• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2018, Vol. 54 ›› Issue (7): 87-96.doi: 10.3901/JME.2018.07.087

• 机械动力学 • 上一篇    下一篇

基于改进堆叠降噪自编码的滚动轴承故障分类

侯文擎1, 叶鸣2, 李巍华1,2   

  1. 1. 华南理工大学机械与汽车工程学院 广州 510640;
    2. 华南理工大学广东省汽车检测工程技术研究中心 广州 510640
  • 收稿日期:2017-04-27 修回日期:2017-11-29 出版日期:2018-04-05 发布日期:2018-04-05
  • 通讯作者: 李巍华(通信作者),男,1973年出生,教授,博士研究生导师。主要研究方向为动态信号处理、智能诊断、车辆NVH性能测试等。E-mail:whlee@scut.edu.cn
  • 作者简介:侯文擎,男,1990年出生。主要研究方向为机械智能诊断,模式识别。E-mail:houwenqing2011@163.com;叶鸣,男,1973年出生,高级工程师。主要从事机动车检测及检测设备开发。
  • 基金资助:
    国家自然科学基金(51475170)和中央高校基本科研业务费专项资金资助项目。

Rolling Element Bearing Fault Classification Using Improved Stacked De-noising Auto-encoders

HOU Wenqing1, YE Ming2, LI Weihua1,2   

  1. 1. School of Mechanical & Automotive Engineering, South China University of Technology, Guangzhou 510640;
    2. Guangdong Vehicle Tesing Technology Research Center, South China University of Technology, Guangzhou 510640
  • Received:2017-04-27 Revised:2017-11-29 Online:2018-04-05 Published:2018-04-05

摘要: 作为一种新兴的机器学习方法,深度学习在故障诊断领域逐渐得到了应用。其中,堆叠降噪自编码(Stacked de-noising auto-encoders,SDAE)算法先对原始数据添加"损伤噪声",然后通过自编码网络进行数据重构,从而得到更鲁棒性的特征表示,易于进行故障分类。然而针对具体的故障诊断问题,网络隐含层节点数、稀疏参数以及输入数据置零比例将直接影响诊断的结果。因此,提出一种改进的SDAE诊断方法,利用粒子群算法(Particle swarm optimization,PSO)对DAE网络超参数进行自适应的选取来确定SDAE网络结构,据此得到故障状态的特征表示,输入到Soft-max分类器中进行故障分类识别。通过变转速工况下的滚动轴承故障仿真和模拟试验对算法进行验证,试验结果表明,基于PSO-SDAE网络的诊断方法在泛化性、故障识别率方面均优于支持向量机(Support vector machine,SVM)、反向传播神经网络(Back propagation,BP)以及深度置信网络(Deep belief network,DBN)。

关键词: 超参数优化, 故障诊断, 降噪自编码, 深度神经网络

Abstract: As a new machine learning method, deep learning has been used gradually in the field of fault diagnosis. Stacked denoising auto-encoders (SDAE), as one of the deep learning algorithms, could acquire more robust feature representation for effective fault classification by adding "corrupted noise" to the original data, and then reconstructing the input data with the auto-encoder network. However, for a specific diagnosis problem, the number of network hidden nodes, sparse parameters and random zero proportion of input data directly affects the diagnosis results. Based on particle swarm optimization (PSO), an improved SDAE algorithm is proposed for SDAE network hyper-parameters adaptive selection. Then, the determined SDAE networks are used to obtain the feature representations of fault conditions, which could be an input to a soft-max classifier for fault classification. Bearing fault simulation and experiments were conducted under varying running conditions to verify the effectiveness of the proposed method. Experimental results demonstrate that, considering the generalization capability and classification performance, the proposed PSO-SDAE algorithm is superior to support vector machine (SVM), artificial neural network (BP), and deep belief network (DBN).

Key words: deep neural network, denoising auto-encoder, fault classification, hyper-parameters optimization

中图分类号: