https://hal-centralesupelec.archives-ouvertes.fr/hal-01330285Lesselier, DominiqueDominiqueLesselierL2S - Laboratoire des signaux et systèmes - UP11 - Université Paris-Sud - Paris 11 - CentraleSupélec - CNRS - Centre National de la Recherche ScientifiqueLambert, MarcMarcLambertGeePs - Laboratoire Génie électrique et électronique de Paris - UP11 - Université Paris-Sud - Paris 11 - UPMC - Université Pierre et Marie Curie - Paris 6 - CentraleSupélec - CNRS - Centre National de la Recherche ScientifiqueOn recent advances and issues ahead in modeling and electromagnetic imaging of perturbed composite laminatesHAL CCSD2016simulationfiber-reinforced composite laminatesmodelingNon-destructive testing[SPI.ELEC] Engineering Sciences [physics]/Electromagnetism[SPI.SIGNAL] Engineering Sciences [physics]/Signal and Image processing[MATH.MATH-MP] Mathematics [math]/Mathematical Physics [math-ph]Lesselier, DominiqueL. Tsang et al.2016-06-10 12:38:112023-03-24 14:53:022016-06-10 12:38:11enConference papers10.1109/PIERS.2016.77342711Non-destructive testing of fiber-reinforced composite laminates as found, e.g., in airplanes, cars and for green energy applications, is challenging from academia to industry. Even if focusing onto electromagnetic testing (not by far the most common), a replete literature exists, encompassing low-frequency modalities tailored to graphite-fiber-based laminates to less explored microwave ones when preocupied with glass-fiber-based materials up to THz regimes to reach submillimetric resolution, leaving aside that not only fibers matter but also the matrix which contains them, and the way the plies in the laminate are arranged, which in effect yields a host of composite laminates and corresponding model complexities. Here is only concerned with two models: a large scale one (large enough local wavelengths vs. main geometric features) involving a locally-averaged complex-valued permittivity dyad (via possibly heuristic homogenization) associated to unixial anisotropy in any given ply ; a small scale one (small enough local wavelength vs. geometry), wherein each ply contains commonly-orientated, periodically-set fibers parallel with its interfaces, the common orientation usuallychanging from ply to ply. For both, damages of various kinds are faced: at large scale, volumetric inclusions mostly; at small scale, missing, misplaced or disorientated fibers. Any well-conceived imaging procedure then must involve proper understanding of the electromagnetic behavior of both sound and damaged structures, and be tailored to data effectively collected (usually in the near¯eld using suitable sources and probes) and defects sought and appraised in electromagnetic and geometric terms. Both direct solutions and inverse ones (one-shot or iterative ones) within exact and first-order frameworks are briefy covered up in the present contribution and illustrated by a few key numerical simulations, enabling also to outline potential research ahead, in harmony with possibly more concrete industrial cases.