高校化学工程学报2026,Vol.40Issue(5):920-929,10.DOI:10.3969/j.issn.1003-9015.2025.00.090
基于多模型协同的MTBE合成工艺优化
Research on optimization of MTBE synthesis process based on multi-model collaboration
摘要
Abstract
To address the challenges of process complexity and optimization difficulty in MTBE synthesis,this study proposes a process design and optimization method based on a machine learning surrogate model and a multi-objective evolutionary algorithm.Three typical process structures,namely the traditional reaction-distillation sequence(TR-TCD),the reactive distillation column(RD),and the reactive dividing-wall column(RDWC),were investigated.Surrogate models were established and integrated with NSGA-Ⅱ for multi-objective optimization.The optimization objectives included total annual cost(TAC),CO2emissions,and exergy efficiency(EE).The results indicate that the RDWC process exhibits the best overall performance among the three processes,with a TAC of 4 783.83 kUSD·a-1,CO2emissions of 38 370.65 kg·a-1,and an EE of 12.44%.In comparison,the TR-TCD and RD processes show certain disadvantages in either economic or energy efficiency aspects.The proposed design and optimization method not only enhances the economic and environmental performance of the process but also provides a new approach for energy-efficiency optimization and low-carbon design of complex reaction-separation coupled systems.关键词
甲基叔丁基醚/反应精馏/隔壁塔/代理模型/XGBoost/NSGA-Ⅱ/多目标优化Key words
methyl tert-butyl ether/reactive distillation/dividing-wall column/surrogate model/XGBoost/NSGA-Ⅱ/multi-objective optimization分类
化学化工引用本文复制引用
高明杰,李开宇,崔哲,田文德..基于多模型协同的MTBE合成工艺优化[J].高校化学工程学报,2026,40(5):920-929,10.基金项目
国家自然科学基金(22578239) (22578239)
山东省自然科学基金(ZR2022QB020,ZR2021MB113). (ZR2022QB020,ZR2021MB113)