UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Reducing operational costs in membrane bioreactors using slug bubbles Starke, Christina 2015

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

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

Item Metadata

Download

Media
24-ubc_2015_february_starke_christina.pdf [ 6.3MB ]
Metadata
JSON: 24-1.0135662.json
JSON-LD: 24-1.0135662-ld.json
RDF/XML (Pretty): 24-1.0135662-rdf.xml
RDF/JSON: 24-1.0135662-rdf.json
Turtle: 24-1.0135662-turtle.txt
N-Triples: 24-1.0135662-rdf-ntriples.txt
Original Record: 24-1.0135662-source.json
Full Text
24-1.0135662-fulltext.txt
Citation
24-1.0135662.ris

Full Text

REDUCING OPERATIONAL COSTS IN MEMBRANEBIOREACTORS USING SLUG BUBBLESbyChristina StarkeB.Sc., Technische Universitaet Berlin, Germany 2013A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFMaster of Applied ScienceinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Civil Engineering)The University of British Columbia(Vancouver)January 2015© Christina Starke, 2015AbstractMembrane bioreactors (MBRs) are commonly used in wastewater treatment pro-cesses. In fact, the demand is expected to increase with more than double digitgrowth annually over the next decade [5]. However, operational costs of MBRs arestill higher compared to operational costs of conventional treatment plants due tothe additional aeration and pumping required in MBRs. This study examines thefeasibility of using excess air that was used to clean the membrane for water con-veyance (known as airlift pump), for a minimized energy use in MBR processes.In order to meet the objective, prototypes of airlift pumps were built with differ-ent dimensions. The experimental results of each prototype were comprehensivelycompared to existing models in the literature. The models were modified for a bet-ter fit of the experimental data.It was determined whether a new apparatus, where many riser tubes were bundledtogether, would behave like many individual riser tubes. While the air was injectedat the bottom of the individual riser tube previously, the bundled riser tubes of thenew apparatus would be attached to a rubber sheet; this, was attached to a frame.iiThe rubber sheet was added to the apparatus in order to trap the air in the tankand lead it to the bundle of riser tubes. Different collector angles of the rubber aswell as different water heights were investigated. The experimental results werecompared to the previously modified models.The last step was to design a system that redirects the pumped water so that it canbe transported back to the head of the MBR plant.The results suggest that air exiting to the atmosphere from an MBR can be used totransport the water. However, the models are only able to predict water flows forindividual airlift pumps that consist of a single riser tube, where the air is injectedat its bottom. Further research needs to be done in order to be able to predict waterflows that can be achieved in systems, such as the one proposed in this presentstudy, which uses a bundle of riser tubes.iiiPrefaceThis dissertation is original, unpublished, independent work by the author, ChristinaStarke.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvList of Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiList of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiv1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.1 Membrane bioreactor ((MBR) . . . . . . . . . . . . . . . . . . . 42.1.1 Membranes . . . . . . . . . . . . . . . . . . . . . . . . . 5v2.1.2 Flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.1.3 Fouling . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.1.4 Air scouring . . . . . . . . . . . . . . . . . . . . . . . . 82.1.5 Benefits of MBRs . . . . . . . . . . . . . . . . . . . . . . 82.1.6 RAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.1.7 Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92.2 Airlift pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.2.1 Airlift pump models . . . . . . . . . . . . . . . . . . . . 122.2.2 Modifications . . . . . . . . . . . . . . . . . . . . . . . . 162.2.3 Friction factor for two-phase flow . . . . . . . . . . . . . 172.2.4 The drift-flux model . . . . . . . . . . . . . . . . . . . . 182.2.5 Nicklin model . . . . . . . . . . . . . . . . . . . . . . . 202.2.6 Reinemann model . . . . . . . . . . . . . . . . . . . . . 212.2.7 De Cachard & Delhaye model . . . . . . . . . . . . . . . 212.2.8 Chexal-Lellouche model . . . . . . . . . . . . . . . . . . 232.2.9 Delano model . . . . . . . . . . . . . . . . . . . . . . . . 262.3 MBR model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272.3.1 Bioreactor tank . . . . . . . . . . . . . . . . . . . . . . . 273 Knowledge gap and objectives . . . . . . . . . . . . . . . . . . . . . 304 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . 324.1 Bench scale system . . . . . . . . . . . . . . . . . . . . . . . . . 324.2 Scale-up from bench to pilot scale . . . . . . . . . . . . . . . . . 364.2.1 Single riser tube system . . . . . . . . . . . . . . . . . . 364.2.2 Air flow calibration . . . . . . . . . . . . . . . . . . . . . 37vi4.2.3 Bundle of riser tubes . . . . . . . . . . . . . . . . . . . . 384.2.4 Water redirection system . . . . . . . . . . . . . . . . . . 424.2.5 MBR system and RAS piping . . . . . . . . . . . . . . . 444.2.6 Calculation of achievable RAS . . . . . . . . . . . . . . . 485 Comparison of model and measured results for single riser tube inbench system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505.1 Comparison of models to experimental results . . . . . . . . . . . 505.2 Modifications to the model . . . . . . . . . . . . . . . . . . . . . 526 Theoretical energy savings for different scenarios . . . . . . . . . . . 566.1 Plant footprint, RAS pumping distance and available air for pumping 566.2 Head loss for RAS pumping . . . . . . . . . . . . . . . . . . . . 596.3 RAS pumping . . . . . . . . . . . . . . . . . . . . . . . . . . . . 616.4 RAS pumping capacity for MBRs of different sizes . . . . . . . . 687 Comparison of model and measured results for multiple riser tubesin a pilot system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717.1 Bundle of riser tubes . . . . . . . . . . . . . . . . . . . . . . . . 717.1.1 Influence of experimental conditions on square and rectan-gular collector geometry . . . . . . . . . . . . . . . . . . 727.1.2 Evaluation of the model, comparison of square and rectan-gular collector geometry . . . . . . . . . . . . . . . . . . 757.1.3 Revised air flow based on water flow . . . . . . . . . . . . 777.1.4 Distribution of water flow . . . . . . . . . . . . . . . . . 797.2 Water redirection system . . . . . . . . . . . . . . . . . . . . . . 82vii8 Conclusions and engineering significance of work . . . . . . . . . . 868.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 868.2 Engineering significance . . . . . . . . . . . . . . . . . . . . . . 878.3 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90A Sludge viscosity calculations . . . . . . . . . . . . . . . . . . . . . . 94B Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97C Air flow calibration in tank used for water collection system . . . . . 107D Achievable RAS flows for all MBR treatment capacities, scenariosand options considered . . . . . . . . . . . . . . . . . . . . . . . . . 110viiiList of TablesTable 2.1 Two-Phase flow parameters as used in [42] . . . . . . . . . . . 13Table 2.3 Parameters assumed for MBR model . . . . . . . . . . . . . . 29Table 4.1 Dimensions of water and air inlet of airlift pump (see Fig. 4.2)in mm. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35Table 4.2 Experimental conditions for the bench scale setup. . . . . . . . 36Table 4.3 Experimental conditions for bundle of riser tube setup. . . . . . 41Table 4.4 Experimental conditions for gas lifted water collection setup. . 44Table 5.1 K-values for D= 3/4 inch. . . . . . . . . . . . . . . . . . . . . 53Table 6.1 MBR dimensions and RAS piping distances. . . . . . . . . . . 58Table 6.2 Head loss calculations for RAS pipes. . . . . . . . . . . . . . . 60Table 6.3 Scenarios considered for RAS pumping for air that can be col-lected from directly above the membrane cassettes.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61Table 6.4 Available air flow [L/min] from membranes. . . . . . . . . . . 63ixTable 6.5 Scenarios and options considered for assumptions made aboutlocation of escaping air. . . . . . . . . . . . . . . . . . . . . . 63Table 6.6 Scenario A, 0.5 MGD (1.89 ∗106 L/day), 100% of air above mod-ule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65Table 6.7 Scenario A, 0.5 MGD (1.89 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 66Table 6.8 Scenario A, 0.5 MGD (1.89 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 67Table 7.1 Extrapolation for air flow per riser tube in a bundle based onmeasured results of rectangular setup, using the Nicklin model. 78Table A.1 Parameters to determine the impact of the sludge viscosity onbubble behaviour. . . . . . . . . . . . . . . . . . . . . . . . . 95Table C.1 Air flow measurements for calibration of pilot scale tank . . . . 109Table D.1 Scenario A, 0.75 MGD (2.84 ∗106 L/day), 100% of air abovemodule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111Table D.2 Scenario A, 0.75 MGD (2.84 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 112Table D.3 Scenario A, 0.75 MGD (2.84 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 113Table D.4 Scenario A, 1 MGD (3.79 ∗106 L/day), 100% of air above module.114Table D.5 Scenario A, 1 MGD (3.79 ∗106 L/day), 90% of air above module,10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 115xTable D.6 Scenario A, 1 MGD (3.79 ∗106 L/day), 75% of air above module,25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 116Table D.7 Scenario A, 2 MGD (7.57 ∗106 L/day), 100% of air above module.117Table D.8 Scenario A, 2 MGD (7.57 ∗106 L/day), 90% of air above module,10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 118Table D.9 Scenario A, 2 MGD (7.57 ∗106 L/day), 75% of air above module,25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 119Table D.10 Scenario A, 5 MGD (18.93 ∗106 L/day), 100% of air above module.120Table D.11 Scenario A, 5 MGD (18.93 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 121Table D.12 Scenario A, 5 MGD (18.93 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 122Table D.13 Scenario A, 10 MGD (37.85 ∗106 L/day), 100% of air abovemodule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123Table D.14 Scenario A, 10 MGD (37.85 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 124Table D.15 Scenario A, 10 MGD (37.85 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 125Table D.16 Scenario B, 0.5 MGD (1.83 ∗106 L/day), 100% of air above mod-ule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126Table D.17 Scenario B, 0.5 MGD (1.83 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 127Table D.18 Scenario B, 0.5 MGD (1.83 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 128xiTable D.19 Scenario B, 0.75 MGD (2.84 ∗106 L/day), 100% of air abovemodule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129Table D.20 Scenario B, 0.75 MGD (2.84 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 130Table D.21 Scenario B, 0.75 MGD (2.84 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 131Table D.22 Scenario B, 1 MGD (3.79 ∗106 L/day), 100% of air above module.132Table D.23 Scenario B, 1 MGD (3.79 ∗106 L/day), 90% of air above module,10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 133Table D.24 Scenario B, 1 MGD (3.79 ∗106 L/day), 75% of air above module,25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 134Table D.25 Scenario B, 2 MGD (7.57 ∗106 L/day), 100% of air above module.135Table D.26 Scenario B, 2 MGD (2.84 ∗106 L/day), 90% of air above module,10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 136Table D.27 Scenario B, 2 MGD (2.84 ∗106 L/day), 75% of air above module,25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 137Table D.28 Scenario B, 5 MGD (18.93 ∗106 L/day), 100% of air above module.138Table D.29 Scenario B, 5 MGD (18.93 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 139Table D.30 Scenario B, 5 MGD (18.93 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 140Table D.31 Scenario B, 10 MGD (37.85 ∗106 L/day), 100% of air abovemodule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141Table D.32 Scenario B, 10 MGD (37.85 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 142xiiTable D.33 Scenario B, 10 MGD (37.85 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 143Table D.34 Scenario C, 0.5 MGD (1.83 ∗106 L/day), 100% of air above mod-ule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144Table D.35 Scenario C, 0.5 MGD (1.83 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 145Table D.36 Scenario C, 0.5 MGD (1.83 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 146Table D.37 Scenario C, 0.75 MGD (2.84 ∗106 L/day), 100% of air abovemodule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147Table D.38 Scenario C, 0.75 MGD (2.84 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 148Table D.39 Scenario C, 0.75 MGD (2.84 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 149Table D.40 Scenario C, 1 MGD (3.79 ∗106 L/day), 100% of air above module.150Table D.41 Scenario C, 1 MGD (3.79 ∗106 L/day), 90% of air above module,10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 151Table D.42 Scenario C, 1 MGD (3.79 ∗106 L/day), 75% of air above module,25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 152Table D.43 Scenario C, 2 MGD (7.57 ∗106 L/day), 100% of air above module.153Table D.44 Scenario C, 2 MGD (7.59 ∗106 L/day), 90% of air above module,10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 154Table D.45 Scenario C, 2 MGD (7.59 ∗106 L/day), 75% of air above module,25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . . . 155Table D.46 Scenario C, 5 MGD (18.93 ∗106 L/day), 100% of air above module.156xiiiTable D.47 Scenario C, 5 MGD (18.93 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 157Table D.48 Scenario C, 5 MGD (18.93 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 158Table D.49 Scenario C, 10 MGD (37.85 ∗106 L/day), 100% of air abovemodule. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159Table D.50 Scenario C, 10 MGD (37.85 ∗106 L/day), 90% of air above mod-ule, 10% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 160Table D.51 Scenario C, 10 MGD (37.85 ∗106 L/day), 75% of air above mod-ule, 25% on sides. . . . . . . . . . . . . . . . . . . . . . . . . 161xivList of FiguresFigure 1.1 Power cost distribution for MBRs, modified from [21]. Theexploded piece of the pie chart represents the energy requiredfor RAS pumping. . . . . . . . . . . . . . . . . . . . . . . . 2Figure 2.1 Process diagramm for CAS and MBR. . . . . . . . . . . . . . 5Figure 2.2 Membrane fouling processes: a) complete pore blocking, b)adsorption, c) partial pore blocking d) cake layer [16]). . . . . 7Figure 2.3 Illustration of an airlift pump. . . . . . . . . . . . . . . . . . 10Figure 2.4 Flow regimes (modified from [31]). . . . . . . . . . . . . . . 11Figure 4.1 Experimental setups of the airlift pump.A) bench scale setup, discussed in Section 4.1, B) setup forpreliminary testing, discussed in Section 4.2.1, C) pilot scalesetup, discussed in Section4.2.3 . . . . . . . . . . . . . . . . 33xvFigure 4.2 Water and air inlet of airlift pump.The shaded areas correspond to the nozzle assembly which wasalso used to provide air. Dimensions can be found in Table 4.1;a) base of riser tube; b) horizontal cross section of riser tubenozzle; c) vertical cross section of riser tube and nozzle. . . . 34Figure 4.3 Airlift riser tubes used in the air collection system. . . . . . . 39Figure 4.4 Experimental setup of bundle of riser tubes attached to air col-lection apparatusa) picture; b) illustration. . . . . . . . . . . . . . . . . . . . . 40Figure 4.5 Illustration of experimental conditions for bundle of riser tubesthat were varied for both, the rectangular and square collectors:submergence ratio α , collector angles and different air flows. . 42Figure 4.6 Configurations A, C and D. . . . . . . . . . . . . . . . . . . . 43Figure 4.7 Top view of a 0.5 MGD plant according to assumptions madein the present study. The bundle of riser tubes setup is illus-trated in the dashed lines, the shaded areas are RAS pipes. . . 47Figure 5.1 Compairison of model and measured results.(Error bars correspond to minimum and maximum values. (D=3/4inch, riser tube length 1m, lift 0.5m)) . . . . . . . . . . . . . 51Figure 5.2 Evaluation of head loss coefficient using riser tubes with D=3/4inch, 1m long.(Error bars correspond to minimum and maximum values. (a:Lift 0.5m, b: Lift 0.6m, c: Lift 0.7m)) . . . . . . . . . . . . . 54xviFigure 5.3 Experimental vs. modeled data for K=1.(Error bars correspond to minimum and maximum values.) . . 55Figure 6.1 Achievable RAS vs plant size, 20 cm submerged, scenario A(see Section 6.3) . . . . . . . . . . . . . . . . . . . . . . . . 68Figure 6.2 Achievable RAS vs plant size, 20 cm submerged, scenario B(see Section 6.3) . . . . . . . . . . . . . . . . . . . . . . . . 69Figure 6.3 Achievable RAS vs plant size, 30 cm submerged, scenario C(see Section 6.3) . . . . . . . . . . . . . . . . . . . . . . . . 69Figure 7.1 Estimated total flow for bundle of riser tubes for rectangularand square collector.(Error bars correspond to estimated minimum and maximumvalues; a: low angle (0°), b: medium angle (6.7°), c: largeangle (13.2°), Alpha = 0.44; the revised data point is based oncalculations presented in Section 7.1.3; model results based inNicklin model) . . . . . . . . . . . . . . . . . . . . . . . . . 73Figure 7.2 Estimated total flow for bundle of riser tubes for rectangularand square collector.(Error bars correspond to estimated minimum and maximumvalues; a: low angle (0°), b: medium angle (6.7°), c: largeangle (13.2°), Alpha = 0.65; model results based in Nicklinmodel) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74Figure 7.3 Water vs. air flow according to Nicklin model.(D=3/4 inch, length =0.54cm, lift = 0.3m; the numbers next tothe lines represent riser tube numbers.) . . . . . . . . . . . . 78xviiFigure 7.4 Water flow distribution of individual riser tubes for the rectan-gular collector (alpha =0.44).(Error bars correspond to minimum and maximum values; airflows of a: 81.2 L/min; b: 142.5 L/min; c: 203.7 L/min) . . . 80Figure 7.5 Water flow distribution of individual riser tubes for the squarecollector (alpha =0.44).(Error bars correspond to minimum and maximum values; airflows of a: 81.2 L/min; b: 142.5 L/min; c: 203.7 L/min) . . . 81Figure 7.6 Effect of water collecting methods.(Error bars correspond to minimum and maximum values; a:medium ; b: large angle. The expected value is based on anextrapolation from 7 individual riser tubes as discussed in Sec-tion 7.1.1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84Figure 7.7 Illustrations of losses in water collection system. . . . . . . . 85Figure A.1 Froude number vs. Eötvös number [10]. . . . . . . . . . . . . 96Figure B.1 Results using riser tubes with D=3/4 inch, 1m long, beforemodification.(Error bars correspond to minimum and maximum values; a:Lift 0.6m; b: Lift 0.7m) . . . . . . . . . . . . . . . . . . . . 98Figure B.2 Results using riser tubes with D=3/4 inch, 2m long, beforemodification.(Error bars correspond to minimum and maximum values; a:Lift 1m; b: Lift 1.1m; c: Lift 1.2m) . . . . . . . . . . . . . . 99xviiiFigure B.3 Results using riser tubes with D=1 inch, 1m long, before mod-ification.(Error bars correspond to minimum and maximum values; a:Lift 0.5m; b: Lift 0.55m) . . . . . . . . . . . . . . . . . . . . 100Figure B.4 Results using riser tubes with D=1 inch, 2m long, before mod-ification.(Error bars correspond to minimum and maximum values; a:Lift 1m; b: Lift 1.1m; c: Lift 1.2m) . . . . . . . . . . . . . . 101Figure B.5 Evaluation of head loss coefficient using riser tubes with D=3/4inch, 2m long.(Error bars correspond to minimum and maximum values; a:Lift 1m; b: Lift 1.1m; c: Lift 1.2m) . . . . . . . . . . . . . . 103Figure B.6 Water flow distribution of individual riser tubes for the rectan-gular setup (alpha =0.64).(Error bars correspond to minimum and maximum values; airflows of a: 81.2 L/min; b: 142.5 L/min; c: 203.7 L/min) . . . 105Figure B.7 Water flow distribution of individual riser tubes for the squaresetup (alpha =0.64).(Error bars correspond to minimum and maximum values; airflows of a: 81.2 L/min; b: 142.5 L/min; c: 203.7 L/min) . . . 106Figure C.1 Representative locations of air flow measurements (top viewof tank). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107xixFigure C.2 Calibration curve for the air flowmeter attached to the tankused for the water collection system.(Error bars correspond to average minimum and maximum val-ues.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108xxList of AcronymsADF Average Daily FlowCAS Conventional Activated SludgeEES Engineering Equation SolverHRT Hydraulic Retention Timelmh Liter per m2 per hour [L/m2h]MBR Membrane BioreactorMF MicrofiltrationMGD Million Gallons per DayMLSS Mixed Liquor Suspended SolidsPVC Poly Vinyl ChlorideRAS Return Activated SludgeSCFM Standard Cubic Feet per MinuteSRT Sludge Retention TimeUF UltrafiltrationxxiList of SymbolsBo Bond numberC0 Liquid slug velocity coefficientρt p Two-phase densitye Average wall roughnessEo Eötvös numberε Gas void ratioεh Homogeneous void fraction (when S=1)εr Pipe roughnessη Dynamic viscosityf ′t p Fanning friction factorft p Two-phase friction factorFr Froude numberg Gravityhl head lossK Nominal head loss coefficientks Half-saturation concentrationLc Chexal-Lellouche fluid parameterxxiiM Morton numberµ Dynamic viscosityµg Maximum specific growth rateN f Archimedes NumberP PressureR Return Activated Sludge Rateφtank Membrane packing densityRe Reynolds numberS Substrateσ Surface tensionΣ Surface tension numberθ Hydraulic Retention Time (HRT)θc Sludge Retention Time (SRT)V VelocityVg j Drift velocityν Kinematic viscosityX BiomassY Biomass yieldxxiiiAcknowledgementsI would like to express my very great appreciation to my supervisor Dr. PierreBérubé. I am thankful for his guidance during this project as well as for supportthroughout my stay at UBC.Thanks are extended to Paula Parkinson, Tim Ma and Bill Leung for all their helpand advice on how to build my setups. Not only knowledge about work, but alsolife wisdoms were shared that embellished my daily routine.I have also been lucky to have the support and input of many friends and labmatessuch as Shona Robinson, Patricia Oka, Jeff MacSween, Joerg Winter and SepidehJankhah. I would like to especially acknowledge the time and effort Syed ZakiAbdullah put into discussions about my work.Finally, I would like to thank Greg Reynen and my family for their love and moralsupport.xxivChapter 1IntroductionAbout 2-3% of the world energy is consumed to treat and convey water and wastew-ater. For wastewater, pumping represents at least 30% of the energy demand.Therefore, reducing pumping costs is of importance [3]. Depending on the wa-ter source (freshwater, seawater or wastewater), the energy required to convey andtreat water is in the range of 0.05 to 5 kWh/m3. Typical energy footprint for conven-tional activated sludge (CAS) treatment ranges from 0.25 to 0.6 kWh/m3. With anadditional nitrification step, it may range from 0.3 to 1.4 kWh/m3 [2] . The energyfootprint for a membrane bioreactor (MBR) is higher and common values rangebetween 0.5 and 2.5 kWh/m3 [2]. Previous studies have demonstrated that MBRscan have lower capital costs than CAS plants [12], while producing a cleaner efflu-ent. The demand for MBR systems is expected to further increase with more thandouble digit growth annually over the next decade [5], due to increasingly stringentregulations and the large demand in water reuse applications.However, operational costs are still high in MBR systems due to the aeration (i.e.1air scouring) and pumping requirements. Over the past 15 years, the energy re-quired for air scouring has decreased by a factor of 15 [12]. However, the costsfor return activated sludge (RAS) pumping remain high. Figure 1.1 illustrates theenergy cost distribution in MBRs, emphasizing the contribution of RAS pumpingto the total operating costs. The present study examines the feasibility of using theenergy associated with the air escaping from the tank after it has been used for airscouring, to convey RAS and minimize energy use in MBR processes.Figure 1.1: Power cost distribution for MBRs, modified from [21]. The explodedpiece of the pie chart represents the energy required for RAS pumping.The structure of this present document is as follows. Background information forMBRs and airlift pumps is provided in Chapter 2. The knowledge gap is iden-2tified, and the objectives are defined in detail in Chapter 3. Chapter 4 presentsthe materials and methods used to address the research objectives. Different airliftpump models were investigated and compared to the data from this study. Chap-ter 5 presents which model provided the best fit, including modifications that weremade to the models. Assumptions for different scenarios (e.g. air is available atdifferent locations in the MBR tank) that were considered, including their possiblecost benefits, are outlined in Chapter 6. Results of the pilot scale air and water col-lection system tests are presented in Chapter 7, and Chapter 8 presents conclusionsand discusses pertaining to the engineering significance of the present study.3Chapter 2Literature review2.1 Membrane bioreactor ((MBR)A conventional activated sludge (CAS) plant typically consists of two systems, asuspended growth activated sludge biological reactor and a secondary clarifier. Thebioreactor relies on microorganisms that metabolize nutrients and organic matterto remove contaminants from the wastewater. In the process, biomass (i.e. solids)is generated [40]. The biomass solids must settle easily for effective solid-liquidseparation in the clarifier. The separated biomass is returned to the bioreactor. Thereturn stream is generally called RAS, and the mixture of biosolids in the bioreac-tor is referred to as mixed liquor suspended solids (MLSS).In a MBR the secondary clarifier is replaced by a membrane that performs thesolid-liquid separation (Fig. 2.1). Typically, membranes are located external of thebioreactor in a separate tank. MBRs require higher RAS rates than CAS systems(300 to 500% compared to 50 to 100% of the average daily flow (ADF) [40]) in4order to avoid an accumulation of solids in the membrane tank.Figure 2.1: Process diagramm for CAS and MBR.2.1.1 MembranesA membrane is a semipermeable barrier that allows at least one component of amixture to pass, while the other components are retained. Membranes generallyseparate mixtures based on molecular weight or particle size, driven either by anegative (vacuum) or positive pressure. The process consists of a feed flow that re-sults in a permeate and a retentate. Membranes can be classified by their material(i.e. synthetic polymers or ceramic), pore size (i.e. microfiltration (MF), ultrafil-5tration (UF), nanofiltration (NF) and reverse osmosis (RO)), module configuration(i.e. flat sheet or hollow fiber) or pressure applied for operation (i.e. high or lowpressure). In addition, membranes can be operated with a constant flux and variablepressure or variable flux and constant pressure. Material flowing to the membraneis generally referred to as feed, the material retained by the membrane as retentate,and material flowing through the membrane as permeate. The flux is defined asthe permeate normalized by the membrane area, and will be further discussed inSection 2.1.2.Both, flat sheet and hollow fibre polymeric MF and UF membranes are typicallyused in MBRs. The pore sizes of MF and UF membranes range between 0.1 to 0.4µm and 0.01 to 0.05 µm, respectively, and are usually operated at low pressures(0.2 to 1 bar).2.1.2 FluxThe flux is the permeate normalized by the membrane area. Lower flux operationsare generally associated with longer membrane life, lower operational risk andless maintenance due to the reduced mass flow and fouling on the membrane sur-face. However, lower flux operations require a larger membrane surface area thanhigh flux operations to treat a given feed flow, causing higher capital costs. Con-versely, higher flux operations are often associated with a shorter membrane life,higher risk for breaches, more maintenance, but less capital costs. When selectingthe flux, the composition of the wastewater, mixed liquor and temperature rangesshould be considered. Plants are usually designed for peak flow conditions [40].A permeate flux of 25 liter per m2 per hour [L/m2h] (lmh) is typical for commercial6MBR systems.2.1.3 FoulingOver time, material (microorganisms, colloids, solutes, and cell debris) retainedby the membrane can accumulate on its surface, forming a cake layer or plug-ging membrane pores (completely or partially). Material can also be adsorbedwithin the membrane pores, increasing the resistance to the permeate flow (Fig.2.2). The process of material accumulations, and the accumulated materials them-selves is generally referred to as fouling and foulants, respectively. Biomass flocsin MBR bioreactors are generally much larger than the membrane pores, and there-fore, tends to form a cake layer on the membrane surface.Figure 2.2: Membrane fouling processes: a) complete pore blocking, b) adsorption,c) partial pore blocking d) cake layer [16]).Different types cleaning procedures can be used to reduce fouling. Hydrauliccleaning, such as air scouring or backwashing, can be used to remove looselyattached foulants on the membrane surface. Chemical cleaning can be used toremove adsorbed foulants. Irreversible fouling cannot be removed by either physi-cal or chemical approaches [27].72.1.4 Air scouringAir scouring is a physical cleaning method used to prevent and reduce fouling ona membrane’s surface. Aeration is applied at the bottom of the membrane mod-ule. As the bubbles rise, shear forces are induced on to the membrane’s surfaceto remove accumulated foulants. Although very effective, membrane scouring hasbeen identified as one of the largest contributors to the high energy consumption ofMBRs [12].The amount of air scouring required for fouling control has substantially decreasedover the last two decades [11]. This evolution was made feasible by increasingmembrane surface area per unit tank volume and improvement of the aerationmethod. This largely resulted from a shift from continuous aeration (1995) tocyclic aeration (2000), then sequential aeration (2006) and, more recently, largepulse bubble aeration (2011) [11, 21].2.1.5 Benefits of MBRsCompared to CAS systems, MBRs can achieve better treated water quality in asmaller footprint because membranes require less space than clarifiers and higherconcentrations of mixed liquor can be maintained. MBRs can also be operatedwith a high degree of automation, which is beneficial if operator attention needsto be minimized, such as in remote treatment systems. In addition, MBRs canmeet more stringent discharge requirements for a number of parameters such assuspended solids, organic matter, total phosphorus or total nitrogen compared totraditional CAS plants [40].82.1.6 RASConcentrated solids in the membrane tank are usually pumped back to the bioreac-tor to maintain high biomass concentrations within the bioreactor tank. The flowof biomass from the membrane tank to the bioreactor, hereafter referred to as RAScan be calculated based on a mass balance around the membrane tank, which yieldsthe relationship presented in Equation 2.1.1.MLSSmembranetank =(R+1)RMLSSbioreactor (2.1.1)To maintain biomass concentrations of approximately 10,000 mg/L,expressed asMLSS), which is typical for MBRs, the recycle ratio should range from approxi-mately 300 to 500% of the incoming flow.2.1.7 CostsYoung et al. [43] developed energy consumption factors for MBRs and CAS plantswith an average daily flow (ADF) rate of 5 MGD (18.925 m3d ) and compared thelife cycle costs between the two systems. Tertiary filtration was required followingCAS treatment to produce an effluent quality comparable to that of an MBR. Theresults indicated that MBRs had higher operation and maintenance costs than CASsystems, mainly because of membrane replacement costs (approximately once ev-ery 10 years), costs associated with fouling control (such as recovery or chemicalcleaning and generally higher level of pre-treatment of the feed), and RAS pump-ing costs. RAS pumping was assumed to be 1.0 and 4.0 times the AFD for theCAS system and the MBR plant, respectively. Nonetheless, the study determinedthat the higher operation and maintenance costs associated with MBR systems was9offset by their lower capital costs [43].2.2 Airlift pumpsAn airlift pump is a vertical riser tube which is partially submerged in liquid intowhich compressed air is injected near the bottom of the riser tube. The risingair in the riser tube entrains the water, causing a gas lift effect (Fig. 2.3). Air-lift pumps are reliable and require limited maintenance compared to mechanicalpumps. Airlift pumps are commonly used in nuclear fuel processing plants due toits robustness in handling corrosive, abrasive and radioactive fluids [9].Figure 2.3: Illustration of an airlift pump.10Two-Phase FlowThe performance of an airlift pump depends the flow pattern or so called “flowregime” when gas and air flow up together in the vertical riser tube. The flowregime is a function of the fluxes of both phases, properties of various factors anddimensions and location of the channel [34]. The basic flow patterns are bubbly,slug, churn and annular, but definitions vary by authors (Fig. 2.2.1). A detailed de-scription of flow maps can be found in [34]. Early but fundamental studies focus-ing on airlift pumps indicated that the slug flow regime (long round nosed bubbles,called “Taylor” bubbles), with the length of the bubbles ranging from roughly thediameter of the tube to several times this value [33], were the most efficient regimefor airlift pumping [29]. Nicklin suggested that each flow-pattern needed a sepa-rate mathematical treatment [29]. All models investigated in this study assume theslug-flow regime, except the Chexal-Lellouche model, which is not flow specific.Figure 2.4: Flow regimes (modified from [31]).112.2.1 Airlift pump modelsThe five semi-mechanistic models considered in the present study are discussedbelow. The models considered were previously compared in [42].The governing equations, which are based on fluid properties, geometrical charac-teristics, and mass and momentum balances can be solved for velocities and hence,liquid flows. The equations for the different airlift models considered are depen-dent on the geometry of the airlift device. The equations presented below weredeveloped for the experimental conditions used in the present study (i.e. verticalcylindrical riser tube partially submerged in water to which air is introduced at thebase of the riser tube). The notation used in the equations is presented in Table 2.1,the indices “g” and “l” stand for gas and water, respectively.12Table 2.1: Two-Phase flow parameters as used in [42]Parameter Unit DefinitionD m Diameter of riser tubeP Pa PressureH m Submergence depthL m Riser tube lengthK Nominal head loss coefficientρg kg/m3 Density of gas phaseρl kg/m3 Density of liquid phaseρh kg/m3 Homogenous densityρt p kg/m3 Two-phase densityA = piD2/4 m2 Total cross sectional area of riser tubeAg m2 Cross sectional area gas occupiesAl = A−Ag m2 Cross sectional area liquid occupiesε = Ag/A - Gas void fraction of the flowεh Homogenous void fraction (S=1)εr mm Pipe roughness∀˙g m3/s Gas volumetric flow rate∀˙l m3/s Liquid volumetric flow rate∀˙= ∀˙g + ∀˙l m3/s Total volumetric flow ratejg = ∀˙g/A m/s Gas superficial velocityjl = ∀˙l/A m/s Liquid superficial velocityj = jg + jl m/s Total average velocity of the flowVm = j m/s Velocity of mixtureVg = jg/ε m/s Velocity of the gasVl = jl/1−ε m/s Velocity of the liquidm˙g kg/s Mass flow rate of gasm˙l kg/s Mass flow rate of liquidm˙ kg/s Total mass flow ratex = m˙g/m˙ - QualityS = Vg/Vl - Slip between phasesRe = jD/ν Reynolds numberft p Two=phase friction factorf ′t p = ftp/4 Fanning friction factor13Based on the momentum equation, the pressure at the inlet of the riser tube can bedetermined:Pa = Psys +ρlgH−ρlV 2l2(2.2.1)where P is the pressure, V is the velocity, g is gravity and the index “a” stands forthe point of the air inlet.Neglecting transition losses and assuming the water at the air inlet conserves itsmomentum as it mixes with the air yields:Pa2 = Pa +ρhVl(Vm−Vl), (2.2.2)where the index “m” stands for the mixture of air and water, “a2” is the state atthe air inlet after mixing. The velocities of both phases are assumed to be ap-proximately equal, therefore a homogenous density (ρh) is used in the momentumequation [42]. It can be calculated with a conservation of mass between the statebefore and after mixing:ρlAVl + m˙g = ρhAVm. (2.2.3)Rearranging yields:ρh =ρlVlA+ m˙gAVm=m˙l + m˙gpi4D2Vm, (2.2.4)The velocity of the mixture describes the total average velocity of both phases andequals the total average velocity of the flow j:14Vm =∀˙g + ∀˙lA=∀˙A= j. (2.2.5)Because the void fraction ε is defined as the average cross sectional area occupiedby the gas divided by the total cross sectional area Ag/A, the average cross sectionalarea occupied by liquid divided by the total cross sectional area Al/A must be 1−ε .Incorporating the above into the momentum equation in the riser tube (from inletto outlet) yields:Pa2 = Psys +ρlgL(1− ε)+ ft p(ρl jl +ρg jg)22ρt p(LD) (2.2.6)where ft p is the two-phase friction factor, which will be discussed in Section 2.2.3,and ρt p is the two-phase density of the mixture. The third term of Equation 2.2.6represents the head loss hl in the riser tube, which is usually in the following form:hl = fρV 22gLD. (2.2.7)The second represents the hydrostatic pressure along the riser tube. However, be-cause the riser tube is only partially filled with water, the term needs to be multi-plied with the fraction of water in the tube (1- ε).Because there is slip between the two phases, the homogenous density cannot beused. The density of the two-phase fluid is:ρt p = ρgε +ρl(1− ε). (2.2.8)15Combining Equations 2.2.2, 2.2.5, 2.2.6 and replacing Vl by jl yields:Pa +ρh jl( j− jl) = Psys + ft p(ρl jl +ρg jg)22ρt p(LD)+ρlgL(1− ε), (2.2.9)adding Equation 2.2.1 yields:Psys +ρlgH−ρlj2l2+ρh jl( j− jl) = Psys + ft p(ρl jl +ρg jg)22ρt p(LD)+ρlgL(1− ε).(2.2.10)Solving Equation 2.2.10 for H/L results in a general equation that describes thesubmergence ratio, which is equivalent to the average pressure gradient along theriser tube ([29]). It is presented in Equation 2.2.11.HL= ft p(ρl jl +ρg jg)22gDρlρt p+j2l2gL+( jlρh( j− jl))ρlgL+1− ε. (2.2.11)2.2.2 ModificationsThe equations presented in the previous section assume, that entrance and exitlosses are negligible. Considering the entrance and exit losses yields:HL= ft p(ρl jl +ρg jg)22gDρlρt p+j2l2gL+( jlρh( j− jl))ρlgL+1− ε +Kj2l2gD, (2.2.12)where K is the nominal head loss coefficient. K is usually assumed to be 0.5 and 1for entrance and exit losses, respectively [32].162.2.3 Friction factor for two-phase flowWhen modeling two-phase flows, predicting two-phase flow friction factors canbe challenging. Reinemann et al. [33] and Cachard and Delhaye [6] calculate thefriction factor using Equation 2.2.13.f =0.316Re0.25, (2.2.13)withRe =jDν , (2.2.14)where ν is the kinematic fluid viscosity of the liquid in m2/s, and Re is the Reynoldsnumber. This is a friction factor for smooth pipes and Reynolds numbers between3000 and 100 000 as suggested by Blasius [19].White [42] uses an approach recommended by Beattie and Whalley [4], that usesthe Colebrook equation to estimate friction factors. The Reynolds number is de-rived from a homogenous model but is calculated using two-phase properties,therefore one Reynolds number was used for modeling which was based on anaverage of each phase’s properties:1√f ′t p= 3.48−4log10[2εrD+9.35Ret p√f ′t p], (2.2.15)where εr is the pipe roughness, 0.0015 mm was used for this study (value for PVCpipes [1]).17The difference between the two friction factors (i.e. Equations 2.2.13 and 2.2.15)was reported to be negligible. As the presented study was in part based on White’swork, the Beattie and Whalley approach was chosen to calculate the friction factor.The parameters needed to solve Equation 2.2.15 are as follows.f ′t p =ft p4, (2.2.16)where f ′t p is the fanning friction factor,εh =xx+ρlρg(1− x), (2.2.17)where εh is the homogeneous void fraction (when S=1, no slip conditions),µt p = εhµg +µl(1− εh)(1+2.5εh), (2.2.18)where µ is the dynamic viscosity in Pas or kg/ms, andRet p =(ρg jg +ρl jl)Dµt p. (2.2.19)2.2.4 The drift-flux modelThe velocity of a slug bubble in a riser tube can be described relative to a movingliquid as presented in Equation 2.2.20:Vt =C0Vm +Vg j, (2.2.20)where Vt is the rise velocity of the bubble, C0 is the liquid slug velocity coefficient18and Vg j is the rise velocity of the same bubble in still fluid in m/s, also called thedrift velocity and can also be defined as:Vg j =Vg− j. (2.2.21)Nicklin et al. [30] suggested that the velocity of the Taylor bubble could also becalculated using Equation 2.2.22.Vt =∀gεA , (2.2.22)where ε is the gas void ratio [30].Combining Equations 2.2.5, 2.2.20 and 2.2.22 results in the following relationship,that summarizes the drift flux model:ε =jgC0( jl + jg)+Vg j. (2.2.23)Zuber and Findlay [44] published the same result for the analysis of two-phaseflows, but used a different approach. Their correlation is applicable to any two-flowregime and takes into account the effect of non-uniform flow and concentrationprofiles which is accounted for by the distribution parameter C0. The effect of thelocal relative velocity is also taken into account by the correlation developed byZuber and Findlay via the weighted mean drift velocity εVgj/ε, which is not readilyapparent from Equation 2.2.23 as described above; further details can be foundin [44]. Both effects depend on the flow regime and appropriate velocities andexpressions for the drift velocity need to be inserted. For the present study, slugflow regime was considered as it was reported to be the most efficient regime to19pump water [29]. Note that the Chexal and Lellouche model is not only applicableto one flow regime. In the present study, the drift flux model was applied to all themodels but the one by Delano.2.2.5 Nicklin modelNicklin et al. [30] formulated an equation to calculate the velocity of slugs in steadytwo-phase flow in 1962. The expression is equivalent to the denominator in thedrift-flux model (i.e.C0( jl + jg)+Vg j, in Equation 2.2.23). This model has beenused so widely that it is often considered the original slug flow model [42]. Nicklinused an approach developed by Dumitrescu [15] and Davies and Taylor [13] todescribe the drift velocity:Vg j = 0.35√g(ρl−ρg)Dρl≈ 0.35√gD. (2.2.24)In addition, Nicklin also assumed:C0 = 1.2. (2.2.25)It is argued that the bubble is transported faster than the average flow as it is locatedin a high-velocity region (i.e. at the centreline) of the riser tube. The value of 1.2was chosen because it describes the maximum to average velocity fully developedturbulent flow regime in cylindrical tubes [44]. The equations were tested for airand water in tubes with riser tube diameter between 0.63 and 2.5 inches (1.6 - 6.35cm), no minimum or maximum pipe length was given [30].202.2.6 Reinemann modelWhite et al. [41] and Zukoski [45] suggested in 1962 and 1966, respectively, thatthe effects of surface tension on the vertical flow becomes increasingly importantwhen the riser tube diameter is decreased below a certain point, without comingup with a model to predict the flow. In 1990, Reinemann et al. [33] published acorrelation for airlift pumps with riser tube diameters between 3 and 25 mm. It wasnoted that previous studies had been performed with riser tube diameters biggerthan 20 mm in which the effect of surface tension is small and can be neglected.Reinemann et al. [33] considered the effects of surface tension, altering the driftvelocity term to Equation 2.2.26:Vg j = 0.352(1−3.18Σ−14.77Σ2), (2.2.26)where Σ is the surface tension number, calculated with:Σ=σρgD2 , (2.2.27)where σ is the surface tension in N/m. C0 remained 1.2. Kouremenos and Staicos[25] also focused on small riser tube diameter airlift pumps (between 12 and 19mm). Their model was not considered for the present study it is not relevant forthe riser tube diameters considered in the present study.2.2.7 De Cachard & Delhaye modelDe Cachard & Delhaye [6] also developed a correlation for small riser tube diam-eter airlift pumps, still using C0 =1.2, but taking surface tension into account byaltering the drift velocity to:21Vg j = 0.345(1− e(−0.01N f0.345))(1− e(3.37−Bom))√gD, (2.2.28)where Bo is the Bond number and N f is the Archimedes number, defined as:Bo =(ρl−ρg)gD2σ ,and (2.2.29)N f =ρl(ρl−ρg)gD3µ2l, (2.2.30)respectively. m is defined for different ranges of N f :when N f > 250:m = 10, (2.2.31a)when 18 < N f <250:m = 69N−0.35f ,and (2.2.31b)when N f <18:m = 25. (2.2.31c)It should be noted that De Cachard & Delhaye also modelled an acceleration com-ponent that was added to the pressure gradient. The component is based on a thinliquid film falling around the Taylor bubble, without interfacial shear stress, insidea vertical cylinder. The acceleration component was calculated for the various ex-perimental conditions used in this study and found to be negligible.22De Cachard & Delhaye suggest to use this model for the design of small diameterriser tubes (up to 40 mm) and tall airlifts, with a length-to-diameter ratio greaterthan 250 [6].2.2.8 Chexal-Lellouche modelThe Chexal-Lellouche correlation was developed for a wide range of pressures,flows, void fractions, fluid types (steam-water, air-water, hydrocarbons and oxy-gen) and is valid for riser tube diameters up to 450 mm [8]. It is continuous anddoes not depend on flow regimes. The drift flux parameters C0 and Vg j are notrelated to the drift flux model and can be determined for both, cocurrent and coun-tercurrent flows.The liquid slug velocity coefficient is defined as presented in Equation 2.2.32.C0 =LcK0 +(1−K0)εr, (2.2.32)where Lc is the Chexal-Lellouche fluid parameter. It varies for different fluids, butfor an air-water mixture it is defined asLc = min(1.15ε0.45,1). (2.2.33)Equation 2.2.32 consists of many parameters, which are defined as follows:K0 = B1 +(1−B1)(ρgρl)0.25, (2.2.34)23r =1+1.57ρgρl1−B1, (2.2.35)B1 = min(0.8,A1),and (2.2.36)A1 =11+ e(−Reν60000) , (2.2.37)where Reν varies with Reg and Rel .If Reg > Rel:Reν = Reg, (2.2.38a)and if Reg ≤ Rel:Reν = Rel. (2.2.38b)The drift velocity is defined as:Vg j = 1.41[(ρl−ρg)σgρ2l]0.25C1C2C3C4, (2.2.39)whereC1 = (1− ε)B1 . (2.2.40)C2 varies with the liquid to gas density ratio; details can be found in [8]. As the24density of the gas and the liquid entering the riser tube are generally constant, C2can be calculated as follows:C2 = 1− e(−C51−C5), (2.2.41)C3 = max(0.5,2e(−Rel300000)),and (2.2.42)C7 =(D2D)0.6, (2.2.43)where D2 = 0.09144 m is a reference diameter. If C7 < 1 then:C4 =11− e−C8, (2.2.44a)else:C4 = 1, (2.2.44b)whereC8 =C71−C7. (2.2.45)For both riser tube diameters considered in the present study, C7 was found to belarger than 1, thus Equation 2.2.44b was used.C5 =√√√√150ρlρg. (2.2.46)252.2.9 Delano modelDelano models an airlift pump as part of the Einstein refrigeration model. Insteadof injecting air at the bottom of the riser tube, the riser tube is heated at the bot-tom, causing vapour bubbles to form and rise. This application is called vapour-liftpump and is a two-phase flow in a vertical riser tube- like an airlift pump, thereforethe same models can be used. Delano’s approach was referenced by other studiesrelated to vapour-lift pumps ([35]). Delano uses the analysis of Stenning and Mar-tin [36]. As mentioned earlier, this model does not use the drift flux model, but thefollowing definition for the void fraction instead:ε = 11+Sjljg(2.2.47)According to Stenning and Martin, the value for the slip (S) is normally between1.5 and 2.5 for the range of flows which yield the best performance [36]. Delanocalculates the head loss in the system using:K =4 f LD. (2.2.48)White [42] suggested to use the constants 2.5 and 17 for S and K, respectively(which was done in this study as well). Instead of Equation 2.2.12, Delano usedthe following equation for the basis of his model [14]:HL=11+[∀˙g∀˙lS] +j2l2gL[K(1+∀˙g∀˙l)+2∀˙g∀˙l+1]. (2.2.49)For the remainder of the present study, the models will be referred to with the26names of the authors.2.3 MBR modelWhen modeling the MBRs, 6 different plant sizes with capacities of 0.5, 0.75, 1,2, 5 and 10 MGD were considered. Before the membrane tank could be modeled,assumptions about the biological parameters had to be made, as discussed below.2.3.1 Bioreactor tankEquation 2.3.1 presents a mass balance of biomass for the bioreactor and mem-brane tank of an MBR.dXdtV = QiXi−QeXe +(µgSks +SX−bX)V, (2.3.1)where X is the biomass in kg/m3, S is the substrate in kg/m3, µg is the maximumspecific growth rate in 1/day, ks is the half-saturation concentration in mg/l, b is theendogenous decay rate in 1/day and the indices i and e stand for influent and efflu-ent, respectively.Assuming steady state conditions, no biomass in the effluent and that the incomingand outgoing flow are equal, replacing Xe by X and after some rewriting, Equation2.3.1 becomes:QV=1θc= µgSks +SX−b, (2.3.2)where θc is the sludge retention ime (SRT) in days.27Similarly, Equation 2.3.3 presents a mass balance of substrate for the bioreactorand membrane tank of an MBR.dSdtV = QiSi−QeSe−(µgYSks +SX)V, (2.3.3)where Y is the biomass yield.Assuming steady state conditions, that the incoming and outgoing flow are equal,replacing Se by S and after some rewriting, Equation 2.3.3 becomes:Si−Sθ =µgYSks +SX , (2.3.4)where θ is the hydraulic retention time (HRT). Using Equations 2.3.2 and 2.3.4, theachievable substrate concentration and the required HRT can be calculated. For theassumed kinematic stoichiometric parameters presented in the Table 2.3, the HRTresults in 0.049 days. This HRT was used to size the MBRs as presented in Chapter6.28Table 2.3: Parameters assumed for MBR modelParameter Sign ValueassumedMaximum specific growth rate [1/day] µg 6Half-saturation concentration [mg/l] ks 40Yield (ratio of mass of microorganisms formed (mea-sured as VSS), to the mass of substrate consumed [-][40]Y 0.4Endogenous decay rate [1/day] b 0.12Substrate in Influent [28] [mg/L] Si 210Values for the decay rate b for conventional activated sludge and aerobic processesis typically in the range of 0.04 to 0.075 1/day [22]. Experiments by Huang et al.[20] suggested a slightly higher decay rate for MBRs, but the chosen value is still inthe appropriate range. Huang et al. also calculated that the biomass yield in MBRsis in the same range as in conventional systems; therefore, the typical value forCAS was used [28]. The value chosen for µg is the typical value for CAS systemspresented in [28], as well, and the value for ks is within the presented range [28].It was assumed that the latter two parameters can be used for MBRs as well.29Chapter 3Knowledge gap and objectivesAs discussed in Section 2.1.7, MBRs have higher operation and maintenance coststhan conventional systems. Previous work focused on this issue by reducing thepower requirement through an optimization of the air scouring [21]. However, af-ter the air is used for cleaning purposes, it is essentially wasted as it escapes intothe atmosphere. The objective of the present study is to determine whether theexcess air (which otherwise escapes to the atmosphere), can be used to airlift andconvey return activated sludge (RAS) within an MBR system. In addition, opti-mum geometries for the design of the application were sought. These included thecollector angle, submergence ratio, shape of air collector (square or rectangular)and the water redirection system.Replacing electrically driven pumps, even if only partially with airlift pumps, pow-ered with escaping air, could result in a significant reduction in energy use. In orderto determine the feasibility of this approach, several tasks had to be accomplished.The tasks that address the objectives are the following:30• Task 1: Model water flow of airlift pumps– Compare the results from exiting models to those obtained experimen-tally– Modify existing models as needed to reliably predict water flow in airlift pumps• Task 2: Model the extent of which RAS can be pumped using waste air forMBR plants with treatment capacities ranging from 0.5 and 10 MGD• Task 3: Design a pilot airlift RAS pump prototype– Design air collector– Design water redirection system• Task 4: Compare results from pilot scale experiments to those obtained bymodeling.31Chapter 4Materials and methodsA bench and a pilot scale setup were build as part of this present study. These wereused to validate the numerical model considered, and to confirm the feasibilityof conveying return activated sludge (RAS) within an MBR system, using airliftpumps. Both setups consist of the following items:• Riser tube• Tank• Air supply• Water flow measurementThe different setups are presented in Figure 4.1 and are discussed below.4.1 Bench scale systemThe experimental setup consisted of a vertical cylindrical riser tube partially sub-merged in a plexiglas tank (Fig. 4.1 A)). The tank dimensions were 15 cm (W) *32Figure 4.1: Experimental setups of the airlift pump.A) bench scale setup, discussed in Section 4.1, B) setup for preliminary testing,discussed in Section 4.2.1, C) pilot scale setup, discussed in Section4.2.33390 cm (L) * 150 cm (D). The riser tube was attached to the side of the tank. Polyvinyl chloride (PVC) riser tubes with lengths of 1 and 2 meter, and riser tube innerdiameters of 3/4 inch (1.905cm) and 1 inch (2.54 cm) were considered.Air was injected through a nozzle located at the bottom of the riser tube (Fig. 4.2).The inner and outer diameter of the air inlet nozzle were 3.6 and 5.7 mm, respec-tively (additional dimensions are presented in Table 4.1). A valve and flowmeterwith an accuracy of 3% (1G08R3, Key Instruments), with glass float was used tocontrol and measure the air flow rate. Air flow rates of 6.0, 7.6, 9.1, 10.7, 12.3,13.9, 15.5, 17.1, 18.7, 20.4, 22.1 L/min, corresponding to rates at standard atmo-spheric conditions, were considered.(a) (b) (c)Figure 4.2: Water and air inlet of airlift pump.The shaded areas correspond to the nozzle assembly which was also used to pro-vide air. Dimensions can be found in Table 4.1; a) base of riser tube; b) horizontalcross section of riser tube nozzle; c) vertical cross section of riser tube and nozzle.A T-connection with a side horizontal tube was placed at the height of the riser34Table 4.1: Dimensions of water and air inlet of airlift pump (see Fig. 4.2) in mm.Riser tube diameter [inch] Sectiona b c d e f3/4 7 13 19.3 4 8 17.71 7 14.25 19.5 4 8 17.7tube where water was to be collected (Fig 4.1 a). Water flow gas lifted throughthe riser tube was determined by measuring the volume of water collected fromthe tube connected to the T-connection over a period of 30 seconds. All waterflow measurements were done in triplicate. Water collected to measure the flowwas returned to the tank. A summary of the experimental conditions considered ispresented in Table 4.2.35Table 4.2: Experimental conditions for the bench scale setup.Riser tubeExperiment number Length [m] Diameter [inch] Submergence ratio α1 1 3/4 0.52 0.43 0.34 1 0.55 0.46 0.37 2 3/4 0.58 0.459 0.410 1 0.511 0.4512 0.4For each condition, air flow rates of 6.0, 7.6, 9.1, 10.7, 12.3, 13.9, 15.5, 17.1, 18.7,20.4, 22.1 L/min were considered. As mentioned in Section 2.2.1, the submergenceratio (α) is the ratio of the submergence depth to the riser tube length (H/L)4.2 Scale-up from bench to pilot scale4.2.1 Single riser tube systemThe experimental setup consisted of a 54 cm long vertical cylindrical PVC risertube, with a riser tube diameter of 3/4 inch (1.905 cm) partially submerged in a36stainless steel tank (Fig. 4.1 B). The tank dimensions were 85 (D) cm *45 (W)cm *213 (H) cm. The tank provides space for a membrane cassette (ZW500, GE,Oakville, Canada). Note that no membrane cassette was present during any exper-iments. The riser tube was held vertically in the tank by hand.The air was supplied to the tank through an air diffuser which was attached toan F-450 flowmeter (blue-white, USA). The diffuser consists of a one inch pipewith 1/4 inch diameter holes drilled into it, each ten cm apart from each other,enabling an evenly distributed air supply in the tank. Air flow rates of 81.2, 142.5and 203.7 L/min, corresponding to rates at standard atmospheric conditions, wereconsidered. The air flow calibration will be discussed in Section 4.2.2. The air wasnot directly injected into the riser tube; it was collected over an area of 50.3 cm2using a cylindrical collector attached to the base of the riser tube (Fig. 4.1 B). Waterflow gas lifted through the riser tube was determined by measuring the volume ofwater collected from the tube connected to the L-connection over a period of 30seconds. All water flow measurements were done in triplicate. Water collected tomeasure the flow was returned to the tank.4.2.2 Air flow calibrationThe cross sectional distribution of air flow at the surface of the water in the tankwas characterized using an inverted 2 L graduated cylinder filled with water. Theair flow was determined based on the volume of air that accumulated in the invertedcylinder over a given time period at nine different locations in the tank (Fig. C.1).Air flow rates of 81.2, 142.5 and 203.7 L/min, corresponding to rates at standardatmospheric conditions, were considered. Measurements at each location were37taken in triplicate. The average air flow measured with the cross sectional area ofthe cylinder (≈ 0.005m2) was then extrapolated to the area in the tank.4.2.3 Bundle of riser tubesThe tank, air addition and air flow rates when considering a bundle of riser tubeswas identical to those for the single riser tube system, which was discussed in Sec-tion 4.2.1. However, the number of riser tubes, the area over which the air wascollected, and the water flow measurements differed for the bundle of riser tubesand the single tube system.A bundle of 19 tubes was considered as an alternative to multiple individual risertubes. As discussed in Chapter 5, tube diameters of 3/4 inch (1.905 cm) could moreeffectively lift liquid than 1 inch (2.54 cm) riser tubes. For this reason, 3/4 inchtubes were used in the present study. A total bundle diameter of 10 inch (25.4cm) was considered to be compatible with piping diameters commonly used atwastewater treatment plants. 19 thin walled aluminum riser tubes could be fit intoa 10 inch diameter pipe. The space between the tubes was sealed using silicone.The riser tube at the centerline of the bundle was placed 0.5 cm higher than theones next to it, while the ones on the perifory of the bundle were another 1 cmlower than the centerline tube (Fig. 4.3 a), allowing a better distribution of air tothe tubes. For the calculations, the slight height differences between the riser tubesat the centerline and perifory were neglected. The riser tubes were all 46 cm long.38(a) Top view (b) Individually measured riser tubes andnumbers they represent(c) Assumed distribution of riser tubesFigure 4.3: Airlift riser tubes used in the air collection system.The lower end of the riser tubes was attached to the air collection apparatus as pre-sented in Figure 4.4. Two geometries were considered for the air collector. Thefirst was rectangular in shape and was sized (83 cm * 43 cm) to be approximatelyequivalent to the surface area of the tank. The second had a square geometry (43cm * 43 cm). The top of the air collector was made of a flexible material, enablingthe angle of the base of the collector to be easily adjusted by lifting the bundle ofriser tubes above the base of the air collector. Three base angles were considered,0°, 6.7° and 13.2°. The collector angle corresponds to tan liftheight/43/2, all dimen-sions in cm. Therefore, for the rectangular collector, only the short collector sidewill be reported (angles 3.2° & 6.7° for long and short side, respectively will berepresented by 6.7°, while angles: 6.3° & 13.2° will be represented by 13.2°), asthe short side of the rectangular collector equals the collector angle of the squaredcollector.39(a) (b)Figure 4.4: Experimental setup of bundle of riser tubes attached to air collectionapparatusa) picture; b) illustration.The rectangular collector was assumed to trap most of the air added to the baseof the tank. Because the size of the collector was slightly smaller than the surfacearea of the tank, some air did escape along the sides of the frame. The amount ofair trapped by the collector was assumed to be equal to 90% of the air added to thebase of the tank. For the square collector, all of the air exiting the top of the tank,over the area occupied by the air collector, was trapped.Because of symmetry, water flow measurements were only performed within onequadrant of the bundled riser tubes as illustrated in Figure 4.3 b), and the resultsextrapolated to all the tubes in the bundle (as illustrated in Figure 4.3 c)). To mea-sure the water flow through an individual riser tube within the bundle, a section offlexible tubing was temporarily placed at the top end of the riser tube to direct thegas lifted water into a container. Note that the flexible tubing increased the lengthof the tube from 46 to 54 cm. This length difference was taken into account whenwater flows were modeled. Two submergence ratios, α = 0.44 (L=54cm, H=24cm),40and α = 0.65 (L=0.54cm, H= 35 cm) were considered. Flow was determined basedon the volume of liquid that accumulated in the container over a given time period.Water collected in the container was returned to the tank to maintain a constantliquid level. Flow measurements were done in triplicate.A summary of the experimental conditions investigated is presented in Table 4.3.The experimental conditions that were considered are also illustrated in Figure 4.5.Table 4.3: Experimental conditions for bundle of riser tube setup.Experimentnr.Air collectionframegeometryRiser tube di-ameter [inch]Submergenceratio (α)Collectorangle [°]1 rectangular 3/4 0.44 02 6.73 13.24 0.65 05 6.76 13.27 square 3/4 0.44 08 6.79 13.210 0.65 011 6.712 13.2For each experiment, the air flow rates 81.2, 142.5 and 203.7 L/min were considered.41Figure 4.5: Illustration of experimental conditions for bundle of riser tubes thatwere varied for both, the rectangular and square collectors: submergence ratio α ,collector angles and different air flows.4.2.4 Water redirection systemThe vertical flow through each of the riser tubes in the bundle must be redirectedto flow horizontally. Four options to achieve this were considered:1. a simple 90 degree elbow (D=4 inches, equivalent to 10.16cm) to redirectthe bulk of the flow from the riser tube bundle (Fig. 4.6 a)),2. a 90 degree elbow (D=4 inches (10.16 cm)) into which nine flexble tubes42were placed to redirect the flow from different sections of the riser tube bun-dle (note: the flexible tubes were not directly connected to the riser tubes),3. flexible tubes joined to the top of each riser tube in a bundle, with approxi-mately half of the flexible tubes grouped together and bent to one side, theother half grouped together and bent to the opposite side, effectively formingtwo bundles of 90° elbows pointing away from each other (Fig. 4.6 b)), and4. flexible tubes joined to the top of each riser tube in a bundle, with all flexibletubes grouped together and bent to one side, effectively forming a bundle of90° elbows (Fig. 4.6 c)).These systems will from now on be referred to as configurations A through D,respectively.(a) (b) (c)Figure 4.6: Configurations A, C and D.The water redirection experiments were examined using the square collector setup.A comparison of these 4 alternative designs to redirect the vertical flow from theriser tubes was performed for a selected set of experimental conditions as presented43in Table 4.4.Table 4.4: Experimental conditions for gas lifted water collection setup.Experimental condition Set-points consideredLength (L) [cm] 62.5Submerged depth (H) [cm] 27.5Submergence ratio (α) 0.44Collector angle [°] 6.7, 13.2Air flow [L/min] 81.2, 142.5, 203.74.2.5 MBR system and RAS pipingThe present study considered pumping mixed liquor using airlift pumps above themembrane tank, enabling it to flow by gravity to the bioreactor tank. Therefore, inaddition to bench and pilot scale equipment used, the present study also consideredfull scale MBR systems. Although full scale MBR systems were not built, it wasnecessary to size them (i.e. determine dimensions of system tanks) to address theresearch objectives.Full scale MBR systems, with capacities of 0.5, 0.75, 1,2,5 and 10 MGD (1 MGD isequivalent to 3.785 ∗106 L/day), were considered. This range was selected becauseit corresponds to the capacities of most existing MBRs. The sizing was impor-tant because the longer the distance between the membrane tank and the bioreactortank, the longer the pipe between these two tanks and therefore the greater thehead loss. As a consequence, the elevation to which the mixed liquor from themembrane tank needs to be air lifted increases with the size of the MBR system.The design of MBRs is highly site, application, membrane type and manufacturerspecific. For the present study, a number of assumptions were considered in deter-44mining the dimensions of the system tank.Based on the biological parameters presented in Section 2.3.1, the tank sizes ofthe MBRs within the considered flow ranges had to be determined. The followingconfigurations and constraints were used.Configurations and constraintsThe membrane area required for the MBRs was calculated using Equation 4.2.1.Amembrane =Flow capacityFlux(4.2.1)A flux of 25 lmh was assumed. Based on the required membrane area, the re-quired number of membrane cassettes was determined. It was assumed that themembrane area in a module is of 31.2 m2 [18] and a cassette consists of 48 mod-ules [17], as is typical for ZW 500 MBRs (GE water and process technologies,Oakville, Canada). The cassettes were assumed to be placed next to each otherwithout spacing between them in a “train”. The maximum number of cassettes pertrain was assumed to be seven.The following conventions were assumed in sizing the membrane tank componentof an MBR.• The foot print area of all cassettes in the tank is 50% of the total foot printarea of the membrane tank (Fig. 4.7).• The distance between the cassettes and the tank wall on the short sides isconsidered to be negligible (shaded area in Fig. 4.7). For example, the foot45print area of the three cassettes required in a 0.5 MGD plant is approximately11 m2, therefore the total foot print area of the membrane tank is approxi-mately 22 m2. Because the volume of the shorter side of the membrane tankis negligible, the longer side is 5.235 m long (which is equivalent the lengthof a cassette).• The distance between the top of the cassette and the liquid level in the tankis 0.35 m.• The distance between the bottom of the cassette and the bottom of the tankis 0.25 m.• The height of the water in the tank is 3.145 m, which is the height of thecassette (2.545 m, [17]), plus the sum of 0.35 m and 0.25 m.• The bioreactor tank wall is located next to the long side of the membranetank, with both having the same length. Figure 4.7 presents the 0.5 MGDplant as an example, where the bioreactor would share the 5.2 m wall.46Figure 4.7: Top view of a 0.5 MGD plant according to assumptions made in thepresent study. The bundle of riser tubes setup is illustrated in the dashed lines, theshaded areas are RAS pipes.The total volume of both, the bioreactor tank and the membrane tank was assumedto be:V = θ ∗V˙ (4.2.2)where V is the volume in m3 and V˙ is the incoming flow in m3/d. θ is the HydraulicRetention Time (HRT), it was calculated in Section 2.3.1.Based on the dimensions of the system tank, it is possible to estimate the distanceover which the RAS must be conveyed. The following assumptions were consid-47ered in determining the distance over which the RAS is conveyed for MBRs withdifferent capacities.• The length of the riser tube is based on the height required to provide enoughhead loss to let the wastewater flow back by gravity plus an additional 10 cmsafety factor. This is discussed in Section 4.2.6.• Head loss is calculated based on the length of the RAS conveyance pipe andassuming that two 90 degree elbows are located within the RAS pipe. Thelength of the pipes transporting the RAS was assumed to be width of themembrane and bioreactor tank, plus a safety factor of one meter.• Several pipes convey the RAS to biological reactor (with D=10.16 cm). Thenumber of pipes assumed differs for different treatment capacities and arepresented in Table 6.2.• RAS pipes are located on the side of the tanks providing easy access to mem-branes if required, as illustrated in Figure 4.7.4.2.6 Calculation of achievable RASWater flow through an air lift pump generally increases as the air flow increases.However, a point of diminishing return occurs above which a further increase inair flow does not result in a proportional increase in water flow. The air flow tothe individual riser tubes in the riser tube bundle was selected as the air flow atthe point of diminishing return. This approach is consistent with that reported byKassab et al. [23] who suggested that the maximum efficiency does not occur atthe maximum water mass flow rate. The resulting total flow to all riser tubes in a48bundle was the sum of the air flow corresponding to the point of diminishing returnfor individual riser tubes. This will be further discussed in Section 6.3.The distribution of the air in the membrane tank is not known. To investigate thepotential effect of the distribution of the air flow, three different scenarios wereconsidered.1. 100% of the air can be collected from a location directly above the membranecassettes2. 90% of the air can be collected from a location directly above the membranecassettes, 10% of the air can be collected from the sides3. 75% of the air can be collected from a location directly above the membranecassettes, 25% of the air can be collected from the sidesThis will be further discussed in Section 6.3.49Chapter 5Comparison of model andmeasured results for single risertube in bench systemThe models presented in Section 2.2.1 were solved using the program engineeringequation solver (EES) for the conditions outlined in Section 4.1. The modeled re-sults were compared with the measured results for the single riser tube experimentsat bench-scale, to assess their validity.5.1 Comparison of models to experimental resultsThe models generally overpredicted the water flow measured experimentally. Arepresentative plot comparing the modeled and measured results is presented inFigure 5.1 . (Note: all model and experimental measurements for all conditions50investigated are presented in appendix B). The discrepancy between the modeledand measured results was lowest over the lower range of air flow considered (below10 L/min), and increased as the air flow increased. The fit of the Delano model tothe data was the poorest. This model overpredicted the liquid flow at low air flowsand underpredicted the liquid flow at high air flows. Although the overall trendbetween modeled and measured data was better for the Reinemann and Chexal-Lellouche models, these models could not be solved for low flows (i.e. air flowslower than approximately 5 L/min). For these reasons, the Delano, Reinemann andChexal-Lellouche models were not further considered.Figure 5.1: Compairison of model and measured results.(Error bars correspond to minimum and maximum values. (D=3/4 inch, riser tubelength 1m, lift 0.5m))51The results for the Nicklin and De Cachard & Delhaye models were similar overthe range of air flows considered. The main difference between the two modelsis the the drift velocity Vg j calculation (see Section 2.2.1). Although the NicklinModel overestimated the liquid flow at a given air flow, the overall trend of in-crease in liquid flow with air flow was consistent with the measured results. Also,the Nicklin model could be solved for low flow conditions, and is simpler than theDe Cachard & Delhaye model. For these reasons, the Nicklin model was consid-ered in the present study.For the remainder of this study, only the smaller riser tube diameter (3/4 inch)was considered because it achieved higher water flows compared to larger risertube diameters. For example, for a submergence ratio of 0.5, higher water flowswere achieved compared to the riser tube with the one inch diameter (Fig. 5.2a)and B.3a)). Also, the smaller riser tube diameter achieved water flows up to asubmergence ratio of α=0.7 (Fig. 5.2c), while 1 inch riser tubes barely lifted waterwhen a low submergence ratio of 0.55 was used (Fig. B.3b)).5.2 Modifications to the modelThe Nicklin model (as well as the other models considered) were developed forlong tubular configurations. For these conditions, entrance and exit losses can beignored. However, when dealing with relatively short tubular configurations, asconsidered in the present study, entrance and exit losses can become significant.The Nicklin model was modified to incorporate a head loss term as presented inSection 2.2.2. Because of the nature of the entrance and exit of the tubular config-52uration considered, the specific entrance or exit loss constant K was unknown. Thevalue of the sum of the entrance and exit loss constant K for the setup consideredwas determined by fitting the modified Nicklin model to the measured results.Estimated best fit values for the different experimental conditions investigated aresummarized in Table 5.1. As indicated, no single K-value could be used to matchthe modeled to the measured results over the entire range of air flows considered.This was unexpected because K-values associated with entrance and exit losses areconsidered to be constants. It is likely that in addition to entrance and exit losses,other mechanisms, which are not accounted for in the models considered, affectbehaviour of the airlift riser tubes.Table 5.1: K-values for D= 3/4 inch.n.a.: Model did not fit data for any K-value considered.Riser tubelength [m]Submergence ratio α0.5 0.4 0.31 0.7 1 1Submergence ratio α0.5 0.45 0.42 Approximatelybetween 0.2 and0.41 n.a.53(a) (b) (c)Figure 5.2: Evaluation of head loss coefficient using riser tubes with D=3/4 inch, 1m long.(Error bars correspond to minimum and maximum values. (a: Lift 0.5m, b: Lift 0.6m, c: Lift 0.7m))54To identify the best K-value to apply, a range of values, corresponding to those inTable 5.1 were considered for the different experimental conditions investigated.As illustrated in Figure 5.2 a), no single head loss coefficient could be used to ac-curately model the measured results over the entire range of air flows considered.At low air flow, the K value had limited effect on model results. This was expectedbecause the head loss associated with entrance and exit losses is proportional tothe square of the flow. However, at higher flows, the modeled and measured resultswere similar, when a K-value of approximately 1 was assumed.For the remainder of this present study, the K-value associated with entrance andexit losses was assumed to be 1. As illustrated in Figure 5.3, assuming a K-value of1 generates results that are consistently similar to measured results (i.e. 1:1 slopeof Figure 5.3).Figure 5.3: Experimental vs. modeled data for K=1.(Error bars correspond to minimum and maximum values.)55Chapter 6Theoretical energy savings fordifferent scenariosThe modified Nicklin model (i.e. including entrance and exit losses, see Section5.2) was used to estimate the amount of mixed liquor in an MBR that could bepumped from the membrane tank to the head of the biological tank, (i.e. the RASflow).6.1 Plant footprint, RAS pumping distance and availableair for pumpingThe RAS flow is expressed as a percentage of the influent flow to the MBR. Sixplant sizes with different capacities were considered (0.5, 0.75, 1, 2, 5 and 10MGD; 1 MGD is equivalent to 3.785 ∗106 L/day). The range of sizes was selectedbecause the most common treatment plants are able to treat 1 to 5 MGD, and in-stallations of sizes of 10 MGD or more are expected to increase [40]. To estimatethe distance over which the RAS needed to be pumped, the footprint of an MBR56was estimated using the configurations and constraints outlined in Section 4.2.5.All of the air introduced for fouling control was assumed to be available for airlifting. As presented in Section 4.2.6, the air present in the tank was assumed to be0.12 m3/m2hr [11]. Thus, the membrane area per MBR size was determined usingEquation 4.2.1. Reactor dimensions, membrane area, supplied air per reactor anddistances over which the RAS must be pumped for the different MBR plant sizesconsidered are summarized in Table 6.1.57Table 6.1: MBR dimensions and RAS piping distances.MBR system size [MGD] (1 MGD is equivalent to 3.785 L/day)0.5 0.75 1 2 5 10Volume [m3] 93.0 139.5 186.0 372.0 930.0 1860.1Volume membrane tank [m3] 69.4 92.8 116.2 232.1 608.9 973.9Volume bioreactor tank [m3] 23.6 46.7 69.8 140.0 321.1 886.2Width membrane tank [m] 4.2 4.2 4.2 8.5 12.2 12.2Length membrane tank [m] 5.2 7.0 8.7 8.7 15.8 25.3Height water level [m] 3.1 3.1 3.1 3.1 3.1 3.1Height tank wall [m] 4.0 4.0 4.0 4.0 4.0 4.0Calculated width of bioreactor tank [m] 1.4 2.1 2.5 5.1 6.4 11.1Length RAS pipe [m] 9.5 10.2 10.7 20.3 30.4 32.7Membrane area [m2] 3154.2 4731.3 6308.3 12616.7 31541.7 63083.3Air supplied in tank [m3/h] 378.5 567.8 757.0 1514.0 3785.0 7570.0586.2 Head loss for RAS pumpingIn order to be conservative, the diameter of the transporting pipes was assumed tobe 4 inches (10.16 cm), and a high average wall roughness height of e= 0.00026m [32] was assumed. Friction factors were determined using a Moody diagram.Results are presented in Table 6.2.59Table 6.2: Head loss calculations for RAS pipes.Multiple pipes were considered in order to reduce head loss.MBR system size [MGD] (1 MGD is equivalent to 3.785 L/day)0.5 0.75 1 2 5 10Single pipe transporting sludgeVelocity [m/s] 0.86 1.29 1.72 3.45 8.62 17.25Reynolds 87266.7 130900.0 174533.3 349066.7 872666.7 1745333.5f factor 0.036 0.032 0.030 0.028 0.026 0.025Head loss [m] 0.13 0.29 0.50 3.55 30.90 128.11Multiple pipes transporting sludgeNumber of pipes 2 2 2 3 9RAS per pipe [L/min] 2628.5 3942.7 5256.9 7009.3 8761.6 11682.1Velocity [m/s] 0.43 0.65 0.86 1.15 1.45 1.91Reynolds 43633.3 65450.0 87266.7 116355.6 145444.5 193925.9f factor 0.045 0.037 0.035 0.032 0.031 0.030Head loss [m] 0.06 0.13 0.22 0.59 1.23 2.27606.3 RAS pumpingThe height to which the activated sludge had to be lifted included the head lossgenerated when RAS flowed from the top of the riser tube bundle (total length(L) - submerged depth (H)) to the head of the biological tank and two 90° elbows,plus a 10 cm safety factor. This sum was calculated for all 6 MBR sizes considered.As discussed in Section 4.2.5, the depth of the liquid above the cassettes in a mem-brane tank is approximately 35 cm. Two different riser tube submergence depths(both ending above the membrane cassette) were considered: 20 cm (scenario Aand B, see Table 6.3) and 30 cm (scenario C, see Table 6.3). These two submergedlengths allow for either 15 or 5 cm between the cassette and riser tubes, to accom-modate an air collector.Table 6.3: Scenarios considered for RAS pumping for air that can be collectedfrom directly above the membrane cassettes.ScenarioParameter A B CSubmerged depth(H) [cm]20 20 30Head loss (hl) cm Depends on MBR size, see Table 6.2Safety factor (SF) 20 10 10Length (L) [cm] 20 + hl +20 20 +hl + 10 30 + hl +10The Length (L) of the riser tubes of each scenario is based on the sum H + hl + SF.In the experiments to study the pilot scale application, a setup with a bundle of risertubes was built to investigate scenario C in a 0.5 MGD plant. This setup had a total61length ‘L’ of 30 cm + 6 cm (hl for 0.5 MGD, see Section 6.2) + 10 cm = 46 cm.This setup was then repurposed with a shorter submerged depth of 20 cm, scenarioA. Since the riser tube setup was the same in scenarios A and C, the effective safetyfactor in scenario a is 20 cm.In order to model the scenarios presented in Table 6.3, the experimental riser tubeconfiguration (H + SF = 40 cm) was considered for or all six plant sizes (i.e. valuesof hl), including the 0.5 MGD size from the experiment.As discussed in Section 4.2.6, it is not known where on the surface of an MBR tankthe air escapes. Table 6.4 presents three possible distributions of air in the differenttank sizes, corresponding to the assumptions made. The fate of the escaping air wasmodeled using scenarios A, B and C for air escaping directly above the cassette.The fate of the escaping air was modeled using options I and II for air escaping atthe sides of the cassette (either submergence of 50 or 60 cm, respectively). OptionsI and II allow for deeper submergence of the riser tube, as there is no cassette inthe way. Thus, for each air distribution and plant size, all possible combinations ofscenarios A,B,C and options I and II were modeled, as summarized in Table 6.5.A range of air flows per riser tube was considered (i.e. discrete values between 2.5and 15 L/min). By applying the modified Nicklin model to this range of air flowsand the riser tube lengths and submergence ratios from scenarios A, B and C, 10values of water flows for individual riser tubes were determined at equal intervals.The air available per condition (see Table 6.4) was then divided by the respectiveconsidered air flows per riser tube, resulting in the number of riser tubes required62Table 6.4: Available air flow [L/min] from membranes.Distributionof airPlant size [MGD]0.5 0.75 1 2 5 10Air flow [L/min]100%abovemodule6308 9613 12617 25233 63083 12616790%abovemodule,5678 8651 11355 22710 56775 11355010% onside631 961 1262 2523 6308 1261775%abovemodule,4731 7209 9463 18925 47313 9462525% onside1577 2403 3154 6308 15770 31542Table 6.5: Scenarios and options considered for assumptions made about locationof escaping air.Distribution of air All plant sizes considered100% above module Scenario A, B and C90% above module, Scenario A, B and C10% on side Option I and II75% above module, Scenario A, B and C25% on side Option I and II63for the tank in each scenario. The calculated water flows were then multiplied bythe respective number of riser tubes to find the total water flow for each condi-tion and scenario. Optimum conditions were those which generated the highestRAS flow. Results for scenario a in 0.5 MGD plant are presented in Tables 6.6to 6.8. The results of the other treatment plant sizes and scenarios considered arepresented in Appendix D. In the tables, optimal conditions are identified.64Table 6.6: Scenario A, 0.5 MGD (1.89 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.46 0.2 6.00 7.00 8.00 9.00 10.00 11.00 12.00Water flow per riser tube [L/min] 0.77 1.12 1.42 1.67 1.89 2.07 2.24Number of riser tubes 1051.00 901.00 788.00 700.00 630.00 573.00 525.00Total RAS flow above module [L/min] 807.77 1008.39 1115.42 1167.61 1187.75 1188.50 1175.89The optimum is presented in bold.65Table 6.7: Scenario A, 0.5 MGD (1.89 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.46 0.2 6.00 7.00 8.00 9.00 10.00 11.00 12.00Water flow per riser tube [L/min] 0.77 1.12 1.42 1.67 1.89 2.07 2.24Number of riser tubes 946.00 811.00 709.00 630.00 567.00 517.00 473.00Total RAS flow above module [L/min] 727.07 907.66 1003.60 1050.85 1068.98 1072.35 1059.42Air on sides, option I, air per riser tube [L/min]0.76 0.5 1.80 2.93 4.07 5.20 6.33 7.47 8.60Water flow per riser tube [L/min] 1.63 2.68 3.32 3.78 4.13 4.41 4.64Number of riser tubes 350.00 215.00 155.00 121.00 99.00 84.00 73.00Total RAS flow from sides [L/min] 569.58 575.16 515.27 457.92 409.35 370.77 338.94Air on sides, option II, air per riser tube [L/min]0.86 0.6 1.80 2.93 4.07 5.20 6.33 7.47Water flow per riser tube [L/min] 2.02 3.01 3.63 4.08 4.41 4.69Number of riser tubes 350.00 215.00 155.00 121.00 99.00 84.00Total RAS flow from sides [L/min] 706.87 647.73 563.12 493.19 437.08 393.62Total RAS option I [L/min] 1647.50Total RAS option II [L/min] 1779.22The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.66Table 6.8: Scenario A, 0.5 MGD (1.89 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.46 0.2 6.00 7.00 8.00 9.00 10.00 11.00Water flow per riser tube [L/min] 0.77 1.12 1.42 1.67 1.89 2.07Number of riser tubes 788.00 675.00 591.00 525.00 473.00 430.00Total RAS flow above module [L/min] 605.64 755.45 836.57 875.71 891.76 891.89Air on sides, option I, air per riser tube [L/min]0.76 0.5 1.80 2.93 4.07 5.20 6.33Water flow per riser tube [L/min] 1.63 2.68 3.32 3.78 4.13Number of riser tubes 876.00 537.00 387.00 303.00 249.00Total RAS flow from sides [L/min] 1425.58 1436.56 1286.51 1146.69 1029.59Air on sides, option II, air per riser tube [L/min]0.86 0.6 2.00 3.00 5.00Water flow per riser tube [L/min] 2.02 3.06 4.00Number of riser tubes 788.00 525.00 315.00Total RAS flow from sides [L/min] 1591.47 1604.61 1258.74Total RAS option I [L/min] 2317.48Total RAS option II [L/min] 2483.36The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.67Even though the present study focuses on pumping mixed liquor in an MBR, thefluid properties of water were used for calculations. The density and viscosity ofwater and mixed liquor are relatively similar for MLSS concentrations less thanapproximately 10,000 mg/L, which is typical for MBRs. However, even at higherMLSS concentrations, the results of the present study are valid as discussed in theAppendix (A).6.4 RAS pumping capacity for MBRs of different sizesFor all scenarios, the achievable RAS, as a percentage of the incoming flow, de-creased as the size of the MBR system increased. The RAS must be transportedover longer distances as the capacities (i.e. size) of the MBR system increases.Figure 6.1: Achievable RAS vs plant size, 20 cm submerged, scenario A (seeSection 6.3)68Figure 6.2: Achievable RAS vs plant size, 20 cm submerged, scenario B (see Sec-tion 6.3)Figure 6.3: Achievable RAS vs plant size, 30 cm submerged, scenario C (see Sec-tion 6.3)69As illustrated in Figures 6.1, 6.2 and 6.3, the higher the submergence ratio, thegreater the extent to which RAS can be pumped. The model predicts scenario a,with the lowest submergence ratio, will perform the poorest. For 0.5 MGD plantsin scenario a, RAS flows between 90 and 189 % can be achieved, depending onwhere the air is located. For scenarios b and c, RAS flows between 163 and 293% and 265 and 332 % can be achieved, respectively. This outcome is promising asmost MBR installations have capacities less than 1 MGD (3.8 MLd ) [40]. Increasingthe submergence depth (i.e. liquid level above the membrane) increases the amountof RAS that can be pumped for a given air flow.As mentioned in Section 2.1.6, the RAS in MBRs is usually between 300 and 500%of the incoming flow. This means that in small plants the pumping costs for waterconveyance can be significantly decreased or even completely eliminated by usingair escaping from the membrane tank for RAS pumping.70Chapter 7Comparison of model andmeasured results for multipleriser tubes in a pilot systemThe following chapter presents water flow data observed in the pilot system. Asdone for bench-scale conditions, the Nicklin model was used to model liquid flowsat pilot scale. To confirm the validity of the Nicklin model, a pilot-scale single risertube experiment was performed (as discussed in Section 4.2.1).7.1 Bundle of riser tubesThe water flows for experiments performed with the riser tube bundle for differentcollectors (rectangular and square), collector angles, submergence ratios (α) andair flows are discussed in the following sections. Results of the air flow calibrationare discussed in the Appendix (C).717.1.1 Influence of experimental conditions on square and rectangularcollector geometryAs discussed in Section 4.2.3, the experimental conditions investigated were twosubmergence ratios (0.44 and 0.65), three collector angles (0°, 6.7° and 13.2°) andthree air flow rates (81.2, 142.5 and 203.7 L/min).Figures 7.1 and 7.2 present the estimates of the total flow for a bundle of 19 risertubes, for the 2 submergence ratios tested. The total measured flow was calcu-lated based on an extrapolation of the water flows measured from the 7 individualriser tubes in the representative quadrant (see Section 4.2.3). The total modeledflow was calculated as the sum of the modeled flow determined using the Nicklinmodel, from each of the 19 individual riser tubes in a bundle, assuming that eachriser tube receives 1/19th of the total air to the collector.Note that for the experiments performed with the rectangular collector at the high-est air flow rate and the low collector angle (0 °), excessive amounts of air escapedbetween the frame and the tank, causing the setup to shake. For these conditions,no measurements were taken and are, therefore, not considered in the analysis. Asimilar behaviour was observed for higher collector angles (6.7 ° and 13.2 °); how-ever, the setup did not shake to the same extent and therefore measurements weretaken and used in the analysis.As expected, for both, the square and the rectangular collector, the low submer-gence ratio resulted in a lower water flow at a given air flow rate than the highsubmergence ratio. This trend is consistent with the observations made at bench72(a)(b)(c)Figure 7.1: Estimated total flow for bundle of riser tubes for rectangular and squarecollector.(Error bars correspond to estimated minimum and maximum values; a: low angle(0°), b: medium angle (6.7°), c: large angle (13.2°), Alpha = 0.44; the revised datapoint is based on calculations presented in Section 7.1.3; model results based inNicklin model)73(a)(b)(c)Figure 7.2: Estimated total flow for bundle of riser tubes for rectangular and squarecollector.(Error bars correspond to estimated minimum and maximum values; a: low angle(0°), b: medium angle (6.7°), c: large angle (13.2°), Alpha = 0.65; model resultsbased in Nicklin model)74scale (see Chapter 5).For the system with a square collector, the total water flow was affected by the col-lector angle. Generally, as the collector angle increased, so did the total water flow.For the system with a rectangular collector, the collector angle did not substantiallyaffect the water flow for different air flows or submergence ratios, except for thelower submergence ratio and low air flow. Under these conditions, the water flowwas highest when considering a low collector angle.For the lower submergence ratio considered, increasing the air flow generally re-sulted in an increase of the total water flow. This was observed for both collectors.However, the modeled increase was much greater than that which was measured.Also, no such correlation between air and water flow was observed for the highersubmergence ratio considered.In general, at a given air flow rate, higher water flows can be achieved with thesquare collector than with the rectangular collector.7.1.2 Evaluation of the model, comparison of square and rectangularcollector geometryIn order to assess whether the asymmetrical geometry of the rectangular collectorcontributed to an uneven distribution of air flow in the bundle of riser tubes, possi-bly affecting the fit of measured to modeled data, results of experiments of a squareand a rectangular collector were compared to the modeled data.75In general, the model more accurately predicts the measurements of the squarecollector, than those of the rectangular collector. For the lower submergence ratioconsidered, there was better agreement between the modeled and measured re-sults for the system with a square collector, especially at the higher air flows andcollector angles considered. For the higher submergence ratio considered, the mea-sured results were also closer to the modeled results for the system with a squarecollector. However, there was never consistent agreement between measured andmodeled results. At high submergence ratios, the model generally overpredictedthe water flows for both collectors. This trend is not consistent with the observa-tions made at bench scale (see Chapter 5), and will be discussed later in this section.For both collectors, at low submergence ratios and low air flow rates, the measuredwater flows were greater than the modeled water flows. This was not observed atthe higher submergence ratio. This may have resulted from the upflow of waterentrained by rising bubbles in the tank. This upward water flow contributes to themomentum of water entering and rising through the tubes, resulting in an increasedflow. At lower submergence ratio, there would be greater opportunity for the up-flow of entrained water than at a higher submergence ratio.As discussed in Section 7.1.1, the water flow increases with increasing submer-gence ratio for both collectors. However, the measured water flow rates for thehigh submergence ratio were significantly lower than what the model predicted forboth collectors. This suggests that a higher submergence ratio does increase thewater flow; however, the extent of increase cannot be predicted with the model76considered.The model results assume that the air flow is equally distributed to all riser tubes,and therefore, the flow from all riser tubes is the same. However, if this is not thecase, then some tubes could be receiving more or less air, which could be out of theoptimum range (as discussed in Section 6.3). To gain insight into the discrepancybetween the model and measured results, the distribution of water flows of the risertubes in the bundles was investigated in two ways, as described in the followingsections.7.1.3 Revised air flow based on water flowTo determine if the discrepancy between the model and measured results was dueto the uneven distribution of air flow into the different riser tubes in a bundle, theair flows of the different riser tubes were estimated based on the measured liquidflow of each tube (see Fig. 7.3). This was done for the measurements collectedfrom the setup with the rectangular collector frame at an air flow rate of 142.5 L/minand a low collector angle.77Figure 7.3: Water vs. air flow according to Nicklin model.(D=3/4 inch, length =0.54cm, lift = 0.3m; the numbers next to the lines representriser tube numbers.)Table 7.1: Extrapolation for air flow per riser tube in a bundle based on measuredresults of rectangular setup, using the Nicklin model.Riser tube number(see Fig. 4.3)Assumed to be repre-sentative of “n” risertubes in the system[n]Estimated air flowper riser tube fromFigure 7.3[L/min]1 1 52 4 63 2 6.84 4 55 4 6.56 2 97 2 10total: 19 126.6As presented in Table 7.1, the total air flow corresponding to the measured water78flow was estimated to be 126.6 L/min. The estimated total air flow is slightly lowerthan the measured flow of air added to the tank (142.5 L/min). The discrepancy islikely due to the gas that escapes the system between the collector angle frame andthe tank walls.When taking into account the uneven distribution of air flow into each riser tube,the model and measured results are in better agreement, as illustrated with an “x”in Figure 7.1.7.1.4 Distribution of water flowIn order to determine whether the shape of the collector had an impact on the dis-tribution of the water flow, water flows of individual riser tubes of both setupswere characterized. As discussed in Section 4.2.3, two submergence ratios (0.44and 0.65), three collector angles (0°, 6.7° and 13.2°) and three air flow rates (81.2,142.5 and 203.7 L/min) were investigated at pilot scale.Figures 7.4 and 7.5 present water flows of the 7 individual riser tubes in the rep-resentative quadrant (see Section 4.2.3), for the rectangular and square collector,respectively, at the low submergence ratio (α =0.44), similar results were observedfor the high submergence ratio (α = 0.65) and are presented in the Appendix B.7.As expected, the measured water flows were more evenly distributed for the squarecollector than the rectangular collector.Although the flow distribution was more homogenous for the square collector than79(a)(b)(c)Figure 7.4: Water flow distribution of individual riser tubes for the rectangularcollector (alpha =0.44).(Error bars correspond to minimum and maximum values; air flows of a: 81.2L/min; b: 142.5 L/min; c: 203.7 L/min)80(a)(b)(c)Figure 7.5: Water flow distribution of individual riser tubes for the square collector(alpha =0.44).(Error bars correspond to minimum and maximum values; air flows of a: 81.2L/min; b: 142.5 L/min; c: 203.7 L/min)81for the rectangular collector, the flow of some of the riser tubes (i.e. number 1 and2) were still less than expected based on the average and the model prediction, forboth submergence ratios. For both collectors, the submergence ratio had no impacton the flow distribution.As discussed in Section 7.1.1, the lowest collector angle resulted in the highesttotal water flow for low and medium air flow ranges for the rectangular collector.However, as presented in Figure 7.4 a, this configuration had the least homogenousdistribution of flows of individual riser tubes. In general, a higher collector angleresults in a better flow distribution for the rectangular collector. When the squarecollector was used, the largest collector angle resulted in the highest water flowindependent of the air flow and submergence ratio. The air flow rate did not affectthe distribution of air in the individual riser tubes for either collector.The results confirm that the shape of the collector has an impact on the distributionof the flow. The square collector resulted in a better distribution and, therefore,generally higher water flows than the rectangular collector. The results also indi-cate that, although the uneven distribution of air flow to the riser tubes contributesto the discrepancy between the measured and modeled results, other mechanismsalso likely contribute to this discrepancy. Further research is required to identifythe mechanisms responsible for the observed results.7.2 Water redirection systemAfter the water is lifted to the required height, it must be transferred into the RASpipes that transport it back to the head of the bioreactor tank. Four different ap-82proaches were investigated to direct the water flow from the vertical riser tubes tothe RAS pipes. The redirection of the lifted water flow was investigated for onlytwo conditions (i.e. two collector angles (6.7 and 13.2°) for one submergence ratioα = 0.44) and for the square collector.As illustrated in Figures 7.6 a and b, the geometry of the flow redirection systemhad a significant effect on the total water flow. The total flow collected when us-ing a simple 90° elbow was the lowest of the different options considered. Thisis likely because some of the lifted water simply hit the top of the elbow and wasredirected downward, rather than horizontally (see Fig. 7.6 a). Inserting flexibletubes in a standard 90° elbow to more effectively redirect the flow did increase thetotal collected flow. However, the total flow was much lower than expected. Whenthe elbow was filled with flexible tubes that were not directly connected to the risertubes, the water likely hit the hoses and was directed back into the riser tubes, in-stead of carried out (Fig.7.7 b).Configuration C and D , where curved tubes are fitted to each riser tube in a bun-dle, resulted in the highest total water flows. The total water flow measured forconfigurations C and D are similar to the total expected water flow (i.e. estimatedfrom the 19 individual riser tubes in a bundle).Although configurations C and D can provide similar water flows, the RAS pipingnetwork is expected to be simpler, and therefore, less costly to build and maintainusing configuration D. This reflects the fact that the flow from a bundle of risertubes is redirected and collected at one side of the bundle, as opposed to configu-83(a)(b)Figure 7.6: Effect of water collecting methods.(Error bars correspond to minimum and maximum values; a: medium ; b: largeangle. The expected value is based on an extrapolation from 7 individual risertubes as discussed in Section 7.1.1)84(a) Elbow (b) Tubes within elbow on top of riser tubesFigure 7.7: Illustrations of losses in water collection system.ration C, for which the water flow is redirected and collected at both sides of thebundle (therefore requiring more piping).85Chapter 8Conclusions and engineeringsignificance of work8.1 ConclusionsAn existing airlift pump model was modified, so that it accurately predicted waterflows for single riser tubes.Using the assumptions for the investigated conditions in the present work, the mod-ified model predicts that airlift pumps, using waste air, can convey RAS (returnativated sludge) flows of up to 332 % of the influent flow, indicating that wasteair from membrane sparging can be used to convey liquid within MBR (membranebioreactor) systems, using airlift pumps.Existing airlift pump models can provide an approximate estimate of the RAS flowthat can be achieved. However, they cannot be used to accurately predict the water86flow that can be achieved.The higher submergence ratio resulted in higher water flows. However, the extentof increase could not be accurately modeled.In general, better flow distribution in a bundle of riser tubes was observed athigher collector angles. However, there was no significant difference in water flowsachieved in systems with small and the large collector angle. .The collector with a square geometry resulted, not only in a better flow distributionin a bundle of riser tubes, but also in higher total water flows than the collector witha rectangular geometry.When redirecting the water from the top of the bundle of riser tubes to the RASpiping network pipes, fitting curved tubes to each riser tube in a bundle resulted inthe highest collected water flows.8.2 Engineering significanceIt was concluded that air exiting to the atmosphere from an MBR can be used toconvey RAS within MBR systems. However, existing airlift pump models couldnot accurately predict water flows that can be achieved.If electrically driven pumps can be completely replaced by airlift pumps, energycosts for an MBR can be decreased by up to 26% (If electrically driven pumpscan be replaced completely, see Fig. 1.1). Replacing electrically driven pumps by87airlift pumps in MBRs would result in lower operating costs.Since both, energy and water demand will increase in the future, solutions thatdecrease the energy demand for water treatment need to be identified. The presentstudy presented an opportunity to do so. However, it is crucial to have accuratemodels that can be used to design air lift RAS pumps. Such models should includefactors like the submergence ratio α , which will likely be given for each MBR dueto a limited space to submerge the riser tubes, and a required height that the waterneeds to be lifted to.8.3 Future workThe outcomes of the present study are specific to the conditions investigated andwere only intended to determine the feasibility of using air escaping from the mem-brane tank to lift and convey RAS. Although the results from the present study arepromising, additional research is required before airlift pumping can be imple-mented at commercial scale for RAS conveyance. The requirements of each indi-vidual MBR will be the key element of the design process, which can be addressedonce a model for this kind of application exists. Limitations can be addressed infuture work, as listed below.1. Further research is needed to develop models that can accurately predict theachievable water flow.2. Build and test several airlift pump applications with varying parameters, suchas the optimum number of riser tubes in a bundle, or the increments betweenthe heights of the individual riser tubes.883. Different types of aeration should be included in the experimental plan suchas cyclic or intermittent aeration.4. Sludge clogging in riser tubes and RAS transportation pipes can be a poten-tial problem. An investigation in to long term maintenance requirements forairlift pumping of RAS needs to be undertaken.5. The potential effect of the mixed liquor solids content/viscosity on the air-lifting of RAS should be addressed.89Bibliography[1] Engineeringtoolbox.com. URL http://www.engineeringtoolbox.com/surface-roughness-ventilation-ducts-d_209.html,May2014.[2] Meeting the challenges of the water-energy nexus: the role of reuse andwastewater treatment. Water 21, Magazine of the International WaterAssociation, April 2012.[3] Water and energy: conflicts and connections. Water 21, Magazine of theInternational Water Association, October 2012.[4] D. R. H. Beattie and P. Whalley. A simple two-phase frictional pressure dropcalculation method. Int. J. Multiphase Flow, 8:83–87, 1982.[5] T. Buer and J. Cumin. Mbr module design and operation. Desalination, 250:1073–1077, 2010.[6] F. D. Cachard and J. M. Delhaye. A slug-churn flow model forsmall-diameter airlift pumps. Int. J. Multiphase Flow, 22(4):627–649, 1996.[7] L. D. Chabalina, J. B. Ruiz, M. R. Pastor, and D. P. Rico. Influence of epsand mlss concentrations on mixed liquor physical parameters of twomembrane bioreactors. Desalination and water treatment, 46:46–59, 2012.[8] L. Chexal, G. Lellouche, G. Horowitz, and J. Haelzer. A void fractioncorrelation for generalized applications. Progress in Nuclear Energy, 27(4):255–295, 1992.[9] N. Clark and R. Dabolt. A general design equation for air lift pumpsoperating in slug flow. American Institue of Chemical Engineers, 32:56–64,1986.[10] R. Clift, J. Grace, and M. Weber. Bubbles, Drops and Particles. AcademicPress, INC., 1978.90[11] P. Côté, Z. Alam, and J. Penny. Hollow fibre membrane life in membranebioreactors (mbr). Desalination, 288:145–151, 2012.[12] P. Côté, T. Young, S. Smoot, and J. Peeters. Energy consumption of mbr formunicipal wastewater treatment: Current situation and potential. InProceedings of the Water Environment Federation, volume 14, pages479–492(14), 2013.[13] R. Davies and G. Taylor. The mechanics of large bubbles rising throughextended liquids and through liquids in tubes.pdf. Proceedings of the RoyalSociety of London, 200A(1062):375–390, 1950.[14] A. D. Delano. Design Analysis of the Einstein Refrigeration Cycle. PhDthesis, Georgia Institute of Technology, June 1998.[15] D. Dumitrescu. Stroemung an einer luftblase im senkrechten rohr.Zeitschrift fuer angewandte Mathematik und Mechanik, 23:139–149, 1943.[16] R. Field. Fundamentals of Fouling, in Membrane Technology: Membranesfor Water Treatment, volume 4 (eds K.-V. Peinemann and S. Pereira Nunes).Wiley-VCH Verlag GmbH und Co. KGaA, Weinheim, Germany, 2010.[17] GE Power and Water. Zeeweed 500d casette specification sheet. Technicalreport, General Electric Company, 2006.[18] GE Power and Water. Zeeweed* 500d module fact sheet. Technical report,General Electric Company, 2013.[19] R. Giles, J. Evett, and C. Liu. Theory and Problems of Fluid Mechanics andHydraulics. McGraw-Hill, Schaum’s Outline Series, third edition, 1994.[20] X. Huang, P. Gui, and Y. Qian. Effect of sludge retention time on microbialbehaviour in a submerged membrane bioreactor. Process Biochemistry, (36):1001–1006, 2001.[21] S. Jankhah. Effect of Bubble Size and Sparging Frequency on the PowerTransferred onto Membranes for Fouling Control. PhD thesis, University ofBritish Columbia, December 2013.[22] S. Judd and C. Judd. MBR Book - Principles and Applications of MembraneBioreactors for Water and Wastewater Treatment. Butterworth-Heinemann,2nd edition, 2011.91[23] S. Kassab, H. Kandil, H. Wards, and W. Ahmed. Air-lift pumpscharacteristics under two-phase flow conditions. International Journal ofHeat and Fluid Flow, 30:88–98, 2009.[24] W. Khongnakorn, C. Wisniewski, L. Pottier, and L. Vachoud. Physicalproperties of activated sludge in a submerged membrane bioreactor andrelation with membrane fouling. Separation and purification technology, 55:125–131, 2007.[25] D. Kouremenos and J. Staicos. Performance of a small air-lift pump.International Journal of Heat and Fluid Flow, 6:217–222, 1985.[26] R. Liu, X. Huang, Y. F. Sun, and Y. Quian. Hydrodynamic effect on sludgeaccumulation over membrane surfaces in a submerged membrane bioractor.Process Biochemistry, 39:157–163, 2003.[27] F. Meng, S. Chae, M. Kraume, A. Drews, F. Yang, and H. Shin. Recentadvances in membrane bioreactors (mbrs): Membrane fouling andmembrane material. Water research, 43(6):1489–1512, 2009.[28] Metcalf and Eddy. Wastewater Engineering Treatment and Reuse.McGraw-Hill, fourth edition, 2003.[29] D. Nicklin. The air-lift pump: Theory and optimization. Trans. Instn. Chem.Engrs., 41:29–39, 1963.[30] D. Nicklin, J. Wilkes, and D. J.F. Two-phase flow in vertical tubes. Trans.Instn. Chem. Engrs., 40:62–68, 1962.[31] J. Orkiszewski. Predicting two-phase pressure drops in vertical pipe. Journalof Petroleum Technology, pages 829–838, 1967.[32] M. C. Potter, D. C. Wiggert, and B. H. Ramadan. Mechanics of Fluids.Cengage learning, 4th si edition, 2012.[33] D. Reinemann, J. Parlange, and T. M.B. Theory of small-diameter airliftpumps. Int. J. Multiphase Flow, 16(1):113–122, 1990.[34] V. Samaras and D. Margaris. Two-phase flow regime maps for air-lift pumpvertical upward gas–liquid flow. International Journal of Multiphase Flow,31:757–766, 2005.[35] L. A. Schaefer. Single pressure absorption heat pump analysis. PhD thesis,Georgia Institute of Technology, 2000.92[36] A. Stenning and C. Martin. An analytical and experimental study of air-liftpump performance. ASME Journal of Engineering for Power, pages106–110, 1968.[37] N. Trixier, G. Guibaud, and M. Baudu. Determination of some rheologicalparameters for the characterization of activated sludge. Bioresourcetechnology, 90:215–220, 2003.[38] R. S. Trussella, R. P. Merlob, S. W. Hermanowiczc, and D. Jenkinsc.Influence of mixed liquor properties and aeration intensity on membranefouling in a submerged membrane bioreactor at high mixed liquor suspendedsolids concentrations. Water research, 41:947–958, 2007.[39] VDI Association. VDI-Waermeatlas. Springer, 10th edition, 2006.[40] Water Environment Federation. Membrane Bioreactors WEF Manual ofPratice No.36. Mc Graw Hill, 2011.[41] E. White and R. Beardmore. Velocity of rise of single cylindrical air bubblesthrough liquids contained in vertical tubes. Chemical Engineering Science,17:351–361, 1962.[42] S. J. White. Bubble pump design and performance. Master’s thesis, GeorgiaInstitute of Technology, August 2001.[43] T. Young, M. Muftugil, S. Smoot, and J. Peeters. Mbr vs. cas- capital andoperation cost evaluation. Water Practice and Technology, 7(04), 2012.[44] N. Zuber and J. Findlay. Average volumentric concentration in two-phaseflow systems. Journal of Heat Transfer, 87:453–468, 1965.[45] E. Zukoski. Influence of viscosity, surface tension and inclination angle onmotion of long bubbles in closed tubes. Journal of Fluid Mechanics, 25:821–832, 1966.93Appendix ASludge viscosity calculationsAs the MLSS concentration in the membrane tank was assumed to be between8000 and 12000 mg/L, these two values were considered as the boundaries for theviscosity calculations. To calculate the viscosity, a correlation by Liu et al. [26]was used:η = 1.61e0.07MLSS. (A.0.1)Note that while for MLSS concentrations below 9 g/L, similar correlations werefound [7, 37]. However, for MLSS concentrations greater than 10 g/L, viscositiesin a wider range were reported (i.e. viscosities up to 46 mPas at 10 g/L [38], oras low as 5 mPas at 15 g/L [24]). When an MBR is run at the upper boundaryMLSS concentration, the viscosity in the membrane tank should be investigated.Since the viscosity does not only depend on the MLSS concentration but also onthe composition of the sludge, the floc size and shape and bound water [24, 37],the appropriateness of the sludge viscosity for the application of the present study94should be investigated.Using Equation A.0.1, the viscosity is between approximately 280 to 370% higherthan the viscosity of pure water (see Table A.1), which was assumed when mod-elling the airlift pump (assuming the viscosity of water is 1.0016 mPas (at 20 °C)[39]).Clift et al. [10] presented a correlation between the Eötvös number Eo, the Froudenumber Fr and the Morton number M, which is presented in Figure A.1. Theparameters needed for this correlation for the lower and upper boundaries assumedcan be found in Table A.1.Table A.1: Parameters to determine the impact of the sludge viscosity on bubblebehaviour.Parameter Unit Lower Boundary Upper BoundaryMLSS mg/L 8000 12000Mo - 1.6∗10−9 4.9∗10−9η mPas 2.82 3.73Eo, Fr and M defined as follows:Eo =g∆ρD2σ (A.0.2)Fr =Vg√ρ∆ρ gD (A.0.3)Mo =gη4∆ρρ2σ3 (A.0.4)95Figure A.1: Froude number vs. Eötvös number [10].For surface tension, the value of water and air at 20°C was assumed (72.74 ∗10−3N/m [39]). The Eötvös number results in Eo = 48.9 for both, the lower andupper boundary. For calculated values for the Morton number of 1.6 ∗ 10−9 and4.9 ∗ 10−9, respectively, Figure A.1 suggests, that bubbles form independently ofviscosity in this range; hence, the assumption of the viscosity of water is justified.96Appendix BResultsResults Models Before Modification97(a) (b)Figure B.1: Results using riser tubes with D=3/4 inch, 1m long, before modifica-tion.(Error bars correspond to minimum and maximum values; a: Lift 0.6m; b: Lift0.7m)98(a) (b) (c)Figure B.2: Results using riser tubes with D=3/4 inch, 2m long, before modification.(Error bars correspond to minimum and maximum values; a: Lift 1m; b: Lift 1.1m; c: Lift 1.2m)99(a) (b)Figure B.3: Results using riser tubes with D=1 inch, 1m long, before modification.(Error bars correspond to minimum and maximum values; a: Lift 0.5m; b: Lift0.55m)100(a) (b) (c)Figure B.4: Results using riser tubes with D=1 inch, 2m long, before modification.(Error bars correspond to minimum and maximum values; a: Lift 1m; b: Lift 1.1m; c: Lift 1.2m)101Results of Evaluation of Head Loss Coefficient102(a) (b) (c)Figure B.5: Evaluation of head loss coefficient using riser tubes with D=3/4 inch, 2m long.(Error bars correspond to minimum and maximum values; a: Lift 1m; b: Lift 1.1m; c: Lift 1.2m)103Water Distribution104(a)(b)(c)Figure B.6: Water flow distribution of individual riser tubes for the rectangularsetup (alpha =0.64).(Error bars correspond to minimum and maximum values; air flows of a: 81.2L/min; b: 142.5 L/min; c: 203.7 L/min)105(a)(b)(c)Figure B.7: Water flow distribution of individual riser tubes for the square setup(alpha =0.64).(Error bars correspond to minimum and maximum values; air flows of a: 81.2L/min; b: 142.5 L/min; c: 203.7 L/min)106Appendix CAir flow calibration in tank usedfor water collection systemAir flow was calibrated as described in Section 4.2.2. Figure C.1 illustrates whereair flow measurements were taken. Detailed results can be found in Figure C.1 andthe calibration curve is presented in Figure C.2.Figure C.1: Representative locations of air flow measurements (top view of tank).107Figure C.2: Calibration curve for the air flowmeter attached to the tank used forthe water collection system.(Error bars correspond to average minimum and maximum values.)108Table C.1: Air flow measurements for calibration of pilot scale tankFlow on scale Air measured Locationcfm [L] 1a 1b 1c 2a 2b 2c 3a 3b 3c5 0.5 25 28 28 25 17 25 21 24 3228 29 27 23 21 20 25 25 26 Total air27 27 23 26 18 22 20 35 23 AverageAverage [s] 26.67 28.00 26.00 24.67 18.67 22.33 22 28 27 [L/min] [L/min]Air flow [L/min] 1.13 1.07 1.15 1.22 1.61 1.34 1.36 1.07 1.11 1.23 93.547 0.5 17.5 17 17 14 13 14 13 15 1016 13 18 14 12 14 17 12.5 13 Total air18 20 16 18 15 13 16 20 15 AverageAverage [s] 17.17 16.67 17.00 15.33 13.33 13.67 15.33 15.83 12.67 [L/min] [L/min]Air flow [L/min] 1.75 1.80 1.76 1.96 2.25 2.20 1.96 1.89 2.37 1.99 151.639 1 18 22 18 17 26 23 15 22 2020 26 17 18 22 20 20 26 17 Total air18 20 20 20 22 23 17 23 15 AverageAverage [s] 18.67 22.67 18.33 18.33 23.33 22.00 17.33 23.67 17.33 [L/min] [L/min]Air flow [L/min] 3.21 2.65 3.27 3.27 2.57 2.73 3.46 2.54 3.46 3.02 229.67109Appendix DAchievable RAS flows for allMBR treatment capacities,scenarios and options considered110Table D.1: Scenario A, 0.75 MGD (2.84 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.53 0.2 7.00 8.00 9.00 10.00 11.00 12.00 18.00Water flow per riser tube [L/min] 0.28 0.58 0.84 1.07 1.27 1.45 2.19Number of riser tubes 1373.00 1201.00 1068.00 961.00 873.00 801.00 534.00Total RAS flow above module [L/min] 390.81 697.82 898.32 1027.55 1108.70 1159.42 1167.54The optimum is presented in bold.111Table D.2: Scenario A, 0.75 MGD (2.84 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.53 0.2 8.00 9.00 10.00 11.00 12.00 18.00Water flow per riser tube [L/min] 0.58 0.84 1.07 1.27 1.45 2.19Number of riser tubes 948.00 843.00 759.00 690.00 633.00 421.00Total RAS flow above module [L/min] 550.82 709.07 811.56 876.29 916.25 920.47Air on sides, option I, air per riser tube [L/min]0.83 0.5 4.07 5.20 6.33 7.47 8.60 9.73Water flow per riser tube [L/min] 1.76 2.32 2.74 3.07 3.34 3.57Number of riser tubes 236.00 184.00 151.00 128.00 111.00 98.00Total RAS flow from sides [L/min] 416.43 426.96 413.90 393.32 371.06 349.65Air on sides, option II, air per riser tube [L/min]0.93 0.6 2.00 3.00 5.00 7.00 13.66 24.45Water flow per riser tube [L/min] 0.32 1.48 2.24 2.77 3.17 3.49Number of riser tubes 480.00 320.00 192.00 137.00 70.00 39.00Total RAS flow from sides [L/min] 155.80 474.14 429.18 379.23 221.92 136.03Total RAS option I [L/min] 1347.43Total RAS option II [L/min] 1394.61The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.112Table D.3: Scenario A, 0.75 MGD (2.84 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.53 0.2 7.00 8.00 9.00 10.00 11.00 12.00Water flow per riser tube [L/min] 0.28 0.58 0.84 1.07 1.27 1.45Number of riser tubes 1029.00 901.00 801.00 720.00 655.00 600.00Total RAS flow above module [L/min] [L/min] 292.89 523.51 673.74 769.86 831.84 868.48Air on sides, option I, air per riser tube [L/min]0.83 0.5 2.93 4.07 5.20 6.33 7.47Water flow per riser tube [L/min] 0.99 1.76 2.32 2.74 3.07Number of riser tubes 819.00 590.00 462.00 379.00 321.00Total RAS flow from sides [L/min] 810.29 1041.08 1072.05 1038.85 986.38Air on sides, option II, air per riser tube [L/min]0.93 0.6 2.00 3.00 4.00 5.00 6.00Water flow per riser tube [L/min] 0.32 1.48 2.24 2.77 3.17Number of riser tubes 1201.00 801.00 600.00 480.00 400.00Total RAS flow from sides [L/min] 389.83 1186.82 1341.19 1328.68 1268.13Total RAS option I [L/min] 1940.53Total RAS option II [L/min] 2209.67The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.113Table D.4: Scenario A, 1 MGD (3.79 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.62 0.2 11.33 13.67 16.00 18.33 20.67 23.00 25.33 27.67Water flow per riser tube [L/min] 0.45 0.84 1.16 1.41 1.61 1.78 1.93 2.05Number of riser tubes 1113.00 923.00 788.00 688.00 610 548 498 456Total RAS flow above module [L/min] 504.92 779.91 910.43 966.90 982.27 975.96 958.71 934.10The optimum is presented in bold.114Table D.5: Scenario A, 1 MGD (3.79 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.62 0.2 11.33 13.67 16.00 18.33 20.67 23.00 25.33 27.67Water flow per riser tube [L/min] 0.45 0.84 1.16 1.41 1.61 1.78 1.93 2.05Number of riser tubes 1001.00 830.00 709.00 619.00 549.00 493.00 448.00 410.00Total RAS flow above module [L/min] 454.11 701.33 819.15 869.93 884.05 878.01 862.46 839.87Air on sides, option I, air per riser tube [L/min]0.92 0.5 5.00 6.00 7.00 8.00 9.00 10.00 11.00Water flow per riser tube [L/min] 0.38 0.80 1.15 1.45 1.71 1.93 2.13Number of riser tubes 252.00 210.00 180.00 157.00 140.00 126.00 114.00Total RAS flow from sides [L/min] 94.72 167.17 207.33 228.20 239.58 243.64 242.50Air on sides, option II, air per riser tube [L/min]1.02 0.6 2.00 3.00 5.00 7.00 13.66 24.45Water flow per riser tube [L/min] 0.41 0.93 1.36 1.71 2.00 2.25Number of riser tubes 630.00 420.00 252.00 180.00 92.00 51.00Total RAS flow from sides [L/min] 255.36 389.77 341.71 307.53 184.25 114.85Total RAS option I [L/min] 1127.68Total RAS option II [L/min] 1273.81The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.115Table D.6: Scenario A, 1 MGD (3.79 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.62 0.2 11.33 13.67 16.00 18.33 20.67 23.00 25.33Water flow per riser tube [L/min] 0.45 0.84 1.16 1.41 1.61 1.78 1.93Number of riser tubes 834.00 692.00 591.00 516.00 457.00 411.00 373.00Total RAS flow above module [L/min] 378.35 584.72 682.82 725.18 735.90 731.97 718.07Air on sides, option I, air per riser tube [L/min]0.92 0.5 5.00 6.00 7.00 8.00 9.00 10.00Water flow per riser tube [L/min] 0.38 0.80 1.15 1.45 1.71 1.93Number of riser tubes 630.00 525.00 450.00 394.00 350.00 315.00Total RAS flow from sides [L/min] 236.80 417.92 518.31 572.68 598.95 609.09Air on sides, option II, air per riser tube [L/min]1.02 0.6 4.00 5.00 6.00 7.00 8.00Water flow per riser tube [L/min] 0.41 0.93 1.36 1.71 2.00Number of riser tubes 788.00 630.00 525.00 450.00 394.00Total RAS flow from sides [L/min] 319.40 584.65 711.90 768.83 789.09Total RAS option I [L/min] 1344.99Total RAS option II [L/min] 1524.99The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.116Table D.7: Scenario A, 2 MGD (7.57 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.99 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]117Table D.8: Scenario A, 2 MGD (7.57 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.99 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]1.29 0.5 18.33 18.67 19.00 19.33 19.67 20.00 20.33Water flow per riser tube [L/min] 2.40 2.43 2.46 2.49 2.52 2.54 2.57Number of riser tubes 137.00 135.00 132.00 130.00 128.00 126.00 124.00Total RAS flow from sides [L/min] 329.19 328.34 324.80 323.48 321.95 320.23 318.31Air on sides, option II, air per riser tube [L/min]1.39 0.6 18.33 18.67 19.00 19.33 19.67 20.00Water flow per riser tube [L/min] 2.93 2.96 2.99 3.01 3.04 3.07Number of riser tubes 137.00 135.00 132.00 130.00 128.00 126.00Total RAS flow from sides [L/min] 401.91 399.81 394.50 391.95 389.20 386.26Total RAS option I [L/min] 329.19Total RAS option II [L/min] 401.91The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.118Table D.9: Scenario A, 2 MGD (7.57 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.99 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]1.29 0.5 18.33 18.67 19.00 19.33 19.67Water flow per riser tube [L/min] 2.40 2.43 2.46 2.49 2.52Number of riser tubes 344.00 337.00 332.00 326.00 320.00Total RAS flow from sides [L/min] 826.57 819.63 816.92 811.19 804.88Air on sides, option II, air per riser tube [L/min]1.39 0.6 18.33 18.67 19.00 19.33 19.67 20.00Water flow per riser tube [L/min] 2.93 2.96 2.99 3.01 3.04 3.07Number of riser tubes 344.00 337.00 332.00 326.00 320.00 315.00Total RAS flow from sides [L/min] 1009.18 998.04 992.23 982.89 973.00 965.66Total RAS option I [L/min] 826.57Total RAS option II [L/min] 1009.18The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.119Table D.10: Scenario A, 5 MGD (18.93 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]1.63 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]120Table D.11: Scenario A, 5 MGD (18.93 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]1.63 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]1.93 0.5 18.00 18.33 18.67 19.00 19.33 19.67 20.00Water flow per riser tube [L/min] 0.22 0.26 0.30 0.33 0.37 0.40 0.43Number of riser tubes 350.00 344.00 337.00 332.00 326.00 320.00 315.00Total RAS flow from sides [L/min] 77.26 88.90 99.46 109.88 119.28 127.98 136.45Air on sides, option II, air per riser tube [L/min]2.03 0.6 18.00 18.33 18.67 19.00 19.33 19.67Water flow per riser tube [L/min] 0.98 1.02 1.05 1.09 1.12 1.15Number of riser tubes 350.00 344.00 337.00 332.00 326.00 320.00Total RAS flow from sides [L/min] 344.59 350.64 354.90 360.56 364.50 367.78Total RAS option I [L/min] 136.45Total RAS option II [L/min] 367.78The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.121Table D.12: Scenario A, 5 MGD (18.93 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]1.63 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]1.93 0.5 18.00 18.33 18.67 19.00 19.33Water flow per riser tube [L/min] 0.22 0.26 0.30 0.33 0.37Number of riser tubes 876.00 860.00 844.00 830.00 815.00Total RAS flow from sides [L/min] 193.38 222.24 249.10 274.69 298.19Air on sides, option II, air per riser tube [L/min]2.03 0.6 18.00 18.33 18.67Water flow per riser tube [L/min] 0.98 1.02 1.05Number of riser tubes 876.00 860.00 844.00Total RAS flow from sides [L/min] 862.46 876.59 888.83Total RAS option I [L/min] 298.19Total RAS option II [L/min] 888.83The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.122Table D.13: Scenario A, 10 MGD (37.85 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]123Table D.14: Scenario A, 10 MGD (37.85 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]2.97 0.5 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 0.09 0.31 0.47 0.58Number of riser tubes 187.00 161.00 141.00 126.00Total RAS flow from sides [L/min] 16.72 50.48 66.01 73.51Air on sides, option II, air per riser tube [L/min]3.07 0.6 56.61 67.35 78.10 88.85 99.6Water flow per riser tube [L/min] 1.03 1.20 1.33 1.43 1.509892Number of riser tubes 222.00 187.00 161.00 141.00 126Total RAS flow from sides [L/min] 227.71 225.24 214.84 202.04 190.2464Total RAS option I [L/min] 73.51Total RAS option II [L/min] 227.71The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.124Table D.15: Scenario A, 10 MGD (37.85 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]2.97 0.5 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 0.09 0.31 0.47 0.58Number of riser tubes 468.00 403.00 354.00 316.00Total RAS flow from sides [L/min] 41.85 126.35 165.74 184.35Air on sides, option II, air per riser tube [L/min]3.07 0.6 56.61 67.35 78.10Water flow per riser tube [L/min] 1.03 1.20 1.33Number of riser tubes 557.00 468.00 403.00Total RAS flow from sides [L/min] 571.33 563.71 537.78Total RAS option I [L/min] 184.35Total RAS option II [L/min] 571.33The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.125Table D.16: Scenario B, 0.5 MGD (1.83 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.36 0.2 3.90 5.13 6.35 7.57 8.79 10.01 11.24 12.46Water flow per riser tube [L/min] 1.07 1.70 2.17 2.53 2.82 3.06 3.26 3.43Number of riser tubes 1615.00 1230.00 993.00 833.00 717.00 629.00 561.00 506.00Total RAS flow above module [L/min] 1734.00 2093.07 2153.66 2108.02 2022.13 1923.61 1827.55 1734.38The optimum is presented in bold.126Table D.17: Scenario B, 0.5 MGD (1.83 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.36 0.2 2.68 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.20 1.07 1.70 2.17 2.53 2.82 3.06Number of riser tubes 2116.00 1454.00 1107.00 894.00 749.00 647.00 566.00Total RAS flow above module [L/min] 427.98 1561.14 1883.77 1938.94 1895.45 1824.71 1730.94Air on sides, option I, air per riser tube [L/min]0.66 0.5 2.68 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 2.49 3.25 3.76 4.14 4.44 4.68 4.88Number of riser tubes 235.00 161.00 123.00 99.00 83.00 71.00 62.00Total RAS flow from sides [L/min] 585.11 522.53 462.27 409.75 368.24 332.13 302.46Air on sides, option II, air per riser tube [L/min]0.76 0.6 2.68 3.90 5.13 6.35 7.57 8.79Water flow per riser tube [L/min] 2.84 3.56 4.05 4.42 4.71 4.94Number of riser tubes 235.00 161.00 123.00 99.00 83.00 71.00Total RAS flow from sides [L/min] 666.58 572.75 498.24 437.47 390.77 350.96Total RAS option I [L/min] 2524.05Total RAS option II [L/min] 2605.52The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.127Table D.18: Scenario B, 0.5 MGD (1.83 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.36 0.2 2.68 3.90 5.13 6.35 7.57 8.79Water flow per riser tube [L/min] 0.20 1.07 1.70 2.17 2.53 2.82Number of riser tubes 1764.00 1211.00 922.00 745.00 624.00 538.00Total RAS flow above module [L/min] 356.78 1300.23 1568.95 1615.79 1579.12 1517.30Air on sides, option I, air per riser tube [L/min]0.66 0.5 2.68 3.90 5.13 6.35 7.57Water flow per riser tube [L/min] 2.49 3.25 3.76 4.14 4.44Number of riser tubes 588.00 403.00 307.00 248.00 208.00Total RAS flow from sides [L/min] 1464.02 1307.94 1153.80 1026.45 922.82Air on sides, option II, air per riser tube [L/min]0.76 0.6 2.00 3.00 5.00Water flow per riser tube [L/min] 2.84 3.06 4.00Number of riser tubes 788.00 525.00 315.00Total RAS flow from sides [L/min] 2235.17 1604.61 1258.74Total RAS option I [L/min] 3079.80Total RAS option II [L/min] 3850.96The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.128Table D.19: Scenario B, 0.75 MGD (2.84 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.43 0.2 3.90 5.13 6.35 7.57 8.79 10.01 11.24 12.46Water flow per riser tube [L/min] 0.12 0.74 1.23 1.62 1.94 2.20 2.42 2.61Number of riser tubes 2462.00 1875.00 1514.00 1269.00 1093.00 959.00 855.00 771.00Total RAS flow above module [L/min] 299.59 1388.52 1864.67 2059.22 2120.03 2111.11 2070.15 2011.28The optimum is presented in bold.129Table D.20: Scenario B, 0.75 MGD (2.84 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.43 0.2 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.12 0.74 1.23 1.62 1.94 2.20Number of riser tubes 1944.00 1481.00 1195.00 1002.00 865.00 758.00Total RAS flow above module [L/min] 236.55 1096.74 1471.78 1625.96 1677.79 1668.63Air on sides, option I, air per riser tube [L/min]0.73 0.5 2.68 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 1.75 2.58 3.14 3.56 3.88 4.14 4.35Number of riser tubes 358.00 246.00 187.00 151.00 126.00 109.00 95.00Total RAS flow from sides [L/min] 626.41 635.38 587.99 537.21 488.75 451.04 413.47Air on sides, option II, air per riser tube [L/min]0.83 0.6 2.00 3.00 5.00 7.00 13.66 24.45Water flow per riser tube [L/min] 2.18 2.97 3.51 3.90 4.21 4.46Number of riser tubes 480.00 320.00 192.00 137.00 70.00 39.00Total RAS flow from sides [L/min] 1045.32 950.84 673.49 534.90 294.98 174.11Total RAS option I [L/min] 2313.17Total RAS option II [L/min] 2723.11The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.130Table D.21: Scenario B, 0.75 MGD (2.84 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.43 0.2 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.12 0.74 1.23 1.62 1.94 2.20Number of riser tubes 1846.00 1406.00 1135.00 952.00 819.00 719.00Total RAS flow above module [L/min] 224.63 1041.20 1397.88 1544.82 1588.56 1582.78Air on sides, option I, air per riser tube [L/min]0.73 0.5 2.68 3.90 5.13 6.35 7.57Water flow per riser tube [L/min] 1.75 2.58 3.14 3.56 3.88Number of riser tubes 896.00 615.00 468.00 378.00 317.00Total RAS flow from sides [L/min] 1567.78 1588.46 1471.56 1344.81 1229.63Air on sides, option II, air per riser tube [L/min]0.83 0.6 2.00 3.00 4.00Water flow per riser tube [L/min] 2.18 2.97 3.51Number of riser tubes 1201.00 801.00 600.00Total RAS flow from sides [L/min] 2615.48 2380.08 2104.65Total RAS option I [L/min] 3177.03Total RAS option II [L/min] 4204.05The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.131Table D.22: Scenario B, 1 MGD (3.79 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.52 0.2 6.35 7.57 8.79 10.01 11.24 12.46 13.68 15.00Water flow per riser tube [L/min] 0.14 0.53 0.86 1.14 1.38 1.59 1.77 1.94Number of riser tubes 1987.00 1666.00 1435.00 1259.00 1122.00 1012.00 922.00 841.00Total RAS flow above module [L/min] 279.49 884.18 1237.74 1440.67 1553.78 1610.95 1633.38 1631.88The optimum is presented in bold.132Table D.23: Scenario B, 1 MGD (3.79 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.52 0.2 6.35 7.57 8.79 10.01 11.24 12.46 13.68 15.00Water flow per riser tube [L/min] 0.14 0.53 0.86 1.14 1.38 1.59 1.77 1.94Number of riser tubes 1788.00 1499.00 1291.00 1133.00 1010.00 911.00 830.00 757.00Total RAS flow above module [L/min] 251.50 795.55 1113.54 1296.49 1398.68 1450.17 1470.40 1468.88Air on sides, option I, air per riser tube [L/min]0.82 0.5 2.68 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.84 1.73 2.34 2.80 3.15 3.43 3.67Number of riser tubes 470.00 323.00 246.00 198.00 166.00 143.00 125.00Total RAS flow from sides [L/min] 392.51 558.17 576.76 554.16 523.01 490.94 458.25Air on sides, option II, air per riser tube [L/min]0.92 0.6 2.00 3.00 5.00 7.00 13.66 24.45Water flow per riser tube [L/min] 1.33 2.20 2.79 3.22 3.56 3.83Number of riser tubes 630.00 420.00 252.00 180.00 92.00 51.00Total RAS flow from sides [L/min] 835.76 923.08 702.71 580.12 327.51 195.38Total RAS option I [L/min] 2047.16Total RAS option II [L/min] 2393.48The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.133Table D.24: Scenario B, 1 MGD (3.79 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.52 0.2 6.35 7.57 8.79 11.24 12.46 13.68Water flow per riser tube [L/min] 0.14 0.53 0.86 1.38 1.59 1.77Number of riser tubes 1490.00 1249.00 1076.00 842.00 759.00 691.00Total RAS flow above module [L/min] 209.58 662.87 928.09 1166.03 1208.21 1224.15Air on sides, option I, air per riser tube [L/min]0.82 0.5 2.68 3.90 5.13 6.35 7.57Water flow per riser tube [L/min] 0.84 1.73 2.34 2.80 3.15Number of riser tubes 1176.00 807.00 615.00 496.00 416.00Total RAS flow from sides [L/min] 982.12 1394.57 1441.91 1388.21 1310.67Air on sides, option II, air per riser tube [L/min]0.92 0.6 2.00 3.00 4.00Water flow per riser tube [L/min] 1.33 2.20 2.79Number of riser tubes 1577.00 1051.00 788.00Total RAS flow from sides [L/min] 2092.06 2309.90 2197.37Total RAS option I [L/min] 2666.06Total RAS option II [L/min] 3534.05The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.134Table D.25: Scenario B, 2 MGD (7.57 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.89 0.2 24.36 35.11 45.86 56.61 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 0.32 0.94 1.29 1.51 1.66 1.78 1.86 1.93Number of riser tubes 1035.00 718.00 550.00 445.00 374.00 323.00 283.00 253.00Total RAS flow above module [L/min] 334.56 674.12 708.04 671.94 622.31 573.77 526.91 488.00The optimum is presented in bold.135Table D.26: Scenario B, 2 MGD (2.84 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.89 0.2 24.36 35.11 45.86 56.61 67.35 78.10 88.85Water flow per riser tube [L/min] 0.32 0.94 1.29 1.51 1.66 1.78 1.86Number of riser tubes 932.00 646.00 495.00 401.00 337.00 290.00 255.00Total RAS flow above module [L/min] 301.26 606.52 637.24 605.50 560.75 515.15 474.78Air on sides, option I, air per riser tube [L/min]1.19 0.5 6.35 7.57 8.79 10.01 11.24 12.46 13.68Water flow per riser tube [L/min] 0.68 1.08 1.41 1.69 1.93 2.13 2.31Number of riser tubes 397.00 333.00 287.00 251.00 224.00 202.00 184.00Total RAS flow from sides [L/min] 268.27 359.21 405.57 424.98 432.39 430.99 424.99Air on sides, option II, air per riser tube [L/min]1.29 0.6 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.12 0.75 1.25 1.65 1.97 2.24Number of riser tubes 646.00 492.00 397.00 333.00 287.00 251.00Total RAS flow from sides [L/min] 78.70 366.85 494.42 548.14 566.03 562.88Total RAS option I [L/min] 1069.62Total RAS option II [L/min] 1203.27The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.136Table D.27: Scenario B, 2 MGD (2.84 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.89 0.2 24.36 35.11 45.86 56.61 67.35 78.10 88.85Water flow per riser tube [L/min] 0.32 0.94 1.29 1.51 1.66 1.78 1.86Number of riser tubes 776.00 539.00 412.00 334.00 280.00 242.00 212.00Total RAS flow above module [L/min] [L/min] 250.84 506.06 530.39 504.33 465.90 429.89 394.72Air on sides, option I, air per riser tube [L/min]1.19 0.5 6.35 7.57 8.79 10.01 11.24Water flow per riser tube [L/min] 0.68 1.08 1.41 1.69 1.93Number of riser tubes 993.00 833.00 717.00 629.00 561.00Total RAS flow from sides [L/min] 671.01 898.58 1013.22 1064.98 1082.90Air on sides, option II, air per riser tube [L/min]1.29 0.6 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.12 0.75 1.25 1.65 1.97 2.24Number of riser tubes 1615.00 1230.00 993.00 833.00 717.00 629.00Total RAS flow from sides [L/min] 196.75 917.13 1236.68 1371.17 1414.10 1410.57Total RAS option I [L/min] 1613.29Total RAS option II [L/min] 1944.49The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.137Table D.28: Scenario B, 5 MGD (18.93 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]138Table D.29: Scenario B, 5 MGD (18.93 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]1.53 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]1.83 0.5 13.61 24.36 35.11 45.86 56.61 67.35 78.10Water flow per riser tube [L/min] 0.33 1.43 1.95 2.25 2.44 2.57 3.21Number of riser tubes 463.00 258.00 179.00 137.00 111.00 93.00 80.00Total RAS flow from sides [L/min] 151.12 368.27 348.60 307.64 270.58 239.12 256.61Air on sides, option II, air per riser tube [L/min]1.93 0.6 13.61 24.36 35.11 45.86 56.61 67.35Water flow per riser tube [L/min] 1.02 2.06 2.54 2.82 2.99 3.12Number of riser tubes 463.00 258.00 179.00 137.00 111.00 93.00Total RAS flow from sides [L/min] 472.17 530.75 454.54 385.74 332.32 289.95Total RAS option I [L/min] 368.27Total RAS option II [L/min] 530.75The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.139Table D.30: Scenario B, 5 MGD (18.93 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]1.53 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]1.83 0.5 13.61 24.36 35.11 45.86 56.61Water flow per riser tube [L/min] 0.33 1.43 1.95 2.25 2.44Number of riser tubes 1158.00 647.00 449.00 343.00 278.00Total RAS flow from sides [L/min] 377.96 923.53 874.43 770.21 677.66Air on sides, option II, air per riser tube [L/min]1.93 0.6 13.61 24.36 35.11Water flow per riser tube [L/min] 1.02 2.06 2.54Number of riser tubes 1158.00 647.00 449.00Total RAS flow from sides [L/min] 1180.93 1331.00 1140.17Total RAS option I [L/min] 923.53Total RAS option II [L/min] 1331.00The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.140Table D.31: Scenario B, 10 MGD (37.85 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]141Table D.32: Scenario B, 10 MGD (37.85 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]2.57 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]2.87 0.5 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 0.09 0.31 0.47 0.58Number of riser tubes 187.00 161.00 141.00 126.00Total RAS flow from sides [L/min] 16.72 50.48 66.01 73.51Air on sides, option II, air per riser tube [L/min]2.97 0.6 56.61 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 1.03 1.20 1.33 1.43 1.51Number of riser tubes 222.00 187.00 161.00 141.00 126.00Total RAS flow from sides [L/min] 227.71 225.24 214.84 202.04 190.25Total RAS option I [L/min] 73.51Total RAS option II [L/min] 227.71The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.142Table D.33: Scenario B, 10 MGD (37.85 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]2.57 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]2.87 0.5 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 0.09 0.31 0.47 0.58Number of riser tubes 468.00 403.00 354.00 316.00Total RAS flow from sides [L/min] 41.85 126.35 165.74 184.35Air on sides, option II, air per riser tube [L/min]2.97 0.6 56.61 67.35 78.10Water flow per riser tube [L/min] 1.03 1.20 1.33Number of riser tubes 557.00 468.00 403.00Total RAS flow from sides [L/min] 571.33 563.71 537.78Total RAS option I [L/min] 184.35Total RAS option II [L/min] 571.33The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.143Table D.34: Scenario C, 0.5 MGD (1.83 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.46 0.3 2.00 3.00 4.00 6.00 8.00 10.00 13.66 24.45Water flow per riser tube [L/min] 0.61 1.56 2.21 3.05 3.58 3.97 4.45 5.16Number of riser tubes 3154.00 2102.00 1577.00 1051.00 788.00 630.00 461.00 258.00Total RAS flow above module [L/min] 1928.36 3269.03 3481.07 3205.34 2824.03 2502.74 2049.69 1331.28The optimum is presented in bold.144Table D.35: Scenario C, 0.5 MGD (1.83 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.46 0.3 4.00 7.00 10.00 13.66 17.00 24.45Water flow per riser tube [L/min] 2.21 3.35 3.96 4.45 4.74 5.16Number of riser tubes 1419.00 811.00 567.00 415.00 334.00 232.00Total RAS flow above module [L/min] 3132.30 2713.28 2247.02 1845.17 1583.16 1196.77Air on sides, option I, air per riser tube [L/min]0.66 0.5 2.00 3.00 4.00 5.00 7.00 13.66 24.45Water flow per riser tube [L/min] 1.84 2.72 3.29 3.71 4.31 4.55 5.45Number of riser tubes 315.00 210.00 157.00 126.00 90.00 46.00 25.00Total RAS flow from sides [L/min] 579.29 571.41 516.88 467.74 387.99 209.15 136.29Air on sides, option II, air per riser tube [L/min]0.76 0.6 2.00 3.00 5.00 7.00 13.66 24.45Water flow per riser tube [L/min] 2.22 3.06 4.00 4.59 5.57 6.38Number of riser tubes 315.00 210.00 126.00 90.00 46.00 25.00Total RAS flow from sides [L/min] 699.30 641.84 503.50 412.72 256.35 159.60Total RAS option I [L/min] 3711.59Total RAS option II [L/min] 3831.60The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.145Table D.36: Scenario C, 0.5 MGD (1.83 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.46 0.3 4.00 10.00 13.66 17.00 24.45Water flow per riser tube [L/min] 2.21 3.96 4.45 4.74 5.16Number of riser tubes 1182.00 473.00 346.00 278.00 193.00Total RAS flow above module [L/min] 2609.15 1874.50 1538.38 1317.72 995.59Air on sides, option I, air per riser tube [L/min]0.66 0.5 2.00 3.00 4.00 5.00 7.00Water flow per riser tube [L/min] 1.84 2.72 3.29 3.71 4.31Number of riser tubes 788.00 525.00 394.00 315.00 225.00Total RAS flow from sides [L/min] 1449.13 1428.53 1297.13 1169.34 969.98Air on sides, option II, air per riser tube [L/min]0.76 0.6 2.00 3.00 5.00Water flow per riser tube [L/min] 2.22 3.06 4.00Number of riser tubes 788.00 525.00 315.00Total RAS flow from sides [L/min] 1749.36 1604.61 1258.74Total RAS option I [L/min] 4058.28Total RAS option II [L/min] 4358.51The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.146Table D.37: Scenario C, 0.75 MGD (2.84 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.53 0.3 2.68 3.90 5.13 6.35 7.57 8.79 10.01 11.24Water flow per riser tube [L/min] 0.40 1.28 1.91 2.38 2.74 3.03 3.27 3.47Number of riser tubes 3584.00 2462.00 1875.00 1514.00 1269.00 1093.00 959.00 855.00Total RAS flow above module [L/min] 1415.74 3153.17 3581.11 3597.96 3474.33 3309.07 3131.93 2963.36The optimum is presented in bold.147Table D.38: Scenario C, 0.75 MGD (2.84 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.53 0.3 2.68 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.40 1.28 1.91 2.38 2.74 3.03 3.27Number of riser tubes 3225.00 1944.00 1481.00 1195.00 1002.00 865.00 758.00Total RAS flow above module [L/min] 1273.93 2489.75 2828.60 2839.87 2743.32 2618.80 2475.50Air on sides, option I, air per riser tube [L/min]0.73 0.5 2.68 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 1.75 2.58 3.14 3.56 3.88 4.14 4.35Number of riser tubes 358.00 246.00 187.00 151.00 126.00 109.00 95.00Total RAS flow from sides [L/min] 626.41 635.38 587.99 537.21 488.75 451.04 413.47Air on sides, option II, air per riser tube [L/min]0.83 0.6 2.68 3.90 5.13 6.35 7.57 8.79Water flow per riser tube [L/min] 2.18 2.97 3.51 3.90 4.21 4.46Number of riser tubes 358.00 246.00 187.00 151.00 126.00 109.00Total RAS flow from sides [L/min] 779.64 730.96 655.95 589.57 530.96 486.61Total RAS option I [L/min] 3475.25Total RAS option II [L/min] 3619.50The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.148Table D.39: Scenario C, 0.75 MGD (2.84 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.53 0.3 2.68 5.43 8.18 10.93 13.68 24.45Water flow per riser tube [L/min] 0.40 2.04 2.89 3.42 3.78 5.16Number of riser tubes 2688.00 1327.00 881.00 659.00 527.00 294.00Total RAS flow above module [L/min] 1061.81 2705.76 2546.22 2253.08 1994.31 1516.59Air on sides, option I, air per riser tube [L/min]0.73 0.5 2.68 3.90 5.13 6.35 7.57Water flow per riser tube [L/min] 1.75 2.58 3.14 3.56 3.88Number of riser tubes 896.00 615.00 468.00 378.00 317.00Total RAS flow from sides [L/min] 1567.78 1588.46 1471.56 1344.81 1229.63Air on sides, option II, air per riser tube [L/min]0.83 0.6 2.68 3.90 5.13Water flow per riser tube [L/min] 2.18 2.97 3.51Number of riser tubes 896.00 615.00 468.00Total RAS flow from sides [L/min] 1951.27 1827.40 1641.62Total RAS option I [L/min] 4294.22Total RAS option II [L/min] 4657.03The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.149Table D.40: Scenario C, 1 MGD (3.79 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.62 0.3 10.01 11.24 12.46 13.68 24.36 35.11 45.86Water flow per riser tube [L/min] 0.19 0.44 0.65 0.85 3.78 4.17 4.39Number of riser tubes 1259.00 1122.00 1012.00 922.00 517.00 359.00 275.00Total RAS flow above module [L/min] 236.20 488.32 661.71 780.89 1955.27 1496.75 1207.48The optimum is presented in bold.150Table D.41: Scenario C, 1 MGD (3.79 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.62 0.3 11.24 12.46 13.68 24.36Water flow per riser tube [L/min] 0.44 0.65 0.85 3.78Number of riser tubes 1010.00 911.00 830.00 466.00Total RAS flow above module [L/min] 439.57 595.67 702.97 1762.39Air on sides, option I, air per riser tube [L/min]0.82 0.5 2.68 3.90 5.13 6.35 7.57 8.79 10.01Water flow per riser tube [L/min] 0.84 1.73 2.34 2.80 3.15 3.43 3.67Number of riser tubes 470.00 323.00 246.00 198.00 166.00 143.00 125.00Total RAS flow from sides [L/min] 392.51 558.17 576.76 554.16 523.01 490.94 458.25Air on sides, option II, air per riser tube [L/min]0.92 0.6 2.68 3.90 5.13 6.35 7.57 8.79Water flow per riser tube [L/min] 1.33 2.20 2.79 3.22 3.56 3.83Number of riser tubes 470.00 323.00 246.00 198.00 166.00 143.00Total RAS flow from sides [L/min] 623.51 709.89 685.98 638.14 590.93 547.83Total RAS option I [L/min] 2339.16Total RAS option II [L/min] 2472.29The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.151Table D.42: Scenario C, 1 MGD (3.79 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.62 0.3 11.24 12.46 13.68 24.36Water flow per riser tube [L/min] 0.44 0.65 0.85 3.78Number of riser tubes 842.00 759.00 691.00 388.00Total RAS flow above module [L/min] 366.46 496.28 585.25 1467.40Air on sides, option I, air per riser tube [L/min]0.82 0.5 2.68 3.90 5.13 6.35 7.57Water flow per riser tube [L/min] 0.84 1.73 2.34 2.80 3.15Number of riser tubes 1176.00 807.00 615.00 496.00 416.00Total RAS flow from sides [L/min] 982.12 1394.57 1441.91 1388.21 1310.67Air on sides, option II, air per riser tube [L/min]0.92 0.6 2.68 3.90 5.13Water flow per riser tube [L/min] 1.33 2.20 2.79Number of riser tubes 1176.00 807.00 615.00Total RAS flow from sides [L/min] 1560.09 1773.64 1714.95Total RAS option I [L/min] 2909.31Total RAS option II [L/min] 3241.03The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.152Table D.43: Scenario C, 2 MGD (7.57 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]0.99 0.3 11.24 12.46 13.61 24.36 35.11 45.86Water flow per riser tube [L/min] 0.18 0.40 0.59 1.65 2.15 2.44Number of riser tubes 2245.00 2025.00 1853.00 1035.00 718.00 550.00Total RAS flow above module [L/min] 404.25 811.23 1087.09 1712.25 1545.85 1341.05The optimum is presented in bold.153Table D.44: Scenario C, 2 MGD (7.59 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.99 0.3 11.24 12.46 13.68 24.36Water flow per riser tube [L/min] 0.18 0.40 0.60 1.65Number of riser tubes 2021.00 1822.00 1660.00 932.00Total RAS flow above module [L/min] 363.91 729.91 991.39 1541.85Air on sides, option I, air per riser tube [L/min]1.19 0.5 6.35 7.57 8.79 10.01 11.24 12.46 13.68Water flow per riser tube [L/min] 0.68 1.08 1.41 1.69 1.93 2.13 2.31Number of riser tubes 397.00 333.00 287.00 251.00 224.00 202.00 184.00Total RAS flow from sides [L/min] 268.27 359.21 405.57 424.98 432.39 430.99 424.99Air on sides, option II, air per riser tube [L/min]1.29 0.6 7.57 8.79 10.01 11.24 12.46 13.68Water flow per riser tube [L/min] 1.65 1.97 2.24 2.47 2.66 2.83Number of riser tubes 333.00 287.00 251.00 224.00 202.00 184.00Total RAS flow from sides [L/min] 548.14 566.03 562.88 553.34 538.29 521.29Total RAS option I [L/min] 1974.24Total RAS option II [L/min] 2107.89The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.154Table D.45: Scenario C, 2 MGD (7.59 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]0.99 0.3 11.24 12.46 13.68 24.36Water flow per riser tube [L/min] 0.18 0.40 0.60 1.65Number of riser tubes 1684.00 1519.00 1383.00 776.00Total RAS flow above module [L/min] 303.23 608.53 825.96 1283.78Air on sides, option I, air per riser tube [L/min]1.19 0.5 6.35 7.57 8.79 10.01 11.24Water flow per riser tube [L/min] 0.68 1.08 1.41 1.69 1.93Number of riser tubes 993.00 833.00 717.00 629.00 561.00Total RAS flow from sides [L/min] 671.01 898.58 1013.22 1064.98 1082.90Air on sides, option II, air per riser tube [L/min]1.29 0.6 7.57 8.79 10.01Water flow per riser tube [L/min] 1.65 1.97 2.24Number of riser tubes 833.00 717.00 629.00Total RAS flow from sides [L/min] 1371.17 1414.10 1410.57Total RAS option I [L/min] 2366.68Total RAS option II [L/min] 2697.88The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.155Table D.46: Scenario C, 5 MGD (18.93 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]1.63 0.3 88.85 99.60Water flow per riser tube [L/min] 0.18 0.32Number of riser tubes 709.00 633.00Total RAS flow above module [L/min] 125.77 204.30The optimum is presented in bold.156Table D.47: Scenario C, 5 MGD (18.93 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]1.63 0.3 88.85 99.60Water flow per riser tube [L/min] 0.18 0.32Number of riser tubes 638.00 500.00Total RAS flow above module [L/min] 113.17 161.37Air on sides, option I, air per riser tube [L/min]1.83 0.5 13.61 24.36 35.11 45.86 56.61 67.35 78.10Water flow per riser tube [L/min] 0.66 1.73 2.23 2.51 2.70 2.83 2.92Number of riser tubes 463.00 258.00 179.00 137.00 111 93 80Total RAS flow from sides [L/min] 304.48 445.49 398.62 344.36 299.52 262.90 233.68Air on sides, option II, air per riser tube [L/min]1.93 0.6 13.61 24.36 35.11 45.86 56.61 67.35Water flow per riser tube [L/min] 0.71 1.78 2.28 2.56 2.75 2.87Number of riser tubes 463.00 258.00 179.00 137.00 111 93Total RAS flow from sides [L/min] 329.00 458.43 407.33 350.91 304.7809 267.282Total RAS option I [L/min] 606.86Total RAS option II [L/min] 619.80The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.157Table D.48: Scenario C, 5 MGD (18.93 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]1.63 0.3 88.85 99.60Water flow per riser tube [L/min] 0.18 0.32Number of riser tubes 532.00 475.00Total RAS flow above module [L/min] 94.37 153.30Air on sides, option I, air per riser tube [L/min]1.83 0.5 13.61 24.36 35.11 45.86 56.61Water flow per riser tube [L/min] 0.66 1.73 2.23 2.51 2.70Number of riser tubes 1158.00 647.00 449.00 343.00 278Total RAS flow from sides [L/min] 761.52 1117.18 999.90 862.15 750.14Air on sides, option II, air per riser tube [L/min]1.93 0.6 13.61 24.36 35.11 45.86 56.61Water flow per riser tube [L/min] 0.71 1.78 2.28 2.56 2.75Number of riser tubes 1158.00 647.00 449.00 343.00 278Total RAS flow from sides [L/min] 822.86 1149.62 1021.74 878.57 763.33Total RAS option I [L/min] 1270.48Total RAS option II [L/min] 1302.93The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.158Table D.49: Scenario C, 10 MGD (37.85 ∗106 L/day), 100% of air above module.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.2Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]159Table D.50: Scenario C, 10 MGD (37.85 ∗106 L/day), 90% of air above module, 10% on sides.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.3Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min]Air on sides, option I, air per riser tube [L/min]2.87 0.5 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 0.09 0.31 0.47 0.58Number of riser tubes 187.00 161.00 141.00 126.00Total RAS flow from sides [L/min] 16.72 50.48 66.01 73.51Air on sides, option II, air per riser tube [L/min]2.97 0.6 56.61 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 1.03 1.20 1.33 1.43 1.51Number of riser tubes 222.00 187.00 161.00 141.00 126.00Total RAS flow from sides [L/min] 227.71 225.24 214.84 202.04 190.25Total RAS option I [L/min] 73.51Total RAS option II [L/min] 227.71The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.160Table D.51: Scenario C, 10 MGD (37.85 ∗106 L/day), 75% of air above module, 25% on sides.L [m] H [m] Air above module, air per riser tube [L/min]2.67 0.3Water flow per riser tube [L/min] Submergence ratio is too high to pump waterNumber of riser tubesTotal RAS flow above module [L/min] [L/min]Air on sides, option I, air per riser tube [L/min]2.87 0.5 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 0.09 0.31 0.47 0.58Number of riser tubes 468.00 403.00 354.00 316.00Total RAS flow from sides [L/min] 41.85 126.35 165.74 184.35Air on sides, option II, air per riser tube [L/min]2.97 0.6 56.61 67.35 78.10 88.85 99.60Water flow per riser tube [L/min] 1.03 1.20 1.33 1.43 1.51Number of riser tubes 557.00 468.00 403.00 354.00 316.00Total RAS flow from sides [L/min] 571.33 563.71 537.78 507.24 477.13Total RAS option I [L/min] 184.35Total RAS option II [L/min] 571.33The optimum is presented in bold. Total RAS flows were calculated by adding the optimum water flow above the moduleand the optimum water flow in the side for option I and II, respectively.161

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0135662/manifest

Comment

Related Items