Otimização de sistemas de conversão de energia eólica de grande porte.

Abstract

This work addresses the static and dynamic optimization of variable speed, variable pitch horizontal axis wind turbines, connected to the grid by an AC-DC-AC converter. The strategy used in the dynamic optimization depends on whether the output power, Pg, is greater than or smaller than the rated power, Px Whenever the turbine is working below rated power (V < Pr), the main goal is to maximize the wind energy capture. This can be achieved by imposing an optimal tip speed ratio \ Q , corresponding to the maximum power coefficient, through the regulation of the generator torque, by using Indirect Oriented Field techniques. On the other hand, when the power output is greater than the rated power (V> Vt), the main goal is to maintain Pg as close as possible to Pt, since there is an exceeding wind power to be neglected. In this case the power regulation of wind turbines can be made by the blade pitch control. In the present work, the power is controled, for V> Vr, by changing both blade pitch and yaw angle. The performance of this new controller is analyzed and compared with the performance of a conventional controller taking into account the forces due to yaw and gyroscopic effects. In the static optimization, the objective is to maximize the electric energy output of several configurations of wind turbines. Lysen (1984) analyses this problem considering rotor type, wind distribution and type of the load as variables. In the Lysen's model, the characteristics of the electric generator are constant. In the present model, the generator electric losses are considered as variables of the problem, what makes the model more general. Furthermore, the field and torque components for the optimal stator current are determined by Direct Oriented Field techniques based on the minimization of the generator losses. The main contribution of this work is related to the performance optimization of wind turbines operating above rated power. The new controller is designed to make the system operate close to the rated power by actuating on the blade pitch and on the yaw devices, simultaneously. Regarding the design of the controller, it is necessary to linearize the rotor torque function with respect to the rotor velocity, co, the pitch, 0, and the yaw angle, 8, around the steady state. Here, co is a state variable and G and 8 are the inputs. The direction and velocity of the wind are considered random variables. Then, the closed loop representation of the wind turbine can be approximated, in the frequency domain, by a second order function. The optimal performance can be achieved by imposing the best value of the damping coefficient. The optimal controller gain is also obtained by other two methods, i. e., one which uses the Riccati equation and the other which computes the minimum value of a cost function that takes into account the deviations from the reference signal and control efforts. It is proved that those three different ways of computing the optimal controller gain lead to close results. The simulation of the wind turbine is then made by a state space representation. An additional analysis is made by taking into account the cyclic forces and moments at the blade root caused by the yaw angle. They are computed by considering the azimuth angle, cp, the tip speed ratio, X, the pitch, 9, and the yaw angle, 8, as variables.