中国电力 ›› 2016, Vol. 49 ›› Issue (8): 12-16.DOI: 10.11930/j.issn.1004-9649.2016.08.012.05

• 电网 • 上一篇    下一篇

基于BP网络算法优化粗糙-Petri网的电网故障诊断

高正中1,龚群英1,刘隆吉2,李世光1,赵丽娜1   

  1. 1. 山东科技大学 电气与自动化工程学院,山东 青岛 266590;
    2. 青岛港湾职业技术学院 电气工程系,山东 青岛 266404
  • 收稿日期:2016-03-16 出版日期:2016-08-10 发布日期:2016-08-12
  • 通讯作者: 中国博士后基金资助项目(2015T80729);青岛市博士后研究人员应用研究项目(2015190)
  • 作者简介:高正中(1971—),男,山东济宁人,副教授,硕士生导师,从事检测技术与自动化装置、电力系统及其自动化研究。

Power System Fault Diagnosis Based on Rough Set and Petri Network Optimized by BP Algorithm

GAO Zhengzhong1, GONG Qunying1, LIU Longji2, LI Shiguang1, ZHAO Lina1   

  1. 1. College of Electrical Engineering and Automation, Shandong University of Science and Technology, Qingdao 266590, China;
    2. Department of Electrical Engineering, Qingdao Harbour Vocational and Technical College, Qingdao 266404, China
  • Received:2016-03-16 Online:2016-08-10 Published:2016-08-12
  • Contact: This work is supported by China Postdoctoral Foundation(No.2015T80729), Applied Research Projects by Postdoctoral Researchers in Qingdao (No. 2015190).

摘要: 为了使电网故障诊断的过程更简洁、快速和直观,提出了一种基于BP网络算法优化粗糙-Petri网的电网故障诊断方法。首先用粗糙集理论对电网的故障征兆数据进行处理,从冗余的故障信息中约简出最小决策表;然后基于得到的最小决策表提取诊断规则并建立最优的 Petri 网模型,利用 Petri 网处理并行推理的能力来实现高效的电网故障诊断。其中引入神经网络中的BP算法对电网故障诊断Petri 网模型的权值参数进行网络优化训练。电网实例分析结果表明该模型能准确找到故障区域,具有较好的快速性、自适应性和一定的容错性。

关键词: 电网, Petri网, 粗糙集, 属性约简, BP网络, 故障诊断

Abstract: In order to improve fault diagnosis process, a power grid fault diagnosis method is proposed based on rough sets and Petri network combined with Back-Propagation algorithm. Firstly, the rough set theory is applied to grid fault symptoms for deduction of minimum decision table from the redundant fault information. Then, diagnosis rules are extracted and optimal Petri net model is created. Petri net’s parallel processing ability is used to implement efficient fault diagnosis. BP algorithm is adopted to train the weight values of Petri network parameters. The results of grid instance show that the proposed model can find fault zone accurately, and has good adaptability, speediness and fault tolerance.

Key words: electric power grid, Petri nets, rough sets, attribute reduction, back propagation, fault diagnosis

中图分类号: