Gas Flow Hydrodynamics in an Oxygen Delignification Retention Tower by Isabelle Pineault B . Ing, Ecole Polytechnique de Montreal, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF CHEMICAL AND BIO-RESOURCE ENGINEERING We accept this thesis as^mfonnin^jQ^ie required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1999 © Isabelle Pineault, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of CLptn*,'**/ I*»H»P r,'„~ The University of British Columbia Vancouver, Canada Date J!ft g ^ j f /??g DE-6 (2/88) II ABSTRACT Gas flow hydrodynamics were characterized as a function of pulp suspension concentration in a laboratory-scale residence tower. Three kraft pulps were studied: never-dried brown stock (kappa 33), a fully bleached kraft pulp (88-89% brightness) that never had been dried, and a dried and reslurried fully bleached kraft pulp (88-89% brightness). The following experiments were performed to charaterize gas flow in pulp retention towers. First, gas flow in a narrow rectangular tower was characterised as a function of suspension mass concentration by calculating the bubble rising velocity. These results were compared with bubble rising velocity measurement made in a pilot-scale column having a L / D ratio o f 6.5:1 (inside diameter of27.94 cm).. The same column was used to measure gas holdup with both the gas disengagement technique [Daly et al., 1992] and measurements of the density variation method along the column height. Finally, mass transfer was evaluated in a small laboratory-scale column (L /D = 4.1:1) using a dissolved oxygen probe. The main results of these experiments are as follows. First is was found that at medium pulp consistency, there are no bubbles formed in the tower due to the lack of free water. Second, at pulp consistencies higher than 5 %, gas is trapped in the fibre network, which increases considerably the gas holdup. Furthermore, this gas holdup is not influenced by the gas flow rate. Third, the bubble size at low consistencies is influenced by the air flow rate as well as the pulp consistency present in the column. Fourth, the type o f pulp used did not have a great influence on the gas holdup in the pilot-scale column. Finally, mass transfer occurs in the laboratory-scale column, with k L a values in the same range as the those reported in bubble columns. iii T a b l e o f C o n t e n t s Abstract i i Table o f Contents i i i List o f Tables viii List o f Figures ix Acknowledgement - Remerciements xxii I N T R O D U C T I O N 1 1. Background 2 1.1 Oxygen delignification process 3 1.2 Environmental considerations 6 1.3 Performance problems 7 1.4 The survey 9 1.5 Purpose of the project • 17 2. Literature review 18 2.1 Theory of three-phase systems 18 2.1.1 Bubble columns 18 2.1.1.1 Flow regimes 19 2.1.2 Fluidized bed bubble dynamics 21 2.1.2.1 Bubble fraction and frequency 21 2.1.2.2 Bubble shape 21 2.1.2.3 Bubble motion 22 2.1.2.4 Bubble coalescence and breakup 22 2.1.3 Low-consistency pulp suspensions 23 2.1.3.1 Fibre suspension characteristics 23 2.1.3.2 Fibre network characteristics 24 iv 2.1.3.4 Bubble behaviour in low consistency pulp flow . . . 24 2.2 Gas holdup 27 2.2.1 Two-phase model 28 2.2.2 Bubble holdup in pulp suspensions 30 2.2.3 Definitions 32 2.2.4 Effect of temperature and gas density on column hydrodynamics . . 34 2.2.5 Effect of column diameter on column hydrodynamics 35 2.2.6 Effect o f superficial, rising gas velocity and liquid velocity on column hydrodynamics 36 2.2.7 Effect of distributor type on column hydrodynamics 37 2.2.8 Effect of static liquid height on column hydrodynamics 37 2.2.9 Effect of bubble size on bubble hydrodynamics 38 2.2.10 Effect of flow regime on column hydrodynamics 38 2.3 Mass transfer 40 3. Experiments methodology and calculations 43 3.1 Materials 43 3.2 2-D Experiments 43 3.2.1 2-D channel apparatus 43 3.2.2 Methodology 45 3.3 3-D Experimental apparatus: the pilot-scale column 45 3.4 Bubble size and rising velocity experiments 52 3.4.1 Principles of fibre optic probes 52 3.4.2 Description of fibre optic probes 54 3.4.3 Methodology 57 3.5 Gas holdup experiments using pressure probes 58 3.5.1 Gas disengagement technique 58 3.5.2 The density variation method . . 62 3.5.3 Methodology 64 3.6 Dissolved oxygen measurement 66 3.6.1 Dynamic method-measurement in the liquid phase o f pulp fibre suspensions 67 3.6.2 Oxygen probe 68 3.6.3 Troubleshooting 70 V 3.6.4 Apparatus 71 3.6.5 Methodology 73 3.6.6 Calculations 73 4. Results and discussion 75 4.1 2-D Experiments 75 4.1.1 Description of the bubble flows 75 4.1.2 2-D channel rising velocity results 78 4.2 3-D Experiments 79 4.2.1 Bubble size and rising velocity experiments 79 4.2.1.1 Description o f the B S T C K and bleached pulp flow 79 4.2.1.2 Size o f the bubbles 103 4.2.2 Gas holdup experiments 106 4.2.2.1 Water experiments 106 4.2.2.2 B S T C K pulp total gas holdup 106 4.2.2.3 Bleached pulp total gas holdup 112 4.2.2.4 Results using the gas disengagement technique with B S T C K pulp 115 4.2.2.5 Comparison o f experimental total gas holdup results with literature results 123 4.3 Mass transfer results 125 5. Industrial retention tower 127 C O N C L U S I O N S 129 Nomenclature 131 References • • • 135 References for bulimia 143 A N N E X E S 145 vi Annex A : Calculation examples 146 Annex A . 1: 2-D Experiments 147 Annex A . 2: Bubble size and rising velocity experiments with fibre optic probes 147 Annex A . 3: Gas disengagement technique 148 Annex A.4 : Density variation method 150 Annex A . 5: Dissolved oxygen measurement 151 Annex A.6 : Design o f the distributors 152 Annex B : 2-D Experimental results 153 Annex B l : Description of the bubble flows in the 2-d channel 154 Annex C: Fibre optic probes results 160 Annex D : Gas disengagement technique results 167 Annex D . l : B S T C K pulp results 168 Annex D . 1.1: Total gas holdup 169 Annex D . 1.2: Dilute phase and large bubble gas holdup 180 Annex D . 1.3: Dense phase gas holdup 190 Annex D . 1.4: Small bubble gas holdup 200 Annex D . 1.5: Large bubble superficial velocity 210 Annex D . 1.6: Small bubble superficial velocity 215 Annex D . 1.7: Large bubble rise velocity 220 Annex D . 1.8: Small bubble rise velocity 225 Annex D.2: Never-dried bleached pulp total gas holdup 230 Annex D . 3: Previously-dried bleached pulp gas holdup 235 Annex D.4: Comparison of the total gas holdup with the three kinds o f pulp 240 Annex E : Density variation method results 245 Annex E . 1: Total gas holdup before the experiments 246 Annex E.2: Total gas holdup during the experiments and comparison with the gas disengagement technique total gas holdup results . . . . 254 Annex E.3 : Total gas holdup after the experiments 262 Annex E.4: Difference between before and after the experiments . . . . 270 Annex F: Comparison o f the experimental total gas holdup with Viswanathan's equation 278 Annex G : Comparison o f the experimental total gas holdup with Zahradnik and VII Kastanek's equation 282 Annex H : Mass transfer results 290 Annex I: Fibre length measurements 292 Annex J: Error analysis 302 Annex J . l : 2-D Experiments 303 Annex J. 2: Bubble size and rising velocity Experiments with fibre optic probes 303 Annex J. 3: Gas disengagement technique and density variation method 303 Annex J.4: Dissolved oxygen measurement 304 VIII L i s t of T a b l e s Table 1.1: Typical operating data ranges for oxygen delignification o f softwood kraft pulp 4 Table 1.2: Oxygen delignification systems: % delignification results (data base performance change over one year) 8 Table 1.3: Oxygen delignification systems: nomber of systems (data base profile change over one year) 8 Table 1.4: Oxygen delignification systems: softwood pulps summary 8 Table 1.5: Survey data: tower characteristics as well as the pulp and gas flows 11 Table 1.6: Survey data: Chemical usage and operating conditions 13 Table 4.1: Bubble size average for B S T C K pulp at 0,1 and 2% consistency in the pilot scale column 103 Table 4.2: Average local bubble rising velocity for B S T C K pulp at 0, 1 and 2 % consistency in the pilot-scale column 103 Table 4.3: Summary of gas holdup values for B S T C K at gas superficial velocity of 0.00809 m/s at 3/4 o f the height o f the column 118 Table B . 1: Rising velocity in the 2-D channel with S G W and B S T C K pulp at Cm = 1, 2 and 4%, at air flow rate from 29.6 to 144.8 E-6 nrVs 159 Table G . 1: Constant A and B values for Zahradnik and Kastanek's equation 283 IX L i s t o f F i g u r e s Figure 1.1: Schematic o f a typical medium-consistency oxygen delignification stage process 3 Figure 1.2: Variability in delignification performance 10 Figure 1.3: Survey: effect of residence time on delignification 15 Figure 1.4: Survey: effect of pulp superficial velocity in oxygen tower on delignification 16 Figure 1.5: Survey: effect of gas superficial velocity in oxygen tower on delignification 16 Figure 2.1: Determination of the flow regime in bubble columns 20 Figure 2.2: Simplistic illustration of the flow regimes in bubble columns 20 Figure 2.3: Schematic o f bubble column flow in churn turbulent flow regime 21 Figure 2.4: Generalized two-phase model 29 Figure 2.5: A i r bubbles adhering to wood pulp fibres 31 Figure 2.6: Bubbles physically trapped in a fibre network 31 Figure 2.7: Typical dynamic gas disengagement experimental data 33 Figure 2.8: Effect o f column diameter on gas holdup in bubble column 36 Figure 3.1: Picture o f 2-D channel with dimensions of900 mm wide by 3 8 mm deep by 1222 mm high used in the 2-D experiments 44 Figure 3.2: Picture of the computer and o f the connections where the probes used were connected to the computer 47 Figure 3.3: Picture o f the pilot-scale column (empty) used to performed the 3 - D experiments 48 Figure 3.4: Picture o f the pilot-scale column (filled with water) used to performed the 3-D experiments 49 Figure 3.5: Picture of the air inlet at the bottom o f the column used to performed the 3-D experiments 50 Figure 3.6: Picture of the pilot-scale column (filled with 1 % consistency B S T C K pulp) used to performed the 3-D experiments 51 Figure 3.7: Schematic of one fibre optic probe used to performed the bubble size and rising velocity experiments 54 Figure 3.8: Picture of the two fibre optic probes used to performed the bubble size and rising X velocity experiments 55 Figure 3.9: Picture of the 300 W D C halogen lamp and the small fan used with the fibre optic probe 56 Figure 3.10: Optic probe signal from the two probes separated by 2 cm, at an air flow rate of 18.4 slpm, using B S T C K pulp; arrows indicate bubble passage 58 Figure 3.11: Gas disengagement technique analysis 60 Figure 3.12: Picture of one pressure probe number E-99999-99 (2 psig) from Cole-Parmer® 65 Figure 3.13: Concentration o f dissolved oxygen at the interface o f the dissolved oxygen probe 69 Figure 3.14: Picture of the small laboratory-scale column used for the mass transfer experiments 72 Figure 3.15: Example of 4 replicated curves used to calculate the k L a values in the small laboratory-scale column with 8% consistency B S T C K pulp and an air rate of 6.1 slpm 74 Figure 4.1: Pictures of gas spherical capped bubbles (5 x 7.5 em) rising through pulp suspensions in the 2-D channel: S G W pulp consistency o f 2% 76 Figure 4.2: Picture of channelling or fingering through B S T C K pulp at 4% consistency in the 2-D channel 77 Figure 4.3: Rising velocities for Stone Ground Wood and B S T C K pulp in the 2-D channel 78 Figure 4.4: Pictures of 1 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 81 Figure 4.5: Pictures of 1 % consistency B S T C K pulp illustrating the flow of pulp at the top o f the pilot-scale column with an air superficial velocity o f 0.01416 m/s . . 82 Figure 4.6: Pictures of 2 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity o f 0.01416 m/s 83 Figure 4.7: Pictures of 2 % consistency B S T C K pulp illustrating the flow o f pulp in the pilot-scale column with an air superficial velocity o f 0.01416 m/s 84 Figure 4.8: Pictures of 2 % consistency B S T C K pulp illustrating the flow of pulp at the top of the pilot-scale column with an air superficial velocity o f 0.01416 m/s . . 85 Figure 4.9: Pictures of 5 % consistency B S T C K pulp illustrating the flow o f pulp in the pilot-scale column with an air superficial velocity of0.01416m/s 86 Figure 4.10: Pictures of 5 % consistency B S T C K pulp illustrating the flow of pulp in the xi pilot-scale column with an air superficial velocity of 0.01416 m/s 87 Figure 4.11: Pictures o f 5 % consistency B S T C K pulp illustrating the flow o f pulp at the top o f the pilot-scale column with an air superficial velocity o f 0.01416 m/s . . 88 Figure 4.12: Pictures of 8.5 % consistency B S T C K pulp illustrating the flow o f pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 89 Figure 4.13: Pictures of 8.5 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of0.01416m/s 90 Figure 4.14: Pictures of 8.5 % consistency B S T C K pulp illustrating the flow o f pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 91 Figure 4.15: Pictures of 10 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of0.01416m/s 92 Figure 4.16: Pictures o f 10 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of0.01416m/s 93 Figure 4.17: Pictures of 10 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 94 Figure 4.18: Pictures of 10 % consistency B S T C K pulp illustrating the flow o f pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 95 Figure 4.19. Pictures of 10 % consistency B S T C K pulp illustrating the flow o f pulp in the pilot-scale column with an air superficial velocity o f 0.01416 m/s 96 Figure 4.20: Pictures of 10 % consistency B S T C K pulp illustrating the separation o f the pulp in the pilot-scale column with an air superficial velocity o f 0.00315 m/s . . 97 Figure 4.21: Pictures of 10 % consistency B S T C K pulp illustrating the separation o f the pulp in the pilot-scale column with an air superficial velocity of 0.00315 m/s . . 98 Figure 4.22: Slugging of the pilot-scale column at 8.5% consistency B S T C K pulp and a gas superficial velocity of0.00315m/s 100 Figure 4.23: Entrapment o f bubbles in the pilot-scale column at 8.5% consistency B S T C K pulp and a gas superficial velocity of 0.00809 m/s 101 Figure4.24: Entrapment ofbubbles in the pilot-scale column at 10% consistency B S T C K pulp and a gas 102 Figure 4.25: Size o f the bubbles in the pilot-scale column with B S T C K pulp and a gas flow rate o f 26 slpm 104 Figure 4.26: Average local bubble rising velocity for B S T C K pulp at 0,1 and 2 % consistency in the pilot-scale column for gas superficial velocity of 0.00809 m/s (26 slpm) 104 XII Figure 4.27: Total gas holdup in water as a function o f gas superficial velocity in the pilot-scale column 107 Figure 4.28: Comparison of the total gas holdup at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 0 ,1 ,2 and 5% consistency as a function of gas superficial velocity 108 Figure 4.29: 10% consistency B S T C K pulp total gas holdup results as a function of the pilot-scale column height and the gas superficial velocity 109 Figure 4.30: Comparison o f the total gas holdup results using the gas disengagement technique and the density variation method using B S T C K pulp at 8.5% consistency as a function o f gas superficial velocity 110 Figure 4.31: Difference between the gas holdup before and after the experiments using 8.5% consistency o f B S T C K pulp as a function o f gas superficial velocity using the density variation method I l l Figure 4.32: Total gas holdup during the experiments with 8.5% B S T C K pulp as a function of gas superficial velocity using the density variation method 112 Figure 4.33: Comparison of the total gas holdup at the top o f the column results using the gas disengagement technique with never-dried bleached pulp at 0, 1 and 2% consistency as a function o f gas superficial velocity 113 Figure 4.34: Comparison of the total gas holdup at the top of the column results using the gas disengagement technique with previously-dried bleached pulp at 0, 1 and 2% consistency as a function o f gas superficial velocity 113 Figure 4.35: Comparison o f the total gas holdup at the top o f the column using the gas disengagement technique with B S T C K , never-dried bleached and previously-dried bleached pulp at 1% consistency as a function o f gas superficial velocity 114 Figure 4.36: Comparison o f the total gas holdup at the top o f the column using the gas disengagement technique with B S T C K , never-dried bleached and previously-dried pulp at 2% consistency as a function of gas superficial velocity . . . 114 Figure 4.37: Comparison of the dilute phase and large bubbles gas holdup at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1, 2, 5, 8.5 and 10% consistency as a function o f gas superficial velocity 116 Figure 4.38: Comparison of the small bubbles gas holdup at 3/4 o f the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1, and 2% XIII consistency as a function o f gas superficial velocity 117 Figure 4.39: Comparison o f the dense phase gas holdup at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 0, 1, 2, 5, 8.5 and 10% consistency as a function of gas superficial velocity 117 Figure 4.40: Comparison o f the small bubble superficial velocity results at 3/4 o f the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function o f gas superficial velocity 120 Figure 4.41: Comparison o f the large bubble superficial velocity results at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of gas superficial velocity 121 Figure 4.42: Comparison of the small bubble rise velocity results at 3/4 o f the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of gas superficial velocity 121 Figure 4.43: Comparison o f the large bubble rise velocity results at 3/4 o f the pilot-scale columnheight using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of gas superficial velocity 122 Figure 4.44: Comparison o f the experimental total gas holdup results at 3/4 o f the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 2% consistency with Viswanathan's equation as a function of gas superficial velocity 123 Figure 4.45: Comparison of the experimental total gas holdup results at 3/4 o f the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 8.5% consistency with Zahradnik and Kastanek's equation 124 Figure 4.46: Mass transfer results: k L a coefficient found with Browns Stock pulp at consistencies of 3 and 8% in the small laboratory-scale column as a function o f gas superficial velocity 125 Figure B . 1: Shape of the bubbles in the 2-D channel 154 Figure C l : Signal from fibre optic probe 6 with 1% B S T C K pulp and a gas superficial velocity of 0.0058 m/s 165 Figure C.2: Signal from fibre optic probe 7 with 1% B S T C K pulp and a gas superficial velocity of 0.0058 m/s 166 Figure D . 1: Total gas holdup as a function of gas superficial velocity for water in the pilot scale column 170 xiv Figure D.2: Total gas holdup as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 171 Figure D.3: Total gas holdup as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 172 Figure D.4: Total gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 173 Figure D.5: Total gas holdup as a function o f gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 174 Figure D.6: Total gas holdup as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 175 Figure D.7: Comparison of the total gas holdup at the middle o f the height of the column using the gas disengagement technique with B S T C K pulp at 8.5 and 10% consistency as a function of the gas superficial velocity 176 Figure D.8: Comparison o f the total gas holdup at 3/4 o f the height o f the column using the gas disengagement technique with B S T C K pulp at 0, 1,2 and 5% consistency as a function of the gas superficial velocity 177 Figure D.9: Comparison o f the total gas holdup at 3/4 o f the height of the column using the gas disengagement technique with B S T C K pulp at 0, 1, 2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity 178 Figure D . 10: Comparison of the total gas holdup at the top o f the height of the column using the gas disengagement technique with B S T C K pulp at 0, 1, 2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity 179 Figure D . 11: Dilute phase and large bubble gas holdup as a function o f gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 181 Figure D . 12: Dilute phase and large bubble gas holdup as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 182 Figure D . 13: Dilute phase and large bubble gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 183 Figure D . 14: Dilute phase and large bubble gas holdup as a function o f gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 184 Figure D . 15: Dilute phase and large bubble gas holdup as a function o f gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 185 Figure D . 16: Comparison o f the dilute phase and large bubble gas holdup at the middle of the height o f the column using the gas disengagement technique with B S T C K pulp XV at 8.5 and 10% consistency as a function of the gas superficial velocity . . 186 Figure D . 17: Comparison of the dilute phase and large bubble gas holdup at 3/4 o f the height of the column using the gas disengagement technique with B S T C K pulp at 1, 2 and 5% consistency as a function of the gas superficial velocity 187 Figure D . 18: Comparison o f the dilute phase and large bubble gas holdup at 3/4 o f the height of the column using the gas disengagement technique with B S T C K pulp at 1,2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity . . 188 Figure D . 19: Comparison o f the dilute phase and large bubble gas holdup at the top of the height of the column using the gas disengagement technique with B S T C K pulp at 1, 2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity 189 Figure D.20: Dense phase gas holdup as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 191 Figure D.21: Dense phase gas holdup as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column . . 192 Figure D.22: Dense phase gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 193 Figure D.23: Dense phase gas holdup as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 194 Figure D . 24: Dense phase gas holdup as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 195 Figure D.25: Comparison o f the dense phase gas holdup at the middle o f the height o f the column using the gas disengagement technique with B S T C K pulp at 8.5 and 10% consistency as a function of the gas superficial velocity 196 Figure D.26: Comparison o f the dense phase gas holdup at 3/4 of the height o f the column using the gas disengagement technique with B S T C K pulp at 1, 2 and 5% consistency as a function of the gas superficial velocity 197 Figure D.27: Comparison o f the dense phase gas holdup at 3/4 of the height o f the column using the gas disengagement technique with B S T C K pulp at 1,2,5,8.5 and 10% consistency as a function of the gas superficial velocity 198 Figure D . 28: Comparison o f the dense phase gas holdup at the top of the height of the column using the gas disengagement technique with B S T C K pulp at 1,2,5,8.5 and 10% consistency as a function o f the gas superficial velocity 199 Figure D.29: Small bubble gas holdup as a function o f gas superficial velocity for a 1% xvi consistency B S T C K pulp in the pilot scale column 201 Figure D.30: Small bubble gas holdup as a function o f gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 202 Figure D.31: Small bubble gas holdup as a function o f gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 203 Figure D.32: Small bubble gas holdup as a function o f gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 204 FigureD.33: Small bubble gas holdup as a function ofgas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 205 Figure D.34: Comparison of the small bubble gas holdup at the middle o f the height o f the column using the gas disengagement technique with B S T C K pulp at 8.5 and 10% consistency as a function o f the gas superficial velocity 206 Figure D.35: Comparison of the small bubble gas holdup at 3/4 of the height of the column using the gas disengagement technique with B S T C K pulp at 1, 2 and 5% consistency as a function of the gas superficial velocity 207 Figure D.36: Comparison o f the small bubble gas holdup at 3/4 o f the height o f the column using the gas disengagement technique with B S T C K pulp at 1,2,5,8.5 and 10% consistency as a function o f the gas superficial velocity 208 Figure D . 3 7: Comparison o f the small bubble gas holdup at the top o f the height o f the column using the gas disengagement technique with B S T C K pulp at 1,2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity 209 Figure D . 3 8: Large bubble superficial velocity as a function o f gas superficial velocity for a 1 % consistency B S T C K pulp in the pilot scale column 211 Figure D.39: Large bubble superficial velocity as a function o f gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 212 Figure D.40: Comparison o f the large bubble superficial velocity at 3/4 of the height o f the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function o f the gas superficial velocity 213 Figure D.41: Comparison of the large bubble superficial velocity at the top o f the height of the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function o f the gas superficial velocity 214 Figure D.42: Small bubble superficial velocity as a function o f gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 216 Figure D.43: Small bubble superficial velocity as a function o f gas superficial velocity 2% XVII consistency B S T C K pulp in the pilot scale column 217 Figure D.44: Comparison o f the small bubble superficial velocity at 3/4 o f the height o f the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function o f the gas superficial velocity 218 Figure D.45: Comparison of the small bubble superficial velocity at the top of the height of the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of the gas superficial velocity 219 Figure D.46: Large bubble rise velocity as a function o f gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 221 Figure D .47: Large bubble rise velocity as a function o f gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 222 Figure D.48: Comparison of the large bubble rise velocity at 3/4 o f the height o f the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function o f the gas superficial 223 Figure D.49: Comparison o f the large bubble rise velocity at the top o f the height of the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function o f the gas superficial velocity 224 Figure D.50: Small bubble rise velocity as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 226 Figure D . 51: Small bubble rise velocity as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 227 Figure D.52: Comparison of the small bubble rise velocity at 3/4 o f the height o f the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function o f the gas superficial velocity 228 Figure D . 53. Comparison of the small bubble rise velocity at the top o f the height o f the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of the gas superficial velocity 229 Figure D.54: Total gas holdup as a function of gas superficial velocity for a 1% consistency never-dried bleached pulp in the pilot scale column 231 Figure D.55: Total gas holdup as a function o f gas superficial velocity 2% consistency never-dried bleached pulp in the pilot scale column 232 Figure D.56: Comparison of the total gas holdup at 3/4 o f the height of the column using the gas disengagement technique with never-dried bleached pulp at 0, 1 and 2% consistency as a function of the gas superficial velocity 233 XVIII Figure D.57: Comparison of the total gas holdup at the top o f the height of the column using the gas disengagement technique with never-dried bleached pulp at 0, 1 and 2% consistency as a function o f the gas superficial velocity 234 Figure D.58: Total gas holdup as a function of gas superficial velocity for a 1% consistency previously-dried bleached pulp in the pilot scale column 236 Figure D.59: Total gas holdup as a function of gas superficial velocity 2% consistency previously-dried bleached pulp in the pilot scale column 237 Figure D.60: Comparison o f the total gas holdup at 3/4 o f the height o f the column using the gas disengagement technique with previously-dried bleached pulp at 0, 1 and 2% consistency as a function of the gas superficial velocity 238 Figure D.61: Comparison of the total gas holdup at the top o f the height o f the column using the gas disengagement technique with previously-dried bleached pulp at 0,1 and 2% consistency as a function of the gas superficial velocity 239 Figure D.62: Comparison of the total gas holdup as a function o f gas superficial velocity for a 1% consistency B S T C K pulp, never-dried and previously-dried bleached pulp at 3/4 o f the height of the pilot scale column 241 Figure D.63: Comparison o f the total gas holdup as a function of gas superficial velocity for a 1% consistency B S T C K pulp, never-dried and previously-dried bleached pulp at the top o f the height of the pilot scale column 242 Figure D.64: Comparison o f the total gas holdup as a function o f gas superficial velocity for a 2% consistency B S T C K pulp, never-dried and previously-dried bleached pulp at 3/4 o f the height of the pilot scale column 243 Figure D.65: Comparison o f the total gas holdup as a function of gas superficial velocity for a 2% consistency B S T C K pulp, never-dried and previously-dried bleached pulp at the top o f the height of the pilot scale column 244 Figure E l : Total gas holdup before the experiments as a function o f gas superficial velocity for water in the pilot scale column 247 Figure E.2: Total gas holdup before the experiments as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 248 Figure E.4: Total gas holdup before the experiments as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 250 Figure E.5: Total gas holdup before the experiments as a function o f gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 251 Figure E.6: Total gas holdup before the experiments as a function of gas superficial velocity xix 10% consistency B S T C K pulp in the pilot scale column 252 Figure E.7: Comparison of the total gas holdup before the experiments between 3/4 and the top of the height of the column with B S T C K pulp at 0, 1, 2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity 253 Figure E.8: Total gas holdup during the experiments as a function of gas superficial velocity for water in the pilot scale column 255 Figure E.9: Total gas holdup during the experiments as a function o f gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 256 Figure E . 10: Total gas holdup during the experiments as a function o f gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 257 Figure E . 11: Total gas holdup during the experiments as a function o f gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 258 Figure E . 12: Total gas holdup during the experiments as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 259 Figure E . 13: Total gas holdup during the experiments as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 260 Figure E . 14: Comparison of the total gas holdup during the experiments between 3/4 and the top o f the height o f the column with B S T C K pulp at 0, 1, 2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity 261 Figure E . 15: Total gas holdup after the experiments as a function o f gas superficial velocity for water in the pilot scale column 263 Figure E . 16: Total gas holdup after the experiments as a function o f gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 264 Figure E . 17: Total gas holdup after the experiments as a function o f gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 265 Figure E . 18: Total gas holdup after the experiments as a function o f gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 266 Figure E . 19: Total gas holdup after the experiments as a function o f gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 267 Figure E.20: Total gas holdup after the experiments as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 268 Figure E.21: Comparison of the total gas holdup after the experiments between 3/4 and the top of the height o f the column with B S T C K pulp at 0, 1, 2, 5, 8.5 and 10% consistency as a function o f the gas superficial velocity 269 XX Figure E.22: Difference in the total gas holdup before and after the experiments (before-after) as a function o f gas superficial velocity water in the pilot scale column . . 271 Figure E.23: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 272 Figure E.24: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 273 Figure E.25: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column . 274 Figure E.26: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 275 Figure E . 27: Difference in the total gas holdup before and after the experiments (before-after) as a function o f gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 276 Figure F. 1: Comparison of the total gas holdup as a function of gas superficial velocity for water in the pilot scale column using the gas disengagement technique and with Viswanathan's equation 279 Figure F.2: Comparison of the total gas holdup as a function o f gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Viswanathan's equation 280 Figure F.3: Comparison of the total gas holdup as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Viswanathan's equation 281 Figure G . 1: Comparison of the total gas holdup as a function of gas superficial velocity with water in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 284 Figure G.2: Comparison of the total gas holdup as a function o f gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation . . 285 Figure G.3: Comparison of the total gas holdup as a function o f gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column using the gas disengagement x x i Figure G.4: Figure G . 5: Figure G.6: Figure 1.1: Figure 1.2: Figure 1.3: Figure J. 1: technique and with Zahradnik and Kastanek's equation 286 Comparison of the total gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 287 Comparison of the total gas holdup as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 288 Comparison o f the total gas holdup as a function o f gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 289 B S T C K pulp fibre length distribution 293 Never-dried bleached pulp fibre length distribution 294 Previously-dried bleached pulp fibre length distribution 295 Comparison o f the average values o f the total gas holdup calculated with the gas disengagement technique with B S T C K pulp at 8.5% consistency and the results from the experiments (un-averaged) as a function o f gas superficial velocity 304 XXII A C K N O W L E D G E M E N T - R E M E R C I E M E N T S First and for most, I would like to give my thanks to my supervisor D r Chad Bennington for is guidance and support through this project. It was a real pleasure working with you Chad. Puis, je voudrais remercier mon directeur de recherche Chad Bennington, le C R S N G et Les Consultants Gest-Eau Inc pour leur support financier sans lequel ce projet n'aurait pas eu lieu. Then I would like to thank the Pulp and Paper Centre staff: Brenda, Lisa, Georgina, Rita, Ken, Brian, Tim and Peter, the mixing group as well as the Chemical and Bio-Resource Engineering Department for their help with day to day life and my numerous problems in the lab. I should also thank Barbouillette for being so patient with me during the experiments. Je voudrais dire un gros merci a mes parents Nicole (mon assistante de recherche) et Grilles pour leur support tant financier qu'emotionnel tout au long de cette maitrise. Sans votre aide je n'aurais jamais termine ce projet. Merci aussi a mon fiance Mark. Merci pour ton soutien, tes encouragements constants et ton amour. Merc i aussi a mes amis: Frederic, Karine, Isabelle, Marie-Claude, Maggie, Mike, Arturo, Jonathan, Don et Edith. Votre amitie est tres importante pour moi. I would also like to give a special thank you to Geoff and Shabina. Without you, I would not have overcome bulimia. Y o u are the best friends anyone can dream off. Another big thank you to my friend Tazim who allowed me to stay with her for the last period of the project. Your friendship is something I cherish a lot. I have now another family in Vancouver. Also, I would like to thank my office mates: John, Steven, Ana, Judy, Heather, Paul, Brihas, Andy and Hyeon. Your humour, support and friendship was greatly appreciated. And last but not least I would like to thank Tim and Sarah (thanks for answering patiently all my questions) and Stan and Marj from Granville Chapel and the U B C Graduate Christian Fellowship group for your prayers through the whole of this project. 1 I N T R O D U C T I O N Oxygen delignification can be defined as the use of oxygen to remove from 35-50% of the lignin content of the brown stock coming into the bleaching process [McDonough, 1996]. This process is normally conducted under pressure (700-800 kPag), alkali conditions (ph 10-10.5) and high temperature (85-105°C). It's purpose is to remove 30-60% of the lignin prior to the core of the bleaching sequence which brightens the pulp. The oxygen delignification system is normally composed of a mixer, a residence tower and a washer. The main benefits of oxygen delignification are environmental. Although this process is becoming very popular among the pulp producing industries, there are several problems related to the operation of such units. In fact, achieving a high rate of delignification can be a struggle for many mills. A priori there seems to be no common trend related to performance of those oxygen delignification units. A survey done by Bennington and Pineault [1999] illustrates the importance of the mass transfer and the hydrodynamics in the delignification process. But, the role of the residence tower and it's importance in the overall delignification process has yet to be assessed. The purpose of this thesis is to characterise the gas hydrodynamics in an 02-delignification retention column. To realize this objective, a reference list of meaningful articles available on bubble columns and three-phase reactors as well as a survey of existing oxygen delignification systems was performed. To study the behaviour of medium-consistency pulp in a retention tower, experiments were conducted as a function of pulp suspension to obtain information in three areas: gas holdup, bubble size and rising velocity and mass transfer. The experiment results have been analysed and interpreted, and the results compared with industrial experience in oxygen delignification systems. 2 ! B A C K G R O U N D Oxygen delignification is one means o f enabling a mill to achieve environmental compliance in terms o f its aqueous discharges. Since the 1990's, oxygen delignification has become very popular in every new mill [Carter, et al., 1997]. 80 % o f the new oxygen delignification processes have come on-line in the last ten (10) years and about 50 % of the total number of mills are as recent as the last five (5) years [Carter, et al., 1997]. Furthermore, in order to keep up to date with environmental regulations, older mills have also implemented this process into their bleaching sequences. A few reasons why oxygen delignification units are so popular are given by Schroderus, et al. [1997] as: ^ The lower lignin content o f pulp to bleaching results in reduced consumption o f bleaching chemicals and in reduced pollution load from bleaching, roughly speaking, in direct proportion to kappa number reduction; ^ Dissolved solids from oxygen delignification are recovered in the liquor cycle which contributes to process economy, that is, the energy and chemical costs; ^ The brightness ceiling is raised, essential in T C F (Totally Chlorine Free) bleaching; The high selectivity of oxygen delignification can be used to gain higher yields in certain conditions. (For example, extended oxygen delignification makes the residual lignin more easily removed in subsequent bleaching, i.e. it requires less chemicals [Fuhrmann, et al., 1997]). In order to better understand how an oxygen delignification process works, a description o f the different types of processes available are given, followed by some environmental considerations. Finally, performance problems are described which will lead to the purpose of this project. 3 1.1 Oxygen delignification process Oxygen delignification has been used commercially since the late 1960s. This is an odourless, relatively pollution-free process used prior to the bleaching sequences [Biermann, 1996]. Lignin is removed from the pulp by oxygen under pressure, high temperature and alkali conditions, as shown in Figure 1.1. Table 1.1 shows the process conditions under which the process operates. 0 2 Reactor F i g u r e 1.1: S chemat i c o f a typ ica l medium-cons i s tency oxygen del igni f icat ion stage process 4 Table 1.1: Typical operating data ranges for oxygen delignification of softwood kraft pulp [McDonough, 1996] Medium-Consistency High-Consistency Consistency, % 10-14 25-28 Retention time, min 50-60 30 Initial temperature, °C 85-105 100-115 Inlet pressure, kPa„ 700-800 415-600 Outlet pressure, kPa e 450-550 415-600 Delignification, % 40-45 45-55 L o w press Steam cons., kg/t 40-110 30-50 Med. Press. Steam cons., kg/t 40-180 75-175 Power consumption, kwh/t 35-45 40-50 Alkali consumption, kg/t 18-28 18-23 Oxygen consumption, kg/t 20-24 15-24 As we can see, the process consists mainly of a mixer (usually a high-shear mixer) and a retention tower followed by efficient washing. Before the process, steam is used to bring the pulp to high temperature. The chemicals can be added to the process either before or into the oxygen mixer. Caustic (NaOH) and magnesium ( M g S 0 4 ) are usually introduced before the oxygen mixer. The trend in new installations is strongly toward the medium-consistency process. Among the reasons cited for this trend are the following [McDonough, 1996]: • lower capital costs, • greater ease o f stock handling with medium-consistency mixing and pumping technology, • improved selectivity in the presence of appreciable amounts ofblack liquor solids. However, here are some of the drawbacks associated with medium-consistency processes 5 [McDonough, 1996]: *• higher chemical consumption than the high-consistency processes, • lower extent of delignification. While, both process types are considered when implementing an oxygen delignification unit, as of 1993, medium-consistency systems accounted for 82 % of installed capacity [McDonough, 1996], The typical cost o f such a medium-consistency process (including all auxiliaries) is typically about 30 million Canadian dollars [Schroderus, et al., 1997]. N o matter which type o f process (medium-consistency or high-consistency) one choses, the main reasons for introducing oxygen delignification is to improve final pulp quality, especially for colour reversion, reduction o f bleaching costs and reduction of pollution load [Tatsuishi, et al., 1987]. Although, in the early 1970's, oxygen delignification units were installed to close the washing and screening system in order to reduce the chemical costs and to reduce effluent B O D (Biochemical Oxygen Demand) and colour. More recently, the emergence o f environmental regulations for dioxins and furans, A O X (Absorbable Organic halide), colour and C O D (Chemical Oxygen Demand) became important. Today, the environmental demand for E C F (Elemental Chlorine Free) and T C F has pressured industry to adopt oxygen delignification treatment [Carter, et al., 1996]. When implementing an oxygen delignification unit in an existing mill, one must be aware of the implications to the mill operation. Oxygen delignification affects liquor cycle operations, i.e. evaporation, recovery boiler and kiln and recausticizing, because o f increased recovery of dissolved solids and the use of oxidized white liquor in the oxygen delignification process [Schroderus, et al., 1997]. The increase in black liquor solids to recovery is about 4.5 to 6 % for 45 % delignification o f softwood pulp (for an incoming kappa number of around 30-32). The dissolved solids from oxygen delignification have a lower heating value than the dissolved solids from cooking due to higher content of inorganic material and the oxidation reactions occurring 6 in the oxygen delignification process [Schroderus, et al., 1997]. Heat generation in the recovery boiler increases about 3-5 %. But the effect on the total mill energy balance is small, as the oxygen delignification system itself consumes about 40 kWh/t electrical power and about 250 kg/t o f steam [Schroderus, et al., 1997]. In order to preserve pulp strength, small amounts o f magnesium ion (0.05-0.1 % on pulp as M g S 0 4 ) must be present in the pulp to protect it. Hence, oxygen, which although very efficient in lignin removal, is also less specific for lignin removal, i.e. it can attack the cellulose content of the fibre [Biermann, 1996]. The magnesium ions allow delignification to occur without a severe lost in the viscosity o f the pulp. The oxygen delignification process decreases the softwood pulp kappa number from 30-35 (20-24 for hardwoods) to 14-18 (7-10 for hardwoods) [Biermann, 1996]. Bennington and Pineault [1999] have found, from their survey done in the pulp and paper industry and from Carter et al. [1996], that most mills achieve a maximum delignification at outcoming kappa number o f 10.5 for softwoods and 7 for hardwoods. Bleaching to kappa numbers below this can lead to unacceptable losses o f the cellulose viscosity [Biermann, 1996]. 1.2 E n v i r o n m e n t a l cons iderat ions Oxygen delignification is known to be an almost pollution free process. Indeed, the filtrate from the effluent o f oxygen delignification systems can be used on the brown stock washers or otherwise ultimately sent to the recovery boiler because there are no chlorine ions present that would lead to high dead loads and corrosion in the recovery boiler [Biermann, 1996]. Furthermore, the inclusion o f an oxygen stage before the bleaching sequence decreases the effluent loading by at least 40% [Liebergott, et al., 1987]. The most obvious beneficial environmental effect o f installing an oxygen delignification unit in a chlorine compound based bleach plant, is a decreased need for oxidizing chemical in the delignifying part o f the bleaching sequence. Lower chemical cost results from this decreased 7 requirement for delignifying chemicals, since oxygen is less expensive, and oxidized white liquor usually provides the necessary alkali for the oxygen stage at low cost. Also, further savings result from a decrease in the chlorine dioxide charge needed for final bleaching stages. Finally, additional savings result from the decrease in the caustic required in the first extraction stage [McDonough, 1996]. I f the bleaching agent used is chlorine or chlorine dioxide, there is a corresponding decrease in chlorinated organic byproducts in the bleach effluent ( A O X ) [McDonough, 1996]. When the delignification stage oxidant is chlorine dioxide, the effect is smaller because chlorine dioxide generates much less A O X , but the chemical cost saving is greater [McDonough, 1996]. In either case, the oxygen stage leads to a major decrease in B O D (around 40-50 % removal, C O D (around 40-50 % removal) and colour (up to 60-85% reduction) [Tench and Harper, 1987]. 1.3 Performance problems Many mills have problems running an oxygen delignification process. Their yield fluctuates and it is hard, for many o f them, to achieve a constant rate o f delignification [Carter, et al., 1996]. There seems to be no common factor related to performance o f those systems [Carter, et al., 1996]. Tables 1.2, 1.3 and 1.4 illustrate the data Carter et al. collected for 78 different bleached Kraft pulp medium-consistency lines around the world. 8 Table 1.2: Oxygen delignification systems: % delignification results (data base performance change over one year) [Carter, et al., 1996] Total systems Inlet Kappa Outlet Kappa % Delignification (average) Softwood 1995 18 26.4 15.7 40.4 Softwood 1996 43 26.3 13.7 47.5 Hardwood 1995 12 15.9 9.9 37.7 Hardwood 199b 35 17.2 10.2 40.2 Table 1.3: Oxygen delignification systems: nomber of systems (data base profile change over one year) [Carter, et al., 1996] H C M C - 1 Stage M C - 2 Stages Softwood 1995 1 12 5 Softwood 1996 6 24 13 Hardwood 1995 0 10 2 Hardwood 1996 2 22 11 Table 1.4: Oxygen delignification systems: softwood pulps summary [Carter, et al., 1996] Average Range Inlet Kappa number 26.4 16-35 Discharge Kappa number 13.7 8-20 % Delignification 47.5 28-67 A s we can see, in 1996, many new oxygen delignification untis were reported. But there appears to be a wide variation between the lowest delignification at 28 % and the highest at 67 % that those units can achieve. In order to try to explain those variations, the oxygen delignification process has to be grouped into high-consistency, 2-stage and 1-stage medium-consistency units. 9 The performance, for softwood, is best at high consistency, followed by the 2-stage and the 1-stage units. It is possible that the better performance of the high-consistency systems is a result o f better washing prior to the oxygen stage because of the inclusion of a feed press [Carter, et al., 1996]. For hardwoods, the 2-stage systems achieve higher performance than the high consistency systems. Fibre source and inherent basic strength may play a key role in the overall balance of delignification in the fibre line, particularly in relation to the desired product [Carter, etal., 1996]. Another observation made by Carter, et al. [1996] is that regardless o f the incoming kappa number, there appears to be a maximum outcoming kappa number to the bleach plant at about 9-10 for softwoods and 7.5-9 for hardwoods. It seems that the mills are comfortable with the final strength and properties at these outcoming kappas. The mills that cook the pulp at lower kappa number also tend to be more consistent in driving the oxygen delignification harder and achieving lower kappa number to the bleach plant [Carter, et al., 1996]. In conclusion, this survey done by Carter, et al. [ 1996] looked at the chemical and temperature variables but did not address the mass transfer and the mixing issues involved in retention tower. There was then a need to investigate further this matter. 1.4 T h e survey Previous surveys did not collect data on mixing or mass transfer related parameters. The tower and the possible role o f gas hydrodynamics within it were not examined. In order to see i f mass transfer played a significant role in oxygen delignification, and to determine typical operating parameters around the retention tower, a survey was done on how pulp mills in North America run their oxygen delignification units (Bennington and Pineault, 1999). It was found that there is also a wide variation in the delignification achieved by the mills, as shown in Figure 1.2 (data shown only for softwood, medium-consistency systems) [Bennington et Pineault, 1999]. 10 B D 2 E F 2 G 1 G 1 H J J K 1 K 2 L 2 M Milk Figure 1.2: Var iab i l i ty in delignification performance [Bennington et Pineault, 1999] N o one-factor taken by itself influenced delignification efficiency. For example, while the amount of oxygen used influences the delignification achieved in the process, since oxygen is usually used in excess, an increase in the amount of oxygen will not always increase delignification [Bennington et Pineault, 1999]. Both a mass transfer parameter (gas void fraction) and the chemical reaction parameters (% 0 2 , % N a O H and temperature) are taken into account and correlated to the "theoretical delignification". The "theoretical delignification" is the delignification calculated with a floor level of lignin that is usually reached in the process: outcoming kappa number of 10.5 for softwoods and 7 for hardwoods. This allows delignification to be compared with the knowledge that i f the entering kappa number is close to the floor kappa number, the there is less lignin to remove than i f the entering kappa number is a lot higher than the floor level [Bennington et Pineault, 1999]. We then obtain, using a multiple regression analysis, a correlation with a coefficient o f r 2 = 71% [Bennington et Pineault, 1999]. Other 11 factors must account for the remaining 30 % of the data variability. A correlation o f 71% shows that both the mass transfer and the chemical reaction are involved in delignification and that they influence one another. But to what extent does the mass transfer and the hydrodynamics in the tower influences the overall delignification? Table 1.5 shows tower operation for all the mills that responded to the survey. The pulp residence time was calculated using the pulp velocity and the height o f the towers. It gives an idea o f how long the pulp actually spends in the tower since few mills actually measure the residence time o f pulp in the tower [Bennington et Pineault, 1999]. Also, Table 1.6 shows the typical operating conditions of the pulp entering the tower as well as the chemical added. From these results, the operating parameters for the pilot-scale column for laboratory experiments were determined. The relative gas superficial velocity was chosen to remain constant in order to represent in the laboratory what is happening in the industrial retention column. Furthermore, the aspect ratio of the pilot-scale laboratory column was chosen to be as close as the aspect ratio o f the towers found in the industry (average aspect ratio o f 9.2 ). Table 1.5: Survey data: tower characteristics as well as the pulp and gas flows [Bennington and Pineault, 1999] Tower Pulp flow Gas Flow Mill Cm(%) Height (m) Ratio Top Pres (kPag) (m/s) Res Time (min) Res Time Theory (min) U g (m/s) Res Time (nun) A s w l 8.5 32.9 11.4 430 0.022 N M 25 0.0031 N M Ajw-2 8.5 32.9 11.4 140 0.022 N M 25 0.0053 N M Bsw 8-10 29.3 8.7 660 0.0074 60 58 0.0017 6-15 11 27.4 6.9 340 0.0081 45 56 0.0016 N M Dsw DHW 10-14 35.7 9.0 230 0.0081 N M 74 0.0027 N M Esw 9 38.7 9.3 300 0.0073 N M 88 0.0024 N M 12 Tower Pulp flow Gas Flow Mill C f f l (%) Height (m) Aspect Ratio Top Prcs (kPag) Up (m/s) Res Time (mm) Res Time Thcorv (mm) Ug(m/s) Res Time (min) Fsw 12 35 8.3 350 0.0068 70 85 0.0021 N M F HW 12 35 8.3 350 0.0091 70 85 0.0026 N M Gsw 9-11 32.3 5.9 400 0.0057 N M 95 0.0012 N M GHW 9-11 32.3 5.9 400 0.0072 N M 75 0.0016 N M H Sw 10-12 35.7 16.7 450 0.0239 N M 25 0.0032 N M Isw 25 10.7 2.9 470 - N M - - N M Jsw 10 36 9.8 340 0.0096 60 62 0.0036 12-15 Kswi 9 22 9.6 400 0.0301 N M 34 0.0040 N M KSW2 10 27 12 400 0.0177 N M 25 0.0026 N M Lswi 10.5-11.5 25.6 8 400 0.0076 N M 56 0.0033 N M LHW2 10-11 25.6 9.3 300 0.0170 60 25 0.0032 N M Msw 10 30.5 8.5 440 0.0094 60 54 0.0023 N M Nswl 11.5 27.4 11.3 650 0.0207 15 22 0.0032 N M Nm-2 11 43 11.2 200 0.0085 60 84 0.00068 N M Here C m is the consistency of the pulp, Top Press is the pressure at the top o f the column, U p is the superficial velocity o f the pulp, Res Time is the residence time in the column, Res Time Theory is the calculated residence time o f the pulp in the column, U g is the superficial velocity of the gas, N M means that the value was not measured, S W means softwood, H W means hardwood, Mil l -1 means the first stage o f a two-stage system, M i l l , means system 1 o f two independent systems at a given mill. 13 Table 1.6: Survey data: Chemical usage and operating conditions [Bennington and Pineault, 1999] M i l l T° (after 0 : mixer) (°C) Pressure (after 0 , mixer) (kPag) pH (into tower) (into mixer) Chemical Usage (% on pulp) NaOH MgSO, other A s w l Agw-2 84 90 925 925 10.5 10.5 0.074 0.061 0.92 0.74 1.4 0 0.042 0 defoamer at decker Bsw 91-99 640 11 0.181 1.6 1.6 0.25 Talc CHW 90 630 10.8 0.144 1.3 - 0 O W L used for caustic Dsw 86 517 10.5-11 0.248 0.248 1.6 1.6 1.7 1.7 0 0 O W L used for part of the caustic used Esw 96 570 11 0.113 1.6 - 0.08 O W L used for part of the caustic used Fsw FHW 95 95 500 500 11 11 0.232 0.214 1.3 1.2 1.6 1.4 0.21 0 GHW 99 99 532 532 10-11 10-11 0.169 0.181 1.2 1.3 2 1.3 0.1-0.25 0 0,45 Talc 0,45 Talc Hsw 80 800 11 0.115 1.1 0.85 0 Isw 115 472 10-11 - 2.1 1.1 0.9 O W L used for part of the caustic used Jsw 98 720 12.1 0.186 2.5 2.7 0.2 O W L used for caustic Kswi KsW2 80 80 690 690 12 12 0.117 0.127 1.1 1.1 0.66 0.48 0.1 0.1 14 M i l l "T"0 (after 0 : Pressure (after 0 : (into tov\cr) (into mixer) Chemical Usage (% on pulp) mixer) <°C) mixer) (kPag) NaOH MgSO, other Lswi LHW2 85 88 1000 670 11.5 11 0.201 0.183 2.5 1.5 -0 0 OWL used for 15% of the caustic used 95 800 10.3 0.197 2 2 0.05 OWL used for caustic Nswl 97.5 1000 12 0.131 1.5 1.2 0.25 N r o-2 99 650 11.8 0.101 0.8 0.3 0.25 Here T° is the temperature, 0 g is the gas void fraction calculated assuming that all the steam added has condensed in the mixer, O W L means oxidised white liquor, S W means softwood, H W means hardwood, Mil l -1 means the first stage o f a two-stage system, Mi l l j means system 1 of two independent systems at a given mill. A s we can see, most mills operate their towers at medium-consistency (8.5-14%). The tower aspect ratio (L/D) is generally above 8 and the pulp residence time is usually 60 minutes. One can notice that there are quite a few mills that have a pulp residence time of about 30 minutes. However, as shown in Figure 1.3, we can also see that a longer residence time does not necessary mean better delignification. The pressure at the top o f the tower remains in the range of300-470 kPa g . The residence time of the gas, where measured, is about 15 minutes. This is much smaller than the residence time of the pulp. It can be an indication of gas channelling through the pulp in the column as well as an indication o f direct bypass o f the pulp by the gas. 15 60,0 50,0 40,0 S 30,0 Q 20,0 10,0 0,0 20 40 60 Residence Time (min) 80 ^ Softwood O Hardwood 100 Figure 1.3: Survey: effect of residence time on delignification [Bennington et Pineault, 1999] The gas and pulp superficial velocities are very small. Based on performance for water data, it should fall in the homogeneous-bubbly flow regime [Viswanathan, 1986]. The pulp superficial velocity is 5 to 10 times higher than the gas superficial velocity. But since the gas residence time is smaller than the pulp residence time, that indicates that the rising velocity of the gas is much higher than it's superficial velocity. Figuresl.4and 1.5 show the influence o f the pulp superficial velocity and gas superficial velocity on delignification. A s we can see, it is impossible to conclude that either affects delignification. It means there is more than one factor that influences the quality of delignification. 16 60,0 50,0 40,0 8 30,0 a Q 20,0 10,0 0,0 0,000 0,005 0,010 0,015 0,020 0,025 Pulp superficial velocity (m/s) • Softwood OHardwood 0,030 0,035 Figure 1.4: Survey: effect of pulp superficial velocity in oxygen tower on delignification [Bennington et Pineault, 1999] 60,0 50,0 40,0 8 30,0 1 0,0 0,0 -I o o • • So ftw o o d O H a r d w o o d 0,001 0,002 0,003 0,004 G a s superficiaI velocity (m/s) 0,005 0,006 Figure 1.5: Survey: effect of gas superficial velocity in oxygen tower on delignification [Bennington et Pineault, 1999] 17 The void fraction in the mixer also influences delignification. Although a small void fraction does not necessarily mean good delignification; i f the gas void fraction was too high in the mixer, the performance of the mixer could be affected and therefore a lower delignification than expected would be achieved [Bennington et Pineault, 1999]. 1.5 Purpose of the project The purpose o f this projet is to better understand the hydrodynamics o f the gas flow in a oxygen delignification retention tower. The results o f our survey (Bennington and Pineault, 1999) show that both chemical reaction (% 0 2 , % N a O H and temperature) and hydrodynamics (gas Void fraction) o f the flow affect the efficiency o f oxygen delignification systems. Mass transfer can then be a key player in optimizing an oxygen delignification process. Indeed, the delignification parameters and processes occurring in the tower are not well understood. Is the tower design helping mass transfer? Is there any mass transfer occurring in the tower? The contribution of this tower to the overall delignification of the process needs to be assessed. Furthermore, the hydrodynamics o f medium-consistency pulp and gas flow have never been studied. For example, how does the oxygen move through the column? H o w does the hydrodynamics of the flow influence the mass transfer of oxygen to the fibre suspension? Answers to these questions may allow the pulp and paper industry to operate their units to overcome the performance problems they encounter when running an oxygen delignification process. 18 2, L I T E R A T U R E R E V I E W Gas-liquid reactions can occur in a number of devices. In order to better understand gas-liquid hydrodynamics in a pulp retention tower, we investigated the standard chemical engineering gas-liquid reactors of the bubble column and fluidized bed reactors. Three aspects of these processes were investigated: bubble flow behaviour, gas holdup and mass transfer. 2.1 Theory of three-phase systems In order to better understand what is happening in an oxygen delignification retention tower, the gas flow has to be characterized. Therefore, bubble columns, fluidized beds and low pulp suspension behaviour have been studied. 2.1.1 B u b b l e co lumns Bubble columns basically consist o f a column wherein liquids and gases are contacted. Both liquid and gas flow can be co-current or countercurrent. The gas usually takes the form o f bubbles. One very simple representation of a bubble column is that it is set up as a core (interior of the column) region where the bubble rise velocity is high and an annular region near the wall where bubble rise velocity is a little smaller [Hyndman and Guy, 1995]. Bubble columns can be operated in batch or continuous mode and there can be one or many stages. Bubble columns are mainly gas-liquid reactors but they can also be utilised simply as an agitation device [Treybal, 1980]. In industry, bubble columns can be used as absorbers, strippers, gas-liquid reactors, etc. [Viswanathan, 1986]. The hydrodynamics are very complex and have been studied extensively. The introduction o f gas creates circulation patterns in the liquid, and the liquid movement by itself influences the gas flow [Godbole, 1986]. The flow regions of 2-phase (gas-liquid) bubble columns can be divided into three axial regions. The first region is located near the sparger where sparger design wil l influence the 19 hydrodynamics of the liquid and the bubbles. In the second region, the bubble behaviour in controlled by the liquid flow. Bubble coalescence occurs in the third region near the top of the column [Godbole, 1986]. 2.1.1.1 Flow regimes In order to understand the hydrodynamics o f the bubble column, the three main flow regimes under which columns typically operate are defined. As illustrated in Figures 2.1 and 2.2, there are three different flow regimes for a bubble column [Viswanathan, 1986]: *• Homogeneous bubbly flow regime, which is generally characterized by small uniformly sized bubbles with a rise velocity in the range of 0.18 to 0.3 m/s. *• Heterogeneous or churn-turbulent regime, which is characterized by the simultaneous presence of large and small bubbles. Figure 2.3 shows a schematic of the bubble column in churn turbulent flow regime. We can see that the large bubbles rise up in a plug flow and the small bubbles are backmixed. The large bubbles can be o f sizes from 0.08 to 0.15 m with rise velocities greater than about 0.8 m/s. About 75 % o f the entering gas passes through the column as large bubbles. • Slug flow regime, where large bubbles (called slugs) occupy the entire column cross section. This flow regime usually occurs in small diameter laboratory bubble columns. 20 1 0,14 0,12 •{ 0,1 0.08 0,06 3 0,04 & 0,02 -I 0 0,01 Slug Flow Heterogene D-US Churn-Turblulent Flow Homogeneous Bubbly Flow I i i i — i — i 1 1 1 1—I—r 0,1 Column diameter (m) Figure 2.1: Determination of the flow regime in bubble columns [Viswanathan, 1986] B u b b l y f l o w C h u r n t u r b u l e n t S l u g f l o w Figure 2.2: Simplistic illustration of the flow regimes in bubble columns [Godbole, 1986] 21 Large Bubbles Plug Flow Liquid Small Bubbles Back-mixing Bubble cluster formation Gas Figure 2.3: Schematic of bubble column flow in churn turbulent flow regime [Godbole, 1986] 2.1.2 Fluidized bed bubble dynamics 2.1.2.1 Bubble fraction and frequency In a turbulent fluidized bed, experiments show that close to the column wall, bubbles are less numerous. The further from the wall, the greater the bubble activity (presence and movement) except at the centre were the activity drops considerably [Farag et al., 1997]. 2.1.2.2 Bubble shape The interaction between a rising bubble and the surrounding liquid-solid medium results in the observed bubble shape and determines the turbulence in the flow field. There are several types o f shapes which a bubble can have, such as spherical, oblate ellipsoidal and spherical-cap 22 [Fan, 1989]. Usually, for small bubbles, bubble shape is spherical because surface tension forces dominate. For intermediate sized bubbles, surface tension forces, the inertia of the liquid flow surrounding the bubbles, liquid viscosity and the presence of surface-active contaminants make the bubble shape more complex: ellipsoidal bubbles. For large bubbles, their shape is like a spherical-cap because the inertial effects of the surrounding flow predominate [Fan, 1989]. 2.1.2.3 Bubble motion The differences in the motion of these bubbles are dependent upon the differences in the flow around them. For a small bubble, the flow around the bubble is called creeping flow for which Stoke's law is appropriate and the wake size is negligible. Therefore the bubble rises steadily in a rectilinear path. For intermediate and large bubbles, the flows around the bubbles can be described as a periodic shedding o f vortices from the bubble wake. So the motion of a intermediate bubble exhibits secondary motion characteristics such as zig-zag or spiral trajectories as they rise. As for large bubbles, they follow a rectilinear path with some rocking and/or base oscillations [Fan, 1989]. 2.1.2.4 Bubble coalescence and breakup A s bubbles rise, coalescence or breakup may occur due to interactions between the rising bubble and other bubbles, individual particles, and the surrounding liquid-solid medium [Fan, 1989]. When two successive bubbles rise in a liquid-solid suspension, the later bubble accelerates and eventually overtakes the preceding bubble. Fan et al. [1989] attribute this behaviour to the fact that there is a suction caused by a low pressure region below the preceding bubble [Fan, 1989]. Bubble coalescence usually occurs in a fluidized bed with small particles where the viscosity and the density o f the medium (the liquid-solid mixture) can be approximated by the viscosity and the density o f the liquid alone [Fan, 1989]. Large bubbles may be deformed and ruptured to form smaller bubbles by turbulent eddies 23 generated by the liquid-solid medium. Bubble breakup can also be associated with horizontal entrainment, flattering, necking and disintegration of the layer bubbles affected by the turbulent shear and vertical flow near the edges of the preceding bubble [Fan, 1989]. Interactions between individual particles and bubbles can also lead to bubble breakup. 2.1.3 Low-consistency pulp suspensions 2.1.3.1 Fibre suspension characteristics The flow o f fibre suspension is a complex, non-Newtonian phenomenon that varies with many factors. Stenuf and Unbehend [1986] listed these factors as: • temperature, • properties o f the matrix (density and viscosity), • volumetric concentration o f the fibres and other suspended materials, • fibre dimensions: length, width and thickness, • fibre density, • fibre flexibility and nonuniformity along the length o f the fibre, • size distribution of fibres, • fibre entanglement and interfibre friction • interfibre attraction or repulsion, surface tension and bonding, • flocculation, • shear forces, • turbulence, • wall friction, • presence o f interaction with other solid particles or fines, • effects o f gas inclusions, • presence o f flocculating agents, deflocculants, electrolytes, etc. A l l these parameters influence the flow of fibre suspensions and gas hydrodynamics. For 24 example, the fibre length can dictate the formation, size and shape of bubbles as they rise in a pulp suspension in a retention tower. Therefore, the type of pulp used in oxygen delignification retention towers needs to be quantified as much as possible. 2.1.3.2 Fibre network characteristics Pulp suspensions are continuous fibre networks which possess structure and strength resulting from interaction between the fibres. With pulp consistency higher than 0.5%, cohesive strength occurs from mechanical forces by the bending and the hooking of fibres. A s the consistency increases, the number of fibre-to-fibre interactions increases which in turn increases the network strength [Bennington, 1996]. Because the strength depends on the number o f fibre-to-fibre interactions, the presence of floes influences the strength o f the network. Floes are regions where the fibre concentration is higher, giving higher strength in that region. Floes may behave as independent entities in a flowing suspension [Bennington, 1996]. 2.1.3.4 Bubble behaviour in low consistency pulp flow The presence of fibres in a reactor wi l l change the gas flow patterns. In an oxygen delignification retention tower, the oxygen needs to be well mixed with the pulp in order for reaction to occur. Lindsay et al. [1995] studied low consistency pulp suspensions in the flotation process in a deinking plant. They observed the following pulp behaviour. A t pulp consistencies lower than 1%, pulp fibres tended to flocculate. The fibre network extent and strength increased with an increase in pulp consistency. In the fibre network, floes were present, creating local areas of higher fibre concentration in the network. So, when air is bubbled into a column filled with low consistency pulp, the bubbles must make their way up around the floes. The floes can then trap the bubbles preventing them from rising through the pulp suspension. In stagnant pulp, bubbles usually rise up through channels where there is lower hydraulic resistance. Furthermore, the 25 presence of fibre enhances coalescence of bubbles. This can cause development of a churn turbulent flow under conditions that would be bubbly flow in water. In this bubbling flow regime, the gas holdup was found to be proportional to the air flow rate, but in the churn turbulent flow regime, radial and vertical dispersion was present. In this churn turbulent flow, very large, fast-moving bubbles appeared in the centre o f the column. They were formed due to a large rate o f coalescence of the bubbles. It was also found that gas holdup increased with the column height. They found that it "may be partially due to the additional entrainment o f gas that occurs via sloshing and churning o f the liquid near the top o f the column". Hence, to compare the results for similar flow conditions in water, the gas holdup in pulp suspensions was found to have lower values. The presence o f fibre in a bubble column appears to enhance a flow regime characterised by a strong bubble coalescence even at low air flow rates (that would give a bubbling flow in pure water). Lindsay et al. [1995] explain the decrease in gas holdup in this manner:"The decreased gas holdup relative to pure water can be explained by increased bubble coalescence caused by a fibrous network. Small bubbles, which normally have long dwell times, are impeded by the fibre network. A bubble held in place by a flow and network structure is soon joined by other rising bubbles, until coalescence yields a bubble having enough buoyant force to break through the restriction. A t this point, the bubble begins to rise rapidly and can have a much smaller dwell time than the original small bubbles would have had in pure water. This results in a net decrease in gas holdup". Furthermore, i f the pulp consistency or the gas flow rates are high enough, the bubbles can bypass the fibre network through channels and thus have a lower residence time and a lower gas holdup value. When the pulp consistency was increased to 2%, channels were more numerous in the column than with the 1% pulp consistency. Again, bubble coalescence was enhanced as was the formation o f large bubbles. The bubble entrapment increases with increasing pulp consistency and inlet air flow rate. The greater the fibre network strenght, the bigger the bubble must be in order to rise up through the pulp suspension. With co-current flow of pulp and air, at 1% pulp consistency, the gas holdup was found 26 to be higher than for water alone. The trapped bubbles can be carried by the pulp up the column before the bubbles have a chance to coalesce. Furthermore, the flow regimes resemble the pure water flow regimes at the same air flow rate. Looking at the gas holdup in the radial direction, Lindsay et al. [1995] found that there was a significant gradient. In fact, the gas holdup along the wall of the column was lower than the gas holdup in the centre o f the column where the value is the highest. That shows that most of the bubbles escape in the centre of the column, not along the wall. Heindel and Monefeldt [1997] have studied the bubble dynamics in a pulp suspension for 0.5, 1 and 1.5 % consistency for flotation deinking and ozone bleaching. They found that increasing the gas flow (from0.5 to 15 slpm) or the pulp consistency changes the flow regime from bubbly to churn-turbulent. Furthermore, increasing the consistency causes the bubbles to be larger which leads to more channelling in the column. But, as the consistency increases, the bubble flow becomes more heterogeneous (like in the churn-turbulent regime) [Heindel and Monefeldt, 1997]. When Heindel and Monefeldt [1997] looked at each set of experiments separately, the following trends were identified. The column they used was 2 cm deep, 20 cm wide and 1 m tall. P U R E W A T E R Large and small bubbles were present in the column. A n increment in the air flow rate increased the size o f the bubbles. This could be explained by an augmentation o f the coalescence o f the small bubbles. Increasing the air flow rate also changed the flow regime from a bubbly flow to a churn-turbulent flow. Furthermore, at high air flow (15 slpm), backmixing occured where larger bubbles rose in the centre o f the column and small bubbles were trapped in the down currents and then travelled downward close to the walls. 27 0.5 % C O N S I S T E N C Y At this consistency, the bubble shape was more spherical than in pure water. Furthermore, the larger bubbles were spread over the entire diameter o f the column but the largest bubbles still rose at the centre of the column creating some backmixing. The bubbles were larger than in pure water and were less numerous. 1 % C O N S I S T E N C Y The bubble flow patterns began to change. Channelling became more predominant. But, the channels formed were not static due to the shifting of the fibre network caused by interaction with intermittent rising bubbles. Additionally, there was an increase in the bubble size and there was more diversity in the bubble size (wider range of bubble size). The small bubbles remained close to the column's bottom due to backmixing. 1.5% C O N S I S T E N C Y Most of the bubbles were large and only a few small bubbles were present in the column. The turbulent flow resulted in uniquely-shaped large bubbles i.e. the bubbles were no longer round, they had a spherical cap shaped or became oval. The large bubbles that formed near the distributor coalesced to form even larger, less numerous bubbles, at higher column locations. 2.2 G a s ho ldup A s we know, the performance o f gas-liquid contractors is greatly influenced by backmixing, gas holdup and interfacial area, which are in turn interrelated with the liquid circulation [Viswanathan, 1986], In fact, gas holdup is important in the determination of gas residence time and the pressure drop across the system. Since heat, mass and momentum transfer depend on all these parameters, gas holdup and it's distribution constitute an important design 28 parameter for any multi phase flow system [Kumar, et al., 1997].This chapter wil l address the gas holdup in the bubble columns and in three-phase (pulp, water and air) reactors. To analyse what is happening in these towers, a two-phase model is used. 2.2.1 Two-phase model Several models can be used to describe a three-phase reactor. For example, Fan [1989] describes a three-phase model: the gas bubble phase, the wake phase and the liquid-solid emulsion phase. This mechanistic model accounts for axial mixing for low density solids in a slurry bubble column (for batch conditions). To characterise bubble-column slurry reactors, the two-phase model is used [Krishna, et al., 1997]. This model can also be used in the characterization o f pulp suspensions in residence towers [Krishna, et al., 1997]. Figure 2.4 illustrate the two-phase model. The two phases are the dilute phase, consisting of the fast-rising large bubbles which traverse the column virtually in plug flow, and the dense phase, which is identified with the liquid phase along with the solid particles and the entrained small bubbles. The dense phase suffers a considerable degree o f backmixing [De Swart and Krishna, 1995] and axial mixing (more than in a bubble column without solids) [Fan, 1989]. In a pulp suspension, which can be represented by this two-phase model, there is also backmixing due to small bubbles flowing downward near the walls. But, as the consistency o f the pulp increases, the backmixing becomes less important. 29 Dilute phase Dense phase Small bubbles Liquid • • phase + suspended solids • • * * • Figure 2.4: Generalized two-phase model [De Swart and Krishna, 1995] With this two-phase model, we can define the following gas holdups. First, the total gas holdup (e) wi l l be define as the volume fraction o f gas present in the pulp suspension. Second, the dilute phase gas holdup (e b) wil l be the gas holdup that is created in the liquid phase alone by the fast-rising large bubbles which traverse the column virtually in plug flow. In medium-consistency pulp suspensions, this phase is hard to visualise since most of the water is absorbed into the fibres. We can then say that the dilute phase gas holdup in a pulp suspension is the gas holdup created by the large bubbles, since it is easier to visualise. It is important to remember that the dilute phase gas holdup is the same as the large bubble gas holdup i.e. the fraction o f gas that exits the reactor as large bubbles. Third, is the dense phase pas holdup (e^) which is created in the liquid phase along with the solid particles and the entrained small bubbles. And fourth, is the small bubbles gas holdup (e8), which is the gas holdup created by the small bubbles alone without the interactions with the solids. Data from De Swart and Krishna's [1995] experiments show that increasing the solids concentration results in a pronounced decrease in the total gas holdup, but the influence on the 30 dilute phase gas holdup is negligeable. The dense phase gas holdup suffers a significant decrease in gas holdup due to enhanced coalescence of the small bubbles resulting from introduction o f a solid. The holdup o f gas in this dense phase is found to be practically independent of the superficial gas velocity in the heterogeneous flow regime. They also found that the small bubbles are strongly dependent on the physical properties of the system and that the dilute phase holdup is a function o f the column diameter between 0.1 to 0.38 m; the dilute phase gas holdup decreases with increasing column diameter [Krishna, et al., 1997]. 2.2.2 Bubble holdup in pulp suspensions Pelton and Piette [1992] found the following trends for pulp suspension. For a given consistency, larger bubbles rise through the column whereas smaller bubbles are held up in the fibre network. The higher the consistency, the greater the bubble diameter required to escape. They also found that there are two criteria for bubble holdup [Pelton and Piette, 1992]: I. sufficient points o f contact on the fibres to prevent channelling, II. the network must be strong enough to withstand the buoyant force exerted by the bubbles. Bubble holdup can occurs in two ways. First, bubbles can adhere to fibres giving a finite contact angle, as shown in Figure 2.5. For this situation to occur, the fibres must have hydrophobic regions on the surfaces. This phenomenon does not usually occur with fibres in a pulp or paper process [Pelton and Piette, 1992]. The second mechanism for bubble holdup is the mechanical entrapment o f bubbles in the fibre network, as illustrated in Figure 2.6. Microscopic observations made by Pelton and Piette [1992] show no instance of bubble adhesion with wood pulp. Therefore, mechanical entrapment o f fibres is the most probable mechanism for bubble holdup in a pulp suspension. Air bubbles adhering to wood pulp fibers Figure 2.6: Bubbles physically trapped in a fibre network [Pelton and Piette, 1992] 32 The presence of fibres in a reactor will also change the gas flow patterns. In an oxygen delignification retention tower, the bubbles need to be well mixed with the pulp in order for the reaction to occur. 2.2.3 Definitions In order to use experimental data, the following definitions must be understood and clarified. The total gas holdup, previously described and defined by e, permits us to obtain the dilute phase gas holdup e b and the dense phase gas holdup from [Krishna, et al., 1997]: (1 - £ V ) + Sat [2-1] The total gas holdup e can be split into two parts [Krishna, et al., 1997]: I. A scale-independent, system and slurry-concentration-dependent dense phase gas holdup e^, II. A scale-dependent, system and slurry-concentration independent dilute phase or large bubble gas holdup e b The gas-disengagement technique was introduced to measure the gas holdup in a column. A more thorough description is presented in the next chapter. Figure 2.7 illustrates the typical results found with this technique. The gas holdup can be calculated from the following definitions from Krishna and Ellenberger [1996]. The different heights present in the definitions are shown in Figure 2.7. The gas disengagement technique mainly consists of bubbling a gas into a liquid-solid suspension. After a steady flow is reached, the air is cut off and the gas escapes from the column. In order to use the gas disengagement technique, all the gas must leave the column during the experiments. Furthermore, the gas disengagement technique does not account for the entrapment o f bubbles. The larger bubbles must escape the column in a plug flow and all the small bubbles must escape the column at the end o f the experiments. 33 Gas shut off at time 0 1 1 1 r - 1 0 - 5 0 5 10 15 20 T i m e (s) Figure 2.7: Typical dynamic gas disengagement experimental data [Krishna and Ellenberger, 1996] Where: Total gas holdup. £ = H ~ H o [2-2] H H — Hi Large bubble and dilute phase gas holdup: Sb = [2-3] H Small bubble gas holdup. £s = £ - £b = — — — [2-4] H ^ , Hi-Ho Dense phase gas holdup. £<# = = [2-5] 1 - Sb Hi Where H is the height o f the pulp with air bubbling through it, FL, is the height of the pulp without air bubbling through it(at the end o f the experiments and no air remaining trapped wothon the suspension) and H , is the height o f the pulp after the escape o f the big bubbles. 34 2.2.4 Effect of temperature and gas density on column hydrodynamics Reilly et al. [1994] have reported that gas holdup is higher in cold flow, which, in return, influences the flow regime that occurs in the column. Also, the gas holdup was found to increase with increasing gas density. The increase in gas holdup was attributed to the formation of smaller bubbles at the higher gas densities [Reilly, et al., 1994]. Letzel et al. [1998] have also studied the effect of gas density on gas holdup in bubble columns. They found that the relative velocity between the bubble phase and the liquid phase influences the growth factors of the bubbles. In fact, on a large bubble, disturbances with "unstable wavelengths" can occur. The bubble surface "moves" under the influence of the external and internal forces. Therefore large bubbles are not stable and they tend to break. But they still are present in bubble columns because of the coalescence of small bubbles together. The two mechanisms are present in the column and an equilibrium is found, i.e. an equilibrium bubble size. They also found that as the gas density increases, the range of "unstable wavelengths" and growth factors increases. The bubbles then tend to break up more easily. Then, the rate of breakups increases as well. But since the pressure has little influence on bubble coalescence, the equilibrium will be shifted towards a smaller bubble equilibrium. Furthermore, as the gas density increases, the velocity of the bubbles will decrease in order to improve the stability of the "wavelengths" in the bubbles. Inversely, if the gas density decreases, the bubble size will increase along with the gas velocity. We can then say: pg\ • U2b\ = pg2 • U2bi and Ub x -4= [2-6] Where U b is the average rising velocity of large bubble, p is the gas density and 1 and 2 are the two different experimental conditions. The velocity is inversely proportional to the square root of the gas density. 35 Furthermore, from these results, Letzel et al. [1998] also found that: Ub, atm Ub and Ub « [2-7] Where V g is the superficial gas velocity, U b is the average rising large bubble gas velocity, p is the gas density and e is the total gas holdup For a turbulent fluidized bed, an increase in the temperature resulted in a shift in the flow regime. In fact, the regime became more homogeneous with smaller bubbles and more gas flowing through the dense-bed emulsion phase [Farag, et al., 1997]. Pressure also had an effect on the gas holdup since it influenced the size of the bubbles [Luo et al., 1997]. According to Luo et al., [ 1997], an increase in pressure had the following effects on a fluidized bed: it slowed down the coalescence of bubbles, increased the liquid viscosity (this seems a questionable result), increased the gas holdup and decreased the surface tension. In fact, an increase in pressure decreased the bubble size which increased the effective fluidizing liquid velocity in the liquid-solid region, hence resulting in an increased bed-expansion. 2.2.5 Effect of column diameter on column hydrodynamics According to Kumar et al.[1997], the column dimensions have no influence on the void fraction i f the diameter is greater than 0.15m (for a bubble column). But as we can see from their data in Figure 2.8 and as mentioned before, the bigger the diameter (after 0.15 m), the higher the void fraction (at U g = 0.08 m/s). Since no error intervals were given; no precise conclusions can be drawn. For column diameters smaller than 0.15 m, Kumar et al. (1997) found that the gas holdup profiles tended to be flatter than for large column diameters. 36 0,25 0,2 c '•§ 0,15 CO "O o > 0,1 0,05 0 • Ug = 0.02 m/s §§ Ug = 0.08 m/s 0 0,1 0,2 0,3 0,4 Column diameter (m) Figure 2.8: Effect of column diameter on gas holdup in bubble column (data taken from Kumar et al. [1997]) 2.2.6 Effect of superficial, rising gas velocity and liquid velocity on column hydrodynamics A s we can see in Figure 2.8, the velocity of the inlet gas has a distinct influence on the void fraction: an increase in the gas velocity causes more gas retention in the bubble column. Furthermore, the higher the velocity, the more parabolic shaped where the gas holdup profiles [Kumar, et al., 1997]. Fan [1989] describes the effect of liquid velocity for slurry bubble column systems. " A t moderately high superficial liquid velocities, the liquid velocity does affect gas holdup; an increase in liquid velocity may increase the absolute bubble rise velocity, decrease the initial bubble size formed at the distributor, or change the coalescence/breakage tendencies." Bubble rising velocity was found to decrease with an increase in the solids concentration and density in 37 a fluidized bed reactor [Bly and Worden, 1992]. Furthermore, the bubble rising velocity increases with bubble size [Bly and Worden, 1992]. 2.2.7 Effect of distributor type on column hydrodynamics The gas holdup for bubble caps, cones and perforated plates were studied by Kumar, et al. [1997]. They found that the perforated plate provides a uniform distribution of the gas. A s for the bubble cap and the cone distributers, the gas moves in large bubbles in the centre o f the column so the regions near the walls receive a lower portion of the gas than the centre. Also, the magnitude o f the holdup generated by a sintered plate is larger than that o f the perforated plate, and this difference increases with an increase in gas velocity [Kumar, et al., 1997]. Furthermore, for a bubble column and a perforated plate distributor, Zahradnik and Kastanek [1979] found that the liquid height had no effect on the tray pressure drop o f the distributor and the critical gas hole velocity for their data. The liquid aspect ratio o f their column was 7. They also found no effect of free plate area (hole concentration) on the value o f the critical gas hole velocity. Hence a gas velocity of % the critical gas hole velocity insured a uniform distribution of the gas in the column. Finally, under a bubbling flow regime, no effect o f the free plate area on the gas holdup values was found [Zahradnik and Kastanek, 1979]. 2.2.8 Effect of static liquid height on column hydrodynamics Depending on the bubble coalescence and breakup characteristics of the system (in a bubble column), the initial liquid height may or may not have a bearing on gas holdup. But, it is now generally accepted that for bubble columns with L / D ratios greater than five (5), the gas holdup is independant of this ratio [Kumar, et al., 1997]. 2.2.9 Effect of bubble size on bubble hydrodynamics 38 Fan [1989] mentions that the initial bubble size, in a bubble column, is affected by: • the hole diameter of the distributor, • the gas velocity through the distributor's holes, • the physical properties of the systems like the surface tension, liquid viscosity and gas and liquid densities. In a bubble column, the bubble size was found to vary as the square root o f the superficial gas velocity [Burns and Rice, 1997]. Furthermore, bubble size was also found to control (along with the wall effects) or influence eddies size [Burns and Rice, 1997]. But in a slurry column, the mean bubble diameter is only slightly affected by the gas velocity. In fact, an small increase in the diameter of the bubbles would be seen with an increase o f the gas velocity [Fan, 1989]. Furthermore, Khare and Niranjan [ 1995] found that the tiny bubbles are responsible for 70-80 % of the total gas holdup in a impeller agitated aerobic reactor. Agitation can influence the coalescence and breakup patterns in bubble and slurry columns and therefore influence the bubble size. 2.2.10 Effect of flow regime on column hydrodynamics In a bubble column, two flow regimes were studied by Zahradnik and Kastanek, [1979]: the bubbling regime and the foam regime. In a bubble column, as summarised in the bubble column flow regime section, there are three typical regimes occurring: the bubbling regime, the churn turbulent regime and the slug flow regime (which only happens in small columns). Hence, in their article, Zahradnik and Kastanek [1979] describe a fourth flow regime which is the foam regime. This flow regime is not a typical flow regime in a bubble column and it is not usually characterised in the literature. The bubbling regime, in a bubble column can be described has a "uniform gas distribution 39 characterised by the presence o f a great number of small-scale eddies of a random nature induced in the layer by rising bubbles where no macroscale currents take place" [Zahradnik and Kastanek, 1979]. In comparison, in the foam regime "the gas-liquid layer is quiet. The turbulent eddies are reduced considerably and the existing of this regime is characterized by a significant increase of the average porosity (gas holdup) of the bed in comparison with the regime o f turbulent bubbling" [Zahradnik and Kastanek, 1979]. Furthermore, in the foam regime "the character of gas-liquid dispersion and the value of porosity were further affected by the geometrical parameters o f distributing plates whereas no effect of plate geometry on the hydrodynamic properties o f the gas-liquid layer was observed under turbulent bubbling conditions" [Zahradnik and Kastanek, 1979]. In the context o f a pulp suspension and the two-phase model, the bubbling regime would be the flow regime induced by the dense phase since its gas holdup is not affected by the geometry o f the column. But i f we only look at the gas holdup values, since the gas holdup o f the dense phase is smaller than the gas holdup induced by the large bubbles i.e. the dilute phase gas holdup, the foam regime would describe better the dilute phase o f the pulp suspension since the foam regime gas holdup is higher than the bubbling regime gas holdup. But this foam regime does not really exist in pulp suspension in retention towers but it does exist in flotation systems. Zahradnik and Kastanek [1979] also found that the foam regime is present at moderate superficial gas velocities (up to 0.15 m/s) and above a certain number of holes per unit area on a plate distributor with a column diameter of 0.152 or 0.292 m. Their data also showed that the maximum value o f porosity ( gas holdup) and the maximum porosity increase compared to turbulent regime conditions is obtained using a perforated distributor with the smallest holes (d 0 = 0.5 mm) [Zahradnik and Kastanek, 1979]. In conclusion, to obtain the highest values o f gas holdup, the foam regime is the most desirable regime to operate at [Zahradnik and Kastanek, 1979], although, it has not been proven that an increase in the gas holdup under a foam regime would also mean an increase in the mass transfer occurring in the column [Zahradnik and Kastanek, 1979]. Hence from their results it is obvious that "with increasing value o f average 40 porosity that is with more developed foam-like character o f the bed, the nonuniformity of porosity distribution is more apparent" [Zahradnik and Kastanek, 1979]. As for the bubbling regime, in turbulent conditions, Zahradnik and Kastanek, [1979] found that the total gas holdup could be described by the following expression (for a superficial gas velocity lower than 0.15 m/s): e = ( U g ) / ( A + B U g ) [2-8] where U g is the superficial gas velocity. The right hand side describes "a functional relation between the rising velocity of unequal-sized bubbles under conditions of mutual interference and the superficial gas velocity". 2.3 Mass transfer The rate at which a component is transferred from one phase to the other depends upon a mass transfer rate or mass transfer coefficient and upon the degree o f departure of the system from equilibrium [Treybal, 1980]. In bubble columns or slurry columns, the gas-side resistence to mass transfer can often be neglected in determining the overall gas-liquid mass transfer due to low solubility o f the gas in the liquid [Fan, 1989]. Thus, the resistance to mass transfer is much larger in the liquid than in the gas phase. The mass transfer coefficient is discussed in terms o f k L a [Fan, 1989]. When material transfers from one phase to the other across an interface, the resistance to mass transfer in each phase causes the creation o f a gradient of concentration on each side of the interface [Perry's, 1984]. The assumption is that thermodynamic equilibrium is reached as soon as the two phases are in contact. In dilute phases, the mass transfer rate is then equal to the difference between the bulk concentration and the concentration at the interface [Perry's, 1984]: NA=kt-(x,-x) [2-9] where N A is the mass transfer rate, k L is the liquid-phase mass transfer coefficient, x is the mole fraction of solute in bulk liquid phase, X; is the mole fraction o f solute in liquid at the interface 41 The interfacial area per unit volume of apparatus is usually very hard to determine [Perry's, 1984]. Therefore, the mass transfer coefficient is usually presented as k L times a; the volumetric mass transfer coefficient, k L a [Perry's, 1984]. Many physical models have been used to describe the gas-liquid interaction at the interface, such as: the film, penetration, surface renewal, film penetration models [Gehlawat, 1986]. The hydrodynamics o f the mixture is then very important because it influences the speed and the process o f the gas reaching the interface of the solid or the liquid. Furthermore, in order to evaluate the mass transfer coefficient, physical and chemical methods have been developed [Gehlawat, 1986]. The technique used to calculate the k L a in a pulp suspension is described in the next section. Gas to liquid mass transfer in a three-phase fluidized bed decreases slightly with the gas superficial velocity [Lee et al., 1997]. The physical and thermal properties o f the three phases as well as the geometric characteristics of the bed and wall influences the mass transfer [Kim and Kang, 1997]. For example, the liquid temperature can influence the diffusivity, the liquid viscosity or the bubble coalescence [Kikuchi et al., 1995] or an increase in the gas surface tension results in a decrease o f the k L a value [Fan, 1989]. The presence o f solid enhances the bubble coalescence and therefore decreases the k L a value [Fan, 1989]. In bubble columns, the presence o f solids decreases the effective area (a) and therefore decreases the k L a values [Gehlawat, 1986]. Furthermore, tiny bubbles were found to not influence mass transfer significally [Khare. andNiranjan, 1995]. Gas-liquid mass transfer in pulp suspension mixing operations was investigated by Bennington et al., [1997]. They found that the mass transfer rates are significantly lower in pulp suspensions of 5 to 10% compared with water. Mass transfer rates decreased with increased pulp consistency. In conclusion, mass transfer is influenced by the hydrodynamics of the mixture. A more 42 thorough investigation of the hydrodynamics o f pulp suspension at medium-consistencies is needed to understand the influence o f mass transfer in the overall delignification in oxygen delignification processes. 3 . E X P E R I M E N T S M E T H O D O L O G Y A N D C A L C U L A T I O N S 4 3 3.1 Materials The three pulps obtained from Canadian Forest Products Ltd. were composed o f 45% spruce, 45% pine and 10% fir. The never-dried brown stock pulp ( B S T C K ) had a kappa number of 33 and an average fibre length of 1.08 mm . The never-dried bleached pulp had 88-89% brightness and came from a DEpD process. It has a fibre length of 0.74 mm. Finally, the previously-dried bleached pulp came from the same bleaching sequence as the never-dried bleached pulp had an average fibre length of 0.73 mm. The complete fibre length distributions are presented in Annex I. A fourth pulp was used only in the 2-D experiments. It is previously-dried stone ground wood (SGW) pulp. It's fibre length is 0.61 mm. 3.2 2-D Experiments The 2-Dimension experiments are experiments done in a very narrow rectangular tower that allowed us to observe bubbles as they rise up the channel. 3.2.1 2-D channel apparatus The 2-D channel is made of two 1.3 cm thick plexiglass plates separated by 3.8 cm.. The inside of the channel has the following dimensions: 90 cm wide by 3.8 cm deep by 122 cm high. Air enters through by a Swagelock® fitting at the bottom-centre of the tower as shown in Figure 3.1. The air tubing of the inlet is a small pipe made of 0.05 cm stainless steel. The air flow is regulated and measured using by a rotameter. The rotameter, model 32915-68 from Cole-Parmer®, has a range of 0 to 50 slpm. 4 4 F i g u r e 3.1: P i c ture o f 2-D channe l w i th d imens ions o f 900 m m wide by 38 m m deep by 1222 m m h igh used in the 2-D experiments 4 5 3.2.2 Methodology The procedure was very simple. First S G W or B S T C K pulp was made up at the desired consistency (C m ) . Then the tower was filled up with the different suspensions and stirred to make sure that the water and the pulp were well mixed in the channel. The different consistencies in the channel were 1, 2, 4 and 7 %. A i r was injected over a range of 1-40 slpm. The tower was lighted by 4 lights placed behind the channel. The light coming through the channel would make the rising up of the bubbles easier to see. Some pictures were taken and everything was filmed with a high speed video camera. After, the experiments were done, the channel was drained and cleaned. Using the frame by frame pictures of the video taken, we calculated the rising velocity o f the bubbles and described the flow behaviour. A n example of the calculations is shown in Annex A . 3.3 3-D Experimental apparatus: the pilot-scale colurnn Figures 3.2 to 3.6 show the apparatus used for the gas holdup measurements and the fibre optic probe experiments. The pilot-scale column is 1.83 m (6 feet) high (that is the level of water or pulp before air is bubbled through the column) and has a 0.28 m (11 inch) inside diameter (L /D ratio o f 6.5:1). The pilot-scale column can be made higher be adding two 0.46 m (1.5 feet) high sections to the top o f the tower. It is made o f 1.27 cm (0.5 inch) thick plexiglass, so that we can see through the column. The different sample points are distributed along the height of the pilot-scale column, allowing measurements to be done at 30.48 cm (1 foot) intervals. The diameter was designed to be large enough so that the gas superficial velocity would be in a turbulent or homogenous regime [Viswanathan, 1986]. At the bottom of the pilot-scale column, the 2.54 cm (1 inch) thick gas distributor, composed of400 1 mm holes (diameter) spreads the gas throughout the column. The holes are evenly distributed on the circular surface of the distributor. The design o f the distributor was made using the formulas for a fluidized bed distributor given by in Kuni i and Levenspiel [1991]. 46 The procedure used to design the distributor is given in Annex A . The pressure is measured by sealed stainless steel pressure transmitters number E-07356-00 (two 15 psig probe) and E-99999-99 (two 2 psig and two 5 psig probe) from Cole-Parmer®. The fixed 24 V D C power supplies, also from Cole-Parmer® (number E-26900-00), complete the pressure transmitter apparatus. The fibre optic probes are described in section 3.4.2. The information acquired from the probes is first converted by the PCI-1200 data acquisition system card from National Instruments™, and then the data acquisition system: "Measure for Windows" from National Instruments™ stores it in a Pentium® computer. Figure 3.2: Picture of the computer and of the connections where the probes used were connected to the computer Figure 3.3: Picture of the pilot-scale column (empty) used to performed the 3-D experiments Figure 3 .4: Picture of the pilot-scale column (filled with water) used to performed the 3-D experiments 50 Figure 3.5: Picture of the air inlet at the bottom of the column used to performed the 3-D experiments 51 Figure 3.6: Picture of the pilot-scale column (filled with 1 % consistency BSTCK pulp) used to performed the 3-D experiments 52 3.4 B u b b l e s ize and rising v e l o c i t y exper iments Bubble velocity, size, frequency and dispersion are valuable parameters needed to understand how mass transfer occurs between the gas and the pulp fibres in an oxygen delignification process. Fibre optic probes have recently been shown to be adequate tools for measuring these local fluid parameters [Chabot, et al.,1992]. Furthermore, fibre optic probes have the potential to be safely applied at high temperature and high pressure. Therefore, these in-situ probes are valuable devices to investigate the fluid dynamics of industrial or pilot-scale fluidizing beds [Chabot, et al., 1992] and could likely be applied to pilot-scale oxygen delignification systems. 3.4.1 Principles of fibre optic probes Optical sensor technology is based on reflection, transmission or refraction principle of light into a fibre. The reflection probe uses one optical fibre to guide the light source for illuminating the investigated reactor volume section and another fibre to receive the reflected light. The intensity of the reflected light determines the concentration of solid present in that part of the reactor. One can also use several probes to have a more complex analysis. The solids velocity and the solids flow direction can then be determined [Chabot, et al., 1992]. The transmission probes also use a light source to penetrate the studied volume. But, a receiving fibre, located on the other side of the reactor, guides the transmitted light to a photo detector. The intensity of the transmitted light (which is a function of light scattering) determines the solids velocity and the solids concentration; this concentration being a influencing parameter in the scattering of the light [Chabot, et al., 1992]. Finally, the refraction based probes have the ability to distinguish the liquid from the gas. The way they operate is that the spherical bulb fibre optic probe is sensitive to changes in refracting indices caused by the alternating presence of gas bubbles and liquid at its detecting 53 tips. The light source is brought to the detecting end and reflected back when it is in contact with the gas phase or refracted and lost when the detecting end is in contact with the liquid [Chabot, et al., 1992]. A draw back of these probes is that some probes can not be used to analyse three-phase systems [Chabot, et al., 1992]. This is not true for all the fibre optic probes since Ouellet et al. [1996] used fibre optic probes to measure the mean residence time o f pulp in refiners. The basic principle used by the refraction or reflection probes is SnelPs law: mjO^mA [3-1] where m, and m 2 are the indexes o f refraction, and @ t and €>2 are the angle o f incidence and refraction o f the light measured from the normal to the surface. For example, i f ml = 1.5 (plastic) and sin0j is 0.87 (60°); all incident light that has an refraction indice (m 2) larger than 1.3 (1.5 X 0.87) wil l be refracted and, i f m 2 is smaller than 1.3, the light wi l l be reflected back into the fibre [Saberi, et al., 1995]. The bubble velocity can be calculated knowing the distance between two probes and dividing it by the delay in time of the perturbation of the light signal induced by the passing of a bubble at the top of each probe. The mean value o f the perturbation o f the light signal from the probes is found by the "peak" reached by the light signal telling when there was a bubble passing at the tip of the probe [De Lasa, et a l , 1984]. Then, knowing the bubble velocity, the bubble size can be calculated from the duration of the light signal generated by the contact between the bubble and the tip probe; the diameter of the bubble is proportional to the duration of the high intensity signal [Saberi, et al., 1995]. While studying bubble or solids flow using any fibre optic probes, these probes disturb the flow to a certain degree. Nevertheless, the studies have shown that the results are within 10 to 20 % [Saberi, et al., 1995] of those obtained by photography over a wide range o f bubble chord lengths (1 to 40 mm) [Chabot, et al., 1992]. However, the column used was small and the wall effects may have affected the larger bubble behaviour [Chabot, et al., 1992]. 3.4.2 Desc r ip t i on o f fibre opt ic probes 54 Fibre optic probes were used to measure the bubble size and velocity. Each probe is made of a central fibre with a diameter of 2 mm and around it are 15 small fibres with a diameter o f 750 um. "The light is carried into the reactor by the small fibres and carried back into the photo detector by the central fibre. The photo detector returns a voltage signal proportional to the intensity of the light. The voltage signal was sampled at 100 Hz . The fibres are made o f acrylic coated with a refractive sheathing. In order to hold all the fibres together, small ribber O-rings were used. The fibres are then coated with optical epoxy glue. The tip of the fibres is capped with a small glass window and bonded inside a stainless steel sleeve." [Ouellet, et al., 1996]. Figure 3.7 to 3.9 illustrate the optic probes and it's light. The 300 W D C halogen lamp was taken from a slide projector and along with it is a small computer fan mounted to the side o f the lamp and blowing air at the fibres to cool them [Ouellet, et al., 1996]. Receiving fibre diameter: 2 mm Transmitting fibres diameter: 0.75 mm Figure 3.7: Schematic of one fibre optic probe used to performed the bubble size and rising velocity experiments 55 Figure 3.8: Picture of the two fibre optic probes used to performed the bubble size and rising velocity experiments 56 Figure 3.9: Picture of the 300 W DC halogen lamp and the small fan used with the fibre optic probe 57 3.4.3 M e t h o d o l o g y The approach was to fill up the pilot-scale column with pulp and inject air at the desired flow rates. Two fibre optic probes were placed 1 cm apart at three quarters of the height of the pilot-scale column. The air flow rates were 5, 10, 18.4, 25.7, 36.8, 45 and 50 slpm which gave superficial velocities between 0.0016 and 0.02 m/s. The air was bubbled into the column for 3 minutes at a sampling rate of 100 H z after the flow had reached steady state. Around 60 bubbles were detected with the fibre optic probe signal in a typical test. Interpretation was based on a minimum of 60 bubbles. The two fibre optic probes were separated by 2 cm. The procedure was done with water, never-dried bleached pulp and previously-dried bleached pulp with consistencies of 1 and 2 %. Never-dried B S T C K pulp was also used at 1,2, 5, 8.5 and 10 % pulp consistency. Evaluation o f bubble size and the bubble rising velocity was done using the following procedure. Both probe signals were compared. When a bubble passed the probe, the signal changed, and made a peak as shown in Figure 3.10. The two peaks from the two signals are matched and the delay between them is calculated (delta t). Knowing the distance between the two probes: 2 cm (delta s), we can then calculate the rising velocity U b o f the passing bubble: Then, we calculated how long the signal is disturbed by the passage of the bubble and multiply it by the rising velocity to get the bubble size. 58 4 5 6 7 8 9 time (s) Figure 3.10: Optic probe signal from the two probes separated by 2 cm, at an air flow rate of 18.4 slpm, using B S T C K pulp; arrows indicate bubble passage 3.5 Gas holdup experiments using pressure probes Gas holdup is an important parameter used to describe and understand the behaviour of gas in a liquid-solid medium. The gas disengagement technique and the density variation method were used to measure and analyse the gas holdup values. 3.5.1 G a s disengagement technique In order to use the gas disengagement technique, the basic requirements are that the holdup structure remains undisturbed by bubble interactions after cutting off the gas supply. Further, the holdup structure is usually assumed to be axially uniform [Shumpe and Grund, 1986]. Also, the rate of disengagement process is assumed to be constant and the cross sectional area occupied by rising bubbles remains constant throughout the disengagement [Daly, et al., 1992]. Another very important assumption is that all the gas can eventually be removed from the system. When air is already present in the system before the experiments, the gas disengagement technique does not account for it. 59 Many problems can be encountered while performing this technique as described by Shumpe and Grund, [1986]. first, during disengagement o f large bubbles, the dispersion height fluctuates strongly and is not well defined because o f splashing o f the liquid. Second, disengagement starts from the bottom so the resulting downward flow of pulp suspension wil l affect the rise velocity of the small bubbles. Third, gas may leave the sparger after the air is cut off. Figure 3.11 illustrates typical data from the gas holdup technique applied to the disengagement of bubbles. Then some definitions are presented in order to calculate the different gas holdup values along with a calculation procedure permitting us to avoid the problems mentioned by Shumpe and Grund [1986]. The calculation procedure is the following: first bubble air into the pilot-scale column until steady state is reached. Then start measuring the pressure and cut off the air. Measure the pressure until all the air has escaped from the column. O f course, with pulp suspension at medium-consistency, this is not possible since air gets trapped in the fibre suspension. As we can see in Figure 3.11, period I represents the escape of the larger bubbles and some of the small bubbles, and period n, the escape of the remaining small bubbles. The H 1 3 height is a calculated value used in the calculation procedure (described after the equations) in order to evaluate the disengagement o f the small bubbles in period I. It does not represent any height on Figure 3.11. The H 1 3 values is quite similar to the H x value, but H 1 3 is calculated with the procedure described after the equations and H I is taken from Figure 3.11. HO HI H2 II Dh/dt i,H2, H i H13) tO tl t2 60 Figure 3.11: Gas disengagement technique analysis [Shumpe and Grund, 1986] Equations: Ho- H2 Total gas holdup: £ = = £ > + £ = £ < , • ( ! - £ * ) + £ # Ho [3-3] Large bubbles and dilute phase gas holdup: £b = Ho- His Ho [3-4] Small bubbles gas holdup: £s = His- H2 Ho [3-5] Dense phase gas holdup: £dj = £ - £ t Hu-Ho 1 - £b H 13 [3-6] Gas superficial velocity: UK = dH dti [3-7] Small bubbles superficial velocity: UGS = dH dtis [3-8] 61 Large bubbles superficial velocity: UGI [3-9] Large bubbles rising gas velocity: UGI UM [3-10] Small bubbles rising gas velocity: UGs Ubs [3-11] Ss Where FL, is the height of the fluid with air bubbling through it, H x is the height of the fluid after the escape of the big bubbles, H 1 3 is the calculated height of the fluid after the escape o f the big bubbles (described in the following paragraph), H 2 is the height of the fluid without air, t 0 is the time when the air is cut off, \ x is the time after the escape o f the big bubbles, t 2 is the time at which all the bubbles have escaped the pilot-scale column, I is the period where big bubbles and some o f the small bubbles escape (between (t0, H 0) and (tj, Hj)) Is is period I where some of the small bubbles escape (between (tj, PL,) and (t l 5 Hi) ) and II is the period where the remaining small bubbles escape (between (t,, H^ and (t2, H2)) Using Figure 3.11, the recommended procedure for analysis o f a disengagment curve is as follows [Shumpe and Grund, 1986]: 1. Determine the mean dispersion height at continuous operation (with air bubbling through the pilot-scale column): Ho, 2. Fit a straight line to the experimental observations during small bubble disengagement (period II), 3. Calculate the intercept (t l 3 H t) o f this line with another line through (to, fL,) with 4 a slope o f - u G (in period I), Use an iterative procedure to determine the value of H 1 3 which is a calculated height, not represented on Figure 3.11. It based on the values of (t u Hj) and the following equation: 62 Change: dH dtb [3-12] Then calculate: His = Hi-ti» (dH\ V dtis) [3-13] until: ti dH dH ( « • - " » ) . ( * , - » , . - ) - X L £ _ dtis dtu Ho* {Hi-Hi) [3-14] Set the superficial flow rate of small bubbles u ^ as: dH dt Is UGS [3-15] 6. Use these last equations (3-3 to 3-11, 3-13, and 3-15) to calculate the fractional holdups, the superficial flow rate o f large bubbles and the bubble rise velocities. B y this approach, the effects of the gas release due to the pressure equilibration in the distributor and the downward liquid flow are taken into account [Shumpe and Grund, 1986]. But, i f there is gas present in the mixture in the tower before the experiments, the gas disengagement technique does not account for it. A s this occurs for pulp suspensions where the pulp consistency is higher or equal 5%, a modified method must be used. 3.5.2 The density variation method In order to account for the fact that air is present in pulp suspension at medium-consistency, the gas holdup can also be calculated by evaluating the difference in pressure between two given heights in the pilot-scale column. The gas holdup is the amount o f gas 63 volume present in the pulp. The amount o f air present in the pulp wil l thereby change the density o f the pulp suspension. We can then evaluate the pulp gas holdup by calculating the difference in density o f the pulp when no air is injected in the pulp and when air is present in the pulp. B y knowing the pressure at two different heights, we can calculate the density of the pulp between those two heights: exp eriments — [3-16] The presence o f air in the pulp influences the density o f the fibre-water suspension. A s the pulp consistency increases, the amount of air trapped in the pulp increases as well [Lindsay and Gullichsen, 1996]. To find out the exact amount of air present in the pulp at the beginning, during or after the experiments, the total gas holdup was calculated using a reference density i.e. the density that the pulp would have without air: _ 1 preference ~ \ [^"17] + Pfibre P mater The gas holdup can then be calculated either before the experiments (to find out how much air is present in the pulp), during the experiments (to find out the gas holdup during the bubbling o f air in the pilot-scale column) or after the experiments (to find out how much air is still remaining in the pulp). The following general equation can be used: Oreference /7 exp eriments £ = [3-18] Where the density o f experiments can either be the density before, during or after the experiments, depending on when the total gas holdup is evaluated. This method has only been used to calculate the total gas holdup o f the pulp. 64 In summary, there are two methods to calculate the total gas holdup: the gas disengagement technique and the density variation method. The gas disengagement technique was developed to calculate the total gas holdup in a slurry bubble column, where all the gas escapes the column and none o f the gas is trapped in the solids. There are, in those systems, two phases: the dilute phase and the dense phase. The dilute phase is composed o f the large bubbles in the liquid. The dense phase is composed of the small bubbles, the liquid and the solids. In pulp suspensions, we need to account for the fact that before the gas disengagement technique is applied, gas is trapped in the pulp fibre network. Therefore, the dense phase gas holdup calculated by the gas disengagement technique wil l not take that amount of gas into account. The dense phase gas holdup of a pulp suspension (e d f p u l p ) would then be the dense phase gas holdup calculated with the gas disengagement technique ( e d f g a s - < f e e n 8 a g e m e n t - t e c h ) p m s a residual gas holdup term eRpvip that takes into account the change in the amount of air trapped in the pulp suspension. pulp (~* gas_disengagement_tech . pulp V} 1 C\~l Go? — Od/ + ox [j-iyj The density variation method can give us the amount gas present in the pulp before and after the experiments. 3.5.3 M e t h o d o l o g y The method used was to fill up the pilot-scale column with B S T C K at 1, 2, 5, 8.5 and 10 % pulp consistency, never-dried bleached pulp and previously-dried bleached pulp at 1 and 2 % pulp consistency and water. A i r was injected in the pilot-scale column until steady state was reached. Then, the air was cut off and the pressure was measured. Figure 3.12 illustrates the pressure probes used. The technique was repeated 6 times for every air flow rate. The air flow rates were 5, 10, 18.4, 25.7, 36.8, 45 and 50 slpm.The pressure is measured by sealed stainless steel pressure transmitters number E-073 56-00 (two 15 psig probe) and E-99999-99 (two 2 psig and two 5 psig probe) from Cole-Parmer® 65 Figure 3.12: Picture of one pressure probe number E-99999-99 (2 psig) from Cole-Parmer® 3.6 D i s s o l v e d o x y g e n measurement 66 In order to understand what is happening in oxygen delignification, mass transfer has to be investigated. To do so, dissolved oxygen measurements in a laboratory-scale column were taken and analysed using an oxygen probe. Several methods can be used to measure the k L a. There are two classes of methods: the steady state method and the dynamic method [Linek, et al., 1987]. Listed below are some of the methods used with gas-liquid systems: Dynamic with measurement in the liquid • Dynamic with measurement in the gas Interrupted dynamic measurement (interrupted gas-in) Sulfite • Glucose oxidase The steady state methods use oxygen absorption in a two-phase continuous flow reactor or in a semi-batch reactor with the liquid being in the batch mode [Linek, et al., 1987]. The dynamic methods measure the oxygen concentration following a step change in the concentration of the gas entering the system. Usually, the oxygen concentration is measured in the liquid phase but it can also be measured in the gas phase [Linek, et al., 1987]. In order to use any of the methods for the measurement o f k L a, one must be able to make the following assumptions [Linek, et al., 1987]: 1. ideal mixing o f the two phases in the reactor (which is not the case with pulp suspensions, particularly at medium-consistency). If ideal mixing is not present under operating parameters, the k L a values found wil l be local values instead of overall values. 2. negligible resistance o f the gas phase to oxygen transfer across the interface 67 One must also consider that with the dynamic methods, "the oxygen concentration in the bubbles may be changed over the whole range, i.e. from zero concentration to that in oxygen containing gas. In the measurements using nitrogen and air as the gas medium , the intensity of coalescence determines whether the bubble contents with the initial and newly supplied gases wil l be ideally mixed " [Linek, et al., 1987]. When ideal mixing fails to occur, the resulting values of k L a will be lower; the lower the bubble coalescence, the higher the k L a values [Linek, et al., 1987]. Because of its simplicity, the dynamic method measuring the oxygen concentration in the liquid was chosen. The gases used wil l be nitrogen and air since they are easily available and not dangerous. 3.6.1 Dynamic method-measurement in the liquid phase of pulp fibre suspensions This dynamic method is based on the measurement, in the transition state, o f the dissolved oxygen concentration variation into the liquid phase o f the pulp suspensions. At first the liquid is de-gasified with nitrogen to make sure that there is no oxygen present in the liquid. With pulp, that is difficult because oxygen might remain trapped in the pulp suspension; so the zero value may not necessarily be zero. In fact, the dissolved oxygen value reached only 8.7% instead of 0%. Then, air is injected into the liquid or pulp suspension until a steady state value o f dissolved oxygen is reached. Nitrogen is reinjected into the pulp suspension, and dissolved oxygen is measured with an oxygen probe. The k L a can easily be calculated with: dC/dt = kLa »(C*-C) [3-20] or kLa = l/(t-t0).ln((C*-C0)/(C*-C)) [3-21] Where C * is the saturated value o f oxygen in the liquid. Since the concentration values C are proportional to the dissolved oxygen (DO) values, we find that C * does not need to be known to do the calculations. C = D O C * [3-22] 68 and k L a = l/(t-to>ln((l-D00)/(l-DO)) [3-23] 3.6.2 O x y g e n probe Oxygen probes allow us to monitor the level of dissolved oxygen present in a liquid or gaseous medium. The probes do not measure the actual dissolved oxygen concentration; they measure the partial pressure of oxygen in the liquid and convert it to the concentration o f oxygen in the liquid. There are two types o f probe: ® the galvanometric or potentiometric probe and © the polarographic probe. In both cases, the probes are made of a set o f anode-cathode immersed in an electrolyte solution. The anode-cathode set is separated from the medium containing the oxygen by a membrane which is permeable only to dissolved oxygen. The probe used for these experiments is a galvanometric probe, model 830 from Cole Parmer. Here are the three steps required for the probes to measure the dissolved oxygen in a liquid or gaseous medium as illustrated in Figure 3.13: 1. 0 2 diffuses through the membrane 2. A t the cathode, 0 2 is reduced 3. The galvanometric probe converts the tension induced and the polarographic probe converts the electric current induced into the percent of saturation o f oxygen. 69 height Film (liquid phase) distance Cathode lectrolyte Figure 3.13: Concentration of dissolved oxygen at the interface of the dissolved oxygen probe [Jolicoeur et al., 1996] Oxygen profile Liquid side: P Q 2 = C L « H e L [3-24] Membrane side: = P Q 2 / Hej^ [3-25] At the interface, the partial pressures of oxygen are the same so: C M at interface = C L # H e L / H e M [3 -26] Where P G 2 is the partial pressure o f oxygen in the liquid, C L is the oxygen concentration in the liquid, C M is the oxygen concentration in the membrane (careful: pulp consistency is C M , not C M ) , H e L is Henry's constant in the liquid and is Henry's constant in the membrane 7 0 3.6.3 Troubleshooting There are two major problems that can occur while measuring the dissolved oxygen concentration with an oxygen probe: X 0 2 has to diffuse through the membrane which means that it is a slow process, X A t the liquid-membrane interface, a stagnant film can add an additional resistance to the diffusion of oxygen through the membrane. In order to be able to deal with and understand those problems, four parameters are important: I the membrane ( D D 2 as a function of the semi-permeable material), II effect of temperature on the permeability o f the membrane (P = k^e"1"71 where k and m are constants and T is the temperature in K ) , III effect o f the dissolved salts on the solubility of oxygen in the liquid medium, I V effect o f the hydrodynamics near the membrane which determines the thickness of the stagnant film. If the probe is a good quality one, those four parameters wil l have been addressed in the construction of the probe for slower mass transfer rates. At high mass transfer rates, the delay of the probe may be a problem. Furthermore, since we are dealing with a pulp suspension, the stagnant film might be thicker than expected. In fact, according to Linek et al. [1987], "it has been shown that in viscous liquids it is not possible to determine the transient probe characteristics i f the film resistance is significant because, no matter which procedure is employed, the hydrodynamic and concentration boundary layer are formed simultaneously. This means that the transient characteristics are measured under hydrodynamic conditions different from the steady state regime in measurements in a test process". Furthermore, the k L a values calculated from the oxygen measurements are not the real values because: l / k L a B d = l / k L a l i i e M y + t [3-27] Where t is the probe response time and t < k ^ ^ / 5 [3-28] 71 One other important requirement is the linearity of the probes. Many probes fail to meet this requirement when the oxygen concentration is very high [Linek, et al., 1987]. The saturated value of oxygen in the water at 20°C is 9.08 mg/L. Finally, when the oxygen concentration is measured in a liquid in which lots o f gas is present; i f the bubble size of the gas is comparable with the cathode size and i f the oxygen concentration in the liquid is not in equilibrium with the oxygen concentration on the bubbles, the location o f the probes becomes very important. If the probe is misplaced, the "bubbles may impinge against the probe membrane, causing a chaotic probe signal that corresponds to the time-average partial pressure of oxygen in front of the membrane" [Linek, et al., 1987]. Therefore, it is necessary to prevent the contact between the bubbles and the membrane. O f course, i f no bubbles are present, the probe location does not influence the probe interactions [Linek, et al., 1987]. 3.6.4 Apparatus Figure 3.14 shows the apparatus. The laboratory-scale column is 0.3937 m (15.5 inches) high and has a 0.1 m (3.75 Inch) inside diameter ( L / D ratio o f 4.1:1). It is made o f 0.64 cm (0.25 inch) thick plexiglass, so that we can see through the column. The different sample points are distributed along the height of the laboratory-scale column, allowing measurements to be done at every 5.08 cm (2 inches). At the bottom of the laboratory-scale column, the 2.54 cm (1 inch) thick gas distributor, composed of forty 1 mm holes distributed the gas evenly throughout the column. The holes are evenly distributed on the circular surface o f the distributor. The design o f the distributor was made using the formulas for a fluidized bed distributor as shown in Kuni i and Levenspiel [1991], as we did for the design of the pilot-scale column. Figure 3.14: Picture of the small laboratory-scale co lumn used for the mass transfer experiments 73 3.6.5 Methodo logy Never-dried B S T C K pulp is used to do these experiments. The procedure was the following: 1. F i l l up the laboratory-scale column with 8% or 3% consistency B S T C K pulp, 2. De-gas the column with nitrogen gas, 3. Inject air into the laboratory-scale column and wait for a steady state of dissolved oxygen value, 4. Change the air inlet pipe for the nitrogen inlet pipe to allow nitrogen to go though the column, 5. Start acquiring data until a steady state value of dissolved oxygen has been reached. 3.6.6 C a l c u l a t i o n s The dissolved oxygen probe sends back the percentage of dissolved oxygen. To calculate the k L a from the D O values, we use the following equations: kLa=—*\n^———{ [3-29] Where delta t is the time during the experiments, D O 0 is the dissolved oxygen value during steady state (at t 0) and D O , is the dissolved oxygen value during the experiment at time t Figure 3.15 illustrates an example o f 4 experimental curves (replicates) obtained with a gas flow rate o f 6.1 slpm in 8% consistency B S T C K pulp. To account for the response delay o f the probe, the first D O values acquired have to be dismissed until the D O values start decreasing. The probe gives 90% o f the reading in 10 seconds. The internal probe dynamics is accounted for in the construction of the probe. 74 Figure 3.15: Example of 4 replicated curves used to calculate the k L a values in the small laboratory-scale column with 8% consistency BSTCK pulp and an air rate of 6.1 slpm 3.7 Error analysis The error bars were evaluated by calculating the standard deviation o f the measured value for repeated runs conducted under identical conditions. In each experiments set, the experimental procedure was repeated several time to see i f the results found are reproducible and to calculate their standard deviation. Details o f the error analysis is presented in Annex J. 75 4, R E S U L T S A N D D I S C U S S I O N 4.1 2 - D Experiments Figures 4.1 and 4.2 illustrate the flow behaviour at different consistencies. As we can see, the type o f pulp, i.e. the suspension rheology and the fibre length have a significant influence on bubble flow behaviour. The following description of the flow behaviour will help develop a picture o f what is happening when bubbles are introduced into pulp. A full description o f the flows as well as the bubble size is presented in Annex B . 4.1.1 Desc r ip t i on o f the bubb le f lows Figure 4.1 illustrate the behaviour 2% consistency S G W (Stone Ground Wood) pulp. The rising bubbles behave like they would in a viscous liquid. A n increase in gas velocity provoked an increased in the size of the oval-shaped bubbles. When the consistency increased to 4%, bubbles are channelling through the pulp. The bubbles are also bigger as they follow their preferential paths. In the fibre network, the floes are small and the bubbles can easily rise up in the 2-D channel. There is also air trapped in the fibre network. But the bubbles are tiny and usually well mixed with the water. When longer fibres are used to make the pulp, i.e. never-dried kraft B S T C K pulp, the behaviour o f the bubbles changes. The bubbles rise up using mostly channels even at low consistency. The bubble size also increases with an increase in the pulp consistency or with the gas velocity (the air flow rate was from 0 to 40 slpm). At 4%, as shown in Figure 4.2, we can see that the pulp is not moving around the bubbles. The fibre network is tighter than with S G W pulp i.e. the fibres floes are bigger and not easily displaced by the bubbles. Furthermore, air is already present in the fibre network. Areas of air bubbles are present in the channel, as we can see in Figure 4.2. At a B S T C K pulp consistency of 7%, we can not see any bubbles rising up because o f the lack of free water in the pulp. Gas channels through the porous network structure, Figure 4.1: Pictures of gas spherical capped bubbles (5 x 7.5 cm) rising through pulp suspensions in the 2-D channel: SGW pulp consistency of 2% 77 Figure 4.2: Picture of channelling or fingering through BSTCK pulp at 4% consistency in the 2-D channel 4.1.2 2-D channel rising velocity results 78 Annex A describes the equations used to calculate the rising velocity and Annex B presents all the results. Figure 4.3 illustrates the results. As we can see, the bubble rise velocity increases with an increase in the.air flow rate, except for the S G W at 4% consistency. Furthermore, the rising velocity of the S G W pulp at 1% is higher than the one at 2%. The 4% results are quite scattered but they are similar to the 1% results. Changing the pulp fibre length and properties also changed the rising velocity. The rising velocity o f the B S T C K pulp is not influenced by the air flow rate and it is a little higher than with the S G W pulp. These results are in agreement with Phillips [1996] results. 0.0E+00 7.0E-05 1.4E-04 2.1E-04 Airflow rate (m3/s) Figure 4.3: Rising velocities for Stone Ground Wood and BSTCK pulp in the 2-D channel 79 4.2 3-D Experiments 4.2.1 Bubble size and rising velocity experiments 4.2.1.1 Description of the BSTCK and bleached pulp flow The flow pattern of the bubbles in pulp changes tremendously as the consistency of the pulp increases. The same behaviour was observed no matter what type of pulp was used (although the pulps used are all very similar). Figures 4.4 to 4.21 show pictures of the pulp and bubbles in the pilot-scale column. In a water bubble column, at low air flow rates (5 to 19 slpm), the bubbles have a spherical-cap shape and the flow is homogeneous. A t higher air flow rates (up to 50 slpm), the flow becomes turbulent and the bubbles have different sizes. There are upflows and downflows and the bubbles are well dispersed in the pilot-scale column. When a 1% consistency pulp is added to the pilot-scale column, the flow changes but resembles the water flow, i.e. the pulp is behaving in the column like water does when air is introduced, as illustrated in Figure 4.4 and 4.5. The bubbles are bigger and they tend to coalesce in order to make their way up in the column. The pulp and the water were well mixed, with upflows and downflows but there are no bubbles rising up after the air is cut off. Channelling is not obvious i f it is present. From the top o f the pilot-scale column, we could see that most of the bubbles rise in the centre o f the column as opposed to rising closer to the walls. Figure 4.5 illustrates this behaviour. These results are in agreement the results found by with Lindsay et al. [1995]. When pulp consistency is increased to 2% as shown in Figure 4.6 to 4.8, the flow becomes less "mobile", i.e the pulp is still moving in the pilot-scale column but a lot less than with 1% consistency pulp present. There are also less currents present than with 1% pulp. Regions of upflow and downflow are still present, but are less numerous compared to 1% consistency pulp. For the same air flow rate, the flow becomes more turbulent (the bubbles shape and size is not 80 homogeneous) than with only water present in the column. The bubbles are bigger. Channelling is present in the pilot-scale column, and bubbles take always the same path as they rise up in the column. When the air flow rate is decreased, channelling is increasing. These results are in agreement with the results found by Lindsay et al. [1995]. If the consistency o f the pulp is increased to 5%, there is no movement o f the pulp. Figures 4.9 to 4.11 illustrate the flow o f pulp in the pilot-scale column. The bubbles are small but we can see that they rise up preferably using the same path along the walls. As before, we can see that the bubbles and the channels are evenly distributed through the pilot-scale column. We can still see a little bit o f the free water moving with the bubbles along the wall o f the column. It is harder to see a clear difference in the height of the pulp when air is present in comparison to the height when there is not air in the column. That is due to the fact that there is a large quantity of air already present in the pulp before we start the experiments. From the top o f the column, we can see that most of the air is coming up in the centre o f the column. At 8.5 and 10% pulp consistency, the bubbles are very hard to see because there is no free water in the pilot-scale column (most o f the water is now located in the pulp fibres). Figures 4.12 to 4.19 illustrate the flow in the pilot-scale column. A t the top of the column, we can hear and feel the air coming out but we can not see bubbles or movement of the pulp. But, i f a very small air flow rate is used for a few minutes and afterward increased to a higher value, the pulp separates in half and the higher half starts rising up as shown in Figure 4.20 and 4.21. When the air is cut off, the higher half o f pulp falls back into it's original place. I f the high flow rate was reintroduced into the column, the pulp separated again and started rising again the first time the air is reintroduced. I f the air is cut off and reintroduced again, the pulp wil l stay in the column without rising up. This behaviour can be explained by the fact that air is trapped in the fibre network and it can not escape. To escape the pilot-scale column, air must agglomerate and form channels. When enough air is trapped together, it can flood the pilot-scale column as we saw (the slugging o f the column is explained in the next section). The air trapped exerts enough forces on the pulp to move it up the column. Figure 4.4: Pictures of 1 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 82 Figure 4.5: Pictures of 1 % consistency BSTCK pulp illustrating the flow of pulp at the top of the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.6: Pictures of 2 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.7: Pictures of 2 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 85 Figure 4.8: Pictures of 2 % consistency BSTCK pulp illustrating the flow of pulp at the top of the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.9: Pictures of 5 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 87 Figure 4.10: Pictures of 5 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.11: Pictures of 5 % consistency BSTCK pulp illustrating the flow of pulp at the top of the pilot-scale column with an air superficial velocity of 0.01416 m/s 89 Figure 4.12: Pictures of 8.5 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.13: Pictures of 8.5 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.14: Pictures of 8.5 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 92 Figure 4.15: Pictures of 10 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 9 3 Figure 4.16: Pictures of 10 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.17: Pictures of 10 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s Figure 4.18: Pictures of 10 % consistency B S T C K pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 96 Figure 4.19: Pictures of 10 % consistency BSTCK pulp illustrating the flow of pulp in the pilot-scale column with an air superficial velocity of 0.01416 m/s 97 Figure 4.20: Pictures of 10 % consistency BSTCK pulp illustrating the separation of the pulp in the pilot-scale column with an air superficial velocity of 0.00315 m/s Figure 4.21: Pictures of 10 % consistency BSTCK pulp illustrating the separation of the pulp in the pilot-scale column with an air superficial velocity of 0.00315 m/s S L U G G I N G OF T H E P I L O T - S C A L E C O L U M N 99 At 5, 8.5 or 10%, we can see that air is being trapped into the fibre network as illustrated in Figure 4.20 to 4.22. After the air has been cut off, the pressure goes down as the bubbles escape out of the column. But after that first decrease in pressure, the pressure goes up again which means that there are some trapped bubbles trying to escape and by doing so push the pulp up again, increasing the pressure read on top of the pressure transducers. This means that there is a danger of slugging the column, i.e. have an excess of air pushing the pulp instead of having a mixture of pulp and air. The slugging has also been seen in the lab experiments when the pulp broke separated into two sections, with the uppermost section rising up the pilot-scale column. E N T R A P M E N T OF G A S I N T H E F I B R E N E T W O R K Figures 4.23 and 4.24 also show the entrapment of air at a consistency o f 8.5 and 10% at a gas flow rate o f 25 slpm (0.00809 m/s). A s we can see, the pressure rises back up after a certain period o f time. The bubbles or air must have agglomerated into the fibre network and pushed the pulp up in order to escape. A l l we see is the pressure going up on the pressure transducer. So when the pressure rises, it shows that the bubbles have agglomerated and tried to escape. The pressure decrease illustrates that the gas disengagement technique can be used to determine the dilute phase gas holdup. But since gas is trapped in the column, pressure does increase after the escape o f the large bubbles. This illustrates that the gas disengagement technique does not account for the presence of air trapped in the pulp both before or after the experiments. o 101 o CM E (6ed) 3inssaj X I X I 3 m 0,5 0 ,45 0,4 0 ,35 0,3 0 ,25 0,2 0 ,15 0,1 0 ,05 0 I C m = 0 % S C m = 1 % • C m = 2 % 0 - 0 . 5 0 . 6 - 1 1.1 - 2 2.1 - 3 3.1 - 4 4.1 Bubb le s i z e (cm) 5 5.1 - 6 6.1 + Figure 4.25: Size of the bubbles in the pilot-scale column with B S T C K pulp and a gas flow rate of 26 slpm 1 >» o o CD > CO m a) xi xi 00 0 , 3 0 , 2 5 0 , 2 0 , 1 5 S 0,1 0 , 0 5 4^ o C m = 0 % • C m = 1 % A C m = 2 % L i n e a r ( C m = 2 % ) L o g . ( C m - 1%) 0 • 0 - 0 . 5 0 . 6 - 1 1.1 - 2 2.1 - 3 3.1 - 4 4 .1 - 5 5.1 - 6 6.1 + B u b b l e s i z e ( c m ) Figure 4.26: Average local bubble rising velocity for B S T C K pulp at 0, 1 and 2 % consistency in the pilot-scale column for gas superficial velocity of 0.00809 m/s (26 slpm) 105 The detection of bubbles is impossible at pulp consistency higher than 2% since the air does not form bubbles: the fibre optic probe signal is all noise. In fact, the smallest bubble size measured is 0.4 cm. Below that value, it is impossible to distinguish the signal noise from the passing o f a bubble. A s we can see, from Tables 4.1 and 4.2 and Figure 4.26, bubble size increases and the bubble size values are more scattered when the pulp consistency increases. Furthermore, the bubble size increases with the air flow rate, except at 2 % pulp consistency. For the rising velocities, we can see that adding pulp to the pilot-scale column decreases the local bubble velocities. From Figure 4.25, we see that the most important fraction o f the bubbles size is between 1 and 2 cm which is in agreement with what was observed in the laboratory. The rising velocities in the water experiments are quite scattered since there was not enough bubbles used to calculate it. Furthermore, the rising velocities used to calculate the bubble sizes (shown in Table 4.2 and in Figure 4.26) are at least 10 times smaller than the rising velocities found with the gas disengagement technique (in the next section) and with the 2-D channel experiments. Indeed, even i f the 2-D channel results show the local bubble rising velocity o f the bubbles, these results are still to high compare (factor of 10) to the ones found in Table 4.2 using the fibre optic probes. The results found with the fibre optic probe agree with what we find in the industrial residence tower. In fact, the residence time o f the gas in industrial towers is between 6 and 20 minutes [Bennington and Pineault, 1999]. Calculating the residence time with the rising velocities found with the fibre optic probes, we find that it is around 10 minutes. The gas disengagement technique results do not account for the fact that there might be air already present in the pulp and that degassing o f air can happen while the experiments are ongoing. That might also explain the difference in range of the two bubble rising velocities since the degassing o f the pilot-scale column may influence the rising velocities o f the bubbles in the column. The rising velocities calculated using the gas disengagement technique Equations (3-10 106 and 3-11) represent an average rising velocity in the pilot-scale column. 4.2.2 G a s holdup experiments The results are summarised and analysed in the following sections. Due to the amount o f results, only a few graphs are presented in these sections. A l l data graphs are presented in Annex D , E , F, and G . In those graphs, some points deviate from the curve set by the other points. We call these deviation points outliers. Since most o f the points on the graphs follow the same curve, the outliers are being excluded from the discussion in the following sections. The outliers are the results of the creation o f preferential channel around the pressure probes and should be ignored since they do not represent the behaviour o f the flow occurring in the whole of the column. 4.2.2.1 Water experiments Before doing the experiments with pulp, the water gas holdup results found with the gas disengagement technique were compared with the results found in the literature. First, this was to make sure that we found the same results with the gas disengagement technique as what was already known in the literature. Second, this comparison allowed us to make sure that the probes were functioning properly. The results shown in Figure 4.27 were compared and are in agreement with Zahradnik and Kastanek [1979], Tsuchiya and Nakanishi [1992] and Fan [1989]. 4.2.2.2 BSTCK pulp total gas holdup Two methods can be used to calculate the total gas holdup in the pilot-scale column. First, the results found using the gas disengagement technique are described. Then, the results found with the density variation method are presented. 107 0,005 0,010 Ug (m/s) 0,015 Top of column - 3/4 of column A Fan [1989] X Tsuchiya and Nakanishi [1992] X Zahradnik and Kastanek [1979] Figure 4.27: Total gas holdup in water as a function of gas superficial velocity in the pilot-scale column G A S H O L D U P R E S U L T S W I T H T H E G A S D I S E N G A G E M E N T T E C H N I Q U E The total gas holdup measures the amount (by volume) o f gas present in the pulp. So by increasing the air flow, we increase the quantity of air present in the pilot-scale column, which increases the gas holdup in the pulp. Also, as we saw in the literature search when one introduces solids into a bubble column, the gas holdup should decrease. A s we can see in Figure 4.28, the difference in gas holdup between the water experiments and the 1, 2 and 5% experiments agree with this. However, with pulps of 8.5 and 10% consistency (results presented in Annex D) , their gas holdup increases again, above that for the water curve. Furthermore, as the consistency o f the pulp increases, the gas holdup increases: thus the gas holdup increases as fibres are added to the suspension. That is easily understood by the fact that the air is getting trapped in the fibre network: the higher the consistency o f the pulp, the tighter the fibre network, the more the bubbles get trapped in it, the higher the gas holdup. Another phenomena that can be observed is that the gas hold up increases with the air superficial velocity until the pulp consistency reaches 5 %. Above this consistency, the gas superficial velocity does not influences gas holdup values (results presented in Annex D) . 108 0,040 0,030 8 0,020 0,010 0,000 0,000 0,005 0,010 Ug (m/s) 0,015 —o— Cm = 0 m Cm = 1 A Cm = 2 x Cm = 5 0,020 Figure 4.28: Comparison of the total gas holdup at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 0,1,2 and 5% consistency as a function of gas superficial velocity With 0 (water), 1 and 2% pulp consistency, the gas holdup is always a little higher at the top o f the pilot-scale column than at three quarters of the column height: € t o p > € 3 / 4 . A t these consistencies, the flow in the column is turbulent. There is movement of the pulp and the bubbles are easy to see as they rise up in the tower. That means that the air bubbles are big enough to push the pulp up the column and that there is enough water around the pulp to permit creation o f bubbles. Therefore, the bubbles are not trapped into the fibre network as they rise up. At a pulp consistency of 5% or higher, the behaviour changes: gas holdup now decreases with the column height: e t o p < e 3 / 4 < €^Me as shown in Figure 4.29. A t these consistencies, the fibre network is tighter and traps the air more efficiently. So the closer a bubble is to the top o f the column, the more chance it has to escape. But, i f a bubble is trapped in the middle o f the column, the chance it has to escape are very small. When air reaches the top of the column, it has agglomerated and it can escape. Therefore, the gas holdup should decrease with the height of the pilot-scale column when there is entrapment of bubbles in the pulp. These results are in agreement with Lindsay et al. [1995] as described in chapter 2. 109 Figure 4.29: 10% consistency B S T C K pulp total gas holdup results as a function of the pilot-scale column height and the gas superficial velocity G A S H O L D U P R E S U L T S W I T H T H E D E N S I T Y V A R I A T I O N M E T H O D Another way to calculate the total gas holdup is to take the differential pressure between two heights. This method is called the density variation method. This method gives results with higher uncertainty since two probes are used to calculate the total gas holdup. Standard deviation is used to calculate the errors. Figure 4.30 shows a comparison o f the total gas holdup results using both ways to calculate it: the gas disengagement technique and the density variation method. 1 1 0 0,30 0,20 0,10 0,00 -0,10 -0,20 T l V f T r i JL 1—1 / 1 ~|_J \ T i i i L 1 i 0,000 0,005 0,010 0,015 Ug (m/s) 0,020 • Density variation! method - •—Gas disengagement (Top of the column) -A- Gas disengagement (Middle column) Figure 4.30: Comparison of the total gas holdup results using the gas disengagement technique and the density variation method using B S T C K pulp at 8.5% consistency as a function of gas superficial velocity The presence of air influences the density o f the fibre-water suspension. A s the pulp consistency increases, the amount o f air trapped in the pulp increases as well. To find out the exact amount of air present in the pulp at any time, the total gas holdup was calculated using a reference density, calculated as the density pulp would have without air. This allows us to evaluate the amount o f air present before, during and after the experiments. Figure 4.32 presents the results for the duration of the experiments with a 8.5% consistency pulp. It was found that for pulp consistencies of 5, 8.5 and 10 % the amount of air present before the experiments is higher than after the experiments. That is why the difference between the gas holdup before and after is negative: there is less air holdup in the pulp after the experiments that there was before the experiments. The fact that air was bubbled into the pulp "degassed" it. This is shown in Figure 4.31. What is happening is that there are some bubbles trapped into the pulp at all time and that when air is introduced into the pulp, it allows the trapped bubbles to leave and therefore, the gas holdup increases the deeper in the column it is calculated. 111 Using the density variation method, the same trends can be seen as with the gas disengagement technique. Although the density variation method usually gives higher gas holdup values than with the gas disengagement technique, that is expected because the gas disengagement technique does not account for air being present in the pulp before and after and that this amount may change overtime. Furthermore, that leads to the total gas holdup being calculated with pulp at 5% consistency is higher than the one calculated with water. 0,00 w -0,05 Density variation method -0,30 0,000 0,005 0,010 Ug (m/s) 0,015 0,020 Figure 4.31: Difference between the gas holdup before and after the experiments using 8.5% consistency of B S T C K pulp as a function of gas superficial velocity using the density variation method 1 1 2 0,30 0,20 0,10 0,00 -0,10 -0,20 0,000 Density variation method 0,005 0,010 Ug (m/s) 0,015 0,020 Figure 4.32: Total gas holdup during the experiments with 8 .5% B S T C K pulp as a function of gas superficial velocity using the density variation method 4.2.2.3 Bleached pulp total gas holdup The gas holdup up experiments were done only at low consistency (1 and 2%) pulp. The purpose o f these experiments was to see i f the flow behaviour and the gas holdup up found was influenced by the pulp having been previously dried or not. We see the same trends as with the B S T C K pulp at the same consistencies. Figures 4.33 and 4.34 show the results found for the never-dried bleached pulp and the previously-dried bleached pulp. Figure 4.35 and 4.36 show the comparison o f the results of the three kinds o f pulp. At a consistency o f 1%, we can see that when the superficial velocity of the air was lower than 0.0125 m/s, the gas holdup is practically the same for the three kinds of pulp. But, at higher superficial velocity, the B S T C K total gas holdup is higher than the bleached pulp gas holdup. Overall, the fact that the bleached pulp was previously-dried or not did not have an effect on the total gas holdup at low consistencies. 113 0,040 0,030 £ 0,020 0,010 0,000 0,000 0,005 0,010 Ug (m/s) 0,015 o Cm = 0 m Cm = 1 A Cm = 2 0,020 Figure 4.33: Comparison of the total gas holdup at the top of the column results using the gas disengagement technique wi th never-dried bleached pulp at 0,1 and 2% consistency as a function of gas superficial velocity 0,040 0,032 0,024 0,016 0,008 0,000 0,000 0,005 0,010 Ug (m/s) 0,015 o Cm = 0 Cm = 1 A Cm = 2 1 1 i 0,020 Figure 4.34: Comparison of the total gas holdup at the top of the column results using the gas disengagement technique wi th previously-dried bleached pulp at 0, 1 and 2% consistency as a function of gas superficial velocity 114 0,0200 0,0160 0,0120 0,0080 0,0040 -| 0,0000 m Brown Stock pulp - A Never-dried bleached pulp -x— Previously-dried bleached pulp 0,0000 0,0050 0,0100 Ug (m/s) 0,0150 0,0200 Figure 4.35: Comparison of the total gas holdup at the top of the column using the gas disengagement technique with B S T C K , never-dried bleached and previously-dried bleached pulp at 1% consistency as a function of gas superficial velocity 0,0200 0,0150 £ 0,0100 0,0050 0,0000 0,0000 0,0070 0,0140 Ug (m/s) m m Si Brown Stock pulp — A — Never-dried A bleached pulp —X- Previously-dried bleached pulp 1 - i 0,0210 Figure 4.36: Comparison of the total gas holdup at the top of the column using the gas disengagement technique with B S T C K , never-dried bleached and previously-dried pulp at 2% consistency as a function of gas superficial velocity 115 4.2.2.4 Results using the gas disengagement technique with BSTCK pulp Using the gas disengagement technique, other results were found from the data acquired: the dilute phase or large bubble gas holdup (equation 3-4), the small bubble gas holdup (equation 3-5), the dense phase gas holdup (equation 3-6), the small bubbles (equation 3-8 and 3-11) and large bubbles (equation 3-9 and 3-10) superficial and rising velocity. One thing that is common to all the results is that they all have great uncertainty: their standard deviation are high as we wil l see in the figures of this section. Most of the results are presented in Annex D . D I L U T E P H A S E and L A R G E B U B B L E S G A S H O L D U P e b The same trends as for total gas holdup were found. Figure 4.37 shows some of the results found. S M A L L B U B B L E S e s and D E N S E P H A S E G A S H O L D U P The results are more scattered than with the total gas holdup. The gas holdup seems the same at the top of the pilot-scale column and at three quarters o f the height of the column at 1 and 2% pulp consistency. But at higher consistencies for the dense phase, we see the same trends in the gas holdup: € t o p < € 3 / 4 < €mjddie- When we compare the results with all the consistencies at every height, we see that at low air flow rates (lower than 18 slpm), the gas holdups o f 1,2 and 5% are almost the same and at higher air flow rates, for the dense phase gas holdup: e,0 > e 8 5 > € 5 > e 2 > € b The small bubbles and the dense phase gas holdup also increases with the air flow rates, but the slope is less steep than with the total gas holdup. Figure 4.38 and 4.39 show some o f the results. 116 Figure 4.37: Comparison of the dilute phase and large bubbles gas holdup at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1, 2, 5, 8.5 and 10% consistency as a function of gas superficial velocity £ s 0,0125 0,0100 0,0075 0,0050 0,0025 H 0,0000 m C m = 1 A Cm = 2 0,000 0,005 0,010 Ug (m/s) 0,015 0,020 Figure 4.38: Comparison of the small bubbles gas holdup at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1, and 2% consistency as a function of gas superficial velocity 1 1 7 0,12 0,10 0,08 8df 0,06 0,04 0,02 0,00 0,000 0,005 0,010 Ug (m/s) 0,015 m Cm = 1 A Cm = 2 x - Cm = 5 —x- Cm = 8.5 -®—Cm = 10 0,020 Figure 4.39: Comparison of the dense phase gas holdup at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 0,1,2, 5, 8.5 and 10% consistency as a function of gas superficial velocity A s discussed in Chapter 3, the gas disengagement technique does not account for the variation in the concentration o f air present before and after the experiments. The dense phase gas holdup o f a pulp suspension (e d f p u l p ) would then be the dense phase gas holdup calculated with the gas disengagement technique (e/as-disen«a8emen,-tech) plus a residual gas holdup (e^) term that takes into account the change in the amount o f air trapped in the pulp suspension. C * P"& C* gps_disengagement_tech . pulp TA ft Odf = Odf + c>r Another way o f analysing the results is to say that the density variation method calculations o f e gives the total gas holdup. In fact, by measuring the differential density between a fixed height and by comparing it to a reference density where no air is present, the total amount o f gas can be calculated. The gas disengagement technique does not give the total gas holdup since this technique does not account for the air present before and after the experiment. The apparent gas holdup represents the holdup that does not account for the amount of gas that remains in the pulp suspension after the experiment. The amount of gas trapped in the fibre network can be calculated using a reference height. This height is calculated using the following 118 equation: = Ho- Ho- £ [4-2] Where H r e f represent the height that we should have found using the gas disengagement technique H 2 . The difference between H 2 and H r e f is the amount o f gas that is trapped into the fibre network: etrappei. We used equation 4-3 to calculated this gas holdup: H i - Href (strapped — [4~3] Ho Table 4.3 shows a summary of the results with a gas superficial velocity of0.00809 m/s. We can see that the total gas holdup increases with pulp consistency. But we also notice that the large bubbles gas holdup represent a big part of the total gas holdup. Furthermore, the amount of gas trapped increases with the consistency, which is in agreement with the results found earlier. Table 4.3: Summary of gas holdup values for B S T C K at gas superficial velocity of 0.00809 m/s at 3/4 of the height of the column (%) e, 6 big t^rapped 0 0.0229 0 0.00229 0 I 0.0071 0.0040 0.0028 0.0003 2 0.0175 0.0034 0.0058 0.0083 5 0.0623 0.0029 0.0149 0.0445 8 5 0.1089 0.0505 0.0264 0.0320 10 0.4969 0.2022 0.0718 0.2229 Here e t is the total gas holdup, e a is the apparent gas holdup, e b j g is the large bubble gas holdup, e s m a B is the small bubbles gas holdup, and e f r a p p e d is the gas holdup that is trapped in the fibre network S M A L L B U B B L E S U P E R F I C I A L V E L O C I T Y 119 At pulp consistencies o f 1 and 2%, the superficial velocities at the top o f the pilot-scale column are the same as in the rest o f the column (all the results are presented in Annex D). The small bubble superficial velocity increases slowly with an increase in the air flow rate. Furthermore the superficial velocity of the pulp at 2% consistency is higher than the one at 1%. Figure 4.40 shows some of the results found. L A R G E B U B B L E S U P E R F I C I A L V E L O C I T Y At pulp consistencies of 1 and 2%, the superficial velocity at the top of the pilot-scale column is the same as in the rest o f the column (all the results are presented in Annex D) . Furthermore the superficial velocity o f the pulp at 2 % consistency is higher than the one at 1%. Figure 4.41 shows some o f the results. A s we can see, the slope is a little bit more steep than with the small bubbles superficial velocity. 120 S M A L L B U B B L E R I S I N G V E L O C I T Y A s opposed to the results o f the superficial velocity, at low consistencyies of 1 and 2%, the rising velocity is smaller at the top o f the pilot-scale column and there is no increase in the rising velocity with an increase in the air flow rate When we compare all the pulp consistency results, for every height we get: U 2 > U j . Figure 4.42 shows some of the results. 0,010 0,008 | 0,006 O 0,004 0,002 0,000 it* m Cm = 1 -A— C m = 2 0,000 0,005 0,010 Ug (m/s) 0,015 0,020 Figure 4.40: Comparison of the small bubble superficial velocity results at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of gas superficial velocity 121 0,0250 0,0200 0,0150 O 0,0100 ZD 0,0050 0,0000 0,000 0,005 0,010 Ug (m/s) 0,015 m Cm = 1 A Cm = 2 0,020 Figure 4.41: Comparison of the large bubble superficial velocity results at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of gas superficial velocity 2,5 2,0 m/s) 1,5 CO 1,0 ZD 0,5 0,0 4 0,000 0,005 0,010 Ug (m/s) 0,015 Cm = 1 A - Cm = 2 0,020 Figure 4.42: Comparison of the small bubble rise velocity results at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of gas superficial velocity 122 One might notice that, aside from the great uncertainty of the rising velocities of the small bubbles, they are comparable to those found for the B S T C K pulp ( C m = 2%) in the 2-D channel, i.e the rise velocity values are around 0.6 m/s at very low flow rates (Ug = 0.00315 m/s). We can then conclude that the narrow tower used to conduct the 2-D experiments are in agreement. But, these results, as mentioned earlier in the bubble size and rising velocity measurement using fibre optic probes, do not represent what we have in the pilot-scale column since the rising velocity calculated with the fibre optic probes gave results 10 times lower than the ones in the 2-D channel and the ones calculated with the gas disengagement technique. L A R G E B U B B L E R I S I N G V E L O C I T Y Also opposed to the results of the superficial velocity, at low consistencies of 1 and 2%, the rising velocity is the same in all the pilot-scale column and there is no increase in the rising velocity with an increase in the air flow rate.. When we compare all the pulp consistency results, for every height we get: U 2 > U i - Figure 4.43 shows some of the results. 4,0 T 3,5 3,0 -55 2,5 & 2,0 - m Cm = 1 A Cm = 2 ::r 0,0 0,000 0,005 0,010 Ug (m/s) 0,015 0,020 Figure 4.43: Comparison of the large bubble rise velocity results at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of gas superficial velocity 123 4.2.2.5 Comparison of experimental total gas holdup results with literature results A s we can see from Figure 4.44, the calculated gas holdup using Viswanathan's [1986] equation predicts the experimental gas holdup very well. Viswanathan's equation: s = 0.5- (~T=)04 • (—)0"8- [4-1] Where U b r is the isolated bubble rising velocity (m/s), R is the pilot-scale column radius (m), U G is the inlet gas superficial velocity (m/s) We need experimental data calculated with the gas disengagement technique such as the bubble rising velocity and the gas superficial velocity to calculate the gas holdup using Viswanathan's equation. It then is more likely that these results are in agreement with the experimental gas holdup although this equation is used to describe bubble columns not pulp suspension in residence towers. 0,020 0,015 £ 0,010 0,005 0,000 0,000 • Calculated gas holdup —•—Exper imenta l gas holdup i—i 0,005 0,010 Ug (m/s) 0,015 0,020 Figure 4.44: Comparison of the experimental total gas holdup results at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 2% consistency with Viswanathan's equation as a function of gas superficial velocity 124 Using Zahradnik and Kastanek's[ 1979] equation for bubble column gas holdup: e g = ( U G ) / (A + B U G ) [4-2] where U G is the superficial gas velocity, we also find that although the values of A and B change for every consistency, this equation describes the gas holdup found experimentally as shown in Figure 4.45, well. The other pulp consistency results are shown in the annexes. 0,10 0,08 0,06 0,04 0,02 0,00 0,000 0,005 0,010 Ug (m/s) • calculated gas holdup -•— Experimental gas j holdup 0,015 0,020 Figure 4.45: Comparison of the experimental total gas holdup results at 3/4 of the pilot-scale column height using the gas disengagement technique with B S T C K pulp at 8.5% consistency with Zahradnik and Kastanek's equation 125 4.3 Mass transfer results Using the procedure done described in Chapter 3, we measured the rate at which nitrogen was replacing oxygen in the laboratory-scale column. The opposite procedure was done twice (oxygen replacing nitrogen) and the same k L a values were found, even i f the two gases switched. Mass transfer coefficients have been reported in the literature for bubble columns [Merchuk and Ben-Zvi, 1992]. k L a values of between 0.002 to 0.007 s"1 are reported for gas superficial velocities between 0.01 to 0.12 m/s [Merchuk and Ben-Zvi, 1992]. We also found similar values as shown in Figure 4.46. 0,25 0,20 0,15 CO _ l 0,10 0,05 0,00 O C m = 8%, kLa • Cm = 8%, kLa A C m = 3%, kLa • n r-i n ° W ° — g &I& -D , 0,000 0,010 0,020 0,030 Gas superficial velocity (m/s) 0,040 Figure 4.46: Mass transfer results: k L a coefficient found with Browns Stock pulp at consistencies of 3 and 8% in the small laboratory-scale column as a function of gas superficial velocity Figure 4.46 shows the k L a values found. The mass transfer has been found to occur in two steps for the experiments with 8% B S T C K pulp. First a very rapid step: k L a (start) where all the oxygen transfers to the water surrounding the fibre. This means that the probe was surrounded by a water film and the mass transfer measured was not the mass transfer occurring in the pulp suspension. Then, it reaches a second stage where the k L a value decreases and is similar to that 126 reported in bubble columns: k L a (end). Since the experiments with 3% pulp give results in the same range as the k L a (end) values, that suggests that when there is sufficient water and pulp around the probe (like with 3% or 8%(end) consistency), the probe does read the dissolved oxygen well. Furthermore, we can see that the k L a values increase with the consistency of the pulp. This can be due to the presence of more pulp or to a larger film of water around the membrane in the 8% consistency pulp than in the 3% consistency pulp. Channelling of the bubbles can also be responsible for this behaviour. Finally, we see that the k L a values are not influenced by the air flow rates for a pulp consistency o f 8%. At 3% consistency, the k L a values increased with the air flow rate. We already know, from the 3-D experiments, that gas holdup values at 8% pulp consistency are higher than with 3% consistency pulp. We find the same behaviour in the k L a values in the laboratory-scale column. N o w how are these k L a values scaled up to the values we would find in the pilot-scale column? The laboratory-scale column is small and therefore, wall effects are more important than in the pilot-scale column. That wil l influence the k L a values. But, in an industrial retention tower, the wall effect are negligible. Furthermore, high pressure, high temperature and the chemicals influence mass transfer as well. 127 5 , I N D U S T R I A L R E T E N T I O N T O W E R Industrial retention towers operate under different conditions than the pilot-scale column. Industrial towers are normally operated under pressure (700-800 kPa g at the bottom o f the tower), alkali conditions (pH 10-10.5) and high temperature (85-105°C) which are not the case for the pilot-scale column. In spite of these differences, the results found in this study can help better understand the gas hydrodynamics occurring industrial retention towers. With medium-consistency pulps, we found that the gas holdup was higher than that o f water and that the gas superficial velocity did not influence the gas holdup values. We also found that air was trapped in the fibre network and that no bubbles were formed since there is not enough free water available. This leads to the understanding that in an industrial retention tower, oxygen directly contacts the pulp fibre and that adding more oxygen wil l not necessarily improve the delignification efficiency. The fact that air gets trapped in the fibre network in the pilot-scale column suggest that, in retention towers, oxygen is in contact with the fibres. But, there is a danger o f slugging the retention tower i f too much air gets trapped into the fibre network. In a tower with a very large diameter, the chances of slugging the column are small. The mixing of the oxygen into the whole fibre network has to be done in the mixer since no movement o f the pulp was observed in the pilot-scale column. Furthermore, since both the pulp and the oxygen are in co-current flow in a residence tower, that can increase the gas holdup values found in the pilot-scale column. In fact, the air wil l get trapped in an upcoming gas and pulp flow since the moving pulp can carry the trapped oxygen and prevent the accumulation (coalescence) o f air into the fibre network. This wi l l lead to more air getting trapped with less chances of it accumulating and finding the path of least resistance and rising up in the tower. Furthermore, in our experiments, we could see that the dense phase gas holdup values (e^) were smaller than the dilute phase gas holdup values (e b). This leads to the understanding that most o f the air escapes as large "amounts" o f air (air accumulation) but that air is still present in the pulp suspension. 128 Finally, we also found that the k L a values resembled those found in bubble columns. This would explain why up to 60% delignification can be achieved by some oxygen delignification systems. The scale up o f the k L a values is hard to do. First, more experiments should be done in the pilot-scale column to see i f the k L a values and behaviour are the same as in the laboratory column. Then, calculations could be done to see of the mass transfer of oxygen in the pulp in the pilot-scale column is the same as the amount o f oxygen added into the pulp in retention towers. With these results, mass transfer in retention towers could be more thoroughly investigated 129 C O N C L U S I O N S The purpose of this thesis was to characterise the hydrodynamics o f gas flow in an oxygen delignification retention column. To do so, several experiments were performed along with a survey and an extensive literature review. From the survey and the literature review, it was found that medium-consistency pulp flow behaviour was not well characterised and understood. N o one had reported hydrodynamics experiments with medium-consistency pulp suspensions. Thus, several types o f experiments were performed. First, gas flow in a narrow rectangular tower (2-D channel) was characterised and bubble rise velocities determined. Then, optic probes were used to calculate the local bubble rise velocity and bubble size in a pilot-scale column (L /D = 6.5:1). In this column, using pressure probes, the gas holdup, superficial and rising velocities were evaluated at a gas flow rates ranging between 5 and 50 slpm. Two techniques were used: the gas disengagement technique and the density variation method. Both techniques evaluate total gas holdup values, but the gas disengagement technique does not account for air already present in the pulp fibre suspension. The density variation method accounts for this specific behaviour by calculating total gas holdup values relative to a reference density o f the pulp suspension where no air is present. Finally, mass transfer was evaluated in a small laboratory-scale column ( L / D = 4.1:1) using a dissolved oxygen probe. From these experiments, the main results found were the following. First it was found that with medium-pulp consistency, no bubbles formed in the tower because of the lack o f free water. Second, we found that, at pulp consistencies higher than 5 %, air gets trapped into the fibre network, increasing considerably the gas holdup and increasing the danger o f slugging the pilot-scale column. The slugging of the column can be defined as the following behaviour: a large amount o f gas accumulates in the pulp suspension and pushes the pulp up in the column. This behaviour was observed in the pilot-scale column. Furthermore, this gas holdup is not influenced by the air flow rate. Third, the bubble size at low consistencies is influenced by the air flow rate as well as the pulp consistency present in the column. Fourth, the type o f pulp used did not have a great influence on the gas holdup in the pilot-scale column. In fact, we saw no difference in the 130 total gas holdup values for never-dried bleached pulp and previously-dried bleached pulp. And finally, mass transfer occurs in the laboratory-scale column with k L a values in the same range as those found in bubble columns (operating with water and air). Most existing mills wil l eventually retrofit an oxygen delignification system. H o w fast this wil l happen depends on various considerations including environmental pressures, market demands, capital availability, recovery boiler and chlorine dioxide plant capacity, and new technical developments [Carter et al., 1997]. Furthermore, the flexibility and opportunity to adjust the individual stages to optimize strength and delignification and to adapt to future technology such as interstage activators wil l probably result in more two-stage installations [Carter et al., 1997]. To do so, a better understanding and control o f the process is critical. More work should be done to better characterise the mass transfer happening in the pilot-scale column. 131 N O M E N C L A T U R E Symbols A constant B constant C concentration of oxygen (mol/litre) C * saturated value of oxygen in the liquid (mol/litre) C L oxygen concentration in the liquid (mol/litre) C m consistency of the pulp C M oxygen concentration in the membrane (mol/litre) H height (m) H height of the fluid with air bubbling through it (m)(in the literature chapter) FLj height o f the fluid with air bubbling through it (m)(in the gas disengagement technique) HQ height o f the fluid without air (m)(in the literature chapter) Hj height o f the fluid after the escape o f the big bubbles (m)(in the literature chapter) H , height o f the fluid after the escape o f the big bubbles (m)(in the gas disengagement technique) H 1 3 calculated height of the fluid after the escape o f the big bubbles (m)(in the gas disengagement technique) H 2 height o f the fluid without air (m)(in the gas disengagement technique) HeL Henry's constant in the liquid H e M Henry's constant in the membrane I period where big bubbles and some o f the small bubbles escape (between (t0, H ^ a n d t t , , H,)) II period where the remaining small bubbles escape (between ( t b H J and (t2, H 2 ) ) 132 k L liquid-phase mass transfer coefficient k L a volumetric mass transfert coefficient (1/s) k ^ a , real mass transfer coefficient (1/s) k i ^ G i y theoretical mass transfer coefficient (1/s) Mil l -1 the first stage o f a two-stage system Mi l l j system 1 of two independent systems at a given mill N A mass transfer rate (l/(s*m 2) P pressure (kPa) P 0 2 partial pressure of oxygen in the liquid (Pa) R column radius (m) probe response time or time during mass transfer experiments (s) o time when the air is cut off (s) time after the escape o f the big bubbles (s) time at which all the bubbles have escaped the column (s) time zero during mass transfer experiments (s) T° temperature (°C) U velocity (m/s) U b average rising large bubble gas velocity (m/s) U b l large bubbles rising gas velocity (m/s) isolate bubble rising velocity (m/s) U b s small bubbles rising gas velocity (m/s) U g superficial velocity o f the gas (m/s) U G inlet gas superficial velocity (m/s) U G 1 large bubbles superficial velocity (m/s) ^ small bubbles superficial velocity (m/s) U p superficial velocity of the pulp (m/s) V g superficial gas velocity (m/s) x mole fraction o f solute in bulk liquid phase X; mole fraction o f solute in liquid at interface 133 Superscripts/subscripts 3/4 at three quarters of the height of the column middle at the middle of the column top at the top of the column total gas holdup apparent gas holdup dilute phase gas holdup large bubbles gas holdup dense phase gas holdup dense phase gas holdup calculated with the gas disengagement technique dense phase gas holdup in pulp dense phase residual gas holdup in pulp small bubbles gas holdup total gas holdup trapped air gas holdup small bubbles gas holdup gas void fraction calculated assuming that all the steam added as condensed into the mixer density of the pulp fibre (kg/m 3) density o f water (kg/m 3) density of the pulp suspension without air present in the pulp (kg/m 3) density of the pulp suspension with air present in the pulp (kg/m 3) Acronyms A O X absorbable organic halide B O D biochemical oxygen demand C O D chemical oxygen demand Geek letters e 6big c- gas_disengagement_tech fcdf pulp fcdf c pulp t^rapped s^mall Pfibre Pwater Preference Pexperiments 134 D O dissolved oxygen (%) delta t time difference between the two peaks from the two signal of the optic probes (s) delta s distance between the two optic probes (m) E C F elemental chlorine free H C high-consistency H W hardwood M C medium-consistency N M the value was not measured O W L oxidised white liquor Res Time residence time in the column (min) Res Time Theory calculated residence time o f the pulp in the column (min) S G W stone ground wood Stand Dev standard deviation S W softwood T C F totally chlorine free T E F totally effluent free Top Press pressure at the top o f the column (kPa) 135 REFERENCES Agarwal, S.B., Cole, B.J .W. , Genco, J . M . and W. Miller ,"Medium Consistency Oxygen Delignification Kinetics and Tower Design", Proceedings, 1997 T A P P I Pulping conference, San Francisco, Oct 19-23, T A P P I Press, Atlanta, G A , pp. 963-966. Andtbacka, S., "The Importance o f Washing in Oxygen Delignification and T C F Bleaching", Pulp and Paper Canada 99 (3): T93-T96 (1998). Balakshin, M . and I. Deineko,"The Reactivity of Individual Lignin Fragment During Wood Delignification with Oxygen in Various Reaction Medium", Inter. Symp. On Wood and Pulping Chemistry, Montreal, Q C , June 5-6, : 13-1 to 13-4 (1997). Benneker, A H , Kronberg, A . E . and K . R . Westerterp,"Influence o f Buoyancy Forces on the Flow o f Gases Through Packed Beds at Elevated Pressures", AIChE J. 44 (2): 263-269 (1998). Bennington, C. P. J., Owusu, G . andD.W. Francis,"Gas-liquid Mass Transfer in Pulp Suspension Mixing Operations", Can. J. Chem. Eng. 75: 53-61 (1997). Bennington, C . P . J , and Isabelle Pineault,"Mass Transfer in Oxygen Delignification Systems: M i l l Survey Results, Analysis and Interpretation", 1999 C P P A Tech. Sect. 84th Ann. Mtg . , Montreal Bly, M . . and M . Worden,"The Effects o f Solids Density and V o i d Fraction on the Bubble Rise Velocity in a Liquid-solid Fluidized Bed", Chem. Eng. Sci. 47 (13/14): 3281-3288 (1992). Bukur, D . B . , Patel, S. A . and J. G . Daly, "Gas Holdup and Solids Dispersion in a Three-phase Slurry Column", A I C h E J. 36 (11): 1731-1735 (1990). Burns, L . F. and R. Rice, "Circulation in Bubble Column", A I C h E J. 43 (6): 1390-1402 (1997). Carter, D . , McKenzie, D . , Johnson, T. and K . Idner. "Oxygen Delignification - a Study of Performance Parameters", 1996 T A P P I Pulping conference Bk. 1, Nashville, oct 27-31, T A P P I Press, Atlanta, G A , pp. 469-479. 136 Carter, D . N . , McKenzie, D . G . , Johnson, A . P. and K . Idner, "Performance Parameters of Oxygen Delignification", T A P P I Journal 80 (10): 111-117 (1997). Chabot, J., Lee, S. L . P., Soria, A . and H . I. De Lasa, "Interaction Between Bubbles and Fiber Optic Probes in a Bubble Column", Can. J. Chem. Eng. 70: 61-68 (1992). Daly, J. G , Patel, S. A . and D . B . Bukur, "Measurement o f Gas Holdups and Sauter Mean Diameters in Bubble Column Reactors by Dynamic Gas Disengagement Method", Chem. Eng. Sci. 47 (13/14): 3647-3654 (1992). De Swart, J. W. A . and R. Krishna, "Influence of Particles Concentration on the Hydrodynamics o f Bubble Column Slurry Reactors", Trans IChemE. 73 (part A ) : 308-313 (1995). De Lasa, H , Lee, S. L . and M . A . Bergougnou, "Bubble Measurement in Three-phase Fluidized Beds Using a U-shaped Optical Fibre", Can. J. Chem. Eng. 62: 165-169 (1984). Farag, H.I . , Ege, P .E. , Grislingas, A . and H.I . De Lasa,"Flow Patterns in a Pilot Plant-scale Turbulent Fluidized Bed Reactor: Concurrent Application o f Tracers and Fiber Optic Sensors", Can. J. Chem. Eng. 75: 851-860 (1997). Fuhrmann, A . , L i , K - L . and R. Rautonien, "Effects of E C F and T C F Bleaching Sequences on the Properties of Solftwood Kraft Pulp", JPPS 23 (10): J487-J493 (1997). Gehlawat, J .K., "Detennining liquid-side mass transfer coefficients and effective interfacial area in gas-liquid contacts", in Encyclopedia o f Fluid mMechanics, Chereminisnoff editor, vol 3, Houston, 1986, pp. 1101-1119. Godbole, S.P. and Y . T. Shah, "Design and operation o f bubble column reactor", in Encyclopedia o f Fluid Mechanics, Chereminisnoff editor, vol 3, Houston, 1986, pp. 1216-123. Guerriere, R. A . , Fayed, M E . and A . J. Matchett, "Mass Transfer in a Turbulent Bed Contacteur", Trans IChemE 73 (Part A ) : 246-252 (1995). 137 Heindel, T. J. and J. L . Monefeldt, "Observations o f the Bubble Dynamics in a Pulp Suspension Using Flash X-ray Radiography", 1997 Engineering & Papermakers conference, Nashville, oct 6-9, T A P P I Press, Atlanta, G A , pp. 1421 -1434. Helwick, J. A . , Dillon, P.O. and M . J . Mccready,"Time Varying Behavior o f Cocurrent Gas-liquid Flows in Packed Beds", Chem. Eng. Sci. 47 (13/14): 3249-3256 (1992). Hyndman, C L . and C. Guy,"Gas Phase Hydrodynamics in Bubble Columns", Trans IChemE 73 (Part A ) : 302-307(1995). Kawase, T., "Operational Experience o f Medium Consistency Oxygen Delignification at Seindai M i l l " , 1987 T A P P I International Oxygen Delignification Conference, San Diego, June 7-12, T A P P I Press, Atlanta, G A , pp. 245-251. Khare, A . S . and K . Niranjan,"Impeller-agitated Aerobic Reactor: the Influence of Tiny Bubbles on Gas Holdup and Mass Transfer in Highly Viscous Liquids", Chem. Eng. Sci. 50: 1091-1105 (1995). Kikuchi , K . I . , Ishida, K . , Enda, S. and H . Takahashi,"Gas-liquid Mass Transfer in Two and Three-phase Upflows Through a Vertical Tube", Can. J. Chem. Eng. 73: 826-832 (1995). K i m , S.D., K i m , H.S . and J. H . Han,"Axial Dispersion Characteristics in Three-phase Fluidized Beds", Chem. Eng. Sci. 47 (13/14): 3419-3426 (1992). K i m , S D . and Y , Kang,"Heat and Mass Transfer in Three-phase Fluidized-bed Reactors-an Overview", Chem. Eng. Sci. 52 (21/22): 3639-3660 (1997). Koide, K , Takazawa, A . , Komura, M . and H . Matsunaga, "Gas Holdup and Volumetric Liquid-phase Mass Transfer Coefficient in Solid-suspended Bubble Columns", J. Chem. Eng. Japan 17 (5): 459-466 (1984). Krishna, R. , De Swart, W. A . , Ellenberger, J., Martina, G . B . and C. Maretto, "Gas Holdup in Slurry Bubble Columns: Effect of Column Diameter and Slurry Concentrations", A I C h E J. 43 (2): 311-316(1997). 138 Krishna, R. and J. Ellenberger, "Gas Holdup in Bubble Column Reactors Operating in the Churn-turbulent Flow Regime", A I C h E J. 42 (9): 2627-2634 (1996). Kumar, S. B . , Moslemian, D . and M . P. Dudukovic, "Gas-holdup Measurements in Bubble Columns Using Computed Tomography", A I C h E J. 43 (6): 1414-1425 (1997). Lamine, A . S . , Col l i Serrano, M . T . and G , Wild,"Hydrodynamics and Heat Transfer in Packed Bed with Cocurrent Upflow", Chem. Eng. Sci. 47 (13/14): 3493-3500 (1992). Lee, J .K., Chun, H.S . and L . W . Shemilt,"Wall-to-liquid Mass Transfer in Liquid-solid and Gas-liquid-solid Fluidized Beds", J. Chem. Eng. Japan 30 (2): 246-252 (1997). Letzel, M . H , Schouten, J.C., Van Den Bleek, C M . and R. Krishna/'Effect o f Gas Density on Large-bubble Holdup in Bubble Column Reactors", A I C h E J. 44 (10): 2333-2336 (1998). Liebergott, N . , van Lierop, B . , Teodorescu, G . and G . J. Kubes, "Oxygen Delignification o f Kraft Pulps - Effect of Variables on Pulp Quality", 1987 T A P P I International Oxygen Delignification Conference, San Diego, June 7-12, T A P P I Press, Atlanta, G A , pp. 231-244. Lindsay, J. D . , Ghiaasiaan, S. M . and S. I. Abdel-Kahlik, "Macroscopic Flow Structure in a Bubbling Paper Pulp-Water Slurry", Ind. Eng. Chem. Res. 34. 3342-3354 (1995). Lindsay J. and Gullichsen, "Pulp Pumping and Hydraulics", Chapter 1, section V I in Pulp Bleaching-Principles and Practice. Dence, C. W . and D . W . Reeve editors, TAPPI Press, Atlanta, 868 pages, 1996. Linek, V . , Vacek, V . and P. Benes,"A Critical Review and Experimental Verification o f the Correct Use of the Dynamic Method for the Determination o f Oxygen Transfer in Aerated Agitated Vessls to Water, Electrolyte Solutions and Viscous Liquids", Chem. Eng. J. 34: 11-34 (1987). Lopes De Figueiredo, M . M . and P H . Calderbank,"The Scale-up o f Aerated Mixing Vessels for Specified Oxygen Dissolution Rates", Chem. Eng. Sci. 34: 1333-1338 (1979). 139 Luo, X . , Jiang, P. and L . S . Fan,"High Pressure Three-phase Fluidization: Hydrodynamics and Heat Transfer", A I C h E J. 43 (10): 2432-2445 (1997). Marchese, M . M . , Uribe-Salas, A . and J. A . Finch,"Measurement of Gas Holdup in a Three- Phase Concurrent Downflow Column", Chem. Eng. Sci. 47 (13/14): 3475-3482 (1992). Marquez, A . L . , Larachi, F. , Wild, G . and A . Laurent,"Mass Transfer Characteristics of Fixed Beds with Cocurrent Upflow and Downflow. A Special Reference to the Effect o f Pressure", Chem. Eng. Sci. 47 (13/14) 3485-3492 (1992). McDonough, T. J., "Oxygen Delignification" Chapter 1, section I V in Pulp Bleaching-Principles and Practice. Dence, C. W. and D . W . Reeve editors, T A P P I Press, Atlanta, 868 pages, 1996. Merchuk, J. C. and S. Ben-Zvi (Yona),"A Novel Approach to the Correlation o f Mass Transfer Rates in Bubble Columns with Non-newtonian Liquids", Chem. Eng. Sci. 47(13/14): 3517-3523 (1992). Muller, F . L . and J.F. Davidson,"On the Contribution o f Small Bubbles to Mass Transfer in Bubble Columns Containing Highly Viscous Liquids", Chem. Eng. Sci. 47 (13/14): 3525-3532 (1992). Ouellet, D . , Bennington, C. P. J. , Senger, J. J., Borisoff, J. F. and J. M Martiskainen, "Measurement o f Pulp Residence Time in a High-consistency Refiner", JPPS 22 (8): J301-J305 (1996). Pelton, R. and R. Piette, " A i r Bubble Holdup in Quiescent W o o d Pulp Suspension", Can. J. Chem. Eng. 70: 660-663 (1992). Phillips, S.,"Gas Channelling in Pulp Residence Towers", Undergraduate Thesis, Department of Chemical Engineering, Apri l 1996. Py, X . , Roizard, C , Bergault, I. and N . Midoux/ 'Physical and Chemical Mass Transfer Enhancement at a Gas-liquid Interface Due to Fine Catalyst Particles", Trans IChemE 73 (Part A ) : 253-262 (1995). 140 Reid, W . D . , Ayton, Y . andT. Mul len , "CPPA Oxygen Delignification Survey", 1998 C P P A T e c h . Sect. 84th Ann. Mtg . , Montreal, pp. A25-A30. Reilly, I. G , Scott, D . S., De Bruijn, T. J. W. and D . Maclntyre, "The Role o f Gas Phase Momentum in Determining Gas Holdup and Hydrodynamic F l o w Regimes in Bubble Column Operations", Can. J. Chem. Eng. 72: 3-12 (1994). Saberi, S., Shakourzadeh, D . , Bastoul, D . and J. Militzer, "Bubble Size and Velocity Measurement in Gas-liquids Systems: Application of Fiber Optic Technique to Pilot Plant Scale", Can. J. Chem. Eng. 73: 253-257 (1995). Schroderus, S. K . , Nguyen, P. and I. Paldy, "Adding Oxygen Delignification in an Existing Kraft M i l l " , Pulp & Paper Canada 98 (9): T294-T297 (1997). Schumpe, A . and G . Grund, "The Gas Disengagement Technique for Studying Gas Holdup Structure in Bubble Columns", Can. J. Chem. Eng. 64: 891-896 (1986). Tan, K . K . and R . B . Thorpe,"Gas Diffusion into Viscous and Non-Newtonian Liquids", Chem. Eng. Sci. 47 (13/14): 3565-3572 (1992). Tatsuishi, H . , Hatano, T., Iwai, T. and K . Kovasin, "Practical Experiences o f Medium-consistency Oxygen Delignification by Rauma-repola and Sumitomo Heavy Industries", 1987 T A P P I International Oxygen Delignification Conference, San Diego, June 7-12, T A P P I Press, Atlanta, G A , pp. 209-215. Tench, L . and S. Harper, "Oxygen Bleaching Practices and Benifits - an Overview", 1987 T A P P I International Oxygen Delignification Conference, San Diego, June 7-12, T A P P I Press, Atlanta, G A , pp. 1-8. Thimmapuram, P R . , Rao, N . S . and S C . Saxena/'Characterization o f Hydrodynamic Regimes in a Bubble Column", Chem. Eng. Sci. 47 (13/14): 3355-3362 (1992). Tsuchiya, K . and O. Nakanishi,"Gas Holdup Behavior in a Tall Bubble Column with Perforated 141 Plate Distributors", Chem. Eng. Sci. 47 (13/14) 3347-3354 (1992). Vasconcelos, J .M.T. , Nienow, A . W . , Martin, T., Alves, S.S. and C M . Mcfarlane/'Alternative Ways of Applying the Hydrogen Peroxide Steady State Method o f K^a Measurement", Trans IChemE 75 (Part A ) : 467-472 (1997). Viswathan, K . , "F low patterns in bubble columns", in Encyclopedia o f Fluid Mechanics, Chereminisnoff editor, vol 4, 1986, pp. 1180-1215. Vunjak-Novakovic, G . , Jovanovic, G , Kundakovic, L j . and B . Obradovic,"Flow Regimes and Liquid Mixing in a Draft Tube Gas-liquid-solid Fluidized Bed", Chem. Eng. Sci. 47 (13/14): 3451-3458(1992). Went, J., Jamialahmadi, M . and H . Muller-steinhagen/'Effect o f Wood Pulp Fibre Concentration on Gas Holdup in Bubble Columns", Apparatetechnik (Chem. Eng. Teck) 65: 306-308 (1992). White D . E . and R. Larsson, "Towers and Reactors", chapter 4, sectionVI in Pulp Bleaching-Principles and Practice. Dence, C. W . and D . W. Reeve editors, T A P P I Press, Atlanta, 868 pages, 1996. Zahradnik, J., Fialova, M . , Kastanek, F. , Green, K . D . and N . H . Thomas,"The Effect o f Electrolytes on Bubble Coalescence and Gas Holdup in Bubble Column Reactors", Trans IChemE 73 (Part A ) : 341-346 (1995). Zahradnik, J. and F. Kastanek,"Gas Holdup in Uniformly Aerated Bubble Column Reactors", Chem. Eng. Commun. 3: 413-429 (1979). Dence, C. W. , and D . W. Reeve eds., Pulp Bleaching-Principles and Practice. T A P P I Press, Atlanta, 868 pages, 1996.. Jolicoeur, M . , Chavarie, C , Kamen, A . and G . Andre, Procede de transfert en systemes chimiques. hiver 96, Ecole Polytechnique de Montreal, 342 pages 142 Fan, L - S . , Gas-Liquid-Solid Fluidization Engineering. Butterworths, Boston, 1989. Kunii , D . and O. Levenspiel, Fluidization Engineering, second edition, Butterworth-Heinemann, Boston, 491 pages, 1991 Perry's Chemical Engineering Handbook, sixth edition, R. H . Perry, New York, 1984 Treybal, R . E . , Mass-transfer Operations, third edition, McGraw-Hi l l , N e w York, 784 pages, 1980 143 R E F E R E N C E S F O R B U L I M I A A B R A H A M , Suzanne and D . Llewellyn-Jones, Exiting Disprders-The Facts, fourth edition, Oxford University Press, Oxford, 1997, 245 pages B U R N S , David D . , The feeling good handbook, A plume book, N e w York, 1989, 587 pages C L A U D E - P I E R R E , Peggy, The secret language ofeating disorders, Random House o f Canada, N e w York, 1997, 288 pages C O O P E R , Peter J . , Bulimia Nervosa and Binge eating-A guide to recovery, New York University Press, New York, 1993, 160 pages D I N G L E , Pauline, Fm dieting as fast as I can, Pauline Dingle, Canada, 1994, 177 pages F A I R B U R N , Christopher, Overcoming Binge Eating, The Guilford Press, N e w York , 1995, 246 pages F A L L O N , P., Katzman, M . A . and S. C. Wooley, Feminist perspectives on eating disorders, The Guilford Press, New York, 1994, 465 pages G A R D N E R , David M . and Paul E . Garfinkel, Handbook of treatment for Eating Disorders, second edition, The Guilford Press, New York, 1997, 528 pages H I R S C H M A N N , Jane R and C. H , Munter, When women stop hating their bodies, Fawcett Columbine Book, N e w York, 1995, 363 pages K A T H E R T N E , Anne, Anatomy of a food addiction, third edition, Gurze Books, C A , 1996,241 pages L E V E N K R O N , Steven, The best little girl in the world, Warner Books Edition, N e w York , 1978, 253 pages 144 P H I L L I P S , Katharine A . , The broken mirror, Oxford University Press, N e w York , 1996, 357 pages R I N Z L E R , Carol Ann, Nutrition for dummies, I D G Books Worldwide, Foster City, 1997, 410 pages SDEGEL, Michel, Brisman, J. And M . Weinshel, Surviving an eating disorder-Strategies for family and friends, revised and updated, HarperPerennial, New York, 1997, 269 pages T R A T T N E R S H E R M A N , Roberta and Ron A . Thompson, Bulimia-A guide for family and friends, Jossey-Bass Publishers, San Francisco, 1997, 155 pages W O L F Noami, The beauty Myth, Vintage Canada, Toronto, 1990, 348 pages 145 ANNEXES Annex A: Calculation exampl 147 Annex A. 1: 2-D Experiments Using the frame by frame function on the video, we can isolate one bubble and follow it's way up the screen. We can convert from apparent distance to actual distance using a scale visible in the video pictures. This scale measured 5 cm. The first thing to do was to measure it on the video screen to know the scale: 5 cm during the experiments = x cm on the screen Then we can calculate the rising velocity with: H 2 - H 1 5 Ug= •- [ A - l ] 12 — 11 x Where U g is the rising velocity, H x is point one when the bubble is at the bottom of the screen (cm), tj is time when the bubble is at the bottom of the screen, read on the video tape (s), H 2 is point one when the bubble is at the top of the screen (cm) and t 2 is time when the bubble is at the top o f the screen, read on the video tape (s). Annex A.2: Bubble size and rising velocity experiments with fibre optic probes We calculated the rising velocity and the bubble size using the two signals coming from the two optic probes. First, we need to identify the two peaks from the signal and then calculate the time delay. Then knowing the distance between the two probes, we calculate the rising velocity: As Ug = —— [A-2] 12 — 11 Where U g is the rising velocity, As is the distance between the 2 optic probes, tj is time when the peak is on the first optic probe signal and t^ is time when the peak reaches the second optic probe signal 148 Knowing the local bubble rising velocity, we can calculate the bubble size by multiplying the rising velocity with the time of the signal when a bubble passes the probe: db=Ug»{ti-t) [A-3] Where d b is the bubble size, U g is the rising velocity, t, is time when the signal start to be perturbed by the passage o f a bubble and t 2 is time when the signal returns to normal Annex A. 3: Gas disengagement technique From the pressure probe signal and graph, we get three heights and three times: H Q which is the height of the fluid with air bubbling through it H X which is the height of the fluid after the escape of the big bubbles H 2 which is the height of the fluid without air t 0 which is the time when the air is cut off t, which is the time after the escape o f the big bubbles t 2 which is the time at which all the bubbles have escaped the column Then follow the procedure: 1. Set: dH Hi-ho value to dth t\ — to [A-4] 2. Calculate: Hn = H\-t\* (dH) v dth) [A-5] Where H 1 3 is the calculated height of the fluid after the escape of the big bubbles t\. 3. Calculated: dH dH ( H . - f f B ) . ( f f i - f f 2 . - ) _ X L £ _ dtis dtu H0.15 [A-22] AJPA Where A P b is the pressure at the bottom of the column di-Uo- p g 2. Using the Reynolds number Re / = [A-23] for the total flow approaching the M distributor, select the orifice coefficient C ^ from the table given on page 105 o f Kunii and Levenspeil [1991] 3. Determine the gas velocity through the orifice, measured at the approached density p g and temperature, using: lAor — (^sdor * V p . J [A-24] The ratio UrJu^ should be less than 10% 4. Set the holes diameter and calculate the number of holes per unit area (N o r ) using: n 2 U o - — • d o r •Uor* Nor [A~25] 4 153 Annex B: 2-D Experimental results 154 The 2-D channel is made of two 1.3 cm plexiglass plates separated by 3.8 cm.. The inside of the channel has the following dimensions: 90 cm wide by 3.8 cm deep by 122 cm high. Before the description of the flow, three definitions must be stated. First, the bubble diameter is defined as d B in the test. When the bubble shape is oval, the height and the width of the bubbles are illustrated in Figure B . 1 O l Height Width Figure B . l : Shape of the bubbles in the 2-D channel Annex B.l: Description of the bubble flows in the 2-d channel Experiment n ° l : SGW pulp at C m = 4% and air flow rate of 29.6 E-6 m3/s • The bubbles are uniform, they have the same shape and the same size • d R ~ 2.5 cm. Experiment n°2: S G W pulp at Cm = 4% and air flow rate of 68.0 E-6 m3/s »• The bubbles are oval-shaped: height « 5 cm, width * 2.5 cm. »• The bubble flow is in the centre of the channel 155 • The bubbles appear and disappear like a sine wave. This is because the bubbles are not touching both walls of the column. Light is absorbed by the pulp and we can not see the bubbles at all times. • It is difficult the measure the velocity Experiment n°3: SGW pulp at C m = 4% and air flow rate of 106.4 E-6 m3/s • The bubbles are bigger than in experiment n°2 and they are oval-shaped: height » 7.5 cm, width » 3 cm. • The flow is not straight. It goes to the left or to the right. The bubbles follow preferential paths • The bubbles are easy to follow. It is easy to calculate their velocity. Experiment n°4: SGW pulp at C m = 4% and air flow rate of 144.8 E-6 m3/s »• The bubbles are bigger than in experiment n°3 and are oval-shaped: height ~ 5 cm, width ~ 10 cm. • The flow is not straight. It goes to the left or to the right. The bubbles follow preferential paths • The bubbles are easy to follow. It is easy to calculate their velocity Experiment n°5: S G W pulp at Cm = 2% and air flow rate of 29.6 E-6 m3/s • The bubbles are small and are oval-shaped: height ~ 2 cm, width * 2.5 cm. *• The flow is in the centre of the channel and the bubbles are coming up one after another in a row • We can see through the pulp. • There is a little coalescence and splitting present in the bubble behaviour Experiment n°6: SGW pulp at Cm = 2% and air flow rate of 68.0 E-6 m3/s • The bubbles are a little bigger than in experiment n°5 and are oval-shaped: height « 2.5 cm, width « 5 cm. 156 • We could see internal circulation in the bubbles »• The flow of bubbles is in the centre of the channel • The bubbles coalesce more than in experiment n°5 Experiment n°7: S G W pulp at Cm = 2% and air flow rate of 106.4 E-6 m3/s • The bubbles are a little bigger than in experiment n°6 and are oval-shaped: height ~ 5 cm, width - 7 . 5 cm. *• We could see internal circulation in the bubbles *• The flow of bubbles is in the centre of the channel *• There is a lot o f coalescence and splitting o f the bubbles Experiment n°8: S G W pulp at Cm = 1% and air flow rate of 29.6 E-6 nrVs *• The bubbles are really small and uniform: d B ~ 1.5 cm • The bubbles are far apart from one another »• The pulp behaves like a liquid »• The flow of bubbles is in the centre of the channel • There is a little coalescence o f the bubbles Experiment n°9: S G W pulp at Cm = 1% and air flow rate of 68.0 E-6 nrVs »• The bubbles are not uniform and are oval-shaped: height ~ 1.75 cm, width - 2 . 5 cm. »• The bubbles are far apart from one another *• The pulp behaves like a liquid • The flow o f bubbles is in the centre o f the channel • There is very little coalescence of the bubbles Experiment n°10: S G W pulp at Cm = 1% and air flow rate of 106.4 E-6 nrVs »• The bubbles are uniform: d B ~ 2.5 cm »• The bubbles are far apart from one another 157 *• The pulp behaves like a liquid *• The flow o f bubbles is in the centre o f the channel • There is very little coalescence o f the bubbles Experiment n ° l l : B S T C K pulp at Cm = 2% and air flow rate of 29.6 E-6 m3/s • The bubbles are small and uniform: d B ~ 1.75 cm • The bubbles are hard to see. • The pulp is very dark as it adsorbs the light • The flow of bubbles is in the centre o f the channel Experiment n°12: B S T C K pulp at Cm = 2% and air flow rate of 68.0 E-6 m3/s *• The bubbles are uniform: d B ~ 2.5 cm *• A s the bubbles rise up in the channel, their path has the shape o f a sine wave »• The flow is not always in the centre *• There are sometimes two parallel bubble flows that rise up Experiment n°13: B S T C K pulp at Cm = 2% and air flow rate of 106.4 E-6 m3/s *• The bubbles are uniform: d B ~ 3 cm • As they rise, some bubbles grow longer (due to coalescence) • A s the bubbles rise up in the channel, their path has the shape of a sine wave • The flow is not always in the centre • There are sometimes two parallel bubble flows that rise up Experiment n°14: B S T C K pulp at Cm = 2% and air flow rate of 144.8 E-6 m3/s • The bubbles are uniform: d B ~ 5 cm - A s they rise, some bubbles grow longer (due to coalescence) *• As the bubbles rise up in the channel, their path has the shape o f a sine wave *• The flow is not always in the centre • There are sometimes two parallel bubble flows that rise up 158 We also did some experiments with some BSTCK pulp at 7% consistency. But at any air flow rates, we could not see any bubbles. The pulp was not moving and we suspect that the air was trapped in the pulp, that some of it escaped along the plexiglass plates (the walls) and that most of the air channelled through the suspension. Table B.1: Rising velocity in the 2-D channel with SGW and BSTCK pulp at Cm = 1,2 and 4%, at air flow rate from 29.6 to 144.8 E-6 m3/s Experiment n° Type of wood C m (%) Air flow (m3/s) Rising velocity (m/s) Average (m/s) 1 Stone Ground Wood 4 29.6E-6 0,563 1 Stone Ground Wood 4 29.6E-6 0,566 1 Stone Ground Wood 4 29.6E-6 0,644 0,591 2 Stone Ground Wood 4 68.0E-6 1,149 2 Stone Ground Wood 4 68.0E-6 1,090 2 Stone Ground Wood 4 68.0E-6 1,417 1,218 3 Stone Ground Wood 4 106.4E-6 0,654 3 Stone Ground Wood 4 106.4E-6 0,812 3 Stone Ground Wood 4 106.4E-6 0,794 3 Stone Ground Wood 4 106.4E-6 0,833 0,773 4 Stone Ground Wood 4 144.8E-6 0,558 4 Stone Ground Wood 4 144.8E-6 0,654 4 Stone Ground Wood 4 144.8E-6 0,491 4 Stone Ground Wood 4 144.8E-6 0,460 0,541 5 Stone Ground Wood 2 29.6E-6 0,380 5 Stone Ground Wood 2 29.6E-6 0,286 5 Stone Ground Wood 2 29.6E-6 0,475 5 Stone Ground Wood 2 29.6E-6 0,304 0,361 6 Stone Ground Wood 2 68.0E-6 0,360 6 Stone Ground Wood 2 68.0E-6 0,344 6 Stone Ground Wood 2 68.0E-6 0,353 0,353 7 Stone Ground Wood 2 106.4E-6 0,484 7 Stone Ground Wood 2 106.4E-6 0,563 7 Stone Ground Wood 2 106.4E-6 0,442 0,496 8 Stone Ground Wood 1 29.6E-6 0,572 8 Stone Ground Wood 1 29.6E-6 0,465 8 Stone Ground Wood 1 29.6E-6 0,538 0,525 9 Stone Ground Wood 1 68.0E-6 0,759 9 Stone Ground Wood 1 68.0E-6 0,591 9 Stone Ground Wood 1 68.0E-6 0,847 0,732 10 Stone Ground Wood 1 106.4E-6 1,006 10 Stone Ground Wood 1 106.4E-6 0,735 10 Stone Ground Wood 1 106.4E-6 0,772 0,838 11 Brown Stock 2 29.6E-6 0,794 11 Brown Stock 2 29.6E-6 0,431 11 Brown Stock 2 29.6E-6 0,389 0,538 12 Brown Stock 2 68.0E-6 0,640 12 Brown Stock 2 68.0E-6 0,476 12 Brown Stock 2 68.0E-6 0,543 0,553 13 Brown Stock 2 106.4E-6 0,586 13 Brown Stock 2 106.4E-6 0,501 13 Brown Stock 2 106.4E-6 0,681 0,589 14 Brown Stock 2 144.8E-6 0,500 14 Brown Stock 2 144,8E-6 0,569 14 Brown Stock 2 144.8E-6 0,661 0,577 160 Annex C: Fibre optic probes results 161 E Q. CO 00 (0 0) to o (0 a o o > CD c to 'CI "C c (0 a> N "5> a> .a (0 e o W 3 O (0 O . N C CO ° o o T - o a i i n s i - f f l f l o m Q N o o t o o i i D ^ o~ o" o" CM" CM" CO" CM" O " O " O " h~" co" r--" co" h-" co" o" CD CO to CM m s O CM h~ CO m ™_ O) h- T - lO CD CM o> o o m o" o" o" T-" T-" CM" ~^ o" o" o CM" CM" • * " co" co" co" o" CO p e w m co CM_ r-_ m g o> g! o o> co co t o - co" io" co" in r-T co" o" ,*r m - •«-" •«*" co" co" ->-" •*" c o ( N C M i - i - i - i - C M C N RA T - T - r- T - CM CM temps start s - ^ O CO. m_ CM. CNI « . S £ CM CO £ - - * « * « m" <° « - $2 £ £ 2 ^" co" oo" °- £ « a Rising velocity m/s 0,008475 0,014815 0,02381 0,01626 0,014815 0,014388 0,016 0,006289 0,011173 0,029412 0,030769 0,013986 0,018349 0,008511 0,025974 0,021739 0,007547 Delta t s ( D m ^ n m o i m n a o a i o n a i i n N C M i f l cocococMcococMT-h~cocO'srocoh-ajcD CM" T-" o" •<-" T-" •«-" T-" co" * - " o" o" T-~ T~" CM" O " O " CM" CM ea 1 " 7,87 27,82 25,91 12,88 14,9 16,6 18,51 22,68 3,72 5,42 6,9 9,76 11,79 14,76 16,15 19,36 26,97 temps 1 s - M". O . CD. in CM. CM. « g £ £ $ _ • - * CO 00 m" $ * - «2 «» £ ? - " * co co" °- 2 * » ctj distance m C M C M C M C M C M C M C M C M C M C M C M C M C M C M C M C M C M O O O O O O O O O O O O O O O O O o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" Pulp o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o c c e c e c c e c c c e e e e e c : J J J J J J J J J J ? S J J J o o o o o o o o o o o o o o o o o mf f i fn fn fn fn f f i f f in r t fn f f i i t imfn inmi in 162 9> c .N C CO ° r~- h-_ O Tf_ O) in C O co ^ ^ ^ * C5 «0 CO (N OO S ^- N O) in N C N CO" T - " C N tN CD •*" T -~ O T - " o" O" O" CO C N C O T - " * - " T - " C N • * " CD ^ T -~ T - " C N O T - " P & E CN co in co cn o i m m ^ T - i - i n M O W i - T - n io T - io CM CD co o o m c o o c o c N c o i - c n i o o c o o ^ n oo CN co T» - CN TtCOS r . . i - r tCO^ , 0)«)^- l f ) lOT -OCOrtCOCOCOfO)NO^intD^N^<0' « N j o m" t o " CM" in o" T - " CM" CO" m" h~" m" t ; CD" o" to" v " ^" S S T - " OO" CM" ^ " " ^ " T - T - T - C N I C N C N C M C M C N C N ^ C N ^ T - T - C M C N ^ ^ C N T - T - ^ -temps start s ( O r t O i i n ^ f l o o w n ^ ® N N C O t - C \ | . n ( n C O C \ | N * ( O i n C \ | N ? ( N " in ^ " l s . n . °- - °- " l °. °l <\ 5°- §5 CN •<- O co 1- o in •*" co" co" m" rC T~r M 1 0 ° r\j " f - T - " o> co" m" co" o" co" o" CM" ^ " " w ^ T - T - T - C M T N , C M C M C M C M C N C N C M 0 r ^ T - T - - C M C N C N C N ' « - ' « -Rising velocity m/s co co cocNcocNcocDcocNooc»T-o5oocoa5-v,r- co - T - a co * Tj- CO m c O C M 0 0 m i - C O ' » - ' * C M i - T - c O C M 0 O f 2 r ^ r ; cs P n s o o 00C0^inCDC3)CNOC0CNinOT-T-CDCNCT)O^inSinc»V^CN'4-C0C0 c o c o ^ . i n c N i c D T - T - o i n c o o i c o T - M n c o i n S f S o c i S r o a c o M i n o o n i n M n c M t o ^ T - i - T - ^ o i N o S o S r . - n S T - o o o T - _ T-_ o_ o_ o_ o_ o_ q o o o T - o o o o a - o °- ° o °. o o o o o" o" o" o" o" o" o" o" o" O " O " O " O " O O " O " O " ° o" ° o" ° o" o" o" o" Delta t s c o i n N ( o c o c o o ) w c » c » « i n ^ c o c > i T - c o n c o s S * < ' > l ' > * ' i - u ) c o T - _ T - *X co co_ CN co a> T - ^ o c\)_ T - ^ N N c» ^ n a^^m — ^ ^ r ^ T - T f o ^ r ^ T - c D i n ^ i n c o — Tj-cocor^-coT-oicocNoo CM C O CD 5 ? °°- T~- w . " l "* <°- 'P. - CD ^ ^ OO CO in O O C O CN T - T -_,: -,c ^ 1- • * " CD" O " T - " CM" •<* CD" r~" JQ r>-" o>" o" in to" co" co" co" T - " OO" CD" •«-" ^""^^T-T-T-CNCNCMCMCMCM t V ICMCM , t -<-T-T-CMCMCMCM-<-^'i-temps 1 s COcoa>inT -^ r o t D w co^tDCMI^00rrCN m m COCMr~'^-CO>OCNr--•<*" co" co" m" rC T " 0 0 1 0 ° r: CM" CO" m" r»" •*" r^-" a>" „ - - T - " a>" co" ui co" o" co" o" CM" distance m CNCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCMCslCMCMCMCMCMCM o_ O O o o o o o_ o_ o_ O O Q O O O O O O O O O O O O O O O o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" 5 as o ^ Pulp o o o o o o o o o o o o o o o o o o o o o o o a o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o c c c c z c c c c c c c c c c z c c c c c c z c c c r c c c c c 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 E [Q [D IQ (D| ]Q I ]Q|]Q[Q| I I B l | ]Q| I l | I l IQ (Q{Q| I ]| I ) |QQ(Q I||{QQa] [n i ]QQ 163 E Q. W IO CN CO o> c "5> ' E "D C (0 CD N n 3 - Q »*-o O) c "__ X J c co CN o> cz> lO CO T-_ h~ in o\ ° . CO C O O I O CD C O CN o r~ L O I O C O o o" O CN C N co" m" in T-" T-" LO" •>ar o" •*" co" J_ of o" co" o" T-~ O T-" T-" O " T-" O " T-" *-" y EE CO 0,0035577 0,0019643 0,0111111 0,0279245 n />o A n o 0,0286486 0,0319643 0,059 0,05 0,0146154 0,0125352 0,0508861 0,0466667 0,0045327 0,0486957 0,0336842 0,11 0,0963158 0,002987 0,0304 0,0046154 0,016 0,008 0,0121429 0,01 0,0065 0,0154054 0,0046154 0,0130435 0,01 de/ra f s co T- ™. r~- o i*- r» o r-_ eo q r - o in » c» oo N n r - r - r . r - T- T- m o ^ o" o" ° o" T-" --" --" --" o" o" CM" •<-" o" o" CN" o" --" o" o" o" o" ° o" o" o" o" o" o" o" CO P C CO T- -J- *R. ° l CD O . - -j- in CN - *H °°- "H ( D « i O < O t _ ( N I > - T - * ( \ | * " * O) ™ J *~ » c\| " 2 £ £ ? ^ CN" <° co" of ° £ £ » °_ <° o" -" CM" CN" co" *•" - f 1 0 co" r-" r-~" Rising velocity m/s m s s ( o s s M O ) T - i o i o r - - n N i D i - ( N N C N C O O > C N N C M S N T - L O c o c o c M L O c o - - c o o o T - c D r ^ i n o > - - c o c o coco C N C O C N O O L O C O C D C O O I ^ O C O C O C D C N O C O C O C O C » C O ' « - C D O > 0 0 - * C O C O - * C D L O O C N O ) C D 0 > N N N S N ( 0 N C B ^ i n i - ^ ( D ^ i - C M ( M O 0 0 n O T - C M O N T - ( D ( 0 o ^ ( o n ( N ^ r t T f r - T - ( M T f o « T - i - i n T - n ' o n n - N i n n - t N i n o o i c o o o o o o o o o o o o o o o - - o o < - ' o - - c - ' o o a o o o o o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" Delta t s o o c N ^ c o T r c M , - c N T r c N o i c o c o c O T - o o < x > T f O T - I O I O N - " . * O ^ S ^ ( M C M S T - ( O i n ( M i O T - N N n " N W C M -CM" I-~ o" o" o" -<-" ° o" T-" r-" o" o" -f" O T-" o" o" T-" o" o" o" o" o" o" ° o" o" o" ° CM CO ^ O ™ " f L O OO O CM C O ro" J £ £ £ ° « o" ^" C N " CN" co" ^ LO" co" c o ^ temps 1 s C O C O - _jT ~ ~ - - CD CN C O jrT - - - - CN CN ^J" L O ~t" O CD - C O O C O co" of £ 2 J £ °2 ™ CM" co" of ° £ £ ~ °2 <° o" -^" CM" CN" co" ^r" 1 0 co" i-" i-" distance m CMCMiNCMCMNCMCMCMCMCMCMCMCMCMCMCMNCMCMCMCMCMCMCMCMCMCMCM o o o o o_ o o o_ o o o o o o o o o o o o o o o o o o o o o o" o" o" o" o" o" o" o" o" o o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" o" £ a? o ^ Pulp o o o o o o o o o o o o o o o o o o 2 2 2 2 2 2 2 2 2 ( - > c ' 0 0 0 ' : ' 0 ( - > ( - > ^ ^ ^ ^ ^ ^ - ' - ^ ^ - ^ ^ -^^U)^ui^^a)^^^oiaiai^a)(oo)2222222222B 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 cacDcacQcacQCQCDCDCDCDaiaiaicDcacDCD 165 Figure C l : Signal from fibre optic probe 6 with 1% B S T C K pulp and a gas superficial velocity of 0.0058 m/s 166 ure C.2: Signal from fibre optic probe 7 with 1% BSTCK pulp and a gas superficial velocity of 0.0058 m/s 167 AnnexD: Gas disengagement technique results 168 Annex D. 1: B STCK pulp results Annex D. 1.1: Total gas holdup 170 Figure D . l : To ta l gas holdup as a function of gas superficial velocity for water in the pilot scale column 171 Figure D .2: Total gas holdup as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 172 Figure D .3 : Total gas holdup as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 173 Figure D .4 : Total gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 174 ure D.5: Total gas holdup as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 175 ure D.6: Total gas holdup as a function of gas superficial velocity 10% consistency BSTCK pulp in the pilot scale column 176 Figure D.7: Comparison of the total gas holdup at the middle of the height of the column using the gas disengagement technique with B S T C K pulp at 8.5 and 10% consistency as a function of the gas superficial velocity 177 Figure D.8: Comparison of the total gas holdup at 3/4 of the height of the column using the gas disengagement technique with B S T C K pulp at 0, 1, 2 and 5% consistency as a function of the gas superficial velocity 178 LO o o CN IO 00 II II II II II II E E E E E E O i O O O O O i I m t o CN O LO O o CO LO o o o o o o LO o LO o LO o LO o CO CO CM CN o o o~ o~ o~ o" o" co o" o" o" CO Figure D.9: Comparison of the total gas holdup at 3/4 of the height of the column using the gas disengagement technique with B S T C K pulp at 0, 1, 2, 5, 8.5 and 10% consistency as a function of the gas superficial velocity 179 Figure D.10: Comparison of the total gas holdup at the top of the height of the column using the gas disengagement technique with B S T C K pulp at 0,1, 2, 5, 8.5 and 10% consistency as a function of the gas superficial velocity Annex D. 1.2: Dilute phase and lar bubble gas holdup 181 Figure D . l l : Dilute phase and large bubble gas holdup as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 182 ure D.12: Dilute phase and large bubble gas holdup as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 183 ure D.13: Dilute phase and large bubble gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 184 Figure D.14: Dilute phase and large bubble gas holdup as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 185 ure D.15: Dilute phase and large bubble gas holdup as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 186 o LO CO o CO — I — LO CN LO o CO T -II II E E O O i o CN o~ .O CO — r -LO o" o CN O LO O II O D) ZD LO o o o CO LO o o o o o o" o Figure D.16: Comparison of the dilute phase and large bubble gas holdup at the middle of the height of the column using the gas disengagement technique with B S T C K pulp at 8.5 and 10% consistency as a function of the gas superficial velocity 187 T— CN LO II II II E E E O O O m < x i — H — i I K - r r H I I h I - X - I \ \ ¥r \«m 1-O CN O LO O CO o CO ZD LO o o o" o o o o o o o o o o o o o 00 CD CN o 00 CD CN o T — T — T — o o o o o o o o o o o o o o o o" o" CO o~ o~ o" CD o" o~ CD CO ure D.17: Comparison of the dilute phase and large bubble gas holdup at 3/4 of the height of the column using the gas disengagement technique with B S T C K pulp at 1, 2 and 5% consistency as a function of the gas superficial velocity 188 LO O CN LO 00 -^II II II II II E E E E E O O O O ! O t f x I f i \ ure D.18: Comparison of the dilute phase and large bubble gas holdup at 3/4 of the height of the column using the gas disengagement technique with B S T C K pulp at 1,2,5,8.5 and 10% consistency as a function of the gas superficial velocity 189 ure D.19: Comparison of the dilute phase and large bubble gas holdup at the top of the height of the column using the gas disengagement technique with B S T C K pulp at 1, 2, 5, 8.5 and 10% consistency as a function of the gas superficial velocity 190 Annex D. 1.3: Dense phase gas holdup 191 ure D.20: Dense phase gas holdup as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 192 Figure D.21: Dense phase gas holdup as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 193 ure D.22: Dense phase gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 194 Figure D.23: Dense phase gas holdup as a function of gas superficial velocity 8.5% consistency BSTCK pulp in the pilot scale column 195 ;ure D.24: Dense phase gas holdup as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 196 ure D.25: Comparison of the dense phase gas holdup at the middle of the height of the column using the gas disengagement technique with B S T C K pulp at 8.5 and 10% consistency as a function of the gas superficial velocity 197 T - CM LO II II II E E E O O O < X =^«-H*_H 1®: I—M 1 l- : LO o o c f o CD O O LO o o c f o C O o o c f LO o o o o o o c f o o o c f (S/LU) son ure D.45: Comparison of the small bubble superficial velocity at the top of the height of the column using the gas disengagement technique with B S T C K pulp at 1 and 2 % consistency as a function of the gas superficial velocity 220 AnnexD.1.7: Large bubble rise velocity 221 Figure D.46: Large bubble rise velocity as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 222 Figure D.47: Large bubble rise velocity as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 223 H < 1 I—«SiH o LO co o co" LO c\T o CN LO —r~ LO O CN O LO O o" o CO Z> LO o o o o o o o~ o" (S/W) iqfl Figure D.48: Comparison of the large bubble rise velocity at 3/4 of the height of the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of the gas superficial velocity 224 LO CO —r— o co" LO CM" -<-H m «?r-LO — I — O o CM O LO T— o II O D> LO o o o" CM LO o" o o o o o" o" (S/LU) iqfl Figure D.49: Comparison of the large bubble rise velocity at the top of the height of the column using the gas disengagement technique with B S T C K pulp at 1 and 2% consistency as a function of the gas superficial velocity 225 Annex D. 1.8: Small bubble rise velocity 226 Figure D.50: Small bubble rise velocity as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 227 Figure D.51: Small bubble rise velocity as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 228 LO CN —r -O CN - o t CD Q . LOCO O O CO o o o o o o o o o o o o o CO CN o 00 CD CN o T— T— o o o o o o o o o o o o o o c f c f c f c f o" c f c f c f c f CO ure D.64: Comparison of the total gas holdup as a function of gas superficial velocity for a 2% consistency B S T C K pulp, never-dried and previously-dried bleached pulp at 3/4 of the height of the pilot scale column 244 o o CM O c f O CO o c f o CM O c f —r~ O CO o o c f o o CM o o LO o c f o w o LO o o o o o o o o o § <=>• o o CO Figure D.65: Comparison of the total gas holdup as a function of gas superficial velocity for a 2% consistency B S T C K pulp, never-dried and previously-dried bleached pulp at the top of the height of the pilot scale column 245 Annex E: Density variation method results Annex E.l: Total gas holdup before the experiments 247 CM O O c f CO co O O o o o o CO o o CD c f c f c f c f o LO c f CO o c f o CO o o CM T -c f c f o o o o c f o Figure E . l : Total gas holdup before the experiments as a function of gas superficial velocity for water in the pilot scale column 248 :ion CO var - Density method o CM O LO o el O O) ZD LO o o o CO o o c f LO o o o co o o c f CO LO o o o o o o c f o o o c f Figure £ .2: Total gas holdup before the experiments as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 249 CO CO 3 ure E.3 : Total gas holdup before the experiments as a function of gas superficial velocity 2 % consistency B S T C K pulp in the pilot scale column 250 Figure £ . 4 : Total gas holdup before the experiments as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 251 o CN t t t • Density variatic method 00 0,005 0,010 0,015 0,0 Ug (m/s) o o oi oo h~ CD LO o CNJ T— T— T— T— T— o o o o o o c f c f c f c f c f c f to Figure E.5: Total gas holdup before the experiments as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 252 t o CM O LO O CO o CD CO 1 LO o o LO o CD LO o CO CO o CM O c f CO oo o cf a> o o o o o o cf o o cf Figure E.6: Total gas holdup before the experiments as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column 253 ® it f <1 i <3 LO o o ^ — CN LO cb T— II II II II II II E E E E E E O i O O i O i O O i I E5i i f X X 1 1 I I ¥ i i i i i X X X I I I X x X X X X X X o CN o co LO v -O cf o o> ZD 0 LO o o c f o o o o LO o LO o LO o LO o LO o o LO NT CO oo CN CN o o o o O o o o o o o o o cf cf cf cf cf cf cf cf cf cf cf CO ure E.7: Comparison of the total gas holdup before the experiments between 3/4 and the top of the height of the column with B S T C K pulp at 0,1, 2, 5, 8.5 and 10% consistency as a function of the gas superficial velocity 254 Annex E.2: Total gas holdup during the experiments and comparison with the gas disengagement technique total gas holdup results 255 Figure E.8: Total gas holdup dur ing the experiments as a function of gas superficial velocity for water in the pilot scale column 256 Figure E .9: Total gas holdup during the experiments as a function of gas superficial velocity for a 1 % consistency B S T C K pulp in the pilot scale column 257 o o o LO o LO o LO o LO o o CO CO CM CM T— T— o o o o o O o o o o cf cf cf cf cf cf cf cf CO Figure £.10: Total gas holdup during the experiments as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column ure E . l l : Total gas holdup during the experiments as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 259 lire E.12: Total gas holdup during the experiments as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column 260 ure E.13: Total gas holdup during the experiments as a function of gas superficial velocity 10% consistency BSTCK pulp in the pilot scale column 261 Figure E.14: Comparison of the total gas holdup during the experiments between 3/4 and the top of the height of the column with BSTCK pulp at 0,1,2, 5, 8.5 and 10% consistency as a function of the gas superficial velocity Annex E.3: 262 Total gas holdup after the experiments 263 o CM O c f LO o cf o cf LO o o o o O o o o CD CO CO v-~ cf cf ef cf o LO c f CO o -NT o" o CO o CM o c f o o o o cf o ure E .15: Total gas holdup after the experiments as a function of gas superficial velocity for water in the pilot scale column 264 rz o -4—< to > CD (D Q E o CN O CO o c f I t O CD 3 LO o o o o o o o o o o o o o o o LO •sr CO CN X — o CN CO o o o O o o o o o o o o o o o o o o c f c f c f c f c f c f c f 1 c f c f CO Figure E.16: Total gas holdup after the experiments as a function of gas superficial velocity for a 1% consistency BSTCK pulp in the pilot scale column 265 Figure E.17: Total gas holdup after the experiments as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 266 Figure £.18: Total gas holdup after the experiments as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column 267 Figure £.19: Total gas holdup after the experiments as a function of gas superficial velocity 8.5% consistency BSTCK pulp in the pilot scale column 268 o CM O O o CO CO LO o o c f o CO o" LO c f o CO c f CO LO c f o o o o c f o c f ure £.20: Total gas holdup after the experiments as a function of gas superficial velocity 10% consistency BSTCK pulp in the pilot scale column 2 6 9 Figure E.21: Comparison of the total gas holdup after the experiments between 3/4 and the top of the height of the column with BSTCK pulp at 0,1, 2, 5, 8.5 and 10% consistency as a function of the gas superficial velocity 270 Annex E.4: Difference between before and after the experiments 271 o CN O • LO O cf o -52 o £ O CO 3 LO o o o o o CD O 00 o o CO o LO o o CO o CM o o o cf cf cf cf cf cf cf cf cf cf o o o c f 3 in eouajaj-HQ Figure E.22: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity water i n the pilot scale column 272 o CM O to o CO § t O CO ZD LO o o o o o o o o o o o o o o o CO h- CO LO CO CM o o o o o o o o o o o o o o o o o o o c f c f c f c f c f c f c f c f c f 3 in eouaiawQ Figure E.23: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity for a 1% consistency B S T C K pulp in the pilot scale column 273 o CN O LO O o " It O CO ZD LO o o o o o LO o LO o LO o LO o LO o o CO CO CM CN -— -— o o o o o o o o o o o o cf cf cf cf cf cf cf cf cf cf 3 U| 90U9J9J4IQ ure E.24: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity 2% consistency B S T C K pulp in the pilot scale column 274 o CN O c f LO O c f I t O) ZD LO o o o o o o o o o o o o o o o o o o o o o o o o o o o o o CM CO 00 o CM CO 00 o o O o o o CN c f c f c f c f c f c f c f c f c f c f o" 3 in eouaiaj^ Q Figure E.25: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity 5% consistency BSTCK pulp in the pilot scale column 275 o CN O LO O it O CO LO o o o o cf LO o cf o I LO o I o CN cf I LO CN cf o o o o cf CO 3 uj aou9J9jj!Q Figure E.26: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity 8.5% consistency BSTCK pulp in the pilot scale column 276 o CN O LO O cf O ^ 5 £ O CO ZD to o o o o o cf o CN cf o CO cf o cf o to o o o o c f CO cf 3 U| 80U9J9J4Q Figure E.27: Difference in the total gas holdup before and after the experiments (before-after) as a function of gas superficial velocity 1 0 % consistency B S T C K pulp in the pilot scale column 277 o cf KJ m K) H KH o o LO o o CN LO CO II II II II II II E E E E E E O i O O O O O i I 3^ f 1 f i t / ; X |V/| I w • l - * H I X- 1 I - X - H X H X i — X — i X o CN O i o CO cf I —r~ o cf 3 U| 90U9J9J4IQ O CN o cf LO o cf it o CJ) ZD LO o o o LO o o o o cf CO cf 1 gure E.28: Comparison of the difference in the total gas holdup before and after the experiments (before-after) measured between 3/4 and the top of the column with B S T C K pulp at 0,1,2,5,8.5 and 10% consistency as a function of the gas superficial velocity Annex F: Comparison of the experimental total gas h o l d u p w i t h Viswanathan's equation 279 Figure F . l : Comparison of the total gas holdup as a function of gas superficial velocity for water in the pilot scale column using the gas disengagement technique and with Viswanathan's equation 280 ure F.2: Comparison of the total gas holdup as a function of gas superficial velocity for a 1 % consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Viswanathan's equation 281 Figure F.3: Comparison of the total gas holdup as a function of gas superficial velocity 2 % consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Viswanathan's equation Annex G: Comparison of the experimental total gas holdup with Zahradnik and Kastanek's equation Table G . l : Constant A and B values for Zahradnik and Kastanek's equation B S T C K pulp consistency 0 1 2 5 8.5 10 A 0.3 0.6 0.6 2 0.3 0.3 B 0 20 15 1 20 20 284 F i g u r e d : Comparison of the total gas holdup as a function of gas superficial velocity with water in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 285 Figure G.2: Comparison of the total gas holdup as a function of gas superficial velocity for a 1 % consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 286 Figure G.3: Comparison of the total gas holdup as a function of gas superficial velocity 2 % consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 287 Figure G.4: Comparison of the total gas holdup as a function of gas superficial velocity 5% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 288 ;ure G.5: Comparison of the total gas holdup as a function of gas superficial velocity 8.5% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation 289 ure G.6: Comparison of the total gas holdup as a function of gas superficial velocity 10% consistency B S T C K pulp in the pilot scale column using the gas disengagement technique and with Zahradnik and Kastanek's equation Annex H: Mass transfer results Cm % Superficial velocity m/s k La (start) 1/s k La (end) 1/s 8 0,010 0,20 0,0505 8 0,010 0,20 0,0869 8 0,013 0,0116 8 0,013 0,21 0,0438 8 0,013 0,18 0,0617 8 0,013 0,0105 8 0,021 0,0026 8 0,021 0,0019 8 0,023 0,19 8 0,023 0,22 0,059 8 0,023 0,23 0,0765 8 0,029 0,0278 8 0,029 0,22 0,0373 3 0,011 0,0048 3 0,011 0,0044 3 0,011 0,0044 3 0,013 0,0063 3 0,013 0,0073 3 0,013 0,0069 Cm % Superficial velocity m/s kLa (start) Average kLa (start) Stand dev kLa(end) Average kLa (end) Stand dev 8 0,010 0,200 0,0015 0,0687 0,0182 8 0,016 0,193 0,0125 0,0319 0,021793 8 0,027 0,0023 0,0004 8 0,030 0,214 0,0149 0,0678 0,0087 8 0,037 0,218 0 0,0326 0,0048 3 0,011 0,0045 0,00019 3 0,013 0,0068 0,00041 292 Annex I: F i b r e l e n g t h measurements 293 K UJ£>. °o-E E CO c a? LL. (%) W6|9M Figure 1.1: BSTCK pulp fibre length distribution 2 9 4 CD CO CD CO CM T -**** A * O E E, __: o> c cu __ -Q (%) W6|a/\A Figure 1.2: Never-dried bleached pulp fibre length distribution 295 rie CL "O , Q-CO CD o x: revii eac CL ~Q • CO LO CO CM (%) W B ! 9 M a o o E E, sz O) c CU Figure 1.3: Previously-dried bleached pulp fibre length distr ibution Kajaani FS-200 results for BSTCK pulp 296 Sample report Sample title: Sampled: Analysed: Upper limit: Lower limit: Length: 0.20 mm B S T C K 16-03-99 21:15 7.20 mm 0.00 mm P = 33.22% Arithmetic A v : L weighted: W weighted: Total fibres: W = 2.04% 1.08 mm 2.27 mm 2.80 mm 17853 Weighted distributi Length on Fibres W G H T D % W C U M L % 0.00 - 0.20 5930 2.04 2.04 0.20 - 0.40 ?565 2.23 4.26 0.40 - 0.60 ?161 2.91 7.17 0.60 - 0.80 965 3.43 10.60 0.80-1.00 784 3.58 14.18 1.00-1.20 702 3.94 18.12 1.20- 1.40 656 4.37 22.49 1.40- 1.60 691 5.31 27.80 1.60-1.80 709 6.19 33.99 1.80-2.00 601 5.88 39.87 2.00 - 2.20 580 6.25 46.11 2.20 - 2.40 632 7.48 53.59 2.40 - 2.60 546 7.05 60.64 2.60 - 2.80 462 6.42 67.07 2.80 - 3.00 431 6.44 73.50 297 Weighted distributi Length on Fibres W G H T D % W C U M L % 3.00-3.20 342 5.46 78.97 3.20-3.40 328 5.58 84.55 3.40-3.60 244 4.40 88.95 3.60-3.80 178 3.40 92.35 3.80-4.00 125 2.51 94.87 4.00 - 4.20 72 1.52 96.39 4.20 - 4.40 55 1.22 97.61 4.40 - 4.60 36 0.84 98.44 4.60 - 4.80 17 0.41 98.86 4.80 - 5.00 13 0.33 99.18 5.00 - 5.20 6 0.16 99.34 5.20 - 5.40 4 0.11 99.45 5.40 - 5.60 4 0.11 99.56 5.60-5.80 3 0.09 99.65 5.80 - 6.00 5 0.15 99.80 6.00 - 6.20 0 0.00 99.80 6.20 - 6.40 4 0.13 99.93 6.40 - 6.60 2 0.07 100.00 6.60 - 6.80 0 0.00 100.00 6.80 - 700 0 0.00 100.00 7.00 - 7.20 0 0.00 100.00 298 Kajaani FS-200 results for never-dried bleached pulp Sample report Sample title: Sampled: Analysed: Upper limit: Lower limit: Length: 0.20 mm B L T C K 16-03-99 21:43 7.20 mm 0.00 mm P = 40.15% Arithmetic A v : L weighted: W weighted: Total fibres: W = 3.61% 0.74 mm 1.76 mm 2.32 mm 9008 Weighted distributi Length on Fibres W G H T D % W C U M L % 0.00 - 0.20 3617 3.61 3.61 0.20 - 0.40 ?031 4.26 7.88 0.40 - 0.60 701 5.12 13.00 0.60 - 0.80 562 5.75 18.74 0.80- 1.00 441 5.85 24.59 1.00- 1.20 425 6.94 31.54 1.20- 1.40 362 7.01 38.55 1.40-1.60 338 7.55 46.10 1.60-1.80 303 7.70 53.80 1.80-2.00 236 6.70 60.50 2.00 - 2.20 217 6.82 67.31 2.20 - 2.40 181 6.21 73.52 2.40 - 2.60 149 5.56 79.09 2.60 - 2.80 117 4.74 83.82 2.80 - 3.00 101 4.37 88.19 299 Weighted distributi Length on Fibres W G H T D % W C U M L % 3.00-3.20 76 3.52 91.71 3.20-3.40 50 2.46 94.18 3.40-3.60 38 1.98 96.15 3.60-3.80 25 1.39 97.54 3.80-4.00 16 0.94 98.48 4.00 - 4.20 5 0.31 98.79 4.20 - 4.40 5 0.32 99.11 4.40 - 4.60 6 0.40 99.51 4.60 - 4.80 1 0.07 99.58 4.80 - 5.00 0 0.00 99.58 5.00-5.20 1 0.08 99.66 5.20 - 5.40 1 0.08 99.74 5.40 - 5.60 1 0.08 99.82 5.60-5.80 0 0.00 99.82 5.80 - 6.00 1 0.09 99.91 6.00 - 6.20 1 0.09 100.00 6.20 - 6.40 0 0.00 100.00 6.40 - 6.60 0 0.00 100.00 6.60 - 6.80 0 0.00 100.00 6.80 - 700 0 0.00 100.00 7.00 - 7.20 0 0.00 100.00 Kaiaani FS-200 results for previously-dried bleached pulp 300 Sample report Sample title: Sampled: Analysed: Upper limit: Lower limit: Length: 0.20 mm D R T C K 16-03-99 21:54 7.20 mm 0.00 mm P = 52.67% Arithmetic A v : L weighted: W weighted: Total fibres: W = 5.48% 0.73 mm 2.22 mm 2.93 mm 17569 Weighted distributi Length on Fibres W G H T D % W C U M L % 0.00 - 0.20 9253 5.48 5.48 0.20 - 0.40 ?716 3.48 8.96 0.40 - 0.60 792 2.96 11.92 0.60 - 0.80 629 3.34 15.25 0 .80-1.00 526 3.60 18.86 1.00- 1.20 502 4.23 23.09 1.20- 1.40 461 4.59 27.68 1.40-1.60 421 4.85 32.53 1.60-1.80 407 5.30 37.83 1.80-2.00 365 5.34 43.17 2.00 - 2.20 343 5.54 48.71 2.20 - 2.40 359 6.34 55.05 2.40 - 2.60 300 5.78 60.83 2.60 - 2.80 242 5.03 65.87 2 .80-3.00 249 5.58 71.44 301 Weighted distribut Length ion Fibres W G H T D % W C U M L % 3.00-3.20 217 5.20 76.64 3.20-3.40 182 4.64 81.29 3.40-3.60 139 3.75 85.04 3.60-3.80 121 3.45 88.49 3.80-4.00 108 3.25 91.74 4.00 - 4.20 58 1.84 93.57 4.20 - 4.40 71 2.36 95.93 4.40 - 4.60 40 1.39 97.32 4.60 - 4.80 21 0.76 98.08 4.80 - 5.00 17 0.64 98.72 5.00 - 5.20 10 0.40 99.12 5.20 - 5.40 6 0.25 99.36 5.40-5.60 4 0.17 99.53 5.60-5.80 1 0.04 99.58 5.80 - 6.00 6 0.27 99.85 6.00 - 6.20 1 0.05 99.90 6.20 - 6.40 1 0.05 99.95 6.40 - 6.60 0 0.00 99.95 6.60 - 6.80 1 0.05 100.00 6.80 - 700 0 0.00 100.00 7.00 - 7.20 0 0.00 100.00 Annex J: Error analysis 303 Annex J. 1: 2-D Experiments For each flow rate, the bubble rise velocity was calculated 3 to 4 times. Then the average value was plotted and the standard deviation o f the results were used as the error. Annex J.2: Bubble size and rising velocity Experiments with fibre optic probes We calculated the rising velocity and the bubble size using the two signals coming from the two optic probes 10-14 times per flow rate for each consistency. The average and the standard deviation were then calculated. But, since there are lots of different size of bubbles, the frequency of each bubble size was also determined and presented in the results chapter. Annex J.3: Gas disengagement technique and density variation method A l l the experiments were repeated 6 times. The total, dense phase, small bubble, large bubble, and dilute phase gas holdup and the small and large bubble superficial and rise velocity average values were calculated along with the standard variation. Figure J. 1 illustrates an example o f the each of the total gas holdup values found with the gas disengagement technique. The pressure probe accuracy is 0.004%. The errors calculated are within the points size and shape. For example, we find an error of0.007 for a gas holdup of 0.2. The maximum error on the pressure readings is 15 Pa and the error on the distance between the two pressure transducers is 0.01 cm. 304 0,09 0,08 0,07 0,06 a. | 0,05 £ 0,04 0,03 0,02 0,01 O results from experiments H Average value with standard deviation 0,00000 0,00300 0,00600 0,00900 0,01200 superficial velocity (m/s) 0,01500 0,01800 Figure J. 1: Comparison of the average values of the total gas holdup calculated with the gas disengagement technique with BSTCK pulp at 8.5% consistency and the results from the experiments (un-averaged) as a function of gas superficial velocity Annex J.4: Dissolved oxygen measurement A l l the experiments were repeated 2-4 times. In Figure 4.45, all the k L a values calculated are presented. The average and the standard deviations are presented in Annex Ff.