摘要
针对道岔转辙机转换油压信号难以提取有效的故障特征且传统的故障诊断方式效果差的问题,本文提出一种基于可视图特征与CatBoost的道岔转辙机故障诊断方法。首先,采用可视图算法将时域信号转换为复杂网络图。然后,提取复杂网络图的5种统计特征,即网络平均度、全局聚类系数、平均路径长度、传递性特征与网络图密度。最后,通过CatBoost算法实现道岔转辙机故障诊断。将该方法分别与其它特征提取方法、故障分类算法进行比较。实验结果表明:可视图特征能够更有效表征道岔转辙机的工作状态,CatBoost算法对道岔转辙机4种工作状态的诊断准确率达到97.5%,验证了该方法的有效性和优越性。
Aiming at the problem that it is difficult to extract effective fault features from oil pressure sig‐nals converted by turnout switch machine and the traditional fault diagnosis method is not effective,a fault diagnosis method turnout switch machine based on visibility graph features and CatBoost is pro‐posed.Firstly,the visual graph algorithm is used to convert the time-domain signal into a complex net‐work graph.Then,the five statistical features of complex network graphs are extracted(i.e.,network average degree,global clustering coefficient,average path length,transitivity feature and network graph density).Finally,the fault diagnosis of turnout switch machine is realized by CatBoost algorithm.This method is compared with other feature extraction methods and fault classification algorithms.The experi‐mental results show that the visibility graph feature can more effectively represent the working state of turnout switch machine.The diagnostic accuracy of CatBoost algorithm for four working states of turnout switch machine reaches 97.5%,which verifies its effectiveness and superiority.
作者
杨喜旺
王智超
黄晋英
荆丽澄
王宇轩
YANG Xiwang;WANG Zhichao;HUANG Jinying;JING Licheng;WANG Yuxuan(School of Information and Communication Engineering,Shanxi Institute of Electronic Science and Technology,Linfen 041000,China;Data Science and Technology,North University of China,Taiyuan 030051,China;School of Mechanical Engineering,North University of China,Taiyuan 030051,China)
出处
《中北大学学报（自然科学版）》
CAS
2024年第1期58-65,共8页
Journal of North University of China(Natural Science Edition)
基金
山西省回国留学人员科研教研资助项目(2022-141)
山西省基础研究计划资助项目(202203021211096)。
作者简介
杨喜旺(1970-),男,副教授,博士,主要从事大数据算法与机械故障诊断方面的研究。;通信作者:黄晋英(1971-),女,教授,博士,主要从事结构动力学分析、检测与控制、故障诊断与健康管理的研究。E-mail:jyhuang@nuc.edu.cn。
