精馏系统的多目标优化设计及分析  

Multi-objective optimization design and analysis of distillation system

作　　者：高伟 景旭亮 陈景 李鸿雄 杨东元 孙育滨 GAO Wei;JING Xuliang;CHEN Jing;LI Hongxiong;YANG Dongyuan;SUN Yubin(Shaanxi Yanchang Petroleum(Group)Co.,Ltd.,Xi′an 710000,Shaanxi Province,China;Shaanxi Yanchang Petroleum(Group)Co.,Ltd.,&Dalian Institute of Chemical Physics Xi′an Clean Energy(Chemical Industry)Research Institute,Xi′an 710000,Shaanxi Province,China)

机构地区：[1]陕西延长石油(集团)有限责任公司,陕西西安710000 [2]陕西延长石油(集团)有限责任公司大连化物所西安洁净能源(化工)研究院,陕西西安710000

出　　处：《化学工程》2024年第11期71-76,共6页Chemical Engineering(China)

基　　金：陕西省青年科技新星项目(2023KJXX-124);陕西省重点研发计划项目(2024GX-YBXM-508)。

摘　　要：相比传统精馏过程多以单塔作为优化对象,将整个精馏系统作为研究对象更有实际意义。将精馏系统总投资费用C_(T)和总热负荷Q作为优化目标,通过计算机COM接口连接Python和Aspen-Plus软件,调用NSGA-Ⅲ算法对精馏系统进行多目标优化。以戊烷分离流程为例,对4塔的理论板数、进料板数8个变量进行优化和分析。优化结果表明,相比较传统优化,多目标优化的表现更优,二者热负荷接近时总投资费用降低5.37%,总投资费用接近时,热负荷降低8.17%。同时分析多目标优化时侧重总投资费用C_(T)的解和热负荷Q的解的不同优化过程。进一步的单变量分析结果表明:由于精馏塔回流比参数的特性,目标函数中热负荷Q受到的影响更大,多目标优化的最优解在热负荷变化平稳后得出。Compared with the traditional distillation process which takes single column as the optimization object,it is more practical to take the whole distillation system as the research object.The total investment cost C_(T)and the total heat load Q of the distillation system were taken as the optimization objectives,Python and Aspen-Plus software were connected through COM interface of the computer,and NSGA-Ⅲalgorithm was invoked for multi-objective optimization of the distillation system.Taking pentane separation process as an example,8 variables of theoretical plate number and feed plate number of 4 towers were optimized and analyzed.The optimization results show that compared with the traditional optimization,multi-objective optimization has a better performance,the total investment cost reduces by 5.37%when the two heat loads are close,and the heat load reduces by 8.17%when the two total investment costs are close.At the same time,different optimization processes focusing on the solution of total investment cost C_(T)and the solution of heat load Q in multi-objective optimization were analyzed.Further univariate analysis results show that due to the characteristics of reflux ratio parameters of distillation column,heat load Q in the objective function is more affected,and the optimal solution of multi-objective optimization is obtained after the heat load changes smoothly.

关 键 词：精馏系统 多目标优化 NSGA-Ⅲ算法 单变量分析 

分 类 号：TQ015.9[化学工程]