作　　者： ; ;

机构地区： 华南理工大学电力学院

出　　处： 《中国电机工程学报》 1996年第6期384-387,共4页

摘　　要： 本文利用ＢＰ型神经网络逼近受控系统逆动态具体设计实现了一种神经网络电力系统稳定器（ＮＮＰＳＳ）。所设计的ＮＮＰＳＳ以四层ＢＰ网络为主体，采用非线性ＰＳＳ的控制响应构成样本进行训练并通过了单、多机仿真的检验。文章深入分析了ＮＮＰＳＳ的性能和特点，重点探讨了设计过程中存在的各种具体问题。ＮＮＰＳＳ装置采用分层结构实现。动模实验结果证实了ＮＮＰＳＳ应用于真实工业环境的可行性。所设计的ＮＮＰＳＳ装置显著改善了电力系统的动、静态稳定性能并为各种大、小扰动提供了有效的阻尼作用。 Designing and implementation of a neural network based power system stablizer (NNPSS) are presented in this paper. In this NNPSS,BP neural network is used to simulate the inverse dynamic of the power system. The NNPSS is trained by samples taken from a nonlinear PSS and tested by digital simulations in signle and multi machine power systems.The performance and features of the NNPSS are analyzed together with the problems faced in its design. The designed NNPSS has been implemented by a hierarchical strcutrue with CPU to simulate the formward inference of the neural network. Results of the laboratory test show the feasibility of using NNPSS in real power system. The designed NNPSS improves the steady state and dynamic stabilities of the power system and provides effective damping to a variety of disturbances.

关 键 词： 神经网络 电力系统 稳定器

领　　域： [电气工程]