Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effect of operating parameters and particle properties on electrostatics in gas-solid fluidized beds Moughrabiah, Wajeeh O. 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2009_fall_moughrabiah_wajeeh.pdf [ 26.01MB ]
Metadata
JSON: 24-1.0058972.json
JSON-LD: 24-1.0058972-ld.json
RDF/XML (Pretty): 24-1.0058972-rdf.xml
RDF/JSON: 24-1.0058972-rdf.json
Turtle: 24-1.0058972-turtle.txt
N-Triples: 24-1.0058972-rdf-ntriples.txt
Original Record: 24-1.0058972-source.json
Full Text
24-1.0058972-fulltext.txt
Citation
24-1.0058972.ris

Full Text

EFFECT OF OPERATING PARAMETERS AND PARTICLE PROPERTIES ON ELECTROSTATICS IN GAS-SOLID FLUIDIZED BEDS  by  Wajeeh O. Moughrabiah  B.Sc., King Saud University, Al-Riyadh, Saudi Arabia, 1993 M.A.Sc., King Saud University, Al-Riyadh, Saudi Arabia, 2002  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Chemical and Biological Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) September, 2009  © Wajeeh O. Moughrabiah, 2009  ABSTRACT The influences of operating pressure, temperature and gas velocity on electrostatics in a fluidized bed of glass beads and different grades of polyethylene resin were investigated in a fluidization column of 150 mm inner diameter and 2.0 m height. Eight collision probes at different levels and radial positions measured the electrostatics in the bed. The electrostatics increased as pressure increased, probably due to an increase in bubble rise velocity, frequency and volume fraction. As the pressure increased, particle-particle and particle-wall collisions near the distributor and wall contributed heavily to charge generation. Temperature also played a role. At higher temperatures (up to 90°C), the polarity of net cumulative charge in the bed reversed. As the superficial gas velocity increased, the electrostatics increased. However, at higher gas velocities, the polarity in the freeboard was opposite to that in the bed, indicating that fines entrained from the column carried charges, resulting in a net charge of opposite polarity to that inside the bed. For Geldart group B particles the degree of electrification in the bed slightly increased with decreasing particle size. Charging for group A particles was significantly greater than for group B particles. For binary mixtures of group A and B particles the electrostatics increased as the proportion of small particles increased. As the relative humidity (RH) of fluidizing air increased, the electrostatics decreased. For the RH range (5-30%) explored, the sensitivity of the charging to RH varied significantly depending on the location of the probes. As the proportion of fine glass beads (<30 µm) increased to 2.0 wt% in a fluidized bed of large glass beads (574 µm), the electrostatics in the bed decreased, likely because the fines acted as spacers between larger particles. The electrostatics decreased as the proportion of an antistatic agent (Larostat) increased from 0.0 to 0.5 wt%, because Larostat tends to adsorb moisture and attach to the surface of the glass beads, consequently enhancing their surface conductivity. However, the degree of electrification increased when the wt% of Larostat exceeded 1.0%, likely due to the tendency of Larostat particles to adsorb water and to agglomerate. ii  TABLE OF CONTENTS ABSTRACT .............................................................................................................................ii TABLE OF CONTENTS ...................................................................................................... iii LIST OF TABLES............................................................................................................... viii LIST OF FIGURES.................................................................................................................x NOMENCLATURE .............................................................................................................xxi ACKNOWLEDGEMENTS .............................................................................................. xxiii DEDICATION .....................................................................................................................xxv Chapter 1: 1.1  1.2  Charge Generation .............................…………………….………………….… ....2 1.1.1  Triboelectrification ....…………………………………………………..... ....2  1.1.2  Frictional charging ……………………………………………………..........6  1.1.3  Thermionic emission …………………………………………….……..........7  1.1.4  Ion collision ……………………………………….……………………... ....8  Fluidization Flow Regimes ......................................................................................8 1.2.1  1.3  1.4  1.5  Introduction …..………............................................................................. ....1  Bubbling flow regime ……………………………………….……............ ....8  Electrostatics in Bubbling Fluidized Beds ……………………….....……..............9 1.3.1  Influence of bubble behaviour ………………………………………….... ....9  1.3.2  Influence of fluidizing gas velocity ………………………….……........... ..11  1.3.3  Influence of particle properties ……………………………………….........11  Electrostatics Measurements and Instrumentation …………….………………....12 1.4.1  Faraday cup …………………………………………………………...........13  1.4.2  Electrostatic probes …………………………………………………...........14  1.4.2.1  Capacitance probes ……………………………….…………........... ..15  1.4.2.2  Induction probes …………………….…………………………..........16  1.4.2.3  Collision probes ………………………………………………......... ..17  Influence of Operating Parameters ……………….………………………........ ..19 1.5.1  Pressure ……………………………………………………………….........19  1.5.1.1  Influence of elevated pressures on gas bubble behaviour ……............19 iii  1.5.1.2 1.5.2 1.6  Influence of elevated pressures on Umf ………………………......... ..20  Temperature …………………………………………………………..........23  Electrostatic Charge Reduction Techniques ………………………………..........24 1.6.1  Gas humidification and ionization ……………………………………........24  1.6.2  Addition of fines, antistatic agents and more conductive particles ............ ..26  1.7  Thesis Objectives ……………………………………………………...................26  1.8  Thesis Outline ........................................................................................................27  Chapter 2:  Experimental Equipment and Approach .........................................….....29  2.1  Elevated-Pressure Fluidization Column ...……..……………………....….……...29  2.2  Measurements and Instrumentation …………........………………………….......31 2.2.1  Electrostatic charge measurement techniques ………..…….….....…….... ..32  2.2.2  Operating variables measurement and control ....……………………....... ..34  2.3  High-Pressure Air System ........…………………………………………………..35  2.4  Bed Materials .......………………………………………………………………..36  2.5  2.4.1  Glass beads (GB) .......……………………………………….…………… ..37  2.4.2  Polyethylene (PE) ………………………………...…...…………….…… ..38  2.4.3  Added fines ………………………………………………….………..........41  Experimental Approach ……………………………………...…...…...…..…… ..41  Chapter 3: 3.1  3.2  Results and Discussion: Effect of Operating Parameters .......................46  Preliminary Experiments ...…...……………………...……………………….... ..46 3.1.1  Charge transfer and charge induction ……….………...……...……...…... ..47  3.1.2  Magnitude of measured charge .....................................................................48  3.1.3  Different charge polarity ...............................................................................50  3.1.4  Probe sensitivity ............................................................................................52  Effect of Operating Parameters ..............................................................................54 3.2.1  Effect of operating pressure ..........................................................................54  3.2.1.1  Glass beads ...........................................................................................54  3.2.1.2  Polyethylene resin ................................................................................63  3.2.2  Effect of operating temperature .................................................................. ..66 iv  3.2.3  3.3  3.2.3.1  Glass beads …………………………………………………………...75  3.2.3.2  Polyethylene resin …………………………………………………....82  Summary ………………………………………………………………………....87  Chapter 4: 4.1  Effect of superficial gas velocity ................................................................ ..75  Results and Discussion: Effect of Particles Properties …………............88  Effect of Average Particle Size ...…………………………………………...........88 4.1.1  Mono-size glass beads ................................................................................ ..88  4.1.2  Binary mixtures of glass beads …………………………………………... ..92  4.1.2.1  Minimum fluidization velocity of a binary mixture ……………….. 101  4.2  Bipolar Charging ………………………………………………………………..104  4.3  Effect of Particle Density and Type …………………………………………….106  4.4  Effect of Particle Chemical Composition ……………………………………… 110  4.5  Summary ………………………………………………………………………..113  Chapter 5:  Electrostatic Charges Reduction Techniques  ………………………... 115  5.1  Humidification ………………………………………………………………….115  5.2  Addition of Fines and Antistatic Agents ………………………………………..120  5.3  Summary ……………………………………………………………………………133  Chapter 6:  Conclusions and Recommendations ……………………………………135  6.1  Conclusions ......……………………………………………………………….... 135  6.2  Recommendations for Future Work …………………………………………….140  Literature Cited ………………………………………………………………...……….141 Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds …………………………………………………………………………... 147 Appendix B: Equipment Photographs and Engineering Drawings ...…………..…...154 B.1 Experimental Apparatus ………………………………………………………...154 v  B.2 KAESER Air Compressor ……………………………………………………... 155 B.3 Fluidization Section ……………………………………………………………. 156 B.4 Sight Glass ……………………………………………………………………... 157 B.5 Control Valves …………………………………………………………………. 158 B.6 Preliminary Experimental Set-up ……………………………………………….159 B.7 Elevated-Pressure Fluidization Column ……………………………………….. 160 Appendix C: Electrostatic Probes Photographs ………………………………………165 C.1 Collision Ball Probes ………......…………………………………………..........165 C.1.1 Collision ball probe parts ..................……….………………………..........165 C.1.2 Collision ball probe positions ……………………………………............ 167 C.2 Faraday Cup …………………………………………………………….............168 Appendix D: Particle Properties ……………………………………………………….169 D.1 Calculation of Minimum Fluidization Velocity of GB-S …………………........ 169 D.2 Particle Size Distribution of Glass Beads ...............……..……………………...170 D.2.1 Particle size distribution of GB-S ………………………………………... 170 D.2.2 Particle size distribution of GB-M ………………………………………..171 D.2.3 Particle size distribution of GB-L ………………………………………...172 D.2.4 Particle size distribution of GB-XL ………………………………............173 D.3 Particles Bulk Density Measurements …………………………………………. 174 D.4 Particle Size Distribution of Polyethylene Particles ……………………………177 D.4.1 Particle size distribution of HDPE ……...…….…………………………..177 D.4.2 Particle size distribution of LLDPE …….………..……………………….178 D.4.3 Particle size distribution of PE-SLD …….………..………………………179 D.4.4 Particle size distribution of PE-X1 ………………………………………. 180 D.4.5 Particle size distribution of PE-X2 ………………………………………. 181 D.5 Particle Size Distribution of Glass Bead Binary Mixtures …………………….. 182 D.5.1 Size distribution of M1 particles ………………………………….............182 D.5.2 Size distribution of M2 particles .................................................................183 D.5.3 Size distribution of M3 particles .................................................................184 vi  D.5.4 Size distribution of M4 particles .................................................................185 D.5.5 Size distribution of M5 particles .................................................................186 D.5.6 Size distribution of M6 particles .................................................................187 Appendix E: Influence of Elevated Pressures on Bubble Behaviour ………………..188 E.1 Influence of Elevated Pressures on Bubble Behaviour …………………………188 Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds ......................190 F.1 Particle-Fluid Interactions (PFIs) ………………………………………............ 190 F.1.1 Drag force ………………………………………………………………... 190 F.1.1.1 Stokes’s law regime …..…………………………………………….191 F.1.1.2 Intermediate regime ………………………………………………... 192 F.1.1.3 Newton’s low regime ……………………………………………….192 F.1.1.4 Empirical drag coefficient expression ……………………………... 193 F.1.1.5 Terminal velocity of particles, Ut ………………………………….. 196 F.1.2 Basset force ………………………..………………………………………197 F.1.3 Lift force …………………………………………………………………. 198 F.1.3.1 Magnus force ……..…..……………………………………………. 198 F.1.3.2 Saffman force ……..………………………………………………...199 F.2 Particle-Particle Interactions (PPIs) …………………………………………….200 F.2.1 Van der Waals forces ……………………………………………………..200 F.2.2 Electrostatics forces ……………………………………………………… 201 F.2.2.1 Triboelectrification ………………………………………………… 202 F.2.2.2 Frictional Charging ……..…………………………………………..202 F.2.3 Collision forces …………………………………………………………...203 F.3 External Field Forces (EFFs) …………………………………………………...206  Appendix G: Estimated Mass Ratio of Larostat in Binary Mixtures .........................210 G.1 Estimated Mass Ratio of Small-to-Large Particles in Binary Mixture for Full Mono-Layer Coverage …….………………………………………....................210  vii  LIST OF TABLES Table 1.1  Polymer work function series of polymers in two studies (adapted from Cross, 1987) ……………………………………………………………….....4  Table 1.2  Triboelectric series (adapted from Cross, 1987) ………………………….....5  Table 2.1  Properties of four types of glass beads (ambient conditions) ........................37  Table 2.2  Properties of polyethylene particles (ambient conditions) ………………....38  Table 2.3  Properties of fine particles added during this study ……..............................41  Table 4.1  Specific surface area, As, of mono-disperse glass beads …………………...90  Table 4.2  Properties of binary particle mixtures. All particle densities are 2500 kg/m3 ............................................................................................................ ..93  Table 4.3  Current flow through ball probes at different axial and radial positions in beds of LLDPE, PE-X1 and PE-X2 particles, dP = 487, 523 and 502 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U- U(mf)M = 0.05 m/s. For probe positions, see Figure 2.9 ……………………………. 112  Table A.1  Summary of previous studies on electrostatics in gas-solid fluidized beds ..149  Table D.2.1  Physical properties of glass beads (GB-S) ..............….…………………....170  Table D.2.2  Physical properties of glass beads (GB-M) ….………………….….……. 171  Table D.2.3  Physical properties of glass beads (GB-L) ..............……….……………....172  Table D.2.4  Physical properties of glass beads (GB-XL) ...............…………………....173  Table D.3.1  Bulk density measurement for (GB-S) using a known-volume container …174  Table D.3.2  Bulk density measurement for (GB-M) using a known-volume container ... 174  Table D.3.3  Bulk density measurement for (GB-L) using a known-volume container …174  Table D.3.4  Bulk density measurement for (GB-XL) using a known-volume container 175  Table D.3.5  Bulk density measurement for (HDPE) using a known-volume container …175  Table D.3.6  Bulk density measurement for (LLDPE) using a known-volume container 175  Table D.3.7  Bulk density measurement for (PE-SLD) using a known-volume container 176 viii  Table D.3.8  Bulk density measurement for (PE-X1) using a known-volume container 176  Table D.3.9  Bulk density measurement for (PE-X2) using a known-volume container 176  Table D.4.1  Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of HDPE particles ……………………………………………………………………177  Table D.4.2  Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of LLDPE particles ……………………………………………………………………178  Table D.4.3  Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of PE-SLD particles …………………………………………………………………... 179  Table D.4.4  Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of PE-X1 particles ……………………………………………………………………180  Table D.4.5  Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of PE-X2 particles ……………………………………………………………………181  Table D.5.1  Physical properties of M1 particles ……………………………………….182  Table D.5.2  Physical properties of M2 particles ……………………………………….183  Table D.5.3  Physical properties of M3 particles ……………………………………….184  Table D.5.4  Physical properties of M4 particles ……………………………………….185  Table D.5.5  Physical properties of M5 particles ……………………………………….186  Table D.5.6  Physical properties of M6 particles ……………………………………….187  Table F.1  Recommended empirical drag coefficient correlations for different ranges of Rep. (adapted from Clift et al., 1978) …………………………………. 194  ix  LIST OF FIGURES Figure 1.1  Closed triboelectric series (adapted from Harper, 1967) ……………………6  Figure 1.2  Charging current plotted against velocity for a polyethylene terephthalate disc rubbed against a metal brush (adapted from Zimmer, 1970) ……….. …7  Figure 1.3  Gas-solid flow regimes as function of fluid velocity (adapted from Grace, 1986) ……………………………………………………………………... …9  Figure 1.4  Faraday cup (adapted from Cross, 1987) …………………………………..13  Figure 1.5  Schematic of Faraday cup fluidization column (adapted from Mehrani, 2005) ……………………....……………………………………………... ..14  Figure 1.6  Capacitance electrostatic probes: (a) (adapted from Guardiola et al. 1992), (b) (adapted from Wolny and Kazmierczak, 1989) ……….………………..15  Figure 1.7  Induction probe with shielded tip (adapted from Boland and Geldart, 1971) ……………………....……………………………………………... ..16  Figure 1.8  Collision electrostatic probes: (a) (adapted from Ciborowski and Wlodarski, 1962), (b) (adapted from Fujino et al. 1985) …………………..18  Figure 1.9  Collision ball probe (adapted from Park et al., 2002b) ………………….…..18  Figure 1.10  Deterioration in measured potential due to adhesion of charged particles on probe tip (adapted from Fujino et al., 1985) ………………………….. ..19  Figure 1.11  Effect of pressure on Umf (adapted from Rowe, 1984) ………………….....23  Figure 1.12  Effect of temperature on Umf (adapted from Yates, 2003) ………………....24  Figure 2.1  Schematic of elevated-pressure column (All dimensions are in m) ………..30  Figure 2.2  Schematic of carbon steel base, which supports the fluidization column .....31  Figure 2.3  Schematic of distributor plates, Nor = 50 orifices, open area = 3.8 % …… ..32  Figure 2.4  Schematic of electrostatic collision ball probe ………....................................33  Figure 2.5  Schematic of Faraday cup …………….........................................................34  Figure 2.6  Schematic of overall layout of elevated-pressure fluidization unit PR: pressure regulator, PVC: pressure control valve, FCV: flow control valve ..36  Figure 2.7  SEM images of polyethylene particles …………………………...................40 x  Figure 2.8  SEM images of fine particles added during this study ................................42  Figure 2.9  Locations of electrostatic collision ball probes ........................................... ..43  Figure 3.1  Schematic of simplified preliminary experimental set-up. (All dimensions are in m.) ……………………………...…………………………………… ..46  Figure 3.2  Ball probe experimental results for a PE tube rubbed against cotton cloth: (a) without contact between the charged object and the probe; (b) with contact …………………..………………………………………………... ..48  Figure 3.3  Faraday cup experimental results for a PE tube rubbed against cotton cloth: (a) without contact between the charged object and Faraday cup; (b) with contact ……………………………………………………………… ..49  Figure 3.4  Ball probe experimental results when: (a) glass tube was rubbed against a cotton cloth; (b) glass tube was rubbed against a PE tube …..................... ..50  Figure 3.5  Faraday cup experimental results when: (a) glass tube was rubbed against a cotton cloth; (b) glass tube was rubbed against a PE tube ...………….... ..50  Figure 3.6  Ball probes experimental results when: (a) plastic ruler was rubbed against hair; (b) ruler was rubbed against cotton …………………………………...51  Figure 3.7  Faraday cup experimental results when: (a) plastic ruler was rubbed against hair; (b) ruler was rubbed against cotton ………………………… ..51  Figure 3.8  Schematic of simple experimental set-up for probe sensitivity experiments. (All dimensions are in m.) …………………………………....52  Figure 3.9  Experimental results for ball probes and Faraday cup obtained by traversing a PE tube which had been rubbed against a cotton cloth, at different probe sensitivities, PS, and different distances from the tip of the probe, λ,: (a) PS = 5 pC/MU and λ = 0.025 m; (b) PS = pC/MU and λ = 0.038 m; (c) PS = 3 pC/MU and λ = 0.025 m; (d) PS = 3 pC/MU and λ = 0.038 m; (e) PS = 1 pC/MU and λ = 0.025 m; (f) PS = 1 pC/MU and λ = 0.038 m ……………………………………............................................... ..53  Figure 3.10  Effect of operating pressure on net cumulative charge as a function of time at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves are absolute pressures (in kPa) at top of column. For probe positions, see Figure 2.9 ...............................................................................55  Figure 3.11  Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure2.9 ..................................................................................................... ..57 xi  Figure 3.12  Charge fluctuations obtained by filtering net cumulative charge curves for probe A versus time using a high-pass filter of cut-off frequency of 10*(1/period) where period is the data sampling range in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves are absolute pressures (in kPa) at top of column. For position of probe A, see Figure 2.9 ………………………………………. ..58  Figure 3.13  Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 ………………………………………………….. ..59  Figure 3.14  Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s, with probe locations interchanged from those in Figure 3.11 ...……………………………….. ..60  Figure 3.15  Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s, with probe locations interchanged from those in Figure 3.13 ………………………... ..60  Figure 3.16  Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of GB-M particles, dp = 365 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 .................................................................................................... ..61  Figure 3.17  Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of GB-M particles, dp = 365 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 ................................................................................61  Figure 3.18  Standard deviation of differential pressure fluctuations across a bed of GB-M particles as a function of pressure, dp = 365 µm,T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s ………..……………….…………………. ..62  Figure 3.19  Effect of pressure on net cumulative charge as a function of time at different axial and radial positions in a bed of HDPE particles, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on the curves are absolute pressures (in kPa) at top of column. For probe positions, see Figure 2.9 ………………………………….…..................... ..64  Figure 3.20  Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of HDPE particles, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 .................................................................................................... ..65  xii  Figure 3.21  Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of HDPE particles, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 ................................................................................65  Figure 3.22  Charge density of HDPE particles measured by Faraday cup, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17%, and U-Umf = 0.05 m/s, particles sampled from ports A, B and C at z = 0.15, 0.31 and 0.55 m, respectively .............. ..66  Figure 3.23  Effect of temperature on net cumulative charge as a function of time for PE-SLD particles, dp = 642 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. Numbers on the curves are bed temperature (in ºC). For probe positions, see Figure 2.9 .............................................................................. ..68  Figure 3.24  Current flow through ball probes as a function of bed temperature at different axial and radial positions in a bed of PE-SLD particles, dp = 642 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 .............................................................................................. ..69  Figure 3.25  Standard deviation of charge fluctuations as a function of temperature at different axial and radial positions in a bed of PE-SLD particles, dp = 642 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 .............................................................................................. ..69  Figure 3.26  Output signals of ball probes as a function of time and temperature in an empty column with air flow rate at 0.15 m/s and temperature range of (a) from ambient to 70ºC and (b) from 70ºC to 32ºC. For probe positions, see Figure 2.9 .................................................................................................... ..70  Figure 3.27  Effect of temperature on net cumulative charge as a function of time when the bed temperature was increased from ambient to 90ºC for GB-XL particles, dp = 693 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. Numbers on the curves are bed temperature (in ºC). For probe positions, see Figure 2.9 .............................................................................................. ..71  Figure 3.28  Effect of temperature on net cumulative charge as a function of time when the bed temperature was cooled from 90ºC to ambient for GB-XL particles, dp = 693 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. Numbers on the curves are bed temperature (in ºC). For probe positions, see Figure 2.9 .............................................................................................. ..72  Figure 3.29  Effect of temperature on net cumulative charge as a function of time and temperature when the bed temperature is raised from ambient to 70ºC and then allowed to cool to ambient again for GB-XL particles, dp = 693 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 .................................................................................................... ..73  xiii  Figure 3.30  Types of behaviours observed for build-up of charge density as a function of time (adapted from Coste and Pechery, 1977) ….…………………….. ..74  Figure 3.31  Effect of superficial gas velocity on electrostatic charges buildup in a bed of GB-XL particles as a function of time at different axial and radial positions, dp= 693 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote U-Umf (in m/s). For probe positions, see Figure 2.9 ..................................................................................................................76  Figure 3.32  Effect of superficial gas velocity on electrostatic charges buildup in a bed of GB-M particles as a function of time at different axial and radial positions, dp= 365 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote U-Umf (in m/s). For probe positions, see Figure 2.9 ..................................................................................................................77  Figure 3.33  Current flow through ball probes as a function of excess gas velocity, at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9 .............................................................................................. ..78  Figure 3.34  Standard deviation of charge fluctuations as a function of excess gas velocity, at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9 ................................................................... ..78  Figure 3.35  Current flow through ball probes as a function of excess gas velocity at different axial and radial positions for GB-M particles, dp = 365 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9 ..................................................................................................................79  Figure 3.36  Standard deviation of charge fluctuations as a function of excess gas velocity at different axial and radial positions for GB-M particles, dp = 365 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9 .............................................................................................. ..79  Figure 3.37  Cumulative charge as a function of time in an empty column with air flow rate at (a) 0.35 m/s and (b) 0.4 m/s .............................................................. ..80  Figure 3.38  Effect of superficial gas velocity on electrostatic charges buildup in a bed of HDPE particles as a function of time at different axial and radial positions, dp= 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote U-Umf (in m/s). For probe positions, see Figure 2.9 ..................................................................................................................83  Figure 3.39  Current flow through ball probes as a function of superficial gas velocity, at different axial and radial positions for HDPE particles, dp = 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9 ..................................................................................................................84 xiv  Figure 3.40  Standard deviation of charge fluctuations as a function of superficial gas velocity, at different axial and radial positions for HDPE particles, dp = 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9 …………………………….……………………....84  Figure 3.41  Electrostatic charges buildup with time in the freeboard at different superficial gas velocities in a bed of HDPE particles, dp= 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote U-Umf (in m/s). For position of probe D, see Figure 2.9 …………………………..85  Figure 3.42  Current flow through ball probes in the freeboard in a bed of HDPE particles as a function of superficial gas velocity, dp = 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For position of probe D, see Figure 2.9 ……………………………………………………………………….... ..86  Figure 3.43  Standard deviation of charge fluctuations as a function of superficial gas velocity, in the freeboard in a bed of HDPE particles, dp = 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For position of probe D, see Figure 2.9 ………………………………………………………………… ..86  Figure 4.1  Effect of average particle diameter on net cumulative charge as a function of time at different axial and radial positions in beds of GB-XL, GB-L, GB-M and GB-S, dp = 693, 574, 365, 65 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.08 m/s. Letters on curves denote glass bead types. For probe positions, see Figure 2.9 ……………............. ..89  Figure 4.2  Current flow through ball probes as a function of average particle diameter at different axial and radial positions in the bed, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.08 m/s. For probe positions, see Figure 2.9 .................................................................................................... ..90  Figure 4.3  Effect of particle size on net cumulative charge as a function of time at different axial and radial positions in beds of binary mixtures, M1, M2, M3 and mono-size GB-XL, P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. Letters on curves denote mixture types. For binary mixture properties, see Table 4.2. For probe positions, see Figure 2.9 ….....95  Figure 4.4  Effect of particle size on net cumulative charge as a function of time at different axial and radial positions in beds of binary mixtures, M4, M5, M6 and mono-size GB-S, P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. Letters on curves denote mixture types. For binary mixture properties, see Table 4.2. For probe positions, see Figure 2.9 ….....96  Figure 4.5  Effect of particle size on net cumulative charge as a function of time in the freeboard in beds of binary mixtures, M1-M6 and of mono-size GB-S and GB-XL, P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. Letters on curves denote mixture types. For binary mixture properties, see Table 4.2. For position of probe D, see Figure 2.9 ………. ..97 xv  Figure 4.6  Current flow through ball probes as a function of wt% of GB-S in binary mixtures at different axial and radial positions, P = 379 kPa, T = 20 ± 2ºC, RH = 13- 17% and U-U(mf)M = 0.08 m/s. For probe positions, see Figure 2.9 …………………………………………………………………………. ..97  Figure 4.7  Current flow through ball probe D as a function of wt% of GB-S in binary mixtures in the freeboard, P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and UU(mf)M = 0.08 m/s. For position of probe D, see Figure 2.9 ………………...98  Figure 4.8  Standard deviation of charge fluctuations as a function of wt% of GB-S in binary mixtures at different axial and radial positions, P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. For probe positions, see Figure 2.9 ………………………………………………………………… ..98  Figure 4.9  Standard deviation of charge fluctuations as a function of wt% of GB-S in binary mixtures in the freeboard. P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. For position of probe D, see Figure 2.9 ………. ..99  Figure 4.10  Charge fluctuations obtained by filtering the net cumulative charge curves for probe A versus time using a high-pass filter of cut-off frequency of 10*(1/period) where period is the data sampling range for GB-XL, M1, M2 and M3 mixtures. P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and UU(mf)M = 0.08 m/s. For binary mixture properties, see Table 4.2. For position of probe A, see Figure 2.9 ………………………………………...99  Figure 4.11  Determination of beginning, minimum and total fluidization velocity, Ubf, Umf, and Utf, respectively (adapted from Yang, 2003) ……………………102  Figure 4.12  Comparison of net cumulative charge as a function of time at different axial and radial positions for GB-L and HDPE particles, dP = 574 and 450 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 ……………………………. 108  Figure 4.13  Comparison of net cumulative charge as a function of time at different axial and radial positions for LLDPE and HDPE particles, dP = 487 and 450 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and UUmf = 0.05 m/s. For probe positions, see Figure 2.9 ……………………... 109  Figure 4.14  Comparison of net cumulative charge as a function of time at different axial and radial positions for LLDPE, PE-X1 and PE-X2 particles, dP = 487, 523 and 502 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 1317% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 ………....111  Figure 4.15  Comparison of net cumulative charge as a function of time in the freeboard region for LLDPE, PE-X1 and PE-X2 particles, dP = 487, 523 and 502 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Letters on the curves denote polyethylene types. For position of probe D, see Figure 2.9 ………………………………………. 112 xvi  Figure 4.16  Comparison of net cumulative charge as a function of time at different axial and radial positions for PE-X1 particles, dP = 523 µm, P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Letters on the curves denote experiment numbers. For probe positions, see Figure 2.9 ……….. 114  Figure 5.1  Effect of relative humidity of fluidizing air on net cumulative charges as a function of time at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC, and U-Umf = 0.04 m/s. Numbers on curves denote % RH. For probe positions, see Figure 2.9 …. 117  Figure 5.2  Current flow through ball probes as a function of fluidizing air relative humidity at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC and U-Umf = 0.04 m/s. For probe positions, see Figure 2.9 ………………………………………. 118  Figure 5.3  Effect of added fines on net cumulative charge as a function of time at different axial and radial positions in bed of binary mixture of GB-L particles and GB-XS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on the curves denote wt% of added fines (GB-XS) in binary mixtures. For probe positions, see Figure 2.9 …………………………………….…………….121  Figure 5.4  Effect of added fines on net cumulative charge as a function of time in the freeboard in beds of binary mixture of GB-L particles and GB-XS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 1317% and U-Umf = 0.05 m/s. Numbers on curves denote wt% of added fines (GB-XS) in binary mixtures. For position of probe D, see Figure 2.9 122  Figure 5.5  Current flow through ball probes as a function of wt% of added fines (GBXS) at different axial and radial positions in bed of binary mixture of GBL particles and GB-XS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 ………………………………………………….. 122  Figure 5.6  Current flow through ball probe D as a function of wt% of added fines (GB-XS) in the freeboard in bed of binary mixture of GB-L particles and GB-XS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For position of probe D, see Figure 2.9 ………………………………………………………………… 123  Figure 5.7  Charge fluctuations obtained by filtering the net cumulative charge curves for probe B versus time using a high-pass filter of cut-off frequency 10*(1/period), where period is the data sampling range, for binary mixtures of GB-L particles GB-XS fines (For physical properties of the GB-L and the added GB-XS fines, see Tables 2.1 and 2.3), P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves denote the wt% of GB-XS. For position of probe B, see Figure 2.9 …….. 124 xvii  Figure 5.8  Standard deviation of charge fluctuations as a function of wt% GB-XS in binary mixtures of GB-L particles and GB-XS fines at different axial and radial positions, (For physical properties of the glass beads and added fines, see Tables 2.1 and 2.3), P = 379 kPa, T = 20 ± 2ºC and RH = 1317%. For positions of probes, see Figure 2.9 ………………...................... 125  Figure 5.9  Standard deviation of differential pressure fluctuations across a bed of binary mixtures of GB-L particles and GB-XS fines as a function of wt% of GB-XS in the binary mixtures (For physical properties of the glass beads and added fines, see Tables 2.1 and 2.3), Pressure = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s ………………………..……...125  Figure 5.10  Differential pressure fluctuations across beds of binary mixtures of GB-L particles and GB-XS fines as a function of time. (For physical properties of the GB-L and added GB-XS fines see Tables 2.1 and 2.3), P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves denote the wt% of GB-XS ……………………………….……………….. 126  Figure 5.11  Power spectrum of charge fluctuations signals and differential pressure signals obtained by fast Fourier transformation (FFT) for binary mixture of GB-L and 1.0 wt% of GB-XS fines. (For physical properties of the GBL and added GB-XS fines, see Tables 2.1 and 2.3), P = 397 kPa, T = 20 ± 2ºC, RH = 13-17%, U-Umf = 0.05 m/s …………………………………… 127  Figure 5.12  Effect of antistatic agent (Larostat 519) on net cumulative charge as a function of time at different axial and radial positions in a binary mixture of GB-M particles and Larostat 519, dp = 365 and 13 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 9-17% and U-Umf = 0.05 m/s. Numbers on curves denote wt% of Larostat 519 in the binary mixtures. For probe positions, see Figure 2.9 ………………………………………………….. 130  Figure 5.13  Effect of antistatic agent (Larostat 519) on net cumulative charge as a function of time in the freeboard in a binary mixture of GB-M particles and Larostat, dp = 365 and 13µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 9-17% and U-Umf = 0.05 m/s. Numbers on curves denote wt% of Larostat 519 in binary mixtures. For position of probe D, see Figure 2.9 ..131  Figure 5.14  Current flow through ball probes as a function of wt% of antistatic agent (Larostat 519) at different axial and radial positions in a binary mixture of GB-M and Larostat 519, dp = 365 and 13 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 9-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9 …………………………………………………………………. 131  Figure 5.15  Current flow through ball probes as a function of wt% of antistatic agent (Larostat 519) in a binary mixture of GB-M and Larostat 519 in the freeboard, dp = 365 and 13 µm, respectively, P =379 kPa, T = 20 ± 2ºC, RH = 9-17% and U-Umf = 0.05 m/s. For position of probe D, see Figure 2.9 …………………………………………………………………………132 xviii  Figure 5.16  SEM images of sample discharged from the bed of binary mixture (GB-M and 0.5 wt% Larostat) after experiments, P =379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s …...………………………………………. 134  Figure B.1  Photograph of experimental apparatus ……………………………………154  Figure B.2  Photograph of high-pressure air system …………………………………..155  Figure B.3  Photograph of fluidization section (side view) ……………………………156  Figure B.4  Photograph of sight glass allowing visualization of fluidization inside column ……….……………………………………………………………157  Figure B.5.1 Photograph of mass flow-meter downstream of vessel …………….……..158 Figure B.5.2 Photograph of control panel of fluidization unit ………………………….158 Figure B.5.3 Photograph of flow control valve downstream of vessel ………………....158 Figure B.5.4 Photograph of pressure regulator valve downstream of vessel …………...158 Figure B.6.1 Photograph of preliminary experimental set-up …………………………..159 Figure B.6.2 Photograph of preliminary experimental set-up (top view) ………………159 Figure B.7.1 Photograph of elevated-pressure fluidization column (expanded section) ..160 Figure B.7.2 Photograph of elevated-pressure fluidization column (straight section) … 160 Figure B.7.3 Photograph of elevated-pressure fluidization Column (wind-box and internal cyclone) ………………………………………………………….. 161 Figure C.1.1 Photograph of collision ball probe parts ………………………………….165 Figure C.1.2 Photograph of collision ball probe positions ……………………………... 167 Figure C.2.1 Photograph of Faraday cup ……………………………………………….168 Figure D.2.1 Size distribution of glass beads (GB-S) ......................................................170 Figure D.2.2 Size distribution of glass beads (GB-M) ...............………………..............171 Figure D.2.3 Size distribution of glass beads (GB-L) ..............………………................172 Figure D.2.4 Size distribution of glass beads (GB-XL) ...........................……………… 173 Figure D.4.1 Size distribution of HDPE particles ……………………………………… 177 xix  Figure D.4.2 Size distribution of LLDPE particles ……….…………………………….178 Figure D.4.3 Size distribution of PE-SLD particles ……….……………………………179 Figure D.4.4 Size distribution of PE-X1 particles ………….…………………………... 180 Figure D.4.5 Size distribution of PE-X2 particles ………….………………………….. 181 Figure D.5.1 Size distribution of M1 particles .................................................................182 Figure D.5.2 Size distribution of M2 particles .................................................................183 Figure D.5.3 Size distribution of M3 particles .................................................................184 Figure D.5.4 Size distribution of M4 particles .................................................................185 Figure D.5.5 Size distribution of M5 particles .................................................................186 Figure D.5.6 Size distribution of M6 particles .................................................................187 Figure E.1  Variation of (a) mean bubble rise velocity, (b) mean pierced length, (c) mean bubble frequency and (d) mean bubble volume fraction with pressure at the mid-level of fluidized bed of 0.2 m x 0.3 m cross-section. Particles: 700 µm sand. (adapted from Olowson and Almstedt, 1990) ….. 189  Figure F.1  Drag Coefficient of Single Sphere as a Function of (Re)p (adopted from Schilichting, 1979) ………………………………………………………..193  Figure F.2  Forces on a rotating and moving sphere …………………………………. 199  Figure F.3  Typical collision between two spherical particles ……………………….. 204  xx  NOMENCLATURE Ar  Archimedes number  AS  Specific surface area of particles (m2/g)  DL  Diameter of smallest circumscribing circle of particle (µm)  Ds  Diameter of largest inscribed circle of particle (µm)  db  Collision ball probe diameter (mm)  dp  Mean particle diameter (µm)  d(p)L  Mean particle diameter of large particles (µm)  d(p)M  Mean particle diameter of mixture (µm)  d(p)S  Mean particle diameter of small particles (µm)  Fcontact  Force of contact (N)  g  Gravitational acceleration (9.81 m/s2)  Ip  Current flow through ball probe (A)  L  Bed height (m)  MU  Mechanical unit (can be unit of force, acceleration or pressure).  Nor  Number of orifices in distributor plate (-)  P  Absolute pressure in freeboard (kPa)  Qtransfer  Charge transferred through ball probe (C)  Qp  Cumulative charge transferred through ball probe (C)  Re  Reynolds number (-)  Remf  Reynolds number at minimum fluidization (-)  RH  Relative humidity (%)  r  Radial distance measured from tip of probe to wall of vessel (m)  T  Bed temperature (ºC)  U  Superficial gas velocity (m/s)  Ubf  Beginning fluidization velocity (m/s)  Umf  Minimum fluidization velocity (m/s)  U(mf)M  Minimum fluidization velocity of mixture (m/s)  Umb  Minimum bubbling velocity (m/s)  Ut  Terminal velocity (m/s)  xxi  x  Mass fraction of particles (-)  xS  Average weight fraction of small particles (-)  z  Vertical coordinate measured from top of distributor (m)  ΔP  Pressure drop across the bed (kPa)  Greek Symbols  ρb  Bulk density of particles (kg/m3)  ρg  Fluidizing gas density (kg/m3)  ρp  Particle density (kg/m3)  ρ(p)L  Particle density of large particles (kg/m3)  ρ(p)M  Particle density of mixture (kg/m3)  ρ(p)S  Particle density of small particles (kg/m3)  Φp  Particle sphericity (-)  λ  Radial distance from tip of probe to charged object (m)  µ  Fluidizing gas viscosity (kg/m .s)  ε  Loosely packed voidage (-)  εmf  Voidage at minimum fluidization (-)  Πr  Dielectric constant (-)  xxii  ACKNOWLEDGEMENTS I would like to express my sincere gratitude to my supervisors, Prof. John Grace and Prof. Xiaotao Bi, for their excellent guidance and continuous support throughout this challenging investigation. Special thanks to you for believing in me and giving me the opportunity to be part of the Fluidization Research Group. To Prof. John R. Grace – You provided me with superior knowledge, valuable advice, strength and brotherly guidance throughout my studies. I am honoured to have the grand opportunity working with you, and I have to admit that throughout the period I worked with you, I was trying hard to learn from your dedication towards your work and your remarkable humility. To Dr. Xiaotao Bi –You have been my guide for many years when dealing with a variety of electrostatics challenges. Your brotherly encouragement, mindful suggestions, effective ideas, and vitality helped me throughout my research and to complete this thesis. I am deeply indebted to my employer SABIC for sponsoring me and for granting me the opportunity to complete my graduate studies. I would like to express my gratitude to my supervisors and dear friends in SABIC R&T, Dr. Fahad Al-Obaidi, Dr. Fahad Al-Sherihi, Dr. Atieh Aburaqabah, Eng. Ahmed Al-Marzogi, Dr. Fahad Al-Khodairi, Eng. Abdulqader Malek, Eng. Mohammed Al-Wakeel, Eng. Fuad Mousa, Eng. Mohammed Al-Ali, Dr. Meteab Al-Otaibi Mr. Abdullah Bin Salama and Mr. Thamer Al-Otaibi for their much appreciated cooperation, continuous support and encouragement during my Ph.D. I also wish to acknowledge the staff of Career Development Department in SABIC Americas Inc. and SABIC Al-Jubail, Mr. Waleed Al-Kowaileet, Mr. Saleh Al-Fawaz, Rosie Moshrefi, Kim Murray, Mr. Ali Al-Garni and Mr. Abdullah Al-Aufi for their help and cooperation. Special thanks to my father-in-low Eng. Mohammed Sami Al-Kodamani for his encouragement and sincere prayers. To my examining committee: Prof. Jim Lim and Prof. John D. Madden – I am grateful for your valuable time and dedication. xxiii  It is my pleasure to have an opportunity to thank my brothers and dear friends, who encouraged me to accomplish this work and made my life at UBC and in beautiful Vancouver such an enjoyable experience. The following list is almost certainly incomplete with my appreciation to: Ziad Moughrabiah, Eng. Uthman Al-Saiari, Dr. Mohammed Abuaish, Eng. Hesham Al-Salman, Eng. Mohammed Al-Rasheed, Prof. Mohammed Gadallah, Dr. Khaled Ibraheem, Dr. Marwan Al-Kishi, Prof. Tamer Khattab, Eng. Tareq AlKhasib, Dr. Khalid Assiri, Dr. Musaed Al-Ghamdi, Eng. Mohammed Al-Sanosi and Prof. Amro Mohammed Salem. I wish to acknowledge and give special mention to Prof. Jim Lim and Prof. Said S.E.H. Elnashaie for their brotherly kindness, valuable advice, wise recommendations, help and encouragement. I am grateful to fellow professors and graduate students for their love and support. My special thanks to my dear friends, Dr. Andres Mahecha-Botero, Monica Danon-Schaffer, Laura Botero-Ramirez, Prof. Poupak Mehrani, Prof. Naoko Ellis, Dr. Ana Stevanova, Dr. Aihua Chen and Omer Muammar, for their ongoing support, useful discussions and sincerely encouragement throughout my Ph.D. I am also grateful to the Natural Sciences and Engineering Research Council of Canada for their financial support of this work. I would like to acknowledge staff of The Chemical Engineering Department, Horace Lam, Alex Thang, Helsa Leong, Amber Lee, Graham Lebelt, Richard Ryoo and Doug Yuen for their expertise, experience, help and friendship. Special thanks to Helsa Leong for her academic advice and assistance. I pray to God to bless her with complete health.  xxiv  DEDICATION This thesis is dedicated to my beloved parents Eng. Osama Moughrabiah and Hend Zabad. I could not have achieved any goal throughout my life without their love, prayers, support, encouragement and fatherly care. It is also dedicated with love to my beloved wife, Noura AlKodamani whose patient and endless love enabled me to complete this work. Special dedication to my children Osama, Hend, Al-Maha and Reem who have filled my life with love and happiness. I also wish to dedicate this work to my beloved sisters Lena and Lama, and to the memory of my beloved brother Dr. Sultan Moughrabiah (1973-1997), who has been a source of brotherly love, support and encouragement to me throughout my life.  xxv  Chapter 1: Introduction  Chapter 1: Introduction  Fluidization is associated with solid particles being transformed into a fluid-like state by an upward flowing fluid. It arrived on the industrial scene in a major way in the early 1940s with Fluid Catalytic Cracking (FCC) and has since been implemented in many other industrial areas, including solid-catalyzed gas-phase reactions, non-catalytic reactions and physical processes (Grace and Matsen, 1980). The excellent gas-solid contacting and heat transfer in gas-solid fluidized beds represent their major strengths for a number of industrial processes, (e.g. fluidized bed combustion, gasification of coal, ore roasting, drying, coating, heat exchange, catalytic processes such as acrylonitrile, aniline and Fischer-Tropsch synthesis, and gas-phase polyolefin processes) (Kunii and Levenspiel, 1991). Electrostatic charging in gas-solid fluidized beds was first reported about 60 years ago in connection with anomalous behaviour encountered in experiments on subjects as diverse as heat transfer (Miller and Logwinuk, 1951), elutriation (Osberg and Charlesworth, 1951), and characteristics of fluidized particles (Lewis et al., 1949). Problems associated with fluidized bed electrification include particle-wall adhesion, inter-particle cohesion and electrostatic discharges. Fluidizing a dielectric material, such as glass beads, polystyrene and polyolefin particles, tends to generate significant electrostatic charges within the fluidized vessel. The charged particles can coat vessel walls, requiring frequent cleaning. They can also interfere with sensors and bed internals. The charged particles and high-voltage electric fields arising from them can significantly alter the bed hydrodynamics. They are also responsible for potentially severe problems in commercial gas-solid fluidized bed facilities, such as agglomeration, sheeting, shank (fusion of solid particles into solid shapes), nuisance discharges and product handling. In addition, unintentional charge accumulation can cause problems in industry ranging from minor nuisance to severe explosion risk (Cross, 1987). The number of factors influencing electrostatic charge generation, accumulation, transfer and dissipation in gas-solid fluidized beds is high, e.g. bubble behaviour, particles rubbing against each other in the region surrounding a rising bubble and against the wall, roughness and condition of surfaces, particle properties, relative velocity of particles, fluid physical 1  Chapter 1: Introduction properties and operating variables such as pressure and temperature. All of these factors, together with the complexity of fluidized media, explain the dearth of studies on fluid bed electrostatics. In recent years, there has been a reawakening of interest in electrostatics, motivated by the need to understand causes of explosions and the rapid development of industrial applications. The central goal of this thesis is to investigate the influence of particles properties and operational variables like pressure, temperature, fluidizing air velocity and relative humidity in order to gain a better understanding of electrostatic phenomena and to characterize electrostatics charge generation, accumulation and dissipation in gas-solid fluidized beds.  1.1  Charge Generation  The mechanism of electrostatic charge generation is quite complex. Electrons or ions can transfer between bodies in contact, forming an electrical double layer consisting of two layers of charges of opposite sign. If the bodies are suddenly pulled apart, the original electronic equilibrium cannot be re-established, and one of the surfaces retains more electrons or ions than before the contact, while the other has acquired a deficit. The net charge of the two surfaces remains constant. However, if one of the surfaces loses its charge (for instance, because it is a better conductor or is earthed), the global result is a net electrical charge (Cross, 1987). Fluidization is, by its very nature, associated with continuous contact and separation, as well as with friction, as particles rub against each other, as well as against the wall. These circumstances favour electrostatics. In gas-solid fluidized beds, triboelectrification, frictional charging, thermionic emission and ion collision in high-temperature processes are known to generate electrostatic charges.  1.1.1  Triboelectrification  Triboelectrification involves the generation of electrical charges due to rubbing between materials. When two materials are in contact, charges move from one to the other based on the energy of the electrons and ions at the surfaces until charge equilibrium occurs (Fan and Zhu, 1998). Upon separation, the particles that have lost electrons become positively charged, whereas those that have gained electrons become negatively charged. 2  Chapter 1: Introduction Triboelectrification, also known as contact electrification, occurs due to the difference in the initial Fermi energy levels of the materials at the contact surface until the energy levels are equalized. The Fermi energy level is the highest occupied energy level at absolute zero temperature. At non-zero temperatures, electrons have thermal energy; therefore some have energies slightly above the Fermi energy, leaving some energy levels empty below the Fermi energy. The energy required to move an electron from the top of the energy distribution, out of the material to infinity, is called the work function (Cross, 1987). For the majority of metals this is ~4 electrons volt (eV) and depends on surface impurities such as adsorbed gases. According to Cross (1987), electron energies in an insulator are a function of position, surface impurities and local atomic structure, as well as the chemical nature of the material. Therefore, the work function of an insulating material should be determined experimentally. Early experiments (Davies, 1969; unpublished results cited in Seanor, 1992) showed that the amount of charge transferred to an insulator, such as polymer or glass beads, in simple contact, by each of a series of metals, was proportional to the metal work function. Therefore, insulators such as polymers can be arranged in work function series. A comparison of two such series is presented in Table 1.1. Work function values represent the charging polarity of two surfaces due to tribocharging. The higher the work function, the more negative the charge. However, it can be seen from Table 1.1 that both estimates of polymer work functions and the order of materials in the work function series differ for different workers. Charge polarities of different materials have been determined by a number of authors and have been arranged in triboelectric series. Table 1.2 summarises a number of triboelectric series (Cross, 1987). The order of materials in a triboelectric series is not always unique, even in one series of tests, and there is a lack of reproducibility when tests are repeated, reflecting the inhomogeneities of surfaces.  3  Chapter 1: Introduction Table 1.1 Polymer work function series of polymers in two studies (adapted from Cross, 1987). Work Function (eV) Material  Davies (1969)  Strella (1970)  PVC  4.85 ± 0.20  5.13  Polyimide  4.36 ± 0.06  Polycarbonate  4.26 ± 0.13  4.8  Teflon (PTFE)  4.26 ± 0.05  5.75  Polyethylene Terepathalate (PET)  4.25 ± 0.10  Polystyrene  4.22 ± 0.07  +  Nylon 66  4.08 ± 0.06  -  Teflon (PTFE)  5.75  Polychlorotrifluoethylene  5.30  Polychlorinated propylene  5.14  PVC  5.13  Polychlorinated Ether  5.11  Poly-4-chlorostyrene  5.11  Poly-4-chloro-4-methoxy-styrene  5.02  Polystyrene  4.90  Polyethylene  4.90  Polycarbonate  4.80  Polyethylene-vinyl acetate  4.79  Polyvinyl acetate  4.38  Polyvinyl butyral  4.3  Poly-2-vinylpyridine-styrene  4.27  -  +  4.3-4.54  Nylon 66  4.03-4.54  Polyethylene oxide  3.95-4.50  4  Chapter 1: Introduction Table 1.2 Triboelectric series (adapted from Cross, 1987).  5  Chapter 1: Introduction The charge polarity can be influenced by various factors such as surface finish, preconditioning, material purity, particle shape and moisture content (Cross, 1987). For example Figure 1.1 represents an exception to the triboelectric series determined by some researchers by changing the rubbing manner (Harper, 1967). As a result, it is impossible to predict with certainty the charge polarity of industrial solids based on published work functions and triboelectric series.  +  -  Zinc  -  +  Silk  Glass  +  -  Filter Paper  + +  -  Cotton  Figure 1.1 Closed triboelectric series (adapted from Harper, 1967).  1.1.2  Frictional charging  In industrial gas-solid fluidized beds, electrostatic charges arise primarily from surface charge polarization due to friction among gas, particles and reactor walls. If the reactor is large enough to neglect wall effects, rubbing of particles against each other becomes the main cause of charge generation (Park et al., 2002a). Boland and Geldart (1971) attributed the generation of electrostatic charges in gas-solid fluidized beds to particles rubbing against each other around rising gas bubbles, particularly in the wake. Frictional charging can occur between similar and dissimilar materials. The quantity of charge generated due to friction between similar materials can be as great as that from dissimilar materials (Cross, 1987). However, the energy of friction affects charge transfer more than the nature of the material. The charge transfer generally increases as the force at contact and the speed of rubbing  6  Chapter 1: Introduction increase (Montgomry, 1959). According to Zimmer (1970) an increase in rubbing velocity may also change the polarity of the charge transferred, as shown in Figure 1.2. This indicates that frictional charging differs from triboelectrification.  2  Charging current  10-10  0  10-10 0  20  40  Rubbing velocity  60  80  100  (cm s-1)  Figure 1.2 Charging current plotted against velocity for a polyethylene terephthalate disc rubbed against a metal brush (adapted from Zimmer, 1970).  1.1.3  Thermionic emission  Another special charging mechanism is thermal electrification. When solid particles are exposed to very high temperatures (e.g. T> 1,000 K) the electrons inside the solid particles can gain sufficient energy from the high-temperature field to overcome their energy barriers and be freed. They are then said to be thermally electrified (Fan and Zhu, 1998). Charge build-up on the particles occurs due to the tendency of the electrons to escape from the solid particles by thermionic emission and increases as the freed electrons are captured by attractive Coulomb forces (Fan and Zhu, 1998).  7  Chapter 1: Introduction  1.1.4  Ion collision  Field charging, diffusion charging and corona charging are three distinct charging processes for particles in an electric field. Field charging is the predominant mechanism for particles larger than 1.0 μm in diameter (Fan and Zhu, 1998). For particles smaller than 0.2 μm, the contribution of an external electric field becomes insignificant relative to random motion of ions, so that the diffusion charging plays an important role (Fan and Zhu, 1998). In addition, when the electric field exceeds the breakdown field of the gas, the surrounding gaseous medium is ionized (Cross, 1987). Ions of single polarity or bi-polarities in gaseous medium are then attracted to different sectors of a particle's surface, polarized due to the electric field (Cross, 1987). This study concentrates on the first two of these mechanisms, because the latter two are not applicable in our case.  1.2  Fluidization Flow Regimes  A gas-solid fluidized bed may be formed when gas passes upwards through a bed of particles supported on a distributor plate. As the gas velocity increases, different flow regimes are encountered, as shown schematically in Figure 1.3. Not all of these regimes are observed for all systems (Grace, 1986). In this study only the bubbling fluidization regime is investigated.  1.2.1  Bubbling flow regime  In gas-solid fluidized beds of Geldart group B particles, the bed is transformed from a fixed bed into a bubbling fluidized bed when the superficial gas velocity (U) is increased beyond the minimum fluidization velocity (Umf). Umf is usually determined by pressure drop measurements. For Geldart group A particles, no bubbles are observed, i.e. the bed expands homogenously, over an interval of superficial gas velocity beyond Umf. Bubbles only appear when U exceeds the minimum bubbling velocity (Umb) (Yang, 2003). Thus the transition point to the bubbling regime is at Umf for particles in groups B and D, while it is at Umb for Geldart group A particles.  8  Chapter 1: Introduction  Figure 1.3 Gas-solid flow regimes as a function of fluid velocity (adapted from Grace, 1986).  1.3 In  Electrostatics in Bubbling Fluidized Beds bubbling  fluidized  beds  operating  at relatively  low  temperatures  (<125ºC)  triboelectrification and frictional charging are considered to be the main mechanisms of particle charging. Several researchers have studied the influence of bubble behaviour, fluidizing gas velocity and particles properties on the generation, accumulation, dissipation and segregation of electrostatic charges in bubbling fluidized beds.  1.3.1  Influence of bubble behaviour  In bubbling fluidized beds particle-particle collisions among particles surrounding bubbles tend to be much more energetic than for particles elsewhere in the bed, potentially leading to higher charge generation there (Chen et al., 2003a). Bubble flow behaviour thus appears to be the key to understanding charge generation and build-up in the bed. In bubbling fluidized beds, bubbles are constantly changing in size. Relatively small bubbles form at the distributor plate and coalesce as they rise. According to Darton et al. (1977), the mean bubble size at a given height increases with increasing gas velocity and catchment area for bubbles  9  Chapter 1: Introduction generated at the distributor plate. Boland and Geldart (1971) concluded that electrification in fluidized beds is generated by the motion of particles around bubbles, and showed that the voltage associated with the passage of the wake is higher than for the nose. They speculated that most particle-particle charging occurs in wakes because the particle motion is much smoother at the bubble nose than in the wake. When they measured the potential generated around bubbles as the bubbles passed an induction probe mounted on the wall of a twodimensional column, they found that the induced charge on the probe increased with increasing bubble diameter because larger bubbles rise more quickly, causing more particle motion than smaller ones. Boland and Geldart (1971) speculated that the charges were of opposite polarity in the wake and nose regions of bubbles. However, Park et al. (2002a) showed that the results can be described by induction effects. The charge distribution on bubbles in bubbling fluidized beds was modeled by Park et al. (2002a) who considered charge induction and transfer to a collision ball probe in bubbling fluidization. Their model, with its assumption of uniform charge polarity on the surface of a spherical bubble, gave better agreement with experimental results than a model with different charge polarity assumptions suggested by previous work. However, the new model has many limitations due to simplifying assumptions and therefore can only be employed as a first approximation. Chen et al. (2003b) modified the Park et al. (2002a) model by taking into account charge build-up on particles remote from the bubble, and the particle charge density distribution near the bubble. The model predictions in the region before the bubble passes the probe were in good agreement with experimental results. Nevertheless, the model needs further modifications to describe the charge distribution around and behind a bubble. In the above studies, electrostatic charges were measured by ball probes mounted inside the fluidization column. However, verification of the charge distribution surrounding rising bubbles required either direct measurement of the specific charge density on particles surrounding the bubble or indirect measurement of the electric field surrounding the bubble (Bi, 2005). A technique has been developed by Chen et al. (2006) to determine the charge distribution around a single rising bubble in a two-dimensional fluidized bed using a number of induction probes located flush with the wall of a Plexiglas column. For probes at the same 10  Chapter 1: Introduction level, vertical resolution was obtained by the rise of the bubble under the assumption that the charge distribution does not change during bubble passage. Their results for glass beads showed that the emulsion phase far from the bubble was charged negatively. There was a decrease in charge density moving inward from the emulsion phase to the bubble interface, with essentially zero charge density inside the bubble. On the other hand, particles in the wake were strongly charged.  1.3.2  Influence of fluidizing gas velocity  Ciborowski and Wlodarski (1962) suggested that increases in charge generation, observed in their experiments, were directly proportional to increases in fluidizing gas velocity. Chen et al. (2003a) suggested that the increase in generated and transferred charges with increasing bubble volume mainly results from the increase in the velocity of particles surrounding faster-rising bubbles. Guardiola et al. (1992) reported that electrostatic charging in fluidized beds increases with increasing fluidization velocity. Guardiola et al. (1996) explained the increase in electrostatic charges generation with increasing fluidization velocity by the larger bubbles causing increased particle movement. This trend is limited by the onset of either, slugging as bubbles grow, restraining particle motion, or turbulent regime. Yao et al. (2002) attributed the increase in the amplitude of voltage signals observed in their experiments to the increase in bubble size and rise velocity.  1.3.3  Influence of particle properties  Triboelectrification in gas-solid fluidized beds is very sensitive to the electronic surface states of materials in contact. Any surface changes at the time and point of contact may influence both the polarity and magnitude of tribocharging (Cross, 1987). Guardiola et al. (1996) reported that the degree of bed electrification increased with increasing particle size. However, establishing a quantitative relationship between particle diameter and the degree of bed electrification is hampered by interaction between relative humidity, air velocity and flow regime of fluidization. Depending on their morphology, surface geometry, purity, finish and preconditioning, frictionally charged materials may contain both positively and negatively charged areas on their surfaces, while the dominant polarity determines the net surface charge (Cross, 1987). 11  Chapter 1: Introduction Bipolar charging has been explained as contact charging between particles of the same material, but different sizes. In this case the larger particles gain charge polarity opposite to those of the smaller ones (Ali et al., 1998; Zhao et al., 2000; Zhao et al., 2003; Mehrani, 2005 and Inculet et al., 2006). Ali et al. (1998) found that for one type of particles, large particles charged negatively and small ones positively, whereas in other cases the polarities were reversed. Zhao et al. (2000) investigated the relation between particle size and polarities for some polymer particles and concluded that smaller particles charged negatively and larger particles positively. Mehrani (2005) conducted experiments with mono-size and binary mixture of particles consisting of relatively large (566 μm mean diameter) and fine (30 μm mean diameter) glass beads. It was found that the entrained fines transported a net charge out of the fluidized bed, thereby leaving a net charge behind. Since fines are always carried over to a greater or lesser extent in fluidized bed processes, and the capture efficiency of entrained fines is always less than 100% in practice, entrainment could be a major source of build-up of net charges inside fluidized beds. Mehrani et al. (2007a) developed a bench-scale shaking experiments and particle-plate contacting tests to determine the charge polarity and mechanisms of charging due to particle-particle and particle-wall contacts. They showed that the charge polarity of the fine particles was opposite to that of the relatively large particles, indicating that charges carried by fines are due to charge separation rather than charge transfer. Mehrani et al. (2007b) found that fines added to an initially charged fluidized bed carry significant, but different, charges out of the column depending on their sizes, physical and chemical surface structure, and the moisture content of the fluidizing gas. Their experiments showed that relatively large glass beads and polyethylene particles charged negatively, whereas entrained fines charged positively. Small particles carried opposite charges to large ones, demonstrating the occurrence of bi-polar charging. Table A.1 in Appendix A summarizes previous work on electrostatics in gas-solid fluidized beds.  1.4  Electrostatic Measurements and Instrumentation  Different measurement techniques have been employed by previous researchers to measure electrostatic charges in gas-solid fluidized beds. The two main techniques have involved Faraday cups and electrostatic probes. 12  Chapter 1: Introduction  1.4.1  Faraday cup  A Faraday cup is a double-wall vessel of any suitable shape. The outer wall is grounded and forms an electrical screen preventing stray external charges from affecting the measurements. The inner wall is connected to an electrometer which measures charge by detecting the voltage built-up across a known capacitance. The principle of the Faraday cup is illustrated in Figure 1.4. When a charged object enters the inner cup, an equal and opposite charge is induced on the wall of the inner cup. This charge is stored on the capacitor in the electrometer and measured (Cross, 1987).  Charged Object  Electrometer Insulator  Figure 1.4 Faraday cup (adapted from Cross, 1987). The Faraday cup method has often been used (e.g. Tardos and Pfeffer 1980; Wolny and Opalinski, 1983; Ali et al., 1998; Zhao et al., 2000; Mehrani, 2005) to measure net electrostatic charge build-up on particles inside fluidized beds. For example, Mehrani (2005) developed an on-line measurement technique based on the Faraday cup principle by constructing a copper fluidization column of diameter 0.1 m and height 2.1 m inside a surrounding copper column of diameter 0.2 m, 1.7 m height and 0.0016 thick as the outer cup (Figure 1.5). The outer column was grounded to eliminate external electrical interference. The fluidization column was insulated from other parts of the inner and the outer columns by Teflon cylinders and Teflon distributor plates, and it was connected to an electrometer to measure the charges induced on the wall. 13  Chapter 1: Introduction  Plexiglas  Teflon Inner Column (copper) Outer Column (copper)  Teflon Distributor Teflon  Electrometer  Figure 1.5 Schematic of Faraday cup fluidization column (adapted from Mehrani, 2005). The Faraday cup method can provide useful results, but has some disadvantages, such as additional charging during the handling of particles before entering the cup and the ability to measure only net (overall) charges.  1.4.2  Electrostatics probes  Electrostatic charges buildup inside fluidized beds has also been measured by electrostatic probes. Three major types of probes have been employed.  14  Chapter 1: Introduction  1.4.2.1  Capacitance probes  Guardiola et al. (1992) used a capacitance probe to measure the degree of electrification in fluidized beds. In their technique, the probe and distributor were considered to be parallel metallic plates, while the bed acted as a dielectric medium. Shown in Figure 1.6a, the fluidization column was made of Perspex tube, 0.052 m in diameter, whereas the probe was made from a copper bar, 5 mm in diameter, except that the tip was coated by silicon rubber. The distributor plate was a grounded stainless steel screen. The probe was placed along the axis of the column, with the exposed surface at a height above the distributor equal to the static bed height. This height was reported by Ciborowski and Woldarski (1962) to correspond to the highest charge build-up in fluidized beds. The probe-to-distributor voltage drop was then measured as a function of time. This method averages the effect of electrostatic charges over most of the fluidized bed. Wolny and Kazmierczak (1989) developed a capacitance probe consisting of an air plate capacitor to measure the electric field inside a three-dimensional column of cross-section 0.20 m by 0.20 m (see Figure 1.6b). The capacitor consisted of two brass plates of 0.06 m by 0.108 m cross-section, 0.07 m apart.  Air  High tension source  +  -  -y-  0.052 m  Air-plates capacitor Pneumatic gun  Air  (a)  0.200 m  (b)  Figure 1.6 Capacitance electrostatic probes: (a) (adapted from Guardiola et al., 1992), (b) (adapted from Wolny and Kazmierczak, 1989).  15  Chapter 1: Introduction  1.4.2.2  Induction probes  Induction probes have been used by many researchers. The fundamental principle is that a real charge induces an image of itself on a conducting surface. There are two main types. One involves a ball or bar probe with a shielded head. Figure 1.7 shows a probe shielded by a layer of wall material used by Boland and Geldart (1971) in a 0.50 m by 0.013 m twodimensional column. Chen et al. (2006) measured the electrical field induced by rising gas bubbles by placing four induction probes flush with the inner surface of a two-dimensional Plexiglas column to eliminate interference with motion in the bed and charge transfer due to collision between particles and the probe. Signals from the probes were used, in conjunction with the bubble position captured by a synchronized digital camera, to reconstruct the charge density distribution around rising gas bubbles. Another probe type often used in pneumatic conveying systems is a ring probe. These non-contacting probes have the advantage of not disturbing the flow since they are not directly exposed to the fluidized material. However, the disadvantage is that they are not directly exposed to the fluidized material. Particle-wall interactions, rather than particle-particle interactions, determine the output.  3.2 mm inner face (of 2-D bed)  1.27 cm back-up wall Nuts and washer for electrical connection Metal tube  Probe head 2.5 mm dim  Coaxial cable  Figure 1.7 Induction probe with shielded tip (adapted from Boland and Geldart, 1971).  16  Chapter 1: Introduction  1.4.2.3  Collision probes  Other researchers have used the same principle of induction and have developed contacting probes made of highly conductive materials inserted along the axis of the bed and connected to electrometers to measure the potential or current generated inside fluidized beds. These are commonly known as collision probes. The most common type of collision probe is a ball probe. Ciborowski and Wlodarski (1962) developed an electrode made of platinum wire shaped as a ball mounted inside the bed by a silk thread connected to an electrometer that measured the potential within the bed. The fluidization column consisted of a glass tube, 0.06 m in diameter and 0.555 m in height, with a grounded steel distributor plate. The whole column was surrounded by a grounded metal screen in an attempt to stabilize the electrical measurements. Fujino et al. (1985) adopted a similar approach by inserting a spherical brass terminal of 6.0 mm in diameter into the fluidized bed supported by a nylon thread and connected to an electrometer (Figure 1.8b). A grounded brass distributor plate eliminated external induced charges. Park et al. (2002b) and Chen et al. (2003b) mounted fixed probes instead of suspended probes to measure charge inducement and transfer due to bubble movement in a two-dimensional fluidized bed. Figure 1.9 shows the ball probe used in both cases. A major disadvantage of these probes is low accuracy. Due to the electrostatic charges generated inside the bed, particles can adhere to the tip of the probe, causing potentials lower than the true potential (see Figure 1.10). Other disadvantages include disturbing the flow since they are suspended inside the bed and introducing extra charging due to particle collisions with the probe.  17  Chapter 1: Introduction 0.06 m Glass tube  scale 0.555 m  Electrode  tower Small reel  electrometer  Metal base  Electrostatic voltmeter  Opening (for thread)  Small rod  Electrometer  Thermometer  (a)  R  gas  (b)  Figure 1.8 Collision electrostatic probes: (a) (adapted from Ciborowski and Wlodarski, 1962), (b) (adapted from Fujino et al., 1985).  Figure 1.9 Collision ball probe (adapted from Park et al., 2002b).  18  Chapter 1: Introduction  V/V max (-)  1.0 GB dp = 250 µm U = 33 cm/s Ψ = 20%  0.5  0  key o Δ  Vmax 1.8kV 0.7kV  He = 11cm  0  30  60  90 Time (s)  120  150  Figure 1.10 Deterioration in measured potential due to adhesion of charged particles on probe tip (adapted from Fujino et al., 1985).  1.5  Influence of Operating Parameters  Most industrial gas-solid fluidized bed reactors operate at temperatures well above ambient, and some, such as those used in the production of polyolefin, operate at elevated pressures and over ranges of fluidizing gas velocities. It is therefore important to know how fluidized beds behave at high temperature, at high pressure and at different superficial gas velocities.  1.5.1  Pressure  High pressures, by increasing gas density, affect gas-particle and particle interactions, and gas-solid flow patterns, affecting gas-solid contacting efficiency. Fluidized beds operating at elevated pressures offer several advantages relative to atmospheric pressure units: augmented heat transfer (Botterill and Desai, 1972), decreased particle segregation (Chen and Keairns, 1975), smaller fluctuations and smaller equipment.  1.5.1.1  Influence of elevated pressures on gas bubble behaviour  Pressure exerts a strong influence on the bubbling behaviour of gas-solid fluidized beds. A general conclusion is that pressurized beds exhibit smoother fluidization, and smaller bubbles (e.g. Li and Kuipers, 2002; Botterill and Desai, 1972; Chiteste et al., 1984; Barreto et al., 1983). For Geldart group A particles, this could be due to (i) more gas flowing through the 19  Chapter 1: Introduction emulsion phase due to an increase in emulsion-phase voidage, or (ii) decreased bubble stability leading to break-up into smaller bubbles (Yates, 1996). Olowson and Almstedt (1990) showed that despite the decreased bubble size, their rise velocity, frequency, volume fraction and visible bubble flow rate increase with increasing pressure and excess gas velocity. This is also supported by visual observations. It has also been reported that at high pressures, the bubble flow is increasingly concentrated towards the centre of the bed (Hoffmann and Yates, 1986). For Geldart group B particles, the bubble size was reported to first increase slightly up to 1600 kPa and then to decrease as the pressure increased up to 6000 kPa (Hoffmann and Yates, 1986). This is consistent with the effect of pressure on the mean pierced length of bubbles observed by Olowson and Almstedt (1990). Harrison et al. (1961) postulated a maximum stable bubble size determined by the internal flow in the bubble. If the bubble becomes too large, then the internal circulating flow could cause particle entrainment from the floor of the rising bubble, leading to bubble break-up from the wake region. CFD predictions of Li and Kuipers (2002) suggested that elevated pressures reduce the particle saltation velocity, resulting in a smaller maximum stable bubble size. Alternatively, the bubble splitting theory of Clift and Grace (1972) proposes that for normal freely bubbling beds, splitting occur from the front, due to Taylor instability of the roof. Based on visual observations, Varadi and Grace (1978) suggested that splitting occurred from the roof, with no evidence of higher splitting frequency at elevated pressures. Newton et al. (2001) reported that at elevated pressures the mean bubble diameter decreases and the bubble velocity increases. Li and Kuipers (2002) found that elevated pressures enhance gas-solid interaction and reduce the particle collision frequency, effectively suppressing the formation of large bubbles.  1.5.1.2  Influence of elevated pressures on Umf  Umf may be found by measuring the pressure drop through a bed of particles as a function of the superficial gas velocity. The frictional pressure drop, ∆P, across a packed bed of spherical particles is given by the Ergun equation:  20  Chapter 1: Introduction 2 ΔP 150(1 − ε) μU 1.75(1 − ε) ρ g U = + L Φpdp ε3 (Φ p d p ) 2 ε3  (1.0)  where U is the superficial gas velocity, L the bed height, dp particle diameter, Φp particle sphericity, µ the gas viscosity, ρg the gas density, and ε the voidage. The first term on the right hand side represents the pressure loss through viscous effects and is the dominant term in the viscous flow regime, whereas the second term is the loss due to turbulent effect and is dominant at higher Reynolds numbers. At Umf the buoyed weight of the bed is fully supported by the flow of the gas, so that the pressure drop across the bed equals the bed weight minus buoyancy per unit area:  ΔP = (1 − ε mf )(ρ p − ρ g )g L  (1.1)  Here εmf is the voidage at minimum fluidization and ρp is the particle density. Umf is found by combining Equations (1.0) and (1.1) to give  Ar =  150(1 − ε mf ) 1.75 Re 2mf + Re mf 3 ε mf Φ p ε 3mf Φ 2p  (1.2)  where the Archimedes number (Ar) and Reynolds number at minimum fluidization (Remf) are defined as:  Ar =  d 3p ρ g (ρ p − ρ g )g μ2  (1.3)  and Re mf =  d p U mf ρ p μ  (1.4)  21  Chapter 1: Introduction  The quadratic can be solved to give Remf and hence Umf. Knowing the effect of pressure and temperature on fluidizing gas density and viscosity, Umf can be calculated over any range of these parameters. For very small particles Equation (1.2) can be simplified to:  U mf =  d 2p (ρ p − ρ g )g ε 3mf Φ 2p 150μ  1 − ε mf  (Ar < 103)  (1.5)  while for large particles Equation (1.2) can be simplified to:  U mf =  d p (ρ p − ρ g )g 1.75ρ g  ε 3mf Φ p  (Ar < 107)  (1.6)  Rowe (1984) rearranged Equation (1.2) to express Umf in terms of the operating variables and showed that for small particles pressure has little or no effect on Umf. As mean particle size increases beyond 500 μm, Umf decreases sharply with pressure up to about 20 bar and more gradually thereafter (Figure 1.11). The same conclusion was reached by King and Harrison (1982) who showed that based on Equation (1.2) Umf is independent of pressure for laminar flow (Remf < 0.5), whereas for turbulent flow (Remf > 500), Umf is inversely proportional to the square root of gas density. Olowson and Almstedt (1991) reported that for Group B, Umf decreases with increasing pressure. Jacob and Weimer (1987) found that with elevated pressure for fine particles (Geldart group A), the minimum bubbling velocity increases. However, the extent of the pressure effect and the reason for it were uncertain. Li and Kuipers (2002) found that increasing pressure reduced Umf. Sidorenko and Rhodes (2004) reported that Umf decreased slightly with increasing pressure for large particles and was independent of pressure for small particles.  22  Minimum fluidization velocity (m/s) …  Chapter 1: Introduction 0.6 0.5 0.4 0.3  dp /microns 1050 550 250 50  0.2 0.1 0 -10 0 10 20 30 40 50 60 70 80 90 100110 Pressure (atm)  Figure 1.11 Effect of pressure on Umf (adapted from Rowe, 1984).  1.5.2  Temperature  Limited work has been devoted to the effect of temperature on fluidization behaviour, and some of what has been published is contradictory. Mii et al. (1973) examined fluidized beds of graphite particles (group B) at temperatures up to 800ºC and found that both the frequency of bubble formation and the quality of fluidization increased with increasing temperature. Similar conclusions were reached by Yoshida et al. (1974) for micro-spherical catalyst particles at 500-1000ºC. Kai and Furasaki (1985) also reported improved quality of fluidization in beds of FCC at temperatures up to 370 ºC. Sittiphong et al. (1981) found a significant increase in the eruption diameter of bubbles in beds of large particles of a refractory material with increasing temperature, contrary to the trend for small particles. The effect of temperature on Umf may be seen by an examination of equation (1.2). Since the density of a gas is inversely proportional to its absolute temperature, gas density decreases with increasing temperature; on the other hand, the viscosity of a gas increases with increasing temperature. The combined effect of changes in density and viscosity results in the Archimedes number decreasing steadily with increasing temperature. Yates (2003) plotted the rearranged equation (1.2) with a square root dependence of gas viscosity on 23  Chapter 1: Introduction  temperature (Figure 1.12), and showed that a rise in temperature causes the Umf of group A particles to decrease somewhat, whereas for group B particles, Umf first increases due to the decrease in gas density, then decreases due to the increase in the viscosity of a gas, as the  Minimum fluidization velocity (m/s) …  effect of gas viscosity starts to dominate.  0.6 0.5 0.4 0.3  dp /microns 1050 550 250 50  0.2 0.1 0 -100 0 100200300400500600700 8009001000 1100 Temperature (C)  Figure 1.12 Effect of temperature on Umf (adapted from Yates, 2003).  1.6  Electrostatic Charge Reduction Techniques  In the past there have been some efforts to prevent or reduce electrostatic charges generation in gas-fluidized beds, but limited fundamental research. As a result, mechanisms of reducing electrostatic charge accumulation are still poorly understood.  1.6.1  Gas humidification and ionization  Increasing the relative humidity of the fluidizing gas is a common means of reducing or eliminating electrostatic charges. Ciborowski and Wlodarski (1962) concluded that the moisture content of the fluidizing stream does not change the rate of static charge generation, but increases the rate of dissipation. Katz and Sears (1969), Boland and Geldart (1971) and Park et al. (2002b) found that moisture in the fluidizing gas reduces electrostatic effects in 24  Chapter 1: Introduction  fluidized beds by increasing surface conductivity, causing more rapid dissipation of electrostatic charges. Guardiola et al. (1996) reported that the effects of relative humidity are complex and depend on the type of fluidization. Their results indicate that when the relative humidity was less than ~10%, electrostatics could not be measured accurately because particles adhered to the electrostatic probe, leading to irreproducible voltages. They also reported that gas relative humidities of approximately 30-65% strongly affected the degree of electrification, with a small increase in humidity in this range causing a sharp drop in generated charges. However, introducing moisture to reduce electrostatic charge accumulation has been found to be ineffective at temperatures above 80°C (Ciborowski and Wlodarski, 1962). In addition, relative humidities >75% can lead to excessive capillary forces, caused by condensation of water vapour on particle surfaces, causing defluidization (Guardiola et al., 1996). Moreover, humidifying the fluidizing gas is not possible for some gas-solid catalytic fluidized bed reactors where humidity poisons the catalyst. Yao et al. (2002) reported that as the relative humidity increased the amplitude of the electrostatic charge voltage signals decreased due to a reduction in charge density buildup in the vicinity of bubbles and a likely decrease in bubble size. Some researchers have investigated the effect of injecting ions into the bed to neutralize static electricity generated in the bed. Revel et al. (2002) reported that charge neutralization by ionized gases is effective in removing charges from particle surfaces when ionized gases of opposite polarity to the particle charges are used. Mehrani et al. (2005) investigated the possibility of gas ionization in their Faraday cup fluidized bed shown above in Figure 1.5. It was concluded that the air leaving the column did not carry appreciable charges, and therefore particle-gas contact had negligible effect on the particle charging mechanism. Air ionization was not expected to significantly affect particle charge dissipation for the conditions studied. In any case, the application of ionized gases for charge neutralization in commercial fluidized bed reactors, such as polymerization reactors, is limited due to the potential product contamination and/or catalyst poisoning by ionized gases.  25  Chapter 1: Introduction  1.6.2  Addition of fines, antistatic agents and more conductive particles  Several researchers have studied the possibility of reducing the generation of electrostatic charges in gas-fluidized beds by changing the physical/electrical properties of the particles (e.g. average particle size, electrical conductivity, morphology etc.) and/or column material of construction. Guardiola et al. (1992) studied a gas-solid fluidized bed of glass beads and mixtures of glass/steel particles. Their results suggested that: (i) If a conducting material such as steel is added to a gas-solid fluidized bed of a dielectric material such as glass beads, the charge in the bed is considerably reduced. (ii) Even for a steel/glass ratio as low as 4%, the electric charge in the bed undergoes strong oscillations, with the bed charging and discharging repeatedly. Wolny and Opalinski (1983) reported that fine particles belonging to Geldart group C changed the space between particles and increased the particle contact area, resulting in increased dissipation of electrostatic charges. When they added conductive, semi-conductive and dielectric fine materials to the fluidized bed, they found that the effect of the fine additives was independent of their electrical properties. Yao et al. (2002) concluded that the addition of antistatic powders was effective in reducing electrostatic charge build-up in a fluidized bed. Ally and Klinzing (1985) reported that the charge-to-mass ratio of copper particles in a glass tube subject to pneumatic transport conditions was much higher than in a Plexiglas tube because glass can extract electrons from copper more readily than Plexiglas. Their column diameter was only 0.024 m, so that most electrostatic charges were likely generated from friction between the particles and the containing wall. Thus their results cannot be extrapolated to larger vessels, where particle-particle interactions are the main source of charge generation.  1.7  Thesis Objectives  A major goal of this thesis is to improve the understanding of electrostatic charges generation, transfer, accumulation and dissipation phenomena in gas-solid fluidized beds in order to reduce their harmful effects and to be able to apply effective charge reduction techniques. Specific tasks contributing to this main goal are:  26  Chapter 1: Introduction •  Design and construction of a high-pressure fluidization unit capable of operating at temperatures above atmospheric and equipped with a proper air compressor and measurement devices to carry out relevant experiments.  •  Establishment of an adequate electrostatic measurement technique to characterize the degree of electrification in the vessel.  •  Investigation of the contribution of bubble movement to the generation and transfer of electrostatic charges at different operating pressures, temperatures and gas velocities  •  Determination of the influence of particle properties (such as size, type, density and chemical composition) on the degree of electrification in the bed.  •  Clarification of the effect of different charge reduction methods such as gas humidification, addition of fine powders and antistatic agents.  1.8  Thesis Outline  Chapter 2 describes the experimental equipment. The first section in Chapter 2 focuses on the high-pressure fluidization column. The second section presents the electrostatic measurement techniques employed in the experiments and describes in detail the collision ball probe immersed in the high-pressure fluidization column, as well as the digital electrometers and logging system. It also highlights the other instrumentation, such as pressure and temperature sensors, differential pressure transducers, and pressure and gas flow control valves. The third section gives details of the high-pressure fluidizing air system, including the air compressor, air dryers, filters and high-pressure buffer tank. The fourth section gives the relevant properties of the solid particles. The last section is devoted to the experimental approach. Chapter 3 provides details of the experimental results on the influence of operating variables on the degree of electrification in the bed. It presents and analyzes the results of experiments carried out at elevated pressure, different temperatures and different fluidizing air velocity.  27  Chapter 1: Introduction  Chapter 4 is concerned with the influence of particle properties on electrostatic charge generation, accumulation, transfer and dissipation. It presents and analyzes experimental results obtained with different particle sizes, shapes, densities and chemical compositions. Chapter 5 focuses on testing different electrostatic reduction techniques such as fluidizing gas humidification, adding fines, and addition of an electrostatics reduction agent. It finishes by discussing the experimental results and the possibility of implementing these methods in industrial fluidized bed reactors. Chapter 6 presents overall conclusions and recommendations for future work. Additional information on forces exerted on particles, experimental results, detailed design and engineering drawing of the high pressure fluidization vessel and photographs of the high-pressure fluidization unit are available in the appendices.  28  Chapter 2: Experimental Equipment and Approach  Chapter 2: Experimental Equipment and Approach  This Chapter describes the experimental facility including the elevated-pressure fluidization column, collision ball probes, high-pressure fluidizing air system, pressure and flow control systems, data acquisition system, pressure and temperature sensors and differential pressure transducers. It also gives the properties of the particles used in experiments and finishes by presenting the experimental methodology.  2.1  Elevated-Pressure Fluidization Column  Freely bubbling fluidization experiments were conducted in a three-dimensional elevatedpressure (up to 10 bar) fluidization column constructed of stainless steel of 150 mm inner diameter and 2.0 m height (straight section), expanding to 200 mm inner diameter over a 0.75 m height (expanded section). The gas velocity decreases in the expanded section at the top of the column, allowing most fines to drop out of the gas and fall back into the bed. Before the gas leaves the column, it passes through a single internal cyclone of diameter 100 mm to capture and return entrained fines via a return dip-leg to a distance of 0.10 m above the distributor plate. The column is equipped with ports of three different NPT sizes (6.35, 9.53 and 12.70 mm) located at different heights to allow the installation of sensors and devices at various levels. The column is also equipped with three sight glasses of diameter 25.4 mm located at three levels (0.229, 0.629 and 1.089 m above the distributor) to allow visual observation. The sight glasses were continuously flushed with high-pressure air delivered by a 3.2 mm (1/8”) tube to prevent and reduce the adhesion of fine particles to their internal surfaces. A schematic of the elevated-pressure column is provided in Figure 2.1. The column was supported inside a carbon steel support base topped with a rubber pad of thickness 10 mm to absorb any vibrations generated during fluidization, (see Figure 2.2). Detailed engineering drawing and photographs of the elevated-pressure column are provided in Appendix B.  29  Chapter 2: Experimental Equipment and Approach  0.699 m  0.1 m  0.2 m  outlet pipe of dia. 0.076 m  inlet open of dia. 0.025 mm  0.025  0.105  0.013 m (1/2”) Fitting  0.08  2.0 m  0.460 m  0.556 m  0.25 m  0.006 m (¼”) Fitting  0.08 0.110 0.149  0.311 m  0.629 m  0.01 m (3/8”) Fitting  0.15 m  inlet pipe of dia. 0.076 m  Figure 2.1 Schematic of elevated-pressure column (All dimensions are in m). As shown in Figure 2.3 the gas distributor consists of a stainless steel perforated-plate (top plate), containing 50 holes of 4 mm diameter, aligned and supported by a second stainless steel perforated-plate (bottom plate) containing 50 holes of 5.6 mm diameter. A steel screen with 38 μm openings is sandwiched between the two plates to prevent fine particles from dropping into the windbox. The distributor plates were designed to have an open area ratio of 3.8% and a pressure drop across them of 4.5 mbar at a superficial gas velocity of 0.45 m/s to  30  Chapter 2: Experimental Equipment and Approach ensure that the fluidizing gas is uniformly distributed as it enters. Detailed engineering design drawings and photographs of the distributor plates are provided in Appendix B.  1. 2  5m  1.65 m  0. 9  2m  Rubber pad 10 mm thick  1.25 m  Figure 2.2 Schematic of carbon steel base, which support the fluidization column.  2.2  Measurements and Instrumentation  The fluidization column was equipped with several sensors, instruments and control systems. Some were directly attached to the column, such as electrostatic collision probes, temperature and pressure sensors, pressure transducers, pressure safety valve and humidity transmitter (hygrometer), while some, such as a pressure control valve, flowmeter and flow control valve, were installed downstream and upstream of the column.  31  Chapter 2: Experimental Equipment and Approach 0.004 m  Top plate 60º 0.02 m  Sandwiched screen bottom plate 0.056 m  60º 0.15 m 0.02 m  Figure 2.3 Schematic of distributor plates, Nor = 50 orifices, open area = 3.8 %.  2.2.1  Electrostatic charge measurement techniques  The degree of electrification in the bed is the main parameter measured and characterized in this study. As mentioned in Chapter 1, different measurement techniques have been employed to measure the degree of electrification in fluidized beds, each with its own advantages and disadvantages. One of the main techniques is the collision probe. The most common type of collision probe is a ball probe like those shown in Figures 1.8 and 1.9, as used by Ciboroweski and Wlodarski (1962), Fujino et al. (1985) and Park et al. (2002b). To measure the degree of electrification, a collision ball probe was developed and employed in our experiments. This consists of a stainless steel ball of 5.3 mm diameter attached to a stainless steel wire of 0.92 mm diameter, isolated by a Teflon tube. A ceramic tube of 5.5 mm outer diameter prevented charge leakage and maintained a high resistance to the ground. The ceramic tube was protected by a polyethylene tube of 8.75 mm outer diameter. A Faraday cage (i.e. brass tube shown in Figure 2.4) of 12.7 mm outer diameter was used to 32  Chapter 2: Experimental Equipment and Approach reduce background current by eliminating disturbances due to build-up of charges on the column walls. The brass tube was threaded from sides, one side mounted to the wall of the column and the other side plugged by a hollow nut through which the Teflon tube passed. The nut also prevented pressure leakage from the clearances between the tubes. Figure 2.4 shows a schematic of the electrostatic collision ball probe.  Brass tube to reduce background current by eliminating disturbances due to build-up of charges on column walls 12.7 mm O.D. Stainless steel ball Teflon tube isolating 5.3 mm dia S. S. wire  Ceramic tube maintaining a high resistance to the ground 5.5 mm O.D.  Polyethylene tube protecting the ceramic tube 8.75 mm O.D.  Copper wire leads to electrometer  Figure 2.4 Schematic of electrostatic collision ball probe. The degree of bed electrification was characterized by measuring the cumulative charges induced and transferred through the collision ball probes. Each probe was connected directly to an electrometer (Kistler model 5010B Digital Electrometer) by a coaxial connector to minimize distortion of the electrostatic potential field. This electrometer can have a charge output ranging from ±10.0 pC to ±1.0 µC, or voltage output from 10 mV to 1 kV. Output signals from the electrometer were logged into a computer using a DAS08 data acquisition card and Lab-Tech data logging software. A Faraday cup measured the net charge density in the bed. It consists of two concentric copper cups, insulated from each other by a small piece of Teflon. The inner cup was 0.152 m in diameter and 0.127 m tall, whereas the outer cup was 0.203 m in diameter and 0.152 m 33  Chapter 2: Experimental Equipment and Approach tall. The Faraday cup is shown schematically in Figure 2.5. Photographs of the collision ball probe, Faraday cup, and electrometer are provided in Appendix C.  Outer cup  Teflon insulator  0.127 m  0.152 m  Inner cup  Electrometer  0.157 m 0.203 m  Figure 2.5 Schematic of Faraday cup.  2.2.2  Operating variables measurements and control  Total pressures inside the fluidized column were measured by heavy duty flush diaphragm pressure transmitters of 12.7 mm (½”) NPT fitting. Each transmitter is provided with a tube to facilitate flushing of the tip of the transmitter to prevent blockage by fine particles. The transmitter was connected to a pressure controller (Omron E5CK Digital controller). The absolute pressure of the column was set by a pressure control valve (Fisher-Rosemount 24000C-series little Scotty control valve) located downstream of the column and connected to the pressure controller. A pressure relief valve was installed downstream of the column to protect it from excessive pressure build-up. Differential pressure fluctuations and local and overall pressure drops across the column were determined by pressure transducers. Bed temperature was measured by three bi-metal dial thermometers immersed inside the column at different levels. The temperature was controlled by electrical heating tapes (Omega HTWC Heater Tape) around the column and the pipes upstream of the column. The  34  Chapter 2: Experimental Equipment and Approach relative humidity and the temperature of the incoming gas were monitored by a hygrometer (Vaisala Model HMP238 Humidity Transmitter) immersed in the bed. Gas flow was determined by a mass flowmeter upstream of the inlet valve. A flow control valve (Fisher-Rosemount 24000C-series little Scotty control valve) at the entrance of the column establishes the superficial gas velocity by controlling the air flow rate.  2.3  High-Pressure Air System  A high-pressure fluid injected rotary screw compressor (KAESER Model SK19) with a power saving Sigma Profile pressurized the column to the required operating pressure and continuously supplied the required air. The compressor is provided with a Sigma control system, which offers the choice of automatic control of the discharged air pressure. It is also provided with built-in protective devices which include safety pressure relief valves, emergency stop bottom and safety interlock switches on maintenance doors. Pressurized air was dried by passing through a refrigerating unit (KAESER Model TA11 Refrigerated Dryer) and vapour-removal filters downstream of the compressor. The air dryer and filters can be totally or partially bypassed, allowing the relative humidity of fluidizing air to be adjusted. A high-pressure buffer tank (120 gallon) rated at 1850 kPa was located upstream of the column inlet valve to provide air at constant pressure. An additional air dryer, containing silica gel desiccant, was installed between the buffer tank and column inlet valve to provide increased protection from vapour. This dryer could again be totally or partially bypassed to assist in setting the relative humidity of the fluidizing air. A schematic of the overall layout of the elevated-pressure fluidization unit appears in Figure 2.6.  35  Chapter 2: Experimental Equipment and Approach  Figure 2.6 Schematic of overall layout of elevated-pressure fluidization unit. PR: pressure regulator, PVC: pressure control valve, FCV: flow control valve.  2.4  Bed Materials  The particles in this study were glass beads (soda lime glass) of different sizes, polyethylene of different properties, and an antistatic agent (Larostat 519). The glass beads were supplied by Manus Abrasive System Inc. and Potters Industries Inc. The polyethylene particles were donated by Saudi Basic Industries Corporation, SABIC R&T, Al-Riyadh, SA. The Larostat 519 was provided by BASF The Chemical Company. The glass beads, which are smooth and spherical, represent ideal particles, whereas the polyethylene particles with non-smooth surfaces are more typical of industrial conditions.  36  Chapter 2: Experimental Equipment and Approach  2.4.1  Glass beads (GB)  The key physical properties of the glass beads particles are provided in Table 2.1. Table 2.1 Properties of four types of glass beads (ambient conditions). GB-S  GB-M  GB-L  GB-XL  45-90  250-425  424-650  600-850  65  365  574  693  Particle density, ρp (kg/m )  2500  2500  2500  2500  Particle bulk density, ρb (kg/m3)  1596  1561  1552  1571  Loosely packed voidage, ε  0.362  0.376  0.379  0.371  Particle sphericity, Φp  ~1.0  ~0.9  ~0.9  ~0.9  Dielectric constant, Πr  5-10  5-10  5-10  5-10  0.005* 0.225 * For detailed calculation of GB-S Umf, see Appendix D.1.  0.310  0.375  Size range (µm) Vol.-weighted mean diameter dp (µm) 3  Umf at 379 kPa (m/s)  The densities of the glass beads were provided by the suppliers. The size distribution and the volume-weighted mean diameters of the particles were obtained by a Malvern Mastersizer 2000 equipped with a wet cell (see Appendix D.2 for particle size distribution charts). This equipment uses laser diffraction to measure particle sizes from 0.02 to 2000 µm. Glass beads particles were suspended in distilled water for the measurements. The bulk densities of the particles were obtained to determine the loose packed voidage, ε, of the glass beads. A graduated container was used to measure the volume of a known mass of particles. The container, with its top end covered, was inverted and returned quickly to its upright position before the particle volume was measured to create a loose packed condition. The loosely packed voidage was then calculated from  ε = 1−  ρb ρp  (2.1)  Several tests were conducted to ensure the reproducibility of these measurements (see Appendix D.3 for the detailed calculations). 37  Chapter 2: Experimental Equipment and Approach The sphericity of the particles was provided by the suppliers. The dielectric constants (relative permittivity) were obtained from Reitz et al. (1993). The minimum fluidization velocities, Umf, were measured experimentally at 379 kPa from pressure drop versus gas velocity plots at the intersection of two straight line portions, as recommended by Kunii and Levenspiel (1991). According to a common powder classification based on an analysis of representative types of fluidization behaviour with gas at atmospheric pressure and temperature (Geldart, 1973), the glass beads (GB-M, GB-L and GB-XL) were group B particles, whereas the glass beads (GB-S) were group A particles.  2.4.2  Polyethylene (PE)  The key physical properties of the polyethylene particles are provided in Table 2.2. Table 2.2 Properties of polyethylene particles (ambient conditions). HDPE  LLDPE  PE-SLD  PE-X1  PE-X2  45-1410  45-1410  500-700  45-2500  45-2500  Sauter mean diameter, d p (µm)  450  487  642  523  502  Particle density, ρp (kg/m3)  965  918  797  918  918  Particle bulk density, ρb (kg/m3)  385  341  458  355  345  Loosely packed voidage, ε  0.601  0.628  0.425  0.613  0.624  Particle sphericity, Φp  ~0.75  ~0.73  ~0.77  ~0.68  ~0.73  Dielectric constant, Πr  2.3  2.3  2.3  N/A*  N/A*  Umf at 379 kPa (m/s)  0.11  0.11  0.10  0.12  0.11  Size range (µm)  * N/A = Not Available. The densities of the polyethylene resins were calculated using a 50 ml pycnometer. Particles of known mass were added to the pycnometer and the total combined weight (Wpy+p) was measured. Ethanol was then added to the pycnometer to make up the total volume to 50 ml. It was necessary that the particles be insoluble in the liquid used. Hence, ethanol was used. The pycnometer containing the polyethylene resin and the ethanol (Wtotal) was then weighed. The mass of the ethanol was determined as the difference between the total weight of the 38  Chapter 2: Experimental Equipment and Approach pycnometer with the polyethylene resin and the ethanol and the weight of pycnometer with the polyethylene resin i.e. (Wtotal - Wpy+p). The volume occupied by this mass (Veth) was determined from the known density of the ethanol at 20ºC. The volume of the polyethylene particles was determined as the difference between the total volume and the volume of the added ethanol i.e. (50 ml - Veth). Knowing the mass and the volume of the polyethylene particles, the particle density was then calculated. The bulk densities of the particles were also obtained, by the method described in section (2.4.1.), to determine the loosely packed voidage, ε, of the polyethylene particles. The Sauter mean particle diameter of the polyethylene particles were obtained from sieve analysis by means of  dp =  1 x ∑di pi  (2.2)  where xi is the mass fraction of particles of average screen particle size dpi. The polyethylene particles were sprinkled with anti-static spray to reduce adhesion between particle-particle and particle-sieve during sieving. Detailed calculations of the Sauter mean particle diameter and particle size distribution are provided in Appendix D.4. The sphericity of the particles was approximated by the Riley sphericity (Riley, 1941)  Φp =  Ds DL  (2.3)  where DL is the diameter of the smallest circumscribing circle and Ds the diameter of the largest inscribed circle. DL and Ds were determined from Scanning Electronic Microscopy (SEM) images of the particles (Figure 2.7). The dielectric constants were obtained from Jiang et al. (1994). The minimum fluidization velocities, Umf, were again measured experimentally at 379 kPa from pressure drop versus superficial gas velocity at the intersection of two straight line portions, as recommended by Kunii and Levenspiel (1991). The polyethylene particles in experiments fall into group B of Geldart classification.  39  Chapter 2: Experimental Equipment and Approach  HDPE  (c) PE-SLD  (b) LLDPE  (d) PE-X1  (e) PE-X2 Figure 2.7 SEM images of polyethylene particles. 40  Chapter 2: Experimental Equipment and Approach  2.4.3  Added fines  The fines particles added during this study were Larostat 519 and fine glass beads (GB-XS). Their key physical properties are provided in Table 2.3. Table 2.3 Properties of fine particles added during this study. GB-XS  Larostat 519  10-80  6-20  30  13  Particle density, ρp (kg/m3)  2500  520  Particle sphericity, Φp  ~0.9  ~0.7-0.9  Dielectric constant, Πr  5-10  N/A*  Terminal velocity, Ut (m/s)  0.063  0.003  Size range (µm) Volume-weighted mean dia dp (µm)  * N/A = Not Available. Larostat 519 is an antistatic agent which reduces static charge build-up, allowing the main particles to be fluidized, blended, conveyed, and sprayed with less dusting. It can be added to particles to reduce problems in fluidizing or handling these particles. Typical applications include the reduction of static explosion hazards in air-conveyed dust-handling systems and spray-on power coatings to achieve uniform coverage. Physical properties, such as surface roughness and sphericity, of the added fines, were analyzed by Scanning Electronic Microscopy (SEM). Figure 2.8a shows that the Larostat particles are non-spherical and have uneven surfaces, whereas Figure 2.8b shows that the fine glass beads are closely spherical.  2.5  Experimental Approach  To investigate the influence of different operating variables and particle properties on electrostatic charge generation, transfer and accumulation associated with movement of bubbles and particles in a fluidized bed and to examine different electrostatic reduction methods, the approaches were as follows:  41  Chapter 2: Experimental Equipment and Approach  (a) Larostat (519)  (b) GB-XS  Figure 2.8 SEM images of fine particles added during this study. •  Pressure experiments  To determine the effect of pressure on the net charge generated inside the fluidized bed, free-bubbling experiments were conducted with two different sizes of mono-size glass beads (GB-M and GB-L) and one size of polyethylene (HDPE), at pressures up to 724 kPa. To eliminate the effect of any other parameters, this set of experiments was performed at room temperature, and the relative humidity of the fluidizing air was maintained between 13 and 17%. The bed was fluidized at an excess gas velocity, (UUmf), of 0.05 m/s to ensure typical particle-particle and particle-wall interactions. Static bed height and probes locations were the same for all experiments. The bed was fluidized for ~1.0 h to achieve steady-state before data were collected. The degree of electrification in the bed was characterized by eight collision ball probes located at four levels inside the bed (0.15, 0.31, 0.55 and 0.89 m above the distributor), and at two radial positions (r = 0 and 50 mm) as indicated in Figure 2.9. Photographs of the electrostatic collision ball probes inside the column are provided in Appendix C. The Faraday cup was also utilized to measure the charge density in the bed by withdrawing particle samples from on-line sampling ports at different levels.  42  Chapter 2: Experimental Equipment and Approach •  Temperature experiments  To determine the effect of temperature on the net charge generated inside the fluidized bed, free bubbling experiments were conducted with mono-size polyethylene particles (PE-SLD) and glass beads (GB-XL) at a freeboard pressure of 379 kPa and at different bed temperatures (up to 90ºC). The operating procedure and conditions were similar to those described above. The temperatures of the bed of GB particles were also varied (from 25 to 90 and back 25ºC) to study the influence of temperature on the degree of electrification in the bed.  75 mm  25 mm  D’  C  C’  B  B’  A  A’  0.15 m  0.16 m  0.24 m  0.34 m  D  Distributor Plate  Figure 2.9 Locations of electrostatic collision ball probes.  43  Chapter 2: Experimental Equipment and Approach •  Gas velocity experiments  In these experiments, free-bubbling fluidization was conducted at a freeboard pressure of 379 kPa and at room temperature. The same experimental procedures as described previously were again followed. The excess air velocity was varied from 0.03 – 0.12 m/s to determine the influence of superficial gas velocity and to investigate different charging mechanisms in the bed. Two particle sizes of mono-size glass beads (GB-M and GB-XL) and one type of polyethylene (HDPE) were used in these experiments. •  Particle properties experiments  Free-bubbling experiments, involved mono-disperse and binary mixtures of different compositions of glass beads, were performed at a freeboard pressure of 379 kPa, and at room temperature to determine the effect of particle size on electrostatic charge generation, transfer and accumulation and to verify the possibility of different charging mechanisms. Polyethylene particles of different densities (LLDPE and HDPE) were tested to investigate the effect of particle density. In addition, two polyethylene types (PE-X1 and PE-X2) with similar physical properties and different chemical compositions were used in free bubbling experiments similar to those described above to determine the effect of the chemical composition of the polyethylene particles. •  Relative humidity of fluidizing gas experiments  To study of the effect of fluidizing gas humidification, free bubbling experiments were performed at a freeboard pressure of 379 kPa and at room temperature with mono-size glass beads (GB-XL). The relative humidity of the fluidizing air was varied from 5 to 30% by passing the incoming air through a refrigerating dryer, vapour-removal filters and an air dryer containing silica gel desiccant. •  Added fines experiments  To investigate the influence of added fines, both types of glass beads (GB-L and GB-XS) were mixed with different percentages (up to 2.0 wt% of fines). Free bubbling experiments were then conducted at a freeboard pressure of 379 kPa and at room  44  Chapter 2: Experimental Equipment and Approach temperature. The bed was fluidized at an excess gas velocity, (U-Umf), of 0.05 m/s to minimize fines entrainment. The relative humidity of the fluidizing gas was maintained between 13 and 17%. The degree of electrification of the bed was characterized by the collision ball probes shown in Figure 2.9. Static bed height and probe locations were the same for all experiments. Free-bubbling experiments were also carried out with glass beads (GB-M) and added Larostat 519 pre-mixed with different percentages (up to 2.0 wt%) of Larostat 519.  45  Chapter 3: Results and Discussion: Effect of Operating Parameters  Chapter 3: Results and Discussion: Effect of Operating Parameters  3.1  Preliminary Experiments  Four preliminary experiments using a simplified experimental set-up consisting of collision ball probes, digital electrometers (Kistler model 5010B Digital Electrometer), a DAS08 data acquisition card and data-logging software were carried out to test the data acquisition system. As shown schematically in Figure 3.1, the experimental set-up included three ball probes directly connected to electrometers using coaxial connectors to minimize distortion of the electrostatic potential field. Output signals from the electrometer were logged into a computer by the data acquisition card and data-logging software. The probes were attached to a movable wooden plate using clamps. The plate was placed on a second wooden plate to measure the distance between the probe and a charged object, which could be changed by moving the top plate. Objects made of different materials such as a plastic ruler, cotton cloth, glass tube and polyethylene tube were used to create positive or negative charges.  Electrometer Probe A  Electrometer Probe B  Electrometer Probe C  λ  0.000 0.025 0.032 0.038 0.044 0.051  Charged ruler  Figure 3.1 Schematic of simple preliminary experimental set-up. (All dimensions are in m.)  46  Chapter 3: Results and Discussion: Effect of Operating Parameters  3.1.1  Charge transfer and charge induction  When an electrostatic ball probe is immersed in a fluidized column, the output signals should reflect a combination of charge changes due to a direct charge transfer when charged particles come into contact with the tip of the ball probe, and charge induction due to an electrical field around the probe induced by rising bubbles and charged particles in the bed. The charge output from direct charge transfer is considered to cause step changes of magnitude equal to the charge transferred at each contact. To create a negatively charged object, a polyethylene tube of inner and outer diameters 9.5 and 12.7 mm was rubbed against a cotton cloth. From Table 1.2, negative charges should be generated on the polyethylene tube when rubbed against cotton. The tube was washed with water and dried before rubbing. It was rubbed against the cotton cloth for 30 s, and placed at λ = 0.05 m where λ is a radial distance from tip of the probe to the tube, and then moved horizontally starting from probe A toward probe C as shown in Figure 3.1. The electrometers were set to measure charges from ±10.0 pC to ±1.0 µC. The output signals were logged into a computer. As shown in Figure 3.2a, three approximately equal peaks were registered as the tube passed each probe. After the tube moved away, the probe output signals returned almost to zero, indicating that no permanent charge had been transferred. In this experiment the probe measured charge induction. The same experiment was next repeated with the tube making contact with the tip of the probe as it passed each probe. In this case three peaks, of different magnitudes were measured. The output signals did not return to the zero, but shifted to new lines, as illustrated in Figure 3.2b, indicating transfer of charge from the tube to the probes. In this case, the measured peaks were due to charge induction, since the tube had to be brought closer to the probe to make contact, and then removed after contact. The differences between the zero line and new lines indicate the net charge transfer. It can be seen in Figure 3.2b that the tube was partially neutralized due to the contact; hence the peak magnitude of the induced charge decreased as the tube moved from probe A to C.  47  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (nC).  11  (a)  (b)  -1 -1  Q transfer  -3 -3  Probe A Probe B Probe C  -5  -5  0  5  10  70  75  80  Time (s) Figure 3.2 Ball probe experimental results for a PE tube rubbed against cotton cloth: (a) without contact between the charged object and the probe; (b) with contact. The PE tube was washed and rubbed against a cotton cloth for 30 s before it entered a Faraday cup for 2 s, and then it was pulled out without contacting the inner cup. This procedure was repeated twice. Figure 3.3a shows two peaks measured by the Faraday cup each time the tube entered the cup. This experiment was then repeated with the tube making contact with the inner cup for ~4 s. It can be seen in Figure 3.3b that the output signals did not return to zero, demonstrating charge transfer from the tube to the Faraday cup.  3.1.2  Magnitude of measured charge  Although the above experiments were qualitative, rather than quantitative, the collision ball probe should be able to indicate the magnitude of measured charges. Experiments were carried out to verify the capability of the data acquisition system to indicate the magnitudes. In these experiments, a glass tube of inner and outer diameters 11.2 and 15.8 mm was rubbed against two different materials, a cotton cloth and a PE tube of inner and outer diameters 9.5 and 12.7 mm, in order to create different magnitudes of charge. According to the triboelectricity series presented by Cutnell and Johnson (1992), the glass tube in both cases should be positively charged, while in the latter case the charge generated should be higher. The glass tube was washed with water and dried before rubbing. The tube was first rubbed against the cotton cloth for 30 s, and placed in a location at λ = 0.05 m, and then moved 48  Chapter 3: Results and Discussion: Effect of Operating Parameters horizontally from probe A toward probe C as shown in Figure 3.1. The same procedure was followed again after the glass tube rubbed against the polyethylene tube for 30 s. As expected, the magnitude of charge generated on the glass tube was higher in the latter case as shown in Figure 3.4b. Similar results were obtained when the experiments were repeated using the Faraday cup (Figure 3.5).  Cumulative charge (nC)..  3  (a)  (b)  Q transfer  -2 -7 -12 -17  0  10  20  80  90  100  Time (s)  Figure 3.3 Faraday cup experimental results for a PE tube rubbed against cotton cloth: (a) without contact between the charged object and Faraday cup; (b) with contact. According to Cross (1987), some materials lose their charges and became neutralized faster than others. This explains the decrease in the peaks shown in Fiugers 3.4 and 3.5, when the glass tube lost its charges very quickly and became partially neutralized by the time it was moved from probe A to probe C. The same happened in the case of the Faraday cup when the tube lost its charge every time it was extracted and then re-inserted into the cup.  49  Chapter 3: Results and Discussion: Effect of Operating Parameters 1.20 Cumulative charge (nC)..  (a)  (b)  Probe A Probe B  0.85  Probe C  0.50 0.15 -0.20 0  5  10  70  75  80  Time (s)  Cumulative charge (nC) ..  Figure 3.4 Ball probe experimental results when: (a) glass tube was rubbed against a cotton cloth; (b) glass tube was rubbed against a PE tube.  12.0  (a)  (b)  8.5 5.0 1.5 -2.0  0  5  10  Time (s)  80  90  100  Figure 3.5 Faraday cup experimental results when: (a) glass tube was rubbed against a cotton cloth; (b) glass tube was rubbed against a PE tube.  3.1.3  Different charge polarity  Depending on their type, surface finish, preconditioning, roughness, purity and physical and chemical properties, materials can gain negative or positive charges. Experiments were carried out by rubbing a plastic ruler against two different types of materials (human hair and a cotton cloth). According to triboelectricity series (Cutnell and Johnson, 1992) negative charges are generated on the ruler when it is rubbed against hair, whereas positive charges 50  Chapter 3: Results and Discussion: Effect of Operating Parameters are generated when it is rubbed against cotton. The same experimental procedure was followed. Figure 3.6a confirms that the ruler became negatively charged when rubbed against hair, while Figure 3.6b confirms that the ruler became positively charged when rubbed against cotton. Figure 3.7 shows the same results using the Faraday cup.  Cumulative charge (nC)  1.0  (a)  (b)  0.5 0.0 Probe A  -0.5  Probe B Probe C  -1.0  0  5  10  70  75  80  Time (s)  Figure 3.6 Ball probes experimental results when: (a) plastic ruler was rubbed against hair; (b) ruler was rubbed against cotton.  Cumulative charge (nC)  10  (a)  (b)  5 0 -5 -10  0  10  20  70  80  90  Time (s)  Figure 3.7 Faraday cup experimental results when: (a) plastic ruler was rubbed against hair; (b) ruler was rubbed against cotton.  51  Chapter 3: Results and Discussion: Effect of Operating Parameters  3.1.4  Probe sensitivity  One of the objectives of this study is to investigate the degree of electrification at different locations in a fluidized bed. Therefore, a number of collision ball probes were immersed at different radial and axial locations in the bed (see Figure 2.9). Since the probe signals could be influenced by the induced electrical field in the bed, it was necessary to limit the sensitivity of the collision ball probes. The sensitivity of the probe was adjusted to be as low as possible when a charged object was placed at λ = 0.038 m and as high as possible at λ = 0.025 m. This helped to localize the measurements for each probe. Experiments were performed with a PE tube of inner and outer diameters 9.5 and 12.7 mm and a cotton cloth. The tube was rubbed against the cotton cloth for 30 s and placed at two different locations (λ = 0.025 and 0.038 m) and then moved horizontally, starting from probe A toward C as shown in Figure 3.8. Figures 3.9a, c and e show three examples of experimental results when the tube shown in Figure 3.8a traversed maintaining λ = 0.025 m. The sensitivities of the probes were set at 5, 3 and 1 pC/Mechanical Unit (MU), respectively. Figures 3.9b, d and f show the corresponding results when the tube shown in Figure 3.8b traversed maintaining λ = 0.038 m, and same sensitivity levels as before. After each experiment, the tube was lowered into Faraday cup in order to measure its degree of electrification and to ensure that the tube was similarly charged in all experiments so that the changes reflected in the probes results were due to the changes in probe sensitivities. Figures 3.9e and f suggest that the best balance was achieved when the sensitivities were set at 1 pC/MU.  A  B  C  (a)  A  B  0.000 0.025 PE tube  PE tube  C  (b)  0.000 0.025 0.038  Figure 3.8 Schematic of simple experimental set-up for probe sensitivity experiments. (All dimensions are in m.)  52  Chapter 3: Results and Discussion: Effect of Operating Parameters  15.0  1.2  10.0  0.6  5.0  0.0  0.0  Cumulative Charge (nC). Ball Probe  (c)  20.0  (d)  1.8 1.2  10.0  0.6 0.0  0.0  2.4  20.0 (e)  (f)  1.8  15.0  1.2  10.0  0.6  5.0  0.0  0.0 0  5  10 Time (s)  Probe A  15  20 0  Probe B  Cumulative Charge (nC). Faraday cup  1.8  2.4  Cumulative Charge (nC). Ball Probe  20.0  (b)  5  10 15 Time (s)  Probe C  Cumulative Charge (pC). Faraday cup  (a)  Cumulative Charge (nC). Faraday cup  Cumulative Charge (nC). Ball Probe  2.4  20  Faraday cup  Figure 3.9 Experimental results for ball probes and Faraday cup obtained by traversing a PE tube which had been rubbed against a cotton cloth, at different probe sensitivities, PS, different distances from the tip of the probe, λ,: (a) PS = 5 pC/MU and λ = 0.025 m; (b) PS = 5 pC/MU and λ = 0.038 m; (c) PS = 3 pC/MU and λ = 0.025 m; (d) PS = 3 pC/MU and λ = 0.038 m; (e) PS = 1 pC/MU and λ = 0.025 m; (f) PS = 1 pC/MU and λ = 0.038 m. 53  Chapter 3: Results and Discussion: Effect of Operating Parameters  3.2  Effect of Operating Parameters  Electrostatic problems in gas-solid fluidized beds have been studied by a number of researchers (see Appendix A). It has been found that the electrostatic charge generation, accumulation, transfer and dissipation in gas-solid fluidized beds are influenced by many factors such as bubble behaviour, particles rubbing against each other in the region surrounding rising bubbles and against the wall, relative velocity of particles, regime transition velocities, particle properties, particle type, and fluid physical properties. All experiments in these previous studies were performed at ambient operating conditions (atmospheric pressure, 25 ±5ºC). Since all of the factors that influence charges are affected by operating variables such as pressure, temperature and fluidizing gas superficial velocity, an investigation of the effect of these variables on the degree of electrification is needed to improve understanding of electrostatic charge generation in gas-solid fluidized beds.  3.2.1 3.2.1.1  Effect of operating pressure Glass beads  In order to investigate the effect of pressure on the degree of electrification in a fluidized bed, freely bubbling experiments were conducted in the three-dimensional elevated-pressure column described in Section 2.1 with collision ball probes at four levels (0.15, 0.31, 0.55 and 0.89 m above the distributor), and at two radial positions (probe tip 25 and 75 mm from the wall) as shown in Figure 2.9. The particles were mono-disperse smooth spherical glass beads (GB-L) of a volume-weighted mean diameter 574 µm. (For physical properties of the GB-L, see Table 2.1.) The static bed height was at ~0.53 m. Umf was found experimentally to be 0.31, 0.28, 0.26 and 0.25 m/s for freeboard absolute pressures of 379, 448, 587 and 724 kPa, respectively. Temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%) in order to isolate the effect of pressure. Excess gas velocity (U-Umf) was also maintained constant in all experiments at 0.05 m/s for the same purpose. The bed was fluidized for about an hour to achieve steady state before collecting data, as recommended by Park (2000). The net cumulative charge in the bed was then measured by the collision ball probes and plotted versus time in Figure 3.10.  54  Cumulative charge (μC)  Cumulative charge (μC)  Cumulative charge (μC)  Chapter 3: Results and Discussion: Effect of Operating Parameters  -1.0  (a)  Probe C  -0.8  (b)  Probe C' 724  586  -0.6  448  724  -0.4  379  586 448  -0.2  379  0.0 -1.0 -0.8  (c)  (d)  Probe B  Probe B'  724  -0.6  724 586  586  -0.4  448  -0.2  379  448  379  0.0 -1.0 -0.8  (e)  (f)  Probe A  Probe A'  586  -0.6  586 724  724  -0.4  448  448  -0.2 0.0 0  100  200  300  Time (s)  379 400 500  0  100  200  300  379  400  500  Time (s)  Figure 3.10 Effect of operating pressure on net cumulative charge as a function of time at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves are absolute pressures (in kPa) at top of column. For probe positions, see Figure 2.9. By definition, current is the rate at which charge is transported through a given surface (Bohn, 1968). Based on the convention that the direction of current is opposite to the direction of electron motion, the actual cumulative charge, Qp, transferred through the ball probe, is given by 55  Chapter 3: Results and Discussion: Effect of Operating Parameters  Q p = − ∫ I p dt  (3.1)  where Ip is the current flow through the ball probe due to direct charge transfer. As previously mentioned there are two components in charge output, induced charge and charges due to direct charge transfer. The induced charge does not affect the net cumulative charge output, since there is no direct charge transfer. Therefore the current flow through the ball probe due to the direct charge transfer, Ip, can be obtained by calculating the slope of the net cumulative charge curves shown in Figure 3.10. Figure 3.11 was obtained by plotting the slopes of the curves versus pressure. On the other hand, the fluctuation of the cumulative charge measured in the bed could be due to several factors such as passage of bubbles, charging instability, charges induced by charged particles in the bed and charge transfer. Figure 3.12 shows an example of the charge fluctuation curves obtained by filtering net cumulative charge curves shown in Figure 3.10e as measured by probe A at different pressures. Hence, the standard deviation of charge fluctuation curves can, to some extent, be influenced by movement of bubbles, charging stability, charge induction and charge transfer. Figure 3.13 plots standard deviation of the charge fluctuation curves versus pressure. The experiments were repeated at the same conditions, but with the probe locations interchanged (probe A to location C, probe C to location B, and probe B to location A) to ensure that the differences are not due to differences in the probes. Figure 3.14 shows the corresponding effect of the pressure on the current flow through the ball probes due to direct charge transfer in a bed of GB-L after the switching of probes, while Figure 3.15 depicts the effect of the pressure on the standard deviation of the charge fluctuation curves. Comparison of Figures 3.13 and 3.15 shows some differences, but trends are similar and the differences are likely due to day-to-day variations when dealing with electrostatics, rather than significant differences in the ball probes.  56  Chapter 3: Results and Discussion: Effect of Operating Parameters  Current (nA) .  -10  Probe A Probe B Probe C  (a)  -7.5  Probe A' Probe B' Probe C'  (b)  C  A'  B  -5  C'  A  -2.5  B'  0 350  450  550 Pressure (kPa)  650  750  350  450  550  650  750  Pressure (kPa)  Figure 3.11 Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9. In order to verify the reproducibility of the elevated-pressure results for different particle sizes, freely bubbling electrification experiments were conducted by fluidizing a bed of mono-disperse glass beads particles (GB-M) of volume-weighted mean diameter 365 µm at the same operating conditions and following the same experimental procedure as above. (For physical properties of the GB-M particles, see Table 2.1). Umf for GB-M was found experimentally to be 0.225, 0.170, 0.161 and 0.154 m/s for freeboard absolute pressures of 379, 448, 587 and 724 kPa, respectively. The excess gas velocity was again maintained constant in all experiments (at U-Umf = 0.05 m/s). Figure 3.16 shows the effect of pressure on the current flow through the ball probe, reflecting the degree of electrification at different locations in the bed, whereas Figure 3.17 shows the effect of pressure on the charge fluctuations in the bed. It can be seen in Figures 3.10-3.17 that as pressure increased the degree of electrification in the bed increased. However, the magnitude of this increase differed from location to location.  57  Chapter 3: Results and Discussion: Effect of Operating Parameters  Charge fluctuations (μC) .  0.12  (b)  379 kPa  448 kPa  0.00  0.0  -0.12  0.1  0.12 0 Charge fluctuations (μC) .  0.1  (a)  2  4  6  8  10  0.1  (d)  (c)  724 kPa  586 kPa  0.00  0.0  -0.12  0.1 0  2  4  6  Time (s)  8  10  0  2  4  6  8  10  Time (s)  Figure 3.12 Charge fluctuations obtained by filtering net cumulative charge curves for probe A versus time using a high-pass filter of cut-off frequency of 10*(1/period) where period is the data sampling range in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves are absolute pressures (in kPa) at top of column. For position of probe A, see Figure 2.9. Kawabata et al. (1981) conducted freely bubbling experiments at elevated pressures (up to 800 kPa) in a two-dimensional fluidization column containing sand of three different mean particle sizes (300, 430 and 600 µm) with particle densities, ρp, of 2630, 2580 and 2590 kg/m3, respectively. Their experiments showed that although bubble size was not influenced by elevated pressure, bubbles became flatter. Olowson and Almstedt (1990) measured bubble characteristics in a bed of particles of 700 µm mean diameter and 2600 kg/m3 density at pressures up to 1600 kPa. They observed that the mean bubble frequency, rise velocity, volume fraction, and visible bubble flow rate increased with increasing pressure. They also found that the mean pierced length of bubbles decreased, after an initial increase, with 58  Chapter 3: Results and Discussion: Effect of Operating Parameters  increasing pressure (see Appendix E). Their experimental results are in full agreement with experimental results reported by Olssen et al. (1995) who measured the mean bubble rise velocity, frequency, and volume fraction at pressures up to 1600 kPa in a bed of sand of 700 µm mean diameter and particle density 2600 kg/m3. On the other hand, Boland and Geldart (1971) suggested that electrostatic charges in fluidized beds are mostly generated by the motion of particles around bubbles, and showed that the degree of charging in a bed of glass beads particles of 200-300 µm mean diameter increased as the bubble size increased. Yao et al. (2002) studied the local instantaneous electrostatic charges in a 0.089 m inner diameter Plexiglas fluidized bed of polyethylene resin particles with the aid of a collision ball probe. Their results again indicated that electrostatic charges in a bubbling fluidized bed are related to the bubble behaviour. It has also been reported by Chen et al. (2003a) that in bubbling fluidized beds, particle-particle collisions among particles surrounding bubbles tend to be much more energetic than for particles elsewhere in the bed, potentially leading to higher charge generation there. From this earlier work it appears likely that the increase in the degree of bed electrification experienced in our elevated-pressure experiments is probably due to the increase in bubble rise velocity, frequency, volume fraction and a slight increase in  Standard deviation (μC) .  bubble size at elevated pressures, as reviewed in Chapter 1. 0.20  (b)  (a) 0.15  C B  0.10 0.05  A' C' B'  A  Probe A' Probe B' Probe C'  Probe A Probe B Probe C  0.00 350  450  550  650  Pressure (kPa)  750  350  450  550  650  750  Pressure (kPa)  Figure 3.13 Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  59  Chapter 3: Results and Discussion: Effect of Operating Parameters  Current (nA) .  -10  Probe A Probe B Probe C  (a)  -7.5  Probe A' Probe B' Probe C'  (b)  C  A'  B  -5  C'  A  B'  -2.5 0 350  450  550  650  750  350  450  Pressure (kPa)  550  650  750  Pressure (kPa)  Standard deviation (μC) .  Figure 3.14 Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 1317% and U-Umf = 0.05 m/s, with probe locations interchanged from those in Figure 3.11. 0.20  B  (a)  (b)  A'  0.15  B'  C  A  0.10 0.05  C' Probe A' Probe B' Probe C'  Probe A Probe B Probe C  0.00 350  450  550  650  Pressure (kPa)  750  350  450  550  650  750  Pressure (kPa)  Figure 3.15 Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of GB-L particles, dp = 574 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s with probe locations interchanged from those in Figure 3.13. Probe A’ indicates that at elevated pressures, particle-particle and particle-wall near the distributor and the wall contributed heavily to static charge generation in the bed, as shown in Figures 3.10f, 3.11b, 3.14b and 3.16b. This can also be attributed to the increase in bubble frequency, in the lower portion of the bed, as pressure increases. Except at the lowest level, 60  Chapter 3: Results and Discussion: Effect of Operating Parameters  the electrostatic charge generation was found to be greater near the column axis than near the wall, as shown by comparing the probe B and C results in Figures 3.10a, 3.10b, 3.11a, 3.14a and 3.16a. Hoffaman and Yates (1986) reported that at atmospheric pressure bubbles are more evenly distributed across the bed, but as pressure increases the bubbles are increasingly concentrated toward the vertical axis, indicating an increase in bubble coalescence. This is consistent with the results measured by probes B and C in our elevated-pressure experiments.  Current (nA) .  -8  Probe A Probe B Probe C  (a)  -6  Probe A' Probe B' Probe C'  (b)  C  C'  B  -4 -2  B'  A  A'  0 350  450  550  650  750  350  450  550  Pressure (kPa)  650  750  Pressure (kPa)  Figure 3.16 Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of GB-M particles, dp = 365 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  Standard deviation (μC) .  0.2  (b)  (a) 0.15  C'  B  0.1  C  A'  A  0.05  B'  Probe A Probe B Probe C  0 350  450  550 Pressure (kPa)  650  750  Probe A' Probe B' Probe C' 350  450  550  650  750  Pressure (kPa)  Figure 3.17 Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of GB-M particles, dp = 365 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9. 61  Chapter 3: Results and Discussion: Effect of Operating Parameters  From the experimental results for probe C presented in Figures 3.10a, 3.11a, 3.14a and 3.16a, it would appear that the maximum static charge was reached at a height of 0.55 m where probe C is located, which is where the maximum bubble size occurs. This height is approximately equal to the static bed height (0.53 m). This observation agrees with those of previous workers (Ciborowski and Woldarski, 1962, Guardiola et al., 1992 and Chen et al., 2003a), who found that a height above distributor equal to the static bed height corresponded to the highest charge build-up in gas-solid fluidized beds. Figures 3.13, 3.15 and 3.17 show that charging stability and charge induction in the bed may have become more stable or even slightly decreased, after an initial increase, as pressure increased. This could be attributed to the fact that pressurized fluidized beds exhibit smoother and more stable fluidization. It can be seen in Figure 3.18 that the standard deviation of the pressure fluctuation across the bed slightly decreased as pressure increased, indicating smoother and more stable fluidization. Although the bed particles became more energetic at elevated pressures potentially leading to generation of more electrostatics in the bed, the induced charge seems to have been influenced by the stability of fluidization rather  Standard deviation (kPa)..  than by the charge generated by energetic particles in the bed.  0.06  0.05  0.04 350  450  550 650 Pressure (kPa)  750  Figure 3.18 Standard deviation of differential pressure fluctuations across a bed of GB-M particles as a function of pressure, dp = 365 µm,T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s.  62  Chapter 3: Results and Discussion: Effect of Operating Parameters  3.2.1.2  Polyethylene resin  Freely bubbling electrification experiments were also conducted in the elevated-pressure column with collision ball probes at four vertical levels and at two radial positions, as shown in Figure 2.9, with polyethylene resin particles (HDPE) of 450 µm Sauter mean diameter and 965 kg/m3 density. (For physical properties of the HDPE particles, see Table 2.2.) The static bed height was ~0.53 m. Freeboard absolute pressures were varied from 101 to 724 kPa. Umf was found experimentally to be 0.137, 0.110, 0.104, 0.095 and 0.088 m/s for freeboard absolute pressures of 101, 379, 448, 587 and 724 kPa, respectively. Temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%) in order to isolate the effect of pressure. Excess gas velocity was also maintained constant at U-Umf = 0.05 m/s. The bed was fluidized for about an hour to achieve steady state before collecting data. The net cumulative charge generated in the bed, measured by the collision ball probes, is plotted versus time in Figure 3.19. Current flow through the ball probes due to direct charge transfer is plotted versus pressure in Figure 3.20, while charge fluctuation components are indicated in Figure 3.21 by plotting the standard deviation of charge fluctuations versus pressure. Figures 3.19-3.21 show that the degree of electrification in the bed of HDPE resin particles increased as pressure increased, probably for the same reasons as discussed above. However, the HDPE particles were oppositely charged at different locations demonstrating bi-polar charging. This can be attributed to the influence of the physical properties of the particles. Probe C and C’ results in Figures 3.19a, 3.19b and 3.20 show that the HDPE particles were positively charged at the highest level in the bed. This is discussed in Chapter 4. Results shown in Figure 3.21 indicate that the charge fluctuation generally increases as pressure increases. Although pressurized fluidized beds exhibit smoother and more stable fluidization, the increase in charge fluctuations in the bed could be due to the occurrence of more than one charging mechanism in the bed. This point is discussed in more detail in Chapter 4.  63  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (nC)  30  Cumulative charge (nC)  724 586  15  Probe C  448  0  (b)  Probe C' 724  586  379 101  101  379  448  -15 -30 30  (c)  Probe B  (d)  15  Probe B' 448  379  586  101  101  0 -15  724  448  586  724  379  -30 30  Cumulative charge (nC)  (a)  (e)  Probe A  15  101  (f)  448 379  379  0  -30 0  101  448  586  -15  724  724 100  200 Time (s)  Probe A'  586 300  400  0  100  200 300 Time (s)  400  Figure 3.19 Effect of pressure on net cumulative charge as a function of time at different axial and radial positions in a bed of HDPE particles, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on the curves are absolute pressures (in kPa) at top of column. For probe positions, see Figure 2.9.  64  Chapter 3: Results and Discussion: Effect of Operating Parameters 0.2  Current (nA) .  Probe A' Probe B' Probe C'  (b)  (a) 0.1  C  C'  0  A  -0.1 -0.2 50  190  330  A'  B'  Probe A Probe B Probe C  B 470  610  750  50  190  Pressure (kPa)  330  470  610  750  Pressure (kPa)  Figure 3.20 Current flow through ball probes as a function of pressure at different axial and radial positions in a bed of HDPE particles, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  Standard deviation (nC) .  6  Probe A Probe B Probe C  (a)  Probe A' Probe B' Probe C'  (b)  C  4 B  2  B'  C'  A  A'  0 50  190  330  470  610  750  Pressure (kPa)  50  190  330  470  610  750  Pressure (kPa)  Figure 3.21 Standard deviation of charge fluctuations as a function of pressure at different axial and radial positions in a bed of HDPE particles, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9. The Faraday cup described schematically in Figure 2.5 was also utilized for the polyethylene particles to measure particle charge density at elevated pressures by withdrawing samples from the same ports A, B and C at z = 0.15, 0.31 and 0.55 m, respectively, as were utilized for insertion of ball probes A, B and C. The charge density of the particles sampled from the ports is plotted versus pressure in Figure 3.22.  65  Particle charge density (nC/g).  Chapter 3: Results and Discussion: Effect of Operating Parameters 26 C  15  Port A Port B Port C  4 -7  B A  -18 350  450  550  650  750  Pressure (kPa)  Figure 3.22 Charge density of HDPE particles measured by Faraday cup, dp = 450 µm, T = 20 ± 2ºC, RH = 13-17%, and U-Umf = 0.05 m/s, particles sampled from ports A, B and C at z = 0.15, 0.31 and 0.55 m, respectively. These Faraday cup results indicate that the magnitude of the charge density increased almost linearly as the freeboard absolute pressure increased. Particles at level C (upper bed) are charged positively, opposite to those at probes A and B (lower levels)  3.2.2  Effect of operating temperature  In order to investigate the effect of temperature on electrostatic charge buildup in a fluidized bed, freely bubbling experiments were conducted in the three-dimensional elevated-pressure column with collision ball probes at four levels (z = 0.15, 0.31, 0.55 and 0.89 m), and at two radial positions as shown in Figure 2.9. Particles were polyethylene (PE-SLD) of mean diameter 642 µm. (For physical properties of the PE-SLD particles, see Table 2.2.) The static bed height was again ~0.53 m. Pressure and the excess gas velocity, (U-Umf) were maintained nearly constant (P = 379 kPa and U-Umf = 0.05 m/s) to isolate the effect of bed temperature. The relative humidity of the fluidizing air was in the range of 17 ±10%; it was difficult to control the relative humidity within a narrower range. Umf for PE-SLD was found experimentally to be 0.10 m/s for this combination. The bed was operated briefly at 85ºC by means of an electrical heating tape. (The maximum bed temperature to avoid PE-SLD particle softening for steady state measurements, requiring ~1 h to achieve steady state, was 60ºC.) The net cumulative charge in the bed was then measured by the collision ball probes 66  Chapter 3: Results and Discussion: Effect of Operating Parameters  and plotted versus time in Figure 3.23. Figure 3.24 shows the effect of bed temperature on the rate of direct charge transfer through the ball probes. Figure 3.25 was obtained by plotting standard deviation of the charge fluctuation curves after filtering the net cumulative charge curves shown in Figure 3.23 for all probes versus temperature using a high-pass filter of cut-off frequency 10*(1/period) where the period is the data sampling range. Newton et al., (2001) studied the effect of temperature on bubble behaviour in gas-solid fluidized beds and reported that increasing temperature from ambient to 85ºC at a constant excess gas velocity resulted in a decrease in bubble frequency from 350/min at 15ºC to 100/min at 85ºC. Increasing the temperature, on the other hand, resulted in bubbles becoming larger for a typical linear low-density polyethylene resin. Our experimental results, depicted in Figures 3.23-3.25, indicate that the bed exhibited smoother fluidization at higher temperatures. This was also supported by visual observation of the fluidized bed behaviour through sight glasses. As shown in Figures 3.23a, b, c and 3.24a bed temperature appeared to play a significant role in determining the degree of bed electrification. As the temperature increased, the magnitude of the electrostatic charges decreased at probes A, B, C and C’, probably due to smaller, slower bubbles as temperature increased. It can also be seen that the effect of temperature was more significant at the axis of the vessel than near the wall. The influence of temperature on the output signals of the collision ball probes was also investigated. The experiments were conducted by introducing air into an empty column with ball probes at four levels in the bed, and at two radial positions as shown in Figure 2.9. Air was introduced into the column at a superficial velocity of 0.15 m/s. The freeboard absolute pressures were maintained constant at 379 kPa, while the temperature in the bed was increased from ambient to 70ºC and then allowed to fall towards ambient. The output signals of the ball probes and temperature are plotted versus time in Figure 3.26. The output signals of the ball probes were found to be insensitive to the change in bed temperature.  67  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (nC)  8  Cumulative charge (nC)  Probe C  4  (b)  Probe C'  20  20 60  60  0 -4 -8  8  (c)  Probe B  (d)  Probe B'  4 60  0 -4  60 20  20  -8 8  Cumulative charge (nC)  (a)  (e)  Probe A  (f)  Probe A'  4  60  0  60  20  20  -4 -8 0  100  200 Time (s)  300  400  0  100  200  300  400  Time (s)  Figure 3.23 Effect of temperature on net cumulative charge as a function of time for PESLD particles, dp = 642 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. Numbers on the curves are bed temperature (in ºC). For probe positions, see Figure 2.9.  68  Chapter 3: Results and Discussion: Effect of Operating Parameters  Current (pA) .  20  Probe A Probe B Probe C  (a)  10  0  Probe A' Probe B' Probe C'  (b)  0  C  0  0  A' B'  A -10  C'  0  B  -20  0 10  25  40 T (C)  55  70  10  25  40 T (C)  55  70  Figure 3.24 Current flow through ball probes as a function of bed temperature at different axial and radial positions in a bed of PE-SLD particles, dp = 642 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  Standard deviation (nC)  2.0  Probe A Probe B Probe C  (a)  1.5  3  Probe A' Probe B' Probe C'  (b)  2  1.0  B  1  C  0.5  B'  A  C'  A'  0.0  0  10  25  40 T (C)  55  70  10  25  40 T (C)  55  70  Figure 3.25 Standard deviation of charge fluctuations as a function of temperature at different axial and radial positions in a bed of PE-SLD particles, dp = 642 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9. In order to investigate the effect of temperatures beyond 60ºC, freely bubbling temperature experiments were conducted by fluidizing a bed of mono-size glass beads particles (GB-XL) of mean diameter 693 µm at the same operating conditions and following the same temperature experimental procedure as described above. (Physical properties of the GB-XL particles are given in Table 2.1.) Umf for GB-XL was found experimentally to be 0.375 m/s 69  Chapter 3: Results and Discussion: Effect of Operating Parameters  for freeboard absolute pressures of 379 kPa. The excess gas velocity was again maintained constant in all experiments (at U-Umf = 0.05 m/s) to isolate the effect of bed temperature. In these experiments the net cumulative charge in the bed was measured by the collision ball probes while increasing the bed temperature through four temperatures ranges (25-30, 40-45, 65-70, 85-90ºC) and plotted versus time in Figure 3.27. For each temperature range, the bed was fluidized for ~1 h to achieve steady state before collecting data. After collecting the data at 85-90ºC all heating tapes were turned-off, and the net cumulative charge in the bed was again measured for the same four temperature ranges (90-85, 70-65, 45-40, 30-25ºC) while the bed cooled from 90ºC to ambient and plotted versus time in Figure 3.28. 75 heating from ambient to 70 ºC Temp  cooling from 70 to 35ºC C  0.15 C' A  0.00 B' -0.15 (a)  B 45  Temp (ºC)  Cumulative charge (pC) ...  0.30  A'  (b)  -0.30  15 0  3000  6000  Time (s) Probe A  Probe B  Probe C  Probe B'  Probe C'  Temp  Probe A'  Figure 3.26 Output signals of ball probes as a function of time and temperature in an empty column with air flow rate at 0.15 m/s and temperature range of (a) from ambient to 70ºC and (b) from 70ºC to 32ºC. For probe positions, see Figure 2.9.  70  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (μC)  Cumulative charge (μC)  0.4  Probe C  0.2  Probe C' 85-90 65-70  85-90  -0.2 -0.4 0.4  (b)  65-70  0.0  40-45  25-30  25-30 40-45  (c)  Probe B  85-90  0.2  (d)  Probe B'  65-70  0.0  40-45  -0.2  25-30  85-90  65-70  40-45 25-30  -0.4 0.4  Cumulative charge (μC)  (a)  (a)  Probe A  0.2  -0.4 0  Probe A'  65-70  85-90  85-90  0.0 -0.2  (f)  25-30  100  35-40  200 Time (s)  300  25-30 400  0  100  200  65-70  40-45 300  400  Time (s)  Figure 3.27 Effect of temperature on net cumulative charge as a function of time when the bed temperature was increased from ambient to 90ºC for GB-XL particles, dp = 693 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. Numbers on the curves are bed temperature (in ºC). For probe positions, see Figure 2.9.  71  Cumulative charge (μC)  Cumulative charge (μC)  Chapter 3: Results and Discussion: Effect of Operating Parameters  0.4  Probe C  0.2  (b)  Probe C'  70-65  90-85 70-65  0.0 45-40  -0.2  30-25  90-85  45-40  30-25  -0.4  0.4  (c)  0.2  90-85 45-40 70-65  (d)  Probe B' 70-65  30-25  0.0  90-85  45-40 30-25  -0.2 Probe B  -0.4 0.4  Cumulative charge (μC)  (a)  (e)  Probe A  0.2  90-85  (f)  Probe A' 70-65 90-85  70-65  0.0 30-25  45-40 -0.2 -0.4 0  45-40  30-25 100  200 Time (s)  300  400  0  100  200  300  400  Time (s)  Figure 3.28 Effect of temperature on net cumulative charge as a function of time when the bed temperature was cooled from 90ºC to ambient for GB-XL particles, dp = 693 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. Numbers on the curves are bed temperature (in ºC). For probe positions, see Figure 2.9.  Other freely bubbling temperature experiments were conducted by fluidizing a bed of monodisperse glass beads (GB-XL) of mean diameter 693 µm. In these experiments the bed was 72  Chapter 3: Results and Discussion: Effect of Operating Parameters  fluidized for ~2.75 h. The bed temperature was raised from ambient to 70ºC and then allowed to cool (with the tape heaters off). The excess gas velocity was again maintained constant in all experiments at U-Umf = 0.05 m/s to isolate the effect of bed temperature. The temperature and net cumulative charges in the bed measured by ball probes are plotted versus time in Figure 3.29.  B  0.5  75  (b) B'  C' A' Temp  A  0.0  C  Temp  45 30  -0.5 -1.0 0  60  2500  5000  7500 10000  0  Time (s)  2500  5000  o  (a)  Temperature ( C)  Cumulative Charge (μC)  1.0  15 7500 10000  Time (s)  Figure 3.29 Effect of temperature on net cumulative charge as a function of time and temperature when the bed temperature is raised from ambient to 70ºC and then allowed to cool to ambient again for GB-XL particles, dp = 693 µm, P = 379 kPa, RH = 7-27% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9. The results in Figures 3.27-3.29 indicate a change in charge polarity as the bed temperature increased. This occurred in all locations tested. These results are consistent with those reported by Zimmer (1970) who studied the charging of plastics when they were rubbed against a metal brush and found that for the polymers tested it was possible to change the polarity of the transferred charge by increasing the temperature. In practice, frictional charging between particles occurs in air rather than in a vacuum. The high field between the surfaces as they separate can cause the air to ionize. The ions neutralize the surface charge, but tend to over-compensate with the result that the portion of the surface where the discharge occurs acquires opposite polarity to the initial frictional charge. The surface may thus end up with areas of different charge (Hull, 1949).  73  Chapter 3: Results and Discussion: Effect of Operating Parameters  On the other hand, as the bed temperature increased, the relative humidity of the fluidizing air decreased because the saturated vapor pressure of moisture in the bed increased with increasing temperature. Coste and Pechery (1977) investigated the effect of moisture content of a polymer. They passed polymer films over metal rollers, one of which could rotate at a speed which differed from that of the film, to provide a source of friction. In their experiments, three distinct types of behaviour were found, as illustrated in Figure 3.30: (i) The charge density increased to a constant value as shown in Figure 3.30a. (ii) The charge density passed through a maximum, then decreased and changed polarity before reaching a new equilibrium state as shown in Figure 3.30b. (iii) The charge density reached a maximum, and then decreased to a new steady state without changing polarity as shown in Figure 3.30c. These results can be interpreted in terms of both the change in temperature of contact surfaces and its tendency to adsorb water.  Charge  (a)  0  (b)  Time  0  (c)  Time  0  Time  Figure 3.30 Types of behaviours observed for build-up of charge density as a function of time (adapted from Coste and Pechery, 1977) Williams (1964) investigated the charge acquired by glass beads as they rolled down a metal plate coated with a polymer film. They measured the charge transferred as a function of resistivity of both the glass beads and polymer, and found a relationship between the charge transfer and the resistivity of the contact surfaces. However, in our experiments it is likely that temperature would affect the electrical properties of particles as well as the wall, because the conductivity of most materials decreases as temperature increases. Hence, temperature seems to play a significant role in determining the charge polarity of the charged surfaces by influencing the physical and electrical properties of the particles, as well as the properties of the wall, leading to different charging mechanisms between the contact surfaces.  74  Chapter 3: Results and Discussion: Effect of Operating Parameters  3.2.3 3.2.3.1  Effect of superficial gas velocity Glass beads  Freely bubbling experiments were next conducted in the elevated-pressure column with collision ball probes at four levels above the distributor, and at two radial positions as shown in Figure 2.9 to study the influence of superficial gas velocity on the electrostatic charge buildup. The particles in these experiments were mono-disperse smooth spherical glass beads (GB-XL and GB-M) of volume-weighted mean diameter 693 and 365 µm, respectively. (For physical properties of these particles, see Table 2.1.) The static bed height was again ~0.53 m, and the freeboard absolute pressure was maintained at 379 kPa. Umf for GB-XL and GBM particles was found experimentally to be 0.38 and 0.23 m/s, respectively, for this pressure. Temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%) in order to isolate the effect of gas velocity. The bed was fluidized for about an hour to achieve steady state before collecting data. The excess gas velocity, U-Umf, was then varied from 0.03 to 0.12 m/s. Figures 3.31 and 3.32 plot the change in net cumulative charges at different axial and radial locations in beds of GB-XL and GB-M, respectively, versus time. Figure 3.33 shows the effect of the superficial gas velocity on the rate of direct charge transfer through the ball probe in the bed of GB-XL, whereas Figure 3.34 shows the effect on the standard deviation of the charge fluctuations. Similarly, Figure 3.35 depicts the effect of the superficial gas velocity on the rate of direct charge transfer through the ball probe in GBM particles, whereas Figure 3.36 plots the effect on the standard deviation of the charge fluctuations. These experimental results show that the degree of electrification in the bed increased as the superficial gas velocity increased. Most importantly, the rate of this increase and its significance depended on the particle size distribution and location in the bed.  75  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (μC)  -0.6  Cumulative charge (μC)  (b)  Probe C  -0.4  Probe C'  0.12  0.09  0.07 0.05  -0.2  0.12 0.07  0.09  0.05  0.03  0.03  0.0  -0.6  (c)  (d)  Probe B  -0.4  Probe B'  0.12  0.09  0.07 0.03  0.12  0.09  -0.2  0.07  0.05 0.05  0.0 -0.6  Cumulative charge (μC)  (a)  (e)  0.09  Probe A 0.12  -0.4  0.03  Probe A' 0.07  0.12  0.07  0.05 -0.2  (f)  0.05 0.09  0.03  0.03  0.0 0  100  200  300  Time (s)  400  500  0  100  200  300  400  500  Time (s)  Figure 3.31 Effect of superficial gas velocity on electrostatic charges buildup in a bed of GB-XL particles as a function of time at different axial and radial positions, dp= 693 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote UUmf (in m/s). For probe positions, see Figure 2.9.  76  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (μC)  -0.6  Cumulative charge (μC)  0.09  -0.4  0.12  0.07  (b)  Probe C' 0.12 0.09  0.05 -0.2  0.07  0.03 0.05 0.03  Probe C  0.0  -0.6  (c)  0.09  (d)  0.07  0.12  -0.4  0.07  0.12  0.09  0.05  0.03 0.05  -0.2  0.03  Probe B  0.0  -0.6 Cumulative charge (μC)  (a)  (e)  Probe B'  (f)  0.12 0.09 0.07  0.12  0.09 0.07  -0.4 0.05  -0.2 0.0 0  0.03 100  200  0.05  Probe A 300 400 500  Time (s)  0.03 Probe A' 0  100  200  300  400  500  Time (s)  Figure 3.32 Effect of superficial gas velocity on electrostatic charges buildup in a bed of GB-M particles as a function of time at different axial and radial positions, dp= 365 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote UUmf (in m/s). For probe positions, see Figure 2.9.  77  Chapter 3: Results and Discussion: Effect of Operating Parameters  Current (nA) .  -2  Probe A Probe B Probe C  (a)  -1.5  (b) B  A  -1  A' B'  C  -0.5  Probe A' Probe B' Probe C'  C'  0 0.02  0.05  0.08  0.10  0.13  0.02  Excess gas velocity, U-Umf (m/s)  0.05  0.08  0.10  0.13  Excess gas velocity, U-Umf (m/s)  Figure 3.33 Current flow through ball probes as a function of excess gas velocity, at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9.  Standard deviaion (μC) .  0.10  Probe A Probe B Probe C  (a) 0.07  B  A  0.03  Probe A' Probe B' Probe C'  (b)  C  A' B'  C' 0.00 0.02  0.05  0.08  0.10  0.13  Excess gas velocity, U-Umf (m/s)  0.02  0.05  0.08  0.10  0.13  Excess gas velocity, U-Umf (m/s)  Figure 3.34 Standard deviation of charge fluctuations as a function of excess gas velocity, at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9.  78  Chapter 3: Results and Discussion: Effect of Operating Parameters  Current (nA) .  -2  (b)  B  (a) C  -1.5 A  A'  -1  B'  -0.5  C'  Probe A Probe B Probe C  0 0.02  0.05  0.08  0.10  0.13  0.02  Excess gas velocity, U-Umf (m/s)  0.05  Probe A' Probe B' Probe C' 0.08  0.10  0.13  Excess gas velocity, U-Umf (m/s)  Figure 3.35 Current flow through ball probes as a function of excess gas velocity at different axial and radial positions for GB-M particles, dp = 365 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9.  Standard deviation (μC) .  0.10  (b)  (a) B C  0.07  A  B'  A'  0.03  0.02  Probe A' Probe B' Probe C'  Probe A Probe B Probe C  0.00 0.05  0.08  0.10  Excess gas velocity, U-Umf (m/s)  0.13  C'  0.02  0.05  0.08  0.10  0.13  Excess gas velocity, U-Umf (m/s)  Figure 3.36 Standard deviation of charge fluctuations as a function of excess gas velocity at different axial and radial positions for GB-M particles, dp = 365 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9. The influence of the superficial gas velocity on the output signals of the collision ball probes was also investigated. The experiments were conducted by introducing air into an empty column with ball probes at four levels in the bed, and at two radial positions as shown in Figure 2.9. The freeboard absolute pressures were maintained constant at 379 kPa, while the temperature and relative humidity of the air were nearly constant (T = 20 ±2ºC and RH = 25 ±2%). Air was introduced to the column at a superficial velocity of 0.35 m/s for ~1000 s, and 79  Chapter 3: Results and Discussion: Effect of Operating Parameters  then increased to 0.4 m/s for a further 1000 s. The output signals of the ball probes are plotted versus time in Figure 3.37. The output signals of the ball probes were found not very sensitive to the small change in fluidizing air velocity.  Cumulative charge (pC) .  0.30  (a)  (b)  0.15 0.00 -0.15 -0.30 0 Probe A Probe A'  1000 Probe B Probe B'  Time (s) Probe C Probe C'  2000 Probe D  Figure 3.37 Cumulative charge as a function of time in an empty column with air flow rate at (a) 0.35 m/s and (b) 0.4 m/s. For Geldart group B particles, the bed transforms from a fixed bed to a bubbling fluidized bed when the gas velocity is increased beyond the minimum fluidization velocity. As the gas velocity is further increased, bubbles grow bigger owing to coalescence, and the bubbling bed can become a slugging bed if the bed diameter is small and the particle diameter is large or a turbulent bed if the bed diameter is large and the particle diameter is small. Geldart (1972) found that the mean bubble size depends on the type of the distributor, the distance above the distributor and the excess gas velocity above that required for minimum fluidization. Mori and Wen (1975) assumed that all gas above the minimum fluidization velocity goes to form a single train of bubbles rising along the vertical axis of the bed and estimated the diameter of the bubbles that would exist at any height in the bed. All equations for the diameter of bubbles in fluidized beds indicate that the average bubble diameter increases monotonically with increasing gas velocity (Davidson and Harrison, 1963; Geldart, 1972; Mori and Wen, 1975; Rowe, 1976; Darton, 1977; Choi, 1998). 80  Chapter 3: Results and Discussion: Effect of Operating Parameters  Thus the increase in the levels of bed electrification with increasing the superficial gas velocity observed in our experiments can be explained by the formation of bigger bubbles that cause higher bubble rise velocities and consequently enhanced motion of particles in the bed, leading to increased electrostatic charge generation. For larger particles (GB-XL), Figures 3.31e, 3.31f, 3.33 and 3.34 show that as the superficial gas velocity increased the maximum electrostatic charge was reached in the region close to the distributor plate, probably due to the increased rate of bubble formation there. Guardiola et al. (1996) investigated the influence of fluidizing gas velocity on bed electrification. They reported that the frequency of bubble formation at the distributor was nearly independent of the air flow rate and they attributed the increase in the observed degree of bed electrification to the increase in bubble size. The differences between their findings and ours could be due to differences between their experimental set-up and ours. Our set-up allows us to measure the degree of electrification locally, while theirs did not allow this. In their experiments they used a capacitance probe technique to characterize the degree of bed electrification, with the probe and distributor considered to be parallel plates, and the bed acting as a dielectric media. The probe-to-distributor voltage drop was then measured. This method averages the effect of electrostatic charges over most of the bed. In our experiments, on the other hand, we were able to measure the net cumulative charge at different locations in the bed. Another main difference between their experimental set-up and ours is that the inner diameter of their column was relatively small (44 mm), so that they could easily reach slugging fluidization with only a small increase in gas velocity. For smaller glass beads GB-M (365 µm), it was found that as the excess gas velocity increased beyond 0.1 m/s, the rate of charge transfer became less sensitive to the superficial gas velocity in the lower portion of the bed and near the wall as indicated by the probe A’ results in Figure 3.35b. It was also observed that the rate of charge transfer became insensitive to the superficial gas velocity at higher levels and near the vertical axis as indicated by the probe B and C results in Figures 3.32a, c and 3.35a. This could be due to reduced contact between particles and ball probe in the higher portion of the bed and near the vertical axis due to lower concentrations of particles there. It could also be due to the reduction in the contacting time between the probes and dense particle phase. Yao et al. 81  Chapter 3: Results and Discussion: Effect of Operating Parameters  (2002) investigated the effect of gas velocity on electrification in a three-dimensional fluidization column of 0.089 m inner diameter and 1.2 m height using a collision ball probe 0.3 m above the distributor in polyethylene particles of mean diameter 378 and 318 µm. They reported that the rate of charge transfer is insensitive to the superficial gas velocity because the increase in charge transfer for higher particle-probe collision velocities is balanced by a reduction in contacting time between the probe and dense particle phase. The difference between their results and ours is that our results allow us to gain a better understanding of how the electrostatic charges change at different locations in the bed.  3.2.3.2  Polyethylene resin  Similar freely bubbling electrification experiments were also conducted in the threedimensional elevated-pressure column for a bed of polyethylene resin particles (HDPE) of 450 µm mean diameter and 965 kg/m3 particle density. Physical properties of these particles appear in Table 2.2. The static bed height was again ~0.53 m and freeboard absolute pressure was constant at 379 kPa. Umf for HDPE particles was found experimentally to be 0.11m/s for a freeboard absolute pressure of 379 kPa. The temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%) in order to isolate the effect of gas velocity. The bed was again fluidized for about an hour to achieve steady state before collecting data. The excess gas velocity, U-Umf, was then varied from 0.03 to 0.12 m/s. The net cumulative charges generated in the bed, measured by the collision ball probes, are plotted versus time in Figure 3.38. Current flow through the ball probes due to direct charge transfer is plotted versus excess gas velocity in Figure 3.39, while Figure 3.40 shows the standard deviation of the charge fluctuations plotted versus excess gas velocity. It can be seen from these experimental results that as the superficial gas velocity increased the degree of electrification in the bed increased, presumably due to enhanced particle motion caused by the increase in bubble size and rise velocity.  82  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (nC)  30  Cumulative charge (nC)  Probe C  0.12  15  0.08  0  (b)  Probe C'  0.12 0.08  0.05  0.05  0.03  0.03  -15 -30 30  (c)  (d)  Probe B  Probe B'  15  0.03  0 0.05 -15  0.03  0.05 0.08  0.12  0.12 0.08  -30  30 Cumulative charge (nC)  (a)  15  (e)  -30 0  Probe A'  0.03  0 -15  (f)  Probe A  0.03  0.05  0.05  0.12 0.08 100  200 Time (s)  0.12  300  400  0  100  0.08 200  300  400  Time (s)  Figure 3.38 Effect of superficial gas velocity on electrostatic charges buildup in a bed of HDPE particles as a function of time at different axial and radial positions, dp= 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote UUmf (in m/s). For positions of probes, see Figure 2.9.  83  Chapter 3: Results and Discussion: Effect of Operating Parameters  The HDPE particles were found to be oppositely charged at different locations in the bed demonstrating bi-polar charging. This can be again attributed to the influence of physical properties of the particles. Figures 3.38a, 3.38b and 3.39 show that as the superficial gas velocity increased, HDPE particles were positively charged at z = 0.55 m in the bed where probes C and C’ were immersed. This phenomenon is discussed in Chapter 4.  Current (nA) .  0.3  (a)  0.15  Probe A Probe B Probe C  Probe A' Probe B' Probe C'  (b) C  0  C'  A  B'  -0.15  A'  B -0.3 0.02 0.05 0.08 0.10 0.13 Excess gas velocity, U-Umf (m/s)  0.02  0.05 0.08 0.10 0.13 Excess gas velocity, U-Umf (m/s)  Figure 3.39 Current flow through ball probes as a function of superficial gas velocity, at different axial and radial positions for HDPE particles, dp = 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9.  Standard deviation (nC) .  3 (a)  (b) A'  B 2  C'  C  1  Probe A Probe B Probe C  0 0.02  B'  A  0.05  0.08  0.10  Excess gas velocity, U-Umf (m/s)  0.13  Probe A' Probe B' Probe C' 0.02  0.05  0.08  0.10  0.13  Excess gas velocity, U-Umf (m/s)  Figure 3.40 Standard deviation of charge fluctuations as a function of superficial gas velocity, at different axial and radial positions for HDPE particles, dp = 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For probe positions, see Figure 2.9.  84  Chapter 3: Results and Discussion: Effect of Operating Parameters  Cumulative charge (nC) .  4  Probe D  3 2  0.12 0.08  1  0.05 0.03  0 -1 0  100  200  300  400  Time (s)  Figure 3.41 Electrostatic charges buildup with time in the freeboard at different superficial gas velocities in a bed of HDPE particles, dp= 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. Numbers on curves denote U-Umf (in m/s). For position of probe D, see Figure 2.9. On the other hand, as shown in Figure 3.41 at higher superficial gas velocities the polarity in the freeboard region was opposite to that at the lower measurement levels inside the bed (levels B and A), indicating that the fines entrained from the bed carried opposite charges, leaving behind a net charge of opposite polarity inside the bed. As shown in Figures 3.42 and 3.43 the rates of charge transfer and charge fluctuations in the freeboard region increased with increasing superficial gas velocity, indicating an increase in particle concentration in the freeboard. According to Cross (1987), it is often found that the amount of charge transferred during any rubbing process between similar or dissimilar materials is affected more by the velocity at the time of contact and by the energy of rubbing than by the nature of the materials. Thus the increase in the rate of charge transferred observed in our experiments could be also attributed to the increase in the velocity of particles at the time of the contact.  85  Chapter 3: Results and Discussion: Effect of Operating Parameters  10.0 Current (pA) .  Probe D 7.5 5.0 2.5 0.0 0.02 0.05 0.08 0.10 0.13 Excess gas velocity, U-Umf (m/s)  Standard deviation (nC) .  Figure 3.42 Current flow through ball probes in the freeboard in a bed of HDPE particles as a function of superficial gas velocity, dp = 450 µm, P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For position of probe D, see Figure 2.9.  0.2  Probe D 0.15 0.1 0.05 0 0.02  0.05  0.08  0.10  0.13  Excess gas velocity, U-Umf (m/s)  Figure 3.43 Standard deviation of charge fluctuations as a function of superficial gas velocity, in the freeboard in a bed of HDPE particles, dp = 450 µm, P = 379 kPa, T = 20 ±2ºC and RH = 13-17%. For position of probe D, see Figure 2.9.  86  Chapter 3: Results and Discussion: Effect of Operating Parameters  3.3  Summary  The effects of operating pressure, temperature and superficial gas velocity on the degree of electrification were investigated in this Chapter. Freely bubbling fluidization experiments indicated that as the pressure increased, the degree of electrification increased, probably due to the increases in bubble rise velocity, frequency and volume fraction. However, the magnitude of the increase differed from location to location. At higher levels electrostatic charge generation was greater near the column axis than near the wall. The maximum charges were found near the height where the maximum bubble size occurred and near the axis. As pressure increased, particle-particle and particle-wall collisions near the distributor and wall contributed heavily to electrostatic charge generation. Freely bubbling experiments with glass beads and polyethylene particles showed that the charge polarity reversed as the bed temperatures increased. At higher temperature the bed exhibited smoother fluidization. Temperature seems to play a significant role in determining charging mechanisms. The degree of electrification in the bed increased as the superficial gas velocity increased, likely due to higher bubble rise velocities. The rate of this increase and its significance depended on the particle size distribution and location. At higher superficial gas velocities the polarity in the freeboard was opposite to that in the bed, indicating that entrained fines carried charges, leaving a net charge of opposite polarity inside the bed.  87  Chapter 4: Results and Discussion: Effect of Particle Properties  Chapter 4: Results and Discussion: Effect of Particle Properties  4.1 4.1.1  Effect of Average Particle Size Mono-size glass beads  In order to investigate the effect of average particle size on electrostatic charge buildup in a fluidized bed, freely bubbling experiments were conducted in the three-dimensional elevated-pressure column described in Section 2.1 with collision ball probes at four levels (0.15, 0.31, 0.55 and 0.89 m above the distributor), and two radial positions (25 and 75 mm from the wall) as shown in Figure 2.9. The particles were mono-disperse smooth spherical glass beads GB-XL, GB-L, GB-M and GB-S of volume-weighted mean diameter 693, 574, 365, 65 µm, respectively. (For physical properties of the glass beads see Table 2.1.) The static bed height was ~0.53 m in all cases. These experiments were conducted at a freeboard absolute pressure of 379 kPa. For this pressure and room temperature, Umf for GB-XL, GB-L and GB-M was found experimentally to be 0.38, 0.31 and 0.23 m/s, respectively, whereas for GB-S it was estimated to be 0.005 m/s. (For estimation of Umf for GB-S, see Appendix D.1.) The temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%) in order to isolate the effect of particle size. The excess gas velocity, (U-Umf), was also maintained constant in these experiments at 0.08 m/s for the same purpose. The bed was fluidized at these conditions for about an hour to achieve steady state before collecting data. The net cumulative charge in the bed was then measured by the collision ball probes. The experimental results depicted in Figures 4.1 and 4.2 indicate that for Geldart group B particles the degree of electrification in the bed slightly increased with decreasing particle size, whereas charging for Geldart group A particles was significantly greater than for group B particles. These increases are probably due to the increase in total surface area of small particles, which led to increases in particle-particle and particle-wall collisions, enhancing contact charging and charge generation in the bed.  88  Cumulative charge (μC)  Cumulative charge (μC)  Cumulative charge (μC)  Chapter 4: Results and Discussion: Effect of Particle Properties  -0.6  (a)  (b)  Probe C  S  -0.4  Probe C'  M L  -0.2  M  S L  XL  XL  0.0 -0.6  (c)  (d)  Probe B  Probe B'  M -0.4 S  XL  L  -0.2  S  M L  XL  0.0 -0.6  (e)  Probe A'  M  -0.4 S  M  XL  S  L  L  -0.2 0.0 0  (f)  Probe A  XL 100  200 Time (s)  300  400  0  100  200  300  400  Time (s)  Figure 4.1 Effect of average particle diameter on net cumulative charge as a function of time at different axial and radial positions in beds of GB-XL, GB-L, GB-M and GB-S, dp = 693, 574, 365, 65 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and UUmf = 0.08 m/s. Letters on curves denote glass bead types. For probe positions, see Figure 2.9.  89  Chapter 4: Results and Discussion: Effect of Particle Properties  lCurrent l (nA ) .  100  Probe A Probe B Probe C  (a)  10  A  Probe A' Probe B' Probe C'  (b) A'  B  B'  1 C  C'  0.1 0  250  500  750  Average particle diameter, dp (μm)  0  250  500  750  Average particle diameter, dp (μm)  Figure 4.2 Current flow through ball probes as a function of average particle diameter at different axial and radial positions in the bed, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.08 m/s. For probe positions, see Figure 2.9. The specific surface areas, As, of the glass beads, determined by a Malvern Mastersizer 2000 equipped with a wet cell (see Appendix D.2), are presented in Table 4.1. Table 4.1 Specific surface area, As, of mono-disperse glass beads  Mean particle diameter dp (µm) Specific surface area, As (m2/g)  GB-S  GB-M  GB-L  GB-XL  65 0.098  365 0.017  574 0.011  693 0.0097  It can be seen from Table 4.1 that the specific surface area is ~10 times larger for GB-S particles than for the GB-XL. This explains the increase in particle-particle collisions as particle size decreased. It also accounts for the substantially higher rate of charge transfer observed in the GB-S experiments compared with the other types of glass beads tested. The higher degree of electrification observed for the group A particles in Figure 4.2 is likely due to charge transfer, caused by an increase in particle-probe collisions as the specific surface area of particles increased with decreasing particle size. In the experiments with GB-S particles, after discharging the particles from the column, a layer of fine dust was observed on the inner wall of the column. This layer, formed due to 90  Chapter 4: Results and Discussion: Effect of Particle Properties fine particles adhering to the wall, could act as an insulator, significantly reducing the dissipation of charges from particles to the grounded column, and consequently enhancing the degree of electrification in the bed. As noted by Grace (1982), for group B particles at low fluidizing velocity, particles in the bed show a gross circulation pattern with up-flow at the wall and down-flow at the bed axis in the lower portion of the column, and a second vortex ring above the original one with upflow at the axis of the bed. At higher gas velocities the flow pattern may reverse because of large rising bubbles, with particle circulation in the upper ring becoming more vigorous and dominating the overall movement of the dense particle phase. In beds of group A particles the transition to up-flow emulsion occurs much closer to Umf than in beds of group B particles. In beds of group B particles, bubbles form at the distributor, then grow and coalesce as they rise. Average bubble volume increases almost linearly with distance above the distributor and excess gas velocity, and it is roughly independent of average particle size, whereas in beds of group A particles, bubbles split and coalesce more frequently as they rise (Yang, 2003). From the above comparison between group B and A particles, it can be seen that the hydrodynamic of bubbling fluidized beds changes significantly when particle size changes from group B to group A particles. Thus the significant differences in bed electrification observed in our experiments when the range of particle size changed from 250-850 to 45-90 µm could also be due to hydrodynamics differences between A and B particles. Guardiola et al. (1996) investigated the effect of average particle size on the degree of bed electrification in a three-dimensional Perspex column of 0.044 m diameter, for nearly monodisperse glass beads of different diameters (i.e. 250-297, 297-350 and 350-420 µm). Their results showed an increase in the average electrical potential with increasing particle size. The difference between their results and ours could be due to several factors: (a) They investigated glass beads of diameters within narrow particle size distributions, while a broad particle size distribution (45-850 µm) was used in our investigation. The three particle types in their experiments were considered as a single particle size in our experiments (i.e. GB-M). (b) In their experiments they used a capacitance probe to average the overall electrical 91  Chapter 4: Results and Discussion: Effect of Particle Properties potential of the bed, while in our experiments, we measured the net cumulative charge locally at different locations in the bed, helping to gain a better understanding of the electrostatic charge generated in different regions. However, probe B’ and A’ results for GB-L and GB-XL shown in Figures 4.1d and f are similar to those of Guardiola et al. (1996). This may be attributable to the increase in charge discharge from particles to the wall of the grounded column due to the increase in particlewall contact for smaller particles. On the other hand, since the Guardiola et al. (1996) experiments were conducted in a relatively small three-dimensional column of 0.044 m inner diameter, the electrostatic charges over most of the bed could have been influenced by particle-wall contacts, explaining the similarity between their results and ours for probes B’ and A’ near the wall. The variation of charges with position observed in these experiments again demonstrates the importance of characterizing bed electrification at different locations in the bed and of differentiating between results at different locations using complementary techniques.  4.1.2  Binary mixtures of glass beads  Electrostatic charge generated due to frictional charging between particles of the same materials, but different sizes, was also investigated by conducting free bubbling experiments in the three-dimensional elevated-pressure column with collision ball probes at four levels, and at two radial positions as shown in Figure 2.9. The particles were binary mixtures, containing different proportions of GB-XL and GB-S as indicated in Table 4.2. The static bed height was ~0.53 m. All these experiments were conducted at a freeboard absolute pressure of 379 kPa and a static bed height of ~0.53 m. The minimum fluidization velocities for the mixtures were estimated by the Goossens et al. (1971) correlation. The temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%) in order to isolate the effect of particle size. Excess gas velocity was also maintained constant (U-U(mf)M = 0.08 m/s).  92  Chapter 4: Results and Discussion: Effect of Particle Properties Key physical properties of the mixtures are also provided in Table 4.2. The densities of the particles in the mixtures were provided by the suppliers. The size distribution, volumeweighted mean diameter and specific surface area of the mixtures were obtained by a Malvern Mastersizer 2000 equipped with a wet cell (see Appendix D.5 for particle size distributions.) The bulk densities of the mixtures were measured to determine the loosely packed voidage, εb, of the particles. Procedures similar to those described in Section 2.4.1 were again followed to determine bulk densities. Table 4.2 Properties of binary particle mixtures. All particle densities are 2500 kg/m3 Mixture wt%  XL M1 M2 M3 M4 M5 M6 S  Vol.-  GBXL  GB-S  d(p)M (µm)  Particle bulk density, ρb (kg/m3)  100 90 80 70 30 20 10 0  0 10 20 30 70 80 90 100  693 516 509 450 192 151 115 65  1573 1566 1583 1592 1601 1589 1603 1596  weighted mean dia.  Loosely packed voidage, εb  Specific surface area (m2/g)  U(mf)M at 379 kPa (m/s)  0.371 0.374 0.367 0.363 0.360 0.364 0.359 0.362  0.0097 0.0313 0.0305 0.0417 0.0806 0.0857 0.0911 0.0980  0.375 0.278 0.213 0.156 0.037 0.023 0.014 0.005  The bed was fluidized for about an hour to achieve steady state before collecting data. The net cumulative charge in the bed was then measured by the collision ball probes. They are plotted versus time in Figures 4.3 to 4.5. Current flow through the ball probes due to direct charge transfer is plotted versus the wt% of GB-S particles in the mixtures in Figures 4.6 and 4.7, while the charge fluctuations are reflected in Figures 4.8 and 4.9 by plotting the standard deviation of charge fluctuations versus the wt% of GB-S particles in the mixtures. The charge fluctuations were obtained by filtering the net cumulative charge curves shown in Figures 4.3 to 4.5 using a high-pass filter of cut-off frequency of 10*(1/period) where the period is the data sampling range. Figure 4.10 shows an example of the charge fluctuations obtained by filtering the net cumulative charge curves shown in Figure 4.3e as registered by probe A for different mixtures. As previously mentioned the standard deviation of charge  93  Chapter 4: Results and Discussion: Effect of Particle Properties fluctuations can be influenced by several factors such as, bubbles behaviour, charging stability, charge induction and charge transfer. Net cumulative charges measured by the probes in beds of binary mixtures, M1-M6, and those of mono-disperse particles GB-XL and GB-S are compared in Figures 4.3 to 4.5. It is seen that the degree of bed electrification generally increased as the wt% of GB-S in the mixture increased. The principal effect of adding small particles, GB-S, to the group B particles, GB-XL, is the reduction in mean particle diameter. As shown in Table 4.2, the specific surface areas, As, of mixtures increased as the proportion of GB-S particles increased. The increase in the degree of electrification in our experiments with increasing wt% of GB-S is probably due to an increase in particle-particle and particle-wall collisions. Figure 4.6 also shows an increase in the direct charge transferred through the ball probe in beds of smaller particles, indicating an increase in the degree of electrification in the bed, probably due to more particle-probe collisions in beds of smaller particles because of their higher specific surface areas.  94  Chapter 4: Results and Discussion: Effect of Particle Properties  Cumulative charge (μC)  -0.6  Cumulative charge (μC)  (b)  M1  M3  Probe C'  M3  M1  -0.4 M2  M2  XL  -0.2  XL  Probe C  0.0 -0.6  (c)  M2  Probe B  (d) M3  M3  -0.4  XL  Probe B'  M2 M1  -0.2 XL  M1  0.0  -0.6 Cumulative charge (μC)  (a)  (e)  Probe A  M3  (f)  M1  -0.4 M2  M3  M2  XL  -0.2 0.0 0  Probe A'  M1  XL  100  200 Time (s)  300  400  0  100  200  300  400  Time (s)  Figure 4.3 Effect of particle size on net cumulative charge as a function of time at different axial and radial positions in beds of binary mixtures, M1, M2, M3 and mono-size GBXL, P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. Letters on curves denote mixture types. For binary mixture properties, see Table 4.2. For probe positions, see Figure 2.9.  95  Chapter 4: Results and Discussion: Effect of Particle Properties  Cumulative charge (μC)  -1.0  (b)  Probe C'  M5 M4  S  -0.5  M4  -0.3  M5  S  M6  0.0  M6 M4  (c)  -0.8  Probe B  (d)  Probe B'  S  -0.5  M5  M4  M5  -0.3  S M6  0.0  -1.0 Cumulative charge (μC)  Probe C  M6  -0.8  -1.0 Cumulative charge (μC)  (a)  (e)  M5  -0.8  S  M4  M6  -0.5  (f)  S  M4 M6  -0.3 0.0 0  M5  Probe A'  Probe A 10  20  30  Time (s)  40  50  0  10  20  30  40  50  Time (s)  Figure 4.4 Effect of particle size on net cumulative charge as a function of time at different axial and radial positions in beds of binary mixtures, M4, M5, M6 and mono-size GBS, P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. Letters on curves denote mixture types. For binary mixture properties, see Table 4.2. For probe positions, see Figure 2.9.  96  Cumulative charge (nC) .  Chapter 4: Results and Discussion: Effect of Particle Properties  -30  0  Probe D  Probe D  -20  M3  M2 -10  M6  0  M5  S  0  XL  M1  M4  0  0  0  50  100  150  200  0  50  100  Time (s)  150  200  Time (s)  Figure 4.5 Effect of particle size on net cumulative charge as a function of time in the freeboard in beds of binary mixtures, M1-M6 and of mono-size GB-S and GBXL, P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. Letters on curves denote mixture types. For binary mixture properties, see Table 4.2. For position of probe D, see Figure 2.9.  Current (nA ) .  -40  Probe A Probe B Probe C  (a)  -25  Probe A' Probe B' Probe C'  (b) C B  B'  A  -10  A' C'  5 0  25  50  75  Proportion of GB-S (wt%)  100  0  25  50  75  100  Proportion of GB-S (wt%)  Figure 4.6 Current flow through ball probes as a function of wt% of GB-S in binary mixtures at different axial and radial positions, P = 379 kPa, T = 20 ± 2ºC, RH = 1317% and U- U(mf)M = 0.08 m/s. For probe positions, see Figure 2.9.  97  Chapter 4: Results and Discussion: Effect of Particle Properties -0.11 Current (nA ) .  Probe D -0.07  -0.03  0.01 0  25  50  75  100  Proportion of GB-S (wt%)  Figure 4.7 Current flow through ball probe D as a function of wt% of GB-S in binary mixtures in the freeboard, P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. For position of probe D, see Figure 2.9.  Standard deviation (μC) .  0.08  (b)  (a) C 0.06  B'  A  C' B  0.04  Probe A Probe B Probe C  0.02 0  25  50  75  Proportion of GB-S (wt%)  100  Probe A' Probe B' Probe C'  A'  0  25  50  75  100  Proportion of GB-S (wt%)  Figure 4.8 Standard deviation of charge fluctuations as a function of wt% of GB-S in binary mixtures at different axial and radial positions, P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. For probe positions, see Figure 2.9.  98  Chapter 4: Results and Discussion: Effect of Particle Properties  Standard deviation (nC) .  6.0  Probe D 4.0  2.0  0.0 0  25  50  75  100  Proportion of GB-S (%)  Figure 4.9 Standard deviation of charge fluctuations as a function of wt% of GB-S in binary mixtures in the freeboard, P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and UU(mf)M = 0.08 m/s. For position of probe D, see Figure 2.9.  Charge fluctuations (μC) .  0.05  Probe A  M3  0.03 0.01 -0.01 -0.03  M2 GB-XL  M1  -0.05 0  1  2  3  4  5  Time (s) GB-XL  M1  M2  M3  Figure 4.10 Charge fluctuations obtained by filtering the net cumulative charge curves for probe A versus time using a high-pass filter of cut-off frequency of 10*(1/period) where period is the data sampling range for GB-XL, M1, M2 and M3 mixtures. P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.08 m/s. For binary mixture properties, see Table 4.2. For position of probe A, see Figure 2.9.  99  Chapter 4: Results and Discussion: Effect of Particle Properties Ali et al. (1998) investigated electrostatic charge generation in gas-solid fluidized beds of binary mixtures. Their experiments were conducted in a PVC three-dimensional column of dimensions 0.25 m x 0.25 m x 0.60 m. Three polymer resins of 98, 88 and 88 µm mean diameter were investigated. A metal scoop coated with polymer was used to sample the particles from the fluidized bed and pour them into nine Faraday pails. The results demonstrated that small particles charged oppositely to larger particles. They concluded that bipolar charging occurs, with small particles charged positively and large particles negatively. In our experiments with binary mixtures of glass beads, no bipolar charging was observed. The differences between the results of Ali et al. (1998) and ours could be due to several factors: (a) Their particles were polymer particles of different shapes, with uneven surfaces and relatively low sphericity compared to our glass beads. Thus the charge polarity of their particles could have been influenced by factors other than particle size, such as particle shape and surface conditions, which could have significantly altered the extent and nature of particle-particle and particle-wall contacts. (b) The small particles in their experiments were <30 µm mean diameter, Geldart group C particles, whose behaviour is greatly influenced by interparticle forces such as van der Waals forces. (For more details on different forces on particles in gas-solid fluidized beds, see Appendix F.) On the other hand, 45 µm was the smallest particle diameter in our experiments. (c) The possibility of changes in charge transfer, generation or dissipation during sampling in their experiments was significant. The bipolarity in their experiments could therefore be due to their measurement techniques. (d) Since the proportion of fine particles in their experiments was small (<3 wt%) and the large particles were charged negatively, fines could have become charged positively by two different mechanisms, charge transfer due to fine-wall collisions and charge separation resulting from fines/large-particle collisions, whereas in our experiments the proportion of small particles was 10 wt% at minimum, large enough for fine-fine collisions to be the dominant charging mechanism. (This point is discussed in more detail in Section 4.2.) Zhao et al. (2000) studied contact charging between particles of different sizes with different added chemicals in a 0.25 m x 0.25 m x 0.60 m PVC column. Three types of polymer particles of mean diameter 84, 79 and 80 µm with different proportions of TiO2 and chemically-combined pigments were tested. A vertical grounded metallic tube, coated with a 100  Chapter 4: Results and Discussion: Effect of Particle Properties layer of the same polymer particles, was inserted through a number of holes in the wall of the tube at different heights to the axis of the column to sample particles. Particles entered the tube and dropped into one of seven Faraday pails located below the column to measure the charge-to-mass ratio. The authors reported bipolar charging, with small particles charged negatively whereas large ones charged positively. Although their experiments were similar to those of Ali et al. (1998), except for the sampling method and chemical properties of the particles, it was found that large particles charged positively, whereas small ones became negatively charged. The inconsistency in relative polarities in the two studies could be due to differences in the chemical properties of the particles. Thus the bipolarity observed in the Zhao et al. (2000) experiments could also be due to factors other than particle size, such as chemical composition. 4.1.2.1  Minimum fluidization velocity of a binary mixture  The minimum fluidization velocity, Umf, for a fluidized bed of mono-size particles of relatively narrow particle size and density distributions is well defined. For mixtures of particles of different sizes, the determination of the minimum fluidization velocity of the mixture, U(mf)M, is less straightforward. According to Yang (2003), although the minimum fluidization velocity of a segregating mixture can still be defined conventionally following the usual procedure for mono-size particles, the minimum fluidization velocity defined in this way loses its physical meaning. Knowlton (1977) and Chiba et al. (1979) suggested that particles in a bed of binary mixture are far from completely supported by the fluidizing gas at this velocity. A study by Chiba et al. (1979) on the minimum fluidization velocity of binary particle mixtures indicated that the conventional fluidization curve shown in Figure 4.11 was not typical, and the beginning fluidization velocity, Ubf, could not be determined accurately. Since the particle mixture started to segregate when the gas velocity exceeded the minimum fluidization velocity of the mixture, the portion of the curve shown in Figure 4.11 obtained by decreasing the gas velocity corresponds to a partially segregated bed. Hence, depending on the rate of particle separation and the time spent to obtain the complete fluidization curve, the descending portion can assume different paths.  101  Chapter 4: Results and Discussion: Effect of Particle Properties  ∆P  0  Ubf Umf  U  Utf  Figure 4.11 Determination of beginning, minimum and total fluidization velocity, Ubf, Umf, and Utf, respectively (adapted from Yang, 2003).  Several equations have been proposed in previous work for calculating the minimum fluidization velocity of binary mixtures (Goossens et al., 1971; Cheung et al., 1974; Chiba et al., 1979). Goossens et al. (1971) modified the equation of Wen and Yu (1966) to account for particle mixtures by substituting the mixture particle density, ρ(p)M, and the mixture particle size, d(p)M, of a binary mixture as follows:  d (p)M U mf ρ p μ  ⎡ d 3(p)M ρ p (ρ (p)M − ρ g )⎤ 2 = ⎢(33.7) + 0.0408 ⎥ μ2 ⎢⎣ ⎥⎦  1  2  − 33.7  (4.1)  where  1 ρ (p)M  =  (  xS 1− xS + ρ (p)S ρ (p)L  )  (4.2)  and ⎛δ ⎞ d (p)M = ⎜ ο ⎟(d (p)L )(d (p)S ) ⎝ δ ⎠  (4.3)  102  Chapter 4: Results and Discussion: Effect of Particle Properties  where  (  )  δ = 1 - x S ρ (p)S d (p)S + x S ρ (p)L d (p)L  (4.4)  and  (  )  δ ο = 1 - x S ρ (p)S + x S ρ (p)L  (4.5)  Here d(p)M is the mixture particle mean diameter, ρ(p)M is the mixture particle density, x S is the average weight fraction of small particles, ρ(p)S and ρ(p)L are the densities of the small and large particles, respectively, and d(p)S and d(p)L are the mean diameters of the small and large particles. They found that the modified Wen and Yu (1966) equation could be applied to binary mixtures of species differing in both size and density for estimation of the mixture minimum fluidization velocities, U(mf)M. In our work, attempts were made to find U(mf)M experimentally, but abandoned due to frequent blockage of the differential pressure taps. In addition, it was difficult to obtain the fluidization curves for beds of binary mixture particles. On the other hand, from the previous studies discussed above, the experimental procedure for determination of the minimum fluidization velocity for mono-size particles is not applicable to binary mixtures. In our experiments with binary mixtures, although the degree of electrification generally increased with increasing the wt% GB-S of the mixture, inconsistent trends were observed for different mixtures as shown in Figures 4.3 to 4.6. This inconsistency could be attributed to the high sensitivity of electrostatics to the superficial gas velocity, as discussed in Section 3.2.3, because in our experiments, the minimum fluidization velocity of binary mixtures was estimated based on a fixed particle size distribution, while the actual size distribution in beds of binary mixtures continually changed with time, depending on the rates of segregation and entrainment.  103  Chapter 4: Results and Discussion: Effect of Particle Properties  4.2  Bipolar Charging  It has often been reported that net charges of one polarity on particles are generated during fluidization (see Table A.1 in Appendix A.) However, few researchers realize that bipolar charging may occur from contact charging between particles of different sizes, with larger particles gaining charges of polarity opposite to that of the smaller particles (Boland & Geldart 1971; Ali et al., 1998; Zhao et al., 2000; Chen et al. 2003b; Zhao et al., 2003; Mehrani et al., 2007b). In gas-solid fluidized beds, charging occurs in fluidizing gas rather than in a vacuum. The high electrical field between the surfaces of charged particles causes the fluidizing gas to ionize as the particles separate. Ions neutralize the particle surface charge, but tend to over-compensate, so that the portion of the surface where the discharge occurs gains opposite polarity to the initial frictional charge. Particle surfaces may thus acquire areas of different polarity (Hull, 1949). According to Cross (1987), surfaces of different polarity can also be produced due to surface non-uniformities, leading to different surface energy at different positions. Wang et al. (2008) investigated the distribution of electrostatic potential in gas-solid fluidized beds in a Plexiglas column of 150 mm inner diameter and 1.0 m height, with polyethylene resin particles of wide particle size distribution (185-1700 µm). Their results show that the electric field in the bed was non-uniform. Furthermore, the voltage polarity reversed near the bed surface. In our work, bipolar charging was observed in two cases: (a) Experiments where polyethylene resin particles constituted the bed materials, as shown in Figures 3.19, 3.20 and 3.38 to 3.41 in Chapter 3. (b) Experiments with binary mixtures of glass beads involving small particles of ~30 µm mean diameter, with <2 wt% of fines in the mixture with particles of 574 µm mean diameter. This section discusses the former case, whereas the latter is discussed in Chapter 5. In gas-solid fluidized beds triboelectrification and frictional charging are known to be the mechanisms of generating electrostatic charges. In beds of broad particle size distributions or binary mixture of different particle sizes, these two charging mechanisms involve fines/fines, large/large-particles, fines/large-particles, fines/wall, and large-particle/wall collisions. Mehrani (2007b) found that large particles tended to charge negatively, whereas fines tended 104  Chapter 4: Results and Discussion: Effect of Particle Properties  to become charged positively. However, fines carried significant, but different, charges out of the column depending on their sizes, physical and chemical surface structure, and the moisture content of the fluidizing gas. Results for probes C, C’, B’, A’, and D in Figures 3.19a, b and f, 3.20, 3.38a, b and d, 3.39, 3.41 and 3.42 indicate bipolar charging. It would appear that large particles charged negatively due to interparticle charge transfer and due to charge separation between them and the wall, whereas fine particles charged positively due to charge transfer between fines and the wall and to charge separation between fines and the large particles in the bed. On the other hand, fine particles formed in the polyethylene resin due to repetitive fracturing and/or grinding of the larger particles in the bed. This may have resulted in differences in the nature and extent of the contact surfaces for the finer particles, thereby influencing the charging mechanisms. Furthermore, according to van Krevelen (1990) some polymer properties are dependent on particle size. For example, polyethylene particles of different sizes have different molecular weights, copolymer contents, surface roughnesses and chemical compositions due to different relative catalyst kernel content. In addition, the tendency of polyethylene particles to adsorb moisture is not the same for small and large particles. This could influence the charging mechanisms in beds of polyethylene resin. It was observed that the results of polyethylene experiments were inconsistent for some cases and at different locations in the bed as shown by the results for probes A’ and B’ in Figures 3.19d, 3.19f, 3.38d and f. Large fluctuations in the net cumulative charges in the bed for some cases and at different locations were also observed for most probes. This probably occurred due to a combination of charging mechanisms in beds of polyethylene particles. For example, fine particles could be charged positively due to charge transfer between the fines and the large particles, as well as to the wall of the column, which could be charged negatively due to charge transfer between the fines and the large particles and to charge separation between the fines and the wall. However one polarity predominates, determining the net charge on the particle surfaces.  105  Chapter 4: Results and Discussion: Effect of Particle Properties  4.3  Effect of Particle Density and Type  Comparison of the results for glass beads and polyethylene particles in our work indicates that the magnitude of net cumulative charges for the glass beads was much higher than for the polyethylene. For example, Figure 4.12 indicates that, for the same pressure (448 kPa), temperature (20 ±2ºC), relative humidity of fluidizing air (13-17%), probe locations as shown in Figure 2.9, excess gas velocity (U-Umf = 0.05 m/s), time (e.g. 400 s) and similar average particle size, the magnitude of the cumulative charge was approximately three orders of magnitude larger for GB-L than for the HDPE particles. The latter are less dense, but the difference is surprising in view of the well-known greater impact of electrostatics in the case of polymerization reactors. Charge transfer generally increases as the force at contact and the speed of rubbing increase (Montgomry, 1959). Haenen (1976) showed that, when a series of metals and polymers were rubbed together at least fifty times, the charge transferred, Qtransfer, could be related to the force of contact, Fcontact, by  Q transfer ∝ F α  (4.6)  where α lies between 0.3 and 1.0 and is a function mainly of the metal but, to a lesser extent, also of the polymer. In gas-solid fluidized beds, collisions occur between particles and between particles and the wall. Particle collisions result from: a) the relative velocity between particles of different sizes or densities due to different responses to the mean gas flow; and b) radial gradients in particle vertical velocity component due to the wall effect and radial gradients in gas velocity. The maximum collision force between two particles is a function of several factors such as particle density, particle relative velocity, coefficient of restitution and particle elasticity (see Appendix F). According to Senior (1992), the collision force increases as the particle density increases. From Tables 2.1 and 2.2, the density of glass beads is approximately 3 times that of polyethylene particles. Thus the differences between the net cumulative charge of HDPE and 106  Chapter 4: Results and Discussion: Effect of Particle Properties  GB-L may be due to the sizable difference between their densities. On the other hand, in our work, bipolar charging was observed in all polyethylene experiments without exception, while it was not observed in the glass bead experiments, except for the added fines experiments discussed in Chapter 5. Although the net cumulative charge at any location in the bed is not determined by the polarity of the particles at that location at any time, it is influenced by bipolar charging in the bed. To further investigate the effect of particle density on the degree of electrification in beds of polyethylene particles of different densities (e.g. high-density polyethylene, HDPE, and linear-low-density polyethylene, LLDPE), freely bubbling experiments were conducted in the elevated-pressure column with collision ball probes at four levels, and at two radial positions as shown in Figure 2.9 with LLDPE particles of 487 µm mean diameter and a static bed height of ~0.53 m. (For physical properties of the HDPE and LLDPE see Table 2.2.) Umf for LLDPE was found experimentally to be 0.10 m/s for a freeboard pressure of 448 kPa. The operating conditions were similar to those for the HDPE experiments for a freeboard pressure of 448 kPa as described in Section 3.2.1.2. The bed was again fluidized for about an hour to achieve steady state before collecting data. The net cumulative charge in the bed was then measured by the collision ball probes. Figure 4.13 compares net cumulative charge in beds of HDPE and of LLDPE particles. These results suggest that the net cumulative charge in beds of polyethylene of different densities and similar mean particle size is likely independent of particle density. The small differences in the net cumulative charges between HDPE and LLDPE particles in Figure 4.13 may be attributable to other factors such as particle shape, surface roughness and concentration of fine particles (see Appendix D.5 for particle size distribution of HDPE and LLDPE). For example, the slightly higher net cumulative charge for HDPE in Figures 4.13a and b could be due to the higher concentration of fines in the bed of HDPE compared with the LLDPE.  107  Chapter 4: Results and Discussion: Effect of Particle Properties  lCumulative chargel (nC)  3  10  (a)  2  10  1  10  (b)  GB-L  Probe C  Probe C' GB-L HDPE  HDPE  0  10  -1  10  lCumulative chargel (nC)  3  10  (c)  2  10  Probe B  (d)  GB-L  Probe B' GB-L  1  10  HDPE  0  10  HDPE  -1  10  lCumulative chargel (nC)  3  10  (e)  2  10  Probe A  (f)  Probe A' GB-L  GB-L  1  10  HDPE  HDPE  0  10  -1  10  0  100  200 Time (s)  300  400  0  100  200  300  400  Time (s)  Figure 4.12 Comparison of net cumulative charge as a function of time at different axial and radial positions for GB-L and HDPE particles, dP = 574 and 450 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  108  Cumulative charge (nC)  Chapter 4: Results and Discussion: Effect of Particle Properties  20  Cumulative charge (nC)  (b)  HDPE  Probe C'  HDPE  10 LLDPE  LLDPE  0 -10  Probe C  -20 20  (c)  Probe B  (d)  Probe B'  10 0  LLDPE  -10  LLDPE  HDPE  HDPE  -20 20  Cumulative charge (nC)  (a)  (e)  Probe A  10  (f)  Probe A'  LLDPE  HDPE  0 LLDPE  -10 HDPE -20 0  100  200 Time (s)  300  400  0  100  200  300  400  Time (s)  Figure 4.13 Comparison of net cumulative charge as a function of time at different axial and radial positions for LLDPE and HDPE particles, dP = 487 and 450 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  109  Chapter 4: Results and Discussion: Effect of Particle Properties  4.4  Effect of Particle Chemical Composition  Freely bubbling experiments were conducted in the elevated-pressure column with collision ball probes at four levels and two radial positions as shown in Figure 2.9 to investigate the effect of particle type on the electrostatic behaviour. Two different types of low density polyethylene resin, PE-X1 and PE-X2, of 523 and 502 µm mean diameter, respectively, were used in the experiments. (For physical properties of PE-X1 and PE-X2 see Table 2.2.) The PE-X1 resin was produced using a new catalyst of different base and active components, whereas PE-X2 was produced by the typical Zieglar-Natta M-catalyst normally used in LLDPE UNIPOL reactors. The static bed height was ~0.53 m in both cases, and the freeboard absolute pressure was 448 kPa. Umf was found experimentally to be 0.10 m/s at this pressure for both PE-X1 and PE-X2. Temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%). The excess gas velocity was also maintained constant at (U-Umf) = 0.05 m/s. In each case, the bed was fluidized for about an hour to achieve steady state before collecting data. The net cumulative charge in the bed was then measured. Figures 4.14 and 4.15 compare the net cumulative charge for beds of LLDPE, PE-X2 and PE-X1. Zhao et al. (2000) reported bipolar for particles of different sizes with different added chemicals. Although they attributed the bipolarity to particle size, it could also have been due to different added chemicals in their experiments. Mehrani et al. (2007b) investigated the electrostatic behaviour of different fines added to a Faraday cup fluidized bed as shown in Figure 1.5. The charges transported by fines of different physical and chemical properties (Larostat 519 antistatic agent, glass beads, silver-coated glass beads, polyethylene catalyst and silica particles) were determined after their injection into an initially charged bed of much larger mono-dispersed particles (glass beads or polyethylene). Entrained fines carried significant charges out of the column, leaving behind a net charge of opposite polarity, with the quantity of charge depending on the size, physical properties and chemical composition of the particles.  110  Cumulative charge (nC)  Chapter 4: Results and Discussion: Effect of Particle Properties  20 10  Cumulative charge (nC)  Probe C  LLDPE  (b)  Probe C' LLDPE  0  PE-X2  PE-X2 PE-X1  -10  PE  PE-X1  -20  20  (c)  (d)  Probe B  10  Probe B' LLDPE  PE-X1 LLDPE  0  PE-X1  -10  PE-X2  PE-X2 -20 20  Cumulative charge (nC)  (a)  (e)  10  Probe A  (f)  PE-X1  Probe A' PE-X1  PE-X2  0 LLDPE PE-X2  -10 -20 0  100  200 Time (s)  300  LLDPE  400  0  100  200  300  400  Time (s)  Figure 4.14 Comparison of net cumulative charge as a function of time at different axial and radial positions for LLDPE, PE-X1 and PE-X2 particles, dP = 487, 523 and 502 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  111  Chapter 4: Results and Discussion: Effect of Particle Properties  Cumulative charge (nC)  2  Probe D PE-X1  1 0 -1 -2 0  PE-X2  LLDPE  100  200  300  400  Time (s) Figure 4.15 Comparison of net cumulative charge as a function of time in the freeboard region for LLDPE, PE-X1 and PE-X2 particles, dP = 487, 523 and 502 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For position of probe D, see Figure 2.9. Table 4.3 Current flow through ball probes at different axial and radial positions in beds of LLDPE, PE-X1 and PE-X2 particles, dP = 487, 523 and 502 µm, respectively. P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-U(mf)M = 0.05 m/s. For probe positions, see Figure 2.9. Current (nA) Probe A Probe B Probe C Probe D Probe A’ Probe B’ Probe C’  LLDPE -0.0059 -0.0179 0.0268 0.0007 -0.0080 0.0052 0.0238  PE-X1 -0.0045  0.0345 -0.0215 0.0024 -0.0047 0.0031 -0.0168  PE-X2 -0.0037 -0.0344 0.0237 0.0001 -0.0011 -0.0106 0.0138  The results for probes B, C and C’ in Figures 4.14a, b and c show that the polarity of the bed of PE-X1 particles at these locations was opposite to those in the beds of LLDPE and PE-X2. This is probably due to the different chemical composition of PE-X1. Similarly, Table 4.3 shows that the direct charge transferred through probes B, C and C’ for PE-X1 particles was of opposite polarity to those for the LLDPE and PE-X2 particles.  112  Chapter 4: Results and Discussion: Effect of Particle Properties  It is also important to note that particles of different chemical compositions differ in their tendency to adsorb moisture. They also react differently with adsorbed water molecules (van Krevelen, 1990). Although, in our experiments, the relative humidity of the fluidizing air was controlled in a narrow range, it is possible that particles were affected differently by moisture, causing their surface composition to differ from the original surface composition, thereby altering chemical and physical properties, and consequently influencing charging. In our work, results for polyethylene resins were less reproducible than for glass beads. For PE-X1 particles it was difficult to reproduce similar results (quantitatively and qualitatively). Figure 4.16 compares the net cumulative charges from three experiments with PE-X1 particles. Although these experiments were conducted with similar operating conditions, the PE-X1 particles charged oppositely in different experiments. This could be due to their surfaces chemical/physical structure, which could have been influenced by the new catalyst. It also could have been caused by the higher fraction of coarse (>2000 µm) particles in the PE-X1 resin compared with the other polyethylene resins in our experiments (see Appendix D.)  4.5  Summary  The influences of particle size, density and chemical composition were investigated in this chapter. Freely bubbling experiments in fluidized beds of four sizes of mono-size glass beads indicated that for Geldart group B particles electrostatic charging increased slightly as particle size decreased, and increased significantly for smaller particles (<100 µm). In experiments with binary mixtures of group A and B particles, electrostatics increased as the average particle size decreased. The rate of these increases and their significance depended on the particle size distribution and location in the bed. The magnitude of net cumulative charges was much higher for the glass beads than for the polyethylene particles, attributable to the sizable difference between their densities and electrical properties. Bipolar charging was observed in all experiments using polyethylene resin. It was found that large particles charged negatively due to interparticle charge transfer and charge separation between them and the wall, whereas fine particles charged positively due to charge transfer between fines and the wall and to charge separation between fines and the large particles in the bed. 113  Cumulative charge (nC)  Chapter 4: Results and Discussion: Effect of Particle Properties  20  Cumulative charge (nC)  Probe C  (b)  Probe C'  10  X1-2 X1-3  X1-2  0  X1-1 X1-3  -10  X1-1  -20 20  (c)  10  (d)  Probe B  X1-1  0  Probe B' X1-3  X1-3  -10  X1-2  X1-1  X1-2  -20 20  Cumulative charge (nC)  (a)  (e)  Probe A  10  (f)  Probe A' X1-3  X1-1  X1-2  0 -10  X1-2  -20 0  100  X1-1  X1-3 200 Time (s)  300  400  0  100  200  300  400  Time (s)  Figure 4.16 Comparison of net cumulative charge as a function of time at different axial and radial positions for PE-X1 particles, dP = 523 µm, P = 448 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  114  Chapter 5: Electrostatic Charge Reduction Technique  Chapter 5: Electrostatic Charge Reduction Technique  5.1  Humidification  Increasing the relative humidity of the fluidizing gas is a well-known method of reducing electrostatic charges in gas-solid fluidized beds. Several researchers (Ciborowski and Wlodarski, 1962; Katz and Sears, 1969; Boland and Geldart, 1971; Wolny and Kazmierczak, 1989; Guardiola et al., 1996; Park et al., 2002b; Yao et al., 2002; Chen et al., 2003b; Mehrani et al., 2007b) investigated the effect of relative humidity on the electrostatic charge build-up in gas-solid fluidized beds. Ciborowski and Wlodarski (1962) reported that the moisture content of the fluidizing air does not change the rate of charge generation, but increases the rate of charge dissipation. The moisture in the fluidizing gas results in rapid dissipation of charges by increasing the surface conductivity (Katz and Sears, 1969). Park et al. (2002b) reported that increasing the relative humidity reduced electrostatic charge accumulation by increasing the surface conductivities of particles, thereby enhancing charge dissipation. Boland and Geldart (1971) showed that charging occurred, even at the highest relative humidity tested in their experiments, and it is therefore unlikely that high relative humidities completely eliminate static effects, but merely allow better dissipation of generated charges as a result of increased surface conductivity.  Ciborowski and Wlodarski (1962) found that the minimum fluidization velocity decreased as the moisture content of the fluidizing gas increased, probably because the bed was easier to fluidize when interparticle forces are reduced. Guardiola et al. (1996) reported that gas relative humidities of approximately 30-65% strongly affected the degree of electrification in a fluidized bed. The voltage measured in their experiments between the probe and the distributor was ten times higher at 50% relative humidity of the fluidizing gas than for 70% relative humidity. Their results also indicate that when the relative humidity was less than ~10%, electrostatics could not be measured accurately because particles adhered to their electrostatic probe, leading to irreproducible voltages. However, the effects of relative humidity were found to be complex and dependent on the operating conditions.  115  Chapter 5: Electrostatic Charge Reduction Technique Park et al. (2002b) investigated the influence of fluidizing air relative humidity on the degree of bed electrification. In their experiments, the relative humidity was varied over a wide range (from 6 to 98%). Their results suggested that for relative humidities less than 20%, charge inducement and transfer decreased as the relative humidity increased. For 20 to 40% relative humidities there was a rapid reduction in electrostatic charging. For 40 to 60% relative humidities the accumulation of electrostatic charges was insignificant, whereas for relative humidities > 80% the signals from their electrostatic ball probe were irregular due to capillary forces. Park et al. (2002b) reported that relative humidities higher than 60% resulted in sufficient charge dissipation to prevent charge accumulation in gas-solid fluidized beds.  Freely bubbling experiments were conducted in the three-dimensional fluidized bed with collision ball probes at four levels and two radial positions as shown in Figure 2.9 to investigate the effect of the relative humidity of fluidizing air on the cumulative charge buildup in the bed. The particles were glass beads, GB-XL, of 693 µm mean diameter. (For physical properties, see Table 2.1.) The static bed height was ~0.53 m. The pressure, temperature and excess air velocity, were maintained nearly constant at 379 kPa, 20 ±2ºC and U-Umf = 0.04 m/s in order to isolate the effect of relative humidity. Umf for GB-XL was found experimentally to be 0.38 m/s for this pressure and temperature. The relative humidity of the fluidizing air was adjusted by passing the air through vapour-removal filters, a refrigerating unit (KAESER Model TA11 Refrigerated Dryer) and an air dryer containing silica gel desiccant upstream of the column. The relative humidity of fluidizing air was varied in this manner from 5 to 30%.  For each experiment, the bed was first fluidized for about an hour to achieve steady state before collecting data. The net cumulative charge was then measured by the collision ball probes during fluidization and plotted versus time as shown in Figure 5.1. Figure 5.2 also indicates the degree of bed electrification by plotting the current (obtained from the slopes of the curves in Figures 5.1) as a function of the relative humidity. As expected, the experimental results in Figures 5.1 and 5.2 show that as relative humidity decreased the accumulation of electrostatic charges increased. Note that the charging depended on the axial and radial location of the probes, as well as the relative humidity of the fluidizing air. 116  Cumulative Charge (μC)  Cumulative Charge (μC)  Cumulative Charge (μC)  Chapter 5: Electrostatic Charge Reduction Technique  -1.0  Probe C  (a)  Probe C'  (b)  -0.8 -0.5 5-8  -0.3  13-17 13-17  5-8 23-28  23-28  0.0  -1.0  Probe B  (c)  Probe B'  (d)  -0.8 -0.5 5-8 -0.3  10-15 5-8  10-15  23-28  23-28  0.0  -1.0  Probe A  (e)  Probe A'  (f)  -0.8 -0.5  5-8  -0.3  5-8 13-17 23-28  13-17 23-28  0.0 0  100  200  300  Time (s)  400  500  0  100  200  300  400  500  Time (s)  Figure 5.1 Effect of relative humidity of fluidizing air on net cumulative charges as a function of time at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC, and U-Umf = 0.04 m/s. Numbers on curves denote % RH. For probe positions, see Figure 2.9.  117  Chapter 5: Electrostatic Charge Reduction Technique -2  Probe A Probe B Probe C  Current (nA) .  (a) -1.5  Probe A' Probe B' Probe C'  (b)  -1  A'  B  A  -0.5  B'  C  C'  0 0  6  12  18  24  RH (%)  30  0  6  12  18  24  30  RH (%)  Figure 5.2 Current flow through ball probes as a function of fluidizing air relative humidity at different axial and radial positions in a bed of GB-XL particles, dp = 693 µm, P = 379 kPa, T = 20 ± 2ºC and U-Umf = 0.04 m/s. For probe positions, see Figure 2.9. Charging between materials is not only caused by electron transfer, but also by positive or negative ion transfer, called “ionic theory” (Cross, 1987). Real surfaces of metal and insulators contain a charged double layer or adsorbed layers of impurities, which are ionic in nature (Cross, 1987). All the forces of molecules or ions on the surface of particles are not balanced by contact with other particles. This imbalance in forces causes the surfaces of the particles to attract and adsorb gases to satisfy their residual forces (Maron and Prutton, 1965). There are two types of adsorption: physical (due to van der Waals forces) and chemical (due to chemical bonds). Physical adsorption is established rapidly, and it is reversible since it requires only a low heat of adsorption (Fan and Zhu, 1998); whereas chemical adsorption, which forms stronger bonds between the gas and the particle surface, is accompanied by much higher heat changes (Maron and Prutton, 1965). In many cases, materials adsorb gas molecules physically at low temperatures and chemically at higher temperatures (Maron and Prutton, 1965). The degree of adsorption is affected by the gas pressure. As the relative humidity of fluidizing gas increases, the partial pressure of water vapour also increases (Park et al., 2002b). Therefore, physical adsorption of water vapour onto the particles increases. According to Cross (1987), most surfaces of solid particles in air are covered by a layer of water ranging in thickness between a monolayer to a macroscopic thin film. As adsorbed moisture forms a continuous film on the solid surface, the water provides a medium for dissociation of ions. 118  Chapter 5: Electrostatic Charge Reduction Technique Except at two locations in the bed (i.e. probes B’ and C’), our experimental results show a decrease in bed electrification with increasing relative humidity of the fluidizing air. The results for Probes C, B, A and A’ in Figures 5.1a, c, e, f, 5.2a and 5.2b indicate that for relative humidities <10%, the accumulation of electrostatic charges in the bed significantly increased as relative humidity decreased, whereas for RH from 13% to 28% the accumulation of electrostatic charges moderately decreased as relative humidity increased. This is probably due to the increase in charge dissipation resulting from the increase of conductivity around the surfaces of particles as the relative humidity of the fluidizing air increased. On the other hand, the results for probes B’ and C’ in Figures 5.1b, d and 5.2b indicate that static charging at higher levels and near the wall was insensitive to fluidizing gas relative humidity. It seems that, for this range of RH (5-30%), humidifying the fluidizing air is very effective when particle-particle charging is the dominant mechanism in the bed, and not effective when particle-wall charging is the dominant mechanism. This probably occurs because the wall is not influenced by the RH of the fluidizing air in the same manner as for the particles. However, it is likely that surfaces with different affinity for moisture behave differently in different ranges of relative humidity. These results are consistent with those of Guardiola et al. (1996), Park et al. (2002b) and Chen et al. (2003b) which indicate that for RH <30%, the electrostatic charge accumulation in the bed was either slightly affected or not influenced by fluidizing air relative humidity. Although Guardiola et al. (1996) attributed this to the difficulties in measuring the degree of bed electrification for this range of relative humidity due to adhesion of particles to either the probe or the wall, it is noteworthy that the Guardiola et al., (1996), Park et al., (2002b) and Chen et al., (2003b) experiments were all conducted in relatively small columns (three-dimensional fluidization column of 0.044 m inner diameter, two-dimensional column of 0.307 m wide, 0.022 m thick and 1.24 m high, and three-dimensional column of 0.088 m inner diameter, respectively) where the charging is greatly influenced by the wall. This may explain the similarity between their results and ours for this range of relative humidity and locations near the wall (e. g. B’ and C’.)  Humidifying the fluidizing gas to reduce the degree of electrification is not suitable for polymerization processes for several reasons: a) most polymerization processes take place between 70 and 110ºC, while humidification has been reported to be ineffective at 119  Chapter 5: Electrostatic Charge Reduction Technique temperatures >80ºC (Ciborowski and Wlodarski, 1962); and b) all polymerization processes use catalysts which are poisoned by humidity.  5.2  Addition of Fines and Antistatic Agents  In order to study the effect of adding small portion of fines such as Geldart group C particles to a bed of large particles (Geldart group B particles), freely bubbling experiments were conducted in the three-dimensional fluidization column with collision ball probes at four levels and at two radial positions as shown in Figure 2.9. The particles were glass beads (GBL) of 574 µm mean diameter and different proportions of added fines (GB-XS) of 30 µm mean diameter. (Physical properties of both the GB-L and GB-XS are provided in Tables 2.1 and 2.3.) The proportion of GB-XS in the binary mixtures was varied from 0.0 to 2.0 wt%. The static bed height was ~0.53 m and the freeboard absolute pressure was maintained at 379 kPa. Temperature and relative humidity of the fluidizing air were maintained nearly constant (T = 20 ±2ºC and RH = 15 ±2%) in order to isolate the effect of added fines. Umf was found experimentally to be 0.310 m/s at this pressure and temperature. Excess air velocity was maintained constant in all experiments at U-Umf = 0.05 m/s. The bed was fluidized for about an hour to achieve steady state before collecting data. The net cumulative charge generated in the bed was then measured by the collision ball probes and plotted versus time in Figures 5.3 and 5.4. Current flows through the ball probes due to direct charge transfer, obtained from the slopes of the cumulative charge curves in Figures 5.3 and 5.4, are plotted versus the wt% of GB-XS in Figures 5.5 and 5.6.  In the current study, results for the binary mixtures of added GB-XS fines in Figures 5.3 and 5.5 show that the degree of electrification decreased as the proportion of GB-XS fines increased from 0.0 to 2.0 wt%. This trend is limited to higher levels in the bed (Level C), where one can see that the electrostatic charging was insensitive to the added fines in the range of 0.0-2.0% as indicated by the results for probes C and C’ in Figures 5.3a, 5.3b, 5.5a and 5.5b. The results for probe D in Figures 5.4 and 5.6 show an increase in the degree of electrification in the freeboard region with increasing wt% of GB-XS fines, probably due to increased entrainment as the proportion of fines in the bed increased. Figures 5.4 and 5.6 also show that fines charged positively. It would appear that the fines charged positively due to 120  Chapter 5: Electrostatic Charge Reduction Technique charge transfer between the fines and the wall and to charge separation between the fines and  Cumulative Charge (μC)  Cumulative Charge (μC)  Cumulative Charge (μC)  large particles in the bed.  -0.20 (a)  Probe C  -0.15  2.0  1.5 1.0  0.5  (b)  Probe C'  1.5  2.0  0.0  1.0  0.5  -0.10  0.0 -0.05 0.00  -0.20  Probe B  (c)  0.0 0.5  -0.15  Probe B' 0.0 0.5 1.0  (d)  1.0  1.5  1.5  -0.10  2.0  2.0  -0.05 0.00 -0.20  Probe A  (e)  0.5 1.5  0.0  -0.15  Probe A'  (f)  1.0 2.0  0.0  -0.10  1.5 0.5 1.0 2.0  -0.05 0.00 0  60  120 Time (s)  180  240  0  60  120  180  240  Time (s)  Figure 5.3 Effect of added fines on net cumulative charge as a function of time at different axial and radial positions in bed of binary mixture of GB-L particles and GB-XS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves denote wt% of added fines (GB-XS) in binary mixtures. For probe positions, see Figure 2.9. 121  Cumulative Charge (nC)  Chapter 5: Electrostatic Charge Reduction Technique  1.2  Probe D  0.9 0.6  1.0 1.5  2.0  0.3  0.5  0.0 0  0.0 180  60  120  240  Time (s) Figure 5.4 Effect of added fines on net cumulative charge as a function of time in the freeboard in beds of binary mixture of GB-L particles and GB-XS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves denote wt% of added fines (GB-XS) in binary mixtures. For position of probe D, see Figure 2.9.  -1.5  .5  Probe A Probe B Probe C  Current (nA ) .  (a) -1.2  Probe A' Probe B' Probe C'  (b) .2  C'  C  -0.9  0.9  A -0.6  A'  0.6  B  -0.3  B'  0.3  0.0  0.5  1.0  1.5  Proportion of fines (wt%)  2.0  0.0  0.5  1.0  1.5  2.0  Proportion of fines (wt%)  Figure 5.5 Current flow through ball probes as a function of wt% of added fines (GB-XS) at different axial and radial positions in bed of binary mixture of GB-L particles and GBXS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  122  Chapter 5: Electrostatic Charge Reduction Technique 0.06 Current (nA) .  Probe D 0.04  0.02  0.00 0.0  0.5  1.0  1.5  2.0  Proportion of fines (wt%)  Figure 5.6 Current flow through ball probe D as a function of wt% of added fines (GB-XS) in the freeboard in bed of binary mixture of GB-L particles and GB-XS fines, dp = 574 and 30 µm, respectively, P = 379 kPa, T = 20 ±2ºC, RH = 13-17% and U-Umf = 0.05 m/s. For position of probe D, see Figure 2.9. As previously mentioned, the charge fluctuations measured by ball probes could be due to several factors such as bubble behaviour, charging instability, charge induced by charged particles and charge transfer. Figure 5.7 provides an example of the charge fluctuation curves obtained by filtering the net cumulative charge curves shown in Figure 5.3c for probe B for binary mixtures of GB-L and GB-XS particles. Figure 5.8 plots standard deviation of the charge fluctuations registered by ball probes versus the wt% of GB-XS in the binary mixtures. Pressure fluctuations in gas-solid fluidized beds result from local bubble passage and pressure waves generated due to bubble formation, coalescence, splitting and eruption at the bed surface, particle oscillations and bed oscillations (Bi et al., 1995). Local differential pressure fluctuations measured over a certain vertical interval of the fluidized bed are mainly induced by the passage of bubbles with most pressure waves filtered (Bi et al., 1995). Hence, the standard deviation of pressure fluctuations across the bed is an indicator of fluidization stability. Figure 5.9 plots the standard deviation of pressure fluctuations across the bed as a function of wt% of GB-XS, whereas the pressure fluctuations across the bed are plotted versus time in Figure 5.10.  123  Chapter 5: Electrostatic Charge Reduction Technique  Charge fluctuations (nC)  0.1  .1  (a)  (b)  0.0  0.5  0.0  .0  -0.1  .1  0.1  1 (d)  Charge fluctuations (nC)  (c) 1.0  1.5  0.0  0  -0.1  1 0  Charge fluctuations (nC)  0.1 (e)  2  4  6  8  10  Time (s)  2.0 0.0  -0.1 0  2  4  6  8  10  Time (s) Figure 5.7 Charge fluctuations obtained by filtering the net cumulative charge curves for probe B versus time using a high-pass filter of cut-off frequency 10*(1/period), where period is the data sampling range, for binary mixtures of GB-L particles GB-XS fines (For physical properties of the GB-L and the added GB-XS fines, see Tables 2.1 and 2.3), P = 397 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s. Numbers on curves denote the wt% of GB-XS. For position of probe B, see Figure 2.9.  124  Chapter 5: Electrostatic Charge Reduction Technique  Standard deviation (nC) .  18.0 (b)  (a) 15.0  C'  C  12.0  B'  A  9.0  Probe A Probe B Probe C  6.0 0.0  0.5  1.0  Probe A' Probe B' Probe C'  B  A'  1.5  2.0  0.0  Proportion of fines (wt%)  0.5  1.0  1.5  2.0  Proportion of fines (wt%)  Figure 5.8 Standard deviation of charge fluctuations as a function of wt% GB-XS in binary mixtures of GB-L particles and GB-XS fines at different axial and radial positions, (For physical properties of the glass beads and added fines, see Tables 2.1 and 2.3), P = 379 kPa, T = 20 ± 2ºC and RH = 13-17%. For positions of probes, see Figure 2.9.  Standard deviation (kPa)..  0.4  0.3  0.2  0.1 0.0  0.5  1.0  1.5  2.0  Proportion of fines (wt%)  Figure 5.9 Standard deviation of differential pressure fluctuations across a bed of binary mixtures of GB-L particles and GB-XS fines as a function of wt% of GB-XS in the binary mixtures (For physical properties of the glass beads and added fines, see Tables 2.1 and 2.3), Pressure = 379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s.  125  Chapter 5: Electrostatic Charge Reduction Technique  0.3  Pressure fluctuations (kPa)  0.3 (a)  (b) 0.5  0.0 0.0  0.0  -0.3  0.3 0.3  Pressure fluctuations (kPa)  0.3  (d)  (c) 1.0  1.5  0.0  0.0  -0.3  0.3 0  Pressure fluctuations (kPa)  0.3  2  4  6  8  10  Time (s)  (e) 2.0  0.0  -0.3 0  2  4  6  8  10  Time (s)  Figure 5.10 Differential pressure fluctuations across beds of binary mixtures of GB-L particles and GB-XS fines as a function of time. (For physical properties of the GB-L and added GB-XS fines see Tables 2.1 and 2.3), P = 397 kPa, T = 20 ± 2ºC, RH = 1317% and U-Umf = 0.05 m/s. Numbers on curves denote the wt% of GB-XS.  126  Chapter 5: Electrostatic Charge Reduction Technique The direct relationship between bubble motion in the bed and the charge signals registered by the ball probes can be obtained by comparing the fluctuations of the charge signals measured by ball probes such as those shown in Figure 5.7 and the differential pressure fluctuations measured by pressure transducers, such as those plotted in Figure 5.10. It can also be verified by comparing the power spectral distributions of the charge and pressure fluctuations signals, obtained by the fast Fourier transformation (FFT) method. For example, as shown in Figure 5.11, the power spectrum of the charge fluctuation signals from probe B and the pressure fluctuation signals from the pressure transducer are very similar, indicating that fluctuations of the charge signals registered by probe B are mainly due to charges induced by passing bubbles and that the passage of bubbles is the dominant cause of fluctuations.  -1  Power (kPa)  9.0x10  Differential pressure signals -1  6.0x10  -1  3.0x10  0.0  0  2  4  6  8  10  -2  Power (nC)  1.2x10  Charge fluctuation signals -3  8.0x10  -3  4.0x10  0.0  0  2  4 6 Frequency (Hz)  8  10  Figure 5.11 Power spectrum of charge fluctuations signals and differential pressure signals obtained by fast Fourier transformation (FFT) for binary mixture of GB-L and 1.0 wt% of GB-XS fines. (For physical properties of the GB-L and added GB-XS fines, see Tables 2.1 and 2.3), P = 397 kPa, T = 20 ± 2ºC, RH = 13-17%, U-Umf = 0.05 m/s. 127  Chapter 5: Electrostatic Charge Reduction Technique The standard deviation of the charge fluctuations of probes A, B, A’ and B’ in Figure 5.8 show that as the proportion of GB-XS increased in the binary mixture, the charge fluctuations generally decreased, indicating more stable fluidization. Figure 5.9 also shows that as the proportion of GB-XS fines increased, the differential pressure fluctuations decreased, again indicating more stable fluidization. This may be due to the fine particles acting as spacers or lubricants, reducing contacts between the large particles. This would enhance the smoothness and stability of fluidization, consequently reducing the generation of electrostatic charges. Another possibility is that the large particles, which were charged negatively, may have been partially neutralized due to their contact with the fines, which carried positive charges as indicated by the probe D results in Figure 5.4. The first of these two possibilities is likely to be more applicable at lower levels (A and B), where the particleparticle contacts are probably the dominant charging mechanism. However, as bubbles rise and approach higher levels (level C), it becomes larger, and charging mechanisms other than particle-particle contacts such as particle-wall contacts may play a greater role. This could explain the results for probes C and C’ in Figure 5.8a and b, located at a higher level, where it can be seen that as the proportion of added GB-XS fines increased, the charge fluctuations increased, indicating unstable fluidization, and enhancing the degree of electrification.  Previous research has been performed on the possibility of preventing or reducing electrostatic charges in gas-solid fluidized beds. One of the methods studied is the addition of fine particles belonging to Geldart group C or antistatic agents, but efforts have generally been limited to minimizing the influence of electrostatics, rather than understanding the phenomena involved and what is happening in different locations. Wolny et al. (1983) investigated the effect of adding conductive (active coal), semi-conductive (titanium dioxide) and dielectric (pigment) fine materials to beds of polystyrene of 1.2 mm mean diameter by measuring the charge on a single particle ejected from the bed into a Faraday cup. They concluded that the addition of fines neutralizes electric charges and improves the hydrodynamics of the fluidized bed, thereby decreasing the effect of electrostatics, independent of the electrical properties of the added fines. Wolny et al. (1989) studied the effect of added fines belonging to Geldart group C such as aluminum powder of 0.1-15 µm diameter or sodium chloride of 1-60 µm diameter on bed behaviour due to electrostatic charge generation by discharging particles from the bed and placing them in an electrical 128  Chapter 5: Electrostatic Charge Reduction Technique field. They reported that the sodium chloride fines exerted the strongest influence on the net charge distribution in the bed. The degree of electrification in a fluidized bed of glass beads and mixtures of glass and steel particles was studied by Guardiola et al. (1992) in order to investigate the effect of added conductive and dielectric fine materials. Their results indicated that if a conducting material such as steel is added to a gas-solid fluidized bed of a dielectric material such as glass beads, the potential difference across the bed, which is a measure of the charge in the bed, is considerably reduced.  Experiments were repeated with glass beads (GB-M) of 365 µm mean diameter and different proportions of an antistatic agent (Larostat 519) of 13 µm mean diameter. (For physical properties of the glass beads particles and the Larostat 519, see Tables 2.1 and 2.3) For these experiments the net cumulative charge in the bed, measured by the collision ball probes, is plotted versus time in Figures 5.12 and 5.13. Current flows through the ball probes due to direct charge transfer, obtained from the slopes of the cumulative charge curves in Figures 5.12 and 5.13, are plotted versus the wt% of Larostat 519 in Figures 5.14 and 5.15.  The addition of antistatic powders has been found to be effective in reducing the electrostatic charge buildup in gas-solid fluidized beds. Yao et al. (2002) found that within 1 h after adding 1% Larostat powder, both the standard deviation of voltage signals and the net charge transfer rate had dropped to insignificant levels. Park et al. (2002b) measured the electric field inside the bed by a wall-mounted ball probe and found that Larostat reduced the electrostatic charge buildup.  129  Cumulative charge (μC)  Cumulative charge (μC)  Cumulative charge (μC)  Chapter 5: Electrostatic Charge Reduction Technique  -0.5  Probe C  (a) -0.4  0.0  2.0  -0.3  Probe C' 1.5  (b)  1.5  -0.2  0.0  2.0 0.5  0.5  -0.1 0.0  -0.5  Probe B 1.5  (c)  -0.4  2.0  Probe B' 1.5 0.0 2.0 0.5  (d)  -0.3 0.0 0.5  -0.2 -0.1 0.0 -0.5  Probe A  (e)  1.5  0.0  -0.3  Probe A'  (f)  2.0  -0.4  2.0  0.5  1.5 0.0 0.5  -0.2 -0.1 0.0 0  60  120  180  Time (s)  240  300  0  60  120  180  240  300  Time (s)  Figure 5.12 Effect of antistatic agent (Larostat 519) on net cumulative charge as a function of time at different axial and radial positions in a binary mixture of GB-M particles and Larostat 519, dp = 365 and 13 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 917% and U-Umf = 0.05 m/s. Numbers on curves denote wt% of Larostat 519 in the binary mixtures. For probe positions, see Figure 2.9.  130  Cumulative charge (μC)  Chapter 5: Electrostatic Charge Reduction Technique  -0.05  Probe D  (e) -0.04 -0.03 2.0  -0.02  1.5 0.5  -0.01  0.0  0.00 0  60  120  180  240  300  Time (s) Figure 5.13 Effect of antistatic agent (Larostat 519) on net cumulative charge as a function of time in the freeboard in a binary mixture of GB-M particles and Larostat, dp = 365 and 13µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 9-17% and U-Umf = 0.05 m/s. Numbers on curves denote wt% of Larostat 519 in the binary mixtures. For position of probe D, see Figure 2.9.  8  -2.8 Current (nA ) .  (a)  Probe A' Probe B' Probe C'  (b)  Probe A Probe B Probe C  A  -2.1  1 C'  B'  B -1.4  4  C  A' 7  -0.7 0.0  0.5  1.0  1.5  Proportion of Larostat 519 (wt%)  2.0  0.0  0.5  1.0  1.5  2.0  Proportion of Larostat 519 (wt%)  Figure 5.14 Current flow through ball probes as a function of wt% of antistatic agent (Larostat 519) at different axial and radial positions in a binary mixture of GB-M and Larostat 519, dp = 365 and 13 µm, respectively, P = 379 kPa, T = 20 ± 2ºC, RH = 917% and U-Umf = 0.05 m/s. For probe positions, see Figure 2.9.  131  Chapter 5: Electrostatic Charge Reduction Technique -32 Current (nA ) .  Probe D -24 -16 -8 0 0.0  0.5  1.0  1.5  2.0  Proportion of Larostat 519 (wt%)  Figure 5.15 Current flow through ball probes as a function of wt% of antistatic agent (Larostat 519) in a binary mixture of GB-M and Larostat 519 in the freeboard, dp = 365 and 13 µm, respectively, P =379 kPa, T = 20 ± 2ºC, RH = 9-17% and U-Umf = 0.05 m/s. For position of probe D, see Figure 2.9. Previous work has exclusively focused on measuring the change in the bed charge due to the addition of fines (Mehrani et al, 2007b). However, to gain a better understanding of the effect of fines on electrostatic charge generation/dissipation inside the bed, it is also necessary to study the changes of the electrostatic behavior of the fines after their addition. Mehrani et al, (2007b) studied the effect of adding fine particles on charge generation/dissipation by investigating the change of the electrostatic behaviour of fines after their addition to a fluidized bed. In their experiments, fines with different physical and chemical surfaces structures such as Larostat 519, glass beads, silver-coated glass beads, catalyst and silica were examined. It was discovered that Larostat 519 and glass bead fines carried positive charges out of the bed of relatively coarse glass beads. It was also found that smaller fines carried more charges out of the bed. On the other hand, it was found that added fines (Larostat 519, catalyst, silica and silver coated glass beads) carried charges of different polarity out of fluidized beds of relatively large polyethylene particles, depending on the relative humidity of the fluidizing gas.  Results in Figures 5.12-5.15, show that the degree of electrification in the bed decreased as the proportion of Larostat 519 increased from 0 to 0.5 wt%. This is because Larostat particles first capture the moisture and then attach themselves to the surface of the large glass beads as 132  Chapter 5: Electrostatic Charge Reduction Technique shown in Figure 5.16 to build a moisture layer on the particle surface, thereby helping the dissipation of charges by the increased effective conductivity and consequently reducing electrostatic charge buildup. However, this trend is limited if the wt% of Larostat is too high, as the results show that the degree of electrification increased as the wt% of Larostat 519 increased beyond 1.0%. This may be due to the tendency of Larostat particles to adsorb water, and therefore as the proportion of Larostat increased in the bed, the degree of electrification in the bed increased as a result of decreasing the adsorbed water on the surface of large particles. In addition, it was noticed that when the Larostat proportion exceeded 1.0%, Larostat particles agglomerated. Hence, they may have lost their ability to reduce electrostatics. Appendix G develops an expression for the ratio of small-to-large particles in a binary mixture for full coverage on the large particle surface by a monolayer of the small particles. When this is applied to Larostat particles on the surface of glass beads, it is shown that 1.5 wt% would be needed for full monolayer coverage. In practice, it is unlikely that full monolayer coverage would be required, so that the maximum occurring at 1.0% is consistent with adding enough antistatic agent to effectively blanket the outside of the particles and prevent direct contacting of the beads.  The results for probe D in Figures 5.13 and 5.15 also show an increase in the degree of electrification in the freeboard region when the proportion of Larostat 519 exceeds a certain level, probably due to increased entrainment as the Larostat proportion in the bed increases. They also show that the entrained Larostat particles carried a charge polarity similar to the large particles in the bed. It would appear that the physical/chemical properties of the Larostat play a significant role in determining the charge reduction mechanisms. This is because Larostat 519 is a hygroscopic material, so that adsorbed moisture could affect the Larostat electrical behaviour as it comes in contact with glass beads.  5.3  Summary  In this Chapter, humidification of the fluidizing air was found to be effective in reducing charge buildup in the bed. Although the degree of electrification in the bed generally decreased with increasing relative humidity, its sensitivity to RH varied significantly, depending on the location of the probes. For the range of relative humidities (5-30%)  133  Chapter 5: Electrostatic Charge Reduction Technique explored, humidifying the fluidizing air was very effective when particle-particle charging was the dominant mechanism, but not very effective when particle-wall charging was the dominant mechanism. Adding small proportion (up to 2.0 wt%) of fine particles (<30 µm mean diameter) to a bed of relatively large glass beads (574 µm mean diameter) decreased electrostatic charging in the bed and enhanced fluidization stability. The degree of electrification in the freeboard increased with increasing proportion of fines. On the other hand, adding fine antistatic powder (Larostat 519) was also found to be effective in enhancing charge dissipation and reducing charge buildup in the bed. However, this trend was limited to a small proportion of Larostat particles (< 0.5 wt%). As the proportion of fine Larostat particles increased beyond 1.0%, they lost their ability to reduce electrostatics, probably due to increased agglomeration.  Figure 5.16 SEM images of sample discharged from the bed of binary mixture (GB-M and 0.5 wt% Larostat) after experiments, P =379 kPa, T = 20 ± 2ºC, RH = 13-17% and U-Umf = 0.05 m/s,  134  Chapter 6: Conclusions and Recommendations  Chapter 6: Conclusions and Recommendations  Electrostatic charges are almost unavoidable in multiphase systems such as gas-solid fluidized beds, due to repeated particle contacts and separation, associated with particles rubbing against each other and the column wall. These charges can be accumulated by dielectric materials such as glass beads or polyethylene resin. The build-up of electrostatic charges in fluidized beds can interfere with normal hydrodynamics of the bed, resulting in particle-wall adhesion, interparticle cohesion, electrostatic discharges, wall sheeting, and even fire and explosion. These can in turn cause serious operational problems, even reactor shutdown. Thus there is a pressing safety and economic incentive to prevent overaccumulation of charges in gas-solid fluidized beds. In order to do this, the relevant phenomena and mechanisms of generation, accumulation, dissipation and separation of electrostatic charges need to be well understood. Appropriate gas-solid fluidization systems equipped with, sensors and control systems, as well as proper measurement techniques, are essential to properly explore and understand charging mechanisms.  The goals of this study were to gain a better understanding of electrostatic phenomena and to characterize electrostatics charge generation, accumulation and dissipation at different locations in gas-solid fluidized beds and to study the influence of particles properties and operational variables such as pressure, temperature, superficial gas velocity and relative humidity. An elevated-pressure fluidization system was built and collision ball probes were developed to achieve these objectives.  6.1  Conclusions  Nearly all industrial gas-solid fluidized bed reactors operate at temperatures well above ambient, and many, such as those used to produce polyolefin (UNIPOL reactors), also operate at elevated pressures, over a range of superficial gas velocities. It is therefore important to know how fluidized beds behave under different operating conditions and how the degree of electrification in the bed influenced by these operating variables.  135  Chapter 6: Conclusions and Recommendations All previous studies on electrostatic in gas-solid fluidized beds were performed at ambient conditions (25 ºC and atmospheric pressure). In this study, the influence of operating pressure on the degree of electrification in the bed was investigated for two types of glass beads and high-density polyethylene resin of 574, 365 and 450 µm mean diameter. Freeboard absolute pressures varied from 379 to 724 kPa, whereas other operating parameters were maintained nearly constant to determine the effect of pressure. As pressure increased the degree of electrification in the bed increased, probably due to an increase in bubble rise velocity, frequency, volume fraction and a slight increase in bubble size. The magnitude of the increase in electrostatics differed from location to location. For example, as pressure increases, particle-particle and particle-wall collisions near the distributor contributed heavily to static charge generation in the bed, probably due to increased bubble frequency in the lower portion of the bed. At higher levels, electrostatic charge generation was greater near the column axis than near the wall as pressure increased, because at elevated pressures bubble flow is increasingly concentrated toward the axis. The maximum static charge was found to occur at a height of 0.55 m, near where the maximum bubble size occurred. Based on these findings it was concluded that bubble behaviour and particles rubbing against each other in the region surrounding rising bubbles and against the wall play a major role in determining the degree of bed electrification.  Freely bubbling experiments with glass beads and low-density polyethylene particles at temperatures up to 90ºC indicated that the polarity of net cumulative charge in the bed reversed as the bed temperature increased. At higher temperatures the bed exhibited smoother fluidization, and the amount of fines adhering to the wall of the column was much less at higher than at lower temperatures. Changes in physical and electrical properties of the particles, as well as the wall of the column were likely responsible for changes in charging mechanisms. However, the effect of temperature is complex due to interaction between temperature and relative humidity of the fluidizing air.  The degree of bed electrification increased as the superficial gas velocity increased. This is due to formation of bigger bubbles that cause higher bubble rise velocities and enhanced the motion of particles. The rate of this increase and its significance depended on the particle size distribution and location in the bed. Charge fluctuations increased as the superficial gas 136  Chapter 6: Conclusions and Recommendations velocity increased. For larger particles, as the superficial gas velocity increased, the maximum electrostatic charge was reached in the region close to the distributor due to the increased bubble formation there. For smaller particles, as the superficial gas velocity increased beyond 0.1 m/s, the rate of charge transfer became less sensitive to the superficial gas velocity in the lower portion of the bed and near the wall. The rate of charge transfer became insensitive to the superficial gas velocity at high levels in the bed and near the vertical axis, possibly due to reduced contact between particles and ball probe in the upper portion of the bed and near the vertical axis due to a lower concentration of particles there, caused by the higher concentration of gas phase in that region.  At higher superficial gas velocities the polarity in the freeboard region was found to be opposite to that at lower measurement levels inside the bed, indicating that entrained fines carried one charge, leaving behind a net charge of opposite polarity inside the bed.  Free-bubbling experiments with four types of mono-disperse glass beads of 693, 574, 365 and 65 µm mean diameter, showed that for Geldart group B particles the degree of electrification in the bed slightly increased with decreasing particle size, whereas charging for Geldart group A particles was significantly greater than for group B particles. These increases are due to the increase in total particle surface area, enhancing contact charging and charge generation in the bed. For group A particles it was also observed that the degree of bed electrification in the bed was further enhanced due to reduced charge dissipation from the charged particle to the wall of the grounded column caused by a layer of fine particles on the inner wall of the column.  The effect of increasing the proportion of small particles (65 µm mean diameter) in binary mixtures of small and large particles of glass beads of 65 and 693 µm mean diameter, was also studied, with the wt% of small particles in the mixtures varied from 10-90%. The degree of electrification in the bed generally increased as the proportion of small particles increased, possibly due to increased particle-particle and particle-wall collisions. In these experiments inconsistent trends were observed for different mixtures. This was attributed to the high sensitivity of electrostatics to the superficial gas velocity, because the actual size distribution  137  Chapter 6: Conclusions and Recommendations in beds of binary mixtures continually changed with time, depending on the rates of segregation and entrainment.  Bipolar charging was observed in all experiments where polyethylene resin particles constituted the bed materials. It was found that large particles charged negatively due to interparticle charge transfer and charge separation between them and the wall, whereas fine particles charged positively due to charge transfer between fines and the wall and to charge separation between fines and the large particles in the bed. On the other hand, large and small particles in the polyethylene resin differ in their nature, extent of contact surfaces, weights, copolymer content, surface roughness, chemical composition and tendency to adsorb moisture. This could significantly influence the charging mechanisms in beds of polyethylene resin. Net cumulative charges of polyethylene were inconsistent for some cases and at different locations in the bed. Large fluctuations were observed for most probes.  Comparison of the results of this work indicated that the magnitude of net cumulative charges for glass beads was much higher than for polyethylene particles. This is attributed to the sizable difference between their densities, which significantly affected the collision forces between particles. The differences in the net cumulative charges between high-density and low-density polyethylene were negligible.  In order to study the effect of particle chemical composition on their charging behaviour, two different types of low-density polyethylene resin of 523 and 502 µm mean diameter, respectively, with different chemical composition were examined. It was found that the polarity was opposite for these two species, probably due to the different chemical composition of the particles, and possibly due to different tendencies to adsorb moisture and react with adsorbed water molecules.  Overall, for polyethylene particles, depending on their physical properties and chemical compositions, charged particles may contain both positively and negatively charged areas on their surfaces, while the dominant polarity determines the net surface charge.  138  Chapter 6: Conclusions and Recommendations Humidifying the fluidizing gas was investigated for 693 µm glass beads. As expected, as the relative humidity of fluidizing air increased, the degree of electrification generally decreased due to increased charge dissipation resulting from increased conductivity around the surfaces of particles. The sensitivity of the degree of electrification in the bed to the effect of relative humidity varied significantly, depending on the axial and radial location of the probes as well as the relative humidity of the fluidizing air. For the range of relative humidities (5-30%) explored, humidifying the fluidizing air was effective when particle-particle charging was the dominant mechanism, but not very effective when particle-wall charging was the dominant mechanism, probably because the wall is not influenced by the relative humidity of the fluidizing air in the same manner as the particles.  The effect of adding fine particles on charge generation and/or dissipation in the bed was studied by investigating the change of the degree of electrification in binary mixtures of 574 µm glass beads and different proportions of 30 µm glass beads. Freely bubbling experiments indicated that the degree of electrification decreased as the proportion of fines increased from 0.0 to 2.0 wt%. On the other hand, an increase in the degree of electrification in the freeboard region with increasing fine wt% was observed, presumably due to increased entrainment. Particles in the freeboard charged oppositely to those in the bed. This was attributed to charge transfer between the fines and the wall and to charge separation between the fines and the larger particles. As the proportion of fines increased from 0.0 to 2.0 wt%, charge fluctuations decreased and the bed exhibited smoother fluidization, possibly because the fines acted as spacers or lubricants between larger particles, reducing contacts between the large particles.  The effect of adding antistatic agents on charge generation and/or dissipation in the bed was also studied for glass beads of 365 µm mean diameter. The degree of electrification decreased as the proportion of Larostat 519 increased from 0.0 to 0.5 wt%. It was concluded that Larostat particles tend to attach to the surface of the large glass beads, thereby reducing the large-particle-particle collisions and consequently reducing electrostatic charge buildup. However, the degree of electrification increased when the wt% of Larostat 519 increased beyond 1.0%. This was attributed to the tendency of Larostat particles to adsorb water and to agglomerate. The degree of electrification in the freeboard region increased as the proportion 139  Chapter 6: Conclusions and Recommendations of Larostat increased, probably due to its increased entrainment. The entrained Larostat particles carried a charge polarity similar to that of the large particles in the bed.  6.2  Recommendations for Future Work  Bed temperature plays a significant role in charging mechanisms in gas-solid fluidized beds. Thus, further work is needed in this area. The effect of bed temperature on charging mechanisms needs to be further investigated by considering: •  Relative humidity of the fluidizing gas shall be well controlled either by proper drying and humidification systems or by using extra dry air to eliminate or reduce the interaction between temperature and relative humidity of the fluidizing air and isolate the effect of temperature.  •  The temperature of the fluidizing gas shall be adjusted to ensure a uniform temperature everywhere in the bed.  Since many commercial fluidization units operate in the turbulent flow regime, it is important to study the effect of different operating conditions on the degree of electrification at higher superficial gas velocities.  Since in the current work the effect of operating variables and particle properties were investigated separately, future work should investigate interactions between operating variables as well as between operating variables and particle properties.  The data collected in this project at different operating conditions and for different particle types should be useful in developing semi-empirical models and/or verifying and validating theoretical models for electrostatics in gas-solid fluidized beds.  140  Literature Cited  Literature Cited  Ali, F.S., Ali, M.A., Ali, R.A. and Inculet, I.I., Minority charge separation in falling particles with bipolar charge, J. Electrostatics, 45, pp.139-155, 1998. Ally, M.R. and Klinzing, G.E., Inter-relation of electrostatic charging and pressure drops in pneumatic transport, Powder Technology, 44, pp.85-88, 1985. Baretto G.F., Yates, J.G. and Rowe, P.N., The effect of pressure on the flow of gas in fluidized beds of fine particles, Chem. Eng. Sci., 38, pp.1935-1945, 1983. Bi, H.T., Electrostatic phenomena in gas-solid fluidized beds, China Particuology, 3, No. 6, pp.395-399, 2005. Bi, H.T., Grace, J.R. and Zhu, J.X., Propagation of pressure waves and forced oscillation of fluidized beds and their effect on measurements of local hydrodynamics, Powder Technology, 82, pp.239-253, 1995. Boland, D. and Geldart, D., Electrostatic charging in gas fluidized beds, Powder Technology, 5, pp.289-297, 1971. Botterill, J.S. and Desai, M., Limited factors in gas-fluidized bed heat transfer, Powder Technology, 6, pp.231-238, 1972. Chen, A.H., Bi, H.T. and Grace, J.R., Measurements of particle charge-to-mass ratios in a gas-solid fluidized bed by a collision probe, Powder Technology, 135-136, pp.181-191, 2003a. Chen, A.H., Bi, H.T. and Grace, J.R. Effect of charge distribution around bubbles on charge induction and transfer to a ball probe in gas-solid fluidized beds, J. of Electrostatics, 58, pp.91-115, 2003b. Chen, A.H., van Willigen, F.K., Bi, H.T., Grace, J.R. and van Ommen, J.R., Measurements of charge distribution around a rising bubble in a 2-D fluidized bed, AIChE Journal, 52, pp.174-184, 2006. Cheung, L., Nienow, A.W. and Rowe P.N., Minimum fluidization velocity of a binary mixture of different sized particles, Chemical Engineering Science, 29, pp.1301-1303, 1974. Chiba, S., Chiba, T., Nienow, A.W. and Kobayashi, H., The minimum fluidization velocity, bed expansion and pressure-drop profile of binary particle mixtures, Powder Technology, 22, pp.255-269, 1979. Chitester, D.C., Kornosky, R.M., Fan, L.S. and Danko, J.P., Characteristics of fluidization at high pressure, Chemical Engineering Science, 39, pp.253-261, 1984. 141  Literature Cited Choi, J. H., Son, J.E. and Kim, S.D., Generalized model for bubble size and frequency in gas-fluidized beds, Ind. Eng. Chem. Res., 37, pp.2559-2564, 1998. Ciborowski, J.S. and Wlodarski, A., On electrostatic effects on fluidized beds, Chemical Engineering Science, 17, pp.23-32, 1962. Clift, R. and Grace J.R., The mechanism of bubble break-up in fluidized beds, Chemical Engineering Science, 27, pp.2309-2310, 1972. Cross, J.A., Electrostatics: Principles, Problems and Applications, Adams Higler, Bristol, 1987. Cutnell, J.D. and Johnson, K.W., Physics, 2nd ed., John Wiley & Sons Inc., New York, pp.497-693, 1992. Darton, R.C., Lanauze, J.F., Davidson, J.F. and Harrison, D., Bubble growth due to coalescence in fluidized beds, Transactions Institution of Chemical Engineers, 55, pp.274-280, 1977. Davies, D.K., Charge generation on dielectric surfaces, J. Physics D: Applied Physics, 2, pp.1533-1540, 1969. Davidson, J.F. and Harrison, D., Fluidized particles, Cambridge University Press, Cambridge, 1963. Fan, L.S. and Zhu, C., Principles of Gas-Solid Flows, Cambridge University Press, New York, 66-70, 1998. Fang, W., Jingdai, W. and Yongrong, Y., Distribution of electrostatic potential in a gas-solid fluidized bed and measurement of bed level, Industrial and Engineering Chemistry Research, 47, pp. 9517-9526, 2008. Fujino, M., Ogata, S. and Shinohara, H., The electric potential distribution profile in a naturally charged fluidized bed and its effects, International Chemical Engineering, 25, pp.149-159, 1985. Geldart, D., The effect of particle size and size distribution on the behaviour of gas-fluidized beds, Powder Technology, 6, pp.201-205, 1972. Geldart, D., Types of gas fluidization, Powder Technology, 7, pp.285-292, 1973. Goossens, W.R.A., Dumont, G.L. and Spaepen G.L., Fluidization of binary mixtures in laminar flow region, Chem Eng Prog Symp Ser, 67(116), pp.38-45, 1971. Grace, J.R. and Matsen, J.M. eds., Fluidization, Plenum Press, New York, 1980. Grace, J.R., Fluidized-bed hydrodynamics, Chapter 8 in Handbook of Multiphase Systems, G. Hetsroni ed., McGraw-Hill, pp.8-7, 1982. 142  Literature Cited Grace, J.R., Contacting modes and behaviour classification of gas-solid and other two-phase suspensions, Canadian Journal of Chemical Engineering, 64, pp.353-363, 1986. Guardiola, J., Ramos, G. and Romero, A., Electrostatic behavior in dielectric/conductor fluidized beds, Powder Technology, 73, pp.11-19, 1992.  binary  Guardiola, J., Rajo, V. and Ramos, G., Influence of particle size, fluidization velocity and relative humidity on fluidized bed electrostatics, J. of Electrostatics, 37, pp.1-20, 1996. Haider, A. and Levenspiel, O., Drag coefficient and terminal velocity of spherical and nonspherical particles, Powder Technology, 58, pp.63-70, 1989. Harper, W.R., Contact and Frictional Electrification, Oxford University Press, London, 1967. Harrison, D., Davidson, J.F., and De Kock, J.W., On the nature of aggregative and particulate fluidization, Transactions Institution of Chemical Engineers, 39, pp.202211, 1961. Haenen H.T.M., Experimental investigation of the relationship between generation and decay of charges on dielectrics, J. of Electrostatics, 2, pp.151-173, 1976. Hoffmann, A.C. and Yates, J.G., Experimental observation of fluidized beds at elevated pressures, Chemical Engineering Communication, 41, pp.133-149, 1986. Hull H.H., A method for studying the distribution and sign of static charges on solid materials, J. Applied physics, 20, pp. 1157-1166, 1949. Inculet, I.I., Castle, G.S.P. and Aartsen, G., Generation of bipolar electric fields during industrial handling of powders, Chemical Engineering Science, 61, pp.2249-2253, 2006. Jacob, K.V. and Weimer, A.W., High-pressure particulate expansion and minimum bubbling of fine carbon powder, A.I.Ch.E., 33, pp.1698-1706, 1987. J. Chemical Engineering Japan, 18, 113-118, 1985. Jiang, P., Bi, H.T., Liang, S.C. and Fan L.S., Hydrodynamic behaviour of circulating fluidized bed with polymeric particles, AIChE Journal, 40, 2, pp.193-206, 1994. Kai, T., and Furusaki, S., Behaviour of fluidized beds of small particles at elevated temperatures, J. Chemical Engineering Japan, 118, pp.113-118, 1985. Katz, H. and Sears, J.T., Electric field phenomena in fluidized and fixed bed, Canadian Journal of Chemical Engineering, 47, pp.50-53, 1969. Kawabata, J., Yumiyama, M., Tazaki, Y., Honma, S., Chiba, T., Sumiya, T. and Ehdo, K., Characteristics of gas fluidized beds under pressure, J. Chemical Engineering Japan, 14, pp.85-89, 1981.  143  Literature Cited King, D.F. and Harrison, D., The dense phase of a fluidized bed at elevated pressure, Transactions Institution of Chemical Engineers, 60, pp.26-30, 1982. Knowlton, T.M., High-pressure fluidization characteristics of several particulate solids, AIChE Symp Ser, 52(61), pp.22-28, 1997. Kunii, D. and Levenspiel, O., Fluidization Engineering, 2nd Edition, Butterworth-Heinemann, USA, 1991. Lewis, W.K., Gilliland, E.R. and Bauer, W.C., Characteristics of fluidized particles, Industrial and Engineering Chemistry, 41, pp.1104-1117, 1949. Li, J. and Kuipers, J.A., Effect of pressure on gas-solid flow behavior in dense gas-fluidized beds: a discrete particle simulation study, Powder Technology, 127, pp.173-184, 2002. Maron, S.H. and Prutton, C.F., Principles of Physical Chemistry, 4th ed., pp.714, MacMillan, New York, 1965. Mehrani, P., Characterization of electrostatic charges in gas-solid fluidized beds, Ph.D. Dissertation, University of British Columbia, 2005. Mehrani, P., Bi, H.T. and Grace, J.R., Bench-scale tests to determine mechanisms of charge generation due to particle–particle and particle–wall contact in binary systems of fine and coarse particles, Powder Technology, 173, pp.73-81, 2007a. Mehrani, P., Bi, H.T. and Grace, J.R., Electrostatic behavior of different fines added to a Faraday cup fluidized bed, J. of Electrostatics, 65, pp.1-10, 2007b. Montgomry, D.J., Static electrification of solids, Solid State Physics, 9, pp.139-196, 1959. Mori, S. and Wen, C.Y., Estimation of bubble diameter in gaseous fluidized beds, AIChE Journal, 21, pp.109-115, 1975. Moughrabiah, W.O., Grace, J.R. and Bi, H.T., Effect of pressure, temperature, and gas velocity on electrostatics in gas-solid fluidized beds, Industrial and Engineering Chemistry Research, 48, pp. 320-325, 2008. Mii, T., Yoshida, K. and Kunii, D., Temperature effect on the characteristics of fluidized beds, J. Chemical Engineering Japan, 6, pp.100-102, 1973. Miller, C.O. and Logwinuk, A.K., Fluidization studies of solid particles, Industrial and Engineering Chemistry, 43, pp.1220-1226, 1951. Newton, D., Fiorention, M. and Smith, G.B., The application of X-ray imaging to the developments of fluidized bed processes, Powder Technology, 120, pp.70-75, 2001. Osberg, G.L. and Charlesworth, D.H., Elutriation in a fluidized bed, Chemical Engineering Progress, 47, pp.566-570, 1951. 144  Literature Cited Olowson, P.A. and Almstedt, A.E., Influence of pressure and fluidization velocity on the bubble behaviour and gas flow distribution in a fluidized bed, Chemical Engineering Science, 45, pp.1733-1741, 1990. Olowson, P.A. and Almstedt, A.E., Influence of pressure on the minimum fluidization velocity, Chemical Engineering Science, 46, pp.637-640, 1991. Olssen, S.E., Wiman, J. and Almstedt A.E., Hydrodynamics of a pressurized fluidized bed with horizontal tubes: influence of pressure, fluidization velocity and tube bank geometry, Chemical Engineering Science, 50, pp.581-592, 1995 Park, A., Electrostatic charging in gas fluidized beds, M.A.Sc. Thesis, University of British Columbia, 2000. Park, A., Bi, H.T. and Grace, J.R., Modeling charge transfer and induction in gas-solid fluidized beds. J. Electrostatics, 55, pp.135-158, 2002a. Park, A., Bi, H.T. and Grace, J.R., Reduction of electrostatic charges in gas-solid fluidized bed, Chemical Engineering Science, 57, pp.153-162, 2002b. Reitz, J.R., Milford, F.J. and Christy, R.W., Foundations of Electronic Theory, 4th ed., pp.1126, Addison-Wesley Publishing Company, Massachusetts, 1993. Revel, J., Gatumel, C., Dodds, J.A. and Taillet, J., Static charge elimination in a slugging fluidized bed, 4th World Congress on Particle Technology, Sydney, pp.1-8, 2002. Riley, N.A., Projection sphericity, J. Sedimentary Petrology, 11, v.2, pp.94-97, 1941. Rowe, P.N., The effect of pressure on minimum fluidization velocity, Chemical Engineering Science, 39, 173-174, 1984. Rowe, P.N., Prediction of bubble size in a gas fluidized bed, Chemical Engineering Science, 31, pp.285-288, 1976 Seanor, D.A., Polymer Science, A.D. Jenkins ed., Amsterdam: North Holland, 2, p.1187, 1972. Senior, R.C., Circulating fluidized bed fluid and particle mechanics: modeling and experimental studies with application to combustion, Ph.D. Dissertation, University of British Columbia, 1992. Sidorenko, I. and Rhodes, J., Influence of pressure on fluidization properties, Powder Technology, 141, pp.137-154, 2004. Sittiphong, N., George A.H. and Bushrell, D., Bubble eruption diameter in a fluidized bed of large particles at elevated temperatures, Chemical Engineering Science, 36, pp.12591260, 1981.  145  Literature Cited Tardos, G. and Pfeffer, R., A method to measure electrostatic charge on a granule in a fluidized bed, Chemical Engineering Communications, 4, pp.665-671, 1980. Varadi, T. and Grace, J.R., High pressure fluidization in two-dimensional bed, in: Davidson, J.F. and Keairns D.L., eds., Fluidization, Cambridge University Press, Cambridge, 5558, 1978. Van Krevelen, D.W., Properties of Polymer; Their Correlation with Chemicals; Their Numerical Estimation and Prediction from Additive Group Contributions, 3rd ed., Elsevier Science, New York, USA, 1990. Wen, C.Y. and Yu, Y.H., A generalized method for predicting the minimum fluidization velocity, AIChE J., 12, pp.610-612, 1966. Wolny, A. and Kazmierczak, W., Triboelectrification in fluidized bed of polystyrene, Chemical Engineering Science, 44, pp.2607-2610, 1989. Wolny, A. and Opaliniski, I., Electric charges neutralization by addition of fines to a fluidized bed composed of coarse dielectric particles. J. Electrostatics, 14, pp.279-289 1983. Yao, L., Bi, H. and Park, A., Characterization of electrostatic charges in freely bubbling fluidized beds with dielectric particles, J. Electrostatics, 56, pp.183-197, 2002. Yang, W. C., Bubbling fluidized bed, Chapter 3 in Handbook of Fluidization and FluidParticle Systems, ed. W.C. Yang, Marcel Dekker, New York, pp.58-111, 2003. Yates, J.G., Effect of temperature and pressure, Chapter 5 in Handbook of Fluidization and Fluid-Particle Systems, ed. W.C. Yang, Marcel Dekker, New York, pp.58-111, 2003. Yoshida, K., Ueno, T. and Kunii, D., Mechanism of bed to wall heat transfer in a fluidized bed at high temperatures, Chem. Eng. Sci., 29, pp.77-82, 1974. Zhao, H., Castle, G.S.P., Inculet, I.I. and Bailey, Bipolar charging in poly-disperse polymer powder in industrial processes, IEEE, 2, pp.835-841, 2000. Zhao, H., Castle, G.S.P. and Inculet, I.I., The measurement of bipolar charging in polydisperse using a vertical array of Faraday pail sensors, J. Electrostatics, 55, pp.261-278, 2002. Zhao, H., Castle, G.S.P., Inculet, I.I. and Bailey, Bipolar charging in polydisperse polymer powder in fluidized beds, IEEE Trans. Ind. Appl., 39, pp.612-618, 2003. Zimmer, E., Die elektrostasche aufladung von Hochpolymeren Isolierstoffen, Kustoffe, 60, pp.465-468, 1970.  146  Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds  Table A.1 Summary of previous studies on electrostatics in gas-solid fluidized beds Bed Reference  material  dp (µm)  Column Hmf (m)  material  dimension dc, (m)  Fluidizing velocity, UF, (m/s)  Measurement technique  Measured variable  RH (%)  Ciborowski Vinyl 490-750 NS* 3-D 0.060 0.10-0.43 Collision ball and polyacetate Glass probe Wlodarski Polystyrol 400-490 (1962) Sand 300-400 Conclusions: • Charge generation is directly proportional to fluidizing gas velocity. • Relative humidity strongly influences rate of charge dissipation. • Introducing moisture is ineffective at temperatures above 80ºC. • Height above the distributor equal to the static bed height corresponds to highest charge build-up.  Avg. charge transfer  0-12 (g/kg)  Katz and Sears (1969)  Bed stability  NS*  electric field potential, current  15-80  Glass beads  250-297  Pyrex Silica gel  88-177 177-297  0.12  3-D Lucite & 3-D Glass  0.075  0.29  Applied electric potential  0.025  Conclusions: • Moisture in fluidizing gas increases surface conductivity, causing dissipation to occur more rapidly. • Particles in a fluidized bed are probably hetero-polar. Boland and Glass beads 100-800 0.60 2-D 0.50 x NS* Shielded Geldart Perspex 0.013 induction probe (1971) Conclusions: • Electrostatics is generated by motion of particles around bubbles, particularly in the wake. • Main particle size influences charge generation. • RH influences strongly rate of charge dissipation. * NS = Not Specified.  147  Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds  Bed Reference Tardos and Pfeffer (1980)  material Porcelain  dp (µm) 2000  Column Hmf (m) 0.26  material Plastic  Fluidizing dimension velocity, UF, (m/s) dc, (m)  0.05  0.92-1.25  Measurement technique Collision Probe Faraday cage  Measured variable Mean current (charge-to-mass ratio)  RH (%) 20-60  Conclusions: • Developed a procedure to measure the mean charge on particles. Ally and Glass beads, 75-314 0.0254 Pneumatic Collision Probe charge-to-mass 10-80 NS* Copper, Plexiglas Klinzing Plexiglas & 145 transport ratio Glass (1985) Cu 196 Conclusions: • Charge-to-mass ratio of copper particles in a glass tube is much higher than for a Plexiglas tube because glass can extract electrons from copper more readily than Plexiglas. • Operated under pneumatic transport conditions. Fujino et al. Glass beads, 200-250 NS* Perspex, 0.067 0.21-0.5 collision probe elect. potential in NS* (1985) neobeads & 200-540 iron (spherical the bed & specific PMMA 540 terminal) charge (chargeto-mass ratio) Conclusions: • Bed electric potential is affected by column material, particle size and nature of the particles. • Gas velocity and RH exert only a weak effect on specific charge of the particles. Wonly and Polystyrene, 475 0.2 NS* 0.2 x 0.2 NS* Air-plate Specific charges 10-70 Opaliniski Al. powder, 0.1-15 (Square) capacitor (charge-to-mass (1989) & ratio), NaCl 1-60 Conclusion: • Particles in a fluidized bed are more probably of hetero-polar type. • There is no relation between net charges on particles and electrostatic effects on fluidized bed hydrodynamics. • RH at 70% is the best modifying agent for electrostatic effects. * NS = Not Specified.  148  Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds  Bed Reference Guardiola et al. (1992)  Column  material  dp (µm)  Hmf (m)  Glass beads steel  500-590 420-500  0.05  material Perspex  Fluidizing dimension velocity, UF, (m/s) dc, (m)  0.052  0.11-0.66  Measurement technique Capacitance probe  Measured variable Average electrical potential  RH (%) 5-40  Conclusions: • When conducting material is added to bed of dielectric material, potential difference across bed is reduced. • Increase in fluidization velocity causes increase in electrostatic generation. Guardiola et al. (1996)  Glass beads  250-297 297-350 350-420  0.09  3-D Perspex  0.044  0.076-0.19 Capacitance probe  Average electrical potential  10-85  Charge of particles (chargeto-mass ratio)  NS*  Conclusions: • Increase in fluidization velocity causes increase in electrostatic generation. • Effects of RH are complex and depend on type of fluidization. • Degree of bed electrification increases with increasing bubble size. Ali et al. (1998)  Polyamid powder, Acrylic based powder, Mixture of both  98 80  NS*  3-D PVC  0.25 x 0.25 x 0.6  NS*  System of nine Faraday pails  80  Conclusions: • A bipolar charged sample of polymer powder can be electrically separated by pouring. • Charge separation is a function of charge on individual particles and height at which sample is poured. * NS = Not Specified.  149  Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds  Bed Reference Zhao et al. (2000)  material Polymer powder & Tio2 Mixture of both  Column dp (µm)  <100  Hmf (m)  material  NS*  3-D PVC  dimension dc, (m)  0.25 x 0.25 x 0.6  Fluidizing velocity, UF, (m/s) 1.14x10-3 Vol. air flow rate (m3/s)  Measurement technique  Measured variable  RH (%)  System of seven Faraday pails mounted vertically in cascade.  Charge of particles (charge-to-mass ratio)  1.6-6.6  Conclusions: • Bipolar charging was observed in all experiments. Small particles charged negatively whereas large particles charged positively, even though the net charge of the powder was negative or positive. Park et al. Glass beads 321 NS* 2-D 0.307 x 0.122 Umf Collision ball Voltage (Vmax,) 17 (2002a) Plexiglas 1.24 background probe (Vmin), 0.022 velocity, charge transfer thick single (Ctransfer) and bubble max magnitude injection of charge (Cmax) Conclusions: • A simple mathematical model was developed. • Model shows that induced charge is insensitive to thickness of the layer of charged particles. • Charge transfer during collision of particles with a probe is influenced by particle velocity. • Model with uniform charge on surface of bubble gives better predictions. Park et al. Glass beads 321 NS* 2-D 0.307 x Collision ball Voltage (Vmax,) 6-98 0.122 Umf (2002b) Plexiglas 1.24 background probe (Vmin), charge Polyethylene 378 0.022 velocity, transfer, 3-D thick (Ctransfer) and 0.085 max magnitude Plexiglas 0.0889 of charge Cmax Conclusions: • Electrostatic charges increase as bubble size increases. • RH of 40-80% reduces the accumulation of charges by increasing surface conductivity, thereby enhancing charge dissipation. * NS = Not Specified. 150  Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds  Bed Reference  material  Column dp (µm)  Hmf (m)  material  dimension dc, (m)  Fluidizing velocity, UF, (m/s)  Yao et al. Polyethylene 318-378 0.35 3-D 0.089 0.124-0.35 (2002) & Larostat Plexiglas Conclusions: • Standard deviation of voltages increases with increasing superficial gas velocity. • Charge build-up in the vicinity of bubble decreases as bubble size decreases. • Addition of antistatic powders is effective in reducing charge build-up. Chen et al. Glass beads 321 0.45 2-D 0.307 x 0.122 Umf (2003a) Plexiglas 1.24 background 0.022 velocity, thick single bubble injection Conclusions: • Particle charge-to-mass ratio is insensitive to size of single bubbles passing probes. • Increase of RH for RH<40% resulted in significant decrease in specific charges. Chen et al. Glass beads 321 0.45 2-D 0.307 x 0.358 (2003b) Plexiglas 1.24 0.022 thick  Measurement technique  Measured variable  RH (%)  Collision ball probe  SD of voltage signals  15-85  Collision ball probe  Specific charges 10-80 (charge-to-mass ratio) using model  Collision ball probe  charge transfer (Ctransfer) and max and min magnitude of charge curve (Cmax), (Cmin)  17  Conclusions: • Previously developed model (Park et al., 2002a) was modified & tested against (Park et al., 2002a) experimental results. • Model simulation results change slightly when background charge density and distribution of charge density around a spherical bubble are introduced into the model. • Significant improvement in agreement between model and experimental results is achieved when contributions of the wake and drift are included. * NS = Not Specified.  151  Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds  Bed Reference  material  dp (µm)  Column Hmf (m)  material  Fluidizing velocity, dimension UF, (m/s) dc, (m)  Measurement technique  Measured variable  RH (%)  Bi (2005) Conclusions: • Level of charge buildup on particles is determined by the balance of charge generation and dissipation rates. • Prediction of the electrostatic forces associated with charged particles in gas-solid fluidized beds requires coupling of charge generation and dissipation mechanisms with hydrodynamic behaviour, particle flow and mixing patterns of fluidized bed. • Highly charged particles due to particle-particle contact in the near-bubble region tend to segregate and concentrate into bubble wake region, causing non-uniform charge density distribution. • Dissipation of charges from particle surfaces in fluidized beds is the rate-limiting step because of the low conductivity of the dense particle mixture. Mehrani et al. Glass beads 10-80 3-D 0.1 diam. 0.21 On-line Faraday Cumulative NS* (2005) 500-600 copper 2.1 high cup fluidized charge, and and 0.01 thick bed system (charge-to-mass Teflon ratio) Conclusions: • Particle–gas contacting had a negligible effect on net charges generated inside a gas–solid fluidized bed. • Net charge inside the bed resulted from the fines entrained from the fluidization column carrying charges. • For the conditions studied, air ionization had a negligible effect on particle charge dissipation. Chen et al. Glass beads 565 0.7 2-D 0.28 x 1.78x10-3 Four induction Changes in NS* (2005) Plexiglas 1.24 Vol. air probes induced charge 0.014 flow rate thick (m3/s) Conclusions: • Emulsion phase far from the bubble was charged negatively for the glass beads used in the experiments. • A decrease of charge density moving inward from the emulsion phase bubble interface, with essentially zero charge density inside the air bubble, was observed. • Charge distribution was shown to be nontrivial for this system as particles in the wake are strongly charged  152  Appendix A: Summary of Previous Studies on Electrostatics in Gas-Solid Fluidized Beds  Bed Reference  material  dp (µm)  Column Hmf (m)  material  Fluidizing velocity, dimension UF, (m/s) dc, (m)  Measurement technique  Measured variable  RH (%)  Inculet et al. Starch 20-2000 Metallic 35 (long) NS* Faraday cup Charge-to-mass NS* (2006) pipe 0.3 diam. ratio Conclusions: • Bipolar charging was observed when the powder was allowed to free fall under gravity from the end of a conveying pipe. • Fine particles charged negatively, whereas large particles charged positively, even though the net charge was positive. Chen et al. (2006)  Glass beads  565  0.78  2-D Plexiglas  0.28 x 1.24 0.014 thick  0.11x10-3 Vol. air flow rate (m3/s)  Eight induction probes  Changes in induced charge  NS*  Conclusions: • Adding one or more probe outside the bubble increases the reconstruction resolution in the region outside the bubble. • To obtain more accurate reconstruction results, probes should be placed from center to the edge of the bubble. • More negatively charged wakes were observed, likely due to bubble splitting and coalescence near the probes. Moughrabiah et al. (2008)  Glass beads LLDPE HDPE  321 600 450  0.55  3-D Stainless steel  0.15 diam. 2.0 high  (U-Umf) = 0.04-0.1 m/s  Eight collision ball probes  Net cumulative charge  Faraday cup  Charge density  9-12  Conclusions: • The degree of bed electrification increased with increasing pressure probably due to an increase in bubble rise velocity, frequency and volume fraction. Maximum static charges were measured at approximately two-thirds of the expanded bed height and near the axis of the column, probably due to an increase in the interactions between bubbles and particles there. • As pressure increased, particle-particle and particle-wall collisions in the region just above the distributor plate and near the wall contributed heavily to increased generation of electrostatic charges. • As the fluidizing gas velocity increased, the degree of electrification increased due to the increase in bubble rise velocity and flow rate. Temperature also was found to play a significant role in charge generation. * NS = Not Specified.  153  Appendix B: Equipment Photographs and Engineering Drawings  Appendix B: Equipment Photographs and Engineering Drawings B.1  Experimental Apparatus  Pressure control valve upstream of vessel Expanded section  Straight section  Flow control valve downstream of vessel  Control panel  Figure B.1 Photograph of experimental apparatus.  154  Appendix B: Equipment Photographs and Engineering Drawings  B.2  KAESER Air Compressor  KAESER rotary screw compressor  Vapour-removal filter Sigma control panel  Refrigerating dryer  Figure B.2 Photograph of high-pressure air system.  155  Appendix B: Equipment Photographs and Engineering Drawings  B.3  Fluidization Section  Bimetal dial thermometer  Heating tape  Collision ball probes  Collision ball probes  Figure B.3 Photograph of fluidization section (side view)  156  Appendix B: Equipment Photographs and Engineering Drawings  B.4  Sight Glass  Sight glass  Figure B.4 Photograph of sight glass allowing visualization of fluidization inside column.  157  Appendix B: Equipment Photographs and Engineering Drawings  B.5  Control Valves  Figure B.5.1 Photograph of mass flowmeter downstream of vessel  Figure B.5.2 Photograph of control panel of fluidization unit  Figure B.5.3 Photograph of flow control valve downstream of vessel  Figure B.5.4 Photograph of pressure regulator valve downstream of vessel  158  Appendix B: Equipment Photographs and Engineering Drawings  B.6  Preliminary Experimental Set-up  Kistler digital Electrometers  Figure B.6.1 Photograph of preliminary experimental set-up  Collision ball probes Figure B.6.2 Photograph of preliminary experimental set-up (top view) 159  Appendix B: Equipment Photographs and Engineering Drawings  B.7  Elevated-Pressure Fluidization Column  Expanded section  Figure B.7.1 Photograph of elevated-pressure fluidization column (expanded section)  Straight section  Figure B.7.2 Photograph of elevated-pressure fluidization column (straight section) 160  Appendix B: Equipment Photographs and Engineering Drawings  Wind-box  Internal cyclone  Figure B.7.3 Photograph of elevated-pressure fluidization Column (wind-box and internal cyclone)  161  Appendix B: Equipment Photographs and Engineering Drawings  162  Appendix B: Equipment Photographs and Engineering Drawings  163  Appendix B: Equipment Photographs and Engineering Drawings  164  Appendix C: Electrostatic Probes Photographs  Appendix C: Electrostatic Probes Photographs C.1  Collision Ball Probes  C.1.1 Collision ball probe parts  (a)  Figure C.1.1 Photograph of collision ball probe parts  165  Appendix C: Electrostatic Probes Photographs (b)  Teflon tube Ceramic tube  Polyethylene tube Stainless steel ball  Brass tube  Figure C.1.1 Photograph of collision ball probe parts. (c)  Hollow nut  Washer  Figure C.1.1 Photograph of more collision ball probe parts.  166  Appendix C: Electrostatic Probes Photographs  C.1.2 Collision ball probes positions  Collision ball probes Collision ball probes  Bimetal Thermometer  Dip-leg Sight glass purging tubes  Figure C.1.2 Photograph of collision ball probe positions.  167  Appendix C: Electrostatic Probes Photographs  C.2 Faraday Cup  Faraday cup  Figure C.2.1 Photograph of Faraday cup  168  Appendix D: Particle Properties  Appendix D: Particle Properties  D.1  Calculation of Minimum Fluidization Velocity of GB-S Ar =  150(1 − ε mf ) 1.75 Re 2mf + Re mf 3 ε mf Φ p ε 3mf Φ 2p  (D.1)  where the Archimedes number (Ar) and Reynolds number at minimum fluidization (Remf) are defined as: Ar =  d 3p ρ g (ρ p − ρ g )g  (D.2)  μ2  and Re mf =  d p U mf ρ p  (D.3)  μ  When εmf and or Φp are not known, one can still estimate Umf by rewriting Eq. (D.1) as  K 1 Re p ,mf + K 2 Re p ,mf = Ar  (D.4)  where K1 =  1.75 ε 3mf φ p  K2 =  and  150(1 − ε mf ) ε 3mf φ 2p  Solving Eq. (D.4) for minimum fluidization conditions and using the values for K1 and K2 recommended by Wen and Yu (1966) for fine particles gives  [  Re p ,mf = (33.7 ) + 0.0494 Ar  ρp = 2500 (kg/m3) g = 9.8 (m/s2)  2  ]  1  2  − 33.7  (D.5)  ρg at 379 kPa = 4.51 (kg/m3) µ = 1.89 x 10-5 (kg/m s)  dp = 65 µm Solving Eq. (D.5) for glass beads of 65 µm diameter gives Umf = 5.0 x 10-3 m/s  169  Appendix D: Particle Properties  D.2  Particle Size Distribution of Glass Beads  D.2.1  Particle size distribution of GB-S  Table D.2.1 Physical properties of glass beads (GB-S) Surface-weighted mean diameter(µm) Volume-weighted mean diameter(µm) Specific surface area (m2/g)  Sample #1 61 65 0.098  Sample #2 60 65 0.099  Sample #3 60 65 0.098  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.2.1 Size distribution of glass beads (GB-S).  170  Appendix D: Particle Properties  D.2.2  Particle size distribution of GB-M  Table D.2.2 Physical properties of glass beads (GB-M) Surface-weighted mean diameter (µm) Volume-weighted mean diameter (µm) Specific surface area (m2/g)  Sample #1 346 366 0.017  Sample #2 349 369 0.017  Sample #3 343 362 0.018  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.2.2 Size distribution of glass beads (GB-M).  171  Appendix D: Particle Properties  D.2.3  Particle size distribution of GB-L  Table D.2.3 Physical properties of glass beads (GB-L) Surface-weighted mean diameter (µm) Volume-weighted mean diameter (µm) Specific surface area (m2/g)  Sample #1 545 574 0.011  Sample #2 544 573 0.011  Sample #3 547 577 0.011  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.2.3 Size distribution of glass beads (GB-L).  172  Appendix D: Particle Properties  D.2.4  Particle size distribution of GB-XL  Table D.2.4 Physical properties of glass beads (GB-XL) Surface-weighted mean diameter (µm) Volume-weighted mean diameter (µm) Specific surface area (m2/g)  Sample #1 658 691 0.0097  Sample #2 658 690 0.0098  Sample #3 663 697 0.0097  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.2.4 Size distribution of glass beads (GB-XL).  173  Appendix D: Particle Properties  D.3  Particles Bulk Density Measurements  Table D.3.1 Bulk density measurement for (GB-S) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  93.33  94.55  93.01  93.63  Wparticle (g)  79.51  80.73  79.19  79.81  Particle bulk density, ρb (g/ml)  1.590  1.615  1.584  1.60  Particle bulk density, ρb (kg/m3)  1590  1615  1584  1596  Table D.3.2 Bulk density measurement for (GB-M) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  91.74  92.45  91.41  91.87  Wparticle (g)  77.92  78.63  77.59  78.05  Particle bulk density, ρb (g/ml)  1.558  1.573  1.552  1.56  Particle bulk density, ρb (kg/m3)  1558  1573  1552  1561  Table D.3.3 Bulk density measurement for (GB-L) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  90.82  91.39  92.03  91.41  Wparticle (g)  77.00  77.57  78.21  77.59  Particle bulk density, ρb (g/ml)  1.540  1.551  1.564  1.55  Particle bulk density, ρb (kg/m3)  1540  1551  1564  1552  174  Appendix D: Particle Properties Table D.3.4 Bulk density measurement for (GB-XL) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  92.74  93.58  90.81  92.38  Wparticle (g)  78.92  79.76  76.99  78.56  Particle bulk density, ρb (g/ml)  1.578  1.595  1.540  1.57  Particle bulk density, ρb (kg/m3)  1578  1595  1540  1571  Table D.3.5 Bulk density measurement for (HDPE) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  33.03  31.28  34.86  33.06  Wparticle (g)  19.21  17.46  21.04  19.24  Particle bulk density, ρb (g/ml)  0.384  0.349  0.421  0.38  384  349  421  385  Particle bulk density, ρb (kg/m3)  Table D.3.6 Bulk density measurement for (LLDPE) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  29.97  30.29  32.34  30.87  Wparticle (g)  16.15  16.47  18.52  17.05  Particle bulk density, ρb (g/ml)  0.323  0.329  0.370  0.34  323  329  370  341  Particle bulk density, ρb (kg/m3)  175  Appendix D: Particle Properties Table D.3.7 Bulk density measurement for (PE-SLD) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  36.76  36.29  37.08  36.71  Wparticle (g)  22.94  22.47  23.26  22.89  Particle bulk density, ρb (g/ml)  0.459  0.449  0.465  0.46  459  449  465  458  Particle bulk density, ρb (kg/m3)  Table D.3.8 Bulk density measurement for (PE-X1) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  31.52  29.01  34.17  31.57  Wparticle (g)  17.70  15.19  20.35  17.75  Particle bulk density, ρb (g/ml)  0.354  0.304  0.407  0.35  354  304  407  355  Particle bulk density, ρb (kg/m3)  Table D.3.9 Bulk density measurement for (PE-X2) using a known-volume container. Wcontainer (g) Volparticle (ml)  13.82 50.00 Trial #1  Trial #2  Trial #3  Avg.  Wcontainer+particle (g)  33.03  28.28  31.86  31.06  Wparticle (g)  19.21  14.46  18.04  17.24  Particle bulk density, ρb (g/ml)  0.384  0.289  0.361  0.34  384  289  361  345  Particle bulk density, ρb (kg/m3)  176  Appendix D: Particle Properties  D.4 D.4.1  Particle Size Distribution of Polyethylene Particles Particle size distribution of HDPE  Table D.4.1 Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of HDPE particles mean sieve opening, dsi (μm)  sieve weight (g)  total weight (g)  particle weight (g)  mass fraction xi  xi / dsi  2360 1400 710 500 250 180 90 45  527.8 422.0 382.8 397.6 351.7 333.4 314.0 315.5  528.3 463.6 528.6 505.3 384.7 348.7 321.9 316.8  0.6 41.6 145.9 107.7 32.9 15.4 7.9 1.3  0.0016 0.1177 0.4131 0.3050 0.0933 0.0435 0.0222 0.0037  0.000001 0.000084 0.000582 0.000610 0.000373 0.000242 0.000247 0.000082  353.1  1.0000  0.00222  SUM Sauter mean diameter, ds = 1/Σ (xi/dsi)  450.4  150  Weight fraction, ix(g)  120  90  60  30  0 0.01  0.10  1.00  10.00  100.00  1000.00  10000.00  Particle size (μm)  Figure D.4.1 Size distribution of HDPE particles  177  Appendix D: Particle Properties  D.4.2  Particle size distribution of LLDPE  Table D.4.2 Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of LLDPE particles. mean sieve opening, dsi (μm)  sieve weight (g)  total weight (g)  Particle weight (g)  mass fraction xi  xi / dsi  2360 1400 710 500 250 180 90 45  527.8 422.0 382.8 397.6 351.7 333.4 314.0 315.5  527.8 460.1 563.9 465.1 412.7 339.5 316.9 315.8  0.0 38.1 181.2 67.5 60.9 6.1 2.9 0.3  0.0000 0.1066 0.5075 0.1891 0.1707 0.0172 0.0081 0.0007  0.00000 0.00008 0.00071 0.00038 0.00068 0.00010 0.00009 0.00002  356.9  1.0000  0.00205  SUM Sauter mean diameter, ds = 1/Σ (xi/dsi)  487.0  200  Weight fraction, ix (g)  150  100  50  0 0.01  0.10  1.00  10.00  100.00  1000.00  10000.00  Particle size (mm)  Figure D.4.2 Size distribution of LLDPE particles 178  Appendix D: Particle Properties  D.4.3  Particle size distribution of PE-SLD  Table D.4.3 Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of PE-SLD particles. mean sieve opening, dsi (μm)  sieve weight (g)  total weight (g)  particle weight (g)  mass fraction xi  xi / dsi  2360 1400 710 500 250 180 90 45  527.8 422.0 382.8 397.6 351.7 333.4 314.0 315.5  528.0 455.5 635.4 437.7 353.0 333.4 317.3 316.2  0.2 33.5 252.7 40.1 1.2 0.0 3.3 0.7  0.0006 0.1009 0.7619 0.1210 0.0037 0.0001 0.0098 0.0021  0.00000 0.00007 0.00107 0.00024 0.00001 0.00000 0.00011 0.00005  331.6  1.0000  0.00156  SUM Sauter mean diameter, ds = 1/S (xi/dsi)  641.7  300  Weight fraction, xi (g)  240  180  120  60  0 0.01  0.10  1.00  10.00  100.00  1000.00  10000.00  Particle size (μm)  Figure D.4.3 Size distribution of PE-SLD particles  179  Appendix D: Particle Properties  D.4.4  Particle size distribution of PE-X1  Table D.4.4 Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of PE-X1 particles. sieve mean sieve opening, total particle mass xi / dsi weight weight (g) weight (g) fraction xi dsi (μm) (g) 3070 551.0 552.8 1.8 0.0051 0.0000017 2360 527.8 563.3 35.5 0.1011 0.0000428 1400 422.0 474.3 52.3 0.1490 0.0001064 710 382.8 477.3 94.5 0.2691 0.0003791 500 397.6 501.7 104.1 0.2965 0.0005929 250 351.7 408.4 56.7 0.1613 0.0006454 180 333.4 337.5 4.2 0.0119 0.0000660 90 314.0 315.7 1.7 0.0048 0.0000538 45 315.5 315.9 0.4 0.0011 0.0000253 SUM 351.2 1.0000 0.00191 Sauter mean diameter, ds = 1/S (xi/dsi)  522.6  150  Weight fraction, xi (g)  120  90  60  30  0 0.01  0.10  1.00  10.00  100.00  1000.00  10000.00  Particle size (μm)  Figure D.4.4 Size distribution of PE-X1 particles  180  Appendix D: Particle Properties  D.4.5  Particle size distribution of PE-X2  Table D.4.5 Detailed calculation of Sauter mean diameter, ds = 1/S (xi/dsi) of PE-X2 particles mean sieve opening, dsi (μm)  sieve weight (g)  total weight (g)  particle weight (g)  mass fraction xi  0.9 0.8 54.2 149.9 78.4 59.6 4.6 1.7 0.4 350.5  0.0026 0.0023 0.1547 0.4276 0.2238 0.1700 0.0131 0.0049 0.0011 1.0000  3070 551.0 551.9 2360 527.8 528.6 1400 422.0 476.2 710 382.8 532.6 500 397.6 476.0 250 351.7 411.3 180 333.4 338.0 90 314.0 315.7 45 315.5 315.9 SUM Sauter mean diameter, ds = 1/S (xi/dsi)  xi / dsi  0.0000008 0.0000010 0.0001105 0.0006022 0.0004475 0.0006798 0.0000729 0.0000539 0.0000254 0.00199 501.5  160  Weight fraction, xi (g)  120  80  40  0 0.01  0.10  1.00  10.00  100.00  1000.00  10000.00  Particle size (μm)  Figure D.4.5 Size distribution of PE-X2 particles  181  Appendix D: Particle Properties  D.5 D.5.1  Particle Size Distribution of Glass Bead Binary Mixtures Size distribution of M1 particles  Table D.5.1 Physical properties of M1 particles Sample #1 Surface-weighted mean diameter (µm) Volume-weighted mean diameter (µm) Specific surface area (m2/g)  Sample #2  609 642 0.00984  Sample #3  198 525 0.0303  186 507 0.0322  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.5.1 Size distribution of M1 particles  182  Appendix D: Particle Properties  D.5.2  Size distribution of M2 particles  Table D.5.2 Physical properties of M2 particles Sample #1 Surface-weighted mean diameter (µm) Volume-weighted mean diameter (µm) Specific surface area (m2/g)  Sample #2  264 535 0.0226  Sample #3  164 490 0.0364  185 502 0.0324  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.5.2 Size distribution of M2 particles  183  Appendix D: Particle Properties  D.5.3  Size distribution of M3 particles  Table D.5.3 Physical properties of M3 particles Sample #1 Surface-weighted mean diameter (µm) Volume-weighted mean diameter (µm) Specific surface area (m2/g)  Sample #2  138 441 0.0435  Sample #3  190 520 0.0315  119 391 0.05  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.5.3 Size distribution of M3 particles  184  Appendix D: Particle Properties  D.5.4  Size distribution of M4 particles  Table D.5.4 Physical properties of M4 particles Sample #1 Surface-weighted mean diameter (µm) Volume-weighted mean diameter (µm) Specific surface area (m2/g)  Sample #2  75 195 0.0798  Sample #3  74 194 0.0807  74 187 0.0813  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.5.4 Size distribution of M4 particles  185  Appendix D: Particle Properties  D.5.5  Size distribution of M5 particles  Table D.5.5 Physical properties of M5 particles Sample #1 Surface-weighted mean (µm) Volume-weighted mean (µm) Specific surface area (m2/g)  Sample #2  71 162 0.0849  Sample #3  73 180 0.0822  67 110 0.0899  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.5.5 Size distribution of M5 particles  186  Appendix D: Particle Properties  D.5.6 Size distribution of M6 particles Table D.5.6 Physical properties of M6 particles Sample #1 Surface-weighted mean (µm) Volume-weighted mean (µm) Specific surface area (m2/g)  Sample #2  67 127 0.0889  Sample #3 61 67 0.098  69 151 0.0866  20  Volume (%)  16  12  8  4  0 0.01  0.1  1  10  100  1000  10000  Particle size (μm)  Figure D.5.6 Size distribution of M6 particles  187  Appendix E: Influence of Pressure on Bubble Behaviour  Appendix E: Influence of Pressure on Bubble Behaviour  E.1  Influence of Pressure on Bubble Behaviour  Pressure exerts a strong influence on the bubbling behaviour of gas-solid fluidized beds. A general conclusion is that pressurized beds exhibit smoother fluidization, and smaller bubbles. (e.g. Botterill and Desai, 1972; Barreto et al., 1983; Chiteste et al., 1984; Olowson and Almstedt, 1990; Li and Kuipers, 2002;). Olowson and Almstedt (1990) measured the visible bubble flow rate and the through-flow velocity of gas inside bubbles in pressurized fluidized bed using capacitance and Pitot-static pressure probes, respectively. In their experiments, a bed of 0.2 x 0.3 m cross-section was operated at pressures up to 1600 kPa and at excess gas velocities between 0.1 and 0.6 m/s. The particulate material was silica sand of 700 µm mean diameter and 2600 kg/m3 density. Their results, shown in Figure E.1, showed that despite the decreased bubble size, their rise velocity, frequency, volume fraction and visible bubble flow rate all increased with increasing absolute pressure.  188  Appendix E: Influence of Pressure on Bubble Behaviour 3.0 2.5  Uf1-Umf=0.1 m/s  (a)  Uf1-Umf=0.2 m/s  0.40  Uf1-Umf=0.2 m/s  Uf1-Umf=0.4 m/s Uf1-Umf=0.6 m/s  2.0  (b)  Uf1-Umf=0.3 m/s  Uf1-Umf=0.3 m/s  Mean pierced length [m]  Mean bubble rise velocity [m/s]  Uf1-Umf=0.1 m/s  1.5 1.0  Uf1-Umf=0.4 m/s Uf1-Umf=0.6 m/s  0.30  0.20  0.10  0.5  0.00  0.0 0  05  1  0  15  05  1  15  4.0 Uf1-Umf=0.1 m/s  Mean bubble volume fraction [%]  Mean bubble frequancy [Hz]  Uf1-Umf=0.3 m/s  3.0  Uf1-Umf=0.1 m/s  (c)  Uf1-Umf=0.2 m/s Uf1-Umf=0.4 m/s Uf1-Umf=0.6 m/s  2.0  1.0  0.0  80.0  (d)  Uf1-Umf=0.2 m/s Uf1-Umf=0.3 m/s Uf1-Umf=0.4 m/s Uf1-Umf=0.6 m/s  60.0  40.0  20.0  0.0  0  0.5  1 Pressure [MPa]  1.5  0  0.5  1  1.5  Pressure [MPa]  Figure E.1 Variation of (a) mean bubble rise velocity, (b) mean pierced length, (c) mean bubble frequency and (d) mean bubble volume fraction with pressure at the mid-level of fluidized bed of 0.2 m x 0.3 m cross-section. Particles: 700 µm sand. (adapted from Olowson and Almstedt, 1990).  189  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  In gas-solid systems, forces controlling the motion of particles may be classified into three groups: •  Particle-Fluid Interactions (PFIs). Forces through the interface between fluid and particles.  •  Particle-Particle interactions (PPIs). Forces due to the interactions between particles.  •  External Field Forces (EFFs). Forces imposed by external fields.  F.1 Particle-Fluid Interactions (PFIs) The motion of a single particle, in a uniform or nonuniform flow field without external fields may be represented by three different forces: i) Drag force, ii) Basset force, iii) Lift force. These are the forces to be considered for PFIs, and can be described as follows:  F.1.1  Drag force  When a gas flows around a submerged particle, drag forces caused by unbalanced pressure distribution as well as skin friction are exerted on the particle surface. The sum of these forces in the direction of relative motion is referred to as the total drag force, FD, (Fan and Zhu, 1998). The magnitude of the drag force on a single particle moving with a constant velocity in a uniform flow field is often expressed as:  FD =  1 C D ρ g U 2r A p 2  (F.1)  where Ap is the projected area of particle perpendicular to the flow direction, ρg density of the gas, Ur velocity of the particle relative to the gas and CD is the drag coefficient, which is a function of the particle shape and Reynolds number (Re)p:  190  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds (Re) p =  U r d pρ g μ  (F.2)  where dp is particle diameter and μ is viscosity of the gas. There are three different regimes distinguishable depending on the magnitude of the particle’s Reynolds number.  F.1.1.1  Stokes’s law regime  The Stokes’s law regime is also known as creeping flow regime. In this regime, (Re)p < 0.2, viscous forces are dominant and theoretical methods of evaluating CD meet with success. Force exerted on the sphere, Fs, can be expressed as  Fs = 3ππμ p U r  (F.3)  From equations (F.1) and (F.3) we have  C D = 24  μ 24 = d p ρ g U r (Re) p  (F.4)  When the gas is of finite extent, the gas streamlines around the particle are constricted by the walls, causing increased drag. Also since the gas is theoretically stationary at a finite distance from the particle, this distorts the flow pattern and increases drag. A simple correction of the force exerted on the sphere can be expressed by introducing the correction factor λ: Fs = 3ππμ p U r λ  (F.5)  dp ⎞ ⎛ ⎟⎟ λ = ⎜⎜1 + k c L w ⎠ ⎝  (F.5a)  where  and Lw is the distance between the center of the particle and the walls, kc = 1.004 for a twodimensional column, and kc = 2.104 for a three-dimensional column (Yang, 2003).  191  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  Other researchers (Oseen, 1910; Proudman and Pearson, 1957) have provided corrections to the Stokes approximation, as shown in equations (F.6 and F.7)  CD =  CD =  F.1.1.2  24 (Re) p  24 (Re) p  ⎡ 3(Re) p ⎤ ⎢1 + ⎥ 16 ⎦ ⎣  ⎡ 3(Re) p 9(Re) 2p ⎧ (Re) p ⎫⎤ + ln ⎨ ⎢1 + ⎬⎥ 16 160 ⎢⎣ ⎩ 2 ⎭⎥⎦  (Re)p << 1.0  (F.6)  (Re)p << 1.0  (F.7)  Intermediate regime  In the intermediate regime, 0.3 < (Re)p < 500, The drag coefficient is a function of the particle Reynolds number. An empirical relationship due to Allen (F.8) is:  CD =  F.1.1.3  18.5 0.6 (Re) p  0.2 < (Re)p < 500  (F.8)  Newton’s low regime  In the Newton’s law regime, (Re)p > 500, CD is relatively constant, and the drag force is largely due to the inertia of the gas rather than to the viscosity of the gas. CD = 0.44  500 < (Re)p < 3x105  (F.9)  An enormous amount of experimental data on CD for a single sphere at various (Re)p is depicted in Figure F.1 (Schilichting, 1979). The sharp reduction in the drag coefficient at high (Re)p (around 3 x 105) corresponds to the transition from laminar to a turbulent boundary layer over the particle. This transition is due to the change of surface pressure distribution around the particle caused by the change of the wake structure behind the particle in the turbulent regime (Fan and Zhu, 1998). The regime beyond (Re)p = 3x105 is not relevant to the particles of interest in my work. It should be pointed out that the curve in  192  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  Figure F.1 was obtained under uniform and undisturbed fluid flow conditions and it applies only to a single particle in an infinite fluid medium.  (Re)p Figure F.1 Drag Coefficient of Single Sphere as a Function of (Re)p (adopted from Schilichting, 1979)  F.1.1.4 Empirical drag coefficient expression There are numerous correlations for the drag coefficient with small numerical differences. For example, Turton and Levenspiel (1986) proposed a single correlation applicable for the complete range of (Re)p and considerably simplified the calculation of the single-particle drag coefficient. The Turton and Levenspiel correlation is:  CD =  [  24 1 + 0.173(Re) 0.657 p (Re) p  0.413 + 1 + 16.30(Re) −p1.09  ] 0.2 < (Re)p < 3x105  (F.10)  Clift et al. (1978) recommended several correlations for different ranges of particle Reynolds numbers, summarized in Table F.1.  193  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  Table F.1 Recommended empirical drag coefficient correlations for different ranges of Rep. (adopted from Clift et al., 1978)  w = log10 (Re)p For non-spherical particles, Haider and Levenspiel (1989) adapted equation (F.10) to cover the non-spherical particles and proposed the following equation:  73.69e −5.0748φ (Re) p 24 − 4.0655φ 0.0964 + 0.5565φ CD = + 1 + 8.1716e (Re) p (Re) p (Re) p + 5.378e 6.2122φ  [ (  )  ]  (F.11)  where φ is sphericity of the particle. For spherical particles, i.e. (φ =1), this expression reduces to:  CD =  0.4607(Re) p 24 + 3.3643(Re) 0.3471 + p (Re) p (Re) p + 2682.5  (F.11a)  194  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  For free fall of a spherical particle in a swarm of other particles, there are three forces acting on it. They are gravitational forces, buoyancy force (net effect of the static pressure gradient of the fluid itself on the particle), and drag. Evaluation of the buoyancy force exerted on the particle in fluidized beds has been a controversial subject. Some authors (Foscolo et al., 1983; Van der Wielen et al., 1996) assert that the buoyancy force should be calculated based on the bulk density of the bed. Others (Nguyen and Grace, 1978; Clift et al., 1987; Jean and Fan, 1992) reject this assertion and conclude that the buoyancy force should be calculated based on the fluid density only, and argue that calculation of the buoyancy based on the bulk density should be restricted only to the case when fluidized particles are much smaller than the falling spherical particle itself. Mostoufi and Chaouki (1999) concluded that there is no significant difference between bulk density correlations and fluid density correlations in dilute beds (volume fraction of the fluid > 0.8). However, when the velocity of the immersed spherical particle reaches a constant value, these forces must be in dynamic equilibrium, i.e., the force of gravity must be equal to the drag force, i.e. π 3 1 π d p (ρ p − ρ g )g = C D eff ρ g U 2r d 2p 6 2 4  (F.12)  giving the following relation for the effective drag coefficient CD eff :  C D eff =  4d p (ρ p − ρ g )g 3ρ g U 2r  (F.13)  The value obtained for the effective drag coefficient CD eff from equation (F.13) is generally greater than that of the drag coefficient, CD, discussed earlier for a freely moving particle in the absence of other particles. It is common to relate these two drag coefficients by a correction factor, f, such that C Deff = f * C D  (F.14)  Wen and Yu (1966) showed that the correction factor, f, which they called the voidage function, is a strong function of voidage of the bed. Some authors (Richardson and Zaki, 195  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  1954; Foscolo et al., 1983; Van der Wielen et al., 1996) presented several functions for relating this correction factor to the concentration of particles. Wen and Yu (1966) and Richardson and Zaki (1954) considered the following simple form for the correction factor f = ε f− m  (F.15)  where m is reported to be a constant or a function of falling particle properties. Mostoufi and Chaouki (1999) showed that the exponent m in equation (F.15) must be a function of physical properties of the falling particle and the fluidized particles. They developed a new correlation for predicting the effective drag coefficient  m = 3.02Ar  0.22 s  (Re)  − 0.33 p  ⎛ dp ⎜⎜ ⎝ ds  ⎞ ⎟⎟ ⎠  0.40  (F.16)  where ds is the diameter of the fluidized particles, dp is the diameter of the falling particle, and Ars is Archimedes number given by  Ar =  F.1.1.5  d 3p ρ g (ρ p − ρ g )g  (F.16a)  μ g2  Terminal velocity of particles, Ut  The terminal velocity of a single particle, Ut, is a significant characteristic of the particle. When a particle falls through a stagnant fluid, forces are balanced when the particle  (  acceleration is zero du  dt  )  = 0 , and the maximum terminal settling velocity of the particle is  then achieved. At its terminal velocity, ut, can be obtained from equation (F.13) by the expression: ⎛ 4d p (ρ p − ρ g )g ⎞ ⎟ Ur = Ut = ⎜ ⎜ ⎟ 3ρ C g D ⎝ ⎠  12  (F.17)  196  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  Haider and Levenspiel (1989) applied the equation of Turton and Levenspiel (1986) to present the following approximation for the terminal velocity:  ⎡ 18 2.335 − 1.744φ ⎤ ⎥ u ∗t = ⎢ + 2 ∗ 0.5 ⎥ ⎢ d ∗p d p ⎦ ⎣  ( )  −1  ( )  (F.18)  where ⎤ ⎡ ρ g2 u = u⎢ ⎥ ⎢⎣ μ (ρ p − ρ g )g ⎥⎦  13  ∗  =  (Re) p Ar1 3  (F.18a)  and ⎡ ρ g (ρ p − ρ g )g ⎤ d = dp ⎢ ⎥ μ2 ⎣ ⎦  13  ∗ p  F.1.2  = Ar1 3  (F.18b)  Basset force  Once a particle is accelerating or decelerating in the fluid, a force known as the Basset history force becomes important. It accounts for the change in the flow field around the particle over time when the motion is unsteady. In a simple model at low Reynolds number with constant acceleration, the ratio of the Basset force to Stokes drag force, RBS, was derived by Wallis (1969) and rearranged by Rudinger (1980) to give the following relation:  R BS =  18ρ g τ s πρ p t  (F.19)  where τs is the Stokes relaxation time defined as  τs =  ρ p d 2p 18μ  (F.20)  197  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  The Basset force tends to be negligible when the fluid-particle density ratio is small (e.g. for solid particles in gases), and when the acceleration time change is much longer than the Stokes relaxation time or the acceleration rate is low.  F.1.3  Lift force  In a region where a velocity gradient exists, for instance near the column wall, in a turbulent boundary layer or in a high shear region, a particle moving at a constant velocity is subjected to a lift force (force perpendicular to the relative velocity between the particle and fluid) caused either by a gas velocity gradient or due to particle rotation imposed from other sources such as collisions of the particle with a solid wall or other particles, or particle rebound from a surface.  F.1.3.1  Magnus force  The lift force, resulting from the rotation of the particles, is termed the Magnus force. The lift is caused by a pressure differential between the sides of the particle resulting from an increase in the velocity on one side of the particle and a decrease on the other side due to rotation. The Magnus force for a spinning sphere in a uniform flow at low Reynolds number is given by: FM =  π 3 d p ρΩU p 8  (F.21)  where Ω is the angular velocity of the sphere, as shown in Figure F.2, and Up is the sphere axial velocity. At high Reynolds numbers, the theoretical analysis of the Magnus force becomes very complex because of the difficulties in obtaining expressions for the pressure and velocity distribution around the surface of the sphere (Fan and Zhu, 1998). Thus, the determination of the Magnus force relies mainly on empirical results.  198  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  FD  FM  Ω  Up  Figure F.2 Forces on a rotating and moving sphere. The lift force due to the particle spin is usually negligible compared to the drag force when the particle size is small or the velocity is low.  F.1.3.2  Saffman force  The lift force due to a velocity gradient in the fluid is termed the Saffman force. Using asymptotic expansions and Fourier transforms of the velocity field, Saffman (1965) obtained an expression for the lift force exerted on a small sphere in a simple shear flow at low Reynolds numbers:  FL,Saf =  6.46 0.5 ρU r d 2p (Gν ) 4  (F.22)  where v is the kinematic viscosity and G is the mean velocity shear. The Saffman lift force is negligible at very small shear rates.  199  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  F.2  Particle-Particle Interactions (PPIS)  Interparticle interactions can occur due to a variety of forces such as van der Waals, electrostatics, capillary and collision forces. Although the effects of interparticle forces on low-velocity fluidized beds have been observed, it is not clear to what extent such forces influence the behaviour of fast-fluidized beds.  F.2.1  Van der Waals Forces  The very rapidly changing dipole of one atom produces an electric field that acts upon the polarizability of neighboring atoms. The induced dipole of the neighboring atoms tends to move in phase with the original dipole, producing an attractive atomic interaction known as the van der Waals force. The van der Waals force, Fvw, can be expressed by  FVW =  ∂E ∂a  (F.23)  where E is the energy of the interaction and a is the separation distance between two atoms. Van der Waals forces exist not only between individual atoms and molecules, but also between solids (Fan and Zhu, 1998). For the case of two spheres, equations for the van der Waals force were deduced by Hamaker (1937) assuming that the molecular forces are additive. This allowed him to calculate the force in terms of the interaction between individual atoms as  FVW =  AR 12a 2  (F.24)  where R is the sphere radius, A is the Hamaker (material-related) constant and a is the surface separation, whose minimum value is of the order of the intermolecular spacing. To circumvent Hamaker’s assumptions of additivity of molecular interactions, Lifshitz (1956) developed a theory for the interaction energy between solid bodies, using bulk material properties only. Hamaker’s constant is expressed by  200  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  A=  3 hϖ 4π  (F.25)  where hϖ is the Lifshitz’s constant. The van der Waals forces are only appreciable when particles come sufficiently close together, i.e. at a separation distance of the order of the size of a molecule, e.g. 0.2 - 1 nm. Moreover, their magnitude becomes negligible compared with that of the gravitational force when the particle size exceeds a certain value, e.g. 30 μm. This occurs because the gravitational force, Fgr, is proportional to the cube of the particle diameter, whereas the van der Waals force is proportional to the diameter itself. According to the Geldart classification, group C covers the range of particles having an average diameter smaller than ∼20 μm. Particles in this group will, by their very nature, demonstrate cohesive van der Waasls effects. Hence, the behaviour of group C particles is determined by interparticle forces. In general, fluidization of the Geldart (group C) particles is extremely difficult, not only because interparticle forces are larger than gravitational forces, but also because the forces which the fluidized gas can exert on these particles are too small for fluidization (Geldart, 1986). In fast-fluidized beds, the high gas velocities and large wall shear and core drag forces suggest that van der Waals forces are unlikely to significantly influence fast-fluidized bed behaviour. However, some aggregation involving very fine particles is possible.  F.2.2  Electrostatic forces  Other attraction or repulsion forces may also operate between the particles in gas-solid fluidized beds, such as electrostatic forces. The mechanism of static charges generation is quite complex. Electrons or ions can transfer between bodies in contact, forming an electrical double layer consisting of two layers of charges of opposite sign. If the bodies are suddenly pulled apart, the original electronic equilibrium cannot be re-established and one of the surfaces retains more electrons or ions than before the contact, while the other has a deficit. The total charge of the two surfaces remains constant. However, if one of the surfaces loses its charge (for instance, because it is a better conductor or is earthed), the global result for the particles is a net electrical charge (Cross, 1987). Fluidization is, by its very nature, associated with continuous particle contact and separation, as well as with  201  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  friction between particles against and between particles and the wall. Such circumstances favour electrostatics. In gas-solid fluidized beds, triboelectrification, frictional charging and thermionic emission in high temperature processes are known to generate electrostatic charges (Cross, 1987).  F.2.2.1  Triboelectrification  Triboelectrification involves the generation of electrical charges due to rubbing between materials. Triboelectrification, also known as contact electrification, occurs due to difference in the initial Fermi energy levels of the materials at the contact surface until the energy levels are equalized (Cross, 1987). Upon separation, the particles that lose electrons become positively charged, whereas those that gain electrons become negatively charged.  F.2.2.2  Frictional Charging  In industrial gas-solid fluidized beds, electrostatic charges arise primarily from surface charge polarization due to friction among gas, particles and reactor walls. If the reactor is large enough to neglect wall effects, particles rubbing against each other become the main cause of charge generation (Park et al., 2002). Boland and Geldart (1971) attributed the generation of electrostatic charges in gas-solid fluidized beds to the motion of particles around gas bubbles, particularly in the wake. The charge transfer generally increases as the force at contact and the speed of rubbing increase (Cross, 1987). Although the mechanism of electrostatics generation in gas-solid fluidized is not fully understood, the well-known Coulomb’s law, which quantifies the electrostatic force, was unraveled as early as two centuries ago (Fan and Zhu, 1998). According to this law, for two charged particles that are much smaller than the distance between them, the electrostatic force, Fe, between them can be expressed by  Fe =  qq ′ 4πε o d 2  (F.26)  202  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  where q and q’ are the charges carried by the particles, d is the distance between their centers and ε o is a constant called the dielectric permittivity of free space. The force between two charged particles in a swarm of other particles is modified by the permittivity of the medium in which they are situated and the Coulomb law in a media of relative permittivity, ε r , then becomes  Fe =  qq ′ 4πε o ε r d 2  (F.27)  The dielectric constant or permittivity of any material is normally expressed as a multiple of the permittivity of free space, ε o , and the relative permittivity of the material, ε r . Thus the permittivity of the material is ε o ε r .  F.2.3  Collision Forces  In gas-solid fluidized beds, collisions occur between particles or between the particle and the wall. Particle collisions result from: 1- The relative velocity between particles of different sizes or densities due to different responses to the mean gas flow. 2- The radial gradient in particle vertical velocity due to wall effects and radial gradients in gas velocity. The collision frequency mainly depends on the particle concentration and particle size. It is also influenced by the flow pattern. Kinetic energy loss of the particles due to the collisions may occur causing frictional heat generation, wall surface erosion, particle attrition and/or particle deformation (Fan and Zhu, 1998). A collision without permanent deformation or heat generation is called an elastic collision. Otherwise, the collision is inelastic and energy loss occurs, mainly in the form of permanent deformation such as particle attrition and frictional heat loss.  203  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  A typical collision between two spherical particles, A and B, of masses mpA and mpB and diameters dpA and dpB is shown in Figure F.3. U1A and U1B are the velocities just before impact, and the rebounding velocities immediately following collision are U2A and U2B. The velocity of particle A relative to particle B is UAB and the unit vector directed from the center of particle A to the center of particle B at the moment of impact is k. If the particle rotation prior to collision is negligible, and if the contact area between particles upon impact is small and the particle surfaces are smooth, then the change in velocity of particle A relative to B occurs only in the direction of k.  U1B U2A  U2B  U1A mpB mpA  Before Collision  mpB  After Collision  Figure F.3 Typical collision between two spherical particles The change in kinetic energy of each particle due to collision is a function of both the energy transferred from one particle to the other and the energy lost in the collision. For particle A the energy change is  ΔE A =  1 m pA (U 2A ⋅ U 2A − U 1A ⋅ U 1A ) 2  (F.28)  For the components of relative velocity in the k-direction, the ratio of relative velocity after collision to that before collision is the coefficient of restitution, e, i.e. k ⋅ U 2AB = −e(k ⋅ U 1AB )  (F.29) 204  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  The restitution coefficient depends not only on the material properties but also on the relative impact velocity. The maximum value, e = 1.0, corresponds to a perfectly elastic material. Typical ranges for e at ambient temperature are 0.8-1.0 for glass particles and 0.2-0.6 for metal particles (Senior, 1992; Goldsmith, 1960). In a two-particle collision, linear momentum is always conserved i.e. m pA U 1A = m pA U 2A − J  (F.30)  m pB U 1B = m pB U 2B + J  (F.31)  where J is the impulse of the forces exerted by particle A on particle B. if the particle velocities before collision, U1A and U1B, are known, equations (F.29), (F.30) and (F.31) may be combined (Senior, 1992) to give the particle velocities after collision:  U 2A = U 1A −  U 2B = U 1B −  m pB m pA − m pB  m pA m pA − m pB  (1 + e )(k ⋅ U1AB )k  (F.32)  (1 + e )(k ⋅ U1AB )k  (F.33)  Substituting equation (F.32) into equation (F.28) and rearranging gives:  ΔE A =  ⎤ m pA m pB (1 + e ) ⎡ m pB (1 + e ) (k ⋅ U1AB )2 − 2(k ⋅ U1AB )(k ⋅ U1A )⎥ ⎢ 2(m pA + m pB ) ⎢⎣ m pA + m pB ⎥⎦  (F.34)  A corresponding equation can be written for particle B. The net loss of kinetic energy in the collision, ∆E = ∆EA + ∆EB, is then  ΔE =  m pA m pB  2(m pA + m pB  ) (e  2  )  − 1 (k ⋅ U 1AB )  2  (F.35)  205  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  For elastic spheres, the maximum collision force in a collinear impact between two particles, Fc, is given by 4 ∗ ∗ ⎛ 15m ∗ U 2AB ⎞ ⎟ Fc = E a ⎜⎜ ∗ ∗ ⎟ 3 ⎝ 16E a ⎠  (F.36)  where m* is the relative mass defined by 1 1 1 = + * m pA m pB m  (F.36a)  a* is the relative radius defined by  1 1 1 = + ∗ aA aB a  (F.36b)  and E* is the contact modulus defined by 1 1 − υ 2A 1 − υ 2B = + EA EB E∗  (F.36c)  EA and EB are the Young's modulus for particles A and B, and υ A and υ B are Poisson's ratios for particles A and B.  F.3  External Field Forces (EFFs)  The external field forces are long-range forces exerted by various fields outside the fluidized beds system. Field forces in a gas-solid fluidized beds system include the gravitational force, Fgr, the electric force, FE, and the magnetic force, FMg.  206  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  When an external electric field is applied to a gas-solid fluidized system the charged particles are subjected to an electric force, Felc, expressed by Felc = qE int  (F.37)  where q is the charge carried by the particle and Eint is the electric field intensity. If particles are magnetism-sensitive, they are subjected to a magnetic force once they are exposed to an electromagnetic field given by Fmag = mμ r B 0  (F.38)  where B0 is the magnetic flux density in a vacuum, m is the number of north magnetic poles and μ r is the relative permeability of the particles.  207  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  Literature Cited in This Appendix Boland, D. and Geldart, D. Electrostatic charging in gas fluidized beds, Powder Technology, 5, pp.289-297, 1971. Clift, R., Grace, J.R. and Weber, M.E., Bubbles, Drops and Particles, New York: Academic Press, 1978. Clift, R., Seville, J.P. K., Moor, S.C. and Chavarie, C., Comments on buoyancy in fluidized beds, Chemical Engineering Science, 42, pp.191-194, 1987. Cross, J.A., Electrostatics: Principles, Problems and Applications, Adams Higler, Bristol, 1987. Fan, L.S. and Zhu, C., Principles of Gas-Solid Flows, Chapter 3, pp.87-129, 1998. Foscolo, P.U., Gibliaro, L.G. and Waldram, S.P., A unified model for particulate expansion of fluidized beds and flow in fixed porous media, Chemical Engineering Science, 38, pp.1251-1260, 1983. Geldart, D., Gas fluidization Technology, Geldart, D. (ed.), Wiley, Chichester, pp.33-38, 1986. Goldsmith W., Impact: the theory and physical behaviour of colliding solids,Arnold, London, 1960. Haider, A. and Levenspiel, O., Drag coefficient and terminal velocity of spherical and nonspherical particles, Powder Technology, 58, pp.63-70, 1989. Hamaker, H.C., The London-Van der Waals attractions between spherical particles, Physica IV, 10, pp.1058, 1937. Jean, R.H. and Fan L.S., On the model equations of Gibliaro and Foscolo with corrected buoyancy force, Powder Technology, 72, pp.201-205, 1992. Lifshitz, E.M., The theory of molecular attractive force between solids, Soviet Physics, 2, pp.73, 1956. Mostoufi, N. and Chaouki J., Prediction of effective drag coefficient in fluidized beds, Chemical Engineering Science, 54, pp.851-858, 1999. Nguyen, T.H. and Grace, J.R., Forces on objects immersed in fluidized beds, Powder Technology, 19, pp.255-264, 1978.  208  Appendix F: Forces Exerted on Particles in Gas-Solid Fluidized Beds  Oseen, C.W., Über die Stokessche Formel und über die verwandte Auphgabe in der Hydrodynamik. Arkiv for Mathimatik Astronomi och Fysik. 6, pp.29, 1910. Park, A., Bi, H.T. and Grace, J.R., Modeling charge transfer and induction in gas-solid fluidized beds. J. Electrostatics. 55, pp.135-158, 2002. Proudman, I. and Pearson, J.R.A., Expansion at small Reynolds numbers for the flow past a sphere and a circular cylinder. Journal of Fluid Mechanic, 2, pp.237-262, 1957. Richardson, J.F. and Zaki, W.N., Sedimentation and fluidization: Part I, Transaction Institution of Chemical Engineers, UK, 32, pp.35-53, 1954. Rudinger, G., Fundamentals of gas-particle flow, Amsterdam: Elsevier Scientific, 1980. Saffaman, P.G., The lift force on a small sphere in a slow shear flow. Journal of Fluid Mechanics, 22, (2), pp.385-400, 1965. Schilichting, H., Boundary Layer Theory, New York: McGraw-Hill, 7th Ed., 1979. Senior R.C., Circulating fluidized bed fluid and particle mechanics: modeling and experimental studies with application to combustion, Ph.D. Dissertation, University of British Columbia, Vancouver, 1992. Turton, R. and Levenspiel, O., A short note on the drag correlation for spheres, Powder Technology, 47, pp.83-86, 1986. Van der Wielen, L.A.M., Van dam, M.H.H. and Luyben, K.Ch.A., On the relative motion of particle in a swarm of different particles, Chemical Engineering Science, 51, pp.9951008, 1996. Wallis, G.B., One-dimensional two-phase flow. New York: McGraw-Hill, 1969. Wen, C.Y. and Yu, Y.H., Mechanics of fluidization. Chemicals Engineering Progress Symposium Series, 62, pp.100-111, 1966. Yang, W.C., Particle Characterization and Dynamics, Chapter 3 in Handbook of Fluidization and Fluid-Particle Systems, ed. W.C. Yang, Marcel Dekker, New York, pp.1-27, 2003.  209  Appendix G: Estimated Mass Ratio of Larostat in Binary Mixtures  Appendix G: Estimated Mass Ratio of Larostat in Binary Mixtures  G.1  Estimated Mass Ratio of Small-to-Large Particles in Binary Mixture for Full Mono-Layer Coverage  Consider the mass ratio of small-to-large particles in a binary mixture to form a mono-layer coverage of the small particles on the surface of one large spherical particle: Surface area of large particles = 4πrL2  (G.1)  Although the cross-sectional area of small particles = πrS2 , the small sphere would at most occupy a fraction π  4  of the total surface area, then for mono-layer coverage  N S * 4rS2 = 4πrL2  (G.2)  where rL is the radius of the large particle, rS is the radius of the fine particles and NS is the number of small particles required for mono-layer coverage  ρ S N S 4 πrS3 3 Mass ratio of small-to-large particles = 4 ρL πr 3 3 L  (G.3)  By obtaining NS from equation (G.2) and substituting it in equation (G.3) ρ S πrL2 rS3 πρ S rS * * = Mass ratio of small-to-large particles = ρ L rS2 rL3 ρ L rL  (G.4)  210  Appendix G: Estimated Mass Ratio of Larostat in Binary Mixtures When the large particles are glass beads of 574 µm diameter and small particles are Larostat particles of 13 µm diameter, then: ρ S = 520 kg/m3  rS = 13/2 µm  ρ L = 2500 kg/ m3  rL = 574/2 µm  π × 520 × 13 Mass ratio of Larostat (519) in binary mixture =  2500 × 574  2 = 1.5% 2  211  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0058972/manifest

Comment

Related Items