MEchanics GAthering –MEGA- Seminar: Space under stress: Numerical challenges in the modeling of dynamic fragmentation of orbital collisions
Event details
| Date | 30.04.2026 |
| Hour | 13:05 › 14:00 |
| Speaker | Thibault Ghesquière-Diérickx (LSMS, EPFL) |
| Location | Online |
| Category | Conferences - Seminars |
| Event Language | English |
Abstract: The increasing density of orbital debris poses a significant threat to the sustainability of the space environment. Hypervelocity collisions and satellite breakups produce fragment clouds that fuel subsequent impacts, driving the cascade effect known as the Kessler syndrome. Physics-based predictions of these events require numerical models of dynamic fragmentation, yet current operational breakup models still rely heavily on empirical parameters.
Such simulations must capture multiple time scales, from crack nucleation and propagation to fragment dispersion and secondary interactions, posing significant challenges for numerical robustness and computational efficiency. Explicit time integration is the standard approach for this multiscale regime, yet prior work showed that combining it with an extrinsic cohesive-zone model and penalty-based contact leads to severe instabilities. Negligible over short durations, these instabilities worsen over time, introducing artificial fragmentation and energy errors that compromise the reliability of fragment statistics.
To address these limitations, we adopt an impulse-based contact treatment that respects the inherently nonsmooth nature of impact. Within an otherwise explicit time-integration scheme, contact constraints are resolved via an implicit solve based on the nonsmooth Newmark-β approach. One-dimensional dynamic fragmentation benchmarks demonstrate that this hybrid approach restores energy conservation and eliminates unphysical fragmentation. Despite the overhead of the implicit contact solve, the enhanced stability permits significantly larger explicit time steps, delivering overall computational efficiency comparable to or exceeding that of penalty-based methods. This establishes impulse-based contact as a robust alternative for the long-duration fragmentation simulations underlying space-debris models.
Bio: After completing my Bachelor's and Master’s degrees in Civil Engineering at EPFL, where I specialized in structural engineering, I joined the Computational Solid Mechanics Laboratory (LSMS) for my PhD. Currently in my third year, I focus on developing robust finite-element methods for dynamic fragmentation applied to hypervelocity orbital collisions, with particular emphasis on bridging fracture and contact mechanics in stable, parallel, open-source simulation tools.
Such simulations must capture multiple time scales, from crack nucleation and propagation to fragment dispersion and secondary interactions, posing significant challenges for numerical robustness and computational efficiency. Explicit time integration is the standard approach for this multiscale regime, yet prior work showed that combining it with an extrinsic cohesive-zone model and penalty-based contact leads to severe instabilities. Negligible over short durations, these instabilities worsen over time, introducing artificial fragmentation and energy errors that compromise the reliability of fragment statistics.
To address these limitations, we adopt an impulse-based contact treatment that respects the inherently nonsmooth nature of impact. Within an otherwise explicit time-integration scheme, contact constraints are resolved via an implicit solve based on the nonsmooth Newmark-β approach. One-dimensional dynamic fragmentation benchmarks demonstrate that this hybrid approach restores energy conservation and eliminates unphysical fragmentation. Despite the overhead of the implicit contact solve, the enhanced stability permits significantly larger explicit time steps, delivering overall computational efficiency comparable to or exceeding that of penalty-based methods. This establishes impulse-based contact as a robust alternative for the long-duration fragmentation simulations underlying space-debris models.
Bio: After completing my Bachelor's and Master’s degrees in Civil Engineering at EPFL, where I specialized in structural engineering, I joined the Computational Solid Mechanics Laboratory (LSMS) for my PhD. Currently in my third year, I focus on developing robust finite-element methods for dynamic fragmentation applied to hypervelocity orbital collisions, with particular emphasis on bridging fracture and contact mechanics in stable, parallel, open-source simulation tools.
Practical information
- General public
- Free
Organizer
- MEGA.Seminar Organizing Committee