Research On Fault Diagnosis Method Of Motor Bearing Based On Optimized Deep Belief Network

The rolling bearing is the core component of the motor rotation system,and it is also the component that is easy to break down.Because the fault of motor bearing has caused many serious casualties.An accurate and effective bearing fault diagnosis method can detect the fault in advance and prevent the occurrence of malignant accidents.However,with the development of electrical equipment towards the direction of complexity and integration,the traditional diagnosis method has been unable to meet the current needs.Therefore,in order to improve the reliability of fault diagnosis results,it has important theoretical significance and application value to study intelligent driving fault diagnosis technology for complex working conditions.In this paper,the rolling bearing of motor is taken as the research object.Aiming at the problem of bearing fault diagnosis under complex working conditions,I propose a fault diagnosis method of rolling bearing based on Variational Mode Decomposition with Double Threshold Central Frequency Method and Deep Belief Network(DTCF-VMD-DBN).In this method,Double Threshold Central Frequency Method(DTCF)is proposed to determine the optimal mode number of Variational Mode Decomposition(VMD).Hilbert transform is used to obtain the envelope spectrum of the intrinsic modal functions.The envelope spectrum is used as the input of Deep Belief Network(DBN)to realize fault diagnosis.Aiming at the influence of DBN initial parameters on diagnosis results,Quantum Difference Algorithm with global optimization capability is introduced.Multi Strategy Improved Quantum Differential Evolution Algorithm(MSIQDE)is proposed to optimize the initial parameters of DBN.Thus,a rolling bearing fault diagnosis method based on MSIQDE is proposed to optimize DBN(MSIQDEDBN).Finally,in view of the influence of the phase difference of the time domain signal on the fault diagnosis results,an adaptive phase signal processing method is proposed.This method can improve the accuracy of fault diagnosis effectively.The validity of the proposed method is verified by the data set of Case Western Reserve University and QPZZ-II experiment platform.The experimental results show that DTCF-VMDDBN can improve the accuracy of bearing fault diagnosis under complex working conditions effectively;MSIQDE-DBN can solve the influence of DBN initial parameters on the diagnosis results,and further improve the accuracy of bearing fault diagnosis;The adaptive phase signal processing method can solve the problem of input signal phase difference and improve the effect of fault diagnosis.