The Wind Turbine Controller Design Based On Additive Decomposition Method

With growing concerns about environmental pollution and a possible energy shortage, Wind energy seems certain to play a major part in the world’s energy future. Steady development of wind turbine technology and the accumulation of wind farm operating experience have resulted in the expanding of wind power. Large wind turbine is a complicated multivariable system with strong nonlinearity. It is difficult to establish the precise mathematical model. PID controller or PI controller are generally applied in actual wind farm,whose control parameters are obtained relying on engineering experience but not theoretical instruction and lack of stability proof. So a method named additive decomposition method is proposed in this paper to design the wind turbine controller.Aerodynamic characteristics of wind turbines are introduced based on the momentum theory, blade element theory and momentum-blade element theory. Then establishing wind model, wind turbine model, transmission model, generator model and pitch actuator model. The simulation platform of wind power system has been set up.Considering from the point of engineering application, there are not standards and formal methods to design the controller parameters. This thesis adopts additive decomposition method to normalize the design parameter of wind turbine controllers. The control problem of wind power generating unit is decomposed into two simple sub-problems(main system and auxiliary system). For main system, the proposed controllers can be finally replaced by PI controllers based on additive decomposition. When the wind speed is lower than the rated wind speed, torque PI controller is designed to capture the maximum power by controlling the wind rotor speed. And the rotor speed is adjusted by generator torque. When the wind speed is higher than the rated wind speed, pitch PI controller is designed to limit the wind energy to stabilize the output power as rated power by adjusting the pitch actuator. For original system, the stability can be proved. Finally, the results are validated through Matlab/Simulink. The method proposed by this thesis will provide theoretical reference for PI controller designing.