基于知识发现技术的电网安全分析新思路.PDFVIP

第 2 卷第 4 期/Vol. 2 No. 4 南方电网技术 交直流并联电网/ pp. 13-36 2008 年 8 月/Aug. 2008 SOUTHERN POWER SYSTEM TECHNOLOGY AC/DC Parallel Grid 文章编号：1674-0629(2008)04-0023-05 中图分类号：TM712; TP182 文献标识码：A 基于知识发现技术的电网安全分析新思路 王同文，管霖，张尧 （华南理工大学电力学院，广州 510640 ） A New Method for Analysis of Power System Security Based on Knowledge Discovery WANG Tong-wen, GUAN Lin, ZHANG Yao （College of Electric Power, South China University of Technology, Guangzhou 510640, China ） Abstract ：This article presents a new method for on-line analysis 稳态运行数据日益丰富和完善。另一方面，电网日 of power system security. The method with hierarchical structure 积月累的不断进行着大量离线、甚至在线的动态安 and based on knowledge discovery, uses steady state load flow 全仿真计算，这些计算条件和结果事实上也构成了 information as input. During off-line learning process, ker- 庞大的安全分析样本数据库。如何充分挖掘这些数 nel-feature subsets of the steady state quantities are obtained 据中蕴涵的安全分析知识，从海量数据中提取出显 through feature reduction process, and the power system secu- 式且易于被人理解的安全分析知识和规则是一项极 rity-related information is preserved as much as possible. The rules 有前景的研究工作，是消除电网运行中“数据海量、 about power gird security states and weak points are then extracted 知识匮乏”现象的关键。基于样本的知识发现技术 in the kernel feature spaces by using the knowledge discovery 为此提供了一种有价值的研究路线[3] 。 technique. With the rules and the steady state load flow quantities, 知识发现（Knowledge Discovery ）是从大量数 the information about power system security state and weak points 据中提取出可信的、新颖的、有效的并能被人理解 can be quickly obtained by the on-line reasoning process. 的模式的高级处理过程。目前，相关领域的研究非 常薄弱。多数该类型的电网安全评估算法运用人工 Key words ：power system security analysis; knowledge 神经网络[4]、模糊理论[5]、模式识别[6]、决策树[7]等 discovery; featur