, R&D by government
and industry has been essential to enabling incremental improvements
in onshore wind energy technology and to driving the improvements
needed in offshore wind energy technology. Finally, for those markets
that are new to wind energy deployment, both knowledge (e.g., wind
resource mapping expertise) and technology transfer (e.g., to develop
local wind turbine manufacturers and to ease grid integration) can help
facilitate early installations.
7.5
Near-term grid integration issues20
As wind energy deployment has increased, so have concerns about the

integration of that energy into electric systems (e.g., Fox et al., 2007).
The nature and magnitude of the integration challenge will be system
specic and will vary with the degree of wind electricity penetration.
Moreover, as discussed in Chapter 8, integration challenges are not
unique to wind energy: adding any type of generation technology to
an electric system, particularly location-constrained variable generation,
presents challenges. Nevertheless, analysis and operating experience
primarily from certain OECD countries (where most of the wind energy
deployment has occurred, until recently, see Section 7.4.2) suggest that,
at low to medium levels of wind electricity penetration (dened here
as up to 20% of total annual average electrical energy demand),21 the
integration of wind energy generally poses no insurmountable technical
barriers and is economically manageable. In addition, increased operating experience with wind energy along with improved technology,
altered operating and planning practices and additional research should
facilitate the integration of even greater quantities of wind energy. Even
at low to medium levels of wind electricity penetration, however, certain
(and sometimes system-specic) technical and/or institutional challenges must be addressed.
20 Non-technology-specic issues related to integration of RE sources in current and
future energy systems are covered in Chapter 8 of this report.
21 This level of penetration was chosen to loosely separate the integration needs for
wind energy in the relatively near term from the broader, longer-term, and non-windspecic discussion of electric system changes provided in Chapter 8. In addition, the
majority of operational experience and literature on the integration of wind energy
addresses penetration levels below 20%.
560
Chapter 7
The integration issues covered in this section include how to address

wind power variability and uncertainty, the possible need for additional
transmission capacity to enable remotely located wind power plants to
meet the needs of electricity demand centres, and the development of
technical standards for connecting wind power plants with electric systems. The focus is on those issues faced at low to medium levels of wind
electricity penetration (up to 20%). Even higher levels of penetration
may depend on or benet from the availability of additional exibility
options, such as: further increasing the exibility of other electricity generation plants (fossil and otherwise); mass-market demand response;
large-scale deployment of electric vehicles and their associated contributions to system exibility through controlled battery charging; greater
use of wind power curtailment and output control or diverting excess
wind energy to fuel production or local heating; increased deployment
of bulk energy storage technologies; and further improvements in the
interconnections between electric systems. The deployment of a diversity of RE technologies may also help facilitate overall electric system
integration. Many of these options relate to broader developments
within the energy sector that are not specic to wind energy, however,
and most are therefore addressed in Chapter 8.
This section begins by describing the specic characteristics of wind
energy that present integration challenges (Section 7.5.1). The section
then discusses how these characteristics impact issues associated with
the planning (Section 7.5.2) and operation (Section 7.5.3) of electric
systems to accommodate wind energy, including a selective discussion
of actual operating experience. Finally, Section 7.5.4 summarizes the
results of various studies that have quantied the technical issues and
economic costs of integrating increased quantities of wind energy.
7.5.1
Wind energy characteristics
Several important characteristics of wind energy are different from

those of many other generation sources. These characteristics must be
considered in electric system planning and operation to ensure the reliable and economical operation of the electric power system.
The rst characteristic to consider is that the quality of the wind resource
and therefore the cost of wind energy is location dependent. As a
result, regions with the highest-quality wind resources may not be situated near population centres that have high electricity demands (e.g.,
Hoppock and Patio-Echeverri, 2010; Liu and Kokko, 2010). Additional
transmission infrastructure is therefore sometimes economically justied (and is often needed) to bring wind energy from higher-quality wind
resource areas to electricity demand centres as opposed to utilizing
lower-quality wind resources that are located closer to demand centres
and that may require less new transmission investment (see Sections
7.5.2.3 and 7.5.4.3).
The second important characteristic is that wind energy is weather
dependent and therefore variablethe power output of a wind power
plant varies from zero to its rated capacity depending on prevailing
Chapter 7
Wind Energy
weather conditions. Variations can occur over multiple time scales, from
shorter-term sub-hourly uctuations to diurnal, seasonal, and ever interannual uctuations (e.g., Van der Hoven, 1957; Justus et al., 1979; Wan
and Bucaneg, 2002; Apt, 2007; Rahimzadeh et al., 2011). The nature of
these uctuations and patterns is highly site- and region-specic. Figure
7.13 illustrates some elements of this variability by showing the scaled
output of an individual wind turbine, a small collection of wind power
plants, and a large collection of wind power plants in Germany over 10
consecutive days. An important aspect of wind power variability for electric system operations is the rate of change in wind power output over
different relatively short time periods; Figure 7.13 demonstrates that the
aggregate output of multiple wind power plants changes much more
dramatically over relatively longer periods (multiple hours) than over very
short periods (minutes). An important aspect of wind power variability
for the purpose of electric sector planning, on the other hand, is the correlation of wind power output with the periods of time when electric
system reliability is at greatest risk, typically periods of high electricity
demand. In this case, the diurnal, seasonal, and even interannual patterns
of wind power output (and the correlation of those patterns with electricity demand) can impact the capacity credit assigned by system planners
to wind power plants, as discussed further in Section 7.5.3.4.
Normalized Wind Power
Third, in comparison with many other types of power plants, wind

power output has lower levels of predictability. Forecasts of wind power
output use various approaches and have multiple goals, and signicant
improvements in forecasting accuracy have been achieved in recent
years (e.g., Costa et al., 2008). Despite those improvements, however,
forecasts remain imperfect. In particular, forecasts are less accurate
over longer forecast horizons (multiple hours to days) than over shorter
periods (e.g., H. Madsen et al., 2005), which, depending on the characteristics of the electric system, can have implications for the ability of
that system and related trading markets to manage wind power variability and uncertainty (Usaola, 2009; Weber, 2010).
The aggregate variability and uncertainty of wind power output
depends, in part, on the degree of correlation between the outputs of
different geographically dispersed wind power plants. This correlation
between the outputs of wind power plants, in turn, depends on the
geographic deployment of the plants and the regional characteristics
of weather patterns, especially wind speeds. Generally, the output of
wind power plants that are farther apart are less correlated with each
other, and variability over shorter time periods (minutes) is less correlated than variability over longer time periods (multiple hours) (e.g.,
Wan et al., 2003; Sinden, 2007; Holttinen et al., 2009; Katzenstein et
al., 2010). This lack of perfect correlation results in a smoothing effect
associated with geographic diversity when the output of multiple
wind turbines and power plants are combined, as illustrated in Figure
7.13: the aggregate scaled variability shown for groups of wind power
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
Single Turbine
0.1
Group of Wind Plants

All German Wind Power
0
7-Dec-06
9-Dec-06
11-Dec-06
13-Dec-06
15-Dec-06
17-Dec-06
Date
Figure 7.13 | Example time series of wind power output scaled to wind power capacity for a single wind turbine, a group of wind power plants, and all wind power plants in Germany
over a 10-day period in 2006 (Durstewitz et al., 2008)
561
Wind Energy
plants over a region is less than the scaled output of a single wind
turbine. This apparent smoothing of aggregated output is due to the
decreasing correlation of output between different wind power plants
as distance between those plants increases. If, on the other hand, the
output of multiple wind turbines and power plants was perfectly correlated, then the aggregate variability would be equivalent to the scaled
variability of a single turbine. With sufcient transmission capacity
between wind power plants, the observed geographic smoothing
effect has implications for the variability of aggregate wind power
output that electric systems must accommodate, and also inuences
forecast accuracy because accuracy improves with the number and
diversity of wind power plants considered (e.g., Focken et al., 2002).
7.5.2
Planning electric systems with wind energy
Detailed system planning for new generation and transmission infrastructure is used to ensure that the electric system can be operated
reliably and economically in the future. Advanced planning is required
due, in part, to the long time horizons required to build new electricity
infrastructure. More specically, electric system planners22 must evaluate the adequacy of transmission to deliver electricity to demand centres
and the adequacy of generation to maintain a balance between supply and demand under a variety of operating conditions. Though not
an exhaustive list, four technical planning issues are prominent when
considering increased reliance on wind energy: the need for accurate
electric system models of wind turbines and power plants; the development of technical standards for connecting wind power plants with
electric systems (i.e., grid codes); the broader transmission infrastructure
needs of electric systems with wind energy; and the maintenance of
overall generation adequacy with increased wind electricity penetration.
7.5.2.1
Electric system models
Computer-based simulation models are used extensively to evaluate the

ability of the electric system to accommodate new generation, changes
in demand and changes in operational practices. An important role of
electric system models is to demonstrate the ability of an electric system
to recover from severe events or contingencies. Generic models of typical synchronous generators have been developed and validated over a
period of multiple decades, and are used in industry standard software
tools (e.g., power system simulators and analysis models) to study how
the electric system and all its components will behave during system
events or contingencies. Similar generic models of wind turbines and
wind power plants are in the process of being developed and validated.
Because wind turbines have electrical characteristics that differ from
typical synchronous generators, this modelling exercise requires signicant effort. As a result, though considerable progress has been made,
22 Electric system planners (or organizations that plan electric systems) is used here as a
generic term that refers to planners within any organization that regulates, operates
components of, or builds infrastructure for the electric system.
562
Chapter 7
this progress is not complete, and increased deployment of wind energy

will require improved and validated models to allow planners to better
assess the capability of electric systems to accommodate wind energy
(Coughlan et al., 2007; NERC, 2009).
7.5.2.2
Wind power electrical characteristics and grid codes
As wind power capacity has increased, so has the need for wind power
plants to become more active participants in maintaining (rather than
passively depending on) the operability and power quality of the electric
system. Focusing here primarily on the technical aspects of grid connection, the electrical performance of wind turbines in interaction with the
grid is often veried in accordance with international standards for the
characteristics of wind turbines, in which methods to assess the impact
of one or more wind turbines on power quality are specied (IEC, 2008).
Additionally, an increasing number of electric system operators have
implemented technical standards (sometimes called grid codes) that
wind turbines and/or wind power plants (and other power plants) must
meet when connecting to the grid to help prevent equipment or facilities
from adversely affecting the electric system during normal operation
and contingencies (see also Chapter 8). Electric system models and
operating experience are used to develop these requirements, which can
then typically be met through modications to wind turbine design or
through the addition of auxiliary equipment such as power conditioning
devices. In some cases, the unique characteristics of specic generation
types are addressed in grid codes, resulting in wind-specic grid codes
(e.g., Singh and Singh, 2009).
Grid codes often require fault ride-through capability, or the ability of
a wind power plant to remain connected and operational during brief
but severe changes in electric system voltage (Singh and Singh, 2009).
The requirement for fault ride-through capability was in response to the
increasing penetration of wind energy and the signicant size of individual wind power plants. Electric systems can typically maintain reliable
operation when small individual power plants shut down or disconnect
from the system for protection purposes in response to fault conditions.
When a large amount of wind power capacity disconnects in response to
a fault, however, that disconnection can exacerbate the fault conditions.
Electric system planners have therefore increasingly specied that wind
power plants must meet minimum fault ride-through standards similar
to those required of other large power plants. System-wide approaches
have also been adopted: in Spain, for example, wind power output may
be curtailed in order to avoid potential reliability issues in the event of
a fault; the need to employ this curtailment, however, is expected to
decrease as fault ride-through capability is added to new and existing wind power plants (Rivier Abbad, 2010). Reactive power control to
help manage voltage is also often required by grid codes, enabling wind
turbines to improve voltage stability margins particularly in weak parts
of the electric system (Vittal et al., 2010). Requirements for wind turbine inertial response to improve system stability after disturbances are
less common, but are under consideration (Hydro-Quebec TransEnergie,
Chapter 7
2006; Doherty et al., 2010). Active power control (including limits on how
quickly wind power plants can change their output) and frequency control are also sometimes required (Singh and Singh, 2009). Finally, controls
can be added to wind power plants to enable benecial dampening of
inter-area oscillations during dynamic events (Miao et al., 2009).
7.5.2.3
system, as indicated by the capacity credit assigned to the resource.23

Although there is not a strict, uniform denition of capacity credit,
the capacity credit of a generator is usually a system characteristic
in that it is determined not only by the generators characteristics but
also by the characteristics of the electric system to which that generator is connected, particularly the temporal prole of electricity demand
(Amelin, 2009).
Transmission infrastructure
As noted earlier, the highest-quality wind resources (whether on- or offshore) are often located at a distance from electricity demand centres.
As a result, even at low to medium levels of wind electricity penetration,
the addition of large quantities of wind energy in areas with the strongest wind resources may require signicant new additions or upgrades
to the transmission system (see also Chapter 8). Transmission adequacy
evaluations must consider any tradeoffs between the costs of expanding the transmission system to access higher-quality wind resources and
the costs of accessing lower-quality wind resources that require less
transmission investment (e.g., Hoppock and Patio-Echeverri, 2010). In
addition, evaluations of new transmission capacity need to account for
the relative smoothing benets of aggregating wind power plants over
large areas, the amount of transmission capacity devoted to managing
the remaining variability of wind power output, and the broader nonwind-specic advantages and disadvantages of transmission expansion
(Burke and OMalley, 2010).
Irrespective of the costs and benets of transmission expansion to
accommodate increased wind energy deployment, one of the primary
challenges is the long time it can take to plan, site, permit and construct new transmission infrastructure relative to the shorter time it
often takes to add new wind power plants. Depending on the legal and
regulatory framework in any particular region, the institutional challenges of transmission expansion, including cost allocation and siting,
can be substantial (e.g., Benjamin, 2007; Vajjhala and Fischbeck, 2007;
Swider et al., 2008). Enabling increased penetration of wind electricity
may therefore require the creation of regulatory and legal frameworks
for proactive rather than reactive transmission planning (Schumacher
et al., 2009). Estimates of the cost of the new transmission required to
achieve low to medium levels of wind electricity penetration in a variety
of locations around the world are summarized in Section 7.5.4.
7.5.2.4
Wind Energy
Generation adequacy
Though methods and objectives vary from region to region, generation

adequacy evaluations are generally used to assess the capability of generation resources to reliably meet electricity demand. Planners often
evaluate the long-term reliability of the electric system by estimating
the probability that the system will be able to meet expected demand in
the future, as measured by a statistical metric called the load-carrying
capability of the system. Each electricity supply resource contributes
some fraction of its nameplate capacity to the overall capability of the
The contribution of wind energy to long-term reliability can be evaluated using standard approaches, and wind power plants are typically
found to have a capacity credit of 5 to 40% of nameplate capacity (see
Figure 7.14). The correlation between wind power output and electrical demand is an important determinant of the capacity credit of an
individual wind power plant. In many cases, wind power output is uncorrelated or is weakly negatively correlated with periods of high electricity
demand, reducing the capacity credit of wind power plants; this is not
always the case, however, and wind power output in the UK, for example, has been found to be weakly positively correlated with periods of
high demand (Sinden, 2007). These correlations are case specic as they
depend on the diurnal, seasonal and yearly characteristics of both wind
power output and electricity demand. A second important characteristic of the capacity credit for wind energy is that its value generally
decreases as wind electricity penetration levels rise, because the capacity credit of a generator is greater when power output is well-correlated
with periods of time when there is a higher risk of a supply shortage.
As the level of wind electricity penetration increases, however, assuming that the outputs of wind power plants are positively correlated, the
period of greatest risk will shift to times with low average levels of wind
energy supply (Hasche et al., 2010). Aggregating wind power plants
over larger areas may reduce the correlation between wind power outputs, as described earlier, and can slow the decline in capacity credit
as wind electricity penetration increases, though adequate transmission
capacity is required to aggregate the output of wind power plants in this
way (Tradewind, 2009; EnerNex Corp, 2010).24
The relatively low average capacity credit of wind power plants (compared to fossil fuel-powered units, for example) suggests that systems
with large amounts of wind energy will also tend to have signicantly
more total nameplate generation capacity (wind and non-wind) to meet
the same peak electricity demand than will electric systems without
large amounts of wind energy. Some of this generation capacity will
operate infrequently, however, and the mix of other generation in an
electric system with large amounts of wind energy will tend (on economic grounds) to increasingly shift towards more exible peaking
23 As an example, the addition of a very reliable 100 MW fossil unit in a system with
numerous other reliable units will usually increase the load-carrying capability of the
system by at least 90 MW, leading to a greater than 90% capacity credit for the fossil
unit.
24 Generation resource adequacy evaluations are also beginning to include the capability of the system to provide adequate exibility and operating reserves to accommodate more wind energy (NERC, 2009). The increased demand from wind energy
for operating reserves and exibility is addressed in Section 7.5.3.
563
Capacity Credit [% Nameplate Wind Power Capacity]
Wind Energy
Chapter 7
45
Germany
40
Ireland ESBNG 5GW
35
Ireland ESBNG 6.5GW

Mid Norway 3 Wind Farms
30
Mid Norway 1 Wind Farm
25
UK 2002
20
UK 2007
US California
15
US Minnesota 2004
10
US Minnesota 2006
US New York On-Off-Shore
0
0
10
20
30
40
50
60
Wind Power Penetration [% of Peak Demand]

Figure 7.14 | Estimates of the capacity credit of wind power plants across several wind energy integration studies from Europe and the USA (Holttinen et al., 2009).
and intermediate resources and away from base-load resources (e.g.,

Lamont, 2008; Milborrow, 2009; Boccard, 2010).
7.5.3
Operating electric systems with wind energy
The unique characteristics of wind energy, and especially power output

variability and uncertainty, also hold important implications for electric
system operations. Here we summarize those implications in general
(Section 7.5.3.1), and then briey discuss three specic case studies
of the integration of wind energy into real electricity systems (Section
7.5.3.2).
7.5.3.1
Integration, exibility and variability
Because wind energy is generated with a very low marginal operating

cost, it is typically used to meet demand when it is available, thereby
displacing the use of generators that have higher marginal costs. This
results in electric system operators and markets primarily dispatching
other generators to meet demand minus any available wind energy (i.e.,
net demand).
As wind electricity penetration grows, the variability of wind energy
results in an overall increase in the magnitude of changes in net demand,
564
and also a decrease in the minimum net demand. For example, Figure
7.15 depicts demand and ramp duration curves for Ireland.25 At relatively
low levels of wind electricity penetration, the magnitude of changes in
net demand, as shown in the 15-minute ramp duration curve, is similar to
the magnitude of changes in total demand (Figure 7.15(c)). At higher levels of wind electricity penetration, however, changes in net demand are
greater than changes in total demand (Figure 7.15(d)). Similar impacts on
changes in net demand with increased wind energy have been reported
in the USA (Milligan and Kirby, 2008). The gure also shows that, at
high levels of wind electricity penetration, the magnitude of net demand
across all hours of the year is lower than total demand, and that in some
hours net demand is near or even below zero (Figure 7.15(b)).
As a result of these trends, wholesale electricity prices will tend to decline
when wind power output is high (or is forecast to be high in the case of
day-ahead markets) and transmission interconnection capacity to other
energy markets is constrained, with a greater frequency of low or even
negative prices (e.g., Jnsson et al., 2010; Morales et al., 2011). As with
25 Figure 7.15 presents demand and ramp duration curves for Ireland with (net demand) and without (demand) the addition of wind energy. A demand duration curve
shows the percentage of the year that the demand exceeds a level on the vertical
axis. Demand in Ireland exceeds 4,000 MW, for example, about 10% of the year. The
ramp duration curves show the percentage of the year that changes in the demand
exceed the level on the vertical axis. The 15-min change in demand in Ireland exceeds 100 MW/15minutes, for example, less than 10% of the year.
Chapter 7
Wind Energy
5,000
(c) Ramp Duration Curve

Ramp [MW/15 minutes]
Power [MW]
(a) Demand Duration Curve

Demand
Net Demand
4,000
3,000
400
Demand
Net Demand
300
200
100
0
-100
2,000
-200
-300
1,000
0
0.2
0.4
0.6
0.8
0.2
0.4
0.6
0.8
Probability
(d) Ramp Duration Curve (Projected)

Ramp [MW/15 minutes]
Power [MW]
(b) Demand Duration Curve (Projected)

6,000
Probability
Demand
Net Demand
4,000
2,000
400
Demand
Net Demand
200
-200
-2,000
-400
0
0.2
0.4
0.6
0.8
Probability
0.2
0.4
0.6
0.8
Probability
Figure 7.15 | Demand duration and 15-minute ramp duration curves for Ireland in (a, c) 2008 (wind energy represents 7.5% of total annual average electricity demand), and (b, d)
projected for high wind electricity penetration levels (wind energy represents 40% of total annual average electricity demand).1 Source: Data from www.eirgrid.com.
Note: 1. Projected demand and ramp duration curves are based on scaling 2008 data (demand is scaled by 1.27 and wind energy is scaled on average by 7). Ramp duration curves
show the cumulative probability distributions of 15-minute changes in demand and net demand.
adding any low marginal cost resource to an electric system, increased

wind electricity penetrations will therefore tend to reduce average
wholesale prices in the short term (before changes are made to the mix
of other generation sources) as wind energy displaces power sources
with higher marginal costs. Price volatility will also tend to increase as
the variability and uncertainty in wind power output ensures that wind
energy will not always be available to displace higher marginal cost
generators. In the long run, however, the average effect of wind energy
on wholesale electricity prices is not as clear because the relationships
between investment costs, O&M costs and wholesale price signals will
begin to inuence decisions about the expansion of transmission interconnections, generator retirement and the type of new generation that
is built (Morthorst, 2003; Frsund et al., 2008; Lamont, 2008; Senz
de Miera et al., 2008; Sensfu et al., 2008; Sder and Holttinen, 2008;
MacCormack et al., 2010).
These price impacts are a reection of the fact that increased wind
energy deployment will require some other generating units to operate in a more exible manner than required without wind energy. At
low to medium levels of wind electricity penetration, the increase in
minute-to-minute variability will depend on the exact level of wind
electricity penetration, the degree of geographic smoothing, and electric

system size, but is generally expected to be relatively small and therefore inexpensive to manage in large electric systems (J. Smith et al.,
2007). The more signicant operational challenges relate to the variability and commensurate increased need for exibility to manage changes
in wind power output over one to six hours (Doherty and OMalley,
2005; Holttinen et al., 2009). Incorporating state-of-the-art forecasting
of wind energy over multiple time horizons into electric system operations can reduce the need for exibility from other generators, and has
been found to be especially important as wind electricity penetration
levels increase (e.g., Doherty et al., 2004; Tuohy et al., 2009; GE Energy,
2010). Nonetheless, even with high-quality forecasts and geographically dispersed wind power plants, additional start-ups and shut-downs,
part-load operation, and ramping will be required from fossil generation units to maintain the supply/demand balance (e.g., Gransson and
Johnsson, 2009; Troy et al., 2010).
This additional exibility is not free, as it increases the amount of time
that fossil fuel-powered units are operated at less efcient part-load
conditions (resulting in lower than expected reductions in production
costs and emissions from fossil generators as described in Sections
565
Wind Energy
7.5.4 and 7.6.1.3, respectively), increases wear and tear on boilers and
other equipment, increases maintenance costs, and reduces power plant
life (Denny and OMalley, 2009). Various kinds of economic incentives
can be used to ensure that the operational exibility of other generators is made available to system operators. Some electricity systems,
for example, have day-ahead, intra-day, and/or hour-ahead markets for
electricity, as well as markets for reserves, balancing energy and other
ancillary services. These markets can provide pricing signals for increased
(or decreased) exibility when needed as a result of rapid changes in or
poorly predicted wind power output, and can therefore reduce the cost of
integrating wind energy (J. Smith et al., 2007; Gransson and Johnsson,
2009). Markets with shorter scheduling periods have also been found
to be more responsive to variability and uncertainty, thereby facilitating wind energy integration (Holttinen, 2005; Kirby and Milligan, 2008;
Tradewind, 2009). In addition, coordinated electric system operations
across larger areas has been shown to benet wind energy integration,
and increased levels of wind energy supply may therefore tend to motivate greater investments in and electricity trade across transmission
interconnections (Milligan and Kirby, 2008; Denny et al., 2010). Where
wholesale electricity markets do not exist, other planning methods or
incentives would be needed to ensure that generating plants are exible
enough to accommodate increased deployment of wind energy.
Planning systems and incentives may also need to be adopted to
ensure that new generating plants are sufciently exible to accommodate expected wind energy deployment. Moreover, in addition to
exible fossil fuel-powered units, hydropower stations, bulk energy
storage, large-scale deployment of electric vehicles and their associated
contributions to system exibility through controlled battery charging,
diverting excess wind energy to fuel production or local heating, and
various forms of demand response can also be used to facilitate the integration of wind energy. The deployment of a diversity of RE technologies
may also help facilitate overall electric system integration. The role of
some of these technologies (as well as some of the operational and
planning methods noted earlier) in electric systems is described in more
detail in Chapter 8 because they are not all specic to wind energy and
because some are more likely to be used at higher levels of wind electricity penetration than considered here (up to 20%). Wind power plants,
meanwhile, can provide some exibility by briey curtailing output to
provide downward regulation or, in extreme cases, curtailing output for
extended periods to provide upward regulation. Modern controls on
wind power plants can also use curtailment to limit or even (partially)
control ramp rates (Fox et al., 2007). Though curtailing wind power output is a simple and often times readily available source of exibility,
there are sizable opportunity costs associated with curtailing plants that
have low operating costs before reducing the output of other plants that
have high fuel costs. These opportunity costs should be compared to the
possible benets of curtailment (e.g., reduced part-load efciency penalties and wear and tear for fossil generators, and avoidance of certain
transmission investments) when determining the prevalence of its use.
566
Chapter 7
7.5.3.2
Practical experience with operating electric systems

with wind energy
Actual operating experience in different parts of the world demonstrates that electric systems can operate reliably with increased
contributions of wind energy (Sder et al., 2007). In four countries, as
discussed earlier, wind energy in 2010 was already able to supply from
10 to roughly 20% of annual electricity demand. The three examples
reported here demonstrate the challenges associated with this operational integration, and the methods used to manage the additional
variability and uncertainty associated with wind energy. Naturally,
these impacts and management methods vary across regions for reasons of geography, electric system design and regulatory structure,
and additional examples of wind energy integration associated with
operations, curtailment and transmission are described in Chapter 8.
Moreover, as more wind energy is deployed in diverse regions and electric systems, additional knowledge about the impacts of wind power
output on electric systems will be gained. To date, for example, there
is little experience with severe contingencies (i.e., faults) during times
with high instantaneous wind electricity penetration. Though existing
experience demonstrates that electric systems can operate with wind
energy, further analysis is required to determine whether electric systems are maintaining the same level of overall security, measured by
the ability of the system to withstand major contingencies, with and
without wind energy, and depending on various management options.
Limited analysis (e.g., EirGrid and SONI, 2010; Eto et al., 2010) suggests that particular systems are able to survive such conditions but,
if primary frequency control reserves are reduced as thermal generation is increasingly displaced by wind energy, additional management
options may be needed to maintain adequate frequency response. The
security of the electric system with high instantaneous wind electricity
penetrations is described in more detail in Chapter 8.
Denmark has the highest wind electricity penetration of any country in
the world, with wind energy supply equating to approximately 20% of
total annual electricity demand. Total wind power capacity installed by
the end of 2009 equalled 3.4 GW, while the peak demand in Denmark
was 6.5 GW. Much of the wind power capacity (2.7 GW) is located
in western Denmark, resulting in instantaneous wind power output
exceeding total demand in western Denmark in some instances (see
Figure 7.16). The Danish example demonstrates the benets of having
access to markets for exible resources and having strong transmission interconnections to neighbouring countries. Denmarks electricity
systems operate without serious reliability issues in part because the
country is well interconnected to two different electric systems. In conjunction with wind power output forecasting, this allows wind energy to
be exported to other markets and helps the Danish operators manage
wind power variability. The interconnection with the Nordic system, in
particular, provides access to exible hydropower resources, and balancing the Danish system is much more difcult during periods when
Chapter 7
Wind Energy
5,000
101%
100
98%
Demand
Wind
% Wind
4,000
3,000
90
80
70
60
50
2,000
40
Wind Penetration [%]
Demand and Wind [MW]
The Electric Reliability Council of Texas (ERCOT) operates a synchronous system with a peak demand of 63 GW and 8.5 GW of wind
power capacity, and with a wind electricity penetration level of 6%
of annual electricity demand by the end of 2009. ERCOTs experience
7.5.4
Results from integration studies
In addition to actual operating experience, a number of high-quality studies of the increased transmission and generation resources required to
accommodate wind energy have been completed, primarily covering
OECD countries. As summarized further below, these studies employ
a wide variety of methodologies and have diverse objectives, but typically seek to evaluate the capability of the electric system to integrate
increased penetrations of wind energy and to quantify the costs and
benets of operating the system with wind energy. The issues and costs
often considered by these studies are reviewed in this section, and
include: the increased operating reserves and balancing costs required
5,000
42%
4,000
40
3,000
30
Demand
2,000
20
20
Wind
% Wind
30
1,000
50
1,000
10
10
0
0
10/01
11/01
12/01
13/01
14/01
15/01
16/01
0
04-Apr-10
05-Apr-10
Figure 7.16 | Wind energy, electricity demand and instantaneous penetration levels in (left) West Denmark for a week in January 2005, and (right) the island of Ireland for two days
in April 2010. Source: Data from (left) www.energinet.dk; (right) www.eirgrid.com and System Operator for Northern Ireland.
567
Wind Penetration [%]
In contrast to the strong interconnections of the Danish system with

other electric systems, the island of Ireland has a single synchronous
system; its size is similar to the Danish system but interconnection
capacity with other markets is limited to a single 500 MW high-voltage
direct current link.The wind power capacity installed by the end of 2009
was capable of supplying roughly 11% of Irelands annual electricity
demand, and the Irish system operators have successfully managed
that level of wind electricity penetration.The large daily variation in
electricity demand in Ireland, combined with the isolated nature of the
Irish system, has resulted in a relatively exible electric system that is
particularly well suited to integrating wind energy; exible natural gas
plants generated 65% of the electrical energy in the rst half of 2010.
As a result, despite the lack of signicant interconnection capacity,
the Irish system has successfully operated with instantaneous levels of
wind electricity penetration of over 40% (see Figure 7.16). Nonetheless,
it is recognized that as wind electricity penetration levels increase further, new challenges will arise. Of particular concern are: the possible
lack of inertial response of wind turbines absent additional turbine
controls, which could lead to increased frequency excursions during
severe grid contingencies (Lalor et al., 2005); the need for even greater
exibility to maintain supply-demand balance; and the need to build
additional high-voltage transmission (AIGS, 2008). Moreover, in common with the Danish experience, much of the wind energy is and will
be connected to the distribution system, requiring attention to voltage
control issues (Vittal et al., 2010). Figure 7.16 illustrates the high levels
of instantaneous wind electricity penetration that exist in Ireland and
West Denmark.
demonstrates the importance of incorporating wind energy forecasts

into system operations, and the need to schedule adequate reserves
to accommodate system uncertainty. On 26 February 2008, a combination of factors, not all related to wind energy, led ERCOT to
implement its emergency curtailment plan, which included the curtailment of 1,200 MW of demand that was voluntarily participating in
ERCOTs Load Acting as a Resource program. The factors involved
in the event included wind energy scheduling errors, an incorrect
day-ahead electricity demand forecast, and an unscheduled outage
of a fossil fuel power plant. With regards to the role of wind energy,
ERCOT experienced a decline in wind power output of 1,500 MW over
a three-hour period on that day, roughly 30% of the 5 GW of installed
wind power capacity in February 2008 (Ela and Kirby, 2008; ERCOT,
2008). The event was exacerbated by the fact that scheduling entitieswhich submit updated resource schedules to ERCOT one hour
prior to the operating hourconsistently reported an expectation of
more wind power output than actually occurred. A state-of-the-art
forecast was available, but was not yet integrated into ERCOT system
operations, and that forecast predicted the wind energy event much
more accurately. As a result of this experience, ERCOT accelerated
its schedule for incorporating the advanced wind energy forecasting
system into its operations.
Demand and Wind [MW]
one of the interconnections is down. Even more exibility is expected to

be required, however, if Denmark markedly increases its penetration of
wind electricity (Ea Energianalyse, 2007).
Wind Energy
Chapter 7
to accommodate the variability and uncertainty in net demand caused by

wind energy; the requirement to maintain sufcient generation adequacy;
and the possible need for additional transmission infrastructure. The studies also frequently analyze the benets of adding wind energy, including
avoided fossil fuel consumption and CO2 emissions, though these benets are not reviewed in this section. This section focuses on the general
results of these studies as a whole; see Chapter 8 for brief descriptions
of individual study results, including some studies that have investigated
somewhat higher levels of wind electricity penetration than considered
here.
integrating up to 20% wind energy into electric systems is, in most cases,
modest but not insignicant. Specically, at low to medium levels of wind
electricity penetration (up to 20% wind energy), the available literature
(again, primarily from a subset of OECD countries) suggests that the
additional costs of managing electric system variability and uncertainty,
ensuring generation adequacy and adding new transmission to accommodate wind energy will be system specic but generally in the range
of US cents2005 0.7 to 3/kWh.26 Concerns about (and the costs of) wind
energy integration will grow with wind energy deployment and, even at
lower penetration levels, integration issues must be actively managed.
7.5.4.1
7.5.4.2
Methodological challenges
Estimating the incremental impacts and costs of wind energy integration is difcult due to the complexity of electric systems and study data
requirements. One of the most signicant challenges in executing these
studies is simulating wind power output data at high time resolutions
for a chosen future wind electricity penetration level and for a sufcient
duration for the results of the analysis to accurately depict worst-case
conditions and correlations of wind and electricity demand. These data
are then used in electric system simulations to mimic system planning
and operations, thereby quantifying the impacts, costs and benets of
wind energy integration.
Addressing all integration impacts requires several different simulation models that operate over different time scales, and most individual
studies therefore focus on a subset of the potential issues. The results
of wind energy integration studies are also dependent on pre-existing
differences in electric system designs and regulatory environments:
important differences include generation capacity mix and the exibility of that generation, the variability of demand and the strength
and breadth of the transmission system. In addition, study results differ
and are hard to compare because standard methodologies and even
denitions have not been developed, though signicant progress has
been made in developing agreement on many high-level study design
principles (Holttinen et al., 2009). The rst-generation integration studies, for example, used models that were not designed to fully reect
the variability and uncertainty of wind energy, resulting in studies that
addressed only parts of the larger system. More recent studies, on the
other hand, have used models that can incorporate the uncertainty of
wind power output from the day-ahead time scale to some hours ahead
of delivery (e.g., Meibom et al., 2009; Tuohy et al., 2009). Integration
studies are also increasingly simulating high wind electricity penetration
scenarios over entire synchronized systems (not just individual, smaller
balancing areas) (e.g., Tradewind, 2009; EnerNex Corp, 2010; GE Energy,
2010). Finally, only recently have studies begun to explore in more depth
the capability of electric systems to maintain primary frequency control during system contingencies with high penetrations of wind energy
(e.g., EirGrid and SONI, 2010; Eto et al., 2010).
Regardless of the challenges of executing and comparing such studies, the
results, as described in more detail below, demonstrate that the cost of
568
Increased balancing cost with wind energy
The additional variability and uncertainty in net demand caused by

increased wind energy supply results in higher balancing costs, in part
due to increases in the amount of short-term reserves procured by system operators. A number of signicant integration studies from Europe
and the USA have concluded that accommodating wind electricity penetrations of up to (and in a limited number of cases, exceeding) 20% is
technically feasible, but not without challenges (R. Gross et al., 2007;
J. Smith et al., 2007; Holttinen et al., 2009; Milligan et al., 2009). The
estimated increase in short-term reserve requirements in eight studies summarized by Holttinen et al. (2009) has a range of 1 to 15% of
installed wind power capacity at 10% wind electricity penetration,
and 4 to 18% of installed wind power capacity at 20% wind electricity penetration. Those studies that predict a need for higher levels of
reserves generally assume that day-ahead uncertainty and/or multi-hour
variability of wind power output is handled with short-term reserves.
In contrast, markets that are optimized for wind energy will generally
be designed so that additional opportunities to balance supply and
demand exist, reducing the reliance on more expensive short-term
reserves (e.g., Weber, 2010). Notwithstanding the differences in results
and methods, however, the studies reviewed by Holttinen et al. (2009)
nd that, in general, wind electricity penetrations of up to 20% can be
accommodated with increased balancing costs of roughly US cents 0.14
to 0.56/kWh27 of wind energy generated (Figure 7.17). State-of-the-art
wind energy forecasts are often found to be a key factor in minimizing
the impact of wind energy on market operations. Although denitions
and methodologies for calculating increased balancing costs differ, and
several open issues remain in estimating these costs, similar results are
reported by R. Gross et al. (2007), J. Smith et al. (2007), and Milligan et
al. (2009).
26 This cost range is based on the assumption that there may be electric systems where
all three cost components (balancing costs, generation adequacy costs and transmission costs) are simultaneously at the low end of the range reported for each of these
costs in the literature or conversely where all three cost components are simultaneously at the high end of the range. As reported below, the cost range for managing
wind energys variability and uncertainty (US cents2005 0.14 to 0.56/kWh), ensuring
generation adequacy (US cents2005 0.58 to 0.96/kWh), and adding new transmission
(US cents2005 0 to 1.5/kWh) sums to roughly US cents2005 0.7 to 3/kWh. Using a
somewhat similar approach, IEA (2010b) developed estimates that are also broadly
within this range.
27 Conversion to 2005 dollars is not possible given the range of study-specic assumptions.
Chapter 7
7.5.4.3
Wind Energy
Relative cost of generation adequacy with wind energy
Increase in Balancing Cost [UScent/kWh Wind]
The benets of adding a wind power plant to an electric system are

often compared to the benets of a base-load, or fully utilized, plant that
generates an equivalent amount of energy on an annual basis (a comparator plant). The comparator plant is typically assumed to have a high
capacity credit, close to 100% of its nameplate capacity. Wind energy,
on the other hand, was shown in Section 7.5.2.4 to have a capacity
credit of 5 to 40% of its nameplate capacity. The resulting contribution
of the wind plant to generation adequacy is therefore often lower than
the contribution of an energy-equivalent comparator plant per unit of
energy generated, and wind energy is typically less valuable than the
comparator plant from the perspective of meeting generation adequacy
targets. Using this framework, R. Gross et al. (2007) estimate that the
difference between the contribution to generation adequacy of a wind
power plant and an energy-equivalent base-load plant can result in a
US cents2005 0.58 to 0.96/kWh generation adequacy cost for wind energy
relative to a comparator plant at wind electricity penetration levels up to
to electricity demand, the geographic distribution of wind power plant

siting and the level of wind electricity penetration will all impact the
capacity credit estimated for wind energy, and therefore the relative
cost of generation adequacy.
7.5.4.4
Cost of transmission for wind energy
Finally, a number of assessments of the need for and cost of upgrading or building large-scale transmission infrastructure between wind
resource regions and demand centres have similarly found modest, but
not insignicant, costs.28 The transmission cost for achieving 20% wind
electricity penetration in the USA, for example, was estimated to add
about USD2005 150 to 290/kW to the investment cost of wind power
plants (US DOE, 2008). The cost of this transmission expansion was
found to be justied because of the higher quality of the wind resources
accessed if the transmission were to be built relative to accessing only
lower-quality wind resources with less transmission expansion. More
0.6
Nordic 2004
Finland 2004
0.5
UK, 2002
UK, 2007
0.4
Ireland
Colorado US
Minnesota 2004
0.3
Minnesota 2006
California US
Pacicorp US
0.2
Greennet Germany
Greennet Denmark
Greennet Finland
0,1
Greennet Norway
Greennet Sweden
0.0
0
10
15
20
25
30
Wind Electricity Penetration [% of Annual Electricity Demand]

Figure 7.17 | Estimates of the increase in balancing costs due to wind energy from several wind energy integration studies in Europe and the USA (Holttinen et al., 2009).1
Note: 1. Conversion to 2005 dollars is not possible given the range of study-specic assumptions.
20%. Using a somewhat different approach, Boccard (2010) provides a

comparable estimate of the generation adequacy cost of wind energy in
several European countries. As discussed earlier, the methodology used
to assess generation adequacy, the correlation of wind power output
28 These costs are distinct from the costs to connect individual wind power plants to the
transmission system; connection costs are often included in estimates of the investment costs of wind power plants (see Section 7.8).
569
Wind Energy
Chapter 7
detailed assessments of the transmission needed to accommodate

increased wind energy deployment in the USA have found a wide range
of results, with estimated costs ranging from very low to sometimes
reaching (or even exceeding) USD2005 400/kW (JCSP, 2009; Mills et al.,
2009a; EnerNex Corp, 2010). Large-scale transmission for cases with
increased wind energy has also been considered in Europe (Czisch and
Giebel, 2000) and China (Lew et al., 1998). Results from country-specic
transmission assessments in Europe have resulted in varied estimates of
the cost of new large-scale transmission; Auer et al. (2004) and EWEA
(2005) identied transmission costs for a number of European studies,
with cost estimates that are somewhat lower than those found in the
USA. Holttinen et al. (2009) reviewed wind energy transmission costs
from several European national case studies, and found costs ranging
from USD2005 0/kW to as high as USD2005 310/kW.
potential to produce some detrimental impacts on the environment and

on human activities and well-being, and many local and national governments have established planning, permitting and siting requirements
to reduce those impacts. These potential concerns need to be taken into
account to ensure a balanced view of the advantages and disadvantages
of wind energy, especially if wind energy is to expand on a large scale.
Transmission expansion for wind energy can be justied by the reduction in congestion costs that would occur for the same level of wind
energy deployment without transmission expansion. A European-wide
study, for example, identied several transmission upgrades between
nations and between high-quality offshore wind resource areas that
would reduce transmission congestion and ease wind energy integration (Tradewind, 2009). The avoided congestion costs associated with
transmission expansion were similarly found to justify transmission
investments in two US-based detailed integration studies of high wind
electricity penetrations (Milligan et al., 2009). At the same time, it is not
always appropriate to fully assign the cost of transmission expansion to
wind energy deployment. In some cases, these transmission expansion
costs can be justied for reasons beyond wind energy, as new transmission can have wider benets including increased electricity reliability,
decreased pre-existing congestion and reduced market power (Budhraja
et al., 2009). Moreover, wind energy is not unique in potentially requiring new transmission investment; other energy technologies may also
require new transmission, and the costs summarized above do not all
represent truly incremental costs.
7.6.1
Notwithstanding these important caveats, at the higher end of the range

from the available literature (USD2005 400/kW), transmission expansion
costs add roughly US cents2005 1.5/kWh to the levelized cost of wind
energy. At the lower end, effectively no new transmission costs would
need to be specically assigned to the support of wind energy.
7.6
Wind energy has signicant potential to reduce (and already is reducing)

GHG emissions, together with the emissions of other air pollutants, by
displacing fossil fuel-based electricity generation. Because of the commercial readiness (Section 7.3) and cost (Section 7.8) of the technology,
wind energy can be immediately deployed on a large scale (Section 7.9).
As with other industrial activities, however, wind energy also has the
29 A comprehensive assessment of social and environmental impacts of all RE sources
570
This section summarizes the best available knowledge about the most
relevant environmental net benets of wind energy (Section 7.6.1),
while also addressing ecological impacts (Section 7.6.2), impacts on
human activities and well-being (Section 7.6.3), public attitudes and
acceptance (Section 7.6.4) and processes for minimizing social and environmental concerns (Section 7.6.5).
Environmental net benets of wind energy
The environmental benets of wind energy come primarily from displacing the emissions from fossil fuel-based electricity generation. However,
the manufacturing, transport, installation, operation and decommissioning of wind turbines induces some indirect negative effects, and the
variability of wind power output also impacts the operations and emissions of fossil fuel-red plants. Such effects need to be subtracted from
the gross benets of wind energy in order to estimate net benets. As
shown below, these latter effects are modest compared to the net GHG
reduction benets of wind energy.
7.6.1.1
Direct impacts
The major environmental benets of wind energy (as well as other forms
of RE) result from displacing electricity generation from fossil fuel-based
power plants, as the operation of wind turbines does not directly emit
GHGs or other air pollutants. Similarly, unlike some other generation
sources, wind energy requires insignicant amounts of water, produces
little waste and requires no mining or drilling to obtain its fuel supply
(see Chapter 9).
Estimating the environmental benets of wind energy is somewhat
complicated by the operational characteristics of the electric system and
the decisions that are made about investments in new power plants to
economically meet electricity demand (Deutsche Energie-Agentur, 2005;
NRC, 2007; Pehnt et al., 2008). In the short run, increased wind energy
will typically displace the operations of existing fossil fuel-based plants
that are otherwise on the margin. In the longer term, however, new
generating plants may be needed, and the presence of wind energy can
inuence what types of power plants are built; specically, increased
wind energy will tend to favour on economic grounds exible peaking/
intermediate plants that operate less frequently over base-load plants
(Kahn, 1979; Lamont, 2008). Because the impacts of these factors are
both complicated and system specic, the benets of wind energy will
also be system specic and are difcult to forecast with precision.
Chapter 7
Wind Energy
Planetary boundary layer research is important for accurately determining wind ow and turbulence in the presence of various atmospheric
stability effects and complex land surface characteristics. Research in
mesoscale atmospheric processes aims at improving the fundamental
understanding of mesoscale and local wind ows (Banta et al., 2003;
Kelley et al., 2004). In addition to its contribution towards understanding
turbine-level aerodynamic and array wake effects, a better understanding of mesoscale atmospheric processes will yield improved wind energy
resource assessments and forecasting methods. Physical and statistical
modelling to resolve spatial scales in the 100- to 1,000-m range, a notable gap in current capabilities (Wyngaard, 2004), could occupy a central
role of this research.
and summarizes forecasts of the potential for further cost reductions

(Section 7.8.4). The economic competitiveness of wind energy in comparison to other energy sources, which necessarily must also include
other factors such as subsidies and environmental externalities, is not
covered in this section.42 Moreover, the focus in this section is on wind
energy generation costs; the costs of integration and transmission are
generally not covered here, but are instead discussed in Section 7.5,
though costs associated with grid connection are sometimes included in
the investment cost gures presented in this section.
Finally, additional research is warranted on the interaction between

global and local climate effects, and wind energy. Specically, work is
needed to identify and understand historical trends in wind resource
variability in order to increase the reliability of future wind energy performance predictions. As discussed earlier in this chapter, further work
is also warranted on the possible impacts of climate change on wind
energy resource conditions, and on the impact of wind energy development on local, regional and global climates.
The levelized cost of energy from on- and offshore wind power plants is
affected by ve primary factors: annual energy production, investment
costs, O&M costs, nancing costs and the assumed economic life of the
plant.43 Available support policies can also inuence the cost (and price)
of wind energy, as well as the cost of other electricity supply options, but
these factors are not addressed here.
Signicant progress in many of the above areas requires interdisciplinary

research. Also crucial is the need to use experiments and observations in
a coordinated fashion to support and validate computation and theory.
Models developed in this way will help improve: (1) wind turbine design;
(2) wind power plant performance estimates; (3) wind resource assessments; (4) short-term wind energy forecasting; and (5) estimates of the
impact of large-scale wind energy deployment on the local climate, as
well as the impact of potential climate change effects on wind resources.
7.8
Cost trends41
Though the cost of wind energy has declined signicantly since the
1980s, policy measures are currently required to ensure rapid deployment in most regions of the world (e.g., NRC, 2010b). In some areas
with good wind resources, however, the cost of wind energy is competitive with current energy market prices (e.g., Berry, 2009; IEA, 2009; IEA
and OECD, 2010). Moreover, continued technology advances in on- and
offshore wind energy are expected (Section 7.7), supporting further cost
reductions. The degree to which wind energy is utilized globally and
regionally will depend largely on the economic performance of wind
energy compared to alternative power sources.
This section describes the factors that affect the cost of wind energy
(Section 7.8.1), highlights historical trends in the cost and performance
of wind power plants (Section 7.8.2), summarizes data and estimates
the levelized generation cost of wind energy in 2009 (Section 7.8.3),
41 Discussion of costs in this section is largely limited to the perspective of private
investors. Chapters 1 and 8 to 11 offer complementary perspectives on cost issues
covering, for example, costs of integration, external costs and benets, economywide costs and costs of policies.
7.8.1
Factors that affect the cost of wind energy
The nature of the wind resource, which varies geographically and temporally, largely determines the annual energy production from a prospective
wind power plant, and is among the most important economic factors
(Burton et al., 2001). Precise micro-siting of wind power plants and even
individual turbines is critical for maximizing energy production. The trend
towards turbines with larger rotor diameters and taller towers has led to
increases in annual energy production per unit of installed capacity, and
has also allowed wind power plants in lower-resource areas to become
more economically competitive. Larger wind power plants, meanwhile,
have led to consideration of array effects whereby the production of
downwind turbines is affected by those turbines located upwind. Offshore
power plants will, generally, be exposed to better wind resources than will
onshore plants (EWEA, 2009).
Wind power plants are capital intensive and, over their lifetime, the initial
investment cost ranges from 75 to 80% of total expenditure, with O&M
costs contributing the balance (Blanco, 2009; EWEA, 2009). The investment cost includes the cost of the turbines (turbines, transportation to site,
and installation), grid connection (cables, sub-station, connection), civil
works (foundations, roads, buildings), and other costs (engineering, licensing, permitting, environmental assessments and monitoring equipment).
Table 7.4 shows a rough breakdown of the investment cost components
for modern wind power plants. Turbine costs comprise more than 70%
of total investment costs for onshore wind power plants. The remaining
investment costs are highly site-specic. Offshore wind power plants are
dominated by these other costs, with the turbines often contributing less
than 50% of the total. Site-dependent characteristics such as water depth
and distance to shore signicantly affect grid connection, civil works and
42 The environmental impacts and costs of RE and non-RE sources are summarized in
Chapters 9 and 10, respectively.
43 Decommissioning costs also exist, but are not expected to be sizable in most instances.
583
Wind Energy
Chapter 7
Table 7.4 | Investment cost distribution for on- and offshore wind power plants (Data sources: Blanco, 2009; EWEA, 2009).
Cost Component
Offshore (%)1
Onshore (%)
Turbine
7176
3749
Grid connection
1012
2123
Civil works
79
2125
Other investment costs
58
915
Note: 1. Offshore cost categories consolidated from original study.
other costs. Offshore turbine foundations and internal electric grids are
also considerably more costly than those for onshore power plants.
The O&M costs of wind power plants include xed costs such as land
leases, insurance, taxes, management, and forecasting services, as well as
variable costs related to the maintenance and repair of turbines, including
spare parts. O&M comprises approximately 20% of total wind power plant
expenditure over a plants lifetime (Blanco, 2009), with roughly 50% of
total O&M costs associated directly with maintenance, repair and spare
parts (EWEA, 2009). O&M costs for offshore wind energy are higher than
for onshore due to the less mature state of technology as well as the
challenges and costs of accessing offshore turbines, especially in harsh
weather conditions (Blanco, 2009).
Financing arrangements, including the cost of debt and equity and the
proportional use of each, can also inuence the cost of wind energy, as
can the expected operating life of the wind power plant. For example,
ownership and nancing structures have evolved in the USA that minimize
the cost of capital while taking advantage of available incentives (Bolinger
et al., 2009). Other research has found that the predictability of the policy
measures supporting wind energy can have a sizable impact on nancing
costs, and therefore the ultimate cost of wind energy (Wiser and Pickle,
1998; Dinica, 2006; Dunlop, 2006; Agnolucci, 2007). Because offshore
wind power plants are still relatively new, with greater performance risk,
higher nancing costs are experienced than for onshore plants (Dunlop,
2006; Blanco, 2009), and larger rms tend to dominate offshore wind
energy development and ownership (Markard and Petersen, 2009).
7.8.2
Historical trends
7.8.2.1
Investment costs
From the beginnings of commercial wind energy deployment to roughly

2004, the average investment costs of onshore wind power plants
dropped, while turbine size grew signicantly.44 With each generation
44 Investment costs presented here and later in Section 7.8 (as well as all resulting levelized cost of energy estimates) generally include the cost of the turbines (turbines,
transportation to site and installation), grid connection (cables, sub-station, connection, but not more general transmission expansion costs), civil works (foundations,
roads, buildings), and other costs (engineering, licensing, permitting, environmental
assessments, and monitoring equipment). Whether the cost of connecting to the grid
is included varies by data source, and is sometimes unclear; costs associated with
strengthening the backbone transmission system are generally excluded.
584
of wind turbine technology during this period, design improvements

and turbine scaling led to decreased investment costs. Historical investment cost data from Denmark and the USA demonstrate this trend
(Figure 7.20). From 2004 to 2009, however, investment costs increased.
Some of the reasons behind these increased costs are described in
Section 7.8.3.
There is far less experience with offshore wind power plants, and the
investment costs of offshore plants are highly site-specic. Nonetheless,
the investment costs of offshore plants have historically been 50 to
more than 100% higher than for onshore plants (BWEA and Garrad
Hassan, 2009; EWEA, 2009). Moreover, offshore wind power plants
built to date have generally been constructed in relatively shallow
water and relatively close to shore (see Section 7.3); higher costs would
be experienced for deeper water and more distant facilities. Figure
7.21 presents investment cost data for operating and announced
offshore wind power plants. Offshore costs have been inuenced
by some of the same factors that caused rising onshore costs from
2004 through 2009 (as well as several unique factors), as described in
Section 7.8.3, leading to a doubling of the average investment cost of
offshore plants from 2004 through 2009 (BWEA and Garrad Hassan,
2009; UKERC, 2010).
7.8.2.2
Operation and maintenance
Modern turbines that meet IEC standards are designed for a 20-year
life, and plant lifetimes may exceed 20 years if O&M costs remain at
an acceptable level. Few wind power plants were constructed 20 or
more years ago, however, and there is therefore limited experience in
plant operations over this entire time period (Echavarria et al., 2008).
Moreover, those plants that have reached or exceeded their 20-year
lifetime tend to have turbines that are much smaller and less sophisticated than their modern counterparts. Early turbines were also designed
using more conservative criteria, though they followed less stringent
standards than todays designs. As a result, early plants only offer limited guidance for estimating O&M costs for more recent turbine designs.
In general, O&M costs during the rst couple of years of a wind power
plants life are covered, in part, by manufacturer warranties that are
included in the turbine purchase, resulting in lower ongoing costs than
in subsequent years. Newer turbine models also tend to have lower initial O&M costs than older models, with maintenance costs increasing
Investment Cost [USD2005 /kW]
Chapter 7
Wind Energy
5,000
4,500
Average Project Investment Cost
4,000
3,500
3,000
2,500
2,000
1,500
1,000
500
2006
2007
2008
2006
2007
2008
2009
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
2005
1982
5,000
4,500
Individual Project Investment Cost

Capacity-Weighted Average Project Investment Cost
4,000
Polynomial Trend Line
3,500
3,000
2,500
2,000
1,500
1,000
500
2009
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
Figure 7.20. Investment cost of onshore wind power plants in (upper panel) Denmark (Data source: Nielson et al., 2010) and (lower panel) the USA (Wiser and Bolinger, 2010).
as turbines age (Blanco, 2009; EWEA, 2009; Wiser and Bolinger, 2010).
Offshore wind power plants have historically incurred higher O&M costs
than onshore plants (Junginger et al., 2004; EWEA, 2009; Lemming et al.,
2009).
7.8.2.3
Energy production
The performance of wind power plants is highly site-specic, and is primarily governed by the characteristics of the local wind regime, which varies
geographically and temporally. Wind power plant performance is also
impacted by wind turbine design optimization, performance, and availability, however, and by the effectiveness of O&M procedures. Improved
resource assessment and siting methodologies developed in the 1970s
and 1980s played a major role in improved wind power plant productivity.
Advances in wind energy technology, including taller towers and larger
rotors, have also contributed to increased energy capture (EWEA, 2009).
Though plant-level capacity factors vary widely, data on average eetwide capacity factors45 for a large sample of onshore wind power
plants in the USA show a trend towards higher average capacity factors over time, as wind power plants built more recently have higher
45 A wind power plants capacity factor is only a partial indicator of performance
(EWEA, 2009). Most turbine manufacturers supply variations on a given generator
capacity with multiple rotor diameters and hub heights. In general, for a given generator capacity, increasing the hub height, the rotor diameter, or the average wind
speed will result in an increased capacity factor. When comparing different wind
turbines, however, it is possible to increase annual energy capture by using a larger
generator, while at the same time decreasing the capacity factor.
585
Wind Energy
Chapter 7
7,000
Investment Cost for Operating European Project
6,000
Announced Investment Cost for Proposed U.S. Project

Announced Investment Cost for Proposed European Project
Capacity-Weighted Average Project Investment Cost
5,000
4,000
3,000
2,000
1,000
0
1990
1995
2000
2005
2010
2015
Figure 7.21 | Investment cost of operating and announced offshore wind power plants (Musial and Ram, 2010).
average capacity factors than those built earlier (Figure 7.22). Higher
hub heights and larger rotor sizes are primarily responsible for these
improvements, as the more recent wind power plants built in this time
period and included in Figure 7.22 were, on average, sited in relatively
lower-quality wind resource regimes.
Capacity Factor [%]
Using a different metric for wind power plant performance, annual

energy production per square meter of swept rotor area (kWh/m2) for a
given wind resource site, improvements of 2 to 3% per year over the last
15 years have been documented (IEA, 2008; EWEA, 2009).
7.8.3
Current conditions
7.8.3.1
Investment costs
The investment costs for onshore wind power plants installed worldwide
in 2009 averaged approximately USD2005 1,750/kW, with many plants
falling in the range of USD2005 1,400 to 2,100/kW (Milborrow, 2010);
data in IEA Wind (2010) are reasonably consistent with this range. Wind
power plants installed in the USA in 2009 averaged USD2005 1,900/
kW (Wiser and Bolinger, 2010). Costs in some markets were lower: for
35
30
25
20
15
10
5
0
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
Projects:
11
17
77
124
137
163
190
217
258
299
242
MW:
701
1.158
2.199
3.955
4.458
5.784
6.467
9.289
11.253
16.068
22.346
Figure 7.22 | Fleet-wide average capacity factors for a large sample of wind power plants in the USA from 1999 to 2009 (Wiser and Bolinger, 2010).
586
Chapter 7
Wind Energy
example, average investment costs in China in 2008 and 2009 were

around USD2005 1,000 to 1,350/kW, driven in part by the dominance of
several Chinese turbine manufacturers serving the market with lowercost wind turbines (China Renewable Energy Association, 2009; Li and
Ma, 2009; Li, 2010).
have been built in relatively shallow water. Offshore plants built in

deeper waters, which are becoming increasingly common and are
partly reected in the costs for announced plants, will have relatively
higher costs.
Wind power plant investment costs rose from 2004 to 2009 (Figure 7.20),
an increase primarily caused by the rising price of wind turbines (Wiser
and Bolinger, 2010). Those price increases have been attributed to a number of factors. Increased rotor diameters and hub heights have enhanced
the energy capture of modern wind turbines, for example, but those
performance improvements have come with increased turbine costs,
measured on a dollar per kW basis. The costs of raw materials, including steel, copper, cement, aluminium and carbon bre, also rose sharply
from 2004 through mid-2008 as a result of strong global economic
growth. The strong demand for wind turbines over this period also
put upward pressure on labour costs, and enabled turbine manufacturers and their component suppliers to boost prot margins. Strong
demand, in excess of available supply, also placed particular pressure on critical components such as gearboxes and bearings (Blanco,
2009). Moreover, because many of the wind turbine manufacturers
have historically been based in Europe, and many of the critical components have similarly been manufactured in Europe, the relative
value of the Euro compared to other currencies also contributed to the
wind turbine price increases in certain countries. Turbine manufacturers and component suppliers responded to the tight supply over this
period by expanding or adding new manufacturing facilities. Coupled
with reductions in materials costs that began in late 2008 as a result
of the global nancial crisis, these trends began to moderate wind
turbine prices in 2009 (Wiser and Bolinger, 2010).
7.8.3.2
Due to the relatively small number of operating offshore wind

power plants, investment cost data are sparse. Nonetheless, the
average cost of offshore wind power plants is considerably higher
than that for onshore plants, and the factors that have increased
the cost of onshore plants have similarly affected the offshore sector. The limited availability of turbine manufacturers supplying the
offshore market and of vessels to install such plants exacerbated
cost increases since 2004, as has the installation of offshore plants
in increasingly deeper waters and farther from shore, and the erce
competition among industry players for early-year (before 2005) demonstration plants (BWEA and Garrad Hassan, 2009; UKERC, 2010).
As a result, offshore wind power plants over 50 MW in size, either
built between 2006 and 2009 or planned for the early 2010s, had
investment costs that ranged from approximately USD2005 2,000 to
5,000/kW (BWEA and Garrad Hassan, 2009; IEA, 2009; Snyder and
Kaiser, 2009a; Musial and Ram, 2010). The most recently installed
or announced plants cluster towards the higher end of this range,
from USD2005 3,200 to 5,000/kW (Milborrow, 2010; Musial and Ram,
2010; UKERC, 2010). These investment costs are roughly 100% higher
than costs seen from 2000 to 2004 (BWEA and Garrad Hassan, 2009;
Musial and Ram, 2010; UKERC, 2010). Notwithstanding the increased
water depth of offshore plants, the majority of the operating plants
Operation and maintenance
Though xed O&M costs such as insurance, land payments and routine maintenance are relatively easy to estimate, variable costs such as
repairs and spare parts are more difcult to predict (Blanco, 2009). O&M
costs can vary by wind power plant, turbine type and age, and the availability of a local servicing infrastructure, among other factors. Levelized
O&M costs for onshore wind energy are often estimated to range from
US cents2005 1.2 to 2.3/kWh (Blanco, 2009); these gures are reasonably
consistent with costs reported in EWEA (2009), IEA (2010c), Milborrow
(2010), and Wiser and Bolinger (2010).
Limited empirical data exist on O&M costs for offshore wind energy,
due in large measure to the limited number of operating plants and the
limited duration of those plants operation. Reported or estimated O&M
costs for offshore plants installed since 2002 range from US cents2005 2 to
4/kWh (EWEA, 2009; IEA, 2009, 2010c; Lemming et al., 2009; Milborrow,
2010; UKERC, 2010).
7.8.3.3
Energy production
Onshore wind power plant performance varies substantially, with capacity factors ranging from below 20 to more than 50% depending largely
on local resource conditions. Among countries, variations in average performance also reect differing wind resource conditions, as well as any
difference in the wind turbine technology that is deployed: the average
capacity factor for Germanys installed plants has been estimated at
20.5% (BTM, 2010); European country-level average capacity factors
range from 20 to 30% (Boccard, 2009); average capacity factors in China
are reported at roughly 23% (Li, 2010); average capacity factors in India
are reported at around 20% (Goyal, 2010); and the average capacity factor for US wind power plants is above 30% (Wiser and Bolinger, 2010).
Offshore wind power plants often experience a narrower range in capacity
factors, with a typical range of 35 to 45% for the European plants installed
to date (Lemming et al., 2009); some offshore plants in the UK, however,
have experienced capacity factors of roughly 30%, in part due to relatively
high component failures and access limitations (UKERC, 2010).
Because of these variations among countries and individual plants,
which are primarily driven by local wind resource conditions but are
also affected by turbine design and operations, estimates of the levelized cost of wind energy must include a range of energy production
estimates. Moreover, because the attractiveness of offshore plants
is enhanced by the potential for greater energy production than for
onshore plants, performance variations among on- and offshore wind
energy must also be considered.
587
Wind Energy
7.8.3.4
Chapter 7
Levelized cost of energy estimates
35
Offshore USD 5,000/kW
30

Onshore USD 2,100/kW
25

20
15
Europe Offshore Projects
10
The levelized cost of on- and offshore wind energy varies substantially,
depending on assumed investment costs, energy production and discount rates. For onshore wind energy, levelized generation costs in
good to excellent wind resource regimes are estimated to average US
cents2005 5 to 10/kWh. Levelized generation costs can reach US cents2005
15/kWh in lower- resource areas. The costs of wind energy in China and
Levelized Cost of Energy [UScent2005 /kWh]
Levelized Cost of Energy [UScent2005 /kWh]
Using the methods summarized in Annex II, the levelized generation cost
of wind energy is presented in Figure 7.23. For onshore wind energy,
estimates are provided for plants built in 2009; for offshore wind energy,
estimates are provided for plants built in 2008 and 2009 as well as those
plants planned for completion in the early 2010s.46 Estimated levelized
costs are presented over a range of energy production estimates to represent the cost variation associated with inherent differences in the wind
resource. The x-axis for these charts roughly correlates to annual average
are used to produce levelized generation cost estimates.48 Taxes, policy

incentives, and the costs of electric system integration are not included
in these calculations.49
35
Offshore Discount Rate = 10%
30

Onshore Discount Rate = 10%
25

20
15
10
5
China
European Low-Medium Wind Areas
US Great Plains
0
15
20
25
30
35
40
45
50
15
20
25
30
Capacity Factor [%]
35
40
45
50
Capacity Factor [%]
Figure 7.23 | Estimated levelized cost of on- and offshore wind energy, 2009: (left) as a function of capacity factor and investment cost* and (right) as a function of capacity factor
and discount rate**.
Notes: * Discount rate assumed to equal 7%. ** Onshore investment cost assumed at USD2005 1,750/kW, and offshore at USD2005 3,900/kW.
wind speeds from 6 to 10 m/s. Onshore investment costs are assumed to

range from USD2005 1,200 to 2,100/kW (with a mid-level cost of USD2005
1,750/kW); investment costs for offshore wind energy are assumed to
range from USD2005 3,200 to 5,000/kW (mid-level cost of USD2005 3,900/
kW).47 Levelized O&M costs are assumed to average US cents2005 1.6/
kWh and US cents2005 3/kWh over the life of the plant for onshore and
offshore wind energy, respectively. A power plant design life of 20 years
is assumed, and discount rates of 3 to 10% (mid-point estimate of 7%)
the USA tend towards the lower range of these estimates, due to lower
average investment costs (China) and higher average capacity factors
(USA); costs in much of Europe tend towards the higher end of the range
due to relatively lower average capacity factors. Though the offshore
cost estimates are more uncertain, offshore wind energy is generally
more expensive than onshore, with typical levelized generation costs
that are estimated to range from US cents2005 10/kWh to more than US
cents2005 20/kWh for recently built or planned plants located in relatively
46 Because investment costs have risen in recent years, using the cost of recent and
planned plants reasonably reects the current cost of offshore wind energy.
48 Though the same discount rate range and mid-point are used for on- and offshore
wind energy, offshore wind power plants currently experience higher-cost nancing
than do onshore plants. As such, the levelized cost of energy from offshore plants
may, in practice, tend towards the higher end of the range presented in the gure, at
least in comparison to onshore plants.
47 Based on data presented earlier in this section, the mid-level investment cost for onand offshore wind power plants does not represent the arithmetic mean between
the low and high end of the range.
49 Decommissioning costs are generally assumed to be low, and are excluded from
these calculations.
588
Chapter 7
Wind Energy
shallow water. Where the exploitable onshore wind resource is limited,

however, offshore plants can sometimes compete with onshore plants.
7.8.4
Potential for further reductions in the cost

of wind energy
The wind energy industry has developed over a period of 30 years.

Though the dramatic cost reductions seen in past decades will not continue indenitely, the potential for further reductions remains given the
many potential areas of technological advances described in Section
7.7. This potential spans both on- and offshore wind energy technologies; given the relatively less mature state of offshore wind energy,
however, greater cost reductions can be expected in that segment. Two
approaches are commonly used to forecast the future cost of wind
energy, often in concert with some degree of expert judgement: (1)
learning curve estimates that assume that future wind energy costs will
follow a trajectory that is similar to an historical learning curve based
on past costs; and (2) engineering-based estimates of the specic cost
reduction possibilities associated with new or improved wind energy
technologies or manufacturing capabilities (Mukora et al., 2009).
7.8.4.1
Learning curve estimates
Learning curves have been used extensively to understand past cost

trends and to forecast future cost reductions for a variety of energy technologies (e.g., McDonald and Schrattenholzer, 2001; Kahouli-Brahmi,
2009; Junginger et al., 2010). Learning curves start with the premise
that increases in the cumulative production of a given technology lead
to a reduction in its costs. The principal parameter calculated by learning curve studies is the learning rate: for every doubling of cumulative
production or installation, the learning rate species the associated
percentage reduction in costs. Section 10.5 provides a more general discussion of learning curves as applied to renewable energy.
A number of published studies have evaluated historical learning rates
for onshore wind energy (Table 7.5 provides a selective summary of the
available literature).50 The wide variation in results can be explained
by differences in learning model specication (e.g., one-factor or multifactor learning curves), variable selection and assumed system boundaries
(e.g., whether investment cost, turbine cost, or levelized energy costs are
explained, whether global or country-level cumulative installations are
used, or whether country-level turbine production is used rather than
Table 7.5 | Summary of learning curve literature for onshore wind energy.
Authors
Learning By Doing Rate

(%)
Neij (1997)
Global or National
Data Years
Independent Variable
(cumulative capacity)
Dependent Variable
Denmark3
Denmark (turbine cost)
19821995
Mackay and Probert (1998)
14
USA
USA (turbine cost)
19811996
Neij (1999)
Denmark3
Denmark (turbine cost)
19821997
Wene (2000)
32
USA2
USA (generation cost)
19851994
Wene (2000)
18
EU
EU (generation cost)
19801995
Miketa and Schrattenholzer (2004)1
10
Global
Global (investment cost)
19711997
Junginger et al. (2005)
19
Global
UK (investment cost)
19922001
Junginger et al. (2005)
15
Global
Spain (investment cost)
19902001
Klaassen et al. (2005)1
Germany, Denmark, and UK
Germany, Denmark, and UK (investment cost)
19862000
Kobos et al. (2006)1
14
Global
19811997
Jamasb (2007)
13
Global
19801998
Sderholm and Sundqvist (2007)
19862000
Sderholm and Sundqvist (2007)1
19862000
Neij (2008)
17
Denmark
Denmark (generation cost)
19812000
Kahouli-Brahmi (2009)
17
Global
19791997
Nemet (2009)
11
Global
California (investment cost)
19812004
Ek and Sderholm (2010)1
17
Global
Germany, Denmark, Spain, Sweden, and UK

(investment cost)
19862002
Wiser and Bolinger (2010)
Global
USA (investment cost)
19822009
Notes: 1. Two-factor learning curve that also includes R&D; others are one-factor learning curves. 2. Independent variable is cumulative production of electricity. 3. Cumulative
turbine production used as independent variable; others use cumulative installations.
50 It is too early to develop a meaningful learning curve for offshore wind energy based
on actual data from offshore plants. Studies have sometimes used learning rates to
estimate future offshore costs, but those learning rates have typically been synthesized based on judgment and on learning rates for related industries and offshore
subsystems (e.g., Junginger et al., 2004; Carbon Trust, 2008b).
589
Wind Energy
installed wind power capacity), data quality, and the time period over
which data are available. Because of these and other differences, the
learning rates for wind energy presented in Table 7.5 range from 4 to
32%, but need special attention to be accurately interpreted and compared. Focusing only on the smaller set of studies completed in 2004
and later that have prepared estimates of learning curves based on total
wind power plant investment costs and global cumulative installations,
the range of learning rates narrows to 9 to 19%; the lowest gure within
this range (9%) is the only one that includes data from 2004 to 2009, a
period of increasing wind power plant investment costs.
There are also a number of limitations to the use of such models to
forecast future costs (e.g., Junginger et al., 2010). First, learning curves
typically (and simplistically) model how costs have decreased with
increased installations in the past, but do not comprehensively explain
the reasons behind the decrease (Mukora et al., 2009). In reality, costs
may decline in part due to traditional learning and in part due to other factors, such as R&D expenditure and increases in turbine, power plant, and
manufacturing facility size. Learning rate estimates that do not account
for such factors may suffer from omitted variable bias, and may therefore
be inaccurate. Second, if learning curves are used to forecast future cost
trends, not only should the other factors that may inuence costs be
considered, but one must also assume that learning rates derived from
historical data can be appropriately used to estimate future trends. As
technologies mature, however, diminishing returns in cost reduction can
be expected, and learning rates may fall (Arrow, 1962; Ferioli et al., 2009;
Nemet, 2009). Third, the most appropriate cost measure for wind energy
is arguably the levelized cost of energy, as wind energy generation
costs are affected by investment costs, O&M costs and energy production (EWEA, 2009; Ferioli et al., 2009). Unfortunately, only two of the
published studies calculate the learning rate for wind energy using a levelized cost of energy metric (Wene, 2000; Neij, 2008); most studies have
used the more readily available metrics of investment cost or turbine
cost. Fourth, a number of the published studies have sought to explain
cost trends based on cumulative wind power capacity installations or
production in individual countries or regions of the world; because the
wind energy industry is global in scope, however, it is likely that much
of the learning is now occurring based on cumulative global installations (e.g., Ek and Sderholm, 2010). Finally, from 2004 through 2009,
wind turbine and power plant investment costs increased substantially,
countering the effects of learning, in part due to materials and labour
price increases and in part due to increased manufacturer protability.
Because production cost data are not generally publicly available, learning curve estimates typically rely upon price data that can be impacted
by changes in materials costs and manufacturer protability, resulting in
the possibility of poorly estimated learning rates if dynamic price effects
are not considered (Yu et al., 2011).
7.8.4.2
reductions associated with specic design changes and/or technical

advances. Though limitations to engineering-based approaches also
exist (Mukora et al., 2009), these models can lend support to learning
curve predictions by dening the technology advances that can yield
cost reductions and/or energy production increases.
These models have been used to estimate the impact of potential technology improvements on wind power plant investment costs and energy
production, as highlighted in Section 7.7. Given the possible technology advances (in combination with manufacturing learning) discussed
earlier, the US DOE (2008) estimates that onshore wind energy investment costs may decline by 10% by 2030, while energy production may
increase by roughly 15%, relative to a 2008 starting point (see Table 7.3,
and the note under that table).
There is arguably greater potential for technical advances in offshore than
in onshore wind energy technology (see Section 7.7), particularly in foundation design, electrical system design and O&M costs. Larger offshore
wind power plants are also expected to trigger more efcient installation
procedures and dedicated vessels, enabling lower costs. Future levelized
cost of energy reductions have sometimes been estimated by associating
potential cost reductions with these technical improvements, sometimes
relying on subsystem-level learning curve estimates from other industries (e.g., Junginger et al., 2004; Carbon Trust, 2008b).
7.8.4.3
Projected levelized cost of wind energy
A number of studies have developed forecasted cost trajectories for onand offshore wind energy based on differing combinations of learning
curve estimates, engineering models, and/or expert judgement. These
estimates are sometimesbut not alwayslinked to certain levels of
assumed wind energy deployment. Representative examples of this literature include Junginger et al. (2004), Carbon Trust (2008b), IEA (2008,
2010b, 2010c), US DOE (2008), EWEA (2009), Lemming et al. (2009),
Teske et al. (2010), GWEC and GPI (2010) and UKERC (2010).
Recognizing that the starting year of the forecasts, the methodological approaches used, and the assumed deployment levels vary, these
recent studies nonetheless support a range of levelized cost of energy
reductions for onshore wind of 10 to 30% by 2020, and for offshore
wind of 10 to 40% by 2020. Some studies focused on offshore wind
energy technology even identify scenarios in which market factors lead
to continued increases in the cost of offshore wind energy, at least in the
near to medium term (BWEA and Garrad Hassan, 2009; UKERC, 2010).
Longer-term projections are more reliant on assumed deployment levels
and are subject to greater uncertainties, but for 2030, the same studies
support reductions in the levelized cost of onshore wind energy of 15 to
35% and of offshore wind energy of 20 to 45%.
Engineering model estimates
Whereas learning curves examine aggregate historical data to forecast

future trends, engineering-based models focus on the possible cost
590
Chapter 7
Using these estimates for the expected percentage cost reduction in

levelized cost of energy, levelized cost trajectories for on- and offshore
wind energy can be developed. Because longer-term cost projections
Chapter 7
are inherently more uncertain and depend, in part, on deployment levels and R&D expenditures that are also uncertain, the focus here is on
relatively nearer-term cost projections to 2020. Specically, Section
7.8.3.4 reported 2009 levelized cost of energy estimates for onshore
wind energy of roughly US cents2005 5 to 15/kWh, whereas estimates
for offshore wind energy were in the range of US cents2005 10 to 20/
kWh. Conservatively, the percentage cost reductions reported above can
be applied to these estimated 2009 levelized generation cost values to
develop low and high projections for future levelized generation costs.51
Based on these assumptions, the levelized generation cost of onshore
wind energy could range from roughly US cents2005 3.5 to 10.5/kWh by
2020 in a high cost-reduction case (30% by 2020), and from US cents2005
4.5 to 13.5/kWh in a low cost-reduction case (10% by 2020). Offshore
wind energy is often anticipated to experience somewhat deeper cost
reductions, with levelized generation costs that range from roughly US
cents2005 6 to 12/kWh by 2020 in a high cost-reduction case (40% by
2020) to US cents2005 9 to 18/kWh in a low cost-reduction case (10%
by 2020).52
Uncertainty exists over future wind energy costs, and the range of
costs associated with varied wind resource strength introduces greater
uncertainty. As installed wind power capacity increases, higher-quality
resource sites will tend to be utilized rst, leaving higher-cost sites for
later development. As a result, the average levelized cost of wind energy
will depend on the amount of deployment, not only due to learning
effects, but also because of resource exhaustion. This supply-curve
effect is not captured in the estimates presented above. The estimates
presented here therefore provide an indication of the technology
advancement potential for on- and offshore wind energy, but should be
used with some caution.
7.9
Potential deployment53
Wind energy offers signicant potential for near- and long-term GHG
emissions reductions. The wind power capacity installed by the end of
2009 was capable of meeting roughly 1.8% of worldwide electricity
demand and, as presented in this section, that contribution could grow
to in excess of 20% by 2050. On a global basis, the wind resource is
51 Because of the cost drivers discussed earlier in this section, wind energy costs in
2009 were higher than in some previous years. Applying the percentage cost reductions from the available literature to the 2009 starting point is, therefore, arguably a
conservative approach to estimating future cost reduction possibilities; an alternative
approach would be to use the absolute values of the cost estimates provided by the
available literature. As a result, and also due to the underlying uncertainty associated
with projections of this nature, future costs outside of the ranges presented here are
possible.
52 As mentioned earlier, the 2009 starting point values for offshore wind energy are
consistent with recently built or planned plants located in relatively shallow water.
53 Complementary perspectives on potential deployment based on a comprehensive
assessment of numerous model-based scenarios of the energy system are presented
in Sections 10.2 and 10.3 of this report.
Wind Energy
unlikely to constrain further deployment (Section 7.2). Onshore wind

energy technology is already being deployed at a rapid pace (Sections
7.3 and 7.4), therefore offering an immediate option for reducing GHG
emissions in the electricity sector. In good to excellent wind resource
regimes, the generation cost of onshore wind energy averages US
cents2005 5 to 10/kWh (Section 7.8), and no insurmountable technical barriers
exist that preclude increased levels of wind energy penetration into electricity
supply systems (Section 7.5). Continued technology advances and cost reductions in on- and offshore wind energy are expected (Sections 7.7 and 7.8),
further improving the GHG emissions reduction potential of wind energy over
the long term.
This section begins by highlighting near-term forecasts for wind energy deployment (Section 7.9.1). It then discusses the prospects for and barriers to wind
energy deployment in the longer term and the potential role of that deployment in reaching various GHG concentration stabilization levels (Section 7.9.2).
Both subsections are largely based on energy market forecasts and GHG and
energy scenarios literature published between 2007 and 2010. The section
ends with brief conclusions (Section 7.9.3). Though the focus of this section is
on larger on- and offshore wind turbines for electricity production, as discussed
in Box 7.1, alternative technologies and applications for wind energy also exist.
7.9.1
Near-term forecasts
The rapid increase in global wind power capacity from 2000 to 2009
is expected by many studies to continue in the near to medium term
(Table 7.6). From the roughly 160 GW of wind power capacity installed
by the end of 2009, the IEA (2010b) New Policies scenario and the
EIA (2010) Reference case scenario predict growth to 358 GW (forecasted electricity generation of 2.7 EJ/yr) and 277 GW (forecasted
electricity generation of 2.5 EJ/yr) by 2015, respectively. Wind energy
industry organizations predict even faster deployment rates, noting
that past IEA and EIA forecasts have understated actual growth by a
sizable margin (BTM, 2010; GWEC, 2010a). However, even these more
aggressive forecasts estimate that wind energy will contribute less
than 5% of global electricity supply by 2015. Asia, North America and
Europe are projected to lead in wind power capacity additions over
this period.
7.9.2
Long-term deployment in the context

of carbon mitigation
A number of studies have tried to assess the longer-term potential

of wind energy, often in the context of GHG concentration stabilization scenarios. As a variable, location-dependent resource with limited
dispatchability, modelling the economics of wind energy expansion
presents unique challenges (e.g., Neuhoff et al., 2008). The resulting differences among studies of the long-term deployment of wind energy
may therefore reect not just varying input assumptions and assumed
policy and institutional contexts, but also differing modelling or scenario
analysis approaches.
591
Climate Change and the EU Emissions

Trading Scheme (ETS): Looking to 2020
Larry Parker
Specialist in Energy and Environmental Policy
January 26, 2010
Congressional Research Service

7-5700
www.crs.gov
R41049
CRS Report for Congress

Prepared for Members and Committees of Congress
Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020
Summary
The European Unions (EU) Emissions Trading Scheme (ETS) is a cornerstone of the EUs
efforts to meet its obligation under the Kyoto Protocol. It covers more than 10,000 energy
intensive facilities across the 27 EU Member countries; covered entities emit about 45% of the
EUs carbon dioxide emissions. A Phase 1 trading period began January 1, 2005. A second,
Phase 2, trading period began in 2008, covering the period of the Kyoto Protocol. A Phase 3 will
begin in 2013 designed to reduce emissions by 21% from 2005 levels.
Several positive results from the Phase 1 learning by doing exercise assisted the ETS in making
the Phase 2 process run more smoothly, including: (1) greatly improving emissions data, (2)
encouraging development of the Kyoto Protocols project-based mechanismsClean
Development Mechanism (CDM) and Joint Implementation (JI), and (3) influencing corporate
behavior to begin pricing in the value of allowances in decision-making, particularly in the
electric utility sector.
However, several issues that arose during the first phase were not resolved as the ETS moved into
Phase 2, including allocation schemes and new entrant reserves, and others. A more
comprehensive and coordinated response by the EU has been made for Phase 3 with harmonized
and coordinated rules being developed by the European Commission.
The United States is not a party to the Kyoto Protocol. However, five years of carbon emissions
trading has given the EU valuable experience in designing and operating a greenhouse gas trading
system. This experience may provide some insight into cap-and-trade design issues currently
being debated in the United States.
The U.S. requires only electric utilities to monitor CO2. The EU-ETS experience
suggests that expanding similar requirements to all facilities covered under a capand-trade scheme would be pivotal for developing allocation systems, reduction
targets, and enforcement provisions.
In the U.S. debate continues on comprehensive versus sector-specific reduction

programs; the EU-ETS experience suggests that adding sectors to a trading
scheme once established may be a slow, contentious process.
As with most EU industries, most U.S. industry groups either oppose auctions
outright or want them to be supplemental to a base free allocation. The EU-ETS
experience suggests Congress may want to consider specifying any auction
requirement if it wishes to incorporate market economics more fully into
compliance decisions.
EU-ETS analysis suggests the most important variables in determining Phase 1

allowance price changes were oil and natural gas price changes; this apparent
linkage raises possible market manipulation issues, particularly with the inclusion
of financial instruments such as options and futures contracts. The EU will
examine the matter in preparation for Phase 3. Congress may consider whether
the government needs enhanced regulatory and oversight authority over such
instruments.
Contents
Overview ....................................................................................................................................1
Results from Phase 1 and 2 .........................................................................................................3
Phase 3 .......................................................................................................................................8
Auctions ...............................................................................................................................9
New Entrant Reserves ......................................................................................................... 11
EC Phase 3 Decision on Eligible Industries ......................................................................... 12
Flexibility Mechanisms and Price Volatility Control ............................................................ 13
Expanding Coverage ........................................................................................................... 14
Summary and Considerations for U.S. Cap-and-Trade Proposals ............................................... 15
Emission Inventories and Target Setting .............................................................................. 15
Coverage ............................................................................................................................ 16
Allocation Schemes ............................................................................................................ 17
Flexibility and Price Volatility............................................................................................. 18
Figures
Figure 1. ECX CFI Futures Contracts: Price and Volume.............................................................5
Figure 2. EU-15 Greenhouse Gas Emissions and Projections for the Kyoto Period: 19902012 ........................................................................................................................................6
Figure 3. Summary of EU-15 Emissions Projection Compared to Projected Kyoto
Protocol Credits .......................................................................................................................7
Tables
Table 1. Proposed Annual ETS Cap Figures for Phase 3 ..............................................................8
Contacts
Author Contact Information ...................................................................................................... 19
Overview1
Climate change is generally viewed as a global issue, but proposed responses typically require
action at the national level. With the 1997 Kyoto Protocol now in force and setting emissions
objectives for 2008-2012, countries that ratified the protocol are implementing strategies to begin
reducing their emissions of greenhouse gases.2 In particular, the European Union (EU) has
decided to use an emissions trading scheme (called a cap-and-trade program), along with other
market-oriented mechanisms permitted under the Protocol, to help it achieve compliance at least
cost.3 The decision to use emission trading to implement the Kyoto Protocol is at least partly
based on the successful emissions trading program used by the United States to implement its
sulfur dioxide (acid rain) control program contained in Title IV of the 1990 Clean Act
Amendments.4
The EUs Emissions Trading System (ETS) covers more than 10,000 energy-intensive facilities
across the 27 EU Member countries, including oil refineries, powerplants over 20 megawatts
(MW) in capacity, coke ovens, and iron and steel plants, along with cement, glass, lime, brick,
ceramics, and pulp and paper installations. In addition, aviation is currently being phased into the
ETS. These covered entities emit about 40%-45% of the EUs total greenhouse gas emissions,
and almost two-thirds of them are combustion installations. The trading program does not cover
either carbon dioxide (CO2) emissions from the transportation sector (except aviation), which
account for about 25% of the EUs total greenhouse gas emissions, or emissions of non-CO2
greenhouse gases, which account for about 20% of the EUs total greenhouse gas emissions. A
Phase 1 trading period ran between January 1, 2005, and December 31, 2007.5 A Phase 2 trading
period began January 1, 2008, covering the period of the Kyoto Protocol, and a Phase 3 has been
finalized to begin in 2013.6
Under the Kyoto Protocol, the then-existing 15 nations of the EU agreed to reduce their aggregate
annual average emissions for 2008-2012 by 8% from the Protocols baseline level (mostly 1990
levels) under a collective arrangement called a bubble. In light of the Kyoto Protocol targets,
the EU adopted a directive establishing the EU-ETS that entered into force October 13, 2003.7
Readers unfamiliar with the workings of the European Union may want to read CRS Report RS21372, The European
Union: Questions and Answers, by Kristin Archick and Derek E. Mix.
2
Six gases are included under the Kyoto Protocol: carbon dioxide, methane, nitrous oxide, hydrofluorocarbons,
perfluorocarbons, and sulfur hexafluoride. The United States has not ratified the Kyoto Protocol and, therefore, is not
covered by its provisions. For more information on the Kyoto Protocol, see CRS Report RL33826, Climate Change:
The Kyoto Protocol, Bali Action Plan, and International Actions, by Jane A. Leggett.
3
Norway, a non-EU country, also has instituted a CO2 trading system which is currently linked with the EU-ETS.
Various other countries and a state-sponsored regional initiative located in the northeastern United States involving
several states are developing mandatory cap-and-trade system programs, but are not operating at the current time. For a
review of these emerging programs, along with other voluntary efforts, see CRS Report RL33812, Climate Change:
Action by States to Address Greenhouse Gas Emissions, by Jonathan L. Ramseur.
4
P.L. 101-549, Title IV (November 15, 1990).
5
For further background on the ETS, see CRS Report RL34150, Climate Change and the EU Emissions Trading
Scheme (ETS): Kyoto and Beyond, by Larry Parker.
6
More information, including relevant directives, on the EU-ETS is available on the European Unions website at
http://europa.eu.int/scadplus/leg/en/lvb/l28012.htm.
7
Directive 2003/87/EC of the European Parliament and of the Council of 13 October 2003 establishing a scheme for
greenhouse gas emissions allowance trading within the Community and amending Council Directive 96/61/EC.
One objective of the second phase of the ETS is to achieve 3.3 percentage points of the 8.0%
reduction required by the EU-15 under the Protocol.8
The importance of emissions trading was elevated by the accession of 12 additional central and
eastern European countries to EU membership from May 2004 through January 2007. For the
new EU-27, the overall ETS emissions cap is set at 2.08 billion metric tons of carbon dioxide
(CO2) annually for the Kyoto compliance period (2008-2012).
The second phase Kyoto compliance stage of the ETS is built on the experience the EU gained
from its preliminary Phase 1. The European Commission (EC) believes that the Phase 1 learning
by doing exercise prepared the community for the difficult task of achieving the reduction
requirements of the Kyoto Protocol. Several positive results from the Phase 1 experience assisted
the ETS in making the Phase 2 process run smoothly, at least so far. First, Phase 1 established
much of the critical infrastructure necessary for a functional emission market, including
emissions monitoring, registries, and inventories. Much of the publicized difficulty the ETS
experienced early in the first phase can be traced to inadequate emissions data infrastructure.9
Phase 1 significantly improved those critical elements in preparation for Phase 2 implementation.
Second, the ETS helped jump-start the project-based mechanismsClean Development
Mechanism (CDM) and Joint Implementation (JI)created under the Kyoto Protocol. 10 As stated
by Ellerman and Buchner:
The access to external credits provided by the Linking Directive has had an invigorating
effect on the CDM and more generally on CO2 reduction projects in developing countries,
especially in China and India, the two major countries that will eventually have to become
part of a global climate regime if there is to be one.11
Third, according to the EC, a key result of Phase 1 was its effect on corporate behavior. An EC
survey of stakeholders indicated that many participants are incorporating the value of allowances
in making decisions, particularly in the electric utility sector, where 70% of firms stated they were
pricing the value of allowances into their daily operations, and 87% into future marginal pricing
decisions. All industries stated that it was a factor in long-term decision-making.12
8
Commission of the European Communities, Communication from the Commission: Progress towards Achieving the
Kyoto Objectives (November 19, 2008). Other reductions are to be achieve through regulatory measures, such as a CO2
emissions standard for automobiles.
9
A. Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations,
and Early Results, Environmental Economics and Policy (Winter 2007), pp. 69-70; and International Emissions
Trading Association, IETA Position Paper on EU ETS Marking Functioning, (no date), p. 3.
10
For more on the effect of the ETS on Kyoto mechanisms, see A. Denny Ellerman and Barbara K. Buchner, The
European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, Environmental Economics and
Policy (Winter 2007), p. 84; and International Emissions Trading Association, IETA Position Paper on EU ETS
Market Functioning (no date), p. 2. For more information on the Kyoto Protocol mechanisms, see CRS Report
RL33826, Climate Change: The Kyoto Protocol, Bali Action Plan, and International Actions, by Jane A. Leggett.
11
A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations,
and Early Results, Environmental Economics and Policy (Winter 2007), p. 84.
12
European Commission, Directorate General for Environment, Review of EU Emissions Trading Scheme: Survey
Highlights, (November 2005), pp. 5-7.
However, several issues that arose during the first phase remained contentious as the ETS
implemented Phase 2, including allocation (including use of auctions and reliance on model
projections), new entrant reserves, and others. In addition, the expansion of the EU and the
implementation of the linking directives created new issues to which Phase 2 has had to
respond. 13 Based on lessons learned in Phase 1 and Phase 2, the EU has taken a substantially
different approach to these issues in Phase 3 that is discussed later.
Results from Phase 1 and 2

It is unclear to what degree the first phase of the ETS achieved real emissions reductions.
Emissions are dynamic over time; a product of a countrys population, economic activity, and
greenhouse gas intensity.14 To capture these dynamics, each Member State of the EU developed
an emissions baseline from models that project future trends in the countrys emissions based on
these and other factors, such as anticipated energy and greenhouse gas policies. 15 During the first
phase, the emissions goal was to put the EU on the path to Kyoto compliancenot actually
comply with the Protocol (which wasnt necessary until the 2008-2012 time period). Thus,
countries developed business as usual baselines based on projected growth in emissions. Such a
projected baseline suffers from two sources of uncertainty: data uncertainties, and forecasting
uncertainties. On data, Phase 1 suffered from uncertainties with respect to data collection and
coverage, in monitoring methods for historic data, and data verification. On projecting future
emissions, Phase 1 faced uncertainties with respect to economic or sector-based growth rates.
Fueled in many cases by over-optimistic economic growth assumptions, these uncertainties
increased the probability of inflated business as usual baselines. 16
The combination of these factors and modest reduction requirements resulted in the emissions
allocations for the 2005-2007 trading period being higher than actua1 2005 emissions. 17 This
result raised questions about how much reductions achieved during Phase 1 were real as opposed
to being merely paper artifacts. On the positive side, verified emissions in 2005 were 3.4% below
the estimated 2005 baseline used during the allocation process. In addition, the allowance prices
for 2005 stayed persistently high, suggesting some abatement was occurring and raising questions
of windfall profits. As stated by Ellerman and Buchner:
First, and most importantly, the persistently high price for EUAs [EU emissions allowances]
in a market characterized by sufficient liquidity and sophisticated players must be considered
as creating a presumption of abatement. It would be startling if power companies did not
13
For a further discussion of Phase 2 implementation issues, see CRS Report RL34150, Climate Change and the EU
Emissions Trading Scheme (ETS): Kyoto and Beyond, by Larry Parker.
14
For more information, see CRS Report RL33970, Greenhouse Gas Emission Drivers: Population, Economic
Development and Growth, and Energy Use, by John Blodgett and Larry Parker.
15
On the role of modeling in the first phase, see A Denny Ellerman and Barbara K. Buchner, The European Union
Emissions Trading Scheme: Origins, Allocations, and Early Results, 1 Environmental Economics and Policy 1
(Winter 2007), pp. 72-73.
16
Regina Betz and Misato Sato, Emissions Trading: Lessons Learnt from the 1st Phase of the EU ETS and Prospects
for the 2nd Phase, 6 Climate Policy (2006), p. 354.
17
For a further discussion, see CRS Report RL33581, Climate Change: The European Union's Emissions Trading
System (EU-ETS), by Larry Parker.
incorporate EUA prices into dispatch decisions that would have shifted generation to less
emitting plants. There is plenty of anecdotal evidence that this was the case, and the
prominent charges of windfall profits assume that the opportunity cost of freely allocated
allowances was being passed on (without noting the implications for abatement). Similarly, it
would be surprising if there were no changes in production processes that could be made by
the operators of industrial plants.18
However, EU emissions allowances (EUAs) during Phase 1 did not maintain value. Phase 1
EUAs were basically worthless during the final six months of 2007. This decline in EUA prices at
least partially reflected the general non-transferability of Phase 1 EUAs to Phase 2. Only Poland
and France included limited banking in their Phase 1 implementation plans (called National
Allocation Plans (NAPs)). The EC further restricted use of Phase 1 EUAs in Phase 2 with a ruling
in November 2006.19 As a result, excess Phase 1 EUAs were worthless at the end of 2007.20
One consequence of the non-transferability of Phase 1 EUAs is that prices for Phase 2 EUAs
remained relatively firm until the 2008-2009 recession reduced demand, as indicated by Figure 1.
Scarcity is critical for the proper functioning of an allowance market. As further indicated by
Figure 1, during 2009, the market firmed up at a much lower level as participants assessed the
impact of the recession on the demand for EUAs. This is a different response than the market had
during Phase 1, and may reflect Phase 2 improvements in the system. In particular, the more
predictable 2009 response may reflect the ability of the EC to certify Phase 2 NAPs using more
verifiable baseline data than were available for Phase 1.21 A major reason the EC rejected ex post
adjustments22 was fear that such adjustments would have a disruptive effect on the marketplace. 23
Phase 1 did not firmly establish this foundation of markets;24 based on the Phase 2 EUA futures
market, further market development appears to be occurring, although, like most commodity
markets, it remains somewhat volatile at times.
18
and Early Results, 1 Environmental Economics and Policy 1 (Winter 2007), p. 83.
19
European Commission, Communication from the Commission to the Council and to the European Parliament on the
assessment of national allocation plans for the allocation of greenhouse gas emission allowances in the second period
of the EU Emissions Trading Scheme, COM(2006) 725 final (November 29, 2006), p. 11.
20
For a further discussion, see Joseph Kruger, Wallace E. Oates, and William A. Pizer, Decentralization in the EU
Emissions Trading Scheme and Lessons for Global Policy, 1 Environmental Economics and Policy 1 (Winter 2007), p.
126; and, Frank J. Convery and Luke Redmond, Market and Price Development in the European Union Emissions
Trading Scheme, 1 Environmental Economics and Policy 1 (Winter 2007), pp. 96-7, 107.
21
International Emissions Trading Association, IETA Position Paper on EU ETS Market Functioning, (no date), p. 2.
22
Once the EC has approved a countrys NAP, including the total number of allowances and the allocation to each
covered entity, the allocations can not be re-visited. Attempts to include provisions permitting such post-approval
adjustments to a facilitys allocation have been uniformly rejected by the EC.
23
European Commission, Communication from the Commission to the Council and to the European Parliament on the
assessment of national allocation plans for the allocation of greenhouse gas emission allowances in the second period
of the EU Emissions Trading Scheme, COM(2006) 725 final (November 29, 2006), p 8; and, A Denny Ellerman and
Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, 1
Environmental Economics and Policy 1 (Winter 2007), p. 71.
24
On the mixed record of the EU-ETS and the need for allowance scarcity to a functioning emissions market, see Eric
Haymann, EU Emission Trading: Allocation Battles Intensifying, Deutsche Bank Research (March 6, 2007). For a
generally positive view of ETS market development, see Frank J. Convery and Luke Redmond, Market and Price
Development in the European Union Emissions Trading Scheme, 1 Environmental Economics and Policy 1 (Winter
2007), pp. 97-106. For a more negative view, see Karsten Neuhoff, Federico Ferrario, Michael Grubb, Etienne Gabel,
and Kim Keats, Emissions Projections 2008-2012 Versus NAPs II, 6 Climate Policy 5 (2006), pp. 395-410.
Figure 1. ECX CFI Futures Contracts: Price and Volume
ECX EUA Futures Contracts: Price and Volume

Total Volume
35
35
30
30
25
20
20
15
15
10
10
06 06 06 06 06 06 06 06 07 07 07 07 07 07 07 07 07 08 08 08 08 08 08 08 08 08 09 09 09 09 09 09 09 09 09
20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20
/
6 7 1 2 1 1 1 1 3 3 6 7 5 6 5 6 6 9 9 2 4 4 4 2 3 3 6 6 8 1 2 2 2 2 1
2/ 3/1 5/ 6/1 7/2 8/3 0/1 1/2 1/ 2/1 3/2 5/ 6/1 7/2 9/ 0/1 1/2 1/ 2/1 4/ 5/1 6/2 8/ 9/1 0/2 12/ 1/1 2/2 4/ 5/2 7/ 8/1 9/2 11/ 2/1
1
1 1
1 1
1
Source: ECX Exchange.
Note: Dec09 Sett: Future contracts with a settlement date of December, 2009.
CRS-5
25
Price per tonne (EUR)
VOLUME (million tonnes CO2)
Dec09 Sett
40
While the environmental performance of Phase 1 may be disputed, the need for additional
reductions to achieve Kyoto is not. For 2008, the EU-15 is estimated to be 6.2% below its baseyear emissions, compared with an 8% five-year average reduction commitment under the Kyoto
Protocol. As indicated by Figure 2, this represents a continuation of reductions by EU-15 over
the past five years. However, as indicated by the pink line, the European Environment Agency
(EEA) projects that the EU-15 existing measures are insufficient to reduce EU-15 emissions to
their Kyoto requirements (represented by the purple line), resulting in a projected 6.9% reduction
from baseline levels. To achieve the Kyoto target the EU projects further actions reductions by
EU-15 countries (represented by the green line in Figure 2), resulting in an overall reduction of
8.5% compared with baseline levels.
Figure 2. EU-15 Greenhouse Gas Emissions and Projections for the Kyoto Period:
1990-2012
Source: European Environmental Agency, Greenhouse Gas Emissions Trends and Projects in Europe 2000, (2009) p.
10.
Note: WEM: with existing measures (measures implemented or adopted). WAM: with additional measures
(planned measures).
In addition to domestic emission reductions, the EU has also projected additional reduction
credits received by activities permitted under the Kyoto Protocol: (1) purchase of project-based
credits by ETS participants and EU governments (e.g., Joint Implementation (JI) and Clean
Development Mechanism (CDM) projects); and, (2) the use of carbon sinks. As indicated in
Figure 3, these activities provide a credit on the EU-15 baseline of 4.6 percentage points. Thus, if
the EU-15 maintains its current path, it would exceed its Kyoto commitment by about 3.5
percentage points (6.9% minus 3.4%). If its planned measures result in the projected 8.5%
reduction below baseline levels, the overachievement of its Kyoto commitment would be 5.1
percentage points (8.5% minus 3.4%).25
Figure 3. Summary of EU-15 Emissions Projection Compared to Projected Kyoto
Protocol Credits
Source: European Environmental Agency, Greenhouse Gas Emissions Trends and Projects in Europe 2000, (2009) p.
11.
Notes: The left section shows the projected emissions considering only domestic measures (existing and
additional) and is showing them as average 2008-2012 emissions (lines) and annual emissions (bars). The right
section shows the projected amount of Kyoto credits accumulated by the end of the commitment period,
including the initial EC assigned amount under the Protocol, the purchase of Kyoto project credits by EU ETS
participants and EU governments, and carbon sink activities.
The EU-27 as a whole does not have an emissions target comparable to the EU-15 bubble. By
2010, EU-27 emissions are projected at 9.6% below Kyoto baseline levels assuming current
policies. This reduction is projected at 11.3% if additional measures are included. Currently, 24 of
the 25 countries with reduction requirements are projected to meet their commitments under the
Kyoto Protocol. 26 Only Austria is not projected to meet its requirements even with additional
planned measures and the use of Kyoto mechanisms. 27
25
European Environment Agency, Greenhouse Gas Emissions Trends and Projections in Europe 2009, (2009) p. 11.
Cyprus and Malta are not Annex 1 countries.
27
European Environmental Agency, Greenhouse Gas Emission Trends and Projections in Europe 2009 (2009), p. 12.
26
Phase 3
The European Union is committed to achieving a 20% reduction in greenhouse gas emissions by
2020 from 1990 levels (or more depending on the actions of other countries). A strategic
component of the effort to achieve this target is a revised ETS that will achieve a 21% reduction
from covered entities from 2005 levels. Table 1 indicates the proposed EU-wide ETS cap for the
next Phase of EU greenhouse gas program (Phase 3) assuming no further international
commitments (the final 2013 cap figure is required by June 30, 2010). As indicated, the EC
envisions a linear reduction in the ETS cap to match the reduction target under the overall 20%
reduction program. These numbers will change as individual countries decide to include more
facilities under the ETS and as the EC expands ETS coverage to include other sectors and nonCO2 greenhouse gases.
Table 1. Proposed Annual ETS Cap Figures for Phase 3
Year
Billion metric tons of

CO2e
Annual limit for Kyoto compliance period

(2008-2012)
2.083
2013
1.974
2014
1.937
2015
1.901
2016
1.865
2017
1.829
2018
1.792
2019
1.756
2020
1.720
Source: European Commission, Questions and Answers on the Commissions Proposal to revise the EU Emissions
Trading System (Brussels, January 23, 2009), response to question 12.
Note: Figures are based on the current Phase 2 scope of the ETS. These need to be adjusted for three reasons:
(1) extensions of ETS scope during phase 2 by Member states; (2) extensions of ETS scope by the EC for third
trading period, and (3) the figures do not include inclusion of aviation, nor the emissions from Norway, Iceland,
and Liechtensteinnon-EU countries that have linked their programs to the ETS.
For Phase 3, the EU is re-shaping the ETS to improve its efficiency and eliminate some of the
problems identified during Phase 1 and 2.28 For Phase 2, the improved emissions inventories
resulting from Phase 1 allowed the EC to harmonize the types of installations covered by the ETS
across the various Member States.29 In addition, the EC imposed a uniform rule on the Member
States preventing the use of ex-post adjustments. However, Phase 2 made little advancement in
28
European Commission, Directive 2009/29/EC of the European Parliament and of the Council of 23 April 2009
amending Directive 2003/87/EC so as to improve and extend the greenhouse gas emission allowance trading system of
the Community (Brussels, April 23, 2009). Hereinafter referred to as the Directive.
29
European Commission, Limiting Global Change to 2 degrees Celsius: The Way Ahead for 2020 and Beyond
(Brussels, January 10, 2007), p. 23.
harmonizing individual countries allocations schemes.30 As with Phase 1, countries continue to

differ widely on several key points.
The critical structural change the EU would make in Phase 3 is eliminating National Allocation
Plans (NAPs), and replacing them with EU-wide rules with respect to allowance availability,
allocations, and auctions. NAPs are central to the EUs effort to achieve its Kyoto obligations
under Phase 2. Each Member of the EU submitted a NAP that lays out its allocation scheme
under the ETS, including individual allocations to each affected unit. These NAPs were assessed
by the EC to determine compliance with 12 criteria delineated in an annex to the emissions
trading directive. 31 Criteria included requirements that the emissions caps and other measures
proposed by the Member State were sufficient to put it on the path toward its Kyoto target,
protections against discrimination between companies and sectors, and delineation of intended
use of CDM and JI credits for compliance, along with provisions for new entrants, clean
technology, and early reduction credits. For the second trading period, the NAP had to guarantee
Kyoto compliance.
This NAP structure will be replaced under Phase 3. There would be one EU-wide cap instead of
the 27 national caps under Phase 1 and 2. Allowances would be allocated under EU-wide, fully
harmonized rules, including those governing: (1) auctions, (2) transitional free allocations for
greenhouse gas intensive, trade-exposed industries, (3) new entrants, and (4) coverage. The EC
proposed a Directive to alter the EU-ETS structure for Phase 3 in January, 2008, 32 and the
Directive was amended and adopted by the European Parliament (EP) and of the Council of the
European Union in April 2009.33
Auctions
Under Phases 1 and 2, allowances generally were and are allocated free to participating entities
under the ETS. During Phase 1, The EU-ETS Directive allowed countries to auction up to 5% of
allowance allocations, rising to 10% under Phase 2.34 Under Phase 1, only 4 of 25 countries used
auctions at all, and only Denmark auctioned the full 5%. The political difficulty in instituting
significant auctioning into ETS allowance allocations is the almost universal agreement by
covered entities in favor of free allocation of allowances and opposition to auctions.35 Free
allocation of allowances represents a one-time transfer of wealth to the entities receiving them
from the government issuing them. 36 The resulting transfer of wealth has been described by
30
Joachim Schleich, Regina Betz, and Karoline Rogge, EU Emissions TradingBetter Job Second Time Around?
Fraunhofer Institute System and Innovation Research (February 2007), p. 23.
31
Commission of the European Communities, Directive 2003/87/EC, available at http://eur-lex.europa.eu/LexUriServ/
LexUriServ.do?uri=OJ:L:2003:275:0032:0046:EN:PDF.
32
European Commission, Proposal for a Directive of the European Parliament and of the Council amending Directive
2003/87/EC so as to improve and extend the greenhouse gas emission allowance trading system of the Community
(Brussels, January 23, 2008).
33
European Commission, Directive 2009/29/EC of the European Parliament and of the Council amending Directive
(Brussels, April 23, 2009).
34
For a further discussion of auctioning and the ETS, see Cameron Hepburn, et. al., Auctioning of EU ETS phase II
allowances: how and why? 6 Climate Policy (2006), pp. 137-160.
35
36
Joseph Kruger, Wallace E. Oates, and William A. Pizer, Decentralization in the EU Emissions Trading Scheme and
(continued...)
several analysts as windfall profits.37 As summarized by Ellerman and Buchner: Allocation in

the EU ETS provides one more example that, notwithstanding the advice of economists, the free
allocation of allowances is not to be easily set aside.38
Despite concerns about windfall profits and economic distortions resulting from the free
allocation of allowances, there was little change in basic allocation philosophy for Phase 2. No
country proposed auctioning the maximum percentage of allowances allowed (10%). Most do not
include auctions at all.39 The unwillingness of governments to employ auctions as an allocating
mechanism revolve around equity considerations, including: (1) inability of some covered entities
to pass through cost because of regulation or exposure to international competition; (2) potential
drag on a sectors economic performance from the up-front cost of auctioned allowances; and (3)
the potential that government will not recycle revenues to alleviate compliance costs,
international competitiveness impacts, or other equity concerns, resulting in the auction costs
being the same as a tax.40
This opposition is mostly overcome for Phase 3 through an EU-wide set of harmonized rules for
allowance allocations and auctions. Under Phase 3, the Directive states:
Auctioning should ... be the basic principle for allocation, as it is the simplest, and generally
considered to be the most economic efficient system. This should also eliminate windfall
profits and put new entrants and economies growing faster than average on the same
competitive footing as existing installations. (paragraph 15)
After nine eastern European Member States threatened to veto an initial proposal to auction 100%
of all allowances, the EU compromised to provide for some free allocation of allowances during
Phase 3 that will begin in 2013.41 Most covered industries will be eligible for some free allocation
of allowances to cover direct emissions under the Phase 3 agreement. The introduction of auction
would be differentiated by sector. In general, for the power sector, full auctioning will begin in
2013. For electric powerplants, most will receive no free allocation of allowances during Phase 3.
However, in a concession to certain eastern European Member States, an optional and temporary
derogation from the no-free-allocation requirement for powerplants is provided to countries that
meet specific energy and economic criteria. Under the optional allocation scheme, the Member
State can allocate allowances equal to 70% of the powerplants Phase 1 emissions free; this
allocation declines to zero in 2020.
(...continued)
Lessons for Global Policy, 1 Environmental Economics and Policy 1 (Winter 2007), p. 114.
37
E.g., Deutsche Bank Research, EU Emission Trading: Allocation Battles Intensifying, (March 6, 2007) pp. 2-3; and
Regina Betz and Misato Sato, Emissions Trading: Lessons Learnt from the 1st Phase of the EU ETS and Prospects for
the 2nd Phase, 6 Climate Policy (2006), p. 353.
38
39
For more information, see CRS Report RL34150, Climate Change and the EU Emissions Trading Scheme (ETS):
Kyoto and Beyond, by Larry Parker.
40
Martina Priebe, Distributional Effect of Carbon-allowance Trading (Cambridge, January 12, 2007). Also, see
Eurochambres, Review of the EU Emission Trading System (June 2007), p. 5.
41
See Position of the European Parliament adopted at the first reading on 17 December 2008 with a view to the
adoption of Directive 2009//EC of the European Parliament and of the Council amending Directive 2003/87/EC so
as to improve and extend the greenhouse gas emission allowance trading system of the Community (December 17,
2008).
10
The auction schedule for most other covered entities is more gradual with 80% of a sectors
allocation provided free in 2013, declining linearly to 30% by 2020, and zero by 2027. As stated
in the final Directive:
For other sectors covered by the Community scheme, a transitional system should be
foreseen for which free allocation in 2013 would be 80% of the amount that corresponded to
the percentage of the overall Community-wide emissions throughout the period 2005 to 2007
that those installations emitted as a proportion of the annual Community-wide total quantity
of allowances. Thereafter, the free allocation should decrease each year by equal amounts
resulting in 30% free allocation in 2020, with a view to reaching no free allocation in 2027.
(paragraph 21)
Because of concern that stringent EU carbon policies may encourage production and related
greenhouse gas emissions to shift to countries without carbon policies (i.e., carbon leakage),
exceptions to this phase-out of free allowances will be made in sectors where carbon leakage may
occur, as discussed later.
Distribution of allowances to be auctioned by the Member States will be determined by a threepart formula (Article 10(2)). Eighty-eight percent of the allowances to be auctioned by each
Member State is distributed to States according to their historic emissions under Phase 1 of the
EU-ETS. Ten percent of the total is distributed to States mostly based in their comparative GDP
per capita within the EU (Annex IIa). Two percent of the total is distributed to nine former
eastern-bloc countries based on the substantial greenhouse gas reductions they have already
achieved (Annex IIb). Auctions will be conducted at the Member State level (cooperative
auctions between States are also allowed) and must be open to any potential buyer. The EC is
directed to develop the appropriate rules for coordinated auctions by June 30, 2010.
Beyond the allocation of allowances, the EU Directive also provides guidelines for the allocation
of revenues from allowance auctions. The Directive states that at least 50% of the proceeds
should be used to fund a variety of climate change related activities, including emission
reductions, adaptation activities, renewable energy, carbon capture and storage (CCS), the Global
Energy Efficiency and Renewable Energy Fund, and assisting developing countries to avoid
deforestation and increase afforestation and reforestation (Article 10(3)).
New Entrant Reserves

Unlike previous cap-and-trade programs, the EU-ETS includes provisions for allocating free
allowances to new entrants to the system. 42 The reasoning behind this decision is based on equity:
(1) it isnt fair to allocate allowances free to existing entities while requiring new entrants to
purchase them, and (2) the EU doesnt want to put Member States at a disadvantage in competing
for new investments.43 These equity concerns trumped concerns about economic efficiency.
42
For example, the U.S. acid rain program provides no allocation of allowances to new entrants; instead, an EPA
sanctioned auction is held annually to ensure that allowances are available to new entrants. New entrants can also
obtain allowances from existing sources willing to sell them, either directly, through the EPA auction, or via a broker.
43
11
As is the case for existing entities, the free allocation of allowances to new entrants is a subsidy.
Under Phase 1 and Phase 2, the size and distribution of this subsidy is left to the individual
Member States. For Phase 1, the reserve varied widely from the average of 3% of total
allowances: Poland set aside only 0.4% of its allocation for new entrants while Malta set aside
26%. For Phase 2, the spread continues with Poland reserving 3.2% of its allowances for new
entrants in contrast to 45% reserved by Latvia.44
The decision to employ a new entrant reserve adds complexity to Member States allocation plans
and influences the investment decisions of covered entities. Rules had to be promulgated with
respect to the reserves size, manner in which the allowances are dispensed, and how to proceed if
the demand either exceeds the supply, or vice versa. As indicated, countries did not harmonize
new entrant reserve rules with respect to size during Phase 1 or 2. Likewise, there is no
standardization on dispensing allowances and replenishing the reserve: first-come, first-serve
with no replenishment is one approach used, but a variety of procedures have been developed
both to dispense allowances and to replenish the reserve if supply is inadequate. Member States
also have different formulas for determining how many allowances a new entrant should receive.
Member States claim to use a form of benchmarking to determine allowance allocationsan
approach based on a standard of best practices or best technology that is applied to the new
entrants anticipated production or capacity. However, the definitions and application of the
benchmarks used by the Member States are not uniform.
This will change under Phase 3. Under Phase 3, the Directive sets an EU-wide cap of 5% of the
total allowance cap for a new entrant reserve, and requires the harmonization of allocation rules.
The EC is to adopt a harmonized rule for applying a new entrant definition contained in the
Directive by December 31, 2010; the Directive expressly excludes any new electricity production
from being defined as a new entrant. The EC is also to determine EU-wide benchmarks for the
allocation of all free allowances. The Directive states that the starting point for setting those
benchmarks shall be the average performance of the 10% most efficient installations in a sector or
subsector in the EU in the years 2007-2008 (Article 10a(2)).
In an attempt to stimulate development of CCS, the Directive also provides that up to 300 million
allowances in the new entrants reserve shall be available through 2015 for aiding construction
and operation of up to 12 demonstration projects. No one project can receive more than 15% of
the allowances allocated for this purpose (Article 10a(8)).
EC Phase 3 Decision on Eligible Industries45

Most studies of the competitiveness impacts of the ETS during Phase 1 have found no impact.
The International Energy Agency (IEA) cites several reasons for this situation:
Experience to date with the EU-ETS does not reveal leakage for the sectors concerned
analysis of steel, cement, aluminum and refineries sectors reveals that no significant changes
in trade flows and production patterns were evident during the first phase (2005-2007) of the
EU-ETS. This is mostly due to the free allocation of allowances, sometimes in generous
44
Karoline Rogge, Joachim Schleich, and Regina Betz, An Early Assessment of National Allocation Plans for Phase 2
of EU Emission Trading, Fraunhofer Institute System and Innovation Research (January 2006).
45
For more information on climate change and competitiveness issues, see CRS Report R40914, Climate Change: EU
and Proposed U.S. Approaches to Carbon Leakage and WTO Implications, by Larry Parker and Jeanne J. Grimmett.
12
quantities, and to the still functioning long-term electricity contracts, which softened the
blow of rising electricity prices. Further, the general boom in prices for most traded products
subject to carbon costswhether direct or indirecthas blurred any effects of the latter.
Finally, the relatively short time span of these policies does not allow observation of the full
potential effects on industry via changes in investment location decisions.46
This conclusion is echoed by Carbon Trust, which states that currently, free allocation of
emissions allowances offset almost all of the additional costs of the ETS; and that conclusion is
echoed by The Climate Group for The German Marshall Fund, which states that companies
surveyed found it difficult to quantify effects on their bottom line in the first phase, or found no
effect at all.47
For energy-intensive, trade-exposed industries, Phase 3 has provisions to provide assistance to
eligible installations to address the direct and indirect impact of emissions control costs. With
respect to direct emissions costs, the EC published a list of installations exposed to a significant
risk of carbon leakage on December 24, 2009, as required under the Directive. 48 The list is
identical to the draft list released in September 2009.49 The decision lists 164 industrial sectors
and subsectors deemed to be exposed sectors under the appropriate European Parliament and
Council directives. Eligible installations will receive allowances sufficient to cover 100% of their
direct emissions, provided they are using the most efficient technology available. This 100%
allocation contrasts with the 80% distribution of free allowances to non-carbon leakage exposed
industries in 2013. Reflecting the fluid nature of the competitive situation and international
negotiations, the EC is to review its decision June 30, 2010, and provide the European Parliament
and Council with any appropriate proposals to respond to the situation.
Assistance for the impact of indirect emissions control costs on exposed industries from higher
electricity prices would be determined by Member States. As stated by the Directive:
Member States may deem it necessary to compensate temporarily certain installations which
have been determined to be exposed to a significant risk of carbon leakage related to
greenhouse gas emissions passed on in electricity prices for these costs. Such support should
only be granted where it is necessary and proportionate and should ensure that the
Community scheme incentives to save energy and to stimulate a shift in demand from grey
to green electricity are maintained. (paragraph 27)
Flexibility Mechanisms and Price Volatility Control

The major flexibility mechanism developed under the EU-ETS has been the Clean Development
Mechanism (CDM) and Joint Implementation (JI) credits permitted under the Kyoto Protocol;
however, this development has proven a controversial process. A major part of the controversy
46
Julia Reinaud, Issues Behind Competitiveness and Carbon Leakage: Focus on Heavy Industry (October 2008), p. 6.
Carbon Trust, EU ETS Impacts on Profitability and Trade (January 2008), p. 4; and The Climate Group, The Effects
of EU Climate Legislation on Business Competitiveness; A Survey and Analysis (September 2009), p. 8.
48
European Commission, Commission Decision of 24 December 2009 determining, pursuant to Directive 2003/87/EC
of the European Parliament and of the Council, a list of sectors and subsectors which are deemed to be exposed to a
significant risk of carbon leakage (Brussels, 2009).
49
European Commission, Draft Commission Decision of 18 September2009 determining, pursuant to Directive
2003/87/EC of the European Parliament and of the Council, a list of sectors and subsectors which are deemed to be
exposed to a significant risk of carbon leakage (Brussels, 2009).
47
13
has been the supplementarity requirement of the Kyoto Protocol to use its flexibility
mechanisms. Supplementarity requires that developed countries, such as most EU countries,
ensure that their use of JI/CDM credits is supplemental to their own domestic control efforts. In
defining supplementarity for Phase 2, the EC used 10% of a countrys allowance allocation as a
rule of thumb in approving NAPswith a greater limit possible based on a countrys domestic
efforts to reduce emissions. This process resulted in some significant reductions in some
countries proposed limits (e.g., Ireland, Poland, Spain), but some increase in others (e.g., Italy,
Latvia, Lithuania). Although these reductions appear substantial in individual cases, most analysts
agree that they do not represent a major barrier to the cost-effective use of JI/CDM. However, the
EU-ETS does not accept credits from land use, land-use change and forestry (LULUCF) projects.
For Phase 3, the EU maintains its ban on using LULUCF credits within the ETS. However, it will
permit up to 50% of the required reductions mandated under Phase 3 to be achieved through
CDM or JI credits. For existing installations, this represents a total of 1.6 billion credits over the
eight-year compliance period. Limits on use of Kyoto credits will be based on a facilitys 20082012 allocation (for an existing facilities) or its verified emissions during Phase 3 (for a new
entrant or sector). The EC estimates that the minimum amount of Kyoto credits an existing
facility will be able to use to comply with Phase 3 will be 11% of its 2008-2012 allocation, while
new entrants and sectors will be able to use a minimum of 4.5% of their verified emissions during
2013-2020 (article 11a(8)). The precise percentages will be determined later.
Another flexibility mechanism, banking, is extended by the Directive from Phase 2 to Phase 3 in
order to prevent a Phase 1 style collapse of allowance prices when the ETS transitions into Phase
3. In addition, the EU hopes that extending the trading period from five years to eight years, along
with the steady, linear emissions reduction schedule, will increase certainty and stability in the
allowance markets.
Phase 3 will introduce two other mechanisms designed to address price volatility. First, the EC is
required under the Directive to examine whether the market for emission allowances is
sufficiently protected from insider dealing or market manipulation. If not, the EC is to present
proposals to ensure such protection to the EP and the Council (article 12(a)).
Second, the Directive provides that if the allowance price is more than three times the preceding
two-year average for more than six consecutive months and the price is not based on market
fundamentals, one of two measures may be taken. The first would allow Member States to shift
forward the auctioning of some of its auctionable allowances. The second would allow Member
States to auction up to 25% of the remaining allowances in the new entrants reserve (article
29(a)).
Expanding Coverage
Despite the ECs interest in expanding the ETS, its coverage in terms of industries included for
Phase 2 is essentially the same as for Phase 1. The exception is for aviation. In December, 2006,
the EC proposed bringing greenhouse gas emissions from civil aviation into the ETS in two
phases. 50 As agreed to by the European Parliament in July 2008, all intra-EU and international
50
2003/87/EC so as to include aviation activities in the scheme for greenhouse gas emission allowance trading within the
Community (Brussels, December 12, 2006).
14
flights will be included under the ETS beginning in 2012. Emissions would be capped at 97% of
average 2004-2006 emissions with 85% of the allowances being allocated free to operators. The
cap would be reduced to 95% in 2013. The cap and auctioning of allowances would be reviewed
as a part of Phase 3 implementation.
Annex I of the Directive identifies three CO2 emitting sectors for inclusion under the ETS:
petrochemicals, ammonia, and aluminum. The ETS will also expand beyond CO2 to include
nitrous oxide (N2O) emissions from nitric, adipic, and glyoxalic acid production, and
perofluorocarbon (PFC) emissions from the aluminum sector. This would expand ETS covered
emissions by 4.6% over Phase 2 allowance allocations, or about 100 million metric tons.51 The
harmonization and codification of eligibility criteria for combustion installations is expected to
increase the coverage by a further 40-50 million metric tons.
To improve the cost-effectiveness of the ETS and reduce administrative costs, the Directive
provides that small installations may be subject to other control regimes (such as carbon taxes)
rather than included under the EU-ETS. Currently, the smallest 1,400 (10% of total installations
covered) installations emit only 0.14% of total emissions covered. The Directive provides that
Member States may opt to exclude installations that emit less than 25,000 metric tons annually
from the EU-ETS (paragraph 11).
Summary and Considerations for U.S. Cap-andTrade Proposals

The United States is not a party to the Kyoto Protocol and no legislative proposal before the
Congress would impose as stringent or rapid an emission reduction regime on the United States
as Kyoto would have. Likewise, U.S. proposals to reduce emissions through 2020 are not as
stringent as that provided in the EU Directive. However, through five years of carbon emissions
trading, the EU has gained valuable experience. This experience, along with the process of
developing Phase 3, may provide some insight into current cap-and-trade design issues in the
United States.
Emission Inventories and Target Setting

The ETS experience with market trading and target setting confirms once again the central
importance of a credible emissions inventory to a functioning cap-and-trade program. 52 The lack
of credible EU-wide data on emissions was a direct cause of the ETS Phase 1 allowance market
collapse in 2006. Arguably, the most important result of Phase 1 was the development of a
credible inventory on which to base future targets and allocations.
51
(Brussels, January 23, 2008), p. 4.
52
As stated by CRS in 1992: For an economic incentive system to be effective, several preconditions are necessary.
Perhaps the most important is data about the emissions being controlled. Such data are important to levy any tax,
allocate any permits, and enforce any limit. CRS Issue Brief IB92125, Global Climate: Proposed Economic
Mechanisms for Reducing CO2, by Larry Parker (archived November 16, 1994), p. 9.
15
In the United States, Section 821 of the 1990 Clean Air Act Amendments requires electric
generating facilities affected by the acid rain provisions of Title IV to monitor carbon dioxide in
accordance with EPA regulations.53 This provision was enacted for the stated purpose of
establishing a national carbon dioxide monitoring system. 54 As promulgated by EPA, regulations
permit owners and operators of affected facilities to monitor their carbon dioxide emissions
through either continuous emission monitoring (CEM) or fuel analysis.55 The CEM regulations
for carbon dioxide are similar to those for the acid rain programs sulfur dioxide CEM
regulations. Those choosing fuel analysis must calculate mass emissions on a daily, quarterly, and
annual basis, based on amounts and types of fuel used. As suggested by the EU-ETS experience,
expanding equivalent data requirements to all facilities covered under a cap-and-trade program
would be the foundation for developing allocation systems, reduction targets, and enforcement
provisions.
Coverage
Despite economic analysis to the contrary, the EU decided to restrict Phase 1 ETS coverage to six
sectors that represented about 40%-45% of the EUs CO2 emissions.56 This restriction was
estimated to raise the cost of complying with Kyoto from 6 billion euro annually to 6.9 billion
euro (1999 euro) compared with a comprehensive trading program. A variety of practical,
political, and scientific reasons were given by the EC for the decision.57
The experience of the ETS up to now suggests that adding new sectors to an existing trading
program is a difficult process. As noted above, a stated goal of the EC is to expand the coverage
of the ETS. However, the experience of Phase 1 did not result in the addition of any new sector
until the last year of Phase 2 when aviation will be included. The EU will expand its coverage
with Phase 3, but the ETS will still cover fewer sectors emitting greenhouse gases than provided
under most U.S. proposals.
U.S. cap-and-trade proposals generally fall into one of two categories. Most bills are more
comprehensive than the ETS, covering 80% to 100% of the countrys greenhouse gas emissions.
At a minimum, they include the electric utility, transportation, and industrial sectors;
disagreement among the bills center on the agricultural sector and smaller commercial and
residential sources. In some cases discretion is provided EPA to exempt sources if serious data,
economic, or other considerations dictate such a resolution.
A second category of bills focuses on the electric utility industry, representing about 33% of U.S.
greenhouse gases and therefore less comprehensive than the ETS. Sometimes including additional
controls on non-greenhouse gas pollutants, such as mercury, these bills focus on the sources with
the most experience with emission trading and the best emissions data. Other sources could be
added as circumstances dictate.
53
Section 821, 1990 Clean Air Act Amendments (P.L. 101-549, 42 USC 7651k).
S.Rept. 101-952.
55
See 40 CFR 75.13, along with appendix G (for CEMs specifications) and appendix F (for fuel analysis specifications.
54
56
For more background, see CRS Report RL33581, Climate Change: The European Unions Emissions Trading
System (EU-ETS), by Larry Parker.
57
Ibid., p 3.
16
As noted, the EUs experience with the ETS suggests that adding sectors to an emission trading
scheme can be a slow and contentious process. If one believes that the electric utility sector is a
cost-effective place to start addressing greenhouse gas emissions and that there is sufficient time
to do the necessary groundwork to eventually add other sectors, then a phased-in approach may
be reasonable. If one believes that the economy as a whole needs to begin adjusting to a carbonconstrained environment to meet long term goals, then a more comprehensive approach may be
justified. The ETS experience suggests the process doesnt necessarily get any easier if you wait.
Allocation Schemes
Setting up a tradeable allowance system is a lot like setting up a new currency. 58 Allocating
allowances is essentially allocating money with the marketplace determining the exchange rate.
As noted above, the free allocation scheme used in the ETS has resulted in windfall profits
being received by allowance recipients. As stated quite forcefully by Deutsche Bank Research:
The most striking market outcome of emissions trading to date has been the power industrys
windfall profits, which have sparked controversy. We are all familiar with the background:
emissions allowances were handed out free of charge to those plant operators participating in
the emissions trading scheme. Nevertheless, in particular the producers of electricity
succeeded in marking up the market price of electricity to include the opportunity-cost value
of the allowances. This is correct from an accounting point of view, since the allowances do
have a value and could otherwise be sold. Moreover, emissions trading cannot work without
price signals.59
The free allocation of allowances in Phase 1 and 2 of the ETS incorporates two other mechanisms
that create perverse incentives and significant distortions in the emissions markets: new entrant
reserves and closure policy. Combined with an uncoordinated and spotty benchmarking approach
for both new and existing sources, the result is a greenhouse gas reduction scheme that is
influenced as much or more by national policy than by the emissions marketplace.
The expansion of auctions for Phase 3 of the ETS could simplify allocations and permit market
forces to influence compliance strategies more fully. Most countries did not employ auctions at
all during Phase 1 and auctions continue to be limited under Phase 2. No country combined an
auction with a reserve price to encourage development of new technology. The EC limited the
amount of auctioned allowances to 10% in Phase 2: a limit no country chose to meet. Efforts to
expand auctions met opposition from industry groups, but attracted support from environmental
groups and economists. The Phase 3 increased use of auctioning through 2020 will represent a
major development for the scheme.
Currently, all U.S. cap-and-trade proposals have some provisions for auctions, although the
amount involved is sometimes left to EPA discretion. Most specify a schedule that provides
increasing use of auctions from 2012 through the mid-2030s with a final target of 66%-100% of
total allowances auctioned. Funds would be used for a variety of purposes, including programs to
encourage new technologies. Some proposals include a reserve price on some auctions to create a
price floor for new technology.
58
59
Unlike a carbon tax which uses the existing currency system to control emissionsbe it euro or dollars.
Deutsche Bank Research, EU Emission Trading: Allocation Battles Intensifying (March 6, 2007), p. 2.
17
Like the situation in the ETS, most U.S. industry groups either oppose auctions outright or want
them to be supplemental to a base free allocation. Given the experience with the ETS where the
EC and individual governments have been unwilling or unable to move away from free
allocation, the Congress, like the EU, may ultimately be asked to consider specifying any auction
requirement if it wishes to incorporate market economics more fully into compliance decisions.
Flexibility and Price Volatility

Despite EU rhetoric during the Kyoto Protocol negotiations, it moved into Phase 2 without a
significant restriction on the use of CDM and JI credits. This embracing of project credits will
significantly increase the flexibility facilities have in meeting their reduction targets. In addition,
Phase 2 includes the use of banking to increase flexibility across time by allowing banked
allowances to be used in Phase 3. Each of these market mechanisms is projected to reduce both
the EUs Kyoto compliance costs and allowance price volatility. These flexibility mechanisms
will be extended into Phase 3 with modifications.
Unfortunately, Phase 1 experience with the ETS did not provide much useful information on the
value of market mechanisms or financial instruments in reducing costs or price volatility. The
combination of poor emissions inventories, non-use of project credits, and time-limited
allowances with effectively no banking resulted in extreme price volatility in Spring 2006, and
virtually worthless allowances by mid-2007. The real test for the mechanisms employed by the
ETS to create a stable allowance market is Phase 2. Initial indications are that a mature market for
allowances appears to be developing, although, like most commodities markets, the allowance
market can still be volatile at times.
Phase 3 is introducing two new mechanisms in the ETS to further respond to volatility not based
on market fundamentals. However, the actual effectiveness of these mechanisms is yet to be
proven.
Like the ETS, U.S. cap-and-trade proposals would employ a combination of devices to create a
stable allowance market and encourage flexible, cost-effective compliance strategies by
participating entities. All include banking. All include use of offsets, although some would place
substantial restrictions on their use. Some proposals have incorporated a safety valve that
would effectively place a ceiling on allowance prices, while others would create a Carbon Market
Efficiency Board to observe the allowance market and implement cost-relief measures if
necessary. Finally, some incorporate strategic reserves auctions, similar in concept to the EU
forward auctioning mechanism, to increase allowance supply without busting the emission cap.
Some see this as a more flexible response with the potential for avoiding or mitigating the
environmental impacts of a safety valve (i.e., increased emissions).
Additionally, concern has been expressed in the United States about the regulation of allowance
markets and instruments. Based on experience with the ETS, the potential for speculation and
manipulation could extend beyond the emission markets. Analysis of ETS allowance prices
during Phase 1 suggests the most important variables in determining allowance price changes
were oil and natural gas price changes.60 This apparent linkage between allowance price changes
and price changes in two commodities markets raises the possibility of market manipulation,
60
Maria Mansanet-Bataller, Angel Pardo, and Enric Valor, CO2 Prices, Energy and Weather, 28 The Energy Journal
3 (2007), pp. 73-92.
18
particularly with the inclusion of financial instruments such as options and futures contracts. The
concern is sufficient for the Directive to require the EC to examine the situation and the current
protections against such activities. Congress may ultimately be asked to consider whether the
Securities and Exchange Commission, Federal Energy Regulatory Commission, the Commodities
Futures Trading Commission, or other body should have enhanced regulatory and oversight
authority over such instruments. 61
Author Contact Information

Larry Parker
Specialist in Energy and Environmental Policy
lparker@crs.loc.gov, 7-7238
61
For a discussion of regulation of allowances as a commodity and implications for a greenhouse gas emissions
market, see CRS Report RL34488, Regulating a Carbon Market: Issues Raised By the European Carbon and U.S.
Sulfur Dioxide Allowance Markets, by Mark Jickling and Larry Parker.
19
Project Finance Primer for

Renewable Energy and Clean Tech Projects
Authors: Chris Groobey, John Pierce, Michael Faber, and Greg Broome
Executive Summary
Investments in the clean technology sector often combine capital intensity with new technologies. Securing project
finance can prove to be a critical step in the path to commercialization. Project finance succeeds best when you have
long-term off-take agreements with quality-credit counterparties (such as power purchase agreements) but commoditybased projects that sell into open markets (such as biofuels) can also benefit from the project finance model.
This primer provides an overview of project finance for renewable energy investors, with a focus on the pros and cons,
as well as a survey of key concepts and requirements, including tax incentives and monetization strategies in the
renewable energy sector, and other key structuring considerations in determining whether to project finance.
Key Points
Project finance has emerged as a leading way to finance large infrastructure projects that might otherwise be
too expensive or speculative to be carried on a corporate balance sheet.
The basic premise of project finance is that lenders loan money for the development of a project solely based
on the specific projects risks and future cash flows. As such, project finance is a method of financing in which
the lenders to a project have either no recourse or only limited recourse to the parent company that develops
or sponsors the project.
For equity investors, the appeal of project finance is that it can maximize equity returns, move significant
liabilities off balance sheet, protect key assets and monetize tax financing opportunities. A wide range of
commercial and legal issues must be addressed to secure adequate returns. Tight credit markets exacerbate
competition for long-term financing, so even small differences in deals can impact the availability of financing
or reduce leverage.
Project financing became particularly important to project development in emerging markets, with participants
often relying on guarantees, long-term off-take or purchase agreements, or other contractual relationships with
the host sovereign or its commercial appendages to ensure the long-term viability of individual projects. These
were typically backstopped by multilateral lending agencies that mitigated some of the political risks to which
the project lenders were exposed. Analogies to alternative energy projects help investors de-risk higher-risk
new technologies.
Austin
New York
Palo Alto
San Diego
San Francisco
Seattle
Shanghai
Washington, D.C.
Part I of the primer introduces project finance to those that may be less familiar with the concept, and asks questions
that will assist investors and developers in determining whether project finance is appropriate for their renewable
energy projects. Part II sets out the legal and contractual structure that will facilitate project financing. Part III
describes the process of obtaining equity investment and some of the important options and considerations that
companies may have in that process. Part IV provides a more in-depth look at what a typical renewable energy project
financing looks like, including fundamental structural components that characterize any project finance transaction.
Finally, Part V outlines key tax incentives currently available in the renewable energy industry, as well as monetization
strategies that may be useful for earlier-stage energy companies unable to directly utilize such tax incentives.
Given the breadth of the current renewable energy landscape, this primer focuses on a hypothetical solar generation
facility (Solar Project) as the primary case study with discussions of other renewable energy projects (wind power
and biofuel projects in particular) as appropriate. In general, once the contracts related to a project are negotiated
(which is described in Part II), the mechanical aspects of raising equity and project financing are likely to be similar
across various renewable technologies, although investor enthusiasm and financing prices and terms are likely to vary
significantly across technologies at any given time.
I. Introduction to Project Finance

A.
What Is Project Finance?
The basic premise of project finance is that lenders loan money for the development of a project solely based on the
specific projects risks and future cash flows. As such, project finance is a method of financing in which the lenders to
a project have either no recourse or only limited recourse to the parent company that develops or sponsors the
project (the Sponsor). Non-recourse refers to the lenders inability to access the capital or assets of the Sponsor to
repay the debt incurred by the special purpose entity that owns the project (the Project Company). In cases where
project financings are limited recourse as opposed to truly non-recourse, the Sponsors capital may be at risk only for
specific purposes and in specific (limited) amounts set forth in the project financing documentation.
Project financing has been used in various ways for many years, but in the 1970s and 1980s it emerged as a leading
way of financing large infrastructure projects that might otherwise be too expensive or speculative for any one
individual investor to carry on its corporate balance sheet. Project financing has been particularly important to project
development in emerging markets, with participants often relying on guarantees, long-term off-take or purchase
agreements, or other contractual relationships with the host sovereign or its commercial appendages to ensure the
long-term viability of individual projects. These were typically backstopped by multilateral lending agencies that
mitigated some of the political risks to which the project lenders (and, sometimes, equity investors) were exposed.
B.
What Underpins Project Finance?
As a general (if not universal) rule, lenders will not forgo recourse to a projects Sponsor unless there is a projected
revenue stream from the project that can be secured for purposes of ensuring repayment of the loans. In the case of
large wind and solar power projects, this revenue is typically generated from a power purchase agreement (PPA)
with the local utility, under which the project may be able to utilize the creditworthiness of the utility to reduce its
borrowing costs. While the wind power market has matured significantly in the past five years, leading to the
successful project financing of merchant projects in the absence of long-term PPAs, Solar Projects are generally not
yet able to be project financed in such a manner. In merchant power projects, lenders are able to receive assurance of
the projects ability to repay its debt by focusing on commodity hedging, collateral values, and the income to be
produced based on historical and forward-looking power price curves and fully developed markets. In non-power
generation contexts, the projects revenue stream may be a long-term operating agreement (e.g., in the case of toll
roads), a capacity purchase agreement (e.g., in the case of transmission lines), a production sharing agreement (e.g.,
in the case of oil field development), or a series of short-term and spot sales into commodity markets (e.g., in the case
of biofuels projects).
While project finance lenders clearly prefer a long-term contract that ensures a relatively consistent and guaranteed
revenue stream (including assured margins over the cost of inputs), in the context of some industries, lenders have
determined that sufficient revenues to support the projects debt are of a high enough probability that they will provide
debt financing without a long-term off-take agreement. Solar Projects, due to their peak period production, high marginal
costs, and lack of demonstrated merchant capabilities, are not at this time viewed as project financeable without PPAs
that cover all or substantially all of their output. Solar Projects lack of merchant viability is exacerbated by the fact that
the southwest United States (the region most appropriate for utility-scale solar power development) does not have a
www.wsgr.com
mature merchant power market that functions in the absence of long-term bilateral sales agreements. The dependence
of large-scale solar projects on the PPA model is not expected to change in the short to intermediate term.
C.
When to Project Finance?
One of the primary benefits of project financing is that the debt is held at the level of the Project Company and not on
the corporate books of the Sponsor. When modeling projects and projected income, the internal rate of return of
Sponsors and other project-level equity investors can increase dramatically once a project is fully leveraged. Sponsors
are frequently able to recover development costs at the closing of the project financing and put their money into other
projects. Another benefit of project financing is the protection of key Sponsor assets, such as intellectual property, key
personnel, and investments in other projects and other assets, in the case of the Project Companys bankruptcy, debt
default, or foreclosure. Moreover, project financing allows for a wide variety of tax structuring opportunities, particularly
in the context of monetizing tax incentives (discussed further in Part V). On the other hand, project financing is
document-intensive, time-consuming, and expensive to consummate. It is not atypical that administrative and closing
costs, when factoring in lenders, consultants, and attorneys fees for all parties, equal several percentage points of the
amount of the loan commitment. Moreover, project financing imposes significant operating restrictions on each Project
Company, including its ability to make equity distributions to the Sponsor prior to the payment of operating expenses,
debt service, and a percentage sweep of additional cash flow (discussed further in Part IV). The result is that the
decision of whether to reinvest cash flow in the project does not rest solely with the Sponsor.
Given the pros and cons of project finance, the most relevant initial inquiry for an investor or developer may be when is
project financing possible or most appropriate? The following questions should be useful in determining if project
financing is a realistic opportunity for any given company:
-
Is there an individual project or group of projects of a sufficient size to make either a standalone or
portfolio project financing worthwhile? Typically lenders will be reluctant to provide project financing if the
total amount of debt is less than US$50 million and, preferably, US$100 million.
Will there be a revenue stream from the project large enough to support a highly leveraged debt financing?
This is a prerequisite for project financing.
Will the receipt of revenue be enforceable under contractual rights against a creditworthy party? This is
not necessarily a prerequisite for all project financings, but the absence of a contract, or questionable
creditworthiness of the purchaser, will prompt lender skepticism and necessitate thorough due diligence
regarding future revenue projections.
Will there be physical assets sufficient to ensure lender repayment in case of foreclosure? Lenders will
want to know that even if the Project Companys projected revenue stream does not materialize, they will
be able to foreclose on the projects assets sufficient in value to make themselves whole, either by selling
the project outright or operating it until the debt is repaid.
Is there a significant level of technology risk? While in many project financings, technology may be
relatively new or cutting edge, project finance lenders almost never want to be the first to finance an
untested technology. Demonstrated successful use in some context will often be necessary to secure
project financing.
Does the project have contractual relationships with reputable companies for services key to the success
of the project or the technology it employs? Lenders will be less likely to lend to a project the success of
which depends solely on a few talented individuals who may depart, leaving the project unable to meet its
potential.
Is the Sponsor ultimately willing to risk the project? In other words, once project financing is completed,
the Sponsor loses the ability to determine how the vast majority of the projects revenue is spent. In the
event a project becomes uneconomic and unable to service its debt, the only option besides refinancing
the debt may be to turn over the project to the lenders (voluntarily or involuntarily), with the corresponding
loss of the Sponsors investment in the project.
Is the Sponsor looking for a quick exit? Once project-financed, divestiture opportunities are complicated
by the requirement of lender consent, and potential purchasers will be thoroughly examined by lenders for
development and operational expertise as well as creditworthiness.
www.wsgr.com
Are Sponsors willing to grant rights of high-level oversight regarding the projects development and
operation to project finance lenders? In many cases the interests of the Sponsor and the lenders will be
aligned, and lenders will tend to defer to the Sponsors developmental expertise. On the other hand,
lenders must be viewed as additional project partners, with veto rights over many significant decisions.
Assuming project financing is a viable option, Part II provides a roadmap to structuring a project financing transaction.
II. Establishing a Project Structure and Negotiating Project Agreements

A.
Project Structure
The project finance structure revolves around the creation of the Project Company that holds all of the projects assets,
including all of its contractual rights and obligations. The Project Company is usually a single-member limited liability
company, although in some cases it may be a limited partnership.
In most cases, the equity interest in the Project Company will be held by at least one intermediate holding company,
usually a limited liability company (the Holdco), created for the purpose of pledging the Project Companys equity to the
lenders in the eventual project financing. While the Holdco will have a separate legal identity, typically it will not have any
business apart from holding the equity of the Project Company. This structure allows for most liability to be contained at
the bankruptcy-remote Project Company level, and thus insulates the Sponsor (including equity investors in the Sponsor)
and the Holdco from liability to either the Project Companys contractual counterparties (Counterparties) or to the
Holdcos lenders. In order to ensure that the Project Company is treated as a separate legal entity, it will be necessary to
have governance mechanisms at the Project Company level that are independent, including designated officers, at least
one independent director, and internal controls and procedures designed to preserve a legal entity distinct from the
Sponsor and the Holdco.
B.
Project Agreements Overview
As a general matter, all contracts related to the development, construction, ownership, and operation of the project will
be entered into by the Project Company (Project Agreements). If development-stage contracts have been executed
by the Sponsor or one of its affiliates, it is important that the contracts allow for their assignment to the Project
Company once the Project Company has been established for the purposes of pursuing project financing.
In addition to the external Project Agreements, there may be several intercompany agreements between the Project
Company and the Sponsor or its affiliates. These may include an Operation and Maintenance Agreement (O&M), an
Administrative Services Agreement (ASA), and a Technology License Agreement (TLA), often with affiliates of the
Sponsor created specifically for the purpose of providing administrative support, operation, and maintenance services
and holding the intellectual property for the benefit of one or more of the Sponsors projects. In other cases, unrelated
third parties may provide these services to the Project Company. If intercompany agreements are used, they should
be structured in such a manner as to track the material commercial terms that the Sponsor could obtain with an
unrelated third party providing the same services.
Intercompany agreements can also have a significant impact on the total return of a project to its investors, so their
economic terms must be carefully crafted. Assuming the O&M, ASA, and TLA are entered into with Sponsor affiliates,
they permit the affiliates to extract arms length fees for the provision of key services and technology to the Project
Company on a monthly or quarterly basis; these fees are frequently paid prior to repayment of debt. The
intercompany-agreement structure also allows the Sponsor, if the project fails following the project financing, to retain
all of its employees that provide services to the Project Company, thereby ensuring that key employees (and knowhow) will not be lost to lenders or a subsequent purchaser out of foreclosure. In such a scenario, the TLA will also
allow the Sponsor to retain ownership of its technology subject only to a license right on the part of the Project
Company which may no longer be affiliated with the Sponsor. These are especially critical points where the Sponsor
has multiple projects that may utilize the same technology, support equipment, and personnel. In addition, the O&M,
ASA, and TLA provide the Project Company's lenders contractual certainty (through the agreements themselves as
well as the corresponding consents to collateral assignment (discussed further in Part IV below)) that key services will
continue if the Project Company defaults, thereby increasing the likelihood of the efficient development, construction,
and operation of the project and the preservation of the value of the lenders' collateral.
There are many other Project Agreements that are typically executed during the course of developing and constructing
a renewable energy project. The Project Agreements may include one or more PPAs, which may have an income
www.wsgr.com
stream payable from an off-taker for energy payments, capacity payments, or both; an Engineering, Procurement, and
Construction Agreement (EPC Agreement); a Site Lease Agreement (if the projects land is not owned by the Project
Company itself); a Renewable Energy Credit Agreement (in states where applicable); an Interconnection Agreement
(for projects tied to the electricity grid); agreements for the provision of utility services; agreements for the provision of
feedstock commodities (in the case of biofuels) and the necessary price and supply hedging; agreements including
equity flip structures to take advantage of the federal tax incentives discussed in Part V below; and other Project
Agreements necessary or desirable to develop, construct, own, or operate the project. In some cases certain
byproducts of production may be sold in addition to the primary product (for example, steam as a byproduct of cogeneration power projects, high protein distillers grains as a byproduct of ethanol production, or carbon dioxide where
markets exist).
Typical Project Finance Structure

Equity Investors
Equity
Contribution $
Operating or Shareholders
Agreement
Sponsor
Equity
Contribution $
Project-Level Equity Investors
Operating or Limited
Partnership
Agreement
Equity
Contribution $
Lenders
Operating or Limited
Partnership
Agreement
Loans $
Loan
Documentation
Project Company
aka The Borrower
Interconnection
Agreement
Interconnecting
Utility
Consent
Lenders
C.
PPA
Purchasing
Utility/
Offtaker
Consent
Lenders
EPC or
EPCM
Agreement
Contractor
Consent
Lenders
Supply
Contracts
REC
Purchase
Agreement
Equipment
Suppliers
Consent
REC
Purchaser
Consent
Lenders
Lenders
O&M
Agreement
O&M
Service
Provider
Consent
Lenders
Admin
Services
Agreement
Technology
License
Agreement
Admin
Services
Provider
Consent
Lenders
Technology
Provider
Consent
Lenders
Key Project Agreement Terms
In the process of negotiating the Project Agreements it will be necessary to consider key project finance principles to
prevent having to revisit contractual terms at the lenders behest in the course of financing the project. One overriding
concept is that lenders will own (and likely seek to immediately transfer) the Project Company in the case of
foreclosure, thus will insist on contractual rights and terms that ensure a seamless transition to the lender or
subsequent owner. To this end, the project lenders will require consents to collateral assignment (Consents) for their
benefit with some if not all of the Counterparties. Therefore, provisions that prevent assignment without Counterparty
consent should be omitted from Project Agreements. Inclusion of contractual language that obligates the Counterparty
to cooperate with the Project Company and its lenders in the course of the financing process will not only expedite the
process of negotiating the Consents but will also reduce the scope for Counterparty intransigence in the context of the
project financing.
The commercial terms of the PPA and the EPC Agreement, together with the market and technology risks, will largely
determine whether lenders view the project as financeable. Foremost among considerations related to the PPA will
be whether or not there is a guaranteed revenue stream (usually energy payments from the actual production of
power) from a creditworthy purchaser that will be sufficient to support the economics of the project, thereby ensuring
www.wsgr.com
prompt repayment of debt and mitigating the risk of default. The PPA term should also be sufficient in length to fully
amortize the contemplated project debt. In contrast to most smaller distributed generation projects, where an off-taker
pays for only the power that is produced, utility scale solar generation facilities may have take or pay PPAs where the
utility is still required to pay the Project Company even if a certain level of power is not purchased. In the case of Solar
Projects, utilities are less likely to deliver capacity payments because power is generally produced only during daylight
hours. However, distributed generation projects may benefit from more certain payments because the distributed
project produces all of the power for a given host.
If a project does not have a PPA or other off-take contract, demonstrated merchant operating histories of similarly
situated plants (more relevant in the context of wind projects) will be necessary to convince lenders of the reliability of
forecast ratios. Even with long-term PPAs, lenders will still look for additional data to support viability such as
meteorological wind data for wind power sites over the course of one to two years, often at installed hub-heights, or
long-term temperature and sun data for Solar Projects. The trend in biofuels project financings is also moving toward
contracted off-take arrangements with a creditworthy purchaser for all of a plants production. At least in the short and
intermediate terms, most project-financed Solar Projects will have PPAs for a significant portion, if not all, of their
generated power. While PPAs for large-scale CSP projects will generally be far more complex than those for smaller
distributed PV projects, in either case, the core economic terms will determine the lenders view of whether a project or
a portfolio of projects is viable from a revenue perspective and, accordingly, financeable on favorable terms.
To the extent a project is not fully constructed by the time project financing is sought, EPC Agreements will be an
integral part of the financing analysis and pricing. While larger developers may be able to finance an entire project on
balance sheet, and subsequently refinance the development to free up invested capital, most developers seek to
leverage their equity and use project finance to construct and operate their projects. Where construction risk is
present, lenders will generally seek corporate parent guarantees, performance bonds, or other forms of performance
surety that ensure that the performance of the contractor is as close to budget and schedule as possible. Warranties
of appropriate substance and duration as well as subsequent maintenance coverage regarding the EPC work and the
equipment purchased will be necessary to convince lenders that significant unbudgeted expenses will not be incurred
by the Project Company. With respect to an EPC contractor, lenders prefer a full wrap EPC agreement because
such an agreement provides a single point of contact with regard to the various risks such an agreement might contain
(warranties and schedule and performance guarantees, among other things). This is particularly the case with newer
and untested technology even if operationally superior to previous generation technology. Liquidated damage
coverage (pre-agreed payments made by the contractor) for schedule and performance delays, inefficiency, or
equipment failures also reassure lenders that a project has the necessary protection against delays or performance
defects that are within the EPC contractors control. How much of the risk an EPC contractor accepts for cost overruns
and design or installation defects, when viewed with other contractual terms, will affect the lenders view of whether a
project is financeable and at what cost. For example, a project that is not financeable at 80% debt due to certain offtake or technology risks may be financeable with 40%60% debt because the lenders are taking less risk with a higher
level of capital pre-paid into the project. Dedicating sufficient resources at the negotiation stage of PPA and EPC
Agreements to achieve commercial and contractual terms as favorable as possible will usually pay dividends at the
financing stage by saving not only money but also costly renegotiation and valuable time toward project completion.
III. Raising Equity

Venture capital and private equity investors also serve as attractive sources for capital raising, as an increasing
number of funds are investing in renewable energy and clean technologies. The large sums of capital required to
initiate and complete renewable energy projects drive not only the selection of appropriate equity investors but also the
structure of such investments. Therefore, Sponsors and investors evaluating equity solar investments should consider
the following action items:
A.
Conduct an internal assessment of capital and budgeting strategies for the investment.
Sponsors and investors should conduct an internal evaluation of whether an equity investment would best serve their
respective strategic objectives. Salient considerations include:
-
Is the companys technology reliable enough to be considered financeable, and is there a realistic potential
pool of equity investors from which to draw?
Are there any gaps in the Sponsors existing organizational structure and operations that an equity
investor would want filled before engaging in substantive discussions and/or closing an equity round of
www.wsgr.com
financing? For example, given the complexity involved in successfully executing Solar Projects, equity
investors will look for experienced management with skill and connections within the industry, as well as
potentially requiring contractual commitments from relevant third parties in the supply chain and customer
base.
B.
How much capital investment is realistically required for the Solar Project? Has Sponsor management, on
the one hand, conducted a thorough analysis of the timing and amount of future capital needs and relevant
burn rates, and has the investor and its syndicate, on the other hand, assessed whether its proposed
financing will be sufficient to either execute the Solar Project or bridge the Sponsor towards its next round
of investment?
What type of investment is ideally suited for the particular Solar Project e.g., is the Sponsor seeking passive
investment, or an active strategic partner that will add value to the organization (as discussed in greater detail
below)?
How would an equity investment impact the Solar Projects existing grants, tax treatment, eligibility for
applicable federal and state incentive programs, and contractual obligations?
As equity investors become shareholders, and in many cases, directors of the Sponsor, how much control
is the Sponsor willing to give to the investor, and what level of control does the investor desire in order to
have an active voice within the organization?
Determine whether the investment will add value to the Solar Project.
Unless a Sponsor is seeking a purely passive equity investment, the Sponsor and its investor should conduct a
thorough assessment of the investors role in driving value to the enterprise by, among other things:
C.
Reviewing the investors existing portfolio companies to determine whether the investor has previously
invested in similar projects or has other relevant experience with alternative energy investments. It is
important to find investors who understand the longer time period required to execute and obtain a return
on investment from renewable energy projects;
Meeting with the investors key decision makers to assess how the investor will add value in addition to the
capital infusion e.g., through participation on the board of directors, introductions to potential customers,
assistance in financial forecasting and planning, and guidance in analyzing potential liquidity events; and
Assessing potential conflicts of interest that may arise to the extent that an investor has, for example, a
competitor as one of its portfolio companies.
Assess the appropriate structure for the equity investment.
Once an appropriate equity investor has been identified, the equity investment typically will proceed to the preparation
of a term sheet that identifies the key terms of the investment, as well as a diligence request and the execution of a
confidentiality agreement to facilitate the exchange of information to the investor for the investors due diligence
purposes. A term sheet is a helpful means of assessing whether the parties truly see eye-to-eye with each other on
the critical aspects of the investment before expending significant time and expense negotiating definitive documents,
and may include the following terms:
-
The identification of the relevant entity that will receive such funds (e.g., will the investment be made into a
special purpose vehicle solely created for the project (for example, a Project Company) or will the
investment be made into the Sponsor which may hold assets unrelated to the project);
The amount of the investment, as the Sponsor should ensure that it receives sufficient capital to minimize
future dilutive cram-down financings, but also not take in more capital than is needed as this also will
have dilutive effects to existing shareholders. Milestone-based investments may help serve to mitigate
Sponsor risk in terms of securing additional future financing, while helping investors stage their investment
to ensure that the Sponsor can meet specific financial and commercial targets before disbursing additional
funds; and
Whether the equity security will be common stock, which is typically issued to founders, optionees, and
angel investors, or preferred stock, which not only is senior to the common stock in preference but also
typically has additional terms and conditions that increase preferred stockholders return on investment
and control over the Sponsor such as:
www.wsgr.com
D.
The right to appoint one or more board members;
Dividend rights;
A liquidation preference, which is the right to receive a preferential return on investment in the event
of a liquidity event such as a merger, asset sale, or change of control;
A redemption right, which is the right to redeem the equity securities at an agreed-upon point in the
future;
Anti-dilution rights, which protect an investor from the dilutive effect of future equity issuances; and
Protective provisions, which allow the preferred holders certain veto rights over key corporate actions.
Have candid discussions to ensure that expectations are aligned on key business issues, including:
-
The market opportunity;
The companys ability to execute its business plan, and the investors commitment to both the initial and
subsequent capital needs during the companys life cycle;
A realistic commercialization timeline, use of proceeds, and the expected internal rate of return of the
Solar Project; and
The appropriate liquidity event, be it an acquisition or an initial public offering, and how the investor can
add value to facilitate a liquidity event (e.g., by assisting in pre-public corporate governance compliance
required by Sarbanes-Oxley, or introductions to key strategic partners, customers, and potential acquirers
in the future).
In summary, both Sponsors and investors analyzing an equity investment should conduct a realistic assessment of the
companys capital needs, structure an investment that can add value to the company and its projects, and seek to
create a mutually beneficial working relationship where expectations on key business issues between the Sponsor and
the investor are aligned.
IV. Time to Project Finance

Before beginning an examination of the project financing process, it is worth noting different options for raising debt in
the context of project development. The three most frequently utilized project financing structures are the syndicated
or club loan, the issuance of project bonds through a private placement, and the issuance of Term B loans.
A.
Syndicated and Club Loans and Project Bonds
Currently the majority of renewable energy projects are financed through the syndicated commercial loan market.
Syndicated loans are loans in which a group of banks each take a portion of a larger loan and thus minimize the risk
that any one individual lender making the same loan would otherwise have. A syndicated loan transaction is usually
coordinated by one or more arranger banks whereas in club deals a handfull of lenders take equal roles in leading
the transaction and lending to the project. An alternative to the syndicated loan market is the private placement of debt
through 144A offerings, which are exempt from registration with the SEC if the purchasers are Qualified Institutional
Buyers as defined in the Securities Exchange Act of 1933. The issuer of 144A bonds could be either the Project
Company or the Sponsor.
Syndicated loan structures are often preferred to accessing the capital markets through 144A offerings, because
capital markets investors are generally less likely to assume construction risk and the disclosure documentation for a
144A offering is generally more extensive than that prepared in connection with syndicating a commercial loan. In
addition, amounts raised through a 144A issuance are all disbursed at closing, which leads to negative carry
implications. Moreover, private placements or corporate level offerings tend to be fixed rate, which, while providing
certainty, removes the upside potential of floating rates that are available pursuant to commercial bank loans. On the
other hand, 144A bond offerings are generally completed more quickly and inexpensively than a syndicated project
loan, the covenants contained in the governing documentation may be less restrictive, and the repayment period of
www.wsgr.com
private placement debt offerings is generally longer. Bonds can also pay interest at tax-exempt rates (lowering the
borrowers borrowing cost), be issued in relatively small amounts (making them ideal for smaller project financings)
and carry implied or explicit credit support from government instrumentalities (again reducing borrowing costs).
B.
Term B Loans
Several years ago, Term B loans emerged as a subset of the project lending market and were characterized by shorter
tenors and lower or delayed amortization, often with bullet payments due at maturity. Correspondingly, Term B loans
carried higher risk profiles and usually were rated non-investment grade. In addition, the terms and conditions of Term
B loans tended to be less onerous than traditional project debt that amortized over a longer period. As a result of the
subprime lending crisis and the resulting credit crunch, the Term B loan market all but disappeared and has yet to reemerge. For purposes of the following discussion, due to the considerations set forth above and the lull in the Term B
market, we assume that a traditional bank syndication model of project financing will be most beneficial to the Sponsor.
Although the terminology may differ from transaction to transaction, the documentation for such a project financing is
governed by a credit or financing agreement (Credit Agreement) and at a minimum will include an asset security and
equity pledge agreement, a mortgage, and various Consents.
C.
Loan Types
Depending on the development stage of the project, and within the project finance framework, the Sponsor may on
behalf of the Project Company seek construction loans, term loans, working capital loans and/or a letter of credit
facility. Construction loans, as the name implies, are utilized only for the period that the project is under construction.
The interest rate can be higher vis-a-vis a term loan (reflecting increased risk to the lenders during the construction
period) but more frequent drawdowns of construction loans are permitted and at the end of the construction loan
availability period, the construction loan usually converts to a term loan. Term loans are characterized by a set and
limited commitment or drawdown period and an extended amortization period. Term loans can have a lower interest
rate than construction loans, and have scheduled (quarterly or otherwise) repayment dates or set amortization
schedules. The conversion from a construction loan to a term loan often coincides with the definition of Substantial
Completion or Final Completion under the EPC Agreement, and a failure to achieve such conversion by a certain
date will cause a default under the construction loan and accelerate the debt due thereunder.
Working capital loans, which are used primarily for ordinary course expenses such as inventory purchases, are
generally sized smaller than construction or term loans and are subject to a maximum available amount tied to the
value of a Project Companys inventory and cash (often 80%). Working capital loans are usually revolving in nature,
meaning that amounts borrowed can be reborrowed once they are repaid. Letters of credit are made available on the
Project Companys behalf usually for the benefit of third parties under the Project Agreements for example, if a letter
of credit is required as credit support under a PPA, an EPC Agreement or for the provision of utility services. Draws by
a third party on an outstanding letter of credit will operate to reduce the amount of working capital loan availability.
The term of a project finance loan will vary depending on the term of the principal off-take agreement. To minimize risk
profile and lower borrowing costs, loans will ideally amortize in full prior to the end of the term of the PPA. The
borrowing costs of a renewable energy project will invariably depend on the risk profile determined by the
characteristics of the project itself, in particular the lenders view of the likelihood that the project will default on its
loans. In addition, exposure to merchant markets or other off-take risk will increase borrowing costs relative to projects
that have a long-term PPA, particularly one that is fixed price with take-or-pay terms. The reduced risks that come with
long-term PPAs prevent Sponsors from taking full advantage of arbitrage opportunities that may become available in
the spot market if power prices rise, as price risk is avoided for the producer. Recent trends have seen a wide swing in
loan terms and it is difficult to provide standard pricing terms, although as a general rule rates are higher and fees
have increased, while internal credit reviews have become more stringent and less forgiving of unmitigated project
risks or even minor holes or errors in Project Agreements. Typical lending fees for a project financing include the
following: (i) two percent (2%) to six percent (6%) of the aggregate loan commitment as an arranging or structuring fee,
(ii) one percent (1%) of the aggregate loan commitment as a syndication fee, (iii) $75,000 administrative agency fee to
be paid annually, (iv) $50,000 collateral agency fee to be paid annually, and (v) facility fees to each lender in the
syndicate in an amount between three-quarters of one percent (.75%) and one and one-half percent (1.5%) of each
lenders commitment. In addition, the Project Company will be required to pay the professional fees and administrative
expenses of each of the lenders in evaluating the transaction, negotiating the loan documents, and providing the loans.
Despite the non-recourse nature of pure project financing, in some transactions lenders will seek guarantees for
certain obligations of the Sponsor or its affiliates, either to ensure construction of the project or to ensure that the
Project Company is sufficiently capitalized to meet its debt service requirements. While by no means a requirement in
www.wsgr.com
all transactions, under certain market conditions, a guaranteed (or limited recourse) project finance structure may be
the only way to finance one or more projects or to obtain reasonably priced project debt.
D.
Security Package
1. Overview
Project finance requires the pledge of a comprehensive collateral security package to the lenders in exchange for the
making of loans. The collateral security package, in the absence of recourse to the Sponsor, serves as the basis for
the lenders securing repayment in the case of default. Specifically, all assets of the Project Company owned at the
time of the loan closing, in addition to those acquired post-closing, will be pledged to the lenders until the loans are
fully repaid. Included in the assets to be pledged will be all of the Project Companys personal property, accounts
receivable, contractual rights, and intellectual property. The Project Companys real property is pledged to the lenders
pursuant to a mortgage. A pledge of the equity interests in the Project Company is executed by the Holdco or any
other entities that directly hold equity interests.
As part of the collateral security package, the lenders will require a Consent from some or all of the Counterparties.
The Consent negotiation process can be time consuming and even contentious, especially if the interests of the
Sponsor and the Project Company on the one hand, and the Counterparty on the other, are not aligned. To
complicate matters, lenders may use the process of Consent negotiation to incorporate amendments to the relevant
Project Agreement, which are likely to benefit the Project Company as well as the lenders, but at which the
Counterparty may balk as a renegotiation of the fundamental business agreement embodied in the Project Agreement.
Even fundamental Consent terms such as the extension of cure periods for defaults for the benefit of the lenders in the
event the Project Company does not cure may be viewed as an unfavorable renegotiation from the perspective of a
Counterparty. In addition, some Counterparties are hesitant to enter into a contractual relationship with a large
financial institution as a putative future partner. The prospect of perceived bargaining asymmetry often complicates
what may be tedious three-way negotiations between the Counterparty, the Project Company, and the lenders, with
the Project Company likely playing the role of honest broker in order to facilitate prompt agreement and closure of the
financing.
2. Distribution of Project Revenues
Almost all project financed loans have what is referred to as the project waterfall. All revenues received by the
Project Company are placed in a master project revenue account, which serves as the top of the metaphorical
waterfall. As the money flows down the waterfall it is siphoned off into segregated secured accounts at each different
level as described in an Accounts or Disbursement Agreement, with any funds remaining at the bottom of the waterfall
being paid, assuming there are no defaults and that certain financial tests are met, to the equity owners of the Project
Company. Typically, the project waterfall is structured (roughly) in a manner as described below, with most
withdrawals from the waterfall occurring on a monthly or quarterly basis as appropriate:
-
The first level of payment would be in an amount necessary to pay costs incurred by the Project
Company (i.e., construction and/or operation and maintenance expenses depending on the projects
stage of development), including pre-approved reasonable amounts paid to the Sponsors affiliates
under the O&M, the ASA, and the TLA;
The second level of payment would be to the lenders to pay (i) loan fees and expenses, (ii) interest
payments, and (iii) principal payments (in this order);
The third level of payment will be used to fill an account segregated for the purposes of paying future
debt service in times of lower project revenues, although once this account has been filled to the level
of the required amount no amounts will be taken out at this level;
The fourth level of payment is often referred to as a cash sweep in which the lenders are repaid
outstanding principal with a certain percentage of the excess cash (generally one-third or half, which
increases in a default scenario) remaining after the operation of the three waterfall levels above;
The fifth level of the waterfall may operate to fill one or more reserve accounts, often designated for
future major maintenance or other purposes, but once the reserve account is filled with the required
amount no amounts will be taken out at this level;
The sixth level of the waterfall may be used to repay the holders of subordinated debt or bondholders,
if applicable; and
10
www.wsgr.com
The seventh level of the waterfall allows for cash remaining after amounts have been removed at the
higher levels to be paid to the equity holders of the Project Company in the form of an equity
distribution, assuming there are no defaults and that financial tests are met.
While every project waterfall will operate somewhat differently and many will have features unique to specific project and
financing arrangements, the waterfall operation outlined above is generally standard in project financing arrangements.
Typical Project Finance Waterfall

Project Revenues Received
$
Project Revenues Account
Project Construction/Operating
Expenses
Construction / Operating Account
Fees, Interest & Scheduled Principal
Debt Payment Account
Maintain Required Debt Service

Reserve Level
Cash Sweep to Lenders
Debt Service Reserve Account
Maintain Required Major Maintenance

Reserve Level
Major Maintenance Reserve Account
Payment of Subordinated Debt

(if any)
Subordinated Debt Account
Remaining amount distributed to equity

holders (assuming no defaults and
financial tests are met)
Distribution Account
E.
Operating Restrictions
Project finance lenders place restrictions and affirmative obligations on the Project Company that significantly impact
its day-to-day operation. While many of the affirmative obligations in particular may seem like ordinary course of
business operations, and the affirmative obligations and restrictions taken individually may not seem particularly
onerous, on a collective basis compliance with these obligations and restrictions requires time and effort from the
Sponsors employees. It is worth noting in connection with the time consuming nature of complying with the covenants
set forth in project financing documentation that there may be certain economies of scale, particularly where the
individual projects are smaller, to arranging project financing on a portfolio basis.
More specifically, project finance lenders will require that the Project Company (i) comply with all laws and regulations,
including permits, (ii) construct and operate the project in accordance with prudent industry standards, (iii) pay its debts
and obligations as they become due, (iv) use proceeds received and cash flow as set forth in the financing
documentation (including operation of the waterfall), (v) maintain pre-determined (and generally quite comprehensive)
insurance coverage, (vi) maintain books and records in accordance with GAAP, (vii) adopt and update budgets,
(viii) permit independent verification by the lenders representatives of performance tests, (ix) maintain in effect all
Project Agreements, (x) preserve title to all assets, (xi) update the financial model, (xii) maintain the liens granted
under the security documentation, and (xiii) enter into pre-approved hedging arrangements both for commodity inputs
for example, natural gas in the case of a power project or feedstock in the case of a biofuel plant and for
purposes of interest rate protection. This list is far from comprehensive in scope or detail. In addition, comprehensive
11
www.wsgr.com
reporting requirements will be set out in the Credit Agreement that obligate the Project Company to provide the lenders
with copies of everything from construction status reports to auditors letters to notices of certain adverse events.
Prohibitions placed on the Project Company by the financing documentation will likely include (i) incurring
indebtedness subject to certain exceptions, (ii) incurring liens subject to certain exceptions, (iii) making investments
subject to certain exceptions, (iii) changing the nature of the business, (iv) issuing equity securities, (v) disposing of
assets outside of the ordinary course of business, (vi) consolidating or merging, (viii) transacting with affiliates subject
to certain exceptions, (ix) opening bank accounts other than those secured under the financing documentation,
(x) creating subsidiaries, partnerships, or joint ventures, (xi) making certain tax elections, (xii) making certain ERISA
elections, (xiii) amending Project Agreements (including EPC change orders) subject to certain exceptions, (xiv)
entering into additional Project Agreements, (xv) suspending or abandoning the project, (xvi) entering into hedging
arrangements not approved by the lenders, (xvii) budgeting changes subject to certain tolerance bands, and (xviii)
making equity distributions outside of the waterfall framework and unless certain criteria are met, including achieving
certain cash available to debt service ratios on a historical and prospective basis (usually between 1.25:1 and 1.5:1).
While this list is not comprehensive, it should again be stressed that lenders will generally tend to be sensitive to the
financial interests of the project and will to some degree tailor a covenant package to the projects expected
construction and operation characteristics. It should also be noted that project finance lenders will often entertain
requests for waivers of obligations set forth in the financing documentation after the closing of the loans, as they are
incentivized to keep the loans performing and out of default.
F.
Potential Defaults
Event of Default is the legal term for the circumstance that allows project finance lenders to exercise their remedies
under the financing documentation, including acceleration of the outstanding debt and foreclosure. Events of Default
may include: (i) nonpayment of fees, interest, or principal due under the financing documentation (usually with a very
short grace period with respect to fees and interest only), (ii) breach of representation or warranty made in the
financing documentation (usually with a grace period if capable of being cured), (iii) non-performance of certain
covenants or obligations under the financing documentation (usually with a grace period if capable of being cured),
(iv) cross-defaults to other debt instruments, (v) non-appealable legal judgments rendered against the Project
Company, (vi) certain events related to ERISA, (vii) bankruptcy or insolvency, (viii) default under or termination of
Project Agreements, (ix) significant delays in construction schedule, (x) failure to obtain or maintain a necessary permit
or government approval, (xi) unenforceability of financing documentation, (xii) certain material environmental matters,
(xiii) loss of or damage to collateral, (xiv) abandonment of the project, and (xv) a change of control. Many Events of
Default have cure periods, which allow the Sponsor or Project Company to take action over the course of a certain
period (usually 30 days but may be less or more) to remedy the non-compliance if the Event of Default is capable of
being cured; for example, a default under another debt instrument may be cured by paying the amount due but a
final, non-appealable legal judgment against the Project Company would be incurable. In addition, during the course
of negotiating the Credit Agreement it will be important for the Project Companys representatives to qualify as many of
the Event of Default provisions with materiality and Material Adverse Effect standards as possible, providing the
Project Company more leeway to avoid an Event of Default and the potential loss of the project.
G.
Conditions to Closing
Project financing lenders will require that a lengthy list of conditions be satisfied in order to close the financing and fund
the loan. While many of the precedent conditions and required documents are shared with other forms of financing, it is
worth mentioning certain of the conditions that constitute particularly long lead time items that must be commenced
months prior to the close of the financing. Specifically, project finance lenders will generally require the delivery of the
following as conditions to closing the loan: (i) a report of an independent engineer that confirms the technology employed
by the project is commercially viable, the reasonableness of budgetary assumptions, the absence of serious
environmental issues, compliance with all necessary permits or approvals, and that financial projections are realistic; (ii) a
power or biofuels market report (if the project will have significant uncontracted off-take) setting forth expected market
conditions over the course of the loan; (iii) an environmental site assessment (at least a Phase I report concluding that
no further environmental investigation is necessary); (iv) an insurance report from the lenders insurance consultant;
(v) land surveys and site descriptions; (vi) a commodity management plan (in the case of biofuels facilities and other
projects where appropriate); (vii) evidence that the required equity component of the project has been contributed or will
otherwise be available when required; and (viii) copies of all third-party and government approvals and permits.
As a final note, depending on the projects funding requirements and the size of the equity contribution and project
finance commitments, it may be possible to include subordinated debt in the financing package. In the case of certain
renewable energy projects (e.g., biofuels production facilities), tax-exempt state bond financing may be available to
12
www.wsgr.com
close any gap between the raised and required equity in a project finance scenario. As a general rule, subordinated
debt will be more expensive than senior debt due to the subordinated lenders higher risk of non-payment. In almost
all circumstances, the subordinated debt will need to be in place prior to finalization of the senior project debt to avoid
the substantial costs that would be incurred to re-document the senior loan. If subordinated debt is employed, an
intercreditor agreement will be negotiated between the agent for the senior lenders and the trustee or agent for the
subordinated debtholders, pursuant to which the senior lenders will obtain standard terms of subordination to ensure
their senior lien and payment positions vis-a-vis the subordinated lenders and any unsecured creditors in the case of
any Event of Default by the Project Company or its bankruptcy or insolvency.
V. Tax Implications
A.
Key Federal Income Tax Incentives for Renewable Energy Projects
The U.S. federal government provides several income tax subsidies to encourage the development of renewable
energy projects. These tax benefits are currently necessary to make renewable energy projects economically
competitive with projects that produce energy from conventional sources, and can finance as much as approximately
60 percent of the capital cost of a project. The following is a brief discussion of some of the key federal income tax
incentives available to developers of, and investors in, renewable energy projects.
1. Production Tax Credits
A production tax credit (PTC) is available for the production and sale of electricity from certain renewable sources.
Renewable sources of energy that qualify for the PTC include wind, biomass, geothermal, municipal solid waste (either
landfill gas or trash), hydropower (in the case of newly installed turbines), and marine and hydrokinetic energy. To
qualify for the PTC, electricity from these sources must be produced at a facility that is placed in service before
(i) January 1, 2013 for a wind facility and (ii) January 1, 2014 for other qualifying facilities. The facility must be located
in the United States.
The PTC is available for 10 years following the date the qualified facility is placed in service. The amount of the credit
for each year is generally determined by multiplying the credit rate by the number of kilowatt hours of electricity
produced by the taxpayer from a qualified facility and sold to an unrelated party. The credit rate is adjusted for inflation
each year and varies based on the type of renewable resource (for 2009, the credit rate for most qualifying facilities
was 2.1 cents per kilowatt hour). The amount of the PTC is reduced by as much as 50 percent to the extent the
project benefits from nontaxable grants, tax-exempt bonds, other subsidized energy financing, or other federal credits.
2. Investment Tax Credit
Most renewable energy projects can qualify for the investment tax credit (ITC), which is based on the cost of the
qualifying property (unlike the PTC, which is based on the amount of electricity generated and sold). The ITC is equal
to the product of the energy percentage and the taxpayers tax basis in its energy property that is placed in service
during the taxable year. Energy property includes, among other things, equipment that uses solar energy to generate
electricity, to heat or cool (or provide hot water for use in) a structure, or to provide solar process heat. Certain fuel cell
power plants that are placed in service before January 1, 2017, also qualify for the ITC. The energy percentage is
30 percent for solar equipment and fuel cell equipment that is placed in service before January 1, 2017. For solar
equipment placed in service after 2016, the energy percentage is 10%, and the ITC is not available for fuel cell
property placed in service after 2016. To be placed in service, the property must be ready for use, which generally
requires that all tests have been completed, all licenses and permits have been obtained, and the project is
synchronized with the transmission system and is operational. Unlike the PTC, the ITC is not reduced by subsidized
energy financing received after 2008.
In 2009, Congress enacted legislation that allows a taxpayer to elect to claim either the ITC or the PTC for facilities
that qualify for the PTC. The ITC for these facilities is 30 percent of tax basis and the election is available for a facility
that is placed in service before the relevant PTC cutoff date: (i) January 1, 2013 for a wind facility, and (ii) January 1,
2014 for other qualifying facilities. In deciding which credit to claim for a renewable energy project that can qualify for
either credit, a Sponsor will consider a number of factors, including the estimated cost of the facility, the expected
power production from the facility over the 10-year PTC period and anticipated power prices (including prices under a
PPA that is in place).
13
www.wsgr.com
To qualify for the ITC, the energy property must satisfy several requirements. The property must be constructed or
acquired by the taxpayer and the original use of the property must commence with the taxpayer. In addition, the
property must be used within the United States and depreciation or amortization must be allowable with respect to the
property. The property must meet any performance and quality standards prescribed by the Internal Revenue Service
(IRS) after consultation with the Department of Energy (none have been proposed to date). The property must also
not be used (including under a lease) by a tax-exempt or governmental entity. Finally, the project must not include
property which is part of a facility the production from which is used to claim PTCs. The taxpayers depreciable basis
in the property is reduced by 50 percent of the amount of the ITC.
The ITC vests 20 percent per year over five years. If the property is disposed of or otherwise becomes ineligible for
the ITC prior to fully vesting, the unvested portion is recaptured.
3. Treasury Grants
A renewable energy project that qualifies for the ITC and that is placed in service in 2009 or 2010 can qualify for a
Department of the Treasury grant in lieu of the ITC (and the PTC). If construction on the project began in 2009 or
2010, the project can also qualify for a grant so long as it is placed in service by the relevant date on which the ITC
terminates. The amount of the grant is calculated in the same manner as the ITC for the qualifying property, but since
the grant is a cash payment from the Department of the Treasury (rather than a credit against federal income tax) a
taxpayer need not have a federal income tax liability to benefit from a grant. Property that is used (including under a
lease) by a tax-exempt or governmental entity may nevertheless qualify for a grant. The taxpayers depreciable basis
in the property is reduced by 50 percent of the amount of the grant.
Treasury grants are also subject to recapture, using the same schedule that is used for the ITC, but a sale or other
disposition of property with respect to which a grant was received is not a recapture event provided that (i) the sale is
to a person eligible to receive a grant and (ii) the buyer of the property agrees to be jointly liable with the seller of the
property for any recapture. An application must be filed for a grant not later than September 30, 2011.
4. Depreciation
Certain equipment used in renewable energy projects may qualify for accelerated depreciation. After the basis of the
property is reduced by 50 percent of the ITC (or Treasury grant), the remaining basis is generally depreciated over five
years following the date the project is placed in service. Bonus depreciation, which allowed a special 50 percent
depreciation deduction, expired for most property (including renewable energy projects) placed in service after 2009.
5. Qualifying Advanced Energy Project Credit
In 2009, Congress enacted a new credit equal to 30 percent of a taxpayer's qualified investment with respect to any
qualifying advanced energy project of the taxpayer. This credit is available only for manufacturing facilities, not for
renewables or energy efficiency installation projects. For the purposes of the credit, a "manufacturing facility" is a
facility that makes, or processes raw materials into, finished products (or accomplishes any intermediate state in that
process). A "qualifying advanced energy project" is a project that re-equips, expands, or establishes a manufacturing
facility to produce property that is designed to (i) be used to produce energy from solar, wind, geothermal, or other
renewable resources; (ii) manufacture fuel cells, microturbines, or an energy storage system for use with electric or
hybrid motor vehicles; (iii) manufacture electric grids to support transmission of intermittent sources of renewable
energy, including storage of such energy; (iv) manufacture carbon capture or sequestration equipment; (v) refine or
blend renewable fuels or produce energy-conservation technologies (including energy-conserving lighting technologies
and smart grid technologies); (vi) manufacture qualified plug-in electric drive motor vehicles, qualified plug-in electric
vehicles, or components designed specifically for such vehicles; or (vii) reduce greenhouse gas emissions (as
determined by the Treasury Department). The IRS must certify that the project is eligible for the credit and the project
cannot produce any property that is used in the refining or blending of any transportation fuel (other than renewable
fuels). A taxpayer's "qualified investment" is the basis of eligible property that the taxpayer places in service during the
taxable year and is part of a qualifying advanced energy project. Property is placed in service in the taxable year in
which it is placed in a condition or state of readiness and availability for its intended purpose. "Eligible property" means
any property of the taxpayer that is (i) necessary for the production of property designed for the uses listed above;
(ii) tangible personal property or other tangible property (not including a building or its structural components), but only
if such property is used as an integral part of the qualified advanced energy project; and (iii) with respect to which
depreciation (or amortization) is allowable. The basis of the property is reduced by the full amount of the credit, and
recapture rules apply to the credit.
14
www.wsgr.com
To claim this credit, among other things a taxpayer must file applications for a project with the Department of Energy
and the IRS and be awarded an amount of credit. A total of $2.3 billion of credits is authorized for awards. In January
2010, the federal government announced that the entire $2.3 billion available had been awarded to projects. However,
because some approved projects might never be completed (or qualify for a lesser amount of credit) and because of
the possibility that an additional several billion dollars of credit might be authorized for future awards, credits might
become available in the future. The techniques described below to monetize tax benefits may also be available for this
credit. Because this credit is awarded to a specific applicant, any monetization technique that involves the transfer of
the property to another person (or a change in tax status of the person awarded a credit) will require that the IRS
approve the transfer of the credit to the successor.
6. Cellulosic Biofuels
An income tax credit is available for a producer of cellulosic biofuels equal to $1.01 for each gallon of qualified fuel
produced. This credit is reduced by the amount of other ethanol credits (which will expire on December 31, 2010), so
that the sum of this credit and those other credits is $1.01 per gallon of qualified fuel produced. In addition, the owner
of a cellulosic biofuel plant placed in service in the United States before January 1, 2013 (or the construction of which
begins after December 20, 2006 and before January 1, 2013) may claim a depreciation deduction equal to 50 percent
of the cost of the plant for the year in which it is placed in service; the remaining 50 percent of the cost of the plant may
qualify for accelerated depreciation.
B.
State Tax Incentives
Although not the focus of this discussion, many states and local municipalities also offer tax incentives, in various
forms (e.g., income tax credits, sales and use tax exemptions, property tax exemptions, and tax abatements), to
promote the development of projects utilizing a wide variety of renewable energy sources. Of course, in budgeting for
and deciding where to site a renewable energy project, Sponsors should also consider the impact of state and local
taxes (such as sales and use tax and property tax) on the proposed project.
An example of an available state income tax credit is the renewable energy property investment tax credit established
by North Carolina to encourage the development and expansion of renewable energy property in that state. This
income tax credit is equal to 35 percent of the cost of renewable energy property constructed, purchased or leased by
a taxpayer and placed in service in North Carolina, with limits of (i) $2.5 million of credit per installation for
nonresidential property and (ii) $1,400 to $10,500 of credit per installation (depending on the type of renewable energy
used) for residential property. If the property serves a single-family dwelling, the credit is taken for the taxable year in
which the property is placed in service; if the property is a multi-family dwelling or is non-residential, the credit is taken
in five equal installments beginning with the year the property is placed in service. Renewable energy property
includes biomass, solar, geothermal, wind, and hydroelectric. An example of a sales and use tax exemption is the
state of New York exemption for the sale and installation of residential solar-energy systems. The exemption applies
to the sale or use of a solar-energy system that utilizes solar radiation to produce energy designed to provide heating,
cooling, hot water, and/or electricity.
C.
Tax Incentive Monetization Structures
For many reasons, a developer of a renewable energy project may not be able to benefit from the various tax
subsidies available to the project. Various strategies have developed that allow a developer to receive value for, or
monetize, the tax incentives the developer would not otherwise be able to utilize. These strategies generally involve
an institutional investor that can benefit from the tax incentives acquiring an equity interest in the project. Two of the
main strategies, the partnership flip and the sale-leaseback, are discussed below. These monetization structures
may also be used for the Treasury grant in lieu of the ITC, described above.
1. Partnership Flip
In a typical partnership flip transaction, an institutional investor will form a partnership with the developer, which will
own the Solar Project or other renewable energy project. The investor will receive an allocation of tax benefits and
cash distributions from the partnership until the investor achieves an agreed-upon after-tax return. Subject to some
limitations, the investor may make its investment in the partnership over time, which effectively allows the investor to
fund its investment in the partnership with reductions in future federal income tax liability.
In the initial stage of the project, the investor generally will receive a disproportionate allocation of the partnerships
income or loss and any tax credits (e.g., PTC, ITC) available to the partnership. When the investors target return is
achieved (the flip-point), the investors allocation of partnership items is reduced to a small portion.
15
www.wsgr.com
The partnership generally will distribute its available cash flow 100 percent to the developer until the developer
recoups its cash investment in the project, and cash would thereafter be distributed 100 percent to the investor until
the flip-point is reached. Following the flip-point, cash distributions would be made in accordance with partnership
allocations (e.g., 95% to the developer and 5% to the investor). The developer will typically have an option,
exercisable on or after the flip-point, to purchase the investors interest in the partnership at its then fair market value.
In late 2007, the IRS published safe harbor guidelines for wind partnership transactions under which it would treat the
investor as a partner in the partnership (rather than a purchaser of tax credits) and respect the disproportionate
allocation of PTCs to the investor. These guidelines were revised in late 2009. The revised guidelines provide that the
safe harbor guidelines are not intended to provide substantive rules and are not to be used as IRS audit guidelines.
Although the safe harbor applies only to wind transactions and the allocation of PTCs, the renewable energy industry
is generally following the safe-harbor guidelines in structuring partnership flip transactions for non-wind projects and for
tax incentives other than the PTC.
2. Sale-Leaseback
The sale-leaseback is another structure utilized to monetize various tax incentives. Although the sale-leaseback
generally may not be used to monetize PTCs, the technique has been used extensively to monetize ITCs in solar
projects. In a typical sale-leaseback transaction involving a solar project, the developer will install, operate, and
maintain the project and a customer will agree to purchase the power generated from the project under a long-term
PPA. The developer will incur all expenses related to the installation, operation, and maintenance of the solar
equipment.
To monetize the ITC and other tax benefits, the developer will sell the facility to an investor within three months after its
in-service date. The investor will lease the project back to the developer for a lease term approximating the term of the
PPA and the developer will typically use the PPA as collateral for its lease payment obligations. The developers
revenue from the PPA is utilized to make rental payments under the lease.
The investor is considered the owner of the project for tax purposes, and it therefore claims the ITC and other tax
benefits. The investor shares its tax savings with the developer in the form of reduced rents. The developer will
typically have an option, exercisable at the end of the lease term, to purchase the project from the investor at its then
fair market value.
For the sale-lease back structure to work, the lease must be structured as a true lease for tax purposes. There is an
extensive body of law addressing the characterization of transactions cast in the form of a lease. In general, under
IRS guidelines, a lease would be respected as a true lease if the lessee does not have an option to purchase the
property for an amount less than its fair market value, the lessor retains the risk that the property will decline in value
(e.g., the lessor does not have the right to require the lessee to purchase the asset at a fixed price), and at the end of
the lease, the leased asset is expected to have a significant residual value (e.g., 20% of its original cost) and a
significant remaining useful life (e.g., 20% of its originally estimated useful life). Case law has held that arrangements
that do not fully comply with the IRSs ruling guidelines nonetheless qualify as leases. Equity investors and lenders
may require that a lease of a renewable energy project qualify as a guideline lease, however.
3. Pass-through Lease
Pass-through leases, used extensively to monetize the rehabilitation tax credit, have been used recently to monetize
solar energy credits (and more recently Treasury grants). Typically, an entity (Owner) would acquire a solar energy
project from a developer at fair market value. The project would include not only the tangible solar assets but also the
contract rights to sell the energy to the off-taker or homeowner (or lease the solar equipment to the offtaker/homeowner). The Owner would be structured as a limited liability company owned 50.01% by the developer (or
an affiliate of the developer) and 49.99% by another limited liability company (Tenant). Tenant typically would be
owned 99.9% by the investor and 0.01% by the developer (or an affiliate). Owner would lease the project to the
Tenant under an arrangement that would qualify as a true lease for income tax purposes. Owner then elects to have
any available federal credits (or Treasury grants) pass though to its lessee, Tenant. (The legislation governing
Treasury grants also permits this election for Treasury grants.) Under the election, the lessees cost basis in the
property (for claiming the credit) is the appraised value of the property, which is not necessarily limited to the lessors
cost basis in the property. The basis of the property is not adjusted for the amount of the credit/Treasury grant, but the
lessee is required to include in income 50% of the amount of the credit/Treasury grant ratably over a 5-year period.
The effect of the election is to bifurcate the credit/Treasury grant from the depreciation, since the investor owns more
than 99% of the lessee, but indirectly owns approximately 50% of the Owner, the entity that will claim depreciation.
16
www.wsgr.com
Thus, this structure is usually proposed by investors who value the credit/grant but place less importance on
depreciation. Developers would typically prefer a partnership flip or sale-leaseback structure, because, among other
things, those structures monetize nearly all of the accelerated depreciation (which developers often cannot use).
In addition to being allocated nearly all tax credits/grants, investors in pass-though structures typically received a fixed
annual equity distribution from the Tenant entity in the range of 2-5% of their invested capital. The economics and
cash flows of these deals are harder to generalize than the cash flows in a sale leaseback (where the investor gets
everything) or flip (where the investor gets nearly everything prior to the flip) and extensive modeling is usually
required to ensure the economics of the deal meet expectations and are consistent with the legal structure.
VI. Conclusion
Companies that are in the business of developing renewable energy projects confront a host of complex and inter-related
commercial and legal issues that must be successfully navigated to ensure a projects success and realize potential
investor returns. Regardless of whether project finance is employed, it is important for Sponsors to assemble a team of
professional advisors that can not only assist in executing a debt or equity transaction, but also analyze the various
options that may exist in the course of developing projects. The currently tight credit environment, which is characterized
by a lack of liquidity in the marketplace and a general risk aversion on the part of lenders, serves only to heighten
competition for available debt. However, in such an environment, the right combination of business model, project scale,
contractual structure, and equity support will still be attractive to project lenders for long-term debt commitments.
Determining whether to pursue project financing in the course of developing renewable projects is one of the most
fundamental decisions that developers must make. An affirmative decision will dictate the legal and contractual
structure of the projects, place certain operational limitations on how the projects operate, and limit the developers
discretion regarding the use of much of the cash flow from the project. On the other hand, a successful project
financing can maximize equity returns through increased project leverage, remove significant liabilities from the
Sponsors balance sheet, capitalize on tax financing opportunities, and protect key Sponsor assets. In order to take
full advantage of project financing opportunities, it is vital that companies invest the time and resources during the
initial development stages to obtain the best possible terms and conditions in commercial agreements which serve as
the foundation to project financing on successful terms.
August 2010
17
www.wsgr.com

Electricity Markets and Renewable Generation

Diunggah oleh

Informasi Dokumen

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Electricity Markets and Renewable Generation

Diunggah oleh

Hak Cipta:

Format Tersedia

2014

Energy Risk Professional

READINGS THAT ARE FREELY AVAILABLE ON THE GARP WEBSITE

Readings for Electricity Markets and Renewable Generation

Compare and contrast the DR programs for various RTOs.

Describe the various methods used to encourage end-user participation in DR programs.

in the creation of DR programs.

ment (DSM) operate.

Global Electricity Market Applications and Current Trends

Johannes P. Pfeifenberger and Kathleen Spees. Evaluation of Market Fundamentals and

Summarize operational decisions associated with the expiration of Power Purchase

fuel-on-fuel competition may have on the electricity market in Alberta.

2014 Global Association of Risk Professionals. All rights reserved.

Energy Risk Professional

Australian Energy Market Operator. An Introduction to Australias National

Understand the roles of AEMO and NEM in the Australian market.

Understand the difference between regulated and unregulated interconnectors and

Describe how power forecasts affect market operation.

Identify the stakeholders in a liberalized electricity market.

Calculate the settlement (payment) for a quantity of dispatched/consumed power

Understand how NORD Pool up-regulation and down-regulation pricing ensures

Describe the deregulation of Singapores electricity market.

Describe vesting contracts, their use and their settlement process.

2014 Global Association of Risk Professionals. All rights reserved.

Energy Risk Professional

Discuss how investments in renewable power generation will be incentivized under

impact of negative prices.

Global Renewable Generation, Carbon Emissions and Project Finance

Nuclear Energy Agency. Nuclear Energy Today, Second Edition (2012).

Compare and contrast different nuclear reactor technologies, including pressurized

Compare low-level, intermediate-level, and high-level radioactive waste and identify

Intergovernmental Panel on Climate Change (IPCC). IPCC Special Report on Renewable

Compare and contrast various solar generation technologies, including photovoltaic

Understand how PV systems are integrated into the power grid.

Understand technologies used to transport and store solar generated electricity.

Explain the smoothing effect with respect to multiple PV solar systems.

Chapter 5 ........................Hydropower (Sections 5.35.5 only)

Compare and contrast run-of-river, storage hydropower, and pumped storage

Describe the characteristics of hydropower generation and understand how these

2014 Global Association of Risk Professionals. All rights reserved.

Energy Risk Professional

Chapter 7.........................Wind Energy (Sections 7.37.5 and 7.8 only)

Understand the dynamics of the power curve for a wind turbine.

the ETS and identify the industries covered.

2014 Global Association of Risk Professionals. All rights reserved.

Addressing Energy Demand

Reprint version of the report prepared for AZURE

Definition of Demand Response ..................................................................................................................1

Benefits Brought by Demand Response .......................................................................................................1

DEMAND-SIDE MANAGEMENT AND THE ROLE OF DR ..................................................................4

Overview of Demand Side Management .....................................................................................................4

Role of DR in Demand-Side Management ....................................................................................................5

Coordination of Demand Response and Energy Efficiency ..........................................................................6

REGULATORY AND POLICY FRAMEWORKS THAT PROMOTE DEMAND RESPONSE .........................7

Wholesale Market Access ......................................................................................................................................... 7

Bilateral Programs with Vertically Integrated Utilities and Network Operators..................................................... 12

Avoid Energy Costs vs. Resource Payments ............................................................................................................ 21

Emerging Trends: Dynamic Pricing ......................................................................................................................... 21

Summary/Comparative Analysis of Policy and Regulatory Frameworks ...................................................22

ENABLING TECHNOLOGY SOLUTIONS FOR DEMAND RESPONSE ................................................. 25

Metering and Control Solutions .................................................................................................................25

Auto-DR and OpenADR (with the AMI linkage) ..........................................................................................27

Smart Meter and Advanced Metering Infrastructure (AMI) and OpenADR...............................................29

BEST PRACTICES AND RESULTS OF DR IMPLEMENTATION .......................................................... 31