节点文献

智能控制技术在水泥工业中的应用

Application of Artificial Intelligence Technology in Cement Industry

【作者】 李志明

【导师】 谭思云;

【作者基本信息】 武汉理工大学 , 控制理论与控制工程, 2003, 硕士

【摘要】 由于水泥回转窑的煅烧过程具有多变量、非线性、纯滞后和强耦合等特点,使得采用传统的控制理论对水泥窑的控制变的十分困难。本学位论文发挥智能控制理论的特点,将神经网络控制和模糊控制应用到水泥窑的优化控制中。本文在阅读了大量的书籍的基础上,将VC++软件和智能控制理论结合起来,用高级语言实现了BP算法、Levenberg-Marguardt算法和Quasi-Newton算法,设计了神经网络系统软件;并将MATLAB软件应用到水泥窑的仿真中,设计了模糊控制仿真软件。将神经网络软件和模糊控制仿真软件应用到水泥窑的优化控制中,结果表明,采用智能控制优化水泥窑控制可显著的降低能耗,提高经济效益,对水泥生产有重要的现实意义。 本文作者首先研究了神经网络的发展、结构和算法,在此基础上,利用VC++软件和动态链接库技术设计并编制了神经网络系统软件,该软件可以采用BP算法、LM算法或拟牛顿算法对一个新的神经网络进行训练,并可以对训练好的神经网络进行测试,还可以采用关系数据库技术对训练样本和测试样本进行管理。将神经网络系统软件应用到水泥窑的优化控制,结果表明,该软件具有一定的使用价值。 其次,作者还研究了模糊控制的基本理论和模糊控制器的设计方法,同时还了解了水泥生产工艺和操作过程,对水泥生产过程中的控制参数和被控参数作了进一步研究,提出了参数的隶属函数,结合中控室操作员的经验提出了水泥窑输入、输出参数之间的模糊控制规则表,然后利用MATLAB软件中的模糊逻辑工具箱设计了模糊控制仿真系统,并设计了GUI界面。将模糊控制仿真软件应用到水泥窑的仿真过程中,取得了比较好的控制效果,对水泥生产的研究和仿真具有一定的意义。 最后,作者对全文进行了总结,并对今后的研究方式和方法作出了展望。

【Abstract】 Because the calcining of cement kiln has characteristics of more variables, non-linearity, pure lag, strong coupling etc., controlling cement kiln with traditional control ways is very difficult. This thesis will bring into play the merits of aritificial intelligence, applying neural networks control and fuzzy control to optimize cement kiln. Based on reading a lot of books, this thesis fell VC++ and aritificial intelligence together, and realized BP algorithm, Levenberg-Marguardt algorithm and Quassi-Newton algorithm by VC++. This thesis also designed fuzzy control imitation software by MATLAB. The applications of neural networks software and fuzzy control imitation software show that aritificial intelligence could reduce energy consumption and increase economic efficience, which have important significance to cement industry.Firstly, the author researched the development, structures and algorithms of neural networks, then programmed the neural networks software by VC++ and DLL. The software could train a new neural network by BP , LM or Quasi-Newton algorithm, test the neural network and manage training swatches and testing swatches. The result of applying neural networks software to cement kiln control shows that it has definite practical worth.Based on researching fuzzy control theory and design way of fuzzy controller, and acquainting cement manufacturing process, the author researched the parameters of producing process in detail, brought forward subjection function and fuzzy control rules list of cement kiln input and output parameters and designed fuzzy control imitation software by fuzzy control toolbox of MATLAB. Applied the software to cement kiln imitating process, it has better control effect and definite significance.At last, the author summarized the whole thesis and made a prospect of research fashions and ways.

【关键词】 水泥回转窑神经网络模糊控制MATLAB
【Key words】 cement kilnneural networksfuzzy controlMATLAB
  • 【分类号】TP273.5
  • 【被引频次】10
  • 【下载频次】363
节点文献中: 

本文链接的文献网络图示:

本文的引文网络