Loss Factors for Small Distributed Wind Turbines Based on Field Data in the United States

While wind energy production loss due to turbine unavailability, environmental impacts, curtailment, and other causes has been studied and characterized at the utility-scale wind farm level, observation-based characterization of project loss is lacking for distributed wind energy, particularly for projects involving small wind turbines. Contemporary tools and research that support pre-construction distributed wind energy characterization present a wide range of default loss factors to convert gross energy estimates to net: 7–18%. Our goal is to use generation observations from operational distributed wind projects to develop more accurate representations of energy loss, along with an improved understanding of year-to-year loss variability, for this understudied sector of wind energy. Using a density-based filtering technique on distributed wind power generation timeseries, we determine periods of typical performance and use them with regression algorithms in a measure-correlate-predict fashion to simulate what the generation would have been during periods of atypical or unreported performance. From there, the actual versus predicted generation leads to the establishment of observation-informed loss factors (median = 17%) for small, single turbine installation distributed wind projects.