Research On Feature Fusion Method Of Aeroengine Rotor Fault Based On Improved VMD

As the "heart" of aircraft,especially rotor system.In the extreme working environment,such as high temperature and high pressure,the rotor system is prone to failure caused by complex vibration.Therefore,it is of great significance to study the early fault diagnosis of aeroengine rotor system to ensure the safety of aviation flight and reduce the occurrence of aviation accidents.Because the early aeroengine fault signal is very weak,often submerged in strong noise,which makes it difficult to extract the fault information.To solve this problem,this paper mainly carries out the research:(1)In this paper,several common types of rotor fault are introduced in detail and analyzed.The limitation of VMD are analyzed.An improved method of VMD is proposed and its superiority compared with original VMD is verified(2)The method of constructing high-dimensional mixed domain fault feature set from time-domain and frequency-domain feature set,sample entropy feature set and ARIMA model is to reflect fault comprehensively in many aspects.The improved variational mode decomposition is combined with the high-dimensional mixed domain feature set,and the high-dimensional fault feature set is reconstructed with the popular learning method of ASL-Isomap,so as to realize the fault feature fusion reduction.An improved feature extraction method of ASL-Isomap is proposed(3)Support vector machine(SVM)is used to identify and classify the data processing results,so as to achieve the fault diagnosis of aeroengine rotor system.The grid search is used to the parameters of SVM,and the fault model of improved VMD mixed feature set ASL-Isomap SVM is established.The experimental results that the fault rate of this model is up to 98.75%,which meets the requirements of diagnosis,thus verifying the of the diagnosis model.Compared with the model of ASL-Isomap SVM,which is the feature set of EWT mixed domain,and the fault diagnosis model of data direct processing,its superiority is verified.(4)The user interface of aeroengine rotor fault system is developed based on guide of MATLAB platform,which makes the whole diagnosis process more convenient and intuitive.