Study On Non-smooth Signal Processing And Fuzzy Recognition Of Deep Drawing Crack

As the main failure form of sheet metal forming parts,the deep drawing cracks often appear in the dangerous area of the forming parts,especially the early cracks.According to the above mentioned problems,this study uses acoustic emission signal detection technology of metal forming deep drawing parts of AE signal detection,and application based on wavelet threshold-EMD synthesis method for crack AE signal decomposition,denoising and recombination,finally using fuzzy clustering method based on fuzzy equivalence relations of all kinds of crack fuzzy recognition.The main research contents are as follows:1)model based on the deep drawing of rectangular box,the drawing of sheet metal forming stress and strain state and process of drawing crack causes are analyzed,and then combined with the simulation of sheet metal drawing bucket results drawing such dangerous areas prone to drawing crack on the detection system and work;the principle of non-destructive testing method of acoustic emission are introduced,in combination with the crack of sheet metal in the process of drawing with the characteristics of acoustic emission signals and acoustic emission signals,to determine the deep drawing of sheet metal parts for deep drawing for drawing crack,and the acoustic emission detection system to monitor the whole process of drawing for containing crack the signal of acoustic emission signals.2)combined with the above theoretical analysis,the acquisition of AE signal of the deep drawing of sheet metal in sheet metal drawing bucket as the research object,the collection and preservation of a large number of no crack,crack,crack extension early AE signal three crack state,and the signals and metal parts bucket corresponding annotation,in order to facilitate the post data processing.3)on the noise reduction method of acoustic emission signal was studied,the comparison and analysis of denoise,EMD wavelet threshold filtering of acoustic emission reducing effect is good two kinds of noise reduction methods,combined with the advantages and disadvantages of proposed wavelet threshold denoisingmethod-EMD,noise reduction processing of sheet metal drawing sound this combined emission signal denoising method in denoising before the first the vanishing moment is 5 Daubechies wavelet signal of AE three layer wavelet decomposition,and then selected according to the signal frequency band,the selected frequency band signal with EMD noise reduction method for noise reduction,but no band signals were selected by using wavelet threshold the noise,and then the two part after the signal denoising for signal reconstruction to obtain pure acoustic emission signals.4)analysis of the non-stationary signal characteristics of sheet metal forming cracks,extract the parameters corresponding to the signal from the pure signal denoising after;the fuzzy clustering algorithms are introduced,fuzzy clustering method based on fuzzy equivalence relations of four kinds of signal recognition,selected independent of the amplitude,RMS voltage(RMS),the average value of signal level(ASL)and energy four parameters as recognition parameters,establish the data matrix,use MATLAB software to extract parameters of sheet metal drawing bucket were simulated and analyzed,the parameters of crack,crack and crack propagation of three kinds of early crack state fuzzy recognition.The research results show that the emission acquisition system of metal parts bucket using acoustic acoustic emission signal,denoising and reconstruction on the drawing signal using wavelet threshold denoising method of-EMD,and then extract the characteristic signal as the parameters of fuzzy clustering in the realization of sheet metal drawing bucket crack status(especially early crack)fuzzy recognition,and has high accuracy.