Applied Energy 147 (2015) 325–334

Contents lists available at ScienceDirect

Applied Energy

journal homepage: www.elsevier .com/locate /apenergy

Performance of a free-air cooling system for telecommunications basestations using phase change materials (PCMs): In-situ tests

http://dx.doi.org/10.1016/j.apenergy.2015.01.0460306-2619/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding authors at: College of Civil Engineering, Hunan University, No. 2Lushan South Road, Yuelu District, Changsha 410082, China. Tel.: +86 731 88821254; fax: +86 731 8882 1005 (Q. Zhang). Civil, Environmental & ArchitecturalEngineering Department, University of Kansas, 1530 W. 15th St., Lawrence, KS, USA.Tel.: + 1 785 864 3604; fax: + 1 785 864 5631 (M.A. Medina).

E-mail addresses: quanzhang@hnu.edu.cn (Q. Zhang), mmedina@ku.edu(M.A. Medina).

Xiaoqin Sun a,b, Quan Zhang a,⇑, Mario A. Medina b,⇑, Shuguang Liao c

a College of Civil Engineering, Hunan University, Changsha 410082, Chinab Dept. of Civil, Env. & Arch. Engineering, University of Kansas, Lawrence 66045, USAc Changsha Maxxom High-tech Co. Ltd., Changsha 410015, China

h i g h l i g h t s

� A free-air cooling system that uses PCMs was developed for TBSs.� PCM stores natural cold energy and release this energy into TBSs.� The operation of A/C is shortened, resulting in energy savings.� The largest energy savings ratio of this unit was 67% with an average value of 50%.� The average annual operating time that this unit replaced traditional A/C use was 83%.

a r t i c l e i n f o

Article history:Received 9 August 2014Received in revised form 11 January 2015Accepted 12 January 2015Available online 16 March 2015

Keywords:Phase change materials (PCMs)Free-air coolingLatent heat storageEnergy savings ratio (ESR)Telecommunications base station (TBS)

a b s t r a c t

A free air cooling system that combines phase change material (PCM) with a natural cold source (i.e., coldair) was developed to reduce the space cooling energy consumption in telecommunications base stations(TBSs). Outside cold air, instead of air conditioning system was used to remove heat in the TBSs. Inaddition, a PCM technology was adopted to improve the mismatch between energy demand and supplyon the electric grid. The proposed system was intended to operate in conjunction with existing airconditioning units within each TBS. Consequently, the running time the air conditioning units wasreduced, resulting in energy and demand savings. A full scale prototype, herein referred to as latent heatstorage unit (LHSU), was designed, built and tested in TBSs located in five different climatic regions insouthwest and eastern China during different seasons. In addition, a mathematical model was developedto simulate the operation of the proposed LHSUs. Energy savings ratio (ESR) was used as the criterion toevaluate LHSU’s energy savings. The estimated average annual ESR in five climatic regions wasapproximately 50%, with a maximum value of 67%. The average percent running time in which theLHSU replaced the operation of conventional air conditioners was 82.6%, with values surpassing 75% infour of the five cities, and one city achieving a value of almost 100%.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

China is the world’s largest mobile phone market. As of the firsthalf of 2013, the number of mobile phones exceeded 1.15 billion[1]. With this vast amount of signal traffic, the thermal manage-ment of telecommunications base stations (TBSs) has become

exceedingly challenging. TBSs are small buildings located in fixedlocations, where reception and transmission equipment is housed.This equipment generally consists of electronic and heat generat-ing hardware used to handle signal traffic between mobile phonesand network subsystems. As the numbers of these stations andnetwork equipment continue to increase, so is the amount of elec-trical energy needed to keep their indoor air at specified tempera-tures and relative humidity [2,3], which is critical for the properfunctioning of electronic equipment. The electrical energy whichis consumed by the air conditioners that condition the air in thesestations accounts for 30–50% of the total energy used in TBSs [4].This is the case because for most part these air conditioners work24 h per day all year round. Therefore, the space cooling of these

Nomenclature

A area (m2)c specific heat (kJ/kg/�C)d diameter (m)G volumetric flow rate (m3/h)Dhm latent heat of fusion (kJ/kg)I electricity consumption (kW h)l length (mm)Nu Nusselt numberP Power (W)Pr Prandtl numberQ heat transfer rate (W)r radius (mm)Re Reynolds numberT temperature (�C)DT temperature difference (�C)V volume (m3)

AbbreviationsESR energy savings ratio (%)LHSU latent heat storage unitLMTD log mean temperature differencePCMs phase change materialsTBS telecommunications base station

Greek symbolsa convective heat transfer coefficient (W/m2 �C)c melting or solidification constante relative error (%)k thermal conductivity (W/m �C)q density (kg/m3)s running time (h)%s percent running time (%)

Subscriptsa airaw between air and waterbc boundary conditioncl cooling loadcon conventional air conditionercp copper pipeec energy charging processed energy discharging processf fanfa fresh air processfin finalin indoorini initialm phase change materialout outdoorp pumps solidw waterwm between water and PCM

Superscriptsed energy discharging processin inletn Prandtl number exponent in Dittus–Boelter equation

(0.3 for cooling and 0.4 for heating)out outlett test

326 X. Sun et al. / Applied Energy 147 (2015) 325–334

stations is considered a primary target for energy management andconservation [3]. Of the technologies based on the use of naturalcold energy, free air cooling has drawn significant attention as aneffective approach to reduce the space cooling energy use in TBSs[5,6]. However, the intermittent and unpredictable nature of thenatural cold energy poses major challenges for its implementation.A possible solution is the use of a thermal storage component tocorrect the gap between energy demand and supply [7].

Kuznik and Virgone [8] and Zhou et al. [9] pointed out thatphase change materials (PCM) could decrease temperature fluctua-tion and reduce energy consumption in buildings. As a result sev-eral thermal energy storage systems have been developed. Forexample, Halawa and Saman [10] and Gracia et al. [11] providedinformation for the development of air based phase change ther-mal storage units, where air passed through PCM slabs. Their theo-retical and parametric studies highlighted first order parametersfor the design and operation of these units. Mahmoud et al. [12]developed six heat sinks with six types of PCMs to cool electronicdevices. The heat sink with single cavity inserting honeycombinside or filled with PCM with a low melting temperature (29 �Cin their research) showed the best performance in terms of tem-perature control. Darkwa and Su [13] simulated microencapsulatedPCM systems in rectangular, triangular and pyramidal config-urations. It was concluded that microencapsulated PCM particledistribution enhanced the effective thermal conductivity byapproximately 10 times, but reduced the energy storage capabilityby 48%. Therefore, other methods for enhancing both the thermalconductivity and energy storage capability had to be developed.To solve the heat transfer problem within the energy storage

system, Tay et al. [14,15] and Amin et al. [16] proposed the effec-tiveness-NTU method to characterize tube-in-tank and PCMencapsulated in spherical thermal energy storage systems.Correlations were established to predict the average thermalcapacity for the systems. Longeon et al. [17] studied the heat trans-fer mechanisms that take place during phase transition processesin annular latent storage units. Based on the above and in combina-tion of the free-air cooling theory and the above heat transfermechanism, Liao et al. [18] developed a free-air cooling systemthat uses latent heat storage in combination with a natural coldsource (i.e., outside cold air) to serve alongside existing airconditioning systems in which case outside cold air is directly usedto cool the interior space of TBSs. The excess cold energy thatremains in the cold air is stored in the PCM, which is then releasedinto the conditioned space when outside cold air is not available.Therefore, the operation of conventional air conditioners can beshortened resulting in significant energy savings. A prototype unit,herein referred to as latent heat storage unit (LHSU), was designed,built, and experimentally verified in an enthalpy difference labo-ratory. The annual adjusted energy efficiency ratio of this unitwas between 11.98 W/W and 14.91 W/W, which varied with out-door and indoor air temperatures [19].

To evaluate the application of this technology in TBSs, in-situtests of the energy savings potential of the free-air coolingsystem were conducted in five different climatic locations duringdifferent seasons. Five cities from around China with a wide rangeof climatic conditions were selected, including Shenyang (severecold region), Zhengzhou (cold region), Kunming (mild region),Guangzhou (hot summer and warm winter region), and Changsha

Table 1Thermal properties of the PCM used in the energy storage modules.

Type Range of transition temperatures (�C) Latent heat of fusion (kJ/kg) Heat conductivity (W/m �C) Specific heat (kJ/kg �C) Density (kg/m3)

Lower Upper Solid Liquid Solid Liquid Solid Liquid

Organic composite 18 20 180 0.72 0.50 2.32 2.11 942 895

X. Sun et al. / Applied Energy 147 (2015) 325–334 327

(hot summer and cold winter region). In addition, a mathematicalmodel was developed to simulate the annual operation of the LHSUacross the five climatic regions. The energy savings ratio (ESR) wasused as the criterion to evaluate the energy savings that resultedfrom integrating the LHSU in TBSs. The correlation between ESRand outdoor air temperature was established to guide the applica-tion of LHSU in different climatic regions.

2. The application of a free air cooling system in TBSs

2.1. Physical parameters of a TBS

TBSs are classified as Class AA through C depending on theirfunction [20]. In this paper, Class C TBSs were studied becausethese are the most commonly used by mobile networks. In thisclass of TBSs, the established indoor air temperature range isbetween 18 and 28 �C; furthermore, this temperature must neverexceed 30 �C. The indoor air relative humidity is between 40%and 70%. On average, the typical overall heat transfer coefficientof Class C TBSs’ walls is around 1.1 W/m2 �C. The average electronicequipment housed in a typical 20-m2 Class C TBS generates heat atthe approximate rate of 3 kW [21].

2.2. Description of the free air cooling system and its various heattransfer processes

The free air cooling system was designed and developed for thepurpose of reducing energy consumption in TBSs [22,23]. As such,outside cold air was brought into the TBS for cooling via the latentheat storage unit (LHSU). PCMs were introduced to store any extracold energy contained in the outside air. This stored energy wouldbe released when needed. That is, when the cold air alone could notkeep up with the space cooling needs. In addition, this system wasdesigned to operate alongside conventional air conditioners.Consequently, the operation of air conditioning system would beshortened, resulting in energy and demand savings. The mainand novel feature of this system was that PCMs, enclosed in modu-lar containers, herein called energy storage modules, in combina-tion with a natural cold source (i.e., outside cold air), providedthe space cooling necessary to keep the indoor air of TBSs at speci-fied temperatures. The energy storage modules (sealed stainlesssteel containers) were filled with PCMs. The thermophysicalparameters of the PCM are shown in Table 1. While in operation,water circulated in a closed loop between the energy storage mod-ules and a heat exchanger, herein referred to as air cooler. Energycharging and discharging processes, as used in this paper, weredefined such that energy charging meant that the incoming outsideair was used to solidify liquid PCM. Conversely, energy dischargingmeant that heat generated by the electronic equipment within theTBS was carried away by the air and into the free air cooling systemto melt the PCM.

The storage capacity of the LHSU in this study was 5 kW h. Thecooling capacity of this unit was 3 kW, which was calculated basedon an indoor air temperature of 28 �C. The dimension of this unitwas 68 cm � 65 cm � 1.68 m. The three processes under whichthe LHSU operated were described in Table 2. This table includesoperating temperature ranges, fan and pump power and flow rates,and LHSU inlet air valve position. The processes underwent by the

LHSU during operation were energy charging (I), fresh air (II), andenergy discharging (III).

2.2.1. Energy charging process (Mode I)During periods when the outside air temperature was lower

than the PCM melting temperature (20 �C), the LHSU was switchedon to cool down both the indoor air and the circulating waterbetween the energy storage modules and the air cooler. This watercharged the PCM with cold energy. This is shown in Fig. 1.

During this process, the energy charging rate in the PCM, Qwm,was calculated based on the volume of solidified PCM on thePCM side and the temperature difference between inlet and outletwater on the water side as shown in Eq. (1). The following assump-tions were made:

1. The sensible heat and natural convection within the PCM wereneglected.

2. The thermal resistance, heat capacity, and axial conduction ofthe water pipe were neglected.

3. The heat loss from the air cooler was neglected.

Heat transfer on the PCM side:

Initially, the inlet and outlet water temperatures (Tinw1,Tout

w1 ) toand from the energy storage modules were set to 20 �C, whichwas also the melting temperature of the PCM.

Qwm ¼5

18qs �

Vs

sec� Dhm ð1Þ

where Vs ¼ r þ 2cs

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiks

qscssec

s !2

� r2

24

35pl� 10�9 ð2Þ

andffiffiffiffipp

cs expðc2s Þerf ðcsÞ ¼

csðTm � TbcÞDhm

ð3Þ

and where the ‘‘5/18’’ in Eq. (1) was related to unit conversionsEqs. (2) and (3) [24] were used to calculate the volume of solidifiedPCM. In these equations, Vs, qs, cs, ks, and cs were the volume,density, solidification constant, thermal conductivity, and specificheat of the solid phase of the PCM, respectively. The Dhm and Tm

were the latent heat of fusion and the PCM solidification tempera-ture, respectively. The energy charging time was given by sec. Thewater pipe length and radius were l and r, respectively. The boundarycondition was given by Tbc, which was calculated using the LMTD(log mean temperature difference) between the PCM and water.

Heat transfer on water side:The convective heat transfer coefficient of the circulating water,

aw, was calculated using iterations between the Dittus–Boelter[25] and Newton’s cooling law equations, as shown in Eqs. (4)and (6). Water temperatures were calculated based upon conver-gence of Eqs. (5) and (6),

Nu ¼ 0:023 Re0:8 Prn ð4Þ

aw ¼ Nukw

dð5Þ

Qwm ¼ awAcpDTwm ð6Þ

Table 2Modes of operation and processes of the latent heat storage unit (LHSU).

Mode Process Temperature (�C) Power (W) Rate (m3/h) Valve position

Indoor Outdoor Fan Pump Fan Pump

I Energy charging Tin > Taout and [18,30) (0,20] 400 50 2000 0.4 ab

II Fresh air Tin > Tout and (20,30) (20,25] 400 – 2000 – aIII Energy discharging [25,30) Tout > 25 400 50 2000 0.4 b

a Tin and Tout are the indoor and outdoor temperatures, respectively.b Refer to Figs. 1–3.

TBS

exhausted aira

water loop

fresh air

outlet air

1234

T1

T2

T3

T4

T2T1 Tm

<<

controller

T4 T3 Tm<<

ACESM

Fig. 1. Energy interactions within the LHSU during energy charging (Mode I).

328 X. Sun et al. / Applied Energy 147 (2015) 325–334

In Eq. (4), the value of n was 0.3 for cooling and 0.4 for heating. Theheat transfer rate between the water and PCM, Qwm, was calculatedusing Eq. (6), where Acp was the outer surface area of the pipe insideof which the water circulated, DTwm was the log mean temperaturedifference between the water and the PCM, kw was the thermal con-ductivity of the circulating water, and d was the diameter of thepipe.

The energy savings ratio (ESR) was developed to evaluate theenergy savings that would result from using the LHSU. The ESRwas defined as the ratio of energy savings attained by using LHSUwhen compared to the energy consumption of conventional airconditioners. In the energy storage mode, the energy savings ratio(ESRec) and electricity consumption (Iec) of LHSU were given by

ESRec ¼Icon � Iec

Icon� 100% ð7Þ

XDIec ¼

XPf

Qcl

Q awDsec þ

XPpDsp

� �� 10�3 ð8Þ

a

water loop

fresh air

T1T3

T4 ACESM

Fig. 2. Energy interactions within the LHSU

where Pp and Pf were the power of the pump and the fan, sp was thepump running time, and Icon was the electrical energy consumptionof the conventional air conditioners.

2.2.2. Fresh air process (Mode II)During this process, fresh outside air was drawn into the tele-

communications base station (TBS), passing through the LHSU,both to ventilate the interior space and to offset the use of the con-ventional air conditioners. Extraction fans vented the inside airfrom the conditioned space to the outside. This is shown inFig. 2. To reduce the heat transfer between the incoming air andthe PCM, the water circulating pump was automatically shut off.

It was assumed that the heat removed by the LHSU was equal tothe indoor space cooling load, Qcl, which was calculated by Eq. (9).The electricity consumption (Ifa) and the energy savings ratio(ESRfa) of the LHSU during this process were

Qcl ¼5

18qaGacaðTin � Tout

a Þ ð9Þ

TBS

exhausted air

outlet air

T2

T2Tm T1< =T3 T4 1

234

controller

=

while in the fresh air mode (Mode II).

breturn air

TBS

water loop

outlet air

T1

T2

T3

T4

Tm T1T2< <T3 T4Tm << 1

234

controllerACESM

Fig. 3. Energy interactions within the LHSU during energy discharging (Mode III).

X. Sun et al. / Applied Energy 147 (2015) 325–334 329

XIfa ¼

XPf Dsfa � 10�3 ð10Þ

ESRfa ¼Icon � Ifa

Icon� 100% ð11Þ

where qa, Ga, and ca were the density, volumetric flow rate, andspecific heat of the incoming air, respectively. Tin and Ta

out werethe temperatures of the indoor air and outlet air of the LHSU,respectively. The duration of this process was given by sfa.

2.2.3. Energy discharging process (Mode III)When the outside air temperature surpassed 25 �C, the LHSU

was not able to remove all the heat generated within the TBS withoutside air. Therefore, the cold energy that had been stored withinthe PCM during the energy charging process was used to cool theinside air. This is shown in Fig. 3.

The duration of this process, sed, was determined by the amount

of stored energy in the PCM and the energy discharging rate,Qedwm,

which was given by

XDsed ¼

PQwmDsec

Q edwm

ð12Þ

The heat transfer rate between the water and the PCM during this

process,Qedwm, was calculated in a similar fashion as with the energy

charging process, but replacing the parameters of the solid PCMwith those of the liquid PCM.

The electricity consumption (Ied) of LHSU and the energy sav-ings ratio (ESRed) during this process were calculated byX

DIed ¼XðPp þ Pf ÞDsed � 10�3 ð13Þ

ESRed ¼Icon � Ied

Icon� 100% ð14Þ

Table 3Climate information for the five selected cities.

City Location Climatic region Annualtemper

Shenyang 41�440N 23�530E Severe cold 8.6Zhengzhou 34�460N 13�390E Cold 14.7Kunming 25�040N 02�410E Mild 15.5Guangzhou 22�300N 14�170E Hot summer and warm winter 22.2Changsha 28�110N 12�580E Hot summer and cold winter 17.4

The total estimated energy savings ratio, ESR, attained by usingthe LHSU was evaluated by

ESR ¼ ESRec �%sec þ ESRfa �%sfa þ ESRed �%sed ð15Þ

where %s was the percent running time of the LHSU for each mode.

3. Experimental evaluation of the energy savings ratio

Energy savings ratios of the LHSU were estimated during theoperation of the unit in five typical Class C telecommunicationsbase stations (TBSs). Each station was located in a city with distinctclimatic conditions. The five cities were Shenyang (severe coldregion), Zhengzhou (cold region), Kunming (mild region),Guangzhou (hot summer and warm winter region) and Changsha(hot summer and cold winter region). The climatic parameters inthese cities are shown in Table 3 [26]. Information related to eachTBS is shown in Table 4.

A schematic diagram of the TBS in which the LHSU and the con-ventional air conditioners were tested is presented in Fig. 4.Outside fresh air was brought into the station across a wall. Anextraction fan was installed as per requirements of using the LHSU.

The testing procedure was as follows:

Step 1: Three TBSs with similar floor plan areas, electronicequipment, and air conditioners in each city were selected. Awatt-hour meter and two self-recording thermo-hygrometerswere installed in each TBS and the thermostats were set tothe same temperature in the three stations. One self-recordingthermo-hygrometer was set to record inside air temperaturesand humidity at intervals of five minutes and the other wasset to record outside air temperatures and humidity at the sametime interval. A watt-hour meter was used to record the elec-tricity consumption of the conventional air conditioners.

meanature (�C)

Annual mean temperature differencebetween daytime and nighttime (�C)

Annual temperaturerange (�C)

11.0 �23.4 to 34.110.0 �9.1 to 37.2

8.6 �2 to 29.97.6 4.7 to 36.67.1 �2 to 38.2

Table 4TBS building construction, existing capacity of air conditioners, and test dates.

Location Enclosuretype

Air conditioner Test date

Coolingcapacity (kW)

Quantity

Shenyang Brick andcement

7.5 1 October 20th–24th

Zhengzhou Brick andcement

5 2 November 9th–13th

Kunming Brick andcement

12 1 April 10th–17th

Guangzhou Brick andcement

7.5 2 February 25th–March 2nd

Changsha Light steelstructure

7.5 2 July 14th–August 9th

T1

1

IT1

T2

τ 3

2τ

1τ

I

T2T3τ

2τ3τ

Electronic hardware

Batteries Data collectorPower supplier

Watt-hour meter

A/C

LHSU

Extraction fan

T3

Fig. 4. Schematic diagram of the TBSs in which the LHSU and conventional A/Cwere evaluated.

330 X. Sun et al. / Applied Energy 147 (2015) 325–334

Step 2: The readings of the watt-hour meters were recordedevery 24 h and the data from the self-recording thermo-hygrometers was transferred to a data collector. Two TBSs withthe most similar temperature distribution and electricity con-sumption were selected out of the original three TBSs. Thelatent heat storage unit (LHSU) was installed in one of the twoselected TBSs. The TBS with conventional air conditioners was

Outdoor air temperature, indoor heat source, TBS enclosure

Calculate indoor air temperature

Identify the running modes of LHSU and air conditioners

Air conditioners Mode I Mode III Mode

Pcon, Icon

Qwm, τp, τec Qcl, τ

Pp, Pf, Iec

τed

Pp, IPp, Pf, Ied

τcon

Calculate ESR, E

edwmQEq.6

Eq.7

Eq.12

Eq.13

Eq.9

Eq.10

Retrofit TBS

Fig. 5. Protocol used to calcu

identified as the control TBS and the TBS with the LHSU (plusconventional air conditioners) was identified as the retrofitTBS. The control TBS served as a benchmark for comparison.Step 3: Two self-recording thermo-hygrometers were installedin the TBSs to record outlet air temperatures (T1) of the spacecooling equipment, indoor air temperatures (T2) and humiditywithin the station, respectively. One more self-recordingthermo-hygrometer was installed to record outdoor air tem-peratures (T3) and humidity.Step 4: Three timers were used for monitoring the runningtimes of the various operating components during the tests,including the conventional air conditioners (s1), the blowerfan (s2), and the extraction fan (s3).

The running times for Mode I (energy charging, sec), Mode II(fresh air, sfa), and Mode III (energy discharging, sed) were calcu-lated by Eqs. (16)–(18), respectively:

sec ¼I � Pcons1 � Pf s3 � ðPf þ PpÞðs2 � s3Þ

Ppð16Þ

sfa ¼ s3 � sec ð17Þ

sed ¼ s2 � s3 ð18Þ

The experimental energy savings ratio (ESRt) was calculated by:

ESRt ¼ðIfin � IiniÞ -ðI0fin � I0iniÞ

Ifin � Iini� 100% ð19Þ

where Pcon was the power of the conventional air conditioners, I wasthe total electricity consumption in the retrofit TBS, Iini and Ifin werethe initial and final electricity consumption readings in the controlTBS and the I0ini and I0fin were the initial and final electricity con-sumption readings in the retrofit TBS.

4. Comparisons between measured and simulated data

4.1. Comparisons between simulated and measured operations of theLHSU and the air conditioners

4.1.1. Simulated operations of the LHSU and the air conditionersTo evaluate the energy savings ratio of the latent heat storage unit

(LHSU) in different climatic regions, the operation of the LHSU inTBSs in five cities for one entire year was simulated. Because of

II

fa

fa

Outdoor air temperature, indoor heat source, TBS enclosure

Air conditioners

Pcon, Icon

τcon

SRec, ESRfa, ESRed

Control TBS

Calculate indoor air temperature

lated the various ESRs.

16:00 20:00 00:00 04:00 08:00 12:00 16:00

4

8

12

16

20

24

28

Retrofit TBS Control TBSL

HSU

and

A/C

out

let a

ir te

mpe

ratu

re(o C

)

Time

Fig. 7. Outlet air temperatures of the LHSU and conventional air conditioners.

X. Sun et al. / Applied Energy 147 (2015) 325–334 331

the varying outside and inside air temperatures, the LHSU changedits operation modes as well as its power demand and consumption.This is shown in Table 2. The procedure used by the simulatingprogram is shown in Fig. 5. The cooling demand of a TBS and theenergy consumption of conventional air conditioners were attainedusing Design Builder [27]. Design Builder is used to model severalbuilding environments. Its outputs, which are based on detailedtime steps using the EnergyPlus simulation engine, include energyconsumption, internal comfort data, and HVAC-related parameters(e.g., component sizes). Design Builder was selected because it canaccurately simulate of many common HVAC types, includingnaturally ventilated buildings. A south-facing TBS model was builtwith a dimension of 3.6 m � 4.1 m � 3.5 m. The overall heat trans-fer coefficient of its enclosure was estimated at 1.122 W/m2 �C.With the meteorological parameters exported from Design Builder,the operation of the LHSU was simulated using ‘‘MATLAB.’’ Theenergy savings ratio was computed via simulation results from boththe conventional air conditioners and the LHSU.

Fig. 6 shows the simulated operations of the LHSU andconventional air conditioners for one testing day in a TBS locatedin Kunming (Mild climate), China, including mode changes, energystorage, and energy demand.

During the test period between 16:00 and 20:00, the outside airtemperature was between 20 �C and 25 �C, which prompted theLHSU to run in Mode II (fresh air process). The TBSs inside airtemperature was maintained between 25 and 28 �C, which wascalculated via energy balance equations.

By 20:00, the outside air temperature decreased below 20 �C.This prompted the LHSU to start running in Mode I. The energystorage percentage was calculated by multiplying the heat transferrate (Eq. (1)) by the pump running time. At around 06:00, the out-side air temperature increased above 20 �C and the unit switchedto Mode II. Under this condition, there was no energy charging ordischarging; consequently, the energy storage percentage wasmaintained at 73.7%. The energy demand was basically that ofthe fan, which was 400 W.

When the outside air temperature was above 25 �C (08:40–10:30), the unit switched to Mode III, under which the PCM startedto melt. This process lasted about 1.8 h until all the energy stored inthe PCM had been released. At this point, the conventional airconditioners were switched on to remove the remaining heat insidethe TBS (10:30–16:00). Under this condition, the energy demand ofthe conventional air conditioners was approximately 1000 W,which was more than twice the energy demand of the LHSU.

26

4.1.2. Measured operations of the LHSU and the air conditioners in theretrofit and control TBSs

The outlet air temperatures of the latent heat storage unit (LHSU)and conventional air conditioners and TBS indoor air temperatures

020406080100

20242832

121620242832

16:00 18:00 20:00 22:00 00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:002004006008001000

Ene

rgy

stor

age

perc

enta

ge (%

)

Indo

or a

irte

mpe

ratu

re (

o C)

Out

door

air

te

mpe

ratu

re (

o C)

Conventional air conditioners power

Mode I Mode II Mode III

Pow

er (W

)

Fig. 6. Simulated operation of the LHSU and air conditioners in a mild climate.

were measured to illustrate the operation of these two units in theretrofit and control TBSs, respectively. Figs. 7 and 8 show the mea-sured operations of the LHSU and air conditioners in one day in thecity of Kunming. The operations of the two units in other testingcities followed a similar pattern, where both the LHSU and airconditioners changed their operating modes as a function of out-side and inside air temperatures.

Fig. 7 shows the experimental outlet air temperatures of theLHSU and conventional air conditioners in the retrofit and controlTBSs, respectively. The conventional air conditioners’ outlet airtemperature fluctuated during their operation. The LHSU’s outletair temperature, however, was stable when this unit was running.The fluctuations in the later periods were caused by the operationof conventional air conditioners. During this period, the outside airtemperature was above 25 �C and all the stored cold energy hadbeen released. Therefore, the LHSU stopped running. In this test,the conventional air conditioners worked for 7 h in the retrofitTBS and 24 h in control TBS. That is, with the application of theLHSU, the operation of conventional air conditioners was short-ened by 17 h in the retrofit TBS.

Fig. 8 shows the experimental indoor air temperatures in retro-fit and control TBSs, respectively. The indoor air temperature fluc-tuated with the outlet air temperature of the conventional airconditioners in control TBS. In contrast, the indoor air temperaturewas stable in the retrofit TBS during the operation of the LHSU. Themean indoor air temperatures were 22.2 �C and 24.6 �C for controland retrofit TBSs, respectively. In other words, part of the energy

16:00 20:00 00:00 04:00 08:00 12:00 16:0018

19

20

21

22

23

24

25

Indo

or a

ir te

mpe

ratu

re(o C)

Retrofit TBS Control TBS

Time

Fig. 8. Indoor air temperatures inside TBSs with LHSU and/or conventional airconditioners.

5 10 15 20 25 30 355

10

15

20

25

30

35

Mode IIIMode IIMode I

TBS indoor air LHSU outlet air LHSU inlet air

Tem

pera

ture

(o C)

LHSU inlet air temperature ( oC)

ShenyangOct.20-Oct.24

ZhengzhouNov.09-Nov.13

KunmingApr.10-Apr.17

GuangzhouFeb.25-Mar.02

Changsha Jul.14-Aug.09

Fig. 9. TBS indoor air and LHSU inlet and outlet air temperatures.

-5-15 -10 0 5 10 15 20 25 300

20

40

60

80

100ShenyangZhengzhou

Kunming Guangzhou Changsha

Ener

gy s

avin

gs ra

tio (%

)

Monthly mean temperature (oC)

Fig. 11. Variation of ESRs as a function of monthly mean temperature.

332 X. Sun et al. / Applied Energy 147 (2015) 325–334

savings achieved from using LHSU was brought by raising theindoor air temperature in the TBS, but still under the desired rangeof specified temperatures. However, it was noted that in othercases the indoor air in retrofit TBS may be colder than it in controlTBS when outside air temperature was lower.

Compared with the simulated results in Fig. 6, it was noted thatthe indoor air temperature fluctuated when the conventional airconditioners were operating during the experiments. The intermit-tent running of air conditioners was considered when calculatingenergy consumption; however, its influence on indoor air tempera-ture was ignored. The indoor air temperature was assumed to be26 �C during the simulations. It was found that the use of LHSUsin TBS decrease indoor air temperature fluctuations as well asenergy consumption.

4.2. Measured temperatures in the five selected cities

Fig. 9 depicts parts of the experimental data from the fiveselected cities during the tests, especially those that representeddifferent LHSU inlet and outlet air temperatures, TBS indoor airtemperature, and running modes. The actual operation of theLHSU switched between the three modes, as shown in Figs. 6–8.

The unit operated in Mode I in the cities of Shenyang andZhengzhou. It appeared that most of the PCM was in the liquidform for Shenyang, which was at the beginning of the energycharging mode; while in Zhengzhou most of the PCM was in the

Shenyang Zhengzhou Kunming Changsha Guangzhou0

15

30

45

60

75

90

Locations

ESR

(%)

ExperimentalCalculated=14.86%

=12.62%

=6.54%

=15.35%

=11.07%

Fig. 10. Comparisons between experimental and simulated ESRs.

solid form, which was at the end of the energy charging mode.The temperature difference between LHSU inlet and outlet airdecreased with the PCM’s solidification [19]. In addition, the out-side air temperature was higher for Zhengzhou, thus also indicat-ing the differences between LHSU inlet and outlet airtemperatures in these two cities. The LHSU outlet air temperatureswere about 3.5 �C and 1.5 �C higher than inlet air temperatures inShenyang and Zhengzhou, respectively. In these cases cold energywas being stored in the PCM. The indoor air temperatures were 4.5and 5 �C higher than the LHSU outlet air temperatures in these twocities. This suggested that the indoor heat was being removed, inpart, by the LHSU. For Kunming, the LHSU outlet and inlet air tem-peratures were the same. This happened when the LHSU operatedin Mode II. The data for Guangzhou indicated that in this locationand during this time of year the LHSU would operate either inMode II or Mode III and/or in a combination of both. This isexplained by the fact that during the test period, the outside airtemperature was near 25 �C, which was the switching pointbetween Mode II and Mode III. All three temperatures had col-lapsed to this temperature, which was when the LHSU operatedin Mode III. The reason for this was that when the average indoorair temperature was 25 �C, the indoor air was not hot enough todrive the heat transfer between the air and the water in the aircooler and the water and the PCM in the energy storage modules.With the increase in indoor air temperature, the heat transfer ratesin both the air cooler and the energy storage modules increaseduntil they reached thermal equilibrium. In Changsha, the LHSUoperated in Mode III (energy discharging). In this case, the LHSUoutlet air temperature was below its inlet air temperature.

4.3. Comparisons between simulated and experimental energy savingsratios

Fig. 10 shows the comparisons between experimental andsimulated values of energy savings ratio for the five cities. Theaverage simulated and experimental energy savings ratios for thefive cities were approximately 47% and 53%, respectively. The aver-age error in value for the different locations was approximately

Table 5Constants used in Eq. (20).

Outdoor air temperature(�C)

a0 b0 c0 D0

T 6 15 88.05 �241.66 1.43 2.81T > 15 2.05 1.35 � 105 1.28 � 10�2 1.57 � 103

Table 6Average energy savings ratios and space cooling equipment running times (including LHSU) for the five cities.

Condition Shenyang Zhengzhou Kunming Guangzhou Changsha

ESR (%) Time (h) ESR (%) Time (h) ESR (%) Time (h) ESR (%) Time (h) ESR (%) Time (h)

Energy charging 50 4150 46 4988 61 7068 63 3132 51 5164Fresh air 76 1251 76 1529 94 1492 77 2506 76 1456Energy discharging 76 83 76 88 76 92 76 55 76 44Avg. ESR and total running time of LHSU 50 5484 44 6605 67 8652 45 5693 43 6664Running time of existing air conditioners – 712 – 1427 – 8 – 3012 – 2041Percent time covered by the LHSU (%) – 89 – 82 – �100 – 65 – 77

Table 7Pump running times and percentages for the energy charging process for the five cities.

City Shenyang Zhengzhou Kunming Guangzhou Changsha

Running time (h) 370 252 381 342 520Percent time in the energy charging mode (%) 8.9 5.1 5.4 10.9 10.1

X. Sun et al. / Applied Energy 147 (2015) 325–334 333

12%. The errors originated from the different amounts of heat thatwas generated by electronic equipment and outdoor air tempera-tures used in simulations and experiments. The simulated andactual space cooling loads of the TBSs differed, which was theresult of the different operations of equipment used in variousTBSs. Moreover, the hour-by-hour temperatures used in the sim-ulations were synthesized in contrast with real-time values of airtemperature and relative humidity used in the experiments. Thelargest error was 15.4% which corresponded to Changsha. The testperiod in Changsha was between July 14th and August 9th, whichwas when the average outside air temperature was around 25 �C atnight and 30 �C in the daytime, the switching points betweenModes II and III. Consequently, there were more mode changesbetween fresh air and energy discharging processes in the actualrunning than in the simulations. In addition, by the definition ofrelative error, larger errors tend to occur when measured andsimulated values are relatively low, as was the case for Changsha.

5. Simulations of annual energy savings ratios

The annual energy savings ratio was simulated using synthe-sized hour-by-hour temperature for one entire year for the fiveselected cities for the purpose of evaluating the energy savingspotential of this unit. The PCM was assumed at liquid state at thebeginning of simulations.

Fig. 11 shows the variations of energy savings ratios (ESR) as afunction of monthly mean temperature for the five cities. The ESRincreased as the monthly mean temperature increased to the pointwhen the outside air temperature was about 8 �C. This was becauseboth the conventional air conditioners and LHSU operated at par-tial loads. The air conditioners would stop running when theindoor air temperature was below 26 �C; while the LHSU wouldstop running when the indoor air temperature was below 18 �C.By way of the definition of energy savings ratio, a lower value ofenergy consumption of the conventional air conditioners wouldyield a lower value. With increasing outside air temperature, theenergy consumption of air conditioners increased; however, theenergy consumption of LHSU remained constant. Consequently,the energy savings ratio increased. Afterward, it reached approxi-mately 88% and decreased when the outdoor air temperaturewas above 20 �C. When outside air temperature was between 20and 30 �C, LHSU operated mainly under Modes II and III duringshort period; and as a result, the potential of the LHSU to saveenergy decreased. The non-linear regression equation that repre-sented the energy savings ratio was given by

ESR ¼ a0 þ b0

c0 � 1:6T þ d0

� �� 100% ð20Þ

where a0, b0, c0 and d0 were constants. Their values are shown inTable 5 and where T represented the outside air temperature. Thenonlinear relationship had an R2 value of 0.982.

Table 6 shows the average energy savings ratios and runningtimes of LHSU and conventional air conditioners in one entire yearfor the five cities. During the energy charging process (Mode I), thepump stopped running when the PCM in the energy storage mod-ules had completely changed phase from liquid to solid. Pump run-ning times and corresponding running time percentages whenoperating in the energy charging process for the five cities areshown in Table 7. The order of pump running times for the regionsmentioned above did not follow a strict pattern from severe coldregion to hot summer and warm winter region because of theinvolvement of other factors, such as, duration of seasons, tem-perature swings between colder and hotter parts of the day, windpatterns, and proximity to large bodies of water. The pump percentrunning time was approximately 8% when the unit operated in theenergy charging process. This indicated that potentially moreenergy could be stored had the amount of PCM in the energy stor-age modules been greater. The LHSU running time under theenergy discharging process was considerably short, which trans-lated to only about 1% of the entire running time of the space cool-ing equipment in one year. This was, in part, the reason why thetemperature difference between daytime and nighttime hoursdid not influence the energy savings ratios. This suggested thatthe amount of PCM should be increased to further shorten the run-ning time of existing air conditioner while still satisfying the peakspace cooling demand. The percent increase in the amount of PCMwould depend on the amount of outside cold air and cooling loadin TBSs. It should be noted, however, that more PCM will increasethe capital cost of this unit [28]. Although the running time underthe energy release mode was shorter, the percent time covered bythe LHSU was up to 100% (largest) in Kunming and 65% (lowest) inGuangzhou.

6. Conclusions

A free air cooling system combining phase change materials anda natural cold source, used to reduce the space cooling energy con-sumption in telecommunications base stations, was evaluated. Atheoretical model was developed to assess the application of thistechnology in different climatic regions. Based on the model, a

334 X. Sun et al. / Applied Energy 147 (2015) 325–334

physical prototype, called latent heat storage unit (LHSU) was builtand tested in stations located in five cities in different climaticregions in southwest and eastern China. The LHSU was able toreduce the space cooling energy consumed in these stations bysubstantially shortening the running time of conventional air con-ditioners. The following conclusions were drawn:

(1) The largest annual energy savings ratio produced by theLHSU was 67% in Kunming. The average annual value acrossthe five cities was close to 50%.

(2) The average annual percent running time of the LHSU was82.6%, with values surpassing 75% in four of the five citiesand one city (Kunming) with a value of almost 100%. It isnoted that since only one system operated at a time, theLSHU permitted substantially shorter running times of theconventional air conditioners. That is, on average the con-ventional air conditioners ran during 17.4% of the time.

(3) The LHSU displayed a best performance in mild climaticregion, followed by severe cold region, hot summer andwarm winter region, cold region and hot summer and coldwinter region.

(4) Because of the low electric demand drawn by the LHSU,these units could be powered by portable generators duringTBS maintenance or in TBSs located in areas not yet reachedby the electric network.

Acknowledgements

The research work of this paper was supported by theInternational Cooperation (2015DFA61170), National 863 Project(2012AA052503), Project in Hunan Province (2013WK2011,K1403142-11) and interdisciplinary project in Hunan University.

References

[1] Lo C. China’s mobile subscribers up 1.2 pct at 1.15 bln in March. Reuters; April26, 2013. <http://www.reuters.com>.

[2] Schaefer C, Weber C, Voss A. Energy usage of mobile telephone services inGermany. Energy 2003;28:411–20.

[3] Lubritto C, Petraglia A, Vetromile C, Curcuruto S, Logorelli M, Marsico G, et al.Energy and environmental aspects of mobile communication systems. Energy2011;36:1109–14.

[4] Tu R, Liu X, Li Z, Jiang Y. Energy performance analysis on telecommunicationbase station. Energy Build 2011;43:315–25.

[5] Sorrentino M, Rizzo G, Genova F, Gaspardone M. A model for simulation andoptimal energy management of telecom switching plants. Appl Energy2010;87:259–67.

[6] Chen Y, Zhang Y, Meng Q. Study of ventilation cooling technology fortelecommunication base stations: control strategy and application strategy.Energy Build 2012;50:212–8.

[7] Oró E, Gracia A, Castell A, Farid M, Cabeza L. Review on phase change materials(PCMs) for cold thermal energy storage applications. Appl Energy2012;99:513–33.

[8] Kuznik F, Virgone J. Experimental assessment of a phase change material forwall building use. Appl Energy 2009;86:2038–46.

[9] Zhou D, Zhao C, Tian Y. Review on thermal energy storage with phase changematerials (PCMs) in building applications. Appl Energy 2012;92:593–605.

[10] Halawa E, Saman W. Thermal performance analysis of a phase change thermalstorage unit for space heating. Renew Energy 2011;36:259–64.

[11] Gracia AD, David D, Castell A, Cabeza LF, Virgone J. A correlation of theconvective heat transfer coefficient between an air flow and a phase changematerial plate. Appl Therm Eng 2013;51:1245–54.

[12] Mahmoud S, Tang A, Toh C, Al-Dadah R, Soo S. Experimental investigation ofinserts configurations and PCM type on the thermal performance of PCM basedheat sinks. Appl Energy 2013;112:1349–56.

[13] Darkwa J, Su O. Thermal simulation of composite high conductivity laminatedmicroencapsulated phase change material (MEPCM) board. Appl Energy2012;95:246–52.

[14] Tay N, Belusko M, Bruno F. Experimental investigation of tubes in a phasechange thermal energy storage system. Appl Energy 2012;90:288–97.

[15] Tay N, Belusko M, Bruno F. An effectiveness-NTU technique for characterizingtube-in-tank phase change thermal energy storage systems. Appl Energy2012;91:309–19.

[16] Amin N, Bruno F, Belusko M. Effectiveness-NTU correlation for lowtemperature PCM encapsulated in sphere. Appl Energy 2012;93:549–55.

[17] Longeon M, Soupart A, Fourmigué J, Bruch A, Marty P. Experimental andnumerical study of annular PCM storage in the presence of natural convection.Appl Energy 2013;112:175–84.

[18] Liao S, Liao X, Liao Y. A latent heat storage unit. Chinese PatentZL201010124208.6, issued May 23, 2012; 2005.

[19] Sun X, Zhang Q, Medina MA, Liu Y, Liao S. A study on the use of phase changematerials (PCMs) in combination with a natural cold source for space coolingin telecommunication base stations (TBSs) in China. Appl Energy2014;117:95–103.

[20] Zhong Z, Ding T. Design and maintenance of air conditioning system in datacenters. Beijing: Post & Telecom Press; 2009 [in Chinese].

[21] DXJS1006-2005. Power, air conditioning, environmental design specificationfor data center room of China Telecom. Beijing, China; 2005 [in Chinese].

[22] Dai J, Das D, Pecht M. Prognostics-based risk mitigation for telecom equipmentunder free air cooling conditions. Appl Energy 2012;99:423–9.

[23] Dai J, Das D, Ohadi M, Pecht M. Reliability risk mitigation of free air coolingthrough prognostics and health management. Appl Energy 2013;111:104–12.

[24] Zhang Y. Phase change energy storage-theory and application. Hefei: Press ofUniversity of Science and Technology of China; 1996 [in Chinese].

[25] Winterton R. Where did the Dittus and Boelter equation come from? Int J HeatMass Transf 1998;41:809–10.

[26] Meteorological information room, meteorological information center,meteorological administration, China, Department of architecture technologyscience in Tsinghua university. Building thermal environment analysis specialmeteorological data in China. Beijing: China Architecture & Building Press;2005 [in Chinese].

[27] � DesignBuilder Software, Ltd., DesignBuilder 1.2 User Manual; 2006.[28] He B, Setterwall F. Technical grade paraffin waxes as phase change materials

for cool thermal storage and cool storage systems capital cost estimation.Energy Convers Manage 2002;43:1709–23.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具