Research On The Technology Of Pitch Control For Direct-Driven Permanent Magnet Wind Turbine

Nowadays, the whole society pays more attention to the environment. For development and utilization of a variety of renewable pollution-free and low cost new energy, especially for the use of wind power, gradually in attention by all of us. The rational utilization and development of wind power resources have a significant effect in some respects of improve the ecological environment, optimize the energy structure, improve social benefits, and promote energy and sustainable development of the society. As the complexity of the wind power system, time-varying and strong interference in the operation process and other factors, how to control the wind power system accurately is still a difficult problem of the present study. Therefore, research on wind power technology, especially the control technology is of great significance to improve the wind power industry.This research topic is the variable pitch control technology of the large direct drive wind turbines. We tried some new methods on variable pitch control include the training platform and machine learning methods applied to wind turbines.According to the characteristics of wind turbine, we established the model of wind system, transmission system, actuator and generator. Through a combination of whole models we get a mathematical model that can reflect the dynamic behavior and is suitable for control requirement. Then, the sliding mode variable structure controller is established based on the mathematical model and the chattering problems are analyzed. After joining the quasi sliding mode control, this controller overcomes the bad effect caused by the chattering problems. Because of the support vector machine have strong robustness and global optimality in solving nonlinear problems, which can be used to optimize the sliding mode controller. To improve the dynamic performance in the operation region of constant power output, a Chattering Free sliding-mode variable structure pitch controller for wind turbine is proposed based on the analyses of system dynamic models, and a new support vector machine(SVM)-based reaching law for reducing the chattering caused by the sign function is presented. A pitch control system mode for the wind turbine system with permanent magnet synchronous generator(PMSG) is built and simulations are performed. The simulation results show that the proposed strategy has many advantages, such as strong ability for eliminating chattering, robust to the parameters variation and fast response.With the emergence of large wind turbines, unified variable pitch control method has already can’t satisfy its control requirements. So we should introduce the independent variable pitch control method for large wind turbines. The variable pitch control is an important component of wind turbine control system. With the megawatt wind turbine generator places in service and in order to ensure it safe, stable and efficient operation, research of variable pitch control technology is of great significance. In order to relieve the aerodynamic load imbalance problem in the operation process of wind turbine, this paper proposes a method of reinforcement learning algorithm based on RBF neural network to weights dynamic compensation control based on the analysis of wind turbine aerodynamic characteristics and the principle of independent variable pitch control. In order to verify its effectiveness, we build the joint simulation model based on the wind power dynamics software FAST and Matlab/Simulink. The simulation results show that the proposed method can reduce the load on the blade and the vibration of the blade flap, improve the efficiency of the wind turbine, on the premise of guarantee the stability of the output power.