Generalized Predictive Temperature Control in Tubular Chemical Reactors by means of Proper Orthogonal Decomposition and Least Squares Support Vector Machine
作者单位：Deportment of Automation Harbin University of Science and Technology Deportment of Mathematics and Statistics Curtin University
会议名称：《第三十九届中国控制会议》
会议日期：2020年
学科分类：080706[工学-化工过程机械] 12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0807[工学-动力工程及工程热物理] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）]
关 键 词：Generalized predictive control Proper orthogonal decomposition Least Squares Support Vector Machine Line kernel function Tubular Chemical Reactors
摘 要：A novel generalized predictive control（GPC） scheme employing proper orthogonal decomposition（POD） and least squares support vector machine（LS-SVM） has been utilized to regulate the distributed temperature profile in tubular chemical reactors to optimize the productivity. The POD technique is used to reduce the high dimensionality of the discretized temperature field, and then the Low dimensional approximate autoregressive exogenous（ARX） prediction models are acquired by ***, A GPC scheme is investigated based on ARX prediction models. A numerical application of the proposed GPC scheme shows that it had satisfied temperature control performance.