Research On Wind Power Generation Performance Monitoring Method Based On Statistical Control Chart

Anomalies in wind turbines may reduce their power generation capacity and affect the power supply to the grid.In order to reduce the losses caused by fan anomalies,wind turbine condition monitoring and anomaly identification are increasingly important.Aiming at some problems left in the field of fault diagnosis,under the premise of wind turbine data analysis,this thesis proposes a method based on statistical control map to monitor wind power performance.Based on long-term research on the fan power curve,it has been recognized that the curvature and shape of the power curve of a normal fan is variable rather than fixed.Previous monitoring methods based on fixed reference power curves may generate false alarms because there is no calculated curve change.Therefore,this thesis mainly studies the statistical control chart based on the fan power curve.Through the statistical control chart,the curvature and shape of the power curve are monitored and analyzed to determine whether the performance of the fan is abnormal,so that the fan can be safely maintained.This thesis first introduces the background and significance of the research,and then introduces the implementation process of the wind power performance monitoring method based on the statistical control chart.Finally,the actual wind turbine data is used to simulate the research method to judge the practicability of the monitoring method.The specific research contents are as follows:Firstly,the method of fitting the wind power curve.According to the characteristics of the fan power curve,two fitting methods are proposed:linear fitting of power curve and Weber fitting of power curve.Through the simulation analysis of the two fitting curves,the fitting effect of the two fitting curves is judged,so that the fitting curve with better effect is selected for use.Secondly,based on the one-dimensional control chart study of the shape of the fan power curve.In order to monitor the change of the shape of the power curve,an I-MR control map and an MR control map are proposed,and the wind power generation performance is identified by monitoring the change of the fan power curve shape.These two control charts are mainly established by statistically analyzing the changes in the residual curves in the fitted curve and the power curve.Thirdly,based on the study of the binary control graph of the curvature curve of the fan power curve.In order to monitor the change of the curvature of the power curve,a binary statistical control chart is proposed to identify the wind power performance by monitoring the curvature change of the power curve.The control chart is mainly established by statistical analysis of the binary parameters obtained by fitting the curve.Fourthly,based on the study of the covariance matrix in the binary control graph.In the monitoring of wind power performance in binary control charts,it is necessary to keep the covariance matrix in a controlled state.In order to monitor the variation of the covariance matrix in the binary control graph,a statistical control graph based on the covariance matrix is proposed,and the covariance matrix is monitored.Changes to identify anomalous data sets.The control chart is mainly established by statistical analysis of the relationship between the matrix feature vector and the feature root.