The Study Of Gear Fault Feature Extraction Of Wind Turbines Gearbox Under Complex Working Condition

With the unsustainable development of fossil fuels and the deterioration of environment,the search for alternative clean energy has become a pressing issue.As a renewable energy,wind power is growing into an irreplaceable position.However,as the service time of wind turbines(WTs)grows,the maintenance cost of WTs is also increasing year by year,and the faults of WTs gearboxes accounts for a large proportion.If the fault of WTs gearboxes can't be detected in time,it will cause undesired downtime and catastrophic failure.At the same time,due to factors such as strong gusts,WTs often operate in non-stationary state conditions.Therefore,the vibration signal collected from the gearbox is non-stationary.In order to reduce downtime and the maintenance fare,extracting the fault characteristics under complex conditions has become top priorities for WTs industry.In the view of the non-stationary failure of gearbox gears at home and abroad and the time consumption of traditional adaptive filtering,the scope of this paper is clarified.Based on the fault feature extraction method under traditional stationary condition,the fault feature extraction of WTs gearbox under non-stationary condition is studied.The fault feature extraction under the complicated working condition of the WTs gearbox is taken as the research focus.Based on the gear failure model under stationary state,the gearbox gear failure model under the non-stationary state is discussed.According to the Weibull distribution,the fault simulation signal of the gear in WTs gearbox under variable wind speed is constructed as well.The traditional adaptive feature extraction methods based on Morlet wavelet are analyzed and summarized.Based on the characteristics of adaptive wavelet filtering under stationary state,the impulsive wavelet is modified to make it more suitable for feature extraction of fault signals.At the same time,the factors affecting the efficiency of impulsive wavelet parameters optimization are analyzed.The adaptive impulsive wavelet filter based on adaptive filtering algorithm is improved from two aspects.Firstly,the fast algorithm combines the correlation coefficient maximum and the wavelet entropy minimum criterion is used to optimize the impulsive wavelet.The parameters optimization of the impulsive is divided into two steps instead of traditional simultaneous optimization method,thereby improving the calculation efficiency.Then,the time shift step is reduced,so as to reduce the amount of calculation.Simulation signals show that the proposed method has unparalleled advantages in terms of feature extraction and computation time compared with the traditional methods.In view of the fault characteristics of the WTs gearbox,an adaptive window segmentation approach is proposed.Firstly,based on the correlation coefficient method,the non-stationary signal is segmented into continuous segments,and then the improved impulsive wavelet algorithm is used to extract the features of the individual segments.Finally,the processed signals are spliced together.The effectiveness of the method is verified by simulation signals.Set up an experimental platform.STM32 microcomputer is used to control the servo motor to simulate the input shaft of the WTs gearbox,and the vibration signals at the end cover of the gearbox are collected.The effectiveness of the proposed method is verified under stable speed,the increasing speed,and varying speed conditions.