Multi-layer carbon fiber reinforced plastic characterization and reconstruction using eddy current pulsed thermography

Abstract

Carbon fibre composite materials are widely used in high-value, high-profit applications, such as aerospace manufacturing and shipbuilding – due to their low density, high mechanical strength, and flexibility. Existing NDT techniques such as eddy current testing suffers from electrical anisotropy in CFRP (carbon fibre reinforced plastics). Ultrasonic is limited by substantial attenuation of signal caused by the multilayer structure. The eddy current pulsed thermography has previously been applied for composites NDE (non-destructive evaluation)such as impact damage, which has the ability for quick and accurate QNDE(quantitative non-destructive evaluation) inspection but can be challenging for detection and evaluation of sub-surface defects, e.g., delamination and debonding in multiple layer structures. Developing QNDE solutions using eddy current thermography for addressing subsurface defects evaluation in multi-layer and anisotropic CFRP is urgently required. This thesis proposes the application of eddy current pulsed thermography (ECPT) and ECPuCT (eddy current pulse compression thermography) for tackling the challenges of anisotropic properties and the multi-layer structure of CFRP using feature-based and reconstruction-based QNDE and material characterisation. The major merit for eddy current heating CFRP is the volumetric heating nature enabling subsurface defect detectability. Therefore, the thesis proposes the investigation of different ECPT and their features for QNDE of various defects, including delamination and debonding. Based on the proposed systems and QNDE approach, three case studies are implemented for delamination QNDE, debonding QNDE, conductivity estimation and orientation inverse reconstruction using the two different ECPT systems and features, e.g., a pulse compression approach to increase the capability of the current ECPT system, the feature-based QNDE approach for defect detection and quantification, and reconstruction-based approach for conductivity estimation and inversion. The three case studies include 1) investigation of delamination with different depths in terms of delamination location, and depth quantification using K-PCA, proposed temporal feature-crossing point feature and ECPuCT system; 2) investigation of debonding with different electrical and thermal properties in terms of non-uniform heating pattern removal and properties QNDE using PLS approaches, impulse response based features