• Extending experimental data on glass fragmentation into the high pre-stress domain (up to 158 MPa) using a fully automated computer vision pipeline capable of digitizing complete fracture patterns containing up to 40,000 fragments. • Demonstrating, for the first time, that mean fragment base area and perimeter approach energy-dependent lower bounds at high pre-stress levels, indicating a transition in the dominant energy dissipation mechanism. • Confirming that support conditions (all-sided, two-sided and bedded) have no significant influence on mean fragment density and local fracture intensity especially at high energy states, providing a basis for simplified modelling assumptions. • Demonstrating good agreement between experimentally observed fragment size distributions and Eigenerosion-based numerical simulations for both thicknesses. • Quantifying three-dimensional fracture surface morphology via CT scanning and box-counting fractal dimension, establishing a statistically significant positive correlation between strain energy and surface roughness.
Bohmann et al. (Mon,) studied this question.