IDENTIFICATION OF DAMAGE MODES IN CERAMIC MATRIX COMPOSITES BY ACOUSTIC EMISSION SIGNAL PATTERN RECOGNITION

The control of damage is a key factor for the durability and reliability of parts in service. Sound identification of damage mechanisms and kinetics is an important issue with a view to predicting remaining lifetime. For that purpose, analysis of acoustic emission (AE) is an interesting and powerful tool. Damage of fiber reinforced ceramic matrix composites involves several phenomena at various length scales including matrix cracking, fiber debonding, fiber failures. The objective of the present paper is to propose a quantitative approach to damage identification based on signal analysis by pattern recognition. An unsupervised classification method was used to differentiate the signals generated during reference fatigue tests performed on C/SiC samples at high temperature. Each class of signals was then associated to relevant damage mode. A library of signals was then used to identify the damage modes generated during fatigue tests performed at various temperatures (supervised classification method).