Fracture of composite materials : the use of constant stress intensity geometries in studies of composite fracture behaviour

Abstract

Composite materials have taken on an increasingly important role in high performance structural applications. For the designers this places an increased emphasis on accurately characterising the failure strengths of such materials. The failure criteria commonly in use are often found lacking and this is a result of the complex failure mechanisms that occur. Due to the non-homogeneous and anisotropic nature of composites, damage can take place in the matrix, interface and fibres of any given lamina. In addition the restraining effect of fibres bridging a matrix crack is known to be a significant factor in the toughness of composites. This thesis concerns itself with testing methods for characterising these failures and evaluating the fibre bridging effect. The advantages of test geometries which exhibit stable crack growth have been previously realised. Using a geometry such as this allows a crack to be grown across a fibre bundle to give subsequent loading of the fibres. This geometry has the potential to evaluate fracture properties of all the constituents in the one test as well as investigating fibre bridging behaviour. In this work two geometries have been evaluated, the double torsion and the double cantilever beam. A Finite Element study of the double torsion geometry has shown the torsional nature of the loading causes interaction between the two cracked ligaments. Contact between the ligaments is responsible for the elliptical shaped crack front profile and produces a deviation from the analytical solution of at least 50% higher stress intensity. The addition of fibres to the geometry further emphasises the contact behaviour and creates problems in accurately characterising the interfacial toughness properties. Behaviour of polymeric thermosetting matrix materials has also been investigated. Unstable crack growth phenomenon has been shown to be a result of a visco-elastic response that manifests itself in the crack propagation behaviour. The unstable propagation will have a significant effect on the failure behaviour of composites and contributes to the scatter in composite failure strengths observed experimentally. Both the Double Torsion and Double Cantilever Beam geometries produce desirable fibre bridging behaviour. The influence of unstable cracks growing across the fibre bundle has been found to have no effect on the subsequent fracture behaviour. Experimentally, the double cantilever beam geometry has been shown to have the advantage of having all the loading in plane. This simplifies the fibre load and allows an analytical solution. Having an analytical solution in turn eliminates the need for additional experimental measures. These are significant advantages for the double cantilever beam geometry and highlights its suitability for further composite fracture studies.