石油学报(石油加工) ›› 2021, Vol. 37 ›› Issue (6): 1235-1249.doi: 10.3969/j.issn.1001-8719.2021.06.004

• 炼油化工技术与模拟优化 • 上一篇    下一篇

基于反应热严格计算的加氢裂化反应器建模

李圣淋,李国庆,王艳,蔡楚轩   

  1. 华南理工大学 化学与化工学院,广东 广州 510640
  • 收稿日期:2020-10-23 修回日期:2021-05-22 出版日期:2021-11-25 发布日期:2021-11-02
  • 通讯作者: 李国庆,男,教授,博士,从事过程能量综合方面研究,E-mail:gqli1@scut.edu.cn E-mail:gqli1@scut.edu.cn
  • 作者简介:李圣淋,男,硕士研究生,从事过程能量综合方面研究,E-mail:824402140@qq.com

Modeling Hydrocracking Reactor Based on Rigorous Calculation of Reaction Heat

LI Shenglin, LI Guoqing, WANG Yan, CAI Chuxuan   

  1. School of Chemistry and Chemical Engineering, South China University of Technology, Guangzhou 510640, China
  • Received:2020-10-23 Revised:2021-05-22 Online:2021-11-25 Published:2021-11-02

摘要: Chevron模型是研究石油馏分窄集总加氢裂化反应动力学的基本模型,但它假设反应热恒定,即不管反应体系和反应条件如何,反应过程每消耗1kg新氢,都将释放21MJ反应热。无疑与实际不符。本文在反应过程氢、碳元素守恒的基础上,基于热力学状态函数法,提出了用反应物和反应产物集总标准燃烧热严格计算反应热的方法,取代恒反应热假设,改进了Chevron模型;同时基于反应温度误差和产品分布误差同时最小,用多目标遗传算法(NSGA-II)拟合模型参数,较之基于反应温度和产品分布总误差最小的传统GA拟合方法,大大提高了模型参数的精度。某200万吨/年蜡油加氢裂化装置反应器计算表明,新模型及NSGA-II参数拟合方法更能仿真实际操作,其产品分布预测精度比现有模型加GA拟合提高了14.0%。

关键词: 加氢裂化, 机理模型, 集总, 反应热, 遗传算法

Abstract: Chevron model has been a fundamental lump kinetic model of petroleum fraction hydrocracking. It assumes that constant quantity of 2.1×104 kJ heat is released when consuming 1 kg of fresh hydrogen during the reaction process for any reaction system and reaction conditions, which is not consistent with industrial practice. In this work, based on thermodynamic state function and hydrogen and carbon material balance, reaction heat was calculated rigorously through the standard combustion heat of lumped reactants and products. This significantly improved prediction accuracy compared with Chevron model. Furthermore, a correlation method aiming at minimizing both temperature and product distribution errors was proposed with using non-dominated sorting genetic algorithm (NSGA-Ⅱ). A case study for a commercial hydrocracking reactor with 2 Mt/a capacity was performed. Calculation results indicated that, with using the proposed model and algorithm, the prediction accuracy of products distribution has improved 14.0%.

Key words: hydrocracking, mechanism model, lump, reaction heat; non-dominated sorting genetic algorithm

中图分类号: