板带在线实时质量预测数据预处理技术  

Data processing technology for on-line real-time quality performance prediction of strip

作　　者：刘博[1] 王学敏[1] LIU Bo;WANG Xue-min(Baoshan Iron&Steel Co.,Ltd.,Shanghai 201900,China)

机构地区：[1]宝山钢铁股份有限公司,上海201900

出　　处：《重型机械》2020年第6期8-12,共5页Heavy Machinery

基　　金：国家工信部智慧制造示范项目“钢铁冷轧数字化车间试点示范”。

摘　　要：对高速生产的板带质量进行实时预测(如板带机械性能),实际上是对板带各个位置的质量进行预测,由于各设备安装空间位置不同,在线实时获得的各工序数据并非对应板带同一位置,因此实时获得的数据并非同一组数据,无法直接使用。本文介绍了一种在线实时流式数据预处理方法,将时序数据集转换到空间位置数据集,实现对各位置的实时质量预测,实践应用表明该技术具有可实施性,可用来对板带质量进行在线实时预测。The real-time prediction of the strip quality(such as the mechanical properties of the strip)is actually to predict the quality in each position of the strip for high speed production.Due to the different installation space of the equipment,the data in on-line real-time obtained is not corresponding to the same position of the strip.Therefore,the data obtained are not the same group of data and cannot be used directly.This paper introduces an on-line real-time streaming data preprocessing method,which transforms the time series data set to the spatial location data set,and realizes the real-time quality prediction of the location.The application case shows that the technology in this paper is feasible,and it can be used in on-line real-time quality prediction of strip.

关 键 词：数字钢卷 数字工厂 智慧制造 大数据与人工智能 性能预测 

分 类 号：TG334.9[金属学及工艺—金属压力加工]