Electromagnetic non-destructive testing of damaged fibered-reinforced composite laminates. From computational modeling to imaging. - Archive ouverte HAL Access content directly
Conference Poster Year :

Electromagnetic non-destructive testing of damaged fibered-reinforced composite laminates. From computational modeling to imaging.

(1) , (2) , (1) , (3)
1
2
3

Abstract

How to model the time-harmonic electromagnetic behavior of possibly damaged fiber-reinforced composite laminates is considered within full-wave and approximated frameworks. This involves so-called super-cell concepts, but also encompasses a variety of first-order models, notably homogenization. Once sound approaches set up, and validated at frequencies from MHz to THz, depending on materials and geometries, with testing as possible by FEM, we focus onto imaging. Emphasis is mostly limited to pin-pointing several missing fibers in the laminate, a number of methodologies being proposed. The latter go from rather straightforward MUSIC-like imaging to sparsity-constrained iterative solutions, those involving to care for the aforementioned sparsity resulting from lack of a few fibers. Very near to more far-field scattered field data in the reflection mode are assumed to be collected in different configurations of potential pratical interest. Particular emphasis is also put on the idea of super-localization as possibly associated to resonances exhibited within the laminate at frequencies such that probing wavelengths are about the fiber radii, while enhancement of resolution is to proceed from low-frequency testing involving to go from the small-scale models to large-scale models.
Not file

Dates and versions

hal-01536720 , version 1 (12-06-2017)

Identifiers

  • HAL Id : hal-01536720 , version 1

Cite

Zicheng Liu, Changyou Li, Dominique Lesselier, Yu Zhong. Electromagnetic non-destructive testing of damaged fibered-reinforced composite laminates. From computational modeling to imaging.. Les Journées COFREND - COFREND Days, May 2017, Strasbourg, France. ⟨hal-01536720⟩
104 View
0 Download

Share

Gmail Facebook Twitter LinkedIn More