您当前位置: 金沙js娱乐场官方网站  >  学术讲座

讲准字12号:姚远学术报告会

发布时间:2015-03-05|浏览次数:

题目:Applications of statistical methods to batch processes

主讲:姚远

时间:3月6日15:00

地点:电气学院415报告厅

主办:电气信息工程学院

 

主讲简介:姚远博士,2001毕业于浙江大学,2004年、2009年分别获浙江大学、香港理工大学硕士、博士学位,现为台湾清华大学化工系副教授。长期从事多变量统计过程控制、软测量、故障检测与诊断等研究工作。发表学术论文50余篇,参编专著1部。

主讲内容:For ensuring operation safety and consistent product quality, multivariate statistical methods, such as multiway PCA/PLS and phase-based PCA/PLS, have been widely adopted in chemical batch process monitoring and quality prediction. However, these methods still have certain limitations, which may not work well in many specific cases. In this presentation, the applications of statistical methods to batch processes will be discussed through three examples: 1. final product quality prediction, 2. batch-to-batch steady state identification and startup monitoring, and 3. online monitoring of resin transfer molding. In the first example, it will be shown that the conventional modeling methods may not fully utilize available process information, which affects both prediction accuracy and model interpretation. In the later two examples, process data do not obey the assumptions of the conventional statistical monitoring methods. This presentation will provide solutions to these problems and show how shallow process knowledge can help statistical modeling of batch processes.

 

欢迎广大师生参加!

XML 地图 | Sitemap 地图