Computer Engineering and Applications ›› 2019, Vol. 55 ›› Issue (12): 250-258.DOI: 10.3778/j.issn.1002-8331.1712-0175

Previous Articles     Next Articles

Research and Application of Prediction Method for Key Indexes of Aluminum Electrolysis

CHEN Yong1,2,3, ZHOU Xiaofeng2,3, LI Shuai1,2,3   

  1. 1.University of Chinese Academy of Sciences, Beijing 100049, China
    2.Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, China
    3.Key Laboratory of Networked Control Systems, Chinese Academy of Sciences, Shenyang 110016, China
  • Online:2019-06-15 Published:2019-06-13


陈  勇1,2,3,周晓锋2,3,李  帅1,2,3   

  1. 1.中国科学院大学,北京 100049
    2.中国科学院 沈阳自动化研究所,沈阳 110016
    3.中国科学院 网络化控制系统重点实验室,沈阳 110016

Abstract: This paper presents a strategy to maintain the stability of aluminum reduction cells through the data mining and modeling of the aluminum factory in order to maintain the continuity of aluminum electrolytic production and ensure the stability of energy consumption and the stability of energy consumption. Firstly, de-noising for data. In view of the unknown data distribution characteristics of aluminum plants, a fuzzy clustering method without parameter adaptation is proposed, and the number and cluster centers of clusters can be obtained through iterative adaptation. According to the clustering results, the data of aluminum factory is labelled according to the actual meaning. A distance based continuous attribute naive Bias algorithm is proposed. The incremental idea of classifiers is used to improve the accuracy of algorithm classification. By using single slot test set data, every index level can be predicted by accumulating method, compared with the previous day, the change of variables are confirmed under each index, and then complete the prediction. It is found that the prediction model can predict the key indicators of aluminum electrolysis, and the proposed clustering and classification algorithms are good in UCI data and aluminum factory data.

Key words: aluminum production, fuzzy clustering, naive Bayes, cumulative method, incremental thinking

摘要: 为维持铝电解生产的持续性,保证电解槽的物耗稳定和能耗稳定,通过对铝厂数据的挖掘与建模,提出一整套维持电解槽稳定的策略方法,并用于指导实际生产。首先为数据去噪,针对铝厂数据分布特征未知的特点,提出一种无参自适应的模糊聚类方法,通过迭代自适应得到类簇个数和簇中心;根据聚类结果,将铝厂数据按实际意义标签化,提出一种基于距离的连续属性朴素贝叶斯算法,对分类器使用增量思想,使算法动态分类准确率得到提高;应用单槽测试集数据,通过累积法完成当天各指标等级趋势的预测,确定各指标下变量相对于前一天的变化量,完成预测。实验发现,预测模型可完成铝电解关键指标的预测;提出的聚类、分类算法在UCI数据及铝厂数据上表现良好。

关键词: 铝生产, 模糊聚类, 朴素贝叶斯, 累积法, 增量思想