Continuous heating of an air-source heat pump during defrosting and improvement of energy efficiency

Continuous heating of an air-source heat pump during defrosting and improvement of energyefficiencyJi Young Jang a ,Heung Hee Bae a ,Seung Jun Lee a ,Man Yeong Ha b ,⇑a SAC R&D Laboratory,LG Electronics,Seongsan-Dong,Changwon City,Gyeongnam,Republic of KoreabSchool of Mechanical Engineering,Pusan National University,Jang Jeon 2-Dong,Geum Jeong Gu,Busan 609-735,Republic of Koreah i g h l i g h t sThe newly designed dual hot gas spray defrosting method was examined. Uninterrupted heating of an air source heat pump during defrosting.We compared newly designed dual hot gas and traditional reverse cycle defrost. Total energy efficiency was increased by 8%compared to traditional method.a r t i c l e i n f o Article history:Received 24September 2012Received in revised form 9April 2013Accepted 10April 2013Available online 6May 2013Keywords:Heat pump DefrostHot gas bypass defrosting Frosta b s t r a c tDuring winter operation,an air-source heat pump extracts heat from the cold outside air and releases the heat into the living space.At certain outside air conditions,when it operates in heating mode,frost can form on the air-cooled heat exchanger,decreasing the heating performance.Conventionally,reverse-cycle defrosting (RCD)has been the common method of frost removal.But this method requires the inter-ruption of heating during defrosting,as well as a period of time to reheat the cooled pipes of the indoor units after defrosting.In this study,a new technology called continuous heating was developed,which utilize only a hot gas bypass valve to remove the frost from the outdoor heat exchanger and thus enabling the supply of hot air to indoors without any interruption.For this,a new high temperature and low pres-sure gas bypass method was designed,which is differentiated from the common high pressure hot gas bypass methods by its use of low pressure.Various refrigerant mass flow distributions were examined,and the most effective defrosting mass flow was 50%in this case.Heating capacity was increased by 17%because of continuous heating,and the cumulated energy efficiency was increased by 8%compared to the traditional reverse cycle defrosting over 4h including two defrost operations.Also,cumulated energy efficiency was increased by 27%compared to electronic heaters that supply the same heating capacity during defrosting.By this new technology,it has been proved that continuous heating and energy savings could be achieved without adopting expensive technologies.Ó2013Elsevier Ltd.All rights reserved.1.IntroductionIn recent years,the variable refrigerant flow heat pump using an air-source has been widely used for space cooling and heating in residential and commercial buildings,because these systems have precise capacity control and individualized thermal comfort capabilities with very high energy efficiency and low initial instal-lation cost compared to any other air conditioning method [1].A variable refrigerant flow heat pump is a refrigerant system that varies the refrigerant flow rate with the help of the variable speed compressor and electronic valves to match the capacity of the system to the space cooling or heating loads in order to maintain the zone air temperature at the set temperature [2].During winters,under certain outside weather conditions,the air-source heat pump often operates with substantial frost forma-tion on the outdoor heat exchanger,which is used as an evapora-tor,and this frost layer has to be melted away periodically.This unavoidable defrosting requirement causes an interruption of in-door heating,lowering energy efficiency and heating capacity [3,4].Many studies have been undertaken in an attempt to over-come this weakness of the air-to-air heat pump,and to delay frost formation.Mei et al.[5]suggested that by adding a moderate amount of heat to the refrigerant steam in the accumulator,the0306-2619/$-see front matter Ó2013Elsevier Ltd.All rights reserved./10.1016/j.apenergy.2013.04.030Corresponding author.Tel.:+82515102440.E-mail address:myha@pusan.ac.kr (M.Y.Ha).evaporator coil temperature can be raised by several degrees.Byun et al.[6]tried to retard the formation and propagation of frost using alternative operation of multi-hot gas bypass line into evaporator inlet.Silva et al.[7]studied the fan characteristics as a frost layer grows with various types of tube–fin type heat exchangers.Mader and Thybo[8]proposed an active distribution valve that is able to feed parallel evaporators passed individually, and that would regularly shut single evaporator circuits off in order to remain frost-free.Kwak and Bei[9]applied an electric heater in front of an outdoor heat exchanger instead of an indoor unit as usual to enhance the heating capacity and delay frost formation. Although many studies have been performed on avoiding frost on evaporator heat exchangers and delaying the defrosting control, the need for defrosting is still one of the serious shortcomings of air to air heat pumps.Defrosting can be carried out in a number of ways:compressor shut down,electric heating,reverse-cycle,hot-gas bypass.Of these, reverse-cycle defrosting(RCD)is the most common method of frost removal,and various papers have studied the dynamic char-acteristics during the defrosting process on the air to air heat pump [10–14].Another widely-used method is hot gas bypass defrosting. Hot gas from compressor discharge can be a good resource to melt frost easily.The advantages of hot gas defrosting are lower refrig-erant noise,smaller indoor temperaturefluctuation,and no cold blowing.But because of its low quantity of heat,defrosting time should be much longer than the RCD time[15].Stoecker[16]inves-tigated the correct pipe size for carrying hot gas to defrost evapo-rators.The correct size will ensure that the resulting increase in specific volume of the hot gas that occurs at low pressure does not cause prohibitive pressure drops in the pipes.Hoffenbecker et al.[17]developed an application of a transient model for pre-dicting the heat and mass transfer effects associated with an indus-trial air-cooling evaporator during a hot gas defrost cycle.Recently, Choi et al.[18]proposed a novel dual hot gas bypass method to re-duce the defrosting time.Despite the large number of studies on the performance of air source heat pumps during defrosting operations,defrosting neces-sarily causes the periodic interruption of indoor heating and degra-dation in winter heating efficiency.As a result,the heating ratio is usually80–85%in the actualfield.Therefore,the objective of this study is to derive a way to improve the heating seasonal perfor-mance and supply uninterrupted heating of an air source heat pump that requires defrosting.diagram of a non-stop heating cycle.An electronic heater that is designed by considering evaporating capacity is generally installed in domain r,and the heat it supplies becomes deficient due to frost.To remove frost from the heat exchanger,a hot gas bypass line that is extracted from the compressor discharge to heat ex-changer is installed in domain s.Both can be installed together or separately depending upon the situation and the design purposes.The application of an electronic heater is a very powerful,sim-ple and direct method for achieving non-stop heating and when a heater situated outdoors performing the role of an evaporator can be more effective than directly locating indoors performing the role of a condenser.But this approach requires a relatively large size and additional expenses if the system capacity is increased. Therefore,almost products supplying continuous heating with a heater are smaller home appliances of5kW or lower.Also,there are some safety issues to consider,such as the potential for short circuit,electric shock andfire.Meanwhile,a hot gas bypass is sim-ple and relatively low cost compared to the heater application,but it reduces the indoor side heating capacity as well as the refriger-antflux for defrosting.Furthermore,it requires more time to re-move frost compared to reverse cycle defrosting.This study proposes a dual spray method using the existing hot-gas bypass method for a continuous cycle design that can also be applied to larger air source heat pump of30kW or higher,as well as smaller ones.3.Applied hot gas defrost using dual spray3.1.Schematics of dual spray hot gas defrosting(DSHG)The most effective and economical non-stop heating cycle with dual hot gas spray method was designed as shown in Fig.2.This method is the same as the general hot gas bypass method that ex-tracts hot gas from the compressor discharge,but the heat exchan-ger was divided into two parts,each having its own extra valve. Also,the existing single electronic expansion valve was increased to two.In normal heating operations,two bypass valves are closed and two expansion valves are working equally.When defrosting starts,the upper expansion valve is closed and the upper hot gas valve is opened to thaw the upper side frost and prevent the inflow of the refrigerant that must be evaporated.After the upper side defrosting isfinished,the lower side defrosting is performed with counter valve operations.When defrosting of one side is in pro-gress,the heat exchanger on the other side acts as a de-magnified evaporator to provide uninterrupted heating.The greatest differ-ence between this method and the existing RCD is that all sides carry out continuous heating,without stopping.Therefore,through RPM control the outdoor fan raises the low pressure of the system on the side of evaporator operating at half-power.Meanwhile the control logic induces the compressor to operate in its maximum allowable frequency and higher pressure to pro-vide maximum heat supply.The magnitude of the indoor heating supply capacity in the defrosting range ultimately depends on the frequency marginal rate of the compressor.The frequency mar-ginal rate is the difference between the present operation fre-quency and the maximum frequency.The high frequency marginal rate attains when the outdoor temperature is high and least indoor heating load.For this reason,it is advantageous for maintaining indoor heating capacity in the defrosting range.On the other hand,the proposed technique has the continuous heating cycle without the frequency marginal rate,or it can restrict the continuous heating operation in the event that the indoor heating can be hardly supplied in the defrosting range,such as at a very low outdoor temperature.Fig.1.General cycle diagram of a non-stop heating model.10J.Y.Jang et al./Applied Energy110(2013)9–16DSHGrefrigerant fluxes are the sum of the indoor and flow._m Defrost ;bypass ð1Þtotal refrigerant flux can be calculated by system parametersD h ð2Þdifference of inlet from outlet enthalpy of system.Whencalculated,indoor mass flow can be determined point.For example,50%of total mass flow means half the heating capacity during the defrost.After is determined,one can choose the valve orifice sufficient range of designed pressure conditions.efficiency including defrost can be expressed as,def 0þQ sys Áb ð2Àa Àa 0ÞW sys Áa þW def ð1Àa Þð3Þwhere a =nominal heating ratio,a 0=heating ratio under non-stop heating,b =heating supply ratio during defrost.A general air source heat pump system’s a =0.8–0.85and b =zero.For example,37.5kW nominal heating capacity and 8.62kW power input model’s nominal COP is 4.35.(a ¼1;a 0¼1;b ¼0).But energy efficiency would be decreased in a manner inversely proportional to the heating ratio,a during the defrost condition (a <1;a 0¼1;b ¼0),as shown in Fig.3.The energy efficiency will be decreased by 15%at a =0.8.On the other hand,from the perspective of non-stop heating (a ¼1;a 0<1;b –0),the energy efficiency will be a function of a 0and b .The estimated efficiency of non-stop heating was shown in Fig.4.The a 0and b were increased along with the total efficiency.Assuming that b is 0.5,the non-stop heat pump can provide 50%of nominal heating during frost with the same efficiency or higher under the same heating ratio (a ¼a 0¼0:8).3.3.The evaluation factors of continuous heatingThere are some indexes to evaluate the durability of continuous heating performance.Fig.5shows normal and abnormal defrost cycle.Fig.5a is the general cycle diagram of RCD and Fig.5b and: Strainer : Check valve : Capillary tube: Electronic expansion valve : Electronic solenoid valve: Pressure switch: Pressure sensor Service valve (Gas)Service valve (Liquid)Compressor 1Oil SeparatorCompressor 2Heat ExchangerSubcool unitFour way valveAccumulatorIDU1IDU2IDU3Indoor Side: CondenserOutdoor Side: EvaporatorIDU4BLDC fanBranch tubeDual Spray LineHEX1HEX2HEX3: Temperature sensor Fig.2.Dual spray hot gas cycle diagram.Fig.3.Heating COP according to heating ratio.J.Y.Jang et al./Applied Energy 110(2013)9–1611c are bad examples for DSHG.The definition of theta,h ;is the slope of the same event during the cycle.For example,as in Fig.5,the slope of the defrost starting point is a very good choice.Ifh has a negative value,that means the cycle will collapse gradually and the heating capacity will decline as well,as shown in Fig.5b.The other index s is the ratio of heating operation.If s is less than one,that means the defrost period will occur too frequently and the cumulated heating capacity could decrease,as shown in (c).So h must be greater than zero and s must be greater than one to maintain the steady heating cycle.s 1¼t 2=t 1;s 2¼t 3=t 2;ÁÁÁð4ÞThe index of each cycle shown in Fig.5is (a)h :0:18;s :0:99,(b)h :À6:31;s :1:02and (c)h :0:24;s :0:71:In conclusion,this in-dex should be fully considered before the evaluation of continuous heating.Experimental setup and test procedure 4.1.Test model and facilityAs shown in Fig.2,the test heat pump has two scroll-type com-pressors,which are composed of one inverter and one constant speed type.A wide louver fin-type heat exchanger,two electronic expansion valves,an accumulator and a reversing valve are theother main components of the test heat pump,and two newly-added solenoid valves control the dual-spray hot gas flow.The test heat pump has a 33.5kW nominal cooling and 37.5kW nominal heating capacity (12hp).The nominal power input is 9.11kW for cooling and 8.62kW for heating.Accordingly,the coefficient of performance of the tested heat pump is 3.68for cooling and 4.35for heating.Four cassette-type indoor units,each with a nominal heating capacity of 10kW,were connected.Each connecting tube length to indoor unit from branch was 30m and total length of pipe was 150m.The test heat pump was installed in a psychrometric calorime-ter as described in Fig.6.The test facility is composed of two cham-bers:an indoor and an outdoor chamber as well.Both chamber can measure cooling capacity up to 87kW and able to control the tem-perature from 60°C to À20°C.To achieve this,it has three cooling coils of 87kW in total for normal temperature and a cooling coil of the same capacity for low temperature.In addition,it has a large heater of the same capacity and a humidifying heater to control the relative humidity within 5–95%.In the indoor side,a large capacity duct of nozzle type is in-stalled to measure the amount of wind.It can measure up to 130CMM.When measuring the specimen capacity,its allowable error was calibrated to be made under 2%.The tested psychromet-ric type calorimeter satisfied the ISO standard.The data acquisition system automatically collected all experiment data,including tem-perature,humidity,wind amount and power consumption,and stored them continuously during the test in order to avoid any hu-man error in data reading.The data recording instrument provided 64data channels,which scanned all data channels consecutively every 2s.To reduce experimental uncertainties,Null points of all temperature and pressure sensors were adjusted with the calibra-tions.The temperature sensors were calibrated with a standard water bath over a temperature range of À30°C to 120°C with an error of ±0.2°C.The relative error of the pressure sensors over the effective measurement range of 0–25bars was ±0.2%.The rel-Fig.4.COP estimation during non-stop heating.Fig.5.New evaluation factors during defrosting.Energy 110(2013)9–16types of defrosting performance,opera-the same specimen.2°C/1°C (DB/WB),under full load con-evaporating exchanger fin was covered with frost air about 80–90min after the system operation was determined by the tempera-the evaporator and the outdoor air.The was another basis for defrost starting.the temperature sensor to check the temperature sensor HEX1of evaporator condition of the defrosting operation temperature of the outdoor heat exchanger.different directions of entrance and exit of during the defrosting operation and the judged based on the exit temperature,used HEX2,3.In DSHG,HEX 2judges upper-side heat exchanger,while HEX 3heat exchanger.RCD was first performed mentioned above.As a result,it took about cycles.The defrosting started when the dif-between the evaporator and the outdoor air was 10°C,and ended when the outdoor heat exchanger tempera-ture was 25°C.To compare with RCD,the newly designed DSHG was performed with same defrost judgment logic and test time,but a different temperature due to its heat limit.After finding the optimized DSHG operation,starting and ending temperatures were maintained as 8°C and5°C respectively over 3min.This earlier stage defrosting concept is referred to in the next chapter.5.Test results 5.1.RCD method resultsThe key factor of RCD defrosting is to change the reverse valve direction from heating to cooling.Once the defrosting operation starts,it is necessary to reduce the rotation speed of the inverter compressor.If a constant-speed compressor is in operation,it shuts down to prevent a drastic pressure change.When the reverse valve is switched to defrosting,the discharge pressure is dropped sharply and the suction pressure is increased.Finally,the high pressure and the low pressure meet to achieve a balance,and it is an inev-itable phenomenon that an outdoor heat exchanger which used as a low pressure evaporator will be switched to a high pressure con-denser.On the other hand,the indoor heat exchanger that operated as a high pressure condenser changes its role to function as a lowpressure evaporator.At this time,the indoor fan shuts down to cut off the supply of cool air into the room.The outdoor fan is also usu-ally turned off to achieve a rapid pressure increase.Fig.7shows the dynamic characteristics of the RCD test cycle.The temperature of the HEX sharply increased due to the incoming superheated hot gas from the compressor discharge.During the defrosting stage,the discharge pressure increased slightly,but the suction pressure decreased noticeably,from 800kPa to 300kPa.As mentioned above,when the fan of the indoor evaporator is not operating,both start and end of defrosting operation will be judged by the temper-ature sensor of HEX1,the defrosting range temperature of HEX1rises rapidly to 0°C and stays there for a certain period of time be-fore it goes up again rapidly.This is because the frost melts into water at the temperature 0°C.Once it reaches 25°C,the set-up temperature for ending defrost,it enters the defrosting end step,lowering the inverter frequency and switching off the 4-way valve to stop the whole defrosting operation.The defrosting time is usu-ally counted up to the switching-off point of the 4-way valve.In the case of this specimen,the defrosting operation was carried out for 5min,from about 400–700s,as shown in Fig.7.This defrosting time is much shorter than that of hot gas defrosting,and it is one of the greatest advantages of the RCD method,in that defrosting can be achieved without much additional cost.Indoor UnitOutdoor UnitFANCooling CoilHeaterHumidifierAir-samplerNozzle chamberInducted FanTemp. & Humid.Measuring sensorPre-HeaterMass flow meter Connecting tubeFig.6.Psychrometric test facility.Fig.7.RCD cycle.J.Y.Jang et al./Applied Energy 110(2013)9–1613The RCD method uses two energy sources forthe power input of the compressor,and the other istion heat of the indoor side.Most of defrostingpower input of compressor,but some of the indoorheat caused by natural convection brings about thedoor air that has been deprived of heat source inoperation range,lowering the indoor temperatureOnce the defrosting is completed and the heatingried out again,a certain period of time is needed toheat exchanger and the ducts.As shown in Fig.7,theof the system is stabilized about700s after the start ofand it takes more than10min until it reaches2500kPathe normal cycle before defrosting.So,we can see thatdefect of RCD type is the duration of a certain timeindoor heating supply in the defrosting range until theindoor temperature and the supply of normal heatinging.The duration time for three cycles of defrostingabout4h in total.For fair comparison and analysis,chamber temperature was sufficiently raised to dry the outdoorheat exchanger and the measuring temperature was reset;then, the test was carried out using the DSHG method for4h.5.2.DSHG method resultsFig.8shows the cycle characteristics of DSHG defrosting range. Like RCD,its defrosting start also judged based on the value given by temperature sensor HEX1on the evaporator side.But,a differ-ent logic is used for the starting condition to enter the defrosting operation earlier than RCD.This is because the DSHG method is limited in terms of its capacity to melt the frost layer due to the fact that less energy is provided to the defrosting heat source than in the RCD method,and so the defrosting process must be carried out before the frost layer gets too thick.Actually,the enthalpy dif-ference in the defrosting range estimated from the temperature defrosting results in more frequent defrosting operations than RCD during the same time duration.When the DSHG method is used,the circulationflux of refriger-ant is,as shown in Eq.(1),divided into the indoorflux for the ac-tual heating and the defrostingflux.The balance of eachflux can be adjusted according to the designed orifice size of dual spray line and the expansion valve of outdoor heat exchanger.The increase of indoorflux will naturally be better for the indoor heating supply in the defrosting range,but it causes a decrease in defrostingflux and elongates the defrosting time,increasing the possibility of defrosting failure.On the other hand,if the indoorflux is reduced, it can shorten the defrosting time and raise the defrosting capacity, but it lowers the indoor heating capacity in the defrosting range, diminishing the benefit of continuous heating.The optimumflux ratio designed through the preliminary checkup is to keep the in-heating capacity in the defrosting range to be50–55%operation right before the start of defrosting,andthe refrigerantflux into about5:5.The defrostingwith opening the hot gas valve on the top and closing expansion valve on the top to remove the frost on the upper exchanger.The defrosting end of the upper side is judged HEX2that checks the outlet temperature of the upperoptimum temperature was set as about5°C.The the defrosting temperature is relatively lower thanthe same as the reason why its defrosting capacity isFig.8.DSHG cycle.Fig.9.Heating capacity results. Fig.10.Heating COP results.14J.Y.enthalpy difference that was mentioned above.Once the upper defrosting is completed,the lower side defrosting is carried out the opposite valves.Likewise,the defrosting end of the lower is judged by sensor HEX2,which checks the outlet tempera-of the lower side.In the case of DSHG,the high pressure of system in the defrosting range drops from2800kPa G to kPa G,as shown in Fig.8.This corresponds to the case in which the heating effusion temperature of the indoor side goes down about10°C,from43°C to33°C.It is a natural consequence the fact that the indoor heatingflux dropped down to about50% regular state,but it may cause unpleasantness for users.So,an additional logic is necessary to reduce the RPM of indoor fan in the defrosting range and raise the outlet temperature.Total time re-quired tofinish defrosting was about10min,while the reverse cy-defrost was about5min with the approach of4-way valve switching.As mentioned above,in RCD,when the defrosting oper-is carried out for5min from the point of4-way valve switch-the indoor heating supply is stopped,and it takes about 10min to resume the normal heating state after defrosting.But, in DSHG,while the defrosting is carried out for about10min,50–55%of heating is supplied to the indoors and the high temper-ature of the system is,as shown in Fig.8,stabilized immediatelyafter the defrosting and recovers the normal heating supply.Ofcourse,its defrosting capacity is lower than that of RCD,and thusthe condition to use DSHG method can be restricted.5.3.Performance comparison and discussionFig.9shows the heating capacity results between common RCDand DSHG that we designed.The instant COP results are alsoshown in Fig.10.The RCD method should shut down the indoorheating periodically to defrost.The designed DSHG method couldprovide about50%of nominal heating during defrost.So,total heating capacity was increased by17%compared to thereverse cycle defrosting,as shown in Fig.9.There were two reversecycle defrosting and four hot gas defrosting over approximately4h.The index noted was h:0:24;s:0:96;this means that the non-stop heating cycle would remain stable.As we have seen inFigs.9and10,there were more defrost operations using DSHGcompared to RCD defrost.The hot gas method takes a long timeto remove frost with a relatively low defrost heat because it shares heat energy with defrost and indoor heating supply.Therefore,to use the hot gas method,as mentioned previously,earlier stage de-frost is required before a full growth of frost,and this is why a higher hot gas defrost frequency appears under the same condi-tions.The cumulated performance indicators are shown in Fig.11.The slope of the graph was almost the same between the two methods,but the slope of the reverse cycle defrosting was zero during defrost because of the indoor heating interruption.On the other hand,the slope of the hot gas defrosting that was designed to supply50%of nominal heating capacity during defrost was uni-form,without variations.The cumulated performance of DSHG was increased by17%compared to the reverse cycle defrosting during 4h.The cumulated power input was increased by7.8%in the same mechanism.So,total cumulated energy efficiency was increased by 8%,and compared with using an electronic heater that covered the lack of heating capacity,and total efficiency was increased by27%, as shown in Fig.12.Fig.13shows the photographs of dual hot gas spray defrost. When heat exchanger frost has grown and the signal has reached the designed condition,one-side defrosting begins.Generally, upper side defrosting should be startedfirst,because the cold water that has been melted from the upper side heat exchangerFig.11.Cumulated performance results(a)capacity and(b)power input.Fig.12.Cumulated COP results.may freeze again at the lower side.But repeated upper side defrosting can have deleterious effects on the lower side heat ex-changer,such as causing frost to remain on the lower side.In some cases,lower side defrosting should be startedfirst,as shown in Fig.13b.6.ConclusionsIn this study,a newly designed dual hot gas spray defrosting method was examined,and the defrost performance was com-pared to a traditional reverse cycle defrosting method.The results can be summarized as follows:(1)A simple and low cost dual hot gas defrosting cycle structurewas introduced and the most effective bypassed hot gasflow rate achieved was about50%of the total refrigerantflow rate during the defrost.(2)The total heating capacity was increased by17%and theinput power was increased by7.8%over4h test period.Finally,the total energy efficiency was increased by8%com-pared to reverse cycle defrosting.(3)The total efficiency could be increased by27%comparedwith system using an electronic heater that covered the lack of heating capacity under the reverse cycle defrosting method.The main achievement of this study is the discovery of a meth-od to supply heating continuously in the large capacity cycle of 30kW or higher without any additional equipment,through the optimum control of hot gases,overcoming the problem of the need to cut off the heating supply in the defrosting stage,which has been the greatest shortcoming of air source heat pump.But,as the indoor heating supply was just about50%of normal operation in the defrosting range when the outdoor temperature was2°C and the heating is provided in all rooms,it may cause complaints by users due to the lowered outlet temperature of the indoor facil-ity.Furthermore,if the outdoor temperature drops lower than2°C, the indoor heating supply in the defrosting range may drop below 50%.Therefore,for the application to the actual product,it may re-quire the control of diverse boundary conditions.For example,its use may be confined to cases when it is not the whole room indoor operation,or when there is some margin in the frequency of the in-verter compressor in normal operation.References[1]Aynur Tolga N.Variable refrigerantflow system.Energy Build2010;42:1106–12.[2]Park Youn Cheol,Kim Young Chul,Min Man-Ki.Performance analysis on amulti-type inverter air conditioner.Energy Convers Manage2001;42:1607–21.[3]Tassou SA,Marquand CJ.Effects of evaporator frosting and defrosting on theperformance of air-to-water heat pumps.Appl Energy1987;28:19–33.[4]Kaygusuz Kamil.Performance of an air-to-air heat pump under frosting anddefrosting conditions.Appl Energy1994;48:225–41.[5]Mei Viung C,Chen Fang C,Domitrovic Ronald E,Kilpatrick Joe K.A frost-lessheat pump.Ashrae Trans2002;108(1).[6]Byun Ju-Suk,Lee Jinho,Jeon Chang-Duk.Frost retardation of an air-source heatpump by the hot gas bybass method.Int J Refrig2008;31:328–34.[7]da Silva Diogo L,Hermes Christian JL,Melo Claudio.Experimental study of frostaccumulation on fan-supplied tube–fin evaporators.Appl Therm Eng 2011;31:1013–20.[8]Mader G,Thybo C.A new method of defrosting evaporator coils.Appl ThermEng2012;39:78–85.[9]Kwak KM,Bai CH.A study on the performance enhancement of heat pumpusing electric heater under the frosting condition.Appl Therm Eng 2010;30:539–43.[10]Liu Zhiqiang,Tang Guangfa,Zhao Fuyun.Dynamic simulation of air-sourceheat pump during hot-gas defrost.Appl Therm Eng2003;23:675–85.[11]Zhiyi Wang,Xinmin Wang,Zhiming Dong.Defrost improvement by heat pumprefrigerant charge compensating.Appl Energy2008;85:1050–9.[12]Chen Yi-guang,Guo Xian-min.Dynamic defrosting characteristics of air sourceheat pump and effects of outdoor air parameters on defrost cycle performance.Appl Therm Eng2009;29:2701–7.[13]Wenju Hu,Yiqiang Jiang,Minglu Qu,Long Ni,Yang Yao,Deng Shiming.Anexperimental study on the operating performance of a novel reverse-cycle hot gas defrosting method for air source heat pumps.Appl Therm Eng 2011;31:363–9.[14]Dong Jiankai,Deng Shiming,Jiang Yiqiang,Xia Liang,Yao Yang.Anexperimental study on defrosting heat supplies and energy consumptions during a reverse cycle defrost operation for an air source heat pump.Appl Therm Eng2012;37:380–7.[15]Huang Dong,Li Quanxu,Yuan parison between hot-gas bypassdefrosting and reverse-cycle defrosting method on an air-to-water heat pump.Appl Energy2009;86:1697–703.[16]Stoecker WF.Selecting the size of pipes carrying hot gas to defrostevaporators.Int J Refrig1984;7(4).[17]Hoffenbecker N,Klein SA,Reindl DT.Hot gas defrost model development andvalidation.Int J Refrig2005;28:605–15.[18]Choi Hwan-Jong,Kim Byung-Soon,Kim Kyung Chun.Defrosting methodadopting dual hot gas bypass for an air-to-air heat pump.Appl Energy 2011;88:4544–55.[19]KS B ISO15042.Multiple split-system air-conditioners and heat air-to-airpumps testing and rating for performance2006.(a) Defrost start(b) Half progress(c) Defrost endingFig.13.Dual spray defrosting results.16J.Y.Jang et al./Applied Energy110(2013)9–16。