基于改进GNG的电解铝生产杂质预警与控制  

Early Warning and Control of Impurities in Electrolytic Aluminum Production Based on Improved GNG

作　　者：刘天松 吴永明[1,2,3] 盛晓静 陈琳升 LIU Tian-song;WU Yong-ming;SHENG Xiao-jing;CHEN Lin-sheng(Key Laboratory of Advanced Manufacturing Technology,Ministry of Education,Guizhou University,Guiyang 550025,China;State Key Laboratory of Public Big Data,Guizhou University,Guiyang 550025,China;School of Mechanical Engineering,Guizhou University,Guiyang 550025,China)

机构地区：[1]贵州大学现代制造技术教育部重点实验室,贵阳550025 [2]贵州大学公共大数据国家重点实验室,贵阳550025 [3]贵州大学机械工程学院,贵阳550025

出　　处：《组合机床与自动化加工技术》2021年第6期72-75,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金资助项目(51505094,61962009);贵州省科学技术基金计划项目[(2016)1037];贵州省科技支撑计划项目[(2017)2029]。

摘　　要：铝电解工业中,铝液中杂质元素含量过高会严重影响铝液质量,Fe和Si含量的预警和控制对保证铝液质量有着重大意义。针对铝液中的杂质含量过高,缺乏用于实时监督的准确预警模型的问题,提出了一种基于加权欧式距离的GNG增量学习模型,构建了阈值机制实现对工业数据流实时监督和预警。通过K-means算法、经典GNG算法与改进GNG算法进行网络拓扑图比较分析。最后,结合铝液中Fe和Si含量的时间序列数据进行动态特征分析,实验结果表明,改进GNG算法对Fe和Si含量数据的实时监督和预警具有可靠性与准确性。In aluminum electrolysis industry,the high content of impurity elements in the liquid aluminum will seriously affect the quality of the liquid aluminum.The early warning and control of Fe and Si content is great significance to ensure the quality of liquid aluminum.In view of the high impurity content in aluminum liquid and the lack of a accuracy warning model for real-time supervision,this paper proposes an improved Growing Neural Gas(GNG)incremental learning model based on weighted European distance,and constructs a threshold mechanism to realize real-time supervision and early warning of industrial data flow.K-means algorithm,classical GNG algorithm and improved GNG algorithm were compared and analyzed.Finally,the dynamic characteristics of Fe and Si content in aluminum solution were analyzed in combination with the time series data.The experimental results showed that the improved GNG algorithm had reliability and accuracy in real-time monitoring and early warning of Fe and Si content data.

关 键 词：预警 电解铝 加权欧式距离 GNG 控制 

分 类 号：TH16[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]