A long-standing objective in applied science and engineering has been to reduce complex nonlinear differential equations to simple, low-dimensional models. A more recent objective is to seek such reduced models directly from numerical or observational data. Examples of areas where all this presents serious challenges include vibration prediction for airplane wings, elimination of sloshing in fluid transportation, controlling of soft robots, identifying material properties of hydrogels, controlling the near-resonant oscillations of MEMS gyroscopes efficiently, or predicting transition to turbulence in pipe flows.
The main topic of these Spring School lectures, model reduction to spectral submanifolds (SSMs), introduces mathematically well-founded solutions to equation- and data-driven model reduction. SSMs are very low dimensional attracting invariant manifolds tangent to eigenspaces (spectral subspaces) of linearizations of nonlinear systems at steady states. These manifolds generally carry the core nonlinear dynamics of the system which range from nontrivial fixed points through periodic or quasiperiodic motions to chaotic attractors. Consequently, the internal dynamics of SSMs provide low-dimensional reduced-order models with which typical system trajectories synchronize exponentially fast. Open-source Matlab and Python codes with a growing library of worked examples are now available for data-driven SSM reduction of physical systems. The lectures introduce the participants to the theory of SSMs and illustrate their use on the specific problems listed as motivation above. The participants will also be guided in carrying out data-driven SSM modeling and prediction on simple benchmark problems using the open-source code SSMLearn (see https://github.com/haller-group).
As previous years, the workshop is meant to be a learning opportunity for doctoral and postdoctoral students, where they get to work interactively in groups on open research problems and in close contact with the organizers and the invited speaker.
Applications of prospective participants should contain:
The applications can be submitted to any of the organizers. Selected candidates will subsequently be notified and will receive a fellowship that covers accommodation. Travel costs to and from the workshop venue have to be carried by the home institute. Registration fee is 200€
We would be grateful if you could disseminate this information to colleagues that could be interested in this event.
ERCOFTAC Montestigliano Spring School is an event for young scientists covering a specific multi-disciplinary and modern topic in fluid mechanics. Its format is based on an interactive teaching concept. Besides the traditional supervision by the student's advisor, the typical education of a PhD student seldom involves hands-on training under the guidance of a true expert in a particular and complementary field. University courses succeed in laying a foundation in rudimentary (and even a few specialized) concepts. Conferences, minisymposia and standard workshops expose young scientists to recent advances. In Montestigliano, students collaborate intensively during a full week in small groups under the guidance of a world-leading expert in a specialized topic.
The workshop will take place at the picturesque estate of Montestigliano in the heart of Tuscany. The village of Montestigliano is composed of 18th century buildings, typically of Tuscan architecture. Located 15 km south-west of Siena, it provides a stunning landscape and a stimulating environment for both intensive work and relaxation. The 18th century houses and farm buildings have been carefully restored to retain their original features common to Tuscan architecture. For further information visit: http://www.montestigliano.it/.