NSGA-Ⅱ用于光气吸收塔的多目标优化  

Multiobjective optimization of phosgene absorber using NSGA-Ⅱ

作　　者：岳金彩[1] 曲波[2] 程华农[2] 韩方煜[2] 

机构地区：[1]华南理工大学化工与能源学院,广东广州510640 [2]青岛科技大学,山东青岛266042

出　　处：《计算机与应用化学》2008年第2期181-184,共4页Computers and Applied Chemistry

摘　　要：光气吸收塔是TDI生产中的一个重要设备,其操作要求塔顶气相出料中的光气含量、塔底液相出料中的HCl含量都尽可能低。选择合适的操作参数和设备参数,使两目标得以兼顾,而归结为一个多目标混合整数非线性规划(MOMINLP)的求解。本文在模块环境(Aspen Plus)下,建立了基于多目标遗传算法NSGA-Ⅱ求解MOMINLP的系统结构。编码为混合编码方式,整型变量采用二进制编码,实数变量采用实型编码。设计了遗传算法与Aspen Plus之间的接口程序,此接口不但传递数据,还具有二进制编码数据、模拟系统错误信息的处理功能。计算结果表明,遗传算法能够较好地给出Perato最优解集,本文提出的系统结构合理,可应用于一般化工流程的多目标优化综合。Phosgene absorber is an important unit in TDI process. There are two operation aims for the absorber, one is to minimize phosgene at top, the other is to minimize HCl at bottom. Optimization these two objectives at same time is a multiobjective mixed integer nonlinear programming （MOMINLP）. Mathematical program and multiobjective genetic algorithm（MOGA） are often used to solve this kind of problem. MOGA is regarded as a prospective method. In this paper we propose a framework based on NSGA- Ⅱ in modular environment （ Aspen Plus） to optimize the phosgene absorber. A interface is designed to not only transfer data between GA and Aspen Plus, but also handle binary code data from GA and error information from Aspen Plus. Results show the system can attain the Perato front of the problem and the framework can be easily used to optimize other processes.

关 键 词：NSGA-Ⅱ 模块环境 吸收塔 多目标混合整数非线性规划 

分 类 号：TQ028[化学工程]