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  

Design of RBF neural network controller for depressing sub-synchronous oscillation of DFIG-based wind farms

作　　者：单馨 王宝华[1,2] 仲振 

机构地区：[1]南京理工大学自动化学院,江苏南京210094 [2]江苏风力发电工程技术中心,江苏南京210023 [3]南京供电公司,江苏南京210008

出　　处：《可再生能源》2017年第10期1506-1513,共8页Renewable Energy Resources

基　　金：江苏省风力发电工程技术中心开放基金资助项目(ZK16-03-02)

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 subsynchronous 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.

关 键 词：双馈机组 次同步振荡 奈奎斯特稳定判据 RBF神经网络 

分 类 号：TK514[动力工程及工程热物理—热能工程]