Diagnosis And Dynamic Analysis Of The Rotor System With Imbalance-Misalignment Coupling Faults

As key components in rotating machinery, to ensure the operation of the rotorsystem to be normal has a great significance to the safe operation of the entireequipment. In rotating machinery, bad manufacturing and assembly or the systemoften in the condition of high speed and over-load, it is easy to appear imbalance,misalignment fault, if these faults cannot be detected and diagnosis, it will causesecondary fault such as rubbing and loosen or even mutual coupling faults, all of thesewill cause significant influence on the machine, however, once the system fall intofailure, often is not in the form of one, but the mutual coupling faults. Therefore,timely and effective way to discover and diagnose the location and type of the faulthas great significance for the normal operation of equipment. In this regard, this paperhas made the following research.At the very first, the mechanism of the imbalance fault and misalignment faultwere studied, the finite element model of the imbalance, misalignment and theimbalance-misalignment coupling faults were built by using the rotor dynamicsanalysis software DyRoBeS.The critical speed and steady state synchronous response of the system wereanalysed in this paper, by using the Newmark_β numerical integral method to obtainthe vibration characteristic of the coupling faults, the moment of inertia, sheardeformation and gyroscopic forces on the system were fully taken into account whenanalysis. And also the influence of unbalance and different rotation speed on thecoupling fault of the whole system were taken into consideration when analysis. Byanalyzing the trajectory of the axis of the rotor system, the time-frequencycharacteristics, the vibration response characteristics at both ends of the bearings andwaterfall chart of the system vibration response, found that with the coupling faults ofrotor system has a wealth of features.Analyzed several typical manifold learning algorithm in the application of faultdiagnosis, in-depth studied of the mechanism of three kinds of manifold learningalgorithm as FDA, LFDA and IOLFDA fault diagnosis, then proposed the new methodwhich is based on wavelet packet and manifold learning method for fault diagnosis.Select the simulation analysis of vibration data and axis trajectory data as a sample,using the five kinds of manifold learning algorithms to intelligent diagnosis samples,to found that the method of LE, LPP can’t distinguish the faults, FDA and LFDA identify better, but can’t completely diagnose specific faults, while IOLFDAdiagnostic work best.The simulated experiment of imbalance, misalignment and imbalance-misalignment coupling faults were done on the basis of the comprehensive faulttest-bed. Through the experimental analysis verify the validity of the simulationresults. The imbalance fault response with one time frequency and the misalignmentwith two times frequency, while the imbalance-misalignment coupling faults featurethe characteristics of the imbalance fault.Finally, applied the method proposed in this paper which is based on waveletpacket and LFDA algorithm of fault diagnosis method to the coupling faults diagnosis,contrast the effect of fault signal diagnosis of the five method with LE, LPP, FDA,LFDA and IOLFDA, to found that the method of LE, LPP can’t distinguish the faults,FDA and LFDA identify better, but can’t completely diagnose specific faults, whilewavelet packet and LFDA algorithm of fault diagnosis method work best, and thisgives a new analysis method for the coupling fault diagnosis.