计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (26): 245-248.DOI: 10.3778/j.issn.1002-8331.2010.26.074

• 工程与应用 • 上一篇    

基于神经网络的氟石粉液化模型分析

陈 刚1,杜树新1,江 丽2,李 晨2,蔡 婧2   

  1. 1.浙江大学 工业控制技术国家重点实验室,杭州 310027
    2.上海出入境检验检疫局 原材料与工业品检测中心,上海 200135
  • 收稿日期:2009-02-19 修回日期:2009-04-07 出版日期:2010-09-11 发布日期:2010-09-11
  • 通讯作者: 陈 刚

Liquefied model analysis of fluorspar powder based on neural network

CHEN Gang1,DU Shu-xin1,JIANG Li2,LI Chen2,CAI Jing2   

  1. 1.State Key Lab of Industrial Control Technology,Zhejiang University,Hangzhou 310027,China
    2.Technical Center of Raw Material Inspection of Shanghai Entry-Exit Inspection and Quarantine Bureau,Shanghai 200135,China
  • Received:2009-02-19 Revised:2009-04-07 Online:2010-09-11 Published:2010-09-11
  • Contact: CHEN Gang

摘要: 在对氟石粉进行实验研究的基础上,应用RBF神经网络,建立了预测氟石粉液化FMP的模型。在该模型中,输入为影响氟石粉液化FMP的各种因素,包括粒径、均匀度、化学性质、真密度及堆积密度等,输出为FMP,预测结果表明,所建RBF神经网络模型预测精度较高,可应用于对氟石粉液化FMP的预测。同时采用逐步反向删除的方法分析输入属性对输出结果的影响,找出了影响氟石粉液化FMP的主要影响因素。最后对每个主要属性进行了灵敏度分析。

关键词: 氟石粉, 液化, 径向基神经网络, 属性分析, 灵敏度

Abstract: A modeling method based on RBF neural network is presented for liquefied FMP of fluorspar powder,where the model input includes the particle size,uniformity of particle,chemical properties,density and deposited density,and the model output is Flow Moisture Point(FMP).The simulation shows that the model has a high prediction accuracy.Finally,the backward stepwise elimination method is used to delete the redundant factors,and the sensitivity of every factor is analyzed.

Key words: fluorspar powder, liquefy, Radial Basis Function(RBF) neural network, analysis of attributes, sensitivity of factors

中图分类号: