Acoustic Emission (AE) Signal Classification from Tensile Tests on Plywood and Layered Wood

Even today, a detailed description of the damage progress in wood or wood materials under tensile loading is still a challenge. The complexity of the damage behaviour results from the various mechanisms occurring simultaneously at several length scales. So far, few studies focused on mechanical behaviour of wood or wood materials by analysing acoustic emission (AE), although AE can provide information on multi-scale damage mechanisms and damage accumulation. The high time resolution of AE measurements is beneficial for detection of micro-mechanisms, their interactions and accumulation leading to macroscopic failure. For the AE analysis presented here, several types of industrial plywood and layered wood materials made from spruce were subjected to quasi-static tensile loads and simultaneously monitored by AE. Since polymer-composites and wood can be assumed to behave analogously, especially regarding their anisotropic properties, application of pattern recognition methods for fibre reinforced polymer-matrix composites are expected to have also high potential for AE signal classification of wood fracture. Such unsupervised pattern recognition, e.g. based on the frequency domain of the AE signals, are purely mathematical approaches to perform signal classification and to identify natural classes of AE signals, respectively. Within the present investigation, a signal classification approach originally developed for fibre-reinforced composite laminates is explored for plywood and layered wood materials. Problems and challenges are identified which have to be solved for a detailed understanding of their damage behaviour. The different layered structures of plywood all yield two AE signal clusters which can roughly be differentiated in signals of relatively high shares of low frequency and high frequency content, respectively. These occur essentially over the whole test duration and yield comparable AE signal amplitudes and energies. The challenge is to assign the features of the detected signals to their microscopic source mechanism. M or e In fo a t O pe n A cc es s D at ab as e w w w .n dt .n et /? id = 17 52 7