Wind turbines convert kinetic energy of the wind into electrical energy. Unfortunately, this process is everything but constant, as the wind source shows large fluctuations with high and low frequencies. This turbulence, together with the wind shear and yawed inflow, excites the turbine structure, thereby driving the loads and the design of turbines in general and blades in particular. In response to this, several control mechanisms have been applied to wind turbines since the generation of stall controlled machines in the 1980s. While collective pitch control was applied first, the control mechanisms have become more localised and act on individual turbine blades, rather than on the rotor as a whole. An advanced control scheme is termed 'smart wind turbine'. These type of wind turbine actively measures vibrations of its blades through a set of distributed sensors throughout the blades and then aims to counteract the vibrations using aerodynamic modifications around the blades' trailing edges close to the tips by means of control surface deflections. This thesis investigates two aspects of the smart rotor concept: the analysis of smart rotors and the design of an autonomous flap concept. For the analysis, a wind turbine analysis tool with special focus on smart rotors and controller implementation has been developed. This code, the Delft University Smart Wind turbine Analysis Tool (DU-SWAT), has been benchmarked not only against conventional wind turbine codes, but a comparison study with the first utility-scale smart rotor experiment, the Sandia National Laboratories Smart Rotor, was performed. The experimentally obtained eigenfrequencies of the test turbine matched closely those of the numerical study. The difference in the first eigenfrequency is 2.7% or 0.1 Hz (4.4 Hz experimentally, 4.5 Hz numerically). A second comparison step was a time domain analysis of the wind turbine response to a step deflection input of the flaps. For the tower response, the frequencies and the amplitudes of the numerical and experimental responses agree very well. For blade vibrations, an increase in damping in the numerical simulations is observed. While for low flap deflection amplitudes, up to 5 degrees, the response amplitude is predicted well. When high step deflections are modelled, the numerical simulations increasingly fail to accurately capture the dynamics of the turbine. In combination with the differences in damping, this leads to the conclusion that vortices, shed from the flap tips, interact with the larger tip vortices, possibly due to the proximity of the flaps to the blade tips. This inaccuracy of high flap deflection angles is however of limited importance, as it was demonstrated that the periodic (1P) load, the most dominant contributor to fatigue damage, could be alleviated effectively even with deflection angles up to 5 degrees. The individual flap controller has been tuned to the NREL 5MW reference turbine and has been used to study both fatigue and extreme loads according to the certification regulations. Failure-free cases were included in the analysis, and loads have been monitored throughout the turbine. The fatigue load reduction of the blade root bending moment of 24\% corresponds well with the findings of previous researchers. Besides this verification, it was also shown that the structural loads increase nowhere in the turbine, with the exception of the blade root torsional moment. Several other loads decrease, for example the tower torsion moments and the bending moments in the turbine shaft. The extreme load reduction is smaller than the fatigue load reduction. Still, the ultimate tip deflection and the ultimate blade root bending moment could be reduced by 7\% and 8\%, respectively. The moments in the tower are also reduced. Besides load alleviation, an additional functionality of the smart rotor was established. The flaps can be used to increase the power production of the turbine by responding to fluctuations in the wind speed and the delays in the adjustment of the rotor speed due to the rotor inertia. An intermediate step of the wind turbine analysis was the development of a suitable structural model. The developed structural dynamics model, which is based on modal equations of motion, is not limited to wind turbine structures, but rather applicable to a broad range of engineering problems concerning structural vibrations. The model closes the gap between modal reductions, which are typically used in linear vibration analysis, and non-linear geometry. For that purpose the structure is segmented and the segments are joined by rigid-body displacements in a co-rotational framework, which introduces geometric non-linearities. This allows modelling of the structural dynamics for large deformations, while maintaining linear stress information of the finite element model of all segments. The basic assumption underlying this approach is that the structural displacement is large, but the strains remain small, which is typically the case for slender structures such as wind turbine blades. The second major topic, which has been addressed in this dissertation, is the physical implementation of a flap system. The described flap system is fully autonomous and is mounted as a free-floating flap, which means that the flap can freely rotate around a hinge axis. The flap is controlled by a trailing edge tab and driven by servo actuators. The flap is mass underbalanced and aeroelastically unstable in interaction with one of the main structural modes. This renders the flap system highly responsive to control inputs, but also to external excitations. When vibrating, the kinetic energy of the flap is converted by electromagnetic harvesters into electric energy. This energy is either stored in a battery or used to power the sensors and the actuators. It was demonstrated that the instability of the flap dramatically increases the amount of harvested energy by, in case of the experiment, a factor of 225 for wind speeds just below and above the flutter speed. The flap system measures the vibrations through accelerometers. When unstable, the vibration amplitude is either limited by structural delimiters or can be actively controlled by the control system. It was shown, that the flap system can be self-sufficient during the controlled limit cycle oscillation. Id est the power produced during limit cycle oscillation is greater than the power consumed to keep the oscillation amplitude constant. The main advantage of the autonomous flap is its improved replaceability compared with non-autonomous ones. As it neither needs a connection to a central control unit and a power system, nor is an integral part of the wind turbine blades like seamless solutions, it can be exchanged easily in case of failure. In conclusion, smart wind turbines have a great potential to improve the cost efficiency by reducing loads for most turbine components as has been shown in this dissertation. This can be achieved using the novel flap concept, which helps, due to its plug-and-play nature, to reduce maintenance costs.