[1]雷 伟,廖光忠,裴 浪.基于改进 DenseNet 模型的滚动轴承故障诊断[J].计算机技术与发展,2024,34(03):207-213.[doi:10. 3969 / j. issn. 1673-629X. 2024. 03. 030]
　LEI Wei,LIAO Guang-zhong,PEI Lang.Fault Diagnosis of Rolling Bearing Based on Improved DenseNet Model[J].,2024,34(03):207-213.[doi:10. 3969 / j. issn. 1673-629X. 2024. 03. 030]
点击复制

基于改进 DenseNet 模型的滚动轴承故障诊断()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
34
期数:
2024年03期
页码:
207-213
栏目:
新型计算应用系统
出版日期:
2024-03-10

文章信息/Info

Title:
Fault Diagnosis of Rolling Bearing Based on Improved DenseNet Model
文章编号:
1673-629X(2024)03-0207-07
作者:
雷 伟1 ; 廖光忠2 ; 裴 浪3
1. 武汉科技大学 计算机科学与技术学院,湖北 武汉 430065;
2. 武汉科技大学 智能信息处理与实时工业系统湖北省重点实验室,湖北 武汉 430065;
3. 武汉晴川学院 计算机科学与技术学院,湖北 武汉 430065
Author(s):
LEI Wei1 ; LIAO Guang-zhong2 ; PEI Lang3
1. School of Computer Science and Technology,Wuhan University of Science and Technology,Wuhan 430065,China;
2. Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System,Wuhan University of Science and Technology,Wuhan 430065,China;
3. School of Computer Science and Technology,Wuhan Qingchuan College,Wuhan 430065,China
关键词:
滚动轴承; 变分模态分解; 多种群差分进化; 密集神经网络; MECANet; 故障诊断
Keywords:
rolling bearing; variational mode decomposition; multi- population differential evolution; dense neural network; MECANet; fault diagnosis
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2024. 03. 030
摘要:
滚动轴承是机械设备的关键部件,为了检测滚动轴承设备的正常运转并且提高识别轴承故障的准确率,提出一种优化变分模态分解( VMD) 结合改进密集神经网络( DenseNet)的故障诊断模型方法。 首先,使用多种群差分进化( MPDE)算法以局部极小包络熵为优化搜索的目标函数,对 VMD 方法中的相关参数进行优化搜索以获取最佳参数组合;然后,使用最佳参数组合优化的 VMD 方法分解处理原始滚动轴承的故障信号,并得到若干本征模态分量信号( IMFs) ;最后,通过引入通道注意力模块( MECANet) 的改进密集神经网络模型对分解得到的 IMF 分量信号进行深层故障特征提取与识别,最终完成滚动轴承的故障诊断。 实验结果表明:提出的优化 VMD 结合改进 DenseNet 模型对滚动轴承故障识别的准确率达到了 99. 23% ,并且对比一些其他常见故障诊断模型的准确率有明显的提升,而且与先进的故障诊断模型对比其准确率存在较小差距,验证了此模型在滚动轴承故障诊断方面的有效性。
Abstract:
Rolling bearing is a key component of mechanical equipment. In order to detect the normal operation of rolling bearingequipment and improve the accuracy of identifying bearing faults,a fault diagnosis model method based on optimized variational modedecomposition ( VMD) and improved dense neural network ( DenseNet) is proposed. Firstly, multi - population differential evolution( MPDE) algorithm is used to search the parameters of VMD method with local minimum envelope entropy as the objective function toobtain the best parameter combination. Then,the original bearing fault signals are decomposed by using the VMD method of optimal parameter combination optimization,and several eigenmode component signals ( IMFs) are obtained. Finally,the improved dense neuralnetwork model with channel attention module ( MECANet) is introduced to extract and identify the deep fault features of the decomposedIMF component signals,and the fault identification of the rolling bearing is completed. The experimental results show that the accuracy ofthe optimized VMD combined with the improved DenseNet model for fault identification of rolling bearings has reached 99. 23% .Compared with some other common fault diagnosis models,the accuracy is significantly improved,and there is a small gap between it andthe advanced fault diagnosis model,which verifies the effectiveness of this model in fault diagnosis of rolling bearings.

相似文献/References:

[1]杨摇 洁,岳美君,曾耀平.基于 VMD 和熵特征的雷达辐射源信号识别[J].计算机技术与发展,2021,31(10):13.[doi:10. 3969 / j. issn. 1673-629X. 2021. 10. 003]
　YANG Jie,YUE Mei-jun,ZENG Yao-ping.Radar Emitter Signal Recognition Based on VMD and Entropy Features[J].,2021,31(03):13.[doi:10. 3969 / j. issn. 1673-629X. 2021. 10. 003]
[2]刘康宁,徐遵义*,李 晨,等.基于 NLMS 和 Autoformer 的滚动轴承 RUL 预测[J].计算机技术与发展,2024,34(03):177.[doi:10. 3969 / j. issn. 1673-629X. 2024. 03. 026]
　LIU Kang-ning,XU Zun-yi*,LI Chen,et al.RUL Prediction of Rolling Bearing Based on NLMS and Autoformer[J].,2024,34(03):177.[doi:10. 3969 / j. issn. 1673-629X. 2024. 03. 026]

更新日期/Last Update: 2024-03-10