Antonio Cuellar Martín

Closed loop control for turbulent flows in aeronautical propulsion.

Supervisors

Stefano Discetti (UC3M) and Andrea Ianiro (UC3M)

Abstract

This thesis is devoted to the development of closed-loop control control algorithms for active control systems for drag reduction and efficiency in wall bounded turbulent flows. Artificial Intelligence (AI) is to be used for this purpose, as a state of art set of tools with a great potential. This control strategy may require to know, or at least to have an estimation of the state of the fluid around the wall; to that end, the first part of the work is focused on the development of algorithms and neural networks to predict the velocity field from wall measurements. Then, some control laws are to be established as black-box systems, using these field predictions, involving a manifold learning task with the most relevant information from the flow. The final step would be to develop control laws as grey-box systems, where there is a direct physical interpretation, and therefore, it would be allowed to scale the setup for alternative configurations.

Project participation

Doctoral Meetings 2021-2022