基于web的水泥预分解窑系统仿真系统的研究
|
摘要
近年来，我国已经建成许多新型干法水泥生产线，这些生产线技术新、自动化水平高，这就意味着水泥行业需要不断培训出大批合格的生产操作人员和工程技术人员。
本文所研究的系统为武汉理工大学“211工程”重点建设项目——水泥生产过程计算机仿真系统(CPPCSS)的子课题之一：水泥窑系统仿真系统。围绕此课题展开，本文主要包括以下内容：
第一，作者通过查阅大量国内外文献资料和实地参观考察等方式对水泥生产工艺技术(主要为窑系统生产过程)及其计算机仿真技术的发展现状进行了研究分析。
第二，根据水泥回转窑系统具有多变量、多回路、非线性、分布复杂和经典数学模型难以建立的特点，运用神经网络与专家系统集成技术对预分解窑系统进行了建模与研究。提出了对传统EBP神经网络算法的改进方法。
本文采用ES(专家系统)与ANN(人工神经网络)集成技术，利用ANN优良的自组织、自学习和自适应能力来解决ES知识获取的“瓶颈问题”，同时运用ES良好的解释机能来弥补ANN中知识表达的“黑箱结构”。将存于ANN内部的系统知识(数据化的结点和权值)运用统计的方法进行规则抽取，使其转换成显示的规则知识并进行解释。
建模过程中，输入输出参数的选取是至关重要的，根据查阅大量资料及结合水泥厂的实际情况。我们在预分解窑仿真系统中，将9个检测参数作为神经网络训练样本的输入数据、5个控制参数作为神经网络训练样本的输出数据。所作的参数选择具有普遍适用性和代表性，足以满足其仿真及生产控制的要求。在此基础上，采集了来自于生产现场的数据，针对样本数据在单位和数量级上的悬殊差别，我们运用极差标准化方法对样本数据进行了归一化预处理，并对神经网络进行了训练与研究。仿真实验结果表明以上方法取得了较好的训练结果，满足了仿真系统建模的要求。
此外，本文还对VB调用Matlab5.1中神经网络工具箱的方法，应用ADO控件实现样本数据库、专家规则库的管理，以及根据水泥预分解窑系统的特点实现的故障诊断功能进行了介绍。
第三，对系统的整体仿真框架进行了分析与研究。提出了系统的功能划分方法，以力求系统的普遍实用性和代表性。对系统基于Web和三层C／S结构的仿真框架进行了阐述，提出系统软件的开发模式。
第四，介绍了运用3D Studio MAX和VRML建立的仿真系统三维模型和动画，对系统模型进行了展示，并阐述了利用3D Studio MAX建立VRML模型的方法。
最后，作者对全文进行了总结，并对今后的研究方式与方法作出了展望。
In recent years, many cement producing lines with new technology and advanced automatic level have been built up in our country. This implies that a great deal of workers and engineers with high abilities should been trained continuously.
Cement Producing Process Computer Simulation System (CPPCSS) is a highlight of 211 Engineering in Wuhan University of Technology. The Rotary Kiln System Computer Simulation described in this paper is one of the subsystems of it. The author discusses the following contents based on this system.
Firstly, according to referring to a lot of references and visiting on the spot, the author study and analyze the cement producing process and its simulation technology, which includes the rotary kiln system producing process.
Secondly, considering the complexity of cement rotary kiln system, we build up the models of this system with the Integration techniques of artificial neural network and expert system. The improved EBP neural network algorithm is used in this system.
In the paper, author tries to integrate artificial neural network with expert system, and achieves the strong complementarity: solving the difficulty in knowledge-acquisition of ES ("bottle nee') by ANN' s excellent capacity of self-organizing, self-learning and self-adaptation, and the weakness of ANN on knowledge-representation ("black-box") by the valuable explain-function of ES.
In the modeling process, the selection of the input and output variables is very important, according to referring to a lot of references and visiting on the spot and considering the applicability and representative of the simulation system. We carefully select nine inspection parameter as input variables of the artificial neural network and five controls parameter as output variables. The author consider that the selection method is enough to meet the requests of the simulation and producing control with variant of the pre-decomposing rotary kiln system. According to the difference of the stylebook data, we deal with the stylebook data by pretreatment. Based on this method, and according to the data gathering from the producing spot, the author trains and researches the neural network. The results indicate that the method was efficient.
Further, the paper introduce the transfer of MatlabS.l by VB and the management of the expert rules database and the Fault Diagnosis of the pre-decomposing rotary kiln system.
Thirdly, the author analyzes and researches this system in detail. And the whole simulation frame of it is introduced. The simulation frame based on Web and C/M/S structure and the method of the system software development are discussed in this paper.
Fourthly, the three-dimensional models of the simulation system based on 3D Studio MAX and VRML are introduced and shown. And then, the transform method between 3D Studio MAX models and VRML ones is applied in this paper.
At last, the author summarizes the whole paper and points out the future direction of the development of the system.
The research in this paper not only has important theoretical and practical values in cement and other industries, but also makes solid basis to the post-development of CPPCSS.
本文所研究的系统为武汉理工大学“211工程”重点建设项目——水泥生产过程计算机仿真系统(CPPCSS)的子课题之一：水泥窑系统仿真系统。围绕此课题展开，本文主要包括以下内容：
第一，作者通过查阅大量国内外文献资料和实地参观考察等方式对水泥生产工艺技术(主要为窑系统生产过程)及其计算机仿真技术的发展现状进行了研究分析。
第二，根据水泥回转窑系统具有多变量、多回路、非线性、分布复杂和经典数学模型难以建立的特点，运用神经网络与专家系统集成技术对预分解窑系统进行了建模与研究。提出了对传统EBP神经网络算法的改进方法。
本文采用ES(专家系统)与ANN(人工神经网络)集成技术，利用ANN优良的自组织、自学习和自适应能力来解决ES知识获取的“瓶颈问题”，同时运用ES良好的解释机能来弥补ANN中知识表达的“黑箱结构”。将存于ANN内部的系统知识(数据化的结点和权值)运用统计的方法进行规则抽取，使其转换成显示的规则知识并进行解释。
建模过程中，输入输出参数的选取是至关重要的，根据查阅大量资料及结合水泥厂的实际情况。我们在预分解窑仿真系统中，将9个检测参数作为神经网络训练样本的输入数据、5个控制参数作为神经网络训练样本的输出数据。所作的参数选择具有普遍适用性和代表性，足以满足其仿真及生产控制的要求。在此基础上，采集了来自于生产现场的数据，针对样本数据在单位和数量级上的悬殊差别，我们运用极差标准化方法对样本数据进行了归一化预处理，并对神经网络进行了训练与研究。仿真实验结果表明以上方法取得了较好的训练结果，满足了仿真系统建模的要求。
此外，本文还对VB调用Matlab5.1中神经网络工具箱的方法，应用ADO控件实现样本数据库、专家规则库的管理，以及根据水泥预分解窑系统的特点实现的故障诊断功能进行了介绍。
第三，对系统的整体仿真框架进行了分析与研究。提出了系统的功能划分方法，以力求系统的普遍实用性和代表性。对系统基于Web和三层C／S结构的仿真框架进行了阐述，提出系统软件的开发模式。
第四，介绍了运用3D Studio MAX和VRML建立的仿真系统三维模型和动画，对系统模型进行了展示，并阐述了利用3D Studio MAX建立VRML模型的方法。
最后，作者对全文进行了总结，并对今后的研究方式与方法作出了展望。
In recent years, many cement producing lines with new technology and advanced automatic level have been built up in our country. This implies that a great deal of workers and engineers with high abilities should been trained continuously.
Cement Producing Process Computer Simulation System (CPPCSS) is a highlight of 211 Engineering in Wuhan University of Technology. The Rotary Kiln System Computer Simulation described in this paper is one of the subsystems of it. The author discusses the following contents based on this system.
Firstly, according to referring to a lot of references and visiting on the spot, the author study and analyze the cement producing process and its simulation technology, which includes the rotary kiln system producing process.
Secondly, considering the complexity of cement rotary kiln system, we build up the models of this system with the Integration techniques of artificial neural network and expert system. The improved EBP neural network algorithm is used in this system.
In the paper, author tries to integrate artificial neural network with expert system, and achieves the strong complementarity: solving the difficulty in knowledge-acquisition of ES ("bottle nee') by ANN' s excellent capacity of self-organizing, self-learning and self-adaptation, and the weakness of ANN on knowledge-representation ("black-box") by the valuable explain-function of ES.
In the modeling process, the selection of the input and output variables is very important, according to referring to a lot of references and visiting on the spot and considering the applicability and representative of the simulation system. We carefully select nine inspection parameter as input variables of the artificial neural network and five controls parameter as output variables. The author consider that the selection method is enough to meet the requests of the simulation and producing control with variant of the pre-decomposing rotary kiln system. According to the difference of the stylebook data, we deal with the stylebook data by pretreatment. Based on this method, and according to the data gathering from the producing spot, the author trains and researches the neural network. The results indicate that the method was efficient.
Further, the paper introduce the transfer of MatlabS.l by VB and the management of the expert rules database and the Fault Diagnosis of the pre-decomposing rotary kiln system.
Thirdly, the author analyzes and researches this system in detail. And the whole simulation frame of it is introduced. The simulation frame based on Web and C/M/S structure and the method of the system software development are discussed in this paper.
Fourthly, the three-dimensional models of the simulation system based on 3D Studio MAX and VRML are introduced and shown. And then, the transform method between 3D Studio MAX models and VRML ones is applied in this paper.
At last, the author summarizes the whole paper and points out the future direction of the development of the system.
The research in this paper not only has important theoretical and practical values in cement and other industries, but also makes solid basis to the post-development of CPPCSS.
引文
[ 1 ] Kuo-Chi Lin, Shih-Wei Tan, Kejun Chen: Study on the dynamical behavior of a real-time distributed simulation system, Simulation, vol. 74(1) ,2000
[2] Alan D. George etc.: An integrated simulation environment for parallel and distributed system prototyping, Simulation, vol. 72(5) . 1999
[3] Christian Schmid: A remote laboratory using virtual reality on the-web, Simulation, vol. 73(3) , 1999
[4] Jari Porras. Jouni Ikonen, Jarmo Harju: Analysing the effect of distribution to the simulation on a network of workstations, Simulation, vol. 72(5) , 1999
[5] M. C. Court: Commercial aircraft-cabin egress: the current state of simulation model development and the need for future research. Simulation, vol. 73(4) , 1999
[6] Wayne J. Davis etc.: Implementing on-line simulation with the world wide web. Simulation, vol. 73(1) . 1999
[7] Changpeng Fan: A simulation study of soft real-time scheduling and handling schemes for multimedia support, Simulation Practice and Theory 7,1999
[8] Judith S. Dahmann etc.: Standards for simulation: as simple as possible but not simpler the high level architecture for simulation. Simulation, vol. 71(6) . 1998
[9] Richard M. Satava, Associate Member. IEEE, and Shaun B. Jones: Current and future application of virtual reality for medicine, Proceedings of the IEEE, vol. 86(3) , 1998
[10] Michael R. Nevins etc: Simulation, animation and visualization of seaport operations. Simulation, vol. 71 (2) , 1998
[11] Chit-Sang Tsang, Carlos R.Villamar: Performance of a backpropagation neural network in diagnostic rhyme test word recognition, Simulation, vol. 70(3) , 1998
[12] Chell A. Roberts, Yasser M. Dessouky: An overview of object-oriented simulation, Simulation, vol.. 70(6) ,1998
[13] Marjan Rihar. The software simulator as an effective tool for testing control algorithms, Simulation, vol. 63(1) , 1994
[14] A.S.Morris: Use of colour to describe system states in a process simulation language, Simulation, vol. 62(1) . 1994
[15] Marisol Delgado, Cesar Pichardo: Use of Matlab and 20-sim to simulate a flash separator. Simulation Practice and Theory. 1999,7
[16] Levent Mollamustafaoglu: Object-oriented design of output analysis tools for simulation languages. Simulation, vol. 60(1) , 1993
[17] Lu Feng, Liu Xiang. Li Feng: Research on Network Database in Cement Producing Process Computer Simulation System. Advances In Systems Science and Applications. 2000． 12
[18] Lu Feng, Liu Xiang, Li Feng: Research on the Application of Web and C/M/S Structure in the Cement Producing Process Computer Simulation System.Proceedings of 7th Joint International Computer Conference.2000,10
[19] Lu Feng, Li Feng, Liu Xiang: Building and Search of the Web Graphics-base of the Cement Producing Process Computer Simulation System Based on Browser/Server Mode.Advances In Systems Science and Applications.2000,12
[20] 陈德全等．水泥预分解技术与热工系统工程．北京：中国建材工业出版社，1998．
[21] 黄稳山、何坚强．水泥厂中央控制室微机及仪表．武汉：武汉工业大学出版社，1995．
[22] 彭礼孝等．3D Studio MAX R3建模基础教程．北京：北京希望电子出版社，2000．
[23] 黄心渊等．3D Studio MAX R2.5学习捷径：(动画、材质和粒子系统篇)．北京：清华大学出版社，1999．
[24] [美]Jon A.Bell等著，潇湘工作室译．3D Studio MAX R2.5大全．北京：机械工业出版社，1999．
[25] 王永骥等．神经元网络控制．北京：机械工业出版社，1998．
[26] 李士勇．模糊控制·神经控制和智能控制论．哈尔滨：哈尔滨工业大学出版社，1998．
[27] 康博创作室．Photoshop 5.X实用教程．北京：人民邮电出版社，1998．
[28] 王宝圣等．Photoshop 5.0制作实例．北京：人民邮电出版社，1998．
[29] [美]Andrea L.Ames David等著，宗志方等译．VRML资源手册．北京：电子工业出版社，1998．
[30] [美]Michael J.Young著，邱仲潘等译．Visual C++6从入门到精通．北京：电子工业出版社，1999．
[31] 官章全等．Visual C++6.0类库大全．北京：电子工业出版社，1999．
[32] 李于剑．Visual C++实践与提高——图形图像编程篇．北京：中国铁道出版社，2001．
[33] 前沿电脑工作室．巧学巧用Dreamweaver、Fireworks、Flash制作网页．北京：人民邮电出版社，1999．
[34] 康博创作室．动态Web应用高级开发指南——ASP、ADO和DHTML编程．北京，人民邮电出版社，1999．
[35] 赵黎．虚拟现实技术是21世纪的关键技术．舰船指挥控制系统，1996(5)：31～37
[36] 冯惠军，冯允成．一个面向对象的仿真建模框架．系统工程理论与实践，1999，(5)：61～66
[37] 陈宗海，许志诠．过程系统训练仿真器的面向对象分析与设计．系统仿真学报，1998．10(3)：50～53
[38] 杨明，夏浩，马萍，王子才．面向对象建模技术开发大型培训仿真系统的一种方法．系统仿真学报，1997，9(4)：24～26
[39] 陈新安．面向对象分析(OOA)方法在数据库系统设计中的应用．计算机工程与应用，2000，5：124～144
[40] 李页清，周晓辉．华新厂5000t/d新型双系列ILC预分解窑的技术分析．水泥工程，2000，3：46～50
[41] 赵黎明，沈艳丰，官晓玲．建筑市场需求预测中的神经网络方法．系统工程理论与实践，2000，10：121～125
[42] 孙晨，周志华，陈兆乾．神经网络规则抽取研究．计算机应用研究，2000，2：34～37
[43] 蔡清方．Web页面两种模型对照研究．计算机应用研究，2001，3：134～136
[44] 李刚，王西．基于Web访问数据库的实现方案．计算机工程与应用，2000，2：112～115
[45] 武苍林，罗晃，夏威，朱建民．基于Intranet的Web数据库发布平台的研究与实现．计算机工程，2000．26(2)：97～99
[46] 陈豪，孙正义，张德富．三层客户／服务器体系结构的一个应用实例．计算机工程与应用，2000，3：173～176
[47] 吕锋，刘传波．水泥悬浮预热器的GM(1.4)动态模型及应用．武汉工业大学学报，1999，12
[48] 吕锋，阮卫东．测试模型参数估计的分块算法及其应用．武汉工业大学学报，1999，12
[49] 吕锋，李道远，闻震．基于Winsock控件实现水泥仿真系统的通讯．中国建材，2000，5
[50] 吕锋，崔晓辉．VRML在水泥生产过程层次仿真模型库研究中的应用．武汉理工大学学报，2001，5
[51] 吕锋，崔晓辉．VRML在水泥计算机仿真中的应用．武汉理工大学学报，2001，23(4)
[52] 吕锋，李锋，刘翔．基于Web的CPPCSS的动画与可视化研究．武汉理工大学学报，2001，23(3)
[2] Alan D. George etc.: An integrated simulation environment for parallel and distributed system prototyping, Simulation, vol. 72(5) . 1999
[3] Christian Schmid: A remote laboratory using virtual reality on the-web, Simulation, vol. 73(3) , 1999
[4] Jari Porras. Jouni Ikonen, Jarmo Harju: Analysing the effect of distribution to the simulation on a network of workstations, Simulation, vol. 72(5) , 1999
[5] M. C. Court: Commercial aircraft-cabin egress: the current state of simulation model development and the need for future research. Simulation, vol. 73(4) , 1999
[6] Wayne J. Davis etc.: Implementing on-line simulation with the world wide web. Simulation, vol. 73(1) . 1999
[7] Changpeng Fan: A simulation study of soft real-time scheduling and handling schemes for multimedia support, Simulation Practice and Theory 7,1999
[8] Judith S. Dahmann etc.: Standards for simulation: as simple as possible but not simpler the high level architecture for simulation. Simulation, vol. 71(6) . 1998
[9] Richard M. Satava, Associate Member. IEEE, and Shaun B. Jones: Current and future application of virtual reality for medicine, Proceedings of the IEEE, vol. 86(3) , 1998
[10] Michael R. Nevins etc: Simulation, animation and visualization of seaport operations. Simulation, vol. 71 (2) , 1998
[11] Chit-Sang Tsang, Carlos R.Villamar: Performance of a backpropagation neural network in diagnostic rhyme test word recognition, Simulation, vol. 70(3) , 1998
[12] Chell A. Roberts, Yasser M. Dessouky: An overview of object-oriented simulation, Simulation, vol.. 70(6) ,1998
[13] Marjan Rihar. The software simulator as an effective tool for testing control algorithms, Simulation, vol. 63(1) , 1994
[14] A.S.Morris: Use of colour to describe system states in a process simulation language, Simulation, vol. 62(1) . 1994
[15] Marisol Delgado, Cesar Pichardo: Use of Matlab and 20-sim to simulate a flash separator. Simulation Practice and Theory. 1999,7
[16] Levent Mollamustafaoglu: Object-oriented design of output analysis tools for simulation languages. Simulation, vol. 60(1) , 1993
[17] Lu Feng, Liu Xiang. Li Feng: Research on Network Database in Cement Producing Process Computer Simulation System. Advances In Systems Science and Applications. 2000． 12
[18] Lu Feng, Liu Xiang, Li Feng: Research on the Application of Web and C/M/S Structure in the Cement Producing Process Computer Simulation System.Proceedings of 7th Joint International Computer Conference.2000,10
[19] Lu Feng, Li Feng, Liu Xiang: Building and Search of the Web Graphics-base of the Cement Producing Process Computer Simulation System Based on Browser/Server Mode.Advances In Systems Science and Applications.2000,12
[20] 陈德全等．水泥预分解技术与热工系统工程．北京：中国建材工业出版社，1998．
[21] 黄稳山、何坚强．水泥厂中央控制室微机及仪表．武汉：武汉工业大学出版社，1995．
[22] 彭礼孝等．3D Studio MAX R3建模基础教程．北京：北京希望电子出版社，2000．
[23] 黄心渊等．3D Studio MAX R2.5学习捷径：(动画、材质和粒子系统篇)．北京：清华大学出版社，1999．
[24] [美]Jon A.Bell等著，潇湘工作室译．3D Studio MAX R2.5大全．北京：机械工业出版社，1999．
[25] 王永骥等．神经元网络控制．北京：机械工业出版社，1998．
[26] 李士勇．模糊控制·神经控制和智能控制论．哈尔滨：哈尔滨工业大学出版社，1998．
[27] 康博创作室．Photoshop 5.X实用教程．北京：人民邮电出版社，1998．
[28] 王宝圣等．Photoshop 5.0制作实例．北京：人民邮电出版社，1998．
[29] [美]Andrea L.Ames David等著，宗志方等译．VRML资源手册．北京：电子工业出版社，1998．
[30] [美]Michael J.Young著，邱仲潘等译．Visual C++6从入门到精通．北京：电子工业出版社，1999．
[31] 官章全等．Visual C++6.0类库大全．北京：电子工业出版社，1999．
[32] 李于剑．Visual C++实践与提高——图形图像编程篇．北京：中国铁道出版社，2001．
[33] 前沿电脑工作室．巧学巧用Dreamweaver、Fireworks、Flash制作网页．北京：人民邮电出版社，1999．
[34] 康博创作室．动态Web应用高级开发指南——ASP、ADO和DHTML编程．北京，人民邮电出版社，1999．
[35] 赵黎．虚拟现实技术是21世纪的关键技术．舰船指挥控制系统，1996(5)：31～37
[36] 冯惠军，冯允成．一个面向对象的仿真建模框架．系统工程理论与实践，1999，(5)：61～66
[37] 陈宗海，许志诠．过程系统训练仿真器的面向对象分析与设计．系统仿真学报，1998．10(3)：50～53
[38] 杨明，夏浩，马萍，王子才．面向对象建模技术开发大型培训仿真系统的一种方法．系统仿真学报，1997，9(4)：24～26
[39] 陈新安．面向对象分析(OOA)方法在数据库系统设计中的应用．计算机工程与应用，2000，5：124～144
[40] 李页清，周晓辉．华新厂5000t/d新型双系列ILC预分解窑的技术分析．水泥工程，2000，3：46～50
[41] 赵黎明，沈艳丰，官晓玲．建筑市场需求预测中的神经网络方法．系统工程理论与实践，2000，10：121～125
[42] 孙晨，周志华，陈兆乾．神经网络规则抽取研究．计算机应用研究，2000，2：34～37
[43] 蔡清方．Web页面两种模型对照研究．计算机应用研究，2001，3：134～136
[44] 李刚，王西．基于Web访问数据库的实现方案．计算机工程与应用，2000，2：112～115
[45] 武苍林，罗晃，夏威，朱建民．基于Intranet的Web数据库发布平台的研究与实现．计算机工程，2000．26(2)：97～99
[46] 陈豪，孙正义，张德富．三层客户／服务器体系结构的一个应用实例．计算机工程与应用，2000，3：173～176
[47] 吕锋，刘传波．水泥悬浮预热器的GM(1.4)动态模型及应用．武汉工业大学学报，1999，12
[48] 吕锋，阮卫东．测试模型参数估计的分块算法及其应用．武汉工业大学学报，1999，12
[49] 吕锋，李道远，闻震．基于Winsock控件实现水泥仿真系统的通讯．中国建材，2000，5
[50] 吕锋，崔晓辉．VRML在水泥生产过程层次仿真模型库研究中的应用．武汉理工大学学报，2001，5
[51] 吕锋，崔晓辉．VRML在水泥计算机仿真中的应用．武汉理工大学学报，2001，23(4)
[52] 吕锋，李锋，刘翔．基于Web的CPPCSS的动画与可视化研究．武汉理工大学学报，2001，23(3)