UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Air-actuated pumping technology in urban drainage 2004

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

Item Metadata

Download

Media
ubc_2004-0178.pdf
ubc_2004-0178.pdf
ubc_2004-0178.pdf [ 5.7MB ]
Metadata
JSON: 1.0063467.json
JSON-LD: 1.0063467+ld.json
RDF/XML (Pretty): 1.0063467.xml
RDF/JSON: 1.0063467+rdf.json
Turtle: 1.0063467+rdf-turtle.txt
N-Triples: 1.0063467+rdf-ntriples.txt
Citation
1.0063467.ris

Full Text

A I R - A C T U A T E D PUMPING T E C H N O L O G Y IN U R B A N D R A I N A G E by D . A a r o n B o h n e n B . A S c , The U n i v e r s i t y o f B r i t i s h C o l u m b i a , 1996 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L M E N T O F 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 O F M A S T E R O F A P P L I E D S C I E N C E i n ' T H E F A C U L T Y O F G R A D U A T E S T U D I E S C I V I L E N G I N E E R I N G W e accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A A p r i l 2004 © D . A a r o n Bohnen , 2004 Library Authorization In presenting this thesis in partial fulfillment 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. I 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. Name of Author (please print) Date ~(dd/mm/yyyy) Title of Thesis: 4'/g^/}^fU/tT&h PUH4P'AJ6r TS^A/rJiFLO fry Degree: Year: Department of CWlL- &KT&? I K I S . ^ : fZ ?r4 ^ The University of British Columbia Vancouver, BC Canada A B S T R A C T A i r l i f t P u m p technology is br ief ly summar ized and the potential appl icat ion o f airl ift technology to low- l i f t , low-submergence, h igh- f low applications such as open channel f l ow i n urban storm drainage, is explored. Four experimental setups are described, i nc lud ing one prototype urban storm drainage installat ion. Three descript ive models for airl ift pump operation are developed and one adopted for appl icat ion i n low- l i f t , l o w - submergence, h igh- f low applications. The mode l a l lows a s imple design procedure for airlift pumps i n this regime. A s imple design hand-calculations procedure is developed, and two personal computer-based implementations are described. A s imple design example is presented and recommendations for further research and development directions are made. TABLE OF CONTENTS Abstrac t i i Table o f Contents i i i L i s t o f Figures v L i s t o f Tables v i Acknowledgemen t s v i i C H A P T E R 1 O v e r v i e w & Summary 1.1 Introduction to A i r l i f t Pumps 1 1.2 Desc r ip t ion o f an A i r l i f t P u m p 6 1.3 A p p l i c a t i o n s o f A i r l i f t Pumps 9 1.4 Project Scope and Rat ionale 11 1.5 Two-Phase F l o w Reg imes 14 1.6 Operat ional E f f i c i ency o f A i r l i f t Pumps 17 1.7 S u m m a r y 20 C H A P T E R 2 Literature R e v i e w 2.1 Introduction 22 2.2 Deve lopment o f A i r l i f t P u m p Theory , 1797 to Present 22 2.3 S u m m a r y 30 C H A P T E R 3 Exper imenta l P rogram 3.1 O v e r v i e w o f the Exper imenta l P rogram 32 3.2 The Exper imenta l Setups 36 3.2 Resul ts o f the Exper imenta l P rogram 56 i i i T A B L E O F CONTENTS (cont'd) C H A P T E R 4 Three A i r l i f t P u m p Theoret ical M o d e l s 4.1 A i r l i f t P u m p M o d e l for F i x e d B u b b l e Ve loc i t i e s 57 4.2 A i r l i f t P u m p M o d e l for V a r i a b l e B u b b l e S l i p V e l o c i t i e s 65 4.3 A i r l i f t P u m p M o d e l for Turbulent M i x i n g 72 4.4 S u m m a r i z i n g the Three M o d e l s 79 C H A P T E R 5 P re l im ina ry D e s i g n o f A i r l i f t Pumps 5.1 P re l imina ry D e s i g n Procedure Calcula t ions 81 5.2 D e s i g n Calcula t ions for Personal Compute r 86 5.3 Prac t ica l Considerat ions for P re l imina ry A i r l i f t P u m p D e s i g n . 91 C H A P T E R 6 Conc lus ions and Research Recommendat ions 6.1 Conc lus ions 94 6.2 Research Recommendat ions 95 R E F E R E N C E S 96 A P P E N D L X 1 99 L I S T O F F I G U R E S Figure 1 - Schemat ic A i r l i f t P u m p Layou t and T e r m i n o l o g y 7 (also 57) F igure 2 - Two-Phase A i r - W a t e r F l o w Regimes 15 F igure 3 - Two-Phase F l o w Regimes characterized by Gas F l u x and M i x t u r e 16 V o i d Frac t ion F igure 4 - F i r s t Labora tory A i r l i f t P u m p i n g Sys tem 37 F igure 5 - Fi rs t Labora tory A i r l i f t P u m p i n g Sys tem Results 38 F igure 6 - Prototype Scale Laboratory A i r l i f t P u m p i n g Sys tem 39 F igure 7 - G i lbe r t R o a d Prototype A i r l i f t Sys tem Layou t 42 F igure 8 - G i lbe r t R o a d Prototype A i r l i f t Sys tem Components 43 F igure 9 - G i lbe r t R o a d Compressed A i r Supp ly Subsystem 45 F igure 10 - G i lbe r t R o a d Prototype A i r l i f t Sys tem in Operat ion 46 F igure 11 - Slot -conf igured A i r l i f t P u m p Sys tem i n Operat ion 51 F igure 12 - R i c h m o n d P u b l i c W o r k s Exper imenta l Setup 54 F igure 13 - B u b b l e R i s e Ve loc i t i e s in S t i l l Wate r 61 F igure 14 - M i x t u r e and Gas F l u x Rates 65 F igure 15 - C o m p a r i s o n o f Exper imenta l and Calcu la ted Performance 78 v L I S T O F F I G U R E S (cont'd) Figure 16 - S i m p l e D e s i g n E x a m p l e Layou t 82 F igure 17 - A i r l i f t P u m p C h u r n F l o w Worksheet 87 F igure 18 - V B A C o d e for C h u r n F l o w A i r l i f t P u m p Des ign 89 F igure 19 - V B A C o d e for C h u r n F l o w A i r l i f t P u m p Des ign 90 L I S T O F T A B L E S Table 1 - A i r l i f t P u m p Nomencla ture and Var iab les 8 Tab le 2 - Prototype Scale Laboratory A i r l i f t P u m p i n g Sys tem Results 40 Tab le 3 - G i lbe r t R o a d Prototype A i r l i f t Sys tem Sample Exper imenta l Da ta 48 Tab le 4 - G i lbe r t R o a d Prototype A i r l i f t Sys tem Sample Exper imenta l 48 Resul ts Tab le 5 - Gi lbe r t R o a d Prototype A i r l i f t Sys tem Leakage Tests 49 Tab le 6 - G i lbe r t R o a d Prototype A i r l i f t Sys tem Sample Exper imenta l 50 Resul ts 2 Tab le 7 - R i c h m o n d P u b l i c W o r k s Sample Exper imenta l Da ta 55 v i A C K N O W L E D G E M E N T S I w o u l d l ike to gratefully acknowledge m y research supervisor, Den i s S. O . R u s s e l l , Professor Emer i tus o f the U n i v e r s i t y o f B r i t i s h C o l u m b i a Department o f C i v i l Engineer ing , for his continuous support throughout this project. Thanks also to Professor J i m Atwa te r and Professor A l a n Russe l l , also o f the U n i v e r s i t y o f B r i t i s h C o l u m b i a Department o f C i v i l Engineer ing . The i r assistance and encouragement was par t icular ly valuable. Thanks also to the Engineer ing Department at the C i t y o f R i c h m o n d , and M r . Ar thu r L o u i e whose interest i n the appl icat ion o f airl ift pumps to urban drainage made this study possible . F i n a l l y , thanks also to the Na t iona l Science and Research C o u n c i l o f Canada for their contr ibut ion dur ing the early phases o f this project. C H A P T E R 1 1.1 - I n t r o d u c t i o n to A i r l i f t P u m p s A i r l i f t pumps are c o m m o n l y considered to be part o f a unique class o f "alternate" p u m p i n g technologies. These alternate p u m p i n g technologies are required when c o m m o n ro todynamic pumps are unsuitable for a g iven project or appl icat ion. S o m e applications that c o m m o n l y benefit f rom alternate pumping technologies i nvo lve f lu id / so l id mixtures , very v iscous f luids, hazardous f luids, l ive organisms suspended i n f luids, low-head or low-submergence situations, scenarios w i t h variable inlet water surface levels , etc. A i r l i f t and other alternate p u m p i n g technologies provide a means for engineers to approach these scenarios. Despi te the success o f the airlift pump i n several other areas, the airlift pump has not gained acceptance i n c i v i l engineering applications. Spec i f i ca l ly it has not been used for management o f s torm waters, p u m p i n g fluids i n open channels, nor i n any other h igh discharge, l o w lift , l o w head applications despite the fact that i n some cases it m a y promise some advantages i n these settings. In fact, an extensive literature search on airlift p u m p research and development found no references at a l l to airlift pumps used i n h igh- discharge, low-head , low- l i f t capacities in c i v i l engineering applications or otherwise. Nevertheless, there are potential applications for l o w head, h igh capacity p u m p i n g o f water i n open channels, and speci f ica l ly o f storm runoff in drainage conduits . T h i s research program is focused on investigating those possibi l i t ies . 1 T h e m a i n feature o f an airl ift pump is a vert ical tube wi th the lower end submerged i n water and a supply o f compressed air p rov ided to the lower end. A s the compressed air f lows into the lower end o f this tube bubbles are formed and a mixture o f water and air bubbles results. S ince this mixture o f air and water is less dense and thus l ighter than water, the l eve l o f the air and water mixture i n the vert ical tube rises above that o f the surrounding water surface. W i t h a suitable phys ica l arrangement, this results in continuous " l i f t i n g " o f the water to a higher leve l than the or ig ina l water surface - i n effect creating an "air l i f t pump" . C a r l Loescher , a G e r m a n m i n i n g engineer, reportedly developed the or ig ina l airl ift p u m p concept i n 1797 and the technology began to gain widespread acceptance i n the m i d d l e 1800's ( W a r d 1924). B y the early 1900's several patents had been issued for various arrangements o f airlift pumps and they were w i d e l y used for p u m p i n g water, often f rom quite deep wel l s , un t i l be ing superceded by reliable e lec t r ica l ly-dr iven submersible ro todynamic pumps. There was a considerable amount o f early research into the airl ift concept, but broad interest on the topic waned as airlift pumps were superceded i n the early 1900's . Despi te having been replaced i n c o m m o n use, airlift pumps have cont inued to be used i n several specia l ized applications, w h i c h are described in more detail later i n this chapter. 2 T h e c i ty o f R i c h m o n d in B r i t i s h C o l u m b i a , Canada is situated in the mouth o f the Fraser R i v e r and experiences an average 1100 m m of rainfal l per year 1 . The average elevat ion o f the c i ty is approximate ly one metre above mean sea leve l . Because o f this very l o w elevat ion, m u c h o f the ci ty w o u l d be submerged under t idal or r iver water dur ing various parts o f the year were it not for the extensive system of levees protecting R i c h m o n d f rom the Fraser R i v e r and the ocean waters o f G e o r g i a Strait. Recent init iat ives to further improve the c i t y ' s protection f rom r iver and sea f lood waters have been to p lan for the instal la t ion o f a so-cal led mid- i s l and d ike to help prevent Fraser R i v e r waters f rom inundat ing central R i c h m o n d i n the event o f a levee breach i n the Eastern region o f the munic ipa l i ty . The average ground slope i n R i c h m o n d is zero and thus the mun ic ipa l stormwater management system is constrained to very l o w slopes in its ' main conduits . The p r o b l e m o f l o w slopes is compounded by the necessary levee system used to protect the c i ty . T h e levees create a need to pump stormwater out o f R i c h m o n d when tides are unfavorable, par t icular ly i n the winter when the tides are relat ively h igh and constant. T h e stormwater management system i n R i c h m o n d relies on l o w tides to a l low the outfall flapgates to open. In the winter months the tides tend to be h igh and constant, w i t h the da i ly second l o w tide s t i l l very h igh . T h i s is unfortunate t im ing since the winter months are often very rainy i n the L o w e r M a i n l a n d . These factors result i n a real and ongoing danger o f winter f lood ing in the c i ty o f R i c h m o n d . 1 City of Richmond Engineering staff graciously provided the background information on their stormwater drainage system as presented in this brief section during various site visits, meetings and conversations that took place from 1997 to 1999 throughout the development of this research project. 3 D u r i n g heavy rains, pumping stations at the perimeter outfalls o f the system can p u l l the l oca l water levels d o w n to shutoff but there can s t i l l be f looding i n central R i c h m o n d because o f the inabi l i ty to move storm water q u i c k l y enough to the outfalls. Recent experience has shown that R i c h m o n d experiences unacceptable s torm water levels and f lood ing i n some areas as frequently as once every two or three years. Because o f these concerns and increasing high-density development i n the urbanized core o f R i c h m o n d , the c i ty has been consider ing options for i m p r o v i n g the capaci ty o f their stormwater management system. Concentrat ion times are short so either faster r emova l o f runoff or detention and storage is required. Detent ion and storage is problemat ic g iven the h igh water table in R i c h m o n d , so the approach has been to consider options focused on increasing the rate o f runoff removal . T h e first op t ion presented was to introduce more and larger conduits. Unfortunately this strategy w o u l d be extremely expensive and very problematic i n pub l ic inconvenience since many o f the ma in stormwater conduits are instal led in buil t -up areas and under m a i n c i ty roads. A d d i t i o n a l l y , installation o f large concrete box culverts has become very unattractive since B r i t i s h C o l u m b i a worker ' s protection legis la t ion concern ing the condi t ions required for their maintenance is so strict that it makes the upkeep o f such conduits imprac t ica l and very expensive. T h e second alternative was to investigate means o f increasing the effective slope o f the system by increasing the water surface grade wi th in the conduits by pumping . T h i s 4 second approach w o u l d accelerate the mean flow veloci t ies and thus remove stormwater f rom the c i ty core at an increased rate. A need for h igh capacity, l o w head pumps that c o u l d be instal led i n s torm drainage conduits to lift s torm water between 1 and three feet (0.3 to 1.0 m) had developed. S u c h pumps w o u l d increase the effective slope and hence the discharge capacity o f the ex is t ing s torm drainage infrastructure. These pumps w o u l d on ly be required for short durations under the combina t ion o f heavy rainfalls and h igh tides. C o m m o n ro todynamic pumps do not conform to this h igh-f low, low-head requirement and i f pump units c o u l d be found to satisfy these requirements they w o u l d s t i l l be expensive to instal l and house in the R i c h m o n d system. Th i s is because o f their need for m i n i m u m submergence levels at their inlets, necessitating substantial excavat ion and placement o f infrastructure in an area w i t h sandy soi ls and a h igh watertable. The diff icul t ies and impract ical i t ies i n both o f the proposed solut ion strategies have effect ively stopped R i c h m o n d ' s progress towards an improved stormwater management system. Despi te the impasse however, the danger o f f looding i n central R i c h m o n d is real and increasing as urban development continues. R e a l i z i n g the need for a way forward, ah alternative pumping technology was sought and this requirement spurred R i c h m o n d into sponsoring the appl ied research program that is described i n this thesis. 5 1.2 - Description of an Airlift Pump A n airl ift pump i tself is compr ised o f f ive major components, namely the air supply apparatus, the air injection or aeration system, the water intake, the riser pipe and the p u m p outlet. F igure 1 shows the ma in elements o f an airlift pump. Nomencla ture used i n that figure and other variables o f interest are defined i n Table 1. A n airl ift pump may also feature a so-cal led "foot piece", a lengthened section o f the m a i n riser pipe located be low the aeration point and i n w h i c h on ly single-phase water f lows . A foot piece a l lows an airlift pump to entrain water f rom a depth greater than i ts ' aeration depth. T h i s a l lows a means for pump units w i th l o w head air-supply apparatus to successfully pump l i q u i d f rom m u c h deeper levels than they w o u l d otherwise be capable of. S ince foot pieces are required on ly i n scenarios i n w h i c h the water to be pumped is to rise f rom a great depth not a l l airlift pumps feature foot pieces. In fact, most short airl ift pumps such as those i n this study, do not use foot pieces. 6 F I G U R E 1 - S c h e m a t i c A i r l i f t P u m p L a y o u t T A B L E 1 - A i r l i f t P u m p N o m e n c l a t u r e a n d V a r i a b l e s A r e a = cross-sectional area o f airlift pump tube b = tuning parameter in air phase ve loc i ty /mixture ve loc i ty relat ionship A a i r = area o f f l ow mixture cross section occupied by air A w = area o f f l ow mixture cross section occupied by water c = tuning parameter i n air phase ve loc i ty /mixture ve loc i ty relat ionship d = tuning parameter i n head loss equation D i a m = diameter o f airlift pump tube Dens = relative density o f the air-water mixture i n the airlift pump tube e = tuning parameter in head loss equation g = acceleration due to gravity Hdrive = d r iv ing head appl ied to airlift pump Hlift = lif t height o f air-water mixture i n airlift pump tube Hioss = head loss i n airlift pump tube Htota = height o f pump lift above aeration point Hfoot = height o f pump tube footpiece be low aeration point H s u b = height o f standing water surface above aeration point Kent = pump tube entrance loss factor Kexit = pump tube exit loss factor Kpipe = pump tube pipe loss factor Ktotal = total pump loss factor Qair = vo lume f low rate o f air in airlift pump tube Qmix = vo lume f low rate o f the air-water mixture in airlift pump tube Qwater = vo lume f low rate o f water i n the airlift pump tube V a i r = ve loc i ty o f the air fraction i n the air-water mixture in airl ift pump tube V v mix = ve loc i ty o f the air-water mixture in airlift pump tube V r e l = relative ve loc i ty o f the air phase to the water phase in the airl ift pump tube Vwater = ve loc i ty o f the water fraction i n the air-water mixture in airlift pump tube Tlsystem = airlift pump system eff iciency Tlairdelivery = airlift pump air del ivery subsystem eff iciency Tlriser = airlift pump riser tube subsystem eff iciency Pair = density o f gas phase Pwater = density o f l i q u i d phase 8 A s ment ioned previous ly , i n this study on ly airlift pumps wi th zero-length footpieces are considered, so Hf 0 0 t = 0 and the total length o f the pump riser tube is equal to the s u m o f the submerged and unsubmerged lengths, H S U b and Hijf t. 1.3 - Applications of Airlift Pumps Despi te hav ing been superceded by submersible ro todynamic pumps i n most c o m m o n applicat ions, airl ift pumps are s t i l l used i n several specia l ized settings. T y p i c a l modern applicat ions o f ex is t ing airlift pump technology include use o f these pumps i n deep water we l l s , where a related system k n o w n as a "geyser p u m p " is also becoming increas ingly c o m m o n where small-diameter pump tubes are feasible. A i r l i f t pumps also s t i l l frequently serve deep shaft and w e l l d r i l l i n g applications. A i r l i f t pumps are also used i n modern w i n d m i l l - d r i v e n pneumatical ly-operated water- w e l l p u m p i n g applicat ions, such as those available as turnkey systems f rom A i r l i f t Technologies o f Redlands , C A . Despi te the fact that m i n i n g technology has developed dramatical ly , airlift pumps are s t i l l a staple i n mine dewatering, and modern examples are remarkably s imi la r to the o r ig ina l system developed by Loescher i n 1797. A i r l i f t pumps are also often used in process applicat ions i n w h i c h corrosive or viscous l iquids such as sand-water slurries, salt solut ion, o i l s and various other waste products make tradit ional ro todynamic pumps less suitable. (Gio t , 1982) T h e o i l industry uses airlift pumps i n retr ieving crude o i l f rom dead wel l s . T h e nuclear industry uses carefully calibrated smal l diameter airlift units to pump fluids i n nuclear fuel retreatment (Clark & Dabo l t 1986). 9 Wastewater treatment plants are currently the most c o m m o n appl icat ion for airlift pumps , where excel lent aeration and subsequent oxygenat ion o f the pumped mixture that is der ived f rom the injected air is a strong benefit. The Sanitaire company o f B r o w n Deer , W I bui lds stainless steel airlift pumps for this applicat ion. A i r l i f t pumps are often used i n aquaculture and f ish farming operations where their l ack o f m o v i n g parts provides necessary safety for f ish and the air introduced into the water c o l u m n improves oxygenat ion (Wurts , M c N e i l l & Overhul ts , 1994). T h e Aquacare company o f B e l l i n g h a m , W A manufactures airlift pumps for fish fa rming applicat ions. The compet ing technologies used in fish farming, namely geyser pumps and propel ler pumps, have the respective disadvantages o f noise and possible damage to f ish safety i n aquaculture applications. Offshore minera l excavat ion and d iamond m i n i n g is an emerging appl ica t ion for airl ift pumps, where the lack o f m o v i n g parts and abi l i ty to handle particulates make them par t icular ly suitable. A i r l i f t pumps are also sometimes used in a s imi la r manner for underwater recovery and salvage operations, where an airlift tube may be r igged and powered f rom the surface, a l l o w i n g divers to place smal l items at the intake o f the p u m p and have the items carr ied to the surface. A i r l i f t pumps for use in deepwater salvage often feature tapered riser pipes, presumably so that as air bubbles increase in size dur ing their rise f rom the aeration point towards the surface the v o i d ratio o f the mixture i n the pump tube does not increase too much and reduce eff iciency. The airlift pump is very w e l l suited to underwater recovery purposes since compressed air is a staple aboard salvage 10 vessels and the turbulent nature o f the f l ow i n the airlift pump tube as w e l l as the upwards-opening shape o f the commonly -used airlift pump barrels i n this appl icat ion are doubtless helpful i n avo id ing any potential j a m m i n g irregularly shaped items may experience i n the pump risers. The f inal c o m m o n appl icat ion o f airlift pumps is i n lake turnover, where these pumps are used to counter the effects o f lake stratification (Parker & Suttle 1987). In lake destratification applications airlift pumps often float on smal l buoys w i t h their outlets at the water surface and compressed air del ivered by f loating supply l ines (Wurts , M c N e i l l & Overhul ts 1994). 1.4 - Project Scope and Rationale T h i s research project aims to investigate the sui tabil i ty and behaviour o f air l if t pumps i n a new class o f applications - namely low- l i f t , h igh- f low, low-submergence scenarios such as p u m p i n g i n open channels and management o f urban storm drainage. Despi te the unorthodox concept, airlift pumps promise many advantages i n such applicat ions. Installed costs are l o w since the pumps are s imple , composed p r imar i ly o f c o m m o n l y avai lable P V C or steel p ipe fittings. A i r l i f t pumps are very robust and nearly maintenance-free since they have no underwater m o v i n g parts (de Cacha rd & De lhaye 1996). A d d i t i o n a l l y , their air supply systems can be located convenient ly above ground to m i n i m i z e instal lat ion costs and facilitate inspection and maintenance. 11 The following discussion of airlift pump efficiency suggests that the low-head, low- submergence, high-flow, necessarily large diameter airlift pumps that would be required in open-channel and urban storm drainage applications would be energy inefficient units. Despite this inefficiency, the author believes that airlift pumps may offer enough other cost and service advantages to offset the operational inefficiency of the airlift pumps that would be applied in these settings. Some of the advantages airlift pumps may offer to urban drainage applications include low installation cost and maintenance cost, very low supporting infrastructure cost, and a possibly huge placement benefit in the potential option for portable pump units and/or portable power units, thus potentially eliminating entirely the need for a pump house or similar infrastructure. The aeration of storm runoff may also be a reason to consider the application of airlift pumps to urban drainage applications. Urban storm runoff often contain high levels of heavy metals, petrocarbons, chemicals from spills, and other roadwash pollutants and tend to create potentially significant environmental impacts to the bodies of water into which they discharge. (Turer, Maynard & Sansalone, 1996). Airlift pumps are used routinely in aquaculture and wastewater treatment because of their significant benefit in aerating the pumped liquid. Using airlift pumps for urban drainage could provide the additional benefit of aerating the storm runoff, thereby mimicking the aeration process used in many municipal mixed-sewage treatment plants, potentially accelerating 1 2 oxida t ion o f the roadwash and other stormwater pollutants, and a l l o w i n g for a decrease in result ing environmental impacts. T h i s c o m p e l l i n g array o f advantages, part icularly the very l o w cost o f instal lat ion and maintenance o f airlift pumps, the lack o f a need for permanently instal led power and control systems w i t h their attendant housing infrastructure, and the potential benefit o f aerating the runoff waters make the investigation o f airlift pumps for these applicat ions very attractive. 13 1.5 - Two-Phase Flow Regimes A i r l i f t pumps are two-phase f l u id f low devices. Gas and l i q u i d ( in most cases air and water) f l ow upwards together in a vert ical pipe. T h i s two-phase f lu id mixture can take several different forms, and the various f l ow patterns o f the two phases i n these forms have s igni f icant ly differ ing hydraul ic behaviours. Th i s is significant to the science and design o f airl ift p u m p i n g systems since any o f these forms o f air-water mixtures are possible , and the fo rm found i n the system o f interest is a very important var iable since phys ica l relationships and der ived mathematical relationships are unique for each form. T h e exact descriptions o f the f low patterns vary somewhat by author, t e rminology is not a lways c o m m o n , and some f lows are described as combinat ions o f patterns (Shel ton & Stewart, 2002) . A summary o f the f ive basic forms observed in the two-phase f l ow o f water and air i n vert ical pipes, a long w i t h their most c o m m o n names are shown in F igure 2. (modi f ied f rom Ta i t e l , B o r n e a & B u c k l e r , 1980). F igure 3 shows the same f low regimes characterized by gas f lux and mixture veloci ty . 14 FIGURE 2 - Two-Phase Air-Water Flow Regimes Bubbly Churn Slug Dispersed Annular Flow "Froth" Flow Annular "Film" Flow "Ripple" Flow Flow 15 F I G U R E 3 - Two Phase Flow Regimes characterized by Gas Flux and Mixture Void Fraction, adapted from Wallis (1969) PERFORATED PLATES 3 0 U o • x NO. OF ORIFICES' DIAMETER(cm) SQUARE ARRAY SPACINGfcm^ I 4 9 1 0 0 2 8 9 4 . 0 6 X 10 4 . 0 6 X 10 1.52 X 10 0.41 X 10 - l - l 6 . 2 5 X 9 . 5 0 X 6 . 2 5 X 10 10 10 -I I -I 2 5 E o o 2 0 X ID CO < CD 15 10 CHURN TURBULENT REGIME « X X X X z UJ _ l QC Z> X o XX X X X X X X x x X X X * X X X x X X X X x A X X X X X XX X X XX IDEAL BUBBLING REGIME 0.1 0 . 2 0 . 3 VOID FRACTION - oc € . 4 UJ Z o u cc z o z <t cc 16 1.6 - Operat ional Efficiency of A i r l i f t Pumps A i r l i f t pump eff ic iency can be defined as the ratio o f energy del ivered to the pump unit to the unit energy output i n the form o f ve loc i ty and head o f the pumped l i q u i d . The overa l l system eff ic iency can be considered a product o f the air de l ivery and airl ift riser subsystem efficiencies. T h e eff iciency o f the air del ivery subsystem depends on the type and configurat ion o f the air supply equipment, p ip ing , conduits and controls . Ef f ic ien t de l ivery o f air through instal led conduits at desired pressures and f low rates is a w e l l - explored and mature branch o f mechanical engineering. C o m m o n w i s d o m i n the design and use o f airlift pumps suggests the eff ic iency o f an airl ift pump riser subsystem is m a x i m i z e d when deep submergence is avai lable , the lift height is l o w , and l i q u i d and air f low rates are l o w . D e Cacha rd & De lhaye (1995) and D e Cacha rd (1989) also suggest a very strong contr ibut ing effect i n the length-to- diameter ratio, namely that slender pumps wi th h igh length-to-diameter ratios are greatly more efficient than their l o w length-to-diameter ratio counterparts. T h e most efficient airlift pumps feature a situation i n w h i c h the air and water phases have very s imi l a r veloci t ies , air bubbles are either spherical and very smal l or are large, dart- shaped T a y l o r bubbles w i t h a cross section near the entire pipe diameter. In both o f these m a x i m a l l y efficient cases, the s l ip ve loc i ty between the air bubbles and water is m i n i m i z e d . 17 A i r l i f t p u m p eff ic iency is further enhanced by use o f the smallest possible stable v o i d fraction - thus p u m p i n g the m a x i m u m amount o f water per amount o f air injected. Aera t ion eff ic iency is also an important factor i n determining the eff ic iency o f short airlift pumps al though it matters less i n l ong pumps ( W a l l i s 1968). Th i s phenomenon appears to occur because the long airlift pumps tend to operate i n s lug f low. S l u g f l o w occurs i n pumps long enough that smal l bubbles can accrete together to fo rm homogeneous ly spaced large T a y l o r bubbles close in cross section to the p ipe diameter (Tai te l & a l . 1980). In this f low regime the f l u id f lows cont inuously i n contact w i t h the pipe wal l s creating losses direct ly dependent on the f lu id ve loc i ty . A t the entrance o f l ong pumps (and i n shorter airlift pumps in w h i c h smal l bubbles do not have the opportunity to accrete into T a y l o r bubbles before ex i t ing the pump riser) the air and water mix ture f low is turbulent and recirculatory. Ta i te l & al (1980) characterize this f l ow regime as "froth" or "churn" f low and identify it by the osci l la tory nature o f the l i q u i d ' s upward and downward mot ion between and around bubbles. A n aerator assembly that diffuses many smal l evenly distributed bubbles into the f l ow f i e ld helps reduce this turbulence and recirculat ion, reducing losses and increasing eff iciency. M o r r i s o n & a l . (1987) suggest that this is also true for the bubbly f low regime where "mul t ipor t injection is more efficient". Despi te early and contradictory observations such as those by W a r d (1924) and Baue r & P o l l a r d (1945) on large diameter airlift pump systems, riser diameter also plays a role i n airlift pump eff iciency for a g iven lift height since larger diameter airlift pumps tend to be 18 less efficient than their smaller counterparts. Th i s is because the larger diameter pumps must be very l ong before the efficient T a y l o r bubble- induced s lug f low regime can s tabi l ize (De Cacha rd & Delhaye 1995). In fact, as the pipe diameter increases the cross sectional area increases even more rapidly , thus d imin i sh ing the abi l i ty o f surface tension forces to h o l d large bubbles intact against the influence o f a complex turbulent shear f i e ld i n the air/water mixture co lumn . D e Cachard & Delhaye (1995) also suggest that surface tension forces in bubbles reduce s l ip veloci t ies between the air water phases. In that case since bubble surface tension forces are increased i n smal l diameter pipes, the reduced eff ic iency o f large diameter pumps may be due to greater s l ip veloci t ies , themselves due to the reduced relative effect o f surface tension forces. Observat ion suggests that as the pipe diameter increases above a m a x i m u m feasible bubble diameter the thickness o f the f i l m i n the annular region surrounding the T a y l o r bubbles i n the s lug f low mixture may begin to thicken rapidly . T h i s rapidly th ickening f i l m c o u l d then provide a dramatical ly increased f low path area for l i q u i d f rom the region ahead o f any T a y l o r bubble to s l ip downwards through the annular-shaped region, past the T a y l o r bubble and into the region behind the bubble. A s the f low rate o f the downward- t rave l ing f l u id in the annulus regions increases, the overa l l fr ict ional shear on the p ipe tube m a y become downward ( W a l l i s 1968). In such cases the overa l l l ift e f f ic iency falls rap id ly . Thus , increasing pipe diameter above the stable bubble diameter for a g iven f l o w f i e ld may reduce efficiencies for l ong airlift pumps operating i n the s lug f low regime. 19 1.7 - Summary T h e mot iva t ion for this w o r k is "Can apply airlift pump technology be prac t ica l ly appl ied to c i v i l engineering works such as open channel drainage o f urban s tormwater?" A n airlift pump is a deceptively s imple two-phase f low device than can operate i n several f l o w regimes, depending on several geometric and f low parameters. A i r l i f t pumps have been the subject o f a smal l amount o f research since their invent ion i n 1797 by C a r l Loescher . S ince then they have been appl ied extensively i n a smal l number o f spec ia l ized applicat ions but not to h igh- f low, low- l i f t , low-submergence c i v i l engineering applicat ions such as open-channel drainage and storm water management. A i r l i f t pump eff ic iency is m a x i m i z e d i n scenarios i n w h i c h submergence is h igh , gas and l i q u i d f l ow rates are l o w and aeration eff iciency is h igh. Despi te the fact that low-head h igh- f low applicat ions do not promise very efficient operation o f airlift pumps there are s ignif icant reasons such as l o w instal led cost, ease o f maintenance, reduction o f environmental impact o f runoff water, etc. to investigate them for these uses. T h i s thesis considers the appl icat ion o f airlift pumps to these c i v i l engineering applicat ions and outlines a four-stage experimental program undertaken at the U n i v e r s i t y o f B r i t i s h C o l u m b i a and the C i t y o f R i c h m o n d , B r i t i s h C o l u m b i a . T h i s study had several objectives. These were namely: to first evaluate the potential for airlift pumps i n urban drainage and other low-s lope open-channel applications, to create a mathematical descript ive mode l o f low-head, h igh discharge airlift pump systems, to develop a pract ical design method for such pumps us ing the mathematical mode l above, and to 20 illustrate the use of that method. To these ends, small-scale and full-scale models were built and tested. Water and air flows and levels were recorded. A broad literature study was undertaken. From this theoretical background and experimental observations, three mathematical models for predicting airlift pump behaviour in these settings are developed. One is suggested as representative. A simple hand-calculator design procedure is explained and two personal computer-based solutions are suggested. A practical design example is presented, and conclusions and recommendations for further development are made. 21 C H A P T E R 2 2.1 - Introduction - Literature Review T h i s chapter describes a br ie f history o f the open literature on airlift pumps, p r o v i d i n g an ove rv iew o f the development and theory behind their operation as w e l l as an ove rv iew o f airl ift theory development to the present day. Th i s literature search first investigated the his tor ical use o f airlift pumps in c i v i l engineering applications. N o ment ion was found. Broaden ing the scope o f the search revealed a niche body o f literature concern ing airl ift pumps i n the process engineering f ie ld , documented ma in ly i n the d isc ip l ines o f Aquacu l tu re and C h e m i c a l Engineer ing . 2.2 - Development of Airlift Pump Theory, 1797 to Present A s noted i n Chapter 1, C a r l Loescher , a German m i n i n g engineer, is thought to have been first invented the airlift pump in 1797 (Gio t 1982). Loescher ' s invent ion was an attempt to s impl i fy the p u m p i n g tasks in deep mines. Submers ible rotomachinery was not avai lable i n the late 1700's and the benefits o f a pneumatical ly-operated system are immedia te ly evident i n that context. A i r l i f t pumps became popular several decades after Loescher ' s first models , dur ing the midd le 1800's ( W a r d 1924). A t this t ime direct pneumatic power was w i d e l y avai lable i n the fo rm o f boi le r steam w h i c h was easi ly generated at h igh pressure. Pneumat ic power was also avai lable f rom steam-powered compressors. Faraday ' s d iscovery o f electromagnetic induct ion i n 1831 led the way to the invent ion o f the electric motor. T h i s 22 and the appearance o f the internal combust ion engine pioneered by R u d o l f D i e s e l and others at the end o f the same century made compressed air a v iable source o f power . S h a w (1920) first suggested a vo lume ratio for the gas and l i q u i d phases i n a long airl ift pump riser tube operating at 100% efficiency: Volumeair = QWMerg-H, \7~1 f rt Volumewaler discharge In (P aerationdepth (1) aerati P discharge J S h a w ' s is the first attempt found i n the open literature to quantify airlift p u m p behaviour on a phys ica l basis. E v i d e n t l y his relationship was successfully used i n design w i t h an eff ic iency mul t ip l i e r added, on the order o f 5 0 % (Zenz 1993). W a r d at the Un ive r s i t y o f W i s c o n s i n d id the first serious experimental study o f air l if t pumps found i n 1924. T h i s study focused on the behaviour o f long airlift pumps and attempted to create functional relationships between the air and water phase f l o w rates, eff ic iency and pump riser geometry such as pump length and diameter. W a r d developed an elaborate curve-f i t t ing a lgor i thm for use i n design but was on ly moderately satisfied w i t h the results and qual i f ied the technique's appl icat ion to the long pump risers i n his study. 23 W a r d presented sixteen summary conclusions in his study. M a n y o f W a r d ' s results and suggestions fo rm the ongoing c o m m o n basis for subsequent use and understanding o f airl ift p u m p i n g systems. Here is a summary o f W a r d ' s eight most salient conclus ions : 1. T h e eff ic iency o f long airlift pumps depends p r imar i ly on f low condi t ions i n the riser pipe, and thus great refinement i n aeration and foot piece design beyond ensuring m i n i m u m f low restriction at the entrance are not necessary in most cases. 2. There is a m a x i m u m eff iciency for every combina t ion o f pump geometry and submergence that depends on water f l ow rate. 3. M a x i m u m eff iciency occurs at submergence ratios o f greater than 7 0 % i n most cases, (ie: when over 7 0 % o f the total length o f the riser tube is submerged) al though very sma l l diameter pumps can operate wi th relat ively high eff ic iency at l ower submergence ratios. 4. H i g h eff ic iency is possible at lower submergence ratios i f the aeration depth is deep. 5. The combined fr ic t ion and s l ip losses due to the f low i n airlift riser pipes f o l l o w a different l aw than those that govern the f low o f water or air in a pipe. 6. There is a relat ively s imple relation between fr ict ional losses and ve loc i ty o f f l o w i n an airlift riser pipe for any particular mixture o f air and water. 7. S m o o t h joints i n airlift riser pipes are necessary to avo id unnecessary losses. Sudden expansion or contraction is very detrimental to efficient operation. 8. A i r lift pumps o f less than forty feet in length are l i k e l y to give results m u c h different to those encountered i n long pumps. Losses that are relat ively insignif icant i n large pumps become important i n short airlift pumps. 24 E i g h t years later i n 1932, Picker t publ ished "The Theory o f the A i r l i f t P u m p " i n an attempt to elaborate on the mechanics o f the f low i n these units. H i s study d i d not present results greatly contr ibutory to the behaviour o f the large diameter, l o w lift , l o w submergence h igh f low pumps o f interest i n this study. M o r e than 25 years passed unt i l G o v i e r , Radford & D u n n ' s "The U p w a r d s V e r t i c a l F l o w o f A i r - W a t e r M i x t u r e s " appeared in 1957. The i r experimental study was based on a 1.025" diameter pump riser 30 ' long (ie: length-to-diameter ratio approximately 350:1). T h e y were able to accurately predict f low pattern, head loss and sl ip ve loc i ty but restricted the appl icat ion o f their results to the behaviour o f pump units o f s imi la r riser tube diameters when p u m p i n g mixtures o f s imi lar gas and l i q u i d properties. D J N i c k l i n ' s "The A i r l i f t P u m p : Theory and O p t i m i z a t i o n " o f 1963 presented the first satisfactory explanat ion o f the behaviour o f small-diameter airlift pumps i n the bubb ly and more important ly, the s lug-f low regimes. N i c k l i n ' s momentum balance, 2-phase drift f lux mode l based on mass f low forms the basis o f the bu lk o f subsequent research into airl ift pumps and s lug f low theory and behaviour. The most broadly used o f N i c k l i n ' s conclus ions (recast here i n consistent terminology for this study) is used to characterize the ve loc i ty o f T a y l o r bubbles i n the s lug f low regime in s t i l l water: (2) V„ taylorbuhble ~ 0.35 • g • Diam 25 N i c k l i n also observed ( l ike Ward ) that although many aerators have been designed to m i n i m i z e bubble size and m a x i m i z e bubble distr ibution, none were successful i n l ong airl ift pumps. H e also first c lar i f ied the one-to-one relationship between the submergence ratio and the average pressure gradient i n the pump riser tube. Mul t i phase f l ow was s t i l l a nascent f ie ld in the 1960's and developments i n this area were happening rapid ly . In 1964, Duck l e r , W i c k s & C l e v e l a n d publ i shed a two-part study "F r i c t i ona l pressure drop i n two-phase f l o w " . The i r results are i l lustrat ive o f the s t i l l - deve lop ing nature o f two-phase f low theory at that t ime. T h e y found the ex is t ing correlations for pressure loss i n two-phase pipe f low to be inadequate and asserted that "There is not even a phenomenologica l understanding o f this type o f f l o w . " A s two-phase f l ow theory was further developed, and due poss ib ly to the exp l i c i t so lu t ion for s lug f low operation as suggested by N i c k l i n , the study o f airlift pumps cont inued to focus increas ingly on the mechanics o f T a y l o r bubbles i n the s lug f low regime, and to an increas ingly lesser degree on the bubbly f low, churning f low and annular f l o w regimes. W a l l i s ' def ini t ive work , One Dimensional Two-Phase Flow, appeared i n 1969. W a l l i s ' text is s t i l l one o f the best sources for a broadly-focused co l lec t ion o f most o f the open theory o f one-dimensional two-phase f low. W a l l i s ' work exposes the tremendous complex i ty i n mult iphase f low behaviour and provides much o f the foundation for two- phase f low as used today. M a n y fr ict ional and ve loc i ty relationships developed by W a l l i s are s t i l l state o f the art in modern two-phase f low theory. 26 T o d o r o s k i , Sato and H o n d a fo l lowed N i c k l i n ten years later i n 1973 w i t h "Performance o f A i r l i f t Pumps" , w h i c h elaborated s l ight ly on N i c k l i n ' s approach to the s lug- f low regime f l o w o f these devices. Todorosk i , Sato and H o n d a modi f i ed N i c k l i n ' s exper imental basis for determination o f the sl ip velocit ies in s lug f low. T h e interpretation o f the various regimes o f vert ical two-phase f l ow was m a i n l y descript ive i n nature unt i l 1980 when Ta i te l , Ba rnea & D u c k l e r publ i shed " M o d e l i n g F l o w Pattern Transi t ions for Steady U p w a r d G a s - L i q u i d F l o w in V e r t i c a l Tubes" . T h e i r study undertook mathematical ly predict ing the transitions between these patterns. T h e y were able to predict w h i c h pattern or regime o f two-phase f low w o u l d occur under a g iven set o f condi t ions, and their approach is s t i l l used today. Ta i te l , Ba rnea & D u c k l e r also provide the best o f the v isua l descriptions o f two-phase f low regimes ( w h i c h suppl ied the basis for F igure 2 in chapter 1). The i r most useful f ind ing for this current study suggests that ( in cases in w h i c h s lug f low can develop) the length o f the turbulent entrance or transit ion zone region f rom the aeration point to the point where s lug f low can develop depends on the mixture veloci ty and pipe diameter: France = 40.6 Diam yjg Diam + 0.22 (3) J In 1982, Marka tos & Singhal produced a numerical analysis process for two-phase f low. T h i s study was focused on bubbly and s lug f l ow , much the same as those that had preceded it. It appears that in the bubbly and part icular ly the s lug f low regimes the mathematical formulat ion for the fr ict ion and loss terms is easier to accompl i sh since the 27 re la t ively f ixed geometry o f the round or T a y l o r bubbles a l l ow a solut ion that requires less experimental data for correlat ion. Marka tos & S ingha l ' s technique was developed for use i n deep water we l l s and depends on the breakdown o f long vert ical risers into smal ler cont iguous segments, i n effect creating a "gradual ly va ry ing f l o w " formulat ion. It is suitable on ly to l ong riser pipes. L o n g , smal l diameter airlift pumps are w ide ly used in nuclear fuel reprocessing. V e r y accurate estimates o f f l ow rates are required i n those settings. In 1986 C l a r k & Dabo l t developed a general set o f design equations for airlift pumps in s lug f low for use i n the nuclear industry. T h e y focused p r imar i ly on accurately predic t ing the f l o w rate behaviour in their applicat ions. Despi te their admitted inabi l i ty to accurately calculate the overa l l f r ic t ional losses i n pipes o f 38 m m diameter they d i d provide an accurate design mode l for such pumps i n the s lug f low regime. Interestingly they also attempted to apply N i c k l i n ' s mode l to a short pump and found that N i c k l i n ' s mode l overpredicted the p u m p eff ic iency, un l ike i ts ' better agreement when appl ied to longer units. C l a r k & D a b o l t ' s general design equation for long , small-diameter pumps does not address pump eff ic iency i n great detail but does provide an accurate and pract ical means o f design for very l o n g slender airl ift pumps. In 1993 Z e n z produced " E x p l o r e Potential o f A i r l i f t Pumps and Mul t iphase Sys tems" p r imar i l y exp lo r ing airlift pumps i n three-phase scenarios. Z e n z ' study was concerned m a i n l y w i t h s lug f low and particulate entrainment, again in l ong pipes. A i r l i f t pump riser pipes are generally considered long when length to diameter ratios are 50:1 or more. 28 Wur t s , M c N e i l l and Overhul ts (1994) prov ided a s imple curve-fi t t ing approach to airl ift pump performance for near-100% submergence i n aquaculture and destratification applicat ions. In 1995 Tramba , T o p a l i d o u , Kas t r inakis , N y c a s , Francois & Scr ivener comple ted their " V i s u a l S tudy o f an A i r l i f t P u m p Operat ing at L o w Submergence Ra t io s " w h i c h is not par t icular ly helpful for this present study since it is ma in ly concerned wi th bubble format ion at a jet inlet and includes no performance data for non-s lug f lows . F o l l o w i n g C l a r k & Dabo l t i n the nuclear fuel reprocessing industry, D e Cacha rd & Ca lhaye created a steady-state mode l for very smal l diameter, l ong lift pumps i n 1995. D e Cacha rd & Ca lhaye ' s is certainly the most extensive study found. It is concerned p r imar i ly w i t h creating an accurate mode l for gravitational and fr ict ional components o f the airl ift pump riser pressure gradient. L i k e C l a r k & Dabo l t ' s work , it is focused on very long "slender" airlift pumps i n the s lug f low regime. D e Cacha rd & Delhaye are not concerned w i t h opt imiza t ion o f energy eff iciency since energy inputs are very sma l l i n their cases o f interest. D e Cachard & De lhaye observed churning f low i n the lower sections o f their study pump units and concur w i th previous researchers that churn f low is a development phase for the s lug f low pattern. Howeve r , they also found that churn f low c o u l d exist as a stable f low pattern at h igh gas f l ow rates. D e Cacha rd & De lhaye developed the most detailed and accurate analysis f ramework avai lable for airl ift pumps o f under 40 m m diameter and wi th length-to-diameter ratios above 250:1 . 29 M o s t recently Nenes , Ass imacopu los , Marka tos , & M i t s o u l i s comple ted " S i m u l a t i o n o f A i r l i f t Pumps for Deep Water W e l l s " i n 1998. The i r analyt ical f ramework involves an interspersed cont inua mode l and solves a system o f differential equations per Marka to s (1982). Results f rom their system are very accurate but unfortunately suitable on ly for very ta l l p ipe units i n w h i c h the pump riser tube may be broken into tens o f internal ly contiguous discrete elements. 2.3 - S u m m a r y o f L i t e r a t u r e R e v i e w A i r l i f t pumps have been a niche interest area i n process, chemica l and mechanica l engineering as w e l l as aquaculture. Publ ica t ion on the topic has been sparse and the literature has tended towards attempts to expla in the behaviour o f these devices i n two dist inct f l o w regimes. N i c k l i n ' s mode l (1963) continues as the base for almost a l l theoretical development whereas the numerica l techniques o f Marka tos , Nenes et a l . (1992) promise a powerfu l toolset for evaluating the behaviour o f l ong airlift units. There has been l i t t le reason to evaluate the h igh- f low, low-head, low-submergence airlift systems and subsequently those applications are s t i l l unexplored f rom theoretical and design standpoints. T h i s has not stopped such pumps f rom being used sporadical ly and often unintent ional ly since in a pract ical sense, s impl i c i ty in f ie ld use has tended towards ins ta l l ing a pipe at an appropriate depth, adding air in an appropriate vo lume and at an appropriate depth to produce the desired results i f possible. In a research sense, the inab i l i ty to accurately measure the relative velocit ies o f the air and water phases except i n the bubb ly and s lug f low regimes have tended towards tuning a theory focused on those 30 regimes alone. H i g h submergence, smal l diameter, short lifts have been t radi t ional ly investigated since low-submergence, high-l if t units provide decreasing eff iciencies. A s W a r d (1924) concluded, there remain inherent gaps in our understanding o f theory that ma in ly arise f rom the fact that the sizes, speeds and distr ibution o f the air bubbles are not k n o w n , yet the s ize and the rate o f ascent o f the bubbles through the air-water mixture are c r i t i ca l variables. Recent research w o r k on air entrainment in fast f l o w i n g water us ing laser opt ica l probe technology w i t h the abi l i ty to measure the behaviour o f the air fraction in a two-phase air-water f low mixture as distinct f rom the overa l l air-water f lu id mixture has recently become avai lable . T h i s technique was employed first by Car te l l ier (1992) and later refined by Serdula & L o e w e n (1998), whose techniques might seem to promise a more rigorous approach to air lift pump design by the direct measurement o f gas and l i q u i d phase veloci t ies and v o i d ratio. Unfortunately, investigation o f the experimental equipment and personal conversations wi th L o e w e n suggest that since the laser system employed is a 'point measurement system it is not suitable for air-water mixtures w i t h recirculatory movement . It cannot resolve differences between upward -mov ing and reci rcula t ing air bubbles and does not function w e l l without distinct boundaries between the air and water phases at the bubble boundaries such as are found i n bubb ly and s lug f low. Unfortunately this means that at least currently the laser measurement technology is inappl icable to the study o f airlift pumps i n the churn f low regime. 31 C H A P T E R 3 3.1 - Overview of the experimental program F o r airl ift pumps to be used effectively i n short lift , h igh f low applications such as s torm drainage conduits , the airlift pump must be able to "lif t" large volumes o f water through a height o f about 0.3 to 0.6 m . F o r example, such an increase in head over a run o f 5000 feet i n re la t ively flat terrain such as exists in R i c h m o n d , B . C i n the 5 ' by 9 ' concrete box culvert leading f rom G r a n v i l l e Street to the Gi lber t R o a d outfall c o u l d easi ly result i n a doubled water surface slope and subsequent dramatic improvements in water f l ow ve loc i ty , and subsequent reduction o f loca l f looding dur ing extreme rainfal l events. A pump system for occasional use i n emergency situations such as might be experienced by R i c h m o n d dur ing the 10-year design f low should ideal ly be inexpensive and maintenance-friendly. S ince the pump units w o u l d run on ly dur ing extreme condi t ions , and since operating costs in extreme condit ions are often accounted for differently than ongoing costs, eff ic iency is on ly important i n so much as it affects the first cost o f the instal lat ion. R u n n i n g costs are less important as the pumps are on ly used dur ing extreme storm events - i n the order o f once every two or three years. Howeve r , instal led cost is important - o f course less expensive is preferred. A i r l i f t pumps promise a very attractive match to these cri teria. The compressed air supply can be dry and out o f the way , i n fact there is no need for the supply apparatus to be permanently located since air can be suppl ied through a f lex ib le pipe. T h i s means that a permanent pumphouse need not necessari ly be used w i t h an airlift system. The pumps themselves can be bu i l t quite 32 inexpens ive ly as they on ly consist o f sets o f tubes and aerators wi th an air supply. Because airl ift pumps are notor iously inefficient compared wi th their ro todynamic counterparts and air compressors are expensive, it is very desirable to have the ratio o f water l if ted to compressed air used be as h igh as possible. These design considerations and potential benefits were investigated, mot iva t ing the current study. P re l iminary calculat ions were made and two series o f p re l iminary quali tat ive experiments were run at the Unive r s i ty o f B r i t i s h C o l u m b i a C i v i l Eng inee r ing H y d r a u l i c s Labora tory to check the concept. C la s s i ca l airlift pump components and layouts were considered and adapted for use i n a low- l i f t situation. S ince the desire was to determine whether a v iable airlift pump cou ld be developed for these applicat ions it was reasonable to start w i th a system that was configured s imi l a r ly to what a w o r k i n g unit might be. F i rs t a small-scale airlift unit was bui l t us ing the fu l l w id th o f a 6 inch wide undergraduate student hydraul ics lab f lume, exp lor ing the conceptual layout for a f u l l - w id th box culvert instal lat ion. Subsequently a s imi la r larger-scale airlift unit was bui l t us ing the fu l l 20-inch w i d t h o f a large hydraul ics lab f lume to check i f the system w o u l d be functional on a larger scale. 33 T h e laboratory experiments p rov ided some valuable insights into the nature o f airl ift performance at l o w lifts, l o w submergence and l o w v o i d ratios. T h e y also showed that airl ift pumps o f this type c o u l d be pract ical . N e x t a sequence o f full-scale prototype experiments was carr ied out at the Gi lbe r t R o a d s torm drainage outfal l i n R i c h m o n d . Several full-scale experimental prototype layouts and systems were planned and tried. Va r ious combinat ions o f upstream and downst ream water levels were investigated wi th va ry ing rates o f air injection. R i se r p ipe diameter was investigated, as was aerator geometry. S ince evaluat ing and m a x i m i z i n g water f l ow for these pump systems was the end goal o f this study, i n a l l cases it was attempted to determine the water flowrate poss ible for a g iven upstream and downstream depth, rate o f air injection, pump riser diameter and aerator geometry. There were many pract ical difficult ies such as unusual ly l o w f lows and water levels in the drainage conduit that was used as the site. Howeve r , these experiments produced several very useful results. T h e y demonstrated that low-head, h igh- f low, l o w - submergence airlift systems d i d provide a viable and practical alternative i n an instal led s torm drainage system. Howeve r , they also showed that there was considerable c i rcula t ion i n the pump tubes and as a result, the water was be ing l if ted several t imes w i t h result ing l o w overa l l eff iciency. T h e y also showed that there was never a steady state situation such as is usual ly assumed in der iv ing theory and formulae. T h e air-water 34 mixture was very turbulent and the air bubbles were increasing and decreasing i n s ize by shearing and coalescence as they rose, in effect a lways in a transient condi t ion . H a v i n g ident i f ied the major p rob lem o f c i rcula t ion and re -cyc l ing , a f inal set o f experiments were set up at the R i c h m o n d P u b l i c W o r k s Y a r d , us ing banks o f tubes 75 to 200 m m i n diameter, as opposed to the 250 and 300 m m tubes used in the Gi lbe r t R o a d prototype experiments. Theory suggested that smaller diameter pipes w o u l d a l l o w less c i rcu la t ion due to a more evenly distributed air phase. A l s o , the riser pipe wal l s w o u l d have a m u c h greater influence over a larger cross-sectional f l ow area. Tr ia l s were also made w i t h the tubes inc l ined instead o f vert ical , as inc l ined tubes w o u l d be easier to instal l and c o u l d be potential ly less cost ly due to a reduced number o f fittings required for construct ion. The results f rom the fourth experimental setup were successful and showed l i t t le c i rcu la t ion , al though there was s t i l l considerable uncertainty as a result o f the transient nature o f the under ly ing phenomenon o f the air bubbles r i s ing, expanding, shearing and coalesc ing. S ince the experimental "pumps" were s imi lar to the proposed f inal design the results were considered acceptably accurate. T h e design concept was thus considered proven and the relationships developed sufficient for design o f a pract ical operating airl ift pump. 35 3.2 - T h e E x p e r i m e n t a l Se tups T h e first experimental laboratory setup is shown i n F igure 4. The purpose o f this first setup was to b u i l d a v i sua l mode l o f the airlift pump as a fu l l -wid th element in a s torm drain scenario. Because o f k n o w n l imitat ions in width , discharge rate, f l ow rate measurement equipment and theoretical knowledge , this first setup was intended to serve as a base for experimentat ion to a id i n understanding airlift pump behaviour , rather than as an instrumented data co l lec t ion experiment. Th i s system was designed to simulate on a very sma l l scale the or ig ina l proposed layout o f an in-culvert airlift p u m p i n g system. Ups t ream water f l owed into the system under a baffle. A n aerator instal led on the base o f the channel suppl ied bubbles to the water co lumn. The air-water mixture then f l owed between the upstream baffle and a downstream baffle, ex i t ing the pump unit at the higher downst ream leve l . The smal l -scale system was instal led i n a s ix - inch wide student hydraul ics lab f lume to test the concept o f a fu l l -wid th airlift system in a rectangular channel . V a r i o u s combinat ions o f upstream and downstream water levels and air vo lume inputs were tr ied i n order to m a x i m i z e the water flowrate g iven any combinat ion o f upstream and downstream levels. Several geometries were tried since a l l o f the various system elements were modular . The most effective layout found is shown i n F igure 4. 36 F I G U R E 4 - F i r s t L a b o r a t o r y A i r l i f t P u m p i n g S y s t e m Preliminary Lab Airlift Pumping System 1 n , o to all dimensions in mm Several aerator designs were also tried, wi th litt le improvement i n eff ic iency. C i r c u l a t i o n was evident as a very important (and previously much under-estimated) effect i n these h igh- f low, low-submergence scenarios. The best per forming setup in the pre l iminary experiment was measured for performance and p rov ided the first clues to the shapes o f pump discharge curves for systems o f this nature. F igure 5 shows the sample data set and resulting discharge curve for this system. T h e second experimental setup was a s imple test intended to determine the poss ib i l i ty o f a larger-scale system based on the same conceptual layout as the first smal l -scale system. The larger scale system bui l t i n a large 20- inch wide hydraul ics f lume at the U n i v e r s i t y o f B r i t i s h C o l u m b i a Department o f C i v i l Engineer ing . Th i s system was designed to determine the feasibi l i ty o f airlift pump technology at prototype scale for very sha l low 37 F I G U R E 5 - First Laboratory Airlift Pumping System Results Preliminary Airlift Pump Curve 0.800 0.700 0.600 0.500 $ 5 0.400 s x 0.300 0.200 0.100 .0  - 0.000 • 4 y - O.O003x J + 0.O< 24x+0.719-\ 10.000 20.000 30.000 Flow Rate, gpm 40.000 50.000 • Q, gpm Poly- (Q. gpm) These operating characteristics are for the model pump which has a width of 156mm and hence art operational lift chamber plan area of 29640 mm2. Lift, mm Q, 1/s Lift, feet Q, gpm 226.000 0.333 0.753 5.283 205.000 0.714 0.683 11.321 182.000 1.176 0.607 " 18.647 162.000 1.667 0.540 26.417 116.000 2.222 0.387 35.222 139.000 2.353 0.463 37.294 67.000 2.500 0.223 39.625 89.000 2.500 0.297 39.625 39.000 2.857 0.130 45.286 7.000 3.077 0.023 48.769 insert ion depths. Several tests were run and results were p romis ing . C i r cu l a t i on was very strongly evident i n the large 18" x 2 0 " lift chamber. Accura te instruments for measur ing the a i r f low i n the system were not avai lable and therefore direct numer ica l results for 38 a i r f low were not col lected. Th i s was not considered a drawback because the larger scale laboratory airl ift unit d i d prove the concept at the prototype scale despite serious submergence l imita t ions and l imitat ions in the air f l ow rate possible f rom the instal led screw compressor-based air de l ivery system. F igure 6 shows the second experimental setup and Tab le 2 shows numer ica l results o f this phase o f the study. F I G U R E 6 - Prototype Scale Laboratory Airlift Pumping System lift chamber length 18" V Qwater i ^ > H u/s H d / s If Qair S e c o n d Phase A i r l i f t P u m p Test Setup U B C C i v i l E n g i n e e r i n g H y d r a u l i c s Labora to ry flume width 20" three-tined 3/4" brass aerator, 3 x 30 ea. orifices 1 mm dia. lift chamber plan view with aerator T 18" _L 20" 39 T A B L E 2 - Prototype Scale Laboratory Airlift Pumping System Results Airlift Flow Test 2 Aaron Bohnen UBC Civil Engineering Hydraulics Lab Flow baffle height Hds Weir crest height Pweir Weir crest width Wweir Coefficient of Discharge = 0.63 for this test. Q=Cd*(2/3)*L*HA(3/2)*SQRT(2*g) 25.5 in 23 in 19.5 in 0.65 m 0.59 m 0.50 m Head d/s Head u/s (Hds, in) (Hus, in) Weir head (Hweir, in) lift (delta H, in) Flow (m A3/s) Flow (l/s) Flow (cfs) 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 25.5 26.0 24.5 24.0 23.0 22.5 22.0 21.5 20.3 18.5 17.0 16.0 15.0 14.5 14.0 5.5 5.3 5.0 4.5 4.3 4.3 4.0 3.5 2.8 2.0 1.5 1.0 0.5 0.3 -0.5 1.0 I. 5 2.5 3.0 3.5 4.0 5.3 7.0 8.5 9.5 10.5 II. 0 11.5 0.050 0.046 0.043 0.037 0.034 0.034 0.031 0.025 0.018 0.011 0.007 0.004 0.001 0.000 49.5 46.2 42.9 36.6 33.6 33.6 30.7 25.1 17.5 10.9 7.1 3.8 1.4 0.5 Notes: Pressure at 100 psi delivered, approx. 4 psi at aerator 20" wide in 39" tall flume four-tined 3/4" diam aerator -1 mm holes 1.75 1.63 1.52 1.29 1.19 1.19 1.09 0.89 0.62 0.38 0.25 0.14 0.05 0.02 T h e l im i t ed a i r f low available f rom the screw-based compressor s lowed development un t i l a four horsepower gasoline engine-powered centrifugal b lower was obtained f rom the univers i ty equipment salvage program. The b lower was overhauled, performance was evaluated and fittings were designed to adapt it to the experimental setup. T h e centrifugal 40 b lower was not instal led i n the second laboratory experimental setup since at this t ime a full-scale prototype loca t ion was selected i n the C i t y o f R i c h m o n d and the experimental program shifted focus to that locat ion. The th i rd set o f prototype experimental equipment was developed, assembled and instal led just upstream o f one o f the f lood boxes at the outlet o f a drainage condui t at the Gi lbe r t R o a d storm water outfall i n R i c h m o n d . The experimental setup was instal led over the winter o f 1997 to 1998. It o r ig ina l ly consisted o f three different p u m p designs a imed at invest igat ing k e y features o f the proposed pumps. The o r ig ina l equipment was compr ised o f two, ten-inch diameter pump units and one, twe lve - inch diameter unit. One o f the ten-inch diameter pump units was constructed f rom clear ac ry l i c p ipe , a l l o w i n g v i sua l inspect ion o f the mixture f l ow regime w i t h i n the pump riser pipe. Water was introduced f rom an upstream chamber over a V - n o t c h w e i r and pumped b y the experimental airlift units into a downstream chamber. Ano the r V - n o t c h w e i r at the output o f the downstream chamber enabled the water p u m p i n g flowrate to be measured. Water levels were read f rom staff gauges. Tests were run un t i l the system had stabi l ized, at w h i c h point measurements o f a l l the water levels and a i r f low rates were taken. The water levels were used to find the flowrates b y convent ional V - n o t c h we i r analysis. Figures 7 and 8 show the or ig ina l prototype layout at the Gi lbe r t R o a d locat ion. A d d i t i o n a l large-scale system and site drawings can be found i n A p p e n d i x 1. 41 F I G U R E 7 - Gi lber t Road Prototype A i r l i f t System Layout F I G U R E 8 - Gilbert Road Prototype Airlift System Components Compressed air was suppl ied to diffusers located beneath each pump unit by a 28 horsepower C o m a i r - R o t r o n posi t ive displacement b lower unit and pipe dis t r ibut ion system. T h i s unit was instal led i n a soundproofed shed and equipped w i t h intake and discharge filters and silencers. Since posi t ive displacement b lowers p rov ide a constant rate o f a i r f low, i n d i v i d u a l valves were instal led at each pump unit and a bypass added. In this w a y each pump unit cou ld be tested ind iv idua l ly . Ventur i -s ty le air ve loc i ty meters were also instal led to a l l o w air f l o w to each pump unit to be i n d i v i d u a l l y moni tored . F igure 9 shows the air supply system at the Gi lber t R o a d site. F igure 10 shows the prototype system i n operation. Large-scale site drawings i n A p p e n d i x 2 show the insta l la t ion o f the air supply subsystem at the Gi lbe r t R o a d locat ion. 44 F I G U R E 9 - Gilbert Road Compressed Air Supply Subsystem 2"«> LONG RADIUS 90" ELBOW (TYP.) 3"0 AIR FLOW METER AIR FLOW 3"0 AIR PIPE 3 0 3-WAY WYE 4"x3" REDUCER 4 « LONG RADIUS 90' ELBOW STEEL LADDER 2"x3" REDUCER 2 0 CONTROL VALVE 4 0 AIR SUPPLY DETAIL Y 45 F I G U R E 10 - G i l b e r t R o a d P r o t o t y p e A i r l i f t S y s t e m i n O p e r a t i o n The o r ig ina l system as tested was moderately successful at pumping water and p rov ided tremendous insight into the operation o f the pump units, par t icular ly by the abi l i ty to observe the mixture flow pattern i n the acry l ic ten-inch diameter unit (the centre unit shown i n Figures 9 and 10). The or ig ina l prototype setup identif ied several weaknesses i n the phys i ca l layout and construction o f the layout at the prototype site. Leakage between upstream and downstream chambers was found to be part icular ly problemat ic . A d d i t i o n a l l y , the abnormal ly l o w levels o f water i n the drainage conduit leading to the site over the winter o f 1997/1998 made testing at on ly one leve l o f upstream f l o w possible . A portable pump was introduced to increase the leve l i n the upstream condui t but this had o n l y l imi t ed success. 46 These diff icul t ies led to delay i n this phase o f the program as several revis ions to the site system layout and air dis t r ibut ion layout were developed, drafted and implemented . T h i s w o r k progressed over m u c h o f the Spr ing o f 1998. The results f rom the tests o f these prototype units were more scattered than expected but c lear ly showed the importance o f design details. A l t h o u g h reduced, leakage between chambers cont inued to be problematic despite the system revis ions , and a b reakdown o f the compressed air de l ivery system created further diff icult ies. Some o f the data and calculat ions f rom the th i rd experimental setup are s h o w n i n Table 3. Table 4 shows sample ve loc i ty and loss calculat ions for the data shown i n Table 3. Table 5 shows the leakage test data col lected at this site. Table 6 shows a sample o f later data and also calculat ions for the loss coefficients o f the three o r ig ina l prototype p u m p systems at the Gi lbe r t R o a d site. W h e n analyzed, the results o f this phase o f testing indicated systemic problems w i t h the alternatives be ing investigated. The large diameters pipes had very l o w p a c k i n g densi ty i n the space-constrained storm drain scenario. A l s o no aerator geometry was found to be h i g h l y successful i n sharply m a x i m i z i n g efficiency. Th i s was l i k e l y due to several factors, p r i m a r i l y the pract ical issue o f leakage and the prevalent rec i rcula t ing f l o w patterns i n the pump risers. 47 T A B L E 3 - Gilbert Road Prototype Airlift System Sample Experimental Data Results march 9,1998 Outside wl = 33 Vee notch = 34 inches Pump U/wl d/swl aircfm Flow pipe dia area qmix vmix headless Mix density vwat vair vrel 2.5 8 35.000 40.25 250 0 676762 0 833 0 54498 4.843429 8.887344 35.7942 0.391054 0.485614 12.55535 12.06974 34.500 39.75 250 0 578271 0 833 0 54498 4.744937 8.70662 35.313 0.395151 0.419289 12.64>t 12.22111 34.500 39 250 0 450973 0 833 0 54498 4.617639 8.473037 33.44365 0.411212 0.340279 12.98521 12.64493 39.000 41.5 . 250 0 974318 0 833 0 54498 5.140985 9.433339 41.45398 0.413587 0.739413 13.0378 12.29839 38.750 41.25 200 0 908704 0 833 0 54498 4.242037 7.783834 28.22426 0.500209 0.834052 12.23798 11.40393 35.250 40.75 200 0 786746 0 833 0 54498 4.120079 7.56005 26.6247 0.45895 0.66255 11.30473 10.64218 35.000 40.5 200 0 730287 0 833 0 54498 4.06362 7.456452 25.9 0.462329 0.619532 11.37578 10.75624 35.000 40.5 150 0 730287 0 833 0 54498 3.230287 5.927345 16.3665 0.552526 0.740398 10.25159 9.511189 35.000 40.75 150 0 786746 0 833 0 54498 3.286746 6.030943 16.94361 0.543329 0.784363 10.04514 9.26078 37.750 41.62 150 1 006934 0 833 0 54498 3.506934 6.434972 19.28984 0.562277 1.038892 10.47997 9.441076 38.000 41.5 100 0 974318 0 833 0 54498 2.640985 4.846018 10.9397 0.675072 1.206897 9.411976 6.205079 34.500 40 100 0 626111 0 833 0 54498 2.292777 4.207082 8.245128 0.658465 0.756489 8.954315 8.197826 40.250 41 250 0 846199 0 833 0 54498 5.012866 9.198249 39.41356 0.406102 0.63056 12.87348 12.24292 33.500 38 250 0 316525 0 833 0 54498 4.483191 8.226335 31.5245 0.356521 0.207068 11.88156 11.6745 35.000 38 200 0 316525 0 833 0 54498 3.649858 6.697228 20.89419 0.443392 0.257522 10.98875 10.73123 39.750 41 200 0 846199 0 833 0 54498 4.179532 7.669142 27.39864 0.471078 0.731449 11.56395 10.8325 37.000 40 150 0 626111 0 833 0 54498 3.126111 5.736189 15.32789 0.524148 0.602177 9.640225 9.038048 35.500 39 150 0 450973 0 833 0 54498 2.950973 5.414823 13.65853 0.5125 0.424095 9.40989 8.985795 T A B L E 4 - Gilbert Road Prototype Airlift System Sample Experimental Results Results March 9,1998 1 2 3 4 5 6 7 8 9 10 11 12 13 Pump U s / w l d/s wl Air Water Vmix Dens Vw Head loss t W m A 2 Vm A 2*Den U V w A 2 -exit/vwA2 ins. ins. cfs cfs ft/sec ft/sec feet 2 35.00 57.75 4.17 0.68 8.96 0.23 5.53 1.44 1.16 5.12 3.03 2.44 34.50 57.75 4.17 0.58 8.78 0.21 5.05 1.46 1.22 5.78 3.69 3.05 34.50 57.75 4.17 0.45 8.54 0.19 4.34 1.55 1.36 7.10 5.29 4.55 39.00 57.75 4.17 0.97 9.51 0.27 6.76 1.60 1.14 4.28 2.26 1.73 1.00 2 38.75 57.75 3.33 0.91 7.83 0.30 5.54 1.42 1.49 4.92 2.98 2.38 35.25 57.75 3.33 0.79 7.61 0.29 5.10 1.21 1.34 4.71 2.99 2.36 35.00 57.75 3.33 0.73 7.51 0.28 i nn. 4.89 1.23 1.40 5.07 3.31 2.65 2 35.00 57.75 2.5 0.73 5.97 1 .UU 0.34 4.00 0.96 1.74 5.16 3.87 3.12 35.00 57.75 2.5 0.79 6.08 0.35 4.19 0.92 1.61 4.63 3.37 2.65 37.75 57.75 2.5 1.01 6.48 0.38 4.87 1.00 1.53 3.99 2.71 2.03 1.00 2 38.00 57.75 1.67 0.97 4.89 0.48 3.79 0.62 1.66 3.49 2.76 1.97 34.50 57.75 1.67 0.63 4.24 0.41 2.81 0.60 2.15 5.22 4.89 3.95 1.00 1 40.25 54.5 4.17 0.85 9.27 0.25 6.27 1.85 1.38 5.54 3.03 2.48 33.50 54.5 4.17 0.32 8.29 0.17 3.43 1.60 1.50 8.80 8.77 7.77 35.00 54.5 3.33 0.32 6.74 0.20 2.86 1.59 2.25 11.00 12.54 11.40 39.75 54.5 3.33 0.85 7.72 0.29 5.32 1.62 1.76 5.97 3.70 3.08 37.00 54.5 2.5 0.63 5.78 0.32 3.64 1.30 2.50 7.86 6.32 5.51 35.50 54.5 2.5 0.45 5.45 0.28 2.93 1.31 2.84 9.98 9.80 8.82 Outside wl = 33 Vee notch = 34 Average #2 1.48 4.96 3.43 2.74 Av. #1 2.04 8.19 7.36 6.51 Sdev #2 0.28 0.91 0.89 0.82 Sdev #1 0.58 2.17 3.69 3.46 CV#2 0.19 0.18 0.26 0.30 CV#1 0.29 0.26 0.50 0.53 48 T A B L E 5 - Gi lber t Road Prototype Ai r l i f t System Leakage Tests Pump Data from May 1 tests No 3 pump-12" Air U/sW,L D/SW.L Weir Weir flow Leaks Total 250 46.5 50.5 41.75 1.127673 0.33775 1.465423 250 44.5 50 41.75 0.974561 0.334608 1.309169 250 37.5 48 41.75 0.489537 0.321734 0.811271 250 38 48.75 41.25 0.769405 0.326621 1.096026 250 46 50 41.25 1.127673 0.334608 1.462281 250 42.5 49.5 41,25 0.974561 0.331436 1.305998 200 45 50 41.75 0.974561 0.334608 1.309169 200 38.25 48 41.75 0.489537 0.321734 0.811271 150 47 50 41.75 0.974561 0.334608 1.309169 150 44.5 50 41.75 0.974561 0.334608 1.309169 150 38 48 41.75 0.489537 0.321734 0.811271 100 45.5 49 41.75 0.707362 0.328234 1.035596 No 2 pump 250 48 49 41.75 0.707362 0.328234 1.035596 250 45 48 41.75 0.489537 0.321734 0.811271 250 37.5 47 41.75 0.317686 0.315099 0.632785 220 39 48 41.25 0.592484 0.321734 0.914218 200 45.5 48.5 41.75 0.592484 0.325 0.917484 150 48.5 49 41.75 0.707362 0.328234 1.035596 150 45.5 48.5 41.75 0.592484 0.325 0.917484 150 39 47 41.75 0.317686 0.315099 0.632785 100 45.5 48 41.75 0.489537 0.321734 0.811271 No 1 pump 250 48 49 41.75 0.707362 0.328234 1.035596 250 45 48 41.75 0.489537 0.321734 0.811271 250 38.5 47 41.75 0.317686 0.315099 0.632785 200 46 48.5 41.75 0.592484 0.325 0.917484 200 38 47 41.75 0.317686 0.315099 0.632785 150 49 49 41.75 0.707362 0.328234 1.035596 150 46 49 41.75 0.707362 0.328234 1.035596 150 39.5 47 41.75 0.317686 0.315099 0.632785 100 46.5 48.5 41.75 0.592484 0.325 0.917484 49 T A B L E 6 - Gilbert Road Prototype Airlift System Sample Experimental Results 2 Tests May 1, 1998 1 2 3 4 5 6 7 8 9 10 11 12 13 Pump U/Swl d/swl Air Water Pipe area V mix Dens Vw Head loss UVm*2 Ud"vmA2 UVw»2 ins. ins. cfs cfs sq. ft. ft/sec ft/sec feet 3 46.50 53.50 4.17 1,80 0.79 7.60 0.38 5.99 1.86 2.07 5.42 3.34 3 44.50 53.50 4.17 1,31 0.79 6.97 0.33 4.98 1.88 2.50 7.46 4.89 3 37.50 53.50 4.17 0.81 0.79 6.34 0.28 3.73 1.53 2.45 8.84 7.08 3 38.00 53.50 4.17 1.10 0.79 6.70 0.31 4.49 1.44 2.06 6.62 4.59 3 46.00 53.50 4.17 1.46 0.79 7.17 0,35 5.31 1.95 2.44 6.96 4.44 3 42.50 53.50 4.17 1.31 0.79 6.97 0.33 4.98 1.72 2.28 6.82 4.47 3 45.00 53.50 3.33 1.70 0.79 6.41 0.43 5.07 1.56 2.44 5.72 3.91 3 38.25 53.50 3.33 1.20 0.79 5.77 0.37 . 4.09 1.21 2.34 6.27 4.65 3 47.00 53.50 2.50 1.31 0.79 4.85 0.46 3.66 1.61 4.41 9.68 7.77 3 44.50 53.50 2.50 1.31 . 0.79 4.85. 0.46 3.66 1.40 3.84 8.43 6.76 3 38.00 53.50 2.50 0.81 0.79 4.22 0.39- .2.65 1.12 4.07 10.45 10.30 3 45.50 53.50 1.67 1.04 0.79 3.44 0.52 2.53 1.22 6.66 12.76 12.37 2 48.00 57.75 4.17 1.04 0.55 9.55 0.28 6:91- 2.31 1.64 5.95 3.12 2 45.00 57.75 4.17 0.81 0.55 9.13 0.2S 6.06 2.19 1.69 6.89 3.84 2 37.50 57.75 4.17 0.63 0.55 8.81 0.22 S.27, 1.67 1.39 6.31 3,89 2 39.00 57.75 3.67 0.91 0.55 8.41 0.28 5.89 1.52 1.39 4.87 2.82 2 45.50 57.75 3.33 0.92 0.55 7.80 0.30 5.52 1.98 2.09 6.86 4.18 2 48.50 57.75 2.50 1.04 0.55 6.49 0.39 4.91 1.87 2.86 7.39 5.01 2 45.50 57.75 2.50 0.92 0.55 6.27 0.37 4.56 1.70 2.78 7.54 5.26 2 39.00 57.75 2.50 0.83 0.55 5.75 0,32 3.63 1.37 2.67 8.33 6.71 2 45.50 57.75 1.67 0.81 0.55 4.55 0.45 3.32 1.36 4.23 9.42 7.94 ! 48.00 54.50 4,17' 1.04 0.55 9.55 0.28 6.91 2.39 1.69 6.14 3.22 1 45.00 54.50 4.17 0.81 0.55 9.13 0.25 6.06 2.26 1.74 7.10 3.96 1 38.50 54.50 4.17 0.63 0.55 8.81 0.22 5.27 1.82 1.51 6.85 4.22 1 46.00 54.50 3.33 0.92 0.55 7.80 0.30 5.52 2.10 2.22 7.29 4.44 1 38.00 , 54.50 3.33 0.63 0.55 7.28 0.26 4.45 1.61 1.96 7.51 5.25 1 49.00 54.50 2.50 1.04 0.55 6.49 0.39 4.91 2.02 3.09 7.97 5.40 1 46.00 54.50 2.50 1.04 0.55 1 6.49 0.39 4.91 1.77 2.71 6.98 4.73 1 39.50 54.50 2.50 0.63 0.55 5.75 0.32 3.63 1.50 2.92 9.11 7.34 1 46.50 54.50 1,67 0.92 0.55 4.74 0.47 3.60 1,49 4.26 9.11 7.38 Average #3 3.13 7.95 6,21 Stdsv #3 1.38 2.17 2.77 CV#3 0.44 0.27 0.45 : Average #2 2.30 7.06 4.75 Std#2 0.93 1.33 1.68 CV#2 0.40 0:19 0.35 Average #1 2.45 7.56 5.10 Std #1 0.88 1.01 1.44 CV#1 0.36 0.13 0.28 A second prototype concept was then designed and built at the Gilbert Road location. This system was conceived in an attempt to maximize the use of the available plan area in the constrained space of a drainage conduit. The concept was based on the original lab model that used the full width of a small flume. A full-width built-in pump assembly was designed. It formed a continuous side-to-side element in the base of the drainage conduit. 5 0 Such a unit w o u l d have superior aeration density and make maximal ly-ef f ic ien t use o f the l im i t ed p lan area i n the base o f the conduit . The result was a lateral slot-based pump and aerator. T h i s unit was bui l t i n the form o f a slot four feet long by one foot w i d e for the air-water mixture . A hor izonta l ly oriented cy l ind r i ca l aerator was designed and instal led at the base o f the unit. I f this proved successful the intention was to adopt the design concept and b u i l d airlift pump units that cou ld span across the entire w i d t h o f rectangular box culvert . Large-scale drawings o f the slot-configured airlift pump system can be found i n A p p e n d i x 2. Figure 11 shows the slot-configured airlift pump i n operation. Figure 11 - Slot-Configured A i r l i f t Pump in Operation 51 This system was tested under adverse conditions with very little upstream depth available. Nevertheless, it identified an unanticipated and major problem with the "slot" design. Water tended to "slosh" from side to side in the pump unit body, but very little was effectively lifted. It was observed that when the water was high at one end of the slot it gave a large enough back pressure to the orifices there to create an increased flow of air from the orifices at the other end. In this way the air tended to escape from the system. When the water sloshed back to the other end of the system the air escaped from the first end. Thus, although a considerable amount of spray was created very little water was pumped. The immediate failure of this design showed convincingly that large capacity airlift pump systems must be designed to prevent this sloshing behaviour, effectively a one-dimensional recirculation effect analogous to that which had been observed in the cylindrical units. This test confirmed the findings of the first test, namely that circulation could easily develop within the pump riser pipes, greatly reducing pump efficiency. The third experimental phase was successful in pointing the way forward. The design concept was revised to minimize the two most severe problems encountered at the Gilbert Road site, namely circulation within the riser pipes and leakage between the test chambers. The large ten and twelve-inch diameter pump barrels were replaced with eight, six, four, and three-inch units. This was considered a useful means of reducing circulation and turned out to be very effective. It was also decided to test the effect of inclining the pump tubes. Wallis (1969) asserts that inclination up to approximately 40 degrees from the vertical does not adversely affect bubble velocity, and inclining the pump riser tubes in this way promised reduced construction costs by requiring fewer pipe fittings. 52 This fourth phase of the experimental program was carried out over the course of several months at the Richmond Public Works Yard. Plastic pipes were set up in a tank with a metered water supply and metered compressed air supplied from a single jet at the bottom of each pump's riser tube. Since aerator configuration had little discernible effect in previous phases of the project, aerator designs were not tested in this phase. The pipes were inclined from zero to thirty degrees from the vertical. The flow of air and water were set and when the water level stabilized in the tank, the stable depth of water in the tank was read. Figure 12 shows the experimental setup for this phase of the project. Various combinations of air and water flow and pipe diameters were tested in an "evolutionary" manner - starting with one pipe at one upstream water level and varying downstream water levels for a given air flow rate. The tests showed that for a given flow of air the inclination away from the vertical of the pump riser pipes within the range o f zero to thirty degrees did not seem to affect the water flow rates. The overall results still evidenced some scatter but this was at least partly due to unavoidable variations in the position of the air jets in the bottom of each pipe riser tube and other minor factors such as the resolution of the meters used, etc. Data from this experimental phase was far more consistent than that from the third experimental setup since upstream and downstream water levels could be precisely controlled and there was no leakage from the sealed tanks. Table 7 shows sample experimental data from this series of experimental tests. 53 FIGURE 12 - Richmond Public Works Experimental Setup f m +-> C O (D H O H > H H CO •a o 3 4H < CO P H O P H T 3 O I P H o a IT) CN CN CN X CN m GO S-H *-> "5b PH 1 *S ' 5 54 T A B L E 7 - R i c h m o n d P u b l i c W o r k s S a m p l e E x p e r i m e n t a l D a t a o 1 3 .a s I I 5? S i PO ^ ?5 ^ s s s s d o o o « n n w w P J d o d d d d CD CD CD O O) S N N N N <r *r v d 6 o' o* d o o o o o o CD <D (O <£> V V f ^ V d d d d d w u> w w i n trj i n i n vn i n jp jn ip jp in i° SP IP IP N N N K K f*. K p» l*« N. K J*. |s. fx. f— N- r— N- h- flCtcooq«p co co <o co" co co iri i r i i n i r i tn s s N s N N N N N W t r t IA n tf) io i n « h~ f» I O in i d » in ifl in in in ui tn in m w to w in « « j w w 13 <3 13 pi S$ <8 i8 !3 5$ $5 S V "t V ^ <o u> m to <o <x> m i n m •« N w n r UJ u> >q « J: e 2 to! CQ m o cn p s m M i n ;*» ro rs.̂ r*. -̂ £> «o P - r* 3 8 8 8 3 8 8 8 8 8 3 8 8 8 S 8 3 8 8 8 S 8 8 8 8 8 ? S ? 8 8 1 l« d o* © d o o d d d o d d d d d d d d d d ddd d d o d o d o i i 2 Q J pj n n n v « n ^ < v 10 n o o o <o co to 03 to co 1 A t this point the head loss relationships described i n equation s (35) and (36) i n chapter 3 were created and used to characterize the system behaviour. A f ina l set o f tests were made w i t h a bundle o f nine, four- inch diameter plast ic pipes at an inc l ina t ion o f 0 to 40 degrees from the ver t ica l . Compressed air was suppl ied through a man i fo ld o f pipes w i t h a one i n c h jet at the centre o f each pump riser pipe. A l t h o u g h some pract ical problems remained, the performance o f this exper imental setup conf i rmed that the relationships developed i n the mode l for turbulent m i x i n g fo rmed a reasonable basis for air l if t pump design i n the churn turbulent regime. T h i s successful phase o f the p rogram resulted i n a reliable data set, p rov id ing the basis for ca l ibra t ion o f the theoretical mode l and pointed the w a y towards a v iable airl ift pump design for the si tuation i n R i c h m o n d . 3.3 - Results of the Exper imental P rogram The lessons f rom the first two laboratory-based phases o f the experimental p rogram indicated the v i a b i l i t y o f the concept o f low- l i f t h igh - f low low-submergence airl ift pumps for urban storm drainage. The Gi lber t R o a d prototype system suggested several pract ical considerations for full-scale applications and led the w a y to the f inal experimental phase. The f ina l phase produced the reliable data set used to calibrate the head loss relat ionships developed i n the th i rd theoretical mode l . Th i s experimental program also l ed to a v iab le pract ical design. The results were also used to ver i fy the theoretical m o d e l developed to exp l a in low- l i f t , low-submergence, h igh- f low airlift pump behaviour i n these scenarios, w h i c h i n turn led to a v iab le pract ical engineering design procedure. 56 C H A P T E R 4 4.1 - Airlift Pump Model for Fixed Bubble Slip Velocities Refer again to Figure 1, reproduced here for convenience: ' j jQwater In static conditions, the pressures inside and outside of the airlift pump tube are equal at the point of aeration. H sub = H t e a l ' D e n S (4) Under dynamic conditions air bubbles are rising through the water column within the airlift pump tube and a driving head must be added to the system as described in (4) to maintain the pumping action: 57 Hsub = H total • + Hdrive Hdrive =Hxllh -Hlolal -Dens which can be rearranged to form (5) For equilibrium, the driving head must be equal to the losses in the system. Hdrive ~ Hhss (6) Fluid flow losses are commonly expressed in the form of: V2 headloss = K (7) 2g so for the case of entrance, pipe and exit losses in the airlift pump system, and assuming that the entrance, pipe and exit losses due to viscosity and fluid friction due to air w i l l be much less than those due to water: v 2 V 2 V 2 TT T/' water . js water , jy- water / o \ ^ loss ~ ^entrance ' + ^ pipe 2g ' 2g or for the case in which loss factors for various entrance, pipe and exit geometries are not explicitly considered separately, a combined loss factor can be used: Kate' Hloss = K,o,al - (9) 2g 58 C o m b i n i n g (3) and (2) and rearranging, get Hsuh-HlolarDens = K, vi water total 2g (10) The goal is to determine the combined loss factors Ktotai for representative geometries so that air l if t p u m p performance can be modeled s imp ly by (7). W e want to determine the loss factor, so rearrange: Equa t ion (11) provides the total loss factor for the airlift pump, dependent o n the submergence and total pump length, as w e l l as the density o f the air-water mix ture and the ve loc i ty o f the water phase. T o solve (11) for the total loss factor we need the density o f the air-water mixture i n the airlift p u m p tube and the ve loc i ty o f the water phase i n the airlift pump tube. Cons ide r ing a representative cross-section o f the air-water mixture f l o w i n g i n the airl ift p u m p tube, the relative density o f the air-water mixture i n the airlift pump tube is g iven by: K«„a, = 28 • Hsuh ~ H „ vi water tolal •Dens (11) Dens = water (12) Area 59 T o solve (11), the area o f the pump cross section occupied by the water phase is also needed. T o obtain the area occupied by the water phase, the ve loc i ty o f the water phase and the ve loc i ty and area occupied b y the air phase are required. T o get the ve loc i ty o f the water fraction o f the mixture , consider cont inui ty o f the v o l u m e flow rates o f the mixture and o f each phase i n the airl ift pump tube: Qmix ^rmix ^mix (13) Qair ^air ^air 0 4) Q water ^water ^water 0^) also the v o l u m e flow rate o f the mixture is composed o f the sum o f the v o l u m e f l o w rates o f the air and water phases: Qmix ^mix ^mix water ^water 0^) and the total cross-sectional area i n the airl ift pump tube is s i m p l y composed o f the sum o f the areas occup ied by the air and water phases: A —A + A mix water air V / N i c k l i n (1962) suggests that i n the s lug flow regime for s t i l l water and where T a y l o r bubble diameter and pipe diameter are very s imi lar , 60 (2) If Taylor bubbles were to be found in the 3 to 12-inch diameter airlift pump riser tubes in this study, Nicklin (1962)'s equation (2) would suggest their rise velocities to be within the 1.1 to 2 foot per second range. Classical observations of bubble rise speeds outside the slug flow regime (i.e.: smaller bubbles not constrained directly by pipe boundaries) suggest the terminal velocity of a single bubble is relatively constant between 25 to 45 cm/s over a broad range of bubble diameters, as shown in Figure 12, here reproduced from Wallis (1969): FIGURE 13 - Bubble Rise Velocities in Still Water, from Wallis (1969) / I 1 1 I I I I I I I I ••!-•• I •'• I L l l 'I l l I I I I I I I I I I I I I I O.OI 0.02 0.04 0.04,0.01,0.1 O.Z O.H OA 0.&/.O 2.0 4.0 Equivalent Radius in centimeters Ry 61 Tai t e l & al.(1980) suggest that above a cr i t ica l diameter (approximately 1.5 m m ) air bubbles tend to deform and adopt an erratic path. Thus the s l ight ly s lower effective bubble rise speeds o f the unconstrained non-Tay lo r bubbles m a y be expla ined by the irregular i ty o f the smal ler bubbles ' rise trajectories compared w i t h the constrained- ver t ica l rise trajectories o f the T a y l o r bubbles i n the s lug f l o w regime. U s i n g the concept o f a constant terminal rise ve loc i ty for bubbles i n s t i l l water, we introduce the relative ve loc i ty o f the air phase to the water phase i n the airl ift p u m p tube, water rel (18) substituting (18) into (14) results i n : Qair = (K water (19) substituting (17) into (19) results i n : Qair = 'water + Kel XAre° " 4 water ) (20) rearranging (15) and substituting into (20): (21) 62 E q u a t i o n (21) provides a functional relationship between the measured f l o w rates o f the air and water phases Qair and Qwaler, the k n o w n cross-sectional area o f the air l if t p u m p tube Area, the k n o w n relative ve loc i ty o f the air phase to the water phase Vrei, and the u n k n o w n ve loc i ty Vwaler o f the water phase. Therefore, under these assumptions the ve loc i ty o f the water phase i n the airlift pump tube can be calculated for any combina t ion o f the measured values. T h i s water phase ve loc i ty can then be used to solve (11) for the desired overa l l loss factor Ktotai. H o w e v e r , to solve (11) w e also need the density o f the air-water mixture i n the air l i f t p u m p tube. Rearranging (15) and substituting into (12): Dens = — (22) Vwater ' A r e a substituting (22) into (11) and rearranging, get: v water ( H 0 ^ _ * x total XL. water V ™* vwater'Ar™; (23) Equa t ion (23) gives the pump loss factor as a funct ion o f the water phase ve loc i ty , diameter, total length and submergence o f the pump tube, vo lume flow rate o f water and ve loc i ty o f the water phase i n the airlift pump tube. The ve loc i ty o f the water phase can be determined f rom (21) and thus the pump loss factor determined for a variety o f flow and submergence condi t ions. 63 In this w a y it was hoped the characteristic behaviour o f the airl ift pump system c o u l d be determined f rom the pump loss coefficient, enabl ing a clear understanding o f the p u m p system operat ion and as w e l l creating a s imple design procedure. W h e n the experimental data was compared w i t h this mode l it became apparent that losses found i n the pump units i n this study were not accurately predicted. The summary values i n c o l u m n 13 o f Table 6 correlate the measured system losses w i t h the water phase ve loc i ty . T h e w i d e spread i n the der ived loss factors as w e l l as the large coefficients o f var ia t ion indicate c lear ly that this mode l is not applicable to the pumps i n this study. Further observations o f the experimental units i n operation and more research suggested that the assumption o f a terminal bubble ve loc i ty as expla ined by F igure 13 was not appl icable i n this case. The assumption o f a terminal bubble speed relative to s t i l l water makes this m o d e l poss ib ly more suitable for l o w v o i d -ratio f lows and l o w mixture f l o w veloci t ies , such as might be encountered i n a long , large diameter riser such as used i n lake aeration or destratification or harbour de- ic ing . A l t h o u g h this mode l for airlift p u m p performance is not he lpfu l i n the case o f low- l i f t , low-head h i g h f l o w pumps such as are considered here it does have promise and m a y prove useful i n analysis o f cases such as those ment ioned above. 64 4.2 - Airlift Pump Model for Variable Bubble Slip Velocities W a l l i s (1968) 's F igure (9.5) i n the section concerning churning f l o w presents the mix ture and gas phase mass f lux rates i n terms o f the mixture mass f lux rate and a gas phase "drif t flux" rate relative to the mixture flux. W a l l i s ' figure is reproduced as F igure 13 here. F I G U R E 14 - Mixture and Gas Flux Rates, Wallis (1969 OVERALL FLUX Jt m/sec 65 It indicates the relat ionship between the mass f l o w rate o f the gas and the mass f l o w rate o f the mix ture , expressed as a mass f lux rate per unit area for the mixture and a relative, or "dr i f t" mass f lux rate per unit area for the gas phase. Inspection o f F igure 13 suggests that the gas phase average ve loc i ty is dependent o n the flux rate o f the mixture . Cons ide r a case such as ours i n w h i c h the densities o f the gas and l i q u i d phases are k n o w n f ixed quantities, the pump geometry is k n o w n and the density o f the gas phase is negl ig ib le compared to that o f both the l i q u i d phase and that o f the mixture . In such a case F igure 9 suggests that the drift f lux rate and mixture flux rate are dependent o n the m i x ratio and component phase veloci t ies only . So for any g iven m i x ratio the straight- l ine relat ionship i n the ratio o f the gas drift flux and mixture f lux rate should be equa l ly representative o f the gas and mixture fraction veloci t ies . In that case for k n o w n fixed densities o f l i q u i d (water) and gas (air) phases, and for a k n o w n v o i d fraction, the ve loc i ty o f the gas phase o f the mixture i n a churn-turbulent two-phase flow depends not o n the ve loc i ty o f the water phase as suggested by Figure 8 and as found i n l o w v o i d - fraction s t i l l water and bubb ly f low, but rather depends o n the ve loc i ty o f the mix ture instead. W a l l i s ' equation 9.36 suggests a different fo rm o f this relat ionship. That is expected since his flow analysis was momentum-based and d i d not require the relative veloci t ies o f the component mixture phases. 66 Nevertheless , the conc lus ion is powerful - namely that i n cases where the gas densi ty is n e g l i g i b l y l o w i n compar i son w i t h the mixture density the gas phase ve loc i ty is greater than, and rises l inear ly w i t h the mixture ve loc i ty . T h i s provides a valuable component m i s s i n g so far i n the analysis o f these short-lift systems. It is reassuring to note that D e Cacha rd & De lhaye (1995) also found a s imi la r result for mixture and gas phase veloci t ies up to approximately 6 m/s i n smal l diameter, long lift p u m p risers. So , expressing the air phase ve loc i ty as a l inear function o f the mixture ve loc i ty : Equa t ion (25) models F igure 13 to remarkably good agreement i n units o f feet per second. T h i s equation fit corresponds w i t h the bubble terminal ve loc i ty o f approximate ly 30 cm/s , w h i c h is approximately equal to 1 foot per second as i n F igure 12 and suggested by W a l l i s for s t i l l water and used i n the first mode l . The \ 2Vmix term is also fami l ia r since it represents the ratio o f the centreline ve loc i ty to the average ve loc i ty i n the fu l ly developed turbulent flow field w i t h i n a c losed pipe. (24) and fi t t ing the l inear relat ionship i n (25) to W a l l i s ' data i n Figure 13 suggests: (25) 67 The fo rm and values o f (25) as interpreted here f rom W a l l i s ' data are ve ry s imi l a r to those suggested b y N i c k l i n (1962) before his w o r k o n airl ift pumps. N i c k l i n suggested for the ve loc i ty o f a s lug bubble r i s ing i n a two-phase mixture at R e y n o l d ' s numbers under In consistent units for a representative pump riser tube o f s ix - inch internal diameter, equation (26) becomes N i c k l i n (1962) qualif ies (28) above as be ing accurate for R e y n o l d ' s numbers b e l o w 8000 and approximate for R e y n o l d ' s numbers over 8000. Furthermore, Fernandes, Semiat & D u c k l e r (1983) independently suggest that the T a y l o r bubble rise ve loc i ty i n larger diameter pipes than those studied by Ta i te l et a l . (1980) is g iven b y Equat ions (27) and (28) are very s imi la r to one another and suggest values for the air phase ve loc i ty for s lug f l o w just s l ight ly greater than suggested by W a l l i s ' experiments 8000: Vair=\.2Vmix+0.35jg-Diam (26) (27) Vair= \.2Wmix +0.35 Jg-Diam (28) 68 for bubb ly f l o w as g iven i n (25). These findings inspire confidence i n this study that the , f o r m and values i n equation (25) are rel iable. Therefore, substituting (25) into (14): a,>=(i.o+i.2F;,fcR (29) Subst i tut ing (13) into (25): 1.0 + 1 . 2 ^ ^mix J (30) Rearranging (30) to solve for the cross sectional area occupied by the air phase, A - Qair 1.0 + 1 . 2 ^ V ^mix J (31) substi tuting : (13) into (28): ( V,, =1.0 + 1.2 'mix Area. (32) 69 also substituting (32) into (14): f 1.0 + 1.2 0,, Area (33) and substituting (33) into (15) and us ing (16): V , =Q water w Area - - 1.0 + 1.2 Q water Qair Area (34) J) Equa t ion (34) w i l l then give the ve loc i ty o f the water phase i n the airl ift p u m p tube as a funct ion o f the measured f l o w rates o f air and water, and the cross sect ional area o f the airl if t p u m p tube for churn-turbulent f lows , assuming the air and water phase f l o w veloci t ies are accurately represented by equation (22) w h i c h was der ived f rom a f ixed - densities and m i x ratios analysis o f W a l l i s (1968) data i n F igure 14 and bolstered b y N i c k l i n (1962). H a v i n g determined the ve loc i ty o f the water phase i n the airlift pump tube f rom equation (34), and k n o w i n g the water phase vo lume f l o w rate and pump geometry, the values can be used to f ind the value for the pump loss coefficient as determined b y equat ion (23): K 2g total Hsuh ~ H water \ total water Vwater ' A r e a J (23) 70 T h i s approach holds more promise than the first for improved and more rel iable results. The experimental data was reanalyzed. Howeve r , even w i t h this more rel iable approach for ca lcu la t ing the ve loc i ty o f the water phase and despite good evidence to support equat ion (29), the head losses were s t i l l found not to be proport ional to the square o f the water phase ve loc i ty . Despi te the fact that this mode l cannot be used to exp la in the behaviour o f the l o w - submergence, low- l i f t , h igh- f low pump units i n this study it does h o l d promise for use i n m i d - v e l o c i t y bubb ly f l o w pump units. In such units head losses are p r i m a r i l y due to p ipe f r ic t ion as suggested by W a r d (1924) and this mode l m a y help provide a s imple analysis too l for that class o f air l if t pump systems. 71 4.3 - Airlift Pump Model for Turbulent Mixing G i v e n the inab i l i ty o f the second mode l to accurately predict the head losses us ing the i m p r o v e d method for calcula t ing the water phase ve loc i ty , a new approach is c lear ly necessary. E v i d e n t l y the assumption that the losses are p r imar i l y due to pipe f r ic t ion, entrance and exi t losses and are dependent on the ve loc i ty o f the water phase must be reexamined. The f l o w o f the mixture i n the airlift pump tube is very turbulent w i t h s ignif icant v i sua l evidence o f churning.and recirculat ion, so the assumption that the influence o f the air phase is negl ig ib le m a y be suspect. W a l l i s (1968) suggests i n passing that the majori ty o f energy dissipated i n the pipe f l o w o f churning two-phase mixtures results f rom internal losses rather than pipe-friction-related causes. C l a r k & Dabol t (1986) also argue that the fr ic t ional head losses are a second-order effect w i t h i n pract ical lengths for non-s lug f lows al though they do not quantify what the fr ict ional head losses are. Further research and passing suggestions i n several other references p rov ide some clues to the mechan i sm o f these losses. W a l l i s (1968) mentions that i n churning f l o w the chaotic movement o f water i n the f l o w mixture causes the most energy loss, and furthermore, that i n the majori ty o f pract ical cases bubb ly f l o w never becomes fu l ly developed and entrance effects dominate the region before s lug f l o w develops. W a r d (1924) mentions that short pumps have losses not important i n l ong pumps. M o r r i s o n & a l . (1987) suggest that churning f l o w is i n fact a transit ion regime usua l ly ex is t ing f rom 15 to 35 pipe diameters away f rom the aeration point, before significant enough bubble 72 accret ion can occur to create s lug f low. D e C a c h a r d & De lhaye (1995) suggest that the length effects f rom the developmental region o f churn flow leading to s tabi l ized s lug flow m a y create higher than predicted head losses up to lengths several hundred t imes the pipe diameter away f rom the entrance. Thus a ful ly s tabi l ized s lug flow regime m a y not develop w i t h i n a length up to even two hundred t imes the pipe diameter. Ta i t e l & a l (1980) quantif ied a m i n i m u m length for the turbulent entrance transit ion zone as Lauras = 4 0 - 6 - D i a m f v. ^ + 0.22 yjg • Diam (35) It occurs that the short p u m p losses mentioned by W a r d (1924) must be due to the entrance and transi t ion zone turbulence. The pumps i n this study are conc lu s ive ly "short" - substantial ly shorter than 15 to 35 to several hundred diameters long , and substantial ly shorter than the entrance lengths suggested by Tai ta l & a l . (1980) by equat ion (35) above. Therefore it is reasonable to assume that the mixture i n the entire pump riser tubes is exc lu s ive ly exper iencing the turbulent transit ion zone f l o w regime. In that case, the losses i n short airl ift pumps such as those i n this study must be p r imar i l y turbulent i n nature - and not pipe f r ic t ion losses dependent o n the water phase ve loc i ty as was assumed i n the first two models , and as c o m m o n l y assumed i n the p r imar i ly s lug- f low models developed, to date. In this case the challenge then becomes h o w to quantify the mixture turbulence and relate the p u m p head losses to that turbulence. 73 It was observed i n experimental trials that the air-water mixture became increas ingly turbulent w i t h increasing mixture veloci t ies , and that very h igh gas phase veloci t ies at h i g h v o i d fractions resulted i n large losses and very li t t le l i q u i d f l ow. These observations suggest that mixture ve loc i ty i n the churning regime is a good indicator o f mixture turbulence and hence o f losses i n these short pump units. Further research d iscovered Ishi i and Zuber ' s (1979) c l a i m that i n turbulent f l o w regimes the bubbles influence the surrounding f lu id and also other bubbles, and that thus bubbles can be entrained i n each others' wakes, and therefore the losses i n such f lows shou ld be considered relative to the mixture ve loc i ty rather than that o f the l i q u i d phase. W a l l i s (1968) also suggests a s imi la r general form. W e therefore propose a functional fo rm for the turbulent head losses i n these short air l if t pumps , dependent o n the mixture density and turbulence, and represented by the densi ty and v e l o c i t y o f the entire mixture rather than the ve loc i ty o f the water phase alone: Hlms=d-Dens-V;iix (36) The tuning parameters d and e w i l l be exper imental ly determined f rom the research program data. The f o r m o f equation (36) w i l l effectively parameterize the head losses but does not promise a great advancement i n terms o f the details o f the head loss mechan ism. T h i s is 74 not entirely surpr is ing since D e Cachard & Calhaye (1995) exp la in that a m o d e l for w a l l f r ic t ion i n churn ing flow is not yet available, and that the chaotic m o t i o n i n churning flow makes empi r i ca l considerations for w a l l f r ic t ion losses a necessity. G o v a n et a l (1991) agree, suggesting that creating a realistic mode l for churn flow mechanics is "par t icular ly cha l leng ing" . D e Cacha rd & Calhaye (1995) recommend a formulat ion s imi la r to equation (36) for l o n g slender pumps, based o n the l i q u i d phase ve loc i ty . F o l l o w i n g their suggestion and us ing their equation [48] and B l a s i u s ' fo rmula for fr ict ional losses i n the boundary layer as g iven i n their equation [11] their solut ion proposes a f r ic t ion loss term for churn flow as: - 0 . 3 1 6 - Dens l i q u i d _„.2S 2 Hhss  =  Hlolal IDiam liquid ^liquid (37) for the R e y n o l d ' s N u m b e r R e based on the ve loc i ty o f the l i q u i d phase Vnquid. The fo rm o f equation (37) is reassuringly s imi la r to the form o f equation (36), arr ived at independently. The p r imary difference between D e Cachard & D e l h a y e ' s fo rm and that suggested i n this study is that their expression is calibrated for sma l l diameter ta l l risers and uses the l i q u i d phase ve loc i ty as was suggested i n the second m o d e l above, whereas equation (36) relies o n the mixture ve loc i ty as an indicator o f turbulence. T o use (36), w e substitute (6) into (5): Hloss=Hsuh-HlolarDens (38) 75 and substitute (36) into (38): d • Dens • Vemix = Hsuh - Hlolal • Dens (39) n o w substituting (13) into (38): d (^water ^ l y e V Area j ' mix n sub n total (A \ \Area j (40) T h i s th i rd m o d e l for airl ift pumps continues to make use o f the relative ve loc i ty o f the air phase to the ve loc i ty o f the mixture as g iven b y the exper imental ly determined equat ion (26) and suggested f rom W a l l i s (1968) 's data and reflected i n F igure 13. So , us ing (29) for the water phase cross sectional area, and substituting into (40): Area- Qa 1 + 1.2 Qm Area J V = M — M ' mix n sub n total Area- Qa, 1 + 1.2 Qm Area (41) F i n a l l y , b y substituting (11) into (41), get: Area — a, 1 + 1.2 Qm Area Qair Qwater Area Y = H„,h-H, sub 1 1 total Area - - Qm 1 + 1.2 V Area (42) 76 B y us ing the measured water and air phase f l o w rates and pump geometry, equation (42) a l lows the tuning parameters d and e to be determined f rom the experimental results. The summaries o f co lumns 10 and 11 i n Table 4 and co lumns 11 and 12 i n Table 6 show good corre la t ion between the head losses i n the pumps and the mixture ve loc i ty as a measure o f turbulence. T h i s correlat ion was found throughout the experimental results, albeit more c o n v i n c i n g l y f rom the last phase o f the program i n w h i c h results were more rel iable than those previous due to factors already discussed. F igure 15 shows a summary o f the experimental data and mode l predict ions. The l ine o f best fit for the experimental data leads to the f o l l o w i n g relat ionship for the head loss as a funct ion o f the mixture f l o w ve loc i ty : constructing a s l igh t ly more conservative curve fit f rom the data lead to the f o l l o w i n g relat ionship for the head loss as a function o f the mixture f l o w ve loc i ty , Equa t ion (44) c o u l d be more suitable for design since it predicts a s l igh t ly higher head loss than the l ine o f best fit and w o u l d thus be a conservative estimate for pump capacity. (43) Hloxs=0.62-dens-Vn (44) 77 Figure 15 - Comparison of Experimental and Calculated Performance: Log(head loss/density) vs. Log(Mixture velocity) T h i s m o d e l for airl ift pump performance performs w e l l i n predic t ing the performance o f the low-submergence, low- l i f t , h igh - f low units investigated i n this project. It also provides the basis for a s imple approach to evaluating the behaviour o f these units and leads to a reasonably direct and practical design approach. 4.4 - Summarizing the three models The first m o d e l described i n 4.1 - Airlift Pump Model for Fixed Bubble Slip Velocities relies o n the assumption o f a re la t ively constant bubble rise speed, and fr ic t ional p u m p losses governed b y the ve loc i ty o f the water phase i n the air-water mixture . Investigations o f the experimental results and further research indicate that this assumption is not v a l i d i n the churn f l o w regime experienced b y the pump units i n this study. T h i s m o d e l does have promise i n applications where the constant bubble rise speed is supported, and m a y find use i n large diameter systems such as are used for lake destratification and harbour de- ic ing . The second m o d e l described i n 4.2 -Airlift Pump Model for Variable Bubble Slip Velocities also assumes losses governed by the ve loc i ty o f the water phase i n the air- water mix ture . It features a refined estimate for the mean bubble ve loc i ty as a funct ion o f the mixture ve loc i ty , a refinement based on the experimental w o r k o f several previous researchers. T h i s m o d e l does not accurately describe the behaviour o f the low- l i f t , h igh - f low, low-submergence pumps i n this study but does h o l d promise for use i n more energetic bubb ly f l o w regime pumps such as those used i n aquaculture and wastewater treatment applicat ions. 79 The th i rd m o d e l as described i n 4.3- Airlift Pump Model for Turbulent Mixing is specif ic to the churn flow regime. The refinement introduced i n the second m o d e l is retained, but a n e w formulat ion for the nature o f the head losses is u t i l i zed . H e a d losses i n the th i rd m o d e l are assumed to be proport ional to the turbulent m o t i o n o f the air and water phases i n the mixture . M i x t u r e ve loc i ty is found to be a good indicator for mix ture turbulence and is thus used as a basis for calcula t ing the turbulent head losses. T h i s m o d e l predicts the behaviour o f the pumps i n this study w i t h good accuracy and forms the basis o f the s imple design procedure presented i n Chapter 5. N o n e o f the procedures suggested i n the open literature are pract ical for the engineer w i s h i n g to des ign a low-submergence, low- l i f t , h igh - f low airlift pump system. W a r d ' s (1924) approach to design o f very long airlift pumps by curve matching includes no data for short length pumps and h igh v o i d fractions. N i c k l i n ' s (1963) technique for des ign o f airl ift pumps i n s lug flow, and a l l o f the suggested refinements to N i c k l i n ' s w o r k suggested b y subsequent researchers do not describe turbulent losses i n a churn ing system. T ramba ' s (1982) and Nenes & al ' s (1995) mul t i -ce l l ed s imulat ion-based numer ica l approaches for deep-wel l airlift pump analysis relies on d i v i d i n g the p u m p pipe riser b o d y into differ ing contiguous sections, each w i t h i nd iv idua l f l o w characteristics, a process not feasible for the short pumps described here. H o w e v e r , 4.3 - Airlift Pump Model for Turbulent Mixing described i n the previous chapter, provides the mi s s ing basis for a s imple and pract ical design procedure a l o w - submergence, low- l i f t h igh - f low airlift pump system. 80 C H A P T E R 5 5.1 - A Preliminary Design Procedure for Low-lift, Low-Submergence Airlift Pumps in the Churn Flow Regime T h i s procedure a l lows a designer to q u i c k l y complete the p re l iminary calculat ions for an airl if t p u m p i n a low-head, h igh- f low, low-submergence appl ica t ion for pract ical p u m p diameters i n the approximately 3 i nch to 12 i n c h range. Since airl ift pumps are inexpensive to construct, a prototype unit m a y then be bui l t and the performance ver i f ied . Because this design approach is s imple and based o n the f r ic t ion and ve loc i ty correlations developed i n this research program it should be used w i t h care i n cases o f m u c h higher lift and m u c h deeper submergence. In those cases the p ipe- f lu id f r ic t ion losses w i l l beg in to p l ay a larger part i n overa l l system behaviour as the bubbles i n the churn f l o w beg in to coalesce into T a y l o r bubbles and arrange themselves into a s lug f l o w pattern. The des ign procedure o f C l a r k & Dabo l t (1986) is recommended for use i n such cases. T h i s des ign procedure is used to predict the v o l u m e f l o w rate o f water expected f rom a low-head , h igh- f low, low-submergence airlift pump operating i n the churn f l o w regime. The des ign parameters required are: • Qair = the intended vo lume f l o w rate o f air, i n cubic feet per second • Qwater = the intended water f l o w rate, i n cubic feet per second 81 • Ups t r = the upstream water leve l , i n feet • Dnstrdes = the desired downstream water l eve l , i n feet The design procedure w i l l be i l lustrated by a s imple example. In this example an engineer desires to aerate and pump 10 cubic feet per second o f water over a 1.5 foot lift us ing a single or mul t ip le -p ipe airl ift pump system w i t h 8 inch diameter riser tubes. T h i s p u m p i n g system is set i n a smal l concrete drainage channel 8 feet w i d e b y 5 feet deep. F igure 16 shows the proposed layout o f the system. FIGURE 16 - Simple Design Example Layout A Simple 10-Step Design Process: 1. T h i s des ign appears to require several pump units to accompl i sh the required flow rate and lift . In such a case, assume an air flow rate and water flow rate for a Qwater 82 single pipe. W i t h no other informat ion avai lable , 5 0 % o f the air f l o w rate is found to be a reasonable starting estimate o f the water f l o w rate. A s s u m i n g 2.5 cfs o f air, and 1.25 cfs o f water: a ,>=2 .5c f s , e w ^ = 1 . 2 5 c f s ( D I ) 2. F o r l o w insert ion depths air can be considered incompressible , so calculate the v o l u m e f l o w rate o f the mixture by adding the vo lume f l o w rates o f the air and water: & * - Qair + QWa,er = (2.5) + ( l .25) = 3.75 cfs (D2) 3. Determine the mixture ve loc i ty by d i v i d i n g the mixture f l o w rate b y the cross- sectional area o f the riser pipe: Qmix (3.75 cfs) 0.785 • (0.75 f t ) : • = 10.7fps (D3) 4. Determine the ve loc i ty o f the air phase i n the pump riser pipe f rom equation (25): Vair = 1 + 1.2-Vmix = 1 +1.2 • (10.7) = 13.9 fps (D4) 83 5. Determine the sectional area occupied by the air phase i n the pump riser p ipe : Vair (13.9 fps) ) 6. Determine the relative density o f the air-water mixture i n the p u m p riser p ipe : Dens = 1 \Area) (0.35 s f j 7. Determine the system head loss f rom equation (44): - 0 . 5 6 - = 0 . 4 8 - ( 1 0 . 7 f p s ) 2 =1 .18f t (D7) 8. Calcula te the expected downstream water level f rom equations (4) and (5): fc,ra,c = fc^) = MzM) = 4 . 7 8 f t (D8) Dens 0.48 9. Compare the calculated downstream water depth f rom equation (D8) i n Step 8 to the desired downstream water depth. I f the calculated water l eve l f rom Step 8 is b e l o w the desired leve l the pump unit cannot provide the desired f l o w rate at the desired lift and g iven air f l o w rate. In such a case the water f l o w rate must be 84 decreased and/or the air f l o w rate must be increased. The reverse is true i f the calculated downstream leve l is above the desired height. 10. In this example w e select a lower water f l o w rate, leave the a i r f low rate as is and re-enter the process at step 1, n o w decreasing our assumption o f the water f l o w rate to Qwater = 1.15 cfs for this pump unit at this a i r f low. C a r r y out the steps again starting at step 1 and check the new result for the downstream water l eve l . Once reasonable agreement has been reached the pre l iminary des ign is complete . In this case the second tr ia l for the water f l o w rate was almost exact and thus a p re l imina ry estimate o f the pump unit performance has been made. In this example a pract ical operating point o f 1.15 cfs o f water and 2.5 cfs o f air i n a single 8" diameter pump w i t h a lift o f 1.5 feet has been established. Other operating points m a y be explored us ing the same technique unt i l a satisfactory operating point is selected. A s s u m i n g the p re l iminary design g iven above is satisfactory, and g iven the design requirement for 10 cfs o f water, a reasonable suggestion w o u l d be to instal l 9 o f the pumps as described, for a total water flowrate o f approximately 10.5 cfs o f water requi r ing approximate ly 23 cfs o f air. G i v e n that step 7 as s h o w n uses the fit values for the parameters i n equation (43) rather than the conservative values o f equation (44) it is reasonable to b u i l d a prototype unit based o n these specifications to check performance. A more conservative approach w o u l d be to use the "envelope curve" parameters from equation (44) i n step 7 instead. D o i n g so 85 results i n a predicted water f l o w rate for the example pump o f 1.05 cfs, ind ica t ing that 10 rather than 9 units cou ld be required. Alternate pipe diameters can be investigated easily, as can the influence o f a greater aeration depth, poss ib ly developed through excavating a sump at the site, etc. F o r example , the m o d e l suggests that the same pump cou ld produce a water flowrate o f 2.14 cfs i f the aerator were p laced i n a three-foot deep sump. H o w e v e r , this increase i n water flowrate at the same air flowrate is not entirely free since the air must be de l ivered at a consequently higher pressure, and subsequently a poss ib ly higher cost. T h i s des ign approach is c lear ly suitable for hand ca lcula t ion and can easi ly be automated by p rog ramming into a pocket calculator such as the Hewle t t -Packard 48 series or others o f s imi la r capabi l i ty . 5.2 - Design Calculations for Personal Computer The des ign procedure out l ined above is also suitable for implementa t ion i n a c o m m o n spreadsheet software such as Mic roso f t E x c e l © . Figure 17 shows a s imple formatted spreadsheet implementa t ion o f this design technique. The user enters des ign values i n the b o x e d cel ls marked as "Input" and the spreadsheet automates the subsequent ca lcu la t ion steps described above. S imp le changes to the system characteristics can be made and effects investigated. F i t parameters for the air phase ve loc i ty to mixture ve loc i ty relat ionship can also be adjusted i f desired, as can the fit parameters for the head loss to mixture ve loc i ty relat ionship. In this w a y the performance predict ions resul t ing f rom the 86 direct fit and conservative parameters can be investigated. The spreadsheet so lu t ion also a l lows for automated fast iteration to accurate solutions by means o f the M i c r o s o f t E x c e l © " S o l v e r " , " G o a l Seek", or equivalent user- implemented system. Figure 17 - Airlift Pump Churn Flow Worksheet Sample Airlift Pump Churn Flow Worksheet AB December 2003 Dens Hloss Qair = volume flow rate of air = 2.50 cfs Input Qw = volume flow rate of water = 1.16 cfs Input Diani = diameter of airlift pump tube = 0.67 ft Input Upstr = upstream water level above aerator = 3.50 ft Input Dnstrdes = desired downstream water level above aerator = 5.00 ft Input grav = acceleration due to gravity = 31.90 fpss Parameter low high fit a = curve fitting parameter in Vair=a+b*Vmix = 1.00 fps Parameter 1 3 1 b = curve fitting parameter in Vair=a+b*Vmix = 1.20 n/a Parameter 1.2 1.29 1.2 d = curve fitting parameter in Hloss = d*Dens*VmixAe = 0.56 n/a Parameter 0.56 0.62 0.56 e = curve fitting parameter in Hloss = d*Dens*VmixAe = 0.62 n/a Parameter 0.62 0.64 0.62 theoretical value for a above given slug flow athy = 0.35*sqrt(grav*Diam) - 1.61 fps Calculated volume flow rate of the mixture Qmix = Qair+Qw = 3.660 cfs Calculated cross sectional area of the airlift pump tube Area = PI()/4*DiamA2 = 0.349 sf Calculated velocity of the mixture Vmix = Qmix/Area = 10.486 fps Calculated velocity of the air phase in the airlift pump tube Vair = a+b*Vmix = 13.583 fps Calculated cross sectional area occupied by the air phase Aair = Qair/Vair = 0.184 sf Calculated relative density of the mixture in the airlift pump tube l-(Aair/Area) = 0.473 n/a Calculated head loss from curve fitting experimental data d*Dens*VmixAe = 1.136 ft Calculated calculated downstream water level Dnstrcalc = (Upstr-Hloss)/Dens = difference in calculated and desired downstream water Dnstrdiff = Dnstrcalc-Dnstrdes = 5.000 ft levels Calculated 0.000 ft Calculated 87 The user selects values for the system inputs and parameters and can explore var ious aspects o f the pump uni t ' s predicted performance. Iteration is s imple as the user adjusts the air and/or water f l o w rate unt i l the desired downstream and calculated downst ream depths are equal . The difference i n these depths is calculated at the bot tom o f the worksheet to facilitate the process. A goal-seeking a lgor i thm or system m a y also be used. The des ign procedure can also be coded into a functional fo rm for i n c l u s i o n i n other spreadsheets. T h i s approach makes the calcula t ion o f airlift pump behaviour immedia te . T h i s approach is also w e l l suited for tabulating predicted airlift pump behaviour and generating predicted performance values for various combinat ions o f des ign variables. The des ign procedure was coded into a set o f Mic roso f t V i s u a l B a s i c for A p p l i c a t i o n s © functions for use w i t h Mic roso f t E x c e l © spreadsheets.. F igures 18 and 19 show the Mic roso f t V i s u a l B a s i c for A p p l i c a t i o n s functions. 88 F i g u r e 18 - V B A © C o d e f o r C h u r n F l o w A i r l i f t P u m p D e s i g n F u n c t i o n Dns t rca lc ( B y V a l Qa i r A s S ingle , B y V a l Q w A s Single , B y V a l D i a m A s S ing le , B y V a l Ups t r A s Single) A s Single ' T h i s funct ion computes the downstream water l eve l g iven the f l o w o f water ( in cfs), ' f l o w o f air ( i n cfs), the pipe riser diameter, and the upstream water l eve l (both i n feet) D i m Q m i x A s S ingle , V m i x A s Single , V a i r A s Single , Dens A s Single D i m H l o s s A s S ingle , A r e a A s Single , A i r A s Single Q m i x = Q a i r + Qwater : A r e a = 0.785 * (D iam) A 2 A a i r = Q a i r / V a i r : Dens = 1 - ( A a i r / Area ) : H l o s s = 0.56 * Dens * ( V m i x A 0.62) Dns t rca lc = (Upstr - H l o s s ) / Dens E n d F u n c t i o n 89 F i g u r e 19 - V B A © C o d e f o r C h u r n F l o w A i r l i f t P u m p D e s i g n F u n c t i o n Q w a t e r ( B y V a l Qa i r A s Single , B y V a l D i a m A s Single , B y V a l Ups t r A s S ingle , B y V a l Dns t r A s Single) A s Single ' T h i s funct ion computes the f l o w o f water g iven the f l o w o f air ( in cfs), the p u m p riser 'diameter, the upstream water l eve l and the downstream water l eve l (a l l i n feet). It sets 'the water f l o w rate to zero and raises it i n smal l steps us ing Dnst rca lc un t i l the calculated ' and desired downst ream levels are equal . D i m Q w l A s S ingle , D n s t r l A s Single Q w l = 0: D n s t r l = Dnst rca lc (Qair , Q w , D i a m , Upst r ) I f D n s t r l <= Dns t r T h e n Qwater = 0 D o U n t i l D n s t r l <= Dnst r Q w l = Q w l + 0.01: D n s t r l = Dnst rca lc(Qair , Q w , D i a m , Upst r ) L o o p Qwater = Q w l E n d F u n c t i o n The disadvantage to the functional form described here is that it isolates the user f rom the intermediate values o f mixture density, air, water, and mixture ve loc i ty , etc. There is a greater opportunity for the user to trust poss ib ly questionable results because o f this disconnect. 90 5.3 - Practical Considerations for Preliminary Airlift Pump Design. Mixture Density: There are several pract ical considerations when us ing this approach. It w i l l become evident by us ing this des ign technique that the low- l i f t , low-submergence churn f l o w airl ift p u m p system is sensitive to the mixture relative density Dens. W h e n the mix ture relative densi ty falls m u c h b e l o w 0.5, d imin i sh ing returns set i n q u i c k l y i n terms o f increased water f l o w rate w i t h increased a i r f low rate. Once the mixture relat ive densi ty has fa l len m u c h b e l o w 0.45, increasing the a i r f low rate even dramat ica l ly w i l l produce very l i t t le increase i n f l o w o f water. In practice, increasing a i r f low past this l eve l w i l l eventual ly reduce the f l o w o f water since the air is d i sp lac ing water i n the pipe riser tube. The m o d e l presented here does not capture this behaviour at very h i g h air f l o w rates. H o w e v e r , that is not considered a fa i l ing because the phenomenon occurs far outside the pract ical range o f design. I f a designer finds h i m or herself attempting to b u i l d an airl if t system to operate i n such a scenario, prototype testing w i l l be required since the p u m p unit w i l l l i k e l y be operating i n the annular or mist f l o w regimes, w h i c h exis t ing airl if t p u m p theory cannot quantify. Air Pressure Required: The air pressure required for an airlift pump system is theoretically equal to the static water pressure at the aeration depth and an a l lowance for losses i n the air d is t r ibut ion system. In practice, i f us ing a mult i -port aerator the aerator ports should contribute a reasonable head loss themselves. P r o v i d i n g a notable pressure drop across the ports helps 91 ensure that a l l ports provide equal a i r f low, thus m a x i m i z i n g the aeration eff ic iency o f the mul t i -por t aerator. Thus the system designer should be prepared to provide a i r f l ow at approximate ly 0.5 to 1 ps i greater than predicted by the aerator submergence and air system dis t r ibut ion losses. Compressor Types Compressed air at l o w pressures and h igh vo lume f l o w rates such as is required b y an airl ift p u m p system o f this type can be obtained by several means. Ene rgy ef f ic iency o f these systems is l o w since m u c h is lost i n turbulence and m i x i n g . Because o f this energy ineff ic iency, requirements for power are reasonably h igh . (Fortunately, portable gas- powered sources are a ve ry v iab le alternative and can be used o n l y w h e n necessary). Centr i fugal b lowers are the most economica l means o f supply ing compressed air to an airl ift p u m p system, p roduc ing h i g h rates o f f l o w at l o w heads, t yp i ca l l y b e l o w 3 to 4 ps i . The V o r t r o n Z 4 0 , for example can easi ly generate 1000 scfm at 3 ps i w i t h a 40 hp motor . The centrifugal units operate at very h igh rotational rates, on the order o f 25 000 rpm, and must be muff led appropriately to avo id excessive noise output. Regenerat ive b lowers are somewhat more expensive than centrifugal b lowers but have the potential for a multistage design. In such systems operating pressures o f up to 9 ps i i n the 200 to 250 scfm range can be reached. The F P Z S C L - 1 1 5 - D H , for example , can generate 475 scfm at 9 p s i w i t h a 40 hp motor. The last type o f air supply machinery suitable for use i n air l if t p u m p i n g systems is the posi t ive displacement b lower . Because o f their des ign these units de l iver a re la t ively constant supply o f air governed b y displacement o f their internal lobes and the 92 rotational speed o f their impel lers . D e l i v e r y pressures up to 15 p s i are possible w i t h single-stage units. F o r example, the Sutorbil t 8 D H can generate 300 scfm at 15 p s i w i t h a 36 hp motor. A n airl ift pump system requir ing an air supply w i t h de l ive ry pressure above 15 p s i w o u l d feature an aerator submergence m u c h greater than those treated i n this study. In such a case air supply w o u l d l i k e l y be suppl ied by a rotary screw compressor (such as that used i n the second experimental phase o f this project). In such a case the design procedure o f C l a r k & Dabo l t (1986) w o u l d be recommended. 93 C H A P T E R 6 6.1 - C o n c l u s i o n s . Interest in low-head, high-flow, low-submergence airlift pump units has historically been low since such pumps are not particularly energy efficient and have been superceded by submersible electric rotomachinery for many decades. Despite having been replaced with more modern technology, airlift pumps are still used in several niche applications and offer some promising potential benefits in the field of urban stormwater management and other open-channel civil-engineering applications. Existing theory was evaluated and found inadequate to describe the behaviour of the low- head, high-flow, low-submergence airlift pumps. A four-stage experimental program was developed and implemented, including a full-scale prototype application in an urban storm drainage application in the city of Richmond, British Columbia. Performance data was collected. Three theoretical models were developed, with one satisfactorily fitting the experimental data. The model was translated into a practical procedure that an engineer may easily use to develop preliminary designs for airlift pumps operating in the churn flow regime. The design procedure was implemented in two personal-computer-based applications and thus can be quickly and easily completed. Some practical considerations for design of airlift pumps operating in the churn flow regime are given. 94 6.2 - R e s e a r c h R e c o m m e n d a t i o n s Now that the behaviour of airlift pumps in the churn flow regime has been modeled, the potential for uses of these systems in other low-lift, high-flow, low-submergence applications than those mentioned in this project should be explored. For example, airlift pumps may be useful in irrigation and other pumping in open channels. If so, methods of optimizing their performance in those scenarios must be developed. Additionally, the aquaculture potential of airlift pumping in shrimp and other invertebrate farming drainage applications should be investigated - the range of lift and flow rates are similar to those in urban drainage and aeration of the water may provide additional productivity benefits and cost savings through reducing the need for aeration equipment. The airlift pump seems to offer many advantages in the urban drainage setting, and the details of those advantages deserve to be investigated. Portable airlift pump units for local flood control may be practical, as might portable or "emergency only" trailer-mounted gasoline-powered air supply subsystems for permanently-installed units. The possibility of reduced environmental impacts in urban drainage subject to aeration as a side effect of airlift pumps should be investigated and subsequent benefits quantified. More work is also needed to better understand the underlying phenomena of the two- phase churn flow regime. Details such as the turbulent fluid behaviour at high void fractions, the manner in which bubbles accrete at high void fractions and the influence of aeration efficiency on regime stability are all unexplored. The development of high-speed 3-dimensional laser imaging technology may provide the necessary tools. 95 References Carte l l ier , A . "Simul taneous V o i d Frac t ion Measurement , B u b b l e V e l o c i t y , and S ize Est imate U s i n g a S ingle Opt i ca l Probe i n G a s - L i q u i d Two-Phase F l o w s " , R e v i e w o f Scient i f ic Instruments 63(11), (1992) C l a r k , N . N . & R . J . Dabol t , " A general design equation for air lift pumps operating i n s lug f l o w " , A m e r i c a n Institute o f C h e m i c a l Engineers Journal 32(1), (1986) D e Cachard , F . & J . M . Delhaye , " A slug-churn mode l for small-diameter air l if t pumps" , International Journal o f Mul t iphase F l o w 22(4), (1996) D u c k l e r A . E . , M . W i c k s & R . G . C leve land , "F r i c t iona l pressure drop i n two-phase f l o w : A . A compar i son o f exis t ing correlations for pressure loss and ho ldup" , A m e r i c a n Institute o f C h e m i c a l Engineers Journal (1964) D u c k l e r A . E . , M . W i c k s & R . G . C leve land , "F r i c t iona l pressure drop i n two-phase f l o w : B . A n approach through s imi la r i ty analysis", A m e r i c a n Institute o f C h e m i c a l Engineers Journal (1964) E l - K a m a s h , M . K . , M . R . L o e w e n & N Rajaratnam, "Measurements o f bubble properties and v o i d fraction o n a stepped chute", Proceedings o f the 1 5 t h Canad ian Hydro techn ica l Converence o f the Canad ian Socie ty for C i v i l Eng inee r ing (2001) Fernandes, R . C . , R . Semiat & A . E . D u c k l e r " H y d r o d y n a m i c m o d e l for gas - l iqu id s lug f l o w i n ver t ica l tubes", A m e r i c a n Institute o f C h e m i c a l Engineers Journal 29 , (1983) G i o t , M . , " H a n d b o o k o f Mul t iphase Systems", G . Hetsroni , ed., H e m i s p h e r e - M c G r a w H i l l , Wash ing ton , D C (1982). 96 G o v a n , A . H . , G . F . Hewi t t , H . J . Rich ter & A . Scott, " F l o o d i n g and churn flow i n ver t ica l pipes", International Journal o f Mul t iphase F l o w 17, 27-44 (1991). G o v i e r , G . W . , B . A . Radfo rd & J . S . C . D u n n , "The upwards ver t ical flow o f air-water mixtures" , Canad ian Journal o f C h e m i c a l Engineer ing (1957). I sh i i , M . & N . Zuber , " D r a g coefficient and relative ve loc i ty i n bubbly , droplet or particulate flows", A m e r i c a n Institute o f C h e m i c a l Engineers Journal 25 , (1979) Jayant i , S. & G . F . Hewi t t , "Pred ic t ion o f the slug-to-churn f l o w transi t ion i n ver t ica l two- phase flow", International Journal o f Mul t iphase F l o w 18(6), (1992) L a w n i c z a k , F . , P . Francois , O . Scrivener, E . G . Kas t r inak i s & S . G . N y c h a s , "The E f f i c i e n c y o f Short A i r l i f t Pumps Operat ing at L o w Submergence Ra t ios" , Canad ian Journal o f C h e m i c a l Engineer ing 77 (1999) Marka tos , N . C . & A . K . S ingha l , " N u m e r i c a l A n a l y s i s o f One -Dimens iona l , Two-Phase F l o w i n a V e r t i c a l C y l i n d r i c a l Passage", A d v a n c e d Engineer ing Software 4, 99- 106(1982) . M o r r i s o n G . L , T . I . Zeineddine , M . Henr iksen & G . B . Tatterson, "Expe r imen ta l A n a l y s i s o f the M e c h a n i c s o f Reverse C i r cu l a t i on i n A i r l i f t Pumps" , Industrial Eng inee r ing C h e m i c a l Research 26 (1987 Nenes , A . , Ass imacopou los , D . , Marka tos , N . & M i t s o u l i s , E . , " S i m u l a t i o n o f A i r l i f t Pumps for Deep Water W e l l s " , Canadian Journal o f C h e m i c a l Eng inee r ing 74, 448-456 ( D A T E ) N i c k l i n , D . J . , "The air-l if t pump: theory and op t imiza t ion" , Transactions o f the Institute o f C h e m i c a l Engineers 41 (1963) 97 N i c k l i n , D . J . , M . A . W i l k e s & M . A . D a v i d s o n , "Two-phase f l o w i n ver t ica l tubes", Transactions o f the Institute o f C h e m i c a l Engineers 40 (1962) P icker t , F . , "The theory o f the air-l if t pump" , Engineer ing , 34 (1932) Serdula , C D . & M . R . L o e w e n , "Exper iments Investigating The U s e o f F ibe r -Op t i c Probes F o r M e a s u r i n g B u b b l e S ize Dis t r ibut ions" , Journal o f Oceanic Eng inee r ing 23 , (1998) Shaw, " U n w a t e r i n g the T i r o Genera l M i n e by A i r l i f t " Transactions o f the A I M E 421 , (1920) Shel ton, S., & S . W . Stewart, " B u b b l e P u m p D e s i g n for Single Pressure A b s o r p t i o n Refr igera t ion C y c l e s " , A S H R A E Transactions 108 (2002) Ta i t e l , Y . , D . B o r n e a & A . E . B u c k l e r , " M o d e l l i n g F l o w Pattern Transi t ions for Steady U p w a r d G a s - L i q u i d F l o w i n V e r t i c a l Tubes" , A m e r i c a n Institute o f C h e m i c a l Engineers Journal 26, 345-354 (1980). Turer , D . , B . J . M a y n a r d & J.J . Sansalone, " H e a v y metal contaminat ion i n soi ls o f urban h ighways : C o m p a r i s o n between runoff and so i l concentrations", Journal o f Water , A i r and S o i l Po l l u t i on , 132(3) (2001). W a l l i s , G . B . , " O n e - D i m e n s i o n a l Two-Phase F l o w " , M c G r a w - H i l l , N e w Y o r k (1969). W a r d , D . "Expe r imen ta l Study o f A i r L i f t P u m p s " B u l l e t i n o f the U n i v e r s i t y o f W i s c o n s i n Engineer ing Services 3, 4 (1924) Wur t s , A . W . , S . G . M c N e i l l & D . G . Overhul ts , "Performance and design characteristics o f air l if t pumps for f ie ld applicat ions", W o r l d Aquacul ture 25(4) (1994) Z e n z , F . A . , " E x p l o r e the Potential o f A i r - L i f t Pumps and Mul t iphase Systems", C h e m i c a l Eng inee r ing Progress, 89(8), A u g u s t (1993) 98 —Si — a 1111111111111 mJIM n n nul l ill"™™" HI ill' ilF"|||i" IIIIIIIIIIII iiiinni1 illiiiiiill1 can lllllllllljl III ill' llll"l»lll IIIIIINllllllll ,1111111111 "llllllllllllll o > C: /AIRPUMP/AIRPUMP- 3.DWG -a o •!'!<i -7 < 5 < 5 < o < 5 < < < < < < •<5<5< < 5 < 5 < < 5 < 5 < <<<<< <<<<< <<<<< <<<<< FLOW TYPE 4 12"x4' PVC I k <<<< <<<< <<<< <<<<' .I > O "O n XI > zi O z o > o X) >

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
United States 8 0
Japan 3 0
China 2 8
Philippines 1 0
Canada 1 0
Turkey 1 0
City Views Downloads
Mountain View 7 0
Tokyo 3 0
Beijing 2 0
Unknown 1 0
Grand Forks 1 0
Cine 1 0
Ashburn 1 0

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

Share

Share to:

Comment

Related Items