UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Comparative analysis of oxygen transfer in full lift and downflow bubble contact hypolimnetic aerators Ashley, Kenneth Ian 2002

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

Item Metadata

Download

Media
831-ubc_2002-731235.pdf [ 35.01MB ]
Metadata
JSON: 831-1.0063683.json
JSON-LD: 831-1.0063683-ld.json
RDF/XML (Pretty): 831-1.0063683-rdf.xml
RDF/JSON: 831-1.0063683-rdf.json
Turtle: 831-1.0063683-turtle.txt
N-Triples: 831-1.0063683-rdf-ntriples.txt
Original Record: 831-1.0063683-source.json
Full Text
831-1.0063683-fulltext.txt
Citation
831-1.0063683.ris

Full Text

C O M P A R A T I V E A N A L Y S I S OF O X Y G E N T R A N S F E R I N F U L L L I F T A N D D O W N F L O W B U B B L E C O N T A C T H Y P O L I M N E T I C A E R A T O R S B y K E N N E T H I A N A S H L E Y B.Sc. The University of British Columbia, 1976 M.Sc . The University of British Columbia, 1981 M . A . S c . The University of British Columbia, 1989 A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF D O C T O R OF P H I L O S O P H Y in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Department of C i v i l Engineering) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A March, 2002 © Kenneth Ian Ashley , 2002 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. 1 further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department DE-6 (2/88) ( A B S T R A C T This research conducted a detailed analysis of design factors influencing the oxygen transfer capabilities of two types of hypolimnetic aerators: full lift and Speece Cone, utilising non steady-state gas transfer methodology in a laboratory scale system. Full lift hypolimnetic aerators use the airlift pump principle to circulate water from the hypolimnion to the surface and return oxygenated water to the hypolimnion. Speece Cones use the Downflow Bubble Contact Aeration (DBCA) principle to oxygenate hypolimnetic water at depth using mechanical pumps to circulate water through the aeration chamber. Four standard units of measure were used to quantify oxygen transfer: (a) K i ^ o (hr - 1), the oxygen transfer coefficient at 20 °C; (b) SOTR (g O2 hr"1), the Standard Oxygen Transfer Rate; (c) S A E (g O2 kWhr"1), the Standard Aeration Efficiency and (d) SOTE (%), the Standard Oxygen Transfer Efficiency. The design factors examined in the full lift experiments were: the effect of diffuser orifice pore diameter (140 u, 400 u and 800 u), diffuser submergence (1.5 and 2.9 m), gas flow rate (10, 20, 30 and 40 1 min" '), oxygen partial pressure (air and Pressure Swing Adsorption generated oxygen) and four in situ modifications (floating surface cover in the separator box, mid-depth inlet tube bubble screens, counter-rotating blades and D B C A ) in the outlet tube. In the full lift experiments, diffuser submergence positively influenced K i ^ o , SOTR, S A E and SOTE and water velocity. Higher gas flow rates increased water velocity, K i ^ o and SOTR but decreased S A E and SOTE. Oxygen partial pressure treatment exerted a significant positive effect on KIAJO, SOTR and SOTE, whereas S A E response decreased on PSA oxygen. Smaller orifice diameters exerted a significant positive effect on K i ^ o , SOTR, S A E and SOTE. The in situ modifications exerted a significant negative effect on K i ^ o , SOTR, S A E and SOTE. The D B C A treatment generated the smallest bubble sizes. The design factors examined in the Speece Cone experiments were the effect of oxygen flow rate (1,2 and 3 1 min"1) and outlet port discharge water velocity (20, 30, 40, 50, 60 and 70 cm sec"1) on K L a 2 0 , SOTR, S A E and SOTE. The inlet water velocity was insufficient to create an efficient gas transfer environment in the Speece Cone experiments. Oxygen flow rate and discharge velocity treatments exerted a significant positive effect on K ^ o and SOTR. S A E increased with oxygen flow rates but discharge velocity had no effect on SAE. SOTE increased with water discharge velocity but responded negatively to increasing oxygen flow, suggesting that oxygen flow rates > 1 1 min"1 were not being fully dissolved inside the Speece Cone. A precipitous decline in SOTE at low oxygen flow rates and discharge velocities > 50 cm sec"1 was due to a déstabilisation of the bubble swarm in the Speece Cone. The Speece Cone, at oxygen flow rates between 1 and 3 1 min"1, delivered K i ^ o and SOTR performance comparable to the full lift system at 10 1 min"1 on oxygen, and outstanding SOTE performance. The D B C A counter-current flow principle of the Speece Cone provides a superior gas transfer environment to the full lift design by positively influencing the fundamental aspects of gas transfer (i.e., a, K L and Q - C L ) . The efficiency of the Pressure Swing Adsorption (PSA) generat-ing units increases with unit capacity, as smaller PSA units are not designed solely for optimum oxygen recovery efficiency; improvements in S A E could be achieved by using lower pressure PSA generators. The relative magnitude of the effects from this study should be transferable to full-scale application, within reasonable limits. Further lab testing, comparative field studies, development of predictive models, reporting in engineering journals and case study reviews of specific installations wil l assist in resolving the engineering aspects of system selection. The Speece Cone and PSA oxygen generating systems represent the most significant advances in hypolimnetic aeration since their inception over half a century ago. Civi l engineers should consider hypolimnetic aeration as a proven technique, which can significantly improve raw water quality, and serve as an integral component of an overall watershed management pro-gram. T A B L E OF CONTENTS Page ABSTRACT ii T A B L E OF CONTENTS v LIST OF T AB L E S xi LIST OF FIGURES xix N O M E N C L A T U R E xxv A C K N O W L E D G M E N T S xxiv C H A P T E R 1: INTRODUCTION 1 1.1 Introduction 1 C H A P T E R 2: L I T E R A T U R E REVIEW 4 2.1 Literature Review 4 2.1.1 History of Hypolimnetic aeration 4 2.1.2 Types of Hypolimnetic Aerators 6 2.1.2.1 Full or Partial Lift Hypolimnetic Aeration with Compressed A i r 6 2.1.2.2 Full or Partial Lift Hypolimnetic Aeration with Pure Oxygen or Oxygen Supplementation 10 2.1.2.3 Downflow Bubble Contact Aeration 11 2.2 Rationale for Current Research 13 C H A P T E R 3: METHODS 17 3.1 System Design 17 3.1.1 Aeration Tank 17 3.1.2 Water Supply 19 3.1.3 A i r Supply 19 3.1.4 Oxygen Supply 20 3.1.5 Gas Delivery and Flow Measurement 20 3.1.6 Water Velocity Measurement 23 3.1.7 Data Logger 23 3.1.8 OxyGuard Probe Design 25 3.1.9 Probe Membrane 25 3.1.10 Water Temperature 26 3.2 Test Procedure 26 3.2.1 Oxygen Probe Calibration 27 3.2.2 Calculation of Daily Oxygen Saturation 27 3.2.3 Deoxygenation Procedure 29 3.3 Full Lift Hypolimnetic Aerator 31 3.3.1 Aerator Dimensions 31 3.3.2 Probe Locations 33 3.3.3 Diffuser Description 34 3.4 Experimental Design 36 3.4.1 Termination of Experiments 43 3.5 Downflow Bubble Contact Aerator (Speece Cone) Hypolimnetic Aerator 44 3.5.1 Aerator Dimensions 44 3.5.2 Tank Mixing and Configuration of Submersible Pumps 46 3.5.3 Probe Locations 47 3.5.4 Diffuser Description 47 3.6 Experimental Design 48 3.6.1 Termination of Experiments 49 3.7 Bubble Size Measurement 50 3.8 Electrical measurements and power calculations 51 3.9 Parameter Estimation Procedure 52 3.9.1 Full Lift Experiments 56 3.9.2 Speece Cone Experiments 60 3.9.3 Statistical Analysis 62 C H A P T E R 4: RESULTS 66 4.1 Full Lift Hypolimnetic Aeration - Oxygen Transfer 66 4.1.1 Depth Effect 66 4.1.1.1 140 u diffuser x air (Note: compares the 140 u. diffuser on air at 1.5 and 2.9 m 66 4.1.1.2 400 u. diffuser x air (Note: compares the 400 u. diffuser on air at 1.5 and 2.9 m) 67 4.1.1.3 800 u. diffuser x air (Note: compares the 800 u. diffuser on air at 1.5 and 2.9 m) 68 4.1.1.4 140 u. diffuser x P S A oxygen (Note: compares the 140 u. diffuser on P S A 0 2 at 1.5 and 2.9 m) 70 4.1.1.5 400 |x diffuser x P S A oxygen (Note: compares the 400 fx diffuser on P S A 0 2 at 1.5 and 2.9 m) 72 4.1.1.6 800 p. diffuser x PSA oxygen (Note: compares the 800 (x diffuser on PSA 0 2 at 1.5 and 2.9 m) 73 4.1.2 Oxygen Partial Pressure Effect 76 4.1.2.1 140 |x diffuser x 1.5 m depth (Note: compares the 140 |x diffuser on airandPSA 0 2 a t 1.5 m) 76 4.1.2.2 400 |x diffuser x 1.5 m depth (Note: compares the 400 |x diffuser on air and PSA 0 2 at 1.5 m) 78 4.1.2.3 800 fx diffuser x 1.5 m depth (Note: compares the 800 |x diffuser on air and PSA 0 2 at 1.5 m) 80 4.1.2.4 140 |x diffuser x 2.9 m depth (Note: compares the 140 JX diffuser on air and PSA 0 2 at 2.9 m) 81 4.1.2.5 400 (x diffuser x 2.9 m depth (Note: compares the 400 |x diffuser on air and P S A 0 2 at 2.9 m) 83 4.1.2.6 800 fx diffuser x 2.9 m depth (Note: compares the 800 u diffuser on air and PSA 0 2 at 2.9 m) 85 4.1.3 Orifice Diameter Effect 87 4.1.3.1 A i r x 1.5 m depth (Note: compares the 140 jx, 400 u. and 800 u. diffusers on air at 1.5 m) 87 4.1.3.2 A i r x 2.9 m depth (Note: compares the 140 [x, 400 jx and 800 u. diffusers on air at 2.9 m) 89 4.1.3.3 PS A oxygen x 1.5 m depth (Note: compares the 140 u, 400 u. and 800 y. diffusers on PSA 0 2 at 1.5 m) 90 4.1.3.4 P S A oxygen x 2.9 m depth (Note: compares the 140 |x, 400 fx and 800 n diffusers on PSA 0 2 at 2.9 m) 93 4.1.4 Full Lift Modifications Effect 95 4.1.4.1 A i r x 140 |x diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 140 |x diffuser on air at 2.9 m) 95 4.1.4.2 A i r x 800 |x diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 800 jx diffuser on air at 2.9 m) 98 4.1.4.3 PSA oxygen x 140 |x diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 140 u diffuser on PSA 0 2 at 2.9 m) 102 4.1.4.4 P S A oxygen x 800 |x diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 800 |x diffuser on PSA 0 2 at 2.9 m) 105 4.1.5 Miscellaneous Tests 109 Full Lift Hypolimnetic Aeration - Water Velocity 111 4.2.1 Gas Effect I l l 4.2.2 Depth Effect 113 4.2.3 Orifice Effect 115 4.2.4 Full Lift Modifications Effect 117 4.3 Full Lift Hypolimnetic Aeration - Bubble Size 119 4.3.1 Depth Effect 124 4.3.1.1 140 | i diffuser at 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m 124 4.3.1.2 400 |x diffuser at 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m 125 4.3.1.3 800 jx diffuser at 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m 126 4.3.2 Orifice Diameter Effect 127 4.3.2.1 1.5/0.5 m with 140 \L, 400 u. and 800 u. diameter orifice 127 4.3.2.2 2.9/0.5 m with 140 |x, 400 \i and 800 |x diameter orifice 129 4.3.2.3 2.9/2.7 m with 140 |x, 400 u. and 800 u diameter orifice 130 4.3.3 Full Lift Modifications Effect 131 4.3.3.1 140 (x orifice with D B C A , Screen, Vanes and Control 131 4.3.3.2 800 |x orifice with D B C A , Screen, Vanes and Control 132 4.4 Full Lift Hypolimnetic Aeration - Oxygen Transfer Locations 135 4.4.1 1 4 0 | x x 2 . 9 m x a i r 135 4.4.2 140 n x 1.5 m x air 138 4.4.3 400 u. x 2.9 m x air 138 4.4.4 400 ^ix 1.5 m x air 138 4.4.5 800 (x x 2.9 m x air 145 4.4.6 800 f ix 1.5 m x air 145 4.4.7 140 [xx 2.9 m x 0 2 145 4.4.8 140 |xx 1.5 m x 0 2 153 4.4.9 400 |x x 2.9 m x 0 2 153 4.4.10 400 |xx 1.5 m x 0 2 153 4.4.11 800 (x x 2.9 m x 0 2 160 4.4.12 800^ix 1 . 5 m x 0 2 160 4.4.13 D B C A x 140 |xx air 160 4.4.14 D B C A x 800 u.x air 167 4.4.15 Screen x 140 u x air 170 4.4.16 Screen x 800 u x air 170 4.4.17 Surface cover x 140 |x x air 170 4.4.18 Surface cover x 800 u. x air 177 4.4.19 Vanes x 140 u,x air 177 4.4.20 Vanes x 800 [i x air 177 4.4.21 D B C A x 140 j x x 0 2 177 4.4.22 D B C A x 800 |x x 0 2 186 4.4.23 Screen x 140 ^ i x 0 2 189 4.4.24 Screen x 800 (X x 0 2 189 4.4.25 Surface cover x 140 |x x 0 2 189 4.4.26 Surface cover x 800 |x x 0 2 196 4.4.27 Vanes x 140 y. x 0 2 196 4.4.28 Vanes x 800 n x 0 2 196 4.5 Speece Cone Hypolimnetic Aeration - Oxygen Transfer 196 4.5.1 PSA 0 2 Experiments 196 4.5.2 A i r Experiments 214 C H A P T E R 5: DISCUSSION 216 5.1 Review of Gas Transfer Theory 216 5.2 Practical Applications of Gas Transfer Theory 220 5.3 Full Lift Hypolimnetic Aeration - Groups 1 to 6 224 5.3.1 Diffuser Depth 224 5.3.2 Oxygen Partial Pressure 227 5.3.3 Orifice Diameter 229 5.3.4 Group 1 to 6 Summary 233 5.4 Full Lift Hypolimnetic Aeration - Groups 7 to 14 234 5.4.1 In situ Modifications 234 5.4.2 Water Velocity 241 5.4.3 Bubble Size 242 5.4.4 Orifice Diameter and Oxygen Partial Pressure 247 5.4.5 Group 7 tol4 Summary 247 5.5 Miscellaneous Tests 249 5.5.1 Pump Only Tests 249 5.5.2 Extended Aeration Tests 251 5.6 Speece Cone 251 5.6.1 Oxygen Flow Rate and Discharge Velocity 252 5.6.2 Speece Cone on A i r 254 5.6.3 Speece Cone Summary 256 5.7 Comparison of Relative Performance of Full Lift vs. Speece Cone 257 5.8 Comparison with Civ i l Engineering Studies 261 5.9 Linkage to Hypolimnetic Aeration Studies 262 5.10 PSA Engineering Design Factors Influencing S A E 267 5.11 Comments on Full Scale Application of Full Lift and Speece Cone Systems 272 5.12 Future Research Requirements 276 C H A P T E R 6: SUMMARY AND CONCLUSIONS 279 6.1 Full Lift Experiments 279 6.1.1 Depth Treatment 279 6.1.2 Oxygen Partial Pressure Treatment 279 6.1.3 Orifice Diameter Treatment.. 280 6.1.4 In situ Modifications Treatment 280 6.2 Speece Cone 281 6.3 Comparison with Civ i l Engineering Studies. 282 6.4 Linkage to Hypolimnetic Aeration Studies 282 6.5 P S A Engineering Design Factors Influencing S A E 282 6.6 Comments on Full Scale Application of Full Lift and Speece Cone Systems 283 6.7 Future Research Requirements 283 6.8 Conclusions 284 BIBLIOGRAPHY 285 LIST OF APPENDICES 293 APPENDIX 1 : RE-AERATION TEST PROCEDURE - F U L L LIFT HYPOLIMNETIC A E R A T O R 294 APPENDIX 2: RE-AERATION TEST PROCEDURE - SPEECE HYPOLIMNETIC A E R A T O R 296 LIST OF TABLES Page Table 3.1. List of treatment groups for full lift hypolimnetic aeration experiments 37 Table 3.2. Experimental combinations for Group 1 38 Table 3.3. Experimental combinations for Group 2 38 Table 3.4. Experimental combinations for Group 3 38 Table 3.5. Experimental combinations for Group 4 39 Table 3.6. Experimental combinations for Group 5 39 Table 3.7. Experimental combinations for Group 6 39 Table 3.8. Experimental combinations for Group 7 (Downflow Bubble Contact Aeration) 40 Table 3.9. Experimental combinations for Group 8 (Downflow Bubble Contact Aeration) 40 Table 3.10. Experimental combinations for Group 9 (Bubble screen) 40 Table 3.11. Experimental combinations for Group 10 (Bubble screen) 41 Table 3.12. Experimental combinations for Group 11 (Surface cover) 41 Table 3.13. Experimental combinations for Group 12 (Surface cover) 41 Table 3.14. Experimental combinations for Group 13 (Counter-rotating blades) 42 Table 3.15. Experimental combinations for Group 14 (Counter-rotating blades) 42 Table 3.16. Experimental combinations for Group Misc. 1 (Pump only /DBCA) 42 Table 3.17. Experimental combinations for Group Misc. 2 (Extended run/DBCA) 42 Table 3.18. List of treatment combinations for the full lift hypolimnetic aeration experiments 43 Table 3.19. List of treatment groups for Speece Cone experiments 48 Table 3.20. List of treatment combinations for Speece Cone experiments 49 Table 3.21. Summary of energy calculations for full lift hypolimnetic aerator experiments 56 Table 3.22. Summary of energy calculations for Speece Cone experiments 60 Table 4.1. Adjusted least squares means (± SE), treatment and interaction effects for the 140 fx diffuser on air at 1.5 m or 2.9 m depth 66 Table 4.2. Adjusted least squares means (± SE), treatment and interaction effects for the 400 |x diffuser on air at 1.5 m or 2.9 m depth 68 Table 4.3. Adjusted least squares means (± SE), treatment and interaction effects for the 800 |x diffuser on air at 1.5 m or 2.9 m depth 69 Table 4.4. Adjusted least squares means (± SE), treatment and interaction effects for the 140 u. diffuser on PSA oxygen at 1.5 m or 2.9 m depth 70 Table 4.5. Adjusted least squares means (± SE), treatment and interaction effects for the 400 LX diffuser on PSA oxygen at 1.5 m or 2.9 m depth 73 Table 4.6. Adjusted least squares means (+ SE), treatment and interaction effects for the 800 u. diffuser on PSA oxygen at 1.5 m or 2.9 m depth 75 Table 4.7. Summary of the adjusted least squares means depth treatment effects showing percentage increase/decrease 76 Table 4.8. Adjusted least squares means (± SE), treatment and interaction effects for the 140 |x diffuser at 1.5 m depth on air or PSA oxygen 77 Table 4.9. Adjusted least squares means (± SE), treatment and interaction effects for the 400 |x diffuser at 1.5 m depth on air or PSA oxygen 78 Table 4.10. Adjusted least squares means (+ SE), treatment and interaction effects for the 800 |x diffuser at 1.5 m depth on air or PSA oxygen 80 Table 4.11. Adjusted least squares means (± SE), treatment and interaction effects for the 140 LX diffuser at 2.9 m depth on air or PSA oxygen 82 Table 4.12. Adjusted least squares means (± SE), treatment and interaction effects for the 400 |x diffuser at 2.9 m depth on air or P S A oxygen 84 Table 4.13. Adjusted least squares means (± SE), treatment and interaction effects for the 800 \JL diffuser at 2.9 m depth on air or PSA oxygen 85 Table 4.14. Summary of oxygen partial pressure treatment effects showing percentage increase/decrease 86 Table 4.15. Adjusted least squares means (± SE), treatment and interaction effects for the 140 LX, 400 |x and 800 LI diffusers at 1.5 m depth on air 87 Table 4.16. Adjusted least squares means (+ SE), treatment and interaction effects for the 140 |x, 400 LX and 800 LX diffusers at 2.9 m depth on air 89 Table 4.17. Adjusted least squares means (+ SE), treatment and interaction effects for the 140 u., 400 u, and 800 |x diffusers at 1.5 m depth on P S A oxygen 91 Table 4.18. Adjusted least squares means (+ SE), treatment and interaction effects for the 140 u, 400 u. and 800 |j, diffusers at 2.9 m depth on P S A oxygen 93 Table 4.19. Summary of diffuser orifice diameter treatment effects showing percentage increase/decrease 94 Table 4.20. Adjusted least squares means (± SE), treatment and interaction effects for the various in situ full lift hypolimnetic aerator treatments on air with the 140 u. diffuser 96 Table 4.21. Matrix of K.La2o pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on air with the 140 n diffuser 97 Table 4.22. Matrix of SOTRpairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on air with the 140 |x diffuser 97 Table 4.23. Matrix of S A E and SOTE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on air with the 140 (i diffuser 98 Table 4.24. Matrix of multiplicative in situ treatment x flow rate interaction effects for K L a 2 o and SOTR for various in situ full lift hypolimnetic aerator treatments on air with the 140 u. diffuser 98 Table 4.25. Adjusted least squares means (± SE), treatment and interaction effects for various in situ full lift hypolimnetic aerator treatments on air with the 800 n diffuser 100 Table 4.26. Matrix of K I A M pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on air with the 800 u, diffuser 100 Table 4.27. Matrix of SOTRpairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on air with the 800 (i diffuser 101 Table 4.28. Matrix of S A E and SOTE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on air with the 800 u. diffuser 101 Table 4.29. Matrix of multiplicative in situ treatment x flow rate interaction effects for K L a 2 o and SOTR for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on air with the 800 u. diffuser 101 Table 4.30. Matrix of multiplicative in situ treatment x flow rate interaction effects for S A E and SOTE for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on air with the 800 |x diffuser 102 Table 4.31. Summary of various in situ full lift hypolimnetic aerator treatments on air with the 140 u. and 800 |x diffuser showing percentage increase/decrease relative to No Treatment 102 Table 4.32. Adjusted least squares means (± SE), treatment and interaction effects for the various in situ full lift hypolimnetic aerator treatments on P S A oxygen with the 140 diffuser 104 Table 4.33. Matrix of K i ^ o and SOTR pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 u. diffuser 104 Table 4.34. Matrix of S A E pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 LX diffuser 104 Table 4.35. Matrix of SOTE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 (x diffuser 105 Table 4.36. Matrix of multiplicative in situ treatment x flow rate interaction effects for K L a 2 o and SOTR for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 (x diffuser 105 Table 4.37. Adjusted least squares means (± SE), treatment and interaction effects for various in situ full lift hypolimnetic aerator treatments on P S A oxygen with the 800 |x diffuser 106 Table 4.38. . Matrix of K i ^ o and SOTR pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on PSA oxygen with the 800 (x diffuser 107 Table 4.39. Matrix of S A E pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on P S A oxygen with the 800 (x diffuser 108 Table 4.40. Matrix of multiplicative in situ treatment x flow rate interaction effects for KLa 2o and SOTR for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on P S A oxygen with the 800 fx diffuser 108 Table 4.41. Matrix of multiplicative in situ treatment x flow rate interaction effects for S A E and SOTE for various in situ full lift hypolimnetic aerator treatments (excluding D B C A ) on PSA oxygen with the 800 LX diffuser 109 Table 4.42. Summary of various in situ full lift hypolimnetic aerator treatments on P S A O2 with the 140 u. and 800 u diffuser showing percentage increase/decrease relative to No Treatment 109 Table 4.43. Mean (+ SE) K L a 2 0 , SOTR, S A E and SOTE values for full lift hypolimnetic aerator operating in pump only D B C A mode on air or P S A oxygen at 3 1 min"1 110 Table 4.44. Mean (± SE) K L a 2 0 , SOTR, S A E and T E values for the full lift hypolimnetic aerator operating in D B C A mode on air or PSA oxygen at 7 and 3 1 min"1 at 2.9 m using a 140 u. diffuser. (Note: 7/3 refers to the gas flow, in 1 min"1, to the diffuser in the inlet tube, and the D B C A diffuser in the separator box) 110 Table 4.45. Laboratory measurements recorded during extended aeration tests 111 Table 4.46. Mean (± SE) water velocity (cm sec"1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - gas effect 113 Table 4.47. . Mean (± SE) water velocity (cm sec"1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - depth effect 114 Table 4.48. Mean (± SE) water velocity (cm sec"1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - orifice effect 116 Table 4.49. Mean (± SE) water velocity (cm sec"1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - full lift in situ modifications effect 118 Table 4.50. Matrix of water velocity pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments with the 140 \i diffuser 118 Table 4.51. Matrix of water velocity pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments with the 800 \i diffuser 119 Table 4.52. Summary statistics for bubble measurements 119 Table 4.53. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 140 u. diffuser at various depths of gas release and bubble measurement 121 Table 4.54. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 400 u, diffuser at various depths of gas release and bubble measurement 121 Table 4.55. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 800 u. diffuser at various depths of gas release and bubble measurement 122 Table 4.56. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 140 jx diffuser with in situ modifications. (Note: 7/3, 17/3, 27/3 and 37/3 refers to the gas flow, in 1 min"1, to the diffuser in the inlet tube, and the D B C A diffuser in the separator box) 122 Table 4.57. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 800 |x diffuser with in situ modifications (Note: 7/3, 17/3, 27/3 and 37/3 refers to the gas flow, in 1 min"1, to the diffuser in the inlet tube, and the D B C A diffuser in the separator box) 123 Table 4.58. Adjusted least squares means (± SE), treatment and interaction effects for the 140 \x diffuser on air at bubble release depth/bubble measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m depths. (* = marginal effect; <x= 0.01) 124 Table 4.59. Adjusted least squares means (± SE), treatment and interaction effects for the 400 |x diffuser on air at bubble release depth/bubble measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m depths 126 Table 4.60. Adjusted least squares means (± SE), treatment and interaction effects for the 800 LX diffuser on air at bubble release depth/bubble measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m depths 126 Table 4.61. Adjusted least squares means (± SE), treatment and interaction effects for the bubble release depth/bubble measurement depth combination of 1.5/0.5 m with 140 |x, 400 \i and 800 |x diameter diffuser orifices 128 Table 4.62. Adjusted least squares means (± SE), treatment and interaction effects for the bubble release depth/bubble measurement depth combination of 2.9/0.5 m with 140 \i, 400 jx and 800 \i diameter diffuser orifices 129 Table 4.63. Adjusted least squares means (± SE), treatment and interaction effects for the air release depth/photo measurement depth combination of 2.9 m/2.7 m with 140 |x, 400 jx and 800 |x diameter diffuser orifices 130 Table 4.64. Mean (+ SE) equivalent bubble diameter (mm) for the air release depth/bubble measurement depth combination of 2.9 m/0.5 m with 140 |x for the full lift hypolimnetic aerator - full lift in situ modifications effect 131 Table 4.65. Mean (± SE) equivalent bubble diameter (mm) for the air release depth/bubble measurement depth combination of 2.9 m/0.5 m with 800 \x for the full lift hypolimnetic aerator - full lift in situ modifications effect: D B C A and vanes 134 Table 4.66. Mean (± SE) equivalent bubble diameter (mm) for the air release depth/bubble measurement depth combination of 2.9 m/0.5 m with 800 |a for the full lift hypolimnetic aerator - full lift in situ modifications effect: Screen and Control 134 Table 4.67. Summary of qualitative observations of the relative magnitude, slope and response to increasing air flows for oxygen differentials in Zones 1-5 152 Table 4.68. Summary of qualitative observations of the relative magnitude, slope and response to increasing P S A oxygen flows for oxygen differentials in Zones 1-5 167 Table 4.69. Summary of qualitative observations of the relative magnitude, slope and response to increasing air flows for oxygen differentials in Zones 1-5 for in situ modifications 186 Table 4.70. Summary of qualitative observations of the relative magnitude, slope and response to increasing PSA 0 2 flows for oxygen differentials in Zones 1-5 for in situ modifications 203 Table 4.71. Adjusted least squares means (± SE), treatment and interaction effects for various PSA oxygen flow rates with the Speece Cone without the 60 cm sec"1 and 70 cm sec"1 water discharge velocities 211 Table 4.72. Adjusted least squares means (± SE), treatment and interaction effects for various PSA oxygen flow rates with the Speece Cone without the 1 min' 1 oxygen flow rate treatment 212 Table 4.73. Mean (± SE) K L a 2 0 , SOTR, S A E and SOTE values for Speece Cone hypolimnetic aerator operating on air at 3 1 min"1 and 50 cm sec"1 compared to PSA oxygen at 3 1 min"1 and 50 cm sec"1 214 Table 5.1. Relationship between K.La (Q - C L ) , the eight operational factors, and the experimental variables 223 Table 5.2. Summary of various full lift in situ modification treatments on water velocity with the 140 p. and 800 |x diffuser, on air, showing percentage increase/decrease relative to No Treatment 235 Table 5.3. Summary of various in situ full lift hypolimnetic aerator treatments on air with the 140 u. and 800 | i diffuser showing percentage increase/decrease relative to No Treatment for K i ^ o , SOTR, S A E and SOTE and water velocity 237 Table 5.4. Decrease in water velocity with the 800 \x. diffusers as compared to 140 u. diffusers on air for the full lift in situ modification treatments (from Table 5.2) 242 Table 5.5. Hypothetical comparison of Pump only, Group 7 and Group 1 values 250 Table 5.6. Adjusted least square treatment means (± SE) comparison between Speece Cone on air at 3 1 min"1 and 50 cm sec"1, with the Group 8 (i.e., 800 u. diffuser, 2.9 m diffuser submergence) full lift D B C A test, running on air with a 7/3 1 min"1 allocation between main and separator box diffuser 255 Table 5.7. Comparison of experimental Speece Cone inlet water velocities and oxygen water ratios to design specifications 258 Table 5.8. Adjusted least square treatment means (± SE) comparison between Speece Cone on P S A 0 2 at 1 and 3 1 min"1 and 50 cm sec"1, with the Group 2 (i.e., 800 u. diffuser, 2.9 m diffuser submergence) full lift test, on air or P S A 0 2 258 Table 5.9. Comparison of experimental system of Mavinic (1973) and this study, Full Lift only on air, Groups 1-6 262 Table 5.10. Estimates of A E and T E from full scale hypolimnetic aeration systems. (Note: A E and T E are not derived from the non steady-state procedure). Speece Cone 1 operated at 22.9 m, and Speece Cone 2 surcharged to equivalent depth of 39.6 m 263 Table 5.11. Decrease of the oxygen content in influent air at various depths during bubble rise to the surface (from Bernhardt, 1967) 265 Table 5.12. A i r input/oxygen output ratio and oxygen recovery efficiencies of selected OGSI Inc. PSA units 268 Table 5.13. Performance comparison of equivalent capacity P S A systems 269 Table 5.14. Calculated effect of variable P S A oxygen recovery efficiency on Group 1 Full Lift S A E performance 270 Table 5.15. Calculated effect of variable PSA feed air compression ratios on Group 1 full lift S A E results using three ranges of P S A oxygen recovery efficiency 271 Table 5.16. General guide to hypolimnetic aeration system selection 275 LIST OF FIGURES Page Figure 2.1. Full lift "Bernhardt" type of hypolimnetic aerator (from Soltero et al., 1994) 7 Figure 2.2. Partial lift hypolimnetic aerator (from Soltero et al., 1994) 9 Figure 2.3 Downflow Bubble Contact Aerator "Speece Cone" hypolimnetic aerator 12 Figure 3.1 Schematic diagram of aeration tank 18 Figure 3.2. Schematic diagram of manifold board with data logger and flow meters 21 Figure 3.3. Schematic diagram of full lift aerator with probe location and aeration zones 32 Figure 3.4 Photograph of 140 u. fine pore diffusers 35 Figure 3.5. Photograph of 400 u. and 800 u. coarse bubble diffusers 35 Figure 3.6. Schematic of Speece Cone aerator with probe locations 45 Figure 3.7. Speece Cone main water pump power consumption vs. rheostat setting 52 Figure 4.1a-d KLa20 (a), SOTR (b), S A E (c) and SOTE (d) for 140 (x diffuser on air at 1.5 and 2.9 m depth 67 Figure 4.2a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 400 n diffuser on air at 1.5 and 2.9 m depth 69 Figure 4.3a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 800 \x diffuser on air at 1.5 and 2.9 m depth 71 Figure 4.4a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 140 u diffuser on PSA 0 2 at 1.5 and 2.9 m depth 72 Figure 4.5a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 400 | i diffuser on PSA 0 2 a t 1.5 and 2.9 m depth 74 Figure 4.6a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 800 \i diffuser on P S A 0 2 a t 1.5 and 2.9 m depth 75 Figure 4.7a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 140 u diffuser at 1.5 m on air or P S A 0 2 . 77 Figure 4.8a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 400 ji diffuser at 1.5 m on air or PSA 0 2 79 Figure 4.9a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 800 \i diffuser at 1.5 m on air or PSA 0 2 . 81 Figure 4.10a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 140 LI diffuser at 2.9 m on air or PSA 0 2 . 82 Figure 4.1 la-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 400 n diffuser at 2.9 m on air or PSA 0 2 . 84 Figure 4.12a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 800 \i diffuser at 2.9 m on air or PSA 0 2 p 86 Figure 4.13a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 140, 400 and 800 \x diffusers on air at 1.5 m 88 Figure 4.14a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 140, 400 and 800 | i diffusers on air at 2.9 m 91 Figure 4.15a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 140, 400 and 800 \i diffusers on P S A 0 2 at 1.5 m 92 Figure 4.16a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for 140, 400 and 800 \i diffusers on PSA 0 2 at 2.9 m 94 Figure 4.17a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for in situ modifications on air with 140 u. diffuser at 2.9 m 96 Figure 4.18a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for in situ modifications on air with 800 u diffuser at 2.9 m 99 Figure 4.19a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for in situ modifications on P S A 0 2 with 140 \L diffuser at 2.9 m 103 Figure 4.20a-d K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) for in situ modifications on PSA 0 2 with 800 [i diffuser at 2.9 m 107 Figure 4.21 a-c Water velocity vs. gas flow rate on air or P S A 0 2 a t 1.5 m using 140 p. (a), 400 p. (b) and 800 n (c) diffusers 112 Figure 4.22a-b Water velocity vs. gas flow rate on air or P S A 0 2 at 2.9 m using 140 LI (a) and 400 n(b) diffusers 112 Figure 4.23a-c Water velocity vs. air flow rate at 1.5 or 2.9 m using 140 \i (a), 400 (J. (b) and 800 |x (c) diffusers 114 Figure 4.24a-b Water velocity vs. air flow rate for 140, 400 and 800 fi diffusers at 1.5 m (a) and 2.9 m (b) 116 Figure 4.25a-b Water velocity vs. air flow rate at 2.9 m using 140 u. (a) or 800 (x (b) diffusers with in situ modifications 118 Figure 4.26a-b Frequency histogram, untranformed (a) and log x 2 transformed (b), of Equivalent Bubble Diameters 120 Figure 4.27 Equivalent Bubble Diameter vs. air flow rate with 140 u, diffuser at diffuser submergence depth and measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m 125 Figure 4.28 Equivalent Bubble Diameter vs. air flow rate with 400 u. diffuser at diffuser submergence depth and measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m 125 Figure 4.29 Equivalent Bubble Diameter vs. air flow rate with 800 u. diffuser at diffuser submergence depth and measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m 127 Figure 4.30 Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combination of 1.5/0.5 m with 140, 400 and 800 (x diffuser. 128 Figure 4.31 Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combination of 2.9/0.5 m with 140, 400 and 800 \i diffuser. 129 Figure 4.32 Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combinations of 2.9/2.7 m with 140, 400 and 800 u diffuser 130 Figure 4.33a-b Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combinations of 2.9/0.5 m with 140 |x diffuser with in situ modifications (a), and without the Control treatment (b) 132 Figure 4.34a-c Equivalent Bubble Diameter on air at diffuser submergence depth and measurement depth combinations of 2.9/0.5 m with 800 \i diffuser and in situ modifications (a), with D B C A and Vanes only (b), and with Control and Screen treatments only (c) 133 Figure 4.35a-d Oxygen differentials in full lift aerator on air at 2.9 m with 140 u. diffuser at 10 1 m i n 1 (a), 20 1 min"1 (b),30 1 min"1 (c),40 1 min"1 (d) 136 Figure 4.36a-d Oxygen differentials in full lift aerator on air at 1.5 m with 140 u. diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 139 Figure 4.37a-d Oxygen differentials in full lift aerator on air at 2.9 m with 400 \i diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c),40 1 min"1 (d) 141 Figure 4.38a-d Oxygen differentials in full lift aerator on air at 1.5 m with 400 jx diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 143 Figure 4.39a-d Oxygen differentials in full lift aerator on air at 2.9 m with 800 u diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 146 Figure 4.40a-d Oxygen differentials in full lift aerator on air at 1.5 m with 800 u, diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 148 Figure 4.41a-d Oxygen differentials in full lift aerator on PSA 0 2 at 2.9 m with 140 |x diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 150 Figure 4.42a-d Oxygen differentials in full lift aerator on PSA O2 at 1.5 m with 140 JJ. diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 154 Figure 4.43a-d Oxygen differentials in full lift aerator on PSA O2 at 2.9 m with 400 p. diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 156 Figure 4.44a-d Oxygen differentials in full lift aerator on PSA O2 at 1.5 m with 400 u. diffuser at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 40 1min"1 158 Figure 4.45a-d Oxygen differentials in full lift aerator on PSA O2 at 2.9 m with 800 u. diffuser at 10 1 min"1 (a), 10 1 min"1 (b), 20 1 min"1 (c), 30 1 m i n 1 (d) 161 Figure 4.46a-d Oxygen differentials in full lift aerator on PSA O2 at 1.5 m with 800 u. diffuser at 10 1 min"1 (a), 20 1 min"1 (b),30 1 min"1 (c),40 1 min"1 (d) 163 Figure 4.47a-d Oxygen differentials in full lift aerator on air at 2.9 m with 140 \i diffuser and D B C A at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min'Cd) 165 Figure 4.48a-d Oxygen differentials in full lift aerator on air at 2.9 m with 800 p. diffuser and D B C A at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min-'(d) 168 Figure 4.49a-d Oxygen differentials in full lift aerator on air at 2.9 m with 140 u. diffuser and Bubble screen at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 I min"1 (d) 171 Figure 4.50a-d Oxygen differentials in full lift aerator on air at 2.9 m with 800 u. diffuser and Bubble screen at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 401 min"'(d) 173 Figure 4.51a-d Oxygen differentials in full lift aerator on air at 2.9 m with 140 p diffuser and Surface cover at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 lmin"'(d) 175 Figure 4.52a-d Oxygen differentials in full lift aerator on air at 2.9 m with 800 u. diffuser and Surface cover at 10 I min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 401 min"1 (d) 178 Figure 4.53a-d Oxygen differentials in full lift aerator on air at 2.9 m with 140 jo. diffuser and Counter-rotating blades at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 401 min"1 (d) 180 Figure 4.54a-d Oxygen differentials in full lift aerator on air at 2.9 m with 800 u. diffuser and Counter-rotating blades at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c),40 1 min"1 (d) 182 Figure 4.55a-d Oxygen differentials in full lift aerator on P S A 0 2 at 2.9 m with 140 LI diffuser and D B C A at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min'Cd) 184 Figure 4.56a-d Oxygen differentials in full lift aerator on P S A O2 at 2.9 m with 800 LL diffuser and D B C A at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min _ 1(d) 187 Figure 4.57a-d Oxygen differentials in full lift aerator on PSA O2 at 2.9 m with 140 u. diffuser and Bubble screen at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 lmin"'(d) 190 Figure 4.58a-d Oxygen differentials in full lift aerator on P S A O2 at 2.9 m with 800 p diffuser and Bubble screen at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 40 1 min"1 (d) 192 Figure 4.59a-d Oxygen differentials in full lift aerator on P S A O2 at 2.9 m with 140 \i diffuser and Surface cover at 10 1 min"1 (a), 20 1 min' 1 (b), 30 1 min"1 (c), 401 min"1 (d) 194 Figure 4.60a-d Oxygen differentials in full lift aerator on P S A O2 at 2.9 m with 800 u. diffuser and Surface cover at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min' 1 (c), 40 1 min' 1 (d) 197 Figure 4.61a-d Oxygen differentials in full lift aerator on P S A O2 at 2.9 m with 140 LL diffuser and Counter-rotating blades at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min"1 (c), 401 min"1 (d) 199 Figure 4.62a-d Oxygen differentials in full lift aerator on PSA 0 2 at 2.9 m with 800 LI diffuser and Counter-rotating blades at 10 1 min"1 (a), 20 1 min"1 (b), 30 1 min' 1 (c), 401 min"1 (d) 201 Figure 4.63a-d Box and whisker plots showing effect of P S A O2 flowrate on K i ^ o (a), SOTR (b), S A E (c) and SOTE (d) 205 Figure 4.64a-d Box and whisker plots showing effect of water discharge velocity on K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) 206 Figure 4.65a-d Effect of PSA 0 2 flow rate and water discharge velocity on K La2o (a), SOTR (b), S A E (c) and SOTE (d) 207 Figure 4.66a-d Effect of PSA 0 2 flow rate and water discharge velocity on K i ^ o (a), SOTR (b), S A E (c) and SOTE (d) using log transformed data 208 Figure 4.67a-d Effect of PSA O2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-50 cm sec"1) on K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) . 209 Figure 4.68a-d Effect of PSA O2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-50 cm sec"1) on K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) using log transformed data 210 Figure 4.69a-d Effect of PSA O2 flow rate (2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on K L a 2 0 (a), SOTR (b), S A E (c) and SOTE (d) using log transformed data 211 Figure 4.70 Effect of PSA O2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on K i ^ o . 213 Figure 4.71 Effect of P S A O2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on SOTR 213 Figure 4.72 Effect of PSA O2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on S A E 215 Figure 4.73 Effect of PSA O2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on SOTE 215 Figure 5.1 Schematic of Two Film Theory gas transfer process (from Eckenfelder and Ford, 1969) 217 Figure 5.2 Graph of bubble rise velocity vs. radius of equivalent spherical bubble (from Barnhart, 1969) 244 Figure 5.3 Graph of K L VS. bubble diameter (all data and graph from Barnhart, 1969) 246 Figure 5.4 Graph of K^a vs. bubble diameter (from Barnhart, 1969) 246 N O M E N C L A T U R E A Interfacial or absorbing surface area of air (m2) A E Field Aeration Efficiency (g O2 kWhf ' ) A S C E American Society of Civi l Engineers A / V a = the interfacial surface area of the air through which diffusion can occur generated by the particular aeration system per unit volume of water (m 2  m"3) B O D Biochemical Oxygen Demand (mg l"1) B P Barometric pressure (mm Hg) BS Bubble screen treatment (Group 9 and 10) C, Dissolved oxygen concentration at time ti (mg l"1) C 2 Dissolved oxygen concentration at time t 2 (mg l"1) Cj Saturation value of dissolved oxygen at the interface between liquid and air bubble (mg f 1 ) C L Average concentration of dissolved oxygen in the bulk liquid (mg 1"') C m * Measured dissolved oxygen air-solubility value (mg l"1) for the ambient barometric pressure, temperature and vapor pressure of water C s * Dissolved oxygen air-solubility value (mg 1"') for the ambient barometric pressure, temperature and vapor pressure of water C s*76o Dissolved oxygen air-solubility value (mg l"1) for the barometric pressure equal to 760.0 mm Hg and ambient temperature C E M E Civ i l Engineering-Mechanical Engineering, U B C C R B Counter-rotating blades treatment (Group 13 and 14) C S V Comma Separated Variable format for data transmission/storage db Mean bubble diameter 2 1 D L Diffusion coefficient for oxygen (m hr" ) D B C A Downflow Bubble Contact Aeration DOBI Deep Oxygen Bubble Injection e Compressor efficiency = 0.80 in adiabatic compression formula E Q D Equivalent Bubble Diameter (mm) E R Extended run treatment (Group Misc. 2) F P A Fixed Point Averaging program for flow meter g Acceleration due to gravity (980 cm sec"2) G s " Gas flow rate G L M General Linear Model in SYSTAT™ G V R D Greater Vancouver Regional District ID Inside diameter (cm) k Unit (1.395) for air (and oxygen) in adiabatic compression formula K L Liquid film coefficient (m hr"1) K.La Overall oxygen transfer coefficient (hr 1 ) IQ^o Oxygen transfer coefficient at 20 °C (hr _ 1 ) K ^ a i Oxygen transfer coefficient at temperature T °C (hr - 1 ) kW Unit of power measurement 1 Liquid film thickness (m ) n (k-l)/k = 0.283 for air (and oxygen) in adiabatic compression formula N m 3 min"1 Normal cubic meters per minute; metric unit measurement of gas flow @ 1 atmosphere pressure, 0% relative humidity and 0.0 °C O T E C Ocean Thermal Energy Conversion pi Absolute inlet pressure before compression (kg cm" ) P2 Absolute inlet pressure after compression (kg cm" ) PH2O Vapor pressure of water for the ambient temperature (mm Hg) psig Pounds per square inch gauge (lb in" ) P w Power input (kW) PSA Pressure Swing Adsorption r Rate of renewal of liquid film (hr"1) r 2 Correlation coefficient r e Equivalent bubble radius (cm) R Engineering gas constant for air = 8.314 kJ/k mol °K S A E Standard Aeration Efficiency (g 0 2 kWhr"1) SC Surface cover treatment (Group 11 and 12) S C F M Standard cubic feet per minute; American unit measurement of gas flow @ 1 atmosphere pressure, 0% relative humidity and 21.1 °C SE Standard error S E M Scanning Electron Micrograph SOTE Standard oxygen transfer efficiency (%) SOTR Standard Oxygen Transfer Rate (g 0 2 hr'1) ti Time at point 1 on the semi-logarithmic plot (hr) t 2 Time at point 2 on the semi-logarithmic plot (hr) T i Absolute inlet temperature before compression (°K) T E Field Oxygen Transfer Efficiency (%) u Bubble terminal rise velocity (cm sec"1) U B C University of British Columbia V Volume of the liquid (m ) w Weight of air flow ( kg s"1) (i.e., 1.2927 g l"1) Wo2 Mass flow rate of oxygen in the gas flow stream (g 0 2 hr"1) W02 air Mass flow rate of oxygen in air in the gas flow stream (g O2 hr"1) W02 PSA Mass flow rate of oxygen in PSA oxygen the gas flow stream (g 0 2 hr"1) A C K N O W L E D G M E N T S This thesis is a continuation of research started in 1977, in which the project planning for the first hypolimentic aerator in Canada was initiated. Two graduate degrees were earned along the way; a M.Sc. in Zoology in 1981 and a M.A.Sc . in Civi l Engineering in 1989, reflecting the multidisciplinary nature of hypolimnetic aeration. This Doctoral dissertation is the latest milestone in this continuing project, in which the information learned from the preceding research was used to forge a link between the engineering and limnological disciplines, hope-fully for the betterment of both. A number of key individuals assisted in this latest endeavor. M y thesis committee, consisting of Drs. Ken Hall, Greg Lawrence, Don Mavinic and Peter Ward were especially supportive and considerate of the difficulties of conducting full time research while working full time; theoretically an impossible situation. The Civil/Mechanical Engineering workshop built the full lift aerator, their construction skills with Plexiglas rivals many artists. Susan Harper kindly ensured lab space was always available and provided equip-ment on short notice, for extended periods of time. Rob Barrait, Dennis Johnson, Brian Hirsch and Michèle Anderson at Point Four Systems were able to decipher my note pad sketches and build a data logging/monitoring and gas manifold system that was reliable and a pleasure to use. Burrard Mechanical built the Speece Cone from drawings that were less than perfect. Bob Land graciously donated weekend time to build the support and access platform. Special thanks are due to Dr. Hassen Yassien for resolving the problems with the data analysis and breaking the logjam on the mountain of data, without which this thesis would not have been completed. Special thanks also to Dr. Tom Johnston who politely endured my numerous statistical inquires which contributed significantly to the timely completion of this thesis. Thanks to fellow gradu-ate students Venkatram Mahendraker and Violeta Martin for their support. Many people provided advice and encouragement to "just get the damm thing done", especially Harvey Andrusak, Joan Ashley, Dr. Jan Barica, Dr. Eddy Carmack, Barry Chilibeck, Dr. Geir Dahl-Hansen, Dr. Paul Hamblin, Dr. Jan Heggenes, Dave Heller, Dr. Staffan Holmgren, Fred Koch, Dr. Marc Labelle, Gillian Larkin, Bob Lindsay, Ian McGregor, Dr. Gôran Milbrink, C.W. Nicol, Dr. Tom Northcote, Dr. Marvin Rosenau, Pat Slaney, Megan Stirling, Dr. John and Ellie Stockner, Dr. Rick Taylor, Maggie Verity, Bruce Ward and Craig Wightman. To the old soldiers who were there at the start, but unable to see the finish (Frank Ashley, Gordon Arm-strong and Bob Simpson), thanks for everything. Finally, to those who made life difficult along the way, and said it could or should not be done, I have only two words, neither of which is "Good Day". C H A P T E R 1: INTRODUCTION 1.1 Introduction Hypolimnetic aeration has developed into an important, but infrequently used, water quality improvement technique, due to its ability to selectively oxygenate the hypolimnion of stratified lakes and reservoirs in situ, while maintaining thermal stratification. This ability has important applications in civil engineering, limnology and inland fisheries management. For example, within the field of lake restoration (i.e., applied limnology), hypolimnetic aeration is often the only feasible water quality improvement option due to the cost and logistical difficulties of controlling point and non-point source nutrient loading. Even if nutrient abatement measures are to be implemented as part of a comprehensive watershed nutrient management plan, several years of monitoring, planning and construction are often required to complete the task. Hypolimnetic aeration can be used as an interim measure that can be rapidly deployed to improve water quality, while longer-term restorative works are being completed. Hypolimnetic aeration can be used to increase the effectiveness of alternative lake restoration techniques, such as enhancing the distribution of liquid alum applied to co-precipitate phosphorus (Soltero et al., 1994). For fisheries management purposes, hypolimnetic aeration is critically important for restoring cold water fish habitat, as artificial circulation (i.e., destratification) during summer months typically elevates hypolimnetic temperatures above the preferred temperature range for sal-monids and eliminates their thermal refuge. A n aerobic hypolimnion permits the existence of "two story" fisheries, where the epilimnion is inhabited by warm water fish species (e.g., centrarchids) and salmonids inhabit the hypolimnion. Public surveys have consistently indi-cated strong support for this type of multi-species fishery throughout temperate regions of North America. The presence of an aerobic hypolimnion allows re-colonization of the profundal sediments by benthic macroinvertebrates and provides a refuge for planktonic macrozooplank-ton, both of which provide an additional food supply for fish and increase the invertebrate species diversity of the water body. In British Columbia, approx. 30 lakes in the Southern Interior Plateau region have been artifi-cially circulated during the winter to provide an ice fishery and prevent overwinter mortality of trout (Ashley et al., 1992). Artificial circulation via destratification has been the preferred approach (Ashley and Nordin, 1999). However, a recent ruling in 1998 by the B . C . Attorney General's Ministry has interpreted the Criminal Code of Canada (Section 263) to indicate that it is illegal to modify the natural integrity of an ice surface during winter months without 100% fencing of the open ice area created by the destratification process: Criminal Code of Canada, Section 263: "Duty to safeguard opening in ice - Excavations on land - Offences: 263 (1) Everyone who makes or causes to be made an opening in ice that is open to or frequented by the public is under legal duty to guard it in a manner that is adequate to prevent persons by falling in by accident and is adequate to warn them that the opening exists. " "Every one who fails to perform duty imposed by subsection (1) is guilty of manslaughter, if the death of any person results therefrom; (b) an offence under Section 269, if bodily harm to any person results therefrom. " This interpretation has lead to the partial cancellation of B.C. 's winter aeration program, which has been in existence since the mid I960's (Halsey, 1968), due to the cost/safety/logistical impracticalities of installing and dismantling temporary barrier fences annually on approx. 25 rural and semi-wilderness lakes. A n alternative technology is required that will provide the oxygenation equivalent of artificial circulation without modifying the integrity of the ice surface. Hypolimnetic aeration, due to its selective circulation ability, does not create polynas as the bubble-water plume is contained with a floating or submerged aeration chamber. However, the oxygen transfer efficiency of the hypolimnetic aerator must be considerably higher than the destratification systems, as the large open ice area for gas transfer is replaced by a much smaller surface exchange area within the inlet tube and aeration chamber. Hypolimnetic aeration, with compressed air, may not have sufficient oxygenation capacity to cost-effectively meet overwinter oxygen demands, which can be as high as 0.1 mg L" 1 d"1 in the Southern Interior Plateau limnological zone (Ashley et al., 1992). Hence, a high efficiency hypolimnetic aeration system has the potential to re-establish B.C. 's winter lake aeration program. From a civil engineering perspective, oxygenation of anoxic or anaerobic hypolimnetic water from eutrophic lakes and reservoirs can significantly improve domestic and industrial water quality by venting reduced gases (e.g., H2S), oxidizing reduced metals (e.g., Fe + 2 , M n + 2 ) , and establishing a positive oxidation-reduction (i.e., redox) potential in the hypolimnion (Speece, 1971). The presence of an oxidizing environment in the hypolimnion is one of the most critical factors in determining lake or reservoir water quality due to the overriding influence of redox potential on water chemistry. Hypolimnetic aeration can be used as a potable water pre-treatment in combination with existing civil engineering works to provide improved water quality at a fraction of the cost of upgrading or expanding conventional water treatment plants. Hypolimnetic aeration can also be used to selectively oxygenate hypolimnetic releases from stratified eutrophic reservoirs, such that the downstream environment does not experience depressed oxygen concentrations from anoxic water releases. C H A P T E R 2: L I T E R A T U R E REVIEW 2.1 Literature Review 2.1.1 History of Hypolimnetic aeration Hypolimnetic aeration was initially developed in post-war Switzerland as an innovative process to improve the water quality of Lake Bret, which is the water supply for Lausanne (Mercier and Perret, 1949). Water was mechanically pumped to a shore based splash basin where it was aerated and allowed to return by gravity flow through a pipe to the hypolimnion. The concept generated little interest for nearly twenty years, then re-emerged in the I960's due to the pio-neering research efforts of Heniz Bernhardt in West Germany, and Richard Speece and Arlo Fast in the U S A . In these systems, water was airlifted to the surface and oxygenated water returned to the hypolimnion via return tubes. Reports on the selective aeration capabilities of hypolimnetic aeration were published in the primary literature (Bernhardt, 1967; Speece, 1971; Fast, 1971) and the process attracted attention from various professional disciplines. The interdisciplinary nature of hypolimnetic aeration was clearly demonstrated in these initial studies. Bernhardt's (1967) interest was in improving potable water quality in eutrophic reser-voirs, whereas Speece (1971) focused on innovative engineering designs for oxygenating water for lakes, reservoirs and fish hatcheries. In contrast, Fast (1971) emphasised the restoration of eutrophic lakes for fisheries management purposes, and stressed the importance of understand-ing the ecological effects of hypolimnetic aeration. Following the limiting aquatic nutrient controversy of the late 1960's, attention in the 1970's focused on reducing nutrient inputs and developing methods for rehabilitating culturally eutro-phied lakes and reservoirs. Many of these water bodies experienced degraded water quality during the prolonged period of economic activity and population growth that followed World War II. Cultural eutrophication is caused by excessive addition of nutrients such as nitrogen and phosphorus to lakes, rivers, reservoirs, estuaries and coastal waters (Wetzel, 1975). Excessive nutrient loading typically results in increased aquatic plant growth, undesirable changes in species composition, oxygen depletions, fish kills, and decreased water quality for domestic, recreational, and industrial use (Lee, 1970). As a result, the interdisciplinary field of lake restoration blossomed in the 1970's as limnologists and engineers developed new methods for restoring eutrophic lakes and reservoirs. Lake restoration refers to " . . . the manipulation of a lake ecosystem to effect an in-lake improvement in degraded, or undesirable conditions." (Dunst et al., 1974). Coincidentally, hypolimnetic aeration had matured to a stage where it was viewed as suitable technology for the rapidly developing field of lake restoration. Hypolimnetic aeration systems were soon installed in several Western European countries (Bernhardt, 1974; Verner, 1984; Steinberg and Arzet, 1984; Jaeger, 1990; Lappalainen, 1994), in Canada (Taggart and McQueen, 1981; Ashley, 1983; McQueen and Lean, 1983; McQueen and Lean, 1984; Taggart, 1984; Ashley, 1985; McQueen and Lean, 1986; Ashley, 1987) and the United States (Fast, 1973; Fast et al., 1973; Fast et al., 1975; Wirth et al., 1975; Fast, 1975; Smith et al. 1975; Fast and Lorenzen, 1976; Fast et a l , 1976; Garrell et al., 1977; Labaugh, 1980; Fast, 1981). By the late 1980's, hypolimnetic aeration was recognized by perceptive environmental engineers and limnologists as a valuable, multi-purpose water quality restoration technique. In the late 1980's and 1990's, research efforts on hypolimnetic aeration had progressed beyond reporting on individual installations, and tended to focus on four general areas of interest: (1) examining the ecological effects of hypolimnetic aeration with liquid oxygen or Pressure Swing Adsorption (i.e., PSA) generation of oxygen (Prepas et al., 1990; Prepas et al., 1997; Prepas and Burke, 1997; Dinsmore and Prepas, 1997a; Dinsmore and Prepas, 1997b; Field and Prepas, 1997; Webb et al., 1997; Lawrence et al., 1997 and Gemza, 1997); (2) developing mathematical models to predict oxygen transfer rates and aerator performance (Little, 1995); (3) full scale testing of innovative designs e.g., Downflow Bubble Contact Aeration systems (i.e., D B C A or "Speece Cone") (Thomas et al., 1994; Doke et al., 1995) and (4) engineering studies of factors influencing oxygen transfer in full lift hypolimnetic aerators (Ashley and Hall, 1990; Ashley et al., 1990). For the purposes of this discussion, hypolimnetic aeration systems are grouped into three general categories: (1) full or partial lift with compressed air; (2) full or partial lift with pure oxygen or oxygen supplementation and (3) D B C A or Speece Cone aerators. 2.1.2 Types of Hypolimnetic Aerators 2.1.2.1 Full or Partial Lift Hypolimnetic Aeration with Compressed Air A variety of hypolimnetic aerators designs have been proposed and at least 13 designs subjected to full scale testing (Fast and Lorenzen, 1976). However, the conventional full lift "Bernhardt" hypolimnetic aerator (Figure 2.1) has emerged as one of the more widely used designs due to its relatively simple design and availability of published data on site specific oxygen transfer and energy efficiency (Lorenzen and Fast, 1977; Taggart and McQueen, 1982; Ashley, 1985; Ashley et al., 1987; Little, 1995). Several design improvements have been suggested for full lift aerators (Lorenzen and Fast, 1977; Taggart and McQueen, 1982; Ashley, 1989), however, relatively few have been subject to thorough laboratory and field testing. Figure 2.1. Full lift "Bernhardt" type of hypolimnetic aerator (from Soltero et al., 1994) Three design parameters relevant to full lift designs that have undergone limited laboratory and field trials include (1) depth of air injection; (2) separator box surface exchange area and (3) diffuser orifice diameter. Varying the surface exchange area of the separator box was found to have no effect on induced water velocity, oxygen input, daily O2 load, aeration efficiency (AE; g CVkWh" 1) and oxygen transfer efficiency (TE; %) (Ashley and Hall , 1990). The effect of diffuser depth on oxygen input was significant, however the effect was less than expected; but orifice diameter did exert a significant influence on oxygen input, daily oxygen load, TE and A E (Ashley et al., 1990). The importance of diffuser design was clearly demonstrated by a diffuser retrofit on a large Bernhardt style full lift hypolimnetic aerator, where the original 3,175 a diameter diffusers were replaced with 140 u diameter fine bubble silica glass diffusers. This increased the overall average daily oxygen input to St. Mary Lake (Saltspring Island, B.C.) from 311 kg-day"1 to 512 kg-day"1 with no corresponding increase in operational costs (Ashley et al., 1990). Partial lift hypolimnetic aerators have not experienced the widespread application of full lift systems, although the concept is technically sound (Figure 2.2). One reason may be the "over-selling" of their capabilities relative to other designs of hypolimnetic aerators. The submerged design approach is particularly useful in high boat traffic areas and ice-prone regions, but it is more difficult to conduct routine inspections and performance monitoring. In extreme cases, undersized partial lift systems may have to be replaced, as they are less amenable to retrofitting than full lift designs. This was the unfortunate situation in Medical Lake, Washington where a partial lift hypolimnetic aeration system, with a coarse bubble diffuser, was unable to meet the hypolimnetic oxygen demand. After retrofitting with a smaller 600 \i orifice diameter diffuser and a higher output compressor, the partial lift unit was still undersized and Figure 2.2. Partial lift hypolimnetic aerator (from Soltero et al., 1994) then experienced complete structural failure. It was replaced with two Bernhardt style full lift hypolimnetic aerators, which successfully oxygenated the hypolimnion and significantly improved the water quality of Medical Lake (Soltero et al., 1994). 2.1.2.2 Full or Partial Lift Hypolimnetic Aeration with Pure Oxygen or Oxygen Supplementation Pure oxygen has been used in several installations to increase the amount of oxygen transferred to the hypolimnion of eutrophic lakes and reservoirs. Although technically classified as a mechanical pump design of hypolimnetic aeration, Fast and Lorenzen (1976) describe two examples of side stream pumping (SSP) where pure oxygen was added to water pumped from a quarry lake and returned to the hypolimnion via a high pressure discharge line. This method increased the hypolimnetic oxygen concentration in Ottoville Quarry, Ohio to 21.5 mg-L"1 which was the highest hypolimnetic oxygen concentration recorded to date in a stratified water body (Fast et al., 1977). Smith et al., (1975) used supplemental oxygen in Mirror Lake, Wis-consin to overcome the high rate of hypolimnetic oxygen demand, and measured oxygen concentrations in the full lift separator box of 13.4 mg-L"1 on pure oxygen and 8.6 mg-L"1, using a blend of 0.45 m3-min"' compressed air and 0.16 m3-min" ] liquid oxygen. Prepas et al., (1997) successfully used liquid oxygen to oxygenate the north basin of naturally eutrophic Amisk Lake, Alberta using a Deep Oxygen Bubble Injection (DOBI) system origi-nally proposed by Speece (1971). This system injected up to 1.3 t-day"1 of fine oxygen bubbles (20-1,500 u diameter) at 33 m and was able to maintain hypolimnetic oxygen concentrations >1.7 mg-L"1 in summer and >5 mg-L"1 in winter. Gemza (1997) used an innovative full lift aerator design with an electric impeller in the separator box to increase water flow and oxygen transfer. Using Pressure Swing Adsorption (i.e., PSA) generated oxygen, this design oxygen-ated anoxic hypolimnetic water up to 8.0 - 11.5 mg-L"1, and was instrumental in improving the water quality in Heart Lake and Whittaker Lake in Southern Ontario. The main limitations and concerns with the use of liquid oxygen are the high operating costs (Fast et a l , 1976), availabil-ity in remote locations, and site storage/security issues. However, the development of PSA oxygen generation has changed the economics and storage/security issues of using oxygen, and increased the opportunities for innovative application in hypolimnetic aeration (Speece et a l , 1990). 2.1.2.3 Downflow Bubble Contact Aeration A promising new type of hypolimnetic aerator is the "Speece Cone" design (Figure 2.3). This concept was originally proposed in 1971 as a Downflow Bubble Contact Aerator (DBCA) with an open cone (Speece, 1971), then re-designed in 1990-91 with a closed cone and field tested for the first time in 1992 in Newman Lake, Washington (Doke et a l , 1995). An innovative design of hypolimnetic aerator was necessary as Newman Lake is large (490 ha) but quite shallow, with a maximum depth of 9.1 m and a mean depth of only 5.8 m, which was too shallow for conventional partial or full lift hypolimnetic aerator designs. The Speece Cone installed in Newman Lake was 2.8 m in diameter, 5.5 m tall, with a 45 kW submerged axial flow pump for water circulation and two 37 kW air compressors supplying compressed air to two Pressure Swing Adsorption (PSA) on-site oxygen generation units. The system was designed to supply 1,360 kg-d"1 of oxygen to the hypolimnion through a specially designed diffuser to avoid unintentional destratification and sediment disturbance (pers. comm., G. Lawrence, U B C , Vancouver, B.C.). Figure 2.3 Downflow Bubble Contact Aerator "Speece Cone" hypolimnetic aerator The system has performed extraordinarily well to date, with measured oxygen concentrations in the outlet pipe >30 mg-L"1 despite being located in only 8.7 m of water. The system increased average summer hypolimnetic oxygen concentrations to 5.5 mg-L"1 in 1992; however, thermal stratification was less stable due to two severe storm events (Thomas et al., 1994). An even larger Speece Cone (7 m in height), capable of supplying 8,000 kg-d"1 of oxygen, was installed and operated in Camanche Reservoir, California in 1993 and 1994 to improve water quality and prevent periodic fish kills in a salmonid hatchery that was relying on hypolimnetic water discharged from the anoxic hypolimnion. Unpublished reports indicate this system has also performed extremely well to date (pers. comm., A . Home, U . California at Berkeley). 2.2 Rationale for Current Research After nearly 50 years of extensive testing and reporting, hypolimnetic aeration is still infre-quently used by civil engineers to improve potable water quality in eutrophic lakes and reser-voirs. This seems paradoxical in light of evolving regulatory requirements for higher quality raw water supplies and the well documented capabilities of hypolimnetic aeration for improving water quality while maintaining thermal stratification. I believe this situation is partially a result of the interdisciplinary nature of hypolimnetic aeration, where much of the research and report-ing has been conducted by limnologists and fisheries biologists. In addition, the conservative tendency for most civil engineers has been to address raw water quality concerns through conventional and updated water treatment facilities, rather than accept risks with less certain "ecological engineering" approaches in lakes or reservoirs. However, with increasingly strin-gent regulations for permissible concentrations of disinfection by-products eliminating such traditional practices as chlorine pre-treatment and pre-oxidation, it is inevitable that water utilities wil l adopt more pro-active watershed management strategies for their raw water sup-plies. This can be achieved through strict control of watershed land use, implementing storm-water best management practices, and by using in-lake techniques such as hypolimnetic aeration to improve and/or maintain high-quality, raw water. Since hypolimnetic aeration is one of the most effective in-lake water quality improvement techniques, it is necessary to provide updated information to water treatment engineers so that hypolimnetic aeration can be integrated into current and future water quality management strategies. A n afternoon session at the 1998 American Water Works Association Water Quality Technology Conference in San Diego was organised on this theme, and several papers were presented reviewing the various methods available for hypolimnetic aeration and the resultant improvements in raw water quality. Unfortunately, only abstracts of these papers were ever published in the conference handbook. To date, a detailed analysis of design factors influencing the oxygen transfer efficiency of conventional and innovative hypolimnetic aerators has not been conducted. Neither has an economic and engineering analysis been attempted comparing different designs of hypolimnetic aerators and alternate sources of oxygen (e.g., compressed air, liquid oxygen and Pressure Swing Adsorption oxygen) to determine which is the most cost effective type of aerator and source of oxygen. Given the variable cost of electricity in the Pacific Northwest and California markets, it is imperative that energy costs be factored into any potential in-lake water quality improvement strategy, as the cumulative operational and maintenance costs of large electric pumps and compressor motors can be substantial. To their credit, Fast and Lorenzen (1976) and Fast et al., (1976) initiated a comparative analysis in the mid 1970's; however, an update is required given the development of PSA technology as an alternative to cryogenic oxygen and the recent, full-scale tests of the Speece Cone in Newman lake and Camanche Reservoir. At present, there are at least three areas of study that require additional information in order to expand the knowledge base for hypolimnetic aeration: (1) optimising oxygen transfer with compressed air; (2) examining the use of PSA generated oxygen, and (3) understanding the oxygenation characteristics of the Downflow Bubble Contact Aerator i.e., the "Speece Cone". In addition, Ashley (1989) suggested three areas of research that may improve the performance of conventional full lift hypolimnetic aeration systems. These involved (1) installing bubble screens in the inlet tube in an attempt to fragment coalesced bubbles in the rising ainwater plume; (2) increasing bubble:water contact time in the inlet tube through the use of deflectors or vanes and (3) using counter-current air or oxygen injection in the outlet tube, in conjunction with co-current upflow in the inlet tube to increase oxygen transfer per unit energy input. Accordingly, this current research effort conducted a detailed analysis of design factors influencing the oxygen transfer capabilities of two different designs of hypolimnetic aerators: full lift and Speece Cone. The experiments utilised non steady-state gas transfer methodology in a laboratory scale system. Specifically, the research examined the effect of several engineering design factors on oxygen transfer using four standard units of measure for quantifying oxygen transfer: (a) K i ^ o (hr _ 1 ) , the oxygen transfer coefficient at 20 °C; (b) SOTR (g O2 hr"1), the Standard Oxygen Transfer Rate; (c) S A E (g O2 kWhr"1), the Standard Aeration Efficiency and (d) SOTE (%), the Standard Oxygen Transfer Efficiency: The design factors examined in the full lift experiments were as follows: 1. the effect of diffuser orifice pore diameter (140 \x, 400 u and 800 u) on K i ^ o , SOTR, S A E and SOTE; 2. the effect of diffuser submergence (1.5 and 2.9 m) on K i ^ o , SOTR, S A E and SOTE; 3. the effect of gas flow rate (10, 20, 30 and 40 1 min"1) on K L a 2 0 , SOTR, S A E and SOTE; 4. the effect of oxygen partial pressure (air and PSA generated oxygen) on K r ^ o , SOTR, S A E and SOTE ; 5. the effect of a floating surface cover in the separator box (present or absent) on K.La2o, SOTR, S A E and SOTE; 6. the effect of a mid-depth inlet tube bubble screen (i.e., fine strand stainless steel mesh screen of 0.64 cm square openings to fragment rising bubbles) on K i ^ o , SOTR, S A E and SOTE; 7. the effect of mid-depth inlet tube counter-rotating blades (i.e., mechanical propeller blades that rotated with the upwelling water velocity and fragmented rising bubbles) on K L a 2 0 , SOTR, S A E and SOTE; 8. the effect of Downflow Bubble Contact Aeration (DBCA) in the outlet tube in combina-tion with co-current upflow in the inlet tube on K.La2o, SOTR, S A E and SOTE. The design factors examined in the Speece Cone experiments were as follows: 1. the effect of oxygen flow rate (1, 2 and 3 1 min"1) on K.La2o, SOTR, S A E and SOTE; 2. the effect of outlet port discharge water velocity (20, 30, 40, 50, 60 and 70 cm sec"1) on K L a 2 0 , SOTR, S A E and SOTE. The purpose of this research was to determine the effect of several key design factors on oxygen transfer, and determine which combination(s) of design factors was most effective at dissolving oxygen into water under laboratory test conditions. The immediate goal of this research is to provide the necessary engineering design analysis, such that practising engineers and applied limnologists can confidently design and construct cost effective and energy efficient full scale hypolimnetic aeration systems. The ultimate goal is to protect and improve raw water quality in stratified lakes and reservoirs. CHAPTER 3: METHODS 3.1 System Design The entire aeration test facility was custom fabricated as there were no existing facilities in the Civi l Engineering-Mechanical Engineering (CEME) laboratory, at the University of British Columbia (UBC), suitable for bench-scale, non steady-state aeration studies. 3.1.1 Aeration Tank The vessel used for the experiments was a custom fibreglass tank built by Condor Engineering Ltd. (Surrey, BC) . The tank dimensions were 0.92 m wide x 0.47 m deep x 3.25 m high (exte-rior dimensions) and 0.905 m x 0.425 m x 3.205 m (interior dimensions). The tank was filled to a depth of 3.105 m for the tests; hence the operational volume of the tank was 1,194 litres. The tank was equipped with a single 2.54 cm drain valve near the bottom on one side, while the other side was fitted with a 16 cm self-sealing inspection port that could be opened from the exterior when the tank was empty. The tank was designed around two custom cast pieces (300 cm x 30 cm x 25.4 mm) of clear acrylic polymer (Reynolds Polymer Technology Inc., Grand Junction, CO) on the front panel to allow viewing of the tank interior (Figure 3.1). The tank developed a noticeable deflection when filled to capacity; hence a steel I-beam (7 mm thickness) was attached to the front of the tank to reduce deflection of the front panel. During initial test fill/drain cycling, the lateral tank deflection was still sufficient to cause a major leak at an acrylic panel seal. Therefore, a pair of flat steel bars (12 mm thickness) was attached to the sides of the tank at heights of 82 cm and 174 cm and connected by 8 mm diameter threaded steel rods to prevent lateral deflection of the tank sidewalls. In addition, a 5 cm wide nylon Figure 3 .I Schematic diagram of aeration tank strap was placed around the entire tank at 1.5 m and tightened as much as possible with a cam ratchet in an attempt to minimize tank deflection. 3.1.2 Water Supply The water used in the tests was taken directly from the water supply system in the C E M E laboratory, which is usually Capilano Reservoir water delivered by the Greater Vancouver Regional District (GVRD) distribution system to U B C . Capilano Reservoir water is low in total dissolved solids (22 mg l"1), slightly acidic (mean pH of 6.1), and typically of high quality, except during high rainfall events when the water can become turbid due to landslides or deltaic erosion in the upper reaches of Capilano Reservoir (Perrin et al., 2000). No significant turbidity events occurred during the test period. 3.1.3 A i r Supply A 44.74 kW Quincy rotary screw compressor (Model Q N W 260-D) rated at 6,800 1 min"1 (240 Standard Cubic Feet per Minute; SCFM) supplied compressed air. A nominal efficiency filter and water-cooled aftercooler was fitted to the discharge end of the compressor. The compressor was located in the basement of the C E M E laboratory, and air was available throughout the C E M E laboratory at several distribution points. The connection to the experi-mental manifold was via ~10 m of 1.9 cm ID compressed air hose at the nearest distribution point. The compressor was regulated to maintain line pressure at 7.03 kg cm" (100 pounds per square inch gauge; psig), and a pressure switch would activate the compressor when line pressure decreased below 6.3 kg cm"2 (90 psig), and shut off when line pressure reached 7.7 kg cm" (110 psig). 3.1.4 Oxygen Supply Oxygen gas for the experiments was produced by an AS-20 PSA oxygen generator, manufac-tured by the AirSep Corporation (Buffalo, New York). The PSA generator separates oxygen from air using the Pressure Swing Adsorption (PSA) air separation process. The PSA process uses packed beds of a molecular sieve (synthetic zeolite) which attracts (adsorbs) nitrogen at high pressure and releases (desorbs) it at low pressure. The unit has two adsorbent beds, and a set of pressure activated solenoids which cycles the two adsorbent beds between high and low pressure so that as one bed is producing oxygen, the other one is regenerating by releasing nitrogen to the atmosphere. The beds are completely regenerative under normal conditions and wil l last indefinitely. The factory specifications for the AS-20 indicate it requires a minimum air pressure and flow of 6.3 kg cm" (90 psig) and 0.14 Normal cubic meters per minute; N m min" (5.3 SCFM) respec-tively (see Nomenclature for S C F M and N m 3 min"1 definitions), and wil l deliver 0.0088 N m 3 min"1 (0.33 SCFM) at 3.2 kg cm' 2 (45 psig). A 227 1 storage receiver was connected to the 3 1 AirSep unit to allow it to temporarily deliver oxygen in excess of 0.0088 N m min" (0.33 SCFM). The purity of oxygen generated ranges from 90 to 95%. A reference cylinder of certified oxygen gas (> 99.99 % oxygen), obtained from Ai r Liquide Canada, was used to calibrate the oxygen gas probe (i.e., Probe 7) and continuously monitor the purity of the oxygen generated by the AirSep unit. 3.1.5 Gas Delivery and Flow Measurement Point Four Systems Inc. manufactured a custom designed manifold board to distribute and regulate the flow and delivery pressure of the various gases being tested (Figure 3.2). To aerator X Flow meter 1 Air regulator^ y. and gauge \s 1 Flow meter 2 To aerator Flow meter 3 a> o valves Ï X iFrom IPSA •unit Rheostat for speece cone Air to diffuser Reference cylinder of oxygen Figure 3.2. Schematic diagram of manifold board with data logger and flow meters Inflowing compressed air from the C E M E laboratory supply was passed through a Wilkerson 5.0 u particulate and oil/water filter, then a 0.01 p particulate filter and oil/water filter. PSA molecular sieves are very sensitive to contamination so it was imperative that no moisture, particulates or compressor oil reached the synthetic zeolite adsorbent beds. Filtered air was then routed to the PSA unit or straight to the manifold via a pair of on-off ball valves. The filtered gas (air or oxygen) was then routed via separate regulators to an array of three mechanical flow indicators, which could be operated independently, or in any combination. Pressure gauges were fitted to each regulator (air or oxygen) on the manifold, on a single stage regulator fitted to the PSA oxygen receiver and on a two stage regulator fitted to the reference cylinder of high purity compressed oxygen. The coarse scale flow meter was a Brooks Sho-rate flow indicator with 150 mm scale, 2 to 46 1 min"1; the medium scale meter was a Brooks Sho-rate flow indicator with 150 mm scale, 2 to 12 1 min"1; and the fine scale meter was a Key Instruments flow indicator with 80 mm scale, 0 to 3.5 1 min"1. The Brooks flow indicators were designed to operate at 3.2 kg cm"2 (45 psig), and the Key Instruments flow indicator was designed for 3.5 kg cm"2 (50 psig), although it was operated at 3.2 kg cm"2. Pressure regulation for the various gases was achieved by adjusting the compressed air regulator on the manifold board, a single stage regulator on the discharge side of the AirSep oxygen receiver, or a two-stage regulator on the reference oxygen cylinder. Flow meter readings were corrected by a specific density correction factor (i.e., Vl .105/1.0) when operated on compressed air. The Key Instruments meter was corrected by a pressure correction factor (i.e., V59.7/64.7) to compensate for the lower operating pressure (i.e., 3.2 vs. 3.5 kg cm"2). A l l of the flow meters were calibrated for 21 °C; hence no temperature correction factor was required. The manifold board was fitted with two 2-way ball valves so that compressed air, PSA oxygen, or high purity reference oxygen could be delivered to the diffusers. The gas supply for the % volume probe cup (i.e., Probe 7) was located downstream of the three flow indicators, however, the volumetric flow rate of gas to the probe cup was so low that it did not introduce a bias to the various flow meter readings. The gas exiting the probe cup was vented into an 11 cm water filled cylinder so that the gas flow could be directly observed in a bubble stream and adjusted to maintain a low, but constant discharge rate (Figure 3.2). 3.1.6 Water Velocity Measurement Water velocity was measured by a Marsh McBirney Model 2000 flow meter. Since fluid dynamics around the sensor electrodes caused some variation in readings, a Fixed Point Averag-ing (FPA) program in the Model 2000 software dampened the output display. This program calculated an average of velocities recorded over a fixed period of time. A time duration of 120 seconds was used for the averaging period, as this was the maximum time duration available in the meter. 3.1.7 Data Logger An eight channel microprocessor based PT4 Monitor (Point Four Systems Inc., Richmond, B.C.) was used to record and log data collected during the experiments. The unit was config-ured with 6 channels for measuring dissolved oxygen in water (Ch. 1-6), one % oxygen channel (Ch. 7) for measuring the purity of the oxygen gas and one temperature channel (Ch. 8). The PT4 Monitor included a communications package enabling RS 232 connectivity to a portable computer for viewing and downloading data from the PT4's internal non-volatile memory. Data transmission is via ASCII Comma Separated Variable (CSV) format. Data logging for all active channels was selectable from 4 seconds to 18 hours, with a maximum capacity of 8,044 samples. Since the unit was equipped with 8 probes (plus time stamp), the maximum memory capacity was 8,044/9 = 894 readings. The PT4 monitor read the channel values and updated the display at 1.5-2.0 second intervals. During the first half of each scan cycle Channels 1-4 were simultaneously read, and Channels 5-8 were read during the second half of the scan cycle. The PT4 Monitor included user defined sample averaging time and log interval settings. The sample averaging time is the period in which channel values are read, accumulated in the internal memory, and the average value taken. The log interval is the time interval between one set of channel values being logged and the next set. The PT4 monitor was programmed for the first set of experiments with the full lift hypolimnetic aerator (Groups 1-6) to average readings for 10 seconds, and log the averaged values at 20 second intervals, so that 3 readings per minute were recorded. This allowed 5 hours of data logging capability, which was required when measuring low air flow rates with 800 p diffusers at the 1.5 m depth. In the second set of experiments with the full lift aerator (Group 7-14) and the Speece Cone (Group 15-21), the averaging time and log interval was reduced by 50% to 5 and 10 seconds, respectively, so that 6 readings per minute were taken. This decreased the maximum logging period to 2.5 hours which was adequate for this phase of the experiments. Data values were logged until at least 60 % saturation had been reached at the intake sampling probe (i.e., Probe 1) during the full lift experiments, and at least 65 % saturation was reached at the intake sampling probe (i.e., Probe 5) in the Speece Cone experiments. Continuous visual checks were made during each test to ensure the upstream gas pressure remained constant at 3.2 kg cm"2, and that gas flow remained at the desired delivery rate. 3.1.8 QxvGuard Probe Design The PT4 monitor was supplied with seven OxyGuard probes; with six probes measuring dissolved oxygen in the water and one measuring the % oxygen volume of the inlet gas (air, PSA oxygen, or calibration oxygen). The OxyGuard design is a membrane covered self polariz-ing galvanic measuring element with built-in temperature compensation. The probes produce a millivolt output in direct proportion to the amount of oxygen present in the medium in which they are placed. Oxygen diffuses through the membrane onto a silver cathode where it reacts chemically and then combines with a zinc anode. This chemical process produces an electrical current which flows through a built in resistor that converts the current (microamps) into millivolts, which is then transmitted via a two conductor cable to the data logger (Point Four Systems, 1997). The minimum current speed required across the membrane for the in-water probes is 2.5 cm s'1. 3.1.9 Probe Membrane Preliminary tests with higher oxygen flows and fine pore diffusers revealed that dissolved oxygen concentrations in the water were changing too quickly to register with the mg l " 1 mem-branes. Therefore, the Oxyguard probes were configured to measure dissolved oxygen in % saturation, rather than mg l " 1 , as the % saturation probe membranes were 50% thinner than the mg l" 1 membranes, resulting in a faster probe response time. 3.1.10 Water Temperature Water temperature was measured with a dedicated stainless steel thermister probe (Channel 8) on the PT4 Monitor, and a temperature probe on the YSI meter. Both probes were typically within ~0.2 °C, and the PT4 temperature reading was used for the probe calibration procedure. The PT4 temperature probe was checked in an ice bath (0 °C) and room temperature (20 °C) water as determined by a mercury calibration thermometer, and found to be accurate within 0.5 °C. 3.2 Test Procedure The test procedure started each morning by draining and refilling the tank with clean water from the C E M E laboratory supply line. Two Little Giant 5-MSP submersible mixing pumps rated at 4,500 1 hr"1 (1,200 gallon hr"1 each at 0.3 m head) located in diagonal corners on the tank floor, were then turned on and allowed to circulate the tank water for 5-6 minutes. Red food coloring dye was added to the tank on several occasions to examine circulation patterns and determine if any stagnant zones existed. The 1,200 litres of water in the tank was well mixed within an average of 2 minutes, thus confirming initial observations that complete mixing was quickly achieved. The monitoring/data logging equipment was then activated. This included the PT4 data logger and Oxyguard probes, a YSI 57 digital dissolved oxygen-temperature meter, and a Toshiba laptop computer for data transfer and storage. A n electronic balance for weighing deoxygena-tion chemicals was activated on the same circuit. The PSA unit was pressurized with com-pressed air from the C E M E laboratory supply line to full operating pressure (7.03 kg cm"2, 100 psig) and allowed to cycle for several minutes, to pressurize the PSA storage receiver to maxi-mum pressure (4.2 kg cm" , 60 psig). 3.2.1 Oxygen Probe Calibration Two replicate samples of water were then collected from the tank surface in 300 ml B O D bottles, and analyzed in the Environmental Engineering laboratory for dissolved oxygen, using the Winkler titration procedure (Azide modification; Lind, 1979). The two Winkler readings were then averaged to provide the reference oxygen concentration to calibrate the PT4 unit and probes for the day. The YSI meter was calibrated in mg l" 1 to the average of the two Winkler readings. 3.2.2 Calculation of Daily Oxygen Saturation The percent saturation for dissolved oxygen on each test day was calculated according to Colt (1984): (3.1) C s * = C s * 7 60 (BP - P H 2 0 ) / (760.0 - P H 2 o) where: C s * = the dissolved oxygen air-solubility value (mg l"1) for the ambient barometric pressure, temperature and vapor pressure of water; C s*760 = the dissolved oxygen air-solubility value (mg l"1) for the barometric pressure equal to 760.0 mm Hg and ambient temperature; B P = barometric pressure in mm Hg; PH20 = vapor pressure of water in mm Hg for the ambient temperature. Values for Cs*76o and PH2O were taken from reference tables in Colt (1984), and the barometric pressure for each test day was taken from the Vancouver weather station on the Environment Canada WebPage at www.weatheroffice.com. The % oxygen saturation of the test water was then determined by dividing the average dis-solved oxygen concentration from the Winkler titrations by the calculated oxygen saturation (i.e., C s*) on each day: (3.2) ^Winkler 1 (mg l' 1) + Winkler 2 (mg f ' ï)/2 x 100 = % oxygen saturation C s * (mg I"1) The PT4 Monitor was then calibrated with this value using the single point calibration proce-dure outlined in the PT4 software (Point Four Systems, 1997). The probe response was exam-ined from 0 % saturation (sulfite bath) to 100 % saturation (AirLiquide certified > 99.99% oxygen gas) and found to be essentially linear, hence this calibration procedure was satisfactory. Once calibrated, the probes were quite stable, and would remain accurate for several days, although they were re-calibrated on each test day to ensure adjustment was made for daily variations in barometric pressure and water temperature. The purity of the oxygen gas (% volume) produced by the PSA unit was monitored by a dedi-cated probe (Ch 7) in the PT4 Monitor. This probe was calibrated daily using the reference cylinder of certified oxygen gas (> 99.99 % oxygen). The experimental manifold contained a 2 way valve and a cup into which the Channel 7 probe was inserted. The valve was set to allow the reference oxygen to flow to probe 7 for several minutes each day while the test tank was mixing. The probe was then calibrated to this reference standard using the single point calibra-tion procedure in the PT4 software (Point Four Systems, 1997). When the PSA unit became pressurized to its maximum operating pressure (4.2 kg cm"2,60 psig), the 2 way valve was turned to allow PSA oxygen to flow to Probe 7, and the % volume oxygen reading was continu-ously recorded during each reaeration test using PSA gas. The probe also monitored the oxygen concentration in air when tests were being conducted on compressed air. 3.2.3 Deoxygenation Procedure The deoxygenation-oxygenation procedure used was the nonsteady-state reaeration test (ASCE, 1992; Boyd and Watten, 1989). The test water was deoxygenated with 0.1 mg l " 1 of cobalt chloride (Certified Grade of C0CI2 6H2O) and 10.0 mg l" 1 of sodium sulfite (Sulftech Catalyzed Na2SC>3; Code 098-3393) for each 1.0 mg 1"' of dissolved oxygen present in the water (Boyd, 1986) plus an additional 10% weight of Na2SÛ3 to ensure rapid deoxygenation at the colder test temperatures. The cobalt chloride was dissolved in a 1 1 flask, added to the tank and allowed to mix into the test water. Sodium sulfite was then mixed into a slurry in a 1 1 flask, added to the tank water and thoroughly mixed by the two submersible pumps. Theoretically only 7.9 mg l" 1 of sodium sulfite is required for each mg l " 1 of dissolved oxygen; however, due to partial oxidation during mixing, it is necessary to add up to 1.5 times the stoichiometric amount (Beak, 1977). The PT4 Monitor low level alarm lighting system was set to activate at 1.0 % saturation, hence it was possible to detect at a distance when probes 1-6 decreased below 1.0 % saturation. A stopwatch was started as soon as the sodium sulfite slurry was added. The YSI meter and PT4 monitor confirmed that the tank water was rapidly deoxygenated, as the % oxygen saturation invariably declined to < 1.0 % within 2 minutes (full lift experiments) and 2.5 minutes (Speece Cone) respectively. The tank was allow to mix for 6 minutes (full lift experiments) and 8 minutes (Speece Cone) respectively, before reaeration treatments were initiated. Since the oxygen differential between the inlet and outlet of the Speece Cone was often quite large, the oxygen flow was terminated when the % saturation at the intake port (i.e., Probe 5) reached 65 % saturation; however, the mixing pumps remained operating. This allowed the pumps to continue circulating water within the tank and displace oxygenated water out of the Speece Cone into the bulk water. The % saturation recorded at Probe 5 (see Speece Cone schematic) following 2 minutes of circulation with zero oxygen flow was then used to represent the terminal % oxygen saturation for the experiment. The 2 minute period following termina-tion of oxygen flow was not included in the various oxygen transfer calculations as no "new" oxygen was added during this procedure, only redistribution of existing oxygen within the Speece Cone and tank. A maximum of six test runs was conducted on each tank of water, to minimise interference from sodium sulfite accumulation (ASCE, 1992; Beak, 1977). The tank was then drained, refilled, and the procedure was repeated. The tank required approx. 40 minutes to drain, and 45 minutes to refill. Since the TDS of the Capilano Reservoir water supply was so low (i.e., 22 mg l"1), the addition of sodium sulfite did not result in a decrease in oxygen saturation due to increased salinity. For example, the greatest amount of Na2SÛ3 was added during the Speece Cone tests, which resulted in a total addition of 635 g (Group 15), and 626 g (Group 17). This increased the total dissolved solids concentration of the test water by 529 and 521 mg l" 1 respectively during the Group 15 and 17 tests, well below the 1,000 mg l" 1 negligible effect concentration noted by Beak (1977). The sodium sulfite imparted a slight brownish tint to the water, which increased with each successive addition of Na2SC>3. There was no foaming of the water, or any other noticeable changes in the test water. A description of the test procedure used for the full lift and Speece Cone experiments is listed in Appendix 1 and Appendix 2 respectively. 3.3 Full Lift Hypolimnetic Aerator 3.3.1 Aerator Dimensions The full lift hypolimnetic aerator was based on a modified Bernhardt (1974) design, as de-scribed in Ashley (1988). The unit was constructed of clear acrylic, and consisted of an open separator box fitted with inlet and outlet tubes (Figure 3.3). The separator box was 82 cm (1) x 36 cm (w) x 28 cm (h) (outside dimensions). The sides of the box were 6.3 mm thick and the bottom section was 9.15 mm. The inlet and outlet tubes were 282.2 cm in length, with an inside diameter of 19.5 cm, and wall thickness of 3 mm. The tube entrance and exit were cut at 90°, as there was insufficient space in the aeration tank to attach an intake flare or discharge elbow. The tubes were located on the centre-line long axis of the separator box, with a 2 cm setback from the endwalls. The separator box was equipped with adjustable supports to suspend it in the tank. The supports were adjusted so that 8 cm of water was present in the separator box with the air flow turned off. The total operating depth of the system was therefore 8 cm + 282.2 cm (tube length) + 0.9 cm (box bottom section thickness) = 291.1 cm. Since the box was filled to a depth of 310.5 cm, there was 19.4 cm clearance between the bottom of the inlet/outlet tubes and the tank floor. A divider was inserted into the tank between the inlet and outlet tubes to prevent short-circuiting from the outlet tube immediately back into the inlet tube. The divider was constructed of clear 6 mm acrylic, and was 200 cm high and 38 cm in width. The divider rested on the tank floor, so Zone 4<v r Zone 5*^ = i tz Probe 4 Probe 3 Probe 5 Probe 2 Probe 6 Probe 1 < 82 cm „ oO oo- • o» ...o-O ro oo ro ho o 3 Figure 3.3. Schematic diagram of full lift aerator with probe location and aeration zones the two tubes were effectively separated for a vertical distance of 200 cm, leaving 110.5 cm for water exchange over the top of the divider. Dye tests indicated that the divider was an effective barrier to short circuiting, as introduced dye could be seen travelling across the separator box, down the outlet tube, then mixing around the bottom of the tank on the outlet tube side of the tank as the dye was discharged into the bulk liquid. The dye would then move upwards over the top of the divider, and flow towards the entrance of the inlet tube and into the inflow tube. The oxygen probes confirmed the visual mixing pattern as the two submersible mixing pumps (4,500 1 hr"1), located in diagonal corners on the tank floor, completely mixed the tank to < 1% saturation, within an average of 2 minutes during actual tests (which was the same time period observed to completely disperse the dye). 3.3.2 Probe Locations The tubes were fitted with oxygen probe insertion ports near both ends of the tubes, and at the mid-point of each tube. The probes were numbered according to their position in the aerator (see Figure 3.3): Probe 1 - 8 cm from the inlet of the inflow tube, operating depth - 283 cm Probe 2 - midpoint of the inflow tube, operating depth - 150 cm Probe 3 - 8 cm from the outlet of the inflow tube, operating depth - 17 cm Probe 4 - 8 cm from the inlet of the outflow tube, operating depth - 17 cm Probe 5 - midpoint of the outflow tube, operating depth - 150 cm Probe 6 - 8 cm from the outlet of the outflow tube, operating depth - 283 cm Probe 7 was the % vol/vol probe, located in a probe cup on the manifold board (see Figure 3.2) and Probe 8 was the temperature probe located at a depth of 3.0 m. The velocity sensor was suspended on a wire frame in the centre of the downflow tube, 28 cm from the junction with the floor of the separator box (see Figure 3.3). 3.3.3 Diffuser Description Four different diffusers were used in the full lift experiments. The fine pore diffusers were made of fused silica glass with a rated maximum orifice diameter of 140 u (Model AS-8, Aquatic Ecosystems Inc., Apopka, FL) or 2 \i (Model M B D 600, Point Four Systems Inc., Richmond, B.C.) (Figure 3.4). Scanning electron micrographs (SEMs) of the diffusers con-firmed the maximum pore size was in the stated size range (Ashley et. al., 1991). The external dimensions of the 140 p diffusers were 7.6 cm L x 3.8 cm W x 3.8 cm D. Each diffuser weighed 0.18 kg and was fitted with a 6.4 mm hose nipple. The 2 u diffuser was 70 cm L x 8.3 cm W x 2.8 cm D, and weighed 2.3 kg. Two sizes of coarse bubble diffusers were used: 397 u (0.4 mm) and 794 u (0.8 mm) diameter pore size. The diffusers were made of 12.7 mm I.D. white Sch. 40 P V C pipe. The diffusers were cross-shaped, with four arms joining into a common center (Figure 3.5). The center was fitted with a 6.4 mm diameter nipple for attaching a 6.4 mm vinyl tubing airline. The outside diameter of the diffusers was 19 cm, and they fit within the inflow tube with 0.3 mm clearance on either side. The diffusers were suspended in the centre of the inflow tube with a 0.45 kg weight, to counter their positive buoyancy and to minimize their movement when discharging gas. Since the surface area of a circle increases four fold as the diameter is doubled, the number of orifices required in each diffuser to maintain an equivalent cross-sectional orifice discharge area was as follows: 397 [i (0.4 mm) = 64 orifices (16 on each arm); 794 u (0.8 mm) = 16 orifices (4 on each arm). A l l three diffusers (140 u, 397 u and 800 u) were suspended at two depths during the experi-ments: 1.5 m and 2.9 m. The 2 u diffuser was positioned on the floor of the separator box, at a depth of 11 cm during the experiments, to introduce gas bubbles into the return flow of the outlet tube by the circulating water. 3.4 Experimental Design The treatments examined in the full lift experiments were as follows: 1. the effect of orifice pore diameter (140 u, 400 u and 800 u); 2. the effect of diffuser depth (1.5 and 2.9 m); 3. the effect of gas flow rate (10, 20, 30 and 40) 1 min"1; 4. the effect of oxygen partial pressure (air and PSA oxygen); 5. the effect of a surface cover in the separator box (present or absent); 6. the effect of a mid-depth inlet tube bubble screen (i.e., fine strand stainless steel mesh screen of 0.64 cm opening); 7. the effect of mid-depth inlet tube counter-rotating blades (i.e., mechanical propeller blades that rotated with the upwelling water velocity); 8. the effect of Downflow Bubble Contact Aeration (DBCA) in the outlet tube. This resulted in a large number of possible treatment combinations. In order to examine all of the various combinations in a realistic time frame, the treatments were sub-divided into 14 experimental groups; Group 1-6 being the "base case" study group where diffuser pore diame-ter, diffuser depth, gas flow rate and gas composition were examined. Groups 7-14 were variations on the "base case" conditions where various hypothesised "enhancements" (i.e., bubble screen, counter-rotating blades and Downflow Bubble Contact Aeration) and surface cover were examined. The experiments in Groups 7-14 were conducted using only two orifice pore diameters (140 \i and 800 u) and one diffuser depth (2.9 m). In addition, two miscellaneous groups were examined. In Misc. 1, the aeration system was operated in D B C A mode with only the mixing pumps to circulate the water without the airlift pump. In Misc. 2, the system was run for an extended period of time to determine how the calculated oxygen saturation (i.e., C* s) compared to the measured oxygen saturation (i.e., C* m ) on air, and the % saturation achievable on oxygen. This resulted in 14 principal treatment groups, and .two minor treatment groups (Table 3.1). Table 3.1. List of treatment groups for full lift hypolimnetic aeration experiments. Group No. Gas flow (I min"1) Orifice diameter (u.) Depth (m) Gas composition Enhancements 1 10, 20, 30, 40 140 2.9 air/oxygen n/a 2 10, 20,30, 40 140 1.5 air/oxygen n/a 3 10, 20, 30, 40 400 2.9 air/oxygen n/a 4 10, 20, 30, 40 400 1.5 air/oxygen n/a 5 10, 20, 30, 40 800 2.9 air/oxygen n/a 6 10, 20, 30, 40 800 1.5 air/oxygen n/a 7 10, 20, 30, 40 140 2.9 air/oxygen D B C A 8 10,20, 30, 40 800 2.9 air/oxygen D B C A 9 10, 20, 30, 40 140 2.9 air/oxygen Bubble screen 10 10, 20, 30, 40 800 2.9 air/oxygen Bubble screen 11 10, 20, 30, 40 140 2.9 air/oxygen Surface cover 12 10, 20, 30, 40 800 2.9 air/oxygen Surface cover 13 10, 20, 30, 40 140 2.9 air/oxygen Rotating blades 14 10, 20, 30, 40 800 2.9 air/oxygen Rotating blades Misc. 1 3 2 n/a air/oxygen Pump/DBCA Misc. 2 10 2/140 2.9 air/oxygen Extended/DBCA The experiments were carried out in a randomised complete block design. Gas flow rate and gas composition could be quickly altered during the experiments hence these two variables were manipulated in each treatment group, while diffuser orifice diameter and diffuser depth (Groups 1-6) or diffuser orifice diameter, diffuser depth and enhancement (Groups 7-14) remained fixed. This resulted in eight possible combinations of gas flow rate x gas composition for each diffuser depth and orifice diameter. The experimental combinations were as follows (Tables 3.2-3.17): Table 3.2. Experimental combinations for Group 1. No. Gas flow (1 min"1) Gas composition Depth (m) Orifice diameter (u.) 1 10 A i r 2.9 140 2 10 PSA 2.9 140 3 20 A i r 2.9 140 4 20 PSA 2.9 140 5 30 A i r 2.9 140 6 30 PSA 2.9 140 7 40 A i r 2.9 140 8 40 P S A 2.9 140 Table 3.3. Experimental combinations for Group 2. No. Gas flow (1 min"1) Gas composition Depth (m) Orifice diameter (u) 1 10 A i r 1.5 140 2 10 PSA 1.5 140 3 20 Ai r 1.5 140 4 20 P S A 1.5 140 5 30 A i r 1.5 140 6 30 PSA 1.5 140 7 40 A i r 1.5 140 8 40 PSA 1.5 140 Table 3.4. Experimental combinations for Group 3. No. Gas flow (1 min"1) Gas composition Depth (m) Orifice diameter (u) 1 10 A i r 2.9 400 2 10 PSA 2.9 400 3 20 A i r 2.9 400 4 20 P S A 2.9 400 5 30 A i r 2.9 400 6 30 P S A 2.9 400 7 40 A i r 2.9 400 8 40 P S A 2.9 400 No. Gas flow (1 min"1) Gas composition Depth (m) Orifice diameter (u.) 1 10 Ai r 1.5 400 2 10 PSA 1.5 400 3 20 A i r 1.5 400 4 20 P S A 1.5 400 5 30 A i r 1.5 400 6 30 PSA 1.5 400 7 40 A i r 1.5 400 8 40 P S A 1.5 400 Table 3.6. Experimental combinations for Group 5. No. Gas flow (1 min"1) Gas composition Depth (m) Orifice diameter (u,) 1 10 A i r 2.9 800 2 10 P S A 2.9 800 3 20 A i r 2.9 800 4 20 PSA 2.9 800 5 30 A i r 2.9 800 6 30 P S A 2.9 800 7 40 A i r 2.9 800 8 40 P S A 2.9 800 Table 3.7. Experimental combinations for Group 6. No. Gas flow (1 min"1) Gas composition Depth (m) Orifice diameter (u) 1 10 A i r 1.5 800 2 10 PSA 1.5 800 3 20 A i r 1.5 800 4 20 PSA 1.5 800 5 30 A i r 1.5 800 6 30 PSA 1.5 800 7 40 Ai r 1.5 800 8 40 P S A 1.5 800 Table 3.8. Experimental combinations for Group 7 (Downflow Bubble Contact Aeration).* No. Gas flow (1 min"1) Gas composition Depth (m) DBCA Orifice diameter (u.) 1 10 (7/3) A i r 2.9 Yes 2/140 2 10 (7/3) P S A 2.9 Yes 2/140 3 20 (17/3) A i r 2.9 Yes 2/140 4 20 (17/3) PSA 2.9 Yes 2/140 5 30 (27/3) Ai r 2.9 Yes 2/140 6 30 (27/3) P S A 2.9 Yes 2/140 7 40 (37/3) A i r 2.9 Yes 2/140 8 40 (37/3) P S A 2.9 Yes 2/140 Table 3.9. Experimental combinations for Group 8 (Downflow Bubble Contact Aeration). No. Gas flow (1 min"1) Gas composition Depth (m) DBCA Orifice diameter (u.) 1 10(7/3) A i r 2.9 Yes 2/800 2 10(7/3) P S A 2.9 Yes 2/800 3 20(17/3) A i r 2.9 Yes 2/800 4 20 (17/3) PSA 2.9 Yes 2/800 5 30 (27/3) A i r 2.9 Yes 2/800 6 30 (27/3) P S A 2.9 Yes 2/800 7 40 (37/3) A i r 2.9 Yes 2/800 8 40 (37/3) P S A 2.9 Yes 2/800 Table 3.10. Experimental combinations for Group 9 (Bubble screen). No. Gas flow (1 min"1) Gas composition Depth (m) Screen Orifice diameter (LI) 1 10 Ai r 2.9 Yes 140 2 10 PSA 2.9 Yes 140 3 20 A i r 2.9 Yes 140 4 20 P S A 2.9 Yes 140 5 30 A i r 2.9 Yes 140 6 30 P S A 2.9 Yes 140 7 40 A i r 2.9 Yes 140 8 40 PSA 2.9 Yes 140 *7/3 refers to gas flow (1 min"1) to the diffuser in the inlet tube, and the 2 LL diffuser in the separator box. No. Gas flow (1 min"1) Gas composition Depth (m) Screen Orifice diameter (n) 1 10 A i r 2.9 Yes 800 2 10 P S A 2.9 Yes 800 3 20 A i r 2.9 Yes 800 4 20 PSA 2.9 Yes 800 5 30 Ai r 2.9 Yes 800 6 30 PSA 2.9 Yes 800 7 40 Ai r 2.9 Yes 800 8 40 P S A 2.9 Yes 800 Table 3.12. Experimental combinations for Group 11 (Surface cover). No. Gas flow (1 min"1) Gas composition Depth (m) SC Orifice diameter (u,) 1 10 Ai r 2.9 Yes 140 2 10 P S A 2.9 Yes 140 3 20 A i r 2.9 Yes 140 4 20 P S A 2.9 Yes 140 5 30 A i r 2.9 Yes 140 6 30 P S A 2.9 Yes 140 7 40 A i r 2.9 Yes 140 8 40 PSA 2.9 Yes 140 Table 3.13. Experimental combinations for Group 12 (Surface cover). No. Gas flow (1 min"1) Gas composition Depth (m) SC Orifice diameter (u.) 1 10 A i r 2.9 Yes 800 2 10 P S A 2.9 Yes 800 3 20 A i r 2.9 Yes 800 4 20 PSA 2.9 Yes 800 5 30 A i r 2.9 Yes 800 6 30 PSA 2.9 Yes 800 7 40 Ai r 2.9 Yes 800 8 40 PSA 2.9 Yes 800 No. Gas flow (1 min"1) Gas composition Depth (m) CRB Orifice diameter (u.) 1 10 Ai r 2.9 Yes 140 2 10 P S A 2.9 Yes 140 3 20 A i r 2.9 Yes 140 4 20 P S A 2.9 Yes 140 5 30 A i r 2.9 Yes 140 6 30 PSA 2.9 Yes 140 7 40 A i r 2.9 Yes 140 8 40 P S A 2.9 Yes 140 Table 3.15. Experimental combinations for Group 14 (Counter-rotating blades). No. Gas flow (1 min"1) Gas composition Depth (m) CRB Orifice diameter (u) 1 10 A i r 2.9 Yes 800 2 10 PSA 2.9 Yes 800 3 20 A i r 2.9 Yes 800 4 20 PSA 2.9 Yes 800 5 30 A i r 2.9 Yes 800 6 30 P S A 2.9 Yes 800 7 40 A i r 2.9 Yes 800 8 40 P S A 2.9 Yes 800 Table 3.16. Experimental combinations for Group Misc. 1 (Pump only/DBCA). No. Gas flow (1 min"1) Gas composition Depth (m) Pump/ DBCA Orifice diameter (V) 1 3 Ai r n/a Yes 2 2 3 PSA n/a Yes 2 Table 3.17. Experimental combinations for Group Misc. 2 (Extended run/DBCA). No. Gas flow (1 min"1) Gas composition Depth (m) ER/ DBCA Orifice diameter (u) 1 10(7/3) A i r 2.9 Yes 2/140 2 10(7/3) PSA 2.9 Yes 2/140 Each of the combinations, within a principal treatment group (i.e., Groups 1-14), was randomly assigned a number from 1 to 8. Each set of 8 combinations was completed, then repeated with a different set of random numbers (Rohlf and Sokal, 1969). The purpose of this design was to remove random error that may occur during any given treatment day. Each principle treatment was replicated 3 times, the minimum number of replicates recommended for non-steady state reaeration tests (ASCE, 1992). A total of 374 individual reaeration tests were completed for the full lift experiments (Table 3.18): Table 3.18. List of treatment combinations for the full lift hypolimnetic aeration experiments. Group No. No. of combinations Replicates Total number of tests 1 8 3 24 2 8 3 24 3 8 3 24 4 8 3 24 5 8 3 24 6 8 3 24 7 8 3 24 8 8 3 24 9 8 3 24 10 8 3 24 11 8 3 24 12 8 3 24 13 8 3 24 13a (repeated) 8 3 24 14 8 3 24 Sub-total 360 Misc. 1 1 7 0 2 /3 air 10 Misc. 2 1 2 0 2 /2 air 4 Sub-total 14 Total 374 3.4.1 Termination of Experiments Each experimental test was terminated when the % saturation recorded at Probe 1 (i.e., the intake to the inlet tube, see Figure 3.3) reached 60% saturation. This represented the % oxygen saturation of the bulk water in the test tank at the end of the test run. The % saturation reading of the remaining probes within the full lift aerator were typically greater than 60% saturation, as they were subject to the experimental combination being tested, which raised the % saturation above the ambient % saturation in the bulk tank water. 3.5 Downflow Bubble Contact Aerator (Speece Cone) Hypolimnetic Aerator 3.5.1 Aerator Dimensions The Downflow Bubble Contact Aerator (DBCA) or "Speece Cone" was based on blueprint drawings of the first full sized D B C A which was installed in Newman Lake, W A in 1992 (Doke et a l , 1995) (Figure 2.3). The unit was constructed by Burrard Mechanical Ltd. (North Van-couver, B.C.) and consisted of a submersible pump and intake port, a D B C A cone and an outlet port (Figure 3.6). The intake port was connected to the submersible pump by a length (40 cm) of clear 5.08 cm dia. vinyl hose. The aluminum housing portion of the intake port tapered from 5.08 cm dia. (entrance) to 6.35 cm dia. (exit) and turned 180° to connect to the top of the cone through a 6.35 cm dia. flange. The cone was 75 cm (H) x 34 cm (base dia.) with a clear acrylic viewing panel (12 cm x 30 cm) in the middle of the tapered portion of the cone. The lower section of the cone was 15 cm high, and the tapered upper section was 60 cm in height, termi-nating at the 6.35 cm diameter connection to the intake port. The outlet port was 25 cm in horizontal length, and 6.35 cm in diameter, with a 90° upward facing elbow at the distal end of the outlet port. The unit was constructed of aluminum, and was attached to a 80 cm x 34 cm base plate. The entire unit was lowered to the bottom of the tank by a pulley system, hence the depth of the intake port (Probe 5) and outlet port (Probe 6) was 239 cm and 295 cm respectively, and the operating depth of the cone ranged from 310.5 cm (cone bottom) to 235.5 cm (cone top). Figure 3.6. Schematic of Speece Cone aerator with probe locations 3.5.2 Tank Mixing and Configuration of Submersible Pumps The same divider used in the full lift experiments (clear 6 mm acrylic, 200 cm H x 38 cm W) was inserted into the tank, to prevent short circuiting of water discharged from the cone outlet port. The separator was positioned on the upper surface of the outlet port between the cone and the 90° elbow of the outlet port, effectively separating the outlet and intake ports over a vertical distance of 200 cm (leaving 110.5 cm for water exchange over the top of the divider). Dye tests indicated the diverter plate was an effective barrier to short circuiting, as dye could be seen discharging from the outlet port, moving upwards on the outlet side of the divider plate, then over the top of the plate and down to the main submersible pump. Initial dye tests indicated the bulk water in the tank was not mixing as quickly as with the full lift hypolimnetic aeration system, as the airlift pump was not operating. Therefore, the location of the two 4,500 1 hr"1 submersible pumps was changed so that both pumps were positioned on the tank floor, on the outlet side of the divider, discharging vertically at depths of 1.22 m and 2.02 m. In addition, a third 4,500 1 hr"1 submersible pump was added near the surface on the inlet side, and oriented so that it discharged vertically downwards at a depth of 0.85 m. The main circulating pump for the Speece Cone was a 10,400 1 hr"1 submersible pump (Little Giant 6-CIM-R) connected to a Variac voltage control unit that allowed pump voltage to be varied from 0 to 110% of rated power. Dye tests and the oxygen probes confirmed the tank (1,194 1) was adequately mixed with all four pumps operating (maximum pumping rate = 23,900 1 hr"1), since the oxygen saturation throughout the tank was reduced to < 1% within an average of 3 minutes, following the addition of a sodium sulfite solution. 3.5.3 Probe Locations The velocity sensor for the Speece Cone was positioned in the centre of the discharge flow, 10 cm from the end of the outlet port. The oxygen probes were arranged in the tank to ensure the tank was completed mixed, as incomplete tank mixing tank invalidates the non steady-state reaeration test (ASCE, 1992). The probes were numbered according to their position in the tank and the Speece Cone (Figure 3.6): Probe 1 - suspended in the tank at 1.5 m; Probe 2 - suspended in the tank at 0.5 m; Probe 3 - suspended in the tank at 1.0 m; Probe 4 - suspended in the tank at 2.0 m; Probe 5 - inlet to the Speece Cone, operating depth - 2.39 m Probe 6 - outlet to the Speece Cone, operating depth - 2.95 m. The remaining two probes were identical to the full lift experiments, i.e., Probe 7 was the % vol/vol probe, located in a probe cup on the manifold board and Probe 8 was the temperature probe located at a depth of 3.0 m (see Figure 3.2). 3.5.4 Diffuser Description There was no diffuser fitted to the Speece Cone. Oxygen entered the cone at an operating depth of 2.3 m, immediately downstream of the 180° elbow in the inlet port, through a 6.4 mm brass fitting and hose nipple connected to the PSA unit (Figure 3.6). 3.6 Experimental Design The treatments examined in the Speece Cone experiments were as follows: 1. the effect of oxygen flow rate (1,2 and 3) 1 min"1 ; 2. the effect of cone discharge velocity (20, 30, 40, 50, 60 and 70 cm sec"1). This was equivalent to pumping rates of 38, 57, 76, 95, 114, and 133 1 min"1. The treatments were arranged into 6 experimental groups, in which the depth of the Speece Cone (i.e., max. tank depth) and gas mixture (i.e., PSA generated oxygen) remained constant while gas flow rate and discharge velocity were varied. In addition, one miscellaneous group was examined where the Speece Cone was operated on compressed air at 3 1 min"1 at a discharge velocity of 50 cm sec"1. This resulted in 6 principal treatment groups, and one minor treatment group (Table 3.19): Table 3.19. List of treatment groups for Speece Cone experiments. Group No. Discharge velocity (cm sec"1) PSA 0 2 flow rate (1 min"1) 15 20 1,2,3 16 30 1,2,3 17 40 1,2,3 18 50 1,2,3 19 60 1,2,3 20 70 1,2,3 Misc. 1 50 3 (air) The experiments were carried out in a randomized complete block design. This resulted in three combinations of gas flow rate x discharge velocity. Each set of 3 combinations was completed, then repeated with a different set of random numbers. The purpose of this design was to remove random error that may occur during any given treatment day. Each treatment was replicated 4 times, 3 being the minimum replicate recommended for non-steadystate reaeration tests (ASCE, 1992). A total of 76 individual reaeration tests were completed (Table 3.20). Table 3.20. List of treatment combinations for Speece Cone experiments. Group No. No. of combinations Replicates Total number of tests 15 3 4 12 16 3 4 12 17 3 4 12 18 3 4 12 19 3 4 12 20 3 4 12 Sub-total 72 Misc. 1 1 4 4 Sub-total 4 Total 76 3.6.1 Termination of Experiments Each experimental test was terminated when the % saturation recorded at Probe 5 (see Figure 3.6) reached 65% saturation. This represented the lowest % oxygen saturation of the bulk water in the test tank at the termination of the test run. The % saturation reading of the remaining probes within the tank and Speece Cone were typically greater than 65% saturation, since they were subject to the experimental combination being tested and influence of the tank divider (which raised the % saturation above the average % saturation in the bulk water on the inlet side). The % saturation recorded at Probe 5, following 2 minutes of pump circulation with zero oxygen flow, was then used to represent the terminal % oxygen saturation for the experiment. This value was always > 65% saturation due to the averaging effect of oxygenated water displaced from the Speece Cone and oxygenated water from the outlet side of the tank. 3.7 Bubble Size Measurement Bubble size was determined by photographing rising bubbles in the tank with a 35 mm camera and flash attachment, synchronized at 1/100 second. A meter ruler graduated in 1 mm incre-ments was suspended in the tank and bubbles were photographed against the ruler for scale. The 35 mm slides were then examined with a Wild dissecting microscope at 25x using an ocular micrometer to determine bubble size. A minimum of 30 bubbles was counted for each orifice size and gas flow rate setting. Only bubbles that were located between the cylinder wall and the ruler, and were either on or touch-ing the ruler were used for measurement. A spreadsheet program was written to calculate the volume of the bubbles. Since many bubbles were oblate spheriod in shape, the following formula was used to calculate volume: (3.3) V = 4/3 na2b Where: V = volume in mm ; a = Vi long axis of the bubble (mm); b = Vi short axis of the bubble (mm); The equivalent diameter of the bubbles was calculated according to: (3.4) where: d = equivalent diameter (mm); V=volume in mm . The mean bubble size, standard deviation and coefficient of variation was also calculated for each orifice size and gas flow rate. 3.8 Electrical measurements and power calculations The electrical current and line voltage for the PSA unit and Speece Cone main pump was measured with a Data Hold digital clamp meter (Model D M 266) and Fluke digital multi-meter. PSA current draw was measured during oxygen production and nitrogen venting cycles, and the average measured amperage (i.e., 0.15 amp) was multiplied by the line voltage (i.e., 110 volts) to determine the energy requirement of the PSA unit (i.e., 16.5 watts). The Speece Cone main pump power requirements were calculated by taking the average of two amperage readings and three voltage readings at each 0.5 increment on the Variac voltage control unit (0-100 scale) from 65 to 85. During the Speece Cone tests, the Variac voltage control unit setting at the end of each test was recorded, and the mean setting determined for each measured water discharge velocity x gas flow combination (i.e, 4 readings). A linear regression of control unit setting and measured energy consumption was then calculated, and the resulting equation used to predict the energy consumption for each of the mean water discharge velocity settings used in the Speece Cone tests (Figure 3.7). The regression equation is as follows: (3.5) where: y = , x = y = 0.0153x- 0.3386; r 2 = 0.99 kW energy consumption of the Speece Cone main pump; Variac voltage control unit setting Speece Cone Power Consumption v - 0.0153X -0.338B R2 = 0.9857 1 20 -i 1 1 1 1 1 1.00 0.80 0.40 0.20 0.00 J 1 1 1 1 1 1 55.0 60.0 65.0 70.0 75.0 80.0 85.0 90.0 Rheostat setting I kW ~ ~ * Linear (kW) | Figure 3.7. Speece Cone main water pump power consumption vs. rheostat setting 3.9 Parameter Estimation Procedure The log deficit method of parameter estimation was used to determine K.La as the experimental tests were terminated at 60-65% saturation, due to the time impracticality of running all tests to 98% of saturation, as is mandatory for the nonlinear regression method of parameter estimation (ASCE, 1992). The nonlinear regression method is the recommended A S C E method of pa-rameter estimation for the measurement of oxygen transfer in clean water; "however, i f the engineer/owner so specifies, the log deficit method described in Annex E shall be permitted in lieu of the nonlinear regression method." (ASCE, 1992). Prior to adopting the log deficit method, a comparative analysis of K.La and SOTR was con-ducted for one group of the full lift experiments (i.e., Group 5) on air and PSA generated oxygen using three methods of parameter estimation: the nonlinear regression method, the log deficit method and a "mass balance" method. The mass balance method used the weighted oxygen values of all six oxygen probes and calculated the mass of oxygen in the test tank at each time stamp of the data logger. The SOTR derived from the log deficit method and nonlinear regres-sion method were compared with the SOTR derived from the actual mass balance method. The results indicated the log deficit method was more accurately associated with the mass balance method than the nonlinear regression method. In addition, a number of other problems were found with the nonlinear regression method: 1. the nonlinear regression method was very sensitive to low end truncation, and required up to 18-20% low end truncation in order to get convergence and minimize the sum of squares. Since the tests were terminated at 60-65% saturation, this deleted too much of the experimental data set; 2. unique estimates of K L a could be obtained simply by selecting different low end trunca-tion points; 3. the nonlinear estimation method did not give realistic results for some probes when calculating K^a for all 6 probes; 4. the nonlinear estimation method did not give realistic results when calculating K^a for the experiments using PSA generated oxygen. In comparison, the log deficit method exhibited minimal sensitivity to low end truncation, worked for all six probes and derived realistic estimates from the experiments using PSA oxygen. The log deficit method is the recommended technique for parameter estimation within the aquacultural and aquatic sciences community (Boyd, 1986; Boyd and Watten, 1989) and was listed as a tentative standard in the 15 t h Edition of Standard Methods ( A P H A , 1980). Therefore, the log deficit method of parameter estimation was selected to determine K^a for these experiments. KLaj was calculated according to: (3.6) K L a T = In iïC/-C1)/C*-C2')1 t2-ti where: Kiaj = oxygen transfer coefficient at the temperature of the test water (hf 1); In = natural logarithm; C s * = the dissolved oxygen air-solubility value (mg l"1) for the ambient barometric pressure, temperature and vapor pressure of water; C i = dissolved oxygen concentration at time ti (mg l"1); C 2 = dissolved oxygen concentration at time t 2 (mg l"1); ti = time at point 1 on the semi-logarithmic plot (hr); t2 = time at point 2 on the semi-logarithmic plot (hr); Ti and t2 are usually chosen as the times at which the measured oxygen concentration is 20% (ti) and 80% (t2) of the saturation value for the test water, corrected for temperature, barometric pressure and vapor pressure of water. This study used approx. 10% (ti) and 60% (t2) (full lift experiments) and 65% (t2) (Speece Cone experiments) saturation values for ti and t2 since a sufficient number of data points (i.e., 11) was collected and it was not practical, from a time perspective, to run each test to 80% saturation. K L ax was corrected to KLa2o according to (ASCE, 1992): (3.7) K L a 2 o = K L a T e ( 2 0 ^ where: T = test water temperature in °C; 0 = 1.024. SOTR was calculated as: (3.8) SOTR = K L a 2 o C s 2 o V J where: SOTR = standard oxygen transfer rate (g 0 2 hr"1); K La2o = oxygen transfer coefficient at 20 °C (hr"1); CS2o = dissolved oxygen concentration (mg l"1) at saturation for 20 °C and standard pressure (760 mm Hg); V = volume of water in the tank (m 3); Standard Aeration Efficiency (SAE) was calculated as: (3.9) S A E = SOTR/power input where: S A E = standard aeration efficiency (g O2 kWhr"1); SOTR = standard oxygen transfer rate (g O2 hr"1); Power input (kW) = total delivered power (kW). The calculation of power input for the various experimental treatments is described in the next section. 3.9.1 Full Lift Experiments For the full lift hypolimnetic aeration experiments, Groups 1-14, on air, the power input was separated into two components. The first component, E i , is the delivered blower power which is the "theoretical power required at a blower discharge to deliver a given mass flow of gas at a given discharge pressure, calculated based upon adiabatic compression" (ASCE, 1992). On air, this pressure was taken as the absolute delivery pressure from the manifold board regulator (i.e., 4.2 kg cm"2). The second component, E2, is the same adiabatic compression formula adjusted to the absolute ambient hydrostatic pressure of gas release (i.e., 1.5 or 2.9 m), rather than the absolute delivery pressure at the manifold board. The sum of E i and E2 equals the total deliv-eredpower (ASCE, 1992) (Table 3.21). Table 3.21. Summary of energy calculations for full lift hypolimnetic aerator experiments. Treatment Ei E 2 E 3 E 4 Full lift - air Adiabatic compression @ 4.2 kg cm"2 Adiabatic compression @ 1.5 or 2.9 m depth Full lift - PSA Adiabatic compression @ 7.36 kg cm"2 Adiabatic compression @ 1.5 or 2.9 m depth Average PSA measured wire power Full lift, pump only - air Adiabatic compression @ 4.2 kg cm"2 Adiabatic compression = 0 Average meas-ured wire power of the two 4,500 1 hr"1 submersible mixing pumps Full lift, pump only - P S A Adiabatic compression @ 7.36 kg cm"2 Adiabatic compression = 0 Average PSA measured wire power Average meas-ured wire power of the two 4,500 1 hr"1 submersible mixing pumps Full lift - extended aeration - air Adiabatic compression @ 4.2 kg cm"2 70% of adiabatic compression @ 2.9 m depth Full lift - extended aeration - PSA Adiabatic compression @ 7.36 kg cm"2 70% of adiabatic compression @ 2.9 m depth Average PSA measured wire power The delivered blower power was calculated according to the adiabatic compression formula (ASCE, 1992): (3.10) P w = wRT,/29.7ne[(p 2/pi) 0 2 8 3-l] where: P w = power input (kW); w = weight of air flow ( kg s"1) (i.e., 1.2927 g l"1); R = engineering gas constant for air = 8.314 kJ/k mol °K; Ti = absolute inlet temperature before compression (°K); pi = absolute inlet pressure before compression = 1.03 kg cm"2; P2 = absolute inlet pressure after compression = 4.2 kg cm"2; n = (k-l)/k = 0.283 for air (and oxygen); k = 1.395 for air (and oxygen); 29.7 = constant for SI conversion; e = compressor efficiency = 0.80. For the full lift hypolimnetic aeration experiments, Groups 1-14, on PSA generated oxygen, the power input was separated into three components plus an expansion factor for the weight of air flow (i.e., w; kg s"1) based on the published air input/oxygen output ratio of the AirSep AS-20 unit (i.e., 15.9:1). The first component, E i , is the delivered blower power required to deliver the mass flow of gas at the absolute minimum pressure requirement of the PSA unit (7.36 kg cm"2). The second component, E 2 , is the same adiabatic compression formula adjusted to the absolute ambient hydrostatic pressure of gas release (i.e., 1.5 or 2.9 m), rather than the absolute minimum pressure requirement of the PSA unit. The third component, E3, is the average measured wire power of the AirSep AS-20 PSA unit (i.e., 0.0165 kW). The sum of E i , E2 and E3 equals the total delivered power (Table 3.21). For the full lift hypolimnetic aeration experiments, Pump Only, on air, the power input was separated into three components. The first component, E i , is the delivered blower power based upon adiabatic compression, from the manifold. The second component, E2, is the same adiabatic compression formula adjusted to the absolute ambient hydrostatic pressure of gas release at the surface (i.e., 11 cm) via the 2 u diffuser, hence this energy component is reduced to zero. The third component is the average measured wire power of the two 4,500 1 hr"1 submersible mixing pumps (i.e., 0.76 kW). The sum of E i , E2 and E3 equals the total delivered power (Table 3.21). For the full lift hypolimnetic aeration experiments, Pump Only, on PSA generated oxygen, the power input was separated into four components plus an expansion factor for the weight of air flow (w; kg s"1) based on the published air input/oxygen output ratio of the AirSep AS-20 PSA unit (i.e., 15.9:1). The first component, E i , is the delivered blower power required to deliver the mass flow of gas at the absolute minimum pressure requirement of the PSA unit (7.36 kg cm"2). The second component, E2, is the same adiabatic compression formula adjusted to the absolute ambient hydrostatic pressure of gas release at the surface (i.e., 11 cm) via the 2 \i diffuser, hence this energy component is reduced to zero. The third component is the averaged measured wire power of the two 4,500 1 hr"1 submersible mixing pumps (i.e., 0.76 kW). The fourth component, E4, is the average measured wire power of the AirSep PSA unit (i.e., 0.0165 kW). The sum of E i , E2, E3 and E4 equals the total delivered power (Table 3.21). For the full lift hypolimnetic aeration experiments, Extended Aeration, on air, the power input was separated into two components. The first component, E i , is the delivered blower power based upon adiabatic compression, from the manifold. The second component, E2, is the same adiabatic compression formula adjusted to different absolute ambient hydrostatic pressures according to the depth of gas release (i.e., 0.08 m and 2.9 m), and the compression energy proportioned to the two gas flow rates (i.e., 70% to 2.9 m and 30% to 0.08 m). Since the absolute ambient hydrostatic pressure of gas release at the surface (i.e., 11 cm) via the 2 u diffuser was so small, 30% of the E2 energy component was reduced to zero. The sum of E i and E2 equals the total delivered power (Table 21). For the full lift hypolimnetic aeration experiments, Extended Aeration, on PSA generated oxygen, the power input was separated into three components plus a correction factor for the weight of air flow (i.e., w; kg s"1) based on the published air input/oxygen output ratio of the AirSep AS-20 unit (i.e., 15.9:1). The first component, E i , is the delivered blower power re-quired to deliver the mass flow of gas at the absolute minimum pressure requirement of the compressor (7.36 kg cm"2). The second component, E2, is the same adiabatic compression formula adjusted to different absolute ambient hydrostatic pressures according to where the gas was released (i.e., 0.08 m and 2.9 m), and the compression energy proportioned between the two gas flow rates (i.e., 70% to 2.9 m and 30% to 0.08 m). Since the absolute ambient hydrostatic pressure of gas release at the surface (i.e., 11 cm) via the 2 u diffuser was so small, 30% of the E2 energy component was reduced to zero. The third component, E3, is the average measured wire power of the AirSep PSA unit (i.e., 0.0165 kW). The sum of E i , E 2 and E3 equals the total delivered power (Table 3.21). For the Speece Cone experiments, Groups 15-20, on PSA generated oxygen, the power input was separated into four components plus an expansion factor for the weight of air flow (i.e., w; kg s"1) based on the published air input/oxygen output ratio of the AirSep AS-20 PSA unit (i.e., 15.9:1). The first component, E i , is the delivered blower power required to deliver the mass flow of gas at the absolute minimum pressure requirement of the PSA unit (7.36 kg cm" ). The second component, E2, is the same adiabatic compression formula adjusted to the absolute ambient hydrostatic pressure of gas release (i.e., 2.9 m) rather than the absolute minimum pressure requirement of the PSA unit. The third component, E3, is the average measured wire power of the main Speece Cone pump, as derived from the linear regression equation between control unit setting and measured energy consumption (i.e., Eqn. 3.5; see Figure 3.7). The fourth component, E4, is the average measured wire power of the AirSep As-20 PSA unit (i.e., 0.0165 kW). The sum of E i , E2, E3 and E4 equals the total delivered power (Table 3.22). Table 3.22. Summary of energy calculations for Speece Cone experiments Treatment Ei E 2 E 3 E 4 Speece Cone -P S A Adiabatic compression @ 7.36 kg cm"2 Adiabatic compression @ 2.9 m depth Average meas-ured wire power of main circula-tion pump Average PSA measured wire power Speece Cone -air Adiabatic compression @ 4.2 kg cm"2 Adiabatic compression @ 2.9 m depth Average meas-ured wire power of main circula-tion pump For the Speece Cone experiments, Group 21, on air, the power input was separated into three components. The first component, E i , is the delivered blower power based upon adiabatic compression, from the manifold board regulator (i.e., 4.2 kg cm"2). The second component, E2, is the same adiabatic compression formula adjusted to the absolute ambient hydrostatic pressure of gas release (i.e., 2.9 m). The third component, E3, is the average measured wire power of the main Speece Cone pump derived from the linear regression equation between control unit setting and measured energy consumption (i.e., Eqn. 3.5, see Figure 3.7). The sum of E i , E2 and E3 equals the total delivered power (Table 3.22). SOTE was calculated as: (3.11) SOTE = S O T R A V 0 2 where: SOTE = standard oxygen transfer efficiency (%); SOTR = standard oxygen transfer rate (g O2 hr"1); W 0 2 = mass flow rate of oxygen in the gas flow stream (g O2 hr"1). The calculation of W 0 2 for the various experimental treatments was as follows: For the Full lift and Speece Cone treatments on air, W 0 2 was calculated as the measured gas flow (1 min"1) (corrected for pressure and density) multiplied by the weight of oxygen in air (0.3 g l"1) and converted to g O2 hr"1 (see Eqn. 3.12): (3.12) W 0 2air (g 0 2 hr"1) = gas flow (1 min"1) x 0.3 g l " 1 x 60 min hr"1 For the Full lift and Speece Cone treatments on PSA generated oxygen, W 0 2 was calculated as the sum of (1) the measured gas flow (1 min"1) (corrected for pressure and density) multiplied by the % of oxygen in the PSA gas (approx. 95 %) as determined by the average reading from Probe 7 for each individual test times the weight of pure oxygen (1.4286 g l"1) and converted to g 02hr"1 and (2) the measured gas flow (1 min"1) (corrected for pressure and density) multiplied by the % of air in the PSA gas (approx. 5%) times weight of oxygen in air (0.3 g l"1) and con-verted to g Chhr"1 (see Eqn. 3.13): (3.13) W 0 2 p s A ( g 0 2 hr"1) = [gas flow (1 min ' ) x % 0 2 x 1.4286 g l ' 1 x 60 min hr'1] + [gas flow (1 min"1) x % air x 0.3 g l " 1 x 60 min hr"1] 3.9.3 Statistical Analysis The statistical model used to analyze the experimental data was the General Linear Model (i.e., G L M ) in the SYSTAT™ 10 statistical package. This model can estimate any univariate or multivariate general linear model, including analysis of variance or covariance (Wilkinson and Coward, 2000). The level of significance was set at a = 0.01 for each statistical test. The first step in the analysis was to plot a frequency distribution of K i ^ o , SOTR, S A E and SOTE to determine if the values were normally distributed. A n initial box and whisker plot of the covariate and treatments (e.g., gas flow rate, diffuser depth, gas composition and orifice diame-ter) was then plotted against K La2o, SOTR, S A E and TE to examine for outliers. A total of four extreme outliers were detected in the 336 trials conducted in the Group 1-14 tests and no outliers were detected in the 72 trials conducted in the Group 15-20 tests. These outliers were a result of equipment malfunction during the tests and were deleted from the data set, resulting in a rejec-tion rate of 0.98%. The approach used in the statistical analysis was to analyze individual groups of experiments which were designed to test for a particular treatment effect (e.g., effect of depth of gas release; effect of oxygen partial pressure; effect of orifice diameter) as a function of gas flow rate. This resulted in a basic analysis of covariance with first order interactions (e.g., depth by gas flow rate). A n alternate approach would have been to develop one large model to analyze all of the data; however, this would have resulted in a 4 (gas flow rates) x 3 (orifice sizes) x 2 (depths) x 2 (% oxygen) matrix design, which would contain first, second and third order interaction effects. Since the experiments were designed as series of discrete tests to be of relevance to the practic-ing engineer, the basic model approach was used. A n initial analysis of covariance model was then used to determine i f there was any significant interaction between the covariate (i.e., gas flow rate for the full lift system, and velocity of water discharge for the Speece Cone) and the experimental treatments. The assumption of parallelism, or homogeneity of slopes, is an important underlying assumption of the analysis of covariance model. The analysis generally indicated a significant interaction effect for K i ^ o and SOTR, and occasionally for S A E and SOTE. The residuals were then plotted to determine the hetero-geneity of the error variance. The residual plots indicated a linear model was appropriate for the data; however, in the majority of cases, the plotted residuals assumed a "right megaphone" shape of distribution. This type of error distribution occurs where an intrinsically positive response varies over a wide range, as the smaller responses have less "room" to vary than larger responses (Weisberg, 1980). As a result of the observed heterogeneity of variance, a variance stabilization transformation was applied to Kjj^o, SOTR, S A E and SOTE. The logarithm is the most commonly used variance stabilizing transformation, and is appropriate when the error standard deviation is a percentage or proportion of the response (Weisberg, 1980). The analysis of covariance model was then re-run with the log-transformed data to examine for homogeneity of slopes in the form: (3.14) model y = constant + treatment + x + treatment*x If the probability (i.e., p) value of the treatment by covariate interaction was >0.01, the assump-tion of homogeneity of slopes was accepted, and the model was re-run without the interaction term in the form: (3.15) model y = constant + treatment + x A n additional benefit of log transforming the data is that it allows interaction effects, which are multiplicative when observed in the untransformed data, to be converted into an additive effect (which is a prerequisite for quantification with the general linear model). This is obtained by calculating the difference between the adjusted least square mean estimates for the treatment effect, and converting the exponent value back to the untransformed value. In the situations where only two treatment effects were being analysed (i.e., gas composition and depth of air release), the statistical analysis was complete at this stage. In situations where there were no significant interaction effects, the null hypothesis was rejected and there was more than one treatment effect being analyzed (i.e., orifice size or the effect of full lift modifications), then Scheffe's test was used as the comparison among means test. Scheffe's test is designed to allow all possible linear combinations of group means to be tested. The result is that Scheffe's test is more conservative than other tests, and a larger difference is required to obtain a significant result (Wilkinson and Coward, 2000). Scheffe's test is the most rigorous a posteriori test, and is recommended by statistical purists for comparison among means tests (Larkin, 1975). In situations where significant interaction effects remained after log transformation, the analysis was re-run by removing the particular treatment which exhibited the interaction, and the remain-ing data analyzed with Scheffe's test. The treatment responsible for the significant interaction effect, and the nature of the interaction was then described and no further statistical analysis conducted. The protocol adopted for presenting the results is to use the untransformed data to describe the observed value of the adjusted least squares mean for a particular treatment effect. However, when a significant treatment or interaction effect was detected, the transformed data was used to determine the level of significance in the covariance analysis, in Scheffe's comparison among means test, and to quantify the nature of the interaction effect, i f appropriate. CHAPTER 4: RESULTS 4.1 Full Lift Hypolimnetic Aeration - Oxygen Transfer 4.1.1 Depth Effect This component of the experiment examined the effect of gas flow variation (10 to 40 1 min"1) at two diffuser depths (1.5 m and 2.9 m) in the full lift inlet tube, using air and PSA oxygen at three different orifice diameters (140 u., 400 u. and 800 u). Within each subset of experiments, the orifice diameter and diffuser depth were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. 4.1.1.1 140 n diffuser x air (Note: compares the 140 fi diffuser on air at 1.5 and 2.9 m. The results for the analysis of covariance for the 140 jx diffuser operating on compressed air are shown in Table 4.1. Significant results (a < 0.01) were obtained for the depth treatment and flow rate covariate effect for K i ^ o , SOTR, S A E and SOTE (Figure 4.1 a-d) with all parameters increasing in value with depth of gas release. A significant depth x flow rate interaction effect was observed for K L ^ O and SOTR (Table 4.1). Table 4.1. Adjusted least squares means (+ SE), treatment and interaction effects for the 140 u. diffuser on air at 1.5 m or 2.9 m depth. Treatment K L a 2 0 (hr 1) SOTR (g 0 2 hr 1 ) SAE (g 0 2 kWhr *) SOTE(%) n 1.5 m 2.1 (0.09) 23.2 (0.95) 227.2 (7.39) 5.2 (0.18) 12 2.9 m 3.3 (0.09) 35.7 (0.95) 323.8 (7.39) 7.9 (0.18) 12 Interaction Yes, 1.52 Yes, 1.52 No No Depth sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The significant depth x flow rate interaction effect indicates that K i ^ o a n d SOTR increased 1.5 fold at the 2.9 m diffuser depth, with increasing gas flow rate, as compared to the 1.5 m diffuser Air Flow (L min-1 ) Air Flow (L min-1 ) Figure 4.1a-d KLa20 (a), SOTR (b), SAE (c) and SOTE (d) for 140 \i diffuser on air at 1.5 and 2.9 m depth. depth (Figure 4.1 a-b). S A E and SOTE were not influenced by the depth x the flow rate interac-tion effect, and declined with increasing gas flow rate. The mean S A E and SOTE at 2.9 m was 323.8 g 0 2 kWhr"1 and 7.9%, a 43 % and 52 % increase over the mean S A E of 227.2 g 0 2 kWhr" 'and SOTE of 5.2 % recorded at 1.5 m (Figure 4.1 c-d and Table 4.1). 4.1.1.2 400 fi diffuser x air (Note: compares the 400/u diffuser on air at 1.5 and2.9 m). The results for the analysis of covariance for the 400 u, diffuser operating on compressed air are shown in Table 4.2. Significant results (a < 0.01) were obtained for the depth treatment and flow rate covariate effect for K i ^ o , SOTR, S A E and SOTE (Figure 4.2 a-d), with all parameters increasing in value with depth of gas release. A significant depth x flow rate interaction effect was observed for K L a2oand SOTR (Table 3.2). Table 4.2. Adjusted least squares means (± SE), treatment and interaction effects for the 400 LI diffuser on air at 1.5 m or 2.9 m depth. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr1) SAE (g 0 2 kWhr'1) SOTE (%) n 1.5 m 1.7 (0.04) 18.1 (0.44) 179.6 (5.9) 4.1 (0.14) 12 2.9 m 2.3 (0.04) 25.5 (0.44) 237.8 (5.9) 5.8(0.14) 12 Interaction Yes, 1.40 Yes, 1.40 No No Depth sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sis, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The significant depth x flow rate interaction effect indicates that K i ^ o and SOTR increased 1.4 fold at the 2.9 m diffuser depth with increasing gas flow rate, as compared to the 1.5 m diffuser depth (Figure 4.2 a-b). S A E and SOTE were not influenced by the depth x the flow rate interac-tion effect, and declined with increasing gas flow rate. The mean S A E and SOTE at 2.9 m was 237.8 g 0 2 kWhr"1 and 5.8%, a 32 % and 41 % increase over the mean S A E of 179.6 g 0 2 kWhr" 'and SOTE of 4.1 % recorded at 1.5 m (Figure 4.2 c-d and Table 4.2). 4.1.1.3 800 JJ. diffuser x air (Note: compares the 800 fx diffuser on air at 1.5 and 2.9 m). The results for the analysis of covariance for the 800 p diffuser operating on compressed air are shown in Table 4.3. Significant results (a < 0.01) were obtained for the depth treatment and flow rate covariate effect for K.La 2o, SOTR, S A E and TE (Figure 4.3 a-d) with all parameters increasing in value with depth of gas release. A significant depth x flow rate interaction effect was observed for KLa2o and SOTR (Table 4.3). & 2-CD LU < CO 400 =£ 300 CM O 200 100 _1_ 10 20 30 Air Flow (L min-1) 40 50 10 20 30 40 Air Flow (L min-1) 50 40 p 30-CM O or o CO 20 O co 10 0 10 20 30 (L min-1) 40 50 -I • • i i DEPTH " • 1.5 m • 2.9 m - • • -- 8 • • • - • 1 i 10 20 30 40 Air Flow (L min-1) 50 Figure 4.2a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 400 p diffuser on air at 1.5 and 2.9 m depth. Table 4.3. Adjusted least squares means (± SE), treatment and interaction effects for the 800 \x diffuser on air at 1.5 m or 2.9 m depth. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr1) SAE (g 0 2 kWhr"') SOTE(%) n 1.5 m 1.3 (0.02) 14.5 (0.21) 139.9 (2.86) 3.2 (0.07) 12 2.9 m 2.0 (0.02) 22.1 (0.21) 197.1 (2.86) 4.8 (0.07) 12 Interaction Yes, 1.50 Yes, 1.50 No No Depth sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The significant depth x flow rate interaction effect indicates that K.T,a2o and SOTR increased 1.5 fold at the 2.9 m diffuser depth with increasing gas flow rate, as compared to the 1.5 m diffuser depth (Figure 4.3 a-b). S A E and SOTE were not influenced by the depth x the flow rate interac-tion effect, and declined with increasing gas flow rate. The mean S A E and SOTE at 2.9 m was 197.1 g 0 2 kWhr' 1 and 4.8%, a 41 % and 50 % increase over the mean S A E of 139.9 g 0 2 kWhr" 'and SOTE of 3.2 % recorded at 1.5 m (Figure 4.3 c-d and Table 4.3). 4.1.1.4 140 jx diffuser x PSA oxygen (Note: compares the 140 LI diffuser on PSA O2 at 1.5 and 2.9 m). The results for the analysis of covariance for the 140 u. diffuser operating on PSA oxygen are shown in Table 4.4. Significant results (a < 0.01) were obtained for the depth treatment and flow rate covariate effect for K La 2o, SOTR, S A E and SOTE (Figure 4.4 a-d) with all parameters increasing in value with depth of gas release. Table 4.4. Adjusted least squares means (± SE), treatment and interaction effects for the 140 u. diffuser on PSA oxygen at 1.5 m or 2.9 m depth. Treatment K L a 2 0 SOTR SAE SOTE n (hr1) (g 0 2 hr1) (g 0 2 kWhr1) (%) 1.5 m 13.6 (0.62) 147.6 (6.71) 63.1 (2.03) 8.0 (0.25) 12 2.9 m 17.5 (0.64) 190.9(7.01) 78.8(2.12) 10.5 (0.26) 11 Interaction No No No No Depth sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 K L a 2 o , S O T R , S A E and T E were not influenced by a depth x flow rate interaction effect. K T ^ O and SOTR increased with increasing gas flow rate. The mean Ki,a2o and SOTR at 2.9 m was 17.5 hr'1 and 190.9 g 0 2 hr'1, a 29% increase over the mean K L a 2 0 of 13.6 hr'1 and SOTR of 147.6 0 2 hr'1 recorded at 1.5 m. S A E and SOTE declined with increasing gas flow rate (Figure 4.4 c-d). The mean S A E and SOTE at 2.9 m was 78.8 g 0 2 kWhr"1 and 10.5%, a 25 % and 32 % increase over the mean S A E of 63.1 g 0 2 kWhr'and SOTE of 8.0 % recorded at 1.5 m (Table 4.4). Figure 4.3a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 800 n diffuser on air at 1.5 and 2.9 m depth. _ 20 SI CO I 10 20 30 40 Oxygen Flow (L min-1) 50 10 20 30 40 Oxygen Flow (L min-1) 300 h 200 CM o s ctt £ i o o h co o CO 10 20 30 40 Oxygen Flow (L min-1) 50 10 20 30 40 Oxygen Flow (L min-1) Figure 4.4a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 140 p diffuser on PSA 0 2 at 1.5 and 2.9 m depth. 4.1.1.5 400 n diffuser x PSA oxygen (Note: compares the 400 JX diffuser on PSA O2 at 1.5 and 2.9 m). The results for the analysis of covariance for the 400 u. diffuser operating on PSA oxygen are shown in Table 4.5. Significant results (a < 0.01) were obtained for the depth treatment and flow rate covariate effect for K L ^ O , SOTR, S A E and SOTE (Figure 4.5 a-d), with all parameters increasing in value with depth of gas release. A significant depth x flow rate interaction effect was observed for K i ^ o and SOTR (Table 4.5). Table 4.5. Adjusted least squares means (+ SE), treatment and interaction effects for the 400 (j. diffuser on PSA oxygen at 1.5 m or 2.9 m depth. Treatment KL320 SOTR SAE SOTE n (hr1) (g 0 2 hr1) (g 0 2 kWhr"1) (%) 1.5 m 10.1 (0.32) 110.2 (3.43) 49.2(1.31) 5.9(0.18) 12 2.9 m 13.9 (0.32) 151.0 (3.43) 62.8(1.31) 8.0(0.18) 12 Interaction Yes, 1.34 Yes, 1.34 No No Depth sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The significant depth x flow rate interaction effect indicates that Ki,a2o and SOTR increased 1.3 fold at the 2.9 m diffuser depth with increasing gas flow rate, as compared to the 1.5 m diffuser depth (Figure 4.5 a-b). S A E and SOTE were not influenced by the depth x the flow rate interac-tion effect, and declined with increasing gas flow rate (Figure 4.5 c-d). The mean S A E and SOTE at 2.9 m was 62.8 g 0 2 kWhr"1 and 8.0%, a 28 % and 35 % increase over the mean S A E of 49.1 g 0 2 kWhr'and SOTE of 5.9 % recorded at 1.5 m (Table 4.5). 4.1.1.6 800 n diffuser x PSA oxygen (Note: compares the 800 LI diffuser on PSA O2 at 1.5 and 2.9 m). The results for the analysis of covariance for the 800 u, diffuser operating on PSA oxygen are shown in Table 4.6. Significant results (a < 0.01) were obtained for the depth treatment and flow rate covariate effect for KLa2o; SOTR, S A E and SOTE (Figure 4.6 a-d), with all parameters increasing in value with depth of gas release. A significant depth x flow rate interaction effect was observed for K.La2o and SOTR (Table 4.6). The significant depth x flow rate interaction effect indicates that K L a 2 o and SOTR increased 1.5 fold at the 2.9 m diffuser depth with increasing gas flow rate as compared to the 1.5 m diffuser depth (Figure 4.6 a-b). S A E and SOTE were not influenced by the depth x the flow rate interac-tion effect, and declined with increasing gas flow rate (Figure 4.6 c-d). The mean S A E and SOTE at 2.9 m was 50.4 g 0 2 kWhr"1 and 6.4%, a 40 % and 45 % increase over the mean S A E of 35.9 g 0 2 kWhf'and SOTE of 4.4 % recorded at 1.5 m (Table 4.6). A summary of the adjusted least squares means depth treatment effects showing the percentage increase/decrease is shown in Table 4.7. -C co i 300 10 20 30 40 Oxygen Flow (L min-1) 50 200 CN O S ce H o CO 100 DEPTH • 1.5 m • 2.9 m 10 20 30 40 Oxygen Flow (L min-1) 50 90 _ 80 I 70 CM O 3 LU < CO 60 50 40 30 I • I I I — DEPTH -• • 1.5 m _ • 2.9 m • • - • •v. • • a I I I i 10 20 30 40 Oxygen Flow (L min-1) 50 m i-o CO 10 20 30 40 Oxygen Flow (L min-1) 50 Figure 4.5a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 400 p diffuser on PSA 0 2 at 1.5 and 2.9 m depth. Table 4.6. Adjusted least squares means (+ SE), treatment and interaction effects for the 800 u. diffuser on PSA oxygen at 1.5 m or 2.9 m depth. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr'1) SAE (g 0 2 kWhr"1) SOTE(%) n 1.5 m 7.6 (0.19) 83.1 (2.05) 35.9 (0.79) 4.8 (0.09) 12 2.9 m 11.5 (0.19) 124.8 (2.05) 50.4 (0.79) 6.4 (0.09) 12 Interaction Yes, 1.47 Yes, 1.47 No No Depth sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 20 15 & 10 DEPTH • 1.5m • 2.9 m 10 20 30 40 Oxygen Flow (L min-1) 50 200 P 150 CM O Q: H o CO 100 50 -DEPTH o 1.5 m • 2.9 m 10 20 30 40 Oxygen Flow (L min-1) 50 60 V 50 g 40r-LU < CO 30 20 DEPTH 10 20 30 40 Oxygen Flow (L min-1) 50 LU r -o to I I I DEPTH • • 1.5 m -• • • 2.9 m • e • • -• • — » I I o I 10 20 30 40 Oxygen Flow (L min-1) 50 Figure 4.6a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 800 n diffuser on PSA 0 2 at 1.5 and 2.9 m depth. Table 4.7. Summary of the adjusted least squares means depth treatment effects showing percentage increase/decrease. Treatment K L a 2 0 SOTR SAE SOTE (hr1) (g 0 2 hr') (g 0 2 kWhr"1) (%) 1.5 m x Air x 140 n 2.1 23.2 227.2 5.2 2.9 m x Air x 140 p 3.3 (+57%) 35.9 (+55%) 323.4 (+42%) 7.8 (+50%) 1.5 m x Air x 400 u 1.7 18.1 179.6 4.1 2.9 m x Air x 400 u. 2.4 (+41%) 25.5 (+41%) 237.8 (+32%) 5.8 (+41%) 1.5 m x Air x 800 u 1.3 14.5 139.9 3.2 2.9 m x Air x 800 u 2.0 (54%) 22.1 (+52%) 197.2 (+41%) 4.8 (+50%) 1.5 m x 0 2 x 140 u 13.5 146.6 63.3 8.0 2.9 m x 0 2 x 140 u 17.5 (+30%) 190.9 (+30%) 79.0 (+25%) 10.5 (+31%) 1.5 m x 0 2 x 400 u. 10.1 110.2 49.2 5.9 2.9 m x 0 2 x 400 u 13.9 (+38%) 151.0 (+37%) 62.8 (+28%) 8.0 (+36%) 1.5 m x 0 2 x 800 u 7.6 83.1 35.9 4.4 2.9 m x 0 2 x 800 u, 11.5 (+51%) 124.8 (+50%) 50.4 (+40%) 6.4 (+45%) 4.1.2 Oxygen Partial Pressure Effect This component of the experiment examined the effect of gas flow variation (10 to 40 1 min"') using air and PSA oxygen, at two diffuser depths (1.5 m and 2.9 m) in the full lift inlet tube at three different orifice diameters (140 u, 400 \i and 800 u). Within each subset of experiments, the orifice diameter and diffuser depth were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. 4.1.2.1 140 JU diffuser x 1.5 m depth (Note: compares the 140 /x diffuser on air and PSA O2 at 1.5 m). The results for the analysis of covariance for the 140 u. diffuser at 1.5 m depth are shown in Table 4.8. Significant results (a < 0.01) were obtained for the gas treatment and flow rate covariate effect for KLa2o, SOTR, S A E and TE (Figure 4.7 a-d), with KLa2o, SOTR and SOTE increasing in value and S A E decreasing with PSA oxygen. A significant gas x flow rate interaction effect was observed for K L a 2 0 , SOTR and S A E (Table 4.8). Figure 4.7a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 140 p diffuser at 1.5 m on air or PSA 0 2 . Table 4.8. Adjusted least squares means (± SE), treatment and interaction effects for the 140 ja diffuser at 1.5 m depth on air or PSA oxygen. Treatment K L a 2 0 SOTR SAE SOTE n (hr1) (g0 2 hr1) (g 0 2 kWhr_1) (%) Air 2.1 (0.33) 23.2 (3.64) 227.2 (2.64) 5.9(0.18) 12 PSA oxygen 13.5 (0.33) 146.6 (3.64) 63.3 (2.64) 8.0(0.18) 12 Interaction Yes, 6.30 Yes, 6.30 Yes, 3.61 No Gas sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The significant gas x flow rate interaction effect indicates that K t ^ O and SOTR increased 6.3 fold on PSA oxygen, with increasing gas flow rate as compared to air, whereas S A E decreased 3.6 fold on PSA oxygen, with increasing gas flow rate as compared to air (Figure 4.7 c). SOTE was not influenced by the gas x flow rate interaction effect, and declined with increasing gas flow rate (Figure 4.7 d). The mean SOTE for PSA oxygen was ~ 8%, a 54% increase over the mean SOTE of 5.9% for air (Table 4.8). 4.1.2.2 400 LI diffuser x 1.5 m depth (Note: compares the 400 LI diffuser on air and PSA O2 at 1.5 m). The results for the analysis of covariance for the 400 u. diffuser at 1.5 m depth are shown in Table 4.9. Significant results (a < 0.01) were obtained for the gas treatment and flow rate covariate effect for K L a 2 0 , SOTR, S A E and SOTE (Figure 4.8 a-d), with K La 2o, SOTR and SOTE increasing in value and S A E decreasing with PSA oxygen. A significant gas x flow rate interaction effect was observed for K.LA20, SOTR, S A E and SOTE (Table 4.9). Table 4.9. Adjusted least squares means (± SE), treatment and interaction effects for the 400 u. diffuser at 1.5 m depth on air or PSA oxygen. Treatment K L a 2 0 SOTR SAE SOTE n (hr1) (g 0 2 hr1) (g 0 2 kWhr1) (%) Air 1.7 (0.18) 18.1 (1.91) 179.6(1.9) 4.1 (0.10) 12 PSA oxygen 10.1 (0.18) 110.2(1.9) 49.2(1.9) 5.9 (0.10) 12 Interaction Yes, 6.15 Yes, 6.15 Yes, 3.68 Yes, 1.44 Gas sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 _ 10 to 200 p 150-o ce H O 100 -20 30 Gas Flow (L min-1) 50 10 20 30 40 Gas Flow (L min-1) 50 10 20 30 40 Gas Flow (L min-1) 10 20 30 40 Gas Flow (L min-1) 50 Figure 4.8a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 400 ]i diffuser at 1.5 m on air or PSA 0 2 . The significant gas x flow rate interaction effect indicates that K La 2oand SOTR increased 6.2 fold on PSA oxygen with increasing gas flow rate, as compared to air (Figure 4.8 a-b). S A E decreased 3.7 fold on PSA oxygen with increasing gas flow rate, as compared to air, while SOTE decreased 1.4 fold on PSA oxygen with increasing gas flow rate as compared to air (Figure 4.8 c-d). 4.1.2.3 800 fx diffuser x 1.5 m depth (Note: compares the 800 fi diffuser on air and PS A O2 at 1.5 m). The results for the analysis of covariance for the 800 u. diffuser at 1.5 m depth are shown in Table 4.10. Significant results (a < 0.01) were obtained for the gas treatment and flow rate covariate effect for K La 2o, SOTR, S A E and TE (Figure 4.9 a-d), with K L a 2 0 , SOTR and SOTE increasing in value and S A E decreasing with PSA oxygen. A significant gas x flow rate interac-tion effect was observed for K.L.a2o, SOTR and S A E (Table 4.10). The significant gas x flow rate interaction effect indicates that K L a 2 o and SOTR increased 5.8 fold on PSA oxygen with increasing gas flow rate, as compared to air, whereas S A E decreased 3.9 fold on PSA oxygen with increasing gas flow rate, as compared to air (Figure 4.9 a-c). SOTE was not influenced by the gas x flow rate interaction effect, and declined with increasing gas flow rate (Figure 4.9 d). The mean SOTE for PSA oxygen was ~ 4.4%, a 38% increase over the mean TE of 3.2% for air (Table 4.10). Table 4.10. Adjusted least squares means (± SE), treatment and interaction effects for the 800 p diffuser at 1.5 m depth on air or PSA oxygen. Treatment K L a 2 0 (hr1) SOTR SAE SOTE n (g 0 2 hr ') (g 0 2 kWhr-1) (%) Air 1.3 (0.13) 14.5(1.42) 139.9(1.22) 3.19(0.08) 12 PSA oxygen 7.6(0.13) 83.1 (1.42) 35.9 (1.22) 4.37 (0.08) 12 Interaction Yes, 5.79 Yes, 5.78 Yes, 3.92 No Gas sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 ~ 10 .c as 200 10 20 30 40 Gas Flow (L min-1) 50 CM O •Si or h-o co LU 1-o CO 10 20 30 40 Gas Flow (L min-1) 50 10 20 30 40 50 Gas Flow (L min-1) I • 1 1 1 • GAS • o Air • Oxygen • - -• • •T——^ 1 1 t i 10 20 30 40 Gas Flow (L min-1) 50 Figure 4.9a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 800 n diffuser at 1.5 m on air or PSA O2 . 4.1.2.4 140 u. diffuser x 2.9 m depth (Note: compares the 140 \x diffuser on air and PSA O2 at 2.9 m). The results for the analysis of covariance for the 140 p. diffuser at 2.9 m depth are shown in Table 4.11. Significant results (a < 0.01) were obtained for the gas treatment and flow rate covariate effect for K L a 2 0 , SOTR, S A E and T E (Figure 4.10 a-d), with K L a 2 0 , SOTR and TE increasing in value and S A E decreasing with PSA oxygen. A significant gas x flow rate interac-tion effect was observed for KLa2o and SOTR (Table 4.11). 10 20 30 40 Gas Flow (L min-1) 50 10 20 30 40 Gas Flow (L min-1) 50 LU I -o CO 10 20 30 40 Gas Flow (L min-1) 50 13 12 11 10 g 8 7 6 I I • I I • \ GAS " - • Air - • • Oxygen • • • • < • I I T • • i 10 20 30 Gas Flow (L min-1) 40 50 Figure 4.10a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 140 p diffuser at 2.9 m on air or PSA O2 . Table 4.11. Adjusted least squares means (+ SE), treatment and interaction effects for the 140 p, diffuser at 2.9 m depth on air or PSA oxygen. Treatment K L a 2 0 (hr1) SOTR SAE SOTE n (g 0 2 hr1) (g 0 2 kWhr"1) (%) Air 3.3 (0.52) 35.9 (5.66) 323.4 (7.89) 7.8 (0.26) 12 PSA oxygen 17.5(0.54) 190.9 (5.91) 79.0 (8.24) 10.5 (0.27) 11 Interaction Yes, 5.33 Yes, 5.33 No No Gas sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The significant gas x flow rate interaction effect indicates that K L a2oand SOTR increased 5.3 fold on PSA oxygen with increasing gas flow rate, as compared to air (Figure 4.10 a-b). S A E and SOTE were not influenced by the gas x flow rate interaction effect, and declined with increasing gas flow rate (Figure 4.10 c-d). The mean SOTE for PSA oxygen was 10.5%, a 34% increase over the mean SOTE of 7.8% for air (Table 4.11). 4.1.2.5 400 /u diffuser x 2.9 m depth (Note: compares the 400 LI diffuser on air and PSA O2 at 2.9 m). The results for the analysis of covariance for the 400 u. diffuser at 2.9 m depth are shown in Table 4.12. Significant results (a < 0.01) were obtained for the gas treatment and flow rate covariate effect for K L a 2 0 , SOTR, S A E and SOTE (Figure 4.11 a-d), with K La 2o, SOTR and SOTE increasing in value and S A E decreasing with PSA oxygen. A significant gas x flow rate interaction effect was observed for K.L,a2o, SOTR and S A E (Table 4.12). The significant gas x flow rate interaction effect indicates that K.La2o and SOTR increased 5.9 fold on PSA oxygen with increasing gas flow rate, as compared to air, whereas S A E decreased 3.8 fold on PSA oxygen with increasing gas flow rate, as compared to air (Figure 4.11 a-b). SOTE was not influenced by the gas x flow rate interaction effect, and declined with increasing gas flow rate (Figure 4.11 d). The mean TE for PSA oxygen was 8.0%, a 38% increase over the mean TE of 5.8% for air (Table 4.12). Table 4.12. Adjusted least squares means (+ SE), treatment and interaction effects for the 400 u, diffuser at 2.9 m depth on air or PSA oxygen. Treatment K L a 2 0 (hr1) SOTR SAE SOTE n (g0 2 hr1) (g 0 2 kWhr"1) (%) Air 2.4 (0.27) 25.5 (2.9) 237.9 (5.75) 5.8 (0.21) 12 PSA oxygen 13.9 (0.27) 151.0 (2.9) 62.8 (5.75) 8.0 (0.21) 12 Interaction Yes, 5.88 Yes, 5.88 Yes, 3.78 No Gas sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 0 10 20 30 40 50 0 10 20 30 40 50 Gas Flow (L min-1 ) Gas Flow (L min-1 ) Figure 4.11a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 400 p diffuser at 2.9 m on air or PSA 0 2. 4.1.2.6 800 n diffuser x 2.9 m depth (Note: compares the 800 /J, diffuser on air and PSA O2 at 2.9 m). The results for the analysis of covariance for the 800 u. diffuser at 2.9 m depth are shown in Table 4.13. Significant results (a < 0.01) were obtained for the gas treatment and flow rate covariate effect for K L a 2 0 , SOTR, S A E and SOTE (Figure 4.12 a-d), with K L a 2 0 , SOTR and SOTE increasing in value and S A E decreasing with PSA oxygen. A significant gas x flow rate interaction effect was observed for K La 2oand SOTR (Table 4.13). Table 4.13. Adjusted least squares means (± SE), treatment and interaction effects for the 800 u. d iffuser at 2.9 m depth on air or PSA oxygen. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr 1) SAE (g 0 2 kWhr"1) SOTE (%) n Air 2.0 (0.14) 22.1 (1.50) 197.1 (2.90) 4.8 (0.09) 12 PSA oxygen 11.5 (0.14) 124.8 (1.50) 50.4 (2.90) 6.4 (0.09) 12 Interaction Yes, 5.68 Yes, 5.68 No No Gas sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The significant gas x flow rate interaction effect indicates that KLa 2oand SOTR increased 5.7 fold on PSA oxygen with increasing gas flow rate, as compared to air (Figure 4.12 a-b). S A E and SOTE were not influenced by the gas x flow rate interaction effect, and declined with increasing gas flow rate (Figure 4.12 c-d). The mean SOTE for PSA oxygen was 6.4%, a 33% increase over the mean SOTE of 4.8% for air (Table 4.13). A summary of the adjusted least squares means oxygen partial pressure treatment effects showing the percentage in-crease/decrease are shown in Table 4.14. 200 10 20 30 40 Gas Flow (L min-1) CM O 3 a. \-o co 10 20 30 40 Gas Flow (L min-1) sz J* CN1 O S LU < CO 300 200 100 10 20 30 Gas Flow (L min-1) O CO 10 20 30 Gas Flow (L min-1) Figure 4.12a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 800 p diffuser at 2.9 m on air or PSA 0 2 . p Table 4.14. Summary of oxygen partial pressure treatment effects showing percentage increase/decrease. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr 1) SAE (g 0 2 kWhr'1) SOTE (%) Air x 140 u, x 1.5 m 2.1 23.2 227.2 5.2 0 2 x 140 \i x 1.5 m 13.5 (+542) 146.6 (+532) 63.3 (-72) 8.0 (+54) Air x 400 u x 1.5 m 1.7 18.1 179.6 4.1 O2x400 u,xl.5m 10.1 (+494) 110.2 (+509) 49.2 (-73) 5.9 (+44) Air x 800 u. x 1.5 m 1.3 14.5 139.9 3.2 O2x800 u,xl.5m 7.6 (+485) 83.1 (+473) 35.9 (-74) 4.4 (+38) Air x 140 u, x 2.9 m 3.3 35.9 323.4 7.8 0 2 x 140 n x 2.9 m 17.5 (+430) 190.9 (+432) 79.0 (-76) 10.5 (+35) Air x 400 n x 2.9 m 2.4 25.5 237.8 5.8 O 2x400px2.9m 13.9 (+479) 151.0 (+492) 62.8 (-74) 8.0 (+38) Air x 800 u x 2.9 m 2.0 22.1 197.2 4.8 O2x800 ux2.9m 11.5 (+475) 124.8 (+465) 50.4 (-)74 6.4 (+33) 4.1.3 Orif iceD iameter Effect This component of the experiment examined the effect of gas flow variation (10 to 40 1 min"1) at three different orifice diameters (140 p, 400 u. and 800 p), using air and PSA oxygen, at two diffuser depths (1.5 m and 2.9 m) in the full lift inlet tube. Within each subset of experiments, the orifice diameter and diffuser depth were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. 4.1.3.1 Airx 1.5 m depth (Note: compares the 140 //, 400 JJ. and 800 /J. diffusers on air at 1.5 m). The results for the analysis of covariance for the 140 u, 400 u, and 800 p diffusers at 1.5 m depth on air are shown in Table 4.15. Significant results (a < 0.01) were obtained for the orifice treatment and flow rate covariate effect for Ki,a2o, SOTR, S A E and SOTE (Figure 4.13 a-d), with K i ^ o , SOTR, S A E and SOTE decreasing in value with increasing orifice diameter. A significant orifice diameter x flow rate interaction effect was also observed for K L ^ O , SOTR, S A E and T E (Table 4.15). Table 4.15. Adjusted least squares means (± SE), treatment and interaction effects for the 140 p, 400 p and 800 p, diffusers at 1.5 m depth on air. Treatment K L a 2 0 (hr1) SOTR SAE SOTE n (g 0 2 hr"') (g 0 2 kWhr"1) (%) 140 p. 2.1 (0.03) 23.2 (0.37) 227.2 (2.55) 5.2 (0.06) 12 400 u 1.7 (0.03) 18.1 (0.37) 179.6(2.55) 4.1 (0.06) 12 800 u. 1.3 (0.03) 14.5 (0.37) 139.9 (2.55) 3.2 (0.06) 12 Interaction Yes, 1.27/1.62 Yes, 1.27/1.62 Yes, 1.62 Yes, 1.62 Orifice sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 CO —I 1 1 1 • 1 ORIF o 140 u _ • 400 u 1 1 1 • 800 u i 300 - 250 SI 5 g 200 S LU «S 150 100 10 20 30 40 Air Flow (L min-1) 50 I • I I I ORIF • 140 u _ • T 400 u • 800 u T • _ • — . • • • —— I I • — I I 10 20 30 40 Air Flow (L min-1) 50 40 P 30 i s: CM or H o w 10 I I I I • / • V^\^^ ORIF » • 140u _ T 400 u I • 800 u i i i 10 20 30 40 Air Flow (L min-1) 50 10 20 30 40 Air Flow (L min-1) 50 Figure 4.13a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 140, 400 and 800 p diffusers on air at 1.5 m. Scheffe's test indicates that orifice diameter values were significantly different from each other (a < 0 .01) for KLa2o, SOTR, S A E and SOTE. The log transformation linearized the orifice diameter x flow rate interaction effect for K i ^ o and SOTR, and indicates that, with increasing gas flow rate, K L ^ O and SOTR from the 140 u, diffuser increased 1.27 fold as compared to the 4 0 0 p diffuser, and 1.62 fold when compared to the 8 0 0 p diffuser. The log transformation did not linearize the orifice diameter x flow rate interaction effect for S A E and SOTE. The analysis was re-run without the 4 0 0 p data, and the results indicate that S A E and SOTE exhibited parallel slopes for the 140 u. and 800 u. diffuser. The plot of the log transformed data indicated that with increasing gas flow rate, the slope of the 400 \x orifice decreased more rapidly than the 140 u. and 800 \x orifices. A n analysis of the 140 and 800 u, treatment pair indicates S A E and SOTE from the 140 u. orifice decreased 1.62 fold, as compared to the 800 \x orifice with increasing gas flow rate. However, the interaction effect could not be separated from the 400 u. and 800 u. treatment pair, and no further statistical analysis was conducted. 4.1.3.2 Air x 2.9 m depth (Note: compares the 140 pt, 400 /x and 800 y. diffusers on air at 2.9 m). The results for the analysis of covariance for the 140 u., 400 u. and 800 p, diffusers at 2.9 m depth on air are shown in Table 4.16. Significant results (a < 0.01) were obtained for the orifice treatment and flow rate covariate effect for KLa2o, SOTR, S A E and SOTE (Figure 4.14 a-d), with KLa2o, SOTR, S A E and SOTE decreasing in value with increasing orifice diameter. A significant orifice diameter x flow rate interaction effect was observed for K.La2o and SOTR (Table 4.16). Table 4.16. Adjusted least squares means (± SE), treatment and interaction effects for the 140 u., 400 u, and 800 u, diffusers at 2.9 m depth on air. Treatment K L a 2 0 (hr1) SOTR SAE SOTE n (g 0 2 hr1) (g 0 2 kWhr1) (%) 140 u. 3.3 (0.07) 35.7 (0.79) 323.8 (7.66) 7.85 (0.19) 12 400 u 2.4 (0.07) 25.5 (0.79) 237.8 (7.66) 5.77 (0.19) 12 800 u 2.0 (0.07) 22.1 (0.79) 197.1 (7.66) 4.78 (0.19) 12 Interaction Yes, 1.37/1.63 Yes, 1.37/1.63 No No Orifice sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Scheffe's test indicates that orifice diameter was significantly different from each other (a < 0.01) for KLa2o, SOTR, S A E and SOTE. The log transformation linearized the orifice diameter x flow rate interaction effect, and indicates that with increasing gas flow rate, K i ^ o and SOTR from the 140 p diffuser increased 1.37 fold, as compared to the 400 p and 1.63 fold, when compared to the 800 p diffuser. S A E and SOTE were not influenced by the orifice diameter x flow rate interaction effect, and declined with increasing gas flow rate (Figure 4.14 c-d). The mean S A E and SOTE for the 140 p diffuser was 323.8 g O2 kWhr 1 and 7.9%, a 36 % increase over the mean S A E of 237.8 g O2 kWhr' 1 and SOTE of 5.8 % for the 400 p diffuser and a 64% increase over the mean S A E of 197.1 g 0 2 kWhr-1 and SOTE of 4.8 % for the 800 p diffuser (Table 4.16). 4.1.3.3 PSA oxygen x 1.5 m depth (Note: compares the 140 LI, 400 LI and 800 /4 diffusers on PSA 02 at 1.5 m). The results for the analysis of covariance for the 140 p, 400 p and 800 p diffusers at 1.5 m depth on PSA oxygen are shown in Table 4.17. Significant results (a < 0.01) were obtained for the orifice treatment and flow rate covariate effect for KLa2o, SOTR, S A E and SOTE (Figure 4.15 a-d), with K i ^ o , SOTR, S A E and TE decreasing in value with increasing orifice diameter. A significant orifice diameter x flow rate interaction effect was also observed for K r ^ o and SOTR (Table 4.17). 500 ^ 400-5 LU $ 200 100 I I I I ORIF • • 140 u _ • 400 u » . » • • 800 u - T • •— • _ t • I T - • I • I I I 10 20 30 40 Air Flow (L min-1) 50 O CO 10 20 30 Air Flow (L min-1) 50 Figure 4.14a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 140, 400 and 800 p diffusers on air at 2.9 m. Table 4.17. Adjusted least squares means (± SE), treatment and interaction effects for the 140 p, 400 p and 800 p diffusers at 1.5 m depth on PSA oxygen. Treatment K L a 2 0 (hr"1) SOTR SAE SOTE n (g0 2 hr 1) (g 0 2 kWhr-1) (%) 140 u 13.5 (0.32) 146.6 (3.53) 63.3 (1.27) 8.0 (0.16) 12 400 p 10.1 (0.32) 110.2 (3.53) 49.2(1.27) 5.9 (0.16) 12 800 u. 7.6 (0.32) 83.1 (3.53) 35.9 (1.27) 4.4 (0.16) 12 Interaction Yes, 1.30/1.76 Yes, 1.30/1.76 No No Orifice sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 250 200 S 150 O S •T i o o H O CO 50 10 20 30 40 Oxygen Flow (L min-1) 50 LU H O co 10 20 30 40 Oxygen Flow (L min-1 ) 50 ORIF • 140 u T 400 u • 800 u 10 20 30 40 Oxygen Flow (L min-1) 50 10 20 30 40 Oxygen Flow (L min-1) 50 Figure 4.15a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 140, 400 and 800 p diffusers on PSA 0 2 at 1.5 m. Scheffe's test indicates that orifice diameter was significantly different from each other (a < 0.01) for K.La2o, SOTR, S A E and SOTE. The log transformation linearized the orifice diameter x flow rate interaction effect, and indicates that with increasing gas flow rate, KLa 2o and SOTR from the 140 u, diffuser increased 1.30 fold, as compared to the 400 u, and 1.76 fold, when compared to the 800 u. diffuser. S A E and SOTE were not influenced by the orifice diameter x flow rate interaction effect, and declined with increasing gas flow rate. The mean S A E and SOTE for the 140 u. diffuser was 63.3 g 0 2 kWhr"1 and 8.0 %, a 29 % and 35 % increase over the mean S A E of 49.2 g 0 2 kWhr"1 and SOTE of 5.9 % for the 400 p diffuser, and a 76 % and 82 % increase over the mean S A E of 35.9 g 0 2 kWhr"1 and SOTE of 4.4 % for the 800 p diffuser (Figure 4.15 c-d and Table 4.17). 4.1.3.4 PSA oxygen x 2.9 m depth (Note: compares the 140 JU, 400 fx and 800 fx diffusers on PSA 02 at 2.9 m). The results for the analysis of covariance for the 140 p, 400 p and 800 p diffusers at 2.9 m depth on PSA oxygen are shown in Table 4.18. Significant results (a < 0.01) were obtained for the orifice treatment and flow rate covariate effect for K La 2o, SOTR, S A E and SOTE (Figure 4.16 a-d), with K La 2o, SOTR, S A E and SOTE decreasing in value with increasing orifice diameter. No significant orifice diameter x flow rate interaction effect was observed for K.La2o, SOTR, S A E and SOTE (Table 4.18). Table 4.18. Adjusted least squares means (± SE), treatment and interaction effects for the 140 u, 400 p. and 800 p, diffusers at 2.9 m depth on PSA oxygen. Treatment K L a 2 0 (hr1) SOTR SAE SOTE n (g 0 2 hr1) (g 0 2 kWhr •) (%) 140 u 17.5 (0.50) 190.5 (5.49) 78.8 (1.70) 10.5 (0.22) 11 400 u. 13.9 (0.48) 151.7 (5.26) 62.7(1.63) 8.0(0.21) 12 800 u 11.5 (0.48) 125.4 (5.26) 50.4(1.63) 6.4 (0.21) 12 Interaction No No No No Orifice sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow rate sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Scheffe's test indicates that orifice diameters were significantly different from each other (a < 0.01) for K L a 2 0 > SOTR, S A E and SOTE. The mean K L a 2 0 , SOTR, S A E and SOTE for the 140 u. diffuser was 26, 26, 26 and 32 % greater, respectively, than the 400 p. diffuser, and 52, 52, 57 and 66 % greater, respectively, than the 800 p diffuser. A summary of the adjusted least squares ce 10 20 30 40 Oxygen Flow (L min-1) 50 10 20 30 40 Oxygen Flow (L min-1) 50 10 20 30 40 Oxygen Flow (L min-1) 50 LU h-O CO 13 12 11 10 g 8 7 6 5 I I 0 I I ORIF • * 140 u _ T • 400 u - • 800 u -T • o - -• • I • I I • I 10 20 30 40 Oxygen Flow (L min-1) 50 Figure 4.16a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for 140, 400 and 800 p diffusers on PSA 02 at 2.9 m. Table 4.19. Summary of diffuser orifice diameter treatment effects showing percentage increase/decrease. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr 1) SAE (g 0 2 kWhr1) SOTE (%) 140 n x Air x 1.5 m 2.1 23.2 227.2 5.2 400 u, x Air x 1.5 m 1.7 (-19) 18.1 (-22) 179.6 (-26) 4.1 (-21) 800 u, x Air x 1.5 m 1.3 (-26) 14.5 (-20) 139.9 (-22) 3.2 (-22) 140 LI x Air x 2.9 m 3.3 35.9 323.4 7.8 400 u. x Air x 2.9 m 2.4 (-27) 25.5 (-29) 237.8 (-26) 5.8 (-29) 800 u, x Air x 2.9 m 2.0 (-17) 22.1 (-13) 197.2 (-17) 4.8 (-17) 140 u, x 0 2 x 1.5 m 13.5 146.6 63.3 8.0 400 u, x 0 2 x 1.5 m 10.1 (-25) 110.2 (-43) 49.2 (-22) 5.9 (-26) 800 LI x 0 2 x 1.5 m 7.6 (-25) 83.1 (-25) 35.9 (-27) 4.4 (-25) 140 p, x 0 2 x 2.9 m 17.5 190.9 79.0 10.5 400 u, x 0 2 x 2.9 m 13.9 (-21) 151.0 (-21) 62.8 (-21) 8.0 (-24) 800 u. x 0 2 x 2.9 m 11.5 (-17) 124.8 (-17) 50.4 (-20) 6.4 (-20) means diffuser orifice diameter treatment effects showing the percentage increase/decrease are shown in Table 4.19. 4.1.4 Full Lift Modifications Effect This component of the experiment examined the effect of gas flow variation (10 to 40 1 min - 1) on four in situ modifications (i.e., bubble screen, counter-rotating vanes, surface cover and Downflow Bubble Contact Aeration in the outlet tube), using air and PSA oxygen with two orifice diameters (140 p and 800 p) at one diffuser depth (2.9 m) in the full lift inlet tube. Within each subset of experiments, the orifice diameter, diffuser depth and in situ modification were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. 4.1.4.1 Air x 140 LI diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 140 LI diffuser on air at 2.9 m). The results for the analysis of covariance for the various in situ modifications of the full lift hypolimnetic aerator with the 140 p diffuser on air are shown in Table 4.20. A l l tests were conducted at 2.9 m. Significant results (a < 0.01) were obtained for the in situ treatment and flow rate covariate effect for K i ^ o , SOTR, S A E and SOTE (Figure 4.17 a-d). The no treatment option exhibited the largest mean K i ^ o , SOTR, S A E and SOTE values, followed in decreasing order by the vanes, D B C A , bubble screen and surface cover treatments. A significant in situ treatment x flow rate interaction effect was observed for KLa2o and SOTR (Table 4.20). Figure 4.17a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for in situ modifications on air with 140 p diffuserai 2.9 m. Table 4.20. Adjusted least squares means (+ SE), treatment and interaction effects for the various in situ full lift hypolimnetic aerator treatments on air with the 140 (i diffuser. Treatment K L a 2 0 (hr 1) SOTR ( g 0 2 hr 1 ) SAE (g 0 2 kWhr SOTE (%) n Bubble screen 2.1 (0.07) 22.5 (0.77) 209.6 (6.70) 5.1 (0.16) 11 D B C A 2.3 (0.07) 24.9 (0.75) 241.3 (6.43) 5.9 (0.16) 12 None 3.3 (0.07) 35.4 (0.73) 324.5 (6.41) 7.9(0.16) 12 Surface cover 1.9 (0.07) 20.6 (0.73) 192.7 (6.41) 4.7 (0.16) 12 Vanes 2.4 (0.07) 26.5 (0.73) 248.1 (6.41) 6.0 (0.16) 12 Interaction Yes Yes No No In situ effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Scheffe's test indicates that the no modifications treatment was significantly greater than the other in situ treatments (a < 0.01) for K i ^ o a n d SOTR (Tables 4.21 and 4.22). In addition, KLa2o and SOTR values for the vanes and D B C A treatments were significantly greater than the surface cover treatment. Table 4.21. Matrix of Ki^o pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on air with the 140 u, diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.324 1.000 None 0.000 0.000 1.000 Surface cover 0.590 0.007 0.000 1.000 Vanes 0.093 0.971 0.000 0.001 1.000 Table 4.22. Matrix of SOTR pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on air with the 140 u. diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.329 1.000 None 0.000 0.000 1.000 Surface cover 0.581 0.007 0.000 1.000 Vanes 0.099 0.974 0.000 0.001 1.000 For S A E and SOTE, Scheffe's test indicates the bubble screen treatment mean was significantly less (a < 0.01) than the D B C A , no modifications and vanes treatment (Table 4.23). In addition, the no treatment mean was significantly greater than the D B C A , surface cover and vanes, and D B C A was significantly greater than the surface cover treatment (Table 4.23). Finally, the vanes treatment was significantly greater than the surface cover treatment. The log transforma-tion linearized the in situ treatment x flow rate interaction effect. Table 4.23. Matrix of SAE and SOTE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on air with the 140 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.010 1.000 None 0.000 0.000 1.000 Surface cover 0.209 0.000 0.000 1.000 Vanes 0.000 0.873 0.000 0.000 1.000 Table 4.24 lists the multiplicative effect (positive or negative) of increasing gas flow rate on K L a 2 o and SOTR. Table 4.24. Matrix of multiplicative in situ treatment x flow rate interaction effects for Ki^o and SOTR for various in situ full lift hypolimnetic aerator treatments on air with the 140 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 0.00 DBCA 1.13 0.00 None 1.54 1.36 0.00 Surface cover 0.91 0.81 0.59 0.00 Vanes 1.18 1.04 0.76 1.29 0.00 S A E and SOTE were not influenced by the in situ treatment x flow rate interaction effect, and declined with increasing gas flow rate, particularly the D B C A treatment. 4.1.4.2 Air x 800 LI diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 800 LI diffuser on air at 2.9 m). The results for the analysis of covariance for the various in situ modifications of the full lift hypolimnetic aerator with the 800 p diffuser on air are shown in Table 4.25. Significant results (a < 0.01) were obtained for the in situ treatment and flow rate covariate effect for K T ^ O , SOTR, S A E and SOTE (Figure 4.18 a-d). The no treatment option exhibited the largest mean KLa2o Figure 4.18a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for in situ modifications on air with 800 p diffuser at 2.9 m. and SOTR values, followed in decreasing order by the D B C A , vanes, bubble screen and surface cover treatments. The no treatment option exhibited the largest mean S A E and SOTE values, followed in decreasing order by the D B C A , bubble screen, vanes and surface cover treatments. A significant orifice diameter x flow rate interaction effect was observed for K i ^ o , SOTR, S A E and SOTE (Table 4.25). Table 4.25. Adjusted least squares means (± SE), treatment and interaction effects for various in situ full lift hypolimnetic aerator treatments on air with the 800 p. diffuser. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr1) SAE (g 0 2 kWhr SOTE (%) n Bubble screen 1.2 (0.02) 13.4 (0.25) 121.9 (3.32) 3.0 (0.08) 12 DBCA 1.5 (0.02) 16.0 (0.25) 154.2 (3.32) 3.7 (0.08) 12 None 2.2 (0.02) 22.1 (0.25) 197.1 (3.32) 4.8 (0.08) 12 Surface cover 1.1 (0.02) 12.1 (0.25) 107.6 (3.32) 2.6 (0.08) 12 Vanes 1.2 (0.02) 13.5 (0.25) 121.0 (3.32) 2.9 (0.08) 12 Interaction Yes Yes Yes Yes In situ effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The log transformation did not linearize the orifice diameter x flow rate interaction effect for K i ^ o , SOTR, S A E and SOTE, hence the analysis was re-run without the D B C A data, since all of the plots showed that the D B C A values decreased substantially with increasing gas flow. The re-analyzed results indicated K i ^ o , SOTR, S A E and SOTE exhibited parallel slopes, so the interaction effect was linearized. Scheffe's test indicates the no modifications treatment was significantly greater than the other in situ treatments (a < 0.01) for KLa2o and SOTR (Tables 4.26 and 4.27). Table 4.26. Matrix of KLa2o pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on air with the 800 u, diffuser. Treatments Bubble screens None Surface Cover Vanes Bubble screens 1.000 None 0.000 1.000 Surface cover 0.105 0.000 1.000 Vanes 0.999 0.000 0.136 1.000 Table 4.27. Matrix of SOTR pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on air with the 800 p diffuser. Treatments Bubble screens None Surface cover Vanes Bubble screens 1.000 None 0.000 1.000 Surface cover 0.108 0.000 1.000 Vanes 1.000 0.000 0.136 1.000 For S A E and SOTE, Scheffe's test indicates the surface cover treatment was significantly less (a < 0.01) than the bubble screen, vane and no treatment combinations. The no treatment option was significantly greater than all other treatments (Table 4.28). Table 4.28. Matrix of SAE and SOTE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on air with the 800 p diffuser. Treatments Bubble screen None Surface cover Vanes Bubble screen 1.000 None 0.000 1.000 Surface cover 0.002 0.000 1.000 Vanes 0.998 0.000 0.004 1.000 The log transformation linearized the in situ treatment x flow rate interaction effect. The multiplicative effects (positive or negative) of increasing gas flow rate on K i ^ o , SOTR, S A E and SOTE are shown in Tables 4.29 and Table 4.30. Table 4.29. Matrix of multiplicative in situ treatment x flow rate interaction effects for KLa2o and SOTR for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on air with the 800 p diffuser. Treatments Bubble screen None Surface cover Vanes Bubble screen 0.00 None 1.62 0.00 Surface cover 0.89 0.55 0.00 Vanes 0.99 0.61 1.12 0.00 Table 4.30. Matrix of multiplicative in situ treatment x flow rate interaction effects for SAE and SOTE for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on air with the 800 p diffuser. Treatments Bubble screen None Surface cover Vanes Bubble screen 0.00 None 1.63 0.00 Surface cover 0.89 0.55 0.00 Vanes 1.00 0.61 1.12 0.00 A summary of the adjusted least squares means diffuser in situ modification treatment effects of the 140 and 800 p on air showing the percentage increase/decrease relative to No Treatment are shown in Table 4.31. Table 4.31. Summary of various in situ full lift hypolimnetic aerator treatments on air with the 140 p and 800 p diffuser showing percentage increase/decrease relative to No Treatment. Treatment K L a 2 0 (hr1) SOTR SAE SOTE n (g 0 2 hr1) (g 0 2 kWhr-1) (%) None x 140 ji 3.3 35.4 324.5 7.9 12 Vanes x 140 u 2.4 (-27%) 26.5 (-25%) 248.1 (-24%) 6.0 (-24%) 12 DBCA x 140 u 2.3 (-30%) 24.9 (-30%) 241.3 (-26%) 5.9 (-25%) 12 Screen x 140 u 2.1 (-36%) 22.5 (-36%) 209.6 (-35%) 5.1 (-35%) 11 Cover x 140 \i 1.9 (-42%) 20.6 (-42%) 192.7 (-41%) 4.7 (-41%) 12 None x 800 p 2.2 22.1 197.1 4.8 12 DBCA x 800 n 1.5 (-32%) 16.0 (-28%) 154.2 (-22%) 3.7 (-23%) 12 Vanes x 800 u 1.2 (-45%) 13.5 (-39%) 121.0 (-39%) 2.9 (-40%) 12 Screen x 800 \i 1.2 (-45%) 13.4 (-39%) 121.9(-38%) 3.0 (-38%) 12 Cover x 800 u 1.1 (-50%) 12.1 (-45%) 107.6 (-45%) 2.6 (-46%) 12 4.1.4.3 PSA oxygen x 140 JU diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 140 fi diffuser on PSA O2 at 2.9 m). The results for the analysis of covariance for the various in situ modifications of the full lift hypolimnetic aerator with the 140 u. diffuser on PSA oxygen are shown in Table 4.32. A l l tests were conducted at 2.9 m. Significant results (a < 0.01) were obtained for the in situ treatment and flow rate covariate effect for K i ^ o , SOTR, S A E and SOTE (Figure 4.19 a-d). The no 30 „ 20 to 1 CM O LU < to 10 MOÔS i I • Screen • + DBCA • y— - • None • Cover • y + + Vanes -i i i 10 20 30 40 Oxygen Flow (L min-1) 50 MODS 10 20 30 40 Oxygen Flow (L min-1) 50 -C CM O •3 en F O to 300 200-100 MO6S i i • Screen / > + DBCA - • None • + • Cover y/ + Vanes I I UJ H O to 10 20 30 40 Oxygen Flow (L min-1) 50 MODS 10 20 30 40 Oxygen Flow (L min-1) 50 Figure 4.19a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for in situ modifications on PSA 0 2 with 140 p diffuser at 2.9 m. treatment option exhibited the largest mean K.La2o and SOTR values, followed in decreasing order by the D B C A , surface cover, vanes and bubble screen treatments. The no treatment option exhibited the largest mean S A E and SOTE values, followed in decreasing order by the D B C A , vanes, bubble screen and surface cover treatments. A significant orifice diameter x flow rate interaction effect was observed for K i ^ o and SOTR (Table 4.32). Scheffe's test indicates the no modifications treatment was significantly greater than the other in situ treatments (a < 0.01) for KL.a2oand SOTR (Table 4.33). There were no significant differences among all other pairwise comparisons. Table 4.32. Adjusted least squares means (± SE), treatment and interaction effects for the various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 p diffuser. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr 1) SAE (g 0 2 kWhr ]) SOTE (%) n Bubble screen 9.9 (0.44) 108.3 (4.75) 46.14(2.07) 5.8 (0.27) 12 DBCA 11.9 (0.44) 130.0 (4.75) 55.71 (2.07) 7.1 (0.27) 12 None 17.5 (0.47) 190.2 (4.97) 78.9 (2.17) 10.5 (0.28) 12 Surface cover 10.0 (0.44) 108.9 (4.75) 45.2 (2.07) 5.7 (0.27) 12 Vanes 10.0 (0.48) 108.3 (5.21) 49.4 (2.27) 6.3 (0.29) 10 Interaction Yes Yes No No In situ effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Table 4.33. Matrix of Ki^o and SOTR pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.038 1.000 None 0.000 0.000 1.000 Surface cover 1.000 0.024 0.000 1.000 Vanes 0.996 0.128 0.000 0.985 1.000 For SAE and SOTE, Scheffe's test indicates the no treatment option was significantly (a < 0.01) greater than the bubble breaker, DBCA, surface cover and vanes treatments (Tables 4.34 and 4.35). There were no significant differences among all other pairwise comparisons. Table 4.34. Matrix of SAE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.026 1.000 None 0.000 0.000 1.000 Surface cover 1.000 0.016 0.000 1.000 Vanes 0.957 0.194 0.000 0.911 1.000 Table 4.35. Matrix of SOTE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.022 1.000 None 0.000 0.000 1.000 Surface cover 1.000 0.012 0.000 1.000 Vanes 0.961 0.166 0.000 0.908 1.000 The log transformation linearized the in situ treatment x flow rate interaction effect on K i ^ o and SOTR. Table 4.36 lists the multiplicative effect (positive or negative) of increasing gas flow rate on K i ^ o and SOTR for various in situ treatments. Table 4.36. Matrix of multiplicative in situ treatment x flow rate interaction effects for KL32O and SOTR for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 140 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 0.00 DBCA 1.20 0.00 None 1.71 1.43 0.00 Surface cover 0.99 0.82 0.58 0.00 Vanes 1.03 0.85 0.60 1.04 0.00 S A E and SOTE were not influenced by the in situ treatment x flow rate interaction effect, and declined with increasing gas flow rate, particularly the D B C A treatment. 4.1.4.4 PSA oxygen x 800 JU diffuser (Note: compares in situ modifications of the full lift hypolimnetic aerator with the 800 [i diffuser on PSA O2 at 2.9 m). The results for the analysis of covariance for the various in situ modifications of the full lift hypolimnetic aerator with the 800 p diffuser on PSA oxygen are shown in Table 4.37. Signifi-cant results (a < 0.01) were obtained for the in situ treatment and flow rate covariate effect for KLa2o, SOTR, S A E and SOTE (Figure 4.20 a-d). The no treatment option exhibited the largest mean K i ^ o , SOTR, S A E and SOTE values, followed in decreasing order by the D B C A , bubble screen, vanes and surface cover treatments. A significant in situ treatment x flow rate interac-tion effect was observed for K L a 2 0 , SOTR, S A E and SOTE (Table 4.37). Table 4.37. Adjusted least squares means (± SE), treatment and interaction effects for various in situ full lift hypolimnetic aerator treatments on PSA oxygen with the 800 p diffuser. Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr 1) SAE (g O z kWhr 1) SOTE (%) n Bubble screen 6.8(0.16) 74.3 (1.69) 30.0 (0.83) 3.8 (0.10) 12 DBCA 8.2 (0.16) 89.7(1.69) 39.3 (0.83) 5.0 (0.10) 12 None 11.5 (0.16) 124.8(1.69) 50.4 (0.83) 6.4 (0.10) 12 Surface cover 6.2 (0.16) 67.2(1.69) 26.4 (0.83) 3.3 (0.10) 12 Vanes 6.7(0.16) 73.3 (1.69) 29. 0 (0.83) 3.7 (0.10) 12 Interaction Yes Yes Yes Yes In situ effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Flow effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 The log transformation did not linearize the orifice diameter x flow rate interaction effect for K L a 2 o, SOTR, S A E and SOTE, hence the analysis was re-run without the D B C A data, since all of the plots showed the D B C A values decreasing more rapidly with increasing gas flow than the other treatments. The re-analyzed results indicated K L a 2 o, SOTR, S A E and SOTE exhibited parallel slopes, so the interaction effect was linearized. Scheffe's test indicates the no modifica-tions treatment was significantly greater than the remaining in situ treatments (a < 0.01) for K L a 2 0 and SOTR (Table 4.38). Figure 4.20a-d K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) for in situ modifications on PSA 0 2 with 800 u diffuserai 2.9 m. Table 4.38. Matrix of KLa 2o and SOTR pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on PSA oxygen with the 800 p diffuser. Treatments Bubble screen None Surface Cover Vanes Bubble screen 1.000 None 0.000 1.000 Surface cover 0.130 0.000 1.000 Vanes 0.939 0.000 0.364 1.000 Scheffe's test indicates that the no treatment S A E and SOTE was significantly greater (a < 0.01) than the bubble screen, vane and surface cover treatments and that S A E and SOTE for the bubble screen treatment was significantly greater than the surface cover treatment (Table 4.39). Table 4.39. Matrix of SAE pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on PSA oxygen with the 800 p diffuser. Treatments Bubble screen None Surface cover Vanes Bubble screen 1.000 None 0.000 1.000 Surface cover 0.000 0.000 1.000 Vanes 0.673 0.000 0.011 1.000 The log transformation linearized the in situ treatment x flow rate interaction effect. The multiplicative effects (positive or negative) of increasing gas flow rate on K i ^ o and SOTR are shown in Table 4.40, and for S A E and SOTE in Table 4.41. Table 4.40. Matrix of multiplicative in situ treatment x flow rate interaction effects for KL32O and SOTR for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on PSA oxygen with the 800 p diffuser. Treatments Bubble screen None Surface cover Vanes Bubble screen 0.00 None 1.68 0.00 Surface cover 0.88 0.53 0.00 Vanes 0.97 0.58 1.10 0.00 A summary of the adjusted least squares means diffuser in situ modification treatment effects of the 140 and 800 p on PSA O2 showing the percentage increase/decrease relative to No Treat-ment are shown in Table 4.42. Table 4.41. Matrix of multiplicative in situ treatment x flow rate interaction effects for SAE and SOTE for various in situ full lift hypolimnetic aerator treatments (excluding DBCA) on PSA oxygen with the 800 p diffuser. Treatments Bubble screen None Surface cover Vanes Bubble screen 0.00 None 1.68 0.00 Surface cover 0.88 0.53 0.00 Vanes 0.97 0.58 1.10 0.00 Table 4.42. Summary of various in situ full lift hypolimnetic aerator treatments on PSA O 2 with the 140 p and 800 p diffuser showing percentage increase/decrease relative to No Treatment. Treatment K L a 2 0 (hr 1) SOTR (g 0 2 hr 1 ) SAE (g 0 2 kWhr 1 ) SOTE(%) n None x140 p 17.5 190.2 78.9 10.5 12 DBCA x 140 p 11.9 (-32%) 130.0 (-32%) 55.7 (-29%) 7.1 (-32%) 12 Vanes x 140 p 10.0 (-43%) 108.3 (-43%) 49.4 (-37%) 6.3 (-40%) 10 Cover x 140 p 10.0 (-43%) 108.9 (-43%) 45.2 (-43%) 5.7 (-46%) 12 Screen x 140 p 9.9 (-43%) 108.3 (-43%) 46.1 (-43%) 5.8 (-45%) 12 None x 800 p 11.5 124.8 50.4 6.4 12 DBCA x 800 \i 8.2 (-29%) 89.7 (-28%) 39.3 (-22%) 5.0 (-22%) 12 Screen x 800 p 6.8 (-41%) 74.3 (-40%) 30.0 (-41%) 3.8 (-41%) 12 Vanes x 800 p 6.7 (-42%) 73.3 (-41%) 29. 0 (-42%) 3.7 (-42%) 12 Cover x 800 p 6.2 (-46%) 67.2 (-46%) 26.4 (-48%) 3.3 (-48%) 12 4.1.5 Miscellaneous Tests The results of the Pump Only full lift tests in D B C A mode were considerably different from the regular performance of the full lift hypolimnetic aerator. The mean values for K.La2o, SOTR and S A E were among the lowest recorded for any of the full lift tests; however, the mean SOTE values were among the highest recorded (Table 4.43). The results of the Extended Aeration tests at the lowest gas flow rate in D B C A mode were similar to the values obtained for K L a 2 o, SOTR, S A E and SOTE during the Group 7 (i.e., 140 u. at 2.9 m) tests at the 7 and 3 1 min"1 gas flow rates (Table 4.44). The standard error of the Extended Aeration tests was larger, due to the smaller sample size and increased variance for the 7/3 1 min"1 - Ext. air tests. Table 4.43. Mean (+ SE) K L a 2 0 , SOTR, SAE and SOTE values for full lift hypolimnetic aerator operating in pump only DBCA mode on air or PSA oxygen at 3 1 min"1. Treatment K L a 2 0 (hr 1) SOTR (g0 2 hr 1 ) SAE (g 0 2 kWhr 1 ) SOTE(%) n Pump only -air 0.6 (0.01) 6.2 (0.14) 8.0 (0.18) 11.3 (0.25) 3 Pump only -PSA 0 2 2.7 (0.07) 29.6 (0.74) 28.4 (0.71) 12.7 (0.33) 7 Table 4.44. Mean (± SE) K L a 2 0 , SOTR, SAE and T E values for the full lift hypolimnetic aerator operating in DBCA mode on air or PSA oxygen at 7 and 3 1 min"1 at 2.9 m using a 140 p diffuser. (Note: 7/3 refers to the gas flow, in 1 min"1, to the diffuser in the inlet tube, and the DBCA diffuser in the separator box). Treatment KLa 2o (hr1) SOTR (2 0 2 hr 1) SAE (g 0 2 kWhr 1) SOTE (%) n 7/3 1 min 1 - Ext. air 1.5(0.32) 15.8(3.48) 359.3 (78.99) 8.4(1.84) 2 7/3 1 min 1 - Grp 7 air 1.4 (0.03) 14.8(0.35) 322.9 (7.57) 7.8(0.18) 3 7/3 1 min 1 - Ext. PSA 0 2 7.1 (0.16) 77.3 (1.78) 77.1 (1.78) 9.7 (0.23) 2 7/3 1 min 1 - Grp 7 PSA O z 7.3 (0.56) 79.0 (6.09) 76.8 (5.92) 9.9 (0.72) 3 The purpose of the extended aeration tests was to determine the oxygen saturation value achieved after an extended period of operation. The lowest gas flow rate setting was required as the PSA unit could not generate high purity oxygen for extend periods at flow rates exceeding 10 1 min"1. The highest oxygen concentration achieved on air was 11.47 mg l" 1 after 247 min-utes of operation, and the % saturation achieved (i.e., 96.4%) exceeded the C * s calculated for the day (i.e., 93.3 %) by 3.1%. The highest oxygen concentration achieved on PSA oxygen was 50.44 mg l" 1 or 446.4 % saturation after 206 minutes of operation, thus exceeding the C* sfor the day for air (i.e., 103.2 %) by a ratio of 4.33 (Table 4.45). Treatment Date Time (min) 0 2 mg l"1 % Sat. c* s Temp (° C) Ext. Air-1 July 19/00 247 11.47 96.4 93.3 11.8 Ext. Air-2 July 20/00 270 11.37 100.6 103.2 12.4 Ext. 0 2 - l July 19/00 183 48.45 407.1 93.3 12.6 Ext. 02-2 July 20/00 206 50.44 446.4 103.2 11.2 4.2 Full Lift Hypolimnetic Aeration - Water Velocity 4.2.1 Gas Effect Note: this component of the experiment examined the effect of gas flow variation (10 to 40 1 min"1) on water velocity in the outlet tube, using air and PSA oxygen, at three different orifice diameters (140 p, 400 p and 800 p), at two diffuser depths (1.5 m and 2.9 m) in the full lift inlet tube. Within each subset of experiments, the orifice diameter and diffuser depth were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. The results for the analysis of covariance for the 140, 400 and 800 p diffusers operating on air or PSA oxygen at 1.5 and 2.9 m depth are shown in Table 4.46. As was expected, their were no significant differences in water velocity in the full lift aerator as a function of air or PSA oxygen among all combinations of diffuser orifice diameter and depth of gas release (Figures 4.21 a-c and 4.22 a-b). Water velocity was significantly influenced (a < 0.01) by gas flow rate in all treatment combinations, increasing in value with increasing gas flow rates. The water velocity measurements for the PSA O2 x 800 p @ 2.9 m treatment were not analyzed as they were too erratic, and not believable due to a measurement error. The Ai r x 800 p @ 2.9 m treatment data was useable; however, the SE was inflated due to the erratic behavior of the PSA O2 x 800 p @ 2.9 m data, which was caused by equipment malfunction. 0 10 20 30 40 50 Gas Flow (L min-1) Figure 4.21a-c Water velocity vs. gas flow rate on air or PSA C*2at 1.5 m using 140 u (a), 400 \i (b) and 800 u (c) diffusers. Figure 4.22a-b Water velocity vs. gas flow rate on air or PSA 0 2 at 2.9 m using 140 u. (a) and 400 u (b) diffusers. Table 4.46. Mean (± SE) water velocity (cm sec'1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - gas effect. Treatment Velocity (cm sec"1) Gas flow sig. Gas composition sig. n Air x 140 p @ 1.5 m 22.0 (0.48) Yes No 12 PSA 0 2 x 140 p @ 1.5 m 22.4 (0.48) 12 Air x 400 p @ 1.5 m 22.9 (0.48) Yes No 12 PSA 0 2 x 400 p @ 1.5 m 22.8 (0.48) 12 Air x 800 p @ 1.5 m 16.4 (0.61) Yes No 12 PSA O z x 800 p @ 1.5 m 17.4(0.62) 12 Air x 140 p @ 2.9 m 32.7 (0.75) Yes No 12 PSA 0 2 x 140 p @ 2.9 m 32.8 (0.75) 12 Air x 400 p @ 2.9 m 30.4 (0.51) Yes No 12 PSA 0 2 x 400 p @ 2.9 m 30.4 (0.51) 12 Air x 800 p @ 2.9 m 32.1 (2.56) Yes No 12 PSA 0 2 x 800 p @ 2.9 m N / A N / A 4.2.2 Depth Effect Note: this component of the experiment examined the effect of gas flow variation (10 to 40 1 min"') on water velocity in the outlet tube, using air only, at two diffuser depths (1.5 m and 2.9 m) in the full lift inlet tube with three orifice diameters (140 p, 400 p and 800 p). Within each subset of experiments, the orifice diameter and diffuser depth were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. Since there was no significant effect of gas composition on water velocity, the remaining depth effect analyses were conducted on air only, since the erratic PSA 0 2 x 800 p @ 2.9 m data set was discarded. The results for the analysis of covariance for the 1.5/2.9 m x 140 p diffuser treatment are shown in Table 4.47. Significant results (a < 0.01) were obtained for the depth and gas flow rate covariate effect (Figure 4.23 a-c), with water velocity increasing in value with depth of gas release and gas flow rate. No depth x gas flow rate interaction effect was observed Figure 4.23a-c Water velocity vs. air flow rate at 1.5 or 2.9 m using 140 p (a), 400 p (b) and 800 p (c) diffusers. Table 4.47. Mean (± SE) water velocity (cm sec-1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - depth effect. Treatment Velocity (cm sec-1) Gas flow sig. Depth sig. Interaction n 1.5 m x 140 u, - Air 22.4 (0.60) Yes Yes No 12 2.9 m x 140 u - Air 33.2 (0.60) 12 1.5 m x 400 u, - Air 23.3 (0.56) Yes Yes Yes, 1.30 12 2.9 m x 400 u. - Air 31.0 (0.56) 12 1.5 m x 800 u - Air 16.8 (0.59) Yes Yes Yes, 1.98 12 2.9 m x 800 u - Air 31.1 (0.59) 12 for the 1.5/2.9 m x 140 p. diffuser treatment (Table 4.46). A significant depth x gas flow rate interaction effect was observed for the 1.5/2.9 m x 400 p, and 1.5/2.9 m x 800 p. diffuser treat-ments, hence the data was log transformed and re-analyzed. The log transformation linearized the interaction effect, and indicates that velocity is significantly influenced by depth of release and gas flow rate, and that water velocity increases 1.30 fold at 2.9 m depth as compared to the 1.5 m depth of gas release with the 400 p diffuser, and increases 1.98 fold at 2.9 m depth as compared to the 1.5 m depth of gas release with the 800 p diffuser (Table 4.47). 4.2.3 Orifice Effect Note: this component of the experiment examined the effect of gas flow variation (10 to 40 1 min - 1) on water velocity in the outlet tube, using air only, with three orifice diameters (140 p, 400 p and 800 p) at two diffuser depths (1.5 m and 2.9 m) in the full lift inlet tube. Within each subset of experiments, the orifice diameter and diffuser depth were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. The results for the analysis of covariance for the 140, 400 and 800 p diffusers operating on air at 1.5 and 2.9 m depth are shown in Table 4.48. At the 1.5 m release depth, water velocity was significantly influenced by orifice diameter and gas flow rate, with velocity increasing with gas flow rate (Figure 4.24 a-b). Scheffe's test indicates the 140 and 400 p. diffusers induced signifi-cantly greater velocities than the 800 p diffuser; however, there was no significant difference between the 140 and 400 p, diffusers. There was no significant interaction effect; however, the graph of the data (Figure 4.24 a) suggests an interaction between gas flow rate and water velocity for the 140 p, diffuser. The significance of the orifice by gas flow rate interaction effect was near the significance level cutoff (i.e., 0.012 vs. 0.01); hence, given the sample size and variance of the response, the test was not able to detect a significant interaction effect. 40 § 30 E è 20 o o CO > CD -i I i ' Orifice (1.5 m) 1 i « 140 u • 400 u • 800 u i i 50 o S 40 30 o o CD > S 20 m 10 20 30 40 Air Flow (I min-1) 50 10 I I I Orifice (2.9 m) a 140 u • 400 u V I • 800 u 10 20 30 40 Air Flow (L min-1) 50 Figure 4.24a-b Water velocity vs. air flow rate for 140, 400 and 800 p diffusers at 1.5 m (a) and 2.9 m (b). Table 4.48. Mean (± SE) water velocity (cm sec"1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - orifice effect. Treatment Velocity (cm sec"1) Gas flow Orifice Interaction n sig. sig. 140 p x 1.5 m - Air 22.4 (0.63) Yes Yes No 12 400 px 1.5 m - A i r 23.3 (0.63) 12 800 px 1.5 m - A i r 16.8 (0.63) 12 140 p x 2.9 m - Air 33.2 (0.58) Yes No No 12 400 p x 2.9 m - Air 31.0(0.58) 12 800 p x 2.9 m - Air 31.1 (0.58) 12 There was no statistically significant effect of orifice diameter on water velocity at the 2.9 m release depth, nor any orifice diameter x gas flow rate interaction effect (Table 4.47). There was a significant gas flow rate effect as would be expected, with water velocity increasing with increasing gas flow rate. 4.2.4 Full Lift Modifications Effect Note: this component of the experiment examined the effect of gas flow variation (10 to 40 1 min"1) on water velocity in the outlet tube, using air only, on four in situ modifications, using two orifice diameters (140 u. and 800 u) at a single diffuser depth (i.e., 2.9 m) in the full lift inlet tube. Within each subset of experiments, the orifice diameter and diffuser depth were held constant, while the gas flow rate and oxygen content (i.e., air or PSA oxygen) were varied. The results for the analysis of covariance for the various in situ modifications of the full lift hypolimnetic aerator with the 140 and 800 u. diffuser are shown in Table 4.49. A l l tests were conducted at 2.9 m. Water velocity was significantly influenced (a < 0.01) by the in situ treatment and flow rate covariate effect (Figure 4.25 a-b). The no treatment option exhibited the highest mean velocity, followed in decreasing order by the surface cover, bubble screen, D B C A and vanes treatments. There were no significant modifications x flow rate interaction effect for velocity. A similar result was obtained for the 800 u diffuser (Figure 4.25 b and Table 4.49). For the 140 u, diffuser, Scheffe's test indicates the no modifications treatment was significantly greater than the D B C A and vanes treatments (a < 0.01) for velocity (Table 4.50). In addition, velocity for the surface cover treatment was significantly greater than the vanes treatment. For the 800 u. diffuser, Scheffe's test indicates the no modifications treatment velocity was significantly greater than the other treatments (a < 0.01) (Table 4.51). In addition, velocity for the bubble screen and D B C A treatment were significantly greater than the vanes treatment. Air Flow (L min-1) Air Flow (L min-1) Figure 4.25a-b Water velocity vs. air flow rate at 2.9 m using 140 p (a) or 800 p (b) diffusers with in situ modifications. Table 4.49. Mean (± SE) water velocity (cm sec"1) measured near the top of the downflow tube in the full lift hypolimnetic aerator - full lift in situ modifications effect. Treatment Velocity (cm sec"1) Gas flow In situ Interaction n sig. sig. Bubble screen x 140 p 29.0(1.04) Yes Yes No 11 DBCA x 140 p 26.7 (0.95) 13 None x140 p 33.2 (0.99) 12 Surface cover x 140 p 30.4 (0.99) 12 Vanes x 140 p 23.9 (0.09) 12 Bubble screen x 800 p 23.8 (0.77) Yes Yes No 12 DBCA x 800 p 24. 5 (0.81) 11 None x 800 p 31.2 (0.77) 12 Surface cover x 800 p 23.2 (0.77) 12 Vanes x 800 p 19.6 (0.77) 12 Table 4.50. Matrix of water velocity pairwise comparison probabilities for Scheffe's test for various /'// situ full lift hypolimnetic aerator treatments with the 140 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.600 1.000 None 0.085 0.001 1.000 Surface cover 0.924 0.139 0.395 1.000 Vanes 0.021 0.414 0.000 0.001 1.000 Table 4.51. Matrix of water velocity pairwise comparison probabilities for Scheffe's test for various in situ full lift hypolimnetic aerator treatments with the 800 p diffuser. Treatments Bubble screen DBCA None Surface cover Vanes Bubble screen 1.000 DBCA 0.989 1.000 None 0.000 0.000 1.000 Surface cover 0.990 0.880 0.000 1.000 Vanes 0.010 0.003 0.000 0.038 1.000 4.3 Full Lift Hypolimnetic Aeration - Bubble Size A total of 1,890 bubble measurements were used in the statistical analysis (Table 4.52). A frequency histogram of the mean Equivalent Bubble Diameter (EQD) indicates the majority of bubbles were < 5 mm in equivalent diameter (Figure 4.26 a), with a mean equivalent bubble diameter of 3.7 mm and a range of 13.6 mm from the smallest to the largest bubble size. Due to the left skewed distribution of the equivalent bubble diameters, the data was transformed using a log x 2 transformation (Figure 4.26 b), and the statistical analyses conducted on the transformed data. The bubble photography was conducted during a 1.5 day photography session after the Group 1-14 tests had been completed, in order to minimize the cost of the photographer. Table 4.52. Summary statistics for bubble measurements. Number of cases 1,890 M i n . E Q D dia. (mm) 0.7 Max. E Q D dia. (mm) 14.3 Range E Q D dia. (mm) 13.6 Mean E Q D dia. (mm) 3.7 Standard Dev. (mm) 1.9 CV (%) 52.4 400 300 200 -100 -o O 3. CD 01 0 5 10 Equivalent Bubble Diameter (mm) 300 200 -100 -3 T3 O a-o CD O) 1 2 3 4 Equivalent Bubble Diameter-Logx2 (mm) Figure 4.26a-b Frequency histogram, untranformed (a) and log x 2 transformed (b), of Equivalent Bubble Diameters. The data are presented in two series of tables. The first series of tables (Tables 4.53, 4.54 and 4.55) illustrates the effect of gas flow rate on equivalent bubble diameter for the 140 p, 400 p and 800 p diffusers measured under three different diffuser arrangements: (1) bubbles generated by each diffuser at 2.9 m and measured at 2.7 m; (2) bubbles generated by each diffuser at 2.9 m and measured at 0.5 m and (3) bubbles generated by each diffuser at 1.5 m and measured at 0.5 m. The purpose of this section of the experiments was to determine if bubble diameter changed as a function of depth of gas release (i.e., 2.9 m vs. 1.5 m), and i f bubble diameter changed as they rose through the water column (i.e., 2.7 m vs. 0.5 m), due to changes in hydrostatic pres-sure and possible coalescence/breakup during ascent through the water column. Table 4.53. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 140 p diffuser at various depths of gas release and bubble measurement. Gas flow Orifice dia. Diffuser Photo Mean equiv. SD C V n (1 m i n 1 ) (P) depth (m) depth (m) Bubble dia. (mm) (mm) (%) 10 140 2.9 2.7 3.6 0.89 24.9 30 20 140 2.9 2.7 3.4 0.98 28.5 30 30 140 2.9 2.7 3.4 0.93 27.8 30 40 140 2.9 2.7 3.7 0.86 23.3 30 10 140 2.9 0.5 3.6 1.54 42.7 30 20 140 2.9 0.5 3.9 1.25 32.1 30 30 140 2.9 0.5 4.1 1.26 30.7 30 40 140 2.9 0.5 4.6 0.94 20.7 30 10 140 1.5 0.5 4.4 0.89 20.2 30 20 140 1.5 0.5 4.4 0.98 22.5 30 30 140 1.5 0.5 4.2 0.75 18.0 30 40 140 1.5 0.5 4.6 1.26 27.3 30 Table 4.54. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 400 p diffuser at various depths of gas release and bubble measurement. Gas flow Orifice dia. Diffuser Photo Mean equiv. SD C V n (1 m i n 1 ) (P) depth (m) depth (m) Bubble dia. (mm) (mm) (%) 10 400 2.9 2.7 4.3 1.01 23.3 30 20 400 2.9 2.7 4.7 1.22 25.8 30 30 400 2.9 2.7 5.0 1.43 28.5 30 40 400 2.9 2.7 4.3 1.93 45.2 30 10 400 2.9 0.5 4.2 1.52 36.2 30 20 400 2.9 0.5 4.1 1.7 41.2 30 30 400 2.9 0.5 4.0 2.13 54.0 30 40 400 2.9 0.5 4.2 2.31 54.8 30 10 400 1.5 0.5 3.3 1.97 60.2 30 20 400 1.5 0.5 4.3 1.87 43.3 30 30 400 1.5 0.5 3.2 1.95 61.8 30 40 400 1.5 0.5 3.4 2.10 61.3 30 Table 4.55. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 800 p diffuser at various depths of gas release and bubble measurement. Gas flow Orifice Diffuser Photo Mean equiv. SD C V n (1 min 1) dia. (p) depth (m) depth (m) Bubble dia. (mm) (mm) (%) 10 800 2.9 2.7 5.7 2.57 45.4 30 20 800 2.9 2.7 3.5 2.26 64.5 30 30 800 2.9 2.7 5.3 2.38 44.7 30 40 800 2.9 2.7 3.5 2.16 62.4 30 10 800 2.9 0.5 2.8 2.06 73.1 30 20 800 2.9 0.5 3.3 2.40 72.3 30 30 800 2.9 0.5 3.5 2.43 70.4 30 40 800 2.9 0.5 3.6 2.05 56.8 30 10 800 1.5 0.5 2.8 2.91 104.2 30 20 800 1.5 0.5 3.3 2.45 74.0 30 30 800 1.5 0.5 3.1 2.45 79.2 30 40 800 1.5 0.5 3.9 3.11 80.5 30 The second series of tables (Tables 4.56 and 4.57) shows the effect of gas flow rate on equiva-lent bubble diameter for the 140 p and 800 p diffusers with a control case (i.e. None) and three in situ modifications: (1) D B C A ; (2) bubble screen and (3) vanes. The bubbles were generated at a fixed depth of 2.9 m and measured at 0.5 m. The purpose of this was to determine if bubble diameter changed as a function of the in situ modifications. Table 4.56. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 140 p diffuser with in situ modifications. (Note: 7/3,17/3,27/3 and 37/3 refers to the gas flow, in 1 min 1 , to the diffuser in the inlet tube, and the DBCA diffuser in the separator box). Group Gas flow (1 min 1) Orifice dia. (p) Diffuser depth (m) Photo depth (m) Mean equiv. Bubble dia. (mm) SD (mm) C V (%) n 1-Cont. 10 140 2.9 0.5 3.6 1.54 42.7 30 1-Cont. 20 140 2.9 0.5 3.9 1.25 32.1 30 1-Cont. 30 140 2.9 0.5 4.1 1.26 30.7 30 1-Cont. 40 140 2.9 0.5 4.6 0.94 20.7 30 7 - D B C A 7/3 140 2.9 0.5 1.7 0.87 50.0 30 7 - D B C A 17/3 140 2.9 0.5 1.7 0.96 55.5 30 Group Gas flow Orifice Diffuser Photo Mean equiv. SD C V n (I min 1) dia. (p) depth (m) depth (m) Bubble dia. (mm) (mm) (%) 7 - D B C A 27/3 140 2.9 0.5 2.0 0.91 45.9 30 7 - D B C A 37/3 140 2.9 0.5 1.5 0.83 55.5 30 9-Screen 10 140 2.9 0.5 4.3 0.96 22.3 30 9-Screen 20 140 2.9 0.5 3.7 1.28 35.1 30 9-Screen 30 140 2.9 0.5 3.7 0.90 24.6 30 9-Screen 40 140 2.9 0.5 4.0 1.43 35.6 30 13-Vane 10 140 2.9 0.5 4.5 0.91 20.4 30 13-Vane 20 140 2.9 0.5 4.5 0.79 17.6 30 13-Vane 30 140 2.9 0.5 4.1 0.90 22.1 30 13-Vane 40 140 2.9 0.5 4.4 1.13 25.6 30 Table 4.57. Mean equivalent bubble diameter, standard deviation and coefficient of variation of bubbles produced by the 800 p diffuser with in situ modifications (Note: 7/3,17/3,27/3 and 37/3 refers to the gas flow, in 1 min 1 , to the diffuser in the inlet tube, and the DBCA diffuser in the separator box). Group Gas flow Orifice Diffuser Photo Mean equiv. SD CV n (1 min-1) dia. (p) depth (m) depth (m) Bubble dia. (mm) (mm) (%) 5-Cont. 10 800 2.9 0.5 2.8 2.06 73.1 30 5-Cont. 20 800 2.9 0.5 3.3 2.40 72.3 30 5-Cont. 30 800 2.9 0.5 3.5 2.43 70.4 30 5-Cont. 40 800 2.9 0.5 3.6 2.05 56.8 30 8-DBCA 7/3 800 2.9 0.5 2.8 1.10 39.4 30 8-DBCA 17/3 800 2.9 0.5 1.8 0.72 41.3 30 8-DBCA 27/3 800 2.9 0.5 1.7 1.02 60.2 30 8-DBCA 37/3 800 2.9 0.5 1.6 0.79 49.6 30 10-Screen 10 800 2.9 0.5 3.7 1.49 40.3 30 10-Screen 20 800 2.9 0.5 3.4 1.53 45.1 30 10-Screen 30 800 2.9 0.5 3.9 1.11 28.2 30 10-Screen 40 800 2.9 0.5 3.9 1.09 28.1 30 14-Vane 10 800 2.9 0.5 5.7 2.61 46.2 30 14-Vane 20 800 2.9 0.5 4.5 2.31 51.0 30 14-Vane 30 800 2.9 0.5 5.0 2.43 48.6 30 14-Vane 40 800 2.9 0.5 3.7 1.21 32.8 30 4.3.1 Depth Effect This component of the experiment examined the effect of gas flow variation (10 to 40 1 min - 1) and three combinations of diffuser/bubble measurement depth (i.e., 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m) in the full lift inlet tube, on EQD, on air, with three orifice diameters (140 u, 400 u. and 800 p). 4.3.1.1 140n diffuser at 1.5/0.5 m, 2.9/0.5 m and2.9/2.7 m The results for the analysis of covariance for the 140 p diffuser x 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m (i.e., bubble release depth/bubble measurement depth) treatments are shown in Table 4.58. Significant results (a < 0.01) were obtained for the depth effect, and a borderline effect (i.e., a = 0.01) was observed for the gas flow rate covariate effect (Figure 4.27). Bubble size was largest for the 1.5/0.5 m combination, then decreased in size at the 2.9/0.5 m and 2.9/2.7 m depth combinations. No depth x gas flow rate interaction effect was observed for any of the treatments (Table 4.58). Table 4.58. Adjusted least squares means (± SE), treatment and interaction effects for the 140 p diffuser on air at bubble release depth/bubble measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m depths. (* = marginal effect; a= 0.01). Treatment EQD dia. (mm) Gas flow sig. Depth sig. Interaction n 1.5/0.5 m 4.4 (0.10) Yes* Yes No 120 2.9/0.5 m 4.0 (0.10) 120 2.9/2.7 m 3.5(0.10) 120 Scheffe's test indicates the equivalent bubble diameter (i.e., EQD) for the 2.9/2.7 m treatment was significantly smaller (a < 0.01) than the 1.5/0.5 m treatment combination. There were no significant differences with the other pairwise comparisons. E 8 £ 7 oS 6 5 4 03 E CO b J D 3 CQ c 0) CO > c r LU Depth |m) • « ; • 1.5m/0.5m • 2.9fh/0.5m J • 2 9m/2.7m 10 20 30 Air Flow (L min-1) 40 50 Figure 4.27 Equivalent Bubble Diameter vs. air flow rate with 140 p diffuser at diffuser submergence depth and measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m. 4.3.1.2 400 n diffuser at 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m The results for the analysis of covariance for the 400 p diffuser x 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m treatments are shown in Table 4.59. A significant result (a < 0.01) was obtained for the depth effect (Figure 4.28). Bubble size was smallest for the 1.5/0.5 m combination, then increased in size at the 2.9/0.5 m and 2.9/2.7 m depth combinations. No depth x gas flow rate interaction effect was observed for any of the treatments (Table 4.59). -P 12 § 8 b 03 Si CQ c .2 CO > « 6 -c r LU 2 -Deptn (m) 1.5m/0.5 m 2.9 m/0.5 m 2.9 m/2.7 m 10 20 30 Air Flow (L min-1) 40 50 Figure 4.28 Equivalent Bubble Diameter vs. air flow rate with 400 p diffuser at diffuser submergence depth and measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m. Table 4.59. Adjusted least squares means (± SE), treatment and interaction effects for the 400 p diffuser on air at bubble release depth/bubble measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m depths. Treatment EQD dia. (mm) Gas flow sig. Depth sig. Interaction n 1.5/0.5 m 3.5(0.17) No Yes No 120 2.9/0.5 m 4.1 (0.17) 120 2.9/2.7 m 4.6 (0.17) 120 Scheffe's test indicates the 2.9/2.7 m treatment was significantly larger (a < 0.01) than the 1.5/0.5 m treatment combination for bubble size. There were no significant differences with the other pairwise comparisons. 4.3.1.3 800n diffuser at 1.5/0.5 m, 2.9/0.5 m and2.9/2.7 m The results for the analysis of covariance for the 800 p diffuser x 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m treatments are shown in Table 4.60. A significant result (a < 0.01) was obtained for the depth effect (Figure 4.29). The results were similar to the 400 p treatment combination, i.e., E Q D was smallest for the 1.5/0.5 m combination, then increased in size at the 2.9/0.5 m and 2.9/2.7 m depth combinations. A depth x gas flow rate interaction effect was observed for the treatments 2.9/2.7 m depth combinations, with bubble size tending to decrease with increasing gas flow rate (Figure 4.29, Table 4.60). Table 4.60. Adjusted least squares means (± SE), treatment and interaction effects for the 800 p diffuser on air at bubble release depth/bubble measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m depths. Treatment EQD dia. (mm) Gas flow sig. Depth sig. Interaction n 1.5/0.5 m 3.3 (0.23) No Yes Yes 120 2.9/0.5 m 3.3 (0.23) 120 2.9/2.7 m 4.5 (0.23) 120 E 1 5 Id E b .O xi m c ai LU 10 I • 1 \ 1 • • Depth J(m) • 1.5 m/0.5 m • 2.9j)i/0.5m • 2.9 m/2.7 m -• • i ? 1 i 10 20 30 Air Flow (L min-1) 40 50 Figure 4.29 Equivalent Bubble Diameter vs. air flow rate with 800 p diffuser at diffuser submergence depth and measurement depth combinations of 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m. Scheffe's test was not conducted due to the confounding effect of the gas flow rate by depth interaction. 4.3.2 Orifice Diameter Effect This component of the experiment examined the effect of gas flow variation (10 to 40 1 min"1) and orifice diameter (140 p, 400 p and 800 p), on E Q D , on air, at three combinations of dif-fuser/bubble measurement depth (i.e., 1.5/0.5 m, 2.9/0.5 m and 2.9/2.7 m) in the full lift inlet tube. 4.3.2.1 1.5/0.5 m with 140 ju, 400 n and 800 pi diameter orifice The results of the analysis of covariance for the 1.5/0.5 m combination with the 140 p, 400 p and 800 p diameter orifice treatments are shown in Table 4.61. A significant result (a < 0.01) was obtained for the orifice diameter effect (Figure 4.30). E Q D was smallest for the 800 p E E. S 0) E CD Q JE r> J 2 3 m c _a> co > '3 CT LU 15 10 Orifice (1.5 m>b.5 m) • 140 u • 400 u • 800 u _l_ 10 20 30 Air Flow (L min-1) 40 50 Figure 4.30 Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combination of 1.5/0.5 m with 140, 400 and 800 p diffuser. diameter diffuser, then increased in size for the 400 p and 140 p treatments. No orifice diameter x gas flow rate interaction effect was observed for any of the treatments (Table 4.61). Table 4.61. Adjusted least squares means (± SE), treatment and interaction effects for the bubble release depth/bubble measurement depth combination of 1.5/0.5 m with 140 p, 400 p and 800 p diameter diffuser orifices. Treatment EQD dia. (mm) Gas flow sig. Orifice sig. Interaction n 140 p 4.4 (0.19) No Yes No 120 400 p 3.5(0.19) 120 800 p 3.3 (0.19) 120 Scheffe's test indicates the equivalent bubble diameter for the 400 p and 800 p diameter orifice treatments were significantly smaller (a < 0.01) than the 140 p diameter orifice treatment; however, they were not significantly different from each other. 4.3.2.2 2.9/0.5 m with 140pi, 400 p and800 p, diameter orifice The results of the analysis of covariance for the 2.9/0.5 m combination with the 140 u, 400 p and 800 p diameter orifice treatments are shown in Table 4.62. A significant result (a < 0.01) was obtained for the orifice diameter effect (Figure 4.31). Bubble size was smallest for the 800 p diameter diffuser, then increased in diameter for the 400 p and 140 p treatments. No orifice diameter x gas flow rate interaction effect was observed for any of the treatments (Table 4.62). Scheffe's test indicates the equivalent bubble diameter for the 800 p diameter orifice treatment was significantly smaller (a < 0.01) than the 140 p and 400 p diameter orifice treatments. The 140 (j, and 400 p orifice diameter treatments were not significantly different from each other. Q X) CO > cr LU 12 10 8 6 4 2 0 1 1 Orifice (2.9 m/0.5 m) e 140«u T 400"\j • 800 u * 10 20 30 Air Flow (L min-1) 40 50 Figure 4.31 Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combination of 2.9/0.5 m with 140, 400 and 800 p diffuser. Table 4.62. Adjusted least squares means (± SE), treatment and interaction effects for the bubble release depth/bubble measurement depth combination of 2.9/0.5 m with 140 p, 400 p and 800 p diameter diffuser orifices. Treatment EQD dia. (mm) Gas flow sig. Orifice sig. Interaction n 140 p 4.0 (0.17) No Yes No 120 400 p 4.1 (0.17) 120 800 p 3.3(0.17) 120 4.3.2.3 2.9/2.7 m with 140 [i, 400 JU and 800 fi diameter orifice The results of the analysis of covariance for the 2.9/2.7 m combination with the 140 p, 400 p and 800 p diameter orifice treatments are shown in Table 4.63. A significant result (a < 0.01) was obtained for the orifice diameter effect (Figure 4.32). Bubble size was largest for the 400 p and 800 p diameter diffusers, then decreased in diameter for the 140 p. treatment. No orifice diameter x gas flow rate interaction effect was observed for any of the treatments (Table 4.63). Scheffe's test indicates the equivalent bubble diameter for the 400 p diameter orifice treatment was significantly larger (a < 0.01) than the 140 p diameter orifice treatment. There were no significant differences with the other pairwise comparisons. E CD b CD m c CO > LU 12 10 8 6 4 2 0 I • • 1 1 Orifice (2.9 m/2.7 m) • • • •140 u • • • •400 u I \ ' I \ " ' S O O U - L I I —1 ! 1 1 1 1 1 1 10 20 30 Air Flow (L min-1) 40 50 Figure 4.32 Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combinations of 2.9/2.7 m with 140, 400 and 800 p diffuser. Table 4.63. Adjusted least squares means (± SE), treatment and interaction effects for the air release depth/photo measurement depth combination of 2.9 m/2.7 m with 140 p, 400 p and 800 p diameter diffuser orifices. Treatment EQD dia. (mm) Gas flow sig. Orifice sig. Interaction n 140 p 3.5 (0.16) No Yes No 120 400 p 4.6 (0.16) 120 800 p 4.5 (0.16) 120 4.3.3 Full Lift Modifications Effect This component of the experiment examined the effect of gas flow variation (10 to 40 1 min"1) and in situ modifications (i.e., D B C A , Vanes and Screen), on E Q D , on air, with the 140 u. and 800 p diffuser in the full lift inlet tube. A l l tests were conducted at a bubble release depth of 2.9 m, and measurement depth of 0.5 m. The Surface Cover treatment was not included in the tests. 4.3.3.1 140 p orifice with DBCA, Screen, Vanes and Control The results of the analysis of covariance for the 140 p orifice with the in situ modification treatments are shown in Table 4.64. A significant interaction effect was observed (Figure 4.33 a), hence the data was re-analyzed without the Control treatment which caused the interac-tion effect (Figure 4.33 b). Bubble size was largest for the Vanes treatment, then decreased slightly in diameter for the bubble breaker treatment, then a large decrease for the D B C A treatment (Table 4.64). No in situ modifications x gas flow rate interaction effect was observed for any of the treatments once the Control treatment data was removed from the analysis (Table 4.64). Table 4.64. Mean (± SE) equivalent bubble diameter (mm) for the air release depth/bubble measurement depth combination of 2.9 m/0.5 m with 140 p for the full lift hypolimnetic aerator - full lift in situ modifications effect. Treatment EQD dia. (mm) Gas flow sig. Mods sig. Interaction n Screen 3.9 (0.09) No Yes No 120 DBCA 1.7 (0.09) 120 Vanes 4.4 (0.09) 120 Scheffe's test indicates the equivalent bubble diameter for the D B C A treatment was signifi-cantly smaller (a < 0.01) than the Screens or Vanes treatment. There were no significant differences with the other pairwise comparisons. 20 30 Air Flow (L min-1) 50 Modifications • Screen T DBCA Vanes 20 30 Air Flow (L min-1) 50 Modifications » Screen • DBCA • None Vanes Figure 4.33a-b Equivalent Bubble Diameter vs. air flow rate at diffuser submergence depth and measurement depth combinations of 2.9/0.5 m with 140 p diffuser with in situ modifications (a), and without the Control treatment (b). 4.3.3.2 800 n orifice with DBCA, Screen, Vanes and Control The results of the analysis of covariance for the 800 p orifice with the in situ modification treatments are shown in Table 4.65. A significant interaction effect was observed (Figure 4.34 a), hence the data was re-analyzed in two data sets. The first data set included the D B C A and Vanes treatments (Figure 4.34 b) without the Screen or Control treatments that caused the E E, £ E CD Q Si Si 3 m "c CD > 'a cr LU 20 30 Air Flow (L min-1) 50 Air Flow (L min-1) 0 10 20 30 40 50 Air Flow (L min-1) Figure 4.34a-c Equivalent Bubble Diameter on air at diffuser submergence depth and measurement depth combinations of 2.9/0.5 m with 800 p diffuser and in situ modifications (a), with DBCA and Vanes only (b), and with Control and Screen treatments only (c). interaction effect. The results were similar to the 140 p analysis, i.e., bubble size was largest for the Vanes treatment, then a large decrease for the D B C A treatment (Table 4.64). No in situ modifications x gas flow rate interaction effect was observed for the D B C A and Vanes treat-ment pair, however, the in situ modifications effect and gas flow effect were both significant (Table 4.65). Table 4.65. Mean (± SE) equivalent bubble diameter (mm) for the air release depth/bubble measurement depth combination of 2.9 m/0.5 m with 800 p for the full lift hypolimnetic aerator - full lift in situ modifications effect: DBCA and vanes. Treatment EQD dia. (mm) Gas flow sig. Mods sig. Interaction n DBCA 1.9 (0.16) Yes Yes No 120 Vanes 4.8(0.16) 120 The second data set included the Screen and Control treatments (Figure 4.34 c) without the D B C A or Vanes treatments. The results indicate EQD was significantly larger for the Screens treatment (Table 4.66). No in situ modifications x gas flow rate interaction effect was observed for the Screen and Control treatment pair, however, the in situ modifications effect was signifi-cant (Table 4.66). Table 4.66. Mean (± SE) equivalent bubble diameter (mm) for the air release depth/bubble measurement depth combination of 2.9 m/0.5 m with 800 p for the full lift hypolimnetic aerator - full lift in situ modifications effect: Screen and Control. Treatment EQD dia. (mm) Gas flow sig. Mods sig. Interaction n Screen 3.7 (0.17) No Yes No 120 Control 3.3 (0.17) 120 4.4 Full Lift Hypolimnetic Aeration - Oxygen Transfer Locations As indicated in the Methods (3.3.2), the six oxygen probe locations for the full lift aerator were located at the top, middle and bottom of the inlet and outlet tubes. This placement was designed to identify five zones of oxygen transfer: Zone 1 - between the entrance (Probe 1) and midpoint of the inlet tube (Probe 2); Zone 2 - between the midpoint (Probe 2) and top of the inlet tube at the entrance to the separator box (Probe 3); Zone 3 - between the entrance (Probe 3) and exit (Probe 4) of the separator box; Zone 4 - between the top of the outlet tube at the exit of the separator box (Probe 4) and midpoint of the outlet tube (Probe 5); and Zone 5 - between the midpoint (Probe 5) and exit of the outlet tube (Probe 6) (see Figure 3.3). The data for each aeration zone was obtained by calculating the differential oxygen concentration from the six probes at each time stamp during the particular test, thus creating a series of graphs showing five lines of differential oxygen concentrations for the duration of each test. The data is pre-sented qualitatively, as the net result of the various experimental treatments has already been quantitatively discussed for K i ^ o , SOTR, S A E and SOTE. In order to minimize the number of graphs presented, only one graph per treatment combination is shown, although the trends were similar for each of the three replicates per treatment. The rationale for presenting this data is to identify patterns or zones of oxygen transfer within the full lift aerator as a function of specific treatments, which document the performance of the current system, and provide insight into future design modifications to potentially improve oxygen transfer. 4.4.1 140 u.x 2.9 m x air The oxygen differential was greatest in Zone 1, followed by Zones 2 and Zone 3 (Figure 4.35 a-d). The oxygen differential was slightly negative in Zones 4 and 5; indicating oxygen concen-trations were decreasing in these zones. A slight downward trend was observed over the I E c i I o 1.0 0.8 06 0 4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 Difference Difference ' " " Difference Difference " " " " " Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 12 16 20 Time (min) 24 28 32 36 -1.0 Concentration reading differences between two adjacent probes 0.8 " Difference Difference • • Difference — Difference Difference CH2-CHI CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 i i 1 1 1 1 r -12 16 n 1 R~ ~I 1 1 r -24 -i r~ 28 20 Time (min) 1.0 T 0.8 -; 0.6 •; O -0.4 •; -0.6 -; Difference Difference • • Difference ™—™" Difference mm"***Km Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -1.0 H 1 1 1 1 1 i 1 1 1 1 1 1 1 1 1 1 « 1 1 1 — 4 8 12 16 20 Time (min) Concentration reading differences between two adjacent probes O -0.4 •; -0.6 •• Difference Difference Difference " " Difference ~" Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 .1.0 T 1 1 1 — — — i 1 1 1 1 1 1 1 1 1 1 r 4 8 12 duration of the test in Zones 1 and 2, whereas Zones 3, 4 and 5 were fairly constant. The general trend with increasing gas flows from 10 to 40 1 min"1 was an increase in the oxygen differential in Zone 1, and an increase in slope for Zones 2 and 3. 4.4.2 140 p x 1.5 m x a i r The oxygen differential was greatest in Zone 2, followed by Zone 3 (Figure 4.36 a-d). The oxygen differential was near zero in Zone 1, and slightly negative in Zones 4 and 5, indicating oxygen concentrations were decreasing in these zones. A downward trend was observed over the duration of the test in Zones 2, whereas Zones 1, 3, 4 and 5 were fairly constant. The general trend with increasing gas flows was an increase in the oxygen differential and slope in Zone 2, and a greater loss of oxygen in Zone 5. 4.4.3 400 u x 2.9 m x air The oxygen differential was greatest in Zone 1, followed by Zones 2 and Zone 3 (Figure 4.37 a-d). The oxygen differential was slightly negative in Zones 4 and 5; indicating oxygen concen-trations were decreasing in these zones. No downward trend was observed over the duration of the test in Zones 1-5. The general trend with increasing gas flows was an increase in the oxygen differential and slope in Zones 1 and 2, and an increased loss of oxygen in Zone 5. 4.4.4 400 p x 1.5 m x a i r The oxygen differential was greatest in Zone 2, followed by Zone 3 (Figure 4.38 a-d). The oxygen differential was near zero in Zone 1, and slightly negative in Zones 4 and 5. A down-ward trend was observed over the duration of the test in Zone 2, whereas Zones 1,3,4 and 5 o.o-j -0.2 •; -0.4 • : -0.6 • : -1.0 0.8 - I " " Difference Difference """ **""*",SH! Difference Difference Difference ' 1 CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 - î — I — I — I — I — I — I — i — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I 1— i 1 — i — I 1—I—r-8 12 16 20 24 28 32 36 Time (min) Concentration reading differences between two adjacent probes Difference Difference Difference Difference ^"""""Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 — i — î — i — i — i — i — i 1 — i — T — i — i — i — i — i — i — i — i — I — I — I — I — i — I — I — I — I — I 1 — I — i — i i 1—I 1—r-4 8 12 16 20 24 28 32 36 Time (min) Concentration reading differences between two adjacent probes -l.o 0.8- Difference Difference ' * " Difference — Difference m ~ ~ m Difference '• , , CH2-CH1 Cl I3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 t — i — — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r 12 16 20 24 28 32 36 Time (min) Concentration reading differences between two adjacent probes ù c 4t M >. 0 1 0 0.8 0.6 04 0.2 0.0 -02 -0 4 -0.6 -0.8 -1.0 Difference Difference 1 " * Difference ~ Difference >m.m,.mm difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 i t i 1 1 1 1 1 1 1 1 r 4 8 -i 1 r- - l 1 1 1 r-20 12 16 -0.6 -0.8 -1.0 " Difference CH2-CH1 "Difference - - - D j f f e r e n c e CH3-CH2 CH4-CH3 Difference CH5-CH4 " Difference CH6-CH5 - i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i -4 8 12 16 20 24 28 32 36 Time (min) Concentration reading differences between two adjacent probes Difference Difference " " Difference ~ Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -1.0 T — I — I — i — I — I — I — I — I — I — I — 1 — I — I — I — I — I — I — I — I — I — I — i — I — I — I — I — I — I — I 1 — i — I — r -4 8 12 16 20 24 28 32 Time (min) 36 1.0 i — ———— 0.8 "I ' ' • r , , , , , r , 0.6 - •> - •- •- < r -, , -, - .- r , , , r -, - ,- ,- , -, - ,- r -, O -OA-; -0.6 • : 0.8- Difference Difference Difference Difference "m m ~ m Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -1.0 T — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — I — i — I — I — I — I — I — I — i — i — i — I — I 1 1 — I — I — I — I — I — I — r 4 8 12 16 20 24 28 32 36 Time (min) were fairly constant. The general trend with increasing gas flows was an increase in the oxygen differential and slope in Zone 2. 4.4.5 800 u x 2.9 m x air The oxygen differentials were similar in Zones 1, 2 and 3, and similar, but negative, in Zones 4 and 5 (Figure 4.39 a-d). As gas flow increased, the differentials become more distinct, with Zone 1 being the highest, followed by Zones 2 and 3; while the Zone 5 differential became increasingly negative. 4.4.6 800 u x 1.5 m x air The oxygen differential was greatest in Zone 2, followed by Zone 3 (Figure 4.40 a-d). The oxygen differential was near zero in Zone 1, and slightly negative in Zones 4 and 5. A slight downward trend was observed over the duration of the test in Zone 2, whereas Zones 1,3,4 and 5 were fairly constant. The general trend with increasing gas flows was an increase in the oxygen differential and slope in Zone 2, an increase in the differential in Zone 3, and an increase in the negative differential in Zone 4. A summary of the qualitative observations of the oxygen differentials, on air, are shown in Table 4.67. 4.4.7 140 f ix 2.9 m xO? The oxygen differential was greatest in Zone 1, followed by Zones 2 and Zone 3 (Figure 4.41 a-d). The oxygen differential was slightly negative in Zones 4 and 5, indicating oxygen concen-trations were decreasing in these zones. A slight downward trend was observed over the duration of the test in Zones 1 and 2, whereas Zones 3, 4 and 5 were fairly constant. The general trend with increasing oxygen flows from 10 to 40 1 min"1 was an increase in the oxygen differential in Zone 1 with no clear trend in slope. 1.0 i 0.8 •[ -. -.- ,- . -. - .- r , , -, - ,- r , , , -, - ,- r -, 0.6 •; •• i -. - .- r -, - ,- r -, - ,- r , -, - ,- ? , , , -, - ,- , -, I M - ; s Difference Difference • " Difference Difference " " " " Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 i i i i i i i 4 8 i i i i i i i 12 16 1 1 1 1 1 1 20 I 1 1 1 1 1 1 24 28 i i i r i i 32 36 Time (min) Concentration reading differences between two adjacent probes l.o i 0.8-; ï - .- r r r r i r -, -.- r , - -, r -, 0 . 6 - '• -, -•- r ' r -, -, - ,- i -, - ,- r i -• - ,- i -, - ,- i -, - ,- f -, - ,- ,- i -, - ,- r -, -0.6 - : •««•««. Q i f f e r e i l c e Difference Difference " " Difference *~ Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -1.0 " 1 — i — i — i — i — ' — i — i — ' — i — i — i — i — i — i — i — 1 — ' — 1 — 1 — 1 — 1 — ' — 1 — 1 — ' — ' — 1 — 1 — 1 — 1 — 1 — 1 — 1 — • — ' — 1 — ' — r 4 8 12 16 20 24 28 32 36 Time (min) o.8 • ; -• - • - - ' -• - • - <• •• - • - •• - - ' -• - • - <• - ' - r -, -, - ,- , 0.6-| -, f -, -,- , , -, -,- r r -, r -, . -, -,-O -0.4 -; -0.6 •; Difference Difference • " Difference Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -1.0 T — I — I — I — I — I — I — I — I — I — I — I — I — I — I — p — I — I — i — I — I — I — I — I — I — I — I — i — i — I 1 1 — I — I — I— I 1 — I — r 4 8 12 16 20 24 28 32 36 Time (min) s l.o-0.8-0.6 • 0.4 • 0.2 • 0.0 • -0.2 • -0.4 • -0.6--0.8--1.0-Difference Difference • * Difference Difference *"* mm* j ) j f f e r e n c e CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 - I— I— I— I— I— I— I— I— I— I— I— I— I— I— I— I— I— I— I 1 1—I—I 1—I 1 — I — r -12 16 20 24 28 32 36 Time (min) Concentration reading differences between two adjacent probes 3 g = o •a b l.o • 0.8 -0.6 -0.4 -0.2 • 0.0 --0.2 --0.4 " -0.6--0.8--1.0 • Difference Difference • " Difference Difference .««M**** Q [ f f e r e n c e CH2-CHI CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r -4 8 12 16 20 24 28 32 36 Time (min) Time (min) Concentration reading differences between two adjacent probes -0.2 V sr. >. H O -0.4 -0.6 -0.8 -1.0 Difference Difference • - Difference Difference m m m m Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 1 — I 1 I T T I 1 I I I I 1 1 1 I I 1 12 16 20 24 28 i i i i — r — r ~ 4 32 36 Time (min) Concentration reading differences between two adjacent probes Difference Difference • " Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -2.0-Time (min) I c o •a s Time (min) Concentration reading differences between two adjacent probes Difference Difference ' - - Difference Difference ™ *"*mtm" Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 Table 4.67. Summary of qualitative observations of the relative magnitude, slope and response to increasing air flows for oxygen differentials in Zones 1-5. Treat-ment Zone 1 Ch. 2-1 Zone 2 Ch. 3-2 Zone 3 Ch. 4-3 Zone 4 Ch. 5-4 Zone 5 Ch. 6-5 10^ 40 1 min"1 140 p x 2.9 m x air Highest differential, Negative slope 2 n d highest differential, less slope 3 rd highest differential, less slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1 differential and slope, increase in slope for Zones 2 and 3 140 p x 1.5 m x air Zero differential, neutral slope Highest differential, Negative slope 2n d highest differential, less slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 2 differential and slope, change from neutral to negative slope in Zone 5 400 p x 2.9 m x air Highest differential, Negative slope 2 n d highest differential, less slope 3 rd highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1 and 2 differen-tial and slope, change from neutral to negative slope in Zone 5 400 p x 1.5 m x air Zero differential, neutral slope Highest differential, Negative slope 2 n d highest differential, less slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 2 differential and slope 800 p x 2.9 m x air Similar differential, neutral slope Similar differential, neutral slope Similar differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1 and 2 differen-tial and slope, change from neutral to negative slope in Zone 5 800 p x 1.5 m x air Zero differential, neutral slope Highest differential, Negative slope 2 n d highest differential, less slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 2 differential and slope, increase in Zone 3 differen-tial, and change from neutral to negative slope in Zone 4 4.4.8 140 p x 1.5 m x O ? The oxygen differential was greatest in Zone 2, followed by Zone 3 (Figure 4.42 a-d). The oxygen differential was near zero in Zone 1, and slightly negative in Zones 4 and 5. A minimal slope was observed over the duration of the test in all zones. The general trend with increasing oxygen flows was an increase in the oxygen differential in Zone 2. 4.4.9 400 u x 2 . 9 m x O > The oxygen differential was greatest in Zones 1 and 2, followed by Zone 3 (Figure 4.43 a-d). The oxygen differential was slightly negative in Zones 4 and 5. Minimal change in slope was observed over the duration of the test in all zones. The general trend with increasing oxygen flows was an increase in the oxygen differential in Zones 1, 2, and 3. 4.4.10 400 u x 1 . 5 m x O ? The oxygen differential was greatest in Zone 2, followed by Zone 3 (Figure 4.44 a-d). The oxygen differential was near zero in Zone 1, and slightly negative in Zones 4 and 5. A slight downward trend was observed over the duration of the test in Zone 2, whereas Zones 1,3,4 and 5 were fairly constant. The general trend with increasing oxygen flows was an increase in the oxygen differential in Zones 2 and 3. Figure 4.42a-d Oxygen differentials in full lift aerator on PSA O 2 at 1.5 m with 140 diffuser at 10 1 min 1 (a), 20 1 min"1 (b), 30 1 min 1 (c), 40 1 min"1 (d). -2.0 Difference Difference * " Difference Difference m m m m m Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 Time (min) Concentration reading differences between two adjacent probes 4 0 • 3 0 • I 2.0-C S l.o • o o.o •• --1.0 •• -2.0-Difference Difference - • Difference Difference " " " " Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 Time (min) Concentration reading differences between two adjacent probes Difference Difference Difference "™ Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -2.0 • Time (min) -2.0 Difference Difference Difference Difference Difference CH2-CH1 cm-cm CH4-CH3 CH5-CH4 CH6-CH5 Time (min) Concentration reading differences between two adjacent probes 40 30 I 2.0 § 1.0 S B 5 U • M 6? o.o o -1.0 ""***""" Difference Difference Difference " • Difference *~ Difference CH2-CH1 cm-cm CH4-CH3 CH5-CH4 CH6-CH5 Time (min) Concentration reading differences between two adjacent probes Difference Difference Difference Difference " r m n * " Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 10 12 Time (min) Figure 4.44a-d Oxygen differentials in full lift aerator on PSA O 2 at 1.5 m with 400 \i diffuser at 10 1 min"1 (a), 20 I min"1 (b),30 1 min"1 (c),40 1 min 1 (d) 40 1 min"1. -2.0 Difference Difference • " Difference m Difference """ Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 Time (min) Concentration reading differences between two adjacent probes Difference Difference " - Difference "~ Difference ' " " Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 4.4.11 800 u x 2.9 m x O? The oxygen differentials were highest in Zones 1 and 2, and next highest in Zone 3. Zones 4 and 5 were negative through the tests (Figure 4.45 a-d). As gas flow increased, the separation between the zone differentials become more distinct, with Zone 1 being the highest, followed by Zones 2 and 3; while Zone 4 and 5 differentials remained negative with a neutral slope. 4.4.12 800 u x l . 5 m x O ? The oxygen differential was greatest in Zone 2, followed by Zone 3 (Figure 4.46 a-d). The oxygen differential was near zero in Zone 1, and slightly negative in Zones 4 and 5. A slight downward trend was observed over the duration of the test in Zone 2, whereas Zones 1, 3, 4 and 5 were fairly constant. The general trend with increasing oxygen flows was an increase in the oxygen differential and slope in Zones 2 and 3, and a minimal change in slope for all zones. A summary of the qualitative observations of the oxygen differentials, on PSA oxygen, are shown in Table 4.68. 4.4.13 D B C A x 140 u x air The oxygen differential was greatest in Zone 1, followed by a tight grouping of Zones 2, 3, and slightly negative in Zone 5 (Figure 4.47 a-d). Minimal trend was observed over the duration of the test in all zones. The general trend with increasing oxygen flows was a slight increase in slope in Zones 1, 2 and 3, and an increase in the negative oxygen differential in Zone 4. Difference Difference " * Difference Difference """" Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 H 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 h-2 4 6 8 10 12 14 16 \i Time (min) Concentration reading differences between two adjacent probes • Difference CH2-CH1 "Difference CH3-CH2 Difference CH4-CH3 "Difference CH5-CH4 "Difference CH6-CH5 10 Time (min) 4.0 T 1 ' ' • ' ' • 1 1 3.0 i • -1.0--Difference CH2-CH1 Difference CH3-CH2 • - Difference CH4-CH3 Difference CH5-CH4 *~™™*"f Dj fference CH6-CH5 i i 2 1 i i 4 6 1 1 X Time (min) Concentration reading differences between two adjacent probes 4.0 T ' ' ' ' 3.0 -• -1.0--Difference Difference • " Difference ™ Difference '**"" """Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -2.0 4 i 1 i i i-2 4 Time (min) 4 0 3.0 I 2.0 1.0 0.0 -1.0 -2.0 Difference Difference Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 10 12 Time (min) Concentration reading differences between two adjacent probes Difference Difference • • Difference Difference m m m m m Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 10 Table 4.68. Summary of qualitative observations of the relative magnitude, slope and response to increasing PSA oxygen flows for oxygen differentials in Zones 1-5. Treatment Zone 1 Ch. 2-1 Zone 2 Ch. 3-2 Zone 3 Ch. 4-3 Zone 4 Ch. 5-4 Zone 5 Ch. 6-5 10^ 40 1 min - 1 140 p x 2.9 m x O2 Highest differential, Negative slope 2 n d highest differential, less slope 3 r d highest differential, less slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1 differential and slope, variable slope 140 p x 1.5 m x O2 Zero differential, neutral slope Highest differential, Negative slope 2 n d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 2 differential and slope 400 p x 2.9 m x O2 Highest differential, Neutral slope Highest differential, neutral slope 2 n d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1, 2and 3 differential 400 p x 1.5 m x O2 Zero differential, neutral slope Highest differential, Minimal slope 2 n d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 2 and 3 differential 800 p x 2.9 m x O2 Highest differential, neutral slope Similar differential, neutral slope Lower differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1, 2 and 3 differential 800 p x 1.5 m x 0 2 Zero differential, neutral slope Highest differential, Neutral slope 2 n d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 2 and 3 differential 4.4.14 D B C A x 800 u x air The oxygen differentials with this treatment combination changed with each increase in gas flow rate. Initially, the oxygen differential was greatest in Zones 3 and 4, followed by Zone 1, Zone 2 and Zone 5, which was positive (Figure 4.48 a-d), with all zones exhibiting a slight positive slope over the duration of the test. At the 17/3 1 min"1 treatment, Zones 1, 2 and 3 were grouped together in a neutral slope, and Zones 4 and 5 behaved similarly. For the 27/3 min"1 treatment, Zones 1, 2 and 3 remained grouped, but the oxygen differential in Zone 5 increased substantially, while Zone 4 became increasingly negative. The pattern of the oxygen differen-tials were similar at the highest flows of 37/3 min"1. i.o î , , 0.8 • ; - - • - > ' - - • - •• • <• - -•- •- ' -• -•• •- - ' -• • • - •- - - i . 0.6 1 r . -, - .- ,- ! -, - ,- , , -, - ,- , , -, - ,- r -. -, - .- 1 H O -0.4 •; -0.6 •; Difference Difference • " Difference Difference '"•"'*""' Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 4 8 12 16 20 24 28 32 36 Time (min) Concentration reading differences between two adjacent probes i.o î — — 1 —— 0.8 ,-0.6 •; -O -0.4 •• -0.6 • : Difference Difference • - Difference Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 4 8 12 16 20 24 28 32 36 4.4.15 Screen x 140 u. x air The oxygen differential was greatest in Zone 1, followed by Zones 2 and 3, and neutral to negative in Zone 5, and negative in Zone 4 (Figure 4.49 a-d). A downward trend was observed over the duration of the test in Zones 1, 2 and 3, whereas Zones 4 and 5 were fairly constant. The general trend with increasing gas flows was an increase in the oxygen differential and slope in Zone 2 and 3, and a decrease in oxygen differential for Zone 5. 4.4.16 Screen x 800 u. x air The oxygen differential was greatest in Zones 1 and 2, followed by Zones 3, and negative in Zones 4 and 5 (Figure 4.50 a-d). Minimal downward trend was observed over the duration of the test in Zones 1, 2 and 3, except at the highest gas flow rates, whereas Zones 4 and 5 exhib-ited a negative differential with neutral slope. The general trend with increasing gas flows was an increase in the oxygen differential and slope in Zones 1 and 2. 4.4.17 Surface cover x 140 p x air The oxygen differential was greatest in Zone 1, followed by Zones 2 and Zone 3 (Figure 4.51 a-d). The oxygen differential was slightly negative in Zones 4 and 5; indicating oxygen concen-trations were decreasing in these zones. A slight downward trend was observed over the duration of the test in Zones 1 and 2, whereas Zones 3, 4 and 5 were fairly constant. The general trend with increasing gas flows from 10 to 40 1 min"1 was an increase in the oxygen differential in Zone 1, and an increase in slope for Zones 2 and 3. 16 20 Time (min) Concentration reading differences between two adjacent probes l i § Concentration reading differences between two adjacent probes -0.6 -0.8 -1.0 Difference Difference " " Difference Difference i w w a w Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 I I — 1 — 1 — I — I — I — I — 1 — 1 — I — I — I — I — I — 1 — I — 1 — I 1 8 12 16 20 24 28 32 -1.0 0.8- Difference Difference Difference ~~ Difference * " ™ Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 1 — I — I — I — I — I — I — I — I — I — I — I — I — I — I 24 28 32 12 16 20 Time (min) 36 Concentration reading differences between two adjacent probes 16 20 Time (min) Concentration reading differences between two adjacent probes 0.8- Difference Difference • " Difference Difference m m m m Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 1 0 I I I I — I i — i — i — i l l — i — i — r — i — i — i — i — i — r — ! — i — i — i — i — i — i i — i — i 1 1 1— i 1—i 1 — i — i — 12 16 20 Time (min) 24 28 32 36 1 E o 12 16 20 Time (min) 24 28 32 36 Concentration reading differences between two adjacent probes -0.2 « >. M O -0.4 -0.6 -0. -1.0 Difference Difference • " Difference Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 T — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — I — I — i — r 4 8 12 16 20 24 28 32 36 40 44 48 52 Time (min) Figure 4.51a-d Oxygen differentials in full lift aerator on air at 2.9 m with 140 \i : _ - l diffuser and Surface cover at 10 1 min - (a), 20 1 min" (b),30 1 min 40 I min 1 (d). 16 20 Time (min) Concentration reading differences between two adjacent probes -i.o 0.8- Difference Difference • • Difference Difference """Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -i 1 1 1 — i — r -32 36 4.4.18 Surface cover x 800 u. x air The oxygen differentials were similar in Zones 1, 2 and 3, and similar, but negative, in Zones 4 and 5 (Figure 4.52 a-d). As gas flow increased, the differentials in Zones 1, 2 and 3 increased but remained tightly grouped; while the Zone 5 differential became less negative. 4.4.19 Vanes x 140 u,x air The oxygen differential was greatest in Zone 1, followed by Zones 2 and Zone 3, and slightly negative in Zones 4 and 5 (Figure 4.53 a-d). A slight downward trend was observed over the duration of the test in Zones 1 and 2, whereas Zones 3, 4 and 5 were fairly constant. No trends were apparent with increasing gas flows. 4.4.20 Vanes x 800 p x air The oxygen differentials were similar in Zones 1, 2 with Zone 3 being lower, and similar, but negative, in Zones 4 and 5 (Figure 4.54 a-d). As gas flow increased, the differentials and negative slope in Zones 1 and 2 increased but remained tightly grouped; while Zones 4 and 5 remained negative with minimal change in slope. A summary of the qualitative observations of the in situ modifications oxygen differentials, on air, are shown in Table 4.69. 4.4.21 D B C A x 140 u x O? The oxygen differential was greatest in Zone 1, followed by Zone 2 and a tight grouping of Zones 3, and 4, and slightly negative in Zone 5 (Figure 4.55 a-d). Minimal trend was observed over the duration of the test in all zones. There was minimal change with increasing oxygen flows in Zones 1,2, 3 and 5, while the Zone 4 oxygen differential decreased. 1 E B -0.2 V CD ft O -0.4 -0.6 -0.8 Difference Difference " Difference Difference mmam Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -1.0 H — i — i — i — i — i — i — ' — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i 4 8 12 16 20 24 Time (min) 28 32 36 Concentration reading differences between two adjacent probes -0.6 -0.8 -1.0 Difference Difference ' - " Difference Difference '" "HM"*'""" ""Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 Figure 4.53a-d Oxygen differentials in full lift aerator on air at 2.9 m with 140 p diffuser and Counter-rotating blades at 10 1 min"1 (a), 20 1 min"1 (b), 301min"1 (c), 401 min"1 (d). -t .0 '—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i— 4 8 12 16 20 24 28 32 36 Time (min) Concentration reading differences between two adjacent probes O -0.4 -; --0.6 -Difference Difference • " Difference Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 1.0 î • — — - — ' — ' — 0.8 - ; • • " " i - . - r , -, - , - r , - , - ,- , , - , - ,- , , - , - ,- r , -, - ,-0.6 " : •- 1 -, - . - f 1 - , - .- . -, - , - r , , - , - ,- r , -, - ,- r , -0.6 • : 0.8- Difference Difference ' • - Difference Difference ,m*mm D j f f e r e n c e CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -1.0 T — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — i — r — i — i — i — i — i — i — i — i — i — r — i — i — i — 4 8 12 16 20 24 28 32 36 Time (min) Figure 4.54a-d Oxygen differentials in full lift aerator on air at 2.9 m with 800 \i diffuser and Counter-rotating blades at 10 1 min"1 (a), 20 1 min 1 (b), 301min"1 (c), 401 min"1 (d). 1 E •a L. E 16 20 Time (min) Concentration reading differences between two adjacent probes E c 5 U Ci 0 " V x - W - W V . ' - - N v V W C ^ - v V ' ' V \ / V • - Difference CH4-CH3 Difference CH5-CH4 """""""" Difference CH6-CH5 Difference CH2-CH1 Difference CH3-CH2 4 8 12 16 20 24 28 32 36 4.0 ' 3.0 • I 2.0-••C S c it u c o U 1.0 -• t? o.o-o -1.0--2.0-10 12 14 16 Time (min) Concentration reading differences between two adjacent probes Time (min) Time (min) Concentration reading differences between two adjacent probes Difference CH4-CH3 'Difference CH5-CH4 ""Difference CH6-CH5 Table 4.69. Summary of qualitative observations of the relative magnitude, slope and response to increasing air flows for oxygen differentials in Zones 1-5 for in situ modifications. Treatment Zone 1 Ch. 2-1 Zone 2 Ch. 3-2 Zone 3 Ch. 4-3 Zone 4 Ch. 5-4 Zone 5 Ch. 6-5 10^ 40 1 min 1 D B C A x 140 p x air Highest differential, Neutral slope 2 n d highest differential, neutral slope 2 n d highest differential, neutral slope 2 n d highest differential, neutral slope Negative differential, neutral slope Increase in slope for Zones 1, 2 and 3, increase in negative differential in Zone 4 D B C A x 800 p x air Highest differential, Neutral slope Highest differential, neutral slope Highest differential, neutral slope Variable differential and slope Variable differential and slope Increase in differen-tial for Zones 1 and 5,increase in negative differential in Zone 4 Screen x 140 p x air Highest differential, Negative slope 2 n d highest differential, negative slope 3rd highest differential, negative slope Negative differential, neutral slope Neutral differential and slope Increase in differen-tial and slope for Zones 2 and 3, decrease in differen-tial in Zone 5 Screen x 800 p x air Highest differential, Neutral slope Highest differential, neutral slope 2 n d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in oxygen differential and slope for Zones 1 and 2 Surface cover x 140 p x air Highest differential, Negative slope 2 n d highest differential, less slope 3 rd highest differential, less slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1 differential and slope, increase in slope for Zones 2 and 3 Surface cover x 800 p x air Similar differential, neutral slope Similar differential, neutral slope Similar differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1, 2 and 3 differential slope, Zone 5 differential becomes neutral Vanes x 140 p x air Highest differential, Minimal slope 2 n d highest differential, less slope 3 r d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Minimal change Vanes x 800 p x air Highest differential, Minimal slope Highest differential, less slope 2 n d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in differen-tial and negative slope in Zones 1 and 2. 4.4.22 D B C A x 800 u x Q 7 The oxygen differentials with this treatment combination changed with each increase in gas flow rate. Initially, the oxygen differential was greatest in Zone 3, followed by Zones 4, 1, 2 and 5, all of which were positive (Figure 4.56 a-d), with all zones exhibiting minimal slope over the duration of the test. At the 17/3 1 min"1 treatment, Zones 1, 2 and 3 were grouped unrerence Difference Difference Difference Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 10 12 Time (min) Concentration reading differences between two adjacent probes 3.0 Difference Difference ' • - Difference Difference " m m m m Di fference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 10 together with a neutral slope, and Zones 4 and 5 behaved similarly, and remained positive. For the 27/3 min"1 treatment, Zones 1, 2 and 3 remained grouped, but the oxygen differential in Zone 5 decreased. The pattern of the oxygen differentials were similar at the highest flows of 37/3 min"1 with the exception of increased oxygen differentials in Zone 2. 4.4.23 Screen x 140 u x O ? The oxygen differential was greatest in Zones 1 and 2, followed by 3, and negative in Zones 4 and 5 (Figure 4.57 a-d). Minimal change in slope was observed over the duration of the test in all zones. There were minimal changes in the general pattern as gas flows increased. 4.4.24 Screen x 800 u x O? The oxygen differential was greatest in Zones 2, followed by Zonel and 3, and negative in Zones 4 and 5 (Figure 4.58 a-d). Minimal downward trend was observed over the duration of the test in all zones. The general trend with increasing gas flows was an increase in the oxygen differential in Zones 2 and 1. 4.4.25 Surface cover x 140 p x O? The oxygen differential was greatest in Zone 1, followed by Zones 2 and Zone 3 (Figure 4.59 a-d). The oxygen differential was slightly negative in Zones 4 and 5; indicating oxygen concen-trations were decreasing in these zones. Minimal change in slope was observed over the duration of the test in all zones 1. The general trend with increasing gas flows was an increase in the oxygen differential in Zone 1. Difference Difference • - Difference ~ ~ Difference *"-***"* Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -2.0-Time (min) Concentration reading differences between two adjacent probes Difference Difference ' • • Difference Difference "*"* Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 -2.0 • H— I—I—h H—I—I I I I I—I—I I I I I I I I—I—I—I—h -f—I- H—I—I—I—I—I I I 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 Time (min) -2.0-Concentration reading differences between two adjacent probes Difference Difference Difference Difference " Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 •+- •+- -+-8 10 Time (min) 12 14 16 10 12 14 Time (min) Concentration reading differences between two adjacent probes - - - -. It ' / ' / ' ' ' \ l ' A " 1.0" Difference Difference • ' Difference Difference m Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 10 5 o.o--2.0-Difference Difference • ' Difference Difference t m m i m m m w m r j j f j e r e n c e CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 M : 2 1 1 4 1 6 8 1 1 10 4.0 • 3.0 • Time (min) Concentration reading differences between two adjacent probes I 2.0-10 • 0 o.o • -1.0 • Difference Difference " " Difference "" ' Di ffer ence *"'™"""™ Di fference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 4.4.26 Surface cover x 800 p x O? The oxygen differentials were similar in Zones land 2, followed by Zone 3, and similar, but negative, in Zones 4 and 5 (Figure 4.60 a-d). As gas flow increased, the differentials in Zones 1 and 2 increased but remained tightly grouped. 4.4.27 Vanes x 140 u x O ? Initially, at the lowest gas flow rates, the oxygen differential was highest in Zone 2. However, in all other flow rates, the oxygen differential was greatest in Zone 1, followed by Zones 2 and Zone 3, and slightly negative in Zones 4 and 5 (Figure 4.61 a-d). Minimal change in slope was observed over the duration of the test in all zones, and no major trends were apparent with increasing gas flows, other than the above noted change as gas flow rates increased above 10 1 min"1. 4.4.28 Vanes x 800 u x O? The oxygen differentials were similar in Zones 1, 2 with Zone 3 being lower, while Zones 4 and 5 were similar, but negative (Figure 4.62 a-d). As gas flow increased, the differentials and neutral slope in Zones 1 and 2 remained similar and tightly grouped; while Zones 4 and 5 remained negative with minimal change in slope. A summary of the qualitative observations of the in situ modifications oxygen differentials, on PSA oxygen, are shown in Table 4.70. 4.5 Speece Cone Hypolimnetic Aeration - Oxygen Transfer 4.5.1 P S A O? Experiments The results for the analysis of covariance for the Speece Cone hypolimnetic aerator are shown in Table 4.70. Unlike the full lift experiments, the Speece Cone was operated at a single depth on 4.0 I ' — , , , , — — 3.0 •• -i.o--Difference Difference • • Difference Difference " " " " " Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 4 6 8 10 12 14 16 8 20 22 24 26 28 30 32 34 36 Time (min) Difference Difference • • • Difference Difference ~ m m m m m Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 i i i 2 4 1 1 6 1 1 8 1 1 1 1 10 12 1 1 14 Time (min) Concentration reading differences between two adjacent probes -1.0" 10 12 '•"***"""" Difference Difference Difference * " ' Difference ~ Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 H 1 1 1 1 1 1 1 1 1 1 h H H 10 12 14 Time (min) Concentration reading differences between two adjacent probes 3.0-4.0 3.0 I 2 0 = -2.0 Difference Difference • " Difference Difference Di fference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 I I I I I I I I I I I I I I I I I I I I I I I I I I I—I—I—I—I—I—I—I—I—I I I 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 Time (min) Concentration reading differences between two adjacent probes -2.0 • - Difference CH4-CH3 Difference CH5-CH4 M 1 1 . 1 1 1 Difference CH2-CH1 Difference CH3-CH2 Difference CH6-CH5 • • i • i • • i i i • • i i • i i • i i i 2 4 • i i 6 8 1 1 1 10 1 1 1 1 12 14 i i i 16 18 Time (min) Difference Difference " * * Difference Difference """"*"*"• Difference CH2-CH1 CH3-CH2 CH4-CH3 CH5-CH4 CH6-CH5 12 Time (min) Concentration reading differences between two adjacent probes Table 4.70. Summary of qualitative observations of the relative magnitude, slope and response to increasing PSA O 2 flows for oxygen differentials in Zones 1-5 for in situ modifications. Treatment Zone 1 Ch. 2-1 Zone 2 Ch. 3-2 Zone 3 Ch. 4-3 Zone 4 Ch. 5-4 Zone 5 Ch. 6-5 10^ 40 1 min"1 D B C A x 140 p x 0 2 Highest differential, Neutral slope 2 n d highest differential, neutral slope 3 r d highest differential, neutral slope Positive differential, neutral slope Negative differential, neutral slope Increase in slope for Zones 1, 2 and 3, decrease in positive differential in Zone 4 D B C A x 800 p x 0 2 Highest differential, Neutral slope Highest differential, neutral slope Highest differential, neutral slope Positive differential, neutral slope Positive differential, neutral slope Minima! change, increase in differential for Zones 2 Screen x 1 4 0 p x O 2 Highest differential, Neutral slope Highest differential, neutral slope 2 n d highest differential, neutral slope Negative differential, neutral slope Neutral differential, neutral slope Minimal change Screen x 800 p x 0 2 2 n d highest differential, Neutral slope Highest differential, neutral slope 3 r d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in oxygen differential and slope for Zones 2 and 1 Surface cover x140 p x 0 2 Highest differential, neutral slope 2 n d highest differential, Neutral slope 3 r d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1 differential Surface cover x 800 p x 0 2 Highest differential, neutral slope Highest differential, neutral slope 2 n d highest differential, Neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in Zone 1 and 2 differential Vanes x 1 4 0 p x O 2 Highest differential, Minimal slope 2 n d highest differential, less slope 3 r d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Minimal change Vanes x 800 p x 0 2 Highest differential, Minimal slope Highest differential, less slope 2 n d highest differential, neutral slope Negative differential, neutral slope Negative differential, neutral slope Increase in differential in Zones 1 and 2. PSA oxygen (except for a single test on air); hence the only treatment effect is oxygen flow rate, and discharge water velocity is the covariate. The initial box and whisker plots of K T ^ O , SOTR, S A E and SOTE, as a function of oxygen flow rate and discharge water velocity, showed an unusual pattern of increased variance for the 1 1 min"1 oxygen flow rate (Figure 4.63) and discharge velocities of 60 and 70 cm sec"1 (Figure 4.64). The untransformed plots of K i ^ o , SOTR, S A E and SOTE show a distinct non-linear "break function" type response for the 1 min"1 treatment (Figure 4.65 a-d). The log-transformed data showed the same response (Figure 4.66 a-d); hence the analysis was re-run in two separate methods. The first approach (Analysis I) was to analyse the untransformed data set without the two highest discharge velocities (i.e., 60 and 70 cm sec"1) in an attempt to determine homogeneity of slopes. The analysis indicated a significant interaction effect (Figure 4.67 a-d); hence the data was log transformed, and re-analyzed (Figure 4.68 a-d). A significant result (a < 0.01) was obtained for the discharge velocity covariate effect for K i ^ o , SOTR, S A E and SOTE, and the oxygen gas flow rate for K i ^ o , SOTR and SOTE (Table 4.71). S A E was not significantly influenced by oxygen gas flow rate. A significant oxygen flow rate x discharge velocity interac-tion effect was observed for K L a 2 0 , SOTR, S A E and SOTE (Table 4.71). Since this approach was unable to linearize the oxygen flow rate x discharge velocity interaction effect, a second analytical approach (Analysis II) was used in which the 1 1 min"1 gas flow rate treatment was removed from the data set, and the log transformed data re-analysed (Figure 4.69 a-d). Significant results (a < 0.01) were obtained for the discharge velocity covariate effect for K L a 2 0 , SOTR and SOTE, and oxygen flow rate for K L a 2 0 , SOTR, S A E and SOTE (Table 4.72). S A E was only significant for oxygen flow rate. A significant oxygen flow rate x discharge 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Oxygen Flow (L min-1 ) CM O D) O 1.0 1.5 2.0 2.5 3.0 Oxygen Flow (L min-1) 3.5 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Oxygen Flow (L min-1) 1.0 1.5 2.0 2.5 3.0 Oxygen Flow (L min-1) ure 4.63a-d Box and whisker plots showing effect of PSA O 2 flowrate on K i ^ o (a) SOTR (b), SAE (c) and SOTE (d). o CM 10 9 8 7 6 5 4 3 2 1 T T V 1 i JL -i 1 10 20 30 40 50 60 70 80 Discharge Velocity (cm sec-1 ) CM O D) O W 110i— 100 -90 -80 -70 -60 -50 -40 -30 -20 -101— 10 ~i 1 1 1 1 r , , X 1 JL -T T T TT i i i 20 30 40 50 60 70 80 Discharge Velocity (cm sec-1) CM O LU < 10 20 30 40 50 60 70 80 Discharge Velocity (cm sec-1 ) LU H O C/> 10 20 30 40 50 60 70 80 Discharge Velocity (cm sec-1 ) Figure 4.64a-d Box and whisker plots showing effect of water discharge velocity on K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d). •|l I I I I I I I 1gl I I I I I I I 10 20 30 40 50 60 70 80 10 20 30 40 50 60 70 80 Discharge Velocity (cm sec-1) Discharge Velocity (cm sec-1 ) 10 Oxygen F\oyi • 1 L min-1. • 2 L min-1 3 L min-1" 80 70 60 g. 50 LU t— O 40-30 20 Oxygen Flow • 1 L min-1 T 2 L min-1 3 L min-1 Discharge Velocity (cm sec-1 ) 10 20 30 40 50 60 70 80 Discharge Velocity (cm sec-1 ) Figure 4.65a-d Effect of PSA O 2 flow rate and water discharge velocity on K i ^ o (a), SOTR (b), SAE (c) and SOTE (d). 1 1 i - 1 T -_ 1 T • • I 1 Oxygen Flow • • 1 L min-1 " - • 2 L min-1 -- x i 3Lmin-1 _ -1 1 i i i i i i 10 20 30 40 50 60 70 Discharge Velocity (cm sec-1 ) 80 2.1 2.0 jf 1.9 n 1.8 0 o> 1.7 1.6 or co 1-5 1.4 1.3 cn o 1.2 1 i i i i i • Y _ * f 1 ~ Oxygen Flow • r\ • t 1 • 1 L min-1 " • • 2 L min-1-- \ ^ • 3 L min-1 _ - 1 " i i i 10 20 30 40 50 60 70 Discharge Velocity (cm sec-1) 80 Discharge Velocity (cm sec-1 ) Discharge Velocity (cm sec-1 ) Figure 4.66a-d Effect of PSA O 2 flow rate and water discharge velocity on K i ^ o (a), SOTR (b), SAE (c) and SOTE (d) using log transformed data. Figure 4.67a-d Effect of PSA O 2 flow rate (1,2 and 3 1 min"1) and water discharge velocity (20-50 cm sec1) on K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d). Discharge Velocity (cm sec-1 ) Discharge Velocity (cm sec-1 ) '10 20 30 40 50 60 ' 10 20 30 40 50 60 Discharge Velocity (cm sec-1) Discharge Velocity (cm sec-1 ) Figure 4.68a-d Effect of PSA O 2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-50 cm sec"1) on K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) using log transformed data. Table 4.71. Adjusted least squares means (± SE), treatment and interaction effects for various PSA oxygen flow rates with the Speece Cone without the 60 cm sec"1 and 70 cm sec"1 water discharge velocities. Treatment K L a 2 0 (hr 1) SOTR (g 0 2 hr 1 ) SAE (g 0 2 kWhr 1 ) SOTE(%) n 11 min 1 4.4 (0.09) 47.9 (0.94) 55.9 (0.92) 61.1 (0.69) 16 2 1 min"1 6.6 (0.09) 71.7 (0.94) 75.1 (0.92) 45.6 (0.69) 16 3 1 min"1 7.7 (0.09) 84.2 (0.94) 80.0 (0.92) 35.8 (0.69) 16 Interaction Yes Yes Yes Yes O 2 flow effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Discharge velocity effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 o CM CD 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Oxygen Flow • 2 L min-1 T 3 L min-1 10 20 30 40 50 60 70 Discharge Velocity (cm sec-1) 80 2.1 2.0 | 1.9 S 1.8 O u> 1.7 £ 1.6 co 1 .0 o 1.4 _i 1.3 1.2 1 Oxygen Flow • 2 L min-1 T 3 L min-1 20 30 40 50 60 70 80 Discharge Velocity (cm sec-1) 2.0 „ 1.9 I 1.8 1.7 CM O 1.6 £ 1.5 £ 1 . 4 o - 1 1.3 1.2 1 I 1 1 • T 1 1 1 1 '• * • ft • 1 -Oxygen Flow 1 1 1 • 2 L min-1 • 3 I min-1 -i i i 20 30 40 50 60 70 Discharge Velocity (cm sec-1) 80 O CO CD o 20 30 40 50 60 70 Discharge Velocity (cm sec-1) 80 Figure 4.69a-d Effect of PSA 0 2 flow rate (2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on K L a 2 0 (a), SOTR (b), SAE (c) and SOTE (d) using log transformed data. velocity interaction effect was observed for K i ^ o and SOTR (Table 4.72). The significant oxygen flow rate x discharge velocity interaction effect indicates that K i ^ o and SOTR increased 1.2 fold, at 3 1 min"1, with increasing discharge velocity as compared to the 2 1 min-1 oxygen flow rate (Figure 4.69 a-b). Of the two analytical approaches, neither is technically correct as some of the actual test data is deleted; however, the latter method appears more useful as the log transformation of the interac-tion effects allows some initial quantification of the performance of the Speece Cone under two different experimental gas flow rate treatments. Table 4.72. Adjusted least squares means (± SE), treatment and interaction effects for various PSA oxygen flow rates with the Speece Cone without the 1 min"1 oxygen flow rate treatment. Treatment K L a 2 0 (hr"1) SOTR (g 0 2 hr 1 ) SAE (g 0 2 kWhr"1) SOTE(%) n 2 1 min"1 6.8 (0.087) 73.6 (0.94) 73.9 (0.96) 46.8 (0.50) 24 3 1 min"1 8.1 (0.087) 88.7 (0.94) 80.9 (0.96) 37.7 (0.50) 24 Interaction Yes, 1.20 Yes, 1.20 No No 0 2 flow effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 Discharge velocity effect sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 sig, p = 0.000 A n effective non-statistical technique to visualise the entire performance envelope of the Speece Cone is to use three-dimensional graphing to examine the behaviour of K i ^ o , SOTR, S A E and SOTE under the full range of experimental treatments. This visualization approach clearly shows the "break function" response as well as smaller change in slope between the 2 and 3 1 min"1 oxygen flow rates. K i ^ o and SOTR (Figure 4.70 and Figure 4.71) increased as oxygen flow rates increased from 1 1 min"1 to 2 1 min - 1 , over the range of velocities from 20 cm sec"1 to 50 cm sec"1, then exhibited reduced slope as oxygen flow increased to 3 1 min"1. However, at Figure 4.71 Effect of PSA 0 2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on SOTR. oxygen flow rates < 2 1 min"1 and discharge velocities > 50 cm sec"', K r ^ o and SOTR decreased precipitously. The S A E response manifold was similar to Kj^oand SOTR response (Fig-ure 4.72). The most dramatic response was for SOTE, which shows increasing oxygen transfer efficiency as oxygen flow rate decreases from 3 1 min"1 to 1 1 min"1 over the range of discharge velocities of 20 cm sec"1 to 50 cm sec - 1 (Figure 4.73). However, at oxygen flow rates < 2 1 min"1 and velocities > 50 cm sec"1, SOTE decreases abruptly, indicating a collapse of the high oxygen transfer efficiencies normally associated with Speece Cone aerators. 4.5.2 A i r Experiments The results of the air only tests with the Speece Cone were considerably different from the performance on PSA generated oxygen. The tests were conducted at 3 1 min"1 air at 50 cm sec"1, as this was among the most efficient oxygen flow rate x discharge velocity combinations for the Speece Cone. However, the mean values obtained for K j ^ o , SOTR, S A E and SOTE at this setting, on air, were lower than any recorded on PSA oxygen (Table 4.73). For comparative purposes, the air results are shown alongside the 3 1 min"1 x 50 cm sec"1 data for PSA oxygen from Table 4.71. Table 4.73. Mean (± SE) K L a 2 0 , SOTR, SAE and SOTE values for Speece Cone hypolimnetic aerator operating on air at 3 1 min"1 and 50 cm sec1 compared to PSA oxygen at 3 1 min"1 and 50 cm sec"1. Treatment K L 8 2 O (hr 1) SOTR (g 0 2 hr 1 ) SAE (g 0 2 kWhr ]) SOTE(%) N 3 1 min"1 air at 50 cm sec"1 1.2 (0.01) 13.4 (0.04) 16.9 (0.05) 24.5 (0.08) 3 31 min"1 PSA 0 2 at 50 cm sec"1 7.7 (0.09) 84.2 (0.94) 80.0 (0.92) 35.8 (0.69) 16 Figure 4.73 Effect of PSA 0 2 flow rate (1, 2 and 3 1 min"1) and water discharge velocity (20-70 cm sec"1) on SOTE. C H A P T E R 5: DISCUSSION 5.1 Review of Gas Transfer Theory It is useful to briefly review gas transfer theory prior to discussing the experimental results, as it will assist in interpreting the data. Several theories have been proposed to describe the mass transfer of a sparingly soluble gas, such as oxygen, to water (Krenkel and Orlob, 1962). These include the Two-Film theory (Lewis and Whitman, 1924) which envisioned a thin film of gas and a film of liquid at the gas-liquid interface; the Film Penetration theory (Higbie, 1935) which postulated that gas transfer occurs during alternate cycles of gas adsorption and mixing; and the Surface Renewal theory (Danckwerts, 1951) which extended the Film Penetration theory by stressing the idea of surface renewal. Although these theories are approximations of the actual physical process, Dobbins (1964) proposed the theory that best appears to describe the gas transfer process. This theory is a combination of the classic Two-Film theory (Lewis and Whitman, 1924) and the Surface Renewal theory (Danckwertz, 1951). According to Dobbins (1964), a gas such as oxygen passes through two films of gas and liquid, respectively, by molecular diffusion, and the mass transfer is driven by a partial pressure gradient in the gas phase and a concentration gradient in the liquid phase. Initially, oxygen molecules from the gas phase are transported to the liquid film surface, resulting in saturation conditions at the interface. For slightly soluble gases such as oxygen, the gas film offers very little resistance and this phase proceeds rapidly. The liquid interface has a finite thickness with unique properties. This layer is estimated to be at least three molecules thick and composed of water molecules (i.e., H 2 O ) oriented with their negative sides (i.e., oxygen) facing the gas phase (Eckenfelder and Ford, 1968). In the second phase, oxygen molecules pass slowly through this film by molecular diffusion. It is presumed that the concentration of oxygen in solution at the interface is at saturation, hence all of the resistance to the passage of oxygen into the water is due to molecular diffusion across the liquid film (Eckenfelder, 1959). In the final stage, oxygen is mixed into the water by diffusion and convection currents. At low energy mixing levels (i.e., laminar flow conditions), the rate of oxygen absorption is regulated by the rate of molecular diffusion through the undis-turbed liquid film, and the Two-Film theory holds true. However, as turbulence levels increase, the surface film is disrupted and renewal of the film becomes responsible for transferring oxygen to the liquid (Eckenfelder, 1969). Thus, renewal of the surface film can be envisioned as the frequency with which fluid with a solute concentration of the bulk liquid (i.e., C L ) is replacing fluid from the interface with a concentration (i.e., Q) derived from its contact with the gas phase. Turbulent eddies from the bulk liquid move to the interface and undergo a brief interval nonsteady-state molecular diffusion before being displaced from the surface film by subsequent eddies. Therefore, mass transfer becomes a function of the eddy exposure time, and the molecular diffusion during the time interval (Eckenfelder and Ford, 1969). This process is graphically depicted in Figure 5.1 (from Eckenfelder and Ford, 1969). INTERFACE LIQUID INTERFACE \LIQUID —<^FILM QUIESCENT TURBULENT Figure 5.1 Schematic of Two Film Theory gas transfer process (from Eckenfelder and Ford, 1969). Dobbins (1964) resolved the dilemma between the Two-Film theory, Penetration theory and Surface Renewal theory by developing the following equation to describe the liquid film coefficient (K L ) (King, 1970): (5.1) KL=VfV c o t h j | ï where: K L = liquid film coefficient (m hr"1); D L = diffusion coefficient for oxygen (m 2 hr"1); r = rate of renewal of liquid film (hr 1); l 2 = liquid film thickness (m2). When the rate of surface renewal (r) is near zero (i.e., laminar flow conditions), equation (5.1) reduces to: (5.2) K L = DL/1 and gas transfer is controlled by molecular diffusion through the liquid film as envisioned by the Two-Film theory. However, as the rate of renewal (i.e., r) increases equation (5.1) reduces to: (5.3) K L =VDj and the mass transfer becomes a function of the rate of surface renewal as described by the Danckwertz Surface Rejuvenation theory. In essence, both theories were correct, and these two models may be regarded as the limits between which both gas transfer mechanisms contribute to the overall oxygen transfer process. In the transitional zone between molecular diffusion and turbulent mixing, the process may be visualized as a transfer, in series, through the diffusional sublayer, at the interface whose boundary layer is subjected to turbulence and subsequent surface renewal (O'Connor, 1982). Mathematically, this theory can be expressed as follows (Mavinic and Bewtra, 1974): (5.4) dm/dt = K L A (Q - C L ) where: dm/dt = time rate of mass transfer (g hr"1); K L = liquid film coefficient (m hr"1); A = interfacial or absorbing surface area of air (m 2); Ci = saturation value of dissolved oxygen at the interface between liquid and air bubble (mg C L = average concentration of dissolved oxygen in the bulk liquid (mg l"1). This mass equation may be expressed in concentration units by introducing the volume of the liquid (V): (5.5) dc/dt = 1/V dm/dt = K L A / V (Q - C L ) = K L a (Q - C L ) where: dc/dt = rate of oxygen transfer (g L" 1 hr"1); V = volume of the liquid (m 3); A / V = a = the interfacial surface area of the air through which diffusion can occur generated by the particular aeration system per unit volume of water (m 2 m"3); Ki^a = overall oxygen transfer coefficient (hr"1). In practice, it is very difficult to measure K L and "a" directly due to the hypothetical nature of the liquid film thickness in K L (according to the Two-Film theory), and the technical difficulties of measuring the interfacial surface area "a" (i.e., A / V ) . Therefore, it has become standard practice to consider the aeration process in terms of the overall oxygen transfer coefficient (KL,a), when characterizing the performance of aeration systems (Nienow, 1980). 5.2 Practical Applications of Gas Transfer Theory The next step to understanding the experimental results is to review the practical applications of the aforementioned gas transfer theory. As previously indicated, mass transfer is driven by (1) a partial pressure gradient in the gas phase and (2) a concentration gradient in the liquid phase, and oxygen absorption is regulated by (3) the rate of molecular diffusion through the undis-turbed liquid film at low energy mixing levels, and (4) the rate of renewal of the liquid film at high energy mixing levels. These four fundamental tenants are the underlying principles of gas transfer, mathematically summarized in Equation 5.5: Ki,a (Q - CjJ. Therefore, all aspects of gas transfer design must consider the effect of various engineering modifications on the liquid film coefficient (K^.the interfacial area for gas transfer (a), the saturation value of dissolved oxygen at the interface between liquid and air bubble (Cj) and the average concentration of dissolved oxygen in the bulk liquid (CL) (Mavinic and Bewtra, 1976). In total, there are at least sixteen environmental and design factors that influence the rate of oxygen transfer into water in diffused aeration systems, all of which must be considered when designing efficient aeration systems (Mavinic and Bewtra, 1974). These include: (1) water temperature; (2) depth of gas release; (3) contact time of the gas bubble in the water; (4) size of the gas bubble; (5) rate of gas flow; (6) type of diffuser; (7) orifice diameter; (8) oxygen concentration gradient; (9) turbulence in and around the gas-liquid interface (10) position of the diffuser; (11) geometry of the mixing tank (12) concentration of dissolved solids; (13) elevation (14) barometric pressure; (15) presence of surface active agents (i.e., surfactants) (16) presence of oil contaminants in air supply. From an operational perspective, some factors can be reduced or eliminated through proper design and maintenance (e.g., geometry of the mixing tank, oil contamination in the air supply), and several factors are beyond the immediate control of the practicing engineer (i.e., water temperature, concentration of dissolved solids, elevation, barometric pressure and presence of surfactants). The remaining nine factors (i.e., depth of release, bubble contact time, bubble size, gas flow rate, diffuser type, orifice diameter, oxygen concentration gradient, turbulence in and around the gas-liquid interface and position of the diffuser) define the range of criteria which must be considered when designing gas transfer systems as they individually, or collec-tively, influence: (1) the hydrodynamics of the system which influence ( K L ) ; (2) the area of contact between the gas and liquid (a); (3) the concentration gradient between the gas and liquid phase (CJ-CL) (Bewtra and Mavinic, 1978;Nienow, 1980). In this set of experiments, one parameter (i.e., the lateral position of the diffuser) was not examined due to the size constraints of the experimental system, thus leaving the remaining eight factors to be considered in the experimental design. At this stage of the discussion, it is useful to review the relationship between Equation 5.5: K^a (Q - C L ) , the eight remaining operational factors, and the experimental variables selected for this series of experiments (Table 5.1). This demonstrates the interconnectedness and complexity of optimizing gas transfer, as altering one variable can have multiple implications. The rationale for selecting the experimental variables to examine in the full lift and Speece Cone hypolimnetic aerators should now be apparent. A n important fact to note is that the four standard units of measure used to quantify oxygen transfer, i.e., (1) KLa2o (hr _ 1 ) , the oxygen transfer coefficient at 20 °C; (2) SOTR (g 0 2 hr"1), the Standard Oxygen Transfer Rate; (3) S A E (g 0 2 kWhr"1), the Standard Aeration Efficiency and (4) SOTE (%), the Standard Oxygen Transfer Efficiency; do not necessarily respond similarly to design modifications. By definition, K i ^ o and SOTR will respond proportionally (see Section 3.9); however, S A E and SOTE are not directly computationally linked to K i ^ o and SOTR. Therefore, it is possible to cause an increase in KIAÎO and SOTR in response to a design modifi-cation, while causing a simultaneous decrease in S A E and SOTE. Thus, the practising engineer must consider all of the design criteria when attempting to design an optimal gas transfer system. Alternatively, one or more units of measure can be intentionally sacrificed in exchange for improved gas transfer performance in the remaining unit(s) of measure in order to achieve a Table 5.1. Relationship between Ki,a (Ci - C L ) , the eight operational factors, and the experimental variables. Full lift experimental variables Effect on K L a ( C , - C L ) Operational design factors Diffuser orifice pore diameter (140 p, 400 p and 800 p) K L , a Bubble size, contact time, diffuser type, orifice diameter Depth of gas release (1.5 and 2.9 m) K L , a, Q - C L Depth of release, bubble contact time, bubble size, oxygen concen-tration gradient Gas flow rate (10, 20, 30 and 40 1 min"1) K L , a Bubble contact time, gas flow rate, turbulence in and around the gas-liquid interface Oxygen partial pressure (air and P S A generated oxygen) C i - C L Oxygen concentration gradient Floating surface cover a Oxygen concentration gradient Bubble screen K L Bubble contact time, bubble size, turbulence in and around the gas-liquid interface Counter-rotating blades K L Bubble contact time, bubble size, turbulence in and around the gas-liquid interface D B C A K L , a, C J - C L Bubble contact time, bubble size, turbulence in and around the gas-liquid interface, type of diffuser, orifice size, oxygen concentration gradient Speece Cone design factors K L a ( C i - C L ) Operational criteria Oxygen flow rate (1,2 and 3 1 min" ) K L , a , Q - C L Bubble contact time, gas flow rate, oxygen concentration gradient, turbulence in and around the gas-liquid interface Outlet port discharge water velocity (20-70 cm sec"1) K L , a Q — C L Bubble contact time, Turbulence in and around the gas-liquid interface desired level of system performance. This contradictory behaviour is the paradox of gas transfer system design. 5.3 Full Lift Hypolimnetic Aeration - Groups 1 to 6 The rationale of the laboratory experiments in Groups 1 to 6 was to examine the effects of diffuser depth (1.5 and 2.9 m), orifice diameter (140 p, 400 p and 800 p), and oxygen partial pressure (air and PSA O2) on K.L.a2o, SOTR, S A E and SOTE, with gas flow rate (10, 20, 30 and 40 1 min"') as the covariate. Water velocity and bubble size were additional data collected to assist in the interpretation, recognising that the bubble size data may be of limited use due to the small sample size relative to the number of bubbles in the system. Although K i ^ o and SOTR are the units of measure which provide basic information on the characteristics of each experi-mental configuration, the two most relevant units of measure for comparing aeration systems are S A E and SOTE (Beak, 1977, Bewtra and Mavinic, 1978). The effect of varying the experimen-tal treatments on KLa2o, SOTR, S A E and SOTE, water velocity and bubble size is discussed below. 5.3.1 Diffuser Depth The diffuser depth treatment exerted a significant effect on K i ^ o , SOTR, S A E and SOTE, with all units of measure showing positive increases in response to an increase in diffuser submer-gence. The magnitude of the responses were all quite large, ranging from 30 to 57% (Table 4.7). The depth of gas release influenced K i ^ o , SOTR and SOTE, mainly by an increase in bubble contact time due to the longer path length the bubbles must take to reach the surface, and a larger deficit or driving force (CJ-CL); this resulted in more oxygen transfer (Mavinic and Bewtra, 1974; Mavinic and Bewtra, 1976). The increase in S A E was unexpected, as similar studies with an upward, co-current, air-water flow observed a decrease in S A E with increased diffuser submergence (i.e., System II; Mavinic and Bewtra, 1976). S A E increased in these experiments, as there was an increase in the bubble-travel distance, an increase in driving force ( Q - C L ) , as well as an increase in water velocity that increased turbu-lence at the liquid interface, and influenced K L . The water velocity data (Table 4.47) clearly demonstrate that velocity increased with depth in each experimental treatment combination, thus supporting this conclusion. The fact that S A E increased with diffuser submergence in this case indicates that, despite the additional air power energy required to deliver air or PSA O2 to greater depths, the energy cost was more than offset by the increased oxygen transfer resulting from the longer bubble contact times, greater turbulence and increased hydrostatic pressure. Unlike the Mavinic and Bewtra (1976) study, there were no water power energy costs in this system, which was a significant energy cost in their Systems II, III and IV. In addition, the outlet tube water from the full lift aerator discharged into the bulk liquid, hence water velocities increased with diffuser submergence, but not to the same extent as the circulating closed-loop system design of Mavinc and Bewtra (1976). This resulted in greater initial turbulence in the inlet tube, as the water entering the inlet tube possessed minimal velocity; hence high shear forces were present as the rising air bubbles mixed with the slower moving water mass entering the inlet tube from the bulk liquid reservoir. Gas flow rate significantly influenced K i ^ o , SOTR, S A E and SOTE in all of the experimental combinations. K i ^ o and SOTR responded positively to increased gas flow rates, a response noted in numerous gas transfer studies (e.g., Bewtra et al., 1970; Schmidt et al., 1978). At least two mechanisms are responsible for this effect. Firstly, higher gas flow rates generate increased turbulence at the liquid interface. Under highly turbulent conditions, mass transfer is regulated by disruption and the rate of renewal of the liquid film as described by the Danckwertz Surface Rejuvenation theory (Dobbins, 1964) (5.3): K L = ^ /D L r . As r (the rate of surface renewal) increases, K L increases. Therefore, one mechanism by which higher gas flow rates increase KLa2o and SOTR is via their positive effect on K L - Secondly, increased gas flow rates increase the number of bubbles present in the water column per unit time. This increases the total interfacial area available for gas transfer to the surrounding medium (Bewtra and Mavinc, 1974). Therefore, increased gas flow rates positively influence K^o and SOTR via their effect on "a". S A E and SOTE responded negatively to increased gas flow rates, a common observation in diffused aeration systems (Bewtra and Nicholas, 1964; Ellis and Stansbury, 1980; Mavinc and Bewtra, 1976). The general explanation for this response is that air bubbles become larger with an increase in gas flow rates. This results in less oxygen transfer due to the reduced ratio of interfacial area to bubble volume. Secondly, as air bubbles become larger, their terminal rise velocities increase, thus reducing the contact time between air bubbles and the surrounding liquid, even though there is a corresponding increase in turbulence at the liquid film. Finally, increased gas flow rates incur additional energy costs for gas compression, both air and PSA 0 2 , which penalises S A E . This is an example of how competing factors interact to determine overall gas transfer rates. As previously indicated, bubble velocity increased significantly with increased diffuser submer-gence, as the longer path length allows bubbles to increase to their terminal rise velocity. Bubble velocity also increased significantly with increased gas flow rate, as the additional volume of air in the inlet tube further decreases the density of the air-water mixture in the inlet tube by increasing the gas volume void fraction; this increases the water velocity in air-lift pumps (Andeen, 1974). Higher gas flow rates also create larger bubbles, and larger bubbles exhibit higher rise velocities (Andeen, 1974). The bubble size data did not show any relationship to gas flow rates. The coefficient of varia-tion (CV) was so large that the broad range in the E Q D data hid any significant effect. The bubble size data indicated the size range between the 2.9/0.5 m and 1.5/0.5 m treatments was only 0.4 mm for the 140 u. and 400 u. treatments (Tables 4.58 and 4.59), and 0.8 mm for the 800 u. treatment (Table 4.60). In general, all of the E Q D values were within ~2 mm in each orifice size class, hence it appears the technique was not sufficiently accurate to detect an effect of depth on EQD. The r 2 for the depth x gas flow rate regression only ranged from 0.28 to 0.34, even with the log transformed data; hence the utility of the E Q D data is minimal. 5.3.2 Oxygen Partial Pressure The oxygen partial pressure treatment exerted a very significant effect on K i ^ n and SOTR, with both units of measure showing positive increases in response to an increase in oxygen concen-tration from the 21% O2 content of air to the -95% O2 content of the PSA gas. The magnitude of the responses were substantial, ranging from 430 to 542% for K La2o, and 432 to 532% for SOTR (Table 4.14). The increase in oxygen partial pressure influenced K i ^ o and SOTR by substantially increasing the driving force (CJ-CL), resulting in more oxygen transfer per unit bubble contact time. Since there was no effect of oxygen partial pressure on water velocity (Table 4.46), the increased concentration gradient accounts for all of the observed effect. Interestingly, the ratio of the observed increase in K i ^ o and SOTR exceeded the PSA gas:air oxygen ratio (i.e., 95%/21% = 4.5) in most cases, suggesting a non-linear effect may be occur-ring at the liquid film; this enhances the diffusion of oxygen molecules above what would be expected on 0 2:air ratios alone. SOTE was also positively influenced by oxygen partial pressure, and although the increases were not as large as for K L ^ O and SOTR, they ranged from 33 to 54%, which is a significant effect. The fact that SOTE did not increase by a ratio of 4.5:1 indicates that a significant fraction of PSA 0 2 gas was not being utilised, and was escaping to the atmosphere, essentially as waste gas. A logical design modification would be to capture the unused PSA 0 2 in the separator box, and re-introduce it to the downflowing water in the outlet tube by a venturi arrangement, in order to scavenge the oxygen without incurring any additional energy costs. The S A E response was strongly negative, and quite uniform, ranging from a 72 to 76% decrease when operating on PSA 0 2 . This indicates that, despite the large increase in (CJ-CL), resulting in more oxygen transfer per unit bubble contact time, the energy costs to produce 95% oxygen via PSA technology more than offset any gains. A detailed discussion of the efficiency of PSA oxygen generation technology and the explanation for the high energy costs is included later in this chapter. Gas flow rate significantly influenced K.La2o, SOTR, S A E and SOTE in all of the experimental combinations, with KLa2o and SOTR responding positively to increased gas flow rates. As before, higher gas flow rates means an increase in the number of bubbles in the system, thus increasing the total interfacial area available for gas transfer to the surrounding medium and the amount of oxygen transfer to the liquid (Bewtra and Mavinic, 1974). In addition, higher gas flows generate increased turbulence, thus increasing turbulence at the liquid interface and the rate of renewal of the liquid film. Even though bubble contact time is reduced by the higher velocities, the tradeoff favors increased gas transfer. S A E and SOTE responded negatively to increased gas flow, as air bubbles become larger with an increase in gas flow rates, resulting in less oxygen transfer due to the reduced ratio of interfa-cial area to bubble volume. The terminal rise velocities of the bubbles also increases, thus reducing the contact time between air bubbles and the surrounding liquid, even though turbu-lence is increased at the liquid film. The increased energy costs required to compress additional air and PSA O2 penalised the S A E unit of measure. There was no effect of gas composition on bubble velocity in any of the experimental combina-tions (Table 4.46). Bubble size was not measured for PSA O2; hence there is no basis for comparison of E Q D as a function of air or PSA 0 2 . 5.3.3 Orifice Diameter Orifice diameter exerted a significant effect on K La2o, SOTR, S A E and SOTE. Each unit of measure increased with a decrease in orifice size (Table 4.19). Smaller orifices produce smaller bubbles, and a reduction in bubble size influences gas transfer in at least four ways: 1. A n increase in surface area per unit bubble volume (Eckenfelder, 1969); 2. A decrease in terminal rise velocity (Stenstrom and Gilbert, 1981); 3. A decrease in the liquid film coefficient (Bewtra and Nicholas, 1964); 4. A n increase in the total number of bubbles in the water column per unit gas discharge. A n increase in bubble surface area per unit volume increases the "a" in K I A and acts to increase K\jà2o and SOTR. A decrease in terminal rise velocity increases the bubble contact time, which acts to increase K L ^ O and SOTR and SOTE, and S A E as no additional energy costs are in-curred. However, the competing response is that a decrease in terminal rise velocity decreases the liquid film coefficient (KL) and turbulence at the liquid interface, which decreases KLa2o, SOTR, S A E and SOTE. Increased number of bubbles in the water column per unit time can saturate the interstitial spaces between the bubbles with oxygen more quickly, i f the bubbles are in a confined column, and reduce the concentration gradient across the liquid film (Mavinic and Bewtra, 1974). The interaction of these opposing factors determines the net effect on KL32O, SOTR, S A E and SOTE. In these experiments, the overall effect was an increase in all units of measure, indicat-ing that the effect of increased surface area and contact time more than compensated for the reduction in K L due to slower rise velocities and increased bubble density in the inlet tube. Since the reduction in orifice diameters did not require an increase in energy, S A E responded positively for the same reasons. If orifice diameter were reduced below a critical minimum diameter (e.g., < 2 p), then additional energy would be required to force gas through the diffuser and overcome surface tension at the orifice-water interface, and S A E would decline accord-ingly. The 2 p diffusers used in the D B C A experiments were in the 2 p size range, and required considerable back pressure before any gas could be seen emerging from the diffuser. In con-trast, the 140 p, 400 p and 800 p diffusers showed minimal backpressure, and responded instantly to fluctuations in gas flow rate. A number of researchers have reported a similar effect of increased SOTE with decreased bubble diameter. For example, Morgan and Bewtra (1960) and Bewtra and Nicholas (1964) both observed increased SOTE with fine bubble (Saran tubes) as compared to coarse bubble diffusers (Spargers). Barnhart (1969) also demonstrated a decrease in K La2owith bubbles diameters of 2 mm and larger. Scheffe's test indicated each orifice size was significantly different from each other for the four units of measure. Since Scheffe's test is conservative by design and the most rigorous a poste-riori test for performing comparison among means (Larkin, 1975), there is little doubt the observed results were real, and not statistical artifacts. Gas flow rate significantly influenced K j ^ o , SOTR, S A E and SOTE in all of the experimental combinations, with K j ^ o and SOTR responding positively to increased gas flow rates, and S A E and SOTE decreasing with increased gas flow rate. The reasons are as before, with higher gas flow rates increasing in the number of bubbles and total interfacial area available for gas transfer to the surrounding medium, increased turbulence at the liquid interface and higher renewal rates of the liquid film. Again, even though the higher velocities reduce bubble contact time, the tradeoff increased K i ^ o and SOTR, thus demonstrating the importance of interfacial area relative to contact time. S A E and SOTE responded negatively to increased gas flow rates. A i r bubbles become larger with an increase in gas flow rates, resulting in less oxygen transfer due to the reduced ratio of interfacial area to bubble volume. The terminal rise velocities of the bubbles also increases, thus reducing the contact time between air bubbles and the surrounding liquid even though turbu-lence is increased at the liquid film. The additional energy costs required to increase air or PSA O2 flow also act to penalise the S A E , hence the net result is a decrease in S A E . The effect of orifice size on water velocity was limited to the 1.5 m submergence treatment, in which the 140 JJ. and 400 u. diameter orifices generated higher velocities than the 800 u. orifice (Table 4.48). The author believes this result was caused by the larger bubbles released by the 800 u. at 1.5 m rising in a rocking motion, then fragmenting into a variety of bubble sizes and shapes. The bubbles generated by the 140 |x and 400 u. diffuser appeared more uniform in shape and diameter, and rose more quickly through the water column. The coefficient of variation (CV) for the 140 u. bubbles at 1.5 m was the lowest observed (Table 4.53), while the C V for the 800 p bubbles at 1.5 m was the largest, with some C V ' s exceeding 100% (Table 4.55). There was no statistical difference in water velocity at the 2.9 m submergence depth (Table 4.53). It is believed this was caused by an averaging effect that takes place during bubble ascent in the inlet tube, in which fragmentation and coalescing of bubbles, under decreasing hydrostatic pressure, tends to "average out" the bubble size, and hence velocity. The bubble size data was contradictory. At the 1.5/0.5 m test, the 140 p. orifice generated the largest bubbles, with 400 u, and 800 \i being smaller and statistically indistinguishable from each other. At 2.9/0.5 (i.e., same photo depth, increased diffuser submergence), a similar trend was observed, i.e., the 140 u. and 400 |x bubbles were larger than the 800 p, but not statistically different from each other. This was reversed for the 2.9/2.7 m treatment, as 140 u. were the smallest, and the 400 u. and 800 p similar to each other. As before, the r 2 of the relationship only ranged from 0.28 to 0.38, hence the relationship is of minimal value. About the only credible information on E Q D is the C V data, which increases with orifice diameter. 5.3.4 Group 1 to 6 Summary The Group 1 to 6 experiments illustrated a number of key aspects of gas transfer. Increased depth of submergence and decreased orifice diameter, within the ranges explored in these tests, are design variables that are beneficial to all units of measure. This demonstrates the impor-tance of contact time, concentration gradient, and increased interfacial bubble area on oxygen transfer. Increased oxygen partial pressure created significant increases in K i ^ o and SOTR, moderate increases in SOTE, but moderate losses in SAE. This indicates the importance of concentration gradient on some aspects of gas transfer, but the additional energy costs to generate the oxygen incurred an energy penalty for S A E . In this case, the design engineer is presented with the option of enhancing some aspects of gas transfer, while sacrificing others. A more detailed discussion on the practical implications of these findings wil l follow in a later section. The effect of increasing gas flow rate was similar across all depths, orifice diameters, and oxygen partial pressures. This clearly demonstrates the overriding importance of higher turbu-lence and interfacial areas for increasing K L ^ O and SOTR, and the negative implications on S A E and SOTE resulting from greater energy costs of gas delivery (air and PSA O2). Again, the balancing act between K L a and SOTR, and S A E and SOTE is readily apparent. Water velocity in the full lift aerator, not surprisingly, was directly related to diffuser depth, and gas flow rate, and was independent of orifice diameter, except at the 1.5 m submergence depth. The bubble size data, expressed as equivalent bubble diameter, was of minimal use due to the significant scatter in the data, and the small sample size relative to the population of bubbles in the system. A n improved procedure for measuring bubble size is required using a more sophis-ticated technique such as an optical image analyzer. The most credible data emerging from the E Q D data set is that C V increases with orifice diameter. This was caused by the larger bubbles from the 800 p orifices fragmenting into a range of bubbles from very large to very small, whereas most of the bubbles emitted from the 140 p orifice were quite uniform in size. The 400 p diffuser, being mid-way between fine and coarse bubble diffuser classification, exhibited C V ' s mid-way between the 140 p and 800 p diffusers 5.4 Full Lift Hypolimnetic Aeration - Groups 7 tol4 The Group 7 tol4 experiments examined the influence of four in situ modifications (DBCA, bubble screen, counter-rotating vanes and surface cover), orifice diameter (140 p and 800 p), and oxygen partial pressure (air and PSA 0 2 ) at a single diffuser depth (2.9 m) on K i ^ o , SOTR, S A E and SOTE, with gas flow rate (10, 20, 30 and 40 1 min"1) as the covariate. Water velocity and bubble size data were collected to assist in the interpretation. The effect of varying the experimental treatments on K i ^ o , SOTR, S A E and SOTE, water velocity and bubble size is discussed below. 5.4.1 In situ Modifications The in situ modifications treatment exerted a significant effect on K L ^ O , SOTR, S A E and SOTE, with all units of measure showing a negative response to the various in situ treatments. The magnitude of the responses were all quite large, ranging from 22 to 50% on air (Table 4.31), and 22 to 48% on PSA 0 2 (Table 4.42). The general trend was that the No Treatment option generated the highest values of K.L.a2o, SOTR, S A E and SOTE (1 s t ranking), the Screen and Cover treatments the lowest values (4 t h and 5 t h ranking), and the Vanes and D B C A treat-ments were in the middle range (2 n d and 3 r d ranking). There are at least two reasons for this effect, the first being that the various in situ modifications caused a reduction in water velocity in the full lift aerator (Table 4.49, and Table 5.2). It is known from earlier discussions that orifice diameter had no effect on water velocity at the 2.9 m submergence depth (5.3.3), and that gas composition had no effect on water velocity, regardless of depth (5.3.2). Diffuser submergence did influence water velocity (5.3.1), but since all of the in situ tests were carried out at 2.9 m depth, the only remaining treatment that can explain the reduction in water velocity is the in situ treatments. It was quite obvious when the tests were being conducted that the in situ treatments were causing a reduction in water velocity by restricting flow either in the inlet tube, the separator box, or the outlet tube. Table 5.2. Summary of various full lift in situ modification treatments on water velocity with the 140 p and 800 p diffuser, on air, showing percentage increase/decrease relative to No Treatment. Treatment Velocity (cm sec"1) None x140 p 33.2 Cover x140 p 30.4 (-8%) Screen x 140 p 29.0 (-13%) DBCA x 140 p 26.7 (-20%) Vanes x 140 p 23.9 (-28%) None x 800 p 31.2 DBCA x 800 p 24.5 (-21%) Screen x 800 p 23.8 (-24%) Cover x 800 p 23.2 (-26%) Vanes x 800 p 19.6 (-37%) The two treatments that were located inside the inlet tubes (screen mesh, counter-rotating vanes) restricted flow by their physical presence, as the gas:bubble mixture had to pass through the various physical constrictions. The small screen mesh size openings (0.64 cm square) presented considerable resistance to the bubble:water mixture, and elongated bubbles could be seen streaming upwards from the upper surface of the screen. The counter-rotating vanes acquired mechanical energy to turn the blades from the flowing water, in addition to possessing signifi-cant cross-sectional area (which caused drag resistance). The counter-rotating design of the vanes forced the directional rotation of the bubble: water mixture to change rapidly between the two vanes. Therefore, it is not surprising that the Vanes treatment exhibited the lowest water velocity. The in situ treatments located outside the inlet tube (DBCA) and separator box (Surface Cover) also reduced water velocity. The Surface Cover treatment restricted the height of the bub-ble:water plume exiting into the separator box. The head differential created by the ainbubble plume is the driving force in air lift pumps (Andeen, 1974). By restricting the height of the ainbubble plume, the surface cover effectively reduced the head differential, hence velocity. The D B C A modification reduced velocity by filling the outlet tube with bubbles, mainly very small (Table 4.56 and 4.57), but some larger bubbles were formed by coalescence in the turbu-lent mixing zone at the entrance to the outlet tube. The presence of a significant number of bubbles in the outlet tube reduced the density of the ainwater mixture and caused an upward buoyant force, as the outlet tube attempted to function as a co-current upflow air lift pump. Since the gas flow rate to the D B C A diffuser was fixed at 3 1 min"1, the co-current air lift pump in the inlet tube dominated the direction of the flow pattern, but the D B C A configuration exacted a price by reducing the net velocity. In addition, since the total gas flow in each treat-ment was fixed at 10, 20, 30 or 40 1 min"1, the 3 1 min"1 portion allocated to the D B C A diffuser in the separator box reduced the gas flow rate, hence energy input, to the inlet tube. Since water velocity is related to gas flow rate in air lift pumps (Andeen, 1974), this contributed to a reduc-tion in water velocity with the D B C A configuration. If water velocity was the single factor by which the in situ treatments influenced K L ^ O , SOTR, S A E and SOTE, then their expected ranking, and % change relative to No Treatment, should be approximately similar to the velocity rankings in Table 5.2. (i.e., No Treatment 1st, Screen and Cover 4 t h and 5 t h, and Vanes and D B C A treatments 2 n d and 3 r d ranking). If one assumes a +/-5% leeway around the % differences, then it appears the % relative reduction in water velocity more or less accounts for the % reduction in K j ^ o , SOTR, S A E and SOTE as a result of the Vanes treatment (Table 5.3). The % reduction in K i ^ o , SOTR, S A E and SOTE from the D B C A treatment seems to be slightly greater than the expected % reduction from water velocity alone. However, since 3 1 min"1 of the gas flow rate was re-directed to the diffuser in the separator box with the D B C A treatment, this would partially explain the additional variance. The two treatments that greatly exceeded the reduction in K i ^ o , SOTR, S A E and SOTE, expected solely as a result of reduced water velocity, are the Screen and Cover treatments. In Table 5.3. Summary of various in situ full lift hypolimnetic aerator treatments on air with the 140 p and 800 p diffuser showing percentage increase/decrease relative to No Treatment for K L a 2 0 , SOTR, SAE and SOTE and water velocity. Treatment Water velocity and % change relative to No Treatment K L a 2 0 (hr1) SOTR (g 0 2 hr"1) SAE (g 0 2 kWhr"1) SOTE(%) None x140 u 33.2 cm sec"1 3.3 35.4 324.5 7.9 Vanes x 140 \i 23.9 (-28%) 2.4 (-27%) 26.5 (-25%) 248.1 (-24%) 6.0 (-24%) DBCA x 140 ji 26.7 (-20%) 2.3 (-30%) 24.9 (-30%) 241.3 (-26%) 5.9 (-25%) Screen x 140 p 29.0 (-13%) 2.1 (-36%) 22.5 (-36%) 209.6 (-35%) 5.1 (-35%) Cover x140 p 30.4 (-8%) 1.9 (-42%) 20.6 (-42%) 192.7 (-41%) 4.7 (-41%) None x 800 u 31.2 cm sec"1 2.2 22.1 197.1 4.8 DBCA x 800 u 24.5 (-21%) 1.5 (-32%) 16.0 (-28%) 154.2 (-22%) 3.7 (-23%) Vanes x 800 p 19.6 (-37%) 1.2 (-45%) 13.5 (-39%) 121.0 (-39%) 2.9 (-40%) Screen x 800 p 23.8 (-24%) 1.2 (-45%) 13.4 (-39%) 121.9(-38%) 3.0 (-38%) Cover x 800 p. 23.2 (-26%) 1.1 (-50%) 12.1 (-45%) 107.6 (-45%) 2.6 (-46%) these treatments, K i ^ o , SOTR, S A E and SOTE were typically 35 to 50% less than the No Treatment, whereas the reduction in water velocity, as a result of these treatments, only ranged from 8 to 26%. This suggests additional factors negatively influenced gas transfer in these treatments. For the Screen treatment, the observations of elongated bubbles streaming upwards from the wire mesh likely de-gassed oxygen, which had been absorbed during the initial rise from the release depth (2.9 m). As will be discussed in a later section, most of the gas transfer occurred in the lower Vi of the inlet tube. If turbulence and low pressure zones around the mesh screen caused oxygen to de-gas, this could explain the additional negative effect of the Screen treatment on K j ^ o , SOTR, S A E and SOTE, above what would have be expected solely on the basis of reduced water velocity. The Cover treatment likely caused a reduction in gas transfer by decreasing the transfer of atmospheric oxygen at the turbulent air-water interface generated by the bursting air bubbles. Neilson (1974) observed reduced rates of oxygenation in similar laboratory experiments using a floating stryrene cover. Therefore, the in situ Cover treatment negatively influenced KLa2o, SOTR, S A E and SOTE by reducing water velocity and likely decreased oxygen transfer in the separator box. As seen in the Group 1 to 6 experiments, gas flow rate significantly influenced K j ^ o , SOTR, S A E and SOTE in all of the experimental combinations. K i ^ o and SOTR responded positively to increased gas flow rates, and S A E and SOTE responded negatively. As before, K i ^ o and SOTR responded as higher gas flow rates increased the number of bubbles in the system and increased the total interfacial area available for gas transfer to the surrounding medium (Bewtra and Mavinc, 1974). The higher gas flows generate increased turbulence at the liquid interface and enhance the rate of renewal of the liquid film. Even though the higher velocities reduce bubble contact time, the tradeoff favored K j ^ o and SOTR. As expected, S A E and SOTE responded negatively to increased gas flow. High gas flow rates increased bubble sizes, result-ing in less oxygen transfer due to the reduced ratio of interfacial area to bubble volume. The terminal rise velocities of the bubbles also increased, thus reducing the contact time between air bubbles and the surrounding liquid, even though turbulence was increased at the liquid film. Finally, the incremental energy costs required to compress additional air and PSA O2 penalised the S A E unit of measure. A significant gas flow rate x in situ modification interaction was observed for all of the K L ^ O , SOTR, S A E and SOTE measurements with the 800 p diffuser, and for the K i ^ o and SOTR with the 140 p diffuser. The K i ^ o and SOTR interaction is the additive effect of gas flow rate on oxygen transfer, caused by increasing turbulence and interfacial area with each increase in gas flow rate. Statistically, this means that the effect of a particular treatment is partially explained by the covariate gas flow rate, and is not solely a function of the treatment being examined. The most interesting interaction was the gas flow rate x D B C A treatment with the 800 p diffuser on S A E and SOTE (Figures 4.18 c and d; 4.20 c and d). A similar, but not statistically significant response, occurred with the 140 p diffuser treatment (Figures 4.17 c and d; 4.20 c and d). The data indicates the D B C A treatment exhibited quite high S A E and SOTE values at the lowest gas flow rate (i.e., 10 1 min"1); however, S A E and SOTE plummeted with increasing gas flow, more so than the other in situ treatments, hence the significant interaction effect. This response was caused by déstabilisation of the D B C A effect occurring in the outlet tube at the lowest gas flow rate. At the lowest gas flow, the downward water velocity in the outlet tube closely matched the upward rise velocity of the D B C A bubble swarm in the outlet tube, signifi-cantly increasing contact time and turbulence at the liquid interface, and generating high values for S A E and SOTE. The data indicates that with the 800 (î orifice diffuser on air, the outlet tube water velocity at 10 1 min"1 was -15 cm sec"1 (Figure 4.25 b). The bubble size for the D B C A treatment with the 800 u. diffuser on air at 10 1 min"1 air flow was approx. 2.8 mm E Q D (Table 4.57). The rise velocity of bubbles in this size range is described by (Haberman and Morton, 1954): r e = equivalent bubble radius (cm) Therefore, 2.8 mm E Q D bubbles would have an upward velocity of -12 cm sec"1 using Eqn. (5.1), and closely match the downward water velocity in the outlet tube. As gas flow rates exceeded 10 1 min-1, water velocity increased to >20 cm sec"1, while bubble size decreased to -1.7 mm E Q D due to additional turbulence and bubble fragmentation in the separator box (Table 4.57); bubble rise velocity would have decreased to - 9 cm sec"1. The net result was a destabilization of the D B C A effect as the bubbles were washed out of the outlet tube into the bulk water in the tank. The loss of the counter-current bubble exchange conditions in the outlet tube, combined with increasing energy costs associated with the higher gas flow rates, resulted in the precipitous decline of SOTE and S A E with gas flow rates >10 1 min"1. A similar, but not as dramatic, effect occurred with the 140 jx diffuser on air. The water velocities in the outlet (5.1) where: u = terminal rise velocity (cm sec" ); g = acceleration due to gravity (980 cm sec"2); tube at 10 1 min"1 were slightly greater than those recorded with the 800 p diffuser (Figure 4.25 a) and the bubble size was smaller (~1.7 mm EQD), hence the D B C A effect did not establish itself as strongly as with the 800 p diffuser, nor was the collapse in S A E and SOTE as dramatic, although it did occur. Mavinic and Bewtra (1974) experienced similar responses with their experiments, and designated Qmm air flow rates for their counter-current design (i.e., System III) and Qmax air flow rates for their co-current downflow design (i.e., System IV) to prevent déstabi-lisation of their systems. 5.4.2 Water Velocity Much of the effect of the in situ modifications on water velocity has been discussed in the previous section. One remaining point to discuss is the observed reduction in water velocity with the 800 p diffuser relative to the 140 p diffuser (see Table 5.2). Orifice diameter has been previously shown to have no effect on water velocity at the 2.9 m submergence depth (5.3.3). This conclusion is supported by the No Treatment velocity data in Table 4.49, which shows similar water velocities for the 140 p and 800 p diffusers at 2.9 m diffuser submergence. However, as shown in Table 5.2, some of the in situ modifications with the 800 p diffuser showed lower velocities than with the 140 p diffuser. The magnitude of the reduction in velocity with the 800 p diffuser treatment, relative to the 140 p diffuser treatment, is shown in Table 5.4. Given the standard error of the velocity data Table 4.49, the D B C A treatment showed the least difference. As previously discussed, the larger bubble size generated by the 800 p diffusers would cause increased upward velocity in the D B C A zone in the outlet tube, which decreased water velocity relative to the 140 u diffuser treatment. Table 5.4. Decrease in water velocity with the 800 p diffusers as compared to 140 p diffusers on air for the full lift in situ modification treatments (from Table 5.2). Full lift treatment Decrease in velocity with 800 u diffuser relative to 140 u diffuser (cm sec"1) None 2.0 Cover 7.2 Screen 5.2 DBCA 2.2 Vanes 4.3 The reduction in velocity caused by the Vanes and Screen treatments with the 800 p diffuser was likely a result of some of the smaller bubbles generated by the 140 p diffuser passing through the screen or rotating blades without striking the mesh screen or blades and fragment-ing. Many large bubbles could be seen colliding with the mesh screen and vanes with the 800 p diffuser, and this likely slowed velocity relative to the 140 p treatment, both of which were slower than No Treatment. The explanation for the greater decrease in velocity caused by the 800 p diffuser with the Cover treatment is not apparent. 5.4.3 Bubble Size The effect of the various in situ treatments on bubble size has already been discussed in detail. The bubble size data shows a clear distinction between D B C A bubble size, and bubble sizes generated by the Screen and Vane treatments (bubble size was not collected for the Cover treatment) (Table 4.64 and 4.65). The E Q D was significantly smaller with the D B C A treatment, as the fine pore diffuser in the separator box used to create the D B C A effect had a rated pore diameter of 2 p. Scanning Electron Microscope (SEM) imaging of this diffuser confirmed the pore size was considerably smaller than the 140 p diameter diffuser, and it was effective at generating demonstrably smaller bubbles when operated within the gas flow range recom-mended by the manufacturer. The size of a bubble formed at low gas flow rates, as they emerge from a single orifice, is the result of a balance between the buoyant force of the bubble and the surface tension attracting the bubble to the diffuser orifice. A n individual bubble increases in size until its buoyancy exceeds the surface tension forces, and it then detaches itself (Bowers, 1955). A t low gas flow rates, bubbles tend to emerge in single formation, with a relatively constant diameter of approx. eleven times the orifice diameter (Haney, 1954). However, as gas flow rates increase, single bubbles cannot carry the gas away quickly enough so the bubbles become larger and leave the orifice in the form of a chain, with adjacent bubbles just touching (Bowers, 1955). The gas flow rate at which chain formation occurs is known as the critical point, above which bubble size becomes dependent on gas flow rate and independent of orifice diameter. Bowers (1955) describes the mathematical derivation of the critical gas flow point for various sized bubbles. The mean bubble diameter produced within the range of gas flow rates typically used in aeration practice is an exponential function of the gas flow rate, with the exponent n varying from 0.10 to 0.44 (Eckenfelder, 1959): (5.6) d b ~ G s n where: db = mean bubble diameter; G s n = gas flow rate. Visual observations of bubble formation for the diffusers used in these experiments indicated bubbles emerged in chain formation, so the gas flow rate per orifice was above the critical flow rate for single bubble formation. As a result, bubble size was dependent on gas flow rate, and the resulting bubble sizes were not markedly different for the 140 p, 400 p and 800 p diffusers. Bubble size was smaller, statistically and visually, with the D B C A configuration, because of the extremely small orifice diameter of the Point Four diffuser (i.e., 2 p), and the nature of the counter-current effect (which tended to maintain the original bubble size and reduce bubble coalescence relative to the co-current flow pattern). Markofsky (1976) observed a similar effect with 90 p and 180 p porous diffusers, and concluded at the experimental gas flow rates he was using, there was no significant difference in oxygen transfer efficiency between the two orifice diameters. Given the competing effects of large and small bubbles on rise velocity, contact time, K L and interfacial area, is there an optimum bubble size which provides the highest overall oxygen transfer coefficient? A n examination of a bubble rise velocity vs. equivalent bubble diameter curve (Figure 5.2, from Barnhart, 1969) reveals a distinct pattern and the answer to this ques-tion. Initially, bubble velocity increases linearly with increasing size, to a local maximum TO S O §>30 UJ20 UJ .S 65 1 0 r 8 UJ 3 2 'o2 OA OS 0 8 Q I 0 .2 0 3 OA Q 6 0 . 8 I O ZO 3 O 4 0 6 0 E Q U I V A L E N T " D I A M E T E R , I N C E N T I M E T E R S V E L O C I T Y O F A I R B U B B L E S I N W A T E R A T 20° C Figure 5.2 Graph of bubble rise velocity vs. radius of equivalent spherical bubble (from Barnhart, 1969). velocity, in water, of ~35 cm sec"1 for single bubbles of 1.2-1.4 mm diameter (Haney, 1954). Bubble velocity then decreases from 1.4 mm to 6 mm before increasing again with bubble diameter. The characteristic shape of the curve is due to interactions between hydrodynamic, viscous and interfacial forces acting on the bubbles (Haberman, and Morton, 1954). Small bubbles assume a spherical shape, as surface tension reduces the surface area to a minimum for a given volume. This spherical shape dictates that small bubbles rise according to Stoke's Law, and the viscosity of the liquid is the most important parameter influencing their rise velocity. As bubble size increase, the viscous and hydrodynamic forces acting on the bubble become more dominant, and flattening of the bubble occurs. This results in an oblate spheroid shape, which has a higher drag coefficient than a sphere of the same volume. The rise velocity de-creases accordingly, and the bubbles rise with a rectilinear rocking motion (Eckenfelder, 1959). As bubble size continues to increase, the viscous and surface tension forces become small relative to hydrodynamic forces and the bubble assumes a spheroid cap shape. The upper surface of these bubbles is essentially spherical, and they rise independently of the properties of the liquid (Haberman and Morton, 1954). A n examination of a liquid film coefficient (i.e., K L ) vs. bubble diameter graph reveals that K L increases rapidly to a peak at 2-2.5 mm diameter, then gradual declines with increasing bubble size (Figure 5.3, from Barnhart, 1969). A plot of surface area vs. bubble volume results in an exponential decay type of curve (Figure 5.4). The net result of these interacting effects is a ^ a response curve similar to the K L vs. bubble diameter relationship. The KLa increases sharply to a peak at 2.0-2.5 mm diameter, then declines exponentially with increasing bubble diameter • -COPPOCK 8 MEIKLEJOHN R E L A T I O N S H I P B E T W E E N L I Q U I D F I L M C O E F F I C I E N T A N D B U B B L E D I A M E -T E R Figure 5.3 Graph of K L vs. bubble diameter (all data and graph from Barnhart, 1969). I ' l l L I I I I L_ O .05 .tO .15 .20 .35 .30 .35 .40 .45 SO .55 .SO .65 .70 d B cm -BUBBLE DIAMETER E F F E C T O F B U B B L E S I Z E O N M A S S T R A N S F E R Figure 5.4 Graph of K^a vs. bubble diameter (from Barnhart, 1969). (Figure 5.5, from Barnhart, 1969). Therefore, the optimum bubble size required to maximize KLa is 2.0-2.5 mm diameter. This confirms that the bubble sizes generated by the 140 p, 400 p and 800 p diffuser were slightly larger than the optimum size (see Tables 4.53, 4.54 and 4.55); however, the D B C A configuration generated bubbles in the correct size range (see Tables 4.56 and 4.57). Eck-enfelder and Ford (1968) state that silicon dioxide, aluminum oxide and Saran diffusers gener-ally produce bubble in the 2.0-2.5 mm ranges when operated at optimal gas flow rates, which explains why these diffusers are consider so efficient at oxygen transfer. However, Downing (1966) states " . . . . That practical difficulties, such as the clogging of orifices of air diffusers, limit the smallest size of bubble that it is feasible to produce by air diffusers alone, to a diameter of about 2-3 mm." Therefore, the lower limit of practical bubble size formation from ceramic diffusers with appropriate gas flow rates is also the optimum size for oxygen transfer. 5.4.4 Orifice Diameter and Oxygen Partial Pressure The effect of orifice diameter and oxygen partial pressure (Table 4.31 and 4.32) on K L ^ O , SOTR, S A E and SOTE were similar to the effects observed in the Group 1 to 6 results, and will not be discussed further. 5.4.5 Group 7 to!4 Summary The Group 7 to 14 tests illustrated several additional aspects of gas transfer. A l l of the in situ modifications treatment exerted a negative effect on K La 2o, SOTR, S A E and SOTE. The general trend was that the No Treatment showed the highest values of KLa2o, SOTR, S A E and SOTE (1 s t ranking), the Vanes and D B C A treatments were in the middle range (2 n d and 3 r d ranking) and the Screen and Cover treatments had the lowest values (4 and 5 ranking). This was caused by a reduction in water velocity in the full lift aerator due to physical flow restric-tions in the inlet tubes (screen mesh, counter-rotating vanes); this restricted the height of the bubble:water plume exiting into the separator box (Cover treatment) and caused an upward buoyant force in the outlet tube ( D B C A treatment). The reduction in water velocity by the Vane and D B C A treatments largely accounted for the observed reduction in KLa2o, SOTR, S A E and SOTE. In addition to reducing water velocity, the Screen and Cover treatments exerted addi-tional negative effects on KLa2o, SOTR, S A E and SOTE, likely by degassing oxygen from the mesh screen and decreasing the transfer of oxygen in the separator box. As seen in the Group 1 to 6 experiments, K i ^ o and SOTR responded positively to increased gas flow rates, due to increased interfacial area and turbulence at the liquid interface. S A E and SOTE responded negatively due the reduced ratio of interfacial area and bubble contact time, and the additional energy costs required to compress additional air and PSA 0 2 . A significant gas flow rate x in situ modification interaction was observed for K i ^ o and SOTR with the 140 p and 800 p diffusers, due to the additive effect of gas flow rate on oxygen transfer resulting from additional turbulence and interfacial area. A very noticeable gas flow rate x D B C A treatment interaction effect occurred on S A E and SOTE with the 800 p diffuser, and marginally with the 140 p diffuser, due to velocity related déstabilisation of the D B C A zone in the outlet tube and increasing energy costs associated with the higher gas flow rates. Water velocity with the 800 p diffuser was slower than with the 140 p diffuser, due to increased upward velocity in the D B C A zone in the outlet tube ( D B C A treatment) and increased collisions with the mesh screen and vanes (Screen and Vane treatments). Bubble sizes created by the D B C A configuration were significantly smaller than the Screen and Vane treatments, due to the 2 u. fine pore diffuser used to create the D B C A effect. The effect of reduced orifice diameter, and increased oxygen partial pressure, exerted similar effects on K i ^ o , SOTR, S A E and SOTE, as seen in the Group 1 to 6 experiments. 5.5 Miscellaneous Tests 5.5.1 Pump Only Tests The Pump Only full lift tests in D B C A mode were conducted as curiosity driven research, to determine if there was any advantage to operating a full lift hypolimnetic aerator using circulat-ing water pumps as the prime mover of the system, rather than air lift pumps. The results were quite interesting as the mean values for K i ^ o , SOTR and S A E were among the lowest recorded for any of the full lift tests; whereas, the mean SOTE values were among the highest recorded (Table 4.43). The low values of K i ^ o and SOTR were related to the low gas flow rate (i.e., 3 1 min"1) used in these experiments, which were —1/3 of the gas flow rates used in the lowest gas flow setting for the Group 7 D B C A tests (i.e., 7/3 1 min-1) or the Group 1 tests (i.e., 10 1 min"1). If one multiplies the K i ^ o , SOTR and S A E values from the Pump Only tests by 3.3 (i.e., 10/3) to standardize the gas flow rates, the hypothetical results, for K i ^ o and SOTR on air or PSA O2, appear more efficient per unit gas flow than the D B C A values from the Group 7 tests (Table 4.44) or the Group 1 tests (Table 5.5). However, when one considers the S A E comparison in Table 5.5, the hypothetical Pump Only S A E values on air were only -7-8% of those recorded during the Group 7 and Group 1 tests. Table 5.5. Hypothetical comparison of Pump only, Group 7 and Group 1 values. Parameter Pump only value Pump only x 3.3 Group 7:7/3 1 min"1 D B C A Mode Group 1: 10 1 min"1 K L a 2 o A i r (hr 1) 0.6 2.0 1.5 1.5 SOTR A i r (g 0 2 hr"1) 6.2 20.7 15.8 16.0 S A E A i r (g 0 2 kWhr"1) 8.0 26.7 359.3 349.4 K L a 2 0 P S A O.Oir- 1) 2.7 9.0 7.1 8.4 SOTR P S A 0 2 ( g 0 2 h r 1 ) 29.6 98.7 77.3 91.6 S A E P S A 0 2 (g 0 2 kWhr" ' ) 28.4 94.7 77.1 89.1 The hypothetical S A E values on PSA 0 2 were greater than those recorded during the Group 1 and Group 7 tests. The explanation for this discrepancy is that the actual S A E values obtained during the Group 1 and 7 tests, on air, used considerably less energy to compress and deliver the air, than was required to operate the circulating pumps in the hypothetical standardized Pump Only S A E comparison. In contrast, due to the higher energy requirements of PSA 0 2 genera-tion, the hypothetical S A E values for the Pump Only tests, on P S A 0 2 , were greater than the actual values recorded during the Group 1 and 7 tests, as the circulating pumps used less energy than PSA generated oxygen. 0 This hypothetical comparison, and the fact that the SOTE values recorded during the Pump Only tests were quite high, indicate that a Pump Only configuration may be a reasonable option i f one wishes to design a system that is high in SOTE at the expense of reduced S A E perform-ance. As wil l be discussed in the next section, a Pump Only configuration, with low oxygen flow rates, high pump energy requirements, and a co-current downflow pattern, is the basis of a D B C A system, and is analogous to a rudimentary Speece Cone, which is discussed in Section 5.6. The extended aeration tests were conducted to determine if the calculated oxygen saturation for each test day bore any relation to an actual oxygen saturation value obtained after an extended period of operation. The results from these two tests, admittedly a small sample size, indicate the oxygen saturation achieved after 247 to 270 minutes of operation on air was within 3% of the calculated value (Table 4.45). This indicates that the procedure used to determine the oxygen saturation for each test day (3.2.2) provided a reasonably accurate estimate of oxygen saturation for this series of tests. The extended aeration on PSA O2 was primarily curiosity driven research, to determine how much oxygen could be dissolved into the test water using PSA O2. The highest oxygen concen-tration achieved on PSA oxygen was 50.4 mg l" 1 or 446% saturation after 206 minutes of operation, thus exceeding the C* sfor the day for air (i.e., 103.2 %) by a ratio of 4.33 (Table 4.45). A t this concentration the water appeared milky and the entire surface of the tank was effervescing, as oxygen was degassing from the surface, and no further tests were run. 5.6 Speece Cone Note: since the Speece Cone in these experiments was improperly designed, the following statements apply only to the experimental system. Due to the interrelation between oxygen flow rates and discharge velocity, and the use of different statistical approaches to analyze the data (i.e., Analysis I and II), both oxygen flow rate and water discharge velocity treatments are discussed together, rather than separately as in the Full Lift discussion sections. 5.6.1 Oxygen Flow Rate and Discharge Velocity The oxygen flow rate and discharge velocity treatments exerted a significant effect on K i ^ o and SOTR, with both units of measure showing positive responses to increased oxygen flow rate and discharge velocity. K t ^ o and SOTR increased with oxygen flow rate, as more oxygen was being introduced into the Speece Cone; hence, a larger deficit of driving force (C; -CL) was available, resulting in more oxygen transfer. In addition, the higher oxygen flow rates created more bubbles in the Speece Cone and more interfacial area was available for oxygen transfer, and KLa2o and SOTR responded accordingly. The increased discharge velocity influenced K L ^ O and SOTR by creating more turbulence inside the Speece Cone, thereby increasing K L . The effect of oxygen flow rate and discharge velocity on S A E depended on which analytical method was used. Analysis I indicated there was no effect of oxygen flow rate on SAE. This result was caused by the differential response of the 1 1 min"1 flow rate, which was independent of flow, whereas S A E responded positively with the 2 and 3 1 min"1 flow rates (Figure 4.68c). Analysis II indicated S A E increased with oxygen flow rates, as the 1 1 min-1 treatment was deleted, and the 60 and 70 cm sec"1 discharge velocities were included. It is at these two higher velocities that the effect of the gas flow rates starts to diverge, hence the treatment effect be-comes statistically significant. The effect of discharge velocity on S A E also differed according to which analytical approach was used. Analysis I indicated S A E increased with discharge velocity, whereas Analysis II indicated discharge velocity had no effect of SAE. Since Analysis II linearized the gas flow rate by discharge velocity interaction effect, it has more credibility. The explanation for the re-sponse is that the additional water pump energy required to provide the higher discharge veloci-ties offset the additional dissolved oxygen gained by the greater turbulence inside the Speece Cone. In addition, since increased discharge rates would act to decrease residence time in the Speece Cone, bubble:water contact time would decrease, which would also act to nullify any gains in S A E resulting from increased discharge velocity. SOTE increased with water discharge velocity regardless of which analytical method was used. This indicates the additional turbulence caused by the increased discharge velocities more than compensated for the reduced bubble :water contact time in the Speece Cone, and the net result was an increase in SOTE. The additional turbulence likely acted in concert with the higher gas flow rates to increase bubble interfacial area, which would also act to increase SOTE. Analysis I indicated a significant oxygen flow rate by discharge velocity interaction, whereas Analysis II did not, hence the effect is not dominant and depends on which data set is analysed. SOTE responded negatively to increasing oxygen flow rates in both analyses. The highest SOTE values were recorded with the lowest oxygen flow rates (except at discharge velocities > 50 cm sec"1), and the effect was consistent at all discharge velocities. The highest SOTE value recorded was 63% with the 1 1 min"1 oxygen flow rate at 50 cm sec"1. This suggests that oxygen flow rates > 11 min"1 were not being fully dissolved inside the Speece Cone, and thus oxygen gas was being wasted. Oxygen bubbles could been seen leaving the discharge port of the Speece Cone under higher gas flow rates. This is another example of the difficulty in optimis-ing gas transfer with competing units of measure. The highest SOTE of 63% was an artifact of inadequate inlet velocity which could not prevent the bubble swarm from collapsing. Properly designed Speece Cones should achieve > 98% oxygen absorption efficiency (pers. Comm, Dr. Richard Speece, Vanderbuilt University). The explanation for the precipitous decline in SOTE at low oxygen flow rates and discharge velocities > 50 cm sec"1 (Figure 4.73) is the same effect that was observed with the D B C A configuration in 5.4.1. Discharge velocities > 50 cm sec"1 destabilised the bubble swarm in the Speece Cone, and forced the oxygen bubbles out the discharge port, thus collapsing the D B C A effect and triggering the precipitous decline in SOTE (and K i ^ o , SOTR and SAE). At oxygen flow rates > 1 1 min"1, there was sufficient positive buoyancy inside the Speece Cone from the additional oxygen bubbles such that discharge velocities > 50 cm sec"1 were not able to purge the bubble swarm. Therefore, the D B C A effect was not destabilised and was able to exchange oxygen, with SOTE values exceeding those recorded at lower discharge velocities (even though many excess oxygen bubbles were purged from the Speece Cone by the higher discharge velocities). The additional turbulence, increased C J - C L and greater interfacial bubble area created by the higher discharge velocities more than compensated for the increased wasting of the air bubbles, and continued to increase SOTE. Operationally, this would present a problem as the mass wasting of bubbles may destratify a lake or reservoir. For this reason, full size Speece Cones must have bubble collector hoods located downstream of the main cone, to collect stray bubbles and vent them to the surface through a discharge pipe, to avoid uninten-tional destratification. Properly designed Speece Cones should discharge very few excess bubbles. 5.6.2 Speece Cone on Ai r The single test of the Speece Cone on air was curiosity motivated research, basically to see what type of performance was possible on air, with an oxygenation system designed to operate on oxygen. The results of the air test were lower than any recorded on PSA oxygen for K j ^ o , SOTR, S A E and SOTE (see Table 4.73). It is not possible to directly compare the performance of the Speece Cone on air with the full lift tests, as the maximum air flow rate to the Speece Cone was ~l /3 of the minimum air flow rate to the full lift aerator. However, it is instructive to compare the relative performance of the Speece Cone on air at 3 1 min"1 and 50 cm sec"1, with the nearest analogue in the full lift configuration to demonstrate the relative difference in performance. In this case, the full lift treatment which crudely approximated a Speece Cone was the Group 8 configuration (i.e., 800 p diffuser, 2.9 m diffuser submergence) full lift D B C A test, running on air with a 7/3 1 min' 1 split between the main diffuser and the separator box diffuser (Table 5.6) Table 5.6. Adjusted least square treatment means (± SE) comparison between Speece Cone on air at 3 1 min"1 and 50 cm sec"1, with the Group 8 (i.e., 800 p diffuser, 2.9 m diffuser submergence) full lift DBCA test, running on air with a 7/3 1 min"1 allocation between main and separator box diffuser. Treatment K L a 2 0 (hr 1) SOTR (S 0 2 hr 1 ) SAE (g 0 2 kWhr 1 ) SOTE(%) N 3 1 min"1 air at 50 cm sec"1 Speece 1.2 (0.01) 13.4 (0.04) 16.9 (0.05) 24.5 (0.08) 4 7/3 1 min" air Full Lift 0.88 (0.06) 9.6 (0.66) 210.2(14.49) 5.10(0.35) 3 The comparison indicates that even at the 1/3 the air flow rate being delivered to the full lift aerator, the Speece Cone on air delivered superior performance for KLa 2o, SOTR, and particu-larly SOTE. The S A E performance was very low, as the higher energy cost required to operate the water pump penalized this key energy related unit of measure. The relative performance of the Speece Cone on air was therefore similar to its performance on oxygen: good K.La2o and SOTR response, low S A E performance, and very good SOTE. For this aeration system, the two most important units of measure for comparing aeration systems (SAE and SOTE) were in contrast to each other, demonstrating the paradox of optimizing oxygen transfer. In practice, a Speece Cone would rarely be operated on air as this would forfeit the higher performance available on oxygen; there would also be concerns about nitrogen supersaturation causing fish mortalities, particularly in salmonids (Rucker, 1972). 5.6.3 Speece Cone Summary K i ^ o and SOTR responded positively to increased oxygen flow rate and discharge velocity, due to a larger deficit of driving force ( Q - C L ) , greater interfacial area and more turbulence inside the Speece Cone. The S A E response depended on which analytical approach was used; however, the general conclusion was that oxygen flow rates > 1 1 min"1 increased S A E . S A E appeared to be independent of discharge velocity as the higher pump energy costs nullified any gain in S A E . SOTE was positively influenced by discharge velocity, as the additional internal turbulence and interfacial area compensated for the reduced bubble:water contact time. SOTE was negatively influenced by oxygen flow rates, as the highest SOTE values were recorded with the lowest oxygen flow rates (except at discharge velocities > 50 cm sec"1). SOTE declined precipitously at oxygen flow rates < 1 1 min"1 and discharge velocities > 50 cm sec"1, as the D B C A bubble swarm in the Speece Cone was destabilised, and forced out the discharge port. Full size Speece Cones must have bubble harvesters to collect stray bubbles and prevent unintentional destratification. A single test on air revealed significantly reduced per-formance relative to an equivalent flow rate of oxygen. However, even on air, the Speece Cone delivered superior K i ^ o , SOTR, and particularly SOTE performance, when compared with a full lift aerator operating on 10 1 min"1 air at the same diffuser depth. In practice, a Speece Cone would rarely be operated on air, for reasons outlined previously. 5.7 Comparison of Relative Performance of Full Lift vs. Speece Cone As previously stated, direct comparisons between the performance of the Full Lift and Speece Cone systems are not possible, as they were not operated at comparable gas flow rates or oxygen partial pressures. In addition, the Speece Cone was not properly scaled-down; hence the performance of the experimental Speece Cone was less than optimal. The design principle of a Speece Cone is to provide a downward water velocity of -305 cm sec"1 at the top of the cone. This high velocity prevents bubbles from escaping out the top of the cone, and the high velocity inlet stream provides the turbulence to maintain the bubble swarm in a dynamically regenerated state, with an exceptionally high gas surface to water volume ratio; this favors high oxygen transfer rates (Speece, 1992). As the cross sectional area of the Speece Cone increases, the net downward velocity of the water decreases to a design velocity of ~ 9 cm sec"1 near the outlet of the cone. Since the buoyant velocity of smaller bubbles is ~ 35 cm sec"1 (see Figure 5.2), the oxygen bubbles are maintained in contact with the water inside the Speece Cone indefinitely. In addition, the relative volumetric flow ratio of oxygen to water should be about 3%, to efficiently dissolve oxygen (Speece, 1992). The Speece Cone used in these experiments did not have sufficient velocity at the top of the cone, and the cross sectional velocity at the base of the cone was too low (Table 5.7). As a result, the dynamic bubble swarm never properly developed inside the cone, but remained as a gas pocket at the top of the cone that the inlet water dropped through, resulting in a drastically reduced gas surface area to water ratio. Therefore, all of the performance measures from this experimental Speece Cone are much lower than would be expected with the correct inlet velocity of 305 cm sec"1 and correct cross sectional velocity at the base of the cone. Table 5.7. Comparison of experimental Speece Cone inlet water velocities and oxygen water ratios to design specifications. Inlet water Water flow Oxygen flow Base cross-sectional 02:water ratio velocity (cm sec"1) (1 min"1) (1 min 1) velocity (cm sec"1) (%) 20 38 1 to 3 0.7 2.6-7.9 30 57 1 to 3 1.0 1.8-5.3 40 76 1 to 3 1.4 1.3-3.9 50 95 1 to 3 1.7 1.1-3.2 60 114 1 to 3 2.1 0 .9-2 .6 70 133 1 to 3 2.4 0 .8-2.3 305 (design velocity 9 (design velocity) 3 (design ratio) However, relative comparisons, qualitative in nature, are instructive for examining the oxygena-tion performance of the two systems; it must be realized that the oxygenation performance of the Speece Cone was much lower than expected, due to inlet water velocities that were too low. In these experiments, the most comparable data sets are those of the Speece Cone on PSA oxygen at 1 or 3 1 min"1 and a discharge velocity of 50 cm sec"1, with the Group 2 full lift tests, on air or PSA oxygen at 10 1 min"1 with an 800 p diffuser (Table 5.8). (Note: the inlet velocity in the Speece Cone is equivalent to the discharge velocity, as the diameter of the inlet and outlet piping was similar, i.e., 6.35 cm). Table 5.8. Adjusted least square treatment means (± SE) comparison between Speece Cone on PSA O 2 at 1 and 3 1 min"1 and 50 cm sec"1, with the Group 2 (i.e., 800 p diffuser, 2.9 m diffuser submergence) full lift test, on air or PSA O 2 . Treatment K L a 2 0 (hr"1) SOTR (g 0 2 hr 1 ) SAE (g 0 2 kWhr"1) SOTE(%) N 11 min"102 at 50 cm sec"1 Speece 4.6 (0.09) 49.8 (0.98) 55.7 (1.09) 63.4(1.23) 4 3 1 min' 10 2 at 50 cm sec"1 Speece 8.7 (0.20) 94.6 (2.14) 86.9(1.96) 40.1 (0.87) 4 10 1 min"1 Full Lift on 0 2 6.4 (0.55) 69.7 (5.97) 70.63 (6.04) 8.97 (0.67) 3 10 1 min 1 Full Lift on Air 1.0 (0.03) 11.3 (0.34) 262.2 (7.88) 5.98 (0.18) 3 The data clearly show the different oxygenation capabilities of the two systems. The Speece Cone, at 1 1 min"1, delivered K L a 2 0 , SOTR and S A E values that were -70-79% of the Full Lift system operating on oxygen, and 460-480% of the K i ^ o and SOTR values of the Full Lift system operating on air, but using only 1/10 of the gas flow rate. The Speece Cone response is not linear, as the values at 3 1 min"1 are not three times that recorded at 1 1 min"1. The S A E of the Speece Cone is considerably less then the Full Lift system on air, whereas the SOTE of the Speece Cone is very high at the lowest gas flow rate. The Speece Cone SOTE was still > 40% at the highest gas flow rates, even though it was obvious that considerable oxygen gas was being wasted from the discharge port at the higher gas flow rates. Therefore, in relative terms, the Speece Cone, at oxygen flow rates between 1 and 3 1 min"1, delivered K i ^ o and SOTR performance comparable to the Full Lift system at 10 1 min"1 on oxygen, and superior performance to the Full Lift system on air. This indicates the Speece Cone design provides increased interfacial area (i.e., a) and a highly turbulent gas exchange environ-ment (KL) that facilitate oxygen transfer. The Speece Cone also has the benefit of increasing the oxygen differential (CJ-CL), as the units are typically located at the lake or reservoir bottom, hence utilizing the ambient hydrostatic pressure to maximize oxygen differential. The most outstanding feature of the Speece Cone design is the SOTE performance, as values > 60% have not been achieved with the full lift design on air or oxygen. This supports the previous conclu-sion that the D B C A counter-current flow principle of the Speece Cone provides a superior gas transfer environment by enhancing the fundamental aspects of gas transfer (i.e., a, K L and Q-C L ) . In these experiments, with an incorrect scale-down, the Achilles heel of the Speece Cone was its S A E performance. Although, the S A E performance of the Speece Cone was comparable to the S A E of the Full Lift system on oxygen, the S A E performance was considerably less than the Full lift design on air. The S A E performance increased significantly as gas flow rate increased from 1 to 3 1 min"1; however, the SOTE declined from > 60% to 40% as the gas flow rate exceeded the dissolution rate in the Speece Cone, thus offsetting the improved S A E perform-ance. However, as previously mentioned, the low S A E performance of this Speece Cone was a reflection of this particular design. Given this, two factors contributed to the low S A E perform-ance: (1) the energy cost of PSA oxygen, and (2) the energy cost of pumping water. Since the Speece Cone is designed to operate on oxygen, the energy costs of delivering oxygen to the unit are considerably higher than delivering compressed air. This topic wil l be discussed in Section 5.10, entitled "PSA Engineering Design Factors Influencing S A E " . The other reason for the reduced S A E performance is the energy costs related to pumping large volumes of water. Unlike the air-lift pump principle of the Full Lift design, which provides oxygen and circulates the water, the Speece Cone requires an additional energy input to circu-late the water though the mixing chamber and out the diffuser. The combination of the addi-tional energy required for oxygen generation and water circulation result in reduced S A E performance, relative to the Full Lift design on air. Since the systems were not operated at comparable gas flow rates, and the Speece Cone was incorrectly scaled-down, any further comparison of the oxygenation performance of these two designs from this data set becomes hypothetical. This indicates the need for direct experimental evaluation of both designs under equivalent gas flow rates. Nonetheless, even an incorrectly sized Speece Cone delivered superior oxygenation performance to the full lift design, in most of the performance compari-sons. 5.8 Comparison with Civil Engineering Studies Although numerous engineering studies have examined various aspects of oxygen transfer, relatively few have systematically examined the effect of diffuser depth, orifice diameter, gas flow rate, and oxygen partial pressure on the four standard unit of measure in an experimental system that is comparable in scale to the present study. Fortunately, the Ph.D. work of Mavinic (1973) and subsequent publications (Mavinic and Bewtra, 1974; Mavinic and Bewtra, 1976) show reasonable similarity, such that performance comparisons can be made. Their study examined the effect of diffuser depth, air flow rate, and circulation pattern on K i ^ o , SOTR, S A E and SOTE (Table 5.9). The configuration, which was most similar to the present study, was System II, the co-current upflow with no water pump. The Mavinic (1973) study used a coarse bubble diffuser (1,600 p diameter) at a similar maximum depth (2.7 vs. 2.9 m), with a smaller diameter tube (10 vs. 19.5 cm) and lower air flow rates (2.8-10.7 vs. 10.5-42.1 1 min"1). By coincidence, the unit air flow rate was almost identical in both systems. The K i ^ o and SOTE values obtained from each experimental system were quite similar; indicating the performance of the current system is comparable to other designs. However, the S A E performance obtained from the current study was considerably lower than the Mavinic (1973) study. The explanation is likely related to the compression ratio in the adiabatic compression formula (ASCE, 1992) used to calculate the energy requirements of the system. This current study operated at a higher pressure than would Table 5.9. Comparison of experimental system of Mavinic (1973) and this study, Full Lift only on air, Groups 1-6. Parameter Mavinic System I Mavinic System II Mavinic System III Mavinic System IV Full lift -Present Study Diffuser depth (m) 0.6-2.7 0.6-2.7 0.6-2.7 0.6-2.7 1.5,2.9 Riser diameter (cm) 10 10 10 10 19.5 Orifice diameter (p) 1,600 1,600 1,600 1,600 140, 400, 800 A i r flow rate (1 min"1) 2.8-10.7 2.8-10.7 2.8-10.7 2.8-10.7 10.5-42.1 Unit air flow rate (1 min"1 cm"2) 0.036-0.136 0.036-0.136 0.036-0.136 0.036-0.136 0.035-0.141 Circulation pattern Single column Ai r , co-current upflow A i r and water pump, counter current flow A i r and water pump, co-current downflow Ai r , co-current upflow Water pump capacity (1 sec"1) n/a 6.3 6.3 6.3 n/a Ki^oOi r" 1 ) 7.8-25.3 0.8-2.4 3.2-7.8 3.0-9.9 0.6-4.8 SOTR (g 0 2 hr"1) n/a n/a n/a n/a 6.8-51.9 S A E ( g 0 2 kWhr"1) 1,202-1,655 1,275-2,971 1,759-2,766 861-1,520 125-349 SOTE (%) 0.7-4.0 2.0-4.9 5.8-17.6 7.4-19.0 2.9-8.5 be required to overcome frictional line losses and hydrostatic pressure, hence this penalized the S A E performance of the system. The sensitivity of S A E performance to compression ratios wil l be discussed in more detail in Section 5.10. 5.9 Linkage to Hypolimnetic Aeration Studies As best as can be determined, this is the first experimental evaluation of a bench-scale, Full Lift and Speece Cone hypolimnetic aerator; hence there is no published data available on similar bench-scale systems for performance comparison. K i ^ o and SOTR data for full-scale systems are not obtainable, as lakes or reservoirs have variable hypolimnetic oxygen concentrations and biochemical oxygen demands, and it is not feasible to conduct non steady-state reaeration tests in a lake or reservoir. However, it is possible to use data collected from field measurements to derive preliminary estimates of oxygen transfer and energy efficiency that are analogous to SOTE and S A E . The data available from the literature show a wide range in aeration efficiency (AE) and transfer efficiency (TE) for full lift and partial lift systems (Table 5.10), although published performance data on the Speece Cone was non-existent. The highest value obtained for a full lift system was from the Wahnbach II system (Bernhardt, 1974), which had an A E of 1.1 kg O2 kWhr"1 and TE Table 5.10. Estimates of A E and T E from full scale hypolimnetic aeration systems. (Note: A E and T E are not derived from the non steady-state procedure). Speece Cone1 operated at 22.9 m, and Speece Cone2 surcharged to equivalent depth of 39.6 m. Hypolimnetic A E (kg 0 2 kWhr 1 ) T E Influent 0 2 Gas Reference aerator (%) concentration design (mgl 1) Full Lift 0.95 50 <4 A i r Bernhardt, 1967 Full Lift 0.77 46 0 A i r Fast, 1971 Full Lift 1.10 50 <4 A i r Bernhardt, 1974 Full Lift 0.25 10 0 A i r Hess, 1975 Full Lift 0.32 21 0 A i r Smith et al., 1975 Full Lift 0.32 10 2-7.5 A i r Smith et al., 1975 Full Lift 0.43-0.52 24-28 n/a A i r Wirthet al., 1975 Full Lift 0.05-0.41 10-49 0.3-4.7 A i r Taggart and McQueen, 1981 Full Lift 0.09 6 0.4-6.7 A i r Ashley, 1983 Full Lift 0.50-0.56 15-17 0.8 A i r Ashley et al., 1987 Partial Lift 0.54 10.6 4 Ai r Fast et al., 1975 Speece Cone 1 3.2 86 n/a 0 2 Speece, 1992 Speece Cone 3.1 100 n/a 0 2 Speece, 1992 of 50%, whereas the lowest performance was an A E of 0.09 kg O2 kWhr"1 and TE of 6% for the Black Lake system (Ashley, 1983). (Note A E and TE indicate units of measure that have not been derived from non steady-state oxygen transfer tests. A E is kg oxygen delivered per hour divided by the nameplate power in kW). One data set was available on a Partial Lift system, which had an A E of 0.54 kg 0 2 kWhr"1 and TE of 10.6 % (Fast et al., 1975). The information on the Speece Cone was taken from an internal report on the proposed Camanche Reservoir system (Speece, 1992). The design A E of this system is 3.1-3.2 kg 0 2 kWhr"1 and the TE is 86 to 100%. These values exceed any reported oxygenation values for full lift and partial lift systems by a very wide margin, and illustrate the superior oxygen transfer capabilities of the Speece Cone design. However, it soon becomes evident that relative comparisons between different studies are of limited value, due to the wide range in environmental and engineering variables capable of influencing oxygen transfer. For example, the concentration of influent oxygen ranged from 0 to 6.7 mg T1, diffuser submergence varied from 4.6 to 42 m, and some systems were degassing H 2 S , which would impart an immediate chemical oxygen demand and negatively influence A E and TE performance. As discussed earlier, the concentration gradient ( Q - C L ) exerts a signifi-cant influence on oxygen transfer, as does hydrostatic pressure. Therefore, in order to derive meaningful comparisons between systems, the different designs need to be studied at a bench-scale in a laboratory setting, as in the current study, or examined in the same body of water under similar conditions. A useful observation from the field studies is that oxygen transfer did not occur uniformly in the full lift systems during ascent through the water column, but the majority of oxygen transfer occurred in the lower portion of the inlet tubes. The same effect was noted in this study (Sec-tion 4.4) as the majority of the oxygen transfer occurred in Zone 1 when diffuser submergence was at 2.9 m, and in Zone 2 when diffuser submergence was at 1.5 m. For example, Bernhardt (1967) noted most oxygen transfer took place in the first 7.3-20.1 m of a 35.9 m bubble rise distance (Table 5.11). Table 5.11. Decrease of the oxygen content in influent air at various depths during bubble rise to the surface (from Bernhardt, 1967). Test dates O 2 volume in O 2 volume at O 2 volume at 0 2 volume at influent air 36.6 m 15.2-20.1 m 0.9 m (%) 8/6/62 20.8 20.4 16.8 16.8 9/3/62 20.8 20.2 16.5 15.8 10/21/62 20.8 20.4 15.5 15.2 Average value 20.8 20.3 16.3 15.9 Smith et al. (1975) observed a similar trend in Mirror Lake, where they measured inlet tube dissolved oxygen concentrations of 0.7 mg l" 1 at 3.0 m, 2.3 mg l" 1 at 6.1 m, but no further increase during the remaining 6.1 m ascent to the surface. These observations suggest that although hydrostatic pressure increases the driving force (i.e., Q - C L ) for oxygen transfer, additional factors modify the magnitude of the effect of hydrostatic pressure. Speece (1975) suggests that declining hydrostatic pressure, decreasing oxygen content of the rising oxygen bubbles and the additive effect of bubble and water velocity in a co-current upflow system are responsible for this effect. Downing (1966) and Mavinic and Bewtra (1974) indicate that the interstitial liquid rising in a confined column with dense bubble clouds becomes saturated with dissolved oxygen very quickly. This reduces the concentration gradient across the liquid film and increases the resistance to lateral diffusion of oxygen, thus negatively influencing K L . This suggests that for Full Lift designs, a key factor for increasing oxygen transfer performance is to optimize diffuser design, as the conditions for gas transfer become progressively less favorable once the bubbles detach from the surface of the diffuser and ascend through the water column. A n examination of bubble formation behavior, in which there are three distinct phases with three different rates of gas transfer, lends support to this conclusion. The first phase occurs during bubble formation at the interstitial space or orifice openings of the diffuser. During formation and growth of the bubble, the liquid-gas interface surrounding the bubble is continu-ally expanding and the concentration gradient in the liquid film remains high, resulting in an unusually high rate of oxygen transfer (Ippen and Carver, 1954). In the second stage, known as the intermediate steady-state, the concentration gradients in the surrounding liquid film attain lower values, hence a reduced transfer of oxygen occurs during the bubble's ascent through the water column. This phase is subject to deviations from the steady-sate, depending on shear and turbulence in the water column (Mancy and Okun, 1960). In the final phase, the bubble bursts when it reaches the surface and releases its gas contents to the surface. The liquid film that surrounded the bubble and contained relatively high concentrations of dissolved oxygen is left behind, and the disturbance of the interface by the bursting bubbles tends to enhance oxygen exchange with the atmosphere at this stage. Therefore, the increase in transfer efficiency arising from smaller orifices occurs during the bubble formation stage, prior to the bubbles coalescing into a heterogeneous mixture. Although the period of bubble formation is quite brief, it is sufficient to allow for increased oxygen transfer. Pasveer (1955) states that the liquid film becomes saturated with oxygen in a time span of 1 x 10"7 seconds, which is considerably faster than the time required for bubble formation. As noted in Section 5.5.3, the overall effect of a decrease in orifice diameter was an increase in all units of measure. Therefore, as demonstrated in these experiments, one area by which full lift hypolimnetic aeration systems can increase their oxygenation performance is by using diffusers with pore diameters of-140 p. Since oxygen transfer in co-current full lift designs is limited by declining hydrostatic pressure, decreasing oxygen content of the rising oxygen bubbles and the additive effect of bubble and water velocity, a highly efficient diffuser is an important design consideration for full lift systems. 5.10 PSA Engineering Design Factors Influencing SAE The effect of PSA oxygen on full lift KLa2o and SOTR performance, relative to air, was quite spectacular (Section 5.3.2); however, S A E performance was penalized by the additional energy requirements of the PSA process. As previously indicated, a variety of factors in addition to oxygen partial pressure influenced the standard aeration efficiency (SAE) of the experimental full lift system (gas flow rate, diffuser depth and orifice size). However, there are at least two engineering design factors external to the oxygen transfer process which are key to improving the S A E of aeration systems using PSA generated oxygen. These include (1) PSA oxygen recovery efficiency and (2) the feed air pressure requirements of the PSA system. If one examines the air input/oxygen output ratios and resulting oxygen recovery efficiencies of PSA units manufactured by Oxygen Generating Systems Inc. (Niagara Falls, New York), a clear trend emerges. The air input/oxygen output ratio decreases and the oxygen recovery efficiency increases from 31.1% to 46% as the units increase in capacity from their Model O G 25 to the O G 1000 (Table 5.12). Table 5.12. Air input/oxygen output ratio and oxygen recovery efficiencies of selected OGSI Inc. PSA units. Model Feed air O2 output Air input/02 O2 recovery No. Output ratio efficiency Nm J h r 1 Nm 3 h r 1 % O G 25 10.2 0.7 15.5 31.3 O G 50 19.2 1.3 14.8 32.7 O G 75 27.0 2.0 13.5 35.8 O G 100 34.8 2.6 13.4 36.1 OG175 50.4 4.6 11.0 44.1 O G 250 70.8 6.6 10.7 45.0 O G 500 138.6 13.1 10.6 45.7 O G 750 210.0 19.7 10.7 45.3 O G 1000 276.0 26.3 10.5 46.0 PSA systems cannot be 100% efficient, as every nitrogen molecule in the feed air does not contact with a molecular sieve particle, and pressurized dry oxygen generated via the PSA process is required to purge saturated nitrogen from the sieves on each depressurization cycle. This oxygen is lost from the process, which places a theoretical, and practical upper limit on the oxygen recovery efficiency of PSA technology. The efficiency of the PSA units increases with unit capacity, as smaller PSA units are not designed solely for optimum oxygen recovery efficiency. The most important aspect of small units is that they produce oxygen for a given application within a cost effective design package. Since their energy use is small, less design attention is focused on optimizing oxygen recovery efficiency. However, when unit capacity increases, energy costs and oxygen recovery effi-ciency become significant operational concerns and the design of the PSA systems is modified to recover as much oxygen as possible (pers. comm., Joe McMahon, OGSI, Chief Engineer, Niagara Falls, N Y ) . For example, the small OGSI system (OG 25) is 31% efficient at recover-ing oxygen, whereas the largest unit (OG 1000) is 46% efficient (Table 5.12). This is achieved by increasing the density of the molecular sieve beds, optimizing the sieve surface area to air contact ratio, and other proprietary design features. A n example of the importance of proprietary design features can be seen by comparing the air input/oxygen output ratio and oxygen recovery efficiency of similar sized PSA units from different manufacturers. Both systems shown in Table 5.13 deliver identical amounts of oxygen at similar purity and air feed pressure (i.e., 90-95% 0 2 @ 6.3 kg cm"2); however, air in-put/oxygen output ratio and oxygen recovery efficiency differ by 19%. Table 5.13. Performance comparison of equivalent capacity PSA systems. Model No. Feed air o 2 output Air input/02 Output ratio 0 2 recovery efficiency Nm3/hr NmW % AirSep 1000 328.2 26.29 12.5 38.7 OGSI1000 276.0 26.30 10.5 46.0 One can then examine the effect of variable PSA oxygen recovery efficiency on the S A E of a given aeration system. The calculated improvement is a direct result of the reduced weight of air used in the standard adiabatic compression formula, which determines the energy component of SAE. For example, the mean S A E from each gas flow rate in the Group 1 PSA oxygen full lift experiments was re-calculated using the range of oxygen recovery efficiencies listed in Table 5.12. The results illustrate that the S A E increases in direct proportion to the oxygen recovery efficiency, and that significant improvements (i.e., 46%) in S A E can be expected when using larger capacity PSA systems (Table 5.14). Table 5.14. Calculated effect of variable PSA oxygen recovery efficiency on Group 1 Full Lift SAE performance. Air input/02 Output ratio O2 recovery efficiency SAE (g 0 2 kWh"1) SAE (g 0 2 kWh"1) SAE (g 0 2 k W h 1 ) SAE (g 0 2 kWh"1) % @ 101 min"1 @ 20 1 min"1 @ 30 1 min"1 @ 401min 1 15.5 31.3 91.3 89.1 79.6 65.1 14.8 32.7 95.6 93.3 83.3 68.2 13.5 35.8 104.6 102.2 91.3 74.8 13.4 36.1 105.4 103.0 91.9 75.3 11.0 44.1 127.8 125.1 111.8 91.7 10.7 45.0 131.3 128.6 115.0 94.2 10.6 45.7 132.5 129.8 116.0 95.1 10.7 45.3 131.3 128.6 115.0 94.2 10.5 46.0 133.8 131.0 117.1 96.0 The pressure requirements of the PSA generator also significantly influences the S A E of a hypolimnetic aeration system as a result of the energy required for feed air compression. One can use this information to examine the theoretical effect of reduced PSA feed air compression ratios on the S A E of a given aeration system. The calculated improvement is a direct result of the reduced compression ratio energy used in the standard adiabatic compression formula (Metcalf and Eddy Inc., 1991). To illustrate, the mean S A E from each flow rate in the Group 1 PSA full lift experiments were re-calculated with compression ratios varying from 6.33 to 0.78 kg cm"2, using three oxygen recovery efficiencies (i.e., 31.3, 36.1 and 46.0%) listed in Table 5.12. The results illustrate that S A E increases dramatically in relation to the feed air compres-sion ratio, and that major improvements in S A E could be achieved by using lower pressure PSA generators (Table 5.15). As discussed in Section 5.8, unnecessarily high compression ratios were likely responsible for the lower S A E performance, on air, of this study relative to the S A E values reported by Mavinic and Bewtra (1976) for their System II co-current upflow configura-tion. Table 5.15. Calculated effect of variable PSA feed air compression ratios on Group 1 full lift SAE results using three ranges of PSA oxygen recovery efficiency. O 2 recovery Feed air SAE SAE SAE SAE efficiency pressure (g 0 2 kWh 1 ) (g 0 2 kWh 1 ) (g 0 2 kWh"1) (g 0 2 kWh"1) % kg cm"2 @ 10 I min 1 @ 20 1 min"1 @ 30 1 min"1 @ 40 1 min"1 31.3 6.33 91.3 89.1 79.6 65.1 31.3 4.92 104.1 101.7 90.8 74.4 31.3 3.52 124.7 122.1 109.1 89.4 31.3 2.11 166.1 163.1 146.0 119.8 31.3 0.70 309.9 308.5 277.3 228.0 36.1 6.33 105.4 103.0 91.9 75.3 36.1 4.92 120.0 117.4 104.9 86.0 36.1 3.52 143.8 141.0 126.0 103.3 36.1 2.11 191.2 188.3 168.6 138.3 36.1 0.70 355.4 355.2 319.8 263.1 46.0 6.33 133.8 131.0 117.1 96.0 46.0 4.92 152.3 149.4 133.6 109.6 46.0 3.52 182.2 179.2 160.5 131.6 46.0 2.11 241.7 239.1 214.5 176.1 46.0 0.70 445.7 449.2 405.7 334.2 In recognition of these potential efficiency gains, AirSep Corp. manufactures a low pressure PSA system, optimized for energy use, which generates oxygen at 1.05 kg cm"2. This model is more energy efficient than their standard pressure units, which generate oxygen at 3.16 kg cm' 2 , essentially by saving energy during the compression cycle. A relatively recent development in oxygen generation technology which takes this concept even further is the Vacuum Swing Adsorption process (i.e., V S A ) , which generates oxygen at 1.55 kg 2 2 cm" and regenerates the molecular sieve in a mild vacuum (0.52 kg cm" ). The rationale for adopting this approach was recognition that the majority of energy costs in standard PSA generators arise from the need to compress large volumes of air at high pressure. By lowering the operating pressure, the entire process becomes more energy efficient. The oxygen recovery efficiency of the molecular sieve is identical to higher pressure units. V S A generating systems (without pressure boosters) are restricted to low pressure applications; however, this develop-ment emphasizes the importance of minimizing energy requirements for the energy efficient oxygenation of water. Therefore, the lower S A E performance on PSA generated oxygen, relative to air, can be substantially improved by utilizing PSA generators with higher recovery efficiencies and lower operating pressures, this enables hypolimnetic aeration systems to provide the highest possible oxygenation performance, without being unnecessarily penalized with reduced Aeration Efficiency (i.e., AE) . As will be discussed in the following section, aeration systems are not selected on A E performance alone. The ability to deliver greatly enhanced oxygenation performance on PSA generated oxygen, using full lift or Speece Cone designs, will likely emerge as the key technological advancement in the field of hypolimnetic aeration. 5.11 Comments on Full Scale Application of Full Lift and Speece Cone Systems One of the most challenging aspects of engineering is the successful transition from bench-scale study to full-scale application. The proper design, installation and operation of hypolimnetic aeration systems are no exception. Given the robustness and consistency of the results obtained in this study, the relative magnitude of the effects should be transferable to full-scale applica-tion, within reasonable limits. For example, the relative effects of diffuser orifice diameter, gas flow rate, oxygen partial pressure and diffuser submergence depth on oxygen transfer should be similar in a full-scale full lift system. Similarly, the excellent oxygenation performance of the Speece Cone wil l occur in full-scale systems, provided that the inlet and outlet water velocities meet realistic design specifications. However, factors that could not be examined at the bench-scale, such as extreme depth, wi l l change the absolute magnitude of some units of measure (i.e., SAE) as a result of the additional energy require to deliver air to greater depths. A more difficult issue is how to select the most appropriate design of a hypolimnetic aerator for a particular installation. A general guide on destratification and hypolimnetic aeration tech-niques was published in the ecological literature in the mid 1970's (Lorenzen and Fast, 1977); however, it requires updating given the recent technological developments (e.g., PSA generators and Speece Cone). In addition, there is no counterpart in the civil engineering literature; hence, civil engineers have a minimal amount of information from which to compare and select systems. Companies who actively market hypolimnetic aeration systems, some with more enthusiasm than scientific credibility or engineering expertise, compound this paucity of infor-mation. The author has personally adjudicated, on more than one occasion, grievances between suppliers and various levels of government, when commercially purchased hypolimnetic aeration systems failed to deliver their advertised level of oxygenation performance. How then, given the paucity of comparative information on the subject, does one select an appropriate hypolimnetic aeration system for a particular project? The short answer is that there is no simple answer. Each installation wil l have different priorities, different infrastructure requirements, and variable amounts of pre-existing data from which to design a system. As has been noted throughout this study, some factors (e.g., orifice diameter) positively influenced all units of measure, while other design factors (e.g., gas flow rate) positively influenced KLa2o and SOTR, while negatively influencing S A E and SOTE. Therefore, in each installation, one unit of measure wil l likely be de-emphasized in order to achieve specified performance targets in another unit of measure. In reality, there are many factors unrelated to gas transfer that play a critical role in system selection. These include obvious factors such as noise level, delivery dates, capital and operating costs, to less quantifiable issues such as esthetics, liability concerns and local politics. There are at least five steps required to assist in resolving the engineering aspects of system selection: 1. Lab Testing - comparative testing of different designs under similar laboratory condi-tions (as per this study); 2. Field testing - comparative testing of different designs under actual field conditions with similar conditions of depth, gas flow rate etc.; 3. Developing predictive models - once sufficient comparative laboratory and field information has been collected, the data can be used to quantify the underlying physical processes and develop predictive models that can be evaluated against systems that were not part of the original data set (to avoid circularity of modeling); 4. Reporting the results in engineering journals and conferences - obvious, but necessary to educate engineers unfamiliar with the process, and the extent to which lake and reser-voir pre-treatment with hypolimnetic aeration can be a viable alternative to conventional in-plant unit processes to improve raw water quality; 5. Case study reviews of specific installations - although some systems may have gone through an initial screening process, very few installations have been reviewed after a few years of operation to examine the systems' actual performance, and the appropriate-ness of the original decision. For obvious reasons this wil l be a difficult task; however, performance audits are particularly useful in reviewing the quality of the original per-formance specifications and decision making process, and would be beneficial in resolv-ing the engineering aspects of system selection, regardless of successes or failures. As discussed in Section 2.2, progress is occurring in some of the aforementioned areas, albeit in an uncoordinated fashion. Nonetheless, based on this study, and existing information, a tenta-tive guide to the engineering and limnological suitability of various hypolimnetic aeration designs would be as follows (subject to change with additional information) (Table 5.16). As expected, there is considerable overlap among the various designs in the selection table, demonstrating that the engineering and limnological selection criteria are quite flexible. In practice, non-engineering and non-limnological site-specific selection criteria wil l generally determine which system is selected. For example, the City of New York examined the potential for hypolimnetic aeration to improve raw drinking water quality in the Croton Water System, Table 5.16. General guide to hypolimnetic aeration system selection. Engineering and limnological scenario Full Lift hypolim-netic aeration Partial Lift hy-polimnetic aeration Speece Cone hypolimnetic aeration Cold water fisheries management Acceptable on air or oxygen Acceptable on oxygen Acceptable on oxygen Improvement of raw potable water quality Acceptable on air or oxygen Acceptable on air or oxygen Acceptable on oxygen Hypolimnetic dam discharges from eutrophic reservoirs Acceptable on air or oxygen Acceptable on oxygen Acceptable on oxygen Hypolimnetic oxygen in lakes < 7 m max. depth Acceptable on oxygen Acceptable on oxygen Acceptable on oxygen Hypolimnetic aeration in ice-prone areas Acceptable i f submerged or removed during winter Acceptable on air or oxygen Acceptable on oxygen Hypolimnetic aeration in high boat traffic areas Not recommended due to liability concerns Acceptable on air or oxygen Acceptable on oxygen Hypolimnetic aeration in reser-voirs with signifi-cant bird populations Not recommended Acceptable on air or oxygen Acceptable on oxygen Hypolimnetic aeration in estheti-cally strategic areas Acceptable depend-ing on esthetic concerns Acceptable on air or oxygen Acceptable on oxygen which is one of three systems (Croton, Catskill and Delaware) that provide drinking water to New York City and upstate communities. The City did not want any floating objects in the reservoir, as it may attract roosting birds, and swimmers, neither of which is desirable in a potable water reservoir. Therefore, a full lift design was not acceptable, and three partial lift hypolimnetic aerators operating on air were selected, based on an assumed daily oxygen input of 1,500 kg. Field testing in the second year of the program revealed the combined input from the three aerators was 900 kg O2 day" , hence the system was undersized, and the reservoir contin-ued to experience significant oxygen depletion and water quality problems. The estimate to oxygenate the entire reservoir with partial lift systems, on air, required an additional 18 aerators, at a cost of $7.2 million (New York City Department of Environmental Protection, 2000). Although not considered, the existing three aerators, on oxygen, likely could have met the required oxygen demand, at a fraction of the cost of the 18 aerator option. Due to liability concerns about oxygen storage, cryogenic oxygen was not acceptable; however, PSA generated oxygen was acceptable, and compatible with the existing infrastructure. Also, a single large Speece Cone, similar to the Camanche Reservoir system, would have easily been able to supply the daily oxygen demand of the reservoir. The various options to improve water quality, including upgrading water treatment plants, are currently under review. 5.12 Future Research Requirements Future research requirements to advance the field of hypolimnetic aeration should focus on comparative testing of bench-scale and full-scale systems as per the five steps outlined in the preceding section. At this stage, most of the bench-scale research has been completed on full lift hypolimnetic aerators; hence, the laboratory research should focus on examining partial lift systems, and exploring the performance envelope of the Speece Cone. Based on observations from this study, varying the height to width ratio of the Speece Cone should influence the behavior of the bubble swarm inside the cone, and may significantly influence the oxygenation performance. In addition, minimal attention has been directed at optimizing diffuser design in Speece Cones, as oxygen is introduced through a pipe connection near the top of the cone. Given the known effect of diffuser orifice diameter, it may be possible to improve the oxygen transfer within the Speece Cone by using more efficient diffusers. However, since bubble size in properly designed Speece Cones is a function of the inlet turbulence and is not related to bubble size when injected, this idea may be of marginal use. In terms of full-scale testing, there is considerable need for additional research. Full-scale comparative testing of the oxygenation performance of full lift, partial lift and Speece Cones in the same body of water under similar conditions (e.g., diffuser depth, gas flow rates, diffuser orifice diameter) would be very informative. As previously indicated, it is not possible to extract comparative performance data from the existing literature due to the wide variation in environmental variables known to influence oxygen transfer. Given the energy requirements of the Speece Cone, a relevant research topic would involve testing of extremely efficient low head pumps to minimize the energy consumption of the main water pump, and further improve the already strong performance of the Speece Cone. Also, as indicated for the bench-scale studies, additional gains in Speece Cone oxygenation performance may be possible by examining the use of fine pore diffusers in the inlet pipe, and considering the use of different aspect ratios to determine if an optimal hydraulic design exists which would enhance oxygen transfer inside the Speece Cone. Given the energy impact on S A E from standard pressure PSA oxygen generators, a useful area of research to pursue would be processes that reduced the energy costs of generating oxygen. This would include the testing of low pressure PSA generators, including the new Vacuum Swing Adsorption process. In addition, a comparative study of PSA generated oxygen and liquid oxygen (LOX) would be useful, although the high rental cost of cryogenic storage facilities, liability concerns and limited availability in rural areas would likely favor PSA oxygen generation. Once sufficient laboratory and comparative field data becomes available, it would be useful to develop mathematical models to predict oxygen transfer rates and aerator performance. This process is already underway (Little, 1995); however, it has been hampered by a lack of pub-lished experimental data (McGinnis and Little, 1998). C H A P T E R 6: SUMMARY AND CONCLUSIONS 6.1 Full Lift Experiments 6.1.1 Depth Treatment 1. The diffuser depth treatment exerted a significant effect on KLa2o, SOTR, S A E and SOTE, with all units of measure showing positive increases in response to an increase in diffuser submergence; 2. The depth of gas release influenced KLa2o, SOTR and SOTE mainly by an increase in bubble contact time due to the longer path length the bubbles must take to reach the surface, and a larger deficit of driving force (CI -CL) , resulting in more oxygen transfer; 3. S A E increased with diffuser submergence, indicating that despite the additional energy required to deliver air or PSA O2 to greater depths, the energy cost was more than offset by the increased oxygen transfer resulting from the longer bubble contact times, greater turbulence and increased hydrostatic pressure; 4. Water velocity increased with depth in each experimental treatment; 5. KLa2o and SOTR responded positively to increased gas flow rates due to increased turbulence at the liquid interface and interfacial area available for gas transfer to the surrounding medium; 6. S A E and SOTE responded negatively to increased gas flow rates due to the reduced ratio of interfacial area to bubble volume, reduced contact time between air bubbles in the surrounding liquid and additional energy costs for gas compression; 7. Water velocity increased significantly with gas flow rate; 8. The bubble size data did not show any relationship to gas flow rates. The coefficient of variation ( C V ) was so large that the variance in the data hid any significant effect; 6.1.2 Oxygen Partial Pressure Treatment 9. Oxygen partial pressure treatment exerted a significant effect on K i ^ o and SOTR by substantially increasing the driving force ( Q - C L ) , resulting in more oxygen transfer per unit bubble contact time; 10. SOTE was positively influenced by oxygen partial pressure; 11. S A E response decreased on PSA oxygen, indicating that the energy costs to produce 95% oxygen via PSA technology offset the gains in oxygen transfer; 12. KLa2o and SOTR responded positively to increased gas flow rates (see 5); 13. S A E and SOTE responded negatively to increased gas flow (see 6); 14. Gas composition had no effect on bubble velocity; 6.1.3 Orifice Diameter Treatment 15. Orifice diameter exerted a significant effect on K i ^ o , SOTR, S A E and SOTE. Each unit of measure increased with a decrease in orifice size due to an increase in surface area per unit bubble volume, a decrease in terminal rise velocity, a decrease in the liq-uid film coefficient and an increase in the total number of bubbles in the water column per unit gas discharge; 16. Gas flow rate significantly influenced KLa2o, SOTR, S A E and SOTE in all of the experimental combinations, with K i ^ o and SOTR responding positively to increased gas flow rates, and S A E and SOTE decreasing with increased gas flow rate; 17. KLa2o and SOTR increased with higher gas flow rates (see 5); 18. S A E and SOTE responded negatively to increased gas flow rates (see 6); 6.1.4 In situ Modifications Treatment 19. The in situ modifications treatment exerted a significant effect on K i ^ o , SOTR, S A E and SOTE, with all units of measure showing a negative response to the treatments; 20. The general trend was that the No Treatment option generated the highest values of KLa2o, SOTR, S A E and SOTE (1 s t ranking), the Screen and Cover treatments the low-est values (4 t h and 5 t h ranking), and the Vanes and D B C A treatments were in the mid-dle range (2 n d and 3 r d ranking); 21. The various in situ modifications caused a reduction in water velocity in the full lift aerator by restricting flow either in the inlet tube, the separator box, or the outlet tube; 22. The Screen and Cover treatments exerted an additional negative effect on K j ^ o , SOTR, S A E and SOTE. Further analysis is required to determine the factors respon-sible; 23. A significant gas flow rate x in situ modification interaction was observed for all of the KLa2o, SOTR, S A E and SOTE measurements with the 800 p diffuser, and for the K L a 2 0 and SOTR with the 140 p diffuser; 24. The KLa2o and SOTR interaction is the additive effect of gas flow rate on oxygen transfer, caused by increasing turbulence and interfacial area with each increase in gas flow rate; 25. The D B C A treatment exhibited quite high S A E and SOTE values at the lowest gas flow rate (i.e., 10 1 min"1) but plummeted with increasing gas flow due to déstabilisa-tion of the D B C A effect in the outlet tube; 26. The D B C A treatment generated the smallest bubble sizes; 6.2 Speece Cone 27. The oxygen flow rate and discharge velocity treatments exerted a significant effect on K.La2o and SOTR, with both units of measure showing positive responses to increased oxygen flow rate and discharge velocity; 28. KLa2o and SOTR increased with oxygen flow rate due to a larger deficit of driving force (CJ-CL) and more interfacial area was available for gas inside the Speece Cone; 29. S A E increased with oxygen flow rates but discharge velocity had no effect of S A E as the additional water pump energy required to provide the higher discharge velocities offset the additional dissolved oxygen gained by the greater turbulence inside the Speece Cone and decreased hydraulic residence time; 30. SOTE increased with water discharge velocity as additional turbulence caused by the increased discharge velocities compensated for the reduced bubble:water contact time in the Speece Cone; 31. SOTE responded negatively to increasing oxygen flow, suggesting that oxygen flow rates > 1 1 min"1 were not being fully dissolved inside the Speece Cone; 32. A precipitous decline in SOTE at low oxygen flow rates and discharge velocities > 50 cm sec" was due to a déstabilisation of the bubble swarm in the Speece Cone; 33. The results of an air test were lower than any recorded on PSA oxygen for K i ^ o , SOTR, S A E and SOTE; however, even at the 1/3 the air flow rate being delivered to the full lift aerator, the Speece Cone delivered superior performance for K.La2o, SOTR, and particularly SOTE; 34. The S A E performance was very low due to the energy cost of PSA oxygen and pumping water, and an incorrectly designed Speece Cone, preventing an efficient counter-current bubble swarm from developing in the cone; 35. Direct comparisons between the performance of the Full Lift and Speece Cone systems are not possible as they were not operated at comparable gas flow rates or oxygen partial pressures; 36. Relative comparisons, qualitative in nature, indicate the Speece Cone, at oxygen flow rates between 1 and 3 1 min"1, delivered K L ^ O and SOTR performance comparable to the Full Lift system at 10 1 min"1 on oxygen, and superior performance to the Full Lift system on air; 37. Even with the unit tested, the Speece Cone delivered outstanding SOTE performance, as values > 60% are unattainable with the full lift design on air or oxygen; 38. The D B C A counter-current flow principle of the Speece Cone provides a superior gas transfer environment to the full lift design by enhancing the fundamental aspects of gas transfer (i.e., a, K L and Q - C L ) ; 39. The Speece Cone S A E performance, on oxygen, is comparable to the S A E of the Full Lift system on oxygen; 40. The Speece Cone S A E performance, on air, is considerably less than the Full lift design on air; 41. The relatively low S A E performance of the Speece Cone was due to three factors: (1) the inlet water velocities were considerably less than the design requirement of-305 cm sec"1, (2) the additional energy cost to generate PSA oxygen, and (3) the energy cost of pumping water; 6.3 Comparison with Civil Engineering Studies 42. The KLa2o and SOTE values obtained from the full lift experiments were quite similar to literature values, indicating the performance of the current system is comparable to other designs; 43. The S A E performance obtained from the full lift design was considerably lower than reported literature values, likely due to the higher compression ratio in the adiabatic compression formula used to calculate the energy requirements of the system; 6.4 Linkage to Hypolimnetic Aeration Studies 44. The data available from the literature show a wide range in A E and TE, and relative comparisons between different studies are of limited value due to the wide range in environmental and engineering variables capable of influencing oxygen transfer; 45. The majority of oxygen transfer occurred in the lower portion of the inlet tubes in the full lift designs, and same effect was noted in this study; 46. Declining hydrostatic pressure, decreasing oxygen content of the rising oxygen bubbles and the additive effect of bubble and water velocity in a co-current upflow limit the efficiency of full lift designs; 6.5 PSA Engineering Design Factors Influencing SAE 47. The efficiency of the PSA units increases with unit capacity, as smaller PSA units are not designed solely for optimum oxygen recovery efficiency; 48. Significant improvements (i.e., 46%) in S A E can be expected when using larger capacity PSA systems; 49. The pressure requirements of the PSA generator significantly influences the S A E of a hypolimnetic aeration system as a result of the energy required for feed air compres-sion; 50. Major improvements in S A E could be achieved by using lower pressure PSA genera-tors; 6.6 Comments on Full Scale Application of Full Lift and Speece Cone Systems 51. The relative magnitude of the effects from this study should be transferable to full-scale application, within reasonable limits; 52. In each hypolimnetic aerator installation, one unit of measure wil l likely be de-emphasized in order to achieve specified performance targets in another unit of meas-ure; 53. There are many factors unrelated to gas transfer that play a critical role in system selection; 54. Lab testing, comparative field studies, development of predictive models, reporting in engineering journals and case study reviews of specific installations will assist in re-solving the engineering aspects of system selection; 6.7 Future Research Requirements 55. Laboratory research should focus on examining partial lift systems, and various designs of Speece Cones; 56. Varying the height to width ratio of the Speece Cone and optimizing diffuser design may improve the oxygen transfer within the Speece Cone; 57. Comparative full-scale testing of the oxygenation performance of full lift, partial lift and Speece Cones in the same body of water under similar conditions (e.g., diffuser depth, gas flow rates, diffuser orifice diameter) would be very informative; 58. Speece Cone oxygenation performance may be improved by using extremely efficient low head pumps, using fine pore diffusers in the inlet pipe, and investigating different aspect ratios to determine if an optimal hydraulic design exists which would enhance oxygen transfer inside the Speece Cone; 59. Once sufficient laboratory and comparative field data becomes available, it would be useful to develop mathematical models to predict oxygen transfer rates and aerator performance. This process is already underway; however, it has been hampered by a lack of published experimental data. 6.8 Conclusions In the 21 s t century, hypolimnetic aeration will become more accepted within the engineering community as a technique for improving raw water quality in eutrophic lakes and reservoirs that receive significant amounts of nutrients from internal loading and uncontrollable external sources. The most widely used designs will likely be (1) Downflow Bubble Contact Aeration using the Speece Cone design and (2) full lift hypolimnetic aerators equipped with high effi-ciency diffusers using compressed air or oxygen. The Speece Cone, with its relatively compact size, exceptional oxygenation performance, and shallow water capability, represents one of the most significant advances in the design of hypolimnetic aeration since their inception over half a century ago. Equally important is the recent development of on-site PSA oxygen generating systems, which eliminates concerns surrounding the use/storage of pure oxygen and allows for the development of smaller and more efficient designs of hypolimnetic aeration systems. Civi l engineers should consider hypolimnetic aeration as a proven technique, which can significantly improve raw water quality, and serve as an integral component of an overall watershed man-agement program. BIBLIOGRAPHY [I] Andeen, G .B . 1974. Bubble pumps. Compressed A i r Magazine 79:16-19. [2] American Public Health Association, American Water Works Association and Water Pollution Control Federation. 1980. Standard methods for the examination of water and wastewater, 15th edition. American Public Health Association, Washington, D.C. [3] A S C E . 1992. Measurement of oxygen transfer in clean water. A N S I / S S C E 2-91, 2 n d edition, June, 1992. American Society of Civ i l Engineers, New York, New York, U S A . [4] Ashley, K . I . 1983. Hypolimnetic aeration of a naturally eutrophic lake: physical and chemical effects. Can. J. Fish. Aquat. Sci. 40(9): 1343-1359. [5] Ashley, K. I . 1985. Hypolimnetic aeration: practical design and application. Water Re-search 19(6):735-740. [6] Ashley, K.I . , S. Hay and G . H . Scholten. 1987. Hypolimnetic aeration: field test of the empirical sizing method. Water Research 21(2):223-227. [7] Ashley, K. I . 1988. Hypolimnetic aeration research in British Columbia. Verh. Internat. Verein. Limnol. 23:215-219. [8] Ashley, K. I . 1989. Factors influencing oxygen transfer in diffused aeration systems and their application to hypolimnetic aeration. M.A.Sc . thesis, Dept. of C iv i l Engineering, University of British Columbia. [9] Ashley, K.I . and K . J . Hall . 1990. Factors influencing oxygen transfer in hypolimnetic aeration systems. Verh. Internat. Verein. Limnol. 24:179-183. [10] Ashley, K.I . , D.S. Mavinic and K . J . Hall . 1990. Oxygen transfer in full lift hypolimnetic aeration systems, pp. 648-659 in Air-Water Mass Transfer: Selected Papers from the Second International Symposium on Gas Transfer at Water Surfaces. S.C. Wilhelms and J.S. Gul l i -ver (Eds.). American Society of Civ i l Engineers, New York, U S A . [II] Ashley, K.I . , K . J . Hall and D.S. Mavinic. 1991. Factors influencing oxygen transfer in fine pore diffuser aeration. Water Research 25(12):1479-1486. [12] Ashley, K.I . , T. Tsumura and B . M . Chan. 1992. Fisheries management of winterkill lakes in Southern Interior British Columbia, pp. 287-297 in Aquatic Ecosystems in Semi-Ar id Regions: Implications for Resource Management. R .D. Robarts and M . L . Bothwell (Eds.). N.H.R.I . Symposium Series 7, Environment Canada, Saskatoon, Canada. [13] Ashley, K . I . and R. Nordin. 1999. Lake aeration in British Columbia: Applications and experiences, pp. 87-108 in Aquatic Restoration in Canada. T. Murphy and M . Munawar (Eds.). Ecovisoon World Monogrpah Series. Backhuys Publishers, Leiden, The Netherlands. [14] Barnhardt, E . L . 1969. Transfer of oxygen in aqueous solutions. J. Sanit. Eng. Div.Proc. A S C E 95 No. SA3:645-661. [15] Beak Consultants Ltd. 1977. C P A R Project Report 542-1. State of the art review: Aeration. Beak Consultants Ltd., Montreal, Quebec. July, 1977. [16] Bernhardt, H . 1967. Aeration of Wahnbach Reservoir without changing the temperature profile. J. American Water Works Assoc. 59:943-964. [17] Bernhardt, H . 1974: Tens years experience of reservoir aeration. Prog. Water Technol. 7:483-495. [18] Bewtra, J.K. and W.R. Nicholas. 1964. Oxygenation from diffused air in aeration tanks. J. Water. Poll . Control Fed. 36(10): 1195-1224. [19] Bewtra, J.K., R .W. Nicholas and L . B . Polkowski. 1970. Effect of temperature on oxygen transfer in water. Water Research 4(1): 115-122. [20] Bewtra, J .K. and D.S. Mavinic. 1978. Diffused aeration systems from theory to design. Can. J. Civ i l Eng. 5(1):32-41. [21] Bowers, R . H . 1955. The mechanics of bubble formation. J. Applied Chemistry 5:542-548. [22] Boyd, C E . 1986. A method for testing aerators for fish tanks. Prog. Fish Cult. 48:68-70. [23] Boyd, C E . and B.J . Watten. 1989. Aeration systems in Aquaculture. Reviews in Aquatic Sciences l(3):425-472. [24] Colt, J. 1984. Computation of dissolved gas concentrations in water as functions of temperature, salinity and pressure. American Fisheries Society Special Publication No. 14. Bethesda, Maryland, U S A . [25] Danckwerts, P .V. 1951. Significance of liquid film coefficient in gas absorption. Indus-trial and Engineering Chemistry 43(6): 1460. [26] Dinsmore, W.P. and E.E. Prepas. 1994a. Impact of hypolimnetic oxygenation on pro-fundal macroinvertebrates in a eutrophic lake in central Alberta. I. Changes in macroinverte-barte abundance and diversity. Can. J. Fish. Aquat. Sci. 54:2157-2169. [27] Dinsmore, W.P. and E.E. Prepas. 1994b. Impact of hypolimnetic oxygenation on profundal macroinvertebrates in a eutrophic lake in central Alberta. II. Changes in Chirono-mus spp. abundance and biomass. Can. J. Fish. Aquat. Sci. 54:2170-2181. [28] Dobbins, W.E. 1964. Mechanisms of gas absorption by turbulent liquids. Advances in Water Research Proc. 1st International Conf. Water Pollution Research. Pergamon Press Ltd., London, England. V o l . 2, p. 61. [29] Doke, J.L., W . H . Funk, S.T.J. Jull, and B .C . Moore. 1995. Habitat availability and benthic invertebrate population changes following alum treatment and hypolimnetic oxygenation in Newman Lake, Washington. J. Freshwater Biol . 10:87-102. [30] Downing, A . L . and G. A . Truesdale. 1955. Some factors affecting the rate of solution of oxygen in water. J. Applied Chemistry 5:570-581. [31] Dunst, R.C. , S .M. Born, S.A. Smith, S.A. Nichols, J.0. Peterson, D.R. Knauer, S.L. Serns, D.R. Winter and T .L . Wirth. 1974. Survey of lake rehabilitation techniques and experiences. Tech. Bul l . No . 75. Wisconsin Dept. Natural Resources, Madison, Wisconsin. 179 pp. [32] Eckenfelder, W.W. 1959. Absorption of oxygen from air bubbles in water. J. Sank. Eng. Div. Proc. A S C E 85 No. SA4:89-99. [33] Eckenfelder, W.W. and D . L . Ford. 1968. New concepts in oxygen transfer and aeration. In Advances in Water Quality Improvement. E.F. Gloyna and W.W. Eckenfelder (Eds.). Water Resource Symposium No. 1. [34] Eckenfelder, W.W. (Ed.) 1969. Oxygen transfer and aeration. In Manual of Treatment Processes, V o l . 1. Environmental Sciences Services Corp., U S A . [35] Ellis , S.E. and R. Stanbury. 1980. The scale-up of aerators. Paper No.3 Symposium on The Profitable Aeration of Waste Water, (London, Apr i l , 1980). B H R A Fluid Engineering, Cranfield, Bedford. U .K . [36] Fast, A . W . 1971. The effects of artificial aeration on lake ecology. E P A Water Pollution Control Research Series 16010 E X E 12/71. [37] Fast, A . W . 1973. Effects of artificial hypolimnion aeration on rainbow trout {Salmo gairdneri Richardson) depth distribution. Trans. A m . Fish. Soc. 73(4):715-722. [38] Fast, A . W . , B . Moss and R .G . Wetzel. 1973. Effects of artifical aeration on the chemis-try and algae of two Michigan lakes. Water Resources Research 9:624-647. [39] Fast, A . W . 1975. Artificial aeration and oxygenation of lakes as a restoration technique, pp. 135-165 in Recovery and Restoration of Damaged Ecosystems. Caimes et al. (Eds.). The University Press of Virginia, Charlottesville, U S A . [40] Fast, A . W . , W.J. Overholtz and R .A . Tubb. 1975. Hypolimnetic oxygenation using liquid oxygen. Water Resources Research 11:294-299. [41] Fast, A . W . and M . W . Lorenzen. 1976. Synoptic survey of hypolimnetic aeration. J. Environ. Eng. Div., Proc. Amer. Soc. Civi l Eng. 106 (EE6):1161-1173. [42] Fast, A . W . , M . W . Lorenzen and J.H. Glenn. 1976. Comparative study with costs of hypolimnetic aeration. J. Environ. Eng. Div., Proc. Amer. Soc. Civi l Eng. 106 (EE6):1175-1187. [43] Fast, A . W . , Overholtz, W.J. and R . A . Tubb. 1977: Hyperoxygen concentrations in the hypolimnion produced by injection of liquid oxygen. Water Resources Res. 13:474-476. [44] Fast, A . W . 1981. Some applications of lake aeration technology and artificial thermal upwelling on fish culture, pp 574-606 in Destratification of Lakes and Reservoirs to Improve Water Quality. Proceedings of a joint United States/Australia Seminar and Workshop, Melbourne, Australia, Feb 19-24, 1979. A . J . Burns and I.J. Powling (Eds.). Australian Government Publishing Service, Canberra, Australia. [45] Field, K . M . and E .E . Prepas. 1997. Increased abundance and depth distribution of pelagic crustacean zooplankton during hypolimnetic oxygenation in a deep, eutrophic Alberta lake. Can. J. Fish. Aquat. Sci. 54: 2146-2156. [46] Garrell, M . H . , J.C. Confer and D. Kirschner. 1977. Effects of hypolimnetic aeration on nitrogen and phosphorus in a eutrophic lake. Water Resources Res. 13:343-347. [47] Gemza, A . F . 1997. Water quality improvements during hypolimnetic oxygenation in two Ontario lakes. Water Qual. Res. J. Canada 32:365-390. [48] Haberman, W . L . and R . K . Morton. 1954. A n experimental study of bubbles moving in liquids. Proc. Amer. Soc. C iv i l . Eng. 80:379-429. [49] Halsey, T .G. 1968. Autumnal and overwinter limnology of three small eutrophic lakes with particular reference to experimental circulation and trout mortality. J. Fish. Res. Bd. Can. 25:81-89. [50] Haney, P.D. 1954. Theoretical principles of aeration. J. American Water Works Assoc. 46(4):353-376. [51] Higbie, R. 1935. The rate of absorption of a pure gas into a still liquid during periods of short exposure. Amer. I. Chem. Eng. Trans. 31:365-390. [52] Hess, L . 1975. The effect of the first year of artificial hypolimnetic aeration on oxygen, temperature and depth distribution of rainbow trout (Salmo gairdneri) in Spruce Knob Lake. West Virginia Dept. of Natural Resources Report F-19-R-3. [53] Ippen, A . T . and C E . Carver. 1954. Basic factors of oxygen transfer in aeration systems. Sewage and Industrial Wastes 26(7):813-829. [54] Jaeger, D. 1990. T I B E A N - a new hypolimnetic water aeration plant. Verh. Internat. Verein. Limnol. 24:184-187. [55] King, D . L . 1970. Reaeration of Streams and Reservoirs; Analysis and Bibliography. Engineering and Research Center, U . S. Bureau of Reclamation R E C - O C E - 70-55. 131 pp. [56] Krenkel, P .A. and G.T. Orlob. 1962. Turbulent diffusion and the reaeration coefficient. J. Sanit. Eng. Div. , Proc. Amer. Soc. Civri l Eng. 88 (SA2):53-83. [57] LaBaugh, J.W. 1980. Water chemistry changes during artificial aeration of Spruce Knob Lake, West Virginia. Hydrobiologia 70: 201-216. [58] Lappalainen, K . M . 1994. Positive changes in oxygen and nutrient contents in two Finnish lakes induced by Mirox hypolimnetic oxygenation method. Verh. Internat. Verein. Limnol. 25:2510-2513. [59] Larkin, P .A. 1975. Biometrics. Biology 300 Course Book. Fall Term, 1975. University of British Columbia, 206 p. [60] Lawrence, G.A. , J . M Burke, T.P. Murphy and E.E. Prepas. Exchange of water and oxygen between the two basins of Amisk lake. Can. J. Fish. Aquat. Sci. 54:2121-2132. [61] Little, J . C , 1995. Hypolimnetic aerators: predicting oxygen transfer and hydrodynamics. Water Research 29:2475-248. [62] Lee, G.F. 1970. Eutrophication. University of Wisconsin Water Resources Center, Eutrophication Information Program. Occasional paper No. 2, Madison, Wisconsin, U S A . [63] Lewis, W . K . and W . G . Whitman. 1924. Principles of gas absorption. Ind. Eng. Chem.l6:1215-1220. [64] Lind, O.T. 1979. Handbook of Common Methods in Limnology. Second Ed. C . V . Mosby. St. Louis, 199 p. [65] Lorenzen, M . and A . W . Fast. 1977. A guide to Aeration/Circulation techniques for lake management. U . S. Environmental Protection Agency 600/3-77-004. [66] Mancy, K . H . and D . A . Okun. 1960. Effects of surface active agents on bubble aeration. J. Water Poll . Control Fed. 32(4):351-364. [67] Markofsky, M . 1979. On the reoxygenation efficiency of diffused air aeration. Water Research 13:1339-1346. [68] Mavinic, D.S. 1973. Optimization of oxygen transfer into water in a coarse-bubble, diffused aeration system. Ph.D. thesis, University of Windsor, Windsor, Ontario, Canada. [69] Mavinic, D.S. and J.K. Bewtra. 1974. Mass transfer of oxygen in diffused aeration systems. Can. J. C iv i l . Eng. 1:71-84. [70] Mavinic, D.S. and J.K. Bewtra. 1976. Efficiency of diffused aeration systems in waste-water treatment. J. Water Poll . Control Fed. 48(10):2273-2283. [71] McGinnis, D.F. and J.C. Little. 1998. Bubble dynamics and oxygen transfer in a Speece Cone. Water Sci. Tech. 37(2):285-292. [72] McQueen, D.J. and D.R.S. Lean. 1983. Hypolimnetic aeration and dissolved gas concentrations: Enclosure experiments. Water Research 17:1781-1790. [73] McQueen, D.J., S.S. Rao and D.R.S. Lean. 1984. Hypolimnetic aerationxhanges in bacterial populations and oxygen demand. Arch. Hydrobiol. 99(4):498-514. [74] McQueen, D.J. and D.R.S. Lean. 1986. Hypolimnetic aeration: A n overview. Water Poll . Res. J. Canada 21(2):205-217. [75] Mercier, P. and J. Perret. 1949: Aeration station of Lake Bret, Monastbull, Schweiz. Ver Gas. Wasser-Fachm 29:25. [76] Metcalf and Eddy Ltd. 1991. Waterwater Engineering: Treatment, Disposal and Reuse, Third Edition. Revised by G. Tchobanoglous and F .L . Burton. McGraw-Hi l l , Inc., New York, New York, U S A . [77] Morgan, P.F. and J.K. Bewtra. 1960. A i r diffuser efficiencies. J. Water Poll . Control Fed. 32(10):1047-1059. [78] Neilson, B.J . 1974. Reaeration dynamics of reservoir destratification. J. Amer. Water Works Assoc. 66:617-620. [79] New York Department of Environmental Protection. 2000. Hypolimnetic Aeration Pilot Project: Final Report. New York City Department of Environmental Protection. Pre-pared by the Division of Drinking Water Quality Control, Valhalla, New York. [80] Nienow, A . W . 1980. The aeration of waste water: Basic physical factors. Paper No. 1. Symp. on the Profitable Aeration of Waste Water, (London, Apr i l , 1980). B H R A Fluid Engineering, Cranfield, Bedford, U .K. [81] O'Connor, D.J. 1982. Wind effects on gas-liquid transfer coefficients. J. Environ. Eng., Proc. A S C E 109(3):731-753. [82] Pasveer, A . 1955. Research on activated sludge V I . Oxygenation of water with air bubbles. Sewage and Industrial Wastes 27(10):1130-1146. [83] Perrin, C.J., K . I . Ashley and G .A . Larkin. 2000. Effect of drawdown on ammonium and iron concentrations in a coastal mountain reservoir. Water Qual. Res. J. Canada 35(2):231-244. [84] Point Four Systems, 1997. PT4 Users Manual U B C Civ i l Engineering Dept. Point Four Systems, Inc., Richmond, B . C . Canada. [85] Prepas, E.E. , D.J . Webb, C . L . K . Robinson and T.P. Murphy. 1990. Impact of liquid oxygen injection on a deep, naturally eutrophic lake: Amisk Lake, Alberta, year one. Verh. Internat. Verein. Limnol. 24:320. [86] Prepas, E .E . and J . M . Burke. 1997. Effects of hypolimnetic oxygenation on water quality in Amisk Lake, Alberta, a deep eutrophic lake with high internal phosphorus loading rates. Can. J. Fish. Aquat. Sci. 54:2111-2120. [87] Prepas, E.E. , K . M . Field, T.P. Murphy., W . L . Johnson., J . M . Burke and W . M . Tonn. 1977: Introduction to the Amisk Lake Project: oxygenation of a deep, eutrophic lake. Can. J. Fish. Aquat. Sci. 54:2105-2110. [88] Rohlf, F.J. and R.R. Sokal. 1969. Statistical Tables. W . H . Freeman and Co. San Fran-cisco, Ca. 253 p. [89] Rucker, R. 1972. Gas bubble disease of salmonids: a critical review. U.S. Fish and Wildlife Service. Tech. Paper 58. [90] Schmit, F .L . , J.D. Wren and D.T. Redmon. 1978. The effect of tank dimensions and diffuser placement on oxygen transfer. J. Water Poll . Control Fed. 50(7): 1750-1767. [91] Smith, S.A., D.R. Knauer and T.Wirth. 1975. Aeration as a lake management technique. Wisconsin Dept. Natural Res., Tech. Bul l . No.87. [92] Soltero, R .A . , L . M . Sexton, K. I . Ashley and K . O . McKee. 1994. Partial and full lift hypolimnetic aeration of Medical Lake, Wa to improve water quality. Water Research 28:2297-2308. [93] Speece, R .E . 1971. Hypolimnion Aeration. J. Amer. Water Works Assoc. 64(l):6-9. [94] Speece, R .E . 1975. Ten Year's experience of reservoir aeration. Reply to H . Bernhardt. Prog. Water Technol. 7:489-494. [95] Speece, R.E. , N . Nirmalakhandan and G. Tchobanoglous. 1990. Commercial oxygen use in water-quality management. Water Environment and Technology 2(7):54-61. [96] Speece, R .E . 1992. Pilot Scale demonstration of Speece Cone for oxygenation of Camanche Reservoir. Internal Report (Revised), Vanderbuilt University. 11 pp. [97] Steinberg, C. and K . Arzet. 1984. Impact of hypolimnetic aeration on abiotic and biotic conditions in a small kettle lake. Environ. Technol. Letters. 5: 151-162. [98] Stenstrom, M . K . and R .G . Gilbert. 1981. Effects of alpha, beta and theta factor upon the design, specification and operation of aeration systems. Water Research 15:643-654. [99] Taggart, C.T. and D.J. McQueen. 1981. Hypolimnetic aeration of a small eutrophic kettle lake: physical and chemical changes. Arch. Hydrobiol. 91(2) 150-180. [100] Taggart, C.T. and D.J. McQueen. 1982. A model for the design of hypolimnetic aerators. Water Research 16:643-654. [101] Taggart, C.T. 1984. Hypolimnetic aeration and zooplankton distribution: a possible limitation to the restoration of cold water fish production. Can. J. Fish. Aquat. Sci. 41:191-198. [102] Thomas, J.A., W.H. Funk, B .C . Moore and W.W. Budd. 1994: Short term changes in Newman Lake following hypolimnetic aeration with the Speece Cone. Lake and Res. Manage. 9:111-113. [103] Verner, B . 1974. Long term effect of hypolimnetic aeration of lakes and reservoirs, with special considerations of drinking water quality and preparation costs. In J. Taggart and L . Moore (Eds.) Lake and Reservoir Management: 134-138. Proc. 3 r d Annual Conf. North Amer. Lake Management Soc, October 18-20, 1983, Knoxville, Tenn., U S A . U . S. E P A 440/5- 84-001. 604 p. [104] Webb, D.J., R .D. Robarts and E.E. Prepas. 1997. Influence of extended water column mixing during the first 2 years of hypolimnetic oxygenation on the phytoplankton community of Amisk lake, Alberta. Can. J. Fish. Aquat. Sci. 54:2133-2145. [105] Weisberg, S. 1980. Applied Linear Regression. John Wiley and Sons, New York, New York, U S A . [106] Wetzel, R .G . 1975. Limnology. W. B . Saunders Co., Philadelphia, Pa. 743 p. [107] Wilkinson, L . and M . Coward. 2000. Linear Models II: Analysis of Variance. Chapter 15 in S Y S T A T 10, Statistics 1. SPSS Inc., Chicago, U S A . [108] Wirth, T.L. , D.R. Knauer and S.A. Smith. 1975. Total and hypolimnetic aeration of lakes in Wisconsin. Verh. Internat. Verien. Limnol. 19:1960-1970. L I S T O F A P P E N D I C E S Page A P P E N D I X 1 : Re-Aeration Test Procedure - Full lift hypolimnetic aerator 294 A P P E N D I X 2: Re-Aeration Test Procedure - Speece hypolimnetic aerator 296 APPENDIX 1: RE-AERATION TEST PROCEDURE - F U L L LIFT HYPOLIMNETIC A E R A T O R 1. Turn on main power bar (on control panel); 2. F i l l tank with water (~ 1200 L) ; 3. Turn on secondary power bar (at top of stairs and activate pump mixers when tank is full); 4. Turn on PT4 data logger/oxygen-temperature probes; 5. Insert YSI probe into tank and turn meter on (zero meter initially); 6. Insert current meter probe into outlet tube and turn meter on; 7. Turn on electronic balance; 8. Let system mix for 5 minutes; 9. Take water sample in 2 x 300 ml B O D bottles and measure temperature of tank water; 10. Do Winkler titration in Civi l Lab within 5 minutes; 11. Calculate % saturation from Winkler, tank water temperature and barometric pressure; 12. Calibrate data logger using single point manual calibration to % saturation value for Chs. 1-6; 13. Check Ch. 7 temperature against YSI temp and thermometer; 14. Calibrate YSI in mg/L from Winkler test; 15. Put on safety glasses and gloves; 16. Calibrate Ch. 7 to 100% for certified oxygen from compressed oxygen cylinder; 17. Attach compressed air line and pressurize manifold; 18. Cycle AirSep unit to full pressure; 19. Calculate amount of cobalt chloride required, and add only once per 5-6 tests dissolved in test water; 20. Calculate amount of sodium sulfite required (+10%), dissolved in flask and add to test water; 21. M i x for 6 minutes, confirm all probes are below 0.5% saturation; 22. Review data logger setting for calibration settings, time stamp and data capture fre-quency; 23. Set up Laptop with data link, power cord and mouse, open Hyperterminal; 24. Record experimental parameters, setting and conditions on record sheet; 25. Confirm YSI is < 0.5 mg/L; 26. Configure data logger to start data collection; 27. Configure manifold to deliver desired (1) gas mixture and (2) gas flow rate and (3) pressure from random number generated treatment schedule; 28. Start data logger after 6 minutes of mixing; 29. Turn mixers off at start of test run; 30. Start air or oxygen flow (If oxygen > 10 S L P M then AirSep on Auto, if > 10 S L P M then AirSep on manual); 31. Adjust regulators to 45 psig; 32. Monitor data logger; 33. Monitor current velocity and record; 34. Terminate test at _% sat; 35. Stop data logger; 36. Prepare computer for data capture; 37. Transfer data to Lap Top; 38. Find file in Excel as .csv file; 39. Save data as unique Excel file; 40. Repeat procedure from Step 20 according to treatment table (Note : max 6 tests per batch of water); 41. Drain tank after 5-6 tests and refill as soon as possible to avoid stressing tank joints at start on new set of 5-6 experiments. APPENDIX 2: RE-AERATION TEST PROCEDURE - SPEECE HYPOLIMNETIC A E R A T O R 1. Turn on main power bar (on control panel); 2. F i l l tank with water (~ 1200 L) ; 3. Turn on secondary power bar (at top of stairs and activate pump mixers only when tank is full); 4. Turn on PT4 data logger/oxygen-temperature probes; 5. Insert YSI probe into tank and turn meter on (zero meter initially); 6. Check current meter probe on discharge tube and turn meter on; 7. Turn on electronic balance; 8. Let system mix for 5 minutes; 9. Take water sample in 2 x 300 ml B O D bottles and measure temperature of tank water; 10. Do Winkler titration in Civi l Lab within 5 minutes; 11. Calculate % saturation from Winkler, tank water temperature and barometric pressure; 12. Calibrate data logger using single point manual calibration to % saturation value for Chs. 1-6; 13. Check Ch. 7 temperature against YSI temp and thermometer; 14. Calibrate YSI in mg/L from Winkler test; 15. Put on safety glasses and gloves; 16. Calibrate Ch. 7 to 100% for certified oxygen from compressed oxygen cylinder; 17. Attach compressed air line and pressurize manifold; 18. Cycle AirSep unit to full pressure; 19. Calculate amount of cobalt chloride required, and add only once per 5-6 tests dissolved in test water; 20. Calculate amount of sodium sulfite required (+10%), dissolved in flask and add to test water; 21. M i x for 8 minutes, confirm all probes are below 0.5% saturation; 22. Review data logger setting for calibration settings, time stamp and data capture fre-quency; 23. Set up Laptop with data link, power cord and mouse, open Hyperterminal; 24. Record experimental parameters, setting and conditions on record sheet; 25. Confirm YSI is < 0.5 mg/L; 26. Configure data logger to start data collection; 27. Configure manifold to deliver desired (1) gas mixture and (2) gas flow rate and (3) pressure from random number generated treatment schedule; 28. Leave mixers on throughout test run; 29. Select current velocity, and adjust rheostat to desired current velocity; 30. Start data logger after 8 minutes of mixing; 31. Start air or oxygen flow (If oxygen > 10 S L P M then AirSep on Auto, if > 10 S L P M then AirSep on manual); 32. Adjust regulators to 45 psig; 3 3. Monitor data logger; 34. Monitor current velocity and record; 3 5. Terminate test at _% sat; 36. Turn off gas flow and turn off main pump to allow system to degas for 5 seconds through vent tube; 37. Turn main mixer back on and let mix for 2 minutes without any gas flow to obtain complete mixing in tank; 38. Stop data logger; 39. Prepare computer for data capture; 40. Transfer data to Lap Top; 41. Find file in Excel as .csv file; 42. Save data as unique Excel file; Repeat procedure from Step 20 according to treatment table (Note : max 6 tests per batch of water); Drain tank after 5-6 tests and refill as soon as possible to avoid stressing tank joints at start on new set of 5-6 experiments. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0063683/manifest

Comment

Related Items