WaveFA-Net: Bearing Fault Diagnosis Model Based on Multi-Domain Blind Deconvolution and Frequency Domain Attention Mechanism
ID:131 Submission ID:58 View Protection:ATTENDEE Updated Time:2024-10-23 10:02:34 Hits:27 Poster Presentation

Start Time:Pending (Asia/Shanghai)

Duration:Pending

Session:[No Session] » [No Session Block]

No files

Abstract
轴承是关键的机械设备部件,会影响系统的可靠性、安全性和效率。传统的故障检测系统依靠传感器来收集振动信号,但在处理多尺度、非线性和高噪声数据时遇到困难,影响了实时性能和精度。该文提出了一种创新方法WaveFA-Net,这是一种集成了多分辨率分解、频率注意力、i域盲反卷积以及用于轴承故障检测的多头和多尺度注意力机制的深度神经网络。 首先,将数据采用离散小波变换(DWT)进行多分辨率分解,提取和增强特征,并利用频率注意力、多域盲反卷积重要特征分量进行筛选此外,网络骨干和颈部模块引入多头和多尺度注意力机制强化信息的全局依赖性和多尺度性质。 最后,采用物理启发的损失函数、数据增强和超参数优化策略 保证鲁棒性和稳定性。实验结果表明在不同信噪比(SNR)条件下,所提方法复杂环境中具有较强的潜力。
 
Keywords
Bearing fault diagnosis,discrete wavelet transform,attention mechanism,data augmentation,hyperparameter optimization
Speaker
MaNing
student Southwest Jiaotong University;Hong Kong Polytechnic University Shenzhen Research Institute

Submission Author
MaNing Southwest Jiaotong University;Hong Kong Polytechnic University Shenzhen Research Institute
LiuWenqiang Hong Kong Polytechnic University Shenzhen Research Institute;Chinese National Rail Transit Electnification and Automation Engineering Technology Researçh Centre
YangHaonan Southwest Jiaotong University
ShiLinjun Southwest Jiaotong University;Hong Kong Polytechnic University Shenzhen Research Institute
LiuZhigang Southwest Jiaotong University
ShanXiangyu Southwest Jiaotong University
Comment submit
Verification code Change another
All comments

Important Dates

15th August 2024   31st August 2024- Manuscript Submission

15th September 2024 - Acceptance Notification

1st October 2024 - Camera Ready Submission

1st October 2024  – Early Bird Registration

 

Contact Us

Website:

https://icsmd2024.aconf.org/

Email:
icsmd2024@163.com