Time-frequency Analysis Of Welding Crack AE Signal Based On S Transform

Welding process structure cracks and status changes arising stress wave is a kind of instantaneous, sudden, not reproducible acoustic emission(AE) signals. Because of the friction of specimen and clamp and arc shock interference in testing, AE signals of actually collected showing a multi-modal, characteristic faint of features, it is a non-stationary signals. Through the research on a time-frequency analysis method of have multi-resolution characteristics in time-frequency domain, applied to welding crack acoustic emission signal analysis and processing, extraction of time-frequency feature for welding crack acoustic emission signals, it provides a method for the effective realization of welding crack acoustic emission signal feature detection, has project practical and theoretical research value.S transform is a time-frequency analysis method of based on STFT and WT,it possess adaptive multi-resolution, time-frequency characteristics of a good aggregation. This paper Based on the analysis of S transform theory, the simulation analysis and improvement for S transform characteristics, upgrade its adaptive of time-frequency feature extraction for welding process acoustic emission signal, and applied it to the welding crack acoustic emission signal time-frequency feature extraction. The main research content was as follows:1. Based on analysis of S transform theory and algorithm, using matlab language achieve programming of the S transform algorithm, through selecting the appropriate simulation signal model simulation comparative analysis for S transform characteristics, and applied to time-frequency analysis for welding crack acoustic emission signal of actual measured, verify the S transform adaptive, multi-resolution,time-frequency concentration, anti-noise of application of welding crack acoustic emission signal.2. Based on the standard on S transform, through lead into two can be adjusted parameters, it achieve the function window size adjustment for standard S transform.Analysis of improved S transform discrete form and achieve algorithm. Study on the influence of different parameters to improve the S transform window function.Comparative simulation analysis improved S transform, determined improved S transform optimum time-frequency resolution parameters, and applied it to the measured AE signals, verify improved S transform has better adaptability and time-frequency aggregation in welding crack acoustic emission signal time-frequencyanalysis.3. Based on S transform frequency analysis method,formulated time-frequency analysis method measure of based on S transform welding crack acoustic emission signal, and applied it to stretch of welded structures of acoustic emission signals testing experiment, calculation and analysis of the various stages of stretch of welded structures acoustic emission signal waveform and spectral characteristics and signal energy distribution. on this basis, welded structure stretch AE signal all phases time-frequency feature as the reference, applied to the welding process crack acoustic emission signals testing experiment, through time-frequency analysis for the period of welding AE signal, achieve the effective detection for the various stages of welded structures crack acoustic emission signals, analysis of welding crack acoustic emission signal at all stages of the evolution of the spectral characteristics of the waveform and time-frequency characteristics.