Estimation and control of turbulent flows

Abstract

Flow control has become increasingly important in the last few years in the aeronautical field, since it has been shown as a promising tool for improvement of aerodynamics and aeroacoustics of aircraft. Hereby, estimation techniques have also assumed greater importance, this time not only for allowing a better comprehension of the flow dynamics itself, but also as a tool to assist flow control. The particular case of jet noise generation shall be addressed by this research, which can imply in highly turbulent flows and are thus intrinsically challenging. To tackle this problem a resolvent-based method shall be used for the estimates of acoustic noise sources. The resolvent analysis allows to deal the non-linearity of a dynamical system as a forcing term, and then, investigate the problem in an input-output manner, where one could identify the optimal forcing terms (near-field turbulence) and their corresponding optimal responses (far-field noise). Thereby, by associating sensor readings to these terms, it is possible to estimate flow statistics related to non-linear terms. To tackle the control problem, the Wiener-Hopf technique is considered here since non-causality is observed in several flow control problems solved in the frequency domain. This method allows to enforce causalitywhen obtaining the control kernel, providing an optimal causal solution, and with this, avoiding the drop in performance that may be observed in flow control applications when using a truncated solution. (AU)