• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2020, Vol. 56 ›› Issue (7): 96-108.doi: 10.3901/JME.2020.07.096

• 机械动力学 • 上一篇    下一篇

扫码分享

基于深度一维残差卷积自编码网络的齿轮箱故障诊断

周兴康, 余建波   

  1. 同济大学机械与能源工程学院 上海 201804
  • 收稿日期:2019-01-19 修回日期:2019-06-27 出版日期:2020-04-05 发布日期:2020-05-12
  • 作者简介:周兴康,男,1996年出生。主要研究方向为深度学习和故障诊断。E-mail:xkzhou@tongji.edu.cn;
    余建波(通信作者),男,1978年出生,博士,教授,博士研究生导师。主要研究方向为故障诊断。E-mail:jbyu@tongji.edu.cn
  • 基金资助:
    国家自然科学基金(71777173)、中央高校基本科研业务费和上海科委创新科技行动计划(19511106303)资助项目。

Gearbox Fault Diagnosis Based on One-dimension Residual Convolutional Auto-encoder

ZHOU Xingkang, YU Jianbo   

  1. School of Mechanical Engineering, Tongji University, Shanghai 201804
  • Received:2019-01-19 Revised:2019-06-27 Online:2020-04-05 Published:2020-05-12

摘要: 一维振动信号常常被用于齿轮箱的监测与故障诊断中,使得能及时地对齿轮箱维护以减少损失。因此,从一维振动信号中提取出关键故障特征决定了故障诊断模型的准确性与可靠性。典型的深度神经网络(deep neural network,DNN),如卷积神经网络已经在故障诊断中表现出良好的性能并得到了广泛的应用,但其监督式训练方式往往需要大量的标签数据而限制了其可应用性。因此,提出一种新的深度神经网络模型,一维残差卷积自编码器(1-dimension residual convolutional auto-encoder,1DRCAE),成功应用于振动信号的无监督学习及故障特征提取,显著提高了齿轮箱的故障诊断率。首先,提出了一维卷积层与自编码器的有效集成方法,形成了深度一维卷积自编码器;其次,引入残差学习机制训练一维卷积自编码器,实现对一维振动信号有效地特征提取;最后,基于编码器提取的特征,使用少量标签数据进行分类微调实现齿轮箱故障模式识别。通过齿轮箱试验台采集的传感器数据进行实验验证表明,这种无监督学习方法具有良好的去噪能力和故障特征提取能力,其特征提取效果好于典型的深度神经网络,如深度置信网络(Deep belief network,DBN)和堆叠自编码网络(Stacked auto-encoders,SAE),同时故障诊断效果也优于一维卷积神经网络(1-dimension convolutional neural network,1DCNN)。

关键词: 齿轮箱故障诊断, 特征学习, 深度学习, 卷积自编码器, 残差学习

Abstract: One-dimension vibration signals are often used for gearbox monitoring and fault diagnosis to perform maintenance timely and then reduce losses. Thus, the accuracy and reliability of the fault diagnosis model are determined by the key fault features extracted from one-dimension vibration signals. Typical deep neural networks (DNNs), e.g., convolutional neural network, have performed well in fault diagnosis and been widely used in machine fault diagnosis. However, supervised learning of CNN often requires a large amount of labeled data thus limits its wide application. So a new DNN, one-dimension residual convolutional auto-encoder (1DRCAE), is proposed for unsupervised learning and feature extraction from vibration signals, which significantly improves diagnosis accuracy of the gearbox fault. First, the effective integration of the one-dimension convolutional layer and the auto-encoder is realized, and a deep one-dimension convolutional auto-encoder is constructed in the 1DRCAE network. Secondly, the residual learning mechanism is introduced to train the one-dimension residual convolutional auto-encoder to learn effective features from one-dimension vibration signals. Finally, based on the features extracted by 1DRCAE, a classifier is constructed using a small amount of labeled data for fine-tuning and then is used for the gearbox fault diagnosis. The vibration signals from a gearbox test rig are used to verify effectiveness of 1DRCAE for gearbox fault diagnosis. The experimental results show that this unsupervised learning of 1DRCAE has good denoising and feature extraction ability. It performs much better on feature extraction than the typical DNNs, e.g., deep belief network (DBN) and stacked auto-encoders (SAE). 1DRCAE also shows better performance on fault diagnosis than that of 1-D CNN (1-DCNN), etc.

Key words: gearbox fault diagnosis, feature learning, deep learning, convolutional auto-encoder, residual learning

中图分类号: