2.793

2018影响因子

(CJCR)

  • 中文核心
  • EI
  • 中国科技核心
  • Scopus
  • CSCD
  • 英国科学文摘

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于ISDAE模型的复杂工业过程运行状态评价方法及应用

褚菲 傅逸灵 赵旭 王佩 尚超 王福利

褚菲, 傅逸灵, 赵旭, 王佩, 尚超, 王福利. 基于ISDAE模型的复杂工业过程运行状态评价方法及应用. 自动化学报, 2021, 45(x): 1−15 doi: 10.16383/j.aas.c200475
引用本文: 褚菲, 傅逸灵, 赵旭, 王佩, 尚超, 王福利. 基于ISDAE模型的复杂工业过程运行状态评价方法及应用. 自动化学报, 2021, 45(x): 1−15 doi: 10.16383/j.aas.c200475
Chu Fei, Fu Yi-Ling, Zhao Xu, Wang Pei, Shang Chao, Wang Fu-Li. Operating performance assessment method and application for complex industrial process based on ISDAE model. Acta Automatica Sinica, 2021, 45(x): 1−15 doi: 10.16383/j.aas.c200475
Citation: Chu Fei, Fu Yi-Ling, Zhao Xu, Wang Pei, Shang Chao, Wang Fu-Li. Operating performance assessment method and application for complex industrial process based on ISDAE model. Acta Automatica Sinica, 2021, 45(x): 1−15 doi: 10.16383/j.aas.c200475

基于ISDAE模型的复杂工业过程运行状态评价方法及应用

doi: 10.16383/j.aas.c200475
基金项目: 国家自然科学基金(61973304, 61503384, 61873049, 62073060), 江苏省六大人才高峰项目(DZXX-045), 江苏省科技计划项目(BK20191339), 徐州市科技创新计划项目(KC19055), 矿冶过程自动控制技术国家重点实验室开放课(BGRIMM-KZSKL-2019-10), 前沿课题专项项目(2019XKQYMS64)资助
详细信息
    作者简介:

    褚菲:中国矿业大学信息与控制工程学院副教授. 2014年获中国东北大学控制理论与控制工程博士学位. 主要研究方向包括复杂工业过程建模、控制与优化、机器学习及运行状态评价等. 本文通信作者. E-mail: chufeizhufei@sina.com

    傅逸灵:中国矿业大学信息与控制工程学院硕士研究生. 2016年获郑州大学电气工程学院学士学位. 主要研究方向为复杂工业过程建模及运行状态评价. E-mail: i11606923@163.com

    赵旭:中国矿业大学信息与控制工程学院硕士研究生. 2017年获三江学院机械与电气工程学院学士学位. 主要研究方向为复杂工业过程运行优化及运行状态评价. E-mail: zhao_xu1994@126.com

    王佩:中国矿业大学信息与控制工程学院硕士研究生. 2019年获合肥师范学院电子信息与电气工程学院学士学位. 主要研究方向为复杂工业过程建模及运行状态评价. E-mail: cumt_aaron@163.com

    尚超:清华大学自动化系助理教授.2016年获清华大学自动化系博士学位. 主要研究方向为大数据解析及工业应用, 过程监控与故障诊断, 工业过程建模等. E-mail: c-shang@tsinghua.edu.cn

    王福利:东北大学教授. 1988年获东北大学自动化系博士学位. 主要研究方向为复杂工业系统的建模、控制与优化, 过程监测和故障诊断等. E-mail: wangfuli@ise.neu.edu.cn

Operating Performance Assessment Method and Application for Complex Industrial Process Based on ISDAE Model

Funds: Supported by National Natural Science Foundation of China (61973304, 61503384, 61873049, 62073060), Selection and Training Project of High-level Talents in the Sixteenth \Six Talent Peaks" of Jiangsu Province (DZXX-045), Science and Technology Plan Project of Jiangsu Province (BK20191339), Science and Technology Innovation Plan Project of Xuzhou (KC19055), Open Foundation of State Key Laboratory of Process Automation in Mining and Metallurgy(BGRIMM-KZSKL-2019-10), and Fundamental Research Funds for the Central Universities (2019XKQYMS64)
  • 摘要: 工业过程运行状态评价通过识别生产过程的运行状态优劣情况, 并通过对非最优因素实时在线追溯, 指导操作人员及时进行生产调整, 保证产品质量, 保障企业的综合经济效益. 针对工业过程中存在强非线性、信息冗余以及不确定性影响而难以建立稳健可靠的运行状态评价模型问题, 提出了一种基于综合经济指标驱动的稀疏降噪自编码器模型(Comprehensive economic index driven sparse denoising autoencoder, ISDAE)的复杂工业过程运行状态评价方法. 首先, 在SDAE(Sparse denoising autoencoder)模型中引入综合经济指标预测误差项, 迫使SDAE学习与综合经济指标相关的数据特征, 建立ISDAE特征提取模型; 其次, 将ISDAE模型所学特征作为输入训练运行状态识别模型, 级联特征提取模型和运行状态识别模型级并通过微调网络结构参数获得运行状态评价模型. 针对非优状态, 提出了一种基于自编码器贡献图算法的非优因素识别方法, 通过计算相应变量的贡献率识别非优因素. 最后, 将所提方法应用于重介质选煤过程的运行状态评价, 实验结果验证了所提方法的有效性和实用性.
  • 图  1  AE模型结构图

    Fig.  1  The structure of AE model

    图  2  基于ISDAE模型的运行状态评价系统框图

    Fig.  2  The system block diagram of ISDAE model based operating performance assessment

    图  3  运行状态在线评价示意图

    Fig.  3  The schematic diagram of online operating performance assessment

    图  4  重介质选煤工艺流程图

    Fig.  4  The process flow diagram of dense medium coal preparation process

    图  5  模型精度与隐层神经元个数的关系图

    Fig.  5  The relationship between the model accuracy and the number of neurons in hidden layer

    图  6  未引入滑动窗口的机理模型数据运行状态在线评价结果

    Fig.  6  Online operating performance assessment result of mechanism model data without introducing sliding window

    图  7  引入滑动窗口的机理模型数据运行状态在线评价结果

    Fig.  7  Online operating performance assessment result of mechanism model data with sliding window

    图  8  实际选煤厂数据分布

    Fig.  8  Data distribution of actual coal preparation plant

    图  9  未引入滑动窗口的实际过程数据运行状态在线评价结果

    Fig.  9  Online operating performance assessment result of field data data without sliding window

    图  10  引入滑动窗口的实际过程数据运行状态在线评价结果

    Fig.  10  Online operating performance assessment result of field data data with sliding window

    图  11  机理模型数据的非优因素识别结果: 第170、211、271个样本为状态“良”的各变量贡献率; 第316、388、409个样本为状态“中”的各变量贡献率; 第460、517、575个样本为状态“差”的各变量贡献率

    Fig.  11  Non-optimal cause identification result of mechanism model data: the contribution rate of each variable of the 170th, 211th, and 271th samples when the state is "fine"; the contribution rate of each variable of the 316th, 388th, and 409th samples, when the state is "medium"; the contribution rate of each variable of the 460th, 517th, and 575th samples, when the state is "poor"

    图  12  实际选煤过程数据的非优因素识别结果

    Fig.  12  The nonoptimal cause identification results of coal preparation field data

    表  1  过程变量选择

    Table  1  The selection of process variable

    编号变量名
    1选煤厂原煤入料(kg/s)
    2双层筛底板筛下流量(kg/s)
    3单层筛顶板上流量(kg/s)
    4混合箱出料密度(kg/m3)
    5混料箱出料流量(m3/s)
    6进入混料箱的重介质密度(kg/m3)
    7旋流器入料压力(Pa)
    8磁性物添加量(kg/s)
    9合格介质桶输出的介质密度(kg/m3)
    10合格介质桶液位(m)
    11合格介质桶出料流量(m3/s)
    下载: 导出CSV

    表  2  机理模型数据运行状态等级划分及等级标签设置

    Table  2  Operating performance level division and level label setting of mechanism model data

    溢流灰分状态等级等级标签
    4.5% ~ 5.5%1
    5.5% ~ 6.7%2
    6.7% ~ 7.7%3
    7.7% ~ 8.7%4
    下载: 导出CSV

    表  3  离线建模数据集中的非优因数设置

    Table  3  Non-optimal factors setting in offline modeling dataset

    状态等级
    采样时刻1 ~ 300301 ~ 400401 ~ 500501 ~ 600601 ~ 700701 ~ 800801 ~ 900901 ~ 10001001 ~ 11001101 ~ 1200
    非优因素变量1变量7变量6变量1变量7变量6变量1变量7变量6
    下载: 导出CSV

    表  4  实际过程数据运行状态等级划分及等级标签设置

    Table  4  Operating performance level division and level label setting of field data

    溢流灰分状态等级等级标签
    6.0% ~ 6.5%1
    6.5% ~ 7.2%2
    7.2% ~ 8.0%3
    8.0% ~ 9.0%4
    下载: 导出CSV

    表  5  模型参数设置

    Table  5  Model parameter setting

    PR$\rho $lr$\beta $$\alpha $$\gamma $
    基于机理模型数据的神经网络模型0.20.10.00120.020.3
    基于实际过程数据的神经网络模型0.10.10.00120.010.1
    下载: 导出CSV

    表  6  测试数据集中的非优因素设置

    Table  6  Non-optimal cause setting in test dataset

    状态等级
    采样时刻0 ~ 150151 ~ 200201 ~ 250251 ~ 300301 ~ 350351 ~ 400401 ~ 450451 ~ 500501 ~ 550551 ~ 600
    非优因素变量1变量7变量6变量1变量7变量6变量1变量7变量6
    下载: 导出CSV

    表  7  TP/FP/FN/TN参数含义

    Table  7  Meaning of parameter TP / FP / FN / TN

    真实情况预测结果
    正例反例
    正例TP(真正例)FN(假反例)
    反例FP(假正例)TN(真反例)
    下载: 导出CSV

    表  8  未引入滑动窗口的运行状态评价结果报告

    Table  8  Report of operating performance assessment results without sliding window

    ISDAESDAEKT-PLS
    精确率召回率F1-值精确率召回率F1-值精确率召回率F1-值
    差(Poor)1.000.850.920.950.810.870.900.620.73
    中(Medium)0.910.970.940.820.900.860.600.680.64
    良(Fine)0.930.970.950.920.950.940.600.710.65
    优(Optimal)0.940.950.940.890.880.890.700.610.65
    宏平均0.940.940.940.900.890.890.700.660.67
    加权平均0.940.940.940.900.890.890.690.660.67
    下载: 导出CSV

    表  9  引入滑动窗口的运行状态评价结果报告

    Table  9  Report of operating performance assessment results with sliding window

    ISDAESDAEKT-PLS
    精确率召回率F1-值精确率召回率F1-值精确率召回率F1-值
    差(Poor)0.990.990.990.991.000.990.960.720.82
    中(Medium)0.990.990.990.980.980.980.760.810.78
    良(Fine)0.980.990.980.980.990.990.730.870.79
    优(Optimal)1.000.970.990.990.960.970.790.700.74
    宏平均0.990.990.990.990.980.980.810.780.78
    加权平均0.990.990.990.980.980.980.800.790.79
    下载: 导出CSV
  • [1] Li W Q, Zhao C H, Gao F R. Linearity evaluation and variable subset partition based hierarchical process modeling and monitoring. IEEE Transactions on Industrial Electronics, 2017, 65(3): 2683−2692
    [2] Jiang Q C, Yan S F, Yan X F, Yi H, Gao F R. Data-Driven Two-Dimensional Deep Correlated Representation Learning for Nonlinear Batch Process Monitoring. IEEE Transactions on Industrial Informatics, 2019, 16(4): 2839−2848
    [3] Xu J, Gu Y, Ma S. Data based online operational performance optimization with varying work conditions for steam-turbine system. Applied Thermal Engineering, 2019, 151: 344−353 doi: 10.1016/j.applthermaleng.2019.02.032
    [4] Hu J, Wu M, Chen X, et al. Hybrid modeling and online optimization strategy for improving carbon efficiency in iron ore sintering process. Information Sciences, 2019, 483: 232−246 doi: 10.1016/j.ins.2019.01.027
    [5] Frangos M. Uncertainty quantification for cuttings transport process monitoring while drilling by ensemble Kalman filtering. Journal of Process Control, 2017, 53: 46−56 doi: 10.1016/j.jprocont.2017.02.008
    [6] 刘洋, 张国山. 基于敏感稀疏主元分析的化工过程监测与故障诊断. 控制与决策, 2016, 31(7): 1213−1218

    Liu Yang, Zhang Guo-shan. Chemical process monitoring and fault diagnosis based on sensitive sparse principal component analysis. Control and Decision, 2016, 31(7): 1213−1218
    [7] Liu Y, Wang F L, Chang Y Q. Operating optimality assessment based on optimality related variations and nonoptimal cause identification for industrial processes. Journal of Process Control, 2016, 39: 11−20 doi: 10.1016/j.jprocont.2015.12.008
    [8] Liu Y, Chang Y Q, Wang F L. Online process operating performance assessment and nonoptimal cause identification for industrial processes. Journal of Process Control, 2014, 24(10): 1548−1555 doi: 10.1016/j.jprocont.2014.08.001
    [9] Liu Y, Wang F L, Chang Y Q, Ma R C. Operating optimality assessment and nonoptimal cause identification for non-Gaussian multimode processes with transitions. Chemical Engineering Science, 2015, 137: 106−118 doi: 10.1016/j.ces.2015.06.016
    [10] Zou X Y, Wang F L, Chang Y Q. Assessment of operating performance using cross-domain feature transfer learning. Control Engineering Practice, 2019, 89: 143−153 doi: 10.1016/j.conengprac.2019.05.007
    [11] Vo H X, Durlofsky L J. Data assimilation and uncertainty assessment for complex geological models using a new PCA-based parameterization. Computational Geosciences, 2015, 19(4): 747−767 doi: 10.1007/s10596-015-9483-x
    [12] Zhang Q C, Yang L T, Chen Z K, Li P. A survey on deep learning for big data. Information Fusion, 2018, 42: 146−157 doi: 10.1016/j.inffus.2017.10.006
    [13] Eren L, Ince T, Kiranyaz S. A generic intelligent bearing fault diagnosis system using compact adaptive 1D CNN classifier. Journal of Signal Processing Systems, 2019, 91(2): 179−189 doi: 10.1007/s11265-018-1378-3
    [14] Wang J J, Ma Y L, Zhang L B, Gao R X, Wu D Z. Deep learning for smart manufacturing: Methods and applications. Journal of Manufacturing Systems, 2018, 48: 144−156 doi: 10.1016/j.jmsy.2018.01.003
    [15] Sun W J, Shao S Y, Zhao R, Yang R Q, Zhang X W. A sparse auto-encoder-based deep neural network approach for induction motor faults classification. Measurement, 2016, 89: 171−178 doi: 10.1016/j.measurement.2016.04.007
    [16] Jiang L, Ge Z Q, Song Z H. Semi-supervised fault classification based on dynamic Sparse Stacked auto-encoders model. Chemometrics and Intelligent Laboratory Systems, 2017, 168: 72−83 doi: 10.1016/j.chemolab.2017.06.010
    [17] Yu W K, Zhao C H. Robust Monitoring and Fault Isolation of Nonlinear Industrial Processes Using Denoising Autoencoder and Elastic Net. IEEE Transactions on Control Systems Technology, 2019, 28(3): 1−9
    [18] 李炜, 宋威, 王晨妮, 张雨轩. 标签约束的半监督栈式自编码器分类算法. 小型微型计算机系统, 2019, 40(3): 488−492 doi: 10.3969/j.issn.1000-1220.2019.03.005

    Li W, Song W, Wang C N, Zhang Y X. Label Regularization Semi-supervised Stacked Autoencoder Classification Algorithm. Journal of Chinese Computer Systems, 2019, 40(3): 488−492 doi: 10.3969/j.issn.1000-1220.2019.03.005
    [19] Chai Z L, Song W, Wang H L, Liu F. A semi-supervised auto-encoder using label and sparse regularizations for classification. Applied Soft Computing, 2019, 77: 205−217 doi: 10.1016/j.asoc.2019.01.021
    [20] Yuan X F, Huang B, Wang Y L, Yang C H, Gui W H. Deep learning-based feature representation and its application for soft sensor modeling with variable-wise weighted SAE. IEEE Transactions on Industrial Informatics, 2018, 14(7): 3235−3243 doi: 10.1109/TII.2018.2809730
    [21] Yuan X F, Zhou J, Huang B, Wang Y L, Yang C H, Gui W H. Hierarchical quality-relevant feature representation for soft sensor modeling: a novel deep learning strategy. IEEE Transactions on Industrial Informatics, 2019, 16(6): 3721−3730
    [22] Sohaib M, Kim J M. Reliable fault diagnosis of rotary machine bearings using a stacked sparse autoencoder-based deep neural network. Shock and Vibration, 2018, 2018.
    [23] Bengio Y. Learning deep architectures for AI[M]. Now Publishers Inc, 2009.
    [24] Chen Z Y, Li W H. Multisensor feature fusion for bearing fault diagnosis using sparse autoencoder and deep belief network. IEEE Transactions on Instrumentation and Measurement, 2017, 66(7): 1693−1702 doi: 10.1109/TIM.2017.2669947
    [25] Lv F Y, Wen C L, Liu M Q, Bao Z J. Weighted time series fault diagnosis based on a stacked sparse autoencoder. Journal of Chemometrics, 2017, 31(9): e2912 doi: 10.1002/cem.2912
    [26] Long W, Gao L, Li X Y. A new deep transfer learning based on sparse auto-encoder for fault diagnosis. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 2017, 49(1): 136−144
    [27] Vincent P, Larochelle H, Lajoie I, Bengio Y, Manzagol P A. Stacked Denoising Autoencoders: Learning Useful Representations in a Deep Network with a Local Denoising Criterion. Journal of Machine Learning Research, 2010, 11(12): 3371−3408
    [28] 刘国梁, 余建波. 知识堆叠降噪自编码器[Online]. 自动化学报: 1-13. available: https://doi.org/10.16383/j.aas.c190375. 2020-11-6

    Liu G L, Yu J B. Knowledge-based stacked denoising Autoencoder[Online]. Acta Automatica Sinica: 1-13, available: https://doi.org/10.16383/j.aas.c190375. November 26, 2020
    [29] 邹筱瑜, 王福利, 常玉清, 郑伟. 基于两层分块GMM-PRS的流程工业过程运行状态评价. 自动化学报, 2019, 45(11): 2071−2081

    Zou X Y, Wang F L, Chang Y Q, Zheng W. Plant-wide Process Operating Performance Assessment Based on Two-level Multi-block GMM-PRS. Acta Automatica Sinica, 2019, 45(11): 2071−2081
    [30] 王慧玲, 宋威, 王晨妮. 稀疏和标签约束半监督自动编码器的分类算法. 计算机应用研究, 2019, 36(9): 2613−2617

    Wang H L, Song W, Wang C N. Semi-supervised auto-encoder using sparse and label regularizations for classification. Application Research of Computers, 2019, 36(9): 2613−2617
    [31] Rumelhart D E, Hinton G E, Williams R J. Learning representations by back-propagating errors. Cognitive modeling, 1988, 5(3): 1
    [32] Miller P, Swanson R E, Heckler C E. Contribution plots: a missing link in multivariate quality control. Applied mathematics and computer science, 1998, 8(4): 775−792
    [33] Yoon S, MacGregor J F. Fault diagnosis with multivariate statistical models part I: using steady state fault signatures. Journal of process control, 2001, 11(4): 387−400 doi: 10.1016/S0959-1524(00)00008-1
    [34] 蒋立. 基于自编码器模型的非线性过程监测[博士论文]. 浙江大学, 中国, 2018.

    Jiang L. Nonlinear Process Monitoring Based On Auto-encoder Model[Ph. D. dissertation]. Zhejiang University, 2018.
    [35] 褚菲, 赵旭, 代伟, 马小平, 王福利. 数据驱动的最优运行状态鲁棒评价方法及应用. 自动化学报, 2020, 46(3): 439−450

    Chu F, Zhao X, Dai W, Ma X P, Wang F L. Data-driven Robust Evaluation Method and Application for the Optimal Operating Status. Acta Automatica Sinica, 2020, 46(3): 439−450
    [36] E J Meyer, I K Craig. The development of dynamic models for a dense medium separation circuit in coal beneficiation. Minerals Engineering, 2010, 23: 791−805 doi: 10.1016/j.mineng.2010.05.020
    [37] Zhang L J, Xia X H, Zhu B. A dual-loop control system for dense medium coal washing processes with sampled and delayed measurements. IEEE Transactions on Control Systems Technology, 2017, 25(6): 2211−2218 doi: 10.1109/TCST.2016.2640946
    [38] Liu Y, Chang Y Q, Wang F L, Ma R C, Zhang H L. Complex process operating optimality assessment and nonoptimal cause identification using modified total kernel PLS. In: Proceedings of the 26th Chinese Control and Decision Conference. IEEE, 2014: 1221−1227.
  • 加载中
计量
  • 文章访问数:  64
  • HTML全文浏览量:  12
  • 被引次数: 0
出版历程
  • 收稿日期:  2020-06-29
  • 录用日期:  2020-11-18
  • 网络出版日期:  2021-01-16

目录

    /

    返回文章
    返回