ALEAS project

Thwarting extreme events in turbulent flows

Impact

Turbulence is a state of fluid motion characterized by swirling movements occurring across a wide range of spatial and temporal scales. It is all around us—in the atmosphere, the oceans, and many other natural systems. Turbulent flows are prone to sudden and unpredictable changes. These natural hazards, increasingly intensified by climate change, can have devastating effects on infrastructure. Examples of seemingly unforeseeable extreme events include unexpected storms, rogue waves that swamp vessels or coastal installations, and violent gusts that destabilize wind turbines. Yet, buried within the complex turbulence of underlying flows are faint early warning signs. Our ability to adapt to climate change hinges on how effectively we can detect and interpret these weak signals. Inspired by birds’ remarkable ability to anticipate major storms, the ALEAS project seeks to learn how to identify subtle precursors of disasters in turbulent environments.

Limitations to overcome

Turbulence fosters the emergence of extreme events that remain poorly understood. Simplified theoretical models fall short in predicting them, their numerical simulation demands immense computing resources, and their rarity makes them hard to study systematically in experiments. The ALEAS project aims to reinvent methods of measurement, simulation, and modeling to address these challenges. It will generate a comprehensive and sustainable database of extreme events, laying the groundwork for a strategy to detect the weak signals that precede them. 

Risks

Artificial intelligence will be employed to guide numerical simulations toward the occurrence of extreme events. A novel biomimetic strategy will focus on changes between states rather than the states themselves. Accounting for interactions with the environment to detect signs of bifurcations represents a conceptual breakthrough.

Innovation potential

The ALEAS project will leverage ultra-high-resolution imaging based on event-driven imaging technology, which detects changes at the pixel level without needing to acquire full images. This approach will enable the development of new, cost-effective, and environmentally sustainable fluid measurement techniques. In the future, the project could also lead to the creation of new neuromorphic AI systems that analyze differences and relationships within the database—ushering in a form of AI that more closely mimics how neurons function.