The Application Of Wavelet Based On Time-frequency Analysis Method In Ultrasonic Signal Processing

It is an important procedure to detect the inner quality of steels in the process of manufacturing. It is also a long-term research direction in the methods of testing signal analysis. In this thesis, the characteristic of ultrasonic transmitting in steel is analyzed. The ultrasonic testing system is modeled and it is built that the mathematical model of ultrasonic emitting signals, ultrasonic transmission and the ultrasonic echo signals. The time-domain and frequency-domain graph of the emission signal and the echo signal is simulated in MATLAB. The ultrasonic launch signal transmits in the internal dissemination of steel, and carries the information of the steel internal structure and noise, and it becomes the ultrasonic echo signal in the receiver. The echo signal has non-stationary characteristics, and we adopt the time-frequency analysis method to analyze the signal.The wavelet analysis is a new time-frequency analysis method. The DB wavelet generating function is selected to de-noise the echo signals using MATLAB and eliminates some high- frequency noise. The echo signals after de-noising is analyzed in wavelet analysis toolbox by the multi-resolution analysis method. Then the signal is decomposed into low- frequency part and high- frequency part. According to the results of wavelet analysis and transmission characteristic of ultrasonic, we determine the internal defects in steels as well as the location of the defects.First, we analyze the principle and characteristics of ultrasonic nondestructive testing, and discuss the basic physics and characteristics of ultrasonic, and design the model for detecting steel. According to the characteristic of ultrasonic transmitting in steel material and comparing to the disadvantage of traditional fourier analysis, we select the wavelet time-frequency analysis method to analyze the echo signals.Secondly, the basic theory of wavelet time-frequency analysis is brief introduced, including the continuous wavelet transform, the discrete wavelet transform and the wavelet fast algorithm. The wavelet generating function is detailed discussed which is used in analyzing the ultrasonic echo signals, the characteristic of wavelet generating function is studied in detail. By comparing the characteristics of different wavelet generating function, we select the DB wavelet as the wavelet generating function for analyzing the ultrasonic echo signal.Thirdly, according to the characteristic of the ultrasonic transmission, we analyze the traits of the ultrasonic launching process, the ultrasonic transmission and the ultrasonic receiving process, and build mathematical model for each process. We analyze the time-domain and frequency-domain traits of the ultrasonic launching signal and the ultrasonic receiving signals. Based on the mathematics model of the launch signals and the echo signals, their waveform is drawn in MATLAB by program.Then, according to the characteristics of the ultrasonic signals, we de-noise the echo signals using the wavelet analysis method. According to the threshold selection criteria, we use wavelet analysis toolbox to draw figure in different threshold. Based on the result of de-noising in different threshold, and we select the heuristic threshold to de-noise the signals. Comparing to the statistical features of de-noising, the DB6 wavelet is selected as the wavelet generating function to analyze the signals by de-noising and multi-resolution analysis.Finally, we analyze the ultrasonic signal using the tool of one-dimensional wavelet under MATLAB environment. The signal is decomposed into 5 layers and gain high-frequency information of each layer and the low-frequency information of the fifth layer. According to the first layer's high-frequency information, we can know the time that the singularity of the echo signal is appearing and calculate the position of the internal structural defects by the ultrasonic transmission speed in steel.