The fracture of polygranular graphites

Abstract Six commonly used models of graphite failure are described and their performance critically reviewed against empirical data derived from a variety of tests to failure in tension and bending. The simpler models based on critical stress, critical strain and critical strain energy density criteria are shown to be remarkably unsuccessful in describing the experimental results. The Weibull model, though versatile in its application to geometry-related effects, is far less useful in treating the influence of microstructural variations. The final two treatments considered, the so-called Rose/Tucker and fracture mechanics models, are much wider ranging in their usefulness, and it is argued that a new theory which embodies the salient points of each should be developed. The microstructural processes which ought to be included in such a description are identified on the basis of observations on subcritical defect development prior to failure.