基于Laplace小波卷积和BiGRU的少量样本故障诊断方法

PDF(2028 KB)

振动与冲击 ›› 2022, Vol. 41 ›› Issue (24) : 41-50.

论文

罗浩1，何超1，陈彪1，路颜萍1，张欣2，张利1

作者信息 +

Small sample fault diagnosis based on Laplace wavelet convolution and BiGRU

LUO Hao1，HE Chao1，CHEN Biao1，LU Yanping1，ZHANG Xin2，ZHANG Li1

Author information +

文章历史 +

摘要

针对滚动轴承通常在复杂条件下工作易发生故障以及训练样本较少等问题。提出一种具有全局平均池化(Global average pooling, GAP)并融合双路Laplace小波卷积和双向门控循环单元(Daul Laplace Wavelet Convolution Bidirectional Gated Recurrent Unit, DLWCB)的故障诊断方法。首先Laplace小波卷积将原始信号从时域转换为频域，接着利用双路卷积和BiGRU挖掘少量样本的多尺度和时空特征；然后设计GAP降低模型的参数量并全面融合各GRU细胞提取的时空特征。其中从优化算法和目标函数入手，引入标签平滑、AdamP等提升DLWCB应对少量样本的能力，最后实现复杂工况下故障诊断。在两种轴承数据集、有限噪声样本下，50秒内便可完成训练，达到98%以上准确率，所提方法具有良好泛化性、鲁棒性和诊断效率。
关键词：拉普拉斯小波卷积核; 双向门控循环单元; 标签平滑；故障诊断；少量样本

Abstract

Targeting the problems that rolling bearings usually work under complex conditions, causing breakdown easily and small training samples. A fault diagnosis method with global average pooling (GAP) and fusion of dual Laplace wavelet convolution and Bidirectional Gated Recurrent Unit(DLWCB) is proposed. Firstly, Laplace wavelet convolution is utilized to transform original signals from time to frequency domain, and then the multi-scale and spatiotemporal characteristics of small samples are mined by dual convolution and BiGRU. In addition, GAP is designed to reduce the amount of parameters of the model and integrate the spatiotemporal characteristics of GRUs. From the optimization algorithms and objective functions, label smoothing and AdamP are introduced to improve the capacity of DLWCB to cope with small samples, and finally achieve fault diagnosis under complex conditions. In two rolling bearing data sets, training can be completed in 50 seconds under limited noise samples, and the accuracy is over 98%. The proposed method has the better capacities of generalization, robustness and diagnosis efficiency.
Key words: Laplace wavelet convolution kernel; bidirectional gated recurrent unit; label smoothing; fault diagnosis; small sample

导出引用

罗浩1，何超1，陈彪1，路颜萍1，张欣2，张利1. 基于Laplace小波卷积和BiGRU的少量样本故障诊断方法[J]. 振动与冲击, 2022, 41(24): 41-50

LUO Hao1，HE Chao1，CHEN Biao1，LU Yanping1，ZHANG Xin2，ZHANG Li1. Small sample fault diagnosis based on Laplace wavelet convolution and BiGRU[J]. Journal of Vibration and Shock, 2022, 41(24): 41-50

参考文献

[1]JIAO Jinyang, ZHAO Ming, LIN Jing, et al. A Comprehensive Review on Convolutional Neural Network in Machine Fault Diagnosis[J]. Neurocomputing, 2020, 417:36-63.
[2]LUO Hao, He Chao, ZHOU Jianing, et al. Rolling Bearing Sub-Health Recognition via Extreme Learning Machine Based on Deep Belief Network Optimized by Improved Fireworks[J]. IEEE Access, 2021, 9:42013-42026.
[3]KE Yun, YAO Chong, SONG Enzhe, et al. An early fault diagnosis method of common-rail injector based on improved CYCBD and hierarchical fluctuation dispersion entropy[J]. Digital Signal Processing, 2021, 114:103049.
[4]何勇，王红，谷穗. 一种基于遗传算法的VMD参数优化轴承故障诊断新方法[J]. 振动与冲击, 2021, 40(6):184-189.
HE Yong, WANG Hong, GU Sui. New fault diagnosis approach for bearings based on parameter optimized VMD and genetic algorithm[J]. JOURNAL OF VIBRATION AND SHOCK, 2021, 40(6): 184-189.
[5]ZHANG Shen, ZHANG Shibo, WANG Bingnan et al. Deep Learning Algorithms for Bearing Fault Diagnostics—A Comprehensive Review[J]. IEEE Access, 2020, 8:29857-29881.
[6]HAN Tian, Ma Ruiyi, ZHENG Jigui. Combination bidirectional long short-term memory and capsule network for rotating machinery fault diagnosis[J]. Measurement, 2021, 176: 109208.
[7]SAUFI S R, AHMAD Z A B, LEONG M S, et al. Gearbox fault diagnosis using a deep learning model with limited data sample[J]. IEEE Transactions on Industrial Informatics, 2020, 16(10):6263-6271.
[8]LI Chenzhong, YANG Kanghua, TANG Haichuan, et al. Fault Diagnosis for Rolling Bearings of a Freight Train under Limited Fault Data: Few-Shot Learning Method[J]. Journal of Transportation Engineering, Part A: Systems, 2021, 147(8):04021041.
[9]ZHANG Ansi, LI Shaobo, CUI Yuxin, et al. Limited data rolling bearing fault diagnosis with few-shot learning[J]. IEEE Access, 2019, 7: 110895-110904.
[10]WANG Cunjun, XU Zili, An intelligent fault diagnosis model based on deep neural network for few-shot fault diagnosis[J]. Neurocomputing, 2021.
[11]WU Jingyao, ZHAO Zhibin, SUN Chuang, et al. Few-shot Transfer Learning for Intelligent Fault Diagnosis of Machine[J]. Measurement, 2020, 166:108202.
[12]沈涛,李舜酩. 针对滚动轴承故障的批标准化CNN-LSTM诊断方法[J]. 计算机集成制造系统, 2021:1-16.
SHEN Tao, LI Shunming. Batch standardized CNN-LSTM method with batch normalization for rolling bearing fault diagnosis[J], Computer Integrated Manufacturing Systems, 2021:1-16.
[13]YANG Jianxi, YANG Fei, ZHOU Yingxin, et al. A data-driven structural damage detection framework based on parallel convolutional neural network and bidirectional gated recurrent unit[J]. Information Sciences, 2021, 566:103-117.
[14]GU Jindong, Tresp V, HU Han. Capsule Network is Not More Robust than Convolutional Network[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021: 14309-14317.
[15]LI Tianfu, ZHAO Zhibin, SUN Chuang, et al. WaveletKernelNet: An Interpretable Deep Neural Network for Industrial Intelligent Diagnosis[J]. IEEE Transactions on Systems, Man, and Cybernetics:Systems, 2021:1-11.
[16]邓飞跃，强亚文，郝如江等. 基于自适应Morlet小波参数字典设计的微弱故障检测方法研究[J]. 振动与冲击, 2021, 40(8):187-193.
DENG Feiyue, QIANG Yawen, HAO Rujiang, et al. A study on a weak fault detection method based on adaptive parametric dictionary design using the Morlet wavelet[J]. JOURNAL OF VIBRATION AND SHOCK, 2021, 40(8): 187-193.
[17]CHEN Xiaohan, ZHANG Beike, GAO Dong. Bearing fault diagnosis base on multi-scale CNN and LSTM model[J]. Journal of Intelligent Manufacturing, 2021, 32(4):971–987.
[18]QIAO Meiying, YAN Shuhao, TANG Xiaxia, et al. Deep Convolutional and LSTM Recurrent Neural Networks for Rolling Bearing Fault Diagnosis Under Strong Noises and Variable Loads[J]. IEEE Access, 2020, 8:66257-66269.
[19]HEO B, SANGHYUK C, SEONG J O, et al. AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-Invariant Weights[C]//The 9th International Conference on Learning Representations (ICLR), 2021.
[20]SHEN Zhiqiang, LIU Zechun, XU Dejia et al. Shen, Zhiqiang, et al. Is Label Smoothing Truly Incompatible with Knowledge Distillation: An Empirical Study[C]//The 9th International Conference on Learning Representations(ICLR), 2021.
[21]Case Western Reserve University(CWRU) Bearing Data Center, [EB/OL], https://csegroups.case.edu/bearingdatacenter/pages/download-data-file/.
[22]ZHANG Wei, PENG Gaoliang, LI Chuanhao, et al. A New Deep Learning Model for Fault Diagnosis with Good Anti-Noise and Domain Adaptation Ability on Raw Vibration Signals[J]. Sensors, 2017, 17(2):425.
[23]ZHAO Zhibin, LI Tianfu, WU Jingyao, et al. Deep learning algorithms for rotating machinery intelligent diagnosis: An open source benchmark study[J]. ISA Transactions, 2020, 107:224-255.
[24]LIN T Y, GOYAL P, GIRSHICK R, et al. Focal Loss for Dense Object Detection[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017, 42(2): 2999-3007.
[25]LI Buyu, LIU Yu, WANG Xiaogang, et al. Gradient Harmonized Single-Stage Detector[C]//Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2019, 33(1): 8577-8584.
[26]DONG Yunjia, LI Yuqing, ZHENG Huailiang, et al. A new dynamic model and transfer learning based intelligent fault diagnosis framework for rolling element bearings race faults: Solving the small sample problem[J]. ISA Transactions, 2021.