When a shock wave reflects off the free surface of a metal sample, it interacts with geometric irregularities (treated as defects such as those caused by a machining tool). At sufficiently high pressures, the shock reflection, followed by the material's expansion into the surrounding medium, can trigger the high-speed ejection (several thousand m/s) of liquid-phase metal sheets from the free surface. These sheets stretch and fragment, creating a particle cloud that can interfere with certain experiments. This process is extremely difficult to characterize experimentally or simulate, as it occurs at challenging spatio-temporal scales and involves material fracture physics that current computational codes based on hydrodynamic equations cannot naturally reproduce.
This thesis aims to address this specific issue. Drawing notably on theoretical analyses, it seeks to incorporate fragmentation processes into simulations where the material, modeled using a continuum (hydrodynamic) approach, is subjected to very high strain rates. One of the objectives is to predict the size distributions of the particles resulting from sheet fragmentation as a function of their velocity and to compare these predictions with available experimental data.