高级检索

  • ISSN 1006-3080
  • CN 31-1691/TQ

基于多数据结构的集成质量监控方法

薛敏 杨健 谭帅 侍洪波

薛敏, 杨健, 谭帅, 侍洪波. 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. doi: 10.14135/j.cnki.1006-3080.20180821002
引用本文: 薛敏, 杨健, 谭帅, 侍洪波. 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. doi: 10.14135/j.cnki.1006-3080.20180821002
XUE Min, YANG Jian, TAN Shuai, SHI Hongbo. Integrated Quality Monitoring Method Based on Multiple Data Structure[J]. Journal of East China University of Science and Technology, 2019, 45(6): 938-945. doi: 10.14135/j.cnki.1006-3080.20180821002
Citation: XUE Min, YANG Jian, TAN Shuai, SHI Hongbo. Integrated Quality Monitoring Method Based on Multiple Data Structure[J]. Journal of East China University of Science and Technology, 2019, 45(6): 938-945. doi: 10.14135/j.cnki.1006-3080.20180821002

基于多数据结构的集成质量监控方法

doi: 10.14135/j.cnki.1006-3080.20180821002
基金项目: 国家自然科学基金(61374140):国家自然科学基金青年基金(61403072)
详细信息
    作者简介:

    薛敏:薛 敏(1996-),女,安徽宿州人,硕士生,研究方向为复杂化工过程的质量监控。E-mail: 2857914155@qq.com

    通讯作者:

    侍洪波,E-mail:hbshi@ecust.edu.cn

  • 中图分类号: TP277

Integrated Quality Monitoring Method Based on Multiple Data Structure

  • 摘要: 考虑到工业过程中不同数据结构特征的提取方式可能会影响质量监控性能,提出了一种融合过程数据集全局与局部结构特征的集成质量监控(Ensemble Learning based Multiple Data Structures Quality Monitoring,E-MDSQM)方法。首先,构建偏最小二乘(Partial Least Square,PLS)、邻域保持回归(Neighborhood Preserving Regression,NPR)、局部全局主成分回归(Local and Global Principal Component Regression,LGPCR)3种基础模型,分别描述过程数据的全局结构、局部拓扑及局部全局混合结构信息;然后,基于一种新的监控指标,采用遗传优化算法求得最优权重,集成融合各统计量并确定控制限;最后,通过田纳西-伊斯曼(Tennessee-Eastman Process,TE)过程仿真,评估集成模型的监控效果,并与PLS、NPR、LGPCR 3种基础算法比较,实验结果表明该集成模型取得了较好的综合效果。

     

  • 图  1  集成模型流程图

    Figure  1.  Flowchart of the integrated model

    图  2  PLS、NPR、LGPCR、E-MDSQM算法在故障8下的检测结果

    Figure  2.  Detection results of fault 8 using PLS, NPR, LGPCR and E-MDSQM algorithms

    表  1  PLS、NPR算法在故障2、6、12、13下的检测效果

    Table  1.   Detection results of fault 2, 6, 12, 13 using PLS and NPR algorithms

    FaultPLSNPR
    FDR/%FAR/%SPFAFDR/%FAR/%SPFA
    298.6250.125171950182
    698.62501671000161
    1277.500018580.1250182
    1392.625020779.1250207
    下载: 导出CSV

    表  2  TE过程质量相关故障检测结果

    Table  2.   Detection results of TE process quality related fault

    FaultPLSNPRLGPCRE-MDSQM
    FDR/%FAR/%FDR/%FAR/%FDR/%FAR/%FDR/%FAR/%
    199.3750.62599.000099.000099.1250
    298.6250.12595.000073.750098.6250
    516.875015.750017.6251.2517.0000
    698.6250100010001000
    729.875033.750029.625032.8750
    890.1251.2584.750088.375092.7500
    1018.750017.500021.250019.3750
    1277.5080.1250.62578.625080.8750
    1392.500079.125081.500089.1250
    1887.3750.62587.375086.625088.0000
    下载: 导出CSV

    表  3  TE过程质量无关故障检测结果

    Table  3.   Detection results of TE process quality independent fault

    FaultPLSNPRLGPCRE-MDSQM
    FDR/%FAR/%FDR/%FAR/%FDR/%FAR/%FDR/%FAR/%
    30.3750.6250.75000.87500.8750
    40.75000.75000.2501.2500.7500
    90.87500.5001.2501.1251.2500.8751.000
    113.00002.75001.75002.1250
    142.25011.37501.37502.6250
    151.62502.87502.75002.6250
    下载: 导出CSV

    表  4  TE过程故障数据集的平均故障检测结果

    Table  4.   Average fault detection results of the TE process fault data set

    AlgorithmQuality related faultQuality unrelated fault
    MFDR/%MFAR/%MFDR/%MFAR/%
    PLS80.510.261.480.1
    NPR77.700.063.160.21
    LGPCR75.80.131.350.25
    E-MDSQM82.2501.310.16
    下载: 导出CSV
  • [1] ZHANG K, HAO H Y, CHEN Z W, et al. A comparison and evaluation of key performance indicator-based multivariate statistics process monitoring approaches[J]. Journal of Process Control, 2015, 33(3): 112-126.
    [2] SONG B, SHI H B. Fault detection and classification using quality supervised double-layer method[J]. IEEE Transactions on Industrial Electronics, 2018, 65(10): 8163-8172. doi: 10.1109/TIE.2018.2801804
    [3] 陶阳, 王帆, 侍洪波. 基于ReliefF的主元挑选算法在过程监控中的应用[J]. 化工学报, 2017, 68(4): 1525-1532.
    [4] 吕照民. 基于数据驱动子空间设计的间歇过程监测[D]. 上海: 华东理工大学, 2017.
    [5] 许圆圆, 杨健, 谭帅, 等. 基于局部特征的多模态过程监控方法[J]. 华东理工大学学报(自然科学版), 2017, 43(2): 260-265.
    [6] 曾九孙, 刘祥官, 罗世华, 等. 主成分回归和偏最小二乘法在高炉冶炼中的应用[J]. 浙江大学学报(理学版), 2009, 36(1): 33-36. doi: 10.3785/j.issn.1008-9497.2009.01.007
    [7] HOERL A E, KENNARD R W. Ridge regression: Applications to nonorthogonal problems[J]. Technometrics, 1970, 12(1): 69-82. doi: 10.1080/00401706.1970.10488635
    [8] HE X F, CAI D, YAN S C, et al. Neighborhood preserving embedding[C]// Proceedings of the 10th IEEE International Conference on Computer Vision. USA: IEEE, 2005: 1208-1213.
    [9] YANG J, ZHANG M S, SHI H B, et al. Dynamic learning on the manifold with constrained time information and its application for dynamic process monitoring[J]. Chemometrics and Intelligent Laboratory System, 2017, 167: 179-189. doi: 10.1016/j.chemolab.2017.04.009
    [10] BELKIN M, NIYOGI P. Laplacian Eigenmaps and spectral techniques for embedding and clustering[J]. Advances in Neural Information Processing Systems, 2002, 14(6): 585-591.
    [11] HE X. Locality preserving projections[J]. Advances in Neural Information Processing Systems, 2005, 45(1): 186-197.
    [12] ROWEIS S T, SAUL L K. Nonlinear dimensionality reduction by locally linear embedding[J]. Science, 2000, 290(5): 2323-2326.
    [13] YU J B. Local and global principal component analysis for process monitoring[J]. Journal of Process Control, 2012, 22(7): 1358-1373. doi: 10.1016/j.jprocont.2012.06.008
    [14] MA Y X, SONG B, SHI H B, et al. Fault detection via local and nonlocal embedding[J]. Chemical Engineering Research and Design, 2015, 94: 538-548. doi: 10.1016/j.cherd.2014.09.015
    [15] KRAWCZYK B, MCINNES B T. Local ensemble learning from imbalanced and noisy data for word sense disambiguation[J]. Pattern Recognition, 2018, 78: 103-119. doi: 10.1016/j.patcog.2017.10.028
    [16] 周志华. 机器学习[M]. 北京: 清华大学出版, 2016: 171-196.
    [17] LIU Y, YANG C, GAO Z L, et al. Ensemble deep kernel learning with application to quality prediction in industrial polymerization processes[J]. Chemometrics and Intelligent Laboratory Systems, 2018, 174: 15-21. doi: 10.1016/j.chemolab.2018.01.008
    [18] HIROMASA K. Automatic outlier sample detection based on regression analysis and repeated ensemble learning[J]. Chemometrics and Intelligent Laboratory Systems, 2018, 177: 74-82. doi: 10.1016/j.chemolab.2018.04.015
    [19] 吴朔枫, 颜学峰. 基于即时学习的集成神经网络及其干点预测[J]. 华东理工大学学报(自然科学版), 2016, 42(5): 696-701.
    [20] LI N, YANG Y. Ensemble kernel principal component analysis for improved nonlinear process monitoring[J]. Industrial & Engineering Chemistry Research, 2015, 54(1): 318-329.
    [21] TONG C D, LAN T, SHI X H. Soft sensing of non-Gaussian processes using ensemble modified independent component regression[J]. Chemometrics and Intelligent Laboratory Systems, 2016, 157: 120-126. doi: 10.1016/j.chemolab.2016.07.006
    [22] HAMZE-ZIABARI S M, BAKHSHPOORI T. Improving the prediction of ground motion parameters based on an efficient bagging ensemble model of M5 and CART algorithms[J]. Applied Soft Computing, 2018, 68: 147-161. doi: 10.1016/j.asoc.2018.03.052
    [23] LIU Y, GE Z Q. Weighted random forests for fault classification in industrial processes with hierarchical clustering model selection[J]. Journal of Process Control, 2018, 64: 62-70. doi: 10.1016/j.jprocont.2018.02.005
    [24] LEO B. Bagging predictors[J]. Machine Learning, 1996, 24: 123-140.
    [25] JIANG Q C, YAN X F, LYV Z M, et al. Independent component analysis-based non-Gaussian process monitoring with preselecting optimal components and support vector data description[J]. International Journal of Production Research, 2014, 52(11): 3273-3286. doi: 10.1080/00207543.2013.870362
  • 加载中
图(2) / 表(4)
计量
  • 文章访问数:  8303
  • HTML全文浏览量:  2350
  • PDF下载量:  23
  • 被引次数: 0
出版历程
  • 收稿日期:  2018-09-13
  • 网络出版日期:  2019-06-06
  • 刊出日期:  2019-12-01

目录

    /

    返回文章
    返回