Microwave NDT&E using open-ended waveguide probe for multilayered structures

Abstract

Microwave NDT&E has been proved to be suitable for inspecting of dielectric structures due to low attenuation in dielectric materials and free-space. However, the microwave responses from multilayered structures are complex as an interrogation of scattering electromagnetic waves among the layers and defects. In many practical applications, electromagnetic analysis based on analytic- and forward structural models cannot be generalised since the defect shape and properties are usually unknown and hidden beneath the surface layer. This research proposes the design and implementation of microwave NDT&E system for inspection of multilayered structures. Standard microwave open-ended rectangular waveguides in X, Ku and K bands (frequency range between 8-26.5 GHz) and vector network analyser (VNA) generating sweep frequency of wideband monochromatic waves have been used to obtain reflection coefficient responses over three types of challenging multilayered samples: (1) corrosion progression under coating, (2) woven carbon fibre reinforced polymer (CFRP) with impact damages, and (3) thermal coated glass fibre reinforced polymer (GFRP) pipe with inner flat-bottom holes. The obtained data are analysed by the selected feature extraction method extracting informative features and verify with the sample parameters (defect parameters). In addition, visualisation methods are utilised to improve the presentation of the defects and material structures resulting in a better interpretation for quantitative evaluation. The contributions of this project are summarised as follows: (1) implementation of microwave NDT&E scanning system using open-ended waveguide with the highest resolution of 0.1mm x 0.1 mm, based on the NDT applications for the three aforementioned samples; (2) corrosion stages of steel corrosion under coating have been successfully characterised by the principal component analysis (PCA) method; (3) A frequency selective based PCA feature has been used to visualise the impact damage at different impact energies with elimination of woven texture influences; (4) PCA and SAR (synthetic aperture radar) tomography together with time-offlight extraction, have been used for detection and quantitative evaluation of flat-bottom hole defects (i.e., location, size and depth). The results conclude that the proposed microwave NDT&E system can be used for detection and evaluation of multilayered structures, which its major contributions are follows. (1) The early stages (0-12month) of steel corrosion undercoating has been successfully characterised by mean of spectral responses from microwave opened rectangular waveguide probe and PCA. (2) The detection of low energy impact damages on CFRP as low as 4 Joules has been archived with microwave opened rectangular waveguide probe raster scan together with SAR imaging and PCA for feature extraction methods. (3) The inner flat-bottom holes beneath the thermal coated GFRP up to 11.5 mm depth has been successfully quantitative evaluated by open-ended waveguide raster scan using PCA and 3-D reconstruction based on SAR tomography techniques. The evaluation includes location, sizing and depth. Nevertheless, the major downside of feature quantities extracted from statistically based methods such as PCA, is it intensely relies on the correlation of the input dataset, and thus hardly link them with the physical parameters of the test sample, in particular, the complex composite architectures. Therefore, there are still challenges of feature extraction and quantitative evaluation to accurately determine the essential parameters from the samples. This can be achieved by a future investigation of multiple features fusion and complementary features.