Research On Fault Diagnosis Of Rotating Machine Based On Spatiotemporal Intrinsic Mode Decomposition

According to the vibration signal fault diagnosis of rotating parts is essential in industrial production,and with the improvement of measurement technology,the collection of multi-channel signals will contain more comprehensive mechanical working condition information,the collected vibration signals are often multi-component nonlinear time-varying data,and the collected signals are often mixed with noise and other vibration sources have nothing to do with the vibration of rotating parts.Therefore,it is difficult to extract the components which can reflect the vibration impact of rotating parts.So it is very important to study the multi-channel nonlinear mechanical vibration signal.In this paper,the spatiotemporal intrinsic mode decomposition method is first applied to the field of mechanical fault diagnosis,and its feature extraction ability and frequency analysis ability for bearing and gear fault signals are studied,and the corresponding characteristic frequencies of various impact components are separated from the fault signals.The main research contents of this paper are as follows:The development history and research status of blind source separation and sparse decomposition are summarized.In view of the non-stationary nonlinear characteristics of multi-channel vibration signals of rotating parts,a spatiotemporal intrinsic mode decomposition method is proposed to process and analyze the fault components of vibration signals.Based on the data-driven time-frequency domain analysis method and the blind source separation method,the spatiotemporal intrinsic mode decomposition method can effectively extract the vibration characteristics of fault points from the mechanical vibration signals of non-stationary nonlinear rotating parts.The overcomplete dictionary base established by the empirical mode decomposition method and the nonlinear matching pursuit method are used to realize the effective separation of multi-channel signals.The separated component signal model can effectively reflect the vibration characteristics of the mechanical structure.Aiming at the characteristics of nonlinear multi-component time-varying signals of multi-channel rolling bearing signals and analyzing the bearing fault characteristics,a rolling bearing fault diagnosis method based on spatiotemporal intrinsic mode decomposition method was proposed.The simulation signal model based on the vibration characteristics of rolling bearings was analyzed,and compared with other blind source separation methods,which showed that the method can effectively separate various signal components in the separation of multi-channel signals,and better reflect the accuracy of decomposition and the effectiveness of impact separation in the time domain index.Through the analysis and comparison of the experimental signal data and the actual engineering data,it is verified that the method can separate the signal mixing component in bearings more clearly and accurately than other methods when dealing with the inner ring fault,the outer ring fault and the inner ring fault in the actual engineering.Aiming at the vibration signal characteristics of multi-channel rolling bearing complex faults containing various shock components and analyzing the vibration signal characteristics of complex faults,a rolling bearing complex fault diagnosis model based on the combination of spatiotemporal intrinsic mode decomposition method and fast spectral kurtosis was proposed.A simulation model was established according to the characteristics of composite fault signals of rolling bearings,and compared with other methods,the proposed method can effectively separate multiple impact components under the condition of low signal-to-noise ratio.In order to further verify the effectiveness of the method,the compound fault signals of bearing inner and outer rings collected in the experiment were analyzed,and the two fault characteristics could be accurately separated by the method.Aiming at the problems of large noise and difficult separation of modulating components in multi-channel gear faults,a gear fault diagnosis method was proposed based on the combination of spatiotemporal intrinsic mode decomposition method and second-order synchronous extrusion short-time Fourier transform method.The simulation signal model was established according to the gear vibration characteristics,and compared with other methods to verify that the proposed method can effectively separate various modulation components in the multi-channel gear mixed signal.By analyzing the experimental signal,the gear crack fault signal is diagnosed,and the fault characteristic frequency can be clearly located in the time-frequency diagram.