Alicia Rodríguez Asensio

PhD Thesis

Identification of actuation manifolds in wall-bounded flows

Supervisors

Stefano Discetti (UC3M) and Andrea Ianiro (UC3M)

Abstract

Closed-loop turbulence control holds the promise of significantly enhancing efficiency and reducing the environmental impact in various domains, including transportation, energy production, and industrial processes. Nevertheless, the complex and multi-scale nature of turbulence poses a challenge for real-time sensing strategies, necessitating a low-dimensional model to efficiently select the most appropriate control measures. In this context, reduced order models and dimensionality reduction become crucial. Reduced order models based on coherent structures have been obtained for several flows and it is often feasible to identify low-dimensional representations of the dynamics of uncontrolled flows. Such low-dimensional representations lead to the identification of n-dimensional surfaces (referred to as manifolds) over which flow systems may “live”. To date, however, this is still a relatively unexplored territory when dealing with flow control scenarios. Identifying actuation manifolds, i.e., manifolds of systems under prescribed control laws, would be a key enabler for the future of closed-loop control of turbulent flows.

This thesis will explore whether an actuation manifold can be identified in wall-bounded flows in presence of control actions, which can be in the form of pulsed jets, plasma actuators, or moving parts. Furthermore, it will be explored the possibility to map the full state of the flow on the manifold coordinates, to use it as a guidance for the definition of optimal control strategies, as well as to interpret black-box strategies such as those based on machine learning techniques. Efforts will be made to ensure that the resulting control laws are straightforward and interpretable, offering valuable insights from laboratory experiments.

Project participation

EXCALIBUR – Extracción de estrategias de aprendizaje automático para el control de flujos turbulentos

Ref.No.: PID2022-138314NB-I00

Funding Entity: Agencia Estatal de Investigación (AEI)

Doctoral meetings