Integrated CO2 system with HVAC and hot water for hotels: Field measurements and performance evaluation

ContentslistsavailableatScienceDirect

International Journal of Refrigeration

journalhomepage:www.elsevier.com/locate/ijrefrig

Integrated CO ₂ system with HVAC and hot water for hotels: Field measurements and performance evaluation

S. Smitt

^∗

, I. Tolstorebrov, A. Hafner

Norwegian University of Science and Technology, Kolbjørn Hejes vei 1D, Trondheim 7491, Norway

a rt i c l e i n f o

Article history:

Received 13 December 2019 Revised 20 March 2020 Accepted 23 March 2020 Available online 8 April 2020 Keywords:

R744 Heat pump HVAC Hot water Thermal storage Hotel energy systems

a b s t r a c t

ThisstudyinvestigatestheperformanceofanintegratedCO2(R744)heatpumpandchillerunitinaNor- wegianhotel.Thesystemconsistsofasingleunitforheating,coolingandhotwaterwithanintegrated thermalstorage.Thethermalsystemofthehotelisdescribedanddatafromthefirstyearofoperation areanalyzed.Usingthefieldmeasurements,hotwaterloadsandCOPsarecalculatedandaveragedto20- minuteintervals.TheheatingandcoolingcapacitiessuppliedbytheR744unitarestudiedonaweekly andmonthlybasistoevaluatetheseasonalbehaviorofthesystem.Thehotwaterstorageholdsanenergy capacityof350kWhatfullychargedconditionsanddemonstratespeakdemandreductionsofmorethan 100kWduringa2-dayperiod.Theresultsshowthatthehotwaterusageaccountsfor52%oftheannual heatloadofthehotel.EnergyefficiencyanalysisoftheintegratedR744systemrevealsanannualsystem SCOPof2.90,and thusanuntappedsystempotential thatcan beexploitedbyincreasingtheAC load deliveredbytheR744unit.Otherfactorsthatgreatlyinfluencetheefficiencyofthesystemarevariations intheambienttemperatureandhigh gascoolerexittemperatures.Thelatterisoftenaresultofhigh temperaturesinthewaterreturningfromthesubsystemsofthehotel.Thiscanbeimprovedbyreducing thenumberofstartsandstopsoftheR744unitandbyinsuringstratificationinhotwatertanks.

© 2020TheAuthor(s).PublishedbyElsevierLtd.

ThisisanopenaccessarticleundertheCCBYlicense.(http://creativecommons.org/licenses/by/4.0/)

Système de CVC et de production d’eau chaude intégré au CO ₂ pour les hôtels : mesures sur le terrain et évaluation des performances

Mots-clés: R744; Pompe à chaleur ; CVC; Eau chaude; Stockage de chaleur; Systèmes énergétiques dans l’hôtellerie

1. Introduction

Inordertosecureasustainablefuture,itisnecessarytoadopt moreefficientmeansofconverting,storingandusingthermalen- ergy. Buildingsaredirectlyresponsibleformorethan40%ofend- useenergyconsumption andCO₂ emissionsinthe EU(EC,2010).

Non-residential buildings, which are largely represented by the commercialsector, accountfor35% oftheenergyuseandrelated emissions(Eurostat,2017).Thepotentialenergysavingswithinthe commercialsector isestimatedto30%,whichcan beachievedby implementing measures to manage demand and increase energy

∗ Corresponding author.

E-mail addresses: [email protected] (S. Smitt), [email protected] (I.

Tolstorebrov), [email protected] (A. Hafner).

efficiency(Economidou etal., 2011; EC,2006). Hotelsare catego- rizedashighenergydemandingbuildings,duetotheiroperational characteristicsandthebehaviorofoccupants(HES,2011).Theap- plication of conventional thermalenergy sources in hotels is ex- tensive,suchasfossilfuelsandelectricboilersforheatinginlarge inefficient central systems (Daltonet al., 2008). Excessive use of electricalpowerbypeakheatingandtheuseoflow-efficiencyair- conditioning(AC)units aggravatethe electricityproblems society isfacing. Existinghotels exhibit the mostsevereproblems ofex- cessively highenergy demand rates, inevitablyrequiring renova- tionsalongwithretrofittingofthermalsystems(Santamourisetal., 1996).

Vapor compression systems are among the most energy- efficient methods of providing heating and cooling in buildings (Liu etal., 2017). An increased focuson environmentally friendly https://doi.org/10.1016/j.ijrefrig.2020.03.021

0140-7007/© 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license. ( http://creativecommons.org/licenses/by/4.0/ )

Nomenclature

COP CoefficientofPerformance DHW DomesticHotWater SH Spaceheating T Temperature[^◦C]

V volume[l]

C_p specificheatcapacity[kWhkg⁻¹ K⁻¹] ref reference

i timestepindex set setpoint E Energy[kWh]

AC AirConditioning F.S. FullScale

Q˙ coolingorheatingload[kW]

˙

m massflowrate[kgs⁻¹] W˙ power[kW]

P Pressure[bar]

HPWH HeatPumpWaterHeater comp compressors

fans evaporatorfans pumps allsystempumps

aux,el auxiliaryelectricalsystems evap evaporation

w supplywater exit exit

gc gascooler

a ambient

avg average SH spaceheating DHW DomesticHotWater AC Airconditioning sys system

min minimum

max maximum

ch DHWcharging nch NoDHWcharging

HVAC Heating,Ventilation,andAir-Conditioning usage consumptionbyendusers

supply supplybyheatpump HFC Hydrofluorocarbon Greeksymbols

^change

ρ

^{density[kgm−3]}

solutions together with a global effort to reduce the application offluorinatedgasesisstrengtheningthepositionofnaturalrefrig- erants(UNEP,2016; EPandEC,2014). Carbondioxide(R744) isa naturalrefrigerant with negligible environmental impact and fa- vorablethermodynamic properties(Lorentzen, 1994; Gulloet al., 2019; Ciconkov, 2018). It is inexpensive, readily available and is neither flammable nor toxic. These qualities make R744 suitable inapplicationswhereothernaturalrefrigerants,such asammonia and propane, are challenging due to safety concerns (Bolaji and Huan, 2013). R744 is firmly established in both heating and re- frigerationapplications and is acceptedas a viable alternative in severalsectors,e.g.supermarket,transportation,domestichotwa- ter(DHW)heatpumpsandindustrialprocesses(Gulloetal.,2018;

Hafner,2015;Nekså etal.,2010).Thedistinctivetemperatureglide of R744 in the gas cooler during transcritical operations allows for efficient heating of water (Nekså, 2002; Nekså et al., 1998), evenuptotemperaturesof90^◦C(Bamigbetanetal.,2017).Inthe Japanesemarketalone,morethan5million R744heatpumpwa- ter heaters(HPWHs) are installed (Shecco, 2016). However, asil-

lustrated by Cecchinato etal. (2005),a suitable R744 heat pump design is imperative to ensure a high efficiency when compared withhydrofluorocarbon(HFC)installations,suchasR134a.Intheir laterwork,Cecchinatoetal.(2010)identifiedcompressorcapacity rate andsecondary fluid temperatures as key influencing factors onoptimumR744highpressure,andthuscycleefficiency.Minetto (2011)presentedexperimentalresultsfromthedevelopmentofan R744air/waterHPWHforresidentialbuildings,andalsoconcluded that optimum operating high-pressure conditions are highly de- pendenton both inlet temperatureandproductionsetpoint tem- peratureforhotwater.Severalother workshavetackledthehigh- pressurecontrolproblemtoachieve maximumcyclecoefficientof performance(COP)(Yangetal.,2015;Huetal.,2015;Wangetal., 2013;Cecchinatoetal.,2012).Thedesignandoperationofthesec- ondarysystem,especiallytheDHWstorage,isequallyimportantto ensurehighefficiencyinR744HPWHinstallations.Itisfirmlyes- tablishedthatreducingthereturntemperaturefromthesecondary systemtothegascoolerwilllimittheR744gascooleroutlettem- perature,andthus enhance cycleCOP (Lorentzen,1994). Thermal stratification of the DHW storage should therefore be employed to reduce mixingandensure thereturn ofcold waterto thegas cooler(Fernandezetal.,2010).Theimpactofthereturntempera- tureofwateroncycleefficiency,withrespecttoambient airand citywatertemperatures,wasillustratedbyYokoyamaetal.(2007). They concluded that the R744 HPWH efficiency doesnot always increase with ambient temperature, as the storage efficiency de- creaseswiththeincreaseofcitywatertemperatures.BesidesDHW, anotherapplicationofthetranscriticalR744heatpumpisacom- binedheatsupplysystemforspaceheating(SH)andDHWbythe means of severalgas coolers in series(Stene,2005; Heinz et al., 2010).

R744systemshavealongtraditioninrefrigerationprocesses.In the European supermarket sector alone, morethan 16,000stores are relying onR744, where14% of theinstallations are operating in the transcritical region (Ska˘canová and De Oña, 2019). Trans- criticalR744systemswithintegratedheatingandcoolingapplica- tionsare traditionallyfoundwithin thissector, whereexcessheat isrecovered asa byproductof therefrigeration process(Pardiñas etal.,2018;Hafner,2017;Girotto,2016).Combinedoperationswith heatrecovery highlyenhancetheperformance oftheR744refrig- eration system(KarampourandSawalha,2017),andcanbe espe- ciallybeneficialinwarmclimateapplications,asdemonstratedby Gullo(2019).However,thecontrolstrategyduringtheseoperations ofheatingandcooling can highlyinfluencetheefficiency (Sarkar etal.,2004;Sarkaretal.,2006).Waterstorageunitscanbeapplied tocompensateforasynchronousheatingandcoolingdemands,and reduce peak load operation (D’Agaro et al., 2019; Polzot et al., 2016). Integratedheating, ventilation,air-conditioning(HVAC)and DHW systems for buildings are widely applied (Fabrizio et al., 2014; Chua et al., 2010; Omer, 2008). However, applications of R744integratedHVACandDHWsystemsoutsidethesupermarket sector are not well-established nor sufficiently documented. The currentstatusofR744systemsprovesthepotentialbenefitsofim- plementingintegratedR744inbuildingswithlargeDHWdemands, suchashotels.Byrneetal.(2009)conductedatheoreticalcompar- isonbetweenanintegratedR744unit forHVACandDHWwitha state-of-the-artR407Asystem,andfoundtheenergyperformances comparable.Minettoetal.(2016)presentedawater-sidereversible R744 HVAC and DHW unit that operated highly efficient during DHW production.However, COP wassignificantly reducedduring the SH heatingmode, due to highreturn temperatures fromthe heatingsystem. Tosato et al.(2019) presented a layout of an in- tegratedHVAC andDHWR744unit for ahotel located inNorth- ern Italy, whereground-water was utilizedas a heat source. Re- sultsfromachargingcycleofthe1.5m³ DHWstoragerevealeda COPof4.1duringtheprocess.Asofyet,nostudieshavebeencon-

ductionoflongtermoperationsofR744systemsinhotels.Atthe sametime,thereisaneedtoevaluatethesesystemswithrespect toDHWstoragecapacitiesduringdifferentoperationalmodes,e.g.

charging and discharging. This paper presents the evaluation of long-term logged data from an integrated R744 unit installed in ahotelinNorway.A6m³DHWstorageisincludedinthethermal systemforpeakloadshaving,theoperationofwhichispresented anddiscussed.

2. Systemdescription

The R744systemanalyzed inthiswork ispartof theexisting heating systemfor a medium-size hotel in Værnes, Norway. The systemprovidesHVACandDHWforafloorareaofapproximately 9000m²,whichincludes157guestrooms.Thehotelwasbuiltin 1987, with an annual heat energy demand of approximately 1.2 GWh prior to the refurbishment of the thermal system in June 2018. The annual heat demand was reduced to approximately 1 GWhfollowingtherefurbishment.Thelocationofthehotelischar- acterizedbycoldclimateconditionswithanormalizedaveragean- nual temperature of5.3 ^◦C and 4276heating degree days(HDD) (average from1961to 1990).HDDforthelocation ofthehotelis calculatedasdescribedinThom(1954)withScandinavianstandard values(Skaugenetal.,2002). Theannualaveragetemperaturefor the first year of operation was recordedto be 6.8 ^◦C with 3860 HDDovertheperiodfromSeptember2018toSeptember2019.

2.1. R744heatpumpandchillerunit

The previous thermal system of the hotel, consisting of an electric-andoilboiler,hasbeenreplacedwiththeR744heatpump and chiller system. The first 6 months of operation revealed a monthlyenergy-savingpotentialof59–69%(Smittetal.,2019).The installed heatingand ACcooling capacityis 280kW and 75kW, respectively.Inthispaper,ACisdefinedastheventilationcooling load.Fig.1illustratestheconfigurationoftheR744heatpumpand chiller unit andsecondary distribution systems. The R744unit is an adapted single-stage supermarket refrigeration unit with heat recovery towards two separate hydronic circuits. The compres- sorrackconsistsoffourparallel compressors(displacementrange from17.8 to21.2 m⁻³h⁻¹ at50Hz). Onecompressoris equipped withavariablespeeddrive(VSD),whilethethreeothercompres- sorsarecontrolledbyON/OFF.Thecompressorsareactivatedbased on the requested capacity. The VSD compressor is always active tomeetthecapacitysetpointbetweentheconstantcapacitysteps providedbytheothercompressors.

The main function of the R744 system is to provide heating and DHW for the hotel, which is achieved with the same strat- egy as for heat recovery in transcritical R744 supermarket units (Danfoss,2015;Danfoss, 2012),with theexception that theheat- ing load, rather than the cooling load, is the controlling param- eter. The capacity control of the R744unit is based on feedback signalsfromthehotel,suchasfromtheDHW,ventilation-andra- diatorcircuits.Thebuildingside suppliesaheatdemandsignalto the R744controller, whichadjusts thesetpoints forthecompres- sorcapacityandthehighpressure.Ifanincreaseincapacityisre- quested,thesetpointfortheevaporationtemperatureistemporar- ilyreducedtoactivateanothercompressorintherack.Thissome- whatunconventionalcontrolisduetotheconversionoftheR744 unit fromasupermarketrefrigerationrig.ThehighpressurePgcis regulated based on the gascooler outlet temperatureT_gc,exit. The control principle of the gascooler pressure is describedin Gullo et al. (2016). The highpressure control valve (EV1) expands the fluiddirectlytotheliquidseparatoratanintermediatepressureof 38–55bar.Fourairevaporators(50kWat−15^◦C)arefedfromthe liquidreceiver. Thermostaticexpansionvalves (EV2–EV5)regulate

thesuperheatattheexitofeachevaporator.Thenumberofactive evaporatorsis dependent on the heatingload. The gas returning fromtheevaporatorsismixedwithflash gasandisdirectedin a passagethroughtheliquidreceiverforheatexchangebeforecom- pression.Also,aheatexchanger(HX)interface(75kWat12/7^◦C) tothechilledwatercircuit(HX6)canbeemployedtorecoverheat ifACisneeded.ThechilledwaterproducedbytheR744systemis usedto supplementthe existingAC chillerunit, andisthus only appliedasanauxiliaryfunctionduringheatgeneration.

2.2.Subsystemsandhotwaterstorage

Thesystem isdesignedto supplyheat forventilationheating, DHWandSH.Heatissuppliedtothehydronicsubsystemsthrough twogascoolersinseries,GC1andGC2asshowninFig.1,athigh (>60^◦C)andmedium(<50^◦C)temperatures.Themediumtem- perature(MT)circuitprovidesheatprimarilytoventilationbatter- iesanda radiator/floorheatingcircuit. Remainingheatisused to preheatDHWthroughHX2fromapproximately8to30^◦C.During winteroperations,HX1intheMTcircuit isappliedfordefrosting oftheevaporatorsthroughabrinecircuit.

The high temperature (HT) circuit mainly supplies heat for DHWreheatthroughHX3.Duringoperationalconditionswithneg- ligible SH demand, the entirety of the DHW production can be covered with HX3. HX2 is then bypassed with valve MV1. Simi- larly,GC1 can be bypassed through the directional valve, DV1, if theDHWstorageisfullychargedandthere isnodemandforHT heat.TheR744unit willintheseinstancesoperateatasubcritical high-pressurelevel.TheHTcircuitalsosuppliesextraheatthought HX5fortheradiators andfloorheatingduringwinterconditions.

Thisistypicallydonewhenthesetpointtemperatureoftheradia- torsexceedsthesetpointoftheMTcircuit.Whenthereturntem- peratureishigherthanthesetpointtemperatureoftheMTcircuit, MV5closesoff thepassagebetweenMTandtheradiators.Ashunt circuit is then established exclusively betweenthe radiators and HX5,topreventanincreaseinwatertemperaturetoGC2.

TheDHWsubsystem consistsofseveraltanksin serieswitha combinedvolume of6 m³. The subsystem is suppliedwith heat from the R744 unit through HX2 and HX3, or from the backup electricboilerthroughHX4.Thesystemcontrolischaracterizedby twodistinctive modesofoperationdependingonthestate ofthe DHWstorageandwhetheractivechargingisneeded.Whenactive chargingofthestorageisunnecessary,themajorityoftheheating loadisallocatedtotheMTcircuittocoverthe moderatetemper- aturedemands,e.g. radiators,floorandventilationheating.Excess heatisallocatedto theDHWsubsystem,usually atalow loadto meettherequiredDHWtemperature.

Thesecondmodeofoperationoccursduringactivechargingof theDHWstorageandisactivatedwhenthetemperaturesintanks 1or3fallbelowathreshold.Afewstepsareinitiatedtostartthe chargingprocess.First,thesetpointofPumpP4ischangedtopro- videahigherflowrate.Then,theheatdemandfromthebuildingis thengivenanoffsetsignaltoinducecharging.Theincreaseinde- mandtriggersanincreaseincompressorload,whichismaintained by temperature insuranceof the HT andMT circuit supply tem- peratures.Duringthechargingprocess,excesshotwaterisstored andmovesthrough the seriesof tanks asthe buffer isgradually chargedfromtankno. 1tono.10. Waterisdrawnfromtank no.

10andissentthroughtheheatingprocess,inthesamemanneras describedbyMinetto(2011).The thermalstorageisfullycharged whenthe normallystratified storagereaches a highanduniform temperature.Duringdischarge,waterisdrawnfromtankno.1and ismixed inMV3 with coldwater to a temperatureof 55^◦C be- foreentering thesupplyline.ThesetpointforDHWproductionis 66^◦C.Onceaweek,thesetpointtemperatureisboostedto86^◦C, duringwhichalltanksmustmeetthesetpointtemperatureforat

Fig. 1. Schematic drawing of the R744 heat pump and chiller unit with thermal storage and secondary system.

leastonehourtopreventlegionellagrowth.Thestartsignalforthe timerisresetifthetemperaturelevelhasbeenreached.As addi- tional insurance, heating elements are installed in each tank for temperatureboostingpurposes.

3. Datacollectionandevaluationmethods

The secondaryhydronicsystemisinstrumentedwithtempera- turesensors(NTC10thermistors, ± 0.2K)andmassflowmeters (oscillatormassflowsensor,class2)ineveryfluidbranch.Temper- aturesensorsintheDHWtanksandsecondarysystemshavebeen validatedto operatewithin a rangeof ± 0.1K. Heatflow me- tersforsecondaryfluidsareinstalledateveryHX(PT500tempera- turesensors,oscillatormassflowsensor,class2).Pressuresensors ( ± 0.3% atfull scale), temperature sensors (PT500 temperature sensors, ± 0.15+0.002T) andelectrical powersupplymonitors (energyanalyzerincontrolunit, ± 2%)areinstalledintheR744 unit.Thereal-timefieldmeasurementsofthehotelhavebeenob- tained via the web-monitoring software IWMAC (IWMAC, 2019).

Themeasurementsareupdatedcontinuously,butthedataatacer- tain time is only loggedby the measurement system ifit differs fromthevalueintheprevioustimestep.Therecordeddatapoints are thereforeregardedasconstantstep valueswithinthe specific time intervaluntil thenext recordedvalue. Allthe recordeddata have been resampled and synchronized to the same time step, using the weighted average of the time intervals. The data used inthisanalysiswere collectedandprocessedfortheperiodfrom September2018toSeptember2019.

3.1. Domestichotwater(DHW)loads

DuetotheabsenceofanenergymeterintheDHWsupplyline, theconsumptionloadandmassflowratearecalculatedusingthe energy balance equation on the DHW subsystem. The reference temperature, T_ref, represents the temperature of the cold supply water,which isfairly stablethroughouttheyear. Thewatertem- perature istherefore assumedto keep a constant temperatureof 8 ^◦C. The energystored in the tanks, E_tanks [kWh],at each time

step,i,iscalculatedbyEq.(1).

E_tanksi=

ρ

^VCp10

j=1

(

^Tji−T_{re f}

)

⁽¹⁾

whereT_jistemperaturemeasured intankj,whichholdsa water volume,V,of600l.Thetemperatureinthestoragetanksnormally variesbetweenT_ref and66^◦C,andaremeasuredinthemiddleof eachtank, whichgivesagoodoverviewofthetemperaturegradi- entacrossthe storage.spaceWaterdensity,

ρ

^{[kg m−3],andspe-}

cificheatcapacity,Cp[kWhkg⁻¹K⁻¹],atthemeanoperationtem- peratureof30^◦Careusedinthecalculations.Applyingtheenergy balanceontheDHWsubsystemyieldsthefollowingequation:

^Etanksi

dt_i =Q˙_HX2i+Q˙_HX3i+Q˙_HX4i−Q˙_DHWi (2) where

^Etanksi=E_tanksi+1−E_tanksi (3)

Q˙_DHW

i [kW] in Eq. (2) is the heat load accompanying DHW usage. Other parameters are explained in Sections 2.1 and 2.2. Change ofenergyinthewaterstorageateach timestep,^Etanks_i

[kWh] (Eq. (3)), is definedasthe difference betweenthe current time step, i,and thenext, i+1.The DHW heatload, (Eq.(4)), is derivedfromEqs.(1)to(3).Theheatlossesfromthestoragetanks are accountedforinQ˙_DHW

i.The averagevalue ofcalculated mea- surementuncertainty[%]ispresentedintheequation.

Q˙DHWi=Q˙HX2i+Q˙HX3i+Q˙HX4i−

ρ

^VCp

dti

10

j=1

(

^Tji+1−Tji

)

±4.5% (4)

TheDHWconsumptionmassflowrate,m˙_DHWi [kgs⁻¹],isthen calculatedas

m˙DHWi= Q˙DHWi

(

^Tset−T_{re f}

)

^{Cp (5)}

whereTsetistheDHWsupplysetpointtemperature(55^◦C).

3.2. Coefficientsofperformance(COPs)

Collected measurements for heating capacities, AC capacities andpowerconsumptionareusedtocalculatedtheCOPsofthein- tegrated thermalsystem. The total systemCOP [-], referred to as COPsys, isdefined asthe ratioof usefulthermalload to thetotal electricityconsumption,usingEq.(6):

COPsys= Q˙GC1+Q˙GC2+Q˙AC

W˙comp+W˙_{f ans}+W˙pumps+W˙_aux,el±6.2% (6) where W˙comp, W˙_{f ans} and W˙pumps [kW] represent the combined electricity consumption forall compressors, evaporationfans and pumps,respectively.W˙_aux,el[kW]is the electricityconsumption for auxiliarysystems,suchascontrolsystems.Q˙_AC istheACloadthat issuppliedthroughHX6.

TheheatpumpCOP,COP_h[-],istheratioofthetotalheatload to theelectricity necessaryto providethe heatingfunctions.This includes electricityconsumption of thecompressors andthefans intheevaporators,asshowninEq.(7).

COP_h= Q˙GC1+Q˙GC2

W˙comp+W˙_{f ans} ±5.7% (7)

TheCOPoftheACchillersystemisnotevaluated asasingular parametersincecoolingisnotacontrollingparameterintheR744 unit,butratherabyproductoftheheatingoperation.

The seasonal coefficient of performance (SCOP) forthe entire heat pump systemwith/without boiler, SCOP_sys+el [-] andSCOPsys

[-], and SCOP for heating, SCOP_h [-], are calculated as the ratio

betweensuppliedheatingand/or ACcoolingenergy[kWh]to the workofcompressorsandauxiliarydevices[kWh],asshowninEqs.

(8)and(9). SCOPsys=

(

^Q˙GC1+Q˙_GC2+Q˙_AC

)

(

^W˙comp+W˙f ans+W˙pumps+W˙aux,el

)

^±6.2%

(8) SCOPsys+el

=

(

^Q˙GC1+Q˙_GC2+Q˙_AC+Q˙_EL

)

(

^W˙comp+W˙f ans+W˙pumps+W˙aux,el+W˙EL

)

^±10.5%

(9)

SCOPh=

(

^Q˙GC1+Q˙_GC2

)

(

^W˙comp+W˙f ans

)

^{±5.7% (10)}

whereQ˙_ELandW˙_EListheheatandpowerassociatedwiththeop- erationoftheelectricboiler,respectively.

4. Systemperformanceanalysis 4.1. Analysisofkeyoperatingparameters

In order to assess the system performance during different operational conditions with variations in heating, DHW and AC loads,key operating parameters of thesystem are evaluated and discussed in this section. Specific periods are categorized based on weather conditions that demonstrate different seasonal per- formance of the system: summer (June through August), winter (Novemberthrough March),andnominalforoperatingconditions representing fall and spring (September through October, April throughMay).

4.1.1. Keysystemoperationalparameters

KeyR744cycleparametersandhigh-sidetemperaturesareana- lyzed.ThestudiedparametersincludeambientairtemperatureTa

[^◦C], R744 evaporationtemperatureTevap [^◦C], aswell as MTand HT supply water temperatures, represented by T_w,MT and T_w,HT [^◦C], respectively.Thehighpressure,P_gc [bar],andgascoolerout- lettemperature, T_gc,exit [^◦C]areincluded inthe analysis.Theper- formance ofthe systemunder fourweeks ofwinter operation is showninFig.2.Theperiodischaracterized bylowambienttem- peratureandmarginalACloads.

As can be observed inFig. 2, Pgc operates in the transcritical pressureregionwithamaximumworkingpressureof100bar.Rel- ativelylargefluctuationsinpressureoccurduringthisperiod,asa resultofvariationsinT_gc,exit.Lowfluidreturntemperaturefromthe secondarythermalsystemswillconsequentlylimitT_gc,exitandthus also Pgc. However, some situations will cause unwanted high re- turntemperaturesfromtheMTcircuittothesecondgascooler:(a) transitionsbetweendifferentmodesofoperation,(b)lowloadop- erationswithmanystartsandstops,and(c)mixinginDHWtanks, whichresultinhighreturntemperatureduringcharging.Theneg- ativeimpactofhighreturntemperaturecanbereducedbyincreas- ingthegascoolerpressure.

Heatis supplied to the secondary thermalsystem atthe two differenttemperaturelevels T_w,MT andT_w,HT.TheMTandHTcir- cuitsetpointtemperaturesareregulatedbasedonoutdoortemper- aturecompensationcurves,whichvariesfrom25to50^◦C and60 to 70^◦C, respectively. Fig. 2 showsthat T_w,HT generally operates between65and70^◦C.

Themassflows throughthefourairevaporatorsarecontrolled accordingtothe superheatatthe exitofeach evaporator.Hence,

Fig. 2. Key operating parameters for winter operations (November 6th to December 4th 2018).

T_evap generallyfollows the pattern of T_a with a temperature dif- ferencedeterminedby thesuperheatcontrol.Thesetpointforsu- perheatis periodicallychanged accordingto temperaturelevel of Ta. For the interval displayedin Fig. 2, the superheat setpoint is fixedtoaminimumof4K.ThesuddendropofTevap isillustrated halfwaythroughweek2.Thisbehavioroccurswhentheheatload isincreased, e.g. duringheat pumpstart-up, capacity increase or duringactivationofadditionalevaporators.Hence,thesetpoint of theevaporation, Tevap, isreduced to boost the dischargetemper- atureandtoincrease thecapacity.The reductionof Tevap isextra workfor thecompressors andcauseexcessive superheat that re- ducesCOPconsiderably.

4.1.2. Domestichotwater(DHW)accumulation

The consumption of hot watertypically follows a certain pat- terndependentonthebehavioroftheresidentsandtheoperation ofthehotelfacilities.Themajorconsumersofhotwaterinhotels areprimarilyguests,kitchens,laundryservicesandspaorpoolfa- cilities(Bohdanowicz,2006;Lawson,2001).Generally,thehotwa- ter consumption inhotel buildingsis characterized by large con- sumptionpeaksforafewhoursduringthemorningsandevenings (NdoyeandSarr,2008;RankinandRousseau,2006;DengandBur- nett,2002).Incircumstances whereno DHWstoragebuffer isin- stalled, the high consumption peaks will be directly reflected in thehotel’spowerconsumption.Fig.3showsthehotwateraverage dailyconsumption profile,DHW_usage [kWh],andtheprofileofen- ergysuppliedbytheheat pumptothestorage,DHW_supply [kWh], overaperiodofoneyear.TheaverageDHWdailyusageduringthis period is 1104 kWh/day. However, significant variations in daily consumptionwererecordedwithmaximumandminimumvalues of2480and480kWh/day.Onaverage,2.3%oftheDHWusageiscov- eredbytheelectricboiler.

As seen in Fig. 3, mostof the DHW consumption occurs be- tweenhour8andmidnight.TheDHWusageduringthistimepe- riodaccountsfor87% ofthedailyconsumption.The activitylevel inthehotelislowbetweenhours0and6,hencetheDHWusage islowerduringthistime.DHWusagepeaksoccurduringthehours 9and23 atvalues around70kWh.However, DHW_supplydoesnot exceed 58 kWh due to the buffer effect granted by the storage, anddemonstrateshowthesystemhandlespowerpeaksonanav- eragebasis. The impact of the storage is the difference between DHWusageandDHW_supply,whichreachesapeak of22kWhduring hour8.Thechargingofthestoragebeginsathour0anddeclines toaminimumaroundhour6,asthestorageisfullycharged.When

Fig. 3. Hourly-average DHW consumption and supply profiles over a one-year pe- riod.

DHWusageincreasestoa peakof73 kWhathour9,thestorageis emptyandactivechargingbegins,holdingavaluebetween50and 55kWhthroughtheday.

The control strategy of the stratified heat storage in an R744 system is essential for successful operation, as described by Tammaroetal.(2016).Detailedoperatingparameters oftheDHW storage areshown witha 20-minute resolution over a2-day pe- riod inFig. 4.Fig. 4Ashowsthe temperaturestratification across the storage, which is illustrated by the temperatures in tanks 1, 3,5,7and10,aslabeled inFig.1.Thestorage loadandthe cor- responding energy in the storage over the period are shown in Fig. 4B and C, respectively. The water temperature of the stor- age fluctuatesbetween 8and 78^◦C during theperiod. The state of the storage can be determined by studying the temperatures Tank1andTank10.As thelast tank inthe series,Tank10issensi- tive to change inDHW mass flow rates entering andexiting the DHWsubsystem.Supplywaterat8^◦C enterstank 10duringdis- chargeandisgraduallypushedthroughthestorageashotwateris drawnfromTank1 andsuppliedto thehotel.The suddendropin all temperaturesinFig.4Aillustratesthedischargeofthestorage

Fig. 4. Operation of the DHW subsystem over a 2-day period showing (a) storage temperature, (b) DHW usage and supply and (c) energy in the storage.

and corresponds to peaks in the DHWusage, ascan be observed in Fig. 4B. The energy potential of the storage is fully exploited when Tank1 reachesits minimum watertemperature. The charg- ingofthestorageisillustratedbytheincreaseinthetemperatures across thebuffer. Hotwateris suppliedtothe storageviatank 1 andiscirculatedthroughthebuffer.Thetemperatureboundarybe- tween hot andcold movesthroughthe storage,astank tempera- turesarelifted.Asaconsequence,thetemperaturesinthemiddle ofthestorage(tanks3–7)canoccasionallybehigherthanthetem- perature oftank1 ifthe hotwatersupplytemperaturefluctuates duringthechargingprocess.Thisbehavior isillustratedby Tank5, whichsometimesishigherthanTank1.Simultaneously,coldwater isdrawnfromtank10forheating,asexplainedinSection2.2.The storage ischargedwhen Tank10reachesits peak temperature. As seen fromFigs. 4A,thereis a 24-hourpatterntothe behavior of the DHW storagetemperatures.The DHW storageenergy isfully exerted andisrechargedtwiceaday,whichisinagreementwith the findings inFig. 3. Fig. 4B showsthat the storageis typically charged for7–10h. The sudden dropinstorage temperaturecan be seen in referenceto the behavior of DHWusage. As shownin Fig.4AandB,largeDHWusagepeaksintherangeof200kWcause a rapid decrease inthe storage temperatures.It can be observed fromFig.4C that ittakesapproximately2hours todischargethe entire storageduring theseperiods. There is still a highdemand for DHWat hour9 each daywhen the storagereaches its mini- mum energypotential.Thehot watergeneratedby theR744unit isthen supplieddirectlytothehotel tocompensate forlarge de- mands.Thissystembehaviorindicates thatthestoragevolumeof 6m³isnotquitesufficienttomeetthepeakDHWdemandsofthe hotel. This is especially evident in the mornings, asTank1 drops belowits setpoint of55^◦C.At fullychargedconditions,the stor- agereachesanenergypotentialofapproximately350kWh.Apos- siblesolution forthe insufficientenergyreserve inthe storageis to store the waterathigher temperatures.Byincreasing the wa- ter temperaturein all tanksto 70^◦C, one couldincrease the en- ergy storage capacity with about 25%. Nevertheless, the storage buffer still provides a beneficial reduction of peak loads. This is representedbythedifferencebetweenDHWusageandDHWsupply, whichismorethan100kWduringpeakhours.Anotherbenefitof the largestorage volumeis higherflexibility inDHW production,

whichallows forlow-intensity DHWgeneration over longertime intervals.

4.2.Evaluationofenergyperformance

The energy efficiency of the system including the provided heating, AC loads and COPs are evaluated in the following sub- sections.

4.2.1. HeatingandACcoolingloads

Seasonalhourly-averagedheatingandACloads,Q˙,oftheinte- gratedR744system, together withhourly-averaged ambienttem- peratures,Ta,avg,andrecordedmaximumandminimumtempera- tures,T_a,max andT_a,min, are showninFig. 5A–C.The specific load forDHW,SHandACareindicatedbysubscripts.Heatingloadsare shown as positive values and the AC loads are shown as nega- tive values. Error bars for the loads indicate the range of values recordedforthatparticularhour.Thehourly-averagedloadsarein- vestigatedover24-hoursduringsummer,winterandnominalpe- riodsofthe year,which definitionsare explained inSections4.1. Fig.5DshowstheannualheatingandACcooling energysupplied by the R744 system, E,and ambient temperatures on a monthly basis.Thehourly-averagedloadsandthemonthlytotalenergycon- sumptionareusedto evaluatetheperformance ofthesystemfor the full range of operation from September 2018 to September 2019.Thetrends forthe differentloads arediscussedindividually inthefollowingparagraphs.ItshouldalsobestatedthattheY-axis temperaturescalefortheseasonalcasesaredifferent.

– ˙Q_SH:TheSHloadisdependentonT_a,avgandvariesinarange of5–100 kWfor thedifferent seasonalscenarios displayed inFig.5A–C.Thelowestrecordedvaluesareobservedinthe summercase,whereQ˙_SHdecreasessignificantlywhenT_a,avg exceeds15^◦C.Inthiscase,theentiretyofQ˙_SH issuppliedto thebatteries inthe ventilationunits.Naturally, thehighest recordedhourly-averagedvaluesareobservedinthewinter scenarioin Fig.5B. Approximately 80% ofQ˙_SH isthen sup- pliedtotheventilationunits,duetotherelativelylargeca- pacityoftheseunits.Thus,onlyasmallportionoftheheat loadisusedtocoverdirectSH,e.g.forfloorheatingandra- diators.

Fig. 5. Hourly-averaged heating and AC loads for (A) summer (May 18th–25th 2019), (B) winter (January 15th to 22nd 2019) and (C) nominal (October 18th–25th 2018).

Total annual energy supplied by the R744 unit on (D) monthly basis (Sep. 2018 to Sep. 2019).

– Q˙_DHW: The DHW loads in Fig. 5A to 5 C drop to the minimum value of 20 to 30 kW at hour 6, followed by a rapid increase in Q˙_DHW between 60 to 80 kW, which stay presentthroughoutthe day.Anoticeable difference in the magnitude of Q˙_DHW is shown in the various seasonal scenarios. These inconsistencies are due to variations in guestloadandareindependentofseasonaloperationalload andTa,avg.

– ˙Q_AC: The ACrefrigeration capacity variesin a limited range of 0 to 24 kW in all seasonal cases.The load is indepen- dent of the hour-of-day and Ta,avg. However, the AC load providedby theR744isnotindependentofTa,avg.Thisun- usual behavior in supplied AC load from the R744 unit is causedbythefact thatitisanauxiliary systemtothepre- installed separate cooling unit. It should be notedthat AC cooling provided by the primary stand-alone chiller is not included in Fig. 5A–D. The separate AC chiller unit is op- erating atfull loadduringthe summerscenario inFig. 5A, though hardly anyAC issupplied by theR744 unit during thistimeduetothelow-sidepressurecontrol.Asshownin Fig.1,theairevaporatorsandHX6intheR744unitoperate atthesamepressurelevel,controlledsolelybytheairevap- orators.TheR744unitthereforeonlysuppliesextraACwhen

theevaporationtemperatureisbelowthe 7^◦C setpointfor ACchilledwater.The largestQ˙_AC capacitiesare observedin thewinterandnominalscenarios inFig. 5BandC, respec- tively.Inthesescenarios,moderateTa,avgenablesoperation ofthechilled waterHXwithin the acceptableevaporation- temperaturerange.

– E_SH: E_SH varies in a range of approximately 10,000 to 55,000 kWh, in close connection to Ta,avg. As shown in Fig.5D,theheatingdemandisstillpresentduringthesum- mer months, due to the relatively cold climate at the ho- tel’slocation.ThelargestrecordedvaluesofE_SHisobserved duringthewintermonthswhenT_a,avgisbelow5^◦C.E_SHis theninarangeof42,000–55,000kWhmonthly,whichisup to5timesthe SHusage forthesummermonths.The total amountofE_SHovertheyearis380,000kWh.

–E_DHW:ThemonthlyenergyforDHWshowninFig.5Disstable throughouttheyearinarangeof30,000to40,000kWhper month, withan average value of 33,600 kWh. The annual energy supplied for DHW over the year is 403,000 kWh.

Thus, 52% of the annualheatingenergy to the hotel isal- locatedtoDHWheating. Therelative consumptionofDHW to totalheating inhotels is typically between40 and70%,

Table 1

COPs for selected intervals in the period from Sep. 2018 to Sep. 2019.

Season Period SCOP sys[-] SCOP sys+el[-] SCOP h[-] T a,avg[ ^◦C]

Winter November–April 2.78 ± 0.17 2.57 ± 0.27 2.69 ± 0.15 0.4 January 15th to 22nd 2.63 ± 0.16 2.37 ± 0.25 2.49 ± 0.14 −5.3 Summer June–September 3.20 ± 0.20 2.75 ± 0.29 3.09 ± 0.18 15.0 May 18th–25th 3.34 ± 0.21 2.97 ± 0.31 3.30 ± 0.19 15.8 Nominal September–November, April–June 2.99 ± 0.19 2.73 ± 0.29 2.90 ± 0.17 8.4

October 18th–25th 3.23 ± 0.20 3.21 ± 0.34 3.05 ± 0.17 6.3 Annual September–September 2.90 ± 0.18 2.64 ± 0.28 2.80 ± 0.16 6.8 Seasonal intervals are from the 1 st to the 1 st in the stated months.

andisdependentonthelocation,buildingenvelopeanduse offacilities(Su,2012;DengandBurnett,2000).

–E_AC:Aspreviouslyexplained,E_ACislargerduringthenominal monthsofoperation.ACisprimarilyusedforclimatecontrol and temperature adjustments in common areas and guest rooms.TheACcoolingcapacityisthereforelargerduringpe- riodswithhighguestloadsandlargeheatingdemands. For theentireyear,only75,500kWh ofACcooling energywas recoveredthroughthechilledwaterHX.

4.2.2. Coefficientofperformance(COPs)

The SCOPs for the scenarios depictedin Fig. 5, together with seasonalandannualvaluesare listedinTable1.Theaverageam- bient temperature, Ta,avg, for the specified intervalsare included inthe table.Predictably, SCOPsys is higherthanSCOP_h forthe in- vestigated scenarios. However, the annual SCOP_sys is only 0.1 or 3.6%higherthantheSCOP_h.Byrneetal.(2009)estimatednumeri- callyaSCOPof3.57foraheatpumpandchillersystemforhotels using R407a.They alsoinvestigated an R744system withsimilar operational conditions and found a SCOP of 3.24. However, sec- ondarysystemsandrealoperatingconditionswerenotaccounted for inthisstudy. The conventional thermalsystems found inthe Nordichotel marketnormally utilizeelectricboilers/districtheat- ing stationsincombinationwithseparate HFC-unitsforAC.Typi- cally,a SCOPsys inthevicinity of1isachieved forthesesystems, dueto the relativelylarge magnitudeof heatingloadtoAC load.

Thesomewhatlow valueofannualSCOPsys fortheintegratedsys- temispartially dueto thelimitedrecovery ofcoldenergytothe ACcoolingcircuit.Thisisalsothecaseforthelong-termseasonal periods,e.g.winter,summerandnominal.Onanannualbasis,ap- proximately 5% of the total heat to the hotel is supplied by the electricboiler.Asaresult,SCOP_sys+el isreducedby9%whencom- pared to SCOPsys. It isexpected that the boiler is applied during the winterseasonto coverpeak heating. However,the low value ofSCOP_sys+el duringthesummerseasonindicatesexcessiveuseof the boilerforDHW heating. Thisis explainedby thehigh return temperatureofwatertoGC2. Atemperatureabove45^◦Ctriggers a signal toreduce the compressorcapacity, due tocompromised efficiency.Consequently,DHWproductionbytheheatpumpisre- ducedandtherequiredloadisthencompensatedbytheboiler.

AllSCOPsarehighlydependentonTa,avgandincreasewithap- proximately 0.4 fromthe winterto the summer season.A larger difference betweenthe specific SCOPsis observed whencompar- ingthesummerandwinterweekscenarios(Fig.5AandB),which canbeattributedtothechangeinT_a,avg.ThenominalweekofOc- tober18th–25threvealsuncharacteristicallyhighvaluesofSCOPsys

whenrelatedtothenominalseason.Moreover,Ta,avgforthisweek is2.1^◦Cbelowtheaveragetemperaturefortheparticularseason.

The high value of SCOPsys duringthis week can be explained by the relatively large utilization ofAC, asdisplayed inFig. 5C. The gain fromQ˙_AC isthereforelargerthanthecontributionfromW˙_{f an} andW˙_aux,el inthecalculationofSCOPsys,asdefinedinEq.(8).

The mean COPsys and COP_h for transcritical operations ( > 73.9 bar), according to specific temperature intervals, are

Fig. 6. Difference between the DHW charging and no charging COPs.

listedinTable2.TheCOPsarecategorizedbywhetherornotDHW chargingistakingplace.Thesubscriptnchincludescircumstances whenthe heat supply tothe hotel iscontrolled by SH demands, and no active charging of the DHW storage is takingplace. Sit- uationswhentheDHWstorageisbeingactively charged,andthe systemiscontrolledaccordingtobothSHandDHWloadsareiden- tifiedbythesubscriptch.Theanalysisofvariance(ANOVA:single factor)wasapplied toanalyze theefficiencyofthe systemunder differentmodesofoperations.Thedifferenceisconsideredsignifi- cantatp < 0.05.

Table2showsasignificantdifferencebetweenchargingandno chargingvaluesofCOP_h,avgattemperaturesbelow0^◦Candabove 15^◦C.COPsys,avgexhibitsignificantdifferencebetweenallintervals, withtheexceptionof−10to−5^◦Cand10to15^◦C.Fig.6depicts therelativechangeinCOP,^{COP[%],fromnocharging(COP}nch)to charging(COP_ch).TheCOPsduringnoDHWchargingaregenerally higherthanchargingmodeatlowtemperatures,whichresultsina decreaseof^{COP.However,both COP}sys,avg andCOP_h,avgincrease considerablyatT_aabove15^◦C.Thisunusualrelationshipbetween the two modes ofoperation can be explained by the magnitude oftheSHloadandthetemperatureofthewaterreturningtothe secondgascooler.Thetemperatureofthefluidreturningfromthe secondarysystemisgenerallyhigherathighvaluesof SH,asthe setpoints ofSH andthusthe returntemperatures areelevated at lowvaluesofT_a.Additionally,DHWchargingprovides atempera- tureliftin the returncircuit whenthe stratification inthe DHW storageis notfully intact, asdiscussed inSections4.1.2.This be- havior wasalso illustrated by Tosato et al. (2019). They noted a reductioninCOPof18%duringthefinalpartoftheDHWcharging process,which wascaused byhighreturntemperatures fromthe

Table 2

System and heating COPs during DHW charging and no charging at different temperature intervals.

T a[ ^◦C] [ −15, −10) [ −10, −5) [ −5,0) [0,5) [5,10) [10,15) [15,20) [20,25) [25,30)

COP sys ch,avg[–] 2.59 (0.50) 2.40 ^a(0.22) 2.53 (0.34) 2.82 (0.44) 3.17 (0.50) 3.38 ^a(0.58) 3.37 (0.61) 3.38 (0.62) 3.28 (0.50)

COP sys nch,avg[–] 2.40 (0.70) 2.41 ^a(0.33) 2.61 (0.56) 2.89 (0.66) 3.32 (0.68) 3.43 ^a(0.92) 3.07 (0.92) 2.80 (0.83) 2.93 (0.85)

COP h ch,avg[–] 2.47 (0.46) 2.32 (0.23) 2.53 (0.34) 2.77 ^a(0.45) 3.03 ^a(0.47) 3.11 ^a(0.50) 3.19 (0.54) 3.23 (0.57) 3.24 ^a(0.53)

COP h nch,avg[–] 2.27 (0.70) 2.38 (0.31) 2.58 (0.49) 2.76 ^a(0.57) 3.04 ^a(0.53) 3.08 ^a(0.80) 2.99 (0.87) 2.93 (0.87) 3.00 ^a(0.85)

T a ch,avg[ ^◦C ] −12.0 (1.3) −7.2 (1.4) −2.0 (1.4) 2.4 (1.5) 7.4 (1.4) 12.3 (1.4) 17.3 (1.5) 21.9 (1.3) 26.9 (1.4)

T a nch,avg[ ^◦C] −12.1 (1.3) −7.5 (1.4) −1.8 (1.4) 2.3 (1.4) 7.1 (1.5) 12.0 (1.4) 17.0 (1.5) 21.6 (1.3) 26.8 (1.4)

a No significant statistical difference ( p > 0.05) between corresponding DHW charging and no charging values. Standard deviation is shown in the brackets.

Values for calculated measurement uncertainties are not included.

storage.Hence,duringseasonswithlowTa,highSHandthusgen- erallyhighreturn fluid temperature, theheating loadofthe CO₂ unitis limitedduetohighT_gc,exit.Thisproblemdiminisheswhen theSH loadis limited, as can be observed in Fig. 6 atT_a above 15^◦C.The COPs duringDHW chargingare generally higherthan thenocharging modeathighambient temperatures,whichis in agreementwiththefindingsinTosatoetal.(2019).Thus,theDHW chargingstrategyofthesystemshouldberegardedasakeyinflu- encingfactortoachievehighefficiency.

5. Conclusions

This work investigated key operating parameters for an R744 heatingand AC cooling unit installed in a Norwegian hotel. The systemis integratedwithHVAC, DHW anda 6 m³ thermalstor- age.Fieldmeasurements fromthehotelwere analyzedforaone- yearperiodandessentialparameterstoevaluatethesystemperfor- mancewere discussed, includingheatingand ACloads, tempera- tures,pressuresandmassflowrates.DHWconsumptionloadsand COPswerecalculatedusingthecollecteddata.TheDHWconsump- tionwasestimatedbytheenergybalanceduetothepeculiarities oftheinstrumentationinstalledbythesupplier.Consequently,the heatlossfromthe storagetankswere includedinthe DHWcon- sumptionrateandthusnot evaluatedinthisstudy.Thesameap- pliestothe existing ACcoolingmachine, asthisunit isnot inte- gratedinthemeasurementandcontrolsystem.

TheheatingandACloadssuppliedbytheR744unitwerestud- iedonaweeklyandmonthlybasistoassesstheseasonalbehavior ofthesystem.TheresultsrevealthattheDHWloadisfairlystable throughouttheyearandisindependentofseasonalambienttem- peratures.TheDHWloadaccountsfor52%oftheannualheatload suppliedto thehotel andfollows a particular24-h pattern, with low consumption betweenmidnight and hour 6.The peak DHW loadoccursaroundhour9andreachesanhourly-averagedvalueof 73kWh.TheDHWstorageholdsanenergycapacityof350kWhat fullychargedconditionsanddemonstratespeakdemandcompen- sationofmorethan 100 kWduringOctober18th–20th 2018. Pe- riodicaldecreaseinstoragetemperaturestovaluesbelowtheset- pointindicatesthatthestorageisnotfullyequippedtohandlethe peakDHWloadsofthehotel.Thiscanbesolvedbyinstallingmore tanksinseriesorbyincreasingthewaterstoragetemperature.

The COPsduring DHWcharging mode are higher when com- paredwithnochargingatambienttemperaturesabove15^◦C,due to limitedSH demands. The SH is highlydependent on ambient temperaturesand variesnoticeably in the differentseasonal sce- narios.ThemonthlysupplyofSHenergyincreasessignificantlyat averageambienttemperaturesbelow5^◦C. TheACcapacitydeliv- eredby the R744 unit is limited and not fully exploited, which isreflected in the moderate annualSCOP_sys of 2.90. Additionally, about5% ofthetotalheat tothehotelis suppliedby theelectric boiler,whichdecreasesoverallSCOPsysby 9%.Thelatterisoftena resultofhighreturntemperaturesfromthebuilding,whichisag- gravated by increasednumber of R744unit starts andstops and mixinginDHWtanks.Other factorsthat greatlyinfluencetheef-

ficiencyofthesystemarevariationsintheambienttemperatures andhightemperaturesatthegascoolerexit.

Observations from this work can be used as a good starting point for modeling and optimization of the existing and similar systems.Future work shouldfocus on increasing thesystemper- formancebychargingthestorageduringlongerperiodsatreduced capacities. The optimalstorage volumefor thistype of systemis animportantissuethatshouldbeprioritized.

DeclarationofCompetingInterest

Theauthorsdeclarethattheyhavenoknowncompetingfinan- cialinterestsorpersonalrelationshipsthatcouldhaveappearedto influencetheworkreportedinthispaper.

Acknowledgments

TheauthorswouldliketoacknowledgetheNorwegianResearch Councilforfundingthisproject.WewouldalsoliketothankKelvin AS for in-depth systemdetails and Scandic HotelHell foraccess totheir systemdata.Also, wewouldliketoacknowledgeYannick Prussforhiscontributiontothisresearchwork.

Supplementarymaterial

Supplementary material associated with this article can be found,intheonlineversion,atdoi:10.1016/j.ijrefrig.2020.03.021. References

Bamigbetan, O. , Eikevik, T.M. , Nekså, P. , Bantle, M. , 2017. Review of vapour compres- sion heat pumps for high temperature heating using natural working fluids. Int.

J. Refrig. 80, 197–211 .

Bohdanowicz, P. , 2006. Responsible Resource Management in Hotels: Attitudes, In- dicators, Tools and Strategies. KTH .

Bolaji, B. , Huan, Z. , 2013. Ozone depletion and global warming: case for the use of natural refrigerant–a review. Renew. Sustain. Energy Rev. 18, 49–54 .

Byrne, P. , Miriel, J. , Lenat, Y. , 2009. Design and simulation of a heat pump for simul- taneous heating and cooling using HFC or CO 2as a working fluid. Int. J. Refrig.

32 (7), 1711–1723 .

Cecchinato, L. , Corradi, M. , Cosi, G. , Minetto, S. , Rampazzo, M. , 2012. A real-time algorithm for the determination of R744 systems optimal high pressure. Int. J.

Refrig. 35 (4), 817–826 .

Cecchinato, L. , Corradi, M. , Fornasieri, E. , Zamboni, L. , 2005. Carbon dioxide as re- frigerant for tap water heat pumps: a comparison with the traditional solution.

Int. J. Refrig. 28 (8), 1250–1258 .

Cecchinato, L. , Corradi, M. , Minetto, S. , 2010. A critical approach to the determi- nation of optimal heat rejection pressure in transcritical systems. Appl. Therm.

Eng. 30 (13), 1812–1823 .

Chua, K.J. , Chou, S.K. , Yang, W. , 2010. Advances in heat pump systems: a review.

Appl. Energy 87 (12), 3611–3624 .

Ciconkov, R. , 2018. Refrigerants: there is still no vision for sustainable solutions. Int.

J. Refrig. 86, 4 41–4 48 .

Council of the European Commission and others (EC) , 2010. Directive 2010/31/EU of the European Parliament and of the council of 19 May 2010 on the energy performance of buildings. Off. J. Eur. Union L 153, 13–35 .

D’Agaro, P. , Coppola, M. , Cortella, G. , 2019. Field tests, model validation and perfor- mance of a CO 2commercial refrigeration plant integrated with HVAC system.

Int. J. Refrig. 100, 380–391 .

Dalton, G. , Lockington, D. , Baldock, T. , 2008. Feasibility analysis of stand-alone renewable energy supply options for a large hotel. Renew. Energy 33 (7), 1475–1490 .