Brian Sebastiani

PhD Thesis

Sparse sensing of controlled turbulent jets: optimal placement and event-based strategies.

Supervisors

Andrea Ianiro (UC3M)

Abstract

Turbulent jets are ubiquitous in aerospace and industrial applications, from aircraft propulsion to noise generation and energy systems. Understanding and controlling their complex dynamics demands measurement and sensing strategies that can capture the essential flow physics in real time while remaining tractable in terms of data volume and computational cost. This thesis aims to develop an event-based framework for sensing and estimating the flow field of turbulent jets, including jets subjected to active flow control. The experimental campaign will combine Particle Image Velocimetry (PIV) for high-fidelity flow characterization with microphone arrays deployed around the jet for acoustic sensing. Active control will be applied using loudspeakers or pulsed-jet actuators, since a central goal of the thesis is to reconstruct low-order representations of the flow field not only in natural conditions but also under the effect of actuation — a prerequisite for any practical closed-loop control system. Building on event-based signal processing paradigms and spiking neural networks, the work will investigate how an array of pressure sensors can be used to reconstruct a simplified, low-order representation of the turbulent flow field, mirroring the way biological systems, such as insects, sense and react to their aerodynamic environment with only a handful of event-driven receptors. The proposed framework pursues sparsity along two complementary dimensions: sparsity in space, through optimal sensor placement strategies that minimize the number and optimize the arrangement of sensors required to achieve a target estimation accuracy; and sparsity in time, through event-based processing that triggers acquisition and inference only when dynamically relevant changes are detected. The research is framed within the ERC Consolidator Grant project SPANDRELS (SParse AND paRsimonious Event-based fLow Sensing) and will be carried out in the Department of Aerospace Engineering at Universidad Carlos III de Madrid (UC3M).