Research On Fault Diagnosis Methods Of Motor Bearings Based On Improved Spectral Entropy Of Mathematical Morphology And SVM

Bearing is one of the core components of motor drive system,and it is also the most vulnerable component to failure.Accidents maded by motor bearing faults often cause great losses.Therefore,it is of great significance to study fault diagnosis technology of motor bearing for early detection of faults,prevention of accidents,reduction of economic losses and casualties.Fault diagnosis mainly includes the study of fault mechanism and fault diagnosis method.Nonlinear dynamics is the basis of studying the mechanism of fault occurrence,development and failure process.The key of fault diagnosis research is to select appropriate methods to identify the type,amount,position and severity of faults based on exploring the fault mechanism.Firstly,based on the study of fault mechanism,the non-linear dynamic model of rotor-bearing system rubbing,misalignment and rubbing-misalignment coupling faults is established,and the dynamic characteristics of the system under different fault degrees are analyzed,and the non-linear dynamic model of motor rotor-bearing system crack fault based on fractional calculus is established.The characteristics of time domain,frequency domain and axis trajectory are analyzed respectively.With the help of time domain waveform bifurcation diagram,the influence of order change on its dynamic characteristics is discussed in depth,and the selection of optimal order is studied.Then,based on the theory of mathematical morphological spectral entropy and the idea of higher-order difference,the identification method of bearing fault damage degree is studied.Combining with the mathematical morphological gradient function,the definition of higher order difference mathematical morphological gradient spectral entropy is proposed and applied to the identification of motor bearing fault damage degree.The simulation and experimental results show that this method not only improves the operation speed,but also improves the discrimination of different fault entropy values.In order to improve the accuracy of motor bearing fault classification,a fault diagnosis model of motor bearing based on empirical mode decomposition,fuzzy entropy,improved particle swarm optimization and support vector machine(CWPSO-SVM)is proposed.Empirical mode decomposition and fuzzy entropy are used to extract vibration signal characteristics of rolling bearings.Variable learning factors and inertia weights are used to improve the particle swarm optimization algorithm and to optimize the model parameters of support vector machine.The experimental results show that the model improves the generalization and classification accuracy of fault diagnosis methods.Finally,aiming at the long operation time of CWPSO-SVM model,a fault diagnosis model of motor bearing based on empirical mode decomposition,fuzzy information entropy,improved particle swarm optimization algorithm and least squares support vector machine is proposed.Because the calculation efficiency of the fuzzy entropy is low,the fuzzy information entropy is used instead of the fuzzy entropy to improve the training speed of the model without affecting the classification accuracy.Empirical mode decomposition and fuzzy information entropy are used to extract vibration signal characteristics.Particle swarm optimization is improved by using particle mutation strategy,variable learning factor and inertia weight.The parameters of least squares support vector machine are optimized by using improved particle swarm optimization.The experimental results show that the model has a faster diagnostic speed when the diagnostic accuracy is similar.