Research On The Gas Path Performance Estimation And Fault Diagnosis Of An Aero-engine

Aero-engine intellectualization is an important trend of development in the future.The aero-engine health management system is the most critical part,and the engine performance testing and fault diagnosis is the core content of the engine health management system.Among the aero-engine fault,the gas path component fault is the highest one,and it is also the main maintenance cost.Therefore,it is very necessary to perform the performance estimation and fault diagnosis of the gas path component fault.Firstly,the essay established aero-engine linear models including the health parameters,and the simulation of the engine model and includes health parameters of engine linear model is proved to maintain the stability of the model,which can not only significantly changes in the controlled quantity,and can reflect the degradation of aero-engine gas path performance.Then,the problem of aero-engine health parameters tracking is discussed.The use of health promotion to Kalman filtering parameters of state variables,suppress the noise interference and tracking parameters of good health,under different flight conditions and failure mode.In order to analyse the trend of estimation problem of aero-engine gas path health parameters,the principle of support vector machine regression analysis and least square regression are used to carry out regression analysis of gas path Kalman tracking parameters.The regression results of the two methods are compared by simulation,it proves that the support vector machine regression analysis method is suitable for the modeling of performance baselines.Finally,a fault diagnosis method based on classical BP neural network fault diagnosis and support vector machine is proposed to diagnose the degradation faults of aero-engine components,and the cross validation method is taken into establish the training model at the same time.By comparing the simulation results of two situations,the results show that the support vector machine can well meet the diagnostic requirements both in time and accuracy.