Study On The Parameters Estimation Of Lamb Wave Time Domain Signals Under Narrow Band Excitation

Lamb wave is one kind of elastic wave propagating in plate-like or layered structures.With the merits of long propagation distance without much energy loss and sensitivity to defects,Lamb wave based inspection techniques are widely considered as effective means to realize repaid defect detection.However,there are still many problems in the Lamb wave applications.One is that the dispersion characteristics and multiple-modes of Lamb wave making signal processing and data analysis complicate;the other is that the operation conditions may lead to dectection failure.In this dissertation,the wave packet models and signal parameters estimation method of Lamb wave time domain signal under narrow-band excitation are studied.The relation among the wave packet characteristics,the theoretical dispersion curve(s)and the propagation distance is revealed in analytical forms.The surface defects localization method in(quasi)isotropic plates is studied by utilizing the relationship.The main research work includes the following aspects:Firstly,based on the general non-linear dispersion theory of Lamb waves,the wave packet models of Lamb wave signals in time domain are derived.Two kinds of the models are considered,the initial excitation propagating in single mode,and bi-mode caused by the mode conversion.In the procedure of model derivation,the analytical relations among model parameters,propagation distance and dispersion characteristics are clearly given,and the two models are verified by the time-frequency results of numerical simulation or(and)experimental signals.Secondly,based on the expectation maximization(EM)algorithm,a new Lamb wave signal processing tool is proposed for the signal parameters estimation.Through the synthetic signal,the convergence,anti-noise ability and accuracy of the parameter estimation results of the tool are verified.At the same time,the relation among the parameters of the single-mode propagation model,the propagation distance and dispersion characteristics is proved by using the experimental signals or(and)numerical simulation signals,and a method for extracting the dispersion characteristics from the signals is proposed by utilizing the relationship.In addition,the bi-mode propagation model is quantitatively verified,and the ellipse location method is improved for the mode conversion signal.Based on the model verification,a multi defects localization method based on the modal conversion signal parameters is proposed.Moreover,the wave packet model and the localization method are further demonstrated by the experimental method.Finally,based on the verification conclusion and using the sensor system with bisurface piezoelectric elements,a scheme of baseline-free defects localization is proposed.This scheme does not need to acquire the material parameters and geometric parameters of the tested materials in advance.The feasibility of the scheme is discussed by numerical simulation,and the experimental verification of the scheme is carried out in a quasiisotropic glass fiber composite plate with unknown parameters.