This thesis examines a methodology to predict the fatigue life of an engineering elastomer component using a fracture mechanics approach. Previous work on the prediction of the fatigue life of elastomer components involved the application of the stress versus number of cycles to failure. However, this would restrict the prediction to a particular geometry and loading conditions, which may not represent the typical service conditions of the part. By applying fracture mechanics, the strain energy release rates are calculated for cracks in a range of different locations for a range of different strains and deformation modes using a finite element analysis technique. These are used in combination with experimental measurements on pure shear test pieces of cyclic crack growth rates to predict the fatigue failure of the component. In order to use a finite element model it was necessary to characterise the materials appropriately and adopt a strain energy function that describes well the behaviour of the particular elastomer. The fatigue life was reliably predicted to within a factor of 2 for a range of deformations and displacements.

The thesis also considers complications that might result in the prediction of fatigue life because of tear rate transitions from initial sharp cuts in elastomer components. It has been observed that cracks initially grow more rapidly from sharp cuts in elastomer materials under stress than is expected from a consideration of the strain energy release rate alone. The pure shear test piece is used in this work to investigate the extent of the transition zone from the initial fast crack growth rate to the steady state conditions using an analytical expression. The changes in the rate of crack growth are also reflected in the fracture surface appearance which roughens as the crack develops. The roughness of the fracture surface reflects the way the crack tip profile developed during the fatigue process. Simple finite element models are adopted to investigate the causes of this surface roughening.

In many cases the actual direction that a flaw will grow is not obvious, so the ability of a finite element analysis package to predict the crack growth path for simple elastomer test pieces is examined. Different approaches are evaluated to determine the crack growth direction. From a consideration of the ease of implementation, contours of the maximum strain energy density around the tip of an existing flaw are seen to be the most appropriate way to predict the direction of crack growth.