Research On The Diagnosis Method Of Rotor System Compound Fault

As the core component of rotating machinery,the fault diagnosis research of rotor system is of great significance to ensure the safe and stable operation of the mechanical system.In engineering practice,the signal collection of rotor system is usually obtained by two sensors of the same type installed vertically on the same section of the rotor.In this paper,with the help of full-spectrum analysis has the advantages of integrating homologous sensors,the faults of the rotor system are simulated by the Bently rotor test bench,and the corresponding full spectrum fault characteristics are also summarized,show the full spectrum analysis method has superiority to traditional Fourier analysis method in rotor system compound fault diagnosis analysis.In addition,combined with the good time-frequency expression characteristics of Hilbert-Huang transform(HHT),a fullHilbert spectral analysis method is constructed.Moreover,aiming at the defect of the fullHilbert spectrum analysis method,an improved method of the full-Hilbert spectrum analysis is proposed.Finally,it is applied to the fault diagnosis of rotor system to extract the characteristic information of rotor fault and some results are obtained.The main research contents are as follows.1.Simulation experiment and full spectrum characteristic of rotor system compound fault.The theoretical basis and numerical expression of the full spectrum method are introduced,and the effectiveness of the dual-channel signal fusion and the good single-channel compatibility are simulated and analyzed.Different types of faults(including single fault and compound fault)in rotor system are simulated by using Bently rotor test rig.The experiment includes crack fault,rub impact fault and crack-rubbing compound fault under the rigid support and oil film instability fault and oil film instability-rub impact compound fault under oil film support.At the same time,the full spectrum characteristics of various fault types are given and compared with the traditional Fourier spectrum analysis method.The analysis results show that the traditional Fourier analysis method is poor to extract the rotor fault feature information,and the full spectrum analysis method can extract the forward and reverse precession of the rotor system by fusing the information of the dual sensor channels in the same section of the rotor system,which is better to characterize the compound fault characteristics of the rotor system.2.An improved full-Hilbert spectrum analysis method is proposed and applied to rotor compound fault analysis.By using the advantages of full spectrum analysis to fuse homologous two-channel information and good representation ability for time-frequency of the Hilbert-Huang,the full-Hilbert spectrum analysis method is constructed by combining the two.Moreover,aiming at the defects that the full spectrum method directly extracted the positive and negative frequency components,an improved method of the full-Hilbert spectrum analysis is proposed Finally,it is applied to rotor fault simulation experiments,including rub impact,oil whirling,oil film oscillation and corresponding compound faults,and compared with the traditional Hilbert spectrum analysis method.The results show that the improved full-Hilbert spectrum method can fuse the homologous dual channel information of the rotor system.In addition to the information provided by the traditional Hilbert spectrum analysis,it can also characterize the forward and reverse precession information of the rotor system and express its time-varying characteristics,provides a more comprehensive and effective composite fault characteristic information,which is more conducive to the compound fault diagnosis of the rotor system.