Multi-model soft sensor for hydrogen purity in catalytic reforming process based on improved fast search clustering algorithm and Gaussian processes regression

doi:10.11949/j.issn.0438-1157.20151854

Abstract

Abstract:

Hydrogen is one of the most important by-products in catalytic reforming process, a hydrogen purity soft sensor will contribute to guiding production. However, the working condition of catalytic reforming process is complex and changeable, a single model soft sensor is hard to ensure the prediction accuracy. Aiming at this problem, this paper present a combined soft sensor model based on modified fast search clustering algorithm and Gaussian processes regression (GPR). The history sample are classified by the novel clustering algorithm and then each sub-model is built through GPR with the classified sub sample. Meanwhile the class identification model has been built by GPR as well. Finally, the combined model soft sensor is established in a switcher form. The combined is applied to a catalytic reformer and the result indicates that the proposed method has a good result and has certain practical value.

Key words: catalytic reforming, hydrogen, model, algorithm, clustering by fast search, Gaussian processes regression, soft sensor

摘要：

氢气是催化重整反应的重要副产物之一,建立氢气纯度软测量模型有助于指导生产。针对催化重整过程工况复杂多变、单一软测量模型难以满足精度要求,提出了一种基于改进的快速搜索聚类算法和高斯过程回归的多模型软测量建模方法。首先,针对快速搜索聚类算法中截断距离是由人为设定的问题,提出了一种截断距离确定方法。并用该改进算法对历史数据进行自动分类,建立各个数据子集的高斯过程回归模型,使各子模型在最大程度上反映不同工况点。然后,针对聚类后得到的带有类别标签的历史数据,建立类别辨识模型,与各子模型相结合,形成开关模式的组合模型。最后,将该建模方法应用于连续催化重整装置,建立了脱氯前氢气纯度的在线计算模型。结果表明,该多模型建模方法具有较高的预测精度,优于传统的单一模型,有一定的实用价值。

关键词: 催化重整, 氢气, 模型, 算法, 快速搜索聚类, 高斯过程回归, 软测量

CLC Number:

TP274

SHUANG Yifan, GU Xingsheng. Multi-model soft sensor for hydrogen purity in catalytic reforming process based on improved fast search clustering algorithm and Gaussian processes regression[J]. CIESC Journal, 2016, 67(3): 765-772.

双翼帆, 顾幸生. 基于改进的快速搜索聚类算法和高斯过程回归的催化重整脱氯前氢气纯度多模型建模方法[J]. 化工学报, 2016, 67(3): 765-772.

References 20

[1]	LIU J L. On-line soft sensor for polyethylene process with multiple production grades[J]. Control Engineering Practice, 2007, 15(7): 769-778.
[2]	唐志杰, 唐朝辉, 朱红求. 一种基于多模型融合软测量建模方法[J]. 化工学报, 2011, 62(8): 2248-2252. DOI: 10.3969/j.issn. 0438-1157. 2011.08.028. TANG Z J, TANG Z H, ZHU H Q. A multi-model fusion soft sensor modeling method[J]. CIESC Journal, 2011, 62(8): 2248-2252. DOI: 10.3969/j.issn.0438-1157.2011.08.028.
[3]	陈贵华, 王昕, 王振雷, 等. 基于模糊核聚类的乙烯裂解深度DE-LSSVM多模型建模[J]. 化工学报, 2012, 63(6): 1790-1796. DOI: 10.3969/j.issn.0438-1157.2012.06.019 CHEN G H, WANG X, WANG Z L, et al. Multiple DE-LSSVM modeling of ethylene cracking severity based on fuzzy kernel clustering[J]. CIESC Journal, 2011, 62(8): 2248-2252. DOI: 10.3969/j.issn. 0438-1157.2012.06.019
[4]	杨慧中, 张文清. 基于特征加权模糊聚类的多模型软测量建模[J]. 控制工程, 2011, 18(4):524-526. DOI: 10.3969/j.issn.1671-7848. 2011.04.009. YANG H Z, ZHANG W Q. Multi-model soft-sensor modeling based on feature-weighted fuzzy clustering[J]. Control Engineering of China, 2011, 18(4): 524-526. DOI: 10.3969/j.issn.1671-7848. 2011. 04.009.
[5]	刘靖明, 韩丽川, 侯立文. 基于粒子群的K均值聚类算法[J]. 系统工程理论与实践, 2005, 25(6): 54-58. LIU J M, HAN L C, HOU L W. Cluster analysis based on particle swarm optimization algorithm[J]. Systems Engineering-Theory & Practice, 2005, 25(6): 54-58.
[6]	SUYKENS J A K, VANDERWALLE J. Least squares support vector machine classifiers[J]. Neural Processing Letters, 1999, 9(3): 293-300.
[7]	钱晓山, 阳春华, 徐丽莎. 基于改进差分进化和最小二乘支持向量机的铝酸钠溶液浓度软测量[J]. 化工学报, 2013, 64(5): 1704-1709. DOI: 10.3969/j.issn.0438-1157.2013.05.027. QIAN X S, YANG C H, XU L S. Soft sensor of sodium aluminate solution concentration based on improved differential evolution algorithm and LSSVM[J]. CIESC Journal, 2013, 64(5): 1704-1709. DOI: 10.3969/j.issn.0438-1157.2013.05.027.
[8]	PATIL P, SHARMA S C, PALIWAL V, et al. ANN modelling of Cu type omega vibration based mass flow sensor[J]. Procedia Technology, 2014, 14(10): 260-265.
[9]	罗健旭, 邵惠鹤. 应用多神经网络建立动态软测量模型[J]. 化工学报, 2003, 54(12):1770-1773. LUO J X, SHAO H H. Development dynamic soft sensors using multiple neural networks[J]. Journal of Chemical Industry and Engineering(China), 2003, 54(12): 1770-1773.
[10]	吴瑶, 罗雄麟, 袁志宏. 多频率系统动态插值神经网络软测量建模[J]. 化工进展, 2009, 28(8):1323-1327. WU Y, LUO X L, YUAN Z H. Soft sensor modeling with dynamic interpolation neural network for multirate system[J]. Chemical Industry and Engineering Progress, 2009, 28(8): 1323-1327.
[11]	李雅芹, 杨慧中. 一种基于Bagging算法的高斯过程集成建模方法[J]. 东南大学学报(自然科学版), 2011, 41(z1):93-96. DOI: 10.3969/j.issn.1001-0505.2011.S1.020. LI Y Q, YANG H Z. Ensemble modeling method based on Bagging algorithm and Gaussian process[J]. Journal of Southeast University (Natural Science Edition), 2011, 41(z1): 93-96. DOI: 10.3969/j.issn. 1001-0505.2011.S1.020.
[12]	雷瑜, 杨慧中. 基于高斯过程和贝叶斯决策的组合模型软测量[J]. 化工学报, 2013, 64(12): 4434-4438. DOI: 10.3969/j.issn.0438-1157. 2013.12.025. LEI Y, YANG H Z. Combination model soft sensor based on Gaussian process and Bayesian committee machine[J]. CIESC Journal, 2013, 64(12): 4434-4438. DOI: 10.3969/j.issn.0438-1157. 2013.12.025.
[13]	ALEX R, ALESSANDRO L. Clustering by fast search and find of density peaks[J]. Science, 2014, 344(6191): 1492-1496.
[14]	WILLIAMS C K I, RASMUSSEN C E. Gaussian Processes for Regression[M]. Cambridge: MIT Press, 2006.
[15]	王华忠. 高斯过程及其在软测量建模中的应用[J]. 化工学报, 2007, 58(11):2840-2845. DOI: 10.3321/j.issn:0438-1157.2007.11.026. WANG H Z. Gaussian process and its application to soft-sensor modeling[J]. Journal of Chemical Industry and Engineering(China), 2007, 58(11): 2840-2845. DOI: 10.3321/j.issn: 0438-1157. 2007. 11.026.
[16]	罗妍. 再接触条件对重整产品收率的影响[J]. 炼油技术与工程, 2013, 43(4):18-23.DOI: 10.3969/j.issn.1002-106X.2013.04.005. LUO Y. Study of impact of re-contacting conditions on catalytic reformer products[J]. Petroleum Refinery Engineering, 2013, 43(4): 18-23. DOI: 10.3969/j.issn.1002-106X.2013.04.005.
[17]	HU J M, WANG J Z. Short-term wind speed prediction using empirical wavelet transform and Gaussian process regression[J]. Energy, 2015, 93(2): 1456-1466.
[18]	LI X L, SU H Y, CHU J. Multiple model soft sensor based on Affinity propagation, Gaussian process and Bayesian committee machine[J]. Chinese Journal of Chemical Engineering, 2009, 17(1): 95-99.
[19]	XIONG Z H, HUANG G H, SHAO H H. soft sensor modeling based on Gaussian process[J]. Journal of Central South University of Technology, 2005, 12(4): 469-471.
[20]	WILLIAMS C K I. Prediction with Gaussian process: from linear regression to linear prediction and beyond[M]//JORDAN M I. Learning and Inference in Graphical Models. Cambridge: MIT Press, 1999: 599-621.