You are on page 1of 6

Fuel Processing Technology 106 (2013) 338343

Contents lists available at SciVerse ScienceDirect

Fuel Processing Technology


journal homepage: www.elsevier.com/locate/fuproc

Fine coal dry cleaning using a vibrated gas-uidized bed


Xuliang Yang , Yuemin Zhao, Zhenfu Luo, Shulei Song, Chenlong Duan, Liang Dong
School of Chemical Engineering & Technology, China University of Mining and Technology, 221008 Xuzhou, China

a r t i c l e

i n f o

Article history:
Received 25 February 2012
Received in revised form 8 June 2012
Accepted 22 August 2012
Available online 19 September 2012
Keywords:
Fine coal
Dry cleaning
Vibrated gas-uidized bed
Segregation
Bubble-driven jigging

a b s t r a c t
Fine coal (6 mm) cleaning in a dry way becomes more important with the extensive deployment of the mechanized mining and progressively serious water shortage, especially in North-West China. In this paper, we
attempted to use the segregation in a vibrated gas-uidized bed of dissimilar particles to provide a solution to
this problem. The effects of several factors including the supercial air velocity, bed height, vibration intensity
and uidizing time on the segregation performance were experimentally studied. The bubble-driven jigging mechanism was proposed to explain the separation process. The results show that the probable error E values of the separation of 6+3 mm and 3+1 mm size fraction of feed coal samples are 0.19 and 0.175 respectively, which
indicates that ne coal separation using a vibrated gas-uidized bed can provide a simple and efcient way for
coal cleaning in dry and cold regions in North-West China.
2012 Elsevier B.V. All rights reserved.

1. Introduction
Dry cleaning of ne coal (6 mm) is an important issue in coal sector, especially for China. Large amounts of ne coal are produced during
coal mining as a result of the extensive deployment of the mechanized
mining technology and should be cleaned with the consideration of energy source conversation and environment protection. In addition,
China's coal reserves are mainly deposited in North-West China where
the arid geological environment and prolonged cold weather per year
present obstacles to the deployment of the coal wet cleaning technologies. Thus, it is urgent to develop a novel and efcient dry cleaning technology for ne coal. The cleaning technologies including air dense
medium uidized bed separator [1,2], air jigging [3] and FGX separator
[4] provide efcient solutions to the dry cleaning of 50+6 mm coal.
Fan et al. [5,6] studied magnetically stabilized uidized beds for separating ne coal (6+ 1 mm). Luo et al. [7] introduced the vibration energy
to an air dense medium uidized bed separator in order to provide a solution to ne coal (6+1 mm) cleaning. Macpherson et al. [8,9] studied
the density-based separations of ne coal (8+1 mm) in the Reux
Classier with an airsand dense-medium and vibration. Although
these three ne coal dry cleaning technologies give good separation
results, they all encounter obstacles in the way of industrial applications
due to the problems of dense medium recovery, product purication,
and low processing capacity. Overall for ne coal (6 mm), there is no
effective dry cleaning technology that can work with great potential for
commercialization.
Granular materials in a uidized bed can segregate due to different material properties such as different densities and/or sizes [10].
Corresponding author. Tel.: +86 15162110730.
E-mail address: yangxuliang@126.com (X. Yang).
0378-3820/$ see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.fuproc.2012.08.019

However, stable uidization of coarse particles (+ 1 mm) that belongs to the type D material in the classication by Geldart [11] is
very difcult by ambient air solely. The introduction of vibration energy to traditional gassolid uidized beds can uidize the coarse particles effectively by enhancing the hydrodynamic interaction between
air and particles and by eliminating the channeling of air ow within
the bed. The segregation processes that occur simultaneously in vibrated gas-uidized beds are solely and entirely by the bubbles [12].
Appropriate bubbling stability is responsible for a stable and effective
segregation behavior due to regular hydrodynamic interaction of the
two-phase ow [13]. In this paper, we focus on utilizing a vibrated
gas-uidized bed to clean 6 + 1 mm ne coal and studying the effects of different operational factors on the separation performance.
2. Mechanism
In a vibrated gas-uidized bed of coarse particles with different densities, the segregation is mainly caused by the bubbles. When a bubble
rises through the granular bed, a temporarily disturbed region having
considerable lower solid volume fraction than the surrounding bulk
phase is formed behind the bubble. In this region, particles with higher
density tend to sink preferentially over the lighter particles, which
leads to local particle segregation. The hindered settling velocity plays a
key role in this segregating process. Briey, high-density particles have
an opportunity to overtake low-density particles by falling rapidly
through the bubbles and also settling faster in the temporarily disturbed
regions. The bubbles impose a periodic expansion and contraction on the
particle bed, thereby causing separation mainly based on density rather
than size. This process is analogous to the separation technique of jigging
[13] and consequently, the aforementioned separation mechanism can
be summarized as the bubble-driven jigging mechanism. Like traditional

X. Yang et al. / Fuel Processing Technology 106 (2013) 338343

jigging techniques, good separation in a vibrated gas-uidized bed also


requires enough bubble-driven jigging cycles. In addition, the achievement of optimal segregation is highly sensitive to several operational factors like the supercial air velocity, vibration intensity, bed height and
uidizing time for their signicant effects on the bubble behavior and
the uidization quality.
3. Experimental
3.1. Coal particle properties
Fine coal is a multi-component granular system and has a considerable wide density distribution generally ranging from 1.2 to 2.4 g/cm3.
Thus, narrow size range of the feed is favorable to the particle segregation
basing on density difference, which will minimize the inuence of particle
size difference on the hindered terminal settling velocity and reduce the
number of dissimilar particles having equal settling velocity. For ne
coal separation using a vibrated gas-uidized bed, the ratio of the upper
limit size to the lower one is approximately no more than 3:1.
In this paper, two size fractions i.e., 6+3 mm and 3+1 mm coal,
were studied and their density distribution is given in Tables 1 and 2
respectively. They both belong to high-ash coal which accounts for a considerable large share of ne coal in North-West China. Both size fractions
have a large amount of high density materials of +1.8g/cm3 of 38.47%
and 37.68% respectively.
3.2. Experimental apparatus
The schematic diagram of the experimental apparatus is shown in
Fig. 1. The experimental apparatus of a vibrated gas-uidized bed consists of three main parts: gas supply system, vibration generating system and uidized bed. The ne coal particles are uidized in a vertical
cylinder with an inner diameter of 110 mm and a height of 400 mm
which is made from transparent plexiglass. Ambient air after ltering
is dispersed by a sintered metal distributor and then enters the vessel
and uidizes the granular particles. Air pressure is controlled over a
range from 0.1 to 0.25 MPa by a valve leading to the atmosphere that
regulates the pressure of the tank. The vibrated bed deployed in this
study is manufactured by China STI Co., LTD. and its operational parameters can be easily adjusted by a digital controller to generate vibration
motions with amplitude ranging from 0 mm to 10 mm and frequency
ranging from 1 Hz to 400 Hz. OLYMPUS i-SPEED 3 high speed camera
system is used to study the particle motions within the bed in order
to reveal the separation process and verify the bubble-driven jigging
mechanism.
3.3. Segregation evaluation
In this paper, the segregation results are evaluated qualitatively and
quantitatively by the segregation pattern and the segregation degree
respectively. The particle bed is rstly uidized for a certain time and
then the uidizing air is suddenly shut down. The static particle bed is

Table 1
Results of the sink-oat experiment of 6 + 3 mm size fraction of coal.

Table 2
Results of the sink-oat experiment of 3 + 1 mm size fraction of coal.
Density fraction Average density (g/cm3) Weight fraction (%) Ash content (%)
(g/cm3)
1.3
+1.31.4
+1.41.5
+1.51.6
+1.61.7
+1.71.8
+1.8
Total

1.25
1.35
1.45
1.55
1.65
1.75
2.2

1.25
1.35
1.45
1.55
1.65
1.75
2.2

4.12
24.64
16.94
6.06
2.92
6.85
38.47
100.00

6.02
10.31
18.34
25.83
35.00
44.52
75.97
40.76

6.02
23.52
16.75
9.65
2.96
3.42
37.68
100.00

5.81
11.38
17.68
24.24
36.81
44.24
76.52
39.76

divided into ve layers evenly in the axial direction and we take samples from each layer to test the ash content. The segregation pattern is
obtained by plotting these ash content data with the dimensionless
bed height H = Hi/H0 as Y-axis where Hi is the average bed height of
the ith sampling layer, and H0 is the total bed height of the static bed.
A statistical indicator, S, is proposed to evaluate the segregation degree
quantitatively and the denition is shown in Eq. (1):

v
uX
n
u
u
Ai =A0 12
t
i1

n1

where Ai is the ash content of coal of the ith sampling point; A0 is the initial ash content of feed coal; and n is the total sampling number. It can
be clearly inferred from Eq. (1) that when a granular system is perfectly
mixed, Ai = A0, S = 0, and a larger value of S indicates better segregation
that is favorable for ne coal separation.
4. Results and discussion
4.1. Effect of supercial air velocity on segregation
In a vibrated gas-uidized bed, the steady state of the granular system is a result of the dynamic equilibrium between the competing processes of mixing and segregation [14]. The supercial air velocity plays
a critical role in achieving the optimal segregation results. Besides, the
dispersion uniformity of the uidizing air also has a signicant inuence
on the bubble characteristics. The optimal air velocity for segregation
highly depends on several factors including particle properties, vibration
energy, and bed height. Thus, a dimensionless supercial air velocity,
U =(UUmb)/Umb, is introduced to study the effect of supercial air velocity on segregation, where U and Umb is the uidizing air velocity and
the minimum bubbling air velocity, respectively.
Fig. 2 depicts the segregation results of the two types of feed coal
at different supercial air velocities. It can be seen that the two curves
both exhibit inverted V-shape and have single peak value that corresponds to the best segregation degree. For 6 + 3 mm size feed coal,
the U value at this peak is equal to 0.2 while it is 0.15 for 3 + 1 mm

5
1

10

11

6
9

Density fraction Average density (g/cm3) Weight fraction (%) Ash content (%)
(g/cm3)
1.3
+1.31.4
+1.41.5
+1.51.6
+1.61.7
+1.71.8
+1.8
Total

339

12

Fig. 1. Schematic diagram of the experimental apparatus 1. Air lter; 2. Roots blower;
3. Tank; 4. Pressure gauge; 5. Valve; 6. Rotameter; 7. Vibrated bed; 8. Air chamber; 9.
Air distributor; 10. Vessel; 11. High-speed camera; 12. Image analysis system.

340

X. Yang et al. / Fuel Processing Technology 106 (2013) 338343

-6+3 mm
-3+1 mm

0.7

0.69

-3+1 mm
-6+3 mm

0.66
0.6
0.63

0.5
0.60

0.4
0.57

0.3

0.54

0.2
0.05

0.10

0.15

0.20

0.25

U*
Fig. 2. Segregation degree at different supercial gas velocities.

size feed coal. When the supercial air velocity is larger than Umb, the
excess air produces bubbles that divide the bed into particulate phase
and bubble phase. Cibilaro and Rowe [12] pointed out that the mixing
and segregation processes in a gassolid uidized bed mainly depend
on the bubbles. Thus, the bubble characteristics including bubble size
and bubble rise velocity signicantly determine the segregation
performance.
Appropriate supercial air velocity can generate the bubbles that
lead to the optimal segregation results. When the bubble size is too
small, there is no enough space for particle settling in the disturbed
region below the rising bubble. When the bubble size is too big, the
bubble rise velocity becomes too fast to provide enough time for particle settling in the disturbed region.
4.2. Effect of vibration intensity on segregation
The vertical vibration may conveniently be characterized by vibration
intensity =A2/g, where A is the amplitude of the oscillation, =2f is
the angular frequency, f is the frequency of the oscillation and is the
ratio of the maximum acceleration of the bed to the acceleration due to
gravity, g. The value range of A and the frequency f are 0.53 mm and
1050 Hz, respectively. The 6+ 3 mm and 3+1 mm size friction
coal is separated at U =0.15,0.20 respectively under different vibration
intensity. The segregation degree of the two types of feed coal is shown
in Fig. 3. It can be seen that the two curves both exist with single peak
value which leads to the optimal segregation results. In a vibrated
gas-uidized bed of coarse particles, the input vibration energy can

0.51

50

60

70

80

90

100

110

H (mm)
Fig. 4. Segregation degree at different bed height.

generate appropriate uidizing conditions for particle segregation by


eliminating the channeling of air ow within the bed and optimizing
the bubble characteristics due primarily to enhanced particle interaction
and particle piercing behavior. However, when the input vibration energy excesses the critical value, the particle motion in the lower section of
the bed becomes severe enough to deteriorate the hindered settling process, which will lead to comparatively good mixing of the particle bed.
4.3. Effect of bed height on segregation
Fig. 4 depicts the segregation results of the two types of feed coal
with different initial bed height (H). It can be seen that when bed
height is smaller than 70 mm, the patterns of the two curves are
at, which indicates that the segregation degrees of the two types
of feed coal are both uniform. But when the bed height is larger
than 70 mm, the segregation performance of the two types of feed
coal both deteriorates dramatically. This change is mainly derived
from the weaknesses of the vibration function to the uidizing condition of the upper section of the particle bed when the bed height exceeds this critical value. In addition, bubble size at the upper section
of the bed with large bed height will become large enough to cause
large-scale particle circulations that deteriorate the segregation
process.
4.4. Effect of uidizing time on segregation
According to the aforementioned separation mechanism, the optimal segregation requires enough bubble-driven jigging cycles in order
to provide enough time for particle segregation primarily depending
on the density differences. Fig. 5 demonstrates the segregation degree
of the two types of feed coal at different uidizing time. We can see
clearly that the segregation degree of the two types of feed coal both increases with the uidizing time until it reaches to a critical value at
2 min and then the curves become at when the uidizing time exceeds 2 min. These curves indicate that the two granular systems both
achieve a stable state of segregation after 2 min uidization and then
there is no improvement when enlarging the uidizing time.
4.5. Verication of bubble-driven jigging mechanism

Fig. 3. Segregation degree at different vibration intensity.

The aforementioned bubble-driven jigging mechanism can be veried


by the results of image analysis of the pictures that are photographed by a
high-speed camera, shown in Fig. 6. Three particles labeled by 1, 2, and 3
with particle size of 3+1 mm represent clean coal, gangue and middlings respectively. These three particles are identied visually by the difference in physical properties like shape, color and luster using the image

X. Yang et al. / Fuel Processing Technology 106 (2013) 338343

0.75

Table 3
Physical properties of the clean coal, middlings and gangue particles.

0.70
0.65
0.60

341

Item

Color

Luster

Shape

Clean coal
Middlings
Gangue

Black
Dark gray
Gray

Submetallic/vitreous luster
Bituminous luster
Dull luster

Round, short-rod, ellipse


Long-rod, cuboid
Sheet, strip

0.55
0.50
0.45

-6+3 mm
-3+1 mm

0.40
0.35

T (min)
Fig. 5. Segregation degree at different uidizing time.

magnication method and the identication is concretely given in


Table 3. At 0 ms, particles in the local region investigated are compact
and there are no visible particle motions. When a bubble passes through
this region since 16.5 ms, particles in this region become loose and different in motion. Particle 1 with lower density moves upwards along with
the bubble due to gas drag force while particle 2 and particle 3 settle
through the bubble quickly. Since 61.5 ms, in the disturbed region
below the bubble, particle 2 settles faster than particle 3. And nally,
after a bubble-driven jigging cycle, the three investigated particles at
106.5 ms are arranged in the axial direction with particle 1 (clean coal)
in the top, particle 3 (middlings) in the middle and particle 2 (gangue)
in the bottom, which is extremely consistent with the above description
of the bubble-driven jigging mechanism.

4.6. Separation performance


Fig. 7 depicts the segregation patterns corresponding to the optimal segregation performance of the two types of feed coal and the

corresponding values of the aforementioned factors are listed in


Table 4. Under this optimal conditions, each feed coal is rstly separated at a higher separation density to discharge the gangue and then
the gangue-free coal is separated at a lower separation density to
produce the clean coal and the middlings. The partition distribution of
three products of each feed coal is examined by carrying out the
sink-oat experiments, as given in Tables 5 and 6 respectively. Based on
these data, the partition curves of the two types of feed coal are plotted
in Fig. 8. The separation results of the two types of feed coal are given in
Table 7. The probable error E values of 6+3 mm and 3+1 mm
size fraction of feed coal are 0.190.225 and 0.1750.195 respectively.
With the comparison of the E value, the 3+1 mm feed coal has a better
separation efciency both at a high separation density and a low one than
the 6+3 mm feed coal. This is because that 3+1 mm feed coal is
easier to achieve uniform uidization with good bubbling behavior than
6+3 mm feed coal. For each feed coal, the E value at a high separation
density is smaller than that at a low one.
5. Conclusion
In this study, we focus on utilizing the segregation in a vibrated
gas-uidized bed to provide a solution to the difcult problem of ne
coal separation in a dry process, which is greatly meaningful with respect to the utilization of ne coal for energy resources in North-West
China. In this process, bubble-driven jigging cycles provide enough
space and time for the particles to segregate in the axial direction
depending on their hindered settling velocity differences. The experimental results show that the operational parameters including supercial air velocity, vibration intensity, bed height and uidizing time have
signicant inuences on the segregation performance and the probably
error E values of 6 + 3 mm and 3 + 1 mm size fraction of feed coal
are 0.190.225 and 0.1750.195 respectively. According to the analysis

Fig. 6. Particle segregation in a bubble-driven jigging cycle.

342

X. Yang et al. / Fuel Processing Technology 106 (2013) 338343

Table 6
The sink-oat results of the products of 6 + 3 mm feed coal.
Density Product weight fraction (%) Calculated feedstock (%)
fraction
(g/cm3)
Gangue Middlings Clean Total
Reconcentration
coal
1.3
1.31.4
1.41.5
1.51.6
1.61.7
1.71.8
+1.8
Total

0.15
1.42
1.10
0.72
0.57
2.50
33.38
39.84

0.54
4.97
5.37
3.17
1.74
3.90
3.85
23.54

3.21
3.90
3.75
18.25
24.64 23.22
8.77
15.24 14.14
3.26
7.16
6.44
1.21
3.52
2.95
1.44
7.84
5.34
0.48
37.70
4.33
36.62 100.00 60.16

Partition
coefcient (%)
High
Low
density density
3.91
5.76
7.21
10.17
16.28
31.85
88.52

14.30
21.41
37.97
49.29
59.15
73.04
88.93

Acknowledgments
The research work involved in this paper received the nancial
support by the National Natural Science Foundation (No. 51134022,
51174203), the Key Project of Chinese National Programs for Fundamental Research and Development (973 program) (No. 2012CB214904), the
National Natural Science Foundation of China for Innovative Research
Group (No. 50921002), the Natural Science Foundation of Jiangsu Province of China (No. BK2010002), the Fundamental Research Funds for
the Central Universities (No. 2010QNB11, 2010ZDP01A06).

100

High-density separation

Fig. 7. The optimal segregation patterns of the two types of feed coal.

Table 4
Operational conditions leading to the optimal segregation performance of the two
types of the feed coal.

H (mm)
T (minutes)

6 + 3 mm

3 + 1 mm

0.2
0.23
70
2

0.15
0.2
70
2

0.20
0.87
0.98
0.86
0.42
2.12
34.77
40.21

0.88
4.55
5.68
4.58
1.53
1.21
3.86
22.29

4.55
5.63
5.44
18.47
23.89 23.02
9.77
16.43 15.45
3.15
8.58
7.72
0.69
2.63
2.22
0.41
3.74
1.62
0.46
39.10
4.33
37.50 100.00 59.79

Partition
coefcient (%)
High
Low
density density
3.46
3.63
5.97
9.97
15.83
56.72
88.93

20

-3+1 mm
1.4

1.6

1.8

2.0

2.2

Density (g/cm3)

Table 5
The sink-oat results of the products of 3 + 1 mm feed coal.

1.3
1.31.4
1.41.5
1.51.6
1.61.7
1.71.8
+1.8
Total

40

1.2

of three product separation results, the ne coal separation in a vibrated


gas-uidized bed system is useful to and more effective for ne coal
cleaning in drought and cold regions.

Density Product weight fraction (%) Calculated feedstock (%)


fraction
(g/cm3)
Gangue Middlings Clean Total
Reconcentration
coal

60

16.26
19.78
36.75
59.27
68.95
74.92
89.26

100
High-density separation
Low-density separation

Partition coefficient (%)

Factors

Partition coefficient (%)

Low-density separation

80

80
60
40
20

-6+3 mm

0
1.2

1.4

1.6

1.8

Density

(g/cm3)

2.0

Fig. 8. Partition curves of the two types of feed coal.

2.2

X. Yang et al. / Fuel Processing Technology 106 (2013) 338343


Table 7
Separation results of ne coal in a vibrated gas-uidized bed.
Separation results

6 + 3 mm

3 + 1 mm

Feed ash content (%)


Low separation density (g/cm3)
High separation density (g/cm3)
Low density separation E
High density separation E
Clean coal ash content (%)
Clean coal yield (%)
Middlings ash content (%)
Middlings yield (%)
Gangue ash content (%)
Gangue yield (%)

40.76
1.55
1.89
0.225
0.19
15.56
36.62
32.47
23.54
70.07
39.84

39.76
1.52
1.77
0.195
0.175
14.42
37.50
30.16
22.29
71.02
40.21

References
[1] Y. Zhao, L. Tang, Z. Luo, C. Liang, H. Xing, W. Wu, C. Duan, Experimental and numerical
simulation studies of the uidization characteristics of a separating gassolid uidized
bed, Fuel Processing Technology 91 (2) (2010) 18191825.
[2] A.K. Sahu, A. Tripathy, S.K. Biswal, A. Parida, Stability study of an air dense medium
uidized bed separator for beneciation of high-ash Indian coal, International Journal
of Coal Preparation and Utilization 31 (34) (2011) 127148.
[3] C.H. Sampaio, W. Aliaga, E.T. Pacheco, E. Petter, H. Wotruba, Coal beneciation of
Candiota mine by dry jigging, Fuel Processing Technology 89 (2) (2008) 198202.

343

[4] B. Zhang, H. Akbari, F. Yang, M.K. Mohanty, J. Hirschi, Performance optimization of


the FGX dry separator for cleaning high-sulfur coal, International Journal of Coal
Preparation and Utilization 31 (34) (2011) 161186.
[5] M. Fan, Q. Chen, Y. Zhao, Z. Luo, Fine coal (61 mm) separation in magnetically
stabilized uidized beds, International Journal of Mineral Processing 63 (4)
(2001) 225232.
[6] M. Fan, Q. Chen, Y. Zhao, Z. Luo, Y. Guan, Fundamentals of a magnetically stabilized
uidized bed for coal separation, International Journal of Coal Preparation and
Utilization 23 (1) (2003) 4755.
[7] Z. Luo, M. Fan, Y. Zhao, X. Tao, Q. Chen, Z. Chen, Density-dependent separation of
dry ne coal in a vibrated uidized bed, Powder Technology 187 (2) (2008)
119123.
[8] S.A. Macpherson, S.M. Iveson, K.P. Galvin, Density based separations in the Reux
Classier with an airCsand denseCmedium and vibration, Minerals Engineering
23 (2) (2010) 7482.
[9] S.A. Macpherson, S.M. Iveson, K.P. Galvin, Density-based separation in a vibrated
Reux Classier with an airsand densemedium: tracer studies with simultaneous
underow and overow removal, Minerals Engineering 24 (10) (2011) 10461052.
[10] P.N. Rowe, A.W. Nienow, Particle mixing and segregation in gas uidised beds, A
review, Powder Technology 15 (2) (1976) 141147.
[11] D. Geldart, Types of gas uidization, Powder Technology 7 (5) (1973) 285292.
[12] L.G. Gibilaro, P.N. Rowe, A model for a segregating gas uidised bed, Chemical
Engineering Science 29 (6) (1974) 14031412.
[13] J.F. Davidson, R. Clift, D. Harrison, Fluidization, 2nd edn Academic Press, Landon,
1985.
[14] N.S. Naimer, T. Chiba, A.W. Nienow, Parameter estimation for a solids mixing/
segregation model for gas uidised beds, Chemical Engineering Science 37 (7)
(1982) 10471057.

You might also like