Non-Linear Control for Floating Wind Turbines

Abstract

As the world is shifting away from fossil fuel-based electricity production, mainly due to concerns for man-made climate change, a sharp rise in demand for wind powered electricity production is seen. In order to meet future demand, the vast amount of off-shore wind resources can be unlocked. A key technology for this task are floating wind turbines (FWTs), as they are less constrained by water depth than fixed wind turbines. One challenge for FWTs is sometimes referred to as ‘negative aerodynamic damping’: interaction between the blade pitch controllers and the fore-aft motions of the floating platform (pitch and surge) can cause instabilities leading to large oscillations in platform motion and rotor speed. These interactions translate to right half-plane (RHP) zeros, limiting the system bandwidth and thereby the performance of the wind turbine. This increases the variability in the power output of the FWT. This thesis looks into a promising type of non-linear control to overcome this limit: reset control. Reset control is characterized by a higher phase compared to linear controllers. This could be used to improve the stability margins of the system and/or to increasethebandwidth. ThisthesisdesignsaCgLpcontrollerspecifically[1]. UsingtheCgLp, a theoretical improvement is seen in the frequency domain. Therefore, a better performance is expected. However, after time domain simulations, it is shown that this controller is not fit for application in blade pitch control. Two different configurations of the CgLp are considered. In the default configuration, large peaks are present in the control signal, leading to excessive motions of the blades. The second configuration performs well under idealized circumstances, but with high frequency disturbance (such as turbulent wind or sensor noise), excessive amounts of resets prevent the controller from reacting appropriately.