Research On Fault Diagnosis Method For Roller Bearing Stator Current Of Wind Turbine

Rolling bearing is an important part of the drive system of wind turbines and one of the components with the highest failure rate.It is of great significance to study the fault diagnosis of rolling bearing to the reliable and stable operation of wind turbine.At present,Traditional gear fault diagnosis methods,for example,vibration diagnosis method needs to install additional vibration sensors on the equipment body,which not only damages the equipment body,but also is greatly affected by the environment and noise.The stator current analysis method can directly diagnose the mechanical faults such as bear by the electrical signals of the system.Its signal acquisition method is simple and robust,and the monitoring system has low cost and high reliability.It is especially suitable for fault diagnosis of wind turbine.This paper first analyzes the failure mechanism and several common failure modes of wind turbine bearing,and deduces the failure characteristic frequency formula of bearing in different positions.The basic principle of ensemble empirical mode decomposition and empirical mode decomposition and the basic principle of fault classification method support vector machine are introduced.EEMD and EMD algorithms are compared and analyzed to verify the superiority of EEMD.Secondly,a simulation model of the generator set system is established in the Matlab / Simulink environment.Through calculation and analysis of the impact torque during the rolling bearing failure process,the bearing is simulated in the fault state and normal state.Then the collected simulation data are analyzed by combining EEMD with envelope spectrum.Finally,a simulated rolling bearing test bed composed of AC motor,rolling bearing and load generator is built.The internal and outer ring faults and normal current signals of rolling bearings at different power supply frequencies are collected respectively,and these current signals are processed by EEMD method.Then the sample entropy is extracted from the first four valid IMF components of the EEMD as fault features,and these four approximate entropy are input to the SVM classifier to complete rolling bearing fault classification.Through simulation analysis and experimental results,it is proved that the instantaneous fluctuation of torque will occur when the ball of rolling bearing passes through the defect position of inner and outer ring,which will cause the change of stator current.By using the signal processing method of EEMD and envelope spectrum,the fault frequency bearing of rolling bearing can be put forward in the weak stator current signal.Finally,the classification accuracy of support vector machine(SVM)for small sample bearing fault samples is verified.