摘要
铝电解过程电流效率的影响因素多且相互耦合,为提高电解槽性能,确保电流高效率运行,需确定各因素之间的最佳匹配。针对上述问题,提出一种基于支持度权重分配数据融合改进算法的方法。首先,利用减聚类改进的模糊C均值算法进行聚类;然后,结合聚类结果,针对传统支持度权重分配数据融合算法存在的问题进行改进,得到待融合的各聚类中心的权重,进而得到铝电解过程电流效率影响因素的最佳匹配。仿真结果表明,与传统的指标最优法和均值法比较,利用数据融合得到的电流效率因素最佳匹配更实用合理,且能有效地指导电解槽操作,以提高电流效率。
As many factors affect the current efficiency of Aluminum electrolysis process and are mutually cou- pled, it needs to determine the best match between the factors to guide cell operation and ensure its high current efficiency. To solve the problem mentioned above, an improved support-degree based weight distribution data fusion method was proposed. First of all, the improved fuzzy C-means clustering algorithm with sub-clustering was used for clustering. Then, the weight of each factor of Clustering centers was determined according to the result of clustering, and the improved support-degree based weight distribution data fusion algorithm was applied to fuse the clustering centers, thereby the best match of influence factors was achieved. Finally, compared with traditional methods, this new method is more practical and reasonable, and can guide more effectively the operation of the cell to improve the current efficiency.
出处
《计算机仿真》
CSCD
北大核心
2015年第5期360-363,415,共5页
Computer Simulation
基金
国家863项目(2013AA040705)
中国科学院重点部署项目(KGZD-EW-302)
关键词
电流效率
最佳匹配
数据融合
模糊聚类
支持度权重分配
Current efficiency
Best match
Data fusion
Fuzzy clustering
Support-degree weight distribution
作者简介
孔磊（1990-），男（汉族），河南驻马店人，硕士生，主要从事铝电解能源管理方面的研究工作；
王卓（1976-），男（汉族），辽宁绥中人，博士，副研究员，主要研究方向为复杂工业过程的建模与控制。
