Acoustic emission (AE) technique is an efficient non-destructive method for detection and identification of various damage mechanisms in composite materials. Discrimination of AE signals related to different damage modes is of great importance in the use of this technique. For this purpose, integration of k-means algorithm and genetic algorithm (GA) was used in this study to cluster AE events of glass/epoxy composite during three-point bending test. Performing clustering analysis, three clusters with separate frequency ranges were obtained, each one representing a distinct damage mechanism. Furthermore, time-frequency analysis of AE signals was performed based on wavelet packet transform (WPT). In order to find the dominant components associated with different damage mechanisms, the energy distribution criterion was used. The frequency ranges of the dominant components were then compared with k-means genetic algorithm (KGA) outputs. Finally, SEM observation was utilized to validate the results. The obtained results indicate good performance of the proposed methods in the damage characterization of composite materials
