Research On Aircraft Engine Wear Limits Based On Data Mining

As the aircraft engine prone to a variety of wear and tear,and the main residual wear and tear in the aircraft lubrication and hydraulic systems,the use of oil analysis technology to monitor the engine wear and fault diagnosis has very important significance.The main contents are as follows:(1)Aiming at the case that the aero-engine oil data does not obey the normal distribution,the limit value method based on the probability density function estimation is studied,which includes maximum entropy method,Parzen window method and K-nearest neighbor method。 The validity and adaptability of the method are verified by simulation data and real data.(2)A fault diagnosis method for oil engine data fusion threshold was proposed.Through the feature fusion of the oil spectral characteristic elements,a health index which can measure the wear state of the aeroengine is obtained,and the threshold value is formulated to realize the purpose of fault diagnosis of the oil data.The limits of the health indicators for the evaluation of aircraft engine wear status is of great significance,respectively,for different elements of the threshold value comparison,more simple and reliable.(3)A fault diagnosis method for automatic extraction of knowledge rules of aero-engine oil data is proposed.Then,the knowledge rules contained in the original data are extracted by weka platform,and the fault diagnosis can be realized by the extracted rules.Finally,a new method is proposed to extract the original data from the original data.Finally,the extraction results of the real aero-engine wear spectrum data show that the knowledge rule extraction based on weka platform has a high recognition rate,which can identify the fault data well and make the fault diagnosis more intelligent and automatic.(4)The Engine Oil Monitoring Expert System(EOMES)was developed using Microsoft Visual C ++ 6.0 and Microsoft Access 2003 database to realize the engine wear monitoring and intelligent diagnosis based on the threshold value and rule extraction.