Improvement of Wind Power Forecasts in Space and Time

• Activity Number: 138
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Lisa Bramer*+ and Petrutza Caragea and Mark Kaiser
• Companies: Iowa State University and Iowa State University and Iowa State University
• Keywords: wind forecasts ; wind power ; bias correction ; hierarchical model

Abstract

The ability to accurately forecast day-ahead wind power production on an hourly level has become increasingly important as more wind power is introduced into electricity markets. Conversion from forecasted wind speed to forecasted energy is typically made using a discrete power curve provided by the turbine manufacturer. Manufacturer's curves were observed to be systematically biased for given wind speeds at several wind farms in Iowa. We model the bias of the manufacturers power curve using a hierarchical structure over space and time. Our model makes use of meteorological covariates related to wind power. We show that the hierarchical model the manufacturer's curve and accounts for the continuous nature of wind speed rather than providing forecasts for a discrete set of wind speeds. Additionally, the hierarchical structure provides the added benefit of obtaining variability estimates, which can provide valuable information for energy companies as they make day-ahead energy bids.