a b s t r a c tGlass fiber-reinforced composite laminates in polyetherimide resin have been studied via terahertz imaging and ultrasonic C-scans. The forced delamination is created by inserting Teflon film between various layers inside the samples prior to consolidating the laminates. Using reflective pulsed terahertz imaging, we find high-resolution, low-artifact terahertz C-scan and B-scan images locating and sizing the delamination in three dimensions. Furthermore, terahertz imaging enables us to determine the thicknesses of the delamination and of the layers constituting the laminate. Ultrasonic C-scan images are also successfully obtained; however, in our samples with small thickness-to-wavelength ratio, detailed ultrasonic B-scan images providing quantitative information in depth cannot be obtained by 5 MHz or 10 MHz focused transducers. Comparative analysis between terahertz imaging and ultrasonic C-scans with regard to spatial resolution is carried out demonstrating that terahertz imaging provides higher spatial resolution for imaging, and can be regarded as an alternative or complementary modality to ultrasonic C-scans for this class of glass fiber-reinforced composites.

No abstract

Ultrasonic guided waves are capable of rapidly interrogating large, plate-like structures for both nondestructive evaluation and structural health monitoring (SHM) applications. Distributed sparse arrays of inexpensive piezoelectric transducers offer a cost-effective way to automate the interrogation process. However, the sparse nature of the array limits the amount of information available for performing damage detection and localization. Minimum variance techniques have been incorporated into guided wave imaging to reduce the magnitude of imaging artifacts and improve the imaging performance for sparse array SHM applications. The ability of these techniques to improve imaging performance is related to the accuracy of a priori model assumptions, such as scattering characteristics and dispersion. This paper reports the application of minimum variance imaging under slightly inaccurate model assumptions, such as are expected in realistic environments. Specifically, the imaging algorithm assumes an isotropic, non-dispersive, single mode propagating environment with a scattering field independent of incident angle and frequency. In actuality, the composite material considered here is not only slightly anisotropic and dispersive but also supports multiple propagating modes, and additionally, the scattering field is dependent on the incident angle, scattered angle, and frequency. An isotropic propagation velocity is estimated via calibration prior to imaging to implement the non-dispersive model assumption. Imaging performance is presented under these inaccurate assumptions to demonstrate the robustness of minimum variance imaging to common sources of imaging artifacts.