Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (18): 245-248.DOI: 10.3778/j.issn.1002-8331.2009.18.074

• 工程与应用 • Previous Articles    

Research on modeling and controlling of cupola melting process based on neural network

LIU Su-qing1,HU Dong-gang1,SUN Zhi-yi1,LIANG Qiu-sheng2   

  1. 1.Taiyuan University of Science and Technology,Taiyuan 030024,China
    2.Foundry Workshop,Jingwei Textile Machinery Co. Ltd,Yuci,Shanxi 030601,China
  • Received:2009-04-10 Revised:2009-05-25 Online:2009-06-21 Published:2009-06-21
  • Contact: LIU Su-qing



  1. 1.太原科技大学,太原 030024
    2.经纬纺织机械股份有限公司 铸造厂,山西 榆次 030601
  • 通讯作者: 刘素清

Abstract: Cupola melting system is very difficult to establish effective quality control model using the conventional methods because this system is characterized as nonlinearity,completed structure and great fluctuation.In this paper,adaptive control approach of cupola based on neural network is brought forward and neural network as identification and controller is used to implement the self-adaption control of temperature of cupola molten iron.The network training and experimental results show that the effect of system control is good and the average error between predictive value and actual value is from -8℃ to 10℃.

Key words: cupola, neural network, intelligent control, modeling

摘要: 对于非线性、结构复杂、干扰波动大的冲天炉熔化系统,很难用传统的方法建立有效的质量控制模型。提出了一种基于神经网络的冲天炉自适应控制方法,利用神经网络作为辨识器和控制器,实现对冲天炉熔炼过程中铁液温度的自适应控制。网络训练与实验结果表明:实测指标与预测值误差在-8℃到10℃之间,系统控制效果良好。

关键词: 冲天炉, 神经网络, 智能控制, 建模