Reduction of electrostatic charges in gas–solid fluidized beds

Chemical Engineering Science 57 (2002) 153–162

www.elsevier.com/locate/ces

Reduction of electrostatic charges in gas–solid #uidized beds Ah-Hyung Park, Hsiaotao Bi ∗ , John R. Grace Department of Chemical and Biological Engineering, The University of British Columbia, 2216 Main Mall, Vancouver, BC, Canada V6T 1Z4 Received 20 February 2001; received in revised form 23 August 2001; accepted 29 August 2001

Abstract Reduction of electrostatic charge accumulation by increasing the humidity of #uidizing gas was investigated using single bubble injection in two- and three-dimensional #uidized beds. Both 321
m glass beads and 378
m polyethylene particles were found to be charged positively when #uidized by air. Electrostatic charges increased as the bubble size increased. Increasing the relative humidity of the #uidizing air to 40 –80% reduced the accumulation of electrostatic charge by increasing the surface conductivity, thereby enhancing charge dissipation. ? 2002 Published by Elsevier Science Ltd. Keywords: Fluidization; Multiphase #ow; Charge; Humidity; Anti-static; Electrostatic

1. Introduction Electrostatics can represent a serious problem for #uidized beds, causing agglomeration, nuisance discharge and even the danger of explosions. In the past, there have been some e
that the e
0009-2509/02/$ - see front matter ? 2002 Published by Elsevier Science Ltd. PII: S 0 0 0 9 - 2 5 0 9 ( 0 1 ) 0 0 3 5 2 - 9

154

A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162 cyclone Air out Air in 3-D fluidization column optical fiber probe electrostatic ball probe pressure regulator

pressure transducers

PC-4 Powder Voidmeter pressure gauge

pressure transducer (alarm purpose) bubble injector rotameter

water in

hygrometer/ thermometer Electrometer

Air heater

Computer

rotameter humidifier water out (drain) dryer rotameter

Fig. 1. Schematic showing overall layout of equipment for three-dimensional column experiments.

introducing moisture to the #uidized bed to reduce electrostatic charges has been found to be ine
study to investigate the e
A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162

155

Fig. 2. SEM micrographs showing surfaces of particles used in this study.

◦

used for the injector with the tip bent 90 to face upward along the axis of the column. Probes used by previous researchers to measure electrostatic charges in #uidized beds can be divided into three major types — capacitance, induction and collision probes. A collision probe, also known as a contacting probe, was employed in the present work. The probe, similar to that of Gidaspow et al. (1986), was inserted into the #uidized bed to make direct electrostatic charge measurements. A glass sleeve maintained a high resistance to the ground, while a brass tube enclosing the glass tube reduced the background current by eliminating disturbances due to buildup of charges on the column walls. The diameter of the stainless steel ball at the tip of the probe was 3:2 mm as suggested by Gidaspow et al. (1986). Alumel wire was securely fastened into a small hole drilled into the stainless steel ball. Local voidage was measured by a Cber optic system, which consists of an optical Cber probe, a light source, a photo-multiplier and an A=D converter. The probe contained two overlapping bundles of Cne optical Cbers, one delivering light from the light source and the other capturing re#ected light. The intensity of back-scattered light depends on the particle concentration in the path of the transmitted light. The particles in this study were glass beads and polyethylene particles. The glass beads were smooth and spherical, whereas the polyethylene particles had non-smooth surfaces and were non-spherical. Fig. 2 shows scanning electron microscope (SEM) images of these particles. The Sauter mean diameters of the glass beads and polyethylene particles were found to be 321 and 378
m, respectively, by sieving. Polyethylene particles were coated with anti-static spray to reduce adhesion between particles during sieving. The particle size distribution of the glass beads was narrower than that of the polyethylene resin. The particle densities, determined by using a 100 ml pycnometer, were 2458 and 715 kg=m3 , while the corresponding bulk densities were 1478 and 412 kg=m3 for the glass beads and polyethylene, respectively. The polyethylene has a dielectric constant of

∼ 2:3, whereas the dielectric constant for glass beads is

between 5 and 10 depending on their chemical composition and surface Cnish (Reitz, Milford, & Christy, 1993; Jiang, Liang, & Fan, 1994). 3. Experimental results from bubble injection in two-dimensional uidized bed 3.1. E5ect of bubble size

Single bubble injection experiments were carried out in both the two- and three-dimensional columns with the background superCcial gas velocity set just below Umf . The probes were located just below the static bed surface (0:42 and 0:21 m above the distributor for the two- and three-dimensional columns) as suggested by Ciborowski and Wlodarski (1962). The vertical distances between the probe and the tip of the bubble injector were 0.205 and 0:16 m, respectively. Typical traces of voltage and charge versus time are plotted in Fig. 3. “Nose” represents arrival of the leading edge of the bubble at the center of the probe, whereas “bubble–wake interface” indicates the moment when the bubble–wake interface passes the probe. The voltage output is very similar to that obtained by Boland and Geldart (1971=72) who also used a Keithley electrometer and a ball probe with a plastic cover. In their case, instead of an optical Cber probe, a Visicorder determined the location of the bubble as a function of time. In our case, regardless of the bubble size, the Crst peak always corresponds to alignment of the center of the bubble with the center of the ball probe. The voltage reading changes its sign as the bubble passes the ball probe because when the bubble is approaching the probe, the intensity of the electromagnetic Celd between the bubble and the probe increases, while the opposite occurs when the bubble moves away from the probe. Consequently, a bubble whose surface contains particles charged with one polarity produces two voltage peaks of opposite sign, with the sign of the Crst peak

156

A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162 V m ax

Voltage [V]

5 4

Ball probe O ptical fiber probe Valve O n/O ff voltage

3 2 half height tim e

1 0 -1

V m in

-2

bubble

C harge [C ]

0.0 -4.0x10

-10

-8.0x10

-10

-1.2x10

bubble injection period

C transfer = charge transfer

C m ax

nose

-9

bubble-w ake interface 3

4

5

6

7

Tim e [s] Fig. 3. Single bubble injection in two-dimensional #uidized bed of 321
m glass beads: Typical base case results (Ug = 0:122 m=s, RH = 17%, ◦ T = 20 C): (a) voltage output; (b) charge induction and transfer.

being the same as the sign of the charges on the particles (Park et al., 2001). The magnitude of the second peak is always smaller than that of the Crst peak due to charge transfer to the probe. In our experiments, the raw data were converted to current and charge as discussed elsewhere (Park et al., 2001). The charge versus time trace corresponding to Fig. 3a is shown in Fig. 3b. The asymptotic value indicates the net direct charge transfer between the probe and the particles (Park et al., 2001). As indicated in Fig. 3, six parameters were used to characterize the electrostatic charges: Vmax ; Vmin ; Cmax ; Ctransfer , and half-height times for the Crst and second peaks. Vmax is proportional to the charges initially distributed on the bubble surface before contacting the probe and the bubble rise velocity. Ctransfer represents the total charges transferred from the particles to the probe during the contact. Cmax represents the maximum induced and transferred charges. The half-height time for the Crst peak is a measure of the induction time when a bubble is approaching the probe, and is a function of bubble rise velocity, which is related to bubble size. These parameters are used to quantify charge reduction throughout this paper. Fig. 4 shows the experimental results for multiple bubble injection in a two-dimensional #uidized bed of 321
m glass beads. Within 15 s, three air pulses of the same size were injected and the output signals were recorded continuously. The voltage and charge signals were quite repeatable, each having forms similar to those in Fig. 3. The di
polyethylene particles, it was not feasible to inject single bubbles since the particles formed agglomerates and adhered to the wall so that they did not move freely, even at high superCcial gas velocities, and the injected air formed a square-nosed slug which rose very slowly. The electrostatic charges produced by di
A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162

157

◦

Fig. 4. Multiple bubble injection into two-dimensional #uidized bed of 321
m glass beads (Ug = 0:122 m=s, RH = 14%, T = 20 C): (a) voltage output; (b) charge induction and transfer.

6

Voltage [V]

4 2 Vmax (1st peak) 0

Vmin (2nd peak)

-2 0.0

Charge [C]

-4.0x10 -8.0x10

(a)

-10

-10

-1.2x10 -1.6x10

-9

Charge transfer

-9

Cmax -2.0x10

(b)

-9

10

15

20

25

30

35

40

Injected volume of bubble [ml] Fig. 5. E
3.2. E5ect of relative humidity of 7uidizing gas The relative humidity of the #uidizing air was increased from 6% to 98% by varying the relative proportions of air passing through the humidiCer and dryer columns. Before each measurement, there was a 2-h waiting period to ensure that steady state had been achieved. The voltage output and charge curves obtained for dif-

ferent relative humidities were similar to the base case shown in Fig. 3. The magnitudes of the six parameters, Vmax ; Vmin ; Cmax ; Cmin , and half-height times for the Crst and second peaks are shown in Figs. 6 –8. These parameters are a
158

A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162 4 B

A

C

E

D

V m ax (1st peak) 3

V m in (2nd peak)

V oltage [V ]

2

1

0

-1 0

20

40

60

80

100

R elative hum idity of fluidizing gas [% ]

Fig. 6. E
2.0x10

B

A

-10

D

C

E

C harge [C ]

0.0

-2.0x10

-10

-4.0x10

-10

-6.0x10

-10

-8.0x10

-10

C m ax C harge transfer 0

20

40

60

80

100

R elative hum idity of fluidizing gas [% ]

Fig. 7. E
relative humidity were maintained constant throughout the experiments. Figs. 6 –8 are each divided into Cve equal zones. In zone A (0 –20% relative humidity), the maximum voltage (Vmax ) decreased while the minimum voltage (Vmin ) remained constant as the relative humidity increased. As seen in Fig. 7, the magnitudes of maximum charge (Cmax ) and charge transfer (Ctransfer ) also decreased as the relative humidity of the #uidizing air increased. There was some disturbance in voltage during bubble injections, but the magnitudes of these disturbances were small com-

pared to the overall voltage signal. Fig. 8 shows that the half-height times for the Crst and second peaks were nearly constant at a low value as the relative humidity increased to 20%. Since the half-height time for the Crst (positive) peak is proportional to the bubble residence time, whereas the half-height time for the second (negative) peak is an indicator of the time required for charge dissipation, the constant value of the Crst half-height time throughout the entire humidity range suggests that the bubble size and rise velocity remained the same as the relative humidity was increased. The constant value of

A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162

159

3.0 A

E

D

C

B

First peak

2.5

H alf height tim e [s]

Second peak 2.0

1.5

1.0

0.5

0.0 0

20

40

60

80

100

R elative hum idity of fluidizing gas [% ]

Fig. 8. E
the second half-height time, on the other hand, indicates that the rate of charge dissipation did not change as the relative humidity of the air increased to 20%. In zone B, the voltage output pattern was very similar to that in zone A. However, Vmin ; Cmax and Ctransfer remained quite constant and low as the relative humidity increased from 20% to 40%. On the other hand, as shown in Fig. 6, the maximum voltage increased slightly as the relative humidity increased, possibly due to a slight increase in surface conductivity, which allowed higher charge transfer between the metallic ball probe and the particles, and among the particles. According to Fig. 8, at relative humidities between 20% and 40%, the time required to dissipate charges through the ball probe (half-height time for the second peak) was almost the same as in zone A. In zone C, extending from 40% to 60% relative humidity, the maximum charge was nearly constant, and the charge transfer was almost zero, indicating that there was little accumulation of static charges (see Fig. 7). Accordingly, the magnitudes of both the minimum and maximum voltage (see Fig. 6) were lower than in zone B. The main characteristic of this zone was that charges dissipated over long periods of time, as indicated by the large half-height time of the second peak, i.e. the relative humidity was suOcient to increase the surface conductivity of the dielectric particles. As a result, even after a bubble burst at the bed surface the accumulated charges around the ball probe continued to dissipate through particle contacts. The large error bars in this zone indicate that the charge dissipation is highly unstable in this humidity range. In zone D (60%–80% RH), most parameters were similar to those in zone C. As shown in Fig. 8, the half-height

times were much smaller than in zone C. In addition, the magnitude of the minimum voltage slightly increased when the relative humidity exceeded 60%. This is likely associated with the high humidity, which signiCcantly increased surface conductivity so that less time was required to dissipate the charges. The signal tended to be unstable, showing a slight discontinuity when the nose of the bubble reached the probe. The absolute values of the charge transfer were close to zero, indicating less accumulation of charges on the particle surface. Finally, there was an unstable zone E due to capillary forces for relative humidities beyond 80%. At 98% relative humidity, a square-nosed slug formed when air was injected into the bed. Such high humidities should clearly be avoided. Instead, relative humidities between 40% and 80% appear to be most advantageous, with faster charge dissipation occurring between 60% and 80% for the conditions studied. In zone C the charge on the particle surface surrounding the bubble was lower than in zones A and B, as re#ected by Vmax . The charge dissipation following the passage of the gas bubble was highly unstable. In zone D, charge accumulation is quite low and the charge dissipation following the bubble passage is quite stable. Therefore, relative humidities between 60% and 80% appear to be optimal for reducing electrostatic charge accumulation, at least in the gas–solid #uidized bed studied. 4. Experimental results for bubble injection in three-dimensional uidized beds Since two-dimensional #uidized beds are subject to substantial wall e
160

A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162 5

V oltage [V ]

4

B all probe O ptical fiber probe

3 2 1 0 -1

C harge [C ]

1.0x10

-10

charge transfer

0.0 -1.0x10

-10

-2.0x10

-10

-3.0x10

-10

-4.0x10

-10

-5.0x10

-10

3

4

5

6

7

8

9

10

Tim e [s]

Fig. 9. Voltage and charge output for bubble injection into three-dimensional #uidized bed of 321
m glass beads at RH = 20% ◦ (Ug = 0:085 m=s; dB ∼ = 28 mm; T = 23 C).

2.5

V m ax (G B )

2.0

V m in (G B ) V m ax (P E I)

V oltage [V ]

1.5

V m in (P E I)

1.0

0.5

0.0

-0.5

-1.0 10

20

30

40

50

60

70

80

90

R elative hum idity of fluidizing gas [% ] Fig. 10. E
were repeated in the three-dimensional column. Fig. 9 shows an example of the voltage output and charge versus time for single bubble injection. The electrostatic ball probe and the optical Cber probe were inserted towards the center of the column at the same height, but perpendicular to each other, with their tips 10 mm apart. Bubble diameters were about 50 mm. Signals from the optical Cber probe indicated that bubbles passed as distinct entities.

Signals in Fig. 9 are similar to those in Fig. 3 for the two-dimensional #uidized bed, except that there is more disturbance corresponding to bubble injection. For both cases, the voltage output reached a maximum when the center of the bubble reached the center of the probe. When the bubble injection experiment was repeated in the three-dimensional bed using polyethylene particles, it was again diOcult to inject single bubbles. The experimental trends were similar to those

A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162 4.0x10

-10

2.0x10

-10

161

C harge [C ]

0.0

-2.0x10

-10

-4.0x10

-10

-6.0x10

-10

-8.0x10

-10

C m ax (G B ) C transfer (G B )

-1.0x10

-9

-1.2x10

-9

C m ax (P E I) C transfer (P E I)

10

20

30

40

50

60

70

80

90

R elative hum idity of fluidizing gas [% ]

Fig. 11. E
with the glass beads, with the particles again charged positively. Figs. 10 and 11 show that an increase in relative humidity of the #uidizing air reduced not only the magnitudes of the positive and negative peaks but also the magnitudes of the maximum charge and charge transfer. The polyethylene particles have porous structures, whereas the glass is totally non-porous, and this may explain why the degree of charge reduction with increasing relative humidity was higher for the polyethylene particles than for the glass beads. 5. Conclusion Electrostatic charges in gas–solid #uidized beds were studied by injecting bubbles into two- and three-dimensional #uidized beds. Both glass beads and polyethylene particles were charged positively. Larger bubbles resulted in higher induction and transfer of electrostatic charges. Increasing the relative humidity reduced the electrostatic charge accumulation by increasing the surface conductivity, thereby enhancing charge dissipation. For the conditions studied, relative humidities between about 40% and 80% were most helpful in eliminating electrostatic charges, while over-humidiCcation led to excessive capillary forces causing de#uidization. Notation Cmax Ctransfer

maximum magnitude of charge, C (see Fig. 3) charge transfer, C (see Fig. 3)

Itotal QP R RH t T Ug Umf V Vmax Vmin

total current, A total charge on probe, C electrical resistance, P relative humidity, % time, s ◦ temperature, C superCcial gas velocity, m=s minimum #uidization velocity, m=s voltage, V maximum voltage of Crst peak (positive peak), V (see Fig. 3) minimum voltage of second peak (negative peak), V (see Fig. 3)

Acknowledgements Financial support from the Natural Sciences and Engineering Research Council of Canada and the Mitsubishi Chemical Corporation is gratefully acknowledged. References Ally, M. R., & Klinzing, G. E. (1985). Inter-relation of electrostatic charging and pressure drops in pneumatic transport. Powder Technology, 44, 85–88. Boland, D., & Geldart, D. (1971). Electrostatic charging in gas #uidized beds. Powder Technology, 5, 289–297. Ciborowski, J. S., & Wlodarski, A. (1962). On electrostatic e
162

A.-H. Park et al. / Chemical Engineering Science 57 (2002) 153–162

Grace, J. R., & Baeyens, J. (1986). Instrumentation and experimental techniques. In D. Geldart (Ed.), Gas 7uidization technology (pp. 415 – 462). New York: Wiley. Guardiola, J., Ramos, G., & Romero, A. (1992). Electrostatic behavior in binary dielectric=conductor #uidized beds. Powder Technology, 73, 11–19. Guardiola, J., Rojo, V., & Ramos, G. (1996). In#uence of particle size, #uidization velocity and relative humidity on #uidized bed electrostatics. Journal of Electrostatics, 37, 1–20. Jiang, P., Bi, H., Liang, S., & Fan, L. (1994). Hydrodynamic behavior of circulating #uidized bed with polymeric particles. A.I.Ch.E. Journal, 40, 193–206. Katz, H., & Sears, J. T. (1969). Electric Celd phenomena in #uidized and Cxed bed. Canadian Journal of Chemical Engineering, 47, 50–53.

Newbold, P. (1991). Statistics for business and economics (3rd ed.) (pp. 301–307). New Jersey: Prentice-Hall. Park, A.-H., Bi, H. T., Grace, J. R., & Chen, A. H. (2001). Modeling charge transfer and inducement in gas–solid #uidized beds. Journal of Electrostatics, in press. Reitz, J. R., Milford, F. J., & Christy, R. W. (1993). Foundations of electromagnetic theory (4th ed.) (pp. 1–126). Massachusetts: Addison-Wesley. Wolny, A., & Kazmierczak, W. (1989). TribelectriCcation in #uidized bed of polystyrene. Chemical Engineering Science, 44, 2607– 2610. Wolny, A., & Opalinski, I. (1983). Electric charge neutralization by addition of Cnes to a #uidized bed composed of coarse dielectric particles. Journal of Electrostatics, 14, 279–289.