Electrostatics of the Granular Flow in a Pneumatic Conveying System

Ind. Eng. Chem. Res. 2004, 43, 7181-7199

7181

Electrostatics of the Granular Flow in a Pneumatic Conveying System Jun Yao,† Yan Zhang,‡ Chi-Hwa Wang,*,†,‡ Shuji Matsusaka,§ and Hiroaki Masuda§ Singapore-MIT Alliance, E4-04-10, 4 Engineering Drive 3, Singapore 117576, Singapore, Department of Chemical and Bimolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117576, Singapore, and Department of Chemical Engineering, Kyoto University, Kyoto 615-8510, Japan

The phenomenon of electrostatic charge generation and its effects on granular flow behavior in a pneumatic conveying system was studied. The main parameters used for quantitative characterization of the phenomenon were the induced current, particle charge density, and equivalent current of the charged granular flow. These were measured using a digital electrometer, Faraday cage, and modular parametric current transformer, respectively. Three different flow patterns corresponding to different electrostatic effects within the pneumatic conveying system were observed, and these were named the disperse flow, half-ring flow, and ring flow patterns. It was found that the induced current, particle charge density, and equivalent current increased with decreasing flow rates. Electrostatic effects generally become stronger with time, and this may lead to clustering behavior occurring even in the disperse flow regime. The effects of several factors such as pipe wall material, particle composition, relative humidity of the conveying air used, and the presence of an antistatic agent in the system were investigated and found to be important in determining the electrostatic charge generation characteristics and granular flow patterns observed. 1. Introduction Pneumatic conveying systems are widely used in the energy, chemical, pharmaceutical, and material processing industries for the transportation of granular material. Solid particulate systems have a natural tendency to acquire electrostatic charges due to collisions with surfaces of a different material type.1 The accumulation of charges on system components may pose possible electrical hazards and lead to compromises in the safety standards of such operations. As such, a comprehensive understanding of the effects of electrostatics on the flow of granular material in a pneumatic conveying system is required. Electrostatic effects and the associated charge generation mechanisms are complex phenomena often dependent on a variety of factors such as the physical, chemical, and electrical characteristics of the material used and ambient conditions. This may give rise to poor reproducibilities and repeatabilities of experiments where such phenomena are the main focus of investigation. The amount of work reported in the literature that involves measuring or calculating electrical charges on particles in granular flow systems has also been limited due to the inherent difficulties in such investigations. Zhang et al.2 carried out a comprehensive and complete study of the effects of electrostatic forces on cohesive particles. On the basis of experimental data on instantaneous velocities of shale particles in a dilute gas/solid system, it was found that the particles oscillate about the mean values in a chaotic manner, the velocity * To whom correspondence should be addressed. Tel.: 656874-5079. Fax: 65-6779-1936. E-mail: [email protected]. † Singapore-MIT Alliance. ‡ National University of Singapore. § Kyoto University.

distribution could be approximated by the Maxwellian distribution function, and kinetic theory was a very promising approach to describe gas/particle flow systems. Matsusaka and Masuda3 developed a formulation for the variation of particle charging caused by repeated impacts on a wall and employed the formulation to particle charging in a granular flow where each particle carried a different amount of charge. They then analyzed theoretically the particle charge distribution. Nieh et al.4 considered the effects of various combinations of air humidity, conveying velocities and particle sizes on generation of electrostatic charges and charge distributions in glass beads flowing in a grounded 2-in. copper pipe loop. They demonstrated that air humidity had a significant effect on particle charges. When the system moisture content exceeded a cutoff relative humidity (RH) of 76%, the charges on glass beads became effectively neutralized. Kanazawa et al.5 investigated the charge generation and accumulation on the inner surface of a pipe in a cascade flow system during the transport of powder. It was found that the surface charge density and charge polarity of the pipe wall depended on the successive number of tests and the material of the pipe, respectively. Furthermore, the charging characteristics of the pipe wall during the transport of glass beads and industrial granules were found to depend on the pipe material and transport method. Diu and Yu6 made an analysis on the deposition of a charged aerosol passing through a two-dimensional bend due to the simultaneous action of inertial and electrostatic forces. The collection efficiency for an idealized rotational flow was obtained, and the results were qualitatively compared to available experimental data. It was found that a charged particle would experience an electrostatic force and the resulting changes in the trajectory of the particle would lead to a change

10.1021/ie049661l CCC: $27.50 © 2004 American Chemical Society Published on Web 09/23/2004

7182

Ind. Eng. Chem. Res., Vol. 43, No. 22, 2004

Figure 1. Experimental setup A; schematic of the pneumatic conveying facility: (1) air control valve, (2) air dryer (silica gel with indicator blue), (3) rotameter, (4) rotary valve, (5) MPCT, (6) induced current measurement, (7) Faraday cage (details shown in view I), (8) electrometer, (9) computer, (10) feed recycle hopper, (11) feed control valve, (12) intermediate hopper, (13) electronic weight indicator, (14) feed control valve. View I: (1) PVC pipe, (2) particle sampling pipe (copper), (3) control valve, (4) Faraday cage, (5) electromagnetic shield.

in the collection efficiency of the bend. Al-Adel et al.7 examined the characteristics and extent of lateral segregation arising from static electrification in riser flows and analyzed the steady and fully developed gasparticle flow in a vertical riser. Wolny and Opalinski8 studied the mechanism of electric charge neutralization by the addition of fines to a fluidized bed. According to this mechanism, the fines change the contact conditions between particles of the bed, and transfer electric charge between particles, causing neutralization of the whole bed. For example, antistatic powder Larostat-519 was used by Zhang et al.2 and Chang and Louge,9 but its antistatic mechanism was never explained. The effects

of electrostatic charge on the flow of granular material have been studied by a few research workers (Smeltzer et al.,10 Nieh and Nguyen,4 Gajewski,11 Joseph and Klinzing12). However, flow patterns of the granular material that can arise due to such effects in a pneumatic conveying system have not been reported. Tsuji and Morikawa13 observed five different granular flow patterns in a horizontal conveying pipe depending on the gas-solid flow rates used but did not analyze the influence of electrostatics on these flow patterns. On the basis of our previous work, it has been established that electrostatic effects play an important role in determining the flow patterns in a vertical pneumatic conveying

Ind. Eng. Chem. Res., Vol. 43, No. 22, 2004 7183 Table 1. Experimental Conditions experimental setup A (Figure 1) particle conveying style particle diameter particle material particle density pipe material pipe diameter (inner) pipe thickness RH air flow rate solid mass flow rate air pressure air superficial velocity temperature film material film thickness

cycle 2.8 mm PP 1123 kg/m3 PVC 40.0 mm 5.0 mm 5% 860-1600 L/min 10.0 ( 2 g/s to 44.4 ( 4 g/s 75 psi 11.4-21.2 m/s 28-30 °C

experimental setup B (Figure 14) no cycle 2.8 mm PP 1123 kg/m3 copper 50.0 mm 1.0 mm 5% 1600 L/min 44.4 ( 4 g/s 75 psi 21.2 m/s 28-30 °C PVC/PE 5.0/0.04 mm

system (Rao et al.,14 Zhu et al.15,16). This paper reports on three flow patterns observed in both vertical and horizontal conveying pipes, which have been named disperse, half-ring, and ring flow. These are believed to arise according to the relative importance between the electrostatic forces and the usual hydrodynamic forces present between the gas and solid phases. 2. Experimental Section The experimental setup used in our present study was modified from the pneumatic conveying system used by Rao et al.14 and Zhu et al.15,16 The schematic diagram of the modified system is given in Figure 1, and the experimental conditions are listed in Table 1. Solid particles are introduced into the rotary valve and entrained by air flowing from the compressor mains. The rotary valve (General Resource Corp., Hopkins, Minnesota) has eight pockets on the rotary and rotates at 30 rpm. The inner diameter of the pipe is 40 mm, and the length of the vertical pipe section between two smooth 90° elbows (R/r ) 2) is about 2.97 m, while the horizontal section is about 4.12 m in length. The conveying pipe is made of transparent poly(vinyl chloride) (PVC) material and has a wall thickness of 5 mm. Induced current measurements were made on the vertical pipe 1.36 m away from the bottom elbow. Static (Gems, model K054205, Basingstone, England) and differential (Stellar Technology Inc., New York, model STI510-15G-000Stellar) pressure transducers were mounted on the vertical pipe at 1.22 and 1.49 m away from the bottom elbow. Pressure data were collected for 30 s, and the mean pressure gradients were determined for different operating conditions. Two types of particles consisting of polypropylene (PP) granules and a mixture of PP granules and glass beads (3 wt %) and one type of antistatic powder were used throughout the experiments. Visual observations of the various solid flow patterns that arise during the experiments were facilitated through the use of transparent pipes. The entire

Figure 2. Typical pattern of particles disperse flow (air flow rate > 1200 L/min): (a) at the beginning; (b) about 2 h later for the case of 1200 L/min; (c) snapshot at a pipe section.

configuration was held in position using metal castings and supports with various pipe segments joined by connectors and reinforced by silicone gel. Air from the compressor mains flowed through the rotary feeder, driving granules fed into the conveying system from the feed hopper. Control valves (labeled 1, 11, and 14 in Figure 1) were used to adjust the solid and air flow rates. The air flow rate was also controlled via a rotameter which allowed a maximum flow rate of 2000 L/min. The air humidity was controlled by the dryer (silica gel with indicator blue, labeled 2 in Figure 1) at RH ) 5% and was checked using a high-

Table 2. Three Characteristic Patterns Developed in a Vertical Conveying Pipea experiment case

flow pattern

air flow rate (L/min)

air superficial velocity (m/s)

solid flow rate (g/s)

Gs (kg/m2 s)

∆P/L (Pa/m)

case 1 case 2 case 3

disperse flow half-ring flow ring flow

1600 1000 860

21.2 13. 3 11.4

44.4 ( 4 17.4 ( 3 10.2 ( 2

35.3 ( 3.2 13.8 ( 2.4 8.1 ( 1.6

132.3 263.7 518.9

a Disperse flow: air superficial velocity > 16.2 (m/s); ∆P/L < 182.3 (Pa/m). Half-ring flow: air superficial velocity, 12.1-15.7 (m/s); ∆P/L, 235.5-389.7 (Pa/m). Ring flow: air superficial velocity < 11.6 (m/s); ∆P/L > 446.3 (Pa/m).

7184


Figure 3. Typical pattern of particles half-ring flow (air flow rate 900-1150 L/min): (a) at the beginning; (b) about 30 min later for the case of 900 L/min; (c) snapshot at a pipe section.

performance digital thermo-hygrometer (RH411, OMEGA Technologics Ltd.) before and after each test. As PP granules were fed back to the recycle solid hopper, they passed through a manual valve (labeled 11 in Figure 1). Closing valve 11 would enable the solid mass flow rate to be measured using the electronic weight indicator (labeled 13 in Figure 1). Valve 14 was used to control the solid feed rate from the solid feed hopper into the conveying system. Ambient temperature was controlled at 28-30 °C, and other experimental conditions are listed in Table 1. During the pneumatic conveying process, collisions between the solid particles and the pipe wall generate electrostatic charges. The current induced along the surface of the pipe wall as a result of these charges was

measured as a function of time. This was done by wrapping an aluminum foil sheet tightly over the outer wall of the PVC pipe (labeled 6 in Figure 1). A coaxial line (connected to the high input end of an electrometer, Advantest R8252 digital electrometer) was connected to the outer surface of the aluminum foil sheet. A polymer film was then wrapped tightly over the aluminum foil sheet to separate this sheet from another aluminum foil sheet whose external surface was connected to the low input end of the coaxial cable. Subsequently, this external layer of aluminum foil sheet was connected to ground and used as an extra electrical shield. The induced current through the pipe wall was measured as a function of time through digital readings from the electrometer and was stored in a computer at

Ind. Eng. Chem. Res., Vol. 43, No. 22, 2004 7185

Figure 4. Typical pattern of particles ring flow (air flow rate < 850 L/min): (a) at the beginning; (b) about 15 min later for the case of 850 L/min; (c) snapshot at a pipe section.

intervals of 0.5 s. Additionally, a modular parametric current transformer (MPCT: Bergoz Instrumentation, France) which allows for current measurement with a resolution of 1 µA using a noninvasive DC beam was also used to measure the induced current. The charge density of the particles was measured by a Faraday cage (TR8031), labeled 7 in Figure 1. The mass of particles collected in the cage was measured using an electronic balance to an accuracy of 10-4 g, and the mass-to-charge ratio of the particles was then calculated. 3. Results and Discussion As mentioned in the Introduction, the solid flow patterns observed in the vertical section of the pneumatic conveying system can be classified into the disperse flow, half-ring flow, and ring flow regimes. The air flow rates and superficial velocities at which each

flow pattern occurs and the corresponding pressure drop per unit length of pipe are shown in Table 2. 3.1. Electrostatics of Granular Flow in the Vertical Pipe. 3.1.1. Granular Flow Pattern. I. Disperse Flow. When the air flow rate exceeds 1200 L/min, particles are transported at high velocities up the vertical pipe and do not concentrate on the pipe wall (Figure 2a). However, after a long time (about 2 h for the case of 1200 L/min), some particles were found to concentrate on the pipe wall from time to time in the form of clusters (Figure 2b,c). Due to the strong air flow, the clusters of particles seemed fairly unstable and exhibited many fluctuations in their vertical positions along the pipe. The clusters were located fairly high up in the pipe and traveled along a curved path by the pipe wall (Figure 2b). These clusters appeared and disappeared intermittently in an unpredictable manner.

7186


Figure 5. Induced current acquired at the vertical pipe: (a) disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); (b) half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s); (c) ring flow (air flow rate 860 L/min, solids feed rate 8.1 ( 1.6 kg/m2 s); (d) comparison of the current value (negative) for the three flows.

II. Half-Ring Flow. When the air flow rate is about 900-1150 L/min, particles tend to move along curvilinear tracks in the vertical pipe (Figure 3a). However, after about half of an hour of cycling through the pneumatic conveying system, particles were observed to cluster on the side of the vertical pipe wall distant from the position of the lower horizontal pipe to form a half-annular ring structure (Figure 3b). This structure was located at a height of about 30 cm above the bottom elbow. III. Ring Flow. When the air flow rate was less than 850 L/min, the effects of gravitational and electrostatic forces became more significant. Initially, particles slid upward in contact with the surface of the pipe (Figure 4a). After about 15 min, the particles were observed to travel in a spiral fashion up the vertical pipe along the pipe wall. This resulted in a ring or annulus structure with high particle concentrations adjacent to the wall and a relatively empty core region (Figure 4b,c). The structure was similarly unstable and fluctuated in its vertical position along the pipe. At an air flow rate of 880 L/min, which was intermediate between those giving rise to the half-ring and ring flow regimes, the observed flow pattern alternated

Table 3. Force on Particles drag (FD) core region force (N) magnitude

boundary region

gravity halfhalf(FG) disperse ring ring disperse ring ring 10-4

10-2

10-3 10-3

10-3

10-4 10-4

between the two characteristic half-ring and ring structures. This might be indicative of an unstable or oscillatory transition between the two flow regimes at this particular air flow rate used. 3.1.2. Dynamics Analysis. In vertical pneumatic conveying, the main forces acting on the particles are the fluid drag force, gravity, and electrostatic forces. If the particles slide on the wall surface, frictional forces would act in the direction opposite to that of its motion. In general, the fluid drag force FD is most dominant in influencing particle behavior (Sommerfeld17):

F BD )

3 Fgmp c (u b -b up)|u bg - b u p| 4 F p Dp D g

If the particle shape is considered as spherical, then the


Figure 7. Comparison of particle charge density (using Faraday cage) for the three flows: disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s); ring flow (air flow rate 860 L/min, solids feed rate 8.1 ( 1.6 kg/m2 s).

Rep )

FgDp|u bg - b u p| µg

(4)

If the particle number density is low, the image force FE (Yu18) acting on the particle due to the presence of the wall is the principal force responsible for particle concentration. For cylindrical geometry,

FE )

Figure 6. Comparison of the induced current acquired at the vertical pipe: (a) charges obtained by integration of the currents (shown in Figure 5a-c) for the three flows, (b) the positive and negative values (absolute) of the induced current of the disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s).

above equation can be simplified as

πFgDp2 F BD ) bg - b up)|u bg - b u p| CD(u 8

(1)

where Fg is the gas density, Fp is the particle density, mp is the mass of the particle, Dp is the diameter of the particles, and b ug and b up is the translational velocity vector of gas and particle, respectively. The drag coefficient, CD, is defined as

CD )

24 24 (1 + 0.15Rep0.687) ) f Rep Rep D CD ) 0.44

Rep > 1000

where Rep can be obtained by

Rep e 1000 (2) (3)

q2r2 16πR20(R - r)2

(5)

where R is the cylindrical vessel radius, r is the radial distance, 0 is the air permittivity, and q is the particle charge. It is known that the drag force acting on a particle varies with the distance of the particle from the pipe wall due to nonuniform gas velocity distribution in the pipe. Within the boundary layer of the pipe wall, the gas velocity decreases sharply, and, by eq 1, the corresponding drag force acting on a particle decreases accordingly. On the basis of the work of a similar problem (Fan et al.19), the magnitude of the drag force estimated at the boundary layer and core region is listed in Table 3. In disperse flow, the air flow rate is sufficiently high for the magnitude of fluid drag forces to be much larger than that of gravitational forces in both the pipe core and the boundary layer regions. At lower air flow rates giving half-ring and ring flows, fluid drag forces within the wall boundary layer are reduced to the same order of magnitude as that of gravitational forces so that a dynamic equilibrium may be established between the two types of forces. Consequently, other types of forces such as the electrostatic force may then emerge as the dominant factor affecting particle behavior. In particular, the image force represented by eq 5 above may be increased significantly due to the Table 4. Mean Induced Current (A) for Three Characteristic Flows disperse flow half-ring flow ring flow

vertical (Figure 6a)

horizontal (Figure 11)

1.77 × 10-9 6.98 × 10-9 1.24 × 10-8

2.80 × 10-10 1.67 × 10-9 2.79 × 10-9

7188


Figure 8. Comparison of charges obtained by integration of 400 MPCT values within a time interval of 20 s acquired at the vertical pipe for the three flows: disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s); ring flow (air flow rate 860 L/min, solids feed rate 8.1 ( 1.6 kg/m2 s).

close proximity of the particles to the pipe wall at low air flow rates or high charge densities acquired by the particles. This suggests that the image force may be responsible for causing particles to stick to the pipe wall and forming the half-ring or ring structure at these flow rates. 3.1.3. Induced Current. In the experiments conducted, attempts have been made to analyze the induced current for the three types of flow described above (Figures 2-4). The induced current detected using the Advantest R8252 digital electrometer at the vertical pipe is illustrated in Figure 5. It can be seen that the induced current fluctuates with time and may be negative or positive in value. It is known that particle-wall collisions in a conveying system lead to charge generation both on the surface of the particles and on the pipe wall.1 On the basis of previous experiments16 using PP particles and PVC pipes, it can be deduced that positive charges are generated on the PP particle surface while equal amounts of negative charges are accumulated on the PVC pipe wall. At the detection segment, if particles slide on the pipe wall as in half-ring or ring flow, charges on the pipe wall are partly neutralized by those on the surface of particles and lead to a lower induced current. On the other hand, when there are no particles sliding on the pipe wall of the detection segment, charges on the pipe wall are retained, leading to a higher induced current. The charge can be modeled by the following equations: i)n

q ) w(

qi/mpi)/n ) wqm ∑ i)1

dq ) qm dw + w dqm I)

dqm dw dq ) qm +w dt dt dt

(6) (7) (8)

where w is the mass of particles at the detection segment, qi is the charge of the ith particle, mpi is the mass of the particle, and qm is the average charge density. Equation 6 says that the charge acquired by

Figure 9. Typical pattern of particles flows at horizontal pipe: (a) disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); (b) half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s); (c) ring flow (air flow rate 860 L/min, solids feed rate 10.2 ( 2. kg/m2 s).

the electrometer from the pipe wall is equal to the product of the particle mass and the charge density at the test section. Equation 7 states that charge variation may arise due to variations in particle mass and charge density; the latter consists of the variation of the initial charge and the transferred charge between the particles and the wall. Equation 8 shows the formulation for induced current obtained by taking the time differential of the previous equation. Generally, eqs 6-8 show that the induced current is a combined effect of the solid mass flow fluctuation and the charge transferred to the pipe wall due to particle collisions. The polarity of the induced current depends on the relative magnitudes of these two effects. From the above analysis, it can be said that the induced current measured by the Advantest R8252 digital electrometer is a composite value resulting from a balance between the electrostatic charges on the particle surface and pipe wall. The absolute value of the induced current indicates the electrostatics of the granular flow.


Figure 10. Induced current acquired at horizontal pipe: (a) disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); (b) half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s); (c) ring flow (air flow rate 860 L/min, solids feed rate 10.2 ( 2 kg/m2 s); (d) comparison of the induced current value (negative) for the three flows.

Figure 5 shows that the magnitude of the induced current increases with decreasing flow rate. Figure 5d shows the absolute values of the negative current sampled at a rate of 200 s-1. It is evident that the absolute current value increases with decreasing flow rate, implying enhanced electrostatics. In addition, to eliminate the fluctuation caused by negative and positive values (Figure 5), induced currents were integrated with respect to time according to eq 9 to obtain the charge Q.

Q)

∫0t I dt

(9)

Figure 6a shows that the charge Q detected at the detection segment increases linearly with time up to 5000 s and the rate of increase seemed constant for each of the three flow patterns (Table 4). Furthermore, the magnitudes of negative currents detected were generally larger than those of positive currents, implying a net electron gain by the pipe wall (Figure 6b). 3.1.4. Particle Charge Density. The variations of the charge-to-mass ratio of particles with respect to time

for the three types of flow are presented in Figure 7 and Tables 5-7. It can be seen that particle charge densities in the half-ring and ring flow are similar, both being larger than those in disperse flow. For all three cases, the general trend for particle charge density to increase with time does indicate possible accumulation of charges on the particles. This may eventually result in strong electrostatic interactions between the particles and the pipe wall and cause formation of the half-ring and ring structures described above. 3.1.5. MPCT Measurement. The equivalent currents for the three flow patterns (disperse, half-ring, and ring) at the vertical section of the pneumatic conveying system are presented in Figure 8, where each data point was obtained by time integration of 400 current readings from the MPCT over a time interval of 20 s using eq 9. It can be seen that MPCT values in the half-ring and ring flows are fairly similar and larger than that in the disperse flow. This is consistent with the trend observed for particle charge densities for the three flow patterns discussed above. Furthermore, Figure 8 also suggests that particles carry positive charges while

7190


Figure 11. Comparison of the charges obtained by integration of the currents acquired at horizontal pipe for the three flows (shown in Figure 10a-c). Table 5. Transient Equivalent Current of Charged Disperse Flow time (min)

Qpa (10-10 C/g)

MPCT (10-8 A)

Icb (10-8 A)

66 78 91 104 116 128 141 157 170 184

3.761 4.406 11.586 8.908 11.790 15.076 13.344 15.248 9.705 12.216

1.713 1.746 1.715 1.548 1.743 1.735 1.944 1.704 1.703 1.763

1.670 1.956 5.144 3.955 5.235 6.694 5.925 6.770 4.309 5.424

a

Qp: particle charge density. b Ic ) Qp(SF); SF ) 44.4 ( 4 g/s.

Table 6. Transient Equivalent Current of Charged Half-Ring Flow time (min)

Qpa (10-9 C/g)

MPCT (10-8 A)

Icb (10-8 A)

12 23 36 47 59 75 88 99 110 124 135 151 162 175 187

1.258 1.946 2.531 2.022 2.911 1.425 2.968 2.146 2.499 2.108 2.679 2.716 3.190 2.097 2.942

2.859 3.124 3.417 2.537 2.841 2.734 2.326 2.096 1.970 1.974 1.881 2.060 2.105 2.058 2.106

2.190 3.386 4.404 3.518 5.065 2.479 5.164 3.735 4.349 3.668 4.661 4.725 5.550 3.650 5.118

a


being transported along the pipe, and this agrees with the contact potential difference measurements by Zhu et al.16 3.2. Electrostatics of Granular Flow in the Horizontal Pipe. 3.2.1. Granular Flow Pattern. The three flow patterns observed in vertical pneumatic conveying were also observed in the horizontal section of the system. At high air flow rates, particles were seen to be transported as a dilute and homogeneous solid phase dispersed in a continuous gas phase (Figure 9a). As the flow rates were decreased, particles started to cluster together intermittently (Figure 9b),

Figure 12. Electrostatics of vertical pipe versus horizontal pipe (disperse flow with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s): (a) charges obtained by integration of induced currents (Figures 5a and 10a); (b) charges obtained by integration of 400 MPCT values within a time interval of 20 s.

and as the flow rates were reduced further, a fairly stable structure was formed which moved along the horizontal pipe at velocities smaller than the superficial air velocity. 3.2.2. Induced Current. The induced currents for the three types of flow in the horizontal pipe segment were similarly measured using the Advantest R8252 digital electrometer. Figure 10 shows that the amplitude of fluctuations in the induced current increases with decreasing flow rate (about (9.0 × 10-8 A in disperse flow, (1.5 × 10-7 A in half-ring flow, and (2.8 × 10-7 A in ring flow). The magnitudes of the negative component of the induced current sampled at a rate of 200 s-1 also increased with decreasing flow rates (Figure 10d), implying stronger electrostatic interactions between the particles and pipe wall at lower flow rates. This can also be deduced from variations of the total accumulated charge on the pipe wall obtained by time integration of the data presented in Figure 10a-c. Figure 11 shows that the rate of charge accumulation is larger at lower air flow rates and vice versa. This

Ind. Eng. Chem. Res., Vol. 43, No. 22, 2004 7191 Table 7. Transient Equivalent Current of Charged Ring Flow time (min)

Qpa (10-9 C/g)

MPCT (10-8 A)

Icb (10-8 A)

2 14 29 63 74 85 97 107 122 135

1.494 2.355 1.189 2.471 2.091 2.387 1.906 1.800 2.653 2.249

2.444 2.609 2.319 2.573 2.172 2.342 2.351 2.416 2.003 2.894

1.524 2.402 1.213 2.520 2.133 2.435 1.944 1.836 2.706 2.294

a


Table 8. Transient Equivalent Current of Charged Disperse Flow for Particle Mixtures: PP and Glass Bead (by 3 wt %) time (min)

Qpa (10-9 C/g)

MPCT (10-8 A)

Icb (10-8 A)

30 40 50 64 77 90 103 116 129 142 156

1.181 1.268 1.457 1.055 1.538 1.281 1.142 1.210 1.298 1.032 1.398

2.135 2.772 2.498 2.779 2.740 2.711 2.990 3.366 3.389 3.707 2.895

5.246 5.629 6.466 4.682 6.829 5.688 5.069 5.374 5.765 4.582 6.207

a


Table 9. Transient Equivalent Current of Charged Half-Ring Flow for Particle Mixture: PP and Glass Bead (by 3 wt %) time (min)

Qpa (10-9 C/g)

MPCT (10-8 A)

Icb (10-8 A)

6 17 28 39 51 63 74 88 99 109 120 131 141 153 165 178

1.358 2.134 1.553 1.283 1.224 1.274 1.436 1.549 1.591 1.969 1.487 1.558 1.909 1.640 1.522 1.582

2.193 1.673 2.093 1.759 2.056 2.420 2.201 2.009 1.935 2.331 3.260 3.139 4.288 3.517 4.098 3.571

2.363 3.715 2.703 2.232 2.130 2.216 2.499 2.695 2.769 3.426 2.588 2.711 3.321 2.854 2.648 2.753

a

Figure 13. Induced current value (negative) of vertical pipe versus horizontal pipe: (a) disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); (b) half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s); (c) ring flow (air flow rate 860 L/min, solids feed rate 10.2 ( 2. kg/m2 s).

implies that the total amount of charge accumulated on the pipe wall after a particular amount of time is larger in the ring and half-ring flow regimes than in the disperse flow regime. 3.3. Comparison of the Electrostatics between the Vertical and Horizontal Pipes. Moreover, in our present work, particle concentration patterns, in terms of cluster, half-ring, and ring structure, were observed at the vertical pipe wall but were not observed at the horizontal pipe wall. A comparison of the electrostatic charge effects occurring at the vertical and horizontal


segment of the system shows larger amounts of charge being accumulated on the vertical pipe. This is seen in both the time-integrated data obtained from the electrometer (Figure 12a) and the MPCT (Figure 12b). The MPCT data also show more fluctuations in vertical than horizontal conveying (Figure 12b). In addition, the absolute values of the negative component of the induced current are also larger in the vertical pipe as compared to those in the horizontal pipe (Figure 13). All of these characteristics seemed to point to the previously stated conclusion that electrostatic effects are much stronger in vertical pneumatic conveying than in horizontal pneumatic conveying. However, it is to be noted that the effects of the pipe bend connecting the two pipe segments have been neglected in this analysis. It is expected that particles impact against the pipe bend with high velocities and large forces and so may cause significant charge generation at this point in the system.

7192


Figure 14. Experimental setup B: (1) air control valve, 2. air dryer (silica gel with indicator blue), (3) rotameter, (4) feed hopper, (5) electronic weight indicator, (6) feed control valve, (7) rotary valve, (8) computer, (9) electrometer, (10) test segment (details shown in view II), (11) Faraday cage (details shown in Figure 1, view I), (12) metal container. View II; cross section of the test segment (inner diameter, 50 mm; length, 72 mm): (1) film (PVC, 5 mm/PE, 0.04 mm), (2) conductive adhesive (0.07 mm), (3) copper electrode (0.10 mm), (4) nonconductive adhesive (0.12 mm), (5) copper electric shield (0.10 mm).

The possible contribution of this toward the stronger electrostatic effects observed in the vertical pipe segment remains the focus of a subsequent study. 3.4. Intercomparison of Electrostatic Characteristics. The equivalent current Ic of a granular flow system due to the motion of charge-carrying particles can be calculated by the following equation:

Ic ) Qp(SF)

(10)

where Qp is the particle charge density obtained using the Faraday cage and SF is the particle mass transported per unit time. The equivalent currents calculated according to eq 10 for the three types of flow are listed

and compared in Tables 5-9. The calculated values are of the same order of magnitude as those measured using the MPCT. The agreement between the two is better at lower flow rates. This may be attributed to the fact that the MPCT is more accurate when electrostatic effects become stronger at lower flow rates or other experimental factors such as fluctuations in solid flow rate, which may become significant at high air flow rates but are not taken into account in eq 10. 4. Mechanism of Charge Generation The phenomenon of electrostatic charge generation due to particle-particle or particle-wall collisions is known as tribocharging. This is a material-dependent


Figure 16. Charges obtained by integration of induced currents (shown in Figure 15a,b) (experimental setup B) PVC versus PE.

Figure 17. Particle charge density determined using Faraday cage (experimental setup B with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s): PVC versus PE.

Figure 15. Induced current for two films (experimental setup B, horizontal pipe with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s): (a) PVC; (b) PE; (c) PVC versus PE (negative value).

process that has important practical applications such as in the design of photo printing and electrostatic coating machines. In the literature, common insulating materials can be ranked according to their charging tendency to form a triboelectric series. Materials higher up in the series tend to become positively charged when rubbed or brought into contact with those lower down in the series.

The experimental setup shown in Figure 14 was used to investigate the tribocharging characteristics of PP particles colliding with surfaces made of PVC and polyethylene (PE). The conveying pipes were made of copper except at the test section where pipes made of PVC or covered with a PE film were used. Figure 14 shows the detailed cross section of the test section following the design used by Matsusaka et al.20 The various layers of materials consisted of PVC (5 mm)/ PE (0.04 mm), conductive adhesive (0.07 mm), copper sheet (0.10 mm) as electrode, nonconductive adhesive (0.12 mm), and copper sheet (0.10 mm) as electric shield. The copper electrode was connected to the electrometer (R8252), and the electric shield was grounded. The whole conveying system was well grounded so that the particle-charging conditions were kept constant during their impacts with the metal wall. When the air pressure and flow rate were adjusted to 75 psi and 1600 L/min, respectively, particles passed through the horizontal pipe, continuously at a rate of 44.4 ( 4.0 g/s. Electrostatic charge was generated at the test section, and the induced current was measured using the electrometer (R8252) and was stored in a computer at intervals of 0.1 s. The charge density of the particles was measured by a Faraday cage (TR8031). All particles were collected in a metal container and grounded for discharging at the end of each experimental run. Figure 15 shows the induced current acquired by the film. The induced current caused by the sliding of a single particle over the film was in the form of a pulse, while that that measured by the electrometer represented the sum of numerous such pulses to form a

7194


charge density measured with the Faraday cage was also found to be higher in the former case than in the latter (Figure 17). To characterize the extent of mechanical interactions between the PP particles and the inner surface of the pipe during a typical conveying process, the PE film was examined using a scanning electron microscope after being used for different lengths of time. Figure 18 shows micrographs of the PE film surface magnified 1000 times after being used for 0, 2, and 10 min, respectively. It is evident that the quality of the film surface deteriorates very quickly during usage with deep gorges forming after 10 min. This confirms the presence of strong sliding effects and frictional forces between particles and the pipe wall. As shown in the sections 3.1.1 and 3.2.1, more particles tend to slide on the pipe wall with the air flow rate decreasing, either at the vertical pipe (Figures 2-4) or at the horizontal pipe (Figure 9). Henthforth, it is reasonable to believe that, for the three granular flows in the pneumatic conveying system, the electrostatic charge generation mechanism of particles and pipe wall mainly depends on the tribroelectrification. 5. Factors Affecting Electrostatics of the Granular Flow

Figure 18. Scanning electron micrographs of PE film (thickness: 0.04 mm) (experimental setup B with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s): (a) fresh film; (b) used for 2 min; (c) used for 10 min.

continuous current due to the motion of a large number of particles. The absolute values of the negative component of the induced current sampled at a rate of 200 s-1 are compared in Figure 15c. It can be seen that the current induced on the PVC wall is higher than that on the PE film. Figure 16 shows that the total accumulated charge for both materials obtained by time integration of Figure 15c increases at a constant rate with respect to time. However, more charges were generated by interactions between the PP particles with the PVC pipe wall than with the PE film. The particle

5.1. Film Material. Figure 19a,b shows that the magnitude of fluctuations of the induced current on the PVC pipe surface ((2.5 × 10-7 A) is larger than that on the PE film surface ((1.0 × 10-7 A). The time-integrated value of the induced current representing the total amount of charge accumulated on the surface of the materials is also larger for the PVC than the PE surface (Figure 20). Thus, the material used for the pipes in a conveying system would be an important factor affecting the charge generation characteristics of the system. 5.2. Humidity. The RH of the ambient air is an important factor to be considered in any electrostatics experiment.4 In the present work, both natural air from the atmosphere with a RH of about 70% and air dried to an RH value of about 5% by passing through a silica gel column (Figure 14) were used. Figure 21 shows that particle charge densities were lower at high humidity levels and vice versa. There were also more fluctuations in both the particle charge density values and the magnitudes of the induced current on the pipe wall at RH ) 5%. As seen previously in Figure 5a, induced currents fluctuated in the range (2.0 × 10-7 A when the RH was 5%, while the present investigation using ambient air resulted in a fluctuation range of (1.0 × 10-7 A (Figure 22a). The absolute values of the induced current (Figure 22b) as well as the total accumulated charge (Figure 23) showed similar trends. In other words, high RH of the conveying gas has a dampening effect on electrostatic charge generation. This claim is also supported by the fact that no half-ring or ring structure was visually observed when natural ambient air was used as the conveying medium, regardless of the flow rate used. 5.3. Electrostatics of the Granular Mixture. The effects of particle material on the electrostatics behavior of the system were investigated by comparing the behavior of the original PP particles with a mixture containing 3 wt % of glass beads with mean diameters of 3.00 mm and PP particles. The hydrodynamic behavior of the mixture of particles was visually verified to be similar to that of the pure PP particles. At flow rates


Figure 19. Induced current acquired for the films (experimental setup A, disperse flow with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s) at the vertical pipe: (a) PVC; (b) PE; (c) negative value (absolute): PVC versus PE.

Figure 20. Charges obtained by integration of induced currents (disperse flow with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s) at the vertical pipe (shown in Figure 19a,b): PVC versus PE.

Figure 21. Particle charge density (determined using Faraday cage) of the half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s): RH ) 5% versus RH ) 70%.

7196


Figure 23. Charges obtained by integration of induced current at the vertical pipe: RH ) 5% (Figure 5b) versus RH ) 70% (Figure 22).

Figure 22. Induced current at the vertical pipe (half-ring flow with air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s): (a) RH ) 70%; (b) RH ) 70% versus RH ) 5%.

of 1600 and 1000 L/min, particle charge densities were found to be larger for the mixture than the pure sample (Figure 24). MPCT measurements presented in Figure 25 where each data point was obtained by time integration of 400 MPCT data over a time interval of 20 s also revealed that the induced current produced by the mixture was larger. Therefore, it seems that the electrostatic properties of a pneumatic conveying system can be altered by changing the composition of the granular material being transported. This has been used by Wolny and Opalinski,8 who added a small proportion (0.1 vol %) of fines to a bulk of polystyrene beads (size: 1.02-1.2 mm) to neutralize electrical charges and improve hydrodynamic and heat-transfer properties of the granular material. 5.4. Antistatic Agent. The ability to effectively control and reduce electrostatic effects in a pneumatic conveying system is an important yet poorly understood research area. A commercially available antistatic agent, Larostat-519 powder (Zhang et al.2), may be a suitable means to achieve this purpose. This is a noninflammable white powder with bulk density of 520 kg/m3, composed of about 60% soyadimethylethylam-

Figure 24. Particle charge density (determined using Faraday cage) of pure particles (PP) versus mixtures (PP and glass beads by weight of 3%): (a) disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); (b) half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s).


Figure 25. Charges obtained by integration of 400 MPCT values within a time interval of 20 s of pure particles (PP) versus mixtures (PP and glass beads by weight of 3%) at the horizontal pipe: (a) disperse flow (air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s); (b) half-ring flow (air flow rate 1000 L/min, solids feed rate 13.8 ( 2.4 kg/m2 s).

monium and 40% ethasulfate amorphous silica. In this section, the effectiveness of this antistatic agent was characterized. In comparison with previous cases where no such agent was used, the induced current measured in the presence of the Larostat-519 powder was observed to be smaller by about 3 orders of magnitude (Figure 26). The same drastic drop in magnitude was also observed in the accumulated charge by comparing Figures 12 and 27. In terms of flow patterns, the Larostat-519 powder has the effect of suppressing the formation of half-ring and ring structures in the vertical pipe segment but not in the horizontal segment. This may be an indication that these structures formed in the horizontal pipe not only as a result of electrostatic effects and other factors may also be contributing to their appearance. Figure 28 shows the scanning electron micrograph of the surface of a PP particle before and after mixing with

Figure 26. Induced current of the antistatic disperse flow measured at the vertical pipe (air flow rate 1600 Ls/min, solids feed rate 35.3 ( 3.2 kg/m2 s): (a) temporal evolution of induced current; (b) absolute values of positive versus negative values.

the Larostat-519 powder. It is seen that the particle surface becomes coated with a thin film of powder after mixing. Here, a parallel can be drawn between a PP particle whose surface is covered with a film of Larostat519 powder colliding with the pipe wall and a clean particle colliding with the wall covered by a film made of a different material. It has been established previously that a pipe wall covered with a PE film leads to very different electrostatics phenomena, and so a similar kind of observation is expected with the use of the Larostat-519 powder. The almost similar readings obtained from the MPCT for the cases where the powder was either absent or present in the system may be due to the inherent inaccuracies of the instrument in detecting low current signals (Figure 29). 6. Conclusions The present work can be summarized as follows. First, air flow rate is the main element determining the

7198


Figure 27. Charges obtained by integration of the induced current (shown in Figure 26a).

Figure 29. Charges obtained by integration of 400 MPCT values within a time interval of 20 s: clean granular flow versus antistatic granular flow (disperse flow at the vertical pipe with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s).

ring and ring in the vertical conveying pipe. Second, electrostatic effects increase with time. The charge accumulated at the pipe wall increases with time, and the rate of increase seems constant for each of the three types of flow. Particle charge density also increases with time, and this may account for clustering behavior occurring at the vertical pipe wall even when a high air flow rate is used and the dominant flow regime is that of disperse flow. Third, pipe wall material has an obvious effect on the electrostatics of the granular flow. The various parameters used throughout this work to characterize electrostatic effects such as the induced current and accumulated charge were shown to be higher when PVC pipe was used without a PE film attached to the inner wall in the conveying system. Electrostatic effects were also dependent on particle composition as shown in the experiments done using PP particles mixed with glass beads. The commercially available antistatic agent, Larostat-519 powder, was found to reduce electrostatic effects within the system effectively. Air with high RH also tends to reduce electrostatic effects when used for pneumatic conveying of particles. No half-ring and ring flow patterns attributable to strong electrostatic effects were observed when ambient air with a RH of about 70% was used in the system. Besides, it is clear that the mechanism of electrostatic charge generation for the granular flow in the pneumatic conveying system mainly depends on tribroelectrification due to the strong force effect on the surface when the particles slide on the pipe wall. Acknowledgment

Figure 28. Scanning electron micrographs of PP particles in the pneumatic conveying system (disperse flow with air flow rate 1600 L/min, solids feed rate 35.3 ( 3.2 kg/m2 s): (a) clean particle; (b) mixed with Larostat-519 powder.

electrostatic behavior of granular flow. The lower is the air flow rate, the higher are the induced current and particle charge density. These in turn lead to particle clustering and the formation of such structures as half-

This project is supported by the National University of Singapore under Grant R279-000-095-112. We are grateful to Professor Sankaran Sundaresan and Dr. Kewu Zhu for many helpful discussions on the project. We also extend our sincere thanks to Eldin Wee Chuan Lim for his assistance in the preparation of this paper. Literature Cited (1) Masuda, H.; Komatsu, T.; Iinoya, K. The static electrification of particles in gas-solid pipe flow. AIChE J. 1976, 22, 558-564.

Ind. Eng. Chem. Res., Vol. 43, No. 22, 2004 7199 (2) Zhang, Y. F.; Yang, Y.; Arastoopour, H. Electrostatic effect on the flow behavior of a dilute gas/cohesive particle flow system. AIChE J. 1996, 42, 1590-1599. (3) Matsusaka, S.; Masuda, H. Electrostatics of particles. Adv. Powder Technol. 2003, 14, 143-166. (4) Nieh, S.; Nguyen, T. Effects of humidity, conveying velocity, and particle size on electrostatic charges of glass beads in a gaseous suspension flow. J. Electrostat. 1988, 21, 99-114. (5) Kanazawa, S.; Ohkubo, T.; Nomoto, Y.; Adachi, T. J. Electrostat. 1995, 35, 47-54. (6) Diu, C. K.; Yu, C. P. Deposition from charged aerosol flows through a two-dimensional bend. J. Aerosol Sci. 1979, 11, 391395. (7) Al-Adel, M. F.; Savile, D. A.; Sundaresan, S. The effect of static electrification on gas-solid flows in vertical risers. Ind. Eng. Chem. Res. 2002, 41, 6224-6234. (8) Wolny, A.; Opalinski, I. Electric charge neutralization by addition of fines to a fluidized bed composed of coarse dielectric particles. J. Electrostat. 1983, 14, 179-289. (9) Chang, H.; Louge, M. Fluid dynamic similarity of circulating fluidized beds. Powder Technol. 1992, 70, 259-270. (10) Smeltzer, E. E.; Weaver, M. L.; Klinzing, G. E. Individual electrostatic particle interaction in pneumatic transport. Powder Technol. 1982, 33, 31-42. (11) Gajewski, A. Measuring the charging tendency of polystyrene particles in pneumatic conveyance. J. Electrostat. 1989, 23, 55-66. (12) Joseph, S.; Klinzing, G. E. Vertical gas-solid transition flow with electrostatics. Powder Technol. 1983, 36, 79-87.

(13) Tsuji, Y.; Morikawa, Y. Flow pattern and pressure fluctuation in air-solid two-phase flow in a pipe at low air velocities. Int. J. Multiphase Flow 1982, 8, 329-341. (14) Rao, M. S.; Zhu, K. W.; Wang, C. H.; Sundaresan, S. Electrical capacitance tomography measurements on the pneumatic conveying of solids. Ind. Eng. Chem. Res. 2001, 40, 4216-4226. (15) Zhu, K. W.; Rao, S. M.; Wang, C. H.; Sundaresan, S. Electrical capacitance tomography measurements on vertical and inclined pneumatic conveying of granular solids. Chem. Eng. Sci. 2003, 58, 4225-4245. (16) Zhu, K. W.; Rao, S. M.; Huang, Q. H.; Wang, C. H.; Matsusaka, S.; Masuda, H. On the Electrostatics of Pneumatic Conveying of Granular Materials Using Electrical Capacitance Tomography. Chem. Eng. Sci. 2004, 59, 3201-3213. (17) Sommerfeld, M. Analysis of collision effects for turbulent gas-particle flow in a horizontal channel: Part I. Particle transport. Int. J. Multiphase Flow 2003, 29, 675-699. (18) Yu, C. P. Precipitation of unipolarly charged particles in cylindrical and spherical vessels. J. Aerosol Sci. 1977, 8, 237-241. (19) Fan, J. R.; Yao, J.; Cen, K. F. Antierosion in a 90° bend by particle impaction. AIChE J. 2002, 48, 1401-1412. (20) Matsusaka, S.; Nishida, T.; Gotoh, Y.; Masuda, H. Electrification of fine particles by impact on a polymer film target. Adv. Powder Technol. 2003, 14, 127-138.

Received for review April 26, 2004 Revised manuscript received August 4, 2004 Accepted August 11, 2004 IE049661L

Electrostatics of the Granular Flow in a Pneumatic Conveying System

Recommend Documents