Towards Data-Driven Control for Modern Wind Turbines

Towards Data-Driven Control for Modern Wind Turbines

Houtzager, I.

Verhaegen, M. (promotor)
Van Wingerden, J.W. (promotor)

Mechanical, Maritime and Materials Engineering

Delft Center for Systems and Control

2011-12-02

Further developments in data-driven control techniques for the load reduction of modern wind turbines can achieve an increased lifetime of components and make the scaling to larger rotor diameters possible, and therefore improve the cost effectiveness of modern wind turbines. Also the success of future rotor designs will heavily depend for their operation on new developments in active control technologies. This thesis proposes a novel control algorithm for the rejection of periodic disturbances. The load disturbances seen by an individual rotor blade are to a large extent periodic and repetitive; for instance, tower shadow, wind shear, yawed error, and gravity are depending on the azimuth angle and rotational speed, and will change slowly over time. The contribution of these periodic disturbances will become even larger in the future, due to the increasing length and mass of the rotor blades and the possible downwind location of a two-bladed rotor. By compensating only the smooth periodic disturbances, the control action would not create to much fatigue damage on the actuators.

lifted repetitive control
wind turbines
smart rotor
predictor-based subspace identification
data-driven control
recursive identification
adaptive tracking
individual pitch control
flutter detection

http://resolver.tudelft.nl/uuid:297ca587-b291-43cb-a5f7-bb89e34d0cf8

9789053354902

Institutional Repository

doctoral thesis

(c) 2011 Houtzager, I.