融合过程控制的多目标实验设计  

Multi-objective experimental design of incorporating process control

作　　者：王阳阳 黄勋[1] 俞辉 赵志明[1] 李辉 刘钢[3] WANG Yang-yang;HUANG Xun;YU Hui;ZHAO Zhi-ming;LI Hui;LIU Gang(College of Mechanical and Electrical Engineering,Shaanxi University of Science and Technology,Xi′an 710021,Shaanxi Province,China;Fujian Institute of Research on the Structure of Matter,Chinese Academy of Sciences,Quanzhou 350000,Fujian Province,China;Xi′an Aerospace Composite Materials Institute,Xi′an 710025,Shaanxi Province,China)

机构地区：[1]陕西科技大学机电工程学院,陕西西安710021 [2]中科院福建物质结构研究所,福建福州350000 [3]西安航天复合材料研究所,陕西西安710025

出　　处：《化学工程》2020年第11期73-78,共6页Chemical Engineering(China)

基　　金：国家自然科学基金青年基金(61806186);“福建省智能物流产业技术研究院建设”项目(2018H2001);机器人与系统国家重点实验室(HIT)(SKLRS-2019-KF-15);泉州市科技计划项目(2019C112和2019STS08);陕西省重点研发项目(2020GT-174)。

摘　　要：为了提高模型参数估计的精度以及模型在特定生产条件下描述系统过程的能力,文中提出了一种结合过程控制与实验设计的方法,通过搭建集成的过程控制与最优实验设计框架,来平衡模型参数估计的精度以及模型预测实际生产过程的能力。以典型的酶促反应过程为例,构建过程控制与参数估计相结合的多目标实验设计优化问题,通过多目标优化算法和模糊优选法确定出最优输入底物浓度。结果表明:当输入底物浓度为0.856 mol/L时,参数估计精度和底物转化率同时达到最优。与单目标最优实验设计进行比较,文中所提方法能以微弱的参数估计精度下降为代价,显著地提高了底物转化率。In order to improve the accuracy of model parameter estimation and the ability to describe the system process under specific production conditions,a new design strategy combining the process control and experimental designwas proposed.By building an integrated process control and optimal experimental design framework,the precision of parameter estimation of the model and the ability of the model to predict the actual production process were balanced.Taking a typical enzyme reaction system model as an example,parameter estimation and process optimization were constructed.The optimal input substrate concentration could be obtained with the proposed design strategy and optimization methods.The results show that the parameter estimation accuracy and the substrate conversion rate reach the optimal simultaneouslywhen the input substrate concentration is 0.856 mol/L.Compared with the single optimal experimental design results,the method proposed in this paper can significantly improve the substrate conversion rate at the cost of weak parameter estimation accuracy.

关 键 词：最优实验设计 过程控制 参数估计 多目标优化 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]