Speaker
Description
One of the main challenges in long-term space debris population modelling is the need to explore multiple scenarios in order to assess system evolution under varying conditions and to evaluate different “what-if” situations. However, traditional approaches rely on sequential simulations, making this process time-consuming and computationally inefficient, as they are typically designed to explore single fixed-parameter scenarios.
Inspired by ensemble techniques widely adopted in the space weather domain, and building upon the promising results of The Aerospace Corporation’s ADEPT work, Politecnico di Milano, in collaboration with Telespazio and GMV under ESA’s EXPRO+ contract (“Ensemble and surrogate modelling for debris environment long-term simulation”), introduces ESMILE: an ensemble-driven Particle-in-a-Box model designed for efficient and automated multi-scenario exploration. The model generates artifacts that serve as interpolating weights for a surrogate model, which is made available through a graphical interface.
This presentation focuses on the architecture of the Particle-in-a-Box model and its ability to maximise computational efficiency through reusability across the simulation campaign. The input parameter space is sampled efficiently to enhance domain exploration, while stochastic events are fully decoupled from the population evolution dynamics. These dynamics are computed once and reused whenever possible across the sampled parameter space.
The model incorporates also techniques to perturb and aggregate data from different modelling methods, in order to add variability and explore the uncertainties and errors in the simulations.
Finally, stochastic processes are handled using a layer-based approach combined with a branch-and-bound strategy that enables the exploration of multiple statistically driven evolutions without repeating the propagation step. This significantly reduces the computational overhead associated with scenario analysis.
As a result, the proposed framework enables the execution of many simulations in significantly less time required by traditional sequential models, providing stakeholders with a powerful tool to assess the impact of mitigation strategies and fragmentation events on the long-term evolution of the space debris environment.
| Which section would you like to submit your abstract to? | Session 2: “Challenges of space debris modelling” |
|---|