Research on frequency-domain characteristics of acoustic emission signals generated during wood damage and fracture process

This research conducts three-point bending tests on air-dried beech, elm, pine, and cypress specimens to research their damage and fracture behaviour. Wavelet transform is utilized to decompose original acoustic emission signals generated during the damage and fracture process into five layers. Then, based on the principle of signal correlation, the wavelet-decomposed detail signals are adaptively reconstructed. The K-means algorithm is used to classify reconstructed AE signals into different groups based on their frequency-domain characteristic. Research results indicate that the peak frequency of low-frequency AE signals generated during the wood damage and fracture process is approximately 21 kHz. These AE signals may be generated by damage, delamination, bucking, and collapse of the cell-wall layer. When wood specimens are in the later stage of nonlinear deformation and in the macroscopic fracture stage, in addition to the mentioned microscopic damage above, there will also be formation and expansion of microcrack damage zones, as well as cell-wall rupture and macroscopic fracture of wood fibres. In addition to the low-frequency AE signals mentioned above, these AE signals are medium-frequency signals with peak frequency in the range of 25-60 kHz, and high-frequency signals with a peak frequency of approximately 165 kHz.