|
摘要: |
为提高发电机机端电压和转速的综合控制性能,设计了附加神经网络电力系统稳定器(NNPSS)的在线学习神经网络逆(OLANNI)励磁控制器。针对多机系统同步发电机组模型,根据逆系统方法得到发电机励磁系统的逆系统的表达形式,并通过离线训练得到发电机励磁系统的神经网络逆系统。借鉴传统的AVR/PSS控制方法,并考虑到其对电力系统不确定性的自适应能力的不足,在离线训练的基础上分别设计了自适应的OLANNI、NNPSS以取代传统的AVR、PSS,给出了基于在线梯度算法的OLANNI和NNPSS的在线学习算法,并根据Lyapunov稳定性理论证明了OLANNI和NNPSS在线学习的收敛性。将设计的控制器应用于一个典型的2区域4机系统,仿真研究结果表明:在系统遭受扰动时,所设计的控制器较AVR/PSS和OLANNI控制器具有更好的综合控制性能。 |
关键词: 神经网络 逆系统 在线学习 励磁控制 神经网络电力系统稳定器 |
DOI: |
分类号:TM571 |
基金项目:国家自然科学基金资助项目(60574097)~~ |
|
Online learning ANN-inversion excitation controller with NNPSS for multi-machine power system |
XU Qinghong DAI Xianzhong
|
Abstract: |
To improve the comprehensive control performance of generator terminal voltage and rotor speed,an OLANNI(Online Learning ANN-Inversion) excitation controller supplemented with NNPSS(Neural Network Power System Stabilizer) is designed.For the synchronous generator set model of multi-machine power system,the inverse system expression of excitation system is deduced using inverse system method,and its ANN-inversion system is derived by offline training.Motivated by the conventional AVR /PSS control scheme and ... |
Key words: neural network inverse system online learning excitation control NNPSS |