Professional Documents
Culture Documents
www.elsevier.com/locate/powtec
Department of Mechanical Engineering, National Pingtung University of Science and Technology, Pingtung 91207, Taiwan, ROC
b
Department of Mechanical Engineering, Czech Technical University, 16607 Prague 6, Czech Republic
c
Department of Mechanical Engineering, National Taiwan University, Taipei 10617, Taiwan, ROC
d
Department of Mechanical Engineering, National Central University, Chung-Li 32054, Taiwan, ROC
Received 11 December 2001; received in revised form 25 November 2002; accepted 5 December 2002
Abstract
The flow patterns and stresses on the wall in a symmetric two-dimensional louvered-wall moving granular filter bed were investigated.
The static wall stress distributions produced by the granular solids were measured and compared with the theoretical prediction using the
differential slice and Runge Kutta (order four) methods. The variations in the dynamic wall stresses with time in a moving granular filter bed
were obtained. In addition, the effect of the louver angle upon the flow patterns and wall stresses was investigated. Four different flow
regions were observed in a moving granular filter bed. As the angle of the louver decreases, the quasi-stagnant zone area adjacent to the side
wall becomes smaller and the static normal stress acting on the convergent section of the side wall becomes larger. The magnitude of the
static normal stress acting on the convergent section is approximately 10 times as large as that acting on the vertical section. When the normal
stress measured by pressure gauge installed on the upper stage decreases to zero, the normal wall stress measured by pressure gauge installed
on the adjacent lower stage then begins to descend and fluctuate under the static normal wall stress during granular material withdrawal.
Employing the results obtained using stress measurements and image processing, the pressure pulsation phenomenon in a symmetric twodimensional louvered-wall moving granular filter bed may be further understood.
D 2002 Elsevier Science B.V. All rights reserved.
Keywords: Moving granular filter bed; Symmetrical louvered-wall; Pressure pulsation
1. Introduction
Hot gas particulate filtration is a key component of
current combined cycle power generation systems based
on the combustion and gasification of coals such as the
integrated gasification combined cycle (IGCC). In these
processes, the gases obtained from the coal must be
expanded through a gas turbine and the gas cleanup must
be carried out without cooling the gases to protect the
downstream heat exchanger and gas turbine components
from fouling and erosion. At the same time, flue gas
cleaning must meet particulate emission standards of 50
mg/mSTP3, which was set by legislation in the mid-1980s for
new and/or existing coal-fired boilers [1].
* Corresponding author. Tel.: +886-8-7703202x7016; fax: +886-87740142.
E-mail addresses: Cschou@mail.npust.edu.tw (C.S. Chou),
Jtkuo@w3.me.ntu.edu.tw (J.T. Kuo), Sshsiau@cc.ncu.edu.tw (S.S. Hsiau).
Particulate removal from a hot gas stream can be accomplished using cyclones, barrier filters, electrostatic precipitators, granular bed filters or scrubbers [2]. The choice of
which filter to use is a complicated optimization problem
involving emissions, reliability and costs. The most promising alternatives seem to be granular bed filters [3] and
barrier filters.
Ceramic barrier filters use the most advanced hot gas
filtration technology with several systems nearly ready for
commercial use in the 250 400 jC temperature range [4
7]. However, problems encountered during recent pilot and
demonstration-scale tests, particularly at high temperatures
(up to 900 jC), have led to concerns over the future
exploitation of this technology. For this reason, alternative
technologies such as granular bed filters continue to be
investigated and developed [8].
Beds of granular solids have been employed for dust
collection for many years. This subject has gained prominence recently as a possible means of simultaneously
0032-5910/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0032-5910(02)00344-3
167
168
Fig. 3. The layout of the pressure gauges and detailed dimensions of the granular filter bed for three tests.
169
The height of the granular bed was 1500 mm, the width
was 455 mm and the depth was 124 mm. The granular bed
width to depth ratio was kept above 3:1 to promote twodimensional behavior. The louvers angle could be adjusted
by placing a wedge shape steel plate between the louver and
the vertical section of the side wall (see Fig. 2). Each of the
side walls had six circular holes, reserved for installing
pressure gauges. The layout of the pressure gauges and
Fig. 4. The dimensions of the pressure gauges. (a) Cylindrical type, (b) hexahedron type.
170
Fig. 4 (continued ).
gauges were attached (see Fig. 5). This ring comprises the
basic dynamometric element of the pressure gauge. The
wiring and locations of the semiconductor strain gauges on
171
and
yit vit kti xit :
Fig. 6. The schematic drawing of the calibration system for pressure gauges.
(a) For normal stress, (b) for shear stress.
the steel ring were made so that the shear stress effect pt was
completely eliminated during the normal stress pn measurement. The shear stress pt measurement is not affected by the
simultaneously acting normal pressure pn.
In this research work, the standard weights were used to
calibrate the pressure gauges. The schematic drawings for
normal stress measurement calibration and shear stress
measurement calibration are shown in Fig. 6(a) (b), respectively. For each pressure gauge, the shear stress measurement calibration was carried out using the following
procedures: (1) a pressure gauge was installed on the bench
and the sensing plate was placed flush to the surface of the
bench; (2) a screw driven into the sensing plate was
wrapped with a cord; (3) the other end of the cord was
used to tie a rod welded to a base; (4) two electrical wires
for transmitting the output voltage of the pressure gauges
were connected to a home made connector; (5) the different
weights were placed on the base one by one and the output
voltage was recorded.
This pressure gauge exhibits linear calibration characteristics (i.e., the linear relationship between the stress and the
output voltage). For each pressure gauge, the calibration
straight lines for normal and tangential stresses were determined, using, respectively,
yin vin kni xin ;
Fig. 7. The calibration of the stress measurement. (a) For normal stress, (b)
for shear stress.
172
Table 1
Test conditions
Test 2
Test 3
50
200
363
455
0.0915
45
200
363
455
0.09
40
200
363
455
0.091
pin
Yni Vni
;
kni
Yti Vti
;
kti
and
pit
Pt
l
Pn
Table 2
Friction angles
Granular
material
Particle
particle
Particle
side steel
wall
Particle
transparent
acrylic wall
PE
26j
8j
12j
173
Fig. 8. The schematic drawing of the stress measurement and video imaging systems.
Fig. 8). The output voltage from the pressure gauges was
amplified using a power supply (TopWard-6303D). A data
acquisition card (ADVANTECH PCL-818HG) was
employed to convert the analog signals into digital signals.
Computer software (VisiDAQ 3.1) was employed to process the data. At the same time, a digital camcorder (SONY
DCR-TRV310) was used to record the development of the
black colored granule flow until no more granules were
174
Pv dPv w dwb
and
K1
p uw
K1 tan
4
2
8
tana
tanuw tana
P v qb g
K2 K3 h
dh
10
175
Fig. 11. The flow history of the black colored granules in the two-dimensional moving granular filter bed under Test 1 (louver angle: 50j). Frames 1 18, time
interval 25 s; frame 18, time 425 s.
176
Fig. 12. The flow history of the black colored granules in the two-dimensional moving granular filter bed under Test 2 (louver angle: 45j). Frames 1 18, time
interval 25 s; frame 18, time 425 s.
177
Fig. 13. The flow history of the black colored granules in the two-dimensional moving granular filter bed under Test 3 (louver angle: 40j). Frames 1 18, time
interval 25 s; frame 18, time 425 s.
178
Fig. 14. The schematic drawing of the four flow regions [20].
Fig. 10(a) (b) shows the static normal wall stress distribution in the granular filter bed under Tests 1 (louver
angle: 50j) and Test 3 (louver angle: 40j), respectively. In
each frame, the solid line and dashed line represent the
theoretical static normal stress distributions with surcharge
and under zero surcharge, respectively. In addition, cross,
diamond, square, circle, asterisk and saltire represent the
static normal stress measured by pressure gauges L3, LS3,
L2, LS2, L1 and LS1, respectively.
In general, at the vertical section of the louvered-wall,
the static normal stress measured by the pressure gauge
agreed well with theoretical prediction obtained using the
differential slice method under zero surcharge (see Fig.
10). For example, at the vertical section of the third stage,
the theoretical and experimental static normal stresses
under Test 3 (louver angle: 40j) were 0.68 and 0.6 kPa,
respectively.
In contrast, at the convergent section of the louveredwall, the static normal stress measured by the pressure
Fig. 15. At 125 s, a comparison of the flow status between Tests 1, 2 and 3. (a) For Test 1 (louver angle: 50j), (b) for Test 2 (louver angle: 45j), (c) for Test 3
(louver angle: 40j).
179
Fig. 16. The variations in dynamic normal and shear stresses with time for Test 1 (louver angle: 50j). The top panel is for pressure gauge RS3; the center panel
is for pressure gauge RS2; the bottom panel is for pressure gauge RS1.
180
Fig. 17. The variations in dynamic normal and shear stresses with time for Test 3 (louver angle: 40j). The top panel is for pressure gauge RS3; the center panel
is for pressure gauge RS2; the bottom panel is for pressure gauge RS1.
181
Fig. 18. The variations in dynamic normal and shear stresses with time for Test 1 (louver angle: 50j). The top panel is for pressure gauge L3; the center panel is
for pressure gauge L2; the bottom panel is for pressure gauge L1.
182
Fig. 19. The variations in dynamic normal and shear stresses with time for Test 3 (louver angle: 40j). The top panel is for pressure gauge L3; the center panel is
for pressure gauge L2; the bottom panel is for pressure gauge L1.
Table 3
Maximum dynamic, mean dynamic and static stresses for Test 1 (louver
angle: 50j)
Pressure
gauge
Normal
Shear
Max.
dynamic
stress
Mean
dynamic
stress
Static
stress
Max.
dynamic
stress
Mean
dynamic
stress
Static
stress
L1
L2
L3
RS1
RS2
RS3
0.62
0.63
0.68
6.1
7.3
7.5
0.45
0.48
0.51
3.5
4.1
3.7
0.42
0.56
0.5
3.54
3.83
3.29
0.19
0.18
0.01
3.1
3.8
3.3
0.13
0.08
0.01
0.8
0.8
0.6
0.15
0.15
0.01
1.11
1.01
0.26
Normal
Max.
dynamic
stress
Mean
dynamic
stress
Static
stress
Shear
Max.
dynamic
stress
Mean
dynamic
stress
Static
stress
L1
L2
L3
RS1
RS2
RS3
0.79
0.83
0.62
9.8
9.8
8.9
0.48
0.51
0.48
5.1
4.3
3.8
0.6
0.59
0.5
4.83
5.85
4.28
0.26
0.21
0.01
2.8
3.1
2.8
0.17
0.1
0.01
1.2
0.8
0.85
0.19
0.13
0.01
0.72
1.25
0.23
183
Acknowledgements
5. Conclusions
The flow patterns and stresses on the wall in a twodimensional moving granular filter bed were investigated.
Filter granules were moved between the two vertical
louvered walls of the filter with no interstitial fluid flow
relative to the solids. The striking feature of the pressure
184
References
[1] J.H. Ludwig, P.W. Spaite, Chemical Engineering Progress 63 (6)
(1967) 82.
[2] J.P. Mustonen, S.J. Bossart, M.W. Durner, Proc. of Int. Conf. on
Fluidized Bed Combustion, Montreal, Canada, vol. 1, 1991, p. 475.
[3] S. Hozumi, J. Fukui, K. Kodama, K. Kamei, T. Mori, N. Izumi, S.
Kusada, Proc. Symp. Gas Cleaning at High Temperature, Pergamon,
Oxford, 1986, p. 385.
[4] R. Chang, J. Sawyer, H. Lips, R. Bedick, R. Dellefield, Proc. Symp.
Gas Cleaning at High Temperature, Pergamon, Oxford, 1986, p. 177.
[5] K. Higashi, Journal of the Society of Powder Technology, Japan 36
(1999) 378.
[6] S. Ito, T. Tanaka, S. Kawamura, Advanced Powder Technology 8 (1)
(1997) 53.
[7] T. Nakayama, H. Matsuda, S. Ito, H. Shirai, M. Kobayashi, T. Tanaka,
Karyoku Genshiryoku Hatsuden 43 (1992) 210.
[8] J.R. Longanbach, . Proc. of 23rd Int. Technical Conf. on Coal Utilization and Fuel System, Clearwater, FL, USA, 1998, p. 69.
[9] J.T. Kuo, J. Smid, S.S. Hsiau, C.S. Chou, Proceedings of the National
Science Council, Republic of China (A) 22 (1) (1998) 17.
[10] J.P.K. Seville, R. Clift, in: J.P.K. Seville (Ed.), Gas Cleaning Demanding Applications, Blackie Academic and Professional, London, 1997,
pp. 170 171, Chapter 9.
[11] A.M. Presser, J.C. Alexander, Proc. of 3rd Symposium on Transfer
and Utilization of Particulate Control Technology, 1982, p. 26.
[12] Avco Co., Final Report, Contract No. PH-86-67-51, Phase Ill, AYATD0107-67-RR, U.S. Department of Commerce, Publication PB 185561,
pp. 8, 48, 67, 1969.
[13] R.G. Reese, Tappi (Journal of the Technical Association of the Pulp
and Paper Industry) 60 (3) (1977) 109.