This article proposes the analysis and mining of SCADA operation data to improve the output performance of wind turbines. By analyzing the operating status and abnormal data types of wind turbines
an improved DBSCAN clustering method is adopted to clean and detect abnormal data caused by environmental changes. By using an improved moving least squares B-spline regression model
the control characteristic parameters corresponding to the maximum unit output at different wind speeds are identified and fed back to the unit control system. On the basis of unit simulation verification
validation and analysis were carried out under actual operating conditions of real units. Actual tests show that this method significantly improved system efficiency at rated wind speed.
WAN S T, CHENG L F, SHENG X L.Effects of yaw error on wind turbine running characteristics based on the equivalent wind speed model[J]. Energies, 2015, 8(7): 6286-6301.