Fault Diagnosis of Wind Turbines Based on Improved Dynamic Network Marker

With the rapid increase in the installed capacity of wind power in China, early warning and identification of defects in wind turbine units has become crucial for the healthy development of the wind power industry. This paper proposes a comprehensive fault warning method for the entire wind turbine unit that is distinct from most current methods that focus solely on partial components. It addresses the complex coupling between various monitoring data from the supervisory control and data acquisition system and the internal defects of the unit. Utilizing a sample covariance matrix and the dynamic network marker theory, the method integrates diverse data inputs to predict potential failures. Case validation reveals that this approach maintains robustness across different operating conditions, issues warnings approximately 6 hours and 40 minutes before faults occur, and effectively identifies fault types. This capability is beneficial for initiative maintenance and rational planning of maintenance schedules in wind turbines.