Research On Sparse Sampling Method Of Coded Ultrasonic Signal Based On FRI

The ultrasonic signal excited by a single pulse has the characteristics of finite rate of innovation,which can be sparsely sampled,and then the key information parameters of the original signal can be accurately recovered from the sparsely sampled data.This theory can effectively reduce the sampling rate of pulsed ultrasonic signals and reduce the amount of collected data,which has a wide range of applications in the field of ultrasonic testing.Coded excitation ultrasonic can effectively increase the average sound power of detection signal without increasing the transmission voltage,then improve the detection range and signal-to-noise ratio.It is also widely used in the field of ultrasonic detection.However,the structure of coded ultrasonic signal waveform is complex and does not have the characteristics of finite rate of innovation.Therefore,it is impossible to directly apply the sampling theory based on finite rate of innovation to sparsely sample it,which restricts the expansion of its application field.In view of this situation,based on the study of the characteristics of coded ultrasonic signal and the theory of sparse sampling with finite rate of innovation,this paper proposes a framework of sparse sampling with finite rate of innovation which is suitable for coded ultrasonic signal,deduces its theoretical system,and implements this framework through the self-designed hardware circuit.The thesis first analyzes the existing sparse sampling theory of finite rate of innovation,and summarizes the current sampling theory of pulsed ultrasonic signal based on finite rate of innovation.The mathematical model of coded ultrasonic signal is established,and the time-frequency characteristics,the excitation method,and the characteristics of transducer response are analyzed through theoretical research and simulation.Combined with the requirements of echo signal parameters in the field of defect detection,a pulse stream conversion method of coded ultrasonic echo signals based on higher-order moment extraction is proposed,which converts the coded ultrasonic echo signals into finite rate of innovation signals and retains the time delay and amplitude parameters.The time delay and amplitude parameters of coded ultrasonic echo signal are accurately reconstructed from the sparsely sampled data.The hardware implementation circuit under the theoretical framework is designed and produced,the input and output characteristics of circuit are tested and analyzed.With the pipeline defect samples as the inspection object,different types and sizes of defects were detected,which can effectively detect the defects.The experimental results show that this method can effectively reduce the sampling rate of coded ultrasonic signal and reduce the amount of data sampled,which provides theoretical basis and hardware support for coded ultrasonic signal sparse sampling based on finite rate of innovation in broadband,multi-sensor array,long-time detection,and occasions requiring high data fast processing.The main research contents are as follows:(1)The principle of sparse sampling with finite rate of innovation is expounded,and the progress of sparse sampling with finite rate of innovation for pulsed ultrasonic signals is summarized.Analyze the application fields and technical characteristics of coded ultrasonic detection,point out the demand field of sparse sampling with coded ultrasonic signals,establish the overall flow of sparse sampling and reconstruction with coded ultrasonic signals.(2)Compare and analyze the common coding forms and their characteristics,propose a binary frequency encoding method,establish its mathematical model,analyze its time-frequency characteristics,transducer response characteristics,and echo signal parameters,It lays the foundation for the subsequent development of sparse sampling with finite rate of innovation based on this coding form.(3)A signal conversion method based on higher-order moments is proposed,the construction framework of the pulsed stream of coded ultrasonic signal is deduced,the basis for selecting the conversion order of pulsed stream signal is given,and the parameter retention characteristics of converted signal are analyzed as the innovation rate sampling signal.The construction of sampling kernel and the sparse sampling method were studied,the sparse sampling data of coded ultrasonic echo signals were obtained,and the time delay and amplitude parameters of echo signal were reconstructed.The hardware circuit under the framework is designed and produced,including binary frequency code generation and excitation circuit,coded ultrasonic signal echo receiving circuit,second-order moment extraction circuit,high-order moment conversion circuit and gain adjustment circuit.Combined with the sampling kernel circuit developed by the research group,the hardware sparse sampling and parameter reconstruction of coded ultrasonic signal are realized.(4)In order to verify the effectiveness of the proposed method and the designed circuit,analyze its performance characteristics,build a pipeline defect coded ultrasonic testing test platform,using 4-bit binary frequency coding form,where "0" represents 3.2MHz low frequency symbol,"1" represents 6.4MHz high-frequency symbol,detects four kinds of defects,and obtains the sparse sampling data of coded ultrasonic echo signal.Using the annihilating filter method to estimate the key information parameter of echo signal,the maximum reconstruction error of time delay parameter is 0.14%,and the maximum reconstruction error of amplitude parameter is 9.32%.According to the conventional Nyquist sampling theory,the sampling rate should be no less than 12.8MHz,and the sparse sampling rate in this experiment is 1.2MHz,the sampling rate is only about 10% of conventional method,the amount of sampled data is reduced by more than 80%,and the common defects in the pipeline can be effectively detected.