The Real-Time Wind Turbine Fault Diagnosis Method Based on Safety Evaluation Model

Abstract

In order to determine the best maintenance time of wind turbines and identify the fault type when it is the best time to do the diagnosis work immediately. The establishment of 4-level safety status model for critical parts of wind turbines, based on wind turbine parts’ significance level, was proposed. According to the corresponding safety level of the wind turbines in real-time working status, you can decide whether the wind turbine needs diagnosis at the time or not. Therefore, we should take measures to monitor the real-time working conditions of the wind turbine’s critical parts, confirming whether the critical part need the fault diagnosis analysis or not according to its real-time working safety status. If it is the right time, then the corresponding fault diagnosis process will be initiated, through which the real online fault diagnosis can be achieved. The multi-scale wavelet decomposition and Hilbert transformation was employed to get the useful parameters such as amplitude, effective value, mean value, kurtosis value and so on of the corresponding waveform to confirm the concrete diagnosis type.