Atmospheric tower soft sensor based on regression and mutual information of kernel slow features

doi:10.11949/j.issn.0438-1157.20161395

Abstract

Abstract:

A novel soft sensor method based on slow feature regression (SFR) was proposed for industrial process with nonlinear and dynamic characteristics. First, a dynamic dataset was built by adding time-delay data and information redundancy was reduced by selecting variables according to mutual information maximization criteria. Then, kernel function was introduced into slow feature analysis(SFA)to improve capability of processing nonlinear data and the kernel slow features were used for regression. Through analysis of sample variation, kernel slow feature analysis(KSFA)could extract components with slowly varying dynamics, characterize trend of industrial process effectively, and improve precision of regression modelling. Finally, effectiveness and feasibility of the proposed method were verified by soft sensor model of constant top oil dry point and constant first line dropping point in atmospheric tower.

Key words: slow feature analysis, mutual information, dynamic modeling, atmospheric tower, petroleum, prediction

摘要：

针对工业过程的非线性及动态特性，提出了一种新的慢特征回归软测量方法。该方法首先通过添加时延数据构造动态数据集，利用互信息最大化准则筛选变量从而减少信息冗余的影响。同时该方法在慢特征分析的基础上引入核函数扩展，加强模型处理非线性数据的能力，并将获得的核慢特征用于回归建模。核慢特征分析通过分析样本的变化，提取具有缓慢变化特征的成分，可以有效地刻画工业过程的变化趋势，提升回归模型精度。最后该方法的有效性在常压塔常顶油干点与常一线初馏点的软测量模型中得到了验证。

关键词: 慢特征分析, 互信息, 动态建模, 常压塔, 石油, 预测

CLC Number:

TP274

JIANG Xinyi, DU Hongbin, LI Shaojun. Atmospheric tower soft sensor based on regression and mutual information of kernel slow features[J]. CIESC Journal, 2017, 68(5): 1977-1986.

蒋昕祎, 杜红彬, 李绍军. 基于核慢特征回归与互信息的常压塔软测量建模[J]. 化工学报, 2017, 68(5): 1977-1986.

References 31

[1]	SHANG C, YANG F, HUANG D, et al. Data-driven soft sensor development based on deep learning technique[J]. Journal of Process Control, 2014, 24(3): 223-233.
[2]	LI Q, DU Q, BA W, et al. Multiple-input multiple-output soft sensors based on KPCA and MKLS-SVM for quality prediction in atmospheric distillation column [J]. International Journal of Innovative Computing, Information and Control, 2012, 8(12): 8215-8230.
[3]	NAPOLI G, XIBILIA M G. Soft sensor design for a topping process in the case of small datasets[J]. Computers & Chemical Engineering, 2011, 35(11): 2447-2456.
[4]	GALICIA H J, HE Q P, WANG J. A reduced order soft sensor approach and its application to a continuous digester[J]. Journal of Process Control, 2011, 21(4): 489-500.
[5]	袁小锋, 葛志强, 宋执环. 基于时间差分和局部加权偏最小二乘算法的过程自适应软测量建模[J]. 化工学报, 2016, 67(3): 724-728. YUAN X F, GE Z Q, SONG Z H. Adaptive soft sensor based on time difference model and locally weighted partial least squares regression[J]. CIESC Journal, 2016, 67(3): 724-728.
[6]	YUAN X F, HUANG B, GE Z Q, et al. Double locally weighted principal component regression for soft sensor with sample selection under supervised latent structure[J]. Chemometrics & Intelligent Laboratory Systems, 2016, 153: 116-125.
[7]	YU J, CHEN K, MORI J, et al. A Gaussian mixture copula model based localized Gaussian process regression approach for long-term wind speed prediction[J]. Energy, 2013, 61(6): 673-686.
[8]	YANG K, JIN H P, CHEN X G, et al. Soft sensor development for online quality prediction of industrial batch rubber mixing process using ensemble just-in-time Gaussian process regression models[J]. Chemometrics & Intelligent Laboratory Systems, 2016, 155: 170-182.
[9]	刘国海, 苏勇, 杨铭, 等. 基于多准则和高斯过程回归的动态软测量建模方法[J]. 东南大学学报(自然科学版), 2015, (6): 1086-1090. LIU G H, SU Y, YANG M, et al. Dynamic soft sensor modeling based on multi-criterion method and Gaussian process regression [J]. Journal of Southeast University (Natural Science Edition), 2015, (6): 1086 -1090.
[10]	阮宏镁, 田学民, 王平. 基于联合互信息的动态软测量方法[J]. 化工学报, 2014, 65(11): 4497-4502. RUAN H M, TIAN X M, WANG P. Dynamic soft sensor method based on joint mutual information [J]. CIESC Journal, 2014, 65(11): 4497-4502.
[11]	WISKOTT L, SEJNOWSKI T. Slow feature analysis: unsupervised learning of invariances[J]. Neural Computation, 2002, 14(4): 715- 770.
[12]	WU C, DU B, ZHANG L. Slow feature analysis for change detection in multispectral imagery[J]. IEEE Transactions on Geoscience & Remote Sensing, 2014, 52(5): 2858-2874.
[13]	FRANZIUS M, WILBERT N, WISKOTT L. Invariant object recognition and pose estimation with slow feature analysis[J]. Neural Computation, 2011, 23(9): 2289-2323.
[14]	ZHANG Z, TAO D. Slow feature analysis for human action recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(3): 436-450.
[15]	MINH H Q, WISKOTT L. Multivariate slow feature analysis and decorrelation filtering for blind source separation.[J]. Image Processing IEEE Transactions on, 2013, 22(7): 2737-2750.
[16]	何会会, 李钢虎, 要庆生, 等. 用慢特征分析算法实现水声信号盲分离[J]. 声学技术, 2014, 33(3): 270-274. HE H H, LI G H, YAO Q S, et al.Blind source separation of underwater acoustic signals by using slowness feature analysis[J].Technical Acoustics, 2014, 33(3): 270-274.
[17]	SHANG C, HUANG B, YANG F, et al. Slow feature analysis for monitoring and diagnosis of control performance[J]. Journal of Process Control, 2016, 39: 21-34.
[18]	SHANG C, YANG, GAO X Q, et al. Concurrent monitoring of operating condition deviations and process dynamics anomalies with slow feature analysis[J]. AIChE Journal, 2015, 61(11): 3666-3682.
[19]	ZHANG H Y, TIAN X M, CAI L F. Nonlinear process fault diagnosis using kernel slow feature discriminant analysis[J]. IFAC-Papers on Line, 2015, 48(21): 607-612.
[20]	SHANG C, YANG F, GAO X Q, et al. Extracting latent dynamics from process data for quality prediction and performance assessment via slow feature regression[C]//American Control Conference (ACC). Chicago: IEEE, 2015: 912-917.
[21]	SHANG C, HUANG B, YANG F, et al. Probabilistic slow feature analysis-based representation learning from massive process data for soft sensor modeling[J]. AIChE Journal, 2015, 61(12): 4126-4139.
[22]	KONEN W, KOCH P. The slowness principle: SFA can detect different slow components in non-stationary time series[J]. International Journal of Innovative Computing & Applications, 2011, 3(3): 3-10.
[23]	王桂增, 叶昊. 主元分析与偏最小二乘法[M].北京: 清华大学出版社, 2012: 106-107. WANG G Z, YE H. Principal Component Analysis and Partial Least Squares [M].Beijing: Tsinghua University Press, 2012: 106-107.
[24]	BOHMER W, GRUNEWALDER S, NICKISCH H, et al. Generating feature spaces for linear algorithms with regularized sparse kernel slow feature analysis[J]. Machine Learning, 2012, 89(1/2): 67-86.
[25]	SOUZA F, SANTOS P, ARAUJO R. Variable and delay selection using neural networks and mutual information for data-driven soft sensors [C]//Emerging Technologies and Factory Automation (ETFA).Spain: IEEE, 2010: 1-8.
[26]	MODDEMEIJER R. On estimation of entropy and mutual information of continuous distributions[J]. Signal Processing, 1989, 16(3): 233-248.
[27]	童楚东, 蓝艇, 史旭华. 基于互信息的分散式动态PCA故障检测方法[J]. 化工学报, 2016, 67(10): 4317-4323. TONG C D, LAN T, SHI X H.Fault detection by decentralized dynamic PCA algorithm on mutual information[J].CIESC Journal, 2016, 67(10): 4317-4323.
[28]	JIN H P, CHEN X G, YANG J W, et al. Adaptive soft sensor modeling framework based on just-in-time learning and kernel partial least squares regression for nonlinear multiphase batch processes[J]. Computers & Chemical Engineering, 2014, 71: 77-93.
[29]	BOLN-CANEDO V, SNCHEZ-MAROO N, ALONSO-BETANZOS A. Feature Selection for High-Dimensional Data[M]. Springer Publishing Company, Incorporated, 2015: 17-24.
[30]	金思毅, 李悦卿, 夏茂森. 常减压装置常压塔塔顶汽油干点的软测量[J]. 化工进展, 2006, 25(s1): 74-76. JIN S Y, LI Y Q, XIA M S. Curde colmun gasoline end point soft-sensing of atmospheric and vacuum unit[J].Chemical Industry and Engineering Progress, 2006, 25(s1): 74-76.
[31]	唐孟海, 胡兆灵. 原油蒸馏[M]. 北京: 中国石化出版社, 2007: 44-45. TANG M H, HU Z L.Crude Oil Distillation [M]. Beijing: China Petro-Chemical Press, 2007: 44-45.