NERC’s reliability oversight is bogged down on two fronts—standard-setting and compliance oversight. Progress depends on improving unwieldy process.
Balance of Power
Large grids can integrate more wind—without major burdens.
intervals instead of the current hourly scheduling procedure. This proposal, if adopted, would greatly enhance the ability of any balancing authority with limited internal flexible resources to integrate wind. The proposed rule would allow a balancing authority to look outside of its own boundaries for resources to offset unanticipated changes in wind output within a given hour. Every 15 minutes (within the hour) the balancing authority could change its intertie schedule by arranging for power supplies outside of its area to offset wind changes within its boundaries.
Early attempts to study the ability of a balancing authority to integrate wind required the modeling of possible future events. Now that several balancing authorities are actually integrating significant amounts of wind, the modeling can be calibrated with real-life experience. Modeling techniques vary from simple spreadsheets of 10-minute loads and resources to more elaborate dispatch models that attempt to replicate the commitment of generating units and dispatch decisions in time frames of less than one hour.
No matter what tool is used, the sub-hourly components are critical: forecast load, actual load, forecast wind and solar, actual wind and solar, and the resources whose output can be changed in slices of time of less than an hour’s duration. Black & Veatch analyzed the wind-integration capabilities and operating reserve requirements at several entities and found important differences between large, small, and wind-only balancing authorities.
• Large Balancing Authority: This example considers a large balancing authority (peak load greater than 6,000 MW) with a minor amount of existing wind. The authority is looking to add approximately 2,500 MW of equal amounts of wind and solar generation. The study was conducted with different assumptions regarding the ability to forecast load and scheduled hourly megawatts from wind and solar.
The study also looked at targeting the 90 percent of allowable control performance standard (CPS) in time periods of 10 minutes (CPS2 value) and targeting a more conservative 95 percent CPS2 value. 2 It was understood at the outset of the study that if all of wind or solar to be added was assumed to have the same hourly pattern, then there wouldn’t be any diversity associated with the added renewables. To avoid this unrealistic result, the study considered the likely locations of wind and solar facilities and the hourly patterns of generation.
For purposes of forecasting load and renewables, it was assumed that forecasts would be persistence-based. A persistence-based forecast derives its forecast value for an upcoming scheduled hour from actual values in a recent historical time period. For example, a two-hour persistence load forecast would estimate the load in the next scheduling hour based on the actual load two hours before the scheduling hour commences.
Cases were run with two-hour and one-hour persistence-based load forecasts and two-hour persistence-based forecasts of wind and solar generation. The study allowed for the possibility that certain percentage improvements could be made in the persistence-based forecasts. For example, a two-hour forecast was assumed without any improvement. Then sensitivity analyses were performed to determine whether this two-hour persistence forecast for load, wind and solar could