摘要
针对大部分多阶段间歇过程,质量多数难以在线测量与控制问题,提出一种新颖的在线质量预测方法,基于子时段的MPLS质量预测方法.首先,采用聚类分析方法对间歇过程时间片矩阵PCA负载矩阵进行分类,依据过程变量相关性的变化,过程被分成几个操作阶段,然后,定出与质量变量最相关的阶段,在该阶段建立sub-MPLS在线质量预测模型,另外,还提出一种闭环质量控制方案.在一个典型的多阶段间歇过程-注塑过程的应用结果表明了所提出方法的有效性.
For most multistage batch processes,the final product quality are usually available at the end of the batch,which makes it difficult to realize on-line quality control.A novel method is thus developed to improve process analysis and on-line quality predictions for multistage batch processes,based on the stage-based subMPLS(multiway partial least square)quality prediction methods.Firstly,using clustering methods based on the stage-based sub-PCA(principal component analysis),the process is divided into severa...
出处
《东南大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2006年第S1期39-42,共4页
Journal of Southeast University:Natural Science Edition
基金
国家自然科学基金资助项目(60374003)
关键词
多阶段间歇过程
质量预测
主元分析
部分最小二乘
注塑过程
multi-stage batch process
quality prediction
principal component analysis(PCA)
partial least square(PLS)
injection molding process