| Flow fields are traditionally manipulated via passive and active open-loop techniques. These strategies are useful for certain situations, but lack desired performance in many cases. My PhD research formulates a method for developing closed-loop direct adaptive control algorithms for complex flow fields. This dissertation research project begins by analyzing the underlying theoretical equations for both fluid dynamics and control system design. Because flow dynamics are highly non-linear, linearizing the governing equations creates dramatic modeling errors for control design. Alternatively, open-loop computational fluid dynamic (CFD) simulations provide the basis for low dimensional, non-linear models which accurately predict and simulate flow states. This low dimensional model allows for simulation of off-design flow cases, reference signals, and feedback control derivation. Adaptive control was chosen to deal with the high non-linearities of fluid mechanics and unforseen uncertainties between model simulations and experiments.;These control design methodologies are then applied to two separate flow fields. The first of them is a simple cylinder wake displaying Von Karman vortex street. The second and the main focus of this paper, addresses the aero optics issue and control of a free, unstable shear layer. The proposed reduced order modeling, state estimation and control development methods are successfully applied to these drastically different flow fields. The project is broken into three main areas: first, flow simulation; second, reduced order modeling which addresses numerical reduction and system identification strategies; third, feedback control which entails sensor placement, state estimation, and control design. |
