Model predictive control of the lift forces of two interacting airfoils

The context of this thesis is to develop a controller which acts on the pitch angles of airfoils in order to output constant lift forces. This lift-force control strategy is divided into two phases: the first focuses on the control of a single airfoil and the second considers the interaction between two airfoils with pitching and heaving capabilities. The later takes into consideration five different vertical position configurations for the airfoils, so the flow interaction is analyzed for different crosswind spacing. Each phase is respectively motivated by the individual pitch control of wind turbines’ blades and by the control of wind turbines in a wind farm arrangement. The control framework is designed as a Dynamic Matrix Control (DMC) scheme, which is a type of a Model Predictive Controller. The airfoil system is nonlinear and is modeled, along with the linear dynamics of the pitch motor, as a Wiener model. Aiming at approximating the system’s model using a linear model, the linearizing pre-compensation method is applied to the airfoil’s static nonlinearity. Hence, a linear DMC is applicable to the proposed strategy. With the purpose of testing the approach experimentally, new concepts of test-sections for a closed-return wind tunnel are developed for the two aforementioned research phases. Both concepts consist of wall-mounted structures, that hold the airfoils horizontally inside the test-section. Stepper motors, which actuate on the pitch angles of the airfoils, are each connected to one side of the airfoils’ shafts. In order to measure the aerodynamic loads on the airfoils, force sensors are mounted at both sides of the walls. Both phases of research are experimentally validated using closed-loop real-time control on a hardware target machine. Furthermore, a comprehensive simulation study is carried out in the Ansys Fluent Computational Fluid Dynamics (CFD) software. Using these CFD simulations, the flow along a pitch-variable airfoil and between two interacting airfoil are visually evaluated. The airfoil system is modeled in accordance to the wind-tunnel experimental cases and is controlled in co-simulation with MATLAB/Simulink