Aeronautical Engineering · PhD Candidate

Gianni Cassoni

Microgravity propellant sloshing · in-orbit refuelling & servicing

The interactive model above couples six-degree-of-freedom spacecraft motion, propellant dynamics and compliant capture in an in-orbit refuelling manoeuvre, exposing the coupled rigid-body, fluid and contact response directly in the browser.

About

I am a PhD candidate developing models of microgravity sloshing for in-orbit refuelling and servicing. My work spans semi-analytical formulations for capillary sloshing and a coupled OpenFOAM–MBDyn framework for ullage-impact loads on spacecraft structures, backed by high-performance computing on CINECA Leonardo. I also supervise thesis students in technical work and research planning.

Dec 2024 – Present

Thales Alenia Space & Roma Tre University

PhD Candidate. Semi-analytical capillary sloshing models and a coupled OpenFOAM–MBDyn framework for ullage-impact loads; thesis-student supervision.

Jul 2022 – Present

Politecnico di Milano

Fellow Researcher. Rotorcraft fleet forecasting to 2050, ATTILA tiltrotor aeroelastic flutter (with Leonardo, NLR, DLR), and rotorcraft–pilot coupling data analysis.

Research highlight

Curvature-corrected sloshing spectra for microgravity. Classical slosh models assume a flat interface. In microgravity, surface tension dominates and the equilibrium meniscus is strongly curved, shifting the natural frequencies and mode shapes. The demo above evaluates a reduced-order model built from a Young–Laplace equilibrium and a modal eigen-solve, then time-steps the coupled rigid-body + slosh dynamics live in your browser. Featured update on LinkedIn.

Selected publications

Journal articles
Conference papers & proceedings

Awards & competitive funding

ISCRA-C CINECA HPC Allocation 2026

100,000 CPU core-hours on Leonardo Data Centric and 10,000 GPU-hours on Leonardo Booster for high-performance studies on microgravity sloshing and reduced-order modeling.

Technical skills

PythonMATLABC++ CFortranJulia CUDAMPI / HPCLaTeX OpenFOAM (VOF)MBDyn deal.IISU2SPARTA Smilei (PIC)Dakota Docker / KubernetesGoogle Cloud FirebaseSolidWorks / Inventor Local LLMs & multi-agent systems

Research interests

Beyond spaceflight fluid dynamics, I am drawn to computational nonlinear dynamics — how organized motion persists in systems governed by simple rules, and when a high-dimensional system genuinely admits a faithful low-dimensional description. The same thread runs through periodic orbits and limit cycles, quasi-periodic tori, bounded chaos, and localized coherent structures — from rotorcraft stability to fluid–structure dynamics to continuous cellular automata like Lenia.

A live multi-kernel Lenia organism. A continuous-space cellular automaton evolved with MAP-Elites: two coupled growth kernels keep a coherent structure alive while it translates. Use the buttons on the canvas to switch pointer mode: Attract pulls the organism toward your cursor, Repel pushes it away, Feed drops food where you click or tap. Click the active mode again to turn interaction off.
A unifying residual view

Recurrence & coherent structures

A recurrent or coherent object can be recognized by applying an evolution operator and measuring a residual: how far the state fails to close on itself, possibly after a symmetry or in a moving frame. One language covers exact periodic motion, quasi-periodic near-returns, bounded chaotic recurrence, and local coherent structures — read together with Floquet multipliers and Lyapunov exponents to tell them apart.

Coherent structures & closure

When is a reduced model actually closed?

I study when a compressed feature evolves on its own, weakly coupled to the unresolved degrees of freedom, through a closure residual — a forward-pass–only diagnostic. It turns “is this reduced description valid?” into a number, and travels from Lenia gliders and fluid–structure features to activation-steering directions in large language models on a single consumer GPU.

Contact

Como, Italy · open to collaboration on microgravity fluid dynamics, reduced-order modeling and coupled CFD–multibody simulation.