《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (S2): 256-260.DOI: 10.11772/j.issn.1001-9081.2023020194

• 前沿与综合应用 • 上一篇    

用于损伤状态识别的极值延拓EMD和LS-SVM算法

刘聪(), 钱坤, 焦准, 丁奇   

  1. 空军工程大学 航空机务士官学校,河南 信阳 464000
  • 收稿日期:2023-03-02 修回日期:2023-09-20 接受日期:2023-09-28 发布日期:2024-01-09 出版日期:2023-12-31
  • 通讯作者: 刘聪
  • 作者简介:刘聪(1988—),男,江西九江人,讲师,博士,主要研究方向:非线性控制理论、非线性系统故障检测及健康管理
    钱坤(1977—),男,湖南常德人,教授,博士,主要研究方向:非线性控制理论、非线性系统故障检测及健康管理
    焦准(1983—),男,河南驻马店人,副教授,硕士,主要研究方向:非线性控制理论、非线性系统故障检测及健康管理
    丁奇(1982—),男,安徽黄山人,讲师,硕士,主要研究方向:非线性控制理论、非线性系统故障检测及健康管理。

Damage state identification algorithm based on extreme value extension empirical mode decomposition and least squares support vector machine

Cong LIU(), Kun QIAN, Zhun JIAO, Qi DING   

  1. School of Aeronautics Maintenance Non?Commissioned Officers,Air Force Engineering University,Xinyang Henan 464000,China
  • Received:2023-03-02 Revised:2023-09-20 Accepted:2023-09-28 Online:2024-01-09 Published:2023-12-31
  • Contact: Cong LIU

摘要:

针对机电系统损伤状态识别问题,提出一种基于极值延拓经验模态分解(EMD)和最小二乘支持向量机(LS-SVM)算法。首先,分析EMD算法的基本原理,针对端点效应利用多项式拟合极值延拓的算法改进设计方案,并利用标准化处理的特征向量设计程式;其次,考虑到机电系统损伤状态数据归属小样本特征,利用LS-SVM算法给出了状态识别的设计程式;最后,开展仿真验证实验。实验结果表明,采用所提算法的损伤状态识别方案,可以确保损伤状态识别的正确率超过96%,满足机电系统工程应用要求。

关键词: 极值延拓, 经验模态分解, 最小二乘支持向量机, 损伤状态, 状态识别

Abstract:

An algorithm based on Empirical Mode Decomposition (EMD) and Least Squares Support Vector Machines (LS-SVM) was proposed for the damage state identification of electromechanical systems. Firstly, the basic principle of EMD was analyzed, the algorithm of fitting extreme value extension was utilized to to tackle the end effect, and the design program of feature vectors was utilized for standardization. Then, considering the damage state data of the electromechanical system was attributed to the characteristics of the small samples, the design program for the state recognition was given by LS-SVM algorithm. Finally, the simulation experiment was carried out. The experimental results show that by using the proposed algorithm, the damage state identification probability is more than 96%, which can meet the application requirements of electromechanical system engineering.

Key words: extreme value extension, Empirical Mode Decomposition (EMD), Least Squares Support Vector Machine (LS-SVM), damage state, state identification

中图分类号: