The Research On Fault Diagnosis Of Motor Bearing Based On VMD-KEI

Motor is an important driving equipment in modern industrial process.Rolling bearing,as one of the key parts of cushioning the complex force of high-speed running parts,is widely used in various motor equipment because of its small friction,easy start-up and speed-up,simple structure and convenient maintenance.Once such an important key part fails,it will seriously affect the stability of the motor,and even cause serious catastrophic accidents.Therefore,in-depth analysis of the rolling bearing failure mechanism,explore the fault diagnosis technology of the rolling bearing for the motor,can do the maintenance work in advance before the bearing has bad performance,which is of great practical engineering significance and application prospect for extending the effective working time of the motor.In this paper,variational mode decomposition(VMD)is introduced into the fault diagnosis of rolling bearings for motors.The feature extraction of non-stationary vibration signals of rolling bearings for motors based on VMD is studied in detail,and the diagnosis model of limit learning machine(ELM)optimized by genetic algorithm(GA)is fused to realize the accurate identification and diagnosis of rolling bearings for motors.Firstly,the basic principle of variational mode decomposition is briefly described.The simulation results show that VMD decomposition can effectively avoid mode aliasing and overdecomposition caused by EMD decomposition.Secondly,aiming at the sensitivity of VMD algorithm to initial parameters,a rough-to-fine parameter optimization strategy is proposed.The experimental data prove the effectiveness and practicability of the optimization strategy.In order to extract the fault feature information of vibration signals of rolling bearings for motor better,a KEI evaluation index based on symbolic dynamic entropy theory and set kurtosis is proposed.Vibration signals are decomposed and screened by variational mode decomposition combined with KEI evaluation index,and the best components are filtered in order to extract the fault feature information of vibration signals of rolling bearings for motor better.Rolling envelope demodulation is used to obtain fault feature information of rolling bearings.The experimental data processing and analysis prove that the proposed VMD-KEI method can extract fault feature information of vibration signals very well.Finally,the principle of extreme learning machine is briefly described.Aiming at the shortcomings of poor stability of extreme learning machine,genetic algorithm is used to optimize the parameters of input layer and hidden layer,and GA-ELM diagnosis model is established.Combining with VMD-KEI feature information extraction method,a fault diagnosis method of rolling bearings for motor based on VMD-KEI and GA-ELM is proposed.Experiments show that this method is feasible.Accurate diagnosis of rolling bearing inner and outer rings and rolling element faults is made.