Sei sulla pagina 1di 27

6/14/2017 Chapter7.

ModellingInductionMotors

Chapter7.ModellingInductionMotors
Prev Next

Chapter7.ModellingInductionMotors
TableofContents
7.1.TheACInductionMotor
7.2.GeneralequationsofACMotors
7.2.1.Machineequationsinanaturalcoordinatesystem
7.2.2.MachineEquationsinacommoncoordinatesystem
7.3.ModifiedEquivalentCircuits
7.4.SettinguptheModel
7.5.StatorCoordinates
7.6.FieldCoordinates
7.7.ModelassumptionsfortheFOCmethod
7.7.1.DecouplingoftheNonlinearity
7.7.2.TheMatlab/SimulinkModeloftheACMotor
7.8.VectorversusScalarcontrol
7.8.1.ScalarControl
7.8.2.VectorControl
7.8.2.1.TheFOCAlgorithmStructure
7.9.Summary
7.9.1.TheMatlab/SimulinkModeloftheFieldOrientedControl

7.1.TheACInductionMotor
Asynchronousmotorsarebasedoninduction.Theleastexpensiveandmostwidelyspreadinductionmotoristhesocalledsquirrelcage
motor.AtypicalasynchronousmotorisshowninFigure71.Ametalringattheendsresultinginashortcircuitconnectsthewiresalong
therotoraxis.Thereisnocurrentsupplyneededfromoutsidetherotortocreateamagneticfieldintherotor.Thisisthereasonwhythis
motorissorobustandinexpensive.Thestatorphasescreateamagneticfieldintheairgaprotatingatthespeedofthestatorfrequency
(1).

Figure7.1.Anasynchronousmotor

Thechangingfieldinducesacurrentinthecagewires,whichthenresultsintheformationofasecondmagneticfieldaroundtherotor
wires.Asaconsequenceoftheforcescreatedbythesetwofields,therotorstartsrotatinginthedirectionofthestatorfieldbutata
slowerspeed().Iftherotorrevolvedatthesamefrequencyasthestator,thentherotorfieldwouldbeinphasewiththestatorfieldandno
inductionwouldbepossible.Thedifferencebetweenthestatorandtherotorfrequencyiscalledslipfrequency(slip=1).Thereareseveral
waystocontrolaninductionmotorintorque,speedorposition,whichcanbecategorizedintotwogroups:therearescalarandvector
controlmethods.BecauseofitsoutstandingrobustnessamongallACMotors,ascanbeseenintheFigure72.,theasynchronous
squirrelcageinductionmotorwaschosenfordetailedanalysisandforcontrolpurposes.Further,theexperimentalapplicationofthis
motortypehasgrowingsignificance.

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 1/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.2.Classificationofelectricmotors

Thefollowingsectionwillbrieflydescribethismotorwiththepurposeofadeeperinsightintoitsbehaviorlateron.

7.2.GeneralequationsofACMotors
TheasynchronousorACinductionmotorisrepresentedbyastatorandaconcentricrotorwiththreephasewindingsandanarrowairgap
betweenthesmoothsurfacesoftherotorandthestator.Onlythestatorwindingsarefedbyavoltageorcurrentsourceinverter.The
connectionbetweenthestatorandtherotorismadebytheflux,whichmeansthereisnomechanicalconnectionbetweenthemapartfrom
thebearings.FordeducingthegeneralequationsoftheACinductionmotor(withsquirrelcagerotor),letasmakethefollowing
assumptions:

Figure7.3.SinusoidalCurrentdensitydistribution

Presumeasymmetricalthreephasewindingssysteminthemachine.
neglectthecooperlossesandtheslotsinthemachine
spatialdistributionoffluxesandamperturnswaveareconsideredsinusoidal,ascanbeseeninFigure73,thecrosssectionofan
InductionmotorisshowninFigure74.
allthecalculationsbasedonthethreephasedvectortheory
allthelossesduetowiring,saturation,andsloteffectscanbeneglected
statorandrotorpermeabilityareassumedtobeinfinite.

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 2/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.4.CrosssectionofanInductionMotor

7.2.1.Machineequationsinanaturalcoordinatesystem
Thevoltageequationsofthestatorarethefollowing:

(7.1)

(7.2)

(7.3)

where arethetotalfluxesofeachphaseincludingthemainfieldofthewindingsandtheleakagefluxconnectionwiththe
otherwindings,and istheonephasewindingresistance.Ifwedescribetheaboveequations(7.1)(7.2)(7.3)according
tothevectortheory,thestatorvectorialvoltageequationwillbethefollowing:

(7.4)

(7.5)

wherethelatterequationisthezerosequencecomponent.

Figure7.5.Threephasestatorandrotorwindingsofanasynchronousmotorinthenaturalcoordinatesystem

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 3/27
6/14/2017 Chapter7.ModellingInductionMotors
Inpractice,thezerosequencecomponentcanbeneglected,sotheequation(7.5)canbeignoredandonlyonevectorialequationwill
describethemotorbehaviorinsteadofthreephaseequations.Writingtheequationforrotorwindingsandproceedinglikebeforewecan
easilyobtainthevectorialvoltageequationoftherotor:

(7.6)

(7.7)

where,forthesamereasonasbefore,theequation(7.7)canbeignored.Thecurrentscirculatinginthethreestatorandthreerotor
windingsgeneratethefluxesinthemotor.Wesupposethattheairgapisconstantandthespatialdistributionofthemainmagneticfieldis
sinusoidal,sothemutualinductanceisproportionalwiththecosinevalueoftheanglebetweentwowindings.Letuschoosetherealaxes
ofthecomplexplanetomatchwiththeaphaseaxes.Inthefollowingdiscussion,wewillcallNaturalCoordinateSystemthecomplex
coordinatesystemfixedinthiswaytotheabcaxesofthemotor.Accordingtothisagreement,thefluxofphaseaofthestatorflux
vectorwillbegivenbytheequation(7.8),(7.9):

(7.8)

(7.9)

Themotorsgeneralequationsaredefinedby(7.4),(7.6),(7.8),(7.9).If aregivenandthetimefunction isalsogiven,thenthe


aboveequationsunambiguouslydefinethemotorcurrents.Theyformalineardifferentialequationsystemwithtimedependent
coefficients.If isnotknownornotgivenandmustbecalculatedfromthetorqueormotionequation,thentheaboveequationsystem
becomesnonlinear.The factorfromthefluxequationsshouldbeeliminatedifthestatorandrotorquantitiesareevaluatedinthesame
coordinatesystem.

7.2.2.MachineEquationsinacommoncoordinatesystem
Intheprevioussectionthestatorquantitiesweregivenintheirowncomplexcoordinatesystemwiththerealaxesfixedtotheaphase
axesofthestatorwindings.Letthe functiondescribetheangularpositionofthenewcommoncoordinatesystem,referringtothe
fixedaphaseaxes.Inthisnewcommoncoordinatesystemthe statorcurrentvectorangle,writtenasacomplexnumber,became
insteadoftheformer .Sothisanglecanbeexpressedas*=xk.

Figure7.6.Thecoordinatesystemsformachineequationtransformation

Nowifwedenotethecurrentvectorwith inthenewreferenceframe,thenthestatorquantitiescanbeexpressedas

(7.10)

(7.11)

Theanglebetweenthetwo(fixedtothestator,common)coordinatesystemsisxkx,sotherotorquantitiescanbeexpressedas

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 4/27
6/14/2017 Chapter7.ModellingInductionMotors
(7.12)

(7.13)

Havingexpressedtheoldquantitiesofthefluxequation,inthenewcoordinatesystem,afterafewsimplifications,weobtainthefollowing
fluxequationsinthenewcommoncoordinatesystem:

(7.14)

(7.15)

Ifwelookcarefullyattheseequations,wecanobservethatthetimedependentcoefficientsarenotpresentintheequationsanymore.So
thefluxequationsoftheasynchronousmotorbecameassimpleasinthecaseofatransformer.Letushavealookatthevoltage
equations,becausetheyhavebecomeslightlycomplicated:

(7.16)

(7.17)

wheredx/dt=istherotorspeedorangularvelocity,anddxk/dt=kisthespeedorangularvelocityofthecoordinatesystem.Bothofthese
shouldbetimedependentvariables.

(7.18)

(7.19)

Lookingattheseequationswecanseethatthetimedependentcoefficientsarepresentinthevoltageequations,dueto .
Supposingthatthespeedoftherotorinmostcasesisconstantorcannotchangedramatically,andchoosingacoordinatesystemwitha
constantspeedweobtainaconstantcoefficientlineardifferentialequationsystem.Lookingatthevoltageequation,itcanbeobserved
thatthefluxhastwocomponents.The componentisthetransformationcomponent,andthefollowing componentisthe
inducedvoltage,resultingfromtherotationandiscommonlycalledtherotationalvoltage,duetotherelative speedofthecoordinate
system,relativetothewindings.Whichpartoftheinducedvoltage writtenformerlyinthenaturalcoordinatesystemis
stemmingfromtransformationalvoltagesandwhichpartisstemmingfromrotationalvoltagesdependsonthechosencommoncoordinate
systemandissubjecttotheactualpointofview.Neglectingthe*symbol,bysupposinganarbitrarycommoncoordinatesystem,the
motorequationsbecomethefollowing:

(7.20)

(7.21)

(7.22)

(7.23)

Inmostcases,accordingtothenatureoftheapplication,wechoosethecommoncoordinatesysteminthreecases:

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 5/27
6/14/2017 Chapter7.ModellingInductionMotors
1.Thecoordinatesystemisatastandstill(fixedtothestator).Inthiscase andthecrossingcomponentwillbe onlyin
therotorvoltageequation,becausetherotorrotatesatthespeed ,so,fromthepointoftherotor,thestatorrotatesbackatthespeed
.

2.Thecoordinatesystemisfixedtotherotor,itrotateswiththesamespeed .Inthiscasewehaveacrossingcomponent
onlyinthestatorvoltageequation.Thisischoseninthecaseofsynchronousmachines,whichisconvenient
becauseoftheasymmetricalaspectoftherotor:

(7.24)

(7.25)

Ifwewritetheseequationsincomplexfrom,usingthefollowingequations: ,wegetthe
followingequations:

(7.26)

(7.27)

ThesearetheParkequations,whichareequivalentwiththevectorialequationof(7.24).Itisimportanttoobservethatinthedaxes
componentwehaveafactor ,duetotheimaginarycomponentoftheequation(7.24).Ifwedonthaveanimaginarycomponent,asin

thecaseofastandstill,thiscrosscouplingfactorinthedandqaxesdoesntappear.Inthecaseofanasynchronousmotorthereisno
differencebetweenthedandqaxesparameters,soitisnotrequiredtodividethevectorialequation(7.24)toitsdandqcomponents.
3.Innormaloperationoftheasynchronousmachineallthevectorsarerotatingwiththesynchronousspeed .Inthiscaseitisdesirable
tochooseacoordinatesystemwhichrotatesatthissynchronousspeed .Inthiscaseallthevectorsareconstantduringnormal
operation.
Thevoltageequationsbecome

(7.28)

(7.29)

From(7.28)and(7.29)thed/dtcomponentwillfalloutinthecaseofnormaloperation,and wheresistheslip.Inthis
casetherotorequationismodifiedasfollows:

(7.30)

Equation(7.30)isnothingelsebuttheequivalentcircuitoftheasynchronousmotorinnormaloperation.Theequivalentcircuitofthe
asynchronousmotorinthiscaseisshowninFigure1.7.
Fromtheabovestatementsitcanbeconcludedthatinnormaloperation,usingacommonreferenceframe,themachineequationsbecome
assimpleasinthecaseofatransformator.Fordynamicortransientanalysesweusetheequivalentcircuitsaccordingtofluxequations,
insteadofequivalentcircuitsaccordingtovoltageequations.Letustransformthefluxequationsnow,usingtheequations ,

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 6/27
6/14/2017 Chapter7.ModellingInductionMotors
and ,wherethefirstequationisthatofvectorialmagnetizingcurrent,whichistheresultofthestatorandrotorfield
induction,when1:1effectiveturnorreductionratioisused,andthesecondisthatoftheequationofthefluxofthemainmagneticfield.
Fromequations(7.22)and(7.23)usingandsubstitutingtheabovestatementsweobtainthefollowingfluxequations:

Figure7.7.EquivalentcircuitoftheACmotorinnormaloperation

(7.31)

(7.32)

Basedonthe(7.31)and(7.32)equationsthefluxequivalentcircuitisshowninFigure78.Thevectorialdiagramofvoltagesandfluxesin
astationarycasecanbeseeninFigure79.AccordingtotheFigure79,insynchronousspeedcoordinatesystem ,allthe
vectorsareconstant,andthevoltageandrotorequationsbecome

,(7.33)

(7.34)

7.3.ModifiedEquivalentCircuits
Theequivalentcircuitsstemmingfromthefluxequationsnormallyuseaperuniteffectiveturnorreductionratio.Inthiscasethe
magnitudesinrelativeunitsare .Inmanycasesitisworthtochooseaturnratiowhichdiffersfromthe1:1ratio,
becauseoftheconsistentsimplificationsinthemotorequations.Letusintroducethefictivereductionfactorafortherotorquantities.

Figure7.8.Equivalentcircuitoffluxes

, , , .Inthiscasethefluxequations(7.31)and(7.32)willbe

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 7/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.9.Vectorialdiagramoffluxesandvoltages

(7.35)

(7.36)

where , , , .Thereductionfactorcanbechoseninsuchawaythat,after
reduction,oneofthereducedleakagefactorsbecomeszero. ,if ,

and ,if .Ifweusetheabovementionednumbersfortheinductanceinthecase

of1:1reductionratio,wegeta1=2.5/2.5+0.110.4anda2=2.5+0.1/2.51.04.Simplifiedfluxequivalentcircuitscanbeusedinthiscase
whenthereductionratioaltersfrom1:1,withafewpercent.InFigure710.,andFigure711.twosimplifiedfluxequivalentcircuitsare
presented.Evaluatingthefluxesmarkedby*,withtheoriginalinductancevalues(inthecaseof1:1reductionratio),furthersimplification
ofthemodelispossible.Thereisobviousneedtointroducethemotortransientinductance ,whichistheresultinginductancefromthe
primarysite,neglectingtheresistorsandconsideringthesecondarypartinashortcircuit.Intheequivalentcircuit(Figure78.)accordingto
fluxes,wecanchoose insteadof :

(7.37)

(7.38)

(7.39)

Figure7.10.Modifiedequivalentcircuitwiththeeliminationofstatorleakage

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 8/27
6/14/2017 Chapter7.ModellingInductionMotors

(7.40)

Figure7.11.Modifiedequivalentcircuitforfluxesiftherotorleakageiseliminated

whereistheresultant(total)leakagecoefficientandkisthecouplingcoefficient.The rotortransientinductanceisthesecondary
resultantinductance,neglectingtheresistorsandconsideringashortcircuitintheprimercircuit.Socalculatingwith weget

Figure7.12.Equivalentcircuitofthemotorvalidalsofortransientoperation

(7.41)

(7.42)

(7.43)

IncaseofFigure710.,wehave

(7.44)

IncaseofFigure711.,wehave

(7.45)

Asitcanbeobservedonthefigures,themainfield,inthecaseofthefirstfigure,is ,while,inthecaseofthesecondfigure,it
is .Knowingthevaluesof , or ,bothequivalentcircuitsofFigure710.orFigure711.canbeuseful.Itisevidentthatin
thiscasethereductionmustbeeffectuatedalsoforthevoltageequations,asfollows:

(7.46)

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 9/27
6/14/2017 Chapter7.ModellingInductionMotors

(7.47)

Thisformmatchestheoriginalequation,soifwechoosethemodeofreductionthe*notationcanbeomitted.TheFigure712.presents
themotorequivalentcircuit,whichisalsovalidfortransientoperationofthemachines.
Takingtheabovestatementsintoconsideration,andregardingthemodifiedequivalentcircuits,forthepurposeoffurtheranalysis,an
equivalentcircuithasbeenchosen,wheretherotorleakageimpedancehasbeenincludedintotheexpressionofthestatortransient
impedance.ThenewequivalentcircuitcanbeseeninFigure713.Inthiscase,aswewillseelateron,theprogramcalculatesthe
voltagescorrespondingtothenecessarycurrentofthemotor,insteadofusingacurrentcontrolloop.Becausethealgorithmrunningtime
islessthantherotortimeconstant,thefluxcanbeconsideredasconstant.Consequently,thechosenequivalentcircuit,wheretherotor
leakageimpedanceisproperlyincludedinthestatortransientimpedance(Ls'),isagoodchoice.

Figure7.13.Thechosenequivalentcircuit

Afterreducingthemotor'sphysicalparametersandtransformingtheequationsmathematically,theexpressionofthestatorvoltagevector
equationwillbeasfollows:

(7.48)

Aswecanseeonthefollowingpages,thisequationwillplayaveryimportantroleinsettingupthecontrolstructures,andalsoin
applications.

7.4.SettinguptheModel
Letuschoosefirstastandstillcoordinatesystem ,whichmeansthatthecommoncoordinatesystemisastationaryone.
Replacingthevalue inthegeneralequationsofthemotorgivenby(7.20),(7.21),(7.22),and(7.23),thefollowingvectorialvoltage
equationsconclude:

(7.49)

(7.50)

Letusfix and asstatevariables.Aftertheeliminationoftheterms and fromtheabovedifferentialequationsystem,weget


thefollowingequations:

(7.51)

(7.52)

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 10/27
6/14/2017 Chapter7.ModellingInductionMotors

(7.53)

Solvingtheaboveequationsaccordingtothechosenstatevariablesweget

(7.54)

(7.55)

Forsimplifyingtheaboveequationsletusmakethenotation.

(7.56)

Thehasbeendefinedin(7.40).Usingagainthecomplexrepresentationofthevectors and

,themodeloftheinductionmotorwhen areconsideredasstatevariablescanbewritteninthe
followingmatrixform:

(7.57)

Insamereferenceswhenthetwophasecoordinatesystemdoesnotrotatethequantities areusedinsteadof
d,qquantities.Sointherestofthethesesthewhen,componentsareusedtheserefertoastationarytwophasecoordinatesystem,when
d,qcomponentsareusedthesewillbethequantitiesinaspecialrotatingtwophasesystem(fieldcoordinates).Themodeloftheinduction
motorwillbedescribedinfieldcoordinates(innoneofthesemodelsoftheAsynchronousMotordowetakethemechanicallossesinto
account),consideringnowthemotionorcinematicequationofthemotor,whichisveryimportantwhenweapplystochasticobserversfor
themotor.Asabasis,nowagain,theequationsdeducedinthepreviousparagraph,butwithnotationsslightlychangedasdescribedalso
in[14],[15]willbeused.Thissystemofequationsisnonlinear.Theindices"r"and"s"meanrotorandstator,respectively:

(7.58)

(7.59)

(7.60)

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 11/27
6/14/2017 Chapter7.ModellingInductionMotors

,where istherotorangularvelocity.(7.61)

Thedefinitionoftheinductancesintheseequationsandthetotalleakagefactorisasfollows:

(7.62)

whereLmisthemaininductanceofthemotor, and aretheresistancesofthewindingsandLls,Llr,aretheleakageinductancesof


thestatorandrotorrespectively.
ThismodelhasbeenrealizedinSIMULINKfortestandsimulationpurposes.SeeFigure715.forthegraphicrealization.Wewillnow
definethemodeloftheasynchronousmotorinfieldcoordinates.Thiswillmakeitpossibletoimplementthecontrollerinfieldco
ordinates.Fieldorientationisdescribedindetailin[16],andinthefollowingsections.Thebasicprincipleisthatweconvertallvaluesto
thecoordinatesystemofthemagneticfield,decomposethestatorcurrentvectorintoafieldgeneratingandatorquegenerating( and
)component.Atthismomenttheactualcontrolbecomesverysimple.Sincetheprogramisdesignedsothatthefieldweakeningrange
willnotbereached,thefieldgeneratingcomponentofthestatorcurrentvoltagevectorcanbekeptataconstantvaluetogeneratea
necessaryfield,andcontrolthetorquegeneratingcomponentaccordingtothespeed.Thismeansthattheoutputofthespeedcontrolleris
thereferencefor .Wewilleliminatetheuseoffluxinthemodel,andwewillusethemagnetizingcurrentinstead.Theconnection
betweenthesetwovariablesisthefollowing:

(7.63)

whereistheangleoftherotormagnetizingcurrentvectorwithrespectofthestationaryaxis.Thisquantitywillplayanimportantrolein
vectorcontrolwhenthespecialreferenceframewillbefixedtotherotorflux.Thetransformationbetweenthesystemswillbedonebythe
followingtransformation:

(7.64)

First,letusconsiderabriefoverviewofthestatevectors(alsocalledstatephasors)representedindifferentcoordinatesystemsas
showninFigure714.

7.5.StatorCoordinates
LetusnowswitchtoaDescartescoordinatesystemandfixittothestator(,stationaryreferenceframe).Inthiscasethespeedofthe
rotatingcoordinatesystemis .Letuschoosethestatorcurrentandtherotorfluxtobethestatevariables.Itmeansthatwehave
toexpressthesequantitiesfromthegeneralmotormodeldescribedinequations(7.20)(7.23).Theresultedscalarmodelequationsare:

(7.65)

(7.66)

(7.67)

(7.68)

Therotortimeconstantisdenotedby .

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 12/27
6/14/2017 Chapter7.ModellingInductionMotors
Theseequationsarenowrewritteninastatespaceformwithstatevariablesis,is,randr.

(7.69)

(7.70)

(7.71)

(7.72)

Figure7.14.Connectionbetweendifferentcoordinatesystems

Thestatortimeconstantisdenotedby .Thekinematicequationsofthemotorinstatorcoordinatesareasfollows:

(7.73)

(7.74)

(7.75)

wheremListhemechanicaltorqueandpisthemagneticpolecount.

7.6.FieldCoordinates
Choosingtherotorfluxasorientationquantity,thatmeansallquantitiesareviewedfromtherotorfluxpointofview,whichareusually
calledfieldcoordinates.Theadvantageofusingthisreferenceframeisthatallquantitiesareconstantinthesteadystate.Thefluxisnow

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 13/27
6/14/2017 Chapter7.ModellingInductionMotors
eliminatedbyintroducinganewquantitycalledmagnetizingcurrent(rotorcurrentcannotbemeasured)seeeq.(7.63).Thestatorcurrent
(infieldcoordinates)andthemagnetizingcurrentarenowchosentobethestatevariables.Theresultedscalarmodelequationsare:

(7.76)

(7.77)

(7.78)

(7.79)

LetusnowrewritetheseequationsinastatespaceformwherestatevariablesiSd,iSqandimRareused.Noticethatbyintroducingthe
magnetizingcurrentweneedonlythreestatevariables.

(7.80)

Where ,thatis,thefluxspeedequalstotherotorspeedplustheslipangularvelocityoftherotorflux.Thekinematic
equationsofthemotorinfieldcoordinatesare:

(7.81)

Theformsofequationsremainsunchangedinthiscoordinatesystemtoo.Equationspresentsthetransformationmatricesbetweenthe
threephaseandtheDescartesstationarycoordinatesystems,describestheconnectionbetweenthestatorandthefieldcoordinates.

(7.82)

(7.83)

Figure714.isasummaryofthestatevectors(alsocalledstatephasors)representedindifferentcoordinatesystems.Notethattherotor
basedvariablesarecompletelyeliminated.

7.7.ModelassumptionsfortheFOCmethod
Usingthefieldorientedmodelofthemotor,ithastheoreticallybecomepossibletorealizeacontrolinfieldorientation,whichmakesit
extremelyeasytocontrolspeedbycontrolling .However,oursystemofequationsisnonlinear.Wewilleliminatethisnonlinearityby
decouplingthenonlinearterms.Thismethodhasbeendescribedin[14],anddetailedtreatmentisalsogivenin[17]:

7.7.1.DecouplingoftheNonlinearity
http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 14/27
6/14/2017 Chapter7.ModellingInductionMotors
Wecanseefrom(7.63),thatthemodeloftheinductionmotorisnonlinearandtimevariant.Letusnowexaminehownonlinearityis
eliminatedbydecouplingtheundesirablenonlinearandcouplingterms.

(7.84)

(7.85)

(7.86)

(7.87)

Wewouldliketocontrolthespeed.Letusrecall.Thetorqueisdirectlyproportionaltoimrandtoisq.BecauseTrislarge,imrcannotbe
usedforaquickcontrol.TheleakagefactorofiSqissmall,thatiswhyiSqcanbeusedtocontrolthetorqueandthespeed.Wegetthe
bestdynamicbehaviourwhenimriskeptconstant,therefore canbeassumed.Thiswillremovethelastnonlinearitybymaking

constant:SupposingthatthemaininductanceLmistimeinvariant,therotorfluxriskeptconstant,too.Underthisassumption,
using,,,and(7.87)wegetthefollowinglinearanddecoupledmodeloftheinductionmotor:

(7.88)

Where

(7.89)

Assumingthat iskeptconstant,KJisalsoconstant.Inthiscasethesystemdescribedbycanbesplitintotwoindependent

parts,thedirect(denotedbyd)andthequadrature(denotedbyq)subsystem:

(7.90)

(7.91)

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 15/27
6/14/2017 Chapter7.ModellingInductionMotors

Afterthisitispossibletocontrolthemotor,providedthatweareabletoconvertallvaluestofieldcoordinates.Thisismadebytheflux
model(seethedescriptionlaterintheSimulinkrealizationinFigure721.andFigure722.).Theformofthefluxmodelwhichhasbeen
realizedhereisbasedon[16],[18]andaSimulinkrealizationdonebydSPACEGmbH.Afterrealizingthefluxmodelcontrolbecomes
possible.Thefirstloopofthecontrolisacontrollerforimr,whichisinfactacontrollerfortherotorflux.Thereisacontrollerforisd,which
willforcethecorrectimr.Thesecondloopbeginswithacontrollerfor ,followedbyacontrollerforisq,whichwillforcethecorrect
torque,andimplicitlywillrealizethecorrectvelocity.TheoverallcontrolstructureisshowninFigure723.andFigure724.

7.7.2.TheMatlab/SimulinkModeloftheACMotor
ThemodeloftheACmotor,whichissometimescalledsimplyASM=AsynchronousMotor,hasbeenimplementedinSimulinkaccording
totheequationsasdescribedin[16]and[14].TheactualmodeloftheASMcanbeseeninFigure715.Thismodelisbasedonthe
mathematicalmodeldescribedin[16].However,ithasbeenadaptedtotherequirementsofSimulink.Forthesakeofsimplicity,allvector
variableshavebeensubstitutedbyscalarvariables.Themodelisinstatorandrotorcoordinates.Theseequationshavetheadvantageof
aneasyimplementation,andgreatsymmetry.ThestatorcoordinatesareindexedwithS,therotorcoordinateswithR,asbefore,andthe
indices anda denotethetwoaxesofthecoordinatesystems.TherealizedequationsofthemodelpresentedinFigure715.arethe
following:

Figure7.15.MATLAB/SIMULINKmodeloftheACmotor

(7.92)

(7.93)

(7.94)

(7.95)

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 16/27
6/14/2017 Chapter7.ModellingInductionMotors

(7.96)

(7.97)

(7.98)

Thebasicfunctionalityoftheabovemodelwastestedwithsinevoltagesignalsexcitationinopenloop.Thespeedsettlednearthespeed
ofthesignals,thatis,therotormovedalmostsynchronouslywiththegeneratedvoltagevector.Thenatorquewasapplied,which
introducedaslipintothesystem.Sincethevoltagesdidnotchange,thespeedoftherotorsanktoalowerlevel.Thischapteralso
describestheadaptationoftheASM(AsynchronousMotor)modelfordifferentkindsofalgorithms,includingsensorlesscontrol,too.The
firststeptodesignaKalmanfilter,forexample,istohaveanappropriatemodel,whichisusedbythefilter.Forthispurpose,wefollowed
themethodpresentedin[19]andderivedthemodel,whichwascomparedwiththeresultsof[14]andtheresultswerethesame.The
modeltakenasabasiswasthefieldorientedmodelofthemotorbasedon[19],describedbytheequationspresentedin.These
equationscanbeformulatedinsuchaformwhichisnonlinear,butcanbeconvertedintoatimedependentmatrixform.Letusseehowto
converttheseequations:

(7.99)

Thiscanbefurthermodifiedbyreplacing basedon(7.106),andsubstituting byTs.Wegetthefollowing:

(7.100)

(7.101)

(7.102)

(7.103)

Theseequationscannowbewritteninmatrixform.Sincewedonothaveasensorfortheloadtorque,itwillbesimplyomitted,and
regardedasadisturbance.Alsonotethatthetransformationmatrixbetweenthefieldandthestatorisintroducedtotheequations,andso
theinputandoutputvariablesareinstatorcoordinates.Thisformiseasiertohandleinthecaseoftheobservingproblem.

(7.104)

(7.105)

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 17/27
6/14/2017 Chapter7.ModellingInductionMotors

Notethatinthismodel,andinallthefollowingmodels,ispartoftheoutputvector.Thisdoesnotmeanthatwemeasureit,butitmustbe
estimatedroughly,andthisestimatedvaluemustbesubstitutedintotheKalmanfilterwherethisoutputvectorisneeded.Thesubstitution
ismadeusingthefollowingformula:

(7.106)

Forpracticalimplementationwewillneedthemodelindiscretetime,soletusconsidertheconversiontodiscretetime.Letusdenotethe
systemmatricesofthecontinuoussystemwithA,BandC,andthoseofthediscretesystemwithAd,Bd,andCd.Ifwesupposethatour
samplingtimeisveryshort,ascomparedtothedynamicsofthesystem(whichisthecaseinmostapplications),wecangeneratethe
systemwiththefollowingapproximateequations(firstorderTaylorseries,see[19]):

(7.107)

(,7.108)

(7.109)

Withthese,nowwecanpresentthesystemindiscretetime.

(7.110)

(7.111)

Thissystemisdependentonthevaluesof and ,whichwereregardedasknownvaluesuntilnow.In[19],however,theKalman


filterispresentedalsoasasubstituteforthefluxmodel,andthefluxisobservedaswell.Thefirstimplementationdoesnotfollowthis
method,butitwillbepresentedbecausefurtherimprovementsmaybebasedonit.Thereasonwhyasmallermodelwaschosenisthat
thedimensionofthemodelisoneofthemostimportantfactorsdeterminingthespeedofcalculation,andthereductionofa6thorder
modelintoa4thorderone,whichisrelativelysimple,canbedoneatagreaterspeedthantheextraneouscalculationofthefluxmodel.To
observetheflux,wemustobserve and .Thiswillbedonebyaddingthesetothesystemasnewstatevariables.Thisincreases
theorderofoursystemto6.Toaddthenewvariables,wewillassumethatthesamplingtimeisshortenoughforthespeedofthefluxto
remainconstantwithinthesamplingperiod.Thatis,wecanassumethat

(7.112)

(7.113)
http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 18/27
6/14/2017 Chapter7.ModellingInductionMotors

Withtheseassumptionswecanestablishthenew6thordersystem,whichcanbethebaseofaKalmanfilter,whichobservesthefluxas
well.Forconveniencethediscretetimemodeloftheinductionmotorinastationaryreferenceframeisalsopresentedwhichwillserve
againforobserverapplication(7.114),wherethespeedoftherotorcanbeaddedasastatevariableofthesystem.IntheequationTisthe
samplingtimeandhastobenotconfusedwiththetimeconstantofthestatorandrotor.

(7.114)

(7.115)

(7.116)

7.8.VectorversusScalarcontrol
Thefieldorientedtheoryisthebaseofaspecialcontrolmethodforinductionmotordrives.Withthiscontrolmethod,inductionmotorscan
successfullyreplaceexpensiveDCmotors.Nowadaysthismethodhasbecomegeneralininductionmotordrivesofhighprecision.The
mainadvantagesofinductionmotorsaretheirsimplicityandtheirpriceaswellastheirgreaterreliability,especiallyinharshindustrial
environments.Inductionmotorsrequireverycomplexcontrolalgorithmsbecausethereisnolinearrelationshipbetweenthestatorcurrent
andeitherthetorqueortheflux.Thismeansthat,becauseofthetransientstates,itisdifficulttocontrolthespeedofthetorqueuntilthe
motorreachesitsnewstationarystate.Controllingtherotorflux,sinceitcannotbemeasured,butonlycomputedcansolvetheproblem.
Thepurposeofthecontrolleristokeeptheamplitudeoftherotorfluxataconstantvaluesothatonlyitsdirectionischanged.Thefield
orientedtheoryoffersasuitablemethodforoptimallycontroloftheinductionmotors.Thecomplexityofthemethodiscompensatedbythe
advantages.Fieldorientedcontrolcanbeusedinawiderangeofindustrialapplications,thankstothedevelopmentofpowerelectronic
componentsandmicroprocessorsystems.Tohaveabasisforcomparison,letusnowintroduceveryshortlythemosttraditionalscalar
control.

7.8.1.ScalarControl
ScalarcontrolmeansthatvariablesarecontrolledonlyinmagnitudeandthefeedbackandcommandsignalsareproportionaltoDC
quantities.Withascalarcontrolmethod,wecanonlydrivethestatorfrequencyusingavoltageoracurrentvalueasacommand.
Regardingthescalarmethodsknowntocontrolaninductionmotor,oneassumesthatbyvaryingstatorvoltagesinproportionwith
frequency,thetorqueiskeptconstant:
http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 19/27
6/14/2017 Chapter7.ModellingInductionMotors

(7.117)

Figure7.16.MaximalandNominalTorquevs.Speed

Figure7.17.Structureofaclosedloopscalarcontrolwithvolts/hertzandslipregulation

Figure717.showsaclosedloopspeedcontrolscheme,whichusesvolts/hertzandslipregulation.
Thecontrolledslipstrategyiswidelyusedbecausetheinductionmachinesinputpowerfactorandtorquetostatorcurrentratiocanboth
bekepthigh,resultinginbetterutilizationoftheavailableinvertercurrent.Whenairgapfluxandslipspeedarebothheldconstant,the
torquedevelopedwillbethesame,buttheefficiencyisnotasgoodasthatobtainedwithairgapfluxandslipheldconstant.Whentheslip
isheldconstant,theslipspeedwillvarylinearlywiththeexcitationfrequencyandtheslopeofthetorquespeedcurveonthesynchronous
speedsidewilldecreasewiththeexcitationfrequency.Thesecontrolmethodshavebeenwellstudiedintheliteratureanditisnotthe
subjectofthisthesistogiveamoreanalyticalinsightintothisfield.Rather,wewillturnnowtothepresentationofthemoreimportant
vectorcontrolprincipleandtheassociatedquestionsinACmotorcontrolbasedonthisprinciple.

7.8.2.VectorControl
Vectorcontrolisreferringnotonlytothemagnitudebutalsotothephaseofthesevariables.Matricesandvectorsareusedtorepresent
controlquantities.Thismethodtakesintoconsiderationnotonlysuccessivesteadystatesbutalsorealmathematicalequationsthat
describethemotoritself.Thecontrolresultsobtainedhaveabetterdynamicfortorquevariationsinawiderspeedrange.Thespace
phasortheoryisamethodtohandletheequations.Thoughtheinductionmotorhasaverysimplestructure,itsmathematicalmodelis
complex,duetothecouplingfactorstobeappliedtoalargenumberofvariables,andtothenonlinearities.FieldOrientedControl(FOC)
offersasolutiontocircumventtheneedtosolvehighorderequationsandtoachieveanefficientcontrolwithhighdynamic.Thisapproach
requiresmorecalculationsthanastandardU/fcontrolscheme.ThiscanbesolvedbytheuseofacalculationunitincludedinaDigital
SignalProcessor(DSP),andhasthefollowingadvantages:
fullmotortorquecapabilityatlowspeed,
betterdynamicbehavior,
higherefficiencyforeachoperationpointinawidespeedrange,
decoupledcontroloftorqueandflux,
shorttermoverloadcapability,fourquadrantoperation.

7.8.2.1.TheFOCAlgorithmStructure

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 20/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.18.ThevectordiagramofFOCprinciple

FOCconsistsofcontrollingthecomponentsofthemotorstatorcurrents,representedbyavector,inarotatingreferenceframed,qaligned
withtherotorflux.Thevectorcontrolsystemrequiresthedynamicmodelequationsoftheinductionmotorandreturnstheinstantaneous
currentsandvoltagesinordertocalculateandcontrolthevariables.Statorcurrentandfluxspacevectorsinthed,qrotatingreference
frameandtherelationshipofthatwiththe,stationaryreferenceframecanbeseeninFigure718.Thecontrolofinductionmachinesmust
takeplaceinanorthogonalreferenceframethatrotateswithrotorfluxangularvelocity( )insuchawaythatthed axescoincide
withtheinstantaneouspositionoftherotorflux .Inthiscasethetorqueexpressionhasjustoneterm,acurrentmultipliedbyoneflux.
Theorthogonalcomponentsofthefluxeswillbezero.Thecurrentinthetorqueexpressionwillnotbeotherthantheprojectionofstator
currentvectortotheq axes,andcanbecharacterizedastheactivecurrent.Thestatorcurrentprojectiontothed axesgivesthe
magnetizingcurrentofthemotor.Inthiswaytheinductionmotorcanbecontrolledbytwoseparateloops:oneforthereactivecurrent
(flux)andonefortheactivecurrent(torqueorspeed),likeaseparatelyexcitedDCmotor.Thefluxidentificationproceduredefinesthetype
ofthecontrolsystem.
Thereareverysimplecaseswheretheorientationfluxanditspositionaredeterminedbymeasurement.Whenthisisnotpossible,inthe
majorityofcases,theorientationquantitiesaredeterminedbyusingthemathematicalmodeloftheinductionmachineexpressedinthe
orthogonalcoordinatesystem.Acomplexcontrolsystemcandefinetheorientationquantitiesinonecoordinatesystemandcontrolthe
motortoanother.InFigure719.thespacephasordiagramoftheinductionmachinewithshortcircuitedrotorwindingsisproposed.The
mostoftenusedmethodistheonewithrotorresultantfluxorientation,becauseofthesimplestructureofthecontrolloopsandeasy
commandvariablecalculation.Thespacephasorofthestatorcurrentissplitintotwocomponents,whichbecomecontrolvariables.
Vectorrotationtechniquesareusedtotransformthreephaseaxesintorotatingtwophase"dq"axes.Thistwophaserotationtechnique
greatlysimplifiestheanalysis,makingitequivalenttoanalyzingseparatelyexcitedDCmotors,becauseinthiscasetherearetwo
independentlycontrollablecurrents:thefieldcurrentandthearmaturecurrent.Inatwophasesystem and representthestator
currentinthestationaryreferenceframewheretheaxesofthetwophasesystemexactlymatchtheaaxis.Tocontrolfieldandtorque
independently,wemustusethemoving"dq"coordinatesystem,rotatedatthesynchronousspeedwithrespecttothestationary
referenceframe.Projecting onto"dq"yieldsthecomponents and .Tocontrolthesequantitiesthestatorcurrentvectormust
beorientedto"dq",where istheorientationangle.Letusomitthepresentationofcoordinatetransformationsrelatedtothevector
controlstrategybecausetheyhavebeenwelldescribedintheliterature.Weonlysummarizeherethetransformationsthatmustbe
effectuated:
(a,b,c)(,)istheClarketransformation,whichoutputsatwocoordinatetimevariantsystem,and
(,)(d,q)istheParktransformation,whichoutputsatwocoordinatetimeinvariantsystem.
ThisisthemostimportanttransformationinFOC.Infact,thisprojectionmodifiesatwophaseorthogonalsystemina(d,q)rotating
referencesystem.

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 21/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.19.BasicschemeofFOCforACmotor

The(d,q)(,)projectionistheinverseParktransformation.
Theoutputsofthisblockarethecomponentsofthereferencevector,whichisthevoltagespacevectortobeappliedtothemotorphases.
Figure720.summarizesthebasicschemeoftorquecontrolwithFOC.Twomotorphasecurrentsaremeasured,whichfeedtheClarke
transformationmodule.Theoutputsofthisprojectionaredesignatedisandis.Thesetwocomponentsofthestatorcurrentaretheinputs
oftheParktransformationthatgivesthecurrentinthed,qrotatingreferenceframe.Theisdandisqcomponentsarecomparedtothe
references(isd)ref(fluxreference)and(isq)ref(torquereference).Atthispoint,thiscontrolstructureshowsaninterestingadvantage:itcan
beusedtocontroleithersynchronousorinductionmachinesbysimplychangingthefluxreferenceandobtainingtherotorfluxposition.As
inthecaseofsynchronouspermanentmagnetmotors,therotorfluxisfixed(determinedbythemagnets)thereisnoneedtocreateone.
Hence,whencontrollingaPMSM,(isd)refshouldbesetatzero.Asinductionmotorsneedrotorfluxcreationinordertooperate,theflux
referencemustnotbezero.Thisconventionallysolvesoneofthemajordrawbacksoftheclassiccontrolstructures:thetransferability
fromasynchronoustosynchronousdrives.Thetorquecommand(isq)refcouldbetheoutputofthespeedregulatorwhenweuseaspeed
FOC.Theoutputsofthecurrentregulators(usd)refand(usq)refareappliedtotheinverseParktransformationblock.Theoutputsofthis
projectionare(us)refand(us)ref,whicharethecomponentsofthestatorvectorvoltageinthe(,)stationaryorthogonalreferenceframe.
ThesearetheinputsoftheSpaceVectorPWM.Theoutputsofthisblockarethesignalsthatdrivetheinverter.Itisworthtoremarkthat
boththeParkandtheinverseParktransformationneedtherotorfluxposition.ObtainingthisrotorfluxpositiondependsontheAC
machinestype(synchronousorasynchronous).Someconsiderationswithregardstorotorfluxpositionformotorsoftheasynchronous
typearethefollowing.Ifthecontrollingmethodisbasedonthefieldorientationprinciple,theidentificationoftheorientedquantitiesisof
crucialimportance.Thisincludesthecalculationorpredictionofthepositionandmagnitudeofthechosenflux,thecontroloftheactive
andreactivecurrentsinthedqreferenceframe,andtherecoveryofthesequantitiesina,b,cquantities,whichcancontrolthemotor
throughastatictypefrequencyconverter.Thestructureofthecontrolsystemandthetypeofthefrequencyconvertermustbetakeninto
considerationwhenmakingtheimportantdecisiononwhichfluxistobeusedinorientation.Thecontrolsystemsdiffernotonlyinthis
respect,thatis,whichfluxisusedinorientation,butalsointheidentificationprocedure.Thereareverysimplecases,whentheorientation
fluxanditspositionaredeterminedbymeasuring(directvectorcontrol).Whenthisisnotpossible,asinthemajorityofcases,the
orientationquantitiesaredeterminedbyusingthemathematicalmodeloftheinductionmachine(indirectvectorcontrol),writteninthe
orthogonalcoordinatesystem.Thislattercoordinatesystemcanbeorientedafteranotherphasor(notonlythefluxphasor),inthescope
ofoperation,withasimplermathematicalmodel.

Figure7.20.BlockdiagramofathreePhasesAsynchronousMotorDriverUsingaFOC

Consequently,ifacomplexcontrolsystemcandefinetheorientationquantitiesinonecoordinatesystem,andcontrolthemotorto
another,thefieldorientationcannotbeusedbyitself.Forthisreason,thevectorcontrolterminologyisappropriateforthissystem,rather
thanthatoffieldorientation.Anotherclassificationmethodologycanalsobeadaptedtoafieldorientedcontrolsystem,basedonthe
determinationofthefieldorientationquantities,whentheparametersareidentifiedonline.Thismethodmeansrealtimeidentificationand
tracingofprocessquantitiesliketheresistanceoftherotorandthechangesoftherotorstimeconstant.Thismethodisapplicabletosuch
http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 22/27
6/14/2017 Chapter7.ModellingInductionMotors
systemswherethetorquehasalowchangingrate.Amoreefficientmethodiswhereparameteridentificationisdeterminedoffline,with
thehelpofamathematicalmodelofthemotor,themethodofpreviouslytestingthemachine,isapplicableonalargeoperationscale.
TheelectrictorqueofanACinductionmotorcanbedescribedbytheinteractionbetweentherotorcurrentsandthefluxwaveresulting
fromthestatorcurrentinduction.Sincetherotorcurrentscannotbemeasuredwithcagemotors,thiscurrentisreplacedbyanequivalent
quantitydescribedinarotatingcoordinatesystemcalledd,qfollowingtherotorflux.Theinstantaneousfluxangle iscalculatedbythe
motorfluxmodel,whichwillbegivenlater.isdandisq,thestatorcurrentcomponentsinthed,qframe,areobtaineddirectlyfromia,iband
ic,thefixedcoordinatestatorphasecurrents,withtheParktransformation:

(7.118)

Insteadystateconditionsthestatorcurrentisdefinedintheabovementionedrotatingsystemisconsideredconstant,aswellasthe
magnetizingcurrentimrrepresentingtherotorflux,isqbeingequivalenttothemotortorque.isdislinkedtoimr[16]and[18]withthe
followingequation:

(7.119)

where istherotortimeconstant.Thissystemtogetherwiththeangletransformations,changestheinductionmotorintoamachinevery
similartoaDCmotorwhereimrcorrespondstotheDCmotorsmainflux,andisqtothearmaturecurrent.Thefieldorientedcontrol
methodachievesthebestdynamicbehavior,wherebytheleadanddisturbancebehaviorcanbeimprovedwithshortercontrolcycletimes.
Thefieldorientedcontrolmethodisadefactostandardtocontrolaninductionmotorinadjustablespeeddriveapplicationswithquickly
changingloadaswellasreferencespeeds.Itsadvantageisthatbytransformingmeasurablestatorvariablesintoasystembasedonfield
coordinatesthecomplexityofthesystemcanbeenormouslyreduced.Asaresult,arelativelysimplecontrolmethod,verysimilartothat
ofaseparatelyexitedDCmotor,canbeapplied.TheroleoftheDSPinsuchasystemistotranslatethestatorvariables(currentsand
angles)intoafluxmodelaswellastocomparethevalueswiththereferencevaluesandupdatethePIcontrollers.Afterback
transformationfromfieldtostatorcoordinates,theoutputvoltagewillbeimpressedtothemachinewithasymmetricorasymmetric
PWM,wherebythepulsepatterniscomputedonlinebytheDSPorahardwaregeneratedspacevectormethod.Insomesystemsthe
positionismeasuredbyanencoder.Thisextracostcanbeavoidedbyimplementinganobservermodel,or,inparticularcases,aKalman
filter.Thesealgorithmsarecomplexand,therefore,requireafastprocessor.AfixedpointDSPisabletoperformtheabovecontrolswith
shortcycletimes.Letusrepeatherethebasicequationsoffieldorientation,whicharededucedin[18].Substitutingtherotorcurrentinthe
rotorvoltageequationandconsideringthemagnetizingcurrentequaltotherotorfluxdividedbymaintheinductance,wecanexpressthe
dqcomponentsofthestatorcurrentphasorasfollows:

(7.120)

Thedcomponentofthestatorcurrentvectorisinfluencedbytherotorfluxamplitudevariation,andtheqcomponentbytherelativespeed
betweentherotorfluxspacephasorandtherotorsspatialposition.Lookingattheequation(7.120),wecannoticethatwhenevertherotor
fluxmagnitudeisnotconstant,thedcomponentofthestatorcurrentcontainsanewelement,proportionaltothisfluxvariation,and
respectively,tothemagnetizingcurrent(delayelement).Theelectromagneticairgaptorquecanbecalculatedfromthefluxoriented
variablesasfollows:

(7.121)

Ifthevariablesareallrotorfluxoriented,thefollowingexpressionscanbeobtained:

(7.122)

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 23/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.21.Fluxidentificationmodelinthecaseofarotorfluxorientedcoordinatesystem

Therotorfluxorientedcomponentsofthestatorcurrentspacephasor canserveforthecontrolofrotorfluxandthemachine
torque,respectively.Letustakealooknowatthecalculationoftheorientationquantitiesbasedontheabovedeductedequations.Asthe
inputstatorcomponentsmustbeorientedandtheorientationvariablesappearonlyattheoutputofthisblock,internalcalculationsto
determinethetrigonometricfunctionsareneeded.ThestructuralschemeofthefluxidentificationisgiveninFigure721.Thestatorcurrent
componentsarefieldoriented.Thetwononlineardifferentialequationsprovetobestableunderalloperatingconditionsandtheir
continuousonlinesolutiononaDSPdoesnotraiseproblems.Itismoreadvantageoustousestatorcurrentsinsteadofvoltages,dueto
theirreducedharmoniccontent.Thisisthereasonwhyatstatorfrequencieslowerthan100Hz,asamplingtimeof1msanda10bit
resolutionA/Dconverterarepermitted.Thestatorresistanceandleakageinductancehavenoinfluenceontheorientationfieldcalculation,
asthefeedbackloopisindependentofthestatorvoltagemodel.Ofrealinterestaretherotorparameters,especiallythetimeconstant ,
influencedbytherotorresistance(varyingwiththetemperature),andtheironcoresaturation.Thetorquecalculationisalsodependenton
Lmandontherotorleakagecoefficient .Inordertoachievehigherspeeds,higherthanthenominalvalues,thestatorfrequencyhasto
beincreased.Inhighfrequenciesthestatorresistancecanbeneglected.Ifthefluxmagnitudevariationisalsoneglected,thevoltage
equationbecomesverysimple: ,where isthesynchronousfrequency.Itisobviousthatifthefluxiskeptconstant,the
limitvoltageoftheconverterisreachedatacertainstatorfrequency.Thusthefluxmustbereducediffurtherspeedincreaseisneeded.
Undersuchoperatingconditions,inthesocalledfieldweakeningregion,thestatorvoltageiskeptapproximatelyconstantatitsmaximum.

7.9.Summary
Whentherotorisshortcircuitedandiftherotorfluxmagnitudeisconstantthentherotorcurrentvectorisperpendiculartothefluxphasor
direction.Intheequation(7.123)thereissummarizedthetwocasesofrotorfluxidentificationaresummarized.First,whentherotorflux
magnitudeiskeptataconstantvalue,and,inthesecondcase,itcanchange.Inthelattercaseatimelagappearsintheequationofthe
fluxgenerationcomponentofthestatorcurrentvector.

(7.123)

where

(.7.124)

Thereactivecurrentisinfluencedbytherotorfluxamplitudevariation,andtheactiveonebytherelativespeedbetweentherotorflux
spacephasorandtherotorsspatialposition.Comparingtheabovetwoequations,itcanbenoticedthatwhenevertherotorfluxamplitude
isnotconstant,thereactivecomponentofthestatorcurrentcontainsanewelement,proportionaltothisfluxvariationandtothe
magnetizingcurrentvariation,respectively.

7.9.1.TheMatlab/SimulinkModeloftheFieldOrientedControl

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 24/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.22.SIMULINKmodeloftherotorfluxidentification

Letusgiveanoverviewfromthepointofviewofsimulation,onhowtheFOCcanbeimplementedinMatlab/Simulinkgraphical
developmentsystem.Thedifferentialequationstreatedin[16],[14],and[18]weretakenintoaccount.Theusageofsymbolswastaken
fromthedSpacerealization,toavoidpossiblemisunderstanding.Forthecompletecontroller,thecontrollingparametersdescribedin[15]
wereadapted.Aslightmodificationtothestructureofthetheoreticalfluxmodelwaseffectuated.ThefluxmodelcanbeseeninFigure7
22.BeforecarryingoutthedivisionbypsiRd( ),wehavetomakesurethatitsvalueisnotzero,otherwiseaninfinitevaluegetsinto
thesystem,which,inturn,causesallfurthercalculationsto"blowup".Tomakesurethatthisdoesnothappen,wemust"part" from
zerobyasmallvalue.Thistaskisdonebytheblock"absmin".Thisblocksetspsi_offsetasaminimalabsolutevaluefor .This
means,thateither <psi_offsetor >psi_offsetmustbetrue.Thevalueofpsi_offsetshouldbesettoasmallvalue,sothat
psi_offset<<psi_Rd_ref,wherepsi_Rd_refisthereference,whichisinourcaseaconstantvalue.Thisisinfactamodification,whichis
implicitlyincludedinthesoftware,sinceinthesoftwareadivisionby0issaturatedtothemaximalortheminimalvalueofthefractional.
Thesineandcosinevaluesofthetransformationangleeps_Rd( )aregiventwice:oncefortransformation,andonceforinverse
transformation.Avoidingthisbycalculatingsineandcosinevaluesonlyonce,andusingthesevaluesdirectlyforbothdirectionsof
calculation,gaveresultstoamuchfasterworkingfluxmodel.LetusshownowthewholecontrolsysteminFigure724.,andtheFOC
partofitinFigure728.

Figure7.23.FOCcontroloftheinductionmotor

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 25/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.24.SIMULINKimplementationoftheFOCcontrolblock

Figure7.25.TheappliedTorque

Figure7.26.componentsofthestatorcurrent

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 26/27
6/14/2017 Chapter7.ModellingInductionMotors

Figure7.27.Rotorfluxcomponents

Figure7.28.ThespeedandreferencespeedinthecaseofFOC

Figure7.29.dqcomponentsofthestatorcurrent

Prev Next
Chapter6.Samplemeasurement Home Chapter8.SensorlessControlwithSpeedEstimators

http://www.mogi.bme.hu/TAMOP/digitalis_szervo_hajtasok_angol/ch07.html 27/27

Potrebbero piacerti anche