Professional Documents
Culture Documents
net/publication/270491263
CITATIONS READS
15 3,766
2 authors:
All content following this page was uploaded by Andre Mesquita on 08 June 2016.
(Submitted: January 26, 2012 ; Revised: May 01, 2013 ; Accepted: May 10, 2013)
Abstract - This paper presents a comparative study of the critical pickup and saltation velocities of particles
in horizontal pipelines for pneumatic conveying design. A comparative study is performed using different
existing correlations in the literature for the determination of the minimum velocity of transport as a function
of the particle and pipe diameter, particle density, solid mass flow rate and particle sphericity. Their
limitations and difficulties in predicting those critical velocities are analyzed. For the pickup velocity, an
experimental study was also carried out in order to support the analysis. Recommendations are presented on
the use of such correlations.
Keywords: Pickup velocity; Saltation velocity; Pneumatic conveying; Gas-solid flow.
several limitations and contradictions, as reviewed Main Correlations Used for Predicting Critical
by Yi et al. (1998) and Rabinovich and Kalman Saltation Velocity
(2008).
The goal of this paper is to provide a critical Rizk (1976)
analysis of the pickup and saltation velocities,
considering the influence of parameters such as Ws ⎛ 1 ⎞
particle and pipeline diameters, particle density and ζ= = ⎜ d ⎟ FrSx (1)
ρf USA ⎝ 10 ⎠
solids loading ratio. For the pickup velocity, an
experimental study was also carried out to support
the analysis. Thus, this paper presents recommenda- where:
tions on the limitations of the existing correlations. FrS = US / g DT (Froude number at the Saltation
velocity), d = 1.44dp + 1.96 , x = 1.1dp + 2.5 , (d and
LITERATURE REVIEW x in mm), ζ is the solids loading ratio, WS is the
solids mass flow rate, dp is the particle diameter, DT
The minimum transport velocity is defined as the is the pipe diameter, US is the saltation velocity, ρf is
lowest velocity at which particles can be transported the fluid density, g is the acceleration of gravity and
inside a pipeline in a stable manner without deposi- A is the cross-sectional area of pipe.
tion on the bottom of the pipe. The saltation velocity
is defined as the gas velocity in a horizontal pipeline Matsumoto et al. (1977)
in which the particles begin to fall from their state of
suspension and are deposited at the bottom of the For:
pipeline.
The pickup velocity is defined as the gas velocity −0.74
⎛ ρp ⎞
necessary to suspend the particles initially at rest in d p > 1.39 DT ⎜ ⎟
the bottom of the pipeline (see Figure 1), or it may ⎝ ρf ⎠
be defined as the fluid velocity required to initiate a (2)
1.06 −3.7 3.61
sliding motion, rolling and suspension of the ⎛ ρp ⎞ ⎛ U ⎞ ⎛ U min ⎞
particles. ζ = 0.373 ⎜ ⎟ ⎜ t ⎟ ⎜⎜ ⎟⎟
⎝ ρf ⎠ ⎜ 10 gd p ⎟ ⎝ 10 gDT ⎠
⎝ ⎠
For:
−0.74
⎛ ρp ⎞
d p < 1.39 DT ⎜ ⎟
⎝ ρf ⎠
(a) (3)
1.43 3.61
⎛ dp ⎞ ⎛ U min ⎞
ζ = 5560 ⎜ ⎟ ⎜⎜ ⎟⎟
⎝ DT ⎠ ⎝ 10 gDT ⎠
Figure 1: (a) Particles at rest in the bottom of the Schade (1987), apud Yi et al. (1998)
pipeline. Air velocity less than the pickup velocity.
(b) Particles are dragged by the air stream. Air 0.025 0.34
US ⎛D ⎞ ⎛ ρp ⎞
velocity equal to or greater than the pickup velocity. = ζ 0.11 ⎜ T ⎟ ⎜ ⎟ (4)
gDT ⎜ dp ⎟ ⎝ ρf ⎠
⎝ ⎠
The main correlations found in the literature that
are used for predicting the critical pickup and Weber (1981)
saltation velocities are presented below. These
correlations are analyzed in this work. For:
For: 2.33
⎛ ⎛ 1 − ε ⎞0.1 ⎞
⎜⎜ ⎟ + 30C0.35
V ⎟ Ar
⎜⎝ ε ⎠ ⎟
Gs ⎝ ⎠
< 47000 (12)
Dt 3/7
(7) ρf d p U mp ⎛⎛ 0.1 ⎞
2.33 ⎞
⎛1− ε ⎞
= 1.1 ⎜ ⎜ ⎜ ⎟ + 30C0.35
V ⎟ Ar ⎟
U min = 1.5G s0.465 DT −0.01μ0.55ρf−0.42 μ ⎜
⎜ ⎝ ε ⎠ ⎟ ⎟
⎝⎝ ⎠ ⎠
For:
where,
Gs
< 47000 ζρf
Dt CV = , ( solids volumetric concentration )
(8) ρp + ζρf
U min = 8.7G s0.302 DT 0.153μ0.55ρf−0.42
Main Correlations Used for Predicting the Critical
where, μ (in kg/m.s) is the dynamic viscosity of the Pickup Velocity
gas and Gs (in kg/m2.s) is the solids flux.
Cabrejos and Klinzing (1992)
Ochi (1991)
U p = ⎡1.27 Ar –1/3 + 0.036 Ar1/3 + 0.45⎤
0.82 ⎣ ⎦
Us ⎛ U ⎞ (13)
= 1.05fs0.47 ⎜ t ⎟ ζ 0.25 (9) ⎡0.70 Ar –1/5 + 1⎤ U p0
gd p ⎜ g dp ⎟ ⎣ ⎦
⎝ ⎠
where fs is the particle friction coefficient with the where Up0 is the pickup velocity of a particle alone
pipeline wall. and,
where, Us0 is the saltation velocity of a single particle. CD is the drag coefficient of the particle.
Brazilian Journal of Chemical Engineering Vol. 31, No. 01, pp. 35 - 46, January - March, 2014
38 L. M. Gomes and A. L. Amarante Mesquita
-2 1
⎡ ⎛ d ⎞1.5 ⎤ ⎛ U7 ⎞ 2
1.54 x10-4 ⎢1- ⎜ p ⎟ ⎥ C Dρf d 4p ⎜ 3 po ⎟
⎢ ⎝ DT ⎠ ⎥ ⎜ v DT ⎟
⎣ ⎦ ⎝ ⎠
⎡ 1⎤ (15)
⎢π ⎛ U 21 ⎞ 8⎥
( )
= f s ⎢ gd p3 ρ p - ρf + 1.302 x 10-6 d p - 6.35 x 10-3ρf d p3 ⎜ 5 po 3 ⎟ ⎥
⎜ v DT ⎟ ⎥
( fine particle )
⎢6 ⎝ ⎠
⎢⎣ ⎥⎦
0.29
0.25 0.75 n ⎛ dp ⎞
Up ⎛D ⎞ ⎛ ρp ⎞ π C D Re p 2 + 0.1fs ⎜ ⎟
= 0.0428 Re p0.175 ⎜ T ⎟ ⎜ ⎟ (16) 8 ⎝ DT ⎠
gd p ⎜ dp ⎟ ⎝ ρf ⎠
⎝ ⎠ (18)
2.02 1
Re − πfs Ar = 0
where Rep is the particle Reynolds number. 6
Valid for:
25 < Rep < 5000, 8 < (DT/dp) < 1340 and 700 < (ρp/ρg) Spherical Coarse Particle (Rolling)
< 4240.
0.29
Kalman et al. (2005) n ⎛ dp ⎞
π C D Re p 2 + 0.1fs ⎜ ⎟
8 ⎝ DT ⎠
First Zone: (19)
2.02 1 dp
3
Re − πfs Ar =0
6 2
Re*p = 5Ar 7 , for Ar > 16.5 (17)
Fine particle
Second Zone:
0.375
Re*p = 16.7 , for 0.45 < Ar < 16.5 ⎛ dp ⎞
0.0277n Re501.96 + 0.00635 ⎜ ⎟
⎝ DT ⎠
Third Zone: (20)
2.625 π 1 A h ρf d p
Re = Ar +
1 6 12 s2μ2
Re*p = 21.8 Ar 3 , for Ar < 0.45
gas at ambient conditions (ρ = 1.18 kg/m3). These 2.5 m/s. Other correlations (Rizk, 1976; Matsumoto
analyses were designed to assess the influence of et al., 1977; Weber, 1981; Ochi, 1991; Cabrejos and
average particle diameter, pipeline diameter and Klinzing, 1994) showed an increase in the saltation
solids mass flow rate at the saltation velocity, thus velocity with the increase in average particle diame-
providing evidence for the indication of the best ter. The results of Shade (1987) showed that the
correlations in specific design situations. saltation velocity decreases with the increase of the
average particle diameter, which is not in accordance
Influence of Particle Diameter with the real physical behavior. The comparison be-
tween the lowest and highest value of the saltation
Figure 2 presents a plot of the saltation velocity velocity results in a value of approximately 13 m/s.
as a function of the average diameter of fine particles
(0-200 μm). The mass flow rate of solids (Ws = 350
kg/h) was kept constant in all simulations. An inner
diameter of 50.4 mm for the horizontal pipeline was
used. Sand particles were used (ρp = 2636 kg/m3) in
all simulations. The result indicate that only the Rizk
(1976) correlation presented a variation with the
increase of the average particle diameter, which sug-
gests that all other correlations are not appropriate
for predicting pickup velocity throughout a wide
range of average particle diameters. It was also
observed that the correlation of Rizk (1976) follows
the physical behavior for the whole range of particle
Figure 3: Velocity of deposition as a function of
diameters, since for small particle sizes the effect of
average particle size. dp: 200 - 4000 micrometers.
cohesive forces is higher, requiring a higher velocity
for the particle entrainment (Cabrejos and Klinzing,
Influence of Pipeline Diameter
1994; Hayden et al. 2003; Rabinovich and Kalman,
2009).
Figure 4 shows a plot of saltation velocity as a
function of pipeline diameter. The simulations were
made with sand particles (ρp = 2636 kg/m3) with
particle diameter (dp = 200 μm) at a solid mass flow
rate of 350 kg/s. A great discrepancy in the results
was found. Some correlations provide a practically
constant saltation velocity estimation (Geldart and
Ling, 1992; Ochi, 1991), showing low values for the
range from 0 to 1 m/s. Other correlations (Rizk,
1976; Shade, 1987; Matsumoto et al., 1977; Cabrejos
and Klinzing, 1994; Kalman et al., 2005) show an
increase with the diameter of the pipeline.
Figure 2: Velocity of deposition as a function of
average particle size. dp: 0 - 200 micrometers.
Conversely, other correlation (Weber, 1981; the following section the testing bench is described,
Ochi, 1991; Geldart and Ling, 1992) show a decrease as well the characteristics of the materials used.
in the saltation velocity when the pipeline diameter
was increased. The difference between the highest The Experimental Setup
and lowest saltation velocity value for a pipe diame-
ter of 200 mm is approximately 19 m/s. An experimental apparatus, see Figure 6, was
developed to determine the pickup velocity, which
Influence of the Solids Mass Flow Rate consists basically a 1.5-meter-long, 50 mm diameter
horizontal steel pipeline; three horizontal PVC pipe-
Figure 5 shows a plot of saltation velocity as a lines (each 6 meters long and 50, 75 and 100 mm in
function of mass flow rate for solid particles of sand diameter), with a butterfly valve at the end of each
(ρp = 2636 kg/m3). The simulations were carried out pipeline; three transparent sections (placed in the
for particle and pipeline diameters of 200 μm and middle of the PVC pipelines, where the visual obser-
50.4 mm, respectively. The range for the solid mass vations were carried out); a root blower (controlled
flow employed was from 50 to 1000 kg/s. Some by a frequency inverter), which provides the gas
correlations (Geldart and Ling, 1992; Ochi, 1991; flow rate and pressure necessary for picking up
Kalman et al., 2008) remained constant in the range the particles; and a solids collector with a paper
of very low saltation velocity values (between 0 and filter bag placed on its top. A complete description
2 m/s) and other correlations show growth with an of this experimental setup can be found in Gomes
increase in solid mass flow (Weber, 1981; Cabrejos (2011). Experiments were carried out using air as the
and Klinzing, 1994; Matsumoto et al., 1977; Rizk, conveying gas at ambient conditions.
1976; Shade, 1987). The saltation velocity values are
very divergent. The difference between the highest
and lowest saltation velocity for the solid mass flow
rate of 1000 kg/s was about 14 m/s.
constant volumetric flow rate through the pipeline range of particle sizes prepared by sieving.
with one of the three butterfly valves opened. The Table 2 shows experimental data used in this
layer started to erode slowly as the gas stream picked paper for the pickup velocity study.
up the top particles.
250
Experiments
1
200 2
3
4
150 5
6
Δm (g)
7
100
50
0
2 3 4 5 6 7 8 Figure 8: Alumina particles, dp = 62 μm.
Average Velocity (m/s)
Brazilian Journal of Chemical Engineering Vol. 31, No. 01, pp. 35 - 46, January - March, 2014
42 L. M. Gomes and A. L. Amarante Mesquita
Table 2: Experimental data used in this paper for the pickup velocity
study and Geldart Classification.
Figure 10 presents the plot of the pickup velocity particle diameters above this limiting value, the
as a function of average particle diameter. A pipe results begin to diverge considerably. The values
diameter of 50.4 mm and a particle diameter ranging obtained with the correlations of Rabinovich and
from 200 to 3500 μm were employed in the simula- Kalman (2009) are usually very low and those ob-
tions. Experimental data are compared for sand parti- tained with the correlation of Cabrejos and Klinzing
cles (ρp = 2636 kg/m3), glass beads (ρp = 2834 kg/m3) (1992) for particle average diameter above 1500 μm
and irregular grains of salt (ρp = 2234 kg/m3) with a rise very quickly. It should be noted here that the
pickup velocity of sand particles (ρp = 2636 kg/m3) correlation of Cabrejos and Klinzing (1992) was
obtained from Cabrejos and Klinzing (1992), obtained from a theoretical model and some of the
Cabrejos and Klinzing (1994), Kalman et al. (2005) parameters were adjusted empirically. The correla-
and correlations from Kalman et al. (2009). tion of Cabrejos and Klinzing (1994) is fully
empirical. Note also that the correlations of Cabrejos
and Klinzing (1994) and Kalman et al. (2005) agree
reasonably well with each other. All correlations
agree reasonably well with the experimental data,
although the correlations by Cabrejos and Klinzing
(1994) and Kalman et al. (2005) show a better fit when
the whole range of particle diameter is considered.
Figure 11 presents the plot of the pickup velocity
as a function of average particle diameter for fine
particles. The simulations were made with data for
sand particles (ρp = 2636 kg/m3) and the results com-
Figure 10: Experimental pickup velocity as a function pared with experimental data for non-spherical glass
of the particle diameter and correlation curves. dp: (ρp = 2834 kg/m3), irregular salt (ρp = 2234 kg/m3), sand
200-3500 μm. (ρp = 2636 kg/m3) and glass beads (ρp = 2834 kg/m3).
The pipeline diameter used in the simulations was
From an analysis of the curves obtained with the 50.4 mm. The equation of Rabinovich and Kalman
use of the above mentioned correlations, it was (2009) presents no correlation with the physical phe-
verified that all of them show an increase in the nomenon, since for the smallest particle diameters
pickup velocity as the average particle diameter is the pickup velocity should increase with the reduc-
increased. However, for a given average particle tion of the average diameter of the particles, which
size, there is a large divergence between the values does not occur. The correlation that provides the best
of pickup velocities obtained and values predicted by agreement with experimental data (which includes an
the correlations. This shows that such correlations average particle diameter ranging from 0 to 200 μm)
present limitations in predicting reliable values of is the correlation of Kalman et al. (2005). It is note-
pickup velocities. It is noted that three correlations worthy to point out that, for the average diameter of
present a fair agreement among themselves for 22 μm, the error of this correlation is approximately
predicting the pickup velocity up to approximately 100%. The correlation of Cabrejos and Klinzing
the average particle diameter of 1500 μm. For (1992) overestimates the results, considerably.
Figure 12: Pickup velocity as a function of pipeline Figure 15 shows the pickup velocity as a function
diameter. of sphericity for pasta (ρp = 1200 kg/m3), spaghetti
Brazilian Journal of Chemical Engineering Vol. 31, No. 01, pp. 35 - 46, January - March, 2014
44 L. M. Gomes and A. L. Amarante Mesquita
(ρp = 1500 kg/m3), lentil (ρp = 1644 kg/m3), rice mm. The materials used were spherical and non-
(ρp = 1590 kg /m3) and bakelite (ρp = 930 kg/m3) spherical glass. The experimental data are obtained
particulate materials. The analysis was performed from Cabrejos and Klinzing (1994), Hayden et al.
only with the Kalman´s correlation (Kalman et al., (2003) and Kalman et al. (2005). The densities are
2005), since it presents a dependence of pickup 2480, 2500 and 2834 kg/m3, respectively. The
velocity on sphericity. To obtain the experimental sphericity of the non-spherical glass is 0.7 (Cabrejos
data, a pipe with a 100 mm diameter was used. The and Klinzing, 1994). Note that the non-spherical
correlation shows reasonable agreement with experi- glass has greater velocities than the glass spheres in
mental data. However, a careful observation of this both pipe diameters of 25 and 50 mm. This is due to
correlation demonstrates that the pickup velocity the fact that, in the case of non-spherical particles,
decreases with decreasing sphericity (keeping other the contact area is larger and therefore there is a
parameters constant), which is in contradiction to greater action of the cohesion forces (Cabrejos and
the physical phenomena involved, since the pickup Klinzing, 1994; Hayden et al., 2003; Kalman et al.,
velocity should increase with the reduction of the 2005). In Kalman’s correlation (Kalman et al., 2005)
sphericity (Cabrejos and Klinzing, 1992; Cabrejos the pickup velocities of spherical particles are always
and Klinzing, 1994; Hayden et al., 2003) and not larger than those for non-spherical particles, thus
the opposite, as is the case of this correlation (see showing that the reasonable agreement obtained in
Figure 16). the pickup velocity from experimental data was
Figure 16 shows the pickup velocity as a function related to its correlation with other parameters and
of particle diameter for pipe diameters of 25 and 52 not necessarily with the sphericity.
CONCLUSION NOMECLATURE
REFERENCES
ACKNOWLEDGEMENTS
Cabrejos, F. J., Klinzing, G. E., Incipient motion of
The authors would like to express their gratitude solid particles in horizontal pneumatic conveying.
for the financial support from CNPq and FAPESPA. Powder Technology, v. 72, pp. 51-61 (1992).
Brazilian Journal of Chemical Engineering Vol. 31, No. 01, pp. 35 - 46, January - March, 2014
46 L. M. Gomes and A. L. Amarante Mesquita
Cabrejos, F. J., Klinzing, G. E., Pickup and saltation phase flow in a horizontal pipe. Proc. The First
mechanisms of solids particles in horizontal ASME-JSME Fluids Engineering Conference,
pneumatic transport. Powder Technology, v. 79, FED-121, pp. 163-166 (1991).
pp. 173-186 (1994). Pirker, S., Kahrimanovic, D., Kloss, C., Popoff, B.,
Chu, K. W., Yu, A. B., Numerical simulation of Braum, M., Simulations coarse particle convey-
complex particle-fluid flows. Powder Technology, ing by a set of Eulerian, lagrangian and hybrid
v. 179, pp. 104-114 (2008). particle models. Powder Technology, v. 204, pp.
Gomes, L. M., Contribution to the Dilute-Phase 203-213 (2010).
Pneumatic Conveying Analysis. PhD Thesis, Rabinovich, E., Kalman, H., Generalized master
Federal University of Pará, Brazil (2011). (In curve for threshold superficial velocities in parti-
Portuguese). cle–fluid systems. Powder Tecnology, v. 183, pp.
Geldart, D., Liung, S. J., Saltation in high pressure 304-313(2008).
conveying of fine coal. Powder Technology. v. Rabinovich, E., Kalman, H., Incipient motion of
69, pp. 157-162 (1992). individual particles in horizontal particle-fluid
Hayden, K. S., Park, K., Curtis, J. S., Effect of systems: B. Theoretical analysis. Powder Tecnol-
particle characteristics on particle pickup velocity. ogy, v. 192, no 3, pp. 326-338 (2009).
Powder Technology, v. 131, pp. 7-14 (2003). Rizk, F., Pneumatic conveying at optimal operation
Huilin, L., Gidaspow, D., Bouillard, J., Wentie, L., conditions and a solution of Bath’s equation. Proc
Hydrodynamic modelling of gas-solid flow in a Pneumotransport, 3, Bath, England, pp. 443-58
riser using the kinetic theory of granular flow. (1976).
Chemical Engineering Journal, v. 95, pp. 1-13, Schade, B., Zum Übergang Sprung-Strähnen-
(2003). förderung bei der Horizontalen Pneumatischen
Kalman, H., Satran, A., Meir, D., Rabinovich, E., Feststoffördrung. Dissertation, University of
Pickup (critical) velocity of particles. Powder Karlsruche (1987) apud Yi, J., Wypych, P. W.,
Technology, v. 160, pp. 103-113 (2005). Pan, R., Minimum Conveying Velocity in Dilute-
Kalman, H., Rabinovich, E., Boundary saltation and Phase Pneumatic Conveying, Powder Handling &
minimum pressure velocities in particle–gas Processing, v. 10, n. 3, pp. 255-261 (1998).
systems. Powder Technology, v. 185, pp. 67-79 Yi, J., Wypych, P. W., Pan, R., Minimum conveying
(2008). velocity in dilute-phase pneumatic conveying.
Matsumoto, S., Hara, M., Saito, S. and Maeda, S., Powder Handling & Processing, v. 10, n. 3, pp.
Minimum transport velocity for horizontal pneu- 255-261 (1998).
matic conveying. Journal of Chemical Engineering Weber, M., Principles of hydraulic and pneumatic
of Japan, v. 7 n. 6, pp. 425-431 (1977). conveying in pipes. Bulk Solids Handling, v. 1,
Ochi, M., Saltation velocity of the gas-solid two- pp. 57-63 (1981).