Kunihiko Taira (UCLA) - Modal and Data-Driven Approaches for Active Flow Control

Controlling the behavior of flows around air, marine, and ground vehicles can greatly enhance their performance, efficiency, and safety. The high-dimensionality, strong nonlinearity, and multi-scale properties of these flows make effective their control a tremendous challenge. Without the reduction of the state variable dimension and extraction of important dynamics, the application of dynamical systems and control theory for flow control remains a difficult task. We focus on developing physics-based approaches to model and control complex fluid flows by leveraging modal analysis, data science, network science, machine learning, and high-performance computing. Equipped with these toolsets, we extract essential dynamics to facilitate the development of sparse and reduced-order models to design flow control techniques for high-dimensional unsteady fluid flows. We discuss some of the challenges and successes in characterizing, modeling, and controlling unsteady bluff-body wakes and stalled flows over wings. The techniques developed here are validated with DNS and LES.

https://www.seas.ucla.edu/fluidflow/