Active control of fluid flows has been the subject of extensive research in fluid dynamics, driven by its potential benefits across a wide range of engineering applications, from renewable energy and industrial processes to transportation optimisation. Control objectives vary significantly, from reducing skin friction drag to managing complex phenomena such as flow separation. In recent years, significant advancements have been made in this field, thanks to improvements in experimental hardware, computational power, sensing and actuation strategies, and the introduction of powerful data-driven methodologies for designing both open- and closed-loop controllers. In particular, reduced-order models, combined with appropriate closure modelling through neural networks or data-driven functional approximation, have shown great promise for providing accurate predictions and potentially enabling control of complex systems. These strategies allow the leverage of first-principles knowledge of physics, augmented by novel methodologies for designing more efficient controllers. Moreover, numerous experimental and numerical studies have demonstrated the potential of reinforcement learning (RL) techniques for control design. In RL, policies are learned directly through interactions with the environment, removing the need for prior models.
Workshop Organisation
The colloquium will focus on two main topics:
- Modeling and Prediction of Fluid Flows
- Model-based and Model-free Strategies for Flow Control
Contributions from numerical, theoretical, and experimental perspectives are welcome. The workshop will emphasise the theoretical aspects and practical applications of model-based and model-free control, with particular attention to the advantages and potential pitfalls associated with each alternative approach.
This colloquium will focus on these recent developments, aiming to foster discussion and the exchange of ideas on the most promising protocols for data-driven control while encouraging networking among leading European and international experts in data-driven fluid mechanics. We look forward to meet in Paris!