Robust Controller for Turbulent and Convective Boundary Layers
Abstract
Linear feedback controllers and estimators have been designed from the governing equations of a channel flow, linearized about the laminar mean flow, and a layer of heated fluid, linearized about the no-motion state. Spectral decomposition involving a two-dimensional Fourier expansion and a Chebyshev-Galerkin projection cast these linearized equations into state-space form that decoupled to independent Fourier wavenumber sub-systems. The control law are designed by applying Linear Quadratic Gaussian (LQG) synthesis to these sub-systems. The size of the controller is reduced by both limiting the number of sub-systems, on which LQG synthesis is applied, as well as applying system theoretic model reduction techniques to each sub-system. This methodology has produced highly effective controllers to suppress convection in a heated fluid layer, but has found only moderate success with the channel flow. While the Oberbeck-Boussinesq equations (heated fluid layer) provides a direct measure of Rayleigh-B enard convection, the Poiseuille flow equations do not. The feedback control laws for channel flow could only indirectly affect viscous drag. An open-loop optimization has been applied to the channel flow control problem in an effort to capture more of the nonlinear dynamics and, thereby, affect the viscous drag directly. During these experiments, it has been discovered that upstream traveling waves of blowing and suction not only reduces the skin-friction drag in a channel but also sustains it below laminar levels.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 2006
- Accession Number
- ADA456687
Entities
People
- J. John Kim
- Jason L. Speyer
Organizations
- University of California, Los Angeles