Research On Key Techniques Of Fault Diagnosis And Prognosis For Aero Engine Rolling Bearing

The rolling bearing is the key part of the aero-engine bearing system,working in high speed,high temperature and high load of variable conditions,extremely prone to failure.Thus,fault diagnosis and residual life prediction of rolling bearings of aero-engine have important theoretical significance and possess certain practical engineering value for the effective implementation of condition-based maintenance and health management of aero-engines.Based on what is stated previously,this paper developed the off line oil abrasive detection technology and the on-line oil liquid debris detection technology.Besides,we have developed a method for detecting the fault diagnosis and the method of residual life prediction of rolling bearing of aero-engine.The feasibility of these methods has been proved by practical engineering applications and experiments.The main research contents of this paper are as follows:(1)Sums up the common failure of models of aircraft engine rolling bearing.Take an aero-engine as an example,the failure cause and failure mechanism of rolling bearing are analyzed by studying typical failure cases of the rolling bearing of main shaft and accessories.Studies have shown that the main failure mode of bearing of aero-engine is rolling contact fatigue failure.Meanwhile,fatigue spalling failure process tests of real rolling bearing of aero-engine,on which the cause of failure of the bearing was analyzed,have been carried out on the accelerated failure test platform,demonstrating that the fatigue failure process of aero-engine bearing can be simulated by the test procedure.In the end,the symptom information of the fatigue failure of the rolling bearing of aero-engine is analyzed,the defects and shortcomings of the existing methods for bearing fault monitoring of aero-engine bearings are discussed and then the failure characteristics of the rolling bearing of aero engine were sorted out,indicating that the fault diagnosis of the rolling aero-engine rolling bearing can be carried out by the off-line monitoring of the small size abrasives in oil as well as the condition assessment and the residual life prediction of the aero-engine rolling bearing can be realized through the on-line monitoring of large size wear debris.(2)Aiming at the problem of difficultly detecting the wear particles larger than 10 microns in the oil sample analysis method,the intelligent detection technology of the oil moving particle is studied and the Multiple Intelligent Debris Classifying System(MIDCS),which can not only calculate the pollution degree of oil solid particles,but also make an analysis of the particles larger than 10μm,by dividing particles into the class of metal and nonmetal while the metal particles are further identified as cutting abrasive grains,severe sliding wear debris,fatigue wear particles,and the non metallic particles were further divided into bubbles,fiber,other non metallic particles,is developed for the more effective diagnosis of aero-engine bearing fatigue failure.Besides,a new method based on genetic algorithm for multi parameter adaptive adjustment of micro imaging system is proposed.In view of the redundancy problem of the wear debris recognition feature in MIDCS,the rule extraction is carried out by using C4.5 algorithm of Weka software,acquiring the expert knowledge rules for identification of abrasive grains,analvzing the rules,and compared with artificial extraction.The results show that the recognition rules of the wear particles proposed in this paper have a high recognition accuracy.(3)A method for detecting the wear fault of aero engine based on the motion of the oil was studied.Discussion and Analysis on the method of setting the boundary value of the wear particle monitoring.MIDCS was used to carry out the actual engine wear monitoring and verification.Because of the MIDCS to 10 m more than the abnormal wear detection,and the rolling bearing early fatigue spalling will produce more than 10 m abnormal wear particles.Therefore,compared with the traditional spectral analysis,MIDCS has a better advantage in the monitoring of the fatigue failure of the rolling bearing of aero engine.MIDCS wear debris detection is used to predict the failure of the aircraft engine successfully,and avoid the occurrence of dangerous faults.(4)A theoretical analysis of the on-line monitoring sensor for oil and liquid on-line wear debris is carried out,and the principle of the magnetic particle detection and the detection of non ferromagnetic particles is discussed.According to the results of theoretical analysis,design and development of on-line oil monitoring system,the system structure and working mode are discussed.The research proposed the wear debris recognition strategy and the algorithm,and carries on the verification.In the aspect of sensor signal processing,the improved median filtering method is proposed,and the proposed method is verified by experiments.(5)Based on the 35-206P1 bearing of the gearbox of the aero engine accessory,the fatigue accelerated test of 15 groups of air bearing was carried out;On the basis of the experiment,vector regression support(SVR)is introduced to train the effective data,Useful Life Remaining(RUL)on the radial load,speed,and the characteristics of the radial load,speed,and the characteristics of the bearing are obtained.The model can be expanded and perfected continuously through experiment;Finally,the bearing data of the real-time data acquisition is predicted by GM(1,1)grey model,and the residual service life of the bearing is obtained by combining the forecasting model.