Ekkad's-Gas turbine cooling

8/10/2019 Ekkad's-Gas turbine cooling

1/36

AdvancedInternalCooling

WithandWithoutEffectofRotation

SrinathV.Ekkad

MechanicalEngineering

VirginiaTech

[email protected]

Collaborators:Dr.DigantaNarzary,JustinLamont,Preston

Stoakes,andDr.MaryAnneAlvin(NETL)


2/36

Pro rammaticRelevance

IGCCplantsplayakeyroleinthefutureofDOEsCleancoalinitiativebyfacilitatingprecombustionCO2capturefromsyngas

GasturbinesusedinIGCCaresubjectedtohighthermalloadsandhightemperaturesalongwithresidualparticulateandvaporcontaminantswhichcouldpotentiallyalterthelifeofprecisionengineeredvanesandbladesinthehotgas

path Theproposedresearchaimstodevelopphysicsbased

modelingtoolstodevelopandpredictnewcoolingstrategiesforhotcomponentsandprovideeffectivecooling

sc emesw owcoo an usagew rec mpac onoverallefficiency

Workshoulddirectlyimpactmaterialsdevelopmentandcoat ngsa so


3/36

Doublewallcoolin usesathin a betweentwo

wallstoenhanceheattransferfromthesurfaceofturbineblades

Doublewallcoolingincreasesareaforheat

transferbetweencoolingfluidandmetal

Impingementjetsandmodifiedsurfacescanbe

usedtoincreaseheattransferontheouterwall

Nothingnewaboutthisconcept hasbeen

aroundforseveraldecades


4/36

PatentsbyBunkeretal.,Ishburg andLee, Jacksonetal.,Liang,Melvinetal.


5/36

Othercoolin o tionsincludin doublewall

Networkoforificesconnected

bysmallpassagestocreate

impingementareasand

Coolairisforceinto double

wallareathroughsmallpassagesandimpingesonthe

outerwall

practiceusedByRRinNorth

America(maturetechnology)


6/36

ribsusedtoincreaseturbulenceandheat

transfer

Coolinggasexhaustedthroughtrailingedgeof

blade

Designdoesnotuseimpingementcooling

LiangG,inventor;2004Oct.26.Coolingsystemforaturbinebladehavinga

doubleouterwall.UnitedStatesPatent

US 6 808 367.


7/36

Motivation&Objectives Gasturbinebladesneedtobeeffectivelycooledtoincrease

componentlifeandreducemaintenancecosts.Thelevelofcoolingisalwaysoffsetbytheamountofcoolantused.Increasedamountofcoolantusagedirectlyimpactstheoveralle c encyo eeng ne.

Usageofdoublewallcoolingschemescanreduceoverallcoolantusagebypushingcoolantclosertotheinsideofthewallexposedtohotgaspath.

een ance eattrans ertot ecoo antt roug t et nwa andalsoduetohighperformanceschemessuchasimpingementwillgreatlybenefitoverallandthermalefficiency

ofthesystem.

todevelopanoptimizationmethodologytodeterminethemosteffectivedoublewall/nearwallschemeforturbineairfoilcooling.Thesecoolinggeometrieswillbeoptimized

pressuredropwithoptimizationsoftwareandCFD.

tostudytheeffectofrotationondoublewallcoolinggeometriestoensureapplicabilitytorotatingblades

Typicalturbinebladeinternal

convectioncoolingconfiguration(Han

et.al.(1986).

workingwithOEMstocomparecoolingeffectivenessof

double

wall

geometries

compared

to

current

cooling

schemes


8/36

Objective: Exploretheuseofdoublewallcoolinginturbineblades,using

impingementcoolingandcombiningwithotherstandardheattransferenhancementtechniques

Developdesignmethodologiesforoptimizeddoublewall

Procedure:

UseCFDtoexploretheeffectivenessofimpingementcoolingin

UseCFDtooptimizedesignpatternofimpingementholesinchannelflow

configurationincomparisontocurrentstandarddesigns

Alsoexperimentallystudyeffectofrotationondoublewallcoolin desi nswithintentiontoo timizewithrotationaleffects


9/36

StationaryOptimizationStudy


10/36

ar w a mp e eome ry

channelwithanimpingementchannel

Mainchannelmeasures1x1foralltest

Impingementchanneldimensionsandlength

o testsect onvar es etweent ecases


11/36

2xN,Nvariesintheoptimizationstudy

Impingementchannel

confinesthespentcoolanttoexitopening

oppositeinlet


12/36

Dynamics(CFD)

foundinliterature

builtforexperiments

includeconductioneffects


13/36

eachonheattransferandpumpingpower

JettoJetSpacingRatio(L/D)

Numberofrowsofholes(N)

r owratevar e w t an toma nta najetReynoldsNumberof10,000


14/36

sectiondesignswereconsideredbychoosing

Range

fourdesignpointsfor

eachparameter D

1/32 1/4

(0.794mm) (6.35 mm)

Totallength,LT, held

constantwhenNis

L/D 2 5

H/D 0.5 4

N 5 11

in1/11 DLDL


15/36

transfercoefficientalongtheimpingement

, , , P

Thetwoparametersarecombinedtoforma

, ,

comparethetestsectiondesigns

0,

0

PP PPhhPer

Alltestsectionscomparedtobaselinedesign


16/36

, ,

spacingratio,H/D,appeartohavethelargest

JettoJetspacingratio,L/D,appearstohave

parameter

etop per orm ng es gnswere u ttovalidateCFDstudy


17/36

Rank D (in) L/D H/D N Per

1 0.125 5 2 5 51.7

2 0.125 2 1 5 51.6

. .

4 0.0625 2 1 5 50.2

5 0.25 2 2 5 50.26 0.125 2 2 5 49.9

7 0.25 3 2 5 49.3

8 0.125 3 2 5 48.9

9 0.0625 4 1 5 48.9

10 0.0625 3 1 5 48.8


18/36

highervelocitythroughlastrowofholes

Finaljetisdeflectedby

cross

flow

of

exhaust

gas

Contoursofvelocityatplaneintersecting

impingementjets

Hig estNusse t num er

valuesoccurunder

Exhaustgasappearsto

exitintwostreamsinline

ContoursofNusselt Numberon

impingementsurface

withimpingementjets


19/36

Experimentsarebeingconductedtovalidate

CFD

simulations

on

initial

test

section Turbulators canbeaddedtotheimpingement

channeltodisruptflowofexhaustgases

Theflowdisru tionshouldallowforamoreeven

heattransferdistribution,aswellasreducejet

deflectionduetocrossflow

Pinfinturbulators willincreaseamountof

conductionintothemainchannelandfurther

increasearea orconvective eattrans er


20/36

cases

gascrossflowwhenjettowallspacingratiois

below1

InitialCFDstudybaseonjettojetspacing

ratio

Comparedinlineandstaggeredarrangement


21/36

Multi le

Arra

Im in ement

for

Low

H/d

double

wall

cases

beincreasedbythewalljetdownstreamofthe

impingementjetand

not

by

the

impingement

e

Undevelopedcorejet

increaseheattransfer


22/36

MultipleArrayImpingementforLowH/ddoublewall

cases

2

ofjetsperformedbetteratalljettojetspacing 1.6

1.8

eter

ratios

Higherjettojetspacing

1.4

nc

ePara

ratioappearsto

increaseperformance1

Perform

Futurestudywill 0.6

.

1.5 2.5 3.5 4.5

Staggered Inline

spacing

ratios

L/D


23/36


24/36

Asimplifiedgeometrywastestedtodeterminetheeffectsof

rotationonbladeinternalcoolantchannelflow

Atypicalcoolantpassagefrom

turbinebladeismodeledinthisLeadingSide(Suction)

geometry.TrailingSide(Pressure)

Typicalturbinebladeinternalconvectioncooling

configuration(Hanet.al.(1986).


25/36

Therotatingrigspinsinterchangeabletestsectionsatdesired

speedswhilelowtemperatureairisinjected

NitrogenGasisventedintotheairpathbeforetestingto

chillcomponents. Thisdropsthetemperatureoftheair

before

reaching

the

test

section

during

a

test.

Acameraismountedtothetestsectiontoviewthecolor

. ,

sothemotorrotationdirectionisreversedwhenfilming

eithertheleadingortrailingside.


26/36

Theeffectofribtypeinrotationaltwopasschannelswas

exploredusingatransientliquidcrystaltechnique

Smoothwall,90Ribbed,andWshaperibbedwallswereexplored

inthestudy.

Eachcasewasheldata

Smooth Wall

,

of250rpm

(Rotationnumber=0.08).

Inletdensityratio=0.10

90 Ribs

Resultsarereportedforthe

trailingside,leadingSide,and

W-shape Ribs

.

comparedtothestationarycase.45

Pitch-to-rib height ratio (P/e)=8

=


27/36

Thehuefromtheliquidcrystaliscalibratedwithtemperatureto

calculatetheheattransferonthesurfacesVideo

Calibration Curve

Typical Temperature response

A

curve

fit

generates

an

algebraicexpressionfor

temperaturewithrespectto

uew c sapp e o e

restofthesurface


28/36

Asemiinfinitemodelisusedtodeterminethelocal

Atransientliquidcrystaltechniqueisusedtodeterminetheheattransfer.

Forthegiventestsection,thesemiinfinitesolidmodelisvalidiftheexperimenttimesdoesnotexceed

convectiveheattransferonthechannelwalls

25minutes.

Averagetesttimeislessthanaminute.Aftercalibratingtheliquidcrystalhuewithtemperature,walltemperaturesforeachpixelateachframe

inthevideoiscalculated.

Temperatureiscalculatedatthesurface(x=0)sothemathematicalmodelusedreducesto:

ththTtT i 2

),0(

kkTT i2

Ti : istheinitialwalltemperatureT(0,t):isthewalltemperature

: isthethermaldiffusivityoftheacrylic

Weknowallparameters,soweareabletosolveforhnumerically


29/36

The

li uid

cr stal

creates

a

continuous

lot

of

heat

transfer

on

thesurfaceofthecoolantchannel

Nu / Nu Nu / Nu

StationarLeadin Trailin

Nu / Nu

TheWribscreatethehighestheattransferinthechannel. The

smoothwallchannelhasthelowestheattransfer.


30/36

Areaaveragesinheattransferalongthelengthofthechannel

giveestimatestothepercentdifferences

Smooth Wall 90 Ribs

W-Shaped Ribs

Smooth Wall 90 Ribs W Ribs

Trailing Side 1st Pass 32 13 0.1

2nd Pass -21 -7 -9.3

Leading Side 1st Pass -19 -13 -3.3

2nd Pass 3 24 17.2


31/36

Theribtypesaredirectlycomparedforthestationary,trailingside,andleadingside

Stationary Trailing Side

Leading Side % Increases

1st Pass 90 Ribs 104 71 116

W-Ribs 260 190 325

2nd Pass 90 Ribs 35 55 76

W-Ribs 138 164 197


32/36

JetImpingementcoolingisanalternativetoribroughenedwalls

tocreatehighheattransfer

Insteadofatwopasschannel,a

radially outwardchannelwith

LeadingSide(Suction)

impingementisstudiedunder

rotation

Currentresultsshowimpingement

Trailing

Side

(Pressure)

channelheighttojetdiameter

ratio(H/d) =2

Pitchtojetdiameterratio

Typicalturbinebladeinternalconvectioncooling

configuration(Hanet.al.(1986).

(P/d)=8

Jetlengthtojetdiameterratio

(b/d)=1

Rotationalspeed=250rpm


33/36

Preliminarystationaryresultsforjetimpingementcoolingwith

crossflow effect

Flowmovesfromrightto

left. Thereisoneoutlet

fortheair,sothelaterjets

feeltheeffectsof

crossflow.

Crossflow bends

the

jets

,

reducingtheeffectiveness.

Theeffectsofcrossflow

,

astheaverageheat

transferreducesandthe

maximumheattransferfor

eachjetreducesasweget

closer

to

the

exit.


34/36

Preliminaryrotationalresultsfortrailingandleadingsideswith

crossflow effect

Inrotation,boththeleadingandtrailingsideresultsarelessthanthestationaryresults

.

Also,thetrailingsidereducedmorethantheleadingside. Thisiscounterintuitive

because

radially outward

flow

for

two

pass

channels

show

an

increase

in

heat

transfer

for

thetrailingsideduetothefavorableeffectsoftheCoriolis force.


35/36

Futurestudiesonribroughenedandjetimpingementcooling

schemesunderrotationRib Roughened Walls Jet Impingement

Further explore the W-shaped Ribs

with variations on flow orientationof ribs and angle of ribs

Explore rotational effects when the

impingement height is varied (H/d),when pitch is varied (P/d), effect of

film coolant extraction vs. crossflow

exit conditions, and Jet angle with

respect to the impingement plate is

varied.

Vs. Capabilities of the Rotating Rig arein the process of being increased.

Higher flow rates, lower

temperatures and higher rotational

speeds will be achievable.


36/36

Firstyearofproject

Understandingvariousparametriceffectsondouble

wallcoolingschemes

rotatingframeforinternalheattransfer

Fundamentalmethodolo ofo timizationachieved Focusingonmorerealisticgeometriesfromhereon

Workingwithindustrytodeterminefactorsof

evaluationanddesignmethodologyfornewcoolinggeometries

Ekkad's-Gas turbine cooling

Documents

Transcript of Ekkad's-Gas turbine cooling