Wind turbine power performance test methods based on least squares prediction

Abstract

Wind turbine power performance is an important process of type approval testing project. Because the period of power performance testing is long and it is difficult to collect complete data in a short time, the wind turbine power performance test methods based on least squares prediction is proposed. Firstly, this paper introduces the testing process, site calibration, uncertainty assessment and common data analysis methods of wind turbine power performance using the least square prediction to achieve linear regression and polynomial regression based on limited data. Then the comparative analysis results are carried out through full curve and predicted curve in the actual case to prove the effectiveness of this method in power performance prediction.