基于径向基神经网络的双馈风力发电机低电压穿越控制研究

根据风力机能量转化机理及风电机组运行状态,建立了双馈感应发电机(DFIG)完整的5阶数学模型,分析了其电流控制方案,提出了一种基于径向基(RBF)神经网络辨识的PI控制器自适应控制算法.利用RBF神经网络进行在线辨识,并根据被控对象的Jacobian信息在线调整PI控制器参数,以改善系统的动态响应特性和提高系统的低电压穿越(LVRT)能力.通过构建系统的Simulink仿真模型进行仿真.结果表明,该控制算法有效地抑制了由电压跌落引起的电流震荡,缩短了系统的故障恢复时间,增加了系统的自适应性和鲁棒性,从而提高了系统的低电压穿越(LVRT)能力.

Research on the Low Voltage Ride-through of Doubly-fed Induction Generator Based on RBF Neural Network

According to energy conversion mechanism and the operation state of wind turbine, the 5th order mathematical model of the doubly-fed induction generator （DFIG） is established and its current control strategy is analyzed. In order to improve the dynamic re- sponse characteristics and the low voltage ride-through （LVRT） capability of system, the algorithm of adaptive PI controller based on RBF neural network identification is proposed, which can realize the on-line identification of system by using RBF neural network, and the PI parameters are adjusted online according to the Jacobian information of controlled object. Building a simulation model of system in Simulink environment, the simulation results show that this control algorithm can effectively restrain the current oscillation caused by the voltage sag, shorten the fault recovery time of system, increase the adaptability and robustness of system, and therefore improve the low voltage ride-through capability of system.