Você está na página 1de 326

2014 Energy Risk Professional

ERP Exam Course Pack


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
in the creation of DR programs.
• Compare and contrast the DR programs for various RTOs.
• Understand how bilateral DR programs like cost recovery and demand-side manage-
ment (DSM) operate.
• Describe the various methods used to encourage end-user participation in DR programs.

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
fuel-on-fuel competition may have on the electricity market in Alberta.
• 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
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 Energy Risk Professional
ERP Exam Course Pack

3. Australian Energy Market Operator. An Introduction to Australia’s 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.
• Describe how power forecasts affect market operation.
• Explain the use of hedge contracts and how these transactions are settled.

4. 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 deregu-
lated 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 Singapore’s creation of “artificial” LNG prices and the rationale for
this policy.
• Describe the deregulation of Singapore’s electricity market.
• Describe the financial flows and contract settlement in both the wholesale and
retail markets.
• Understand the types of offers made by a generator, how these offers are accepted,
market prices established and offers cleared by the market.
• Describe vesting contracts, their use and their settlement process.

© 2014 Global Association of Risk Professionals. All rights reserved.

2014 Energy Risk Professional
ERP Exam Course Pack

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
impact of negative prices.
• Discuss how investments in renewable power generation will be incentivized under
the new market design, including capacity-based mechanisms.
• Define the balancing market and understand the relationship between wind power
production and balancing costs, and explain how balancing costs are reduced.
• 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.

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 3.3–3.5 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 5.3–5.5 only)

• Compare and contrast run-of-river, storage hydropower, and pumped storage
• 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 Global Association of Risk Professionals. All rights reserved.

2014 Energy Risk Professional
ERP Exam Course Pack

Chapter 7.........................Wind Energy (Sections 7.3–7.5 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 hydrocar-
bon-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
the ETS and identify the industries covered.
• Understand how an auction system can address the issue of windfall profits that often
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

© 2014 Global Association of Risk Professionals. All rights reserved.


Addressing Energy Demand

through Demand Response:
International Experiences and

Bo Shen, Girish Ghatikar, Chun Chun Ni, and

Junqiao Dudley
Environmental Energy Technologies Division
Lawrence Berkeley National Laboratory

Phil Martin and Greg Wikler


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.

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. INTRODUCTION ..........................................................................................................................1
1.1. Definition of Demand Response ..................................................................................................................1

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

2. 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


3.1. Demand Response Enabling Policies ............................................................................................................7
3.1.1 Wholesale Market Access ......................................................................................................................................... 7

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

3.2. Encouraging End-User Participation: The Role of Incentives .....................................................................16

3.2.1 Lack of Sufficient Incentives from Standard and TOU Pricing: Experience with Interruptible Tariffs ...................... 19

3.2.2 Cost and Risks–How 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

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

4. 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

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

5.1. 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


List of Acronyms

AESO Alberta Energy System Operator

AMI advanced metering infrastructure
AMP aggregator managed portfolio
Auto-DR automated demand response
BRA base residual auction
C&I commercial and industrial
CAP Climate Action Plan
CEC California Energy Commission
CPP critical peak pricing
CSPs curtailment service providers
DECC U.K. Department of Energy and Climate Change
DLC direct load control
DOE U.S. Department of Energy
DPCR5 distribution price control review
DR demand response
DRAS demand response automation server
DRRC Demand Response Research Center
DSM demand side management
EE energy efficiency
EEPS Energy Efficiency Portfolio Standard
EILS emergency interruptible load service
EISA 2007 Energy Independence and Security Act of 2007
EMCS energy management control systems
EPACT The Energy Policy Act of 2005
ERCOT Electric Reliability Council of Texas
FCM forward capacity market
FERC U.S. Federal Energy Regulatory Commission
FRCC Florida Reliability Coordinating Council
HVAC heating ventilation and air-conditioning
I/C Interruptible/curtailable
IA incremental auctions
IESO independent electricity system operator
IOUs investor-owned utilities
IT information technology
LMP locational marginal price
M&V measurement and verification
MRO Midwest Reliability Organization
NOCs network operation centers
NPCC Northeast Power Coordinating Council, Inc.
OPA Ontario Power Authority
PG&E California Pacific Gas and Electric
PPA power purchase agreement
PTR peak time rebates
RFC ReliabilityFirst Corporation
RPM reliability pricing model
RTDR real time demand response
RTEG real time emergency generation
RTO regional transmission organization
RTP real-time pricing
SCE Southern California Edison
SERC Southeastern Electric Reliability Council
SPP Southwest Power Pool
SRM sycnhronized reserve market
STOR short term operating reserves
TOU time-of-use
TVA Tennessee Valley Authority
WECC Western Electricity Coordinating Council
WEM whosesale electricity market

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 US-
based 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 MW…For 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.

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

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
Brattle Group, “The Power of Five Percent,” May 16, 2007.

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.



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

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)

Annual Growth Rate (%)

25,000 10.0%

10,000 -10.0%
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. (
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).

(Unit: %)

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.
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
 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.

Cappers, Peter, C. Goldman, and D. Kathan (2009), Demand Response in U.S. Electricity Markets: Empirical Evidence,
Lawrence Berkeley National Laboratory, LBNL-2124E.


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

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 state’s 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.

The current size of the ISO-New England system is approximately 26 GW.
The current size of the PJM system is approximately 165 GW.
Federal Energy Regulatory Commission (2011). “Assessment of Demand Response & Advanced Metering Staff Report.”
November 2011. Table 2.

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 country’s 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

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 cost-
effective 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

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.
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. PJM’s 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 PJM’s 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 PJM’s
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)

PJM’s 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 PJM’s BRA, the ISO-NE FCM also allows both
dispatchable demand response and energy efficiency measures to participate in the market.

ISO-NE’s 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

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.

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 nation’s 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

For example, the Fast Reserves (FR) program has a 50 MW minimum requirement for participation.

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

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.

The Reserve Capacity Cycle (RCC) is the process that is used in Australia to procure DR resources as part of the Reserve
Capacity Mechanism.
Note that to date this situation has never been experienced so no auctions have been called.

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. DR3’s 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 next-
generation 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,

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

Perhaps the most basic and essential enabling policy is a cost-recovery mechanisms. Under a cost-
recovery 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 California’s
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 state’s highest-
priority 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,
California’s 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 state’s 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

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 state’s 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 state’s utilities to achieve a 15% reduction in forecast
electricity usage by the year 2015. The state’s utilities are implementing aggressive EE and DR programs
in order to meet that goal, which specifies that each of the state’s 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
state’s 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 firm’s 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

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

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&E’s 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 Boston-
based 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 10-
year 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

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.

Program Size Up to 560 MW
Advanced Notification 30 minutes
Dispatch Trigger TVA’s discretion
Availability Window April – October: 12:00-20:00, Mon-Fri
November – March: 5:00-13:00, Mon-Fri
Maximum Cumulative Dispatches 40 hours per annum
Term Length 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

U.S. Department of Energy (February 2006).

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).

Table 1: Demand Response Options and Related Pricing Mechanisms
Price-Based (Voluntary)
(Contractually Mandatory)
 Time-of-use (TOU): a rate with different unit  Direct load control: a program by which the
price for usage during different blocks of time, program operator remotely shuts down or cycles
usually defined for a 24 hour day. TOU rates a customer’s electrical equipment (e.g., air
reflect the average cost of generating and conditioner, water heater) on short notice. Direct
delivering power during those time periods. load control programs are primary offered to
 Real-time pricing (RTP): a rate in which the residential or small commercial customers.
price for electricity typically fluctuates hourly  Interruptible/curtailable (I/C) service: curtailment
reflecting changes in the wholesale price of options integrated into retail tariffs that provide a
electricity. Customers are typically notified of rate discount or bill credit for agreeing to reduce
RTP prices on a day-ahead or hour-ahead basis. load during system contingencies. Penalties
 Critical Peak Pricing (CPP): CPP rates are a maybe assessed for failure to curtail. Interruptible
hybrid of the TOU and RTP design. The basic programs have traditionally been offered only to
rate structure is TOU. However, provision is the largest industrial (or commercial) customers.
make for replacing the normal peak price with  Demand Bidding/Buyback Program: customers
a much higher CPP event price under specified offer bids to curtail based on wholesale electricity
trigger conditions (e.g., when system reliability market prices or an equivalent. Mainly offered to
is compromised or supply prices are very high). large customers (e.g., one megawatt [MW] and
 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

U.S. Federal Energy Regulatory Commission (December 2008), Assessment of Demand Response and Advanced Metering,
Washington D.C.

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. Non-
residential 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 utility’s 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 utility’s 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 pre-
specified amount, during system reliability problems in return for an incentive payment or a similar rate

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 Risks–How Load Aggregators have Removed Traditional Barriers to DR
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
don’t 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

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 don’t
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

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 five-
times 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-

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 UK’s “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 vertically-
integrated 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 peak-
shaving 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

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.

4. Enabling Technology Solutions for Demand Response

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 facility’s 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) Manual DR: This involves manually turning off or changing comfort set points, lights, or
processes or each equipment, switch, or controller.
2) 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.
3) 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

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

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 cost-
prohibitive to deploy at individual customer sites.

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.

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
The OpenADR Primer, White paper by the OpenADR Alliance (http://www.openadr.org/).

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 pre-
programmed 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 set-
point that a load should go to).
 Levels: These are simple representations of the price and reliability signals such as NORMAL,

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 (so-
called 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


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

Figure courtesy: Ron Hoffman, California Energy Commission.
McParland, Charles. Home Network Technologies and Automating Demand Response. LBNL, 2008. LBNL-3093E.

Evaluation of Market Fundamentals and
Challenges to Long-Term System
Adequacy in Alberta’s 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.


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.
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.  AESO’s 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 

The Alberta Electric System Operator (“AESO”) asked The Brattle Group to review long-term
challenges to resource adequacy in Alberta’s electricity market and assess the following four
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
Alberta’s 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

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 recently-
experienced 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 Alberta’s 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 Alberta’s, 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

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 demand-
response, 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-of-
market 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 Alberta’s 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 long-
term for the simple reason that Alberta’s 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.


The Alberta Electric System Operator (“AESO”) asked The Brattle Group to review long-term
challenges to resource adequacy in Alberta’s 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
Alberta’s 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-of-
market” or “backstop” mechanisms. The scope of our analysis does not include challenges
related to transmission planning, system operations, and short-term market design initiatives.


Alberta’s 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

Designs without Explicit Resource Designs With Explicit Resource

Type of Adequacy Requirement Adequacy Requirement
Capacity With Capacity
Market Payments or Short-Term Forward
AESO, Australia’s Argentina, Chile, SPP, former
NEM, ERCOT, Colombia, Peru, power pools
Great Britain, Spain, South (NYPP, PJM,
NordPool Korea, Ontario NEPOOL)
Voluntary Midwest ISO
NYISO, former
Mandatory PJM, Australia’s
NE, 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

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”), Australia’s National Electricity Market (“NEM”), Australia’s 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
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
Australia’s 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

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 administratively-
determined 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 energy-
only 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. Alberta’s 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 VOLL-
based 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

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
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.

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.


Prior to deregulating, Alberta’s 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

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 long-
term 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.

Alberta’s 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, Alberta’s 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 heavy-
handed 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.

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).

However, Alberta’s energy-only market design and market fundamentals also differ from other
energy-only markets in a number of respects. Alberta’s 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 Alberta’s centralized market is an ex-post real
time market, with no day-ahead market, hour-ahead market, or centralized generation unit

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.


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, Alberta’s 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.


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

For example, Australia’s 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
For a comprehensive review of these issues, see our review of market design for DR in AESO,
Pfeifenberger and Hajos (2011).

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

Figure 1
Monthly Electric and Gas Prices in Alberta
$14 $200
Historic Future
$12 Gas Electric
(Left Axis) (Right Axis)

$10 $140

Electric Price ($/MWh)

Gas Price ($/GJ)


$4 $60


$0 $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 2009-
2010 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

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).

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.


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

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.

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 gas-
fired 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

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).

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
Historic Future

7 Hydro
Gas CT

Capacity (GW) &

2 Sundance 1 & 2
Retire Prior to PPA Expiration

1 Coal (Now Retired)

Gas ST
Gas ST (Now Retired)
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.

See Balancing Pool (2009), pp. 39-40; Alberta Government (2007b), Section 7.

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 AESO’s 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.


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. Alberta’s 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 Environment’s strategy includes a 139 MT CO2e

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

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
450 450
Carbon Capture and Sequestration
Greening Energy Production Business as Usual
400 Conservation and Energy Efficiency 400

350 350
MegaTonnes of CO 2 e &

50 MT 139 MT
300 300
200 MT

250 37 MT 250
2005 GHG Output 24 MT
200 Alberta Plan 200

150 150

100 100

50 Historic Electric Emissions 50

0 0
2005 2010 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 Alberta’s 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

generators and 1.0% for gas generators. Calculation based on emissions rates from Alberta Environment
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

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 Alberta’s total GHG output in 2008 as shown in
Figure 3.

Table 2
Alberta Registered GHG Emissions by Sector, 2008

Number of
Total Sector Percent of
Sector Reporting
Emissions, kT Total
Chemical Manufacturing 15 10,270 9.3%
Coal Mining 3 497 0.4%
Conventional Oil and Gas Extraction 29 6,845 6.2%
Mineral Manufacturing 6 2,403 2.2%
Oil Sands In Situ Extraction 13 10,927 9.9%
Oil Sands Mining and Upgrading 5 23,848 21.5%
Paper Manufacturing 4 478 0.4%
Petroleum Refineries 3 3,862 3.5%
Pipeline Transportation 4 2,797 2.5%
Utilities 26 48,903 44.1%
Waste Management 1 90 0.1%
Total 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 Alberta’s 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:

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
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

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 Alberta’s 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 Strategy’s 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 energy-
only market either in terms of prices or resource adequacy.
Overall, however, Alberta’s 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

the electric sector’s 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

Table 3
Large-Scale CCS Projects under Development in Alberta

Generation Government Annual CO2

Project Description Online Date
Capacity Awards, $M Sequestration

Swan Hills Synfuels [1] Underground coal gasification. 2015 300 MW $285 Alberta 1.3 MT
Above ground separation of syngas
from CO2 and impurities. Power
production from syngas in separate
combined cycle facility.
Alberta Carbon Trunk Line [2] 240 km CO2 pipeline from near Fort 2012 n/a $495 Alberta 14.6 MT Capacity
Saskatchewan south to Clive, to be $63 Federal 1.8 MT Initially
used for EOR. Initial CO2 will
come from Agrium Redwater
Complex and the North West
Upgrading facility once completed.

Shell Quest Project [3] Scotford bitumen upgrader facility. 2015 n/a $745 Alberta 1.2 MT
$120 Federal
Project Pioneer on Keephills 3 [4] Carbon capture retrofit to coal plant. 2011 Power Output 450 MW $436 Alberta 1 MT
2015 Carbon Capture $343 Federal

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 CCS-
enabled 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

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.

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

Alberta’s 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%,

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 unit’s 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.

 Contribute $15/tonne of CO2e to the Climate Change and Emissions Management
(“CCEM”) Fund, which invests these funds in projects to reduce emissions
 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.

Figure 4
Alberta GHG Specified Gas Emitters Regulation Compliance by Category
CO2 e Reductions or Payments, MT&


10 CCEM Fund Payments

6 Alberta Offsets

Performance Credits
Operational Improvements
2007 2008 2009
(half year)
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 Alberta’s generation fleet.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).

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

Fuel Cost, $/GJ [1] 6 6 1
Heat Rate, GJ/MWh [2] 7.7 13.2 10.4
GHG Rate, kg/GJ [3] 56.6 56.6 103.1
CO2e Cost, $/tonne [4] $15 $15 $15
Fraction of CO2e Output Charged [5] 12% 12% 12%

Costs, $/MWh
Fuel [6] $46.29 $79.43 $10.37
VOM [7] $2.22 $3.84 $4.93
CO2e [8] $0.79 $1.35 $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).
[4] Alberta Government (2007).
[5] Alberta Government (2007).
[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

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 sector’s 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 Alberta’s. 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

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.

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


Capacity (GW) &


Units Subject to Retirement Mandate



Sundance 1 & 2
Retire Prior to Mandate




















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).


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

See CASA (2003).
See CASA (2010), p. 1; Alberta Environment (2010c).

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 5-
year 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.

Alberta’s 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

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).

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

Gas Peakers
2.5 Gas

Capacity (GW) &




0.5 Sundance 1 & 2

Retire Prior to Air
Quality Mandate

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).


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

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.

certificates to allow utilities in California to meet the state’s ambitious renewable energy

Figure 7
Historic and Potential Future Wind Capacity Growth
Historic Projected
Wind Nameplate Capacity (GW) &


Proposed Additions

Permitted Additions

Under Construction
Historic Additions

Winter 2000/01 Capacity

















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. Alberta’s capacity factor during peak times is higher than in many other systems,
simply because Alberta’s 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

See Herndon (2011).
See AESO (2010b); AESO (2010h).

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 turbine’s 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 low-
wind 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

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.

Calculated from hourly wind data and wind installed capacity data from AESO (2010c).
See Midwest ISO (2009), pp. 3, 13.

Figure 8
Price Duration Curve at Different Levels of Wind Output



AESO Price ($/MWh)



Higher Prices
Wind < 100 MW

All Hours

$50 Lower Prices

Wind > 400 MW










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 AESO’s 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

See AESO (2010b); AESO (2010h).

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 Fleet-Wide
Active Standby Total Capability

Regulating 160 131 290 1,785

Spinning 243 106 350 3,148
Supplemental 243 37 280 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

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.


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

In addition to restoring intertie ATC to design rating, there is one intertie project that will further
expand Alberta’s 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-

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 AESO’s 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
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.

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 Alberta’s 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

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.

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.

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

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).

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
AESO’s 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

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 lower-
cost energy market prices over the long run, even after accounting for the capital cost recovery required
through Alberta’s energy market (compared to Alberta’s 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


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

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



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) &


Operating Reserves

Energy Margins


Cost of New Plant



Fixed O&M
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).

Figure 33
Historic Gas Cogen Operating Margins vs. Fixed Costs
Operating Reserves

Energy Margins
Operating Margins and Fixed Costs ($/kW-y)



Cost of New Plant



Fixed O&M

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).

Figure 34
Historic Hydro Operating Margins vs. Fixed Costs

Other AS
Operating Margins and Fixed Costs ($/kW-y) &

TMR Spinning
Cost of New Plant

Energy Margins




Fixed O&M
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).
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).

Figure 35
Historic Wind Operating Margins vs. Fixed Costs
Cost of New Plant
Operating Margins and Fixed Costs ($/kW-y) &

$200 Energy Margins



$50 Fixed O&M

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).
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).

Figure 36
Historic Gas CT Operating Margins vs. Fixed Costs (Including TMR Units)

Other AS
Operating Margins and Fixed Costs ($/kW-y) &


Cost of New Plant


Operating Reserves

Energy Margins
Fixed O&M
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) and PJM (2009); and by inflation between 2009 and 2010 from Bank of Canada (2010).


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 Reserves Revenues vs. Perfect Dispatch Margins
a * (perfect energy margins) + b = (actual energy margins) a * (perfect energy margins) + b = (OR revenues)
a b R2 a b R2
($/kW-yr) ($/kW-yr)

Coal 0.726 19.86 0.710 Coal 0.001 0.219 0.030

Gas Cogen 0.667 -21.68 0.817 Gas Cogen 0.033 -0.076 0.081
Gas CC 0.748 -42.91 0.918 Gas CC 0.031 2.877 0.104
Gas CT 0.484 -26.91 0.655 Gas CT 0.209 2.500 0.167
Hydro 0.295 -10.81 0.440 Hydro 0.290 -60.828 0.628
Wind 0.287 n/a n/a Wind 0.000 0.000 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.

Figure 37
Historic Gas CT Operating Margins vs. Perfect Dispatch Margins
$800 Energy
y = 0.484x + -26.91 100%
R-Squared: 0.7095 100%
Actual Energy Margins ($/kW-yr) &







Opeating Reserves
$0 $100 $200 $300 $400 $500 $600 $700 $800
OR Revenue ($/kw-yr)

$800 Operating Reserves

$600 y = 0.209x + 2.5
R-Squared: 0.1666

$0 $100 $200 $300 $400 $500 $600 $700 $800
Perfect Dispatch Energy Margins ($/kW-yr)
Sources and Notes:
Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c).

Figure 38
Historic Gas CC Operating Margins vs. Perfect Dispatch Margins
$900 yy==Energy
0.484x –+ 42.914
0.748x -26.91 100%
R-Squared: 0.7095 100%
Actual Energy Margins ($/kW-yr) &

R-Squared: 0.9177







Opeating Reserves
$0 $100 $200 $300 $400 $500 $600 $700 $800 $900 $1,000
OR Revenue ($/kw-yr)

Operating Reserves
y = 0.031x + 2.877
$600 R-Squared: 0.104

$0 $100 $200 $300 $400 $500 $600 $700 $800 $900 $1,000
Perfect Dispatch Energy Margins ($/kW-yr)
Sources and Notes:
Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c).

Figure 39
Historic Gas Cogen Operating Margins vs. Perfect Dispatch Margins
$900 y =Energy
y 0.484x + -26.91
= 0.667x -21.684 100%
R-Squared: 0.7095 100%
Actual Energy Margins ($/kW-yr) &

R-Squared: 0.8174







Opeating Reserves
$0 $100 $200 $300 $400 $500 $600 $700 $800 $900 $1,000
OR Revenue ($/kw-yr)

Operating Reserves
y = 0.033x + -0.076
$600 R-Squared: 0.0805

$0 $100 $200 $300 $400 $500 $600 $700 $800 $900 $1,000
Perfect Dispatch Energy Margins ($/kW-yr)
Sources and Notes:
Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c).

Figure 40
Historic Coal Operating Margins vs. Perfect Dispatch Margins
$1,000 y =Energy
y0.484x + -26.91
= 0.726x + 19.865 100%
R-Squared: 0.7095 100%
Actual Energy Margins ($/kW-yr) &

R-Squared: 0.7095





Opeating Reserves
$0 $200 $400 $600 $800 $1,000
OR Revenue ($/kw-yr)

$1,000 Operating Reserves

$800 y = 0.001x + 0.219 100%
$600 R-Squared: 0.0302

$0 $200 $400 $600 $800 $1,000
Perfect Dispatch Energy Margins ($/kW-yr)
Sources and Notes:
Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c).

Figure 41
Historic Hydro Operating Margins vs. Perfect Dispatch Margins
0.484x +– -26.91
0.295x 10.806 100%
R-Squared: 0.7095 100%
Actual Energy Margins ($/kW-yr) &

$1,000 R-Squared: 0.9177





Opeating Reserves
$0 $200 $400 $600 $800 $1,000 $1,200
OR Revenue ($/kw-yr)

$1,000 Operating Reserves
y = 0.29x + -60.828
R-Squared: 0.6277

$0 $200 $400 $600 $800 $1,000 $1,200
Perfect Dispatch Energy Margins ($/kW-yr)
Sources and Notes:
Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c).

Figure 42
Historic Wind Operating Margins vs. Perfect Dispatch Margins
$900 Energy
yy == 0.484x
0 100%
R-Squared:0.7095 100%
Actual Energy Margins ($/kW-yr) &








Opeating Reserves
$0 $100 $200 $300 $400 $500 $600 $700 $800 $900 $1,000
OR Revenue ($/kw-yr)

Operating Reserves
y = 0x + 0
$600 R-Squared: n/a

$0 $100 $200 $300 $400 $500 $600 $700 $800 $900 $1,000
Perfect Dispatch Energy Margins ($/kW-yr)
Sources and Notes:
Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c).


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.

Figure 44
Projected Gas Cogen 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.

Figure 45
Projected Hydro 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.

Figure 46
Projected Wind 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.

JULY 2010

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, AEMO’s liability is limited, at AEMO’s 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.

The Electricity Supply Industry 2
The National Electricity Market 4
The Australian Energy Market Operator 5
National Electricity Law and Rules 6
The Spot Market 7
Key Parameters for NEM Operation 8
Operating the NEM 9
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


Sectors of the electricity supply industry are involved Electricity can be converted readily to heat and light How is Electricity Produced?
with the generation, transmission, distribution and and used to power machines. It can also be transported
retail sale of electricity. Australia’s social, industrial and with relative ease. These characteristics make electricity Electricity can be produced by either chemical means or
commercial success depends on the reliability of the a convenient and manageable form of energy, and mechanical action. Electricity produced by chemical means
electricity supply. In this way, the industry contributes contribute both to its value as a commodity and its relies on a flow of charged particles from cells in a battery.
significantly to the national economy. versatility as a source of power. While this type of electricity has some very important
applications in modern society, it is an expensive
A unit of power is referred to as a watt. The number of production process and can meet only limited, specific
What is Electricity? watts, or wattage, of an electrical appliance indicates the requirements for electricity. The generators in modern
Electricity is a form of energy produced by the flow of rate at which the appliance converts electrical energy to power stations produce electricity by the mechanical
electrons in a substance known as a conductor. The best another form of energy such as heat or light. One watt action of large, powerful magnets that spin rapidly inside
conductors are metals such as copper and aluminium, is equivalent to one joule of work per second. Both the the huge coils of conducting wire driven by steam, gas
and are commonly used in electrical wiring. electrical pressure (voltage) and the number of electrons or water turbines.
flowing (current) determine the electrical power or rate
Energy exists in many forms. Electricity is a secondary of energy conversion. A 60-watt light globe uses 60 watts More than 90 per cent of Australia’s electricity production
energy source as it is produced by the conversion of other of electricity to produce light, and a typical electric kettle relies on the burning of fossil fuels - coal, gas and oil.
energy sources like the chemical energy in coal, natural uses 2400 watts to produce heat. The chemical energy stored in these fuels is used to heat
gas and oil. Other primary sources of energy, like the water and produce steam. The steam is then forced under
sun and wind, are increasingly being used to produce great pressure through a turbine that drives a generator
electricity. A quantity of energy can be changed or to produce electricity. The complete process involves
converted, but can never be created or destroyed. 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.


One megawatt (MW) One gigawatt (GW) One megawatt hour A 100 MW generator A 600 MW generator
is equal to one million is equivalent to one (MWh) is the energy will power one million has sufficient capacity
watts (W). thousand megawatts. required to power ten 100 W light globes to service 200,000
thousand 100 W light simultaneously. domestic customers.
globes for one hour.



How is Electricity Transported?
Transmission lines Distribution lines
Electricity travels along a conductor at close to the carry electricity carry low voltage
speed of light. When an appliance is switched on, power long distances electricity to
is instantly transmitted from a power station to the consumers
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 Power plant Transformer Substation transformer Homes, offices and
the power outlets in homes, offices and factories. generates converts low voltage converts high voltage factories use electricity
electricity electricity to high electricity to low voltage for lighting and
voltage for efficient for distribution heating and to power
transport appliances


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.

The National Electricity Market (NEM) began operating as Some assets that comprise the NEM’s infrastructure Electricity is an ideal commodity to be traded using pool
a wholesale market for the supply of electricity to retailers are owned and operated by state governments, and arrangements because of two of its unique characteristics.
and end-users in Queensland, New South Wales, the some are owned and operated under private business Electricity cannot be stored for future use, so supply must
Australian Capital Territory, Victoria and South Australia arrangements. vary dynamically with changing demand. And because one
in December 1998. Tasmania joined the NEM in 2005 and unit of electricity is indistinguishable from all other units,
operations today are based in five interconnected regions Exchange between electricity producers and electricity it is impossible to determine which generator produced
that largely follow state boundaries. consumers is facilitated through a pool where the output which electricity.
from all generators is aggregated and scheduled to meet
The NEM operates on the world’s longest interconnected demand. The electricity pool is not a physical location; Sophisticated information technology systems underpin
power system – from Port Douglas in Queensland to Port rather it is a set of procedures that AEMO manages the operation of the NEM. The systems balance supply
Lincoln in South Australia – a distance of around 5,000 according to the provisions of National Electricity Law and with demand, maintain reserve requirements, select which
kilometres. More than $10 billion of electricity is traded Statutory Rules (the Rules) and in conjunction with market components of the power system operate at any one
annually in the NEM to meet the demand of more than participants and regulatory agencies. time, determine the spot price, and thereby facilitate the
eight million end-use consumers. financial settlement of the physical market.



NATURAL GAS3: 12.2 %

METALS: 18.3%
1 Excludes embedded and
non-grid private generation
2 Includes generation from
semi-scheduled and large Source: Electricity Gas Australia 2010 ESAA
non-scheduled intermittent
3 Includes generation from
coal seam methane

The Australian Energy Market Operator


The Australian Energy Market Operator (AEMO) was AEMO operates on a cost recovery basis as a corporate With respect to the electricity market AEMO has two
established to manage the NEM and gas markets from entity limited by guarantee under the Corporations Law. core roles:
1 July 2009. Its membership structure is split between government
and industry, respectively 60 and 40 percent, with • Power System Operator
AEMO’s core functions can be grouped into the this arrangement to be reviewed after three years of • Market Operator
following areas: operation. Government members of AEMO include
The market requirements determine how the power
• Electricity Market - Power System and Market Operator the Queensland, New South Wales, Victorian, South
system is operated.
• Gas Markets Operator Australian and Tasmanian state governments, the
• National Transmission Planner Commonwealth and the Australian Capital Territory. AEMO’s electricity market and system operation
• Transmission Services responsibilities include:
• Energy Market Development AEMO and the NEM
•  anagement of the NEM
Created by the Council of Australian Governments A key aim of AEMO is to provide an effective • Overseeing reliability and security of the NEM
(COAG) and developed under the guidance of the infrastructure for the efficient operation of the wholesale • Ensuring supply reserve to meet reliability standards
Ministerial Council on Energy (MCE), AEMO strengthens electricity market, to develop the market and improve • Directing generators to increase production during
the national character of energy market governance by its efficiency and to coordinate planning of the periods of supply shortfall
drawing together under the one operational framework interconnected power system. • Instruction of load shedding to rebalance supply and
responsibility for electricity and gas market functions, demand and protect power system operations
NEM system operations, management of Victoria’s gas AEMO’s primary responsibility is to balance the demand • Implementation of reserve trading to maintain supply
transmission network and national transmission planning. and supply of electricity by dispatching the generation and reliability levels through demand-side response
necessary to meet demand. • National transmission planning for the electricity
AEMO carries out the electricity functions previously transmission grid and production of a National
undertaken by the National Electricity Market AEMO’s key financial objective of being self-funding is Transmission Network Development Plan
Management Company (NEMMCO) with respect to the achieved through the full recovery of its operating costs • Publication of the Electricity Statement
NEM and the planning responsibilities of the Electricity from fees paid by market participants. of Opportunities
Supply Industry Planning Council (ESIPC, South Australia). • Electricity emergency management
The National Electricity Law and the Rules were amended
Additionally, AEMO assumed the retail and wholesale gas • Facilitation of Full Retail Competition.
to replace NEMMCO with AEMO as the national
market responsibilities of the Victorian Energy Networks
electricity market and system operator. AEMO manages the market and power system from two
Corporation (VENCorp), Retail Energy Market Company
(REMCO), Gas Market Company (GMC) and Gas Retail AEMO’s functions are prescribed in the National control centres in different states. Both centres operate
Market Operator (GRMO). Electricity Law while procedures and processes for market around the clock, and are equipped with identical
operations, power system security, network connection communication and information technology systems.
As part of its gas market functions, AEMO is responsible The entire NEM, or individual regions within it, can be
and access, pricing for network services in the NEM and
for the establishment of a Short Term Trading Market, due operated from either or both centres. This arrangement
national transmission planning are all prescribed in
to commence in 2010 (initially in the New South Wales ensures continuous supply despite the risks posed by
the Rules.
and South Australia hubs), which sets a daily wholesale natural disasters or other critical events, and provides
price for natural gas. AEMO with the flexibility to respond quickly to dramatic
changes in the market or the power system.

When the NEM commenced, a National Electricity Wholesale trading in electricity is conducted as a spot
Code provided guidelines for how the market was to market where supply and demand are instantaneously
operate. These guidelines were developed following matched in real-time through a centrally-coordinated
comprehensive consultation and extensive trials dispatch process. Generators offer to supply the
conducted between governments, the electricity market with specific amounts of electricity at particular
supply industry and electricity users as part of a prices. Offers are submitted every five minutes
government-driven deregulation and reform agenda. of every day. From all offers submitted, AEMO
determines the generators required to produce
In June 2005, the National Electricity Code was replaced electricity based on the principle of meeting prevailing
by the National Electricity Law and Rules. The Law and demand in the most cost-efficient way. AEMO then
Rules were recently amended to replace NEMMCO dispatches these generators into production.
with AEMO as the national electricity market and
system operator. A dispatch price is determined every five minutes, and six
dispatch prices are averaged every half-hour to determine
AEMO’s functions are prescribed in the National the spot price for each trading interval for each of the
Electricity Law while procedures and processes for market regions of the NEM. AEMO uses the spot price as the
operations, power system security, network connection basis for the settlement of financial transactions for all
and access, pricing for network services in the NEM and energy traded in the NEM.
national transmission planning are all prescribed
in the Rules. 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 Trends in spot price movement provide signals for future
investment in generation and transmission infrastructure
The Rules place a limit on the maximum spot price at any in the NEM. As the capacity of available generation to
regional reference node. This limit is called the Market meet demand diminishes, relative scarcity will lead to an AEMO
Price Cap and is $12,500 per MWh. The spot price may be increase in the spot price, and new generation or network
set to the cap value when it is necessary to involuntarily capacity will be attracted into the market. High spot
interrupt electricity supplies (ie load shedding) to Market prices during periods of supply scarcity may also act as an
Customers in order to balance the overall electricity incentive for consumers to reduce their demand.
supply and demand.
The NEM is a wholesale market. Up to 50 percent of GENERATOR
Market Floor Price the price paid by domestic and business consumers for
electricity supply is accounted for by the direct cost of
The Rules place a limit on the minimum spot price. This the energy. Additional charges are added to retail
limit is called the Market Floor Price and is currently set at accounts for network usage, service fees, market charges,
-$1,000 per MWh. retail charges and GST. TRANSMISSION NETWORK
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 TOTAL ENERGY SENT OUT 2008/09
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
SA 6.5%
consumed through electrical resistance and the heating
NSW 38%
up of conductors. Secondly, electricity being transported
along certain elements of the network may encounter
VIC 25.1%
technical constraints on capacity or bottlenecks.
QLD 25.4%

Source: ESAA





AEMO is required to operate the power system The minimum reserve levels across the different During a period of load shedding, supply is withdrawn
efficiently and ensure agreed standards of security and NEM regions are listed in the Electricity Statement of from those NEM regions affected by the shortfall in
reliability are maintained. Opportunities on the AEMO website. proportion to the demand levels at the time the shortfall
began. The proportioning process determines the amount
Security of Supply Managing Security and Reliability of load shedding for each affected region up to the point
where interconnectors are operating to their maximum
AEMO’s highest priority as power system and market In all but extraordinary circumstances, market forces keep transfer capacity. Once the interconnectors reach their
operator of the NEM is the management of power system supply and demand in the NEM in balance. However, maximum transfer capacity, the importing region must
security. Security of electricity supply is a measure of the during periods of supply shortfall when system security or bear any additional load shedding locally.
power system’s capacity to continue operating within reliability of supply is threatened, the Rules endow AEMO
defined technical limits despite the disconnection of a with authority to use a variety of tools to restore supply By implementing load shedding, AEMO protects the
major power system element, such as a generator or and demand balance. The tools include demand side integrity of power system operation so that widespread
interconnector. management, the power of direction, load shedding and and long-lasting blackouts are avoided. It also ensures
reserve trading. that the hardship caused by a sustained supply shortfall is
The maintenance of power system security ensures the shared in an equitable fashion.
power system is operated in a way that does not overload
or damage any part of it or risk overload or damage after Security and Reliability Directions
Reserve Trading
a credible event. AEMO has the power to direct registered generators into
When there is sufficient notice of an upcoming shortfall
production when a supply shortfall is expected and some
of supply that threatens to compromise minimum reserve
Power System Reliability generators are known to have withheld some of their total
margins, AEMO may tender for contracts for electricity
capacity from the market. AEMO only uses this power
Reliability is a measure of the power system’s capacity to supply from sources beyond those factored into AEMO’s
of direction to protect power system security or supply
continue to supply sufficient power to satisfy customer usual forecasting processes. At these times, emergency
demand, allowing for the loss of generation capacity. generators and other generators connected directly to
The shortfall of supply against demand is referred to as the distribution network who submit tenders may enter
unserved energy. Reliability standards are established in Load Shedding contracts to boost supply in the NEM so the widespread
the NEM that determine that unserved energy per year supply interruptions that may otherwise have occurred can
In the event that demand in a region exceeds supply
for each region must not exceed 0.002 percent of the total be avoided. In the same way, some electricity consumers
and all other means to satisfy demand have been
energy consumed in that region that year. may offer for a financial consideration to decrease their
implemented, AEMO can instruct network service
demand at times of supply shortfall so that demand and
providers to shed some customer load. This action is only
supply are brought into balance.
Supply Reserve taken when there is an urgent need to protect the power
system by reducing demand and returning the system to
The power system is required to be operated at all balance. Load shedding involves a temporary suspension
times with a certain level of reserve in order to meet of supply to customers in a specific part or region of the
the required standard of supply reliability across the NEM where system security is at risk.
NEM. Calculation of the minimum reserve requirements
recognises reserve sharing in a national context.



Operating the NEM involves conducting a sequence A typical level of demand for electricity across the NEM
of activities to facilitate trade between the producers is approximately 25,000 megawatts on a business day of

and wholesale consumers of electricity. These activities average temperatures. There is ample supply available in 9000
include establishing demand levels, receiving offers the system to meet this level of demand. In fact, supply
to supply from generators, scheduling generators, only comes under extreme pressure for a few hours on
dispatching generators into production, calculating just a few days of extreme high temperature that occur
the spot price, measuring electricity use and financially each year. Further, because peak demand does not occur


settling the market. simultaneously in all regions, total supply can be shared
between regions using the interconnected power network. 6000
Demand in the Victorian and South Australian regions of
AEMO conducts forecasts of expected electricity demand the NEM is characterised by short-term demand peaks
in order to operate the NEM. Demand varies from region during the summer months. It makes economic and
to region depending on population, temperature, and the market sense that these extreme peaks of demand be
industrial and commercial needs. It also varies throughout met by special arrangements rather than having excess 3000
the day, with daily demand peaks (driven by domestic base-load generation capacity in the system at all times.


activity) typically occurring between 7:00 am and 9:00 am The peaks are currently being met by a combination of
and between 4:00 pm and 7:00 pm. generators that have been specifically built to service
extreme demand periods (peak generators), and demand
side participation, where consumers voluntarily and 0
temporarily withdraw from the market when the spot price





reaches a threshold level.



Source: ESAA


Supply Scheduling and Dispatching Generators

The delivery of electricity to market customers comprises From the bids submitted, AEMO’s systems determine
a sequence of distinct processes that AEMO manages which generators are required to satisfy demand, at
according to strict timetables. 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
Submitting Offers to Supply
production. The use of the rising-price stack means that
To enable AEMO’s systems to facilitate supply, scheduled more expensive generators are scheduled into production
NEM generators are required to submit to AEMO offers as total demand for electricity increases.
indicating the volume of electricity they are prepared to
At times, the technical capacity of the transmission
produce for a specified price.
network may determine which generators are scheduled
There are three types of bids or offers to supply – daily to meet demand. In such a situation, generators may be
bids, re-bids and default bids. Daily bids are submitted scheduled out of price order so that demand in
before 12:30 pm on the day before supply is required, a particular area supplied through the network may
and are reflected in pre-dispatch forecasts. Generators be satisfied.
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 ‘commercial-
in-confidence’ nature and, in general, reflect the base
operating levels for generators.
From the bids
submitted, AEMO’s
systems determine
which generators are
required to satisfy



500 Bids to produce electricity received by AEMO D. By 4:20pm, demand has increased to the
are stacked in ascending price order for point that Generator 5 is just required to meet

D each dispatch period. Generators are then demand, and the price increases to $38 per MWh.
400 F progressively scheduled into production to
C E. At 4.25 pm, Generators 1-4 are fully
meet prevailing demand, starting with the
B least-cost generation option.
dispatched and Generator 5 partly dispatched.
A $37 The price remains at $38 per MWh.

300 A. In order to supply demand for power at 4:05

$35 F. By 4:30 pm, demand has fallen. Generator
pm, Generators 1 and 2 are dispatched to their
5 (the most expensive generator) is no longer
full bid capacity, and Generator 3 is only partly
required, and Generator 4 is only partly
200 dispatched. The price is $35 per MWh.
dispatched. The price returns to $37 per MWh.
$28 B. At 4:10 pm, demand has increased:
The spot price for the trading period is
Generators 1, 2 and 3 are fully dispatched,
100 calculated as the average of the six dispatch
and Generator 4 is partly dispatched.
prices. That is, $(35+37+37+38+38+37) per MWh
$20 The price is $37 MWh.
divided by six, or $37 per MWh. This is the price
0 C. At point C (4:15 pm) demand has increased all generators receive for production during this
4:05 4:10 4:15 4:20 4:25 4:30 a further 30 MW. Generators 1, 2, 3 and 4 period, and the price market customers pay for
continue producing power and the price remains electricity they consume from the pool during
at $37 MWh. this period.


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

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 low fuel cost with plentiful suitable for remote and
operating cost water supply; production stand-alone applications;
severely affected by drought batteries may be used to
store power


Setting the Spot Price Factors that contribute to variations in the spot price by local network service providers. These service providers
in different regions of the NEM include limits on are responsible for measuring the volume of electricity
AEMO issues dispatch instructions to generators at interconnector capacity and reliance on differing fuel supplied, validating the data from the meters, applying
five-minute intervals throughout each day based on the sources for local supply in different NEM regions. Because distribution loss factors, and forwarding the information to
offers generators have submitted in the bidding process. gas is a more expensive fuel than coal or water, electricity AEMO for use in calculating and preparing accounts for
In this way, there are 288 dispatch intervals every day. produced using gas will generally cost more than financial settlement.
The dispatch price represents the cost to supply the last electricity produced by the other means. Other factors
megawatt of electricity to meet demand, and applies to – including total system load, plant outages, frequency
all generators scheduled into production regardless of
Settling the Market
control, voltage control, testing and transmission outages
the level of their original offer. – also affect the dispatch and spot prices. During 2007-08, AEMO calculates the financial liability of all market
the average daily spot price across all regions of the NEM participants on a daily basis and settles transactions for all
A trading interval in the NEM is a half-hour period. Hence,
was $52 per MWh. trade in the NEM weekly. This involves AEMO collecting
there are 48 trading intervals in the market each day. The
all money due for electricity purchased from the pool
spot price of electricity for all 30-minute trading intervals
Measuring Electricity Use from market customers, and paying generators for the
each day is the average of the six dispatch prices during
electricity they have produced. The spot price is the basis
the preceding half-hour. There is a separate spot price for
All market customers are required to install equipment for all these financial transactions.
each trading interval in each of the NEM’s five regions.
to record their electricity consumption. AEMO registers,
accredits and audits a range of metering services provided 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
THE GENERATOR DISPATCH CYCLE arrangements and a robust risk management program in
place. As part of this, AEMO requires the deposit of bank
Scheduling guarantees and security deposits against an established
• Ranking bids maximum credit limit for each market customer. AEMO
• Identifying the dispatch levels closely monitors the activities of all participants in the
of generating units market and has a firm timetable in place for the entire
settlement process.

The settlement process involves determining the financial

Data input
• Establishing current liabilities, issuing accounts, and settling amounts payable
operational status of and receivable for electricity sold to and purchased
generating units from the pool. The settlement price for both generators
• Assessing demand and market customers is equal to the amount of energy
forecasts produced or consumed multiplied by both the spot price
• Applying loss factors Dispatch that applies in the region of their operation and any loss
• Determining system • Issuing dispatch instructions
factors that apply.
conditions to generators

If a market participant breaches their maximum credit limit
on any one day of trading, a call notice for rectification of All market customers are
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
required to install equipment
a market participant who fails to respond adequately
to a default notice, and to reinstate that market participant to record their electricity
only when their required financial position
is re-established.
consumption. AEMO registers,
Demand Side Participation
accredits and audits a range
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
of metering services provided
high spot prices. This is a deliberate action taken when
demand for power drives spot prices high. by local network service
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.


Ancillary services are those services used by AEMO NCAS are primarily used to:
to manage the power system safely, securely and
reliably. Ancillary services maintain key technical • C ontrol the voltage at different points of the electrical
characteristics of the system, including standards network to within the prescribed standards; or
for frequency, voltage, network loading and system • Control the power flow on network elements to within
re-start processes. the physical limitations of those elements

AEMO operates eight separate markets for the delivery SRAS are reserved for contingency situations in which
of Frequency Control Ancillary Services (FCAS), and there has been a major supply disruption or where the
purchases Network Control Ancillary Services (NCAS) electrical system must be restarted.
and System Restart Ancillary Services (SRAS) under
Ancillary service costs are dependant upon the amount
agreements with service providers.
of service required at any particular time and, as these
FCAS providers bid their services into the FCAS markets amounts can vary significantly from period to period, costs
in a similar way to how generators bid into the energy will also vary.
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

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.



The NEM comprises five interconnected electrical Regulated Interconnectors
regions. There is a designated region reference node in
each region where the regional spot price of electricity A regulated interconnector is an interconnector that has
is set. The Queensland, New South Wales, Victoria, passed the ACCC-devised regulatory test and has been
Tasmania and South Australia regions all contain both deemed to add net market value to the NEM. Having
major generation and demand centres. 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
Interconnectors usage. The revenue is collected as part of the network
The high-voltage transmission lines that transport charges included in the accounts of electricity end-users.
electricity between adjacent NEM regions are called At present, regulated interconnectors operate between
interconnectors. Interconnectors are used to import all adjacent regions of the NEM, except Tasmania.
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.

AEMO’s 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
interconnector’s capacity is reached, the interconnector REGIONAL REFERENCE NODE
is said to be constrained. For example, if prices are REGULATED INTERCONNECTOR
very low in one region and high in an adjacent region,
electricity can be sent from the first to the second region QLD
across an interconnector up to the maximum technical SOUTH PINE
capacity of the interconnector. AEMO’s systems will then NSW-QLD
dispatch local generators with the lowest price offers
from within the second region to meet the outstanding TERRANORA
consumer demand. VIC-SA NSW WEST SYDNEY


Unregulated Interconnectors Loss of Energy in the System

Unregulated (or market) interconnectors derive revenue As electricity flows through the transmission and REGION A
by trading in the spot market. They do this by purchasing distribution networks, energy is lost due to electrical
energy in a lower price region and selling it to a higher resistance, and the heating of conductors. The losses REGION A
price region, or by selling the rights to revenue generated are equivalent to approximately 10 per cent of the total NODE
by trading across the interconnector. Unregulated electricity transported between power stations and
interconnectors are not required to undergo regulatory market customers.
test evaluation. REGION
Energy losses on the network must be factored in at all
An unregulated interconnector – Basslink – operates stages of electricity production and transport to ensure
between the Tasmanian and Victorian regions of the the delivery of adequate supply to meet prevailing
NEM. Murraylink and Directlink were built as unregulated demand and maintain the power system in balance.
interconnectors between Victoria and South Australia, In practical terms, this means that more electricity must
and New South Wales and Queensland respectively. be generated than indicated in demand forecasts in order REFERENCE
They successfully applied to the ACCC for conversion to allow for this loss during transportation. NODE
to regulated status.
The impact of network losses on spot prices is
mathematically represented as transmission and CUSTOMER 1
distribution loss factors. Loss factors within each region
of the NEM are calculated based on forecast demand, REGION B
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 INTRA-REGIONAL LOSS FACTOR APPLIES
calculated and reflect the operating conditions at the time
of the transmission of the electricity. 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
customer’s connection point to the grid. In the diagram,
As electricity flows through the customer C1 would require more energy to be imported
to receive the same amount of supply as customer C2,
transmission and distribution networks, because C2 is closer to the regional reference node.

energy is lost due to electrical resistance,

and the heating of conductors.


AEMO uses a variety of forecasting processes to Pre-dispatch Forecasting Projected Assessment of System Adequacy
determine the level of demand for every dispatch
interval in the NEM. Then using the submitted offers Pre-dispatch is a short-term forecast of supply and AEMO monitors the future adequacy of generating
to generate electricity, AEMO produces a schedule or demand in the market. It is used to estimate the price capacity based on the predicted availability of generating
timetable of generation to ensure that the forecast and demand for the upcoming trading day, and the units at power plants. AEMO produces both seven-day
demand will be met based on the requirements that volume of electricity expected to be supplied through the and two-year forecasts because of the variability
the least expensive generators are dispatched into interconnectors between regions. of demand for electricity. These forecasts are called the
production and the power system remains in a secure short-term and medium-term Projected Assessments of
Generators and network operators are required to notify System Adequacy, or PASA, respectively. They are used
operating state.
AEMO of their maximum supply capacity and availability, by AEMO to ensure that adequate levels of reserve are
As a prerequisite for maintaining supply and demand in and this information is matched against regional demand in the system at all times, and by generators and network
balance, it is important for AEMO’s planning processes forecasts. All offers to supply are then collated so that operators to plan augmentation, maintenance and
to be informed in advance of any limits on the capacity potential shortfalls of supply against demand can be other outages.
of generators to supply electricity or networks to identified and published. Participants in the market use
transport electricity. This enables the remainder of market this information as the basis for any re-bids of the capacity
they wish to bring to the market.
participants to respond to potential supply shortfalls by AEMO PRODUCES TWO PASA FORECASTS
increasing their generation or network capacity to the
market. Market participants are able to signal upcoming Five-minute Matching of Supply and Demand Forecast Forecast Updated/
limitations on supply by means of a variety of forecasting Period Published
tools designed to improve the overall efficiency of Generators are scheduled and dispatched into production
to match supply with prevailing demand every five minutes Short-term 7 days 2-hourly from
the market.
PASA 4:00am
of every day. This process, in turn, produces dynamic
price signals that guide market participants as they bid to Medium-term 2 years 2:00pm every
supply electricity to the market. PASA Tuesday



24 60
18 6 45 15


1 HOUR (288 PER DAY)


Electricity Statement of Opportunities National Transmission Network

Development Plan
AEMO publishes a 10 year forecast called the electricity
Statement of Opportunities (SOO) each year. This AEMO is the National Transmission Planner for the
publication provides information to assist market electricity transmission grid. A core component of this
participants assess the future need for electricity transmission planning responsibility involves preparing
generating capacity, demand side capacity and annual network development plans to guide investment in
augmentation of the network to support the operation the power system.
of the NEM.
In 2009 an interim National Transmission Statement (NTS)
It also contains forecasts of ancillary service requirements, replaced the previous Annual National Transmission
minimum reserve levels, and economic and operational Statement produced by NEMMCO. This document will
data to assist potential investors gain a full understanding be superseded by the National Transmission Network
of the NEM. Development Plan (NTNDP) from 2010.

The Electricity SOO brings together information supplied The NTNDP will:
to AEMO by the planning bodies in each jurisdiction of
the NEM. A year-by-year annual supply-demand balance • p rovide historical data and projections of network
is presented for each region in the SOO as a snapshot utilisation and congestion;
forecast of the capacity of generation and distribution to • summarise emerging reliability issues and potential
satisfy demand for electricity into the future. 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.



Since the commencement of the NEM, electricity One of AEMO’s responsibilities under the Rules is to Market participants include market generators, market
consumers have progressively gained the right to register participants in the NEM. There are six main network service providers and market customers. A market
choose their own supplier. This has meant that AEMO’s categories of registered participant, including those participant must be separately registered in each category
responsibilities have extended from managing the who participate directly in trading activities, and other of the market in which it participates. For example, a
wholesale market to providing the systems and processes participants who provide services essential for the business that participates as a generator (a peaking
to support competition and choice for all end-users in the operation of the market. The categories are generator, generating plant for instance) and as a market customer
retail electricity market. Delivering full retail competition market customer, intending participant, network service (retailer of electricity to end-use customers) would be
(FRC), or contestability, has required new information provider, trader, reallocator and special participant. required to be registered as both a generator and a
technology systems to process transfers of customers market customer.
between registered retailers in the NEM. The systems that
facilitate this function contain one of the largest metering The registration of participants is a formal process, strictly
databases in the world. They accept data from a variety of defined in the Rules. Registered participants are required
electricity meter types and have the capacity to process to pay participant fees that are levied to recover the costs
information from up to 10 million meters. associated with managing the market.

AEMO’s systems are set up to provide key meter REGISTERED PARTICIPANTS

installation details to support a simple and rapid
information transfer process. Different metering processes Market Participants Other Registered Participants
are required for different types of meters used in the
Registered to participate in the National Electricity Market
NEM, to support consumer transfer and core settlement Transmission Network Service Provider
Owner and operator of the high-voltage transmission towers
procedures and to calculate load profiles. The cost of Market Generators Market Network Service
and wires that transport electricity.
electricity consumed is then calculated according to Sell entire electricity output Providers
through the spot market Own and operate a network Distribution Network Service Provider
a user profile that approximates the pattern of use in Owner and operator of substations and the wires that transport
and receive the spot price at linked to the national grid at
a typical situation. settlement. two terminals in different NEM
from distribution centres to end-use consumers. Also provider
of technical services, including construction of power lines,
Scheduled: aggregate regions. Pay market participant inspection of equipment, maintenance and street lighting.
By June 2009, approximately 6.3 million customer transfers fees and obtain revenue from
generation capacity of more than
from one retailer to another had taken place. As a result or equal to 30 megawatts. trading in the NEM. Reallocator
Registered with AEMO to participate in ‘reallocation transactions’
of the introduction of full retail competition, electricity Semi-scheduled: aggregate under clause 3.15.11 of the National Electricity Rules.
Market Customers
retailers are increasingly competing, and creating new generation capacity of more than
or equal to 30 megawatts where Purchase electricity supplied to Special Participant
and unique products as a means to increase their System operators or agents appointed to perform power security
output is intermittent. a connection point on a NEM
customer bases. Non-scheduled: aggregate transmission or distribution functions. Distribution system operators and controllers or
operators of any portion of the distribution system.
generation capacity of less than system for the spot price.
30 megawatts. Electricity Retailers: buy Intending Participant
electricity at spot price and Must reasonably satisfy AEMO of intention to perform activity that
would entitle it to be a registered participant.
on-sell it to end-use customers.
End-use Customers: buy directly Trader
from the market for own use. Party registered to participate in the settlement residue auction.


Participants in the NEM require a means of managing

In this example of a hedge contract,
the financial risks associated with the significant $160
the two parties have agreed to set a
degree of spot price volatility that occurs during price (the strike price) of $40/MWh.
trading periods. They typically achieve this by $140 SPOT PRICE The graph shows the strike price
using financial contracts that lock in a firm price for and the actual spot price over a
hypothetical day’s trading.
electricity that will be produced or consumed at a
$120 When the spot price is below the
given time in the future. These contracts serve to
substantially reduce the financial exposure of market strike price, the market customer pays
the difference in these prices to the
participants and contribute to spot market stability. $100 generator. In this case, when the spot
They are known as derivatives, and include swaps BUYER PAYS SELLER price is $17, the market customer pays

or hedges, options and futures contracts. DIFFERENCE BETWEEN the generator $23/MWh.
AGREED STRIKE PRICE AND SPOT PRICE When the spot price exceeds the
strike price, the generator pays the
Hedge Contracts $60 market customer the extra required
to purchase electricity from the pool.
Hedge contracts are typically agreements between In this case, when the spot price rises
generators and customers that operate independently of $40 to $145/MWh, the generator pays
both the market and AEMO’s administration. The details the market customer the $105/MWh
of hedge contracts are not factored into the balancing difference.
of supply and demand, and are not regulated under the (AGREED CONTRACT PRICE)
Rules. These contracts can be entered into under either
long-term or short-term arrangements that set an agreed,
0400 0800 1200 1600 2000 0000
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. HEDGE CONTRACTS IN THE NEM

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
Hedge contracts are typically
price is above the strike price. When the spot price is
below the strike price, customers pay generators the
agreements between generators
difference between the spot price and the strike price. and customers that operate
independently of both the market
and AEMO’s administration.


Auctions of Inter-region Settlement Residues

The spot price for electricity in each region of the NEM is REGION A
determined by a number of factors, including supply and
demand, the physical limitations of interconnectors, and SPOT PRICE
the loss factors for both the transmission and distribution $100/MWH
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 GENERATORS

region where it is generated and its value if sold in another ARE PAID AT
region is called the inter-regional settlement residue. $100/MWH
The settlement residue that accumulates is made available
to the market by the conduct of an auction. The auction REGION BOUNDARY
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 SETTLEMENT RESIDUE BUY AT
price outcomes between trading regions. ACCRUED IN REGION B: $120/MWH
Registered participants who purchase auction units obtain FACTORS)
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. $120/MWH





The range of technologies for the generation of In late 2008, the Australian Wind Energy Forecasting AEMO’s role as the manager of both the power system
electricity is expanding to accommodate alternative System (AWEFS) was implemented to forecast the and the electricity market means that it is sometimes
energy sources such as wind energy. energy likely to be produced by major wind farms in the asked about issues of environmental management, the
NEM. This system enabled the generator classification sustainability of the market and the electricity supply
Large wind generators are typically registered as semi- ‘semi-scheduled’ to be introduced in the NEM, allowing industry in general.
scheduled generators (rather than scheduled) because intermittent generators (including wind farms) to compete
their energy source is intermittent and their generation in the NEM through the bidding process. Under the Rules, AEMO’s charter focuses specifically
cannot increase on demand. The market is designed to on efficiency, security and reliability of power supply,
allow intermittent generators to participate and share the The success of AWEFS and the semi-scheduled category and excludes favouring one fuel source over any other.
same power system and the same consumers. will help the energy market respond to an ongoing Consequently, AEMO has neither the power nor the
expansion in renewable generation while maintaining authority to make decisions based on considerations of
The NEM’s base-load generators are scheduled the security of supply that has been the hallmark sustainability and balance in resource management.
according to bids, and production from each generating of the industry.
unit is controlled by operators. The changeability and The various state regulators ensure that environmental
unpredictability of wind means that wind generators impact assessments are conducted as part of any power
cannot be scheduled in the usual way. industry planning initiatives. The regulators also monitor
operations at industry sites within their jurisdictions, and
The integration of wind and other intermittent generators the industry itself operates and audits waste reduction and
to the NEM must take account of AEMO’s responsibility to recycling programs.
maintain power system security, and be managed during
each five-minute dispatch interval. The variation of output
associated with wind generators (with individual output Renewable Energy Target
that can change by as much as 50 per cent in a five-minute The Federal Government’s expanded Renewable Energy
dispatch interval) may require interconnectors to operate Target (RET) will result in changes to the generation mix in
at lower limits to avoid overloads, and hence reduce the the NEM over the next decade.
total supply capacity available to the market.
AEMO will continue to provide advice as requested by
government and the Ministerial Council on Energy with
regard to these changes in the NEM’s 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.


AEMO is not responsible for market regulation. Since The AER has responsibility for the enforcement of
mid-2005, the Australian Energy Market Commission and monitoring compliance with the Rules, as well
(AEMC) and the Australian Energy Regulator (AER) as responsibility for economic regulation of electricity
have had responsibility for oversight and regulation transmission. The AER issues infringement notices for
of the National Electricity Market. certain breaches of the National Electricity Law and
Rules, and is the body responsible for bringing court
The AEMC is responsible for rule making and market proceedings in respect of breaches.
development. The rule-making role does not involve
initiating changes to the Rules other than where the A Memorandum of Understanding between the ACCC,
change involves correcting minor errors or where the AER and the AEMC guides interaction between these
the change is of a non-material nature. Rather, the three bodies and their function in the Australian energy
role involves managing the rule change process, and industry. The regulatory bodies have been created under
consulting and deciding on rule changes proposed by the auspices of the Ministerial Council on Energy and take
others. In regard to its market development function, the over many of the electricity regulatory arrangements
AEMC conducts reviews at the request of the Ministerial that were previously the responsibility of state government
Council on Energy or at its own volition on the operation authorities.
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.


Renewable generation Networks (Transmission)

Energy conversion techniques including wind, solar, hydro Transmission lines carry high voltage electricity to
and geothermal. The primary commercial sources at this substation transformers where it is changed to low voltage
time are hydro and wind power. for distribution.

Hedge contracts Networks (Distribution)

Long term or short term arrangements that set a strike Distribution lines carry low voltage electricity to
(agreed) price for electricity traded through the pool. consumers who access it through the power outlets
in homes, offices and factories.
The high-voltage transmission lines that transport Transformers
electricity between adjacent NEM regions. Convert the electricity produced at a generation plant
from low to high voltage to enable its efficient transport
Intermittent generators on the transmission system. When the electricity arrives at
Where the changeability and unpredictability of the source the location where it is required, a substation transformer
(such as wind) means the generators cannot be scheduled changes the high voltage electricity to low voltage for
to operate in the same way as conventional generators distribution to end users.
(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.














132 / 110 KV LINE

66 KV LINE 2

DC 2


3 2




2 2


2 2 2
2 2

VICTORIA 2 2 2 3
2 2



2 2 2 2


2 2 2

2 2

2 2 2

2 2 2
3 4 2
2 2 3
2 2 4




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
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

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

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 producer’s
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 retailer’s 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

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

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 retailer’s 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.

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

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 state-
owned grid company Statnett. In Sweden, the TSO is the state-owned grid company
Svenska Kraftnät. 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 down-
regulation orders, e.g. consumers being able to reduce consumption. All regulating
power orders submitted to the TSO’s 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.
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

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

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).

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

Figure 5: The retailer’s 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 retailer’s 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 retailer’s total trading and the
retailer’s 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
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

during this hour). Normally, the down-regulating price will be lower than
the market price.

In a different case, assume the retailer’s 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 retailer’s 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 up-
regulation, 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 Spot’s Day-ahead Auction Market

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

Elspot is Nord Pool Spot’s 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

- 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.
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

- 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

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..
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

At noon, Nord Pool Spot’s 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

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 TSO’s 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.

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

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.

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

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.

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

Available transmission capacity 0 MW exchange 50 MW exchange

Bidding area [NOK/MWh] [NOK/MWh]

Norway 200 233.33
Sweden 300 283.33

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.

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

Figure 12: Market coupling between the Nordic, German and Central Western European exchange

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

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 TSO’s.

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.

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

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

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 contract’s
so-called “delivery period” is a specific month (for instance, it may be June 2014).

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

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 contract’s 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.
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

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 other’s 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

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
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

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 Spot’s 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 market’s 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 year’s 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.

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

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.


Alex Wong
T +65 6302 2557

Amy Lee
T +65 6302 2558

Ming Hui Chock

Senior Associate
T +65 6302 2560

Adrian Wong
Senior Associate
T +65 6302 2568

This note is written as a general guide only. It should not be

relied upon as a substitute for specific legal advice

Introduction  Recent developments in the Singapore energy industry 1

Schedule 1  Summary of Singapore electricity market deregulation and wholesale market operations 4

Schedule 2  Summary of Singapore gas market deregulation and the Gas Network Code 15
1 Singapore energy market

Introduction – Recent developments in the Singapore energy industry

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.


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 Singapore also continues to encourage energy efficiency as it
generation as they reposition themselves in a more moves towards a greener future. The most direct
competitive wholesale market for power generation. Today, manifestations of this include the continued steady growth of
more than 80% of Singapore’s electricity is fuelled by piped the National Environment Agency’s waste to energy plants
natural gas. 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
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
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.

Full retail contestability has still not been achieved so household and small
Policy on Self-Supply of Electricity – Information Paper, Energy Market consumers retain the Market Support Services Licensee as their default
Authority of Singapore (“EMA”), 21 April 2008 power retailer.
2 Singapore energy market

Introduction – Recent developments in the Singapore energy industry


Increased competition in the generation industry however and East Africa and shale gas in North America), LNG prices
almost immediately translates into a battle for the cheapest are set to fall in the medium to long term in spite of strong
fuel source for power generation. For a large part of the demand from China and Japan. At the same time, the West
foreseeable past, this has been piped natural gas from Natuna and South Sumatra gas fields in Indonesia appear to
Indonesia and Malaysia. This dominance (which not only be depleting at a higher rate than originally envisaged and
fuels power generation but also major industries in the this, coupled with pressure on the Indonesia government to
petrochemicals, electronics and biomedical sectors) has been reserve more of its precious resources for domestic
deliberately challenged in Singapore in the recent past by the consumption, would likely mean a far less reliable future for
importance the Singapore government places in piped natural gas imports.
diversification of fuel sources and by the global fall in LNG
prices (particularly long-term prices). It is Singapore’s long- This has led to BG’s 3 Mtpa franchise for aggregating LNG for
term strategy for fuel mix diversification for power generation Singapore having a strong uptake. With 2.65 Mtpa of LNG
and industry that will dominate this discussion in the coming having been taken up by February 2012, the EMA expects the
months. entire franchise to be fulfilled by 2013 at the latest. This begs
the immediate question of Singapore’s LNG import policy
LNG AND ITS IMPACT ON THE SINGAPORE GAS beyond BG’s franchise. The EMA has sought feedback
MARKET through a consultation exercise that will aim to determine
whether Singapore’s future LNG import framework will be
Singapore’s current gas import consists almost exclusively of carried out through a regulated sole importer framework
piped natural gas imported through four pipelines from South (“BG+1”) or a multiple aggregator framework (“BG+3”).
Sumatra and West Natuna gas fields in Indonesia and the gas
imports from Malaysia. Under the BG+1 framework, the EMA will appoint a Regulated
Sole Importer (“RSI”) who will import all incremental LNG
beyond BG’s 3 Mtpa supply. The EMA will regulate the RSI’s
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 BG’s 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
Significant focus has been placed by the Singapore selected by EMA (either sequentially over time or
government in reducing the reliance on these piped natural concurrently) and awarded LNG import licences through a
gas imports and increasing gas imports from other sources. competitive RFP process. Each importer would be awarded a
This led to the Singapore government moving ahead with the franchise to import a specific amount of LNG. Under the
development of its first LNG import and regasification terminal market-driven aggregation model, competition between
and the appointment of BG Asia Pacific Pte Limited (“BG”) as players in the LNG import sector would give rise to natural
Singapore’s first LNG aggregator. aggregation of demand into a few dominant players (e.g. the
formation of a few buyer groups). EMA would set entry
At the time this decision was made, piped natural gas into criteria that importers must fulfil to qualify for access to the
Singapore was significantly less expensive than the LNG terminal.
comparable LNG price but the decision to proceed was
nonetheless made on strategic grounds and ignoring the With the LNG terminal able to support 7 LNG storage tanks
immediate economics of the decision. Singapore had to and up to 15Mtpa of LNG demand, the Singapore government
create an ʻartificialʼ demand for LNG in order to wean the is also open to seeing the LNG terminal being used as a
market off its reliance on piped natural gas. This ʻartificialʼ trans-shipment and trading hub. As of today, a number of
market was created by simultaneously requiring power international LNG trading companies (such as Shell, GDF
generators to accept LNG as part of their fuel portfolio as well Suez, ConocoPhillips and BP) have set up LNG trading
as imposing a moratorium on future piped natural gas imports. offices in Singapore and they will be looking to tap the strong
regional growth of spot and short-term LNG contracts. There
In hindsight, the strategy appears to have been prescient. should be no doubt though that the government sees the LNG
With new sources of gas flooding the global market (both
traditional gas fields in Australia (which is set to overtake 3
LNG Procurement Framework – Consultation Paper, EMA, 30 March 2012
Qatar as the largest LNG exporter by the end of the decade)
3 Singapore energy market

Introduction – Recent developments in the Singapore energy industry


terminal’s first priority as that of helping to ensure security of selection regime, any electricity import selection will likely be
fuel supply for Singapore. subject to a similar tender process.

With decisions on the LNG import strategy likely to be made In summary, the Singapore energy market is best described
in the coming months, the EMA will likely also have to resolve as a ʻmanaged privatised’ market. Singapore is adamant not
the wider question on the long term fuel mix strategy and the to fall into the cycle of underinvestment and high energy
on-going question of how LNG imports will affect the import of prices that befalls many other privatised energy markets
piped natural gas. Given the long term investments and around the world but yet wants to take full advantage of
decisions required with respect to new piped natural gas deregulated market practices. The strategy appears to have
import contracts, the EMA is aware that it will not be able to worked so far but its continued success will depend on the
make sharp changes in direction once its long term policy and ability of the EMA and the Singapore government as a whole
strategy have been settled. Amongst other things, the EMA to balance the supply and demand pressures of the merchant
will have to make challenging decisions on issues such as market with the strategic priorities of the city state.
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 EMA’s 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.


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
government’s 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
ʻsyngasʼ facility to provide fuel for power generation is
nonetheless subject to EMAʼs 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-for-
differences pricing structure. Similar to the aggregator
4 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations4


The electricity and piped gas industries in Singapore have Electricity Act
traditionally been vertically integrated and government-owned.
The Public Utilities Board (“PUB”) was formed in May 1963 to The Electricity Act is the principal legislation governing the
undertake the supply of water, electricity and piped gas to the electricity industry and the NEMS. The principal rights and
population of Singapore. obligations of the participants in the wholesale and retail
electricity markets are set out in the Singapore Electricity
In 1995, the Government began to implement a number of Market Rules (the “Market Rules”), the electricity licences
changes to deregulate the electricity industry. On 1 October and in the codes of practice (the “Codes of Practice”) issued
1995, the PUB transferred its electricity and gas activities to by the EMA.
Temasek Holdings. Within Temasek Holdings, Singapore
The Singapore Wholesale Market Rules
Power was created as the holding company for several other
new companies including the generation companies, The Market Rules are effectively contracts between each
PowerSenoko (now known as Senoko Power) and market participant and the EMC under section 49 of the
PowerSeraya; the transmission company, PowerGrid; and SP Electricity Act. This ensures that market participants have the
Services Ltd, the electricity supply and utilities support option to take legal action against the EMC for damages
services company. A further generator, Tuas Power, was set sustained as a result of the non-observance of the Market
up as an independent company directly under Temasek Rules by the EMC and vice versa. The Market Rules also
Holdings. contain dispute resolution procedures.
THE NATIONAL ELECTRICITY MARKET OF SINGAPORE The objectives of the Market Rules are:
 to establish and govern efficient, competitive and reliable
On 1 April 2001, the Government established a body
markets for the wholesale selling and buying of electricity
corporate, the EMA, under the Ministry of Trade and Industry
and ancillary services in Singapore;
(“MTI”), to regulate, among others, the electricity industry. In
that same year, PowerGrid transferred its system operator  to provide market participants and the Market Support
function to the Power System Operator (“PSO”), and the 5
Services Licensee (the “MSSL”) with non-discriminatory
market operator function and pooling and settlement
access to the transmission system; and
responsibilities to the Energy Market Company Pte Ltd
(“EMC”), which was formed as a subsidiary of the EMA to  to facilitate competition in the generation of electricity.
operate the Singapore Electricity Pool (“SEP”) and
subsequently the wholesale electricity market in the NEMS. Electricity Licenses
The NEMS consists of a wholesale electricity market and a Under the Electricity Act, an entity may not engage in certain
retail electricity market. The wholesale market consists of two electricity-related activities unless it has been issued with an
markets: the “real-time” (or the “spot market”) for energy, electricity licence by the EMA (or it has been exempted from
reserve and regulation; and the “procurement market” for holding one). The electricity-related activities that require an
other ancillary services. The sale of energy, reserve and electricity licence are:
regulation are done through price/offer quantities submitted by
generation companies every half hour.  operation of any wholesale electricity market;

In addition, as part of the government’s policy of separating  generation of electricity;

ownership of electricity generation assets from ownership of
the Transmission and Distribution Systems, Singapore Power  transmission of electricity;
divested its ownership interests in Senoko Power and
PowerSeraya to Temasek Holdings. Temasek Holdings  provision of market support services (such as meter
subsequently divested Tuas Power, Senoko Power and reading and meter data management);
PowerSeraya to private sector investors.
 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
EMA – Introduction to the National Electricity Market of Singapore (version subject to most of the same obligations as market participants are under
6) – October 2010. the Market Rules.
5 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

 trading in the wholesale electricity market; and conditions, the Market Rules and Codes of Practice. The
table in page 6 sets out the various market agreements and
 importing or exporting electricity. contracts in the NEMS.

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

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

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
6 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

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 MSSL, generation licensees This agreement provides for meter reading services for
Agreement and direct market participants wholesale settlement. The generation licensees pay directly for
(“DMP”) the services with fees set by the MSSL.

PSO – Market Participant PSO, generation licensees, This agreement establishes a contractual relationship between
Agreement RELs and DMP 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 This agreement gives effect to the obligations that must exist
licensees, party wanting DMP between SP PowerAssets and the party wanting connection
and electricity consumers service.

Retailer Use of System SP PowerAssets and RELs This agreement is for the collection of Use of System charges
Agreement (“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 This agreement is a contract between the EMC and a market
and the generation licensees participant (usually a generation licensee) supplying ancillary

Source: EMA
7 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

ELECTRICITY INDUSTRY STRUCTURE  facilitate the planning and augmentation of the

Transmission System;
The EMA was established in April 2001 pursuant to the EMA  provide information and other services to facilitate
Act as an independent regulator overseeing the electricity and decisions for investment and the use of resources in the
gas industries in Singapore. Under section 3 of the EMA Act, electricity industry; and
the EMA is charged with the general administration of the
EMA Act, and its functions and duties include:  exercise and perform the functions, powers and duties
assigned to the EMC under the EMA Act, its electricity
 to perform the interests of consumers with regard to licence, the Market Rules and applicable Codes of
prices, reliability and quality of services; Practice.

 to perform the functions of economic and technical The EMC is a 51:49 joint venture between the EMA and M-Co
regulator; (The Marketplace Company) Pte Limited (“M-Co Singapore”).
M-Co Singapore is a related company of The Marketplace
 to ensure that electricity licensees provide an efficient Company Limited, which developed, implemented and
service; currently operates the wholesale electricity market in New
 to ensure security of supply of electricity to consumers
and to arrange for the secure operation of the PSO
transmission system; The role of the PSO (a division of EMA) is to ensure the
security of supply of electricity to consumers and to arrange
 to protect the public from dangers arising from electricity-
for the secure operation of the electricity system.
related activities;
The functions of the PSO include:
 to create an economic and regulatory framework for the
electricity sector that promotes competitive, fair and  maintaining the reliability of the electricity system;
efficient market conduct and prevents the misuse of
monopoly or market power;  forecasting and reporting on conditions on the
Transmission System;
 to advise the Government on matters relating to the
electricity system; and
 coordinating the outages of generation facilities;
 In fulfilling these functions, the EMA has at its disposal a
number of regulatory tools and powers. These include the  providing Transmission System status and load
authority to issue, suspend, revoke or modify an forecasting to the EMC for the purposes of market
electricity licence; the power to issue and modify codes of clearing;
practice and other standards of performance; the power
to issue directions to electricity licensees; the power to  coordinating the actions of the EMC and market
fine electricity licensees; and the authority to investigate participants during emergencies; and
and sanction anti-competitive conduct.
 dispatching generation facilities.
Market Participant
The EMC is licensed to operate the wholesale electricity
market in the NEMS. The EMC’s functions are to: A Market Participant in the NEMS is defined as a person (that
is, an entity or organisation, as well as people) that:
 operate and administer the wholesale electricity market in
the NEMS;  has an electricity licence issued by the EMA; and

 prepare schedules for generation facilities, loads (that is,  has been registered with the EMC as a market
the withdrawal of electricity from the Transmission participant.
System) and the Transmission System;
The wholesale electricity market is a mandatory market in the
 settle accounts of market participants; 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:
8 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

 the Transmission Licensee; RELs

The retail electricity market does not come under the
 generation licensees; jurisdiction of the Market Rules and the EMC. It is created
and regulated under the EMA Act, the electricity licenses and
 RELs; the Codes of Practice issued by the EMA.
 persons, other than the generation licensees and RELs, RELs may be market participants who purchase electricity
who have been licensed to trade in the wholesale directly from the wholesale electricity market or purchase
electricity market; and through the MSSL. Since the RELs are permitted to trade in
electricity and are not subject to the same degree of
 any department of the Government that generates regulation as the MSSL, they may offer contestable
electricity before 1 April 2001. consumers contracts different from those available from the
MSSL. RELs can bundle energy and other charges into a
A MSSL is not a market participant. single invoice, charge a price other than the Uniform
Singapore Energy Price (the “USEP”) for energy, and offer
In the NEMS, it is mandatory for all generation facilities above
additional services to consumers.
to be licensed by the EMA. It is also mandatory for
generation facilities above 10MW to be registered for dispatch Consumers
by the PSO. Mandatory registration ensures that all
generation facilities of any significant size are subject to the Consumers are classified as either contestable or non-
Market Rules. contestable, depending on their electricity usage.
Contestable consumers are entitled to purchase electricity
Transmission Licensee from a REL, or directly from the wholesale electricity market,
or indirectly through MSSLs. Non-contestable consumers are
SP PowerAssets is currently the sole Transmission Licensee
supplied by MSSLs.
in Singapore.
Currently, consumers with a monthly usage of 10,000kWh
The responsibilities of SP PowerAssets and of the persons
and above are contestable. The EMA continues to study
whose facilities are connected to the Transmission System
when full contestability of all retail consumers will be allowed.
are set out in the Transmission Code and the connection
agreements. The Market Rules also contain specific Relationships between the Market Participants
provisions for SP PowerAssets and the PSO’s obligations in
respect of the reliability and security of the Transmission The figure on page 9 shows the financial flows between the
System. 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

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.

Please refer to page 13 for further discussions on the USEP.
9 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

Financial Flows between the Market Participants in the NEMS

An important role for the EMA is the monitoring of market In the NEMS, the real-time dispatch of electricity (scheduling
behaviour to guard against the concentration of market power, generators to supply energy, reserve and regulation) is
abuse of dominance and exercise of anticompetitive determined by the operation of a wholesale spot market run
behaviours such as collusion and capacity withdrawal to drive every half-hour. Generators offer their capacity (specifying
up prices. The EMA is also empowered to prohibit anti- price/quantity pairs) to the market and the PSO provides a
competitive practices including restrictions on mergers and prediction of the expected load along with any system
acquisitions (that is, there will be cross ownership limitations). constraints for that half-hour. The market then determines the
least-cost dispatch quantities and the corresponding market
Significantly, the anti-competition provisions in the Electricity clearing prices based on the offers made by generators.
Act have retrospective effect. Where the EMA finds that a
person has acted in an anti-competitive manner, the EMA can The wholesale electricity market consists of two markets:
give a direction for the cessation of such conduct and impose
a financial penalty on the person.  the spot market for energy, reserve and regulation; and

The electricity industry is excluded from the Competition Act  the procurement market for other ancillary services.
2004 as the EMA will continue to regulate anti-competitive
practices and other competition issues in the electricity Every half-hour, the Market Clearing Engine (“MCE”), a linear
industry. programming computer model, determines the spot market
outcomes for:

EMA – Introduction to the National Electricity Market of Singapore (version
6) – October 2010.
10 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

 the dispatch quantity produced by each generation electricity. There is no overall cheaper dispatch available, in
facility; terms of the offers that have been made to the market, by
providing energy, reserve and regulation from different
 the reserve and regulation capacity required to be sources, or in different quantities from the same sources.
maintained by each generation facility; and This is the minimum cost market dispatch.

 the corresponding wholesale spot market prices for The supply of energy, reserve and regulation is specified by
energy, reserve and regulation. means of offers that contain price/quantity tranches indicating
the quantity of energy, reserve or regulation that each
Quantities and prices are based on price/quantity offers made dispatchable generation facility is willing to supply at the
by the generators and load forecasts prepared by the EMC corresponding energy, reserve or regulation prices.
based on demand forecast information received from the
PSO. With the dispatch based on estimates of the demand for the
coming dispatch period, the market prices are similarly the
RESERVE AND REGULATION prices set according to those estimates. This form of price
setting, called ex ante pricing (pricing before the event), gives
Reserve is unused capacity that has to be made available to the market participants certainty about market prices even if
the electricity system quickly to correct any imbalance and 8
dispatch quantities differ from those scheduled .
maintain reliable supply in case of an unexpected outage of a
scheduled generation facility. This capacity must be able to The MCE does not produce a single market energy price
be in production within a short timeframe, depending on the because of the effects of losses and congestion on the
arrangement. There are three reserve classes: primary transmission system. Different energy prices apply to
reserve (eight seconds’ response), secondary reserve (30 different nodes on the transmission system.
seconds’ response) and contingency reserve (10 minutes’

Regulation, or “load-following”, is a normal operational Energy, Reserve and Regulation Offers

requirement to cover second-to-second variations in load Generators make offers to supply energy, reserve, and
away from estimated load. regulation for each of their units in each half-hourly dispatch
period in which they want to operate. They are similarly
DISPATCH SCHEDULING permitted to offer interruptible load to supply reserve. Offers
Dispatch scheduling is the process of matching the generation can vary for each half-hour, and are assumed to stand, unless
capacity needed to meet forecast demand. It is at the heart of modified, from the time they are made through to dispatch.
running an electricity system. To enable the MCE to generate The market does not distinguish between offers used for the
the dispatch schedule, the PSO and the generation facilities’ market outlook, pre-dispatch and real-time processes. It
dispatch coordinators need to know in advance when each simply uses the most recent offer made for each half-hour.
generation facility will be operating and how much output is
expected from each. The dispatch schedule comes from the Key features of the generator offer process are:
 standing offers are required. Generators are required to
The PSO instructs the generation facilities to conform to the make standing offers into the market. The standing
dispatch schedule. Any deviations from the estimated load offers form a pattern for a week. The use of standing
and corresponding schedule are handled by the PSO using offers is particularly valuable for smaller generators, since
ancillary services. it eases the administrative burden of participating in the
 continuous adjustment of offers. Market participants are
Every half-hour, the MCE is run to determine the dispatch allowed to continually adjust their offers up to gate
schedule and the associated energy market prices for the 9
closure ;
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 8

For all but plants providing regulation, energy actually injected should not
instructions that minimises the cost of supplying load at all differ greatly from scheduled energy, under normal circumstances.
nodes (injection or exit points) of the Transmission System, as 9
Although offers modified within the last hour may be subject to scrutiny
well as meeting the reserve and regulation requirements for from the market surveillance panel.
11 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

 up to 10 price/energy quantity bands. Generators may offs between the commodities so that a facility will not be
make energy offers consisting of up to 10 price/quantity scheduled to produce more energy, reserve and
bands (tranches) for each facility for each half-hour; regulation than it can simultaneously manage; and

 up to five price/reserve and regulation quantity bands.  offers at a node. Energy offers for each generation
Generators and interruptible load may make reserve facility are made at the node where that facility is located.
offers (of different classes) and generators may make
regulation offers if they are registered to do so; Market Clearing Process in the Spot Market
Prices in S$ per megawatt hour (“MWh”)
 combined offers. Energy, reserve and regulation are all
offered simultaneously, and are co-optimised by the
market clearing model. The model respects the trade-
12 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

MARKET CLEARING PROCESS (600MW thermal units, for example) are large relative to the
total load.
The figure on page 11 shows a simplified example of the
market clearing process in the spot market. A generator would wish to receive payment for the reserve
and regulation it provides because it forgoes the opportunity
In this example, there are three generators (A, B and C) of being dispatched fully, by being partially available for PSO
whose offers consist of four price/quantity tranches each. The to call on it for reserve and regulation, as and when required.
tranches are arranged in ascending price order. The market-
clearing price is found at the point where total demand from Reserve in Singapore can be provided by generation facilities
consumers is met by the offer tranches. In this example, the 10
and load . For a facility to provide reserve as quickly as in
third tranche from Company C sets the market-clearing price eight seconds or even in 30 seconds, it needs to be already
with its offer price of S$85/MWh. The total demand is a spinning and synchronised. In most instances, that requires
forecast of the load for that period. the unit also to be supplying energy, with reserve capability
coming from its ability to ramp up its scheduled output very
Offers below the Market Clearing Price are accepted and quickly. Since not all plants have this technical capability, a
those generation facilities are dispatched, in full, to the offers. plant has to be certified as meeting the requirements for
Offers above the Market Clearing Price are not accepted, and registration to provide reserve before it can be offered in the
so the generation capacity represented by those offers is not reserve market.
taken at all. At the margin, the offer that sets the price is
usually only partially dispatched. The generation facility at the There are also different reserve provider groups for each
margin is called the marginal unit (the “Marginal Unit”). class of reserve. These groups represent the reliability of
different reserve sources in providing reserve, and their
effectiveness in curtailing falls in system frequency.
The NEMS is a self-commitment market. This means that unit
commitment is the responsibility of each generation company, Since a facility’s capacity may be available for both energy
and no start-up or shutdown payments are made (generation and reserve/regulation, the MCE must consider the optimal
companies are expected to factor these payments in to their trade-off between the offers for reserve, regulation and
offering strategy). This fact is important because some energy. In solving the markets for each class of reserve and
generation units require a significant period of time to warm regulation, the MCE simultaneously finds the lowest cost
up before they can produce electricity and hence need to be solution (in terms of the offers made) that trades off between
committed some time in advance. The PSO needs to know these products for the various facilities. Within the MCE,
and account for the ability of the generation unit to ramp up or optimisation of the supply of energy must account for the
down. This information is part of the standing capability data minimum running level of facilities that provide reserve. The
required from each generation unit. overall optimal solution may result in a unit being run “out of
merit” for energy so that the unit is available for reserve .
Offers for reserve from a generator can only be made in
Energy Price association with a corresponding offer for energy. Part of the
standing capability date for the plant is a function relating its
In common with many modern electricity markets, the NEMS
reserve capability to its energy capability. This relationship is
uses a form of energy pricing referred to as nodal pricing,
entered into the MCE.
meaning that prices at each node in the network will be
influenced by the physical properties and constraints of the The cost of regulation is recovered from consumers and
transmission system. This results in the price of energy generation facilities. The cost is allocated on a S$ per MW
differing at different physical locations on the network. basis across all MW of consumption in a dispatch period plus
the first 10MW of electricity dispatched by each generation
The MCE automatically produces a different price at each
facility in that dispatch period.
node on the network. Dispatchable generators are paid the
nodal price at their point of injection.

Reserve and Regulation Price 10

An Interruptible Load scheme was introduced into the Singapore market in
Reserve is generation capacity that is required in case of an 2004.
unexpected outage of scheduled plant. Because generating 11
Since often a facility must be running in order to be available for reserve, it
units may fail without warning, some reserve capacity has to may be dispatched for energy even though its energy offer is higher than
be made available to the system to correct any imbalance 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
quickly and maintain reliable supply. Reserve is a significant reserve price received by such a plant will compensate it for the shortfall
factor in the Singapore system since some generating units between its energy offer price and the energy spot price.
13 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

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.

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

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 Singapore energy market

Schedule 1 – Summary of Singapore electricity market deregulation and

wholesale market operations (cont’d)

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
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 EMC’s
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 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.


The Energy Supply Theme of UKERC

UKERC‟s 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 UK‟s 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.

1. SUMMARY .......................................................................................................... 3


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


4.1 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


5.1 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 UK‟s 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. Tomorrow‟s 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

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
Night Peak Low wind

Price l (low wind


CCGT & Coal

Price h (high wind)
Wind & nuclear


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.

1Renewable 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.

2Conventional 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.

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

3 Notable examples of “energy only” markets are Australia (NEM), ERCPT, Nordpool & Ontario

4New 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

5 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 UK‟s 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

6 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 UK‟s 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 high-
carbon 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

7Balancing 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.

approach involving incentives, payments or restrictions in investing in conventional

Increasing centralisation

Option A Option B Option C Option D Option E

Greater carbon Support low Regulate to limit Separate low Single buyer
price certainty carbon in the high carbon carbon market agency
alone current market generation

Minimum carbon Additional Regulate to drive Long term Single agency is

price guarantee at incentives for low decarbonisation payments to low the only purchaser
currently expected carbon generation carbon generators of electricity
level price above carbon to provide revenue generation –
price Competitive certainty existing and new
Competitive market framework Conventional plant low and high
market framework Competitive as today trades in carbon- and only
as today market framework competitive seller of this on to
as today market framework suppliers

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 DECC‟s 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-

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

both security of supply and investment in low carbon technologies, albeit at the cost of
some complexity in design. 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; 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 capacity–based 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 carbon-
intensity 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).

8For example, Southwest & Southwest Power Pool covering most of the Southern & South-western
states except California and Texas.

9Independent System Operator (ISO). A not for profit entity, charged with the operation of the
electricity network and who may also administer the electricity market.

13 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

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).

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 co-
exist 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. 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 non-
delivery 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.

11LOLP 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.

12 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

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

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 low-
carbon 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

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 EMA‟s 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


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.

13Gateclosure – 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

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.

15Operational Trading Reports are available at


settlement “production account”16 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

16For 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.

17The GBSO contracts for Short Term Operational Reserve (STORR) via auctions held three times a

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

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.

Aggregating wind output over a wide geographic area significantly reduces wind output forecast error
together with associated reserve and capacity requirements. See for example

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

19EU 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.

sequestrated generation is likely to have higher high marginal costs than non-
sequestrated 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 19770‟s & 1980‟s. 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

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.

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

20The margin of generation capacity over demand is expected to rise from historic levels of around
24% to nearer 90% by 2020

21Congestion 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.

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 “cheviot”22 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.

22The “Cheviot” boundary is that which cuts the four transmission circuits connecting Scotland with
England, currently having a capacity of around 2.4 GVA.

23 http://www.opsi.gov.uk/acts/acts2010/pdf/ukpga_20100027_en.pdf

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.






1 2 3 4 5 6 7 8 9 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.

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

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, over-
investing due to inappropriate market signals or design rules would impose unnecessary
costs on customers and could ultimately undermine the case for connecting renewable

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.

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.

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.

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 Ofgem‟s „Embedding Financeability in a New Regulation"


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.

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

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


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.

Chapter 3

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 Jäger-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 Jürgen Schmid (Germany)

This chapter should be cited as:

Arvizu, D., P. Balaya, L. Cabeza, T. Hollands, A. Jäger-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. Schlömer, C. von Stechow (eds)],
Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

Chapter 3 Direct Solar Energy

the primary source of information, but their accuracy is inherently lower 3.2.4 Possible impact of climate change on resource
than that of a well-maintained and calibrated ground measurement. potential
Therefore, satellite radiation products require validation with accurate
ground-based measurements (e.g., the Baseline Surface Radiation Climate change due to an increase of greenhouse gases (GHGs) in the
Network). Presently, the solar irradiance at the Earth’s surface is esti- atmosphere may influence atmospheric water vapour content, cloud
mated with an accuracy of about 15 W/m2 on a regional scale (ISCCP cover, rainfall and turbidity, and this can impact the resource potential
Data Products, 2006). The Satellite Application Facility on Climate of solar energy in different regions of the globe. Changes in major cli-
Monitoring project, under the leadership of the German Meteorological mate variables, including cloud cover and solar irradiance at the Earth’s
Service and in partnership with the Finnish, Belgian, Dutch, Swedish and surface, have been evaluated using climate models and considering
Swiss National Meteorological Services, has developed methodologies anthropogenic forcing for the 21st century (Meehl et al., 2007; Meleshko
for irradiance data from satellite measurements. 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
Various international and national institutions provide information the globe, and it varies from model to model. Currently, there is no other
on the solar resource, including the World Radiation Data Centre evidence indicating a substantial impact of global warming on regional
(Russia), the National Renewable Energy Laboratory (USA), the National solar resources. Although some research on global dimming and global
Aeronautics and Space Administration (NASA, USA), the Brasilian brightening indicates a probable impact on irradiance, no current evi-
Spatial Institute (Brazil), the German Aerospace Center (Germany), the dence is available. Uncertainty in pattern changes seems to be rather
Bureau of Meteorology Research Centre (Australia), and the Centro de large, even for large-scale areas of the Earth.
Investigaciones Energéticas, Medioambientales y Tecnológicas (Spain),
National Meteorological Services, and certain commercial companies.
Table 3.2 gives references to some international and national projects 3.3 Technology and applications
that are collecting, processing and archiving information on solar irradi-
ance resources at the Earth’s surface and subsequently distributing it in This section discusses technical issues for a range of solar technologies,
easily accessible formats with understandable quality metrics. 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 Earth’s 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 Radiation Data Centre, Saint Petersburg, Russian
world. Federation (wrdc.mgo.rssi.ru)

National Solar Radiation Database that includes 1,454 ground locations for 1991 to 2005. The satellite-modelled solar National Renewable Energy Laboratory, USA (www.nrel.gov)
data for 1998 to 2005 provided on 10-km grid. The hourly values of solar data can be used to determine solar resources for

European Solar Radiation Database that includes measured solar radiation complemented with other meteorological data Supported by Commission of the European Communities,
necessary for solar engineering. Satellite images from METEOSAT help in improving accuracy in spatial interpolation. Test National Weather Services and scientific institutions of the
Reference Years were also included. European countries

The Solar Radiation Atlas of Africa contains information on surface radiation over Europe, Asia Minor and Africa. Data Supported by the Commission of the European Communities
covering 1985 to 1986 were derived from measurements by METEOSAT 2.

The solar data set for Africa based on images from METEOSAT processed with the Heliosat-2 method covers the period 1985 Ecole des Mines de Paris, France
to 2004 and is supplemented with ground-based solar irradiance.

Typical Meteorological Year (Test Reference Year) data sets of hourly values of solar radiation and meteorological parameters National Renewable Energy Laboratory, USA.
derived from individual weather observations in long-term (up to 30 years) data sets to establish a typical year of hourly data. National Climatic Data Center, National Oceanic and
Used by designers of heating and cooling systems and large-scale solar thermal power plants. 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 International Energy Agency (IEA) Solar Heating and Cooling
solar radiation resources at Earth’s surface in easily accessible formats with understandable quality metrics. The task focuses Programme (SHC). (swera.unep.net)
on development, validation and access to solar resource information derived from surface- and satellite-based platforms.

Solar and Wind Energy Resource Assessment (SWERA) project aimed at developing information tools to simulate RE Global Environment Facility-sponsored project. United Nations
development. SWERA provides easy access to high-quality RE resource information and data for users. Covered major areas Environment Programme (swera.unep.net)
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.

Direct Solar Energy Chapter 3

active heating and cooling, PV electricity generation, CSP electricity release, and maintaining satisfactory thermal comfort conditions by
generation and solar fuel production. Each section also describes appli- limiting the maximum rise in operative (effective) room temperature
cations of these technologies. (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
3.3.1 Passive solar and daylighting technologies the cavity of the wall system and the room. However, this system has
not gained much acceptance because it limits views to the outdoor
Passive solar energy technologies absorb solar energy, store and dis- environment through the fenestration. Hybrid thermal storage with
tribute it in a natural manner (e.g., natural ventilation), without using active charging and passive heat release can also be employed in
mechanical elements (e.g., fans) (Hernandez Gonzalvez, 1996). The term part of a solar building while direct-gain mass is also used (see, e.g.,
‘passive solar building’ is a qualitative term describing a building that the EcoTerra demonstration house (Figure 3.2, left panel), which
makes significant use of solar gain to reduce heating energy consump- uses solar-heated air from a building-integrated photovoltaic/ther-
tion based on the natural energy flows of radiation, conduction and mal system to heat a ventilated concrete slab). Isolated thermal
convection. The term ‘passive building’ is often employed to emphasize storage passively coupled to a fenestration system or solarium/sun-
use of passive energy flows in both heating and cooling, including redis- space is another option in passive design.
tribution of absorbed direct solar gains and night cooling (Athienitis and
Santamouris, 2002). • Well-insulated opaque envelope appropriate for the climatic condi-
tions can be used to reduce heat transfer to and from the outdoor
Daylighting technologies are primarily passive, including windows, sky- environment. In most climates, this energy efficiency aspect must be
lights and shading and reflecting devices. A worldwide trend, particularly integrated with the passive design. A solar technology that may be
in technologically advanced regions, is for an increased mix of passive used with opaque envelopes is transparent insulation (Hollands et
and active systems, such as a forced-air system that redistributes pas- al., 2001) combined with thermal mass to store solar gains in a wall,
sive solar gains in a solar house or automatically controlled shades that turning it into an energy-positive element.
optimize daylight utilization in an office building (Tzempelikos et al.,
2010). • Daylighting technologies and advanced solar control systems, such
as automatically controlled shading (internal, external) and fixed
The basic elements of passive solar design are windows, conservatories shading devices, are particularly suited for daylighting applica-
and other glazed spaces (for solar gain and daylighting), thermal mass, tions in the workplace (Figure 3.2, right panel). These technologies
protection elements, and reflectors (Ralegaonkar and Gupta, 2010). With include electrochromic and thermochromic coatings and newer
the combination of these basic elements, different systems are obtained: technologies such as transparent photovoltaics, which, in addition
direct-gain systems (e.g., the use of windows in combination with walls to a passive daylight transmission function, also generate electric-
able to store energy, solar chimneys, and wind catchers), indirect-gain ity. Daylighting is a combination of energy conservation and passive
systems (e.g., Trombe walls), mixed-gain systems (a combination of solar design. It aims to make the most of the natural daylight that
direct-gain and indirect-gain systems, such as conservatories, sunspaces is available. Traditional techniques include: shallow-plan design,
and greenhouses), and isolated-gain systems. Passive technologies are allowing daylight to penetrate all rooms and corridors; light wells in
integrated with the building and may include the following components: the centre of buildings; roof lights; tall windows, which allow light
to penetrate deep inside rooms; task lighting directly over the work-
• Windows with high solar transmittance and a high thermal resis- place, rather than lighting the whole building interior; and deep
tance facing towards the Equator as nearly as possible can be windows that reveal and light room surfaces to cut the risk of glare
employed to maximize the amount of direct solar gains into the liv- (Everett, 1996).
ing space while reducing heat losses through the windows in the
heating season and heat gains in the cooling season. Skylights are • Solariums, also called sunspaces, are a particular case of the direct-
also often used for daylighting in office buildings and in solaria/ gain passive solar system, but with most surfaces transparent, that
sunspaces. is, made up of fenestration. Solariums are becoming increasingly
attractive both as a retrofit option for existing houses and as an
• Building-integrated thermal storage, commonly referred to as ther- integral part of new buildings (Athienitis and Santamouris, 2002).
mal mass, may be sensible thermal storage using concrete or brick The major driving force for this growth is the development of new
materials, or latent thermal storage using phase-change materials advanced energy-efficient glazing.
(Mehling and Cabeza, 2008). The most common type of thermal stor-
age is the direct-gain system in which thermal mass is adequately Some basic rules for optimizing the use of passive solar heating in build-
distributed in the living space, absorbing the direct solar gains. ings are the following: buildings should be well insulated to reduce
Storage is particularly important because it performs two essential overall heat losses; they should have a responsive, efficient heating sys-
functions: storing much of the absorbed direct solar energy for slow tem; they should face towards the Equator, that is, the glazing should

Chapter 3 Direct Solar Energy

BIPV/T Exhaust
Roof Fan q
Variable External
Speed Fan Rolling Tilted
Shutter Slats
Air Geothermal
Inlet Pump

Passive Light
Slab Shelf

Ventilated 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 office interior space (Athienitis, 2008).

be concentrated on the equatorial side, as should the main living rooms, of existing homes as well. Many homes also add a solarium during
with rooms such as bathrooms on the opposite side; they should avoid retrofit. The new glazing technologies and solar control systems allow
shading by other buildings to benefit from the essential mid-winter sun; the design of a larger window area than in the recent past.
and they should be ‘thermally massive’ to avoid overheating in the sum-
mer and on certain sunny days in winter (Everett, 1996). 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
Clearly, passive technologies cannot be separated from the building itself. for mechanical cooling may often be eliminated by designing for pas-
Thus, when estimating the contribution of passive solar gains, the follow- sive cooling. Passive cooling techniques are based on the use of heat
ing must be distinguished: 1) buildings specifically designed to harness and solar protection techniques, heat storage in thermal mass and heat
direct solar gains using passive systems, defined here as solar buildings, dissipation techniques. The specific contribution of passive solar and
and 2) buildings that harness solar gains through near-equatorial facing energy conservation techniques depends strongly on the climate (UNEP,
windows; this orientation is more by chance than by design. Few reliable 2007). Solar-gain control is particularly important during the ‘shoul-
statistics are available on the adoption of passive design in residential der’ seasons when some heating may be required. In adopting larger
buildings. Furthermore, the contribution of passive solar gains is miss- window areas—enabled by their high thermal resistance—active solar-
ing in existing national statistics. Passive solar is reducing the demand gain control becomes important in solar buildings for both thermal and
and is not part of the supply chain, which is what is considered by the visual considerations.
energy statistics.
The potential of passive solar cooling in reducing CO2 emissions
The passive solar design process itself is in a period of rapid change, has been shown recently (Cabeza et al., 2010; Castell et al., 2010).
driven by the new technologies becoming affordable, such as the recently Experimental work demonstrates that adequate insulation can reduce
available highly efficient fenestration at the same prices as ordinary glaz- by up to 50% the cooling energy demand of a building during the hot
ing. For example, in Canada, double-glazed low-emissivity argon-filled season. Moreover, including phase-change materials in the already-
windows are presently the main glazing technology used; but until a insulated building envelope can reduce the cooling energy demand in
few years ago, this glazing was about 20 to 40% more expensive than such buildings further by up to 15%—about 1 to 1.5 kg/yr/m2 of CO2
regular double glazing. These windows are now being used in retrofits emissions would be saved in these buildings due to reducing the energy

Direct Solar Energy Chapter 3

consumption compared to the insulated building without phase-change including: low-energy house, high-performance house, passive house
material. (‘Passivhaus’), zero-carbon house, zero-energy house, energy-savings
house, energy-positive house and 3-litre house. Concepts that take into
Passive solar system applications are mainly of the direct-gain type, account more parameters than energy demand again use special terms
but they can be further subdivided into the following main application such as eco-building or green building.
categories: multi-story residential buildings and two-story detached or
semi-detached solar homes (see Figure 3.2, left panel), designed to have Another IEA Annex—Energy Conservation through Energy Storage
a large equatorial-facing façade to provide the potential for a large solar Implementing Agreement (ECES IA) Annex 23—was initiated in
capture area (Athienitis, 2008). Perimeter zones and their fenestration November 2009 (IEA ECES, 2004). The general objective of the Annex is
systems in office buildings are designed primarily based on daylighting to ensure that energy storage techniques are properly applied in ultra-
performance. In this application, the emphasis is usually on reducing low-energy buildings and communities. The proper application of energy
cooling loads, but passive heat gains may be desirable as well during storage is expected to increase the likelihood of sustainable building
the heating season (see Figure 3.2, right panel, for a schematic of shad- technologies.
ing devices).
Another passive solar application is natural drying. Grains and many
In addition, residential or commercial buildings may be designed to use other agricultural products have to be dried before being stored so that
natural or hybrid ventilation systems and techniques for cooling or fresh insects and fungi do not render them unusable. Examples include wheat,
air supply, in conjunction with designs for using daylight throughout rice, coffee, copra (coconut flesh), certain fruits and timber (Twidell and
the year and direct solar gains during the heating season. These build- Weir, 2006). Solar energy dryers vary mainly as to the use of the solar
ings may profit from low summer night temperatures by using night heat and the arrangement of their major components. Solar dryers
hybrid ventilation techniques that utilize both mechanical and natural constructed from wood, metal and glass sheets have been evaluated
ventilation processes (Santamouris and Asimakopoulos, 1996; Voss et extensively and used quite widely to dry a full range of tropical crops
al., 2007). (Imre, 2007).

In 2010, passive technologies played a prominent role in the design

of net-zero-energy solar homes—homes that produce as much elec- 3.3.2 Active solar heating and cooling
trical and thermal energy as they consume in an average year. These
houses are primarily demonstration projects in several countries cur- Active solar heating and cooling technologies use the Sun and mechani-
rently collaborating in the International Energy Agency (IEA) Task 40 of cal elements to provide either heating or cooling; various technologies
the Solar Heating and Cooling (SHC) Programme (IEA, 2009b)—Energy are discussed here, as well as thermal storage.
Conservation in Buildings and Community Systems Annex 52—which
focuses on net-zero-energy solar buildings. Passive technologies are
essential in developing affordable net-zero-energy homes. Passive solar Solar heating
gains in homes based on the Passive House Standard are expected to
reduce the heating load by about 40%. By extension, systematic pas- In a solar heating system, the solar collector transforms solar irra-
sive solar design of highly insulated buildings at a community scale, diance into heat and uses a carrier fluid (e.g., water, air) to transfer
with optimal orientation and form of housing, should easily result in that heat to a well-insulated storage tank, where it can be used when
a similar energy saving of 40%. In Europe, according to the Energy needed. The two most important factors in choosing the correct type
Performance of Buildings Directive recast, Directive 2010/31/EC (The of collector are the following: 1) the service to be provided by the
European Parliament and the Council of the European Union, 2010), all solar collector, and 2) the related desired range of temperature of the
new buildings must be nearly zero-energy buildings by 31 December heat-carrier fluid. An uncovered absorber, also known as an unglazed
2020, while EU member states should set intermediate targets for 2015. collector, is likely to be limited to low-temperature heat production
New buildings occupied and owned by public authorities have to be (Duffie and Beckman, 2006).
nearly zero-energy buildings after 31 December 2018. The nearly zero
or very low amount of energy required should to a very significant level A solar collector can incorporate many different materials and be man-
be covered by RE sources, including onsite energy production using ufactured using a variety of techniques. Its design is influenced by the
combined heat and power generation or district heating and cooling, to system in which it will operate and by the climatic conditions of the
satisfy most of their demand. Measures should also be taken to stimu- installation location.
late building refurbishments into nearly zero-energy buildings.
Flat-plate collectors are the most widely used solar thermal collectors
Low-energy buildings are known under different names. A survey car- for residential solar water- and space-heating systems. They are also
ried out by Concerted Action Energy Performance of Buildings (EPBD) used in air-heating systems. A typical flat-plate collector consists of an
identified 17 different terms to describe such buildings across Europe, 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, Unglazed Solar Collectors
only a single plate is used as an absorber (Figure 3.3b). Flat-plate col-
lectors demonstrate a good price/performance ratio, as well as a broad Tube-on-Sheet Collector
range of mounting possibilities (e.g., on the roof, in the roof itself, or

Flat Plate Collectors Single or

Glazing Pump
Glazing Frame
Metal Deck


Box Serpentine Plastic

Pipe Collector
Flow Passages

Absorber Insulation
Plate Backing

Figure 3.3a | Schematic diagram of thermal solar collectors: Glazed flat-plate. 1 1/2’’ ABS Pipe

Pump Flow
Evacuated-tube collectors are usually made of parallel rows of trans-
parent glass tubes, in which the absorbers are enclosed, connected to
a header pipe (Figure 3.3c). To reduce heat loss within the frame by
Figure 3.3b | Schematic diagram of thermal solar collectors: Unglazed tube-on-sheet
convection, the air is pumped out of the collector tubes to generate
and serpentine plastic pipe.
a vacuum. This makes it possible to achieve high temperatures, useful

Evacuated-Tube Collectors
ed W

Solar fer Copper
Heat Sleeve in
fe r
Heat Aluminium
r Ris es to Header
Vapo Casing
etur ns to
Liquid R
Cond Copper

Vacuum Indicator Evacuated Glass Tube Evacuated Heat Pipe

Figure 3.3c | Schematic diagram of thermal solar collectors: Evacuated-tube collectors.

Direct Solar Energy Chapter 3

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

A Close-Coupled Collector Solar
Solar Water Energy
Solar Energy

To Taps

Tank Controller


Arrows Show
Direction of Water Water
Flow Through Copper Feed
Pipes When the Sun
Heats the Collector Panels.

Figure 3.4 | Generic schematics of thermal solar systems. Left: Passive (thermosyphon). Right: Active system.

Chapter 3 Direct Solar Energy

Depending on the size of the combisystem installed, the annual space include the heat recovery units, heat exchangers and humidifiers. Liquid
heating contribution can range from 10 to 60% or more in ultra-low sorption techniques have been demonstrated successfully.
energy Passivhaus-type buildings, and even up to 100% where a large
seasonal thermal store or concentrating solar thermal heat is used. Thermal storage Solar cooling Thermal storage within thermal solar systems is a key component to
ensure reliability and efficiency. Four main types of thermal energy stor-
Solar cooling can be broadly categorized into solar electric refrigera- age technologies can be distinguished: sensible, latent, sorption and
tion, solar thermal refrigeration, and solar thermal air-conditioning. thermochemical heat storage (Hadorn, 2005; Paksoy, 2007; Mehling and
In the first category, the solar electric compression refrigeration uses Cabeza, 2008; Dincer and Rosen, 2010).
PV panels to power a conventional refrigeration machine (Fong et al.,
2010). In the second category, the refrigeration effect can be produced Sensible heat storage systems use the heat capacity of a material. The
through solar thermal gain; solar mechanical compression refrigeration, vast majority of systems on the market use water for heat storage. Water
solar absorption refrigeration, and solar adsorption refrigeration are the heat storage covers a broad range of capacities, from several hundred
three common options. In the third category, the conditioned air can be litres to tens of thousands of cubic metres.
directly provided through the solar thermal gain by means of desiccant
cooling. Both solid and liquid sorbents are available, such as silica gel Latent heat storage systems store thermal energy during the phase
and lithium chloride, respectively. change, either melting or evaporation, of a material. Depending on the
temperature range, this type of storage is more compact than heat stor-
Solar electrical air-conditioning, powered by PV panels, is of minor inter- age in water. Melting processes have energy densities of the order of
est from a systems perspective, unless there is an off-grid application 100 kWh/m3 (360 MJ/m3), compared to 25 kWh/m3 (90 MJ/m3) for sen-
(Henning, 2007). This is because in industrialized countries, which have sible heat storage. Most of the current latent heat storage technologies
a well-developed electricity grid, the maximum use of photovoltaics is for low temperatures store heat in building structures to improve ther-
achieved by feeding the produced electricity into the public grid. mal performance, or in cold storage systems. For medium-temperature
storage, the storage materials are nitrate salts. Pilot storage units in the
Solar thermal air-conditioning consists of solar heat powering an absorp- 100-kW range currently operate using solar-produced steam.
tion chiller and it can be used in buildings (Henning, 2007). Deploying
such a technology depends heavily on the industrial deployment of low- Sorption heat storage systems store heat in materials using water
cost small-power absorption chillers. This technology is being studied vapour taken up by a sorption material. The material can either be a solid
within the IEA Task 25 on solar-assisted air-conditioning of buildings, (adsorption) or a liquid (absorption). These technologies are still largely
SHC program and IEA Task 38 on solar air-conditioning and refrigera- in the development phase, but some are on the market. In principle,
tion, SHC program. sorption heat storage densities can be more than four times higher than
sensible heat storage in water.
Closed heat-driven cooling systems using these cycles have been known
for many years and are usually used for large capacities of 100 kW Thermochemical heat storage systems store heat in an endothermic
and greater. The physical principle used in most systems is based on chemical reaction. Some chemicals store heat 20 times more densely
the sorption phenomenon. Two technologies are established to produce than water (at a ΔT≈100°C); but more typically, the storage densities
thermally driven low- and medium-temperature refrigeration: absorp- are 8 to 10 times higher. Few thermochemical storage systems have
tion and adsorption. been demonstrated. The materials currently being studied are the salts
that can exist in anhydrous and hydrated form. Thermochemical systems
Open cooling cycle (or desiccant cooling) systems are mainly of interest can compactly store low- and medium-temperature heat. Thermal stor-
for the air conditioning of buildings. They can use solid or liquid sorp- age is discussed with specific reference to higher-temperature CSP in
tion. The central component of any open solar-assisted cooling system Section 3.3.4.
is the dehumidification unit. In most systems using solid sorption, this
unit is a desiccant wheel. Various sorption materials can be used, such Underground thermal energy storage is used for seasonal storage and
as silica gel or lithium chloride. All other system components are found includes the various technologies described below. The most frequently
in standard air-conditioning applications with an air-handling unit and used storage technology that makes use of the underground is aquifer

Direct Solar Energy Chapter 3

thermal energy storage. This technology uses a natural underground layer medium-temperature heat and are often necessary in areas with high
(e.g., sand, sandstone or chalk) as a storage medium for the temporary solar irradiance and high energy costs.
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 modi- Some process heat applications can be met with temperatures deliv-
fied temperature level at a separate location nearby. Most applications ered by ‘ordinary’ low-temperature collectors, namely, from 30°C to
are for the storage of winter cold to be used for the cooling of large 80°C. However, the bulk of the demand for industrial process heat
office buildings and industrial processes. Aquifer cold storage is gain- requires temperatures from 80°C to 250°C.
ing interest because savings on electricity bills for chillers are about
75%, and in many cases, the payback time for additional investments Process heat collectors are another potential application for solar
is shorter than five years. A major condition for the application of this thermal heat collectors. Typically, these systems require a large capac-
technology is the availability of a suitable geologic formation. ity (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 Active solar heating and cooling applications very early stage of development and no products are available on an
industrial scale. In addition to ‘concentrating’ collectors, improved flat
For active solar heating and cooling applications, the amount of hot collectors with double and triple glazing are currently being devel-
water produced depends on the type and size of the system, amount of oped, which could meet needs for process heat in the range of up
sun available at the site, seasonal hot-water demand pattern, and instal- to 120°C. Concentrating-type solar collectors are described in Section
lation characteristics of the system (Norton, 2001). 3.3.4.

Solar heating for industrial processes is at a very early stage of develop- Solar refrigeration is used, for example, to cool stored vaccines. The
ment in 2010 (POSHIP, 2001). Worldwide, less than 100 operating solar need for such systems is greatest in peripheral health centres in rural
thermal systems for process heat are reported, with a total capacity of communities in the developing world, where no electrical grid is
about 24 MWth (34,000 m² collector area). Most systems are at an exper- available.
imental stage and relatively small scale. However, significant potential
exists for market and technological developments, because 28% of the Solar cooling is a specific area of application for solar thermal tech-
overall energy demand in the EU27 countries originates in the industrial nology. High-efficiency flat plates, evacuated tubes or parabolic
sector, and much of this demand is for heat below 250°C. Education and troughs can be used to drive absorption cycles to provide cooling. For
knowledge dissemination are needed to deploy this technology. a greater coefficient of performance (COP), collectors with low con-
centration levels can provide the temperatures (up to around 250°C)
In the short term, solar heating for industrial processes will mainly be used needed for double-effect absorption cycles. There is a natural match
for low-temperature processes, ranging from 20°C to 100°C. With tech- between solar energy and the need for cooling.
nological development, an increasing number of medium-temperature
applications—up to 250°C—will become feasible within the market. A number of closed heat-driven cooling systems have been built,
According to Werner (2006), about 30% of the total industrial heat using solar thermal energy as the main source of heat. These systems
demand is required at temperatures below 100°C, which could theoreti- often have large cooling capacities of up to several hundred kW. Since
cally be met with solar heating using current technologies. About 57% the early 2000s, a number of systems have been developed in the
of this demand is required at temperatures below 400°C, which could small-capacity range, below 100 kW, and, in particular, below 20 kW
largely be supplied by solar in the foreseeable future. and down to 4.5 kW. These small systems are single-effect machines
of different types, used mainly for residential buildings and small com-
In several specific industry sectors—such as food, wine and beverages, mercial applications.
transport equipment, machinery, textiles, and pulp and paper—the
share of heat demand at low and medium temperatures (below 250°C) Although open-cooling cycles are generally used for air conditioning
is around 60% (POSHIP, 2001). Tapping into this low- and medium- in buildings, closed heat-driven cooling cycles can be used for both air
temperature heat demand with solar heat could provide a significant conditioning and industrial refrigeration.
opportunity for solar contribution to industrial energy requirements. A
substantial opportunity for solar thermal systems also exists in chemi- Other solar applications are listed below. The production of potable
cal industries and in washing processes. water using solar energy has been readily adopted in remote or
isolated regions (Narayan et al., 2010). Solar stills are widely used
Among the industrial processes, desalination and water treatment in some parts of the world (e.g., Puerto Rico) to supply water to
(e.g., sterilization) are particularly promising applications for solar households of up to 10 people (Khanna et al., 2008). In appropriate
thermal energy, because these processes require large amounts of isolation conditions, solar detoxification can be an effective low-cost

Chapter 3 Direct Solar Energy

treatment for low-contaminant waste (Gumy et al., 2006). Multiple-

effect humidification (MEH) desalination units indirectly use heat
from highly efficient solar thermal collectors to induce evaporation Anti-Reflection Coating

and condensation inside a thermally isolated, steam-tight container.

n-Type Semiconductor
These MEH systems are now beginning to appear in the market. Also
see the report on water desalination by CSP (DLR, 2007) and the dis-
Front Contact
cussion 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 flow 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
Electron (-) Hole (+)
in developing countries (Lahkar and Samdarshi, 2010) though might still
be considered an early stage commercial product due to limited overall Recombination

deployment in comparison to other cooking methods. A solar cooker

uses sunlight as its energy source, so no fuel is needed and operating p-Type Semiconductor +
Back Contact
costs are zero. Also, a reliable solar cooker can be constructed easily and
quickly from common materials.
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 multicrystalline silicon wafer PV (including ribbon technologies) are the
the photovoltaic effect. Light shining on a semiconductor such as sili- dominant technologies on the PV market, with a 2009 market share
con (Si) generates electron-hole pairs that are separated spatially by an of about 80%; thin-film PV (primarily CdTe and thin-film Si) has the
internal electric field created by introducing special impurities into the remaining 20% share. Organic PV (OPV) consists of organic absorber
semiconductor on either side of an interface known as a p-n junction. materials and is an emerging class of solar cells.
This creates negative charges on one side of the interface and positive
charges are on the other side (Figure 3.5). This resulting charge separa- Wafer-based silicon technology includes solar cells made of monocrys-
tion creates a voltage. When the two sides of the illuminated cell are talline or multicrystalline wafers with a current thickness of around 200
connected to a load, current flows from one side of the device via the μm, while the thickness is decreasing down to 150 μm. Single-junction
load to the other side of the cell. The conversion efficiency of a solar cell wafer-based c-Si cells have been independently verified to have record
is defined as a ratio of output power from the solar cell with unit area energy conversion efficiencies of 25.0% for monocrystalline silicon
(W/cm2) to the incident solar irradiance. The maximum potential effi- cells and 20.3% for multicrystalline cells (Green et al., 2010b) under
ciency of a solar cell depends on the absorber material properties and standard test conditions (i.e., irradiance of 1,000 W/m2, air-mass 1.5,
device design. One technique for increasing solar cell efficiency is with a 25°C). The theoretical Shockley-Queisser limit of a single-junction cell
multijunction approach that stacks specially selected absorber materials with an energy bandgap of crystalline silicon is 31% energy conversion
that can collect more of the solar spectrum since each different material efficiency (Shockley and Queisser, 1961).
can collect solar photons of different wavelengths.
Several variations of wafer-based c-Si PV for higher efficiency have
PV cells consist of organic or inorganic matter. Inorganic cells are based been developed, for example, heterojunction solar cells and interdigi-
on silicon or non-silicon materials; they are classified as wafer-based cells tated back-contact (IBC) solar cells. Heterojunction solar cells consist
or thin-film cells. Wafer-based silicon is divided into two different types: of a crystalline silicon wafer base sandwiched by very thin (~5 nm)
monocrystalline and multicrystalline (sometimes called ‘polycrystalline’). amorphous silicon layers for passivation and emitter. The highest-effi-
ciency heterojunction solar cell is 23.0% for a 100.4-cm2 cell (Taguchi
et al., 2009). Another advantage is a lower temperature coefficient. The Existing photovoltaic technologies efficiency of conventional c-Si solar cells declines with elevating ambi-
ent temperature at a rate of -0.45%/°C, while the heterojunction cells
Existing PV technologies include wafer-based crystalline silicon (c-Si) show a lower rate of -0.25%/°C (Taguchi et al., 2009). An IBC solar
cells, as well as thin-film cells based on copper indium/gallium disul- cell, where both the base and emitter are contacted at the back of the
fide/diselenide (CuInGaSe2; CIGS), cadmium telluride (CdTe), and cell, has the advantage of no shading of the front of the cell by a top
thin-film silicon (amorphous and microcrystalline silicon). Mono- and electrode. The highest efficiency of such a back-contact silicon wafer

Direct Solar Energy Chapter 3

cell is 24.2% for 155.1 cm2 (Bunea et al., 2010). Commercial module efficiency of 20.1% (Green et al., 2010b). Due to higher efficiencies and
efficiencies for wafer-based silicon PV range from 12 to 14% for multi- lower manufacturing energy consumptions, CIGS cells are currently in
crystalline Si and from 14 to 20% for monocrystalline Si. the industrialization phase, with best commercial module efficiencies
of up to 13.1% (Kushiya, 2009) for CuInGaSe2 and 8.6% for CuInS2
Commercial thin-film PV technologies include a range of absorber (Meeder et al., 2007). Although it is acknowledged that the scarcity of
material systems: amorphous silicon (a-Si), amorphous silicon-germa- In might be an issue, Wadia et al. (2009) found that the current known
nium, microcrystalline silicon, CdTe and CIGS. These thin-film cells have economic indium reserves would allow the installation of more than 10
an absorber layer thickness of a few μm or less and are deposited on TW of CIGS-based PV systems.
glass, metal or plastic substrates with areas of up to 5.7 m2 (Stein et al.,
2009). High-efficiency solar cells based on a multijunction technology using
III-V semiconductors (i.e., based on elements from the III and V columns
The a-Si solar cell, introduced in 1976 (Carlson and Wronski, 1976) with of the periodic chart), for example, gallium arsenide (GaAs) and gallium
initial efficiencies of 1 to 2%, has been the first commercially successful indium phosphide (GaInP) , can have superior efficiencies. These cells
thin-film PV technology. Because a-Si has a higher light absorption coef- were originally developed for space use and are already commercial-
ficient than c-Si, the thickness of an a-Si cell can be less than 1 μm—that ized. An economically feasible terrestrial application is the use of these
is, more than 100 times thinner than a c-Si cell. Developing higher effi- cells in concentrating PV (CPV) systems, where concentrating optics are
ciencies for a-Si cells has been limited by inherent material quality and used to focus sunlight onto high efficiency solar cells (Bosi and Pelosi,
by light-induced degradation identified as the Staebler-Wronski effect 2007). The most commonly used cell is a triple-junction device based on
(Staebler and Wronski, 1977). However, research efforts have success- GaInP/GaAs/germanium (Ge), with a record efficiency of 41.6% for a
fully lowered the impact of the Staebler-Wronski effect to around 10% lattice-matched cell (Green et al., 2010b) and 41.1% for a metamorphic
or less by controlling the microstructure of the film. The highest stabi- or lattice-mismatched device (Bett et al., 2009). Sub-module efficien-
lized efficiency—the efficiency after the light-induced degradation—is cies have reached 36.1% (Green et al., 2010b). Another advantage of
reported as 10.1% (Benagli et al., 2009). the concentrator system is that cell efficiencies increase under higher
irradiance (Bosi and Pelosi, 2007), and the cell area can be decreased in
Higher efficiency has been achieved by using multijunction technologies proportion to the concentration level. Concentrator applications, how-
with alloy materials, e.g., germanium and carbon or with microcrystal- ever, require direct-normal irradiation, and are thus suited for specific
line silicon, to form semiconductors with lower or higher bandgaps, climate conditions with low cloud coverage.
respectively, to cover a wider range of the solar spectrum (Yang and
Guha, 1992; Yamamoto et al., 1994; Meier et al., 1997). Stabilized
efficiencies of 12 to 13% have been measured for various laboratory Emerging photovoltaic technologies
devices (Green et al., 2010b).
Emerging PV technologies are still under development and in laboratory
CdTe solar cells using a heterojunction with cadmium sulphide (CdS) or (pre-) pilot stage, but could become commercially viable within the
have a suitable energy bandgap of 1.45 electron-volt (eV) (0.232 aJ) next decade. They are based on very low-cost materials and/or processes
with a high coefficient of light absorption. The best efficiency of this and include technologies such as dye-sensitized solar cells, organic solar
cell is 16.7% (Green et al., 2010b) and the best commercially available cells and low-cost (printed) versions of existing inorganic thin-film
modules have an efficiency of about 10 to 11%. technologies.

The toxicity of metallic cadmium and the relative scarcity of tellurium Electricity generation by dye-sensitized solar cells (DSSCs) is based on
are issues commonly associated with this technology. Although several light absorption in dye molecules (the ‘sensitizers’) attached to the very
assessments of the risk (Fthenakis and Kim, 2009; Zayed and Philippe, large surface area of a nanoporous oxide semiconductor electrode (usu-
2009) and scarcity (Green et al., 2009; Wadia et al., 2009) are available, ally titanium dioxide), followed by injection of excited electrons from the
no consensus exists on these issues. It has been reported that this poten- dye into the oxide. The dye/oxide interface thus serves as the separator
tial hazard can be mitigated by using a glass-sandwiched module design of negative and positive charges, like the p-n junction in other devices.
and by recycling the entire module and any industrial waste (Sinha et The negatively charged electrons are then transported through the semi-
al., 2008). conductor electrode and reach the counter electrode through the load,
thus generating electricity. The injected electrons from the dye molecules
The CIGS material family is the basis of the highest-efficiency thin-film are replenished by electrons supplied through a liquid electrolyte that
solar cells to date. The copper indium diselenide (CuInSe2)/CdS solar penetrates the pores of the semiconductor electrode, providing the elec-
cell was invented in the early 1970s at AT&T Bell Labs (Wagner et al., trical path from the counter electrode (Graetzel, 2001). State-of-the-art
1974). Incorporating Ga and/or S to produce CuInGa(Se,S)2 results in the DSSCs have achieved a top conversion efficiency of 10.4% (Chiba et
benefit of a widened bandgap depending on the composition (Dimmler al., 2005). Despite the gradual improvements since its discovery in 1991
and Schock, 1996). CIGS-based solar cells have been validated at an (O’Regan and Graetzel, 1991), long-term stability against ultraviolet light

Chapter 3 Direct Solar Energy

irradiation, electrolyte leakage and high ambient temperatures continue Committee, 2001; Navigant Consulting Inc., 2006; EU PV European
to be key issues in commercializing these PV cells. Photovoltaic Technology Platform, 2007; Kroposki et al., 2008; NEDO,
Organic PV (OPV) cells use stacked solid organic semiconductors, either
polymers or small organic molecules. A typical structure of a small- At the component level, BOS components for grid-connected applications
molecule OPV cell consists of a stack of p-type and n-type organic are not yet sufficiently developed to match the lifetime of PV modules.
semiconductors forming a planar heterojunction. The short-lived nature Additionally, BOS component and installation costs need to be reduced.
of the tightly bound electron-hole pairs (excitons) formed upon light Moreover, devices for storing large amounts of electricity (over 1 MWh
absorption limits the thickness of the semiconductor layers that can be or 3,600 MJ) will be adapted to large PV systems in the new energy
used—and therefore, the efficiency of such devices. Note that excitons network. As new module technologies emerge in the future, some of the
need to move to the interface where positive and negative charges can ideas relating to BOS may need to be revised. Furthermore, the quality
be separated before they recombine. If the travel distance is short, the of the system needs to be assured and adequately maintained according
‘active’ thickness of material is small and not all light can be absorbed to defined standards, guidelines and procedures. To ensure system qual-
within that thickness. ity, assessing performance is important, including on-line analysis (e.g.,
early fault detection) and off-line analysis of PV systems. The knowledge
The efficiency achieved with single-junction OPV cells is about 5% (Li et gathered can help to validate software for predicting the energy yield of
al., 2005), although predictions indicate about twice that value or higher future module and system technology designs.
can be achieved (Forrest, 2005; Koster et al., 2006). To decouple exciton
transport distances from optical thickness (light absorption), so-called To increasingly penetrate the energy network, PV systems must use
bulk-heterojunction devices have been developed. In these devices, technology that is compatible with the electric grid and energy supply
the absorption layer is made of a nanoscale mixture of p- and n-type and demand. System designs and operation technologies must also be
materials to allow excitons to reach the interface within their lifetime, developed in response to demand patterns by developing technology to
while also enabling a sufficient macroscopic layer thickness. This bulk- forecast the power generation volume and to optimize the storage func-
heterojunction structure plays a key role in improving the efficiency, to tion. Moreover, inverters must improve the quality of grid electricity by
a record value of 7.9% in 2009 (Green et al., 2010a). The developments controlling reactive power or filtering harmonics with communication in
in cost and processing (Brabec, 2004; Krebs, 2005) of materials have a new energy network that uses a mixture of inexpensive and effective
caused OPV research to advance further. Also, the main development communications systems and technologies, as well as smart meters (see
challenge is to achieve a sufficiently high stability in combination with Section 8.2.1).
a reasonable efficiency. Photovoltaic applications Novel photovoltaic technologies
Photovoltaic applications include PV power systems classified into two
Novel technologies are potentially disruptive (high-risk, high-potential) major types: those not connected to the traditional power grid (i.e., off-grid
approaches based on new materials, devices and conversion concepts. applications) and those that are connected (i.e., grid-connected applica-
Generally, their practically achievable conversion efficiencies and cost tions). In addition, there is a much smaller, but stable, market segment
structure are still unclear. Examples of these approaches include inter- for consumer applications.
mediate-band semiconductors, hot-carrier devices, spectrum converters,
plasmonic solar cells, and various applications of quantum dots (Section Off-grid PV systems have a significant opportunity for economic appli-
3.7.3). The emerging technologies described in the previous section pri- cation in the un-electrified areas of developing countries. Figure 3.6
marily aim at very low cost, while achieving a sufficiently high efficiency shows the ratio of various off-grid and grid-connected systems in the
and stability. However, most of the novel technologies aim at reaching Photovoltaic Power Systems (PVPS) Programme countries. Of the total
very high efficiencies by making better use of the entire solar spectrum capacity installed in these countries during 2009, only about 1.2% was
from infrared to ultraviolet. installed in off-grid systems that now make up 4.2% of the cumulative
installed PV capacity of the IEA PVPS countries (IEA, 2010e). Photovoltaic systems Off-grid centralized PV mini-grid systems have become a reliable alter-
native for village electrification over the last few years. In a PV mini-grid
A photovoltaic system is composed of the PV module, as well as the system, energy allocation is possible. For a village located in an isolated
balance of system (BOS) components, which include an inverter, storage area and with houses not separated by too great a distance, the power
devices, charge controller, system structure, and the energy network. The may flow in the mini-grid without considerable losses. Centralized
system must be reliable, cost effective, attractive and match with the systems for local power supply have different technical advantages con-
electric grid in the future (US Photovoltaic Industry Roadmap Steering cerning electrical performance, reduction of storage needs, availability

Direct Solar Energy Chapter 3
Installed PV Power [MWp]

an existing structure; and the PV array itself can be used as a cladding
or roofing material, as in building-integrated PV (Eiffert, 2002; Ecofys
Cumulative Grid-Connected
Netherlands BV, 2007; Elzinga, 2008).
Cumulative Off-Grid

An often-cited disadvantage is the greater sensitivity to grid intercon-

15,000 nection 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
5,000 have the function of the anti-islanding effect.

Grid-connected centralized PV systems perform the functions of cen-

‘92 ‘93 ‘94 ‘95 ‘96 ‘97 ‘98 ‘99 ‘00 ‘01 ‘02 ‘03 ‘04 ‘05 ‘06 ‘07 ‘08 ‘09 tralized power stations. The power supplied by such a system is not
associated with a particular electricity customer, and the system is
Figure 3.6 | Historical trends in cumulative installed PV power of off-grid and grid-
not located to specifically perform functions on the electricity network
connected systems in the OECD countries (IEA, 2010e). Vertical axis is in peak megawatts.
other than the supply of bulk power. Typically, centralized systems are
mounted on the ground, and they are larger than 1 MW.
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 The economical advantage of these systems is the optimization of instal-
have a diesel generator set as an optional balancing system or operate lation and operating cost by bulk buying and the cost effectiveness of
as a hybrid PV-wind-diesel system. These kinds of systems are relevant the PV components and balance of systems at a large scale. In addition,
for reducing and avoiding diesel generator use in remote areas (Munoz the reliability of centralized PV systems can be greater than distributed
et al., 2007; Sreeraj et al., 2010). PV systems because they can have maintenance systems with monitor-
ing equipment, which can be a smaller part of the total system cost.
Grid-connected PV systems use an inverter to convert electricity from
direct current (DC)—as produced by the PV array—to alternating cur- Multi-functional PV, daylighting and solar thermal components involv-
rent (AC), and then supply the generated electricity to the electricity ing PV or solar thermal that have already been introduced into the built
network. Compared to an off-grid installation, system costs are lower environment include the following: shading systems made from PV
because energy storage is not generally required, since the grid is used and/or solar thermal collectors; hybrid PV/thermal (PV/T) systems that
as a buffer. The annual output yield ranges from 300 to 2,000 kWh/ generate electricity and heat from the same ‘panel/collector’ area; semi-
kW (Clavadetscher and Nordmann, 2007; Gaiddon and Jedliczka, 2007; transparent PV windows that generate electricity and transmit daylight
Kurokawa et al., 2007; Photovoltaic Geographic Information System, from the same surface; façade collectors; PV roofs; thermal energy roof
2008) for several installation conditions in the world. The average annual systems; and solar thermal roof-ridge collectors. Currently, fundamen-
performance ratio—the ratio between average AC system efficiency and tal and applied R&D activities are also underway related to developing
standard DC module efficiency—ranges from 0.7 to 0.8 (Clavadetscher other products, such as transparent solar thermal window collectors, as
and Nordmann, 2007) and gradually increases further to about 0.9 for well as façade elements that consist of vacuum-insulation panels, PV
specific technologies and applications. panels, heat pump, and a heat-recovery system connected to localized
Grid-connected PV systems are classified into two types of applications:
distributed and centralized. Grid-connected distributed PV systems are Solar energy can be integrated within the building envelope and with
installed to provide power to a grid-connected customer or directly to energy conservation methods and smart-building operating strategies.
the electricity network. Such systems may be: 1) on or integrated into Much work over the last decade or so has gone into this integration,
the customer’s premises, often on the demand side of the electricity culminating in the ‘net-zero’ energy building.
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 Much of the early emphasis was on integrating PV systems with thermal
4 kW for residential systems, and 10 kW to several MW for rooftops on and daylighting systems. Bazilian et al. (2001) and Tripanagnostopoulos
public and industrial buildings. (2007) listed methods for doing this and reviewed case studies where
the methods had been applied. For example, PV cells can be laid on
These systems have a number of advantages: distribution losses in the the absorber plate of a flat-plate solar collector. About 6 to 20% of the
electricity network are reduced because the system is installed at the solar energy absorbed on the cells is converted to electricity; the remain-
point of use; extra land is not required for the PV system, and costs ing roughly 80% is available as low-temperature heat to be transferred
for mounting the systems can be reduced if the system is mounted on to the fluid 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 gas, nuclear, oil or biomass—comes from creating a hot fluid. CSP sim-
two conversion devices had been mounted side by side and worked ply provides an alternative heat source. Therefore, an attraction of this
independently. PV cells have also been developed to be applied to win- technology is that it builds on much of the current know-how on power
dows to allow daylighting and passive solar gain. Reviews of recent generation in the world today. And it will benefit not only from ongoing
work in this area are provided by Chow (2010) and Arif Hasan and advances in solar concentrator technology, but also as improvements
Sumathy (2010). continue to be made in steam and gas turbine cycles.

Considerable work has also been done on architecturally integrating the Any concentrating solar system depends on direct-beam irradiation
solar components into the building. Any new solar building should be as opposed to global horizontal irradiation as for flat-plate systems.
very well insulated, well sealed, and have highly efficient windows and Thus, sites must be chosen accordingly, and the best sites for CSP are
heat recovery systems. Probst and Roecker (2007), surveying the opin- in near-equatorial cloud-free regions such as the North African desert.
ions of more than 170 architects and engineers who examined numerous The average capacity factor of a solar plant will depend on the quality
existing solar buildings, concluded the following: 1) best integration is of the solar resource.
achieved when the solar component is integrated as a construction ele-
ment, and 2) appearance—including collector colour, orientation and Some of the key advantages of CSP include the following: 1) it can be
jointing—must sometimes take precedence over performance in the installed in a range of capacities to suit varying applications and condi-
overall design. In describing 16 case studies of building-integrated pho- tions, from tens of kW (dish/Stirling systems) to multiple MWs (tower
tovoltaics, Eiffert and Kiss (2000) identified two main products available and trough systems); 2) it can integrate thermal storage for peaking
on the architectural market: façade systems and roof systems. Façade loads (less than one hour) and intermediate loads (three to six hours);
systems include curtain wall products, spandrel panels and glazings; 3) it has modular and scalable components; and 4) it does not require
roofing products include tiles, shingles, standing-seam products and exotic materials. This section discusses various types of CSP systems and
skylights. These can be integrated as components or constitute the thermal storage for these systems.
entire structure (as in the case of a bus shelter).
Large-scale CSP plants most commonly concentrate sunlight by reflec-
The idea of the net-zero-energy solar building has sparked recent inter- tion, as opposed to refraction with lenses. Concentration is either to a
est. Such buildings send as much excess PV-generated electrical energy line (linear focus) as in trough or linear Fresnel systems or to a point
to the grid as the energy they draw over the year. An IEA Task is consid- (point focus) as in central-receiver or dish systems. The major features of
ering how to achieve this goal (IEA NZEB, 2009). Recent examples for each type of CSP system are illustrated in Figure 3.7 and are described
the Canadian climate are provided by Athienitis (2008). Starting from a below.
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 In trough concentrators, long rows of parabolic reflectors concentrate
heating or as a source for a heat pump. Solar water-heating collectors the solar irradiance by the order of 70 to 100 times onto a heat collec-
are included, as is fenestration permitting a large passive gain through tion element (HCE) mounted along the reflector’s focal line. The troughs
equatorial-facing windows. A key feature is a ground-source heat pump, track the Sun around one axis, with the axis typically being oriented
which provides a small amount of residual heating in the winter and north-south. The HCE comprises a steel inner pipe (coated with a solar-
cooling in the summer. selective surface) and a glass outer tube, with an evacuated space in
between. Heat-transfer oil is circulated through the steel pipe and heated
Smart solar-building control strategies may be used to manage the col- to about 390°C. The hot oil from numerous rows of troughs is passed
lection, storage and distribution of locally produced solar electricity through a heat exchanger to generate steam for a conventional steam
and heat to reduce and shift peak electricity demand from the grid. An turbine generator (Rankine cycle). Land requirements are of the order of
example of a smart solar-building design is given by Candanedo and 2 km2 for a 100-MWe plant, depending on the collector technology and
Athienitis (2010), where predictive control based on weather forecasts assuming no storage. Alternative heat transfer fluids to the synthetic oil
one day ahead and real-time prediction of building response are used to commonly used in trough receivers, such as steam and molten salt, are
optimize energy performance while reducing peak electricity demand. being developed to enable higher temperatures and overall efficiencies,
as well as integrated thermal storage in the case of molten salt.

3.3.4 Concentrating solar power electricity generation Linear Fresnel reflectors use long lines of flat or nearly flat mirrors, which
allow the moving parts to be mounted closer to the ground, thus reduc-
Concentrating solar power (CSP) technologies produce electricity by ing structural costs. (In contrast, large trough reflectors presently use
concentrating direct-beam solar irradiance to heat a liquid, solid or gas thermal bending to achieve the curve required in the glass surface.) The
that is then used in a downstream process for electricity generation. The receiver is a fixed inverted cavity that can have a simpler construction
majority of the world’s electricity today—whether generated by coal, than evacuated tubes and be more flexible in sizing. The attraction of

Direct Solar Energy Chapter 3

(a) (b)

Absorber Tube Curved Mirrors Curved Mirrors

Solar Field Piping

Absorber Tube and


(c) (d)

Central Receiver



Figure 3.7 | Schematic diagrams showing the underlying principles of four basic CSP configurations: (a) parabolic trough, (b) linear Fresnel reflector, (c) central receiver/power tower,
and (d) dish systems (Richter et al., 2009).

linear Fresnel reflectors is that the installed costs on a per square metre temperature is a benefit because higher-temperature thermodynamic
basis can be lower than for trough systems. However, the annual optical cycles used for generating electricity are more efficient. This technology
performance is less than that for a trough. uses an array of mirrors (heliostats), with each mirror tracking the Sun
and reflecting the light onto a fixed receiver atop a tower. Temperatures
Central receivers (or power towers), which are one type of point-focus of more than 1,000°C can be reached. Central receivers can easily gen-
collector, are able to generate much higher temperatures than troughs erate the maximum temperatures of advanced steam turbines, can use
and linear Fresnel reflectors, although requiring two-axis tracking as high-temperature molten salt as the heat transfer fluid, and can be used
the Sun moves through solar azimuth and solar elevation. This higher to power gas turbine (Brayton) cycles.

Chapter 3 Direct Solar Energy

Dish systems include an ideal optical reflector and therefore are suitable not be renewable (unless it is biomass-derived), it provides significant
for applications requiring high temperatures. Dish reflectors are paraboloid operational benefits for the turbine and improves solar yield.
and concentrate the solar irradiation onto a receiver mounted at the
focal point, with the receiver moving with the dish. Dishes have been CSP applications range from small distributed systems of tens of kW to
used to power Stirling engines at 900°C, and also for steam genera- large centralized power stations of hundreds of MW.
tion. There is now significant operational experience with dish/Stirling
engine systems, and commercial rollout is planned. In 2010, the capac- Stirling and Brayton cycle generation in CSP can be installed in a wide
ity of each Stirling engine is small—on the order of 10 to 25 kWelectric. range from small distributed systems to clusters forming medium- to
The largest solar dishes have a 485-m2 aperture and are in research large-capacity power stations. The dish/Stirling technology has been
facilities or demonstration plants. under development for many years, with advances in dish struc-
tures, high-temperature receivers, use of hydrogen as the circulating
In thermal storage, the heat from the solar field is stored prior to working fluid, as well as some experiments with liquid metals and
reaching the turbine. Thermal storage takes the form of sensible or improvements in Stirling engines—all bringing the technology closer
latent heat storage (Gil et al., 2010; Medrano et al., 2010). The solar to commercial deployment. Although the individual unit size may only
field needs to be oversized so that enough heat can be supplied to be of the order of tens of kWe, power stations having a large capacity
both operate the turbine during the day and, in parallel, charge the of up to 800 MWe have been proposed by aggregating many modules.
thermal storage. The term ‘solar multiple’ refers to the total solar field Because each dish represents a stand-alone electricity generator, from
area installed divided by the solar field area needed to operate the tur- the perspective of distributed generation there is great flexibility in
bine at design point without storage. Thermal storage for CSP systems the capacity and rate at which units are installed. However, the dish
needs to be at a temperature higher than that needed for the work- technology is less likely to integrate thermal storage.
ing fluid of the turbine. As such, system temperatures are generally
between 400°C and 600°C, with the lower end for troughs and the An alternative to the Stirling engine is the Brayton cycle, as used by
higher end for towers. Allowable temperatures are also dictated by gas turbines. The attraction of these engines for CSP is that they are
the limits of the media available. Examples of storage media include already in significant production, being used for distributed generation
molten salt (presently comprising separate hot and cold tanks), steam fired with landfill gas or natural gas. In the solarized version, the air is
accumulators (for short-term storage only), solid ceramic particles, instead heated by concentrated solar irradiance from a tower or dish
high-temperature phase-change materials, graphite, and high-tem- reflector. It is also possible to integrate with a biogas or natural gas
perature concrete. The heat can then be drawn from the storage to combustor to back up the solar. Several developments are currently
generate steam for a turbine, as and when needed. Another type of underway based on solar tower and micro-turbine combinations.
storage associated with high-temperature CSP is thermochemical stor-
age, where solar energy is stored chemically. This is discussed more Centralized CSP benefits from the economies of scale offered by large-
fully in Sections 3.3.5 and 3.7.5. scale plants. Based on conventional steam and gas turbine cycles,
much of the technological know-how of large power station design
Thermal energy storage integrated into a system is an important attri- and practice is already in place. However, although larger capacity has
bute of CSP. Until recently, this has been primarily for operational significant cost benefits, it has also tended to be an inhibitor until
purposes, providing 30 minutes to 1 hour of full-load storage. This recently because of the much larger investment commitment required
eases the impact of thermal transients such as clouds on the plant, from investors. In addition, larger power stations require strong infra-
assists start-up and shut-down, and provides benefits to the grid. structural support, and new or augmented transmission capacity may
Trough plants are now designed for 6 to 7.5 hours of storage, which is be needed.
enough to allow operation well into the evening when peak demand
can occur and tariffs are high. Trough plants in Spain are now operat- The earliest commercial CSP plants were the 354 MW of Solar Electric
ing with molten-salt storage. In the USA, Abengoa Solar’s 280-MW Generating Stations in California—deployed between 1985 and
Solana trough project, planned to be operational by 2013, intends 1991—that continue to operate commercially today. As a result of the
to integrate six hours of thermal storage. Towers, with their higher positive experiences and lessons learned from these early plants, the
temperatures, can charge and store molten salt more efficiently. trough systems tend to be the technology most often applied today as
Gemasolar, a 17-MWe solar tower project under construction in Spain, the CSP industry grows. In Spain, regulations to date have mandated
is designed for 15 hours of storage, giving a 75% annual capacity fac- that the largest capacity unit that can be installed is 50 MWe to help
tor (Arce et al., 2011). stimulate industry competition. In the USA, this limitation does not
exist, and proposals are in place for much larger plants—280 MWe in
Thermal storage is a means of providing dispatchability. Hybridization the case of troughs and 400-MWe plants (made up of four modules)
with non-renewable fuels is another way in which CSP can be based on towers. There are presently two operational solar towers of
designed to be dispatchable. Although the back-up fuel itself may 10 and 20 MWe, and all tower developers plan to increase capacity in

Direct Solar Energy Chapter 3

line with technology development, regulations and investment capital. Figure 3.8 illustrates possible pathways to produce H2 or syngas from
Multiple dishes have also been proposed as a source of aggregated water and/or fossil fuels using concentrated solar energy as the source
heat, rather than distributed-generation Stirling or Brayton units. of high-temperature process heat. Feedstocks include inorganic com-
pounds such as water and CO2, and organic sources such as coal,
CSP or PV electricity can also be used to power reverse-osmosis plants biomass and natural gas (NG). See Chapter 2 for parallels with bio-
for desalination. Dedicated CSP desalination cycles based on pres- mass-derived syngas.
sure and temperature are also being developed for desalination (see
Section 3.3.2). 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,
3.3.5 Solar fuel production the overall solar-to-hydrogen energy conversion efficiency ranges
between 10 and 14%, assuming electrolyzers working at 70% effi-
Solar fuel technologies convert solar energy into chemical fuels, which ciency and solar electricity being produced at 15% (PV) and 20%
can be a desirable method of storing and transporting solar energy. They (CSP) annual efficiency. The electricity demand for electrolysis can be
can be used in a much wider variety of higher-efficiency applications significantly reduced if the electrolysis of water proceeds at higher
than just electricity generation cycles. Solar fuels can be processed into temperatures (800° to 1,000°C) via solid-oxide electrolyzer cells
liquid transportation fuels or used directly to generate electricity in (Jensen et al., 2007). In this case, concentrated solar energy can be
fuel cells; they can be employed as fuels for high-efficiency gas-turbine applied to provide both the high-temperature process heat and the
cycles or internal combustion engines; and they can serve for upgrading electricity needed for the high-temperature electrolysis.
fossil fuels, CO2 synthesis, or for producing industrial or domestic heat.
The challenge is to produce large amounts of chemical fuels directly Thermolysis and thermochemical cycles are a long-term sustainable
from sunlight in cost-effective ways and to minimize adverse effects on and carbon-neutral approach for hydrogen production from water. This
the environment (Steinfeld and Meier, 2004). route involves energy-consuming (endothermic) reactions that make
use of concentrated solar irradiance as the energy source for high-
Solar fuels that can be produced include synthesis gas (syngas, i.e., temperature process heat (Abanades et al., 2006). Solar thermolysis
mixed gases of carbon monoxide and hydrogen), pure hydrogen (H2) requires temperatures above 2,200°C and raises difficult challenges
gas, dimethyl ether (DME) and liquids such as methanol and diesel. The for reactor materials and gas separation. Water-splitting thermochemi-
high energy density of H2 (on a mass basis) and clean conversion give it cal cycles allow operation at lower temperature, but require several
attractive properties as a future fuel and it is also used as a feedstock for chemical reaction steps and also raise challenges because of ineffi-
many industrial processes. H2 has a higher energy density than batteries, ciencies associated with heat transfer and product separation at each
although batteries have a higher round-trip efficiency. However, its very step.
low energy density on a volumetric basis poses economic challenges
associated with its storage and transport. It will require significant new Decarbonization of fossil fuels is a near- to mid-term transition path-
distribution infrastructure and either new designs of internal combustion way to solar hydrogen that encompasses the carbothermal reduction
engine or a move to fuel cells. Additionally, the synthesis of hydrogen of metal oxides (Epstein et al., 2008) and the decarbonization of fossil
with CO2 can produce hydrocarbon fuels that are compatible with exist- fuels via solar cracking (Spath and Amos, 2003; Rodat et al., 2009),
ing infrastructures. DME gas is similar to liquefied petroleum gas (LPG) reforming (Möller et al., 2006) and gasification (Z’Graggen and
and easily stored. Methanol is liquid and can replace gasoline without Steinfeld, 2008; Piatkowski et al., 2009). These routes are being pur-
significant changes to the engine or the fuel distribution infrastructure. sued by European, Australian and US academic and industrial research
Methanol and DME can be used for fuel cells after reforming, and DME consortia. Their technical feasibility has been demonstrated in concen-
can also be used in place of LPG. Fischer-Tropsch processes can produce trating solar chemical pilot plants at the power level of 100 to 500
hydrocarbon fuels and electricity (see Sections 2.6 and 8.2.4). kWth. Solar hybrid fuel can be produced by supplying concentrated
solar thermal energy to the endothermic processes of methane and
There are three basic routes, alone or in combination, for producing biomass reforming—that is, solar heat is used for process energy only,
storable and transportable fuels from solar energy: 1) the electrochemi- and fossil fuels are still a required input. Some countries having vast
cal route uses solar electricity from PV or CSP systems followed by an solar and natural gas resources, but a relatively small domestic energy
electrolytic process; 2) the photochemical/photobiological route makes market (e.g., the Middle East and Australia) are in a position to pro-
direct use of solar photon energy for photochemical and photobiological duce and export solar energy in the form of liquid fuels.
processes; and 3) the thermochemical route uses solar heat at moderate
and/or high temperatures followed by an endothermic thermochemical Solar fuel synthesis from solar hydrogen and CO2 produces hydrocar-
process (Steinfeld and Meier, 2004). Note that the electrochemical and bons that are compatible with existing energy infrastructures such as
thermochemical routes apply to any RE technology, not exclusively to the natural gas network or existing fuel supply structures. The renew-
solar technologies. able methane process combines solar hydrogen with CO2 from the

Chapter 3 Direct Solar Energy

Solar Energy

H2O H2O Splitting Decarbonization Fossil Fuels

(NG, Oil, Coal)

Solar Thermochemical Solar Electricity & Solar

Solar Thermolysis Solar Reforming Solar Cracking
Cycle Electrolysis Gasification

Optional CO2/C
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 gasification 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 cata- solar fuel conversion (technical photosynthesis) with an efficiency of
lyst. In this way, a substitute for natural gas is produced that can be 10% (Sterner, 2009) and via solar-driven thermochemical dissociation
stored, transported and used in gas power plants, heating systems of CO2 and H2O using metal oxide redox reactions, yielding a syngas
and gas vehicles (Sterner, 2009). mixture of carbon monoxide (CO) and H2, with a solar-to-fuel effi-
ciency approaching 20% (Chueh et al., 2010). This approach would
Solar methane can be produced using water, air, solar energy and a provide a solution to the issues and controversy surrounding existing
source of CO2. Possible CO2 sources are biomass, industry processes biofuels, although the cost of this technology is a possible constraint.
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 3.4 Global and regional status of market and
permanently. Thus, carbon sink energy systems powered by RE can industry development
be created (Sterner, 2009). The first pilot plants at the kW scale with
atmospheric CO2 absorption have been set up in Germany, proving the This section looks at the five key solar technologies, first focusing on
technical feasibility. Scaling up to the utility MW scale is planned in installed capacity and generated energy, then on industry capacity
the next few years (Specht et al., 2010). and supply chains, and finally on the impact of policies specific to
these technologies.
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 transporta- 3.4.1 Installed capacity and generated energy
tion. The main advantages of these solar fuels are the same range
as fossil fuels (compared to the generally reduced range of electric This subsection discusses the installed capacity and generated energy
vehicles), less competition for land use, and higher per-hectare yields within the five technology areas of passive solar, active solar heating
compared to biofuels. Solar energy can be harvested via natural pho- and cooling, PV electricity generation, CSP electricity generation, and
tosynthesis in biofuels with an efficiency of 0.5%, via PV power and solar fuel production.

Direct Solar Energy Chapter 3

For passive solar technologies, no estimates are available at this time for collectors. In terms of per capita use, Cyprus is the leading country in
the installed capacity of passive solar or the energy generated or saved the world, with an installed capacity of 527 kWth per 1,000 inhabitants.
through this technology.
The type of application of solar thermal energy varies greatly in differ-
For active solar heating, the total installed capacity worldwide was ent countries (Weiss and Mauthner, 2010). In China (88.7 GWth), Europe
about 149 GWth in 2008 and 180 GWth in 2009 (Weiss and Mauthner, (20.9 GWth) and Japan (4.4 GWth), flat-plate and evacuated-tube col-
2010; REN21, 2010). lectors mainly prepare hot water and provide space heating. However,
in the USA and Canada, swimming pool heating is still the dominant
In 2008, new capacity of 29.1 GWth, corresponding to 41.5 million m2 of application, with an installed capacity of 12.9 GWth of unglazed plastic
solar collectors, was installed worldwide (Weiss and Mauthner, 2010). collectors.
In 2008, China accounted for about 79% of the installations of glazed
collectors, followed by the EU with 14.5%. The biggest reported solar thermal system for industrial process heat
was installed in China in 2007. The 9 MWth plant produces heat for a tex-
The overall new installations grew by 34.9% compared to 2007. The tile company. About 150 large-scale plants (>500 m2; 350 kWth)1 with a
growth rate in 2006/2007 was 18.8%. The main reasons for this growth total capacity of 160 MWth are in operation in Europe. The largest plants
were the high growth rates of glazed water collectors in China, Europe for solar-assisted district heating are located in Denmark (13 MWth) and
and the USA. Sweden (7 MWth).

In 2008, the global market had high growth rates for evacuated-tube In Europe, the market size more than tripled between 2002 and 2008.
collectors and flat-plate collectors, compared to 2007. The market for However, even in the leading European solar thermal markets of Austria,
unglazed air collectors also increased significantly, mainly due to the Greece, and Germany, only a minor portion of residential homes use
installation of 23.9 MWth of new systems in Canada. solar thermal. For example, in Germany, only about 5% of one- and two-
family homes are using solar thermal energy.
Compared to 2007, the 2008 installation rates for new unglazed, glazed
flat-plate, and evacuated-tube collectors were significantly up in Jordan, The European market has the largest variety of different solar thermal
Cyprus, Canada, Ireland, Germany, Slovenia, Macedonia (FYROM), applications, including systems for hot-water preparation, plants for
Tunisia, Poland, Belgium and South Africa. space heating of single- and multi-family houses and hotels, large-scale
plants for district heating, and a growing number of systems for air-
New installations in China, the world’s largest market, again increased conditioning, cooling and industrial applications.
significantly in 2008 compared to 2007, reaching 21.7  GWth. After a
market decline in Japan in 2007, the growth rate was once again posi- Advanced applications such as solar cooling and air conditioning
tive in 2008. (Henning, 2004, 2007), industrial applications (POSHIP, 2001) and desal-
ination/water treatment are in the early stages of development. Only a
Market decreases compared to 2007 were reported for Israel, the Slovak few hundred first-generation systems are in operation.
Republic and the Chinese province of Taiwan.
For PV electricity generation, newly installed capacity in 2009 was
The main markets for unglazed water collectors are still found in the about 7.5 GW, with shipments to first point in the market at 7.9 GW
USA (0.8 GWth), Australia (0.4 GWth), and Brazil (0.08 GWth). Notable (Jäger-Waldau, 2010a; Mints, 2010). This addition brought the cumu-
markets are also in Austria, Canada, Mexico, The Netherlands, South lative installed PV capacity worldwide to about 22 GW—a capacity
Africa, Spain, Sweden and Switzerland, with values between 0.07 and able to generate up to 26 TWh (93,600 TJ) per year. More than 90%
0.01 GWth of new installed unglazed water collectors in 2008. 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%)
Comparison of markets in different countries is difficult due to the (Jäger-Waldau, 2010b). These markets are dominated by grid-connected
wide range of designs used for different climates and different demand PV systems, and growth within PV markets has been stimulated by
requirements. In Scandinavia and Germany, a solar heating system various government programmes around the world. Examples of such
will typically be a combined water-heating and space-heating system, programmes include feed-in tariffs in Germany and Spain, and various
known as a solar combisystem, with a collector area of 10 to 20 m2. In mechanisms in the USA, such as buy-down incentives, investment tax
Japan, the number of solar domestic water-heating systems is large, but credits, performance-based incentives and RE quota systems. For 2010,
most installations are simple integral preheating systems. The market in
Israel is large due to a favourable climate, as well as regulations man- 1 To enable comparison, the IEA’s Solar Heating and Cooling Programme, together
with the European Solar Thermal Industry Federation and other major solar thermal
dating installation of solar water heaters. The largest market is in China, trade associations, publish statistics in kWth (kilowatt thermal) and use a factor of
where there is widespread adoption of advanced evacuated-tube solar 0.7 kWth/m2 to convert square metres of collector area into installed thermal capacity

Chapter 3 Direct Solar Energy

the market is estimated between 9 and 24 GW of additional installed 661/2007 has been a major driving force for CSP plant construction and
PV systems, with a consensus value in the 13 GW range (Jäger-Waldau, expansion plans. As of November 2009, 2,340 MWe of CSP projects had
2010a). 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
Figure 3.9 illustrates the cumulative installed capacity for the top eight agreement contracts. The different contracts specify when the projects
PV markets through 2009, including Germany (9,800 MW), Spain (3,500 must start delivering electricity between 2010 and 2015 (Bloem et al.,
MW), Japan (2,630 MW), the USA (1,650 MW), Italy (1,140 MW), Korea 2010). More than 10,000 MWe of new CSP plants have been proposed in
(460 MW), France (370 MW) and the People’s Republic of China (300 the USA. More than 50 CSP electricity projects are currently in the plan-
MW). By far, Spain and Germany have seen the largest amounts of ning phase, mainly in North Africa, Spain and the USA. In Australia, the
growth in installed PV capacity in recent years, with Spain seeing a huge federal government has called for 1,000 MWe of new solar plants, cover-
surge in 2008 and Germany having experienced steady growth over the ing both CSP and PV, under the Solar Flagships programme. Figure 3.10
last five years. shows the current and planned deployment to add more CSP capacity
in the near future.
Cumulative Installed Capacity [MW]

Hybrid solar/fossil plants have received increasing attention in recent
8,000 Germany Italy years, and several integrated solar combined-cycle (ISCC) projects
7,000 Spain Korea have been either commissioned or are under construction in the
6,000 Japan France Mediterranean region and the USA. The first plant in Morocco (Ain
5,000 USA China Beni Mathar: 470  MW total, 22  MW solar) began operating in June
4,000 2010, and two additional plants in Algeria (Hassi R’Mel: 150 MW total,
3,000 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 plant’s 31,000-m2 parabolic trough solar field
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 will be the first to use molten salt as the heat transfer fluid (SolarPACES,
Figure 3.9 | Installed PV capacity in eight markets. Data sources: EurObserv’ER (2009);
IEA (2009c); REN21 (2009); and Jäger-Waldau (2010b).
Solar fuel production technologies are in an earlier stage of develop-
ment. The high-temperature solar reactor technology is typically being
Concentrating photovoltaics (CPV) is an emerging market with about 17 developed at a laboratory scale of 1 to 10 kWth solar power input.
MW of cumulative installed capacity at the end of 2008. The two main
tracks are high-concentration PV (>300 times or 300 suns) and low-
Installed Capacity [MW]

to medium-concentration PV with a concentration factor of 2 to about

300 (2 to ~300 suns). To maximize the benefits of CPV, the technology South Africa
requires high direct-beam irradiance, and these areas have a limited China
geographical range—the ‘Sun Belt’ of the Earth. The market share of Israel

CPV is still small, but an increasing number of companies are focusing Jordan

on CPV. In 2008, about 10 MW of CPV were installed, and market esti- Egypt
mates for 2009 are in the 20 to 30 MW range; for 2010, about 100 MW Algeria

are expected. Morocco

6,000 Abu Dhabi
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
4,000 USA
majority of installed plants use parabolic trough technology. Central-
receiver 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 first major CSP plants
came on line with Nevada Solar One (64 MWe, USA) and PS10 (11 MWe, 1990 2000 2006 2007 2008 2009 2010 2012 2015
Spain). In Spain, successive Royal Decrees have been in place since 2004
Figure 3.10 | Installed and planned concentrated solar power plants by country (Bloem
and have stimulated the CSP industry in that country. Royal Decree et al., 2010).

Direct Solar Energy Chapter 3

Scaling up thermochemical processes for hydrogen production to the different countries has improved the design capabilities (Athienitis and
100-kWth power level is reported for a medium-temperature mixed Santamouris, 2002).
iron oxide cycle (800°C to 1,200°C) (Roeb et al., 2006, 2009) and for
the high-temperature zinc oxide (ZnO) dissociation reaction at above The integration of passive solar systems with the active heating/cool-
1,700°C (Schunk et al., 2008, 2009). Pilot plants in the power range of ing air-conditioning systems both in the design and operation stages
300 to 500 kWth have been built for the carbothermic reduction of ZnO of the building is essential to achieve good comfort conditions while
(Epstein et al., 2008), the steam reforming of methane (Möller et al., saving energy. However, this is often overlooked because of inadequate
2006), and the steam gasification of petcoke (Z’Graggen and Steinfeld, collaboration for integrating building design between architects and
2008). Solar-to-gas has been demonstrated at a 30-kW scale to drive engineers. Thus, the architect often designs the building envelope based
a commercial natural gas vehicle, applying a nickel catalyst (Specht et solely on qualitative passive solar design principles, and the engineer
al., 2010). Demonstration at the MW scale should be warranted before often designs the heating-ventilation-air-conditioning system based
erecting commercial solar chemical plants for fuels production, which on extreme design conditions without factoring in the benefits due to
are expected to be available only after 2020 (Pregger et al., 2009). solar gains and natural cooling. The result may be an oversized system
and inappropriate controls incompatible with the passive system and
Direct conversion of solar energy to fuel is not yet widely demonstrated that can cause overheating and discomfort (Athienitis and Santamouris,
or commercialized. But two options appear commercially feasible in the 2002). Collaboration between the disciplines involved in building design
near to medium term: 1) the solar hybrid fuel production system (includ- is now improving with the adoption of computer tools for integrated
ing solar methane reforming and solar biomass reforming), and 2) solar analysis and design.
PV or CSP electrolysis.
The design of high-mass buildings with significant near-equatorial-facing
Australia’s Commonwealth Scientific and Industrial Research window areas is common in some areas of the world such as Southern
Organisation is running a 250-kWth reactor and plans to build a Europe. However, a systematic approach to designing such buildings is
MW-scale demonstration plant using solar steam-reforming technology, still not widely employed. This is changing with the introduction of the
with an eventual move to CO2 reforming for higher performance and passive house standard in Germany and other countries (PHPP, 2004),
less water usage. With such a system, liquid solar fuels can be produced the deployment of the European Directives, and new national laws such
in sunbelts such as Australia and solar energy shipped on a commercial as China’s standard based on the German one.
basis to Asia and beyond.
Glazing and window technologies have made substantial progress in
Oxygen gas produced by solar (PV or CSP) electrolysis can be used for the last 20 years (Hollands et al., 2001). New-generation windows result
coal gasification and partial oxidation of natural gas. With the combined in low energy losses, high daylight efficiency, solar shading, and noise
process of solar electrolysis and partial oxidation of coal or methane, reduction. New technologies such as transparent PV and electrochromic
theoretically 10 to 15% of solar energy is incorporated into the metha- and thermochromic windows provide many possibilities for designing
nol or DME. Also, the production cost of the solar hybrid fuel can be solar houses and offices with abundant daylight. The change from regu-
lower than the solar hydrogen produced by the solar electrolysis process lar double-glazed to double-glazed low-emissivity argon windows is
only. presently occurring in Canada and is accelerated by the rapid drop in
prices of these windows.

3.4.2 Industry capacity and supply chain The primary materials for low-temperature thermal storage in passive
solar systems are concrete, bricks and water. A review of thermal stor-
This subsection discusses the industry capacity and supply chain within age materials is given by Hadorn (2008) under IEA SHC Task 32, focusing
the five technology areas of passive solar, active solar heating and cool- on a comparison of the different technologies. Phase-change material
ing, PV electricity generation, CSP electricity generation and solar fuel (PCM) thermal storage (Mehling and Cabeza, 2008) is particularly
production. promising in the design, control and load management of solar build-
ings because it reduces the need for structural reinforcement required
In passive solar technologies, people make up part of the industry for heavier traditional sensible storage in concrete-type construction.
capacity and the supply chain: namely, the engineers and architects Recent developments facilitating integration include microencapsulated
who collaborate to produce passively heated buildings. Close collabo- PCM that can be mixed with plaster and applied to interior surfaces
ration between the two disciplines has often been missing in the past, (Schossig et al., 2005). PCM in microencapsulated polymers is now on
but the dissemination of systematic design methodologies issued by the market and can be added to plaster, gypsum or concrete to enhance

Chapter 3 Direct Solar Energy

the thermal capacity of a room. For renovation, this provides a good Figure 3.11 plots the increase in production from 2000 through 2009,
alternative to new heavy walls, which would require additional struc- showing regional contributions (Jäger-Waldau, 2010a). The compound
tural support (Hadorn, 2008). annual growth rate in production from 2003 to 2009 was more than
In spite of the advances in PCM, concrete has certain advantages for
thermal storage when a massive building design approach is used, as 12,000

Annual PV Production [MW]

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 Rest of World
effective than thermal storage without this added function. United States

For active solar heating and cooling, a number of different collector Europe

technologies and system approaches have been developed due to dif- Japan

ferent applications—including domestic hot water, heating, preheating 6,000

and combined systems—and varying climatic conditions.

In some parts of the production process, such as selective coatings,
large-scale industrial production levels have been attained. A number of 2,000
different materials, including copper, aluminium and stainless steel, are
applied and combined with different welding technologies to achieve
a highly efficient heat-exchange process in the collector. The materi- 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

als used for the cover glass are structured or flat, low-iron glass. The Figure 3.11 | Worldwide PV production from 2000 to 2009 (Jäger-Waldau, 2010b).
first antireflection coatings are coming onto the market on an industrial
scale, leading to efficiency improvements of about 5%.
In general, vacuum-tube collectors are well-suited for higher-temperature The announced production capacities—based on a survey of more
applications. The production of vacuum-tube collectors is currently dom- than 300 companies worldwide—increased despite very difficult eco-
inated by the Chinese Dewar tubes, where a metallic heat exchanger is nomic conditions in 2009 (Figure 3.12) (Jäger-Waldau, 2010b). Only
integrated to connect them with the conventional hot-water systems. published announcements from the respective companies, not third-
In addition, some standard vacuum-tube collectors, with metallic heat party information, were used. April 2010 was the cut-off date for the
absorbers, are on the market. information included. This method has the drawback that not all com-
panies announce their capacity increases in advance; also, in times of
The largest exporters of solar water-heating systems are Australia, financial tightening, announcements of scale-backs in expansion plans
Greece and the USA. The majority of exports from Greece are to Cyprus are often delayed to prevent upsetting financial markets. Therefore, the
and the near-Mediterranean area. France also sends a substantial capacity figures provide a trend, but do not represent final numbers.
number of systems to its overseas territories. The majority of US exports
are to the Caribbean region. Australian companies export about 50% In 2008 and 2009, Chinese production capacity increased over-
of production (mainly thermosyphon systems with external horizontal proportionally. In actual production, China surpassed all other countries,
tanks) to most of the areas of the world that do not have hard-freeze
conditions. 70,000
Annual Production/Production Capacity [MW]

PV electricity generation is discussed under the areas of overall solar 60,000
cell production, thin-film module production and polysilicon production.
50,000 South Korea
The development characteristic of the PV sector is much different than USA
the traditional power sector, more closely resembling the semicon- 40,000
ductor market, with annual growth rates between 40 to 50% and a Europe
high learning rate. Therefore, scientific and peer-reviewed papers can Japan
be several years behind the actual market developments due to the 20,000

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. 0
Estimated Planned Planned Planned Planned
Production Capacity Capacity Capacity Capacity
2 Solar cell production capacities mean the following: for wafer-silicon-based solar 2009 2009 2010 2012 2015
cells, only the cells; for thin films, the complete integrated module. Only those com-
panies that actually produce the active circuit (solar cell) are counted; companies Figure 3.12 | Worldwide annual PV production in 2009 compared to the announced
that purchase these circuits and then make modules are not counted. production capacities (Jäger-Waldau, 2010a).

Direct Solar Energy Chapter 3

estimated in 2009 at between 5.4 and 6.1 GW (including 1.5 to 1.7 GW 2010b). The first thin-film factories with GW production capacity are
production in the Chinese province of Taiwan), Europe had 2.0 to 2.2 GW, already under construction for various thin-film technologies.
and was followed by Japan, with 1.5 to 1.7 GW (Jäger-Waldau, 2010b).
In terms of production, First Solar (USA/Germany/France/Malaysia) was The rapid growth of the PV industry since 2000 led to the situation
number one (1,082  MW), followed by Suntech (China) estimated at between 2004 and early 2008 where the demand for polysilicon out-
750 MW and Sharp (Japan) estimated at 580 MW. stripped the supply from the semiconductor industry. This led to a silicon
shortage, which resulted in silicon spot-market prices as high as USD2005
If all these ambitious plans can be realized by 2015, then China will 450/kg (USD2005, assumed 2008 base) in 2008 compared to USD2005 25.5/
have about 51% (including 16% in the Chinese province of Taiwan) of kg in 2003 and consequently higher prices for PV modules. This extreme
the worldwide production capacity of 70 GW, followed by Europe (15%) price hike triggered the massive capacity expansion, not only of estab-
and Japan (13%). lished companies, but of many new entrants as well.

Worldwide, more than 300 companies produce solar cells. In 2009, The six companies that reported shipment figures delivered together
silicon-based solar cells and modules represented about 80% of the about 43,900 tonnes of polysilicon in 2008, as reported by Semiconductor
worldwide market (Figure 3.13). In addition to a massive increase in pro- Equipment and Materials International (SEMI, 2009a). In 2008, these
duction capacities, the current development predicts that thin-film-based companies had a production capacity of 48,200 tonnes of polysili-
solar cells will increase their market share to over 30% by 2012. con (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
70,000 that not all new capacity actually produced polysilicon at nameplate
Production Capacity [MW/yr]

capacity in 2008, it was estimated that 62,000 tonnes of polysilicon
Crystalline Wafer Silicon

Thin Films
could be produced. Subtracting the needs of the semiconductor industry
50,000 and adding recycling and excess production, the available amount of
silicon for the PV industry was estimated at 46,000 tonnes of polysili-
con. With an average material need of 8.7 g/Wp (p = peak), this would
30,000 have been sufficient for the production of 5.3 GW of crystalline silicon
PV cells.

10,000 The drive to reduce costs and secure key markets has led to the emer-
gence of two interesting trends. One is the move to large original design
2006 2009 2010 2012 2015 manufacturing units, similar to the developments in the semiconductor
Figure 3.13 | Actual (2006) and announced (2009 to 2015) production capacities of
industry. A second is that an increasing number of solar manufacturers
thin-film and crystalline silicon-based solar modules (Jäger-Waldau, 2010b). move part of their module production close to the final market to dem-
onstrate the local job creation potential and ensure the current policy
support. This may also be a move to manufacture in low-cost or subsi-
dized markets.
In 2005, production of thin-film PV modules grew to more than 100 MW
per year. Since then, the compound annual growth rate of thin-film PV The regional distribution of polysilicon production capacities is as fol-
module production was higher than that of the industry—thus increas- lows: China 20,000 tonnes, Europe 17,500 tonnes, Japan 12,000 tonnes,
ing the market share of thin-film products from 6% in 2005 to about and USA 37,000 tonnes (Service, 2009).
20% in 2009. Most of this thin-film share comes from the largest PV
company. In 2009, solar-grade silicon production of about 88,000 tonnes was
reported, sufficient for about 11 GW of PV assuming an average materi-
More than 150 companies are involved in the thin-film solar cell produc- als need of 8 g/Wp (Displaybank, 2010). China produced about 18,000
tion process, ranging from R&D activities to major manufacturing plants. tonnes or 20% of world demand, fulfilling about half of its domestic
The first 100-MW thin-film factories became operational in 2007, and demand (Baoshan, 2010).
the announcements of new production capacities accelerated again in
2008. If all expansion plans are realized in time, thin-film production Projections of silicon production capacities for solar applications in 2012
capacity could be 20.0 GW, or 35% of the total 56.7 GW in 2012, and span a range between 140,000 tonnes from established polysilicon pro-
23.5  GW, or 34% of a total of 70  GW in 2015 (Jäger-Waldau, 2009, ducers, up to 250,000 tonnes including new producers (e.g., Bernreuther

Chapter 3 Direct Solar Energy

and Haugwitz, 2010; Ruhl et al., 2010). The possible solar cell produc- transfer fluids such as molten salts. The accepted standard to date has
tion will also depend on the material use per Wp. Material consumption been to use large heliostats, but many of the new entrants are pursuing
could decrease from the current 8 g/Wp to 7 g/Wp or even 6 g/Wp (which much smaller heliostats to gain potential cost reductions through high-
could increase delivered PV capacity from 31 to 36 to 42 GW, respec- volume mass production. The companies now interested in heliostat
tively), but this may not be achieved by all manufacturers. development range from optics companies to the automotive industry
looking to diversify. High-temperature steam receivers will benefit from
Forecasts of the future costs of vital materials have a high-profile history, existing knowledge in the boiler industry. Similarly, with linear Fresnel,
and there is ongoing public debate about possible material shortages a range of new developments are occurring, although not yet as devel-
and competition regarding some (semi-)metals (e.g., In and Te) used in oped as the central-receiver technology.
thin-film cell production. In a recent study, Wadia et al. (2009) explored
material limits for PV expansion by examining the dual constraints of Dish technology is much more specialized, and most effort presently
material supply and least cost per watt for the most promising semicon- has been towards developing the dish/Stirling concept as a commercial
ductors as active photo-generating materials. Contrary to the commonly product. Again, the technology can be developed as specialized compo-
assumed scarcity of indium and tellurium, the study concluded that nents through specific industry know-how such as the Stirling engine
the currently known economic reserves of these materials would allow mass-produced through the automotive industry.
about 10 TW of CdTe or CuInS2 solar cells to be installed.
Within less than 10 years prior to 2010, the CSP industry has gone from
In CSP electricity generation, the solar collector field is readily scalable, negligible activity to over 2,400 MWe either commissioned or under
and the power block is based on adapted knowledge from the existing construction. A list of new CSP plants and their characteristics can be
power industry such as steam and gas turbines. The collectors themselves found at the IEA SolarPACES web site.3 More than ten different com-
benefit from a range of existing skill sets such as mechanical, structural panies are now active in building or preparing for commercial-scale
and control engineers, and metallurgists. Often, the materials or compo- plants, compared to perhaps only two or three who were in a position to
nents used in the collectors are already mass-produced, such as glass build a commercial-scale plant three years ago. These companies range
mirrors. from large organizations with international construction and project
management expertise who have acquired rights to specific technolo-
By the end of 2010, strong competition had emerged and an increas- gies, to start-ups based on their own technology developed in-house. In
ing number of companies had developed industry-level capability to addition, major independent power producers and energy utilities are
supply materials such as high-reflectivity glass mirrors and manufac- playing a role in the CSP market.
tured components. Nonetheless, the large evacuated tubes designed
specifically for use in trough/oil systems for power generation remain The supply chain does not tend to be limited by raw materials, because
a specialized component, and only two companies (Schott and Solel) the majority of required materials are bulk commodities such as glass,
have been capable of supplying large orders of tubes, with a third steel/aluminium, and concrete. The sudden new demand for the specific
company (Archimedes) now emerging. The trough concentrator itself solar salt mixture material for molten-salt storage is claimed to have
comprises know-how in both structures and thermally sagged glass mir- impacted supply. At present, evacuated tubes for trough plants can be
rors. Although more companies are now offering new trough designs produced at a sufficient rate to service several hundred MW per year.
and considering alternatives to conventional rear-silvered glass (e.g., However, expanded capacity can be introduced readily through new fac-
polymer-based reflective films), the essential technology of concentra- tories with an 18-month lead time.
tion remains unchanged. Direct steam generation in troughs is under
demonstration, as is direct heating of molten salt, but these designs are Solar fuel technology is still at an emerging stage—thus, there is no
not yet commercially available. As a result of its successful operational supply chain in place at present for commercial applications. However,
history, the trough/oil technology comprised most of the CSP installed solar fuels will comprise much of the same solar-field technology being
capacity in 2010. deployed for other high-temperature CSP systems, with solar fuels
requiring a different receiver/reactor at the focus and different down-
Linear Fresnel and central-receiver systems comprise a high level of stream processing and control. Much of the downstream technology,
know-how, but the essential technology is such that there is the poten- such as Fischer-Tropsch liquid fuel plants, would come from existing
tial for a greater variety of new industry participants. Although only a expertise in the petrochemical industry. The scale of solar fuel dem-
couple of companies have historically been involved with central receiv- onstration plants is being ramped up to build confidence for industry,
ers, new players have entered the market over the last few years. There which will eventually expand operations.
are also technology developers and projects at the demonstration level
(China, USA, Israel, Australia, Spain). Central-receiver developers are 3 See: www.solarpaces.org.
aiming for higher temperatures, and, in some cases, alternative heat

Direct Solar Energy Chapter 3

Hydrogen has been touted as a future transportation fuel due to its Chapter 1. Solar technologies differ in levels of maturity, and although
versatility, pollutant-free end use and storage capability. The key is a some applications are already competitive in localized markets, they
sustainable, CO2-free source of hydrogen such as solar, cost-effective generally face one common barrier: the need to achieve cost reductions
storage and appropriate distribution infrastructure. The production of (see Section 3.8). Utility-scale CSP and PV systems face different bar-
solar hydrogen, in and of itself, does not produce a hydrogen economy riers than distributed PV and solar heating and cooling technologies.
because many factors are needed in the chain. The suggested path to Important barriers include: 1) siting, permitting and financing challenges
solar hydrogen is to begin with solar enhancement of existing steam to develop land with favourable solar resources for utility-scale projects;
reforming processes, with a second generation involving solar electricity 2) lack of access to transmission lines for large projects far from electric
and advanced electrolysis, and a third generation using thermolysis or load centres; 3) complex access laws, permitting procedures and fees for
advanced thermochemical cycles, with many researchers aiming for the smaller-scale projects; 4) lack of consistent interconnection standards
production of fuels from concentrated solar energy, water, and CO2. In and time-varying utility rate structures that capture the value of distrib-
terms of making a transition, solar hydrogen can be mixed with natu- uted generated electricity; 5) inconsistent standards and certifications
ral gas and transported together in existing pipelines and distribution and enforcement of these issues; and 6) lack of regulatory structures
networks to customers, thus enhancing the solar portion of the global that capture environmental and risk mitigation benefits across technolo-
energy mix. gies (Denholm et al., 2009).

Steam reforming of natural gas for hydrogen production is a con- Through appropriate policy designs (see Chapter 11), governments have
ventional industrial-scale process that produces most of the world’s shown that they can support solar technologies by funding R&D and by
hydrogen today, with the heat for the process derived from burning a providing incentives to overcome economic barriers. Price-driven instru-
significant proportion of the fossil fuel feedstock. Using concentrated ments (see Section 11.5.2), for example, were popularized after feed-in
solar power, instead, as the source of the heat embodies solar energy in tariff (FIT) policies boosted levels of PV deployment in Germany and
the fuel. The solar steam-reforming of natural gas and other hydrocar- Spain. In 2009, various forms of FIT policies were implemented in more
bons, and the solar steam-gasification of coal and other carbonaceous than 50 countries (REN21, 2010) and some designs offer premiums for
materials yields a high-quality syngas, which is the building block for a building-integrated PV. Quota-driven frameworks such as renewable
wide variety of synthetic fuels including Fischer-Tropsch-type chemicals, portfolio standards (RPS) and government bidding are common in the
hydrogen, ammonia and methanol (Steinfeld and Meier, 2004). USA and China, respectively (IEA, 2009a). Traditional RPS frameworks
are designed to be technology-neutral, and this puts at a disadvantage
The solar cracking route refers to the thermal decomposition of natural many solar applications that are more costly than alternatives such as
gas and other hydrocarbons. Besides H2 and carbon, other compounds wind power. In response, features of RPS frameworks (set-asides and
may also be formed, depending on the reaction kinetics and on the credits) increasingly are including solar-specific policies, and such pro-
presence of impurities in the raw materials. The thermal decomposition grams have led to increasing levels of solar installations (Wiser et al.,
yields a carbon-rich condensed phase and a hydrogen-rich gas phase. 2010). In addition to these regulatory frameworks, fiscal policies and
The carbonaceous solid product can either be sequestered without CO2 financing mechanisms (e.g., tax credits, soft loans and grants) are often
release or used as material commodity (carbon black) under less severe employed to support the manufacturing of solar goods and to increase
CO2 restraints. It can also be applied as reducing agent in metallurgical consumer demand (Rickerson et al., 2009). The challenge for solar proj-
processes. The hydrogen-rich gas mixture can be further processed to ects to secure financing is a critical barrier, especially for developing
high-purity hydrogen that is not contaminated with oxides of carbon; technologies in market structures dominated by short-term transactions
thus, it can be used in proton-exchange-membrane fuel cells without and planning.
inhibiting platinum electrodes. From the perspective of carbon seques-
tration, it is easier to separate, handle, transport and store solid carbon Most successful solar policies are tailored to the barriers posed by spe-
than gaseous CO2. Further, thermal cracking removes and separates cific applications. Across technologies, there is a need to offset relatively
carbon in a single step. The major drawback of thermal cracking is the high upfront investment costs (Denholm et al., 2009). Yet, in the case
energy loss associated with the sequestration of carbon. Thus, solar of utility-scale CSP and PV projects, substantial and long-term invest-
cracking may be the preferred option for natural gas and other hydro- ments are required at levels that exceed solar applications in distributed
carbons with a high H2/C ratio (Steinfeld and Meier, 2004). markets. Solar heating and cooling technologies are included in many
policies, yet the characteristics of their applications differ from electric-
ity-generating technologies. Policies based on energy yield rather than
3.4.3 Impact of policies4 collector surface area are generally preferred for various types of solar
thermal collectors (IEA, 2007). See Section 1.5 for further discussion.
Direct solar energy technologies support a broad range of applications,
and their deployment is confronted by many of the barriers outlined in Similar to other renewable sources, there is ongoing discussion about
the merits of existing solar policies to spur innovation and accelerate
4 Non-technology-specific policy issues are covered in Chapter 11 of this report. deployment using cost-effective measures. Generally—and as discussed

Chapter 3 Direct Solar Energy

in Chapter 11—the most successful policies are those that send clear, 3.5.2 District heating and other thermal loads
long-term and consistent signals to the market. In addition to targeted
economic policies, government action through educationally based Highly insulated buildings can be heated easily with relatively low-
schemes (e.g., workshops, workforce training programs and seminars) temperature district-heating systems, where solar energy is ideal, or
and engagement of regulatory organizations are helping to overcome quite small quantities of renewable-generated electricity (Boyle, 1996).
many of the barriers listed in this section. 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 sys-
3.5 Integration into the broader energy tem becomes a viable solution to the cooling demand of buildings. There
system5 are already many DCS installations in the USA, Europe, Japan and other
Asian countries because this system has many advantages compared to
This section discusses how direct solar energy technologies are part of a decentralized cooling system. For example, it takes full advantage of
the broader energy framework, focusing specifically on the following: economy of scale and diversity of cooling demand of different buildings,
low-capacity energy demand; district heating and other thermal loads; reduces noise and structure load, and saves considerable equipment area.
PV generation characteristics and the smoothing effect; and CSP gen- It also allows greater flexibility in designing the building by removing the
eration characteristics and grid stabilization. Chapter 8 addresses the cooling tower on the roof and chiller plant in the building or on the roof,
broader technical and institutional options for managing the unique and it can provide more reliable and flexible services through a special-
characteristics, production variability, limited predictability and loca- ized professional team in cold-climate areas (Shu et al., 2010). For more
tional dependence of some RE technologies, including solar, as well as on RE integration in district heating and cooling networks, see Section
existing experience with and studies associated with the costs of that
In China, Greece, Cyprus and Israel, solar water heaters make a significant
contribution to supplying residential energy demand. In addition, solar
3.5.1 Low-capacity electricity demand 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.,
There can be comparative advantages for using solar energy rather than Australia, Canada and the USA), the impact of numerous solar domestic
non-renewable fuels in many developing countries. Within a country, the water heaters on the operation of the power grid depends on the util-
advantages can be higher in un-electrified rural areas compared to urban ity’s load management strategy. For a utility that uses centralized load
areas. Indeed, solar energy has the advantage, due to being modular, of switching to manage electric water heater load, the impact is limited to
being able to provide small and decentralized supplies, as well as large fuel savings. Without load switching, the installation of many solar water
centralized ones. For more on integrated buildings and households, see heaters may have the additional benefit of reducing peak demand on the
Section 8.3.2. 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
In a wide range of countries, particularly those that are not oil producers, a capacity benefit from load displacement of electric water heaters. Large-
solar energy and other forms of RE can be the most appropriate energy scale deployment of solar water heating can benefit both the customer
source. If electricity demand exceeds supply, the lack of electricity can and the utility. Another benefit to utilities is emissions reduction, because
prevent development of many economic sectors. Even in countries with solar water heating can displace the marginal and polluting generating
high solar energy sustainable development potential, RE is often only con- plant used to produce peak-load power.
sidered to satisfy high-power requirements such as the industrial sector.
However, large-scale technologies such as CSP are often not available to Combining biomass and low-temperature solar thermal energy could pro-
them due, for example, to resource conditions or suitable land area avail- vide zero emissions and high capacity factors to areas with less frequent
ability. In such cases, it is reasonable to keep the electricity generated near direct-beam solar irradiance. In the short term, local tradeoffs exist for
the source to provide high amounts of power to cover industrial needs. areas that have high biomass availability due to increased cloud cover
Applications that have low power consumption, such as lighting in rural and rainfall. However, solar technology is more land-efficient for energy
areas, can primarily be satisfied using onsite PV—even if the business plan production and greatly reduces the need for biomass growing area and
for electrification of the area indicates that a grid connection would be biomass transport cost. Some optimum ratio of CSP and biomass supply
more profitable. Furthermore, the criteria to determine the most suitable is likely to exist at each site. Research is being conducted on tower and
technological option for electrifying a rural area should include benefits dish systems to develop technologies—such as solar-driven gasification of
such as local economic development, exploiting natural resources, creat- biomass—that optimally combine both these renewable resources. In the
ing jobs, reducing the country’s dependence on imports, and protecting longer term, greater interconnectedness across different climate regimes
the environment. 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
5 Non-technology-specific issues related to integration of RE sources in current and system.
future energy systems are covered in Chapter 8 of this report.

Direct Solar Energy Chapter 3

3.5.3 Photovoltaic generation characteristics and the Wiemken et al. (2001) used data from actual PV systems in Germany
smoothing effect to demonstrate that five-minute ramps in normalized PV power output
at one site may exceed ±50%, but that five-minute ramps in the nor-
At a specific location, the generation of electricity by a PV system varies malized PV power output from 100 PV systems spread throughout the
systematically during a day and a year, but also randomly according to country never exceed ±5%. Ramachandran et al. (2004) analyzed the
weather conditions. The variation of PV generation can, in some instances, reduction in power output fluctuation for spatially dispersed PV systems
have a large impact on voltage and power flow of the local transmission/ and for different time periods, and they proposed a cluster model to
distribution system from the early penetration stage, and on supply- represent very large numbers of small, geographically dispersed PV sys-
demand balance in a total power system operation in the high-penetration tems. Results from Curtright and Apt (2008) based on three PV systems
stage (see also Section 8.2.1 for a further discussion of solar electricity in Arizona indicate that 10-minute step changes in output can exceed
characteristics, and the implications of those characteristics for electricity 60% of PV capacity at individual sites, but that the maximum of the
market planning, operations, and infrastructure). aggregate of three sites is reduced. Kawasaki et al. (2006) similarly
analyzed the smoothing effect within a small (4 km by 4 km) network
Various studies have been published on the impact of supply-demand of irradiance sensors and concluded that the smoothing effect is most
balance for a power system with a critical constraint of PV systems inte- effective during times when the irradiance variability is most severe—
gration (Lee and Yamayee, 1981; Chalmers et al., 1985; Chowdhury and particularly days characterized as partly cloudy.
Rahman, 1988; Jewell and Unruh, 1990; Bouzguenda and Rahman, 1993;
Asano et al., 1996). These studies generally conclude that the economic Murata et al. (2009) developed and validated a method for estimating
value of PV systems is significantly reduced at increasing levels of system the variability of power output from PV plants dispersed over a wide
penetration due to the high variability of PV. Today’s base-load generation area that is very similar to the methods used for wind by Ilex Energy
has a limited ramp rate—the rate at which a generator can change its out- Consulting Ltd et al. (2004) and Holttinen (2005). Mills and Wiser (2010)
put—which limits the feasible penetration of PV systems. However, these measured one-minute solar insolation for 23 sites in the USA and char-
studies generally lack high-time-resolution PV system output data from acterized the variability of PV with different degrees of geographic
multiple sites. The total electricity generation of numerous PV systems in diversity, comparing the variability of PV to the variability of similarly
a broad area should have less random and fast variation—because the sited wind. They determined that the relative aggregate variability of PV
generation output variations of numerous PV systems have low correla- plants sited in a dense ten by ten array with 20-km spacing is six times
tion and cancel each other in a ‘smoothing effect’. The critical impact on less than the variability of a single site for variability on time scales
supply-demand balance of power comes from the total generation of the of less than 15 minutes. They also found that for PV and wind plants
PV systems within a power system (Piwko et al., 2007, 2010; Ogimoto et similarly sited in a five by five grid with 50-km spacing, the variability
al., 2010). of PV is only slightly more than the variability of wind on time scales of
5 to 15 minutes.
Some approaches for analyzing the smoothing effect use modelling
and measured data from around the world. Cloud models have been Oozeki et al. (2010) quantitatively evaluated the smoothing effect in a
developed to estimate the smoothing effect of geographic diversity load-dispatch control area in Japan to determine the importance of data
by considering regions ranging in size from 10 to 100,000 km2 (Jewell accumulation and analysis. The study also proposed a methodology to
and Ramakumar, 1987) and down to 0.2 km2 (Kern and Russell, 1988). calculate the total PV output from a limited number of measurement
Using measured data, Kitamura (1999) proposed a set of specifications data using Voronoi Tessellation. Marcos et al. (2010) analyzed one-
for describing fluctuations, considering three parameters: magnitude, second data collected throughout a year from six PV systems in Spain,
duration of a transition between clear and cloudy, and speed of the ranging from 1 to 9.5 MWp, totalling 18 MW. These studies concluded
transition, defined as the ratio of magnitude and duration; he evalu- that over shorter and longer time scales, the level of variability is nearly
ated the smoothing effect in a small area (0.1 km by 0.1 km). A similar identical because the aggregate fluctuation of PV systems spread over
approach, ‘ramp analysis’, was proposed by Beyer et al. (1991) and the large area depends on the correlation of the fluctuation between
Scheffler (2002). PV systems. The correlation of fluctuation, in turn, is a function both
of the time scale and distance between PV systems. Variability is less
In a statistical approach, Otani et al. (1997) characterized irradiance correlated for PV systems that are further apart and for variability over
data by the fluctuation factor using a high-pass filtered time series of shorter time scales.
solar irradiance. Woyte et al. (2001, 2007) analyzed the fluctuations of
the instantaneous clearness index by means of a wavelet transform. To Currently, however, not enough data on generation characteristics exist
demonstrate the smoothing effect, Otani et al. (1998) demonstrated that to evaluate the smoothing effect. Data collection from a sufficiently
the variability of sub-hourly irradiance even within a small area of 4 large number of sites (more than 1,000 sites and at distances of 2 to 200
km by 4 km can be reduced due to geographic diversity. They analyzed km), periods and time resolution (one minute or less) had just begun
the non-correlational irradiation/generation characteristics of several PV in mid-2010 in several areas in the world. The smoothed generation
systems/sites that are dispersed spatially. characteristics of PV penetration considering area and multiple sites will

Chapter 3 Direct Solar Energy

be analyzed precisely after collecting reliable measurement data with 3.6.1 Environmental impacts
sufficient time resolution and time synchronization. The results will con-
tribute to the economic and reliable integration of PV into the energy No consensus exists on the premium, if any, that society should pay for
system. 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,
3.5.4 Concentrating solar power generation 2001; Bickel and Friedrich, 2005; NEEDS, 2009; NRC, 2010). Solar energy
characteristics and grid stabilization has been considered desirable because it poses a much smaller environ-
mental burden than non-renewable sources of energy. This argument
In a CSP plant, even without integrated storage, the inherent thermal has almost always been justified by qualitative appeals, although this
mass in the collector system and spinning mass in the turbine tend to is changing.
significantly reduce the impact of rapid solar transients on electrical out- Results for damage costs per kilogram of pollutant and per kWh were
put, and thus, lead to less impact on the grid (also see Section 8.2.1). By presented by the International Solar Energy Society in Gordon (2001).
including integrated thermal storage systems, base-load capacity factors The results of studies such as NEEDS (2009), summarized in Table 3.3
can be achieved (IEA, 2010b). This and the ability to dispatch power on for PV and in Table 3.4 for CSP, confirm that RE is usually comparatively
demand during peak periods are key characteristics that have motivated beneficial, though impacts still exist. In comparison to the figures pre-
regulators in the Mediterranean region, starting with Spain, to support sented for PV and CSP here, the external costs associated with fossil
large-scale deployment of this technology with tailored FITs. CSP is suit- generation options, as summarized in Chapter 10.6, are considerably
able for large-scale 10- to 300-MWe plants replacing non-renewable higher, especially for coal-fired generation.
thermal power capacity. With thermal storage or onsite thermal backup
(e.g., fossil or biogas), CSP plants can also produce power at night or Considering passive solar technology, higher insulation levels provide
when irradiation is low. CSP plants can reliably deliver firm, scheduled many benefits, in addition to reducing heating loads and associated
power while the grid remains stable. costs (Harvey, 2006). The small rate of heat loss associated with high
levels of insulation, combined with large internal thermal mass, creates
CSP plants may also be integrated with fossil fuel-fired plants such as a more comfortable dwelling because temperatures are more uniform.
displacing coal in a coal-fired power station or contributing to gas- This can indirectly lead to higher efficiency in the equipment supply-
fired integrated solar combined-cycle (ISCC) systems. In ISCC power ing the heat. It also permits alternative heating systems that would not
plants, a solar parabolic trough field is integrated in a modern gas and
steam power plant; the waste heat boiler is modified and the steam
Table 3.3 | Quantifiable external costs for photovoltaic, tilted-roof, single-crystalline sili-
turbine is oversized to provide additional steam from a solar steam con, retrofit, average European conditions; in US2005 cents/kWh (NEEDS, 2009).
generator. Better fuel efficiency and extended operating hours make
2005 2025 2050
combined solar/fossil power generation much more cost-effective than
Health Impacts 0.17 0.14 0.10
separate CSP and combined-cycle plants. However, without including
thermal storage, solar steam could only be supplied for some 2,000 of Biodiversity 0.01 0.01 0.01

the 6,000 to 8,000 combined-cycle operating hours of a plant in a year. Crop Yield Losses 0.00 0.00 0.00

Furthermore, because the solar steam is only feeding the combined-cycle Material Damage 0.00 0.00 0.00

turbine—which supplies only one-third of its power—the maximum Land Use N/A 0.01 0.01

solar share obtainable is under 10%. Nonetheless, this concept is of Total 0.18 0.17 0.12
special interest for oil- and gas-producing sunbelt countries, where solar
power technologies can be introduced to their fossil-based power mar-
ket (SolarPACES, 2008). Table 3.4 | Quantifiable external costs for concentrating solar power; in US2005 cents/
kWh (NEEDS, 2009).

2005 2025 2050

3.6 Environmental and social impacts6 Health Impacts 0.65 0.10 0.06

Biodiversity 0.03 0.00 0.00

This section first discusses the environmental impacts of direct solar Crop Yield Losses 0.00 0.00 0.00
technologies, and then describes potential social impacts. However, an Material Damage 0.01 0.00 0.00
overall issue identified at the start is the small number of peer-reviewed
Land Use N/A N/A N/A
studies on impacts, indicating the need for much more work in this area.
Total 0.69 0.10 0.06

6 A comprehensive assessment of social and environmental impacts of all RE sources

covered in this report can be found in Chapter 9.

Chapter 5

5 Hydropower

Coordinating Lead Authors:

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),
Ånund Killingtveit (Norway), Zhiyu Liu (China)

Contributing Authors:
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)

This chapter should be cited as:

Kumar, A., T. Schei, A. Ahenkorah, R. Caceres Rodriguez, J.-M. Devernay, M. Freitas, D. Hall, Å. 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. Schlömer, C. von Stechow (eds)], Cambridge University Press, Cambridge,
United Kingdom and New York, NY, USA.

Chapter 5 Hydropower

significant effect on the power output. Increased risks of landslides 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
and glacial lake outbursts, and impacts of increased variability, are
A1B scenario in 12 different climate models (Milly et al., 2008), UNEP world regions and
of particular concern to Himalayan countries (Agrawala et al., 2003). data for the hydropower system in 2005 (US DOE, 2009) as presented in Hamududu and
The possibility of accommodating increased intensity of seasonal pre- Killingtveit (2010).
cipitation by increasing storage capacities may become of particular
Power Generation Capacity (2005) Change by 2050
importance (Iimi, 2007). REGION
TWh/yr (PJ/yr)
GW TWh/yr (PJ/yr)
Africa 22 90 (324) 0.0 (0)
In Europe, by the 2070s, hydropower potential for the whole of Europe
Asia 246 996 (3,586) 2.7 (9.7)
has been estimated to potentially decline by 6%, translated into a 20
Europe 177 517 (1,861) -0.8 (-2.9)
to 50% decrease around the Mediterranean, a 15 to 30% increase in
North America 161 655 (2,358) 0.3 (≈1)
northern and Eastern Europe, and a stable hydropower pattern for west-
South America 119 661 (2,380) 0.3 (≈1)
ern and central Europe (Lehner et al., 2005).
Oceania 13 40 (144) 0.0 (0)

In New Zealand, increased westerly wind speed is very likely to enhance TOTAL 737 2931 (10,552) 2.5 (9)

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 In general the results given in Table 5.2 are consistent with the (mostly
rainfall to snowfall, and to increase river flows in winter and early qualitative) results given in previous studies (IPCC, 2007b; Bates et al.,
spring. This is very likely to increase hydroelectric generation during the 2008). For Europe, the computed reduction (-0.2%) has the same sign,
winter peak demand period, and to reduce demand for storage. 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
In Latin America, hydropower is the main electrical energy source for changes by 2070, so a direct comparison is difficult.
most countries, and the region is vulnerable to large-scale and persistent
rainfall anomalies due to El Niño and La Niña, as observed in Argentina, It can be concluded that the overall impacts of climate change on the
Colombia, Brazil, Chile, Peru, Uruguay and Venezuela. A combination of existing global hydropower generation may be expected to be small, or
increased energy demand and droughts caused a virtual breakdown of even slightly positive. However, results also indicated substantial varia-
hydroelectricity in most of Brazil in 2001 and contributed to a reduction tions in changes in energy production across regions and even within
in gross domestic product (GDP). Glacier retreat is also affecting hydro- countries (Hamududu and Killingtveit, 2010).
power generation, as observed in the cities of La Paz and Lima.
Insofar as a future expansion of the hydropower system will occur incre-
In North America, hydropower production is known to be sensitive to mentally in the same general areas/watersheds as the existing system,
total runoff, to its timing, and to reservoir levels. During the 1990s, for these results indicate that climate change impacts globally and aver-
example, Great Lakes levels fell as a result of a lengthy drought, and in aged across regions may also be small and slightly positive.
1999, hydropower production was down significantly both at Niagara
and Sault St. Marie. For a 2°C to 3°C warming in the Columbia River Still, uncertainty about future impacts as well as increasing difficulty of
Basin and BC Hydro service areas, the hydroelectric supply under worst- future systems operations may pose a challenge that must be addressed
case water conditions for winter peak demand is likely to increase (high in the planning and development of future HPP (Hamududu et al., 2010).
confidence). Similarly, Colorado River hydropower yields are likely to
decrease significantly, as will Great Lakes hydropower. Northern Québec Indirect effects on water availability for energy purposes may occur if
hydropower production would be likely to benefit from greater pre- water demand for other uses such as irrigation and water supply for
cipitation and more open-water conditions, but hydropower plants in households and industry rises due to the climate change. This effect is
southern Québec would be likely to be affected by lower water levels. difficult to quantify, and it is further discussed in Section 5.10.
Consequences of changes in the seasonal distribution of flows and in
the timing of ice formation are uncertain.
5.3 Technology and applications
In a recent study (Hamududu and Killingtveit, 2010), the regional and
global changes in hydropower generation for the existing hydropower Head and also installed capacity (size) are often presented as criteria for
system were computed, based on a global assessment of changes in the classification of hydropower plants. The main types of hydropower,
river flow by 2050 (Milly et al., 2005, 2008) for the SRES A1B scenario however, are run-of-river, reservoir (storage hydro), pumped storage,
using 12 different climate models. The computation was done at the and in-stream technology. Classification by head and classification by
country or political region (USA, Canada, Brazil, India, China, Australia) size are discussed in Section 5.3.1. The main types of hydropower are
level, and summed up to regional and global values (see Table 5.2). presented in Section 5.3.2. Maturity of the technology, status and

Hydropower Chapter 5

current trends in technology development, and trends in renovation This broad spectrum in definitions of size categories for hydropower may
and modernization follow in Sections 5.3.3 and 5.3.4 respectively. 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 define eligibility for specific support schemes (e.g.,
5.3.1 Classification by head and size US Renewable Portfolio Standards). It clearly illustrates that different
countries have different legal definitions of size categories that match
A classification by head refers to the difference between the upstream their local energy and resource management needs.
and the downstream water levels. Head determines the water pressure
on the turbines that together with discharge are the most important Regardless, there is no immediate, direct link between installed capac-
parameters for deciding the type of hydraulic turbine to be used. ity as a classification criterion and general properties common to all
Generally, for high heads, Pelton turbines are used, whereas Francis HPPs above or below that MW limit. Hydropower comes in manifold
turbines are used to exploit medium heads. For low heads, Kaplan and project types and is a highly site-specific technology, where each project
Bulb turbines are applied. The classification of what ‘high head’ and is a tailor-made outcome for a particular location within a given river
‘low head are varies widely from country to country, and no generally basin to meet specific needs for energy and water management services.
accepted scales are found. 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
Classification according to size has led to concepts such as ‘small hydro’ operation worldwide. Similarly, while larger facilities will tend to have
and ‘large hydro’, based on installed capacity measured in MW as the lower costs on a USD/kW basis due to economies of scale, that ten-
defining criterion. Small-scale hydropower plants (SHP) are more likely dency will only hold on average. Moreover, one large-scale hydropower
to be run-of-river facilities than are larger hydropower plants, but res- project of 2,000 MW located in a remote area of one river basin might
ervoir (storage) hydropower stations of all sizes will utilize the same have fewer negative impacts than the cumulative impacts of 400 5-MW
basic components and technologies. Compared to large-scale hydro- hydropower projects in many river basins (Egré and Milewski, 2002).
power, however, it typically takes less time and effort to construct and For that reason, even the cumulative relative environmental and social
integrate small hydropower schemes into local environments (Egré and impacts of large versus small hydropower development remain unclear,
Milewski, 2002). For this reason, the deployment of SHPs is increasing in and context dependent.
many parts of the world, especially in remote areas where other energy
sources are not viable or are not economically attractive. All in all, classification according to size, while both common and admin-
istratively simple, is—to a degree—arbitrary: general concepts like
Nevertheless, there is no worldwide consensus on definitions regarding ‘small’ or ‘large hydro’ are not technically or scientifically rigorous indi-
size categories (Egré and Milewski, 2002). Various countries or groups cators of impacts, economics or characteristics (IEA, 2000c). Hydropower
of countries define ‘small hydro’ differently. Some examples are given projects cover a continuum in scale, and it may be more useful to evalu-
in Table 5.3. From this it can be inferred that what presently is named ate a hydropower project on its sustainability or economic performance
‘large hydro’ spans a very wide range of HPPs. IJHD (2010) lists several (see Section 5.6 for a discussion of sustainability), thus setting out more
more examples of national definitions based on installed capacity. realistic indicators.

Table 5.3 | Small-scale hydropower by installed capacity (MW) as defined by various countries

Small-scale hydro as defined by

Country Reference Declaration
installed capacity (MW)

Brazil ≤30 Brazil Government Law No. 9648, of May 27, 1998

Natural Resources Canada, 2009: canmetenergy-canmetenergie.nrcan-rncan.gc.ca/eng/renewables/

Canada <50

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
US National Hydropower Association. 2010 Report of State Renewable Portfolio Standard Programs (US
USA 5–100