Rotor Rub-impact Fault Identification Based On Acoustic Emission Signal

As the main mechanical equipment at present,rotating machinery is frequently used in power generation,metallurgy,aviation,military industry and many other fields.As one of the key components,whether the rotor system can operate normally and safely is closely related to the lifespan of the equipment.With the increasing speed and efficiency of rotating machinery,the rotor system is prone to rub-impact failure.If the failure is not found in time at the early stage,it will cause abnormal operation of mechanical equipment,damage internal structure,reduce work efficiency and production quality,and even cause serious consequences such as casualties.Therefore,in order to ensure the stable operation of machinery and equipment,as well as the health of personnel and property safety,it is extremely important to prevent the rotor rub-impact failure and identify it at the initial stage.Taking the rotor system in rotating machinery equipment as the research object,this thesis studies the rub-impact phenomenon.In this thesis,acoustic emission detection technology is used to analyze the mechanism of rotor rub-impact failure and the acoustic emission characteristics of metal materials.It designs experiments and builds a simulation experimental platform for rotor rub-impact failure,which collects the acoustic emission signals of rotor without rub-impact,slight rub-impact and severe rub-impact at three different speeds,and then analyzes the waveform characteristics of signals in time and frequency domain.This thesis proposes a signal decomposition method based on Sparrow Search Algorithm(SSA)to optimize the parameters of Variational Modal Decomposition(VMD),which solves the problems of modal aliasing and unobvious component characteristics in classical Empirical Modal Decomposition.The energy value,kurtosis value and energy entropy of each component of the acoustic emission signal decomposed by SSA-VMD are extracted to construct feature vectors for subsequent fault identification.The results show that the characteristics of each modal component are clear and obvious after SSA-VMD decomposition.This thesis adopts the method of support vector machine(SVM)in pattern recognition to identify the rotor rub-impact failure,and the radial basis function(RBF)is used as its kernel function.In order to improve the recognition accuracy of the SVM model,a method of optimizing SVM parameters and RBF parameters by SSA is proposed.By training and classifying the extracted feature samples,the results show that the SSA-SVM method has high accuracy in rotor rub-impact failure identification and can be applied to practical work.