用于双馈风电场次同步振荡抑制的RBF神经网络控制器设计

摘要

针对含双馈机组DFIG(Doubly-Fed Induction Generator)风电场经串联补偿线路并网引起的次同步振荡问题,运用基于阻抗的奈奎斯特稳定判据进行筛选分析,验证风速、串联补偿度对风电场次同步振荡的影响.利用RBF(Radial Basis Function Neural Network)神经网络自学习能力在线调整PID参数,设计了用于次同步振荡抑制的RBF神经网络控制器.为验证控制器抑制效果,在Matlab平台上编程搭建了系统仿真模型,将传统PI控制和RBF神经网络控制效果进行对比,结果表明RBF神经网络控制器对次同步振荡具有良好的抑制效果.%Considering synchronous oscillation caused by DFIG wind energy system interconnected with a series compensated electric network, impedance-based Nyquist stability criterion was applied to analyze and verify the influence of wind farm's sub-synchronous oscillation caused by wind velocity, series compensation level. The ability of self-study of RBF was applied to adjust PID parameters online, and the RBF neural network controller was designed to suppress the sub-synchronous oscillation. For the purpose of verifying the suppress effect of the RBF neural network controller, the results of the simulation model built on the MATLAB platform, show that has a good effect on Sub-synchronous oscillation suppression compared to the traditional PI controller.