Active aerodynamic load control on wind turbines: Aeroservoelastic modeling and wind tunnel

This thesis investigates particular concepts and technologies that can alleviate fatigue loads on wind turbines by using distributed active aerodynamic devices on the blades, a concept briefly referred to as `smart blades'. Firstly, published research work on smart control devices is reviewed, and the pros and cons for the application on wind turbines is discussed. Then an introduction to the quantification of unsteady loads on the blades of a representative 5MW reference wind turbine is given. Based on this, certain requirements for the load control concepts to alleviate the fatigue loads are presented. Moreover, an overview of possible smart rotor concepts with a view towards application to wind turbines is provided. The choice is made for active load control by trailing edge flaps. The theoretical background for the unsteady aerodynamics modeling of airfoils with trailing edge flaps, is included, focusing on wind turbine blades. Moreover, results of 2D aerodynamic load control with aps are shown. Experimental wind tunnel work on a aeroelastically scaled non-rotating blade, where feedback control using piezoelectric-material based flexible flaps is also performed. A developed full wind turbine aeroservoelastic code is presented, capable of simulating rotors with span-wise distributed trailing edge flaps and various sensor and control capabilities. Furthermore, novel wind tunnel experimental work on a scaled rotor equipped with flexible flaps is shown, providing measured load alleviation results. Numerical predictions, correlated with the scaled rotor experiment, using the developed aeroservoelastic tool are shown, giving a reasonable correspondence. Various numerical predictions with the use of the aeroservoelastic model for the case of a representative 5MW wind turbine are presented. Aeroelastic behaviour of smart wind turbine rotors and design of various complexity controllers for active flaps are analysed. The fatigue load can be lowered up to 27%. Conclusions from this research work are finally drawn, with a view towards future work in the field.