摘要
针对一类具有多阶段特性的间歇过程操作曲线优化问题,提出了一种基于过程正常运行批次的数据驱动型操作曲线多阶段融合优化方法。首先通过时间片矩阵的划分以及主元分析,计算加权时间片载荷矩阵,并对其进行聚类分析得出过程的阶段;然后在每一阶段下利用时段变量与指标变量的相似度修正操作曲线;最后通过计算各阶段与指标变量的相关系数,获取阶段的权重从而实现操作曲线融合。将该方法应用到某化工产品的间歇结晶过程中,结果验证了所提方法的有效性。
A fusion optimization method based on daily operation batches was proposed to solve trajectory optimization problems of multi-stage batch processes. Time-slice matrices were divided and time-slice loading matrices were acquired by PCA（principal component analysis） algorithm. Clustering algorithm was applied to weighted time-slice loading matrices to obtain process stages. The operating trajectories were then corrected based on the similarity between time-segment variables and index variables at each stage. Weight coefficients were acquired by calculating correlation coefficients between each stage and index variables, and the stage trajectories were fused with the weight coefficient. This method was employed in a batch crystallization process and the results illustrate the effectiveness and benefits.
作者
仇力
栾小丽
刘飞
QIU Li;LUAN Xiao-li;LIU Fei(Key Laboratory of Advanced Process Control for Light Industry of the Ministry of Education,Jiangnan University,Wuxi 214122,China)
出处
《高校化学工程学报》
EI
CAS
CSCD
北大核心
2018年第3期628-635,共8页
Journal of Chemical Engineering of Chinese Universities
基金
国家自然科学基金(61473137
61722306)
关键词
间歇过程
多阶段
操作曲线优化
主元分析
batch processes
multi-stage
trajectory optimization
principal component analysis
