摘要
铟金属在电解精炼过程中,随着电解液成分以及电解条件的变化容易发生阳极歧化反应、阴极析氢效应等异常情况,如果不及时处理,就会严重影响产品质量。以铟电解槽的pH值作为时间序列建立ARIMA模型,实现对铟电解槽的异常预测。首先对数据进行平稳性判断和白噪声检验,然后运用平稳数据绘制自相关图和偏自相关图,再利用赤池信息量准则和贝叶斯信息准则确定模型的最佳参数,最后确认残差序列为白噪声,证明了模型的有效性。实验结果表明:该模型能提早6 h进行预测,平均准确率达到93.28%;相对于传统的在线监测判断电解异常而言,该方法能提早预测电解液的pH值走势,从而及早作出改善措施,以减少在电解精炼过程中的损失。
During the electrorefining process of indium metal,with the change of electrolyte composition and electrolysis conditions,abnormal behaviors such as anode disproportionation reaction and cathode hydrogen evolution effect are prone to occur.If not processed in time,product quality will be seriously affected.In response to this problem,this paper uses the pH data of the indium electrolytic cell as a time series to establish an ARIMA model to predict the abnormal electrolysis process.First,perform the stationarity judgment and white noise test on the data.Second,draw the auto-correlation graph and partial auto-correlation graph on stationary data.Then,use Akaike information criterion and Bayes information criterion to determine the best parameters of the model.Finally,confirm the residual sequence is white noise.This proves effectiveness of the model.Experimental results show that the model can complete the prediction 6 hours earlier,with an average accuracy of 93.28%.Compared with traditional online monitoring to determine electrolysis abnormalities,this method can predict the pH trend of the electrolyte earlier,so that improvement measures can be taken early to reduce losses in the electrolytic refining process.
作者
杨聪
彭巨擘
伍美珍
张合生
YANG Cong;PENG Juqing;WU Meizhen;ZHANG Hesheng(School of Mechatronics Engineering&Automation,Shanghai University,Shanghai 200444,China;Yunnan Tin Group(Holding)Company Limited,Kunming 650000,China;Shanghai University(Zhejiang Jiaxing)Emerging Industries Institute,Jiaxing 314006,China)
出处
《仪表技术》
2022年第3期30-34,共5页
Instrumentation Technology
基金
云南省重大科技专项(202002AB080001-2)。
关键词
铟电解精炼
ARIMA模型
异常预测
indium electrolytic refining
ARIMA model
anomaly prediction
