Evaluation strategy and mass balance for making decision about the amount of aluminum fluoride addition based on superheat degree
Evaluation strategy and mass balance for making decision about the amount of aluminum fluoride addition based on superheat degree
复制标题
基于过热度决策氟化铝添加量的评价策略和质量平衡
DOI：
10.3934/jimo.2018169
复制
发表时间：
2020
影响因子：
1.3
通讯作者：
Yongfang Xie
中科院分区：
文献类型：
--
作者：
Weichao Yue;Weihua Gui;Xiaofang Chen;Zhaohui Zeng;Yongfang Xie
添加氟化铝（AlF3）的目的是调节铝还原过程中的结晶度（SD）。由于其固有的复杂性，确定添加的AlF3的适当量长期以来一直是一个具有挑战性的工业问题。由于经验丰富的技术人员数量减少，人工添加AlF3通常不准确，这容易导致电池状态不稳定。本文提出了一种基于SD的AlF3添加评价策略。设计了一种扩展的朴素贝叶斯分类器（ENBC）来估计分别代表当前和潜在电池状态的SD状态及其趋势，然后基于模糊理论对估计结果进行评价，从而对过程进行分级。根据评价等级将还原过程划分为几种情况，并引入物料平衡法确定每种情况下AlF3的加入量。实验结果表明，该方法是可行的，与现有方法相比，AlF3的添加效果得到了提高。此外，自动AlF3添加是有前途的基于所提出的策略。
The purpose of aluminum fluoride (AlF3) addition is to adjust the superheat degree (SD) in the aluminum reduction process. Determining the appropriate amount of AlF3 to add has long been a challenging industrial issue as a result of its inherent complexity. Because of the decreasing number of experienced technicians, the manual addition of AlF3 is usually inexact, which easily leads to an unstable cell condition. In this paper, an evaluation strategy based on the SD for AlF3 addition is proposed. An extended naive Bayesian classifier (ENBC) is designed to estimate the states of SD and its trends that represent the current and potential cell condition respectively, and then the process is graded by evaluating the estimated results based on fuzzy theory. The reduction process is divided into a few situations based on the evaluation grades, and mass balance is introduced to determine the amount of AlF3 addition in each situation. The results of experiments show that the proposed strategy is feasible, and the effectiveness of AlF3 addition is improved compared to the existing method. Moreover, automatic AlF3 addition is promising based on the proposed strategy.