Control theory for infinite dimensional dynamical systems and applications to falling liquid film flows

Abstract

In this thesis, we study the problem of controlling the solutions of various nonlinear PDE models that describe the evolution of the free interface in thin liquid films flowing down inclined planes. We propose a control methodology based on linear feedback controls, which are proportional to the deviation between the current state of the system and a prescribed desired state. We first derive the controls for weakly nonlinear models such as the Kuramoto-Sivashinsky equation and some of its generalisations, and then use the insight that the analytical results obtained there provide us to derive suitable generalisations of the controls for reduced-order long-wave models. We use two long-wave models to test our controls: the first order Benney equation and the first order weighted-residual model, and compare some linear stability results with the full 2-D Navier-Stokes equations. We find that using point actuated controls it is possible to stabilise the full range of solutions to the generalised Kuramoto-Sivashinsky equation, and that distributed controls have a similar effect on both long-wave models. Furthermore, point-actuated controls are efficient when stabilising the flat solution of both long-wave models. We extend our results to systems of coupled Kuramoto-Sivashinsky equations and to stochastic partial differential equations that arise by adding noise to the weakly nonlinear models.