化工进展 ›› 2023, Vol. 42 ›› Issue (2): 684-691.DOI: 10.16085/j.issn.1000-6613.2022-1409

• 化工过程与装备 • 上一篇    下一篇

基于社区发现算法的复杂网络关键反应路线提取

陈土杰1(), 毕可鑫1, 邱彤2,3, 吉旭1(), 戴一阳1   

  1. 1.四川大学化学工程学院,四川 成都 610065
    2.清华大学化学工程系,北京 100084
    3.清华大学北京市工业大数据重点实验室,北京 100084
  • 收稿日期:2022-07-27 修回日期:2022-10-26 出版日期:2023-02-25 发布日期:2023-03-13
  • 通讯作者: 吉旭
  • 作者简介:陈土杰(1996—),男,硕士研究生,研究方向为反应网络。E-mail:scuctj@163.com
  • 基金资助:
    国家自然科学基金(U1462206);国家重点研发计划(2021YFB4000500)

Extraction of important reaction pathways for complex reaction network based on community detection algorithm

CHEN Tujie1(), BI Kexin1, QIU Tong2,3, JI Xu1(), DAI Yiyang1   

  1. 1.School of Chemical Engineering, Sichuan University, Chengdu 610065, Sichuan, China
    2.Department of Chemical Engineering, Tsinghua University, Beijing 100084, China
    3.Beijing Key Laboratory of Industrial Big Data System and Application, Tsinghua University, Beijing 100084, China
  • Received:2022-07-27 Revised:2022-10-26 Online:2023-02-25 Published:2023-03-13
  • Contact: JI Xu

摘要:

在分子自由基尺寸上对化工过程建立准确的机理过程模型,是目前世界“分子炼油”发展的重要方向。分子炼油过程体系的复杂性主要来源于化学反应网络的耦合性和多尺度性,这对深入了解化工生产过程提出了挑战。对复杂化学反应网络进行关键信息挖掘与表达,有助于工程师深入理解化学反应过程机理,实现机理透明化。由于炼油炼化过程的复杂反应网络存在以关键反应物质为中心的模块化、社区化特征,本文采用Leiden社区发现算法,从介尺度上对蒸汽热裂解制乙烯的反应网络进行反应社区划分,并基于分子自由基尺度从简化后的反应社区提取出对应的关键反应路线,为宏观反应网络到微观物质的相互作用提供一种可解释性的桥梁,助力揭示物质转化过程的知识传递机制。

关键词: 反应, 网络, 算法, 介尺度, 社区发现, 反应路线

Abstract:

The development of accurate mechanistic process models of processes at the molecular radical scale is an important direction in the development of "molecular refining" in the world. The complexity of molecular refining process systems is mainly due to the coupling and multiscale of chemical reaction networks, which poses a challenge to the in-depth understanding of chemical production processes. The key information mining and representation of the complex reaction network can help engineers to understand the process mechanism in depth and realize the transparency of the mechanism. Since the complex reaction networks of oil refining processes have modular and community-based characteristics centered on key reaction substances, this paper adopted the Leiden community detection algorithm to divide the reaction networks of steam cracking into reaction communities at mesoscale. The corresponding key reaction pathways were extracted from the reduced reaction communities at the molecular free radical scale. It provided an interpretable bridge from macroscopic reaction networks to microscopic substance interactions and helped to reveal the knowledge transfer mechanism of the substance transformation process.

Key words: reaction, network, algorithm, mesoscale, community detection, reaction pathway

中图分类号: 

京ICP备12046843号-2;京公网安备 11010102001994号
版权所有 © 《化工进展》编辑部
地址:北京市东城区青年湖南街13号 邮编:100011
电子信箱:hgjz@cip.com.cn
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持:support@magtech.com.cn