Composite materials are experiencing a huge expansion in use on high-value components for their high strength and low weight. To increase damage tolerance, complex architectures such as 3D weaving are being used. Existing analysis techniques are focussed on laminated composites, and are unable to capture the complex behaviour and realize the full potential of 3D weaved material. The aim of this study was to develop a technique by which a 3D weaved component could be simulated to determine damage intiation and growth, and understand the response of the structure. Multi-scale techniques have been employed to achieve this in a practical, accurate and efficient manner. A model has been produced by extending established ‘layerwise’ finite element methods used for laminated (1D/2D) composites in order to generate globally accurate displacements and selective locally accurate stresses, and combining with models to simulate delamination and transverse tow pullout. Additionally, asymptotic expansion homogenization models are used to resolve stresses to the tow scale, enabling an accurate assessment of damage accumulation at the micro-scale and the resulting effect at the component scale. The model has been implemented in MATLAB, and the constituent parts of the model have been validated against analytical, test and published data. It has been proven to produce good results at a lower computational cost or greater resolution than other equivalent models. The capability of the unified model has been demonstrated by simulating the failure process of a double cantilever beam constructed from a 3D weaved composite.
|Date of Award||8 Oct 2014|
|Supervisor||Michele Meo (Supervisor)|