THE EFFECT OF SOFTENING ON THE PREDICTED STRENGTH OF BRITTLE MATERIALS USING A CONTINUUM DAMAGE MECHANICS FAILURE MODEL

Li Shi,Haiyan Li,S. L. Fok,B. J. Marsden,Suyuan Yu
2004-01-01
Abstract:A continuum damage mechanics (CDM) failure model, which can predict both the initiation and propagation of cracks in brittle materials such as ceramics and nuclear graphite, has been developed and implemented into the commercial finite element code ABAQUS. In this model, both the stress-based failure criterion and fracture-mechanics-based failure criterion are used to define the damage surface, and a softening parameter 'n' is adopted to control the rate of damage development. Specifically, a change in the softening parameter changes the shrinkage rate of the damage surface, thus giving rise to different traction-displacement behaviors at the interface where crack initiation and propagation are taking place. This paper studies the effect of the degree of softening in the CDM model on the predicted strength of brittle materials subjected to loadings with different stress gradients. 0. FOREWORD The failure of brittle materials, such as nuclear graphite and ceramics, is not only a function of component geometry and size but also of loading mode. Therefore, full-size component testing is often required to determine the failure loads. Although various computational methods have been developed to predict the lifetime and failure load of brittle components, the absence of a satisfactory failure model is still an impediment to accurate failure predictions. Several commonly used failure models of graphite were reviewed in Ref. (1). The authors showed that simple models based on critical stress, critical strain and critical strain energy density criteria were remarkably unsuccessful in describing the experimental results. A continuum damage mechanics (CDM) model (2) has been developed for nuclear graphite at the Manchester School of Engineering (MSE), University of Manchester. The CDM model can be viewed as an extension or generalization of the cohesive zone model (3), in that it couples together the effects of mode-I and mode-II loadings. An interface with no thickness is introduced into the continuum solid at positions where potential crack surfaces may form. Within the interfacial constitutive law, which relates tractions and displacements across the interface, a damage parameter is introduced to take account of the effects of damage such as mcro-cracking. A damage surface is then constructed using both the conventional stress-based and fracture-mechanics-based failure criteria. Therefore, the CDM model can predict both the initiation and propagation of cracks in components with high stress concentration or even stress singularity.
What problem does this paper attempt to address?