Bearing fault diagnosis method based on SAVMD and CNN

SONG ChunSheng; LIANG YaRu; LU NiFang; DU Gang; JIA Bo

doi:10.16579/j.issn.1001.9669.2024.03.001

您当前的位置：

首页 >

文章列表页 >

Bearing fault diagnosis method based on SAVMD and CNN

更新时间：2024-07-05

- Bearing fault diagnosis method based on SAVMD and CNN
- Journal of Mechanical Strength Vol. 46, Issue 3, Pages: 509-517(2024)
- 作者机构：
  
  1. 武汉理工大学机电工程学院
- 作者简介：
- 基金信息：
- DOI：10.16579/j.issn.1001.9669.2024.03.001
  CLC： TH133.3
- Published：15 June 2024，
- 稿件说明：
移动端阅览
SONG ChunSheng, LIANG YaRu, LU NiFang, et al. Bearing fault diagnosis method based on SAVMD and CNN. [J]. Journal of Mechanical Strength , 2024,46(3):509-517.
DOI：

SONG ChunSheng, LIANG YaRu, LU NiFang, et al. Bearing fault diagnosis method based on SAVMD and CNN. [J]. Journal of Mechanical Strength , 2024,46(3):509-517. DOI： 10.16579/j.issn.1001.9669.2024.03.001.

摘要

由于长期受到复杂环境和设备的噪声信号干扰，旋转机械中轴承的早期不稳定、非线性的故障信号被掩盖，无法准确捕捉原始振动信号的故障特征，正确识别故障类型。针对现在故障状态的缺点，通过研究变分模态分解（Variational Mode Decomposition

VMD）法中的参数[

]

，使用物理系统退火的方式来选取全局优解，提出了模拟退火变分模态分解（Simulated Annealing Variational Mode Decomposition

SAVMD）的特征提取方法。用加权峭度评价指标重构本征模态函数（Intrinsic Mode Function

IMF）分量，将重构信号输入到卷积神经网络（Convolutional Neural Network

CNN）模型进行故障分类，再通过美国凯斯西储大学公开的轴承数据集验证所提方法，诊断准确率达99.28%。引入-6 dB的白噪声模拟高斯噪声环境进行验证，准确率达93.6%，结果表明能够在复杂干扰工况下实现故障轴承的诊断分类。

Abstract

Non-stationary and nonlinear vibrating signals in rotating mechanical equipment are always disturbed by noises from working environment and other devices. Bearing fault signals in mechanical systems are reduced which makes it difficult to extract fault characteristic

diagnose fault type. Selection of global optimal solution was solved by studying the parameter [

]

about variational mode decomposition (VMD) and simulating physical system annealing， that was simulated annealing variational mode decompos

ition (SAVMD) extract characteristic value. The intrinsic mode function (IMF) components were reconstructed by weighted kurtosis evaluation index

put them into convolutional neural network (CNN) model for fault classification. Finally

the proposed method was verified by numerical simulations with the open bearing data of Case Western Reserve University. The accuracy rate is 99.28%. After introducing -6 dB white noise to simulate the noisy environment

the accuracy rate is 93.6%. These results show that this method can be used to achieve bearing fault diagnosis in complex working environment.

关键词

轴承故障变分模态分解模拟退火特征提取故障诊断

Keywords

Bearing faultVariational mode decompositionSimulated annealingFeature extractionFault diagnosis

references

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

ROLLING BEARING FAULT DIAGNOSIS BASED ON ADAPTIVE FEATURE SELECTION k-SUB CONVEX HULL

RESEARCH ON ROLLING BEARING FAULT DIAGNOSIS BASED ON IMPROVED HHO-LSTM

改进MOMEDA方法及其在滚动轴承故障特征增强中的应用

基于XGBoost的冷水机组不平衡数据故障诊断

FAULT DIAGNOSIS OF RECIPROCATING COMPRESSOR ON THE RESONANCE-BASED SPARSE SIGNAL DECOMPOSITION WITH OPTIMAL Q-FACTOR

Related Author

CHENG JunSheng

YANG Yu

WU ZhanTao

HU AiRu

ZHU SiJia

SHAO LiangShan

程尧

张卫华

Related Institution

College of Mechanical and Vehicle Engineering, Hunan University

Hunan Provincial Key Laboratory of Equipment Service Quality Assurance, Hunan University

System Engineering Institute,Liaoning Technical University

School of Business Administration,Liaoning Technical University

西南交通大学牵引动力国家重点实验室

⁰