Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)
计算机科学技术     
基于模型参数聚类的分解炉温度建模
张进锋, 杨强, 颜文俊
浙江大学 系统科学与工程学系,浙江 杭州 310027
Model parameter clustering based modeling approach for calciner temperature
ZHANG Jin-feng, YANG Qiang, YAN Wen-jun
Department of System Science and Engineering, Zhejiang University, Hangzhou 310027, China
 全文: PDF(1054 KB)   HTML
摘要：

针对流程工业多工况、非线性特性所导致的建模困难问题,提出一种基于模型参数聚类的预测建模方法.结合员工操作、环境变化以及数据特征等因素选择涵盖多种工况的典型历史数据；根据典型历史数据,利用受限最小二乘法,分段建立多个脉冲响应模型；以各模型参数为特征,采用K-均值方法对各分段模型进行子空间聚类,生成K类聚类模型；在实际控制阶段,根据校正预测效果选择合适的聚类模型,并采用该模型进行实时控制.研究结果表明：该聚类建模方法能反映分解炉的运行状态,预测效果好,鲁棒性强,能够适应多种工况；应用该模型对分解炉温度进行实时控制可获得满意的效果.
Abstract:

A predictive modeling approach based on model parameter clustering was presented to solve the modeling problems brought out by the multi-point and nonlinear nature of the process industry. According to the operation on calciner, environment changes and data feature, typical historical measurement data that covers various operating modes are chosen. Based on the obtained typical historical measurement data, multiple piecewise impulse response models are established by the use of constraint least squares method. With the parameters of the piecewise models, K-means method is adopted to the process subspace clustering against each piecewise model, and K clustering models are achieved. In the practical control phase, the appropriate model is chosen to carry out the real-time control actions based on the correction prediction result. The proposed solution has been examined through a set of simulation experiments, and the numerical result demonstrates that the suggested modeling can perform well with satisfactory prediction, strong robustness, as well as adaptability to a range of operational scenarios. In addition, the modeling method is adopted for the calciner’s real-time control, and the result demonstrates that the calciner temperature can be constrained within a very small range.
出版日期: 2014-12-01
:  TP 279  
基金资助:

国家“863”高技术研究发展计划资助项目（2012AA051704）
通讯作者: 颜文俊,男,教授,博导     E-mail: yanwenjun@zju.edu.cn
作者简介: 张进锋（1983—）,男,博士生,从事工业控制与人工智能研究. E-mail: zhangjinfeng01@163.com
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章　　

引用本文:

张进锋, 杨强, 颜文俊. 基于模型参数聚类的分解炉温度建模[J]. 浙江大学学报(工学版), 10.3785/j.issn.1008-973X.2014.12.006.

ZHANG Jin-feng, YANG Qiang, YAN Wen-jun. Model parameter clustering based modeling approach for calciner temperature. JOURNAL OF ZHEJIANG UNIVERSITY (ENGINEERING SCIENCE), 10.3785/j.issn.1008-973X.2014.12.006.

链接本文:

http://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2014.12.006    或    http://www.zjujournals.com/eng/CN/Y2014/V48/I12/2139

［1］ MAYNE D Q, RAWLINGS J B, RAO C V, et al. Constrained modelpredictivecontrol: stability and optimality ［J］. Automatica, 2000, 36(6): 789-814.
［2］ GANG F. A survey on analysis and design of model-basedfuzzycontrolsystems ［J］. IEEE Transactions on Fuzzy Systems, 2006, 14(5): 676-697.
［3］ YU H L, LIU W L, DONG H J.Research on recognition of working condition for calciner and grate cooler based onexpertsystem ［C］∥ 12th International Conference on Control, Automation, Robotics and Vision. \[S.l.\]: IEEE, 2012: 1733-1737.
［4］ FIDAROS D K, BAXEVANOU C A, DRITSELIS C D, et al.Numerical modeling of flow and transport processes in acalcinerfor cement production ［J］. Powder Technology, 2007, 17(12): 81-95.
［5］ MUJUMDAR K S, GANESH K V, KULKARNI S B, et al. Rotary cement kiln simulator RoCKS): Integrated modeling of pre-heater,calciner, kiln and clinker cooler ［J］. Chemical Engineering Science, 2007, 62(9): 2590-2607.
［6］ 梅书霞.水泥分解炉结构参数优化与煤粉燃烧的数值模拟［D］.武汉：武汉理工大学, 2008: 86-112.
MEI Shu-xia. Numerical simulations of gas-solid flow field and coal combustion in precalciners of cement industry for optimization ［D］. Wuhan: Wuhan University of Technology, 2008: 86-112.
［7］ SAIDUR R, HOSSAIN M S, ISLAM M R, et al. Areviewonkilnsystemmodeling ［J］. Renewable & Sustainable Energy Reviews, 2011, 15(5): 2487-2500.
［8］ 邹健,诸静.模糊预测函数控制在水泥回转窑分解炉温控系统中的应用研究［J］.硅酸盐学报, 2001, 29(4): 318-321.
ZOU Jian, ZHU Jing. Fuzzy predictive functional control strategy for decomposing furnace temperature system of cement rotary kiln ［J］. Journal of the Chinese Ceramic Society, 2001, 29(4): 318-321.
［9］ KOUMBOULIS F N, KOUVAKAS N D. Indirect adaptive neural control for precalcination in cement plants ［J］. Mathematics and Computers in Simulation, 2002, 60(3-5): 325-334.
［10］ 熊会思.新型干法烧成水泥熟料设备设计、制造、安装与使用［M］.北京:中国建材工业出版社, 2004: 143.
［11］ ZHANG Jin-feng, PANG Bo, CAI Ning, et al. Predictive modeling of cement decomposing furnace temperature ［C］∥ Intelligent Control and Automation. ［S.l.］:IEEE, 2010: 1552-1556.
［12］ 董会君.水泥生产过程预分解系统工况识别的研究［D］.济南: 济南大学, 2012: 24-29.
DONG Hui-jun. Research of working recognition for cement production precalcining system ［D］. Jinan: University of Jinan, 2012: 24-29.
［13］ FAYYAD U, REINA C, BRADLEY P S. Initialization of iterative refinement clustering algorithms［C］∥ Proc of the Fourth International Conference on Knowledge Discovery and Data Mining. ［S.l.］: AAAI, 1998: 194-198.
No related articles found!