Quality pattern monitoring for polymerization process based on Bayesian statistical learning

doi:10.16085/j.issn.1000-6613.2018-0595

Abstract

Abstract: Due to the strong non-linearity and complexity of mechanisms, the quality control of the polymerization process becomes a difficult and hot topic in the field of process control. By combining the characteristics of the polymerization process and using the process parameter data related to the physical properties of the polymer produced in the process of polymerization production, the concept of quality pattern monitoring was firstly introduced. This paper innovatively proposed a Bayesian statistical learning method based on pattern index to solve the quality pattern monitoring problem of the polymerization process. Firstly, the principal component analysis was used to project the essential features of the observational spatial data into the low-dimensional space to obtain the pattern index. Then, the Bayesian statistical learning method was applied to distinguish the constructed quality pattern based on posterior probability. Finally the proposed method was verified using the production data of the polymerization reactor provided by a chemical plant. Compared with the quality index such as conversion rate and polymerization rate, the pattern index can better reflect the consistency and product quality of the polymerization process. Therefore, the Bayesian discriminant method based on the pattern index has higher accuracy in comparison with the process parameters, and it is more effective on quality monitoring for polymerization process.

Key words: polymerization process, pattern monitoring, quality control, Bayesian, principal component analysis, experimental validation

摘要： 聚合过程由于反应机理复杂、过程非线性强等特点，使得其质量监控问题成为过程控制领域的研究难点和热点。本文结合聚合反应过程的特点，利用聚合生产过程中与聚合物的物性参数相关的过程参数数据，引入质量模式监测的概念，创新性地提出了基于模式指标的贝叶斯统计学习方法，以解决聚合过程的质量模式监测问题。首先利用主成分分析将观测空间数据的本质特征投影到低维空间，得到模式指标；然后，基于贝叶斯统计学习算法，根据后验概率对构建的质量模式进行判别；最后利用某化工厂提供的聚合反应釜的生产数据对所提方法进行验证。由于模式指标相比于质量指标如转化率、聚合速率等，更能反映聚合反应过程的一致性和产品品质，因此基于模式指标的贝叶斯判别法较基于参数指标的贝叶斯判别法更加准确，更能实现对聚合过程的质量监控。

关键词: 聚合过程, 模式监测, 质量监控, 贝叶斯, 主元分析, 实验验证

CLC Number:

TQ325.3

GAO Shuang, ZHENG Niannian, LUAN Xiaoli, LIU Fei. Quality pattern monitoring for polymerization process based on Bayesian statistical learning[J]. Chemical Industry and Engineering Progress, 2018, 37(12): 4558-4564.

高爽, 郑年年, 栾小丽, 刘飞. 聚合过程贝叶斯统计学习质量模式监测[J]. 化工进展, 2018, 37(12): 4558-4564.

References

[1] BINDLISH R. Nonlinear model predictive control of an industrial polymerization process[J]. Computers & Chemical Engineering, 2015, 73:43-48.
[2] LUCIA S, FINKLER T, ENGELL S. Multi-stage nonlinear model predictive control applied to a semi-batch polymerization reactor under uncertainty[J]. Journal of Process Control, 2013, 23(9):1306-1319.
[3] SMITH D,GLEN K,THOMAS R. Automated image analysis with the potential for process quality control applications in stem cell maintenance and differentiation[J]. Biotechnology Progress, 2015, 32(1):215-223.
[4] HOSEN M A, KHOSRAVI A, NAHAVANDI S, et al. Prediction interval-based neural network modelling of polystyrene polymerization reactor——a new perspective of data-based modelling[J]. Chemical Engineering Research & Design, 2014, 92(11):2041-2051.
[5] LUO L, ZHANG N, XIA Z, et al. Dynamics and stability analysis of gas-phase bulk polymerization of propylene[J]. Chemical Engineering Science, 2016, 143:12-22.
[6] NELE M, PINTO J C. Dynamic behavior of a continuous autothermal isobutylene polymerization reactor[J]. Journal of Applied Polymer Science, 2015, 65(7):1403-1413.
[7] ZHOU L, CHEN J, SONG Z, et al. Probabilistic latent variable regression model for process-quality monitoring[J]. Chemical Engineering Science, 2014, 116:296-305.
[8] YIN J, ZHAN X, ZHENG Y, et al. Impact of quality control of satellite soil moisture data on their assimilation into land surface model[J]. Geophysical Research Letters, 2015, 41(20):7159-7166.
[9] HOSEN M A, HUSSAIN M A, MJALLI F S, et al. Performance analysis of three advanced controllers for polymerization batch reactor:an experimental investigation[J]. Chemical Engineering Research & Design, 2014, 92(5):903-916.
[10] JIANG B, WENG Y, ZHANG S, et al. Kinetics and mechanism of ethylene polymerization with TiCl₄/MgCl₂, model catalysts:effects of titanium content[J]. Journal of Catalysis, 2018, 360:57-65.
[11] 苏鑫, 吴迎亚, 裴华健, 等. 大数据技术在过程工业中的应用研究进展[J]. 化工进展, 2016, 35(6):1652-1659. SU X, WU Y Y, PEI H J, et al. Development in the application of big data technology in process industry.[J]. Chemical Industry and Engineering Progress, 2016, 35(6):1652-1659.
[12] WANG Z, LIU D. Stability analysis for a class of systems:from model-based methods to data-driven methods[J]. IEEE Transactions on Industrial Electronics, 2014, 61(11):6463-6471.
[13] ROUBINET D, TARTAKOVSKY D M. Hybrid modeling of heterogeneous geochemical reactions in fractured porous media[J]. Water Resources Research, 2013, 49(12):7945-7956.
[14] NASCIMENTO C, GIUDICI R, SCHERBAKOFF N. Modeling of industrial nylon-6,6 polymerization process in a twin-screw extruder reactor. Ⅱ. Neural networks and hybrid models[J]. Journal of Applied Polymer Science, 2015, 72(7):905-912.
[15] WILLIAMS B J, COLE B. Mining monitored data for decision-making with a Bayesian network model[J]. Ecological Modelling, 2013, 249:26-36.
[16] KAYABOL K, KUTLUK S. Bayesian classification of hyperspectral images using spatially-varying Gaussian mixture model[J]. Digital Signal Processing, 2016, 59:106-114.
[17] WANG X Z, HE Y L, WANG D D. Non-naive Bayesian classifiers for classification problems with continuous attributes[J]. IEEE Transactions on Cybernetics, 2013, 44(1):21-39.
[18] SANTOS F L C D, PACI M, NANNI L, et al. Computer vision for virus image classification[J]. Biosystems Engineering, 2015, 138:11-22.
[19] PROCHAZKA A, VYSATA O, VALIS M, et al. Bayesian classification and analysis of gait disorders using image and depth sensors of microsoft kinect[J]. Digital Signal Processing, 2015, 47:169-177.
[20] PAGE G, BHATTACHARYA A, DUNSON D. Classification via Bayesian nonparametric learning of affine subspaces[J]. Journal of the American Statistical Association, 2013, 108(501):187-201.
[21] JIANG Q, WANG B, YAN X. Multi-block independent component analysis integrated with Hellinger distance and Bayesian inference for non-Gaussian plant-wide process monitoring[J]. Industrial & Engineering Chemistry Research, 2015, 54(9):2497-2508.
[22] JIANG Q, YAN X, HUANG B. Performance-driven distributed PCA process monitoring based on fault-relevant variable selection and Bayesian inference[J]. IEEE Transactions on Industrial Electronics, 2015, 63(1):377-386.
[23] TONG C, SONG Y, YAN X. Distributed statistical process monitoring based on four-subspace construction and Bayesian inference[J]. Industrial & Engineering Chemistry Research, 2013, 52(29):9897-9907.
[24] CHENG Z, LIU X. Optimal online soft sensor for product quality monitoring in propylene polymerization process[J]. Neurocomputing, 2015, 149(3):1216-1224.
[25] 朱鹏飞, 夏陆岳, 潘海天, 基于软测量技术的间歇聚合过程质量控制[J]. 计算机与应用化学, 2015, 32(8):959-963. ZHU P F, XIA L Y, PAN H T. Quality control for batch polymerization process based on soft sensor technology[J]. Computers and Applied Chemistry, 2015, 32(8):959-963.
[26] JIANG B, ZHU X, HUANG D, et al. A combined canonical variate analysis and Fisher discriminant analysis (CVA-FDA) approach for fault diagnosis[J]. Computers & Chemical Engineering, 2015, 77:1-9.
[27] CHENG J, GREINER R. Comparing Bayesian network classifiers[J]. IEEE Transactions on Vehicular Technology, 2013, 63(5):2002-2012.
[28] LIU Z J, GUO S L, LI T Y, et al. Comparative study of Bayesian probabilistic flood forecasting models[J]. Journal of Hydraulic Engineering, 2014, 45(9):1019-1028.
[29] OHSHIMA M, TANIGAKI M. Quality control of polymer production processes[J]. Journal of Process Control, 2000, 10(2):135-148.
[30] COBLE J B, FRAGA C G. Comparative evaluation of preprocessing freeware on chromatography/mass spectrometry data for signature discovery[J]. Journal of Chromatography A, 2014, 1358:155-164.