岳爽¹， 于伟¹， 张宁²， 袁方洋¹

1.江南大学 机械工程学院， 江苏 无锡 214122； 2.中国科学院上海应用物理研究所， 上海 201800引用格式：

岳爽， 于伟， 张宁， 等. 熔盐堆尾气处理系统内扰流柱对颗粒聚集的影响［J］. 中国粉体技术， 2027， 33（2）： 1-13.

Yue Shuang， Yu Wei， Zhang Ning， et al. Influence of disturbance columns on particle agglomeration in off-gas treatment system of a molten-salt reactor［J］. China Powder Science and Technology， 2027， 33（2）： 1-13.

DOI：10.13732/j.issn.1008-5548.2027.02.004

收稿日期： 2026-07-01， 修回日期： 2026-08-27， 上线日期： 2026-09-29。

基金项目： 国家自然科学基金项目，编号：12172152。

第一作者： 岳爽（2003—），男，硕士生，研究方向为微化工机械。E-mail：15255837533@163.com。

通信作者： 于伟（1987—），男，副教授，博士，硕士生导师，研究方向为微化工机械。E-mail：yuwei0301@163.com。

摘要： 【目的】 明确不同扰流结构对干态微米颗粒接触聚集行为的影响，量化颗粒接触关系及其持续特征。【方法】 采用计算流体力学-离散元法（computational fluid dynamics-discrete element method，CFD-DEM）双向耦合方法，对圆柱、波浪柱和十字形柱3种单扰流结构开展数值模拟；基于JKR（Johnson-Kendall-Roberts）接触模型和Rocky输出信息，统计颗粒-颗粒接触对（contact pair，CP）、接触团簇（contact cluster，CC）、持续接触对（persistent contact pair，PCP）及压降归一化指标。【结果】 十字形柱形成更宽的尾流扰动区和更复杂的剪切层，稳定统计时间0.04~0.10 s内平均CP数量约为7 849，较圆柱和波浪柱分别提高87.4%和99.4%；颗粒数大于等于4的颗粒CC为295.2 个/帧，平均最大CC尺寸为23.2。粒径分辨结果表明，粒径为20 μm的颗粒更易参与接触，粒径为10 μm-20 μm和20 μm-20 μm组合贡献较高。【结论】 十字形柱可显著增强表观接触聚并倾向，但伴随较高压降；流向臂加长的十字形柱（Cross-2）综合表现最优，可作为后续结构优化重点方案。

关键词： 微米颗粒； 扰流柱； 接触聚集； 计算流体力学-离散元法； 接触团簇

Abstract

Objective This study investigates how the geometry of disturbance columns affects the contact-based agglomeration tendency of dry micron-sized particles in the off-gas treatment system of a molten salt reactor. Rather than relying solely on macroscopic quantities such as outlet concentration or mean particle size， particle-scale contact relationships extracted from discrete element method outputs are analyzed. The objectives are to quantify contact pairs （CP）， contact clusters （CC）， persistent contact pairs （PCP）， size-resolved contact participation， and the pressure-drop cost of contact enhancement， and to identify a cross-shaped column for subsequent low-resistance optimization. These contact-based indicators characterize particle contact opportunities and contact network development rather than irreversible physical agglomeration.

Methods A two-way coupled computational fluid dynamics-discrete element method （CFD-DEM） model was established using ANSYS Fluent and Rocky DEM to compare the performance of a circular cylinder， a wavy column， and a cross-shaped column under identical operating conditions. The computational domain was 25 mm × 15 mm×2 mm， the inlet velocity was 1.5 m/s， and the column center was located 5 mm downstream of the inlet. Particles with diameters of 2， 5， 10， and 20 μm and a density of 2 000 kg/m³ were injected at 2×10⁶ particles/s for each size. The gas-phase flow was resolved using the SST k-ω-based stress-blended eddy simulation （SBES） model， with the wall-adapting local eddy-viscosity （WALE） model applied in the scale-resolving region. Pressure-velocity coupling was performed using the PISO algorithm， and dry adhesive contacts were described by the Johnson-Kendall-Roberts （JKR） model with a particle-particle surface energy of 1 J/m². Mesh independence was assessed using three mesh densities for each geometry， and the medium-density meshes were selected to balance wake resolution and computational cost. The total simulation time was 0.10 s， with statistics calculated over 0.04-0.10 s after the initial filling stage. A CP was defined as a pair of particles in direct contact within an output frame； a CC was defined as a connected particle-contact network； and a PCP was defined as a unique particle pair that remained continuously in contact for at least 2， 3， or 5 consecutive frames. Additional indicators included particle-size-specific contact participation rates， particle-size-combination contributions， pressure-normalized contact-pair and contact-cluster indices， and a power-normalized contact-pair index. Five cross-shaped candidates， Cross-0—Cross-4， were further evaluated using the CRITIC-TOPSIS method.

Results and Discussion The three disturbance columns produced distinct wake structures and particle-contact responses. The circular cylinder generated a conventional bluff-body wake with a localized velocity deficit and alternating shear layers， whereas the wavy column extended the disturbed region downstream but provided limited transverse entrainment. The cross-shaped column introduced multiple separation points and interacting shear layers， creating a wider and more heterogeneous wake that increased particle migration， local residence time， and trajectory overlap. During the stable statistical window， the average numbers of contact pairs were 4 187.6， 3 937.4， and 7 849.2 for the circular cylinder， wavy column， and cross-shaped column， respectively. Therefore， the cross-shaped column increased the average CP number by 87.4% and 99.4% relative to the other two structures. It also promoted the formation of multi-particle contact networks， producing 460.6 three-particle clusters and 295.2 clusters with at least four particles per frame. Its average maximum cluster size was 23.2 particles， with an instantaneous maximum of 33 particles. Although most contacts were short-lived， the cross-shaped column maintained substantially more long-duration contacts. Under the criterion of at least five consecutive frames， 2 747 unique PCPs were identified， compared with 13 for the circular cylinder and 10 for the wavy column.Contact participation increased with particle diameter. For the cross-shaped column， the participation rates of 2， 5， 10， and 20 μm particles were 2.82%， 3.72%， 6.64%， and 19.04%， respectively. Different-size contacts dominated for all geometries， while the cross-shaped column increased the contribution of same-size contacts to 32.40%. The 10 μm-20 μm and 20 μm-20 μm pairs showed the strongest enhancement， reaching （2 488.6±175.3） and （2 161.0±191.2） CPs， respectively. This enhancement resulted in a higher hydraulic penalty， with pressure drops of 0.498， 1.970， and 2.437 Pa for the circular cylinder， wavy column， and cross-shaped column， respectively. Consequently， the circular cylinder achieved the highest contact-pair number per unit pressure drop， whereas the cross-shaped column provided the greatest absolute enhancement in CPs and CCs. These normalized indicators represent contact opportunities relative to flow resistance rather than the net particle-collection efficiency. Among the five cross-shaped candidates，Cross-2， with a longer streamwise arm，achieved the highest CRITIC-TOPSIS score of 1.000 and the best overall performance in pressure-normalized contact-pair enhancement，maximum cluster size，persistent contacts，and contacts involving 20 μm particles.

Conclusion The cross-shaped column most effectively strengthens the contact-based agglomeration tendency of dry micron-sized particles by widening the wake，increasing particle trajectory overlap，and sustaining larger contact networks.Its advantage is especially evident for contacts involving 20 μm particles and persistent contacts， but it is accompanied by a higher pressure drop. The circular cylinder is more economical when contact enhancement is evaluated per unit pressure drop. Under the present operating conditions and evaluation framework， Cross-2 provides the best overall balance among the screened cross-shaped columns and is recommended as the reference configuration for further low-resistance optimization while maintaining multiple shear layers and a broad wake.

Keywords： micron-sized particle； disturbance column； contact-based agglomeration； computational fluid dynamics-discrete element method； contact cluster

参考文献（References）

［1］刘含笑， 郦建国， 姚宇平， 等. PM_2.5湍流聚并方法研究进展［J］. 中国环保产业， 2013（4）： 27-30.

Liu Hanxiao， Li Jianguo， Yao Yuping， et al. Research progress on PM_2.5 turbulent flows and assembling method［J］. China Environmental Protection Industry， 2013（4）： 27-30.

［2］刘忠， 刘含笑， 冯新新， 等. 湍流聚并器流场和颗粒运动轨迹模拟［J］. 中国电机工程学报， 2012， 32（14）： 71-75.

Liu Zhong， Liu Hanxiao， Feng Xinxin， et al. Simulation for the flow field of the turbulence coalescence device and the trajectory of particles［J］. Proceedings of the CSEE， 2012， 32（14）： 71-75.

［3］Rosenthal M W. An account of Oak Ridge National Laboratory’s thirteen nuclear reactors： ORNL/TM-2009/181［R/OL］. 2009-08-01. https：//doi.org/10.2172/970897.

［4］张书斌， 陈占秀， 杨历， 等. 绕流圆柱对气固两相流场颗粒聚并效果的研究［J］. 节能， 2016， 35（11）： 24-28.

Zhang Shubin， Chen Zhanxiu， Yang Li， et al. Study on particle agglomeration in gas-solid two-phase flow around a circular cylinder［J］. Energy Conservation， 2016， 35（11）： 24-28.

［5］王国昌，刘玺璞， 米建春.聚并元件结构和来流颗粒粒径对细颗粒物湍流聚并的影响［J］.环境工程学报，2021，15（1）：253-261.

Wang Guochang， Liu Xipu， Mi Jianchun. Effect of agglomeration element structure and incoming particle size on turbulent agglomeration of fine particles［J］. Chinese Journal of Environmental Engineering， 2021， 15（1）： 253-261.

［6］章鹏飞， 米建春， 潘祖明. 装置元件排列间距和颗粒浓度对细颗粒湍流聚并的影响［J］. 中国电机工程学报， 2016， 36（6）： 1625-1632.

Zhang Pengfei， Mi Jianchun， Pan Zuming. Influences of elemental arrangement and particle concentration on fine particle amalgamation［J］. Proceedings of the CSEE， 2016， 36（6）： 1625-1632.

［7］李正鸿， 鹤欣， 杨富鑫， 等. 扰流件排列对亚微米颗粒湍流团聚效率影响研究［J］. 洁净煤技术， 2020， 26（5）： 173-180.

Li Zhenghong， Liu Hexin， Yang Fuxin， et al. Study on the effect of disturbing structure arrangement on the turbulent agglomeration efficiency of the submicron particulate［J］. Clean Coal Technology， 2020， 26（5）： 173-180.

［8］Liu Hexin， Yang Fuxin， Tan Houzhang， et al. Experimental and numerical investigation on the structure characteristics of vortex generators affecting particle agglomeration［J］. Powder Technology， 2020， 362： 805-816.

［9］Sun Zongkang， Yang Linjun， Chen Shuai， et al. Promoting the removal of fine particles by turbulent agglomeration with the coupling of different-scale vortexes［J］. Powder Technology， 2020， 367： 399-410.

［10］Zhang Yumeng， Chen Xianying， Wei Di， et al. Investigation on fine particle agglomeration and separation promoted by different bluff bodies［J］. Journal of Cleaner Production， 2022， 374： 134039.

［11］Wang Hainuo， Zhang Yumeng， Yan Shijun， et al. Promotion of particle agglomeration by different vortex generator angles based on PIV-PDPA experiment［J］. Chemical Engineering Journal， 2024， 486： 150308.

［12］Zheng Kaixin， Yan Xiaokang， Wang Lijun， et al. Turbulent effects of vortex generators on the separation of fine particles［J］. Chemical Engineering Journal， 2021， 418： 129373.

［13］王爽， 浦航， 尚妍， 等. 扰流柱对微米颗粒湍流团聚特性的影响研究［J］. 工程热物理学报， 2022， 43（12）： 3305-3311.

Wang Shuang， Pu Hang， Shang Yan， et al. Study on the influence of disturbance columns on turbulent agglomeration characteristics of micron particles［J］. Journal of Engineering Thermophysics， 2022， 43（12）： 3305-3311.

［14］Wang Shuang，Mu Lin，Li Xue， et al. Turbulent agglomeration of microparticles in a cylinder wake flow using LES-DEM： focusing on the effect of the Reynolds number［J］. Journal of Thermal Science， 2025， 34（1）： 34-49.

［15］Wang Shuang， Mu Lin， Wang Chu， et al. Modeling and simulation of micron particle agglomeration in a turbulent flow： impact of cylindrical disturbance and particle properties［J］. ACS Omega， 2024， 9： 49302-49315.

［16］徐止恒， 李政权， 王贻得， 等. 基于CFD-DEM的湿颗粒气力输送数值模拟［J］. 中国粉体技术， 2024， 30（2）： 12-23.

Xu Zhiheng， Li Zhengquan， Wang Yide， et al. Numerical simulation of pneumatic conveying of wet particles based on CFD-DEM［J］. China Powder Science and Technology， 2024， 30（2）： 12-23.

［17］陈伟， 张佩， 孙永昌， 等. 基于CFD-DEM的非球形颗粒水力输送数值模拟［J］. 中国粉体技术， 2022，28（5）： 82-91.

Chen Wei， Zhang Pei， Sun Yongchang， et al. Numerical simulation of hydraulic transport of non-spherical particles based on CFD-DEM［J］. China Powder Science and Technology， 2022， 28（5）： 82-91.

［18］李东晖， 柳波， 张晓仪， 等. 撞击流中颗粒运动行为的CFD-DEM模拟［J］. 中国粉体技术， 2021， 27（5）： 11-19.

Li Donghui， Liu Bo， Zhang Xiaoyi， et al. CFD-DEM simulation of particle motion behavior in impinging stream［J］. China Powder Science and Technology， 2021， 27（5）： 11-19.

［19］Kuang Shibo， Zhou Mengmeng， Yu Aibing. CFD-DEM modelling and simulation of pneumatic conveying： a review［J］. Powder Technology， 2020， 365： 186-207.

［20］Kuang Shibo， Li Ke， Yu Aibing. CFD-DEM simulation of large-scale dilute-phase pneumatic conveying system［J］. Industrial & Engineering Chemistry Research， 2020， 59（9）： 4150-4160.

［21］Trofa M， D'Avino G， Sicignano L， et al. CFD-DEM simulations of particulate fouling in microchannels［J］. Chemical Engineering Journal， 2019， 358： 91-100.

［22］Liu Daoyin， van Wachem B G M， Mudde R F， et al. Characterization of fluidized nanoparticle agglomerates by using adhesive CFD-DEM simulation［J］. Powder Technology， 2016， 304： 198-207.

［23］Chen Zhifan， Duan Angxu， Liu Yang， et al. Discrete element contact model and parameter calibration of sticky particles and agglomerates［J］. Journal of Terramechanics， 2024， 116： 100998.

［24］Liu Xin， Hessels C J M， Deen N G， et al. CFD-DEM investigation on the agglomeration behavior of micron-sized combusted iron fines［J］. Fuel， 2023， 346： 128219.

［25］Zou Yi， Zou Ruiping， Wu Yongli. CFD-DEM study on agglomeration and spout-assisted fluidization of cohesive particles［J］. Powder Technology， 2024， 436： 119512.

［26］Saparbayeva N， Balakin B V. CFD-DEM model of plugging in flow with cohesive particles［J］. Scientific Reports， 2023， 13： 17188.

［27］Chen Jian， Krengel D， Nishiura D， et al. A force-displacement relation based on the JKR theory for DEM simulations of adhesive particles［J］. Powder Technology， 2023， 427： 118742.

［28］Hærvig J， Kleinhans U， Wieland C， et al. On the adhesive JKR contact and rolling models for reduced particle stiffness discrete element simulations［J］. Powder Technology， 2017， 319： 472-482.