Whitepaper. Electricity Markets and Variable Generation Integration

Transcription

1 Whitepaper Variable Generation Subcommittee Marketing Workgroup Electricity Markets and Variable Generation Integration January 6, 2011 Jack Ellis, Resero Consulting Brenda Anderson, Bonneville Power Administration Paul Arnold, ColumbiaGrid Jamie Austin, PacifiCorp Ty Bettis, Portland General Electric Peter Blood, Columbia Energy Partners Tom Cooper, Salt River Project Mike Evans, Shell Energy North America Wally Gibson, Pacific Northwest Planning Council Matthew Hunsaker, WECC Brendan Kirby, NREL Gary Lawson, SMUD David Lemmons, Xcel Michael Milligan, NREL Michelle Mizumori, WECC Steve Norris, APS Farrokh Rahimi, OATI David Schiada, Southern California Edison Jagjit Singh, OATI

2 Table of Contents 1. Introduction Definition of Variable Generation (VG) System Balancing What Policy Objectives Are Achieved By Integrating VG? Supporting State and Federal Renewable Energy Policies Minimizing the Cost of Achieving High Levels of Renewable Generation Penetration What Are the Challenges of Integrating VG? VG Increases the Combined Variability of Load and Generation VG Output is Difficult to Accurately Predict and Control Current Operating, Scheduling and Business Practices Lack Flexibility Federal Renewable Energy Tax Incentives Favor Production Commercial Terms and Conditions Discourage Curtailment How a Resource Flexibility Market Could Address VG Challenges Current Impediments to Realizing Additional Benefits from Market-Based Solutions Administrative Constraints Lack of Consistent, Well-Defined Service Requirements for VG Integration Transmission Rate Pancaking Transmission Limitations Between BAs Overcoming Barriers Develop Market Products that Meet the Needs of VG Integration Re-Examine Operating Practices Develop Revenue-Sharing and Other Tariff Mechanisms That Eliminate Transmission Rate Pancaking Standardize Business Practices Current Voluntary Efforts...14 WECC Electricity Markets and Variable Generation Integration 2

3 10. Recommendations Technical Appendix Relevant Characteristics of VG for System Operators System Operation with Wind Efficient Integration of VG Wind Generation and Efficient Power System Integration Aggregation and Large Balancing Area Size The Value of Energy Markets Load Following as a Byproduct of Fast Energy Markets or the Economic Dispatch Stack Market Evidence from New York Independent System Operator (NYISO) Data The Benefits of Large Balancing Areas and Sub-Hourly Energy Markets for Wind Wind Integration Costs Wind Integration is Facilitated by Energy Markets in Europe Adopting Limited Inter-Balancing Area Variability-Sharing System Evaluation Tool Summary Appendix on Markets References...48 WECC Electricity Markets and Variable Generation Integration 3

4 1. Introduction Variable Generation (VG) creates new challenges for efficient power system operations. Successfully integrating VG alongside more traditional supply resources requires creative thinking and new solutions, because the more traditional operational paradigm does not adequately account for the greater levels of uncertainty that underlie VG. The purpose of this Whitepaper is to: Identify some of the technical and institutional barriers that impede the efficient, economic and operational integration of VG. Suggest how new market mechanisms for buying and selling system flexibility can improve access to physically available flexibility from generators, loads, and energy storage facilities to facilitate VG integration. The specific mechanisms that are established to access this flexibility must not compromise reliability. Although this paper may outline some design choices, and discuss some of their advantages and disadvantages, specific design recommendations are outside its scope. 2. Definition of Variable Generation (VG) VG technologies are forms of power generation that depend on a primary energy source, which varies over time and cannot be stored. It is important to distinguish between renewable energy sources and variable generation sources. For example, geothermal and biomass generation use renewable fuels, but geothermal energy and biofuels can be stored in much the same way that traditional fuels are stored and are then used when needed to produce power. Variable generation is renewable, but it depends on fuels such as wind, sunlight, or water, which cannot be stored (although separate thermal or electric storage technologies can be applied). VG technologies typically deliver energy on an as-available basis, and they increase the level of variability and uncertainty in power system operations. The different VG technologies exhibit these attributes in different ways, but the fundamental challenges these technologies pose to power system operations remain rooted in this variability and uncertainty. Even if uncertainty could be eliminated (by perfectly predicting these energy delivery profiles in advance), the fundamental variability would alter at least some aspects of power system operation. The inability to predict precisely the output of VG over various time frames is not unlike the problem power system operations face today with respect to forecasting electric demand. However, on a per-unit basis, it is more difficult to predict VG delivery than it is to predict load. The ability to predict VG output varies among the various technologies, and while it is expected to mature over time, VG forecasts will never be perfect. The fundamental variable generation technologies that are identified in the North American Electric Reliability Corporation (NERC) Integration of Variable Generation Task Force (IVGTF) Summary Document 1 are: 1 Special Report: Accommodating High Levels of Variable generation, April 16, 2009, page 12, WECC Electricity Markets and Variable Generation Integration 4

5 Wind energy is converted to electricity when wind passes through a rotating turbine and generator. Presently, wind energy is the fastest-growing variable generation technology. Solar energy can be divided into two variations: o o Photovoltaic (PV) technology converts solar radiation directly to electricity. The level of power output depends on the properties of the solar cell and the amount of solar radiation at any given time. Solar thermal technology consists of a collector system that converts solar energy to heat, and a power block that converts the heat into electricity. Concentrating solar power (CSP) plants collect solar energy with a concentration of mirrors or lenses. A CSP plant collects thermal energy (heat) that it converts to mechanical energy and then electrical energy. Thermal storage can be added to the system to extend production to the early morning and evening hours, and to reduce the production variability and uncertainty inherent in a system without thermal storage capability. Hydrokinetic energy has three, principal variants: o o o Hydroelectric power utilizes energy from water flowing through river basins. Hydro power can consist of a combination of run-of-river production that is largely uncontrollable, and energy delivery that is controllable with the use of storage reservoirs. Storage hydro is not VG, whereas run-of-river hydro is. Wave power converts wave energy into electricity. This technology has not yet matured to the point that it is commercially viable, but it may be developed in the future. Specific characteristics of the energy delivery to the power system will likely be different in degree, but not in kind, to mature VG technologies. Tidal generation harnesses the energy latent in predictable tidal flows. Because tidal flows are very predictable, there is a high degree of production certainty compared to other VG technologies. This technology is not yet mature, but there are several demonstration projects operating today. The impact that wind energy has on a power system is becoming better understood. The production characteristics of other forms of variable generation, notably photovoltaic solar, have been analyzed to a lesser extent, and their impacts on power system operation and planning are not as well known. NERC has formed the IVGTF, which is analyzing issues that have been divided into 13 work streams. Some of the information from the IVGTF s ongoing work is included here. This whitepaper regards wind energy as the primary variable generation technology, because it has been studied much more extensively than other forms. However, in a broad sense, wind, WECC Electricity Markets and Variable Generation Integration 5

6 solar, and other VG technologies bring an increased level of variability and uncertainty to power system planning and operations. 3. System Balancing When considering system balance, it is neither necessary nor desirable to counter each movement in an individual wind or solar plant with a counter-movement from another generator. This is a direct consequence of the interconnected system that does not require specific loads and generators to be balanced individually. Instead, it is the aggregate of load and variable generation that must be balanced (subject to various constraints) through Balancing Authorities (BAs), within the bounds of established tolerances. The increased variability from VG implies that other controllable forms of generation will increase their cycling duty to maintain system balance. This increased cycling will occur over a variety of time scales ranging from a few minutes to several hours. Changes to the level of committed capacity could also be required, depending on the level of wind output and installed capacity. The uncertainty attribute increases the need for various forms of reserves, which may be needed over a variety of time scales. Uncertainty about VG levels day-ahead also has an impact on the unit commitment process and the commitment stack. Accurate VG forecasts can help achieve a cost-effective commitment stack, which maintains system reliability at target levels by reducing required levels of reserves. There are many important conclusions that can be drawn from the large amount of recent research and analyses of wind integration impacts. Most or all of the qualitative insights will apply to other forms of VG, although the quantitative impacts may differ. 4. What Policy Objectives Are Achieved By Integrating VG? 4.1. Supporting State and Federal Renewable Energy Policies A number of states in the Western Interconnection have established goals for renewable energy penetration, 2 while others have established voluntary goals. Targets and goals are shown in Figure 1. WECC has studied the transmission-related impacts of supplying as much as 15 percent of end-use sales from renewable energy resources, 3 though the study did not assess the operating impacts. 2 It is notable that California has a legislatively mandated target of 20% of end-use sales by 2012, and an administratively established target of 33% by The Database of State Incentives for Renewables & Efficiency can be located at 3 TEPPC 2009 Annual Report, table 5.7. WECC Electricity Markets and Variable Generation Integration 6

7 2019 RPS Targets and Goals for WECC States and Provinces Percent of Energy Sales Subject to RPS Figure 1 - Source: TEPPC 2009 Annual Report Most of the renewable resources that will be employed to meet these state mandates will be variable generation technologies, namely wind and solar. Federal renewable energy policies are still evolving, but it is reasonable to anticipate that a national renewable energy portfolio standard may be enacted in the next few years along the lines of the state goals. 4.2 Minimizing the Cost of Achieving High Levels of Renewable Generation Penetration A key objective is to minimize the delivered cost of renewables. This includes the cost of grid services and resources that are employed to deal with variability and uncertainty, while maintaining reliability. Delivering on this cost objective requires that existing and available resource flexibility across the Western Interconnection be used to the maximum extent possible; and that new resources are used effectively and efficiently. 5. What Are the Challenges of Integrating VG? 5.1 VG Increases the Combined Variability of Load and Generation Load (customer demand for electricity) is inherently variable and somewhat uncertain. It is influenced by weather trends, which are difficult to forecast with certainty. VG production is also variable and uncertain (particularly solar and wind), but because production is typically WECC Electricity Markets and Variable Generation Integration 7

8 dominated by local weather conditions, the degree of variability and uncertainty is much greater. Sometimes the correlation between load and VG variability can be complementary, as in the case of high solar production on hot summer days. In other circumstance, the correlation is negative, as what occurs when wind production falls off on hot summer days. Conventional generating plants and other sources of system flexibility must be committed and dispatched to serve what s known as the net load, or electric demand less VG production. The degree of variability and uncertainty around VG and its relative impact on individual BAs also will vary. This variance depends on the amount of VG, the size of the BA, its geographic location, and the degree to which VG is dispersed across its service area in terms of technology and geography. The BA s ability to react to this variability and uncertainty is affected by the types and amounts of flexibility available from conventional resources and other sources of system flexibility, which can be deployed to balance supply with net demand. 5.2 VG Output is Difficult to Accurately Predict and Control Because VG production is so weather-dependent, it is difficult to predict. Recent improvements in wind forecasting technology have led to improved accuracy and less uncertainty. Methods for forecasting solar production also are improving rapidly, though they are not yet as robust or as accurate as wind forecasts. Unlike conventional generation, VG is less easily controlled. VG captures energy from the wind or the sun as it is available, and if the wind isn t blowing or the sun isn t shining, there s no way to increase production from these resources. For wind turbines, too much wind can be as problematic as too little, since production typically has to be shut down entirely in cases where wind speeds exceed the design capability of the wind turbine. Often, feathering blades on wind machines, or tilting mirrors or panels on solar projects (de-focusing) can decrease production. However, this is accomplished at the expense of reducing overall VG production. Feathering for wind machines and de-focusing for solar also provide the means to limit rapid increases in VG production, which require corresponding decreases in the output of other supply resources. Increased net load variability can increase the operating costs associated with maintaining a given level of grid reliability. That s because compensating for variability requires operating flexible resources with high fuel costs in place of less flexible resources with lower fuel costs. As with inherent variability, VG uncertainty also has a cost in terms of increased requirements for reserve capacity. The specific nature of the additional reserves is dependent on the degree of uncertainty with regards to predictability and the speed of unexpected ramps in VG output. The more uncertain the level of output, the more reserve capacity must be carried. The faster the unexpected ramps, the faster the reserve capacity must be. Additional reserve capacity and increased reserve response speed increase the costs of VG integration. In a recent study by the California ISO, titled Integration of Renewable Resources Operational Requirements and Generation Fleet Capability at 20% RPS, it was demonstrated that the combination of solar and wind resources can lessen operational requirements, because solar resources are ramping up when wind resources are ramping down, and vice-versa. WECC Electricity Markets and Variable Generation Integration 8

9 5.3 Current Operating, Scheduling and Business Practices Lack Flexibility Current practices for operating and scheduling transmission and generation are based on hourly time frames. However efficient integration of VG requires more flexibility because of the intrahour variability. A combination of scheduling, operating and business practices that allow withinhour updates would have the following benefits: Reduce the amount of reserve capacity that the source BA must hold; Transfer some of the responsibility for managing the variability and uncertainty of the resource to the sink BA, through the use of dynamic schedules or by operating as a pseudo-tie resource; and Allow VG generators to assume some of the burden of managing variability and uncertainty themselves, by selling production that exceeds scheduled volumes and by buying production shortfalls rather than placing the entire burden on the source BA. Today, intra-hour deviations are caused largely by load variability. As variable energy resources are added, flexible resources will have to be operated to manage the combined intra-hour deviations of load and VG. These deviations are expected to grow with increasing amounts of wind and solar energy, 4 requiring a greater focus on intra-hour operations. 5.4 Federal Renewable Energy Tax Incentives Favor Production To the extent that curtailment of the output of variable generating resources is technically viable, production-based tax incentives discourage variable renewable generators from curtailing their own resources as part of an economic solution to variable resource integration. Prior to enactment of the American Recovery and Reinvestment Act of 2009 (ARRA), tax incentives, as applied to wind and solar electric generating facilities, were production-based (Production Tax Credit). ARRA introduced the concept of a grant in lieu of the Production Tax Credits at rate of 30% of the capital cost of the project. The ARRA approach to tax incentives removes part of the economic disincentive for variable renewable generators to curtail production when necessary. 5.5 Commercial Terms and Conditions Discourage Curtailment Common terms and conditions in renewable energy purchase contracts require the supplier to produce renewable energy whenever possible, and payments are structured on a take-and-pay basis. To minimize costs, utilities want as much output as possible from any given renewable resource in order to meet renewable portfolio mandates. This commercial incentive may conflict occasionally with a need for VG integration, which will need to be curtailed at times to assist in grid reliability management. Currently, the projects often are financed on the basis of being able to sell maximum production from the variable generators. 4 See an example in the California ISO Report, Integration of Renewable Resources - Operational Requirements and Generation Fleet Capability At 20% RPS. WECC Electricity Markets and Variable Generation Integration 9

10 6. How a Resource Flexibility Market Could Address VG Challenges Using a centralized market to buy and sell flexibility could maximize VG output while minimizing the costs of integrating VG. Markets provide incentives for parties that have surplus flexibility to make it available to those who need flexibility. Those who need flexibility can use markets to determine the most cost-effective means of acquiring it. Markets establish a fair price for that flexibility, and provide an efficient means for parties to enter into transactions and exchange compensation for flexibility services. A key assumption that underlies this Whitepaper is that existing resource flexibility could be more efficiently utilized across the Western Interconnection to maintain reliability. This also would deliver value in the form of higher revenues for those entities that have flexible resources and lower costs for those that need access to flexible resources. Developing a common understanding of what sort of flexibility VG requires, and what is available, should guide participants in the development of products, contracts, trading mechanisms, regulations, and business practices that increase the economic benefit for each entity. Within a market, buyers and sellers have a natural incentive to maximize their economic benefit. Therefore, when buyers and sellers cannot increase their economic benefit, either there is no more benefit to be had or access to that benefit is in some way impeded. The purpose of the market is to bring together buyers and sellers to identify and act on opportunities to improve each side s economic benefit. Within the context of VG, the concept of supply and demand references resource flexibility on the part of generation, transmission, and load. While this paper stops short of quantifying the additional amount of resource flexibility that exists, there are many reasons why improving access to the flexibility that does exist would benefit entities in the Western Interconnection. For many BAs, VG presents new levels of challenges to balancing system loads and resources, while operating within associated reliability metrics such as frequency, stability, intertie flows, transmission line loadings, and control performance standards such as CPS-1 and CPS-2. Historically, most of the variability and uncertainty within BAs stemmed from the variability of load and the occasional loss of generation or transmission capacity. VG production is inherently more variable and more uncertain, and if the VG build-out is large compared to the BA s load, operating reliably presents a significant and potentially expensive challenge. Having access to flexible resources reduces the scale, scope, and cost of meeting this challenge because the variability of VG resources can be spread across several BAs. To the extent that multiple BAs could absorb the volatility more cost effectively than the host BA, and be adequately compensated for doing so, the market has successfully achieved its purpose of improving each entity s economics. Owners of transmission and generation assets also can benefit from market access to resources that could help meet the reliability demands of VG. During certain times, transmission and generation assets are not utilized to their full potential. At other times, transmission congestion, environmental limits, and other limits placed on generation, decrease system flexibility to the point where it is necessary to access remote system flexibility. Beyond physical constraints, such as outages that limit utilization, one major barrier to increased utilization is simple economics: The marginal benefit of increased utilization must be greater than the marginal cost for increased utilization to make sense. A market for resource flexibility would help WECC Electricity Markets and Variable Generation Integration 10

11 participants efficiently allocate costs and benefits among market participants to increase efficient transmission utilization. Expanding the visibility of flexible resource needs and the availability of flexible resources, while extending markets across a wider footprint, would improve access to more efficient resources. Expanded access to flexible resources that could help meet the reliability demands of VG may provide opportunities for increasing the marginal revenue side of this relationship. This would improve asset utilization and the assets overall return on investment. Given the ability to serve the reliability demands of VG, and the sufficient latitude provided in structuring a transaction, a seller s perspective may change from, I cannot provide that flexibility, to, I can provide lots of flexibility if the price is right. This ultimately may identify efficient solutions to the challenge of VG integration that avoid arbitrary administrative remedies. Developments in the market may also encourage participation from energy storage resources and increase participation by those resources already in the market. Some of the increased flexibility needed to integrate VG may come from demand-side and storage resources if market mechanisms and eligibility rules are amended accordingly. Finally, market mechanisms may prove beneficial in bringing together VG resources to leverage their geographical and technological dispersion in a manner that reduces the combined variability of load and generation. It may be the case that one BA s load and resource volatility is equal to that of another, but their combined profile may be less volatile. If the contractual arrangement underlying such a combination is economically beneficial for both parties, then the market has been successful in achieving its stated objective. 7. Current Impediments to Realizing Additional Benefits from Market-Based Solutions A number of institutional, market, and operational barriers stand in the way of leveraging markets to share flexibility. These include: 7.1 Administrative Constraints It is currently difficult to quickly initiate dynamic transfers, 5 which are a means of moving variability from one BA to another. This limitation also precludes the development of a broadly traded market for certain kinds of flexibility products between BAs. In addition, dynamic transfer limits may be kept too low by a lack of automation of reliability processes, such as voltage control and RAS arming. The existing processes for implementing and approving interchange and transmission schedules are oriented around hourly time intervals, which is not sufficiently granular for the operational demands of VG. Many of the related business processes still require manual 5 These include dynamic schedules, which are interchange schedules that can be adjusted withinhour on intervals as small as five minutes, and pseudo-tie resources, which are typically supply resources that are physically located in one BA but are considered to be located in a different BA for operational control purposes. WECC Electricity Markets and Variable Generation Integration 11

12 intervention at various stages. As noted later in this paper, there are efforts underway to permit within-hour schedules and to further automate many of the attendant business processes. Similarly, the existing processes for making adjustments to schedules create administrative constraints. Although a limited number of BAs currently allow schedules to be changed after the start of the hour, schedules can be changed only for a very limited number of reasons, and when allowed, the number of schedule changes is limited to one per hour. This limitation often leads to poor utilization of available transmission and imposes severe constraints on the ability of the BA or the VG itself to manage the variability and forecast uncertainty. As noted earlier, many of the business processes related to operations and scheduling still require substantial manual effort. The ability to efficiently and cost-effectively integrate VG will require investments to refine and automate more of the related business processes. Many BAs, especially smaller ones, likely are reluctant to invest the amounts of time and money in staffing, business process modernization and new systems that will be required. 7.2 Lack of Consistent, Well-Defined Service Requirements for VG Integration Although much is known about the general operational requirements for integrating large amounts of VG, consistent and well-defined requirements for the relevant types of services have not yet been established. The general requirements include fast, intra-hour markets for energy and generation capacity that can reduce the amounts of reserve capacity that a BA must hold in order to deal with VG variability and uncertainty. It also requires intra-hour markets for transmission service that can increase the utilization of existing transmission capacity and potentially reduce the need for additional transmission infrastructure. Parties must be able to schedule or change schedules on time frames shorter than an hour, though neither minimum nor likely optimal amounts of adjustment flexibility (30 minutes, 15 minutes or perhaps even less) are yet known. For the relevant business practices that do exist, there is a lack of uniformity among BAs in terms of implementation. Each BA also continues to develop its own rules regarding whether intra-hour schedule changes are allowed, the circumstances under which they are allowed, and the way in which within-hour schedule changes are submitted and approved. However, a number of discussions are underway in various forums in an effort to seek more consistent practices among BAs in the Western Interconnection. Technical and other eligibility requirements for flexible resources that can meet operational requirements continue to be linked, either explicitly or implicitly, to the capabilities and features of specific technologies. Certain kinds of storage devices, for example, are currently limited in providing regulating reserves unless they can sustain their full charging or discharging rate for at least one hour. This may no longer be necessary if fast energy markets allow BAs to hold less regulation, which need only be operated for a single, sub-hourly market interval. 7.3 Transmission Rate Pancaking Pancaked transmission rates limit the ability to economically share flexibility, because they require transactions that cross multiple transmission service provider boundaries to pay a WECC Electricity Markets and Variable Generation Integration 12

13 separate fee each time a boundary is crossed. Low-cost, highly flexible resources located far away may end up being uneconomical compared with higher-cost resources that are located nearby, after factoring in the cost of multiple transmission fees. This is especially problematic in the Western Interconnection with its relatively large number of Transmission Providers. A notable experiment in reducing transmission service fees is the WestConnect Regional Transmission Pricing Experiment. It is a two-year pilot program that eliminates transmission rate pancaking for certain types of transactions within its footprint. The experiment began on July 1, Transmission Limitations Between BAs Limitations on the amount of available transmission may prevent one BA from purchasing resource flexibility located in another, distant BA, whether by a conventional interchange schedule or via dynamic transfer. In some cases, sufficient physical transfer capability is available, but other incompatible operational and scheduling practices can limit the ability to move flexibility from where it can be supplied to where it is needed Overcoming Barriers There are a number of activities that currently are being undertaken, or may be undertaken in the future, to remove existing barriers that act to inhibit the Western Interconnection s ability to accommodate large amounts of VG at reasonable cost. Broader participation among WECC members in these activities will be most helpful in removing existing barriers. 8.1 Develop Market Products that Meet the Needs of VG Integration WECC member entities can all play an important role in developing market products that facilitate VG integration by bringing stakeholders together and marshalling the right subject matter experts. Specific tasks that could be undertaken in connection with this development include: Bringing together experts from the National Laboratories (NREL, PNNL and others), the VG development community, BAs, and owners and operators of conventional resources to develop a common understanding of the reliability and operational requirements of VG. Consulting a body of research that includes: Existing regional wind integration studies, the work underway by the California ISO to capture the impacts of solar technologies; market and operational design work currently underway at ERCOT; and papers collected by the Utility Wind Integration Group, which contains a great deal of useful background information. 6 For example, the path rating at COI is 4800 MW, but only 500 MW of dynamic transfers from VG can be accommodated until upgrades that include automation of the voltage control equipment and automated methods for arming and disarming remedial action schemes can be implemented. WECC Electricity Markets and Variable Generation Integration 13

14 Participating in industry efforts to structure commercial arrangements for products sourced from the bilateral markets that meet the identified operational needs. These products have to fit within existing commercial structures (WSPP contract, California ISO markets, OATTs). They have to be defined clearly and unambiguously so they meet reliability objectives, and do not create unnecessary commercial or regulatory risks. They should also be flexible enough to accommodate a wide range of needs and to encourage innovation. Support promising initiatives both inside and outside WECC that seek to design and implement solutions. These include, but are not necessarily limited to: Supporting efforts being undertaken by the Joint Initiative to address dynamic transfers, intrahour scheduling and automated transaction processing; exploring further development of the Efficient Dispatch Toolkit or suitable alternatives; and tracking the effectiveness of the WestConnect Transmission Rate experiment that is outlined in this Whitepaper. 8.2 Re-Examine Operating Practices WECC member entities also may take a fresh look at existing operating and scheduling practices. Many of today s operating practices and rules initially were developed at a time when computing power was scarce, communication was expensive, many of the analysis and visualization tools that are available today weren t even contemplated, and wholesale markets were less complex. 8.3 Develop Revenue-Sharing and Other Tariff Mechanisms That Eliminate Transmission Rate Pancaking WECC member entities may benefit from discussing mutually beneficial revenue-sharing solutions that allow transmission providers to recover their revenue requirements but don t compromise reliability. A regional transmission tariff might be one solution. There are likely to be others. 8.4 Standardize Business Practices Finally, WECC can provide a forum for identifying and discussing common business practices that would be beneficial and appropriate. Among other things, common business practices should meet the needs of all market participants. Business practices should be designed with enough flexibility in mind so that worthwhile enhancements can be added when and if they make sense. Business practices should not have an adverse impact on reliability or unduly discriminate in ways that favor some parties at the expense of others. Although VG is perceived to be the driver, most business process improvements that are likely to result from this initiative would be worthwhile with or without VG. 9. Current Voluntary Efforts WECC member companies already are undertaking a number of voluntary efforts to address some of the VG integration challenges discussed throughout this paper. WECC Electricity Markets and Variable Generation Integration 14

15 The Joint Initiative (JI) is a series of collaborative efforts among the ColumbiaGrid, Northern Tier Transmission Group and WestConnect transmission planning groups to develop standardized business practices and implement supporting business systems. The JI currently has three projects underway: Intra-Hour Transmission Purchasing and Scheduling, which is designed to facilitate intra-hour schedule changes to address unanticipated changes in generation, and to allow better use of capacity within and from outside BAs by developing common business practices that allow shorter time frames for scheduling. Intra-Hour Transaction Accelerator Platform (I-TAP), which facilitates bilateral transactions from both within and outside a BAA. Dynamic Scheduling System (DSS), which provides a more agile delivery mechanism for dynamic energy products, including dynamic schedules and pseudotie resources. Although the short-term drive for these three efforts is wind integration, members recognize that their value goes beyond managing VG to expanding the market for flexibility products and services. The WSPP Intra-Hour Supplemental Power service schedule (Schedule E) is a new service that provides a within-hour capacity product that is callable by the purchaser under the WSPP form agreement. The Wind Integration Study Team (WIST) is a collaborative effort among the ColumbiaGrid and Northern Tier Transmission Group to facilitate the integration of VG into the northwest transmission grid. It is currently examining two issues: System constraints on the increased use of dynamic scheduling, and how to balance the cost of new transmission capacity against the value of increased amounts of delivered wind energy. The WestConnect Regional Transmission Pricing Experiment is a two-year pilot program that eliminates transmission rate pancaking for certain types of transactions. It began on July 1, Area Control Error (ACE) Diversity Interchange (ADI) is a pilot program that allows a group of BAs to share the burden of managing ACE, by allowing BAs to net portions of their ACE deviations. The WECC Reliability-Based Control (RBC) field trial is a proof-of-concept effort to assess an alternative method of managing interconnection frequency. It requires BAs to limit the duration of operation outside a variable ACE limit that gets tighter as frequency deviates further from 60 Hz. RBC may reduce the amount of movement that is required from flexible resources in order to meet NERC reliability standards. The Efficient Dispatch Toolkit (EDT) initiative is intended to provide a voluntary, centrally dispatched, within-hour market for sharing resource flexibility across the Western Interconnection. It is based on a similar service that currently operates in the SPP. WECC is WECC Electricity Markets and Variable Generation Integration 15

16 currently preparing to undertake a cost-benefit study of the EDT and is in early discussions regarding budgets, governance and operations. Under the direction of the VGS, WECC is undertaking a Western Interconnect-wide study of VG integration to assess the value of virtual BA consolidation and related operating strategies. 10. Recommendations The VGS MWG has identified further actions that will bring together stakeholders to promote a common understanding of the challenges of VG integration, and jointly develop approaches that expand access to resource flexibility. These recommend efforts include: Developing a 10-year plan that addresses future needs for scheduling, operations, forecasting, and markets. Continuing to provide a forum for discussion and further development of methods to promote more efficient access to and utilization of transmission. Supporting continued development of tools such as ADI, DSS, and RBC that help BAs continue to fulfill their reliability obligations in the presence of large amounts of VG. Supporting the development of market tools such as I-TAP and the EDT. Completing VG Integration studies proposed by the VGS to gain insights on new approaches to managing large amounts of VG across the Western Interconnection, including virtual BA consolidation techniques. Initiating studies where necessary to establish regulation, load following and ramping needs for VG as a function of VG penetration, of scheduling interval and of forecast accuracy. Development of a best-practice methodology for evaluating these requirements could be a high priority of VGS (probably the Operating Work Group). This information is precedent to understanding the parameters of potential market solutions to the problems raised in VG integration. Reviewing WECC policies regarding the use of its data and infrastructure, and making necessary changes to broaden the use of these assets by members who are developing new tools for planning, BA operations, and markets. Providing a bibliography of current and ongoing studies and experiences from around the world on potential solutions for VG integration. Providing periodic forums that highlight current work in WECC and in the industry atlarge for educational purposes. Exploring technology to more fully automate tagging and facilitate sub-hourly scheduling of generation and transmission capacity. WECC Electricity Markets and Variable Generation Integration 16

17 Identifying gaps and inefficiencies that hinder the market-based solutions, and facilitating specific improvements in business practices and OATTs that minimize administrative seams to expand access to resource flexibility. Retaining the VGS to help WECC achieve the above-stated objectives and suggesting ways to improve reliable and efficient integration of VG. WECC Electricity Markets and Variable Generation Integration 17

18 11. Technical Appendix This section contains technical background and other supporting information. It was prepared by Michael Milligan of the National Renewable Energy Laboratory (NREL) and contains sections of an NREL report written by Michael Milligan and Brendan Kirby entitled Market Characteristics for Efficient Integration of Variable Generation in the Western Interconnection. 12. Relevant Characteristics of Variable Generation (VG) for System Operators Variable generation itself has four principal characteristics that are important for power system planning and operations: the generation output is variable because the primary energy source (wind, solar, or hydro flow) itself is variable, many forms of variable generation (wind, solar, hydro) have a zero or near-zero variable cost, wind and some forms of variable generation are difficult to forecast precisely. In many cases, the resource is locationally constrained, and the best sites are often located far from load centers. Wind and solar are often described as intermittent, but since that term implies sudden changes in output. Since VG output varies over a period of from minutes to hours longer than the instantaneous outages that affect conventional units we use the term variable. Failure to recognize these characteristics of VG will raise the financial and environmental costs of the nation s overall electricity usage. It is important to note that the characteristics of VG are not unlike those of load. However, the magnitude of VG s variability is greater than the variability of load on a per-unit basis, and depending on the time frame of interest, VG may somewhat more difficult to forecast than load. Hence, the primary differences are more of degree than of kind. 7 VG is capable of being dispatched to a limited degree, and there is interest in exploring options that would make this flexibility available over certain times frames. Given the high cost of fossil fuels today, the primary value of VG is in supplying energy, not capacity, although the different technologies can be counted upon to meet peak demands to varying degrees. VG saves fuel and reduces emissions. It is also a resource with a high capital cost, but very low variable costs (fuel or operating costs). It almost always saves money to reduce production at fossil-fired power plants (given sufficient flexibility in the non-vg generation fleet), and instead generate as much as possible at VG plants. It can also make sense to reduce generation at hydro plants if they have a reservoir for storing water to take advantage of wind or solar generation. Most reservoir-based hydro systems are energy limited, so saving water with wind can increase the capacity value of the hydro system and may increase the value of the hydro energy as well. Although modern wind turbine technology allows an operator to precisely curtail wind production when necessary, spilling free wind and other VG output is usually only desirable when it is 7 We note that some non-conforming loads such as arc furnaces, impose an extremely large regulation burden on the power system. Arc furnaces cycle from zero to full output nearly instantaneously, whereas VG such as wind move much more slowly over the same time frame. Anecdotal evidence from system operators indicates that these large changes in load, which can be up to 100 MW, can be adopted into standard system operations practice with experience. Similarly, there is anecdotal evidence that system operators adapt to wind generation with experience, resulting in more efficient system operation over time. WECC Electricity Markets and Variable Generation Integration 18

19 required for reliability reasons, or in the event that wind could supply ancillary services in the future. Similarly, spilling solar or hydro energy is generally not desirable Since much more analysis has been done on wind generation than solar, and because of the likelihood of significant wind development in the WECC over the next few years, much of the following discussion centers on wind. However, most of these issues apply similarly to solar (especially PV), with some (likely lesser) application to concentrating solar with storage. While annual wind and solar energy production can be forecast with reasonable accuracy, it is more difficult to predict wind generation output a few hours or a few days in advance. Wind generator operators cannot commit to and follow hourly generation schedules like conventional generation operators. Wind forecasts are getting better, however, and power system costs can be reduced when power system operators have an advanced wind forecast available in their control room. The best wind resources are often located far from load centers and are frequently far from existing transmission lines. Investment in new wind plants must often be accompanied by a significant investment in transmission additions before the full amount of new wind generation capacity can be accepted onto the grid and delivered to loads. Where new wind capacity precedes transmission expansion, the incremental wind generation may have to compete for transmission access with other generators and be delivered using non-firm capacity; such generation may often be curtailed due to reliability or transmission capacity limitations. A robust transmission system (coupled with sub-hourly scheduling) that allows wind and load variability to be aggregated over large geographic areas further reduce the cost of wind integration. Therefore, wind greatly benefits from a long-term, regional approach to transmission planning. Of course, power system customers benefit from a robust regional transmission network with or without wind. 13. System Operation with Wind Wind s characteristics can present challenges to the power system operator. Wind variability in itself is not unique; system operators continuously deal with load variability over all time frames from seconds to seasons. However, because wind is variable, it does add to aggregate variability. While a power system does not need to respond to the variability of each individual wind turbine, it does need to meet the North American Electric Reliability Corporation (NERC) reliability standards by balancing the difference between aggregate load and net VGS production 8 with aggregate conventional generation. Adding wind to the generation mix will increase the control actions the conventional generators must take. Fortunately wind and load variability tend to be uncorrelated, so they do not add linearly, greatly reducing the net flexibility required from the conventional generators. This is based on the principle of statistical independence, described below. If additional flexibility is not valued by the market or incorporated into system plans, then sufficient response capability may not be offered, making it 8 To meet RPS goals and minimize fuel burn and operating costs, it is usually optimal to use all VG energy that is available, adjusting the balance of generation to meet the remaining load. This load is often called the net load which is the native load plus exports, minus the VG. This net load is what the remainder of the system operates to. WECC Electricity Markets and Variable Generation Integration 19

20 more difficult to balance the system. This is a key issue: there may be sufficient flexible generation that could respond, but in the absence of a market or other institutional framework, there may be no way to access that generation. Forecast error is not unique to wind either. System operators regularly deal with load forecast uncertainty. A one-degree weather forecast error, for example, can result in a 1000-MW summer peak-load forecast error in a BA with a 45,000 MW peak demand. As with variability, wind forecast errors add to load forecast errors, increasing the required conventional generator flexibility. Fortunately, as with variability, wind and load forecast errors tend to be uncorrelated, reducing the total required conventional generation flexibility. 14. Efficient Integration of VG Capturing the full environmental and economic benefits of variable generation requires looking at the power system differently than in the past. These differences encompass resource portfolio strategies and operational efficiencies. One important difference is that energy production should be valued as well as capacity. This requires examining total annual fuel requirements (and emissions) as well as peak generation needs. Accommodating VG involves adjusting the power system structure to accept energy from VG sources when and where it is available. This involves a combination of physical attributes and institutional support, along with incentives for conventional generation to be built with desired physical attributes and operating capabilities. Adding VG to an existing power system does not affect the need for installed capacity (which depends on load), but may change the type of capacity: more flexible generation with the ability to start, stop and ramp more quickly and achieve lower turn-down rates (in the case of high wind penetration) may eventually replace base-load generation that is not as flexible. 9 A common objective for efficient integration is to minimize the cost of electricity, given the resource mix, and subject to holding reliability constant (as compared to a no-vg case). There may be trade-offs involved in this complex optimization. For example, the reliability constraint may cause the system operator to spill VG energy to avoid a reliability problem. Spilling involves an economic cost that generally should be avoided, but reductions in reliability have their own costs that may significantly exceed the cost of spillage. Other objectives include meeting required resource mixes that may be mandated by state or other governments. In the discussion that follows, the goal is to examine how to maximize the energy capture from a fleet of VG resources, given the objective of minimizing the cost of operating a power system that has significant variable generation. The discussion begins with an examination of physical characteristics that help with efficient integration, and continues into a discussion of the market and institutional approaches that can help deliver the required physical capability to the power system operator. Recognizing that the goal of WECC is to maintain system reliability, we argue that institutional constraints, such as the lack of large, fast energy markets, can impede reliability because access to needed generation is limited. As stated above, much more analysis 9 Or, if economics justify the retrofitting of inflexible generation, with the ability to achieve lower turndown levels or ramp faster, those efforts may be undertaken. WECC Electricity Markets and Variable Generation Integration 20

21 has been performed on wind generation than solar. We expect that there will be similar impacts from PV, and to a lesser extent CSP, as from wind, although the precise magnitude of impacts on variability and uncertainty will certainly vary somewhat from those of wind. The difference is believed to be more of degree than of kind. 15. Wind Generation and Efficient Power System Integration Large, flexible power systems make it easier to economically manage high penetrations of variable generation. Physical size is important because the correlation between the production levels from multiple wind plants diminishes as those plants are geographically farther apart (Figure 2). A similar characteristic occurs in solar generation, and presumably other forms of VG. In the graph below, correlation between wind farms is lowest (approaches 0) when distances between wind plants are large. 12-hour 4h Average 2h Average 1h Average 30min Average 5min Average Figure 2. Wind generator variability loses correlation as the distance between machines increases and as the time frame of interest decreases (Ernst, 1999). To illustrate the geographic diversity impacts within the wind plant, Figure 3 illustrates the perunit variability from a group of 15 turbines and a nearby group of 200 turbines. The graph is normalized by the average wind output. The lower panel shows the group of 15 turbines, and it is apparent that this grouping has much more variability per unit as compared to the upper panel. The coefficient of variation (standard deviation divided by mean) is for the small cluster, and for the large cluster. WECC Electricity Markets and Variable Generation Integration 21

22 Figure 3. Geographic smoothing within a wind plant has a considerable smoothing impact using 1-second data for nearly 9 hours. Preliminary results from the Eastern Wind Integration Study (EWITS) provide dramatic evidence of the benefits of large area aggregation (Figure 4). EnerNex studied the wind variability from 71,671 MW of wind generation spread over most of the eastern interconnection. Meso scale wind modeling was used to generate three years of historic wind data on a ten-minute basis and a 2-kilometer grid spacing. Note how the variability of the total wind fleet (red bar) is well below the variability of the same wind resource when dealt with on an individual regional basis. Note also that these are large regions to start with. Total wind production is expected to vary by 10% from hour to hour less than 14 times per year. When 300,000 MW of wind are studied a 10% variation is expected less than four times per year. Large geographic dispersion, coupled with a robust transmission grid and market rules that support sub-hourly transactions allow wind aggregation over great distances and reduce the net variability, which must be compensated for. WECC Electricity Markets and Variable Generation Integration 22

23 Hourly Wind Variability By Region Reference Scenario Hours per Year MAPP NYISO ISO-NE MISO TOTAL PJM SERC SPP TVA % -7.5% -5.0% -2.5% 0.0% 2.5% 5.0% 7.5% 10.0% Hourly Production Changes on Nameplate Figure 4. Enernex study performed for the Eastern Wind Integration and Transmission Study (EWITS) shows a dramatic reduction in wind variability when transmission supports aggregation over a large geographic area (Zavadil, 2009, draft results). As with load, larger geographic and electrical size also makes forecasting easier. Table 1 shows that the wind forecasting error is reduced significantly when wind output from all four regions of Germany are compared with wind output from a single region. (Rohrig, 2005) This conclusion is reinforced by Ahlstrom (2008) when aggregated over a broad geographic region, wind forecast errors can be reduced by as much as 30%-50%. Thus, power system operators can more accurately predict and plan for changes in wind output when their systems are larger. Not surprisingly, forecasting accuracy also improves closer to real time. It is easier to forecast for short periods ahead, compared to longer periods in the future. Markets that operate closer to real-time take advantage of the improved forecasting accuracy by allowing more frequent generator schedule changes. Hour-ahead markets better accommodate wind than day-ahead markets. Sub-hourly markets have the least forecast error. A coordinated series of regularly clearing markets over a range of time frames (e.g., day-ahead, hour-ahead, and at fast, subhourly time frames) provides the best ability for conventional generation to adjust to changing wind conditions at least cost. RTOs and ISOs typically operate sub-hourly markets while other regions do not. This feature is one of the drivers of costs of software systems for RTOs and ISOs. WECC Electricity Markets and Variable Generation Integration 23

24 The Principle of Statistical Independence Electric power systems are comprised of a very large number of components. A typical utility service territory (or market area) includes many thousands of individual customers. The behavior of these customers exhibits some statistical correlation over some time periods, but has little correlation over other periods. During the morning load pickup, customers are generally increasing their usage of electrical devices, leading to an overall increase in electric demand. However, during very short periods of time, such as seconds to minutes, some loads are increasing at the same time that other loads are decreasing. There is no correlation between these random events; one customer turns on the lights at the same time as another customer turns off the lights. These events, when they occur simultaneously, have no net impact on electrical demand. Wind turbines have a similar statistical property. During the short time periods of seconds or minutes, one wind turbine may be experiencing an increase in wind speed, resulting in more wind power output from the turbine. At the same moment, another wind turbine may experience a decline in wind speed and power output. The random nature of these events can be captured statistically, and are formally described as uncorrelated events. It is important to note that if wind turbine A always runs counter to wind turbine B, then they are perfectly negatively correlated (correlation coefficient is -1). But if sometimes the turbines move together, and other times move in opposite directions, this lack of correlation has important implications for balancing requirements. The principle of statistical independence is the reason why each increase in customer demand (resulting from a switched on light, for example) does not need to be matched by a corresponding increase in generation. Because other customers are switching off their lights at the same time, statistical methods can be used to calculate the amount of generation required to match the aggregate change in load. This principle of statistical independence over short time frames applies to loads, wind turbines, and to load and wind combined. This article illustrates this concept in several different contexts: load, wind, load and wind, and wind forecasts are all subject to the principle of statistical independence. WECC Electricity Markets and Variable Generation Integration 24

25 Table 1. Wind forecasting accuracy improves when larger geographic areas are considered. NRMSE Error % Forecasting Germany (all 4 control zones) ~1000 km 1 German Control Zone ~350 km Day ahead hours ahead hours ahead Aggregation and Large Balancing Area Size Utilities have taken advantage of aggregation for decades. Since each balancing area only has to compensate for the variability in its aggregate load, and since random variations in individual loads partially cancel each other out, larger balancing areas require relatively less system balancing through regulation service than smaller balancing areas. As shown above, the same principle applies to integrating wind: Larger BAs are better able to integrate large amounts of wind because the random variability of individual wind generators and individual loads partially cancel each other out. This is based on the principle of statistical independence. If multiple remote wind plants are grouped and operated together within a single balancing area, their overall per-unit variability falls and it costs less to integrate their production into grid operations. Having a deep pool of flexible generation that can respond to variations in wind output helps system operators and reduces the cost of system balancing. Larger balancing areas have larger generation pools. Greater flexibility is a function of the generation mix, but larger pools always provide greater flexibility than smaller pools of the same generation mix. Therefore there are two beneficial impacts of larger balancing areas: (1) the net per-unit variability declines with aggregation; equivalently, variability adds less than linearly, and (2) ramping capability, the ability to supply variability from the generation fleet, adds linearly with aggregation. These two effects, taken together, show that larger balancing areas can manage more variability, and that the amount of variability from VG that must be offset in the combined balancing areas increases more slowly than the supply of flexible resources. As an example of the benefit of the larger balancing areas, a recent NREL technical report analyzed the consequences of balancing area consolidation in Minnesota, both with and without wind (Milligan and Kirby, 2007). Neighboring balancing areas will sometimes need to redispatch their generation in different directions at the same time. This happens when the load in one balancing area is increasing during a period when the load is decreasing in another balancing area. During such times, it would be beneficial for both systems to net their load ramping requirements, which would result in less ramping of generation in both balancing areas. Using hourly data, the authors calculated the ramping that could be eliminated if the four balancing areas in Minnesota were to combine (which subsequently happened as part of the expansion of the Midwest Independent System Operator footprint). The graph for one full year of hourly load data is shown in Figure 5. Opposite ramping does not occur in all hours, but it is apparent from WECC Electricity Markets and Variable Generation Integration 25

26 the graph that 50 MW/hr or more can be reduced during much of the year, resulting in approximately a 14% reduction in ramping requirements (both up and down) annually if operations are combined. This reduction in load ramping requirements translates into lower cost of serving loads in all affected balancing areas. In Figure 5 below, the top graph shows the load ramping movements that would cancel out and need not be performed if the four Minnesota balancing areas were combined. Cancellation happens whenever one balancing area is ramping up while another is ramping down. Benefits are spread throughout the year, but can be seen to vary from hour to hour. The lower portion of Figure 5 reorganizes the same ramping information into a ramp-duration curve, which shows that, absent balancing area consolidation, there is as much as 75 MW of costly, unnecessary load-following generation in Minnesota attempting to compensate for the net variability of loads. This graph is based on loads only; there is no wind in the system portrayed by this graph. Figure 5. Physical ramping requirements can be reduced by consolidating balancing areas (hourly load data). Combining balancing areas provides multiple benefits for loads, as seen in Figure 5. Because wind is also subject to the principle of statistical independence, wind variability declines on a per unit basis when more wind is added to the system. An example of this benefit for a large wind penetration is shown in Figure 6, where the benefits of consolidated operations over a large geographical and electrical footprint is more significant than portrayed for load alone in Figure 5. WECC Electricity Markets and Variable Generation Integration 26

27 What these figures show is that excess ramping, which is unneeded and costly, is significant when balancing areas operate independently. Some balancing areas must ramp generation up at the same time that other balancing areas are ramping down. If operations could be coordinated, much of this ramping, and the associated costs, could be eliminated. The figure shows that, with wind, the maximum unnecessary ramp is approximately 400 MW, and is matched by a -400 MW ramp. This bi-directional ramp requirement could be eliminated if the balancing areas combined their operations. Balancing areas can be consolidated either physically or virtually. Physically combining balancing areas is straightforward, but may not always be desirable. Two or more balancing areas can retain their autonomy and still capture much of the aggregation benefit by electronically combining their Area Control Errors (ACE). Each can control to an allocated portion of the combined ACE, assuring that reliability is met at lower cost. 10 Sub-hourly scheduling can provide fast access to neighboring markets, allowing trading of opposite ramping requirements. Figure 6. Combining balancing areas can reduce ramping requirements for systems that have significant wind and load. 10 There are several ways to allocate the combined ACE among the participating balancing areas but in all cases the required control actions are smaller than if the ACE signals were not combined. WECC Electricity Markets and Variable Generation Integration 27

28 Large, infrequent 5-minute ramps can be avoided when BA s are combined (either actually or virtually). 11 Figure 7 shows that this impact does not only apply to regions with large wind penetrations, but also may apply to loads only. Figure 7. Large, infrequent ramps can be significantly reduced by combining balancing areas. These results are corroborated by the New York State wind integration study which found that combined operation of the eleven zones in the New York State power system reduces hourly load variability by 5% and five-minute load variability by 55%. 12 (GE Energy, 2005). Hourly wind variability is reduced by 33% and five-minute wind variability is reduced by 53% with state-wide operations. Hourly system variability is further reduced by 10% and five-minute system 11 An Analysis of Sub-Hourly Ramping Impactsof Wind Energy and Balancing Area Size, Milligan and Kirby, 2008: 12 Hourly load variability shows the smallest reduction in variability (5%) when statewide operations are compared with zonal operations because loads are highly correlated on an hourly basis. Most loads increase in the morning and decrease in the evening. Statewide wind does not show that same similar pattern. WECC Electricity Markets and Variable Generation Integration 28

29 variability is reduced by 15% when wind and load are considered together. Note that while operating large balancing areas helps reduce the cost of wind integration, it also helps reduce the cost of serving load with or without wind (as pointed out in Figure 5 for the no-wind case). The benefits of large electricity markets apply to systems around the world. In a recent report for the International Energy Agency (Holttinen et al, 2007), the authors conclude Larger balancing area size and wind aggregation: both load and generation benefit from the statistics of large numbers as they are aggregated over larger geographical areas. Larger balancing areas make wind plant aggregation possible. The forecasting accuracy improves as the geographic scope of the forecast increases; due to the decrease in correlation of wind plant output with distance, the variability of the output decreases as more plants are aggregated. On a shorter time scale, this translates into a reduction in reserve requirements; on a longer time scale, it produces some smoothing effect on the capacity value. Larger balancing areas also give access to more balancing units. (page 107) The Value of Energy Markets Markets help economically and reliably integrate wind both in how they treat wind generators and in how they treat conventional generators. Markets that allow variable resources to sell excess energy or purchase shortages at transparent and fair prices accommodate the natural characteristics of wind while reflecting the true real-time cost of maintaining reliability. More generally, generation scheduling rules and the energy market structure itself are the most important factors in tapping the physical flexibility of the conventional generation fleet. Subhourly energy markets provide economic signals that make it profitable for conventional generators to respond to fluctuations in load and wind. Scheduling rules that restrict generators to hourly movements artificially constrain the conventional generation fleet, resulting in lost opportunities for those generators and increased costs for all. Markets that encourage conventional generation movement when it helps increase reliability and also allow generators to change output more frequently than once per hour reduce costs. Markets can provide direct economic incentives for generation flexibility if the current fleet does not have enough. RTOs and ISOs in the U.S. have fast energy markets, which result in a new economic dispatch every 5 to 15 minutes, depending on the market. It is these fast energy markets (not the RTO or ISO structure, per se) that make it possible to hold the regulating units closer to their preferred operating point because they can be brought back to the mid-point of their operating range much faster than if the redispatch did not occur for an hour. Therefore, there is less need for regulation in faster energy markets. This leads to a significant reduction in costs because regulation is typically the most expensive ancillary service. Thus, when calculating wind integration costs, features that reduce balancing costs will generally lead to lower wind integration costs. We emphasize that it is not the presence of an ISO/RTO that is critical for this analysis; rather it is the nature of the energy markets operating sub-hourly, along with the large geographic scope. The same basic operating improvements could be achieved by inter-ba coordination, involving fast energy markets and transfers between BAs. WECC Electricity Markets and Variable Generation Integration 29

30 Enhancing the flexibility of the conventional generation fleet helps to accommodate wind. This can involve valuing physical flexibility as other generators are built: fast start generators, lower minimum load capability, and high ramp rates are all valuable. Valuing this flexibility properly should result in (1) sufficient levels of installed flexible resources, with the ability to provide the desired ramping and minimum generation levels, and (2) the economic incentive to induce the operation of these flexible units when needed. These represent long-run and short-run price signals. We include two analyses: a relatively detailed analysis of the NYISO market, followed by discussion of results from several markets. We then turn to summaries of wind integration experience in Europe. 17. Load Following as a Byproduct of Fast Energy Markets or the Economic Dispatch Stack When wind is providing energy to its native BA, Milligan and Kirby (2009) have shown that there is no need for additional capacity relative to the no-wind case. It is clear that wind will likely impose additional ramping (dc/dt, the derivative of capacity with respect to time) requirements on the system. Ramping does not require additional capacity; rather, it involves using the existing capacity differently. Before discussing the impact of wind on ramping in more detail, we first examine the relationship between load following and energy markets. Our discussion can easily be applied to vertically integrated utilities that practice economic dispatch, however, for the discussion we focus on the case of an open energy market. To sell energy into a market, it is necessary for the generator to be maneuvered to the appropriate level of output. Base-load units are typically not required to maneuver much because they primarily supply energy at a constant rate. The unit commitment process selects the unit for operation; the unit is started and brought to its rated capacity over a period of hours. The low operating cost of base-load units (below the market clearing price) allows them to participate in energy markets by continuously operating at their full economic output. Intermediate and peaking units operate differently. Their somewhat higher operating cost is not always below the energy market clearing price so they do not participate in the market unless the clearing price is at or above their cost of generation. These units may not run at night when prices are low, for example, but instead operate and sell energy during the day. In order to sell into a given hourly market, the intermediate unit must position itself so that it can sell the prescribed energy for the duration of the market period. In subsequent market periods, the generator must again move to a new output level if it will sell more or less energy than in the first period, depending on its operating cost and the market clearing price. Failure to respond to market prices seriously hurts an intermediate unit. The unit incurs a lost opportunity if the market price is high and it does not turn on and ramp up to sell. Similarly, the intermediate unit loses money if the market price is below its operating cost and it does not turn off or minimize production. WECC Electricity Markets and Variable Generation Integration 30

31 Energy markets are able to obtain a great deal of load following response from intermediate and peaking generators at little or no cost to customers. 13 This is very different from regulation where the system operator must purchase a specific regulation ancillary service. Most of the time the intermediate and peaking units can position themselves as needed, and each interconnection has procedures to allow for the required ramps so that at the top of the hour, units have achieved their desired level of output. However, there are times when the generators can t move quickly enough, which results in very high-energy prices for a short duration of time. This situation can arise if the dispatch stack is not sufficiently deep or if sufficient ramping capability does not exist. Figure 8 shows an example of a base-load unit that is on the margin and is unable to ramp quickly enough between 8:30 and 9:00 to meet the load, which increases quickly during this time. Instead, fast maneuverable generation, such as a peaking unit with very high marginal cost, must be dispatched to cover the ramp. Once the base-load unit catches up to the load, the peaking unit is no longer needed. If the energy price is set based on the marginal unit, the price will rise from $10.00/MWh at 8:30 to $90/MWh from 8:30-12:30, and then fall back to $10/MWh. In this case, the price spikes because the marginal generator is not nimble enough to fully participate in the energy market starting at 9: Enery Price Clearly $10/MWH What is the Correct Energy Price? Enery Price Clearly $10/MWH MW 2800 Peaking - $90/MWh Total Load 4:00 4:45 5:30 6:15 7:00 7:45 8:30 9:15 10:00 Base Load - $10/MWh 10:45 11:30 12:15 13:00 13:45 14:30 15:15 16:00 16:45 Figure 8 Load following is a distinct service that is needed if ramping capability impacts the energy market unit selection. 17:30 13 This does not mean that generators provide the service for free. Instead, they profit from the resulting intra-hour price volatility WECC Electricity Markets and Variable Generation Integration 31

32 Conversely, if the load had followed the capability of the base-load unit, the required load following would have simply been extracted from the energy market. The generator would have been able to position itself to provide energy for load, and no distinct load following service would have been needed. Introducing wind into the example, Figure 9 shows that the ramp requirement may increase further. We have purposely provided a challenging case; the solid line shows the net load after wind energy is applied to the needed load. At 5:30, wind generation increases, which in turn decreases the capacity and energy required from the base-load unit. At 6:30, the wind reaches its maximum generation level, and remains there until 8:30, at which time the wind falls off quickly Enery Price Clearly $10/MWH What is the Correct Energy Price? Enery Price Clearly $10/MWH MW Total Load Wind Energy 4:00 4:45 5:30 6:15 7:00 7:45 Peaking - $90/MWh Base Load - $10/MWh 8:30 9:15 10:00 10:45 11:30 12:15 13:00 13:45 14:30 15:15 16:00 16:45 Figure 9 Wind energy can increase ramping requirements and provides additional energy, but does not require more capacity. When the wind picks up at 5:30, the base-load generation reduces its output. Although in our example we assume that there is sufficient downward flexibility, this need not be the case. When the wind begins to reduce output at 8:30, the ramp requirement is even steeper than in the no-wind case, which required a peaking unit to pick up the ramp. Clearly, in this example, wind has increased the need for load following and additional peaking capacity (or other fastramping unit) is required to maintain system balance. In this example, no additional capacity is needed to maintain system balance. The wind reduces the conventional capacity usage from 5:30-9:00 and imposes an additional downward ramp requirement on the base unit. Wind further exacerbates the ramping shortfall from 8:30-9:00, but does not require additional capacity. Instead, wind imposes a need for more flexible capacity. If that flexibility cannot be provided from the energy market, it must be provided by a load 17:30 WECC Electricity Markets and Variable Generation Integration 32

33 following market or by emergency provisions. 18. Market Evidence from New York Independent System Operator (NYISO) Data Existing energy markets provide evidence that load following can be provided by energy market movements. Clearly, all hourly markets follow the rise and fall in generation required to match the daily load pattern without resorting to a special load-following charge. Energy prices during the afternoon are simply higher than energy prices in the middle of the night. Data from subhourly energy markets show that faster generator response is also provided from energy markets without resorting to a special load-following charge. An examination of sub-hourly energy markets shows that the energy markets themselves provide an incentive for generators to respond to power system needs for movement, and they do it without incurring costs to customers. ERCOT, PJM, NYISO, ISONE, MISO, and CAISO 14 all operate sub-hourly energy markets, which are capable of responding to wind and load variability, and forecast error. We examined a year of sub-hourly price data from NYISO. From the analysis, we can conclude that fast energy markets, coupled with fast economic dispatch (5-15 minutes), provide a by-product capacity service in many times of the year. This may need to be supplemented by ramping markets as the penetration of VG increases substantially, and more work is needed on this issue. Wind does not move fast enough to constitute a contingency event (typically moving on the order of 1000 MW/hour even in overspeed events see Ela and Kirby) compared to a large generator which can trip within cycles. Therefore, a non-spinning product in the minute time frame may be needed to help manage large penetrations of wind or other VG (Walling, Banunarayanan and Miller, 2008). 19. The Benefits of Large Balancing Areas and Sub-hourly Energy Markets for Wind Of the various utility structures operating in the United States today, ISOs and RTOs provide the best environment for wind generation development. The reasons have nothing to do with the ISO/RTO structure itself. Instead, it is the large electrical and geographic footprints coupled with the presence of large open markets that provides these benefits for variable generation integration. The sub-hourly balancing markets can tap the physical maneuvering capabilities of the conventional generators. Balancing payments are typically based upon cost causation rather than on arbitrary penalties. 15 A summary of utility industry research by the Utility Wind Integration Group (UWIG - states that well-functioning hour-ahead and day- 14 Electric Reliability Council of Texas (ERCOT), PJM Interconnection (PJM), Independent System Operator of New England (ISONE), Midwest Independent System Operator (MISO), California Independent System Operator (CAISO) 15 See Milligan and Kirby and Wan: Cost-Causation-Based Tariffs for Wind Ancillary Service Impacts, American Wind Energy Association, WindPower 2006, June WECC Electricity Markets and Variable Generation Integration 33

34 ahead markets provide the best means of addressing the variability in wind plant output. This assertion is based on evidence from wind integration studies such as the one conducted by GE Energy (2005) for the NY ISO. Markets are also cited as helping with wind integration in the International Energy Agency (IEA) report on wind integration on large power systems (Holtinnen, 2007) and by Smith et. al., (2007). The UWIG document also says that, consolidation of balancing areas or the use of dynamic scheduling can improve system reliability and reduce the cost of integrating additional wind generation into electric system operation. This is also based on evidence from NY (GE Energy, 2005) and MN (Zavadil, 2006, and Milligan & Kirby, 2007). In this report we do not specify the time period over which the energy market should run. Markets in the U.S. have been run at intervals that include 5 minutes, 10 minutes, and hourly. Generally, faster markets reduce the institutional constraints for obtaining flexibility from physical generation. For example, a 15-minute market performs a new economic dispatch every 15 minutes. Units that participate in this dispatch hold their output flat during the 15-minute period, with the exception of ramps that are required to maneuver into the required position for the next market period. Other units will either hold output constant (because of market bids and generator costs) or will not participate because their cost exceeds the market price. All changes in demand that occur within the market period (15 minutes, in this example) must be met by units on AGC that provide regulation. Because the dispatch stack is frozen for the market period, all sub-market-period movements must be met by regulating units. This restricts the available pool of generation that can move within the market period to those operating as regulating units. For the purposes of this discussion, we call this the physically available but institutionally constrained generation (PAIC). Should there be excessively large ramps that exceed the capability of the regulating units, no mechanism would exist (without resorting to emergency measures) of tapping the PAIC generators in the economic dispatch stack. Shorter markets require more frequent dispatch. Many parts of the U.S. now run 5-minute markets. With this short market period, adjustments to the dispatch stack are made more frequently. Now consider a large load ramp (possibly accentuated by a simultaneous ramp in VG) that lasts for 30 minutes or more. With a 5-minute market, the output of the economic dispatch stack can be changed in 5 minutes, allowing the regulating units to return to their preferred operating points after each 5-minute period. However, with a 15-minute market period, the regulating units would have to chase the ramp for 15 minutes before being relieved by units on the dispatch stack. Consequentially, the regulating units must diverge from their POPs much more than before, and the PAIC units are constrained from ramping to help mitigate the ramp for a longer period of time. It is clear from this simple example that longer market periods impose additional institutional constraints, imposing longer time periods during which the economic dispatch stack cannot assist in chasing large ramps. The general principle is therefore that shorter market periods provide better institutional access to flexibility than longer market periods. WECC Electricity Markets and Variable Generation Integration 34

35 A recent study required by the Minnesota legislature to assess the reliability and cost of providing 20 percent of the state s electricity from wind provides a good example of how open markets can facilitate wind integration: The MISO [Midwest Independent System Operator] energy market also played a large role in reducing wind generation integration costs. Since all generating resources over the market footprint are committed and dispatched in an optimal fashion, the size of the effective system into which the wind generation for the study is integrated grows to almost 1200 individual generating units. The aggregate flexibility of the units on line during any hour is adequate for compensating most of the changes in wind generation. (See (Zavadil, 2006) 20. Wind Integration Costs Wind integration studies typically show lower wind integration costs for ISO and RTO markets than for non-iso/rto areas. As we pointed out above, it is not the inherent ISO/RTO structure that provides the benefit; rather it is the large, organized energy market that runs at short subhourly intervals and over a broad geographic area. These studies quantify the costs of additional reserves, changes in unit commitment and dispatch, gas nominations, etc. Integration costs are separate from energy and emissions benefits. Table 2 shows results from several recent wind integration studies (Smith et. al, 2007, Northwest Wind Integration Action Plan, 2007). 16 In general, the studies show lower integration costs in large market areas than in smaller, single-utility service areas. The integration costs for the three studies carried out in large energy markets range from zero to $4.41/MWh of wind, while the integration costs for the two non-market studies range from $8.84 to $16.16/MWh. The primary reason for these results is that the three large BAs operate sub-hourly markets, i.e. they dispatch generation on a five to fifteen-minute time frame, while the two small BAs require generators to follow hourly schedules and obtain all sub-hourly balancing from regulating units. The second reason for these results is the large size of ISOs and RTOs, which means there is much more conventional generation with ramping capability available to respond to changes in wind output while maintaining the balance between generation and load, thereby reducing wind integration costs. (ISO/RTO Council, 2007). 16 The quoted integration costs are actually the operating cost impacts. Some studies quantify additional wind related costs. WECC Electricity Markets and Variable Generation Integration 35

36 Table 2. Wind Integration Cost Study Results. Date Study Large BA/ Large Markets? Wind Capacity Penetration Integration Cost: $/MWh of Wind Output Energy Market Interval 3/05 NYISO Yes 10% Very Low 5 minute 12/06 Minnesota/MISO Yes 31% $ minute 2/07 GE/Pier/CAIAP (a) Yes 33% $0-$ minute 3/07 Avista No 30% $ hour 3/07 Idaho Power (b) No 30% $ hour (a) Includes two-thirds wind and one-third solar and includes cost increases of regulation and load following assigned to regulation. (b) Reduced from $16.16 in September, 2007, settlement proceedings. 21. Wind Integration is Facilitated by Energy Markets in Europe Evidence from the United States is corroborated by evidence in Europe. Denmark, Germany, and Spain have all integrated large amounts of wind generation into their power systems. Their experience points to the benefits of operating in a region with a robust spot electricity market (Holttinen, 2007). West Denmark already receives 24% of its electric energy from wind. Participation in the Nordpool spot market greatly helps wind integration. Holttinen notes that reserve requirements in the Nordic countries would have to be doubled if they operated as single countries compared to operating as a combined pool. Pool operation also reduces the need to curtail wind when there is excess production in one country. Denmark has not needed to increase the amount of operating reserves because of wind, but it does use the reserves more often. Holttinen (2007) shows that West Denmark has experienced periods when wind production exceeded load. Figure 10, taken from the IEA publication, illustrates one such time period. Note that wind generation exceeding load during light-load nighttime conditions at hours 13, and Wind generation nearly exceeded load during hours 156, 204 and 228. During the high-wind/low-load events, West Denmark was able to export the surplus wind energy to markets in Norway and Sweden using DC transmission connections. Since Norway and Sweden have hydroelectric generation with water storage capabilities, they use the imported Danish wind in lieu of hydropower and use the hydropower at other times. WECC Electricity Markets and Variable Generation Integration 36

37 4000 West Denmark January 3-15, 2005 MW Wind Load hour Figure 10. West Denmark can export excess wind power to other energy markets when wind exceeds load. North Germany receives 33% of its energy from wind. As noted earlier, (Table 1) forecasting errors are reduced by integrating wind generation across the four German regions. Transmission ties are being strengthened to increase the size of the region over which wind variability and load-following generation can be aggregated. Germany too has seen an increase in the use of operating reserves, but no increase in the amount of reserves needed. Spain, with 24% to 30% of electric energy coming from wind in various regions, has not required increased operating reserves though it also uses the reserves more often. Spain also derives large benefits from integrated interregional operations. The European Wind Energy Association (EWEA) advocates regional markets as an important policy for the integration of wind: the capacity of the European power system to absorb significant amounts of wind power is determined more by economics and regulatory rules than by technical or practical constraints. The European Commission cites the need for increased cross-border transmission links and increased liquidity in wholesale electricity markets as barriers to increased wind integration in its 2005 Benchmarking report. EWEA also notes that the large geographical spread of wind power will reduce variability, increase predictability, and decrease the occurrences of near-zero or peak wind output. (Van Hulle, 2005) 22. Adopting Limited Inter-Balancing Area Variability-Sharing In large parts of the western part of the U.S., there are many small balancing areas, and energy markets are not robust and are not operated on sub-hourly time frames. However, there is significant interest in developing cooperative agreements among balancing areas that would provide some of the benefits of consolidation. The Northern Tier Transmission Group (NTTG) ( developed an ACE Diversity Interchange (ADI) pilot program that allows for the sharing of regulation across regions. There is significant interest in this project. WestConnect WECC Electricity Markets and Variable Generation Integration 37

38 ( has joined the NTTG ADI project, and continues to investigate wholesale market enhancements and seams issues in the footprint. The National Renewable Energy Laboratory (NREL) has completed a large Western Wind and Solar Integration Study. 17 The focus of the study is the WestConnect footprint, but the entire U.S. portion of the Western Electricity Coordinating Council (WECC) was modeled, and high wind and solar penetrations were analyzed. Figure 11 shows the study footprint. Part of the analysis examined the benefit of consolidated balancing area operations for integrating a high penetration of renewable energy sources. In the Pacific Northwest, the Bonneville Power Administration convened stakeholders and utilities to examine how the region could best position itself to integrate up to 6,000 MW of wind that may be developed in the next several years. The result of this effort is the Northwest Wind Integration Action Plan ( Among the items on the agenda for the Northwest Wind Integration Action Plan are (1) developing more cooperation between regional utilities to spread the variability of wind more broadly; (2) developing markets that will reward entities who choose to market their surplus flexibility. Other parts of the report indicate a need for developing more robust markets for control area services that will provide needed electric services for smaller control areas with substantial wind resources (page 13) WECC Electricity Markets and Variable Generation Integration 38

39 Figure 11. NREL's Western Wind and Solar Integration Study focuses on the WestConnect footprint and models the U.S. portion of the WECC footprint. Although the outcomes of these various initiatives cannot be precisely predicted, they are further indication that when analysts consider how to integrate wind, market structure and design changes can offer significant benefits. The combination of regions in the Northwest and in WestConnect covers nearly the entire West that is not currently part of the California ISO or the Southwest Power Pool (parts of eastern New Mexico). 23. System Evaluation Tool Kirby, Milligan, Gramlich, and Goggin (2009) developed the System Evaluation Tool (shown in Figure 12), a spreadsheet based instrument for assessing how accommodating the structure of a balancing area (BA) or region is to the integration of large amounts of wind generation. It is especially useful in comparing regions or BAs. The judgments are necessarily subjective but the subjectivity is limited and provide some structure. The evaluator provides a numeric rating between 1 and 10 for each BA in the ten areas discussed previously: one is the poorest performance and ten is the best. WECC Electricity Markets and Variable Generation Integration 39