东北大学学报（自然科学版）

东北大学学报(自然科学版) ›› 2010, Vol. 31 ›› Issue (12): 1673-1676.DOI: -

• 论著 •    下一篇

磨矿过程产品粒度软测量的比较研究

周平;柴天佑;   

  1. 东北大学流程工业综合自动化教育部重点实验室;
  • 收稿日期:2013-06-20 修回日期:2013-06-20 出版日期:2010-12-15 发布日期:2013-06-20
  • 通讯作者: -
  • 作者简介:-
  • 基金资助:
    中央高校基本科研业务费专项资金资助项目(N090608001);;

Comparative study on soft-sensor of grinding particle size

Zhou, Ping (1); Chai, Tian-You (1)   

  1. (1) Key Laboratory of Integrated Automation of Process Industry, Ministry of Education, Northeastern University, Shenyang 110004, China
  • Received:2013-06-20 Revised:2013-06-20 Online:2010-12-15 Published:2013-06-20
  • Contact: Zhou, P.
  • About author:-
  • Supported by:
    -

RichHTML

0

PDF (PC)

502

摘要： 磨矿粒度(GPS)难以进行直接在线检测且化验过程滞后,必须采用软测量技术对其进行在线估计以及在此基础上的闭环控制.针对一个实际的两段式磨矿回路流程,分别基于多元回归、案例推理(CBR)和神经网络(NN)技术建立了三种磨矿粒度软测量模型,并对其进行了基于工业试验的比较研究.结果表明,基于CBR的磨矿粒度软测量方法优于BP神经网络软测量(BP-NN)方法和多元回归模型估计方法.

关键词: 磨矿粒度(GPS), 软测量, CBR, BP-NN, 多元回归

Abstract: It is difficult to measure the grinding particle size (GPS) online directly, while the offline analysis with samples in lab will cause long time delay. The soft-sensor technique is therefore required to achieve the online estimation and closed-loop control of the GPS. Taking a practical two-stage grinding circuit as the objective to investigate, three GPS soft-sensor models were developed based on multiple regression, cased-based reason (CBR), and neural network (NN), respectively. These soft-sensor models were discussed comparatively via industrial tests. The results showed that the CBR soft-sensor model is superior to both the BP-NN soft-sensor model and the multiple regression soft-sensor model.

中图分类号: