Research On Fault Diagnosis Of Rolling Bearing Based On Multi Domain Features Fusion

The rolling bearing is the key part of the CNC machine tool to support the revolving parts.Once a bearing fails,it will seriously affect the normal operation of machine tools,and may cause major safety accidents when severe.The data show that at present,thirty of the faults in CNC equipment are caused by bearing failures.Therefore,it is of great economic and practical significance to carry out the fault diagnosis research of the rolling bearing of CNC machine tools and monitor the running state of the bearing of the machine tool.The traditional fault diagnosis method of rolling bearing is usually based on the time domain characteristics of the vibration signal,or simply from the frequency domain,so it can not reflect the characteristics of the fault signal.In addition,the fault feature is not a simple linear relationship with the failure mode.Therefore,it is difficult to solve the problem of rolling bearing fault diagnosis by single domain feature method or experience.This topic comprehensively and systematically studies and analyzes the vibration signal characteristics and fault mechanism of the rolling bearings.The wavelet packet technology is used to denoise the vibration signal of the bearing.The characteristics of signal in time domain and frequency domain are comprehensively analyzed.The dimensionless parameters in the two domain of time and frequency are selected as the feature parameters of fault diagnosis comprehensively,so as to effectively and comprehensively reflect the fault characteristics,and lay the foundation for the accurate diagnosis of faults.The bearing fault mode is identified by the self-learning,self-adaptive and powerful nonlinear ability of the neural network.The HS71909-C-T-P4 S angle contact ball bearing in the transmission system of a vertical machining center of a certain enterprise model LGMazak VTC-16 A is studied.According to the characteristic parameters of the bearing fault and the characteristics of the fault type,the corresponding BP neural network model is designed and established.The AIC9000 multi-function rotor system experiment platform produced by the Beijing Aerospace intelligent control and monitoring technology research institute is used to collect the experimental data and construct the training samples of the network model.Using MATLAB software platform to train and test the network model,verify the feasibility and effectiveness of the fault diagnosis method in this paper.The experimental results show that the rolling bearing fault diagnosis method is effective and can effectively identify the fault types of bearings.