化工学报 ›› 2022, Vol. 73 ›› Issue (7): 3156-3165.doi: 10.11949/0438-1157.20220349

• 过程系统工程 • 上一篇    下一篇

深度融合特征提取网络及其在化工过程软测量中的应用

周乐1(),沈程凯1,吴超1,侯北平1(),宋执环2   

  1. 1.浙江科技学院自动化与电气工程学院，浙江 杭州 310024
    2.浙江大学工业控制技术国家重点实验室，浙江 杭州 310027
  • 收稿日期:2022-03-07 修回日期:2022-04-05 出版日期:2022-07-05 发布日期:2022-08-01
  • 通讯作者: 侯北平 E-mail:zhoule@zust.edu.cn;bphou@zust.edu.cn
  • 作者简介:周乐（1987—），男，博士，副教授，zhoule@zust.edu.cn
  • 基金资助:
    国家自然科学基金项目(62173306);浙江省重点研发计划项目(2022C04012)

Deep fusion feature extraction network and its application in chemical process soft sensing

Le ZHOU1(),Chengkai SHEN1,Chao WU1,Beiping HOU1(),Zhihuan SONG2   

  1. 1.School of Automation and Electrical Engineering, Zhejiang University of Science and Technology, Hangzhou 310024, Zhejiang, China
    2.State Key Laboratory of Industrial Control Technology, Zhejiang University, Hangzhou 310027, Zhejiang, China
  • Received:2022-03-07 Revised:2022-04-05 Published:2022-07-05 Online:2022-08-01
  • Contact: Beiping HOU E-mail:zhoule@zust.edu.cn;bphou@zust.edu.cn

RichHTML

6

PDF (PC)

83

摘要：

复杂化工过程的观测数据往往同时包含非线性和强动态特性，而传统的化工过程软测量方法无法准确提取观测数据的非线性动态特征，以至影响数据建模和质量预报的准确性。提出了一种基于变分自编码器的深度融合特征提取网络(deep fusion features extraction network， DFFEN)。在变分自编码器框架下，通过构建潜隐特征信息传递通道，提取非线性动态潜隐变量。并利用自注意力机制(self-attention)融合关键的隐层信息，优化因信息传递通道过长而导致的潜在特征被遗忘的问题。此外，在后端网络构建潜隐变量和关键质量变量之间的回归模型，以实现关键质量变量的预报。最后，通过数值案例和实际的合成氨过程验证了所提出的DFFEN模型的可行性和有效性。

关键词: 过程控制, 非线性动态建模, 神经网络, 深度融合特征, 合成气

Abstract:

The observation data of complex chemical processes often contain both nonlinear and strong dynamic characteristics, and the traditional soft sensing method of chemical process cannot accurately extract the nonlinear dynamic characteristics of the observation data, so as to affect the accuracy of data modeling and quality prediction. In this paper, a deep fusion features extraction network (DFFEN) is proposed. Under the framework of variational auto encoder, the nonlinear dynamic latent variables are extracted by constructing latent feature information transfer channel. In addition, a self-attention mechanism is used to fuse key hidden layer information and optimize the problem that the potential features are forgotten, which is mainly caused by the excessively long information transmission channel. Then, the regression model between latent variables and key quality variables is constructed in the backend network to achieve the prediction of key quality variables. Finally, the feasibility and effectiveness of the proposed DFFEN model are verified by numerical cases and an actual ammonia synthesis process.

Key words: process control, nonlinear dynamic modeling, neural networks, deep fusion feature extraction, syngas

中图分类号: 

  • TP 183

图1

DFFEN模型结构"

图2

基于DFFEN的软测量流程图"

图3

T与评价指标R2和RMSE的关系"

表1

数值案例不同模型预测结果"

ModelsR2RMSE
PPCR0.89442.0062
SSAE0.91871.7601
SNDS0.92331.7098
DFFEN
T=20.92021.7472
T=30.93431.5838
T=40.93641.5573
T=50.92821.6561

图4

数值案例中不同模型的软测量结果"

图5

各模型的预测误差"

图6

一段转化炉流程图"

表2

一段转化炉变量描述"

编号变量描述
FR03001流入03B001的燃气流量
FR03002流入03B001的外置燃气流量
PC0300203E005出口处燃气外置燃料压力
PC0300703B001炉膛出口烟气压力
TI0300103E005出口处的燃料放气温度
TI0300903B002E06出口燃气的温度
TR0301203B001入口处的加工气体温度
TI0301303B001左上角炉膛烟气温度
TI0301403B001右上角炉膛烟气温度
TR0301503B001炉顶混炉烟气温度
TR0301603B001左出口的转化气体温度
TR0301703B001右出口的转化气体温度
TR0302003B001出口转化气体的温度
AR03001炉顶氧浓度

表3

各模型在合成氨过程的预测结果"

ModelsR2RMSE
PPCR-0.32911.1737
SSAE0.53890.6913
SNDS0.11210.8833
DFFEN0.69460.5181

图7

各模型在合成氨过程的预测结果"

图8

各模型在合成氨过程中的预测误差"

1 陈忠圣, 朱梅玉, 贺彦林, 等. 基于分位数回归CGAN的虚拟样本生成方法及其过程建模应用[J]. 化工学报, 2021, 72(3): 1529-1538.
Chen Z S, Zhu M Y, He Y L, et al. Quantile regression CGAN based virtual samples generation and its applications to process modeling[J]. CIESC Journal, 2021, 72(3): 1529-1538.
2 李海波, 柴天佑, 岳恒. 浮选工艺指标KPCA-ELM软测量模型及应用[J]. 化工学报, 2012, 63(9): 2892-2898.
Li H B, Chai T Y, Yue H. Soft sensor of technical indices based on KPCA-ELM and application for flotation process[J]. CIESC Journal, 2012, 63(9): 2892-2898.
3 Kadlec P, Gabrys B, Strandt S. Data-driven soft sensors in the process industry[J]. Computers & Chemical Engineering, 2009, 33(4): 795-814.
4 Yuan X F, Huang B, Wang Y L, et al. Deep learning-based feature representation and its application for soft sensor modeling with variable-wise weighted SAE[J]. IEEE Transactions on Industrial Informatics, 2018, 14(7): 3235-3243.
5 Ge Z Q, Song Z H, Ding S X, et al. Data mining and analytics in the process industry: the role of machine learning[J]. IEEE Access, 2017, 5: 20590-20616.
6 Massy W F. Principal components regression in exploratory statistical research[J]. Journal of the American Statistical Association, 1965, 60(309): 234-256.
7 Ma M, Khatibisepehr S, Huang B. A Bayesian framework for real-time identification of locally weighted partial least squares[J]. AIChE Journal, 2015, 61(2): 518-529.
8 Fortuna L, Giannone P, Graziani S, et al. Virtual instruments based on stacked neural networks to improve product quality monitoring in a refinery[J]. IEEE Transactions on Instrumentation and Measurement, 2007, 56(1): 95-101.
9 Kaneko H, Funatsu K. Application of online support vector regression for soft sensors[J]. AIChE Journal, 2014, 60(2): 600-612.
10 Ku W F, Storer R H, Georgakis C. Disturbance detection and isolation by dynamic principal component analysis[J]. Chemometrics and Intelligent Laboratory Systems, 1995, 30(1): 179-196.
11 Dong Y N, Qin S J. Regression on dynamic PLS structures for supervised learning of dynamic data[J]. Journal of Process Control, 2018, 68: 64-72.
12 Chen J, Liu K C. On-line batch process monitoring using dynamic PCA and dynamic PLS models[J]. Chemical Engineering Science, 2002, 57(1): 63-75.
13 Ge Z Q, Chen X R. Dynamic probabilistic latent variable model for process data modeling and regression application[J]. IEEE Transactions on Control Systems Technology, 2019, 27(1): 323-331.
14 Everett B. An Introduction to Latent Variable Models[M]. Berlin:Springer Science & Business Media, 2013.
15 Dong Y N, Qin S J. Dynamic latent variable analytics for process operations and control[J]. Computers & Chemical Engineering, 2018, 114: 69-80.
16 Zhou L, Li G, Song Z H, et al. Autoregressive dynamic latent variable models for process monitoring[J]. IEEE Transactions on Control Systems Technology, 2017, 25(1): 366-373.
17 Zhou L, Zheng J Q, Ge Z Q, et al. Multimode process monitoring based on switching autoregressive dynamic latent variable model[J]. IEEE Transactions on Industrial Electronics, 2018, 65(10): 8184-8194.
18 Schölkopf B, Smola A, Müller K R. Nonlinear component analysis as a kernel eigenvalue problem[J]. Neural Computation, 1998, 10(5): 1299-1319.
19 Rosipal R, Trejo L J. Kernel partial least squares regression in reproducing kernel Hilbert space[J]. Journal of Machine Learning Research, 2001, 2: 97-123.
20 Kingma D P, Welling M. Auto-encoding variational Bayes[J/OL]. [2022-03-07]..
21 McCoy J T, Kroon S, Auret L. Variational autoencoders for missing data imputation with application to a simulated milling circuit[J]. IFAC-PapersOnLine, 2018, 51(21): 141-146.
22 Xie R M, Jan N M, Hao K R, et al. Supervised variational autoencoders for soft sensor modeling with missing data[J]. IEEE Transactions on Industrial Informatics, 2020, 16(4): 2820-2828.
23 Odiowei P P, Cao Y. Nonlinear dynamic process monitoring using canonical variate analysis and kernel density estimations[J]. IEEE Transactions on Industrial Informatics, 2010, 6(1): 36-45.
24 Yuan X F, Li L, Wang Y L. Nonlinear dynamic soft sensor modeling with supervised long short-term memory network[J]. IEEE Transactions on Industrial Informatics, 2020, 16(5): 3168-3176.
25 Yuan X F, Li L, Shardt Y A W, et al. Deep learning with spatiotemporal attention-based LSTM for industrial soft sensor model development[J]. IEEE Transactions on Industrial Electronics, 2021, 68(5): 4404-4414.
26 Yao L, Jiang X Y, Huang G P, et al. Virtual sensing f-CaO content of cement clinker based on incremental deep dynamic features extracting and transferring model[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-10.
27 Huang Y, Chen C H, Huang C J. Motor fault detection and feature extraction using RNN-based variational autoencoder[J]. IEEE Access, 2019, 7: 139086-139096.
28 Geng Z Q, Chen Z W, Meng Q C, et al. Novel transformer based on gated convolutional neural network for dynamic soft sensor modeling of industrial processes[J]. IEEE Transactions on Industrial Informatics, 2022, 18(3): 1521-1529.
29 Lee Y S, Ooi S K, Tanny D, et al. Developing soft-sensor models using latent dynamic variational autoencoders[J]. IFAC-PapersOnLine, 2021, 54(3): 61-66.
30 Yao L, Ge Z Q. Dynamic features incorporated locally weighted deep learning model for soft sensor development[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-11.
31 Shen B B, Ge Z Q. Supervised nonlinear dynamic system for soft sensor application aided by variational auto-encoder[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 69(9): 6132-6142.
32 Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need [C]//Advances in Neural Information Processing Systems 30 (NIPS 2017. LongBeach, CA, USA, 2017: 5998-6008.
33 Zhang M X, Yang Y, Ji Y L, et al. Recurrent attention network using spatial-temporal relations for action recognition[J]. Signal Processing, 2018, 145: 137-145.
[1] 王琨, 侍洪波, 谭帅, 宋冰, 陶阳. 局部时差约束邻域保持嵌入算法在故障检测中的应用[J]. 化工学报, 2022, 73(7): 3109-3119.
[2] 孙哲, 金华强, 李康, 顾江萍, 黄跃进, 沈希. 基于知识数据化表达的制冷空调系统故障诊断方法[J]. 化工学报, 2022, 73(7): 3131-3144.
[3] 赵涛岩, 曹江涛, 李平, 冯琳, 商瑀. 区间二型模糊免疫PID在环己烷无催化氧化温度控制系统中的应用[J]. 化工学报, 2022, 73(7): 3166-3173.
[4] 孟博, 刘艳萍, 蒋新科, 韩一帆. Fe5C2-MnO x 尺度调控及催化合成气制烯烃性能研究[J]. 化工学报, 2022, 73(6): 2677-2689.
[5] 戚子豪, 钟文琪, 陈曦, 周冠文, 赵小亮, 辛美静, 陈翼, 朱永长. 基于混合建模的水泥生料分解过程动态特性研究[J]. 化工学报, 2022, 73(5): 2039-2051.
[6] 王建松, 许锋, 罗雄麟. 化工过程多回路PID控制系统模式切换参数自整定[J]. 化工学报, 2022, 73(4): 1647-1657.
[7] 任嘉辉, 刘豫, 刘朝, 刘浪, 李莹. 基于分子指纹和拓扑指数的工质临界温度理论预测[J]. 化工学报, 2022, 73(4): 1493-1500.
[8] 袁聪, 蒲舸, 高杰, 贾帅辉. 改性BaFe2O4载氧体生物质化学链气化特性[J]. 化工学报, 2022, 73(3): 1359-1368.
[9] 刘立邦, 杨颂, 王志坚, 贺欣欣, 赵文磊, 刘守军, 杜文广, 米杰. 基于改进WOA-LSTM的焦炭质量预测[J]. 化工学报, 2022, 73(3): 1291-1299.
[10] 张成, 潘立志, 李元. 基于加权统计特征KICA的故障检测与诊断方法[J]. 化工学报, 2022, 73(2): 827-837.
[11] 张兴硕, 罗雄麟, 许锋. 催化裂化装置反再系统动态模拟精细化与控制系统“工艺优先”配对设计[J]. 化工学报, 2022, 73(2): 747-758.
[12] 袁壮, 凌逸群, 杨哲, 李传坤. 基于TA-ConvBiLSTM的化工过程关键工艺参数预测[J]. 化工学报, 2022, 73(1): 342-351.
[13] 王乾浩, 赵璐, 孙付琳, 房克功. ZSM-5催化剂与低温等离子体协同转化H2S-CO2制合成气[J]. 化工学报, 2022, 73(1): 255-265.
[14] 唐晓婕, 杨博, 李宏光. 复杂化工过程调控操纵策略的深度学习方法[J]. 化工学报, 2021, 72(9): 4830-4837.
[15] 颜建国, 郑书闽, 郭鹏程, 张博, 毛振凯. 基于GA-BP神经网络的超临界CO2传热特性预测研究[J]. 化工学报, 2021, 72(9): 4649-4657.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 2019 《化工学报》编辑部
北京市东城区青年湖南街13号（100011）
电话：010-64519489，010-64519362，010-64519485，010-64519490，010-64519451
E-Mail：hgxb@cip.com.cn
京ICP备12046843号-7
京公网安备 11010102001995号
本系统由北京玛格泰克科技发展有限公司设计开发