Power and Speed Control of Wind Turbines using Rotor Speed Estimates

Fig. 6 Pitch angles In this work a generic 4.8MW wind turbine model is used. Details of the model can be seen in [1]. This model includes a number of faults which is deactivated in this work. The model consists of: a dynamic drive train model, dynamic generator & converter model, a dynamic pitch system model and a static aerodynamic model, in addition measurement noise is used to model the sensors. In the model the wind turbine is controlled by a two mode controller, which uses a fixed pitch angle at power optimization (partial load) and uses a PI controller to control the pitch angle during full load control. This controller has the measured generator speed as input. An overview of this model is given in Fig.1. Standard industrial solutions for power & speed control of wind turbines are based on measurements of the generator speed, even though it is actually the rotor speed which is the true control objective. The reason is that the measurement of the rotor speed is highly influenced by measurement noises. In this work it is proposed to estimate the rotor speed based on a number of measurements from the wind turbine. The proposed scheme uses a Kalman estimator based on a model of the wind turbine and measurements from the wind turbine. Rotor and generator speed estimations have previously been used in more advanced state space based control schemes. This will, however, require an entire redesign of the wind turbine controller. Instead this work proposes to use the rotor speed estimate as input to standard PID controller solution, and thereby acts directly on the variable it should control and not a low pass filtered version of it, as the generator speed is. The proposed scheme is applied to a wind turbine model for test in order to illustrate the potential of the proposed scheme. This test shows that proposed scheme achieves similar power and speed control with less and more smooth pitch actuation.