On-line Monitoring And Particle Filter Based Fatigue Crack Growth Prediction

The prognostics and health management(PHM)system can monitor the health status of the aircraft on-line to achieve life prediction of key components,and ultimately direct the task decision and logistical maintenance.It shows great application potential and economic benefit in shortening the aircraft downtime,prolonging the service life and enhancing the maintenance.Fatigue crack is one of the common failure types of key aircraft components,which is the focus of PHM systems.However,the fatigue crack growth process is a non-linear non-Gaussian dynamic stochastic process,which involves a variety of uncertainties.Actual crack initiation and growth is usually deviated from the result obtained with fracture mechanics analysis.In order to effectively monitor and predict fatigue crack initiation and growth overcoming the influence of uncertainties and promote the development of the PHM technology,in this dissertation,the active Lamb wave based technology is integrated for crack monitoring,the state space model which dscribes crack growth is established and improved,and finally the particle filter and its improved ones are adopted to peroform crack growth life prediction in various aeronautical structures.The main works are as follows:1.The state space model describing fatigue crack growth includes an evolution equation which is established by Paris law and an observation equation which is constructed by the active Lamb wave online monitoring technique.Based on this state space model,a Lamb wave-particle filter based on-line prediction method is proposed,which overcomes the influence of uncertainties.Finally,the effectiveness of this method is validated by on-line prediction of hole-edge crack growth.2.The method of establishing the observation equation against fatigue crack growth based on active Lamb wave is put forward.The four aeronautical structures: surface-opening,hole edge,attach lug and horizontal stabilizer shaft have been applied to study the damage indices of the Lamb wave response signal with the amplitude,phase and signal correlation of the crack propagation.Dmage index results from the same type of specimens are statistically analyzed,at the mean time,variance of the damage index is used as an indicator to compare the effective of various observation methods of damage indices.An observation equation describing the quantification relationship between the damage index and the crack size is constructed,and the respond signal characters,the optimization guidance of sensor arrangement and experimental design are summarized.3.The method for establishing the evolution equation of fatigue crack growth based on fracture mechanics is proposed.Applicability of Paris law,Walker model and Forman model is discussed under fatigue tests of four kinds of aeronautical structures.The calculation of stress intensity factor range and the uncertainty of material parameters are studied.Combining with the uncertainty of fatigue crack growth,two kinds of state noise distribution are considered,which can describe the stochastic dynamic process of fatigue crack growth with higher robustness.4.From the perspective of optimizing the importance density function,a Lamb wave-auxiliary particle filter based on-line prediction method is proposed,which can mitigate the lack of particle diversity.The proposed method has been verified experimentally with the fatigue test of hole-edge crack specimens.The result shows the advancement of prediction accuracy.5.From the perspective of optimizing the resampling technology,a Lamb wave-deterministic resampling particle filter based on-line prediction method is proposed,which can overcome the lack of particle diversity.This proposed method has been verified experimentally via fatigue test of attachment lugs.Compared with the standard particle filter and auxiliary particle filter,the Lamb wavedeterministic resampling particle filter possesses higher prediction accuracy.