Bond-based peridynamics model of 3-point bend tests of ceramic-composite interfaces

The need for ceramic composites is increasing over a vast range of industrial applications, including energy and nuclear technology (structural materials for fission and fusion applications); space (thermal shields) or chemical sectors. Current finite element (FE) models are not able to reproduce the aleatory nature of crack generation and propagation through materials and interfaces. To tackle the shortcomings of FE models, a nonlocal bond-based peridynamics model implemented in the Abaqus FE code has been updated to reproduce the behavior of interfaces between ceramic materials. Two- and three-dimensional simulations have been used to simulate three-point bend tests for a range of ceramic composites and compared with experiments available in the literature. The model reproduced most of the experimental failure loads, as well as correctly reproducing the failure modes and crack patterns. In particular, the model was able to predict two mechanisms for interface failure, cohesive failure, in which the two materials separate exactly at the interface, and cohesive failure of the substrate, in which a thin layer of the weaker material remains attached to the stronger material and a crack propagates in the weaker material, rather than at the interface. Both failure modes are experimentally observed and still difficult to reproduce with currently available FE simulations.