Multilayer structured media are widely used in many fields because of their excellent characteristics,such as large thickness metal-nonmetal multilayer structured workpieces,which are typically represented by solid rocket motors.Due to the large thickness of the workpiece high frequency ultrasound is often difficult to penetrate,the use of low and medium frequency ultrasound pulse reflection method for the inspection of such large thickness of multilayer structure workpiece.However,due to the non-uniform thickness of each layer of the multilayer structure workpiece,coupled with the large difference in acoustic impedance of each layer of the material,the obtained ultrasonic detection signal has the phenomenon of mixing of different components,which will affect the extraction of useful information of the target signal,and because of the large bandwidth of the low-frequency ultrasonic pulse,the difference in frequency shift of the ultrasonic signal in the multilayer medium due to attenuation is not obvious,and it is difficult to be decoupled by the way of time-frequency domain filtering.decoupling by time-frequency domain filtering.Therefore,it is important to realize the decoupling separation of low and medium frequency ultrasonic mixed signals to improve the detection accuracy of such multilayer structure workpieces.In view of the above-mentioned problems,this thesis adopts the Radon domain adaptive matching algorithm to realize the study of the decoupling method for ultrasonic mixed signals.The main research contents of this thesis are:(1)A multi-physics field multilayer structure model was constructed to study the propagation characteristics of medium and low frequency ultrasonic waves in multilayer structure medium,and the indistinguishability of the reflected echo signal in the time-frequency domain and the differential distribution in the space-time domain were analyzed;(2)Based on the idea that the spatial distribution of direct interference waves is highly correlated and the correlation of effective echoes is relatively weak,an adaptive matching algorithm model is constructed to process the signal from a spatial and temporal perspective.According to the simulation results,the direct adaptive matching processing of the signal can play a role in suppressing the direct interference wave to a certain extent,but due to the non-uniform degree of mixing throughout the matching object,matching is difficult,resulting in a poor separation effect;(3)To address the shortcomings of direct adaptive matching of signals,a high-resolution Radon domain adaptive matching method is constructed by combining the linear projection focusing property of Radon transform to narrow the matching area.The applicability and reliability of the adaptive matching signal separation method combined with the high-resolution Radon transform are verified by processing the signal data from the mathematical model,the physical model simulation,and the measured data. |