Smoothed Particle Hydrodynamics in pkdgrav3 for Shock Physics Simulations. II. Shear Strength

TL;DR

This paper integrates a shear strength model into the pkdgrav3 SPH code, validating it against experiments and analyzing its impact on planetary collision outcomes, revealing significant effects on disruption thresholds and morphology.

Contribution

It introduces a comprehensive strength model into a GPU-accelerated SPH code for planetary impacts, enabling realistic simulations of solid mechanics at large scales.

Findings

Strength increases disruption thresholds in low-mass collisions.

Convergence to fluid behavior occurs near 0.7 Earth masses.

Including strength modestly impacts computational performance.

Abstract

Material strength effects have been recently shown to be significant in giant impacts even at scales of planetary collisions. Despite this, their effects are often neglected in numerical giant impact simulations. We present an implementation of a basic strength model (pressure dependent shear strength) in the massively parallel smoothed particle hydrodynamics code pkdgrav3. The model includes elastic deviatoric stresses, plasticity with pressure-dependent yield strength, and thermal softening, and is fully integrated into the GPU-accelerated framework introduced in Paper I, preserving its scalability and performance characteristics. We validate the implementation against laboratory experiments of granular cliff collapse and our simulation results are in excellent agreement. We then determine the catastrophic disruption threshold, $Q_{RD}^*$, over a wide mass range of the colliding bodies using simulations performed both with and without material strength. Consistent with prior work, we find that strength substantially increases $Q_{RD}^*$ in the low-mass regime, while convergence toward the fluid limit occurs only near $R_{C1} \sim 10^7$ m ($\sim 0.7,M_\oplus$), well above the often assumed $\sim 100$ km size limit. Entropy production and remnant morphology likewise remain sensitive to rheology at intermediate masses. Performance measurements show that including strength introduces only modest computational overhead while maintaining favorable scaling, thereby enabling realistic solid mechanics in large-scale impact simulations.