Lessons Learned from the Share-03 Exercise of the Power Curve Working Group
10 Sep 19
5:00 PM
-
6:00 PM
Tracks:
Data Science and Digitalization
The Share-03 exercise is the most recent intelligence-sharing exercise of the Power Curve Working Group, which focuses on advancing the modeling of wind-turbine performance in the real-world atmosphere. The goal of the exercise is to search for modeling methods that reduce error and uncertainty in power prediction during times of abnormal power production with extreme wind shear and turbulence intensity. Herein, this analysis includes a total of 55 submissions from 9 contributing organizations and various statistical tests. We assess the improvements of implementing four proposed candidate methods upon the Baseline method, which uses the conventional definition of power curve with wind speed and air density at hub height to predict power. Specifically, with statistical significance, the Density and Augmented Turbulence method reduces power-prediction error from Baseline in high winds during abnormal atmospheric condition, and the Density and Three-Dimensional Power Deviation Matrix method decreases both the errors and uncertainty in power prediction compared to Baseline when both the wind speed and turbulence are low. Moreover, in most of the submissions, the 10-minute data in non-average atmospheric conditions are at least twice as common as those identified as average, based on the current definition of partitioning the two conditions. Based on these findings, we recommend using a combination of power-prediction modeling methods to address different atmospheric scenarios.