Higher Order Spectra Analysis And Its Applications In Wind Turbine Fault Diagnosis

Wind turbines are usually operated in harsher environment.Low speed and heavy-duty operation conditions frequently cause malfunctions in the key components,such as the generator and the gear box.The external radiation signals of wind turbines are complex,non-stationary and non-Gauss.There are many nonlinear coupling signals.It is difficult to separate the fault characteristics by traditional processing methods.In order to effectively extract the fault features,this paper introduces high order spectrum into the fault diagnosis of wind turbine generator,and obtains the following results:The characteristics of the vibration signal of the gear box of the wind turbine are studied and analyzed.It is found that the coupling of vibration signals between multiple parts of wind turbine gearbox and the coupling of many pairs of gears is the main cause of a large number of nonlinear signals.The necessity and feasibility of extracting fault characteristics by analyzing and identifying the phase coupling of the signal is demonstrated.Compared with the commonly used detection technology based on temperature and vibration signal,the diagnosis method based on motor stator current analysis has the advantages of economy and easy implementation.Motor bearing failure causes rotor radial vibration and torsional vibration,making the motor stator current frequency modulates by phase modulation,and generates side band structure on the current spectrum.The analysis of the phase coupling information of these special frequency components can effectively diagnose the fault of the wind generator.High order spectrum is a powerful tool to deal with non Gauss and nonlinear signals.It describe the two phase coupling phenomenon quantitatively and completely suppressing Gauss noise in theory.Simulation and industrial actual signal analysis show that the improved bispectrum can effectively identify the sidebands generated by the signal amplitude modulation,and accurately extract the fault features.Summarizing the above research results,this paper proposes a fault feature extraction method that uses high-order spectrum analysis of wind turbine gearbox vibration signal and generator stator current signal and successfully applied to fault diagnosis of V52-850kW wind turbine generator.Industrial test results show that this method can effectively diagnose various types of wind turbine faults.