The Study Of Aero Engine Rotor System's Vibration Fault Diagnosis Based On Clustering Of Fast Search And Find Density Peak

With the continuous progress of science and technology,the construction of aero engines has become more and more complex.The consequent engine failure probability is also increasing.Therefore,the requirement for improving aero engine stability and reliability are becoming increasingly high.As an important part of the aero engine,the rotor system is one of the most vulnerable parts in the aero engine.Therefore,the research on the fault diagnosis method of the aero engine rotor system not only has important academic research value,but also has practical significance to the development of aviation industry and travel safety.Firstly,the method of signal denoising in fault diagnosis has been studied in this paper,and the specific implementation process of lifting wavelet threshold denoising method is introduced.The construction of the lifting wavelet scheme and the number of lifting wavelet decomposition level which affect the denoising effect are discussed and analyzed.A specific denoising scheme is obtained and a comparative analysis was performed from the perspective of the denoising effect on fault diagnosis feature extraction.It was verified that signal denoising greatly improves the accuracy of fault diagnosis.In this paper,the Clustering by fast search and find of density peaks(CFSFDP)algorithm introduced into the fault classification of fault diagnosis.Aiming at some defects of CFSFDP algorithm still exists,an automatic search and find density peaks clustering algorithm based on Mahalanobis distance is proposed.The Mahalanobis distance is introduced as the distance measure between data points,which improves the clustering caused the quantities and dimensions of clustering when truncation distance selection difficult problems.A comprehensive reference value ? is introduced to help determine the density peak.The peak value of density is obtained,and the clustering center is automatically achieved.The problem of manually determining the peak value of the cluster and manually determining the cluster center are solved,so that the subsequent clustering operation can be performed and the purpose of automatic clustering can be achieved.Through the comparison of the improved CFSFDP algorithm and the original algorithm,it is found that the improved CFSFDP algorithm greatly improves the clustering accuracy;in comparison with other clustering methods,found that the improved CFSFDP algorithm has certain advantages in accuracy.Finally,the aeroengine rotor system simulation experiment device is introduced.The rotor system fault simulation experiment is carried out.The collected aeroengine rotor system's running status signal is analyzed by examples to verify the effectiveness of the algorithm and other methods are added for comparison.It shows that the clustering algorithm based on the Mahalanobis distance found density clustering algorithm has a good effect in the application of fault diagnosis.The accuracy of fault diagnosis is higher than the general high diagnostic method.