欢迎访问过程工程学报, 今天是

过程工程学报 ›› 2022, Vol. 22 ›› Issue (9): 1181-1191.DOI: 10.12034/j.issn.1009-606X.221420

• 研究论文 • 上一篇    下一篇

双分散颗粒湍动床反应器的多尺度模拟及细粉夹带优化

段宏霖1,2， 杜承哲2,3， 鲁波娜2,3*， 许友好4*， 王维2,3*， 周建文5， 徐莉4， 谢英鹏1
  

  1. 1. 沈阳化工大学化学工程学院，辽宁 沈阳 110142 2. 中国科学院过程工程研究所，北京 100190 3. 中国科学院大学，北京 101408 4. 中国石化石油化工科学研究院，北京 100083 5. 中国石化金陵分公司，江苏 南京 210033
  • 收稿日期:2021-12-14 修回日期:2021-12-30 出版日期:2022-09-28 发布日期:2022-10-09
  • 通讯作者: 鲁波娜 bnlu@ipe.ac.cn
  • 作者简介:段宏霖(1996-)，男，山东省宁津县人，硕士研究生，化学工程专业，E-mail: 2300164139@qq.com；通讯联系人，鲁波娜，E-mail: bnlu@ipe.ac.cn；许友好，E-mail: xuyouhao.ripp@sinopec.com；王维，E-mail: wangwei@ipe.ac.cn.
  • 基金资助:
    国家自然科学基金;中国石油化工股份有限公司;多相复杂系统国家重点实验室自主研究课题

Multiscale CFD simulation of bidisperse turbulent bed reactors with optimization of fine powder entrainment

Honglin DUAN1,2,  Chenzhe DU2,3,  Bona LU2,3*,  Youhao XU4*,  Wei WANG2,3*,  Jianwen ZHOU5,   Li XU4,  Yingpeng XIE1   

  1. 1. School of Chemical Engineering, Shenyang University of Chemical Technology, Shenyang, Liaoning 110142, China 2. Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China 3. University of Chinese Academy of Sciences, Beijing 101408, China 4. Sinopec Research Institute of Petrochemical Technology, Beijing 100083, China 5. Sinopec Jinling Company, Nanjing, Jiangsu 210033, China
  • Received:2021-12-14 Revised:2021-12-30 Online:2022-09-28 Published:2022-10-09
  • Supported by:
    National Natural Science Foundation of China;China Petroleum & Chemical Corporation LTD;State Key Laboratory of Multiphase Complex Systems

RichHTML

PDF

摘要： S-zorb吸附脱硫工艺技术是生产超低硫含量汽油的关键技术。目前，S-zorb湍动床反应器内的细粉夹带严重，容易堵塞其顶部过滤器，是影响反应器使用寿命和工艺流程运行周期的主要原因。因此，深入研究其反应器内的吸附剂颗粒和细粉颗粒的运动行为对优化反应器和提升工艺运行效率非常重要。本研究采用多尺度计算流体力学方法对双分散颗粒S-zorb湍动床反应器进行了模拟研究，考察了减少细粉夹带的优化方案。首先，比较了不同非均匀EMMS-bubbling曳力模型和EMMS-ANN模型对双分散颗粒运动行为的影响，发现采用上述两个模型都能合理预测吸附剂颗粒的流动行为和浓度分布，但EMMS-ANN模型对细粉颗粒的夹带量预测与工厂实验值的偏差大于前者。考察了一系列设计改造方案对减少细粉夹带的影响，结果表明，适当增大沉降段圆柱高度和半径以及增设固相出口横管都有助于减少细粉夹带，其中在反应区和沉降段之间的过渡区增设固相出口横管以及增大沉降段半径的方案对减少细粉夹带最有效。研究结果将有助于S-zorb工艺的进一步优化和升级。

关键词: 双分散颗粒, S-zorb, 模拟, 曳力, 能量最小多尺度, 人工神经网络

Abstract: S-zorb adsorption desulfurization technology is one of the most important technologies to produce the ultra-low sulfur content gasoline. However, the short operation cycle of the turbulent bed reactor limits the wide application and efficient production of S-zorb process. The main reason is that the fine particles produced by the circulation and attrition of absorbent particles are easily carried upwards to the disengager and then gradually deposited on the filter, leading to the blocking and shutdown of the top filter. In order to improve the design of S-zorb reactor and reduce the entrainment of fine particles, a deep understanding of hydrodynamic behaviors in the reactor is very necessary. In this work, a series of multiscale CFD simulations of the S-zorb reactor with absorbent large particles and fine particles were carried out and the optimization schemes were proposed. As the drag force played very important role in predicting heterogeneous gas-solid fluidized flows, the effects of drag models, i.e., the EMMS-bubbling model and EMMS-ANN model were first investigated. It was found that using both the EMMS-bubbling model and EMMS-ANN model can reasonably predict the flow distribution of absorbent particles and segregation behaviors. Compared to using the EMMS-bubbling model, using the EMMS-ANN model over-predicted the entrainment of fine particles. Then a series of design modification schemes for reducing entrainment of fine powder were investigated. It was found that increasing the height of the disengager from 0.848 to 1.150, the radius of the disengager from 0.909 to 1.212 or adding a horizontal pipe for discharging solids can help reduce the carryover of fine powders. Among these modification schemes, the increase in the disengager radius and adding a horizontal pipe at the transitional section between the reaction zone and disengager were the most effective to reduce the carryover of fine particles. These findings were very helpful for optimization of S-zorb fluidized reactor and process upgrading.

Key words: Bidispersed particle, S-zorb, simulation, Drag Force, EMMS, artificial neural network