Tfe SQ010607

ExperimentalStudyandModelingofaLow
TemperatureRankineCycleforSmall
ScaleCogeneration.
SylvainQuoilin
ThesisSubmittedinPartialFulfillment
oftheRequirementsfortheDegreeof
ElectroMechanicalEngineer,
(Energeticengineering)
May2007
UNIVERSITY OF LIEGE
FACULTY OF APPLIED SCIENCES
AEROSPACE AND MECHANICAL ENGINEERING
DEPARTMENT
THERMODYNAMICS LABORATORY
2
Acknowledgements
IamverygratefultoProfessorLebrun,whomadethisworkpossible.Ithankhimfor
histrustandthegreatopportunitieshegaveme.
AspecialthanksgoestoVincentLemortforassistingmeduringthewholeperiodof
thiswork. Hisadvisesandhisconstantavailabilityeveninthedistancehavebeen
essentialandthisworkowesmuchtohim.
I would like to thank the members of the laboratory, C. Cuevas, B. Georges, J.
HannayandV.Teodoresefortheiravailabilityandtheirhelpinmanyoccasions.
This project wouldn't have been possible either without Jos Concha, Richard
LabendaandJeanMarieRalet,andthecaretheytooktopreparethetestbench.
IalsothankJessicaSchroufffortherereadingofthework.
Lastly,IwanttothankStphaneBertagnolio,LudovicBuckinx,LaurieDetrouxand
Pierre Gustin for their friendship and the excellent working atmosphere in the
laboratory.
Abstract
TheaimofthisworkisthestudyandthemodelingofanOrganicRankineCycle(ORC).
OrganicRankineCyclesareusedinlowgradeheatrecoveryapplications. Theirworkingfluidisa
refrigerantorahydrocarbonwhosepropertiesareadaptedtotheconditionsinwhichtheheatrecovery
isperformed.
AnexperimentalstudyiscarriedoutonanORCrecoveringheatfromhotairatatemperatureranging
from150to200C.Theexpanderusedforthisstudyisavolumetricscrollexpander.Theworking
fluidselectedisR123becauseofitshighefficiencyinORCapplications.
Duringthetestingperiod,thetestbenchisimproved:theefficiencyofthecycleanditsoutputshaft
powerareincreased.Thelimitsofthecycleareinvestigatedbymodifyingsomeparametersofthe
testbench:hotairsourcetemperature,expanderrotationalspeed,refrigerantcharge,etc.
Itturnsoutthattherefrigerantchargehasatremendousimportanceinthebehaviorofthecycleand
thatitsoptimalvaluehastobefoundinordertomaximizetheperformances.
Themeasurementuncertaintiesandtheincoherenciesdetectedduringthetestsareexplainedandan
uncertainty propagation study is performed. The bench is also adapted to minimize these
uncertainties.
Amodelofeachcomponentofthecycleisproposedandvalidated.Thewholecycleissimulatedin
ordertohaveagoodunderstandingofitsbehaviorunderdifferentworkingconditions.Themodelis
then used to optimize the cycle, startingfrom a basiccase and improving it inarealistic wayto
evaluatethesystempotential.
Finally,severalimprovementsforthetestbenchareproposed,thatcouldbeappliedinfutureworks.
3
Rsum
L'objetdel'tudeprsenteiciestlamiseaupointetlamodlisationd'uncycledeRankineorganique.
Les cycles de Rankine organiques (ORC) sont gnralement utiliss dans des applications de
rcuprationdechaleurbassetemprature.Leurfluidedetravailestunrfrigrantouuncomposant
hydrocarbon, dont les proprits sont adaptes au conditions dans lesquelles la rcupration de
chaleuresteffectue.
UnetudeexprimentaleestmenebiensuruncycledeRankineorganiquercuprantlachaleur
d'unesourced'airchaudunetempraturevariantde150200C.L'expanseurutilispourlebanc
d'essai est un expanser scroll. Le fluide de travail slectionn est le R123, possdant un bon
rendementdanslesapplicationsd'ORC.
Durantlapriodedetests,lebancd'essaiestamlior:lerendementducycleainsiquelapuissance
produite sont augments. Les "limites" du cycle sont explores en modifiant les diffrents
paramtres,telsquelatempraturedelasourcechaude,lavitessederotationdel'expanseur,lacharge
derfrigrant,etc.
Lachargederrigrantse rvletre unparamtretrsimportantpourlebonfonctionnement du
cycle.Cellecidoittreoptimiseafind'atteindredebonnesperformances.
Leserreursdemesureetlesincohrencesdtectesaulongdecestestssontexpliquesetunetude
de propagation d'erreur est effectue. Le banc d'essai est en outre adapt afin de minimiser ces
erreurs.
Unmodledechaquelmentdubancd'essaiestproposetvalid,cequipermetdesimulerlecycle
tout entier afin de bien comprendre son comportement sous diffrentes conditionsdetravail. Le
modleestensuiteutilispouroptimiserlecycle,enpartantd'uncasdebaseetenl'amliorantde
faonralisteafind'valuerlepotentieldusystme.
Enfin,plusieursamliorationspourlebancd'essaisontproposes,quipourraienttreappliqueslors
detravauxfuturs.
4
TableofContents
1Introduction......................................................................................................................... ........................8
2LowTemperatureHeatRecovery........................................................................................................ ........9
2.1TheRankinecycle.................................................................................................................. ......10
2.1.1Improvementstothecycle:........................................................................................ ..........11
2.1.2Realcycle:............................................................................................... ............................11
2.2ThelowtemperatureRankinecycle:....................................................................... ...................11
2.2.1Improvementsofthecycle:..................................................................................... .............12
2.3LowtemperatureRankinecycleapplications:................................................... .........................13
2.4WorkingfluidsforthelowtemperatureRankineCycle............................................................. ..14
2.5Othersolutionsforheatrecovery:...................................................................... .........................18
3DescriptionoftheRankineCycleTestBench............................................................... ...........................19
3.1Configurationofthecycle:............................................................................................ ..............20
3.2Scrollexpander........................................................................................................... ..................22
3.2.1Leakages............................................................................................................................ ....22
3.2.2Conversionofthescrollcompressorintoanexpander........................................................23
3.2.3Justificationofthechoiceofthescrollexpander...................................................... ...........25
3.2.4Pump......................................................................................................................... .................27
3.3Condenser................................................................................................................. ....................29
3.4Evaporator......................................................................................................................... ............31
3.5Draincock....................................................................................................................... ..............31
3.6Asynchronousmachine......................................................................................................... ........31
3.7Measurementdevices......................................................................................... ..........................32
3.7.1Thermocouples............................................................................................................ ..........32
3.7.2Pressuremeasurement........................................................................................................ ...33
3.7.3Torquemeter............................................................................................. ............................33
3.7.4Waterflowmeter.............................................................................................................. .....34
3.7.5Wattmeter......................................................................................................... ....................34
3.7.6Electricitycounter......................................................................................................... ........35
3.7.7Acquisitioncards............................................................................................................... ....35
3.8Coriolisflowmeter.................................................................................................................. .....36
3.8.1Generalprincipleofthecurvedtubeflowmeter.................................................................. 37
3.8.2Workingprincipleofthemeasurementdevice.................................................................. ...37
3.8.3Devicecharacteristics........................................................................................... ................39
4Descriptionoftheexperiments...................................................................................... ...........................41
4.1Firstsetoftests:.................................................................................................... ......................42
4.1.1Configurationofthetestbench:.................................................................................... ......42
4.1.2Descriptionofthetests:............................................................................. ..........................42
4.1.3Startingprocedure...................................................................................... .....................43
4.1.4Parametersandtestresults...................................................................... ........................43
4.2Secondsetoftests........................................................................................................... ........45
4.2.1Modificationsonthetestbench......................................................................... .............45
4.2.2Descriptionofthetests............................................................................. ......................45
4.3Criticalanalysisofthetwofirstsetsofmeasurements..................................... .....................47
4.3.1Airflowrate.............................................................................................. ......................47
4.1.2Refrigerantflowrate.................................................................................................... ....51
4.1.3Refrigerantlosses....................................................................................... .....................53
4.2Thirdsetoftests............................................................................................................... .......55
4.2.2Generalpurposeofthenewmeasurements:.............................................................. ....55
4.2.3Modificationsonthetestbench................................................................................ ......56
4.2.4Preparationofthenewtests................................................................................ ............58
4.2.4.1Refrigerantchargecalculation....................................................... .........................58
4.2.4.2Expansiondevicecalculation..................................................... ............................59
4.2.4.3Checkoutofthepumpbehaviorbymakingitruninstandalonemode.................62
4.2.4.4Calibrationofthewaterflowmeter................................................................ ........64
4.3Descriptionofthetests....................................................................................... ....................65
4.4Analysisoftheresults..................................................................................................... ........68
4.4.2Heatbalances.............................................................................................. ....................68
4.4.3Pumpandrefrigerantflowrateanalysis............................................................ .............70
4.4.4Refrigerantchargeandsubcooling........................................................ .........................71
4.4.5Asynchronousmachineefficiency....................................................................... ...........72
4.4.6Presenceofnoncondensablegasinthecycle.......................................................... ......73
4.5Conclusion.............................................................................................................................. .74
5Models.............................................................................................................................................. ....75
5.1Introduction.......................................................................................................... ...................76
5.2Expandermodel.................................................................................................... ..................77
5.3Condensermodel............................................................................................... .....................85
5.3.2Onezonemodel................................................................................................... ...........85
5.3.3Threezonemodel...................................................................................... .....................86
5.4Evaporatormodel.............................................................................................................. ......90
5.4.2Firstmodel............................................................................................. .........................90
5.4.3Secondmodel............................................................................................................ ......92
5.4.4Singlephasemodel........................................................................................ .................94
5.5Pumpmodel........................................................................................................ ....................97
5.6Linesmodel........................................................................................................................... ..98
5.7Refrigerantchargemodel.................................................................................................... ....98
5.8Globalmodelofthecycle......................................................................................... ............100
5.9Optimizationofthecycle............................................................................................... .......103
5.10Optimizedcycle......................................................................................................... ..........108
6RecommendationsandPerspectives............................................................................................ ......109
6
7Conclusion......................................................................................................................................... .113
8Bibliography................................................................................................................................... ....115
Appendix1:Othersolutionsforlowgradeheatrecovery....................................................... .............118
1.Thetranscriticalcycle................................................................................................... .......118
2.Thewaterammoniacycle................................................................................................... .119
3.TheStirlingcycle.......................................................................................... ......................122
4.TheEricssonCycle:....................................................................................... ...................124
5.TheThermoelectriceffect................................................................................................ ...125
Appendix2:Firstseriesoftests.................................................................................... .......................128
Appendix3:Secondseriesoftests.................................................................................................. .....129
Appendix4:Thirdseriesoftests..................................................................................................... .....130
7
1 Introduction
Over the last century, the world economical growth has accelerated dramatically. The industrial
development, the increasing number of vehicles on the road and the multiplication of energy
consumingdomesticequipmentshavecausedanimportantgrowthoftheenergydemand.
Unfortunately,thisdemandhasbeenmostlycoveredbyamassiveconsumptionoffossilfuels,which
causesmanyseriousenvironmentalproblems,suchasglobalwarmingoratmosphericpollution.
New energy conversion technologies are required in order to insure the production of electricity
withoutgeneratingenvironmentalpollution.Amongthem,lowgradeheatsourcesareconsideredas
appropriatecandidatesforthenewenergysources.
Theinterestforlowgradeheatrecoveryhasgrowndramaticallyinthepastfewyears.Animportant
numberofnewsolutionswereproposedtogenerateelectricityfromlowtemperaturesources.Those
solutionscanbeappliedtoverydiversifiedfieldssuchassolarthermal,biologicalwasteheat,engine
exhaust gases, small scale cogeneration, domestic boilers, etc. They can be decentralized and
producesmallamountsofenergy.
Thepurposeofthisworkistogiveanoverviewoftheexistingtechnologies,andtoanalyzeapractical
applicationbasedonthemodelingandtestresultsofaRankineCycletestbench.
Thefirstpartoftheworkisdedicatedtothestudyofthebestsuitedcyclesforheatrecovery,with
theiradvantages,drawbacks,andpossibleapplications.Themainparameterstakenintoaccountare
efficiency,specificwork,andsimpleness.
Acomparisonbetweendifferentworkingfluidsisperformedtodeterminethemostappropriateand
environmentfriendlyones.
The expander having a crucial importance, a chapter is dedicated to its study and a comparison
betweendifferentexpandertypesisperformed.Thechoiceofthescrollmachineselectedforthistest
benchisjustified.
Thetestbenchisthendescribed,withacompleteoverviewofitscomponentsandofitsmeasurement
devices.Theexperimentalprocessisexplainedaswellasthepracticaldifficultiesandtheproposed
solutions.
Acriticalanalysisofthemeasurementisproposed,inordertoexplaintheincoherenciesdetectedin
thetests.Amodelofeachelementofthecycleisdescribedandvalidated.Asimplifiedmodelofthe
refrigerantchargeanditseffectsonthecycleisalsoproposed.
The possible improvements on the test bench andonthe models arefinally evaluated,and afew
recommendationsareexpressed.
8
SylvainQuoilin Chapter2:Lowtemperatureheatrecovery
Chapter2:
LowTemperatureHeatRecovery:
StateofTheArt
2 LowTemperatureHeatRecovery
9
Thischaptergivesabrief theoreticalintroductiontotheOrganicRankineCyclesandtotheother
possiblesolutionsforheatrecovery.
AnanalysisoftheoptimalworkingfluidfortheORCisproposedandafewpossibleapplicationsare
detailed.
2.1 TheRankinecycle.
ThewellknownRankinecycleisthemostusedintraditionalpowerplants. Theworkingfluidis
pumpedtoaboilerwhereitisevaporated,passesthroughaturbineandisfinallyrecondensed.
In the ideal cycle, the expansion and compression are isentropic, and the heat transfers in the
evaporatorandthecondenserareisobaric.
Figure1:IdealRankineCycle
Theshapeofathermodynamiccyclecanbeusedtogrosslyevaluateitsefficiencyandspecificwork:
TheTsdiagramcangiveanideaofthecycleefficiency.Themorerectangularthecycle,the
closertotheidealCarnotcycle,andthehighertheefficiency.
Thepvdiagramcangiveanideaofthespecificworkproducedbythecycle. Thisworkis
givenbytheareabetweenthecurvesofthecycle(yellowarea).
Thetwoconditions(cycleefficiencyandhighspecificwork)areantagonist:amorerectangularcycle
usuallymeansaflatterpvdiagram.
InthecaseofatraditionalRankinecycle,thepvdiagramshowsagoodspecificworkbutthets
diagramisnotrectangularanditsefficiencyisthuslowered.
10
2.1.1 Improvementstothecycle:
ThetwomainimprovementsusuallybroughttotheRankinecyclearethefollowing:
Regeneration:Thevaporintheturbineisusedtopreheatthefluidattheexhaustofthepumpin
ordertospendlessenergyintheboiler.
Reheating during theexpansion : theexpansion isfractioned andthefluidisreheated, which
permitskeepingahighsupplytemperaturebeforeeachexpansion.
2.1.2 Realcycle:
Theeffectoftheirreversibilitiesinthecycleisareductionofcycleefficiencyandofusefulwork
output.
Themainirreversibilitiesare:
Lossesinthepumpandintheexpander(friction,leakage,etc.)
Pressuredropsintheheatexchangers
Inefficienciesintheheatexchangers
Thepressuredropsandtheinefficienciesintheheatexchangerslowerwiththesizeoftheexchanger,
butincreaseinthesameprocessthecostandthesizeofthefacility.
2.2 ThelowtemperatureRankinecycle:
TheusualworkingfluidforRankinecyclesiswaterunderpressure.Inthecaseofalowtemperature
Rankine cycle, the boiling temperature is much lower, and the water/steam working fluid is not
appropriatebecauseofitslowefficiencyundertheseconditions.Wateralsoshowsahighvaporization
specificvolumethatimposeslargerinstallations.
Thisleadstotheuseofrefrigerantsorhydrocarbonsasworkingfluid.Thephysicalpropertiesofthe
working fluid are of key importance and influence the whole behavior of the cycle. A detailed
analysisofthedifferentworkingfluidsisperformedinsection2.4.
Variousconfigurations areconceivablefortheRankinecycle,dependingofthenatureoftheheat
sourceandoftheconditionsinwhichtheheatrecoveryisperformed.
11
2.2.1 Improvementsofthecycle:
TheORCefficiencycanbeimprovedbytheadditionofaregenerator(figure2).
Theuseofaregeneratorisjustifiedwhentheslopeofthesaturationvaporcurveispositive(seefigure
3),ormoregenerallywhenthefluidisstillstronglyoverheatedaftertheexpansion.Theregenerator
issituatedattheexhaustofthepumponthehighpressureside,andbetweentheexpanderandthe
condenseronthelowpressureside.
12
Figure2:ORCwithregeneration
2.3 LowtemperatureRankinecycleapplications:
3maincategoriesofapplicationsareconceivablefortheOrganicRankineCycle(ormoregenerally
forthelowtemperatureRankinecycle):
Wasteheatrecovery
Where the ORC is most useful is in the recovery and use of waste heat. Two primary
applicationsincludeCombinedHeatandPowerplants(especiallythoseutilizingbiomassas
fuel),andgeneralheatrecoveryapplicationsfrommanypotentialsources.
Acombinedheatandpowerplantcanforexamplebeasmallscalecogenerationplantona
domesticwaterheater.Inthisperspective,twooptionsareconceivable:theprioritycanbe
giventothepowercycle,thelatterrecoveringtheheatproduceddirectlyintheboilerandthe
hotwaterbeingproducedatthecondenseroftheORCatalowertemperature,orthepriority
canbegiventothehotwater,inwhichcasetheheatsourceofthecycleistheheatrecovered
onthegasesattheexhaustofthecombustionchamber.
Heat recovery can also be performed on industrial or farming processes such as organic
products fermentation, hot exhausts from ovens or furnaces, exhaust gases from vehicles,
intercoolingofacompressor,condenserofapowercycle,etc.
Solarthermalpower
OrganicRankinecyclescanbeusedinthesolarparabolicthroughtechnologyinsteadofthe
usualsteamRankinecycle.TheORCallowsalowercollectortemperature,abettercollecting
efficiency(reducedambientlosses)andhencethepossibilityofreducingthesizeofthesolar
fields.
SeveralexampleofparabolictroughsusingORCareavailableintheliterature:
S. Canada presented a parabolic trough ORC Solar Power Plant, using npentane as
workingfluidandwithaninlettemperatureof204C[Canada,2005].
E.H.MalickKane,inhisdoctoratethesis,studiedtheintegrationoftwosuperposedORC
onaparabolicthoughsolarcollector,thetoppingcycleusingR123asworkingfluid,and
thebottomingcycleusingR134a[Kane,2002].
Geothermalplants
Geothermic heat source vary in temperature from 50 to 350C. For lowtemperature
geothermalsources(typicallylessthan100C),thepowerplantefficiencyisverydependent
ontheambienttemperature,thatdeterminatestheheatsinktemperature.
13
2.4 WorkingfluidsforthelowtemperatureRankineCycle.
TheselectionoftheworkingfluidisofkeyimportanceinlowtemperatureRankineCycles.Because
of the low temperature, heat transfer inefficienciesarehighlyprejudicialtotheefficiency. These
inefficiencies depend very strongly on the thermodynamic characteristics of the fluid and on the
operatingconditions.
Optimalcharacteristicsoftheworkingfluid:
Isentropicsaturationvaporcurve.
SincethepurposeoftheORCfocusesontherecoveryoflowgradeheatpower,asuperheated
approachlikethetraditionalRankinecycleisnotappropriate.Therefore,asmallsuperheating
at the exhaust of the evaporator will always be preferred, which has consequences on the
optimalthermodynamicpropertiesofthefluid:
ThesaturationcurveintheTsdiagramcanhaveapositive,negativeorverticalslope(figure
3).
Thepreferred characteristic for lowtemperatureRankinecyclesistheisentropicsaturation
vapor curve (R11 in figure 3). Indeed, the saturation curve is parallel to the line of the
14
Figure3:Wet,isentropicanddrysaturationvaporcurve
isentropicexpansionandfitsverywellthecycle.
Inthecaseofnegativeslopesaturationcurve(R22infigure3),thefluidbecomessaturated
beforetheendoftheexpansion,whichisnotoptimalbecausethecondensateimposesathreat
ofdamagetotheexpander.
Inthecaseofapositiveslopesaturationcurve(isopentaneinfigure3). Thefluidhastobe
cooleddownattheexhaustoftheexpanderbeforeenteringthetwophasestate.Thiscanbe
doneby theuse of aregeneratorbetweentheexhaustofthepumpandtheexhaust of the
expander.
Lowfreezingpoint,highstabilitytemperature
Unlikewater,organicfluidsusuallysufferchemicaldeteriorationsanddecompositionathigh
temperatures.Themaximumhotsourcetemperatureisthuslimitedbythechemicalstability
oftheworkingfluid.
Thefreezingpointhastobelowerthanthelowesttemperatureinthecycle.
Highheatofvaporisationanddensity
Afluidwithahighlatentheatanddensitywillabsorbmoreenergyfromthesourceinthe
evaporator and thus reduce the required flow rate, the size of the facility, and the pump
consumption.
Lowenvironmentalimpact
ThemainparametersaretheOzonedepletingpotential(ODP)andgreenhousewarming
potential(GWP).
Safety
Thefluidhastobenoncorrosive,nonflammable,andnontoxic.Dependingonthelocation
ofthefacility,thisparametercanbeofveryhighimportance(e.g.anammoniacyclewillnot
beinstalledinadwellingwithoutanyprecaution).TheASHRAEsafetyclassificationofthe
refrigerantisagoodindicatorofthedangerousnesslevelofthefluid.
Goodavailabilityandlowcost
TraditionalrefrigerantsusedinORCareexpensive. Thiscostcouldbereducedbyamore
massiveproductionofthoserefrigerants,orbytheuseoflowcosthydrocarbons.
Acceptablepressures
Veryhighpressureshaveanegativeimpactonthereliabilityofthecycle. Theyleadtothe
needofmoreresistantandmoreexpensivefacilities.
15
Aquick review of the literature aboutlowtemperatureRankinecyclesgivesanideaofthe usual
workingfluidsusedinORCsystems.Table1and2givethemaincharacteristicsofthosefluids,from
anenvironmentalandthermodynamicpointofview.
Refrigerant Atmospheric
lifetime
ASHRAELevel
ofsafety
ODP NetGWP100year
(2102)
PhaseoutYear
R11 45 A1 1 3660 1996
R22 12.0 A1 0.034 1710 2020
R113 85 A1 0.90 5330 1996
R123 1.3 B1 0.012 53 2030
R134a 14.0 A1 ~0 1320
R245fa 7.6 B1 ~0 1020
R717(ammonia) B2 ~0 <1
R601(npentane) ~0 ~20
R601a(isopentane) ~0 ~20
Table1:Environmentaldataforhistorical,current,andcandidatechillerrefrigerants
A first selection can be achieved among those fluids given the working temperature range. The
applicationenvisagedinthepresentworkhaveaheatsourcetemperaturerangingfrom100to200C
andanheatsinktemperaturerangingfrom10to50C. ItisobviousthatR22andR134acanbe
eliminated,theirtemperaturerangebeingtoolow(seetable2).Inthesamemanner,water,benzene,
tolueneandpxylenehaveatoohightemperaturerangeandcanbeeliminatedaswell.
From anenvironmental pointof view,R11andR113haveaveryhighozonedepletingpotential
(ODP),andwerephasedoutin1996bytheMontrealprotocol. Therefore,theywon'tbetakeninto
accountinthiswork.
16
Workingfluid Slopeofsaturation
vaporline
Criticalpoint Vaporization heat
at 1 atm.
[Kj/kg]
Boiling
temperature at 1
atm.
Safety
Water Wet 374C220bar 2258 100C Nonflammable
R11 Isentropic 198C44.1bar 180.4 23.5C Nonflammable
R22 Wet 96.1C49.9bar 232.7 41.1C Nonflammable
R113 Dry 214C34.4bar 143.9 47.8C Nonflammable
R123 Isentropic 184C36.7bar 171.5 27.7C Nonflammable
R134a Wet 101C40.6bar 217.2 26.4C Nonflammable
R245fa Isentropic 154C36.4bar 197.5 14.6C Nonflammable
R601(npentane) Dry 196C33.6bar 358.7 35.5C Flammable
R601a(isopentane) Dry 187C33.7bar 342.8 27.5C Flammable
C6H6(benzene) Dry 289C49bar 395.4 79.8C Flammable
C7H8(Toluene) Dry 319C41bar 362.5 110.4C Flammable
C8H10(pxylene) Dry 343C35bar 339.9 66.65C Flammable
Table2:Thermodynamicpropertiesofafewworkingfluids
Theworkingfluidsremainingfromthisfirstselectionare:R123,R245fa,npentaneandisopentane.
Eachofthemhasitsadvantagesanddrawbacks:
R123andR245faarenonflammableandisentropic,whilenpentaneandisopentanehaveahigher
vaporizationheat,alowerGWP,andamuchlowercost.
R123alsohasthedrawbackofanonnullODPandwillbeforbiddenin2030.
Themainparametertotakeintoaccountistheefficiencyofthecyclewitheachoneofthosefluids.
Severalsimulationsperformedwithdifferentworkingfluidsareproposedintheliteratureandgivean
ideaoftheirpotentialefficiencies:
Liuetal[Liu,2002]performedasimulationonanORCwithvariousworkingfluidsforahot
temperatureof150Candacoldtemperatureof30C.ItturnedoutinthissimulationthatR
123hadaslightlybetterefficiencythanisopentane. R245faandnpentanewerenottaken
intoaccountinthatstudy.
V.Lemort[2007]comparedthreeworkingfluidsonanORCapplication(hotsidetemperature
:130C;coldsidetemperature:30C).The3workingfluidswereR123,R245fa,R134a
andnpentane.Thecalculatedefficiencieswererespectively:9.71,9.3,7.86and9.74%.
H.D. Madhawa [Hettiarachchia, 2007], simulated a geothermal ORC with a hot source
temperatureof70to90C,atemperaturedifferencerangingfrom40to60degreesandvarious
workingfluids. Amongthoseworkingfluids,R123andnpentaneoutperformedtheother
fluids(ammoniaandPF5050)withahigherefficiencyof9.8and9.9%respectively.
17
2.5 Othersolutionsforheatrecovery:
TheORCisnottheonlyconceivablepossibilitytorecoverlowgradeheat. Sometechnologiesare
availableandshowgoodefficiencies,someothertechnologiesarestillunderdevelopment butseem
promising.Amongthosetechnologies,themostinterestingonesarethefollowing:
The waterammonia cycle. Thiscycleusesamixtureofwaterandammoniaasworking
fluid.Ithasanonazeotropictemperatureprofileintheheatexchangers,whichimprovesthe
heattransfer.ItturnsouttohaveabetterefficiencythantheORCinsomeconditions,butits
designismorecomplexandcomprisestwopumpsandmoreheatexchangers.
The supercritical CO2 cycle. The working fluid is carbon dioxide. The heat source
temperatureisabovethecriticaltemperature,whichavoidsthepinchpointlimitationinthe
evaporator.Thiscyclealsoshowsagoodefficiency,thedrawbackcomingfromtheveryhigh
pressure,andfromtheincreaseoftheinstallationcost.
TheStirlingandtheEricssoncycleshaveaverygoodtheoreticalefficiency.Thenumberof
theirpracticalapplicationsishoweverlimited,sincetheyrequirehighlyefficientheattransfers
intheregeneratorandattheheatsource/sinklevel.
Thethermoelectricgenerator. Itiscomposedofmodulesassociatedinseries,generating
electricity,thankstotheSeebeckeffect. Itshowsa lowerefficiencycomparedtotheabove
cycles,buthasahighpotentialofimprovementandcouldturnouttobecomeeconomically
satisfactoryinthefuture.Itisalsoquitesimple,withnomovingparts,andmightthushavea
betterdurability.
A complete description of the working principles concerning those technologies is proposed in
appendix1.
In this work, the Organic Rankine cycle (ORC) is preferred to those technologies, because of its
simplicityanditslimitednumberofcomponents,allofthembeingverycommonandcommercially
available.
18
SylvainQuoilin Chapter3:DescriptionoftheTestBench
Chapter3:
DescriptionoftheRankine
CycleTestBench
3 DescriptionoftheRankineCycle
TestBench
19
ThischapterdescribestheexperimentalstudycarriedoutonaheatrecoveryRankinecyclebench.
Forthepurposeofthisstudy,atestbenchissetupinthethermodynamicslaboratoryinLige.The
aimistoprovethefeasibilityofanORC,toevaluatethepotentialheatrecoveryefficiency,andto
validatethemodelsofthedifferentcomponents.Figure4showsaschematicviewofthefacility.
3.1 Configurationofthecycle:
TheworkingfluidusedisR123.
Theheatsourceiscomposedoftwohotairflowsatatemperaturerangingfrom150to190C.One
flowgoestotheheatexchangersHX1andHX2andtheotherflowgoestotheexchangerHX3.
Theexpanderprototypeisanoilfreeairscrollcompressoradaptedtoberuninexpandermode.Itcan
beisolatedandbypassedthanksto3bulbvalves,inordertostartthetestinappropriateconditions.
Theexpanderiscoupledtoanasynchronousmachinebymeansofatorquemeterandtwobeltand
pulleycouplings.Theasynchronousmachineworkseitherinmotormodeorbrakemode,depending
onthepowerdevelopedbytheexpander.Anasynchronousmachineischosenforitsabilitytoimpose
therotationalspeedtotheexpander.
Thecondenserisfedwithtapwater.Figure4showsitsconfigurationwith2heatexchangersputin
parallel.However,severaldispositionsweretestedandwillbediscussedinthefollowingchapters.
Theliquidcondensateleavingthecondenseristhenpumpedtowardsthehigherpressureboiler.
20
21
figure4:ViewoftheRankineCycletestbench
3.2 Scrollexpander
Thescrollexpanderisapositivedisplacementmachine.Itisinitiallyascrollcompressor,adaptedto
workinexpandermode. Theoriginalscrollcompressorisanoilfreeaircompressor,withaswept
volumeof148cmandaninternalbuiltinvolumeratioof4.1.
Operatingprincipleofascrollcompressor.
Ascrollcompressoriscomposedoftwospirals,afixedoneandamobileone. Themobilescroll
orbitseccentricallywithoutrotating,therebytrappingandcompressingpocketsoffluidbetweenthe
scrolls. As shown in Figure 5, in compressor mode, the fluid is trapped in two pockets at the
periphery of the two spirals. As the outer spiral orbits, the volume of the two trapped pockets
decreases, and the fluid is moved towards the center. Simultaneously, the pressure of the fluid
increases.Thecompressedfluidisfinallydischargedthroughthedischargeportlocatedatthecenter
ofthetwospirals.Inexpandermode,thefluidflowsfromthecentertotheperiphery.
Thescrollcompressorisverywidespreadinrefrigerationapplications.Ithasfewermovingpartsthan
reciprocatingcompressors,whichimprovesitsreliabilityandreducesthesoundpollution.
Scrollcompressorsareknowntobeverycompactandtooperateverysmoothly,astheirvibration
levelislimited.
3.2.1 Leakages
There aretwo types of leakage in a scrollcompressor:theflankleakageisduetotheclearance
betweentheflanksofthetwoscrolls,andtheradialleakageisduetheclearancebetweenthetipof
onescrollandtheplateoftheotherone[Chenetal,2002].
22
figure5:Workingprincipleofthescrollcompressor
In compression mode, the leakage reduces the volumetric efficiency and increases the specific
compressionwork,asthefluid,passingfromahighpressureregiontoalowpressureregion,needsto
berecompressed.
Inthesameway,theleakagereducestheoutputpowerofascrollmachineworkingasexpander,asthe
fluidflowsdirectlyfromthehighpressureregiontothelowpressureregionwithoutproducingany
usefulwork.
3.2.2 Conversionofthescrollcompressorintoanexpander.
Scrollcompressorsmaybelubricatedornot.Lubricationreducesthefrictionbetweenthetwoscrolls
andreducestheleakagearea. However,modifyingalubricatedscrollcompressorintoanexpander
raisesseveralproblems:
Therotationtakingplaceintheoppositedirection,theoilpumpmightnotworkanymoreifit
isdirectlyconnectedtotheshaftofthecompressor.Inthiscase,aseparateoilcircuithasto
beadded.
The compatibility of the working fluid with the lubricating oil is not guaranteed if the
compressorhasnotbeendesignedforORCfluids.
In ordertocope with these problemsthescrollmachineselectedforthistestbenchisanoilfree
compressor.
Anotherconsiderationtakenintoaccountforthechoiceofthecompressoristheinternalbuiltinratio:
ithasto beadaptedtotherange of pressureratiosimposedtotheexpander. Withrespecttothe
applicationstudiedhere,amachinewithahighinternalbuiltinvolumeratioisselected.
Aschematicviewofthescrollcompressorusedisgivenisfigure7.
23
figure6:Leakagesinascrollmachine
This machine presents another particularity. In order to reduce leakages, two types of seals are
embeddedinthecompressor:
Aninternalseal,nettledinthetipofbothscrolls(seefigure8).Itsroleistoreducetheradial
leakage.
Acircularperipheralseal,locatedonthefixedscroll,whoseroleistoisolatetheinsideofthe
scroll from the outside and to avoid eitherexternal infiltrationsor leakages to the outside
(dependingonthepressureintheperipheralpockets).(figure8).
Figure8showsaviewofthespiralsofthetwoscrolls.Theorbitingscrollissituatedontherightand
thefixscrollontheleft.Thesealsarevisibleatthetipofeachscroll.
Incompressormode,astheworkingfluidisair,theairtightnessbetweentheinsideandtheoutsideof
thecompressorinnotprimordial.Theperipheralsealismostlyusefultoavoidtheinfiltrationofdust
fromtheoutside.
24
figure7:Oilfreescrollaircompressor
figure8:Fixandorbitingscrolls
Inexpandermode,theworkingfluidisnotairanymoreandisathigherpressurethantheatmospheric
pressure.Anyleakagetotheoutsideofthescrollmachineisalossofworkingfluidforthecycleand
mustbeavoided.
In order to reduce this loss, an extra 0.8 mm layer (Reinzit 200 material) is inserted under the
peripheralsynthetictube.Thecontacteffortbetweenthefixedandorbitingscrollbodiesisincreased
(thiseffortcanbeadjustedbymeansofAllenscrews),andasealantadhesiveisappliedbetweenthe
fixedscrollbodyandthemobilescroll.Thedrawbackofthisoperationistheincreaseofthefriction
torquebetweenthetwoparts.
Another modification brought to the compressor is the obstruction of its aircooling circuit (air
channel along the finned external envelope of the scroll). Indeed, the cooling of the gas is
advantageousforacompression,butnotforanexpansion.
3.2.3 Justificationofthechoiceofthescrollexpander.
Afirstdistinctionmustbedonebetweentwomaintypesofexpanders:Theturbomachinesandthe
displacementtypemachines.
Thefirst type is the most used in traditional power plants. However,turbomachineshave several
drawbackswhenusedinlowtemperatureheatrecoveryapplications:
Theperformancesofmostrotarymachinesarerelatedtotheirperipheralspeed(ortipspeed)
U [m/s], rather than directly to the shaft speed. They have an optimal tip speed, usually
independentfromthemachinesize.Forscrollcompressorthisvaluerangestypicallyfrom1to
10m/s,whileforturbomachines,thisvalueiscloseto300m/s.
The tip speed is given by
U=
2nNR
60
, R being the radius of the rotary machine.
When used in smaller units, the turbomachines have a lower radius R, and their optimal
rotationalspeedisthereforeincreased. Thisveryhighshaftspeedcauseshighmechanical
stresses(e.gduetocentrifugalloading),bearingfrictionlosses,diminutionofthebearinglife,
necessityforhigherreductiongear,etc.Agoodexampleisthetypicalvehicleturbocharger:it
usuallyrunsat100000rpmorhigher.However,theturbochargerhasnomechanicalcoupling
to the engine or other mechanical unit, and the high speed in this case is not a problem.
In contrast, the tip speed of a displacement type machine is inherently lower, and the
drawbackspresentedabovedisappear.[Platell,1993]
Thepressureratioofasinglestageturbomachinehasalowvalue(typically1.5),whilethe
displacementmachinecanhaveashighpressureratiosasdesired.Thislattersolutionishence
preferredforthesinglestageexpansionusuallyusedinthelowtemperatureRankinecycle.
[Hung, 1996]
25
Volumetricmachinesaremuchmoreresistanttoaneventualliquidphaseinthefluidthan
turbines: their rugged design and their low rotational speed make them less sensitive to
contaminationbyliquiddroplets.
The scroll machine has been selected among all the displacement type machines for its reduced
numberofmovingparts,reliability,wideoutputpowerrange,andgoodavailability[Zanelli,1994].
Compared to the piston compressor, the scroll also shows theadvantage of not having admission
valves(thatworksperfectlyincompressormodeasacheckvalves,butthatneedstobesynchronized
inexpandermode,likeinaninternalcombustionengine).
However,thescrollexpanderisnottheonlysolutionsuitablefortheORC.Afewpapersalsopresent
theWankelengineandthescrewexpanderasappropriatetechnologiesfororganicRankinecycles
[Badr,1991;Persson,1994].
26
3.2.4 Pump
Thepumpisahydraulicdiaphragmtypepump.
The piston transmits the power to the fluid by the intermediary of a
flexiblemembrane(hydraulicdiaphragm).Thereisthereforenocontact
betweenthefluidandthepiston,whichallowstheuseofabrasiveor
corrosiveliquids.ThecheckvalvesarecomposedofanINOXballthat
blocksthefluidduringtheadmissionorcompressionstroke(seefigure
9).
Theflowratecanbeadjustedbyamanualgraduatedgearboxattherear
ofthepump.Itmodifiesthesweptvolumeandthustheamountoffluid
displacedateachstroke.
This type of pump is designed to ensure accuracy and is engineered for continuous metering
applicationsinthefieldofwatertreatmentandpetrochemicalengineering.
AspecialattentionispaidtothechoiceofthispumpbecauseithastoresisttotheR123working
fluid,whichisverycorrosivewithsyntheticmaterials.
27
figure9:Schematicviewofthepump
Pumpcharacteristics:
Strokefrequency 112strokes/minat
50Hz
Maximalflowrate 210l/hat50Hz
Maximalpressure 20bars
Motor 0.55kW
1400rpm
Pistondiameter 40mm
Diaphragmdiameter 120mm
Strokelength 25mm
Netweight 26kg
28
3.3 Condenser
Thecondenseriscomposedofplateheatexchangers.
Plateheatexchangersaremadeofastackofcorrugatedplatesgasketedorbrazedtogether(seeFigure
10).Theyhaveamajoradvantageoverconventionalheatexchangersinthat,thefluidsbeingspread
outallovertheplates,theyareexposedtoamuchlargersurfacearea.Theseexchangersaretherefore
verycompact.
Animportantparameterforplateheatexchangersisthechevronangle,correspondingtotheangleof
thecorrugationsontheplate.Thelowerthatangle,thehighertheheatexchangecoefficient,butalso
thehigherthepressuredrops.
Thetwofluidsexchangeheatincounterflowmode.However,severaldifferentconfigurationscanbe
consideredasshowninfigure11. Thetwopassmodeincreasestheexchangeareabutreducesthe
totalpassagearea.
Figure11:Oneortwopassconfiguration
Forthefirstsetoftests,thecondenserwascomposedoftwoorthreeplateheatexchangersassembled
inseriesandonepassmode.Thisconfigurationhavinghighpressuredrops,itwaschangedtotwo
29
figure10:Brazedplate
heatexchanger
heatexchangersassociatedinparallelforthesecondsetoftests(figure12).
figure12:Former(left)andnew(right)configurationofthecondenser
Characteristicsoftheheatexchangers:
Model A B
Numberofplates 75 35
Totalvolume[l] 3,75 1,75
Dimensions[mm] 250x112x189 250x112x93
Chevronangle 60 30
Working
temperatures
160Cto175C 160Cto
175C
Workingpressures 0to32bars 0to32bars
30
3.4 Evaporator
Theevaporatoriscomposedofthreeplateheatexchangers(Amodel).Thetwofirstsarefedwiththe
firsthotairsourceandthethirdoneisfedwiththesecondhotairsource.
3.5 Draincock
Inordertoremovetheairthatmayhaveleakedintothecyclewhileitisnotrunning,adrain
cockisinstalledattheexhaustofthepump.Theairispurgedeverytimebeforestartingthe
cycle.
3.6 Asynchronousmachine
Theelectricitygeneratingdeviceisa7.5kWasynchronousmotor. Itisconnectedtothetriphasic
electricitygridanditsoutputpowerismeasured.
Itscharacteristicsaregiveninthefollowingtable:
Poles 4
Power 7,5kW
Rotationspeed(motor) 1440rpm
Efficiency 87%
cosu
0,85
Ratedcurrent(380V) 15,4A
Weight 79kg
This motor is oversized for the present application, whose output power does not exceed 2 kW.
Consequently,itmightnothaveanoptimalefficiency.Itdoesnotconstituteaproblem,sincetheshaft
powerisalsomeasuredrightaftertheexpanderbymeansofatorquemeter.
31
3.7 Measurementdevices
3.7.1 Thermocouples
ThethermocouplesusedareTypeTthermocouples(Copper/Constantan),withatoleranceclass1.
They are suitable for measurements in the 200 to 400C range. Type T thermocouples have a
sensitivityof~43V/C
Theiraccuracyisgivenby:
Thermocouple Tolerance Classes as per IEC 584-2 (1989)
Reference point at 0%C
Type Tolerance class 1
(C)
Tolerance class 2
(C)
Tolerance class 3
(C)
Type T
Temperature -40<t<125 -40<t<133 -67<t<40
Tolerance 0,5 1 1
Temperature 125<t<350 133<t<350 -200<t<67
Tolerance 0,004|t| 0,0075|t| 0,015|t|
Thistoleranceisthetoleranceovertheabsolutetemperaturegivenbythethermocouplewiththe0C
reference. When usedtomeasureatemperaturedifferencebetweentwopointswithtwodifferent
thermocouplesandthesamereference,thetoleranceisevaluatedto0.2K
Allthethermocouplesareputinelbows,inordertoreducetheconductionparasiticheatfluxandthe
errorlinkedtoit.
The0Creferenceisgivenbyanicebath,whoseaccuracyisevaluatedto0.05K(Ngendakumana,
2006).
Thethermocouplesarenotcalibrated,thisoperationbeingquitecomplexbecauseitshighcostandof
thetoolsrequired(highresolutionmultimeter,benchmark,etc.).
32
3.7.2 Pressuremeasurement
Thepressuremeasurementsareperformedbydifferentialtransducersandpressuresensors.
Differentialtransducers
Twodifferentialtransducersareused:oneformeasuringthepressuredifferentialovertheexpander
(Sensotec010 bar)andthe other todeterminetheairflowratebymeansofthestandardnozzle
(Sensotec0500mbar).Accuracyofthesepressuretransducersis0.5%fullscale.
Pressuresensors
Pressuresensors(Keller010bar,accuracyof0.5%fullscale)areusedtomeasurethepressuresat
thepositionsindicatedonfigure13.
Theirreferencepressureistakeninaninternalchamberatanabsolutepressureof1bar.
3.7.3 Torquemeter
Inordertomeasuretheexpandermechanicalpower,atorquemeterisinstalledbetweentheexpander
and the asynchronous machine. The accuracy on the measurement of the torque is 0.1 N.m. A
proximitysensoralsomeasurestherotationalspeedofthetorquemetershaft.
33
figure13:Differentialpressuretransducer(pressure
differentialovertheexpander)andpressuresensors
(expandersupplyandexhaust)
Themechanicalpoweriscalculatedby:
n
cour
beingtheefficiencyofthetransmissionbelt.
3.7.4 Waterflowmeter
Thewaterflowmeterisasinglejetflowmeter,whoseworkingconditions
range from 0 to 2.5 m
3
/ h ,0to16barandamaximumtemperature of
120C.Theflowrateinformationissenttotheacquisitiondeviceunderthe
formofelectricalimpulses(4impulsesperliter).
Its accuracy is not given by the constructor, but its calibration in section
4.2.4.4showsanerrorofapproximately2to3%.
3.7.5 Wattmeter
Inordertomeasuretheamountofelectricalenergyproducedbytheasynchronousmachine,aGossen
wattmeterisused(maxpower:3kW,class:0.5)
34
figure14:Expander,Torquemeter
andAsynchronousmachine
3.7.6 Electricitycounter
An electricity counter is added in November to measure the pump electrical
consumption.
3.7.7 Acquisitioncards
ThevariablesarerecordedusingaSolartron34951AIMP/PCsystemwithasamplingrateof1s.The
measuredvalueisaDCvoltagecomingfromthepressureandpowersensors.Thesensitivityforthis
rangeofvoltageis 100uV andtheaccuracyisgivenby:(0.01%reading+0.01%fullscale).
TheerrorontheacquisitionofthethermocouplesvoltagesisgivenbytheIEC584standard:
Type MidRange Error FullRange Error Sensitivity
T 100to400 <0.3 200to400 <0.5 0.1
SincetheSolartron34951cardscannotrecordimpulseinputs,anadditional35952Acardisusedto
measurethevariablescomingfromthewaterflowmeter.
35
3.8 Coriolisflowmeter
TheCoriolisflowmeter,alsoknownasmassflowmeter,isadevicethatmeasureshowmuchfluidis
flowingthroughatube.Itdoesnotmeasurethevolumeofthefluidbuttheamountofmassflowing
throughthedevice.
The advantagesof theCoriolisflow meterovertraditionalflowratemeasurementdevices arethe
following:
Highlyaccurateandrepeatableflowratemeasurement(typicallyfrom0.05to0.5%).
Autocalibration:nocalibrationorzeropointresearchisneeded,aseverythingisdonebythe
transmitter.
Reducedmaintenance:sincetherearenomovingparts,thedurabilityofthedeviceisvery
good,itonlyneedstobecleanedfromtimetotime.
Abilitytomeasurethefluiddensityortemperature,inadditiontotheflowrate.
Therearetwobasicconfigurations:thecurvedtubeflowmeterandthestraighttubeflowmeter.
Thestraighttubeflowmeterismoreaccurate,andremainsaccurateevenunderchangingdensitiesof
processfluid,externalvibrationandmountingstresses.Itisalsomoreexpensive.
SincetheCoriolisflowmeterusedonthetestbenchisacurvedtubeflowmeter,onlythistypeof
flowmeterwillbediscussedinthiswork.
36
figure15:ViewoftheCoriolis
flowmeter
3.8.1 Generalprincipleofthecurvedtubeflowmeter
Toillustratetheprinciple,arotatingdispositionoftubesasshownonfigure16canbeimagined.
figure16:WorkingprincipleoftheCoriolisflowmeter
Sincetheflowhasaradialcomponent,aCoriolisaccelerationappearsinthefoursidelegsofthe
tubes.Theforceisgivenby:
F
c
=2 o
v
TheresultoftheCorioliseffectisatorsionofthedeviceasshownontherightsideoffigure16.This
torsionhastobemeasuredtocalculatethefluidmassvelocity
v .
3.8.2 Workingprincipleofthemeasurementdevice
In manufactured flow meters, there is no rotation, but two tubes oscillating in opposition to one
anotherbytheeffectofamagnetandcoilassembly,asshownonfigure17.Pickoffcoilsaremounted
onthesidelegsofoneflowtube,andmagnetsaremountedonthesidelegsoftheopposingflow
tube.
37
figure17:Oscillatinglegsofthe
flowmeter
Whennofluidisflowingthrough,thetwolegsoscillatesynchronically.Asthefluidstartstoflowin
thetubes,theCorioliseffectcausesaphasedifferencebetweenthetwoparts,asshownonfigure18
(viewfrombelow).
Thephasedifferencebetweenthetwocoilscanbemeasured. Figure19showsthevoltageoneach
device,inthecaseofanullandnonnullflowrate.
figure19:Phasedifferenceatthetwocoils
The AT betweenthetwophasescanbecorrelatedveryaccuratelytothemassflowrate.
38
figure18:Corioliseffectontheoscillatingtubesoftheflowmeter
3.8.3 Devicecharacteristics
Thecharacteristicsofthedevicearethefollowing:
Maximum flow range 0 to 2180 kg/h (0 to 0.6
kg/s)
Accuracy liquid
Seefigure20
Repeatability liquid Up to +/-0,05%of rate
Accuracy gas Up to +/-0,35% of rate
Repeatability gas Up to +/-0,20% of rate
Process temperature -240 up to 204C
Accuracy density Standard 0,0005 g/cc
Repeatability density +/-0,0002 g/cc
Maximumoperatingpressure 100bar
Figure20showsthatagoodaccuracyisachievedalongthemeasurementrange.Therangeinwhich
theflowmeterisusedonthistestbenchisindicatedinblue.
39
figure20:Coriolisflowmeteraccuracyalongitsmeasurementrange
Table3showsthepressuredropsofthedeviceasafunctionoftheflowrate.Theyremainverylowin
therangeofflowratesusedinthescopeofthiswork.
Turndownfrom
maximumflowrate
Pressuredrop
[bar]
40:1 0.01
15:1 0.03
2:1 0.98
Table3:Pressuredropintheflowmeter
TheCoriolissensorisconnectedwitha4wirescabletoatransmitter,whosefunctionisto:
Transform the information received from the sensor into an analogic or impulse output
readablebyanyacquisitiondevice.
Performanautomaticcalibrationandzeroresearch
Transmitsimultaneouslyflowrate,temperatureanddensityinformationaboutthefluid.
Perform a verification of the Coriolis meter (erosion, corrosion, etc) by comparing its
mechanicalandelectricalpropertiesagainstabaselineestablishedatthefactory.
40
SylvainQuoilin Chapter4:Descriptionoftheexperiments
Chapter4:
Descriptionoftheexperiments
4 Descriptionoftheexperiments
41
This chapter describes the 3 series of tests performed on the test bench. For each set of test, a
summaryandacriticalanalysisoftheresultsisproposed.Theevolutionandtheimprovementsofthe
cycleduringthosetestsaredescribed.
4.1 Firstsetoftests:
Afirstsetof13testsiscarriedoutinSeptemberOctober2006.
4.1.1 Configurationofthetestbench:
For this first series of tests, the condenser is composed of three plate heat exchangers
associatedinseries.
TheCoriolisflowmeterisnotyesinstalled.
Theevaporatorissetupasdescribedinsection3.4.
Thewaterflowmeterisnotconnectedtoanyacquisitiondevice. Theinstantaneouswater
flowrateisthusnotknown.
4.1.2 Descriptionofthetests:
Theexpanderissetupandadaptedasdescribedinchapter3.
Apreliminarytestiscarriedoutinordertoevaluatetheperipheralleakagearea:theexpanderis
connected toa pressurized air tank, whosepressureisinitially4bars. Thetimeevolution ofthe
pressureinsidethetankismeasuredandallowstheevaluationoftheleakagefromtheinsideofthe
expanderthroughtheperipheralseal.
Themaindifficultyencounteredinthisfirstsetoftestsisthelackofflowmeterontherefrigerant
side. Theflowrateiscalculatedusingtheheatbalanceoverthecondenser. Inordertoachievean
acceptableaccuracyonthisflowrate,the AT onthewatersidecannotbetoosmallandthusa
maximumwaterflowratecannotbesurpassed. Thislimitedflowratereducestheperformancesof
theheatexchangeinthecondenserandtheglobalefficiencyofthecycle.
Sincethepressureinthewaternetworkcanvary,thewaterflowratehastobemeasuredmanuallyas
oftenaspossible.
42
4.1.3 Startingprocedure.
Allthemeasurementsdevicesareturnedon.Theicebathreferenceforthethermocouplesis
set.
Theacquisitionofthedataisstarted.
Thewatervalveisopenedtofeedthecondenserwithwater.
Thetwovalvesisolatingtheexpanderareclosed,thebypassvalveisopened.
Thepumpisswitchedon.Theworkingfluidcirculatesinthecycleafewminutesbeforeany
furtheroperation.
Thehotairgeneratorisactivated,theairflowsthroughtheplateexchangers
Asthepressureincreasesinthecycle,theairthatmayhaveenteredthecycleispurgedatthe
draincock.
When the refrigerant temperature reaches approximately 100 C at the exhaust of the
evaporator,theexpanderisstarted:thisisachievedbyopeningprogressivelythetwoisolating
valvesattheinletandattheoutletoftheexpander,andbyslowlyclosingthebypassvalve.
Whenthespeedoftheasynchronousmachinereachesapproximately1500rpm,theswitchis
turnedontoconnectthemachinetotheelectricgrid.
Thisprocedureremainsunchangedforthethreesetsoftests.
4.1.4 Parametersandtestresults
Forthisfirstseriesoftests,threemainparametersofthecyclearemodified:therefrigerantflowrate,
thewaterflowrateatthecondenserandthehotairtemperature.
Theexpanderrotationspeedissettoaconstantvalueof2200rpm(imposedbytheasynchronous
machine)
Thehotairflowrateiskeptalmostconstantaroundavalueof85g/s.
Theconditionsinwhichthetestsareperformedarethefollowing:
Parameter Minimumvalue Maximumvalue
Firsthotairsource
temperature
176.5C 188.5C
Secondhotairsource
temperature
135.2C 160.4C
X
pp
70% 90%
Condenserwaterflowrate 0.21l/s 0.59l/s
43
Mainsresults:
Minimumvalue Maximumvalue
Pressureratioattheexpander 3.2 3.8
Refrigerantflowrate 61.3g/s 73.7g/s
Outputshaftpower 0.71kW 1.07kW
Cycleefficiency 4.0% 5.4%
Expanderisentropic
effectiveness
53% 62%
Carnotefficiency 29.5% 33.7%
Exergeticefficiency 13.4% 16.1%
Theoutputshaftpoweriscalculatedby:
W
sh ,exp
=

W
sh , tqm
n
transmission
n
transmission
=0.95
Where
W
sh ,tqm
isthemeasuredpoweratthetorquemeter
n
transmission
isthetransmissionbeltefficiency,evaluatedto0.95
Andthecycleefficiencyiscalculatedby:
n
cycle
=

W
sh, exp

W
pp
Q
a, hx123
Where
W
pp
isthepumpconsumption
Q
a, hx123
istheairsideheattransferoverthewholeevaporator
Acompletetableofallthetestsperformedandoftheperformancesreachedisavailableinappendix
2.
Amoreexhaustivedescriptionandcriticalanalysisofthosetestsisproposedinthenextsection.
44
4.2 Secondsetoftests
Asecondseriesof19testsisperformedinNovember2006. Thepurposeistovalidatethemodels
overawiderrangeofparameters,andtoevaluatethepossibilityofimprovingtheperformancesofthe
cycle.
4.2.1 Modificationsonthetestbench
Inordertodecreasethepressuredropsonthecondenser,thelatterismodified. Insteadof
threeplateexchangersinseries,twoheatexchangersareusedinparallel,asshowninfigure
12.
Inordertoimprovethemeasurementofthewaterflowrateatthecondenser(whichisofkey
importance), the water is deviated to a water tank mounted on a very accurate balance
(sensitivity100gr).Thetimetofillthetankwith100kgofwaterismeasuredandtheflow
rateisdeduced.Thesemeasurementsalsopermitthecalibrationofthewaterflowmeter.Itis
foundthattheflowrategivenbythewaterflowmeterhastobecorrectedby2%.
Anelectricitycounterisaddedtomeasurethepumpelectricalconsumption.
4.2.2 Descriptionofthetests
Moreparametersaremodifiedforthissecondseriesoftests.Inadditiontotherefrigerantflowrate,
waterflowrateandhotairsourcetemperature,theexpanderrotationspeedandthehotairflowrate
aremodifiedaswell.
Themodificationoftheexpanderrotationalspeedisachievedbymodifyingthediameterofthepulley
ontheexpandershaft.Threediametersareused:180mm,140mm,and112mm
Theexpanderspeedisproportionaltothediameterofthepulley.
Thismodificationcanobviouslynotbeperformedwhilethecycleisworking.Acompletestoppingof
45
Figure21:3differentpulleysusedtomodifytheexpander
rotationalspeed
theexpanderisnecessarybeforechangingthepulley.Afrequencyinvertercouldbeusedtovarythe
rotationalspeedinasimplerway.
Theconditionsinwhichthetestsareperformedarethefollowing:
Firsthotairsource
temperature
172.6C 182.3C
Secondhotairsource
temperature
135.9C 179.1C
X
pp
50% 100%
Expanderrotationspeed 1776rpm 2664rpm
Airflowrate 0.084kg/s 0.113kg/s
Mainsresults:
Refrigerantflowrate 44g/s 52g/s
Cycleefficiency 4% 6.1%
Expanderisentropic
effectiveness
53% 69%
Carnotefficiency 30% 36.8%
Theexhaustivedescriptionofthetestsparametersandresultsisavailableinappendix3.
Althoughhighervaluesoftheshaftpowerwereexpectedforthissecondsetoftests,thecontraryis
observed:theoutputshaftpowerisdecreased.Thisismainlyduetoalowerworkingfluidflowrate:
itwillbeshowninsection4.1.2thatthepumpisnotabletoimposetheflowrateforthosetests,and
lessheatisrecoveredfromthehotairsource.
Nevertheless,theefficiencyofthecycleisincreasedfromamaximumof5.4%inthefirsttestsupto
6.1%.
46
4.3 Criticalanalysisofthetwofirstsetsofmeasurements
Thisanalysisisbasedonthe32testsperformedfromOctober13
th
toNovember21
st
anddescribed
above.Thefirst13testswerealreadytreatedsuccessivelybyVincentLemortandCristianCuevasin
three internal reports. The purpose ofthischapter isto give an explanation tothe incoherencies
detectedinthosetestsandtoextendthisanalysistothe19furthertests.
4.3.1 Airflowrate
Theredundanciesinthecalculationandthemeasurementoftheairflowratearecomparedtogetherin
ordertodeterminethemostaccuratemethod:
Theairflowrateismeasuredwithanozzleattheexhaustofthethirdheatexchanger. The
nozzleiscalibratedandtheairflowrateiscalculatedwiththehelpoftheISO5167standard.
Asecondmethodtoevaluatethisairflowrateistheconstructordataofthehotairgenerator:
theairflowratecanbededucedfromthesettingsofthishotairgenerator.
Thethirdmethodistheheatbalanceontheevaporator.Thismethodpresentsthedrawbackof
high uncertainties on the refrigerant flow rates (discussed in the next section) and the
uncertaintiesonthemeasuredtemperatures.
Theairsflowratesmeasuredatthenozzleiscomparedtotheonecalculatedwiththeconstructordata
infigure22:
Themeasurementuncertaintiesareindicatedbytheverticallines.
Amoreorlessconstanterrorof10%appearsbetweenthetwoairflowrates,thatcannotbeexplained
47
Figure22:NozzleairflowratevsConstructorairflowrate
bytheuncertaintiesonthemeasurementsatthenozzle. The(pessimistic)uncertaintiestakeninto
accountare:
0.3Konthenozzleairsupplytemperature(noncalibratedthermocouple)
Ahypothetical1%erroronthepressuretransducers.
Inordertodeterminewhichmethodisthebesttomeasuretheairflowrate,theheatfluxesoverthe
evaporatorarecalculatedandcomparedwitheachmethod:
Threedifferentheattransfersarecompared:
Overthefirstexchanger(hx1).Inthiscase,onlythe13firsttestsaretakenintoaccount,the
fluidbeingintwophasestateattheexhaustofhx1intheothertests.(cfrfigure26)
Overthetwofirstexchangers(hx12).Inthiscase,ontheotherhand,onlythe13lastpointsare
takenintoaccount,inordertobesurethatthefluidisinsuperheatedvaporstateattheexhaust
ofthesecondexchanger.
Overthewholeevaporator(hx123)
Theyarecalculatedasfollows:
48
Iftheairflowrateiscalculatedwiththe
constructordata:
Iftheairflowisthenozzlemeasuredairflow:
Figure23:Heatfluxthroughexchangerhx1,hx12,hx123,refrigerantsidevsairside
49
Figure23showsthattheagreementbetweenthevaluesoftheheatfluxisbetterinthefirstcase.In
thesecondcase,theheatfluxcalculatedattherefrigerantsideissystematicallysuperior,whichwould
mean a negative ambient loss at the evaporator. On the other hand, when calculated with the
constructorairflowrate,theheatfluxesshowagoodagreementandalmostallthedifferencescanbe
explainedbythemeasurementuncertainties.
Therefore,theairflowratetakenintoaccountinthisworkistheconstructorflowrate.
50
4.1.2 Refrigerantflowrate
Therefrigerantflowrateisimposedbythepump.Thefluidisinliquidstateattheexhaustofthe
condenserandisincompressible.Therefore,itsflowratedependsdirectlyonthevolumetricflowrate
ofthepump.Sincethepumpisameteringpump,thisvolumetricflowratecanbeadjustedwithan
acceptableaccuracy.
Therefrigerantflowrateisnotdirectlymeasured. Itiscalculatedfromtheheatbalanceoverthe
condenser. Itisthus necessarytomeasurethewaterflowrateandthetemperatureswithagood
accuracy.
Inthiswork,theuncertaintyonthewaterflowrateissetto1%ofthemeasuredvalue.
Theuncertaintyonthethermocouplesissetto0,3K.
The refrigerant flow rates given by the pump model and the heat balance on the condenser are
comparedinfigure24:
Inthisfigure,Istatesforthetestnumber,inchronologicalorder. Itisputonthexaxistogivea
descriptionofthechronologicalevolutionoftheflowrate.
Itisobviousthattheagreementbetweenthetwovaluesisonlyacceptableatthebeginning:Theflow
rategivenbythepump(inblue)isclosetotheflowratecalculatedatthecondenser.Butaroundtest
1617,thedifferencesbetweenthetwoflowratescannotbeexplainedbytheuncertaintiesanymore.
Fromtest21,thisdifferenceincreasesdramatically:thecalculatedflowrateevendecreaseswhilethe
pumpflowrateincreases.Thepumpinthiscasedoesn'tactuateasexpected:itdoesn'timposethe
51
Figure24:Condensercalculatedfluidflowrateandpumpflowratevstestnumber
flowrateanymore.
Thepossibleoriginsfortheunexpectedflowrateare:
Apumpmalfunction
Pumpcavitationduringtheaspirationofthefluid
Presenceofairinthecycle:SincethevaporpressureoftheHCFC123at20Cis0.76bar,
thereismostlikelyairenteringthroughtheleaksinthecyclewhenitisstopped. Thisair
shouldberemovedbythedraincockbutthereisapossibilitythatsomeairremainsinthe
cycle.
Ifthereisreallypresenceofcavitation,itmightbecorrelatedtoadiminutionofthesubcoolingatthe
exhaustofthecondenser.Figure25showsthatthisisnotthecaseandconsequentlythishypothesis
cannotbeconsideredasrelevant.
Thevalueofthesubcoolingisgivenby:
T
ex,cd
= T
sat
( fluid$ , P =p
f,ex,cd
) T
f,ex,cd
Sincethesubcoolingcannotexplainthelowflowrate,anotherexplanationhastobeinvestigated.This
isdoneindetailinthenextchapters.
52
Figure25:Subcoolingatcondenserexhaustvstestnumber
4.1.3 Refrigerantlosses
Itseemsthat,duringtheperiodofthetests,somerefrigerantlossesoccurred.Thiscanbecausedby
leaksattheexpanderlevel(whichisnothermetic)orelsewhereinthecycle.
Figure26showsthedifferencebetweentheeffectivetemperatureandthesaturationtemperatureatthe
exhaustofeachheatexchanger:
Thevalueofthesuperheatingisgivenby:
T
ex,hx1
= T
f,ex,hx1
T
sat
( fluid$ , P =p
f,ex,hx1
)

T
ex,hx12
= T
f,ex,hx12
T
sat
( fluid$ , P =P
f,ex,hx12
)

T
ex,hx123
= T
f,ex,ac
T
sat
( fluid$ , P =p
f,ex,ac
)
Figure26showsthat:
Attheexhaustofthethirdexchangerhx123,thevaporwassuperheatedforallthetests.
Attheexhaustofthesecondexchangerhx12,thefluidisintwophasestateuntiltest18andin
superheatedvaporstateafter.
Attheexhaustofthefirstexchangerhx1,thefluidisinliquidstateuntiltest20,andintwo
phasestatefortheothertests.
53
Figure26:Superheatingattheexhaustofeachexchangervstestnumber
Itisobviousthat,duringtheperiodoftheexperiments,thelevelofliquidintheevaporatorlowered.
Itisverylikelythatthisloweringwascausedbyrefrigerantlosses.Averyinterestingexperimentis
theonecorrespondingtothepointsi=20,21and22.Itwasperformedonthe8
th
ofNovemberwitha
fasterexpanderrotatingspeedthanfortheothertests.
Figure27showstheevolutionofthesupplyandexhaustexpandertemperatureswithtime.Thetwo
verticallinescorrespondtoamodificationof
X
pp
successivelyfrom60to70andthen80%.After
a change of
X
pp
, the temperatures should lower asymptotically to the new steady state
temperatures. Figure 27 shows that the temperature lowers as expected but then rises without
stabilizing. Thisismostlikelyduetotherefrigerantlosses,increasedbythehighvelocityofthe
expanderinthistest.
Thehypothesisofrefrigerantlossesiscorroboratedbytheemptyingofthecyclebeforethethirdsetof
tests.Indeed,only5.5kgofrefrigerantarerecovered,although10kgwereintroducedinthecyclein
September2006.
54
Figure27:Evolutionofexpandertemperaturesforpointsn20,21and22
4.2 Thirdsetoftests
This third set of tests takes place in AprilMay 2007. The test bench is modified and improved
accordingtotherecommendationsformulatedaftertheanalysisofthetwofirstsetsoftests.
4.2.2 Generalpurposeofthenewmeasurements:
Validationofthemodels :
Thefirstmodelswerevalidatedon13points,butwithanalmostconstantworkingfluidflow
rate,andsmallvariationsoftheotherparameters.Itisnecessarytovalidatethesemodelona
widerrangeofparameters.
Moreaccuratemeasurementswillalsoimprovethequalityofthosemodels.Forexample,in
themodelingoftheevaporator,afictitiousheatsourcehadtobeaddedbeforetheaftercooler
tocompensate the measured differencebetweentheheatfluxesontheairsideandonthe
refrigerantside.
Demonstration of the possibility to get more shaft power from the Rankine Cycle :
InJanuary2006,CristianCuevasshowedthatitwaspossibletoincreasethecycleefficiency
from5toalmost12%.
Hisrecommendationstoachievethisimprovementare:
Reductionofthecondenserpressuredropsandsubcooling.
Oversizingoftheevaporator.
Enhancementoftheexpandervolumeratio,andreductionofthefrictiontorqueandofthe
leakagearea
Anotherveryimportantparameterthatneedstobecontrolledandoptimizedistherefrigerant
flowrate.Asexplainedabove,ithasbeenverydifficultduringtheteststohaveagoodcontrol
of this parameter : the pump was not capable of imposing the flow rate and the latter
decreaseddramaticallyduringtheperiodoftheexperiments,decreasinginthesameprocess
theshaftpower.
Onthetestbench,onlyalimitednumberofphysicalparametersisaccessibleinorderto
achievethoseimprovements:
Thesubcoolingcanbecontrolledbymodifyingtherefrigerantcharge.
Theflowratecanandshouldbecontrolledbythepump.
55
4.2.3 Modificationsonthetestbench
The experience of the first tests permits to determine the weakest points of the test bench. The
uncertaintiesonthefluidflowratearehighlyprejudicialandresultindifficultiesinmodelingthe
differentelementsofthecycle.Theambientlossesaredetrimentaltotheheatbalances,andnocontrol
oftherefrigerantchargeispossible.
Thefollowingmodificationsareperformedinordertoimprovethebehaviorofthetestbench:
Theheatexchangers,thepipesandtheexpanderarecompletelyinsulatedinordertoreduce
theambientlosses. Thisoperationisperformedtoimprovethequalityoftheheatbalances
overeachelement,andtoimprovetheefficiencyofthecyclebyreducingtheheatlosses.The
insulatingmaterialisanelastomericrubberinsulation,withathicknessof13mmanda0C
conductivityof0.036W/(mK).
Thewaterflowmeterisconnectedtoanimpulsiondataacquisitionsystemtomeasurethe
instantaneouswaterflowrate(whichisnotconstantbecauseofslightvariationsinthewater
supplypressure).
InstallationoftheCoriolisflowmeter,inordertoachieveanaccuratemeasurementofthe
refrigerantflowrate.Thisparameterisindeedprimordial,asitisconsideredinthemodeling
ofeachelement.
56
Figure28:Viewofthe
insulatedexpander
Installationofanexpansiontankattheexhaustofthepump.Thistankhastwogoals:
Thefirstgoalistoabsorbthevariationsoftherefrigerantflow
rate due to the volumetric piston pump in order to make its
measurementpossible.
Thismeasurementisindeedlimitedbytheacquisitionfrequency
ofthesystem:
Thepumpstrokeshaveafrequencyof1,87Hz
Theacquisitionfrequencyofthecoriolisflow meteris
150Hz
Thefrequencyoftheacquisitioncardis1Hz
TheCoriolisflowmeterishenceperfectlycapableofmeasuring
thevariationsoftheflowrateattheexhaustofthepump,but
theacquisitionsystemhasatoolowacquisitionfrequencythat
willresultinirregularvariationsinthemeasuredflowrate.The
role of the expansion tank is to absorb the variations and to
smooththeflowratecurve.
Thesecondroleofthetankistoallowthemodificationofthe
refrigerant charge in the cycle : the tank is hanged to a
piezoelectric force transducer that measures the weight of the
tankandthustheamountofrefrigerantinthedevice.
Theexpansion tank is halffilledwithrefrigerantandtheotherhalfisfilledwithnitrogen
underpressure.Thelowerendisconnectedtotheexhaustofthepumpandtheupperendis
connectedtoagascylinderthatallowstheadjustmentofthepressure(seefigure29).Ifmore
fluid is required in the cycle, the nitrogen pressure is increased and the level in the tank
lowers.
Avalveisaddedinthecircuitdownstreamtotheexpansiontank,toincreasethepressuredrop
when required. Increasing thispressuredropwillincreasetheabsorptionoftheflow rate
variationsbytheexpansiontank.
Asimulationoftheexpansiontankisproposedinsection4.2.4.2.
57
4.2.4 Preparationofthenewtests.
4.2.4.1 Refrigerantchargecalculation
Asthecircuitiscompletelyemptiedandmodifiedafterthesecondsetoftests,anewevaluationofthe
refrigerantchargeisneeded.
Thevolumeofeachelementhastobecalculated.Theregionsofthecyclewherethefluidisinvapor
statearenottakenintoaccount,thedensityofthevaporbeingnegligiblecomparedtothatofthe
liquid.
Thevolumestakenintoaccountarethefollowing:
1. Volumeofthepump
It is evaluated by the product between the diaphragm area and the stroke length :
V
pump
=
nd
diaphr
2
4
l=
n0.120
2
4
0.025=2.8310
4
m
3
2. Volumeofthepiping.
Thetubesconcernedarethetubesbetweentheoutletofthecondenserandtheinletofthe
evaporator. Their total lengthisapproximately4mandtheirinternaldiameteris 16 mm.
V
tubes
=
n0.016
2
4
4=8.0410
4
m
3
3. Volumeoftheevaporator:
The total volume of the heat exchangers is given in section 3.3. The working fluid only
occupies half of that volume, the other half being filled with hot air.
It is assumed that
1
2
of the evaporator volume is filled with liquid. This assumption
arbitrary,thepurpose beingonlytoputenoughrefrigerantinordertofillthecircuit. The
adjustmentoftherefrigerantchargetogetthedesiredlevelisdonebyaddingorremoving
refrigerantduringthetest.
Thevolumeisthusexpressedby:
V
ev
=
1
2
Vol
tot , hx1
2
+
Vol
tot , hx2
2
+
Vol
tot , hx3
2
)=2.31l =2.3110
3
m
3
4. Volumeofthecondenser:
Theapproachisthesamethanfortheevaporator.Inthiscase,itisassumedthat
3
4
ofthe
condenserisfilledwithliquid.
V
cond
=
3
4
Vol
hx, cd
2
+
Vol
hx, cd
2
)=2.8110
3
m
3
Theamountofrefrigerantthatneedstobeintroducedinthecircuitisgivenby:
m
r123
=
r123
V
pump
+V
tubes
+V
ev
+V
cond
)=9.2kg
Therefrigerantcharge to introduce in thecircuitisthusevaluatedto 10 kg, inorderto take into
accounttherefrigerantmassinvaporstateandthepossiblerefrigerantlossesduringthetransfer.
58
4.2.4.2 Expansiondevicecalculation.
Asimulationoftheexpansiontankisperformedinordertocalculatetheminimumvolumerequired
togetagoodsmoothingoftheflowratecurve.
The instantaneous theoretical flow rate

V of the pump is a
semisinusoidcalculatedonthebasisofthedisplacedvolume
andonthestrokefrequency.Itissplitintotwoflows

V
1
and
V
2
:
Theflowrate

V
1
enteringtheexpansiontankislinkedbythe
liquidlevelinthetankbytherelation:
Theflowrate

V
2
leavingthesystemdependsonthehydraulic
resistanceofthevalve:
Itcanbeconsideredthattheheatexchangebetweenthetankand
theambientisnegligible. Theexpansion/compressionofthe
gasissupposedreversible,adiabatic,andthusisentropic:
p and v beingthereferencepressureandspecificvolumecorrespondingtoan8barpressureand
atemperatureof25C.
Thehydraulicresistanceoftheliquidlinepumpevaporatorisidentifiedfromthefirstsetoftests.Its
approximate value is
R
openvalve
=0.810
13 Pas
2
m
6 . Theeffectofclosingthevalveismodeledby
multiplyingthisresistancebyafactorK.K=1correspondstoawideopenvalve.
ThedifferentialequationissolvedusingEESona5secondsperiod.
Figure30showstheresultsofthesimulationforthefollowingconditions:
Tankvolume:5liters
Thetankisinitiallyemptyandatapressureof1bar
Thepressuredropregulationvalveiswideopen
When the tank is connected to the circuit, it is filled by the refrigerant and the pressure is
increasedupto8bars.Thesimulationisstartedwhenthispressureisreached.
59
Figure29:Expansiontank
Figures30showsthatthestabilizationoftheflowrateisquitegood:theflowraterangesbetween
valuesof 0.000045and 0.000075 m
3
/ s insteadofa00.00015 m
3
/ s rangeofvalues without
expansion tank.
Startingfromapressureof1barcorrespondstoaconservativehypothesis:ifthepressuresuddenly
fallsinthecircuit(forinstanceifthepumpstopsworking),nogaswillflowfromthetanktothe
circuit, since the pressure of the cycle neverfallsbelow1bar. Itisthusmuchsaferbuthasthe
drawbackofasmallnitrogenvolumeat8bars.
Since thetank isalso usedformodifyingtherefrigerantcharge,thissafeconditionwill not be
respectedinthetests.Itisindeednecessarytointroduceahighquantityofpressurizednitrogeninthe
tankinordertoadjustthecharge.
Theinfluenceofthepressuredropatthevalvelevelon

V
2
isgiveninfigure31:
60
Figure30:Flowrateatthepumpexhaust,inthetankandintheflow
meter
Asexpected,thecurveissmootherwhenthevalveisclosed(whenKincreases).ForK=4andK=
8,thediagramshowsthat

V
2
istoolowatthebeginningofthesimulation.Itisexplainedbythe
increaseofthepressureinthetankduetohighpressuredrops:apartoftheflowisusedtofillthe
tankandincreasethepressure.
Amaximumpressureof10.8barisreachedattheexhaustofthepumpinthecaseK=8.Thispressure
canperfectlybehandledbythepump,whosemaximumoutputpressureis20bar.
Itisthusconcludedthata5litertankissufficienttostabilizetheflowrateinthecoriolisflowmeter.
Thetankfinallyinstalledonthetestbenchhasatotalvolumeof6.2l,connectionpipeincluded.
61
Figure31:InfluenceofthevalveopeningonV2
4.2.4.3 Checkoutofthepumpbehaviorbymakingitruninstandalonemode
Inordertounderstandthetoolowflowratedetectedinthefirstseriesoftests,thepossibilityofa
pumpmalfunctionisinvestigated.Theexpanderisisolatedfromthecircuit,theheatsourceandthe
heatsinkareactivatedandthepumpisrun.Theoutletpressureofthepumpisadaptedbymodifying
theopeningofthevalveattheCoriolisflowmetersupply.
Theresultsofthistestaregiveninfigure32.
Thepumpseemstodelivertheexpectedquantityoffluidformostofthetests,whateveritsexhaust
pressure.
Animportantobservationismadeduringthosetests:forthepointcircledinred,thefluidisintwo
phasestateattheinletofthepump(presenceofbubblesdetectedintheliquidindicator).Itexplains
thattheflowratedeliveredbythepumpislowerinthiscase.Inordertomakethistwophasestate
disappear,theairflowrateoftheheatsourcehastobeincreased.Thesecondgreenpointisthen
obtained.
Thistwophasestatemostlikelyexplainsthelowflowratethatwasobtainedduringthesecondsetof
tests.
62
Figure32:Measuredvstheoreticalflowrateatthepump
4.2.4.4 Calibrationofthewaterflowmeter.
AfirstcalibrationofthewaterflowmeterwasalreadyperformedinOctober2006.
Sincethewatercounterisnowconnectedtoanacquisitionsystem,asecondcalibrationisrealized.
Thewaterisdeviatedtoaweightedtankandthetimetofillitwith100kgofwaterismeasured.The
resultsofthecalibrationaregiveninfigure33.Theyshowthatthereisnozeroerror,buttheflow
ratehastobecorrectedby2,4%.
63
Figure33:Calibrationofthewatercounter
4.3 Descriptionofthetests.
Forthisthirdseriesoftests,39pointsarerealizedinaoneweekperiod.Themainimprovementon
thisnewtestbenchistheabilitytochangetherefrigerantcharge.Thisaspectwillbeexploitedin
ordertounderstandhowthecyclereactstoarefrigerantchargemodification.
Allthetestsarecarefullydefinedbeforestartingthetestbenchforthefirsttimeinordertoavoidany
improvisationduringthetestperiod.Theirpurposeistoevaluatethecycleresponseonawiderange
ofparametersandonparametersthatwerenottakenintoaccountintheprevioustests.
Thetestperiodismadeasshortaspossibleinordertominimizetherefrigerantlosses.
Thetestsareperformedasfollows:
1
st
May2007: Thepurposeofthistestistoevaluatethegeneralbehaviorofthetestbenchonafew
veryclassicalparameters.Theairflowrateissetto3differentvalues.Foreachair
flowratevalue,therefrigerantflowrateismodifiedto3values.
3
rd
May2007: Afirstseriesofpointsisrealizedmodifyingthehotairsourcetemperatures.These
temperaturesaresetto2mainvalues.Foreachvaluetherefrigerantflowrateis
modified.(pointsAtoH)
Asecondseriesisrealizedbymodifyingtherefrigerantchargeandtherefrigerant
flowrate.(pointsItoL)
pointN:attempttomaximizetheoutputpower
pointOandP:attempttomaximizethecycleefficiencybyreducingthe
superheatingattheevaporatorexhaustandthesubcoolingatthecondenserexhaust.
4
th
May2007: Therefrigerantchargeismodifiedataconstantrefrigerantflowratefromits
minimumvalue(twophasestateatthepumpsupply)uptoitsmaximumvalue(two
phasestateattheexpandersupply).
5
th
May2007: Theexpanderrotationspeedissetto3differentvalues.Foreachrotationalspeed
value,thefluidflowrateissetto3values,andforeachflowrate,therefrigerant
chargeissettothepointofmaximumefficiency(subcoolingreducedatits
minimum).
64
Rangeofvaluesfortheparameters:
Firsthotairsource
temperature
176.8C 191.8C
Secondhotairsource
temperature
120.1C 173.8C
X
pp
50% 100%
Expanderrotationspeed 1771rpm 2660rpm
Airflowrate 0.071kg/s 0.90kg/s
Refrigerantcharge 10kg 16kg
Mainresults:
Refrigerantflowrate 45g/s 86g/s
Cycleefficiency 2,6% 7,4%
Expanderisentropic
effectiveness
43% 68%
Carnotefficiency 27.7% 36.4%
Themaximumoutputpowerincreasedbyalmost70%inthislastsetoftestscomparedtothe
previoustests(from1.1kWto1.8kW).Itisimportanttokeepinmindthatthepumpconsumption
needstobedeductedfromthisvalueinordertogetthenetoutputpower.
Themaximumcycleefficiency(takingintoaccountthepumpconsumption)isincreasedfrom6%to
7.4%.
Acompletedescriptionoftheresultsforthisthirdsetoftestsisavailableinappendix4.
65
Thesegoodresultshavetwomainexplanations:
Theinsulationofthecircuitdecreasedtheheatlossesintheevaporatorandintheexpander.
Thepossibilityofadjustingtherefrigerantchargepermittedtohaveahigherrefrigerantflowrate
(disappearingofthetwophasestateatthepumpsupply),increasinginthesameprocessthe
pressureratioandtheshaftpower.
66
4.4 Analysisoftheresults.
4.4.2 Heatbalances.
Inordertoevaluatetheaccuracyofthemeasurements,aheatbalanceoverthecomponentsofthe
circuitisrealized.Figure34and35showthisheatbalancefortheevaporatorandforthecondenser.
67
Figure34:Airsideheattransfervsrefrigerantsideheattransferoverthe
firstexchangeroftheevaporatorandoverthewholeevaporator
Theheatbalanceovertheevaporatorshowsaverygoodagreementbetweentheheattransfersoverthe
firstexchanger.Thefivecircledpointscorrespondtoatwophasestateattheexhaustofthefirst
exchanger.Theheattransferontherefrigerantsideisthusunderestimated(itcouldbeoverestimated
ifthetemperaturewasjustabovethesaturationtemperature).
Theheatbalanceforthewholeevaporatorisnotasgoodasthepreviousone,butremainsacceptable.
Thedifferencescanbeexplainedbyparasiticheattransfers,measurementsuncertainties,andairflow
rateuncertainties.
Theheatbalanceoverthecondenseralsoshowsaverygoodagreement.
68
Figure35:Watersideheattransfervsrefrigerantsideheat
transferoverthecondenser
4.4.3 Pumpandrefrigerantflowrateanalysis.
Figure36showsthemeasuredrefrigerantflowrateasafunctionofthetheoreticalflowratethat
shouldbeprovidedbythepump.Asaforementioned,somepointsshowaverylowflowratedueto
thetwophasestateatthesupplyofthepump.Thiseffectisespeciallyvisibleathighpumpflowrate
(
X
pp =100%).
Thistwophasestateseemstobeverydependentontherefrigerantcharge.Duringthetests,theonly
waytomakeitdisappearwastoaddrefrigeranttothecircuit.
However,thistwophasestateisnotcorrelatedtoanullsubcoolingattheexhaustofthecondenser.
Forallthetests,thesubcoolingdegreewassuperiorto15K,whichmeansthatthesystemisnotin
thermodynamicalequilibrium.
Figure37showstheratiobetweenthepumptheoreticalflowrateandthemeasuredflowrateasa
functionofthesubcooling.Acorrelationbetweenthetwovaluescanhardlybededucedfromthis
graph.
69
Figure36:Measuredflowratevspumptheoreticalflowrate
Apossibleexplanationforthisphenomenonisthecondenserasymmetry:thetwoheatexchangers
composingthecondenserareinstalledinparallelontherefrigerantsidebutinseriesonthewater
side.Theconsequenceisahigherwatertemperatureinoneofthecondenser(about3C)andhencea
reducedheatexchange.Itisthuspossiblethatoneoftheheatexchangercondensesthefluidandeven
subcoolsit,whiletheotheroneisnotcapableofcompletingthecondensation.Whenthetwoflows
aremixedattheexhaustofthecondenser,itcanoccurtheparadoxicalsituationinwhichatwophase
stateremainsinthesubcooledfluid,thevapornothavingtimetobecondensedbeforeenteringthe
pump.
4.4.4 Refrigerantchargeandsubcooling.
InRankinecycles,itisgenerallyassumedthattherefrigerantchargeimposesthesubcooling:the
liquidoccupyingmorespaceinthecondenserandintheevaporator,ithasahigherexchangesurface
tobesubcooled.
Figure38showsthesubcoolingatthecondenserexhaustasafunctionoftherefrigerantcharge.The
subcoolingiscalculatedbytwodifferentmethods:
Thetraditionalway,withthesaturationtemperaturedefinedattheexhaustpressureofthe
condenser
Asecondpossibilityistodefinethesaturationtemperatureatthesupplypressureofthe
condenser.
Onlythelast15pointsaretakenintoaccount,sincetheweightmeasurementoftheexpansiontank
wasnotsetupforthe24firsttests.
70
Figure37:Errorinthepumpflowratevssubcooling
M
f , bal istherefrigerantmassintheexpansiontank.Theloweritsvalue,thehighertherefrigerant
chargeinthecycle.
Figure38showsthatthesubcoolingisindeedfunctionoftherefrigerantcharge.Itishowevermuch
morevisibleifthesubcoolingiscalculatedwiththesecondmethod.
4.4.5 Asynchronousmachineefficiency.
Theshaftpowerismeasuredwiththetorquemeterwhiletheasynchronousmachineoutputpoweris
measuredwiththewattmeter.Itisthereforepossibletocalculatetheefficiencyofthesystem
asynchronousmachine/beltandpulleyscoupling.Thisisdoneinfigure39:
71
Figure38:Subcoolingatthecondenserexhaustasafunctionoftherefrigerant
chargeintheexpansiontank
Theasynchronousmachineefficiencydecreaseswiththeshaftpower,andseemstostabilizetoavalue
between80and90%forhigheroutputpowers.Thenominalpoweroftheasynchronousmotoris7,5
kW.Itisthereforeobviousthatthemachineisnotworkinginitsoptimalrangeofpower.
The(verysmall)variationoftheefficiencyatdifferentshaftrotationspeedscanbeexplainedbya
betterorworsealignmentofthepulley,asthelatterhadtobechangedtomodifytheshaftrotation
speed.
4.4.6 Presenceofnoncondensablegasinthecircuit.
Thefollowingobservationleadstotheconclusionofthepresenceofanoncondensablegasinthe
circuit:
Whenthecycleisstoppedseveralhours,itisobservedthatitspressurestabilizesatameanvalueof
1.3to1.4barandatatemperatureclosetotheambienttemperature(around22C).Thevapor
pressureoftheworkingfluidatthistemperatureis0.82bar.Itisthereforeconcludedthatanon
condensablegas(mostlikelyair)ispresent,whosepartialvaporpressureinthecycleisaround0.5
bar.
Thepresenceofairinthecircuitisexplainedbythenegativerelativevaporpressureoftherefrigerant
atlowtemperatures.Itisverylikelythatthisairentersthecircuitthroughsmallleakswhenthe
temperatureisminimal(duringthenight).Thepressureitthenincreaseduptheobservedpressure
whenthetemperatureofthecircuitincreasesduringtheday.
Toremovetheair,thepumpisrunatthebeginningofeachtestwithoutactivatingtheheatsourceand
theheatsink.Apartoftheairispurgedatthedraincock.Itisobservedthatthemeanpressureofthe
cycledecreasesfrom1.4bardownto1.1bar.
Thispresenceofairinthecircuitcan'texplaintheobservationofthetwophasestateatthepump
supply,sinceasecondliquidindicatorinstalledattheexhaustofthepumpdoesnotdetectanybubble.
72
Figure39:Asynchronousmachine
4.5 Conclusion
Atotalof71testsisperformedonthetestbench:
The two first series of tests show acceptable results, they demonstrate the feasibility ofsuch an
organicRankinecycle.
Inthosetests,theefficienciesreachedarequitelowandveryfarfromtheCarnotefficiency. They
couldeasilybeimprovedbyabetterdesignoftheexpanderandanoptimizationofthecycle.
Themeasurementsshowmanyincoherencies,especiallyfortheairflowrate:themeasuredflowrate
seemswrongandisabandonedinfavorofthetheoreticalflowrateofthehotairgenerator.
Ahighpressuredropisalsomeasuredinthecondenser.Thispressuredropbeingdetrimentaltothe
cycle,itistobeimprovedinthenexttests.
Therangeofparametersusedforthefirstseriesoftestisquitelimited.Itiswidenedinthesecond
series.
Itisobviousthatsomerefrigerantlossesoccurredduringthetestperiod.Theconsequenceisavery
lowrefrigerantflowrateinthesecondseriesoftest,andapooroutputpower.
Thethirdseriesoftestisperformedinagreementwiththeexperienceandtherecommendationsof
thetwofirstsetsoftests.Thecircuitisinsulated,acoriolisflowmeterisinstalled,andanexpansion
tankissetup.
Theresultsshowaverygoodimprovementoftheoutputpowerandofthecycleefficiency.
Anotherveryimportantimprovementonthetestbenchistheabilitytomodifytherefrigerantcharge.
Thischargeturnsouttobeaprimordialparametertomaximizethecycleoutputpower.
Thelowflowratedetectedinthesecondseriesoftestsisexplainedbytheapparitionofatwophase
stateattheexhaustofthecondenser.Thistwophasestateoccursevenifthefluidissubcooled,andis
explainedbythecondenserasymmetry.
Thepresenceofanoncondensablegasishighlightedandshowsthatthecircuitisnotcompletely
hermetic.
73
SylvainQuoilin Chapter5:Models
Chapter5:
ModelingtheComponents
oftheCycle
5 Models
Essentially,allmodelsarewrong,
butsomeareuseful
GeorgeBox
74
5.1 Introduction
Foreachcomponentofthecycle,amodelisdeveloped.Itisvalidatedonthemeasurementsandthe
correlationbetweenthepredictedandmeasuredvaluesispresented.
Thesimplifiedmodelsusedforthefirstsetoftestsarenotnecessarilysuitableforthefurthertests:
sincetheinputsofthecyclerunoverawiderrangeofvalues,thehypothesisofthefirstmodelsare
notalwayssatisfiedfortheothertests,andnewmodelshavetobedeveloped.Theevolutionofthe
modelsisthusdescribed.
Aglobalmodelofthecycleisfinallydevelopedbyconnectingtheaforementionedmodelstogether.
ThemodelsaredevelopedunderEES(EngineeringEquationSolver).Themethodologyusedisthe
following:
Eachmodelisdevelopedinaseparatesheet.Theparameters,theinputsandtheoutputsofthe
modelareseparatedintheworkingsheet.
Theparametersofthemodelareidentifiedandthemodelisvalidatedbycomparisonwith
measurementscarriedoutontheRankinecycletestbench.
Themodelisnextplacedintoamoduleinthesamesheet,inordertobeeasilytransportableto
anothersheet.
Allthemodulescontainingthemodelsarefinallycopiedintoaseparatesheet.Theguess
valuesofthemodulesarealsoimported,inorderforthosemodelstoconvergeonceintegrated
intothenewworkingsheet.
Themodulesareconnectedtogetherinordertogettheglobalmodelofthecycle.
Thismethodpresentstheadvantageofcreatingalibraryofmodules,eachonecontainingamodel.
Thesemodulescanveryeasilybecopiedandusedinotherworkingsheetsorfortotallydifferent
projects.Thisavoidsthenecessityofrewritingthetotalityoftheequations.
Anotherimportantguidelinefollowedduringthisworkistheuseofproceduresasmuchaspossible.
Proceduresareequivalenttomodulesattheexceptionthatalltheequationstheycontainareexplicit.
Everytimeitispossible,amodeloragroupofequationisputinaprocedureinsteadofamodule.
Thisreducesthecalculationtimeandsimplifiesdramaticallytheconvergenceoftheresolution.
75
5.2 Expandermodel
ThescrollexpandermodelisbasedoncompressormodelproposedbyE.Winandyin1999.
[Winandy,1999].ThemodeloftheexpanderisvalidatedwithwatersteambyV.LemortinhisDEA
thesis,[Lemort,2006].
Theevolutionofthefluidstatethroughtheexpanderisdecomposedintothefollowingsteps:
1. Pressuredrop(su su,11)
Itiscalculatedby:
C
thr , su
isthefluidvelocityinthefictitiousnozzlemodelingthepressuredrop.
Thispressuredropisnottakenintoaccountforthetwofirstsetoftests.
2. Coolingdownandpressuredrop(su,11 su,1)
Vaporenteringtheexpanderexperiencesfirstapressuredropandisnextcooleddownatthe
contactofthemetallicmassoftheexpander(insupplyline).Forthepurposeofthe
calculation,thisheattransferisassumedtooccurbetweenthefluidandafictitiousisothermal
envelope(wall)atatemperature
T
w
.
Thecalculationoftheheatexchangeiscarriedoutwiththeclassical cNTU method:
76
3. Isentropicexpansion(su,1 ad)
Sincetheexpanderisavolumetricmachine,thepressureratioisaconstant(ataconstant
supplyspecificvolume).Therewillthereforebeanadaptedpressuregivenby:
Where
P
r, in, exp istheadaptedpressure.
Theexpansionbetweenthesupplypressureandtheadaptedpressureisassumedtobe
isentropic.
4. Expansionatafixedvolume(ad ex,2)
Astheadaptedpressureimposedbythepressureratioisnotnecessarilyequaltotheexhaust
pressure,asecondexpansion(thatcanalsobeacompression)takesplacefromtheadapted
pressuretotheexhaustpressure.Thissecondexpansionisrealizedatafixedvolume.The
workproducedisthusexpressedby:
5. Mixingbetweensuctionflowandleakageflow(ex,2 ex,1)
Theleakageflowpassesdirectlyfromthehighpressurezonetothelowpressurezonewithout
producinganyusefulwork.Itishencebledatthe(su,1)levelandmixedwiththeoutletofthe
expansion.
Theleakageflowrateischaracterizedbyafictitiousleakagearea.
6. Coolingdownorheatingup(ex,1 ex)
Theflowpassesthroughasecondfictitiousheatexchangerattheexhaustoftheexpander.
Thewalltemperatureisassumedtobethesamethanthatoftheinlet.Theheatexchangeis
calculatedwiththesameequationthanforstepn2.
Inaddition,ambientlossesfromthewalltemperaturetotheambienttemperaturearedefined:
77
Themechanicallossesaredefinedby:
where
T
m
isthefrictiontorque.
Thefictitiouswalltemperatureiscalculatedbyaheatbalancebetweenthemechanicallosses,theheat
exchangesbetweenthewallandthefluidandbetweenthewallandtheambientair:
Themechanicalpowerisequaltotheinternalexpansionpowerminusthemechanicallosses:
Parametersofthemodel:
Thismodelisbasedona8parametersthatdependonthephysicalcharacteristicsoftheexpander.
Thoseparametersare:
Swept volume of the scroll machine, in compressor mode. This volume is

linkedtothesweptvolumeinexpandermodebythebuiltinvolumeratio.
Builtinvolumeratio
Fictitiousleakagearea
Nominalheattransfercoefficientforthesupplyfictitiousheatexchanger.This
coefficientisdefinedforanominalflowrate.Ithastobeadaptedtothereal
flowrate.
Nominalheattransfercoefficientfortheexhaustfictitiousheatexchanger.
Frictiontorqueoftheexpander,determiningthemechanicallosses.
Heattransfercoefficientfortheambientlosses.
Throatdiameterofthefictitiousnozzleintroducedtomodelthepressuredrops
78
Inputsandoutputsofthemodel:
Itcanbeconsideredthatthemodeltakesasinput:
Expandersupplytemperature
Expanderexhaustpressure
Refrigerantflowrate
Ambienttemperature
Theoutputsofthemodelsare:
Expanderexhausttemperature.
Expandershaftoutputpower
Expandersupplypressure:sincetheexpanderisavolumetricmachine,itcan
absorbafixedvolumetricflowrate.Inordertomakethisvolumetricflowrate
fittothemassflowrate,thesupplypressurehastobeadapted.Thisexplains
whythesupplypressureisalsoanoutputofthemodel.
Identificationoftheparameters.
Themodelcontains3typesofparameters:
Knownparameters Theseparametersaregivenbytheconstructororcanbedirectly
measuredonthedevice.Thisisthecaseforthebuiltinvolumeratio
forexample.Theseparametersdon'thavetobeidentified.
Calculableparameters Theseparameterscanbeevaluated.Thisisthecasefortheambient
lossesheattransfercoefficientforexample.
Unknownparameters Theseparametersarenotgivenandcan'tbecalculated.Thisisthecase
forthecoefficientofthewallfictitiousheattransferforexample.Those
parametershavetobeidentifiedfromthemeasurements.
Theknownparametersarefixedtotheirvalue,andthecalculableandunknownparametersare
identifiedtomakethepredictedvaluesfittothemeasurementsresults.
79
Inordertodiscoverthebestparametersforthewholeexperiments,anoptimizationscriptis
developed:foragivensetofparameters,themodelisrunsuccessivelywiththeinputscorresponding
toeachtest.Theoutputsarethencomparedtothemeasuredvaluesofeachtestandameanerror
functionisdefined.
Inthecaseoftheexpander,thisfunctionhasthefollowingform:
Theerrorfunctionitthusaweightedmeanvaluebetweenthequadraticerrorsofeachpredicted
outputvalues.Thisfunctioniscalculatedforeachpoint.
Thebestsetofparametersforthemodelcorrespondstotheoneminimizingthemeanvalueofthe
errorfunctioncalculatedoneachpoint.Inordertominimizethatfunction,ageneticalgorithmis
used,thegenesoftheindividualsbeingtheparametersofthemodel.
Thereasonjustifyingtheuseofthegeneticalgorithmisitscharacteristicsofnotconvergingtolocal
minimums.Thesetofparameterresultingfromthisoptimizationishencesupposedtobethebestset
possible.
Forthisoptimization,thenumberofindividualsissetto32,andthenumberofgenerationto64.
Thisoptimizationisperformedtwotimes:
Onetimeforthetwofirstsetoftests
Asecondtimeforthethirdsetoftests,sincetheexpanderhasbeeninsulatedandits
parametershavebeenmodified.
80
Theresultsoftheidentificationarethefollowing:
Twofirstsetsoftests Thirstsetoftests
Forthetwofirstsetoftests,thesupplypressuredropwasnoincludedinthemodel.Itexplainsthe
absenceoftheparameter
d
thr , su,exp
inthefirstcolumn.
Inthethirdsetoftests,theambientheatlossescoefficient
AU
amb, exp
islowerthanforthefirst
tests.Itisofcourseexplainedbytheinsulationoftheexpander.
Thelowerfrictiontorqueandhigherleakageinthethirdsetoftestsmightbeexplainedbyareduction
ofthepressurebetweenthetwoscrollsorbyareductionoftheelasticityofthesealingmaterial
duringthemonthsbetweenthesecondandthirdsetoftests.
Itishoweververydangeroustoconcludeanythingfromidentifiedparameters,especiallywhenthe
modelhaschangedlikeinthiscase.Thoseparametershaveindeedalimitedphysicalmeaning,asthe
modelissimplifiedanddoesn'ttakeintoaccountallthephysicalphenomenas.
Validationofthemodel.
Oncetheparametersareidentified,thepredictedandmeasuredvaluesarecomparedforeachoutput
ofthemodel.Theresultsofthiscomparisonareshowninfigure40.
Forthetwofirstsetsoftests,figure40showsthattheexhausttemperatureisfairlywellpredicted.
However,thepredictedsupplypressureisnearlyalwaysbelowthemeasuredvalueandamaximum
errorof15%isdetectedintheprevisionoftheexpanderoutputpower.
Theerrorinthepredictionoftheoutputworkismostlikelyduetothebadpredictionofthesupply
pressure.Inordertosolvethisproblem,apressuredropmodelisaddedattheexpandersupplyinthe
modelusedinthethirdsetoftests.
Thethirdsetoftestsshowsaverygoodagreementbetweenallthepredictedandmeasuredoutputs.It
isthusconcludedthattheadditionofthepressuredropinthemodelishighlyadvantageous.
81
Twofirstsetsoftests: Thirdsetoftests:
Expanderexhausttemperature:
Expandersupplypressure:
Expanderoutputpower:
Figure40:Predictedvsmeasuredexpanderoutputsforthe3setsoftests
82
Themodeloftheexpanderpermitstoinvestigatemoredeeplythemechanicallossesintheexpander
andinthetransmissionbeltbetweentheexpandershaftandthetorquemeter.
Figure41givesthemeasuredpoweratthetorquemeterasafunctionoftheinternalpowergivenby
theexpandermodel.The3distinctrotationalspeedsusedareseparated.
Figure41showsthatforeachrotationalspeed,aconstantpowerlossoccurs,aswellasproportional
losses.Theconstantlossesaredifferentforeachrotationalspeedandaremodeledbythefriction
torque
T
m
parameter.Since
T
m
isaconstant,thehighertherotationalspeed,thehigherthe
constantlosses.
Theproportionallosseshaveseveralorigins,butthemainoneisthebeltefficiency.Inthemodel,itis
settoanhypotheticalvalueof0.95.Figure41showsaproportionalityfactorforeachrotationalspeed
(0.95;0.91and0.91).Thevariationsofthisproportionalityfactormighthave2origins:
Betterorworsealignmentofthepulleys,ormodificationofthetensioninthebelteachtime
thepulleyischanged.
Presenceofproportionallossesintheexpander.
Theselectionbetweenthosetwohypothesesisnotpossiblewiththemeasurements.Thisproblem
maybesolvedbyplacingthetorquemeterdirectlyontheexpandershaft.Thiswouldpermitanexact
calculationoftheproportionallosses,andtheintegrationofthoseproportionallossesinthemodelof
theexpander.
83
Figure41:Measurepoweratthetorquemetervsinternalpower
5.3 Condensermodel.
5.3.2 Onezonemodel.
Thisfirstmodelisasemiisothermalmodel.Thismodelassumesthatthefluidisundertwophase
stateinthewholecondenser.Thevaporandliquidphasesareneglected.
Theheatexchangeiscalculatedbythe cNTU method:
Apressuredropisdefinedontherefrigerantside.Itischaracterizedbyafictitiousnozzlediameter
d
r , su, cd
:
Theheattransferiscalculatedby:
Ambientlossesaredefinedonthewaterside,attheexhaustofthecondenser.Theyarecharacterized
bytheexchangecoefficient
AU
cd, amb
Therefore,theparametersofthismodelare:
AU
cd , amb
, AU et
d
r , su, cd
.
84
Validationofthemodel:
Thevalidationofthemodelforthepredictionof
P
r , ex, cd
isgiveninfigure42.
Figures2showsthatthemodelisacceptableforthefirstsetoftests,butnotforthesecondone.This
ismostlikelyduetoavariationoftherefrigerantchargeduetorefrigerantlosses:themodelcan't
predicttheheattransfer(andthusthepressure)ifthelevelofliquidismodifiedinthecondenser.
Asecondmodelishenceproposed.
5.3.3 Threezonemodel
Inthismodel,thecondenserisdividedinto3zones:aliquid(subcooled)zone,atwophasezone,and
avapor(superheated)zone.Thesizeofthese3zonesisvariableanddeterminedbythemodel.
Themeanheatexchangecoefficientofeachzoneissupposedconstantandisidentifiedfromthe
measurements.
85
Figure42:Predictedvsmeasuredvalueofthecondenserexhaustpressure
Thewatersideambientlossesarenotshownonthesketchbuttheyaredefinedinthesamewaythan
forthefirstmodel,atthecondenserexhaust.
Theresistancesforeachzonearecalculatedby:
Wherecfandhfstandforcoldfluidandhotfluid.Cisaproportionalitycoefficient,parameter
ofthemodelandidentifiedfromtheexperiments.
Theexponentofthemassflowrateischosenasn=0.8forboththewaterandtherefrigerantside,
sincetheflowcanbeconsideredasturbulent.
TheAUforeachzoneisgivenbytheabovecalculatedresistancesandthe o factor(thethermal
resistanceofthemetallicplateisneglected):
The o foreachzonearelinkedtogetherbytherelation:
Sincethepropertiesofthefluidaredifferentwhenitsstateischanged,adifferentpressuredropis
definedforeachzone.Itisdefinedbythediameteroftheequivalentnozzleareaandisponderedby
therelativevolume o ofthezoneinthecondenser:
AP
r , cd
=o
cd , sh
AP
r , cd , sh
+o
cd , tp
AP
r , cd, tp
+o
cd, sc
AP
r , cd, sc
86
Therefore,thetotalsetofparameterswiththeiridentifiedvaluesis:
Proportionalitycoefficientfortherefrigerantsideresistanceinthe
liquidzone
twophasezone
vaporzone
Proportionalitycoefficientforthewatersideresistance
Ambientlossescoefficient
Diameterofthefictitiousareaforthepressuredroponthetwo
phasezone
Throatdiameterofthefictitiousnozzlemodelingthepressuredrop
intheliquidzone
Throatdiameterofthefictitiousnozzlemodelingthepressuredrop
inthevaporzone
InthemodeloftheRankinecycle,itisconsideredthatthecondenserimposesitspressure(boththe
supplyandtheexhaustpressure,asthemodeloftheexpanderaccountsforthepressuredrop),its
supplyandexhausttemperaturesbeinggiven.
87
Figure43:Predictedvsmeasuredcondensersupplypressure
Figure43showstheagreementbetweenthepredictedandmeasuredvalueofthesupplypressurein
thecondenserforthethirdsetoftests.
Theerrorbetweenthepredictedandmeasuredvaluesishighformanypoints.Inparticular,thepoint
circledinredshowsaveryhigherror.Itcorrespondstothetest040507E,inwhichtherefrigerant
chargeisincreasedtoitsmaximumacceptablevalue.Itispossiblethatthesupplyofthecondenseris
intwophasestateforthistests,whichcouldexplaintheveryhigherror.
Itisobviousthatthecondenserdoesn'tactasexpectedbythemodel.Thisunexpectedbehaviorofthe
condenserhasalreadybeenhighlightedduringtheexperiments,inchapter4.4.3.Adeeper
investigationisneededheretounderstandtheseresults.
Figures44showstheagreementbetweenthepredictedandmeasuredvaluesofthepressuredropover
thecondenserforthethirdsetoftests:
Theagreementbetweenthetwovaluesisacceptableexceptforafewpoints:
Thetwopointscircledattherightofthepicturecorrespondtoexperimentsinwhichan
importantrefrigerantchargeisintroducedinthecircuit.Thiserroristhuslinkedtotheone
explainedabove.
Agroupofpointscorrespondingtothetest030507showsaveryhigherrorontheleft.
However,theproblemseemstocomefromthemeasurementandnotfromthemodel.The
pressuredropisindeednegativefor2ofthosepoints,whichisofcourseunrealistic.Thiserror
isduetoamalfunctionofthepressuretransducermeasuring
p
r ,ex ,cd
.Thepressure
transducerisreplacedafterthattest.
88
Figure44:Predictedvsmeasuredpressuredropoverthecondenser
5.4 Evaporatormodel
Proposingasimpleandaccuratemodelfortheevaporatorisacomplextask.Itisindeedcomposedof
twohotairsourcesandthreeheatexchangers.Dependingonthetest,thetwophasezonemoves
betweentheexchangers,makingitverydifficulttoproposeaglobalmodelforalltheexperiences.
Inputsofthemodel:
T
a,su, hx12
: Firsthotairsourcetemperature
T
a,su, ac
: Secondhotairsourcetemperature
M
a
: Airflowrate
M
f
: Refrigerantflowrate
T
f , su, hx1
: Evaporatorsupplytemperature,refrigerantside
p
f ,ex , hx3
: Evaporatorexhaustpressure
T
amb
: Ambienttemperature
Outputsofthemodel:
T
f , ex, ev
Exhaustrefrigeranttemperature
Q
ev
Totalheattransferovertheevaporator
Ap
ev
Pressuredropovertheevaporator
5.4.2 Firstmodel
Afirstmodelisproposedforthefirstseriesoftests.Itassumesthatthetwophasezoneislocated
overthetwolastexchangersHX2andHX3,thefirstexchangerHX1beingfilledonlywithliquidand
workingasapreheater.
Inthesamemannerasforthecondenser,heatambientlossesaredefinedontheairsideattheexhaust
ofeachexchanger.Twopressuredropsaredefinedontherefrigerantsideatthesupplyandatthe
exhaustoftheevaporator.
89
Theparametersofthismodelare:
Theheattransfercoefficientforeachexchanger
ThevalueofAUforthe3ambientlossesand
Thethroatdiametersofthefictitiousnozzlesmodelingthepressuredrops.
Inorderforthismodeltogivegoodresults,theevaporatorhastoworkinveryspecificconditions.It
isthereforeverylimited.
Theagreementbetweenthepredictedandmeasuredheattransferforthefirstseriesoftestisgivenin
figure45:
Themodelsshowsagoodagreementforthefirstsetoftests.However,whenappliedtothethirdset
90
Figure45:Predictedvsmeasuredhattransferoverthefirstheat
exchangerhx1
oftests,themodelisnotcapableofpredictingtheexhausttemperatureanymore(seefigure46):
Thisisduetothevariationoftherefrigerantlevelintheevaporator:sincethetwophasezonemoves
betweenthethreeheatexchangers,thehypothesisofatwophasezonelocatedonthetwolast
exchangersisnotvalidanymore.Another,morecompletemodelhastobedevelopedforthesecond
andthirdseriesoftests.
5.4.3 Secondmodel.
Theconfigurationoftheevaporatorisveryparticular:itiscomposedofthreeplateheatexchangers
andtwoheatsources.Developinganappropriatemodelforthisevaporatorconfigurationisadifficult
task:amultizonemodelhastobeimplementedforeachheatexchanger,andthemodelcan'tbeused
forallthepoints:ithastobemodifieddependingonthelocationofthetwophasezoneinthe
evaporator.
Inordertomodelthistwophasezone,eachheatexchangerhastobediscretized,totakeintoaccount
themodificationofthefluidpropertiesasafunctionofitsquality.Thisdiscretizationmakesthe
convergenceofthemodelmorecomplicated.
Thepurposeofthepresentworkistointegrateamodeloftheevaporatorintotheglobalmodelofthe
cycle.Thismodelmustthereforebesuitableforallthetests,andconvergeoverawiderangeof
parameters.
Theideaofadiscretized,multizonemodelishenceabandoned,andtwoothersolutionsaretested:
Linearregression:Alinearregressionisperformedinordertoexpresstheoutputofthe
91
Figure46:Predictedvsmeasureexhaustevaporatortemperature,firstmodel
modelasapolynomialfunctionoftheinputs.A2
nd
orderpolynomialwithcrosstermsis
used.
Suchapolynomialwithoneoutputandtwoinputswouldbewritten:
output=C0+Ainput1+Binput2+Cinput1input2+Dinput1
2
+Einput2
2
C0,A,B,C,DandEbeingidentified.
Figure47showsafairlygoodagreementbetweenthepredictedandmeasuredvalues.
However,thelinearregressionhastobeusedverycautiously:adeeperanalysisshowsthat
thissolutionisnotacceptable,sinceitdivergesassoonasoneparameterismodifiedfromits
initialvaluetoanotherveryclosevalue.Thissolutionisthusrejected.
Identifiedcorrelation:inordertoexpresstheoutputasafunctionoftheinputsinabetter
way than with the linear regression, all the values are transformed into nondimensional
numbersandthefollowingrelationisinvestigated:
T
f , ex, ev, pred
T
ref
=C1

M
a
M
ref
)
C2
T
a , su, hx3
T
ref
)
C3
T
a, su, hx12
T
ref
)
C4
T
f , su, hx1
T
ref
)
C5
p
ref
p
f , ex,hx3
)
C6
Where
T
ref
=298K

M
ref
=0.1kg/ s
p
ref
=10
5
Pa arethereferencetemperature,flow
rateandpressureandarechosenarbitrarily. AllthetemperaturesareexpressedinKelvin.
C1,C2,C3,C4,C5andC6areidentifiedtominimizetheerrorbetweenthemodelandthe
measurements.Theiridentifiedvaluesare:
C1=1.027;C2=0.1545;C3=0.8302;C4=0.07308;C5=2.201;C6=0.02383
Figure48showstheagreementbetweenthepredictedandmeasuredvaluesifthiscorrelation
isused.
Theagreementisnotasgoodaswiththelinearregression,butnodivergenceofthepredicted
92
Figure47:Predictedvsmeasuredevaporatorexhausttemperature
valueisstatediftheparametersaremodified.
This second solution is thus chosen to predict the evaporator exhaust temperature.
Itisimportanttokeepinmindthatthegoodbehaviorofthecorrelationcannotbeguaranteed
outsidetherangeofparametersonwhichitwasdefined.Awarningisthusintegratedinthe
programifthecorrelationisusedoutofitsrangeofparameters.
Figure48:Predictedvsmeasureevaporatorexhausttemperature
5.4.4 Singlephasemodel.
Inthefirstsetoftests,thefirstheatexchangeroftheevaporator,hx1,isfilledwithrefrigerantin
liquidstateforallthepoints.Thosepointsareusedtodevelopaheatexchangemodelforthesingle
phasezone.
ThismodelusestheMartincorrelationinordertopredicttheNusseltnumber[Martin,1995].This
correlationissuitedforcorrugatedplateheatexchangers.Itisdeducedfromatheoreticalapproach,
takingintoaccounttheChevronangle,theplatethickness,theamplitudeofthecorrugationsandthe
platesize.
TheNusseltiscalculatedby:
WhereKisacorrectionfactor.
93
fistheFanningfrictionfactorandiscalculatedby:
With:
CalculationoftheReynoldsnumber:
Theparametersofthemodelare:
Amp=210
3
m AmplitudeoftheCorrugations
\=710
3
m Wavelengthofthecorrugations
d=2.410
3
m Distancebetweentwoplates
e=5.310
4
m Meanthicknessofoneplate
=30
Chevronangle
u
m
Viscosityofthefluidatthemeantemperature
u
w
Viscosityofthefluidatthewalltemperature
AcorrectionfactorKhastobeaddedtotheNusseltcorrelationinordertomakethepredictedvalues
fittothemeasuredvalues.AvalueofK=0.18isidentified.
TheMartincorrelationisusedontheairandontherefrigerantside.Thethermalresistanceofthe
plateistakenintoaccount.TheAUiscalculatedby:
94
TheagreementbetweenthepredictedvalueoftheMartincorrelationandthemeasuredvalueisgiven
infigure49:
Thisagreementbetweenthetwovaluesisfairlygood.Thismodelcanthuspotentiallybeusedin
furtherworksinordertomodelproperlytheevaporatorwithamultizonemodel.
Itcanalsobeadaptedtootherplateheatexchangersveryeasilybychangingthegeometrical
propertiesinthecorrelation.
95
Figure49:Predictedvs.measuredvalueoftheheatexchangeoverthe
firstexchangerforthefirstsetoftests
5.5 Pumpmodel
Inordertopredictthepumpconsumption,aconstantefficiencyisassumed.
Thepowertransmittedtothefluidisgivenby:
W
sh , pump
=

M
r
v
r , su, pump
p
r , ex, pump
p
r , su , pump
)
Andtheelectricalconsumptionisexpressedby:
W
elec, pump
=

W
sh , pump
pump
Sincethepumpconsumptionmeasurementisonlyperformedfortwopointsofthesecondsetoftests,
avalidationofthismodelisnotpossible.Thepumpefficiencyisidentifiedfromthesetwopointsand
theresultingvalueis:
n
pump
=0.15
Thisvalueisinagreementwiththeconstructordata:themaximumoutputpressureofthepumpis20
barsandthemaximumenginepoweris550W.Ifthesemaximumvaluesaretakenasapointof
functioning,theefficiencyidentifiedonthispointisequalto0.14
Theflowrate,asaforementioned,isimposedbythepumpwiththerelation:
V
r , pump
=X
pump

V
r ,max, pump
with
V
r ,max, pump
=210l / h
Thepumpshowsapoorefficiency.Itisindeednodesignedtomaximizetheefficiency,buttodeliver
amodifiableandaccurateflowrate.Thisisveryappropriateforthistestbench,sincetheflowrate
hastobemodified.However,amoreefficientpumpshouldbeintegratedinacommercialapplication
oftheORC,increasinginthesameprocessthecycleefficiency.
96
5.6 Linesmodel
Thepressuredropsandtheambientlossesofthepipinglinesarealsomodeled.Onlytwolinesare
takenintoaccount:thepumpevaporatorlineandtheevaporatorexpanderline.Thesizeoftheother
linesbeingverysmall,theambientlossesandthepressuredropsareneglected.
Theidentifiedvaluesoftheheatexchangecoefficientanddiameteroftheequivalentnozzleareaare
givenintable4forthethirdsetoftests.
EvaporatorExpander PumpEvaporator
AU
amb
2.1W/K 0.8W/K
d
r , su, line
0.0011 0.005
Table4:Parametersofthelinemodelforthethirdsetoftests
5.7 Refrigerantchargemodel
Sincenodetailedsuitablemodeloftheevaporatorisavailable,anaccuratemodeloftherefrigerant
chargeinthecycle,takingintoaccountthevoidfractionintheheatexchangersisnotfeasible.
Itishoweverpossibletoevaluatetherefrigerantchargebymeansofseveralsimplifyinghypothesis.
Thepurposeistodemonstratethattheglobalevolutionofthepredictedrefrigerantchargecorresponds
totheexpectations
.
Liquidlevelintheevaporator:
Thefollowingassumptionsarerealized:
Thethreezones(liquid,twophase,vapor)arewellseparatedintheevaporator.
Eachzoneoccupiesavolumeintheevaporatorproportionaltotheheatexchangedinthezone
97
Forinstance,thevolumeoftheliquidzoneiscalculatedasfollows:
Where:
isthespecificheatexchangeintheliquidzone
istheidentifiedcoefficientofproportionality
Therefrigerantmassintheevaporatorisroughlyevaluated,assumingthathalfofthetwophasezone
isfilledwithliquid.Themassofthevaporphaseisneglected:
Theidentifiedvaluesofthecoefficientsare:
Liquidlevelinthecondenser:
Thecondensermodel(seechapter5.3)calculatestherelativevolumeofeachzoneinthecondenserby
meansofthe o coefficient.Therefrigerantchargeinthecondenseriscalculatedthesamewayas
theevaporatorrefrigerantcharge:
Thetotalrefrigerantchargecanthusbecalculated.
Thepredictedrefrigerantchargeasafunctionoftheamounttherefrigerantchargeisgiveninfigure
50.Onlythe15lastpointsaretakenintoaccount,sincetheweightmeasurementoftheexpansion
tankwasnotsetfortheothertests.
98
Theagreementbetweenthetwovaluesisacceptableforsuchasimplifiedmodel.Thebadprevision
ofthechargeforthetwopointsontherightisduetotheproblemsdetectedinthecondensermodelfor
highrefrigerantcharges(seefigure43)
5.8 Globalmodelofthecycle.
Inordertosimulatethewholecycle,thedifferentmodelsexplainedaboveareconnectedtogether.
Therefrigerantchargemodelisnotintroducedintheglobalmodel,sinceithasbeenvalidatedona
limitednumberofpointsandwithsimplifyinghypothesis.Thesubcoolingisthusimposedtothe
model.
Thefollowingblockdiagramshowsthevariablesconnectingthecomponents.
99
Figure50:Predictedvsmeasuredrefrigerantcharge
Eachcomponentissupposedtoimposeoneorseveraloutputs:
thepumpimposestherefrigerantflowrate,
theexpanderimposestheevaporatorexhaustpressureandthecondensersupplytemperature,
theevaporatorimposestherefrigerantoverheatingandthepumpexhaustpressure,
thecondenserimposestheexpanderexhaustpressureandthepumpsupplypressure.
Thepredictionoftheoutputshaftpowerbytheglobalmodelisshowninfigure52.
Forpoints22,24,35,and36,themodeldidnotconverge.Thosepointsarethusnotrepresentedon
thefigure.
Withtheexceptionofthetwopointscircledinred,figure52showsamaximumerrorof20,1%
betweenthetwovalues.The2pointscorrespondtothetestsinwhichahighrefrigerantchargewas
introduced,andinwhichabadpredictionofthepressurebythecondensermodelwasdetected(see
figure43).
100
Figure51:Blockdiagramoftheglobalmodelofthecycle
Figure52:Predictedvsmeasuredshaftpower,cyclemodel
Figure53showsthepvdiagramofthemeasuredandpredictedvaluesofthecyclefortestn8.Itis
obviousthatthemainerrorcomesfromabadpredictionofthecondenserpressure.Themodelofthe
condenseristhusindeedtheweakestlinkoftheglobalmodelandhastobeimproved.
101
Figure53:Predictedandmeasuredpvdiagramofthecycle
5.9 Optimizationofthecycle
Thischapterdescribeshowtheglobalmodelcanbeusedtooptimizethecycle.
Astartingbasiccaseisdefined,andtheevolutionoftheperformancesofthisbasiccaseisshown
foreachpossibleimprovement.Anoptimizedcycleisfinallydefinedanditsperformancesare
comparedtothebasiccase.
Thetwoparametersusedfortheevaluationoftheperformancesarethenetoutputpowerandthecycle
efficiency.Thenetoutputpowerisgivenbythedifferencebetweentheshaftpowerandthepump
consumption:
Thebasiccaseisdefinedasfollows:
Pumpsetting:
Ambienttemperature[C]:
Airflowrate[kg/s]:
Firstandsecondhotairsourcetemperature[C]:
Waterflowrate[kg/s]:
Subcooling[kg/s]:
Thepredictedperformancesofthecycleinthoseconditionsarethefollowing:
Netoutputpower[W]:
Cycleefficiency:
Expanderinternalbuiltinratio:
Theeffectoftheinternalbuiltinratioontheperformancesofthecycleisstudiedinfigure54:
102
Figure54showsthatthereisanoptimalbuiltinratiofortheconditionsofthesimulation.Itsvalueis
around5andcorrespondstoanisentropiconlyexpansion(theexpansionataconstantvolumeis
null)
Pumpsetting:
Figure55showsthatthehighertheflowrateimposedbythepump,thehigherthecycleefficiency
andnetoutputpower.Thecycleefficiencyseemstostabilizeathighflowrates.Itislikelythata
maximumwouldappearforevenhigherflowrateswhiletheoutputpowercontinuesincreasing.
103
Figure54:Influenceoftheinternalbuiltinratio
Figure55:Influenceofthepumpflowrate
Sweptvolumeoftheexpander:
Thesweptvolumeoftheexpandershowsanoptimal,likethebuiltinvolumeratio.Thisvalue
correspondstoanisentropiconlyexpansion.
Pressureratio:
Inordertostudytheinfluenceofthepressureratioatagivenflowrate,thispressureratioisimposed
andthesweptvolumeismodifiedinordertokeepthesamesupplypressure.Thissweptvolumehasto
bereducedinordertoachievehigherpressureratios.
104
Figure56:Influenceofthesweptvolume(incompressormode)
Evaporatorexhaustsuperheating:
Intheory,thehigherthesuperheatinglevelattheexpandersupply,thehighertheoutputpower,but
thelowerthecycleefficiency.Figure58showsthatthisisnotthecaseforthiscycle:theoutput
powerincreasesasexpected,buttheefficiencyincreasesaswell.Thisisexplainedbytheconstant
frictionlossesintheexpander:theirproportionintheoutputpowerislowerifthatoutputpoweris
increased,andtheefficiencyincreasesinthesameprocess.However,theincreaseofthecycle
efficiencyremainsquitelow(from0.0585to0.0625).
105
Figure57:Influenceofthepressureratio
Figure58:Influenceofthesuperheating
Condenserexhaustsubcooling:
Figure59shows,asexpected,thatifthesubcoolingisreduced,thecycleefficiencyandtheoutput
powerareincreased.However,sinceasubcoolinglowerthan15Kcouldnotbeachievedduringthe
testsbecauseofthetwophasestateatthepumpsupply,thisvalueofthesubcoolingissetasthe
minimumvalueachievable.
Improvementoftheexpanderleakageareaandofthefrictiontorque:
Asexplainedabove,theexpandercouldeasilybeimproved,sincethemachineusedforthistestbench
isinitiallydesignedtoworkasacompressor,usingairasworkingfluid.
Toevaluatethepossibleimprovements,theleakageareaandthefrictiontorquearemultipliedbya
factor
K
improve
whosevaluerangesbetween0and1.
106
Figure59:Influenceofthesubcooling
Figure60showsthattheimprovementoftheexpander(andespeciallyoftheleakagearea)hasavery
highpotentialofimprovementfortheoutputpower.
5.10 Optimizedcycle
Thisexampleshowsthepossibleeffectofcumulatedrealisticimprovementonthebasiccase.An
optimizedcaseisdefinedwiththefollowingparameters:
Theflowrateiskeptconstant
Theinternalbuiltinpressureratioissettoitsoptimalvalue(5)
Thesubcoolingattheexhaustofthecondenserissetto15K
Theleakageareaandthefrictiontorquearereducedby50%.
Thepumpefficiencyisincreasedfrom15%upto40%
Thepredictedoutputpowerandcycleefficiencyinthoseconditionsare:
Whichcorrespondstoanimprovementofthenetoutputpowerof83%,andanimprovementofthe
cycleefficiencyof86%.
107
Figure60:Influenceoftheleakageareandthefrictiontorque
SylvainQuoilin Chapter6:Recommendations
Chapter6:
RecommendationsandPerspectives
6 Recommendationsand
Perspectives
108
TheOrganicRankineCycletestbenchhasbeenimprovedoverthe3setsofexperiments.Those
improvementsresultedinmoreaccuratemeasurements,bettercontroloftheparameters,higher
efficiencyandoutputpower.
However,somenewimprovementsonthetestbenchandonthemodelscanstillbeperformed.The
followingnonexhaustivelistdetailsthemostimportantachievableimprovements.
Improvementsonthetestbench:
Changingtheorderoftheheatsourcesintheevaporator:
Themeantemperatureofthefirstheatsourceisapproximately185C,whilethemean
temperatureofthesecondheatsourcedoesn'texceed160C.Theheatexchangewouldhence
bebetterifthefirstheatsourcewaslocatedonthelastexchangeroftheevaporator.The
temperatureprofilewouldbeimproved,andtheexergeticlosseswouldlower.Intest34,the
secondheatsourceevendecreasesthetemperatureoftherefrigerant,whichisofcoursean
aberration.
Changingtheconfigurationofthecondenser:
Asexplainedinthepreviouschapters,theasymmetryofthecondenser(seriesonthewater
sideandparallelontherefrigerantside)resultedinatwophaseflowatthepumpsupplyand
inhighdifficultiesforthemodeling.
Thebestconfigurationwouldbeaparallelassociationforbothsides,inordertokeepthe
pressuredropsaslowaspossible,andtoeliminatethetwophaseflowattheexhaust.
Additionofadraincockattheleveloftheevaporator:
Theactualdraincockislocatedattheexhaustofthepump,intheliquidline.Aseconddrain
cock(whichcouldbeasimplepurgingvalve)betweentheevaporatorandtheexpanderwould
beveryusefultopurgethenoncondensablegasespresentinthecycle.
Installationofthetorquemeterdirectlyontheexpandershaft.Thiswillpreventthe
uncertaintieslinkedtotheunknownefficiencyofthetransmissionbeltandthusallowabetter
measurementoftheexpandershaftpower.
Improvementoftheexpander:
Thislastpointisprobablythemostimportant.Asaforementioned,theexpanderisnot
optimizedatall.Itisthecomponentofthecyclethatshowsthehighestpotentialof
improvement.Theinternalleaksmaybereduced,aswellastheexternalleaksthatcause
refrigerantlosses.Thefrictiontorquemayalsobeeasilyreduced.
Thebestoptionwouldbeahermeticexpander,whichshows2mainadvantages:
Noexternalleaks,andthusnorefrigerantlosses.Thiswouldalsopermittheuseof
flammablerefrigerantssuchasnpentane,whichwerenotuseduntilnowbecauseofthe
leakagepossibility.
Reductionoftheheatlosses:theelectricitygeneratingdevicebeingintegratedintothe
expander,itsinefficiencieswillresultinheatthatmaybetransmittedtothefluid.Itis
109
howeverimportanttoimplementageneratingdevicecapableoffunctioninginahigh
rangeoftemperatures.
Thelubricationofthescrollexpanderwouldalsobeadvantageous,astheleakageareaandthe
frictiontorquewouldbereduced.However,thelubricationraisesseveraltechnicalproblems
alreadymentionedinchapter3.2.2thatmustbeovercome.
Improvementsofthemodels:
Developmentofanappropriatemodelfortheevaporator.Thismodelshouldbeadiscretized
model,withcorrelationsforthevoidfraction,theNusseltnumber,andthepressuredrops.
Thiswouldpermitthedevelopmentofamorecompleterefrigerantchargemodel.
Checkingofthecondensermodel.Thecondensermodelwillhavetoberetestedwhenthe
configurationofthecondenserchanges.Apurelyparallelconfigurationshouldresultina
betteragreementbetweenthepredictedandmeasuredvalue.
Appropriatemodelsoftheheatexchangers,takingintoaccountthefluidproperties,would
makepossiblethemodelingofthecyclewithdifferentworkingfluids,andwouldbeavery
interestingindicatorforthechoiceofthebestworkingfluid.
110
SylvainQuoilin Conclusion
Conclusion
111
7 Conclusion
Thefirstpartofthisworkspresentsthedifferentsolutionsforlowgradeheatrecovery.Amongthem,
theOrganicRankineCycleshowsadvantageousproperties,namelyitssimplicityandgoodefficiency.
Afewdistinctworkingfluidsarecomparedtogetherandtheiridealpropertiesaredetermined:
verticalsaturationcurve,highcriticaltemperature,goodthermalproperties,highheatofvaporization,
safety,andenvironmentalimpactsarethekeycharacteristicsfortheworkingfluidselection.
R123turnsouttobeoneofthebestadaptedfluidsforahotsourcetemperaturebetween100and
200Candisselectedforthetestbenchsetupinthelaboratory.
Thetestbenchisthendescribed,withallitscomponentsandtheirworkingprinciple.Theselectionof
ascrollcompressorforthetestbenchisjustifiedbythebetterbehaviorofthevolumetricexpandersin
smallscaleunitsandbyitrobustnessevenundertwophaseflowconditions.
The3setsoftestscarriedoutonthecyclearedescribed.Themotivationsforeachsetoftestsare
explainedandtheresultsarepresented.Theexperienceacquiredduringeachseriesoftestisusedto
improvethetestbenchforthenexttests.
Thetwofirstsetoftestsshowveryhighmeasurementuncertainties,mainlyduetothelackofflow
meterinthecircuit.Refrigerantlossesarealsodetectedandtheirnegativeinfluenceontheoutput
powerisstated.
Inthethirdsetoftests,thetestbenchisinsulated,therefrigerantchargeiscontrolledandmodified,
andmuchhigherefficienciesandoutputpowersareobtained.Themeasurementuncertaintiesarealso
reduced,facilitatingthemodelingofthecycle.
Themodelsproposedforeachcomponentofthecyclearedescribed.Thereasonsfortheeventualbad
previsionofthemeasuredvaluesareexplained.Theirevolutionisalsodescribed,andthenecessityof
developingmorecomplexmodelsisjustified.Thosemodelsarevalidatedonthemeasurementsof
thefirst,secondand/orthirdsetoftests.
Themodeloftheexpandershowsaverygoodagreementbetweenthepredictedandmeasuredvalue.
Thenecessityofaddingapressuredroptothemodelishighlighted.
Thefirstmodeloftheevaporatorisvalidonlyiftherefrigerantchargeisnotmodified.Asamodel
takingintoaccounttherefrigerantchargeisaverycomplextask,acorrelationisidentifiedtopredict
theexhausttemperatureoftheevaporator.
Despitetheimprovementsperformedonthecondensermodel,thelattershowsbadresultsand
predictsthelowpressureofthecyclewithahigherrorlevel.Thisisexplainedbyanasymmetryin
thecondenserassembly,modifyingitsbehavior.
112
Theglobalmodelisfinallyusedtooptimizethecycle.Itturnsoutthat,withafewrealistic
improvements,thecycleefficiencyandtheoutputpowermaybeincreasedbynearly100%.
Afewrecommendationsarefinallydetailed,inordertoimprovethetestbenchandthemodelsin
futureworks.Amongthem,thenecessitytoimprovetheexpanderisoverriding.
113
SylvainQuoilin Bibliography
8 Bibliography
1. S.Bonnet,M.Alaphilippe,P.Stouffs,Energy,exergyandcostanalysisofamicro
cogenerationsystembasedonanEricssonengine,InternationalJournalofThermalSciences
4411611168,PAU,2005.
2. L.Borel,D.Favrat,Thermodynamiqueetnergtique,del'nergiel'exergie,pressses
polytechniquesetuniversitairesromandes,2005
3. J.M.Calm,Comparativeefficienciesandimplicationsforgreenhousegasemissionsofchiller
refrigerants,InternationalJournalofRefrigeration,Issue5,Pages833841,2006
4. J.M.Calm,G.C.Hourahan,RefrigerantDataSummary,EngineeredSystems,November2001
5. S.Canada,G.Cohen,R.Cable,D.Brosseau,andH.Price,ParabolicTroughOrganic
RankineCycleSolarPowerPlant,DOESolarEnergyTechnologiesProgramReviewMeeting,
Denver,2004
6. Y.A.Cengel,M.A.Boles,Thermodynamics,AnEngineeringApproach,McGrawHillHigher
Education,2006
7. R.Chammas,D.Clodic,CombinedCycleforHybridVehicles,SAEInternational,2005
8. YuChen,NilsP.Halm,EckhardA.GrollandJamesE.Braun,Mathematicalmodelingof
scrollcompressorspartI:compressionprocessmodeling,InternationalJournalof
Refrigeration,Volume25,Issue6,Pages731750,September2002.
9. T.C.Hung,T.Y.Shai,S.K.Wang,AreviewoforganicRankinecyclesfortherecoveryoflow
gradewasteheat,KaohsiungPolytechnicInstitute,Taiwan,1996
10. M.Jonsson,J.Yan,Ammoniawaterbottomingcycles:acomparisonbetweengasenginesand
gasdieselenginesasprimemovers,Energy26(2001)3144,DepartmentofChemical
EngineeringandTechnology,Stockholm,Sweden,2000
11. D.Kadikoff,Refrigerantindexandretrofitguidelines,HVACMechanic.com,
http://www.hvacmechanic.com/refrigerants/refrigerants.htm,November2003
114
12. ElH.M.KANE,Intgrationetoptimisationthermoconomique&environomiquede
centralesthermiquessolaireshybrides,PhDThesis,LaboratoiredEnergtiqueIndustrielle,
EcolepolytechniqueFdraledeLausanne,Suisse,2002
13. V.Lemort,C.Cueva,J.Lebrun,I.V.Teodorese,S.Quoilin,Developmentandexperimental
validationofanorganicRankinecycleModel,HeatSet2007,2007
14. V.Lemort,C.Cuevas,I.VTeodorese,J.Lebrun,Contributionaltudedescyclesderankine
dercuprationdechaleur,ColloqueInteruniversitaireFrancoQubcoissurlaThermique
desSystmes,2007
15. V.Lemort,Testingandmodelingscrollcompressorswithaviewtointegratingthemas
expandersintoaRankinecycle,DEAthesis,UniversitdeLige,2006
16. Lebrun,J.,Machinesetsystmesthermiques:notesdecours,coursMECA00061,University
ofLige,2005
17. BoTauLiu,KuoHsiangChien,ChiChuanWang,Effectofworkingfluidsonorganic
Rankinecycleforwasteheatrecovery,IndustrialTechnologyResearchInstitute,Taiwan,2002
18. H.D.MadhawaHettiarachchia,MihajloGolubovica,WilliamM.Woreka,YasuyukiIkegamib,
OptimumdesigncriteriaforanOrganicRankinecycleusinglowtemperaturegeothermalheat
sources,Energy,Chicago,2007
19. H.Martin,Atheoreticalapproachtopredicttheperformanceofchevrontypeplateheat
exchangers,YhermischeVe~fahrenstechnik,UniversitiitKarlsruhe,Germany,1995
20. P.Ngendakumana,Mesuredesgrandeursthermofluides,UniversitdeLige,Facultdes
sciencesappliques,20052006
21. P.Ngendakumana,Echangeursdechaleur,UniversitdeLige,Facultdessciences
appliques,20052006
22. J.G.PERSSON,Screwexpandersforsmallscalecogeneration,VDIBerichte,nr.1135,
Duseldorf,1994
23. P.Platell,Displacementexpandersforsmallscalecogeneration,Licenciatethesis,Royal
InstituteofTechnology,Stockholm,1993
115
24. S.Quoilin,Lescentralessolairesconcentration,Travaildanslecadreducoursd'nergies
renouvelables,UniversitdeLige,2007.
25. D.M.RoweandGaoMin,Evaluationofthermoelectricmodulesforpowergeneration,
JournalofPowerSources,Volume73,Pages193198,1998
26. EFThacher,BTHelenbrook,MAKarri1,andCJRichter,Testingofanautomobileexhaust
thermoelectricgeneratorinalighttruck,J.AutomobileEngineering,Proc.IMechEVol.221
PartD,2007
27. J.Vazquez,M.ASansBobi,R.Palacios,A.Arenas,StateoftheArtofThermoelectric
GeneratorsBasedonHeatRecoveredfromtheExhaustGasesofAutomobiles,Escuela
TecnicaSuperiordeIngenieria,Madrid,2002
28. E.L.WINANDY,Contributiontotheperformanceanalysisofreciprocatingandscroll
refrigerationcompressors,PhDThesis,UniversityofConcepcin,Chile,1999
29. YangJ.,Opportunities&ChallengesofThermoelectricWasteHeatRecoveryinthe
AutomotiveIndustry.2005DieselEngineEmissionsReduction(DEER)Conference,Chicago,
2005
30. R.ZanelliandD.Favrat,Experimentalinvestigationofahermeticscrollexpandergenerator.
SwissFederalInstituteofTechnologyofLausanne,1994
116
SylvainQuoilin Appendix1:Othersolutionsforlowgradeheatrecovery
Appendix1:Othersolutionsforlowgradeheatrecovery.
1.Thetranscriticalcycle
Thetranscriticalcycleworkswithafluidinsupercriticstateintheheatsourceexchanger.Ithasthe
advantageof a goodtemperatureprofile(nopinchlimitation)andhasthereforelessirreversibility
(figure62).
Theworkingfluidusuallyusediscarbondioxide.Itsmaindrawbackisthepressure:theelementsof
thecyclearesubmittedtoaveryhighpressure,whichcancauseunacceptablestressesandleaks.
ChenetAlproposedacomparativestudybetweenatranscritialCO2cycleandatraditionalORCwith
R123asworkingfluid[Chenetal,2002].ThecalculationswereperformedusingEES,followingthe
configurationshowedinfigure61.Thehotsidetemperaturewassetto150Candthemassflowrate
to0,4Kg/s.
TheresultsofthesimulationshowedthattheCO2cyclehasaslightlyhigherefficiencythantheORC.
Further,thepowersystemwithCO2isalsomorecompactandmoreenvironmentalfriendlythanthe
onewithR123.
117
Figure62:R123(left)vsCO2(right)heattransfer
Figure61:CO2transcriticalcycle
2.Thewaterammoniacycle
Theparticularityofthiscycleisitsnonazeotropicworkingfluid.Itshowsagoodtemperatureprofile
intheexchangersandhasthereforelessthermalirreversibilities.
Thenonisothermalevaporationorcondensationisduetothepropertiesofbinarymixtures:
ThefluidiscomposedofNH3andwater(infigure63,xwouldbethemolefractionofwaterinNH3).
Whenheatedfrompoint1topoint3,thefluidisfirstheatedinliquidphase,andthenentersthetwo
phase state where it is evaporated at a nonconstant temperature, modifying the concentration of
ammoniainthevaporandliquidphases(figure63).
Figure64showstheadvantageofthenonazeotropicfluidoverthetraditionalonecomponentfluid.
Thisadvantageisvalidintheevaporatoraswellasinthecondenser.
118
Figure63:Evaporationofthewaterammoniamixture
Anotheradvantageistheflexibilityofthecycle:TheNH3concentrationcanbeadaptedtogetthe
desiredboilingtemperature.TheboilingtemperatureatatmosphericpressureofpureNH3isindeed
33Candthatofwateris100C. IncreasingthefractionofNH3willthereforereducetheboiling
temperature.
ThebestknowncycleusingtheammoniawatermixtureistheKalinacycle. Ithasbeenshownina
numberofstudiesthattheammoniawatercyclecangeneratemorepowerthantheRankinecyclefor
certainkindsofheatsources.
Fore Instance, Jonsson and Yan, 2000 studied the ammoniawater bottoming cycle designed for
exhaustgasesheatrecoveryondifferentenginetypes.Dependingontheconfiguration,theyshowed
thattheKalinacyclecouldincreasetheoutputpowerby5to50%withanexhaustgasestemperature
ofabout400C.
119
Figure64:Heatprofileintheevaporator
Figure65:WorkingprincipleoftheKalinacycle
Thiscycle,initsbasicconfiguration,iscomposedoftwopumpsandtwocondensers.TheNH3/H20
weightratiovariesthroughthecycle:
Inordertohavealowevaporatingpressure,theconcentrationofammoniaintheevaporatorhastobe
relativelyhigh.Themixturepassingthoughtheevaporatoratsuchconcentrationiscalledtheworking
mixture. The workingmixture is then expandedandhastobecondensed. DuetoitshighNH3
concentration,the condensation is not possible at the usual cooling fluid temperatures and at the
expanderexhaustpressure. ItishencenecessarytoreducetheNH3concentration.Thisisdoneby
mixingtheworkingmixturewithanNH3leansolutiontoobtainthesocalledbasicmixture. This
basicmixtureiscondensed,andpressurizedwiththeLPpump.
Torecovertheworkingmixturefromthebasicmixture,thelatterhastobereenriched.Thisisdone
bymixingitwithNH3enrichedvapor.Thenewworkingmixtureisnextcondensedasecondtime(at
ahigherpressure)andthenpressurizedtoreturntothebeginningoftheloop.
InordertoobtaintheNH3leansolutionandtheNH3enrichedvapor,apartofthebasicmixtureis
tappedaftertheLPpumpandpassesthroughthereheatertoreachthetwophase state(point2in
figure63).Intheseparator,thewaterenrichedsolution(NH3leansolution)andthewaterleanvapor
(NH3enrichedvapor)areseparatedandsentbacktotheirrespectivemixingvalves.
Several configurations can be conceived for this cycle : various heat sources can be used, heat
exchangerscanbemultiplied,withoneorvariousseparators,etc.
ThemaindrawbackoftheKalinacycleisthemultiplicationofthecomponents:atleastfourheat
exchangersarerequired,aswellastwopumpsandoneseparator.Thisincreasestheweightandthe
size of the facility and hence makes it more difficult to incorporate in mobile applications (e.g.
vehicles).Thecostincreasesaswellwiththenumberofcomponents.
Another nonnegligible drawback is the high toxicity of ammonia. The system has to be highly
secured,inordertoavoidanyriskofleakage.
120
3.TheStirlingcycle
TheStirlingcycleisrepresentedinfigure66.Itcomprises4transformations:
12:Isothermalcompression
23:Isochoricheating
34:Isothermalexpansion
41:Isochoriccooling
The ideal Stirling cycle presents a very good efficiency : the expansion being isothermal, the
efficiencyistheCarnotefficiency. Thisefficiencyisachievedattheexpenseofalowerspecific
work,thepvdiagrambeingflatter.
TheStirlingengineisdescribedinFigure67.
121
Figure67:WorkingprincipleoftheStirlingengine
Figure66:IdealStirlingcycle
ThemaincomponentsoftheStirlingengineare:
Powerpiston:Performsthecompressionandtheexpansionoftheworkingfluid
Displacerpiston:Itsroleistointerrupttheinfluenceofthehotsourceorthatoftheheatsink
Regenerator:storestheheatwhentheworkingfluidmovesdownandresituatesitwhenit
movesup.
Hotsideheatexchanger:Heatsourceofthecycle,continuouslyheated(
T
h )
Coldsideheatexchanger:Heatsinkofthecycle,continuouslycooled(
T
c )
TheStirlingengineischaracterisedbyaworkingfluidtotallyisolatedfromitsenvironment.
This working fluid is usually hydrogen because of its low viscosity, good specific heat and high
thermalconductivity,butcanbereplacedwithhelium,whichhastheadvantageofbeinginert(not
explosive)anddoesn'tdiffusethoughthemetallicshelloftheengine.
Realcycle:
TheStirlingcycleisdifficulttoachieveinpractisebecauseitinvolvesreversibleheattransferinallthe
components,includingtheregenerator.Thiswouldrequireprovidinganinfinitelylargeheatexchange
areaandinfiniteheattransfertime.
Becauseoftheinefficiencies,Stirlingcycleshavebeenofonlytheoreticalinterestforalongtime.
Nowadays,thereisanewinterestinthiscycle,withsomeapplicationsalreadycommercialized.
Applications:
Stirlingcyclecanbeusedforsmallscalecogeneration. Forexample,WhisperGen,aNew
Zealandfirm,developsan"ACMicroCombinedHeatandPower"Stirlingcycleengine.These
microCHPunitsaregasfiredcentralheatingboilerswhichsellpowerbackintotheelectricity
grid.
Concentrating solar power : Stirling engines have been used in various dish solar
collectors,theenginebeingsituatedatthefocusoftheparabolicdish.
Otherapplicationssuchasgeothermalenergyrecovery,cryogenicapplications,heatpumping
areenvisagedbutnotyetcommerciallyavailable.
122
4.TheEricssonCycle:
TheEricssoncycleisverysimilartotheStirlingcycle.Itcomprises4transformations:
12:Isothermalcompression
23:Isobaricheating
34:Isothermalexpansion
41:Isobariccooling
Figure68:Ericssonidealcycle
Thiscycle,liketheStirlingcycle,usesaregeneratorforthetransformations23(isobariccooling)and
41 (isobaric heating). The isothermal compression and expansion are achieved by permanently
cooling(resp.heating)thecompressor(resp.turbine).
The efficiency of the ideal cycle is the Carnot efficiency but, like the Stirling cycle, it involves
reversibleheattransfers,whichisnotpracticallyfeasible. Aparticularlydifficultissueisaproper
heattransferatthecompressorandturbinelevelbetweentheworkingfluidandtheheatsources.
123
Figure69:WorkingprincipleoftheEricssonEngine
5.TheThermoelectriceffect.
Thethermoelectriceffect,alsocalledtheSeebeckeffect,isthedirectconversionofheatenergyinto
electricenergy.
Whenasemiconductorissubmittedtoatemperaturegradient,heatflowsthoughitandputsinmotion
the chargecarriers : the holes for the ptypeand theelectronsforthe ntype runintheopposite
direction,tocreateanentropyfluxopposedtotheonecreatedbytheheatflux.Thecurrentinthen
typesemiconductorishenceinthesamedirectionthantheheatflux,andintheoppositedirectionfor
the ptype. Therefore, connecting the semiconductors as indicated on figure 70 is equivalent to
connectingasetofcurrentgeneratorsinseries.
Figure70:Workingprincipleofthethermoelectricgenerator
The thermoelectric effect is characterised by the Seebeck coefficient, which formulates the
electromagneticforcebetweentwodifferentmaterialssubmittedtoatemperaturegradient:
S
ab
=
dV
dT
=S
a
S
b
Theperformancesofthethermoelectricconversionfromheattoelectricityareexpressedbya non
dimensionalcoefficientZTcalledfigureofmerit.Thefirstthermoelectricmaterialshadafigureof
meritofabout1to1.5. Nowadays,theusualvaluesrunfrom2to3andresearchershopetoreach
ZT=4inaclosefuture.
124
Figure71showsthechronologicalevolutionoftheZTparameterandtheperspectivesforthiskindof
systemswiththeprogressofmaterialscience[Yang,2005].
Thetheoreticalefficiencyofthethermoelectricconversionisgivenby:
TheefficiencyisequaltotheCarnotefficiencymultipliedbyasecondefficiencyfactor
n
seeb
Figure76showstheevolutionofthetheoreticalthermoelectricefficiencywiththeZTforahotside
temperature
T
h
=500K
andacoldsidetemperature
T
c
=300K
125
Figure71:ChronologicalevolutionofZT
Figure72:ThermoelectricconversionefficiencyvsZT
ThethermoelectricefficiencyremainsfarbelowtheCarnotefficiencywithvaluesrunningfrom0.06
to0.15fortheusualZTrangeofvalues.
Applications:
Thermoelectricmoduleshavelongbeenusedonlyforcoolingapplications.Nevertheless,anumber
ofstudiesshowedthatwiththerecentimprovements,theycouldalsobeusedforheatrecoveryatan
acceptablecost[Rowe,1998].
Theapplicationsconsideredareaimedatenginesexhaustgasesheatrecovery.
A complete review of several existing thermoelectric systems applied to vehicles is proposed by
Vazquezetal[Vazquez,2002].
Table5showstheexperimentalresultssummarizedinhispaper:
Testdescription Max electric
power
[W]
Max
efficiency
[%]
Inlettemperature
[C]
Weight
[kg]
Nissan,3000cc,60km/hclimbhill 35,6 0,9 499C
23lengine,60km/hclimbhill 193 2,9 583C
Heavy truck, engine NTC 350, 300
hp,1700rpm
1068 44kg
Takanose, 65 km/h climb hill, 3244
rpm
131,5 2,3 866
Rustonengine3YDA,38hp 42,3 600
Table5:Thermoelectricperformancesofvehicleapplications
AnotherexperimentalstudywasperformedonaSierraPickuptruckbyThatcher etAl [Thatcher,
2006].Theefficienciesofthemodules(electricpowervsexhaustgasesextractedpower)ranfrom1.2
to1.5,from2.1to2.2andfrom2.8to2.9%respectivelyforspeedsof30,50and70mphandinlet
gastemperaturesrangingfrom350to600C.
These results show that the efficiencies reached by the thermoelectric modules are quite low.
Nevertheless,animportantpotentialofimprovementremains,thatcouldincreasetheseefficiencies.
126
Appendix2:Firstseriesoftests
127
Test Air flow rate Pump setting
[kg/s] [kg/s] [kg/s] [C] [C] [C] [K] [K]
061013A 0.073 0.093 0.587 0.8 14.78 155.1 185 -18.06 20.08
061017A 0.074 0.093 0.592 0.8 14.7 154.6 184.5 -17.72 17.04
061020A 0.065 0.092 0.588 0.7 14.37 158.2 187.8 -22.02 34.3
061020B 0.064 0.092 0.412 0.7 14.37 138.2 176.5 -19.43 18.58
061024A 0.064 0.092 0.586 0.7 14.38 156.6 186.2 -20.09 32.9
061024B 0.064 0.092 0.411 0.7 14.34 136.9 176.9 -20.1 17.91
061025A 0.064 0.092 0.589 0.7 14.32 157.1 187.2 -22.34 34
061025B 0.064 0.091 0.414 0.7 14.28 137 178.3 -21.52 19.65
061027A 0.063 0.091 0.327 0.7 14.54 156.1 188.5 -20.99 34.39
061027B 0.061 0.091 0.208 0.7 14.45 135.2 178.6 -20.82 21
061031A 0.066 0.096 0.255 0.8 14.4 159.2 185.9 -15.83 32.32
061031B 0.067 0.095 0.252 0.85 14.38 159.9 187.7 -16.72 31.26
061031C 0.068 0.096 0.249 0.9 14.36 160.4 188 -16.08 31.64
Test
[C] [C] [W] [W]
061013A 121.7 835615 1025 686 0.583 0.049 0.328 0.151 3.63
061017A 118.4 827503 1010 668.9 0.580 0.048 0.327 0.148 3.62
061020A 131.7 761584 856.6 530.2 0.558 0.045 0.334 0.134 3.36
061020B 114.1 729325 733.4 410.8 0.537 0.040 0.301 0.134 3.21
061024A 130.1 757731 850.8 522.9 0.564 0.045 0.331 0.135 3.34
061024B 113.5 728960 736.5 418.9 0.539 0.041 0.299 0.137 3.21
061025A 131.5 763748 932.8 601.1 0.591 0.051 0.332 0.153 3.5
061025B 115.1 727003 826.4 502.8 0.557 0.048 0.299 0.161 3.5
061027A 131.9 764184 886.5 575.6 0.575 0.048 0.330 0.146 3.48
061027B 115.8 717174 710.7 428.7 0.530 0.042 0.296 0.141 3.27
061031A 132 800923 1034 706.9 0.619 0.053 0.335 0.159 3.68
061031B 131.8 817999 1067 758.1 0.616 0.054 0.336 0.160 3.76
061031C 132.4 819594 1066 742.4 0.610 0.053 0.337 0.157 3.77
Refrigerant
flow rate
Water flow
rate in the
condenser
Cold source
temperature
Second hot
air source
temperature
First hot air
source
temperature
Subcooling
at the
condenser
exhaust
Superheating
at the
evaporator
exhaust
Refrigerant
temperature
at the
expander
supply
Refrigerant
pressure at
the expander
supply
Shaft output
power
Asynchronous
machine
output power
Expander
isentropic
effectiveness
Cycle
efficiency
Carnot
efficiency
Exergetic
efficiency
Pressure ratio
over the
expander
Appendix3:Secondseriesoftests.
128
[kg/s] [kg/s] [kg/s] [C] [C] [C] [K] [K]
061107 A 0.065 0.107 0.247 0.8 13.41 164.5 184.6 -17.79 36.01
B 0.060 0.107 0.247 0.7 13.39 165.5 185.9 -19.52 47.15
061108 A 0.051 0.108 0.190 0.6 13.12 164.6 184.5 -17.94 62.52
B 0.054 0.108 0.192 0.7 13.1 164.2 184.2 -16.48 57.35
C 0.058 0.108 0.192 0.8 13.09 164.3 185.1 -16.26 53.34
D 0.054 0.107 0.185 0.9 13.08 165.4 186.2 -16.29 60.09
E 0.053 0.108 0.185 1 13.06 165.7 186.3 -16.32 61.26
061110 A 0.049 0.108 0.391 0.7 13.07 166.7 185.6 -19.79 61.29
B 0.045 0.108 0.329 0.7 13.06 135.9 172.6 -20.55 44.38
C 0.045 0.108 0.381 0.7 13.04 138.3 174.4 -20.94 46.31
D 0.048 0.107 0.418 0.7 13.03 167.8 187.6 -20.84 64.55
E 0.047 0.107 0.540 0.7 13.03 168.3 188.1 -21.48 65.15
F 0.049 0.107 0.704 0.7 13.03 168.7 188.4 -21.93 65.62
G 0.048 0.107 0.126 0.7 13.03 170.4 189.8 -15.47 66.51
H 0.047 0.107 0.166 0.7 13.03 170.1 189.7 -18.16 66.93
I 0.048 0.107 0.226 0.7 13.03 170.1 189.8 -20.3 67.18
J 0.045 0.107 0.226 0.5 13.03 174.7 192.3 -22.72 73.73
061114 A 0.052 0.123 0.217 0.8 12.71 179.1 189.3 -16.57 75.09
06AA21 A 0.044 0.096 0.217 0.8 12.38 170.6 190.4 -16.91 73.55
[C] [C] [W] [W]
061107 A 135 788000 1034 685.2 0.626 0.049 0.345 0.143 3.62
B 143.4 747967 933.8 602 0.602 0.048 0.347 0.139 3.60
061108 A 152.5 650396 708.7 373.9 0.535 0.040 0.346 0.117 3.44
B 148.8 673685 798.5 448 0.553 0.043 0.346 0.125 3.56
C 146.1 692323 856 507 0.551 0.045 0.346 0.129 3.66
D 151.3 671189 821.8 475.7 0.550 0.043 0.348 0.123 3.77
E 152.2 667834 843.7 491 0.573 0.043 0.348 0.124 3.75
061110 A 157.7 762394 1015 657.8 0.674 0.061 0.349 0.176 4.22
B 134.8 660733 780.7 454.7 0.667 0.055 0.301 0.184 3.66
C 137 666315 803.4 467.1 0.673 0.056 0.305 0.185 3.69
D 160.3 752295 988.3 640.2 0.680 0.061 0.351 0.173 4.16
E 160.9 751929 995 650.1 0.697 0.061 0.352 0.173 4.16
F 161.3 751032 995.5 653.8 0.666 0.061 0.352 0.173 4.16
G 162.7 759006 879.7 548.9 0.666 0.053 0.355 0.149 3.60
H 162.7 753457 902.5 569.6 0.666 0.055 0.355 0.154 3.75
I 162.8 749936 928.7 595 0.657 0.057 0.355 0.160 3.93
J 167 715351 754.6 428.6 0.631 0.050 0.361 0.137 3.39
061114 A 168.2 706340 991.7 620.1 0.657 0.052 0.368 0.142 3.75
06AA21 A 160.1 613113 765.2 406.6 0.609 0.048 0.357 0.136 3.67
Refrigerant
flow rate
Water flow
rate in the
condenser
Cold air
source
temperature
Second hot
source
temperature
First hot air
source
temperature
Subcooling
at the
condenser
exhaust
Superheating
at the
evaporator
exhaust
Refrigerant
temperature
at the
expander
supply
Refrigerant
pressure at
the expander
supply
Shaft output
power
Asynchronous
machine
output power
Expander
isentropic
effectiveness
Cycle
efficiency
Carnot
efficiency
Exergetic
efficiency
Pressure ratio
over the
expander
Appendix4:Thirdseriesoftests
129
[kg/s] [kg/s] [kg/s] [C] [C] [C] [K] [K]
010507A 0.046 0.077 0.719 0.5 12.24 140 177.2 -21.14 38.84
B 0.057 0.076 0.719 0.75 12.12 138.1 176.8 -22.38 19.93
C 0.071 0.091 0.719 1 12.04 151.6 188.1 -23.78 21.9
D 0.066 0.091 0.719 0.75 12.01 152.6 188.5 -24.94 29.83
E 0.054 0.091 0.719 0.6 11.98 156.7 190.7 -25.46 52.92
F 0.054 0.106 0.719 0.6 11.95 161.6 188.3 -27.37 60.33
G 0.068 0.106 0.719 0.8 11.94 159.4 186.7 -26.53 37.06
H 0.071 0.106 0.719 1 11.93 158.4 186.4 -25.63 33.27
030507A 0.055 0.122 0.717 0.6 11.27 173.4 188.8 -35.3 71.48
B 0.070 0.122 0.722 0.8 11.23 169.3 187.7 -34.63 40.49
C 0.072 0.122 0.723 1 11.21 168.8 187.8 -33.98 36.49
D 0.074 0.122 0.722 1 11.21 168.5 187.7 -33.79 38.88
E 0.073 0.122 0.718 1 11.19 120.1 183.3 -34.26 10.22
F 0.070 0.122 0.713 1 11.18 141.6 185.3 -33.96 23
G 0.063 0.121 0.717 0.7 11.17 144 186.4 -35.62 37.92
H 0.062 0.121 0.722 0.7 11.16 173.2 190.2 -35.8 58.32
I 0.056 0.106 0.717 1 11.11 157.9 185.9 -16.57 52.97
J 0.070 0.106 0.721 1 11.08 155.4 185.5 -15.96 31.25
K 0.081 0.106 0.722 1 11.06 151.7 184.7 -15.51 8.86
L 0.082 0.106 0.714 1 11.04 151.2 184.4 -15.71 4.16
M 0.072 0.106 0.716 0.8 11.02 152.9 185.3 -17.86 27.37
N 0.085 0.121 0.721 1 10.98 173.8 189.1 -17.53 31.89
O 0.074 0.106 0.719 0.89 10.96 127.4 181.9 -17.75 3.19
P 0.072 0.106 0.715 0.87 10.94 126.5 181.7 -17.83 5.6
040507A 0.062 0.107 0.859 0.7 11.66 154.1 189.1 -20.11 34.28
B 0.062 0.106 0.863 0.7 11.62 154.3 189.9 -20.83 32.83
C 0.063 0.106 0.858 0.7 11.6 154.1 190.2 -21.11 31.62
D 0.063 0.106 0.865 0.7 11.57 154.1 190.6 -21.48 31.91
E 0.063 0.106 0.864 0.7 11.56 154.2 190.9 -21.78 31.79
F 0.061 0.106 0.863 0.7 11.5 155.2 191.6 -20.23 37.14
G 0.058 0.106 0.864 0.7 11.51 155.9 191.8 -19.51 44.16
050507A 0.065 0.107 0.898 0.75 12.58 149.9 183.1 -18.18 26.08
B 0.054 0.107 0.896 0.6 12.24 153.7 185.5 -19.07 45.83
C 0.045 0.107 0.900 0.5 12.15 162.7 190.2 -17.74 69.86
D 0.082 0.121 0.901 1 12.1 171.9 188.6 -19.82 26.41
E 0.081 0.111 0.902 1 12.05 166.2 189.5 -20.43 17.9
F 0.086 0.121 0.842 1 12.03 172.1 187.9 -19.93 31.36
G 0.071 0.106 0.841 0.8 12.15 156.6 185.8 -20.25 33.67
H 0.054 0.106 0.839 0.6 12.02 160.2 187.8 -18.33 63.64
Test
[C] [C] [W] [W]
010507A 121 540000 382.1 74 0.43 0.026 0.309 0.084 2.7
B 109.8 640000 736.9 410 0.58 0.049 0.307 0.158 3.37
C 122.6 820000 1188 800 0.64 0.062 0.329 0.189 4.1
D 128.1 780000 1065 703 0.64 0.058 0.330 0.175 3.73
E 143.9 669890 737.4 400 0.57 0.044 0.337 0.132 3.19
F 152.2 680000 767.2 401 0.57 0.043 0.344 0.126 3.24
G 136.5 800000 1184 770 0.66 0.059 0.341 0.172 3.81
H 134.4 830000 1275 863 0.65 0.063 0.340 0.185 4.15
030507A 165.2 719611 801.8 428 0.58 0.042 0.363 0.116 3.14
B 141.9 833966 1259 850 0.66 0.055 0.357 0.153 3.9
C 139 851360 1339 921 0.67 0.057 0.357 0.161 4.09
D 142.7 879402 1404 980 0.67 0.061 0.356 0.173 4.16
E 110.1 802622 1134 738 0.64 0.056 0.277 0.200 3.88
F 123.2 810026 1191 790 0.65 0.056 0.315 0.177 4.03
G 134.6 755426 1021 636 0.63 0.052 0.319 0.163 3.67
H 156.4 783801 1150 755 0.65 0.056 0.363 0.154 3.83
I 145.6 694675 1154 758 0.63 0.068 0.341 0.199 5.03
J 132.2 825565 1413 992 0.65 0.071 0.337 0.211 4.92
K 112.9 877928 1446 1021 0.65 0.062 0.331 0.188 4.73
L 107.7 868783 1426 1001 0.65 0.059 0.331 0.178 4.61
M 129.2 840134 1401 981 0.66 0.069 0.333 0.207 4.74
N 141.6 1.00E+006 1769 1306 0.66 0.070 0.364 0.192 4.98
O 101.7 774893 1189 797 0.62 0.057 0.291 0.196 4.45
P 103.8 773085 1188 794 0.62 0.059 0.289 0.204 4.52
040507A 130.1 736310 1180 785 0.65 0.058 0.334 0.174 4.6
B 128.8 740628 1167 775 0.65 0.056 0.334 0.168 4.44
C 127.9 744804 1106 721 0.64 0.052 0.334 0.156 4.13
D 128.1 744158 939.2 569 0.61 0.043 0.334 0.128 3.51
E 128.1 745060 647.3 301 0.51 0.026 0.334 0.078 2.8
F 132.5 730526 1164 776 0.65 0.057 0.336 0.170 4.44
G 138 709944 1151 762 0.65 0.059 0.337 0.177 4.58
050507A 130.3 890365 1394 1011 0.68 0.074 0.325 0.228 5.28
B 143.5 778760 1167 811 0.67 0.070 0.332 0.210 4.94
C 160.6 675326 974.7 640 0.65 0.070 0.346 0.201 4.69
D 141.2 1.11E+006 1820 1391 0.68 0.073 0.359 0.203 5.39
E 130.8 1.06E+006 1708 1294 0.68 0.071 0.351 0.202 5.26
F 138.9 951243 1648 1239 0.65 0.064 0.360 0.178 4.54
G 132.5 787230 1269 902 0.63 0.062 0.336 0.184 4.28
H 152.4 638289 934.3 601 0.59 0.056 0.342 0.163 4.09
Refrigerant
flow rate
Water flow
rate in the
condenser
Cold source
temperature
Second hot
air source
temperature
First hot air
source
temperature
Subcooling
at the
condenser
exhaust
Superheating
at the
evaporator
exhaust
Refrigerant
temperature
at the
expander
supply
Refrigerant
pressure at
the expander
supply
Shaft output
power
Asynchronous
machine
output power
Expander
isentropic
effectiveness
Cycle
efficiency
Carnot
efficiency
Exergetic
efficiency
Pressure ratio
over the
expander

Tfe SQ010607

Загружено:

Сведения о документе

Исходное описание:

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Tfe SQ010607

Загружено:

Авторское право:

Доступные форматы

ExperimentalStudyandModelingofaLow

Swept volume of the scroll machine, in compressor mode. This volume is

Вам также может понравиться