Damage Identification Algorithms for Composite Structures

Composites present additional challenges for inspection due to their heterogeneity and anisotropy, the fact they fail by interacting modes, and since often damage occurs beneath their surface. Currently successful laboratory non-destructive methods, such as X-ray and C-scans, are impractical for inspection of large integrated structures. It is clear that new approaches for inspection of composites need to be developed. To resolve this issue, during past BAA and SBIR work with the NRO, AFOSR, NSF and NASA, the Metis Design Corporation (MDC) has developed a structural health monitoring system components for damage detection in composites using Lamb waves. This technique has provided reliable information about the presence, location and type of damage successfully for simple laboratory specimens, and during the course of this SBIR MDC proposes to continue to develop the coded algorithms to increase accuracy and precision. Specifically, MDC will be using pattern recognition techniques in order to improve the reliability of identifying damage type, severity and location. This rule-based method hold potential over traditional logic-based methods for increased robustness due to the constant machine-learning process to further classify (and sub-classify) damage states.