Professional Documents
Culture Documents
A R T I C L E I N F O A B S T R A C T
Keywords: The micro-scale model of fibrous media was simulated to obtain viscous resistance in clean state, which was
Fibrous media provided to the CFD porous media. A macro-scale model for predicting the pressure drop during dust loading was
CFD porous Media established based on the Euler-Lagrangian method by implementing the UDF in ANSYS-Fluent. The trapping
Pressure drop
mechanism programs were developed to realize different distribution forms of particles in the porous media. The
Dust loading
macro-empirical model was used to calibrate the macro-scale model under different SVFs. The 592*592*292-4
Coupling simulation
V–F9 filter was simplified to CFD porous media and the numerical simulation was carried out during dust
loading. The results show that the pressure drop shows different growth trends under different trapping mech
anism programs. Very good agreement was found between the experimental and simulated pressure drop during
dust loading. Thus, the macro-scale model broadens the way of studying the pressure drop of fibrous media in
depth filtration.
* Corresponding author.
E-mail address: fpingqian@ahut.edu.cn (F. Qian).
https://doi.org/10.1016/j.buildenv.2021.108015
Received 10 March 2021; Received in revised form 18 May 2021; Accepted 26 May 2021
Available online 7 June 2021
0360-1323/© 2021 Elsevier Ltd. All rights reserved.
B. Cao et al. Building and Environment 202 (2021) 108015
performance under different structure and operating parameters during conducted a numerical simulation on particle deposition characteristics
dust loading. In summary, a large number of researchers have been used of fiber filter based on porous media, and then estimated the relationship
the microstructure model to directly carry out the numerical simulation between pressure drop and dust deposition. Yousefi et al. [24] devel
of filtration performance during dust loading. However, the amount of oped a new method for the microstructure model of dust deposition with
calculation is very large based on the current numerical calculation arbitrary shape and multi dispersed particles.
capacity. In this work, due to the limitation of CFD porous media, we adopted
At present, numerical simulation is widely used to analyze the the user defined function in ANSYS-Fluent to develop it again, and then
filtration performance of fiber filtration equipment (e.g. bag-type and established a macro-scale model to simulate the pressure drop of fibrous
cartridge-type dust collector, air filter). However, the structure of fiber media in the depth filtration. As shown in Fig. 1, The technical route of
filtration equipment involves a huge scale span, it will cause a huge this research is as follows: (1) the flow field in the micro-scale model of
amount of numerical calculation to establish the fibrous material fibrous media was numerically simulated to obtain its viscous resistance
directly and simulate its gas-solid two phase flow based on the micro in clean state, which was provided to the CFD porous media as the initial
structure. With the development and application of multi-scale analysis parameter; (2) A macro-scale model was established based on the Euler-
method, the study of complex flow is also promoted to the level of multi- Lagrangian method by means of using the user defined function in
scale analysis. The numerical simulation are used to analyze the flow at ANSYS-Fluent; (3) Different forms of trapping and loading of particles in
each scale, and an effective data transfer mechanism between scales is the CFD porous media were realized, so as to calculate the pressure drop
established. This method can not only meet the requirements of system in the depth filtration; (4) The macro-empirical model in the literature
analysis of complex flow problems under multi-scale coupling condi was used to calibrate the macro-scale model under different SVFs.
tions, but also greatly reduce the amount of numerical calculation. In
order to simulate the flow field in the fiber filtration equipment, the 2. Micro-flow simulation of fibrous media
fibrous material is usually simplified as a CFD porous media [21]. The
biggest advantage of this method is that the cost of simulation calcula 2.1. 3D modeling of fibrous media
tion is low, but the disadvantages are shown in the two aspects: (1) This
method does not pay attention to the microstructure characteristics of The modeling method is based on the characteristics of layered
the fibrous media; (2) It is impossible to capture particles and calculate fibrous media [18], and such model accounts for the bending, defor
the pressure drop in the porous media, which will seriously affect the mation, penetration characteristics and arrangement of fibers. The
flow field distribution in the fiber filtration equipment. geometric topology of layered fibrous media is modeled programmati
Therefore, three key issues need to be solved to simulate the gas-solid cally by in MATLAB environment, and the point, line, surface and vol
two phase flow: (1) the trapping of particles in the porous media; (2) the ume elements needed to establish the 3D model are saved as txt script
dynamic change of the pressure drop in the porous media during dust files. Then, the pre-processor of ANSYS-Fluent calls the log file com
loading; (3) the different distribution forms of dust in the porous media mand to read the script file to build the fibrous media. As shown in
are realized based on the characteristics of surface filtration and depth Fig. 2, it is an improved micro-model of fibrous media. The fiber
filtration. The filtration mechanism of high-efficiency filters is mainly diameter and layer spacing of the reconstructed model are both 15 μm.
so-called surface filtration regime, where most of the dust accumulates By changing the number of fibers, a total of 6 models with SVFs of
on the surface of the filter, and then forms a dust-cake (initial adhesion 8.77%, 11.71%, 15.87%, 17.51%, 19.22%, and 21.98% were obtained.
layer) [22]. According to the characteristics of surface filtration, some
researchers have developed the calculation method for fibrous media
during dust loading based on CFD porous media. Saleh et al. [23]
2
B. Cao et al. Building and Environment 202 (2021) 108015
2.2. Calculation of flow field As mentioned above, the arrangement of fibers in the computational
domain is irregular. Hence, ICEM CFD was used to generate an un
Since the fiber size and operating parameters considered in this structured tetrahedral grid to realize the discretization of the compu
study, and the Reynolds number is less than unity. Therefore, it is tational domain [29]. In order to accurately simulate the flow field of the
considered that the air flow through the fibrous media is laminar and fibrous media, the grid independence must be verified. There are two
incompressible [7]. The Stokes flow is assumed in the computational settings for the global element seed size of the computational domain.
domain and solved by the finite volume method in ANSYS-Fluent. The mesh at the fiber needs to be locally densified, and the maximum
Continuity equation and the Navier-Stokes (N–S) equations are size of the fiber at the three settings is defined. The Detailed parameter
expressed as follows [25]: settings of the grid are listed in Table 1. As the number of grids increases,
the pressure drop tends to stabilize, as shown in Fig. 4. Therefore, the
∇⋅→
u (1)
third grid setting is adopted in this study to balance the computational
efficiency and reliability.
d→u
ρ = − ∇p + η∇2 →
u + ρ→
g (2)
dt
2.3. Dimensionless pressure drop
where → u is the fluid velocity vector, ρ, p, η are the velocity vector,
density, pressure, and dynamic viscosity of the air respectively. → g is the The gas through the fibrous media is laminar flow, so the pressure
gravitational acceleration. The computational domain and boundary drop of fibrous media follows Darcy’s law [2,29], which can be sum
conditions for simulation were shown in Fig. 3. For the continuous marized as the follows.
phase, the air flow inlet boundary was imposed “velocity inlet”, and the
boundary condition of “pressure outlet” was adopted at the outlet. At the
Δp ηV
= f (α) 2 (3)
same time, the uniform inlet and outlet boundary conditions should be L df
placed far away from the region with larger velocity and pressure
where f (α) is dimensionless pressure drop, η is dynamic viscosity of air,
gradient [26]. The pressure drop of fibrous media was extracted through
df is fiber diameter and L is fibrous media thickness. According to Dar
the “interior”. Since the fiber in the computational domain is taken from
cy’s law, there is a quantitative relationship between the slope of the
the overall fiber filter, and there is no significant lateral airflow in the
fitting line (S) and the structural parameters of the fibrous media (see
fibrous media, the symmetry boundary was used for the surrounding
Fig. 5), and the dimensionless pressure drop can be obtained by the
surface of the computational domain [27]. In this study, the Knudsen
following equation.
number of the fiber is far less than unity, so the fiber surface was set to be
the no-slip boundary condition [28]. Sdf2
f (α) = (4)
ηL
As shown in Fig. 6, the f (α) ~ SVF fitting curve of fibrous media was
obtained by six groups of flow field simulation data. It can be seen that
with the increase of SVF, the internal structure of the fibrous media
becomes more complicated, and the deviation between the numerical
results and the Davies experimental values [30] increases. However, the
fitting result of this study and Davies experimental value are in good
agreement with SVF ranging from 5% to 25%. Therefore, it is reliable to
provide the fitting results in this range as initial parameters to the CFD
porous media.
Table 1
Parameter setting of grid.
Mesh Global element seed size (mm) Maximum mesh size of the fibers (mm)
1 0.032 0.008
2 0.032 0.006
3 0.032 0.004
4 0.016 0.004
3
B. Cao et al. Building and Environment 202 (2021) 108015
two terms: one due to viscous loss effect and the other due to inertial
effect [21].
( )
∑3 ∑3
1
Si = − Dij ηVi + Tij ρ|V|Vj (8)
j=1 j=1
2
1
C1 = (10)
K
∑3
η
With current numerical simulation capabilities, the computational Δpy = vj Δny (13)
costs are very huge to simulate the particle deposition characteristics on j=1
K yj
4
B. Cao et al. Building and Environment 202 (2021) 108015
Fig. 5. Fitting line of filtration velocity and pressure drop under different SVFs.
after the current time step is completed, and record the data into UDM3. in the program, the enumeration form and its physical meaning of UDMs
The porosity and viscous resistance were recorded by UDM4 and UDM5 were given in Table 2.
respectively. The viscous resistance of porous media in each direction is It is worth noting that the position of trapped particles in porous
defined by using Define_Profile Macro. After a time step is completed, media will have a great impact on the pressure drop. Fig. 9 shows the
the porosity of each the cell in the porous media was recalculated to trapping mechanism of particles in the porous media in this study. A
update the viscous resistance coefficient. In order to further explain the threshold (X) was set for the porosity of the cell in porous media to
UDMs used in the modeling process and make it convenient to call UDMs realize different deposition forms along the thickness direction. For
5
B. Cao et al. Building and Environment 202 (2021) 108015
completely trapped in the surface cell, and the threshold (X) is greater
than 0 and less than the porosity in the clean state. When UDM4 reaches
this threshold, the particles are no longer trapped by the current cell and
move towards the downstream cell. In this section, we have compared
four kinds of trapping mechanism programs (X = 0.75ε, 0.80ε, 0.85ε and
0.90ε). As shown in Fig. 10, the mass flow rate of particles was 1*10− 6
kg/s, and the different particle distribution forms along the thickness
direction of the porous media under four trapping mechanism programs
were given. It can be seen that with the increase of threshold, more
Table 2
The enumeration types and physical meanings of UDM in this study.
UDM Enumerated types Physical meanings
6
B. Cao et al. Building and Environment 202 (2021) 108015
particles will be trapped by the downstream cell, and the thickness of model for dynamic prediction of pressure drop for multi-layer fibrous
dust layer will increase. It is worth mentioning that because of the use of media. Based on Model A, Thomas et al. [35] presented a quantitative
CFD porous media, it is only a technical method to simulate the particle computational model for predicting the dynamic pressure drop of
capture in the filter, which is different from the real filtration process. multi-layer fibrous media. Leung and Hung [34] proposed a quantitative
mathematical for predicting the dynamic change of pressure drop for
fibrous media. On the other hand, based on the Davies model, it is
3.3. Solver parameter setting assumed that the fiber diameter and solid volume fraction are increased
due to dust loading (Model B), Zhao et al. [36] proposed a quantitative
The simulation of the micro-scale model and the gas-solid two phase mathematical model for the dynamic change of the pressure drop of the
flow of the macro-scale model were performed on a computer with Intel fibrous media which is suitable for the depth filtration. The expression of
Xeon CPU E5-26700@2.60GHz and memory of 32 GB. For transient the dynamic pressure drop is as follows.
calculation, the monitoring object are velocity and continuity in x, y and
z directions. The convergence residuals of each time step in all directions Δp = f (α)t ηVL (16)
are lower than 10− 6. The parameter settings of the solver are listed in ( )
Table 3. f (α)t = F αt , dft (17)
αt = Kd2 α (18)
4. Results and discussion
dft = Kd df (19)
4.1. The macro-empirical model
( /( ))1/2
The effect of particle the deposition pattern should be considered Kd = 1 + M αρp L (20)
when predicting the dynamic change of pressure drop. In order to
where f(α)t is the equivalent dimensionless pressure drop, αt is the
establish a mathematical model to dynamically predict the pressure
equivalent solid volume fraction, dft is the equivalent fiber diameter, Kd
drop of fibrous media during dust loading, two ideal deposition models
(Model A and Model B) [34] were proposed by simplifying the deposi is the deposition coefficient, and M is the particle deposition per unit
tion model of particles. In Model A, it is assumed that dust loading is area (kg/m2). The macro-empirical model was used to calibrate the
considered as “new fibers”. In Model B, it is assumed that dust loading threshold (X) in the particle trapping mechanism under different SVFs.
changes the structural parameters of the original fiber. Model A was In this process, three kinds of fibrous media with different SVFs (10%,
suitable for a larger number of deposited particles during dust loading 15%, 20%) were selected. Firstly, the viscous resistance in clean state
(the surface filtration) and Model B was more appropriate for the initial was provided to the porous media. Then, three kinds of macro-scale
stage of the dust loading [18]. models were simulated under four different particle trapping mecha
However, few researchers proposed the quantitative mathematical nisms. Finally, the best particle trapping mechanism was selected for
different macro-scale models through error analysis.
Table 3
Key parameter settings in the simulation. 4.2. Calibration of the macro-scale model with different SVFs
Parameters Values
It is shown that the method is reliable to extract the viscous resis
Particle density (kg⋅m− 3) 1550
tance from the simulation results of the micro-scale model of fibrous
Particle diameter (μm) 5
Total flow rate (kg⋅s− 1) 1*10− 6 media and provide it to the porous media as the initial parameter. As
Time step for CFD (s) 0.02 shown in Fig. 11, during dust loading, with the increase of the particle
Time step for DPM (s) 0.02 loading, the pressure drop under different trapping mechanism pro
Simulation time (s) 7 grams presents different growth trends. For the macro-scale model with
Number of user-defined memory locations 6
the same SVF, with the increase of the threshold (X), more dust will be
7
B. Cao et al. Building and Environment 202 (2021) 108015
8
B. Cao et al. Building and Environment 202 (2021) 108015
captured by the downstream cell in porous media, which makes the dust
distribution more uniform, so the growth rate of pressure drop de
creases. The macro-scale model established in this study was calibrated
by using the calculation value of the macro-empirical model of pressure
drop during dust loading. The macro-scale model with 10%, 15% and
20% SVF was calibrated. It is more reasonable to select the threshold (X)
of porous media cell to be 0.85ε, 0.85ε and 0.80ε respectively for
selecting the trapping mechanism. For the macro-scale model with
different SVFs after calibration, the deviation between the numerical
simulation results and the calculation values of macro-empirical model
is less than 20%. It is worth mentioning that with the increase of SVF, the
threshold (X) decreases, and more dust will be captured by the surface
cell in porous media. The results of CFD-DEM simulation in Ref. [18] can
also support this explanation.
Fig. 13. The filter performance test bench.
9
B. Cao et al. Building and Environment 202 (2021) 108015
Table 4 and (2) the dust generation test is conducted at the air volume of 3400
The test data of pressure drop for F9 filter material in clean state. m3/h (face velocity of 2.81 m/s), and the filtration velocity is about 5.4
Sample Filtration velocity’ Pressure drop cm/s. The first dust generation was 30 g, and the pressure drop of the
filter was calculated after the dust generation; (3) the air blowing was
(L/min) (cm/s) (mmH2O) (Pa)
stopped for 5 min, and the weight gain of the filter was recorded; (4) the
F9 filter material 0 0.0 0.0 0.0 pressure drop of the filter was 150, 250, 350 and 450 pa by four times of
21.0 3.5 0.9 8.8
32.4 5.4 1.3 12.7
dust generation, and the weight gain of the filter after each dust gen
43.4 7.2 1.7 16.7 eration was recorded.
54.4 9.1 2.1 20.6 Table 6 shows the pressure drop test data of 592*592*292-4 V–F9
64.5 10.8 2.4 23.5 filter during dust loading, and the corresponding relationship between
76.5 12.8 2.8 27.5
pressure drop and dust holding capacity was obtained. As shown in
Fig. 17, the fitting curve between the pressure drop of dust layer and the
particle deposition per unit area was obtained by fitting the five groups
of data in Table 6. The expression is shown in equation (22).
the rated air volume of 3400 m3/h, the filtration velocity is about 5.4 Table 5
cm/s. Pressure drop test data of 592*592*292-4 V–F9 filter under different air volume.
The test of the dust loading of the filter is divided into the following Air volume (m3/h) 1700 2550 3400 4250
steps: (1) as shown in Table 5, the pressure drop of the filter is tested
Pressure drop (Pa) 26.5 46.2 72.2 103.7
under four operating air volumes (1700, 2550, 3400 and 4250 m3/h);
10
B. Cao et al. Building and Environment 202 (2021) 108015
Table 6
Pressure drop test data of 592*592*292-4 V–F9 filter under different dust
holding capacity.
Test Single dust Cumulative Average Cumulative Pressure
emission dust emission weight weight gain drop
(g) (g) gain (g) (g) (Pa)
1 0 0 0 0 72.2
2 30 30 27.50 27.50 89.0
3 55 85 50.42 77.92 150.0
4 60 145 55.01 132.93 250.0
5 53 198 48.58 181.51 350.0
6 41 239 37.59 219.10 450.0
Fig. 19. Comparison of numerical simulation results and the test values.
(1) The flow field of the micro-scale model of the layered structure
fibrous media was numerically simulated, and the viscous resis
tance under the clean state is obtained, which is provided to the
porous media as the initial parameter. This method is accurate
and reliable.
(2) A macro-scale model was established based on the CFD porous
Fig. 17. Fitting curve of pressure drop and dust holding capacity of dust layer. media by means of the user defined function to simulate the
pressure drop in the depth filtration. In other words, the function
of particle trapping in porous media and the dynamic change of
pressure drop during dust loading are realized.
(3) By controlling the threshold of the cell in porous media, the
different trapping mechanism programs are developed to realize
the different distribution forms of particles along the thickness
direction. Under the same dust loading, with the increase of
threshold, more particles will be trapped by the downstream cell.
With the increase of the dust loading, the pressure drop under
different trapping mechanism programs presents different
growth trends.
(4) The 592*592*292-4 V–F9 filter used in the test was simplified to
CFD porous media and the numerical simulation was carried out
during dust loading. The deviation between the numerical
simulation results and the test values is basically within 30%.
Thus, it provides guidance for the simulation of gas-solid two
phase flow in dust collector and filter.
11
B. Cao et al. Building and Environment 202 (2021) 108015
Acknowledgment [17] X.D. Liu, X.Y. Ding, C. Chen, R.R. An, W.Y. Guo, W. Zhang, Investigating the
filtration behavior of metal fiber felt using CFD-DEM simulation, Engineering
Applications of Computational Fluid Mechanics 13 (2019) 426–437, https://doi.
This study was financially supported by Anhui Provincial Scientific org/10.1080/19942060.2019.1608306.
and Technological Major Project (Grant no. 18030801109). [18] F.P. Qian, N.J. Huang, J.L. Lu, Y.L. Han, CFD–DEM simulation of the filtration
performance for fibrous media based on the mimic structure, Comput. Chem. Eng.
71 (2014) 478–488, https://doi.org/10.1016/j.compchemeng.2014.09.018.
Appendix A. Supplementary data [19] C. Yue, Q. Zhang, Z.Q. Zhai, Numerical simulation of the filtration process in
fibrous filters using CFD-DEM method, J. Aerosol Sci. 101 (2016) 174–187,
Supplementary data to this article can be found online at https://doi. https://doi.org/10.1016/j.jaerosci.2016.08.004.
[20] R. Tao, M.M. Yang, S.Q. Li, Effect of adhesion on clogging of microparticles in fiber
org/10.1016/j.buildenv.2021.108015. filtration by DEM-CFD simulation, Powder Technol. 360 (2020) 289–300, https://
doi.org/10.1016/j.powtec.2019.09.083.
References [21] J.L. Li, P. Wang, D.S. Wu, D.R. Chen, Numerical study of opposing pulsed-jet
cleaning for pleated filter cartridges, Separ. Purif. Technol. 234 (2020) 116086,
https://doi.org/10.1016/j.seppur.2019.116086.
[1] W.J. Ji, Z.Z. Liu, C.H. Liu, C.W. Wang, X.F. Li, Characteristics of fine particulate
[22] S. Bourrous, L. Bouilloux, F.X. Ouf, P. Lemaitre, P. Nerisson, D. Thomas, et al.,
matter and volatile organic compounds in subway station offices in China, Build.
Measurement and modeling of pressure drop of HEPA filters clogged with ultrafine
Environ. 188 (2021) 107502, https://doi.org/10.1016/j.buildenv.2020.107502.
particles, Powder Technol. 289 (2016) 109–117, https://doi.org/10.1016/j.
[2] E.Z. Tian, J.H. Mo, Toward energy saving and high efficiency through an optimized
powtec.2015.11.020.
use of a PET coarse filter: the development of a new electrostatically assisted air
[23] A.M. Saleh, S. Fotovati, H.V. Tafreshi, B. Pourdeyhimi, Modeling service life of
filter, Energy Build. 186 (2019) 276–283, https://doi.org/10.1016/j.
pleated filters exposed to polydispersed aerosols, Powder Technol. 266 (2014)
enbuild.2019.01.021.
79–89, https://doi.org/10.1016/j.powtec.2014.06.011.
[3] I. Nezis, G. Biskos, K. Eleftheriadis, O.L. Kalantzi, Particulate matter and health
[24] S.H. Yousefi, H.V. Tafreshi, Novel approach to model microstructure of dust-
effects in offices - a review, Build. Environ. 156 (2019) 62–73, https://doi.org/
deposits comprised of polydisperse particles of arbitrary shapes, Separ. Purif.
10.1016/j.buildenv.2019.03.042.
Technol. 244 (2020) 116844, https://doi.org/10.1016/j.seppur.2020.116844.
[4] J. Cai, Air Filtration ABC, China Architecture and Building Press, 2002.
[25] J.J. Wu, O. Akampumuza, P.H. Liu, Z.Z. Quan, H.N. Zhang, X.H. Qin, et al., 3D
[5] R.R. Cai, H. Lu, L.Z. Zhang, Evaluation the effect of fiber alignment on particle
structure design and simulation for efficient particles capture: the influence of
collection performance of mechanical/electret filters based on Voronoi
nanofiber diameter and distribution, Materials Today Communications 23 (2020)
tessellations, Chem. Eng. Sci. 197 (2019) 109–119, https://doi.org/10.1016/j.
100897, https://doi.org/10.1016/j.mtcomm.2020.100897.
ces.2018.12.015.
[26] H. Bai, X.M. Qian, J.T. Fan, Y. Qian, Y.S. Liu, Y.C. Duo, et al., Micro-scale layered
[6] M.D.R. Kok, R. Jervis, T.G. Tranter, M.A. Sadeghi, D.J.L. Brett, P.R. Shearing, et al.,
structural filtration efficiency model: probing filtration properties of non-uniform
Mass transfer in fibrous media with varying anisotropy for flow battery electrodes:
fibrous filter media, Separ. Purif. Technol. 236 (2020) 116037, https://doi.org/
direct numerical simulations with 3D X-ray computed tomography, Chem. Eng. Sci.
10.1016/j.seppur.2019.116037.
196 (2019) 104–115, https://doi.org/10.1016/j.ces.2018.10.049.
[27] L. Zhang, J.W. Zhou, B. Zhang, W. Gong, Semi-analytical and computational
[7] R.R. Cai, L.Z. Zhang, Y.Y. Yan, Performance prediction of PM 2.5 removal of real
investigation of different fibrous structures affecting the performance of fibrous
fibrous filters with a novel model considering rebound effect, Appl. Therm. Eng.
media, Sci. Prog. 103 (2019) 1–25, https://doi.org/10.1177/0036850419874231.
111 (2017) 1536–1547, https://doi.org/10.1016/j.applthermaleng.2016.07.162.
https://orcid.org/0000-0002-2327-3031.
[8] W.Y. Zhang, S.M. Deng, Y.X. Wang, Z.P. Lin, Modeling the surface filtration
[28] W. Sambaer, M. Zatloukal, D. Kimmer, 3D modeling of filtration process via
pressure drop of PTFE HEPA filter media for low load applications, Build. Environ.
polyurethane nanofiber based nonwoven filters prepared by electrospinning
177 (2020) 106905, https://doi.org/10.1016/j.buildenv.2020.106905.
process, Chem. Eng. Sci. 66 (2011) 613–623, https://doi.org/10.1016/j.
[9] E. Sikorska, J.M. Gac, L. Gradoń, Performance of a depth fibrous filter at
ces.2010.10.035.
particulate loading conditions. Description of temporary and local phenomena with
[29] F.P. Qian, N.J. Huang, X.J. Zhu, J.L. Lu, Numerical study of the gas-solid flow
structure development, Chem. Eng. Res. Des. 132 (2018) 743–750, https://doi.org/
characteristic of fibrous media based on SEM using CFD-DEM, Powder Technol.
10.1016/j.cherd.2018.02.020.
249 (2013) 63–70, https://doi.org/10.1016/j.powtec.2013.07.030.
[10] D. Thomas, S. Pacault, A. Charvet, N. Bardin-Monnier, J.C. Appert-Collin,
[30] C.N. Davies, Air Filtration, Academic Press, London, 1973, https://doi.org/
Composite fibrous filters for nano-aerosol filtration: pressure drop and efficiency
10.1016/0021-9797(74)90095-2.
model, Separ. Purif. Technol. 215 (2019) 557–564, https://doi.org/10.1016/j.
[31] Q. Wang, B. Maze, H.V. Tafreshi, B. Pourdeyhimi, A case study of simulating
seppur.2019.01.043.
submicron aerosol filtration via lightweight spun-bonded filter media, Chem. Eng.
[11] S.A. Hosseini, H.V. Tafreshi, Modeling particle-loaded single fiber efficiency and
Sci. 61 (2006) 4871–4883, https://doi.org/10.1016/j.ces.2006.03.039.
fiber drag using ANSYS–Fluent CFD code, Comput. Fluids 66 (2012) 157–166,
[32] B.W. Cao, S.L. Wang, W. Dong, J.J. Zhu, F.P. Qian, J.L. Lu, et al., Investigation of
https://doi.org/10.1016/j.compfluid.2012.06.017.
the filtration performance for fibrous media: coupling of a semi-analytical model
[12] O. Filippova, D. Hänel, Lattice-Boltzmann simulation of gas-particle flow in filters,
with CFD on Voronoi-based microstructure, Separ. Purif. Technol. 251 (2020)
Comput. Fluids 26 (1997) 697–712, https://doi.org/10.1016/S0045-7930(97)
117364, https://doi.org/10.1016/j.seppur.2020.117364.
00009-1.
[33] S. Park, Y.H. Joe, J. Shim, H. Park, W.G. Shin, Non-uniform filtration velocity of
[13] H.M. Wang, H.B. Zhao, Z.L. Guo, C.G. Zheng, Numerical simulation of particle
process gas passing through a long bag filter, J. Hazard Mater. 365 (2019)
capture process of fibrous filters using Lattice Boltzmann two-phase flow model,
440–447, https://doi.org/10.1016/j.jhazmat.2018.10.098.
Powder Technol. 227 (2012) 111–122, https://doi.org/10.1016/j.
[34] W.F. Leung, C.H. Hung, Investigation on pressure drop evolution of fibrous filter
powtec.2011.12.057.
operating in aerodynamic slip regime under continuous loading of sub-micron
[14] S.Q. Li, J.S. Marshall, Discrete element simulation of micro-particle deposition on a
aerosols, Separ. Purif. Technol. 63 (2008) 691–700, https://doi.org/10.1016/j.
cylindrical fiber in an array, J. Aerosol Sci. 38 (2007) 1031–1046, https://doi.org/
seppur.2008.07.015.
10.1016/j.jaerosci.2007.08.004.
[35] D. Thomas, P. Contal, V. Renaudin, et al., Modelling pressure drop in hepa filters
[15] A.M. Saleh, S.A. Hosseini, H.V. Tafreshi, B. Pourdeyhimi, 3-D Microscale
during dynamic filtration, J. Aerosol Sci. 30 (1999) 235–246, https://doi.org/
simulation of dust-loading in thin flat-sheet filters: a comparison with 1-D
10.1016/S0021-8502(98)00036-6.
macroscale simulations, Chem. Eng. Sci. 99 (2013) 284–291, https://doi.org/
[36] Z.M. Zhao, I.T. Gabriel, R. Pfeffer, Separation of airborne dust in electrostastically
10.1016/j.ces.2013.06.007.
enhanced fibrous filters, Chem. Eng. Commun. 108 (1991) 307–332, https://doi.
[16] Z.Y. Shu, F.P. Qian, J.J. Zhu, J.L. Lu, Numerical simulation of gas-solid flow
org/10.1080/00986449108910964.
characteristic of particles in fibrous media using OpenFOAM, Indoor Built Environ.
[37] European Committee for Standardization, European Committee for
29 (2020) 921–930, https://doi.org/10.1177/1420326X19862016.
Standardization, European, 2012. ⟪EN779: 2012⟫.
12