Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Shear stress and fouling control in hollow fiber membrane systems under different gas sparging conditions 2012

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2012_fall_ye_dongying.pdf [ 5.47MB ]
ubc_2012_fall_ye_dongying.pdf
Metadata
JSON: 1.0050877.json
JSON-LD: 1.0050877+ld.json
RDF/XML (Pretty): 1.0050877.xml
RDF/JSON: 1.0050877+rdf.json
Turtle: 1.0050877+rdf-turtle.txt
N-Triples: 1.0050877+rdf-ntriples.txt
Citation
1.0050877.ris

Full Text

 SHEAR STRESS AND FOULING CONTROL IN HOLLOW FIBER MEMBRANE SYSTEMS UNDER DIFFERENT GAS SPARGING CONDITIONS   by   Dongying Ye   B.Eng, National University of Singapore, 2009   A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF APPLIED SCIENCE   in   The Faculty of Graduate Studies (Civil Engineering)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August, 2012  © Dongying Ye, 2012    ii  ABSTRACT   The shear stress created by gas sparging has been widely recognized as a controlling factor in the fouling of submerged membrane systems. Effective gas sparging can significantly reduce the fouling and improve the membrane performance. Several factors, such as membrane module configuration, gas sparging tank configuration, gas sparging pattern and temperature, affect the hydrodynamic conditions around the membrane. Chan et al. (2011) reported that different types of shear profiles exist inside the submerged hollow fiber membrane module, and the different types of shear conditions have different effects on fouling control. In this thesis, the relationship between the shear stress generated by conventional sparging and a novel sparging approach on fouling control were studied. The results indicate that, to achieve the similar fouling control (fouling rate), only a quarter of energy (i.e air flow rate) input was required for the novel slug bubble sparger, compared to the conventional coarse bubble sparger.                   iii  TABLE OF CONTENTS   ABSTRACT ............................................................................................................................................................. ii TABLE OF CONTENTS ........................................................................................................................................ iii LIST OF TABLES ................................................................................................................................................... v LIST OF FIGURES ................................................................................................................................................ vii ACKNOWLEDGEMENTS ..................................................................................................................................... x 1. INTRODUCTION ................................................................................................................................................ 1 2. LITERATURE REVIEW ..................................................................................................................................... 3 2.1 Membrane Filtration Process .......................................................................................................................... 3 2.2 Membrane Performance ................................................................................................................................. 5 2.3 Membrane Fouling ......................................................................................................................................... 6 2.3.1 Type of fouling ........................................................................................................................................ 6 2.3.2 Particle blocking model ........................................................................................................................... 6 2.4 Air Sparging for Fouling Control ................................................................................................................... 9 2.5 Knowledge Gap and Research Objective...................................................................................................... 11 3. MATERIALS AND METHODS ....................................................................................................................... 13 3.1 Configuration of Membrane Modules .......................................................................................................... 13 3.1.1 Membrane modules used for shear stress measurements ....................................................................... 13 3.1.2 Membrane module used for filtration tests ............................................................................................ 16 3.2 Experimental Apparatus ............................................................................................................................... 20 3.2.1 Experimental apparatus for shear stress measurements ......................................................................... 20 3.2.2 Experimental apparatus for filtration tests ............................................................................................. 23 3.3 Experimental Program .................................................................................................................................. 26 3.4 QA/QC ......................................................................................................................................................... 29 3.4.1 QA/QC for shear stress measurements .................................................................................................. 29 3.4.2 QA/QC for filtration tests ...................................................................................................................... 29 4. RESULTS AND DISCUSSIONS ....................................................................................................................... 30 4.1 Shear Stress Measurements Results .............................................................................................................. 30 4.1.1 Shear stress measurements results for D14 tank experiments ................................................................ 30 4.1.2 Shear stress measurements results for D20 tank experiments ................................................................ 32 4.2 Shear Stress Measurements Discussions ...................................................................................................... 35 iv  4.2.1 Effect of diffuser type ............................................................................................................................ 35 4.2.2 Effect of system tank configuration ....................................................................................................... 38 4.2.3 Effect of gas flow rate ........................................................................................................................... 41 4.2.4 Effect of gas flow pattern ...................................................................................................................... 42 4.2.5 Effect of bundle slackness ..................................................................................................................... 43 4.2.6 Summary of observations ...................................................................................................................... 44 4.3 Filtration Tests Results ................................................................................................................................. 45 4.4 Filtration Tests Discussions .......................................................................................................................... 48 4.4.1 RMS and fouling rate ............................................................................................................................ 48 4.4.2 Effect of diffuser type ............................................................................................................................ 50 4.4.3 Effect of bulk liquid flow pattern........................................................................................................... 53 4.4.4 Effect of gas flow pattern ...................................................................................................................... 54 4.4.5 Summary of observations ...................................................................................................................... 55 5. CONCLUSIONS ................................................................................................................................................ 56 5.1 Overall Conclusions ..................................................................................................................................... 56 5.2 Contribution and Engineering Significance .................................................................................................. 57 REFERENCES ....................................................................................................................................................... 58 APPENDIX 1. PRINCIPLE OF ELECTROCHEMICAL SHEAR PROBE .......................................................... 62 APPENDIX 2. SHEAR PROBE CALIBRATIONS ............................................................................................... 66 APPENDIX 3. THE EFFECT OF TEMPERATURE TO LIMITING CURRENT ................................................ 75 APPENDIX 4. SHEAR RATE IN THE SIDE LOOP SETUP ............................................................................... 76 APPENDIX 5. SIDE LOOP STEADY-STATE CALIBRATION RESULTS ....................................................... 79 APPENDIX 6. CONVERSION BETWEEN STEADY-STATE CALIBRATION COEFFICIENT AND VOLTAGE-STEP CALIBRATION COEFFICIENT............................................................................................. 80 APPENDIX 7. VOLTAGE-STEP CALIBRATION RESULTS ............................................................................ 81 APPENDIX 8. EXPERIMENTAL PROCEDURES FOR THE SHEAR STRESS MEASUREMENT ................. 82 APPENDIX 9. EXPERIMENTAL PROCEDURE FOR FILTRATION TEST ..................................................... 84 APPENDIX 10. QA/QC RESULTS ....................................................................................................................... 86 APPENDIX 11. SHEAR STRESS MEASUREMENT RESULTS IN D14 TANK ............................................... 91 APPENDIX 12. RMS, MEAN AND STD FOR ALL SHEAR STRESS MEASUREMNTS............................... 118 APPENDIX 13. FILTRATION TESTS RESULTS ............................................................................................. 120   v  LIST OF TABLES   Table 2-1.  General characteristics of membrane processes.................................................................4 Table 3-1. Bulk liquid velocity under different air flow rates...........................................................22 Table 3-2. Experimental conditions investigated for shear stress measurements..............................27 Table 3-3. Experimental conditions investigated for filtration tests..................................................28 Table 4-1. Shear stress measurements results of D14 tank experiments...........................................31 Table 4-2. Shear stress measurements results of D20 tank experiments...........................................35 Table 4-3. Fine bubble diffuser vs. coarse bubble diffuser................................................................35 Table 4-4. Pulse bubble diffuser vs. coarse bubble diffuser..............................................................36 Table 4-5. Slow pulse diffuser vs. fast pulse diffuser........................................................................38 Table 4-6. Baffled vs. unbaffled system tank.....................................................................................39 Table 4-7. Cross-flow vs. stagnant.....................................................................................................40 Table 4-8. Gas flow rate: low vs. medium vs. high...........................................................................41 Table 4-9. Gas flow pattern: continuous vs. fast intermittent vs. slow intermittent..........................42 Table 4-10. Bundle slackness: 1.3% vs. 2.2%.....................................................................................44 Table 4-11. Fouling rates of the filtration tests under different conditions..........................................48 Table 4-12. The filtration results and the corresponding 1  shear stress measurements results.............49 Table 4-13. The RMS and filtration results of coarse bubble diffuser vs. pulse bubble diffuser........50 Table 4-14. The RMS and filtration results of slow pulse diffuser vs. fast pulse diffuser...................52 Table 4-15. The RMS and filtration results of stagnant vs. cross-flow...............................................53 vi  Table 4-16. Continuous gas flow pattern vs. intermittent gas flow pattern.........................................54                        vii  LIST OF FIGURES   Figure 2-1.  Definition sketch for a membrane process.........................................................................3 Figure 2-2. The four classical filtration models....................................................................................7 Figure 3-1. The top view of the fibers before roll up..........................................................................14 Figure 3-2. (a) One end of the fiber pad with electrical tap. (b) The middle of the fiber pad with shear probes......................................................................................................................14 Figure 3-3. Cross-section view of the bundle......................................................................................14 Figure 3-4. (a) Image of the fiber bundle used for shear stress measurements. (b) Image of the top end of the bundle. (c) Image of bottom end of the bundle...............................................15 Figure 3-5. The structure of membrane module used for shear stress measurements........................15 Figure 3-6. (a) Cross-section view of the active membrane fibers. (b) Image of the section view of the active membrane fibers...............................................................................................16 Figure 3-7. (a) Schematic of the active membrane fibers. (b) Image of the active membrane fibers.................................................................................................................................17 Figure 3-8. (a) The top view of the fibers before roll up. (b) Image of one end of the fibers before roll up................................................................................................................................18 Figure 3-9. (a) Cross-section view of the membrane bundle. (b) Image of bottom end of the bundle. (c) Image of the top end of the bundle..............................................................................18 Figure 3-10. Image of the membrane bundles used for filtration tests..................................................19 Figure 3-11. The structure of the membrane module used for filtration tests.......................................19 Figure 3-12. The schematic diagram of the shear stress measurements apparatus...............................20 Figure 3-13. Three different system tanks configuration......................................................................21 viii  Figure 3-14. Filtering mode of the filtration setup................................................................................24 Figure 3-15. Backwash mode of the filtration setup.............................................................................25 Figure 4-1. Shear stress measurements result of slow pulse diffuser, low gas flow rate, without baffle, stagnant, continuous gas flow and 1.3% slackness in D14 tank...........................30 Figure 4-2. Shear stress measurements result of slow pulse diffuser, 3500 mL/min gas flow rate, with baffle, cross-flow, continuous gas flow and 1.8% slackness in D20 tank................33 Figure 4-3. Shear stress measurements result of fast pulse diffuser, 3500 mL/min gas flow rate, with baffle, cross-flow, continuous gas flow and 1.8% slackness in D20 tank........................34 Figure 4-4. Fine bubble diffuser vs. coarse bubble diffuser................................................................36 Figure 4-5. Pulse bubble diffuser vs. coarse bubble diffuser..............................................................37 Figure 4-6. (a) Probe1 test 1 shear profile of CMOSC1.3. (b) Probe1 test 1 shear profile of PMOSC1.3........................................................................................................................37 Figure 4-7. Slow pulse diffuser vs. fast pulse diffuser........................................................................38 Figure 4-8. Baffle vs. unbaffle system tank........................................................................................39 Figure 4-9. Cross-flow vs. stagnant.....................................................................................................40 Figure 4-10. Gas flow rate: low vs. medium vs. high...........................................................................42 Figure 4-11. Gas flow pattern: continuous vs. fast intermittent vs. slow intermittent..........................43 Figure 4-12. Bundle slackness: 1.3% vs. 2.2%.....................................................................................44 Figure 4-13. Averaged filtration results in D14 tank............................................................................46 Figure 4-14. Averaged filtration results in D20 tank............................................................................47 Figure 4-15. RMS vs. fouling rate.........................................................................................................50 Figure 4-16. Filtration results of coarse bubble diffuser vs. pulse bubble diffuser...............................51 Figure 4-17. Filtration results of slow pulse diffuser vs. fast pulse diffuser.........................................52 ix  Figure 4-18. Filtration results of stagnant vs. cross-flow......................................................................53 Figure 4-19. Filtration results of continuous gas flow pattern vs. intermittent gas flow pattern…......54                      x  ACKNOWLEDGEMENTS   First of all, I would like to thank my thesis supervisors, Dr. Pierre Bérubé for his valuable insight, guidance, and trust throughout my M.A.Sc. study. Without his continuous support, I wouldn’t be able to overcome all the obstacles encountered during different stages of this research.  I would also like to thank Bill Leung and Scott Jackson for their genuine help in building the experiment apparatus and electrical set-up for the shear stress measurements and filtration tests. Special thanks to Paula Parkinson and Tim Ma for their technical assistance and advice in the Environmental laboratory, and their patience to bear with me for my demanding requirements of pipe fittings.  In addition, I would like to thank PhD candidate Sepideh Jankhah for teaching me how to fabricate and calibration the shear probe and PhD candidate Syed Abdhullah for sharing his knowledge of the membrane filtration.  I am sincerely grateful to my parents, for all they have done for me so that I can be where I am today. I am forever grateful for their boundless love and sacrifices.  Last but not least, without the support from my girlfriend Corrine, I wouldn’t be able to live with my passion of research and commit myself to my M.A.Sc. study. Thank you for your unconditional and care all along the way. I am really grateful to have you with me.        1  1. INTRODUCTION   The tightening of regulations on drinking water quality and wastewater discharge limits has led to an increasing demand for advanced water and wastewater treatment technologies. With its excellent filtration capability and decreasing cost, membrane technology attracts the growing attention of researchers in the field of water and wastewater treatment.  The use of membranes in biological treatment of wastewater is becoming one of the most important applications of membranes. Compared to the conventional technologies, membrane bioreactors (MBR) offer many advantages, including smaller footprint, better treated water quality, higher volumetric loadings and less sludge production.  Despite the many benefits of membrane technologies, the unavoidable problem of membrane fouling remains during membrane filtration. Membrane fouling is a process where solute or particles deposit onto a membrane surface or into membrane pores. Membrane fouling causes a reduction in the permeate yield and a decrease quality of water produced (Metcalf and Eddy, 1991, 2003), which leads to the increase of operational and capital costs of the entire membrane filtration processes.  Shear forces are recognized to promote the mass transfer of foulants away from membrane surfaces and therefore contribute to fouling control. In the nanofiltration (NF) and reverse osmosis (RO) system, the cross-flow operation is typically used to induce shear stresses at the membrane surface for fouling control. In submerged microfiltration (MF) and ultrafiltration (UF) membrane systems, the gas sparging under the membrane module is commonly used to induce shear stresses at the membrane surface for fouling control. However, due to the complex conditions in the gas sparging membrane systems, the mechanisms of fouling control by sparging are poorly understood.  In the recent studies, Bérubé et al. (2006), Chan et al. (2007)  and Fulton et al. (2011) demonstrated that air sparging can create different kind of shear profiles in submerged hollow fiber systems. Chan et al. (2011) also did a comprehensive study investigating the effects of sparger location, fiber movement and physical contact between fibers on the shear stress induced on the hollow fiber membrane surface. This research indicated that hydrodynamic conditions will affect the mass transfer at the membrane surface and affect the fouling control. The present thesis builds on these previous studies and investigates the 2  relationship between the shear stress generated by conventional sparging and novel sparging approach on fouling control.                            3  2. LITERATURE REVIEW   In this chapter, the concepts of the membrane filtration process, transmembrane pressure, membrane fouling and shear stress are reviewed.   2.1 Membrane Filtration Process  During the filtration process, the membrane serves as a selective barrier that only allows the passage of certain constituents and will retain other constituents in the liquid. Referring to Figure 2-1, the influent to the membrane module is known as feed water (also known as feed stream). The liquid that passes through the semi-permeable membrane is known as permeate and the liquid containing the retained or rejected constituents is known as concentrate (also known as retentate, reject, or waste stream).        Figure 2-1. Definition sketch for a membrane process.   Membrane processes include microfiltration (MF), ultrafiltration (UF), nanoflitration (NF), reverse osmosis (RO), dialysis, and electrodialysis (ED) (Metcalf and Eddy, 2003). The first four membrane processes are commonly used in water and wastewater treatment and a hydraulic pressure provides the driving force. Dialysis involves the transport of constituents through a semi-permeable membrane on the basis of concentration differences. Electrodialysis involves the use of an electromotive force and ion- selective membranes to achieve the separation of charged ionic species. The focus of the following sections is on pressure-driven membrane processes. Table 2-1 illustrates the general characteristics of the four pressure driven membrane processes. Feedwater Container for membrane modules Membrane Permeate Concentrate 4  Table 2-1. General characteristics of membrane processes. Membrane Process Typical operating pressure, bar Separation mechanism Pore size Typical operating range, μm Typical constituents removed Microfiltration 0.1-2 Sieve Macropores (>50 nm) 0.08-2.0 TSS, turbidity, protozoan oocysts and cysts, some bacteria and viruses Ultrafiltration 0.1-2 Sieve Mesopores (2-50 nm) 0.005-2.0 Macromolecules, colloids, most bacteria, some viruses, proteins Nanofiltration 4-20 Sieve + solution/ diffusion +exclusion Micropores (<2 nm) 0.001-0.01 Small molecules, some hardness, viruses Reverse osmosis 10-30 Sieve + solution/ diffusion +exclusion Dense (< 2 nm) 0.0001- 0.001 Very small molecules, color, hardness, sulphates, nitrates, sodium, other ions   Membranes used for water and wastewater application can be made from a number of organic and inorganic materials. Organic membranes are typically made of synthetic polymers such as polyvinylidene difluoride (PVDF), polyethylsulfone (PES), polyethylene (PE) and polypropylene (PP) (Judd, 2006). Inorganic membranes such as ceramic and metallic membranes have been used, although they are not as commonly used as organic membrane. One reason is that the cost of inorganic membranes can be up to ten times higher than that of the organic membrane (Owen et al., 1995). However, inorganic membrane can usually tolerate higher temperatures, higher pressure and harsher chemical conditions associated with the feed water or cleaning solutions. 5  2.2 Membrane Performance  The driving force for membrane filtration is the transmembrane pressure (TMP), defined as the difference between the feed water pressure and the permeate pressure:                                            TMP = Pfeed  - Ppermeate = ∆P  (2-1)  Referring to the figure 1-1, the permeate flow obtained during filtration can be described by Darcy’s Law (Belfort et al., 1994):   T R A)P( Q                   (2-2)  where Q [m 3 ·s -1 ] is the permeate flow rate through the membrane surface, ∆P [Pa] is the TMP, ∆π [Pa] is the osmotic pressure of the solution, A [m 2 ] is the membrane area, µ [Pa·s] is the solution viscosity and RT [m -1 ]is the total membrane resistance to the permeate flow. For a dilute solution, the osmotic pressure of the retained material can be calculated using the van’t Hoff equation (Belfort et al, 1994):   M vcRT                                          (2-3)  where v [dimensionless] is the molar fraction of species, c [g·L -1 ] is the solute concentration, R [Pa·L·mol -1 ·K -1 ] is the universal gas constant, T [K] is the temperature and M [g·mol -1 ] is the molecular weight of the solute. In microfiltration and ultrafiltration, the molecular weight of the solute in the feed water is usually high as they colloidal in size. As a result, the osmotic pressure is typically negligible in MF and UF.  Viscosity is an important parameter that will affect the performance of the membrane. In wastewater treatment applications, the viscosity of activated sludge can be characterized as shear-thinning and non- Newtonian (Seyssiecq et al., 2003). In water treatment applications, the viscosity of solution is typically similar to that of pure water. Temperature, which significantly affects viscosity, is therefore an important parameter affecting the performance of membranes systems.  The total membrane resistance (RT) to the permeate flow is the sum of intrinsic membrane resistance and the resistance due to fouling. Membrane fouling is defined as a “process resulting in loss of performance 6  of a membrane due to the deposition of suspended or dissolved substances on its external surfaces, at its pore openings, or within its pores” (Koros et al., 1996). Overtime, as fouling increases and the resistance to the permeate flow becomes too high, the membrane must be physically and/or chemically cleaned to remove the accumulated foulants.   2.3 Membrane Fouling  Membrane fouling is an unavoidable consequence of membrane filtration. It causes an increase of resistance to the permeate flow, which leads to an increase in the TMP or a decrease in permeate flow over time.  2.3.1 Type of fouling  Membrane fouling in general can be divided into two categories – reversible and irreversible. Reversible fouling refers to the resistance that can be removed after hydraulic or chemical cleaning. The main cause of reversible fouling is the accumulation of retained material on the membrane surface or in the membrane pores.  Irreversible fouling is defined as the increase of membrane resistance that is permanent and cannot be removed, even after hydraulic and/or chemical cleaning. Irreversible fouling is caused by the formation of chemical precipitates and/or the permanent damage to the membrane surface due to the biological or chemical agents, which may result in permanent blockage of pores and changes to the membrane’s physical/ chemical properties.  The extent of irreversible fouling is dependent on the frequency of backwashing, as well as the strength of adhesion of the clogging particles and precipitates on the membrane surfaces (Chang et al., 2002). The focus of the remainder of the discussion in this literature review is on the hydraulically reversible fouling.  2.3.2 Particle blocking model  Hermans and Bredée (1935) proposed filtration laws to describe the accumulation of foulants on membranes. Gonsalves (1950) made a critical study of the physical models used to derive these laws. The accumulation of foulants was categorized into four filtration models, namely 1) complete pore 7  blocking model, 2) intermediate pore blocking model, 3) standard pore blocking model and 4) cake layer formation model. Figure 2-1 illustrates each type of model schematically.                             (a)                                 (b)                              (c)                               (d)  Figure 2-2. The four classical filtration models a) complete pore blocking b) intermediate pore blocking c) standard pore blocking d) cake layer formation.  In the complete pore blocking model, every particle can block one membrane pore when it reaches the membrane surface. This causes a decrease in the available pores and results in an increase in the resistance. In the intermediate pore blocking model, not all of the particles block the membrane pore. Some particles can overlap existing particles. As with partial pore blocking, the available pores is reduced. In the standard blocking model, small particles deposit/adsorb onto the membrane surface as well as on the inner walls of the membrane pores. The deposition/adsorption of the particles on the membrane walls decreases the pore diameter and results in an increase in the resistance. In the cake formation model, particles accumulate, forming a cake layer on the membrane surface and results in an increase in the resistance.  Hermia (1982) developed a mathematical model that describes the fouling of membranes base on the four filtration laws for the constant pressure filtration conditions. The characteristic form of the model is:                    n 2 2 ) dV dt (k dV td ni                     (2-4)  where the t [s] and V [m 3 ] are the filtration time and cumulative permeate volume, dt/dV is the inverse of the permeate flow rate and the rate of variation of the dt/dV with respect to filtrate volume V is defined as the resistance coefficient which is d 2 t/dV 2 . The n value is different for the different filtration laws: for complete blocking, n = 2; for standard blocking, n = 1.5; for intermediate blocking, n = 1; and for cake layer formation, n = 0. ki n  is the blocking constant and is different for each filtration law. 8  For a constant flow and variable pressure systems, the Hermia’s model yields equations 2-5 to 2-8 for the four different fouling laws.  Cake fouling:                       VkPP c0t                            (2-5) Intermediate blocking:                       Vk 0t iePP                                  (2-6) Standard blocking:  2 s 0 t )Vk 2 1 1( P P                           (2-7) Complete blocking:  Vk1 P P b 0 t                                  (2-8)  where the Pt [Pa] is the TMP at time t and P0 is the initial pressure, V is the filtered volume and kc, ki, ks, and kb are the four different blocking constant.  The Hermia’s model was developed assuming dead-end filtration with no back transport. Field et al. (1995) modified the Hermia’s model for the cake formation, intermediate blocking and complete blocking filtration laws, by assuming that under these three laws, the particle can transport back into the bulk solution as presented in equation 2-9.   )J-J(kJ dt dJ - *2-n ni                        (2-9)  where the J [m 3 ·m -2 ·s -1 ] is the  permeate flux (dV/dt/A), J *  [m 3 ·m -2 ·s -1 ] is the critical flux defined by Field et al. (1995) that fouling will not happen when the permeate flux below the critical flux and the ki n  is the blocking constant for different filtration laws.  The back transportation of the particle has been suggested to occur predominantly due to three mechanisms for which models have been designed. In the shear-induced diffusion model, particles subjected to a shear flow randomly “bump into” and tumble over each other as well as the membrane 9  surface, resulting in net displacement of the particles away from the membrane surface (Zydney and Colton, 1986). In the inertial lift model, particles subjected to a laminar cross-flow velocity distribution near the membrane surface undergoes rotation, and reusling in a lateral movement away from the membrane due to the differential pressure induced by the velocity gradient of the rotating particles (Green and Belfort, 1980). The surface transport models suggest that the particle at the membrane surface roll along the surface due to tangential flow along the membrane surface (Beloft et al., 1994).  The three models suggest that particle back transport is strongly affected by the shear rate (as well as the particle size and concentration) (Beloft et al., 1994). However, the particle transport models were developed for constant laminar cross-flow conditions at a membrane surface. In a gas sparged membrane systems, the hydrodynamic conditions at the surface are highly variable and turbulent. Therefore, the above models cannot be applied directly to gas sparged membrane systems. Nonetheless, the models do suggest that particle back transport in gas sparged systems is likely to be affected by the share rate.   2.4 Air Sparging for Fouling Control  As previously discussed, shear rate is one of the critical parameters governing the particle transportation and enhancing permeate flux during membrane filtration. Shear rate is related to the viscosity and the shear stress at the membrane surface, which can be controlled by the temperature and the hydrodynamic conditions near the membrane. Gas spraging is typically used to induce favourable hydrodynamic conditions in the water and wastewater application.  The use of air sparging for fouling control in membrane processes has been gaining popularity over the past decade, particularly in the areas of drinking water production and biological wastewater treatment (i.e. membrane bioreactor) and macromolecular separation (Mercier-Bonin et al., 2003). Compared to cross-flow filtration (without air sparging), air sparging has been widely acknowledged to reduce the extent of fouling by 30 to 300%, depending on the applications, the operating conditions, membrane configuration (tubular, hollow fiber or flat sheet membranes) and the characteristics of the liquid being filtered (Bellara et al., 1996; Ueda et al., 1997; Bérubé and Lei, 2006; Cabassud et al., 1997; Cheng et al., 1998; Cui and Wright, 1994; Li et al., 1998).  10  Unconfined modules (e.g. submerged hollow fibers or flat sheet membranes) are those for which the membranes are submerged in the liquid to be filtered and then the membrane surfaces are scoured by sparged air bubbles. For these modules, the liquid typically permeates through the membrane in an outside-in flow configuration. Confined modules (e.g. tubular cross-flow membranes) are those for which the liquid to be filtered and sparged air bubbles are confined within the membrane. For these modules, the liquid typically permeates through the membrane in an inside-out flow configuration.  Although the surface shear stresses induced by a rising air sparged bubble in a confined vertical duct have been extensively characterized both numerically and experimentally, very limited research has focused on characterizing the hydrodynamics of air sparged bubbles in unconfined systems such as submerged flat sheet or hollow fiber membranes. This is in part due to the extremely complex nature of the geometry and flow path of air sparged bubbles in unconfined systems.  Some researchers have assumed that since the packing density of fibers in a submerged membrane system is relatively high, the shear stresses induced onto membrane surfaces by air-sparged bubbles rising between fibers is similar to that induced by an air slug in a confined system (Cui et al., 2003; Busch et al., 2007; Chang and Fane, 2000). However, the validity of this assumption is questionable given the complex and ever-changing nature of the hydrodynamic conditions in submerged hollow fiber membrane systems. In submerged hollow fiber membrane systems, the fibers are normally held loosely and typically sway extensively. This irregular fiber configuration continuously changes the flow path of rising air-sparged bubbles.  Bérubé et al. (2006) measured the shear stresses on the surface of hollow fibers in a submerged membrane system using a non-directional electrochemical probe. The frequency, duration and amplitude of the shear events were highly variable, and significantly different from the shear profiles of air slugs rising in a confined tubular system. Nagaoka et al. (2006) used a two-direction load sensor to record the shear stress profile acting over the entire length of the hollow fiber during gas sparging. Their reported profiles were similar to those observed by Bérubé et al. (2006), i.e. highly variable frequency, duration and amplitude of shear events. Yeo et al. (2006) used particle image velocimetry (PIV) to measure liquid velocities in an air sparged hollow fiber module. They observed that the axial velocities inside a hollow fiber module were up to ten times lower than those outside of the module where sparged air bubbles were introduced. This observation is similar to those of Bérubé et al. (2006) who suggested that a tightly configured multi-fiber module could potentially shield certain areas of a fiber from both the 11  bulk liquid flow and sparged gas bubbles. Fulton et al. (2011) used the electrochemical shear probe to map out the shear profile in a pilot scale submeraged hollow fiber membrane module. They found that the distributions of surface shear forces in the sytem were highly affected by the system geometry (e.g. module spacing, tank configuration and diffuser nozzle size). Chan et al. (2011) did a comprehensive study of the hydrodynamic conditions in the air sparged hollow fiber membrane modules. The effect of fiber tension, fiber packing density, sparging rate and the location of the aerator on the shear stress at the membrane surface within a bench-scale submerged hollow fiber system was investigated by using an electrochemical shear probe. Chan observed that increasing sparging rate increased both cross-flow velocity (‘baseline shear’) as well as the frequency of high shear events, while fiber tension and packing density affected both the ability of bubbles to penetrate and stay within the fiber bundle. The closer the bubble was to the membrane surface, the higher were the shear stress observed. In the study, using an electrochemical shear probe together with a high speed camera, they observed that the high shear events caused by the passage of bubbles in a multi-fiber system were approximately 0.02-0.1 seconds in duration. Their results also indicated that shear profiles at short duration and high frequency were preferred for fouling control than those of longer duration and low frequency. However, an optimal frequency was observed above which no additional benefit for fouling control could be observed.   2.5 Knowledge Gap and Research Objective  Base on the recent studies by Bérubé et al. (2006), Fulton et al. (2011) and Chan et al. (2011), air sparging can create different kinds of shear profiles in submerged hollow fiber systems, which will lead to different hydrodynamic conditions around the membrane and affect the fouling control efficiency. In addition, the shielding effect of the membrane bundle can reduce the shear stress inside the membrane. Chan et al. (2011) used a mechanical method to create different shear profiles in a bench-scale system and carried out several fouling testes under those conditions. The study by Chan et al. (2011) indicated that variable shear conditions were better than constant shear conditions for fouling control.  However, in their study, the shear conditions were generated mechanically with an impeller, rater tan with air sparging in full-scale submerged membrane system.  The first objective of the present study is to generate the variable shear conditions identified by Chan et al. (2011) to be optimal for fouling control, by using gas sparging. The second objective of the present study is to confirm if these variable shear conditions do in fact result in optimal fouling control in an air 12  sparging system with full-scale hollow fibers. To achieve the objects, the shear profiles of different air sparging conditions are measured and analyzed by an electrochemical shear method, followed by the fouling test for the selected conditions.                           13  3. MATERIALS AND METHODS   In the present thesis, both shear stress measurements and filtration tests were carried out. This chapter describes the experiment setups and programs used for the shear stress measurements and the filtration tests. In order to create similar conditions during shear stress measurements and filtration tests, similar experiment steups (i.e. membrane modules and air sparging apparatus) were used throughout the study.   3.1 Configuration of Membrane Modules  Since it was essential to create the comparable hydrodynamic conditions for the shear stress measurements and the filtration tests, the configuration of the modules needed to be similar for all experiments. In this section, details of the configurations of the modules, as well as the processes used to construct the modules, are presented.  3.1.1 Membrane modules used for shear stress measurements  The shear probes used in the experiments were similar to those used by Fulton et al. (2011). The principle of the shear probe method was presented in Appendix 1. Before the probes were used, they were calibrated as presented in a calibration process in Appendix 2 to 7. Refer to Fulton et al. (2011) for more details about the shear probe calibration procedure.  The module was constructed using two test fibers on to which shear probes had been installed as well as 48 full-length (1.8 m) inactive membrane fibers. The module was assembled using the following procedure.  1) Lay all the inactive 48 fibers and two test fibers closely together on a flat surface, as illustrated in Figure 3-1. The two test fibers were placed at the side of the fibers and shear probe was located at the middle of the modules. Electrical tape was used to fix both ends of all the inactive fibers and test fibers as presented in Figure 3-2.  14   Figure 3-1. The top view of the fibers before roll up.                                                     (a)                                                                  (b) Figure 3-2. (a) One end of the fiber pad with electrical tap. (b) The middle of the fiber pad with shear probes.  (2) Roll the inactive fibers and test fibers. Secure the ends using electrical tape. The Figure 3-3 and Figure 3-4 illustrates the rolled bundle.  Figure 3-3. Cross-section view of the bundle. inactive Rolled bundle diameter approx. 15 mm 180 cm 2 cm 2 cm 15   (a)                                                 (b)                                                                        (c) Figure 3-4. (a) Image of the fiber bundle used for shear stress measurements. (b) Image of the top end of the bundle. (c) Image of bottom end of the bundle.  (3) Slide both ends of the bundle into the 15 mm inner diameter bulkheads. The bulkheads were made of PVC as the experiment requires a metal free environment except for the cathode and anode. Zip ties were used to secure the bundle to the bulkheads. Figure 3-5 illustrates the structure of the final membrane module used for shear stress measurements.  The test fibers were placed at the center of the module to examine the potential shielding effects of fibers. The shear probes were placed at the middle of the test fibers, where maximum fiber sway is expected.   Figure 3-5. The structure of membrane module used for shear stress measurements. 16  3.1.2 Membrane module used for filtration tests  The configuration of the membrane module used for filtration tests was similar to that of the module used for shear stress measurements. However, instead of placing test fibers at the center of the module active membrane fibers were placed at this location.  Filtration was performed through these three active membrane fibers. The total filtration area for the three fibers was 0.0293 m 2 . The end of the three active membrane fibers were potted into quarter inch plastic tubes as presented in Figure 3-6 and 3-7. Epoxy was used for potting the fibers into the tubes. The plastic tube at the top of the membrane fibers was connected to a peristaltic pump (Masterflex ® ) and a suction pressure was applied during the filtration. The plastic tube at the bottom of the fibers was sealed. Care was taken to keep the active membrane fibers wet during the entire procedure.                                         (a)                                                                                          (b) Figure 3-6. (a) Cross-section view of the active membrane fibers. (b) Image of the section view of the active membrane fibers.  17   (a)  (b) Figure 3-7. (a) Schematic of the active membrane fibers. (b) Image of the active membrane fibers.  The process to assemble the fiber bundle for the filtration tests was similar to that used to assemble the fiber bundle for the shear stress measurements. The major difference was that the bundle used for the shear stress measurements contained 50 fibers (2 test fibers + 48 inactive membrane fibers) while the bundle used for the filtration tests was only 40 fibers (3 active membrane fibers + 37 inactive membrane fibers). In both cases, the outer diameter of the bundle was approximately 15 mm. The active membrane fibers were incorporated into the modules as follows.  1) Lay all the 37 inactive fibers and the 3 active membrane fibers close together on a flat surface, fix both end of all the fibers using electrical tape as illustrated in figure 3-8. To permeate pump Top plastic tube Bottom plastic tube 18                                                                   (a)                                                                                      (b) Figure 3-8. (a) The top view of the fibers before roll up. (b) Image of one end of the fibers before roll up.  (2) Roll the 37 inactive fibers around the active membrane fibers as illustrated in Figure 3-9. Secure both ends using electrical tape.   (a)                                                      (b)                                                            (c) Figure 3-9. (a) Cross-section view of the membrane bundle. (b) Image of bottom end of the bundle. (c) Image of the top end of the bundle.  2 cm 180 cm 2 cm Active membrane fibers Rolled bundle diameter approx. 15 mm 19   Figure 3-10. Image of the membrane bundles used for filtration tests.  (3) Slide both end of the bundle into the 15 mm inner diameter bulkheads and secure with zip ties as previously discussed. Figure 3-11 illustrates the structure of the membrane module used for the filtration tests. Care was taken to keep the active membrane fibers wet during the entire procedure.   Figure 3-11. The structure of the membrane module used for filtration tests.  20  3.2 Experimental Apparatus  A similar experimental apparatus was used for both the shear stress measurements and the filtration tests as presented in the following sections.  3.2.1 Experimental apparatus for shear stress measurements  The overall system used for the shear stress measurements is presented in Figure 3-12.   Figure 3-12. The schematic diagram of the shear stress measurements apparatus.  The three types of diffusers used in the shear stress measurements were located at the bottom of the system tank (i.e. fine bubble diffuser, coarse bubble diffuser and pulse bubble diffuser). The fine bubble diffuser had 6 holes of 2 mm diameter each. The coarse bubble diffuser had only one 5 mm diameter hole. The total area of the holes for the above two diffusers were very close to each other so that when System tank Masterflex® Pump 21  the same flow rate was used, the velocity of air entering the sparging tank was the same for both diffusers. The third type of diffuser was a pulse bubble diffuser. This diffuser is a proprietary system from GE Water and Process Technology (i.e. LEAD system diffuser). This diffuser is able to generate large intermittently released bubbles. The configuration of the pulse bubble diffuser can be modified to generate bubbles of various sizes released at different frequencies. In the present study, two pulse bubble diffuser configurations were tested, one is fast pulse and another one is slow pulse. For similar gas flow rates, the frequency of bubble released from the fast pulse diffuser was approximately 2 times faster than that from the slow pulse diffuser.    (a) Baffled cross-flow                          (b) Baffled stagnant                              (c) Unbaffled Figure 3-13. Three different system tanks configuration.  Three types of system tanks configurations were used, as presented in Figure 3-13. The first type was the baffled cross-flow as presented in Figure 3-13a. For this configuration, a concentric open-ended cylindrical baffle (diameter 6.5 cm and length 200 cm) was placed in the system tank. This promoted convective flow in the system. The size of the tank was 14 cm in diameter and 260 cm in length. The liquid depth was 220 cm to fully cover the cylindrical baffle. Note that for this configuration the bubble created by the pulse diffuser was larger than the diameter of the cylindrical baffle, and as a result, the bubble behaved as a taylor type slug bubble. To prevent the bubble confinement, a larger system tank with 20 cm in diameter and 260 cm in length was constructed. The larger tank also had a concentric cylindrical baffle which was 14 cm in diameter and 200 cm in length. The larger tank was used only baffle 22  with the pulse bubble diffuser. In the following discussion, the small tank (diameter 14 cm) will be referred as D14 tank and the larger tank (diameter 20 cm) will be referred as D20 tank. For some of the experiments, a doughnut shape baffled was added to the D14 tank as illustrated in Figure 3-13b to prevent convective flow and promote stagnant conditions. Comparison of the results with stagnant and cross-flow conditions of above two configurations enabled the assessment of the effect of convective flow. The third type of system tank configuration was an unbaffled D14 tank as illustrated in Figure 3- 13c. For this configuration, the concentric cylindrical baffle was removed and the membrane module was simply fixed with top and bottom supports. This system tank configuration promoted stagnant conditions and also prevented large pulse bubbles from being confined by a cylindrical baffle.  Nitrogen gas was used for sparging. The flow to the diffusers was monitored using a flow meter and adjusted using a pressure regulator. For shear measurements, three flow rates, low flow rate, medium flow rate and high flow rate with 3500 mL/min, 6000 mL/min and 8000 mL/min respectively, were used in the D14 tank. In the D20 tank, only the low flow rate (3500 mL/min) was considered. All the flow rates were measured under a pressure of 15 psi. For conditions with convective flow (Figure 3-13a), the gas sparging generated a cross-flow in the system tank. The velocity of this cross-flow was estimated by dye testing (Table 3-1).  Table 3-1. Bulk liquid velocity under different air flow rates. Gas flow rate (ml/min) Bulk liquid flow rate (cm/s) D14 tank D20 tank 3500 43 18 6000 50 - 8000 58 -  Two intermittent sparging conditions were considered for the fine and coarse bubble diffusers. For fast intermittent conditions, the sparging flow was on for 3 seconds then off for 3 seconds. For slow intermittent conditions, the sparging flow was on for 6 seconds then off for 6 seconds.  For the shear stress measurements, the current between the anode and cathode was measured and recorded using the Labview ®  software, similar to that used for the side loop setup in appendix 1. The sampling time was 1 minute for each experimental condition investigated. The cathode was the shear probe located on the test fiber at the center of the membrane module and the anode was located in the system tank with much larger surface area to ensure that the limiting current conditions was achieved. 23  DO meter was used to measure the dissolve oxygen in system tank and the temperature was also monitored.  The electrolyte solution used for the shear measurement experiments had the same compositions (0.003 M ferricynaide, 0.006 M ferrocyanide and 0.3 M potassium chloride) as that used for the shear calibration in appendix 1. The electrolyte was stored in the reservoir before the experiments as presented in Figure 3-12. Nitrogen gas and a fine diffuser were used to purge the oxygen out of the solution as the experiment required an oxygen free environment. Peristaltic pump (Masterflex ® ) was used to transfer the electrolyte between the system tank and the reservoir.  The membrane module was located at the center of the system tank. The slackness of the fiber bundle was defined as follows:  %100) fiber  theofLength bulkheads obetween tw Distance -(1  slackness Bundle                        (3-1)  In the D14 tank, two bundle slackness, 1.3% and 2.2% were used.  In the D20 tank, only 1.3% slackness was used.  3.2.2 Experimental apparatus for filtration tests  The overall systems used for the filtration tests are presented in Figure 3-14 and Figure 3-15.   24   Figure 3-14. Filtering mode of the filtration setup. Masterflex® Pump Masterflex® Pump 25   Figure 3-15. Backwash mode of the filtration setup.  The two types of diffusers used in the filtration tests were located at the bottom of the system tank (i.e. coarse bubble diffuser and pulse bubble diffuser), and were the same diffusers as those used during the shear stress measurements. For the pulse bubble diffuser, both fast and slow pulse configurations were investigated.  Two types of system tanks configurations were used as presented in Figure 3-13a and Figure 3-13c. The baffled cross-flow system tank (Figure 3-13a) in filtration tests was only used for the D20 tank and the Unbaffled system tank was only used for D14 tank. The dimensions of the system tanks used during the filtration tests were identical as those used during the shear stress measurements.  Masterflex® Pump Masterflex® Pump 26  Air was used for sparging. The flow to the diffusers was monitored using a flow meter and adjusted using a pressure regulator. In the D14 tank, two flow rates, 7000 mL/min and 3500 mL/min were used. In the D20 tank, three flow rates, 3500 mL/min, 1750 mL/min and 875 mL/min were used to investigate the effect of the coarse and pulse diffusers on membrane fouling.  Two intermittent sparging conditions were considered for the fine and coarse bubble diffusers. For the first condition, sparging flow was on for 3 seconds and off for 3 seconds (i.e. similar to the conditions used in shear stress measurements). For the second condition, sparging flow was on for 2 seconds and off for 2 seconds, to create a similar frequency as the slow pulse diffuser.  A bentonite solution was used as the feed water for the filtration tests. The concentration of bentonite was 250 mg/L and tap water was used to prepare the solution. A peristaltic pump (Masterflex ® ) was used to transfer the prepared solution or to drain the tank (Figure 3-14 and 3-15).  The membrane module use for the filtration tests (described in section 3.1.2) was also located at the center of the system tank. The bundle slackness was kept constant at 18%. The permeate side was connected to a permeate pump (Masterflex ® ) which created a vacuum during filtration (Figure 3-14). The vacuum pressure was measured using a pressure transducer and recorded using a Labview ®  software. The flow rate of the pump was set as 60 mL/min so that the permeate flux used in the experiment was 125 L/m 2 /hour. Backwash was performed after every filtration test. During the backwash mode (Figure 3-15), the direction of the permeate pump (Masterflex ® ) was reversed and a lower flow rate was used so that the backwash pressure was kept below 3 psi. A thermometer was used to monitor the temperature in the system tank.   3.3 Experimental Program  Due to the limited time and resource, only selected combinations of the conditions mentioned in the previous section were considered. For the shear stress measurements, 54 different conditions using the D14 tank and 2 conditions using D20 tank were investigated. For the filtration tests, 4 conditions using the D14 tank and 6 conditions using D20 tank were investigated. All the experiment conditions investigated are summarized in Table 3-2 and 3-3.  27  Table 3-2. Experimental conditions investigated for shear stress measurements. A) D14 Tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Experiment name Fine Diffuser (F) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.3 FLWCC1.3 Stagnant (S) Continuous (C) 1.3 FLWSC1.3 6000 mL/min Medium (M) With baffle (W) Cross-flow (C) Continuous (C) 1.3 FMWCC1.3 Stagnant (S) Continuous (C) 1.3 FMWSC1.3 8000 mL/min High (H) With baffle (W) Cross-flow (C) Continuous (C) 1.3 FHWCC1.3 Stagnant (S) Continuous (C) 1.3 FHWSC1.3 Coarse Diffuser (C) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.3 CLWCC1.3 Fast alter (F) 1.3 CLWCF1.3 Slow alter (S) 1.3 CLWCS1.3 Stagnant (S) Continuous (C) 1.3 CLWSC1.3 Fast alter (F) 1.3 CLWSF1.3 Slow alter (S) 1.3 CLWSS.13 Without baffle (O) Stagnant (S) Continuous (C) 1.3 CLOSC1.3 2.2 CLOSC2.2 Fast alter (F) 1.3 CLOSF1.3 2.2 CLOSF2.2 Slow alter (S) 1.3 CLOSS1.3 2.2 CLOSS2.2 6000 mL/min Medium (M) With baffle (W) Cross-flow (C) Continuous (C) 1.3 CMWCC1.3 Fast alter (F) 1.3 CMWCF1.3 Slow alter (S) 1.3 CMWCS1.3 Stagnant (S) Continuous (C) 1.3 CMWSC1.3 Fast alter (F) 1.3 CMWSF1.3 Slow alter (S) 1.3 CMWSS1.3 Without baffle (O) Stagnant (S) Continuous (C) 1.3 CMOSC1.3 2.2 CMOSC2.2 Fast alter (F) 1.3 CMOSF1.3 2.2 CMOSF2.2 Slow alter (S) 1.3 CMOSS1.3 2.2 CMOSS2.2 8000 mL/min High (H) With baffle (W) Cross-flow (C) Continuous (C) 1.3 CHWCC1.3 Fast alter (F) 1.3 CHWCF1.3 Slow alter (S) 1.3 CHWCS1.3 Stagnant (S) Continuous (C) 1.3 CHWSC1.3 Fast alter (F) 1.3 CHWSF1.3 Slow alter (S) 1.3 CHWSS1.3 Without baffle (O) Stagnant (S) Continuous (C) 1.3 CHOSC1.3 2.2 CHOSC2.2 Fast alter (F) 1.3 CHOSF1.3 2.2 CHOSF2.2 Slow alter (S) 1.3 CHOSS1.3 2.2 CHOSS2.2 Slow Pulse Diffuser (P) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.3 PLWCC1.3 Stagnant (S) Continuous (C) 1.3 PLWSC1.3 Without baffle (O) Stagnant (S) Continuous (C) 1.3 PLOSC1.3 2.2 PLOSC2.2 6000 mL/min Medium (M) With baffle (W) Cross-flow (C) Continuous (C) 1.3 PMWCC1.3 Stagnant (S) Continuous (C) 1.3 PMWSC1.3 Without baffle (O) Stagnant (S) Continuous (C) 1.3 PMOSC1.3 2.2 PMOSC2.2 8000 mL/min High (H)  With baffle (W) Cross-flow (C) Continuous (C) 1.3 PHWCC1.3 Stagnant (S) Continuous (C) 1.3 PHWSC1.3 Without baffle (O) Stagnant (S) Continuous (C) 1.3 PHOSC1.3 2.2 PHOSC2.2 28         B)  D20 Tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Experiment name Slow Pulse Diffuser (SP) 3500 mL/min Low (L) With baffle (W) Cross-flow (C)  Continuous (C)  1.8 SPLWCC1.8 Fast Pulse Diffuser (FP) 3500 mL/min Low (L) With baffle (W) Cross-flow (C)  Continuous (C) 1.8 FPLWCC1.8  Table 3-3. Experimental conditions investigated for filtration tests. A) D14 Tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Slow Pulse Diffuser 3500 mL/min Without baffle  Stagnant Continuous  1.8  Coarse Diffuser  3500 mL/min Without baffle Stagnant  Continuous  1.8 Stagnant  3 sec on/off  1.8 7000 mL/min Without baffle  Stagnant  2 sec on/off  1.8  B) D20 Tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Slow Pulse Diffuser  3500 mL/min With baffle  Cross-flow  Continuous  1.8 1750 mL/min With baffle  Cross-flow  Continuous  1.8 875 mL/min With baffle  Cross-flow  Continuous  1.8 Coarse Diffuser 3500 mL/min With baffle  Cross-flow  Continuous 1.8 1750 mL/min With baffle  Cross-flow  Continuous  1.8 875 mL/min With baffle Cross-flow  Continuous  1.8  For the shear stress measurements, the procedure used was similar to that used for the calibration of the electrochemical probe. The detail procedure is presented in Appendix 8. Refer to Fulton et al. (2011) for more details. For the filtration tests, the experiment procedures were developed and presented in detail in Appendix 9 to ensure the stable performance of the membrane. Note that the performance of new fiber modules was very unstable during the first few filtration tests. This was likely due to the presence of residual glycerin. After approx. 4 filtration tests, the performance became stable. Results for the first few filtration tests with new fiber modules were not used in the analysis.      29  3.4 QA/QC  3.4.1 QA/QC for shear stress measurements  In order to generate reliable data, it was critical to do a proper calibration of the shear probes. Appendices 2 to 7 presents the full processes used to do the shear probe calibration. The shear probes were calibrated by the steady-state calibration approach at the beginning and at the end of the study to make sure the performance of the shear probe was stable. In addition, the voltage step calibration approach was carried out after every four shear stress measurements to monitor the performance of the shear probes. Other factors including temperature, oxygen content and gas flow rate were also monitored, to ensure good performance of the shear probes (Appendices 2 to 8).  For the shear stress measurements in the D14 tank, the experiments were repeated in duplicate, and for those in the D20 tank, the experiments were repeated in triplicate. The repeated data confirmed the reputability of the shear stress measurements. QA/QC results for shear stress measurements are presented in Appendix 10.  3.4.2 QA/QC for filtration tests  In order to generate reliable data, it was critical to do a proper calibration of the pressure transducer. The calibration results of the pressure transducer are presented in Appendix 10. In addition, the particle size distribution of bentonite solution was presented. Other factors including temperature, gas flow rate and integrity test were also monitored to ensure good performance of the filtration tests (Appendix 10).  For the filtration tests in the D14 tank, the experiments were repeated in triplicate, and for those in the D20 tank, the experiments were repeated in duplicate. The repeated data confirmed the reputability of the filtration tests. QA/QC results for filtration tests are presented in Appendix 10.     30  4. RESULTS AND DISCUSSIONS   4.1 Shear Stress Measurements Results  4.1.1 Shear stress measurements results for D14 tank experiments  A total of 54 experiments were performed using the D14 tank. Typical experimental results are presented in Figure 4-1.   Figure 4-1. Shear stress measurements result of slow pulse diffuser, low gas flow rate, without baffle, stagnant, continuous gas flow and 1.3% slackness in D14 tank.  For each experimental condition, two duplicate tests were performed (test 1 and test 2). Cumulative distributions were used to compare the data between replicate tests. As presented in Figure 4-1, the 31  overlap of the cumulative distributions indicated that the results from the two experiments were similar. Results for all replicate experiments are presented in Appendix 11.  Although the relationship between surface shear stress and fouling control remain unclear, statistically significant correlations have been reported between the root mean square (RMS) of the surface shear stress over time induced by gas sparging and the extent of fouling in membrane systems (Yeo et al., 2006; Chan et al., 2011). For this reason, in the following discussion, the results are presented in terms of RMS. The RMS values for all 54 experiments are presented in Table 4-1. The RMS for all replicates are presented in Appendix 12.  Table 4-1. Shear stress measurements results of D14 tank experiments. Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Experiment name Shear stress (Pa) RMS 1 STD 2 Fine Diffuser (F) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.3 FLWCC1.3 1.027 0.090 Stagnant (S) Continuous (C) 1.3 FLWSC1.3 0.797 0.095 6000 mL/min Medium (M) With baffle (W) Cross-flow (C) Continuous (C) 1.3 FMWCC1.3 1.614 0.171 Stagnant (S) Continuous (C) 1.3 FMWSC1.3 1.045 0.186 8000 mL/min High (H) With baffle (W) Cross-flow (C) Continuous (C) 1.3 FHWCC1.3 1.717 0.143 Stagnant (S) Continuous (C) 1.3 FHWSC1.3 1.275 0.123 Coarse Diffuser (C) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.3 CLWCC1.3 1.216 0.226 Fast alter (F) 1.3 CLWCF1.3 0.904 0.106 Slow alter (S) 1.3 CLWCS1.3 0.760 0.083 Stagnant (S) Continuous (C) 1.3 CLWSC1.3 0.909 0.082 Fast alter (F) 1.3 CLWSF1.3 0.715 0.065 Slow alter (S) 1.3 CLWSS.13 0.668 0.075 Without baffle (O) Stagnant (S) Continuous (C) 1.3 CLOSC1.3 0.508 0.082 2.2 CLOSC2.2 0.629 0.065 Fast alter (F) 1.3 CLOSF1.3 0.482 0.095 2.2 CLOSF2.2 0.536 0.043 Slow alter (S) 1.3 CLOSS1.3 0.465 0.095 2.2 CLOSS2.2 0.467 0.094 6000 mL/min Medium (M) With baffle (W) Cross-flow (C) Continuous (C) 1.3 CMWCC1.3 1.231 0.047 Fast alter (F) 1.3 CMWCF1.3 1.035 0.087 Slow alter (S) 1.3 CMWCS1.3 0.958 0.071 Stagnant (S) Continuous (C) 1.3 CMWSC1.3 1.170 0.169 Fast alter (F) 1.3 CMWSF1.3 0.963 0.090 Slow alter (S) 1.3 CMWSS1.3 1.014 0.129 Without baffle (O) Stagnant (S) Continuous (C) 1.3 CMOSC1.3 0.589 0.093 2.2 CMOSC2.2 0.701 0.118 Fast alter (F) 1.3 CMOSF1.3 0.585 0.094 2.2 CMOSF2.2 0.642 0.023 Slow alter (S) 1.3 CMOSS1.3 0.584 0.139 2.2 CMOSS2.2 0.631 0.109 8000 mL/min High (H) With baffle (W) Cross-flow (C) Continuous (C) 1.3 CHWCC1.3 1.430 0.074 Fast alter (F) 1.3 CHWCF1.3 1.265 0.075 Slow alter (S) 1.3 CHWCS1.3 1.139 0.019 Stagnant (S) Continuous (C) 1.3 CHWSC1.3 1.310 0.191 Fast alter (F) 1.3 CHWSF1.3 1.058 0.194 32   Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Experiment name Shear stress (Pa) RMS 1 STD 2 Coarse Diffuser (C) 8000 mL/min High (H) With baffle (W) Stagnant (S) Slow alter (S) 1.3 CHWSS1.3 1.168 0.122 Without baffle (O) Stagnant (S) Continuous (C) 1.3 CHOSC1.3 0.755 0.099 2.2 CHOSC2.2 0.814 0.060 Fast alter (F) 1.3 CHOSF1.3 0.708 0.115 2.2 CHOSF2.2 0.720 0.090 Slow alter (S) 1.3 CHOSS1.3 0.605 0.034 2.2 CHOSS2.2 0.714 0.078 Slow Pulse Diffuser (P) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.3 PLWCC1.3 0.997 0.091 Stagnant (S) Continuous (C) 1.3 PLWSC1.3 0.926 0.065 Without baffle (O) Stagnant (S) Continuous (C) 1.3 PLOSC1.3 0.685 0.050 2.2 PLOSC2.2 0.824 0.173 6000 mL/min Medium (M) With baffle (W) Cross-flow (C) Continuous (C) 1.3 PMWCC1.3 1.286 0.099 Stagnant (S) Continuous (C) 1.3 PMWSC1.3 1.002 0.188 Without baffle (O) Stagnant (S) Continuous (C) 1.3 PMOSC1.3 0.840 0.030 2.2 PMOSC2.2 0.910 0.041 8000 mL/min High (H)  With baffle (W) Cross-flow (C) Continuous (C) 1.3 PHWCC1.3 1.464 0.035 Stagnant (S) Continuous (C) 1.3 PHWSC1.3 1.135 0.139 Without baffle (O) Stagnant (S) Continuous (C) 1.3 PHOSC1.3 1.013 0.053 2.2 PHOSC2.2 0.857 0.095 1 RMS value presented is the average from the four experiments (duplicates from Probe 1 and 2 ). 2 STD is calculated from the four RMS mentioned in 1 .  4.1.2 Shear stress measurements results for D20 tank experiments  2 experiments were performed using the D20 tank. Experiment results are presented in Figure 4-2 and 4- 3. For each experimental condition, three replicate tests were performed (test 1, test 2 and test 3). Cumulative distributions were used to compare the data between replicate tests. As presented in Figure 4-2 and 4-3, the overlap of the cumulative distributions indicated that the results from the replicate experiments were similar.  The RMS values for the 2 experiments are presented in Table 4-1. The RMS for all replicates are presented in Appendix 12. 33   Figure 4-2. Shear stress measurements result of slow pulse diffuser, 3500 mL/min gas flow rate, with baffle, cross-flow, continuous gas flow and 1.8% slackness in D20 tank. 34   Figure 4-3. Shear stress measurements result of fast pulse diffuser, 3500 mL/min gas flow rate, with baffle, cross-flow, continuous gas flow and 1.8% slackness in D20 tank. 35  Table 4-2. Shear stress measurements results of D20 tank experiments. Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Experiment name Shear stress (Pa) RMS 1 STD 2 Slow Pulse Diffuser (SP) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.8 SPLWCC1.8 0.916 0.072 Fast Pulse Diffuser (FP) 3500 mL/min Low (L) With baffle (W) Cross-flow (C) Continuous (C) 1.8 FPLWCC1.8 0.883 0.112 1 RMS value presented is the average from the six experiments (replicates from Probe 1 and 2). 2 STD is calculated from the four RMS mentioned in 1 .  4.2 Shear Stress Measurements Discussions  In the following sections, the effects of the different experimental conditions investigated on the RMS are compared. 4.2.1 Effect of diffuser type  4.2.1.1 Fine bubble diffuser vs. Coarse bubble diffuser A total of 6 experimental condition sets were considered for which fine bubble and coarse bubble diffusers were used and all other experimental conditions were similar (Table 4-3). The data presented in Table 4-3 is summarized in Figure 4-4.  Table 4-3. Fine bubble diffuser vs. coarse bubble diffuser. Series Number Fine Bubble Shear stress (Pa) Coarse bubble Shear stress (Pa) RMS STD RMS STD 1 FLWCC1.3 1.027 0.090 CLWCC1.3 1.216 0.226 2 FLWSC1.3 0.797 0.095 CLWSC1.3 0.909 0.082 3 FMWCC1.3 1.614 0.171 CMWCC1.3 1.231 0.047 4 FMWSC1.3 1.045 0.186 CMWSC1.3 1.170 0.169 5 FHWCC1.3 1.717 0.143 CHWCC1.3 1.430 0.074 6 FHWSC1.3 1.275 0.123 CHWSC1.3 1.310 0.191 (Note: Data presented is from Table 4-1 and/or Table 4-2) 36   Figure 4-4. Fine bubble diffuser vs. coarse bubble diffuser (Error bar corresponds to the STD).  No consistent and statistically significant difference was observed between the RMS for fine bubble and coarse bubble diffusers. However, in general, the RMS was higher for the fine bubble diffuser than for the coarse bubble diffuser for conditions with the cross-flow and higher sparging rates. It is likely that the greater number of fine bubbles than coarse bubbles at a given gas sparging rate resulted in a higher entrained cross-flow velocities. However, the effect of the speed of cross-flow velocity on the RMS was not investigated as part of the present study.  4.2.1.2 Coarse bubble diffuser vs. Pulse bubble diffuser A total of 12 experimental condition sets were considered for which coarse bubble and pulse bubble diffusers were used and all other experimental conditions were similar (Table 4-4). The data presented in Table 4-4 is summarized in Figure 4-5.  Table 4-4. Pulse bubble diffuser vs. coarse bubble diffuser. Series Number Pulse Bubble Shear stress (Pa) Coarse bubble Shear stress (Pa) RMS STD RMS STD 1 PLWCC1.3 0.997 0.091 CLWCC1.3 1.216 0.226 2 PLWSC1.3 0.926 0.065 CLWSC1.3 0.909 0.082 3 PLOSC1.3 0.685 0.050 CLOSC1.3 0.508 0.082 4 PLOSC2.2 0.824 0.173 CLOSC2.2 0.629 0.065 5 PMWCC1.3 1.286 0.099 CMWCC1.3 1.231 0.047 6 PMWSC1.3 1.002 0.188 CMWSC1.3 1.170 0.169 7 PMOSC1.3 0.840 0.030 CMOSC1.3 0.629 0.093 8 PMOSC2.2 0.910 0.041 CMOSC2.2 0.701 0.118 9 PHWCC1.3 1.464 0.035 CHWCC1.3 1.430 0.074 10 PHWSC1.3 1.135 0.139 CHWSC1.3 1.310 0.191 11 PHOSC1.3 1.013 0.053 CHOSC1.3 0.755 0.099 12 PHOSC2.2 0.857 0.095 CHOSC2.2 0.814 0.060 (Note: Data presented is from Table 4-1 and/or Table 4-2) 37   Figure 4-5. Pulse bubble diffuser vs. coarse bubble diffuser (Error bar corresponds to the STD).  No consistent and statistically significant difference was observed between the RMS for coarse bubble and pulse bubble diffusers. However, in general, the RMS was higher for the pulse bubble diffuser than for the coarse bubble diffuser when the system tank was operated with an unbaffled configuration (Figure 3-13c). This difference is likely due to the fact that without a baffle, the coarse bubbles were not confined to a zone that was close to the fibers and therefore could rise in the tank without contacting the fibers. On the other hand, the pulse bubbles were generally always in contact with the fibers. Typical results for coarse bubble and pulse bubble diffusers for the unbaffled system tank are presented in Figure 4-6.                                                  (a)                                                                         (b) Figure 4-6. (a) Probe1 test 1 shear profile of CMOSC1.3. (b) Probe1 test 1 shear profile of PMOSC1.3. 38  4.2.1.3 Slow pulse diffuser vs. Fast pulse diffuser Only one experimental condition set was considered for which slow pulse and fast pulse diffusers were used and all other experimental conditions were similar (Table 4-5). The data presented in Table 4-5 is summarized in Figure 4-7.  Table 4-5. Slow pulse diffuser vs. fast pulse diffuser. Series Number Slow Pulse Diffuser Shear stress (Pa) Fast pulse diffuser Shear stress (Pa) RMS STD RMS STD 1 SPLWCC1.8 0.916 0.072 FPLWCC1.8 0.883 0.112 (Note: Data presented is from Table 4-1 and/or Table 4-2)   Figure 4-7. Slow pulse diffuser vs. fast pulse diffuser (Error bar corresponds to the STD).  No consistent and statistically significant difference was observed between the RMS for slow pulse and fast pulse diffusers. Ongoing research by others is investigating the effect of pulse bubble frequency on the RMS values.  4.2.2 Effect of system tank configuration  4.2.2.1 Baffled system tank v.s. Unbaffled system tank A total of 12 experimental condition sets were considered for which baffled and unbaffled system tanks were used and all other experimental conditions were similar (Table 4-6). The data presented in Table 4- 6 is summarized in Figure 4-8. For each series, all experimental conditions other than the tank setup are similar.      39  Table 4-6. Baffled vs. unbaffled system tank. Series Number Baffle Shear stress (Pa) Unbaffle Shear stress (Pa) RMS STD RMS STD 1 CLWSC1.3 0.909 0.082 CLOSC1.3 0.508 0.082 2 CLWSF1.3 0.715 0.065 CLOSF1.3 0.482 0.095 3 CLWSS1.3 0.668 0.075 CLOSS1.3 0.465 0.095 4 CMWSC1.3 1.170 0.169 CMOSC1.3 0.629 0.093 5 CMWSF1.3 0.963 0.090 CMOSF1.3 0.585 0.094 6 CMWSS1.3 1.014 0.129 CMOSS1.3 0.584 0.139 7 CHWSC1.3 1.310 0.191 CHOSC1.3 0.755 0.099 8 CHWSF1.3 1.058 0.194 CHOSF1.3 0.708 0.115 9 CHWSS1.3 1.168 0.122 CHOSS1.3 0.605 0.034 10 PLWSC1.3 0.926 0.065 PLOSC1.3 0.685 0.050 11 PMWSC1.3 1.002 0.188 PMOSC1.3 0.840 0.030 12 PHWSC1.3 1.135 0.139 PHOSC1.3 1.013 0.053 (Note: Data presented is from Table 4-1 and/or Table 4-2)   Figure 4-8. Baffle vs. unbaffle system tank (Error bar corresponds to the STD).  A consistant and generally statistically significant difference was observed between the RMS for baffled and unbaffled conditions. This was expected because the baffle confined thte sparged bubbles to a zone closer to the fibers, promoting greater contact between the bubbles and fibers.  4.2.2.2 Cross-flow vs. Stagnant A total of 16 experimental condition sets were considered for which cross-flow and stagnant bulk liquid patterns were used and all other experimental conditions were similar (Table 4-7). The data presented in Table 4-7 is summarized in Figure 4-9.   40  Table 4-7. Cross-flow vs. stagnant. Series Number Cross-flow Shear stress (Pa) Stagnant Shear stress (Pa) RMS STD RMS STD 1 FLWCC1.3 1.027 0.090 FLWSC1.3 0.797 0.095 2 FMWCC1.3 1.614 0.171 FMWSC1.3 1.045 0.186 3 FHWCC1.3 1.717 0.143 FHWSC.13 1.275 0.123 4 CLWCC1.3 1.216 0.226 CLWSC1.3 0.909 0.082 5 CLWCF1.3 0.904 0.106 CLWSF1.3 0.715 0.065 6 CLWCS1.3 0.760 0.083 CLWSS1.3 0.668 0.075 7 CMWCC1.3 1.231 0.047 CMWSC1.3 1.170 0.169 8 CMWCF1.3 1.035 0.087 CMWSF1.3 0.963 0.090 9 CMWCS1.3 0.958 0.071 CMWSS1.3 1.014 0.129 10 CHWCC1.3 1.430 0.074 CHWSC1.3 1.310 0.191 11 CHWCF1.3 1.265 0.075 CHWSF1.3 1.058 0.194 12 CHWCS.13 1.139 0.019 CHWSS1.3 1.168 0.122 13 PLWCC1.3 0.997 0.091 PLWSC1.3 0.926 0.065 14 PMWCC1.3 1.286 0.099 PMWSC1.3 1.002 0.188 15 PHWCC1.3 1.464 0.035 PHWSC1.3 1.135 0.139 16 1 SPLWCC1.8 0.916 0.072 PLOSC2.2 0.824 0.173 (Note: Data presented is from Table 4-1 and/or Table 4-2) 1 This condition set compared the different conditions between D14 and D20 tank. As D20 tank has a 14 cm diameter baffle, the effect of baffled (W) or unbaffled (O) tank setup is considered as similar.   Figure 4-9. Cross-flow vs. stagnant (Error bar corresponds to the STD).  No consistent and statistically significant difference was observed between the RMS for cross-flow and stagnant bulk liquid. However, in general, the RMS was higher for the cross-flow than stagnant bulk liquid in most of the conditions. It is likely that, although the shear stresses generated by cross-flow are 41  relatively small compare to those generate by gas bubbles, they still contribute to the overall magnitude of the shear stresses.  4.2.3 Effect of gas flow rate  A total of 18 experimental condition sets were considered for which low, medium and high gas flow rates were used and all other experimental conditions were similar (Table 4-8). The data presented in Table 4-8 is summarized in Figure 4-10.  Table 4-8. Gas flow rate: low vs. medium vs. high. Series Number Low Shear stress (Pa) Medium Shear stress (Pa) High Shear stress (Pa) RMS STD RMS STD RMS STD 1 FLWCC1.3 1.027 0.090 FMWCC1.3 1.614 0.171 FHWCC1.3 1.717 0.143 2 FLWSC1.3 0.797 0.095 FMWSC1.3 1.045 0.186 FHWSC.13 1.275 0.123 3 CLWCC1.3 1.216 0.226 CMWCC1.3 1.231 0.047 CHWCC1.3 1.430 0.074 4 CLWCF1.3 0.904 0.106 CMWCF1.3 1.035 0.087 CHWCF1.3 1.265 0.075 5 CLWCS1.3 0.760 0.083 CMWCS1.3 0.958 0.071 CHWCS.13 1.139 0.019 6 CLWSC1.3 0.909 0.082 CMWSC1.3 1.170 0.169 CHWSC1.3 1.310 0.191 7 CLWSF1.3 0.715 0.065 CMWSF1.3 0.963 0.090 CHWSF1.3 1.058 0.194 8 CLWSS1.3 0.668 0.075 CMWSS1.3 1.014 0.129 CHWSS1.3 1.168 0.122 9 CLOSC1.3 0.508 0.082 CMOSC1.3 0.629 0.093 CHOSC1.3 0.755 0.099 10 CLOSC2.2 0.629 0.065 CMOSC2.2 0.701 0.118 CHOSC2.2 0.814 0.060 11 CLOSF1.3 0.482 0.095 CMOSF1.3 0.585 0.094 CHOSF1.3 0.708 0.115 12 CLOSF2.2 0.536 0.043 CMOSF2.2 0.642 0.023 CHOSF2.2 0.720 0.090 13 CLOSS1.3 0.465 0.095 CMOSS1.3 0.584 0.139 CHOSS1.3 0.605 0.034 14 CLOSS2.2 0.467 0.094 CMOSS2.2 0.631 0.109 CHOSS2.2 0.714 0.078 15 PLWCC1.3 0.997 0.091 PMWCC1.3 1.286 0.099 PHWCC1.3 1.464 0.035 16 PLWSC1.3 0.926 0.065 PMWSC1.3 1.002 0.188 PHWSC1.3 1.135 0.139 17 PLOSC1.3 0.685 0.050 PMOSC1.3 0.840 0.030 PHOSC1.3 1.013 0.053 18 PLOSC2.2 0.824 0.173 PMOSC2.2 0.910 0.041 PHOSC2.2 0.857 0.095 (Note: Data presented is from Table 4-1 and/or Table 4-2)  42   Figure 4-10. Gas flow rate: low vs. medium vs. high (Error bar corresponds to the STD).  No consistently and statistically significant difference was observed between the RMS for low, medium and high gas flow rate. However, for all conditions investigated, the RMS was generally highest for the high gas flow rate and lowest for the low gas sparging rate.  4.2.4 Effect of gas flow pattern  A total of 12 experimental condition sets were considered for which continuous, fast intermittent and slow intermittent were used and all other experimental conditions were similar (Table 4-9). The data presented in Table 4-9 is summarized in Figure 4-11.  Table 4-9. Gas flow pattern: continuous vs. fast intermittent vs. slow intermittent. Series Number Continuous Shear stress (Pa) Fast intermittent Shear stress (Pa) Slow intermittent Shear stress (Pa) RMS STD RMS STD RMS STD 1 CLWCC1.3 1.216 0.226 CLWCF1.3 0.904 0.106 CLWCS1.3 0.760 0.083 2 CLWSC1.3 0.909 0.082 CLWSF1.3 0.715 0.065 CLWSS1.3 0.668 0.075 3 CLOSC1.3 0.508 0.082 CLOSF1.3 0.482 0.095 CLOSS1.3 0.465 0.095 4 CLOSC2.2 0.629 0.065 CLOSF2.2 0.536 0.043 CLOSS2.2 0.467 0.094 5 CMWCC1.3 1.231 0.047 CMWCF1.3 1.035 0.087 CMWCS1.3 0.958 0.071 6 CMWSC1.3 1.170 0.169 CMWSF1.3 0.963 0.090 CMWSS1.3 1.014 0.129 7 CMOSC1.3 0.629 0.093 CMOSF1.3 0.585 0.094 CMOSS1.3 0.584 0.139 8 CMOSC2.2 0.701 0.118 CMOSF2.2 0.642 0.023 CMOSS2.2 0.631 0.109 9 CHWCC1.3 1.430 0.074 CHWCF1.3 1.265 0.075 CHWCS.13 1.139 0.019 10 CHWSC1.3 1.310 0.191 CHWSF1.3 1.058 0.194 CHWSS1.3 1.168 0.122 11 CHOSC1.3 0.755 0.099 CHOSF1.3 0.708 0.115 CHOSS1.3 0.605 0.034 12 CHOSC2.2 0.814 0.060 CHOSF2.2 0.720 0.090 CHOSS2.2 0.714 0.078 (Note: Data presented is from Table 4-1 and/or Table 4-2) 43   Figure 4-11. Gas flow pattern: continuous vs. fast intermittent vs. slow intermittent (Error bar corresponds to the STD).  No consistent and statistically significant difference was observed between the RMS for continuous, fast intermittent and slow intermittent gas flow pattern. However, in general, the RMS was higher for continuous than intermittent (fast and slow) gas flow pattern. This was expected since, at a given gas flow rate, because sparging flow is alterated between on and off, only half of the volume of gas is added to the system for intermittent compared to continuous gas sparging.  4.2.5 Effect of bundle slackness  A total of 12 experimental conditions were considered for which 1.3% and 2.2% bundle slackness were used and all other experimental conditions were similar (Table 4-10). The data presented in Table 4-10 is summarized in Figure 4-12.          44  Table 4-10. Bundle slackness: 1.3% vs. 2.2%. Series Number 1.3% Shear stress (Pa) 2.2% Shear stress (Pa) RMS STD RMS STD 1 CLOSC1.3 0.508 0.082 CLOSC2.2 0.629 0.065 2 CLOSF1.3 0.482 0.095 CLOSF2.2 0.536 0.043 3 CLOSS1.3 0.465 0.095 CLOSS2.2 0.467 0.094 4 CMOSC1.3 0.629 0.093 CMOSC2.2 0.701 0.118 5 CMOSF1.3 0.585 0.094 CMOSF2.2 0.642 0.023 6 CMOSS1.3 0.584 0.139 CMOSS2.2 0.631 0.109 7 CHOSC1.3 0.755 0.099 CHOSC2.2 0.814 0.060 8 CHOSF1.3 0.708 0.115 CHOSF2.2 0.720 0.090 9 CHOSS1.3 0.605 0.034 CHOSS2.2 0.714 0.078 10 PLOSC1.3 0.685 0.050 PLOSC2.2 0.824 0.173 11 PMOSC1.3 0.840 0.030 PMOSC2.2 0.910 0.041 12 PHOSC1.3 1.013 0.053 PHOSC2.2 0.857 0.095 (Note: Data presented is from Table 4-1 and/or Table 4-2)   Figure 4-12. Bundle slackness: 1.3% vs. 2.2% (Error bar corresponds to the STD).  No consistent and statistically significant difference was observed between the RMS for 1.3 % and 2.2 % bundle slackness. However, in general, the RMS was higher for the module with looser fibers (2.2%), compared to those with tighter fibers (1.3%). This is consistent with results from previous studies that indicate that looser fiber will have higher shear stress than tighter fiber (Fulton et al., 2011).  4.2.6 Summary of observations  1) A fine bubble diffuser is able to create higher shear stress than a coarse bubble diffuser, when baffle was in use, while the pulse bubble diffuser is able to create a higher shear stress than the coarse bubble 45  when the baffle was not in use. The difference between slow pulse and fast pulse diffuser was not significant. 2) The cylindrical baffle in the system tank can significantly increase the shear stress on the membrane surface. Cross-flow liquid, in general, was able to generate higher shear stress than the stagnant liquid. 3) Gas flow rate had a significant effect on the shear stress, the higher the flow rate, the higher the shear stress. Continuous flow can generate higher shear stress than the intermittent flow, due to the fact that only half of the volume of gas is added to the system for intermittent compared to continuous gas sparging. 4) Looser fiber (2.2%) will have a slightly higher shear stress than the tighter fiber (1.3%). However, the results were not statistically significant.  4.3 Filtration Tests Results  A total of 10 different experimental conditions were considered. The results of each filtration test (including replicate) are presented in Appendix 13. The time average (i.e. from replicate experiments) results for each condition are presented in Figures 4-13 and 4-14.            46       (a)                                                                                       (b)     (c)                                                                                        (d) Figure 4-13. Averaged filtration results in D14 tank: (a) Slow pulse diffuser, stagnant bulk liquid and 3500 mL/min continuous gas flow; (b) Coarse bubble diffuser, stagnant bulk liquid and 3500 mL/min continuous gas flow; (c) Coarse bubble diffuser, stagnant bulk liquid and 3500 mL/min 3 sec on 3 sec off gas flow; and (d) Coarse bubble diffuser, stagnant bulk liquid and 7000 mL/min 2 sec on 2 sec off gas flow.          47       (a)                                                                                       (b)                                                     (c)                                                                                        (d)                                                     (e)                                                                                        (f) Figure 4-14. Averaged filtration results in D20 tank: (a) Slow pulse diffuser, cross-flow bulk liquid and 3500 mL/min continuous gas flow; (b) Slow pulse diffuser, cross-flow bulk liquid and 1750 mL/min continuous gas flow; (c) Slow pulse diffuser, cross-flow bulk liquid and 875 mL/min continuous gas flow; (d) Fast pulse diffuser, cross-flow bulk liquid and 3500 mL/min continuous gas flow; (e) Fast pulse diffuser, cross-flow bulk liquid and 1750 mL/min continuous gas flow; and (f) Fast pulse diffuser, cross-flow bulk liquid and 875 mL/min continuous gas flow.  The filtration laws (see section 2.3.2) were fitted to the filtration tests data. The complete blocking, the intermediate blocking and the cake fouling laws could all model the filtration data relatively well (see Appendix 13). However, for some cases, the cake fouling law could model the filtration data more 48  comprehensively than the other blocking models (see Appendix 13). For this reason, in the discussion that follows, the fouling rate for the cake fouling law is used to compare the results from the different filtration tests (Table 4-11).  Table 4-11. Fouling rates of the filtration tests under different conditions. A) In D14 Tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Fouling rate (psi/L) Mean 1 STD 2 Slow Pulse Diffuser 3500 mL/min Without baffle Stagnant Continuous 1.8 0.1966 0.0575 Coarse Diffuser  3500 mL/min Without baffle Stagnant  Continuous  1.8 1.1946 0.1022 Stagnant  3 sec on/off  1.8 1.5298 0.1108 7000 mL/min Without baffle  Stagnant  2 sec on/off 1.8 0.8783 0.1539 B) In D20 Tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Fouling rate (psi/L) Mean 1  STD 2 Slow Pulse Diffuser 3500 mL/min With baffle  Cross-flow Continuous  1.8 0.3088 0.0139 1750 mL/min With baffle  Cross-flow  Continuous  1.8 0.7940 0.0213 875 mL/min With baffle  Cross-flow  Continuous  1.8 1.0826 0.0853 Fast Pulse Diffuser  3500 mL/min With baffle  Cross-flow Continuous  1.8 0.5388 0.0054 1750 mL/min With baffle  Cross-flow  Continuous  1.8 0.7974 0.0457 875 mL/min With baffle  Cross-flow  Continuous  1.8 1.2290 0.0098 1 Mean value presented is the averaged cake fouling constant from replicate tests (Appendix 13). 2 STD is calculated from the replicate mean values mentioned in 1 .   4.4 Filtration Tests Discussions  4.4.1 RMS and fouling rate  For six of the experimental conditions investigated, both shear stress measurements and filtration data are available. The RMS of surface shear stress and the fouling rate for these experimental conditions are summarized in Table 4-12 and Figure 4-15.        49  Table 4-12. The filtration results and the corresponding 1  shear stress measurements results. A) In D14 tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness 2 (%) Fouling rate (psi/L)  Shear stress (Pa)  Mean STD RMS STD Slow Pulse Diffuser 3500 mL/min Without baffle Stagnant Continuous 1.8 0.1966 0.0575 0.824 0.173 Coarse Diffuser  3500 mL/min Without baffle Stagnant  Continuous  1.8 1.1946 0.1022 0.629 0.065 Stagnant  3 sec on/off  1.8 1.5298 0.1108 0.536 0.043 7000 mL/min 3 Without baffle Stagnant  2 sec on/off 4  1.8 0.8783 0.1539 0.720 0.090  B) In D20 tank Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness 2 (%) Fouling rate (psi/L)  Shear stress (Pa)  Mean STD RMS STD Slow Pulse Diffuser 3500 mL/min With baffle Cross- flow Continuous 1.8 0.3088 0.0139 0.916 0.072 Fast Pulse Diffuser 3500 mL/min With baffle Cross- flow Continuous  1.8 0.5388 0.0054 0.883 0.112 (Note: Data presented is from Table 4-11, Table 4-1, and/or Table 4-2) 1 As the conditions in the filtration tests and shear stress measurements were not exactly the same, the shear stress results were selected by the closest conditions as explained in follows. 2 1.8% slackness was only used in filtration tests. The data of 2.2% slackness in shear stress measurements were selected for comparison. 3 7000 mL/min gas flow rate was only used in filtration tests. The data of 8000 mL/min gas flow rate in shear stress measurements was selected for comparison. 4 2 sec on/off gas flow pattern was only used in filtration tests. The data of 3 sec on/off gas flow pattern in shear stress measurement was selected for comparison.  50   Figure 4-15. RMS vs. fouling rate (Error bar corresponds to the STD).  From Figure 4-15, it can be observed, that when the RMS values increases, the fouling rate decreases. The results are consistent with those from previous studies (Yeo et al., 2006; Fulton et al., 2011) and indicate that RMS can be used to provide an estimate of the extent of fouling control that can be achieved when inducing a given amount of shear stress using gas sparging.  4.4.2 Effect of diffuser type  4.4.2.1 Coarse bubble diffuser vs. Pulse bubble diffuser The RMS and fouling rate were obtained for a total of 4 experimental conditions, for which coarse bubble diffuser and pulse bubble diffuser were used (Table 4-13). The average filtration results of replicate experiments for these experimental conditions are summarized in Figure 4-16.  Table 4-13. The RMS and filtration results of foarse bubble diffuser vs. pulse bubble diffuser. Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Fouling rate (psi/L)  Shear stress (Pa)  Mean STD RMS STD Coarse Diffuser 3500 mL/min Without baffle Stagnant Continuous 1.8 1.1946 0.1022 0.629 0.065 Slow Pulse Diffuser 3500 mL/min Without baffle Stagnant Continuous 1.8 0.1966 0.0575 0.824 0.173 875 mL/min With baffle Cross- flow Continuous 1.8 1.0826 0.0853 - - Fast Pulse Diffuser 875 mL/min With baffle Cross- flow Continuous  1.8 1.2290 0.0098 - - (Note: Data presented is from Table 4-11, Table 4-1, and/or Table 4-2) 51   (a)  (b) Figure 4-16. Filtration results of coarse bubble diffuser vs. pulse bubble diffuser: (a) Conditions with the same gas flow rate. (b) Conditions with the similar fouling rate.  For a given gas flow rate, pulse sparging resulted in a 10 fold reduction in the rate of fouling, compared to the rate of fouling achieved with coarse bubble sparging (Table 4-13 and Figure 4-16a). To achieve a given rate of fouling control, Approximately 25% of the gas flow rate required for continuous sparging was needed for pulse sparging (Table 4-13 and Figure 4-16b). These results clearly indicate that a significantly lower gas flow rate is required to achieve a given fouling rate with pulse sparging, compared to continuous coarse bubble sparging.     52  4.4.2.2 Slow pulse diffuser vs. Fast pulse diffuser The RMS and fouling rate were obtained for a total of 6 experimental conditions, for which slow pulse diffuser and fast pulse diffuser were used (Table 4-14). The average filtration results of replicate experiments for these experimental conditions are summarized in Figure 4-17.  Table 4-14. The RMS and filtration results of slow pulse diffuser vs. fast pulse diffuser. Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Fouling rate (psi/L)  Shear stress (Pa)  Mean STD RMS STD Slow Pulse Diffuser 3500 mL/min With baffle Cross-flow Continuous 1.8 0.3088 0.0139 0.916 0.072 1750 mL/min With baffle  Cross-flow  Continuous  1.8 0.7940 0.0213 - - 875 mL/min With baffle  Cross-flow  Continuous  1.8 1.0826 0.0853 - - Fast Pulse Diffuser 3500 mL/min With baffle  Cross-flow Continuous  1.8 0.5388 0.0054 0.883 0.112 1750 mL/min With baffle  Cross-flow  Continuous  1.8 0.7974 0.0457 - - 875 mL/min With baffle  Cross-flow  Continuous  1.8 1.2290 0.0098 - - (Note: Data presented is from Table 4-11, Table 4-1, and/or Table 4-2)  Figure 4-17. Filtration results of slow pulse diffuser vs. fast pulse diffuser.  For a relatively high gas flow rate (i.e. 3500 mL/min), the fouling rate for the fast pulse diffuser was about 44% higher than that for the slow pulse diffuser. This was expected because the size of the bubble generated by the slow pulse bubble diffuser was approximately twice the size of that generated by the fast bubble diffuser. For lower gas flow rates (i.e. 1750 and 875 mL/min), no significant difference was observed between the slow and fast pulse diffusers. These results indicate that, in addition to the 53  magnitude of the shear stress, the frequency of shear events also likely affect fouling control. Further research is required to confirm this hypothesis.  4.4.3 Effect of bulk liquid flow pattern  The RMS and fouling rate were obtained for a total of 2 experimental conditions, for which stagnant and cross-flow bulk liquid flow pattern were used (Table 4-15). The average filtration results of replicate experiments for these experimental conditions are summarized in Figure 4-18.  Table 4-15. The RMS and filtration results of stagnant vs. cross-flow. Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Fouling rate (psi/L)  Shear stress (Pa)  Mean STD RMS STD Slow Pulse Diffuser 3500 mL/min Without baffle Stagnant Continuous 1.8 0.1966 0.0575 0.824 0.173 3500 mL/min With baffle Cross-flow Continuous 1.8 0.3088 0.0139 0.916 0.072 (Note: Data presented is from Table 4-11, Table 4-1, and/or Table 4-2)  Figure 4-18. Filtration results of stagnant vs. cross-flow.  For a given sparging flow rate, the fouling rate for stagnant bulk liquid was slightly lower than that for a bulk liquid with cross-flow. Future experiments are required to verify the result.     54  4.4.4 Effect of gas flow pattern  The RMS and fouling rate were obtained for a total of 4 experimental conditions for which continuous and intermittent gas flow pattern were used (Table 4-16). The average filtration results of replicate experiments for these experimental conditions are summarized in Figure 4-18.  Table 4-16. Continuous gas flow pattern vs. intermittent gas flow pattern. Diffuser type Gas flow rate Tank setup Bulk flow pattern Gas flow pattern Slackness (%) Fouling rate (psi/L)  Shear stress (Pa) Mean STD RMS STD Coarse Diffuser  3500 mL/min Without baffle Stagnant  Continuous  1.8 1.1946 0.1022 0.629 0.065 Stagnant  3 sec on/off  1.8 1.5298 0.1108 0.536 0.043 7000 mL/min Without baffle Stagnant  2 sec on/off 1.8 0.8783 0.1539 0.720 0.090 Slow Pulse Diffuser 3500 mL/min Without baffle Stagnant Continuous 1.8 0.1966 0.0575 0.824 0.173 (Note: Data presented is from Table 4-11, Table 4-1, and/or Table 4-2)   Figure 4-19. Filtration results of continuous gas flow pattern vs. intermittent gas flow pattern.  For a given gas flow rate, continuous coarse bubble sparging was significantly better at fouling control than intermittent coarse bubble sparging. However, since the gas flow rate is added during the on cycle for intermittent sparging, the amount (i.e. volume) of gas delivered to the system is approximatly half of the amount delivered when using continuous sparging. When the same amount of gas is delivered, better fouling control was achieved with intermttent coarse bubble sparging (7000 mL/min with flow turn off 50% of time) than conitnuous coarse bubble sparging (3500 mL/min continuously). The pulse diffuser, 55  which can be considered as an extreme intermittent sparging device because air is actually only added periodically when the large bubbles are released, resulted in better fouling control than either continuous coarse bubble or intermittent coarse bubble sparging. These results clearly indicate that for a given volume of sparged gas applied, better fouling control can be achieved with intermittent sparging than continous sparging.  4.4.5 Summary of observations  1) The RMS shear stress had a strong correlation with the fouling rate, which can be used to provide an estimate of the extent of fouling control. 2) Pulse sparging had a very significant reduction in the rate of fouling, compared to the rate of fouling achieved with coarse bubble sparging. A significantly lower gas flow rate is required to achieve a given fouling rate with pulse sparging, compared to continuous coarse bubble sparging. 3) Slow pulse diffuser had a better fouling control than the fast pulse diffuser, when the gas flow rate was relatively high; however, when the flow rate reduced, the difference between slow pulse and fast pulse diffuser was reduced. The frequency of shear events also likely affected fouling control. 4) For a given sparging flow rate, the fouling rate for stagnant bulk liquid was slightly lower than that for a bulk liquid with cross-flow. 5) For a given volume of sparged gas applied, better fouling control can be achieved with intermittent sparging than continuous sparging.            56  5. CONCLUSIONS   5.1 Overall Conclusions  In this thesis, the relationship between the shear stress and the fouling control was investigated in a comprehensive manner. Shear stresses induced by different gas sparging conditions were measured and the effect of these sparging conditions on membrane fouling was quantified. The following are the main conclusions from this study.  Shear stress measurements For the six sparging conditions used in the shear stress measurements, flow rate and tank setup can significantly increase the shear stress on the membrane surface. The ideas of these two conditions were similar, which are increasing the volume of gas close to the membrane fibers (by either increasing sparging rate or by adding baffle to concentrate the gas) to promote shear. However, these two conditions required high energy input and high capital cost, which is not preferred in industrial applications.  For the rest of the four conditions, diffuser type, bulk liquid flow pattern, gas flow pattern and bundle slackness are more preferable to optimize the gas sparging, since they don’t required much capital and operational cost. The results indicated that: 1) The pulse bubble diffuser presented to be a better sparger than coarse bubble diffuser when a baffle was not in use. 2) Cross-flow liquid in general was able to generate higher shear stress than the stagnant liquid. 3) Continuous flow can generate higher shear stress than the intermittent flow due to the fact that only half of the volume of gas is added to the system for intermittent, compared to continuous gas sparging. 4) Looser fiber (2.2%) will have a slightly higher shear stress than the tighter fiber (1.3%). However, the results were not statistically significant.     57  Filtration tests The filtration tests were designed to firstly examine the relationship between the shear stress and fouling rate and also to investigate the effect of different sparging conditions. Due to this reason, the tank setup and bundle slackness were fixed as constants in the filtration tests. The results indicate that: 1) The RMS shear stress had a strong correlation with the fouling rate, which can be used to provide an estimate of the extent of fouling control. 2) Pulse sparging was overall better than the conventional coarse sparging. A 10 fold reduction was observed when applying the same gas flow to pulse diffuser than the coarse diffuser. Only one quarter of the gas was required to achieve the same fouling control when using a pulse diffuser. 3) For a given sparging flow rate, the fouling rate for stagnant bulk liquid was slightly lower than that for a bulk liquid with cross-flow. 4) For a given volume of sparged gas applied, better fouling control can be achieved with intermittent sparging than continuous sparging.   5.2 Contribution and Engineering Significance  This research successfully created the variable shear conditions identified by Chan et al. (2011) using the gas sparging. The different shear profiled was recorded by the electrochemical shear method. In addition, filtration tests examine the different sparging conditions to the fouling rate to further confirm which sparging approach can actually improve the fouling control in a submerged hollow fiber system.  Pulse bubble diffuser proved to be a very effective way to introduce the gas into the sparging system. A 10 fold reduction was observed when applying the same gas flow to pulse diffuser than the coarse diffuser. Only one quarter of the gas was required to achieve the same fouling control when using a pulse diffuser. By using the pulse diffuser, the energy input of gas sparging can be significantly reduced. In addition, the backwash cycle can be prolonged during the filtration, which can also significantly reduce the operational costs.      58  REFERENCES   Belfort, G., Davis, R.H. and Zydney, A.L. (1994) The behavior of suspensions and macromolecular solutions in cross-flow microfiltration. Journal of Membrane Science 96(1-2), 1-58 Bellara, S.R., Cui, Z.F. and Pepper, D.S. (1996) Gas sparging to enhance permeate flux in ultrafiltration using hollow fibre membranes. Journal of Membrane Science 121(2), 175-184 Berger, F.P. and Ziai, A. (1983) Optimisation of experimental conditions for electrochemical mass transfer measurements. Chemical Engineering Research and Design 61a, 377 – 382 Bérubé, P.R., Afonso, G., Taghipour, F. and Chan, C.C.V. (2006) Quantifying the shear at the surface of submerged hollow fiber membranes. Journal of Membrane Science 279(1-2), 495-505 Bérubé, P.R. and Lei, E. (2006) The effect of hydrodynamic conditions and system configurations on the permeate flux in a submerged hollow fiber membrane system. Journal of Membrane Science 271(1- 2), 29-37 Busch, J., Cruse, A. and Marquardt. W. (2007) Modeling submerged hollow-fiber membrane filtration for wastewater treatment. Journal of Membrane Science 288(1-2), 94-111 Cabassud, C., Laborie, S. and Laine, J.M. (1997) How slug flow can improve ultrafiltration flux in organic hollow fibres. Journal of Membrane Science 128(1), 93-101 Chan, C.C.V.,  Bérubé, P.R. and Hall, E.R. (2011) Relationship between types of surface shear stress profiles and membrane fouling. Water Research 45, 6403-6416 Chan, C.C.V., Berube, P.R. and Hall, E.R. (2007) Shear profiles inside gas sparged submerged hollow fiber membrane modules. Journal of Membrane Science 297 (1-2), 104-120 Chang, S. and Fane, A.G. (2000) Characteristics of microfiltration of suspensions with inter-fiber twophase flow. Journal of Chemical Technology and Biotechnology 75(7), 533-540 Chang, I.S., Le Clech, P., Jefferson, B. and Judd, S. (2002) Membrane fouling in membrane bioreactors for wastewater treatment. Journal of Environmental Engineering-Asce 128(11), 1018-1029 Cheng, T.W., Yeh, H.M. and Gau, C.T. (1998) Enhancement of permeate flux by gas slugs for crossflow ultrafiltration in tubular membrane module. Separation Science and Technology 33(15), 2295-2309 Cui, Z.F., Chang, S. and Fane, A.G. (2003) The use of gas bubbling to enhance membrane processes. Journal of Membrane Science 221(1-2), 1-35 Cui, Z.F. and Wright, K.I.T. (1994) Gas-Liquid 2-Phase Cross-Flow Ultrafiltration of Bsa and Dextran Solutions. Journal of Membrane Science 90(1-2), 183-189 59  Field, R.W., Wu, D., Howell, J.A. and Gupta, B.B. (1995) Critical flux concept for microfiltration fouling. Journal of Membrane Science 100(3), 259-272 Fulton, B.G., Redwood, J., Tourais, M. and J. Berube, P.R. (2011) Distribution of surface shear forces and bubble characteristics in full-scale gas sparged submerged hollow fiber membrane modules. Desalination 281, 128-141 Gaucher, C., L., P., Jaouen, P., Comiti J. and Pruvost, J. (2002) Hydrodynamics study in a plane ultrafiltration module using an electrochemical method and particle image velocimetry visualization. Experiments in Fluids 32(3), 283-293 Govan, A.H., Hewitt, G.F., Owen, D.G, Burnett, G. (1989) Wall shear stress measurements in vertical air-water annular two-phase flow. Journal of Multiphase Flow 15(3), 307-325 Gonsalves, V.E. (1950)   crit ical investigation on the viscose filtration process. Recueil des Travaux Chimiques des Pays-Bas 69(7), 873–903 Green, G., and Belfort, G. (1980) Fouling of ultrafiltration membranes: lateral migration and the particle trajectory model. Desalination 35, 129-147 Hermans, P.H. and Bredée, H.L. (1935) Zur Kenntnis der Filtrationsgesetze. Recueil des Travaux Chimiques des Pays-Bas 54 (9) 680–700 Hermia, J. (1982) Constant pressure blocking filtration laws - application to ower-law non-newtonian fluids. Transactions in Chemical Engineering 60, 183-187 Judd., S. (2006) The MBR Book: Principles and Applications of Membrane Bioreactors for Water and Wastewater Treatment. Elsevier Science Kissinger, P. T. and Heineman, W. R. (1996) Laboratory techniques in electroanalytical chemistry (2nd ed.). Mercel Dekker Inc. New York Koros, W.J., Ma, Y.H., and Shimidzu, T. (1996) Terminology for membranes and membrane processes (IUPAC Recommendation 1996). Journal of Membrane Science 120(2), 149-159 Legrand, J., Dumont, E., Comiti, J. and Fayolle, F. (2000). Diffusion coefficients of ferricyanide ions in polymeric solutions — comparison of different experimental methods. Electrochimica Acta 45(11), 1791-1803 Li, Q.Y., Ghosh, R., Bellara, S.R., Cui, Z.F. and Pepper, D.S. (1998) Enhancement of ultrafiltration by gas sparging with flat sheet membrane modules. Separation and Purification Technology 14(1-3), 79-83 Mercier-Bonin, M., Gesan-Guiziou, G. and Fonade, C. (2003) Application of gas/liquid two-phase flows during crossflow microfiltration of skimmed milk under constant transmembrane pressure conditions. Journal of Membrane Science 218(1-2), 93-105 60  Metcalf and Eddy, Inc. et al. (1991) Wastewater engineering, treatment, disposal, and reuse. McGraw- Hill, New York, NY Metcalf and Eddy, Inc. et al. (2003) Wastewater Engineering Treatment and Reuse (4 th  Edition). McGraw-Hill, New York, NY Nagaoka, H., Kurosaka, M., Shibata, N. and Kobayashi, M. (2006) Effect of bubble flow velocity on drag-force and shear stress working on submerged hollow fibre membrane. Water Science and Technology 54(10), 185-192 Nakoryakov, V. E., Kashinsky, O. N., Burdukov, A. P. and Odnoral, V. P. (1981) Local characteristics of upward gas-liquid flows. International Journal of Multiphase Flow 7(1), 63-81. Nakoryakov, V.E., Kashinsky, O.N. and Kozmenko, B.K. (1986) Experimental study of gas-liquid slug flow in a small diameter vertical pipe. Journal of Multiphase Flow 12(3), 337-355 Owen, G., Bandi, M., Howell, J.A. and Churchouse, S.J. (1995) Economic assessment of membrane processes for water and waste water treatment. Journal of Membrane Science 102, 77-91 Reiss, P.L. and Hanratty, T.J. (1962) Measurement of instantaneous rates of mass transfer to a small sink on a wall. American Institute of Chemical Engineers Journal 8(2), 245-247 Reiss, L. P. and Hanratty, T. J. (1963) An experimental study of the unsteady nature of the viscous sublayer. American Institute of Chemical Engineers Journal 9(2), 154-160 Rode, S., Midoux, N., Latifi, M.A. and Storck, A. (1994) Hydrodynamic and liquid-solid mass transfer mechanisms in packed beds operating in cocurrent gas-liquid downflow: an experimental study using electrochemical shear rate sensors. Chemical Engineering Science 49(9), 1383-1401 Rosant, J.M. (1994) Liquid-wall shear stress in stratified liquid/gas flow. Journal of Applied Electrochemistry 24(7), 612-618 Selman, J.R. and Tobias, C.W. (1978) Mass-transfer measurements by the limiting-current technique. Advances in Chemical Engineering, 10, 211-318 Seyssiecq, I., Ferrasse, J.-H. and Roche, N. (2003) State of the art: rheological characterisation of wastewater treatment sludge, Biochem. Eng. J 16, 41–56 Sobolik, V., Tihon, J., Wein, O. and Wichterle K. (1998) Calibration of electrodiffusion friction probes using a voltage-step transient. Journal of Applied Electrochemistry 28(3), 329-335 Taama, W. M., Plimley, R. E. and Scott, K. (1996) Influence of supporting electrolyte on ferricyanide reduction at a rotating disc electrode. Electrochimica Acta 41(4), 549-551 Ueda, T., Hata, K., Kikuoka, Y. and Seino, O. (1997) Effects of aeration on suction pressure in a submerged membrane bioreactor. Water Research 31(3), 489-494 61  Yeo, A.P.S., Law, A.W.K. and Fane, A.G. (2006) Factors affecting the performance of a submerged hollow fiber bundle. Journal of Membrane Science 280(1-2), 969-982 Zabaras, G., Dukler, A.E. and Moalemmaron, D. (1986) Vertical upward cocurrent gas-liquid annular flow. Aiche Journal 32(5), 829-843 Zydney, A.L. and Colton, C.K. (1986) A concentration polarization model for the filtrate flux in cross- flow microfiltration of particulate suspensions. Chemical Engineering Communications 47(1-3), 1- 21                         62  APPENDIX 1. PRINCIPLE OF ELECTROCHEMICAL SHEAR PROBE   As membrane surface shear stress has been recognized as playing a significant role in fouling reduction within membrane systems, many methods have been used to examine and quantify this parameter. Electrochemical shear probes are not only uniquely well suited for measuring shear stress within two- phase flow systems, but have the additional benefits of rapid response and relatively simple construction. The measurement of the shear stress using electrochemical methods was first performed by Reiss and Hanrraty (1962), and has been used extensively by other researchers to study mass transfer and local hydrodynamic properties in fluidized reactors, two phase pipe flows etc. (Zabaras, 1986; Govan, 1989; Nakoryakov, 1986; Rode, 1994). The principle of the electrochemical method for the shear stress measurement is that the current between cathode and anode in an electrochemical cell depends on the change of flow rate. Usually, in a cell the shear probe is operated as the cathode which is also the controlling electrode to determine the current strength. In the following sections, the principle of electrochemical shear method and the theory using for shear probe measurement and calibration would be discussed.   A1.1 Limiting Current Condition  Figure A1-1 shows the relationship between measured current density and the applied potential in an electrochemical cell. In Zone I, the redox reaction rate at the shear probe is less than the rate of arrival of the reacting species to the shear probe. Increasing applied potential results in an increase in the reaction rate at the shear probe, and therefore an increase in the current density. In Zone II, the reaction rate at the shear probe is greater than the rate of arrival of the reacting species to the shear probe. The reaction rate at the shear probe is so high that the concentration of the reacting species at the shear probe surface is zero. Under this condition, the current is not a function of applied potential. Further increase in the applied potential does not have any effect on the reaction rate. The measured current is dependent only on the diffusion and migration of the reacting species from the bulk solution to the shear probe. This condition is called limiting current condition, or the limiting diffusion condition. In Zone III, the applied 63  potential exceeds the discharge potential of the solvent, causing the secondary reaction (such as hydrogen evolution) at the shear probe, therefore the current increases.   Figure A1-1. Typical current-potential relationship in an electrochemical cell when a potential is applied between two electrodes.  The limiting current condition defined in Zone II is the region where the electrochemical shear method is applicable.  Under the limiting current condition, the magnitude of the current passing through the shear probe is a consequence of just three variables: (1) the diffusion rate of ions to/from the shear probe surface, (2) the shear probe surface area, and (3) the concentration of the active ionic species within the solution (Kissinger and Heineman, 1996). To prevent the reaction at the probe surface from being limited by the reaction rate at its counter-electrode, the counter-electrode (anode) must have a much larger active surface area than the probe (Nakoryakov et al., 1981; Reiss and Hanratty, 1963). In addition, the range of Zone II depends on the Reynolds number in the system. The larger the Reynolds number, the smaller the range of plateau. As a result, limiting current test of the system is required before the electrochemical measurement to make sure that the applied potential is the right range to create limiting current condition.   64  A1.2 The Leveque Equation and Steady-state Calibration  As the measured current is dependent on the probe surface area and it is difficult to fabricate a probe with an area of an exact value, to compare the measured value between two probes or two different experiments, the calibration of the shear probe is essential.  The limiting current (I) can be calculated using the theory of convective diffusion of a single species. If the Péclet number is so high that the effects of longitudinal and transverse diffusion can be neglected, the equation of convective diffusion under steady state can be solved. The resulting Leveque equation to calculate the limiting current I is presented in equation A1-1 (Sobolik et al., 1998).                                                                  I = KLevγ 1/3                                           (A1-1)  where the KLev [A·s 1/3] is the Leveque coefficient and γ [s-1] is the shear rate or velocity gradient on the probe surface. Leveque coefficient can be estimated from the following equation if the shear probe is circular:  KLev = nFAckD 2/3 d -1/3                           (A1-2)  where n is a dimensionless number indicate the amount of electrons involved in reaction, F [96485C·mol -1 ] is the Faraday constant, A [m 2 ] is the area of the shear probe surface, c [mol·m -3 ] is the mol concentration of bulk solution, k [0.862 for circular probe] is a dimensionless represent the shape factor of the shear probe, D [m 2 ·s -1 ] is the diffusion coefficient and d [m] is the diameter of the shear probe. If the electrochemical solution is a Newtonian fluid, applying the Newtonian law of viscosity, the shear rate can be calculated by equation A1-3 and A1-4.  4)-(A1 μ τ γ 3)-(A1                                       γμ δy δv μτ x     where τ [Pa] is the shear stress and μ [Pa·s] is the dynamic viscosity. Combining equations A1-1 to A1-4 we can get the equation A1-5 (Sobolik et al., 1998).                                                                               1/31/3 Lev τμKI                                                 (A1-5)  65  Although the exact probe diameter and area are not known, the KLev can be estimated experimentally for a given probe by inducing a known shear stress and measuring the resulting current through the probe (for a given electrochemical solution and temperature). Knowing KLev, the shear stress can be estimated based on the current measurements through the probe.  A1.3 The Cottrell Equation and Voltage Step Calibration  As an alternative, an in-situ voltage step calibration approach can be used to convert the measured current I to a shear stress. The calibration starts from the zero-current state with the corresponding equilibrium voltage. At t = 0 a constant voltage V is applied which is high enough to guarantee the limiting diffusion regime under the ultimate steady-state condition. If the additional transport (ohmic and faradaic) resistances are negligible at any moment of the transient process, the applied voltage step causes a step change of the surface depolarizer concentration from bulk concentration zero (or a negligible fraction of bulk concentration).The transient current for a very short time after this step, can be fitted to Cottrell equation (Sobolik et al., 1998) to determine KCot (equation A1-6), from which the probe area can be derived using equation A1-7.  1/2 CottK=I                                                (A1-6) KCot = nFAcD 1/2π -1/2                                                        (A1-7)  where KCot is [A·s -1/2 ] the Cottrell coefficient and t [s] is the time after the voltage is applied. Other variables n, F, A, c and D is the same as in equation A1-2. However, the estimated area of the shear probe is highly depends on the current measured in the short time interval so that the noise may have significant effect. Sampling frequency will also affect the results.  Applying the theory and the equation that mentioned above, the electrochemical method can be used to measure the shear stress on the membrane surface, which would be the key parameter of the hydrodynamic conditions in the submerged hollow fiber system.      66  APPENDIX 2. SHEAR PROBE CALIBRATIONS   Before the shear probes were, they were calibrated in a side loop setup. The side loop set up was able to create a controlled hydrodynamic condition, which is required for the steady-state calibration process. This setup has been used by Fulton et al. (2011) for the probe calibration.   A2.1 Structure of Electrochemical Shear Probe  Electrochemical probes have been used extensively by other researchers to study mass transfer and hydrodynamic properties in fluidized reactors. The surface area of a shear probe is usually very small as measurements reflect the average of the shear stresses experienced across the entire probe surface, which for large probes may prevent an accurate reading of the real shear stresses experienced at any specific point (Reiss and Hanratty, 1963). Nickel and platinum have both commonly been used as electrode materials with the ferri-ferrocyanide couple for mass transfer measurements, but results of Taama et al. (1996) indicate that platinum is far superior for this purpose, as it requires less frequent cleaning to maintain its ability to provide accurate readings.  The electrochemical shear probes used in the present study were essentially the same as those used by Fulton et al. (2011), a 0.5 mm diameter round platinum wire sanded flush with the outer surface of a 1.8 mm hollow Teflon tube (see Figure A2-1). Once completed, test fibers were similar in shape and flexibility to Zeeweed-500 PVDF membrane fibers (see Figure A2-2) (Fulton et al., 2011).   Figure A2-1 Top (a) and side (b) views of a shear probe on a test fiber. 67    Figure A2-2. A test fiber shown with a Zeeweed-500 PVDF membrane fiber.   A1.2 Choice of Electrolyte  Ferri-ferrocynaide couple. There are four criteria determining the choice of electrolyte in the electrochemical shear measurement: 1) chemical stability, 2) high solubility, 3) electrode potential different from hydrogen, and 4) low cost (Selman and Tobias, 1978). The ferri-ferrocyanide electrolytic couple is commonly used with nickel or platinum since both metals are sensitive to the present of cyanide ions. The reaction at the working electrode (cathode) is:  Fe(CN)6 3-   +  e -  → Fe(CN)6 4-   As there is no precipitation in the product of the reaction, this electrolytic couple is better than other couple that can produce precipitation on the probe surface, such as copper, which may cause the change of the shear probe surface area.  Concentration of inner salt. One of the assumptions in the limiting current equation is that the current generated by the migration effect due to the applied potential can be eliminated. For this reason, inert 68  salts such as NaCl, KCl are normally added in excess to reduce/eliminate the migration effect. The following equation can be used to determine the concentration of the salt solution required for a 1% migration current relative to the total migration current (Berger and Ziai, 1983):  t salt i mig IC C 2 1 =I   (A1-1)  where Imig is the migration current, It is the total current, Ci is the concentration of reacting species and Csalt is the concentration of inner salt. For example, for a 0.01M ferri and ferrocyanide solution, a concentration of 0.5M of salt solution is sufficient to reduce the migration current to 1% of the total current.  Special considerations during experiment. Special considerations are needed when preparing the electrolyte solution and during experiment, as discussed below,  i. Photochemical decomposition Ferricyanide and ferrocyanide are sensitive to light and will undergo photochemical decomposition when exposed to light: OHCNFe(CN)OHFe(CN) 2 3 5 light 2 4 6      OHHCNOHCN light2 The decomposition byproduct, hydrogen cyanide (HCN), may poison the electro surface (Berger and Ziai, 1983). The decomposition of the ferri/ferrocyanide will also change the measured current and cause the error during the measurement. It is recommended that the solution not be exposed to light, and the solution be stored overnight in opaque container.   ii. Exposure to oxygen Oxygen can oxdize ferricyanide and ferrocyanide if these are exposed to air.  The oxidized ferricyanide and ferrocyanide byproducts may also poison the electrode surface (Berger and Ziai, 1983). In addition, the oxygen in the solution can react with the ferri/ferrocyanide. Isolation of the electrochemical system from air is recommended throughout the experiment and nitrogen gas can use to purge out the oxygen in the solution.  69   iii. Exposure to acid The ferricyanide and ferrocyanide have very low toxicity. However, under strongly acidic conditions, highly toxic hydrogen cyanide gas can be released. This is possible to happen during the preparation of the solution in the lab where the strong acid can accidentally mix with the ferri/ferrocyanide solution. Make sure the absent of strong acid from the chemical throughout the experiment.  iv. Effect of temperature Diffusivity of ferricyanide is sensitive to temperature changes. Berger and Ziai (1983) showed that a 3 C°  temperature increase can result in a 20% increase in the limiting current of the system. It is important to monitor the temperature of the measurement system and keep the time of the measurement short enough to prevent significant change of the temperature.  Rosant (1994) showed that the 0.003 M ferricynaide and 0.006 M ferrocyanide had been successfully applied in electrochemical shear measurements. According to the equation 3-6 by applying the value of ferrocyanide concentration, 0.006 M, the 0.3 M potassium chloride was used as the inner salt. The actual chemical used to provide the ferrocyanide ion was potassium ferrocyanide trihydrate. Solvent used for the experiment is Reverse Osmosis water and three liters of solution was used in total. The physical properties of the electrolytes for the shear probe calibration were shown in Table A2-1.  Table A2-1. Physical properties of the electrolyte solution. Compound Concentration (M) Molecular weight (g/mol) Amount Potassium ferricyanide (K3[Fe(CN)6]) 0.003 329.26 2.936g Potassium ferrocyanide trihydrate (K[Fe(CN)6] 3H2O) 0.006 422.41 7.603g Potassium chloride (KCl) 0.3 74.56 67.104g Reverse Osmosis water - - 3L Other physical properties at 20 o C (Rosant, 1994): Density = 1016 kg/m 3  Viscosity = 0.001 kg/m/s Ferricyanide diffusivity = 7.14 × 10 -10  m 2 /s  70  A1.3 Side Loop System Setup  The side loop membrane setup was built to create the well-controlled shear rate on the shear probe surface. Figure A2-3 and A2-4 presents an overview of the side loop system. There was a Plexiglas cylinder 1.2 m in length with an inner diameter of 0.025 m, within which ferri/ferrocyanide solution could be circulated using a Masterflex peristaltic pump. The flow could be varied from 0 - 3.5 liters/min, though only flow rates of 0.5 - 2.0 liters/min were used (equal to cross-flow velocities of 0.017 - 0.065 m/sec) to ensure laminar flow conditions. The Plexiglas cylinder was constructed such that a single test fiber (with shear probe) could be accurately placed at the centreline of the cylinder, resulting in an annular type geometry. The shear probe was located at the middle of the cylinder during the measurement. Figure A2-5 presents a the front view and the cross section view of the Plexiglas cylinder and the test fiber.  Nitrogen gas was used to purge out the oxygen from the solution before the calibration. A fine bubble diffuser was placed inside the electrolyte reservoir to increase the efficiency of the purging. The dissolve oxygen (DO) meter was used to monitor the temperature and the oxygen content of the electrolyte solution. The electrical circuit in the system is discussed in section A1.4.  71   Figure A2-3. The schematic diagram of the side loop system setup.   Figure A2-4.  Photo of the Plexiglas column used in the side loop setup. 72                                    (a)                                                                              (b) Figure A2-5. (a) Side view of the Plexiglas cylinder with the test fiber and shear probe. (b) Cross section A-A of the Plexiglas cylinder with the test fiber and shear probe.   A1.4 Electrical Circuit  The cathode of the electrochemical system was the shear probe, and the anode was a piece of stainless steel pipe fitting located on one end of the Plexiglas cylinder (Figure A2-3). The surface area of the stainless steel fitting was much larger than the surface area of the probes to insure the limiting diffusion conditions only occured on the shear probe surface. The power supply between the anode and cathode was adjustable. Before the calibration, limiting current test was carried out to make sure that the applied voltage was in Zone II as discussed before. The test result showed that the current didn’t change when voltage increase from 400 mV to 600 mV. As a result, the voltage was set to 500 mV for the calibration test.  During the measurement, the current in the circuit passed through a 100 Ω resistor. The voltage drop across the resistor was measured and amplified (gain =1000). The data acquisition system acquired the signals and recorded real time using a custom Labview application. Data acquisition frequency was adjustable and 1000 Hz acquisition frequency was used for the steady-stay calibration of the two test fibers. Figure A2-6 presents a schematic diagram of the electrical circuit in the side loop setup.  73   Figure A2-6. The schematic diagram of the electrical circuit in the side loop setup.   A1.5 Calibration Procedure and Results  Before the shear stress measurements were taken, the prepared electrolyte solution was purging by nitrogen gas for about 15 minutes until the oxygen saturation of the solution reduced to less than 7% to minimize the effect of oxygen. The two shear probes were thoroughly washed using de-ionized water and Kimwipes® (non-abrasive laboratory-grade paper towel) to ensure their cleanliness. As each time only one test fibre can be placed in the Plexiglas cylinder, the measurement of the actual limiting current was carried out sequentially. In addition, the four different flow rates (0.5, 1.0, 1.5, 2.0 L/min) of the pump were calibrated by a stop watch and a measurement cylinder.  During the measurement, steady-state calibration was carried out first. Four corrected flow rates were used and the data for each flow rate was recorded. Each flow rate was operated for about 1 min to ensure the steady state condition and the sampling period was about 10 seconds. Temperature was closely monitored as it can significantly affect the value of limiting current. Appendix 3 presents the relationship between the temperature and the theoretical coefficient of the limiting current. Appendix 4 presents the detail of the calculations and results. After the steady-state calibration, the voltage-step calibration was also carried out. This was done by switching off the applied voltage first, and then began the signal recording. About 3 seconds later, the applied voltage was switched back on. The measured voltage will have a sudden increase and then decrease to stable value. This transient behavior was recorded for 5 seconds. The conversion between the steady-state calibration coefficient and voltage-step calibration coefficient is presented in Appendix 6. 74  After the measurements, the shear probe was washed using de-ionized water and Kimwipes® immediately. The corrected data was analysed and the average calibration factor for each probe was obtained. Then the two test fibres were ready for the shear measurement in the full length membrane module. The results of steady-state calibration are presented in Appendix 5. The results of the voltage- step calibration were presented in Appendix 7.  Calibration was done again after the shear stress measurements were completed to examine if the properties of the shear probes had changed over time.  As one of the quality control method, voltage-step calibrations were performed in the large scale system.                     75  APPENDIX 3. THE EFFECT OF TEMPERATURE TO LIMITING CURRENT   Temperature will affect the viscosity of the water as well as the diffusion rate of the ion. It is important to correct the temperature effect of the system to avoid error in the experiment. In this thesis, two assumptions were made for correcting the temperature effect. The first one is that viscosity changes linearly with the temperature. As the temperature range of all the experiments is quite small (17-23°C), we assume that viscosity changes linearly in this range. The second one is that the value of the product between the diffusion coefficient and the dynamic viscosity overs the temperature is constant (Legrand et al., 2000),                               T D  constant                       (A3-1) The data from Gaucher et al. (2002) was used as a standard value at 30°C and we assume the solution at 20°C has a same viscosity as water. The shear probe diameter is assumed as 0.0005 m and the concentration is 3 mol/m 3 . The theoretical coefficient values of a ferri/ferrocyanide solution were calculated in the Table A3-1.  Table A3-1. Theoretical coefficient values of the electrochemical solution under different temperatures Temperature (°C) Temperature  (K) Viscosity (Pas) Diffusion Coefficient (m 2 /s) Leveque coefficient KLev Cottrell coefficient KCot 30 303.15 8.1300E-04 8.3600E-10 5.4777E-07 9.2713E-07 29 302.15 8.3190E-04 8.1431E-10 5.3826E-07 9.1502E-07 28 301.15 8.5080E-04 7.9359E-10 5.2908E-07 9.0330E-07 27 300.15 8.6970E-04 7.7376E-10 5.2024E-07 8.9195E-07 26 299.15 8.8860E-04 7.5478E-10 5.1169E-07 8.8094E-07 25 298.15 9.0750E-04 7.3659E-10 5.0344E-07 8.7026E-07 24 297.15 9.2640E-04 7.1914E-10 4.9546E-07 8.5989E-07 23 296.15 9.4530E-04 7.0239E-10 4.8773E-07 8.4982E-07 22 295.15 9.6420E-04 6.8630E-10 4.8025E-07 8.4003E-07 21 294.15 9.8310E-04 6.7083E-10 4.7301E-07 8.3050E-07 20 293.15 1.0020E-03 6.5594E-10 4.6598E-07 8.2123E-07 19 292.15 1.0209E-03 6.4160E-10 4.5917E-07 8.1221E-07 18 291.15 1.0398E-03 6.2778E-10 4.5255E-07 8.0341E-07 17 290.15 1.0587E-03 6.1445E-10 4.4612E-07 7.9484E-07 16 289.15 1.0776E-03 6.0160E-10 4.3988E-07 7.8648E-07  With this table, after we measured the limiting current and the corresponding shear stress, we would be able to calibrate the shear probe. 76  APPENDIX 4. SHEAR RATE IN THE SIDE LOOP SETUP   The flow condition in the side loop column is an annulus flow from bottom to top (Figure A2-5). Figure A4-1 shows the cylindrical coordination for the annulus flow in the side loop column.                   Figure A4-1. Cylindrical coordination for the annulus flow  We assume the flow is parallel with the wall and flow is in steady state, so Vr = Vθ = 0 and 0z/Vz  . Also, for steady, axisymmetric flow, Vz is not a function of θ. So the velocity, Vz, is only a function of the radial position within the tube, that is Vz = Vz(r). Under these conditions, the Navier–Stokes equations become   r p 0                                                            (A4-1)     p r 1 0                                                      (A4-2)                                                            r V r rr 1 g z p 0 z                    (A4-3)  Equation (A4-1) and (A4-2) indicate that pressure is not a function of r or θ, and the pressure is constant at any particular cross section, and the z component of the pressure gradient, z/p  , is not a function of g z θ r Vz 77  r or θ as well. So now we can integrate equation ( 4-3) and treat z/p   as constant in the integration. The result of the integration is:                                        21z CrlnCg z p 4 1 V                               (A4-4) where C1 and C2 are integration constant. By applying the boundary condition that, at the wall ro, the velocity is zero and at the fiber surface ri, the velocity is also zero. The equation will become:                          o i o 2 o 2 i2 o 2 z r r ln r r ln rr rrg z p 4 1 V        (A4-5) Integrating Vz from the r direction; we get the flow rate:                           i o 22 i 2 o4 i 4 o r r ln )rr( rrg z p 8 Q        (A4-6) The shear rate is defined as the change of the velocity, so γ = r Vz   . We can differentiate equation (A4-5) respected to r and get:                             i o 2 o 2 iz r r lnr rr r2g z p 4 1 r V            (A4-7) If we know the flow rate Q, we can use equation (A4-6) to fine the pressure drop g z p     and then calculate the shear rate from equation (A4-7).  nd the Shear rate γ can be calculated as                               i o 22 i 2 o4 i 4 o i o 2 o 2 i r r ln )rr( rr r r lnr rr r2 Q2               (A4-8) In the side loop system, ro = 0.025 m and ri = 0.0018 m. As we only interested in the shear rate at the probe surface, we let the r = ri in equation (A4-8). For a given flow rate Q, we would be able to find out 78  the theoretical γ. In the experiment, four different flow rates were used, 0.5, 1.0, 1.5 and 2 L/min.  s the flow rate is critical to calculate the shear rate, the four flow rates were calibrated by a stop watch and a measurement cylinder. Table A4-1 shows the corresponding shear rate of the given flow rate.  Table A4-1. Shear rate γ under given flow rate. Designed flow rate (m 3 /s) Measured flow rate (m 3 /s) Shear rate (s -1 ) 8.33E-06 8.28E-06 22.09 1.67E-05 1.67E-05 44.44 2.50E-05 2.46E-05 65.76 3.33E-05 3.26E-05 87.02  One thing to note that all the assumption is based on the laminar flow condition. By checking the Relynos number in the system, the maximum flow rate to maintain laminar flow in the cylinder is about 21.5 L/min, which would be ten times larger than the flow rate used in the experiment. So the laminar flow assumption is valid.                     79   APPENDIX 5. SIDE LOOP STEADY-STATE CALIBRATION RESULTS   Combing the theoretical value of shear rate and Leveque coefficient, we can use equation A1-5 to calculate the theoretical limiting current. For example, if the flow rate is 3.26E-05 m3/s and the temperature is 20°C, the shear rate would be 87.02 s -1  and KLev = 4.6598E-07 A·s 1/3 . Applying equation A2-5, Istandard = 4.6598E-07×(87.02 1/3 ) = 2.065E-06 A. What we measured in the experiments was not the current but the voltage. Referring to figure A2-6, the voltage we get was the current times the resistant 100 Ω and times the amplifier gain 1000, Vstandard = IR×1000 = 2.065E-06×100×1000 = 0.2065 V. Under four different flow rates we will get four measured value, then we can get the ratio between the theoretical value and the measurement value called α, so that Vstandard = αVmeasure. The α was also called steady state calibration coefficient in this thesis. The calibration experiments were done in triplicate and averaged value of α before and after the experiments are presented in Table  5-1.  Table A5-1. Steady-state calibration results of the shear probe before and after the experiments. Q (m 3 /s) Probe 1 α Probe 2 α Before After Before After 8.28E-06 8.413E-01 8.667E-01 7.060E-01 7.088E-01 1.67E-05 8.300E-01 8.978E-01 7.002E-01 6.928E-01 2.46E-05 8.127E-01 8.926E-01 6.726E-01 6.711E-01 3.26E-05 7.950E-01 8.989E-01 6.506E-01 6.780E-01 Average value 8.197E-01 8.890E-01 6.823E-01 6.877E-01  The results showed that two probes remained consistent conditions throughout the study. This gave us the confidence that the experiment data is reliable.    80  APPENDIX 6. CONVERSION BETWEEN STEADY-STATE CALIBRATION COEFFICIENT AND VOLTAGE-STEP CALIBRATION COEFFICIENT   The different between the measured voltage and the theoretical voltage is due to the difference in probe area if other conditions are all the same, referring to the equation A1-1 and A1-2, the steady state calibration coefficient α can be expressed as:  3/5 3/5 3/5 3/5 3/1 3/1 1/31/3-2/3 1/31/32/3 measure standard measure standard 'd d 'd4/1 d4/1 'd'A Ad γd'ckDnFA' γdnFAckD I I V V α             (A6-1)  where d ’ is the actual diameter of the probe. Referring to the equation A1-6 and A1-7, the voltage-step calibration coefficient β can be expressed as:  2 2 2 2 2/11/2 2/11/2 measure-Cot -standardCot 'd d 'd4/1 d4/1 'A A cDnFA' nFAcD K K                 (A6-2)  Combining equation A6-1 and A6-2, we can easily get that α = β5/6. This helps us to compare the calibration result between steady-state calibration and voltage-step calibration.            81  APPENDIX 7. VOLTAGE-STEP CALIBRATION RESULTS   The voltage-step calibration was periodically performed throughout the study to monitor the property of the electrolyte. When there was a change of the sparging tank and the electrolyte needed to be drained and refilled, the calibration would be done. In the following table, all the voltage-step calibration results were presented.  Table A7-1. Voltage-step calibration results.  Probe 1 Probe 2 β α β α Side loop(before) 0.951 0.959 0.779 0.812 Throughout the entire experiment 1.199 1.163 0.832 0.858 1.095 1.079 0.694 0.738 1.004 1.003 0.619 0.671 1.004 1.003 0.833 0.859 0.988 0.990 0.711 0.753 1.147 1.121 0.811 0.840 1.024 1.020 0.723 0.763 1.067 1.055 0.797 0.828 1.052 1.043 0.755 0.792 1.119 1.098 0.734 0.773 1.113 1.093 0.791 0.823 0.998 0.998 0.714 0.755 0.999 0.999 0.748 0.785 Side loop(after) 0.983 0.986 0.714 0.755 Average 1.049 1.041 0.751 0.787  The two averaged α value from the voltage-step calibration results are slightly larger than from the steady-state calibration results. The voltage-step calibration helps to exam the condition of the probes and the electrolyte throughout the entire experiment.         82  APPENDIX 8. EXPERIMENTAL PROCEDURES FOR THE SHEAR STRESS MEASUREMENT   The volume of the electrolyte required in the D14 tank was about 32 L and for the D20 tank was about 60 L. Slightly more electrolyte solution was prepared to cover the volume of the pipe and in case of any leakage. The concentration of each chemical was the same as listed in Appendix 2. Due to the large amount of water required, the solute used in the D20 tank was tap water instead of RO water. Because there were a lot of solution and many experimental runs to be carried on, it was important to monitor the condition of the solution. The following experiment procedure was developed to ensure the quality of measurement results.  1) Clean the two shear probes using de-ionized water and Kimwipes® before put the membrane module into the sparging tank. 2) Open the valve on top of the electrolyte reservoir, and then open the nitrogen gas line connected to the diffuser in the reservoir. Use the master flux pump to pump the solution into the reservoir. 3) Keep the nitrogen gas going for about 10 mins to purge out the oxygen in the water. After that, turn on the nitrogen gas line connected to the sparging tank, and then turn on the pump to start pumping the electrolyte to the sparging tank. 4) When the tank was filled, turn off the both line nitrogen gas line as well as the pump. Turn off the valve on top of the electrolyte reservoir. Use the DO meter to check the oxygen saturation. If the oxygen saturation is larger than 7%, turn on the nitrogen gas line connected to the sparging tank again. Monitor the DO value until it reaches less than 7%. 5) After turn on the nitrogen gas line connected to the sparging tank, adjust the pressure regulator to 15 psi and set the flow rate to 8000 ml/min. After that, do the limiting current test under the higher flow rate condition and make the necessary adjustment. 6) Do the voltage-step test. After that start sampling. The sampling period for each experiment run would be 1 min. Change the air flow pattern and air flow rate and do the test again. These two conditions can easily be changed without empty the sparging tank. Keep the pressure at 15 psi all the time. In addition, there was also not requirement to empty the sparging tank if want to switch between fine bubble diffuser and coarse bubble diffuser. 83  7) When there is a need to change conditions of the tank setup, bulk flow pattern, bundle slackness as well as the switch of pulse diffuser, the electrolyte need to be pumped back to reservoir. Open the valve on top of the electrolyte reservoir and turn on the nitrogen gas line connected to the diffuser in the reservoir. Use the master flux pump to pump back the solution into the reservoir. After that, turn off the valve on top of the electrolyte reservoir and the nitrogen gas line connected to the diffuser in the reservoir. 8) When changing the setup of the sparging tank, use the de-ionized water and Kimwipes® to wash the shear probes. 9) After change the setup of the sparging tank, turn on the valve on top of the electrolyte reservoir and the nitrogen gas line connected to the diffuser in the reservoir. Turn on the nitrogen gas line connected to the sparging tank and then turn on the pump to start pumping the electrolyte to the sparging tank. 10) Repeat the step (4), (6), (7), (8) and (9) until all the conditions were tested (step (5) only do once at the beginning of the experiment).  The electrolyte solution was stored in an opaque reservoir when not in use. The voltage-step was carried out when there was a change of the sparging tanks setup or when it was felt necessary.  After the measurements, the shear probe was washed by de-ionized water and Kimwipes® immediately. The corrected data was analyzed and the average calibration factor for each probe was obtained. Then the two test fibres would be removed from the membrane module, and moved back to the side loop setup to do the calibration again. This was to examine the quality of the shear probes. The results of the voltage-step calibration were shown in Appendix 7. With those results, we were able to track the quality of the shear probes as well as the quality of electrolyte.         84  APPENDIX 9. EXPERIMENTAL PROCEDURE FOR FILTRATION TEST   For the filtration test, the following procedures were developed to ensure the stable performance of the membrane.  1) After the active membrane fibers were assembled (see section 3.1.2), an integrity test was performed. This was done by submerging the active membrane fibers into a water tank, and using the permeate pump (in a reverse flow direction) to slowly add air into the fibers to achieve a pressure of 3 psi. If the pressure drop inside the active membrane fibers was less that 2 % in 5 mins, the membrane fibers will be considered not to be breached. If the pressure drops significantly, new active membrane fibers need to be assembled. 2) If the active membrane fibers pass the integrity test, the membrane module can be assembled (section 3.1.2). After the module is assembled, a second integrity test is performed to ensure that the active membrane fibers were not damaged. 3) Place the membrane module in the system tank and turn on the permeate pump. Set the flow rate of the pump to 60 mL/min and measure the initial TMP of the membrane. At the room temperature of approximately 22 ºC, the TMP for clean water should be approximately 4.5 psi. 4) Prepare the bentonite solution. Measure the amount required for the filtration test. For the D14 tank, 8 g bentonite is required and for the D20 tank 15 g is required. Use a small container of approx. 3 L to prepare a bentonite slurry and make sure that all the bentonite power is dispersed. Hot water can be used to accelerate the dispersing process. Approximately 10 – 15 mins are required before the bentonite becomes fully disperse. 5) Use the tap water to rinse the container and then transfer the rinse water to the tank. Do this 3 times to make sure that all bentonite has been transferred into the tank. Fill up the tank with tap water. Turn on the air to mix the solution and make sure that the temperature of water in the tank is approx. 22 -27 ºC. 6) When the bentonite solution is ready, the filtration test can be started. Use the pressure regulator and the flow meter to adjust the flow rate of the sparging air. The pressure regulator was set to 15 psi. And the air flow rate is adjusted for different experiment conditions. The flow rate of the permeate pump is set to 60 mL/min. Use the pressure transducer and data logger to record the 85  pressure change in the permeate line. Stop the filtration when the TMP reach 10 psi. If after 160 mins and the TMP still doesn’t reach the 10 psi, the experiment will also stop. 7) After the membrane is fouled (i.e. at the end of the filtration test), a backwash is performed. The bentonite solution is drained from the tank which is then fill up with clean water from hot water tap (temperature 40ºC). The system is then backwashed. A relatively low flow rate is used during backwash so that the pressure does not exceed 3 psi. An air flow rate of 10000 mL/min is used during the backwash. After 10 mins, the tank is drained and it is then filled up again with hot water. A second backwash is then performed and tank is drained after. 8) The tank is then filled up with tap water and a clean water test is performed. A mixture of cold and hot tap water is used to achieve a water temperature of approximately 22-27 ºC in the tank. In filtration mode, monitor the TMP when filtering clean water. If the TMP is larger than 4.5 psi, clean the membrane again. If the TMP is significantly less than 4.5 psi, do the integrity test to check any leakage. When the TMP is approx.. 4.5 psi, the membrane is considered to be clean and next filtration test can begin. 9)  As the filtration experiments lasted for a few weeks, there was the possibility that bacterial growth occurred in/on the membrane. If the clean water TMP is greater than 4.5 psi and backwash could not reduce it, a chemical cleaning was performed. This was done by soaking the membrane module in a bleach solution with a concentration of 1000 ppm for 16 hours. After 16 hours, a 100 ppm bleach solution was filtered for 30 mins. The fibers were then rinsed by filtering with clean water for another 30 mins. A chemical cleaning was usually performed every other week. 10) Integrity tests were done every other week as well as at the end of the filtration test to ensure the membrane integrity was not compromised.           86  APPENDIX 10. QA/QC RESULTS   A10.1 QA/QC Results for Shear Stress Measurement  Calibration was one of the most important quality control methods during the shear stress measurement. It helped to monitor the quality of the shear probe as well as the electrolyte. The calibrations for the shear probes in D14 and D20 experiments were done separately due to the different time and location of these two experiments. A parameter, α, was introduced as the calibration coefficient which indicated the different value between the measured value and the standard value. For the full definition of α, please refer to appendix 5. The calibration results for D14 and D20 were presented separately.  A10.1.1 Calibration results of the shear probes in D-140 tank experiments  Referring to appendix 5 table A5-1, under different flow rate the α value remained relatively constant.  s α did not relate to the change of flow rate, this result gave us the confidence that the system and the shear probe was in stable conditions. Compared the α between probe 1 and probe 2, probe 1 has a bigger α than probe 2, which means that, probe 2 has a larger diameter than probe 1. The results between before and after the shear measurement showed that, there was a small increase of α. For probe 1 was 8.5% and for probe 2 was 0.8%. This may due to the different bulk chemical concentration or some effect of electro-field in the surrounding environment. But according to equation A1-1, the limiting current is the product of the KLev and γ 1/3 . For the calibration, we try to calibrate the effect of KLev. And in the shear measurement, the different of the shear rate can be 10-100 times. So only a 8% different of the α value would not really effect the result of the limiting current we were measuring. We would consider that the probe was in a relatively stable condition throughout the experiment.  For the voltage-step calibration, the following graph was generated from appendix 7 table A7-1.  87   Figure A10-1. Voltage-step calibration results throughout the experiment.  Throughout the study, there were 15 voltage-step calibrations performed. From the figure we could tell that the α value remained stable throughout the experiment. The small variation is likely due to experimental errors. This indicated that the characteristics of the electrolyte remained stable from begin to the end and the electrolyte was well kept in a oxygen free environment.  Table A10-1 presents the combination of results of the steady-state calibration and voltage-state calibration. In the table, the average α values for probe 1 and probe 2 under different calibration methods were presented. From the table we can see that steady-state calibration would give us lower α value than voltage-step calibration. This may due to the different assumptions and other factors that cannot be quantified. For example, the steady-state calibration assumes that the shear probe was at the centre of the plexigals cylinder. This may not be true in the real situation. In addition, the shear probe surface may be not perfectly smooth and circular shape. For the voltage-step calibration, we need to choose the target time interval for the linear regression, but the choice of this time interval is still under research, different time interval would lead to different KCot value. All of these can lead to small errors of calibration. However, small differences of the α doesn’t really effect the measurement results since we average the value between steady-state and voltage-step calibration and would use these two number for the following calculation and discussion.    88   Table A10-1. Summary of the steady-state and voltage-step calibrations in D14 tank.  Probe 1 α Probe 2 α Steady-state 0.854 0.685 Voltage-step 1.041 0.787 Average 0.948 0.736  A10.1.2 Calibration results of the shear probes in D20 tank experiments  The voltage-step calibration results of the two shear probes in the D20 tank were presented in Table A10-2. Table A10-2. Voltage-step calibration results of the probe 1 and probe 2 in D20 tank. Probe 1 α Probe 2 α 0.872 0.622   A10.2 QA/QC Results for Filtration Test  The data obtained from the pressure transducer was actually a voltage measurement. To convert the voltage signal to pressure, a calibration was done. The results are presented in Figure A10-2. The calibration was done again after the filtration process to make sure that there was not significant change of the pressure transducer.  89   Figure A10-2. Calibration of the pressure transducer Particle size distribution of the bentonite solution was measured (Figure A10-3). From the figure we can see that peak size distribution happened in the size range from 0.631 to 0.724, which is 14.04% of the total number of the particle. The detectable particle size was about 0.4 µm. This is 10 times larger than the membrane pore size which is about 0.04 µm. The bentonite particle would not likely to go inside the membrane.    Particle Size Distribution  0.01  0.1  1  10  100  1000  3000 Particle Size (µm) 0  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17 N u m b e r (% ) bentonite-, Wednesday, March 23, 2011 2:45:09 AM Figure A10-3. Particle size distribution of the bentonite solutions.  Temperature throughout the whole experiment was carefully monitored. The thermometer was used to check the temperature in the tank. The ratio between the cold water and hot water was about 3:4. After the mixing, the temperature was checked and if the temperature was too high or too low, the air would be turned on to mix up the water, which was to help the temperature in the tank reach the room 90  temperature. The initial temperature and final temperature of each filtration run was recorded and the different between them was within 1 ºC. The room temperature was around 24 ºC. For all the tests, the recorded temperature was between 22.9 to 26.9 ºC. As a result, the effect of the viscosity different was minimized.  There were altogether six integrity tests done throughout the entire filtration experiments. All the pressure drop was less than 0.05 psi when pressurized the membrane to 3 psi. Due to the diffusion of the gas into the water, slightly pressure drop was reasonable and we assume that the fiber was not leak throughout the experiment.                      91  APPENDIX 11. SHEAR STRESS MEASUREMENT RESULTS IN D14 TANK   Figure A11-1. Result of PLOSC1.3.  Figure A11-2. Result of PLOSC2.2. 92   Figure A11-3. Result of PMOSC1.3.      Figure A11-4. Result of PMOSC2.2.            93    Figure A11-5. Result of PHOSC1.3.   Figure A11-6. Result of PHOSC2.2 94   Figure A11-7. Result of PLWSC1.3.   Figure A11-8. Result of PMWSC1.3.  95    Figure A11-9. Result of PHWSC1.3.   Figure A11-10. Result of PLWCC1.3. 96     Figure A11-11. Result of PMWCC1.3.    Figure A11-12. Result of PHWCC1.3. 97     Figure A11- 13. Result of CLOSC1.3.    Figure A11-14. Result of CLOSC2.2. 98     Figure A11-15. Result of CLOSF1.3.   Figure A11-16. Result of CLOSF2.2. 99   Figure A11-17. Result of CLOSS1.3.    Figure A11-18. Result of CLOSS2.2.  100   Figure A11-19. Result of CMOSC1.3.   Figure A11-20. Result of CMOSC2.2.  101    Figure A11-21. Result of CMOSF1.3.   Figure A11-22. Result of CMOSF2.2. 102     Figure A11-23. Result of CMOSS1.3.    Figure A11-24. Result of CMOSS2.2. 103    Figure A11-25. Result of CHOSC1.3.    Figure A11-26. Result of CHOSC2.2. 104   Figure A11-27. Result of CHOSF1.3.   Figure A11-28. Result of CHOSF2.2.  105   Figure A11-29. Result of CHOSS1.3.   Figure A11-30. Result of CHOSS2.2.  106   Figure A11-31. Result of CLWSC1.3.   Figure A11-32. Result of CLWSF1.3.  107   Figure A11-33. Result of CLWSS1.3.   Figure A11-34. Result of CMWSC1.3.  108   Figure A11-35. Result of CMWSF1.3.    Figure A11-36. Result of CMWSS1.3. 109   Figure A11-37. Result of CHWSC1.3.   Figure A11-38. Result of CHWSF1.3.  110   Figure A11-39. Result of CHWSS1.3.   Figure A11-40. Result of CLWCC1.3.  111   Figure A11-41. Result of CLWCF1.3.   Figure A11-42. Result of CLWCS1.3.  112   Figure A11-43. Result of CMWCC1.3.  Figure A11-44. Result of CMWCF1.3.   113   Figure A11-45. Result of CMWCS1.3.  Figure A11-46. Result of CHWCC1.3.   114   Figure A11-47. Result of CHWCF1.3.   Figure A11-48. Result of CHWCS1.3.  115   Figure A11-49. Result of FLWSC1.3.    Figure A11-50. Result of FMWSC1.3. 116   Figure A11-51. Result of FHWSC1.3.   Figure A11-52. Result of FLWCC1.3.  117   Figure A11-53. Result of FMWCC1.3.    Figure A11-54. Result of FHWCC1.3. 118  APPENDIX 12. RMS, MEAN AND STD FOR ALL SHEAR STRESS MEASUREMNTS.  Table A12-1. The RMS, mean and STD of shear stress values for the 54 experiments in D14 tank.  Probe1 Test1 Probe2 Test1 Probe1 Test2 Probe2 Test2 RMS Mean Std RMS Mean Std RMS Mean Std RMS Mean Std PLOSC1.3 0.635 0.370 0.516 0.708 0.350 0.615 0.654 0.359 0.547 0.744 0.419 0.615 PLOSC2.2 0.615 0.367 0.494 0.997 0.574 0.815 0.754 0.444 0.609 0.930 0.537 0.759 PMOSC1.3 0.822 0.496 0.656 0.810 0.506 0.632 0.850 0.557 0.642 0.876 0.540 0.690 PMOSC2.2 0.939 0.606 0.718 - - - 0.881 0.590 0.654 - - - PHOSC1.3 0.990 0.672 0.727 1.092 0.678 0.856 0.985 0.643 0.747 0.985 0.655 0.735 PHOSC2.2 0.953 0.614 0.730 0.823 0.513 0.643 0.912 0.608 0.679 0.739 0.511 0.534 PLWSC1.3 0.940 0.716 0.608 0.942 0.712 0.617 0.986 0.761 0.626 0.834 0.635 0.541 PMWSC1.3 1.173 0.941 0.700 0.876 0.615 0.624 1.151 0.862 0.762 0.806 0.506 0.627 PHWSC1.3 1.280 0.938 0.870 1.012 0.672 0.757 1.227 0.942 0.787 1.020 0.691 0.750 PLWCC1.3 1.115 0.849 0.723 0.927 0.786 0.491 1.022 0.870 0.536 0.923 0.778 0.496 PMWCC1.3 1.315 1.109 0.706 1.208 1.047 0.602 1.414 1.203 0.743 1.208 1.052 0.594 PHWCC1.3 1.496 1.293 0.753 1.455 1.267 0.715 1.487 1.300 0.723 1.419 1.246 0.678 CLOSC1.3 0.423 0.299 0.299 0.468 0.325 0.337 0.612 0.426 0.440 0.528 0.378 0.368 CLOSC2.2 0.709 0.521 0.480 0.571 0.397 0.410 0.655 0.476 0.451 0.580 0.410 0.410 CLOSF1.3 0.524 0.357 0.383 0.379 0.219 0.310 0.592 0.396 0.440 0.433 0.273 0.336 CLOSF2.2 0.510 0.355 0.366 0.490 0.310 0.380 0.558 0.392 0.398 0.584 0.384 0.440 CLOSS1.3 0.573 0.370 0.437 0.508 0.294 0.414 0.422 0.251 0.340 0.358 0.197 0.299 CLOSS2.2 0.463 0.290 0.360 0.374 0.222 0.301 0.597 0.369 0.469 0.433 0.240 0.361 CMOSC1.3 0.657 0.487 0.441 0.568 0.399 0.404 0.747 0.559 0.496 0.543 0.389 0.379 CMOSC2.2 0.751 0.552 0.510 0.574 0.413 0.398 0.840 0.626 0.560 0.638 0.453 0.449 CMOSF1.3 0.613 0.389 0.474 0.547 0.331 0.435 0.699 0.440 0.543 0.480 0.283 0.387 CMOSF2.2 0.631 0.429 0.463 0.621 0.393 0.482 0.675 0.459 0.494 0.641 0.422 0.482 CMOSS1.3 0.733 0.443 0.584 0.460 0.267 0.374 0.672 0.396 0.543 0.470 0.268 0.385 CMOSS2.2 0.759 0.438 0.620 0.497 0.313 0.386 0.659 0.415 0.513 0.610 0.380 0.477 CHOSC1.3 0.901 0.643 0.631 0.719 0.518 0.498 0.680 0.509 0.451 0.719 0.488 0.527 CHOSC2.2 0.860 0.641 0.574 0.792 0.592 0.527 0.866 0.645 0.578 0.739 0.518 0.528 CHOSF1.3 0.793 0.466 0.641 0.557 0.345 0.437 0.803 0.539 0.595 0.677 0.414 0.535 CHOSF2.2 0.813 0.546 0.603 0.659 0.407 0.519 0.779 0.494 0.602 0.629 0.407 0.480 CHOSS1.3 0.618 0.375 0.491 0.638 0.387 0.508 0.605 0.375 0.475 0.558 0.315 0.460 CHOSS2.2 0.763 0.470 0.602 0.608 0.329 0.512 0.782 0.495 0.605 0.702 0.424 0.559 CLWSC1.3 0.962 0.762 0.587 0.959 0.731 0.620 0.925 0.718 0.583 0.788 0.580 0.534 CLWSF1.3 0.765 0.525 0.556 0.647 0.422 0.490 0.777 0.540 0.559 0.672 0.483 0.468 CLWSS1.3 0.776 0.496 0.597 0.607 0.380 0.474 0.631 0.370 0.511 0.656 0.426 0.498 CMWSC1.3 1.336 1.085 0.779 1.051 0.817 0.661 1.292 1.017 0.797 1.000 0.765 0.645 CMWSF1.3 1.018 0.684 0.754 0.829 0.550 0.621 0.992 0.677 0.726 1.012 0.690 0.739 CMWSS1.3 1.054 0.705 0.783 0.980 0.657 0.727 1.163 0.746 0.893 0.857 0.498 0.697 CHWSC1.3 1.474 1.181 0.881 1.164 0.923 0.709 1.476 1.153 0.921 1.126 0.890 0.691 CHWSF1.3 1.241 0.796 0.952 0.872 0.543 0.683 1.209 0.839 0.871 0.911 0.603 0.683 CHWSS1.3 1.237 0.826 0.920 1.045 0.678 0.795 1.302 0.839 0.996 1.087 0.664 0.861 CLWCC1.3 1.076 0.886 0.611 1.227 1.040 0.652 1.030 0.879 0.536 1.530 1.301 0.805 CLWCF1.3 0.809 0.610 0.530 1.048 0.804 0.672 0.841 0.621 0.568 0.918 0.688 0.607 CLWCS1.3 0.870 0.579 0.649 0.777 0.526 0.572 0.695 0.436 0.541 0.696 0.431 0.546 CMWCC1.3 1.269 1.081 0.665 1.255 1.069 0.657 1.234 1.019 0.696 1.164 0.974 0.636 CMWCF1.3 0.959 0.713 0.641 0.961 0.658 0.700 1.121 0.880 0.695 1.097 0.848 0.696 CMWCS1.3 1.006 0.655 0.764 0.983 0.685 0.705 0.990 0.664 0.734 0.853 0.571 0.633 CHWCC1.3 1.512 1.314 0.747 1.447 1.238 0.748 1.425 1.202 0.766 1.334 1.105 0.747 CHWCF1.3 1.300 0.985 0.849 1.206 0.927 0.772 1.355 1.048 0.859 1.200 0.961 0.718 CHWCS.13 1.157 0.796 0.840 1.141 0.823 0.791 1.144 0.791 0.827 1.113 0.780 0.794 119    Probe1 Test1 Probe2 Test1 Probe1 Test2 Probe2 Test2 RMS Mean Std RMS Mean Std RMS Mean Std RMS Mean Std FLWSC1.3 0.867 0.680 0.538 0.796 0.611 0.511 0.862 0.658 0.557 0.662 0.509 0.423 FMWSC1.3 1.178 0.926 0.727 0.956 0.746 0.598 1.218 0.948 0.765 0.826 0.637 0.526 FHWSC.13 1.403 1.111 0.855 1.182 0.939 0.717 1.356 1.054 0.853 1.157 0.902 0.724 FLWCC1.3 1.056 0.882 0.581 1.071 0.857 0.642 0.893 0.715 0.534 1.087 0.916 0.585 FMWCC1.3 1.585 1.390 0.763 1.654 1.428 0.835 1.401 1.202 0.720 1.815 1.580 0.894 FHWCC1.3 1.618 1.394 0.822 1.591 1.373 0.804 1.756 1.527 0.867 1.901 1.658 0.929  Table A12-2. The RMS, mean and STD of shear stress values for the 2 experiments in D20 tank.  Test 1 Test 2 Test 3 Slow pulse RMS Mean Std RMS Mean Std RMS Mean Std Probe1 0.953 0.562 0.770 0.918 0.523 0.754 1.040 0.571 0.871 Probe2 0.861 0.500 0.701 0.847 0.480 0.698 0.877 0.481 0.734 Fast pulse Probe1 1.010 0.644 0.784 0.993 0.610 0.784 0.947 0.610 0.724 Probe2 0.782 0.496 0.604 0.789 0.484 0.623 0.778 0.497 0.598                   120  APPENDIX 13. FILTRATION TESTS RESULTS    Figure A13-1. Filtered volume vs. TMP under the condition of slow pulse diffuser, stagnant bulk liquid and 3500 mL/min continuous gas flow in D14 tank.   Table A13-1. Filtration model fitting of the results under the condition of slow pulse diffuser, stagnant bulk liquid and 3500 mL/min continuous gas flow in D14 tank.  Model Test 1 Test 2 Test 3 ki ki R 2  ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.0270 0.2860 0.0324 0.1326 0.0419 0.6954 0.0338 0.0075 Intermediate blocking Vk 0t iePP   0.0294 0.4223 0.0366 0.3274 0.0488 0.7865 0.0383 0.0098 Cake fouling VkPP c0t   0.1455 0.5367 0.1854 0.4830 0.2588 0.8668 0.1966 0.0575   121   Figure A13-2. Filtered volume vs. TMP under the condition of coarse bubble diffuser, stagnant bulk liquid and 3500 mL/min continuous gas flow in D14 tank.   Table A13-2. Filtration model fitting of the results under the condition of coarse bubble diffuser, stagnant bulk liquid and 3500 mL/min continuous gas flow in D14 tank.  Model Test 1 Test 2 Test 3 ki ki R 2  ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.1440 0.9915 0.1373 0.9900 0.1625 0.9938 0.1479 0.0130 Intermediate blocking Vk 0t iePP   0.1894 0.9896 0.1808 0.9956 0.2133 0.9909 0.1945 0.0168 Cake fouling VkPP c0t   1.1646 0.9549 1.1107 0.9595 1.3084 0.9489 1.1946 0.1022        122   Figure A13-3. Filtered volume vs. TMP under the condition of coarse bubble diffuser, stagnant bulk liquid and 3500 mL/min 3 sec on 3 sec off gas flow in D14 tank.   Table A13-3. Filtration model fitting of the results under the condition of coarse bubble diffuser, stagnant bulk liquid and 3500 mL/min 3 sec on 3 sec off gas flow in D14 tank.  Model Test 1 Test 2 Test 3 ki ki R 2  ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.2039 0.9965 0.1742 0.9917 0.1908 0.9946 0.1896 0.0149 Intermediate blocking Vk 0t iePP   0.2675 0.9896 0.2298 0.9914 0.2500 0.9887 0.2491 0.0189 Cake fouling VkPP c0t   1.6401 0.9453 1.4186 0.9489 1.5307 0.9412 1.5298 0.1108    123   Figure A13-4. Filtered volume vs. TMP under the condition of coarse bubble diffuser, stagnant bulk liquid and 7000 mL/min 2 sec on 2 sec off gas flow in D14 tank.   Table A13-4. Filtration model fitting of the results under the condition of coarse bubble diffuser, stagnant bulk liquid and 7000 mL/min 2 sec on 2 sec off gas flow in D14 tank.  Model Test 1 Test 2 Test 3 ki ki R 2  ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.1260 0.9913 0.0878 0.9618 0.1116 0.9912 0.1085 0.0193 Intermediate blocking Vk 0t iePP   0.1661 0.9926 0.1162 0.9958 0.1463 0.9935 0.1429 0.0251 Cake fouling VkPP c0t   1.0238 0.9544 0.7172 0.9813 0.8938 0.9580 0.8783 0.1539       124   Figure A13-5. Filtered volume vs. TMP under the condition of slow pulse diffuser, cross-flow bulk liquid and 3500 mL/min continuous gas flow in D20 tank.   Table A13-5. Filtration model fitting of the results under the condition of slow pulse diffuser, cross-flow bulk liquid and 3500 mL/min continuous gas flow in D20 tank.  Model Test 1 Test 2 ki ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.0482 0.8168 0.0459 0.9076 0.0470 0.0016 Intermediate blocking Vk 0t iePP   0.0580 0.9144 0.0548 0.9648 0.0564 0.0023 Cake fouling VkPP c0t   0.3186 0.9666 0.2990 0.9885 0.3088 0.0139         125   Figure A13-6. Filtered volume vs. TMP under the condition of slow pulse diffuser, cross-flow bulk liquid and 1750 mL/min continuous gas flow in D20 tank.   Table A13-6. Filtration model fitting of the results under the condition of slow pulse diffuser, cross-flow bulk liquid and 1750 mL/min continuous gas flow in D20 tank.  Model Test 1 Test 2 ki ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.0981 0.9832 0.1008 0.9543 0.0995 0.0019 Intermediate blocking Vk 0t iePP   0.1281 0.9973 0.1323 0.9967 0.1302 0.0030 Cake fouling VkPP c0t   0.7789 0.9663 0.8090 0.9847 0.7940 0.0213        126   Table A13-7. Filtration model fitting of the results under the condition of slow pulse diffuser, cross-flow bulk liquid and 875 mL/min continuous gas flow in D20 tank.   Table A13-7. Filtration model fitting of the results under the condition of slow pulse diffuser, cross-flow bulk liquid and 875 mL/min continuous gas flow in D20 tank.  Model Test 1 Test 2 ki ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.1436 0.9921 0.1278 0.9840 0.1357 0.0111 Intermediate blocking Vk 0t iePP   0.1876 0.9959 0.1674 0.9988 0.1775 0.0143 Cake fouling VkPP c0t   1.1429 0.9602 1.0222 0.9705 1.0826 0.0853        127   Table A13-8. Filtration model fitting of the results under the condition of fast pulse diffuser, cross-flow bulk liquid and 3500 mL/min continuous gas flow in D20 tank.   Table A13-8. Filtration model fitting of the results under the condition of slow pulse diffuser, cross-flow bulk liquid and 3500 mL/min continuous gas flow in D20 tank.  Model Test 1 Test 2 ki ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.0666 0.9530 0.0671 0.9558 0.0668 0.0003 Intermediate blocking Vk 0t iePP   0.0873 0.9956 0.0882 0.9951 0.0878 0.0006 Cake fouling VkPP c0t   0.5350 0.9808 0.5426 0.9745 0.5388 0.0054        128   Table A13-9. Filtration model fitting of the results under the condition of fast pulse diffuser, cross-flow bulk liquid and 1750 mL/min continuous gas flow in D20 tank.   Table A13-9. Filtration model fitting of the results under the condition of slow pulse diffuser, cross-flow bulk liquid and 1750 mL/min continuous gas flow in D20 tank.  Model Test 1 Test 2 ki ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.0936 0.9586 0.1017 0.9743 0.0977 0.0057 Intermediate blocking Vk 0t iePP   0.1239 0.9979 0.1343 0.9995 0.1291 0.0074 Cake fouling VkPP c0t   0.7651 0.9821 0.8297 0.9753 0.7974 0.0457        129   Table A13-10. Filtration model fitting of the results under the condition of fast pulse diffuser, cross-flow bulk liquid and 875 mL/min continuous gas flow in D20 tank.   Table A13-10. Filtration model fitting of the results under the condition of slow pulse diffuser, cross-flow bulk liquid and 875 mL/min continuous gas flow in D20 tank.  Model Test 1 Test 2 ki ki R 2  ki R 2  Mean STD Complete blocking Vk1 P P b 0 t    0.1539 0.9921 0.1544 0.9956 0.1541 0.0003 Intermediate blocking Vk 0t iePP   0.2020 0.9958 0.2011 0.9924 0.2016 0.0006 Cake fouling VkPP c0t   1.2359 0.9590 1.2220 0.9536 1.2290 0.0098           

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
China 6 0
Japan 5 0
United States 4 1
Germany 4 0
France 2 10
Republic of Korea 2 1
Iran 2 33
United Kingdom 2 0
Ukraine 2 0
Canada 1 6
City Views Downloads
Unknown 12 1
Tokyo 5 0
Shenzhen 3 0
Beijing 3 0
Isfahan 2 33
Mountain View 2 1
Reston 1 0
Vancouver 1 6
Ashburn 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}

Share

Share to:

Comment

Related Items