See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/232747389

A CFD Study on the Prediction of Cyclone Collection Efficiency

Article  in  International Journal for Computational Methods in Engineering Science and Mechanics · May 2005
DOI: 10.1080/15502280590923649

CITATIONS READS
65 7,206

4 authors, including:

Jolius Gimbun Abdullah Luqman Chuah

Universiti Malaysia Pahang Universiti Putra Malaysia
141 PUBLICATIONS   1,794 CITATIONS    420 PUBLICATIONS   10,330 CITATIONS   

SEE PROFILE SEE PROFILE

A. Fakhru'l-Razi
National Defence University of Malaysia
206 PUBLICATIONS   6,559 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Hydrogen Production from CuCl-HCl Electrolysis View project

methyl ester sulphonates View project

All content following this page was uploaded by Jolius Gimbun on 27 May 2014.

The user has requested enhancement of the downloaded file.

 International Journal for Computational Methods in Engineering Science and Mechanics, 6(3): 161–168, 2005
Copyright
c Taylor & Francis Inc.
ISSN: 1550–2287 print / 1550–2295 online
DOI: 10.1080/15502280590923649

A CFD Study on the Prediction of Cyclone

Collection Efficiency
Jolius Gimbun1, Thomas S. Y. Choong2, T. G. Chuah2, and A. Fakhru’l-Razi2
1
Faculty of Chemical and Natural Resources Engineering, University College of Engineering and
Technology Malaysia, KUKTEM, MEC Town, 25200 Kuantan, Pahang D. M., Malaysia.
2
Department of Chemical and Environmental Engineering, Faculty of Engineering, Universiti
Putra Malaysia, Selangor D. E., Malaysia.

ditions includes the removal of coal dust in a power plant and

This work presents a Computational Fluid Dynamics calcula- the use as a spray dryer or gasification reactor.
tion to predict and to evaluate the effects of temperature, operat- Engineers are generally interested in two parameters in order
ing pressure and inlet velocity on the collection efficiency of gas to carry out an assessment of the design and performance of a cy- 40
10 cyclones. The numerical solutions were carried out using spread-
sheet and commercial CFD code FLUENT 6.0. This paper also clone. These parameters are the collection efficiency of particle
reviews four empirical models for the prediction of cyclone col- and pressure drop through the cyclone. An accurate prediction
lection efficiency, namely Lapple [1], Koch and Licht [2], Li and of cyclone efficiency is very important because an inaccuracy in
Wang [3], and Iozia and Leith [4]. All the predictions proved to be the efficiency prediction may result in an inefficient design of
15 satisfactory when compared with the presented experimental data. the cyclone separator. CFD has a great potential to predict the 45
The CFD simulations predict the cyclone cut-off size for all oper-
ating conditions with a deviation of 3.7% from the experimental flow field characteristics and particle trajectories inside the cy-
data. Specifically, results obtained from the computer modelling clone as well as the pressure drop [6]. The complicated swirling
exercise have demonstrated that CFD model is the best method of turbulent flow in a cyclone places great demands on the numeri-
20 modelling the cyclones collection efficiency. cal techniques and the turbulence models employed in the CFD
codes when modelling the cyclone pressure drop. 50
Keywords Cyclone, CFD, Efficiency, Temperature, Inlet Velocity, This study presents an application of computational fluid dy-
Cut-Off Size namics, in the prediction of cyclone efficiency. This study also
reviews the prediction of four different empirical models for cy-
clone efficiency, namely Lapple [1], Koch and Licht [2], Li and
1. INTRODUCTION
Wang [3], and Iozia and Leith [4]. The simulation results are 55
25 Cyclones are devices that employ a centrifugal force gen- then compared with experimental data found in the literature
erated by a spinning gas stream to separate particles from the for different inlet flow rates, pressures and temperatures. In this
carrier gas. Their simple design, low capital cost and nearly study, the CFD calculations are carried out using a commercial
maintenance-free operation make them ideal for use as pre- finite volume code, FLUENT 6.0, and the empirical models are
cleaners for more expensive final control devices such as bag- performed in Microsoft Excel spreadsheet. 60
30 houses or electrostatic precipitators. Cyclones are particularly
well suited for high temperature and pressure conditions be-
cause of their rugged design and flexible component materials. 2. CYCLONE DESIGN
Cyclone collection efficiencies can reach 99% for particles big- Many different types of cyclones have been built but the re-
ger than 5 µm [5], and can be operated at very high dust loading. verse flow cyclone with tangential inlet in Fig. 1 is most often
35 Cyclones are used for the removal of large particles for both air used for industrial gas cleaning [3, 7]. In this study, the numerical
pollution control and process use. Application in extreme con- simulation deals with the standard case of reverse flow cyclone 65
with a tangential rectangular inlet. Cyclone dimensions used in
this simulation are as shown in Table 1.
Received 6 January 2004; in accepted 25 May 2004.
Address correspondence to Jolius Gimbun, Faculty of Chemical
3. COMPUTATIONAL FLUID DYNAMICS APPROACH
& Natural Resources Engineering, University College of Engineering
& Technology Malaysia, MEC Town, 25200 Kuantan, Pahang D.M., FLUENT is a commercially available CFD code that utilizes
Malaysia. E-mail: [email protected] the finite volume formulation to carry out coupled or segregated 70

161
 162 J. GIMBUN ET AL.

RNG k-
model underestimates the variation of the axial veloc-

ity profile across the radial direction and also overestimates the
magnitude of the tangential velocity and the cyclone pressure 100
drop.
The Reynolds stress model requires the solution of transport
equations for each of the Reynolds stress components as well
as for dissipation transport without the necessity to calculate an
isotropic turbulent viscosity field. The Reynolds Stress turbu- 105
lence model yields an accurate prediction of swirl flow pattern,
axial velocity, tangential velocity and pressure drop on cyclone
simulations [8–10].
The finite volume method has been used to discretize the
partial differential equations of the model using the SIMPLE 110
method for pressure-velocity coupling and the Second Order
FIG. 1. Tangential cyclone configuration. Upwind scheme to interpolate the variables on the surface of
the control volume. The segregated solution algorithm was se-
calculations (with reference to the conservation of mass, mo- lected. The Reynolds stress (RSM) turbulence model was used
mentum and energy equations). It is ideally suited for incom- in this model due to the anisotropic nature of the turbulence in 115
pressible to mildly compressible flows. The conservation of cyclones. Standard Fluent wall functions were applied and high
mass, momentum and energy in a fluid flow are expressed in order discretization schemes were also used.
75 terms of non-linear partial differential equations that generally Under the RSM second order upwind for discretization there
defy solution by analytical means. The solution of these equa- is a difficulty to reach the convergence in simulation [11].
tions has been made possible by the advent of powerful worksta- The residuals may exhibit cyclic tendencies, which means that 120
tions, opening avenues towards the calculation of complicated the transient pattern occurs. In this instance, the solver must
flow fields with relative ease. be changed to a transient solver and this makes the time step
80 For the turbulent flow in a cyclone the key to the success of something in the region of 0.025 seconds or a tiny fraction of
CFD lies with the accurate description of the turbulent behavior the residence time of the cyclone. The simulation is then solved
of the flow [6]. To model the swirling turbulent flow in a cyclone with a coupling of unsteady and steady state solvers in FLUENT. 125
separator, there are a number of turbulence models available in For the simulation using RNG k-
model the steady state solver
FLUENT. These range from the standard k-
model to the more is sufficient to reach convergence.
85 complicated Reynolds stress model (RSM). The comparison of To calculate the trajectories of particles in the flow, the dis-
the different RANS-based turbulence models available in FLU- crete phase model (DPM) was used to track individual particles
ENT 6.0 is presented in Table 2. The k-
model involves the so- through the continuum fluid. The particle loading in a cyclone 130
lution of transport equations for the kinetic energy of turbulence separator is typically small (3–5%), and therefore it can be safely
and its dissipation rate and the calculation of a turbulent contri- assumed that the presence of the particles does not affect the
90 bution to the viscosity at each computational cell. The standard flow field (one-way coupling). The equation of motion for an
k-
, RNG k-
and Realizable k-
models were not optimized for individual particle can be written as Crowe et al., [12]
the strongly swirling flows typically found in cyclones [8, 9].
Turbulence may be stabilized or destabilized in the parts of flow dv f
= (u − v) + g [1]
domain where strong streamline curvature is present. However, dt τv
95 to reduce the computational effort, the RNG k-
model can be
where the other contributions to the force on the particle (buoy- 135
used with about 12% deviation on experimental data [6]. The
ancy, virtual mass and Basset term) are negligible because of
numerical studies carried out by Fredriksson [10] reveal that the
the small fluid-to-particle density ratio. The response time of
the particle, τv is defined in terms of the particle density, particle
TABLE 1 diameter and the viscosity of the air as:
Cyclone geometry used in this simulation
Geometry a/D b/D De /D S/D h/D H/D B/D ρ p d 2p
τv = [2]
18µ
Stairmand High 0.5 0.2 0.5 0.5 1.5 4 0.375
Efficiency The drag factor f is defined as: 140
Kim and Lee 0.33 0.225 0.257 1.157 1.447 3.05 0.482
(1990) cyclone I C D Rer
Bohnet (1995) 0.533 0.133 0.333 0.733 0.693 2.58 0.333 f = [3]
24
Q A CFD STUDY ON THE PREDICTION OF CYCLONE 163

TABLE 2
Comparison of the different turbulence models in FLUENT 6.0

Model Strength Weaknesses

Standard k-ε Robust, economical and reasonably accurate Mediocre results for complex flow involving severe
Many sub-models available, i.e. combustion, pressure gradients, strong streamline curvature,
buoyancy, compressibility, etc. swirl and rotation.
RNG k-ε Good for moderately complex behavior like jet Subjected to limitations due to isotropic eddy
impingement, separating flows, swirling flows, and viscosity assumption.
secondary flows.
Realizable k-ε Offers largely the same benefits as RNG; resolves Subjected to limitations due to isotropic eddy
round jet anomaly. viscosity assumption.
RSM Physically most complete model (history, transport and Requires more CPU effort (2–3 times); limited near
anisotropy of turbulent stresses are all accounted wall modelling options; tightly coupled momentum
for). Most suitable for complex 3D flows with strong and turbulence equations.
streamline curvature, swirl and rotation.

with Pentium IV 2.8 GHz HP workstation XW8000 with 512 cache-

memory, 1 GB RAM-memory, and 110 GB hard-disc memory. 160
ρg d p |u − v|
Rer = [4]
µg 4. CYCLONE EFFICIENCY EMPIRICAL MODELS
where Rer is the relative Reynolds number and C D is the drag co- 4.1. Iozia and Leith Model
efficient. In FLUENT, the drag coefficient for spherical particles Iozia and Leith [4] logistic model is a modified version of
is calculated by using the correlations developed by Morsi and Barth [15] Model, which is developed based on force balance.
145 Alexander [13]. For non-spherical particles, the correlation was The model assumes that a particle carried by the vortex endures 165
developed by Haider and Levenspiel [14]. The ordinary differ- the influence of two forces: a centrifugal force, Z and a flow
ential equation (Eq. (1)) was integrated along the trajectory of an resistance, W . The collection efficiency ηi of particle diameter
individual particle. Collection efficiency statistics were obtained d pi can be calculated from
by releasing a specified number of monodispersed particles at
1
150 the inlet of the cyclone and by monitoring the number escap- ηi = [5]
ing through the underflow. Collisions between particles and the 1 + (d pc /d pi )β
walls of the cyclone were assumed to be perfectly elastic (coef- β is an expression for slope parameter derived based on the
ficient of restitution is equal to 1). statistical analysis of experimental data of a cyclone with D = 170
The numerical calculation was made with a fine numerical 0.25 m given as
155 grid as shown in Fig. 2. The numerical grid of cyclone A, B

d pc


ab

2
ab
and C contains 28871, 33056, and 18045 nodes respectively, to β = 0.62 − 0.87 ln + 5.21 ln + 1.05 ln
yield a reasonable prediction. The details of the CFD setting are 100 D2 D2
presented in Table 3. The CFD simulation was performed with a [6]

FIG. 2. CFD surface mesh of cyclone used in the simulations.

 164 J. GIMBUN ET AL.

TABLE 3 • The tangential velocity is related to the radius of cy-

Detail on CFD setting clone by: u R n = constant.

Boundary condition The concentration distribution in a cyclone is given as:

Inlet Velocity inlet  
1

Outlet Outflow c0 (rw − rn ) exp −λ K (1+n) r 1+n

c(r, θ ) = [12]
Cyclone wall Standard wall function  rw 

1 1+n dr
Viscous rn exp K (1+n) r

Turbulence Reynolds stress model (RSM) where 190

RNG k-

Discretization (1 − n)(ρ p − ρg )d 2 Q
K = [13]
18 µb rw1−n − rn1−n

Pressure Presto!
Pressure-velocity coupling SIMPLE
Momentum 2nd order upwind and
Turbulence kinetic energy 2nd order upwind
(1 − α)K ww
Turbulence dissipation rate 2nd order upwind λ= [14]
Reynolds stresses 2nd order upwind Dr rwn
Discrete phase modelling The resultant expression of the collection efficiency for particle
Assumption Spherical particle of any size is given as
Maximum number of step
(phase integration) 20000
ηi = 1 − exp{−λθ1 } [15]

where
and d pc is the 50% cut size given by Barth [15]

 0.5 θ1 = 2π(S + L)/a [16] 195

9 µQ
d pc = [7]
πρ p z c vt2max
4.3. Koch and Licht Model
where core length, z c , and core diameter, dc , are given as Koch and Licht [2] collection theory recognized the inher-
ently turbulent nature of cyclones and the distribution of gas
 
H−S residence times within the cyclone. Koch and Licht described
z c = (H − S) − [(dc /B) − 1] for dc > B [8a] particle motion in the entry and collection regions with the ad- 200
(D/B) − 1
zc = H − S for dc < B [8b] ditional following assumptions:

−0.25
1.4 • The tangential velocity of a particle is equal to the tan-

ab De gential velocity of the gas flow, i.e. there is no slip in the
dc = 0.47D [9]
D2 D tangential direction between the particle and the gas.
• The tangential velocity is related to the radius of cy- 205
175 4.2. Li and Wang Model clone by: u R n = constant.
The Li and Wang [3] model includes particle bounce or re-
entrainment and turbulent diffusion at the cyclone wall. A two- A force balance and an equation on the particles collection yields
dimensional analytical expression of particle distribution in the the grade efficiency ηi
cyclone is obtained. Li and Wang model was developed based  0.5/(n+1) 
180 on the following assumptions: Gτi Q
ηi = 1 − exp −2 (n + 1) [17]
D3
• The radial particle velocity and the radial concentration
profile are not constant for uncollected particles within where
the cyclone.
8K c
• Boundary conditions with the consideration of turbu- G= [18]
185 lent diffusion coefficient and particle bounce re- K a2 K b2
entrainment on the cyclone wall are: (12D)0.14 T + 460 0.3
  
n = 1− 1− [19]
c = c0 , at θ = 0 [10] 2.5 530
2
∂c ρ p d pi
Dr = (1 − α)wc, at r = D/2 [11] τi = [20]
∂r 18µ
 A CFD STUDY ON THE PREDICTION OF CYCLONE 165

FIG. 3. Particle trajectories from CFD simulation of different particle size in the Bohnet cyclone at T = 1073 K.

210 G is a factor related to the configuration of the cyclone, n is The efficiency of collection of any size of particle is given by
related to the vortex and τ is the relaxation term.
1
ηi = [23]
4.4. Lapple Model 1 + (d pc /d¯ pi )2
Lapple [1] model was developed based on force balance with-
out considering the flow resistance. Lapple assumed that a par-
215 ticle entering the cyclone is evenly distributed across the inlet 5. RESULT AND DISCUSSION
opening. The particle that travels from inlet half width to the
wall in the cyclone is collected with 50% efficiency. The semi 5.1. Grade Efficiency Prediction under Ambient
empirical relationship developed by Lapple [1] to calculate a Temperature and Pressure
50% cut diameter, d pc , is An accurate prediction of cyclone efficiency under ambi- 225
ent temperature and pressure is important since there are a lot
  12 of applications of cyclone under these conditions. Application
9 µb of the cyclone under room temperature includes the removal
d pc = [21]
2π Ne vi (ρ p − ρg ) of sawdust, grain dust and rock dust. Kim and Lee [16] and
Dirgo and Leith [17] presented experimental data obtained at 230
220 where Ne is the number of revolutions room temperature. The calculated trajectories of 1, 2, 2.5 and
  6 µm particles in the Bohnet cyclone are shown in Fig. 3.
1 H −h While, the CFD flow field simulation on Bohnet cyclone is
Ne = h+ [22]
a 2 presented in Fig. 4. The comparisons between the presented

FIG. 4. CFD flow field simulation on Bohnet cyclone (vi = 8 m/s, T = 293 K).
 166 J. GIMBUN ET AL.

FIG. 5. Calculated and measured collection efficiencies for Kim and Lee [16] FIG. 7. Calculated and measured collection efficiencies for Stairmand high
cyclone (P = 1 Bar, T = 293 K, vi = 4.25 m/s, D = 0.311 m). Data point efficiency cyclone (P = 1 Bar, T = 293 K, vi = 5 m/s, D = 0.305 m). Data
from Kim and Lee (1990). point from Dirgo and Leith [17].

experimental data, empirical models and CFD prediction are the grade efficiency for D = 0.0311 m, as shown in Fig. 5. The
235 shown in Figs. 5 to 7. reason for this disagreement may be caused by the generalized 260
The Li and Wang empirical model prediction is found to form of core length, z c in the Iozia and Leith model, which is de-
agree much better with the data from Kim and Lee, and Dirgo veloped based on the statistical analysis of experimental cyclone
and Leith, compared to the other models developed by Koch and data from cyclone of D = 0.25 m. Therefore, the prediction of
Licht, Iozia and Leith, and Lapple (Figs. 5 to 7). Lapple’s model the model is only satisfactory for cyclone diameter around this
240 yields less accurate fitting to the experimental data (curves are range. 265
flatter at higher particle size), as does the Koch and Licht model. The CFD simulations yielded very good predictions on cy-
Both models considerably underestimate efficiency for large par- clone collection efficiency under ambient temperature and pres-
ticles and overestimate efficiency for small particles. The Lapple sure operating condition, as shown in Figs. 5 to 7. The accu-
model is unable to fit well with any experimental data. This is racy of the CFD prediction on cyclone collection efficiency is
245 possibly because the Lapple model simply assumes that parti- comparable to the Li and Wang model in all types and size of 270
cles that enter the cyclone are evenly distributed across the inlet cyclones evaluated in this study. There is a slight discrepancy
opening and a particle that travels from the inlet half width to on the CFD prediction as shown in Fig. 5. However, the CFD
the cyclone wall is collected with 50% efficiency. Unjustified as- result still yielded an accurate prediction on cut size diameter,
sumptions of complete and uniform mixing of uncollected dust D pc , of each cyclone under ambient temperature and pressure
250 at any height in the cyclones may also contribute to the dis- condition (Table 4). 275
crepancy between the experimental data and the Koch and Licht
predictions. Mothes and Loffler [18] experimental findings fur-
ther support the fact that there is indeed a concentration gradient 5.2. Grade Efficiency Prediction under Different
in the radial direction of the cyclones. Operating Conditions
Iozia and Leith logistic model predicted the efficiency satis- Ray et al. [19] and Bohnet [20] have done an experiment
255
factory for cyclone of diameter 0.305 m as shown in Fig. 6 and under high temperature and pressure operating conditions. The
7. For smaller cyclone diameters, the prediction of the Iozia and comparison between the experimental data, CFD and the four 280
Leith model is not satisfactory. It considerably overestimates selected empirical model predictions is shown in Figs. 8 and 9.
The prediction of the Li and Wang model under high pressure

FIG. 6. Calculated and measured collection efficiencies for Stairmand high FIG. 8. Calculated and measured collection efficiencies for Stairmand high
efficiency cyclone (P = 1 Bar, T = 293 K, vi = 15 m/s, D = 0.305 m). Data efficiency cyclone (P = 1.7 Bar, T = 293 K, vi = 11 m/s, D = 0.4 m). Data
point from Dirgo and Leith [17]. point from Ray et al. [19].
 A CFD STUDY ON THE PREDICTION OF CYCLONE 167

TABLE 4
Comparison of measured and predicted cut-off size of different cyclones

Models

Cyclone type and experiment value CFD Li and Wang Iozia and Leith Koch and Licht Lapple
Kim and Lee [16] 2.86 2.91 3.05 1.7 0.82 2.52
Dirgo and Leith [17] 5 m/s 6.24 6.14 5.91 6.73 4.72 8.22
Dirgo and Leith [17] 15 m/s 3.06 3.27 3.06 3.34 2.43 4.19
Ray et al. [19] 2.61 2.54 2.67 2.84 2.46 3.57
Bohnet [20] 873 K 2.52 2.75 3.38 1.85 1.54 2.48
Bohnet [20] 1073 K 3.12 3.12 3.83 1.96 1.91 2.48
Average deviation (%) 0 3.67 11.85 21.69 33.28 23.24

operating conditions is good compared to the experimental data 3.7% to the measured value. The Li and Wang, Lapple, Iozia and
as shown in Fig. 8. CFD results and the Iozia and Leith model Leith, and Koch and Licht models were found to be inconsistent
285 also yield a reasonably good prediction on cyclone efficiency in the cut-off size prediction with the deviation ranging from 310
under this operating condition. 11.9 to 33.3% from the measured value.
The data presented by Bohnet [20] concerns experiments at
temperatures above 1000 K. It appears that the CFD code shows 6. CONCLUSIONS
good prediction of cyclone efficiency under extremely high tem- The Li and Wang model and CFD code both predict very
well the cyclone efficiency and cut-off size for any operational
290 peratures, as shown in Fig. 9. The model of Dirgo is found to
overestimate the cyclone collection efficiency under the high conditions. The Li and Wang model and FLUENT CFD code
temperature operating condition (Fig. 9). The models of Koch produce a better fit to the Ray, Dirgo and Leith, and Kim and 315
and Licht, and Lapple still show a reasonably good prediction Lee experimental data respectively. In all operating conditions
under this extreme condition. Meanwhile, Li and Wang model is and cyclone types the FLUENT CFD and Li and Wang model
were found to be much closer to the experimental measurement.
295 found to underestimate the cyclone collection efficiency under
the extreme operating temperatures. However, only the FLUENT CFD code is consistently predicts
the cyclone cut-off size. Therefore, both the Li and Wang model 320
and FLUENT CFD code can be used to evaluate the collection
5.3. Cut-Off Size Prediction efficiency in the cyclone design except for the extreme operat-
Cyclones have been characterized by a cut size (d50 ), which ing temperatures, which is Li and Wang model is less accurate.
defines the particle size for which the cyclone collection effi- The Lapple and Koch and Lich models considerably underesti-
300 ciency is 50%. It is important to know the cyclone cut-off di- mate the efficiency for large particles and overestimate efficiency 325
ameter under certain operational conditions and geometry. The for small particles. Iozia and Leith logistic model show a good
comparison between the experimental data, CFD and the four agreement with an experimental data for the cyclone size range
selected empirical models prediction is shown in Table 4. of D = 0.25–0.4 m, but it is unable to predict correctly the ef-
The simulation results obtained from the computer modelling ficiency for small cyclone (D < 0.1 m). Iozia and Leith model
305 exercise have demonstrated that CFD code is the best method of is only suitable for efficiency prediction of cyclone diameter 330
modelling the cyclones cut-off size with the average deviation of around 0.25 m.

FIG. 9. Separation efficiency of Bohnet (1995) cyclone at high temperature (P = 1 Bar, vi = 8.61 m/s, D = 0.15 m). Data point from Bohnet [20].
 168 J. GIMBUN ET AL.

ACKNOWLEDGEMENT u, v Velocity magnitude (m/s)

The authors would like to thank Dr. Tom Fraser, Fluent India Rer relative Reynolds number
and Fluent Europe UK for their guidance and support. CD drag coefficient
RANS Reynolds Average Navier Stokes

NOMENCLATURE Greek Letters 375

τv particle response time (s)
335 L natural length (m)
a cyclone inlet height (m)
µg gas viscosity (m2 /s)
β slope parameter
b cyclone inlet width (m)
D cyclone body diameter (m) τ relaxation time (s)
cyclone gas outlet diameter (m) θ angular coordinate 380
De
340 H cyclone height (m) α particle bounce or re-entrainment coefficient
h cyclone cylinder height (m) λ characteristic value
ηi grade efficiency of particle size at mid-point of in-
S cyclone gas outlet duct length (m)
ternal i (%)
B cyclone dust outlet diameter (m)
c0 , c1 particle inlet and outlet concentration (kg/m3 )
ρg gas density (kg/m3 ) 385
ρp particle mass density (kg/m3 )
345 d p particle diameter (m)
Dr radial turbulent diffusion coefficient
d pc cut particle diameter collected with 50% efficiency REFERENCES
1. Lapple, C. E., Chem. Eng. 58, 144 (1951).
(m) 2. Koch, W. H., and Licht, W., Chem. Eng., 7, 80 (1977).
n cyclone vortex exponent (0.5 < n < 1) 3. Li Enliang, and Wang Yingmin, A.I.Ch.E. J. 35, 666 (1989). 390
350 Q volumetric gas flow rate (m3 /s) 4. Iozia, D. L., and Leith, D., Aerosol Sci. Technol. 12, 598 (1990).
r radial dimension, rw = D/2 and rn = De /2(m) 5. Silva, P. D., Briens, C., Bernis, A., Powder Technol. 131, 111 (2003).
6. Griffiths, W. D., and Boysan, F., J. Aerosol Sci. 27, 281 (1996).
R radius (m)
7. Altmeyer, S., Mathieu, V., Jullemier, S., Contal, P., Midoux, N., Rode, S.,
T absolute temperature (K) and Leclerc, J.-P., Chem. Eng. Process 43, 511 (2004). 395
w radial particle velocity (rad/s) 8. Reddy, M., Fluent India, Personal Communication, [email protected]
355 wn , ww radial particle velocity at r = rn and r = rw (rad/s) (2003).
d pi diameter of particle in size range i(m) 9. Fraser, T., Personal Communication, [email protected], www.cfd-
online.com (2003).
g gravity acceleration (m/s2 )
10. Fredriksson, C., Exploratory experimental and theoretical studies of cy- 400
G cyclone configuration factor clone gasification of wood powder. Doctoral thesis, Lulea University of
i subscript donates interval n particles size range Technology, Sweden (1999).
360 K a a/D 11. Gimbun, J., Chuah, T. G., Fakhru’l-Razi, A., and Thomas S. Y. Choong,
Kb b/D Chem. Eng. Process (2004) (in press).
Kc cyclone volume constant 12. Crowe, C. T., Sommerfeld, M., and Tsuji, Y., Multiphase Flow with Droplets 405
and Particles. CRC Press, Boca Raton (1998).
Ne number of revolutions Ne of gas spins through a in 13. Morsi, S. A., and Alexander, A. J., J. Fluid Mech. 55, 193 (1972).
the outer vortex 14. Haider, A., and Levenspiel, O., Powder Technol. 58, 63 (1989).
365 vi inlet velocity (m/s) 15. Barth, W., Brennstoff-Wärme-Kraft, 8, 1 (1956).
K cyclone configuration and operating condition con- 16. Kim, J. C., and Lee, K. W., Aerosol Sci. Technol. 12, 1003 (1990). 410
stant 17. Dirgo, J., and Leith, D., Aerosol Sci. Technol. 4, 401 (1985).
18. Mothes, H., and Loffler, F., J. Aerosol Sci. 13, 184 (1982).
zc core length (m) 19. Ray, M. B., Hoffmann, A. C., and Postma, R. S., J. Aerosol Sci. 31, 563
dc core diameter (m) (2000).
370 vt max maximum tangential velocity (m/s) 20. Bohnet, M., Chem. Eng. Process 34, 151 (1995). 415

View publication stats