某焦化厂供水网络多目标建模及优化.pdf

学兔兔 化 工 进 展 2016年第35卷第4期 CHEMICAL INDUSTRY AND ENGINEERING PROGRESS ·1033· 某焦化厂供水网络多目标建模及优化 韩庚，曹萃文，顾幸生 (华东理工大学化工过程先进控制与优化技术教育部重点实验室，上海 200237) 摘要：对某焦化厂的供水网络进行了详细的分析，在用水供需平衡满足生产要求的前提下，以新鲜水供给量最 小以及新鲜水供给成本与循环水及除盐水的浪费成本之和最小作为两个优化目标，建立了多目标混合整数非线 性规划(MINLP)的数学模型。利用文献提出的Pareto规则下的多目标改进文化差分进化算法对模型进行了求 解。所得的计算结果与现场实时操作数据相比，模型及算法实现的优化操作减少了新鲜水消耗量，降低了新鲜 水供给成本与循环水及除盐水的浪费成本，达到了很好的优化效果。 关键词：焦化厂；供水网络；混合整数非线性规划；多目标改进文化差分；多目标优化 中图分类号：TP29 文献标志码：A 文章编号：1000—6613(2016)04—1033—09 DoI：10．16085~．issn．1000．6613．2016．04．011 Multi·objective modeling and optimization for the water supply network of a coking plant HANGeng，CAO Cuiwen，GUXingsheng (Key Laboratory ofAdvanced Control and Optimization for Chemical Process，Ministry ofEducation，East China University of Science and Technology，Shanghai 200237，China) Abstract：In this paper，the water supply network system of a real—world cooking plant was deeply analyzed．Under the balance between water supply and demand，a bi—objective mixed-integer nonlinear programming(MINLP)model was built．Minimizing the fresh water supply and the total cost of flesh water，recycle water and desalting water were the two objectives．The Pareto based multi—objective cultural differential evolution(MO—CDE)algorithm of reference was selected to solve this mode1．The computing results showed that the bi—objective optimal results were better than current operational parameters in reducing the flesh water supply and the cost of flesh water，recycle water and desalting wate