Research On Signal Inversion Of X90 Steel Corrosion Acoustic Emission Source Based On Blind Deconvolution

The identification of X90 steel corrosion source signal characteristics based on acoustic emission technology is quite complicated.Due to the randomness and uncertainty of the corrosion process,the complex corrosion environment has couplings between multiple corrosion types,and the corrosion acoustic emission signal attenuation and distortion will occur in the process of propagation.Therefore,the acoustic emission signal analysis method based on the blind deconvolution algorithm is particularly important for the inversion and recovery of the corrosion source signal and the exploration of source signal characteristics.In this paper,the method of combining numerical simulation and experimental research is used to discuss the inversion of the signal characteristics of the X90 steel corrosion acoustic emission source.The main research work and results are as follows:The COMSOL software is used to carry out numerical simulation research on X90 steel corrosion acoustic emission signal.In the X90 steel two-dimensional,three-dimensional and three-dimensional welded sheet model,a modulated sinusoidal signal with a center frequency of 150 k Hz is used to replace the corrosion acoustic emission signal for wave field law.Research has found that the waveform will be refracted and reflected at the boundary of the sheet during the propagation process to generate an echo signal.When the signal passes through the weld,the waveform energy is significantly reduced and the wave mode also changes,the farther away from the sound source,the more complex the signal type is.The signal collected at each observation point is subjected to wavelet transform and blind deconvolution algorithm for blind recovery.The research has found that the recovered signal is close to the source signal,especially the signal characteristics are obviously recovered,it again verifies the feasibility of the blind deconvolution algorithm.The X90 steel corrosion acoustic emission experiment device was constructed.The experiment chose the dilute sulfuric acid solution with p H=4.0 as the corrosion liquid,and set up the corrosion experiment and the corrosion experiment with weld defects respectively to determine the corrosion acoustic emission signal of X90 steel and study the process attenuation characteristics after the weld.Experiments have found that the acoustic emission signals generated by corrosion of X90 steel are mainly low-frequency signals,which are mainly distributed in the frequency domain at 100 to 200 k Hz.Wavelet denoising is performed on the signal.The db4 wavelet is used for 5-level wavelet decomposition and reconstruction,the peak value in the frequency domain is around 170 k Hz after removing the interference signal.When the signal propagates through multiple welds,the waveforms measured by the sensor become more complicated.The mode of the signal wave is transformed mainly due to the superposition or cancellation of various acoustic emission waves and echoes when propagating through the welds.The signal propagates encountering obstacles in the process causes the energy to gradually decrease,and high-frequency signals attenuate faster than low-frequency signals.Blind deconvolution restoration based on wavelet transform is performed on the signals of each observation point to better recovery the original signal characteristics,the low-frequency components of the signal are mainly concentrated in 130 to 170 k Hz,the higher frequency part of the signal frequency peak is 200 k Hz and 250 k Hz.The attenuated high-frequency signal is recovered to a certain extent,which makes the measured signal closer to the real corrosion source signal of X90 steel.