Assessment of a proposed hybrid photovoltaic array maximum power point tracking method

Assessment of a proposed hybrid photovoltaic array maximum power point tracking method Water Science ScienceDirect Water Science 30 (2016) 108–119 journal homepage www elsevier com/locate/wsj Full Len[.]

ScienceDirect

Water Science 30 (2016) 108–119 journal homepage: www.elsevier.com/locate/wsj

Full Length Article

aNational Research Center, Ministry of Water Resources and Irrigation, Egypt

bElectrical and Machine Power Department, Cairo University, Egypt

Received 30 July 2016; received in revised form 14 October 2016; accepted 19 October 2016

Available online 5 December 2016

Abstract

Photovoltaicarrayshavelimitedconversionefficiencyandthus,amaximumpowerpointtrackingtechniqueisessential.This makesthemaximumpower pointtracking(MPPT) requirepriorpredictionofthe mentionedpointin spiteofthe undeniable changesintheenvironment.InthismanuscriptanintroductionandassessmentofthedifferenttechniquesofMPPTispresented ThecategorizationschemeoftheMPPTtechniquesisaccordingtoeitherthepredefinitionofoperatingpointswithoutsystemdata update(offlinemethods)orcontinuoussamplingofsystemvariables,toupdatethePVmodulemeasurements(onlinemethods) Whereashybridmethodisacombinationofboth.AnumberoftechniquesfromeachclassweresimulatedinMATLAB/Simulink environmentinordertocomparetheirperformance.Moreover,thehybridmethodwassimulatedintwosuccessivestepswithout pre-assumptionoftheoutputoftheofflinemethod.Theresultsdemonstratedtherelevanceofthehybridmethodwhenappliedtoa photovoltaicsystemduetoitsgoodperformance,fastresponseandlessfluctuations,whensubjectedtosuddenclimaticchanges

©2016NationalWaterResearchCenter.Productionandhostingby ElsevierB.V.ThisisanopenaccessarticleundertheCC BY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/)

Keywords:PV systems; MPPT; Online methods; Offline methods; Hybrid methods

1 Introduction

Recentlytheuseofsolarenergyhasbeenemerging.Themainadvantagesofphotovoltaic(PV)systemsarezero greenhousegasemission,lowmaintenancecosts,fewerlimitationswithregardtositeofinstallationandabsenceof mechanicalnoisearisingfrommoving parts(Reisietal., 2013).Theglobal PVmarket reached173GW in2014 However,therearethreemajorlimitationsinphotovoltaicgenerationsystems: theconversionefficiency toelectric powerislow(9–17%),thevariabilityofelectricpowergeneratedduetotheweatherconditionsandsunlighthoursat

∗Correspondingauthor.

E-mail addresses:yasmin.adel1982@hotmail.com (Y Adel), rameens@hotmail.com (R Abdelhady), drahmed@nahdetmisr.com

(A.M Ibrahim).

Peer review under responsibility of National Water Research Center.

http://dx.doi.org/10.1016/j.wsj.2016.10.004

1110-4929/© 2016 National Water Research Center Production and hosting by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Fig 1 Single diode equivalent circuit of a solar cell.

daytime(Xiaoetal.,2006),andtherelativelyhighercost(12.5¢/kWh)ascomparedtothatproducedbyconventional powergeneration(7.4¢/kWhfornaturalgascombinedcycle)systemsoreventootherrenewablesourcessuchashydro electricalenergy(8.4¢/kWh)(EnergyInnovation,2015)

Asolarcell(alsocalledphotovoltaiccellorphotoelectriccell)isasolidstatedevicethatconvertstheenergyof sunlightdirectlyintoelectricitybythephotovoltaiceffect(Zhangetal.,2011).Assembliesofcellsareusedtomakesolar modules,alsoknownassolarpanels.Photovoltaicsystemusesvariousmaterialsandtechnologiessuchascrystalline Silicon(c-Si),Cadmiumtelluride(CdTe),Galliumarsenide(GaAs),chalcopyritefilmsofCopper-Indium-Selenide (CuInSe2)

Solarcellsexhibitanon-linear current–voltagecharacteristicsthat dependonsolarradiationsandtemperatures Thereisapointonthecharacteristicscurvewheretheoutputpowerfromthearray(astringofpanels)hasamaximum value.Solarcellsareusuallyassessedbymeasuringthe currentvoltagecharacteristicsof thedeviceunderspecific conditionsofilluminationandthendeterminingasetofparameters.Inordertoensureefficientoperationofthesolar arraythemaximumpowerpoint(MPP)ofthearrayhastobetracked.MPPT,inadditiontorisingthepowerdelivered fromthePVmoduletotheload,isconsideredasaPVsystemlifetimebooster(Bahgatetal.,2005)

1.1 PV system

For simplicityin analyzing characteristics of solar cells, electrical equivalentcircuits are used in representing them.Researchersusednumerousequivalentcircuitstohelpinpredictingthebehaviorundervariousenvironmental conditions,andfurtherinobtaining(I–V)and(P–V)characteristiccurves.Themostcommonlyusedequivalentcircuits arethesinglediodemodel(DeBlasetal.,2002)(showninFig.1),thedoublediodemodel(CabestanyandCastaner,

1983)andthethreediodemodel(Khannaetal.,2015)

According totheabove circuitthe solar cellcan berepresentedby: acurrent generator,adiodeindicatingthe recombinationlosses,ashuntresistancesymbolizinglossesfromcurrentsthatreturnacrossthejunctionandaseries resistancedenotingresistancelosses.ByapplyingasimpleKirchhoff’scurrentlaw(KCL);therelationofthecurrent

IandthevoltageVisgivenas:

I=I ph−I o (e q(V nkT +IRs) −1)−V +IR s

whereI ph isthephotocurrent,I o isthesaturationcurrent ofthe diode,R s istheseriesresistance,R sh istheshunt resistance,nisthediodeidealityfactor,kisBoltzmann’sconstant(1.4×10−23),qistheelectroncharge(1.6×10−19),

TistheabsolutetemperatureinKelvin

110 Y Adel et al / Water Science 30 (2016) 108–119

Fig 2 Typical IV/PV characteristic of a solar cell at 1 kW/m2and different temperatures.

Fig 3 PV characteristic of a solar cell at 75 ◦Canddifferentilluminationlevels.

ByplottingtheI–VcurveintherelationinEq.(1),theexistenceof auniquepointisnoticed;thespotnearthe kneeoftheI–Vcurve(Fig.2).Thisisthepointatwhichtheproductofcurrentandvoltageachievesitsmaximum, whichisnoticedmoreontheP–Vcurves(Figs.2and3).Thispoint,asseenfromthefiguresvariesaccordingtothe environmentalconditions(illuminationandtemperature).ManyapproacheshavebeenintroducedtodeducetheMPP fromEq.(1)whichimpliesdeterminingthemajorparameters:thediodesaturationcurrent,theseriesresistance,the idealityfactor,thephotocurrentandtheshuntresistance.SincetheIVequationisimplicit(currentexistsonbothsides

oftheequation),extractingtheparametersbythesimpleleastsquaremethodisnotpossible.Alsothedetermination

Fig 4 System block diagram.

requiresmeasuredIVcharacteristicsatthespecifiedinsolationandtemperature.Zhangetal.(2011)usedanexplicit analyticexpressionforIwiththehelpofLambertωfunctiontobeabletoutilizetheconventionalcurvefittingmethods (Eq.(2))

I−

(1/ (a+b)) V − (b (c+d) / (a+b)) + (e/a) × lambertω



((adb) / (e (ab))) exp

e (a+b)



((ca) + (da) + V )



(2)

where:a = R s , b = R sh , c = I L , d = I oande = (nkT)/q.

AfterconcludingtheseparameterstheMPPcanbededucedusingtheaboveequationandthefactthat∂P/∂V V =Vm=

0atVmax

ApracticalsolutionrequirestheintroductionofatrackerwhichisinsertedbetweenthePVsystemandtheload

ADC–DCconverterisusedtomanagethepowerandacontrollerisintroducedwithinthetrackertocompensatethe parametervariationduetoenvironmentalconditions

VariousMPPTmethodshavebeendeveloped.Thesemethodscanbegroupedbasedondifferentfeatures.Inessence, MPPTmethodsarecategorizedinto:offlinemethodswhicharedependentonsolarcellmodels,onlinemethodswhich areusuallyreferredtoasthemodel-freemethodandhybridmethodwhichrepresentsasequenceoftheofflineand onlinemethods.Inthismanuscriptahybridmethodwasassessedbysimulationintwosuccessivestepswithout pre-assumptionoftheoutputoftheofflinemethod.Alsosimulationofdifferenttechniquesunderonlineandofflinemethods andcomparisonwiththehybridmethodispresented

2.1 System overview

Ingeneral,anyPVsystemconsistsof:PVarray,DC–DCconverter,MPPTcontrollerandabattery.Thebatteryis discardedasthemainobjectiveistheMPPTofthesolararray.Fig.4showstheblockdiagramofthesystemusedin thissimulationmodel

ThePVarrayused inthismodel isof typeSunPowerSPR-305-WHTwhichconsistsof20parallelstringsand

5series-connectedsolarmodulesperstring.ThemodulecharacteristicsunderSTC(1kW/m2and25◦C)are:64.2V

V OC,5.96AI SC,54.7VV MPPand5.58AI MPP

TheDC–DCconvertersareusedformatchingthecharacteristicsoftheloadwiththoseofthesolarpanelsi.e.to balancethesystem.Inourmodelthesimulatedconverterisa5kHz–500Vboostconverteranda100ohmresistance

isusedasload.ThecontrolsignalgeneratedbyseveralMPPTmethodsfeedtheboostconverterswitch

2.2 The MPPT control

Inthisresearchcomparisonofdifferentconventionaltrackingtechniqueswiththeintroducedmethodandassessment

ofallroutinesunderdifferentclimaticconditions,iscarriedout.Theradiationandrapidchangeinradiation(shadowing) weretakenintoaccount.Thetemperaturewasdiscardedavoidingcomplexity.Foragivensolarradiation,whenthe celltemperatureincreases,theV oc,dropsslightly,whiletheI sccurrentincreasesconsiderablywhichmakesthetotal effectmarginal(Salmietal.,2012)

112 Y Adel et al / Water Science 30 (2016) 108–119

TheMPPTmethodsweresimulatedinMATLAB/Simulinksoftwareenvironment.MATLAB/Simulinkisselected, duetoitsreusability,extendibility,andflexibilityinsuchsystems.Amongthe offlinetechniquessimulated inthis researchweretheopencircuitvoltage(OCV)andtheshortcircuitcurrent(SCC)techniquesdiscussedbelow.MPPT basedontheonlinemethodsofperturbandobserve(P&O)andincrementalconductance(IncCond)alsohavebeen performedtocomparebetweenofflineandonlinetechniquesingeneral.Theproposedhybridmethodisthensimulated, thecontrolsignalsconsistsof twostages:setpointfiguringandtuning.Thefirst stageisthe setpointloopwhich approximatesthemaximumpowerbytheVOCofflinemethod.ThenextstageisusingtheinputfromtheVOCmethod

asaninitialconditiontotheP&Oonlinemethod,whichisthesecondphaseoftrackingtheMPP(finetuningloop)

2.2.1 Offline methods

Offlinemethodsgenerallyrequiretheknowledgeofoneormoreofthesolarpanelparametersvalues,suchasthe opencircuitvoltage(V OC),shortcircuitcurrent(I SC),temperatureandradiation.Thesevaluesgeneratethecontrol signalnecessaryfordrivingthesolarcelltoitsMPP.Inthetrackingoperation,thiscontrolsignalremainsconstantif ambientconditionscanberegardedasfixedandtherearenoattemptstoregulatetheoutputpowerofthePVsystem (Reisietal.,2013)

Offlinetechniquesinclude:opencircuitvoltagemethod(OCV),shortcircuitcurrentmethod(SCC)andartificial intelligence(AI)

TheOCVtechniqueisoneofthemoststraightforwardofflinemethods.Itusesthealmostlinearrelationshipbetween theopencircuitvoltage(V OC)andtheMPPvoltage(V MPP)underchangedclimaticcircumstancesasdescribedbythe followingequation:

wherekisaconstant,dependingonthesolarcellcharacteristics

TheconstantkisderivedempiricallyaftermeasuringseveralV OCandV MPPforanyspecificcellunderdifferent climaticconditions.Valuesfortheconstantk,ingeneral,varybetween0.73to0.80(SchoemanandWyk,1982),here

inourspecificcase,theconstantkisassumedtobe0.8

Despitethesimpleimplementation(hastodowithhardwareusedtoimplementthismethod)andlowcosts,this techniquesuffersfromtwomaindrawbacks:theMPPtrackingisnotpreciseandmeasurementofV OCnecessitates periodicsheddingoftheload.TherecanbenoneedtoshedtheloadinordertomeasuretheV OCaspilotcells(Brunelli

etal.,2009),whosecharacteristicsrepresentthoseoftheoriginalPVarray,mightbeused

AnotherofflinetechniqueistheSCCwhichdependsonthefactthatI MPP isapproximatelylinearlyrelatedtothe

I SCofthePVarray.AlthoughthismethodismoreaccuratethantheOCVapproachyettheimplementationismore complexandtheperiodicsheddingoftheloadortheexploitationofpilotcells,tomeasuretheI SChasnotbeenavoided ThethirdofflinetechniqueisAIwhichhasseveraldisciplinesunderitincludingartificialneuralnetworks(ANNs)and fuzzylogic(FL)

2.2.2 Online methods

Inonlinemethods,alsoknownasmodel-freemethods,usuallytheinstantaneousvaluesofthePVoutputvoltage

orcurrentareusedtogeneratethecontrolsignals(Reisietal.,2013).Thetechniquesundertheonlinemethodsare basedontheprincipleoftheoptimalcontroltheory.Thesetechniquesincludeperturbationandobservationtechnique (P&O),andtheincrementalconductancetechnique(IncCond)

TheP&Otechnique,isoneofthemostsimpleonlinemethodswhich,hasbeentakenintoaccountbyanumberof researchers(Wasynczuk,1983;HuaandLin,1996;Huaetal.,1998).P&Oisimplementedbyapplyingsmallconstant perturbationstothevoltageorthecurrentsignalofthesolarpanel(V ref,I ref).Itsworkingprincipledependsonthe samplingmethod.Thearrayvoltagehereisdecreasedbyincreasingthedutycycleoftheboostconverterandvice versa.Thedutycycleoftheboostconverterisbetween0and1andtheincrementused todecreaseorincreasethe voltageisintheorderof10−4fromliterature.

Aftereachperturbation,theoutputpowerismeasured.Iftheoutputpowerisgreaterinvaluethanthepower of the preceding step,powerwillmovetowardtheMPP(theleftsideofthePVcurve),thereforeavoltageperturbation

ofthesamesignmustbeappliedinthesucceedingstep.Asmallervalueofpower,onthecontrary,infersthatpower hasdeviatedfromMPP(therightsideofthePVcurve),andaperturbationofoppositesignwillhavetobeapplied

Fig 5 Flowchart of P&O technique.

(Reisietal.,2013).ThesecontinuousperturbationseventuallycausetheapproachtotheMPP.Thisflowchartofthis procedureisshowninFig.5

TheP&Otechniquehastwodrawbacks:thecontinuousperturbationscausestheoscillationaroundtheMPPand neverreachingitactually.AlsoP&Otechniquecanfailundervaryingatmosphericconditions(Reisietal.,2013).In theP&Othevalueofthe perturbationsappliedtothesystemisthe mainfactorcontrollingthe convergenceofthe outputpowertotheMPP

Intheincrementalconductancetechniqueas thealgorithmtakestwosamplesofvoltageandcurrenttocompute MPP.ThistechniqueisbasedonthefactthattheslopeofthePVarraypowercurveiszeroattheMPP,positivefor valuesofoutputpowersmallerthanMPP,andnegativeforvaluesoftheoutputpowergreaterthanMPP(Irisawaetal., 2000;HohmandRopp,2003;KoizumiandKurokawa,2005;Harada andZhao,1993).Themainlimitationofthe IncCondmethodisthatitobligescomplexcontrolcircuitrywhichmighthaveresultedinahighcostsystempreviously

2.2.3 Hybrid method

Thehybridmethodisacombinationofboththeabovemethods.Itisdivided intwosuccessivesignals:thefirst dependingonanotaccurateofflinepredictionoftheMPPTtakingintoconsiderationtheatmosphericconditionsthen applyingoneoftheonlinetechniques,followingthesameworkflowinFig.5,toaccuratelydeducetheMPPT.This methodeliminatedthedrawbacksoftheothertwo

114 Y Adel et al / Water Science 30 (2016) 108–119

Fig 6 Output generated power from PV system without MPPT control at different radiation levels.

Inthismanuscriptthehybridmethodisapplied,twocontrolsignalsaregeneratednotjustapre-assumptionofthe outputofthe offlinemethod.The firstsignalisgeneratedfromtheOCVtechniqueandthesecond fromtheP&O techniquefollowingtheworkflowinFig.5.Thiscombinationeliminatedthedisadvantages ofbothtechniques:the inaccuracyoftheOCVtechnique,sincethemaximumpointreachedisnotthefinalandmisestimatingtheinitialpoint

(V ref)intheP&Otechnique,asitistheoutputfromthefirsttechnique

Also,comparisonofdifferentconventionaltrackingtechniqueswiththeintroducedmethodandassessmentofall methodsunderdifferentclimaticconditions,isproposed(VOCtechnique)(P&Otechnique)

3 Results

MPPTisanessential component inthePV systemsinceit allowsanincrease inthe powerdeliveredfrom the

PVmoduletotheload,anditalsobooststhefunctioninglifetimeofthePVsystem.Fig.6showsthePVgenerated powerperformancewhenthereisnoMPPTcontroltechniqueappliedtothesystem.Inthefigurethepowergenerated diminishestozeroforanincrementofasecond(reaching0.1sfor1000W/m2),thegeneratedpowerfromthearray thenrisestoalevelnotmorethanonetenthoftheMPPinhighradiationlevels

ComparisonoftheefficiencyofdifferentMPPTmethodsunderdifferenttechniquesandtheproposedtechnique wascarriedoutatconstanttemperature(25◦C)andthreevariationsinradiationlevel(250,750,1000W/m2).The efficiencywascalculatedbythefollowingformulashowninEq.(4):

η=



P max.Tracked

P max.



whereηistheefficiencyofthetrackingmethod,Pmax.TrackedistheMPPtrackedandPmax.istheMPPasmeasuredin datasheetofthearray

SimulationresultsoftheOCVtechniqueandtheSCCtechnique(Figs.7and8)showthattheOCVmethodmaintains

arangeof85%efficiencyathighilluminationlevelsyettheefficiencydropsslightlyto82%atlowilluminationlevels

Ontheotherhand,SCCmethodexhibitsanextremelylowefficiencyatlowilluminationlevels

Fortheonlinetechniques(Figs.9and10);theefficiencyoftheP&Oreached97%.TheP&Omethod’sefficiency showeddirectproportionalitytotheilluminationlevels,reaching92%at250W/m2.TheIncCondmethodshowedthe

Fig 7 Output generated power from PV system with OCV MPPT control at various radiation levels.

Fig 8 Output generated power from PV system withSCCMPPT control at various radiation levels.

highestefficiencywhichreached99%atlowilluminationlevels.Theintroducedhybridmethodshowedthehighest efficiencyofallreaching99%,sincethetrackingmethodisamingleofbothmethods:theonlineandoffline(Fig.11)

Figs.7–9showfluctuationssincethesimulationdurationis3s.Forsometechniquesittakesmoretimethanothers

toreachsteadystate.AsfortheP&OandIncCondtechniquesthefluctuationsexistregardlesstheirhighefficiency

116 Y Adel et al / Water Science 30 (2016) 108–119

Fig 9 Output generated power from PV system with P&O MPPT control at various radiation levels.

Fig 10 Output generated power from PV system with IncCond MPPT control at various radiation levels.

Fig 11 Output generated power from PV system with hybrid MPPT control at various radiation levels.

Fig 12 Output generated power from PV system with hybrid MPPT control at rapid change in radiation level.