Research On Fault Diagnosis Method Of Rolling Bearing Based On Wavelet Analysis And WOA-PNN

Rolling bearings are typical precision parts,but they are prone to failure in the process of use.Usually,the equipment failure caused by this part accounts for about 30%.In recent years,industry and science and technology develop very rapidly,and rolling bearings are developing toward high speed and large scale.A large number of studies have shown that the enhanced diagnosis of the component can better reduce its maintenance investment and thus obtain higher economic benefits.The analysis of the limitations of the current relevant diagnostic system has a positive effect on improving the maximum service life of the component and reducing the operating cost to the greatest extent,etc.,and also has significant practical value.At diagnosis in rolling bearing fault types relatively mature application is specific to only in the fault location,also is the common fault diagnosis model of four classification,and in the design engineering practice is not as long as the bearing has certain fault needs to be replaced,some equipment need bearing damage reaches a certain degree to be replaced,this requests us to design a high precision of the judgment of the rolling bearing damage degree of the model.In this study,based on the common four categories and taking rolling bearings as the object,the damage degree diagnosis model was built for them.Then,the existing fault diagnosis technologies were fully analyzed and a more scientific method was proposed,that is,WOA-PNN and wavelet transform were integrated to make scientific diagnosis of their faults.The experimental data in this paper are the rolling bearing vibration signals measured from Western Reserve University(USA)in the public data set.With the help of wavelet denoising technology,the real information of the signals is extracted to improve the signal-to-noise ratio.Then,with the aid of wavelet packet decomposition,different frequency band coefficients obtained,on the normalized processing,concrete can be defined as its characteristic parameters,it is used as the input vector of the PNN,and use of the latter's classification and adaptive function at the same time,with the aid of fault model of self-learning ability to constitute the part,then use the new testing,vibration signal to verified the validity of the model,and to optimize the ANN,to further improve diagnosis precision.The results show that the diagnosis method based on the combination of wavelet analysis and whale optimization probabilistic neural network can effectively detect and diagnose the fault damage degree of rolling bearings.