基于LCD和改进SVM的轴承故障诊断方法

针对滚动轴承极易损伤,振动信号表现出非线性、非平稳性等特点,提出一种基于局部特征尺度分解（LCD）和改进支持向量机（SVM）的滚动轴承故障诊断算法。首先对采集到的轴承振动信号进行LCD,分解得到一系列内禀尺度分量（ISC）,通过与经验模态分解（EMD）对比研究,证明了LCD方法的优越性;然后计算所有分量的能量熵值,提取出轴承信号的敏感特征集,输入到经过遗传算法（GA）进行参数优选后的SVM识别模型进行轴承状态的诊断识别。实验研究表明,基于LCD和改进SVM的轴承诊断算法能较好地提取出轴承故障特征信息,对4种轴承状态的识别率高达90%,是一种较为有效的轴承故障诊断方法。...

Full description

Saved in:
Bibliographic Details
Published in电子技术应用 Vol. 42; no. 6; pp. 81 - 83
Main Author 张美玲 胡晓
Format Journal Article
LanguageChinese
Published 中国矿业大学信息与电气工程学院,江苏徐州,221116%中国矿业大学机电工程学院,江苏徐州,221116 2016
Subjects
局部特征尺度分解
支持向量机
故障诊断
滚动轴承
遗传算法
genetic algorithm
支持向量机
滚动轴承
遗传算法
fault diagnosis
support vector machine
故障诊断
rolling bearing
局部特征尺度分解
local characteristic-scale decomposition
Online AccessGet full text
ISSN0258-7998
DOI10.16157/j.issn.0258-7998.2016.06.022

Cover

More Information
Summary:针对滚动轴承极易损伤,振动信号表现出非线性、非平稳性等特点,提出一种基于局部特征尺度分解（LCD）和改进支持向量机（SVM）的滚动轴承故障诊断算法。首先对采集到的轴承振动信号进行LCD,分解得到一系列内禀尺度分量（ISC）,通过与经验模态分解（EMD）对比研究,证明了LCD方法的优越性;然后计算所有分量的能量熵值,提取出轴承信号的敏感特征集,输入到经过遗传算法（GA）进行参数优选后的SVM识别模型进行轴承状态的诊断识别。实验研究表明,基于LCD和改进SVM的轴承诊断算法能较好地提取出轴承故障特征信息,对4种轴承状态的识别率高达90%,是一种较为有效的轴承故障诊断方法。
Bibliography:Zhang Meiling, Hu Xiao （1 .College of Information and Electrical Engineering, China University of Mining and Technology, Xuzhou 221116, China; 2.College of Mechanical and Electrical Engineering, China University of Mining and Technology, Xuzhou 221116, China）
rolling bearing; local characteristic-scale decomposition; genetic algorithm; support vector machine; fault diagnosis
11-2305/TN
According to the characteristics of nonlinear and non- stationary of rolling bearing vibration signals, a new method of bearing fault diagnosis based on local characteristic-scale decomposition（LCD） and improved support vector machine（SVM） is proposed.Firstly, bearing vibration signals were decomposed by LCD, and a series of intrinsic scale components（ ISC） were obtained. Through comparative study with empirical mode decomposition（EMD）, the superiority of the LCD method was proved. Then the energy entropy of all components were calculated and bearing signal sensitive feature sets were extracted, which were input to SVM after param
ISSN:0258-7998
DOI:10.16157/j.issn.0258-7998.2016.06.022