Optimised ultrasonic data acquisition and surface-echo suppression for 3D characterisation of composite materials

Student thesis: Doctoral ThesisEngineering Doctorate (EngD)

Abstract

The predominant non-destructive method for inspecting fibre-reinforced composites is normal-incidence pulse-echo ultrasound, although evolving designs, combined with the anisotropic nature of composites, demand the ongoing development of more advanced signal-processing techniques and testing equipment. During this EngD program, the focus has been the development of methods that improve pulse-echo ultrasonic inspection and advanced 3D characterisation of composite materials.
The specific aims involve: i) improving ultrasonic defect detection and material characterisation by developing post-processing algorithms that suppress the impedance-mismatch effects at the front and back interfaces between the coupling medium and the laminate, whilst retaining all information from the material itself; ii) assessing and optimising the quality of raw ultrasonic data using newly defined data-quality metrics. The motivation from which these aims arose include problems frequently encountered in industry, specifically: i) the masking effect of the strong front- and back-surface reflections, which can prevent the detection of near-surface defects and ii) the absence of standards for ultrasonic inspection of composites aiming to assess and improve ultrasonic data quality in order to facilitate 3D characterisation.
The suppression of the front- and back-surface reflections from each ultrasonic analytic-signal response was accomplished by developing an algorithm for construction, and then subtraction, of the individual front- and back-surface echoes, isolating the response of the material itself. The independent application to each response waveform and the ability to retain all information from the material are crucial advances over previous attempts to do this. They allow higher contrast of defects and the application to non-uniform surfaces. The method was tested in a variety of composite material configurations, both simulated and real, with different types of defects, testing the bounds of applicability and the ability to improve detectability of defects and tracking of near-surface plies.
Metrics aimed at assessing data quality, giving confidence to delay processing and analysis of the data, have also been shown to allow optimised 3D characterisation of composites by analysing the effect on quality of several ultrasonic data-acquisition parameters. The metrics were developed and tested using different materials (real and simulated) across a range of acquisition parameters such as centre frequency, bandwidth, sample rate, scan step size and focal length versus probe size. As a result, guidance has been given regarding the recommended ranges of acquisition parameters based on material properties such as ply spacing and fibre-tow width.
Finally, in a study merging the two projects, the new data-acquisition metrics were employed to assess quantitatively the benefits, in terms of data quality, of applying the surface-suppression algorithm.
Date of Award19 Mar 2024
Original languageEnglish
Awarding Institution
  • The University of Bristol
SupervisorRobert A Smith (Supervisor), Paul D Wilcox (Supervisor) & Steve Eichhorn (Supervisor)

Cite this

'