Acoustic emission signal processing based on breaking lead test
                                                                           GUO Li, HUO Ke ke
                  (College of Mechanical and Vehicle Engineering, Hunan University, Changsha 410082, China)

Abstract: Aiming at the present situation of wide application and no in depth analysis of the acoustic emission experimental signal of pencil lead breaking,in order to reveal the characteristics of pencil lead break acoustic emission (AE) signal in a deeper level and to explore the difference of pencil lead break AE signal under different analysis methods, empirical mode decomposition (EMD) and 5 layer discrete wavelet transform (DWT) were respectively used to decompose and reconstruct pencil lead breakAE signals. The time domain and frequency domain features of the decomposed and reconstructed AE signals were analysed. In the EMD 6 intrinsic modal functions were obtained by using the correlation coefficient method to analyse the AE signal in depth.The results indicate that low frequency noise in the AE signal is filtered and the lead fracture characteristic information can be accurately extracted by EMD and DWT. At the same time, the analysis of pencil lead break AE by short time Fourier transform also find that the noise is distributed in the similar frequency range, which further proves the accuracy of the low frequency noise filtered in the AE signal filtered.EMD and 5 layer wavelet transform are used to decompose and reconstruct pencil lead break AE signal and the low frequency noise signalin the AE signal is filtered out, which can extract the lead fracture characteristic information. At the same time, the analysis of pencil lead break AE by short time Fourier transform also proves its accuracy.

Key words: acoustic emission; pencillead break; wavelet analysis; empirical mode decomposition（EMD）; short time Fourier transform