Research On Intelligent Pitch Control And Optimization For Wind Turbines

As a kind of clean renewable energy, wind energy has gained worldwide attention in recent years. Wind power generation, which is the main utilization form of wind energy, has very broad development prospects because it is an environmentally friendly, economical and sustainable mode of new energy production. Generally, the large wind turbine units now all have pitch-adjustable blades, which can control the absorption of the wind by rotating the blades around their ordinate axes. In this way, the wind turbine units can change their running states and control the output power more actively. Therefore, researching the pitch control technology in depth is of great significance for guaranteeing the safe, stable and optimal running of the wind turbines.The instability of wind and the complex structure of the wind turbines make the pitch control objet has big disturbances, nonlinearity, multiple variables and complicated working states, making the conventional PI control strategy hard to achieve the control objective that utilizing the wind energy reasonably and providing stable, high-quality electric energy to the electricity grid. How to build suitable models and implement effective control becomes the focus of pitch control research. Therefore, in the paper, pitch control and optimization strategies for wind turbines are the main research contents, and the following researches are carried out:Firstly, the operating characteristics and typical working states of the pitch-adjustable wind turbines are studied firstly. Based on the pitch control tasks under different working states, the focuses of the paper are determined, which are generator speed control strategy when the wind turbines begin to integrate into the electricity grid and power control strategy when the wind speed is above rated wind speed.Secondly, in view of the limitations when use traditional mechanism modeling methods to build the wind turbine models, a neural network modeling method based on field data is studied. The generator speed model under lower wind speed and the output power model under higher wind speed are built respectively using the actual field data. Simulation results show the accuracy of the neural network models, so they can be used as the controlled plants models in the pitch control simulations.Thirdly, considering that the generator speed must be controlled to rise to the synchronous speed smoothly and precisely from its startup to grid integration, a novel neural network model predictive control strategy based on small-world optimization algorithm is proposed. The simulation results show that the system can forecast the change of generator rotational speed based on the wind speed disturbance, making the controller act ahead to eliminate the impact of system delay. And the system output can track the reference trajectory well, making sure that the system can integrate into the electricity grid steadily.Finally, for the constant power control when the wind speed is above the rated speed, a self-adaptive PI control strategy based on the small-world optimization algorithm on-line tuning is proposed. The strategy uses BP neural network for the on-line identification of controlled object, in order to provide accurate real-time information for the parameter tuning. Meanwhile, the small-world optimization algorithm is used to optimize the PI control parameters dynamically, seeking for the best control effect. Simulation results show that using this control strategy in the pitch control can effectively stabilize the output power and the control effects of the modified PI control is better than conventional PI control.