Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Impact of hydrodynamic conditions and membrane configuration on the permeate flux in submerged membrane… Lei, Xiaoling E. 2005

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

Item Metadata

Download

Media
831-ubc_2005-0520.pdf [ 6.72MB ]
Metadata
JSON: 831-1.0063308.json
JSON-LD: 831-1.0063308-ld.json
RDF/XML (Pretty): 831-1.0063308-rdf.xml
RDF/JSON: 831-1.0063308-rdf.json
Turtle: 831-1.0063308-turtle.txt
N-Triples: 831-1.0063308-rdf-ntriples.txt
Original Record: 831-1.0063308-source.json
Full Text
831-1.0063308-fulltext.txt
Citation
831-1.0063308.ris

Full Text

IMPACT OF H Y D R O D Y N A M I C CONDITIONS A N D M E M B R A N E CONFIGURATION O N THE P E R M E A T E F L U X IN S U B M E R G E D M E M B R A N E SYSTEMS FOR DRINKING W A T E R T R E A T M E N T  By  X I A O L I N G E. LEI B.A.Sc. , Environmental Science and Engineering, Tsinghua University, 1991 M.A.Sc, Environmental Science and Engineering, Tsinghua University, 1994  A THESIS SUBMITTED IN P A R T I A L F U L F I L M E N T OF THE REQUIREMENTS FOR T H E D E G R E E OF M A S T E R OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Civil Engineering)  THE UNIVERSITY OF BRITISH C O L U M B I A June, 2005 ©  Xiaoling E. Lei, 2005  ABSTRACT  ii  ABSTRACT  Submerged membrane systems are increasingly being used in drinking water treatment applications. One of the major reasons driving this increase is the relatively low operating cost associated with submerged membrane systems, compared to their external counterparts. However, the operating cost for submerged membrane systems is still relatively high when compared to that associated with conventional treatment technologies (such as sand filtration). The need to maintain a high permeate flux is one of the most significant factors affecting the capital and operating cost associated with submerged membrane systems. A i r sparging is often needed to maintain a high permeate flux in submerged membrane systems. However, the air sparging mechanism that enhance the permeate flux is poorly understood. As a result, the infrastructure and control of air sparging in submerged membrane systems,are typically designed based on a time and capital extensive trial-and-error approach. This research was undertaken to investigate the air sparging mechanisms that promote a high permeate flux in a submerged membrane system.  The results indicate that the decline of the permeate flux over filtration time could be characterized by an initial short period of fast permeate flux decline, followed by a longer period of slower permeate flux decline, according to an exponential equation.  The hydrodynamic conditions and the system configuration had a significant impact on permeate flux. The crossflow of water along the membrane surface, which occurs as a result  ABSTRACT  iii  of the rising sparged air bubbles, significantly enhanced the permeate flux. The contact between the sparged air bubbles and the membrane surface had a great impact on enhancing the permeate flux. In addition, the physical contact between the membrane fiber in the submerged system, which varied with the tension of the fiber and the intensity of the air sparging, had a significant impact on the permeate flux.  The hydrodynamic conditions and the system configuration had a significant impact on the pseudo-steady-state  permeate flux. The maximum pseudo-steady state permeate flux was  achieved when the membrane system was operated under dual phase crossflow, with physical contact between the fiber in the membrane module. However, there was no significant benefit of providing a crossflow velocity in excess of 0.2m/s, in terms o f maintaining a high, pseudosteady-state, permeate flux. The hydrodynamic conditions and the membrane configuration did not have a consistently significant impact on the initial permeate flux decline coefficient. The hydrodynamic conditions and the system configuration had a significant impact on the pseudo-steady-state  permeate flux decline coefficient. The minimum,  pseudo-steady-state,  permeate flux decline coefficient was achieved when the membrane system was operated under dual phase crossflow, with the physical contact between the membrane fiber in the membrane module. Significant physical contact only occured when fibers in the membrane module were in a loose configuration.  The pseudo-steady-state permeate flux decline coefficient was proportional to the inverse of pseudo-steady-state permeate flux for all of the experimental conditions investigated in this present study.  iv  TABLE OF CONTENTS  TABLE OF CONTENTS ABSTRACT  ii  TABLE OF CONTENTS  iv vii  LIST O F T A B L E S LIST O F FIGURES-  viii  LIST O F S Y M B O L S  xii  xiii  ACKNOWLEDGEMENTS 1  INTRODUCTION  1.1  Introduction  1  Aim and Objectives  3  Thesis Structure  4  LITERATURE REVIEW  5  .1.2 1.3 2  1  AND RESEARCH OBJECTIVES  2.1  Definition of Membrane—  2.2  Membrane Classification  -  5 —  2.2.1  Classification based on a pore size of membrane material—  6  2.2.2  Classification based on configuration of membrane module  8  2.2.3  Classification based on the physical arrangement in a particular treatment system  3  6  9  2.3  Material Accumulation and Permeate Flux Decline  2.4  Air Sparging in Submerged Hollow Fiber Membrane System  —  -—10 14  EXPERIMENTAL DESIGN  18  3.1  18  Bench-scale Membrane System  TABLE OF  v  3.1.1  System tank  -  19  3.1.2  Membrane module  21  3.1.3  Aeration system  24  3.1.4  System vacuum  31  3.1.5  Permeate flux collection and measurement system  32  3.2  Source Water to be Filtered  3.3  Experimental Program  34  3.4  Routine Membrane Integrity Testing and Maintenance  36  3.5  4  CONTENTS  T  33  3.4.1  Membrane integrity testing  36  3.4.2  Membrane cleaning—  37  Analytical Methods  38  3.5.1  Temperature  38  3.5.2  Transmembrane pressure  38  3.5.3  Permeate  flux—-  38 41  RESULTS AND DISCUSSIONS 4.1  Permeate Flux Decline over Time  4.2  Permeate Flux Decline with Volume Filtered  43  4.3  Modeling of Experimental Results  45  —41  4.3.1  Model of experimental data  45  4.3.2  Estimate of the initial permeate  4.3.3  Estimate of pseudo-steady-state permeate flux, initial permeate flux decline  flux  coefficient and pseudo-steady-state permeate flux decline coefficient  47  48  TABLE OF CONTENTS 4.4  vi  Effect of Hydrodynamic Conditions and Membrane Configurations on Pseudo Steady State Permeate Flux, Initial Permeate Flux Decline Coefficient, and Pseudo Steady State Permeate Flux Decline Coefficient  4.5 5  49  4.4.1  Impact of single phase crossflow on permeate flux  49  4.4.2  Impact of dual phase crossflow with air sparging on permeate flux  52  4.4.3  Impact of physical contact between membrane fibers on the permeate flux-56  4.4.4  Impact of fiber tension in membrane modules on permeate flux  Discussion  —  -  CONCLUSIONS  60 66 72  REFERENCES  75  APPENDIX: EXPERIMENT RESULTS OF T O T A L ORGANIC CARBON  82  vii  LIST OF TABLES  LIST O F T A B L E S  2.1 Membrane classification and application  —  —6  3.1 Physical characteristics of the membranes used in the present study 3.2 Operating bulk crossflow velocity and air flowrate  21 —  29  3.3 Key characteristics of hydrodynamic condition and membrane configuration of each experiment  -  -35  4.1 Multiple coefficient of determination of experimental results of non-linear regression analysis using Equation 4.1  45  4.2 Estimate of the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient and the pseudo-steady-state permeate flux decline coefficient  A . l : Experiment Results of TOC Removal  -  —  49  82  LIST OF FIGURES  vm  LIST OF FIGURES  2.1  Selective permeation across a membrane  5  2.2  Filtration application guide for different materials removal  7  2.3  Schematic representation of the four principal types of membrane modules  8  2.4  Typical operation mode for external membrane systems  9  2.5  Typical operation mode for internal membrane systems  10  2.6  Materials accumulation during membrane filtration process  12  2.7  Schematic representation of permeate flux decline  13  2.8  Reductions in permeate flux over time with periodic cleaning  14  2.9  Schematic of sludge accumulation processT  15  2.10 Comparison of suction pressure profiles between loose and tight hollow fiber membrane systems  16  2.11 Increase in suction pressure with tight and loose fiber bundles versus different gas flow rates  17  3.1  Schematic of the laboratory bench-scale set-up  18  3.2  Picture of the bench-scale experimental set-up  —19  3.3  Schematic of system tank with cylindrical baffle  3.4  Schematic of membrane module  21  3.5  Pictures of membrane fibers and membrane modules  22  3.6  Schematic of membrane modules  23  —  20  ix  LIST OF FIGURES 3.7  Schematic of loose and tight configuration for the membrane modules  24  3.8  Central aerator  25  3.9  Peripheral aerator  25  3.10 Schematic of aeration system  .  26  3.11 Schematic of system crossflow  28  3.12 Relationships between bulk crossflow velocity and air flowrate  —28  3.13 Pictures of bulk crossflow velocity measurement  29  3.14 Dye injection for bulk crossflow velocity measurement  30  3.15 Typical trend of transmembrane pressure-  ~  32  3.16 Schematic of the membrane integrity testing apparatus  36  4.1  Permeate flux versus filtration time  42  4.2  Permeate flux versus volume  4.3  Initial permeate flux  48  4.4  Impact of single phase bulk crossflow velocity on the permeate flux  50  4.5  Pseudo-steady-state permeate flux versus bulk crossflow velocity (single phase  filtered  44  crossflow) 4.6  51  Initial permeate flux decline coefficient versus bulk crossflow velocity (single phase crossflow)  4.7  51  Pseudo-steady-state permeate flux decline coefficient versus bulk crossflow velocity (single phase crossflow)  52  4.8  Impact of dual phase bulk crossflow velocity on the permeate flux  53  4.9  Pseudo-steady-state permeate flux versus bulk crossflow velocity (dual phase crossflow)  —  54  LIST OF FIGURES  x  4.10 Initial permeate flux decline coefficient versus bulk crossflow velocity (dual phase crossflow)  —  —  55  4.11 Pseudo-steady-state permeate flux decline coefficient versus bulk crossflow velocity (dual phase crossflow)  56  4.12 Impact of physical contact on the permeate flux  57  4.13 Pseudo-steady-state permeate flux versus bulk crossflow velocity (dual phase crossflow with physical contact)  58  4.14 Initial permeate flux decline coefficient versus bulk crossflow velocity (dual phase crossflow with physical contact)  59  4.15 The pseudo-steady-state permeate flux decline coefficient versus bulk crossflow velocity (dual phase crossflow with physical contact) 4.16 Impact of fiber tension in membrane modules on permeate flux  60 —  61  4.17 Impact of fiber tension in membrane modules on pseudo-steady-state permeate flux-62 4.18 Impact of fiber tension in membrane modules on initial permeate flux decline coefficient  —  —63  4.19 Impact of fiber tension in membrane modules on pseudo-steady-state permeate flux decline coefficient  63  4.20 Impact of hydrodynamic conditions and membrane configuration on pseudo-steadystate permeate flux  66  4.21 Impact of hydrodynamic conditions and membrane configuration on initial permeate flux decline coefficient  -  67  4.22 Impact of hydrodynamic conditions and membrane configuration on the pseudosteady-state permeate flux decline coefficient  67  LIST OF FIGURES  xi  4.23 Impact of the inverse of pseudo-steady-state permeate flux on the pseudo-steady-state permeate flux decline coefficient  71  A . 1 T O C concentrations of raw water and filtrate  -83  A.2 T O C removal efficiency  83  xii  LIST OF SYMBOLS  LIST OF SYMBOLS  Jv  the permeate flux [L/m hr]  a, b, c, d and b'  empirical constants  a  the initial permeate flux decline coefficient [L" ]  b  the pseudo-steady-state permeate flux decline coefficient [L" ]  1  1  2  1  b'  a modified pseudo-steady-state permeate flux decline coefficient [m" hr" ]  Jo  the initial permeate flux [L/m""hr]  J'ss . the pseudo-steady-state permeate flux [L/m hr] V. corresponding to the filtration volume [L]. Qp filtrate flow rate.[L/hr] A  membrane surface area [m ]  V  volume filtered over a given time period [L]  / AP  filtration period [hr] the transmembrane pressure [N/m ]  //  the absolute viscosity of the water being filtered [N's/m ]  Rm  the hydraulic resistance of the clean membrane to permeate flow [Ns/m ]  /?/•  the hydraulic resistance caused by retained material to permeate flow, [Ns/m ]  2  •y  3  AP(op) the actual operated transmembrane pressure [psi]  3  xiii  LIST OF SYMBOLS 2  Jv(op) the permeate flux measured at AP(OP) (calculated from Equation 2.1 in units of L/m hr) Jv(4j)  the permeate flux corrected to a reference transmembrane pressure of 4. lpsi (27.7 kPa)  using Equation 3.1. Jvi5  the permeate flux normalized to 25°C [L/m hr]  T  the temperature of the liquid matrix being filtered [°C]  2  A CKNO WLEDGEMENTS  xiv  ACKNOWLEDGEMENTS  This research would not have been possible without the assistance and guidance of my supervisor Dr. Pierre Berube. I sincerely appreciate his precious instruction to me throughout this research. The financial support for this research was provided by the Natural Science and Engineering Research Council of Canada (NSERC).  Special thanks go to the following  people for their help throughout the present research: o  Yubo Liu for his encouragement and backup throughout this research  o  Paula Parkinson for all her help in finding and fitting parts and TOC analysis in the laboratory  «  Susan Harper for providing me with everything that I needed.in the laboratory  © Bill for building my reactor and miscellaneous parts •  Scott for setting up the data loading system  o  Wendong Tao for his help with collecting Raw water from Jericho pond  o  Adeline S. Chin for her help with collecting Raw water from Seymour  •  James Craig Bradley for his help in conducting the crossflow velocity test  This thesis is dedicated to my father Zicai Lei whose memory kept me going throughout my studies and my mother Guilian Zhang whose encouragement helped me through the hardship.  CHAPTER 1 INTRODUCTION AND RESEARCH OBJECTIVES  1  CHAPTER 1 INTRODUCTION AND RESEARCH OBJECTIVES 1.1 Introduction Since the development of synthetic asymmetric membranes in 1960, the international interest in research, development, commercial activity, and full-scale application of membrane processes for water and wastewater treatment has grown steadily (Mallevialle et al., 1996). Amongst all of these activities, the commercialization of backwashable hollow-fiber micro filtration (MF) and ultrafiltration (UF) membrane processes for the removal of particulate matter (i.e., turbidity and microorganisms) have had the most profound impact on the use, acceptance, and regulation of all types of membrane processes for drinking water treatment (US EPA, 2003).  The use of submerged membrane processes has increased rapidly for drinking water treatment applications. One of the major reasons driving this increase is the relatively low operating cost associated with submerged membrane systems, compared to their external counterparts. However, the operating costs for submerged membrane systems are still relatively high, when compared to those associated with conventional treatment technologies, such as sand filtration. The capital and operating costs associated with membrane systems are typically proportional to the membrane permeate flux that can be achieved (Mallevialle et al., 1996). Thus, the permeate flux that can be achieved, and the factors that impact the permeate flux are central considerations in determining membrane performance and cost. Moreover, the permeate flux  CHAPTER 1 INTRODUCTION AND RESEARCH OBJECTIVES  2  that can be achieved is largely governed by the extent of fouling that occurs on the membrane surface (Mallevialle et al., 1996).  The impact of air sparging on confined membrane systems, such as a conventional cross-flow tubular membrane system, has been extensively investigated. In general, under slug air flow conditions, an increase in the extent of air sparging results in an increase in the permeate flux that can be achieved in a membrane system (Judd et al., 2001). This increase in the permeate flux that can be achieved is likely due to the high shear forces that are generated at the surface of a membrane, as a slug of air travels along this surface (Cabassud et al., 2001). These high shear conditions increase the mass transfer of retained material away from the membrane surface. There are two main mechanisms that contribute to the high surface forces that are generated by air sparging in confined (e.g. tubular) membrane systems. The first mechanism results from the high liquid velocities that occur in the falling film which forms between the rising slug of air and the membrane surface. Based on a numerical analysis, Ghosh and Cui (1999) reported that, for a given bulk cross-flow in a tubular membrane system, the mass transfer that is generated at a membrane surface by the falling film is approximately two orders of magnitude greater under air sparged (i.e. slug air flow) conditions, when compared to single phase (i.e. water only) flow. The second mechanism results from the significant eddies that are generated in the wake of an air slug. Based on a numerical analysis, Judd et al. (2001) reported that short periods of very high surface shear conditions were associated with the passage of the wake of an air slug along a membrane surface. Depending on the size of the air slug, the magnitude of the surface shear force associated with the wake was twice as large as that associated with the falling film.  Similar trends were observed experimentally by  CHAPTER 1 INTRODUCTION AND RESEARCH OBJECTIVES  3  Nakoryakov et al. (1989), when using an electrochemical cell to measure the mass transfer in the wake of an upward slug flow in a pipe. A review of the impact of air sparging on confined membrane systems is presented in Cui et al. (2003).  As with their confined counterparts, unconfined membrane systems are also significantly impacted by air sparging. However, in an unconfined system, such as a submerged hollow fiber membrane system, other mechanisms can also potentially contribute to the high shear forces that are generated at a membrane surface, notably the shear forces resulting from the swaying of a membrane fiber in the wake of an air slug. The lateral velocity of a swaying fiber can be relatively high, resulting in high shear forces at the membrane surface (Cui et al., 2003). The physical contact between swaying fibers can also potentially generate high shear forces at the membrane surface. The magnitude of the surface shear force, resulting from the lateral movement and the physical contact, is dependent upon a number of complex parameters relating to the configuration of the submerged membrane system and the air sparging practices. Unfortunately, it is not possible to accurately estimate the magnitude of these forces using currently available computational methods. As a result, the relative contribution of these mechanisms on the resulting improvements in permeate flux is unknown, and the impact of air sparging on submerged hollow fiber systems is poorly understood.  1.2 Aim and Objectives  The present study was designed to address the current knowledge gap and undertaken to investigate the impact of air sparging and the system configuration on the permeate flux in a submerged hollow fiber membrane system. This study was designed to investigate the relative  CHAPTER 1 INTRODUCTION AND RESEARCH OBJECTIVES  4  impact of liquid crossflow along membrane surface, air bubble contact with membrane surface and physical contact between membrane fibers on the permeate flux in a submerged membrane system.  Throughout this study, a better knowledge of air sparging mechanisms will enable submerged member systems to be designed and operated to maximize the permeate flux.  1.3 Thesis Structure In this thesis, the study aims and objectives are articulated above. Chapter 2 presents a review of relevant literature. Chapter 3 describes general experimental set-up and experimental approach. Chapter 4 deals with the experimental results and discussions. The conclusions of this study are presented in Chapter 5.  CHAPTER 2 L1TERATURE REVIEW  5  CHAPTER 2 LITERATURE REVIEW 2.1 Definition of M e m b r a n e  A membrane is defined as a thin film separating two phase and acting as a selective barrier to the transport of matter (Mallevialle et al., 1996, as illustrated in Figure 2.1. The different types of membranes are typically classified based on the pore size of the membrane surface, the configuration of the membrane modules and their physical arrangement in a particular treatment system. Membrane  Driving force(AC, AP, AW) Q  Larger size material which is retained by the membrane  •  Smaller size material which can go through the membrane Figure 2.1: Selective permeation across a membrane  CHAPTER  2  LITERA  TURE  6  REVIEW  2.2 Membrane Classification 2.2.1 Classification based on a pore size of membrane material Membranes are usually classified based on their pore sizes. Microfiltration (MF) and Ultrafiltration (UF) membranes, also referred to as low pressure systems, are characterized by their ability to remove suspended or colloidal particles via a sieving mechanism; this is based on the size of the membrane pores relative to that of the particulate matter. Nanofiltration (NF) and Reverse Osmosis (RO) membranes, also referred to as high pressure systems, are characterized by their ability to remove dissolved containments, as well as larger materials (i.e. suspended or colloidal particles). The type of membrane used for a specific application is dictated by the type of material that must be separated. A summary of the categories of membrane typically used in drinking water treatment applications, and their relevant characteristics, is presented in Table 2.1 and Figure 2.2. Table 2.1: Membrane classification and application Classification  Microfiltration (MF)  Ultrafiltration (UF)  Pore size (pm) Mechanism  0.1-0.2 Sieving  0.01-0.05 Sieving  Required Driving Pressure (Bar) Permeate Flux (L/m .hr) Retained Materials  0.1-0.3  Reverse Osmosis (RO) 0.0001-0.001 Sieving, Diffusion 10-100  Nanofiltration (NF) 0.0007-0.01 Sieving, Diffusion 3-30  50-500  20-200  4-40  8-80  Particulate matter, Bacteria.  Colloidal material and macromolecultes, viruses.  Suspended and dissolved species except for water and some small dissolved salts.  Organics  2  (Adapted from Hillis, 2000)  7  CHAPTER 2 LITERA TURE RE VIEW  Micromolecular organics Material of Concern  Colloids Macromolecular organic  Dissolved salts  Viruses  Bacteria Yeasts  Size ((im) Molecular Weight (Daltons)  0.0001  0.001  0.01  200  20,000  10  100  200,000  Algae Drinking Water Pathogens  Giardia  L  Cryptosporidiihm  Membrane Filtration Process  Sand filtcn  Mr  J\ZI..I..:::U:"": Figure 2.2: Filtration application guide for different materials removal (Adapted from US EPA (2003) and Hillis (2000))  CHAPTER  2  LITERATURE  8  REVIEW  2.2.2 Classification based on configuration of membrane module A number of different types of membrane modules are commonly found on the market. In addition to the membrane material itself, the membrane modules include the support structure, inlet and outlet ports, and permeate ports. In drinking water applications, the four main types of membrane modules that are commonly used are illustrated in Figure 2.3. A detailed description of these types of membrane modules can be found in Mallevialle et al. (1996). (a)  Membrane -support plate- membrane assembly Feed in •» / , , Concentrate  (b)  Raw water  ^r  Feed side Spacer—z^XXXXXXXXXV  =ieB  Permeate * " ; .-Concentrate Permeate  Permeate collector Membrane  (c) Raw water  Concentrate Permeate (d)  Permeate Concentrate  Feed-*s^  W^^^<\ "Potting  \ Hollow fibers  Perforated sleeve Figure 2.3: Schematic representation of the four principal types of membrane modules  CHAPTER  2 LITERATURE  9  REVIEW  (a) plate and frame; (b) spiral module; (c) tubular module; (d) hollow fiber. (Adapted from Mallevialle et al., 1996) As presented in Chapter 3, outside-in flow, hollowfibre ultrafiltration membrane modules were used in this present study.  2.2.3 Classification based on the physical arrangement in a particular treatment system Membrane systems for drinking water treatment are typically loosely defined either as external or internal systems, a) External systems: In external systems (Figure 2.4), the liquid matrix to be filtered is pumped from a feed water tank through a membrane filter. Some of the liquid permeates from the inside of the membrane and is collected on the outside (i.e. inside-out flow). The pressure generated by the pump provides the driving pressure differential (i.e. transmembrane pressure) that allows the permeate to flow through the membrane. The material that does not permeate through the membrane is returned to the feed water tank. The tangential flow of liquid along the membrane surface provides the high shear conditions needed to maintain a high permeate flux. Some of the recirculated material is purged to prevent excessive build-up of retained material in the system.  Raw water  .  \Z  Purge  *  Recirculation loop  i  Feed water  Permeate Recirculation pump  Membrane filter  CHAPTER 2 LITERATURE REVIEW Figure 2.4: Typical operation mode for external membrane systems  b) Internal systems: In internal systems (Figure 2.5), the membrane is submersed in the feed water tank. The liquid permeates from the outside of the membrane and is collected on the inside of the membrane (i.e. outside-in flow). A vacuum pump provides the driving pressure differential that allows the permeate to flow through the membrane. Air sparging is used to provide the high shear conditions needed to maintain a high permeate flux. Some of the liquid in the tank is purged to prevent excessive build-up of retained material in the system. Internal systems are also commonly referred to as submerged systems. Raw water  Vacuum pump Permeate  Air sparging  Figure 2.5: Typical operation mode for internal membrane systems As presented in Chapter 3, a submerged, hollow-fiber ultrafiltration membrane system was used in the present study. Unless indicated otherwise, all of the discussions in the remainder of the present thesis focus on submerged, hollow-fiber, untrafiltration membrane systems.  2.3 Material Accumulation and Permeate Flux Decline  11  CHAPTER 2 LITERATURE REVIEW  Membrane productivity is typically defined in terms of permeate flux, which is defined as the filtrate flow per unit of membrane area, as shown in Equation 2.1 (Mallevialle et al., 1996).  jv=ee=-il A  (in  t*A  Where, Jv is the permeate flux [L/m hr]; Qp is filtrate flow rate [L/hr]; A is membrane 2  surface area [m ]; V is volume filtered over a given time period [L]; and / is filtration period 2  [hr]. The permeate flux (Jv) of clean water across a membrane can be estimated using Darcy's law (Mallevialle etal., 1996). Jv =  (2.2)  p * Rm  Where, AP is the transmembrane pressure [N/m ]; p is the absolute viscosity of the water 2  being filtered [Ns/m ]; and Rm is the hydraulic resistance of the clean membrane to permeate 2  flow [Ns/m ]. 3  However, when filtering a liquid matrix, materials accumulate on and within the membrane by blocking or constricting pores and by forming a layer. Overtime, material that is retained at the membrane surface can accumulate, as illustrated in Figure 2.6.  Material accumulation, which is also commonly referred to as fouling, may arise from particle deposits on the membrane surface, macromolecules adsorbing onto the membrane surface, or within the membrane pores. As a result, the membrane resistance to the permeated flow increases over time. For a given operating transmembrane pressure and liquid matrix, an  CHAPTER 2 LITERA TURE REVIEW  12  increase in the resistance due to the accumulation of retained material results in a decline in the permeate flux, as suggested by Equation 2.3 (Hermia, 1982; Field et al., 1995). Jv =  AP  (2.3)  fi * (Rm + Rr)  Where, Rr is the hydraulic resistance caused by retained material to permeate flow [Ns/m ] 3  Concentrate fluid flow  macromolecule  Membrane  Permeate Figure 2.6: Materials accumulation during membrane filtration process (Adapted from Mallevialle et al., 1996) Considering that overtime, the extent of material that accumulates on the membrane surface increases, therefore, the permeate flux through a membrane decreases with time. Earlier studies by Hermia (1982) have shown that the decline in the permeate flux occurs in two stages, as illustrated in Figure 2.7. The initial stage is characterized by a short term rapid flux decline due to pore blocking and cake formation. The second stage is characterized by a long  CHAPTER 2 LITERA TURE REVIEW  13  term gradual flux decline, due to cake compaction precipitative fouling, and/or adsorptive fouling.  Initial sharp flux decline LpJ!£ term gradual flux dec line  Filtration process over time or volume filtered Figure 2.7: Schematic representation of permeate flux decline The rate at which the permeate flux declines is governed by the rate at which material accumulates at the membrane surface. Under low shear conditions (i.e. low mass transfer rate away from the membrane surface) material accumulates relatively rapidly. As discussed in Section 2.4, air sparging is used in submerged hollow fiber systems to induce shear at the membrane surface. Under high shear conditions (i.e. high mass transfer rate away from the membrane surface) material accumulates at a relatively low rate. Therefore, the rate at which the permeate flux declines can be controlled by varying the magnitude of the shear that is imposed onto the membrane surface. The accumulated material can also be periodically removed by relaxation and/or backwash cycles (Figure 2.8). During a relaxation cycle, the permeate flow is interrupted (i.e. driving force is removed), while the applied shear is maintained. During a backwash cycle, the direction of the permeate is reversed while the  CHAPTER 2 LITERA TURE REVIEW  14  applied shear is maintained. However, even by applying high shear forces along the membrane surface, back flushing and/or relaxing the membrane, some of the accumulated foulants cannot be removed. As a result, the permeate flux that can be achieved following a backflush cycle and/or a relax cycle decreases over time (Figure 2.8).  irreversible (fouling)  x 3  CD  Figure 2.8: Reductions in permeate flux over time with periodic cleaning (Adapted from Mallevialle et al., 1996) 2.4 A i r Sparging in Submerged Hollow Fiber Membrane System  One of the main limitations for membrane system is the permeate flux decline that occurs over time. This decline is associated with material accumulation on the membrane surface. Air sparging is used in submerged membrane systems to minimize the extent of material accumulation on the membrane surface. Air sparging induces shear forces onto the surface of a hollow-fiber, submerged membrane surface.  Ozaki and Yamamoto (2001) investigated the effects of crossflow velocity induced by airsparging on the foulant accumulation process in a bioreactor, using flat sheet membrane modules. They observed that foulant accumulation rate was initially very low, followed by an  CHAPTER  2  LITERATURE  REVIEW  15  increase in foulant accumulation rate. After approx 6 h of filtration, the accumulation rate decreased gradually. The rate of foulant accumulation is schematically shown in Figure 2.9.  Sludge accumulation (mg»cm) Equivalent sludge accumulation  Maximum sludge accumulation rate Lag phase  »  Time(hr)  Figure 2.9: Schematic of sludge accumulation process Ozaki and Yamamoto (2001) further observed that the extent of sludge accumulation, the sludge accumulation rate, and the lag phase were dependent on aeration intensity. Based on their investigation, the following conclusions were reached. 1) The extent of sludge accumulation decreased significantly with an increase in the aeration intensity, whereas the effect of changes in mixed-liquor suspended solids (MLSS) concentration on the extent of sludge accumulation was substantially smaller. 2) The sludge accumulation rate decreased as the aeration intensity increased. The mixedliquor suspended solids concentration did not have a substantial impact on the sludge accumulation rate. 3) The duration of the lag phase increased with aeration intensity. This means that, initially, sludge is less susceptible to attach itself to the membrane surface at higher aeration intensities.  CHAPTER 2 LITERATURE REVIEW  16  Therefore, the extent of sludge accumulation and the sludge accumulation rate are significantly impacted by the extent of air sparging.  Chang and Fane (2002) investigated the impact of air sparging intensity on the vacuum pressure needed to maintain a constant permeate flux in a submerged hollow fiber membrane system. They observed that the extent of the increase in the vacuum pressure (i.e. over time) was lower at higher air sparging intensities. They also observed that the beneficial impact of air sparging on reducing the extent of the increase in the vacuum pressure plateaued when the air sparging intensity was increased above a certain value.  In addition to the sparging intensity, the configuration of submerged hollow fiber membrane systems can also significantly impact the permeate flux. Chang and Fane (2002) investigated the impact of looseness of the fibers in a submerged membrane system. As illustrated in Figure 2.10, the vacuum pressure, needed to maintain a given permeate flux, increased more rapidly for membrane systems with tightly configured modules.  0  60  120 180 time (min)  240  300  CHAPTER 2 LITEM PURE REVIEW  17  Figure 2.10: Comparison of suction pressure profiles between loose and tight hollow, fiber membrane system (Adapted from Chang and Fane, 2002)  In addition, Chang and Fane (2002) observed that the beneficial impact of loosely configured membrane modules was most significant at lower sparging intensities (Figure 2.11).  200 I light  160 1/5 <U  a  3  CO  I loose  120 80 40 0 1  Figure 2.  1.8  _  11  Gas flow rate (dm /min)  Increase in suction pressure with tight and loose fiber bundles versus different gas flow rates (Adapted from Chang et al., 2002)  Commercial manufactures  of submerged hollow-fiber  systems have recognized the  importance of a certain degree of fiber looseness in the membrane module configuration (e.g. K. Ohkubo, 1988; Ebara Corp, 1989).  18  CHAPTER 3 EXPERIMENTAL DESIGN  CHAPTER 3  EXPERIMENTAL DESIGN  3.1 Bench-scale membrane system  This study was designed to investigate the relative impact of the liquid crossflow along a membrane surface, the air bubble contact with a membrane surface and the physical contact between membrane fibers on the permeate flux in a submerged membrane system. The study was conducted using the bench-scale submerged membrane system illustrated in Figure 3.1 and Figure 3.2. L_J Vacuum System  System Tank  P e r m e a t e Flux Collection and  Aeration System  Measurement Pressure Gauge  Vacuum Valve Thermometer Vacuum Line  Vacuum Regulator  Vacuum  Vacuum  Chamber  Chamber  (I)  (II)  Mesh  ^Cylindrical Tank Cylindrical  Filtrate C o l l e?ctor ctor  111  Baffle  Air Flowmeter Air  Computer Peripheral  Scale  N  • Membrane  Aerator-  -XCentral Aerator  Drain  Figure 3.1: Schematic of the laboratory bench-scale set-up The submerged membrane system consisted of a system tank, membrane modules, an aeration system, a system vacuum, and a permeate flux collection and measurement system (Figure  CHAPTER 3 EXPERIMENTAL DESIGN  19  3.1). The main components of the bench scale membrane system include: System tank, Membrane module, Aeration system, Vacuum system, and Data collection system.  (a) Laboratory experimental setup  (b) Membrane state  Figure 3.2: Picture of the bench-scale experimental set-up 3.1.1 System tank The system tank consisted of an open top, Plexiglas cylindrical tank with a diameter of 0.14m and a height of 1.4m. The system tank had a working depth of 1.0 m, resulting in a working volume of 16L. A Plexiglas cylindrical baffle, with a diameter of 0.07m and a height of 0.8m, was placed in the cylindrical tank along the vertical center line of the system tank, and 0.025m from the bottom of the cylindrical tank, as illustrated in Figure 3.3.  CHAPTER 3 EXPERIMENTAL DESIGN  20  Figure 3.3: Schematic of system tank with cylindrical baffle  The cylindrical baffle enabled the hydrodynamic conditions in the submerged membrane tank to be controlled (see Section 3.1.3). A valve with a diameter of 8mm, located 20mm from the bottom of the cylindrical tank, enabled the liquid in the tank to be drained between each experiment.  CHAPTER  3  EXPERIMENTAL  21  DESIGN  3.1.2 Membrane module The membranes used in this study were submerged (i.e. outside-in flow) hollow fiber membranes manufactured by Zenon Environmental Inc. (Oakville, ON). The physical characteristics of the membranes are presented in Table 3.1. Table 3.1: Physical characteristics of the membranes used in the present study Configuration  Outside diameter  Outside-in hollow-fiber  1.8mm  Surface Nominal pore diameter properties Non-ionic & 0.04pm hydrophilic  Typical operating transmembrane pressure 1-8 psig  The membrane modules were placed at the vertical centre-line of the system tank, as illustrated in Figure 3.4.  i  B  Figure 3.4: Schematic picture of membrane module  CHAPTER  3  EXPERIMENTAL  22  DESIGN  3.1.2.1 Membrane modules configuration Two different membrane modules were used in this research. The first consisted of a singlefiber membrane module. The second consisted of a multi-fiber membranes bundle. For the multi-fiber bundle configuration, although 8 fibers were assembled together, only one of the fibers was actively used to filter. This fiber was placed at the centre of the bundle. The other seven "inactive" fibers were placed surrounding the "active'" fiber. The length of the fibers in each membrane module was 0.42m long, with a total effective membrane surface area of 0.0023m . The configurations of the membrane modules are shown in Figure 3.5. 2  • • •  *""  1  ••'  3-.  (a) Single-fiber module  n  (b) Multi-fiber module  Figure 3.5: Pictures of membrane fibers and membrane modules  3.1.2.2 Potting of the membrane modules The membrane fibers were fixed (i.e. potted) into a top and a bottom bracket as illustrated in Figure 3.6.  Epoxy glue was used to pot the membrane fibers into the brackets. The end of the membrane fiber(s) that was potted into the bottom bracket was sealed. For the single fiber membrane module, the end of the membrane fiber, that was potted into the top bracket, was open. For the multi-fiber membrane module, only the 'active' fiber of the bundle was open. The permeate  CHAPTER 3 EXPERIMENTAL  DESIGN  23  was collected from the open liber end in the top bracket. The 'inactive' libers were potted into the top bracket and sealed as illustrated in Figure 3.6. Vacuum  Vacuum  0.125rrJ ^ J o p bracket (00.013m)  111  0.11m  hpoxy-s  61  'active' fiber 'active' liber 'inactive' fiber  Epoxy,  5m J  Bottom bracket  40mm r~-~77>  5mm  13 mm Single-liber module  Multi-fiber module  Figure 3.6: Schematic of membrane modules  3.1.2.3 Loose and tight membrane fibers Both loose and tight membrane module configurations were considered. For the tight configuration, the spacing between the brackets, when the membrane module was installed in  CHAPTER 3 EXPERIMENTAL DESIGN  24  the system tank, was set to 42cm, as illustrated in Figure 3.7. For the loose configuration, the distance between the brackets was set to 41cm, as illustrated in Figure 3.7.  (a) Membrane module in tight configuration (b) Membrane module in loose configuration Figure 3.7: Schematic of loose and tight configuration for the membrane modules 3.1.3  A eration system  The aeration system was used to control the hydrodynamic conditions in the system tank. Two aerators were located in the submerged membrane tank. The central aerator consisted of a perforated plate, with 12-2mm diameter openings (Figure 3.8).  25  CHAPTER 3 EXPERIMENTAL DESIGN  Base of the system tank . Perforated plate  Figure 3.8: Central aerator The peripheral aerator consisted of a perforated PVC ring, with 18 openings of different diameters (Figure 3.9). The diameter of the openings in the peripheral aerator ranged from 1mm to 3mm. At the edges, where compressed air was added, the diameter of the openings was 1mm. As the radial distance from the edges where the compressed air was added increased, so did the diameter of the openings, as illustrated in Figure 3.9. This configuration enabled the air flow from the peripheral aerator to be evenly distributed throughout the ring. <D3mm  Compressed air Figure 3.9: Peripheral aerator  CHAPTER  3  EXPERIMENTAL  DESIGN  26  The central aerator was located directly below the cylindrical baffle, at the base of the system tank. The peripheral aerator was located between the cylindrical tank and the cylindrical baffle, at a height of 0.17m above the base of the submerged membrane tank. A schematic of the aeration system is presented in Figure 3.10.  S o n. <u Q 3  Flow Control Valve Compressed Air Line Figure 3.10: Schematic of aeration system  CHAPTER  3  EXPERIMENTAL  DESIGN  27  The air flow rate to each aerator was controlled using an Air Valve (VERIFLO IR 500D35001273) and an Air Flowmeter (Cole-Parameter: N034-39(ST)).  3.1.3.1 Control of system crossflow Both single phase (water only) and dual phase (water + air sparging) crossflow, along the membrane surface, were studied. Dual phase crossflow along the membrane surface was generated by aerating the system tank with the central diffuser. The rising air bubbles, which were confined to the inside of the cylindrical baffle, entrained water upwards along the inside of the cylindrical baffle. A schematic of the flow path for the dual phase crossflow is illustrated in Figure 3.11(a). Single phase crossflow along the membrane surface was generated by aerating the system tank with the peripheral aerator. The rising air bubbles entrained water upwards along the outside of the cylindrical baffle, and subsequently downwards through the inside of the cylindrical baffle. A mesh screen located at the top of the cylindrical baffle ensured that no air bubbles were entrained downwards into the cylindrical baffle. A schematic of the flow path for the single phase crossflow along the membrane surface is illustrated in Figure 3.11 (b).  For both single phase and dual phase crossflow operation, the bulk liquid crossflow velocity at the membrane surface could be controlled by varying the flow rate of air into the respective aerators. The relationship between the air flow rate and the bulk crossflow velocity at the membrane surface, for both single phase and dual phase crossflow, is presented in Figure 3.12.  CHAPTER 3 EXPERIMENTAL  DESIGN  Mesh Water Surface .Cylindrical baffle  Membrane  ^^Cylindrical tank Peripheral aeration Air . Central aeration Aii-  (a) Dual phase crossflow Figure 3.11  (b) Single phase crossflow  Schematic of system crossflow (arrows represent direction of bulkflow)  A tr-tto w r a t e  (p s i G )  Figure 3.12: Relationships between bulk crossflow velocity and air flowrate  28  CHAPTER 3 EXPERIMENTAL DESIGN  29  For the present study, bulk crossflow velocities of 0.2, 0.3 and 0.4 m/s were considered. The corresponding air flow rates (and Air flow meter readings), to achieve these bulk crossflow velocities for both the single phase and the dual phase crossflow systems, are presented in Table 3.2. Table 3.2: Operating bulk crossflow velocity and air flowrate Bulk crossflow  Single phase crossflow system  Dual phase crossflow system  velocity (m/s) 0.2  Air flowrate (ml/min) 3800  Air flow meter readings 30  Air flowrate (ml/min) 3200  Air flow meter readings 25  0.3  6400  51  5300  42  0.4  10200  82  7500  60  3.1.3.2 Control of bulk crossflow velocity The relationships between the bulk crossflow velocities at membrane surface and the air flowrate readings for single phase and dual phase crossflow (i.e. Figure 3.12) were determined using dye injection and time lapse imaging. See Figure 3.13 for typical pictures of bulk crossflow velocity measurement under single phase crossflow.  1  2  3  4  Figure 3.13: Pictures of bulk crossflow velocity measurement (overall mass highlighted)  CHAPTER 3 EXPERIMENTAL DESIGN  30  For the bulk crossflow velocity measurement for single phase crossflow, dye (Rhodomine dye) drops were added at the water surface at the center of the cylindrical tank, as shown in Figure 3.14 (a). Time lapse images were taken using a digital camera (Cannon G3) in rapid sequence while the dye drops traveling downward. Based on the distance traveling by the dye over time, it was possible to estimate the overall bulk velocity.  For the bulk crossflow velocity measurement for dual phase crossflow, the bulk crossflow velocities were measured in a similar manner as described above. However, the dye drops were injected through a tube inserted into the drain line at the bottom of the system tank, as shown in Figure 3.14 (b).  (a) Single phase-water only crossflow  (b) Dual phase-water and air crossflow  Figure 3.14: Dye injection for bulk crossflow velocity measurement  CHAPTER 3 EXPERIMENTAL DESIGN  3.1.4  31  System vacuum  The system vacuum provided the driving pressure differential (i.e. transmembrane pressure) that enabled the permeate to flow through the membrane. All experiments were performed at a constant transmembrane pressure and the permeate flux was monitored over time (See section 3.1.5).  The vacuum system consisted of a vacuum regulator (Bellofram:77MOD 960-502-00), two vacuum chambers (4L) and a vacuum valve (VERIFLO IR500D3500-1273) (see Figure 3.1). The vacuum system was directly connected to the vacuum line in the laboratory. Initially, significant fluctuations in the vacuum occurred. A vacuum regulator and two buffers were used to minimize these fluctuations. With this configuration, the fluctuations were drastically reduced, and the vacuum varied by less than 0.15 psi during each test run.  The transmembrane pressure (TMP) was adjusted using the vacuum regulator and monitored using a pressure gauge installed on the vacuum line between the membrane and the permeate collection container. Care was taken to maintain the transmembrane pressure constant during filtration. However, at the beginning of each experiment, a lag-phase was observed during which time a lower transmembrane pressure existed for a period of approx 15 minutes. After this lag-phase, the transmembrane pressure remained relatively constant and varied by less than +0.15psi during each experiment. Throughout this study, the average transmembrane pressure value of each test ranged between -4.0psi (-27.7 kPa) and -4.3psi (-29.7 kPa). A typical transmembrane pressure distribution during a filtration experiment is shown as Figure 3.15.  CHAPTER  3  EXPERIMENTAL  32  DESIGN  The magnitude of the vacuum that was applied is equivalent to the pressure differential that is typically applied to a full-scale constant flow (variable pressure) submerged membrane system at the start of a filtration cycle.  Lag-phase  \  0  1  1  1  1  1  1  20  40  60  80  100  120  140  Time (min)  Figure 3.15: Typical trend of transmembrane pressure  3.1.5  Permeate flux collection and measurement system  The flow measurement system consisted of a collection container, a digital electrical scale, and a computer (see Figure 3.1). The digital scale (Scout Pro 4000) was used to measure the weight of permeate which accumulated in the collection container over time. The cumulative weight was automatically logged into a computer at one-minute intervals, and automatically connected to a permeate flow rate. Periodically (i.e. every 1 hour), a volume of raw water,  CHAPTER 3 • EXPERIMENTAL DESIGN  33  equivalent to that which was collected over the past hour, was added to the system tank to maintain a relatively constant liquid level.  3.2 Source Water to be Filtered The source water used in the present study consisted of a mixture of raw water from the Seymour Reservoir and Jericho Pond. The Seymour Reservoir is located in North Vancouver and supplies almost 40% of the drinking water requirements for the Great Vancouver Regional District (GVRD). Raw water from the Seymour Reservoir is characterized by a low turbidity (<2NTU) and low organic content (<3mg/L as Total Organic Carbon-TOC). Because of its relatively high quality, raw water from the Seymour Reservoir has a very low fouling potential. Preliminary tests indicated that it would take a number of days (to weeks) for a significant amount of material to accumulate and foul the membrane surface, when filtering raw water from the Seymour Reservoir. To enable a greater number of experimental conditions to be considered within a shorter timeframe, raw water from the Seymour Reservoir was mixed with raw water from the Jericho Pond at a 5:1 ratio. Raw water from the Jericho Pond has a much higher organic content than that from the Seymour Reservoir. The resulting raw water mixture had a TOC of approximately lOmg/L, and a relatively higher fouling potential. When filtering the mixture from Seymour Reservoir and Jericho Pond, it was possible to significantly foul the membrane surface (i.e. achieve a significant reduction in the permeate flux) within one to two days.  Raw water from the Seymour Reservoir was collected using 15-20L containers on August 20 , th  2003. The containers were stored in a refrigerator at 4°C in the Environmental Engineering  CHAPTER 3 EXPERIMENTAL DESIGN  34  Laboratory at UBC. Raw water from the Jericho Pond was collected using 6-20L containers on Sept. 27 , 2003, and stored along with the raw water from Seymour Reservoir on Oct. 3 , th  rd  2003.  Algae were observed at the surface of the raw water from Jericho Pond. The algae were removed by removing the surface water sample from the stored containers.  The supernatants of the raw water samples from the Seymour Reservoir and the Jericho Pond water were then mixed together in 5:1 ratio, and stored in a 500L storage tank. As a result, it was possible to conduct all of the experiments (see Section 3.3) is in the same raw water mixture. The storage tank was kept in the refrigerator at 4 °C during this study. Before each filtration test, 16L of raw water was collected from the tank and brought to room temperature.  3.3 Experimental Program  Five series of experiments were performed as part of this study. Two series of experiments were designed to investigate the impact of the hydrodynamic conditions on the permeate flux in a submerged membrane system. Another two series of experiments were designed to investigate the impact of the membrane configuration on the permeate flux in a submerged membrane system. A fifth series of experiments were performed under static water (i.e. no crossflow). This fifth series of experiments represent datum conditions against which all other series of experiments were to be compared. Key hydrodynamic condition and membrane configuration characteristics of each experiments studied in this research are summarized in Table 3.3.  35  CHAPTER 3 EXPERIMENTAL DESIGN  For the test names in Table 3.3, the first character in the experimental name corresponds to the number of fiber in the membrane module (i.e. 1 for single fiber modules and 8 for multifiber modules); the second character corresponds to the hydrodynamic conditions (i.e. 1 for single phase/water only crossflow, and 2 for dual-phase/air sparged crossflow); the third character corresponds to the tightness of the membrane module (i.e. T for tight and L for loose); and the last character corresponds to the bulk crossflow velocity (i.e. 0 for static condition and A, B, and C for crossflows of 0.2, 0.3 and 0.4 m/s, respectively).  Table 3.3: Key characteristics of hydrodynamic condition and membrane configuration of each experiment  Experiment series  Purpose of experiments  Hydrodynamic condition  Membrane configuration  1  Impact of single phase crossflow on permeate flux Impact of dual phase crossflow on permeate flux Impact of physical contact on permeate flux  Water (single phase) crossflow only along membrane surface  Single-tight fiber  Water flow + air sparging (dual phase) crossflow along membrane surface Water flow + air sparging (dual phase) crossflow on membrane surface + physical contact between membrane fibers Water crossflow only on membrane surface + limited physical contact between membrane fibers Water flow + air sparging dual phase crossflow on membrane surface + limited physical contact between membrane fibers Static water  Single-tight fiber  2  3  4  5  Impact of membrane fiber configuration on permeate flux  Datum condition  Bulk crossflow velocity (m/s) 0.2  Test name  0.3 0.4  1-1-T-B 1-1-T-C  0.2 0.3  1-2-T-A 1-2-T-B  0.4  1-2-T-C  7 loose fibers surrounding one loose "active" fiber  0.2 0.3 0.4  8-2-L-A 8-2-L-B 8-2-L-C  7 tight fibers surrounding one tight "active" fiber  0.3  8-1-T-B (compari ng to 1 -1 T-B)  7 tight fibers surrounding one tight "active" fiber  0.3  8-2-T-B (compari ng to 1-2T-B)  Single-tight fiber  0  1-1-T-0  1-1-T-A  CHAPTER 3 EXPERIMENTAL DESIGN  36  3.4 Routine Membrane Integrity Testing and Maintenance  3.4.1 Membrane integrity testing Before starting each experiment, the integrity of the membrane module used was verified using a bubble test. If air bubbles were observed to escape from a leak in the membrane system, the integrity of the membrane was compromised and the membrane module was discarded. A schematic of membrane integrity testing apparatus is illustrated in Figure 3.16.  Pressure Gauge  Figure 3.16 Schematic of the membrane integrity testing apparatus  CHAPTER 3 EXPERIMENTAL DESIGN  37  The membrane fibers used in the present study are designed to operate under vacuum with outside-in flow. For the integrity testing, the membrane fibers were pressured. Based on conversations with the membrane manufacture, the fibers should not be exposed to a pressure exceeding lOpsi for an extended period of time. Exposure to higher pressures for an extended period of time can damage the membrane and result in breaches in the integrity of the membrane. To ensure that the integrity testing itself would not damage the membrane fiber, the integrity testing was done at a pressure of 6psi. Based on discussions with the membrane manufacture, breaches in the integrity of the membrane, that are larger than approximately 5um, can be detected by performing an integrity test at a pressure of 6psi.  Fortunately, no breaches in the integrity of any of the membrane modules assembled for the present study were found.  3.4.2 Membrane cleaning As recommended by the membrane manufacture, after each experiment, the membrane fiber was cleaned using a sodium hypochlorite (NaCIO) solution (diluted from 6.0% Domestic Miroclean Bleach). The membrane cleaning procedure used during the present study was as follows: 1. The used membrane fiber was soaked into a 750ppm solution of sodium hypochlorite (NaCIO) solution (diluted from 6.0% Domestic Miroclean Bleach) for 16 hours; 2. The 750ppm solution of NaCIO was filtered through the membrane fiber at a -4.1 psi vacuum for a period of 20 minutes;  CHAPTER 3 EXPERIMENTAL DESIGN  38  3. The membrane fiber was transferred into a fresh 50ppm solution of NaClO. And the solution was then filtered through the membrane fiber at -4.1 psi for a period of 20 minutes; 4. The cleaned membrane fiber was then stored by soaking it in a 50ppm solution of NaClO; 5. Before a cleaned membrane fiber was used for filtration test, the membrane fiber was rinsed three times using distilled water. Distilled water was then filtered through the membrane fiber at -4.1 psi for a period of 20 minutes.  3.5 Analytical Methods  3.5.1 Temperature The water temperature in the system tank was monitored using a standard alcohol based thermometer (Fisher Scientific 14-997).  3.5.2 Transmembrane pressure The transmembrane pressure was monitored using a pressure gauge (Cole-Parameter) which was installed between the permeate collection container and the submerged membrane. The pressure gage was located 0.3m above the water level in the system tank.  3.5.3 Permeate flux  3.5. 3.1 Measurement In this study, the permeate flux was described as the volumetric rate of liquid flow through the membrane surface in units of L/m hr.  39  CHAPTER 3 EXPERIMENTAL DESIGN  As discussed in Section 3.1.5.1, the cumulative mass of liquid filtered was automatically logged into a computer at one minute intervals. The permeate flux was calculated using Equation 2.1.  3.5.3.2 Transmembrane pressure correction The experiments were performed at a relatively constant transmembrane pressure. However, the transmembrane pressure varied by approximately ±0.15psi. throughout each experiment. As presented in Equation 2.3, permeate flux is predicted to increase proportionally with an increase in transmembrane pressure. The transmembrane pressure has a direct impact on the permeate flux. To remove the impact of pressure aviations on the permeate flux, the measured permeate flux was corrected to a reference transmembrane pressure of 4.1 psi (27.7 kPa) using Equation 3.1. Jv(op)* 4.1 <4,)  *P{op)  (  U  Where, AP(op) is the actual operated transmembrane pressure [psi]; Jv(op) is the permeate flux measured at AP(op) (calculated from Equation 2.1 in units of L/m hr); Jv( j) is the permeate 4  flux corrected to a reference transmembrane pressure of 4.1 psi (27.7 kPa) using Equation 3.1. 3.5.3.3 Temperature correction All filtration experiments were performed at room temperature. Prior to the start of each experiment, the water mixture to be filtered was equilibrated to this temperature. The water mixture temperature for these different filtration experiments ranged from 18.5°C to 21.5°C.  C  H  A  P  T  E  R 3  E  X  P  E  R  I  M  E  N  T  A  L D  E  S  I  G  N  Temperature affects the viscosity of the liquid mixture being filtered. The viscosity of the water mixture has a direct impact on the permeate flux as presented in Equation 2.3. Since water temperature can have a significant impact on membrane permeate flux, it is common practice to "normalize" the permeate flux to a reference temperature during operation, for the purpose of monitoring system productivity independent of changes in water temperature. Using a reference temperature of 25°C, Pohland (1988) reports that the following expression for the temperature correction factor is correct within approximately 3 percent. Jv = 25  25  — 1.03*(r-25)  (3.3)  Where, Jv?j is the permeate flux normalized to 25°C, [L/m hr]; T is the temperature of the 2  liquid matrix being filtered, [°C].  All of the permeate flux valves presented in this study are standardized to a transmembrane pressure of 4.1 psi and temperature of 25°C.  CHAPTER 4 RESULTS AND DISCUSSIONS  CHAPTER 4  41  RESULTS A N D DISCUSSIONS  This chapter presents the experimental results obtained during the present study. As discussed in Section 3.3, and summarized in Table 3.3, five series of experiments were performed. For each of these series of experiments, the permeate flux was monitored over time.  4.1 Permeate Flux Decline over Time The results for all of the series of experiments are presented in Figure 4.1 in terms of permeate flux versus filtration time. The following overall trends can be observed. 1) For all experimental conditions, the permeate flux initially decreased rapidly. 2) The rate of initial rapid decrease in the permeate flux increased in the following order. i. Experimental series with dual phase crossflow with physical contact between fibers with loose configuration; ii. Experimental series with dual phase crossflow with tight fibers configuration; iii. Experimental series with dual phase crossflow; iv. Experimental series with single phase crossflow; v. Experimental series with static conditions. 3) True steady state conditions were never reached. For all experimental conditions, the permeate flux continued to slowly decrease over time after the initial rapid decrease. The magnitude of the permeate flux, following the initial period of rapid decrease, was defined as the pseudo steady state permeate flux. 4) For all experimental conditions, pseudo steady state conditions were reached at approximately 260 minutes after the start of the filtration experiment.  42  CHAPTER 4 RESULTS AND DISCUSSIONS  5) The magnitude of pseudo steady state permeates flux for the different experimental conditions increased in the following order. i. Experimental series with static conditions; ii. Experimental series with single phase crossflow; iii. Experimental series with dual phase crossflow; iv. Experimental series with dual phase crossflow with tight fibers configuration; v. Experimental series with dual phase crossflow with physical contact between fibers in loose. 6) Similar results were observed for the experiment conducted with tight multi-fiber modules and tight single-fiber modules. These results suggest that, for the physical contact between membrane fibers to occur, and have a beneficial impact on the permeate flux, the multi-fiber modules must be in a loose configuration 130  160  140  H  i  c xo 0  120 i rOQ  A O +  1-1-T-O 1-1-T-A 1-1-T-B  • 0 V • G t x • •  1-1-T-C 1-2-T-A 1-2-T-B 1-2-T-C 8-2-LA 8-2-L-B 8-2-L-C 8-1-T-B 8-2-T-B  60  E  <D SO O. ' "  40 N  0  °  0  0  0  0 O O o 0o O 0 0 ,  0  OOD •  •  •  •  •  O  D  D  D  D  D  D  D  Q  D  20 A A A A A A  0-  20c  4CC  i3CC  SCO  ixo  1200  1400  Tirre(rTin)  Figure 4.1: Membrane permeate flux versus filtration time  1600  CHAPTER 4 RESULTS AND DISCUSSIONS  43  4.2 Permeate Flux Decline with Volume Filtered As suggested by Hermia (1982), depending on the type of fouling, the extent of the decline in the permeate flux can be related to the amount of permeate filtered, rather than the filtration time. The experimental data is also presented in terms of the volume filtered. The results for all of the series of experiments are presented in Figure 4.2, in terms of permeate flux versus volume filtered. The following overall trends were observed: 1) The trend observed for the permeate flux versus volume filtered is relatively similar to that observed for the permeate flux versus filtration time. 2) For all experimental conditions, the permeate flux initially decreased rapidly. 3) The rate of initial rapid decrease in the permeate flux increased in the following order. i. Experimental series with dual phase crossflow with physical contact between fibers with loose configuration; ii. Experimental series with dual phase crossflow with tight fibers configuration; iii. Experimental series with dual phase crossflow; iv. Experimental series with single phase crossflow; v. Experimental series with static conditions. 4) True steady state conditions were never reached. For all experimental conditions, the permeate flux continued to slowly decrease over time after the initial rapid decrease. The magnitude of the permeate flux, following the initial period of rapid decrease, was defined as the pseudo steady state permeate flux. 5) For all experimental conditions, pseudo steady state conditions were reached approximately when 0.7L filtrate filtered. 6) The magnitudes of pseudo steady state permeate flux for the different experimental conditions increased in the following order.  44  CHAPTER 4 RESULTS AND DISCUSSIONS i. Experimental series with static conditions; ii. Experimental series with single phase crossflow; iii. Experimental series with dual phase crossflow; iv. Experimental series with dual phase crossflow with tight fibers;  v. Experimental series with dual phase crossflow with physical contact between fibers in loose. 7) Similar results were observed for the experiment conducted with tight multi-fiber modules and tight single-fiber modules. These results suggest that, for the physical contact between membrane fibers to occur, and have a beneficial impact on the permeate flux, the multifiber modules must be in a loose configuration.  °o  S3*  12C  Hft, x  100  -f  o  x  »#  x  x  x  x  X  •  «°  % a O  1-1-T-B  0  1-1-T-C  O  1-2-T-A  7  88°  80  1-1-T-O 1-1-T-A  <  0 „ fflx  3  » 0  1-2-T-B  B  1-2-T-C  O  &-24.-A  t  &-2-4.-B  X  8-2-4.-C  •  8-1-T-B  •  8-2-T-B  X  ^»J& , •  ^HO  g  •  •  V  7  SI  a a oo  CO »o  o  40  20  00  0.5  1.0  1.5  Volume filtered(L)  Figure 4.2: Permeate flux versus volume filtered  2.0  2.5  CHAPTER  45  4 RESULTS AND DISCUSSIONS  4.3 Modeling of Experimental Results  4.3.1 Model of experimental data A number of empirical equations were fitted to the collected data using linear regression (Sigma Plot 8.0 SPSS Science, Inc.). The best fit was observed for an experimental equation of the form, presented below. Jv = c *exp a(t  v)  + d* exp  b<t  (4.1)  v>  Where, Jv is the permeate flux [L/m hr]; c, a, d and b are empirical constants; t and V 2  corresponding to the filtration time and the filtration volume, respectively. Equation 4.1 was fitted to the data collected in the present study, with respect to both the cumulative filtration time and the cumulative filtration volume. As presented in Table 4.1, fitting Equation 4.1 with respect to the cumulative volume filtered, typically resulted in better results. Table 4.1: Multiple coefficient of determination of experimental results of non-linear regression analysis using Equation 4.1 Experiment series  Test name  1  1-1-T-A 1-1 -T-B 1-1-T-C 1-2-T-A 1-2-T-B 1-2-T-C 8-2-L-A 8-2-L-B 8-2-L-C 8-1-T-B 8-2-T-B 1-1-T-O  2  3  4 5  Value of multiple coefficient of determination (I ) for non-linear regression using 4.1 V (fit with Equation 4.1 Jv versus / analysis (fit with using Equation 4.1Equation Jv versus to experimental data) to experimental data) 0.9852 0.9850 0.9323 0.9333 0.9198 0.9097 0.9901 0.9948 0.9536 0.9587 0.9486 0.9495 0.9817 0.9953 0.9033 0.9191 0.9722 0.9876 0.9796 0.9899 0.9765 0.9766 0.9793 0.9953  CHAPTER 4 RESULTS AND DISCUSSIONS  46  As a result, the decline in the permeate flux was modeled with respect to the cumulative filtration volume. It should be noted that all of the experimental measurements are within a 99% confidence interval associated with the fit of Equation 4.2 to the collected data.  The first term on the right hand side of Equation 4.1 corresponds to the initial rapid permeate flux decline phase. Field et al. (1995) and Hermia (1992) also reported that, depending on the fouling mechanism, the decline in the permeate flux can follow an exponential relationship. Field et al. (1995) indicated that the magnitude of the decline in the permeate flux can be proportion to the difference between the initial permeate flux (Jo) and the steady state (or critical) permeate flux (Jss). However, the relationships developed by Field et al. (1995) assume that steady state conditions can be reached. Once a steady state condition was reached, the permeate flux remained constant. Based on this relationship, the second term on the right hand side of Equation 4.1 is constant at the steady state permeate flux. However, true steady state conditions were never reached in the present study. For the present study, the magnitude of the decline in the permeate flux that occurs during the initial phase (i.e. c in Equation 4.1) was defined as the difference between the initial permeate flux and the pseudo-steady-state permeate flux (J'ss). The rate of the decline in the permeate flux that occurs during the initial phase was defined by an exponential coefficient (a). A similar exponential coefficient was introduced by Field et al. (1995).  The most significant modification to the fouling models introduced by Field et al. (1995) is the introduction of a term that describes the slow permeate flux decline that occurs over time. As with the initial rapid permeate flux decline, the subsequent long term slow permeate flux  47  CHAPTER 4 RESULTS AND DISCUSSIONS  decline could be modeled using an exponential relationship, (i.e. second term on the right hand side of Equation 4.1). The magnitude of the decline that can occur in this phase was defined as the pseudo-steady-state permeate flux. The rate of the decline in the permeate flux, during the long term permeate flux decline, was defined in terms of the pseudo steady state permeate flux reduction coefficient (b). The resulting modifications to Equation 4.1 are presented in Equation 4.2. (4.2)  Jv = (Jo-J'ss) *exp + J'ss* exp aV  bv  2  2  Where, Jv is the permeate flux [L/m hr]; Jo is the initial permeate flux [L/m hr]; J'ss is the pseudo-steady-state permeate flux [L/m hr]; a is the initial permeate flux decline coefficient 2  [L ]; and b is the pseudo-steady-state permeate flux decline coefficient [L" ]; Vcorresponding 1  1  to the filtration volume [L]. The absolute values associated with the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient and the pseudo-steady-state permeate flux decline coefficient were used to measure and assess the magnitude of the impact of hydrodynamic conditions and membrane configuration on the permeate flux. All comparison related to the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient and the pseudo-steady-state permeate flux decline coefficient are based on a 90% confidence, in terms of the measurements made.  4.3.2 Estimate of the initial permeate flux The initial permeate flux is the permeate flux through a membrane in the absence of any fouling. The initial permeate flux was measured using a clean membrane when filtering clean  48  CHAPTER 4 RESULTS AND DISCUSSIONS  water (distilled water). The values obtained for the initial permeate flux are shown in Figure 4.3. A l l permeate' flux data collected during the present study was normalized to a transmembrane pressure of 4.1 psi (27.7kPa) and liquid temperature of 25°C. The initial permeate flux was determined to be 240.5L/m hr. 2  3 0 0 -,  250  . .  .  .  .  .  .  .  •  '  .  .  .  200 -  150 -  100 -  50 -  0  -I  0  1  1  100  200  1  1  300  400  50  T i m e(m in)  Figure 4.3: Initial permeate flux 4.3.3 Estimate of the pseudo-steady-state permeate flux, the initial permeate flux decline and the pseudo-steady-state permeate flux decline coefficient Non-linear regression (Sigmaplot 8.0, SPSS Science, Inc.) was used to fit Equation 4.2 to the experimental data presented in Figure 4.2..The resulting estimates for the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient and the pseudo-steady-state permeate flux decline coefficient are summarized in Table 4.2 (also see Figures 4.4, 4.5, 4.6 and 4.7 in Section 4.4).  49  CHAPTER 4 RESULTS AND DISCUSSIONS  Table 4.2: Estimate of the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient and the pseudo-steady-state permeate flux decline coefficient Experiment series  Test name  Pseudosteadystate permeate flux  Initial permeate flux decline coefficient  Pseudosteady-state permeate flux decline coefficient  1  1-1-T-A  42.9±1.68  8.39±0.27  0.673±0.045  Multiple coefficient of determination for experimental results of non-linear regression analysis using Equation 4.2 0.9850  1-1-T-B  38.0±4.47  8.28±0.97  0.463±0.120  0.9333  1-1-T-C  44.2±1.29  7.77±0.32  0.359±0.025  0.9097  1-2-T-A  65.0±1.30  5.36±0.17  0.235±0.014  0.9948  1-2-T-B  91.8±4.24  7.78±1.12  0.309±0.049  0.9587  1-2-T-C  93.7±1.00  5.36±1.31  0.268±0.096  0.9495  8-2-L-A  98.1±2.80  6.68±0.63  0.288±0.026  0.9953  8-2-L-B  103±4.55  3.30±0.66  0.070±0.021  0.9191  8-2-L-C  104±5.62  6.52±1.19  0.110±0.051  0.9876  8-1-T-B  37.3±2.03  8.25±0.40  0.539±0.060  0.9899  8-2-T-B  78.5±10.4  4.13±0.68  0.077±0.108  0.9766  1-1-T-O  25.7±1.29  6.93±0.16  0.839±0.043  0.9953  2  3  4  5  4.4 Effect of Hydrodynamic Conditions and Membrane Configurations on Pseudo Steady State Permeate Flux, Initial Permeate Flux Decline Coefficient, and Pseudo Steady State Permeate Flux Decline Coefficient  4.4.1 Impact of single phase crossflow on the permeate flux The impact of single phase (i.e. water only) crossflow on the permeate flux was investigated by monitoring the permeate flux at different bulk crossflow velocities, under single phase flow with a single fiber membrane module. Overall, as presented in Figure 4.4, an increase in the bulk crossflow velocity impacted the permeate flux.  CHAPTER 4 RESULTS AND DISCUSSIONS  50  Bulk crossflow velocity « OrrVs O 0.2rrVs * 0.3nYs o 0.4rrVs  '  (D%  cftr °  O  D  °  D  ° ° ° ° o  D  D  a  o  a  D D  D  D  a  a  o  o  °ooOo  D  ooao  D  D  Volume filtered(L)  Figure 4.4: Impact of single phase bulk crossflow velocity on the permeate flux  The pseudo-steady-state permeate flux, the initial permeate flux decline coefficient, and the pseudo-steady-state permeate flux decline coefficient for the experiments conducted, using single phase crossflow, are presented in Figure 4.5, 4.6, and 4.7, respectively.  The bulk crossflow velocity had a significant impact on the pseudo-steady-state permeate flux as presented in Figure 4.5. The pseudo-steady-state permeate flux was 67% higher at a bulk crossflow velocity of 0.2m/s compared to that at Om/s. However, there was no significant difference in pseudo-steady-state permeate flux for bulk crossflow velocities of 0.2, 0.3 and 0.4 m/s.  CHAPTER  4 RESULTS AND  DISCUSSIONS  51  E  CD Q.  0.2  0.3  0.4  0.5  Bulk crossflow velocity(m/s)  Figure 4.5:  Pseudo-steady-state permeate flux versus bulk crossflow velocity (Single phase crossflow)  (Error bars represent the standard error of the estimated parameter)  The bulk crossflow velocity also had a significant impact on the initial permeate flux decline coefficient as Figure 4.6. Again, there was no significant difference in the initial permeate flux decline coefficient for bulk crossflow velocities of 0.2, 0.3 and 0.4 m/s.  0.2  0.3  Bulk crossflow velocity(m/s)  Figure 4.6:  Initial permeate flux decline coefficient versus bulk crossflow velocity (Single phase crossflow)  (Error bars represent the standard error of the estimated parameter)  CHAPTER 4 RESULTS AND DISCUSSIONS  52  As shown in Figure 4.7, the bulk crossflow velocity had a significant impact on the pseudosteady-state permeate flux decline coefficient. As the bulk crossflow velocity increased, the pseudo-steady-state permeate flux decline coefficient significantly decreased. As a result, the permeate flux was enhanced.  0.0  0.1  0.2  0.3  0.4  0.5  Bulk crossflow velocity(m/s)  Figure 4.7: Pseudo-steady-state permeate flux decline coefficient versus bulk crossflow velocity (Single phase crossflow) (Error bars represent the standard error of the estimated parameter) 4.4.2 Impact of dual phase crossflow with airsparging on the permeate flux The impact of dual phase (i.e. water and airsparging) crossflow on the permeate flux was investigated by monitoring the permeate flux at different bulk crossflow velocities, under dual phase flow, with a single-fiber membrane module. Overall, as presented in Figure 4.8, an increase in the bulk crossflow velocity impacted the permeate flux. The pseudo-steady-state permeate flux, the initial permeate flux decline coefficient, and the pseudo-steady-state permeate flux decline coefficient, for the experiments conducted using dual phase crossflow, are presented in Figure 4.9, 4.10, and 4.11, respectively.  CHAPTER  4  RESULTS  AND  53  DISCUSSIONS  Bulk crossflow velocity » • a  si  1-1-T-0m/s 1-2-T-O 2nv's 1-2-T-0.3m/s 1-2-T-O 4m/s  %  "  V  % V . • »•-  B  • •»• • , 1  10  5 a (  ,  1.5  Volume filtered(L)  Figure 4.8: Impact of dual phase bulk crossflow velocity on the permeate flux  The bulk crossflow velocity had a significant impact on the pseudo-steady-state permeate flux as presented in Figure 4.9. A s the crossflow velocity increased, the pseudo-steady-state permeate flux increased. The pseudo-steady-state permeate flux was 153% higher at a bulk crossflow velocity of 0.2m/s, compared to that at Om/s. A t 0.3m/s, the pseudo-steady-state permeate flux was 257% higher compared to that at Om/s. However, there was no significant difference in the pseudo-steady-state permeate flux for bulk crossflow velocities of 0.3 and 0.4 m/s.  CHAPTER 4 RESULTS AND DISCUSSIONS  54  120  a.  o4 0.0  , 0.1  Bulk  ,  ,  ,  0.2  0.3  0.4  .  1 0.5  crossflow velocity(m/s)  Figure 4.9: Pseudo-steady-state permeate flux versus bulk crossflow velocity (Dual phase crossflow) (Error bars represent the standard error of the estimated parameter) As presented in Figure 4.10, there is no consistent trend in terms of the magnitude of the initial permeate flux decline coefficient versus the bulk crossflow velocity. The initial permeate flux decline coefficient decreased when the bulk crossflow velocity was increased, from Om/s (i.e. static) to 0.2m/s. However, the initial permeate flux decline coefficient increased when the bulk crossflow velocity was increased from 0.2m/s to 0.3m/s. The magnitude of the initial permeate flux decline coefficient at a bulk crossflow velocity of 0.3m/s was not significantly different from that at Om/s. When the bulk crossflow velocity was further increased to 0.4m/s, the initial permeate flux decline coefficient once again decreased. At bulk crossflow velocity of 0.4m/s, the initial permeate flux decline coefficient was not significantly different from that observed for bulk crossflow velocities of 0.2m/s.  CHAPTER 4 RESULTS AND DISCUSSIONS  55  10 .-  u  o .  «  2 -  0 J  0.0  1  0.1  1  0.2  ,  1  0.3  0.4  0.5  Bulk crossflow velocity(m/s)  Figure 4.10: Initial permeate flux decline coefficient versus bulk crossflow velocity (Dual phase crossflow) (Error bars represent the standard error of the estimated parameter) As shown in Figure 4.11, the pseudo-steady-state permeate flux decline coefficient (b) for dual phase crossflow was significantly impacted by the bulk crossflow velocity. The pseudosteady-state permeate flux decline coefficient (b) was significantly lower for bulk crossflow velocities of 0.2m/s when compared to that observed for static conditions (i.e. Om/s). The pseudo-steady-state permeate flux decline coefficient was 72% less at a bulk crossflow velocity of 0.2m/s, compared to that at Om/s. However, there was no significant difference in pseudo-steady-state permeate flux decline coefficient for bulk crossflow velocities of 0.2, 0.3 and 0.4m/s.  56  CHAPTER 4 RESULTS AND DISCUSSIONS  0.2  0.3  0.5  Bulk crossflow velocity [m/s]  Figure 4.11: The pseudo-steady-state permeate flux decline coefficient (b) versus bulk crossflow velocity (Dual phase crossflow) (Error bars represent the standard error of the estimated parameter)  4.4.3 Impact ofphysical contact between membrane fibers on the permeate flux The impact of the physical contact between membrane fibers on the permeate flux was investigated by monitoring the permeate flux for loose multi-fiber modules, at different bulk crossflow velocities, under dual phase flow. Overall, as presented in Figure 4.12, an increase in the bulk crossflow velocity impacted the permeate flux. The pseudo-steady-state permeate flux, the initial permeate flux decline coefficient, and the pseudo-steady-state permeate flux decline coefficient for the experiments conducted to investigate the impact of physical contact, are presented in Figure 4.13, 4.14, and 4.15, respectively.  CHAPTER 4 RESULTS AND DISCUSSIONS  57  180  Bulk crossficw velocity 160 X  X  ;-io  0  X  0  I  X  °0  8-2-L-0.3rrVs 8-2-L-0.4rrVs  x  0  X  T  x  *  x  x  X  y  *  x  e *  ' x  X  jt • 100  1-1-T-Orrfe 8-2-L0.2rrYs  x  0 120  ' o  » X  x  X  x  , '  X  X  0  X  00  ea  ' ' ' '  T  v  X  x  *  *  * 'x  X  _ X  x  '  '  '  ,  '  Oo  T  0 ^  o  o 0  o  Q  o  <x>-  0  «• "as  20  * *%  O'I  10  1.5  2.0  Volume filtered(L)  Figure 4.12: Impact of physical contact on the permeate flux  flie crossflow had a significant impact on the pseudo-steady-state permeate flux, as presented in Figure 4.13. The pseudo-steady-state permeate flux was 282% higher at a bulk crossflow velocity of 0.2m/s compared to that at Om/s. However, there was no significant difference in the pseudo-steady-state permeate flux) for bulk crossflow velocities of 0.2, 0.3 and 0.4 m/s.  CHAPTER  4  RESULTS  0.2  AND  58  DISCUSSIONS  0.3  0.5  Bulk crossflow velocity(m/s)  Figure 4.13: Pseudo-steady-state permeate flux versus bulk crossflow velocity (Dual phase crossflow with physical contact) (Error bars represent the standard error of the estimated parameter) As presented in Figure 4.14, there is no consistent trend in the terms of the magnitude of the initial permeate flux decline coefficient versus bulk crossflow velocity. At the bulk crossflow velocity of 0.2m/s, the magnitude of the initial permeate flux decline coefficient was not significantly different from that at Om/s. However, the initial permeate flux decline coefficient decreased significantly, when the bulk crossflow velocity was increased from 0.2m/s to 0.3m/s. When the bulk crossflow velocity was further increased to 0.4m/s, the initial permeate flux decline coefficient increased significantly, once again. However, at a bulk crossflow velocity of 0.4m/s, the initial permeate flux decline coefficient was not significantly different from that at 0 m/s or 0.2m/s.  CHAPTER  4  RESULTS  AND  59  DISCUSSIONS  '_5 0 \ 0.0  ,  ,  ,  ,  0.1  0.2  0.3  0.4  0.5  Bulk crossflow velocity(m/s)  Figure 4.14: Initial permeate flux decline coefficient versus bulk crossflow velocity (Dual phase crossflow with physical contact) (Error bars represent the standard error of the estimated parameter) As shown in Figure 4.15, the pseudo-steady-state permeate flux decline coefficient (b) for a multi-fiber module, in dual phase crossflow, was significantly impacted by the bulk crossflow velocity. As the bulk crossflow velocity increased, the pseudo-steady-state permeate flux decline coefficient (b) decreased significantly. The pseudo-steady-state permeate flux decline coefficient (b) was 65.7% lower at a bulk crossflow velocity of 0.2m/s, compared to that at Om/s. At 0.3m/s, the pseudo-steady-state permeate flux decline coefficient (b) was 91.7% lower, compared to that at Om/s. However, there was no significant difference in the pseudosteady-state permeate flux decline coefficient (b) for crossflows of 0.3 and 0.4 m/s.  CHAPTER  4  RESULTS  AND  60  DISCUSSIONS  JT- 1.0  a-  0.0 0.2  0.3  0.5  Bulk crossflow velocity(m/s)  Figure 4.15: The pseudo-steady-state permeate flux decline coefficient (b) versus bulk crossflow velocity (Dual phase crossflow with physical contact) (Error bars represent the standard error of the estimated parameter) 4.4.4 Impact offiber tension in membrane modules on the permeate flux The impact of fiber tension on the permeate flux was investigated by monitoring the permeate flux for loose and tight single and multi-fiber modules, under single and dual phase crossflow, at a bulk crossflow velocity of 0.3m/s. Overall, as presented in Figure 4.16, the tightness of the fiber bundles had an impact on the permeate flux. The pseudo-steady-state permeate flux, the initial permeate flux decline coefficient, and the pseudo-steady-state permeate flux decline coefficient for the experiments conducted are presented in Figures 4.17, 4.18, and 4.19, respectively.  CHAPTER  4  RESULTS  AND  DISCUSSIONS  61  Single-tight membrane module for single phase v  Single-tight membrane module for dual phase  T  Multi-loose membrane module for dual phase  •  Multi-tight membrane module for single phase  #  Multi-tight membrane module for dual phase  £ 100 3  Volume filteredfLI  Figure 4.16: Impact of fiber tension in membrane modules on permeate flux (at bulk crossflow velocity of 0.3m/s) 4.4.4.1 Comparison between loose and tight multi-fiber membrane modules for dual phase crossflow The results presented in Figure 4.16 indicate that, for dual phase crossflow, the permeate flux under a loose configuration, is significantly larger than that under tight configuration. As shown in Figure 4.17, the fiber tension had a significant impact on the pseudo-steady-state permeate flux. The pseudo-steady-state permeate flux was 31.2% higher for loose multi-fiber membrane modules configuration, compared to that for tight, multi-fiber membrane modules configuration, at a bulk crossflow velocity of 0.3m/s with dual phase crossflow. As presented  C H A P T E R  in  4  R E S U L T S  A N D  D I S C U S S I O N S  62  Figure 4.18, there is no significant difference on the initial permeate flux decline  coefficient between loose multi-fiber membrane modules configuration and tight multi-fiber membrane modules configuration, at a bulk crossflow velocity of 0.3m/s with dual phase crossflow. As shown in Figure 4.19, the fiber tension also had no significant impact on the pseudo-steady-state permeate flux decline coefficient. Therefore, the beneficial impact of the physical contact between the membrane fibers can only be achieved when the membrane modules are in a loose configuration.  120  E 13  100  3 to  80  I  CD Ei OJ Q.  60  •o  40  CO <D  o  T3  20  A T  CD oo CL  Single phase crossflow  tight multi-fiber loose multi-fiber  Dual phase crossflow  Hydrodynamic conditions  Figure 4.17: Impact of fiber tension on pseudo-steady-state permeate flux (Error bars represent the standard error of the estimated parameter)  CHAPTER  4  RESULTS  AND  63  DISCUSSIONS  10  tight multi-fiber loose multi-fiber  •a  Single phase crossflow  Dual phase crossflow  Hydrodynamic condition  Figure 4.18: Impact of fiber tension on initial permeate flux decline coefficient (Error bars represent the standard error of the estimated parameter) 1.0  0.8  0.6  A  tight m u l t i - f i b e r  •  loose m u l t i - f i b e r  I  0.4  0.2 -\  i 0.0  Single phase crossflow  Dual phase crossflow  Hydrodynamic condition  Figure 4.19: Impact of fiber tension on pseudo-steady-state permeate flux decline coefficient (Error bars represent the standard error of the estimated parameter)  CHAPTER 4 RESULTS AND DISCUSSIONS  64  4.4.4.2 Comparison between single and multi-fiber membrane modules with a tight configuration under dual phase crossflow The results presented in Figure 4.16 indicate that for dual phase crossflow, the permeate flux is similar for both single and multi-fiber membrane modules, when the modules have a tight configuration. Also, the permeate flux for both single and multi-fiber membrane modules, with a tight configuration, are significantly lower than that for multi-fiber membrane modules with a loose configuration.  As presented in Table 4.3, there is no significant difference on the pseudo-steady-state permeate flux between single and multi-fiber module, in a tight configuration at 0.3m/s with dual phase crossflow. However, the initial permeate flux decline coefficient for tight multifiber module is less than that for tight single-fiber membrane modules at 0.3m/s (see Table 4.2). However, as discussed in Section 4.4.2, there was no consistent trend in terms of the magnitude of the initial permeate flux decline coefficient versus the bulk crossflow velocity (see Figure 4.10). As a result, the initial permeate flux decline coefficient for a tight multifiber module is within the spread of the observations for the initial permeate flux decline coefficient for tight single fiber module (see Figure 4.10). Therefore, it can be concluded that there is no substantial difference between the initial permeate flux decline coefficient for single and multi-fiber membrane modules with a tight configuration under dual phase crossflow. The pseudo-steady-state permeate flux decline coefficient for tight multi-fiber module is also slightly less than that for tight single fiber module at 0.3m/s (see Table 4.3). However, as discussed in Section 4.4.2, there was no significant difference between pseudo-  CHAPTER 4 RESULTS AND DISCUSSIONS  65  steady-state permeate flux decline coefficient for tight single-fiber modules at bulk crossflow velocities of 0.2, 0.3 and 0.4m/s (see Figure 4.11). The pseudo-steady-state permeate flux decline coefficient for tight multi-fiber modules at 0.3m/s is not significantly different than that for tight single-fiber modules at 0.4m/s. Therefore, it can be concluded that there is no substantial difference between the pseudo-steady-state permeate flux decline coefficient for single and multi-fiber membrane modules, with a tight configuration under dual phase flow.  These results suggest that significant physical contact between fibers likely only occurs when the membrane modules are in a loose configuration. This result is consistent with those presented in Section 4.4.4.1, which indicate that the beneficial effects of the physical contact between membrane fibers on the permeate flux only occurs when the membrane modules are in a loose configuration.  4.4.4.3 Comparison between single and multi-fiber membrane modules with a tight configuration under single phase crossflow The results presented in Figure 4.16 indicate that, for single phase crossflow, the permeate flux relatively similar for both single and multi-fiber membrane modules when the modules have a tight configuration. This result is consistent with the conclusions for dual phase crossflow.  As presented in Table 4.3, there is no significant difference on the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient, and the pseudo-steady-state permeate flux decline coefficient, for single and multi-fiber membrane modules in a tight  CHAPTER 4 RESULTS AND DISCUSSIONS  66  configuration, at a bulk crossflow velocity of 0.3m/s, with single phase crossflow. These results are similar to those observed between single and multi-fiber membrane modules, with a tight configuration, under dual phase crossflow (see Section 4.4.4.2).  4.5 Discussion  The results for the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient and the pseudo-steady-state permeate flux decline coefficient, presented in Section 4.4, are summarized in Figure 4.20, 4.21 and 4.22, respectively. 120  Bulk crossflow velocity(m/s)  Figure 4.20: Impact of hydrodynamic conditions and membrane configuration on pseudosteady-state permeate flux (Error bars represent the standard error of the estimated parameter)  CHAPTER 4 RESULTS AND  67  DISCUSSIONS  10  Single phase crossflow Dual phase crossflow Dual phase crossflow with physical contact 0.0  0.1  0.2  0.3  0.4  0.5  Bulk crossflow velocity(m/s)  Figure 4.21: Impact of hydrodynamic conditions and membrane configuration on initial permeate flux decline coefficient (Error bars represent the standard error of the estimated parameter)  CD O  Single phase crossflow Dual phase crossflow Dual phase crossflow with physical contact  0 O  o  <D  c  "o CD  0)  ro <D  E I  Q.  £  ro  TJ  ro o  TJ 0) (/> Q.  0.2  0.3  0.4  0.5  Bulk crossflow velocity(m/s)  Figure 4.22: Impact of hydrodynamic conditions and membrane configuration on the pseudo-steady-state permeate flux decline coefficient (Error bars represent the standard error of the estimated parameter)  CHAPTER 4 RESULTS AND DISCUSSIONS  68  Based on the results presented in Figure 4.17, 4.18 and 4.19, the following observations can be made.  First, single phase crossflow had a significant impact on pseudo-steady-state permeate flux (see Figure 4.20), when the bulk crossflow velocity was increased from Om/s to 0.2m/s. The pseudo-steady-state permeate flux was approximately 67% higher at a bulk crossflow velocity of 0.2m/s compared to that at Om/s. However, there was no significant difference in pseudosteady-state permeate flux for bulk crossflow velocities of 0.2, 0.3 and 0.4 m/s. Air sparging and physical contact enabled a high pseudo-steady-state permeate flux to be maintained. The maximum pseudo-steady state permeate flux that could be reached with air sparging (i.e. dual phase crossflow) was approximately 135% higher than that which could be achieved without air sparging (i.e. sigle phase crossflow). The maximum pseudo-steady-state permeate flux that could be reached when enabling physical contact was approximately 15% higher than that which could be achieved when physical contact did not occur. Therefore, to maximize the pseudo-steady-state permeate flux, the membrane should be operated under dual phase crossflow and the physical contact between membrane fibres should be promoted.  In addition, the results indicate that there is no significant benefit of providing a bulk crossflow velocity in excess of 0.2 m/s, in terms of maintaining a high pseudo-steady-state permeate flux for dual phase crossflow, when the physical contact between membrane fibers is promoted. The membrane modules must be in a loose configuration to promote the physical contact between fibers.  CHAPTER 4 RESULTS AND DISCUSSIONS  69  Second, for the experimental conditions used in the present study, the hydrodynamic conditions and the membrane configuration did not have a consistently significant impact on the initial permeate flux decline coefficient. However, further investigations are needed to better understand the reason behind the high variability in the measured initial permeate flux decline coefficient.  Third, as presented in Figure 4.22, single phase crossflow had a significant impact on the pseudo-steady-state permeate flux decline coefficient. As the bulk crossflow velocity increased from Om/s to 0.4m/s, the pseudo-steady-state permeate flux decline coefficient decreased approximately 57%. However, the decrease of the pseudo-steady-state permeate flux decline coefficient was only approximately 20% when the bulk crossflow velocity was increased from Om/s to 0.2m/s. Air sparging and physical contact enabled a lower pseudosteady-state permeate flux decline coefficient to be maintained. The minimum pseudo-steadystate permeate flux decline coefficient that could be reached with air sparging (i.e. dual phase crossflow) was approximately 70% less than that which could be achieved under statistic condition (at the bulk crossflow velocity of Om/s). The minimum pseudo-steady-state permeate flux decline coefficient that could be reached, when enabling physical contact, was approximately 92% less than that which could be achieved under statistic condition. When physical contact did not occur, the minimum pseudo-steady-state permeate flux decline coefficient that could be reached was approximately 77% less than that which could be achieved under static conditions. Therefore, to minimize the pseudo-steady-state permeate flux decline coefficient, the membrane should be operated under dual phase crossflow and that the physical contact between membrane fibres should be promoted. In addition, the  CHAPTER 4 RESULTS AND DISCUSSIONS  70  results indicate that there is no significant benefit in providing a bulk crossflow velocity in excess of 0.3m/s, in terms of maintaining a low pseudo-steady-state permeate flux decline coefficient for dual phase crossflow, when the physical contact between membrane fibers is promoted.  Further numerical analysis revealed that the pseudo-steady-state permeate flux decline coefficient may be related to the pseudo-steady-state permeate flux. As presented in Figure 4.23, the pseudo-steady-state permeate flux decline coefficient was proportional to the inverse of pseudo-steady-state permeate flux (\/J'ss) for all of the experimental conditions investigated in this present study. The inverse of pseudo-steady-state permeate flux is proportional to the time that the permeating liquid spends in a membrane pore. It is not clear why it appears to be a relationship between the pseudo-steady-state permeate flux decline coefficient and the contact time. One could hypothesize that at longer contact times, a larger amount of foulants can adsorb to the surface of a membrane pore, resulting in a higher rate of permeate flux decline. Further research is required to confirm this hypothesis.  Equation 4.2 was modified to include the linear relationship between pseudo-steady-state permeate flux decline coefficient and pseudo-steady-state permeate flux as presented in Equation 4.3. Jv=(J -J' )exp (  0  ss  aV)  + J' exp  bV(I/rss)  ss  (4.3)  Where, b' is a modified pseudo-steady-state permeate flux decline coefficient and is equal to 21.44m" hr"' for the experimental conditions considered in the present study and the standard 2  error of the estimate is 2.141.  CHAPTER 4 RESULTS AND DISCUSSIONS  71  If the modified pseudo-steady-state permeate flux decline coefficient is related to the rate of adsorption of foulants on to the surface of a membrane pore, then it would be expected to be impacted by the characteristics of the raw water being filtered. Similarly, the modified pseudo-steady-state permeate flux decline coefficient would also be expected to be impacted by pre-treatment processes (e.g. pre coagulation) that would impact the raw water characteristics. Further studies are required to identify the raw water characteristics and pretreatment processes that can potentially impact the modified pseudo-steady-state permeate flux decline coefficient. • 1.0  =i  I  -a  0.8 A  0.6-  x  I  0.4  J  t5  0.0 0.00  0.01  0.02  0.03  0.04  0.05  l/J'ss [L m hr]  Figure 4.23: Impact of pseudo-steady-state permeate flux decline coefficient on the inverse of pseudo-steady-state permeate flux  72  CHAPTER 5 CONCLUSIONS  C H A P T E R 5  C O N C L U S I O N S  1. The decline of the permeate flux over filtration time could be characterized by an initial short period of fast permeate flux decline, followed by a longer period of slower permeate flux decline according to the following equation.  Jv = (Jo-J 'ss) *exp~ + J 'ss * exp' av  hv  Where, Jv is the permeate flux [L/m hr]; Jo is the initial permeate flux [L/m hr]; J'ss is 2  2  the pseudo-steady-state permeate flux [L/m hr]; a is the initial permeate flux decline 2  coefficient [L" ]; and b is the pseudo-steady-state permeate flux decline coefficient [L ]; V 1  1  corresponding to the filtration volume [L]. 2. In general, the hydrodynamic conditions and the system configuration had a significant impact on the permeate flux, the pseudo-steady-state permeate flux, the initial permeate flux decline coefficient and the pseudo-steady-state permeate flux decline coefficient. 3. Single phase crossflow (i.e. water only) had a significant impact on pseudo-steady-state permeate flux. The pseudo-steady-state permeate flux was approximately 67% higher at a bulk crossflow velocity of 0.2m/s compared to that at Om/s. However, there was no significant benefit fo providing a crossflow velocity in excess of 0.2m/s, in terms of maintaining a high pseudo-steady-state permeate flux under single phase crossflow. 4. Air sparging and physical contact enabled a high pseudo-steady state permeate flux to be maintained. The maximum pseudo-steady state permeate flux was achieved when the membrane system was operated under dual phase crossflow with physical contact between the fibres in the membrane module. The maximum pseudo-steady state permeate flux that  CHAPTER 5 CONCLUSIONS  73  could be reached with air sparging (i.e. dual phase crossflow) was approximately 135% higher than that which could be achieved without air sparging (i.e. sigle phase crossflow). The maximum pseudo-steady-state permeate flux that could be reached when enabling physical contact was approximately 15% higher than that which could be achieved when physical contact did not occur. 5. The pseudo-steady-state permeate flux increased with the extent of air sparging. However, there is no significant benefit in providing a bulk crossflow velocity in excess of 0.2 m/s, in terms of maintaining a high pseudo-steady-state permeate flux for dual phase crossflow, when the physical contact between membrane fibers is promoted. 6. For the experimental conditions used in the present study, the hydrodynamic conditions and the membrane configuration did not have a consistently significant impact on the initial permeate flux decline coefficient. However, further investigations are needed to better understand the reason behind the high variability in the measured initial permeate flux decline coefficient. 7. Single phase crossflow had a significant impact on the pseudo-steady-state permeate flux decline coefficient. As the bulk crossflow velocity increased from Om/s to 0.4m/s, the pseudo-steady-state permeate flux decline coefficient decreased approximately 57%. 8. Air sparging and physical contact enabled a lower pseudo-steady-state permeate flux decline coefficient to be maintained. The minimum pseudo-steady-state permeate flux decline coefficient was achieved when the membrane system was operated under dual phase crossflow with the physical contact between the membrane fibres in the membrane module. The minimum pseudo-steady-state permeate flux decline coefficient that could be reached with air sparging (i.e. dual phase crossflow) was approximately 70% less than that  CHAPTER 5 CONCLUSIONS  74  which could be achieved under statistic condition. The minimum pseudo-steady-state permeate flux decline coefficient that could be reached when enabling physical contact was approximately 92% less than that which could be achieved under statistic condition (at a bulk crossflow velocity of Om/s). 9. There is no significant benefit in providing a bulk crossflow velocity in excess of 0.3m/s, in terms of maintaining a low pseudo-steady-state permeate flux decline coefficient for dual phase crossflow when the physical contact between membrane fibers is promoted. 10. Single phase bulk crossflow did not significantly contribute to maintaining a high permeate flux. On the other hand, the interactions between sparged air bubbles and the membrane fibers significantly contributed to maintaining a high permeate flux. The physical contacts between the membrane fibers also significantly contribute to maintaining a high permeate flux. Significant physical contact only occured when fibers in the membrane module were in a loose configuration. 11. The pseudo-steady-state permeate flux decline coefficient was proportional to the inverse of pseudo-steady-state permeate flux for all of the experimental conditions investigated in this present study. The decline of the permeate flux over time was modified to the following equation. Jv=(J -J' )exp - > + (  0  ss  aV  J' exp '  b v(,/Jss)  ss  Where, b' is a modified pseudo-steady-state permeate flux decline coefficient and is equal to 21.44m" hr"' for the experimental conditions considered in the present study. 2  REFERENCES  75  REFERENCES 1. Aimar P. and Aptel P., Membrane Processes, Proceedings of Euromembranes 92, Lavoisier, Paris, 1992. 2. Aimar P. and Howell J. A., Effects of Concentration Boundary Layer Development on the Flux Limitations in Ultrafdtration, Chem. Eng. Res. Des., 67: 255, 1989. 3. APHA, A WW A and WEF, Standard Methods for the Examination of Water and Wastewater, 18 ed., APHA, A W W A & WEF, 1992. th  4. Buckley C.A., Membrane Technology for the Treatment of Dyehouse Effluents, Wat. Sci. Tech. 25(10): 203, 1992. 5. Bixler H.J. and Rappe G.C., Increasing the Mass-Transport Rate Across Ultrafdtration Membranes, US Patent 3 541 006, 1970. 6. Baker R.J., Fane A.G., Fell C.J.D. and Yoo B. H., Factors Affecting Flux in Crossflow Filtration, Desalination, 53: 81, 1985. 7. Bellara S.R., Cui Z.F. and Pepper D.S., Gas Sparging to Enhance Permeate Flux in Ultrafdtration Using Hollow Fibre Membranes, J. Membr. Sci. 121: 175, 1996. 8. Costa A.R. and Fane A.G., Effect of Configuration on Fluid Flow Path and Untrafdtration Flux, Ind. Eng. Chem. Res. 33 (7): 1845, 1994. 9. Cabassud C , Laborie S., Durand-Bourlier L. and Laine J.M., Air Sparging in Ultrafiltration  Hollow  Fibers:  Relationship  between Flux  Enhancement, Cake  Characteristics and Hydrodynamic Parameters, J. Membr. Sci. 181 (1): 57, 2001.  REFERENCES  76  10. Chiemchaisri C , Wong Y.K., Urase T. and Yamamoto K., Organic Stabilization and Nitrogen Removal in Membrane Separation Bioreactor for Domestic Wastewater Treatment, Wat. Sci. Tech. 25: 231, 1992. 11. Crespo J. G. and Boddeker K. W., Membrane Processes in Separation and Purification, NATOASI Series E : Applied Sciences, vol. 272, Kluwer Academic Publishers, Dordrecht, 1994. 12. Cheryan M., Ultrafiltration Handbook, Technomic Publishing Co., USA, 1986. 13. Cartwright P.S., Industrial Wastewater Treatment with Membranes-A United States Perspective, Wat. Sci. Tech. 25(10): 373, 1992. 14. Chang S. and Fane A.G., Characteristics of Microfiltration of Suspensions with inter-fiber two-phaseflow,Journal of Chemical Technology and Biotechnology 75: 533, 2000. 15. Chang S. and Fane A.G., Filtration of Biomass with Axial Inter-fibre Upward Slug-flow: Performance and Mechanism, J. Membr. Sci. 180 (1): 57, 2000. 16. Chang S. and Fane A.G., the Effect of Fibre Diameter on Filtration and Flux DistributionRelevance to submerged Hollow Fibre Modules, J. Membr. Sci. 184 (2): 221, 2001. 17. Chang S. and Fane A.G., Filtration of Biomass with Laboratory-Scale Submerged Hollow Fibre Modules -Effect of Operating Conditions and Module Configuration, Journal of Chemical Technology and Biotechnology 77:1030, 2002. 18. Cheng T.W, Yeh H.M. and Gau C.T., Enhancement of Permeate Flux by Gas Slugs for Crossflow Ultrafiltration in Tubular Membrane Module, Sep. Sci. Technol. 33: 2295, 1998. 19. Cheng T.W., Yeh H.M. and Wu J.H., Effects of Gas Slugs and Inclination Angle on the Ultrafiltration Flux in Tubular membrane Module, J. Membr. Sci. 158 (1-2): 223, 1999.  REFERENCES  11  20. Cheng T.W., Influent of Inclunation on Gas-sparged Cross-flow Ultrafdtration through an Inorganic Tubular Membrane, J. Membr. Sci. 196 (1): 103, 2002. 21. Cui Z.F., Experimental Investigation on Enhancement of Crossflow Ultrafdtration with Air Sparging, in: R. Paterson (Ed.), Effective Membrane Processes-New Perspectives, Mechanical Engineering Publications Ltd., London, 237, 1993. 22. Cui Z.F., Bellara S.R. and Homewood P., Airlift Crossflow Membrane Filtration-A Feasibility Study with Dextran Ultrafdtration, J. Membr. Sci. 128: 83, 1997. 23. Cui Z.F., Chang S. and Fane A.G., The Use of Gas Bubbling to Enhance Membraen Processes, J. Mem. Sci., 121:1, 2000. 24. Cui Z.F., Ghosh R., Yu J. and Luan S.D., An Experimental Study of Flux Enhancement with Air Sparging in A Horizontal Tubular Membrane Module, in: Proceedings of the Sixth World Congress on Chemical Engineering, Melbourne, 2001. 25. Cui Z.F. and Wright K.L.T., Gas-liquid Two-phase Crossflow Ultrafdtration of Dextrans and BSA Solution, J. Membr. Sci. 90: 183, 1994. 26. Cui Z.F. and Wright K.L.T., Flux Enhancement with Gas Sparging in Downwards Crossflow Ultrafdtration: Performance and Mechanism, J. Membr. Sci. 117: 109, 1996. 27. Ducom G., Puech F.P. and Cabassud F.P., Air Sparging with Flat Sheet Nanofdtration: A Link Between Wall Shear Stress And Flux Enhancement, Desalination 145: 97, 2002. 28. Ebara Corp, Hollow Fibre Filter Device, US Patent 4,876,006, 1989. 29. EPA, Membrane Filtration Guidance Manual, 815-D-03-008, 2003. 30. Field R.W., Wu D., Howell J. A. and Gupta B.B., Critical Flux Concept for Microfiltration Fouling, J. Membr. Sci., 100: 259, 1995.  REFERENCES  78  31. Ghosh R. and Cui Z.F. Mass Transfer in Gas Sparged Ultrafiltration: Upward Slug Flow in Tubular Mambranes, J. Membr. Sci., 162: 91, 1999. 32. Hermia J., Constant Pressure Blocking Filtration Laws-Application to Power-law NonNewtonian Fluids, Institut de Genie Chimique UCL, Louvain -la-Neuve, Belgium, 1982. 33. Hout R.V., Gulitsaki A., Barnea D. and Shemer L., Experimental Investigation of the Velocity Field Induced by A Taylor Bubble Rising in A Stagnant Water, Int. J. Multiphase Flow 28: 579, 2002. 34. Hong S.P., Bae T.H., Tak T.M., Hong S. and Randall A., Fouling Control in Activated Sludge Submerged Hollow Fiber Membrane Bioreactors, Desalination, 143: 219, 2002. 35. Imasaka T., So H., matsushita K., Kurukawa T. and Kanekuni N., Application of Gasliquid Two-phase Crossflow filtration to pilot-scale methane fermentation, Drying Techno. 11: 769, 1993. 36. Judd S.J., Le-Clech P., Taha T. and Cui Z.F., Theoretical and Experimental Representation of A Submerged Membrane Bio-reactor System, Feature, Membrane Technology, No. 135,2001. 37. Kubie J., Bubble Induced Heat Transfer in Two-Phase Gas-liquid flow, Int. J. Heat Mass Transfer 18: 537, 1975. 38. Lee C , Chang W. and Ju Y., Air Slugs Entrapped Cross-flow Filtration of Bacteria Suspension, Biotechnol, Bioeng, 41: 525, 1993. 39. Li Q.Y., Cui Z.F. and Pepper D.S., Effect of Bubble Size and Frequency on the Permeate Flux of Gas Sparged Ultrafiltration with Tubular membranes, Chem. Eng. J. 67 (1): 71, 1997.  REFERENCES  79  40. Laborie S., Cabassud C , Bourlier L.D. and Laine J.M., Flux Enhancement by A Continuous Tangential Gas Flow in Untrqfiltration Hollow Fibers For Drinking Water Production: Effects of Slug-flow on Cake Structure, Filtration Sep. 34(8): 887, 1997. 41. Laborie S., Cabassud C , Bourlier L. D. and Laine J.M., Characterisation of Gas-liquid Two-phase Flow Inside Capillaries, Chem. Eng. Sci. 54 (23): 5723, 1999. 42. Mallevialle J., Odendaal P.E. and Wiesner M.R., Water Treatment Membrane Processes. McGraw-Hill, 1996. 43. Mulder M., Basic Principle of Membrane Technology, Kluwer Academic Publishers, Dordrecht, 1991. 44. Nakoryakov V. E., Kashinsky O.N., Petukhov A.V. and Gorelik R. S., Study of Local Hydrodynamic Characteristics of Upward Slug Flow, Experiments in Fluids, 7: 560, 1989. 45. Onishi K. and Futamura O., Integrated Membrane Filtration Activated Sludge Wastewater Treatment System, in: Proceedings of the International Symposium on Fibre Science and Technology, Yokohama, 1994. 46. Ohkubo K., Hayashi T. and Nagai H., Hollow Fiber Filter Device, Ebara, US Patent 4 876 006, 1988. 47. Ozaki N. and Yamamoto K., Hydraulic Effects on Sludge Accumulation on Membrane Surface in Crossflow Filtration, Wat. Res., 35, 13: 3137, 2001. 48. Pohland H.W., Theory of Membrane Processes, Proceedings of the A W W A Annual Conference, Orlando, Fla., 1988. 49. Romero C A . and Davis R.H., Global Model of Crossflow Microfiltration Based on Hydrodynamic Particle Diffusion, J. Membr. Sci., 39: 157, 1988. 50. Rautenbach R., and Albrecht R., Membrane Processes, John Wiley, New York, 1989.  REFERENCES  80  51. Sur H.W., Li Q.Y., and Cui Z.F., Gas Sparging to Enhance Crossflow Ultrafiltration in Turbulent Flow, The IChemE Research Events, Paper-159 (CD-ROM), IChemE, UK, 1998. 52. Suda K., Shibuya S., Itoh Y., and Kohno T., Development of A Tank-Submerged Type Membrane Filtration System, Desalination 119: 151, 1998. 53. Smith T. and Derowitsch R., Membranes Decrease  Energy Demand in RO System,  A W W A ( l ) : 132, 1994. 54. Tarleton  E.S. and  Microfiltration,  Wakerman  R.J., Understanding  Flux  Decline  in  Crossflow  Part 1, Effects of Particle and Pore Size, Trans. Ind. Chem. Eng., 71 A,  1993. 55. Taha T. and Cui Z.F., CFD Modelling  of Gas Sparged Ultrafiltration  in  Tubular  Membranes, J. Membr. Sci. 210: 13, 2002. 56. Um M.J., Yoon S.H., Lee C.H., Chung K.Y. and Kim J.J., Flux Enhancement with Gas Injection in Crossflow Ultrafiltration of Oily Wastewater, Pergamon, (01)00155-5, 2001. 57. Ueda T., Hata K., Kikuoka Y. and Seino O., Effects of Aeration on Suction Pressure in A Submerged Membrane Bioreactor, Wat. Res., 31,3: 489, 1997. 58. Verberk J.Q.J.C., Hoogeveen P.E., Futselaar H. and Dijk J.C., Combined Air-water  Flush  in Dead-end Ultrafdtration, Wat. Sci. and Tech., Water Supply, 1, 5/6: 393, 2001. 59. Verberk J.Q.J.C, Hoogeveen P.E., Futselaar H. and Dijk J.C., Hydraulic Distribution of Water and Air Over a Membrane Module Using Airflush, Wat. Sci. and Tech., Water Supply, 2, 2: 297, 2002.  REFERENCES  81  60. Vera L., Delgado S. and Elmaleh S., Gas Sparged Cross-flow Microfiltration of Biologically Treated Wastewater, Membrane Technology in Environmental Management, Tokyo, 1999. 61. Vera L., Delgado S., and Elmaleh S., Dimensionless Numbers for The Steady-state Flux of Cross-flow Microfiltration and Ultrafiltration with Gas Sparging, Chem. Eng. Sci. 55, 3419, 2000.  APPENDIX  EXPERIMENT RESULTS OF TOTAL ORGANIC CARBON  82  A P P E N D I X : E X P E R I M E N T RESULTS O F T O T A L O R G A N I C C A R B O N  Analysis of Total organic carbon (TOC) All glassware used was subjected to a rigorous cleaning regime to minimize contamination. All glassware was washed with tap water and then further rinsed with distilled water several times. All the sample collection vials (40mL) were dried under room temperature before use. Samples for measuring TOC were collected in 40mL vials. TOC was determined by the Persulfate-Ultraviolet Oxidation Method 53IOC (APHA, AWWA, and WEF  1992) with the  aid of the Dohrman Phoenix 8000 UV-Persulfate analyzer (Dohrman).  TOC concentrations of raw water and filtrate of each test were shown in Table A . l and Figure A . l . TOC removal efficiency of each test was shown in Figure A.2. Table A . l : Experiment Results of TOC Removal Experiment condition  Bulk crossflow velocity(m/s)  Static water Water flow only  0 0.2 0.3 0.4 0.2 0.3 0.4 0.2 0.3 0.4 0.4(with air sparging) 0.3(water flow only)  Water flow + air sparging Water flow + air sparging + physical contact Water flow + limited physical contact  Raw water TOC(mg/L) After Before Average test test 8.6 8.7 8.65 8.8 9.2 9.0 11.1 11.8 11.45 9.4 9.3 9.5 8.6 9.0 8.8 9.9 11.6 10.75 9.1 9.9 9.5 8.3 8.8 8.55 9.4 8.7 10.1 9.1 9.9 9.5 9.4 10.5 9.95  Filtrate TOC (mg/L) 6.9 4.5 4.9 4.4 5.0 4.8 4.4 5.3 5.0 5.1 5.0  Efficiency of TOC removal* (%) 20.2 50.0 57.2 53.2 43.2 55.3 53.7 38.0 46.8 46.3 49.7  9.9  4.9  50.8  10.0  9.95  Note: * Efficiency of TOC removal (%) =100%x(Average raw water TOC-Filtrate TOC)/ Average raw water TOC  APPENDIX EXPERIMENT RESULTS OF TOTAL ORGANIC CARBON  Experimental condition Figure A . 2 : T O C removal efficiency  83  APPENDIX  EXPERIMENT RESULTS OF TOTAL ORGANIC CARBON  84  Data in Table A . l and Figure A . l , Figure A.2 illustrate that: 1. All the experimental raw water had similar TOC concentrations around 9.0mg/L throughout this study. 2. Under each experimental condition, TOC concentration of raw water slightly increased (almost 1 .Omg/L) along with membrane filtration process. 3  During each membrane filtration process, TOC in raw water was significantly removed except for static water system. TOC concentrations of filtrate with static water system were 6.9mg/L, which was much higher than others (around 4.8mg/L). As a result, TOC removal efficiency of static water system (only 20.2%) was much lower than others (around 50%).  During the filtration process, since mostly solute free water can permeate through the pores of the membrane, the solute material is left behind. Therefore, the concentration of the solute adjacent to the membrane is higher than in the bulk solution. The remained solute forms a boundary layer. The thickness of the boundary layer (at steady state) is function of the convective transport of the solute material to the membrane surface and the rate of backdiffusion into the bulk solution. The increase in the concentration of the solute at the membrane surface is called the concentration polarization effect, which can be used to explain why the filtrate TOC concentration of static water system was much higher than others and correspondingly its TOC removal efficiency was much lower than others.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0063308/manifest

Comment

Related Items