Damage classification and evolution in composite under low-velocity impact using acoustic emission, machine learning and wavelet packet decomposition

The primary objective of this research is to conduct an exhaustive experimental investigation into the damage identification and damage evolutionary mechanisms in laminated CFRP under lowvelocity impact. The recorded acoustic signals were classified by Principal Component Analysis, Pearson correlation analysis and K-means++ clustering algorithm and four damage modes including matrix cracking, delamination, fiber-matrix debonding and fiber failure are identified. The Wavelet Packet Decomposition is used to analyze the characteristics of the signal in the time-frequency domain. Matrix cracking is the fundamental damage mode, with its characteristic frequency band being FB1. An increase in the FB3 of matrix cracking signals indicates the tendency to the inter-laminar region, leading to delamination damage. Fiber-matrix debonding is a two-phase damage mode. The transition of the energy components from FB1 to FB7, as indicated by the scanning electron microscopy (SEM) results, signifies the evolution of Fiber-matrix debonding. The initial stage is dominated by matrix peeling off the fiber surface, while the second half of the loading shows matrix peeling accompanied by fiber fracture. Fiber breakage failure is caused by local stress concentrations due to matrix cracking. In the later stages of loading, fibers fail by brittle fracture due to mutual compression without the involvement of matrix cracking as reflected by SEM and WPD.