摘要
质量交换网络(MEN)综合发展至今,已经成为了过程工业中实现污染预防的有效工具。当板式塔作为质量交换器来综合MEN时,得到最优网络结构后,塔板数通常需要向上圆整到最近的整数,这将增大MEN的设备投资费用。在质量交换网络设计中,当设备投资费用增大时,以质量分离剂费用为主的操作费用应减小。然而在传统的优化模型中,尽管在塔板数向上取整后增大了设备投资费用,操作费用依然被认为是定值,因此得到的最低年度总费用是偏大的。为了解决这一问题,本工作在目标函数中加入了取整函数。将取整函数应用在基于MEN多级超结构的非线性(NLP)模型中,并将修改后的新模型应用于焦炉气的处理的案例中,得到了比其他文献费用更低的质量交换网络。结果表明:取整函数在综合MEN时,能够解决由于对塔板数取整而造成的最优MEN结构的最小年度总费用偏大的问题,并且能够获得更加准确,费用更低的质量交换网络。
Mass exchange network(MEN)has become an effective tool for pollution prevention in the process industry.When the staged column was employed as mass exchanger to synthesize a MEN,the number of trays must be rounded up to the next largest integer number after obtaining an optimal design,which would increase the capital cost.In designing a staged column,when the capital cost increased,the operating cost should decrease.However,the operating cost was regarded as a fixed value in the traditional optimization model even if the capital cost increased after rounding up to the number of trays,so the total annual cost(TAC)was higher than the real optimal objective.To solve this problem,an integral function was added to the mathematical model in this research.This function was applied in the nonlinear programming(NLP)based on the stage-wise superstructure.The modified model was applied in the case of coke-oven gas treatment,leading to a MEN with lower cost than data from other literatures.This paper showed that the integer function can solve the inaccurate minimum TAC result of the optimal MEN due to the rounding of the number of trays.A more accurate MEN structure with lower TAC could obtained than those previously reported in similar studies.
作者
侯创
罗明生
徐文星
HOU Chuang;LUO Mingsheng;XU Wenxing(Beijing Institute of Petrochemical Technology,Beijing 102617,China;Beijing University of Chemical Technology,Beijing 100029,China)
出处
《化学反应工程与工艺》
CAS
北大核心
2020年第2期108-116,共9页
Chemical Reaction Engineering and Technology
关键词
质量交换网络
多级超结构
非线性模型
工艺合成
工艺优化模拟
mass exchange network
multilevel superstructure
nonlinear programming
process synthesis
process optimization and modeling