Simulations of experimental crack development in ceramic material with notched specimen

Sammanfattning: Low-density armour often uses ceramic materials, as their high compressive strength and relatively low weight make the material ideal for moving in the field. The ballistic protection works by redirecting the plastic flow of the projectile radially, which reduces penetration. The main drawback of using ceramics is their brittle behaviour, causing rapid crack propagation. When subjected to ballistic impacts, the ceramic fails catastrophically by spalling and cracking. The fracture behaviour is essential to model correctly to predict the effectiveness of armor systems. In this experiment, a notched specimen was used to investigate mode I fracture. The setup is similar to an open edge on impact experiment where the fracture propagation is initiated by the impact of a projectile fired by an air pressure gun. To avoid spallation, the projectile velocities were kept low to only initiate crack development from bending. For velocities below 15 m/s, a single crack develops across the specimen. For higher velocities between 15-75 m/s, the behaviour is different for short and long projectiles. For short projectiles, a straight crack develops before bifurcating when the projectile and ceramic lose contact. For longer projectiles, the crack bifurcates at the initiation. It forms a pointed ellipse up to the point when the projectile and ceramic lose contact, at which point it behaves the same as for shorter projectiles. Simulations were conducted using LS-Dyna and Impetus AFEA with four different material models. Two different fracture formulations were also used. Erosion was used in LS-Dyna, while erosion and node-splitting were used in Impetus AFEA. Simulations in both LS-Dyna and Impetus AFEA could reproduce experimental results. Simulations show that bifurcation highly depends on the materials' tensile strength. The transition from straight to bifurcating cracks emerges in a small velocity range, allowing for the possibility of calibrating the simulations to experiments. A possibility for comparing the crack development in simulations and the experimental setup besides crack patterns is the crack mouth opening displacement (CMOD). Measurements of CMOD with photon doppler velocimetry (PDV) correlated well with the simulations.