Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Role of external carbon and metal salt dosing in membrane bioreactor system to achieve limits of technology… Pattanayak, Soubhagya Kumar 2013

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

Item Metadata

Download

Media
24-ubc_2014_spring_pattanayak_soubhagya.pdf [ 5.56MB ]
Metadata
JSON: 24-1.0072134.json
JSON-LD: 24-1.0072134-ld.json
RDF/XML (Pretty): 24-1.0072134-rdf.xml
RDF/JSON: 24-1.0072134-rdf.json
Turtle: 24-1.0072134-turtle.txt
N-Triples: 24-1.0072134-rdf-ntriples.txt
Original Record: 24-1.0072134-source.json
Full Text
24-1.0072134-fulltext.txt
Citation
24-1.0072134.ris

Full Text

Role of External Carbon and Metal Salt Dosing in Membrane Bioreactor System to Achieve Limits of Technology Nutrient Removal from Municipal Wastewater  by  Soubhagya Kumar Pattanayak  B.E., Malaviya National Institute of Technology, Jaipur, India, 2002   M.A.Sc., University of British Columbia, Vancouver, Canada, 2007    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY   in    The Faculty of Graduate and Postgraduate Studies  (Civil Engineering)          THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   December 2013  ? Soubhagya Kumar Pattanayak, 2013    ii  Abstract  Membrane bioreactor (MBR) technology in conjunction with conventional biological nutrient removal has been demonstrated to be successful in recent years. However, the limits of technology (LoT) effluent goal, ? 3 mg TN/L (total nitrogen) and ? 0.1 mg TP/L (total phosphorus), could potentially push a system to the limits of its capability. The broad objective of the long-term PhD study was to investigate role of external dosing of alum in a membrane biological nutrient removal (MBNR) system targeting LoT effluent nutrient levels. Two parallel MBNR systems, modified Bardenpho configuration, were operated under similar process conditions with metal salt addition being the only difference.   The continuous flow MBNR system performance data signified the importance of external methanol and alum dosing in accomplishing the LoT nutrient removal goal. The stoichiometric methanol ratio, i.e. mg methanol required / mg NO3-N removed, was calculated to be 6.1 in reducing average permeate NO3-N concentration to 1.4 mg/L. Similarly, an average molar Al/TP ratio of 1.9 was required to reduce PO4-P concentration to 0.07 mg/L in the permeate.   Chemical phosphorus removal did not have any influence on COD removal, nitrification (except for a brief period) and denitrification. The relationship between chemical P removal and enhanced biological phosphorus removal (EBPR) was dynamic and was dependent on alum dosage concentration. At high dosage levels (i.e. 80 mg/L), alum supplementation competed with and finally, inhibited EBPR until the MBNR system was converted to a chemical P removal system.   Activated sludge modeling was undertaken to analyze its suitability in predicting the performance of an MBNR system targeting LoT goals. The model was successful in predicting nitrogen removal, while parameter calibration was required for fitting of the    iii  measured suspended solids and EBPR data. Moreover, the model could not predict the relationship between the simultaneous biological and chemical P removal accurately.     A direct batch DON measurement method, batch anion exchange resin adsorption followed by persulfate digestion, was developed and validated successfully. Using the method, the DON contribution to permeate total nitrogen was observed to vary from 7 percent to 96 percent in the parallel MBNR systems, when permeate TN concentrations were less than 3 mg/L.     iv  Preface          This statement confirms that the author of this thesis is the primary person responsible for the research contained. All experimental designs and procedures were conceived by the author with input from the supervisory committee, namely Dr. Eric Hall, Dr. Donald Mavinic, Dr. Barry Rabinowitz, and Dr. Victor Lo. The specific names of those who assisted in conducting several of the experiments have been gratefully recognized in the Acknowledgments section of this thesis.                                    v  Table of Contents  ABSTRACT .................................................................................................................................................. ii PREFACE .................................................................................................................................................... iv TABLE OF CONTENTS ............................................................................................................................. v LIST OF TABLES ..................................................................................................................................... viii LIST OF FIGURES ..................................................................................................................................... xi LIST OF ABBREVIATIONS ................................................................................................................... xvi ACKNOWLEDGEMENTS ...................................................................................................................... xix DEDICATION ........................................................................................................................................... xxi 1 INTRODUCTION ............................................................................................................................... 1 1.1 ENGINEERING PROBLEM-NUTRIENT CHALLENGE FOR WATER SYSTEMS ......................................... 1 1.2 NUTRIENT REMOVAL IN WASTEWATER TREATMENT SYSTEMS ........................................................ 1 1.2.1 Nitrogen .................................................................................................................................. 1 1.2.2 Phosphorus ............................................................................................................................. 2 1.3 NUTRIENT REMOVAL WITH MEMBRANE BIOREACTOR TECHNOLOGY ............................................... 4 2 BACKGROUND AND LITERATURE REVIEW ........................................................................... 6 2.1 INTRODUCTION ................................................................................................................................. 6 2.2 MEMBRANE BIOLOGICAL NUTRIENT REMOVAL PROCESS CONFIGURATIONS ................................... 7 2.3 NITROGEN REMOVAL IN BNR PLANTS ........................................................................................... 11 2.3.1 General introduction ............................................................................................................ 11 2.3.2 Fundamentals of biological nitrification .............................................................................. 12 2.3.2.1 Stoichiometry ...................................................................................................................................12 2.3.2.2 Microbiology ....................................................................................................................................13 2.3.2.3 Parameters influencing nitrification .................................................................................................14 2.3.3 Fundamentals of biological denitrification .......................................................................... 14 2.3.3.1 Stoichiometry ...................................................................................................................................14 2.3.3.2 Microbiology ....................................................................................................................................15 2.3.3.3 Parameters influencing denitrification .............................................................................................16 2.3.4 Fundamentals of dissolved organic nitrogen ........................................................................ 17 2.3.4.1 Parameters influencing DON removal .............................................................................................19 2.3.4.2 Measurement of DON ......................................................................................................................20 2.4 LIMITS OF TECHNOLOGY NITROGEN REMOVAL .............................................................................. 22 2.5 PHOSPHORUS REMOVAL IN BNR SYSTEMS ..................................................................................... 22 2.5.1 General introduction ............................................................................................................ 22 2.5.2 Enhanced biological phosphorus removal (EBPR) process ................................................. 24 2.5.2.1 Biochemical pathways of EBPR ......................................................................................................24 2.5.2.2 Microbiology ....................................................................................................................................26 2.5.2.3 Parameters influencing EBPR removal ............................................................................................27 2.5.3 Chemical phosphorus removal ............................................................................................. 31 2.5.3.1 General introduction ........................................................................................................................31 2.5.3.2 Mechanism .......................................................................................................................................31 2.5.3.3 Parameters influencing chemical phosphorus removal ....................................................................33 2.5.4 Interaction between EBPR and chemical phosphorus removal ............................................ 36 2.6 LIMIT OF TECHNOLOGY PHOSPHORUS REMOVAL ............................................................................ 38 2.7 MBR PROCESS MODELING ............................................................................................................. 39 2.7.1 Introduction .......................................................................................................................... 39 2.7.2 Application of ASM modeling in MBR Systems .................................................................... 43 2.8 CONCLUSIONS ................................................................................................................................. 49    vi  3 RESEARCH OBJECTIVES ............................................................................................................ 52 4 MATERIALS AND METHODS ..................................................................................................... 53 4.1 INTRODUCTION ............................................................................................................................... 53 4.2 MBNR SYSTEM .............................................................................................................................. 53 4.2.1 Design and operation ........................................................................................................... 53 4.2.2 Sample analysis .................................................................................................................... 58 4.2.3 Process start-up .................................................................................................................... 59 4.2.4 Recycle rates and nutrient supplementation ......................................................................... 59 4.3 OFF-LINE BATCH TESTS .................................................................................................................. 64 4.3.1 Methodology ......................................................................................................................... 64 4.3.2 Chemical Analysis ................................................................................................................ 65 4.4 MBNR SYSTEM MODELING AND SIMULATION ............................................................................... 66 4.4.1 General introduction ............................................................................................................ 66 4.4.2 MBNR system configuration in BioWinTM............................................................................. 67 4.4.3 Simulation strategy ............................................................................................................... 69 4.4.3.1 Data collection, analysis and steady state simulation (with default parameters) ..............................69 4.4.3.2 Sensitivity analysis ...........................................................................................................................70 4.4.3.3 Calibration of model with steady state simulation ...........................................................................71 4.4.3.4 Validation of model by dynamic simulation of MBNR (Biological) system ...................................71 4.4.3.5 Dynamic modeling of MBNR (Chemical) system ...........................................................................72 4.5 METHOD DEVELOPMENT FOR DON MEASUREMENT ...................................................................... 72 4.5.1 Batch anion exchange resin method ..................................................................................... 72 4.5.2 Persulfate digestion method.................................................................................................. 74 4.5.3 Nitrate analysis ..................................................................................................................... 75 4.5.4 Sampling for DON profiling ................................................................................................. 75 5 PERFORMANCE OF PARALLEL MBNR SYSTEMS TARGETING LOT NUTRIENT REMOVAL ................................................................................................................................................. 76 5.1 INTRODUCTION ............................................................................................................................... 76 5.2 INFLUENT WASTEWATER CHARACTERIZATION ............................................................................... 76 5.3 PROCESS PERFORMANCE ................................................................................................................. 77 5.3.1 COD profiling ....................................................................................................................... 77 5.3.2 Nitrification .......................................................................................................................... 81 5.3.3 Denitrification ...................................................................................................................... 85 5.3.4 Phosphorus removal ............................................................................................................. 90 5.3.4.1 MBNR (Biological) system ..............................................................................................................90 5.3.4.2 MBNR (Chemical) system .............................................................................................................100 5.4 SUSPENDED SOLIDS DATA ............................................................................................................ 107 5.5 LOT GOAL-MBNR (BIOLOGICAL) SYSTEM .................................................................................. 111 5.6 LOT GOAL-MBNR (CHEMICAL) SYSTEM ..................................................................................... 113 5.7 CONCLUSIONS ............................................................................................................................... 115 6 BATCH STUDIES FOR COMPARATIVE EVALUATION OF EBPR KINETICS AND STOICHIOMETRY OF THE PARALLEL MBNR SYSTEMS .......................................................... 118 6.1 INTRODUCTION ............................................................................................................................. 118 6.2 BATCH TESTS - PHASE II ............................................................................................................... 118 6.3 BATCH TESTS - PHASE III ............................................................................................................. 123 6.4 BATCH TESTS - PHASE IV ............................................................................................................. 130 6.5 BATCH TESTS - PHASE V ............................................................................................................... 139 6.6 KINETIC AND STOICHIOMETRIC PARAMETER EVALUATION .......................................................... 148 6.7 CONCLUSIONS ............................................................................................................................... 150 7 STEADY STATE AND DYNAMIC MODELING OF THE PARALLEL MBNR SYSTEMS 152 7.1 INTRODUCTION ............................................................................................................................. 152 7.2 STEADY STATE MODELING OF MBNR (BIOLOGICAL) SYSTEM WITH DEFAULT PARAMETERS ..... 153    vii  7.3 SENSITIVITY ANALYSIS ................................................................................................................. 154 7.4 CALIBRATION OF MODEL .............................................................................................................. 158 7.4.1 Parameter modification ...................................................................................................... 158 7.4.2 Measured and calibrated model suspended solids data ..................................................... 159 7.4.3 Measured and calibrated model nitrogen data ................................................................... 160 7.4.4 Measured and calibrated model phosphorus data .............................................................. 161 7.5 DYNAMIC MODELING OF THE MBNR (BIOLOGICAL) SYSTEM ...................................................... 164 7.6 APPLICATION OF THE MODEL TO THE MBNR (CHEMICAL) SYSTEM ............................................. 170 7.7 ASSESSMENT OF THE ASM-BASED MODEL .................................................................................. 176 7.8 CONCLUSIONS ............................................................................................................................... 178 8 NOVEL BATCH METHOD FOR DIRECT MEASUREMENT OF DISSOLVED ORGANIC NITROGEN IN THE PARALLEL MBNR SYSTEMS ........................................................................ 180 8.1 INTRODUCTION ............................................................................................................................. 180 8.2 QUALITY CONTROL FOR BATCH ANION EXCHANGE RESIN METHOD ............................................ 180 8.2.1 Nitrate removal ................................................................................................................... 181 8.2.2 Ammonium recovery ........................................................................................................... 183 8.2.3 DON recovery ..................................................................................................................... 184 8.3 QUALITY CONTROL FOR PERSULFATE DIGESTION METHOD ......................................................... 184 8.3.1 Ammonium conversion ........................................................................................................ 185 8.3.2 DON conversion ................................................................................................................. 185 8.3.3 Nitrate recovery .................................................................................................................. 186 8.4 PARALLEL MBNR SYSTEM PERMEATE TN SPECIATION ............................................................... 187 8.5 PARALLEL MBNR SYSTEM REACTOR DON PROFILING ............................................................... 189 8.6 CONCLUSIONS ............................................................................................................................... 191 9 CONCLUSIONS AND RECOMMENDATIONS ........................................................................ 193 9.1 RESEARCH CONCLUSIONS ............................................................................................................. 193 9.2 ENGINEERING SIGNIFICANCE ........................................................................................................ 196 9.3 RECOMMENDATIONS FOR FUTURE RESEARCH .............................................................................. 197 REFERENCES ......................................................................................................................................... 199 APPENDICES .......................................................................................................................................... 222 APPENDIX A: SIZING OF REACTORS OF PARALLEL MBNR SYSTEMS ..................................................... 222 APPENDIX B:  BATCH TEST DATA .......................................................................................................... 233             viii  List of Tables  Table 2.1 Dissimilatory and assimilatory nitrate reduction (Adapted from Zumft, 1997) ......................................................................................................................................... .15 Table 2.2 Phosphorus species in wastewater (Denham, 2007; Neethling et al., 2007; Thistleton, 2000) ............................................................................................................... 23 Table 2.3 Possible limits for phosphorus removal technologies (Adapted from Barnard, 2006) ................................................................................................................................. 39 Table 2.4 Model parameters from literature on MBR in municipal wastewater treatment (Modified from Fenu et al., 2010) .................................................................................... 47 Table 4.1 Design operating parameters of the bench-scale MBNR system ..................... 56 Table 4.2 Influent/Effluent monitoring program .............................................................. 57 Table 4.3 Reactor scan schedule ....................................................................................... 57 Table 4.4 Detail of sample analysis procedure ................................................................. 58 Table 4.5 Dosing set-points for acetate, methanol and alum supplementation during different phases of MBNR operation ................................................................................ 63 Table 4.6 Input data for BioWinTM influent specifier ....................................................... 70 Table 5.1 Influent wastewater characteristics ................................................................... 77 Table 5.2 Average permeate NO3-N concentration for different methanol dosages ........ 87 Table 5.3 Average TSS values in MBNR (Biological) and MBNR (Chemical) system.. ........................................................................................................................................ 110 Table 5.4 Average % VSS/TSS ratios in MBNR (Biological) and MBNR (Chemical) system ............................................................................................................................. 111 Table 6.1 Parallel MBNR system P-profiling during the period of batch studies (Phase II) ........................................................................................................................................ 122 Table 6.2 Parallel MBNR system P-profiling during the period of batch studies (Phase III) ................................................................................................................................... 130 Table 6.3 Parallel MBNR system P-profiling during the period of batch studies (Phase IV) ................................................................................................................................... 139    ix  Table 6.4 Parallel MBNR system P-profiling during the period of batch studies (Phase V) ........................................................................................................................................ 148 Table 6.5 Kinetics and stoichiometry of EBPR sludge acclimatized to municipal wastewater (Modified from Monti, 2006) ...................................................................... 149 Table 6.6 Literature data for ?Mg+2/ ?P and ?K+/ ?P (Modified from Barat et al., 2005) ........................................................................................................................................ 150 Table 7.1 Measured and predicted steady state (with default BioWinTM parameters) TSS data for the MBNR (Biological) system ......................................................................... 153 Table 7.2 Measured and predicted steady state (with default BioWinTM parameters) VSS data for the MBNR (Biological) system ......................................................................... 154 Table 7.3 Calibration summary ...................................................................................... 159 Table 7.4 Measured and predicted steady state (with calibrated BioWinTM parameters) TSS data for the MBNR (Biological) system ................................................................. 160 Table 7.5 Measured and steady predicted steady state (with calibrated BioWinTM parameters) VSS data for the MBNR (Biological) system ............................................. 160 Table 7.6 Measured and predicted steady state (with calibrated BioWinTM parameters) NH4-N and NO3-N data for the MBNR (Biological) system .......................................... 161 Table 7.7 Measured and predicted steady state (with calibrated BioWinTM parameters) PO4-P data for the MBNR (Biological) system .............................................................. 162 Table 7.8 2nd level calibration of BioWinTM (EBPR parameters) ................................... 163 Table 7.9 Measured and predicted steady state (with 2nd level calibration of BioWinTM) PO4-P data for the MBNR (Biological) system .............................................................. 163 Table 8.1 NO3-N removal efficiency of batch anion exchange resin method (samples with different initial NO3-N concentrations) .......................................................................... 182 Table 8.2 NO3-N removal efficiency of batch anion exchange resin method (MBNR system mixed liquor and permeate samples) .................................................................. 183 Table 8.3 NH4-N recovery efficiency of batch anion exchange resin method ............... 183 Table 8.4 Urea recovery efficiency of batch anion exchange resin method ................... 184 Table 8.5 NH4-N conversion efficiency of persulfate digestion method ........................ 185 Table 8.6 Urea conversion efficiency of persulfate digestion method ........................... 186 Table 8.7 Glutamic acid conversion efficiency of persulfate digestion method ............ 186    x  Table 8.8 NO3-N recovery efficiency of persulfate digestion method ........................... 187       xi  List of Figures  Figure 2.1 Modified Luzdack-Ettinger (MLE) MBNR system .......................................... 8 Figure 2.2 Modified Johannesburg MBNR System............................................................ 9 Figure 2.3 Post-denitrification MBNR System (Adapted from Lesjean et al., 2005) ...... 10 Figure 2.4 MBNR System (Adapted from Fleischer et al., 2005) .................................... 10 Figure 2.5 DON model (conceptualized by Parkin and McCarty, 1981a; Parkin and McCarty, 1981b) (Figure adapted from Bratby et al., 2008) ............................................ 18 Figure 2.6 Information flow between real world and modeling (Adapted from WEF MOP 31) ..................................................................................................................................... 40 Figure 2.7 Bio-P mechanism described in ASM2 (Xpp: Polyphosphate; SPO4: Orthophosphorus; SO: Oxygen) (Adapted from Henze et al., 1995) ................................ 42 Figure 4.1 Schematic of parallel MBNR systems ............................................................. 54 Figure 4.2 NO3-N profiles in anaerobic reactor of parallel MBNR systems .................... 60 Figure 4.3 NO3-N profiles in pre-anoxic reactors of parallel MBNR systems ................. 61 Figure 4.4 Different phases of MBNR operation w.r.t. acetate (anaerobic reactor), methanol (post-anoxic reactor) and alum (membrane tank) dosing ................................. 63 Figure 4.5 Batch reactor schematic ................................................................................... 65 Figure 4.6 Parallel MBNR system configuration in BioWinTM ........................................ 68 Figure 4.7 Rotating mixer with centrifuge tubes .............................................................. 74 Figure 5.1 Measured raw influent COD concentrations ................................................... 78 Figure 5.2 Estimated raw influent COD concentrations ................................................... 79 Figure 5.3 COD concentrations and removal efficiencies in parallel MBNR systems .... 80 Figure 5.4 Permeate COD concentrations with and without alum addition ..................... 81 Figure 5.5 Influent and effluent NH4-N concentrations in parallel MBNR systems ........ 82 Figure 5.6 NH4-N reactor scan data for MBNR (Biological) system ............................... 84 Figure 5.7 NH4-N reactor scan data for MBNR (Chemical) system ................................ 85 Figure 5.8 Influent and effluent NO3-N concentrations in parallel MBNR systems ........ 86 Figure 5.9 NO3-N data for anaerobic, pre-anoxic and aerobic reactors of the parallel MBNR systems ................................................................................................................. 89    xii  Figure 5.10 NO3-N data for post-anoxic and membrane reactors of the parallel MBNR systems .............................................................................................................................. 90 Figure 5.11 PO4-P removal in the MBNR (Biological) system ........................................ 91 Figure 5.12 Reactor PO4-P profile in the MBNR (Biological) system ............................. 91 Figure 5.13 Reactor PO4-P release (-)/uptake (+) profile in the MBNR (Biological) system ............................................................................................................................... 92 Figure 5.14 pH profile of influent, anaerobic and aerobic reactors of the MBNR (Biological) system ........................................................................................................... 96 Figure 5.15 Reactor suspended solids distribution in MBNR (Biological) system .......... 99 Figure 5.16 PO4-P removal in the MBNR (Chemical) system ....................................... 100 Figure 5.17 Reactor PO4-P profile in the MBNR (Chemical) system ............................ 102 Figure 5.18 Reactor PO4-P release (-)/uptake (+) profile in the MBNR (Chemical) system ........................................................................................................................................ 102 Figure 5.19 Raw influent and anaerobic VFA profile of the parallel MBNR systems ... 104 Figure 5.20  Alum-induced phosphorus removal in MBNR (Chemical) system ........... 106 Figure 5.21 TSS concentration in MBNR (Biological) system ...................................... 109 Figure 5.22 TSS concentration in MBNR (Chemical) system ....................................... 109 Figure 5.23 % VSS/TSS ratio in MBNR (Biological) system ........................................ 110 Figure 5.24 % VSS/TSS ratio in MBNR (Chemical) system ......................................... 111 Figure 5.25 Effluent TP and TN concentrations in the MBNR (Biological) system ...... 113 Figure 5.26 Effluent TP and TN concentrations in the MBNR (Chemical) system ....... 114 Figure 6.1 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase II) ........................................................................................................................................ 120 Figure 6.2 Batch test acetate profile of the parallel MBNR systems (Phase II) ............. 120 Figure 6.3 Batch test maximum specific phosphorus release and uptake profile (Phase II) ........................................................................................................................................ 121 Figure 6.4 Batch test P-released/VFAs-consumed profile (Phase II) ............................. 122 Figure 6.5 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase III) ........................................................................................................................................ 123 Figure 6.6 Batch test acetate profile of the parallel MBNR systems (Phase III) ............ 124 Figure 6.7 Batch test Mg+2 and K+1 profile of the parallel MBNR systems (Phase III) ..125    xiii  Figure 6.8 Batch test maximum specific phosphorus release and uptake profile (Phase III) ........................................................................................................................................ 126 Figure 6.9 Batch test P-released/VFAs-consumed profile (Phase III) ............................ 126 Figure 6.10 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 4 of MBNR (Biological) system (Phase III) ........................................................................................................... 128 Figure 6.11 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 4 of MBNR (Chemical) system (Phase III) ........................................................................................................... 128 Figure 6.12 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 5 of MBNR (Biological) system (Phase III) ........................................................................................................... 129 Figure 6.13 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 5 of MBNR (Chemical) system (Phase III) ........................................................................................................... 129 Figure 6.14 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase IV) ................................................................................................................................... 131 Figure 6.15 Batch test acetate profile of the parallel MBNR systems (Phase IV) ......... 132 Figure 6.16 Batch test Mg+2 and K+1 profile of the parallel MBNR systems (Phase IV). ........................................................................................................................................ 133 Figure 6.17 Batch test maximum specific phosphorus release and uptake profile (Phase IV) ................................................................................................................................... 134 Figure 6.18 Batch test P-released/VFAs-consumed profile (Phase IV) ......................... 134 Figure 6.19 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 7 of MBNR (Biological) system (Phase IV) ........................................................................................................... 136 Figure 6.20 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 7 of MBNR (Chemical) system (Phase IV) ........................................................................................................... 136 Figure 6.21 Mole K+1 and mole Mg+2 vs. mole P in Batch Test of MBNR (Biological) system (Phase IV) ........................................................................................................... 137 Figure 6.22 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 8 of MBNR (Chemical) system (Phase IV) ........................................................................................................... 137 Figure 6.23 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 9 of MBNR (Biological) system (Phase IV) ........................................................................................................... 138 Figure 6.24 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 9 of MBNR (Chemical) system (Phase IV) ........................................................................................................... 138    xiv  Figure 6.25 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase V) ........................................................................................................................................ 140 Figure 6.26 Batch test acetate profile of the parallel MBNR systems (Phase V) ........... 141 Figure 6.27 Batch test Mg+2 and K+1 profile of the parallel MBNR systems (Phase V).. ........................................................................................................................................ 142 Figure 6.28 Batch test maximum specific phosphorus release and uptake profile (Phase V) .................................................................................................................................... 143 Figure 6.29 Batch test P-released/VFAs-consumed profile (Phase V) ........................... 144 Figure 6.30 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 10 of MBNR (Biological) system (Phase V) ............................................................................................................ 145 Figure 6.31 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 10 of MBNR (Chemical) system (Phase V) ............................................................................................................ 145 Figure 6.32 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 11 of MBNR (Biological) system (Phase V) ............................................................................................................ 146 Figure 6.33 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 11 of MBNR (Chemical) system (Phase V) ............................................................................................................ 146 Figure 6.34 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 12 of MBNR (Biological) system (Phase V) ............................................................................................................ 147 Figure 6.35 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 12 of MBNR (Chemical) system (Phase V) ............................................................................................................ 147 Figure 7.1 Sensitivity analysis of ISS for the MBNR (Biological) system .................... 155 Figure 7.2 Sensitivity analysis of Fup for the MBNR (Biological) system ..................... 156 Figure 7.3 Sensitivity analysis of YH for the MBNR (Biological) system ..................... 157 Figure 7.4 Sensitivity analysis of YPO4 for the MBNR (Biological) system .................. 157 Figure 7.5 Sensitivity analysis of bH for MBNR (Biological) system ............................ 158 Figure 7.6 MBNR (Biological) system measured and dynamic modeling (with calibrated BioWinTM parameters) suspended solids data ................................................................ 166 Figure 7.7 MBNR (Biological) system measured and dynamic modeling (with calibrated BioWinTM parameters) NH4-N data ................................................................................ 167 Figure 7.8 MBNR (Biological) system measured and dynamic modeling (with calibrated BioWinTM parameters) NO3-N data ................................................................................ 168    xv  Figure 7.9 MBNR (Biological) system measured and dynamic modeling (with calibrated BioWinTM parameters) PO4-P data ................................................................................. 169 Figure 7.10 MBNR (Chemical) system measured and dynamic modeling (with calibrated BioWinTM parameters) suspended solids data ................................................................ 172 Figure 7.11 MBNR (Chemical) system measured and dynamic modeling (with calibrated BioWinTM parameters) NH4-N data ................................................................................ 173 Figure 7.12 MBNR (Chemical) system measured and dynamic modeling (with calibrated BioWinTM parameters) NO3-N data ................................................................................ 174 Figure 7.13 MBNR (Chemical) system measured and dynamic modeling (with calibrated BioWinTM parameters) PO4-P data ................................................................................. 175 Figure 8.1 Permeate DON concentration of the parallel MBNR systems ...................... 188 Figure 8.2 TN speciation of the parallel MBNR systems ............................................... 189 Figure 8.3 DON concentration in individual reactors of the MBNR (Biological) system ........................................................................................................................................ 190 Figure 8.4 DON concentration in individual reactors of the MBNR (Chemical) system ....................................................................................................................................... .191              xvi  List of Abbreviations   ADP  Adenosine Diphosphate  AOB  Ammonia-Oxidizing Bacteria ASDM  Activated Sludge/Anaerobic Digestion Model ASM  Activated Sludge Model ATP   Adenosine Triphosphate BNR  Biological Nutrient Removal BOD  Biochemical Oxygen Demand bsCOD Biodegradable Soluble Chemical Oxygen Demand  CAS  Conventional Activated Sludge CEBPR Conventional Enhanced Biological Phosphorus Removal COD  Chemical Oxygen Demand CSTR  Continuous Stirred Tank Reactor DGGE  Denaturing Gradient Gel Electrophoresis DO  Dissolved Oxygen DIN  Dissolved Inorganic Nitrogen DON  Dissolved Organic Nitrogen EBPR   Enhanced Biological Phosphorus Removal  EDTA  Ethylenediaminetetraacetic Acid  EMP  Embden-Meyerhoff-Parnas EPS  Extracellular Polymeric Substances  FISH  Fluorescence In Situ Hybridization  GAO  Glycogen-Accumulating Organism HAO  Hydrous Aluminum Oxide HAP  Hydroxy-Apatite  HDP  Hydroxy-Dicalcium-Phosphate HFO  Hydrous Ferric Oxide HRT  Hydraulic Retention Time  ICP  Inductively Coupled Plasma    xvii  ISS  Inorganic Suspended Solids LoT  Limits of Technology MEBPR Membrane Enhanced Biological Phosphorus Removal MBRs  Membrane Bioreactors MBNR Membrane Bioreactor Nutrient Removal MLE  Modified Luzdack-Ettinger MLSS  Mixed Liquor Suspended Solids MW  Molecular Weight NMP  Nuclear Magnetic Resonance NOB   Nitrite-Oxidizing Bacteria  OLR  Organic Loading Rate  PAC  Polyaluminum Chloride PAO  Phosphate-Accumulating Organism PHA  Polyhydroxyalkanoate PHB  Poly-?-Hydroxybutyrate  PHV  Polyhydroxyvalarate QA  Quality Assurance QC  Quality Control RAS   Return Activated Sludge rDON  Recalcitrant Dissolved Organic Nitrogen SEM  Scanning Electron Microscopy SERC  Staging Environmental Research Centre SMP  Soluble Microbial Product SPE  Solid-Phase Extraction SRT  Solids Retention Time TDN  Total Dissolved Nitrogen TEM  Scanning Electron Microscopy TKN    Total Kjedhal Nitrogen TN  Total Nitrogen TP  Total Phosphorus TSS  Total Suspended Solids     xviii  UCT  University of Cape Town VIP  Virginia Initiative Plant VFA   Volatile Fatty Acid VSS  Volatile Suspended Solids WAS  Waste Activated Sludge WRF  Water Reclamation Facility WRP  Water Reclamation Plant  WWTP Wastewater Treatment Plant           xix  Acknowledgements  I take this opportunity to thank my teacher and research supervisor Dr. Eric R. Hall. I had a long association with Eric and in the process learnt a great deal about guidance, critical thinking and integrity in work environment. These lessons will definitely hold me good in my future professional endeavors. Eric also had a big hand in transforming me to a confident analytical researcher. I enjoyed our meetings where I would go off-topic and seek his advice on real life issues. Eric, thank you very much. I am really glad that I did my masters and doctoral research work under your guidance.        Another person who influenced my doctoral work immensely is Dr. Venkat Mahendraker. His passion for research is really infectious. More importantly, he is a wonderful person. I really do miss his mentoring. My prayers are with Venkat for quick recovery. I am also thankful to Dr. Barry Rabinowitz, Dr. Donald Mavinic and Dr. Victor Lo for their valuable advices in the project and serving as committee members.    I would like to acknowledge National Sciences and Engineering Research Council of Canada (NSERC) for the financial support provided for this doctoral program. I extend my thanks to GE Water & Process Technologies, industry collaborators.   I express my special gratitude to Fred Koch for his suggestions on biological processes. One wise man once told me that if you want to learn everything about phosphorus removal, take Fred for a drink. I definitely agree. Moreover, he made me realize that as researchers we have the big responsibility of creating sustainable technologies. I am also thankful to Paula Parkinson and Timothy Ma for their help. I cannot emphasize enough about how much they helped me in Environmental Engineering Laboratory.   During my doctoral program, I have been very lucky to connect with students from all over the world. Alessandro, Kathy, Zaki, Isabel and Mehrnoush, thanks for your    xx  friendship and for making my time at UBC memorable. I also appreciate help from Connor, Peter, Suranjit, Parssa, Niloufar and Sam in my research work.   I am grateful to my brother (Sabyasachi) and sister (Pragyan) for all the support they provided in my studies and life.  Finally, I thank my dear wife Sainy for all her sacrifices in my PhD journey. She has been my rock solid companion and friend in the best and worst of times. Thank you sweetheart.           xxi  Dedication           Baba & Maa,  For their unconditional support, motivation & belief in me.                     1  1 Introduction  1.1 Engineering Problem-Nutrient Challenge For Water Systems  Nitrogen and phosphorus are essential nutrients for any biological growth, including algae and aquatic plants in rivers, lakes, and shallow embayed areas of the marine environment. When discharged to surface water, these nutrients may promote eutrophication which (1) adversely affects fish growth, (2) causes undesirable tastes and odours and (3) reduces the value of water for domestic, industrial, agricultural and recreational use (Oldham and Rabinowitz, 2001). Recently, a study concluded that freshwater eutrophication-related economic loss was approximately $2.2 billion per year in the United States (Dodds et al., 2009). Therefore, eutrophication control by limiting discharges of nitrogen and phosphorus to surface water is an important and challenging task for environmental engineers and scientists. The focus of this PhD research work was on wastewater treatment technology designed to remove nitrogen and phosphorus and the use of membranes for separation (filtration) of microbial solids from treated water.   1.2 Nutrient Removal in Wastewater Treatment Systems  1.2.1 Nitrogen  Municipal and industrial wastewaters contain significant amounts of nitrogen and phosphorus, and removal of these nutrients has become one of the major goals for wastewater treatment processes. Biological, chemical and physical treatment methods can be used to accomplish nitrogen removal from wastewater. Nitrogen removal can be achieved by processes such as ammonia stripping, ion exchange and membrane separation. Nonetheless, nitrogen removal by the activated sludge process is the most common method employed by environmental engineers and scientists due to the suitability of the process to most wastewaters and the relatively low cost of the application. In the activated sludge process, nitrogen removal is achieved by two    2  sequential biochemical reactions - nitrification and denitrification. During nitrification, ammonia is oxidized to nitrite and then to nitrate by two groups of autotrophic bacteria (ammonia-oxidizing bacteria ? AOB and nitrite-oxidizing bacteria - NOB) under aerobic conditions.  Denitrification, facilitated principally by heterotrophic microorganisms, occurs by the reduction of nitrate to nitrogen gas under anoxic conditions.  1.2.2 Phosphorus  Phosphorus removal by the activated sludge process was first reported in the 1950s. Greenburg et al. (1955) reported that the activated sludge process biomass could take up phosphorus in excess of its normal microbial growth requirements. Later, Srinath et al. (1959) observed that soluble phosphorus concentration could be reduced to less than 1 mg/L in batch experiments. Levin and Shapiro (1965) were the first researchers to report enhanced biological phosphorus removal (EBPR) during their work at the District of Columbia activated sludge plant. EBPR is defined as the ability of activated sludge microorganisms to remove a greater mass of phosphorus from the wastewater than that required for their basic metabolic purposes (Oldham and Rabinowitz, 2001). In the last two decades, different configurations of suspended growth biological processes have been used to remove phosphorus from wastewater. All of these biological processes include the basic steps of an anaerobic zone followed by an aerobic zone in a single sludge process. This practice is based on the original breakthrough in enhanced biological P removal technology reported by Barnard (1974). Barnard (1974) suggested that anaerobic contact between activated sludge and influent wastewater was required before aerobic degradation in order to accomplish enhanced biological phosphorus removal. The work by Barnard (1974) laid the foundation for the development of the Phoredox and Bardenpho process configurations, which form the basis of most of the EBPR process configurations in use today (Oldham and Rabinowitz, 2001).   Although enhanced biological phosphorus removal has been proven to be highly efficient, it can be sensitive to influent wastewater characteristics (especially the concentration of volatile fatty acids (VFA) in the influent and organic material that can be    3  fermented to VFA within the anaerobic zone), and subject to fluctuations in performance due to changes in environmental conditions and operation of the system.  Therefore, in many cases, physiochemical treatment has been used with biological treatment to meet effluent phosphorus limits. Chemical methods like ion exchange, crystallization and metal salt addition have been used for removal of phosphorus from water and wastewater (Clark and Stephenson, 1998). In fact, the addition of metal salt has been adopted as a common practice in full-scale wastewater treatment plants due to the efficiency of phosphorus removal and the low cost of operation. The main metals used for chemical phosphorus removal are calcium hydroxide, iron (II), iron (III) and aluminum sulphates and chlorides (Clark and Stephenson, 1998). Chemical addition of precipitants can take place at one of three locations in a wastewater treatment process namely: pre-precipitation (before secondary treatment), co-precipitation (during secondary treatment) or post-precipitation (after secondary treatment) (de Haas et al., 2000). Pre-precipitation and post-precipitation methods have very little direct impact on the biological process. However, co-precipitation can have multiple direct effects on the biological process in an activated sludge system. Some of the advantages of co-precipitation, compared to the other two precipitation methods, include: improved sludge settleability, reduced chemical consumption, relatively reduced sludge production and efficient removals of phosphorus (Clark and Stephenson, 1998). In some instances, a combination of the above approaches is employed with multiple points for metal injection for better control and operation of the process.   The effects of chemical addition on activated sludge process performance in terms of carbon oxidation, nitrification and denitrification are dependent on the chemical added, the amount of chemical used and the biological process configuration (Clark and Stephenson, 1998). Co-precipitation might have an adverse impact on the enhanced biological phosphorous removal process. The benefits of an EBPR system may be lost if simultaneous addition of chemical precipitant to achieve low effluent phosphorus concentration significantly inhibits the enhanced biological P removal mechanism (Clark and Stephenson, 1998; de Haas et al., 2000).        4  1.3 Nutrient Removal with Membrane Bioreactor Technology   In recent years, membrane filtration technology has been successfully implemented in biological wastewater treatment processes. Increasingly stringent effluent discharge standards, scarcity of land in urban areas, and the need to reclaim and reuse water are promoting the use of membrane bioreactors (MBR) as alternatives to conventional wastewater treatment processes (Ahn et al., 1999; Judd, 2006). MBRs offer some key advantages over conventional treatment processes that utilize secondary clarification ? a gravity separation process. The complete retention of biomass in an MBR system decouples the solids retention time (SRT) from the hydraulic retention time (HRT), allowing biomass concentrations to increase in the reactor, resulting in a smaller reactor and a higher organic loading rate (OLR) (Adham et al., 2001; Liao et al., 2006). Also, the MBR process not only retains the solids but also some of the macromolecules in the mixed liquor (Monti et al., 2006). The retention of macromolecules may impact the biological process as well as the membrane fouling characteristics of the mixed liquor. This may lead to degradation of slowly biodegradable retained organics, thus, leading to higher quality of effluents. However, it is also possible that the retained organics could influence the microbial population or the kinetics of the process. For example, it has been reported that nitrification kinetics are reduced in MBRs, potentially due to the inhibitory effect of retained organics and soluble microbial products (SMP) present in the MBRs (Ekama, 2009).   Many full scale MBR facilities have been designed and built for nitrification/denitrification and chemical phosphorus removal (Phagoo et al., 2005). More recently, there has been significant interest in coupling the enhanced biological phosphorus removal process to MBR technology to capitalize on the benefits of the two advanced technologies (Barnard, 2006; Fleischer et al., 2005; Monti et al., 2006; Patel et al., 2005; Peeters et al., 2010; Sibag and Kim, 2012; Smith et al., 2013). This combination is unique, because the EBPR process provides excess phosphorus accumulation in the biomass and the MBR process provides excellent solids-liquid separation which ensures that virtually no solids are present in the treated effluent    5  (Phagoo et al., 2005). Nonetheless, many membrane biological nutrient removal systems (MBNR) employ supplemental metal salt and methanol addition to enhance phosphorus (P) and nitrogen (N) removal respectively (Crawford et al., 2006; Fleischer et al., 2005; Judd, 2006). The application of MBNR is a new development and many aspects of this technology are not very well understood. In addition, many questions with regards to the biological and chemical P removal processes, and the interactions between these two phosphorus removal mechanisms in a single process still remain to be answered, especially in the context of MBNR processes.     6  2 Background and Literature Review  2.1 Introduction  Membrane filtration technology along with the suspended growth activated sludge process can provide excellent nitrogen and phosphorus removal from wastewaters. However, the selection of an activated sludge process is dependent on influent quality and the effluent nutrient limits placed on a specific treatment plant. Nutrient limits can apply to either effluent nitrogen, or effluent phosphorus, or both, depending on the location and the limiting nutrient associated with the receiving environment. There is increasing pressure to achieve very low levels of nitrogen and phosphorus in effluents (Barnard et al., 2008; Fleischer et al., 2005; Scherrenberg et al., 2008). This objective has been referred to as ?limits of technology (LoT)? and is presently defined as effluent concentrations of ? 3 mg TN/L (total nitrogen) and 0.1 mg TP/L (total phosphorus) (Barnard et al., 2008).  The selection of an MBR process configuration is, therefore, crucial for achieving combined low effluent N and P concentrations. Research work has been focusing on extremely high nutrient removal in MBR systems treating municipal wastewater (Barnard et al., 2008; Crawford et al., 2006; Fleischer et al., 2005; Lesjean et al., 2005; Meng et al., 2012; Patel et al., 2005; Phagoo et al., 2005; Sun et al., 2013).   Another very important research area is the application of activated sludge modeling (ASM) to MBNR systems targeting LoT nutrient removal. ASM models have been successfully applied in design and operation of conventional activated sludge systems (CAS) for the last two decades. The general argument is that the basic process models are similar for both systems and the conventional activated sludge process model can be converted to that of an MBR by replacing the secondary clarifier with membrane filtration.  Nevertheless, MBR models can have different kinetic and stoichiometric values due to the elevated sludge retention times and high mixed liquor concentrations applied, the accumulation of soluble microbial products (SMP) rejected by the membrane    7  filtration step, and the high aeration rates used for membrane fouling control (Fenu et al., 2010).  The application of MBR technology for LoT nutrient removal is still an active area of research with many fundamental questions yet to be answered. The objective of this literature review is to summarize and critically evaluate the current state of MBR technology for membrane enhanced biological nutrient removal in wastewater treatment systems.   2.2 Membrane Biological Nutrient Removal Process Configurations  Initial MBR process designs often incorporated the Modified Ludzack-Ettinger (MLE) process for biological nitrogen removal in wastewater treatment plants (Barnard, 2006; Crawford et al., 2006). Depending on the influent chemical oxygen demand (COD) and total nitrogen concentrations, the MLE process has the ability to achieve effluent total nitrogen (TN) concentrations between 5 and 10 mg/L (Tchobanoglous et al., 2003). However, the major limitation of earlier MLE process designs was that the oxygen-rich recycle stream from the membrane tank consumed much of the readily available COD in the influent. As a result, the denitrification potential of the process was significantly reduced (Barnard, 2006; Crawford et al., 2006). To overcome this limitation, in the current membrane-coupled MLE process, the RAS (return activated sludge) is sent to the aerobic zone, where the high dissolved oxygen (DO) is utilized. However, this change in the MBR process configuration results in the addition of an internal biomass recirculation loop, which increases the capital and operating costs of the system. Figure 2.1 shows a schematic diagram of the membrane-coupled Modified Ludzack-Ettinger (MLE) processes.    8    Figure 2.1 Modified Luzdack-Ettinger (MLE) MBNR system    The inclusion of biological phosphorus removal along with nitrogen removal added further complexity to the design of the MBR processes. Basic BNR configurations such as the A2O, Johannesburg process, five-stage Bardenpho, University of Cape Town (UCT), modified UCT (MUCT) and Virginia Initiative Plant (VIP) processes can be coupled with membranes to achieve low nutrient concentrations in the effluent. The necessity of meeting increasingly stringent N and P effluent discharge permit requirements at wastewater treatment facilities has led to improvements of MBNR process configurations.   The Cauley Creek Water Reclamation Facility (WRF) in Georgia, USA, upgraded its MLE MBR process with chemical phosphorus removal to become an enhanced MBNR facility by employing the modified Johannesburg process (Phagoo et al., 2005). In the conventional Johannesburg configuration, the return activated sludge (RAS) flows to an initial pre-anoxic zone. As a result, nitrate is minimized in the anaerobic reactor and biological phosphorus removal potential is optimized in the following zones. In the modified Johannesburg configuration, the process has been improved by introducing an additional recycle stream between the pre-anoxic zone and the anaerobic zone as shown in Figure 2.2 (Dr. James L. Barnard, pers. comm.). The additional biomass recycle improves denitrification potential by providing the denitrifying bacteria with residual biodegradable organic compounds not consumed by the phosphorus-accumulating    9  organisms (PAOs) in the anaerobic zone. Figure 2.2 provides a schematic of a modified Johannesburg MBR configuration.    Figure 2.2 Modified Johannesburg MBNR System  Lesjean et al. (2005) studied two parallel MBNR configurations (pre-denitrification and post-denitrification without external carbon addition) in pilot scale facilities. The rationale behind using a post-denitrification configuration (Figure 2.3) offered by these authors included (1) higher and more stable nitrogen removal expected due to independence from the actual influent N/COD ratio and (2) lower mixed liquor suspended solids concentration at the membranes. The outcome from their study showed that very high P removal (0.05 mg/L TP in effluent) was feasible using both configurations. However, the wastewater used in the study had high concentrations of calcium and ferric ions (approximately 130 mg Ca/L and 10 mg Fe/L), which might have contributed significantly to the natural precipitation of phosphorus in the biological reactor. The same study also reported improved nitrogen removal (down to 5 mg/L TN in effluent) at high solids retention times (SRTs) in the post-denitrification MBNR configuration. Lesjean et al. (2005) postulated that the utilization of internally stored glycogen by some denitrifying bacteria in the anoxic zone resulted in improvement of denitrification rates. The next goal of their research was to achieve less than 0.05 mg TP/L and 5 mg TN/L in the effluent by using a post-denitrification MBNR configuration (Figure 2.3).      10    Figure 2.3 Post-denitrification MBNR System (Adapted from Lesjean et al., 2005)  Fleischer et al. (2005) investigated the feasibility of a multistage MBNR system targeting extremely low effluent TN and TP concentrations. As shown in Figure 2.4, the six stage MBNR system had two pre-anoxic reactors and one post-anoxic reactor for optimization of nitrogen removal. Furthermore, methanol was added to provide an additional carbon source for denitrifying bacteria in the post-anoxic reactor. Phosphorus removal was achieved by employing both biological and chemical (alum) phosphorus removal processes. The study demonstrated the capability of the MBNR system to reliably produce effluent with TN less than 3 mg/L and TP less than 0.1 mg/L.   Figure 2.4 MBNR System (Adapted from Fleischer et al., 2005)  A review of the recently designed multi-stage MBNR systems demonstrates that membrane bioreactor technology has the capability to meet stringent nutrient discharge limits. However, most full scale MBNR plants still make provision for phosphorus    11  removal by supplemental metal addition, even though they may be designed to remove phosphorus via the EBPR process (Crawford et al., 2006; Phagoo et al., 2005). Studies have indicated that over-dosing of metal salt can precipitate too much of the available soluble phosphorus, and as a result, reduce the competitive advantage of the EBPR mechanism (Crawford et al., 2006; de Haas et al., 2000; R?ske and Sch?nborn 1994a; R?ske and Sch?nborn 1994b). Therefore, controlling metal salt addition to the process is a key operational issue for MBNR systems. Similarly, for improving nitrogen removal potential, external carbon (e.g. methanol, acetate) is used in most MBNR plants to enhance denitrification. The use of chemicals in wastewater treatment plants designed for membrane biological nutrient removal adds complexity to the fundamental mechanisms occurring within the processes. Therefore, a comprehensive review of the current nitrogen and phosphorus removal technologies will help in understanding and optimizing treatment performance in MBR plants and identifying the gaps in fundamental understanding of the processes involved in N and P removal by membrane-coupled processes, especially, the dynamics between the chemical and biological processes.  2.3 Nitrogen Removal in BNR Plants  2.3.1 General introduction  Nitrogen in influent wastewater consists of ammonium (~60%) and organic nitrogen (~40%) (Tchobanoglous et al., 2003). Typically, less than 1 percent is present as nitrate or nitrite unless the plant receives influent from industrial sources. Ammonium-nitrogen is removed by nitrification and the products of nitrification, nitrite and nitrate, are removed by denitrification in biological nutrient removal plants. On the other hand, removal of organic nitrogen is dependent on the biodegradability of the particular nitrogenous compound. Generally, the biodegradable organic nitrogen (both particulate and dissolved) fraction is removed when it is converted to ammonium-nitrogen by hydrolysis and mineralization in wastewater treatment plants (WWTPs) (Paredes et al., 2007). Additionally, the non-biodegradable particulate organic nitrogen fraction is removed by efficient solids-liquid separation processes such as clarifiers, filters and    12  membrane systems. However, non-biodegradable dissolved nitrogen or recalcitrant dissolved organic nitrogen (rDON) remains in the effluent of BNR processes.  2.3.2 Fundamentals of biological nitrification  2.3.2.1 Stoichiometry  Nitrification can be defined as a two step biological process in which ammonium (NH4-N) is oxidized to nitrite (NO2-N) by autotrophic ammonia-oxidizing bacteria (AOB) and nitrite is oxidized to nitrate (NO3-N) by nitrite-oxidizing bacteria (NOB) (Tchobanoglous et al., 2003). The reactions for ammonium oxidation to nitrate are as follows (Henze et al., 1996):  Ammonia oxidation: NH4+ + 3/2O2        2NO2- + 4H+ + 2H2O             (1) ?Go(W) = -270 kJ/mol NH4+-N  Nitrite oxidation: NO2- + 1/2O2        NO3-               (2) ?Go(W) = -80 kJ/mol NO2--N  Total ammonia oxidation reaction: NH4+ + 2O2      NO3- + 2H+ + H2O              (3)  Equation (3) illustrates a simplified picture of nitrification in wastewater treatment systems. However, nitrification is a biological process that requires external sources of carbonate and nitrogen compounds for the growth of autotrophic bacteria (Marquot, 2006). Bicarbonate (HCO3-) and ammonium act as carbonate and nitrogen sources respectively during nitrification. Therefore, the nitrification reaction with respect to bacterial (C5H7NO2) growth is as follows (Marquot, 2006):     13  NH4+ + 1.86O2 + 1.98 HCO3-        0.02 C5H7NO2 + 0.98 NO3- + 1.88 H2CO3+ H2O       (4)  2.3.2.2 Microbiology   In theory, nitrification can be carried out by both autotrophic and heterotrophic microorganisms, although under normal wastewater treatment conditions, nitrification is achieved mainly by the autotrophic microorganisms (Joo et al., 2005; Joo et al., 2007; Lin et al., 2006; Su et al., 2006). Nitrosomonas europaea is the most common species of autotrophic ammonia-oxidizing bacteria (AOB) in wastewater treatment plants (Henze et al., 1996). Other autotrophic AOB genera include: Nitrosococus, Nitrosospira, Nitrosolobus, and Nitrosorobrio (Painter, 1970). Manser (2005) reported that nitrosomonads (including Nitrosococcus mobilis), not nitrosospiras (encompassing the genera Nitrosospira, Nitrosolobus and Nitrosovibrio) are in fact the predominant AOBs in wastewater treatment plants. Nonetheless, the advances in microbiological techniques and gene analysis in recent years has revealed that various other autotrophic bacteria have the ability to act as AOB in the activated sludge environment of WWTPs (Marquot, 2006; Tchobanoglous et al., 2003).   Nitrobacter, belonging to the Alphaproteobacteria, was typically considered to be the most common species of autotrophic nitrite-oxidizing bacteria (NOB) in wastewater treatment plants (Henze et al., 2002). Wagner et al. (1996) however, contradicted this notion when they reported that Nitrobacter could not be detected in samples from nine different WWTPs.  In fact, Wagner et al. (2002) postulated that Nitrospira bacteria are dominant in most WWTPs due to their higher affinities towards nitrite and oxygen. Over the years, other researchers have also found NOB species like Nitrospira and Nitrosococcus to be prevalent in activated sludge treatment processes (Daims et al., 2001; Daims et al., 2000; Teske et al., 1994).        14  2.3.2.3 Parameters influencing nitrification  Since autotrophic nitrifiers have a lower specific growth rate than heterotrophic denitrifying microorganisms, nitrification typically limits the overall biological nitrogen removal capacity of a system. In addition, the sensitivity of nitrifying microorganisms to factors like temperature, pH, alkalinity, ammonium or nitrite and heavy metals contributes to the variability of the nitrification mechanism (Carrera et al., 2004; Hu et al., 2004; Hu et al., 2004; Tchobanoglous et al., 2003). During nitrification, ammonium oxidation (first step) has been found to be the rate-limiting step, as AOB have lower specific growth rates than NOB (Henze et al., 1996). For that reason, nitrification was considered to be a single step process (NH4 to NO3) in early process models. However, research has indicated that both reactions can be limiting at different stages of the process. Therefore it is necessary to consider individual oxidation reactions for the modeling of nitrification kinetics (Chandran and Smets, 2000a; Chandran and Smets, 2000b; Chandran and Smets, 2005).  2.3.3 Fundamentals of biological denitrification  2.3.3.1 Stoichiometry  Denitrification is carried out by facultative heterotrophic microorganisms when they use oxidized nitrogen (NO2--N or NO3--N) as electron acceptors in respiration (Henze et al., 1996). Many facultative denitrifying microorganisms have the ability to use both oxygen and oxidized nitrogen as electron acceptors due to similarities in metabolic pathways. Nevertheless, microorganisms prefer oxygen as the final electron acceptor if both oxygen and nitrate are present in the reactor. Therefore, in a single-sludge process, oxygen should be absent for part of the process if denitrification by nitrate reduction is the design objective.    Microorganism-induced nitrate reduction can be either assimilatory or dissimilatory as illustrated in Table 2.1 (Zumft, 1997). During assimilatory reduction, nitrate is reduced    15  to ammonium for synthesis of microorganisms. The dissimilatory transformation of nitrate or nitrite to a gaseous species concomitant with energy conservation is the phenomenological definition of the denitrification process (Zumft, 1997).   Table 2.1 Dissimilatory and assimilatory nitrate reduction (Adapted from Zumft, 1997) Dissimilatory branch Assimilatory branch Denitrification (energy conservation, electron sink) Ammonification  (electron sink, detoxification, energy conservation) Assimilation  (biosynthesis of nitrogen-containing compounds) Respiratory nitrate reduction NO3-        NO2- (nitrite excreted or reduced further) Assimilatory nitrate reduction NO3-       NO2- (nitrite reduced further)     Denitrification sensu stricto, nitrate and nitric oxide respiration  NO3-       NO2-       N2O (gases may be set free)  Ammonifying nitrite reduction   NO2-      NH4+  (ammonia excreted) Assimilatory nitrite reduction NO3-      NO2- (nitrite reduced further)  Assimilatory nitrite reduction NO2-        NH4+ (ammonia incorporated)  Nitrous oxide respiration N2O        N2  Denitrification associated with both branches        NO3-? NO2-        N2O  The stoichiometric equation for denitrification in which organic matter in wastewater is used as the energy and carbon source for microorganisms is shown below (Henze et al., 1996).   1/70 C18H19O9N + 1/5 NO3- + 1/5 H+        1/10 N2 + 17/20 CO2 + 1/70 HCO3- + 1/70 NH4+ + 1/5 H2O                     (5)  2.3.3.2 Microbiology   Denitrification can be achieved by both heterotrophic and autotrophic microorganisms (Zumft, 1997). However, the majority of the denitrification in wastewater treatment processes is achieved by heterotrophic organisms. The heterotrophic denitrifying bacteria genera include: Achromobacter, Acinetobacter, Agrobacterium, Alcaligenes, Arhtrobacter, Bacillus, Chromobacterium,    16  Corynebacterium, Flavobacterium, Hypomicrobium, Halobacterium, Methanomonas Moraxella, Neisseria, Paracoccus, Propionibacterium, Pseudomonas, Rhizobium, Rhodopseudomonas Spirillum, and Vibrio (Gayle, 1989; Payne, 1981). It is generally thought that Pseudomonas species is the most common microorganism for denitrification in WWTPs (Janda et al., 1988; Gray, 1990; Lazarova et al., 1992; Payne, 1981).   Recent research indicates the presence of two distinct types of denitrifiers (true denitrifiers and incomplete denitrifiers) in activated sludge treatment processes (Glass et al., 1997; Drysdale et al., 1999). Drysdale et al. (1999) found that even though many different heterotrophic bacteria contributed to denitrification in the Darvill BNR process, the majority of them were incomplete denitrifiers. The incomplete denitrifiers can only reduce nitrates to nitrites with no further reduction of the nitrites. According to the authors, denitrification was achieved via interactive associations between the complete and incomplete denitrifying heterotrophic bacteria. Their study also concluded that Pseudomonas was the most prevalent denitrifying species in the Darvill BNR process.    2.3.3.3 Parameters influencing denitrification  Reactor configuration and the nature and concentration of carbon substrates are two important parameters that typically influence denitrification rates (Henze et al., 1996; Tchobanoglous et al., 2003). Both pre- and post-anoxic reactor configurations have been used in wastewater treatment plants (Barnard et al., 2008; Crawford et al., 2006; Lesjean et al., 2005). Denitrification by pre-anoxic design is achieved by creating an anoxic zone at the head end of the process and recycling the nitrified mixed liquor from the aerobic zone of the process. On the other hand, post-anoxic denitrification is achieved by either endogenous respiration or the addition of an external carbon source, by locating the reactor downstream of the aerobic zone and such an arrangement eliminates the need for recycling the nitrified biomass.   Denitrification by endogenous respiration becomes inadequate when the total nitrogen (TN) concentration is high in the influent wastewater (Barnard et al., 2008). For    17  that reason, an external carbon source is essential for treatment plants with strict nitrogen discharge limits. Chemicals that have been used as external carbon substrates include methanol, acetate, ethanol, sucrose solution, glycerol, high-fructose corn syrup and commercially available organic carbon (Barnard et al., 2008). Methanol is the most common external carbon substrate used in the denitrification process. Methanol is also preferred in wastewater treatment plants because it is widely available, cost-effective and the PAOs cannot use it for the EBPR mechanism (Baytshtok et al., 2008; Dold et al., 2008; Louzeiro et al., 2003). On the other hand, methanol-utilizing microorganisms are slow growing, particularly at low temperatures and can be washed out at short anoxic hydraulic retention times (HRTs) and solids retention times (SRTs) (Dold et al., 2008).   2.3.4 Fundamentals of dissolved organic nitrogen  Dissolved organic nitrogen (DON) is typically the main form of nitrogen in the effluent of an enhanced nitrification-denitrification process (Pehlivanoglu-Mantas and Sedlak, 2008). Pagilla et al. (2008) studied nitrogen speciation in primary and secondary effluents of three nitrifying plants in the United States and four BNR plants in Poland. The key objective of the study was to quantify the dissolved fraction of organic nitrogen in both influent and effluent. The authors observed DON fractions ranging from 9 to 50 percent of the TN in the treated effluent. Bratby et al. (2008) reported DON concentrations varying from 0.4 to 2.2 mg/L by reviewing data from their work as well other studies. The highly variable effluent DON concentration poses a challenge to future nutrient removal initiatives, as it is very difficult to remove DON in conventional BNR plants. Hence, characterization of DON along with its fate in wastewater treatment plants should be given more attention.   In the late 1970?s and early 1980?s, Stanford University researchers focused on factors affecting DON production and removal during activated sludge treatment (Parkin and McCarty, 1981a; Parkin and McCarty, 1981b; Randtke, 1977). The key contribution of their work was the development of a model which conceptualized the distribution of    18  DON constituents. The model is based on experimental work done in batch activated sludge systems and is illustrated in Figure 2.5.      Figure 2.5 DON model (conceptualized by Parkin and McCarty, 1981a; Parkin and McCarty, 1981b) (Figure adapted from Bratby et al., 2008)  The various individual DON constituents shown in the Figure 2.5 are the following. ? Influent wastewater DON that can be categorized as biodegradable DON (DONB) and   refractory DON (DONR). DONB is utilized rapidly in the activated sludge system while DONR passes through the system unchanged.  ? DONEQ representing the DON excreted by microorganisms to maintain concentration equilibrium across the cellular membrane. ? DONG representing the DON produced during the growth of microorganisms. ? DOND represents production of DON related to microorganism decay. The model postulates that DOND is refractory.     19  In recent years, researchers have been expending effort to understand the structure and behaviour of wastewater-derived DON, even though challenges remain due to the difficulties associated with the measurement of DON (Pehlivanoglu-Mantas and Sedlak, 2006; Pehlivanoglu-Mantas and Sedlak, 2008). Pehlivanoglu-Mantas and Sedlak (2008) reported that total amino acids and ethylenediaminetetraacetic acid (EDTA) accounted for less than 30 percent of the DON in effluent. Approximately, 70 percent of the DON could not be identified in their work. The authors put forward the idea that the unknown DON compounds consisted of a complex suite of partially metabolized compounds of biogenic origin. Using solid-phase extraction (SPE) and molecular weight (MW) fractionation techniques, Pehlivanoglu-Mantas and Sedlak (2008) concluded that most of the DON compounds were of low-molecular weight and hydrophilic in nature. Westerhoff et al. (2006) estimated that 78 percent of the DON was less than 1,000 Daltons in size and approximately 7.5 percent was larger than 10,000 Daltons in effluent from the Mesa WWTP in Arizona.   2.3.4.1 Parameters influencing DON removal  The fate of DON in the activated sludge process can be influenced by operating conditions (solids retention time (SRT), hydraulic retention time (HRT)), membrane vs conventional treatment and coagulant addition. According to Parkin and McCarty (1981a), SRT is a crucial operating parameter as it controls DONB to DONR ratio in the effluent. Longer SRTs will reduce DONB in the effluent, whereas, decay may increase the DONR fraction. Therefore, SRT optimization is imperative to minimize concentrations of DON in the effluent. The authors suggested an optimum range of 6 to 10 days for SRT in wastewater treatment systems. Sattayatewa et al. (2009) investigated the role of reactor HRT in DON transformation in a four stage Bardenpho-type WWTP. The researchers found that DON was produced in the pre-anoxic zone and to a lesser extent in the first aerobic zone.   Membrane bioreactors can influence DON removal because (1) complete retention of suspended solids improves degradation of slowly biodegradable retained organics which    20  may include DONs and (2) MBRs are operated with longer SRTs than conventional activated sludge systems. Kim and Nakhla (2010) observed lower effluent DON concentrations in MBR systems as compared to a parallel conventional system. The research work was conducted on a UCT-MBR system, their patented MBR system and a conventional A2O system. The authors also concluded that there was no correlation between different MBR process configurations, system biomass concentration (ranged from 1.6 g/L to 2.7 g/L) and system HRT (6 to 8 hours) and effluent DON concentration.     Aluminum- or iron-based coagulants are used in wastewater treatment plants with the primary objective of removing phosphorus from the effluent. Depending on the structure of DON, these coagulants can also be helpful in reducing DON concentration in the effluent. Coagulants are typically more efficient in removing higher MW (molecular weight) compounds. The current knowledge indicates that most of the effluent DON consists of low molecular weight compounds, which may limit the effectiveness of coagulant addition. By reviewing the relationship between coagulant addition and DON removal, Bratby et al. (2008) observed a wide spread in effluent DON concentrations. They hypothesized that different wastewaters contain a wide variety of DON constituents and it is particularly evident at lower DON concentrations. Nonetheless, currently a lot of work is needed to fully understand the role of coagulants with lower effluent DON concentrations.  2.3.4.2 Measurement of DON  DON in WWTP effluent is typically estimated by subtracting ammonium-nitrogen from soluble total Kjeldhal nitrogen (soluble-TKN). Although this method gives satisfactory results for secondary treatment plant effluent, it has drawbacks when effluent is derived from a nitrifying treatment plant or effluent with very low total nitrogen concentrations. Effluent samples with NO3-N concentrations higher than 6 mg/L have been reported to interfere with the TKN test (Bratby et al., 2008). The interference occurs in the digestion step, in which a reaction between ammonium and nitrate causes reduction of ammonium. During the reduction reaction, ammonium is converted to nitrous oxide    21  (N2O) and disappears with the digestion gas (Bratby et al., 2008). As a result, estimated TKN concentrations can be smaller than ammonium concentrations determined separately. When effluent TN concentration is around, or less than, 3 mg/L, the detection limit is the major issue with the TKN method. DON can also be determined by subtracting dissolved inorganic nitrogen (DIN) from the total dissolved nitrogen (TDN) concentration. However, when effluent concentrations are close to limits of technology levels (? 3 mg TN), the individual constituents are present at very low concentrations. This will present significant analytical challenges in terms of detection limits, measurement precision and expensive instrumental methods (Sattayatewa and Pagilla, 2008).   The methodologies discussed above have been adopted by most of the studies related to DON determination in wastewater effluent. Nonetheless, the move towards very low TN concentration means that the indirect methods may not provide reliable DON data. For that reason, Sattayatewa and Pagilla (2008) modified the column anion exchange resin-persulfate digestion method proposed by Crumption et al. (1992) for wastewater effluent DON determination. The steps of the Crumption et al. (1992) method are the following.  ? Acidification (reduce pH < 2) water samples with HCl. This pH is well below the pKa of most of the organic constituents and as a result, they will not be adsorbed to basic anion resin. ? Passage of the acidified water samples through a very basic anion exchange resin column for NO3-N removal. ? Digestion of the nitrate-free samples with persulfate reagents during which all forms of nitrogen are converted to nitrate. ? Measurement of nitrate concentration by one of several different methods such as diphenylamine spot plates, spectrophotometric, nitrate-selective electrode, high-performance liquid chromatography, and cadmium reduction.       22  The results from the Sattayatewa and Pagilla (2008) study showed excellent nitrate removal by ion exchange treatment.  In addition, they reported organic nitrogen recoveries ranging between 96.9 to 105.8 percent, when anion exchange resin was applied to solutions containing urea, glycine, glutamic acid, sulphanilamide and methionine. Although the column method described above can be very useful in determining DONs in WWTP, it is very tedious and requires considerable time to analyze large numbers of samples.  2.4 Limits of Technology Nitrogen Removal  The likelihood of achieving the LoT total nitrogen concentration (TN ? 3 mg/L) in wastewater treatment plants is significantly limited by the recalcitrant dissolved organic nitrogen concentration. In fact, in order to achieve TN ? 3 mg/L, the wastewater effluent should not contain combined ammonium, nitrates, nitrites and residual degradable dissolved organic nitrogen concentrations of more than 1 to 1.5 mg/L (Barnard et al., 2008). The key design parameters towards that objective are longer SRTs for ammonium removal and external carbon addition for nitrate and nitrite removal. Depending on the recalcitrant nature of the DON, more external carbon will be needed for nitrate removal at greater cost. Research in the area of bioavailability and biodegradability characteristics of DON will help in shaping the future realistic effluent TN regulations.  2.5 Phosphorus Removal in BNR Systems  2.5.1 General introduction  Municipal and industrial wastewater discharges are two of the major point-sources of phosphorus to surface water. In order to remove phosphorus effectively in wastewater treatment plants, it is necessary to understand the various forms of phosphorus in the influent wastewater, as summarized in Table 2.2.        23  Table 2.2 Phosphorus species in wastewater (Denham, 2007; Neethling et al., 2007; Thistleton, 2000) Phosphate Group Species Liquid/Solid Remark Orthophosphate PO43-, HPO42-, H2PO4-,  H3PO4 (depending on pH) Liquid ? Weak acid, most dominant form, reactive;  ? Final product of the phosphorus cycle;  ? Most readily available for biological utilization Polyphosphate/ Condensed Phosphate Pyrophosphate, tripolyphosphate, metaphosphate, intracellular polyphosphate granules Liquid/solid ? Complex large molecule;  ? Precipitate in condensed form or hydrolysis to orthophosphate; ? Found in agriculture, water treatment processes and cleaning compounds  Organic phosphate Cell material and organic material Solid/liquid ? Associated with biological growth;  ? Can exist in particulate or soluble form  Chemically bound phosphorus Phosphorus precipitants, typically Al, Ca, Fe, struvite and others; adsorption to sorbent or to metal hydroxides, form complex Solid ? Includes precipitated, co-precipitated and adsorbed forms  Table 2.2 indicates that there is no gaseous form of phosphorus through which it can be removed from wastewater treatment systems. Biological, physical and chemical reactions convert different forms of phosphorus to particulates that can be removed from wastewater in a solids-liquid separation step. In this review, biological and chemical phosphorus removal processes will be discussed comprehensively to identify potential knowledge gaps.              24  2.5.2 Enhanced biological phosphorus removal (EBPR) process  2.5.2.1 Biochemical pathways of EBPR  The concept of bacterial-induced enhanced biological phosphorus removal was first published by Fuhs and Chen (1975) during their study on the Phostrip process. The authors proposed that an excess biological phosphorus removal mechanism occurred through complex biochemical pathways, mediated by microorganisms that were later referred to as phosphate-accumulating organisms (PAOs). The key biochemical reactions in the EBPR process (with acetate as substrate) are presented below. Acetate has been identified as the model carbon substrate in the vast majority of EBPR biochemical pathways studies. This can be attributed to acetate being typically the most common VFA species present in influent wastewater (Ahn and Speece, 2006; Comeau et al., 1987; Cuevas-Rodr?guez and Tejero-Monz?n, 2003; Thomas et al., 2003; Zeng et al., 2006).   Anaerobic Metabolism ? Under anaerobic conditions, PAOs assimilate acetate that has been produced by fermentation of biodegradable soluble chemical oxygen demand (bsCOD) (Tchobanoglous et al., 2003). ? Activation of acetate to acetyl-CoA follows the assimilation phase. The activation step happens simultaneously with adenosine triphosphate (ATP) catabolism. In the anaerobic reactor, ATP generation takes place mostly due to the conversion (or hydrolysis) of energy-rich phosphoric group (i.e. polyphosphate) to adenosine diphosphate (ADP) (Oehmen et al., 2007).  ? K+, Mg+2, and H2PO4- are released to the external medium during the hydrolysis of polyphosphates (Oehmen et al., 2007). ? Two molecules of acetyl-CoA condense to form acetoacetyl-CoA, and subsequently they are reduced by NADH to form 3-hydroxybutyryl-CoA. The final product of anaerobic metabolism is 3-hydroxybutyryl-CoA, also known as poly-?-hydroxybutyrate (PHB) (Oehmen et al., 2007).      25  Aerobic Metabolism ? Cell growth of PAOs occurs by utilizing stored PHB as an energy and carbon source under aerobic conditions (Mino et al., 1998; Tchobanoglous et al., 2003). ? Energy released from PHB oxidation facilitates formation of intracellular polyphosphate bonds following the uptake of soluble orthophosphate from the solution (Tchobanoglous et al., 2003). ? Additionally, some stored PHB and energy produce glycogen in EBPR process (Oehmen et al., 2007; Tchobanoglous et al., 2003).   During anaerobic metabolism, reduction power is required for conversion of acetate to polyhydroxyalkanoate (PHA) compounds (Mino et al., 1998). To explain the production of reducing equivalents in EBPR processes, two different biochemical models have been proposed (Mino et al., 1998; Oehmen et al., 2007). According to the Comeau/Wentzel model, the TCA cycle functions to oxidize a part of the acetate to CO2 and generates reducing power in the form of NADH under anaerobic conditions (Comeau et al., 1986; Wentzel et al., 1986). Alternatively, the Mino model proposed that the reducing equivalents in the form of NADH are produced when internal carbohydrate (i.e. glycogen) is converted to pyruvate via the Embden-Meyerhoff-Parnas (EMP) pathway. Also, pyruvate is converted to acetyl-CoA and CO2 in the EBPR process (Mino et al., 1987). Over the years, several experimental research studies have supported the Mino model by demonstrating the involvement of glycogen in the EBPR process (Maurer et al., 1997; Satoh et al., 1992; Satoh et al., 1996; Smolders et al., 1994b).   Using in vivo 13C and 31P nuclear magnetic resonance (NMR) techniques, Pereira et al. (1996) offered new insights on TCA and glycogen utilisation in EBPR processes. Based on redox balance reactions, the authors concluded that the reducing power generated by degradation of glycogen cannot solely account for the PHA production. Therefore, Pereira et al. (1996) suggested that both glycogen and the TCA cycle were used for the production of reducing equivalents, merging the two initial models. In recent years, researchers have studied other metabolic pathways involving utilization of glucose and propionate substrates (Lemos et al., 2003; Wang et al., 2002).      26  2.5.2.2 Microbiology  The early research identified microorganisms involved in EBPR based on culture-dependent techniques (Seviour et al., 2003). Using these techniques, Fuhs and Chen (1975) identified Acinetobacter sp. as the primary organism responsible for the EBPR process. The observations by Fuhs and Chen (1975) were subsequently supported by other microbiologists who tried to identify EBPR microorganisms (Wenztel et al., 1988). However, the use of more sophisticated techniques such as biomarkers, fluorescence in situ hybridization (FISH), 16s rRNA-based clone libraries or denaturing gradient gel electrophoresis (DGGE) has shown that Acinetobacter is not primarily responsible for EBPR (Mino et al., 1998; Oehmen et al., 2007; Seviour et al., 2003). Although many studies reported isolation of pure PAO cultures, none of them exhibited all of the EBPR characteristics, which they should theoretically possess (Mino et al., 1998). In most cases, anaerobic acetate metabolism was the key characteristic lacking in the isolated pure cultures (Jenkins and Tandoi, 1991). Initially, it was thought that only a single dominant group of microorganisms could exhibit enriched P removal in an EBPR process. However, subsequent evidence indicated that the microbial community of the EBPR process is phylogenetically very diverse (Loy et al., 2002). Using a genus-specific FISH probe ACA23a, Wagner et al. (1994) found that the bacterial population was dominated by Actinobacteria (36%) and Betaproteobacteria (36%) groups. On the other hand, Acinetobacter accounted for only 3-6 percent of the total bacterial community.   In the last decade, a great deal of EBPR microbial population research has focussed on the Rhodocyclus group from subclass 2 of the Betaproteobacteria. Bond et al. (1995) were the first to highlight presence of the Rhodocyclus group in clone libraries from the EBPR biomass. The authors postulated that Rhodocyclus organisms played a role in EBPR, as it represented the major difference between EBPR and non-EBPR microbial communities. Using FISH, Bond et al. (1999) later showed that subclass 2 of the Betaproteobacteria comprised 55 percent of all bacteria in a laboratory scale EBPR reactor. This particular phosphate-removing community is known as Candidatus Accumulibacter phosphatis (Hesselmann et al., 1999) or often abbreviated to    27  Accumulibacter in the literature (Oehmen et al., 2007). Accumulibacter has been found to be an abundant PAO in both lab scale (Onuki et al., 2002) and full scale studies (Lee et al., 2002; Saunders et al., 2003; Zilles et al., 2002).   Genomics has been proposed as a new tool for gathering information on the metabolism of an individual cell or a microbial population, along with an understanding of the associated phylogeny (Oehmen et al., 2007). Martin et al. (2006) obtained almost the complete genome of Accumulibacter from two enriched EBPR mixed liquors. By using the metagenomic data, they were able to analyze the important metabolic processes such as: (1) the transport of polyphosphate through the cell membrane and its degradation/generation under anaerobic/aerobic conditions, (2) the metabolic pathways for glycogen degradation and for the generation of additional reducing power necessary for anaerobic PHA production, (3) extracellular polymeric substances (EPS) production and (4) denitrification and nitrogen fixation ability. The current knowledge regarding metagenomics of dominant populations of EBPR systems is still limited and development in this area of research will provide full scale plants with an important tool for evaluating the EBPR potential of a sludge population.  2.5.2.3 Parameters influencing EBPR removal  Successful operation of an EBPR process depends on numerous environmental and operational factors. Review of the literature indicates that parameters such as (1) microbial competition between glycogen-accumulating organisms (GAOs) and phosphate-accumulating organisms (PAOs), (2) cation concentration and (3) solids retention time (SRT) and hydraulic retention time (HRT) are crucial for EBPR processes, in addition to the need for VFAs in the anaerobic zone.  PAOs and GAOs competition  The link between EBPR failure and microbial competition was first reported by Cech and Hartmann (1993), who noticed large numbers of ?G-Bacteria? in a glucose-fed    28  reactor. The ?G-bacteria? term in the literature now commonly refers to glycogen-accumulating organisms (GAOs). Due to their influence on EBPR performance, the biochemical pathways of GAOs have been studied extensively over the years. Several of these studies concluded that PAOs and GAOs have about the same functional pathways (Mino et al., 1998; Oehmen et al., 2007; Seviour et al., 2003). Like PAOs, GAOs can take up VFAs and store them as PHA under anaerobic conditions. However, the key difference between the two organisms is the source of energy; PAOs use polyphosphate to generate energy whereas GAOs ferment glycogen to PHA and CO2 to generate energy. Recently, Erdal et al. (2004) noted that the two organisms store different forms of PHA. According to the authors, the main storage product of PAOs is poly-?-hydroxybutyrate (PHB) while that in GAOs is polyhydroxyvalerate (PHV). Lately, studies have focussed on the factors influencing the microbial competition between PAOs and GAOs (Filipe et al., 2001c; Oehmen et al., 2005; Lu et al., 2006). Influent composition, pH and temperature have been reported in the literature as the major parameters that can influence microbial competition dynamics in EBPR processes (Oehmen et al., 2007).   The presence of volatile fatty acids (VFAs) such as acetate, propionate and butyrate is a prerequisite for biological phosphorus removal (Henze et al., 1996). Wastewaters with high concentrations of VFAs or a large fraction of readily fermentable substrate possess high phosphorus removal potential (Tanyi, 2006). This is the case because VFAs constitute the substrate for PAOs and subsequently provide the PAOs a competitive advantage over the GAOs. Therefore, VFA content of the influent is the most important design parameter in the sizing of the anaerobic zone in an EBPR process (Monti, 2006). VFAs are typically present in the feed or are produced through fermentation (hydrolysis and acidogenesis) of readily biodegradable COD (CODRb) in the anaerobic zone or in a side-stream sludge fermenter.  In EBPR systems, the pH value strongly affects the PAO-GAO competition. Typically, an increase in pH enhances PAOs selection over GAOs (Oehmen et al., 2007). The selection of PAOs can be explained by the fact that a higher pH will cause higher energy demand for acetate and simultaneously, this will negatively affect the ability of    29  GAOs to take up acetate. Since polyphosphate is the only extra energy source for PAOs as compared to GAOs, the PAOs will expend polyphosphate to meet higher energy demand at increased pH values (Filipe et al., 2001b). Additionally, Filipe et al. (2001c) proposed that a pH value of 7.25 is the critical point in an anaerobic zone of an EBPR process. Below the critical pH value, GAOs have the capability to take up VFAs faster than the PAOs. In the aerobic zone, a higher pH (7-7.5) is also beneficial for the PAOs, as low pH (6.5) will inhibit PHA utilization and biomass growth (Filipe et al., 2001a). Similar results have been reported by other researchers, who observed shifting of microbial populations from GAOs to PAOs and subsequent higher P removal at increased pH values (from ? 7 to 7.5?8.5) (Bond et al., 1999; Jeon et al., 2001; Schuler and Jenkins, 2002; Serafim et al., 2002).   The influence of temperature on microbial competition has been investigated in both low and high temperature operations. Erdal (2002) reported that low temperature (particularly less than 10 0C) improved biological phosphorus removal by providing the PAOs with an advantage over the GAOs. This can be attributed to the fact that the glycolysis reaction in the GAOs is more temperature sensitive than the energy reaction in the PAOs (Dr. James L. Barnard, pers. comm.). Consequently, PAOs outperformed the GAOs at 5 0C, even though metabolism of PAOs was slower at 5 0C than at 20 0C (Erdal, 2002). On the other hand, EBPR performance tends to slow down or diminish completely at warmer temperatures (Panswad et al., 2003; Rabinowitz et al., 2004). Predominance of GAOs at higher temperatures has been reported as the cause for the decline in phosphorus release and uptake in EBPR systems (Whang et al., 2007).  Cation concentration  Cations such as magnesium (Mg2+) and potassium (K+) must be present above critical concentrations in influent wastewater for a successful EBPR process (Machnicka et al., 2004). This is because Mg2+ and K+ are essential counter-ions for polyphosphate in the cell, and play a major role in energy generation (Sch?nborn et al., 2001; Van Groenestijn    30  et al., 1988). The role/fate of cations in biological phosphorus removal is explained as follows. ? Potassium defines cell membrane permeability and is critical for phosphate transport between the surrounding environment and the cell (Medveczky and Rosemberg, 1971). ? The enzyme polyphosphate kinase catalyzes polyphosphate biosynthesis in the presence of magnesium ions (Machnicka et al., 2004).  ? Magnesium is taken up and released simultaneously with phosphate in the EBPR process (Machnicka et al., 2004).  Solids retention time (SRT)/ hydraulic retention time (HRT)  Barnard (1988) noted that EBPR systems can operate with an SRT in the range from 2 to 40 days. However, efficient phosphorus uptake typically requires a minimum SRT of 3 to 4 days and this value can vary with mixed liquor temperature in the EBPR process (Dr. James L. Barnard, pers. comm.). Wentzel et al. (1988) proposed that successful EBPR process operation at high SRT could be attributed to the low endogenous decay rates of PAOs as compared to those of aerobic heterotrophic bacteria. At high SRTs, a proportionally higher percentage of active biomass will consist of PAOs; as a result, the phosphorus content of the biomass will increase with an increase in SRT. On the other hand, if SRT is increased to a level at which the endogenous biomass decay rate is significant, secondary phosphorus release may lead to decreased effluent quality in EBPR systems (Dr. James L. Barnard, pers. comm.).   Hydraulic retention time (HRT) selection can influence both the formation of PHAs in the anaerobic zone and phosphorus uptake rate in the aerobic zone of an EBPR process. If sufficient HRT is not allowed in the anaerobic zone, formation of PHA will not be adequate to support the desired phosphorus uptake in the aerobic zone. Similarly, if aerobic HRT is too small, phosphorus uptake could be limited (Dr. James L. Barnard, pers. comm.). However, Erdal (2002) noted that a long aerobic HRT actually reduced    31  EBPR efficiency. The author indicated that long HRT in the aerobic phase causes depletion of glycogen reserves, which can limit PHA storage in the anaerobic zone.   2.5.3 Chemical phosphorus removal  2.5.3.1 General introduction  Chemical phosphorus removal relies on the transformation of soluble, colloidal and quasi-colloidal forms of phosphorus to a particulate form and the subsequent removal of this form (along with any phosphorus already present in particulate form) by solids-liquid separation processes (Tak?cs et al., 2006). The advantage of chemical phosphorus removal is its simplicity of operation and ease of implementation in wastewater treatment systems. Depending on the physical configuration of the plant, chemical cost factors, and the effluent quality requirements, phosphorus removal can be accomplished by pre-precipitation, co-precipitation (simultaneous precipitation) or post-precipitation at wastewater treatment facilities (Nutt, 1991). Co-precipitation along with filtration is the most commonly used precipitation process that can meet effluent concentrations of 0.5 mg/L or lower on a consistent basis (Denham, 2007). For co-precipitation to occur, metal salts can be added to (1) the effluent from primary sedimentation facilities, (2) the biological reactor and (3) the effluent from a biological treatment process before secondary sedimentation (Tchobanoglous et al., 2003). Aluminum (III) and ferric (III) salts are typically used for phosphorus precipitation in wastewater treatment facilities. Aluminum (III) compounds used for phosphorus removal include alum [Al2(SO4)3. 18 H20], sodium aluminate (NaAlO2), and polyaluminum chloride (PAC). Ferric (III) is used as ferric chloride (FeCl3) or ferric sulphate [Fe2(SO4)3] in wastewater treatment facilities (WEF and ASCE, 2005).  2.5.3.2 Mechanism  The classic model describing precipitation of phosphorus with aluminum (III) and ferric (III) is as follows (Tchobanoglous et al., 2003; WEF and ASCE, 2005).    32  Precipitation with aluminum (III):  Al3+ + HnPO43-n         AlPO4 + nH+              (7)  Precipitation with ferric (III):  Fe3+ + HnPO43-n         FePO4 + nH+              (8)  From Equation (7) and (8), it can be observed that 1 mole aluminum (III) or ferric (III) will precipitate 1 mole of phosphate. However, Maurer and Boller (1999) and Smith et al. (2008) reported that the above precipitation reactions cannot explain phosphorus removal under the conditions in a wastewater treatment plant. The precipitation of phosphorus in a liquid environment is a complex phenomenon due to the formation of a range of metal phosphorus and hydroxyl complexes, as well as adsorption and co-precipitation of phosphorus onto the precipitates and complexes (Neethling et al., 2007). The different simultaneous pathways that are most likely responsible for phosphorus removal by aluminum (III) and ferric (III) species include (Maurer and Boller, 1999; Smith et al., 2008; Tak?cs et al., 2006):   ? Adsorption (surface complexation) of phosphates and organic dissolved P onto hydrous aluminum oxide (HAO) or hydrous ferric oxide (HFO); HAO/HFO is formed by rapid precipitation of acidic aluminum/ferric solution in wastewater in the presence of sufficient alkalinity; ? Co-precipitation of phosphate into the HAO and HFO structure;  ? Precipitation of ferric phosphate and mixed cation phosphates (i.e. calcium, magnesium, aluminum phosphate or hydroxyphpsphates);  ? Coagulation/flocculation of primary precipitate colloidal particles and organic colloidal phosphorus.        33  2.5.3.3 Parameters influencing chemical phosphorus removal  In spite of its widespread application, chemical P removal is not well understood and is a complex process that is difficult to predict with respect to the net chemical reactions and their results. This can be attributed to the influence of variables like metal dose-to-phosphorus ratio, contact time, mixing intensity, pH, alkalinity and age of flocs on the precipitation process (Maurer and Boller, 1999; Smith et al., 2008; Thomas et al., 1996). Szab? et al. (2008) conducted batch and continuous tests to understand the influence of the different parameters on phosphorus precipitation. A brief review of their findings along with other related research work is presented in the following paragraphs.  Metal/phosphorus ratio  Generally, the dose of metal required for phosphorus removal is dependent on the effluent phosphorus discharge permit and the design specifications of the wastewater treatment plant. Other factors that can affect the stoichiometric ratio include (Bratby, 2006): ? Phosphorus speciation, i.e. influent wastewater dominated by orthophosphates is more readily removed by chemical precipitation than condensed and organic phosphorus;  ? pH correction, i.e. acid/base should always be added before metal addition, as precipitation reactions are irreversible; and ? Efficiency of mixing at the point of metal addition.   During their investigation, Szab? et al. (2008) observed a linear relationship between dosage of coagulant and residual phosphorus at small metal dosages. However, the authors found that specific phosphorus removal decreased with an increase in concentration of metal salt. Maurer and Boller (1999) hypothesized that higher dosages typically result in an over-saturation of metals in water and therefore, will lead to an accelerated growth of precipitated flocs. Accelerated floc growth will facilitate fast conversion of micro flocs to macro flocs, including preferential binding of hydroxide    34  groups into the precipitates. Subsequently, specific phosphorus removal per dose decreases during high metal salt dosages (Maurer and Boller, 1999).    Contact time  Lijklema (1980) demonstrated that HFO flocs continued to adsorb orthophosphates after nearly 1,000 hours of contact time. Similar results were reported by Szab? et al. (2008), who observed the adsorption of orthophosphate onto HFO flocs even after over 100 hours of contact time. Furthermore, the authors observed two distinct phosphorus removal mechanisms during chemical precipitation. The initial mechanism was ?instantaneous? whereby very fast P removal occurred under ideal mixing conditions. The second mechanism was a ?slow removal or polishing step?, where a further significant decrease in soluble phosphorus occurred via precipitation. The slow removal can take a few hours to days to achieve the target residual phosphorus concentrations. The slow removal step or precipitation contact time in wastewater treatment systems can be reduced by using higher metal/P ratios. However, using continuous flow reactors under similar mixing conditions, Szab? et al. (2008) demonstrated that a system with longer HRT and SRT can provide more efficient P removal than one with a shorter HRT and SRT. Therefore, it was recommended that additional removal of residual phosphorus should be reached by a slow removal step, rather than by adding massive dosages of excess metal salts.    Mixing energy  Complete mixing along with mixing intensity are key parameters for phosphorus precipitation reactions.  Smith et al. (2008) suggested that mixing should be done at the site of addition of the acidic metal solution to attain the required complexation. Otherwise, incomplete mixing could result in low sorptive capacity and thus affect the efficiency of chemical precipitation. Similarly, rapid mixing is critical, especially for the initial ?instantaneous? step, as it provides the metal and phosphate ions ample opportunity for complexation (Szab? et al., 2008). In their experimental work, Gillberg et    35  al. (1996) demonstrated that rapid mixing significantly increased the percentage of orthophosphate removed, as compared to slow mixing. Moreover, rapid mixing was found to be more important to both aluminum and ferric ions at higher pH values.   pH  The effect of pH during chemical precipitation is very important in achieving extremely low soluble phosphorus residuals in wastewater effluent. Under favourable conditions, excellent phosphorus removal can be accomplished over a wide pH range for both water (phosphorus solution) and raw wastewater (Szab? et al., 2008). Nonetheless, the prospect of complexes forming with phosphorus is highest for an optimum pH value (Bratby, 2006). Szab? et al. (2008) suggested that the most favourable orthophosphate removal can be accomplished with pH values between 5 and 7. On the other hand, for pH values in the acidic range, a soluble phosphate complex was the most predominant form along with a limited quantity of metal hydroxide complexes. In addition, higher metal salt addition (subsequent pH decrease) can dissolve already precipitated phosphate compounds (Bratby, 2006; Szab? et al., 2008). For pH values between 7 and 10 (alkaline range), the formation of negatively charged soluble iron hydroxide [i.e. Fe(OH)4] adversely impacts phosphorus removal in wastewater (Szab? et al., 2008). Altundogan and T?men (2001) reported a similar mechanism in their work whereby reverse desorption of orthophosphate occurred from bauxite under increased pH conditions. For pH values greater than 10, calcium and magnesium can form precipitates with orthophosphate (Fettig et al., 1990) and consequently, phosphorus removal can be attained without iron or coagulant addition.   Alkalinity  Szab? et al. (2008) investigated the role of alkalinity by conducting a series of jar tests on model wastewater (phosphorus solution). The authors reported that for a specific value of pH, higher alkalinity resulted in higher residual soluble phosphorus concentrations. Szab? et al. (2008) suggested that in higher alkalinity waters, the    36  probability of metal hydroxide precipitation is greater than that of co-precipitation of phosphate and metal hydroxides. Nonetheless, the impact of alkalinity in a chemical phosphorus removal process has not been researched extensively and it requires further evaluation and analysis.   Age of flocs  According to various researchers, the aging of flocs adversely affects the long term slow phosphorus removal mechanism (Berkowitz, 2006; Lijklema, 1980; Szab? et al., 2008; Smith et al., 2008). Lijklema (1980) demonstrated that one-day old floc has half the sorption capacity of fresh HFO flocs. Similarly, Szab? et al. (2008) reported that the sorption capacity of HFO flocs decreased by 25 percent after 30 minutes of aging. The decrease in sorption capacity of old flocs might be due to the following factors:  ? The HFO molecules become denser with age (Dzombak and Morel, 1990); and ? Higher density limits the ability of orthophosphate diffusion within the molecular structure (Makris et al., 2004). To understand the above mechanisms better, Smith et al. (2008) employed scanning electron microscopy (SEM) and transmission electron microscopy (TEM) techniques to examine HFO particles of different ages. The image analysis results indicated that fresh HFO flocs were indeed much less dense than older flocs, confirming the hypothesis.  2.5.4 Interaction between EBPR and chemical phosphorus removal  Since the early days of EBPR operation, it has been discussed and debated whether, and to what extent, chemical precipitation contributes to the phosphorus removal that is seen in the process (Arvin, 1983; Marais et al., 1983; Rabinowitz and Marais, 1980). Although the occurrence of simultaneous P removal mechanisms may be beneficial for many wastewater treatment facilities, a knowledge gap still exists concerning the mechanisms involved in achieving target effluent phosphorus concentrations.     37  Rabinowitz and Marais (1980) were probably one of the first groups to study the influence of simultaneous metal salt dosing on EBPR systems. The key findings from their work included: (1) chemical phosphorus removal and biological phosphorus removal mechanisms were independent of each other and (2) chemical phosphorus removal was observed (persistence effect) even after cessation of metal dosing in the reactors. However, L?tter (1991) observed that sustained iron dosing reduced biological phosphorus removal capability in activated sludge treatment processes. The author did not provide any explanation regarding the exact mechanism for the decline in phosphorus removal performance. Boyd and L?tter (1993) later hypothesized that inhibition of biological phosphorus removal by ferric salt in EBPR systems was caused by the formation of ferric hydroxide precipitates. The ferric hydroxide precipitates take up hydroxyl ions which are necessary for the hydroxyl-mediated transport process of phosphate across bacterial cell membranes (Boyd and L?tter, 1993). During their research on interactions between the chemical and enhanced biological phosphorus elimination processes, R?ske and Sch?nborn (1994a) concluded that biological phosphorus removal was not affected by low Fe and Al concentrations (up to 3 mg/L) but was out-competed at Al concentrations of more than 6 mg/L. Their conclusion was based on the development of an analytical P-fractionation technique, which could distinguish between biologically and chemically bound phosphate in activated sludge systems. In their second paper of the series, R?ske and Sch?nborn (1994b) reported that the extent of phosphorus release (to the supernatant) under anaerobic conditions was lower in systems operating with simultaneous Fe addition. However, the authors were not able to determine whether this was due to a purely biological effect or to a chemical effect whereby biologically-released P formed a complex with Fe salt in the sludge mass. de Haas et al. (2000) have also studied the impact of Al and Fe salts on the EBPR mechanism in activated sludge systems. The authors found that chemical precipitation improved the net P removal of the EBPR processes. However, similarly to the previous research findings, de Haas et al. (2000) observed a negative influence of metal salts on the biological phosphorus removal mechanism. Additionally, de Haas et al. (2000) noted that inhibition of biological P removal was greater during periods of phosphate limitation (i.e. low effluent P concentration) conditions.      38  The review of the literature indicates that the research work on interactions between EBPR and chemical phosphorus removal has focused on conventional BNR systems. However, very little is known regarding the influence of metal salt addition on biological phosphorus removal in MBNR systems, where the membrane forms a physical barrier to passage of particulates and the systems are operated at higher MLSS concentrations. Also, a much higher concentration of metal salt will be needed than reported in the studies for LoT effluent TP concentrations. Therefore, research into simultaneous phosphorus removal mechanisms in MBNR systems will provide useful insight for understanding and optimizing phosphorus removal in wastewater treatment facilities.  2.6 Limit of Technology Phosphorus Removal  A total phosphorus (TP) concentration of 0.01 to 0.02 mg/L in effluent is being proposed as a future nutrient removal goal in wastewater treatment facilities (Barnard, 2006; Neethling et al., 2007). However, current biological systems can only achieve effluent TP less than 1.0 mg/L reliably in full scale systems (Neethling et al., 2007). In fact, the lowest effluent TP observed in biological wastewater treatment plants is 0.1 to 0.3 mg/L (Neethling et al., 2007). For that reason, researchers are currently focusing on the coupling of the biological P removal process with other advanced processes to achieve extremely low residual P concentrations. A summary of various process combinations reported in the literature along with the effluent phosphorus limits achievable are shown in Table 2.3.             39  Table 2.3 Possible limits for phosphorus removal technologies (Adapted from Barnard, 2006) Process Configurations Phosphorus Limits (mg/L) Biological treatment with chemical addition + filters 0.09-0.1 Biological or chemical treatment with post chemical + filters 0.05 Membrane reactors with biological and/or chemical treatment 0.04-0.05 Biological treatment plus iron oxide-coated sand filters 0.01-0.02 Reverse osmosis <0.01  Based on the information in Table 2.3, it can be concluded that apart from the application of reverse osmosis, technological challenges persist in achieving LoT TP limits. Alternatively, Neethling et al. (2007) proposed that fundamental understanding of different phosphorus species could help in interpreting and optimizing phosphorus removal technologies. Using standard filtration and chemical analysis, Neethling et al. (2007) assessed the different phosphorus species in water and wastewater and compared phosphorus speciation values from different treatment processes. The key conclusions from their work were: (1) advanced tertiary treatment processes achieved approximately 0.02 mg/L effluent TP and (2) refractory dissolved organic phosphorus (rDOP) compounds were the most dominant in the effluent. However, currently, very little is known about the characterization and treatability of rDOP compounds. Therefore, Neethling et al. (2007) proposed that rDOP should be given more attention in future phosphorus removal research initiatives.  2.7 MBR Process Modeling  2.7.1 Introduction  The development of the activated sludge process has expanded from carbon oxidation alone, to nitrification, denitrification and enhanced biological phosphorus removal (EBPR). These mechanisms added further complexity via involvement of three different    40  groups of microorganisms (PAOs, non-PAOs and nitrifying autotrophs) and three distinct environmental regimes (anaerobic, anoxic and aerobic) (WEF MOP 31). Modeling has therefore, become an inherent part of the design of a wastewater treatment plant. The advantages of modeling are the provision of insight into plant performance, process design, trouble shooting and operator training. Nonetheless, successful implementation of models is dependent on the information flow between the models and real world systems as demonstrated in Figure 2.6. Influent data, physical sizes of facilities, operating data, and effluent data are the information engineers/designers can obtain from real world systems. These data can be used in a model (through influent fractioning and plant configuration interfaces) to achieve specific objectives. Subsequently, model information can be used to compare and improve real world system performance. Models can be classified as mechanistic, when based on physical description of the process, or empirical, when based on quantitative description of the process. Mechanistic models are generally used in activated sludge system modeling.    Figure 2.6 Information flow between real world and modeling (Adapted from WEF MOP 31)     41  The early efforts in developing models for wastewater systems utilized only two state variables where degradation of substrate and formation of biomass was considered with first order kinetics (McKinney, 1962). Research progress in the area of activated sludge enabled modelers to incorporate additional state variables and process descriptions. The model structure was based on Monod-type kinetics.  The model was developed for both steady state (Marais and Ekama, 1976) and dynamic simulation environments (Dold et al., 1980). The above work had set the platform for development of the ASM series of models.  The Activated Sludge Model No. 1 (ASM1) is typically considered as the reference model for wastewater treatment systems. ASM1 was developed to describe carbon oxidation, nitrification and denitrification in activated sludge wastewater treatment systems. Influent carbon and nitrogenous compounds were subdivided into different fractions based on biodegradability and solubility. Chemical oxygen demand (COD) was selected to represent the concentration of organic matter in the model. ASM2 was developed to explain biological phosphorus removal in activated sludge systems. The major principles of the bio-P mechanism according to ASM2 are (Figure 2.7): (1) Growth of PAOs (XPAO) can only occur with cell internal organic matter (XPHA).  (2) The storage of cell internal matter is possible when fermentation products like acetate (SA) are present in the system. This implies that XPHA storage will only take place in the anaerobic zone of real world activated sludge systems.     42   Figure 2.7 Bio-P mechanism described in ASM2 (Xpp: Polyphosphate; SPO4: Orthophosphorus; SO: Oxygen) (Adapted from Henze et al., 1995)  ASM2d was later developed to include the denitrifying abilities of PAOs. Additionally, chemical phosphorus removal via precipitation was introduced in the ASM2 series models. The purpose of ASM3 development was to address three major defects in the ASM1 model (Gujer et al., 1999). First, a single decay process (lysis) had been used in the ASM1 model to explain the decay process in both aerobic and anoxic conditions, while endogenous respiration was the selected mechanism in ASM3. Second, ASM3 recognizes the importance of storage, i.e. all readily biodegradable substrate (SS) is taken up and stored as XSTO in the activated sludge process. Third, the circular growth-decay-growth model (also known as death regeneration model) in the case of ASM1 was replaced with an easy-to-calibrate growth-endogenous respiration model in ASM3 (Gernaey et al., 2004).  According to Koch et al. (2000), ASM3 performs better when the storage of readily biodegradable substrate is significant (industrial wastewater) or in the case of wastewater treatment plants with substantial non-aerated zones.   Note: The above discussion only addresses the high level concepts of the ASM series models. Complete information regarding model development, structure and influent COD, TN and TP fractionation can be found in Henze et al. (2000). Furthermore, a recent review of the models can be found in Hauduc et al. (2013).    43  2.7.2 Application of ASM modeling in MBR Systems   ASM models have been developed and implemented successfully in conventional activated sludge systems (CAS) for the last two decades. In recent years, ASMs have also been used to model membrane bioreactors (Jiang et al., 2009; Lobos et al., 2009; Ng and Kim, 2007; Nopens et al., 2007; Sp?randio and Espinosa, 2008; Wintgens et al., 2003). It is important to note that ASMs have been developed for CAS systems operating with SRTs in the range of 3-15 d, an HRT range of 3-5 h and an MLSS range of 1.5-4 g/L (Tchobanoglous et al., 2003). On the other hand, Itokawa et al. (2008) reported that MBR plants are operating with HRT between 4 and 6 h (13 plants), MLSS in the range of 7 to 13.5 g/L (11 plants), and an SRT between 15 and 40 days (7 plants) in European municipal plants. Fenu et al. (2010) subsequently suggested that MBRs can have different kinetic and stoichiometric values due to high sludge retention times, high mixed liquor concentration, accumulation of soluble microbial products (SMP) rejected by the membrane filtration step, and high aeration rates for scouring purposes. It has also been hypothesized that complete sludge retention can affect biomass population selection, settling characteristics and growth kinetics (Parco et al., 2006) and biomass with a higher substrate affinity and lower growth rate may have a competitive advantage over those with a lower substrate affinity and higher growth rate in MBR systems (Jiang et al., 2009). Due to the fundamental difference in operating conditions, various research groups have evaluated and estimated (both experimentally and with a trial-error approach) the suitability of ASM in MBR solids production, nitrification, denitrification and phosphorus removal.    Net sludge production typically depends on non-biodegradable particulate COD (XI), heterotrophic yield (YH) and heterotrophic decay rate (bH). As sludge production estimation is an important goal of modeling, the determination of correct values for the above parameters is crucial for MBR systems. According to Fenu et al. (2010), the high SRTs in MBRs can cause hydrolysis of the particulate COD component that is generally considered to be inert in CAS and as a result, can influence the suspended solids concentration. However, Witzig et al. (2002) did not observe hydrolysis in their work and    44  postulated that bacteria went into maintenance mode at high sludge ages. From a modeling perspective, the maintenance mode impacts the sludge yield in the same manner as hydrolysis. Nonetheless, accurate determination of XI can be done effectively by comparing measured and predicted sludge production values in wastewater treatment system (Henze et al., 2000).   Jiang et al. (2005) reported the YH value to be 0.72 g COD/g COD (at 23 0C) by conducting respirometric measurements with acetate. On the contrary, Zhang and Hall (2006) experimentally determined (with influent municipal wastewater) a YH value of 0.5 g COD/g COD in their membrane enhanced biological phosphorus removal (MEBPR) system. They reported a higher heterotrophic yield value (=0.59 g COD/g COD) for a parallel conventional enhanced biological phosphorus removal (CEBPR) system. Since the nature of the carbon source present in influent wastewater can impact the YH value, Fenu et al. (2010) suggested that a range of 0.63-0.67 g COD/g COD can be used for MBR systems. Zhang and Hall (2006)  also determined heterotrophic decay rate (bH)  with the batch respirometric method of Ekama et al. (1986), and reported 0.24 d-1 and 0.31 d-1 for MEBPR and CEBPR respectively. These values are lower than the default for bH in ASM2 (=0.4 d-1). Jiang et al. (2005) observed a similarly low bH (= 0.25 d-1) value in their MBR system. A summary of studies on MBR system non-biodegradable particulate COD (XI), heterotrophic yield (YH) and decay rate (bH) is presented in Table 2.4.   Due to the very sensitive nature of autotrophs to environmental conditions, their performance in nitrogen removal has been widely studied in MBR systems. While Monti and Hall (2008) observed that nitrification rate was 15 to 75 percent greater in CAS as compared to a parallel MBR system, other researchers have reported that nitrifier growth was significantly higher in a submerged MBR and nitrification was more effective and stable than in a CAS (Gao et al., 2004; Munz et al., 2008; Parco et al., 2006). Though Manser et al. (2005) did not find any difference in MBR and CAS system maximum specific ammonium uptake rates, the MBR system performed better during transient shock loads, especially at low temperature and relatively low dissolved oxygen (DO).    45  The discrepancy in nitrifier activity in MBR vs CAS can be explained by hypotheses such as difference in microbial population selection, bioavailability of substrates due to the smaller size of flocs in MBRs (Manser et al., 2005) and their tendency to grow in clusters in different areas of the floc (Fenu et al., 2010). Along with the experimental work, researchers have attempted to model MBR nitrification with default ASM parameters. Jiang et al. (2005) observed autotrophic yield value (YA) of 0.25 g N/g COD, which is closer to the ASM default value of 0.24 g N/g COD. Manser et al. (2005) reported similar ammonium-oxidizing bacteria (AOB) decay values (bA) (0.13 d-1) for both CAS and MBR systems. The authors however found nitrate-oxidizing bacteria (NOB) decay rate to be slightly different with 0.28 d-1 for CAS and 0.17 d-1 for MBR systems. The influence of autotrophic growth rate (?A) and decay rate (bA) was studied by Sperandio and Espinosa (2008), whereby they calibrated these parameters for a wide range of SRTs.  They concluded that ASM1 default values (0.8 d-1; 0.04 d-1) overestimated ammonium removal for all the SRTs studied, whereas ASM3 (1 d-1, 0.15 d-1) gave better results but minimized the SRT influence. The authors proposed new values for nitrifier growth rate (?A = 0.45 d-1) and decay rates (bA = 0.04 d-1) for MBRs. Another approach to nitrification modeling has been to calibrate half saturation constants KNH and KOA, which directly influence the model effluent ammonium concentration. The calibration is primarily based on the principle that lower transfer resistance is observed in MBRs due to smaller floc size. Nonetheless, Fenu et al. (2010) suggested that selection of KNH and KOA values should depend on careful examination of operational parameters such as SRT, MLSS concentration, viscosity, oxygen concentration and floc size distribution. Table 2.4 summarizes different studies focusing on MBR nitrification modeling along with their calibration data.  Modeling of denitrification in MBR systems with default ASM parameters has received limited attention from researchers. The general conclusion has been that denitrification is not affected by membrane configuration and hence, default values for reduction factor for anoxic growth, and the anoxic heterotrophic yield can be used for modeling (Fenu et al., 2010; Parco et al., 2007). Parco et al. (2007) demonstrated that the denitrification rate in an MBR system was similar to that of CAS process (= 0.25 mg    46  NO3/mg SS.d) while operating with an SRT of 20 days and mass load of 0.14 g COD/g MLSS.d. However, it is important to note that process configurations with sludge recirculation directed from the DO-saturated membrane tank to the anoxic tank can negatively impact denitrification potential (Sarioglu et al., 2008).  This mechanism can be addressed by calibrating the parameter KOH (Table 2.4).          47  Table 2.4 Model parameters from literature on MBR in municipal wastewater treatment (Modified from Fenu et al., 2010)  Model Parameter Unit Default ASM1 Jiang et al., 2005 Zhang  and Hall, 2006 Spe?randio et al., 2005 Manser et al., 2005 Jiang 2007 Sarioglu et al., 2008 Delrue, 2008 Jimenez et al., 2008 Erftverband 2001, 2004 RWTH 2008 Range of values     ASM1  ASM2 ASM1; ASM3 ASM1 ASM2d ASM1 endogenous decay model  ASM1 ASM1 modified  ASM1 ASM1   SRT D  20 17-25 10-110 20  38 30-60 15    Nitrification ?maxA d-1 0.8   0.45   1  0.8   0.45-1.00 ba d-1 0.05-0.15 0.08  0.04  0.055 0.06  0.15   0.04-0.15 KNH mg N-NH4/l 1   0.25-0.6  0.2 2 1  0.1  0.10-2.00 KOA Mg O2/l 0.4    0.18 0.2 1.25 1    0.18-1.25 YA g COD/   g N 0.24 0.25          0.24 Denitrification, COD oxidation, Sludge production % XI %  COD 15        17.5    YH g  COD/ g COD 0.67 0.72 0.50    0.66   0.67 0.52-0.9 0.63-0.67 bH d-1 0.62 0.25 0.24   0.4 0.24    0.03-0.47 0.24-0.4 KO,NOB mg N/l 0.5    0.13  2  1   0.13-2 KOH mg O2/l 0.2    0.05  1  0.22   0.05-1    48  Research efforts in the area of MBR EBPR modeling with default ASM parameters have provided mixed results (Cosenza et al., 2013; Zuthi et al., 2013). Parco et al. (2007) found that MBR system anaerobic P-release rate, acetate consumption rate, anoxic P-uptake rate and aerobic P-uptake rate were very close to CAS EBPR rates with mixed cultures. In addition, they reported that different volatile suspended solids concentrations have no impact on the above mentioned rates. The authors subsequently concluded that EBPR kinetic parameters are comparable in both MBR and CAS systems. On the other hand, modeling work by Jiang et al. (2008) with default ASM2d parameters resulted in overestimation of nitrate concentration and underestimation of phosphorus concentration. Using a trial and error calibration approach, they simultaneously reduced anaerobic acetate production and the aerobic/anoxic phosphorus uptake rate (qfe = 1 d-1, qpp =1.1 d-1 and ?NO3,PAO = 0.4) for data fitting.   Both EBPR and chemical phosphorus removal is being used increasingly in WWTPs and studies related to their interplay have been summarized in Section 2.5.4. Modeling work in this area can provide further insight into the process phosphorus removal performance, optimization of metal salt dosage and the viability of EBPR when higher concentrations of metal salt are added for achieving the effluent objectives (in the context of LoT phosphorus removal). Liu et al. (2011) used an activated sludge model combined with a chemical precipitation model to study the impact of alum on biological phosphorus removal, nitrification and denitrification in an MBR system targeting 0.025 mg/L TP in the effluent. The authors conducted their experimental work in a pilot scale UCT MBR system (SRT of 51 days and alum dosage of 17.5 mg/L) and modeling work was completed in BioWinTM. The two major conclusions from their work were (1) sludge production, COD, nitrification and denitrification performance were predicted reasonably by calibrating only AOB growth rate and (2) alum dosing, as predicted, inhibited EBPR while the measured data did not provide such evidence. However, the authors did not provide detailed information on EBPR activity such as anaerobic release and aerobic uptake potential of PAOs before and after alum addition.       49  This literature review shows that current MBR process modeling is predominantly based on experience with conventional activated sludge models and researchers have successfully implemented these with calibration of stoichiometric and kinetic parameters. However, LoT nutrient removal will require modeling of nitrification, denitrification by both influent and external carbon addition, EBPR and chemical phosphorus removal. This very complex task will require detailed understanding of the predicted vs measured data, careful calibration of the parameters based on current literature and process knowledge and conclusions regarding the process mechanisms and performance.  2.8 Conclusions  Based on the literature review presented above, following conclusions can be drawn.   Currently known  ? The future BNR technology will move in the direction of achieving very low effluent nitrogen (i.e. ? 3 mg TN/L) and phosphorus (i.e. ? 0.1 mg TP/L) concentrations. Membrane biological nutrient removal (MBNR) is a novel technology that can contribute to the achievement of these goals. Nevertheless, external carbon and metal salt supplementation has been proposed as necessary to meet stringent nitrogen and phosphorus discharge limits respectively, in MBNR systems.  ? Metal salt addition is imperative when an EBPR system is targeting very low effluent TP concentrations. The relationship between the two simultaneous phosphorus removal mechanisms has been investigated in conventional BNR systems and some results indicate inhibitory effects on the bio-P mechanism. It is however important to mention that the impact is dependent on the added metal salt concentration. ? ASM models have been developed and implemented successfully in conventional activated sludge systems (CAS). MBRs, on the other hand, may require different kinetic and stoichiometric values due to the long sludge retention times, high mixed liquor suspended solids concentrations, the accumulation of soluble microbial products (SMP) rejected by the membrane filtration step, and the high aeration rates    50  used for membrane scouring purposes. Researchers have used both experimental and trial-error approach to estimate parameters for sludge production, nitrification, denitrification, EBPR and chemical phosphorus removal in MBR systems. ? Effluent dissolved organic nitrogen (DON) concentration becomes the dominant nitrogen fraction (varying from 0.4 to 2.2 mg/L) for wastewater treatment systems targeting LoT nitrogen removal. That means effluent should not contain combined ammonium, nitrates or nitrites of more than 1 to 1.5 mg/L. For design considerations, complete nitrification and denitrification, aided by supplementation with an external carbon source, are necessary. ? DON in WWTP effluent can either be calculated by subtracting ammonium from soluble Kjeldahl nitrogen (sol-TKN) or by subtracting dissolved inorganic nitrogen (DIN) from the total dissolved nitrogen (TDN) concentration. However, when effluent concentrations are close to the limits of technology level (? 3 mg TN), this will represent significant analytical challenges in terms of detection limits, measurement precision and expensive instrumental methods. Recently, a combined column anion exchange resin (for residual nitrate removal) and persulfate digestion (for conversion of DON to nitrate) method has been proposed for reliable measurement of effluent DON concentrations.   Knowledge gaps  ? MBRs in conjunction with conventional nitrification/denitrification and EBPR have been demonstrated to be successful in recent years. However, the LoT effluent goal could potentially push a system to limits of its capability. Currently, little information exists regarding performance of an MBR system in such a scenario. Moreover, the significance of external carbon and metal salt dosing for enhanced denitrification and phosphorus removal respectively, has not been explored in the context of the LoT objective.            ? The interactions between simultaneous biological and chemical phosphorus removal are often very complex and poorly understood. Due to the poor reliability of EBPR in meeting ? 0.1 mg/L TP goal consistently, metal salt might be added at greater than    51  stoichiometric requirements. In addition, the longer SRTs maintained in MBNR systems will enable build up of metal salt inventory. The impact of such levels of metal salt dosing on EBPR and nitrogen removal is currently unknown.   ? ASM modeling can be very useful in predicting nitrogen and phosphorus removal capabilities of a selected MBNR configuration. In particular, the requirement for and the efficiency of external carbon and metal salt dosing can be assessed. Nevertheless, there is very little in the current literature describing efforts to model an MBNR system targeting LoT effluent concentrations.  ? The importance of the DON fraction in LoT level TN effluent has been discussed extensively. However, a knowledge gap exists regarding how it is produced or utilized in reactors of an MBNR system.  Moreover, high alum dosing, originally meant for LoT phosphorus removal, can potentially reduce DON concentrations in permeate (via a coagulation effect). This theory has not been investigated yet. Another key issue regarding DON is the development of a direct and reliable measurement method, which can be used to process bulk samples from reactors and permeates. The direct measurement method that has been proposed recently is ill equipped for the processing of large numbers of samples.       52  3 Research Objectives  Four fundamental research questions were developed following the literature review.  1. Is external carbon and metal salt dosing a significant requirement in an MBNR system targeting LoT TN and TP removal?  i. Could LoT level nitrogen and phosphorus removal be achieved in an MBNR system without supplemental additions of carbon and/or metal salt?  ii. If external carbon dosing is needed, is the stoichiometric requirement known for achieving extremely low effluent nitrate levels? iii. If  external metal salt dosing is needed, a. Is the stoichiometric requirement known for achieving extremely low effluent phosphorus levels? b. Is the relationship with EBPR defined by dosing concentrations? c. Does it influence COD removal, nitrification and denitrification?  2. Do EBPR kinetics become progressively inhibited when alum addition to the mixed liquor of an MBNR system is increased from small dose to a high dose in a stepwise manner?   3. Does ASM-based modelling successfully predict the influence of external carbon and metal salt dosing on COD, nitrogen and biological phosphorus removal performance of an MBNR system targeting LoT effluent goals?  4. Is there a direct method that might be applied to measure DON concentrations reliably in an MBNR system targeting LoT nitrogen goals?      53  4 Materials and Methods  4.1 Introduction  The methodology for the project is presented in four parts: (1) design, construction and operation of two parallel laboratory scale MBNR systems (2) off-line sequential anaerobic/aerobic batch tests, (3) modeling of the MBNR system and (4) development of a reliable and direct technique for DON measurement. The following sections provide detailed descriptions of the methodologies.  4.2 MBNR System  4.2.1 Design and operation  There are a number of process configurations that can be used for enhanced BNR along with membranes for solids-liquid separation. For the present study, the chosen process configuration was a modified Bardenpho-type reactor with five different zones: (1) anaerobic, (2) pre-anoxic, (3) aerobic, (4) post-anoxic and (5) membrane tank. The following modifications were made to the conventional configuration: ? The 2nd aerobic reactor was replaced by a membrane tank which also functioned as solids-liquid separator; ? The return activated sludge (RAS) flow from the membrane tank was directed to the aerobic zone, rather than to the upstream anaerobic zone; ? A second recirculation flow was added from the pre-anoxic zone to the anaerobic zone.   The two parallel MBNR systems were operated under similar process conditions, with metal salt addition being the only difference as shown in Figure 4.1.     54   a. MBNR (Biological) system  b. MBNR (Chemical) system Figure 4.1 Schematic of parallel MBNR systems    55  The MBNR systems were studied at lab-scale with municipal wastewater as the influent to each treatment line. The influent was municipal wastewater collected from the Staging Environmental Research Centre (SERC), South Campus, University of British Columbia. Initially, the wastewater was collected two times per week and stored in the refrigerator at 4 0C. However, unsteady influent VFA concentrations (which impact EBPR performance) and the need for data for steady state modeling necessitated the shifting of influent collection to once per week on operation day 193. The influent supply to the MBNR systems was renewed every day. The common influent tank had a mixer operating continuously that provided gentle mixing to keep particles in suspension. Also, the influent line was provided with a fine screen (pore size 1 mm) to hold back material that might clog the membranes. The whole set-up was operated in a temperature controlled-room at a constant temperature of 20 0C that was maintained during the whole study period. All the reactors were cylindrical in shape and built from plexiglass. The forward flow of the wastewater was achieved by gravity and peristaltic pumps were installed for influent flow, recycle flow, permeation and nutrient addition. The flow rate of the influent was controlled by level sensors employed in the membrane tank. The sensors facilitated on/off control of both the influent and permeate pumps depending on the mixed liquor level in the membrane tank. The rationale was to stop overflow of mixed liquor or emptying of the membrane tank during off-hours and weekends. Mixers were placed in all the reactors to achieve ideal continuous stirred tank reactor (CSTR) conditions, with the exception of the membrane tanks. Air supply to the aerobic reactor was intermittent in the MBNR systems. Using an on-line dissolved oxygen (DO) probe (calibrated), solenoid valve and a controller, the set point was fixed at 2 mg/L. The on-line DO probe reading was compared with an external DO probe value on a daily basis and it was cleaned periodically to avoid potential oxygen depletion in the aerobic reactor. The membrane tank received a constant air supply of 8 L/minute which resulted in oxygen-saturated mixed liquor in the membrane tank.   The design operating conditions of the bench-scale MBNR systems are summarized in Table 4.1. The reactor volumes were calculated based on the design net flow and the desired HRT. The MBNR systems were operated with constant SRT and variable MLSS    56  concentration during the whole study period. APPENDIX A provides detailed information on the sizing of the reactors based on net flow and HRT and mixer dimensions. Furthermore, photographs are included to illustrate different component of bench scale parallel MBNR systems.   Table 4.1 Design operating parameters of the bench-scale MBNR system Parameter Value Net flow (Q) 33.45 L/day Anaerobic reactor volume 2.5 L Pre-anoxic reactor volume 5.6 L Aerobic reactor volume 14.65 L Post-anoxic reactor volume 5.6 L Membrane tank volume 1.5 L Total hydraulic retention time (HRT) 13.83 h Anaerobic HRT 1.5 h Pre-anoxic HRT 3 h Aerobic HRT 6 h Post-anoxic HRT 3 h Membrane HRT 1/3 h Solids retention time (SRT) 25 days Temperature   20 0C Membrane module  ZW-1, submersible Membrane pore size (nominal) 0.04 ?m Membrane flux Membrane Operation 15 L/m2.hr Relaxation Mode (5 min ON/1 min OFF)  During the experimental work, MBNR operating data were collected manually each day and recorded in log sheets. The log sheets were also used to record major operations and maintenance events for reference. The parameters monitored were as follows: ? Date ? pH (influent, reactors and permeate) (by using portable pH Testr BNC (r)) ? DO (aerobic reactor) (by using portable Hach HQ30d DO probe) ? Permeate Flux  ? Transmembrane pressure (TMP) before/during/after relaxation     57  ? Sludge wasting  ? Recycle rates (once/week)  The sampling schedule for the MBNR systems is shown in Table 4.2 and Table 4.3. The schedule was developed to obtain a consistent and comprehensive evaluation of the performance of the MBNR systems. Influent and effluent grab samples were collected on Mondays, Wednesdays and Fridays for COD, ammonium, nitrate/nitrite and orthophosphate analysis. The same schedule was followed for VFA in the influent and the anaerobic zone of both MBNR systems. On the other hand, samples for TKN, TP and reactor scan were collected twice i.e. on Mondays and Fridays. Measurement of total suspended solids (TSS)/ volatile suspended solids (VSS) was undertaken only on Fridays.   Table 4.2 Influent/Effluent monitoring program Parameter Number of sampling events/week Influent Permeate Total suspended solids (TSS)/ Volatile suspended solids (VSS) 1 1 - Volatile fatty acid (VFA) 3 1 - Total COD (CODTot)/ Filtered Flocculated COD (CODFF)/ Soluble COD (CODSol) 3 1 1 Total Kjeldahl nitrogen (TKN)/ Dissolved TKN 2 1 1 NH4-N 3 1 1 NO3-N 3 1 1 Total phosphorus (TP)/ Dissolved TP 2 1 1 PO4-P  3 1 1  Table 4.3 Reactor scan schedule Parameter Number of  sampling events/week Anaerobic Zone Pre-anoxic Zone Aerobic Zone Post- anoxic Zone Membrane Tank  VFA 3 1 - - - - NH4-N 2 1 1 1 1 1 NO3-N 2 1 1 1 1 1 PO4-P 2 1 1 1 1 1       58  4.2.2 Sample analysis  Table 4.4 shows detailed information on final sample volume, preservation and chemical analysis procedures   Table 4.4 Detail of sample analysis procedure Parameter Sample Volume Filtration (0.45 ?m)  Preservation Chemical Analysis  (APHA et al., 2005) Equipment Information CODTot/ CODSol/ CODFF* 2 mL No for CODTot/; Yes for CODSol/CODFF  5220 D. Closed Reflux, Colorimetric Method  Hach DR/2000 Direct Reading Spectrophotometer VFA 1-2 mL Yes 2% Phosphoric acid solution (H3PO4) 5560 D. Gas Chromatography Method  HP 5890 Series II Gas Chromatograph FID (Flame Ionization Detector) NH4-N 5-8 mL Yes 5 % Sulfuric acid solution (H2SO4) 4500-NH3 G. Automated Phenate Method Lachat QuikChem 8000 Flow Injection Analyzer NO3-N 5-8 mL Yes 0.1 g mercuric acetate in 20 mL acetone and 80mL water solution  4500-NO3-F. Automated Cadmium Reduction Method Lachat QuikChem 8000 Flow Injection Analyzer PO4-P 5-8 mL Yes 0.1 g mercuric acetate in 20 mL acetone and 80mL water solution 4500-P F. Automated Ascorbic Acid Method Lachat QuikChem 8000 Flow Injection Analyzer TKN 10 mL- Influent 20 mL- Effluent  No 5 % Sulfuric acid solution (H2SO4) 4500-NOrg D. Block Digestion and Flow Injection Analysis  Lachat QuikChem 8000 Flow Injection Analyzer TP 10 mL- Influent 20 mL- Effluent No 5 % Sulfuric acid solution (H2SO4) 4500-P H. Manual  Digestion Method  Lachat QuikChem 8000 Flow Injection Analyzer TSS/VSS 10 mL Yes (with 1.2 ?m filter) Analyzed directly after sampling 2540D. Total Suspended Solids / 2540 E. Fixed and Volatile Solids   *: CODFF (flocculated filtered COD) sample preparation was completed by adopting the Mamais et al. (1993) methodology. The first step was flocculation of 100 mL influent wastewater samples with 1 mL of 100 g/L ZnSO4. Then, the pH of the mixed sample was adjusted to approximately 10.5 with 6 M sodium hydroxide solution. After a few minutes of settling, the final step was filtration with 0.45 ?m membrane filter.    59  4.2.3 Process start-up  The bench scale system was seeded on 18th February, 2010 with mixed liquor from the full scale BNR WWTP at Salmon Arm, British Columbia. The Salmon Arm treatment plant employs a trickling filter and a suspended growth system combination for ammonium and phosphorus removal respectively. The volume of mixed liquor transported to UBC was sufficient to fill the reactors to their designed HRT (Table 4.1) levels. The MBNR systems were operated in continuous mode from day 1 with municipal wastewater providing the required nutrients. pH, DO and flow rates were closely monitored from the beginning. Initially, the systems shown in Figure 4.1 were six stage processes with a second aerobic tank between the post-anoxic reactor and the membrane tank. After review of the literature, it was decided that the 2nd aerobic tanks would not contribute to LoT nutrient removal and they were removed on 10th May, 2010. In addition, there were many modifications in the process recycle rate, the VFA addition to the anaerobic tank and the membrane tank sizing during the period between 18th February, 2010 and 10th May, 2010. Since these modifications were expected to impact steady state operation, the results for that period are not discussed here. The two systems were reset by stoppage of wasting on 11th May, 2010 and hence, became the de facto operating day 1 of this project. Wasting did not begin again until day 45. The intention was to operate a high MLSS steady state system, considered typical for MBR configurations.  4.2.4 Recycle rates and nutrient supplementation  Determination of recycle rate is a very important design consideration for MBNR system nutrient removal objectives. For better solids distribution in the reactors and reduced solids load on the membrane tank, anoxic (IR1), aerobic (IR2) and returned activated sludge (IR3) recycle rates were initially kept high, with values of 2Q, 3Q and 4Q respectively. On the other hand, higher recycle rates increase the probability of elevated nitrate concentrations in the anaerobic tank which could inhibit EBPR activity.    60  For that reason, nitrate was monitored in the anaerobic and pre-anoxic reactors with the results shown in Figure 4.2 and Figure 4.3 respectively.  Anoxic Recycle = 2QAerobic Recyle = 3QRAS = 4QDays0 100 200 300 400 500NO3-N (mg/L)01234567Anaerobic Reactor-MBNR (Biological)Anaerobic Reactor-MBNR (Chemical)Anoxic Recycle = QAerobic Recyle = 1.5QRAS = 2Q Figure 4.2 NO3-N profiles in anaerobic reactor of parallel MBNR systems     61  Anoxic Recycle = 2QAerobic Recycle = 3QRAS = 4QDays0 100 200 300 400 500NO3-N (mg/L)0246810121416Pre-anoxic Reactor-MBNR (Biological)Pre-anoxic Reactor-MBNR (Chemical) Anoxic Recycle = QAerobic Recycle = 1.5QRAS = 2Q Figure 4.3 NO3-N profiles in pre-anoxic reactors of parallel MBNR systems  The figures clearly demonstrate the presence of nitrate in both zones until day 72, which indicates that the systems were not optimized for EBPR activity. This led to the decrease in all the recycle rates by 50 percent on Day 73. The reduced recycle rates were maintained for the remainder of the studies. The modification resulted in very low and stable nitrate concentrations in both reactors thereafter.   The parallel MBNR systems were supplemented with methanol (external carbon) and acetate (VFA), vital nutrients for enhanced denitrification and biological phosphorus removal respectively. Post-anoxic reactor methanol dosing began on day 80 and the initial target concentration after addition was 24 mg/L. The selection of the initial methanol supplementation concentration was based on measured average permeate nitrate concentrations of the parallel MBNR systems during first 79 days of operation (pre-methanol dosing period). As described in Section 5.3.3., the average permeate nitrate concentrations for that period were 11.3 and 13 mg/L for the MBNR (Biological) and MBNR (Chemical) systems respectively. To begin with, an approximate methanol dosing    62  to permeate nitrate ratio of 2 was chosen and was increased in a step-wise manner to 3 (methanol dosing of 36 mg/L), 4 (methanol dosing of 48 mg/L) and finally 6 (methanol dosing of 72 mg/L).  Acetate supplementation was utilized from day 1 so that EBPR activity would not be affected by carbon-limited conditions. Initially, acetate was added to the mixed liquor directly to produce a nominal initial concentration of 80 mg COD /L in the anaerobic reactor of the parallel MBNR systems. Selection of acetate dosing concentration was done with the singular objective of promoting EBPR activity in a VFA-rich environment.  However, past pilot research work at UBC has demonstrated excellent EBPR with only 40 mg COD/L of acetate added to the anaerobic zone (Monti, 2006). Although the previous study was carried out with a modified UCT system, the wastewater characteristics were comparable to those of the present study. From day 73 onwards, acetate addition was reduced to 40 mg COD/L in both systems. Alum was the preferred metal salt for the MBNR system shown in Figure 4.1b. For this project, it was postulated that a functional EBPR process would be necessary to analyze the impact of alum in the MBNR (chemical) system. Accordingly, addition of alum did not start until day 226. The initial membrane tank alum dosage was 20 mg/L and was increased in a step-wise mode until effluent TP ~ 0.1 mg/L was achieved. Figure 4.4 and Table 4.5 illustrate the different phases and concentrations of the MBNR system operation in terms of acetate, methanol and alum supplementation respectively. Nitrogen and phosphorus removal performance of the parallel MBNR systems will be discussed with respect to each phase of operation.    63  Days0 100 200 300 400 500Concentration (mg/L)020406080100Acetate (mg COD/L)MethanolAlum for MBNR (Chemical) I II IIIIVV Figure 4.4 Different phases of MBNR operation w.r.t. acetate (anaerobic reactor), methanol (post-anoxic reactor) and alum (membrane tank) dosing  Table 4.5 Dosing set-points for acetate, methanol and alum supplementation during different phases of MBNR operation Parameter Days Recycle Flow Acetate (mg/L) Methanol (mg/L) Alum (mg/L) in MBNR (Chemical) Phase I 1-73 Anoxic = 2Q Aerobic = 3Q RAS = 4Q 80 0 0 Phase II 74-225 Anoxic = Q Aerobic = 1.5Q RAS = 2Q 40 24-72 0 Phase III 226-276 Anoxic = Q Aerobic = 1.5Q RAS = 2Q 40 72 20 Phase IV 277-345 Anoxic = Q Aerobic = 1.5Q RAS = 2Q 40 72 40 Phase V 346-480 Anoxic = Q Aerobic = 1.5Q RAS = 2Q 40 72 80      64  4.3 Off-line Batch Tests  The relationship between enhanced biological phosphorus removal and chemical phosphorus removal was also investigated by conducting off-line batch tests in the laboratory. Mixed liquor from the aeration tanks of the parallel MBNR systems was subjected to sequential anaerobic-aerobic conditions during the tests. The first tests were conducted in Phase II of the program when the MBNR (Chemical) system was not supplemented with alum. Further tests were conducted in Phase III, IV and V to evaluate EBPR under a step-wise increase in alum addition. Three batch tests were conducted in each phase to document the EBPR activity with time. The batch test set-up is shown in Figure 4.5 and the methodology is described below.  4.3.1 Methodology  The batch test set-up was comprised of a 1.0 L Erlenmeyer flask, magnetic stirring bar, a rubber stopper seal with septa, pH probe, sampling tube and nitrogen-filled balloon. A nitrogen-filled balloon was used to compensate for the volumes of sampled liquid and to prevent oxygen intrusion to maintain anaerobic conditions (Comeau, 1989). The test temperature and pH were constant at 20 0C and 7.0 respectively. Depending on the measured mixed liquor pH value, either 0.1 N HCl or 0.1 N NaOH was added to maintain the set-point pH. A VWR portable probe was used for mixed liquor pH and temperature measurement. An initial non-aerated period of 2-4 hours was maintained for endogenous denitrification (with mixing). The duration of the non-aerated period was dependent on the initial nitrate concentration in the mixed liquor.  The objective of this procedure was to prevent any denitrification-related carbon consumption during the tests themselves. After the completion of endogenous denitrification, N2 was introduced to the batch reactor to rapidly establish anaerobic conditions.   The anaerobic period was maintained for 2 hours. Acetate was added to the mixed liquor to produce a nominal initial concentration of 100 mg COD /L for optimal EBPR performance. The first sample was collected 1-2 minutes after substrate addition for    65  measuring acetate, NO3-N, Mg2+, K+ and PO4-P concentrations. Further samples were taken every 30 minutes. Following the anaerobic period, an aerobic phase was imposed of 3 hours duration. The N2 gas was replaced by air sparging and the DO concentration was maintained between 2.0-3.0 mg/L during the aerobic period. Sampling parameters and frequency were similar to those of the anaerobic period.      Figure 4.5 Batch reactor schematic  4.3.2 Chemical Analysis  The batch tests required determination of NO3-N, Mg2+, K+ and PO4-P concentrations in the liquid phase of the mixed liquor. The procedure for NO3-N and PO4-P analysis has been described in Table 4.4. Mg2+ and K+ samples were prepared and analyzed in general accordance with 3120 B. Inductively Coupled Plasma (ICP) Method (APHA et al., 2005).    66  A Perkin Elmer Optima 7300 DV Optimal Emission Spectrometer was used for final analysis.  4.4 MBNR System Modeling and Simulation  4.4.1 General introduction  The BioWinTM (version 3.1) simulator was used for MBNR process modeling. The simulator is based on a combined Activated Sludge/Anaerobic Digestion (ASDM) model which the developers refer to the as the BioWin General Model (EnviroSim Associates Ltd, Hamilton, Ontario, Canada). In total, the General Model has fifty state variables and sixty process expressions. The activated sludge component is primarily based on ASM1 for nitrification and denitrification and on work by Wentzel et al. (1989a and b) for biological phosphorus removal. The General Model also incorporates a fermentation process which converts readily biodegradable COD to short chain fatty acids (assuming a loss of COD from the system), hydrolysis of enmeshed slowly biodegradable COD under anoxic and anaerobic conditions and anoxic growth of PAO organisms. The rationale for the modifications and their values can be found in the article written by Barker and Dold (1997). The BioWinTM simulator also enables the user to model a process using the default ASM series models. The other important modeling options that have been integrated into the main simulator are:  ? Oxygen modeling  ? pH modeling (includes the option of pH limitation on sludge kinetic equations)  ? Chemical precipitation modeling for struvite, hydroxy-dicalcium-phosphate (HDP) and hydroxy-apatite (HAP) ? Chemical phosphorus precipitation modeling (options include alum or ferric)  ? Settler modeling (options include modified Vesilind or double exponential)   BioWinTM can be used for both steady state and dynamic modeling of activated sludge systems. In steady state modeling, equilibrium relationships between model variables are independent of time. On the other hand, model variables are described by    67  differential equations for dynamic modeling. Steady state modeling can provide vital information on design and capacity evaluation of wastewater treatment plants. Dynamic modeling is typically employed to address issues related to daily or seasonal variations in a wastewater treatment plant, for instance changing flow rate, influent concentrations, internal pumping rates or aeration patterns varying over time.  4.4.2 MBNR system configuration in BioWinTM  The first step in the simulation involved setting up the parallel MBNR systems in the BioWinTM interface. This was done by inserting dimensions for each bioreactor in the simulator. Then the bioreactors were connected with pipes and splitters and provided with accurate flow rates. The General Model was selected in both MBNR systems for simulation of suspended/ volatile solids production, nitrogen removal and EBPR. The Chemical Phosphorus Precipitation Model was the addendum for simulation of alum dosing in the MBNR (Chemical) system. Figure 4.6 shows the configurations of the MBNR systems set up in the BioWinTM interface along with separate element units for influent, permeate, waste activated sludge (WAS), acetate, methanol and alum.     68   a. MBNR (Biological) system   b. MBNR (Chemical) system Figure 4.6 Parallel MBNR system configuration in BioWinTM  Anaerobic PermeateInfluent AerobicPre-Anoxic Post-anoxic Membrane WASMethanolAcetateAnaerobic PermeateInfluent Aerobic Pre-Anoxic Post-anoxic Membrane WASMethanolAcetateAlum   69  4.4.3 Simulation strategy  Different simulation strategies have been recommended in the literature for activated sludge system modeling (Cosenza et al., 2013; Melcer et al., 2003; Vanrolleghem et al., 2003; Langergraber et al., 2004). In the current project, the ?Biomath-Calibration? protocol (Vanrolleghem et al., 2003) was broadly adopted for simulation of the parallel MBNR systems. The four major steps of the ?Biomath-Calibration? are: (1) definition of objectives, (2) comprehensive data collection and analysis from an activated sludge system, (3) steady state calibration and (4) dynamic calibration and evaluation of results. In the current project, definition of objectives and evaluation of simulation results can be found in Chapter Three and Chapter Seven respectively. This section describes methodologies used for data collection, sensitivity analysis, calibration and validation.       4.4.3.1 Data collection, analysis and steady state simulation (with default parameters)  Steady state simulation was initially carried out for the MBNR (Biological) system with default BioWinTM kinetic and stoichiometric parameters. Since the MBNR (Biological) system was a reference for evaluating the influence of chemical phosphorus removal on EBPR, it was the default choice for steady state simulation and calibration. The assumption was that a calibrated MBNR (Biological) system would host the same EBPR mechanism as the MBNR (Chemical) system before the addition of alum.   The influent data collection period for the modeling task was from operating days 236 to 356. During that period, influent municipal wastewater was collected once per week from the Staging Environmental Research Centre (SERC), South Campus, University of British Columbia to maintain a relatively steady input to the model. The selection of input parameters was driven by the requirements of the BioWinTM influent specifier element and their average values are summarized in Table 4.6.        70  Table 4.6 Input data for BioWinTM influent specifier Parameter Number of Samples (n) Concentration (Min.-Max.) CODTotal (mg/L) 50 302 (?65) (189-483) CODSol (mg/L) 50 122 (?34) (67-205) CODFF (mg/L) 50 84 (?24) (45-139) BODTotal (mg/L) 20 95 (?17.7) (64-129) BODSol (mg/L) 20 48 (?12) (23-70) TKN (mg/L) 35 43 (?7.4) (31.5-57) NH4 -N (mg/L) 50 31.2 (?5.9) (20.1-44.9) TP (mg/L) 35 4.9 (?7.4) PO4-P (mg/L) 50 3 (?0.7) (1.8-5.1) VFA (mg/L) 50 19.3 (?7.7) (2.1-36.4) Alkalinity (mg CaCO3/L) 23 147 (?10.4) (128-165) TSSRaw (mg/L) 16 103 (?37.5) (40-160) pH  67 7.6 (?0.4) (7.1- 8.6) Ca2+ (mg/L) 18 19.2 (?6.8) (12- 34) Mg2+ (mg/L) 18 2.1 (?0.6) (1.1- 3.2)       ?: Standard Deviation                                                                                                                       Data Period: Operating days 236 to 356  The sampling schedule and analytical procedures for the above parameters (except for alkalinity, BOD5 and Ca2+) were presented in Table 4.2 and Table 4.4 respectively. Sampling for alkalinity, BOD5 and Ca2+ tests was done on Monday and Friday of each week during the data period. Alkalinity and BOD5 were subsequently determined by using the 2320 B. Titration Method and 5210 B. 5-Day BOD Test method respectively (APHA et al., 2005). The BOD samples were analyzed using an YS1 52 Dissolved Oxygen Meter. Ca2+ samples were prepared and analyzed in general accordance with 3120 B. Inductively Coupled Plasma (ICP) Method (APHA et al., 2005). A Perkin Elmer Optima 7300 DV Optimal Emission Spectrometer was used for final analysis.   4.4.3.2 Sensitivity analysis  Preliminary steady state simulation of the MBNR (Biological) system demonstrated (results in Chapter Seven) discrepancy in suspended solids concentration and EBPR mechanism when compared to measured data from the real system. A sensitivity analysis was therefore completed to understand the importance of specific kinetic and stoichiometric parameters on suspended solids concentration and EBPR. In modeling    71  work, the main objective of a sensitivity analysis is to assess the influence of specific parameters on the model outputs. A kinetic or stoichiometric model parameter is classified as highly sensitive if a small change in its value can cause a large change in model prediction (Liwarska-Bizukojc and Biernacki, 2010). On the contrary, a low sensitivity parameter can be varied significantly without much impact on model output. In the current project, the selection of modeling parameters for sensitivity analysis was based on information available in current MBNR modeling literature. Modeling parameter selection rationale and results of sensitivity analysis is described in detail in Chapter Seven.   4.4.3.3 Calibration of model with steady state simulation  Once the most important parameters were identified in sensitivity analysis, their values were adjusted to calibrate a steady state model against measured data. The first step of calibration required the identification of kinetic and stoichiometric parameters that significantly influenced suspended solids concentration in the simulated system. The calibration of suspended solids concentration is typically the first step, which is then followed by calibration of nitrogen removal (nitrification and denitrification) and EBPR in modeling of activated sludge systems (Brdjanovic et al., 2000; Meijer et al., 2001; Hulsbeek et al., 2002; Petersen et al., 2002). Since nitrogen removal model predictions matched very well with MBNR (Biological) system measured data with default nitrification and denitrification kinetic and stoichiometric parameter values, the second step was calibration of kinetic and stoichiometric parameters relevant to the EBPR mechanism.  Detailed data on the outcome of steady state simulation (with calibrated model parameters) can be found in Chapter Seven.  4.4.3.4 Validation of model by dynamic simulation of MBNR (Biological) system  After successful calibration of the model under steady state conditions, a dynamic simulation of MBNR (Biological) system was conducted for operating days 359 to 449. The motivation behind this exercise was to evaluate predictive capability under    72  moderately dynamic influent conditions. Detailed data on the outcome of dynamic simulation of the MBNR (Biological) system can be found in Chapter Seven.  4.4.3.5 Dynamic modeling of MBNR (Chemical) system  Finally, dynamic simulation of the MBNR (Chemical) system was conducted to investigate the impact of step wise increments of alum concentrations on EBPR activity. This exercise also included a comparative analysis of suspended solids production, nitrification and denitrification as predicted by the model relative to measured data. The simulation period was between operating days 226 to 470.  Discussion on the potential negative influence of alum on EBPR as predicted by the model versus MBNR (Chemical) system measured data is detailed in Chapter Seven.  4.5 Method Development for DON Measurement  Method development for reliable DON measurement was one of the main requirements for understanding effluent total nitrogen speciation. The two key needs of a method were the removal of nitrate to the background levels (by anion exchange resin) and conversion of residual DON to nitrate (by persulfate digestion). The adequacies of the above methods were investigated by quality control (QC) experiments. Final nitrate analysis did not pose a big challenge as it could be done by any of the methods described in Standard Methods. Once the methodology was fully developed, samples were collected from the reactors and permeate of parallel MBNR systems for DON profiling.   4.5.1 Batch anion exchange resin method  Removal of nitrate by an anion exchange resin method was put forward by Crumption et al. (1992) and their methodology is documented in Section 2.3.4.2. For the present project, a batch method was developed to remove residual nitrate from the effluent samples of the parallel MBNR systems. The batch method offered some key advantages, i.e. many samples could be processed simultaneously, very little expert training was    73  required for the experiments and it was much less expensive than the burette column method of Sattayatewa and Pagilla (2008). The step by step by procedure for the batch anion exchange resin test is described in the next paragraph.    The first step involved the determination of the weight of ion exchange resin required for different ranges of NOx-N concentrations. After experiments with standards of known nitrate concentrations and permeates from both MBNR systems, it was concluded that 0.75 g and 1.25 g of dry ion exchange resin could extract NOx-N in the concentration range of 0 ? 5 mg/L and 5 ? 10 mg/L respectively (detailed results is discussed in Chapter Eight). During the experiments, depending on the sample being analyzed, a fixed weight of dry ion exchange resin (Acros Organics Dowex 1X8 50-100 mesh; 3.2 meq/dry g total capacity) (0.75 or 1.25 g) was carefully poured into the bottom of a 50 mL clean centrifuge tube. In addition, 5 mL of distilled water was transferred to the centrifuge tube to ensure that resin sticking to the side of the tube was rinsed to the bottom of the tube for soaking. A tube cap was screwed on securely and the tube was stored upright in a rack overnight at room temperature. Mixing was not required for the stored aliquots of resin. These above steps were replicated with multiple centrifuge tubes for simultaneous anion exchange experiments. Subsequently, 10 mL of sample was filtered (0.45 ?m) and poured into a graduated cylinder. Process sampling was done in triplicates and multiple samples were prepared at the same time with designated graduated cylinders. The next step was acidification of the samples (target pH < 2), achieved by the addition of three drops of 3N HCl to each graduated cylinder. The samples were then poured into their dedicated resin-water-containing centrifuge tubes and the tubes were placed into a rotating mixer. The rotating mixer was operated at 10 RPM for 1 hour. After switching off the mixer, approximately 0.65 mL of 0.5 N NaOH was added to each sample for neutralization. The mixer was then switched on again for a minute to mix and neutralize the samples. Aliquots of 5 mL of each neutralized supernatant samples (without any resin) were carefully poured into 10 mL test tubes or sample bottles. Finally, the samples were stored in refrigerator at 4 0C for processing with the persulfate digestion method. Figure 4.7 illustrates a typical set-up with centrifuge tubes (with soaked resin and samples) in a rotating mixer.    74      Figure 4.7 Rotating mixer with centrifuge tubes  The anion exchange resin quality control (QC) tests were completed with standards of known nitrate concentration for analyzing removal consistency and with ammonium and urea (representing DON) for recovery consistency. It was hypothesized that no ammonium or urea would be removed during the nitrate removal phase of the DON analysis. The test results are discussed in Chapter Eight.  4.5.2 Persulfate digestion method  The persulfate digestion method was used to convert all non-adsorbed forms of nitrogen to nitrate. In the present project, the digestion utilized the 4500-N C. Persulfate Method (APHA et al., 2005). The method required alkaline oxidation of samples at temperatures in the range from 100 to 110 0C. Temperatures in this range accelerate persulfate (K2S2O8) auto decomposition, thus generating the O2 needed for the oxidation of N.    Persulfate digestion QC tests were undertaken to study the conversion consistency of standards with known concentrations of ammonium, urea and glutamic acid to nitrate. In addition, recovery tests were conducted on standards containing known concentrations of nitrate. The results are discussed in the Chapter Eight.     75  4.5.3 Nitrate analysis  Nitrate was determined by using 4500-NO3-F. Automated Cadmium Reduction Method (APHA et al., 2005) and the sampling procedure is detailed in Table 4.4.   4.5.4 Sampling for DON profiling  Sampling for DON was completed during Phase V of the experimental program. For permeate, sampling began on operating day 403 and was finished on operating day 464. The permeate DON measurement was expected to provide information on the range of concentrations in the parallel MBNR systems and the impact of up to 80 mg/L of alum addition on DON removal. Reactor DON sampling was conducted between operating days 405 and 457. Sample volume was typically 50 mL for both permeate and reactor samples. The only difference was that filtration (0.45 ?m) was required for reactor samples.       76  5 Performance of Parallel MBNR Systems Targeting LoT Nutrient Removal   5.1 Introduction  Two parallel modified Bardenpho MBNR systems were operated with the goal of accomplishing permeate TN ? 3 mg/L and TP ? 0.1 mg/L. Alum dosing for enhanced phosphorus removal was the only difference between the two systems. During continuous operation over a period of 478 days, the parallel MBNR systems were evaluated for COD removal, nitrification, denitrification, EBPR and chemical phosphorus removal performance. These data were then used to assess the relationship between the simultaneously occurring EBPR and chemical P removal mechanisms in the MBNR (Chemical) system. An important objective of the continuous flow system operation was to determine the significance of external dosing of alum and methanol in the realization of LoT phosphorus and nitrogen targets, respectively. This chapter provides an estimate of their stoichiometric requirements.            5.2 Influent Wastewater Characterization  The parallel laboratory scale MBNR systems were operated for approximately sixteen months with municipal wastewater from the SERC at the University of British Columbia. The average influent wastewater characteristics for that period are summarized in Table 5.1.           77  Table 5.1 Influent wastewater characteristics Parameter Concentration (Min.-Max.) CODTotal (mg/L) 290 (?85) (125-584) CODSol (mg/L) 117 (?37) (40-231) CODFF (mg/L) 89 (?32) (28-217) TKN (mg/L) 39 (?10.2) (15-69) NH4 -N (mg/L) 33 (?6.7) (19-57) TP (mg/L) 4.5 (?1.2) (1.6-9.4) PO4-P (mg/L) 3.4 (?1.0) (1.5-7.4) VFA (mg/L) 26 (?18.7) (0-144) TSSRaw (mg/L) 140 (?76) (10-380) pH  7.4 (?0.4) (6.1- 9.0)       ? : Standard Deviation      Data collected between May 10, 2010 and August 31, 2011.   5.3 Process Performance  5.3.1 COD profiling  Measurement of influent and permeate COD was undertaken to investigate the carbon removal efficiency in the parallel MBNR systems and the stoichiometric suitability of the wastewater for enhanced denitrification and biological phosphorus removal. Influent characterization was conducted for total COD, soluble COD and CODFF concentrations and the results are shown in Figure 5.1. The average values for the study period along with standard deviations are presented in Table 5.1. The influent total COD concentration was observed to be variable with the highest average value of 316 mg/L occurring during Phase II and the lowest average value of 235 mg/L in Phase V. From Figure 5.1, it can also be seen that CODSol and CODFF data are not available for the period between day 100 and 200. This occurred as a result of a systematic error in which filter paper with a pore size of 1.2 ?m was used instead of 0.45 ?m during this period.         78  Days0 100 200 300 400 500COD (mg/L)0100200300400500600700Raw InfluentSoluble  Raw InfluentFF Raw Influent I II III IV V Figure 5.1 Measured raw influent COD concentrations  Influent particulate COD and readily biodegradable COD (CODRb) concentrations were estimated in the study and their values are shown in Figure 5.2.  Particulate COD was determined by subtracting soluble COD from total COD values. The particulate COD value ranged between 30 mg/L and 288 mg/L, with an average concentration of 164 mg/L during the study period. Depending on biodegradability, the wide range for particulate COD was expected to significantly impact MBNR system suspended solids concentrations. The CODRb measurement was based on the theory that the influent soluble non-biodegradable COD fraction passes through the MBNR systems without being produced or utilized in the system. Therefore, the difference between CODFF and permeate total COD was, in principle, the readily biodegradable fraction. The CODRb values (Figure 5.2) were calculated by using permeate total COD values from the MBNR (Biological) system. The CODRb values ranged between 13 mg/L to 195 mg/L, with an    79  average concentration of 64 mg/L during the study period. CODRb is a crucial influent wastewater component for achieving successful biological phosphorus removal.  Days0 100 200 300 400 500COD (mg/L)0100200300400500600700 Particulate Raw Influent  Readily Biodegradable Raw Influent I II III IV V Figure 5.2 Estimated raw influent COD concentrations  COD removal in the parallel MBNR systems is illustrated in Figure 5.3. Treatment performance was very similar in both systems, with average removal efficiencies of 91 percent calculated for both systems. Average permeate total COD concentrations of 24 and 25 mg/L were observed for the MBNR (Biological) and MBNR (Chemical) systems, respectively. In addition, permeate total COD concentration was steady in all five phases of the operation in the parallel systems. High COD removal efficiencies were expected, as the MBNR systems were operated with relatively long HRTs and SRTs under ambient room temperature conditions.      Some studies have demonstrated superior and consistent COD removal with the addition of alum in MBR systems (Fleischer et al., 2005; Holbrook et al., 2004; Lee et al., 2001).  Lee et al. (2001) hypothesized that coagulation reduces the concentration of soluble organics in alum-supplemented sludge and as a result, lower COD concentration    80  is observed in the permeate. Holbrook et al. (2004) found that protein and polysaccharide concentrations were lower during a period of alum addition when compared to a period without alum addition. However, no enhanced COD removal was observed in the MBNR (Chemical) system (Figure 5.4), even with the highest alum doses applied. This observation could be attributed to the role of alum in the aggregation of organic particles of specific particle sizes and the pore size of the membrane filter. Holbrook et al. (2004) observed in their study that alum-induced aggregation rates for larger particle sizes (15 and 7.5 ?m) were higher than those for 3.5 ?m particles. Fan et al. (2007) reported that different alum dosages had greater coagulation impact on particle sizes > 10 ?m in their lab scale submerged MBR system. Since alum does not coagulate particles smaller than 0.04 ?m (nominal pore size of ZW1 membrane), therefore, in the present study, it is not surprising that reduced permeate COD was not observed for the MBNR (Chemical) system. This is in agreement with the data presented in Figure 5.4.   MBNR (Biological)Days0 100 200 300 400 500COD (mg/L)020040060080010001200Percentage (%) 020406080100120Raw Influent Permeate COD RemovalI II III IV V MBNR (Chemical)Days0 100 200 300 400 500COD (mg/L)020040060080010001200Percentage (%) 020406080100120Raw Influent Permeate COD RemovalI II III IV V Figure 5.3 COD concentrations and removal efficiencies in parallel MBNR systems    81  Days250 300 350 400 450 500COD (mg/L)020406080100Concentration (mg/L)020406080100Pemeate MBNR (Biological)Permeate MBNR (Chemical)AlumIII IV V Figure 5.4 Permeate COD concentrations with and without alum addition  5.3.2 Nitrification  The parallel MBNR systems nitrification performance was assessed by measuring influent and permeate NH4-N concentrations and the results are presented in Figure 5.5. From the figure, it can be concluded that NH4-N removal efficiency was close to 100 percent in all five phases of operation for both systems. The MBNR (Biological) system permeate NH4-N concentration was higher than 1 mg/L only once, on day 363. This datum was not an outlier because reactor scan data (Figure 5.6) for that day also showed incomplete nitrification in the aerobic tank. The two key parameters, i.e. process temperature and SRT were constant throughout the study period and cannot be the reasons for failure. Previous research work in SERC has reported low alkalinity in the influent stream and sodium bicarbonate supplementation was a requirement for successful nitrification (Monti, 2006). However, sodium bicarbonate was not added in the current project as enhanced methylotrophic denitrification can theoretically recover any loss in alkalinity due to nitrification.      82  MBNR (Chemical)Days0 100 200 300 400 500NH4-N (mg/L)0102030405060Raw InfluentPermeate I II III IV V Figure 5.5 Influent and effluent NH4-N concentrations in parallel MBNR systems  Although NH4-N removal efficiencies were similarly high in the MBNR (Chemical) system, unusual permeate NH4-N concentrations were observed between day 354 and day 380, at the beginning of Phase V of the operation (Figure 5.5). As described above, low influent alkalinity could have contributed to incomplete nitrification during that period. However, the extended period of failure, as compared to the parallel MBNR (Biological) system, required focus on the potential impact of alum addition on nitrification. Alum dosage was increased from 40 mg/L to 80 mg/L on day 346 of the operation. The role of alum in nitrification inhibition has been reported in the literature (Lee et al., 2001; Liu et al., 2011). Also, it is known that 1 mg/L of alum consumes 0.5 mg/L of alkalinity (as CaCO3) in water (Fleischer et al., 2005). Therefore, a combination of low influent alkalinity and increased alum addition might have inhibited nitrification efficiency of the MBNR (Chemical) system between day 354 and day 380 of the operation. Nonetheless, swift recovery and consistent performance was observed thereafter as far as nitrification was concerned.   NH4-N was also measured in the individual reactors of the two MBNR systems to track removal performance and to improve understanding of individual reactor performance. The MBNR (Biological) and MBNR (Chemical) NH4-N scan results are MBNR (Biological)Days0 100 200 300 400 500NH4-N (mg/L)0102030405060Raw InfluentPermeate I II III IV V   83  summarized in Figure 5.6 and Figure 5.7, respectively. The profiles are very similar for the entire study period with the only exception being very small periods of failure in Phase V of the operation. As expected, the internal recycles caused dilution of NH4-N in the anaerobic and pre-anoxic reactors of the parallel MBNR systems. The scan data also illustrate that the majority of the NH4-N was removed in the aerobic reactors. The post-anoxic reactor NH4-N profiles were very similar to those of the aerobic reactor. There were signs of nitrification in the membrane tank, as residual NH4-N from the aerobic reactor was reduced to below the detection limit. One of the key advantages of the reactor scan was validating whether higher permeate NH4-N concentration was a result of sampling and analysis error, or due to reactor performance. In the MBNR (Biological) system, the breakthrough on day 363 can be attributed to incomplete nitrification in the aerobic reactor (Figure 5.6). Similarly, Figure 5.7 demonstrates the presence of elevated concentrations of NH4-N in the aerobic reactor of the MBNR (Chemical) system between operation day 354 and day 380.     84  MBNR (Biological)Days0 100 200 300 400 500NH4-N (mg/L)01020304050AnaerobicPre-anoxicAerobicPost-anoxicMembranePermeateI II III IVV Figure 5.6 NH4-N reactor scan data for MBNR (Biological) system    85  MBNR (Chemical)Days0 100 200 300 400 500NH4-N (mg/L)01020304050AnaerobicPre-anoxicAerobic Post-anoxicMembranePermeateI II III IVV Figure 5.7 NH4-N reactor scan data for MBNR (Chemical) system  5.3.3 Denitrification  Dentrification performance of the parallel MBNR systems is shown in Figure 5.8. Influent nitrate concentration was close to zero, which is the case with most municipal wastewaters. The effluent nitrate profile was always dependent on the extent of methanol addition. In Phase I, i.e. without methanol addition, the average permeate NO3-N concentration was 11.3 and 13 mg/L for the MBNR (Biological) and MBNR (Chemical) systems respectively. The average influent total COD/TKN ratio was 12 for the present study, which was higher than the recommended ratios of 6 (Barnard, 1988) or 8.6 (Ekama et al., 1984) for excellent denitrification performance. It could therefore be said that denitrification performance was not optimized in the two MBNR systems. Two possible reasons can be offered for the sub-optimum performance. First, the aerobic recycle was set at a relatively low rate for creating favorable conditions for PAOs in the upstream reactors. This decision might have caused underloading of the pre-anoxic reactor.    86  Second, fermentation of influent CODRb to VFA and the subsequent use by PAOs in the anaerobic zone most probably reduced the available COD in the pre-anoxic zone.  MBNR (Biological)Days0 100 200 300 400 500NO3-N (mg/L)0510152025Methanol (mg/L)20406080100Raw Influent PermeateMethanolI II III IV V MBNR (Chemical)Days0 100 200 300 400 500NO3-N (mg/L)0510152025Methanol (mg/L)20406080100Raw Influent PermeateMethanolIII III IV V Figure 5.8 Influent and effluent NO3-N concentrations in parallel MBNR systems  Methanol supplementation improved denitrification performance considerably in the parallel MBNR systems (Figure 5.8).  To obtain a better understanding of the impact of different dosages of methanol addition, average permeate NO3-N concentrations along with minimum, maximum and standard deviation values were calculated and these are summarized in Table 5.2. The table demonstrates the impact of increased methanol addition on the reduction of residual nitrate in the permeate to very low concentrations. With 72 mg/L of methanol supplementation, average permeate NO3-N concentrations of 2.0 mg/L and 1.4 mg/L were observed in the MBNR (Biological) and MBNR (Chemical) systems, respectively. In fact, permeate NO3-N concentration was below the detection limit a number of times in both systems. The data in Figure 5.8 demonstrate the capability of both MBNR systems in removing nitrate to extremely low levels in the effluents. Another observation from Figure 5.8 is the difference in variability of effluent NO3-N concentration in the parallel systems. NO3-N concentration was consistently below 5 mg/L in the MBNR (Chemical) system permeate during operation with  72 mg/L of methanol supplementation (Figure 5.8). On the other hand, the MBNR (Biological) permeate NO3-N concentration was variable and was particularly high at some times    87  during Phase V of operation (Figure 5.8). It is important to understand why the system denitrification was not optimal, even with very high methanol supplementation. Although methanol was added to the post-anoxic reactor of both processes, foam was observed only in the MBNR (Biological) system. The problem was addressed by opening the reactor on a daily basis and mixing the foam into the mixed liquor vigorously. However, this task could not be maintained consistently during the periods of failure. Since methanol was added from the top of the post-anoxic reactor, it is thought that foam prevented the effective mixing of methanol with the mixed liquor in the post-anoxic reactor. As a result, nitrate concentration was high in the permeate. Data for the subsequent days indicated that methanol-induced denitrification recovered quickly and was optimal with the resumption of foam mixing.     Table 5.2 Average permeate NO3-N concentration for different methanol dosages Methanol Concentration (mg/L) Sample Size MBNR (Biological)  MBNR (Chemical) Average (mg NO3-N /L)   (Min.-Max.) mg methanol/ mg NO3-N Average (mg NO3-N /L)  (Min.-Max.) mg methanol/ mg NO3-N 0 34 11.3 (?2.4)  (6.9-17.3)  13.0 (?3.8)  (7.4-26.5)  24 12 8.9 (?1.9)  (5.6-11.3) 9.9 12.0 (?2.3)  (8.4-11.3) 23.7 36 18 5.2 (?2.3)  (1.5-8.8) 5.8 6.7 (?2.8)  (1.3-10.5) 5.6 48 9 5.7 (?3.2)  (0.0-10.2) 8.6 6.2 (?4.3)  (0.1-12.3) 7.1 72 119 2.0 (?2.5)  (0.0-11.8) 7.6 1.4 (?1.5)  (0.0-5.7) 6.1 ? : Standard Deviation  The efficiency of post-anoxic denitrification was evaluated by estimating the mg of methanol added/ mg NO3-N removed ratio in the present study (Table 5.2). For this calculation, it was assumed that the difference between average permeate NO3-N concentration for the period without methanol addition and the average permeate NO3-N concentration for a specific methanol addition rate was solely achieved by methylotrophic denitrifiers. The ratios are illustrated in Table 5.2. The values in the table were higher than the range of 3 to 3.5 suggested by McCarty et al. (1969) in their work. Philips et al. (2010) suggested that inefficient denitrification can occur due to poor contact between bacteria, substrate and nitrate and air entrapment in the floc of the anoxic reactor. For the    88  present project, the main objective of the mixer in the post-anoxic reactor was to keep the solids in suspension. It is unknown whether the mixing was ideal for methanol-induced denitrification. Also, since the post-anoxic reactor was placed after the aerobic reactor, it received mixed liquor with elevated dissolved oxygen from the membrane tank. Therefore, the presence of oxygen in the floc could have been a potential factor leading to a methanol dosage requirement.    The reactor NO3-N profiles for the parallel MBNR systems are shown in Figure 5.9 and Figure 5.10.  Although the anaerobic NO3-N concentration was variable during Phase I, very little nitrate was observed for rest of the study period. The observed low NO3-N concentration in the anaerobic reactor was crucial for EBPR viability. Similarly, the pre-anoxic NO3-N concentration was usually < 0.5 mg/L in both MBNR systems. Figure 5.9 also confirms that nitrification in the aerobic reactor caused high NO3-N concentrations. The profiles were almost identical in the parallel MBNR systems, with NO3-N concentrations in the range of 5-10 mg/L. It is essential to note that the nitrification-related absolute NO3-N concentrations in the aerobic reactor were diluted by flow (Q) from the pre-anoxic reactor and the recycle flow (2Q) from the membrane tank. A review of the data in Figure 5.9 and Figure 5.10 confirmed low NO3-N concentrations in the two reactors and hence, diluted aerobic NO3-N values.       89  MBNR (Biological)Days0 100 200 300 400 500NO3-N (mg/L)05101520AnaerobicPre-anoxic Aerobic I II III IVVMBNR (Chemical)Days0 100 200 300 400 500NO3-N (mg/L)05101520AnaerobicPre-anoxic Aerobic I II III IV V Figure 5.9 NO3-N data for anaerobic, pre-anoxic and aerobic reactors of the parallel MBNR systems  The post-anoxic NO3-N data in Figure 5.10 definitively illustrated the influence of methanol in enhanced denitrification. In fact, the post-anoxic NO3-N concentrations were close to zero in the two MBNR systems during periods of operation. On the other hand, the NO3-N profile in Figure 5.10 further confirmed the earlier observation of foam-related methylotrophic denitrification failure in the MBNR (Biological) system. NO3-N data for the membrane tank were very similar and sometimes, higher than in the post-anoxic reactor. The higher NO3-N value was not entirely unexpected and can be attributed to endogenous decay and nitrification activity in the membrane tank (WEF MOP 36, 2011).        90  MBNR (Biological)Days0 100 200 300 400 500NO3-N (mg/L)05101520Methanol (mg/L)20406080100Post-anoxicMembraneMethanolI II III IV V  MBNR (Chemical)Days0 100 200 300 400 500NO3-N (mg/L)05101520Methanol (mg/L)20406080100Post-anoxicMembraneMethanolI II III IV V Figure 5.10 NO3-N data for post-anoxic and membrane reactors of the parallel MBNR systems  5.3.4 Phosphorus removal  The phosphorus removal performance was investigated in the context of understanding the capabilities of the parallel MBNR systems and the relationship between simultaneous biological and chemical phosphorus removal mechanisms. As mentioned before, EBPR and simultaneous EBPR-chemical phosphorus removal mechanisms were fostered in the MBNR (Biological) and MBNR (Chemical) systems respectively.   5.3.4.1 MBNR (Biological) system  A phosphorus removal performance summary for the MBNR (Biological) system is shown in Figure 5.11. The average permeate PO4-P concentration and removal efficiency were 2 mg/L and 41 percent respectively, for the entire study period. Clearly, EBPR performance was variable in the different phases of operation in the MBNR (Biological) system. The performance was also reflected in PO4-P concentration profiling (Figure 5.12) and PO4-P release/uptake profiling (Figure 5.13) of each reactor. PO4-P release/uptake in Figure 5.13 was calculated by mass balance in each reactor. If the difference between inflow and outflow of soluble PO4-P was negative, phosphorus    91  release was occurring in the reactor. On the other hand, uptake was the prevalent mechanism if the difference between inflow and outflow soluble PO4-P was positive.    MBNR (Biological)Days0 100 200 300 400 500PO4-P (mg/L)012345678Raw Influent PermeateI II III IV VMBNR (Biological)Days0 100 200 300 400 500% Removal -100-50050100PO4-PI II III IV V Figure 5.11 PO4-P removal in the MBNR (Biological) system  MBNR (Biological)Days0 100 200 300 400 500PO4-P (mg/L)05101520AnaerobicPre-anoxic Aerobic I II IIIIV V MBNR (Biological)Days0 100 200 300 400 500PO4-P (mg/L)05101520Post-anoxicMembraneI II III IV V Figure 5.12 Reactor PO4-P profile in the MBNR (Biological) system     92  MBNR (Biological)Days0 100 200 300 400 500PO4-P Release/Uptake (mg/day)-30-20-100102030AnaerobicPre-anoxic Aerobic I II III IV V MBNR (Biological)Days0 100 200 300 400 500PO4-P Release/Uptake (mg/day)-30-20-100102030Post-anoxicMembraneI II III IV V Figure 5.13 Reactor PO4-P release (-)/uptake (+) profile in the MBNR (Biological) system  During the first 44 days of operation (no excess sludge wasting), both PO4-P removal and release in the anaerobic reactor improved gradually in the MBNR (Biological) system, as shown in Figure 5.12 and Figure 5.13. However, the PO4-P release/uptake profiles in the other reactors (Figure 5.13) showed that the classic EBPR mechanism was not established in the MBNR (Biological) system. In the aerobic reactor, unexpected PO4-P release was observed, while significant PO4-P uptake occurred in the pre-anoxic reactor and the membrane tank. Additionally, PO4-P release was observed in the post-anoxic reactor. Although uptake of some PO4-P was expected in the pre-anoxic reactor and the membrane tank, usually P uptake takes place in the aerobic reactor of all well-functioning EBPR processes. Interestingly, once sludge wasting was resumed on day 45, there was an immediate decline in PO4-P release in the anaerobic zone, as well as in the P uptake that occurred in the pre-anoxic reactor and the membrane tank. This was also reflected in elevated permeate PO4-P concentrations (Figure 5.12). It can be argued that the MBNR (Biological) system was not operating at steady state for the first 95 days due to the absence of sludge wasting for 44 days, the modification of the MBNR recycling rates (Refer Materials and Methods Chapter), a change in acetate supplementation, the augmentation of methanol addition and the presence of NO3-N in the anaerobic reactor. As a result, the initiation of EBPR mechanism was negatively impacted in the MBNR (Biological) system.     93  After approximately 2 SRTs, the PO4-P concentrations steadily increased in both the anaerobic and pre-anoxic reactors until operating day 140, at which time these concentrations decreased unexpectedly until day 221. Between operating days 95 and 140, the pre-anoxic PO4-P concentration was sometimes higher than that observed in the anaerobic reactor (Figure 5.12). Figure 5.13 furthermore confirms higher PO4-P release in the pre-anoxic reactor during that period. The absence of nitrate in the pre-anoxic reactor (Figure 5.9) and very high VFA in the raw influent and in the anaerobic zone (Figure 5.19) might have caused unusually high PO4-P concentrations in the pre-anoxic reactor. The most probable explanation is that the pre-anoxic zone behaved like an anaerobic zone in the absence of nitrate (underloading in pre-anoxic zone) and caused further phosphorus release in the presence of abundant VFAs. In addition to the unexpectedly high PO4-P release in the pre-anoxic reactor, both uptake and release were observed in the aerobic reactor between operating days 95 and 140 (Figure 5.13). On the other hand, PO4-P uptake was fairly consistent in the membrane tank, as illustrated in Figure 5.13. Depending on the PO4-P uptake performance in the membrane tank, a low permeate PO4-P concentration was achieved intermittently in the MBNR (Biological) system between operating days 95 and 140 (Figure 5.12).    Aerobic PO4-P release in the MBNR (Biological) system was a cause of concern and the reason for this was further investigated. Various researchers have observed aerobic PO4-P release in their experimental work and this was often attributed to the presence of acetate in the aerobic tank (Bradjanovic et al., 1998; Guisasola et al., 2004; Randall and Chaplin, 1997). Different theories have been offered regarding the impact of aerobic zone acetate on EBPR performance. While Bradjanovic et al. (1998) and Guisasola et al. (2004) suggested that ATP generation due to oxidative phosphorylation under aerobic conditions is supplemented by PO4-P release, Randall and Chaplin (1997) proposed that the presence of acetate under aerobic conditions promoted excessive growth of filamentous bacteria that resulted in washout of PAOs. In the current study, uncharacteristically high VFA concentrations observed in the influent between days 95 and 170 might have progressed to the aerobic reactor and caused P release (Figure 5.19). All three literature studies predicted failure of EBPR in such a scenario. That was exactly    94  what happened in the present study, as the anaerobic PO4-P release deteriorated and complete failure was observed between days 196 and 221 (Figure 5.13).     The variability  of influent VFA concentrations and the steady state modeling task required modification in the wastewater collection schedule from twice per week to once per week on day 193 of operation. The other known change in the process was an increment in methanol dosing to 72 mg/L on operating day 175. Thereafter, the system recovered and EBPR performance was steady for the rest of the study period. The PO4-P profiles in Figure 5.12 resembled those of a functioning EBPR from day 224 until the end of the study. The average effluent PO4-P concentration and removal efficiency were 1.7 mg/L and 46 percent, respectively, between day 224 until the end of operation. This was a marked improvement as compared to the EBPR performance in the first 223 days of operation (average effluent PO4-P concentration and removal were 2.48 mg/L and 36 percent respectively). More importantly, Figure 5.13 provides evidence of major PO4-P release and uptake taking place in the anaerobic and the aerobic reactors, respectively. Furthermore, consistent PO4-P uptake was observed in the pre-anoxic reactor while no uptake or release was observed in the post-anoxic reactor and membrane tank after day 224. This period included a phase, between days 415 and 443, during which the system effluent PO4-P concentration was less than 0.5 mg/L, indicating excellent EBPR. On average, 87 percent phosphorus removal was observed during that period. It is not clear why the MBNR (Biological) system EBPR performed so well during that period. Nonetheless, performance could not be maintained and the permeate phosphorus values increased to approximately 1 mg/L for rest of the study.   A review of the reactor and effluent PO4-P profiles illustrates that performance was most consistent from day 224 until the last day of operation. Moreover, there were no modifications in the process recycle rates, operation or methanol addition during that period. Hence, it can be concluded that the EBPR performance from day 224 until the last day of operation is the best that could be achieved in the MBNR (Biological) system. Evaluation of the MBNR (Biological) system phosphorus removal performance during that period and the potential factors associated with it was expected to provide valuable    95  information from the stand point of the classic EBPR mechanism. The important factors could have been either the influent wastewater characteristics or the process operating conditions.    One of the important parameters for successful EBPR operation is the availability of influent COD in adequate concentrations in the anaerobic zone. The average COD/TP ratio was 70 in the current project, more than the recommended value of 40 for reducing phosphorus to less than 1 mg/L in the effluent (Dr. James L. Barnard, pers. comm.). Additionally, the average CODRb/ TP ratio was 12.5 in the present study.  A CODRb/TP ratio in the range of 10 to 16 is typically sufficient (Dr. James L. Barnard, pers. comm.) to meet an effluent TP goal of 1 mg/L in wastewater treatment plants. Moreover, in the present study, to avoid large impacts from variation in influent wastewater characteristics, the MBNR (Biological) system was always supplemented with acetate, added directly to the anaerobic reactor. As mentioned before, acetate was the chosen external VFA for the initial 268 days of operation. However, continued inefficient EBPR performance prompted replacement of acetate by a mixture of acetate and propionate (4:1 ratio) in the parallel MBNR systems for the rest of the study period. This step was taken based on the finding that PAOs have competitive advantage over GAOs when supplemented with a solution of acetic and propionic acid (Dr. James L. Barnard, pers. comm.). However, the EBPR performance did not show any noticeable improvement as a result of the modification in external VFA composition.   The other two parameters that could have influenced EBPR activity were temperature and pH. Since the MBNR (Biological) system was operated at a constant temperature of 20 0C, no major influence was expected with respect to variation in EBPR performance. pH is another key parameter, as it has been found to impact anaerobic phosphorus release (Filipe et al., 2001a; Kuba et al., 1997a; Liu et al., 1996; Smolders et al., 1994a; WEF and ASCE, 2006) as well as aerobic uptake efficiencies (Filipe et al., 2001a). From the studies mentioned above, a pH of 7.0 to 8.0 is the recommended range for optimum anaerobic phosphorus release efficiency.  As far as P uptake by PAOs in the aerobic reactor is concerned, Filipe et al. (2001a) observed in batch studies that efficiency was    96  essentially the same at pH 7.0 and 7.5, but decreased greatly at pH 6.5. The pH profiles of the influent, anaerobic and aerobic reactors of MBNR (Biological) system are shown in Figure 5.14. pH was consistently above 7.0 in the anaerobic reactor except for the last two months of operation. It can, therefore, be assumed that the observed anaerobic pH did not have a significant effect on phosphorus release efficiency. On the other hand, the pH of the aerobic reactor varied from the lowest value of 5.8 to the highest value of 8.5 during the study (Figure 5.14). Influent wastewater variations and nitrification were the two major factors controlling pH in the aerobic reactor of the MBNR (Biological) system. The average pH was 7.1 between day 1 to day 224, while it was 6.6 for the rest of the study period. The lower aerobic pH coincided with the steadiest, yet non-optimal, biological phosphorus removal observed in the MBNR (Biological) system. Moreover, as shown in Figure 5.12, EBPR performance was mostly limited by the uptake capability of PAOs in the aerobic reactor. Hence, it could be concluded that low aerobic pH might be one of the major reasons for inefficient EBPR in the MBNR (Biological) system.   MBNR (Biological)Days0 100 200 300 400 500pH6789InfluentAnaerobicAerobicI IIIII IV V Figure 5.14 pH profile of influent, anaerobic and aerobic reactors of the MBNR (Biological) system    97   The presence of nitrate and oxygen in an EBPR anaerobic zone can severely inhibit EBPR activity in BNR systems. In the present study, the anaerobic reactor NO3-N concentration in the MBNR (Biological) system was below the detection limit except for the initial days of operation (refer Figure 5.9). Similarly, the DO concentration was always observed to be zero from intermittent DO measurements in the anaerobic tank (data not shown here). The impact of the above two parameters on EBPR performance was, therefore, concluded to be negligible.   Hydraulic retention time (HRT), which remained constant in the MBNR (Biological) system, has been studied extensively to evaluate its impact on EBPR efficiency.  Anaerobic HRT is particularly important, as adequate time is required for the formation of PHAs. Researchers have recommended that an anaerobic hydraulic residence time (HRT) of 0.25 to 1.0 hour to be sufficient to induce the target metabolisms (Barnard, 1984; Barnard and Fothergill, 1998; Grady et al., 1999). Coats et al. (2011) further noted that anaerobic HRT in the range of 1 to 3 hours would enrich the mixed liquor with PAOs and enable successful operation. In the context of present study, the set anaerobic HRT of 1.5 hours was in the recommended range and was not expected to inhibit phosphorus release. Aerobic HRT is generally set for achieving nitrification requirements. However, a very short HRT may impede phosphorus uptake efficiency in EBPR systems. Neethling et al. (2005) reported that an aerobic HRT to anaerobic HRT ratio between 3 and 4 was optimal for phosphorus removal. In the MBNR (Biological) system, aerobic HRT to anaerobic HRT ratio was maintained at 4 throughout the present study.      The MBNR (Biological) system was operated with a relatively long SRT of 25 days, generally not applied in conventional activated sludge systems. Long SRT has been reported to degrade effluent quality because endogenous biomass activity prompts secondary phosphorus release. Nonetheless, Lesjean et al. (2002) operated pilot MBR systems with an SRT of 26 days and accomplished 99.3 percent and 99.2 percent TP removal in pre-denitrification and post-denitrification modes respectively. Ersu et al. (2010) evaluated MBR system TP removal performance with different SRTs of 10, 25, 50    98  and 75 days. These authors reported the highest TP removal efficiency of 80 percent at a 50 day SRT and the lowest removal efficiency of 60 percent at a 75 day SRT.  The results from these studies suggest that low TP removal efficiency in the MBNR (Biological) system was unlikely caused by the applied SRT.       Internal recirculation rates in the MBNR (Biological) system did impact the mixed liquor suspended solids distribution, particularly in the upstream anaerobic reactor. In contrast to conventional activated sludge systems in which biomass is more evenly distributed throughout the system, the filtration barrier used in MBRs produces suspended solids accumulation in the downstream membrane tank. The specification of recirculation rates is, therefore, very important in MBR systems as far as maintaining upstream anaerobic reactor solids inventories and phosphorus release is concerned (Crawford et al., 2006; Erdal et al., 2010). In the present study, Figure 5.15 shows reactor suspended solids distribution profiles for the MBNR (Biological) system. The fractional suspended solids distributions were calculated as follows.  %	ft al. = [(?ff ? ?????t?	?t ???) ? ? (?ff	 ? ?????t?	?t ???)r????????? 	] ? 100     (9)                                   99  MBNR (Biological)Days0 100 200 300 400 500%  Solids 0102030405060AnaerobicPre-anoxicAerobicPost-anoxic Membrane TankI II III IVV Figure 5.15 Reactor suspended solids distribution in MBNR (Biological) system  From Figure 5.15, suspended solids distribution among the anaerobic, pre-anoxic, aerobic, post-anoxic and membrane zone was approximately 5 percent, 17 percent, 49 percent, 24 percent and 5 percent respectively. It is interesting to note that the anaerobic zone suspended solids fraction was very low relative to those of the other reactors. Though in-depth studies have not been conducted regarding the ideal anaerobic solids fraction requirement, data are available for a few full scale MBR plants with comparable process configurations. Daigger et al. (2009) reported that anaerobic solids fraction was 10 percent in the Traverse City modified VIP MBR plant and 13 percent in Broad Run modified Bardenpho MBR Water Reclamation Plant (WRP). Both the Traverse City and Broad Run plants primarily employed EBPR to meet effluent TP concentration of 0.5 mg/L and 0.1 mg/L respectively. In comparison, the anaerobic fraction of the total biomass (in other words, anaerobic SRT) was significantly lower in the present MBNR (Biological) system. Grady et al. (1999) reported that low anaerobic SRT may have an adverse affect on EBPR performance.  It is therefore postulated that a low anaerobic    100  suspended solids fraction played a key role in limiting the efficiency of EBPR in the MBNR (Biological) system. Nevertheless, detailed investigation is still needed to improve understanding of the relationship between anaerobic solids fraction and EBPR performance.       5.3.4.2 MBNR (Chemical) system  MBNR (Chemical) system phosphorus removal performance is summarized in Figure 5.16. This system was operated for biological P removal only, for the first 225 days of operation (Phase I and Phase II), after which time alum was incrementally supplemented for the rest of the study. During EBPR-only operation, the average effluent PO4-P concentration was 2.8 mg/L and removal efficiency was 27 percent. It can be argued that performance was non-optimal during that period (Figure 5.16). In fact, the MBNR (Chemical) system EBPR performance mimicked that of the MBNR (Biological) system during the first 225 days of operation (Figure 5.17 and Figure 5.18). The reasons for the poor EBPR efficiency were discussed above and the focus of this section will be on the period of operation with simultaneous biological and chemical phosphorus removal mechanisms.    MBNR (Chemical)Days0 100 200 300 400 500PO4-P (mg/L)012345678Alum (mg/L)020406080I II III IV VRaw Influent PermeateAlumMBNR (Chemical)Days0 100 200 300 400 500 % Removal-100-50050100Alum (mg/L)020406080PO4-PAlumI II III IV V Figure 5.16 PO4-P removal in the MBNR (Chemical) system     101  There were two clear objectives of alum addition in the MBNR (Chemical) system: (1) accomplishing effluent TP ? 0.1 mg/L and (2) analyzing the relationship of chemical P removal to the existing EBPR mechanism. Initially, a low dosage of alum was added to the membrane tank mixed liquor to realize a nominal initial concentration 20 mg/L, for 2 SRTs (Phase III). PO4-P removal performance immediately improved with an average effluent concentration of 0.80 mg/L and removal efficiency of 71 percent (Figure 5.16). Of course, the total observed P removal was attributable to combined EBPR and chemical phosphorus removal mechanisms. Although the MBNR system did not recover completely from the previous high influent VFA period (Figure 5.19), the existence of EBPR was clearly evident as phosphorus release occurred in the anaerobic reactor in Phase III of operation (Figure 5.18). Phosphorus release was also observed in the aerobic and post-anoxic reactors while uptake took place in the pre-anoxic reactor and the membrane tank. A key conclusion from Phase III was that alum supplementation offered value in stabilizing overall phosphorus removal while EBPR was not optimal in the MBNR system. The PO4-P profile in Phase IV, as demonstrated in Figure 5.17 and Figure 5.18, provided evidence that functional EBPR (anaerobic release and aerobic uptake) was starting to take place in the MBNR (Chemical) system. Therefore, 20 mg/L of alum supplementation did not seem to inhibit EBPR activity in the MBNR (Chemical) system. This observation was further corroborated by batch studies in which the mixed liquor EBPR potential of the MBNR (Biological) system was found to be comparable to that of the MBNR (Chemical) system. The results of the batch studies are discussed in Section 6.3.       102  MBNR (Chemical)Days0 100 200 300 400 500PO4-P (mg/L)05101520Alum (mg/L)020406080AnaerobicPre-anoxic Aerobic AlumI II III IV V   MBNR (Chemical)Days0 100 200 300 400 500PO4-P (mg/L)05101520Alum (mg/L)020406080Post-anoxicMembraneAlumI II III IV V Figure 5.17 Reactor PO4-P profile in the MBNR (Chemical) system  MBNR (Chemical)Days0 100 200 300 400 500PO4-P Release/Uptake (mg/day)-40-30-20-10010203040Alum (mg/L)102030405060708090Post-anoxic MembraneAlumI II III IV VMBNR (Chemical)Days0 100 200 300 400 500PO4-P Release/Uptake (mg/day)-40-30-20-10010203040Alum (mg/L)102030405060708090AnaerobicPre-anoxic Aerobic AlumI II III IV VFigure 5.18 Reactor PO4-P release (-)/uptake (+) profile in the MBNR (Chemical) system  Alum dosage concentration was increased to 40 mg/L on operating day 277 in the MBNR (Chemical) system and this dosage was maintained for 70 days (Phase IV). During that period, the average effluent phosphorus concentration and removal efficiency that could be attained were 0.42 mg/L and 89 percent respectively. At the same time, EBPR performance stabilized with consistent phosphorus release and uptake in the anaerobic and aerobic reactors respectively (Figure 5.18). Figure 5.18 further illustrates that phosphorus uptake occurred in the pre-anoxic reactor while all activities stopped in the post-anoxic reactor and the membrane tank. Interestingly, PO4-P concentration    103  gradually decreased in all the reactors except the anaerobic reactor during the period of 40 mg/L of alum supplementation (Figure 5.17). It is postulated that as alum reacted with phosphorus to form precipitated compounds, the soluble phosphorus concentration became progressively lower in the MBNR (Chemical) system. As far as reaction in the anaerobic reactor is concerned, it is unclear whether biological phosphorus release and alum-induced phosphorus uptake was occurring simultaneously. The average PO4-P concentration in the MBNR (Chemical) anaerobic reactor was 6 mg/L between operating days 277 and 346, whereas it was 6.45 mg/L in the MBNR (Biological) anaerobic reactor during the same period. The comparable data in both systems suggest that insignificant chemical phosphorus removal was taking place in the anaerobic reactor. It can be concluded that both alum and EBPR were responsible for the low PO4-P concentration in the MBNR (Chemical) system effluent in Phase IV of operation. Regarding inhibition of EBPR at the increased alum concentration of 40 mg/L, there was no evidence of it in the MBNR (Chemical) system reactor PO4-P scan profile (Figure 5.17), the PO4-P release (-)/uptake (+) profile (Figure 5.18) or the batch test results (Section 6.4) in Phase IV of operation. The conclusion that could be derived is that sufficient phosphorus was available in the MBNR (Chemical) system for both chemical phosphorus removal and EBPR without interference. As mentioned earlier, the EBPR performance was not efficient in the parallel MBNR systems. For that reason, more phosphorus was available for alum complexation. This scenario might have been different with a higher functioning EBPR system.  It is also important to mention that 20 mg/L and 40 mg/L of alum addition were continued only for a short period of time. In a hypothetical situation, if the MBNR (Chemical) system had been supplemented with either of the above dosages for longer periods, the impact on EBPR might have been different. This is because the alum inventory would have steadily increased in the system such that eventually, there would have been substantial free alum available for complex formation. If both PAOs and alum targeted the same soluble phosphorus, chemical phosphorus removal would have been expected to be the dominant removal mechanism due to faster reaction rates.   Alum dosage was doubled to 80 mg/L on day 347 and this level of addition was continued until the end of the study (Phase V). The average effluent PO4-P concentration    104  was 0.07 mg/L and removal efficiency was 98 percent during that period (Figure 5.16). In fact, there were periods in which effluent PO4-P concentration was below the detection limit (detection limit was 0.01 mg/L) for the MBNR (Chemical) system. It seems that the high alum dosage produced extremely low effluent PO4-P concentrations along with consistent removal, as demonstrated in Figure 5.16. On the other hand, the EBPR mechanism gradually declined until it was almost negligible in the MBNR (Chemical) system (Figure 5.17 and Figure 5.18). Figure 5.18 also shows that both anaerobic release and aerobic uptake mechanisms were affected. The inhibition of EBPR was most likely caused by competition for soluble phosphorus with alum. Without phosphorus, the aerobic growth of PAOs stopped and hence, their concentration slowly decreased in the MBNR (Chemical) system mixed liquor. Results from batch studies during Phase V confirmed the observations from the continuous flow system (Section 6.5). The overall phosphorus removal occurring in the MBNR (Chemical) system during the period of 80 mg/L dosing was initially attributable to simultaneous biological and chemical removal, and then subsequently to chemical phosphorus removal only for the rest of period.  MBNR (Biological)Days0 100 200 300 400 500VFA (mg COD/L)020406080100120140160Acetate (mg COD/L)020406080Raw Influent AnaerobicExternal Acetate AdditionI II III IV V     MBNR (Chemical)Days0 100 200 300 400 500VFA (mg COD/L)020406080100120140160Acetate (mg COD/L)020406080Raw Influent AnaerobicExternal Acetate AdditionI II III IV V Figure 5.19 Raw influent and anaerobic VFA profile of the parallel MBNR systems  Although chemical phosphorus removal helped to achieve the low observed effluent PO4-P concentrations, further analysis is required to understand the efficiency of treatment. Figure 5.20 summarizes the mole Al (added)/mole TP (removed) and mg alum    105  (added) /mg TP (removed) ratios observed in the MBNR (Chemical) system. The average molar Al/TP ratios were 0.6 (Phase III), 0.8 (Phase IV) and 1.9 (Phase V). The value of 1.9 is comparable to the data published by Daigger et al. (2009), who reported that an Al/TP molar dose between 1.0 and 2.5 at the Broad Run water reclamation plant (WRP) achieved less than 0.05 mg/L TP in the effluent. The Broad Run MBNR WRP was configured to produce effluent TN < 3 mg/L and TP < 0.1 mg/L. The plant design incorporated simultaneous biological and chemical phosphorus removal. Nonetheless, it is postulated that alum dosage could have been further optimized in the present MBNR (Chemical) system. In the present system, alum was directly added to the membrane tank. High shearing conditions in the tank most probably hindered flocculation of aluminum precipitates and microbial floc. As a result, more alum may have been added than was required to reduce effluent TP concentrations. In full scale plants, therefore, it has been recommended to add alum in the mixed liquor pipe between the bioreactor and membrane tank. In a recent development, Liu et al. (2011) added a small alum contact tank (10 minute HRT) to optimize chemical phosphorus removal in a pilot UCT MBR process.    Figure 5.20 also illustrates that specific phosphorus removal (mg alum added/mg TP removed) decreased with an increase in alum dosing. This observation has been documented in other studies (Szab? et al., 2008) and the rationale for it can be found elsewhere (Maurer and Boller, 1999). A high alum requirement would be a significant cost driver for wastewater treatment plants targeting LoT TP goals. For plants designed for simultaneous phosphorus removal mechanisms, an optimally functioning EBPR capability is, therefore, a crucial requirement. Once it is achieved in an MBNR system, the alum requirement would be minimized with no affect on final effluent TP concentrations.      106  MBNR (Chemical)Days250 300 350 400 450 500Molar Al (added)/TP (removed)0.00.51.01.52.02.53.0Alum (mg/L)020406080III IV V  MBNR (Chemical)Days250 300 350 400 450 500mg Alum (added)/mg TP (removed)0510152025303540Alum(mg/L)020406080III IV V Figure 5.20  Alum-induced phosphorus removal in MBNR (Chemical) system  The preceding discussion of the phosphorus removal performance of the MBNR (Chemical) system provides important insight into the relationship between the simultaneous biological and chemical phosphorus removal mechanisms. It seems that both mechanisms can co-exist if sufficient phosphorus is available in soluble form. Once there is competition for phosphorus, it is most likely that chemical phosphorus removal would become the dominant mechanism. As a result, EBPR would be inhibited and the system would evolve to become a chemical phosphorus removal process only. For successful operation of simultaneous phosphorus removal, the alum dosage requirement could be calculated by evaluating P removal performance of the designed EBPR system with respect to effluent discharge limits. The case of LoT phosphorus removal is very important. It is highly unlikely that engineers would design a standalone EBPR system when the objective is ? 1 mg/L TP in the effluent. They would probably opt for either a combined simultaneous phosphorus removal system (e.g. Broad Run WRP) or a chemical removal system (e.g. Spokane WRF) (Daigger et al., 2009). In case of a combined system targeting LoT phosphorus removal, the contribution of EBPR in TP removal should be evaluated first for the chosen configuration. This could be achieved by using either ASM-based simulators or pilot operation with the same influent wastewater, or the combination of both. A thorough investigation of EBPR removal potential should be conducted as PAOs are susceptible to dynamic influent conditions and seasonal changes. Following that, the stoichiometric alum requirement could be calculated. During operation of the plant, anaerobic phosphorus release and aerobic uptake data should be profiled regularly, to monitor the impact of chemical phosphorus removal. However, sometimes, unexpected    107  suboptimal EBPR might be encountered. If so, higher alum dosing would be needed to meet effluent discharge criteria. In such a scenario, the potential inhibition of EBPR and the capability of the system in a subsequent revival of optimum EBPR and the timeframe for it, are not very well understood. Further research is needed in this direction, to assess the recovery of a chemical phosphorus removal-dominated system to that of a functioning EBPR system, with reduction or complete stoppage of alum addition. Additionally, the timeframe (i.e. number of SRTs) for such a conversion would provide important information on the persistence effect of metal salt in the mixed liquor.    5.4 Suspended Solids Data  Suspended solids concentration was related to the design SRT of the parallel MBNR systems. Based on the selected SRT of 25 days, mixed liquor wasting was done on a daily basis to maintain steady state conditions in the two systems.  Mixed liquor wasting was calculated in the present study by the following formula:    V = 	[(? (???))????????????]? 	?????	??                    (10)              where,   ? = Mixed liquor wasting rate (L/day)   ?= Reactor volume (L)  f = Reactor suspended solids concentration (mg/L)  ?? = Solids retention time (SRT) (days)   ?= Flow rate (L/day)  f?	= Effluent suspended solids concentration (mg/L)  f?= Suspended solids concentration from the wasting reactor (aerobic) (mg/L)  Total suspended solids (TSS) data for the MBNR (Biological) and MBNR (Chemical) systems are shown in Figure 5.21 and Figure 5.22, respectively. In addition, individual reactor average TSS and standard deviation values are summarized in Table 5.3. The data    108  show that TSS values were comparable in both systems when alum was not being added to either. With alum addition, TSS gradually increased in the MBNR (Chemical) system (Figure 5.22).  Although some alum was wasted during normal suspended solids wasting, the high final dosage (80 mg/L) of alum applied contributed significantly to the final MLSS concentration in the MBNR (Chemical) system.   Volatile suspended solids (VSS) to TSS percentage ratio profiles of the parallel systems are shown in Figure 5.23 and Figure 5.24 respectively. Also, individual reactor average % VSS/TSS ratio, along with standard deviation values, is summarized in Table 5.4. The data demonstrate that the %VSS/TSS ratio was higher than typically observed in other activated sludge systems. It was most probably due to efficient settling of influent wastewater particulate material at different stages before it entered the bioreactors:  (1) influent wastewater was collected after a preliminary settling tank at the SERC, University of British Columbia, (2) further settling took place in the transport containers and (3) finally, daily settling was observed in the influent holding tank. All of these resulted in loss of significant fraction of the particulate solids. Figure 5.24 and Table 5.4 also illustrate the expected decline of % VSS/TSS in the presence of alum.      109  MBNR (Biological)Days0 100 200 300 400 500Total suspended solids (TSS) (mg/L)020004000600080001000012000AnaerobicPre-anoxicAerobicPost-anoxic Membrane III III IV V Figure 5.21 TSS concentration in MBNR (Biological) system  MBNR (Chemical)Days0 100 200 300 400 500Total suspended solids (TSS) (mg/L)020004000600080001000012000Alum (mg/L)020406080100120AnaerobicPre-anoxicAerobicPost-anoxicMembraneAlumI II IIIIV V Figure 5.22 TSS concentration in MBNR (Chemical) system    110  Table 5.3 Average TSS values in MBNR (Biological) and MBNR (Chemical) system  Average TSS (mg/L)  MBNR (Biological)  MBNR (Chemical) Alum = 0 mg/L) MBNR (Chemical) Alum = 20 mg/L MBNR (Chemical) Alum = 40 mg/L MBNR (Chemical) Alum = 80 mg/L Anaerobic 1452 (?135) 1594 (?568) 1342 (?214) 1574 (?338) 2071 (?425) Pre-anoxic 2355 (?559) 2468 (?705) 2865 (?947) 2574 (?483) 3413 (?569) Aerobic 3290 (?571) 3549 (?894) 3384 (?262) 3548 (?262) 4913 (?554) Post-anoxic 3230 (?603) 3505 (?872) 3375 (?209) 3620 (?223) 4780 (?568) Membrane 4564 (?858) 4734 (?1528) 4896 (?651) 6402 (?1824) 7653 (?1388) ?: Standard Deviation (SD)  MBNR (Biological)Days0 100 200 300 400 500%  VSS/TSS 60708090100110120AnaerobicPre-anoxicAerobicPost-anoxic MembraneI IIIII IV V Figure 5.23 % VSS/TSS ratio in MBNR (Biological) system     111  MBNR (Chemical)Days0 100 200 300 400 500%  VSS/TSS60708090100110120Alum (mg/L)020406080100AnaerobicPre-anoxicAerobicPost-anoxicMembraneAlumIII III IV V Figure 5.24 % VSS/TSS ratio in MBNR (Chemical) system  Table 5.4 Average % VSS/TSS ratios in MBNR (Biological) and MBNR (Chemical) system  Average % VSS/TSS ratio  MBNR (Biological) MBNR (Chemical) Alum = 0 mg/L MBNR (Chemical) Alum = 20 mg/L MBNR (Chemical) Alum = 40 mg/L MBNR (Chemical) Alum = 80 mg/L Anaerobic 93 (?5) 93 (?4)          91(?3) 89 (?7) 76 (?3) Pre-anoxic 90 (?3) 91 (?3) 86 (?2) 83 (?4) 75 (?2) Aerobic 88 (?3) 88 (?3) 85 (?2) 81 (?3) 74 (?3) Post-anoxic 88 (?3) 88 (?3) 85 (?2) 81 (?3) 75 (?2) Membrane         87 (?2) 87 (?2) 84 (?2) 80 (?3) 75 (?2) ?: Standard Deviation (SD)  5.5 LoT Goal-MBNR (Biological) System  The MBNR (Biological) system effluent TP and TN (TKN + NO3-N) profiles are shown in Figure 5.25. It is evident from the TP data that the LoT objective could not be accomplished by EBPR alone in the MBNR (Biological) system. As far as TN data are    112  concerned, the LoT goal of 3 mg/L was accomplished. The average effluent TN concentration was 3 mg/L with 72 mg/L of methanol supplementation in the MBNR (Biological) system. An in-depth discussion has already been presented to explain the rationale for the nitrogen and phosphorus removal performance of the MBNR (Biological) system. Nonetheless, it was thought to be imperative to determine whether LoT goal could realistically be realized in processes comparable to the MBNR (Biological) system (i.e. employ biological nitrogen and phosphorus removal principles). Lesjean et al. (2005) reported effluent TP and TN concentrations of 0.05 mg/L (90 percent of the time) and 5 mg/L (some periods) respectively in a pilot scale 3-stage post-denitrification MBR system. These authors attributed LoT phosphorus removal to EBPR along with phosphorus precipitation with calcium and ferric ions present in the influent wastewater (approximately 130 mg Ca/L and 10 mg Fe/L). Nitrogen removal, specifically denitrification, was excellent for some periods when N-loading was constant in the process. However, variable N-loading caused fluctuations in denitrification performance and hence, impacted TN effluent concentrations. The key conclusions from their study were that chemical precipitation of phosphorus aided EBPR in meeting the LoT TP goal, whereas the LoT TN goal was not realized. Monti (2006) achieved effluent TP in the range of 0.1-0.2 mg/L by employing EBPR in a pilot scale modified UCT-MBR configuration. Although EBPR was broadly successful in that study, the authors reported periods of failure with elevated TP concentrations in the MBR effluent. In addition, effluent TN was consistently high with an average concentration of 10 mg/L in the effluent. No external carbon was added to improve denitrification during that work. The Monti study provided further evidence of the difficulty of achieving LoT nutrient removal without external addition of carbon and metal salt.          113  MBNR (Biological)Days0 100 200 300 400 500TP (mg/L)0.010.1110PermeateTarget LoT TP I II III IV V  MBNR (Biological)Days0 100 200 300 400 500TKN + NO3-N (mg/L)05101520Methanol (mg/L)20406080100EffluentTarget LoT TNMethanolI II III IV V Figure 5.25 Effluent TP and TN concentrations in the MBNR (Biological) system  In conclusion, EBPR performance could hypothetically be improved in MBNR (Biological) system with pH control in the aerobic reactor and an increased suspended solids inventory in the anaerobic reactor. Similarly, resolving the issue of foam mixing in the post-anoxic reactor could help achieving LoT TN goal in the MBNR (Biological) system more consistently.   5.6 LoT Goal-MBNR (Chemical) System  The MBNR (Chemical) system effluent TP and TN (TKN + NO3-N) profiles are shown in Figure 5.26. From the figure, it could be concluded that alum was the major driver in accomplishing extremely low effluent TP concentrations in the MBNR (Chemical) system. With the highest alum dosage of 80 mg/L, the average effluent total P concentration was 0.19 mg/L. Moreover, effluent TP was observed to be less than 0.1 mg/L on some days during that period. These data demonstrate the capability of the MBNR (Chemical) system in meeting the LoT TP goal. During the same period, effluent PO4-P concentrations were very low with an average concentration of 0.07 mg/L, indicating that most of the total P concentration was composed of other forms of phosphorus. It is imperative to understand the other fractions of phosphorus in the LoT effluent. According to Neethling et al. (2007), effluent from chemical phosphorus removal processes consists primarily of ortho-phosphorus and organic phosphorus.    114  Organic phosphorus is refractory in nature and typically present in the range of 0.01 to 0.05 mg/L in secondary and tertiary wastewater treatment plant effluents. Furthermore, the authors commented that determination of the characteristics of refractory organic phosphorus and its treatability is a key factor in pursuing the LoT TP goal. In the present study, due to the use of membrane-based solids-liquid separation, only rDOP (refractory dissolved organic phosphorus) was present in the permeate. The average permeate rDOP concentration was 0.12 mg/L (0.19 mg/L TP -0.07 mg/L PO4-P) in the MBNR (Chemical) system, which was higher than the range reported above (Neethling et al., 2007). The value shows that removal of rDOP would be crucial in meeting the LoT TP goal. Alum has been reported to remove rDOP (Arnaldos and Pagilla, 2010). For the MBNR (Chemical) system, there could be a requirement for significantly higher alum dosing to accomplish very low rDOP concentrations.      The TN profile in Figure 5.26 demonstrates the ability of the MBNR (Chemical) system in meeting LoT goal. The average permeate TN concentration was 2.4 mg/L with 72 mg/L of methanol supplementation, which was lower than the target LoT effluent TN of 3 mg/L.     MBNR (Chemical)Days0 100 200 300 400 500TP (mg/L)0.010.1110Alum (mg/L)020406080100EffluentTarget LoT TPAlumI II III IV V   MBNR (Chemical)Days0 100 200 300 400 500TKN + NO3-N (mg/L)051015202530Methanol (mg/L)20406080100EffluentTarget LoT TNMethanolI II III IV V Figure 5.26 Effluent TP and TN concentrations in the MBNR (Chemical) system  The key conclusion was that the MBNR (Chemical) system had the capability of meeting the LoT TP and TN goals. Nonetheless, external addition of methanol and alum were key factors in realizing the goal for the chosen process configuration.     115  5.7 Conclusions  The key conclusions on performance of the parallel MBNR systems are the following.  ? Permeate COD concentrations were similar in both MBNR systems with average removal efficiencies of 91 percent for the study.  ? Nitrification was essentially complete in the MBNR (Biological) system. Similar performance was achieved in the MBNR (Chemical) system, except when the average permeate ammonium concentration was 3.6 mg/L between operating day 354 and 380. ? Denitrification performance was primarily driven by methanol dosing. Without methanol addition, average permeate NO3-N concentrations were 11.3 mg/L and 13 mg/L in the MBNR (Biological) system and MBNR (Chemical) system respectively. With the highest methanol dosing applied (i.e. 72 mg/L), permeate NO3-N concentrations were very low with average values of 2.0 mg/L and 1.4 mg/L in the MBNR (Biological) system and MBNR (Chemical) system respectively.  ? The stoichiometric methanol ratio, i.e. mg methanol required / mg NO3-N removed, was calculated to be 7.6 and 6.1 for reducing the average permeate NO3-N concentration to 2 mg/L and 1.4 mg/L for the MBNR (Biological) system and MBNR (Chemical) system respectively.   ? EBPR employed in the MBNR (Biological) system (whole study period) and MBNR (Chemical) system (first two phases of operation) accomplished an average PO4-P removal efficiency of 41 percent and 27 percent respectively. The sub-optimum phosphorus removal performance was attributed to low observed aerobic reactor pH and low anaerobic SRT.  ? Permeate PO4-P removal steadily improved to 71 percent, 89 percent and 98 percent during the 20 mg/L, 40 mg/L and 80 mg/L alum dosing periods respectively, in the MBNR (Chemical) system.  ? In the MBNR (Chemical) system, an average molar Al/TP ratio of 1.9 was required to reduce the PO4-P concentration to 0.07 mg/L in the permeate. The ratio is in the recommended range of 1.0 to 2.5 for accomplishing < 0.05 mg/L TP in the effluent.     116  ? Operation of the MBNR (Chemical) system demonstrated that the relationship between chemical P removal and EBPR is dynamic and is dependent on alum dosage concentration. Alum dosing up to a concentration of 40 mg/L complemented EBPR in improving permeate P removal. On the other hand, 80 mg/L of alum supplementation competed with and finally, inhibited EBPR until the MBNR (Chemical) system was converted fully to a chemical P removal system. The dynamic relationship, i.e. complimentary or competitive, needed further confirmation, particularly the related impact on EBPR kinetics. A series of sequential anaerobic-aerobic batch tests have been conducted and the results are discussed in Chapter Six.      ? Alum addition did not have any influence on COD removal and denitrification in the MBNR (Chemical) system when compared with the MBNR (Biological) system. Similar conclusion can be made for nitrification except when permeate ammonium concentration was observed to be at elevated levels between operating day 354 and day 380, at the beginning of the period of 80 mg/L of alum dosing. It is thought that low influent alkalinity in combination with alum-induced alkalinity deficit conditions in the mixed liquor impeded nitrification temporarily.       ? The parallel MBNR system performance data for the period without external methanol and alum dosing provided evidence of the inability of the selected modified Bardenpho process configuration (as well as internal recycle rates) in accomplishing the LoT nutrient removal goal.     ? The MBNR (Biological) system average permeate TP concentration was 2.1 mg/L, demonstrating the difficulty of using only EBPR to achieve extremely low effluent TP concentrations. On the other hand, when supplemented with 72 mg/L of external methanol, the observed average permeate TN concentration was 3 mg/L, thus meeting the LoT TN effluent target. ? The MBNR (Chemical) system demonstrated LoT TP and TN treatment capability in the present study. The average permeate TP concentration was 0.19 mg/L (including periods when TP concentration was < 0.1 mg/L) with 80 mg/L of alum dosing. Similarly, an average permeate TN concentration of 2.4 mg/L was achieved with 72 mg/L of methanol dosing. The extremely low permeate TN and TP values in the    117  MBNR (Chemical) system established the significance of external alum and methanol requirement in realizing LoT effluent discharge goals.        118  6 Batch Studies for Comparative Evaluation of EBPR Kinetics and Stoichiometry of the Parallel MBNR Systems  6.1 Introduction  Batch tests were conducted on mixed liquor from the parallel MBNR systems for deeper understanding of the relationship between biological and chemical phosphorus removal at different alum dosing levels. Mixed liquor was withdrawn periodically from the aerobic reactors of the parallel MBNR systems for the tests. The tests were conducted in 1.0 L Erlenmeyer flasks and the temperature and pH were constant at 20 0C and 7.0 respectively. Before the start of the tests, a non-aerated period of 2-4 hours was maintained to remove residual nitrate present in the aerobic mixed liquor. This step was taken to avoid interference in EBPR activity test from denitrification-related VFA consumption. Following the non-aerated period, a sequential anaerobic (for 2 hours) and aerobic period (for 3 hours) were maintained to monitor phosphorus release and uptake capabilities of the mixed liquor, respectively. The batch tests were done in Phases II, III, IV and V to analyze the role of increased alum dosing rates on EBPR potential in the MBNR (Chemical) system (APPENDIX B). Three tests were conducted in each phase of operation. The EBPR potential of the MBNR (Chemical) system was assessed relative to the MBNR (Biological) system and conclusions were derived.  Finally, the batch tests were also expected to provide key information on EBPR kinetics and stoichiometry of both systems at different stages of operation.   6.2 Batch Tests - Phase II  Batch test NO3-N and PO4-P profiles of the parallel MBNR systems are shown in Figure 6.1. NO3-N was present at the beginning of the anaerobic period of all the tests. This was unexpected as a non-aerated period of 2 hours had been allowed prior to the batch test for endogenous denitrification. The presumption was that 100 percent of the nitrate would be removed during that time period. The results in Figure 6.1 indicated    119  otherwise and hence, the length of the non-aerated phase was increased to 4 hours in subsequent batch tests. The NO3-N concentration was reduced to 0 mg/L at the time of collection of the first samples from the anaerobic period (i.e. after ? hour) in all the batch tests. It is believed that acetate added at the start of anaerobic period was used for both denitrification and phosphorus release. The presence of residual acetate after ? hour of the anaerobic period (Figure 6.2) indicated that 100 mg COD/L of acetate was sufficient for denitrification and phosphorus release.  Hence, it was concluded that nitrate did not inhibit phosphorus release in the anaerobic phase of the batch tests. Nonetheless, almost all the acetate was utilized at the end of anaerobic period in five of the tests, with the only exception being Batch Test 1 with MBNR (Chemical) system mixed liquor. Also, from Figure 6.2, maximum acetate utilization took place in the first ? hour of the anaerobic phase.   The PO4-P profile in Figure 6.1 illustrates functioning EBPR in the parallel MBNR systems with classic anaerobic P release and aerobic P uptake.  As expected, phosphorus release coincided with acetate utilization (Figure 6.2). However, EBPR potential was highly dynamic in both MBNR systems with varied phosphorus release and uptake values observed in each of the three batch tests. A key observation from Figure 6.1 was that the EBPR potential in the MBNR (Biological) system was greater than that of the MBNR (Chemical) system during that specific period of operation. It should be reiterated that there was no difference in the operation of the parallel MBNR systems during this experimental phase.        120  Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)01020304050NO3-N (Day 106) (Batch Test 1)NO3-N (Day 122) (Batch Test 2)NO3-N (Day 134) (Batch Test 3)PO4-P (Day 106) (Batch Test 1)PO4-P (Day 122) (Batch Test 2)PO4-P (Day 134) (Batch Test 3)   Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentration (mg/L)01020304050NO3-N (Day 108) (Batch Test 1)NO3-N (Day 127) (Batch Test 2)NO3-N (Day 142) (Batch Test 3)PO4-P (Day 108) (Batch Test 1)PO4-P (Day 127) (Batch Test 2)PO4-P(Day 142) (Batch Test 3) Figure 6.1 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase II)  Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 106) (Batch Test 1)Acetate (Day 122) (Batch Test 2)Acetate (Day 134) (Batch Test 3)Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 108) (Batch Test 1)Acetate (Day 127) (Batch Test 2)Acetate (Day 142) (Batch Test 3) Figure 6.2 Batch test acetate profile of the parallel MBNR systems (Phase II)  EBPR kinetic and stoichiometric parameters were estimated from the batch studies. Figure 6.3 and Figure 6.4 summarize kinetic parameter values, i.e., maximum specific phosphorus release (mg P/(g VSS? hr) and uptake (mg P/(g VSS? hr) rates and the associated stoichiometric parameter values, i.e.  P-released/VFAs-consumed (g P/g COD)    121  respectively, in the parallel MBNR systems. Both kinetic and stoichiometric values were the lowest in Batch Test 1 in both the MBNR systems. Subsequent tests resulted in higher values, although Batch Test 2 exhibited slightly higher parameter values than Batch Test 3. More importantly, the kinetic rates validate the earlier observation (Figure 6.1) of EBPR potential being significantly higher in the MBNR (Biological) system as compared to the MBNR (Chemical) system.          The stoichiometric parameter, P-released/VFAs-consumed, is typically expected to be of constant value in batch tests. This was not the case in the tests summarized in Figure 6.4. One explanation might be the occurrence of denitrification-related acetate utilization. A second factor could be presence of other microorganisms in the mixed liquor (for example, GAOs) which could utilize acetate under anaerobic conditions.  Batch test number0 1 2 3 4Max. sp. P-release rate (mg P/(gVSS.hr)01234567MBNR (Biological)MBNR (Chemical)62 %33 %24 %  Batch test number0 1 2 3 4Max. sp. P-uptake rate (mg P/(gVSS.hr)012345MBNR (Biological)MBNR (Chemical)12 %36 % 44 % Figure 6.3 Batch test maximum specific phosphorus release and uptake profile (Phase II)        122  Batch test number0 1 2 3 4 P-released/VFAs-consumed (g P/(g COD)0.00.10.20.30.40.5MBNR (Biological)MBNR (Chemical)41%36 %19 % Figure 6.4 Batch test P-released/VFAs-consumed profile (Phase II) (*VFA consumption related to denitrfication has not been subtracted)  The EBPR potential observed in the batch tests was evaluated with respect to the continuous system performance for the same period of study. Table 6.1 summarizes average anaerobic, aerobic and permeate PO4-P concentrations of the parallel MBNR systems. The data show that anaerobic PO4-P concentrations were higher in the MBNR (Biological) system at the times of the batch studies, whereas, aerobic and permeate   PO4-P concentration was higher in the MBNR (Chemical) system. Hence, EBPR potential differences observed in the batch studies were consistent with the phosphorus removal performances of the parallel MBNR systems.              Table 6.1 Parallel MBNR system P-profiling during the period of batch studies (Phase II)  MBNR (Biological) system  PO4-P (mg/L)  (between day 105-133) MBNR (Chemical) system  PO4-P (mg/L)  (between day 109-140) Anaerobic Reactor 12.2 (?2.9) 10.1 (?1.2) Aerobic Reactor 6.9 (?1.6) 7.5 (?1.8) Permeate 1.4 (?1.1) 2.3  (?1.5) ?: standard deviation    123  6.3 Batch Tests - Phase III  NO3-N and PO4-P profiles from the Batch Tests 4, 5 and 6, conducted with the aerobic zone mixed liquor from the parallel MBNR systems are summarized in Figure 6.5. The measured NO3-N concentrations were very low in both periods of all the tests. This established that a pre-test non-aerated phase of 4 hours was sufficient for removing residual nitrate present in the aerobic mixed liquor.   As far as PO4-P profile was concerned, anaerobic P release and aerobic P uptake were observed in all the batch tests. Figure 6.5 illustrates that similar EBPR potential was observed in all three batch tests of both the MBNR (Biological) system and the MBNR (Chemical) system. Moreover, the EBPR potential of the MBNR (Chemical) system was comparable to that of the MBNR (Biological) system. The key points are that (1) EBPR potential was steady in the parallel MBNR systems in Phase III of operation (unlike Phase II) and (2) EBPR potential was not inhibited by 20 mg/L of alum supplementation in the MBNR (Chemical) system. However, PO4-P release at the end of the 2 hour anaerobic period was lower in Batch Tests 5 and 6 (Figure 6.5) than in Batch Tests 2 and 3 (Figure 6.1) for both MBNR systems.   AnaerobicAerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)0510152025NO3-N (Day 246) (Batch Test 4)NO3-N (Day 260) (Batch Test 5)NO3-N (Day 274) (Batch Test 6)PO4-P (Day 246) (Batch Test 4)PO4-P (Day 260) (Batch Test 5)PO4-P (Day 274) (Batch Test 6)    Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentrattion (mg/L)0510152025NO3-N (Day 247) (Batch Test 4)NO3-N (Day 261) (Batch Test 5)NO3-N (Day 274) (Batch Test 6)PO4-P (Day 247) (Batch Test 4)PO4-P (Day 261) (Batch Test 5)PO4-P (Day 274) (Batch Test 6) Figure 6.5 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase III)    124  Batch test acetate utilization profiles are shown in Figure 6.6. One set of data (batch test on day 274 with MBNR (Chemical) mixed liquor) was not reported, as samples were lost during analysis. Nevertheless, the data in Figure 6.6 illustrate that a significant amount of acetate was still present at the end of the anaerobic period. This was in contrast to the results of the Phase II batch tests (Figure 6.2) in which essentially all of the acetate had been consumed at this point. As mentioned in Section 6.2, the consumption of some of the acetate could be attributed to the presence of nitrate at the beginning of anaerobic phase in Phase II batch tests. In addition, it is hypothesized that a higher abundance of PAOs in the mixed liquor during Phase II of operation caused more extensive acetate uptake and hence, more phosphorus release (Figure 6.1). Nonetheless, in the Phase III tests, residual acetate was rapidly oxidized in the aerobic period as demonstrated in Figure 6.6. More acetate seemed to be utilized in the anaerobic periods of the MBNR (Biological) system batch tests than for the MBNR (Chemical) system batch tests (Figure 6.6).    Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 246) (Batch Test 4)Acetate (Day 260) (Batch Test 5)Acetate (Day 274) (Batch Test 6)AnaerobicAerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 247) (Batch Test 4)Acetate (Day 261) (Batch Test 5) Figure 6.6 Batch test acetate profile of the parallel MBNR systems (Phase III)  EBPR potential was also assessed by profiling Mg+2 and K+1 in the anaerobic and aerobic phase of the batch tests (Figure 6.7). The samples were collected for the first two batch tests only. Nonetheless, cation concentrations mimicked the PO4-P profiles with    125  release and uptake taking place in the anaerobic and aerobic periods respectively. Figure 6.7 further demonstrates that the cation profile was comparable in the two batch tests as well as in the parallel MBNR systems. This provides supplemental evidence of steady EBPR in Phase III of operation.  Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)0510152025Mg+2 (Day 246) (Batch Test 4)Mg+2 (Day 260) (Batch Test 5) K+1 (Day 246) (Batch Test 4)K+1 (Day 260) (Batch Test 5)    Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentration (mg/L)0510152025 Mg+2 (Day 247) (Batch Test 4) Mg+2 (Day 261) (Batch Test 4) K+1 (Day 247) (Batch Test 4) K+1 (Day 261) (Batch Test 5) Figure 6.7 Batch test Mg+2 and K+1 profile of the parallel MBNR systems (Phase III)  EBPR kinetic and stoichiometric parameters were estimated for the Phase III batch studies. Figure 6.8 and Figure 6.9 summarize the kinetic parameter values, i.e., maximum specific phosphorus release (mg P/(g VSS?hr) and uptake (mg P/(g VSS?hr) and the values of the stoichiometric parameter, P-released/VFAs-consumed (g P/gCOD), respectively, in the parallel MBNR systems. The maximum specific phosphorus release values for both the MBNR systems were comparable in all three batch tests, although estimated values were lower in Batch Test 6 than in the previous two tests. Similarly, there was very little difference in the parallel MBNR systems as far as maximum specific phosphorus uptake data were concerned. In conclusion, the batch kinetic data provide additional confirmation of the earlier observation (refer Figure 6.5) that EBPR potential in the MBNR (Biological) system was comparable to that of the MBNR (Chemical) system during Phase III.      126  Interestingly, P-released/VFAs-consumed (g P/g COD) values were higher in the MBNR (Chemical) mixed liquor in batch test 4 and 5. Also, the values were different in the two batch tests for both the MBNR systems.  Batch test number3 4 5 6 7Max. sp. P-release rate (mg P/(gVSS.hr)012345MBNR (Biological)MBNR (Chemical)7 %18 %-7 %   Batch test number3 4 5 6 7Max. sp. P-uptake rate (mg P/(gVSS.hr)0.00.51.01.52.02.53.0MBNR (Biological)MBNR (Chemical)8 %-1  % Figure 6.8 Batch test maximum specific phosphorus release and uptake profile (Phase III) Batch test number3 4 5 6 7 P-released/VFAs-consumed (g P/(g COD)0.00.20.40.60.81.0MBNR (Biological)MBNR (Chemical)-59 %-223 % Figure 6.9 Batch test P-released/VFAs-consumed profile (Phase III)  Cation data shown in Figure 6.7 were further examined to elucidate relationships with respect to soluble orthophosphorus in the anaerobic and aerobic periods of the batch tests.    127  Figures 6.10 to Figure 6.13 illustrate molar K+1 and molar Mg+2 vs molar P values for the four batch tests. By examining the R2 values of the figures, it can be said that a linear model is a reasonably good fit as far as relationship between the cations and phosphorus is concerned. This conclusion is applicable to both the anaerobic and aerobic periods of the batch tests. However, the R2 value was lower in Batch Test 4 of the MBNR (Biological) system (Figure 6.10) and Batch Test 5 of the MBNR (Chemical) system (Figure 6.13). Another important observation was that mole K+1 / mole P and mole Mg+2 / mole P values were not constant and varied between periods, batch tests and between the parallel MBNR systems. Typically, the observed ratio was between 0.20 to 0.35 for both cations, which is in the range reported by several authors in literature (Table 6.6).      128   MBNR (Biological)Mole P0.2 0.4 0.6 0.8Mole K+10.250.300.350.400.450.500.550.608382.02217204360=+=R.x.yAnaerobic MBNR (Biological)Mole P0.2 0.3 0.4 0.5 0.6 0.7 0.8Mole K+10.450.500.550.607887.023688032310=+=R.x.yAerobic MBNR (Biological)Mole P0.2 0.4 0.6 0.8Mole Mg+20.000.050.100.150.200.250.307969.020202034240=-=R.x.yAnaerobic MBNR (Biological)Mole P0.2 0.3 0.4 0.5 0.6 0.7 0.8Mole Mg+20.100.150.200.250.307969.020527032550=+=R.x.yAerobic Figure 6.10 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 4 of MBNR (Biological) system (Phase III)  MBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole K+0.350.400.450.500.559599.023649023360=+=R.x.yAnerobic MBNR (Chemical)Mole P0.1 0.2 0.3 0.4 0.5 0.6 0.7Mole K+10.400.450.500.550.600.650.705875.023575025720=+=R.x.yAerobicMBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole Mg+20.000.050.100.150.200.250.309973.020625028580=+=R.x.yAnaerobicMBNR (Chemical)Mole P0.1 0.2 0.3 0.4 0.5 0.6 0.7Mole Mg+20.100.150.200.250.309827.020474031170=+=R.x.yAerobic Figure 6.11 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 4 of MBNR (Chemical) system (Phase III)        129  MBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8Mole K+10.250.300.350.400.450.500.550.600.659271.022723042330=+=R.x.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8Mole K+10.400.450.500.550.609169.0238030880=+=R.x.yAerobicMBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8Mole Mg+20.000.050.100.150.200.250.306662.020168030770=-=R.x.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8Mole Mg+20.100.150.200.250.307964.020937.021470=+=Rx.yAerobic Figure 6.12 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 5 of MBNR (Biological) system (Phase III)   MBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole K+10.400.450.500.550.600.650.706495.023688.037150=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.1 0.2 0.3 0.4 0.5 0.6 0.7Mole K+10.450.500.550.600.653851.025166.01140=+=Rx.yAerobicMBNR (Chemical)Mole P0.0 0.2 0.4 0.6Mole Mg+20.000.050.100.150.200.250.305148.020396.024220=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.1 0.2 0.3 0.4 0.5 0.6 0.7Mole Mg+20.100.150.200.250.309019.020927.020870=+=Rx.yAerobic Figure 6.13 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 5 of MBNR (Chemical) system (Phase III)    130  Table 6.2 shows average PO4-P concentrations in the anaerobic reactors, aerobic reactors and permeates of the parallel MBNR systems during the period of Phase III batch studies. Anaerobic PO4-P concentrations were slightly higher in the MBNR (Biological) system whereas aerobic PO4-P concentrations were similar in both systems between operating days 245 and 277. This observation is consistent with the batch test findings, which demonstrated comparable PO4-P release/uptake profiles in the parallel MBNR systems. The average permeate PO4-P concentration, on the other hand, was 1 mg/L lower in the MBNR (Chemical) system. This was attributed to the 20 mg/L of alum addition in Phase III of operation.    Table 6.2 Parallel MBNR system P-profiling during the period of batch studies (Phase III)  MBNR (Biological) system  PO4-P (mg/L)  (between day 245-277) MBNR (Chemical) system  PO4-P (mg/L)  (between day 245-277) Anaerobic Reactor 5.8 (?0.5) 5.0 (?0.6) Aerobic Reactor 2.0 (?0.4) 1.7 (?0.4) Permeate 1.6 (?0.4) 0.6 (?0.2) ?: standard deviation  6.4 Batch Tests - Phase IV  Figure 6.14 summarizes NO3-N and PO4-P profiles of the batch tests 7, 8 and 9, conducted with aerobic mixed liquor from the parallel MBNR systems during Phase IV of the study. NO3-N concentrations were extremely low in the anaerobic periods of each test. As observed earlier, a non-aerated period of 4 hours was sufficient for removal of the residual nitrate present in mixed liquor samples. However, significant NO3-N concentrations were observed in the aerobic phases of all the batch tests. Monti (2005) reported similar observations in his batch test work. He postulated that microorganisms exposed to batch conditions for long periods of time undergo lysis and release ammonium as a by-product. The ammonium is subsequently converted to nitrate under aerobic conditions. This was most likely the case for the present batch tests, as significant NO3-N concentrations also were observed in the aerobic phases of the batch tests conducted in Phase II (Figure 6.1) and Phase III (Figure 6.5) of operation.      131  The PO4-P profiles in Figure 6.14 show anaerobic P release and aerobic P uptake in all the batch tests, thus demonstrating EBPR capability of the parallel MBNR systems. This is particularly important for the MBNR (Chemical) system as, at this time, it was being supplemented with 40 mg/L of alum. In the MBNR (Biological) system, the PO4-P profile was slightly different in Batch Test 7 than in Batch Tests 8 and 9. On the other hand, the PO4-P profiles were almost identical in all three MBNR (Chemical) system batch tests. For that reason, it could be assumed that the EBPR potential of the parallel MBNR systems was consistent during Phase IV. Figure 6.14 also indicates that the PO4-P concentrations after the completion of the 2 hour anaerobic period were somewhat higher in the MBNR (Biological) system than in the MBNR (Chemical) system for Batch Tests 8 and 9.  Although alum-mediated inhibition is a possibility, an accurate assessment could only be done after cation profiling along with comparative EBPR kinetic rate analysis.   Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)051015202530NO3-N (Day 288) (Batch Test 7)NO3-N (Day 303) (Batch Test 8)NO3-N (Day 345) (Batch Test 9)PO4-P (Day 288) (Batch Test 7)PO4-P (Day 303) (Batch Test 8)PO4-P(Day 345) (Batch Test 9)   Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentrattion (mg/L)051015202530NO3-N (Day 288) (Batch Test 7)NO3-N (Day 303) (Batch Test 8)NO3-N (Day 345) (Batch Test 9)PO4-P (Day 288) (Batch Test 7)PO4-P (Day 303) (Batch Test 8)PO4-P (Day 345) (Batch Test 9) Figure 6.14 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase IV)  Acetate utilization in the Phase IV batch tests is shown in Figure 6.15. Two interesting points to note are that significant amounts of acetate remained unutilized at the end of the anaerobic phases (similar to Phase III) and each batch test exhibited a different    132  residual acetate after the anaerobic phase. The least acetate utilization was observed in MBNR (Biological) system Batch Test 7, while the greatest acetate utilization was observed in the MBNR (Chemical) system Batch Test 9. It was difficult to identify the cause for such a phenomenon except that the possibility of continuously dynamic mixed liquor with non-PAO cultures having the ability to utilize acetate in anaerobic conditions. Nonetheless, acetate was oxidized in the first ? hour of the aerobic phase of all the batch tests.    Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 288) (Batch Test 7)Acetate (Day 303) (Batch Test 8)Acetate (Day 345) (Batch Test 9)AnaerobicAerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 288) (Batch Test 7)Acetate (Day 303) (Batch Test 8)Acetate (Day 345) (Batch Test 9) Figure 6.15 Batch test acetate profile of the parallel MBNR systems (Phase IV)  Figure 6.16 shows Mg+2 and K+1 concentrations for the batch tests conducted in Phase IV of operation. From this figure, it can be observed that cation profiles mimicked those of phosphorus, as cation release occurred in the anaerobic periods followed by uptake in the aerobic periods. Moreover, Mg+2 and K+1 profiles were very similar in all three batch tests of the MBNR (Biological) system as well as those of the MBNR (Chemical) system. This supports the earlier observation that EBPR potential was consistent in both the MBNR systems throughout Phase IV batch studies. Since the cation profiles of the MBNR (Chemical) system were also comparable to those of the MBNR (Biological) system, the likely conclusion is that 40 mg/L of alum did not inhibit EBPR during the period of Phase IV batch tests.       133  AnaerobicAerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)0510152025Mg+2 (Day 288) (Batch Test 7)Mg+2 (Day 303) (Batch Test 8)Mg+2 (Day 345) (Batch Test 9)K+1 (Day 288) (Batch Test 7)K+1 (Day 303) (Batch Test 8)K+1 (Day 345) (Batch Test 9)   Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentrattion (mg/L)0510152025Mg+2 (Day 288) (Batch Test 7)Mg+2 (Day 303) (Batch Test 8)Mg+2 (Day 345) (Batch Test 9)K+1 (Day 288) (Batch Test 7)K+1 (Day 303) (Batch Test 8)K+1 (Day 345) (Batch Test 9) Figure 6.16 Batch test Mg+2 and K+1 profile of the parallel MBNR systems (Phase IV)  Maximum specific phosphorus release (mg P/g VSS? hr) and maximum specific phosphorus uptake (mg P/g VSS? hr) were determined for the batch studies and these are summarized in Figure 6.17. The maximum specific phosphorus release data did not suggest any specific trend with higher values in the MBNR (Chemical) mixed liquor for Batch Tests 7 and 9 and lower value for Batch Test 8. On the other hand, maximum specific phosphorus uptake was higher in the MBNR (Chemical) mixed liquor for all the batch tests. Therefore, it is absolutely evident that 40 mg/L of alum addition did not adversely influence EBPR kinetics in the MBNR (Chemical) system during the Phase IV batch studies.   The stoichiometric parameter (P-released/VFAs-consumed; g P/g COD) values for the stage IV batch tests are shown in Figure 6.18. The values were variable between each batch test for the MBNR (Biological) system as well as the MBNR (Chemical) system. Figure 6.18 also demonstrates considerably higher P-released/VFAs-consumed values for the MBNR (Biological) system in Batch Tests 7 and 9. On the contrary, a slightly lower value was observed in Batch Test 8. The variability of the stoichiometric parameter    134  values in the batch tests could be due to the presence of non-PAOs that could utilize acetate anaerobically.    Batch test number6 7 8 9 10Max. sp. P-release rate (mg P/(gVSS.hr)012345MBNR (Biological)MBNR (Chemical)-12 %22 %-4 %Batch test number6 7 8 9 10Max. sp. P-uptake rate (mg P/(gVSS.hr)01234MBNR (Biological)MBNR (Chemical)-13 %-16 %-8 % Figure 6.17 Batch test maximum specific phosphorus release and uptake profile (Phase IV) Batch test number6 7 8 9 10 P-released/VFAs-consumed (g P/(g COD)0.00.20.40.60.81.01.2MBNR (Biological)MBNR (Chemical)43 %-9 %32 % Figure 6.18 Batch test P-released/VFAs-consumed profile (Phase IV)     135  Molar K+1 and Mg+2 vs. molar P plots for the batch tests are illustrated in the six figures below (Figure 6.19 to Figure 6.24). The profiles include data from both the anaerobic and aerobic periods of the tests.  A linear relationship exists between cation and phosphorus concentrations in both phases, although the R2 values were different in the three batch tests. In Batch Tests 7 and 9, R2 was mostly greater than 0.95 in the MBNR (Biological) system (Figure 6.19 and Figure 6.23) as well as in the MBNR (Chemical) system (Figure 6.20 and Figure 6.24), demonstrating excellent fit between Molar K+ and Mg+2 vs. molar P. On the other hand, the R2 value was predominantly less than 0.90 in Batch Test 8 for both the MBNR systems (Figure 6.21 and Figure 6.22).  As far as mole K+1/ mole P and mole Mg+2/ mole P value was concerned, the typical range was between 0.20 to 0.35, with very few exceptions. Once again the observed range was in accordance with the literature values shown in Table 6.6.                        136   MBNR (Biological)Mole P0.3 0.4 0.5 0.6 0.7 0.8Mole K+10.350.400.450.500.550.608851.022854.031530=+=Rx.yAnaerobic MBNR (Biological)Mole P0.2 0.3 0.4 0.5 0.6 0.7 0.8Mole K+10.300.350.400.450.500.550.609721.022934.03050=+=Rx.yAerobicMBNR (Biological)Mole P0.3 0.4 0.5 0.6 0.7 0.8Mole Mg+20.100.150.200.250.309861.020312.028130=+=Rx.yAnaerobicMBNR (Biological)Mole P0.2 0.3 0.4 0.5 0.6 0.7 0.8Mole Mg+20.100.150.200.250.30991.020419.027020=+=Rx.yAerobic Figure 6.19 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 7 of MBNR (Biological) system (Phase IV)    MBNR (Chemical)Mole P0.2 0.3 0.4 0.5 0.6 0.7 0.8Mole K+10.350.400.450.500.550.609912.023191.025970=+=Rx.yAnaerobic MBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole K+10.250.300.350.400.450.500.559628.022832.028490=+=Rx.yAerobicMBNR (Chemical)Mole P0.2 0.3 0.4 0.5 0.6 0.7 0.8Mole Mg+20.100.150.200.250.309789.020544.027830=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.2 0.3 0.4 0.5 0.6 0.7 0.8Mole Mg+20.100.150.200.250.309553.020445.027270=+=Rx.yAerobic Figure 6.20 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 7 of MBNR (Chemical) system (Phase IV)      137  MBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole K+10.300.350.400.450.500.550.608988.022907.021710=+=Rx.yAnaerobicMBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8 1.0Mole K+10.250.300.350.400.450.500.557458.022594.020090=+=Rx.yAerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole Mg+20.050.100.150.200.259174.020183.020890=+=Rx.yAnaerobicMBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8 1.0Mole Mg+20.050.100.150.200.258925.020268.018350=+=Rx.yAerobic Figure 6.21 Mole K+1 and mole Mg+2 vs. mole P in Batch Test of MBNR (Biological) system (Phase IV)    MBNR (Chemical)Mole P0.4 0.6 0.8 1.0 1.2Mole K+10.300.350.400.450.500.550.60799.021925.031170=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8 1.0 1.2Mole K+10.300.350.400.450.500.550.606731.023188.015640=+=Rx.yAerobicMBNR (Chemical)Mole P0.4 0.6 0.8 1.0 1.2Mole Mg+20.000.050.100.150.200.250.308607.02051.026280=-=Rx.yAnaerobicMBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8 1.0 1.2Mole Mg+20.000.050.100.150.200.250.307143.02021.015420=+=Rx.yAerobic Figure 6.22 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 8 of MBNR (Chemical) system (Phase IV)       138  MBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8 1.0Mole K+10.350.400.450.500.559272.023068.020760=+=Rx.yAnaerobicMBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8 1.0Mole K+10.300.350.400.450.500.559719.022509.028040=+=Rx.yAerobicMBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8 1.0Mole Mg+20.050.100.150.200.250.308897.020762.021550=+=Rx.yAnaerobicMBNR (Biological)Mole P0.0 0.2 0.4 0.6 0.8 1.0Mole Mg+20.100.150.200.250.309054.020771.020540=+=Rx.yAerobic Figure 6.23 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 9 of MBNR (Biological) system (Phase IV)    MBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole K+10.300.350.400.450.500.550.609712.023148.034430=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole K+10.300.350.400.450.500.550.609874.023029.033930=+=Rx.yAerobicMBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole Mg+20.050.100.150.200.250.30942.020634.030410=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.0 0.2 0.4 0.6 0.8Mole Mg+20.050.100.150.200.250.309582.020503.028350=+=Rx.yAerobic Figure 6.24 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 9 of MBNR (Chemical) system (Phase IV)     139  The conclusion from the Phase IV batch studies was that alum addition at the dose applied did not influence PAO kinetics. It was imperative to know whether the same trend was observed in the continuous flow system performance during that period. Table 6.3 shows average PO4-P concentrations in the anaerobic reactors, aerobic reactors and permeates of the parallel MBNR systems. There was very little difference in PO4-P concentration as far as anaerobic reactors were concerned. Therefore, it could be theorized that phosphorus release in the MBNR (Chemical) system was as good as that in the MBNR (Biological) system and was not inhibited by alum addition. On the other hand, the aerobic reactor and permeate PO4-P concentrations were significantly lower in the MBNR (Chemical) system during stage IV batch studies. As discussed in Section 5.3.4.2, the higher removal efficiency observed was a contribution of combined chemical phosphorus removal and EBPR mechanisms.   Table 6.3 Parallel MBNR system P-profiling during the period of batch studies (Phase IV)  MBNR (Biological) system  PO4-P (mg/L)  (between day 287-345) MBNR (Chemical) system  PO4-P (mg/L)  (between day 287-345) Anaerobic Reactor 6.5 (?0.9) 6.0 (?0.8) Aerobic Reactor 2.6 (?0.7) 0.7 (?0.4) Permeate 2.2 (?0.7) 0.3 (?0.3)  6.5 Batch Tests - Phase V  The NO3-N and PO4-P profiles of the batch tests 10, 11 and 12, conducted in Phase V are illustrated in Figure 6.25.  As discussed previously, the NO3-N concentrations were consistently close to zero during the 2 hours of the anaerobic period of each batch tests. In the aerobic period, elevated NO3-N concentrations were observed in all the batch tests. This was particularly prominent in Batch Test 10 of the MBNR (Chemical) system, where approximately 5 mg/L of NO3-N was observed in the mixed liquor (Figure 6.25). Cell lysis-related ammonium release and subsequent conversion to nitrate in the aerobic phase was the key reason for the elevated nitrate concentrations as shown in Figure 6.25.          140  PO4-P profiling of the batch tests, as illustrated in Figure 6.25, yielded noteworthy results. Although anaerobic release and aerobic uptake was observed in all batch tests, it was obvious that EBPR potential was significantly lower in the MBNR (Chemical) system than in the MBNR (Biological) system. As reported in the continuous flow system performance discussion (Section 5.3.4.2), inhibition of EBPR was most likely caused by alum addition. Alum dosing had been increased to 80 mg/L on day 346, which was 28 days before the first batch test of Phase V, i.e. Batch Test 10. During the intervening period, it is presumed that PAOs in the MBNR (Chemical) system could not compete with alum-based P complexation to accumulate soluble PO4-P for their growth. As a consequence, the abundance of PAOs apparently declined in the mixed liquor as reflected in the diminishing PO4-P profiles of Batch Tests 10 to 12 (Figure 6.25). On the other hand, the three MBNR (Biological) system batch tests exhibited much more extensive and consistent phosphorus release and uptake profiles, thus indicating steady EBPR potential in this system during the period of batch tests.  Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)05101520253035NO3-N (Day 374) (Batch Test 10)NO3-N (Day 390) (Batch Test 11)NO3-N (Day 404) (Batch Test 12)PO4-P (Day 374) (Batch Test 10)PO4-P (Day 390) (Batch Test 11)PO4-P (Day 404) (Batch Test 12)   Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentrattion (mg/L)05101520253035NO3-N (Day 374) (Batch Test 10)NO3-N (Day 390) (Batch Test 11)NO3-N (Day 404) (Batch Test 12)PO4-P(Day 374) (Batch Test 10)PO4-P (Day 390) (Batch Test 11)PO4-P (Day 404) (Batch Test 12) Figure 6.25 Batch test NO3-N and PO4-P profile of the parallel MBNR systems (Phase V)  Figure 6.26 shows acetate utilization profiles of the batch tests. Due to sample contamination, acetate concentrations could not be reported for Batch Test 12 of the    141  MBNR (Biological) system.  Acetate utilization took place anaerobically, although it was variable in each batch test. To be specific, anaerobic utilization was greater in Batch Tests 10 and 11 of the MBNR (Biological) system than for the MBNR (Chemical) system. More extensive acetate consumption in the two batch tests was consistent with greater PO4-P release in the MBNR (Biological) system (refer to Figure 6.25). However, the greatest anaerobic acetate utilization was observed in batch test 12 of the MBNR (Chemical) system. The puzzling fact was that the lowest PO4-P release occurred concurrently in the same batch test (Figure 6.25). The hypothesis of non-PAOs using acetate anaerobically is the most relevant explanation for the observation. Nevertheless, acetate remaining after the anaerobic phases was oxidized completely in first ? hour of the aerobic phases.      Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 374) (Batch Test 10)Acetate (Day 390) (Batch Test 11)AnaerobicAerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentration (mg/L)020406080100Acetate (Day 374) (Batch Test 10)Acetate (Day 390) (Batch Test 11)Acetate (Day 404) (Batch Test 12) Figure 6.26 Batch test acetate profile of the parallel MBNR systems (Phase V)  K+1 and Mg+2 concentrations in Phase V batch tests are summarized in Figure 6.27.  The broad conclusion from this figure is that both cations were anaerobically released and aerobically taken up in all the batch tests. In the case of the MBNR (Biological) system, identical profiles were observed for K+1 and Mg+2 in the three batch tests. This was not the case for the MBNR (Chemical) system, for which greater cation release was observed in Batch Test 10 as compared to the next two tests. Furthermore, the release and uptake was more extensive for the MBNR (Biological) system than for the MBNR (Chemical)    142  system. This provides additional affirmation of the adverse influence of alum addition on EBPR in the MBNR (Chemical) system.     Anaerobic AerobicMBNR (Biological)Hours0 1 2 3 4 5Concentration (mg/L)0510152025Mg+2 (Day 374) (Batch Test 10)Mg+2 (Day 390) (Batch Test 11)Mg+2 (Day 404) (Batch Test 12)K+1 (Day 374) (Batch Test 10)K+1 (Day 390) (Batch Test 11)K+1 (Day 404) (Batch Test 12)  Anaerobic AerobicMBNR (Chemical)Hours0 1 2 3 4 5Concentrattion (mg/L)0510152025Mg+2 (Day 374) (Batch Test 10)Mg+2 (Day 390) (Batch Test 11)Mg+2 (Day 404) (Batch Test 12)K+1 (Day 374) (Batch Test 10)K+1 (Day 390) (Batch Test 11)K+1 (Day 404) (Batch Test 12) Figure 6.27 Batch test Mg+2 and K+1 profile of the parallel MBNR systems (Phase V)  Figure 6.28 shows maximum specific phosphorus release (mg P/(g VSS? hr) and maximum specific phosphorus uptake (mg P/(g VSS? hr) values for the batch studies. Maximum specific phosphorus release was steady for the MBNR (Biological) system, while it gradually decreased in the MBNR (Chemical) system. In addition, there was a significant difference between the kinetic values of the parallel MBNR systems. A similar observation was made with the maximum specific phosphorus uptake profiles.  This indicates that while EBPR kinetic rates were consistent in the MBNR (Biological) system, they progressively declined in the MBNR (Chemical) system during the Phase V batch studies. The conclusion is very much in accordance with PO4-P profile (Figure 6.25) and cation profile (Figure 6.27), which confirm alum-related inhibition of EBPR in the MBNR (Chemical) system.       143  Figure 6.29 shows P-released/VFAs-consumed (g P/g COD) profile for the Phase V batch tests. There was large difference between the parallel MBNR systems as far as P-released/VFAs-consumed values were concerned. Also, the values were different in each batch test. The presence of non-PAOs utilizing acetate anaerobically has been identified in the previous batch tests as a potential contributor to the variability of the stoichiometric parameter and it holds true for the Phase V batch tests.  Batch test number9 10 11 12 13Max. sp. P-release rate (mg P/(gVSS.hr)012345MBNR (Biological)MBNR (Chemical)69 % 80 %83 %Batch test number9 10 11 12 13Max. sp. P-uptake rate (mg P/(gVSS.hr)0.00.51.01.52.02.53.0MBNR (Biological)MBNR (Chemical)66 %78 %86 % Figure 6.28 Batch test maximum specific phosphorus release and uptake profile (Phase V)     144  Batch test number9 10 11 12 13 P-released/VFAs-consumed (g P/(g COD)0.00.10.20.30.40.50.6MBNR (Biological)MBNR (Chemical)65 %32 % Figure 6.29 Batch test P-released/VFAs-consumed profile (Phase V)  Figures 6.30 to Figure 6.35 summarize molar K+1 and Mg+2 vs. molar P plots for Phase V batch tests. The broad agreement from the six figures is that a linear relationship existed between the cations and phosphorus in the anaerobic and aerobic periods of the batch tests. R2 values were usually higher than 0.90 and provided evidence for the above hypothesis. Molar K+1 and Mg+2 vs. molar P ratios were in the range of 0.20 to 0.35 for the MBNR (Biological) system, whereas higher ratios were observed for the MBNR (Chemical) system. It can be proposed that although linear relationships existed, alum adversely influenced the equilibrium between the cations and phosphorus.                  145   MBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole K+10.350.400.450.500.550.609915.023375.022680=+=Rx.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole K+10.350.400.450.500.550.609657.022495.029400=+=Rx.yAerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole Mg+20.100.150.200.250.300.350.409665.020642.027160=+=Rx.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole Mg+20.100.150.200.250.300.350.409838.020488.027320=+=Rx.yAerobic Figure 6.30 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 10 of MBNR (Biological) system (Phase V)      MBNR (Chemical)Mole P0.0 0.1 0.2 0.3 0.4Mole K+10.300.350.400.450.500.55986.023312.037180=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.0 0.1 0.2 0.3 0.4Mole K+10.200.250.300.350.400.450.509822.02282.052370=+=Rx.yAerobicMBNR (Chemical)Mole P0.0 0.1 0.2 0.3 0.4Mole Mg+20.050.100.150.200.259928.020971.031660=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.0 0.1 0.2 0.3 0.4Mole Mg+20.050.100.150.200.250.308356.020884.040930=+=Rx.yAerobic Figure 6.31 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 10 of MBNR (Chemical) system (Phase V)      146  MBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole K+10.300.350.400.450.500.550.609935.023032.027760=+=Rx.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole K+10.300.350.400.450.500.550.609665.022432.031340=+=Rx.yAerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole Mg+20.050.100.150.200.250.300.359796.020276.027260=+=Rx.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole Mg+20.050.100.150.200.250.300.359917.020214.027340=+=Rx.yAerobic Figure 6.32 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 11 of MBNR (Biological) system (Phase V)     MBNR (Chemical)Mole P0.00 0.05 0.10 0.15 0.20 0.25 0.30Mole K+10.300.350.400.450.509453.02332.031120=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.00 0.05 0.10 0.15 0.20 0.25 0.30Mole K+10.250.300.350.400.458322.022845.052340=+=Rx.yAerobicMBNR (Chemical)Mole P0.00 0.05 0.10 0.15 0.20 0.25 0.30Mole Mg+20.050.100.150.200.25964.020817.028960=+=Rx.yAnaerobicMBNR (Chemical)Mole P0.00 0.05 0.10 0.15 0.20 0.25 0.30Mole Mg+20.050.100.150.200.259442.020588.034830=+=Rx.yAerobic Figure 6.33 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 11 of MBNR (Chemical) system (Phase V)        147  MBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole K+10.30.40.50.60.79203.023485.027140=+=Rx.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole K+10.30.40.50.60.79218.02271.034890=+=Rx.yAerobic MBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole Mg+20.050.100.150.200.250.300.359650.020226.03040=+=Rx.yAnaerobicMBNR (Biological)Mole P0.2 0.4 0.6 0.8 1.0Mole Mg+20.050.100.150.200.250.300.359591.020169.030630=+=Rx.yAerobic Figure 6.34 Mole K+1 and mole Mg+2 vs. mole P in Batch Test 12 of MBNR (Biological) system (Phase V)      MBNR (Chemical)Mole P0.00 0.05 0.10 0.15 0.20Mole K+10.250.300.350.400.45

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.24.1-0072134/manifest

Comment

Related Items