高级检索

  • ISSN 1006-3080
  • CN 31-1691/TQ

基于多块卷积变分信息瓶颈的多变量动态过程故障诊断

何雨旻 侍洪波

何雨旻, 侍洪波. 基于多块卷积变分信息瓶颈的多变量动态过程故障诊断[J]. 华东理工大学学报(自然科学版). doi: 10.14135/j.cnki.1006-3080.20201022001
引用本文: 何雨旻, 侍洪波. 基于多块卷积变分信息瓶颈的多变量动态过程故障诊断[J]. 华东理工大学学报(自然科学版). doi: 10.14135/j.cnki.1006-3080.20201022001
HE Yumin, SHI Hongbo. Multivariate Dynamic Process Fault Diagnosis Based on Multi-Block Convolutional Variational Information Bottleneck[J]. Journal of East China University of Science and Technology. doi: 10.14135/j.cnki.1006-3080.20201022001
Citation: HE Yumin, SHI Hongbo. Multivariate Dynamic Process Fault Diagnosis Based on Multi-Block Convolutional Variational Information Bottleneck[J]. Journal of East China University of Science and Technology. doi: 10.14135/j.cnki.1006-3080.20201022001

基于多块卷积变分信息瓶颈的多变量动态过程故障诊断

doi: 10.14135/j.cnki.1006-3080.20201022001
基金项目: 国家自然科学基金(61673173, 61703161);上海市自然科学基金(19ZR1473200)
详细信息
    作者简介:

    何雨旻(1995—),男,贵州贵阳人,硕士生,主要研究方向为基于深度学习的故障诊断。E-mail:yuminhe91@163.com

    通讯作者:

    侍洪波,E-mail:hbshi@ecust.edu.cn

  • 中图分类号: TP277

Multivariate Dynamic Process Fault Diagnosis Based on Multi-Block Convolutional Variational Information Bottleneck

  • 摘要: 针对多变量动态过程的故障诊断,采用局部提取、全局整合的特征提取策略,提出了一种多块卷积变分信息瓶颈(Multi-Block Convolutional Variational Information Bottleneck,MBCVIB)模型。首先,根据过程机理,对所有变量分块,将同一操作单元的变量划分为同一子块,再利用一维卷积神经网络(One-Dimensional Convolutional Neural Network,1-D CNN)提取过程中每个子块的局部特征,从而考虑样本间的时序相关性;然后,整合局部特征得到全局特征表示,在全局特征的基础上,根据变分信息瓶颈(Variational Information Bottleneck,VIB)原理进一步提取与故障最相关的信息;最后,采用连续搅拌釜反应器(Continuous Stirred Tank Reactor,CSTR)和田纳西-伊士曼过程(Tennessee Eastman Process,TEP)对模型的有效性进行了验证。结果显示本文模型在CSTR上实现了0.983的平均故障诊断准确率,在TEP上实现了0.955的平均故障诊断准确率。

     

  • 图  1  1-D CNN网络结构图

    Figure  1.  Network structure of ordinary 1-D CNN

    图  2  并行多块1-D CNN

    Figure  2.  Parallel multiblock 1-D CNN

    图  3  MBCVIB网络结构

    Figure  3.  Network structure of MBCVIB

    图  4  基于MBCVIB的故障诊断方法流程

    Figure  4.  Procedure of fault diagnosis method based on MBCVIB

    图  5  特征可视化结果

    Figure  5.  Consequences of feature visualization

    表  1  CSTR故障描述

    Table  1.   Faults description in CSTR

    Fault numberFault description
    1Bias in the sensor of the output temperature T
    2Bias in the sensor of the inlet temperature T0
    3Bias in inlet reactant concentration CAA
    4Drift in the sensor of CAA
    5Slow drift in reaction kinetics
    6Drift in E/R
    7Drift in Uac
    8Drift in CAA, CAS, FS, and T0
    9Step change in T0
    下载: 导出CSV

    表  2  CSTR故障诊断模型结构

    Table  2.   Detailed fault diagnosis model structures in CSTR

    ModelStructure
    Stacked AEInput(1×(9·10))-FC(30)-FC(20)-FC(10)-Softmax
    LSTMInput(9×10)-LSTM(50)-LSTM(50)-FC(30)-FC(10)-Softmax
    2-D CNNInput(9×10)-Conv2D(64)-Conv2D(64)-Maxpool(2×2)-Conv2D(128)-FC(128)-FC(10)-Softmax
    MBCVIBInput(9×10)- (Conv(60)·2)·4-Concatenation- FC(100)×2-FC(100)-FC(10)-Softmax
    下载: 导出CSV

    表  3  CSTR故障诊断模型参数

    Table  3.   Detailed fault diagnosis model parameters in CSTR

    ModelWindow widthActivation functionConvolutional kernelLearning rateBatch sizeEpoch
    Stacked AE10SELU1×10−32420
    LSTM10SELU1×10−32420
    2-D CNN10SELU3×31×10−32420
    MBCVIB10SELU(Vb×3)1×10−3 2420
    下载: 导出CSV

    表  4  CSTR分块结果

    Table  4.   Block division result in CSTR

    BlockProcess variables
    1T0, CAS, FS
    2FA, CAA
    3FC, TC
    4 CA, T
    下载: 导出CSV

    表  5  不同模型在CSTR上的故障诊断准确率

    Table  5.   Fault diagnosis accuracy of different models on CSTR

    FaultAccuracy
    SVMStacked AELSTM2-D CNNMBCVIB
    Normal0.2990.6880.4801.0000.992
    10.5730.9380.8300.9191.000
    20.9920.9940.9920.9920.994
    30.9980.9930.9770.9960.997
    40.5250.7100.9310.8500.876
    50.9730.9730.9840.9830.995
    60.9120.9580.9560.9680.998
    70.8110.9380.9320.9650.985
    80.9940.9940.9890.9940.998
    90.9450.9610.9880.9601.000
    Average0.8020.9150.9060.9630.983
    下载: 导出CSV

    表  6  TEP实验时延数据维度

    Table  6.   Time-delay data dimension of TEP

    Status indexTrainingTesting
    Normal481×(52×20)981×(52×20)
    Fault 1~5, 7~20461×(52×20)781×(52×20)
    Fault 6121×(52×20)121×(52×20)
    下载: 导出CSV

    表  7  TEP故障诊断模型结构

    Table  7.   Detailed fault diagnosis model structures in TEP

    ModelStructure
    Stacked AEInput((1×(52·20))-FC(500)-FC(250)-FC(18)-Softmax
    LSTMInput(52×20)-LSTM(128)-LSTM(128)-FC(128)-FC(64)-FC(18)-Softmax
    2-D CNNInput(52×20)-Conv2D(64)-Conv2D(256)-Maxpool(2×2)-Conv2D(128)-Maxpool(2×2)-FC(384)-FC(18)-Softmax
    MBCVIBInput(52×20)- (Conv(60)·6)×4-Concatenation-FC(100)×2 -FC(100)-FC(18)-Softmax
    下载: 导出CSV

    表  8  TEP故障诊断模型参数

    Table  8.   Detailed fault diagnosis model parameters in TEP

    ModelWindow widthActivation functionConvolutional kernelLearning rateBatch sizeEpoch
    Stacked AE20SELU${\rm{1}} \times {\rm{1}}{{\rm{0}}^{ - 3}}$2420
    LSTM20SELU${\rm{1}} \times {\rm{1}}{{\rm{0}}^{ - 3}}$2420
    2-D CNN20SELU3×3${\rm{1}} \times {\rm{1}}{{\rm{0}}^{ - 3}}$2420
    MBCVIB20SELU(${V_b}$×3)${\rm{1}} \times {\rm{1}}{{\rm{0}}^{ - 3}}$2420
    下载: 导出CSV

    表  9  TEP中每个操作单元的变量

    Table  9.   The variables of each operation unit in TEP

    Operation unitProcess variables
    ReactorXMEAS(1, 2, 3, 4, 5, 6, 7, 8, 9, 21, 23, 24, 25, 26, 27, 28); XMV(1, 2, 3, 10)
    CondenserXMV(11)
    Recycle compressorXMEAS(20); XMV(5)
    SeparatorXMEAS(10, 11, 12, 13, 14, 22, 29, 30, 31, 32, 33, 34, 35, 36); XMV(6, 7)
    stripperXMEAS(15, 16, 17, 18, 19, 37, 38, 39, 40, 41; XMV(4, 8, 9)
    下载: 导出CSV

    表  10  TEP分块结果

    Table  10.   Block division result of TEP

    BlockProcess variables
    1XMEAS(1, 2, 3, 4, 5, 6, 7, 8, 9, 21, 23, 24, 25, 26, 27, 28); XMV(1, 2, 3, 10)
    2XMV(5, 11); XMEAS(20)
    3XMEAS(10, 11, 12, 13, 14, 22, 29, 30, 31, 32, 33, 34, 35, 36); XMV(6, 7)
    4XMEAS(15, 16, 17, 18, 19, 37, 38, 39, 40, 41); XMV(4, 8, 9)
    下载: 导出CSV

    表  11  不同模型在TEP中的故障诊断准确率

    Table  11.   Fault diagnosis accuracy of different models on TEP

    FaultAccuracy
    SVMStacked AELSTM2-D CNNMBCVIB
    Normal0.4280.3300.2430.4140.942
    11.0000.9971.0001.0000.997
    20.9960.9970.9950.9950.990
    41.0000.9960.8580.3191.000
    51.0000.9920.8550.2230.996
    61.0001.0000.9751.0001.000
    71.0001.0000.9671.0001.000
    80.4060.5300.5350.7380.813
    100.2690.2720.0730.4260.887
    110.10.4600.5380.8760.982
    120.6900.6400.6970.8170.942
    130.8770.9160.9390.9170.939
    14110.9990.9991.000
    160.3470.3840.2920.3640.955
    170.8760.8990.8670.8410.941
    180.9320.9330.9350.9310.949
    190.6200.4440.6700.8870.947
    200.700.8350.6180.6940.910
    Average0.7360.7570.7250.7470.955
    下载: 导出CSV
  • [1] 薛敏, 杨健, 谭帅, 等. 基于多数据结构的集成质量监控方法[J]. 华东理工大学学报(自然科学版), 2019, 45(6): 938-945.
    [2] 吕铮, 杨健, 侍洪波, 等. 基于T-TSNPR的动态过程质量监控[J]. 华东理工大学学报(自然科学版), 2019, 45(6): 946-953.
    [3] YANG J, LYU Z, SHI H, et al. Performance monitoring method based on balanced partial least square and statistics pattern analysis[J]. ISA Transactions, 2018, 81: 121-131. doi: 10.1016/j.isatra.2018.07.038
    [4] TAO Y, SHI H B, SONG B, et al. Parallel quality-related dynamic principal component regression method for chemical process monitoring[J]. Journal of Process Control, 2019, 73: 33-45. doi: 10.1016/j.jprocont.2018.08.009
    [5] AMIN M T, IMTIAZ S, KHAN F. Process system fault detection and diagnosis using a hybrid technique[J]. Chemical Engineering Science, 2018, 189: 191-211. doi: 10.1016/j.ces.2018.05.045
    [6] ZHANG Y. Fault detection and diagnosis of nonlinear processes using improved kernel independent component analysis (KICA) and support vector machine (SVM)[J]. Industrial & Engineering Chemistry Research, 2008, 47(18): 6961-6971.
    [7] RATO T, RESI M, SCHMITT E, et al. A systematic comparison of PCA‐based Statistical Process Monitoring methods for high‐dimensional, time‐dependent Processes[J]. AIChE Journal, 2016, 62(5): 1478-1493. doi: 10.1002/aic.15062
    [8] ZHU Z B, SONG Z H, Fault diagnosis based on imbalance modified kernel Fisher discriminant analysis[J]. Chemical Engineering Research & Design, 2010, 88(8): 936-951.
    [9] YIN G, ZHANG Y T, LI Z N, et al. Online fault diagnosis method based on Incremental support vector data description and extreme learning machine with incremental output structure[J]. Neurocomputing, 2014, 128: 224-231. doi: 10.1016/j.neucom.2013.01.061
    [10] 张祥, 崔哲, 董玉玺, 等. 基于VAE-DBN的故障分类方法在化工过程中的应用[J]. 过程工程学报, 2018, 18(3): 590-594.
    [11] 王翔, 任佳. 基于多注意力机制的深度神经网络故障诊断算法[J]. 浙江理工大学学报(自然科学版), 2020, 43(2): 224-231.
    [12] 王翔, 柯飂挺, 任佳. 样本重构多尺度孪生卷积网络的化工过程故障检测[J]. 仪器仪表学报, 2019, 40(11): 184-191.
    [13] ZHU J Z, SHI H B, SONG B, et al. Deep neural network based recursive feature learning for nonlinear dynamic process monitoring[J]. The Canadian Journal of Chemical and Engineering, 2020, 98(4): 919-933. doi: 10.1002/cjce.23669
    [14] ZHANG Z H, JIANG T, ZhAn C J, et al. Gaussian feature learning based on variational autoencoder for improving nonlinear process monitoring[J]. Journal of Process Control, 2019, 75: 136-155. doi: 10.1016/j.jprocont.2019.01.008
    [15] 于志强, 余正涛, 黄于欣, 等. 基于变分信息瓶颈的半监督神经机器翻译[J/OL]. 自动化学报, 2020, 46(x): 1-12. https://doi.org/10.16383/j.aas.c190477.
    [16] ALEMI A, FISCHER I, JOSHUA, DILLON V, et al. Deep variational information bottleneck[EB/OL]. arX-iv, (2016-12-01), [2020-10-21]. arXiv: 1612.00410.
    [17] WU H, ZHAO J. Deep convolutional neural network model based chemical process fault diagnosis[J]. Computers & Chemical Engineering, 2018, 115: 185-197.
    [18] CHEN S M, YU J B, WANG S J. One-dimensional convolutional auto-encoder-based feature learning for fault diagnosis of multivariate processes[J]. Journal of Process Control, 2020, 87: 54-67. doi: 10.1016/j.jprocont.2020.01.004
    [19] YIN J, YAN X F. Mutual information−dynamic stacked sparse autoencoders for fault detection[J]. Industrial & Engineering Chemistry Research, 2019, 58: 21614-21624.
    [20] TISHBY N, ZASLAVSKY N. Deep learning and the information bottleneck principle[C]// IEEE Information Theory Workshop (ITW). Jerusalem: IEEE, 2015: 1-5.
    [21] ALCALA C F, QIN S J. Reconstruction-based contribution for process monitoring[J]. Automatica, 2009, 45(7): 1593-1600. doi: 10.1016/j.automatica.2009.02.027
    [22] 杨健, 宋冰, 谭帅, 等. 时序约束NPE算法在化工过程故障检测中的应用[J]. 化工学报, 2016, 67(12): 5131-5139.
    [23] ZHANG Z, ZHAO J S. A deep belief network based fault diagnosis model for complex chemical processes[J]. Computers & Chemical Engineering, 2017, 107: 395-407.
    [24] CHIANG L H, RUSSELL E L, BRAATZ R D, Fault Detection and Diagnosis in Industrial Systems[M]. London: Springer, 2001.
    [25] HINTON G E. Visualizing high-dimensional data using t-SNE[J]. Journal of Machine Learning Research, 2008, 9(2): 2579-2605.
  • 加载中
图(5) / 表(11)
计量
  • 文章访问数:  484
  • HTML全文浏览量:  290
  • PDF下载量:  9
  • 被引次数: 0
出版历程
  • 收稿日期:  2020-10-22
  • 网络出版日期:  2021-01-19

目录

    /

    返回文章
    返回