In order to obtain aerodynamic stability and high thrust ratio,the twin spool rotor structures are widely adopted in modern aero engines,and the inter-shaft bearings are used to support the rotor system.The working condition of the inter-shaft bearing is complex that could lead to failure of bearings.Therefore,the inter-shaft bearing is important for reliability of aero engines.Accurate diagnosis of inter-shaft bearing fault is significant for the operation of aero engines.In this paper,the main content includes the following points:In the first place,it is hard to obtain bearing fault features for the long transfer path.The inter-shaft bearings locate between high speed rotor and low speed rotor,and then leading a heave noise of background.In this paper,the wavelet de-noising method is adopted to reduce the noise of vibration signals and acoustic emission signals.The vibration signals and acoustic emission signals are decomposed into several narrow intrinsic mode function(IMF)by ensemble empirical mode decomposition(EEMD),and the singular entropy and sample entropy are analyzed,then the fault features are extracted.Due to the high dimensionality of the fault feature matrix,fusion and dimensionality reduction of the extracted singular entropy and sample entropy by kernel principal component analysis(KPCA)are used,so that the feature of inter-shaft bearing fault can be better reflected.Further,the support vector machine(SVM)is used to classify and diagnose the fusion fault features.In this paper,the measured signals are processed and analyzed to obtain vibration characteristics,acoustic emission characteristics,and fusion characteristics.The diagnosis results of the SVM show that the fusion rate of the support vector machine was the highest.At the same time,the validity of the integrated diagnosis method for the inter-shaft bearing faults in the paper was verified.By analyzing the processing signals,the fault characteristics of the weak signals of the inter-shaft bearing were obtained.A high feature diagnosis rate was obtained using SVM.The EEMD-KPCA method proposed in this paper provides a new idea for the fault diagnosis of the inter-shaft bearings of the aero engines. |