Research On The Key Technologies Of Magnetostrictive Longitudinal Mode Guided Wave Signal Processing

In order to meet the accuracy and precision demands of actual magnetostriction guided wave testing, taking magnetostrictive longitudinal mode guide wave as research object, some key technologies of guided wave signal processing such as joint time-frequency analysis, signal denoise and locating reference extraction were studied in this paper, The detail work was presented as follows:(1) The importance of signal processing in magnetostrictive guided wave testing was discussed, the bottlenecks of current signal processing technology were analyzed, on the basis of those,the key technologies of magnetostrictive longitudinal mode guided wave were summarized and the content and direction of this research were determined.(2) On the basis of magnetostrictive guided wave testing theory and the longitudinal mode guide wave detection method, characteristics of guided wave and the technical difficulties of signal processing were mainly discussed, then a magnetostrictive guided wave detection system was built, by this system rockbolts in the air and concrete were detected and the testing signals were analyzed.(3)In order to analyze the dispersion and multi-modal characteristics of magnetostrictive longitudinal mode guide wave more effectively, the application of short time fourier transform and Wigner-Vill distribution in guided wave signal processing was discussed, the time-frequency analysis ability of those two methods was not only discussed in theory but also in experiment. To solve the problems in the signal processing using those two methods,some optimizational window was applied in short time fourier transform, also pseudo Wigner-Vill distribution was introduced into the signal processing of magnetostrictive longitudinal mode guide wave signal.(4) The magnetostictive guided wave signals processed by traditional filters are not satisfactory, in view of those problems, a non-linear adaptive filter algorithm is proposed, wavelet decomposition, neural networks and particle swarm optimization algorithm are introduced into the traditional adaptive filter. A new non-linear adaptive filter is realized by the presented method. Actual magnetostictive guided wave signals were processed by this algorithm, results show this algorithm can improve the signal-to-noise ratio more effectively compared to the traditional approach, and it avoids the problem of the distortion of filtered signal.(5) In order to locate the magnetostrictive longitudinal mode guided wave detection signals more accurately, two locating reference extraction algorithms were proposed: reference extraction algorithms based on wavelet singularity detection,eference extraction algorithm based on Teager envelope algorithm. Those two algorithms were applied in the signal processing of magnetostrictive longitudinal mode guide wave signal, results showed the positioning accuracy of detecting signals were improved by using those two algorithms and those two algorithms have the characteristics of simple calculation process, low requirement on the SNR.