Variability and uncertainty are the two attributes of wind generation that underlie most of the concerns related to power system operations and reliability. In day-ahead planning, whether it be for conventional unit commitment or offering generation into an energy market, forecasts of the demand for the next day will drive the process. In realtime operations, generating resources must be maneuvered to match the ever-changing demand pattern. To the extent that wind generation adds to this variability and uncertainty, the challenge of meeting demand at the lowest cost while maintaining system reliability is increased.
The primary focus of this study is to determine how the real-time operation of Idaho Power’s Hells Canyon Complex would be impacted by the addition of significant amounts of wind generation. Enough has been learned about the behavior of large amounts of wind generation – in either a few large plants consisting of many dozens to hundreds of turbines, or an equivalent number of turbines spread out over a large geographic area – to state that the impacts of wind generation uncertainty and variability on the bulk power system are primarily economic, and manifested in increased system costs. These costs are a consequence of the additional controllable generation capacity that must be allocated to manage the incremental variability of the Balancing Authority area due to wind generation, and the increased uncertainty that must be dealt with in operations planning.
The wind integration study ordered by the IPUC evaluates the operational and financial impacts of four levels of wind generation on three differing hydro condition years. The four wind generation levels are 300MW, 600MW, 900MW and 1,200MW. The three years are 1998, 2000 and 2005 which represent high, median and low hydro conditions respectively. The objective of the study is to evaluate the changes in operations and the resultant costs that wind variability and uncertainty introduces into the system at the four generation levels for each hydro year selected.
The major tasks in preparing the study consisted of:
1. Gathering wind data and building wind generation profiles.
2. Gathering and analyzing current generation and load data without wind.
3. Analyzing combined wind and load data and determining operational changes.
4. Modeling operational changes to determine economic impacts.
5. Evaluating the results.
DEVELOPING WIND GENERATION PROFILES
The study required detailed wind data from numerous locations over several years. Wind speed data of sufficient coverage and temporal resolution for constructing the chronological wind generation profiles is not available. Detailed meteorological simulations for the study years were developed for this analysis. The wind resource in Idaho Power service area is distributed across the southern third of the state (an Idaho wind resource map from the National Renewable Energy Laboratory (NREL) is included in the appendices). The wind generation scenarios considered for this study assess possible development projects throughout this area, along with potential projects in eastern Oregon.
The meteorological simulations employ a three-dimensional, physics-based model of the atmosphere based on the MM5 mesoscale model. When used for weather forecasting, the MM5 model is initialized with all known and available information about the current state of the atmosphere. This includes data from meteorological stations, balloon soundings, sensors on commercial aircraft, and satellite data. The model is then run forward through time, and variables relevant to the forecast such as temperature, humidity, wind speed, etc. are saved periodically for locations of interest. While the number of hours over the year at the highest wind generation production levels is limited, the amount of wind generation relative to system load over any period may still be significant. Figure 1 shows the amount of hourly wind generation relative to system load, sorted in descending order. This “wind generation penetration curve” shows that for over 4,000 hours in calendar year 2000, the amount of wind generation is greater than 5% (for the 300 MW scenario) to 20% (for 1,200 MW of installed capacity) of the hourly system load. At the extreme, wind generation can be as high as
25% (for 300 MW) to almost 80% (for 1200 MW) of the system load.
ASSESSING WIND GENERATION IMPACTS ON REAL-TIME OPERATION OF THE IDAHO POWER’S SYSTEM
The study evaluates the operational changes wind’s characteristics would have on Idaho Power’s current system. It is therefore important to understand the current operations and Balancing Authority compliance to be able to evaluate how wind generation changes the operations to maintain those compliance levels. Hourly load and hydro data were obtained from Idaho Power records for 1998, 2000 and 2005. The 1998 and 2000 loads are scaled to 2005.
Operationally, Idaho Power’s generating resources must have the flexibility during the course of an hour to manage:
• Variability in load and wind, and
• Differences between forecast and actual load and wind.
Maintaining this flexibility is essential in assuring system reliability and compliance with NERC performance standards (CPS1 and CPS2). Idaho Power presently maintains this operational flexibility to respond to unexpected and/or variable load conditions. With increased variability and short-term uncertainty due to wind generation, the required operational flexibility will also need to increase to maintain current levels of system control performance.
Using wind generation time-series data developed for this study and load data collected from Idaho Power Company archives, incremental requirements due to wind generation were determined for three categories of reserves:
• Regulating reserve – minute-by-minute requirement – about 1% of the installed wind generation capacity
• Load following capability – hourly load following
• Additional Operating Reserves – during the hour flexibility to cover deviations between the forecast and actual control area demand with their own supply resources. Wind generation adds to this short-term uncertainty, thereby increasing operating reserve.
Results of this analysis for the no wind case and the four penetration levels are shown in Table 1.
SIMULATING ANNUAL OPERATIONS OF IDAHO POWER’S SYSTEM WITH WIND GENERATION
While there is no formal or rigorous definition, “integration cost” is the term used to describe the economic impact of wind generation variability and uncertainty on the utility company charged with accepting and delivering that energy. The term applies to the operational time frame, which comprises the real-time management of conventional generating units and the short-term planning for demand over the coming day or days.
The Vista DSS Model is a hydro optimization program that simulates the operating characteristics of Idaho Power’s system. The model has detailed generating unit definitions, a simplified bus level transmission architecture and hourly inputs for hydro inflows, loads, electricity prices, reserve requirements and energy contracts. This software is capable of optimizing generation scheduling for the Hells Canyon Complex, while observing hydraulic, transmission and regulatory constraints on the system. The generation scheduling computed by the Vista DSS Model for the Hells Canyon hydro facilities includes generation from other Idaho Power resources as well as off-system market transactions.
The economic consequences of managing wind generation are computed through an analytical procedure that replicates an hourly simulation of real-time operation To elicit the integration costs attributable to wind generation, a reference or base case is developed where wind generation is stripped of the attributes responsible for integration cost – variability and uncertainty. This transformation of wind generation results in the delivery of a precisely-known amount of energy in a way that imposes the minimum burden on planning and real-time operations. This has been interpreted to be a flat block of known energy for the day, although the amount can vary day-to-day. Economic metrics from this case are compared to one where wind generation exhibits variability and uncertainty, and requires that additional generating capacity be set aside for its management. The difference between the cases, then, is attributed to wind generation integration cost.
Integration costs determined via comparison of cases with variable wind and equivalent wind energy delivery by a resource with no uncertainty or variability are summarized in Table 2. Simulations for calendar year 2000 revealed very high integration costs, which after additional analysis were determined to be a function of the anomalous market prices that were the result of the California power crisis.
The hourly simulations with Vista DSS show a strong correlation between wind generation integration cost and the market prices for electric energy. When normalized by the average annual market price, the integration cost curves for calendar year 1998 and calendar year 2005 are track more consistently.
It should again be noted that the focus of this study was the increased cost of Idaho Power’s daily operations due to the variability and uncertainty of wind generation. In all of the simulations, wind generation was treated as a “must run” resource, and the price paid per MWh of wind energy was not considered. The integration cost is simply the difference in the end-of-year operating economics between a case where wind generation is “ideal” and another where it exhibits the normal variability and uncertainty.
Previous studies of this type have found that “integration costs” for wind generation as defined here can be sensitive to the assumptions made. In addition, there are always uncertainties about the future, in terms of physical characteristics such a load levels and resource capabilities or institutional constructs which ultimately dictate how the power system must be operated. These uncertainties all have potential to influence the economics associated with managing a variable and uncertain resource.
It is appropriate, therefore, to recount some of the assumptions made to guide the analysis as well as other uncertainties that could affect integration cost. These include:
• Relicensing of Hydro Power Projects. In its long-term resource planning and for this study, Idaho Power is assuming that no reduction of the available capacity or operational flexibility of the Hells Canyon Complex will result from the relicensing process.
• Effect of Load Growth. Peak load in the Idaho Power Company service territory is growing twice as fast as the annual energy requirement. Going forward, then, this growth will lead to higher ramp rate requirements in the summertime and less available hydro capacity for managing the system. The cost of reserves would then likely increase, which could increase the integration cost for wind.
• Market Prices in the Pacific Northwest. The cost for managing wind generation is related to the market prices for energy, as shown in the analysis, and specially by the analysis for calendar year 2000 with the very high market prices. It was assumed for this study that the addition of wind generation in Idaho would not influence market prices, so that historical profiles could be used to represent other companies in the region. As more wind is considered and eventually developed in the Pacific Northwest, this assumption would not be correct. It is difficult to even conjecture how significant wind throughout the region (and even the interconnection) would influence energy prices, but it is almost a certainty that there would be significant effects.
• Market Structure and Operating Agreements in the Pacific Northwest. A major finding of this study was how the structure for in-the-day transactions with other utilities leads to an increased requirement for operating reserves with wind generation. As more utilities in the region ponder how the effects of wind generation can be managed, there is a possibility that operating agreements which seek to utilize geographic diversity and large aggregates of demand could substantially reduce integration costs. Possible changes were not included in this study.
• Improvements in Wind Generation Forecasting. Over time, improvements in wind forecasting could reduce integration costs.
• Transmission Limitations. In addition to constraining the hourly operation of Idaho Power’s system, transmission limitations can affect the provision and delivery of ancillary services. For example low-cost regulating resources in the region may not have accessible transmission capacity to Idaho Power’s system.
• Nature of Wind Generation Development in Southern Idaho. The wind generation scenarios constructed for this study are well distributed across the wind resource areas in the southern half of Idaho. Geographic diversity has a dramatic impact on the aggregate variability in the operational time frames. If actual wind generation development in Idaho is more concentrated in just one or two of the most favorable areas, the variability and uncertainty of the aggregate generation would be higher than what was considered in the study, almost
certainly increasing integration costs.
• Modeling of Reserve Requirements. In the hourly dispatch simulations, it was assumed that the total operating reserves were the same for each hour of the year. By using an average, the reserve amounts used are high when wind generation is low or zero, and short of what would be required during periods of substantial wind generation.
The previous uncertainties aside, there are substantial conclusions that can be drawn from the work reported here. These include:
• The costs identified in this study are the opportunity cost and generator loading inefficiencies introduced by adding operating reserves. These operating reserves are needed to cover the increased variability and uncertainty of Idaho Power’s Balancing Authority Area demand with wind.
• In general, incremental operating reserves required for wind generation fall into three major categories: 1) regulating reserve; 2) load following reserve; and 3) additional reserves to cover the expected short-term wind generation forecast error over the next hour. Due to the prevalence of hydroelectric generation, the current practice in the Pacific Northwest essentially combines regulating and load following reserve since they are generally provided by the same hydroelectric units.
• Wind integration costs are sensitive to hydro conditions and AGC constraint violations. AGC constraint violations are correlated to hydro conditions as shown in Table 2. Although more violations occurred at the higher penetration levels in all scenarios it was especially pronounced for 2005 which was the low water year studied. The Hells Canyon Complex did not have the fuel supply to manage the 900 MW and 1,200 MW wind scenarios.
• The results also suggest integration costs are partly a function of the disparity between heavy and light load pricing; given that one of the primary effects of the enhanced reserve requirement is an increase in light load generation and a decrease in heavy load generation.