This paper presents an efficient technique to uniquely identify damage through the creation of a multi-scale map of material deformations and behavior within a representative volume element (RVE) of the host structure. Optimally distributed, embedded fiber optic sensors provide strain, strain gradient, and integrated strain fields throughout the RVE. As a demonstration, an isotropic, homogeneous RVE is modeled instrumented with an evenly spaced grid of sensing elements. The multi-scale damage identification technique and an equivalent single-scale method are evaluated on the basis of damage detection and identification. A large number of induced, stochastic damage cases are analyzed, generated by introducing a crack defined by three random variables: center location, length, and orientation angle. The multi-scale sensing capability is shown to provide a higher quality strain map of the RVE from the distributed sensors, resulting in significantly improved damage identification.

This content is only available via PDF.
You do not currently have access to this content.