| Rotor bearing system is the core component of rotating machinery.Due to the dual effects of unbalanced magnetic tension and fatigue crack caused by eccentric mass of generator rotor,the rotor bearing system shows complex nonlinear dynamic response.Furthermore,the system vibration signal is often mixed with noise signal,the original vibration signal distortion makes it difficult to extract fault characteristics.In view of this,this paper takes the rotor-bearing system with transverse cracks as a whole,establishes its dynamic model and motion differential equation,calculates the differential equation by using the fourth-order Runge Kutta method,and obtains the dynamic characteristics of the system with different parameters.At the same time,the method of combining variational mode decomposition with local tangent space permutation algorithm is used to solve the problem of noise signal adulterated in rotor lateral vibration signal;the fault characteristic extraction is further realized by combining the local mean decomposition of the one-dimensional signal after denoising with energy entropy.The specific work contents are as follows:(1)Discussed the calculation procedure of unbalanced magnetic tension and the expression with stiffness of the crack model,and the mathematical model of cracked rotor bearing system is also determined.The fourth-order Runge Kutta method is used to solve the motion differential equations of the system considering the unbalanced magnetic pull,and finally the dimensionless equations are obtained.(2)It is found that considering the influence of unbalanced magnetic pull can effectively delay the chaos or quasi period of the system,and the system can run smoothly in the middle and low speed region.The influence of speed,electromagnetic stiffness and crack angle on the non-linear vibration of the system is studied.The results show that the existence of electromagnetic stiffness is beneficial to the system in the middle and low speed region Stable operation;when the mass eccentricity is taken as the control parameter,the result only has the characteristics of double period and quasi period,and with the increase of mass eccentricity,the vibration amplitude of the system increases gradually;when the crack angle is taken as the parameter,the time domain diagram and Poincare diagram of the system are respectively 1/3 and 1/2 of the critical speed after the two ends of the cracked rotor system are added with the bearing affected by the nonlinear oil film force The crack characteristics can't be found on the axis locus and spectrum.(3)In this paper,the simulation signal composed of the system lateral vibration signal and Gaussian white noise signal is studied.It is decomposed into three modal components by the variational modal decomposition algorithm,and then an optimal component can be obtained by the double optimization of correlation coefficient and kurtosis.C-C method is used to choose the parameters of reconstruction phase space.According to the chosen component,a high-dimensional space is constructed and the high-dimensional data set of rotor fault is obtained.(4)Using the local tangent space arrangement algorithm to reduce the dimension of the high-dimensional data set,the main manifold is reconstructed after the low-dimensional manifold is obtained,and then the one-dimensional manifold is reversed and the time-domain waveform and frequency spectrum are made to show the effect of noise reduction.The results show that the algorithm can effectively reduce the noise signal and improve the signal-to-noise ratio.(5)The research of fault feature extraction based on local mean decomposition of noise reduced signals shows that the components after local mean decomposition not only have advantages in fitting the frequency of the original signal,but also can find the main set of fault features on the PF2 component by calculating the energy value and energy entropy of each component.The validity of the method based on local mean decomposition combined with energy entropy is verified by using the normal energy entropy.Therefore,in this paper,based on the consideration of the balanced magnetic pull,the nonlinear analysis of the rotor sliding bearing system with transverse cracks is carried out,and then the transverse vibration signal with white noise is decomposed by the variational mode and the appropriate components are selected for reconstruction,then the noise is eliminated by the method of local tangent space arrangement,and the one-dimensional time-domain signal after noise reduction is taken as the original signal Based on the combination of local mean decomposition and energy entropy,the research results can provide theoretical reference for vibration control,vibration signal noise reduction and feature extraction. |
PDF Full Text Request