You are on page 1of 334

2014

Energy Risk Professional ERP Exam Course Pack

READINGS THAT ARE FREELY AVAILABLE ON THE GARP WEBSITE


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

Readings for Electricity Markets and Renewable Generation


System Reliability and Demand Response 1. Bo Shen, Girish Ghatikhar, Chun Chun Ni, and Junqiao Dudley. Addressing Energy Demand Through Demand Response. (Berkeley National Laboratory, June 2012). (Sections 1 to 4 only) Define demand response (DR) and understand how DR works to curtail shortages on a power grid. Understand how government policy and market deregulation have been instrumental 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 management (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 Australias National Electricity Market. Understand the roles of AEMO and NEM in the Australian market. Identify the Market Price Cap and Market Floor Price and understand their use in electricity transactions. Describe the scheduling of generators for a given hour under NEM rules and calculate the market price. Understand the difference between regulated and unregulated interconnectors and the auction of inter-region settlement residues. 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 deregulated market. Summarize the process a TSO uses to manage supply and demand on the electric grid. Calculate the settlement (payment) for a quantity of dispatched/consumed power under NORD Pool rules. Understand how NORD Pool up-regulation and down-regulation pricing ensures efficient market operation.

5.

Hogan, Lovells, Lee & Lee. Singapore Energy Market (Schedule 1: Summary of Singapore Electricity Market Deregulation and Wholesale Market Operations). Understand Singapores creation of artificial LNG prices and the rationale for this policy. Describe the deregulation of Singapores 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 3.33.5 only) Compare and contrast various solar generation technologies, including photovoltaic (PV) and concentrating solar power (CSP). Understand how PV systems are integrated into the power grid. Understand technologies used to transport and store solar generated electricity. Explain the smoothing effect with respect to multiple PV solar systems.

Chapter 5 ........................Hydropower (Sections 5.35.5 only) Compare and contrast run-of-river, storage hydropower, and pumped storage technologies. Understand factors that impact the efficiency of a hydropower plant, and compare the efficiency of hydropower turbines over a range of discharge levels. Describe the characteristics of hydropower generation and understand how these characteristics can contribute to the reliable operation of a power grid.

2014 Global Association of Risk Professionals. All rights reserved.

2014

Energy Risk Professional ERP Exam Course Pack

Chapter 7.........................Wind Energy (Sections 7.37.5 and 7.8 only) Understand the dynamics of the power curve for a wind turbine. Compare and contrast the generation technology and output associated with onshore and offshore wind facilities. Understand the challenges associated with planning electric power systems that include wind energy, including the integration of wind power to the grid. Define capacity credits and compare the capacity credit of wind power to hydrocarbon-based power. Explain how incremental wind power capacity impacts price formation, volatility, and balancing costs in an electricity market. Understand the factors that impact the economics of wind power generation, including the capacity factor, the levelized cost of energy, and the effect of government policies. 9. Larry Parker (U.S. Congressional Research). Climate Change and the EU-Emissions Trading Scheme (ETS): Looking to 2020. Understand the ETS system; assess the greenhouse gas reduction commitment under 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 depreciation.

2014 Global Association of Risk Professionals. All rights reserved.

LBNL-5580E

Addressing Energy Demand through Demand Response: International Experiences and Practices
Bo Shen, Girish Ghatikar, Chun Chun Ni, and Junqiao Dudley Environmental Energy Technologies Division Lawrence Berkeley National Laboratory Phil Martin and Greg Wikler ENERNOC, INC.

Reprint version of the report prepared for AZURE INTERNATIONAL , December 1, 2011 .

June 2012
This work was supported by AZURE INTERNATIONAL, CESP and Energy Foundation through the U.S. Department of Energy under Contract No. DE-AC02-05CH11231.

Disclaimer

This document was prepared as an account of work sponsored by the United States Government. While this document is believed to contain correct information, neither the United States Government nor any agency thereof, nor The Regents of the University of California, nor any of their employees, makes any warranty, express or implied, or assumes any legal responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by its trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof, or The Regents of the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof, or The Regents of the University of California. Ernest Orlando Lawrence Berkeley National Laboratory is an equal opportunity employer.

Table of Contents
1. INTRODUCTION ..........................................................................................................................1
1.1. 1.2. Definition of Demand Response ..................................................................................................................1 Benefits Brought by Demand Response .......................................................................................................1

2.

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


2.1. 2.2. 2.3. Overview of Demand Side Management .....................................................................................................4 Role of DR in Demand-Side Management ....................................................................................................5 Coordination of Demand Response and Energy Efficiency ..........................................................................6

3.

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


3.1.
3.1.1 3.1.2

Demand Response Enabling Policies ............................................................................................................7


Wholesale Market Access ......................................................................................................................................... 7 Bilateral Programs with Vertically Integrated Utilities and Network Operators..................................................... 12

3.2.
3.2.1 3.2.2 3.2.3 3.2.4

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


Lack of Sufficient Incentives from Standard and TOU Pricing: Experience with Interruptible Tariffs ...................... 19 Cost and RisksHow Load Aggregators have Removed Traditional Barriers to DR Participation ........................... 20 Avoid Energy Costs vs. Resource Payments ............................................................................................................ 21 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. 4.2. 4.3. Metering and Control Solutions .................................................................................................................25 Auto-DR and OpenADR (with the AMI linkage) ..........................................................................................27 Smart Meter and Advanced Metering Infrastructure (AMI) and OpenADR...............................................29

5.

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


5.1.
5.1.1 5.1.2 5.1.3 5.1.4 5.1.5 5.1.6 5.1.7

DR Strategies in Commercial and Industrial Buildings ...............................................................................32


DR Strategies for Commercial Buildings ................................................................................................................. 33 DR Strategies for Industrial Facilities ...................................................................................................................... 33 DR Strategies in Water or Wastewater Facilities .................................................................................................... 33 DR Strategies in Refrigerated Warehouse Facilities ................................................................................................ 34 DR Strategies in Food Processing Facilities ............................................................................................................. 34 DR Strategies in Data Centers ................................................................................................................................. 35 DR Strategies in Heavy Industry .............................................................................................................................. 35

6.

RECOMMENDATIONS AND KEY PRINCIPLES FOR DESIGNING AND IMPLEMENTING DR IN CHINA. 36

List of Acronyms
AESO Alberta Energy System Operator

AMI AMP Auto-DR BRA


C&I

advanced metering infrastructure


aggregator managed portfolio automated demand response base residual auction commercial and industrial Climate Action Plan

CAP CEC CPP CSPs


DECC

California Energy Commission critical peak pricing


curtailment service providers U.K. Department of Energy and Climate Change

DLC DOE DPCR5 DR


DRAS DRRC

direct load control U.S. Department of Energy


distribution price control review

demand response
demand response automation server Demand Response Research Center

DSM EE EEPS EILS EISA 2007 EMCS EPACT ERCOT FCM FERC FRCC
HVAC

demand side management energy efficiency


Energy Efficiency Portfolio Standard emergency interruptible load service

Energy Independence and Security Act of 2007


energy management control systems

The Energy Policy Act of 2005 Electric Reliability Council of Texas


forward capacity market

U.S. Federal Energy Regulatory Commission Florida Reliability Coordinating Council


heating ventilation and air-conditioning

I/C IA IESO IOUs


IT

Interruptible/curtailable
incremental auctions independent electricity system operator

investor-owned utilities
information technology

LMP
M&V

locational marginal price


measurement and verification

MRO
NOCs

Midwest Reliability Organization


network operation centers

NPCC OPA PG&E PPA PTR RFC RPM RTDR RTEG RTO RTP SCE SERC SPP SRM STOR TOU TVA WECC WEM

Northeast Power Coordinating Council, Inc.


Ontario Power Authority California Pacific Gas and Electric power purchase agreement peak time rebates

ReliabilityFirst Corporation
reliability pricing model real time demand response real time emergency generation regional transmission organization

real-time pricing
Southern California Edison

Southeastern Electric Reliability Council Southwest Power Pool


sycnhronized reserve market short term operating reserves

time-of-use
Tennessee Valley Authority

Western Electricity Coordinating Council


whosesale electricity market

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

1.2. Benefits Brought by Demand Response


There are a variety of benefits brought by DR, ranging from the environmental to the economic. Environmental benefits By reducing electric demand to ensure the sufficiency of existing supply, rather than increasing supply to meet rising demand, DR avoids power plant operation and its associated emissions. Moreover, because DR capacity is distributed, there are added benefits due to the avoidance of electrical losses in the transmission and distribution lines typically experienced from centrally-generated utility power. The USbased energy consultancy Synapse Energy Economics addressed this issue in its study of DR and air emissions in the US market of ISO New England: when DR operates it reduces system line losses relative to reference case system operation. This is because when energy is provided to customers from the grid it often comes from power plants a considerable distance from the point of end use, and energy is lost in transmission. Usually line losses are in the range of 5 to 10 percent, but they can be higher during periods when transmission lines are heavily loaded. In contrast, the DR resource be it a load reduction or a generator is located at the site of energy use, so no energy is lost in transmission. Because DR avoids line losses, a DR resources of five MW is comparable to a grid-connected asset of slightly larger than five MWFor example, a five-MW DR resource might be credited as

U.S. Department of Energy (February 2006), Benefits of Demand Response in Electricity Markets and Recommendations for Achieving Them: A Report to the United States Congress Pursuant to Section 1252 of The Energy Policy Act of 2005, pp. 11-12. (http://eetd.lbl.gov/ea/ems/reports/congress-1252d.pdf). 2 http://www.energy.ca.gov/2011_energypolicy/documents/2011-07-06_workshop/background/Metrics_July_IEPR_DR_v1.pdf.

providing 5.5 MW of reserve capacity is average system line losses during DR events were determined to be roughly 10 percent.3 In addition, because DR is often procured on a forward basis, it may not only offset the operation of power plants but also their very construction. In this manner, the environmental benefits of DR extend to the avoided emissions associated with the construction of the materials for the power plant itself (i.e. cement, steel, etc.), as well as the potential ecological impact that may have resulted should the unit have been constructed. The use of DR for non- peak-shaving purposes such as for ancillary services, also comes with significant environmental benefits, despite the very short duration dispatches of such resources. In many systems, ancillary services (also known as reserves), are primarily provided by plants in running operating mode, as there may be an insufficient number of quick-start generating units able to start, synchronize, and export power to the grid in the requisite period of time. These plants tend to be fueled by diesel or oil, which add to local and regional pollution. Increased use of quick-response DR can reduce the need for power plants to run in operating mode, as well as potentially lead to a more efficient overall use of resources within the system. Economic benefits The economic benefits of DR oftentimes may be more significant than the environmental benefits. While there is a clear environmental benefit to avoiding or reducing power plant operation, the targeted usage of DR will not save the same amount of energy as permanent load reductions that come from energy efficiency measures. As peak periods are relatively infrequent, so too tends to be the use of DR. Yet, the infrequent spikes in demand have a significant economic impact: in many systems, 10% (or more) of costs are incurred to meet demands which occur less than 1% of the time.4 Reducing this peak demand through DR programs means that the capacity requirements which drive investments in generation, transmission, and distribution assets can also be proportionally reduced. The US-based energy consultancy the Brattle Group, in its 2007 paper The Power of Five Percent, found that a 5% reduction in peak demand would have resulted in avoided generation and T&D capacity costs of $2.7 billion per year.5 In addition, the use of DR during peak periods can result in significant savings in terms of energy expenditure. In wholesale markets, spot energy prices during peak periods can skyrocket due to increased demand. Similarly, energy prices in vertically integrated, non-wholesale market systems can also increase during peak periods as less efficient units (i.e. with a higher heat rate) are utilized in order to meet the rising demand. As retail energy rates tend to not reflect the true cost of energy during peak periods, the expensive utilization of generation during these times is socialized among all customers. By reducing the need to purchase high-priced power, all customers in a system are positively impacted. The
3 4

Synapse Energy Economics, Modeling Demand Response and Air Emissions in New England, September 4, 2003. Page 16. EnerNOC, Inc. Analysis of US and Australian Electricity System Data.; Brattle Group, The Power of Five Percent, May 16, 2007 5 Brattle Group, The Power of Five Percent, May 16, 2007.

aforementioned Brattle report also identified energy savings on the order of $300 million per year from the same 5% reduction in the peak demand of the US as a whole. Figure 1 further illustrates the point that the avoided capacity costs far outweigh the avoided energy and avoided T&D costs.

Figure 1: Annual Benefits of 5% Remand Response in the US6 Indeed, it is important to recognize the financial benefits participants in these programs receive, which are the sum of both the avoided energy costs (and demand charges) as well as the direct incentive payments for participation and successful performance.

Id.

2. DEMAND-SIDE MANAGEMENT AND THE ROLE OF DR


2.1. Overview of Demand Side Management
Demand-side management (DSM) consists of a broad range of planning, implementing and monitoring of activities designed to encourage end-users to modify their levels and patterns of electricity consumption. DSM programs and initiatives are typically implemented to achieve two basic objectives: energy efficiency (EE) and load management. EE is primarily achieved through programs that reduce overall energy consumption of specific end-use devices and systems by promoting high-efficiency equipment use and building design. Conversely, load management programs are designed to achieve reductions in consumption primarily during times of peak demand, rather than on a permanent or ongoing basis. Load management programs can include permanent-load shifting and peak-shaving activities traditionally associated with demand response. With improvements in technology, load management programs are also increasingly dispatched on a level playing field with supply-side resources. Figure 2 presents the total peak load reduction through DSM from 1998 to 2009 in the US. Peak load reduction through energy efficiency programs increased from 13,591 MW in 1998 to 19,766 MW in 2009, but decreased from 13,640 MW to 11,916 MW during the same period by load management programs.7

35,000 30,000

25.0% 20.0%

Peak Load Reduction (MW)

25,000 20,000 15,000 10,000 5,000 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Energy Efficiency Annual Growth Rate of Energy Efficiency Load Management Annual Growth Rate of Load Management

10.0% 5.0% 0.0% -5.0% -10.0% -15.0%

-20.0%
-25.0%

Figure 2: Peak Load Reductions from DSM Programs by Program Category8

U.S. Energy Information Administration (November 23, 2010), Demand-Side Management Actual Peak Load Reduction by Program Category. (http://205.254.135.24/cneaf/electricity/epa/epat9p1.html) 8 US Energy Information Agency. 2009.

Annual Growth Rate (%)

15.0%

2.2. Role of DR in Demand-Side Management


Demand Response is normally included as part of utility DSM program or a potential DSM program solution which helps make the electric grid much more efficient and balanced by assisting the electric grid's commercial and industrial customers in reducing their electric peak demands, and/or shifting the time period when they use their electricity, and/or prioritizes the way they use electricity, and in return reduces their overall energy costs. A key difference between DR and EE is the energy reductions for DR are time-dependent, whereas reductions for EE are not. Demand response programs yield reductions in demand at critical times, which typically corresponds to time of peak power demand, while EE programs yield permanent energy savings. However, the two programs have overlapping effects: EE can permanently reduce demand including those occurred during the peak time while demand response with well-targeted control strategies can also produce energy savings.9 Up to 2003, EE programs in the US contributed to more than 60% of actual peak load reduction;10 however, its share dropped by almost 10% from 59.3% in 2004 to 49.9% in 2009. Meanwhile, contributions from load management had increased by about 12% from 37.6% to 50.1% during the same period (see Figure 3).

100.0% 90.0%

80.0%
70.0%

(Unit: %)

60.0%
50.0%

40.0%
30.0% 20.0% 10.0% 0.0% 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Energy Efficiency Load Management

Goldman, Charles, M. Reid, R. Levy, and A. Silverstein (January 2010), Coordination of Energy Efficiency and Demand Response, Lawrence Berkeley National Laboratory, LBNL-3044E. (http://eetd.lbl.gov/ea/ems/reports/lbnl-3044e.pdf) 10 Peak load reductions are categorized as potential or actual. Potential peak load reductions are the amount of load available for curtailment through load control programs such as direct load control, interruptible load control, other load management, or other DSM programs. Actual peak load reductions are the amount of reduction that is achieved from load control programs that are put into force at the same time as peak load and the amount of reductions that result from energy efficiency programs at the time of peak load.

Figure 3: Share of Total Actual Peak Load Reduction by Program Category11

2.3. Coordination of Demand Response and Energy Efficiency


Demand response and energy efficiency programs could be coordinated at the customer level at least in the following four ways:12 Offering combined programs: Although separating energy efficiency and demand response programs are quite common, customers could be presented with both opportunities at the same time. Furthermore, technologies that are commonly used for DR such as energy monitoring, building automation systems and load control equipment can also be leveraged to help inform about opportunities that would lead to energy efficiency improvements. Coordinating program marketing and education: Program sponsors could package and promote demand response and energy efficiency in a closely coordinated way. Because the two programs are quite complicated to customers, program sponsors could help customers addresses both topics under a broad DSM and management theme. Market-driven coordination services: Effective coordination can be done not only by utilities and Independent System Operators (a.k.a. ISO), but also by the initiative of private firms that find a market among customers who are interested in reducing their energy costs or receiving incentives.

Incorporating Building codes and appliance standards: Building codes and appliance efficiency standards can incorporate demand response and energy efficiency functions into the design of buildings, infrastructure, and power-consuming appliances/equipments. Integrating those codes and standards can lead to significant reduction in the costs to customers of integrating demand response and energy efficiency strategies and measures.

11 12

Id. Cappers, Peter, C. Goldman, and D. Kathan (2009), Demand Response in U.S. Electricity Markets: Empirical Evidence, Lawrence Berkeley National Laboratory, LBNL-2124E. (http://eetd.lbl.gov/ea/ems/reports/lbnl-2124e.pdf).

3. REGULATORY AND POLICY FRAMEWORKS THAT PROMOTE DEMAND RESPONSE


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

3.1.1 Wholesale Market Access


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

13 14

The current size of the ISO-New England system is approximately 26 GW. The current size of the PJM system is approximately 165 GW. 15 Federal Energy Regulatory Commission (2011). Assessment of Demand Response & Advanced Metering Staff Report. November 2011. Table 2.

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

16

Load aggregation is the process by which individual energy users band together in an alliance to secure more competitive prices than they might otherwise receive working independently. Oftentimes, load aggregator companies are formed to represent the interests of these groups of customers. 17 Cappers, Peter, C. Goldman, and D. Kathan (2009).

PJM (Pennsylvania-New Jersey-Maryland) Interconnection is a regional transmission organization (RTO) that coordinates the movement of wholesale electricity in all or parts of Delaware, Illinois, Indiana, Kentucky, Maryland, Michigan, New Jersey, North Carolina, Ohio, Pennsylvania, Tennessee, Virginia, West Virginia and the District of Columbia. PJM is the largest market in the US and allows DR to participate in all of its markets types capacity, energy, and ancillary services. Today, more than 60 entities serve as Curtailment Service Providers (CSPs), or load aggregators, in the PJM system. Like in most systems, the bulk of the DR in PJM participates in the capacity market. Capacity markets are particularly well-suited to peaking resources like DR which operate for relatively few hours a year and may have trouble accessing the proper price signals from an energy-only market. PJMs current capacity market, the Reliability Pricing Model (RPM), was instituted in 2007. In the RPM, those resources include not only generating stations, but also demand response actions and energy efficiency measures by consumers to reduce their demand for electricity. In this manner, demand side management is directly integrated into the wholesale capacity market structure. Every year PJM conducts a Base Residual Auction (BRA) for delivery of capacity three years in the future. The BRA is held in May and the delivery year begins 3 years later on June 1st and ends on May 31st of the following year. In addition to the BRA, PJM conducts three Incremental Auctions (IA) that are held in advance of each corresponding delivery year. The purpose of the IA is to balance any changes in the load forecast and to allow suppliers of capacity resources to adjust their positions. In PJMs most recent Base Residual Auction in May 2011, the market procured 149,974 MW of capacity for the 2014/2015 delivery year. Of note, 14,118 MW of this capacity or 9.4% of the total came from demand response resources. Once cleared through the capacity market, these DR resources become participants in PJMs Emergency Load Response Program. In PJM, qualifying DR resources can also participate in the wholesale energy market (both day-ahead and real-time) as well as various ancillary service markets (primarily, the synchronized reserve market). However, these markets are not the same drivers of DR growth that the capacity market is. The energy market does not feature capacity incentives, and therefore requires significantly more participation to garner the same financial opportunity. The Synchronized Reserve Market, on the other hand, requires full response within 10-minutes of a dispatch signal and generation-grade telemetry, limiting the pool of potential participants. ISO New England (ISO-NE) PJMs counterpart to the north, ISO New England, also operates a forward capacity market (FCM) in which DR can participate alongside generation, and which accounts for the majority of demand side participation within the New England system. Similar to PJMs BRA, the ISO-NE FCM also allows both dispatchable demand response and energy efficiency measures to participate in the market. ISO-NEs use of demand response may be the clearest example of a resource designed specifically for reliability and/or emergency prevention purposes. While the ELRP in PJM is also designed for similar
9

purposes, the trigger for usage in ISO-NE is even more defined. ISO-NE treats DR provided by curtailment and on-site generation as distinct resources, labeling the former Real Time Demand Response (RTDR) and the latter Real Time Emergency Generation (RTEG). RTDR may be called by ISO-NE only when the system reaches an emergency level known as Operating Procedure 4 Action 9 (OP4 Action 9). RTEG, on the other hand, cannot be dispatched until a further level of emergency has been reached, OP4 Action 12. For customers that utilize both load curtailment and on-site generation to provide DR capacity, they (or their DR provider), must be able to call those distinct loads separately in order to comply with ISO-NE requirements. Today, approximately 2,000 MW, or 8% of the resources in the capacity market, are dispatchable demand response. This figure grows to 3,400 MW, or 10% of the ISO-NE system, in 2014/15. Demand response resources can also provide energy to the ISO-NE market through the Real-Time Price Response and Day-Ahead Load Response Programs. As with energy market participation in PJM, these programs are relatively unpopular compared to the capacity market, as they require much more frequent participation and have comparatively lower economic benefit. In both markets, sites that participate in the energy programs tend to be among the most flexible participants in the capacity markets who are looking for an additional economic opportunity, rather than the energy program serving as the sole method of DR participation in the market. While ISO-NE formerly had a pilot program testing the ability for DR to provide ancillary services the Demand Response Reserve Pilot (DRRP) it no longer has an active mechanism for DR to provide operating or spinning reserves. DR participation in these markets is now under active consideration, in part due to the aforementioned FERC Order 719. UNITED KINGDOM National Grid Short Term Operating Reserves (STOR) Market Demand response resources also enjoy wholesale market access in the United Kingdom, albeit in a much more limited context. Market-based opportunities for demand-side resources in the UK are currently restricted to ancillary service markets, primarily the Short Term Operating Reserves Market. While others exist, the parameters result in low levels of participation and or a small addressable market.18 DR cannot access the nations wholesale energy market and unlike PJM and ISO-NE, there is no capacity market in the UK. That said, the government, spearheaded by the Department of Energy and Climate Change (DECC), is pushing forward legislation to launch one in the coming years and which will also allow for demand side participation. STOR is essentially a supply and demand balancing service that meets the need of the grid as demand changes and as traditional power plants come online and ramp up and down, similar in many ways to the Sycnhronized Reserve Market (SRM) in the PJM Interconnection. While not a capacity market per se, cleared resource receive an availability payment for each hour they are in the market and available to be dispatched. Utilization (energy) payments are also given for the actual load reduction provided. Both features are also present in the aforementioned SRM. As a balancing market, STOR is called much more
18

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

10

frequently than the capacity programs in PJM and ISO-NE which are used primarily to address emergency conditions. STOR participants, on average, must be prepared to respond to a dispatch every week. Such frequent participation requires the employment of different curtailment strategies than those that are found in capacity programs designed to shave consumption only during infrequent, peak periods. AUSTRALIA Independent Market Operator Wholesale Electricity Market (WEM) Another successful example of DR participation in wholesale markets can be found in the South West Interconnected System of Western Australia, run by the Independent Market Operator (IMO). There are two wholesale markets in Australia; the Whosesale Electricity Market (WEM) in Western Australia, and the National Electricity Market (NEM) in the eastern states (except for the Northern Territory). The NEM is an energy-only market and has very low levels of DR participation for the reasons mentioned earlier, whereas the WEM is a capacity market similar in many ways to ISO-NE and PJM, and with a significant penetration of DR. In the most recent Reserve Capacity Cycle, more than 8% of the capacity procured came from demand-side resources.19 The IMO-administered WEM procures system resources through the Reserve Capacity Mechanism, in which capacity can be traded bilaterally to the IMO directly, or to retailers. Unlike in PJM and ISO-NE where capacity prices are the result of competitive offers, the IMO sets a price for all capacity based on the avoided cost of a marginal new peaking unit, specifically a 160 MW open cycle gas turbine. Auctions are only triggered if the bilateral trading mechanism secures insufficient capacity.20 As in the previously discussed capacity markets of the US, DR and generation receives the same exact market payment. As with PJM and ISO-NE, DR is assumed to have different levels of dispatch capability than traditional supply-side resources. The RCM has 4 Availability Classes; Generation must all list itself as Class 1, or available for more than 96 hours a year; DR meanwhile can offer at between 24-96 hours of dispatch. One important distinction about DR in WA is that unlike PJM and ISO-NE; DR in the WEM can be dispatched when it is deemed to be economic and is not dependent on emergency conditions. The system operator is required to first utilize the plants of the former state-owned generation company, but afterwards all dispatch is determined by the market energy price offered by the resource. Unlike its counterparts in the US, DR in WA is limited to participation in the wholesale capacity market. While a wholesale energy market also exists in WA, the STEM, DR resources do not have access to the market. Meanwhile, a competitive balancing market is only just now being designed, and access for DR is expected once the market is fully operational in the coming years.

19

The Reserve Capacity Cycle (RCC) is the process that is used in Australia to procure DR resources as part of the Reserve Capacity Mechanism. 20 Note that to date this situation has never been experienced so no auctions have been called.

11

Capacity-based Programs in Energy-only Markets Some wholesale market operators have taken a slightly different approach, creating DR-specific opportunities outside of the standard wholesale markets themselves. For the most part, these are energy-only markets, where the underlying structure is not as conducive to peaking resources like demand response, which operate for relatively few hours per year. In the Canadian Province of Ontario, the Independent Electricity System Operator (IESO) and the Ontario Power Authority (OPA) launched a large scale DR program (DR3) in 2007 to provide additional capacity to the market due to planned retirement of coal-fired power plants in the province. DR3s inclusion of a capacity payment represented a departure from previous DR programs in Ontario that failed to gain traction, primarily due to incentives being limited to energy payments. Another example is the Emergency Interruptible Load Service (EILS) program in the Texas market of ERCOT. EILS is essentially a standalone markets that exists alongside the wholesale markets open to generation resources in ERCOT. EILS is designed to provide reserve capacity to the energy-only ERCOT market, and is procured during four separate markets spaced evenly throughout the year. Unlike DR3, pricing in EILS is the result of offers made by DR providers, similar to how the capacity markets in PJM and ISO-NE work. A further similarity with PJM is that ERCOT allows EILS participating loads to provide operating reserves in the ERCOT ancillary service markets and receive the same payment as generation resources, similar to the SRM.

3.1.2 Bilateral Programs with Vertically Integrated Utilities and Network Operators
Access to existing wholesale markets are just one mechanism for creating and leveraging demand response resources. In recent years, much growth in the industry has been found in bilateral programs with vertically integrated utilities in traditionally regulated environments, and with network (T&D) operators located within a liberalized market structure. These bilateral programs are most often used as a way to avoid or defer investments in generation and/or T&D infrastructure, and tend to look similar in structure to a power purchase agreement (PPA) that a utility might sign with an independent power producer. These utility programs are likely better proxies for how the implementation of nextgeneration demand response could manifest itself in China, given the lack of a wholesale market. There are a number of enabling policies that have encouraged the development of bilateral DR programs throughout North America, the UK and Australia. These policies include: Cost recovery and DSM funds Loading orders and similar regulations Peak demand mandates and energy efficiency portfolio standards Cost Recovery and DSM Funds Whether in the US, the UK, or Australia, vertically integrated utilities and distribution network operators are regulated monopolies whose revenues are dependent on government policy and regulation. As such,
12

it is essential to understand the regulatory environments in which these utilities operate in order to understand how regulatory policies have both contributed to, and hindered, the growth of demand response. Perhaps the most basic and essential enabling policy is a cost-recovery mechanisms. Under a costrecovery mechanism, a utility can recover prudently-incurred costs of DR and EE investments on a dollar-for-dollar basis, typically through a rider or customer surcharge. Cost recovery is designed to make a utility whole on its DR and EE investments. However, there are challenges with this approach. First, cost recovery alone will not address the lost margin revenue the utility will face due to reduced energy sales from DR and EE programs. Second, cost recovery does not factor in opportunity costs: DR and EE investments displace supply-side investments for which the utility can earn a profit. Given these opportunity costs, absent a statutory or regulatory mandate, program cost recovery alone will generally not attract utility interest in DR and EE programs. However, in some jurisdictions, utilities are authorized to recover additional costs associated with the lost revenue due to the energy efficiency measures. There are also provisions for earning a fair rate of return on the DSM investment, typically at levels that are equivalent to allowable returns on power generation assets. Loading Orders and Similar Regulations Loading orders are governmental proclamations that define the priority order in which resources are to be developed. To underscore the importance of energy efficiency and demand response in Californias future energy picture, the state government developed the Energy Action Plan established a loading order of preferred resources, placing energy efficiency and demand response as the states highestpriority procurement resource, and set aggressive long-term goals for energy efficiency and demand response resources. In addition, energy efficiency and demand response strategies were implemented to address greenhouse gas emission reduction targets specified by AB32, a law adopted in California to create regulatory policy mechanisms to combat global warming. As a result of these policies, Californias energy efficiency and demand response efforts have proven to be very successful. California leads the nation in term of energy saved. The state invests nearly $3 billion per year in energy efficiency and demand response programs that target electricity and natural gas customers to install high efficiency equipment, take measures to reduce their peak demands, and establish time-sensitive price structures that are more in line with the actual cost of providing the electricity. Resources such as renewable generation, distributed generation, and traditional generation are considered as the second and third priorities, respectively in the loading order, and should only be considered once all energy efficiency and demand response resources are exhausted. In Massachusetts, a law known as the Green Communities Act was passed in 2008 and implemented shortly thereafter. The law requires the states utilities to procure all available energy efficiency resources that cost less than traditional energy sources do. The law in effect prioritizes energy efficiency as being at the top of the loading order, ahead of renewable energy, and more traditional forms of generation. Among the major provisions is a requirement for utilities to invest in energy efficiency when it is less expensive than buying power. Previously companies purchased more power when demand
13

increased. The effect of the law is that the state is seeing significant investments in energy efficiency, leading toward the ultimate goal of reducing the states use of fossil fuels in buildings by 10% and overall greenhouse gas emissions by 20% in the year 2020. Peak Demand Mandates, Energy Efficiency Portfolio Standards Peak demand mandates and energy efficiency portfolio standards have recently emerged as another mechanism to encourage DR outside of market-based opportunities. Perhaps most well known is a mandate in the state of Pennsylvania, the so-called Act 129 legislation, signed into law in October 2008, which requires all electric distribution companies to achieve peak demand reduction targets of 4.5% and energy efficiency reductions of 4% by 2015. While the legislation does not expressly encourage DR over other types of peak reduction such as energy efficiency and or solar PV, Pennsylvania utilities appear to have determined C&I DR was the most cost effective way to reach compliance and several large deals with aggregators have already been publicly announced. Other states with peak demand mandates that are similar to Pennsylvania include New York, Colorado, Michigan and Ohio. In New York, the Public Service Commission established an Energy Efficiency Portfolio Standard (EEPS) which ordered the states utilities to achieve a 15% reduction in forecast electricity usage by the year 2015. The states utilities are implementing aggressive EE and DR programs in order to meet that goal, which specifies that each of the states utilities realize specific MWh and peak MW reduction amounts by 2015. In Colorado, the Climate Action Plan (CAP) sets carbon reduction goals for the state and proclaims that energy efficiency programs are the most important responses to the carbon-reduction challenge. In response, the Colorado Public Utilities Commission has ordered the states utilities to implement EE and DR programs to meet that goal. Michigan and Ohio have similar statutory mandates to lower energy usage and peak demand. Parity of Treatment Traditional utility regulation favors supply-side resources over DR and EE resources. First, utilities earn a rate of return on investments in generation, transmission and distribution infrastructure. The absence of a parallel incentive for DR and EE investments creates a bias against demand-side resources. This has been described in the economic literature as the Averch-Johnson Effect. That is, where a firms profits are linked to its capital investment, as is the case with utilities under traditional regulatory structures, there is an embedded incentive for the firm to increase its capital outlay in a manner that does not necessarily maximize producer and consumer surplus. Stated another way, traditional regulatory frameworks create a disincentive for utilities to meet resource needs using approaches that are less capital intensive. Thus, faced with otherwise equivalent alternatives of building a power plant that contributes to profitability or making investments in DR and EE that allow for cost-recovery only, a utility would generally prefer to build a power plant (or T&D). The government of the United Kingdom recently recognized and addressed this very challenge. In the 2010-2015 Distribution Price Control Review 5 (DPCR5), Ofgem the national electricity and gas regulator instituted the so-called Equalisation Incentive which establishes parity in the treatment of
14

capital and operating expenditures by distribution utilities. Thus, any utility acting in its own rational economic interest will clearly pursue the most cost-effective way to meet network needs and reliability requirements, whether that is through traditional investments in infrastructure or through non-network alternatives like DSR. As a result of this new regulation, one local distribution network operator Electricity North West has already deployed a commercial scale DR program in which an aggregator is deploying DR on specific circuits in order to defer investments in substations. Other distribution network operators, such as UK Power Networks, are also conducting pilot projects using DR for distribution relief as they hope to prepare themselves to launch commercial-scale programs under this new regulatory framework. Example Utility DR Programs There are several examples of bilateral DR programs. In California, Pacific Gas and Electric (PG&E) implements the Aggregator Managed Portfolio (AMP) program. AMP is a non-tariff program that consists of bilateral contracts with aggregators to provide PG&E with price-responsive demand response. The program can be called at PG&Es discretion. Each aggregator is responsible for designing and implementing their own demand response program, including customer acquisition, marketing, sales, retention, support, event notification and payments. To participate, customers must enroll through a load aggregator. The customer in turn authorizes the aggregator to act on their behalf with respect to all aspects of AMP, including receipt of notification of an event, receipt of incentive payments and/or penalties. Southern California Edison (SCE) operates the Demand Response Contracts (DRC) program. SCE has contracted with several aggregator companies to provide SCE with price-responsive and/or demand response events that SCE may call at its discretion. Each aggregator designs their own programs, and offers demand response program structures and options that may not be directly available through SCE. Customers may select an aggregator with services that best meet their business needs. More common are arrangements where a utility contracts with a single DR load aggregator for a program in their territory (or a single provider per customer class). For example, EnerNOC, a Bostonbased load aggregator has a program in place with the Tennessee Valley Authority (TVA) in the southeastern US, the largest public power company in the country. TVA procured a long-term, 560 MW resource from EnerNOC which it is required to deliver in line with contract requirements over the 10year contract length. There are many other load aggregator companies operating in the various electricity markets throughout North America. As with the aforementioned DR programs in California, the load aggregator is responsible for all roles from customer acquisition through resource dispatch and settlement. As with similar DR programs, TVA has purchased a guaranteed firm resource. In addition to identifying and enabling DR capacity in line with contract milestones, the load aggregator must also meet performance standards when dispatched by TVA. Should the load aggregator fail to do either, financial penalties against the aggregator may be assessed. In this manner, TVA can depend on its DR-based virtual power plant in the same way its system planners and operators can trust a traditional generation resource. Figure 4 provides a summary of the TVA bi-lateral program parameters.
15

Program Size Advanced Notification Dispatch Trigger Availability Window Maximum Cumulative Dispatches Term Length

Up to 560 MW 30 minutes TVAs discretion April October: 12:00-20:00, Mon-Fri November March: 5:00-13:00, Mon-Fri 40 hours per annum 10 years

Figure 4: TVA Bi-lateral DR Program Parameters Other vertically-integrated utilities in the US that have implemented similar programs include: Arizona Public Service, Idaho Power, NV Energy, Public Service Company of New Mexico, Puget Sound Energy, Salt River Project, San Diego Gas & Electric, Tampa Electric, Tucson Electric Power, and Xcel Energy. Per the aforementioned equalisation incentive now in effect in the UK, distribution network operators (DNOs) in the country are also now deploying demand response programs to defer or avoid investments in network infrastructure. Electricity North West (ENW), one of the 14 regulated DNOs in the UK with a network that includes the Greater Manchester and Cumbria areas, has recently launched a DR program along with a third-party load aggregator. Under this program, DR is deployed within specified circuits in the network, allowing demand to be controlled on a geographically-targeted basis that will prevent the need to upgrade the substations on those portions of the network. The program, announced in May 2011, is set to last for five years. Such network support contracts are also commonly found in Australia, particularly in New South Wales. The same Distribution Price Control Review that launched the equalisation incentive, also included funds from Ofgem the UK electric regulator for Low Carbon Network (LCN) projects that will pilot new technologies and facilitate the development of an environmentally-friendly electricity system in the country. Many DNOs throughout the UK have successfully applied for LCN funding to pilot the use of DR in their networks, including UK Power Networks (UKPN) and Northern Powergrid (formerly CE Electric). ENW has also recently been awarded LCN funding from Ofgem to pilot the use of DR in new ways within their system that, if successful, would reduce the amount of network capacity DNOs would need to have in order to comply with reliability standards.

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


The U.S. Department of Energy classifies demand response into two categories, i.e. price-based demand response and incentive-based demand response.21 Each category has its own subcategories. Pricing mechanisms vary on each subcategory as shown in Table 1. Price-based demand response refers to changes in usage by customers in response to changes in the prices they pay and include real-time pricing, critical-peak pricing, and time-of-use rates. If the price differentials between hours or time periods are significant, customers can respond to
21

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

16

the price structure with significant changes in energy use, reducing their electricity bills if they adjust the timing of their electricity usage to take advantage of lower-priced periods and/or avoid consuming when prices are higher. Customers load use modifications are entirely voluntary (Table 1) Incentive-based demand response programs are established by utilities, load-serving entities, or a regional grid operator. These programs give customers load-reduction incentives that are separate from, or additional to, their retail electricity rate, which may be fixed (based on average costs) or time-varying. The load reductions are needed and requested either when the grid operator thinks reliability conditions are compromised or when prices are too high. Most demand response programs specify a method for establishing customers baseline energy consumption level, so observers can measure and verify the magnitude of their load response. Some demand response programs penalize customers that enroll but fail to respond or fulfill their contractual commitments when events are declared (Table 1).

17

Table 1: Demand Response Options and Related Pricing Mechanisms


Price-Based (Voluntary) Time-of-use (TOU): a rate with different unit price for usage during different blocks of time, usually defined for a 24 hour day. TOU rates reflect the average cost of generating and delivering power during those time periods. Real-time pricing (RTP): a rate in which the price for electricity typically fluctuates hourly reflecting changes in the wholesale price of electricity. Customers are typically notified of RTP prices on a day-ahead or hour-ahead basis. Critical Peak Pricing (CPP): CPP rates are a hybrid of the TOU and RTP design. The basic rate structure is TOU. However, provision is make for replacing the normal peak price with a much higher CPP event price under specified trigger conditions (e.g., when system reliability is compromised or supply prices are very high). Incentive-Based (Contractually Mandatory) Direct load control: a program by which the program operator remotely shuts down or cycles a customers electrical equipment (e.g., air conditioner, water heater) on short notice. Direct load control programs are primary offered to residential or small commercial customers. Interruptible/curtailable (I/C) service: curtailment options integrated into retail tariffs that provide a rate discount or bill credit for agreeing to reduce load during system contingencies. Penalties maybe assessed for failure to curtail. Interruptible programs have traditionally been offered only to the largest industrial (or commercial) customers. Demand Bidding/Buyback Program: customers offer bids to curtail based on wholesale electricity market prices or an equivalent. Mainly offered to large customers (e.g., one megawatt [MW] and over). Emergency Demand Response Programs: programs that provide incentive payments to customers for load reductions during periods when reserve shortfall arise. (e.g. ERCOT EILS) Capacity Market Programs: customers offer load curtailments as system capacity to replace conventional generation or delivery resources. Customers typically receive day-of notice of events. Incentives usually consist of up-front reservation payments, and face penalties for failure to curtail when called upon to do so. (e.g. PJM ELRP, IMO WA) Ancillary Services Market Program: customers bid load curtailments in ISO/RTO markets as operating reserves. If their bids are accepted, they paid the market price for committing to be on standby. If their load curtailments are needed, they are called by the ISO/RTO, and may be paid the spot market energy price. (e.g. PJM SRM, UK STOR)

Source: DOE (2006), p.12.

In addition to federal regulation as described in Section 3.1 and economic benefits described in Section 3.2, numbers of the U.S. utilities have taken action to expand their retail demand response programs. One incentive factor for many of them has been concern about peak load growth and rising energy prices.22
22

U.S. Federal Energy Regulatory Commission (December 2008), Assessment of Demand Response and Advanced Metering, Washington D.C. (http://www.ferc.gov/legal/staff-reports/12-08-demand-response.pdf).

18

3.2.1. Lack of Sufficient Incentives from Standard and TOU Pricing: Experience with Interruptible Tariffs
Many utilities have offered a variety of traditional DR programs for many years. These legacy programs are typically referred to as load management programs. There are three types of legacy load management programs: direct load control (DLC), time-of-use (TOU) rates, and interruptible contracts. Each of these programs use some form of incentive to encourage customers to participate. However, the amount of the incentives or the nature of the incentives has not been sufficient to bring about meaningful levels of demand reductions. DLC programs allow the utility to directly control customer end-uses during certain periods when the electrical system is under strain. The customer end-uses are directly controlled by the utility and when events are called, those loads are either shut down, cycled on and off, or moved to a lower consumption periods. Residential DLC programs often target air conditioners or electrical water heaters. Nonresidential DLC programs include air conditioner systems, lighting and in some regions irrigation control. There are a number of challenges with DLC programs. First, customers tend to become frustrated with effects of the service interruptions and oftentimes will leave the program if they are called too frequently. Second, the incentives offered by the utilities have been insufficient to encourage their sustained participation. TOU rates are tariff schedules that are typically offered to residential and small business customers on a voluntary basis and are mandatory for the largest commercial and industrial customers. The TOU rates are structured to charge lower rates during a utilitys off-peak and partial-peak periods and higher rates during seasonal and daily peak demand periods. By charging more during the peak period, when incremental costs are highest, TOU rates send accurate marginal-cost price signals to customers. TOU rates encourage customers to shift energy use away from peak periods to partial-peak or off-peak periods and enable customers to lower their electricity bills. There are two common challenges with TOU rates. First, the utilities have often set the TOU peak periods to be for long periods at a time, thus limiting customers abilities to shift their loads to the lower price off-peak periods. Second, the TOU rate programs tend to be static in nature in that the peak and off-peak prices do not change regardless of system conditions and the true costs required to deliver electricity to customers. Because of the static nature of the TOU rates, they cannot be counted on for meeting the peak demand needs of the utility. In addition, the utilities often design these tariffs to be revenue neutral. That is, the price differentials between on-peak and off-peak are intended to not change the utilitys overall revenue. This goal oftentimes is inconsistent with a goal of maximizing customer participation in order to have meaningful peak demand reductions as a result of the TOU tariff. Interruptible tariffs are contractual arrangements set up between the utility and large non-residential customers. Customers agree to reduce their electrical consumption to a pre-specified level, or by a prespecified amount, during system reliability problems in return for an incentive payment or a similar rate
19

discount. Customers are given the incentive regardless of whether reliability events are called. In the past, these programs were developed mostly for customer retention as the utilities assured customers that reliability events were so rare and would never be called. However, as reliability problems are becoming more acute, utilities are calling more interruptible events. As a result, many customers are opting to negotiate an exit to their contractual obligations for these programs as they cannot tolerate the volume interruptions to their businesses.

3.2.2. Cost and RisksHow Load Aggregators have Removed Traditional Barriers to DR Participation
Complicated tariff structures and insufficient incentives are just a few of the challenges utilities face when trying to garner customer interest in traditional, non-aggregator-based DR programs. Equally important are the costs and risks customers must bear in order to participate. While the costs for metering and load control equipment may not always be borne by the customer in these situations, the exposure to performance penalties remains essentially a constant. Without an aggregator to guarantee the load response, utilities have no choice but to penalize customers if they dont fully comply with a dispatch in order to ensure proper response. However, C&I loads are inherently volatile and customers may not always be able to participate and consequently customers may need to be willing to face a strong likelihood of penalties if they seek to participate. Using load aggregators is one proven approach to removing many of these traditional barriers to DR participation. It is typical that the load aggregator pays all costs for the installation of metering and load control equipment, making participation for the customer a no-cost proposition. More importantly, because load aggregators are measured on the total load reduction their entire portfolio of sites provides, and not on a site-by-site basis, they are able to pool resources in a way that ensures that contract performance requirements can be met. In the event that performance penalties are assessed on the aggregator, many will still refrain from passing these onto the customer. Figure 5 illustrates this concept.

20

Load Aggregator

Utility/TSO

Figure 5: Aggregated Performance and Risk-Shielding

3.2.3. Avoid Energy Costs vs. Resource Payments


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

3.2.4. Emerging Trends: Dynamic Pricing


Dynamic pricing refers to a category of rates that offer customers time-varying electricity prices on a day-ahead or real-time basis. Prices are higher during peak periods to reflect higher-than-average cost
21

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

3.3. Summary/Comparative Analysis of Policy and Regulatory Frameworks


The global survey of demand response programs in this report illustrates a variety of regulatory constructs under which DR can thrive. Fundamentally, all of these regulations and policies in one way or another attempt to change the traditional paradigm that has historically lead to investments in additional supply-side infrastructure rather than load management. Clearly, one method that has been incredibly successful in this regard is the wholesale capacity market. By removing the type of resource from the decision-making equation altogether and rather basing procurement decision on price alone, any resource that can meet the necessary market requirements can be purchased. With more 8-10%, or more, of system capacity met by DR in the PJM Interconnection, ISO New England, and the Western Australia Independent Market Operator, these markets have shown a clear ability to drive significant penetration of demand response. Outside of the established liberalized markets where DR is present, there tend to be more fragmented regulatory efforts to mitigate but not eliminate the disparity in incentives between supply-side and demand-side investments by utilities. In fact, one could argue that these multitude of policies and initiatives are required because in most areas, the underlying financial drivers that encourage a supply-

22

side-focused perspective have not been modified: utility revenue is still tied to the amount of kWh sold, and the amount of capital they invest in generation and/or network infrastructure. In many regions, utilities are only allowed to recover their DSM expenditures, but cannot earn a rate of return in the same manner as they would for supply-side investments. Because of this unequal treatment, some jurisdictions require their utilities to first pursue DSM programs before they can build generation assets to ensure solutions that may be cost-effective, but not financially beneficial, are considered. In other areas, utilities are mandated to reduce the peak demands (and energy consumption) or face penalties such as in Pennsylvania where there is no financial driver for the utility to do anything other than build more and more infrastructure. It is within this environment that the UKs equalisation incentive, is significant as it demonstrates a way to create true parity of treatment outside of a wholesale market context. While wholesale generation is competitive in the UK, distribution network operation is not they are regulated monopolies in the same manner as vertically-integrated utilities in traditionally-regulated markets. Moreover, in traditionally regulated areas, such a mechanism could be applied to all investments so that generation (or alternatives to it) were also covered. In many ways, it is the concept of the equalisation incentive that is most important, and not its exact methodology. A multitude of regulatory mechanisms could likely be developed that would result in equal financial treatment between supply-side and demand-side investments, and it is important to not prescribe specific methodologies that may be better suited for one system than another. This global survey demonstrates that good program designs are crucial to the success of demand response, perhaps more so than the existence of a formal market structures. Regardless of how DR programs or opportunities are engendered, programs must have the essential elements outlined in this paper in order to be sustainable, whether they are in liberalized markets or operated by verticallyintegrated utilities. In the wholesale capacity markets profiled in this paper, the programs found in the investor-owned utilities of California, as well as the program for the public utility TVA, clear similarities are evident. All such programs and markets are capacity-based, in which demand response resources are paid an ongoing payment for being available to provide capacity. In addition, all these examples are mainly targeted at the infrequent, yet expensive, top peak hours of the year. While there is indeed the ability for DR to provide more frequent response, such as in ancillary service markets, these general peakshaving or emergency-prevention programs are suitable for the widest number of participants and can therefore lead to the highest levels of customer penetration. Lastly, the inclusion of demand response load aggregators is another key recipe for success. In wholesale markets, often only the largest industrial customers can participate directly and aggregators are a mechanism for small and medium sized C&I customers to participate as well. Yet, even in such conditions, it is common for customers that could otherwise directly access the market do so instead through aggregators for the risk-mitigation benefits discussed in this paper. And in both the wholesale markets and among the regulated utility environments that are indeed more similar to the landscape in
23

China, we see aggregators play two other key roles that contribute to the success of DR. From the utility or system operator perspective is the ability to provide guaranteed capacity. Once reliability can be ensured, system planners and operators are subsequently able to depend on the DR resource and reduce the usage of, or construction of, supply-side infrastructure. Put another way, without these guarantees, there would be limited ability for investments in demand response that lead to opportunities for participation among end-users. Equally if not more important is the behind-the-meter expertise that aggregators offer. With specialized staff and technology able to implement repeatable curtailment strategies that do not negatively impact commercial business operations, aggregators can both identify and leverage more capacity, and achieve higher levels of customer participation.

24

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 facilitys potential to participate in Automated Demand Response (Auto-DR) and to maximize load reduction savings without affecting day-to-day business or operations. The three key ways a DR program can be implemented are: 1) 2) Manual DR: This involves manually turning off or changing comfort set points, lights, or processes or each equipment, switch, or controller. Semi-Automated DR: This involves automation of HVAC or one or several processes or systems within a facility using Energy Management Control Systems (EMCS) or centralized control system, with the remainder of the facility under manual operations. Fully Automated DR: This involves automation of an entire facility, with integration of end use loads into an EMCS and centrally managed with no human intervention.

3)

Regardless of the type, technology plays an important role in the reliable operation of demand response.

4.1. Metering and Control Solutions


Metering Granular meter data is essential to the successful operation of a demand response program. First and foremost, it is the foundation of accurate measurement and verification (M&V), which is necessary for both proper measurement of the performance of the DR resource as well as financial settlement. Ensuring that DR performs as expected requires real-time data, so that the actual consumption of participating facilities can be compared to accurate forecasts of what their consumption would have been should a dispatch not have occurred. Furthermore, real-time metering and data presentment allows for performance monitoring during a dispatch. For aggregators, this enables them to ensure their entire portfolio is cumulatively delivering the load reduction required, and if not, allows them to utilize other resources to provide the proper level of curtailment. From an end-user perspective, particularly among larger sites responding with some level of manual action, real-time data also allows for them to ensure they have taken the proper steps necessary to comply with their intended response. That said, it is important to differentiate between real-time meter data for DR and typically advanced metering infrastructure (AMI). First, while AMI can facilitate DR it is by no means a prerequisite for successful deployments. In fact, because AMI deployments are in their early stages, and often focused on the residential customer classes, most of the technology-enabled DR present today utilizes real-time meter data by the installation of additional technology. This typically includes directly accessing a utility meter through analog pulse or digital serial outputs, as well as metering/sub-metering specific loads such as a generator, and then transferring this information back to the DR aggregator using existing broadband and wireless infrastructure. Even when and where AMI is present, it may be insufficient. Most smart meters and their supporting infrastructure are designed primarily with automated meter reading, and not DR in mind. As such, it is
25

common for these new smart meters to read data every half-hour or hour, and then backhaul the consumption data once a day. Such infrequent and delayed measurements, while appropriate for AMR purposes, do not provide the needed functionality for DR aggregators whom need to ensure delivery standards are met in real time. In this manner, the installation of additional or specialized metering equipment is likely required even where AMI is present. Load Control Load control hardware is another essential component of modern-day, technology-enabled DR deployments and is often part of the same advanced metering kit that is installed on customer premises. Many customer types require some level of automation in order to be able to respond to a dispatch signal. A grocery store, for example, will typically not have an energy or facilities manager on staff able to initiate curtailment measures. Even if personnel was present, without automation, they would likely be unable to manually enact common strategies for this customer segment, including HVAC cycling, partial lighting curtailment, and anti-sweat heater (condensation) control. In other situations, it is the program requirements that require load control in order to comply with the response time. Ancillary service programs, and some bilateral utility programs, can have response times of ten minutes, or less. In fact, frequency responsive DR programs can have even shorter response times. For example, the Alberta Energy System Operator (AESO) just launched a DR program with a 200-millisecond response time. With such requirements, automation and load control is an absolute necessity. Yet even in traditional peak management programs, remote load control is increasingly being utilized for customer convenience and enhanced resource reliability. The aforementioned metering/gateway devices installed are often the foundation for initiating load control as they feature two-way communication. Such devices may toggle relays attached to specific circuits, send scripts to Building Energy Management Systems (BEMS) to begin pre-defined curtailment actions, or attach directly to industrial control equipment. Dispatch, Monitoring and Management In order to successfully leverage the metering and load control hardware described above, DR providers commonly deploy Network Operation Centers (NOCs) to utilize the aforementioned foundation technologies. It is from these NOCs that load aggregators can initiate automatic dispatch notifications to participating customers, remotely control customer loads and generation, monitor performance in order to ensure performance compliance, and coordinate technicians in the field. Centralized control centers also allow DR to comply with telemetry requirements in a cost-effective way. Some grid operators require resource in some of their markets (e.g. PJM Synchronized Reserves, National Grid STOR) to be directly integrated into their respective control rooms with remote terminal units, or other similar equipment. Such generation-grade hardware is expensive, and would be costprohibitive to deploy at individual customer sites.

26

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


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

23 24

http://collaborate.nist.gov/twiki-sggrid/bin/view/SmartGrid/OpenADR. Piette, M.A., G. Ghatikar, S. Kiliccote, E. Koch, D. Hennage, P. Palensky, and C. McParland. 2009. Open Automated Demand Response Communications Specification (Version 1.0). California Energy Commission, PIER Program. CEC-500-2009-063 and LBNL-1779E. 25 The OpenADR Primer, White paper by the OpenADR Alliance (http://www.openadr.org/).

27

Figure 6: OpenADR Structure


Box 1: OpenADR Features Continuous, Secure and Reliable Provides continuous, secure, and reliable two-way communications infrastructures where the clients at the end-use site receive and acknowledge to the DR automation sever upon receiving the DR event signals. Translation Translates DR event information to continuous Internet signals to facilitate DR automation. These signals are designed to interoperate with Energy Management and Control Systems, lighting, or other end-use controls. Automation Receipt of the external signal is designed to initiate automation through the use of preprogrammed demand response strategies determined and controlled by the end-use participant. Opt-Out Provides opt-out or override function to participants for a DR event if the event comes at a time when reduction in end-use services is not desirable. Complete Data Model Describes a rich data model and architecture to communicate price, reliability, and other DR activation signals. Scalable Architecture Provides scalable communications architecture to different forms of DR programs, end-use buildings, and dynamic pricing. Open Standards Open standards-based technology such as Simple Object Access Protocol (SOAP) and Web services form the basis of the communications model.

During a Demand Response event, the utility or RTO/ISO provides information to the DRAS about what has changed and on what schedule, such as start and stop times. A typical change would specify one or more of the following: Price signals: This would include a price multipler, a price relative, or an absoulte price Reliability signals: This would include the load amount to be shed (difference, load level, or setpoint that a load should go to). Levels: These are simple representations of the price and reliability signals such as NORMAL, MODERATE, and HIGH.

28

The standard also specifies considerable additional information that can be exchanged related to DR and Distributed Energy Resources (DER) events, including event name and identification, event status, operating mode, various enumerations (a fixed set of values characterizing the event), reliability and emergency signals, renewable generation status, market participation.

Widespread adoption of OpenADR will accelerate the successful implementation of DR programs andDER, thereby providing the following four major benefits for all stakeholders: Lower Costs Standardization lowers development and support costs for vendors and, ultimately, their utility customers. Standardization also fosters technology innovation and competition, which expands product choices for both utilities and end users. Assured Interoperability Electricity providers and consumers alike benefit from being able to choose from among a wide range of different products and services without concern for any incompatibility or inevitable obsolescence. Greater Reliability Products based on robust standards function dependably under normal circumstances and are able to recover from any anticipated error conditions to deliver dependable operation. Enhanced Flexibility OpenADR has been designed to work with existing DR equipment (socalled backwards compatibility), as well as with newer, more sophisticated systems offering advanced feature sets. Commercial, industrial and residential customers, and energy aggregators, will all be able to reduce costs, time and risk in the selection and deployment of products and systems based on the OpenADR standard. Work being performed by the OpenADR Alliance26will educate these stakeholders about the benefits of DR, and will increase their confidence in the available solutions with rigorous testing and certification programs. As a result, equipment vendors and systems integrators will be able to accelerate the time-to-market for, and lower the development costs of, innovative products and services, while electric utilities, ISOs and RTOs will gain faster access to the market, experience lower capital and operational expenditures, and achieve greater success with DR programs. Even regulatory agencies will benefit from knowing that the introduction of new pricing policies will not be undermined by incompatibilities or other end-to-end impediments in the marketplace.

4.3. Smart Meter and Advanced Metering Infrastructure (AMI) and OpenADR
As the use of OpenADR for commercial and industrial facilities has gained significant traction in California and other parts of the U.S., the Advanced Metering Infrastructure (AMI) and smart home technologies are currently being implemented on a large-scale basis in residences. The AMI system wide implementation in California residences by the utilities together with development of the supporting
26

http://www.openadr.org/.

29

technologies has provided opportunity for wide range of system operation and customer management applications, including communicating DR information through the AMI communication channels. The AMI communication is not open and accessible outside the utility network. AMI infrastructure can include smart meter, which is a revenue-qualified device from which charges can be derived. Other means of measuring power may be used, but they would generally not be qualified for revenue use from the residence point-of-view, the advanced meter contains valuable information about current and past power usage. This advanced infrastructure is however, not needed in most DR programs. While the traditional electrical meters only measure total consumption and as such provide no information of when the energy is consumed, an interval meter can usually record consumption of electricity in intervals of an hour or less and communicate that information at least daily back to the utility for monitoring and billing purposes. In some DR programs, an interval meter is all that is needed for a customer to be qualified to participate. LBNL is working with the utilities and other stakeholders to provide an external non-AMI based OpenADR interface to the residential technologies and home automation networks (HAN). These interfaces coexist with the AMI infrastructure that the utilities plan to use for their metering and billing purposes. Figure 7 shows these interfaces where OpenADR can be used as a means of communication directly with the residential gateway or the end-use devices such as the appliances:27

Figure 7: AMI-HAN Interface Further details on the home automation technologies and its use within the DR context are available from previous LBNL studies.28

27 28

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

30

Evaluation of Market Fundamentals and Challenges to Long-Term System Adequacy in Albertas Electricity Market

April 2011

Johannes P. Pfeifenberger Kathleen Spees

Prepared for

Copyright 2011 The Brattle Group, Inc. This material may be cited subject to inclusion of this copyright notice. Reproduction or modification of materials is prohibited without written permission from the authors.

Acknowledgements The authors would like to thank the AESO staff for their cooperation and responsiveness to our many questions and requests. We would also like to acknowledge the research and analytical contributions of Lucas Bressan and Robert Carlton. Opinions expressed in this report, as well as any errors or omissions, are the authors alone.

TABLE OF CONTENTS I. II. Executive Summary ...........................................................................................................1 Background ........................................................................................................................4 A. Market Designs to Address Resource Adequacy ........................................................ 4 1. Energy-Only Markets............................................................................................ 5 2. Market Designs Based on Administrative Capacity Payments............................. 7 3. Market Designs with Resource Adequacy Requirements ..................................... 9 B. AESOs Energy-Only Market Design ........................................................................ 9 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 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 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 Findings and Recommendations.....................................................................................66

III.

IV.

V.

VI.

Bibliography .................................................................................................................................68 List of Acronyms ..........................................................................................................................75 Appendices ....................................................................................................................................77 A. Generator Operating Margins versus Fixed Costs .................................................. A-1 B. Method For Projecting Operating Margins ............................................................. B-1 C. Projection of Generator Operating Margins versus Fixed Costs ............................ C-1

I.

EXECUTIVE SUMMARY

The Alberta Electric System Operator (AESO) asked The Brattle Group to review long-term challenges to resource adequacy in Albertas electricity market and assess the following four questions: 1. Is the market design sustainable in its current state? 2. Is the energy-only market design sustainable with minor changes? 3. Are major changes required to maintain resource adequacy? 4. What long-term adequacy metrics can be used as milestones for change? The challenges that Alberta will face over the coming decade include: (1) the potential introduction of new environmental regulations that could force aging plants to retire or incur significant capital expenditures; (2) the expiration of power purchase arrangements, which may trigger accelerated retirement partly due to decommission cost recovery regulations; (3) the addition of wind generation capacity, which suppresses energy prices and increases price volatility; (4) expanded interconnections with neighboring markets, which has the potential to reduce reliability in Alberta if they lead to the province becoming dependent on interties for resource adequacy, and (5) the continued long-term outlook of low natural gas and power prices, which result in low operating margins and limits investment cost recovery particularly for coal and hydro plants. Individually, each of these challenges may impose a manageable downward pressure on reserve margins and consequently upward pressure on market prices, ultimately resulting in relatively stable levels of market prices and reliability. However, the combined impact of these factors might create a resource adequacy challenge for the Alberta electricity market large enough to result in unacceptably low levels of reliability or higher, more volatile power prices. The overall challenge is amplified to the extent that the market will be exposed to all of these pressures simultaneously over a relatively short period of time. Most electricity markets around the world face a similar set of challenges, although some of these challenges are unique to Alberta. For example, most US electricity markets do not rely on market mechanisms to determine the desired level of reliability, but instead impose resource adequacy standards that ensure a specific reserve margin. By doing so, reserve capacity becomes valuable and power plants can earn revenues through bilateral or centralized capacity markets. Other power markets offer regulated capacity payments to encourage investment. In contrast, no similar capacity-related revenue sources are available to power plants or demand-side resources in Alberta. Rather, investment costs need to be recovered solely through revenue earned in Albertas energy and ancillary service markets. Note that these energy revenues may also be hedged through short-term and long-term bilaterally contracted sales, although the prices agreed upon in these contracts will be ultimately informed and driven by energy spot prices from the centralized wholesale market. This energy-only market design creates significant uncertainties about whether the market will maintain resource adequacy in the presence of the identified challenges. In fact, some other energy-only markets, such as in Great Britain, are in the midst of significant market redesign efforts to address these challenges. We find that the identified challenges will come about gradually and increase the rate of plant retirements and investment needs. However, with the possible exception of accelerated 1

retirements related to decommissioning cost recovery, the identified challenges should not result in substantial simultaneous retirements of existing plants. The rate of plant retirement will most likely average 220 MW per year over the next two decades, which is 1.5 times the 150 MW of annual retirements experienced during the last decade. Considering both these retirements as well as the anticipated load growth of 3.2% per year and an associated reserve margin requirement increase, this would require the addition of 740 MW per year over the next 20 years. This is almost twice the rate of historic generation additions, which averaged 380 MW over the past decade. We conclude that the current market design should be able to support this higher and consequently more challenging rate of generation additions. Our analysis shows that the Alberta market design is generally well-functioning, with energy and ancillary service prices that have been relatively low when reserve margins were high, but that have increased enough to attract new plant additions when system-wide reserve margins declined. We also find that the Alberta market design will likely be able to retain existing resources and attract new entry without dramatic price increases or a significant reduction in resource adequacy. Our projections of future energy and ancillary service prices based on recentlyexperienced market conditions show that only modest increases in market prices, consistent with projected increases in natural gas and carbon emission costs, should be sufficient to avoid premature retirement of existing resources and, importantly, support investments in new generation. We find that projected future market prices based on current fundamentals strongly favor a shift in the resource mix from coal generation to natural-gas-fired power plants, which are more flexible and have lower capital costs. The entry of additional wind turbines and coal plants with carbon capture and storage may be supported by government policies and through the value of green attributes. As a result, and perhaps contrary to our initial expectations, we currently see no compelling need for major changes in Albertas electricity market design. However, the outlook for resource adequacy remains uncertain and sensitive to changes in market fundamentals and continued evolution of the identified challenges, which must not be underestimated. It also needs to be recognized that an energy-only market design will not be able to guarantee that a certain reserve margin will be maintained. In fact, in a small system such as Albertas, the lack of coordination between the retirement and online dates of individual units can cause transitional reliability concerns and price spikes, as has been highlighted by the recently announced, unexpected potential early retirements of Sundance 1 and 2. Overall, we offer the following recommendations. The AESO should carefully monitor market fundamentals in light of the identified challenges. In addition to the already ongoing monitoring of resource adequacy metrics based on a 24-month outlook, we recommend monitoring: (1) trends in market heat rates and the long-term outlook for technology-specific operating margins; (2) retirement schedules and associated system reserve margins; (3) market price impacts of wind generation as more wind power plants come on line; and (4) the impact of interties as they are expanded and market rules related to the use of these interties evolve. Alberta policy makers should consider relaxing or revising the existing decommissioning cost recovery rule to reduce the risk of large simultaneous plant retirements in 2020 when most of the existing purchase power arrangements expire. More generally, policy makers 2

should avoid introducing regulations that could result in large simultaneous retirements, which are difficult to manage in any market or regulated environment. We recommend that the AESO consider increasing the current price cap from $1,000/MWh to the lower end of estimates for the value of lost load, which tend to be in the range of approximately $3,000/MWh. We also recommend reducing the price floor below zero to a level where generators, including wind plants, would have an incentive to shut down when it is economic to do so. These adjustments would also allow for economically efficient prices during reliability events, stimulate demandresponse, facilitate entry of resources at lower average annual market prices, and make the level of the price cap more consistent with those in other energy-only markets, such as Texas and Australia. Coincidentally with increasing its price cap, the AESO should consider revising its mechanism for setting administrative prices under emergency conditions when out-ofmarket reliability actions become necessary. Under these conditions, prices should be set to reflect the marginal cost of any out-of-market actions. The AESO should carefully consider the long-term resource adequacy implications of its efforts to refine the Alberta market design, which include: (1) the integration of additional wind generation; (2) refining ancillary service markets and market designs for demand response; and (3) the expansion of interconnections with neighboring systems. Overall we conclude that Albertas energy-only market is generally well-functioning and sustainable, although its efficiency and effectiveness can be improved with some design changes. However, we caution that the current positive outlook cannot guarantee resource adequacy longterm for the simple reason that Albertas market design, like other energy-only markets, does not include a resource adequacy requirement. For this reason the AESO must continue to monitor potential challenges to resource adequacy over time.

II.

BACKGROUND

The Alberta Electric System Operator (AESO) asked The Brattle Group to review long-term challenges to resource adequacy in Albertas electricity market and assess the sustainability of the current energy-only market design from a long-term resource adequacy perspective. This report assesses the possible impact of these challenges to the long-term sustainability of Albertas energy-only market and explores options that may help reduce the risk of highly undesirable outcomes. In this context, our report explores four questions: 1. Is the market design sustainable in its current state? 2. Is the energy-only market design sustainable with minor changes? 3. Are major changes required to maintain resource adequacy? 4. What long-term adequacy metrics can be used as milestones for change? Our evaluation defines a sustainable market design as one that will provide long-term resource adequacy through pricing signals that are sufficient to attract and retain capacity when needed. A sustainable design can provide an efficient level of reliability without reliance on out-ofmarket or backstop mechanisms. The scope of our analysis does not include challenges related to transmission planning, system operations, and short-term market design initiatives. A. M ARKET D ESIGNS TO A DDRESS R ESOURCE A DEQUACY Albertas energy-only market design lies within a spectrum of resource adequacy constructs that have been implemented in North America and around the world, as summarized in Table 1. Table 1 describes four different electricity market design approaches: (1) energy-only markets, which are usually accompanied by a set of ancillary services markets, but without an explicit resource adequacy requirement; (2) markets in which resource adequacy is ensured through administratively determined capacity payments made directly to suppliers; (3) markets with explicit resource adequacy requirements that mandate the procurement of reserve capacity by retail suppliers on a short-term basis (e.g., for the next peak season); and (4) market designs that mandate procurement of reserve capacity by retail suppliers on a forward basis (e.g., one to several years prior to the year when the capacity is needed).

Table 1 Spectrum of Approaches to Resource Adequacy1 Type of Centralized Capacity Market Designs without Explicit Resource Adequacy Requirement Energy-Only Markets AESO, Australias NEM, ERCOT, Great Britain, NordPool With Capacity Payments or PPAs Argentina, Chile, Colombia, Peru, Spain, South Korea, Ontario Designs With Explicit Resource Adequacy Requirement Short-Term SPP, former power pools (NYPP, PJM, NEPOOL) Midwest ISO NYISO, former PJM, Australias SWIS PJM, ISONE, Brazil Forward

None

CAISO

Voluntary Mandatory

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), Australias National Electricity Market (NEM), Australias South West Interconnected System (SWIS), PJM Interconnection (PJM), ISO New England (ISO-NE), New York Power Pool (NYPP), New England Power Pool (NEPOOL), and New York Independent System Operator (NYISO). For a full discussion of the theoretical basis for pure energy-only markets, see Hogan (2005) and Joskow and Tirole (2004). In many energy-only markets, there often are market interventions through the system operator or government entities in the case of insufficient resources. Such out-of-market interventions can take the form of backstop procurement mechanisms, government-built generation, or out-of-market cost recovery such as government-supported long-term power purchase arrangements. These out-of-market interventions damage the function of the energy-only market by artificially suppressing energy-market

levels of energy prices punctuated by occasional severe price spikes. This is because sufficient resources are available most of the time, and competitive market forces depress prices towards the production cost of the most expensive unit dispatched. These prices near marginal production costs are below the price levels needed for full investment cost recovery for marginal resources. However, there will also be occasional conditions in which supplies become scarce and energy prices increase (or even spike) to include a scarcity premium that provides generators with the operating margins needed to recover their investment and other fixed costs. These occasional price spikes must be large enough and frequent enough to allow the full recovery of fixed operations and maintenance and investment costs if capacity resources are to be attracted to and retained in the market. Revenues received from the ancillary services markets, which tend to track with prices in the energy market, also help determine when and which types of new capacity investments are attractive. While such scarcity-based price spikes are inherent to the design of energy-only markets, they can impose economic impacts on retail customers that create political challenges to maintaining the market design. However, retail suppliers have the option to hedge against the economic impact of this price volatility, a practice that is widespread in some energy-only markets, such as Australias National Electricity Market (NEM).4 For buyers and sellers that are fully hedged with long-term contracts for power, the hourly energy price has no effect other than as a settlement tool, or as a benchmark helping to determine a reasonable price for a new long-term contract. Occasional high scarcity prices also motivate demand reductions through price-responsive demand (PRD) and interruptible retail services. The price during a scarcity event must rise until supply and demand are balanced. If that happens, the scarcity price represents an economically efficient and accurate representation of the value customers place on consuming peak power and avoiding interruptions in service. Energy suppliers, likewise, have an efficient price signal indicating whether or not to invest in capacity without any administrativelydetermined resource adequacy standard. The ability to rely on customers to choose their own desired level of reliability through the marketplace, rather than relying on administrative determinations, is one of the (at least theoretical) advantages of energy-only markets. Demand can adequately adjust to balance the system during supply shortages only if: (1) a large enough fraction of the load is exposed to and is responsive to market prices; and (2) prices are allowed to rise to the high value that customers place on reliability. In most real-world energyonly markets, there is not yet sufficient price response or interruptible load to realize the theoretical model of how the market should behave under scarcity conditions. Instead, during a scarcity event, the system administrator may have to rely on out-of-market actions such as expensive off-system power purchases, voluntary emergency load shedding contracts, or resort to involuntary load curtailments. In some markets, the actions of the system operator to increase

prices and tend to be self-perpetuating. A well-functioning energy-only market should not require such interventions. See Pfeifenberger, et al. (2009), pp. 19-38. Albertas backstop reliability mechanism, instituted as part of its Long-Term Adequacy Rules, allows the market operator to intervene to procure sufficient capacity when the 2-year supply outlook is insufficient to maintain a reliability threshold, see AESO (2008). See AER (2007), Ch. 3.

supply through out-of-market actions during emergency events can actually have the undesirable effect of artificially suppressing the market price. Finally, in the extreme event of firm load shed, the market price has to be set to an administratively-determined level because the market clearing price is an undefined quantity during a rationing event. In Alberta the price is set at the price cap under such conditions. The theoretically efficient price during emergency operations is the marginal cost of the next emergency procedure. For example, if voluntary curtailments are required, the pool price should be set equal to the per-MWh cost of the most expensive load-shed contract called upon during the emergency.5 If involuntary curtailments of firm load are required, the most efficient price during the rationing event is the estimated price that the average interrupted customer would have been willing to pay to avoid interruption. This price level is referred to as the Value of Lost Load (VOLL).6 Estimates of VOLL vary widely depending partly on the makeup of the customer base and partly on uncertainty in estimation methods, but usually are at least in the range of $3,000-$10,000/MWh.7,8 Administrative scarcity pricing at the VOLL crudely approximates a demand curve for energy.9 More advanced administrative scarcity pricing schemes, as used by the Midwest ISO for example, gradually increase the price toward the VOLL as the necessity of involuntary curtailments becomes more likely.10 When the potential for exercise of generator market power is a concern, administrative scarcity pricing can also allow the system operator to maintain a generator bid cap below the VOLLbased price cap, without undermining efficiently high prices during scarcity events. For example, prices can increase to the generator bid cap as the supply stack runs out. At even higher levels of scarcity, a combination of high-priced demand bids (which can be higher than the generator bid cap) and administrative scarcity pricing can tie the prevailing market price directly to the marginal cost of demand interruptions or the marginal cost of out-of-market emergency operations. 2. Market Designs Based on Administrative Capacity Payments Some energy market designs mitigate or otherwise suppress market prices to levels far below the VOLL, such that they do not include a sufficient scarcity premium. As a result, suppliers are generally unable to recover their fixed costs solely through energy and ancillary services

6 7 8

10

For example, if the load-shed contract for a 1 MW reduction costs $10,000/year and stipulates an expected 5 hours of curtailment per year, then the hourly system-wide price for any hour when the contract is enacted should be $2,000/MWh. See Joskow and Tirole (2004) and Hogan (2005). See Centolella and Ott (2009) and Midwest ISO (2006). Note that some sectors of industry, such as mining, place an extremely high value on lost load, exceeding $50,000/MWh, see Midwest ISO (2006). However, a system-wide estimate of average VOLL does not need to include the full VOLL of these customers if they exceed the cost of private investments in back-up generation. Note that if there actually were sufficient levels of demand response and interruptibility in the market, the outcome during a scarcity event would be much more efficient because customers would self-select reductions from low-value uses of power. Under involuntary curtailments, high and low value applications for power are indiscriminately interrupted. See Hogan (2005) and Newell, et al. (2010), Section IV.A.4.

markets, resulting in missing money relative to what is needed to attract and retain sufficient capacity to meet reliability targets.11 In a market design with administrative capacity payments as shown in Table 1, the system operator makes direct payments to suppliers or signs PPAs with suppliers of capacity. The system administrator then recovers the costs associated with these capacity payments via an uplift charge assessed to customers.12 There has been great variation in the determination of administrative capacity payments and the designation of eligible suppliers. The most widely-used capacity payment design is similar to the one first implemented in Chile in 1982.13 This was an availability-based compensation mechanism under which any supplier bidding into the energy market would receive a capacity payment whether or not the unit was dispatched. These capacity payments would be set such that, over the course of the year, they would cover the annual investment costs of a peaking unit as long as the plant demonstrated sufficient availability during months of peak demand or capacity shortage.14 The major criticism of capacity payment systems is that they rely on administrative judgment rather than market forces.15 In a capacity payment system, the system administrator is extensively involved in determining the size of the payments that will be made and the type of capacity resources that would be eligible. However, the quantity that will be supplied in response of such payments can remain uncertain, which can lead to excess capacity or reliability levels that remain below targets despite the administrative payments. Maintaining target levels of resource adequacy by making administratively-set capacity payments available only to new resources is sometimes viewed as a more cost-effective solution than providing capacity payments to all resources. However, attempts to limit payments only to new resources, while implemented in some places such as Spain, will not likely result in lower costs in the long run, particularly in cases where re-investing in existing facilities would have been lower in cost than building new facilities. Such an approach also generally risks higher long-term costs because capacity payments are generally not made available to low-cost capacity supply from demand-side resources, capacity uprates, or postponed retirements. Finally, the cost of these payments is not generally reflected in market prices during peak load conditions, which means that efficient levels of demand-response cannot be achieved even in the absence of other barriers to demand-response.

11

12 13 14

15

For additional discussion and explanation of the meaning of missing money and how baseload, intermediate, and peaking capacity is affected, see Hogan (2005), pp. 2-7. See Adib, et al. (2008), pp. 336-337. See Batlle (2007), p. 4547; Larsen (2004); Rudnick (2002). In Chile, the peak demand months are May-September; in Colombia, the payments are made during the dry season of December-April when hydro capacity is limited, see p. 161, Rudnick (2002). Sometimes the capacity payments are differentiated depending on the type of resource, for example, in order to incent investments in thermal capacity after a period of drought and associated electric shortages, Colombia introduced increased capacity payments for thermal units. However, the units would have to make at least some energy margins to be profitable overall, see Larsen, (2004). For example, both the South Korean and Colombian systems have been criticized for lack of transparency and predictability. See Park (2007), pp. 5821-22; Larsen, et al. (2004), p. 1772.

3. Market Designs with Resource Adequacy Requirements The approach to ensuring resource adequacy used in most of the United States is based on reserve margin requirements imposed on retail suppliers. Under this market design, the regulator or system administrator determines the amount of capacity each retail supplier must procure to ensure resource adequacy. For example, to ensure a system-wide reserve margin of 15%, each retail supplier would be required to procure capacity amounting to 115% of its projected coincident peak load. These reserve margin requirements can be imposed on a current or forward basis. As shown in Table 1, the Midwest ISO, SPP, and NYISO require that sufficient capacity commitments are demonstrated immediately prior to each delivery period (e.g., prior to each delivery month). In contrast, PJM, ISO-NE, and CAISO all require capacity procurement on an annual or 3-year forward basis. Another key difference among these markets is whether the design relies exclusively on bilateral contracting and self-supply, or whether the system operator facilitates procurement through a centralized capacity market. While the creation of a resource adequacy requirement always creates a bilateral market for capacity, centralized capacity markets are not a necessary design element. For example, in SPP, retail suppliers procure capacity through self-supply or bilateral contracting.16 The Midwest ISO operates in largely the same way, but it also administers a Voluntary Capacity Auction (VCA) through which market participants can buy or sell capacity on a voluntary basis.17 In both Midwest ISO and SPP, retail suppliers are solely responsible for procuring capacity, and the system operator would not intervene to fill deficiencies if any existed. This is unlike California, where the CAISO will bilaterally procure capacity when needed to fill any deficiencies that remain beyond what LSEs have already procured and submitted in their capacity procurement plans. In PJM, NYISO, and ISO-NE, the ISOs procure capacity deficits through their centralized capacity auctions. Participation in the centralized market for procuring residual capacity is mandatory in PJM, NYISO, and ISO-NE.18 Under these designs, retail suppliers have the option to self-supply or contract bilaterally, and the RTO will procure any residual capacity requirements through the mandatory centralized auction and assign responsibility for payment to retail providers. Participation in the capacity auction is also mandatory for all existing capacity with competitive bid levels overseen by the market monitor. B. AESO S E NERGY-O NLY M ARKET D ESIGN Prior to deregulating, Albertas electric sector consisted of vertically-integrated regulated investor-owned utilities as well as municipalities and cooperatives. Under this market structure, supply adequacy was ensured by regulator-approved cost recovery for assets needed for

16

17 18

Member utilities in SPP are mandated to fulfill the 12% capacity margin. The RTO oversees but does not enforce this provision, with overall resource adequacy and enforcement handled by state regulators. See NERC (2008), p. 222; SPP (2009), pp. 2.2-2.4. See Midwest ISO (2009). See PJM (2009); NYISO (2009).

reliability. In the 1990s, Alberta began a deregulation initiative to create competition in the electric sector. In 1996, Alberta introduced a power pool, creating the wholesale energy market, and over 1998-2001 Alberta deregulated its electric generation fleet.19 With these reforms, Alberta transitioned to an energy-only market in which new generation investments would not be mandated by regulators but rather would be attracted by market incentives. After the first few years of experience with the energy-only market, the Alberta Department of Energy initiated a market design review to determine whether major market modifications were required for longterm adequacy, including the option of imposing adequacy obligations on retail suppliers. The review concluded that such a redesign was not necessary at the time, noting that the market had attracted more than 3,500 MW of competitive new generation between 1998 and 2005.20 However, the review did recommend an initiative toward the long-term adequacy (LTA) rules. Albertas energy-only market design is implemented along with a set of ancillary services markets including operating reserves to ensure sufficient operating flexibility. The energy and ancillary markets are also accompanied by a dispatch down service (DDS) settlement mechanism to mitigate against energy price distortions from out-of-market transmission must run (TMR) dispatch. Like other energy-only markets such as those in Great Britain, Scandinavia, Texas, and Australia, Albertas electricity market design does not offer capacity payments and does not have a mandated resource adequacy requirement. Also similarly to other market designs, Alberta has out-of-market backstop mechanisms for providing reliability when in-market signals have failed to provide sufficient supply for reliability. One backstop mechanism is the option to sign TMR contracts with generation units that are needed for locational resource adequacy or voltage stability although they are uneconomic to operate as market-based assets. The LTA rule sets out another set of backstop mechanisms that may be implemented if the two-year supply outlook appears insufficient to maintain a reliability threshold. In this case AESO can engage in out-of-market reliability contracts for load shedding, back-up generation, or the temporary installation of emergency portable generation.21 The need to rely on out-of-market backstop reliability contracts such as these could be a strong indicator of problems in the market design, which may not be providing sufficient price signals for supply investments. Out-of-market reliability contracts can also add to the reliability problem by suppressing market prices during periods of scarce supply, unless they are managed carefully. Like other energy-only markets, Alberta takes a carefully restrained approach to mitigation of market power, allowing energy prices to spike sufficiently in response scarcity events to attract and retain generation and demand response investments.22 This differs from the more heavyhanded price mitigation in U.S. and other electricity markets with resource adequacy standards, which mitigate energy market prices to much lower levels but supplement suppliers cost recovery with a capacity market or capacity payments.

19 20 21 22

See AESO (2006), pp. 4, 7-9. See DOE (2005), pp.3-4, 26-35. See AESO (2006), pp. 10-11. For example, see MSA (2010a) and (2010b).

10

However, Albertas energy-only market design and market fundamentals also differ from other energy-only markets in a number of respects. Albertas market price limits are more restrictive than in other markets, with a price floor at zero and a $1,000/MWh price cap that is far below reasonable estimates of VOLL.23 The relatively low price cap along with a high load factor and other features likely combine to limit the potential for demand response (DR), which may only choose to respond at much higher energy prices.24 Albertas centralized market is an ex-post real time market, with no day-ahead market, hour-ahead market, or centralized generation unit commitment. Importantly, Alberta is a relatively small market which naturally limits the number of market participants and the extent to which competitive locational submarkets could be maintained. The Alberta energy-only market is also surrounded by non-market-based regions with resource adequacy requirements, which creates some unusual challenges at the market seams. Finally, provincial regulations, the upcoming expiration of power purchase arrangements (PPA), and high dependence on coal under a potential federal coal retirement mandate all create challenges unique to Alberta. These unique factors also limit the extent to which experience in other markets can be directly applied in our analysis. III. LONG-TERM SYSTEM ADEQUACY CHALLENGES FACED IN ALBERTA Like other electric markets around the world, Alberta faces a series of challenges to resource adequacy over the coming decades. Existing generators will face retirement pressures from a number of directions, including the potential federal coal retirement mandate, Albertas carbon and air quality emissions standards, the expiration of PPAs for most of the coal generation fleet, and reduced operating margins caused by low electric prices. Low electric prices are driven by the economic turndown, low natural gas prices, the growth of wind power, and the potential expansion of interties with neighboring power markets where generators do not need to rely only on energy market revenues to recover investment costs. In particular, the growth of wind power and increased intertie capacity may reduce energy prices without substantially contributing to dependable capacity available for resource adequacy. We describe each of these challenges here, document the scale of impact that these challenges may have on the Alberta market, and discuss the potential resource adequacy implications. A. L OW N ATURAL G AS AND E LECTRIC P RICES The price of natural gas directly impacts the production cost and offer prices of gas generators in the wholesale electricity market. Because natural gas generators are the price-setting suppliers in many hours, the price of natural gas also has a strong impact on the market clearing price for

23

24

For example, Australias NEM currently has a price floor of -$1001/MWh (-$1000 AUD/MWh) and a price cap at their estimated VOLL of $12,512/MWh ($12,500 AUD/MWh). See AEMC (2010). Exchange rate of 1.0009 CAD/AUD is the December, 2010 monthly average exchange rate from FRB (2011). For a comprehensive review of these issues, see our review of market design for DR in AESO, Pfeifenberger and Hajos (2011).

11

electric energy.25 The close relationship between natural gas and electricity prices can be seen in Figure 1. The figure shows historic spot and futures prices for gas at Alberta Energy Company (AECO) C Hub, along with the AESO electricity prices over the same period. While the relationship is not one-for-one, the impact of natural gas prices on electric energy prices can be seen clearly during several periods of high gas prices, including early 2003, late 2005, and early 2008.26 Figure 1 Monthly Electric and Gas Prices in Alberta
$14 Historic Future $12 Gas
(Left Axis)

$200

Electric
(Right Axis)

$180 $160

$8

$120 $100

$6

$80 $60 $40

$4

$2
$20

$0 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16

$0

Sources and Notes: Historic and Future gas prices from Bloomberg (2010); AESO electric prices from Ventyx (2010). AECO C basis futures shown through 2010; AECO price in out years estimated by AESO (2010c) based on Henry Hub futures.

More recently, the combination of economic downturn and rapid increases in shale gas production have resulted in lower prices for natural gas and electricity.27 AECO C Hub prices dropped to $3.82/GJ in 2009-10 from an average of $6.14/GJ for 2003 through 2008. Coincident with this drop in natural gas prices, electric prices have also dropped to $50.86/MWh for 20092010 from an average of $70.83/MWh for 2003 through 2008. Given the changed fundamentals of the natural gas industry due to shale gas developments, these low gas prices are expected to

25

26

27

For example, in 2009, gas and cogen units submitted price-setting bids in 40% of hours while coal units submitted price-setting bids in 60% of all hours. Note that this means that gas-based generators have a disproportionately large impact on the average price. See AESO (2010e), p. 6. Note however, that gas prices are not the only or necessarily the dominant reason for many of the observed variations in electric prices. For example, the Alberta energy price spike in May 2010 was caused by transmission outages. Note also that while monthly energy and gas prices have a relatively strong correlation, hourly energy prices have a relatively weaker relationship to gas prices, with most of the volatility explained by short-term fluctuations in supply and demand. See, for example, Saur and Wallace (2011).

12

Electric Price ($/MWh)

$10

$140

Gas Price ($/GJ)

continue for the foreseeable future, as also indicated by the futures market for gas for the next several years. AECO C gas prices will make only a modest recovery to approximately $4.90/GJ by 2015 as shown in Figure 1. These low gas and electric prices have already greatly reduced the operating margins of existing and potential new generators as discussed in Section V.A.6. The impact is particularly pronounced for baseload coal generators that have low operating costs but high fixed costs. These generators require higher operating margins to cover the capital costs of new coal units and fixed costs of existing baseload coal units. Given the additional environmental challenges that coal generators face, and associated environmental upgrades that may be required to keep existing units operating, these low energy margins may not only deter new entry but also force some existing units to retire early. The likely impact that these low gas and electric prices have had on existing coal generators in Alberta are discussed further in Section V.A.6. B. E XPIRATION OF P OWER P URCHASE A RRANGEMENTS As part of the transition to a competitive wholesale electricity market, 7,600 MW or approximately 78% of the Alberta electric generating fleet were placed under PPAs in 2001.28 These PPAs were introduced to assure that generation assets built under the previous regulated, rate-of-return regime would be able to recover their costs, while still allowing for a transition to a competitive wholesale market. Under the terms of these PPAs, the original generation suppliers retained ownership of the facilities but were provided with PPAs that ensured full cost recovery through the remainder of the assets lifetime. The buyer of the PPA was obligated to make the agreed-upon payments to the asset owner and, in return, gained the right to schedule sales and collect revenues from the wholesale market.29 These PPAs expire over the 2003-2020 period. Rights to 4,460 MW of PPAs covering thermal capacity were sold at a competitive auction in August 2000, with auction proceeds returned to retail customers. An additional 2,350 MW of thermal capacity failed to sell in the auction and 790 MW of hydroelectric capacity were not placed in the auction.30 These unsold PPAs were transferred to the Balancing Pool, an entity created by the Alberta government in 1998, which manages these assets as the PPA buyer and returns any net revenues to retail customers in Alberta.31 A potential resource adequacy challenge is created by the possibility that a substantial proportion of the units currently operating under PPAs may retire after the PPAs expire. For example, some asset owners may be operating facilities that are recovering their fixed costs under the terms of the PPA even though those units would not be economically viable without the PPA payments. Additionally, asset owners need to make continuous investments into their facilities over time to

28

29 30

31

List of units originally under PPA and their MW ratings are from Appendix D, AESO (2006), AESO (2010c), and Balancing Pool (2004), p. 9. Percentage is based on a total installed fleet of 9,400 MW in 2000 and a hydro derate to 67% of installed capacity value. List of units online in 2000 and MW ratings from Ventyx (2010) and AESO (2010c). See AESO (2006), pp. 12-13, 58-60. List of units sold at auction and their MW ratings are from AESO (2006), Appendix D; AESO (2010c); and Balancing Pool (2004), p. 9. See Balancing Pool (2009), p. i.

13

maintain the assets and extend the operating lives of the plants beyond PPA termination. They may, however, choose not to make these investments if they do not expect to be able to recoup the costs once the PPAs expire. Finally, by December 31, 2018 asset owners need to determine whether or not to decommission the facility within one year of PPA expiration to be eligible for payment of decommissioning costs.32 The payment of these decommissioning costs may be a major factor in the retirement decision for units with large environmental liabilities such as ash or asbestos cleanup, particularly if these units would expect to operate only a few years beyond PPA expiration in any case. For these reasons, the expiration of PPAs is an important factor to consider when assessing long-term resource adequacy in the Alberta electricity market. Figure 2 shows the historic and future PPA expiration dates and generating capacities by unit type. The figure also shows the PPA capacity that has already retired. The experience to date shows that generation retirement after PPA expiration is a possibility. In fact, among coal units with already expired PPAs, 540 MW out of 680 MW have since retired, and among natural gasfired steam turbines (STs) all 840 MW have since retired.33 The figure shows some delayed retirement dates for some of the capacity with expired PPAs in light green and light purple for past years. However, several of these retirements may have been delayed not because they were economically viable, but rather because they were awarded temporary (non-market-based) TMR contracts by the AESO to avoid local reliability problems that would have been introduced by their retirement.34 Finally, the figure also shows the potential early retirement and PPA termination of the Sundance 1 and 2 units, which reportedly developed mechanical problems so substantial that they may not be resolved to fulfill the PPA term.35 A key problem introduced by the scheduled PPA expirations is the fact that a large proportion of them occur at the same time. Of the 5,400 MW of capacity currently still operating under PPAs, 4,300 MW of coal and 780 MW of hydro PPAs will expire on December 31, 2020.36 This large quantity of simultaneous PPA expirations represents 41% of the currently-available generation fleet, and may represent 28% of the fleet in 2020.37 Fortunately, the simultaneous retirement of all of these units after their PPAs expire in 2020 is unlikely. As discussed further in Section V.A.6, these units generally earn sufficient returns in

32

33 34

35 36

37

Units that apply to the Alberta Utilities Commission (AUC) for retirement within one year of PPA expiration are entitled to receive payments for any decommissioning costs unrecovered from the PPA or from consumers prior to PPA commencement. See Balancing Pool (2009), pp. 39-40; Alberta Government (2007b), Section 7; and Alberta Government (2003), Section 5. Retirement dates from Ventyx (2010) and AESO (2010c). For example, the Rainbow gas CTs and Rossdale gas STs received substantial non-market TMR contract payments that may have contributed to their delayed retirement dates, AESO (2010c). See TransAlta (2011). Future PPA expiration dates from AESO (2010c). Current PPA capacity number is after the potential early Sundance 1 and 2 retirement and PPA termination. Dependable capacity value of hydro and wind units derated to 67% and 0% of installed capacity respectively. Calculation is based on a current effective installed capacity of 11,730 MW and an estimated 2020 installed capacity of 17,440 assuming that future capacity will be large enough to meet projected peak load and a 15% reserve margin. List of units online in 2010 and MW ratings from Ventyx (2010) and AESO (2010c). The AESO projection of 2020/21 winter peak Alberta Internal Load is 15,162, see AESO (2010a).

14

the energy market to cover their ongoing fixed costs even at relatively low market prices. Therefore, unless faced with significant investment needs or near-term decommissioning costs, most units will have sufficient economic incentive to continue operating beyond the PPA expiration for the remainder of the economic life of the plant. Figure 2 Historic and Future PPA Expirations by Unit Type
8
Historic Future

Hydro Gas CT

6 Capacity (GW) &

Coal
2

Sundance 1 & 2
Retire Prior to PPA Expiration

Coal (Now Retired) Gas ST Gas ST (Now Retired)


2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022

Sources and Notes: Dependable capacity rating reported above for hydro is 67% of installed capacity. Unit online and retirement dates, MW rating, and future PPA expiration dates from Ventyx (2010) and AESO (2010c). Historic PPA expiration dates from Appendix D, AESO (2006). Clover Bar and HR Milner PPAs originally expired in 2010 and 2012, but are reported above at their early termination dates of 2005 and 2001 respectively, CRA (1999); p.3, Balancing Pool (2001), p.9 (2004), p. 3 (2005).

An additional challenge is that a substantial portion of these coal units likely will be forced into retirement within a few years after 2020 regardless of the PPA expiration. These retirements will be driven by a combination of factors discussed below, chiefly the pending federal coal retirement mandate, large capital expenditures that might be required to life-extend an aging unit, or capital investments that may be required for environmental upgrades. If these units would be forced into retirement within a few years of PPA expiration, there is an increased risk that the owners will opt to accelerate retirement by a few years in order to recover decommissioning costs.38 For example, the federal coal retirement mandate would force 1,170 MW into retirement over the 2020-25 period if enacted, which could lead to large simultaneous retirements if these facilities were to accelerate retirement to recover decommissioning costs.

38

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

15

Overall, these factors combine to introduce a substantial risk of a step change of an unusually large number of retirements in 2020 and 2021. If these retirements were phased in on a more gradual basis, it would be less challenging for market-based investments to replace the units without introducing temporary reliability problems. A large step change in the number of retirements may be too abrupt for the market to absorb without administrative intervention. The number of retirements in 2020 thus should be monitored and the decommissioning cost rule stipulated in the Power Purchase Arrangements Regulation may have to be reexamined and relaxed to spread retirements over several years.39 The AESOs forward-looking supply adequacy review, which summarizes suppliers announced retirement and online dates, will also be a helpful mitigation factor. However, the AESO cannot modify announced retirements without a resource adequacy requirement or market interventions. C. A LBERTA AND F EDERAL C ARBON L EGISLATION Both Alberta and the Canadian federal government have greenhouse gas (GHG) reduction goals that could substantially affect plant retirement, resource adequacy, and the operation of the energy-only market over the next 20 years. The Federal GHG reduction target is a 17% reduction below 2005 levels by 2020. This compares to a less ambitious Alberta reduction target of 21% above 2005 levels. Albertas major carbon policy initiatives are $2 billion in investments in carbon capture and storage (CCS) technology and the Specified Gas Emitters Regulation. Both of these efforts as well as the implications of the overall GHG strategy are discussed below. Federal policy on GHG has yet to be codified, although the recently proposed strict carbon emissions standard for coal would effectively require either a CCS retrofit or retirement, and could significantly impact resource adequacy in Alberta as discussed in Subsection III.C.2. 1. Alberta Carbon Policy The government of Alberta has a Climate Change Strategy for reducing GHG emissions in the province as a whole, which will require large contributions from the electricity sector. The two current initiatives that may have the largest impact on the wholesale electricity market are the carbon capture and sequestration objectives and the Specified Gas Emitters Regulation a. Alberta Climate Change Strategy In January 2008, Alberta Environment published its climate change strategy, laying out a policy framework for reducing GHG emissions in the province.40 The strategy sets a GHG reduction target of 15% below a business-as-usual (BAU) case by 2020, and 50% below BAU by 2050. This is equivalent to 21% above 2005 CO2-equivalent (CO2e) output levels in 2020 and 14% below 2005 levels by 2050.41 Alberta Environments strategy includes a 139 MT CO2e

39

40 41

See Alberta Government (2007b), Section 7. Note that AESO does not have authority to revise the decommissioning cost recovery rule, which may need to be reviewed by the Department of Energy and the Alberta Utilities Commission. See Alberta Environment (2008). The unit for measuring GHG emissions used by the Alberta government and in this report is tonnes of CO2e. Under this unit non-CO2 greenhouse gases are converted into the equivalent global warming potential of CO2. For the electric sector, the non-CO2 emissions covered are methane (CH4) and nitrous oxide (N2O), which typically contribute approximately 0.6% of the total CO2e emissions for coal

16

reduction (70% of total reductions) through CCS, a 37 MT reduction (19%) through greening energy production, and a 24 MT reduction (12%) through conservation and energy efficiency as shown in Figure 3.42,43 Figure 3 Alberta Climate Strategy GHG Reductions Plan
450 400
MegaTonnes of CO 2 e &
Carbon Capture and Sequestration Greening Energy Production Conservation and Energy Efficiency

450
Business as Usual

400 350

350 300 250


2005 GHG Output
50 MT 139 MT

200 MT 300

37 MT 24 MT

250 200 150 100

200 150 100 50 0 2005 2010 2020 2030


Historic Electric Emissions

Alberta Plan

50 0 2040 2050

Sources and Notes: Alberta Environment (2008), pp. 23-24. Historic calculated from AESO (2010c)

A large fraction of these emissions reductions will be achieved within the electric sector, which accounted for 44.1% of Albertas registered GHG emissions as of 2008 as shown in Table 2, although only approximately 21% of total emissions as shown in Figure 3.44 Table 2 shows that the utilities sector contributes more registered emissions than any other sector.45 The high

42

43 44

45

generators and 1.0% for gas generators. Calculation based on emissions rates from Alberta Environment (2010a). Id., pp. 23-24. Year 2020 50 MT reduction was explicitly reported, but percentage numbers are estimated from a graphic representation. One MT is equivalent to one million tonnes or one megatonne. Total Alberta 2008 emissions of approximately 243 MT from visual inspection of figure in Alberta Environment (2008), pp. 23-24. Total reported and unreported 2008 electric sector emissions were approximately 51.4 MT as explained in footnote 45. While the utilities sector is almost totally comprised of electric generation plants, emissions from each sector are not strictly separated in all cases. In particular for cogeneration units, the electric-related emissions and industrial process emissions are generally reported together and may be included under either utilities or under another industry such as oil sands mining or petroleum refining. See Alberta Environment (2010a). In an independent calculation of the sector GHG emissions based on AESO data and separating out the cogen emissions attributable to electricity generation, the electric sector emissions

17

registered proportion from electricity is partly because more than 99% of the GHG emissions in the electricity sector are from large point sources emitting more than 100 kT/yr, while sources from some other sectors, such as transportation, are from diffuse sources and therefore are not covered under the current reporting rules.46 Including unregistered emissions, the electric sector accounted for only approximately 21% of Albertas total GHG output in 2008 as shown in Figure 3. Table 2 Alberta Registered GHG Emissions by Sector, 2008
Sector Chemical Manufacturing Coal Mining Conventional Oil and Gas Extraction Mineral Manufacturing Oil Sands In Situ Extraction Oil Sands Mining and Upgrading Paper Manufacturing Petroleum Refineries Pipeline Transportation Utilities Waste Management Total Number of Total Sector Percent of Reporting Total Emissions, kT Facilities 15 3 29 6 13 5 4 3 4 26 1 109 10,270 497 6,845 2,403 10,927 23,848 478 3,862 2,797 48,903 90 110,921 9.3% 0.4% 6.2% 2.2% 9.9% 21.5% 0.4% 3.5% 2.5% 44.1% 0.1% 100%

Sources and Notes: Alberta Environment (2010a), p. 7. Total Alberta GHG emissions are not represented in this table; only facilities outputting more than 100 kT of CO2e annually must report their emissions.

Meeting these targets of 15% CO2e reductions below BAU by 2020 and 50% below BAU by 2050 will have a large impact on Albertas generation fleet and its energy-only market. Many of the impacts can only be inferred, however, because the measures that will be enacted to meet these goals have not yet been specified. Nevertheless, the most immediate impacts on Alberta wholesale electricity prices and resource adequacy will come from the 2020 goals, toward which the electric sector may have to contribute approximately 15 MT of reductions from CCS, 4 MT from greening production, and 3 MT from efficiency.47 These reduction targets may have the following impacts:

46

47

were approximately 51.4 MT in 2008, 51.0 MT of which were from units emitting more than 100 kT/yr. This calculation would put electric sector emissions at 45.9% of all registered emissions in Alberta, AESO (2010c). One kT is equivalent to one thousand tonnes or one kilotonne. One MT or megatonne is equal to 1,000 kT or one million tones. Based on approximate BAU emissions estimate, see Footnote 42. Assumes that contributions toward CO2e reductions by category are the same in 2020 as in 2050, or 35 MT CCS, 9 MT greening production, and 6 MT efficiency over all of Alberta. Electric sector reductions are assumed to be achieved in proportion to

18

Carbon Capture and Sequestration Achieving a 15 MT reduction of CO2e in the electricity sector by 2020 may require approximately 3,620 MW of CCS-enabled coal generation, compared to a coal fleet of 5,780 MW in 2010. The scale and implications of this goal and current large scale CCS projects are discussed in the next subsection. Greening Energy Production The largest initiative enacted to date toward achieving 4 MT of CO2e reductions through greening energy production is Albertas Specified Gas Emitters Regulation. This regulation requires GHG reductions below a per-unit historic baseline or else requires payments on excess emissions as discussed further below. Alberta is investing revenue collected under this regulation into carbon-reducing programs and renewable energy sources, the potential AESO impacts of which are discussed further in Section III.E. Energy Efficiency and Conservation The Climate Change Strategys efficiency goal amounts to approximately 4% consumption reductions below BAU by 2020.48 This goal could reduce energy load growth from the forecasted 4.6% to 4.1% annually between 2010 and 2020.49 While large efficiency gains would tend to reduce wholesale energy prices and relieve resource adequacy concerns if implemented quickly on a large scale, the gradual introduction planned is unlikely to substantially impact the Alberta energyonly market either in terms of prices or resource adequacy. Overall, however, Albertas climate change strategy may require significant changes to the makeup of the generation fleet, impacting the wholesale electricity prices and resource adequacy. b. Carbon Capture and Sequestration The Government of Alberta has awarded $2 billion in financial commitments to developing four large CCS projects in Alberta, along with $526 million in federal investments as shown in Table 3. Together, these four projects are expected to achieve 5 MT of annual CO2 sequestration by 2015.

48

49

the electric sectors share of registered GHG emissions, or 44.1% of the total as shown in Table 2. In reality, the share of reductions required from the electric sector may be greater than from other sectors. Percentages assume that efficiency gains in the electric sector will be proportional to its share of the registered Alberta GHG emissions or 3 MT by 2020. From Table 2, the registered CO2e rate in the electric sector was 48.9 MT in 2008 over 69,947 GWh of AIL from AESO (2010a). At this same emissions rate of 0.70 kT/GWh, a 3 MT reduction by 2020 would require a 4,291 GWh reduction in AIL by 2020 or 3.9% of the current projection of 108,638 GWh. Compound annual growth rates calculated from AESO projected Alberta Internal Load (AIL) energy of 72,459 GWh in 2010 to 113,652 GWh in 2020 and an alternative 2020 load reduced by 2%. AESO (2010a).

19

Table 3 Large-Scale CCS Projects under Development in Alberta


Project Description Online Date Generation Government Capacity Awards, $M 300 MW $285 Alberta Annual CO2 Sequestration 1.3 MT

Swan Hills Synfuels

Alberta Carbon Trunk Line

[1] Underground coal gasification. Above ground separation of syngas from CO2 and impurities. Power production from syngas in separate combined cycle facility. [2] 240 km CO2 pipeline from near Fort Saskatchewan south to Clive, to be used for EOR. Initial CO2 will come from Agrium Redwater Complex and the North West Upgrading facility once completed. [3] Scotford bitumen upgrader facility.

2015

2012

n/a

$495 Alberta 14.6 MT Capacity $63 Federal 1.8 MT Initially

Shell Quest Project

2015

n/a 450 MW

Project Pioneer on Keephills 3 [4] Carbon capture retrofit to coal plant. 2011 Power Output 2015 Carbon Capture

$745 Alberta $120 Federal $436 Alberta $343 Federal

1.2 MT 1 MT

Sources and Notes: [1 - 4] Alberta Government (2010a); Natural Resources Canada (2010). [1] Alberta Government (2009). [2] Enhance Energy, (2010). Agrium fertilizer plant retrofit is planned to provide 0.25-0.55 MT of CO2 annually, Agrium (2010). North West Upgrading will produce 1.3 MT annually from each of 3 identical phases with Phases I and II online 2013 and 2018, North West Upgrading (2010a-b). 1.8 MT initial output assumes both Agrium and North West Upgrading Phase 1 are operational. [4] TransAlta (2010).

Two of these planned CCS projects are planned for new coal generation facilities with a total capacity of 750 MW and an expected 2.3 MT of total annual CO2 sequestration. Once energy consumption of the CCS equipment is accounted for, these projects may contribute approximately 1.7 MT of net avoided CO2 or 11% of the 2020 GHG reduction target for the electric sector.50 If the rate of avoided CO2 emissions can be improved on future projects to 81% below the emissions rate of a new coal plant without CCS, then an additional 2,880 MW of CCSenabled coal generation may have to be built or retrofitted by 2020.51 Combined with the projects currently under way, the potential 3,630 MW of CCS-enabled coal generation by 2020 compares to an existing coal fleet of approximately 5,780 MW as of 2010. This ambitious CCS goal represents a massive build-out of capacity that will likely be too aggressive to achieve. However, if this target is met, CCS-enabled coal will represent two thirds of the current coal fleet

50

51

A fraction of the CO2 sequestered is not counted toward net avoided CO2 emissions. Net avoided CO2 emissions are approximately 72%-76% of captured CO2 emissions for pulverized coal plants because CCS technology consumes power itself, and therefore decreases the net plant capacity rating for power deliverable to the grid. See IPCC (2005), Table 8.3a. Calculation assumes that 15 MT of the 2020 CCS coal must be met in the electric sector, or a fraction proportional to the currently registered emissions from the utilities sector, of which 1.7 MT will be met by the two electric projects already funded as described in Table 3. Also assumes that 620 kg CO2/MWh can be avoided and units would operate at 85% capacity factor. See IPCC (2005), Table 8.3a.

20

and about 21% of the entire Alberta generation fleet by 2020. For comparison, coal currently accounts for 49% of the generation fleet.52 While this estimate of the required CCS-enabled coal generation is only a rough approximation of the investment required to meet the Climate Strategy targets, it can be used to infer the scale of impacts on the AESO electricity market. Large governmental investments in CCS-enabled coal over the coming decade could boost resource adequacy in Alberta by supporting new generation additions or possibly enabling the retrofit and refurbishment of coal units that otherwise would be retired. These CCS-enabled coal plants may also impact wholesale electricity market prices by operating as must-run units to achieve high levels of CO2 sequestration. If operating as must-run generation, they are likely to bid into the wholesale energy market at or near zero, thereby tending to suppress market prices. During peak hours, this price suppression may not be a problem, especially if peak prices are allowed to rise to levels that can support new entry. During off-peak hours, however, this addition of must-run units could potentially increase the frequency of surplus supply conditions. At these times, wholesale electricity prices can drop to zero and must-run units will operate at a loss because they are unable to reduce output without incurring even larger shutdown-related costs. During some low-load conditions, the AESO must force these units to ramp down or shut down to maintain system stability, regardless of additional costs. Note that the efficient market price during such events would be negative because generators would rather pay some amount (up to their shut-down-related costs) than be forced to reduce output further, as discussed further in Section III.D. The economics of must-run coal, cogeneration and, increasingly, wind generation already result in occasional surplus supply conditions. Due to the low dispatch flexibility of CCS plants, the frequency and severity of these conditions could increase as CCS generation expands.53 c. Specified Gas Emitters Regulation Albertas Specified Gas Emitters Regulation went into effect July 1, 2007, requiring emissions reductions from all Alberta facilities outputting more than 100 kT of CO2e annually.54 These facilities were assigned a GHG emissions intensity reduction target of 12% below their baseline output established over 2003-2005.55 For electric generators, this target is a requirement to reduce the quantity of CO2e emitted per MWh produced. In order to comply with the regulation, facilities have four options: Improve the efficiency of operations to reduce per-unit output by 12%,

52

53 54 55

Year 2020 percentage assumes installed capacity will be equal to projected Alberta Internal Load of 15,160 MW plus a 15% reserve margin; current effective installed capacity is 11,730 MW assuming that hydro dependable capacity is 67% of installed capacity and wind dependable capacity is 0%. From AESO (2010a), AESO (2010c). For a full analysis of minimum generation conditions in AESO, see AESO (2010b). See Alberta Government (2007a), pp. 5-7. A new units baseline is determined from the 3rd year of operations, with the efficiency requirement ramped up to the full 12% by the 9th year of operations. See Alberta Government (2007a), pp. 7, 17.

21

Contribute $15/tonne of CO2e to the Climate Change and Emissions Management (CCEM) Fund, which invests these funds in projects to reduce emissions elsewhere, Purchase offset credits for CO2e emissions from Alberta-based projects that reduced output but are not covered by the regulation, or Purchase performance credits from other Alberta GHG emitters that exceeded their 12% GHG reduction target.56 Year 2009 program results show that Alberta-wide reductions targets were approximately 11 MT in total and approximately 4.9 MT for the electricity generation sector. Of the total from all sectors, 38% or 4.2 MT of the reductions target were met via payments to the CCEM, 28% or 3.1 MT were met by either operational improvements or performance credits to facilities covered by the regulation, and the remaining 34% or 3.8 MT were supplied by Alberta-based CO2e offsets as shown in Figure 4. Figure 4 Alberta GHG Specified Gas Emitters Regulation Compliance by Category
CO2 e Reductions or Payments, MT & 12 10 8 6 4 2 0 2007 (half year)
Sources and Notes: Alberta Government (2010c).

CCEM Fund Payments

Alberta Offsets

Performance Credits Operational Improvements 2008 2009

These compliance requirements put a cost burden on large GHG emitters participating in the Alberta market. As of 2009, there were approximately 4,090 MW of natural gas and 6,060 MW of coal plants subject to this regulation, representing 85% of Albertas generation fleet.57

56 57

See Alberta Government (2010b), pp. 2-3. Based on 2009 effective installed capacity of 11,920 after accounting for a 67% net dependable capacity rating for hydro and 0% dependable for wind. Determination of units covered by the regulation is based on a calculation of estimated GHG output in 2009 for each unit from AESO internal generation data and estimated heat rates, AESO (2010c), Ventyx (2010).

22

Covered suppliers are likely to pass these increased production costs through to the wholesale electricity market in the form of increased offer prices to the extent that their offer prices are based on their marginal production cost rather than strategic bidding. To scale the total impact that this regulatory requirement may have on the market, we can examine a case in which suppliers meet their entire regulated efficiency reduction through $15/tonne CO2e payments. Table 4 shows the approximate impact that paying full price for these emissions would have on the production cost for gas and coal units. As the table shows, if the full $15/tonne compliance cost is paid, then this regulation increases production costs by approximately 2% for natural gas-fired combustion turbine (CT) and combined cycle (CC) plants, and by approximately 13% for coal plants. Table 4 Approximate Production Cost Impact of Alberta CO2 Emissions Regulation
Gas CC Assumptions Fuel Cost, $/GJ Heat Rate, GJ/MWh GHG Rate, kg/GJ CO2e Cost, $/tonne Fraction of CO2e Output Charged Costs, $/MWh Fuel VOM CO2e Total Cost w/o CO2e Charges Total Cost w/ CO2e Charges % Cost Increase w/ CO2e Charges Gas CT Coal

[1] [2] [3] [4] [5]

6 7.7 56.6 $15 12%

6 13.2 56.6 $15 12%

1 10.4 103.1 $15 12%

[6] [7] [8] [9] [10] [11]

$46.29 $2.22 $0.79 $48.50 $49.29 1.6%

$79.43 $3.84 $1.35 $83.26 $84.61 1.6%

$10.37 $4.93 $1.92 $15.30 $17.22 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 23

volatility of gas prices.58 Given the small scale of this impact, it appears that the Specified Gas Emitters Regulation is unlikely to substantially impact retirements or resource adequacy. The cost of emitting CO2e and the portion of the sectors output that is covered by the regulation would have to be increased substantially before the regulation would have a material impact on the economics of existing units or the wholesale electricity market. 2. Federal Carbon Policy Federal GHG reduction goals are substantially more ambitious than Albertas. The Canadian federal government has committed to reducing GHG output to 17% below 2005 levels by 2020, compared to the Alberta goal of 21% above 2005 levels by 2020.59 The federal government announced this regulatory framework for GHG emissions reductions in 2007, but the framework has not been translated into binding regulation in time to meet the original 2010 reductions goals. Although the Alberta regulation discussed in Section III.C.1.c remains the only binding GHG regulation currently affecting Alberta, the more ambitious federal commitment highlights the possibility of substantial federal mandates. A federal policy that would have a large impact on Alberta is the coal generation performance standard proposed by the Minister of the Environment on June 23, 2010, for which draft regulations may be published in spring 2011.60 The proposed regulation would effectively phase out all coal generation in Canada without CCS. The proposal would require coal plants to meet a strict performance standard based on CCS technology, or else retire after the later of its 45th operating year or PPA expiration. In order to build new coal plants or extend the lives of existing coal plants, operators would have to meet a GHG emissions performance standard of approximately 360 to 420 kg of CO2e/MWh, putting it in the range of the emissions rate of a natural gas-fired CC or a CCS-enabled coal unit with an overall 50% rate of net avoided CO2e.61 The entire Alberta coal fleet would be affected by this regulation, but the impact would be phased in over the next twenty years, as shown in Figure 5. The figure shows Alberta coal capacity that would be subject to retirement under the regulation. These retirements add up to an overall retirement rate of approximately 210 MW per year starting in 2015. This is 1.5 times the retirement rate observed in Alberta over the past decade.62 The retirements driven by this

58

59 60 61

62

For example, in 2009 daily AECO C gas prices had a standard deviation of 26% of the average annual value, while monthly gas prices had a standard deviation of 25% of the average. Bloomberg (2010). See Environment Canada (2010a). See Environment Canada (2010a-b). Note that the total rate of avoided CO2e is lower than the rate of captured CO2e because of the efficiency losses associated with CCS. For comparison, the emissions rate of a typical gas CCs is approximately 344 to 379 kg/MWh and the emission rate for a new coal unit without CCS is approximately 736 to 811 kg/MWh. See IPCC (2005), Table 8.1. Over 2001 through 2010, the annual retirement rate in AESO has been approximately 150 MW per year. AESO (2010c). Note that these numbers are different from those reported on page 2. Page 2 reports the total retirements over 2011-29 projected by AESO in their long-term planning activities, including Sundance 1 and 2 and some adjustments to assumed retirements timing as informed by factors other than just the federal coal mandate. The numbers here cover a shorter time span, exclude Sundance 1 and 2, and include only retirements that would be driven by the federal coal mandate over 2015-29.

24

potential federal mandate, along with retirements that may occur for other reasons, would noticeably increase the rate of new plant additions required to maintain resource adequacy. Figure 5 Alberta Coal Units Subject to Proposed Federal Coal CO2 Emissions Standard
4.0

3.5

3.0 Capacity (GW) &

2.5

2.0

Units Subject to Retirement Mandate

1.5

1.0

Sundance 1 & 2
Retire Prior to Mandate

0.5

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029

Sources and Notes: Sundance 1 & 2 retirement mandates would have been in 2017 and 2018 respectively. No units will be under PPA past their 45th year. Terms of retirement mandate from Environment Canada (2010a-b). Unit online date and MW rating from Ventyx (2010) and AESO (2010c).

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

63 64

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

25

after the 60th operating year for gas peaking units.65 The allowed emissions rates are set based on a determination of the best available technology economically achievable (BATEA), which may improve over time and will therefore result in more strict emissions standards over time. These emissions standards somewhat increase the costs of building new units by requiring that new generators have pollution controls. For aging units past their design life, these standards are likely to require a retrofit installation of emissions controls for the unit to continue operating. For many units, the costs associated with these upgrades may be too high relative to potential going-forward operating margins to remain viable. For this reason, these control standards could force some aging units to retire before these controls would need to be installed. Figure 6 shows the timeline over which the existing AESO coal and gas fleet will be subjected to these SO2, NOx, and PM standards. This 50-year coal retirement timeline corresponds to a 5year delay relative to when the federal coal retirement mandate would force coal plants to retire. For this reason, if the federal coal mandate is enacted, the Alberta air quality standard will have no incremental effect on coal retirements or resource adequacy. In either case, the standard will have an incremental impact on natural gas units past their 40th operating year or past the 60th operating year for peaking units. Overall, these standards could force 1,630 MW of coal and 460 MW of natural gas plant retirements by 2029.66 While the quantity of capacity affected is large, the resource adequacy impact is likely to be very limited because of the gradual timeline and the imposition of standards only on older units that are already past their design life. Albertas Mercury Emissions from Coal-Fired Power Plants Regulation is an additional emissions standard based on output levels under BATEA.67 However, the mercury standard is imposed on all existing generators at the same time regardless of the unit age. In order to comply with the mercury standard, coal generators had to meet the following requirements by the first of the year: Continuous monitoring equipment installed by January 1, 2010 70% mercury capture by January 1, 2011 80% mercury capture by January 1, 2013 The mercury standard was implemented with some flexibility that allowed some older units to avoid installing mercury controls as long as they committed to retiring by unit-specific deadlines over the 2012-17 timeframe.68 Of these, Sundance 1 and 2 may already be considered retired for unrelated reasons, but all other coal facilities in Alberta will meet the mercury standard.69 The recent CASA review contained recommendations for increasing flexibility in meeting the requirement by allowing credits for early reductions that could later be used at the same facility

65

66

67 68

69

Peaking units are regulated based on an annual total emissions limit that assumes approximately 1500 MW of operation per year, see CASA (2010), Sections 3 and 6; Alberta Environment (2005). Note that Sundance 1 and 2 are excluded from this total as they may have already retired, pending determination of whether repowering will occur. See Alberta Government (2006). Specifically, HR Milner would have had to retire by 2012, Battle River 3 and 4 by 2015, and Sundance 1 and 2 by 2017. See Alberta Government (2006), p. 4. Confirmed via personal communication with Alberta Environment staff director of the mercury program. Note that Sundance 1 and 2 will retire early in advance of the air quality mandate because of unrelated large investment costs that would be required for continued operation. See TransAlta (2011).

26

if it had equipment problems.70 Overall, it appears that the mercury standard has been successfully implemented without imposing any resource adequacy concerns and even without imposing any incremental retirements. Figure 6 Capacity Subject to Provincial Air Quality Emissions Standards
3.0

2.5

Gas Peakers Gas


(Non-Peakers)

Capacity (GW) &

2.0

Coal
1.5

1.0

0.5

Sundance 1 & 2
Retire Prior to Air Quality Mandate

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029

Sources and Notes: Peaking units identified as units operating at less than 17% capacity factor, see p. 4, Alberta Environment (2005) and AESO (2010c). Unit online date and MW rating from Ventyx (2010), AESO (2010c), and Alberta Environment (2005).

E. W IND I NTEGRATION AND A NCILLARY S ERVICES Wind generation capacity in Alberta has increased quickly over the past decade, from 30 MW in 2000 to 630 MW in 2010 as shown in Figure 7. These increases in wind capacity can be expected to continue, with 240 MW of wind currently under construction and another 1300 MW either permitted or proposed. While many of the proposed or even permitted projects may never get built, they do indicate a high level of continued investor interest. The large increases in wind penetration have been driven by a variety of policies subsidizing wind. For example, wind suppliers are eligible to create and sell Alberta-based GHG offsets under the Specified Gas Emitters Regulation discussed in Section III.C.1.c and are now able to sell renewable energy

70

Assuming the recommendations are adopted, 50% of the early reductions above 75% capture would earn a credit starting 2011, while 50% of reductions above 80% would earn a credit starting 2013. These credits could not be used to delay controls upgrades or transferred to other facilities, but they could be used to offset excess emissions caused by maintenance or operational issues. All credits would expire by the end of 2015. See CASA (2010), p. 4.

27

certificates to allow utilities in California to meet the states ambitious renewable energy standards.71 Figure 7 Historic and Potential Future Wind Capacity Growth
2.5

Historic Projected

Wind Nameplate Capacity (GW) &

2.0

Proposed Additions
1.5

1.0

Permitted Additions Under Construction

0.5

Historic Additions

0.0

Winter 2000/01 Capacity 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Sources and Notes: Unit online and retirement dates and MW rating from Ventyx (2010) and AESO (2010c). Units under construction, permitted, and proposed from AESO (2010d).

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

71 72

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

28

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

73 74

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

29

Figure 8 Price Duration Curve at Different Levels of Wind Output


$500 $450 $400 $350 AESO Price ($/MWh) $300 $250 $200 $150 $100 $50 $0 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Higher Prices Wind < 100 MW

All Hours 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 AESOs qualified provider list. Note that the total capability for reserves is less than the sum of the capability for each individual type of reserves because suppliers cannot supply their maximum capability for more than one type of reserves at a time. Overall, the fleet capability is 6 times the currently

75

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

30

required level of regulating reserves, 9 times the required spinning reserves, 12 times the required supplemental reserves, and 4 times the total simultaneously required reserves. Table 5 Operating Reserves Need and Fleet-Wide Capability
Average Scheduled in 2009 Active Standby Total MW MW MW Regulating Spinning Supplemental Total Sources and Notes: AESO (2010c). 160 243 243 646 131 106 37 274 290 350 280 920 Fleet-Wide Capability MW 1,785 3,148 3,502 3,780

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

76 77

See AESO (2010h), pp. 15-16. This option is discussed in a forthcoming Brattle study for AESO, Demand Response Review, by Johannes Pfeifenberger and Attila Hajos.

31

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

78 79

80 81 82 83

See AESO (2009a), pp. 10, 14, 26-27. The north-south transmission constraint between Calgary and Edmonton is the primary constraint reducing intertie ATC values to below design ratings. Both the BC Hydro and Saskatchewan interties are connected south of AESOs north-south transmission cut plane, which places a limit on the total imports and exports that can be scheduled. The loss of an intertie would represent too large a contingency for the system south of the cut plane to absorb by itself. From communication with AESO staff and AESO (2009a), Section 4.3. Current value from AESO staff; design rating from AESO (2009a). Current value from AESO staff; design rating from AESO (2009a), p. 303. Based on communication with AESO staff and AESO (2009a), p. 39. The expansion paths being considered include: southern Alberta to Saskatchewan and Manitoba, southern Alberta to the US Pacific Northwest, northern Alberta to northern BC, and northern Alberta to northern Saskatchewan. See AESO (2009a), p. 36 and Section 4.9.3.

32

priced locations. Because Alberta is typically an exporter at night and an importer during the day, expanded interconnections may increase off-peak prices while decreasing on-peak prices.84 The combined impact from both effects is likely to be a suppression of Alberta prices overall, given that Alberta prices are higher on average than are those of its neighbors as shown in Figure 9. Part of the reason for the lower energy prices in neighboring markets is that these other markets are cost-of-service regulated, meaning that suppliers recover the capacity costs of their generation through regulated retail rates or through public ownership as in British Columbia rather than solely through wholesale market prices for energy. Other more distant markets, such as California, have resource adequacy requirements that allow suppliers to earn capacity payments through a bilateral market to supplement their energy market revenues.85 As a consequence, the energy prices of neighboring regions typically reflect only short-term generation dispatch costs without sufficient contributions to recover investment costs. This is in contrast to AESO, where suppliers need to recover their investment costs through the wholesale energy and ancillary service markets.86 Price suppression in Albertas energy-only market through expanded interties is likely to be magnified by an increase in imports from zero-marginal-cost technologies, such as new wind generation. For example, the MATL developer has predicted that adding the new transmission line will allow for the development of a large wind farm at its source in Montana.87 Similarly, increased intertie capacity with BC Hydro will interconnect Alberta more heavily with a market that is planned to become a large exporter of green power, likely from wind and hydro, which could further depress Alberta energy prices because of lower energy prices in British Columbia.88 These low-marginal-cost imports would directly benefit Alberta customers in the near term. However, in the long term, the price suppression would reduce the profitability of generation resources in Alberta, which would make it less likely that new resources would be built while increasing the likelihood that existing generators would retire prematurely. These impacts could tend to reduce the reserve margin within Alberta and make the system more dependent on the interties for resource adequacy.

84 85 86

87 88

See AESO (2009a), p. 50. See CPUC (2006); Blakes (2008), Section VI. See Section II.A for more discussion of the difference between energy-only markets and markets with resource adequacy standards. See Puckett (2009). See BC Government (2010), Section 2.n.

33

Figure 9 Monthly On-Peak Energy Prices in Alberta, Northern California, and Mid-Columbia

Sources and Notes: Bloomberg (2010); Ventyx (2010); and UCEI (2003). Prices for North Path 15 (NP15) are zonal prices until April 2009, after which NP15 Gen Hub is shown. NP15 and Mid C prices converted to CAD from USD, Bloomberg (2010).

While expanded interconnections increase the capability for importing power, they do not guarantee that external supplies will be available for import when they are needed most. This is because generators in neighboring markets tend to be obligated to supply their local customers during peak load conditions. If such peak load or emergency conditions occur simultaneously in Alberta and neighboring markets, Alberta will not be able to import the needed supplies no matter how much intertie capacity is available and no matter how high the AESO price. It also means that the resource adequacy value of the increased interties is limited to a probabilistic value that depends on the extent to which neighboring markets are over-built beyond their own resource adequacy requirements and the extent to which those markets experience system peak conditions at times different from Alberta. In addition, shortage conditions in neighboring markets can introduce supply shortages in Alberta. Because Alberta generators do not have the obligation to serve Alberta load, they have the option to export power to neighboring systems even during peak conditions and would rationally choose to do so any time external power prices are higher.89 The converse is not true, however. Because generators in neighboring markets will typically not be able to sell power into

89

This impact applies to peak conditions only prior to the initiation of emergency procedures, as AESO will intervene to curtail exports to zero during peak load conditions. See AESO (2010f).

34

Alberta during emergencies, this means that neighboring markets are largely insulated from any resource adequacy challenges in Alberta.90 The scale of the impact that shortages in external markets may have on Alberta can be gauged to some extent by examining the level of correlation that already exists between prices in Alberta and those in neighboring markets. Figure 9 shows electricity prices in Alberta compared with electricity prices in Northern California and Mid-Columbia in Washington, the two closest liquid power trading hubs with transparent market prices. While the figure shows a strong correlation among the electric prices in the three locations, this is somewhat misleading since a significant portion of the common price movements starting in 2001 have been driven by changes in the regional price for natural gas.91 The extreme high prices in the year 2000 affecting all three locations were caused by the California power crisis.92 These extreme price conditions during 2000 were driven partly by tight supply conditions and partly by market power abuses. The market power abuses have been the subject of substantial litigation, investigation, and damages settlements before the Federal Energy Regulatory Commission (FERC) in the United States, and also have been investigated for their impact on Alberta by the Market Surveillance Administrator (MSA).93 The experience does highlight the point that real and even manipulated shortages in neighboring markets can substantially impact Alberta. Expanding interconnections will further increase AESOs exposure to the market fundamentals and potential shortages in neighboring markets in the future. At the same time, as Figure 9 also shows, the price spikes that occurred in the Alberta energy market since 2006, with proximate causes often related to short-term supply adequacy problems, had virtually no impact on the market prices in these neighboring regions. These factors mean that the expansion of interconnections with neighboring markets will require the AESO to increase the extent to which it monitors for potential shortage conditions, including assigning a realistically low capacity value to total import capability. It also means that the AESO should maintain its procedures for limiting exports during scarcity conditions, and not introduce firm export transmission service without careful consideration of the potential resource adequacy consequences.94

90

91

92

93

94

If Alberta had the potential to become a large net exporter of power, then increased interties would have the potential to increase reliability in Alberta by incenting generation build-out in excess of supply needs to meet export demand. This scenario would only materialize if Alberta had structural potential for lowercost energy market prices over the long run, even after accounting for the capital cost recovery required through Albertas energy market (compared to Albertas neighbors which award cost recovery outside the energy market). Excluding the year 2000, the R2 values of predicting AESO prices from Mid-Columbia and Northern California power prices are 0.122 and 0.221 respectively, while AESO, Mid-Columbia, and Northern California power prices have R2 values of 0.227, 0.119, and 0.302 when predicted by gas prices. Natural gas prices were also high during the year 2000, but no higher than in other years with moderately high electric prices such as 2005. While the investigation into the bidding strategies and intertie conduct of Enron Canada Corporation and Powerex Corporation did not uncover prohibited behaviors, it did result in the revision of rules governing intertie conduct that would prevent those behaviors in the future. See MSA (2005); FERC (2010). No similar concern would be introduced by allowing for firm import capability, which would allow for the possibility that suppliers would have the firm option to sell into Alberta. However, only external suppliers

35

LIST OF ACRONYMS ACE AECO AESO AEMC AIES AIL ATC AUC AUD BATEA BAU CAD CAISO CASA CC CCEM CCS CH4 CO2 CO2e CONE CPUC CT DDS DR EDC EIA ERCOT FERC FOM FRB GHG Area Control Error Alberta Energy Company Alberta Electric System Operator Australian Energy Market Commission Alberta Internal Electric System Alberta Internal Load Available Transfer Capability Alberta Utilities Commission Australian Dollars Best Available Technology Economically Achievable Business as Usual Canadian Dollars California Independent System Operator Clean Air Strategic Alliance Combined Cycle Climate Change Emissions Management Carbon Capture and Sequestration Methane Carbon Dioxide Carbon Dioxide Equivalent Cost of New Entry California Public Utilities Commission Combustion Turbine Dispatch Down Service Demand Response EDC Associates, Ltd. Energy Information Administration Electric Reliability Council of Texas Federal Energy Regulatory Commission Fixed Operations and Maintenance Federal Reserve Board Greenhouse Gas 75

GJ GW GWh IPCC ISO ISO-NE kT MATL MT MW MWh MSA N2O NOX NEM NEPOOL NERC NP15 NYPP O&M PJM PM PPA PRD SO2 SPP SWIS TMR VCA VOLL

Gigajoules Gigawatts Gigawatt-hours Intergovernmental Panel on Climate Change Independent System Operator ISO New England Kilotonnes Montana-Alberta Tie Limited Megatonnes Megawatts Megawatt-hours Market Surveillance Administrator Nitrous Oxide Mono-nitrogen Oxides National Electricity Market New England Power Pool North American Electric Reliability Corporation North Path 15 New York Power Pool Operations and Maintenance PJM Interconnection, LLC Particulate Matter Power Purchase Arrangement Price-Responsive Demand Sulfur Dioxide Southwest Power Pool South West Interconnected System Transmission Must Run Voluntary Capacity Auction Value of Lost Load

76

APPENDICES

77

A. G ENERATOR O PERATING M ARGINS VERSUS F IXED C OSTS Figure 32 through Figure 36 show the estimated generator operating margins and fixed costs over the past decade for natural gas CCs, natural gas cogen, hydro, wind, and natural gas CTs including TMR units. A discussion of the implications of this information is in Section V.A.6, along with similar figures for coal units and gas CTs without TMR units. Figure 32 Historic Gas CC Operating Margins vs. Fixed Costs
$250

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

DDS
$200

Operating Reserves Energy Margins

$150

Cost of New Plant


$100

$50

Fixed O&M
$0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sources and Notes: Energy margins represent revenues minus estimated operating costs in energy market. Cost of New Plant includes capital costs and FOM. Unit-specific volumes and revenues as well as 2010 VOM, CONE, and FOM by unit type are from AESO (2010c). Gas prices and exchange rates from Bloomberg (2010). Heat rates estimated from Ventyx (2010), AESO (2010c), Alberta Environment (2010a-b). Historic CONE and FOM numbers are inflated according to the Handy-Whitman Index (converted from USD to CAD) between 2000 and 2009 from Whitman, et al. (2008) andPJM (2009); and by inflation between 2009 and 2010 from Bank of Canada (2010).

A-1

Figure 33 Historic Gas Cogen Operating Margins vs. Fixed Costs


$250

Operating Reserves Energy Margins

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

$200

$150

Cost of New Plant


$100

$50

Fixed O&M
$0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sources and Notes: 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


$400

Other AS
Operating Margins and Fixed Costs ($/kW-y) &
$350

$300

TMR

Supplemental Spinning Cost of New Plant Regulating Energy Margins

$250

$200

$150

$100

$50

Fixed O&M
$0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sources and Notes: Energy margins represent revenues minus estimated operating costs in energy market. Cost of New Plant includes capital costs and FOM. Unit-specific volumes and revenues as well as 2010 VOM, CONE, and FOM by unit type are from AESO (2010c). Gas prices and exchange rates from Bloomberg (2010). Historic CONE and FOM numbers are inflated according to the Handy-Whitman Index (converted from USD to CAD) between 2000 and 2009 from Whitman, et al. (2008) andPJM (2009); and by inflation between 2009 and 2010 from Bank of Canada (2010).

A-2

Figure 35 Historic Wind Operating Margins vs. Fixed Costs


Cost of New Plant
Operating Margins and Fixed Costs ($/kW-y) &
$250

$200

Energy Margins

$150

$100

$50

Fixed O&M

$0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sources and Notes: 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)
Operating Margins and Fixed Costs ($/kW-y) &
$200

Other AS

$150

TMR Cost of New Plant


$100

Operating Reserves

$50

Energy Margins Fixed O&M


$0

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Sources and Notes: 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).

A-3

B. M ETHOD F OR P ROJECTING O PERATING M ARGINS Section V.A.6 describes our approach to projecting future operating reserves revenues and energy margins as a linear function of perfect dispatch margins that could be achieved by a plant with no startup costs, outages, or dispatch constraints. The parameters of these linear relationships are shown in Table 8. Table 8 Energy Margins and Operating Reserves Revenue versus Perfect Dispatch Margins (Linear Relationships Based on Historic Monthly Data)
Energy Margins vs. Perfect Dispatch Margins a * (perfect energy margins) + b = (actual energy margins) a Coal Gas Cogen Gas CC Gas CT Hydro Wind 0.726 0.667 0.748 0.484 0.295 0.287 b ($/kW-yr) 19.86 -21.68 -42.91 -26.91 -10.81 n/a R2 0.710 0.817 0.918 0.655 0.440 n/a
Coal Gas Cogen Gas CC Gas CT Hydro Wind Reserves Revenues vs. Perfect Dispatch Margins a * (perfect energy margins) + b = (OR revenues) a 0.001 0.033 0.031 0.209 0.290 0.000 b ($/kW-yr) 0.219 -0.076 2.877 2.500 -60.828 0.000 R2 0.030 0.081 0.104 0.167 0.628 n/a

Sources and Notes: Calculated from historic monthly unit-level data over January 2008 through October 2010 from AESO (2010c). Wind is calculated as a simple percentage.

These relationships were developed based on the historic relationship between historic energy margins and operating reserves revenue calculated from AESO internal data as described in Section V.A.6 and a theoretical back-cast of perfect dispatch margins. These data are represented at the unit level for each month from January 2008 through October 2010. Figure 37 through Figure 42 are scatter plots of the data used to determine these linear relationships. Note that some data points show zero historic energy margins, which is an indication that the unit was on outage during that month.

B-1

Figure 37 Historic Gas CT Operating Margins vs. Perfect Dispatch Margins


$800 Actual Energy Margins ($/kW-yr) & $700 $600 $500 $400 $300 $200 $100 $0

Energy
y = 0.484x + -26.91 Energy R-Squared: 0.7095

100%

100%

Opeating Reserves
$0
OR Revenue ($/kw-yr) &&

$100

$200

$300

$400

$500

$600

$700

$800

$800 $600 $400 $200 $0 $0

Operating Reserves
y = 0.209x + 2.5 R-Squared: 0.1666

100%

$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


$1,000 $900 Actual Energy Margins ($/kW-yr) & $800 $700 $600 $500 $400 $300 $200 $100

Energy
= 0.484x + 42.914 -26.91 Energy yy= 0.748x R-Squared: 0.7095 R-Squared: 0.9177

100% 100%

Opeating Reserves
$0 $0
OR Revenue ($/kw-yr) &&

$100

$200

$300

$400

$500

$600

$700

$800

$900

$1,000

$1,000 $800 $600 $400 $200 $0 $0

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

100%

$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).

B-2

Figure 39 Historic Gas Cogen Operating Margins vs. Perfect Dispatch Margins
$1,000 $900 Actual Energy Margins ($/kW-yr) & $800 $700 $600 $500 $400 $300 $200 $100

Energy y= + -26.91 y 0.484x = 0.667x -21.684 Energy

R-Squared: 0.7095 R-Squared: 0.8174

100%

100%

Opeating Reserves
$0 $0
OR Revenue ($/kw-yr) &&

$100

$200

$300

$400

$500

$600

$700

$800

$900

$1,000

$1,000 $800 $600 $400 $200 $0 $0

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

100%

$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


$1,000 Actual Energy Margins ($/kW-yr) &

Energy y= 0.484x + -26.91 Energy y = 0.726x + 19.865


R-Squared: 0.7095 R-Squared: 0.7095

100%

100%

$800

$600

$400

$200

Opeating Reserves
$0 $0
OR Revenue ($/kw-yr) &&

$200

$400

$600

$800

$1,000

$1,000 $800 $600 $400 $200 $0 $0

Operating Reserves
y = 0.001x + 0.219 R-Squared: 0.0302

100%

$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).

B-3

$1,200 Actual Energy Margins ($/kW-yr) &

Figure 41 Historic Hydro Operating Margins vs. Perfect Dispatch Margins


Energy = 0.484x + Energy yy= 0.295x -26.91 10.806
R-Squared: 0.9177
R-Squared: 0.7095

$1,000

100% 100%

$800

$600

$400

$200

Opeating Reserves
$0 $0
OR Revenue ($/kw-yr) &&

$200

$400

$600

$800

$1,000

$1,200

$1,200 $1,000 $800 $600 $400 $200 $0 $0

Operating Reserves
y = 0.29x + -60.828 R-Squared: 0.6277

100%

$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


$1,000 $900 Actual Energy Margins ($/kW-yr) & $800 $700 $600 $500 $400 $300 $200 $100

Energy y 0.484x -26.91 Energy y= = 0.287x++ 0

R-Squared: 0.7095 R-Squared: n/a

100% 100%

Opeating Reserves
$0 $0
OR Revenue ($/kw-yr) &&

$100

$200

$300

$400

$500

$600

$700

$800

$900

$1,000

$1,000 $800 $600 $400 $200 $0 $0

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

100%

$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).

B-4

C. P ROJECTION OF G ENERATOR O PERATING M ARGINS VERSUS F IXED C OSTS Figure 43 through Figure 46 show a projection of future generator operating margins and fixed costs under baseline CO2e and gas assumptions using historic heat rates from 2006-10. The figures show results for natural gas CCs, natural gas cogen, hydro, and wind power plants. A discussion of the implications of this information is in Section V.B.2, along with similar figures for coal units and gas CTs. Figure 43 Projected Gas CC Operating Margins vs. Fixed Costs

Sources and Notes: Future price duration curve is calculated based on heat rates from 2006-10 and baseline gas and CO2e price forecasts. Cost of new plant includes FOM and real levelized capital costs from Section V.A.5, future escalation at 2.4% annually.

C-1

Figure 44 Projected Gas Cogen Operating Margins vs. Fixed Costs

Sources and Notes: 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.

C-2

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.

C-3

AN INTRODUCTION TO AUSTRALIAS NATIONAL ELECTRICITY MARKET


JULY 2010

Disclaimer This document is made available to you on the following basis: (a)  Purpose - This document is provided by the Australian Energy Market Operator Limited (AEMO) to you for information purposes only. You are not permitted to commercialise it or any information contained in it. (b)  No Reliance or warranty - This document may be subsequently amended. AEMO does not warrant or represent that the data or information in this document is accurate, reliable, complete or current or that it is suitable for particular purposes. You should verify and check the accuracy, completeness, reliability and suitability of this document for any use to which you intend to put it and seek independent expert advice before using it, or any information contained in it. (c)  Limitation of liability - To the extent permitted by law, AEMO and its advisers, consultants and other contributors to this document (or their respective associated companies, businesses, partners, directors, ofcers or employees) shall not be liable for any errors, omissions, defects or misrepresentations in the information contained in this document, or for any loss or damage suffered by persons who use or rely on such information (including by reason of negligence, negligent misstatement or otherwise). If any law prohibits the exclusion of such liability, AEMOs liability is limited, at AEMOs option, to the re-supply of the information, provided that this limitation is permitted by law and is fair and reasonable. 2010 - All rights reserved.

ELECTRICITY
3

NATIONAL ELECTRICITY MARKET

The Electricity Supply Industry The National Electricity Market The Australian Energy Market Operator National Electricity Law and Rules The Spot Market Key Parameters for NEM Operation Operating the NEM Ancillary Services Inter-regional Trade Market Forecasts Full Retail Competition Registered Participants Financial Contracts for Electricity Alternative Generation Technologies AEMO and the Environment Regulatory Arrangements Glossary

2 4 5 6 7 8 9 14 15 17 19 19 20 22 22 23 24

CONTENTS
1

THE ELECTRICITY SUPPLY INDUSTRY


Sectors of the electricity supply industry are involved with the generation, transmission, distribution and retail sale of electricity. Australias social, industrial and commercial success depends on the reliability of the electricity supply. In this way, the industry contributes signicantly to the national economy. Electricity can be converted readily to heat and light and used to power machines. It can also be transported with relative ease. These characteristics make electricity a convenient and manageable form of energy, and contribute both to its value as a commodity and its versatility as a source of power. A unit of power is referred to as a watt. The number of watts, or wattage, of an electrical appliance indicates the rate at which the appliance converts electrical energy to another form of energy such as heat or light. One watt is equivalent to one joule of work per second. Both the electrical pressure (voltage) and the number of electrons owing (current) determine the electrical power or rate of energy conversion. A 60-watt light globe uses 60 watts of electricity to produce light, and a typical electric kettle uses 2400 watts to produce heat.

How is Electricity Produced?


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

What is Electricity?
Electricity is a form of energy produced by the ow of electrons in a substance known as a conductor. The best conductors are metals such as copper and aluminium, and are commonly used in electrical wiring. Energy exists in many forms. Electricity is a secondary energy source as it is produced by the conversion of other energy sources like the chemical energy in coal, natural gas and oil. Other primary sources of energy, like the sun and wind, are increasingly being used to produce electricity. A quantity of energy can be changed or converted, but can never be created or destroyed.

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

THE ELECTRICITY SUPPLY INDUSTRY CONTINUED

NATIONAL ELECTRICITY MARKET

How is Electricity Transported?


Electricity travels along a conductor at close to the speed of light. When an appliance is switched on, power is instantly transmitted from a power station to the appliance. Although this occurs instantaneously, a specic 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 efcient transport on the transmission system. When the electricity arrives at the location where it is required, a substation transformer changes the high voltage electricity to low voltage for distribution. Distribution lines then carry low voltage electricity to consumers who access it through the power outlets in homes, ofces and factories. Transmission lines carry electricity long distances Distribution lines carry low voltage electricity to consumers

Power plant generates electricity

Transformer converts low voltage electricity to high voltage for efcient transport

Substation transformer converts high voltage electricity to low voltage for distribution

Homes, ofces and factories use electricity for lighting and heating and to power appliances

TRANSPORT OF ELECTRICITY

Energy exists in many forms. Electricity is a secondary energy source as it is produced by the conversion of other energy sources like the chemical energy in coal, natural gas and oil. Other primary sources of energy, like the sun and wind, are increasingly being used to produce electricity.
3

THE NATIONAL ELECTRICITY MARKET


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

GENERATION BY FUEL TYPE1


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

ELECTRICITY CONSUMPTION BY SECTOR


TRANSPORT AND STORAGE: 1.0% MINING: 9.4% MANUFACTURING: 9.1% ALUMINIUM SMELTING: 11.0% METALS: 18.3%

NUMBER OF CUSTOMERS BY SECTOR

AGRICULTURE: 0.8% RESIDENTIAL: 27.7%

BUSINESS: 12% DOMESTIC: 88%

COMMERCIAL: 22.8%

Source: Electricity Gas Australia 2010 ESAA

The Australian Energy Market Operator


The Australian Energy Market Operator (AEMO) was established to manage the NEM and gas markets from 1 July 2009. AEMOs core functions can be grouped into the following areas: Electricity Market - Power System and Market Operator Gas Markets Operator National Transmission Planner Transmission Services Energy Market Development AEMO operates on a cost recovery basis as a corporate entity limited by guarantee under the Corporations Law. Its membership structure is split between government and industry, respectively 60 and 40 percent, with this arrangement to be reviewed after three years of operation. Government members of AEMO include the Queensland, New South Wales, Victorian, South Australian and Tasmanian state governments, the Commonwealth and the Australian Capital Territory. AEMO and the NEM A key aim of AEMO is to provide an effective infrastructure for the efcient operation of the wholesale electricity market, to develop the market and improve its efciency and to coordinate planning of the interconnected power system. AEMOs primary responsibility is to balance the demand and supply of electricity by dispatching the generation necessary to meet demand. AEMOs key nancial objective of being self-funding is achieved through the full recovery of its operating costs from fees paid by market participants. The National Electricity Law and the Rules were amended to replace NEMMCO with AEMO as the national electricity market and system operator. AEMOs functions are prescribed in the National Electricity Law while procedures and processes for market operations, power system security, network connection and access, pricing for network services in the NEM and national transmission planning are all prescribed in the Rules. With respect to the electricity market AEMO has two core roles: Power System Operator Market Operator The market requirements determine how the power system is operated. AEMOs electricity market and system operation responsibilities include:  anagement of the NEM M Overseeing reliability and security of the NEM  Ensuring supply reserve to meet reliability standards  Directing generators to increase production during  periods of supply shortfall Instruction of load shedding to rebalance supply and  demand and protect power system operations Implementation of reserve trading to maintain supply  and reliability levels through demand-side response National transmission planning for the electricity  transmission grid and production of a National Transmission Network Development Plan Publication of the Electricity Statement  of Opportunities Electricity emergency management Facilitation of Full Retail Competition.

NATIONAL ELECTRICITY MARKET

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

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

NATIONAL ELECTRICITY LAW AND RULES


When the NEM commenced, a National Electricity Code provided guidelines for how the market was to operate. These guidelines were developed following comprehensive consultation and extensive trials conducted between governments, the electricity supply industry and electricity users as part of a government-driven deregulation and reform agenda. In June 2005, the National Electricity Code was replaced by the National Electricity Law and Rules. The Law and Rules were recently amended to replace NEMMCO with AEMO as the national electricity market and system operator. AEMOs functions are prescribed in the National Electricity Law while procedures and processes for market operations, power system security, network connection and access, pricing for network services in the NEM and national transmission planning are all prescribed in the Rules.

THE SPOT MARKET


Wholesale trading in electricity is conducted as a spot market where supply and demand are instantaneously matched in real-time through a centrally-coordinated dispatch process. Generators offer to supply the market with specic amounts of electricity at particular prices. Offers are submitted every ve minutes of every day. From all offers submitted, AEMO determines the generators required to produce electricity based on the principle of meeting prevailing demand in the most cost-efcient way. AEMO then dispatches these generators into production. A dispatch price is determined every ve minutes, and six dispatch prices are averaged every half-hour to determine the spot price for each trading interval for each of the regions of the NEM. AEMO uses the spot price as the basis for the settlement of nancial transactions for all energy traded in the NEM. The Rules set a maximum spot price, also known as a Market Price Cap, of $12,500 per megawatt hour (MWh). This is the maximum price at which generators can bid into the market and is the price automatically triggered when AEMO directs network service providers to interrupt customer supply in order to keep supply and demand in the system in balance.

THE SPOT MARKET CONTINUED

NATIONAL ELECTRICITY MARKET

Market Price Cap The Rules place a limit on the maximum spot price at any regional reference node. This limit is called the Market Price Cap and is $12,500 per MWh. The spot price may be set to the cap value when it is necessary to involuntarily interrupt electricity supplies (ie load shedding) to Market Customers in order to balance the overall electricity supply and demand. Market Floor Price The Rules place a limit on the minimum spot price. This limit is called the Market Floor Price and is currently set at -$1,000 per MWh. The Reliability Panel reviews the level of the Market Floor Price and Market Price Cap every two years. Two aspects of the transmission network contribute to varying costs of electricity supply within different areas of the NEM. Firstly, losses are incurred as power is transported from where it is produced to where it is consumed through electrical resistance and the heating up of conductors. Secondly, electricity being transported along certain elements of the network may encounter technical constraints on capacity or bottlenecks.

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

AEMO

GENERATOR

TRANSMISSION NETWORK SERVICE PROVIDER

TOTAL ENERGY SENT OUT 2008/09


TAS 4.9% SA 6.5% NSW 38% VIC 25.1% QLD 25.4%

DISTRIBUTION NETWORK SERVICE PROVIDER

MARKET CUSTOMER

Source: ESAA

DISPATCH INSTRUCTIONS PHYSICAL ELECTRICITY FLOW FINANCIAL FLOWS

ENERGY AND FINANCIAL FLOWS

KEY PARAMETERS FOR NEM OPERATION


AEMO is required to operate the power system efciently and ensure agreed standards of security and reliability are maintained. The minimum reserve levels across the different NEM regions are listed in the Electricity Statement of Opportunities on the AEMO website. During a period of load shedding, supply is withdrawn from those NEM regions affected by the shortfall in proportion to the demand levels at the time the shortfall began. The proportioning process determines the amount of load shedding for each affected region up to the point where interconnectors are operating to their maximum transfer capacity. Once the interconnectors reach their maximum transfer capacity, the importing region must bear any additional load shedding locally. By implementing load shedding, AEMO protects the integrity of power system operation so that widespread and long-lasting blackouts are avoided. It also ensures that the hardship caused by a sustained supply shortfall is shared in an equitable fashion.

Security of Supply
AEMOs highest priority as power system and market operator of the NEM is the management of power system security. Security of electricity supply is a measure of the power systems capacity to continue operating within dened technical limits despite the disconnection of a major power system element, such as a generator or interconnector. The maintenance of power system security ensures the power system is operated in a way that does not overload or damage any part of it or risk overload or damage after a credible event.

Managing Security and Reliability


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

Security and Reliability Directions


AEMO has the power to direct registered generators into production when a supply shortfall is expected and some generators are known to have withheld some of their total capacity from the market. AEMO only uses this power of direction to protect power system security or supply reliability.

Reserve Trading
When there is sufcient notice of an upcoming shortfall of supply that threatens to compromise minimum reserve margins, AEMO may tender for contracts for electricity supply from sources beyond those factored into AEMOs usual forecasting processes. At these times, emergency generators and other generators connected directly to the distribution network who submit tenders may enter contracts to boost supply in the NEM so the widespread supply interruptions that may otherwise have occurred can be avoided. In the same way, some electricity consumers may offer for a nancial consideration to decrease their demand at times of supply shortfall so that demand and supply are brought into balance.

Power System Reliability


Reliability is a measure of the power systems capacity to continue to supply sufcient power to satisfy customer demand, allowing for the loss of generation capacity. The shortfall of supply against demand is referred to as unserved energy. Reliability standards are established in the NEM that determine that unserved energy per year for each region must not exceed 0.002 percent of the total energy consumed in that region that year.

Load Shedding
In the event that demand in a region exceeds supply and all other means to satisfy demand have been implemented, AEMO can instruct network service providers to shed some customer load. This action is only taken when there is an urgent need to protect the power system by reducing demand and returning the system to balance. Load shedding involves a temporary suspension of supply to customers in a specic part or region of the NEM where system security is at risk.

Supply Reserve
The power system is required to be operated at all times with a certain level of reserve in order to meet the required standard of supply reliability across the NEM. Calculation of the minimum reserve requirements recognises reserve sharing in a national context.

OPERATING THE NEM


Operating the NEM involves conducting a sequence of activities to facilitate trade between the producers and wholesale consumers of electricity. These activities include establishing demand levels, receiving offers to supply from generators, scheduling generators, dispatching generators into production, calculating the spot price, measuring electricity use and nancially settling the market. A typical level of demand for electricity across the NEM is approximately 25,000 megawatts on a business day of average temperatures. There is ample supply available in the system to meet this level of demand. In fact, supply only comes under extreme pressure for a few hours on just a few days of extreme high temperature that occur each year. Further, because peak demand does not occur simultaneously in all regions, total supply can be shared between regions using the interconnected power network. Demand in the Victorian and South Australian regions of the NEM is characterised by short-term demand peaks during the summer months. It makes economic and market sense that these extreme peaks of demand be met by special arrangements rather than having excess base-load generation capacity in the system at all times. The peaks are currently being met by a combination of generators that have been specically built to service extreme demand periods (peak generators), and demand side participation, where consumers voluntarily and temporarily withdraw from the market when the spot price reaches a threshold level.

NATIONAL ELECTRICITY MARKET

9000

6,004

8,984 5,935 1,542

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

6000

3000 1,158

NSW

VIC

SA

AVERAGE DEMAND

AVERAGE DEMAND (MW) 2008/09


Source: ESAA

QLD

TAS

OPERATING THE NEM CONTINUED

Supply
The delivery of electricity to market customers comprises a sequence of distinct processes that AEMO manages according to strict timetables. Submitting Offers to Supply To enable AEMOs systems to facilitate supply, scheduled NEM generators are required to submit to AEMO offers indicating the volume of electricity they are prepared to produce for a specied price. There are three types of bids or offers to supply daily bids, re-bids and default bids. Daily bids are submitted before 12:30 pm on the day before supply is required, and are reected in pre-dispatch forecasts. Generators may submit re-bids up until approximately ve minutes prior to dispatch. In doing so, they can change the volume of electricity from what it was in the original offer, but they cannot change the offer price. Default bids are standing bids that apply where no daily bid has been made. These bids are of a commercialin-condence nature and, in general, reect the base operating levels for generators.

Scheduling and Dispatching Generators From the bids submitted, AEMOs systems determine which generators are required to satisfy demand, at what time, and their production levels in a process called scheduling. Offers to generate are stacked in order of rising price, and are then scheduled and dispatched into production. The use of the rising-price stack means that more expensive generators are scheduled into production as total demand for electricity increases. At times, the technical capacity of the transmission network may determine which generators are scheduled to meet demand. In such a situation, generators may be scheduled out of price order so that demand in a particular area supplied through the network may be satised.

From the bids submitted, AEMOs systems determine which generators are required to satisfy demand.

10

OPERATING THE NEM CONTINUED

NATIONAL ELECTRICITY MARKET

500
TOTAL DEMAND OF ELECTRICITY FROM THE POOL (MW)

400

D C A B

E F

$38 $37 $35

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

D. By 4:20pm, demand has increased to the point that Generator 5 is just required to meet demand, and the price increases to $38 per MWh. E. At 4.25 pm, Generators 1-4 are fully dispatched and Generator 5 partly dispatched. The price remains at $38 per MWh. F. By 4:30 pm, demand has fallen. Generator 5 (the most expensive generator) is no longer required, and Generator 4 is only partly dispatched. The price returns to $37 per MWh. The spot price for the trading period is calculated as the average of the six dispatch prices. That is, $(35+37+37+38+38+37) per MWh divided by six, or $37 per MWh. This is the price all generators receive for production during this period, and the price market customers pay for electricity they consume from the pool during this period.

300

200

$28
100

$20
0

4:05
GENERATOR:

4:10
ONE

4:15
TWO

4:20
THREE

4:25
FOUR FIVE

4:30

5-MINUTE PERIODS THROUGHOUT A HALF HOUR TRADING PERIOD

SCHEDULING OF NEM GENERATORS


Characteristic Gas and Coal-red Boilers Time to re-up generator from cold Degree of operator control over energy source Use of non-renewable resources Production of greenhouse gases Other characteristcs 8-48 hours high high high medium-low operating cost Gas Turbine 20 minutes high high medium-high medium-high operating cost Type Water (Hydro) 1 minute medium nil nil low fuel cost with plentiful water supply; production severely affected by drought Renewable (Wind/Solar) dependent on prevailing weather low nil nil suitable for remote and stand-alone applications; batteries may be used to store power

CHARACTERISTICS OF GENERATORS
11

OPERATING THE NEM CONTINUED

Setting the Spot Price


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

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

by local network service providers. These service providers are responsible for measuring the volume of electricity supplied, validating the data from the meters, applying distribution loss factors, and forwarding the information to AEMO for use in calculating and preparing accounts for nancial settlement.

Settling the Market


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

Measuring Electricity Use


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

THE GENERATOR DISPATCH CYCLE


Scheduling Ranking bids  Identifying the dispatch levels of generating units

Data input  Establishing current operational status of generating units  Assessing demand forecasts Applying loss factors  Determining system conditions

Dispatch  Issuing dispatch instructions to generators

12

NATIONAL ELECTRICITY MARKET

If a market participant breaches their maximum credit limit on any one day of trading, a call notice for rectication of the situation and then a default notice may be issued to ensure that AEMO is able to settle the market according to its xed timetable. AEMO has the authority to suspend a market participant who fails to respond adequately to a default notice, and to reinstate that market participant only when their required nancial position is re-established.

Demand Side Participation


Demand side participation refers to the situation where market customers reduce their consumption of electricity in response to a change in market conditions, such as high spot prices. This is a deliberate action taken when demand for power drives spot prices high. Under similar arrangements scheduled loads, such as smelters, may elect to withdraw from the market when the spot price reaches a particular threshold, and resume trading when the price falls to the level of their bids again. This strategy is benecial 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 specic demand is intentionally moved to a time when there is lower overall demand and consequent lower spot prices. Off-peak hot water arrangements are an example of the deliberate shifting of demand for electricity to a low-demand period.

All market customers are required to install equipment to record their electricity consumption. AEMO registers, accredits and audits a range of metering services provided by local network service providers.

13

ANCILLARY SERVICES
Ancillary services are those services used by AEMO to manage the power system safely, securely and reliably. Ancillary services maintain key technical characteristics of the system, including standards for frequency, voltage, network loading and system re-start processes. AEMO operates eight separate markets for the delivery of Frequency Control Ancillary Services (FCAS), and purchases Network Control Ancillary Services (NCAS) and System Restart Ancillary Services (SRAS) under agreements with service providers. FCAS providers bid their services into the FCAS markets in a similar way to how generators bid into the energy market. The FCAS markets were introduced to the NEM in September 2001 and provide simpler, more dynamic and transparent arrangements that have further increased competition and contributed to improved overall market efciency. 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. NCAS are primarily used to: C  ontrol the voltage at different points of the electrical network to within the prescribed standards; or  Control the power flow on network elements to within the physical limitations of those elements SRAS are reserved for contingency situations in which there has been a major supply disruption or where the electrical system must be restarted. Ancillary service costs are dependant upon the amount of service required at any particular time and, as these amounts can vary signicantly from period to period, costs will also vary.

14

INTER-REGIONAL TRADE
The NEM comprises ve interconnected electrical regions. There is a designated region reference node in each region where the regional spot price of electricity is set. The Queensland, New South Wales, Victoria, Tasmania and South Australia regions all contain both major generation and demand centres.

NATIONAL ELECTRICITY MARKET

Regulated Interconnectors
A regulated interconnector is an interconnector that has passed the ACCC-devised regulatory test and has been deemed to add net market value to the NEM. Having passed the test, a regulated interconnector becomes eligible to receive a xed annual revenue set by the ACCC and based on the value of the asset, regardless of actual usage. The revenue is collected as part of the network charges included in the accounts of electricity end-users. At present, regulated interconnectors operate between all adjacent regions of the NEM, except Tasmania.

Interconnectors
The high-voltage transmission lines that transport electricity between adjacent NEM regions are called interconnectors. Interconnectors are used to import electricity into a region when demand is higher than can be met by local generators, or when the price of electricity in an adjoining region is low enough to displace the local supply. AEMOs ability to schedule generators to meet demand using an interconnector to facilitate importing electricity is sometimes limited by the physical transfer capacity of the interconnector. When the technical limit of an interconnectors capacity is reached, the interconnector is said to be constrained. For example, if prices are very low in one region and high in an adjacent region, electricity can be sent from the rst to the second region across an interconnector up to the maximum technical capacity of the interconnector. AEMOs systems will then dispatch local generators with the lowest price offers from within the second region to meet the outstanding consumer demand.

INTERCONNECTORS IN THE NEM

REGIONAL REFERENCE NODE REGULATED INTERCONNECTOR MARKET NETWORK SERVICE PROVIDER

QLD SOUTH PINE


NSW-QLD (QNI)

SA

NSW-QLD TERRANORA VIC-SA NSW WEST SYDNEY (MURRAYLINK)

TORRENS ISLAND

VIC
VIC-SA (HEYWOOD)

VIC-NSW

THOMASTOWN
TAS-VIC (BASSLINK)

GEORGE TOWN TAS

15

INTER-REGIONAL TRADE CONTINUED

Unregulated Interconnectors
Unregulated (or market) interconnectors derive revenue by trading in the spot market. They do this by purchasing energy in a lower price region and selling it to a higher price region, or by selling the rights to revenue generated by trading across the interconnector. Unregulated interconnectors are not required to undergo regulatory test evaluation. An unregulated interconnector Basslink operates between the Tasmanian and Victorian regions of the NEM. Murraylink and Directlink were built as unregulated interconnectors between Victoria and South Australia, and New South Wales and Queensland respectively. They successfully applied to the ACCC for conversion to regulated status.

Loss of Energy in the System


As electricity ows through the transmission and distribution networks, energy is lost due to electrical resistance, and the heating of conductors. The losses are equivalent to approximately 10 per cent of the total electricity transported between power stations and market customers. Energy losses on the network must be factored in at all stages of electricity production and transport to ensure the delivery of adequate supply to meet prevailing demand and maintain the power system in balance. In practical terms, this means that more electricity must be generated than indicated in demand forecasts in order to allow for this loss during transportation. The impact of network losses on spot prices is mathematically represented as transmission and distribution loss factors. Loss factors within each region of the NEM are calculated based on forecast demand, and xed for a period of 12 months to facilitate efcient scheduling and settlement processes in the NEM. Loss factors between regions of the NEM are dynamically calculated and reect the operating conditions at the time of the transmission of the electricity. REGION BOUNDARY REGION A
REGION A REFERENCE NODE

REGION B REFERENCE NODE CUSTOMER 2 CUSTOMER 1

REGION B
INTER-REGIONAL LOSS FACTOR APPLIES INTRA-REGIONAL LOSS FACTOR APPLIES

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

As electricity ows through the transmission and distribution networks, energy is lost due to electrical resistance, and the heating of conductors.
16

LOSS OF ENERGY IN THE POWER SYSTEM

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

NATIONAL ELECTRICITY MARKET

Pre-dispatch Forecasting
Pre-dispatch is a short-term forecast of supply and demand in the market. It is used to estimate the price and demand for the upcoming trading day, and the volume of electricity expected to be supplied through the interconnectors between regions. Generators and network operators are required to notify AEMO of their maximum supply capacity and availability, and this information is matched against regional demand forecasts. All offers to supply are then collated so that potential shortfalls of supply against demand can be identied and published. Participants in the market use this information as the basis for any re-bids of the capacity they wish to bring to the market.

Projected Assessment of System Adequacy


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

AEMO PRODUCES TWO PASA FORECASTS


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

Five-minute Matching of Supply and Demand


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

A DAY IN THE NEM


12.00 MIDNIGHT START/ END OF SETTLEMENT DAY PRE-DISPATCH FORECAST PUBLISHED

24 24 HOUR CLOCK

18

4.00AM (EST) TRADING DAY STARTS AND ENDS 30 MINUTE TRADING INTERVAL (48 PER DAY)

60 1 HOUR CLOCK

45

15
5 MINUTE DISPATCH INTERVAL (288 PER DAY) PRE-DISPATCH FORECAST PUBLISHED

1 HOUR 12:30PM DEADLINE FOR DAILY BIDS FOR NEXT TRADING DAY (RE-BIDS CAN BE UP TO 5 MINUTES PRIOR TO DISPATCH

17

MARKET FORECASTS CONTINUED

Electricity Statement of Opportunities


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

National Transmission Network Development Plan


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

18

FULL RETAIL COMPETITION


Since the commencement of the NEM, electricity consumers have progressively gained the right to choose their own supplier. This has meant that AEMOs responsibilities have extended from managing the wholesale market to providing the systems and processes to support competition and choice for all end-users in the retail electricity market. Delivering full retail competition (FRC), or contestability, has required new information technology systems to process transfers of customers between registered retailers in the NEM. The systems that facilitate this function contain one of the largest metering databases in the world. They accept data from a variety of electricity meter types and have the capacity to process information from up to 10 million meters. AEMOs systems are set up to provide key meter installation details to support a simple and rapid information transfer process. Different metering processes are required for different types of meters used in the NEM, to support consumer transfer and core settlement procedures and to calculate load proles. The cost of electricity consumed is then calculated according to a user prole that approximates the pattern of use in a typical situation. By June 2009, approximately 6.3 million customer transfers from one retailer to another had taken place. As a result of the introduction of full retail competition, electricity retailers are increasingly competing, and creating new and unique products as a means to increase their customer bases. One of AEMOs responsibilities under the Rules is to register participants in the NEM. There are six main categories of registered participant, including those who participate directly in trading activities, and other participants who provide services essential for the operation of the market. The categories are generator, market customer, intending participant, network service provider, trader, reallocator and special participant. Market participants include market generators, market network service providers and market customers. A market participant must be separately registered in each category of the market in which it participates. For example, a business that participates as a generator (a peaking generating plant for instance) and as a market customer (retailer of electricity to end-use customers) would be required to be registered as both a generator and a market customer. The registration of participants is a formal process, strictly dened in the Rules. Registered participants are required to pay participant fees that are levied to recover the costs associated with managing the market.

NATIONAL ELECTRICITY MARKET

REGISTERED PARTICIPANTS
Market Participants
Registered to participate in the National Electricity Market
Market Generators Sell entire electricity output through the spot market and receive the spot price at settlement. Scheduled: aggregate generation capacity of more than or equal to 30 megawatts. Semi-scheduled: aggregate generation capacity of more than or equal to 30 megawatts where output is intermittent. Non-scheduled: aggregate generation capacity of less than 30 megawatts. Market Network Service Providers Own and operate a network linked to the national grid at two terminals in different NEM regions. Pay market participant fees and obtain revenue from trading in the NEM. Market Customers Purchase electricity supplied to a connection point on a NEM transmission or distribution system for the spot price. Electricity Retailers: buy electricity at spot price and on-sell it to end-use customers. End-use Customers: buy directly from the market for own use.

Other Registered Participants


Transmission Network Service Provider Owner and operator of the high-voltage transmission towers and wires that transport electricity. Distribution Network Service Provider Owner and operator of substations and the wires that transport from distribution centres to end-use consumers. Also provider of technical services, including construction of power lines, inspection of equipment, maintenance and street lighting. Reallocator Registered with AEMO to participate in reallocation transactions under clause 3.15.11 of the National Electricity Rules. Special Participant System operators or agents appointed to perform power security functions. Distribution system operators and controllers or operators of any portion of the distribution system. Intending Participant Must reasonably satisfy AEMO of intention to perform activity that would entitle it to be a registered participant. Trader Party registered to participate in the settlement residue auction.

19

FINANCIAL CONTRACTS FOR ELECTRICITY


Participants in the NEM require a means of managing the nancial risks associated with the signicant degree of spot price volatility that occurs during trading periods. They typically achieve this by using nancial contracts that lock in a rm price for electricity that will be produced or consumed at a given time in the future. These contracts serve to substantially reduce the nancial exposure of market participants and contribute to spot market stability. They are known as derivatives, and include swaps or hedges, options and futures contracts.
$160 $140 $120 $100 $80 $60 $40 $20 $0 STRIKE PRICE (AGREED CONTRACT PRICE) BUYER PAYS SELLER DIFFERENCE BETWEEN AGREED STRIKE PRICE AND SPOT PRICE SPOT PRICE In this example of a hedge contract, the two parties have agreed to set a price (the strike price) of $40/MWh. The graph shows the strike price and the actual spot price over a hypothetical days trading. When the spot price is below the strike price, the market customer pays the difference in these prices to the generator. In this case, when the spot price is $17, the market customer pays the generator $23/MWh. When the spot price exceeds the strike price, the generator pays the market customer the extra required to purchase electricity from the pool. In this case, when the spot price rises to $145/MWh, the generator pays the market customer the $105/MWh difference.

SELLER PAYS BUYER DIFFERENCE BETWEEN AGREED STRIKE PRICE AND SPOT PRICE

PRICE

Hedge Contracts
Hedge contracts are typically agreements between generators and customers that operate independently of both the market and AEMOs administration. The details of hedge contracts are not factored into the balancing of supply and demand, and are not regulated under the Rules. These contracts can be entered into under either long-term or short-term arrangements that set an agreed, or strike, price for electricity traded through the pool. In this way, hedge contracts are nancial instruments that participants can use to manage the nancial risk that results from potential volatility of the spot price. The basic form of a hedge contract exists where two parties agree to exchange cash so that a dened quantity of electricity over a nominated period is effectively valued at an agreed strike price. Under such an agreement, generators pay customers the difference when the spot price is above the strike price. When the spot price is below the strike price, customers pay generators the difference between the spot price and the strike price.

0400

0800

1200

1600 TIME

2000

0000

HEDGE CONTRACTS IN THE NEM

Hedge contracts are typically agreements between generators and customers that operate independently of both the market and AEMOs administration.

20

FINANCIAL CONTRACTS FOR ELECTRICITY CONTINUED

NATIONAL ELECTRICITY MARKET

Auctions of Inter-region Settlement Residues


The spot price for electricity in each region of the NEM is determined by a number of factors, including supply and demand, the physical limitations of interconnectors, and the loss factors for both the transmission and distribution networks. This means that there may be signicant differences in the spot price for any trading interval across NEM regions. The difference between the value of electricity in the region where it is generated and its value if sold in another region is called the inter-regional settlement residue. The settlement residue that accumulates is made available to the market by the conduct of an auction. The auction process establishes the market value of the residue, and contributes to inter-regional trade by providing registered generators, market customers and traders with a mechanism to manage the risk associated with different price outcomes between trading regions. Registered participants who purchase auction units obtain access to a share of the residue. In this way, the premium paid for the auction units provide protection against high price differences between regions in the wholesale market. REGION A
SPOT PRICE $100/MWH

GENERATORS ARE PAID AT $100/MWH REGION BOUNDARY

SETTLEMENT RESIDUE ACCRUED IN REGION B: $120-$100=$20 (EXCLUDING LOSS FACTORS)

CUSTOMERS BUY AT $120/MWH

SPOT PRICE $120/MWH

REGION B
POWER FLOW

INTER-REGIONAL SETTLEMENTS RESIDUE

21

ALTERNATIVE GENERATION TECHNOLOGIES


The range of technologies for the generation of electricity is expanding to accommodate alternative energy sources such as wind energy. Large wind generators are typically registered as semischeduled generators (rather than scheduled) because their energy source is intermittent and their generation cannot increase on demand. The market is designed to allow intermittent generators to participate and share the same power system and the same consumers. The NEMs base-load generators are scheduled according to bids, and production from each generating unit is controlled by operators. The changeability and unpredictability of wind means that wind generators cannot be scheduled in the usual way. The integration of wind and other intermittent generators to the NEM must take account of AEMOs responsibility to maintain power system security, and be managed during each ve-minute dispatch interval. The variation of output associated with wind generators (with individual output that can change by as much as 50 per cent in a ve-minute dispatch interval) may require interconnectors to operate at lower limits to avoid overloads, and hence reduce the total supply capacity available to the market. In late 2008, the Australian Wind Energy Forecasting System (AWEFS) was implemented to forecast the energy likely to be produced by major wind farms in the NEM. This system enabled the generator classication semi-scheduled to be introduced in the NEM, allowing intermittent generators (including wind farms) to compete in the NEM through the bidding process. The success of AWEFS and the semi-scheduled category will help the energy market respond to an ongoing expansion in renewable generation while maintaining the security of supply that has been the hallmark of the industry.

AEMO AND THE ENVIRONMENT


AEMOs role as the manager of both the power system and the electricity market means that it is sometimes asked about issues of environmental management, the sustainability of the market and the electricity supply industry in general. Under the Rules, AEMOs charter focuses specically on efciency, security and reliability of power supply, and excludes favouring one fuel source over any other. Consequently, AEMO has neither the power nor the authority to make decisions based on considerations of sustainability and balance in resource management. The various state regulators ensure that environmental impact assessments are conducted as part of any power industry planning initiatives. The regulators also monitor operations at industry sites within their jurisdictions, and the industry itself operates and audits waste reduction and recycling programs.

Renewable Energy Target


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

Large wind generators are typically registered as semi-scheduled generators because their energy source is intermittent and their generation cannot increase on demand.
22

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

NATIONAL ELECTRICITY MARKET


23

GLOSSARY
Renewable generation Energy conversion techniques including wind, solar, hydro and geothermal. The primary commercial sources at this time are hydro and wind power. Hedge contracts Long term or short term arrangements that set a strike (agreed) price for electricity traded through the pool. Interconnectors The high-voltage transmission lines that transport electricity between adjacent NEM regions. Intermittent generators Where the changeability and unpredictability of the source (such as wind) means the generators cannot be scheduled to operate in the same way as conventional generators (coal, gas or oil). Load shedding AEMO can request network service providers to disconnect some customers when demand in a region exceeds supply. This action is only taken when there is an urgent need to reduce demand and return the system to balance. Market Price Cap The maximum price at which generators can bid into the market. Networks (Transmission) Transmission lines carry high voltage electricity to substation transformers where it is changed to low voltage for distribution. Networks (Distribution) Distribution lines carry low voltage electricity to consumers who access it through the power outlets in homes, ofces and factories. Transformers Convert the electricity produced at a generation plant from low to high voltage to enable its efcient transport on the transmission system. When the electricity arrives at the location where it is required, a substation transformer changes the high voltage electricity to low voltage for distribution to end users.

24

REGIONS AND NETWORKS IN AUSTRALIAS NATIONAL ELECTRICITY MARKET

2 2 2 2 2 2 2 2 2 2

TRANSMISSION INFRASTRUCTURE
POWER STATION SUBSTATION WINDFARM 500 KV TRANSMISSION LINE 330 KV TRANSMISSION LINE 275 KV TRANSMISSION LINE 220 KV TRANSMISSION LINE 132 / 110 KV LINE 66 KV LINE DC LINK MULTIPLE CIRCUIT LINES

2 2 2

2 2

DC 3 2

REGIONAL BOUNDARIES
SYD WEST

2 2 2

REGIONAL REFERENCE NODE QUEENSLAND NEW SOUTH WALES VICTORIA SOUTH AUSTRALIA TASMANIA

2 2

2 2 2 2 2 2 2 2 2

2 2

2 3

SOUTH PINE
DC

2 2

2 2
2

DC 2 2 2 2

TORRENS ISLAND
2 2 2 2 2 2 2 2 2

WEST SYDNEY

2
2

2 2

3 2

4 2

2 2

THOMASTOWN
DC

2 2 2 2

GEORGE TOWN

2 2

25

ELECTRICITY
AEMO GPO Box 2008 Melbourne VIC 3001 Website: www.aemo.com.au INFORMATION CENTRE Telephone: 1300 361 011 ISBN 0-646-41233-7

The Nordic Electricity Exchange and The Nordic Model for a Liberalized Electricity Market

1 The market When the electricity market is liberalized, electricity becomes a commodity like, for instance, grain or oil. At the outset, there is as in all other markets a wholesale market and a retail market and there are the three usual players: the producers, the retailers and the end users. However, for electricity, a more advanced trading pattern quickly develops. New players enter the scene: the traders and the brokers (Figure 1).

Figure 1: The commercial players and the electricity exchange

A trader is a player who owns the electricity during the trading process. For example, the trader may buy electricity from a producer and subsequently sell it to a retailer. The trader may also choose to buy electricity from one retailer and sell it to another retailer and so forth: there are many routes from the producer to the end user. The brokers play the same part in the electricity market as the estate agent in the property market. The broker does not own the commodity he acts as an intermediary. A retailer may, for example, ask the broker to find a producer who will sell a given amount of electricity at a given time. The Nordic electricity exchange Nord Pool Spot covers Denmark, Finland, Sweden, Norway , Estonia and Lithuania. Nord Pool Spot is an exchange primarily servicing the players at the wholesale market for electricity. The customers on Nord Pool Spot are
Nord Pool Spot AS, Tel +47 6710 9100, Fax +47 6710 9101, PO Box 121, NO-1325 Lysaker, Norway, info@nordpoolspot.com, org nr. NO 984 058 098 MVA, www.nordpoolspot.com 2

the producers, retailers, and traders who choose to trade on the electricity exchange. In addition, large end users trade on the electricity exchange. In this article, the term the Nord Pool Spot exchange area denotes Denmark, Finland, Sweden, Norway, Estonia and Lithuania. 2 The Point Tariff System In Figure 2, the water illustrates the electrical power and the walls of the tanks illustrate the transmission grid.

Figure 2: Illustration of the Nordic Point tariff system

The idea of the system of point tariff is that the producers pay a fee to the grid owner for each kWh they pour into the grid. Correspondingly, the end users pay a fee for each kWh they draw from grid. This means for example, that a retailer in Southern Sweden may buy electricity from a producer in Northern Sweden. Of course, such a deal does not cause the producers electricity to travel all the way from Northern Sweden to Southern Sweden. The principle is simply that for each hour somewhere a producer has to pour an amount of electricity to the grid which corresponds to the amount the retailers customers have tapped from the grid. 3 The non-commercial players The roads in the Nordic countries are operated by monopolies: The municipalities, the counties and the state. For electricity, the grid functions like the roads transporting the power. Correspondingly, the grid is operated by non-commercial monopolies (Figure 3). For each local area, there is a local grid operator who handles the local low-voltage grid (cf. the municipalities and counties operating the local roads). The high-voltage grid is operated by the transmission system operator (TSO) just as the motorways are operated by the state. In addition to owning and operating the high-voltage grid, the TSO is responsible for the security of supply in its country. Consequently, the TSO rules and controls the electricity system in his country. Basically, the physical control and maintenance of
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 3

the electricity system is done in the same way, whether you have market economy or planned economy.

Figure 3: The grid connection between producers and End-users

Only the financial organization is changed when we shift from planning economy to market economy. This is because the laws of nature are the same whether we have planned economy or market economy. This also holds for corn flakes: the machine filling the corn flakes into cartons does not care whether there is market economy or planned economy. It makes no difference to the physics whether there is planned economy or market economy. The commercial players are not and cannot be responsible for the security of supply. If a South Swedish retailer, for example, has bought electricity from a North Swedish producer, the North Swedish producer cannot guarantee that there will be electricity in the plug at the retailers customers. What the commercial players deliver to each other and the end users are only the prices (and the bills). Hence, the commercial players deliver financial services only. The commercial players work in the domain which is changed when the electricity market is liberalized: the financial domain.

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

4 The transmission system operator (TSO) The TSO is responsible for keeping the respective area electrically stable. Technically, this means that the frequency must be kept at 50 Hz. In other words, the TSO is responsible for the commodity (electricity) arriving at the end users sites. The TSO must be a non-commercial organization, neutral and independent of commercial players. The TSOs in the exchange areas thus have the responsibility for both the high-voltage grid and the security of supply. In Norway, the TSO is the stateowned grid company Statnett. In Sweden, the TSO is the state-owned grid company Svenska Kraftnt. The TSO in Finland is the grid company Fingrid and is owned partly by the Finnish State and partly by Finnish insurance companies. In Denmark, the TSO is the state-owned grid company Energinet.dk. Energinet.dk is TSO for both electricity and gas. The TSO in Estonia is Elering and is fully owned by the Estonian state. In Lithuania the TSO, Litgrid, is also stately owned. 5 Regulating power market The regulating power market is managed by the TSO in order to obtain stable frequency in the transmission grid. It may happen that the consumption exceeds the generation. In this case, the frequency of the alternating current will fall to a value below 50 Hz. When this happens, the TSO must ensure that one or more producers deliver(s) more electricity to the grid (Figure 2). In this case, the TSO buys more electrical power from producer(s) who has proclaimed excess generation capacity. We say that the TSO is procuring up regulation. The generation of electricity may also be too big exceeding the consumption. In this case, the frequency will rise to a value above 50 Hz. Now, the TSO must ensure that one or more producers reduce(s) the generation of electricity. In this case, the TSO is selling electrical power to the producers thereby causing the producers to reduce their generation. We say that the TSO is procuring down regulation. The electricity, which the TSO in this way trades with selected market players, is called regulating power. Hence, the regulating power is traded by the TSO in order to regulate the frequency to keep it at 50 Hz. To illustrate the setting of prices in the regulating power market an example is presented. The green rectangles in Figure 6 illustrate up-regulation orders, e.g. producers with available generation capacity. The orange rectangles illustrate downregulation orders, e.g. consumers being able to reduce consumption. All regulating power orders submitted to the TSOs are ranked with increasing price (merit-order). Assume there is a need for 400 MW up regulations. All the up-regulation orders with lowest prices are activated until 400 MW is reached. The price of the last up regulated MW sets the up-regulation price. The orders with prices below the up-regulation price have a profit, equal to the difference between final regulation price and the offered price. The same procedure is used to find the down-regulation price.
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 5

Figure 4: Price setting in the regulating power market

6 Balancing Power In the wholesale market electricity is bought and sold hourly. Figure 4 illustrates an example where a retailer buys electricity for one particular hour at one specific date. The hour during which the power is delivered and consumed is called the hour of operation. In the example, the retailer has two contracts of 30 MWh and 70 MWh, respectively: the retailer expects that his customers will consume 100 MWh during this hour of operation (1 MWh is 1,000 kWh).

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 6

Figure 5: The retailers purchase of electricity for the hour 1pm 2pm on September 23th 2010

Before the hour of operation, the purchases must be made. After the hour of operation, the settlement is done (Figure 5). The retailer pays the suppliers for the 30 MWh and the 70 MWh. Assume that the retailers customers have only used 85 MWh during this hour of operation. In this case, the retailer has per definition sold the 15 MWh to the TSO. The TSO pays the retailer for the 15 MWh.

Figure 6: Settling the consumption of electricity

This trade with the TSO creates a balance between the retailers total trading and the retailers customers consumption. The electricity, which the retailer trades with the TSO, is therefore called balancing power, or often referred to as regulating power. If the TSO had to procure up-regulation during this hour, the TSO will pay the retailer the up-regulating price for the balancing power (i.e. the retailer will get the same price as the producers, who sold up-regulating power to the TSO during this hour). Normally, the up-regulating price will be higher than the market price (in this article, the market price is the day-ahead exchange price for this hour). If the TSO had to procure down-regulation during this hour, the TSO will pay the retailer the down-regulating price for the balancing power (i.e. the retailer will get the same price as the producers, who bought down-regulating power from the TSO
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 7

during this hour). Normally, the down-regulating price will be lower than the market price. In a different case, assume the retailers customers have used 110 MWh during this hour of operation. This is 10 MWh more than the retailer bought before the hour of operation. In this case, the retailer has to buy the additional 10 MWh from the TSO. In this situation, the TSO will invoice the retailer for the 10 MWh.

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

Elspot is Nord Pool Spots day-ahead auction market, where electrical power is traded. Players, who want to trade power on the Elspot market, must send their purchase orders to Nord Pool Spot at the latest at noon the day before the power is delivered to the grid. Correspondingly, participants who want to sell power to Elspot must send their sale offers to Nord Pool Spot at the latest at noon the day before the power is delivered to the grid (i.e. gate closure is 12.00).

Figure 7: Bid/Offer from one player for the hour 1pm 2pm of tomorrow.

The orders and offers are sent electronically to Nord Pool Spot in Oslo: the participants send the orders to Nord Pool Spot via the Internet. Figure 7 shows an example of orders submitted by a retailer for one hour of the following day. The retailer expects that his customers will consume 50 MWh during this hour. This retailer has his own generation facility. Hence, he can choose whether he will either: - buy the 50 MWh from the exchange and therefore not produce anything himself. - buy some of the electricity from the exchange and produce the rest himself. - produce precisely 50 MWh.
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 9

- or sell electricity to the exchange and consequently produce more than 50 MWh. The retailer in the example has informed the electricity exchange that he will buy 50 MWh from Elspot, if the exchange price for this hour turns out to be 20 EUR/MWh or less. If the exchange price for this hour turns out to be 40 EUR/MWh, the retailer will buy 10 MWh. In this case, the retailer will produce the remaining 40 MWh at his own generation facility. The retailer will sell 10 MWh if the price turns out to be 50 EUR/MWh. If the price is between 50 and 60 EUR/MWh, the retailer will sell an amount corresponding to the sloping curve. If the price is 60 EUR/MWh or more the retailer will sell 30 MWh. At Nord Pool Spot, the purchase orders are aggregated to a demand curve. The sale offers are aggregated to a supply curve (Figure 8). The intersection of the two curves gives the market price for one specific hour.

Figure 8: Aggregated supply and demand curves

Nord Pool Spot calculates a price for each hour. Elspot is a day-ahead market, as this is trading for the following day. This way of calculating the price is called a double auction, as both the buyers and the sellers have submitted orders (for many other auction types, only the buyers submit orders). Hence, Elspot is called a day-ahead auction market (as the word double is cut out from the type description). Figure 9 shows the prices during one specific day, July 14th 2006..
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 10

At noon, Nord Pool Spots computer in Oslo starts calculating the day-ahead prices. Having finished the calculation, Nord Pool Spot publishes the prices. At the same time, Nord Pool Spot reports to the participants how much electricity they have bought or sold for each hour of the following day. These reports on buying and selling are also sent to the TSOs in the Nord Pool Spot area. The TSOs use this information, when they later calculate the balancing power for each player.

Figure 9: System price Thursday September 22nd 2011

There is a standard Elspot trading fee in EUR/MWh which is paid by both buyers and sellers. 9 Bidding areas Actually, chapter 6 describes how the so-called System Price is calculated. The System Price is the theoretical, common price we would have in the Nordic area if there were no grid bottlenecks. Due to the bottlenecks, the Nord Pool Spot exchange area is divided into a number of bidding areas. For example, when a producer in Eastern Denmark sends his orders to Nord Pool Spot, he must specify that these orders are submitted for delivery in the bidding area Eastern Denmark. The TSOs decides the number of bidding areas its boundaries. Eastern Denmark and Western Denmark are always treated as two different bidding areas. Sweden constitutes one bidding area until November 2011 when it is to be divided into four bidding areas. Also, Finland, Estonia and Lithuania constitutes one bidding area while Norway currently (August 2011) has five bidding areas.
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 11

Nord Pool Spot calculates a price for each bidding area for each hour of the following day. Naturally, there are often hours, where neighboring bidding areas have the same price. Likewise there may also be hours, where the whole exchange area has the same price: for example, during 2010, the whole exchange area had the same day-ahead price during 10 % of the hours.

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

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 12

Figure 10: Export of 50 MW from Norway to Sweden

Figure 11: Import of 50 MW to Sweden from Norway

Assume there is 50 MW available transmission capacity between Norway and Sweden. The price in Sweden would be lowered to 233.33 NOK/MWh due to additional available production. The price in Norway would increase to 283.33 NOK/MWh due to higher consumption.

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 13

Available transmission capacity

0 MW exchange

50 MW exchange CONGESTION [NOK/MWh] 233.33 283.33

Bidding area Norway Sweden

[NOK/MWh] 200 300

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 14

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

12 Cross-border trading Inside the Nord Pool Spot exchange area, all the transmission capacity on the external transmission lines is handled by Nord Pool Spot through implicit auction during price calculation. Two Nordic commercial players situated in different bidding areas cannot trade electricity with each other. This is because Nord Pool Spot handles all the trading capacity on the cross-border links, on behalf of the Nordic TSOs. In order to trade with each other, Nordic players in different bidding areas can use the financial electricity market (Figure 13). The two players can trade the power on Nord Pool Spot or with a player situated in their own bidding area (i.e. the power is traded locally). In addition the two players have a settlement in accordance with the financial contract.

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 15

Figure 13: The capacity on the Nordic bottlenecks is given to E.E. (Electricity Exchange). How can a producer P and a retailer R trade, if they are separated by one or more bottleneck(s)? Answer: They trade the power with E.E. or with another local counterpart. Furthermore, they have a financial contract.

The idea of this principle is the following: you can always buy or sell electrical power. For example, you can trade on the electricity exchange. Hence, what is interesting for the commercial players is only the price. However by means of a financial contract, the players can lock the price. 13 The financial electricity market At the financial electricity market you cannot trade one single kWh. As mentioned above, the financial market is used for price hedging and risk management. Figure 14 illustrates how a financial contract works. The example illustrates a financial contract of the type called a futures contract. In the example, a retailer and a supplier have entered into a futures contract with a volume of 4 MWh and a hedge price of 65 EUR/kWh. In the example, the contracts so-called delivery period is a specific month (for instance, it may be June 2014).

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 16

Figure 14: Producer and retailer sign a future contract with hedge price 65 EUR/MWh; If, for instance, the average system price in the month concerned turns out to be 66 EUR/MWh; The producer pays the retailer 1 EUR/MWh * 4 MWh. If, for instance, the average system price in the month concerned turns out to be 63 EUR/MWh; the retailer pays the producer 2 EUR/MWh * 4 MWh. In the example, the parties have cleared the contract. Hence, the settlement runs via clearing house.

The parties have a mutual insurance (and a mutual obligation). Suppose the average system price for the month in question turns out to be 66 EUR/MWh. A high price on the wholesale market is obviously disadvantageous for the retailer. However in this situation, the supplier will compensate the retailer. The supplier pays the retailer 1 EUR/MWh * 4 MWh = 4 EUR. Suppose instead the average system price for the month in question turns out to be 63 EUR/MWh. A low price on the wholesale market is obviously disadvantageous for the supplier. In this situation, the retailer will compensate the supplier. The retailer pays the supplier 2 EUR/MWh * 4 MWh = 8 EUR. The contract is therefore settled by comparing the hedge price of the contract with the average system price for the period in question. The difference in price is multiplied by the contracts volume. Eventually, this amount of money is transferred between the parties. It is important to note that the parties of a financial contract are not delivering physical power with each other. Only money is exchanged between them (therefore, the name financial electricity market). However, in addition, the retailer may submit a purchase order with an unspecified price to Elspot. The retailer can notify Elspot that he will buy 5 MWh each hour during the month irrespective of the price. With a purchase of 5 MWh each hour during the whole month, the retailer will in total have bought 3600 MWh by the end of a 30-days month: 5 MWh/h * 24 h * 30 days = 3600 MWh.
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 17

The retailer does not need to worry about the price. If it is higher than 65 EUR/MWh, he will be compensated. On the other hand, if the price is lower than 65 EUR/MWh, he has to compensate the opposite party of the futures contract. The retailer, therefore, has two trade arrangements: a purchase on Elspot and a futures contract. In total, the two trade arrangements guarantee his price for the 3600 MWh will be 65 EUR/MWh. 14 Clearing of financial contracts The two parties of a financial contract can choose to clear the contract using a clearing house. In this case, the clearing house takes care of the settlement of the contract (Figure 14). Furthermore, the clearing house guarantees the settlement: the clearing house will ensure that the settlement is carried out, even if one of the parties cannot fulfill his obligations. If the parties have entered the contract via a financial electricity exchange, clearing is mandatory. This is because the trading at the financial exchange is anonymous: the parties do not know each others identity. Hence, the contract must be cleared, so the clearing house sits between the parties. 15 Long-term contracts At Elspot, the commercial players can trade power day-ahead. Now, let us take a look at the market for long-term contracts. For example, let us consider a retailer who has sold 100 MWh to an end user at a price of 67 EUR/MWh for the following year. The retailer now has to make a corresponding purchase on the wholesale market. However, the retailer does not need to buy the power immediately. In order to hedge his position, all the retailer needs now is a futures contract. For example, the retailer has earned 2 EUR/MWh if he enters into a futures contract with a hedge price of 65 EUR/MWh. Next year, the retailer can simply buy the power from Elspot or from a local supplier. Therefore, the financial market is also the market for long-term contracts. 16 The day-ahead price must be reliable As it appears, the Elspot day-ahead price is used, when the financial contracts are settled. We say that the day-ahead price is the underlying reference for the financial contracts.
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 18

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

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 19

Singapore energy market


September 2012

Further information If you would like further information on any aspect of this client note, please contact a person mentioned below.

Contact Alex Wong Partner T +65 6302 2557 alex.wong@hllnl.com Amy Lee CEO T +65 6302 2558 amy.lee@hllnl.com Ming Hui Chock Senior Associate T +65 6302 2560 minghui.chock@hllnl.com Adrian Wong Senior Associate T +65 6302 2568 adrian.wong@hllnl.com

This note is written as a general guide only. It should not be relied upon as a substitute for specific legal advice

Contents

Introduction Recent developments in the Singapore energy industry Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations Schedule 2 Summary of Singapore gas market deregulation and the Gas Network Code

1 4 15

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. POWER MARKET UPDATE The Singapore power market has witnessed, over the last few years, both expansion and consolidation as it looks forward to the coming years of energy market development. From a consolidation perspective, almost all of the large incumbent power generators have undergone repowering exercises and moved almost exclusively towards CCGT plants for power generation as they reposition themselves in a more competitive wholesale market for power generation. Today, more than 80% of Singapores electricity is fuelled by piped natural gas.

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

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

2 1

Policy on Self-Supply of Electricity Information Paper, Energy Market Authority of Singapore (EMA), 21 April 2008

Full retail contestability has still not been achieved so household and small consumers retain the Market Support Services Licensee as their default power retailer.

Singapore energy market

Introduction Recent developments in the Singapore energy industry (contd)

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

and East Africa and shale gas in North America), LNG prices are set to fall in the medium to long term in spite of strong demand from China and Japan. At the same time, the West Natuna and South Sumatra gas fields in Indonesia appear to be depleting at a higher rate than originally envisaged and this, coupled with pressure on the Indonesia government to reserve more of its precious resources for domestic consumption, would likely mean a far less reliable future for piped natural gas imports. This has led to BGs 3 Mtpa franchise for aggregating LNG for Singapore having a strong uptake. With 2.65 Mtpa of LNG having been taken up by February 2012, the EMA expects the entire franchise to be fulfilled by 2013 at the latest. This begs the immediate question of Singapores LNG import policy beyond BGs franchise. The EMA has sought feedback 3 through a consultation exercise that will aim to determine whether Singapores future LNG import framework will be carried out through a regulated sole importer framework (BG+1) or a multiple aggregator framework (BG+3). Under the BG+1 framework, the EMA will appoint a Regulated Sole Importer (RSI) who will import all incremental LNG beyond BGs 3 Mtpa supply. The EMA will regulate the RSIs returns as well as its LNG procurement, gas sales prices and contract terms. A variant of this framework is adopted by Asian importers like South Korea, Taiwan and Thailand. It is expected that the appointment will be conducted through a similar RFP exercise that resulted in BGs appointment. The BG+3 framework would see up to 4 large LNG importers serving up to 15 Mtpa of LNG capacity (which is the projected LNG demand by 2024) in Singapore. Under the government mandated aggregation variant of this model, importers will be selected by EMA (either sequentially over time or concurrently) and awarded LNG import licences through a competitive RFP process. Each importer would be awarded a franchise to import a specific amount of LNG. Under the market-driven aggregation model, competition between players in the LNG import sector would give rise to natural aggregation of demand into a few dominant players (e.g. the formation of a few buyer groups). EMA would set entry criteria that importers must fulfil to qualify for access to the LNG terminal. With the LNG terminal able to support 7 LNG storage tanks and up to 15Mtpa of LNG demand, the Singapore government is also open to seeing the LNG terminal being used as a trans-shipment and trading hub. As of today, a number of international LNG trading companies (such as Shell, GDF Suez, ConocoPhillips and BP) have set up LNG trading offices in Singapore and they will be looking to tap the strong regional growth of spot and short-term LNG contracts. There should be no doubt though that the government sees the LNG
3

Significant focus has been placed by the Singapore government in reducing the reliance on these piped natural gas imports and increasing gas imports from other sources. This led to the Singapore government moving ahead with the development of its first LNG import and regasification terminal and the appointment of BG Asia Pacific Pte Limited (BG) as Singapores first LNG aggregator. At the time this decision was made, piped natural gas into Singapore was significantly less expensive than the comparable LNG price but the decision to proceed was nonetheless made on strategic grounds and ignoring the immediate economics of the decision. Singapore had to create an artificial demand for LNG in order to wean the market off its reliance on piped natural gas. This artificial market was created by simultaneously requiring power generators to accept LNG as part of their fuel portfolio as well as imposing a moratorium on future piped natural gas imports. In hindsight, the strategy appears to have been prescient. With new sources of gas flooding the global market (both traditional gas fields in Australia (which is set to overtake Qatar as the largest LNG exporter by the end of the decade)

LNG Procurement Framework Consultation Paper, EMA, 30 March 2012

Singapore energy market

Introduction Recent developments in the Singapore energy industry (contd)

terminals first priority as that of helping to ensure security of fuel supply for Singapore. With decisions on the LNG import strategy likely to be made in the coming months, the EMA will likely also have to resolve the wider question on the long term fuel mix strategy and the on-going question of how LNG imports will affect the import of piped natural gas. Given the long term investments and decisions required with respect to new piped natural gas import contracts, the EMA is aware that it will not be able to make sharp changes in direction once its long term policy and strategy have been settled. Amongst other things, the EMA will have to make challenging decisions on issues such as LNG imports into the nearby Malaysian LNG import terminals. Where regasified LNG is piped into Singapore from these terminals, they will likely be considered to be piped natural gas and thus subject to EMAs current moratorium. The extent to which the moratorium may be lifted for such future imports (and thus a potential direct challenge to the business of the Singapore LNG terminal) will be a difficult decision that the EMA will have to make in terms of striking a balance between market forces and price competitiveness and security of supply. SINGAPORE ENERGY MARKET INVESTMENTS IN THE NEAR FUTURE The business of the Singapore LNG terminal is, of course, of significant interest to the EMA and the Singapore government as a whole. The initial intention was for the LNG terminal to be a self-sustaining business model to be run independent of government subsidy (but albeit subject to regulation by the EMA). The decision for EMA to take over terminal development reflected the urgency and importance with which the government saw the terminal as a cornerstone of its energy strategy. That said, it remains the intention of the government to, at an appropriate juncture, divest the ownership of the LNG terminal to the private sector (perhaps similar to the way in which the previous Temasek owned generating companies were divested in 2008) a decision that will no doubt attract significant market interest. Whilst natural gas dominates the discussions on energy policy and strategy, there remain other elements of the governments overall energy diversification strategy that deserve a mention. The EDB has released an RFP for a synthetic gas facility (through coal gasification) to be developed on Jurong Island to provide feedstock for industry and potentially also for power generation. The ability of the syngas facility to provide fuel for power generation is nonetheless subject to EMAs overall study and policy finalisation on gas import control. Electricity imports also remain on the horizon (with a possible 4-5 year timeline from 2012). Subject to a restriction of 600MW per country (under the ASEAN Grid masterplan) any such import will have to be licensed by the EMA and will likely involve a contracts-fordifferences pricing structure. Similar to the aggregator

selection regime, any electricity import selection will likely be subject to a similar tender process. In summary, the Singapore energy market is best described as a managed privatised market. Singapore is adamant not to fall into the cycle of underinvestment and high energy prices that befalls many other privatised energy markets around the world but yet wants to take full advantage of deregulated market practices. The strategy appears to have worked so far but its continued success will depend on the ability of the EMA and the Singapore government as a whole to balance the supply and demand pressures of the merchant market with the strategic priorities of the city state.

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations4

HISTORICAL BACKGROUND The electricity and piped gas industries in Singapore have traditionally been vertically integrated and government-owned. The Public Utilities Board (PUB) was formed in May 1963 to undertake the supply of water, electricity and piped gas to the population of Singapore. In 1995, the Government began to implement a number of changes to deregulate the electricity industry. On 1 October 1995, the PUB transferred its electricity and gas activities to Temasek Holdings. Within Temasek Holdings, Singapore Power was created as the holding company for several other new companies including the generation companies, PowerSenoko (now known as Senoko Power) and PowerSeraya; the transmission company, PowerGrid; and SP Services Ltd, the electricity supply and utilities support services company. A further generator, Tuas Power, was set up as an independent company directly under Temasek Holdings. THE NATIONAL ELECTRICITY MARKET OF SINGAPORE 4 (NEMS) On 1 April 2001, the Government established a body corporate, the EMA, under the Ministry of Trade and Industry (MTI), to regulate, among others, the electricity industry. In that same year, PowerGrid transferred its system operator function to the Power System Operator (PSO), and the market operator function and pooling and settlement responsibilities to the Energy Market Company Pte Ltd (EMC), which was formed as a subsidiary of the EMA to operate the Singapore Electricity Pool (SEP) and subsequently the wholesale electricity market in the NEMS. The NEMS consists of a wholesale electricity market and a retail electricity market. The wholesale market consists of two markets: the real-time (or the spot market) for energy, reserve and regulation; and the procurement market for other ancillary services. The sale of energy, reserve and regulation are done through price/offer quantities submitted by generation companies every half hour. In addition, as part of the governments policy of separating ownership of electricity generation assets from ownership of the Transmission and Distribution Systems, Singapore Power divested its ownership interests in Senoko Power and PowerSeraya to Temasek Holdings. Temasek Holdings subsequently divested Tuas Power, Senoko Power and PowerSeraya to private sector investors.

REGULATORY FRAMEWORK Electricity Act The Electricity Act is the principal legislation governing the electricity industry and the NEMS. The principal rights and obligations of the participants in the wholesale and retail electricity markets are set out in the Singapore Electricity Market Rules (the Market Rules), the electricity licences and in the codes of practice (the Codes of Practice) issued by the EMA. The Singapore Wholesale Market Rules The Market Rules are effectively contracts between each market participant and the EMC under section 49 of the Electricity Act. This ensures that market participants have the option to take legal action against the EMC for damages sustained as a result of the non-observance of the Market Rules by the EMC and vice versa. The Market Rules also contain dispute resolution procedures. The objectives of the Market Rules are: to establish and govern efficient, competitive and reliable markets for the wholesale selling and buying of electricity and ancillary services in Singapore; to provide market participants and the Market Support 5 Services Licensee (the MSSL) with non-discriminatory access to the transmission system; and to facilitate competition in the generation of electricity.

Electricity Licenses Under the Electricity Act, an entity may not engage in certain electricity-related activities unless it has been issued with an electricity licence by the EMA (or it has been exempted from holding one). The electricity-related activities that require an electricity licence are:

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

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

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

trading in the wholesale electricity market; and importing or exporting electricity.

conditions, the Market Rules and Codes of Practice. The table in page 6 sets out the various market agreements and 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 System. Regulated Supply Service Code The Regulated Supply Service Code is binding on MSSLs, which is currently SP Services only. It sets out the minimum conditions that a MSSL must meet in carrying out its obligations to procure the supply of electricity and provide market support services to non-contestable consumers under section 21 of the Electricity Act. Market Support Services Code The Market Support Services Code (the MSS Code) is binding on MSSLs. It sets out the minimum conditions that a MSSL must meet in carrying out its obligations to provide market support services to Retail Electricity Licensees (RELs) and contestable consumers, and facilitate their access to the wholesale electricity market. Metering Code The Metering Code is binding on the Transmission Licensee, generation licensees and MSSLs, and sets out the minimum conditions that a metering equipment service provider must meet in carrying out its obligations to install and maintain meters. It also sets out the roles and obligations of the meter reader and meter data manager. Codes of Practice for RELs The Codes of Practice for RELs sets out the minimum standards of behaviour that a REL must observe in retailing to consumers. Market Agreements and Contracts Most market participants in the NEMS are required to enter into a number of agreements and contracts. These are generally a consequence of their respective licence

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

Agreements and Contract Parties Operating Agreement PSO and SP Power Assets

Purpose 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. 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. This agreement provides for meter reading services for wholesale settlement. The generation licensees pay directly for the services with fees set by the MSSL. This agreement establishes a contractual relationship between the PSO and the market participant and provides that the Market Rules will have the effect of contract as between the PSO and the market participant in so far as it applies to them. 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. This agreement gives effect to the obligations that must exist between SP PowerAssets and the party wanting connection service. This agreement is for the collection of Use of System charges (UoS Charges), that is, the transmission and distribution tariffs, when a REL opts for consolidated billing, that is, when a REL assumes the payment responsibility of its customers for the transmission charges. This agreement is an agency agreement for the provision of UoS Charges collection services for SP PowerAssets. This agreement is a contract between the EMC and a market participant (usually a generation licensee) supplying ancillary services.

MSSL EMC Agreement

EMC and MSSLs

MSSL Market Participant Agreement

MSSL, generation licensees and direct market participants (DMP) PSO, generation licensees, RELs and DMP

PSO Market Participant Agreement

MSSL Agreement

MSSL and RELs

Connection Agreement

SP PowerAssets, generation licensees, party wanting DMP and electricity consumers SP PowerAssets and RELs

Retailer Use of System Agreement

Agency Agreement

MSSL and SP PowerAssets

Ancillary Services Agreement

EMC on behalf of the PSO and the generation licensees

Source: EMA

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

ELECTRICITY INDUSTRY STRUCTURE The EMA The EMA was established in April 2001 pursuant to the EMA Act as an independent regulator overseeing the electricity and gas industries in Singapore. Under section 3 of the EMA Act, the EMA is charged with the general administration of the EMA Act, and its functions and duties include: to perform the interests of consumers with regard to prices, reliability and quality of services; to perform the functions of economic and technical regulator; to ensure that electricity licensees provide an efficient service; to ensure security of supply of electricity to consumers and to arrange for the secure operation of the transmission system; to protect the public from dangers arising from electricityrelated activities; to create an economic and regulatory framework for the electricity sector that promotes competitive, fair and efficient market conduct and prevents the misuse of monopoly or market power; to advise the Government on matters relating to the electricity system; and In fulfilling these functions, the EMA has at its disposal a number of regulatory tools and powers. These include the authority to issue, suspend, revoke or modify an electricity licence; the power to issue and modify codes of practice and other standards of performance; the power to issue directions to electricity licensees; the power to fine electricity licensees; and the authority to investigate and sanction anti-competitive conduct.

facilitate the planning Transmission System;

and

augmentation

of

the

provide information and other services to facilitate decisions for investment and the use of resources in the electricity industry; and exercise and perform the functions, powers and duties assigned to the EMC under the EMA Act, its electricity licence, the Market Rules and applicable Codes of Practice.

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

maintaining the reliability of the electricity system; forecasting and reporting Transmission System; on conditions on the

coordinating the outages of generation facilities; providing Transmission System status and load forecasting to the EMC for the purposes of market clearing; coordinating the actions of the EMC and market participants during emergencies; and dispatching generation facilities.

The EMC The EMC is licensed to operate the wholesale electricity market in the NEMS. The EMCs functions are to: operate and administer the wholesale electricity market in the NEMS; prepare schedules for generation facilities, loads (that is, the withdrawal of electricity from the Transmission System) and the Transmission System; settle accounts of market participants;

Market Participant A Market Participant in the NEMS is defined as a person (that is, an entity or organisation, as well as people) that:

has an electricity licence issued by the EMA; and has been registered with the EMC as a market participant.

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

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

the Transmission Licensee; generation licensees; RELs; persons, other than the generation licensees and RELs, who have been licensed to trade in the wholesale electricity market; and any department of the Government that generates electricity before 1 April 2001.

RELs The retail electricity market does not come under the jurisdiction of the Market Rules and the EMC. It is created and regulated under the EMA Act, the electricity licenses and the Codes of Practice issued by the EMA. RELs may be market participants who purchase electricity directly from the wholesale electricity market or purchase through the MSSL. Since the RELs are permitted to trade in electricity and are not subject to the same degree of regulation as the MSSL, they may offer contestable consumers contracts different from those available from the MSSL. RELs can bundle energy and other charges into a single invoice, charge a price other than the Uniform 6 Singapore Energy Price (the USEP) for energy, and offer additional services to consumers. Consumers Consumers are classified as either contestable or noncontestable, depending on their electricity usage. Contestable consumers are entitled to purchase electricity from a REL, or directly from the wholesale electricity market, or indirectly through MSSLs. Non-contestable consumers are supplied by MSSLs. Currently, consumers with a monthly usage of 10,000kWh and above are contestable. The EMA continues to study when full contestability of all retail consumers will be allowed. Relationships between the Market Participants The figure on page 9 shows the financial flows between the Market Participants in the NEMS.

A MSSL is not a market participant. In the NEMS, it is mandatory for all generation facilities above to be licensed by the EMA. It is also mandatory for generation facilities above 10MW to be registered for dispatch by the PSO. Mandatory registration ensures that all generation facilities of any significant size are subject to the Market Rules. Transmission Licensee SP PowerAssets is currently the sole Transmission Licensee in Singapore. The responsibilities of SP PowerAssets and of the persons whose facilities are connected to the Transmission System are set out in the Transmission Code and the connection agreements. The Market Rules also contain specific provisions for SP PowerAssets and the PSOs obligations in respect of the reliability and security of the Transmission System. The transmission network transports electricity at high voltage from generators to the low voltage distribution network (or, in a small number of cases, directly to large industrial consumers). SP PowerAssets, being the monopoly provider of transmission services, is not permitted to compete in the energy market, whether as a generator, retailer or trader (either directly or indirectly by ownership of companies engaged in such activities), because opportunities exist for it to afford a preference to its competitive activities or its competitive affiliates. Market Support Services Licensee SP Services is currently the sole MSSL and provides market support services to the majority of electricity consumers in Singapore. SP Services charges regulated fees, as approved by the EMA, for market support services provided.

Please refer to page 13 for further discussions on the USEP.

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

Financial Flows between the Market Participants in the NEMS

Competition An important role for the EMA is the monitoring of market behaviour to guard against the concentration of market power, abuse of dominance and exercise of anticompetitive behaviours such as collusion and capacity withdrawal to drive up prices. The EMA is also empowered to prohibit anticompetitive practices including restrictions on mergers and acquisitions (that is, there will be cross ownership limitations). Significantly, the anti-competition provisions in the Electricity Act have retrospective effect. Where the EMA finds that a person has acted in an anti-competitive manner, the EMA can give a direction for the cessation of such conduct and impose a financial penalty on the person. The electricity industry is excluded from the Competition Act 2004 as the EMA will continue to regulate anti-competitive practices and other competition issues in the electricity industry.

OVERVIEW OF WHOLESALE MARKET OPERATIONS

In the NEMS, the real-time dispatch of electricity (scheduling generators to supply energy, reserve and regulation) is determined by the operation of a wholesale spot market run every half-hour. Generators offer their capacity (specifying price/quantity pairs) to the market and the PSO provides a prediction of the expected load along with any system constraints for that half-hour. The market then determines the least-cost dispatch quantities and the corresponding market clearing prices based on the offers made by generators. The wholesale electricity market consists of two markets:

the spot market for energy, reserve and regulation; and the procurement market for other ancillary services.

Every half-hour, the Market Clearing Engine (MCE), a linear programming computer model, determines the spot market outcomes for:

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

10

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

the dispatch quantity produced by each generation facility; the reserve and regulation capacity required to be maintained by each generation facility; and the corresponding wholesale spot market prices for energy, reserve and regulation.

electricity. There is no overall cheaper dispatch available, in terms of the offers that have been made to the market, by providing energy, reserve and regulation from different sources, or in different quantities from the same sources. This is the minimum cost market dispatch. The supply of energy, reserve and regulation is specified by means of offers that contain price/quantity tranches indicating the quantity of energy, reserve or regulation that each dispatchable generation facility is willing to supply at the corresponding energy, reserve or regulation prices. With the dispatch based on estimates of the demand for the coming dispatch period, the market prices are similarly the prices set according to those estimates. This form of price setting, called ex ante pricing (pricing before the event), gives the market participants certainty about market prices even if 8 dispatch quantities differ from those scheduled . The MCE does not produce a single market energy price because of the effects of losses and congestion on the transmission system. Different energy prices apply to different nodes on the transmission system. THE OFFER PROCESS Energy, Reserve and Regulation Offers Generators make offers to supply energy, reserve, and regulation for each of their units in each half-hourly dispatch period in which they want to operate. They are similarly permitted to offer interruptible load to supply reserve. Offers can vary for each half-hour, and are assumed to stand, unless modified, from the time they are made through to dispatch. The market does not distinguish between offers used for the market outlook, pre-dispatch and real-time processes. It simply uses the most recent offer made for each half-hour. Key features of the generator offer process are: standing offers are required. Generators are required to make standing offers into the market. The standing offers form a pattern for a week. The use of standing offers is particularly valuable for smaller generators, since it eases the administrative burden of participating in the market; continuous adjustment of offers. Market participants are allowed to continually adjust their offers up to gate 9 closure ;

Quantities and prices are based on price/quantity offers made by the generators and load forecasts prepared by the EMC based on demand forecast information received from the PSO. RESERVE AND REGULATION Reserve is unused capacity that has to be made available to the electricity system quickly to correct any imbalance and maintain reliable supply in case of an unexpected outage of a scheduled generation facility. This capacity must be able to be in production within a short timeframe, depending on the arrangement. There are three reserve classes: primary reserve (eight seconds response), secondary reserve (30 seconds response) and contingency reserve (10 minutes response). Regulation, or load-following, is a normal operational requirement to cover second-to-second variations in load away from estimated load. DISPATCH SCHEDULING Dispatch scheduling is the process of matching the generation capacity needed to meet forecast demand. It is at the heart of running an electricity system. To enable the MCE to generate the dispatch schedule, the PSO and the generation facilities dispatch coordinators need to know in advance when each generation facility will be operating and how much output is expected from each. The dispatch schedule comes from the MCE. The PSO instructs the generation facilities to conform to the dispatch schedule. Any deviations from the estimated load and corresponding schedule are handled by the PSO using ancillary services. THE MARKET CLEARING ENGINE Every half-hour, the MCE is run to determine the dispatch schedule and the associated energy market prices for the upcoming dispatch period. The MCE also determines which generation facility is on reserve and regulation duty along with the market prices for reserve and regulation. The objective of the MCE is to find a set of dispatch instructions that minimises the cost of supplying load at all nodes (injection or exit points) of the Transmission System, as well as meeting the reserve and regulation requirements for

For all but plants providing regulation, energy actually injected should not differ greatly from scheduled energy, under normal circumstances. Although offers modified within the last hour may be subject to scrutiny from the market surveillance panel.

11

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

up to 10 price/energy quantity bands. Generators may make energy offers consisting of up to 10 price/quantity bands (tranches) for each facility for each half-hour; up to five price/reserve and regulation quantity bands. Generators and interruptible load may make reserve offers (of different classes) and generators may make regulation offers if they are registered to do so; combined offers. Energy, reserve and regulation are all offered simultaneously, and are co-optimised by the market clearing model. The model respects the trade

offs between the commodities so that a facility will not be scheduled to produce more energy, reserve and regulation than it can simultaneously manage; and offers at a node. Energy offers for each generation facility are made at the node where that facility is located.

Market Clearing Process in the Spot Market Prices in S$ per megawatt hour (MWh)

12

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

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

(600MW thermal units, for example) are large relative to the total load. A generator would wish to receive payment for the reserve and regulation it provides because it forgoes the opportunity of being dispatched fully, by being partially available for PSO to call on it for reserve and regulation, as and when required. Reserve in Singapore can be provided by generation facilities 10 and load . For a facility to provide reserve as quickly as in eight seconds or even in 30 seconds, it needs to be already spinning and synchronised. In most instances, that requires the unit also to be supplying energy, with reserve capability coming from its ability to ramp up its scheduled output very quickly. Since not all plants have this technical capability, a plant has to be certified as meeting the requirements for registration to provide reserve before it can be offered in the reserve market. There are also different reserve provider groups for each class of reserve. These groups represent the reliability of different reserve sources in providing reserve, and their effectiveness in curtailing falls in system frequency. Since a facilitys capacity may be available for both energy and reserve/regulation, the MCE must consider the optimal trade-off between the offers for reserve, regulation and energy. In solving the markets for each class of reserve and regulation, the MCE simultaneously finds the lowest cost solution (in terms of the offers made) that trades off between these products for the various facilities. Within the MCE, optimisation of the supply of energy must account for the minimum running level of facilities that provide reserve. The overall optimal solution may result in a unit being run out of 11 merit for energy so that the unit is available for reserve . Offers for reserve from a generator can only be made in association with a corresponding offer for energy. Part of the standing capability date for the plant is a function relating its reserve capability to its energy capability. This relationship is entered into the MCE. The cost of regulation is recovered from consumers and generation facilities. The cost is allocated on a S$ per MW basis across all MW of consumption in a dispatch period plus the first 10MW of electricity dispatched by each generation facility in that dispatch period.

10

An Interruptible Load scheme was introduced into the Singapore market in 2004. Since often a facility must be running in order to be available for reserve, it may be dispatched for energy even though its energy offer is higher than that of the marginal plant for energy. This is acceptable because there is no cheaper energy and reserve solution for the system as a whole. The reserve price received by such a plant will compensate it for the shortfall between its energy offer price and the energy spot price.

11

13

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

USEP While the generation facilities are paid their nodal price, buyers from the wholesale electricity market pay a uniform overall average price so that no customers are disadvantaged by location. The USEP is calculated from the weighted average of the nodal prices at all of the exit nodes on the Transmission System. The nodal energy price at each node is weighted by the energy withdrawn from that node. VESTING CONTRACTS In the transition to the NEMS, the EMA had concerns with the degree of market power that will exist in the wholesale electricity market. The EMA has addressed these concerns using Vesting Contracts without interfering with the structure of the wholesale electricity market. Vesting Contracts are contracts for differences (CfDs) vested on the large incumbent electricity generation companies, for a transitional period. In Singapore, the Vesting Contracts take the following form: generators will be required to enter into Vesting Contracts with the MSSL, who is the counterparty to all of the Vesting Contracts. The MSSL will then distribute debits and credits associated with the contracts to consumers, both contestable and non-contestable; Vesting Contracts have a contract price (or strike price) set at about the economic or long-run marginal cost (LRMC) of a new entry generator (i.e. the electricity price that a new investor in base-load capacity would require in order to cover its fixed and variable costs at a reasonable return to shareholders over the life of the plant). The same strike price applies to all generators; the contract quantity will be set to keep the market power of large generators at an acceptable level. During peak load times, the contract quantity will be a larger proportion of total load, while in off-peak times it will be a smaller proportion. The average contract quantity will reduce over time as new capacity is built to mitigate the market power of incumbents; and the contract quantities for each generator are based on the generation capacity for each company.

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

14

Singapore energy market

Schedule 1 Summary of Singapore electricity market deregulation and wholesale market operations (contd)

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

Electricity Market Design for a Low-carbon Future


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

The Energy Supply Theme of UKERC UKERCs energy supply research activities are being undertaken by the University of Cardiff, Imperial College and the University of Exeter.

Energy Research Centre

UKERC/WP/ESM/2005/004

This paper considers GB electricity market and network regulatory arrangements in the context of transitioning to a low carbon electricity system. By considering some of the primary features of a low carbon electricity system and building on themes raised by a previous UKERC Supply Theme paper (Baker, 2009), the paper attempts to identify what characteristics an appropriate market and regulatory framework would need to posses. The paper goes on to consider how existing market arrangements perform in these areas and the possible need for change. The aim of the paper is to contribute to the debate on energy market reform that is now underway. Currently, discussion seems to be focussing primarily on how to ensure adequate investment in low carbon and, in the medium term, conventional generation to meet the UKs climate change and security of supply goals. Delivering the necessary generation capacity is clearly crucial and by reviewing some of the mechanisms that could be used to encourage investment, this paper attempts to contribute in this area. However, the paper also addresses other areas where reform may be required but that have, to date, received less attention; issues such as arrangements to ensure efficient dispatch and energy balancing, efficient mechanisms to deal with network congestion and measures necessary to facilitate demand side participation. The approach taken by the paper is incremental in nature, focussing on how current market arrangements may need to develop in the coming years, rather than proposing radical change. It is likely that successfully decarbonising the electricity sector may ultimately require a fundamentally different market design and that change, particularly in relation to low-carbon investment, may be required sooner rather than later. However, the transition to a low carbon electricity system will be gradual and arguably best served by incremental change in response to demonstrated need.

Contents
1. SUMMARY .......................................................................................................... 3 2. A LOW CARBON ELECTRICITY SECTOR AND ITS IMPLICATIONS FOR MARKET DESIGN .. 4 3. GENERATION INVESTMENT ................................................................................... 5 3.1 THE GENERATION INVESTMENT CHALLENGE ......................................................................5 3.2 IMPACT OF WIND ON ENERGY PRICES ..............................................................................6 3.3 ENERGY-ONLY MARKETS ............................................................................................8 3.4 ENCOURAGING GENERATION INVESTMENT ......................................................................10 3.4.1 Output-based mechanisms. ........................................................................11 3.4.2 Capacity-based mechanisms ......................................................................12 3.5 A SINGLE BUYER .....................................................................................................16 4. ENERGY DISPATCH AND BALANCING IN A LOW CARBON ELECTRICITY SYSTEM ......... 17 4.1 PROBLEMS WITH EXISTING MARKET ARRANGEMENTS. .........................................................17 4.2 SYSTEM RESERVES ...................................................................................................19 4.3 MARKET LIQUIDITY .................................................................................................19 4.4 A SEPARATE MARKET FOR INTERMITTENT GENERATION? .....................................................20 4.5 THE CASE FOR A MORE INTEGRATED APPROACH TO MARKET DESIGN ......................................21 4.6 INTEGRATED MARKETS AND THE NEED FOR PRIORITY DISPATCH ............................................21 4.7 BALANCING MARKET SIGNALS WITH DEPLOYMENT RISKS FOR WIND. .......................................22 5.NETWORK CONGESTION AND APPROPRIATE NETWORK INVESTMENT SIGNALS........... 23 5.1 CONGESTION VOLUME .............................................................................................23 5.2 MINIMISING CONGESTION COSTS .................................................................................25 5.3 TRANSMISSION INVESTMENT SIGNALS ...........................................................................26 6. ENCOURAGING DEMAND-SIDE PARTICIPATION..................................................... 26 6.1 REDUCING CAPACITY AND RESERVE ..............................................................................27 6.2 SETTLEMENT IMPACTS ..............................................................................................27 7. NETWORK REGULATION .................................................................................... 29 7.1 ENSURING EFFICIENT NETWORK UTILISATION...................................................................30 7.2 DELIVERING NETWORK INVESTMENT IN A TIMELY FASHION ..................................................31 7.3 FUNDING NECESSARY NETWORK INVESTMENT ..................................................................32 8. CONCLUSIONS .................................................................................................. 33 8.1 ENCOURAGING GENERATION INVESTMENT ......................................................................34 8.2 ENERGY DISPATCH AND BALANCING .............................................................................36 8.3 DEMAND RESPONSE ................................................................................................38 8.4 NETWORK REGULATION............................................................................................39

1. Summary
The electricity system will have a pivotal role in delivering the UKs climate change obligations and longer term aspirations. The need to accommodate large amounts of renewable, mainly intermittent, generation by 2020, replace generation expected to decommission in the same timescales and the need to effectively decarbonise the electricity system by 2030 through the introduction of new low-carbon technologies, represent huge challenges to be overcome. In addition, the need to partially electrify the heat and transport sectors with the introduction of heat pumps and electric vehicles will require further investment in generation capacity and could, if not adequately managed, place additional strains on the electricity infrastructure. If these challenges are to be met, some aspects of electricity network regulation and market arrangements will need to change. While the current arrangements have arguably served us well, delivering secure electricity supplies and driving out unnecessary cost, they are designed around controllable conventional generation capacity serving demand that varies in a predicta ble fashion. Tomorrows electricity sector will, however, look very different with a more flexible demand base required to accommodate a low carbon generation fleet that contains large amounts of high capital cost, intermittent and inflexible capacity. By considering some of the primary characteristics of a low carbon electricity sector, this paper attempts to identify things that an appropriate market and regulatory regime will need to do well. The analysis suggests that the highly disaggregated, energy-only and illiquid nature of the current market, reinforced by asymmetrical and non cost-reflective imbalance charges, may not be the most appropriate arrangement for dealing with intermittent renewable generation. A return to a more integrated market design is proposed, operating seamlessly down to real time in order to provide the liquidity and near real time balancing opportunities necessary to accommodate intermittent generation technologies such as wind. A more integrated electricity market would allow reserves, energy, and potentially, network requirements, to be optimised simultaneously. The introduction of wind and other intermittent generation technologies will cause energy prices to fall on average, but become far more volatile. This will make financing both capital intensive low carbon and peaking generation more difficult and, given the general scepticism over the ability of emissions trading to fully internalise the costs of carbon, there would seem to be a need for additional measures to support investment. Options include the extension of existing supplier-based obligations, Feed in Tariffs

(FiTs), capacity obligations or capacity payments. In addition, the more radical option of creating a central entity to procure both capacity and energy has been suggested. The need to retain substantial amounts of conventional plant to back-up intermittent generation can be expected to significantly increase network congestion. The current market arrangements, where network requirements are only considered one hour before real time, arguably encourage practices that increase congestion volumes and make the resolution of that congestion unnecessarily expensive. In addition to incurring unnecessary costs, which are ultimately borne by electricity customers, the current arrangements also cause network investment to appear overly attractive. The adoption of more integrated market arrangements would allow earlier consideration of network requirements and a more cost effective resolution of network congestion. The development of a more responsive demand side to accommodate a partially intermittent and inflexible generation fleet will be facilitated by the introduction of advanced or smart metering, where domestic and small commercial customers are metered on a half-hourly basis. This will require fundamental changes to the settlement processes and it will be necessary to ensure that increased data retrieval, handling and aggregation requirements do not impose unnecessary burdens on small customers or impede the development of a more responsive demand base. Finally, the paper briefly addresses some regulatory issues and makes the case for a regulatory environment that more effectively supports innovation and equalizes incentives for network investment and operational alternatives. Regulation will also need to ensure that network investments necessary to accommodate renewable and low carbon generation can be delivered in a timely fashion and that those investments can be adequately financed.

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

sequestration technology will be required by 2030 in order to decarbonise the electricity sector. Wind, nuclear and CCS technologies all have high capital costs and, as such, may not be plant that investors would necessarily choose to support. An appropriate electricity market design will, therefore, need to ensure that sufficient low-carbon capacity is brought forward. Furthermore, as much of this low-carbon capacity will be intermittent in nature, with output difficult to predict with any accuracy until close to real time, the electricity market will need to accommodate the increased short-term trading and balancing activity necessary to maintain security of supply. In fact it seems likely that these two issues, i.e. the need to ensure adequate investment in low-carbon technologies and accommodate high levels of short-term trading and balancing activity, will be the principle determinants in designing an electricity market for the future. Other issues influencing market design will be the need to generally minimise emissions, manage network congestion efficiently and accommodate a flexible and price-sensitive demand base. Minimising emissions will require that low-carbon generation has priority of use over conventional technologies and that overall dispatch efficiency is maximised. The need for an effective means of managing network congestion stems from the intermittent nature of technologies such as wind and the consequent need for conventional generation back-up, which would make the provision of sufficient network capacity to accommodate the simultaneous operation of all generation capacity unnecessary and prohibitively expensive. The requirement for market arrangements to facilitate the development of a flexible demand base is also associated with need to minimise the impacts of intermittency, both in terms of generation investment and energy balancing, and to allow the partial electrification of the heat and surface transport sectors to be accomplished in a cost effective fashion.

3. Generation investment
3.1 The generation investment challenge
The UK will need to invest heavily in generation capacity over the coming years. Deploying sufficient renewable and low-carbon generation to meet our climate change obligations while replacing plant expected to close as a result of E U Large Combustion Plant and Industrial Emissions Directives, is likely to require some 140 billion of investment by 2025 (Ernst & Young, 2009). Delivering the necessary investment will be all the more challenging given that many other countries will be embaking on similar programmes. It is estimated that global investment in generation could run at around $550 billion/year until 2030, with investment in Europe running at some 60 billion/year over the same period (E.on, 2009). This international dimension is particularly relevant given that the UK will be heavily dependent on large European-

based energy companies, operating away from their home markets, to deliver the investment in generation capacity required. The UK will, therefore, need to maintain a regulatory and market environment that is attractive to these companies, who clearly have choices in terms of where they invest. The following paragraphs in this section consider how the introduction of intermittent generation technologies such as wind make the investment challenge more difficult and why the current GB energy only electricity market may not be the most appropriate design to deliver the investment required to achieve our climate change goals. The section then moves on to consider the various options available for encouraging necessary investment, drawing on experience from the UK and overseas.

3.2 Impact of wind on energy prices


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

High wind /Mwh Price l (low wind ) OCGT Night Peak Low wind

CCGT & Coal Price h (high wind) Wind & nuclear

MWh Figure 1. Impact of wind on energy prices Indeed, as wind penetration increases, spot electricity prices may fall to zero and even go negative on those occasions when windy conditions coincide with periods of low

demand and wind generators attempt to retain access to operational subsidies 1. The negative impact of wind generation on wholesale electricity prices has been observed in countries such as Denmark, Germany and Spain, which have installed large amounts of wind generation as a proportion of their peak electrical demand (Poyry, 2010). In fact, the impact of wind energy on electricity prices could be even more pronounced in GB, due the island nature of the electricity system with little interconnection currently available to smooth variations in supply. While the injection of large amounts of zero-marginal cost energy will tend to reduce average wholesale electricity prices, the intermittent nature of that energy will introduce some additional, offsetting, costs. Intermittent generation such as wind cannot be relied upon to be available at any particular point in time and contributes little to security of supply. Back up resources in the form of flexible conventional generation or alternatives such as demand response, storage or support from adjacent systems via interconnection capacity, therefore need to be retained on almost a MW for MW basis 2 in order to operate when wind output is low. Conventional generation, typically CCGTs, operating in this role will experience decreasing utilisation as wind capacity builds, but be expected to operate more flexibly, i.e. starting and stopping more frequently and being part-loaded in order to provide both upwards and downwards reserv e. These modes of operation will introduce operational inefficiencies and associated costs, which will need to be spread over a reducing number of running hours. In addition to the impact of these operational costs, consumers will also need to bear the costs of retaining conventional back up plant in service and of eventually funding its replacement. A sustainable electricity system with high levels of intermittent renewable generation will require far more generating capacity than there is peak demand to be supplied and the fixed costs of this additional capacity will need to be supported through more volatile energy prices or, alternatively, mechanisms that reward capacity explicitly.

Renewable generation receives Renewable Obligation Certificates (ROCs) for each MW of energy generated. These can be sold on to suppliers to help meet their obligation to purchase energy from renewable sources, thereby creating an income stream. During periods when the combination of renewable output and that of inflexible sources such as nuclear exceed demand, it is worth renewable generation paying suppliers to take energy in order to retain access to the ROC income stream. In these circumstances the spot price of energy would enter negative territory.
1

Conventional generation is generally held to a have a 95% availability forecast error over peak demand periods. For wind to have a similar firmness, its capacity would need to be factored down to approximately 4% of installed capacity.
2

3.3 Energy-only markets


In common with many other jurisdictions both in Europe and elsewhere, GB operates an energy-only electricity market where non-subsidised, conventional generation relies on the difference between energy prices and variable cost to service its investment and other fixed costs3. Energy-only markets rely on the theory of peak load pricing (Kleindorfer), which describes how generation investment can be optimised through efficient pricing signals. For the majority of time, available generation capacity will exceed demand and wholesale energy prices will reflect the variable costs of the marginal plant. Low variable cost generation such as nuclear or wind will receive excess income when prices are set by higher-variable cost plant such as CCGTs or coal, contributing to their fixed costs. However, marginal and peaking generation will need to rely on periods of high energy prices associated with tight capacity margins, which may only occur for just a few hours per year. To work effectively, energy-only markets require demand to be sensitive to price. Spikes in energy prices caused by plant scarcity will be attenuated by price sensitivity and the value that different classes of demand place on an additional MWh of supply will be exposed. In this way, the market effectively determines how much generation capacity is required, rather than compliance with some arbitrary generation adequacy standard. In the absence of demand price sensitivity, as is the case in GB, the energy market may become distorted with the GBSO having to impose voltage reductions or physical disconnection to curtail demand in the face of generation shortages. The need to impose operational measures to curtail demand in the absence of any natural response to increasing price may result in some customers loosing access to supplies before prices have reached the level where they would restrict consumption voluntarily. Despite these general concerns, the GB energy-only market in the form of NETA and BETTA4 has been relatively effective in bringing forward new capacity. While there was a sharp drop in generation commissioning following the introduction of NETA in 2000, this was probably in part a reaction to the increased plant margins that applied in the latter years of the England & Wales Electricity Pool. The 16 GW of new, non renewable, capacity that is forecast5 to commission by 2015/16, suggests that the current arrangements are capable of dealing with the immediate requirements for generation

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

New Electricity Trading Arrangements (NETA). Introduced to replace the E&W Electricity Pool in April 2000. The bilateral trading arrangements introduced by NETA were extended to Scotland in April 2005 with the introduction of the British Electricity Trading & Transmission Arrangements (BETTA).
4 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 UKs climate change goals, which would inject a large amount of controllable demand and allow further demand smoothing. Increased interconnection and storage would have a similar effect. Notwithstanding this mitigation, the increase in zero and low marginal cost generation will undoubtedly challenge the ability of an energy-only market to deliver adequate levels of generation investment. There must be a limit to the extent to which price sensitivity, particularly during periods of cold weather that often coincide with calm conditions, can be expected to limit electrical demand. Moreover, energy price spikes will still be necessary to adequately reward low merit and peaking plant and there is a concern that periods of extreme, if temporary, energy prices may prove to be unacceptable from a political or regulatory point of view. Consumers exposed to real time electricity prices seem likely to press for prices to be capped and, given the year to year variability suggested by Poyry, it might be difficult to distinguish between justified price spikes and those resulting from an abuse of market power (Poyry, 2009). There is a danger, therefore, that regulatory or political interventions may result in measures that prevent energy prices from rising to the levels necessary to justify investment in new capacity. Indeed, regulatory and political pressures have resulted in the application of measures to contain wholesale prices in

Defined as the Value of Lost Load or VOLL, currently assumed to be around 4000/MWh.

many electricity markets and, while not currently applied in GB, price caps have been applied in the past. It is also worth noting that there are other mechanisms at work within BETTA that tend to attenuate spot prices, for example the use of contracted reserve contracts by the GBSO as an alternative to accepting more expensive Balancing Mechanism7 offers to adjust output in real time. In conclusion therefore, energy only markets can claim to have the virtue of relative simplicity, with reliability and generation investment set by market participants, rather than arbitrary rules. However, reliance on scarcity pricing to recover fixed costs increases investment uncertainties and finance risk. Furthermore, the ability of scarcity pricing to stimulate adequate investment will be tested to the extreme by the introduction of zero-marginal cost intermittent generation technologies, with the consequent decline in average energy prices and increased price uncertainty and volatility. While studies such as that carried out by Poyry do not, of themselves, make the case against energy-only markets and the need to reward for generation capacity explicitly, they do clearly demonstrate the challenges to be faced.

3.4 Encouraging generation investment


There appears to be an emerging consensus that existing market arrangements are unlikely to deliver the low carbon investment necessary to satisfy the UKs cl imate change ambitions (Ofgem, 2010), (Committee on Climate Change, 2010), (HM Treasury, 2010). Whereas none of this analysis currently goes beyond presenting options, a common theme is the need to introduce some form of mechanism, external to the main energy market, to encourage investment in low carbon generation capacity. In their Energy Market Assessment (EMA), HM Treasury/DECC set out five possible models for market reform, shown in figure 2. These models escalate in terms of intervention from simply adding a carbon floor price to existing market arrangements, the provision of additional low carbon incentives, regulation to limit investment in highcarbon technologies, the provision of long-term low carbon payments, and finally the creation of a central buyer for all generation capacity and output. In discussing the relative merits of these five options, the EMA concludes that Option A, which would introduce a floor price for carbon while leaving other market arrangements as they are, is unlikely to drive the pace and scale of investment required. At the other end of the intervention spectrum, the single buyer proposed by EMA Option E is discounted as not having sufficient benefits over less interventionist options that retain a competitive

Balancing Mechanism (BM). The BM commences at market closure, one hour before real time. Generators (or demand) submit bids and offers to vary output (or demand) and these may be accepted by the GBSO to ensure final energy balancing and that network congestion is resolved.
7

10

approach involving incentives, payments or restrictions in investing in conventional technologies.

Increasing centralisation
Option A Option B Option C Option D
Separate low carbon market

Greater carbon price certainty alone


Minimum carbon price guarantee at currently expected level Competitive market framework as today

Support low carbon in the current market


Additional incentives for low carbon generation price above carbon price Competitive market framework as today

Regulate to limit high carbon generation


Regulate to drive decarbonisation Competitive market framework as today

Option E
Single buyer agency

Long term payments to low carbon generators to provide revenue certainty Conventional plant trades in competitive market framework

Single agency is the only purchaser of electricity generation existing and new low and high carbon- and only seller of this on to suppliers

Figure 2. EMA market reform options The following paragraphs consider some of the alternatives for encouraging sufficient low-carbon and conventional generation capacity in the context of the Energy Markets Assessment Options B & D. The mechanisms considered include energy or output based obligations, such as the GB Renewable Obligation (RO), capacity obligations and capacity payments. In addition, and despite DECCs rather summary dismissal of their Option E, the possible merits of a single-buyer concept are considered.

3.4.1 Output-based mechanisms.


Since the demise of the Non Fossil Fuel Obligation, support for the deployment of renewable generation in the UK has been via the Renewable Obligation (RO) and, from April this year, via a feed in tariff (FiT) for smaller generation. Both are output-based support mechanisms that reward generators for producing renewable energy, and EMA Option B appears to propose extending this type of arrangement to other forms of low carbon generation. The Committee on Climate Change (CCC, 2010) also recommends an extension of FiTs and the introduction of a low carbon obligation to ease uncertainties over cost recovery, thereby reducing investment costs. While output based mechanisms have been effective in bringing forward low carbon investment worldwide and provide strong delivery incentives, they are not without problems. Quantity based output mechanisms such as the RO, for example, suffer from uncertainties over future ROC prices due to headroom issues (i.e. prices decline as the specified quantity is achieved) that increase investment risk and can also over -

11

reward the cheapest low carbon technologies (BERR, 2008). Price based mechanisms, such as FiTs, suffer from uncertainty in terms of actual response to the guaranteed price and, if that price is incorrectly set, can result in over or under supply (Cory, 2009). Output based mechanisms also have the potential to distort the energy market. As indicated in 3.2, increasing wind and nuclear capacity will give rise to the possibility of wind becoming the marginal plant when periods of low demand coincide with high wind output. During these periods, wind generation will seek to retain access to ROC income, driving energy prices into negative territory. Some analysis (Strbac, 2008) suggests that, taking into account the need to carry addition reserves on part-load thermal plant, up to 25% of wind energy may need to be rejected when wind capacity exceeds 30 GW. Clearly, this could have a serious impact of the financial viability of wind as well as other low carbon technologies such as nuclear, which will be dependent on high energy prices to recover investment costs. In order to avoid or reduce the prospect of negative prices, measures such as curtailing ROC or FiT payments during periods of excess low carbon output could be considered. Some possibility of low and damaging energy prices would remain however and attention is likely to turn to the deployment of additional storage or interconnection capacity in order to artificially boost demand. An alternative approach, albeit involving a degree of central planning, would be to take a more strategic view of the interactions between nuclear and wind generation and attempt to optimise the capacity mix. The premium FiTs discussed in EMA Option B, which top up revenues from the energy market and provide investors with a guaranteed income, have the advantage over standard FiT designs of keeping generation involved and interested in the energy market. Dispatchable renewable generation would be able to respond to energy price signals and therefore be less likely to contribute to negative price problems. However, intermittent renewable technologies such as wind are not dispatchable in any meaningful way and are less able to respond to price signals (Poyry, Elementenergy, 2009). By supplementing energy market revenues, premium FiTs therefore could still act in a similar fashion to the RO, particularly in the case of intermittent technologies, in encouraging negative biding during periods of excess low carbon output.

3.4.2 Capacity-based mechanisms


An alternative approach to encouraging investment in low carbon technologies, which would fit comfortably in EMA options B or D, would be to focus on low-carbon capacity rather than output. There are numerous examples world-wide where obligations are placed on suppliers to procure sufficient generation capacity to satisfy demand to some standard of supply security. To date, the focus of such mechanisms has been security of supply alone and they have not therefore been technology specific. However, there seems no reason in principle why such obligations could not be broadened to deliver

12

both security of supply and investment in low carbon technologies, albeit at the cost of some complexity in design. 3.4.2.1 Capacity obligations on suppliers There are numerous examples of supplier-based obligations in Europe (often referred to as Public Service Obligations), the US and elsewhere. In the US, capacity obligations are a carryover from the old regional power pool structures in which all participating suppliers, referred to as Load serving Entities or LSE s, were required to acquire sufficient capacity to serve their peak demand plus a reliability margin set by the pool. With the restructuring of the US electricity system in the mid-1990s, many of these arrangements developed into organized capacity markets (see 3.4.2.2); however, some of the original pooling arrangements remain8, with capacity being traded bilaterally to meet obligations in response to demand movements or diversity in demand peaks. Penalties usually apply in the event of an LSE failing to acquire sufficient generation capacity to satisfy the obligation, although there is concern that there may not necessarily be time for the Independent System Operator (ISO) 9, to access capacity in the event of shortages. Capacity obligations can and often do allow demand-side participation however, allowing relief in shorter timescales. Some jurisdictions, i.e. California, have addressed this problem by imposing forward obligations to ensure potential capacity shortages are identified in good time. As with all capacitybased obligations that flow from an administered reliability standard, there are concerns that the value of reliability may not be adequately balanced against the cost of providing that reliability. Concerns have also been raised in the US about a lack of market liquidity and that the prices paid for capacity are not always transparent (Brattle Group, 2009). Addressing the issue of how capacity obligations might be applied in GB, and taking a cue from the Renewable Obligation, suppliers could be required to purchase capacity certificates in proportion to their demand or pay a buyout price, with the proceeds distributed to certificate holders. As the object would be to ensure both security of supply and decarbonisation, the obligation would need to recognise the carbonintensity of different technologies, possibly through premium payments for low carbon technologies or selecting successful bids on the basis of low-carbon emissions as well as bid price (Gottstein, 2010). For example, Southwest & Southwest Power Pool covering most of the Southern & South-western states except California and Texas.
8

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

13

3.4.2.2 Capacity markets Placing obligations on suppliers will naturally lead to capacity trading, as individual parties seek to satisfy specific capacity requirements. However, where a more organised approach is required, or where doubts exist as to the effectiveness of supplier obligations in bringing forward sufficient generation investment, a more reliable option may be to place an obligation on the System Operator. There are a number of examples in Europe, the US and elsewhere of such obligations, which typically involve the System Operator establishing a generation capacity requirement sufficiently far ahead to encourage new investment and procuring the capacity necessary to meet that requirement. PJM, ISO10 New England, ISO New York and ISO Midwest are examples in the US of where trading around the original regional pool-based LSE capacity obligations developed into centralised capacity markets, administered by the ISO. Early market designs delivered mixed results for a number of reasons, including an initial focus on short-term supply reliability at the expense of signalling the need for investment in new capacity and the distorting impact of price caps. This short term focus resulted in bipolar pricing (Gottstein, 2010), with prices collapsing when a surplus of capacity existed but rising to high levels in the event of capacity shortfalls. Market designs developed to overcome these initial problems and now include forward capacity auctions, typically three years before the year of delivery. LSEs retain their supply obligation and can choose to contract bilaterally for capacity. However, where participation in the market is mandated, this capacity must be input to the auction with the ISO effectively procuring any residual capacity required. Both existing and new capacity, and in the case of PJM & ISO New York demand response, bid into the auction and the clearing price is paid to all successful bidders. The costs of procuring the required capacity are allocated to LSEs on a pro-rata basis. There is also a locational element to the auctions to ensure that transmission constraints are respected. Capacity markets operating in the US and elsewhere are essentially non-technology specific in nature, focussing on security of supply alone. However, as is the case with simple capacity obligations, there seems to be no reason why capacity markets could not be designed to take into account the carbon-intensity of generation in order to deliver both security of supply and low carbon objectives It can be argued that the System Operator or ISO is better placed than individual suppliers to anticipate future system demand and optimize the shape of the generation

10

Independent System Operator (ISO).

14

portfolio. Individual suppliers may, for example, be prepared to shed market share rather than commit to new capacity in an uncertain world and might favour one technology over another. An independent System Operator may well take a more holistic view but would be sourcing capacity on the basis of generation adequacy rather than market signals. In the context of the current GB market arrangements, it is difficult to see how investment to satisfy some non-market based adequacy requirement could coexist with investment on a commercial basis. There is a danger therefore, that placing an obligation on the System Operator to procure capacity, even as provider of last resort, could deter normal commercial investment. 3.4.2.3 Capacity payments A limitation of some early capacity-based obligations and markets was a lack of incentives to ensure real time plant availability. Although financial penalties for nondelivery are now common, these penalties do not always reflect the real value of availability during times of system stress. Capacity payment designs are considered more effective in providing real time availability incentives (Oren, 2000). Capacity payment mechanisms normally involve a payment being made to every generator that is available to meet demand for each trading period irrespective of whether or not the generator is actually required to run. Payments, which are normally funded via an uplift on suppliers, are usually linked to the value of capacity, i.e. when the supply position is tight payments will be higher. Capacity payment mechanisms are invariably associated with more integrated scheduling and dispatch arrangements such as the England & Wales Pool, which operated from privatisation of the GB electricity sector in 1990 to the introduction of NETA in 2001. In the case of the England & Wales Electricity Pool, capacity payments were a function of the loss of load probability (LOLP)11 and the value of lost load (VOLL)12. If the supply situation became very tight and LOLP approached unity, then capacity payments could reach very high levels, capped only by the value of VOLL (currently assumed to be around 4000/MWh). The all-Ireland Single Electricity Market introduced in 2007 also incorporates a capacity payment mechanism but utilising a softer link to the fluctuating value of capacity based on the cost of building efficient open cycle gas fired plant.

LOLP is a measure of the likelihood of insufficient generation capacity being available to meet peak demand, varying from zero when there is no risk to unity when there is certainty.
11

VOLL is an estimate of the maximum price a customer is prepared to pay to maintain access to electricity supplies. In practice, VOLL will vary between customer groups and on other circumstances.
12

15

While linking payments to the scarcity value of capacity incentivises additional provision during periods of system stress, perversely it also provides an incentive on portfolio generators to withdraw capacity in order to increase those payments. To some extent this is a criticism that can be levelled at all capacity mechanisms, however the more granular nature of capacity payments compared with capacity obligations provides rather more scope for abuse. Problems associated with the withholding of capacity were a principal driver behind the abandonment of the E&W Electricity Pool and the development of a bilateral market with incentives to contract forward (Patrick, 2001). Other criticisms of capacity payments relate the value of VOLL, which is estimated rather than set by the market, and the usually simplistic methods used for calculating of the value of LOLP. In combination, these issues are almost certain to result in a mismatch between the value of capacity payments and that which would arise from a capacity market. As is the case with capacity obligations and markets, the application of capacity payment mechanisms to date has been linked exclusively to maintaining security of supply. There seems to be no reason however why a capacity payment mechanism could not be designed to recognise the carbon intensity of generation in order to address both decarbonisation and security of supply objectives.

3.4.3 A single buyer


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

16

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

4. Energy dispatch and balancing in a low carbon electricity system


4.1 Problems with existing market arrangements.
Currently, the GB market arrangements make no organised attempt to optimise generation dispatch. Generators and suppliers trade energy in advance on a bilateral basis and, at gate closure13, present the GBSO with generation or demand schedules necessary to deliver contractual commitments a process referred to as self dispatch. Furthermore, as the majority of energy is produced by vertically integrated utilities with both generation and supply businesses, much of this trading is internal - i.e. these utilities self supply to a significant extent. Self supply limits the competitive pressures on which bilateral markets depend to ensure efficiency (Sioshansi, 2009) and the combination of self dispatch and self supply creates the potential for non-optimised dispatch outcomes.

13Gate

closure 1 hour before real time, the energy markets close and the Balancing Mechanism commences. Contractual positions at gate closure are compared with outturn in order to determine imbalance.

17

An additional issue to be considered in the context of dispatch efficiency is the tendency for generators to self insure. The asymmetrical and non cost -reflective cash out prices14 applied to residual imbalances resolved by the GBSO in the Balancing Mechanism, which operates from gate closure, encourages generators to carry reserves in order to minimise imbalance. These reserves are additional to those specified by the GBSO to cover demand or generation uncertainties and result in an energy market that is predominately long, with connected generation capacity exceeding the demand to be supplied. Consequently, more generation is part-loaded than is actually required, reducing overall efficiency and causing unnecessary emissions. An indication of the extent to which generation companies self insure is given by the monthly Trading Operation Report15 published by Elexon, which suggests that there is typically between 1000 and 2000MW of unused reserve available on part-loaded plant over demand peaks and considerably more during other periods. The extent to which the combination of self dispatch, self supply and self insurance reduces the overall efficiency of dispatch is unclear. However, there is evidence from both the US (Sioshansi, 2009) and GB (ILEX, March 2002) to suggest that fuel inputs could be around 3 or 4% higher than would be the case if an organised attempt were made to fully optimise generation dispatch. Further anecdotal evidence that the current GB market arrangements may produce generation dispatch outcomes that differ from the optimum is given by analysis undertaken for Elexon in developing a mechanism to account for transmission losses (Siemens, 2009). This analysis demonstrated that, in some cases, transmission line loss factors calculated from actual line flows differed from those produced using a load flow model that dispatched plant on the basis of marginal cost. In other words, the disposition of generation resulting from existing market arrangements appears to differ to some extent from that which might be delivered by a truly optimised dispatch process based on actual marginal costs. Although the inefficiencies introduced by self dispatch might currently be of a low order, they do result in unnecessary cost and carbon emissions. Dispatch inefficiency is also likely to increase with the growth of intermittent generation. With relatively little wind capacity connected, portfolio generating companies can hide intermittency within their

14Imbalances

that add to the net system imbalance are treated differently than those that reduce net imbalance. For example, a generator whose imbalance adds to system imbalance is exposed to the balancing costs incurred by the GBSO. Generators whose imbalances reduce net imbalance pay/receive prices related to the short term energy prices. This asymmetry encourages parties to self balance and penalizes inflexible or intermittent generators. Operational Trading Reports are available at http://www.elexon.co.uk/search/default.aspx?qs=operational%report
15

18

settlement production account16 relatively easily. However, as wind capacity builds, internalising the impact of intermittency within a generation portfolio will become more difficult. Generators will attempt to trade out intermittency close to real time as wind forecasts become more accurate, however to limit imbalance risk and exposure to cash out prices, generators will need to carry more reserve as the intermittent capacity within their portfolio increases.

4.2 System reserves


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

4.3 Market liquidity


As suggested in 4.1, the growth in wind capacity will significantly increase short term trading as generators attempt to match commitments to updated and more accurate wind output forecasts. This increased trading close to real time will require efficient, liquid short term markets and there are factors which suggest that the current market structures may not be best placed to provide that liquidity or deal with the challenges associated with large amounts of wind generation. The GB market is the least liquid of all comparable European markets (Weber, 2009) due primarily to the vertically For the purposes of settlement, a generating company has a single production account. In the case of a portfolio generator, imbalance prices are applied to the aggregated production account imbalance, rather than the imbalance of each individual generator in that account.
16

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

19

integrated nature of the sector and the high level of internal trading. Liquidity is also reduced by the continuous nature of trading and the existence of alternative trading platforms, which tend to disperse trading activity. This lack of short term market liquidity acts against the interests of both intermittent generation such as wind and also small independent players, who have a greater need for balancing in the shorter term. Prompted by these and more general concerns about market efficiency, Ofgem has consulted on measures to improve market liquidity and has threatened action by the end of 2010 if the situation has not sufficiently improved (Ofgem, 2010).

4.4 A separate market for intermittent generation?


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

Aggregating wind output over a wide geographic area significantly reduces wind output forecast error together with associated reserve and capacity requirements. See for example www.nationalgrid.com/.../GBSQSSIntegratedReliabilityAndEconomicsAssessment.pdf

18

20

4.5 The case for a more integrated approach to market design


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

4.6 Integrated markets and the need for priority dispatch


If the costs of carbon emissions are fully internalised, an integrated dispatch process that attempts to minimise the overall cost of meeting demand securely should also minimise carbon emissions. However, if the cost of carbon remains low, this will not necessarily be the case. While intermittent wind and nuclear generation plant have zero or low marginal costs and will always be dispatched before carbon emitting generation, renewable technologies such as biomass 19 have non-trivial marginal costs and carbon EU Directive 2009/28/EC requires that member states introduce regulations to ensure that renewable generation is given priority in dispatch over other forms of generation. The UK has not introduced regulations to give effect to priority dispatch as, in the GB electricity market, all generation can self dispatch and therefore achieve priority unilaterally.
19

21

sequestrated generation is likely to have higher high marginal costs than nonsequestrated plant due to the associated efficiency penalty. There is a possibility therefore, that an integrated dispatch process may not minimise carbon emissions if carbon is incorrectly priced. In the transition to a low-carbon electricity system, the introduction of a more integrated dispatch process would need to be accompanied by some means of prioritising low carbon generation. Rather than dispatching generation on the basis of cost, low carbon generation would need to be dispatched on the basis of carbon emissions, or some function of marginal cost and carbon emissions, to ensure that overall emissions were minimised. While low carbon capacity remained at modest levels, there would be little need to differentiate between individual generators or technology for the purpose of dispatch. However, as capacity increased, network or energy-related constraints would become more frequent, and some means of differentiation would be required to ensure that carbon emissions were minimised at the lowest possible cost. Differentiation between technologies could be achieved on the basis of emissions, while differentiation within technologies could be achieved when necessary on the basis of marginal cost or some other measure, such as transmission losses. It is interesting to note that there is some experience of non-marginal cost related generation dispatch in the UK, albeit in a rather different context. The Central Electricity Generating Board (CEGB ), which operated a highly detailed centralised dispatch optimisation process, was able to move seamlessly from dispatching on the basis of marginal cost to a heat rate based dispatch during the frequent fuel emergencies of the 19770s & 1980s. Dispatching fossil fired generation on the basis of heat rate rather than marginal cost resulted in a significant reduction in fuel inputs and, as a consequence, would have reduced carbon emissions/MWh of electrical energy generated.

4.7 Market signals v deployment risks for wind.


While a more integrated market design, coupled with priority in dispatch, would create a more benign environment for intermittent technologies such as wind, additional measures may be required given the characteristics of wind generation and the scale of deployment required. The increasingly volatility of energy prices, with wind always likely to be on the wrong side of the balancing argument - attempting to sell energy when wind output is high and energy prices low (or even negative) - will decrease the value that wind can extract from the energy market (Redpoint, 2009) over time. To this erosion of value can be added system integration costs that will also rise steadily as wind capacity builds, further impacting on the viability of future wind projects.

22

While these market signals may simply reflect the economic consequences of intermittency, they could have a negative impact on deployment or at least on the need for subsidy to maintain the level of deployment required to deliver climate change goals. The extent to which wind generation is exposed to market signals varies across Europe. For example GB chooses to draw no distinction on the basis of generation technology, exposing the full costs of balancing and imposing technical requirements that require wind to behave as any other generation even though a system approach may result in lower overall cost. Germany, on the other hand, protects wind generation from the full rigour of market and technical signals with the costs of integration falling mainly on the System Operator. A question to be addressed by the UK and indeed all jurisdictions that intend to connect large amounts of wind or other intermittent renewable generation is, therefore, how to balance the exposure of that generation to market signals with the risks to deployment inherent in those signals particularly given the limited ability of wind to respond.

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

5.1 Congestion volume


In terms of controlling the magnitude of network congestion, the current market arrangements are deficient in two respects. Firstly, market participants can trade energy bilaterally in forward markets without the need to consider the costs that those trades will impose on the electricity grid. The implications of this unconstrained trading are

20

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

Congestion arises when potential power flows exceed the capability of network circuits or boundaries. Congestion is resolved either by adjusting generation patterns to reduce power flows, or by increasing network capacity either by investment in primary assets or by operational means.
21

23

presented to the GBSO in the form of individual generator dispatch schedules one hour before real time, at which point the GBSO is required to establish counter-flows via the Balancing Mechanism (BM) to ensure that actual power flows do not exceed network capacity. Secondly, the cost of resolving this congestion is recovered via Balancing Use of System (BUSoS) charges paid by all trading parties on a per kWh basis and there is therefore no incentive on parties causing that congestion to modify their behaviour. In fact, it can be argued that market arrangements currently encourage behaviour that leads increased network congestion. The separation of energy trading and congestion management into distinct markets prompts generators to consider how best they can maximise returns. Consider for example a portfolio generating company with a large installed wind capacity in Scotland and conventional generation assets on both sides of the cheviot22 network boundary. The company would contract ahead to supply energy assuming, due to its intermittent nature, a modest contribution from wind. If, approaching gate closure, it appeared that wind output would be high, the generator would need to decide what conventional generation to stand down plant on the export side of the boundary or plant on the import side. If the company stands plant down on the export side, potential congestion across the boundary is eased but that plant earns no income. If however, the company stands down conventional plant in E&W, then congestion across the boundary is increased and the GBSO is likely to accept bids from Scottish conventional plant to reduce output. As these bids are invariably less than variable cost of generation, the Scottish plant earns income by not producing or by producing less. Furthermore, the conventional plant in E&W that was stood down is now free to offer replacement energy at a significant premium to market prices. Market rules therefore allow, in fact encourage, companies to maximise income by acting in a fashion which is detrimental to the efficient operation of the system (LECG Consulting, 2010). Ofgem appears to consider that such behaviour amounts to an abuse of market power and a Market Power License Condition (MPLC)
23was

passed into law by

the 2010 Energy Act. The new Condition gives Ofgem the power to penalise the withholding or manipulation of output however, given the short term market volatility that the deployment of wind at scale will bring, deliberate manipulation of output to exploit network constraints will become more difficult to demonstrate. Furthermore, even if demonstrated, such behaviour is arguably no more than the expected commercial response to a particular set of flawed market arrangements. The MPLC therefore addresses the symptoms of the problem, rather than the problem itself.

The Cheviot boundary is that which cuts the four transmission circuits connecting Scotland with England, currently having a capacity of around 2.4 GVA.
22 23

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

24

5.2 Minimising congestion costs


Not only are existing GB market arrangements ineffective in managing congestion volume, they make dealing with congestion particularly expensive. This stems from the energy only nature of the electricity market and the need for mid -merit plant to recoup a proportion of their fixed costs by extracting a discount or premium on forward market prices through Balancing Mechanism bids and offers. The need for mid merit plant to attempt to recover fixed costs through the BM in this fashion is reinforced by the fact that peaking plant is able to partially recover investment and other fixed costs through pre-gate closure energy contracts. By utilising contracted plant over demand peaks, the GBSO is able to reduce spikes in energy prices, therefore reducing the income available to non-contracted mid merit plant (SEDG, 2009). The impact of fixed cost recovery through the BM can be seen in figure 3, which illustrates the relationship between accepted offers and bids to the market index price (MIP)24. It can be seen that accepted BM offers are invariable at a significant premium to MIP, while accepted bids are invariably discounted. As the cost of resolving congestion is the difference between the associated bids and offers, these costs can on occasion exceed 150/MWh.

250

200

150

/MWh
100 50 0 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.

25

It is instructive to compare the costs of resolving congestion under current market rules with those observed under previous market regimes, for example the England & Wales Electricity Pool, which precede the introduction of NETA/BETTA. Under the old Pool rules, the cost of resolving congestion was essentially the difference between the offers made by generation that was ultimately constrained and offers made by replacement plant at the day-ahead schedule stage. For similar technologies, i.e. where coal plant was displaced by other coal plant in order to resolve a network constraint, the difference in day ahead offers may only have been a few pounds and maybe around 15/MWh where CCGT plant was replaced by coal. Resolving network congestion under BETTA is therefore around ten times as expensive as was the case under the E&W Electricity Pool rules or the old CEGB merit order arrangements which existed before the industry was privatised.

5.3 Transmission Investment signals


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

6. Encouraging demand-side participation


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

26

6.1 Reducing capacity and reserve


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

6.2 Settlement impacts


The delivery of domestic and small commercial sector demand response will have implications for the electricity market settlement process. As demand less than 100kW is currently metered on a summation basis, it is input to the electricity market via a profiling process, where customers are allocated to one of eight demand profiles for the purposes of settlement. Rather than being charged for the actual half-hourly consumption of their smaller customers, suppliers are charged on the deemed consumption given by these profiles.

27

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

28

References
Abbad, J. R. (2010). Electricity Market Participation of wind Farms:the success story of

the Spannish pragmitism. Energy Policy.


Baker, Mitchell & Woodman. (2009). The Extent to Which Economic Regulation Enables

the Transition to a Sustainable Electricity System.


Brattle Group. (2009). A Comparison of PJM's RPM with Alternative Energy and Capacity

Market Designs.
Committee on Climate Change. (2010). Meeting Carbon Budgets - ensuring a low

carbon recovery.
E.on. (2009). http://www.eon.com/en/downloads/E.ON_Capital_Market_Day_09.pdf. ENA. (2010). ENA Response to Ofgems Embedding Financeability in a N ew Regulation"

Consultation.
Ernst & Young. (2009). Securing the UK's Energy Future. Geen Investment Bank Commission. (June, 2010). Unlocking Investment to Deliver

Britain's low carbon future.


Gottstein, S. (2010). the Rol of Forward Capacity Markets in Increasing Demand-Side and

Other Low Carbon Resources: Experiences and Prospects.


HM Treasury. (2010). Energy Market Assessment. House of Commons Business and Enterprise Select Committee. (2009). Energy Policy -

Future challenges.
IEA. (2008). Design of Power Systems with Large Amounts of Wind Power. Wind Task 25

Phase 1 Final Report.


ILEX. (March 2002). NETA- the Next Phase. Investco Perpetual. (2010). Open Letter to Lord Mogg from Neil Woodford. Kleindorfer, C. &. The Economics of ublic tility egulation. 1986: Macmillan Press. LECG Consulting. (2010). The UK Transmission Congestion Problem.

41

Littlechild. (2007). A Proposal for a Balancing Market to Determine Cash Out Prices. Natioanl Grid. (2009). Grid Code Working Group - Intermittent Generation Data. National Grid. (2010). Letter from Paul Whittaker to Scott Phillips, Ofgem; RPI-X@20

Current Thinking Working Paper - Financeability.


National Grid. (2009). Operating the Electricity Transmission Networks in 2020. Ofgem. (2010). Liquidity Proposals for the GB wholesale Electricity Market. Ref 22/10. Ofgem. (2010). Project Discovery. Ofgem. (2009). Project Discovery: Energy Market Scenarios consultation, apendix 2. Ofgem. (2010). Regulation energy networks for the future: RPI-X@20 Recomendations:

Ref 91/10.
Ofgem. (2008). Transmission Access Review - Initial Consultation on Enhanced

Investment Incentives. Document 175/08.


Poyry. (2009). Impact of Intermittency - How Wind Variability could Change the Nature

of the British & Irish Electricity Markets.


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

Areport to the EWEA.


Redpoint. (2009). Decarbonising the GB Power Sector, evaluating investment pathways,

generation patterns and emissions through to 2030. A report to the Committee on Climate Change.
Redpoint Energy. (2006). Dynamics of GB Electricity Generation Investment. Redpoint. (2010). Improving Grid Access: Modeling the Impacts of the consultation

Options.
SEDG. (2009). Probabalistic Network Operation and Design standards to Support the

Development of the UK Low carbon Electricity System.


Siemens. (2009). MP229 Load Flow Modelling Service. Report to Elexon. Sioshansi, O. O. (2009). The cost of Anarchy in Self-commitment Based Electricity

Markets.

42

Strbac. (2008). Transmission Systems with Wind & Nuclear. Weber. (2009). Adequate Intraday market desig to Allow the integration of wind energy

into the European power systems. Elssvier.

43

Chapter 3

Direct Solar Energy


Coordinating Lead Authors:
Dan Arvizu (USA) and Palani Balaya (Singapore/India)

Lead Authors:
Luisa F. Cabeza (Spain), K.G. Terry Hollands (Canada), Arnulf Jger-Waldau (Italy/Germany), Michio Kondo (Japan), Charles Konseibo (Burkina Faso), Valentin Meleshko (Russia), Wesley Stein (Australia), Yutaka Tamaura (Japan), Honghua Xu (China), Roberto Zilles (Brazil)

Contributing Authors:
Armin Aberle (Singapore/Germany), Andreas Athienitis (Canada), Shannon Cowlin (USA), Don Gwinner (USA), Garvin Heath (USA), Thomas Huld (Italy/Denmark), Ted James (USA), Lawrence Kazmerski (USA), Margaret Mann (USA), Koji Matsubara (Japan), Anton Meier (Switzerland), Arun Mujumdar (Singapore), Takashi Oozeki (Japan), Oumar Sanogo (Burkina Faso), Matheos Santamouris (Greece), Michael Sterner (Germany), Paul Weyers (Netherlands)

Review Editors:
Eduardo Calvo (Peru) and Jrgen Schmid (Germany)

This chapter should be cited as: Arvizu, D., P. Balaya, L. Cabeza, T. Hollands, A. Jger-Waldau, M. Kondo, C. Konseibo, V. Meleshko, W. Stein, Y. Tamaura, H. Xu, R. Zilles, 2011: Direct Solar Energy. In IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation [O. Edenhofer, R. Pichs-Madruga, Y. Sokona, K. Seyboth, P. Matschoss, S. Kadner, T. Zwickel, P. Eickemeier, G. Hansen, S. Schlmer, C. von Stechow (eds)], Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

333

Chapter 3

Direct Solar Energy

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

3.2.4

Possible impact of climate change on resource potential

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

3.3

Technology and applications

This section discusses technical issues for a range of solar technologies, organized under the following categories: passive solar and daylighting,

Table 3.2 | International and national projects that collect, process and archive information on solar irradiance resources at the Earths surface. Available Data Sets
Ground-based solar irradiance from 1,280 sites for 1964 to 2009 provided by national meteorological services around the world. National Solar Radiation Database that includes 1,454 ground locations for 1991 to 2005. The satellite-modelled solar data for 1998 to 2005 provided on 10-km grid. The hourly values of solar data can be used to determine solar resources for collectors. European Solar Radiation Database that includes measured solar radiation complemented with other meteorological data necessary for solar engineering. Satellite images from METEOSAT help in improving accuracy in spatial interpolation. Test Reference Years were also included. The Solar Radiation Atlas of Africa contains information on surface radiation over Europe, Asia Minor and Africa. Data covering 1985 to 1986 were derived from measurements by METEOSAT 2. The solar data set for Africa based on images from METEOSAT processed with the Heliosat-2 method covers the period 1985 to 2004 and is supplemented with ground-based solar irradiance. Typical Meteorological Year (Test Reference Year) data sets of hourly values of solar radiation and meteorological parameters derived from individual weather observations in long-term (up to 30 years) data sets to establish a typical year of hourly data. Used by designers of heating and cooling systems and large-scale solar thermal power plants. The solar radiation data for solar energy applications. IEA/SHC Task36 provides a wide range of users with information on solar radiation resources at Earths surface in easily accessible formats with understandable quality metrics. The task focuses on development, validation and access to solar resource information derived from surface- and satellite-based platforms. Solar and Wind Energy Resource Assessment (SWERA) project aimed at developing information tools to simulate RE development. SWERA provides easy access to high-quality RE resource information and data for users. Covered major areas of 13 developing countries in Latin America, the Caribbean, Africa and Asia. SWERA produced a range of solar data sets and maps at better spatial scales of resolution than previously available using satellite- and ground-based observations.

Responsible Institution/Agency
World Radiation Data Centre, Saint Petersburg, Russian Federation (wrdc.mgo.rssi.ru) National Renewable Energy Laboratory, USA (www.nrel.gov)

Supported by Commission of the European Communities, National Weather Services and scientic institutions of the European countries Supported by the Commission of the European Communities

Ecole des Mines de Paris, France

National Renewable Energy Laboratory, USA. National Climatic Data Center, National Oceanic and Atmospheric Administration, USA. (www.ncdc.noaa.gov) International Energy Agency (IEA) Solar Heating and Cooling Programme (SHC). (swera.unep.net)

Global Environment Facility-sponsored project. United Nations Environment Programme (swera.unep.net)

343

Direct Solar Energy

Chapter 3

active heating and cooling, PV electricity generation, CSP electricity generation and solar fuel production. Each section also describes applications of these technologies.

3.3.1

Passive solar and daylighting technologies

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

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

344

Chapter 3

Direct Solar Energy

BIPV/T Roof Variable Speed Fan Dryer Air Inlet Geothermal Pump Passive Slab

Exhaust Fan External Rolling Shutter

Solar

Tilted Slats

HRV Light Shelf DHW Internal Rolling Shutter

Ventilated Slab

Blinds

Side-Fin

Figure 3.2 | Left: Schematic of thermal mass placement and passive-active systems in a house; solar-heated air from building-integrated photovoltaic/thermal (BIPV/T) roof heats ventilated slab or domestic hot water (DHW) through heat exchanger; HRV is heat recovery ventilator. Right: Schematic of several daylighting concepts designed to redistribute daylight into the ofce interior space (Athienitis, 2008).

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

of existing homes as well. Many homes also add a solarium during retrot. The new glazing technologies and solar control systems allow the design of a larger window area than in the recent past. In most climates, unless effective solar gain control is employed, there may be a need to cool the space during the summer. However, the need for mechanical cooling may often be eliminated by designing for passive cooling. Passive cooling techniques are based on the use of heat and solar protection techniques, heat storage in thermal mass and heat dissipation techniques. The specic contribution of passive solar and energy conservation techniques depends strongly on the climate (UNEP, 2007). Solar-gain control is particularly important during the shoulder seasons when some heating may be required. In adopting larger window areasenabled by their high thermal resistanceactive solargain control becomes important in solar buildings for both thermal and visual considerations. The potential of passive solar cooling in reducing CO2 emissions has been shown recently (Cabeza et al., 2010; Castell et al., 2010). Experimental work demonstrates that adequate insulation can reduce by up to 50% the cooling energy demand of a building during the hot season. Moreover, including phase-change materials in the alreadyinsulated building envelope can reduce the cooling energy demand in such buildings further by up to 15%about 1 to 1.5 kg/yr/m2 of CO2 emissions would be saved in these buildings due to reducing the energy

345

Direct Solar Energy

Chapter 3

consumption compared to the insulated building without phase-change material. Passive solar system applications are mainly of the direct-gain type, but they can be further subdivided into the following main application categories: multi-story residential buildings and two-story detached or semi-detached solar homes (see Figure 3.2, left panel), designed to have a large equatorial-facing faade to provide the potential for a large solar capture area (Athienitis, 2008). Perimeter zones and their fenestration systems in ofce buildings are designed primarily based on daylighting performance. In this application, the emphasis is usually on reducing cooling loads, but passive heat gains may be desirable as well during the heating season (see Figure 3.2, right panel, for a schematic of shading devices). In addition, residential or commercial buildings may be designed to use natural or hybrid ventilation systems and techniques for cooling or fresh air supply, in conjunction with designs for using daylight throughout the year and direct solar gains during the heating season. These buildings may prot from low summer night temperatures by using night hybrid ventilation techniques that utilize both mechanical and natural ventilation processes (Santamouris and Asimakopoulos, 1996; Voss et al., 2007). In 2010, passive technologies played a prominent role in the design of net-zero-energy solar homeshomes that produce as much electrical and thermal energy as they consume in an average year. These houses are primarily demonstration projects in several countries currently collaborating in the International Energy Agency (IEA) Task 40 of the Solar Heating and Cooling (SHC) Programme (IEA, 2009b)Energy Conservation in Buildings and Community Systems Annex 52which focuses on net-zero-energy solar buildings. Passive technologies are essential in developing affordable net-zero-energy homes. Passive solar gains in homes based on the Passive House Standard are expected to reduce the heating load by about 40%. By extension, systematic passive solar design of highly insulated buildings at a community scale, with optimal orientation and form of housing, should easily result in a similar energy saving of 40%. In Europe, according to the Energy Performance of Buildings Directive recast, Directive 2010/31/EC (The European Parliament and the Council of the European Union, 2010), all new buildings must be nearly zero-energy buildings by 31 December 2020, while EU member states should set intermediate targets for 2015. New buildings occupied and owned by public authorities have to be nearly zero-energy buildings after 31 December 2018. The nearly zero or very low amount of energy required should to a very signicant level be covered by RE sources, including onsite energy production using combined heat and power generation or district heating and cooling, to satisfy most of their demand. Measures should also be taken to stimulate building refurbishments into nearly zero-energy buildings. Low-energy buildings are known under different names. A survey carried out by Concerted Action Energy Performance of Buildings (EPBD) identied 17 different terms to describe such buildings across Europe,

including: low-energy house, high-performance house, passive house (Passivhaus), zero-carbon house, zero-energy house, energy-savings house, energy-positive house and 3-litre house. Concepts that take into account more parameters than energy demand again use special terms such as eco-building or green building. Another IEA AnnexEnergy Conservation through Energy Storage Implementing Agreement (ECES IA) Annex 23was initiated in November 2009 (IEA ECES, 2004). The general objective of the Annex is to ensure that energy storage techniques are properly applied in ultralow-energy buildings and communities. The proper application of energy storage is expected to increase the likelihood of sustainable building technologies. Another passive solar application is natural drying. Grains and many other agricultural products have to be dried before being stored so that insects and fungi do not render them unusable. Examples include wheat, rice, coffee, copra (coconut esh), certain fruits and timber (Twidell and Weir, 2006). Solar energy dryers vary mainly as to the use of the solar heat and the arrangement of their major components. Solar dryers constructed from wood, metal and glass sheets have been evaluated extensively and used quite widely to dry a full range of tropical crops (Imre, 2007).

3.3.2

Active solar heating and cooling

Active solar heating and cooling technologies use the Sun and mechanical elements to provide either heating or cooling; various technologies are discussed here, as well as thermal storage.

3.3.2.1

Solar heating

In a solar heating system, the solar collector transforms solar irradiance into heat and uses a carrier uid (e.g., water, air) to transfer that heat to a well-insulated storage tank, where it can be used when needed. The two most important factors in choosing the correct type of collector are the following: 1) the service to be provided by the solar collector, and 2) the related desired range of temperature of the heat-carrier uid. An uncovered absorber, also known as an unglazed collector, is likely to be limited to low-temperature heat production (Dufe and Beckman, 2006). A solar collector can incorporate many different materials and be manufactured using a variety of techniques. Its design is inuenced by the system in which it will operate and by the climatic conditions of the installation location. Flat-plate collectors are the most widely used solar thermal collectors for residential solar water- and space-heating systems. They are also used in air-heating systems. A typical at-plate collector consists of an absorber, a header and riser tube arrangement or a single serpentine

346

Chapter 3

Direct Solar Energy

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

Unglazed Solar Collectors


Tube-on-Sheet Collector

Flat Plate Collectors

Single or Double Glazing Glazing Frame

Pump Flow

Metal Deck Rubber Grommet Box Flow Passages

Serpentine Plastic Pipe Collector

Absorber Plate

Insulation Backing

Figure 3.3a | Schematic diagram of thermal solar collectors: Glazed at-plate.

Typically 1 1/2 ABS Pipe

Evacuated-tube collectors are usually made of parallel rows of transparent glass tubes, in which the absorbers are enclosed, connected to a header pipe (Figure 3.3c). To reduce heat loss within the frame by convection, the air is pumped out of the collector tubes to generate a vacuum. This makes it possible to achieve high temperatures, useful

Pump Flow

Figure 3.3b | Schematic diagram of thermal solar collectors: Unglazed tube-on-sheet and serpentine plastic pipe.

Evacuated-Tube Collectors
Heat ater ed W

Insulation

Solar

Energ

y Heat fe Trans r r

Heat r Ris Vapo ensed Liq es to Top ns to Botto m

fe Trans

Copper Sleeve in Manifold Aluminium Header Casing

etur uid R

Cond

Copper Manifold

Vacuum Indicator

Evacuated Glass Tube

Evacuated Heat Pipe

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

347

Direct Solar Energy

Chapter 3

for cooling (see below) or industrial applications. Most vacuum tube collectors use heat pipes for their core instead of passing liquid directly through them. Evacuated heat-pipe tubes are composed of multiple evacuated glass tubes, each containing an absorber plate fused to a heat pipe. The heat from the hot end of the heat pipes is transferred to the transfer uid of a domestic hot water or hydronic space-heating system. Solar water-heating systems used to produce hot water can be classied as passive or active solar water heaters (Dufe and Beckman, 2006). Also of interest are active solar cooling systems, which transform the hot water produced by solar energy into cold water. Passive solar water heaters are of two types (Figure 3.4). Integral collector-storage (ICS) or batch systems include black tanks or tubes in an insulated glazed box. Cold water is preheated as it passes through the solar collector, with the heated water owing to a standard backup water heater. The heated water is stored inside the collector itself. In thermosyphon (TS) systems, a separate storage tank is directly above the collector. In direct (open-loop) TS systems, the heated water rises from the collector to the tank and cool water from the tank sinks back into the collector. In indirect (closed-loop) TS systems (Figure 3.4, left), heated uid (usually a glycol-water mixture) rises from the collector to an outer tank that surrounds the water storage tank and acts as a heat exchanger (double-wall heat exchangers) for separation from potable water. In climates where freezing temperatures are unlikely, many collectors include an integrated storage tank at the top of the collector. This design has many cost and user-friendly advantages compared to a system that uses a separate standalone heat-exchanger tank. It is also appropriate in

households with signicant daytime and evening hot water needs; but it does not work well in households with predominantly morning draws because sometimes the tanks can lose most of the collected energy overnight. Active solar water heaters rely on electric pumps and controllers to circulate the carrier uid through the collectors. Three types of active solar water-heating systems are available. Direct circulation systems use pumps to circulate pressurized potable water directly through the collectors. These systems are appropriate in areas that do not freeze for long periods and do not have hard or acidic water. Antifreeze indirect-circulation systems pump heat-transfer uid, which is usually a glycol-water mixture, through collectors. Heat exchangers transfer the heat from the uid to the water for use (Figure 3.4, right). Drainback indirect-circulation systems use pumps to circulate water through the collectors. The water in the collector and the piping system drains into a reservoir tank when the pumps stop, eliminating the risk of freezing in cold climates. This system should be carefully designed and installed to ensure that the piping always slopes downward to the reservoir tank. Also, stratication should be carefully considered in the design of the water tank (Hadorn, 2005). A solar combisystem provides both solar space heating and cooling as well as hot water from a common array of solar thermal collectors, usually backed up by an auxiliary non-solar heat source (Weiss, 2003). Solar combisystems may range in size from those installed in individual properties to those serving several in a block heating scheme. A large number of different types of solar combisystems are produced. The systems on the market in a particular country may be more restricted, however, because different systems have tended to evolve in different countries.

A Close-Coupled Solar Water Heater Solar Energy

Solar Collector

Solar Energy

To Taps Tank Controller

Boiler

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

Cold Water Feed Pump

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

348

Chapter 3

Direct Solar Energy

Depending on the size of the combisystem installed, the annual space heating contribution can range from 10 to 60% or more in ultra-low energy Passivhaus-type buildings, and even up to 100% where a large seasonal thermal store or concentrating solar thermal heat is used.

include the heat recovery units, heat exchangers and humidiers. Liquid sorption techniques have been demonstrated successfully.

3.3.2.3 3.3.2.2 Solar cooling

Thermal storage

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

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

349

Direct Solar Energy

Chapter 3

thermal energy storage. This technology uses a natural underground layer (e.g., sand, sandstone or chalk) as a storage medium for the temporary storage of heat or cold. The transfer of thermal energy is realized by extracting groundwater from the layer and by re-injecting it at the modied temperature level at a separate location nearby. Most applications are for the storage of winter cold to be used for the cooling of large ofce buildings and industrial processes. Aquifer cold storage is gaining interest because savings on electricity bills for chillers are about 75%, and in many cases, the payback time for additional investments is shorter than ve years. A major condition for the application of this technology is the availability of a suitable geologic formation.

medium-temperature heat and are often necessary in areas with high solar irradiance and high energy costs. Some process heat applications can be met with temperatures delivered by ordinary low-temperature collectors, namely, from 30C to 80C. However, the bulk of the demand for industrial process heat requires temperatures from 80C to 250C. Process heat collectors are another potential application for solar thermal heat collectors. Typically, these systems require a large capacity (hence, large collector areas), low costs, and high reliability and quality. Although low- and high-temperature collectors are offered in a dynamically growing market, process heat collectors are at a very early stage of development and no products are available on an industrial scale. In addition to concentrating collectors, improved at collectors with double and triple glazing are currently being developed, which could meet needs for process heat in the range of up to 120C. Concentrating-type solar collectors are described in Section 3.3.4. Solar refrigeration is used, for example, to cool stored vaccines. The need for such systems is greatest in peripheral health centres in rural communities in the developing world, where no electrical grid is available. Solar cooling is a specic area of application for solar thermal technology. High-efciency at plates, evacuated tubes or parabolic troughs can be used to drive absorption cycles to provide cooling. For a greater coefcient of performance (COP), collectors with low concentration levels can provide the temperatures (up to around 250C) needed for double-effect absorption cycles. There is a natural match between solar energy and the need for cooling. A number of closed heat-driven cooling systems have been built, using solar thermal energy as the main source of heat. These systems often have large cooling capacities of up to several hundred kW. Since the early 2000s, a number of systems have been developed in the small-capacity range, below 100 kW, and, in particular, below 20 kW and down to 4.5 kW. These small systems are single-effect machines of different types, used mainly for residential buildings and small commercial applications. Although open-cooling cycles are generally used for air conditioning in buildings, closed heat-driven cooling cycles can be used for both air conditioning and industrial refrigeration. Other solar applications are listed below. The production of potable water using solar energy has been readily adopted in remote or isolated regions (Narayan et al., 2010). Solar stills are widely used in some parts of the world (e.g., Puerto Rico) to supply water to households of up to 10 people (Khanna et al., 2008). In appropriate isolation conditions, solar detoxication can be an effective low-cost

3.3.2.4

Active solar heating and cooling applications

For active solar heating and cooling applications, the amount of hot water produced depends on the type and size of the system, amount of sun available at the site, seasonal hot-water demand pattern, and installation characteristics of the system (Norton, 2001). Solar heating for industrial processes is at a very early stage of development in 2010 (POSHIP, 2001). Worldwide, less than 100 operating solar thermal systems for process heat are reported, with a total capacity of about 24 MWth (34,000 m collector area). Most systems are at an experimental stage and relatively small scale. However, signicant potential exists for market and technological developments, because 28% of the overall energy demand in the EU27 countries originates in the industrial sector, and much of this demand is for heat below 250C. Education and knowledge dissemination are needed to deploy this technology. In the short term, solar heating for industrial processes will mainly be used for low-temperature processes, ranging from 20C to 100C. With technological development, an increasing number of medium-temperature applicationsup to 250Cwill become feasible within the market. According to Werner (2006), about 30% of the total industrial heat demand is required at temperatures below 100C, which could theoretically be met with solar heating using current technologies. About 57% of this demand is required at temperatures below 400C, which could largely be supplied by solar in the foreseeable future. In several specic industry sectorssuch as food, wine and beverages, transport equipment, machinery, textiles, and pulp and paperthe share of heat demand at low and medium temperatures (below 250C) is around 60% (POSHIP, 2001). Tapping into this low- and mediumtemperature heat demand with solar heat could provide a signicant opportunity for solar contribution to industrial energy requirements. A substantial opportunity for solar thermal systems also exists in chemical industries and in washing processes. Among the industrial processes, desalination and water treatment (e.g., sterilization) are particularly promising applications for solar thermal energy, because these processes require large amounts of

350

Chapter 3

Direct Solar Energy

treatment for low-contaminant waste (Gumy et al., 2006). Multipleeffect humidication (MEH) desalination units indirectly use heat from highly efcient solar thermal collectors to induce evaporation and condensation inside a thermally isolated, steam-tight container. These MEH systems are now beginning to appear in the market. Also see the report on water desalination by CSP (DLR, 2007) and the discussion of SolarPACES Task VI (SolarPACES, 2009b). In solar drying, solar energy is used either as the sole source of the required heat or as a supplemental source, and the air ow can be generated by either forced or free (natural) convection (Fudholi et al., 2010). Solar cooking is one of the most widely used solar applications in developing countries (Lahkar and Samdarshi, 2010) though might still be considered an early stage commercial product due to limited overall deployment in comparison to other cooking methods. A solar cooker uses sunlight as its energy source, so no fuel is needed and operating costs are zero. Also, a reliable solar cooker can be constructed easily and quickly from common materials.

Anti-Reection Coating n-Type Semiconductor

Front Contact

Electron (-) Recombination Hole (+)

p-Type Semiconductor Back Contact

Figure 3.5 | Generic schematic cross-section illustrating the operation of an illuminated solar cell.

3.3.3

Photovoltaic electricity generation


multicrystalline silicon wafer PV (including ribbon technologies) are the dominant technologies on the PV market, with a 2009 market share of about 80%; thin-lm PV (primarily CdTe and thin-lm Si) has the remaining 20% share. Organic PV (OPV) consists of organic absorber materials and is an emerging class of solar cells. Wafer-based silicon technology includes solar cells made of monocrystalline or multicrystalline wafers with a current thickness of around 200 m, while the thickness is decreasing down to 150 m. Single-junction wafer-based c-Si cells have been independently veried to have record energy conversion efciencies of 25.0% for monocrystalline silicon cells and 20.3% for multicrystalline cells (Green et al., 2010b) under standard test conditions (i.e., irradiance of 1,000 W/m2, air-mass 1.5, 25C). The theoretical Shockley-Queisser limit of a single-junction cell with an energy bandgap of crystalline silicon is 31% energy conversion efciency (Shockley and Queisser, 1961). Several variations of wafer-based c-Si PV for higher efciency have been developed, for example, heterojunction solar cells and interdigitated back-contact (IBC) solar cells. Heterojunction solar cells consist of a crystalline silicon wafer base sandwiched by very thin (~5 nm) amorphous silicon layers for passivation and emitter. The highest-efciency heterojunction solar cell is 23.0% for a 100.4-cm2 cell (Taguchi et al., 2009). Another advantage is a lower temperature coefcient. The efciency of conventional c-Si solar cells declines with elevating ambient temperature at a rate of -0.45%/C, while the heterojunction cells show a lower rate of -0.25%/C (Taguchi et al., 2009). An IBC solar cell, where both the base and emitter are contacted at the back of the cell, has the advantage of no shading of the front of the cell by a top electrode. The highest efciency of such a back-contact silicon wafer

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

3.3.3.1

Existing photovoltaic technologies

Existing PV technologies include wafer-based crystalline silicon (c-Si) cells, as well as thin-lm cells based on copper indium/gallium disulde/diselenide (CuInGaSe2; CIGS), cadmium telluride (CdTe), and thin-lm silicon (amorphous and microcrystalline silicon). Mono- and

351

Direct Solar Energy

Chapter 3

cell is 24.2% for 155.1 cm2 (Bunea et al., 2010). Commercial module efciencies for wafer-based silicon PV range from 12 to 14% for multicrystalline Si and from 14 to 20% for monocrystalline Si. Commercial thin-lm PV technologies include a range of absorber material systems: amorphous silicon (a-Si), amorphous silicon-germanium, microcrystalline silicon, CdTe and CIGS. These thin-lm cells have an absorber layer thickness of a few m or less and are deposited on glass, metal or plastic substrates with areas of up to 5.7 m2 (Stein et al., 2009). The a-Si solar cell, introduced in 1976 (Carlson and Wronski, 1976) with initial efciencies of 1 to 2%, has been the rst commercially successful thin-lm PV technology. Because a-Si has a higher light absorption coefcient than c-Si, the thickness of an a-Si cell can be less than 1 mthat is, more than 100 times thinner than a c-Si cell. Developing higher efciencies for a-Si cells has been limited by inherent material quality and by light-induced degradation identied as the Staebler-Wronski effect (Staebler and Wronski, 1977). However, research efforts have successfully lowered the impact of the Staebler-Wronski effect to around 10% or less by controlling the microstructure of the lm. The highest stabilized efciencythe efciency after the light-induced degradationis reported as 10.1% (Benagli et al., 2009). Higher efciency has been achieved by using multijunction technologies with alloy materials, e.g., germanium and carbon or with microcrystalline silicon, to form semiconductors with lower or higher bandgaps, respectively, to cover a wider range of the solar spectrum (Yang and Guha, 1992; Yamamoto et al., 1994; Meier et al., 1997). Stabilized efciencies of 12 to 13% have been measured for various laboratory devices (Green et al., 2010b). CdTe solar cells using a heterojunction with cadmium sulphide (CdS) have a suitable energy bandgap of 1.45 electron-volt (eV) (0.232 aJ) with a high coefcient of light absorption. The best efciency of this cell is 16.7% (Green et al., 2010b) and the best commercially available modules have an efciency of about 10 to 11%. The toxicity of metallic cadmium and the relative scarcity of tellurium are issues commonly associated with this technology. Although several assessments of the risk (Fthenakis and Kim, 2009; Zayed and Philippe, 2009) and scarcity (Green et al., 2009; Wadia et al., 2009) are available, no consensus exists on these issues. It has been reported that this potential hazard can be mitigated by using a glass-sandwiched module design and by recycling the entire module and any industrial waste (Sinha et al., 2008). The CIGS material family is the basis of the highest-efciency thin-lm solar cells to date. The copper indium diselenide (CuInSe2)/CdS solar cell was invented in the early 1970s at AT&T Bell Labs (Wagner et al., 1974). Incorporating Ga and/or S to produce CuInGa(Se,S)2 results in the benet of a widened bandgap depending on the composition (Dimmler and Schock, 1996). CIGS-based solar cells have been validated at an

efciency of 20.1% (Green et al., 2010b). Due to higher efciencies and lower manufacturing energy consumptions, CIGS cells are currently in the industrialization phase, with best commercial module efciencies of up to 13.1% (Kushiya, 2009) for CuInGaSe2 and 8.6% for CuInS2 (Meeder et al., 2007). Although it is acknowledged that the scarcity of In might be an issue, Wadia et al. (2009) found that the current known economic indium reserves would allow the installation of more than 10 TW of CIGS-based PV systems. High-efciency solar cells based on a multijunction technology using III-V semiconductors (i.e., based on elements from the III and V columns of the periodic chart), for example, gallium arsenide (GaAs) and gallium indium phosphide (GaInP) , can have superior efciencies. These cells were originally developed for space use and are already commercialized. An economically feasible terrestrial application is the use of these cells in concentrating PV (CPV) systems, where concentrating optics are used to focus sunlight onto high efciency solar cells (Bosi and Pelosi, 2007). The most commonly used cell is a triple-junction device based on GaInP/GaAs/germanium (Ge), with a record efciency of 41.6% for a lattice-matched cell (Green et al., 2010b) and 41.1% for a metamorphic or lattice-mismatched device (Bett et al., 2009). Sub-module efciencies have reached 36.1% (Green et al., 2010b). Another advantage of the concentrator system is that cell efciencies increase under higher irradiance (Bosi and Pelosi, 2007), and the cell area can be decreased in proportion to the concentration level. Concentrator applications, however, require direct-normal irradiation, and are thus suited for specic climate conditions with low cloud coverage.

3.3.3.2

Emerging photovoltaic technologies

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

352

Chapter 3

Direct Solar Energy

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

Committee, 2001; Navigant Consulting Inc., 2006; EU PV European Photovoltaic Technology Platform, 2007; Kroposki et al., 2008; NEDO, 2009). At the component level, BOS components for grid-connected applications are not yet sufciently developed to match the lifetime of PV modules. Additionally, BOS component and installation costs need to be reduced. Moreover, devices for storing large amounts of electricity (over 1 MWh or 3,600 MJ) will be adapted to large PV systems in the new energy network. As new module technologies emerge in the future, some of the ideas relating to BOS may need to be revised. Furthermore, the quality of the system needs to be assured and adequately maintained according to dened standards, guidelines and procedures. To ensure system quality, assessing performance is important, including on-line analysis (e.g., early fault detection) and off-line analysis of PV systems. The knowledge gathered can help to validate software for predicting the energy yield of future module and system technology designs. To increasingly penetrate the energy network, PV systems must use technology that is compatible with the electric grid and energy supply and demand. System designs and operation technologies must also be developed in response to demand patterns by developing technology to forecast the power generation volume and to optimize the storage function. Moreover, inverters must improve the quality of grid electricity by controlling reactive power or ltering harmonics with communication in a new energy network that uses a mixture of inexpensive and effective communications systems and technologies, as well as smart meters (see Section 8.2.1).

3.3.3.5 3.3.3.3 Novel photovoltaic technologies

Photovoltaic applications

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

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

3.3.3.4

Photovoltaic systems

A photovoltaic system is composed of the PV module, as well as the balance of system (BOS) components, which include an inverter, storage devices, charge controller, system structure, and the energy network. The system must be reliable, cost effective, attractive and match with the electric grid in the future (US Photovoltaic Industry Roadmap Steering

353

Direct Solar Energy

Chapter 3

Installed PV Power [MWp]

25,000

Cumulative Grid-Connected 20,000 Cumulative Off-Grid

an existing structure; and the PV array itself can be used as a cladding or roong material, as in building-integrated PV (Eiffert, 2002; Ecofys Netherlands BV, 2007; Elzinga, 2008). An often-cited disadvantage is the greater sensitivity to grid interconnection issues, such as overvoltage and unintended islanding (Kobayashi and Takasaki, 2006; Cobben et al., 2008; Ropp et al., 2008). However, much progress has been made to mitigate these effects, and today, by Institute of Electrical and Electronics Engineers (IEEE) and Underwriter Laboratories standards (IEEE 1547 (2008), UL 1741), all inverters must have the function of the anti-islanding effect. Grid-connected centralized PV systems perform the functions of centralized power stations. The power supplied by such a system is not associated with a particular electricity customer, and the system is not located to specically perform functions on the electricity network other than the supply of bulk power. Typically, centralized systems are mounted on the ground, and they are larger than 1 MW. The economical advantage of these systems is the optimization of installation and operating cost by bulk buying and the cost effectiveness of the PV components and balance of systems at a large scale. In addition, the reliability of centralized PV systems can be greater than distributed PV systems because they can have maintenance systems with monitoring equipment, which can be a smaller part of the total system cost. Multi-functional PV, daylighting and solar thermal components involving PV or solar thermal that have already been introduced into the built environment include the following: shading systems made from PV and/or solar thermal collectors; hybrid PV/thermal (PV/T) systems that generate electricity and heat from the same panel/collector area; semitransparent PV windows that generate electricity and transmit daylight from the same surface; faade collectors; PV roofs; thermal energy roof systems; and solar thermal roof-ridge collectors. Currently, fundamental and applied R&D activities are also underway related to developing other products, such as transparent solar thermal window collectors, as well as faade elements that consist of vacuum-insulation panels, PV panels, heat pump, and a heat-recovery system connected to localized ventilation. Solar energy can be integrated within the building envelope and with energy conservation methods and smart-building operating strategies. Much work over the last decade or so has gone into this integration, culminating in the net-zero energy building. Much of the early emphasis was on integrating PV systems with thermal and daylighting systems. Bazilian et al. (2001) and Tripanagnostopoulos (2007) listed methods for doing this and reviewed case studies where the methods had been applied. For example, PV cells can be laid on the absorber plate of a at-plate solar collector. About 6 to 20% of the solar energy absorbed on the cells is converted to electricity; the remaining roughly 80% is available as low-temperature heat to be transferred to the uid being heated. The resulting unit produces both heat and

15,000

10,000

5,000

0 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09

Figure 3.6 | Historical trends in cumulative installed PV power of off-grid and gridconnected systems in the OECD countries (IEA, 2010e). Vertical axis is in peak megawatts.

of energy, and dynamic behaviour. Centralized PV mini-grid systems could be the least-cost options for a given level of service, and they may have a diesel generator set as an optional balancing system or operate as a hybrid PV-wind-diesel system. These kinds of systems are relevant for reducing and avoiding diesel generator use in remote areas (Munoz et al., 2007; Sreeraj et al., 2010). Grid-connected PV systems use an inverter to convert electricity from direct current (DC)as produced by the PV arrayto alternating current (AC), and then supply the generated electricity to the electricity network. Compared to an off-grid installation, system costs are lower because energy storage is not generally required, since the grid is used as a buffer. The annual output yield ranges from 300 to 2,000 kWh/ kW (Clavadetscher and Nordmann, 2007; Gaiddon and Jedliczka, 2007; Kurokawa et al., 2007; Photovoltaic Geographic Information System, 2008) for several installation conditions in the world. The average annual performance ratiothe ratio between average AC system efciency and standard DC module efciencyranges from 0.7 to 0.8 (Clavadetscher and Nordmann, 2007) and gradually increases further to about 0.9 for specic technologies and applications. Grid-connected PV systems are classied into two types of applications: distributed and centralized. Grid-connected distributed PV systems are installed to provide power to a grid-connected customer or directly to the electricity network. Such systems may be: 1) on or integrated into the customers premises, often on the demand side of the electricity meter; 2) on public and commercial buildings; or 3) simply in the built environment such as on motorway sound barriers. Typical sizes are 1 to 4 kW for residential systems, and 10 kW to several MW for rooftops on public and industrial buildings. These systems have a number of advantages: distribution losses in the electricity network are reduced because the system is installed at the point of use; extra land is not required for the PV system, and costs for mounting the systems can be reduced if the system is mounted on

354

Chapter 3

Direct Solar Energy

electricity and requires only slightly more than half the area used if the two conversion devices had been mounted side by side and worked independently. PV cells have also been developed to be applied to windows to allow daylighting and passive solar gain. Reviews of recent work in this area are provided by Chow (2010) and Arif Hasan and Sumathy (2010). Considerable work has also been done on architecturally integrating the solar components into the building. Any new solar building should be very well insulated, well sealed, and have highly efcient windows and heat recovery systems. Probst and Roecker (2007), surveying the opinions of more than 170 architects and engineers who examined numerous existing solar buildings, concluded the following: 1) best integration is achieved when the solar component is integrated as a construction element, and 2) appearanceincluding collector colour, orientation and jointingmust sometimes take precedence over performance in the overall design. In describing 16 case studies of building-integrated photovoltaics, Eiffert and Kiss (2000) identied two main products available on the architectural market: faade systems and roof systems. Faade systems include curtain wall products, spandrel panels and glazings; roong products include tiles, shingles, standing-seam products and skylights. These can be integrated as components or constitute the entire structure (as in the case of a bus shelter). The idea of the net-zero-energy solar building has sparked recent interest. Such buildings send as much excess PV-generated electrical energy to the grid as the energy they draw over the year. An IEA Task is considering how to achieve this goal (IEA NZEB, 2009). Recent examples for the Canadian climate are provided by Athienitis (2008). Starting from a building that meets the highest levels of conservation, these homes use hybrid air-heating/PV panels on the roof; the heated air is used for space heating or as a source for a heat pump. Solar water-heating collectors are included, as is fenestration permitting a large passive gain through equatorial-facing windows. A key feature is a ground-source heat pump, which provides a small amount of residual heating in the winter and cooling in the summer. Smart solar-building control strategies may be used to manage the collection, storage and distribution of locally produced solar electricity and heat to reduce and shift peak electricity demand from the grid. An example of a smart solar-building design is given by Candanedo and Athienitis (2010), where predictive control based on weather forecasts one day ahead and real-time prediction of building response are used to optimize energy performance while reducing peak electricity demand.

gas, nuclear, oil or biomasscomes from creating a hot uid. CSP simply provides an alternative heat source. Therefore, an attraction of this technology is that it builds on much of the current know-how on power generation in the world today. And it will benet not only from ongoing advances in solar concentrator technology, but also as improvements continue to be made in steam and gas turbine cycles. Any concentrating solar system depends on direct-beam irradiation as opposed to global horizontal irradiation as for at-plate systems. Thus, sites must be chosen accordingly, and the best sites for CSP are in near-equatorial cloud-free regions such as the North African desert. The average capacity factor of a solar plant will depend on the quality of the solar resource. Some of the key advantages of CSP include the following: 1) it can be installed in a range of capacities to suit varying applications and conditions, from tens of kW (dish/Stirling systems) to multiple MWs (tower and trough systems); 2) it can integrate thermal storage for peaking loads (less than one hour) and intermediate loads (three to six hours); 3) it has modular and scalable components; and 4) it does not require exotic materials. This section discusses various types of CSP systems and thermal storage for these systems. Large-scale CSP plants most commonly concentrate sunlight by reection, as opposed to refraction with lenses. Concentration is either to a line (linear focus) as in trough or linear Fresnel systems or to a point (point focus) as in central-receiver or dish systems. The major features of each type of CSP system are illustrated in Figure 3.7 and are described below. In trough concentrators, long rows of parabolic reectors concentrate the solar irradiance by the order of 70 to 100 times onto a heat collection element (HCE) mounted along the reectors focal line. The troughs track the Sun around one axis, with the axis typically being oriented north-south. The HCE comprises a steel inner pipe (coated with a solarselective surface) and a glass outer tube, with an evacuated space in between. Heat-transfer oil is circulated through the steel pipe and heated to about 390C. The hot oil from numerous rows of troughs is passed through a heat exchanger to generate steam for a conventional steam turbine generator (Rankine cycle). Land requirements are of the order of 2 km2 for a 100-MWe plant, depending on the collector technology and assuming no storage. Alternative heat transfer uids to the synthetic oil commonly used in trough receivers, such as steam and molten salt, are being developed to enable higher temperatures and overall efciencies, as well as integrated thermal storage in the case of molten salt. Linear Fresnel reectors use long lines of at or nearly at mirrors, which allow the moving parts to be mounted closer to the ground, thus reducing structural costs. (In contrast, large trough reectors presently use thermal bending to achieve the curve required in the glass surface.) The receiver is a xed inverted cavity that can have a simpler construction than evacuated tubes and be more exible in sizing. The attraction of

3.3.4

Concentrating solar power electricity generation

Concentrating solar power (CSP) technologies produce electricity by concentrating direct-beam solar irradiance to heat a liquid, solid or gas that is then used in a downstream process for electricity generation. The majority of the worlds electricity todaywhether generated by coal,

355

Direct Solar Energy

Chapter 3

(a)

(b)

Reector Absorber Tube

Curved Mirrors

Curved Mirrors

Solar Field Piping

Absorber Tube and Reconcentrator

(c)
Central Receiver

(d)

Reector

Receiver/Engine Heliostats

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

linear Fresnel reectors is that the installed costs on a per square metre basis can be lower than for trough systems. However, the annual optical performance is less than that for a trough. Central receivers (or power towers), which are one type of point-focus collector, are able to generate much higher temperatures than troughs and linear Fresnel reectors, although requiring two-axis tracking as the Sun moves through solar azimuth and solar elevation. This higher

temperature is a benet because higher-temperature thermodynamic cycles used for generating electricity are more efcient. This technology uses an array of mirrors (heliostats), with each mirror tracking the Sun and reecting the light onto a xed receiver atop a tower. Temperatures of more than 1,000C can be reached. Central receivers can easily generate the maximum temperatures of advanced steam turbines, can use high-temperature molten salt as the heat transfer uid, and can be used to power gas turbine (Brayton) cycles.

356

Chapter 3

Direct Solar Energy

Dish systems include an ideal optical reector and therefore are suitable for applications requiring high temperatures. Dish reectors are paraboloid and concentrate the solar irradiation onto a receiver mounted at the focal point, with the receiver moving with the dish. Dishes have been used to power Stirling engines at 900C, and also for steam generation. There is now signicant operational experience with dish/Stirling engine systems, and commercial rollout is planned. In 2010, the capacity of each Stirling engine is smallon the order of 10 to 25 kWelectric. The largest solar dishes have a 485-m2 aperture and are in research facilities or demonstration plants. In thermal storage, the heat from the solar eld is stored prior to reaching the turbine. Thermal storage takes the form of sensible or latent heat storage (Gil et al., 2010; Medrano et al., 2010). The solar eld needs to be oversized so that enough heat can be supplied to both operate the turbine during the day and, in parallel, charge the thermal storage. The term solar multiple refers to the total solar eld area installed divided by the solar eld area needed to operate the turbine at design point without storage. Thermal storage for CSP systems needs to be at a temperature higher than that needed for the working uid of the turbine. As such, system temperatures are generally between 400C and 600C, with the lower end for troughs and the higher end for towers. Allowable temperatures are also dictated by the limits of the media available. Examples of storage media include molten salt (presently comprising separate hot and cold tanks), steam accumulators (for short-term storage only), solid ceramic particles, high-temperature phase-change materials, graphite, and high-temperature concrete. The heat can then be drawn from the storage to generate steam for a turbine, as and when needed. Another type of storage associated with high-temperature CSP is thermochemical storage, where solar energy is stored chemically. This is discussed more fully in Sections 3.3.5 and 3.7.5. Thermal energy storage integrated into a system is an important attribute of CSP. Until recently, this has been primarily for operational purposes, providing 30 minutes to 1 hour of full-load storage. This eases the impact of thermal transients such as clouds on the plant, assists start-up and shut-down, and provides benets to the grid. Trough plants are now designed for 6 to 7.5 hours of storage, which is enough to allow operation well into the evening when peak demand can occur and tariffs are high. Trough plants in Spain are now operating with molten-salt storage. In the USA, Abengoa Solars 280-MW Solana trough project, planned to be operational by 2013, intends to integrate six hours of thermal storage. Towers, with their higher temperatures, can charge and store molten salt more efciently. Gemasolar, a 17-MWe solar tower project under construction in Spain, is designed for 15 hours of storage, giving a 75% annual capacity factor (Arce et al., 2011). Thermal storage is a means of providing dispatchability. Hybridization with non-renewable fuels is another way in which CSP can be designed to be dispatchable. Although the back-up fuel itself may

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

357

Direct Solar Energy

Chapter 3

line with technology development, regulations and investment capital. Multiple dishes have also been proposed as a source of aggregated heat, rather than distributed-generation Stirling or Brayton units. CSP or PV electricity can also be used to power reverse-osmosis plants for desalination. Dedicated CSP desalination cycles based on pressure and temperature are also being developed for desalination (see Section 3.3.2).

Figure 3.8 illustrates possible pathways to produce H2 or syngas from water and/or fossil fuels using concentrated solar energy as the source of high-temperature process heat. Feedstocks include inorganic compounds such as water and CO2, and organic sources such as coal, biomass and natural gas (NG). See Chapter 2 for parallels with biomass-derived syngas. Electrolysis of water can use solar electricity generated by PV or CSP technology in a conventional (alkaline) electrolyzer, considered a benchmark for producing solar hydrogen. With current technologies, the overall solar-to-hydrogen energy conversion efciency ranges between 10 and 14%, assuming electrolyzers working at 70% efciency and solar electricity being produced at 15% (PV) and 20% (CSP) annual efciency. The electricity demand for electrolysis can be signicantly reduced if the electrolysis of water proceeds at higher temperatures (800 to 1,000C) via solid-oxide electrolyzer cells (Jensen et al., 2007). In this case, concentrated solar energy can be applied to provide both the high-temperature process heat and the electricity needed for the high-temperature electrolysis. Thermolysis and thermochemical cycles are a long-term sustainable and carbon-neutral approach for hydrogen production from water. This route involves energy-consuming (endothermic) reactions that make use of concentrated solar irradiance as the energy source for hightemperature process heat (Abanades et al., 2006). Solar thermolysis requires temperatures above 2,200C and raises difcult challenges for reactor materials and gas separation. Water-splitting thermochemical cycles allow operation at lower temperature, but require several chemical reaction steps and also raise challenges because of inefciencies associated with heat transfer and product separation at each step. Decarbonization of fossil fuels is a near- to mid-term transition pathway to solar hydrogen that encompasses the carbothermal reduction of metal oxides (Epstein et al., 2008) and the decarbonization of fossil fuels via solar cracking (Spath and Amos, 2003; Rodat et al., 2009), reforming (Mller et al., 2006) and gasication (ZGraggen and Steinfeld, 2008; Piatkowski et al., 2009). These routes are being pursued by European, Australian and US academic and industrial research consortia. Their technical feasibility has been demonstrated in concentrating solar chemical pilot plants at the power level of 100 to 500 kWth. Solar hybrid fuel can be produced by supplying concentrated solar thermal energy to the endothermic processes of methane and biomass reformingthat is, solar heat is used for process energy only, and fossil fuels are still a required input. Some countries having vast solar and natural gas resources, but a relatively small domestic energy market (e.g., the Middle East and Australia) are in a position to produce and export solar energy in the form of liquid fuels. Solar fuel synthesis from solar hydrogen and CO2 produces hydrocarbons that are compatible with existing energy infrastructures such as the natural gas network or existing fuel supply structures. The renewable methane process combines solar hydrogen with CO2 from the

3.3.5

Solar fuel production

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

358

Chapter 3

Direct Solar Energy

Concentrated Solar Energy

H2O

H2O Splitting

Decarbonization

Fossil Fuels (NG, Oil, Coal)

Solar Thermolysis

Solar Thermochemical Cycle

Solar Electricity & Electrolysis

Solar Reforming

Solar Cracking

Solar Gasication

Optional CO2/C Sequestration Solar Fuels (Hydrogen, Syngas)

Figure 3.8 | Thermochemical routes for solar fuels production, indicating the chemical source of H2: water (H2O) for solar thermolysis and solar thermochemical cycles to produce H2 only; fossil or biomass fuels as feedstock for solar cracking to produce H2 and carbon (C); or a combination of fossil/biomass fuels and H2O/CO2 for solar reforming and gasication to produce syngas, H2 and carbon monoxide (CO). For the solar decarbonization processes, sequestration of the CO2/C may be considered (from Steinfeld and Meier, 2004; Steinfeld, 2005).

atmosphere or other sources in a synthesis reactor with a nickel catalyst. In this way, a substitute for natural gas is produced that can be stored, transported and used in gas power plants, heating systems and gas vehicles (Sterner, 2009). Solar methane can be produced using water, air, solar energy and a source of CO2. Possible CO2 sources are biomass, industry processes or the atmosphere. CO2 is regarded as the carrier for hydrogen in this energy system. By separating CO2 from the combustion process of solar methane, CO2 can be recycled in the energy system or stored permanently. Thus, carbon sink energy systems powered by RE can be created (Sterner, 2009). The rst pilot plants at the kW scale with atmospheric CO2 absorption have been set up in Germany, proving the technical feasibility. Scaling up to the utility MW scale is planned in the next few years (Specht et al., 2010). In an alternative conversion step, liquid fuels such as Fischer-Tropsch diesel, DME, methanol or solar kerosene (jet fuel) can be produced from solar energy and CO2/water (H2O) for long-distance transportation. The main advantages of these solar fuels are the same range as fossil fuels (compared to the generally reduced range of electric vehicles), less competition for land use, and higher per-hectare yields compared to biofuels. Solar energy can be harvested via natural photosynthesis in biofuels with an efciency of 0.5%, via PV power and

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

3.4

Global and regional status of market and industry development

This section looks at the ve key solar technologies, rst focusing on installed capacity and generated energy, then on industry capacity and supply chains, and nally on the impact of policies specic to these technologies.

3.4.1

Installed capacity and generated energy

This subsection discusses the installed capacity and generated energy within the ve technology areas of passive solar, active solar heating and cooling, PV electricity generation, CSP electricity generation, and solar fuel production.

359

Direct Solar Energy

Chapter 3

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

collectors. In terms of per capita use, Cyprus is the leading country in the world, with an installed capacity of 527 kWth per 1,000 inhabitants. The type of application of solar thermal energy varies greatly in different countries (Weiss and Mauthner, 2010). In China (88.7 GWth), Europe (20.9 GWth) and Japan (4.4 GWth), at-plate and evacuated-tube collectors mainly prepare hot water and provide space heating. However, in the USA and Canada, swimming pool heating is still the dominant application, with an installed capacity of 12.9 GWth of unglazed plastic collectors. The biggest reported solar thermal system for industrial process heat was installed in China in 2007. The 9 MWth plant produces heat for a textile company. About 150 large-scale plants (>500 m2; 350 kWth)1 with a total capacity of 160MWth are in operation in Europe. The largest plants for solar-assisted district heating are located in Denmark (13 MWth) and Sweden (7 MWth). In Europe, the market size more than tripled between 2002 and 2008. However, even in the leading European solar thermal markets of Austria, Greece, and Germany, only a minor portion of residential homes use solar thermal. For example, in Germany, only about 5% of one- and twofamily homes are using solar thermal energy. The European market has the largest variety of different solar thermal applications, including systems for hot-water preparation, plants for space heating of single- and multi-family houses and hotels, large-scale plants for district heating, and a growing number of systems for airconditioning, cooling and industrial applications. Advanced applications such as solar cooling and air conditioning (Henning, 2004, 2007), industrial applications (POSHIP, 2001) and desalination/water treatment are in the early stages of development. Only a few hundred rst-generation systems are in operation. For PV electricity generation, newly installed capacity in 2009 was about 7.5 GW, with shipments to rst point in the market at 7.9 GW (Jger-Waldau, 2010a; Mints, 2010). This addition brought the cumulative installed PV capacity worldwide to about 22 GWa capacity able to generate up to 26 TWh (93,600 TJ) per year. More than 90% of this capacity is installed in three leading markets: the EU27 with 16 GW (73%), Japan with 2.6 GW (12%), and the USA with 1.7 GW (8%) (Jger-Waldau, 2010b). These markets are dominated by grid-connected PV systems, and growth within PV markets has been stimulated by various government programmes around the world. Examples of such programmes include feed-in tariffs in Germany and Spain, and various mechanisms in the USA, such as buy-down incentives, investment tax credits, performance-based incentives and RE quota systems. For 2010,
1 To enable comparison, the IEAs Solar Heating and Cooling Programme, together with the European Solar Thermal Industry Federation and other major solar thermal trade associations, publish statistics in kWth (kilowatt thermal) and use a factor of 0.7 kWth/m2 to convert square metres of collector area into installed thermal capacity (kWth).

360

Chapter 3

Direct Solar Energy

the market is estimated between 9 and 24 GW of additional installed PV systems, with a consensus value in the 13GW range (Jger-Waldau, 2010a). Figure 3.9 illustrates the cumulative installed capacity for the top eight PV markets through 2009, including Germany (9,800 MW), Spain (3,500 MW), Japan (2,630 MW), the USA (1,650 MW), Italy (1,140 MW), Korea (460 MW), France (370MW) and the Peoples Republic of China (300 MW). By far, Spain and Germany have seen the largest amounts of growth in installed PV capacity in recent years, with Spain seeing a huge surge in 2008 and Germany having experienced steady growth over the last ve years.

661/2007 has been a major driving force for CSP plant construction and expansion plans. As of November 2009, 2,340 MWe of CSP projects had been preregistered for the tariff provisions of the Royal Decree. In the USA, more than 4,500 MWe of CSP are currently under power purchase agreement contracts. The different contracts specify when the projects must start delivering electricity between 2010 and 2015 (Bloem et al., 2010). More than 10,000 MWe of new CSP plants have been proposed in the USA. More than 50 CSP electricity projects are currently in the planning phase, mainly in North Africa, Spain and the USA. In Australia, the federal government has called for 1,000 MWe of new solar plants, covering both CSP and PV, under the Solar Flagships programme. Figure 3.10 shows the current and planned deployment to add more CSP capacity in the near future. Hybrid solar/fossil plants have received increasing attention in recent years, and several integrated solar combined-cycle (ISCC) projects have been either commissioned or are under construction in the Mediterranean region and the USA. The rst plant in Morocco (Ain Beni Mathar: 470 MW total, 22 MW solar) began operating in June 2010, and two additional plants in Algeria (Hassi RMel: 150 MW total, 30 MW solar) and Egypt (Al Kuraymat: 140 MW total, 20 MW solar) are under construction. In Italy, another example of an ISCC project is Archimede; however, the plants 31,000-m2 parabolic trough solar eld will be the rst to use molten salt as the heat transfer uid (SolarPACES, 2009a). Solar fuel production technologies are in an earlier stage of development. The high-temperature solar reactor technology is typically being developed at a laboratory scale of 1 to 10 kWth solar power input.
12,000
South Africa

Cumulative Installed Capacity [MW]

10,000 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Germany Spain Japan USA Italy Korea France China

Figure 3.9 | Installed PV capacity in eight markets. Data sources: EurObservER (2009); IEA (2009c); REN21 (2009); and Jger-Waldau (2010b).

Concentrating photovoltaics (CPV) is an emerging market with about 17 MW of cumulative installed capacity at the end of 2008. The two main tracks are high-concentration PV (>300 times or 300 suns) and lowto medium-concentration PV with a concentration factor of 2 to about 300 (2 to ~300 suns). To maximize the benets of CPV, the technology requires high direct-beam irradiance, and these areas have a limited geographical rangethe Sun Belt of the Earth. The market share of CPV is still small, but an increasing number of companies are focusing on CPV. In 2008, about 10 MW of CPV were installed, and market estimates for 2009 are in the 20 to 30MW range; for 2010, about 100MW are expected. Regarding CSP electricity generation, at the beginning of 2009, more than 700 MWe of grid-connected CSP plants were installed worldwide, with another 1,500 MWe under construction (Torres et al., 2010). The majority of installed plants use parabolic trough technology. Centralreceiver technology comprises a growing share of plants under construction and those announced. The bulk of the operating capacity is installed in Spain and the south-western United States. In 2007, after a hiatus of more than 15 years, the rst major CSP plants came on line with Nevada Solar One (64 MWe, USA) and PS10 (11 MWe, Spain). In Spain, successive Royal Decrees have been in place since 2004 and have stimulated the CSP industry in that country. Royal Decree

Installed Capacity [MW]

10,000

China Israel Jordan Egypt

8,000

Algeria Morocco Tunesia

6,000

Abu Dhabi Australia Spain

4,000

USA

2,000

0 1990 2000 2006 2007 2008 2009 2010 2012 2015

Figure 3.10 | Installed and planned concentrated solar power plants by country (Bloem et al., 2010).

361

Direct Solar Energy

Chapter 3

Scaling up thermochemical processes for hydrogen production to the 100-kWth power level is reported for a medium-temperature mixed iron oxide cycle (800C to 1,200C) (Roeb et al., 2006, 2009) and for the high-temperature zinc oxide (ZnO) dissociation reaction at above 1,700C (Schunk et al., 2008, 2009). Pilot plants in the power range of 300 to 500 kWth have been built for the carbothermic reduction of ZnO (Epstein et al., 2008), the steam reforming of methane (Mller et al., 2006), and the steam gasication of petcoke (ZGraggen and Steinfeld, 2008). Solar-to-gas has been demonstrated at a 30-kW scale to drive a commercial natural gas vehicle, applying a nickel catalyst (Specht et al., 2010). Demonstration at the MW scale should be warranted before erecting commercial solar chemical plants for fuels production, which are expected to be available only after 2020 (Pregger et al., 2009). Direct conversion of solar energy to fuel is not yet widely demonstrated or commercialized. But two options appear commercially feasible in the near to medium term: 1) the solar hybrid fuel production system (including solar methane reforming and solar biomass reforming), and 2) solar PV or CSP electrolysis. Australias Commonwealth Scientic and Industrial Research Organisation is running a 250-kWth reactor and plans to build a MW-scale demonstration plant using solar steam-reforming technology, with an eventual move to CO2 reforming for higher performance and less water usage. With such a system, liquid solar fuels can be produced in sunbelts such as Australia and solar energy shipped on a commercial basis to Asia and beyond. Oxygen gas produced by solar (PV or CSP) electrolysis can be used for coal gasication and partial oxidation of natural gas. With the combined process of solar electrolysis and partial oxidation of coal or methane, theoretically 10 to 15% of solar energy is incorporated into the methanol or DME. Also, the production cost of the solar hybrid fuel can be lower than the solar hydrogen produced by the solar electrolysis process only.

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

3.4.2

Industry capacity and supply chain

This subsection discusses the industry capacity and supply chain within the ve technology areas of passive solar, active solar heating and cooling, PV electricity generation, CSP electricity generation and solar fuel production. In passive solar technologies, people make up part of the industry capacity and the supply chain: namely, the engineers and architects who collaborate to produce passively heated buildings. Close collaboration between the two disciplines has often been missing in the past, but the dissemination of systematic design methodologies issued by

362

Chapter 3

Direct Solar Energy

the thermal capacity of a room. For renovation, this provides a good alternative to new heavy walls, which would require additional structural support (Hadorn, 2008). In spite of the advances in PCM, concrete has certain advantages for thermal storage when a massive building design approach is used, as in many of the Mediterranean countries. In this approach, the concrete also serves as the structure of the building and is thus likely more cost effective than thermal storage without this added function. For active solar heating and cooling, a number of different collector technologies and system approaches have been developed due to different applicationsincluding domestic hot water, heating, preheating and combined systemsand varying climatic conditions. In some parts of the production process, such as selective coatings, large-scale industrial production levels have been attained. A number of different materials, including copper, aluminium and stainless steel, are applied and combined with different welding technologies to achieve a highly efcient heat-exchange process in the collector. The materials used for the cover glass are structured or at, low-iron glass. The rst antireection coatings are coming onto the market on an industrial scale, leading to efciency improvements of about 5%. In general, vacuum-tube collectors are well-suited for higher-temperature applications. The production of vacuum-tube collectors is currently dominated by the Chinese Dewar tubes, where a metallic heat exchanger is integrated to connect them with the conventional hot-water systems. In addition, some standard vacuum-tube collectors, with metallic heat absorbers, are on the market. The largest exporters of solar water-heating systems are Australia, Greece and the USA. The majority of exports from Greece are to Cyprus and the near-Mediterranean area. France also sends a substantial number of systems to its overseas territories. The majority of US exports are to the Caribbean region. Australian companies export about 50% of production (mainly thermosyphon systems with external horizontal tanks) to most of the areas of the world that do not have hard-freeze conditions. PV electricity generation is discussed under the areas of overall solar cell production, thin-lm module production and polysilicon production. The development characteristic of the PV sector is much different than the traditional power sector, more closely resembling the semiconductor market, with annual growth rates between 40 to 50% and a high learning rate. Therefore, scientic and peer-reviewed papers can be several years behind the actual market developments due to the nature of statistical time delays and data consolidation. The only way to keep track of such a dynamic market is to use commercial market data. Global PV cell production2 reached more than 11.5 GW in 2009.
2 Solar cell production capacities mean the following: for wafer-silicon-based solar cells, only the cells; for thin lms, the complete integrated module. Only those companies that actually produce the active circuit (solar cell) are counted; companies that purchase these circuits and then make modules are not counted.

Figure 3.11 plots the increase in production from 2000 through 2009, showing regional contributions (Jger-Waldau, 2010a). The compound annual growth rate in production from 2003 to 2009 was more than 50%.
12,000

Annual PV Production [MW]

10,000

Rest of World United States China

8,000

Europe Japan

6,000

4,000

2,000

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 3.11 | Worldwide PV production from 2000 to 2009 (Jger-Waldau, 2010b).

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

Annual Production/Production Capacity [MW]

70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 Estimated Production 2009 Planned Capacity 2009 Planned Capacity 2010 Planned Capacity 2012 Planned Capacity 2015 ROW India South Korea USA China Europe Japan

Figure 3.12 | Worldwide annual PV production in 2009 compared to the announced production capacities (Jger-Waldau, 2010a).

363

Direct Solar Energy

Chapter 3

estimated in 2009 at between 5.4 and 6.1GW (including 1.5 to 1.7GW production in the Chinese province of Taiwan), Europe had 2.0 to 2.2GW, and was followed by Japan, with 1.5 to 1.7GW (Jger-Waldau, 2010b). In terms of production, First Solar (USA/Germany/France/Malaysia) was number one (1,082 MW), followed by Suntech (China) estimated at 750MW and Sharp (Japan) estimated at 580MW. If all these ambitious plans can be realized by 2015, then China will have about 51% (including 16% in the Chinese province of Taiwan) of the worldwide production capacity of 70GW, followed by Europe (15%) and Japan (13%). Worldwide, more than 300 companies produce solar cells. In 2009, silicon-based solar cells and modules represented about 80% of the worldwide market (Figure 3.13). In addition to a massive increase in production capacities, the current development predicts that thin-lm-based solar cells will increase their market share to over 30% by 2012.

2010b). The rst thin-lm factories with GW production capacity are already under construction for various thin-lm technologies. The rapid growth of the PV industry since 2000 led to the situation between 2004 and early 2008 where the demand for polysilicon outstripped the supply from the semiconductor industry. This led to a silicon shortage, which resulted in silicon spot-market prices as high as USD2005 450/kg (USD2005, assumed 2008 base) in 2008 compared to USD2005 25.5/ kg in 2003 and consequently higher prices for PV modules. This extreme price hike triggered the massive capacity expansion, not only of established companies, but of many new entrants as well. The six companies that reported shipment gures delivered together about 43,900 tonnes of polysilicon in 2008, as reported by Semiconductor Equipment and Materials International (SEMI, 2009a). In 2008, these companies had a production capacity of 48,200 tonnes of polysilicon (Service, 2009). However, all polysilicon producers, including new entrants with current and alternative technologies, had a production capacity of more than 90,000 tonnes of polysilicon in 2008. Considering that not all new capacity actually produced polysilicon at nameplate capacity in 2008, it was estimated that 62,000 tonnes of polysilicon could be produced. Subtracting the needs of the semiconductor industry and adding recycling and excess production, the available amount of silicon for the PV industry was estimated at 46,000 tonnes of polysilicon. With an average material need of 8.7 g/Wp (p = peak), this would have been sufcient for the production of 5.3 GW of crystalline silicon PV cells. The drive to reduce costs and secure key markets has led to the emergence of two interesting trends. One is the move to large original design manufacturing units, similar to the developments in the semiconductor industry. A second is that an increasing number of solar manufacturers move part of their module production close to the nal market to demonstrate the local job creation potential and ensure the current policy support. This may also be a move to manufacture in low-cost or subsidized markets. The regional distribution of polysilicon production capacities is as follows: China 20,000 tonnes, Europe 17,500 tonnes, Japan 12,000 tonnes, and USA 37,000 tonnes (Service, 2009). In 2009, solar-grade silicon production of about 88,000 tonnes was reported, sufcient for about 11GW of PV assuming an average materials need of 8g/Wp (Displaybank, 2010). China produced about 18,000 tonnes or 20% of world demand, fullling about half of its domestic demand (Baoshan, 2010). Projections of silicon production capacities for solar applications in 2012 span a range between 140,000tonnes from established polysilicon producers, up to 250,000tonnes including new producers (e.g., Bernreuther

Production Capacity [MW/yr]

70,000 60,000 50,000 40,000 30,000 20,000 10,000 0


Crystalline Wafer Silicon Thin Films

2006

2009

2010

2012

2015

Figure 3.13 | Actual (2006) and announced (2009 to 2015) production capacities of thin-lm and crystalline silicon-based solar modules (Jger-Waldau, 2010b).

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

364

Chapter 3

Direct Solar Energy

and Haugwitz, 2010; Ruhl et al., 2010). The possible solar cell production will also depend on the material use per Wp. Material consumption could decrease from the current 8g/Wp to 7g/Wp or even 6g/Wp (which could increase delivered PV capacity from 31 to 36 to 42 GW, respectively), but this may not be achieved by all manufacturers. Forecasts of the future costs of vital materials have a high-prole history, and there is ongoing public debate about possible material shortages and competition regarding some (semi-)metals (e.g., In and Te) used in thin-lm cell production. In a recent study, Wadia et al. (2009) explored material limits for PV expansion by examining the dual constraints of material supply and least cost per watt for the most promising semiconductors as active photo-generating materials. Contrary to the commonly assumed scarcity of indium and tellurium, the study concluded that the currently known economic reserves of these materials would allow about 10TW of CdTe or CuInS2 solar cells to be installed. In CSP electricity generation, the solar collector eld is readily scalable, and the power block is based on adapted knowledge from the existing power industry such as steam and gas turbines. The collectors themselves benet from a range of existing skill sets such as mechanical, structural and control engineers, and metallurgists. Often, the materials or components used in the collectors are already mass-produced, such as glass mirrors. By the end of 2010, strong competition had emerged and an increasing number of companies had developed industry-level capability to supply materials such as high-reectivity glass mirrors and manufactured components. Nonetheless, the large evacuated tubes designed specically for use in trough/oil systems for power generation remain a specialized component, and only two companies (Schott and Solel) have been capable of supplying large orders of tubes, with a third company (Archimedes) now emerging. The trough concentrator itself comprises know-how in both structures and thermally sagged glass mirrors. Although more companies are now offering new trough designs and considering alternatives to conventional rear-silvered glass (e.g., polymer-based reective lms), the essential technology of concentration remains unchanged. Direct steam generation in troughs is under demonstration, as is direct heating of molten salt, but these designs are not yet commercially available. As a result of its successful operational history, the trough/oil technology comprised most of the CSP installed capacity in 2010. Linear Fresnel and central-receiver systems comprise a high level of know-how, but the essential technology is such that there is the potential for a greater variety of new industry participants. Although only a couple of companies have historically been involved with central receivers, new players have entered the market over the last few years. There are also technology developers and projects at the demonstration level (China, USA, Israel, Australia, Spain). Central-receiver developers are aiming for higher temperatures, and, in some cases, alternative heat

transfer uids such as molten salts. The accepted standard to date has been to use large heliostats, but many of the new entrants are pursuing much smaller heliostats to gain potential cost reductions through highvolume mass production. The companies now interested in heliostat development range from optics companies to the automotive industry looking to diversify. High-temperature steam receivers will benet from existing knowledge in the boiler industry. Similarly, with linear Fresnel, a range of new developments are occurring, although not yet as developed as the central-receiver technology. Dish technology is much more specialized, and most effort presently has been towards developing the dish/Stirling concept as a commercial product. Again, the technology can be developed as specialized components through specic industry know-how such as the Stirling engine mass-produced through the automotive industry. Within less than 10 years prior to 2010, the CSP industry has gone from negligible activity to over 2,400 MWe either commissioned or under construction. A list of new CSP plants and their characteristics can be found at the IEA SolarPACES web site.3 More than ten different companies are now active in building or preparing for commercial-scale plants, compared to perhaps only two or three who were in a position to build a commercial-scale plant three years ago. These companies range from large organizations with international construction and project management expertise who have acquired rights to specic technologies, to start-ups based on their own technology developed in-house. In addition, major independent power producers and energy utilities are playing a role in the CSP market. The supply chain does not tend to be limited by raw materials, because the majority of required materials are bulk commodities such as glass, steel/aluminium, and concrete. The sudden new demand for the specic solar salt mixture material for molten-salt storage is claimed to have impacted supply. At present, evacuated tubes for trough plants can be produced at a sufcient rate to service several hundred MW per year. However, expanded capacity can be introduced readily through new factories with an 18-month lead time. Solar fuel technology is still at an emerging stagethus, there is no supply chain in place at present for commercial applications. However, solar fuels will comprise much of the same solar-eld technology being deployed for other high-temperature CSP systems, with solar fuels requiring a different receiver/reactor at the focus and different downstream processing and control. Much of the downstream technology, such as Fischer-Tropsch liquid fuel plants, would come from existing expertise in the petrochemical industry. The scale of solar fuel demonstration plants is being ramped up to build condence for industry, which will eventually expand operations.
3 See: www.solarpaces.org.

365

Direct Solar Energy

Chapter 3

Hydrogen has been touted as a future transportation fuel due to its versatility, pollutant-free end use and storage capability. The key is a sustainable, CO2-free source of hydrogen such as solar, cost-effective storage and appropriate distribution infrastructure. The production of solar hydrogen, in and of itself, does not produce a hydrogen economy because many factors are needed in the chain. The suggested path to solar hydrogen is to begin with solar enhancement of existing steam reforming processes, with a second generation involving solar electricity and advanced electrolysis, and a third generation using thermolysis or advanced thermochemical cycles, with many researchers aiming for the production of fuels from concentrated solar energy, water, and CO2. In terms of making a transition, solar hydrogen can be mixed with natural gas and transported together in existing pipelines and distribution networks to customers, thus enhancing the solar portion of the global energy mix. Steam reforming of natural gas for hydrogen production is a conventional industrial-scale process that produces most of the worlds hydrogen today, with the heat for the process derived from burning a signicant proportion of the fossil fuel feedstock. Using concentrated solar power, instead, as the source of the heat embodies solar energy in the fuel. The solar steam-reforming of natural gas and other hydrocarbons, and the solar steam-gasication of coal and other carbonaceous materials yields a high-quality syngas, which is the building block for a wide variety of synthetic fuels including Fischer-Tropsch-type chemicals, hydrogen, ammonia and methanol (Steinfeld and Meier, 2004). The solar cracking route refers to the thermal decomposition of natural gas and other hydrocarbons. Besides H2 and carbon, other compounds may also be formed, depending on the reaction kinetics and on the presence of impurities in the raw materials. The thermal decomposition yields a carbon-rich condensed phase and a hydrogen-rich gas phase. The carbonaceous solid product can either be sequestered without CO2 release or used as material commodity (carbon black) under less severe CO2 restraints. It can also be applied as reducing agent in metallurgical processes. The hydrogen-rich gas mixture can be further processed to high-purity hydrogen that is not contaminated with oxides of carbon; thus, it can be used in proton-exchange-membrane fuel cells without inhibiting platinum electrodes. From the perspective of carbon sequestration, it is easier to separate, handle, transport and store solid carbon than gaseous CO2. Further, thermal cracking removes and separates carbon in a single step. The major drawback of thermal cracking is the energy loss associated with the sequestration of carbon. Thus, solar cracking may be the preferred option for natural gas and other hydrocarbons with a high H2/C ratio (Steinfeld and Meier, 2004).

Chapter 1. Solar technologies differ in levels of maturity, and although some applications are already competitive in localized markets, they generally face one common barrier: the need to achieve cost reductions (see Section 3.8). Utility-scale CSP and PV systems face different barriers than distributed PV and solar heating and cooling technologies. Important barriers include: 1) siting, permitting and nancing challenges to develop land with favourable solar resources for utility-scale projects; 2) lack of access to transmission lines for large projects far from electric load centres; 3) complex access laws, permitting procedures and fees for smaller-scale projects; 4) lack of consistent interconnection standards and time-varying utility rate structures that capture the value of distributed generated electricity; 5) inconsistent standards and certications and enforcement of these issues; and 6) lack of regulatory structures that capture environmental and risk mitigation benets across technologies (Denholm et al., 2009). Through appropriate policy designs (see Chapter 11), governments have shown that they can support solar technologies by funding R&D and by providing incentives to overcome economic barriers. Price-driven instruments (see Section 11.5.2), for example, were popularized after feed-in tariff (FIT) policies boosted levels of PV deployment in Germany and Spain. In 2009, various forms of FIT policies were implemented in more than 50 countries (REN21, 2010) and some designs offer premiums for building-integrated PV. Quota-driven frameworks such as renewable portfolio standards (RPS) and government bidding are common in the USA and China, respectively (IEA, 2009a). Traditional RPS frameworks are designed to be technology-neutral, and this puts at a disadvantage many solar applications that are more costly than alternatives such as wind power. In response, features of RPS frameworks (set-asides and credits) increasingly are including solar-specic policies, and such programs have led to increasing levels of solar installations (Wiser et al., 2010). In addition to these regulatory frameworks, scal policies and nancing mechanisms (e.g., tax credits, soft loans and grants) are often employed to support the manufacturing of solar goods and to increase consumer demand (Rickerson et al., 2009). The challenge for solar projects to secure nancing is a critical barrier, especially for developing technologies in market structures dominated by short-term transactions and planning. Most successful solar policies are tailored to the barriers posed by specic applications. Across technologies, there is a need to offset relatively high upfront investment costs (Denholm et al., 2009). Yet, in the case of utility-scale CSP and PV projects, substantial and long-term investments are required at levels that exceed solar applications in distributed markets. Solar heating and cooling technologies are included in many policies, yet the characteristics of their applications differ from electricity-generating technologies. Policies based on energy yield rather than collector surface area are generally preferred for various types of solar thermal collectors (IEA, 2007). See Section 1.5 for further discussion. Similar to other renewable sources, there is ongoing discussion about the merits of existing solar policies to spur innovation and accelerate deployment using cost-effective measures. Generallyand as discussed

3.4.3

Impact of policies4

Direct solar energy technologies support a broad range of applications, and their deployment is confronted by many of the barriers outlined in
4 Non-technology-specic policy issues are covered in Chapter 11 of this report.

366

Chapter 3

Direct Solar Energy

in Chapter 11the most successful policies are those that send clear, long-term and consistent signals to the market. In addition to targeted economic policies, government action through educationally based schemes (e.g., workshops, workforce training programs and seminars) and engagement of regulatory organizations are helping to overcome many of the barriers listed in this section.

3.5.2

District heating and other thermal loads

3.5

Integration into the broader energy system5

This section discusses how direct solar energy technologies are part of the broader energy framework, focusing specically on the following: low-capacity energy demand; district heating and other thermal loads; PV generation characteristics and the smoothing effect; and CSP generation characteristics and grid stabilization. Chapter 8 addresses the broader technical and institutional options for managing the unique characteristics, production variability, limited predictability and locational dependence of some RE technologies, including solar, as well as existing experience with and studies associated with the costs of that integration.

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

3.5.1

Low-capacity electricity demand

There can be comparative advantages for using solar energy rather than non-renewable fuels in many developing countries. Within a country, the advantages can be higher in un-electried rural areas compared to urban areas. Indeed, solar energy has the advantage, due to being modular, of being able to provide small and decentralized supplies, as well as large centralized ones. For more on integrated buildings and households, see Section 8.3.2. In a wide range of countries, particularly those that are not oil producers, solar energy and other forms of RE can be the most appropriate energy source. If electricity demand exceeds supply, the lack of electricity can prevent development of many economic sectors. Even in countries with high solar energy sustainable development potential, RE is often only considered to satisfy high-power requirements such as the industrial sector. However, large-scale technologies such as CSP are often not available to them due, for example, to resource conditions or suitable land area availability. In such cases, it is reasonable to keep the electricity generated near the source to provide high amounts of power to cover industrial needs. Applications that have low power consumption, such as lighting in rural areas, can primarily be satised using onsite PVeven if the business plan for electrication of the area indicates that a grid connection would be more protable. Furthermore, the criteria to determine the most suitable technological option for electrifying a rural area should include benets such as local economic development, exploiting natural resources, creating jobs, reducing the countrys dependence on imports, and protecting the environment.
5 Non-technology-specic issues related to integration of RE sources in current and future energy systems are covered in Chapter 8 of this report.

367

Direct Solar Energy

Chapter 3

3.5.3

Photovoltaic generation characteristics and the smoothing effect

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

Wiemken et al. (2001) used data from actual PV systems in Germany to demonstrate that ve-minute ramps in normalized PV power output at one site may exceed 50%, but that ve-minute ramps in the normalized PV power output from 100 PV systems spread throughout the country never exceed 5%. Ramachandran et al. (2004) analyzed the reduction in power output uctuation for spatially dispersed PV systems and for different time periods, and they proposed a cluster model to represent very large numbers of small, geographically dispersed PV systems. Results from Curtright and Apt (2008) based on three PV systems in Arizona indicate that 10-minute step changes in output can exceed 60% of PV capacity at individual sites, but that the maximum of the aggregate of three sites is reduced. Kawasaki et al. (2006) similarly analyzed the smoothing effect within a small (4 km by 4 km) network of irradiance sensors and concluded that the smoothing effect is most effective during times when the irradiance variability is most severe particularly days characterized as partly cloudy. Murata et al. (2009) developed and validated a method for estimating the variability of power output from PV plants dispersed over a wide area that is very similar to the methods used for wind by Ilex Energy Consulting Ltd et al. (2004) and Holttinen (2005). Mills and Wiser (2010) measured one-minute solar insolation for 23 sites in the USA and characterized the variability of PV with different degrees of geographic diversity, comparing the variability of PV to the variability of similarly sited wind. They determined that the relative aggregate variability of PV plants sited in a dense ten by ten array with 20-km spacing is six times less than the variability of a single site for variability on time scales of less than 15 minutes. They also found that for PV and wind plants similarly sited in a ve by ve grid with 50-km spacing, the variability of PV is only slightly more than the variability of wind on time scales of 5 to 15 minutes. Oozeki et al. (2010) quantitatively evaluated the smoothing effect in a load-dispatch control area in Japan to determine the importance of data accumulation and analysis. The study also proposed a methodology to calculate the total PV output from a limited number of measurement data using Voronoi Tessellation. Marcos et al. (2010) analyzed onesecond data collected throughout a year from six PV systems in Spain, ranging from 1 to 9.5 MWp, totalling 18 MW. These studies concluded that over shorter and longer time scales, the level of variability is nearly identical because the aggregate uctuation of PV systems spread over the large area depends on the correlation of the uctuation between PV systems. The correlation of uctuation, in turn, is a function both of the time scale and distance between PV systems. Variability is less correlated for PV systems that are further apart and for variability over shorter time scales. Currently, however, not enough data on generation characteristics exist to evaluate the smoothing effect. Data collection from a sufciently large number of sites (more than 1,000 sites and at distances of 2 to 200 km), periods and time resolution (one minute or less) had just begun in mid-2010 in several areas in the world. The smoothed generation characteristics of PV penetration considering area and multiple sites will

368

Chapter 3

Direct Solar Energy

be analyzed precisely after collecting reliable measurement data with sufcient time resolution and time synchronization. The results will contribute to the economic and reliable integration of PV into the energy system.

3.6.1

Environmental impacts

3.5.4

Concentrating solar power generation characteristics and grid stabilization

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

No consensus exists on the premium, if any, that society should pay for cleaner energy. However, in recent years, there has been progress in analyzing environmental damage costs, thanks to several major projects to evaluate the externalities of energy in the USA and Europe (Gordon, 2001; Bickel and Friedrich, 2005; NEEDS, 2009; NRC, 2010). Solar energy has been considered desirable because it poses a much smaller environmental burden than non-renewable sources of energy. This argument has almost always been justied by qualitative appeals, although this is changing. Results for damage costs per kilogram of pollutant and per kWh were presented by the International Solar Energy Society in Gordon (2001). The results of studies such as NEEDS (2009), summarized in Table 3.3 for PV and in Table 3.4 for CSP, conrm that RE is usually comparatively benecial, though impacts still exist. In comparison to the gures presented for PV and CSP here, the external costs associated with fossil generation options, as summarized in Chapter 10.6, are considerably higher, especially for coal-red generation. Considering passive solar technology, higher insulation levels provide many benets, in addition to reducing heating loads and associated costs (Harvey, 2006). The small rate of heat loss associated with high levels of insulation, combined with large internal thermal mass, creates a more comfortable dwelling because temperatures are more uniform. This can indirectly lead to higher efciency in the equipment supplying the heat. It also permits alternative heating systems that would not

Table 3.3 | Quantiable external costs for photovoltaic, tilted-roof, single-crystalline silicon, retrot, average European conditions; in US2005 cents/kWh (NEEDS, 2009). 2005 Health Impacts Biodiversity Crop Yield Losses Material Damage Land Use Total
0.17 0.01 0.00 0.00 N/A

2025
0.14 0.01 0.00 0.00 0.01

2050
0.10 0.01 0.00 0.00 0.01

0.18

0.17

0.12

Table 3.4 | Quantiable external costs for concentrating solar power; in US2005 cents/ kWh (NEEDS, 2009). 2005 2025
0.10 0.00 0.00 0.00 N/A

2050
0.06 0.00 0.00 0.00 N/A

3.6

Environmental and social impacts6

Health Impacts Biodiversity

0.65 0.03 0.00 0.01 N/A

This section rst discusses the environmental impacts of direct solar technologies, and then describes potential social impacts. However, an overall issue identied at the start is the small number of peer-reviewed studies on impacts, indicating the need for much more work in this area.
6 A comprehensive assessment of social and environmental impacts of all RE sources covered in this report can be found in Chapter 9.

Crop Yield Losses Material Damage Land Use Total

0.69

0.10

0.06

369

Chapter 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), nund 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, . Killingtveit, Z. Liu, 2011: Hydropower. In IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation [O. Edenhofer, R. Pichs-Madruga, Y. Sokona, K. Seyboth, P. Matschoss, S. Kadner, T. Zwickel, P. Eickemeier, G. Hansen, S. Schlmer, C. von Stechow (eds)], Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

437

Chapter 5

Hydropower

signicant effect on the power output. Increased risks of landslides and glacial lake outbursts, and impacts of increased variability, are of particular concern to Himalayan countries (Agrawala et al., 2003). The possibility of accommodating increased intensity of seasonal precipitation by increasing storage capacities may become of particular importance (Iimi, 2007). In Europe, by the 2070s, hydropower potential for the whole of Europe has been estimated to potentially decline by 6%, translated into a 20 to 50% decrease around the Mediterranean, a 15 to 30% increase in northern and Eastern Europe, and a stable hydropower pattern for western and central Europe (Lehner et al., 2005). In New Zealand, increased westerly wind speed is very likely to enhance wind generation and spill over precipitation into major South Island watersheds, and to increase winter rain in the Waikato catchment. Warming is virtually certain to increase melting of snow, the ratio of rainfall to snowfall, and to increase river ows in winter and early spring. This is very likely to increase hydroelectric generation during the winter peak demand period, and to reduce demand for storage. In Latin America, hydropower is the main electrical energy source for most countries, and the region is vulnerable to large-scale and persistent rainfall anomalies due to El Nio and La Nia, as observed in Argentina, Colombia, Brazil, Chile, Peru, Uruguay and Venezuela. A combination of increased energy demand and droughts caused a virtual breakdown of hydroelectricity in most of Brazil in 2001 and contributed to a reduction in gross domestic product (GDP). Glacier retreat is also affecting hydropower generation, as observed in the cities of La Paz and Lima. In North America, hydropower production is known to be sensitive to total runoff, to its timing, and to reservoir levels. During the 1990s, for example, Great Lakes levels fell as a result of a lengthy drought, and in 1999, hydropower production was down signicantly both at Niagara and Sault St. Marie. For a 2C to 3C warming in the Columbia River Basin and BC Hydro service areas, the hydroelectric supply under worstcase water conditions for winter peak demand is likely to increase (high condence). Similarly, Colorado River hydropower yields are likely to decrease signicantly, as will Great Lakes hydropower. Northern Qubec hydropower production would be likely to benet from greater precipitation and more open-water conditions, but hydropower plants in southern Qubec would be likely to be affected by lower water levels. Consequences of changes in the seasonal distribution of ows and in the timing of ice formation are uncertain. In a recent study (Hamududu and Killingtveit, 2010), the regional and global changes in hydropower generation for the existing hydropower system were computed, based on a global assessment of changes in river ow by 2050 (Milly et al., 2005, 2008) for the SRES A1B scenario using 12 different climate models. The computation was done at the country or political region (USA, Canada, Brazil, India, China, Australia) level, and summed up to regional and global values (see Table 5.2).

Table 5.2 | Power generation capacity in GW and TWh/yr (2005) and estimated changes (TWh/yr) due to climate change by 2050. Results are based on an analysis using the SRES A1B scenario in 12 different climate models (Milly et al., 2008), UNEP world regions and data for the hydropower system in 2005 (US DOE, 2009) as presented in Hamududu and Killingtveit (2010). REGION
Africa Asia Europe North America South America Oceania TOTAL

Power Generation Capacity (2005) GW


22 246 177 161 119 13 737

TWh/yr (PJ/yr)
90 (324) 996 (3,586) 517 (1,861) 655 (2,358) 661 (2,380) 40 (144) 2931 (10,552)

Change by 2050 TWh/yr (PJ/yr)


0.0 (0) 2.7 (9.7) -0.8 (-2.9) 0.3 (1) 0.3 (1) 0.0 (0) 2.5 (9)

In general the results given in Table 5.2 are consistent with the (mostly qualitative) results given in previous studies (IPCC, 2007b; Bates et al., 2008). For Europe, the computed reduction (-0.2%) has the same sign, but is less than the -6% found by Lehner et al. (2005). One reason could be that Table 5.2 shows changes by 2050 while Lehner et al. (2005) give changes by 2070, so a direct comparison is difcult. It can be concluded that the overall impacts of climate change on the existing global hydropower generation may be expected to be small, or even slightly positive. However, results also indicated substantial variations in changes in energy production across regions and even within countries (Hamududu and Killingtveit, 2010). Insofar as a future expansion of the hydropower system will occur incrementally in the same general areas/watersheds as the existing system, these results indicate that climate change impacts globally and averaged across regions may also be small and slightly positive. Still, uncertainty about future impacts as well as increasing difculty of future systems operations may pose a challenge that must be addressed in the planning and development of future HPP (Hamududu et al., 2010). Indirect effects on water availability for energy purposes may occur if water demand for other uses such as irrigation and water supply for households and industry rises due to the climate change. This effect is difcult to quantify, and it is further discussed in Section 5.10.

5.3

Technology and applications

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

449

Hydropower

Chapter 5

current trends in technology development, and trends in renovation and modernization follow in Sections 5.3.3 and 5.3.4 respectively.

5.3.1

Classication by head and size

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

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

Table 5.3 | Small-scale hydropower by installed capacity (MW) as dened by various countries Country
Brazil Canada China EU Linking Directive India Norway Sweden USA

Small-scale hydro as dened by installed capacity (MW)


30 <50 50 20 25 10 1.5 5100

Reference Declaration
Brazil Government Law No. 9648, of May 27, 1998 Natural Resources Canada, 2009: canmetenergy-canmetenergie.nrcan-rncan.gc.ca/eng/renewables/ small_hydropower.html Jinghe (2005); Wang (2010) EU Linking directive, Directive 2004/101/EC, article 11a, (6) Ministry of New and Renewable Energy, 2010: www.mnre.gov.in/ Norwegian Ministry of Petroleum and Energy. Facts 2008. Energy and Water Resources in Norway; p.27 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 RPS)

Norwegian Water Resources and Energy Directorate, Water resource act and regulations, 2001.

450

Chapter 5

Hydropower

5.3.2

Classication by facility type

HPPs is relatively inexpensive and such facilities have, in general, lower environmental impacts than similar-sized storage hydropower plants.

Hydropower plants are often classied in three main categories according to operation and type of ow. Run-of-river (RoR), storage (reservoir) and pumped storage HPPs all vary from the very small to the very large scale, depending on the hydrology and topography of the watershed. In addition, there is a fourth category called in-stream technology, which is a young and less-developed technology.

5.3.2.2

Storage Hydropower

5.3.2.1

Run-of-River

A RoR HPP draws the energy for electricity production mainly from the available ow of the river. Such a hydropower plant may include some short-term storage (hourly, daily), allowing for some adaptations to the demand prole, but the generation prole will to varying degrees be dictated by local river ow conditions. As a result, generation depends on precipitation and runoff and may have substantial daily, monthly or seasonal variations. When even short-term storage is not included, RoR HPPs will have generation proles that are even more variable, especially when situated in small rivers or streams that experience widely varying ows. In a RoR HPP, a portion of the river water might be diverted to a channel or pipeline (penstock) to convey the water to a hydraulic turbine, which is connected to an electricity generator (see Figure 5.5). RoR projects

Hydropower projects with a reservoir are also called storage hydropower since they store water for later consumption. The reservoir reduces the dependence on the variability of inow. The generating stations are located at the dam toe or further downstream, connected to the reservoir through tunnels or pipelines. (Figure 5.6). The type and design of reservoirs are decided by the landscape and in many parts of the world are inundated river valleys where the reservoir is an articial lake. In geographies with mountain plateaus, high-altitude lakes make up another kind of reservoir that often will retain many of the properties

Dam

Penstock Powerhouse Switch Yard

Tailrace Desilting Tank Headrace Forebay Intake Diversion Weir

Figure 5.6 | Typical hydropower plant with reservoir.

Penstock

Powerhouse

Tailrace

Stream

of the original lake. In these types of settings, the generating station is often connected to the lake serving as reservoir via tunnels coming up beneath the lake (lake tapping). For example, in Scandinavia, natural high-altitude lakes are the basis for high pressure systems where the heads may reach over 1,000 m. One power plant may have tunnels coming from several reservoirs and may also, where opportunities exist, be connected to neighbouring watersheds or rivers. The design of the HPP and type of reservoir that can be built is very much dependent on opportunities offered by the topography.

5.3.2.3
Figure 5.5 | Run-of-river hydropower plant.

Pumped storage

may form cascades along a river valley, often with a reservoir-type HPP in the upper reaches of the valley that allows both to benet from the cumulative capacity of the various power stations. Installation of RoR

Pumped storage plants are not energy sources, but are instead storage devices. In such a system, water is pumped from a lower reservoir into an upper reservoir (Figure 5.7), usually during off-peak hours, while ow is reversed to generate electricity during the daily peak load period or at

451

Hydropower

Chapter 5

5.3.3
Upper Reservoir

Status and current trends in technology development

Pumping Pumped Storage Power Plant

Generating

Lower Reservoir

Hydropower is a proven and well-advanced technology based on more than a century of experiencewith many examples of hydropower plants built in the 19th century still in operation today. Hydropower today is an extremely exible power technology with among the best conversion efciencies of all energy sources (~90%, water to wire) due to its direct transformation of hydraulic energy to electricity (IEA, 2004). Still, there is room for further improvements, for example, by improving operation, reducing environmental impacts, adapting to new social and environmental requirements and by developing more robust and cost-effective technological solutions. The status and current trends are presented below, and options and prospects for future technology innovations are discussed in Section 5.7.

Figure 5.7 | Typical pumped storage project.

5.3.3.1

Efciency

other times of need. Although the losses of the pumping process make such a plant a net energy consumer overall, the plant is able to provide large-scale energy storage system benets. In fact, pumped storage is the largest-capacity form of grid energy storage now readily available worldwide (see Section 5.5.5).

The potential for energy production in a hydropower plant is determined by the following parameters, which are dependent on the hydrology, topography and design of the power plant: The amount of water available; Water loss due to ood spill, bypass requirements or leakage;

5.3.2.4

In-stream technology using existing facilities

The difference in head between upstream intake and downstream outlet; Hydraulic losses in water transport due to friction and velocity change; and The efciency in energy conversion of electromechanical equipment. The total amount of water available at the intake will usually not be possible to utilize in the turbines because some of the water will be lost or will not be withdrawn. This loss occurs because of water spill during high ows when inow exceeds the turbine capacity, because of bypass releases for environmental ows, and because of leakage. In the hydropower plant the potential (gravitational) energy in water is transformed into kinetic energy and then mechanical energy in the turbine and further to electrical energy in the generator. The energy transformation process in modern hydropower plants is highly efcient, usually with well over 90% mechanical efciency in turbines and over 99% in the generator. The inefciency is due to hydraulic loss in the water circuit (intake, turbine and tailrace), mechanical loss in the turbogenerator group and electrical loss in the generator. Old turbines can

To optimize existing facilities like weirs, barrages, canals or falls, small turbines or hydrokinetic turbines can be installed for electricity generation. These basically function like a run-of-river scheme, as shown in Figure 5.8. Hydrokinetic devices being developed to capture energy from tides and currents may also be deployed inland in both free-owing rivers and in engineered waterways (see Section 5.7.4).

Spillway

Diversion Canal

Irrigation Canal

Switch Yard

Powerhouse Tailrace Channel

Figure 5.8 | Typical in-stream hydropower plant using existing facilities.

452

Chapter 5

Hydropower

have lower efciency, and efciency can also be reduced due to wear and abrasion caused by sediments in the water. The rest of the potential energy is lost as heat in the water and in the generator. In addition, some energy losses occur in the headrace section where water ows from the intake to the turbines, and in the tailrace section taking water from the turbine back to the river downstream. These losses, called head loss, reduce the head and hence the energy potential for the power plant. These losses can be classied either as friction losses or singular losses. Friction losses depend mainly on water velocity and the roughness in tunnels, pipelines and penstocks. The total efciency of a hydropower plant is determined by the sum of these three loss components. Hydraulic losses can be reduced by increasing the turbine capacity or by increasing the reservoir capacity to get better regulation of the ow. Head losses can be reduced by increasing the area of headrace and tailrace, by decreasing the roughness in

these and by avoiding too many changes in ow velocity and direction. The efciency of electromechanical equipment, especially turbines, can be improved by better design and also by selecting a turbine type with an efciency prole that is best adapted to the duration curve of the inow. Different turbine types have quite different efciency proles when the turbine discharge deviates from the optimal value (see Figure 5.9). Improvements in turbine design by computational uid dynamics software and other innovations are discussed in Section 5.7.

5.3.3.2

Tunnelling capacity

In hydropower projects, tunnels in hard and soft rock are often used for transporting water from the intake to the turbines (headrace), and from the turbine back to the river, lake or fjord downstream (tailrace). In addition, tunnels are used for a number of other purposes when the power station is placed underground, for example for access, power cables, surge shafts

Efciency () [%]

100

Francis
90

Kaplan

80

Pelton

70

60

Crossow

50

Propeller

40

30

20

10

10

20

30

40

50

60

70

80

90

100

110

120

Relative Discharge [%]


Figure 5.9 | Typical efciency curves for different types of hydropower turbines (Vinogg and Elstad, 2003).

453

Hydropower

Chapter 5

and ventilation. Tunnels are increasingly favoured for hydropower construction as a replacement for surface structures like canals and penstocks. Tunnelling technology has improved greatly due to the introduction of increasingly efcient equipment, as illustrated by Figure 5.10 (Zare and Bruland, 2007). Today, the two most important technologies for hydropower tunnelling are the drill and blast method and the use of tunnel-boring machines (TBM). The drill and blast method is the conventional method for tunnel excavation in hard rock. Thanks to the development in tunnelling technology, excavation costs have been reduced by 25%, or 0.8%/yr, over the past 30 years (see Figure 5.10). TBMs excavate the entire cross section in one operation without the use of explosives. TBMs carry out several successive operations: drilling, support of the ground traversed and construction of the tunnel. The diameter of tunnels constructed can be from <1 m (micro tunnelling) up to 15 m. The excavation progress of the tunnel is typically from 30 up to 60 m/day.

In addition, for HPPs with reservoirs, their storage capacity can be lled up by sediments, which requires special technical mitigation measures or plant design. Lysne et al. (2003) reported that the effects of sediment-induced wear of turbines in power plants can be, among others: Generation loss due to reduction in turbine efciency; Increase in frequency of repair and maintenance; Increase in generation losses due to downtime; Reduction in lifetime of the turbine; and Reduction in regularity of power generation.

Tunnel Excavation Costs [USD/m]

3,000

2,500

All of these effects are associated with revenue losses and increased maintenance costs. Several promising concepts for sediment control at intakes and mechanical removal of sediment from reservoirs and for settling basins have been developed and practised. A number of authors (Mahmood, 1987; Morris and Fan, 1997; ICOLD, 1999; Palmieri et al., 2003; White, 2005) have reported measures to mitigate the sedimentation problems by better management of land use practices in upstream watersheds to reduce erosion and sediment loading, mechanical removal of sediment from reservoirs and design of hydraulic machineries aiming to resist the effect of sediment passing through them.

2,000

5.3.4
1,500

Renovation, modernization and upgrading

1,000

500

0 1975 1979 1983 1988 1995 2005

Figure 5.10 | Developments in tunnelling technology: the trend in excavation costs for a 60 m2 tunnel, in USD2005 per metre (adapted from Zare and Bruland, 2007).

5.3.3.3

Technical challenges related to sedimentation management

Although sedimentation problems are not found in all rivers (see Section 5.6.1.4), operating a hydropower project in a river with a large sediment load comes with serious technical challenges. Specically, increased sediment load in the river water induces wear on hydraulic machinery and other structures of the hydropower plant. Deposition of sediments can obstruct intakes, block the ow of water through the system and also impact the turbines. The sediment-induced wear of the hydraulic machinery is more serious when there is no room for storage of sediments.

Renovation, modernization and upgrading (RM&U) of old power stations is often less costly than developing a new power plant, often has relatively smaller environment and social impacts, and requires less time for implementation. Capacity additions through RM&U of old power stations can therefore be attractive. Selective replacement or repair of identied hydro powerhouse components like turbine runners, generator windings, excitation systems, governors, control panels or trash cleaning devices can reduce costs and save time. It can also lead to increased efciency, peak power and energy availability of the plant (Prabhakar and Pathariya, 2007). RM&U may allow for restoring or improving environmental conditions in already-regulated areas. Several national programmes for RM&U are available. For example, the Research Council of Norway recently initiated a program with the aim to increase power production in existing hydropower plants and at the same time improve environmental conditions.10 The US Department of Energy has been using a similar approach to new technology development since 1994 when it started the Advanced Hydropower Turbine Systems Program that emphasized simultaneous improvements in energy and environmental performance (Odeh, 1999; Cada, 2001; Sale et al., 2006a). Normally the life of hydroelectric power plants is 40 to 80 years. Electromechanical equipment may need to be upgraded or replaced after 30 to 40 years, however, while civil structures like dams, tunnels
10 Centre for Environmental Design of Renewable Energy: www.cedren.no/.

454

Chapter 5

Hydropower

etc. usually function longer before they requires renovation. The lifespan of properly maintained hydropower plants can exceed 100 years. Using modern control and regulatory equipment leads to increased reliability (Prabhakar and Pathariya, 2007). Upgrading hydropower plants calls for a systematic approach, as a number of hydraulic, mechanical, electrical and economic factors play a vital role in deciding the course of action. For techno-economic reasons, it can also be desirable to consider up-rating (i.e., increasing the size of the hydropower plant) along with RM&U/life extension. Hydropower generating equipment with improved performance can also be retrotted, often to accommodate market demands for more exible, peaking modes of operation. Most of the existing worldwide hydropower equipment in operation will need to be modernized to some degree by 2030 (SER, 2007). Refurbished or up-rated hydropower plants also result in incremental increases in hydropower generation due to more efcient turbines and generators. In addition, existing infrastructure without hydropower plants (like existing barrages, weirs, dams, canal fall structures, water supply schemes) can also be reworked by adding new hydropower facilities. The majority of the worlds 45,000 large dams were not built for hydropower purposes, but for irrigation, ood control, navigation and urban water supply schemes (WCD, 2000). Retrotting these

with turbines may represent a substantial potential, because only about 25% of large reservoirs are currently used for hydropower production. For example, from 1997 to 2008 in India, about 500 MW have been developed on existing facilities. A recent study in the USA indicated some 20 GW could be installed by adding hydropower capacity to 2,500 dams that currently have none (UNWWAP, 2006).

5.4
5.4.1

Global and regional status of market and industry development


Existing generation

In 2008, the generation of electricity from hydroelectric plants was 3,288 TWh (11.8 EJ)11 compared to 1,295 TWh (4.7 EJ) in 1973 (IEA, 2010a), which represented an increase of roughly 25% in this period, and was mainly a result of increased production in China and Latin America, which reached 585 TWh (2.1 EJ) and 674 TWh (2.5 EJ), respectively (Figures 5.11 and 5.12). Hydropower provides some level of power generation in 159 countries. Five countries make up more than half of the worlds hydropower production: China, Canada, Brazil, the USA and Russia. The

1973

2008

Africa 2.2%

Asia 7.2%

Latin America 7.2%

Asia 25.5%

Latin America 20.5%

Non-OECD Europe 2.1%

Former USSR 9.4%

Non-OECD Europe 1.5%

Africa 3% Former USSR 7.3%

Middle East 0.3%

Middle East 0.3% OECD 72% OECD 41.9%

1,295 TWh (4,662 EJ)

3,288 TWh (11,836 EJ)

Figure 5.11 | 1973 and 2008 regional shares of hydropower production (IEA, 2010a).

11 These gures differ slightly from those presented in Chapter 1.

455

Hydropower

Chapter 5

Rest of World; 1007; 31% Sweden; 69; 2% Venezuela; 87; 3% PR of China; 585; 18% Japan; 83; 2% India; 114; 3% Norway; 141; 4% Canada; 383; 12% Brasil; 370; 11% United States; 282; 9%

possibility of drawing on the hydroelectric resource was important for the introduction and consolidation of the main electro-intensive sectors on which the industrialization process in these countries was based during a considerable part of the 20th century. Despite the signicant growth in hydroelectric production, the percentage share of hydroelectricity on a global basis has dropped during the last three decades (1973 to 2008), from 21 to 16%. This is because electricity demand and the deployment of other energy technologies have increased more rapidly than hydropower generating capacity.

Russia; 167; 5%

5.4.2

The hydropower industry

Figure 5.12 | Hydropower generation in 2008 by country, indicating total generation (TWh) and respective global share (IEA, 2010a).

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

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

Table 5.4 | Major hydroelectricity producer countries with total installed capacity and percentage of hydropower generation in the electricity mix. Source: IJHD (2010). Country
China Brazil USA Canada Russia India Norway Japan France Italy Rest of the world

Installed Capacity (GW)


200 84 78.2 74.4 49.5 38 29.6 27.5 21 20 301.6

Country Based on Top 10 Producers


Norway Brazil Venezuela Canada Sweden Russia India China Italy France Rest of the world1

Percent of Hydropower in Total Domestic Electricity Generation (%)


99 83.9 73.4 59.0 48.8 19.0 17.5 15.5 14.0 8.0 14.3

World Note: 1. Excluding countries with no hydropower production.

926.1

World

15.9

456

Chapter 5

Hydropower

5.4.3

Impact of policies12

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

investment capital, and international carbon markets offer one possible route to that capital. Out of the 2,062 projects registered by the CDM Executive Board (EB) by 1 March 2010, 562 were hydropower projects. When considering the predicted volumes of Certied Emission Reductions to be delivered, registered hydropower projects are expected to avoid more than 50 Mt of carbon dioxide (CO2) emissions per year by 2012. China, India, Brazil and Mexico represent roughly 75% of the hosted projects.

5.4.3.2

Project nancing

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

5.4.3.3

Administrative and licensing process

5.4.3.1

International carbon markets

As with other carbon reduction technologies, carbon credits can benet hydropower projects by bringing additional funding and thus helping to reduce project risk and thereby secure nancing. Though the Clean Development Mechanism (CDM) is not unique to hydropower, hydropower projects are one of the largest contributors to the CDM and Joint Implementation (JI) mechanisms and therefore to existing carbon credit markets. In part, this is due to the fact that new hydropower development is targeted towards developing countries that are in need of
12 Non-technology-specic policy issues are covered in Chapter 11 of this report.

Hydropower is often regarded as a public resource (Sternberg, 2008), emphasized by the operating life of a reservoir that may be more than 100 years. Legal frameworks vary from country to country, however, including practices in the award and structuring of concessions, for instance, regarding concession periods, royalties, water rights etc. Environmental licensing procedures also vary greatly. With growing involvement of the private sector in what was previously managed by public sector, contractual arrangements surrounding hydropower have become increasingly complex. There are now more parties involved and much greater commercial accountability, with a strong awareness of environmental and social indicators and licensing processes. Clearly, the policies and procedures established by governments in granting licenses and concessions will impact hydropower development outcomes.

457

Hydropower

Chapter 5

5.4.3.4

Classication by size

Finally, many governments and international bodies have relied upon various distinctions between small and large hydro, as dened by installed capacity (MW), in establishing the eligibility of hydropower plants for certain programs. While it is well known that large-scale HPPs can create conicts and concerns (WCD, 2000), the environmental and social impacts of a HPP cannot be deduced by size in itself, even if increasing the physical size may increase the overall impacts of a specic HPP (Egr and Milewski, 2002; Sternberg, 2008). Despite their lack of robustness (see Section 5.3.1), these classications have had signicant policy and nancing consequences (Egr and Milewski, 2002). In the UK Renewables Obligation,13 eligible hydropower plants must be below 20 MW in size. Likewise, in several countries, feed-in tariffs are targeted only towards smaller projects. For example, in France, only projects with an installed capacity not exceeding 12MW are eligible,14 and in Germany, a 5 MW maximum capacity has been established.15 In India, projects below 5 and 25 MW in capacity obtain promotional support that is unavailable to projects of larger sizes. Similar approaches exist in many developed and developing countries around the world, for example, in Indonesia.16 Because project size is neither a perfect indicator of environmental and social impact nor of the nancial need of a project for addition policy support, these categorizations may, at times, impede the development of socially benecial projects. Similar concerns have been raised with respect to international and regional climate policy. Though hydropower is recognized as a contributor to reducing GHG emissions and is included in the Kyoto Protocols exible mechanisms, those mechanisms differentiate HPPs depending on size and type. The United Nations Framework Convention on Climate Change (UNFCCC) CDM EB, for example, has established that storage hydropower projects are to follow the power density indicator (PDI), W/m2 (installed capacity/reservoir area), to be eligible for CDM credits. The PDI indicates tentative GHG emissions from reservoirs. The CDM Executive Board stated (February 2006) that Hydroelectric power plants with power densities greater than 4 W/m2 but less than or equal to 10 W/m2 can use the currently approved methodologies, with an emission factor of 90 g CO2eq/kWh for projects with reservoir emissions, while less than or equal to 4 W/m2 cannot use current methodologies. There is little link, however, between installed capacity, the area of a reservoir and the various biogeochemical processes active in a reservoir. Hypothetically, two identical storage HPPs would, according to the PDI, have the same emissions independent of climate zones or of inundated
13 The Renewables Obligation Order 2006, No. 1004 (ROO 2006): www.statutelaw. gov.uk. 14 Dcret n2000-1196, Decree on capacity limits for different categories of systems for the generation of electricity from renewable sources that are eligible for the feed-in tariff: www.legifrance.gouv.fr. 15 EEG, 2009 - Act on Granting Priority to Renewable Energy and Mineral Sources: bundesrecht.juris.de/eeg_2009/. 16 Regulation of the Minister of Energy and Mineral Resources, No.31, 2009.

biomass and carbon uxes (see Section 5.6.3). As such, the PDI rule may inadvertently impede the development of socially benecial hydropower projects, while at the same time supporting less benecial projects. The European Emission Trading Scheme and related trading markets similarly treat small- and large-scale hydropower stations differently.17

5.5

Integration into broader energy systems

Hydropowers large capacity range, exibility, storage capability when coupled with a reservoir, and ability to operate in a stand-alone mode or in grids of all sizes enables hydropower to deliver a broad range of services. Hydropowers various roles in and services to the energy system are discussed below (see also Chapter 8).

5.5.1

Grid-independent applications

Hydropower can be delivered through national and regional interconnected electric grids, through local mini-grids and isolated grids, and can also serve individual customers through captive plants. Water mills in England, Nepal, India and elsewhere, which are used for grinding cereals, for lifting water and for powering machinery, are early testimonies of hydropower being used as captive power in mechanical and electrical form. The tea and coffee plantation industries as well as small islands and developing states have used and still make use of hydropower to meet energy needs in isolated areas. Captive power plants (CPPs) are dened here as plants set up by any person or group of persons to generate electricity primarily for the person or the groups members (Indian Electricity Act, 2003). CPPs are often found in decentralized isolated systems and are generally built by private interests for their own electricity needs. In deregulated electricity markets that allow open access to the grid, hydropower plants are also sometimes installedfor captive purposes by energy-intensive industries such as aluminium smelters, pulp and paper mills, mines and cement factories in order to weather short-term market uncertainties and volatility (Shukla et al., 2004). For governments of emerging economies such as India facing shortages of electricity, CPPs are also a means to cope with unreliable power supply systems and higher industrial tariffs by encouraging decentralized generation and private participation (Shukla et al., 2004).

5.5.2

Rural electrication

According to the International Energy Agency (IEA, 2010c), 1.4 billion people have no access to electricity (see Section 9.3.2). Related to the discussion in Section 5.5.1, small-scale hydropower (SHP) can sometimes be an economically viable supply source in these circumstances, as SHP can provide a decentralized electricity supply in those rural areas
17 Directive 2004/101/E, C article 11a(6), www.eur-lex.europa.eu.

458

Chapter 5

Hydropower

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

hydropower, on the other hand, can largely decouple the timing of hydropower generation and variable river ows. For large storage reservoirs, the storage may be sufcient to buffer seasonal or multi-seasonal changes in river ows, whereas for smaller reservoirs the storage may buffer river ows on a daily or weekly basis. With a very large reservoir relative to the size of the hydropower plant (or very consistent river ows), HPPs can generate power at a nearconstant level throughout the year (i.e., operate as a base-load plant). Alternatively, in the case that the hydropower capacity far exceeds the amount of reservoir storage, the hydropower plant is sometimes referred to as energy-limited. An energy-limited hydropower plant would exhaust its fuel supply by consistently operating at its rated capacity throughout the year. In this case, the use of reservoir storage allows hydropower generation to occur at times that are most valuable from the perspective of the power system rather than at times dictated solely by river ows. Since electrical demand varies during the day and night, during the week and seasonally, storage hydropower generation can be timed to coincide with times where the power system needs are the greatest. In part, these times will occur during periods of peak electrical demand. Operating hydropower plants in a way that generates power during times of high demand is referred to as peaking operation (in contrast to base-load). Even with storage, however, hydropower generation will still be limited by the size of the storage, the rated electrical capacity of the hydropower plant, and downstream ow constraints for irrigation, recreation or environmental uses of the river ows. Hydropower peaking may, if the outlet is directed to a river, lead to rapid uctuations in river ow, water-covered area, depth and velocity. In turn this may, depending on local conditions, lead to negative impacts in the river (see Section 5.6.1.5) unless properly managed. Hydropower generation that consistently occurs during periods with high system demand can offset the need for thermal generation to meet that same demand. The ratio of the amount of demand that can be reliably met by adding hydropower to the nameplate capacity of the hydropower plant is called the capacity credit. Even RoR hydropower that consistently has river ows during periods of high demand can earn a high capacity credit, while adding reservoir storage can increase the capacity credit to levels comparable to thermal power plants (see Section 8.2.1.2). In addition to providing energy and capacity to meet electrical demand, hydropower generation often has several characteristics that enable it to provide other services to reliably operate power systems. Because hydropower plants utilize gravity instead of combustion to generate electricity, hydropower plants are often less susceptible to the sudden loss of generation than is thermal generation. Hydropower plants also offer operating exibility in that they can start generating electricity with very short notice and low start-up costs, provide rapid changes in generation, and have a wide range of generation levels over which power can be generated efciently (i.e., high part-load efciency) (Haldane and Blackstone, 1955; Altinbilek et al., 2007). The ability to rapidly change

5.5.3

Power system services provided by hydropower

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

459

Hydropower

Chapter 5

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

hydropower being a net energy consumer. A traditional storage HPP may also be retrotted with pump technologies to combine the properties of storage and pump storage HPPs (SRU, 2010). The use and benet of pumped storage hydropower in the power system will depend on the overall mix of existing generating plants and the architecture of the transmission system. Pumped storage represents about 2.2% of all generation capacity in the USA, 10.2 % in Japan and 18.7 % in Austria (Deane et al., 2010). Various technologies for storing electricity in the grid are compared by Vennemann et al. (2010) in Figure 5.13 for selected large storage sites in different parts of the world. In addition to hydropower supporting fossil and nuclear generation technologies, hydropower can also help reduce the challenges of integrating variable renewable resources. In Denmark, for example, the high level of variable wind (>20% of the annual energy demand) is managed in part through strong interconnections (1 GW) to Norway, where there is substantial storage hydropower (Nordel, 2008). More interconnectors to Europe may further support increasing the share of wind power in Denmark and Germany (SRU, 2010; see also Section 11.6.5). From a technical viewpoint, Norway alone has a long-term potential to establish pumped storage facilities in the 10 to 25 GW range, enabling energy storage over periods from hours to several weeks in existing reservoirs, and more or less doubling the present installed capacity of 29 GW (IEAENARD, 2010). Increasing variable generation will also increase the amount of balancing services, including regulation and load following, required by the power system (e.g., Holttinen et al., 2009). In regions with new and existing hydropower facilities, providing these services with hydropower may avoid the need to rely on increased part-load and cycling of thermal plants to provide these services. Similarly, in systems with high shares of variable renewable resources that provide substantial amounts of energy but limited capacity, the potential for a high capacity credit of hydropower can be used to meet peak demand rather than requiring peaking thermal plants.

5.5.4

Hydropower support of other generation including renewable energy

Electricity systems worldwide rely upon widely varying amounts of hydropower today. In this range of hydropower capabilities, electric system operators have developed economic dispatch methodologies that take into account the unique role of hydropower, including coordinating the operation of hydropower plants with other types of generating units. In particular, many thermal power plants (coal, gas or liquid fuel, or nuclear energy) require considerable lead times (often four hours for gas turbines and over eight hours for steam turbines) before they attain an optimum thermal efciency at which point fuel consumption and emissions per unit output are minimum. In an integrated system, the considerable exibility provided by storage HPPs can be used to reduce the frequency of start ups and shut downs of thermal plants; to maintain a balance between supply and demand under changing demand or supply patterns and thereby reduce the load-following burden on thermal plants; and to increase the amount of time that thermal units are operated at their maximum thermal efciency. In some regions, for instance, hydroelectric power plants are used to follow varying peak load demands while nuclear or fossil fuel power plants are operated as base-load units. Pumped hydropower storage can further increase the support of other resources. In cases with pumped hydropower storage, pumps can use the output from thermal plants during times that they would otherwise operate less efciently at part load or be shut down (i.e., low load periods). The pumped storage plant then keeps water in reserve for generating power during peak period demands. Pumped storage has much the same ability as storage HPPs to provide balancing and regulation services. Pumped storage hydropower is usually not a source for energy, however. The hydraulic, mechanical and electrical efciencies of pumped storage determine the overall cycle efciency, ranging from 65 to 80% (Egr and Milewski, 2002). If the upstream pumping reservoir is also used as a traditional reservoir the inow from the watershed may balance out the energy loss caused by pumping. If not, net losses lead to pumped

5.5.5

Reliability and interconnection needs for hydropower

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

460

Chapter 5

Hydropower

Storage Capacity per Cycle [MWh]

1,000,000

Pumped Storage Plants (PSP) 100,000


PSP, Limberg I, II, III, AT, 2018 PSP, Linth Limmern, CH, 2015 PSP, Bath County, US, 1985

10,000

Compressed Air/Thermal

PSP, Goldisthal, DE, 2003 CAES, McIntosh, US, 1991 AA-CAES, RWE/GE, Project

PSP, Atdorf, DE, 2018 PSP, Dinorwig, GB, 1984 PSP, Vianden, LU, 1964

1 month

1,000

Batteries

Molten Salt, Andasol I, ES, 2008

7d

100

3d 32 h 16 h 8h

NaS, Rokkasho, JP, 2008 Lead Acid, Chino, US, 1988

10
4h 2h 1h

VRB, Tomamae, JP, 2000 NiCd, Fairbanks, US, 2003

1 1 10 100 1,000 10,000

Generation Power [MW]


Figure 5.13 | Storage and installed capacity of selected large electricity storage sites (Vennemann et al., 2010). Note: PSP = Pumped storage plants; CAES = compressed air energy storage, AA-CAES = advanced adiabatic compressed air energy storage; Batteries: NaS = sodium-sulphur, NiCd = nickel cadmium, VRB = vanadium redox battery.

that adequate total annual energy demand can be met. Strong interconnections between diverse hydropower resources or between hydro-dominated and thermal-dominated power systems have been used in existing systems to ensure adequate energy generation (see Section 8.2.1.3). In the future, interconnection to other renewable resources could also ensure adequate energy. Wind and direct solar power, for instance, can be used to reduce demands on hydropower, either by allowing dams to save their water for later release in peak periods or letting storage or pumped storage HPPs consume excess energy produced in off-peak hours.

environmental side, hydropower may have a signicant environmental footprint at local and regional levels but offers advantages at the macroecological level. With respect to social impacts, hydropower projects may entail the relocation of communities living within or nearby the reservoir or the construction sites, compensation for downstream communities, public health issues etc. A properly designed hydropower project may, however, be a driving force for socioeconomic development (see Box 5.1), though a critical question remains about how these benets are shared. Because each hydropower plant is uniquely designed to t the sitespecic characteristics of a given geographical site and the surrounding society and environment, the magnitude of environmental and social impacts as well as the extent of their positive and negative effects is highly site dependent. Though the size of a HPP is not, alone, a relevant criterion to predict environmental performance, many impacts are related to the impoundment and existence of a reservoir, and therefore do not apply to all HPP types (see Table 5.5). Section 5.6.1 summarizes

5.6

Environmental and social impacts19

Like all energy and water management options, hydropower projects have negative and positive environmental and social impacts. On the
19 A comprehensive assessment of social and environmental impacts of all RE sources covered in this report can be found in Chapter 9.

Under Construction/ Project

CAES, Huntorf, D, 1978


In Operation

Pumped Storage Compressed Air Batteries Thermal Storage

461

Chapter 7

Wind Energy
Coordinating Lead Authors:
Ryan Wiser (USA), Zhenbin Yang (China)

Lead Authors:
Maureen Hand (USA), Olav Hohmeyer (Germany), David Ineld (United Kingdom), Peter H. Jensen (Denmark), Vladimir Nikolaev (Russia), Mark OMalley (Ireland), Graham Sinden (United Kingdom/Australia), Arthouros Zervos (Greece)

Contributing Authors:
Nam Darghouth (USA), Dennis Elliott (USA), Garvin Heath (USA), Ben Hoen (USA), Hannele Holttinen (Finland), Jason Jonkman (USA), Andrew Mills (USA), Patrick Moriarty (USA), Sara Pryor (USA), Scott Schreck (USA), Charles Smith (USA)

Review Editors:
Christian Kjaer (Belgium/Denmark) and Fatemeh Rahimzadeh (Iran)

This chapter should be cited as: Wiser, R., Z. Yang, M. Hand, O. Hohmeyer, D. Ineld, P. H. Jensen, V. Nikolaev, M. OMalley, G. Sinden, A. Zervos, 2011: Wind Energy. In IPCC Special Report on Renewable Energy Sources and Climate Change Mitigation [O. Edenhofer, R. Pichs-Madruga, Y. Sokona, K. Seyboth, P. Matschoss, S. Kadner, T. Zwickel, P. Eickemeier, G. Hansen, S. Schlmer, C. von Stechow (eds)], Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA.

535

Wind Energy

Chapter 7

loading for offshore plants, and changes in sea ice and/or permafrost conditions may also inuence access for performing wind power plant O&M (Laakso et al., 2003). One study focusing on northern Europe found substantial declines in sea ice under reasonable climate change scenarios (Claussen et al., 2007). Other meteorological drivers of turbine loading may also be inuenced by climate change but are likely to be secondary in comparison to changes in resource magnitude, weather extremes, and icing issues (Pryor and Barthelmie, 2010). Additional research on the possible impact of climate change on the size, geographic distribution and variability of the wind resource is warranted, as is research on the possible impact of climate change on extreme weather events and therefore wind turbine operating environments. Overall, however, research to date suggests that these impacts are unlikely to be of a magnitude that will greatly impact the global potential of wind energy deployment.

Power in Wind Rated Power

Power Captured

Rotor RPM Power

Wind Speed Cut-In Speed Rated Speed Cut-Out Speed

Region I

Region II

Region III

7.3

Technology and applications

Figure 7.3 | Conceptual power curve for a modern variable-speed wind turbine (US DOE, 2008).

Modern, commercial grid-connected wind turbines have evolved from small, simple machines to large, highly sophisticated devices. Scientic and engineering expertise and advances, as well as improved computational tools, design standards, manufacturing methods, and O&M procedures, have all supported these technology developments. As a result, typical wind turbine nameplate capacity ratings have increased dramatically since the 1980s (from roughly 75 kW to 1.5 MW and larger), while the cost of wind energy has substantially declined. Onshore wind energy technology is already being manufactured and deployed on a commercial basis. Nonetheless, additional R&D advances are anticipated, and are expected to further reduce the cost of wind energy while enhancing system and component performance and reliability. Offshore wind energy technology is still developing, with greater opportunities for additional advancement. This section summarizes the historical development and current technology status of large grid-connected on- and offshore wind turbines (7.3.1), discusses international wind energy technology standards (7.3.2), and reviews power conversion and related grid connection issues (7.3.3); a later section (7.7) describes opportunities for further technical advances.

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

7.3.1 7.3.1.1

Technology development and status Basic design principles

Generating electricity from the wind requires that the kinetic energy of moving air be converted to mechanical and then electrical energy, thus the engineering challenge for the wind energy industry is to design costeffective wind turbines and power plants to perform this conversion. The amount of kinetic energy in the wind that is theoretically available for extraction increases with the cube of wind speed. However, a turbine only captures a portion of that available energy (see Figure 7.3).

550

Chapter 7

Wind Energy

conditions has become common as wind turbines, wind power plants and the wind energy market have all increased in size; site-specic conditions that can impact turbine and plant design include geographic and temporal variations in wind speed, site topography and access, interactions among individual wind turbines due to wake effects, and integration into the larger electricity system (Burton et al., 2001). Wind turbine and power plant design also impacts and is impacted by noise, visual, environmental and public acceptance issues (see Section 7.6).

7.3.1.2

Onshore wind energy technology

In the 1970s and 1980s, a variety of onshore wind turbine congurations were investigated, including both horizontal and vertical axis designs (see Figure 7.4). Gradually, the horizontal axis design came to dominate, although congurations varied, in particular the number of blades and whether those blades were oriented upwind or downwind of the tower (EWEA, 2009). After a period of further consolidation, turbine designs largely centred (with some notable exceptions) around the three-blade, upwind rotor; locating the turbine blades upwind of the tower prevents the tower from blocking wind ow onto the blades and producing extra aerodynamic noise and loading, while three-bladed machines typically have lower noise emissions than two-bladed machines. The three blades are attached to a hub and main shaft, from which power is transferred (sometimes through a gearbox, depending on design) to a generator. The main shaft and main bearings, gearbox, generator and control system are contained within a housing called the nacelle. Figure 7.5 shows the components in a modern wind turbine with a gearbox; in wind turbines without a gearbox, the rotor is mounted directly on the generator shaft.

In the 1980s, larger machines were rated at around 100 kW and primarily relied on aerodynamic blade stall to control power production from the xed blades. These turbines generally operated at one or two rotational speeds. As turbine size increased over time, development went from stall control to full-span pitch control in which turbine output is controlled by pitching (i.e., rotating) the blades along their long axis (EWEA, 2009). In addition, a reduction in the cost of power electronics allowed variable speed wind turbine operation. Initially, variable speeds were used to smooth out the torque uctuations in the drive train caused by wind turbulence and to allow more efcient operation in variable and gusty winds. More recently, almost all electric system operators require the continued operation of large wind power plants during electrical faults, together with being able to provide reactive power: these requirements have accelerated the adoption of variable-speed operation with power electronic conversion (see Section 7.3.3 for a summary of power conversion technologies, Section 7.5 for a fuller discussion of electric system integration issues, and Chapter 8 for a discussion of reactive power and broader issues with respect to the integration of RE into electricity systems). Modern wind turbines typically operate at variable speeds using full-span blade pitch control. Blades are commonly constructed with composite materials, and towers are usually tubular steel structures that taper from the base to the nacelle at the top (EWEA, 2009). Over the past 30 years, average wind turbine size has grown signicantly (Figure 7.6), with the largest fraction of onshore wind turbines installed globally in 2009 having a rated capacity of 1.5 to 2.5 MW; the average size of turbines installed in 2009 was 1.6 MW (BTM, 2010). As of 2010, wind turbines used onshore typically stand on 50- to 100-m towers, with rotors that are often 50 to 100 m in diameter; commercial

Horizontal-Axis Turbines

Vertical-Axis Turbines

Figure 7.4 | Early wind turbine designs, including horizontal and vertical axis turbines (South et al., 1983).

551

Wind Energy

Chapter 7

Rotor Blade Wind Direction D ire ectio io on on For F or an an Upwind Rotor

Wind Direction Nacelle with Gearbox and Generator Low-Speed Shaft

Rotor Diameter

Swept Area of Blades

Pitch

Gear Box Generator

Hub

Rotor

Controller

Anemometer

Wind Direction Brake Hub Height Yaw Drive Blades Yaw Motor Tower Transformer Tower High-Speed Shaft Nacelle Wind Vane

Figure 7.5 | Basic components of a modern, horizontal-axis wind turbine with a gearbox (Design by the National Renewable Energy Laboratory (NREL)).

machines with rotor diameters and tower heights in excess of 125 m are operating, and even larger machines are under development. Modern turbines operate with rotational speeds ranging from 12 to 20 revolutions per minute (RPM), which compares to the faster and potentially more visually disruptive speeds exceeding 60 RPM common of the smaller turbines installed during the 1980s.13 Onshore wind turbines are typically grouped together into wind power plants, sometimes also called wind projects or wind farms. These wind power plants are often 5 to 300 MW in size, though smaller and larger plants do exist. The main reason for the continual increase in turbine size to this point has been to minimize the levelized generation cost of wind energy
13 Rotational speed decreases with larger rotor diameters. The acoustic noise resulting from tip speeds greater than 70 to 80 m/s is the primary design criterion that governs rotor speed.

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

552

Chapter 7

Wind Energy

Hub Height (m)

320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 1980- 19901990 1995 19952000 20002005 20052010 2010-? 2010-? Future
17m 75kW 30m 300kW Rotor Diameter (m) Rating (kW) 50m 750kW

250m 20,000kW

Past and Present Wind Turbines


150m 10,000kW 125m 5,000kW 100m 3,000kW 80m 1,800kW

Future Wind Turbines

70m 1,500kW

Future

Figure 7.6 | Growth in size of typical commercial wind turbines (Design by NREL).

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

technology transfer in wind turbine design, component manufacturing and wind power plant siting exist. Extreme environmental conditions, such as icing or typhoons, may be more prominent in some of these markets, providing impetus for continuing research. Other aspects unique to less-developed countries, such as minimal transportation infrastructure, could also inuence wind turbine designs if and as these markets grow.

7.3.1.3

Offshore wind energy technology

The rst offshore wind power plant was built in 1991 in Denmark, consisting of eleven 450 kW wind turbines. Offshore wind energy technology is less mature than onshore, and has higher investment and O&M costs (see Section 7.8). By the end of 2009, just 1.3% of global installed wind power capacity was installed offshore, totalling 2,100 MW (GWEC, 2010a). The primary motivation to develop offshore wind energy is to provide access to additional wind resources in areas where onshore wind energy development is constrained by limited technical potential and/or by planning and siting conicts with other land uses. Other motivations for developing offshore wind energy include: the higher-quality wind resources located at sea (e.g., higher average wind speeds and lower shear near hub height; wind shear refers to the general increase in wind speed with height); the ability to use even larger wind turbines due to avoidance of certain land-based transportation constraints and the potential to thereby gain additional economies of scale; the ability to build larger power plants than onshore, gaining plant-level economies of scale; and a potential reduction in the need for new, long-distance, land-based transmission infrastructure

553

Wind Energy

Chapter 7

to access distant onshore wind energy15 (Carbon Trust, 2008b; Snyder and Kaiser, 2009b; Twidell and Gaudiosi, 2009). These factors, combined with a signicant offshore wind resource potential, have created considerable interest in offshore wind energy technology in the EU and, increasingly, in other regions, despite the typically higher costs relative to onshore wind energy. Offshore wind turbines are typically larger than onshore, with nameplate capacity ratings of 2 to 5 MW being common for offshore wind power plants built from 2007 to 2009, and even larger turbines are under development. Offshore wind power plants installed from 2007 to 2009 were typically 20 to 120 MW in size, with a clear trend towards larger turbines and power plants over time. Water depths for most offshore wind turbines installed through 2005 were less than 10 m, but from 2006 to 2009, water depths from 10 to more than 20 m were common. Distance to shore has most often been below 20 km, but average distance has increased over time (EWEA, 2010a). As experience is gained, water depths are expected to increase further and more exposed locations with higher winds will be utilized. These trends will impact the wind resource characteristics faced by offshore wind power plants, as well as support structure design and the cost of offshore wind energy. A continued transition towards larger wind turbines (5 to 10 MW, or even larger) and wind power plants is also anticipated as a way of reducing the cost of offshore wind energy through turbine- and plant-level economies of scale. To date, offshore turbine technology has been very similar to onshore designs, with some modications and with special foundations (Musial, 2007; Carbon Trust, 2008b). The mono-pile foundation is the most common, though concrete gravity-based foundations have also been used with some frequency; a variety of other foundation designs (including oating designs) are being considered and in some instances used (Breton and Moe, 2009), especially as water depths increase, as discussed in Section 7.7. In addition to differences in foundations, modication to offshore turbines (relative to onshore) include structural upgrades to the tower to address wave loading; air conditioned and pressurized nacelles and other controls to prevent the effects of corrosive sea air from degrading turbine equipment; and personnel access platforms to facilitate maintenance. Additional design changes for marine navigational safety (e.g., warning lights, fog signals) and to minimize expensive servicing (e.g., more extensive condition monitoring, onboard service cranes) are common. Wind turbine tip speed could be chosen to be greater than for onshore turbines because concerns about noise are reduced for offshore power plantshigher tip speeds can sometimes lead to lower torque and lighter drive train components for the same power output. In addition, tower heights are sometimes
15 Of course, transmission infrastructure is needed to connect offshore wind power plants with electricity demand centres, and the per-kilometre cost of offshore transmission typically exceeds that for onshore lines. Whether offshore transmission needs are more or less extensive than those needed to access onshore wind energy varies by location.

lower than used for onshore wind power plants due to reduced wind shear offshore relative to onshore. Lower power plant availabilities and higher O&M costs have been common for offshore wind energy relative to onshore wind both because of the comparatively less mature state of offshore wind energy technology and because of the inherently greater logistical challenges of maintaining and servicing offshore turbines (Carbon Trust, 2008b; UKERC, 2010). Wind energy technology specically tailored for offshore applications will become more prevalent as the offshore market expands, and it is expected that larger turbines in the 5 to 10 MW range may come to dominate this market segment (EU, 2008). Future technical advancement possibilities for offshore wind energy are described in Section 7.7.

7.3.2

International wind energy technology standards

Wind turbines in the 1970s and 1980s were designed using simplied design models, which in some cases led to machine failures and in other cases resulted in design conservatism. The need to address both of these issues, combined with advances in computer processing power, motivated designers to improve their calculations during the 1990s (Quarton, 1998; Rasmussen et al., 2003). Improved design and testing methods have been codied in International Electrotechnical Commission (IEC) standards, and the rules and procedures for Conformity Testing and Certication of Wind Turbines (IEC, 2010) relies upon these standards. Certication agencies rely on accredited design and testing bodies to provide traceable documentation of the execution of rules and specications outlined in the standards in order to certify turbines, components or entire wind power plants. The certication system assures that a wind turbine design or wind turbines installed in a given location meet common guidelines relating to safety, reliability, performance and testing. Figure 7.7(a) illustrates the design and testing procedures required to obtain a wind-turbine type certication. Plant certication, shown in Figure 7.7(b), requires a type certicate for the turbine and includes procedures for evaluating site conditions and turbine design parameters associated with that specic site, as well as other site-specic conditions including soil properties, installation and plant commissioning. Insurance companies, nancing institutions and power plant owners normally require some form of certication for plants to proceed, and the IEC standards therefore provide a common basis for certication to reduce uncertainty and increase the quality of wind turbine products available in the market (EWEA, 2009). In emerging markets, the lack of highly qualied testing laboratories and certication bodies limits the opportunities for manufacturers to obtain certication according to IEC standards and may lead to lower-quality products. As markets mature and design margins are compressed to reduce costs, reliance on internationally recognized standards is likely to become even more widespread to assure consistent performance, safety and reliability of wind turbines.

554

Chapter 7

Wind Energy

(a) Wind Turbine Type Certication Procedure

(b) Wind Project Certication Procedure

Design Basis Evaluation

Site Conditions Assessment Type Certicate Design Basis Evaluation

Design Evaluation

Foundation Design Evaluation

Integrated Load Analysis

Manufacturing Evaluation

Foundation Manufacturing Evaluation

Wind Turbine/RNA Design Evaluation

Support Structure Design Evaluation

Other Installations Design Evaluation

Wind Turbine/RNA Manuf. Surveillance

Support Structure Manuf. Surveillance

Other Installations Manuf. Surveillance

Type Testing Transportation & Install Surveillance Type Characteristics Measurements

Commissioning Surveillance

Project Characteristics Measurements Final Evaluation Optional Module Final Evaluation Optional Module Type Certicate Project Certicate Operational & Maintainance Surveillance

Figure 7.7 | Modules for (a) turbine type certication and (b) wind power plant certication (IEC, 2010). Notes: RNA refers to Rotor Nacelle Assembly. The authors thank the IEC for permission to reproduce information from its International Standard IEC 61400-22 ed.1.0 (2010). All such extracts are copyright of IEC, Geneva, Switzerland. All rights reserved. Further information on the IEC is available from www.iec.ch. IEC has no responsibility for the placement and context in which the extracts and contents are reproduced by the authors, nor is IEC in any way responsible for the other content or accuracy therein. Copyright 2010 IEC Geneva, Switzerland, www.iec.ch.

7.3.3

Power conversion and related grid connection issues

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

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

555

Wind Energy

Chapter 7

output in synchronism with system power imbalances. This lack of inertial response is an important consideration for electric system planners because less overall inertia in the electric system makes the maintenance of stable system operation more challenging (Gautam et al., 2009). Wind turbine manufacturers have recognized this lack of intrinsic inertial response as a possible long-term impediment to wind energy and are actively pursuing a variety of solutions; for example, additional turbine controls can be added to provide inertial response (Mullane and OMalley, 2005; Morren et al., 2006).

7.4

Global and regional status of market and industry development

The majority of the capacity has been installed onshore, with offshore installations constituting a small proportion of the total market. About 2.1 GW of offshore wind turbines were installed by the end of 2009; 0.6 GW were installed in 2009, including the rst commercial offshore wind power plant outside of Europe, in China (GWEC, 2010a). Many of these offshore installations have taken place in the UK and Denmark. Signicant offshore wind power plant development activity, however, also exists in, at a minimum, other EU countries, the USA, Canada and China (e.g., Mostafaeipour, 2010). Offshore wind energy is expected to develop in a more signicant way in the years ahead as the technology advances and as onshore wind energy sites become constrained by local resource availability and/or siting challenges in some regions (BTM, 2010; GWEC, 2010a). The total investment cost of new wind power plants installed in 2009 was USD2005 57 billion (GWEC, 2010a). Direct employment in the wind energy sector in 2009 has been estimated at roughly 190,000 in the EU and 85,000 in the USA. Worldwide, direct employment has been estimated at approximately 500,000 (GWEC, 2010a; REN21, 2010). Despite these trends, wind energy remains a relatively small fraction of worldwide electricity supply. The total wind power capacity installed by the end of 2009 would, in an average year, meet roughly 1.8% of worldwide electricity demand, up from 1.5% by the end of 2008, 1.2% by the end of 2007, and 0.9% by the end of 2006 (Wiser and Bolinger, 2010).

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

7.4.2

Regional and national status and trends

7.4.1

Global status and trends

Wind energy has quickly established itself as part of the mainstream electricity industry. From a cumulative capacity of 14 GW at the end of 1999, global installed wind power capacity increased 12-fold in 10 years to reach almost 160 GW by the end of 2009, an average annual increase in cumulative capacity of 28% (see Figure 7.8). Global annual wind power capacity additions equalled more than 38 GW in 2009, up from 26 GW in 2008 and 20 GW in 2007 (GWEC, 2010a).

The countries with the highest total installed wind power capacity by the end of 2009 were the USA (35 GW), China (26 GW), Germany (26 GW), Spain (19 GW) and India (11 GW). After its initial start in the USA in the 1980s, wind energy growth centred on countries in the EU and India during the 1990s and the early 2000s. In the late 2000s, however, the USA and then China became the locations for the greatest annual capacity additions (Figure 7.9). Regionally, Europe continues to lead the market with 76 GW of cumulative installed wind power capacity by the end of 2009, representing 48% of the global total (Asia represented 25%, whereas North America

Annual Capacity Additions [GW]

50 40 30 20 10 0

Annual Wind Power Installations (Left Axis) Cumulative Wind Power Capacity (Right Axis)

150 120 90 60 30 0

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 7.8 | Global annual wind power capacity additions and cumulative capacity (Data sources: GWEC, 2010a; Wiser and Bolinger, 2010).

556

Cumulative Capacity [GW]

60

180

Chapter 7

Wind Energy

Cumulative Wind Power Capacity [GW]

40 35 30 25 20 15 10 5 0 US China Germany Spain India Italy France UK Portugal Denmark


2006 2007 2008 2009

Figure 7.9 | Top-10 countries in cumulative wind power capacity (Date source: GWEC, 2010a).

represented 24%). Notwithstanding the continuing growth in Europe, the trend over time has been for the wind energy industry to become less reliant on a few key markets, and other regions of the world have increasingly become the dominant markets for wind energy growth. The annual growth in the European wind energy market in 2009, for example, accounted for just 28% of the total new wind power additions in that year, down from over 60% in the early 2000s (GWEC, 2010a). More than 70% of the annual wind power capacity additions in 2009 occurred outside of Europe, with particularly signicant growth in Asia (40%) and North America (29%) (Figure 7.10). Even in Europe, though Germany and Spain have been the strongest markets during the 2000s, there is a trend towards less reliance on these two countries. Despite the increased globalization of wind power capacity additions, the market remains concentrated regionally. As shown in Figure 7.10, Latin America, Africa and the Middle East, and the Pacic regions have installed

relatively little wind power capacity despite signicant technical potential in each region, as presented earlier in Section 7.2. And, even in the regions of signicant growth, most of that growth has occurred in a limited number of countries. In 2009, for example, 90% of wind power capacity additions occurred in the 10 largest markets, and 62% was concentrated in just two countries: China (14 GW, 36%) and the USA (10 GW, 26%). In both Europe and the USA, wind energy represents a major new source of electric capacity additions. From 2000 through 2009, wind energy was the second-largest new resource added in the USA (10% of all gross capacity additions) and EU (33% of all gross capacity additions) in terms of nameplate capacity, behind natural gas but ahead of coal. In 2009, 39% of all capacity additions in the USA and 39% of all additions in the EU came from wind energy (Figure 7.11). In China, 5% of the net capacity additions from 2000 to 2009 and 16% of the net additions in 2009 came from wind energy. On a global basis, from 2000 through 2009,

Annual Capacity Additions, by Region [GW]

16 14 12 10 8 6 4 2 0
2006 2007 2008 2009

Europe

North America

Asia

Latin America

Africa & Middle East

Pacic

Figure 7.10 | Annual wind power capacity additions by region (Data source: GWEC, 2010a). Note: Regions shown in the gure are dened by the study.

557

Wind Energy

Chapter 7

201 GW

324 GW

26 GW

25 GW

Total Additions

Percent of Nameplate Electric Capacity Additions [%]

100 90 80 70 60 50 40 30 20 10 0 EU US EU US
Other Coal Natural Gas Wind

A number of countries are beginning to achieve relatively high levels of annual wind electricity penetration in their respective electric systems. Figure 7.12 presents data for the end of 2009 (and the end of 2006, 2007 and 2008) on installed wind power capacity, translated into projected annual electricity supply, and divided by electricity consumption. On this basis, and focusing only on the 20 countries with the greatest cumulative wind power capacity, at the end of 2009, wind power capacity was capable of supplying electricity equal to roughly 20% of Denmarks annual electricity demand, 14% of Portugals, 14% of Spains, 11% of Irelands and 8% of Germanys (Wiser and Bolinger, 2010).17

7.4.3

Industry development

2000-2009

2009

Figure 7.11 | Relative contribution of electricity supply types to gross capacity additions in the EU and the USA (Data sources: EWEA, 2010b; Wiser and Bolinger, 2010). Note: The other category includes other forms of renewable energy, nuclear energy, and fuel oil.

The growing maturity of the wind energy sector is illustrated not only by wind power capacity additions, but also by trends in the wind energy industry. In particular, major established companies from outside the traditional wind energy industry have become increasingly involved in the sector. For example, there has been a shift in the type of companies developing, owning and operating wind power plants, from relatively small independent power plant developers to large power generation

Projected Wind Electricity as a Proportion of Electricity Consumption [%]

22 20 18 16 14 12 10 8 6 4 2 Denmark Portugal Spain Ireland Germany Greece Netherlands UK Italy India Austria US Sweden France Australia Canada Turkey China Brazil Japan TOTAL 0
Approximate Wind Penetration, End of 2009 Approximate Wind Penetration, End of 2008 Approximate Wind Penetration, End of 2007 Approximate Wind Penetration, End of 2006

Figure 7.12 | Approximate annual average wind electricity penetration in the twenty countries with the greatest installed wind power capacity (Wiser and Bolinger, 2010).

roughly 11% of all newly installed net electric capacity additions came from new wind power plants; in 2009 alone, that gure was probably more than 20%.16
16 Worldwide capacity additions from 2000 through 2007 come from historical data from the US Energy Information Administration. Capacity additions for 2008 and 2009 are estimated based on historical capacity growth from 2000 to 2007. The focus here is on capacity additions in GW terms, though it is recognized that electricity generation technologies often have widely divergent average capacity factors, and that the contribution of wind energy to new electricity demand (in GWh terms) may differ from what is presented here.

companies (including electric utilities) and large independent power plant developers. With respect to wind turbine and component manufacturing, the increase in the size and geographic spread of the wind energy market, along with manufacturing localization requirements in some countries, has brought in new players. The involvement of these new players has, in turn, encouraged a greater globalization of the industry. Manufacturer product strategies are shifting to address larger
17 Because of interconnections among electricity grids, these percentages do not necessarily equate to the amount of wind electricity consumed within each country.

558

Chapter 7

Wind Energy

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

(though not all) regions of the world, wind energy is more expensive than current energy market prices, at least if environmental impacts are not internalized and monetized (NRC, 2010a). Wind energy also faces a number of other challenges, some of which are somewhat unique to wind energy or are at least particularly relevant to this sector. Some of the most critical challenges include: (1) concerns about the impact of wind energys variability on electricity reliability; (2) challenges to building the new transmission infrastructure both on- and offshore (and within country and cross-border) needed to enable access to the most attractive wind resource areas; (3) cumbersome and slow planning, siting and permitting procedures that impede wind energy deployment; (4) the technical advancement needs and higher cost of offshore wind energy technology; and (5) lack of institutional and technical knowledge in regions that have not experienced substantial wind energy deployment to this point. As a result of these challenges, growth in the wind energy sector is affected by and responsive to political frameworks and a wide range of government policies. During the past two decades, a signicant number of developed countries and, more recently, a growing number of developing nations have laid out RE policy frameworks that have played a major role in the expansion of the wind energy market. These efforts have been motivated by the environmental, fuel diversity, and economic development impacts of wind energy deployment, as well as the potential for reducing the cost of wind energy over time. An early signicant effort to deploy wind energy at a commercial scale occurred in California, with a feed-in tariff and aggressive tax incentives spurring growth in the 1980s (Bird et al., 2005). In the 1990s, wind energy deployment moved to Europe, with feed-in tariff policies initially established in Denmark and Germany, and later expanding to Spain and then a number of other countries (Meyer, 2007); renewable portfolio standards have been implemented in other European countries and, more recently, European renewable energy policies have been motivated in part by the EUs binding 20%-by-2020 target for renewable energy. In the 2000s, growth in the USA (Bird et al., 2005; Wiser and Bolinger, 2010), China (Li et al., 2007; Li, 2010; Liu and Kokko, 2010), and India (Goyal, 2010) was based on varied policy frameworks, including renewable portfolio standards, tax incentives, feed-in tariffs and government-overseen bidding. Still other policies have been used in a number of countries to directly encourage the localization of wind turbine and component manufacturing (Lewis and Wiser, 2007). Though economic support policies differ, and a healthy debate exists over the relative merits of different approaches, a key nding is that both policy transparency and predictability are important (see Chapter 11). Moreover, though it is not uncommon to focus on economic policies for wind energy, as noted above and as discussed elsewhere in this chapter and in Chapter 11, experience shows that wind energy markets are also dependent on a variety of other factors (e.g., Valentine, 2010). These include local resource availability, site planning and approval procedures, operational integration into electric systems, transmission grid expansion, wind energy technology improvements, and the availability of institutional and technical knowledge in markets unfamiliar with

7.4.4

Impact of policies

18

The deployment of wind energy must overcome a number of challenges that vary in type and magnitude depending on the wind energy application and region.19 The most signicant challenges to wind energy deployment are summarized here. Perhaps most importantly, in many
18 Non-technology-specic policy issues are covered in Chapter 11 of this report. 19 For a broader discussion of barriers and market failures associated with renewable energy, see Sections 1.4 and 11.1, respectively.

559

Wind Energy

Chapter 7

wind energy (e.g., IEA, 2009). For the wind energy industry, these issues have been critical in dening both the size of the market opportunity in each country and the rules for participation in those opportunities; many countries with sizable wind resources have not deployed signicant amounts of wind energy as a result of these factors. Given the challenges to wind energy listed earlier, successful frameworks for wind energy deployment might consider the following elements: support systems that offer adequate protability and that ensure investor condence; appropriate administrative procedures for wind energy planning, siting and permitting; a degree of public acceptance of wind power plants to ease implementation; access to the existing transmission system and strategic transmission planning and new investment for wind energy; and proactive efforts to manage wind energys inherent output variability and uncertainty. In addition, R&D by government and industry has been essential to enabling incremental improvements in onshore wind energy technology and to driving the improvements needed in offshore wind energy technology. Finally, for those markets that are new to wind energy deployment, both knowledge (e.g., wind resource mapping expertise) and technology transfer (e.g., to develop local wind turbine manufacturers and to ease grid integration) can help facilitate early installations.

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

7.5

Near-term grid integration issues20

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

7.5.1

Wind energy characteristics

Several important characteristics of wind energy are different from those of many other generation sources. These characteristics must be considered in electric system planning and operation to ensure the reliable and economical operation of the electric power system. The rst characteristic to consider is that the quality of the wind resource and therefore the cost of wind energy is location dependent. As a result, regions with the highest-quality wind resources may not be situated near population centres that have high electricity demands (e.g., Hoppock and Patio-Echeverri, 2010; Liu and Kokko, 2010). Additional transmission infrastructure is therefore sometimes economically justied (and is often needed) to bring wind energy from higher-quality wind resource areas to electricity demand centres as opposed to utilizing lower-quality wind resources that are located closer to demand centres and that may require less new transmission investment (see Sections 7.5.2.3 and 7.5.4.3). The second important characteristic is that wind energy is weather dependent and therefore variablethe power output of a wind power plant varies from zero to its rated capacity depending on prevailing

560

Chapter 7

Wind Energy

weather conditions. Variations can occur over multiple time scales, from shorter-term sub-hourly uctuations to diurnal, seasonal, and ever interannual uctuations (e.g., Van der Hoven, 1957; Justus et al., 1979; Wan and Bucaneg, 2002; Apt, 2007; Rahimzadeh et al., 2011). The nature of these uctuations and patterns is highly site- and region-specic. Figure 7.13 illustrates some elements of this variability by showing the scaled output of an individual wind turbine, a small collection of wind power plants, and a large collection of wind power plants in Germany over 10 consecutive days. An important aspect of wind power variability for electric system operations is the rate of change in wind power output over different relatively short time periods; Figure 7.13 demonstrates that the aggregate output of multiple wind power plants changes much more dramatically over relatively longer periods (multiple hours) than over very short periods (minutes). An important aspect of wind power variability for the purpose of electric sector planning, on the other hand, is the correlation of wind power output with the periods of time when electric system reliability is at greatest risk, typically periods of high electricity demand. In this case, the diurnal, seasonal, and even interannual patterns of wind power output (and the correlation of those patterns with electricity demand) can impact the capacity credit assigned by system planners to wind power plants, as discussed further in Section 7.5.3.4. Third, in comparison with many other types of power plants, wind power output has lower levels of predictability. Forecasts of wind power

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

Normalized Wind Power

1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 7-Dec-06
Single Turbine Group of Wind Plants All German Wind Power

9-Dec-06

11-Dec-06

13-Dec-06

15-Dec-06

17-Dec-06

Date
Figure 7.13 | Example time series of wind power output scaled to wind power capacity for a single wind turbine, a group of wind power plants, and all wind power plants in Germany over a 10-day period in 2006 (Durstewitz et al., 2008)

561

Wind Energy

Chapter 7

plants over a region is less than the scaled output of a single wind turbine. This apparent smoothing of aggregated output is due to the decreasing correlation of output between different wind power plants as distance between those plants increases. If, on the other hand, the output of multiple wind turbines and power plants was perfectly correlated, then the aggregate variability would be equivalent to the scaled variability of a single turbine. With sufcient transmission capacity between wind power plants, the observed geographic smoothing effect has implications for the variability of aggregate wind power output that electric systems must accommodate, and also inuences forecast accuracy because accuracy improves with the number and diversity of wind power plants considered (e.g., Focken et al., 2002).

this progress is not complete, and increased deployment of wind energy will require improved and validated models to allow planners to better assess the capability of electric systems to accommodate wind energy (Coughlan et al., 2007; NERC, 2009).

7.5.2.2

Wind power electrical characteristics and grid codes

7.5.2

Planning electric systems with wind energy

Detailed system planning for new generation and transmission infrastructure is used to ensure that the electric system can be operated reliably and economically in the future. Advanced planning is required due, in part, to the long time horizons required to build new electricity infrastructure. More specically, electric system planners22 must evaluate the adequacy of transmission to deliver electricity to demand centres and the adequacy of generation to maintain a balance between supply and demand under a variety of operating conditions. Though not an exhaustive list, four technical planning issues are prominent when considering increased reliance on wind energy: the need for accurate electric system models of wind turbines and power plants; the development of technical standards for connecting wind power plants with electric systems (i.e., grid codes); the broader transmission infrastructure needs of electric systems with wind energy; and the maintenance of overall generation adequacy with increased wind electricity penetration.

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

7.5.2.1

Electric system models

Computer-based simulation models are used extensively to evaluate the ability of the electric system to accommodate new generation, changes in demand and changes in operational practices. An important role of electric system models is to demonstrate the ability of an electric system to recover from severe events or contingencies. Generic models of typical synchronous generators have been developed and validated over a period of multiple decades, and are used in industry standard software tools (e.g., power system simulators and analysis models) to study how the electric system and all its components will behave during system events or contingencies. Similar generic models of wind turbines and wind power plants are in the process of being developed and validated. Because wind turbines have electrical characteristics that differ from typical synchronous generators, this modelling exercise requires signicant effort. As a result, though considerable progress has been made,
22 Electric system planners (or organizations that plan electric systems) is used here as a generic term that refers to planners within any organization that regulates, operates components of, or builds infrastructure for the electric system.

562

Chapter 7

Wind Energy

2006; Doherty et al., 2010). Active power control (including limits on how quickly wind power plants can change their output) and frequency control are also sometimes required (Singh and Singh, 2009). Finally, controls can be added to wind power plants to enable benecial dampening of inter-area oscillations during dynamic events (Miao et al., 2009).

system, as indicated by the capacity credit assigned to the resource.23 Although there is not a strict, uniform denition of capacity credit, the capacity credit of a generator is usually a system characteristic in that it is determined not only by the generators characteristics but also by the characteristics of the electric system to which that generator is connected, particularly the temporal prole of electricity demand (Amelin, 2009). The contribution of wind energy to long-term reliability can be evaluated using standard approaches, and wind power plants are typically found to have a capacity credit of 5 to 40% of nameplate capacity (see Figure 7.14). The correlation between wind power output and electrical demand is an important determinant of the capacity credit of an individual wind power plant. In many cases, wind power output is uncorrelated or is weakly negatively correlated with periods of high electricity demand, reducing the capacity credit of wind power plants; this is not always the case, however, and wind power output in the UK, for example, has been found to be weakly positively correlated with periods of high demand (Sinden, 2007). These correlations are case specic as they depend on the diurnal, seasonal and yearly characteristics of both wind power output and electricity demand. A second important characteristic of the capacity credit for wind energy is that its value generally decreases as wind electricity penetration levels rise, because the capacity credit of a generator is greater when power output is well-correlated with periods of time when there is a higher risk of a supply shortage. As the level of wind electricity penetration increases, however, assuming that the outputs of wind power plants are positively correlated, the period of greatest risk will shift to times with low average levels of wind energy supply (Hasche et al., 2010). Aggregating wind power plants over larger areas may reduce the correlation between wind power outputs, as described earlier, and can slow the decline in capacity credit as wind electricity penetration increases, though adequate transmission capacity is required to aggregate the output of wind power plants in this way (Tradewind, 2009; EnerNex Corp, 2010).24 The relatively low average capacity credit of wind power plants (compared to fossil fuel-powered units, for example) suggests that systems with large amounts of wind energy will also tend to have signicantly more total nameplate generation capacity (wind and non-wind) to meet the same peak electricity demand than will electric systems without large amounts of wind energy. Some of this generation capacity will operate infrequently, however, and the mix of other generation in an electric system with large amounts of wind energy will tend (on economic grounds) to increasingly shift towards more exible peaking
23 As an example, the addition of a very reliable 100 MW fossil unit in a system with numerous other reliable units will usually increase the load-carrying capability of the system by at least 90 MW, leading to a greater than 90% capacity credit for the fossil unit. 24 Generation resource adequacy evaluations are also beginning to include the capability of the system to provide adequate exibility and operating reserves to accommodate more wind energy (NERC, 2009). The increased demand from wind energy for operating reserves and exibility is addressed in Section 7.5.3.

7.5.2.3

Transmission infrastructure

As noted earlier, the highest-quality wind resources (whether on- or offshore) are often located at a distance from electricity demand centres. As a result, even at low to medium levels of wind electricity penetration, the addition of large quantities of wind energy in areas with the strongest wind resources may require signicant new additions or upgrades to the transmission system (see also Chapter 8). Transmission adequacy evaluations must consider any tradeoffs between the costs of expanding the transmission system to access higher-quality wind resources and the costs of accessing lower-quality wind resources that require less transmission investment (e.g., Hoppock and Patio-Echeverri, 2010). In addition, evaluations of new transmission capacity need to account for the relative smoothing benets of aggregating wind power plants over large areas, the amount of transmission capacity devoted to managing the remaining variability of wind power output, and the broader nonwind-specic advantages and disadvantages of transmission expansion (Burke and OMalley, 2010). Irrespective of the costs and benets of transmission expansion to accommodate increased wind energy deployment, one of the primary challenges is the long time it can take to plan, site, permit and construct new transmission infrastructure relative to the shorter time it often takes to add new wind power plants. Depending on the legal and regulatory framework in any particular region, the institutional challenges of transmission expansion, including cost allocation and siting, can be substantial (e.g., Benjamin, 2007; Vajjhala and Fischbeck, 2007; Swider et al., 2008). Enabling increased penetration of wind electricity may therefore require the creation of regulatory and legal frameworks for proactive rather than reactive transmission planning (Schumacher et al., 2009). Estimates of the cost of the new transmission required to achieve low to medium levels of wind electricity penetration in a variety of locations around the world are summarized in Section 7.5.4.

7.5.2.4

Generation adequacy

Though methods and objectives vary from region to region, generation adequacy evaluations are generally used to assess the capability of generation resources to reliably meet electricity demand. Planners often evaluate the long-term reliability of the electric system by estimating the probability that the system will be able to meet expected demand in the future, as measured by a statistical metric called the load-carrying capability of the system. Each electricity supply resource contributes some fraction of its nameplate capacity to the overall capability of the

563

Wind Energy

Chapter 7

Capacity Credit [% Nameplate Wind Power Capacity]

45
Germany

40
Ireland ESBNG 5GW

35

Ireland ESBNG 6.5GW Mid Norway 3 Wind Farms Mid Norway 1 Wind Farm

30

25
UK 2002

20

UK 2007 US California US Minnesota 2004

15

10
US Minnesota 2006

US New York On-Off-Shore

0 0 10 20 30 40 50 60

Wind Power Penetration [% of Peak Demand]


Figure 7.14 | Estimates of the capacity credit of wind power plants across several wind energy integration studies from Europe and the USA (Holttinen et al., 2009).

and intermediate resources and away from base-load resources (e.g., Lamont, 2008; Milborrow, 2009; Boccard, 2010).

7.5.3

Operating electric systems with wind energy

The unique characteristics of wind energy, and especially power output variability and uncertainty, also hold important implications for electric system operations. Here we summarize those implications in general (Section 7.5.3.1), and then briey discuss three specic case studies of the integration of wind energy into real electricity systems (Section 7.5.3.2).

and also a decrease in the minimum net demand. For example, Figure 7.15 depicts demand and ramp duration curves for Ireland.25 At relatively low levels of wind electricity penetration, the magnitude of changes in net demand, as shown in the 15-minute ramp duration curve, is similar to the magnitude of changes in total demand (Figure 7.15(c)). At higher levels of wind electricity penetration, however, changes in net demand are greater than changes in total demand (Figure 7.15(d)). Similar impacts on changes in net demand with increased wind energy have been reported in the USA (Milligan and Kirby, 2008). The gure also shows that, at high levels of wind electricity penetration, the magnitude of net demand across all hours of the year is lower than total demand, and that in some hours net demand is near or even below zero (Figure 7.15(b)). As a result of these trends, wholesale electricity prices will tend to decline when wind power output is high (or is forecast to be high in the case of day-ahead markets) and transmission interconnection capacity to other energy markets is constrained, with a greater frequency of low or even negative prices (e.g., Jnsson et al., 2010; Morales et al., 2011). As with
25 Figure 7.15 presents demand and ramp duration curves for Ireland with (net demand) and without (demand) the addition of wind energy. A demand duration curve shows the percentage of the year that the demand exceeds a level on the vertical axis. Demand in Ireland exceeds 4,000 MW, for example, about 10% of the year. The ramp duration curves show the percentage of the year that changes in the demand exceed the level on the vertical axis. The 15-min change in demand in Ireland exceeds 100 MW/15minutes, for example, less than 10% of the year.

7.5.3.1

Integration, exibility and variability

Because wind energy is generated with a very low marginal operating cost, it is typically used to meet demand when it is available, thereby displacing the use of generators that have higher marginal costs. This results in electric system operators and markets primarily dispatching other generators to meet demand minus any available wind energy (i.e., net demand). As wind electricity penetration grows, the variability of wind energy results in an overall increase in the magnitude of changes in net demand,

564

Chapter 7

Wind Energy

(a) Demand Duration Curve


5,000 Demand Net Demand

(c) Ramp Duration Curve


Ramp [MW/15 minutes]
400 300 200 100 0 -100 Demand Net Demand

Power [MW]

4,000

3,000

2,000 -200 1,000 0 0.2 0.4 0.6 0.8 1 -300 0 0.2 0.4 0.6 0.8 1

Probability

Probability

(b) Demand Duration Curve (Projected)


6,000

(d) Ramp Duration Curve (Projected)


Ramp [MW/15 minutes]
400

Power [MW]

Demand Net Demand

Demand Net Demand

4,000

200

2,000

-200

-2,000 0 0.2 0.4 0.6 0.8 1

-400 0 0.2 0.4 0.6 0.8 1

Probability

Probability

Figure 7.15 | Demand duration and 15-minute ramp duration curves for Ireland in (a, c) 2008 (wind energy represents 7.5% of total annual average electricity demand), and (b, d) projected for high wind electricity penetration levels (wind energy represents 40% of total annual average electricity demand).1 Source: Data from www.eirgrid.com. Note: 1. Projected demand and ramp duration curves are based on scaling 2008 data (demand is scaled by 1.27 and wind energy is scaled on average by 7). Ramp duration curves show the cumulative probability distributions of 15-minute changes in demand and net demand.

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

electricity penetration, the degree of geographic smoothing, and electric system size, but is generally expected to be relatively small and therefore inexpensive to manage in large electric systems (J. Smith et al., 2007). The more signicant operational challenges relate to the variability and commensurate increased need for exibility to manage changes in wind power output over one to six hours (Doherty and OMalley, 2005; Holttinen et al., 2009). Incorporating state-of-the-art forecasting of wind energy over multiple time horizons into electric system operations can reduce the need for exibility from other generators, and has been found to be especially important as wind electricity penetration levels increase (e.g., Doherty et al., 2004; Tuohy et al., 2009; GE Energy, 2010). Nonetheless, even with high-quality forecasts and geographically dispersed wind power plants, additional start-ups and shut-downs, part-load operation, and ramping will be required from fossil generation units to maintain the supply/demand balance (e.g., Gransson and Johnsson, 2009; Troy et al., 2010). This additional exibility is not free, as it increases the amount of time that fossil fuel-powered units are operated at less efcient part-load conditions (resulting in lower than expected reductions in production costs and emissions from fossil generators as described in Sections

565

Wind Energy

Chapter 7

7.5.4 and 7.6.1.3, respectively), increases wear and tear on boilers and other equipment, increases maintenance costs, and reduces power plant life (Denny and OMalley, 2009). Various kinds of economic incentives can be used to ensure that the operational exibility of other generators is made available to system operators. Some electricity systems, for example, have day-ahead, intra-day, and/or hour-ahead markets for electricity, as well as markets for reserves, balancing energy and other ancillary services. These markets can provide pricing signals for increased (or decreased) exibility when needed as a result of rapid changes in or poorly predicted wind power output, and can therefore reduce the cost of integrating wind energy (J. Smith et al., 2007; Gransson and Johnsson, 2009). Markets with shorter scheduling periods have also been found to be more responsive to variability and uncertainty, thereby facilitating wind energy integration (Holttinen, 2005; Kirby and Milligan, 2008; Tradewind, 2009). In addition, coordinated electric system operations across larger areas has been shown to benet wind energy integration, and increased levels of wind energy supply may therefore tend to motivate greater investments in and electricity trade across transmission interconnections (Milligan and Kirby, 2008; Denny et al., 2010). Where wholesale electricity markets do not exist, other planning methods or incentives would be needed to ensure that generating plants are exible enough to accommodate increased deployment of wind energy. Planning systems and incentives may also need to be adopted to ensure that new generating plants are sufciently exible to accommodate expected wind energy deployment. Moreover, in addition to exible fossil fuel-powered units, hydropower stations, bulk energy storage, large-scale deployment of electric vehicles and their associated contributions to system exibility through controlled battery charging, diverting excess wind energy to fuel production or local heating, and various forms of demand response can also be used to facilitate the integration of wind energy. The deployment of a diversity of RE technologies may also help facilitate overall electric system integration. The role of some of these technologies (as well as some of the operational and planning methods noted earlier) in electric systems is described in more detail in Chapter 8 because they are not all specic to wind energy and because some are more likely to be used at higher levels of wind electricity penetration than considered here (up to 20%). Wind power plants, meanwhile, can provide some exibility by briey curtailing output to provide downward regulation or, in extreme cases, curtailing output for extended periods to provide upward regulation. Modern controls on wind power plants can also use curtailment to limit or even (partially) control ramp rates (Fox et al., 2007). Though curtailing wind power output is a simple and often times readily available source of exibility, there are sizable opportunity costs associated with curtailing plants that have low operating costs before reducing the output of other plants that have high fuel costs. These opportunity costs should be compared to the possible benets of curtailment (e.g., reduced part-load efciency penalties and wear and tear for fossil generators, and avoidance of certain transmission investments) when determining the prevalence of its use.

7.5.3.2

Practical experience with operating electric systems with wind energy

Actual operating experience in different parts of the world demonstrates that electric systems can operate reliably with increased contributions of wind energy (Sder et al., 2007). In four countries, as discussed earlier, wind energy in 2010 was already able to supply from 10 to roughly 20% of annual electricity demand. The three examples reported here demonstrate the challenges associated with this operational integration, and the methods used to manage the additional variability and uncertainty associated with wind energy. Naturally, these impacts and management methods vary across regions for reasons of geography, electric system design and regulatory structure, and additional examples of wind energy integration associated with operations, curtailment and transmission are described in Chapter 8. Moreover, as more wind energy is deployed in diverse regions and electric systems, additional knowledge about the impacts of wind power output on electric systems will be gained. To date, for example, there is little experience with severe contingencies (i.e., faults) during times with high instantaneous wind electricity penetration. Though existing experience demonstrates that electric systems can operate with wind energy, further analysis is required to determine whether electric systems are maintaining the same level of overall security, measured by the ability of the system to withstand major contingencies, with and without wind energy, and depending on various management options. Limited analysis (e.g., EirGrid and SONI, 2010; Eto et al., 2010) suggests that particular systems are able to survive such conditions but, if primary frequency control reserves are reduced as thermal generation is increasingly displaced by wind energy, additional management options may be needed to maintain adequate frequency response. The security of the electric system with high instantaneous wind electricity penetrations is described in more detail in Chapter 8. Denmark has the highest wind electricity penetration of any country in the world, with wind energy supply equating to approximately 20% of total annual electricity demand. Total wind power capacity installed by the end of 2009 equalled 3.4 GW, while the peak demand in Denmark was 6.5 GW. Much of the wind power capacity (2.7 GW) is located in western Denmark, resulting in instantaneous wind power output exceeding total demand in western Denmark in some instances (see Figure 7.16). The Danish example demonstrates the benets of having access to markets for exible resources and having strong transmission interconnections to neighbouring countries. Denmarks electricity systems operate without serious reliability issues in part because the country is well interconnected to two different electric systems. In conjunction with wind power output forecasting, this allows wind energy to be exported to other markets and helps the Danish operators manage wind power variability. The interconnection with the Nordic system, in particular, provides access to exible hydropower resources, and balancing the Danish system is much more difcult during periods when

566

Chapter 7

Wind Energy

one of the interconnections is down. Even more exibility is expected to be required, however, if Denmark markedly increases its penetration of wind electricity (Ea Energianalyse, 2007). In contrast to the strong interconnections of the Danish system with other electric systems, the island of Ireland has a single synchronous system; its size is similar to the Danish system but interconnection capacity with other markets is limited to a single 500 MW high-voltage direct current link.The wind power capacity installed by the end of 2009 was capable of supplying roughly 11% of Irelands annual electricity demand, and the Irish system operators have successfully managed that level of wind electricity penetration.The large daily variation in electricity demand in Ireland, combined with the isolated nature of the Irish system, has resulted in a relatively exible electric system that is particularly well suited to integrating wind energy; exible natural gas plants generated 65% of the electrical energy in the rst half of 2010. As a result, despite the lack of signicant interconnection capacity, the Irish system has successfully operated with instantaneous levels of wind electricity penetration of over 40% (see Figure 7.16). Nonetheless, it is recognized that as wind electricity penetration levels increase further, new challenges will arise. Of particular concern are: the possible lack of inertial response of wind turbines absent additional turbine controls, which could lead to increased frequency excursions during severe grid contingencies (Lalor et al., 2005); the need for even greater exibility to maintain supply-demand balance; and the need to build additional high-voltage transmission (AIGS, 2008). Moreover, in common with the Danish experience, much of the wind energy is and will be connected to the distribution system, requiring attention to voltage control issues (Vittal et al., 2010). Figure 7.16 illustrates the high levels of instantaneous wind electricity penetration that exist in Ireland and West Denmark. The Electric Reliability Council of Texas (ERCOT) operates a synchronous system with a peak demand of 63 GW and 8.5 GW of wind power capacity, and with a wind electricity penetration level of 6% of annual electricity demand by the end of 2009. ERCOTs experience

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

7.5.4

Results from integration studies

In addition to actual operating experience, a number of high-quality studies of the increased transmission and generation resources required to accommodate wind energy have been completed, primarily covering OECD countries. As summarized further below, these studies employ a wide variety of methodologies and have diverse objectives, but typically seek to evaluate the capability of the electric system to integrate increased penetrations of wind energy and to quantify the costs and benets of operating the system with wind energy. The issues and costs often considered by these studies are reviewed in this section, and include: the increased operating reserves and balancing costs required

Demand and Wind [MW]

Wind Penetration [%]

Demand and Wind [MW]

4,000

Demand Wind % Wind

90 80 70 60 50

42% 4,000 40

3,000

3,000 Demand 2,000 Wind % Wind 1,000

30

2,000

40 30

20

1,000

20 10

10

0 10/01 11/01 12/01 13/01 14/01 15/01 16/01

0 04-Apr-10 05-Apr-10

Figure 7.16 | Wind energy, electricity demand and instantaneous penetration levels in (left) West Denmark for a week in January 2005, and (right) the island of Ireland for two days in April 2010. Source: Data from (left) www.energinet.dk; (right) www.eirgrid.com and System Operator for Northern Ireland.

567

Wind Penetration [%]

5,000

101%

98%

100

5,000

50

Wind Energy

Chapter 7

to accommodate the variability and uncertainty in net demand caused by wind energy; the requirement to maintain sufcient generation adequacy; and the possible need for additional transmission infrastructure. The studies also frequently analyze the benets of adding wind energy, including avoided fossil fuel consumption and CO2 emissions, though these benets are not reviewed in this section. This section focuses on the general results of these studies as a whole; see Chapter 8 for brief descriptions of individual study results, including some studies that have investigated somewhat higher levels of wind electricity penetration than considered here.

integrating up to 20% wind energy into electric systems is, in most cases, modest but not insignicant. Specically, at low to medium levels of wind electricity penetration (up to 20% wind energy), the available literature (again, primarily from a subset of OECD countries) suggests that the additional costs of managing electric system variability and uncertainty, ensuring generation adequacy and adding new transmission to accommodate wind energy will be system specic but generally in the range of US cents2005 0.7 to 3/kWh.26 Concerns about (and the costs of) wind energy integration will grow with wind energy deployment and, even at lower penetration levels, integration issues must be actively managed.

7.5.4.1

Methodological challenges

7.5.4.2

Increased balancing cost with wind energy

Estimating the incremental impacts and costs of wind energy integration is difcult due to the complexity of electric systems and study data requirements. One of the most signicant challenges in executing these studies is simulating wind power output data at high time resolutions for a chosen future wind electricity penetration level and for a sufcient duration for the results of the analysis to accurately depict worst-case conditions and correlations of wind and electricity demand. These data are then used in electric system simulations to mimic system planning and operations, thereby quantifying the impacts, costs and benets of wind energy integration. Addressing all integration impacts requires several different simulation models that operate over different time scales, and most individual studies therefore focus on a subset of the potential issues. The results of wind energy integration studies are also dependent on pre-existing differences in electric system designs and regulatory environments: important differences include generation capacity mix and the exibility of that generation, the variability of demand and the strength and breadth of the transmission system. In addition, study results differ and are hard to compare because standard methodologies and even denitions have not been developed, though signicant progress has been made in developing agreement on many high-level study design principles (Holttinen et al., 2009). The rst-generation integration studies, for example, used models that were not designed to fully reect the variability and uncertainty of wind energy, resulting in studies that addressed only parts of the larger system. More recent studies, on the other hand, have used models that can incorporate the uncertainty of wind power output from the day-ahead time scale to some hours ahead of delivery (e.g., Meibom et al., 2009; Tuohy et al., 2009). Integration studies are also increasingly simulating high wind electricity penetration scenarios over entire synchronized systems (not just individual, smaller balancing areas) (e.g., Tradewind, 2009; EnerNex Corp, 2010; GE Energy, 2010). Finally, only recently have studies begun to explore in more depth the capability of electric systems to maintain primary frequency control during system contingencies with high penetrations of wind energy (e.g., EirGrid and SONI, 2010; Eto et al., 2010). Regardless of the challenges of executing and comparing such studies, the results, as described in more detail below, demonstrate that the cost of

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

568

Chapter 7

Wind Energy

7.5.4.3

Relative cost of generation adequacy with wind energy

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

to electricity demand, the geographic distribution of wind power plant siting and the level of wind electricity penetration will all impact the capacity credit estimated for wind energy, and therefore the relative cost of generation adequacy.

7.5.4.4

Cost of transmission for wind energy

Finally, a number of assessments of the need for and cost of upgrading or building large-scale transmission infrastructure between wind resource regions and demand centres have similarly found modest, but not insignicant, costs.28 The transmission cost for achieving 20% wind electricity penetration in the USA, for example, was estimated to add about USD2005 150 to 290/kW to the investment cost of wind power plants (US DOE, 2008). The cost of this transmission expansion was found to be justied because of the higher quality of the wind resources accessed if the transmission were to be built relative to accessing only lower-quality wind resources with less transmission expansion. More

Increase in Balancing Cost [UScent/kWh Wind]

0.6
Nordic 2004

0.5

Finland 2004 UK, 2002 UK, 2007

0.4

Ireland Colorado US Minnesota 2004

0.3

Minnesota 2006 California US Pacicorp US

0.2

Greennet Germany Greennet Denmark Greennet Finland

0,1
Greennet Norway Greennet Sweden

0.0 0 5 10 15 20 25 30

Wind Electricity Penetration [% of Annual Electricity Demand]


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

20%. Using a somewhat different approach, Boccard (2010) provides a comparable estimate of the generation adequacy cost of wind energy in several European countries. As discussed earlier, the methodology used to assess generation adequacy, the correlation of wind power output

28 These costs are distinct from the costs to connect individual wind power plants to the transmission system; connection costs are often included in estimates of the investment costs of wind power plants (see Section 7.8).

569

Wind Energy

Chapter 7

detailed assessments of the transmission needed to accommodate increased wind energy deployment in the USA have found a wide range of results, with estimated costs ranging from very low to sometimes reaching (or even exceeding) USD2005 400/kW (JCSP, 2009; Mills et al., 2009a; EnerNex Corp, 2010). Large-scale transmission for cases with increased wind energy has also been considered in Europe (Czisch and Giebel, 2000) and China (Lew et al., 1998). Results from country-specic transmission assessments in Europe have resulted in varied estimates of the cost of new large-scale transmission; Auer et al. (2004) and EWEA (2005) identied transmission costs for a number of European studies, with cost estimates that are somewhat lower than those found in the USA. Holttinen et al. (2009) reviewed wind energy transmission costs from several European national case studies, and found costs ranging from USD2005 0/kW to as high as USD2005 310/kW. Transmission expansion for wind energy can be justied by the reduction in congestion costs that would occur for the same level of wind energy deployment without transmission expansion. A European-wide study, for example, identied several transmission upgrades between nations and between high-quality offshore wind resource areas that would reduce transmission congestion and ease wind energy integration (Tradewind, 2009). The avoided congestion costs associated with transmission expansion were similarly found to justify transmission investments in two US-based detailed integration studies of high wind electricity penetrations (Milligan et al., 2009). At the same time, it is not always appropriate to fully assign the cost of transmission expansion to wind energy deployment. In some cases, these transmission expansion costs can be justied for reasons beyond wind energy, as new transmission can have wider benets including increased electricity reliability, decreased pre-existing congestion and reduced market power (Budhraja et al., 2009). Moreover, wind energy is not unique in potentially requiring new transmission investment; other energy technologies may also require new transmission, and the costs summarized above do not all represent truly incremental costs. Notwithstanding these important caveats, at the higher end of the range from the available literature (USD2005 400/kW), transmission expansion costs add roughly US cents2005 1.5/kWh to the levelized cost of wind energy. At the lower end, effectively no new transmission costs would need to be specically assigned to the support of wind energy.

potential to produce some detrimental impacts on the environment and on human activities and well-being, and many local and national governments have established planning, permitting and siting requirements to reduce those impacts. These potential concerns need to be taken into account to ensure a balanced view of the advantages and disadvantages of wind energy, especially if wind energy is to expand on a large scale. This section summarizes the best available knowledge about the most relevant environmental net benets of wind energy (Section 7.6.1), while also addressing ecological impacts (Section 7.6.2), impacts on human activities and well-being (Section 7.6.3), public attitudes and acceptance (Section 7.6.4) and processes for minimizing social and environmental concerns (Section 7.6.5).

7.6.1

Environmental net benets of wind energy

The environmental benets of wind energy come primarily from displacing the emissions from fossil fuel-based electricity generation. However, the manufacturing, transport, installation, operation and decommissioning of wind turbines induces some indirect negative effects, and the variability of wind power output also impacts the operations and emissions of fossil fuel-red plants. Such effects need to be subtracted from the gross benets of wind energy in order to estimate net benets. As shown below, these latter effects are modest compared to the net GHG reduction benets of wind energy.

7.6.1.1

Direct impacts

The major environmental benets of wind energy (as well as other forms of RE) result from displacing electricity generation from fossil fuel-based power plants, as the operation of wind turbines does not directly emit GHGs or other air pollutants. Similarly, unlike some other generation sources, wind energy requires insignicant amounts of water, produces little waste and requires no mining or drilling to obtain its fuel supply (see Chapter 9). Estimating the environmental benets of wind energy is somewhat complicated by the operational characteristics of the electric system and the decisions that are made about investments in new power plants to economically meet electricity demand (Deutsche Energie-Agentur, 2005; NRC, 2007; Pehnt et al., 2008). In the short run, increased wind energy will typically displace the operations of existing fossil fuel-based plants that are otherwise on the margin. In the longer term, however, new generating plants may be needed, and the presence of wind energy can inuence what types of power plants are built; specically, increased wind energy will tend to favour on economic grounds exible peaking/ intermediate plants that operate less frequently over base-load plants (Kahn, 1979; Lamont, 2008). Because the impacts of these factors are both complicated and system specic, the benets of wind energy will also be system specic and are difcult to forecast with precision.

7.6

Environmental and social impacts29

Wind energy has signicant potential to reduce (and already is reducing) GHG emissions, together with the emissions of other air pollutants, by displacing fossil fuel-based electricity generation. Because of the commercial readiness (Section 7.3) and cost (Section 7.8) of the technology, wind energy can be immediately deployed on a large scale (Section 7.9). As with other industrial activities, however, wind energy also has the
29 A comprehensive assessment of social and environmental impacts of all RE sources covered in this report can be found in Chapter 9.

570

Chapter 7

Wind Energy

Planetary boundary layer research is important for accurately determining wind ow and turbulence in the presence of various atmospheric stability effects and complex land surface characteristics. Research in mesoscale atmospheric processes aims at improving the fundamental understanding of mesoscale and local wind ows (Banta et al., 2003; Kelley et al., 2004). In addition to its contribution towards understanding turbine-level aerodynamic and array wake effects, a better understanding of mesoscale atmospheric processes will yield improved wind energy resource assessments and forecasting methods. Physical and statistical modelling to resolve spatial scales in the 100- to 1,000-m range, a notable gap in current capabilities (Wyngaard, 2004), could occupy a central role of this research. Finally, additional research is warranted on the interaction between global and local climate effects, and wind energy. Specically, work is needed to identify and understand historical trends in wind resource variability in order to increase the reliability of future wind energy performance predictions. As discussed earlier in this chapter, further work is also warranted on the possible impacts of climate change on wind energy resource conditions, and on the impact of wind energy development on local, regional and global climates. Signicant progress in many of the above areas requires interdisciplinary research. Also crucial is the need to use experiments and observations in a coordinated fashion to support and validate computation and theory. Models developed in this way will help improve: (1) wind turbine design; (2) wind power plant performance estimates; (3) wind resource assessments; (4) short-term wind energy forecasting; and (5) estimates of the impact of large-scale wind energy deployment on the local climate, as well as the impact of potential climate change effects on wind resources.

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

7.8.1

Factors that affect the cost of wind energy

The levelized cost of energy from on- and offshore wind power plants is affected by ve primary factors: annual energy production, investment costs, O&M costs, nancing costs and the assumed economic life of the plant.43 Available support policies can also inuence the cost (and price) of wind energy, as well as the cost of other electricity supply options, but these factors are not addressed here. The nature of the wind resource, which varies geographically and temporally, largely determines the annual energy production from a prospective wind power plant, and is among the most important economic factors (Burton et al., 2001). Precise micro-siting of wind power plants and even individual turbines is critical for maximizing energy production. The trend towards turbines with larger rotor diameters and taller towers has led to increases in annual energy production per unit of installed capacity, and has also allowed wind power plants in lower-resource areas to become more economically competitive. Larger wind power plants, meanwhile, have led to consideration of array effects whereby the production of downwind turbines is affected by those turbines located upwind. Offshore power plants will, generally, be exposed to better wind resources than will onshore plants (EWEA, 2009). Wind power plants are capital intensive and, over their lifetime, the initial investment cost ranges from 75 to 80% of total expenditure, with O&M costs contributing the balance (Blanco, 2009; EWEA, 2009). The investment cost includes the cost of the turbines (turbines, transportation to site, and installation), grid connection (cables, sub-station, connection), civil works (foundations, roads, buildings), and other costs (engineering, licensing, permitting, environmental assessments and monitoring equipment). Table 7.4 shows a rough breakdown of the investment cost components for modern wind power plants. Turbine costs comprise more than 70% of total investment costs for onshore wind power plants. The remaining investment costs are highly site-specic. Offshore wind power plants are dominated by these other costs, with the turbines often contributing less than 50% of the total. Site-dependent characteristics such as water depth and distance to shore signicantly affect grid connection, civil works and
42 The environmental impacts and costs of RE and non-RE sources are summarized in Chapters 9 and 10, respectively. 43 Decommissioning costs also exist, but are not expected to be sizable in most instances.

7.8

Cost trends41

Though the cost of wind energy has declined signicantly since the 1980s, policy measures are currently required to ensure rapid deployment in most regions of the world (e.g., NRC, 2010b). In some areas with good wind resources, however, the cost of wind energy is competitive with current energy market prices (e.g., Berry, 2009; IEA, 2009; IEA and OECD, 2010). Moreover, continued technology advances in on- and offshore wind energy are expected (Section 7.7), supporting further cost reductions. The degree to which wind energy is utilized globally and regionally will depend largely on the economic performance of wind energy compared to alternative power sources. This section describes the factors that affect the cost of wind energy (Section 7.8.1), highlights historical trends in the cost and performance of wind power plants (Section 7.8.2), summarizes data and estimates the levelized generation cost of wind energy in 2009 (Section 7.8.3),
41 Discussion of costs in this section is largely limited to the perspective of private investors. Chapters 1 and 8 to 11 offer complementary perspectives on cost issues covering, for example, costs of integration, external costs and benets, economywide costs and costs of policies.

583

Wind Energy

Chapter 7

Table 7.4 | Investment cost distribution for on- and offshore wind power plants (Data sources: Blanco, 2009; EWEA, 2009). Cost Component
Turbine Grid connection Civil works Other investment costs

Onshore (%)
7176 1012 79 58

Offshore (%)1
3749 2123 2125 915

Note: 1. Offshore cost categories consolidated from original study.

other costs. Offshore turbine foundations and internal electric grids are also considerably more costly than those for onshore power plants. The O&M costs of wind power plants include xed costs such as land leases, insurance, taxes, management, and forecasting services, as well as variable costs related to the maintenance and repair of turbines, including spare parts. O&M comprises approximately 20% of total wind power plant expenditure over a plants lifetime (Blanco, 2009), with roughly 50% of total O&M costs associated directly with maintenance, repair and spare parts (EWEA, 2009). O&M costs for offshore wind energy are higher than for onshore due to the less mature state of technology as well as the challenges and costs of accessing offshore turbines, especially in harsh weather conditions (Blanco, 2009). Financing arrangements, including the cost of debt and equity and the proportional use of each, can also inuence the cost of wind energy, as can the expected operating life of the wind power plant. For example, ownership and nancing structures have evolved in the USA that minimize the cost of capital while taking advantage of available incentives (Bolinger et al., 2009). Other research has found that the predictability of the policy measures supporting wind energy can have a sizable impact on nancing costs, and therefore the ultimate cost of wind energy (Wiser and Pickle, 1998; Dinica, 2006; Dunlop, 2006; Agnolucci, 2007). Because offshore wind power plants are still relatively new, with greater performance risk, higher nancing costs are experienced than for onshore plants (Dunlop, 2006; Blanco, 2009), and larger rms tend to dominate offshore wind energy development and ownership (Markard and Petersen, 2009).

of wind turbine technology during this period, design improvements and turbine scaling led to decreased investment costs. Historical investment cost data from Denmark and the USA demonstrate this trend (Figure 7.20). From 2004 to 2009, however, investment costs increased. Some of the reasons behind these increased costs are described in Section 7.8.3. There is far less experience with offshore wind power plants, and the investment costs of offshore plants are highly site-specic. Nonetheless, the investment costs of offshore plants have historically been 50 to more than 100% higher than for onshore plants (BWEA and Garrad Hassan, 2009; EWEA, 2009). Moreover, offshore wind power plants built to date have generally been constructed in relatively shallow water and relatively close to shore (see Section 7.3); higher costs would be experienced for deeper water and more distant facilities. Figure 7.21 presents investment cost data for operating and announced offshore wind power plants. Offshore costs have been inuenced by some of the same factors that caused rising onshore costs from 2004 through 2009 (as well as several unique factors), as described in Section 7.8.3, leading to a doubling of the average investment cost of offshore plants from 2004 through 2009 (BWEA and Garrad Hassan, 2009; UKERC, 2010).

7.8.2.2

Operation and maintenance

7.8.2 7.8.2.1

Historical trends Investment costs

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

Modern turbines that meet IEC standards are designed for a 20-year life, and plant lifetimes may exceed 20 years if O&M costs remain at an acceptable level. Few wind power plants were constructed 20 or more years ago, however, and there is therefore limited experience in plant operations over this entire time period (Echavarria et al., 2008). Moreover, those plants that have reached or exceeded their 20-year lifetime tend to have turbines that are much smaller and less sophisticated than their modern counterparts. Early turbines were also designed using more conservative criteria, though they followed less stringent standards than todays designs. As a result, early plants only offer limited guidance for estimating O&M costs for more recent turbine designs. In general, O&M costs during the rst couple of years of a wind power plants life are covered, in part, by manufacturer warranties that are included in the turbine purchase, resulting in lower ongoing costs than in subsequent years. Newer turbine models also tend to have lower initial O&M costs than older models, with maintenance costs increasing

584

Chapter 7

Wind Energy

Investment Cost [USD2005 /kW]

5,000 4,500 Average Project Investment Cost 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008 2008

Investment Cost [USD2005 /kW]

5,000 4,500 4,000 3,500 3,000 2,500 2,000 1,500 1,000 500 0 Individual Project Investment Cost Capacity-Weighted Average Project Investment Cost Polynomial Trend Line

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

Figure 7.20. Investment cost of onshore wind power plants in (upper panel) Denmark (Data source: Nielson et al., 2010) and (lower panel) the USA (Wiser and Bolinger, 2010).

as turbines age (Blanco, 2009; EWEA, 2009; Wiser and Bolinger, 2010). Offshore wind power plants have historically incurred higher O&M costs than onshore plants (Junginger et al., 2004; EWEA, 2009; Lemming et al., 2009).

and 1980s played a major role in improved wind power plant productivity. Advances in wind energy technology, including taller towers and larger rotors, have also contributed to increased energy capture (EWEA, 2009). Though plant-level capacity factors vary widely, data on average eetwide capacity factors45 for a large sample of onshore wind power plants in the USA show a trend towards higher average capacity factors over time, as wind power plants built more recently have higher
45 A wind power plants capacity factor is only a partial indicator of performance (EWEA, 2009). Most turbine manufacturers supply variations on a given generator capacity with multiple rotor diameters and hub heights. In general, for a given generator capacity, increasing the hub height, the rotor diameter, or the average wind speed will result in an increased capacity factor. When comparing different wind turbines, however, it is possible to increase annual energy capture by using a larger generator, while at the same time decreasing the capacity factor.

7.8.2.3

Energy production

The performance of wind power plants is highly site-specic, and is primarily governed by the characteristics of the local wind regime, which varies geographically and temporally. Wind power plant performance is also impacted by wind turbine design optimization, performance, and availability, however, and by the effectiveness of O&M procedures. Improved resource assessment and siting methodologies developed in the 1970s

585

2009

2009

Wind Energy

Chapter 7

Investment Cost [USD2005 /kW]

7,000

Investment Cost for Operating European Project

6,000

Announced Investment Cost for Proposed U.S. Project Announced Investment Cost for Proposed European Project

5,000

Capacity-Weighted Average Project Investment Cost

4,000

3,000

2,000

1,000

0 1990

1995

2000

2005

2010

2015

Figure 7.21 | Investment cost of operating and announced offshore wind power plants (Musial and Ram, 2010).

average capacity factors than those built earlier (Figure 7.22). Higher hub heights and larger rotor sizes are primarily responsible for these improvements, as the more recent wind power plants built in this time period and included in Figure 7.22 were, on average, sited in relatively lower-quality wind resource regimes. Using a different metric for wind power plant performance, annual energy production per square meter of swept rotor area (kWh/m2) for a given wind resource site, improvements of 2 to 3% per year over the last 15 years have been documented (IEA, 2008; EWEA, 2009).

7.8.3 7.8.3.1

Current conditions Investment costs

The investment costs for onshore wind power plants installed worldwide in 2009 averaged approximately USD2005 1,750/kW, with many plants falling in the range of USD2005 1,400 to 2,100/kW (Milborrow, 2010); data in IEA Wind (2010) are reasonably consistent with this range. Wind power plants installed in the USA in 2009 averaged USD2005 1,900/ kW (Wiser and Bolinger, 2010). Costs in some markets were lower: for

Capacity Factor [%]

35 30 25 20 15 10 5 0

1999
11 701

2000
17 1.158

2001
77 2.199

2002
124 3.955

2003
137 4.458

2004
163 5.784

2005
190 6.467

2006
217 9.289

2007
258 11.253

2008
299 16.068

2009
242 22.346

Projects: MW:

Figure 7.22 | Fleet-wide average capacity factors for a large sample of wind power plants in the USA from 1999 to 2009 (Wiser and Bolinger, 2010).

586

Chapter 7

Wind Energy

example, average investment costs in China in 2008 and 2009 were around USD2005 1,000 to 1,350/kW, driven in part by the dominance of several Chinese turbine manufacturers serving the market with lowercost wind turbines (China Renewable Energy Association, 2009; Li and Ma, 2009; Li, 2010). Wind power plant investment costs rose from 2004 to 2009 (Figure 7.20), an increase primarily caused by the rising price of wind turbines (Wiser and Bolinger, 2010). Those price increases have been attributed to a number of factors. Increased rotor diameters and hub heights have enhanced the energy capture of modern wind turbines, for example, but those performance improvements have come with increased turbine costs, measured on a dollar per kW basis. The costs of raw materials, including steel, copper, cement, aluminium and carbon bre, also rose sharply from 2004 through mid-2008 as a result of strong global economic growth. The strong demand for wind turbines over this period also put upward pressure on labour costs, and enabled turbine manufacturers and their component suppliers to boost prot margins. Strong demand, in excess of available supply, also placed particular pressure on critical components such as gearboxes and bearings (Blanco, 2009). Moreover, because many of the wind turbine manufacturers have historically been based in Europe, and many of the critical components have similarly been manufactured in Europe, the relative value of the Euro compared to other currencies also contributed to the wind turbine price increases in certain countries. Turbine manufacturers and component suppliers responded to the tight supply over this period by expanding or adding new manufacturing facilities. Coupled with reductions in materials costs that began in late 2008 as a result of the global nancial crisis, these trends began to moderate wind turbine prices in 2009 (Wiser and Bolinger, 2010). Due to the relatively small number of operating offshore wind power plants, investment cost data are sparse. Nonetheless, the average cost of offshore wind power plants is considerably higher than that for onshore plants, and the factors that have increased the cost of onshore plants have similarly affected the offshore sector. The limited availability of turbine manufacturers supplying the offshore market and of vessels to install such plants exacerbated cost increases since 2004, as has the installation of offshore plants in increasingly deeper waters and farther from shore, and the erce competition among industry players for early-year (before 2005) demonstration plants (BWEA and Garrad Hassan, 2009; UKERC, 2010). As a result, offshore wind power plants over 50 MW in size, either built between 2006 and 2009 or planned for the early 2010s, had investment costs that ranged from approximately USD2005 2,000 to 5,000/kW (BWEA and Garrad Hassan, 2009; IEA, 2009; Snyder and Kaiser, 2009a; Musial and Ram, 2010). The most recently installed or announced plants cluster towards the higher end of this range, from USD2005 3,200 to 5,000/kW (Milborrow, 2010; Musial and Ram, 2010; UKERC, 2010). These investment costs are roughly 100% higher than costs seen from 2000 to 2004 (BWEA and Garrad Hassan, 2009; Musial and Ram, 2010; UKERC, 2010). Notwithstanding the increased water depth of offshore plants, the majority of the operating plants

have been built in relatively shallow water. Offshore plants built in deeper waters, which are becoming increasingly common and are partly reected in the costs for announced plants, will have relatively higher costs.

7.8.3.2

Operation and maintenance

Though xed O&M costs such as insurance, land payments and routine maintenance are relatively easy to estimate, variable costs such as repairs and spare parts are more difcult to predict (Blanco, 2009). O&M costs can vary by wind power plant, turbine type and age, and the availability of a local servicing infrastructure, among other factors. Levelized O&M costs for onshore wind energy are often estimated to range from US cents2005 1.2 to 2.3/kWh (Blanco, 2009); these gures are reasonably consistent with costs reported in EWEA (2009), IEA (2010c), Milborrow (2010), and Wiser and Bolinger (2010). Limited empirical data exist on O&M costs for offshore wind energy, due in large measure to the limited number of operating plants and the limited duration of those plants operation. Reported or estimated O&M costs for offshore plants installed since 2002 range from US cents2005 2 to 4/kWh (EWEA, 2009; IEA, 2009, 2010c; Lemming et al., 2009; Milborrow, 2010; UKERC, 2010).

7.8.3.3

Energy production

Onshore wind power plant performance varies substantially, with capacity factors ranging from below 20 to more than 50% depending largely on local resource conditions. Among countries, variations in average performance also reect differing wind resource conditions, as well as any difference in the wind turbine technology that is deployed: the average capacity factor for Germanys installed plants has been estimated at 20.5% (BTM, 2010); European country-level average capacity factors range from 20 to 30% (Boccard, 2009); average capacity factors in China are reported at roughly 23% (Li, 2010); average capacity factors in India are reported at around 20% (Goyal, 2010); and the average capacity factor for US wind power plants is above 30% (Wiser and Bolinger, 2010). Offshore wind power plants often experience a narrower range in capacity factors, with a typical range of 35 to 45% for the European plants installed to date (Lemming et al., 2009); some offshore plants in the UK, however, have experienced capacity factors of roughly 30%, in part due to relatively high component failures and access limitations (UKERC, 2010). Because of these variations among countries and individual plants, which are primarily driven by local wind resource conditions but are also affected by turbine design and operations, estimates of the levelized cost of wind energy must include a range of energy production estimates. Moreover, because the attractiveness of offshore plants is enhanced by the potential for greater energy production than for onshore plants, performance variations among on- and offshore wind energy must also be considered.

587

Wind Energy

Chapter 7

7.8.3.4

Levelized cost of energy estimates

Using the methods summarized in Annex II, the levelized generation cost of wind energy is presented in Figure 7.23. For onshore wind energy, estimates are provided for plants built in 2009; for offshore wind energy, estimates are provided for plants built in 2008 and 2009 as well as those plants planned for completion in the early 2010s.46 Estimated levelized costs are presented over a range of energy production estimates to represent the cost variation associated with inherent differences in the wind resource. The x-axis for these charts roughly correlates to annual average

are used to produce levelized generation cost estimates.48 Taxes, policy incentives, and the costs of electric system integration are not included in these calculations.49 The levelized cost of on- and offshore wind energy varies substantially, depending on assumed investment costs, energy production and discount rates. For onshore wind energy, levelized generation costs in good to excellent wind resource regimes are estimated to average US cents2005 5 to 10/kWh. Levelized generation costs can reach US cents2005 15/kWh in lower- resource areas. The costs of wind energy in China and

Levelized Cost of Energy [UScent2005 /kWh]

Levelized Cost of Energy [UScent2005 /kWh]

35
Offshore USD 5,000/kW

35
Offshore Discount Rate = 10%

30

Offshore USD 3,900/kW Offshore USD 3,200/kW

30

Offshore Discount Rate = 7% Offshore Discount Rate = 3% Onshore Discount Rate = 10%

25

Onshore USD 2,100/kW Onshore USD 1,750/kW Onshore USD 1,200/kW

25

Onshore Discount Rate = 7% Onshore Discount Rate = 3%

20

20

15

15

10

Europe Offshore Projects

10

5
China European Low-Medium Wind Areas US Great Plains

0 15 20 25 30 35 40 45 50 Capacity Factor [%]

0 15 20 25 30 35 40 45 50

Capacity Factor [%]

Figure 7.23 | Estimated levelized cost of on- and offshore wind energy, 2009: (left) as a function of capacity factor and investment cost* and (right) as a function of capacity factor and discount rate**. Notes: * Discount rate assumed to equal 7%. ** Onshore investment cost assumed at USD2005 1,750/kW, and offshore at USD2005 3,900/kW.

wind speeds from 6 to 10 m/s. Onshore investment costs are assumed to range from USD2005 1,200 to 2,100/kW (with a mid-level cost of USD2005 1,750/kW); investment costs for offshore wind energy are assumed to range from USD2005 3,200 to 5,000/kW (mid-level cost of USD2005 3,900/ kW).47 Levelized O&M costs are assumed to average US cents2005 1.6/ kWh and US cents2005 3/kWh over the life of the plant for onshore and offshore wind energy, respectively. A power plant design life of 20 years is assumed, and discount rates of 3 to 10% (mid-point estimate of 7%)
46 Because investment costs have risen in recent years, using the cost of recent and planned plants reasonably reects the current cost of offshore wind energy. 47 Based on data presented earlier in this section, the mid-level investment cost for onand offshore wind power plants does not represent the arithmetic mean between the low and high end of the range.

the USA tend towards the lower range of these estimates, due to lower average investment costs (China) and higher average capacity factors (USA); costs in much of Europe tend towards the higher end of the range due to relatively lower average capacity factors. Though the offshore cost estimates are more uncertain, offshore wind energy is generally more expensive than onshore, with typical levelized generation costs that are estimated to range from US cents2005 10/kWh to more than US cents2005 20/kWh for recently built or planned plants located in relatively
48 Though the same discount rate range and mid-point are used for on- and offshore wind energy, offshore wind power plants currently experience higher-cost nancing than do onshore plants. As such, the levelized cost of energy from offshore plants may, in practice, tend towards the higher end of the range presented in the gure, at least in comparison to onshore plants. 49 Decommissioning costs are generally assumed to be low, and are excluded from these calculations.

588

Chapter 7

Wind Energy

shallow water. Where the exploitable onshore wind resource is limited, however, offshore plants can sometimes compete with onshore plants.

7.8.4.1

Learning curve estimates

7.8.4

Potential for further reductions in the cost of wind energy

The wind energy industry has developed over a period of 30 years. Though the dramatic cost reductions seen in past decades will not continue indenitely, the potential for further reductions remains given the many potential areas of technological advances described in Section 7.7. This potential spans both on- and offshore wind energy technologies; given the relatively less mature state of offshore wind energy, however, greater cost reductions can be expected in that segment. Two approaches are commonly used to forecast the future cost of wind energy, often in concert with some degree of expert judgement: (1) learning curve estimates that assume that future wind energy costs will follow a trajectory that is similar to an historical learning curve based on past costs; and (2) engineering-based estimates of the specic cost reduction possibilities associated with new or improved wind energy technologies or manufacturing capabilities (Mukora et al., 2009).

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

Table 7.5 | Summary of learning curve literature for onshore wind energy. Authors
Neij (1997) Mackay and Probert (1998) Neij (1999) Wene (2000) Wene (2000) Miketa and Schrattenholzer (2004)1 Junginger et al. (2005) Junginger et al. (2005) Klaassen et al. (2005)1 Kobos et al. (2006)1 Jamasb (2007)1 Sderholm and Sundqvist (2007) Sderholm and Sundqvist (2007)1 Neij (2008) Kahouli-Brahmi (2009) Nemet (2009) Ek and Sderholm (2010)1 Wiser and Bolinger (2010)

Learning By Doing Rate (%)


4 14 8 32 18 10 19 15 5 14 13 5 4 17 17 11 17 9

Global or National Independent Variable (cumulative capacity)


Denmark3 USA Denmark3 USA2 EU
2

Dependent Variable
Denmark (turbine cost) USA (turbine cost) Denmark (turbine cost) USA (generation cost) EU (generation cost) Global (investment cost) UK (investment cost) Spain (investment cost) Germany, Denmark, and UK (investment cost) Global (investment cost) Global (investment cost) Germany, Denmark, and UK (investment cost) Germany, Denmark, and UK (investment cost) Denmark (generation cost) Global (investment cost) California (investment cost) Germany, Denmark, Spain, Sweden, and UK (investment cost) USA (investment cost)

Data Years
19821995 19811996 19821997 19851994 19801995 19711997 19922001 19902001 19862000 19811997 19801998 19862000 19862000 19812000 19791997 19812004 19862002 19822009

Global Global Global Germany, Denmark, and UK Global Global Germany, Denmark, and UK Germany, Denmark, and UK Denmark Global Global Global Global

Notes: 1. Two-factor learning curve that also includes R&D; others are one-factor learning curves. 2. Independent variable is cumulative production of electricity. 3. Cumulative turbine production used as independent variable; others use cumulative installations. 50 It is too early to develop a meaningful learning curve for offshore wind energy based on actual data from offshore plants. Studies have sometimes used learning rates to estimate future offshore costs, but those learning rates have typically been synthesized based on judgment and on learning rates for related industries and offshore subsystems (e.g., Junginger et al., 2004; Carbon Trust, 2008b).

589

Wind Energy

Chapter 7

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

reductions associated with specic design changes and/or technical advances. Though limitations to engineering-based approaches also exist (Mukora et al., 2009), these models can lend support to learning curve predictions by dening the technology advances that can yield cost reductions and/or energy production increases. These models have been used to estimate the impact of potential technology improvements on wind power plant investment costs and energy production, as highlighted in Section 7.7. Given the possible technology advances (in combination with manufacturing learning) discussed earlier, the US DOE (2008) estimates that onshore wind energy investment costs may decline by 10% by 2030, while energy production may increase by roughly 15%, relative to a 2008 starting point (see Table 7.3, and the note under that table). There is arguably greater potential for technical advances in offshore than in onshore wind energy technology (see Section 7.7), particularly in foundation design, electrical system design and O&M costs. Larger offshore wind power plants are also expected to trigger more efcient installation procedures and dedicated vessels, enabling lower costs. Future levelized cost of energy reductions have sometimes been estimated by associating potential cost reductions with these technical improvements, sometimes relying on subsystem-level learning curve estimates from other industries (e.g., Junginger et al., 2004; Carbon Trust, 2008b).

7.8.4.3

Projected levelized cost of wind energy

A number of studies have developed forecasted cost trajectories for onand offshore wind energy based on differing combinations of learning curve estimates, engineering models, and/or expert judgement. These estimates are sometimesbut not alwayslinked to certain levels of assumed wind energy deployment. Representative examples of this literature include Junginger et al. (2004), Carbon Trust (2008b), IEA (2008, 2010b, 2010c), US DOE (2008), EWEA (2009), Lemming et al. (2009), Teske et al. (2010), GWEC and GPI (2010) and UKERC (2010). Recognizing that the starting year of the forecasts, the methodological approaches used, and the assumed deployment levels vary, these recent studies nonetheless support a range of levelized cost of energy reductions for onshore wind of 10 to 30% by 2020, and for offshore wind of 10 to 40% by 2020. Some studies focused on offshore wind energy technology even identify scenarios in which market factors lead to continued increases in the cost of offshore wind energy, at least in the near to medium term (BWEA and Garrad Hassan, 2009; UKERC, 2010). Longer-term projections are more reliant on assumed deployment levels and are subject to greater uncertainties, but for 2030, the same studies support reductions in the levelized cost of onshore wind energy of 15 to 35% and of offshore wind energy of 20 to 45%. Using these estimates for the expected percentage cost reduction in levelized cost of energy, levelized cost trajectories for on- and offshore wind energy can be developed. Because longer-term cost projections

7.8.4.2

Engineering model estimates

Whereas learning curves examine aggregate historical data to forecast future trends, engineering-based models focus on the possible cost

590

Chapter 7

Wind Energy

are inherently more uncertain and depend, in part, on deployment levels and R&D expenditures that are also uncertain, the focus here is on relatively nearer-term cost projections to 2020. Specically, Section 7.8.3.4 reported 2009 levelized cost of energy estimates for onshore wind energy of roughly US cents2005 5 to 15/kWh, whereas estimates for offshore wind energy were in the range of US cents2005 10 to 20/ kWh. Conservatively, the percentage cost reductions reported above can be applied to these estimated 2009 levelized generation cost values to develop low and high projections for future levelized generation costs.51 Based on these assumptions, the levelized generation cost of onshore wind energy could range from roughly US cents2005 3.5 to 10.5/kWh by 2020 in a high cost-reduction case (30% by 2020), and from US cents2005 4.5 to 13.5/kWh in a low cost-reduction case (10% by 2020). Offshore wind energy is often anticipated to experience somewhat deeper cost reductions, with levelized generation costs that range from roughly US cents2005 6 to 12/kWh by 2020 in a high cost-reduction case (40% by 2020) to US cents2005 9 to 18/kWh in a low cost-reduction case (10% by 2020).52 Uncertainty exists over future wind energy costs, and the range of costs associated with varied wind resource strength introduces greater uncertainty. As installed wind power capacity increases, higher-quality resource sites will tend to be utilized rst, leaving higher-cost sites for later development. As a result, the average levelized cost of wind energy will depend on the amount of deployment, not only due to learning effects, but also because of resource exhaustion. This supply-curve effect is not captured in the estimates presented above. The estimates presented here therefore provide an indication of the technology advancement potential for on- and offshore wind energy, but should be used with some caution.

unlikely to constrain further deployment (Section 7.2). Onshore wind energy technology is already being deployed at a rapid pace (Sections 7.3 and 7.4), therefore offering an immediate option for reducing GHG emissions in the electricity sector. In good to excellent wind resource regimes, the generation cost of onshore wind energy averages US cents2005 5 to 10/kWh (Section 7.8), and no insurmountable technical barriers exist that preclude increased levels of wind energy penetration into electricity supply systems (Section 7.5). Continued technology advances and cost reductions in on- and offshore wind energy are expected (Sections 7.7 and 7.8), further improving the GHG emissions reduction potential of wind energy over the long term. This section begins by highlighting near-term forecasts for wind energy deployment (Section 7.9.1). It then discusses the prospects for and barriers to wind energy deployment in the longer term and the potential role of that deployment in reaching various GHG concentration stabilization levels (Section 7.9.2). Both subsections are largely based on energy market forecasts and GHG and energy scenarios literature published between 2007 and 2010. The section ends with brief conclusions (Section 7.9.3). Though the focus of this section is on larger on- and offshore wind turbines for electricity production, as discussed in Box 7.1, alternative technologies and applications for wind energy also exist.

7.9.1

Near-term forecasts

7.9

Potential deployment53

Wind energy offers signicant potential for near- and long-term GHG emissions reductions. The wind power capacity installed by the end of 2009 was capable of meeting roughly 1.8% of worldwide electricity demand and, as presented in this section, that contribution could grow to in excess of 20% by 2050. On a global basis, the wind resource is
51 Because of the cost drivers discussed earlier in this section, wind energy costs in 2009 were higher than in some previous years. Applying the percentage cost reductions from the available literature to the 2009 starting point is, therefore, arguably a conservative approach to estimating future cost reduction possibilities; an alternative approach would be to use the absolute values of the cost estimates provided by the available literature. As a result, and also due to the underlying uncertainty associated with projections of this nature, future costs outside of the ranges presented here are possible. 52 As mentioned earlier, the 2009 starting point values for offshore wind energy are consistent with recently built or planned plants located in relatively shallow water. 53 Complementary perspectives on potential deployment based on a comprehensive assessment of numerous model-based scenarios of the energy system are presented in Sections 10.2 and 10.3 of this report.

The rapid increase in global wind power capacity from 2000 to 2009 is expected by many studies to continue in the near to medium term (Table 7.6). From the roughly 160 GW of wind power capacity installed by the end of 2009, the IEA (2010b) New Policies scenario and the EIA (2010) Reference case scenario predict growth to 358 GW (forecasted electricity generation of 2.7 EJ/yr) and 277 GW (forecasted electricity generation of 2.5 EJ/yr) by 2015, respectively. Wind energy industry organizations predict even faster deployment rates, noting that past IEA and EIA forecasts have understated actual growth by a sizable margin (BTM, 2010; GWEC, 2010a). However, even these more aggressive forecasts estimate that wind energy will contribute less than 5% of global electricity supply by 2015. Asia, North America and Europe are projected to lead in wind power capacity additions over this period.

7.9.2

Long-term deployment in the context of carbon mitigation

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

591

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020
Larry Parker Specialist in Energy and Environmental Policy January 26, 2010

Congressional Research Service 7-5700 www.crs.gov R41049

CRS Report for Congress


Prepared for Members and Committees of Congress

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

Summary
The European Unions (EU) Emissions Trading Scheme (ETS) is a cornerstone of the EUs efforts to meet its obligation under the Kyoto Protocol. It covers more than 10,000 energy intensive facilities across the 27 EU Member countries; covered entities emit about 45% of the EUs carbon dioxide emissions. A Phase 1 trading period began January 1, 2005. A second, Phase 2, trading period began in 2008, covering the period of the Kyoto Protocol. A Phase 3 will begin in 2013 designed to reduce emissions by 21% from 2005 levels. Several positive results from the Phase 1 learning by doing exercise assisted the ETS in making the Phase 2 process run more smoothly, including: (1) greatly improving emissions data, (2) encouraging development of the Kyoto Protocols project-based mechanismsClean Development Mechanism (CDM) and Joint Implementation (JI), and (3) influencing corporate behavior to begin pricing in the value of allowances in decision-making, particularly in the electric utility sector. However, several issues that arose during the first phase were not resolved as the ETS moved into Phase 2, including allocation schemes and new entrant reserves, and others. A more comprehensive and coordinated response by the EU has been made for Phase 3 with harmonized and coordinated rules being developed by the European Commission. The United States is not a party to the Kyoto Protocol. However, five years of carbon emissions trading has given the EU valuable experience in designing and operating a greenhouse gas trading system. This experience may provide some insight into cap-and-trade design issues currently being debated in the United States. The U.S. requires only electric utilities to monitor CO2. The EU-ETS experience suggests that expanding similar requirements to all facilities covered under a capand-trade scheme would be pivotal for developing allocation systems, reduction targets, and enforcement provisions. In the U.S. debate continues on comprehensive versus sector-specific reduction programs; the EU-ETS experience suggests that adding sectors to a trading scheme once established may be a slow, contentious process. As with most EU industries, most U.S. industry groups either oppose auctions outright or want them to be supplemental to a base free allocation. The EU-ETS experience suggests Congress may want to consider specifying any auction requirement if it wishes to incorporate market economics more fully into compliance decisions. EU-ETS analysis suggests the most important variables in determining Phase 1 allowance price changes were oil and natural gas price changes; this apparent linkage raises possible market manipulation issues, particularly with the inclusion of financial instruments such as options and futures contracts. The EU will examine the matter in preparation for Phase 3. Congress may consider whether the government needs enhanced regulatory and oversight authority over such instruments.

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

Contents
Overview ....................................................................................................................................1 Results from Phase 1 and 2 .........................................................................................................3 Phase 3 .......................................................................................................................................8 Auctions ...............................................................................................................................9 New Entrant Reserves ......................................................................................................... 11 EC Phase 3 Decision on Eligible Industries ......................................................................... 12 Flexibility Mechanisms and Price Volatility Control ............................................................ 13 Expanding Coverage ........................................................................................................... 14 Summary and Considerations for U.S. Cap-and-Trade Proposals ............................................... 15 Emission Inventories and Target Setting .............................................................................. 15 Coverage ............................................................................................................................ 16 Allocation Schemes ............................................................................................................ 17 Flexibility and Price Volatility............................................................................................. 18

Figures
Figure 1. ECX CFI Futures Contracts: Price and Volume.............................................................5 Figure 2. EU-15 Greenhouse Gas Emissions and Projections for the Kyoto Period: 19902012 ........................................................................................................................................6 Figure 3. Summary of EU-15 Emissions Projection Compared to Projected Kyoto Protocol Credits .......................................................................................................................7

Tables
Table 1. Proposed Annual ETS Cap Figures for Phase 3 ..............................................................8

Contacts
Author Contact Information ...................................................................................................... 19

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

Overview1
Climate change is generally viewed as a global issue, but proposed responses typically require action at the national level. With the 1997 Kyoto Protocol now in force and setting emissions objectives for 2008-2012, countries that ratified the protocol are implementing strategies to begin reducing their emissions of greenhouse gases.2 In particular, the European Union (EU) has decided to use an emissions trading scheme (called a cap-and-trade program), along with other market-oriented mechanisms permitted under the Protocol, to help it achieve compliance at least cost.3 The decision to use emission trading to implement the Kyoto Protocol is at least partly based on the successful emissions trading program used by the United States to implement its sulfur dioxide (acid rain) control program contained in Title IV of the 1990 Clean Act Amendments.4 The EUs Emissions Trading System (ETS) covers more than 10,000 energy-intensive facilities across the 27 EU Member countries, including oil refineries, powerplants over 20 megawatts (MW) in capacity, coke ovens, and iron and steel plants, along with cement, glass, lime, brick, ceramics, and pulp and paper installations. In addition, aviation is currently being phased into the ETS. These covered entities emit about 40%-45% of the EUs total greenhouse gas emissions, and almost two-thirds of them are combustion installations. The trading program does not cover either carbon dioxide (CO2) emissions from the transportation sector (except aviation), which account for about 25% of the EUs total greenhouse gas emissions, or emissions of non-CO2 greenhouse gases, which account for about 20% of the EUs total greenhouse gas emissions. A Phase 1 trading period ran between January 1, 2005, and December 31, 2007.5 A Phase 2 trading period began January 1, 2008, covering the period of the Kyoto Protocol, and a Phase 3 has been finalized to begin in 2013.6 Under the Kyoto Protocol, the then-existing 15 nations of the EU agreed to reduce their aggregate annual average emissions for 2008-2012 by 8% from the Protocols baseline level (mostly 1990 levels) under a collective arrangement called a bubble. In light of the Kyoto Protocol targets, the EU adopted a directive establishing the EU-ETS that entered into force October 13, 2003.7

Readers unfamiliar with the workings of the European Union may want to read CRS Report RS21372, The European Union: Questions and Answers, by Kristin Archick and Derek E. Mix. 2 Six gases are included under the Kyoto Protocol: carbon dioxide, methane, nitrous oxide, hydrofluorocarbons, perfluorocarbons, and sulfur hexafluoride. The United States has not ratified the Kyoto Protocol and, therefore, is not covered by its provisions. For more information on the Kyoto Protocol, see CRS Report RL33826, Climate Change: The Kyoto Protocol, Bali Action Plan, and International Actions, by Jane A. Leggett. 3 Norway, a non-EU country, also has instituted a CO2 trading system which is currently linked with the EU-ETS. Various other countries and a state-sponsored regional initiative located in the northeastern United States involving several states are developing mandatory cap-and-trade system programs, but are not operating at the current time. For a review of these emerging programs, along with other voluntary efforts, see CRS Report RL33812, Climate Change: Action by States to Address Greenhouse Gas Emissions, by Jonathan L. Ramseur. 4 P.L. 101-549, Title IV (November 15, 1990). 5 For further background on the ETS, see CRS Report RL34150, Climate Change and the EU Emissions Trading Scheme (ETS): Kyoto and Beyond, by Larry Parker. More information, including relevant directives, on the EU-ETS is available on the European Unions website at http://europa.eu.int/scadplus/leg/en/lvb/l28012.htm. Directive 2003/87/EC of the European Parliament and of the Council of 13 October 2003 establishing a scheme for greenhouse gas emissions allowance trading within the Community and amending Council Directive 96/61/EC.
7 6

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

One objective of the second phase of the ETS is to achieve 3.3 percentage points of the 8.0% reduction required by the EU-15 under the Protocol.8 The importance of emissions trading was elevated by the accession of 12 additional central and eastern European countries to EU membership from May 2004 through January 2007. For the new EU-27, the overall ETS emissions cap is set at 2.08 billion metric tons of carbon dioxide (CO2) annually for the Kyoto compliance period (2008-2012). The second phase Kyoto compliance stage of the ETS is built on the experience the EU gained from its preliminary Phase 1. The European Commission (EC) believes that the Phase 1 learning by doing exercise prepared the community for the difficult task of achieving the reduction requirements of the Kyoto Protocol. Several positive results from the Phase 1 experience assisted the ETS in making the Phase 2 process run smoothly, at least so far. First, Phase 1 established much of the critical infrastructure necessary for a functional emission market, including emissions monitoring, registries, and inventories. Much of the publicized difficulty the ETS experienced early in the first phase can be traced to inadequate emissions data infrastructure.9 Phase 1 significantly improved those critical elements in preparation for Phase 2 implementation. Second, the ETS helped jump-start the project-based mechanismsClean Development Mechanism (CDM) and Joint Implementation (JI)created under the Kyoto Protocol. 10 As stated by Ellerman and Buchner:
The access to external credits provided by the Linking Directive has had an invigorating effect on the CDM and more generally on CO2 reduction projects in developing countries, especially in China and India, the two major countries that will eventually have to become part of a global climate regime if there is to be one.11

Third, according to the EC, a key result of Phase 1 was its effect on corporate behavior. An EC survey of stakeholders indicated that many participants are incorporating the value of allowances in making decisions, particularly in the electric utility sector, where 70% of firms stated they were pricing the value of allowances into their daily operations, and 87% into future marginal pricing decisions. All industries stated that it was a factor in long-term decision-making.12

8 Commission of the European Communities, Communication from the Commission: Progress towards Achieving the Kyoto Objectives (November 19, 2008). Other reductions are to be achieve through regulatory measures, such as a CO2 emissions standard for automobiles.

A. Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, Environmental Economics and Policy (Winter 2007), pp. 69-70; and International Emissions Trading Association, IETA Position Paper on EU ETS Marking Functioning, (no date), p. 3. 10 For more on the effect of the ETS on Kyoto mechanisms, see A. Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, Environmental Economics and Policy (Winter 2007), p. 84; and International Emissions Trading Association, IETA Position Paper on EU ETS Market Functioning (no date), p. 2. For more information on the Kyoto Protocol mechanisms, see CRS Report RL33826, Climate Change: The Kyoto Protocol, Bali Action Plan, and International Actions, by Jane A. Leggett. 11 A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, Environmental Economics and Policy (Winter 2007), p. 84. European Commission, Directorate General for Environment, Review of EU Emissions Trading Scheme: Survey Highlights, (November 2005), pp. 5-7.
12

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

However, several issues that arose during the first phase remained contentious as the ETS implemented Phase 2, including allocation (including use of auctions and reliance on model projections), new entrant reserves, and others. In addition, the expansion of the EU and the implementation of the linking directives created new issues to which Phase 2 has had to respond. 13 Based on lessons learned in Phase 1 and Phase 2, the EU has taken a substantially different approach to these issues in Phase 3 that is discussed later.

Results from Phase 1 and 2


It is unclear to what degree the first phase of the ETS achieved real emissions reductions. Emissions are dynamic over time; a product of a countrys population, economic activity, and greenhouse gas intensity.14 To capture these dynamics, each Member State of the EU developed an emissions baseline from models that project future trends in the countrys emissions based on these and other factors, such as anticipated energy and greenhouse gas policies. 15 During the first phase, the emissions goal was to put the EU on the path to Kyoto compliancenot actually comply with the Protocol (which wasnt necessary until the 2008-2012 time period). Thus, countries developed business as usual baselines based on projected growth in emissions. Such a projected baseline suffers from two sources of uncertainty: data uncertainties, and forecasting uncertainties. On data, Phase 1 suffered from uncertainties with respect to data collection and coverage, in monitoring methods for historic data, and data verification. On projecting future emissions, Phase 1 faced uncertainties with respect to economic or sector-based growth rates. Fueled in many cases by over-optimistic economic growth assumptions, these uncertainties increased the probability of inflated business as usual baselines. 16 The combination of these factors and modest reduction requirements resulted in the emissions allocations for the 2005-2007 trading period being higher than actua1 2005 emissions. 17 This result raised questions about how much reductions achieved during Phase 1 were real as opposed to being merely paper artifacts. On the positive side, verified emissions in 2005 were 3.4% below the estimated 2005 baseline used during the allocation process. In addition, the allowance prices for 2005 stayed persistently high, suggesting some abatement was occurring and raising questions of windfall profits. As stated by Ellerman and Buchner:
First, and most importantly, the persistently high price for EUAs [EU emissions allowances] in a market characterized by sufficient liquidity and sophisticated players must be considered as creating a presumption of abatement. It would be startling if power companies did not
For a further discussion of Phase 2 implementation issues, see CRS Report RL34150, Climate Change and the EU Emissions Trading Scheme (ETS): Kyoto and Beyond, by Larry Parker. 14 For more information, see CRS Report RL33970, Greenhouse Gas Emission Drivers: Population, Economic Development and Growth, and Energy Use, by John Blodgett and Larry Parker.
15 On the role of modeling in the first phase, see A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, 1 Environmental Economics and Policy 1 (Winter 2007), pp. 72-73. 16 Regina Betz and Misato Sato, Emissions Trading: Lessons Learnt from the 1st Phase of the EU ETS and Prospects for the 2nd Phase, 6 Climate Policy (2006), p. 354. 17 For a further discussion, see CRS Report RL33581, Climate Change: The European Union's Emissions Trading System (EU-ETS), by Larry Parker. 13

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

incorporate EUA prices into dispatch decisions that would have shifted generation to less emitting plants. There is plenty of anecdotal evidence that this was the case, and the prominent charges of windfall profits assume that the opportunity cost of freely allocated allowances was being passed on (without noting the implications for abatement). Similarly, it would be surprising if there were no changes in production processes that could be made by the operators of industrial plants.18

However, EU emissions allowances (EUAs) during Phase 1 did not maintain value. Phase 1 EUAs were basically worthless during the final six months of 2007. This decline in EUA prices at least partially reflected the general non-transferability of Phase 1 EUAs to Phase 2. Only Poland and France included limited banking in their Phase 1 implementation plans (called National Allocation Plans (NAPs)). The EC further restricted use of Phase 1 EUAs in Phase 2 with a ruling in November 2006.19 As a result, excess Phase 1 EUAs were worthless at the end of 2007.20 One consequence of the non-transferability of Phase 1 EUAs is that prices for Phase 2 EUAs remained relatively firm until the 2008-2009 recession reduced demand, as indicated by Figure 1. Scarcity is critical for the proper functioning of an allowance market. As further indicated by Figure 1, during 2009, the market firmed up at a much lower level as participants assessed the impact of the recession on the demand for EUAs. This is a different response than the market had during Phase 1, and may reflect Phase 2 improvements in the system. In particular, the more predictable 2009 response may reflect the ability of the EC to certify Phase 2 NAPs using more verifiable baseline data than were available for Phase 1.21 A major reason the EC rejected ex post adjustments22 was fear that such adjustments would have a disruptive effect on the marketplace. 23 Phase 1 did not firmly establish this foundation of markets;24 based on the Phase 2 EUA futures market, further market development appears to be occurring, although, like most commodity markets, it remains somewhat volatile at times.

18 A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, 1 Environmental Economics and Policy 1 (Winter 2007), p. 83. 19 European Commission, Communication from the Commission to the Council and to the European Parliament on the assessment of national allocation plans for the allocation of greenhouse gas emission allowances in the second period of the EU Emissions Trading Scheme, COM(2006) 725 final (November 29, 2006), p. 11. 20 For a further discussion, see Joseph Kruger, Wallace E. Oates, and William A. Pizer, Decentralization in the EU Emissions Trading Scheme and Lessons for Global Policy, 1 Environmental Economics and Policy 1 (Winter 2007), p. 126; and, Frank J. Convery and Luke Redmond, Market and Price Development in the European Union Emissions Trading Scheme, 1 Environmental Economics and Policy 1 (Winter 2007), pp. 96-7, 107. 21 International Emissions Trading Association, IETA Position Paper on EU ETS Market Functioning, (no date), p. 2. 22 Once the EC has approved a countrys NAP, including the total number of allowances and the allocation to each covered entity, the allocations can not be re-visited. Attempts to include provisions permitting such post-approval adjustments to a facilitys allocation have been uniformly rejected by the EC. 23 European Commission, Communication from the Commission to the Council and to the European Parliament on the assessment of national allocation plans for the allocation of greenhouse gas emission allowances in the second period of the EU Emissions Trading Scheme, COM(2006) 725 final (November 29, 2006), p 8; and, A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, 1 Environmental Economics and Policy 1 (Winter 2007), p. 71. 24 On the mixed record of the EU-ETS and the need for allowance scarcity to a functioning emissions market, see Eric Haymann, EU Emission Trading: Allocation Battles Intensifying, Deutsche Bank Research (March 6, 2007). For a generally positive view of ETS market development, see Frank J. Convery and Luke Redmond, Market and Price Development in the European Union Emissions Trading Scheme, 1 Environmental Economics and Policy 1 (Winter 2007), pp. 97-106. For a more negative view, see Karsten Neuhoff, Federico Ferrario, Michael Grubb, Etienne Gabel, and Kim Keats, Emissions Projections 2008-2012 Versus NAPs II, 6 Climate Policy 5 (2006), pp. 395-410.

Congressional Research Service

Figure 1. ECX CFI Futures Contracts: Price and Volume

ECX EUA Futures Contracts: Price and Volume


Total Volume Dec09 Sett

40 35 VOLUME (million tonnes CO2) 30 25

35 30 25 20 Price per tonne (EUR)

20 15 15 10 5 0
06 06 06 06 06 06 06 06 07 07 07 07 07 07 07007007008008 008008008008008008008009009 009009009009009009 009 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 /20 / 6 7 1 2 1 1 1 1 3 3 6 7 5 6 5 6/ 6/ 9/ 9/ 2/ 4/ 4/ 4/ 2/ 3/ 3/ 6/ 6/ 8/ 1/ 2/ 2/ 2/ 2/ 1/ 2/ 3/1 5/ 6/1 7/2 8/3 0/1 1/2 1/ 2/1 3/2 5/ 6/1 7/2 9/ 0/1 1/2 1/ 2/1 4/ 5/1 6/2 8/ 9/1 0/2 12/ 1/1 2/2 4/ 5/2 7/ 8/1 9/2 11/ 2/1 1 1 1 1 1 1

10 5 0

Source: ECX Exchange. Note: Dec09 Sett: Future contracts with a settlement date of December, 2009.

CRS-5

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

While the environmental performance of Phase 1 may be disputed, the need for additional reductions to achieve Kyoto is not. For 2008, the EU-15 is estimated to be 6.2% below its baseyear emissions, compared with an 8% five-year average reduction commitment under the Kyoto Protocol. As indicated by Figure 2, this represents a continuation of reductions by EU-15 over the past five years. However, as indicated by the pink line, the European Environment Agency (EEA) projects that the EU-15 existing measures are insufficient to reduce EU-15 emissions to their Kyoto requirements (represented by the purple line), resulting in a projected 6.9% reduction from baseline levels. To achieve the Kyoto target the EU projects further actions reductions by EU-15 countries (represented by the green line in Figure 2), resulting in an overall reduction of 8.5% compared with baseline levels. Figure 2. EU-15 Greenhouse Gas Emissions and Projections for the Kyoto Period: 1990-2012

Source: European Environmental Agency, Greenhouse Gas Emissions Trends and Projects in Europe 2000, (2009) p. 10. Note: WEM: with existing measures (measures implemented or adopted). WAM: with additional measures (planned measures).

In addition to domestic emission reductions, the EU has also projected additional reduction credits received by activities permitted under the Kyoto Protocol: (1) purchase of project-based credits by ETS participants and EU governments (e.g., Joint Implementation (JI) and Clean Development Mechanism (CDM) projects); and, (2) the use of carbon sinks. As indicated in Figure 3, these activities provide a credit on the EU-15 baseline of 4.6 percentage points. Thus, if the EU-15 maintains its current path, it would exceed its Kyoto commitment by about 3.5 percentage points (6.9% minus 3.4%). If its planned measures result in the projected 8.5%

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

reduction below baseline levels, the overachievement of its Kyoto commitment would be 5.1 percentage points (8.5% minus 3.4%).25 Figure 3. Summary of EU-15 Emissions Projection Compared to Projected Kyoto Protocol Credits

Source: European Environmental Agency, Greenhouse Gas Emissions Trends and Projects in Europe 2000, (2009) p. 11. Notes: The left section shows the projected emissions considering only domestic measures (existing and additional) and is showing them as average 2008-2012 emissions (lines) and annual emissions (bars). The right section shows the projected amount of Kyoto credits accumulated by the end of the commitment period, including the initial EC assigned amount under the Protocol, the purchase of Kyoto project credits by EU ETS participants and EU governments, and carbon sink activities.

The EU-27 as a whole does not have an emissions target comparable to the EU-15 bubble. By 2010, EU-27 emissions are projected at 9.6% below Kyoto baseline levels assuming current policies. This reduction is projected at 11.3% if additional measures are included. Currently, 24 of the 25 countries with reduction requirements are projected to meet their commitments under the Kyoto Protocol. 26 Only Austria is not projected to meet its requirements even with additional planned measures and the use of Kyoto mechanisms. 27

25 26

European Environment Agency, Greenhouse Gas Emissions Trends and Projections in Europe 2009, (2009) p. 11. Cyprus and Malta are not Annex 1 countries. 27 European Environmental Agency, Greenhouse Gas Emission Trends and Projections in Europe 2009 (2009), p. 12.

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

Phase 3
The European Union is committed to achieving a 20% reduction in greenhouse gas emissions by 2020 from 1990 levels (or more depending on the actions of other countries). A strategic component of the effort to achieve this target is a revised ETS that will achieve a 21% reduction from covered entities from 2005 levels. Table 1 indicates the proposed EU-wide ETS cap for the next Phase of EU greenhouse gas program (Phase 3) assuming no further international commitments (the final 2013 cap figure is required by June 30, 2010). As indicated, the EC envisions a linear reduction in the ETS cap to match the reduction target under the overall 20% reduction program. These numbers will change as individual countries decide to include more facilities under the ETS and as the EC expands ETS coverage to include other sectors and nonCO2 greenhouse gases. Table 1. Proposed Annual ETS Cap Figures for Phase 3
Year Annual limit for Kyoto compliance period (2008-2012) 2013 2014 2015 2016 2017 2018 2019 2020 Billion metric tons of CO2e 2.083 1.974 1.937 1.901 1.865 1.829 1.792 1.756 1.720

Source: European Commission, Questions and Answers on the Commissions Proposal to revise the EU Emissions Trading System (Brussels, January 23, 2009), response to question 12. Note: Figures are based on the current Phase 2 scope of the ETS. These need to be adjusted for three reasons: (1) extensions of ETS scope during phase 2 by Member states; (2) extensions of ETS scope by the EC for third trading period, and (3) the figures do not include inclusion of aviation, nor the emissions from Norway, Iceland, and Liechtensteinnon-EU countries that have linked their programs to the ETS.

For Phase 3, the EU is re-shaping the ETS to improve its efficiency and eliminate some of the problems identified during Phase 1 and 2.28 For Phase 2, the improved emissions inventories resulting from Phase 1 allowed the EC to harmonize the types of installations covered by the ETS across the various Member States.29 In addition, the EC imposed a uniform rule on the Member States preventing the use of ex-post adjustments. However, Phase 2 made little advancement in

28 European Commission, Directive 2009/29/EC of the European Parliament and of the Council of 23 April 2009 amending Directive 2003/87/EC so as to improve and extend the greenhouse gas emission allowance trading system of the Community (Brussels, April 23, 2009). Hereinafter referred to as the Directive. 29 European Commission, Limiting Global Change to 2 degrees Celsius: The Way Ahead for 2020 and Beyond (Brussels, January 10, 2007), p. 23.

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

harmonizing individual countries allocations schemes.30 As with Phase 1, countries continue to differ widely on several key points. The critical structural change the EU would make in Phase 3 is eliminating National Allocation Plans (NAPs), and replacing them with EU-wide rules with respect to allowance availability, allocations, and auctions. NAPs are central to the EUs effort to achieve its Kyoto obligations under Phase 2. Each Member of the EU submitted a NAP that lays out its allocation scheme under the ETS, including individual allocations to each affected unit. These NAPs were assessed by the EC to determine compliance with 12 criteria delineated in an annex to the emissions trading directive. 31 Criteria included requirements that the emissions caps and other measures proposed by the Member State were sufficient to put it on the path toward its Kyoto target, protections against discrimination between companies and sectors, and delineation of intended use of CDM and JI credits for compliance, along with provisions for new entrants, clean technology, and early reduction credits. For the second trading period, the NAP had to guarantee Kyoto compliance. This NAP structure will be replaced under Phase 3. There would be one EU-wide cap instead of the 27 national caps under Phase 1 and 2. Allowances would be allocated under EU-wide, fully harmonized rules, including those governing: (1) auctions, (2) transitional free allocations for greenhouse gas intensive, trade-exposed industries, (3) new entrants, and (4) coverage. The EC proposed a Directive to alter the EU-ETS structure for Phase 3 in January, 2008, 32 and the Directive was amended and adopted by the European Parliament (EP) and of the Council of the European Union in April 2009.33

Auctions
Under Phases 1 and 2, allowances generally were and are allocated free to participating entities under the ETS. During Phase 1, The EU-ETS Directive allowed countries to auction up to 5% of allowance allocations, rising to 10% under Phase 2.34 Under Phase 1, only 4 of 25 countries used auctions at all, and only Denmark auctioned the full 5%. The political difficulty in instituting significant auctioning into ETS allowance allocations is the almost universal agreement by covered entities in favor of free allocation of allowances and opposition to auctions.35 Free allocation of allowances represents a one-time transfer of wealth to the entities receiving them from the government issuing them. 36 The resulting transfer of wealth has been described by
30 Joachim Schleich, Regina Betz, and Karoline Rogge, EU Emissions TradingBetter Job Second Time Around? Fraunhofer Institute System and Innovation Research (February 2007), p. 23. 31 Commission of the European Communities, Directive 2003/87/EC, available at http://eur-lex.europa.eu/LexUriServ/ LexUriServ.do?uri=OJ:L:2003:275:0032:0046:EN:PDF. 32 European Commission, Proposal for a Directive of the European Parliament and of the Council amending Directive 2003/87/EC so as to improve and extend the greenhouse gas emission allowance trading system of the Community (Brussels, January 23, 2008). 33 European Commission, Directive 2009/29/EC of the European Parliament and of the Council amending Directive 2003/87/EC so as to improve and extend the greenhouse gas emission allowance trading system of the Community (Brussels, April 23, 2009). 34 For a further discussion of auctioning and the ETS, see Cameron Hepburn, et. al., Auctioning of EU ETS phase II allowances: how and why? 6 Climate Policy (2006), pp. 137-160. 35 A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, 1 Environmental Economics and Policy 1 (Winter 2007), p. 73. 36 Joseph Kruger, Wallace E. Oates, and William A. Pizer, Decentralization in the EU Emissions Trading Scheme and (continued...)

Congressional Research Service

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

several analysts as windfall profits.37 As summarized by Ellerman and Buchner: Allocation in the EU ETS provides one more example that, notwithstanding the advice of economists, the free allocation of allowances is not to be easily set aside.38 Despite concerns about windfall profits and economic distortions resulting from the free allocation of allowances, there was little change in basic allocation philosophy for Phase 2. No country proposed auctioning the maximum percentage of allowances allowed (10%). Most do not include auctions at all.39 The unwillingness of governments to employ auctions as an allocating mechanism revolve around equity considerations, including: (1) inability of some covered entities to pass through cost because of regulation or exposure to international competition; (2) potential drag on a sectors economic performance from the up-front cost of auctioned allowances; and (3) the potential that government will not recycle revenues to alleviate compliance costs, international competitiveness impacts, or other equity concerns, resulting in the auction costs being the same as a tax.40 This opposition is mostly overcome for Phase 3 through an EU-wide set of harmonized rules for allowance allocations and auctions. Under Phase 3, the Directive states:
Auctioning should ... be the basic principle for allocation, as it is the simplest, and generally considered to be the most economic efficient system. This should also eliminate windfall profits and put new entrants and economies growing faster than average on the same competitive footing as existing installations. (paragraph 15)

After nine eastern European Member States threatened to veto an initial proposal to auction 100% of all allowances, the EU compromised to provide for some free allocation of allowances during Phase 3 that will begin in 2013.41 Most covered industries will be eligible for some free allocation of allowances to cover direct emissions under the Phase 3 agreement. The introduction of auction would be differentiated by sector. In general, for the power sector, full auctioning will begin in 2013. For electric powerplants, most will receive no free allocation of allowances during Phase 3. However, in a concession to certain eastern European Member States, an optional and temporary derogation from the no-free-allocation requirement for powerplants is provided to countries that meet specific energy and economic criteria. Under the optional allocation scheme, the Member State can allocate allowances equal to 70% of the powerplants Phase 1 emissions free; this allocation declines to zero in 2020.

(...continued) Lessons for Global Policy, 1 Environmental Economics and Policy 1 (Winter 2007), p. 114. 37 E.g., Deutsche Bank Research, EU Emission Trading: Allocation Battles Intensifying, (March 6, 2007) pp. 2-3; and Regina Betz and Misato Sato, Emissions Trading: Lessons Learnt from the 1st Phase of the EU ETS and Prospects for the 2nd Phase, 6 Climate Policy (2006), p. 353. 38 A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, 1 Environmental Economics and Policy 1 (Winter 2007), p. 85.
39

For more information, see CRS Report RL34150, Climate Change and the EU Emissions Trading Scheme (ETS): Kyoto and Beyond, by Larry Parker.

Martina Priebe, Distributional Effect of Carbon-allowance Trading (Cambridge, January 12, 2007). Also, see Eurochambres, Review of the EU Emission Trading System (June 2007), p. 5. 41 See Position of the European Parliament adopted at the first reading on 17 December 2008 with a view to the adoption of Directive 2009//EC of the European Parliament and of the Council amending Directive 2003/87/EC so as to improve and extend the greenhouse gas emission allowance trading system of the Community (December 17, 2008).

40

Congressional Research Service

10

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

The auction schedule for most other covered entities is more gradual with 80% of a sectors allocation provided free in 2013, declining linearly to 30% by 2020, and zero by 2027. As stated in the final Directive:
For other sectors covered by the Community scheme, a transitional system should be foreseen for which free allocation in 2013 would be 80% of the amount that corresponded to the percentage of the overall Community-wide emissions throughout the period 2005 to 2007 that those installations emitted as a proportion of the annual Community-wide total quantity of allowances. Thereafter, the free allocation should decrease each year by equal amounts resulting in 30% free allocation in 2020, with a view to reaching no free allocation in 2027. (paragraph 21)

Because of concern that stringent EU carbon policies may encourage production and related greenhouse gas emissions to shift to countries without carbon policies (i.e., carbon leakage), exceptions to this phase-out of free allowances will be made in sectors where carbon leakage may occur, as discussed later. Distribution of allowances to be auctioned by the Member States will be determined by a threepart formula (Article 10(2)). Eighty-eight percent of the allowances to be auctioned by each Member State is distributed to States according to their historic emissions under Phase 1 of the EU-ETS. Ten percent of the total is distributed to States mostly based in their comparative GDP per capita within the EU (Annex IIa). Two percent of the total is distributed to nine former eastern-bloc countries based on the substantial greenhouse gas reductions they have already achieved (Annex IIb). Auctions will be conducted at the Member State level (cooperative auctions between States are also allowed) and must be open to any potential buyer. The EC is directed to develop the appropriate rules for coordinated auctions by June 30, 2010. Beyond the allocation of allowances, the EU Directive also provides guidelines for the allocation of revenues from allowance auctions. The Directive states that at least 50% of the proceeds should be used to fund a variety of climate change related activities, including emission reductions, adaptation activities, renewable energy, carbon capture and storage (CCS), the Global Energy Efficiency and Renewable Energy Fund, and assisting developing countries to avoid deforestation and increase afforestation and reforestation (Article 10(3)).

New Entrant Reserves


Unlike previous cap-and-trade programs, the EU-ETS includes provisions for allocating free allowances to new entrants to the system. 42 The reasoning behind this decision is based on equity: (1) it isnt fair to allocate allowances free to existing entities while requiring new entrants to purchase them, and (2) the EU doesnt want to put Member States at a disadvantage in competing for new investments.43 These equity concerns trumped concerns about economic efficiency.

42 For example, the U.S. acid rain program provides no allocation of allowances to new entrants; instead, an EPA sanctioned auction is held annually to ensure that allowances are available to new entrants. New entrants can also obtain allowances from existing sources willing to sell them, either directly, through the EPA auction, or via a broker. 43 A Denny Ellerman and Barbara K. Buchner, The European Union Emissions Trading Scheme: Origins, Allocations, and Early Results, 1 Environmental Economics and Policy 1 (Winter 2007), p. 75.

Congressional Research Service

11

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

As is the case for existing entities, the free allocation of allowances to new entrants is a subsidy. Under Phase 1 and Phase 2, the size and distribution of this subsidy is left to the individual Member States. For Phase 1, the reserve varied widely from the average of 3% of total allowances: Poland set aside only 0.4% of its allocation for new entrants while Malta set aside 26%. For Phase 2, the spread continues with Poland reserving 3.2% of its allowances for new entrants in contrast to 45% reserved by Latvia.44 The decision to employ a new entrant reserve adds complexity to Member States allocation plans and influences the investment decisions of covered entities. Rules had to be promulgated with respect to the reserves size, manner in which the allowances are dispensed, and how to proceed if the demand either exceeds the supply, or vice versa. As indicated, countries did not harmonize new entrant reserve rules with respect to size during Phase 1 or 2. Likewise, there is no standardization on dispensing allowances and replenishing the reserve: first-come, first-serve with no replenishment is one approach used, but a variety of procedures have been developed both to dispense allowances and to replenish the reserve if supply is inadequate. Member States also have different formulas for determining how many allowances a new entrant should receive. Member States claim to use a form of benchmarking to determine allowance allocationsan approach based on a standard of best practices or best technology that is applied to the new entrants anticipated production or capacity. However, the definitions and application of the benchmarks used by the Member States are not uniform. This will change under Phase 3. Under Phase 3, the Directive sets an EU-wide cap of 5% of the total allowance cap for a new entrant reserve, and requires the harmonization of allocation rules. The EC is to adopt a harmonized rule for applying a new entrant definition contained in the Directive by December 31, 2010; the Directive expressly excludes any new electricity production from being defined as a new entrant. The EC is also to determine EU-wide benchmarks for the allocation of all free allowances. The Directive states that the starting point for setting those benchmarks shall be the average performance of the 10% most efficient installations in a sector or subsector in the EU in the years 2007-2008 (Article 10a(2)). In an attempt to stimulate development of CCS, the Directive also provides that up to 300 million allowances in the new entrants reserve shall be available through 2015 for aiding construction and operation of up to 12 demonstration projects. No one project can receive more than 15% of the allowances allocated for this purpose (Article 10a(8)).

EC Phase 3 Decision on Eligible Industries45


Most studies of the competitiveness impacts of the ETS during Phase 1 have found no impact. The International Energy Agency (IEA) cites several reasons for this situation:
Experience to date with the EU-ETS does not reveal leakage for the sectors concerned analysis of steel, cement, aluminum and refineries sectors reveals that no significant changes in trade flows and production patterns were evident during the first phase (2005-2007) of the EU-ETS. This is mostly due to the free allocation of allowances, sometimes in generous
Karoline Rogge, Joachim Schleich, and Regina Betz, An Early Assessment of National Allocation Plans for Phase 2 of EU Emission Trading, Fraunhofer Institute System and Innovation Research (January 2006). For more information on climate change and competitiveness issues, see CRS Report R40914, Climate Change: EU and Proposed U.S. Approaches to Carbon Leakage and WTO Implications, by Larry Parker and Jeanne J. Grimmett.
45 44

Congressional Research Service

12

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

quantities, and to the still functioning long-term electricity contracts, which softened the blow of rising electricity prices. Further, the general boom in prices for most traded products subject to carbon costswhether direct or indirecthas blurred any effects of the latter. Finally, the relatively short time span of these policies does not allow observation of the full potential effects on industry via changes in investment location decisions.46

This conclusion is echoed by Carbon Trust, which states that currently, free allocation of emissions allowances offset almost all of the additional costs of the ETS; and that conclusion is echoed by The Climate Group for The German Marshall Fund, which states that companies surveyed found it difficult to quantify effects on their bottom line in the first phase, or found no effect at all.47 For energy-intensive, trade-exposed industries, Phase 3 has provisions to provide assistance to eligible installations to address the direct and indirect impact of emissions control costs. With respect to direct emissions costs, the EC published a list of installations exposed to a significant risk of carbon leakage on December 24, 2009, as required under the Directive.48 The list is identical to the draft list released in September 2009.49 The decision lists 164 industrial sectors and subsectors deemed to be exposed sectors under the appropriate European Parliament and Council directives. Eligible installations will receive allowances sufficient to cover 100% of their direct emissions, provided they are using the most efficient technology available. This 100% allocation contrasts with the 80% distribution of free allowances to non-carbon leakage exposed industries in 2013. Reflecting the fluid nature of the competitive situation and international negotiations, the EC is to review its decision June 30, 2010, and provide the European Parliament and Council with any appropriate proposals to respond to the situation. Assistance for the impact of indirect emissions control costs on exposed industries from higher electricity prices would be determined by Member States. As stated by the Directive:
Member States may deem it necessary to compensate temporarily certain installations which have been determined to be exposed to a significant risk of carbon leakage related to greenhouse gas emissions passed on in electricity prices for these costs. Such support should only be granted where it is necessary and proportionate and should ensure that the Community scheme incentives to save energy and to stimulate a shift in demand from grey to green electricity are maintained. (paragraph 27)

Flexibility Mechanisms and Price Volatility Control


The major flexibility mechanism developed under the EU-ETS has been the Clean Development Mechanism (CDM) and Joint Implementation (JI) credits permitted under the Kyoto Protocol; however, this development has proven a controversial process. A major part of the controversy
46 47

Julia Reinaud, Issues Behind Competitiveness and Carbon Leakage: Focus on Heavy Industry (October 2008), p. 6. Carbon Trust, EU ETS Impacts on Profitability and Trade (January 2008), p. 4; and The Climate Group, The Effects of EU Climate Legislation on Business Competitiveness; A Survey and Analysis (September 2009), p. 8.

European Commission, Commission Decision of 24 December 2009 determining, pursuant to Directive 2003/87/EC of the European Parliament and of the Council, a list of sectors and subsectors which are deemed to be exposed to a significant risk of carbon leakage (Brussels, 2009). 49 European Commission, Draft Commission Decision of 18 September2009 determining, pursuant to Directive 2003/87/EC of the European Parliament and of the Council, a list of sectors and subsectors which are deemed to be exposed to a significant risk of carbon leakage (Brussels, 2009).

48

Congressional Research Service

13

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

has been the supplementarity requirement of the Kyoto Protocol to use its flexibility mechanisms. Supplementarity requires that developed countries, such as most EU countries, ensure that their use of JI/CDM credits is supplemental to their own domestic control efforts. In defining supplementarity for Phase 2, the EC used 10% of a countrys allowance allocation as a rule of thumb in approving NAPswith a greater limit possible based on a countrys domestic efforts to reduce emissions. This process resulted in some significant reductions in some countries proposed limits (e.g., Ireland, Poland, Spain), but some increase in others (e.g., Italy, Latvia, Lithuania). Although these reductions appear substantial in individual cases, most analysts agree that they do not represent a major barrier to the cost-effective use of JI/CDM. However, the EU-ETS does not accept credits from land use, land-use change and forestry (LULUCF) projects. For Phase 3, the EU maintains its ban on using LULUCF credits within the ETS. However, it will permit up to 50% of the required reductions mandated under Phase 3 to be achieved through CDM or JI credits. For existing installations, this represents a total of 1.6 billion credits over the eight-year compliance period. Limits on use of Kyoto credits will be based on a facilitys 20082012 allocation (for an existing facilities) or its verified emissions during Phase 3 (for a new entrant or sector). The EC estimates that the minimum amount of Kyoto credits an existing facility will be able to use to comply with Phase 3 will be 11% of its 2008-2012 allocation, while new entrants and sectors will be able to use a minimum of 4.5% of their verified emissions during 2013-2020 (article 11a(8)). The precise percentages will be determined later. Another flexibility mechanism, banking, is extended by the Directive from Phase 2 to Phase 3 in order to prevent a Phase 1 style collapse of allowance prices when the ETS transitions into Phase 3. In addition, the EU hopes that extending the trading period from five years to eight years, along with the steady, linear emissions reduction schedule, will increase certainty and stability in the allowance markets. Phase 3 will introduce two other mechanisms designed to address price volatility. First, the EC is required under the Directive to examine whether the market for emission allowances is sufficiently protected from insider dealing or market manipulation. If not, the EC is to present proposals to ensure such protection to the EP and the Council (article 12(a)). Second, the Directive provides that if the allowance price is more than three times the preceding two-year average for more than six consecutive months and the price is not based on market fundamentals, one of two measures may be taken. The first would allow Member States to shift forward the auctioning of some of its auctionable allowances. The second would allow Member States to auction up to 25% of the remaining allowances in the new entrants reserve (article 29(a)).

Expanding Coverage
Despite the ECs interest in expanding the ETS, its coverage in terms of industries included for Phase 2 is essentially the same as for Phase 1. The exception is for aviation. In December, 2006, the EC proposed bringing greenhouse gas emissions from civil aviation into the ETS in two phases. 50 As agreed to by the European Parliament in July 2008, all intra-EU and international
50 European Commission, Proposal for a Directive of the European Parliament and of the Council amending Directive 2003/87/EC so as to include aviation activities in the scheme for greenhouse gas emission allowance trading within the Community (Brussels, December 12, 2006).

Congressional Research Service

14

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

flights will be included under the ETS beginning in 2012. Emissions would be capped at 97% of average 2004-2006 emissions with 85% of the allowances being allocated free to operators. The cap would be reduced to 95% in 2013. The cap and auctioning of allowances would be reviewed as a part of Phase 3 implementation. Annex I of the Directive identifies three CO2 emitting sectors for inclusion under the ETS: petrochemicals, ammonia, and aluminum. The ETS will also expand beyond CO2 to include nitrous oxide (N2O) emissions from nitric, adipic, and glyoxalic acid production, and perofluorocarbon (PFC) emissions from the aluminum sector. This would expand ETS covered emissions by 4.6% over Phase 2 allowance allocations, or about 100 million metric tons.51 The harmonization and codification of eligibility criteria for combustion installations is expected to increase the coverage by a further 40-50 million metric tons. To improve the cost-effectiveness of the ETS and reduce administrative costs, the Directive provides that small installations may be subject to other control regimes (such as carbon taxes) rather than included under the EU-ETS. Currently, the smallest 1,400 (10% of total installations covered) installations emit only 0.14% of total emissions covered. The Directive provides that Member States may opt to exclude installations that emit less than 25,000 metric tons annually from the EU-ETS (paragraph 11).

Summary and Considerations for U.S. Cap-andTrade Proposals


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

Emission Inventories and Target Setting


The ETS experience with market trading and target setting confirms once again the central importance of a credible emissions inventory to a functioning cap-and-trade program. 52 The lack of credible EU-wide data on emissions was a direct cause of the ETS Phase 1 allowance market collapse in 2006. Arguably, the most important result of Phase 1 was the development of a credible inventory on which to base future targets and allocations.

51 European Commission, Proposal for a Directive of the European Parliament and of the Council amending Directive 2003/87/EC so as to improve and extend the greenhouse gas emission allowance trading system of the Community (Brussels, January 23, 2008), p. 4. 52 As stated by CRS in 1992: For an economic incentive system to be effective, several preconditions are necessary. Perhaps the most important is data about the emissions being controlled. Such data are important to levy any tax, allocate any permits, and enforce any limit. CRS Issue Brief IB92125, Global Climate: Proposed Economic Mechanisms for Reducing CO2, by Larry Parker (archived November 16, 1994), p. 9.

Congressional Research Service

15

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

In the United States, Section 821 of the 1990 Clean Air Act Amendments requires electric generating facilities affected by the acid rain provisions of Title IV to monitor carbon dioxide in accordance with EPA regulations.53 This provision was enacted for the stated purpose of establishing a national carbon dioxide monitoring system. 54 As promulgated by EPA, regulations permit owners and operators of affected facilities to monitor their carbon dioxide emissions through either continuous emission monitoring (CEM) or fuel analysis.55 The CEM regulations for carbon dioxide are similar to those for the acid rain programs sulfur dioxide CEM regulations. Those choosing fuel analysis must calculate mass emissions on a daily, quarterly, and annual basis, based on amounts and types of fuel used. As suggested by the EU-ETS experience, expanding equivalent data requirements to all facilities covered under a cap-and-trade program would be the foundation for developing allocation systems, reduction targets, and enforcement provisions.

Coverage
Despite economic analysis to the contrary, the EU decided to restrict Phase 1 ETS coverage to six sectors that represented about 40%-45% of the EUs CO2 emissions.56 This restriction was estimated to raise the cost of complying with Kyoto from 6 billion euro annually to 6.9 billion euro (1999 euro) compared with a comprehensive trading program. A variety of practical, political, and scientific reasons were given by the EC for the decision.57 The experience of the ETS up to now suggests that adding new sectors to an existing trading program is a difficult process. As noted above, a stated goal of the EC is to expand the coverage of the ETS. However, the experience of Phase 1 did not result in the addition of any new sector until the last year of Phase 2 when aviation will be included. The EU will expand its coverage with Phase 3, but the ETS will still cover fewer sectors emitting greenhouse gases than provided under most U.S. proposals. U.S. cap-and-trade proposals generally fall into one of two categories. Most bills are more comprehensive than the ETS, covering 80% to 100% of the countrys greenhouse gas emissions. At a minimum, they include the electric utility, transportation, and industrial sectors; disagreement among the bills center on the agricultural sector and smaller commercial and residential sources. In some cases discretion is provided EPA to exempt sources if serious data, economic, or other considerations dictate such a resolution. A second category of bills focuses on the electric utility industry, representing about 33% of U.S. greenhouse gases and therefore less comprehensive than the ETS. Sometimes including additional controls on non-greenhouse gas pollutants, such as mercury, these bills focus on the sources with the most experience with emission trading and the best emissions data. Other sources could be added as circumstances dictate.

53 54

Section 821, 1990 Clean Air Act Amendments (P.L. 101-549, 42 USC 7651k). S.Rept. 101-952. 55 See 40 CFR 75.13, along with appendix G (for CEMs specifications) and appendix F (for fuel analysis specifications.
56

For more background, see CRS Report RL33581, Climate Change: The European Unions Emissions Trading System (EU-ETS), by Larry Parker. 57 Ibid., p 3.

Congressional Research Service

16

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

As noted, the EUs experience with the ETS suggests that adding sectors to an emission trading scheme can be a slow and contentious process. If one believes that the electric utility sector is a cost-effective place to start addressing greenhouse gas emissions and that there is sufficient time to do the necessary groundwork to eventually add other sectors, then a phased-in approach may be reasonable. If one believes that the economy as a whole needs to begin adjusting to a carbonconstrained environment to meet long term goals, then a more comprehensive approach may be justified. The ETS experience suggests the process doesnt necessarily get any easier if you wait.

Allocation Schemes
Setting up a tradeable allowance system is a lot like setting up a new currency. 58 Allocating allowances is essentially allocating money with the marketplace determining the exchange rate. As noted above, the free allocation scheme used in the ETS has resulted in windfall profits being received by allowance recipients. As stated quite forcefully by Deutsche Bank Research:
The most striking market outcome of emissions trading to date has been the power industrys windfall profits, which have sparked controversy. We are all familiar with the background: emissions allowances were handed out free of charge to those plant operators participating in the emissions trading scheme. Nevertheless, in particular the producers of electricity succeeded in marking up the market price of electricity to include the opportunity-cost value of the allowances. This is correct from an accounting point of view, since the allowances do have a value and could otherwise be sold. Moreover, emissions trading cannot work without price signals.59

The free allocation of allowances in Phase 1 and 2 of the ETS incorporates two other mechanisms that create perverse incentives and significant distortions in the emissions markets: new entrant reserves and closure policy. Combined with an uncoordinated and spotty benchmarking approach for both new and existing sources, the result is a greenhouse gas reduction scheme that is influenced as much or more by national policy than by the emissions marketplace. The expansion of auctions for Phase 3 of the ETS could simplify allocations and permit market forces to influence compliance strategies more fully. Most countries did not employ auctions at all during Phase 1 and auctions continue to be limited under Phase 2. No country combined an auction with a reserve price to encourage development of new technology. The EC limited the amount of auctioned allowances to 10% in Phase 2: a limit no country chose to meet. Efforts to expand auctions met opposition from industry groups, but attracted support from environmental groups and economists. The Phase 3 increased use of auctioning through 2020 will represent a major development for the scheme. Currently, all U.S. cap-and-trade proposals have some provisions for auctions, although the amount involved is sometimes left to EPA discretion. Most specify a schedule that provides increasing use of auctions from 2012 through the mid-2030s with a final target of 66%-100% of total allowances auctioned. Funds would be used for a variety of purposes, including programs to encourage new technologies. Some proposals include a reserve price on some auctions to create a price floor for new technology.

58 59

Unlike a carbon tax which uses the existing currency system to control emissionsbe it euro or dollars. Deutsche Bank Research, EU Emission Trading: Allocation Battles Intensifying (March 6, 2007), p. 2.

Congressional Research Service

17

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

Like the situation in the ETS, most U.S. industry groups either oppose auctions outright or want them to be supplemental to a base free allocation. Given the experience with the ETS where the EC and individual governments have been unwilling or unable to move away from free allocation, the Congress, like the EU, may ultimately be asked to consider specifying any auction requirement if it wishes to incorporate market economics more fully into compliance decisions.

Flexibility and Price Volatility


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

Congressional Research Service

18

Climate Change and the EU Emissions Trading Scheme (ETS): Looking to 2020

particularly with the inclusion of financial instruments such as options and futures contracts. The concern is sufficient for the Directive to require the EC to examine the situation and the current protections against such activities. Congress may ultimately be asked to consider whether the Securities and Exchange Commission, Federal Energy Regulatory Commission, the Commodities Futures Trading Commission, or other body should have enhanced regulatory and oversight authority over such instruments. 61

Author Contact Information


Larry Parker Specialist in Energy and Environmental Policy lparker@crs.loc.gov, 7-7238

61 For a discussion of regulation of allowances as a commodity and implications for a greenhouse gas emissions market, see CRS Report RL34488, Regulating a Carbon Market: Issues Raised By the European Carbon and U.S. Sulfur Dioxide Allowance Markets, by Mark Jickling and Larry Parker.

Congressional Research Service

19

Project Finance Primer for Renewable Energy and Clean Tech Projects
Authors: Chris Groobey, John Pierce, Michael Faber, and Greg Broome

Executive Summary
Investments in the clean technology sector often combine capital intensity with new technologies. Securing project finance can prove to be a critical step in the path to commercialization. Project finance succeeds best when you have long-term off-take agreements with quality-credit counterparties (such as power purchase agreements) but commoditybased projects that sell into open markets (such as biofuels) can also benefit from the project finance model. This primer provides an overview of project finance for renewable energy investors, with a focus on the pros and cons, as well as a survey of key concepts and requirements, including tax incentives and monetization strategies in the renewable energy sector, and other key structuring considerations in determining whether to project finance.

Key Points Project finance has emerged as a leading way to finance large infrastructure projects that might otherwise be too expensive or speculative to be carried on a corporate balance sheet. The basic premise of project finance is that lenders loan money for the development of a project solely based on the specific projects risks and future cash flows. As such, project finance is a method of financing in which the lenders to a project have either no recourse or only limited recourse to the parent company that develops or sponsors the project. For equity investors, the appeal of project finance is that it can maximize equity returns, move significant liabilities off balance sheet, protect key assets and monetize tax financing opportunities. A wide range of commercial and legal issues must be addressed to secure adequate returns. Tight credit markets exacerbate competition for long-term financing, so even small differences in deals can impact the availability of financing or reduce leverage. Project financing became particularly important to project development in emerging markets, with participants often relying on guarantees, long-term off-take or purchase agreements, or other contractual relationships with the host sovereign or its commercial appendages to ensure the long-term viability of individual projects. These were typically backstopped by multilateral lending agencies that mitigated some of the political risks to which the project lenders were exposed. Analogies to alternative energy projects help investors de-risk higher-risk new technologies.

Austin

New York

Palo Alto

San Diego

San Francisco

Seattle

Shanghai

Washington, D.C.

Part I of the primer introduces project finance to those that may be less familiar with the concept, and asks questions that will assist investors and developers in determining whether project finance is appropriate for their renewable energy projects. Part II sets out the legal and contractual structure that will facilitate project financing. Part III describes the process of obtaining equity investment and some of the important options and considerations that companies may have in that process. Part IV provides a more in-depth look at what a typical renewable energy project financing looks like, including fundamental structural components that characterize any project finance transaction. Finally, Part V outlines key tax incentives currently available in the renewable energy industry, as well as monetization strategies that may be useful for earlier-stage energy companies unable to directly utilize such tax incentives. Given the breadth of the current renewable energy landscape, this primer focuses on a hypothetical solar generation facility (Solar Project) as the primary case study with discussions of other renewable energy projects (wind power and biofuel projects in particular) as appropriate. In general, once the contracts related to a project are negotiated (which is described in Part II), the mechanical aspects of raising equity and project financing are likely to be similar across various renewable technologies, although investor enthusiasm and financing prices and terms are likely to vary significantly across technologies at any given time.

I. Introduction to Project Finance


A. What Is Project Finance? The basic premise of project finance is that lenders loan money for the development of a project solely based on the specific projects risks and future cash flows. As such, project finance is a method of financing in which the lenders to a project have either no recourse or only limited recourse to the parent company that develops or sponsors the project (the Sponsor). Non-recourse refers to the lenders inability to access the capital or assets of the Sponsor to repay the debt incurred by the special purpose entity that owns the project (the Project Company). In cases where project financings are limited recourse as opposed to truly non-recourse, the Sponsors capital may be at risk only for specific purposes and in specific (limited) amounts set forth in the project financing documentation. Project financing has been used in various ways for many years, but in the 1970s and 1980s it emerged as a leading way of financing large infrastructure projects that might otherwise be too expensive or speculative for any one individual investor to carry on its corporate balance sheet. Project financing has been particularly important to project development in emerging markets, with participants often relying on guarantees, long-term off-take or purchase agreements, or other contractual relationships with the host sovereign or its commercial appendages to ensure the long-term viability of individual projects. These were typically backstopped by multilateral lending agencies that mitigated some of the political risks to which the project lenders (and, sometimes, equity investors) were exposed. B. What Underpins Project Finance?

As a general (if not universal) rule, lenders will not forgo recourse to a projects Sponsor unless there is a projected revenue stream from the project that can be secured for purposes of ensuring repayment of the loans. In the case of large wind and solar power projects, this revenue is typically generated from a power purchase agreement (PPA) with the local utility, under which the project may be able to utilize the creditworthiness of the utility to reduce its borrowing costs. While the wind power market has matured significantly in the past five years, leading to the successful project financing of merchant projects in the absence of long-term PPAs, Solar Projects are generally not yet able to be project financed in such a manner. In merchant power projects, lenders are able to receive assurance of the projects ability to repay its debt by focusing on commodity hedging, collateral values, and the income to be produced based on historical and forward-looking power price curves and fully developed markets. In non-power generation contexts, the projects revenue stream may be a long-term operating agreement (e.g., in the case of toll roads), a capacity purchase agreement (e.g., in the case of transmission lines), a production sharing agreement (e.g., in the case of oil field development), or a series of short-term and spot sales into commodity markets (e.g., in the case of biofuels projects). While project finance lenders clearly prefer a long-term contract that ensures a relatively consistent and guaranteed revenue stream (including assured margins over the cost of inputs), in the context of some industries, lenders have determined that sufficient revenues to support the projects debt are of a high enough probability that they will provide debt financing without a long-term off-take agreement. Solar Projects, due to their peak period production, high marginal costs, and lack of demonstrated merchant capabilities, are not at this time viewed as project financeable without PPAs that cover all or substantially all of their output. Solar Projects lack of merchant viability is exacerbated by the fact that the southwest United States (the region most appropriate for utility-scale solar power development) does not have a

www.wsgr.com

mature merchant power market that functions in the absence of long-term bilateral sales agreements. The dependence of large-scale solar projects on the PPA model is not expected to change in the short to intermediate term. C. When to Project Finance?

One of the primary benefits of project financing is that the debt is held at the level of the Project Company and not on the corporate books of the Sponsor. When modeling projects and projected income, the internal rate of return of Sponsors and other project-level equity investors can increase dramatically once a project is fully leveraged. Sponsors are frequently able to recover development costs at the closing of the project financing and put their money into other projects. Another benefit of project financing is the protection of key Sponsor assets, such as intellectual property, key personnel, and investments in other projects and other assets, in the case of the Project Companys bankruptcy, debt default, or foreclosure. Moreover, project financing allows for a wide variety of tax structuring opportunities, particularly in the context of monetizing tax incentives (discussed further in Part V). On the other hand, project financing is document-intensive, time-consuming, and expensive to consummate. It is not atypical that administrative and closing costs, when factoring in lenders, consultants, and attorneys fees for all parties, equal several percentage points of the amount of the loan commitment. Moreover, project financing imposes significant operating restrictions on each Project Company, including its ability to make equity distributions to the Sponsor prior to the payment of operating expenses, debt service, and a percentage sweep of additional cash flow (discussed further in Part IV). The result is that the decision of whether to reinvest cash flow in the project does not rest solely with the Sponsor. Given the pros and cons of project finance, the most relevant initial inquiry for an investor or developer may be when is project financing possible or most appropriate? The following questions should be useful in determining if project financing is a realistic opportunity for any given company:
-

Is there an individual project or group of projects of a sufficient size to make either a standalone or portfolio project financing worthwhile? Typically lenders will be reluctant to provide project financing if the total amount of debt is less than US$50 million and, preferably, US$100 million. Will there be a revenue stream from the project large enough to support a highly leveraged debt financing? This is a prerequisite for project financing. Will the receipt of revenue be enforceable under contractual rights against a creditworthy party? This is not necessarily a prerequisite for all project financings, but the absence of a contract, or questionable creditworthiness of the purchaser, will prompt lender skepticism and necessitate thorough due diligence regarding future revenue projections. Will there be physical assets sufficient to ensure lender repayment in case of foreclosure? Lenders will want to know that even if the Project Companys projected revenue stream does not materialize, they will be able to foreclose on the projects assets sufficient in value to make themselves whole, either by selling the project outright or operating it until the debt is repaid. Is there a significant level of technology risk? While in many project financings, technology may be relatively new or cutting edge, project finance lenders almost never want to be the first to finance an untested technology. Demonstrated successful use in some context will often be necessary to secure project financing. Does the project have contractual relationships with reputable companies for services key to the success of the project or the technology it employs? Lenders will be less likely to lend to a project the success of which depends solely on a few talented individuals who may depart, leaving the project unable to meet its potential. Is the Sponsor ultimately willing to risk the project? In other words, once project financing is completed, the Sponsor loses the ability to determine how the vast majority of the projects revenue is spent. In the event a project becomes uneconomic and unable to service its debt, the only option besides refinancing the debt may be to turn over the project to the lenders (voluntarily or involuntarily), with the corresponding loss of the Sponsors investment in the project. Is the Sponsor looking for a quick exit? Once project-financed, divestiture opportunities are complicated by the requirement of lender consent, and potential purchasers will be thoroughly examined by lenders for development and operational expertise as well as creditworthiness.

www.wsgr.com

Are Sponsors willing to grant rights of high-level oversight regarding the projects development and operation to project finance lenders? In many cases the interests of the Sponsor and the lenders will be aligned, and lenders will tend to defer to the Sponsors developmental expertise. On the other hand, lenders must be viewed as additional project partners, with veto rights over many significant decisions.

Assuming project financing is a viable option, Part II provides a roadmap to structuring a project financing transaction.

II. Establishing a Project Structure and Negotiating Project Agreements


A. Project Structure The project finance structure revolves around the creation of the Project Company that holds all of the projects assets, including all of its contractual rights and obligations. The Project Company is usually a single-member limited liability company, although in some cases it may be a limited partnership. In most cases, the equity interest in the Project Company will be held by at least one intermediate holding company, usually a limited liability company (the Holdco), created for the purpose of pledging the Project Companys equity to the lenders in the eventual project financing. While the Holdco will have a separate legal identity, typically it will not have any business apart from holding the equity of the Project Company. This structure allows for most liability to be contained at the bankruptcy-remote Project Company level, and thus insulates the Sponsor (including equity investors in the Sponsor) and the Holdco from liability to either the Project Companys contractual counterparties (Counterparties) or to the Holdcos lenders. In order to ensure that the Project Company is treated as a separate legal entity, it will be necessary to have governance mechanisms at the Project Company level that are independent, including designated officers, at least one independent director, and internal controls and procedures designed to preserve a legal entity distinct from the Sponsor and the Holdco. B. Project Agreements Overview

As a general matter, all contracts related to the development, construction, ownership, and operation of the project will be entered into by the Project Company (Project Agreements). If development-stage contracts have been executed by the Sponsor or one of its affiliates, it is important that the contracts allow for their assignment to the Project Company once the Project Company has been established for the purposes of pursuing project financing. In addition to the external Project Agreements, there may be several intercompany agreements between the Project Company and the Sponsor or its affiliates. These may include an Operation and Maintenance Agreement (O&M), an Administrative Services Agreement (ASA), and a Technology License Agreement (TLA), often with affiliates of the Sponsor created specifically for the purpose of providing administrative support, operation, and maintenance services and holding the intellectual property for the benefit of one or more of the Sponsors projects. In other cases, unrelated third parties may provide these services to the Project Company. If intercompany agreements are used, they should be structured in such a manner as to track the material commercial terms that the Sponsor could obtain with an unrelated third party providing the same services. Intercompany agreements can also have a significant impact on the total return of a project to its investors, so their economic terms must be carefully crafted. Assuming the O&M, ASA, and TLA are entered into with Sponsor affiliates, they permit the affiliates to extract arms length fees for the provision of key services and technology to the Project Company on a monthly or quarterly basis; these fees are frequently paid prior to repayment of debt. The intercompany-agreement structure also allows the Sponsor, if the project fails following the project financing, to retain all of its employees that provide services to the Project Company, thereby ensuring that key employees (and knowhow) will not be lost to lenders or a subsequent purchaser out of foreclosure. In such a scenario, the TLA will also allow the Sponsor to retain ownership of its technology subject only to a license right on the part of the Project Company which may no longer be affiliated with the Sponsor. These are especially critical points where the Sponsor has multiple projects that may utilize the same technology, support equipment, and personnel. In addition, the O&M, ASA, and TLA provide the Project Company's lenders contractual certainty (through the agreements themselves as well as the corresponding consents to collateral assignment (discussed further in Part IV below)) that key services will continue if the Project Company defaults, thereby increasing the likelihood of the efficient development, construction, and operation of the project and the preservation of the value of the lenders' collateral. There are many other Project Agreements that are typically executed during the course of developing and constructing a renewable energy project. The Project Agreements may include one or more PPAs, which may have an income

www.wsgr.com

stream payable from an off-taker for energy payments, capacity payments, or both; an Engineering, Procurement, and Construction Agreement (EPC Agreement); a Site Lease Agreement (if the projects land is not owned by the Project Company itself); a Renewable Energy Credit Agreement (in states where applicable); an Interconnection Agreement (for projects tied to the electricity grid); agreements for the provision of utility services; agreements for the provision of feedstock commodities (in the case of biofuels) and the necessary price and supply hedging; agreements including equity flip structures to take advantage of the federal tax incentives discussed in Part V below; and other Project Agreements necessary or desirable to develop, construct, own, or operate the project. In some cases certain byproducts of production may be sold in addition to the primary product (for example, steam as a byproduct of cogeneration power projects, high protein distillers grains as a byproduct of ethanol production, or carbon dioxide where markets exist).

Typical Project Finance Structure


Equity Investors

Equity Contribution $

Operating or Shareholders Agreement

Sponsor

Project-Level Equity Investors

Lenders

Equity Contribution $

Operating or Limited Partnership Agreement

Equity Contribution $

Operating or Limited Partnership Agreement

Loans $

Loan Documentation

Project Company aka The Borrower Interconnection Agreement Interconnecting Utility EPC or EPCM Agreement REC Purchase Agreement Admin Services Agreement Technology License Agreement

PPA

Supply Contracts

O&M Agreement

Purchasing Utility/ Offtaker

Contractor

Equipment Suppliers

REC Purchaser

O&M Service Provider

Admin Services Provider

Technology Provider

Consent

Consent

Consent

Consent

Consent

Consent

Consent

Consent

Lenders

Lenders

Lenders

Lenders

Lenders

Lenders

Lenders

Lenders

C.

Key Project Agreement Terms

In the process of negotiating the Project Agreements it will be necessary to consider key project finance principles to prevent having to revisit contractual terms at the lenders behest in the course of financing the project. One overriding concept is that lenders will own (and likely seek to immediately transfer) the Project Company in the case of foreclosure, thus will insist on contractual rights and terms that ensure a seamless transition to the lender or subsequent owner. To this end, the project lenders will require consents to collateral assignment (Consents) for their benefit with some if not all of the Counterparties. Therefore, provisions that prevent assignment without Counterparty consent should be omitted from Project Agreements. Inclusion of contractual language that obligates the Counterparty to cooperate with the Project Company and its lenders in the course of the financing process will not only expedite the process of negotiating the Consents but will also reduce the scope for Counterparty intransigence in the context of the project financing. The commercial terms of the PPA and the EPC Agreement, together with the market and technology risks, will largely determine whether lenders view the project as financeable. Foremost among considerations related to the PPA will be whether or not there is a guaranteed revenue stream (usually energy payments from the actual production of power) from a creditworthy purchaser that will be sufficient to support the economics of the project, thereby ensuring

www.wsgr.com

prompt repayment of debt and mitigating the risk of default. The PPA term should also be sufficient in length to fully amortize the contemplated project debt. In contrast to most smaller distributed generation projects, where an off-taker pays for only the power that is produced, utility scale solar generation facilities may have take or pay PPAs where the utility is still required to pay the Project Company even if a certain level of power is not purchased. In the case of Solar Projects, utilities are less likely to deliver capacity payments because power is generally produced only during daylight hours. However, distributed generation projects may benefit from more certain payments because the distributed project produces all of the power for a given host. If a project does not have a PPA or other off-take contract, demonstrated merchant operating histories of similarly situated plants (more relevant in the context of wind projects) will be necessary to convince lenders of the reliability of forecast ratios. Even with long-term PPAs, lenders will still look for additional data to support viability such as meteorological wind data for wind power sites over the course of one to two years, often at installed hub-heights, or long-term temperature and sun data for Solar Projects. The trend in biofuels project financings is also moving toward contracted off-take arrangements with a creditworthy purchaser for all of a plants production. At least in the short and intermediate terms, most project-financed Solar Projects will have PPAs for a significant portion, if not all, of their generated power. While PPAs for large-scale CSP projects will generally be far more complex than those for smaller distributed PV projects, in either case, the core economic terms will determine the lenders view of whether a project or a portfolio of projects is viable from a revenue perspective and, accordingly, financeable on favorable terms. To the extent a project is not fully constructed by the time project financing is sought, EPC Agreements will be an integral part of the financing analysis and pricing. While larger developers may be able to finance an entire project on balance sheet, and subsequently refinance the development to free up invested capital, most developers seek to leverage their equity and use project finance to construct and operate their projects. Where construction risk is present, lenders will generally seek corporate parent guarantees, performance bonds, or other forms of performance surety that ensure that the performance of the contractor is as close to budget and schedule as possible. Warranties of appropriate substance and duration as well as subsequent maintenance coverage regarding the EPC work and the equipment purchased will be necessary to convince lenders that significant unbudgeted expenses will not be incurred by the Project Company. With respect to an EPC contractor, lenders prefer a full wrap EPC agreement because such an agreement provides a single point of contact with regard to the various risks such an agreement might contain (warranties and schedule and performance guarantees, among other things). This is particularly the case with newer and untested technology even if operationally superior to previous generation technology. Liquidated damage coverage (pre-agreed payments made by the contractor) for schedule and performance delays, inefficiency, or equipment failures also reassure lenders that a project has the necessary protection against delays or performance defects that are within the EPC contractors control. How much of the risk an EPC contractor accepts for cost overruns and design or installation defects, when viewed with other contractual terms, will affect the lenders view of whether a project is financeable and at what cost. For example, a project that is not financeable at 80% debt due to certain offtake or technology risks may be financeable with 40%60% debt because the lenders are taking less risk with a higher level of capital pre-paid into the project. Dedicating sufficient resources at the negotiation stage of PPA and EPC Agreements to achieve commercial and contractual terms as favorable as possible will usually pay dividends at the financing stage by saving not only money but also costly renegotiation and valuable time toward project completion.

III. Raising Equity


Venture capital and private equity investors also serve as attractive sources for capital raising, as an increasing number of funds are investing in renewable energy and clean technologies. The large sums of capital required to initiate and complete renewable energy projects drive not only the selection of appropriate equity investors but also the structure of such investments. Therefore, Sponsors and investors evaluating equity solar investments should consider the following action items: A. Conduct an internal assessment of capital and budgeting strategies for the investment.

Sponsors and investors should conduct an internal evaluation of whether an equity investment would best serve their respective strategic objectives. Salient considerations include:
-

Is the companys technology reliable enough to be considered financeable, and is there a realistic potential pool of equity investors from which to draw? Are there any gaps in the Sponsors existing organizational structure and operations that an equity investor would want filled before engaging in substantive discussions and/or closing an equity round of

www.wsgr.com

financing? For example, given the complexity involved in successfully executing Solar Projects, equity investors will look for experienced management with skill and connections within the industry, as well as potentially requiring contractual commitments from relevant third parties in the supply chain and customer base.
-

How much capital investment is realistically required for the Solar Project? Has Sponsor management, on the one hand, conducted a thorough analysis of the timing and amount of future capital needs and relevant burn rates, and has the investor and its syndicate, on the other hand, assessed whether its proposed financing will be sufficient to either execute the Solar Project or bridge the Sponsor towards its next round of investment? What type of investment is ideally suited for the particular Solar Project e.g., is the Sponsor seeking passive investment, or an active strategic partner that will add value to the organization (as discussed in greater detail below)? How would an equity investment impact the Solar Projects existing grants, tax treatment, eligibility for applicable federal and state incentive programs, and contractual obligations? As equity investors become shareholders, and in many cases, directors of the Sponsor, how much control is the Sponsor willing to give to the investor, and what level of control does the investor desire in order to have an active voice within the organization?

B.

Determine whether the investment will add value to the Solar Project.

Unless a Sponsor is seeking a purely passive equity investment, the Sponsor and its investor should conduct a thorough assessment of the investors role in driving value to the enterprise by, among other things:
-

Reviewing the investors existing portfolio companies to determine whether the investor has previously invested in similar projects or has other relevant experience with alternative energy investments. It is important to find investors who understand the longer time period required to execute and obtain a return on investment from renewable energy projects; Meeting with the investors key decision makers to assess how the investor will add value in addition to the capital infusion e.g., through participation on the board of directors, introductions to potential customers, assistance in financial forecasting and planning, and guidance in analyzing potential liquidity events; and Assessing potential conflicts of interest that may arise to the extent that an investor has, for example, a competitor as one of its portfolio companies.

C.

Assess the appropriate structure for the equity investment.

Once an appropriate equity investor has been identified, the equity investment typically will proceed to the preparation of a term sheet that identifies the key terms of the investment, as well as a diligence request and the execution of a confidentiality agreement to facilitate the exchange of information to the investor for the investors due diligence purposes. A term sheet is a helpful means of assessing whether the parties truly see eye-to-eye with each other on the critical aspects of the investment before expending significant time and expense negotiating definitive documents, and may include the following terms:
-

The identification of the relevant entity that will receive such funds (e.g., will the investment be made into a special purpose vehicle solely created for the project (for example, a Project Company) or will the investment be made into the Sponsor which may hold assets unrelated to the project); The amount of the investment, as the Sponsor should ensure that it receives sufficient capital to minimize future dilutive cram-down financings, but also not take in more capital than is needed as this also will have dilutive effects to existing shareholders. Milestone-based investments may help serve to mitigate Sponsor risk in terms of securing additional future financing, while helping investors stage their investment to ensure that the Sponsor can meet specific financial and commercial targets before disbursing additional funds; and Whether the equity security will be common stock, which is typically issued to founders, optionees, and angel investors, or preferred stock, which not only is senior to the common stock in preference but also typically has additional terms and conditions that increase preferred stockholders return on investment and control over the Sponsor such as:

www.wsgr.com

D.

The right to appoint one or more board members; Dividend rights; A liquidation preference, which is the right to receive a preferential return on investment in the event of a liquidity event such as a merger, asset sale, or change of control; A redemption right, which is the right to redeem the equity securities at an agreed-upon point in the future; Anti-dilution rights, which protect an investor from the dilutive effect of future equity issuances; and Protective provisions, which allow the preferred holders certain veto rights over key corporate actions.

Have candid discussions to ensure that expectations are aligned on key business issues, including:
-

The market opportunity; The companys ability to execute its business plan, and the investors commitment to both the initial and subsequent capital needs during the companys life cycle; A realistic commercialization timeline, use of proceeds, and the expected internal rate of return of the Solar Project; and The appropriate liquidity event, be it an acquisition or an initial public offering, and how the investor can add value to facilitate a liquidity event (e.g., by assisting in pre-public corporate governance compliance required by Sarbanes-Oxley, or introductions to key strategic partners, customers, and potential acquirers in the future).

In summary, both Sponsors and investors analyzing an equity investment should conduct a realistic assessment of the companys capital needs, structure an investment that can add value to the company and its projects, and seek to create a mutually beneficial working relationship where expectations on key business issues between the Sponsor and the investor are aligned.

IV. Time to Project Finance


Before beginning an examination of the project financing process, it is worth noting different options for raising debt in the context of project development. The three most frequently utilized project financing structures are the syndicated or club loan, the issuance of project bonds through a private placement, and the issuance of Term B loans. A. Syndicated and Club Loans and Project Bonds

Currently the majority of renewable energy projects are financed through the syndicated commercial loan market. Syndicated loans are loans in which a group of banks each take a portion of a larger loan and thus minimize the risk that any one individual lender making the same loan would otherwise have. A syndicated loan transaction is usually coordinated by one or more arranger banks whereas in club deals a handfull of lenders take equal roles in leading the transaction and lending to the project. An alternative to the syndicated loan market is the private placement of debt through 144A offerings, which are exempt from registration with the SEC if the purchasers are Qualified Institutional Buyers as defined in the Securities Exchange Act of 1933. The issuer of 144A bonds could be either the Project Company or the Sponsor. Syndicated loan structures are often preferred to accessing the capital markets through 144A offerings, because capital markets investors are generally less likely to assume construction risk and the disclosure documentation for a 144A offering is generally more extensive than that prepared in connection with syndicating a commercial loan. In addition, amounts raised through a 144A issuance are all disbursed at closing, which leads to negative carry implications. Moreover, private placements or corporate level offerings tend to be fixed rate, which, while providing certainty, removes the upside potential of floating rates that are available pursuant to commercial bank loans. On the other hand, 144A bond offerings are generally completed more quickly and inexpensively than a syndicated project loan, the covenants contained in the governing documentation may be less restrictive, and the repayment period of

www.wsgr.com

private placement debt offerings is generally longer. Bonds can also pay interest at tax-exempt rates (lowering the borrowers borrowing cost), be issued in relatively small amounts (making them ideal for smaller project financings) and carry implied or explicit credit support from government instrumentalities (again reducing borrowing costs). B. Term B Loans

Several years ago, Term B loans emerged as a subset of the project lending market and were characterized by shorter tenors and lower or delayed amortization, often with bullet payments due at maturity. Correspondingly, Term B loans carried higher risk profiles and usually were rated non-investment grade. In addition, the terms and conditions of Term B loans tended to be less onerous than traditional project debt that amortized over a longer period. As a result of the subprime lending crisis and the resulting credit crunch, the Term B loan market all but disappeared and has yet to reemerge. For purposes of the following discussion, due to the considerations set forth above and the lull in the Term B market, we assume that a traditional bank syndication model of project financing will be most beneficial to the Sponsor. Although the terminology may differ from transaction to transaction, the documentation for such a project financing is governed by a credit or financing agreement (Credit Agreement) and at a minimum will include an asset security and equity pledge agreement, a mortgage, and various Consents. C. Loan Types

Depending on the development stage of the project, and within the project finance framework, the Sponsor may on behalf of the Project Company seek construction loans, term loans, working capital loans and/or a letter of credit facility. Construction loans, as the name implies, are utilized only for the period that the project is under construction. The interest rate can be higher vis-a-vis a term loan (reflecting increased risk to the lenders during the construction period) but more frequent drawdowns of construction loans are permitted and at the end of the construction loan availability period, the construction loan usually converts to a term loan. Term loans are characterized by a set and limited commitment or drawdown period and an extended amortization period. Term loans can have a lower interest rate than construction loans, and have scheduled (quarterly or otherwise) repayment dates or set amortization schedules. The conversion from a construction loan to a term loan often coincides with the definition of Substantial Completion or Final Completion under the EPC Agreement, and a failure to achieve such conversion by a certain date will cause a default under the construction loan and accelerate the debt due thereunder. Working capital loans, which are used primarily for ordinary course expenses such as inventory purchases, are generally sized smaller than construction or term loans and are subject to a maximum available amount tied to the value of a Project Companys inventory and cash (often 80%). Working capital loans are usually revolving in nature, meaning that amounts borrowed can be reborrowed once they are repaid. Letters of credit are made available on the Project Companys behalf usually for the benefit of third parties under the Project Agreements for example, if a letter of credit is required as credit suppo