Research On Active Power Control Of Wind Farm Under Uncertainty

With the development of large-scale wind farms, the wind-farm-generated power with the characteristic of randomness and volatile significantly influences the active power balance of power grid, which in turn seriously affects the stability and security of power grid. As the increasing penetration of the large-scale wind farm in power grid, the power system operator now require that wind farms have the active power control capability by implementing the active power control system.The active power control system of wind farm generally has hierarchical structure with both a wind farm central control level and a wind turbine local control level. The performance of active power control system is affected by the active power control of each wind turbine local control level, the active power allocation and the active power control of the central wind farm level. Motivated by the ever-growing active power control requirements put on wind farm by power system operator, developing control method and allocation strategy for the active power control of wind farm has become the focus of ongoing research.Due to the nonlinearity, strong coupling and uncertainty of wind turbine, and the uncertain disturbance from the complex wind farm environment, the performance of active power control system of wind farm is degraded. Thus, this dissertation focuses on the research of active power control of wind farm under the uncertainty mentioned above, and the main contributions of this dissertation are as follows:1. A high gain observer based maximum power tracking control method is proposed for the wind turbine local control level. Firstly, a high gain observer is designed to estimate the aerodynamic torque, from which wind speed is deduced by using the Newton algorithm. With the estimated wind speed, a backstepping control based controller is then designed to achieve the maximum power point tracking. Based on Lyapunov stability theory, it is proved that the closed-loop control system is uniformly ultimately bounded with the tracking errors converging to the neighborhoods of the origin exponentially. The simulation results show that the proposed control method can effectively achieve maximum power point tracking without anemometer. 2. A nonlinear tracking differentiator based active power control method is proposed for wind turbine to track the power reference specified by the wind farm central control level. To avoid the adverse effects caused by power reference jump, a nonlinear tracking differentiator is designed to arrange a transient process for power reference. Then, a dynamic surface based adaptive controller is designed for the power tracking of wind turbine, and the parameter uncertainty caused by wind disturbance is compensated by an adaptive law, which is derived based on Lyapunov stability theory. The simulation results show that the proposed method can effectively reduce the adverse effects caused by power reference jump and wind disturbance.3. An improved GM(1,1) based ultra-short term active power prediction method is proposed to provide the available power of wind turbine for the wind farm active power allocation. Firstly, the weaken buffering operator is utilized to preprocess the historical wind speed data and a GM(1,1) rolling model is built to predict wind speed from the preprocessed data. With the predicted wind speed, the available active power is then obtained according to the wind power curve. The experiment results show that the proposed method could effectively estimate the available power of each wind turbine, which can be used to optimize the wind farm active power allocation.4. Considering the system uncertainty and wind disturbance, a sliding mode variable structure based active power control method is proposed for the wind farm central control level. First, the dynamic model of wind farm is established based on the analysis of the power tracking dynamic of the individual wind turbine. Secondly, by using the exponential reaching law method, the variable structure control law is designed for the controller in the wind farm central control level. Based on Lyapunov stability theory, it is proved that the power tracking error of wind farm will asymptotically converges to zero. The simulation results show that the proposed sliding mode controller provides good active power tracking performance and robustness against the system uncertainty and wind disturbance.