While both NERC and the NERC regional councils (known today as the Electric Reliability Organization) have standards and guidelines for resource adequacy and system reliability, much of the specificity as to how interruptible (e.g., demand-side) and intermittent resources (e.g., wind) are included is left up to the individual ISO/RTOs, states, provinces, etc. In fact, the various regions across North America each seem to have their own methodology for incorporating these resources into their resource adequacy and reserve-margin calculations.
As the North American energy industry escalates its desire to reduce greenhouse-gas emissions through the expanded use of demand-side resources and intermittent renewables, the importance of this topic also will escalate.
Alberta approaches its reserve margin differently than do all other ISO/RTOs. Instead of defining a reserve margin based on installed capacity, Alberta establishes a reserve margin based on effective generation capacity. Starting with total installed capacity, the Alberta Electric System Operator (AESO) subtracts out the following: mothballed capacity, a derating for hydro not available in the peak winter season, and a derating for wind because of its low capacity factor and variable output. For December 2005, the AESO established the reserve margin in this manner at 10 percent. What this figure means to the AESO is that when the differential between peak load and effective capacity falls to 10 percent, generation owners will add new capacity in response to market-price signals.
Wind generation in Alberta typically has a capacity factor in the 30 to 35 percent range, with output being higher during the winter peak. However, in calculating effective capacity, the AESO includes wind at a value of 20 percent of nameplate capacity. The AESO’s calculation of its total effective generating capacity does not appear to incorporate interruptible resources.
AESO has performed two in-depth wind integration impact studies that are available on the AESO Web site.1 The first study, completed in November 2005, identified the following operational and performance issues related to wind development:
• Increasing amounts of wind power on the interconnected system increases operational uncertainty;
• A significant portion of the wind power capacity may appear or disappear over a three- to five-hour period due to a “ramping effect”;
• Increases in installed wind capacity will increase the magnitude of the ramping effect and result in operational performance issues on the Alberta system; and
• For the AESO, operational performance issues are apparent at the 900-MW level of wind development (which is 10 percent of the AESO’s total installed capacity) and mitigating measures are required to maintain system performance at acceptable levels. These mitigating measures include: developing wind power forecasting techniques, increasing regulating reserves, increasing transmission reliability margins, and placing constraints on wind power facilities.
The second study, completed in July 2006, investigated each of these mitigating measures. After much discussion, AESO determined that the success of wind-power integration requires wind-power forecasting, although the other mitigation measures were not discounted. Forecasting was seen as the best first step. The majority of AESO stakeholders agreed, and most supported a centralized forecasting system, while others supported a decentralized forecasting system which may better support wind bidding into the market. The opinions as to who should pay for such a forecasting system were wide and varied.2
The perceived benefits of a wind-forecasting system include: improved system operation and reliability, more efficient use of regulating reserves and wind following, and less chance of constraining wind-power potential. In October 2006, the AESO issued two requests for proposals for a wind-power forecasting pilot project. The first phase of the pilot will forecast meteorological data and convert it into wind power data for specific locations in Alberta for one year. The second phase of the pilot will analyze the results obtained in Phase I. This first-of-its-kind wind forecasting pilot could establish new precedents, protocols, and procedures for the interconnection of large quantities of wind capacity onto the grids of North America.
To sell power into the New York market, NY-ISO requires all generators over 10 MW to meet certain requirements (e.g., net dependable capability, outage schedules), this includes both interruptible (demand-side) and intermittent (wind and solar) power resources. NY-ISO determines each generator’s unforced capacity (UCAP) based upon the data it submits. If a generator has committed to meet a certain amount of unforced capacity, but fails to meet that commitment, the ISO will impose various financial penalties.
Unlike the New York ISO, ISO- New England (ISO-NE) does not have a required reserve-margin criterion. Required resources are planned based on meeting the standard LOLE/LOLP reliability criteria of not disconnecting any firm load due to resource deficiencies, on average, more than once in 10 years.
ISO-NE uses a multi-area reliability model, developed by the ISO-NE staff, to assess the resource adequacy of the New England bulk power system. This model considers forecasts for demand, generation, imports, transmission, and demand-side management. The model considers hydro, but does not appear to consider wind generation.
New England currently has only two operational wind-energy projects, which produce about 12 GWh annually. As of June 2006, 12 wind projects, totaling 924 MW are in the ISO-NE generator interconnection queue. ISO-NE estimates that the capacity factor for the 11 onshore wind generators will be 25 percent, and the capacity factor for the one offshore wind generator will be 38 percent. ISO-NE does not believe that the renewable projects currently in its generator queue are adequate to meet the state renewable requirements of Connecticut, Massachusetts, Rhode Island, and Vermont. While the ISO-NE 10-year Regional System Plan, which is updated annually, discusses the future incorporation of wind resources, it does not appear to include wind in the modeling process.
PJM performs its Reserve Requirement Study annually using two probabilistic models (GE MARS model and an in-house PRISM model) to determine the reserve margin that satisfies the NERC reliability council LOLE standard of one-day-in-10-years.3 PJM applies Mid Atlantic Area Council criteria across its entire region, regardless of NERC reliability council boundaries.4 This model considers load, generation, imports and exports, and the reliability value of load-management programs. Inputs to this model are supplied by stakeholders (e.g., generators and distribution companies.) Generation statistics generally are based on the most recent five years of historical performance. New generating units roll actual performance data into their historical base as it becomes available. PJM’s current installed reserve margin is 15 percent.5 PJM classifies intermittent resources as either mature or immature. Mature resources are those with three or more years of operational history, and immature resources are those with less history. The effective capacity of mature wind farms is the mean average of three single-year capacity factors. For immature wind farms, PJM assigns an effective capacity of 20 percent, although this percentage is subject to change upon operational data to the contrary.6
The Midwest ISO encompasses all, or portions of, at least four NERC regions: ECAR, MAIN, MAPP, and SPP, and about 15 states and the province of Manitoba.7 The Midwest ISO is in the process of consolidating its 26 balancing authorities, which may change the way the region’s reserve margin is calculated. However, for the foreseeable future, the following applies.
The Midwest ISO does not require a set reserve margin for its entire territory, although discussions are underway to do so eventually. Instead, each load-serving market participant is required to continue to follow the reserve requirements of their individual state(s) and NERC region(s).
Each generation resource that sells power into the Midwest ISO market must pass a deliverability test and be designated as a network resource. As long as this test is passed, any generating resource can participate in the Midwest ISO market. The only specific criterion for wind generation is that “all wind farms’ maximum output should be reduced to 20 percent in the model because only 20 percent of wind farm’s maximum output can be counted for capacity purposes, unless demonstrated otherwise.”8 The Midwest ISO considers interruptible demand as an alternative capacity resource and requires only that the resource meet the adequacy requirements of the state, province, and NERC region where the resource is located.
The minimum reserve margin for ERCOT is 12.5 percent, as set by the Texas Public Utility Commission (PUC). ERCOT calculates the reserve margin annually. ERCOT’s calculated reserve margin for 2007 is about 15.2 percent, but is expected to decline over the next few years until new proposed generation is built.9 ERCOT complies with the same LOLE/LOLP standards in calculating its reserve margin as the other NERC regions.
To calculate its reserve margin, ERCOT considers all installed generation, private co-generation facilities, imported capacity, large loads under interruptible contracts, mothballed units (assuming a 1-year restart time frame), and planned units with a signed interconnection agreement. While wind generation is considered, ERCOT currently counts only 2.9 percent of a wind farm’s capacity toward reserve margin because the maximum wind output from West Texas does not correspond to peak demand (4 p.m. to 6 p.m. in July and August).10 In its 2006 report to the Texas Legislature, the Texas PUC says that wind generation historically has supplied, on average, only 2.6 percent of its rated capacity during summer peaks.11 ERCOT re-evaluates the wind-farm capacity percentage every year based on historical data. In 2005, ERCOT was using a 10-percent effective capacity for wind. In December 2006, Global Energy did a study for ERCOT that used a probabilistic model to estimate that the effective capacity of wind generation is roughly 8.7 percent of nameplate capacity.
Under California’s proposed Market Redesign Technology Upgrade (MRTU), the coordinators for load-serving entities (LSEs) will have to submit annual and monthly resource adequacy and supply plans to the ISO, which will validate and coordinate all plans. The types of resources that qualify for resource-adequacy consideration include: Generation within and outside of the Cal-ISO control area, resources under construction, demand-side resources, a portfolio of physical resources with a common bus-bar (e.g., multiple wind generators), and more. To qualify for consideration, a resource must pass a deliverability test (e.g., the output can reach load under peak conditions.) Each LSE must meet at least a 15-percent reserve requirement. The supply plans must include details about the resources the LSE has contracted for and each generating unit’s net qualifying capacity value. Intermittent resources, such as wind and solar, are considered limited-use, non-dispatchable resources and are expected to schedule into the day-ahead and hour-ahead markets, rather than the real-time market. Cal-ISO will not penalize these units for failing to meet their expected commitment. However, the ISO will track the performance of these resources and use that to understand the variances between planned and actual contributions. This tracking is vitally important because California has some of the nation’s most ambitious targets for the reduction of greenhouse gases through the employment of both interruptible (e.g., demand-side) and intermittent (e.g., wind) resources.
1. “Wind Integration Impact Studies,” Phase I, November 2005, and Phase II, July 2006.
2. “Wind Integration Stakeholder Workshop,” posted presentation, AESO Web site, July 19, 2006.
3. The PRISM model is a legacy FORTRAN program which has been used by PJM for adequacy studies since the mid-1960’s. PRISM stands for: Probabilistic Reliability Index Study Model.
4. MAAC is a NERC region called Mid-Atlantic Area Council. It recently has been incorporated into the new, larger NERC region called Reliability First Corporation.
5. 2006 Reserve Requirement Study, letter to the PJM Planning Committee from the Capacity Adequacy Planning Department, April 19, 2006.
6. PJM Manual 21, Rules and Procedures for Determination of Generating Capability, Appendix B-1: Determination of Wind Generator Capacity Values, Aug. 15, 2005.
7. The ECAR and MAIN regions recently have joined and are now called RFC, Reliability First; and the MAPP region is now called MRO, Midwest Reliability Organization.
8. “MISO Generation Deliverability Study Method,” Midwest ISO, Version 2, Aug. 11, 2006.
9. “Electric Reliability and Resource Adequacy Update,” Presentation given to the Senate Committee on Natural Resources, ERCOT, July 13, 2006.
10. Generation Adequacy Task Force Report to the Technical Advisory Committee (TAC), ERCOT Web site, April 2007.
11. TX PUC Report to the 80th Texas Legislature, “Need for Transmission and Generation Capacity in Texas,” p. 13, December 2006.