Abstract: In wind power production, wind power generation is limited by load demand variation. This usually leads to wind curtailment. It is very important to identify wind curtailment data because their existence has huge impact on wind power prediction and wind farm equivalency modeling. Characteristics of normal operation data and wind curtailment data are analyzed and compared. Based on standard wind turbine speed to power transfer curve, concept of viscosity range is proposed according to the mean and standard deviation of normal distribution. A wind curtailment data identification and elimination method is established. The application on one of wind farm in Northeast area validates the effectiveness of proposed method.

Key words: wind power, power curve, viscosity range, wind curtailment, probability density