Definition and Response Validation of a Simplified Aero-Servo-Elastic Wind Turbine Model

Abstract

Currently, the foundation design of offshore wind turbines is based on a simple cantilever beam model with a mass representing the rotor and nacelle, or by making use of wind turbine simulation software such as Bladed, which are built for rotor design. The simple models are unrealistic since the dynamic interaction between the aerodynamic loads, hydrodynamic loads, control system and the structure is not included which leads to a lack of understanding of the behavior of the structure and possibly a poor design. The simulation software models, on the other hand, require input that the wind turbine designing party is not willing to provide, are computationally expensive, and since these are 'black box' models, give no insight into the system. In between these methods, there is currently unoccupied room for a hybrid model which includes relatively advanced rotor aerodynamics but remains computationally inexpensive while providing good insight into the behavior of the system.
This thesis aims to define such a model based on modification of models present in literature and validate it by comparing its behavior to commercial wind turbine design software. First, a model of only the rotor with a rigid frictionless drivetrain shaft and blade pitch control system (the model restricts itself to the above rated regime) is considered and validated by comparison with Bladed. The blades of the rotor are assumed to be identical and rigid, the flow is assumed to be attached and the wind velocity field is uniform and has only one directional component. The effect of the use of different wake models is tested and it is concluded that for a step wind input the equilibrium wake model is most suited, while for a turbulent wind input the dynamic wake model is the best option. Simplification of the model results in the conclusion that the lift and drag coefficients can be evaluated for the mean wind velocity and chosen to be time independent without having any significant effect, while the induction factor cannot be chosen to be time independent without it significantly affecting the behavior of the model. The aerodynamic torque is linearized with respect to wind velocity, rotational rotor velocity, pitch angle and induction factor, which results in an acceptable approximation while the operating conditions are within reasonable proximity of the chosen mean operation state. Afterwards, a tower structure and flexible drivetrain shaft are added to the model, which again is validated by using Bladed. Frequency domain analysis shows that the tower motions of both the model and Bladed are similar, thus validating the model. The aerodynamic excitation is linearized with respect to wind velocity, structural motion, rotational rotor velocity, pitch angle and induction factor and applied to the model including a tower structure and flexible drivetrain shaft. Lastly, it is concluded that after simplification of the model and linearization of the aerodynamic excitation the model results in a good approximation of the wind turbine simulation software.