Issue |
JNWPU
Volume 41, Number 2, April 2023
|
|
---|---|---|
Page(s) | 344 - 353 | |
DOI | https://doi.org/10.1051/jnwpu/20234120344 | |
Published online | 07 June 2023 |
A novel bearing fault diagnosis method based on multi-scale transfer symbolic dynamic entropy and support vector machine
基于多尺度迁移符号动力学熵和支持向量机的轴承诊断方法研究
School of Mechatronic Engineering, Xi'an Technological University, Xi'an 710021, China
Received:
27
October
2021
In view of the problem that the generalization ability of traditional data-driven fault diagnosis model declines or even fails in mechanical system diagnosis, a fault diagnosis method based on multi-scale transfer symbolic dynamic entropy and support vector machine is proposed based on the idea of transfer learning. Firstly, multi-scale symbolic dynamic entropy is used to extract fault features from measured vibration signals. And then a feature projection technique based on transfer learning is proposed, which reduces the data distribution difference. Secondly, the parameters of the multi-scale transfer symbol dynamic entropy method are optimized to improve the final fault identification rate. Then, the support vector machine can implement the fault identification. Finally, through the test of bearing fault experimental signals, the rolling bearing diagnosis method based on multi-scale transfer symbol dynamic entropy can effectively improve the generalization ability of data-driven model and realize accurate identification of different fault types of rolling bearing under a small number of samples.
摘要
针对传统数据驱动故障诊断模型在机械系统诊断中存在的泛化能力下降甚至失效的问题, 应用迁移学习的思想, 提出了基于多尺度迁移符号动力学熵和支持向量机的故障识别算法。采用多尺度符号动力学熵提取故障特征, 在此基础上提出基于迁移学习的特征映射技术, 使非同分布数据的特征在映射后分布差异减小。对多尺度迁移符号动力学熵方法的参数进行优选, 将其输入支持向量机中, 进一步提高最终的故障识别率。通过轴承故障实验信号的测试证明, 基于多尺度迁移符号动力学熵的滚动轴承诊断方法能够有效提升数据驱动故障诊断模型的泛化能力, 实现少量样本下滚动轴承不同故障位置的准确识别。
Key words: symbolic dynamic entropy / feature extraction / transfer learning / rolling bearing / fault diagnosis
关键字 : 多尺度迁移符号动力学熵 / 特征提取 / 迁移学习 / 故障诊断 / 滚动轴承
© 2023 Journal of Northwestern Polytechnical University. All rights reserved.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.