基于RBF神经网络的电网脆弱性评估及其趋势估计  

Power grid vulnerability assessment based on RBF neural network and its trend estimation

作　　者：王耀升 张英敏[1] 王畅 漆万碧 Wang Yaosheng;Zhang Yingmin;Wang Chang;Qi Wanbi(School of Electrical Engineering and Information, Sichuan University, Chengdu 610065,China)

机构地区：[1]四川大学电气信息学院,成都610065

出　　处：《电测与仪表》2019年第9期49-55,共7页Electrical Measurement & Instrumentation

摘　　要：建立了分层网状拓扑结构下的电网脆弱性评价体系,针对该体系提出了基于径向基函数(Radial Basis Function,RBF)神经网络的电网脆弱性评估方法。将电网综合脆弱性分为状态脆弱性和结构脆弱性,并与相应的子指标构成脆弱性网状评价体系,同时以高斯(Gauss)函数作为RBF神经网络函数的核函数解决指标间的非线性问题。通过MATLAB中的RBF神经网络函数对IEEE14母线系统计算分析,验证了该方法的全面性与有效性。最后,针对节点多个测量周期的脆弱性测度建立自回归(Auto Regression,AR)模型,通过判定AR模型的差分方程稳定性,分析了节点脆弱性测度的发展趋势。This paper establishes the hierarchical network topology of network vulnerability evaluation system. The system was proposed based on radial basis function (RBF) neural network method of grid vulnerability assessment. The comprehensive vulnerability of the power grid is divided into the state vulnerability and structural vulnerability, and the corresponding sub-indexes constitute the vulnerability network evaluation system. Meanwhile, taking Gauss functions as the kernel function of RBF neural network function to solve nonlinear problem between the indicators. The calculation and analysis of IEEE14-bus system is carried out to verify the comprehensiveness and effectiveness of the method through using the RBF neural network function in MATLAB. Finally, the auto regressive (AR) model is established according to the multiple nodes of the measurement cycle vulnerability, the AR model is to determine the stability of difference equation and analyzes the development trend of node vulnerability measure.

关 键 词：电网脆弱性 非线性 脆弱性指标 神经网络 AR模型 趋势估计 

分 类 号：TM711[电气工程—电力系统及自动化]