MEchanics GAthering -MEGA- Seminar: Talk 1 - Weak nonlinearity for strong nonnormality; Talk 2 - Network-based modeling of turbulent fows
Event details
Date | 21.10.2021 |
Hour | 16:15 › 17:30 |
Speaker | Yves-Marie Ducimetière (LFMI, EPFL), Daniel Fernex (UNFoLD, EPFL) |
Location | Online |
Category | Conferences - Seminars |
Event Language | English |
Talk 1: Weak nonlinearity for strong nonnormality, by Yves-Marie Ducimetière (LFMI, EPFL)
Abstract We propose a theoretical approach to derive amplitude equations governing the weakly nonlinear evolution of nonnormal systems, when they experience transient growth or respond to harmonic forcing. This approach reconciles the non-modal nature of these growth mechanisms and the need for a center manifold to project the leading-order dynamics. Under the hypothesis of strong nonnormality, the methodology is outlined for a generic nonlinear dynamical system, and two application cases highlight two common nonnormal mechanisms in hydrodynamics: the flow past a backward-facing step, subject to streamwise convective nonnormal amplification, and the plane Poiseuille flow, subject to lift-up nonnormality.
Bio I am a PhD student at the Laboratory of Fluid Mechanics and Instabilities (LFMI). I am interested in weakly nonlinear effects on complex systems, typically fluid flows, and subject to stochastic/harmonic forcing and/or transient growth.
Talk 2: Network-based modeling of turbulent fows, by Daniel Fernex (UNFoLD, EPFL)
Abstract Complex nonlinear dynamics govern many fields of science and engineering. Dynamic modeling for the long-term features is a key enabler for understanding, modeling, prediction, control, and optimization. Here we present the cluster-based network modeling bridging machine learning, network science, and statistical physics. Cluster-based network modeling describes short- and long-term behavior and is fully automatable, as it does not rely on application-specific knowledge. This method is demonstrated on numerous examples including a high-dimensional boundary layer flow.
Bio Daniel Fernex is a postdoctoral fellow in the Unsteady Flow Diagnostics Laboratory (UNFoLD) at EPFL. During his PhD at the Technical University Braunschweig he developed network-based techniques to derive simple models from higly non-linear and high-dimensional physical systems. At EPFL, he extends his methods to the field of unsteady aerodynamics to model, understand and optimize systems such as vertical axis wind turbines and pitching airfoils.
Abstract We propose a theoretical approach to derive amplitude equations governing the weakly nonlinear evolution of nonnormal systems, when they experience transient growth or respond to harmonic forcing. This approach reconciles the non-modal nature of these growth mechanisms and the need for a center manifold to project the leading-order dynamics. Under the hypothesis of strong nonnormality, the methodology is outlined for a generic nonlinear dynamical system, and two application cases highlight two common nonnormal mechanisms in hydrodynamics: the flow past a backward-facing step, subject to streamwise convective nonnormal amplification, and the plane Poiseuille flow, subject to lift-up nonnormality.
Bio I am a PhD student at the Laboratory of Fluid Mechanics and Instabilities (LFMI). I am interested in weakly nonlinear effects on complex systems, typically fluid flows, and subject to stochastic/harmonic forcing and/or transient growth.
Talk 2: Network-based modeling of turbulent fows, by Daniel Fernex (UNFoLD, EPFL)
Abstract Complex nonlinear dynamics govern many fields of science and engineering. Dynamic modeling for the long-term features is a key enabler for understanding, modeling, prediction, control, and optimization. Here we present the cluster-based network modeling bridging machine learning, network science, and statistical physics. Cluster-based network modeling describes short- and long-term behavior and is fully automatable, as it does not rely on application-specific knowledge. This method is demonstrated on numerous examples including a high-dimensional boundary layer flow.
Bio Daniel Fernex is a postdoctoral fellow in the Unsteady Flow Diagnostics Laboratory (UNFoLD) at EPFL. During his PhD at the Technical University Braunschweig he developed network-based techniques to derive simple models from higly non-linear and high-dimensional physical systems. At EPFL, he extends his methods to the field of unsteady aerodynamics to model, understand and optimize systems such as vertical axis wind turbines and pitching airfoils.
Practical information
- General public
- Free
Organizer
- MEGA.Seminar Organizing Committee