The success of wind power as a renewable energy source depends on its cost of energy. Wind turbine control has attracted much attention in the controls community due to its potential impact on the cost of wind power. However, novel methods in the literature have not transitioned well to industry. This is because the potential cost benefits of these methods are not well understood. There is a need for basic research to address this issue. This thesis is one step toward transitioning of advanced control methods in literature to the industry. Particularly, we aim to understand the limits of performance. The potential performance improvements of the advanced methods should be large enough to justify their cost and complexity. We investigate the optimal trade-offs between multiple turbine performance goals. We also explore the use of a novel wind preview sensor in closed-loop control laws. The impact of this novel sensor on the optimal turbine performance is investigated.
The specific contributions of this thesis can be grouped in three categories. First, we present a preliminary, nonlinear optimization based controller design and analysis framework. This framework can simplify the design of the advanced multivariable controllers for nonlinear systems. It can also be used to investigate the optimal design trade-offs between nonlinear performance constraints and objectives. Second, engineering insight is provided into turbine design trade-offs. Third, we provide mathematical tools that quantify the limits of turbine performance in presence of preview wind measurements. Optimization tools that can analyze the trade-off between preview time and operating condition dependent turbine performance objectives are presented. In low wind speeds, our results show that simultaneous power capture improvements and structural load reductions can be obtained. In high wind speeds, a short amount of preview wind information can be used to overcome the fundamental performance limitations imposed by actuator rate constraints. We provide analytical formulas that quantify these preview time requirements and performance limitations. A convex optimization framework is also presented for the analysis of extreme operating conditions that are defined by deterministic wind disturbance trajectories.