Understanding the damage accumulation and tensile strength in carbon fibre reinforced polymers using high resolution in situ computed tomography

Currently, composite components are widely adopted in aerospace applications but typically over-designed due to the lack of reliable predictive models for their mechanical properties. The objective of this thesis is to reach a higher level of understanding of the damage accumulation processes occurring in tensile loaded carbon fibre reinforced polymer systems to advance the development of predictive models. In situ loading combined with high-resolution Synchrotron Radiation Computed Tomography (SRCT) imaging has allowed a data-rich investigation of the key mechanisms leading to the final failure in different material systems. An extensive database of performance and damage behaviour has been compiled and systematic comparisons performed for material systems consisting of aerospace and industrial grade fibres, as well as different levels of adhesion of the fibre/matrix interface, obtained through changes in the sizing agent and fibre surface treatment. Focus was given to the damage mode that drives tensile failure in unidirectional layers loaded in the fibre direction: i.e. fibre failure, as this is often considered to be the final failure event in the application of multi-layered components. Clusters of fibre breaks (hereby indicated as multiplets) are believed to play a significant role in the provision of a critical crack site that propagates to final failure. Both qualitative and quantitative analyses have been performed, indicating that the accumulation of fibre breaks does not have a simple correlation to the macroscopic properties of the material, such as the ultimate tensile strength (UTS) and the fibre type but particularly the fibre/matrix interface have been observed to affect the multiplet formation. The morphology of local damaged sites has been investigated in a novel statistical approach, needed to distinguish very similar fibre arrangements. Automated tools have been specifically developed to extract fibre breaks and the fibre shapes from low contrast images with high fibre volume fraction. The fibre misorientation in damaged sites is seen to differ statistically from that in non-damaged sites (using well recognised statistical tools) and the single fibre misorientation distributions show a consistently higher standard deviation in orientation when compared to intact fibre distributions, even though locally damaged sites did not exhibit a peculiar fibre packing arrangement. The research provides a unique database of data as well as automated statistically inferred tools that can provide a better insight into the fundamental mechanisms leading to tensile failure in longitudinally loaded composites, supporting the model development in two different phases: (a) at the initial stages, identifying phenomena and influences that should be included in accurate but parsimonious model formulation and (b) at a quantitative calibration/verication point, providing critical but previously unavailable numerical descriptions of micro-mechanical processes.