Research Of Aero-engine Fault Diagnosis Based On Filtering Methods

Aero-engine gas path fault is one of the most common faults of the engine, and itoften occurs in the high temperature and pressure parts. The fault will generally lead tothe decrease of engine performance parameters (flow ability and efficiency), and willaffect the safety of the engine performance and reliable performance in some serioussituation. Aero-engine gas path fault diagnosis technology can examine the engine gaspath components in real-time monitoring and testing, and it can also examine the faulttimely and accurately and handle it according to its type and order of severity. Thetechnology can effectively increase the economical efficiency and reliability of engine,and reduce the cost of engine operation greatly. So the research of this technology hasthe very high practical value. In this paper, the mutation type and gradual type fault arediagnosed based on the nonlinear model of aero-engine, using the method of kalmanfilter and particle filter.First, the simplified nonlinear gas path fault diagnosis model is established on thebasis of the aircraft engine component level model based on Matlab/Simulink. And itsaccuracy is verified by means of comparing the results of the simulation model withcomponent level model. The model is built using the fitting method, and the enginehealth parameters are introduced in the model. In the process of modeling the idea of thenormalization is adopted, which can greatly reduce the amount of calculation.Starting from the mechanism, the mechanism of kalman filter is research throughthe introduction of average filter and low-pass filter, and the factors which can affect theeffect of kalman filter are analyzed. Then simple examples are introduced to illustrate it.On the basis of the standard kalman filter, the extended kalman filter (EKF) andunscented kalman filter (UKF) are introduced. The application range and the influencefactors of the two methods are discussed. Based on the established nonlinear model, twokinds of fault (the mutation type and gradual type) are diagnosed using EKF and UKFrespectively. The simulation results show that for weak nonlinear system, UKF has notobvious advantage over EKF, and UKF is suitable for the type of gas path faultdiagnosis of mutations.Finally, the particle filter method is introduced into the field of aero-engine gas path fault diagnosis, and the applicability of the method in the field is discussed. Firstly,the basic ideas of particle filter (PF) and influencing factors are put forward, and on thisbasis the modified method (EKPF) is discussed. Then the mutation type and gradualtype fault are diagnosed using two methods respectively. The results show that twokinds of method can meet the basic requirements of diagnosis, but the PF is not suitablefor fault diagnosis of mutations. EKPF gets better diagnosis results than PF because itcan overcome particle degradation problem which exits in the particle filter method.