Reserarch On Fault Diagnosis Method For Gear Box Of Wind Turbine

Under the dual pressures of energy crisis and environmental pollution,vigorous development of renewable energy is a new development trend.The wind power industry is a successful commercialization case in the field of new energy and has developed rapidly in recent years.Wind turbines operate in harsh environments such as low temperatures,dust,and snow storms.Due to the gradual and random wind speed,the wind wheel and the blade will be impacted by the time-varying load.Because the components on the chain are rigidly connected,each component on the chain will also be subjected to frequent load shocks.Since wind turbines mainly adjust their control strategies according to changes in wind speed,the transrmission chain(especially gears and rolling bearings)is subjected to ever-changing load shocks and the impact strength is relatively large.According to relevant statistics,gears and rolling bearings in gearboxes are prone to failure in wind turbines.It is of great significance for wind turbine gearboxes to study fault diagnosis methods for gears and rolling bearings.The main research contents of this paper are as follows:(1)This paper mainly studies the gears and rolling bearings in the gearbox of wind turbines.It mainly analyzes the common failure modes of gears and bearings,and the vibration mechanism of gears and bearings and the vibration signals of different faults.Features were analyzed.(2)The modal aliasing phenomenon occurs when the non-stationary signal is analyzed for the empirical mode decomposition method,which directly affects the decomposition result.Therefore,this paper uses the decomposition method of the set empirical mode to eliminate the modal aliasing phenomenon.When components(gears,rolling bearings)in the gearbox fail,in most cases,different degrees of modulation occur,so that low-frequency fault information is modulated onto the high-frequency signal.Hilbert envelope demodulation and energy operator demodulation are introduced respectively,which can be used to extract low frequency fault information.The experimental results show that the error of the energy operator is small.Therefore,the emphasis of energy operator demodulation is used to extract the low-frequency fault information.(3)Due to the failure of the gearbox,the fault signal is susceptible to other factors(background noise)during the transmission process,so the difficulty of fault feature extraction is increased.For this problem,the adaptive random vibration method is used to enhance the problem.The weak fault feature is based on the noise of the vibration signal itself to reduce the effect of noise on extracting fault features.The vibration signal contains continuous vibration continuous oscillation component and transient impact component.For fault diagnosis,the transient impact component is more concerned.Therefore,the Q-switched wavelet transform based on signal oscillation characteristics is used to obtain the transient impact component,which reduces the difficulty of fault feature extraction.(4)It is difficult to obtain the fault samples in the field of wind power gearbox fault diagnosis.At present,there are few researches on the diagnosis of fault severity,so a gearbox fault diagnosis with improved K-means clustering algorithm is adopted.Since the K-means clustering algorithm is easy to fall into the local optimal solution,the particle mean group is used to optimize the K-means clustering algorithm.Usually,a sample can be described by using representative features,so the feature information is reduced by using principal component analysis.