MEchanics GAthering –MEGA- Seminar: Variational computation of invariant solutions in wall-bounded chaotic flows
Event details
| Date | 16.05.2024 |
| Hour | 16:15 › 17:15 |
| Speaker | Omid Ashtari (ECPS, EPFL) |
| Location | Online |
| Category | Conferences - Seminars |
| Event Language | English |
Abstract:
The dynamics of a chaotic fluid flow is supported by invariant solutions embedded in the chaotic attractor of the system. These are non-chaotic solutions to the governing equations with a simple dependence on time, such as equilibria and periodic orbits. Despite their significance, the identification of invariant solutions remains a computational challenge, rendering many solutions inaccessible.
We introduce a new family of methods for computing invariant solutions. We recast the computation of an invariant solution as a minimization problem in the space of all sets of the same topological structure as the sought-after solution, e.g., the space of all loops when searching for a periodic orbit. The deviation of a trial set from satisfying the definition of the objective solution is penalized by a non-negative cost function. The minimization of the cost function evolves a guess until at a global minimum, where the cost function takes a value of zero, an invariant solution is found. This eliminates the need for any time-marching of the flow, resulting in a robust convergence compared to state-of-the-art methods.
In this talk, the feasibility of the method and its superior convergence are demonstrated by computing several invariant solutions of the 1D Kuramoto-Sivashinsky dynamics as well as 3D Couette flow.
Bio: Omid is a PhD candidate at the Emergent Complexity in Physical Systems (ECPS) lab at EPFL. Before joining ECPS, he completed his B.Sc. and M.Sc. at the University of Tehran, Iran, where he specialized in computational fluid dynamics. His current research focuses on developing algorithms for nonlinear and chaotic dynamical systems, aimed at applications in fluid turbulence.
Bio: Omid is a PhD candidate at the Emergent Complexity in Physical Systems (ECPS) lab at EPFL. Before joining ECPS, he completed his B.Sc. and M.Sc. at the University of Tehran, Iran, where he specialized in computational fluid dynamics. His current research focuses on developing algorithms for nonlinear and chaotic dynamical systems, aimed at applications in fluid turbulence.
Practical information
- General public
- Free
Organizer
- MEGA.Seminar Organizing Committee