Modelling and Control of a Wind Turbine Julian Genov, Gancho Venkov, Bogdan Gilev 1. Introduction. The cost of electricity from wind turbines (WT) depends essentially on the initial investment and very expensive maintenance. Therefore keeping the operating mode and reducing dynamic loads leading to accidents is essential to reduce the price of produced electricity. Wind turbines control is done through changing the electrical characteristics of the generator and changing the pitch angle. Also wind turbines are equipped with systems for regulating the orientation angles of the axis of the turbine against the wind and against the skyline. Control systems based on the PID controllers are applied for power control to several large WT [12]. LQR controllers and state space observer to control wind turbine power are used, by changing the angle of attack of blades [7,13]. The dynamic model of the WT has significant nonlinear characteristics, usually associated with aerodynamic interactions. Thus controller designed for one operating point of the turbine may be ineffective and may even have worse performance in others. The design of multiple controllers for different working points is a possible solution [4] when controllers are switched with the change of the respective conditions. This is associated by undesirable transient response and therefore is used approach for a smooth transition [2]. There are also used adaptive controllers [6], predictive controllers [5] and those based on the use of neural networks, fuzzy logic and genetic algorithms [2.11]. 2. Dynamic model of wind generator. Scheme of the constituting the generator is shown in Figure 1. 1. Air flow measurement 2. Hydraulic braker 3. Fundamental 4. Generator 5. Control panel 6. Aviation beacon 7. Dampers 8. Coupling 9. Front cone 10. Gear box 11. Oil cooler 12. Actuator 13. Rotor shaft 14. Bearing 15. Drive hub 16. Rotor hub 17. Ventilation 18. Heat exchanger Figure1. Structure of wind turbine 2.1 Model of interaction between the wind and wind turbine. - Modelling the wind speed. The wind speed is modelled as the sum of the following components: (1) , () () () () a r g taw t V (t) v v t v t v t v t ∞ = + + + + where: – average initial speed of the wind; a v (t) v r - component of linear increase in amplitude A r ; (t) v g - component of the wind gust speed, (under extreme conditions it is modelled as a stochastic component, depending on regional weather); (t) v taw - component takes into account aerodynamic "shadow" of a tower. It causes harmonic components of the wind speed with frequencies multiple of number of blades. In 3-blade turbine the component with frequency three times greater of the turbine speed is essential; (t) v t - turbulent component due to roughness of the blades and wrapping processes. It is seen as a random process with spectral density which is determined by several methods: Relation (1) is approximated as normally distributed random signal (white noise) which is passed through a Kaimal filter [10]. After this filter are included harmonic filters of second order, which increase the influence of harmonic components multiple to 3 (for 3- blade generators). 1 wind f(u) sin fun f(u) cos fun ZERO_Harmonic_Filter 1/z 1/z Kaimal_Filter_3 Kaimal_Filter_2 Kaimal_Filter_1 [wind] 2*pi/60 Gain [wind] [wind] [wind] Clock wn White Noise wn White Noise wn White Noise 3rd_Harmonic_Filter2 3rd_Harmonic_Filter1 2 omg_wt 1 avg_wind Figure 2. Program module in MATLAB-Simulink to generate the wind speed through the Kaimal filter [10] Structural diagram that realizes the wind speed according to (1) is shown in Figure 2. Figure 2. Structural diagram of wind speed mode Model of the aerodynamic interaction between wind and the blades