AE Signal Feature Analysis And Experimental Research Of Early Crack In Deep Drawing Parts Forming

Crack defects in the process of deep drawing is one of the deep drawing parts main failure form, the early crack which contained in crack defects is difficult to detect by traditional method. Aiming at this problem, this paper gather the deep drawing AE signal by acoustic emission testing technology, and accomplished AE signal processing and analysis by using empirical mode decomposition(EMD) and wavelet analysis method.Finally got the acoustic emission signal characteristic of deep drawing stamping early crack defects. The main research contents and results are as follows:1) Analyzed the reasons of crack defects in deep drawing stamping process,introduced the basic principle of acoustic emission technology. Determined the method of obtain crack defects AE signal characteristics by analysis and processing collected AE signal, which on the basis of the characteristic of acoustic emission signal appear with extension crack in metal forming process, and combining with the characteristics of AE signals;2) Studied the AE signal de-noising method, advantages and disadvantages between the threshold filtering of wavelet de-noising and EMD de-noising in AE signal de-noising;wavelet threshold-EMD noise reducing method was proposed and be applied to drawing part forming AE Signal de-noising, which choose the Daubechies wavelet for three layers of wavelet decomposition of AE signal and cancel lost moment is 5; IMFs signal-to-noise boundary was determined on the basis of the continuous square error criterion determine;3) The wavelet decomposition coefficients and wavelet feature can spectral coefficient were extracted from AE signals after noise reduction and decomposition as the feature parameter by wavelet transform analysis method, which based on wavelet multi resolution analysis(MRA);Four kinds AE signal characteristic of deep drawing forming were summarized by forming AE signal wavelet feature spectral coefficients in each scale distribution, wavelet decomposition coefficients statistical distribution and the analysis of the frequency spectrum of reconstruction signal, especially AE signals characteristics of early cracks;4) The drawing parts forming AE signal characteristics experimental study plan was enacted and implemented. A drawing bucket as research object, 32 sets of AE signal data was collected, which included no cracks, early crack, micro crack, short crack four kinds of forming state, and according to the above theoretical method, the signal data of each group were processed by MATLAB software, the characteristics of AE signal under different forming conditions are obtained by the way of using the wavelet feature spectrum coefficient map, wavelet decomposition coefficient distribution chart and various types of signal spectrum.Research results show that: the signal data collected by acoustic emission technique were processed and analyzed by MATLAB software with wavelet threshold-EMD de-noising method and wavelet feature parameters analysis method, feature difference of four forming state(especially early crack) AE signals can be obtained, achieve the aim of early crack AE signal characteristics analysis in deep drawing.