交通运输系统工程与信息 ›› 2013, Vol. 13 ›› Issue (6): 105-111.

• 决策论坛 • 上一篇    下一篇

改进的FP-tree算法在动车组故障诊断中的应用研究

钟雁*1,马海漫2,张春1,赵怀昕1   

  1. 1. 北京交通大学 城市交通复杂系统理论与技术教育部重点实验室,高速铁路网络管理教育部工程研究中心,北京 100044;2. 天津铁路职业技术学院,天津 300240
  • 收稿日期:2013-05-22 修回日期:2013-08-08 出版日期:2013-12-24 发布日期:2014-01-14
  • 作者简介:钟雁(1959-),男,广东梅县,教授.
  • 基金资助:

    国家高技术研究发展计划(863计划)项目(2012AA040912).

Application Research on Improved FP-tree Algorithm in EMU’s Fault Diagnosis

ZHONG Yan, MA Hai-man, ZHANG Chun, ZHAO Huai-xin   

  1. 1. Engineering Research Center of Network Management Technology for High Speed Railway MOE, MOE Key Laboratory for Urban Transportation Complex Systems Theory and Technology, Beijing Jiaotong University, Beijing 100044, China;2. Tianjin Railway Technical Vocational College, Tianin 300240, China
  • Received:2013-05-22 Revised:2013-08-08 Online:2013-12-24 Published:2014-01-14

摘要:

从利用动车组海量运维数据获取故障诊断知识的角度出发,通过对数据挖掘中关联规则——FP-tree算法的研究,结合动车组故障诊断和提高动车组运营安全的要求,在生成树和搜索频繁项两个阶段对FP-tree算法进行改进,提出了一种改进算法——整枝FP-tree算法.改进算法在生成树阶段将故障信息置于顶层,在搜索频繁项阶段将所有的搜索都搜索到树的顶层.最后,将整枝FP-tree算法应用到动车组故障信息和状态信息的关联规则的抽取中,通过对改进算法的具体分析以及实际测试,表明该算法输出结果满足要求,并且对故障诊断知识获取的时间消耗和空间消耗有较大的降低.

关键词: 智能交通, 故障诊断, FP-tree算法, 动车组, 关联规则

Abstract:

From the point view of acquiring fault diagnosis knowledge using EMU’s mass operation and maintenance data, through the study of FP-tree algorithm which is an association rule algorithm of data mining, combining with the requirement of EMU’s fault diagnosis, the FP-tree algorithm is improved from the aspect of spanning tree and searching for frequent items, and an improved algorithm is put forward that called NP-FP-tree algorithm. In the stage of spanning tree, EMU’s fault information is put in the top, in searching for frequent items, all searches are carried out to the top. Finally, the NP-FP-tree algorithm is applied to mining of association rule of EMU’s fault information and state information. Through carefully analyzing and testing the performance of NP-FP-tree algorithm, it proves that the improved algorithm is better than the original one in the aspect of time consumption and space consumption.

Key words: intelligent transportation, fault diagnosis, FP-tree algorithm, EMU, association rule

中图分类号: