Acoustic emission characterization of failure mechanisms in oriented strand board using wavelet-based and unsupervised clustering methods
The acoustic emission (AE) technique has been used for decades to detect damage onset and propagation in complex materials such as composites. In this research, acoustic emission along with mechanical testing were applied for a better understanding and characterizing of micro- and macrostructural damage mechanisms leading to overall failure and fracture of specimens. For this purpose, strands and also unidirectional laboratory-made oriented strand board specimens were loaded under tensile and bending loads. Parametric analysis of AE, unsupervised pattern recognition algorithm of k-means, and wavelet packet transform were applied to the analysis of the acquired data. The results of the analysis of AE waveforms revealed that the fingerprint of frequency range of strands AE response has been detected in the strand board specimens. The data analysis led to detection and discrimination of different failure mechanisms such as microcracks of strands, strands failure, debonding, and delamination of strand boards using acoustic emission technique. The frequency ranges of strand failure mechanisms, local debonding of strands, and delamination were defined as 0–200, 200–300, and 300–400 kHz, respectively.