Research On Detection And Fault Identification Technology Of Typical Rotor Vibration Signal

Rotary machinery is the most important type of machinery in modern production.which will cause great loss in case of accident.The rotor,which is the core component of a rotating machine,is also the main part of the failure.It often accompanies abnormal vibrations in the event of a failure.Therefore,the research of rotor vibration signal detection and fault recognition technology is of great significance for ensuring the stable and reliable operation of rotating machinery,avoiding huge economic losses and casualties,and improving economic benefits.Firstly,this paper proposes a rotor fault signal detection and identification method based on EEMD-SVM,and uses a rotor test bench to simulate and collect fault signals such as rotor imbalance,misalignment,looseness,rubbing,oil film whirl fault signals and analyzes the mechanism and characteristic manifestations of various faults.Combining the median-wavelet denoising method,an improved filtering method is proposed,which successfully suppresses the impulse noise and the random noise in the signal.Secondly,this paper analyzes the Empirical Mode Decomposition(EMD)method and principle,according to the EMD modal aliasing problem,put forward to the Ensemble Empirical Mode Decomposition(EEMD)method to suppress the modal aliasing phenomenon.The simulation signals are used to compare the advantages and disadvantages of EMD and EEMD in the suppression of modal aliasing.Based on the EEMD method,the rotor fault signal is analyzed.Meanwhile,a method based on EEMD energy entropy and Intrinsic Modal Function(IMF)energy feature to form feature vectors is proposed to achieve the purpose of dimensionality reduction of feature vectors and improve the classification efficiency.Then,the SVM multi-fault classifier was designed based on support vector machine(SVM)using directed acyclic graph(DAG)multi-classification method.Using the fault eigenvectors to build training samples and test samples to train and test the classifier,a good classification recognition effect is obtained.The effect of EEMD and EMD on the classification results was compared based on SVM.The BP neural network model is established,and the classification effect of SVM and BP neural network is compared.The results show that the EEMD-SVM method has a better effect on the recognition and classification of rotor fault signals.Finally,using LabVIEW as a software development platform,combined with sensors,data acquisition card and other hardware,a "rotor vibration signal detection and identification system" was designed and developed.That can achieve data acquisition,data analysis and processing,data saving and playback of rotor vibration signals.In addition,combining EEMD and SVM,through the mixed programming of LabVIEW and MATLAB to achieve the identification of the rotor fault signal.The function of the system was verified by a rotor test bench.The results show that the system can be well applied in the process of signal detection and fault identification of the rotating machinery rotor.