摘要
针对动态多变量过程中难以提取明确的过程变量的动态关系问题,本文提出基于多数据集动态潜变量分析(MSDLV)的在线性能分级评估的方法.首先将性能相近的过程历史数据段划分为不同性能等级的集合,然后运用MSDLV方法提取性能级之间的公共基向量,保留训练数据中性能相关的过程变化,将性能相关的特有变化分解为动态部分与静态部分,提取动态自相关过程的动态因素.建立动态潜变量与性能等级之间的离线模型后,在线评估当前过程性能以及判断其所处状态.最后,将该方法运用于乙烯裂解炉反应过程,结果表明该方法具有良好的准确度.
In the dynamic multivariate process, the dynamic relations among process variables are implicit and difficult to interpret. An online performance grading assessment method based on multiset dynamic latent variables(MSDLV)is proposed in this paper to solve the problem. First, a similar historical dataset is divided into different sets according to performance grades. Then, variations related to performance are reserved in the training data due to common basis vector obtained through MSDLV algorithm and decomposed into dynamic part and static part. The dynamic factors in auto-correlated process are extracted, the offline model of latent variables and performance grades is established. Current performance can be assessed online, the steady-state performance grades and the transition performance grades are recognized and distinguished. Finally, the method is applied to the online performance assessment of ethylene cracking process,which illustrates the accuracy of proposed performance assessment method.
作者
曹晨鑫
王昕
王振雷
CAO Chen-xin;WANG Xin;WANG Zhen-lei(Key Laboratory of Advanced Control and Optimization for Chemical Processes,East China University of Science and Technology,Shanghai 200237,China;Electrical and Electronic Experimental Teaching Center,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《控制理论与应用》
EI
CAS
CSCD
北大核心
2020年第3期658-666,共9页
Control Theory & Applications
基金
国家自然科学基金项目(61673268)
国家自然科学基金重点项目(61533003)
国家自然科学基金重大项目(61590922)
中央高校基本科研业务费(222201814043)资助.
关键词
在线性能分级评估
多数据集动态潜变量
神经网络
动态自相关
乙烯裂解
online performance grading assessment
multiset dynamic latent variables
neural network
dynamic autocorrelation
ethylene cracking