A novel fuzzy logic pitch angle controller with genetic algorithm optimization for wind turbines

Official URL: https://risc01.sabanciuniv.edu/record=b2862148_(Table of contents)

As one of the most preferred renewable energy sources in the contemporary world, wind turbine technology has grown in importance. Blades are among the most crucial parts of a modern horizontal-axis wind turbines. They extract dynamic energy from the wind and convert it to rotational mechanical energy for the turbine. Blades play a significant role in the safety, stability and control of the wind power plant. Blade pitch angles are controlled online via electrical or hydraulic actuators to safeguard the turbine from hazards of extreme wind conditions. The same actuation mechanisms are also active during power production for control purposes. Turbines operate with prespecified generated power references. In order to keep the production at this reference pitch angles are position-controlled with feedback from the power output. Conventionally, linear control methodologies are applied. Recently, soft computing techniques and especially fuzzy logic controllers are applied in this field with promising success. The fuzzy rule base, the employed inputs and parameter values play important roles in the controller performance. This dissertation presents the design of a novel fuzzy logic blade pitch angle controller. Power regulation is carried out by this system which evaluates power error, rate of change of power error and generator speed. This set of inputs, different from the majority of the studies reported in the literature, creates flexibility in the design of fuzzy rules which compute pitch angle references to be applied to the blade actuators. Tuning the many parameters of the three-dimensional rule base, however, proves to be an elaborate task. Evolutionary computing is applied in this thesis for the tuning of these parameters. The controller is tested with dynamic simulations of a 2 MW wind turbine model under fluctuating wind profiles and over nominal wind speeds. The performance of the novel controller is contrasted to a number of traditional pitch angle control techniques. Also tested are these conventional techniques when they are tuned by genetic algorithms. Simulation studies and data from the literature indicate superior performance of the proposed technique. An energy production improvement of 1.1 % is achieved when compared with conventional pitch control technique.