Research On Improving Signal-to-Noise Ratio Of Aluminum Alloy FSW Planar Defects Ultrasonic Testing Based On OMP

Aluminum alloys are widely used in industry.Friction stir welding(FSW)technique is usually used for welding.However,improper selection of FSW welding parameters will cause defects.The internal defect interface gap represented by weak bonding is narrow and close to the acoustic impedance of aluminum alloy materials,resulting in insufficient signal-to-noise ratio in ultrasonic testing and restricting defect identification and quantification.In order to solve the above problems,based on the orthogonal matching pursuit(OMP)algorithm and particle swarm optimization(PSO)algorithm,this paper carries out the research on improving the signal-to-noise ratio of ultrasonic detection of closed defects.The main contents include:(1)For 2219-T6 aluminum alloy FSW samples with a thickness of 35 mm,acoustic parameters such as longitudinal and transverse wave velocity,attenuation coefficient and signal-to-noise ratio were measured.Through fatigue test,machining vertical and horizontal defects.On this basis,a 32 element linear array probe with a center frequency of 5 MHz is used to collect array signals in combination with a 55 ° shear wave wedge,which is used for subsequent experimental detection research.(2)For Gaussian signals with different signal-to-noise ratios(-12 d B,-6 d B,0 d B,6 d B),the noise suppression effects before and after OMP combined with PSO algorithm are compared.Analyze the influencing factors of the algorithm,compare the processing effect of Gabor dictionary and discrete cosine transform(DCT)dictionary,and study the influence of OMP iteration times,repetition times and initial value rule on the processing effect.At the same time,several groups of Gaussian signals,sinusoidal signals and signals with unknown prior parameters are processed,and the influence of the parameter range of atoms on the improvement of signal-to-noise ratio is compared.On this basis,the signal-to-noise ratio,mean square error(RMSE)and smoothness(R)are introduced for evaluation.The results show that when OMP and PSO algorithm are combined to process noise signals,optimizing the influencing factors of the algorithm and ultrasonic detection can improve the signal quality and efficiency,the denoising effect is stable,and the efficiency can be increased by more than 11 times.It has good processing results for 0 d B and 6 d B signals,and the RMSE value can be reduced by more than 66%.(3)Change the probe position to detect the primary and secondary waves of defects.Collect the full matrix data with 619 and 1886 signal points.After applying noise,6 d B and 0d B array signals are obtained.Carry out OMP and PSO processing.Evaluate by using relevant evaluation indicators and the arrival time of defect waves.Simulation and experimental results show that this method can effectively reduce the signal RMSE and R value,and the arrival time and quantitative evaluation of defect wave are more accurate.(4)The double Gauss dictionary is introduced to replace the Gabor dictionary,and the processing results of the array signals obtained from the primary and secondary wave detection are compared.On this basis,expand the range of signal-to-noise ratio(-6 d B,-3 d B,0 d B,3 d B,6 d B)to further judge the applicability of the double Gaussian dictionary.The simulation and experimental results show that the RMSE is reduced by 49% and 68% on average compared with the original signal when the array signals detected by primary and secondary waves are processed by double Gauss dictionary.Even under the condition of-6 d B,the average RMSE of the signal after 1000 times of processing by the double Gauss dictionary decreases by 55%,which is 65% lower than that of the Gabor dictionary.The double Gauss dictionary can suppress the noise and retain the effective waveform components to a greater extent.