气固流化床中颗粒团聚物的静态特征和湮灭规律研究

doi:10.11949/0438-1157.20251102

摘要/Abstract

摘要：

气固流化系统中，介尺度结构的形成和动态演化对反应器的三传一反过程具有重要的影响。本文采用CFD-DEM耦合模拟与正交试验设计方法，系统研究了快速流化床中颗粒团聚物的静态（直径、浓度）与动态（寿命）特征及其影响机制。建立了基于DBSCAN和自适应大津法相结合的颗粒尺度的团聚物识别方法，提出了考虑破碎/聚并事件的团聚物寿命定义，并考察气速、颗粒通量、提升管几何结构及颗粒密度等多因素的耦合效应。分析表明气速对团聚物直径影响最大，提升管几何结构对浓度影响最显著，寿命主要由气速与几何结构共同控制；本文建立了团聚物直径预测关联式，揭示了寿命随直径呈先增后减、随浓度单调升高的非线性规律。

关键词: 流化床, 离散颗粒模拟, 计算流体力学, 聚团, 介尺度, 试验设计

Abstract:

In gas-solid fluidization systems, the formation and dynamic evolution of mesoscale structures have significant impacts on the heat and mass transfer and reaction processes. In this article, CFD-DEM coupled simulation and orthogonal experiment design were employed to systematically investigate the static (diameter, concentration) and dynamic (lifetime) characteristics of particle clusters and their influencing mechanisms in a fast fluidized bed. A particle-scale cluster recognition method combining DBSCAN and adaptive Otsu methods was established, and a definition of cluster lifetime considering breakage and coalescence events was proposed. The coupled effects of multiple factors such as gas velocity, solid flux, riser geometry, and particle density were examined. The analysis shows that gas velocity has the greatest impact on cluster diameter, riser geometry has the most significant effect on concentration, and lifetime is mainly controlled by gas velocity and geometric structure. A correlation for predicting cluster diameter was established, and further study reveals that lifetime first increases and then decreases with diameter but monotonically increases with concentration.

Key words: fluidized bed, discrete particle simulation, computational fluid dynamics, cluster, mesoscale, experiment design

中图分类号:

TQ 028.1

周芷康, 胡善伟, 刘新华. 气固流化床中颗粒团聚物的静态特征和湮灭规律研究[J]. 化工学报, DOI: 10.11949/0438-1157.20251102.

Zhikang ZHOU, Shanwei HU, Xinhua LIU. Investigation on the static properties and annihilation characteristics of particle clusters in gas-solid fluidized beds[J]. CIESC Journal, DOI: 10.11949/0438-1157.20251102.

图/表 26

图1 实验装置及模拟计算域示意图

Fig. 1 Schematic diagrams of the experimental and the computational domain

表1 DEM模拟参数表

Table 1 Parameters in DEM simulations

名称	符号	值
颗粒直径（μm）	d_p	850
气体密度（kg/m³）	ρ_g	1.225
气体黏度（Pa·s）	μ_g	1.8×10^-5
CFD步长（s）	∆t_gas	5.0×10^-5
DEM步长（s）	∆t_DEM	1/20×T_collide
网格分辨率（-）	-	3d_p
弹性恢复系数（颗粒-颗粒）	e_p-p	0.96
弹性恢复系数（颗粒-墙壁）	e_p-w	0.86
摩擦系数（颗粒-颗粒）	μ_fr	0.15
弹簧刚度系数（N/m）	k_n	1600

表2 无量纲工况参数

Table 2 Dimensionless operating parameters

试验号	$W / d p$	$H / W$	$U g / g d p 1 / 2$	$ρ p / ρ g$	$G s / (ρ g U g)$	时均颗粒数
1	60	12	55	1600	6	27820
2	60	14	60	1800	8	36019
3	60	16	65	2000	10	39803
4	60	18	70	2200	12	45105
5	70	12	65	2200	8	38044
6	70	14	70	2000	6	26055
7	70	16	55	1800	12	71497
8	70	18	60	1600	10	60931
9	80	12	70	1800	10	41477
10	80	14	65	1600	12	50581
11	80	16	60	2200	6	79376
12	80	18	55	2000	8	95094
13	90	12	60	2000	12	80448
14	90	14	55	2200	10	125792
15	90	16	70	1600	8	46484
16	90	18	65	1800	6	74453

表2 无量纲工况参数

Table 2 Dimensionless operating parameters

试验号	$W / d p$	$H / W$	$U g / g d p 1 / 2$	$ρ p / ρ g$	$G s / (ρ g U g)$	时均颗粒数
1	60	12	55	1600	6	27820
2	60	14	60	1800	8	36019
3	60	16	65	2000	10	39803
4	60	18	70	2200	12	45105
5	70	12	65	2200	8	38044
6	70	14	70	2000	6	26055
7	70	16	55	1800	12	71497
8	70	18	60	1600	10	60931
9	80	12	70	1800	10	41477
10	80	14	65	1600	12	50581
11	80	16	60	2200	6	79376
12	80	18	55	2000	8	95094
13	90	12	60	2000	12	80448
14	90	14	55	2200	10	125792
15	90	16	70	1600	8	46484
16	90	18	65	1800	6	74453

表3 实际操作参数

Table 3 Operating parameters corresponding to Table 2

试验号	$W / m$	$H / m$	$U g / (m / s)$	$ρ p / (k g / m 3)$	$G s / (k g / m 2 s)$
1	0.051	0.612	5.0	1960	36.9
2	0.051	0.714	5.5	2205	53.7
3	0.051	0.816	5.9	2450	72.7
4	0.051	0.918	6.4	2695	94.0
5	0.060	0.714	5.9	2695	58.2
6	0.060	0.833	6.4	2450	47.0
7	0.060	0.952	5.0	2205	73.8
8	0.060	1.071	5.5	1960	67.1
9	0.068	0.816	6.4	2205	78.3
10	0.068	0.952	5.9	1960	87.3
11	0.068	1.088	5.5	2695	40.3
12	0.068	1.224	5.0	2450	49.2
13	0.077	0.918	5.5	2450	80.5
14	0.077	1.071	5.0	2695	61.5
15	0.077	1.224	6.4	1960	62.6
16	0.077	1.377	5.9	2205	43.6

表3 实际操作参数

Table 3 Operating parameters corresponding to Table 2

试验号	$W / m$	$H / m$	$U g / (m / s)$	$ρ p / (k g / m 3)$	$G s / (k g / m 2 s)$
1	0.051	0.612	5.0	1960	36.9
2	0.051	0.714	5.5	2205	53.7
3	0.051	0.816	5.9	2450	72.7
4	0.051	0.918	6.4	2695	94.0
5	0.060	0.714	5.9	2695	58.2
6	0.060	0.833	6.4	2450	47.0
7	0.060	0.952	5.0	2205	73.8
8	0.060	1.071	5.5	1960	67.1
9	0.068	0.816	6.4	2205	78.3
10	0.068	0.952	5.9	1960	87.3
11	0.068	1.088	5.5	2695	40.3
12	0.068	1.224	5.0	2450	49.2
13	0.077	0.918	5.5	2450	80.5
14	0.077	1.071	5.0	2695	61.5
15	0.077	1.224	6.4	1960	62.6
16	0.077	1.377	5.9	2205	43.6

图2 DBSCAN算法示意图

Fig. 2 Schematic diagram of DBSCAN algorithm

图3 团聚物寿命定义示意图

Fig. 3 Schematic diagram of the definition of cluster lifetime

图4 不同气速下时均颗粒浓度的轴向分布注:(a) Ug=5.55 m/s (b) Ug=5.95 m/s (c) Ug=6.35 m/s (d) Ug=6.74 m/s

Fig. 4 Time-averaged axial profiles of solid volume fraction at different gas velocities

图5 不同气速下时均颗粒浓度的横向分布注:(a) Ug=5.55 m/s (b) Ug=5.95 m/s (c) Ug=6.35 m/s (d) Ug=6.74 m/s(a) Ug=5.55 m/s (b) Ug=5.95 m/s (c) Ug=6.35 m/s (d) Ug=6.74 m/s

Fig. 5 Time-averaged lateral profiles of solid volume fraction at different gas velocities

图6 算例1中不同时刻下的r.d.f函数

Fig. 6 r.d.fs at different instants in case 1

图7 minPts的确定

Fig. 7 Determination of minPts

图8 瞬时团聚物识别结果(a)与性质统计(b)

Fig. 8 Instantaneous recognition results (a) and distributions of cluster properties (b).

图9 正交数值模拟试验算例颗粒浓度瞬时分布（为了方便对比，对反应器大小进行了缩放）

Fig. 9 Instantaneous contours of solid volume fraction in the central plane in different cases (note the size of the reactors has been scaled).

图10 团聚物直径分布

Fig. 10 Distributions of cluster diameters for different cases

图11 团聚物直径箱线图。其中箱体内的中心线和矩形分别表示中位数和均值，箱体上下边界分别为上、下四分位数，须线延伸至1.5倍四分位距内的最值，圆点为离群值。

Fig. 11 Boxplots of cluster diameters. The center line and rectangle within the box represent the median and mean, respectively. The upper and lower boundaries of the box are the upper and lower quartiles. The whiskers extend to the maximum values within 1.5 times the interquartile range, and the dots indicate outliers.

图12 各因素对团聚物直径影响的极差图

Fig. 12 Range values of different factors on cluster diameters under different operating conditions

表4 团聚物平均直径的多元线性回归方程系数表

Table 4 Multiple linear regression equation of average cluster diameter

	式（18）	式（19）	式（20）
常数项	29.34	22.42	22.38
$H / W$	-	-	0.0027
$G s / (U g ρ g)$	0.44	0.44	0.44
$U g / g d p 1 / 2$	-0.18	-0.18	-0.18
$ρ p / ρ g$	-	0.0036	0.0036
p值	0.023	0.012	0.032
显著性	显著	显著	显著
Pearson系数	0.66	0.77	0.77

表4 团聚物平均直径的多元线性回归方程系数表

Table 4 Multiple linear regression equation of average cluster diameter

	式（18）	式（19）	式（20）
常数项	29.34	22.42	22.38
$H / W$	-	-	0.0027
$G s / (U g ρ g)$	0.44	0.44	0.44
$U g / g d p 1 / 2$	-0.18	-0.18	-0.18
$ρ p / ρ g$	-	0.0036	0.0036
p值	0.023	0.012	0.032
显著性	显著	显著	显著
Pearson系数	0.66	0.77	0.77

图13 基于式（19）的团聚物平均直径预测值与真实值对比

Fig. 13 Comparison between the predicted and actual cluster diameters based on Eq. (19)

图14 团聚物颗粒浓度分布

Fig. 14 Distributions of cluster solid holdup for different cases

图15 团聚物颗粒浓度箱线图

Fig. 15 Boxplots of cluster solid holdup for different cases

图16 各因素对团聚物颗粒浓度影响的极差图

Fig. 16 Range values of different factors on cluster solid holdup under different operating conditions

图17 典型团聚物的动态演化过程：(a)破碎 (b)聚并 (c)生长与湮灭

Fig. 17 Dynamic evolution processes of typical clusters: (a)breakup (b)coalescence (c)growth and dissipation

图18 两种不同定义下的团聚物寿命分布

Fig. 18 Distributions of cluster lifetimes under the two definitions: (a) Liu’s method; (b) this study

图19 不同寿命定义方式下各因素对团聚物寿命影响的极差图

Fig. 19 Range values of different factors on cluster lifetime under the two definitions

图20 团聚物在床层内的位置分布及其与寿命的关系

Fig. 20 Relationship between the initial locations of cluster and their lifetimes

图21 团聚物平均寿命沿反应器宽度方向的分布

Fig. 21 Distributions of the averaged cluster lifetime along the bed width

图22 团聚物寿命随团聚物特征的变化趋势

Fig. 22 Variations of the averaged cluster lifetime with cluster properties: (a) cluster diameter (b) cluster solid holdup

参考文献 47

[1]	Afsahi F A, Sotudeh-Gharebagh R, Mostoufi N. Clusters identification and characterization in a gas–solid fluidized bed by the wavelet analysis[J]. The Canadian Journal of Chemical Engineering, 2009, 87(3): 375-385.
[2]	Cocco R, Shaffer F, Hays R, et al. Particle clusters in and above fluidized beds[J]. Powder Technology, 2010, 203(1): 3-11.
[3]	Boyd J W R, Varley J. The uses of passive measurement of acoustic emissions from chemical engineering processes[J]. Chemical Engineering Science, 2001, 56(5): 1749-1767.
[4]	Liu X H, Gao S Q, Li J H. Characterizing particle clustering behavior by PDPA measurement for dilute gas–solid flow[J]. Chemical Engineering Journal, 2005, 108(3): 193-202.
[5]	Yang T Y, Leu L P. Multiresolution analysis on identification and dynamics of clusters in a circulating fluidized bed[J]. AIChE Journal, 2009, 55(3): 612-629.
[6]	Liu H P, Zhang L Y, Chen T P, et al. Experimental study on the fluidization behaviors of the superfine particles[J]. Chemical Engineering Journal, 2015, 262: 579-587.
[7]	Harris A T, Davidson J F, Thorpe R B. The prediction of particle cluster properties in the near wall region of a vertical riser (200157)[J]. Powder Technology, 2002, 127(2): 128-143.
[8]	Teplitskii Y S. On the cluster structure of a circulating fluidized bed[J]. Journal of Engineering Physics and Thermophysics, 2005, 78(2): 316-322.
[9]	Horio M, Morishita K, Tachibana O, et al. Solid distribution and movement in circulating fluidized beds[M]//Circulating Fluidized Bed Technology. Amsterdam: Elsevier, 1988: 147-154.
[10]	Mostoufi N, Chaouki J. Flow structure of the solids in gas–solid fluidized beds[J]. Chemical Engineering Science, 2004, 59(20): 4217-4227.
[11]	Wang C X, Lan X Y, Sun Z N, et al. Cluster identification by a k-means algorithm-assisted imaging method in a laboratory-scale circulating fluidized bed[J]. Industrial & Engineering Chemistry Research, 2022, 61(1): 942-956.
[12]	Xu J, Zhu J. A new method for the determination of cluster velocity and size in a circulating fluidized bed[J]. Industrial & Engineering Chemistry Research, 2012, 51(4): 2143-2151.
[13]	Lin Q, Wei F, Jin Y. Transient density signal analysis and two-phase micro-structure flow in gas–solids fluidization[J]. Chemical Engineering Science, 2001, 56(6): 2179-2189.
[14]	Yu J H, Wang S, Luo K, et al. CFD-DEM modeling of dense gas-solid reacting flow in the framework of GPU[J]. Chemical Engineering Journal, 2024, 484: 149480.
[15]	Banaei M, Jegers J, van Sint Annaland M, et al. Tracking of particles using TFM in gas-solid fluidized beds[J]. Advanced Powder Technology, 2018, 29(10): 2538-2547.
[16]	Gao X, Li T W, Rogers W A. Assessment of mesoscale solid stress in coarse-grid TFM simulation of Geldart A particles in all fluidization regimes[J]. AIChE Journal, 2018, 64(10): 3565-3581.
[17]	Xiong Q G, Li B, Zhou G F, et al. Large-scale DNS of gas–solid flows on Mole-8.5 [J]. Chemical Engineering Science, 2012, 71: 422-430.
[18]	Deen N G, Peters E A J F, Padding J T, et al. Review of direct numerical simulation of fluid–particle mass, momentum and heat transfer in dense gas–solid flows[J]. Chemical Engineering Science, 2014, 116: 710-724.
[19]	Yang L N, Han C X, Xu J, et al. Role of mesoscale structure in gas–solid fluidization: Comparison between continuum and discrete approaches[J]. Chemical Engineering Journal, 2023, 454: 139979.
[20]	Jin N N, Hu S W, Liu X H, et al. Machine learning assisted characterization of local bubble properties and its coupling with the EMMS bubbling drag[J]. Industrial & Engineering Chemistry Research, 2024, 63(10): 4631-4646.
[21]	Kumar B K, Ganesh S, Goyal H. Capturing mesoscale structures in computational fluid dynamics simulations of gas-solid flows[J]. AIChE Journal, 2024, 70(5): e18360.
[22]	Wang T Y, Deng A M, He Y R, et al. Artificial intelligence-based approach for cluster identification in a CFB riser[J]. Chemical Engineering Science, 2023, 268: 118379.
[23]	Yin S Y, Zhong W Q, Song T, et al. Clusters identification and meso-scale structures in a circulating fluidized bed based on image processing[J]. Advanced Powder Technology, 2019, 30(12): 3010-3020.
[24]	Petersen A J, Baker L, Coletti F. Experimental study of inertial particles clustering and settling in homogeneous turbulence[J]. Journal of Fluid Mechanics, 2019, 864: 925-970.
[25]	Liu X W, Wang L M, Ge W. Meso-scale statistical properties of gas–solid flow: a direct numerical simulation (DNS) study[J]. AIChE Journal, 2017, 63(1): 3-14.
[26]	Kong L K, Xu J, Wang J W, et al. Nonequilibrium characteristics and spatiotemporal long-range correlations in dense gas-solid suspensions[J]. International Journal of Multiphase Flow, 2021, 142: 103731.
[27]	Zhang H D, Sun Z N, Zhang M Z, et al. Comparison of the flow structures and regime transitions between a cylindrical fluidized bed and a square fluidized bed[J]. Powder Technology, 2020, 376: 507-516.
[28]	Carlos Varas A E, Peters E A J F, Kuipers J A M. CFD-DEM simulations and experimental validation of clustering phenomena and riser hydrodynamics[J]. Chemical Engineering Science, 2017, 169: 246-258.
[29]	Cabezas-Gómez L, da Silva R C, Navarro H A, et al. Cluster identification and characterization in the riser of a circulating fluidized bed from numerical simulation results[J]. Applied Mathematical Modelling, 2008, 32(3): 327-340.
[30]	Lu H L, Sun Q Q, He Y R, et al. Numerical study of particle cluster flow in risers with cluster-based approach[J]. Chemical Engineering Science, 2005, 60(23): 6757-6767.
[31]	Tsuo Y P, Gidaspow D. Computation of flow patterns in circulating fluidized beds[J]. AIChE Journal, 1990, 36(6): 885-896.
[32]	Tsuji Y, Tanaka T, Yonemura S. Cluster patterns in circulating fluidized beds predicted by numerical simulation (discrete particle model versus two-fluid model)[J]. Powder Technology, 1998, 95(3): 254-264.
[33]	Hu S W, Liu X H. A CFD-PBM-EMMS integrated model applicable for heterogeneous gas-solid flow[J]. Chemical Engineering Journal, 2020, 383: 123122.
[34]	Wang H F, Chen Y P, Wang W. Particle-level dynamics of clusters: Experiments in a gas-fluidized bed[J]. AIChE Journal, 2022, 68(3): e17525.
[35]	Tagawa Y, Mercado J M, Prakash V N, et al. Three-dimensional Lagrangian Voronoï analysis for clustering of particles and bubbles in turbulence[J]. Journal of Fluid Mechanics, 2012, 693: 201-215.
[36]	Liu Y Q, Shen L, Zamansky R, et al. Life and death of inertial particle clusters in turbulence[J]. Journal of Fluid Mechanics, 2020, 902: R1.
[37]	Kong L K, Xu J, Wang J W, et al. Characterizing particle clustering behavior in dense gas–solid suspensions[J]. Industrial & Engineering Chemistry Research, 2023, 62(45): 19145-19160.
[38]	Niu L, Chu Z M, Cai M, et al. Modified force balance model of estimating agglomerate sizes in a gas–solid fluidized bed[J]. Industrial & Engineering Chemistry Research, 2019, 58(19): 8472-8483.
[39]	Lu L Q, Gao X, Dietiker J F, et al. Machine learning accelerated discrete element modeling of granular flows[J]. Chemical Engineering Science, 2021, 245: 116832.
[40]	Cundall P A, Strack O D L. A discrete numerical model for granular assemblies[J]. Géotechnique, 1979, 29(1): 47-65.
[41]	Carlos Varas A E, Peters E A J F, Kuipers J A M. Experimental study of full field riser hydrodynamics by PIV/DIA coupling[J]. Powder Technology, 2017, 313: 402-416.
[42]	Garg R, Galvin J, Li T W, et al. Open-source MFIX-DEM software for gas–solids flows: Part I: Verification studies[J]. Powder Technology, 2012, 220: 122-137.
[43]	Tian T, Wang H, Ge W, et al. Detecting particle clusters in particle-fluid systems by a density based method[J]. Communications in Computational Physics, 2025, 26(5): 1617-1630.
[44]	Fortune S. A sweepline algorithm for Voronoi diagrams[J]. Algorithmica, 1987, 2(1): 153-174.
[45]	Wu X, Leung D Y C. Optimization of biodiesel production from camelina oil using orthogonal experiment[J]. Applied Energy, 2011, 88(11): 3615-3624.
[46]	Goldhirsch I I, Zanetti G. Clustering instability in dissipative gases[J]. Physical Review Letters, 1993, 70(11): 1619-1622.
[47]	Fullmer W D, Hrenya C M. The clustering instability in rapid granular and gas-solid flows[J]. Annual Review of Fluid Mechanics, 2017, 49: 485-510.