Chapter-2

Literature <R.gview

2. Litera111re Review

LITERATURE REVIEW

Numerous papers arc available in the literature on the study of airlin pump. /\t

the present time, no workable theory is known which can act:urately predict the

phenomena of airlift pump in configuration of practical interest. To derive the

maximum benefit from the pump it is most essential to have an insight into the

operating parameters of the pump. /\n allempt has been made in this direction

in the present study.

2. 1) Literature on operating cri teria of airlift pump:

A good and thorough review of the slate of the art on airl i fl design was found

in White and Shelton and White (5). A total or 49 contributions are cited g iving

a detailed overview on the subject.

In the govern ing equations the submergence ratio of the airlift is used to

describe the average pressure gradient along the lift tube which can he

expressed in terms of velocities, geometrical parameters, and lluid propcrtics by

using momentum and mass balances. The two-phase fr iction factor and the two­

phase density in the lift tube, both based on average properties of the liquid and

gas. are as equation parameters. Thus most calculations of airlifts arc based on

assuming two-phase slug flow in vertical tube (Clark and Dabolt, 6).

Many of authors have inve~tigatcd the design of airlifts, whi le two-phase

vertical flow has been stud ied si nce almost the beginning of thc last century

(Stepanoff 7). Although the studies are avai lable that dcal with the optimization

of airl i ft design, hardly one can be applied in practice (White).

Pickert (8) and many other authors developed the basit:s for two-phase llow

theory. Nicklin (9) speculated that the efficiency of small diameter tub<:~ (< 20

mm) and of low now rates might be high (along with White and Beardmore

S a A111rm·ati University 10

2. /.iterot11re Rel'iew

and Zukoski)( IO, I I). Stcnning and Martin introduced the basic principle~ of

two-phase flow ( 12). Delano produces the same plots in his study (operation at

the maximum water llow rate)( 13 ). Kouremenos and Staicos devised the

experiments to obtain perfect slug How (14). Clark and DabulL pul>li:>hcd u

model to predict the height to which an airlift pump operating in the slug flow

regime can Ii fl a given volumetric tlow rate of liquid, given the air llow rate and

the pressure at the point of gas introduction. de Cachard and Oclhayc ( 15) used

an approximation of the general T ,ockhart-Martinelli correlation( 16 ). Chisholm

provides a good starling point for basic definitions and ll<l\\ pallcrns

encountered in vertical pipe flows (but no optimization of' the design)( 17).

Collier and I home presented the basic flow patterns: bubbly, slug, churn and

annular ( 18).

White describes the models available to calculate airlifLc; as IOllo\\cc.I: /\long

with Delano's model, several empirical models from previous worl-s arc

compared to tbe experimenta l data to see ii' they are valid in the range or

c.liamctcrs, submergence ratios. for all models the theoretical model or Stc1111i11g

an<l Manin is used as the starting point lo setup the relationship bctwcc11 the

suboergence ratio and the velocities (through momentum and mass balances).

Additionally, each model (except for Delano who uses an empirical conslanl)

uses the method of Be::attie and Whollcy ( 19) to find the two-phase friction

factor, and all but Delano uses the drift Oux method (Zuber and Findlay (20) to

find the gas void fraction. The differcm.:c between models is lhc v11lm: of th<:

coefficients used in the drift nux model. Lockhart and Martinelli <ldinc<l a lwo­

phasc multiplier, used with a single-phase pre!)sun.: drop calculation. White

Dt:a!lie and Whalley recommend using the Colebrook equation ror f'rtction

using two-phase properties in the Reynolds number.

S. G .-l111rm·ati Unfrersity 11

2. f . i1e r o11we J?eview

The drift nux model - th e most widely accepted m e thod - fo rmali 7.cd by Z ube r

and f"indlay provides a means to accoun t for the effects of the local relative

velocity between the phases as well a s the e ffec ts of non-unif<mn phase ve loc ity

and co11centration distributions. This theory is the basis lo describe the model

used today (Chcxal et a l )(2 1 ).

T hree of these mode ls a rc m eant for s lug regime only (N ick lin c t a l (22). de

Cachard a nd Dclhaye , and Reinemann e t a l (23)). The fourth model (Chcxal

and L e llouche ) is not fl ow reg ime sp ee i fie (24 ).

N ickl in et al formula ted an equation based on the work of Du111itrescu(25) and

O ev ics and T aylor(26) . This equation is o ften cons idered the original s lug flow

model. De Cachard and Oclhayc took surfoee tension e ffects in to account

(a lte rnati ve empirical corre lation's, using distribution para m eter Co - 1.2).

Rcinemann e l a l a r·e a lte ring th e su rface te ns ion c ffoc ts. C hcxa l and Le llouehc

devised an empirical program, in o rder to be able to use one model for a ll

regimes. Disadva ntage of the derived equations is that they arc rather complex.

The model of Dela no u ses the analysis of Stenning and Marlin .

A n optimum efficiency to operate at for a specific diameter and submergence

ratio was found very close lo the trans ition from s lug to churn flow. (I Jewill and

Roberts (27)). Jayanti and I Iewitt (28) concluded tha t the s lug to churn

transition is attributed to the fl ooding of the liquid film surrounding the Tay lor

bubble in s lug flow. McQui llan and Wha!Jey(29) , Nicklin and D <:1viclson .

Wnllis (30) • <1n<l C iovan et a l (31) came t<.> the sam e cu11clusion . From

experimen ts it is expected that d iameters less tha n 6 mm will have a high

effic ie ncy (Schaffer )(32).

The five different mode ls - ( I ) N ic kl in, (2) de Cachard, (3) Rcincmnnn. (4)

Chexal and L e llouch c a nd (5) De la no were imple mented in the Engim:cring

Eq ua ti on Solver (EES) by Klein and A lvarado (33). The resume is no single

S Ci. Amrm'lffi U11il 'l!rsi1y 12

2. f ,iterature Rel'iell'

model works for the entire range o f submergence ratio an<l/ or diameter!) und is

not useful for installation point of view.

Pickert concluded that the narrow diameter p ipes arc heller frorn cnicicncy

point or view. Martin C.13. challenged the resu lts obtained hy Pickcrt with

regard to the diameter of rising main describing that the larger diameter pipes

arc suitable for efficient airlift pump. The ambiguity was observed in the

experimental results obtained by Pickert and Marti n C.H. in terms or diameter

of the rising main with concern lo efficiency of pump.

As a pumping device, however, the airlift has wide variations in efficiency and

reduction of now below its optimal range increases hydraulic losses.

irrespective of lower pipe friction loss. Due to this reason. a famil iarity" ith the

hydrau lic pcrfonnancc of airlift is much more important tha n the knowll:dge or

pipe friction Joss. l fence the head loss due to friction is not focussed in the work

hccm1se of the concern for the difficulty of isolating and measuring these losses.

All the studies depend on ei ther an cxperimenra l data or empirical correlation

factors pertaining to specific categories of practical application. The present

work is focussed on the experimental analysis of the various parameters

affecting the performance of airli ft pump lo predict the pumping performance in

relation to the gas now rate and to gel the efficient operat ion o f the airlift pump.

The study is leading lo development generalized mathematical functions that

will find its usefulness in the optimum design of pump for its installation.

2.2) Literature on injecting device fo r airlift pump:

·1 he air injector design has a considerahle effect on the water <lisdiarge

as well as on the whole performances of the air lift fHlmp. Cl . .I. Parker (34)

reported the s tudy or small airlift pump made from 24.3111111-borc g lass tube

using l\\'o different designs of air injection foot piece. In one (Ilic air-jacket

S. G A111rm·a1i lJ11ii'ers11y 13

2. Literature Review

des ign), air was injected radially inwards, and in the o ther (the noi'.zlc des ig n).

a ir· was injected axia lly at inlet to the r iser. l·:ach design has bct:n tt:sled using a

va r-icty of injection ho le s izes un<..1 num bt:rs. With the a ir-j acket d t:sign. the

pump discharge cha racte ris tic was found to be independen t or the number an<l

s izes o f the injection holes. The 1107.7.le des ign showed greater pumping

capability at high airflow rates and with s m all ori fi ce area, but the efficiency

then was very low. Some comparisons with the theor·cti cal m ode l o r Stcnnino -=-

and Martin have been made, and the model has been extended to take account

of the m omentum o f ai r injected in the nozzle foo t piece.

D. A. Kouremcnos and J. S taicos focussed on th e pumping or liquids

us ing two-phast: Oow. The performance has been examined experimentally in

small airlifl pumps with 12- 19 mm bore P lexig las tubes . An ai r injecti on system

was devised to create and maintain ' p e r fect' s lug flow in the vertica l ri ser tu b1.:.

An equation has been derive d, based on momentum consen ·at ion

considerations, which correlates well with the measurements obta ined . Slip

variation, or liquid ho ldup, between the two phases and the formation or the

' entrance' scecion part of the pump (sut:tiun pipe) were taken into cons idera tion .

Unlike its predecessors, this equatio n predicts the reversa l in the pump

performance curve observed experimentally.

Airli rt pump p erformance was investigated experimentally hy M .F.

Khalil et a l (35) for different s ubmcrge::m.:c ra tio (the rati o be tween the

immersed leng th of th e riser and its total le ng th) using diffe rent designs of a ir

injection foot piece. For this purpose airlift pump w ith a ri st:r 200 cm long and

2.54 c m in diameter. was des igned and tested. Nine different air injection l(mt

piece designs were used at four-subm ergence ratios w ith d i fferent air injn:tion

pressu res (from 0.2 to 0 .4 bar). An area of J Osq.mm was chosen and di' idccl

into nine injections hole a rrangem ents ( 1, 2, 3, 4, 6, 1 5, 25, 34 and 4 8 ho ks) to

S. G. Amrornti U11il•ersi1y 14

2. Literature l?e1•iew

cover the whole experimental range. Four submergence ratios were used for this

work: 0.75, 0.7, 0.6 and 0.5. The experimental results showed a marked effc.:cl

on the pump performance when operated with different types or inj ectors at

different submergence ratios. The results indicated that the disk with three holes

(0 3) gives the highest efficiency at nearly all submergence ratios. Moreover, it

is foun<l that there is a suitable disk design for maximum water flow rate al

every submergence..: ratio. Further, the highest efficiency resulted at the largest

used submergence ratio, namely 0. 75. Tht: pump capacity and c.:flicicncy were

found to be functions ofairtlow rate, lirt ratio, and injection pressu re.:.

G.J.Parkar focused on the effect of foot piece (air jacket) with number or holes of same diameter. The foot piece used in this study was ori lice plate.: & it

was observed that there is no effect of foot piece hole configuration on

performance (discharge characteristics). Parkar had not focused on the issue or effect or different injection devices on performance of airlirt pump.

The injection device proposed by Parkar is an integration of nozzle &

orifice. This integrated device follow the different characteristic & orilicc of

smaller diameter results in the high pumping rate & quantity & it continues Lo

increase with decrease in orifice area. Parkar docs not study the elTeCl of

separate injection device. The foot piece proposed in the present study is

different than the foot piece proposed by Parker. The submergence ratio

parameter is neglected in the analysis of Parker.

M.F.Khalil focused on the effect of orifice plate as an injc..:ction dcvicL:

with number of holes of the same diameter. The effect of suhmcrgencc ratio

neglected by G.J.Parkar, which was presented by M.F.Khalil but the crlC.:ct

different injecting device is not co11sidcred in the study. The most important

parameter for studying the characteristics of the airlift pump is rate or flow or water. In the study presented by Khalil this parameter was ignored.

S. ( ;. A111raw11i University 15

2. literature Reviell'

T.K.Theyyunni (36) focussed the applicati on of Ouid technology in

nuclear chemical plant. The author has proposed an empirical formula to

determine diameter of orifice plate through which air can be introduced into the

foot piece. The equation is 02 = Q/435/ (P l-P2) P2rfl.

The equation developed is based on theoretical consideration of the work

done & expansion, which can not be justified for work of compression &

~xpansion of air in foot piece. An adiabatic process justifies the comprcssior) &

expansion of air.

The foot piece design has a considerab le effect on the water discharge as

well as on the complete performance of the airlift pump. The·individual effect

of design of foot piece is not yet reported in literature. The characteristic curves

of one injecting device arc not applicable to different shapes of injecting device.

Therefore, the present study is focused on the experimental analysis of d i ITcrent

designs of foot piece. The ignored parameters of rate of flow of water (Kha lil)

& submergence ratio (Parkar) arc also studied experimentally. An adiabatic

process justi fies the compression & expansion of air. Hence, the analysis is

carried out considering an adiabatic compression process of a compressor.

The published test results are not focused on the effective diameter of

nozzle (injecting device) su itable to diameter of the rising main. The present

,,·ork is leading to tbe development of mathematical function to determ ine the

dfcctivc diameter of nozzle. Also the new design of inj ecting device is

proposed to improve the effectiveness of a pump.

2.3) Literature on centrifugal pump handl ing two-phase now:

Requirements of handling two phase mixtures are increasing day by day

:md published test resu lts are few and are noL sui table from the insta ll ation point

,)f view. The abi lity or the centrifuga l pump to handle entrai ned gases is very

5. G. Amra1•ati University 16

2. Lit<?rature Rel'iew

limited. If the air / gas percentage increases beyond the 14 percent i11 volumetric

ratio, the pump stops work ing and is said to be air bound. Entrained air in a

centri l"ugal pump has undesirable effects on the pump operation and knowledge

of the pump performance under air admitting conditions is n<.:<.:dcd for actual

purpose.

In order to investigate the performance, a semi-open impeller pump with

a transparent casing was employed and the behavior o f the entrained air bubbles

in the pump was observed. The impeller work and hydraulic loss of the pump.

together with energy loss for air delivering arc d iscussed. The total :.1.:ad of the

pump is decreased by the a ir admission, but the head developec.J by the impeller

remains substantially constant. The drop of the pump head is mainly caused h)

the energy consumption for air delivering within the range or qs/Q <= 0.02, and

when qs/Q is increased beyond this range the hydrau lic loss due Lo the now

resistance of air bubbles increases its weight in the head drop. These relations

are given in the empirical formulas proposed by the M i.tsukiyo Murakami and

Kiyoshi Minemura(37).

Sim:i Sato cl a l (38) described the uir/watc1· two phase flow !'Crformam:c

of a centrifugal pump for five kinds of closed impellers, each of which has

different outlet or inlet angle. Th<.: r<.:sults showed that (i) s ucl c.kn pump head

c.Jegradati on due to gas accumulation in the impeller occur:> al a lower air Oo\\

rate with incr<.:asing in let o r outlet b lade angle of the impeller. (ii) l"hc pump

head remains high even under the condition of gas accumulation when the

impeller has large blade outlet angle. These r<.:sults are qualitatively d iscuss1.:c.J

including numerical calculations of one-dimensional two-phase flow.

l\.ir/wacer two-phase flow performance of centrifugal pump impellers

with various blade angles were investigated through experimenLs and

theoretical and calculations. T he results are summari zed as fo llows:

S. G. A111ravati University 17

2. Literature Re1·iell'

In the case of high incidence angle for an oncoming flow al low watcr now rate

and large blade inlet angk, the two-phase now pattern is changcd

discontinuously from a bubbly now along the blade prc~~un; surrai..:c 10 ii

separated flow with an air cavity on the blade suction surface al a certain Oow

rate ratio of air/water. at which the pump head drops sudden ly. In the case or

low incidencc ang le, air cavity appears on the blade pressure surface and head

is de-graded continuously. The abrupt pump head degradation due to appcarnncc

of the air cavity occurs at lower air flow rate, as the blade outlet angle of the

impeller becomes larger. However since water is able to flow with higher

Euler's head out of the impeller in a two-phase separated flow when the

impeller has a larger blade outlet angle. the pump head remains high even with

the ex istence of the air cavity.

The mechanism of air accumulation on the blade pressure surface in the

case of low incidence angle can be explained from a bubbly llow calculation as

an inviscid flow field. However, this calculation is powerlu l for pr1,;dicting the

discontinuous appearance o f the air cavity on the blade suction surface. In ortkr

to e lucidate its mechanism. the bubhle behavior in the shear fl ow with now

separation near the suction surface should be considered.

The use of centrifugal pump for handling two-phase mixtures is limited. The

two-phase mixtures of chemicals arc to be transported in many of the chemical

industries. The example in the process industry is that of pumping the mixture

of vi nyl chloride after oxychlorination of ethylcn1,;. Another example is that ol'

the process hydration of propylene with sulphuric acid for the production of'

isopropanol. The characteristics of two-phase centrifugal pump cannot be

applied to design the airlift pump. llence the comparative study of the con11no11

parameters to airlift pump and centrifugal pump handling two phase mixtun:s is

focussed in th1,; present work.

S. G. A11w<11·ati Univer.l'ify 18

2 l.iteratur<' Rel"iew

2.4) Lite rature on Computational Fluid Dynamics (CFD) aprlicd

to two-phase Ilow:

The phys ics in the bubble columns and airli n pump is simil ar. Shah YT.

Godbolc SP, Deckwer WO (39) has g iven a very good review of supcrli<.:ial gas

velocities in bubble column. Superficial gas velocity is the average velocity of

the gas that is sparged into the column which is, s imply expressed as the

volumetric Oow rate divided by the cross-sectiona l area or the column. ()as

holdup in bubble columns depends mainly on superfi cial gas velocity.

Hyndman CL( 40) and Schum pc A, Grund (J ( 41) have reported that for both

bubble columns and s lurry bubble columns, gas holdup has been found to

increase with increasing superficial gas velocity. Allhough the systems

investigated in these studies are qu ite di ffercnl fru111 each other. al I <.:oncludc

that the gas holdup increases with increasing superfi cial gas velocity

Recently Veera et al.(42) reported an experimen tal study based 011

investigation or effect of gas velocity on gas holdup profiles in roaming

liquids. They observed that the superficial gas velocity has a large influence on

rad ial holdup profi le at high foaming agent concentrations. /\t present time no

literature is available for application of CFO simulation to airlift pump. The

clear understanding of the physics of two-phase fl ow in th<:: airliti pump is only

possible w ith the CFD soHware. Tht: present work is addressed to the

development of two-phase CFO model to understand th<:: of air water llo\\' in

airlift pump. The CfD analysis is carried out w ith the help of' /\NSYS-CFX

software version 5.7. 1. The author has received a research grant of Rs. 8.75

lacs from A ICTE (TAPTEC Scheme), New Delhi. for establishing the Cr:D

centre nt the research laboratory of mechanical engineering department.

S. G. A111rurnti University 19

2. literot11re Rel'ie u·

The purpose of present investigation is to analyze the two-phase llO\\ and to

cval:rntc the performance for the design of airlifl pump. The recent pap~rs

without experimental data were of no use in the present work. IL is not claimed

that the author has processed data from all invest igations. It \vas rather dillicult

LO procure the data from many investigations because quite a large number of

such investigations were carried out long ago. The author has tried lo procure as

much data as possible from

i) The journals available.

ii) The literature available in the library of IlT, Llaus Khas. Delhi and llT,

Powai, Mumbai and llSc, Bangalore.

iii) ll1e British Library Services, London

The literature cited at the end of the list on references was extremely helpful to

the author in understanding the subject.

S G . .tl111ro1•m i Unfrersity 20