Abnormal State Analysis of Wind Turbines Based on the Power Curve
A. Li Yuan, B. Huang Qiujuan, Chenle Yi, D. Xing Zuoxia · 2018
For the difficulty of operation and maintenance of wind turbines, anomaly detection technology was derived to identify faults early. In the research of power curve modeling, multivariable and nonlinear problems need to be involved. This paper proposes a wind generator abnormal state analysis method based on the power characteristic curve. Based on the data of wind turbine SCADA system, double threshold is used for data cleaning and Bin method is used for data packet in this method. Using nonlinear fitting to set up a parametric Logistic mathematical mode of the power curve. The control pattern is used for the operating condition's monitoring and abnormal analysis, and limits of the control pattern are obtained by calculating the average residual and standard deviation. Finally, simulation results verify the effectiveness of the wind generator abnormal state analysis method based on the power characteristic curve. In conclusion, to identify the abnormal state of the wind turbines, this thesis studies correlation between the historical data of the wind farm and power characteristics of wind turbines, and the modeling method. It is not only an attempt and exploration of the related theories and technical methods to wind power big data, but also provides a basis for the performance evaluation of wind turbines and has practical application value.