Methodology for monitoring the damage in composite materials under tensile loading by acoustic emissions clustering

Acoustic emissions (AE) registration allows detection of micro damage events in materials and structures and potentially identification of the damage nature. In the case of textile reinforced polymers the evolution of cumulative energy of AE events was adopted for the identification of damage initiation and propagation during loading. But methods for the identification of the damage modes based on the features of the AE events are still under investigation. Cluster analysis is a powerful methodology to analyze multi-parametrical AE signals. The analysis classifies events and creates a framework for subsequent identification of the links between the established cluster event classification and the physical nature of the damage. In this work, the AE cluster analysis methodology, previously used for 2D and 3D glass woven composites, is applied to investigate the peculiarities of the AE evolution and clustering in thermoset and thermoplastic reinforced 2D carbon and glass textiles. One thermoset (Epoxy) and two thermoplastic (Polyphenylenesulphide, PPS, and Polyetheretherketone, PEEK) resins were considered. The epoxy was reinforced with a balanced twill 2x2 carbon fabric and a balanced twill 2x2 glass fabric. The PPS and the PEEK were reinforced with a carbon plain weave textile. Tensile tests were assisted with two acoustic sensors connected to the Vallen AMSY-5 system. Two high resolution cameras were adopted: the first for the evaluation of the full filed strain by digital image correlation technique, the second for the local damage observation in the thickness. The comparison of the cumulative energy of AE events show that the thermoplastics reinforced composites have an early beginning of the damage and rapid saturation while a later beginning and more continuous growth of the damage was monitored in the textiles reinforced epoxy materials. From the primary nine AE features, three were selected for the cluster analysis: peak amplitude (A), peak frequency (PF) and frequency centroid of gravity (FCOG) of the signal. The cluster analyses highlighted that the thermoset and thermoplastic reinforced composites have a net separation of the clusters. The three different types of clusters can be subdivides as: CL1 low amplitude and low frequency; CL2 high amplitude and low frequency; CL3 high frequency. The recorded limits of the three clusters are similar to the previously found in literature. The crack monitoring in the thickness allowed the preliminary correlation of the transverse cracks in the PPS reinforced composite to the CL1 cluster containing AE events of low amplitude and low frequency.