自然科学版 英文版
自然科学版 英文版
自然科学版 英文版

您目前所在的位置:首页 - 期刊简介 - 详细页面

中南大学学报(自然科学版)

Journal of Central South University

第43卷    第2期    总第210期    2012年2月

[PDF全文下载]    [Flash在线阅读]

    

文章编号:1672-7207(2012)02-0548-06
铝电解生产过程的多目标优化
郭俊,桂卫华,文新海

(中南大学 信息科学与工程学院,湖南 长沙,410083)

摘 要: 针对某铝厂铝电解生产过程中的摩尔比、电解温度和极距难以根据不同工况进行实时调整,无法达到期望综合生产目标的问题,提出以电流效率最高、槽电压最低为优化目标,以氧化铝浓度指标和生产工艺要求为约束条件的基本优化思想。首先采用多元线性回归与改进BP神经网络方法,建立铝电解生产过程非线性不等式约束的多目标优化模型;然后,采用比例加权系数法与广义简约梯度组合算法求解多目标优化问题,获得摩尔比、电解温度和极距的优化设定值。研究结果表明,在工况正常时,模型优化结果的相对误差在5%左右;工况异常时,模型优化结果的相对误差在10%以内;本文所用方法的优化精度较高,基于实际生产数据的仿真结果的有效率都在90%以上,在工况正常情况下可以达到97%,此结果能很好地满足企业的要求。

 

关键字: 铝电解;BP神经网络;多目标优化;比例加权系数法

Multi-objective optimization for aluminum electrolysis production process
GUO Jun, GUI Wei-hua, WEN Xin-hai

School of Information Science and Engineering, Central South University, Changsha 410083, China

Abstract:Based on the fact that in aluminum electrolysis production process, it’s difficult to adjust molecular ratio, electrolytic temperature and polar distance timely according to different conditions, the anticipant integrated production target can’t be reached, a basic optimization idea was proposed using the maximum of current efficiency and minimum of aluminum cell voltage as the optimization objective and using concentration of alumina and technological requirements as the constraint conditions. Firstly, the multiple linear regression and improved BP neural network methods were used to build a multi-objective optimization model with nonlinear inequality constraints for aluminum electrolysis production process.Then, the proportion weighted coefficient method and GRG (Generalized reduced gradient) combinatorial algorithm were applied to resolve the multi-objective optimization problem. Finally, the optimized target values of molecular ratio, electrolytic temperature and polar distance were obtained. The results show that the relative error of the proposed model is about 5% under normal producing condition and is less than 10% during abnormal production, which indicates the approach owns high precision. What’s more, the efficiency of the optimization results based the producing data is higher than 90%, and can reach 97% during normal production so the approach can meet the enterprise needs well.

 

Key words: aluminum electrolysis; BP neural networks; multi-objective optimization; proportion weighted coefficient method

中南大学学报(自然科学版)
  ISSN 1672-7207
CN 43-1426/N
ZDXZAC
中南大学学报(英文版)
  ISSN 2095-2899
CN 43-1516/TB
JCSTFT
版权所有:《中南大学学报(自然科学版、英文版)》编辑部
地 址:湖南省长沙市中南大学 邮编: 410083
电 话: 0731-88879765(中) 88836963(英) 传真: 0731-88877727
电子邮箱:zngdxb@csu.edu.cn 湘ICP备09001153号