lable at ScienceDirect

Energy 150 (2018) 591e600

Contents lists avai

Energy

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

Investigation on the charging process of a multi-PCM latent heatthermal energy storage unit for use in conventional air-conditioningsystems

Xiao-Yan Li*, Liu Yang, Xue-Lei Wang, Xin-Yue Miao, Yu Yao, Qiu-Qiu QiangSchool of Energy and Building Engineering, Harbin University of Commerce, No. 1 Xuehai Street, Songbei District Harbin, Heilongjiang 150028, China

a r t i c l e i n f o

Article history:Received 27 August 2017Received in revised form19 February 2018Accepted 20 February 2018Available online 3 March 2018

Keywords:Conventional air-conditioning systemThermal energy storageMultiple phase change materialsNumerical simulationCharging characteristics

* Corresponding author. Tel.: þ86 18645042691.E-mail address: mylxy6168@sina.com (X.-Y. Li).

https://doi.org/10.1016/j.energy.2018.02.1070360-5442/© 2018 Elsevier Ltd. All rights reserved.

a b s t r a c t

In order to improve the performance of thermal energy storage (TES) systems, a multiple phase changematerial (multi-PCM) based TES unit for use in conventional air-conditioning systems was studied. ThreePCMs (PCM-1, PCM-2, and PCM-3) with phase change temperatures of 5.3 �C, 6.5 �C and 10 �C, respec-tively, were used. Water was used as the heat transfer fluid (HTF). A three-dimensional model (3D) wasdeveloped in ANSYS FLUENT to investigate the charging process of multi-PCM TES unit. In order tovalidate the model, an experimental system was set up. The effect of volume ratio of multi-PCM, HTFinlet temperature and flow rate on the charging process of TES unit was investigated. The simulationresults indicate that TES unit using multi-PCM with volume ratio 1:2:3 intensify the charging process incomparison with using single-PCM. The total charging capacity of multi-PCM TES unit with 1:2:3 pro-portions was 3637.2 kJ and increased by approximately 32.22% as compared to the single-PCM. For theHTF flow rate of 0.3 kg/s, decreasing the inlet temperature of HTF sped up the charging capacity, andobviously shortened the complete charging time of TES unit. However, the HTF inlet temperature did notappreciably change the total charging capacity.

© 2018 Elsevier Ltd. All rights reserved.

1. Introduction

The use of conventional central air-conditioning systems hascontinued to increase in recent years, which leads to high electricalcosts. Air-conditioning systems using TES seem to be the rightsolution to this problem. Compared with the sensible heat TES, theuse of ice as the PCM in latent heat thermal energy storage (LTES)has high energy storage density and isothermal phase transition(small temperature swing) [1e3]. Therefore, ice-based latent heatthermal energy storage is considered more efficient [4]. PCMs storeenergy during the off-peak electrical load periods, and release itduring the on-peak electrical load periods, which shifts the on-peak demand to off-peak periods, thus, resulting in the use ofcheaper electric power [5e7]. However, the phase change tem-perature of ice is low (0 �C), and the coefficient of performance(COP) of a refrigerator decreases with the lowering of PCM's phasechange temperature. Therefore, there is an urgent need to identifyan appropriate PCM, whose phase change temperature is higher

than the ice. Li et al. [8] investigated a kind of PCM for use inconventional central air-conditioning systems, in which, the stor-age tank was composed of spherical capsules filled with PCM. Themelting temperature of the PCM was 8.5 �C, which increased theCOP and reduced the costs of such a system. However, the latentheat and the thermal conductivity of PCM used in conventionalair-conditioning systems are small. A decrease in latent heat andthermal conductivity leads to a decrease in the charging/discharging capacity and the rate of charging/discharging, respec-tively [9e11]. The charging/discharging capacity represents thecold storage/release capacity of the TES unit. To overcome thisproblem, air conditioning using multi-PCM TES has been proposed.In recent years, several studies have investigated the advantages ofusing multi-PCM instead of single-PCM [12e14]. Aldoss et al. [15]investigated the performance of a latent thermal energy storagesystem using a single-PCM design and a multi-PCM design of twoand three stages. The results showed that the multi-PCM designattained higher performance both in charging and discharging cy-cles. Peiro et al. [16] provided an experimental evaluation of theadvantages of using multi-PCM in TES systems. The results showedthat the multi-PCM configurations could obtain higher heat

Nomenclature

AbbreviationHTF heat transfer fluidPCM phase change materialmulti-PCM multiple phase change materialTES thermal energy storageLTES latent heat thermal energy storageDSC differential scanning calorimeterCOP coefficient of performance

SymbolsT temperature (�C)k thermal conductivity (W/m �C)L PCM latent heat (kJ/kg)cp constant-pressure specific heat (kJ/kg �C)t time (s)v velocity (m/s)H enthalpy (kJ)

h sensible enthalpy (kJ)p pressure (N/m2)Amush mushy zone constantg gravity vector (m/s2)m mass (kg)Q charging capacity of the TES unit (kJ)W flow rate (kg/s)

Greek symbolsr density (kg/m3)f liquid fractionm dynamic viscosity coefficient (N/m2s)

Subscriptsl liquids solidref reference valueini initialfin finalw wall

X.-Y. Li et al. / Energy 150 (2018) 591e600592

transfer rate, and achieved an effectiveness enhancement of 19.36%,when compared with the single PCM configuration. However, thecharging/discharging performance of the multi-PCM TES systemhas not been further discussed. The operating parameters to opti-mize the thermal performance of multi-PCM TES system needfurther study [15,16]. To examine the performance of multi-PCMTES system, Adine et al. [17] developed a mathematical model topredict the thermal performance of latent heat storage unit thatconsisted of a shell-and-tube and used two phase change materials.They found that with the increase in HTF inlet temperature andmass flow rate, the melting required shorter time, whereas theshortest melting time corresponded to an optimum length of0.33m (L1¼0.33m). Furthermore, several simulations have beenconducted to study the effects of operating and geometric param-eters on the thermal performance of heat storage units. Mosaffaet al. [18,19] presented a 2-dimensional (2D) numerical analysis ofthe performance of a free cooling system using multi-PCM design,and investigated the effect of design parameters on the storageperformance. The results showed that higher inlet air temperatureand air flow rate increased the heat transfer rate, and shortened thecharging time. Wang et al. [20] established a 2D mathematicalmodel to investigate the charging behavior of a zigzag configura-tion heat exchange device containing a multi-PCM. The resultsindicated a larger phase change temperature difference betweenthe multi-PCM, which remarkably improved the charging process.The larger the melting temperature differences between multiplePCMs, the more PCM are melted in the LTES unit. The modelingresults also indicated that, for a given input power, an optimal fluidvelocity existed for obtaining a high rate of melting. According tothe presented literature, many researchers have investigated thethermal performance of TES system using multi-PCM by 2D nu-merical analysis, and showed the importance of improving thecharging/discharging processes of multi-PCM TES systems. How-ever, the complete charging time of amulti-PCM TES system cannotbe shown by 2D numerical analysis, thus making the analysis lessaccurate. Compared with the 2D numerical analysis, the 3D nu-merical analysis is more suitable to predict the complete chargingtime and temperature distribution in the radial directionwithin theheat transfer unit [21e23]. However, the 3D numerical analysis forthe dynamic characteristics of TES systemwith multi-PCM is rarelyreported in literature. Especially, a 3D numerical analysis on the

charging process of a TES unit consisting of heat transfer tubes withmulti-PCM for use in conventional air-conditioning systems has notbeen investigated. In addition, a multi-PCM with higher latent heatand higher phase change temperature to improve the charging/discharging capacity of TES unit has not been provided. Therefore,the present computational study was undertaken.

In the first part of the paper, an inorganic PCMwith higher latentheat and an organic PCM (i.e., PCM-1 and PCM-2, respectively) aredeveloped in the laboratory. In addition, an organic PCM-3 isselected. Subsequently, a TES unit consisting of heat transfer tubesand filled with the three kinds of PCMs (PCM-1, PCM-2, and PCM-3)is studied. A 3D numerical model is developed using ANSYS FLUENTto simulate the charging process of multi-PCM TES unit. The modelis verified through experiments. Finally, an optimum proportionbetween multi-PCM is identified to obtain the maximum thermalenergy charging rate in the heat transfer tubes. Meanwhile, theeffect of HTF inlet flow rate and inlet temperature on the volumefraction, charging capacity, and charging ratewere also numericallyinvestigated.

2. Thermal performance of the PCM

A novel inorganic PCM-1 (HS-W1) and an organic PCM-2(HS-W2) were developed for use in a multi-PCM TES unit, andtheir thermal performances were investigated. PCM-3 was analready existing material (Paraffin C15). The melting temperatureand the value of latent heat of HS-W1 and HS-W2 were measuredusing differential scanning calorimeter (DSC). While analyzing inDSC, the freezing is delayed due to certain kinetics-related factors.Therefore, the freezing temperatures of HS-W1 and HS-W2 weremeasured using a cooling curve method.

The result of DSC measurement is a curve of heat flux versustemperature. Figs. 1 and 2 show the DSC curves for HS-W1 andHS-W2, respectively. The curve can be used to calculate the latentheat by integrating the peak area, while the melting point isconsidered to be the peak temperature. Fig. 1 shows that themelting temperature and the phase-change latent heat of HS-W1were 5.3 �C and 271.2 kJ/kg. Fig. 2 shows that the melting temper-ature and the phase-change latent heat of HS-W2 were 6.5 �C and226.2 kJ/kg, respectively. The freezing temperatures of HS-W1 andHS-W2, which were measured using the cooling curve method,

Fig. 1. DSC curves of PCM-1 HS-W1.

Fig. 2. DSC curves of PCM-2 HS-W2.

X.-Y. Li et al. / Energy 150 (2018) 591e600 593

were 4.2 �C and 6 �C, respectively. In addition, the thermal con-ductivity and specific heat of HS-W1 and HS-W2 were measuredusing a non-stable method (transient plane source method) [24]and DSC dynamic measurement method [25], respectively. In themeasurement of thermal conductivity, a constant electric power issupplied to increase the temperature of PCM. The relationship be-tween the supplied energy and temperature increase can be used tocalculate the thermal conductivity of PCM. To measure the specificheat capacity of PCM, the baseline heat flow curve of empty cruciblehad acquired. Then the sapphire and PCMweremeasured, and aftersubtracting the baseline curve, two DSC signals are obtained. Byanalyzing the DSC signals, the specific heat capacity of PCM can becalculated. Furthermore, Density Balance Precisa/XB220Awas usedto measure the density of HS-W1 and HS-W2. Table 1 lists thethermos-physical properties of the three PCMs.

Table 1Thermo-physical properties of the three PCMs.

PCM-1(HS-W1)

PCM-2(HS-W2)

PCM-3(Paraffin C15)

Melting temperature (ºC) 5.3 6.5 10Heat of fusion (kJ/kg) 271.2 226.2 205Thermal conductivity[s] (W/m ºC) 1.318 0.51 0.48Thermal conductivity[l] (W/m ºC) 0.762 0.29 0.27Specific heat capacity [s] (kJ/kg ºC) 1.362 2.113 2.4Specific heat capacity [l] (kJ/kg ºC) 1.175 1.594 1.8Density[s] (kg/m3) 1645.9 851.4 860Density[l] (kg/m3) 1120 774 774.65

3. Physical and mathematical models

3.1. Physical model

The TES unit shown in Fig. 3 is composed of 36 heat transfertubes, each containing the three PCMs (PCM-1, PCM-2, and PCM-3)with different phase change points. The cold HTF enters throughthe inlet into the TES unit, exchanging heat with the PCM inside thetubes, and then flows out of the TES unit. The PCM with higherphase change temperature is located at the top in the tube.Therefore, a relatively high temperature difference between PCMand HTF can be maintained to ensure a high heat transfer rateduring the charging process.

3.2. Mathematical model

In this study, a 3D model has been established to simulate theheat transfer in heat transfer tubes of the TES unit, which are filledwith multi-PCM. The finite-volume formulation was solved usingNavier-Stokes equations. The solidification and melting modelwas based on enthalpy method, and used to simulate the solidi-fication process. A mushy freeze/melt zone existed between theupper melting temperature and the lower freezing temperature.Once the temperature of PCM dropped below the freezing tem-perature, the phase change was considered to have completed.The laminar viscous model was introduced to simulate thestratification.

The governing equations for the present model are based on thefollowing assumptions:

(1) The flow of HTF is steady, incompressible, and fullydeveloped.

(2) The heat losses to surroundings are negligible.(3) The phase segregation, subcooling, and hysteretic phenom-

ena are neglected.(4) The thermo-physical properties of PCM and HTF are inde-

pendent of temperature.(5) The thermo-physical properties of multi-PCM are equal in

solid and liquid phases, except for their phase changetemperature.

(6) Natural convection in the liquid PCM is neglected. In naturalconvection, the flow intensity can be determined throughRayleigh number. For the condition (small size of geome-try), the Rayleigh number is small, and therefore, it isreasonable to assume that the natural convection isinsignificant [26].

3.2.1. Governing equations

3.2.1.1. Mass conservation equation. The equation for conservationof mass, or continuity equation, can be written as follows:

vr

vtþ Vðr n!Þ ¼ 0 (1)

where r is the density, t is the time and n! is the velocity.The Reynolds number for the HTF should be calculated as:

Re ¼ rvlm

(2)

where m is the dynamic viscosity, and l is the characteristic length ofthe shell-tube tank, which is equivalent to hydraulic diameter D.The hydraulic diameter is defined as four times of the hydraulicradius.

Y-Z plane, X = 120 X-Y plane, Z = 0

Fig. 3. Physical model of the TES unit.

X.-Y. Li et al. / Energy 150 (2018) 591e600594

The hydraulic radius is formulated as below:

D ¼ 4AP

(3)

where A is the cross-section area of the TES, and P is the wettedperimeter of the cross-section. The hydraulic diameter D is calcu-lated to be 0.0296m for the proposed system.

3.2.1.2. Energy conservation equation. The energy conservationequation is:

v

vtðrHÞ þ Vðr n!HÞ ¼ V$ðkVTÞ þ Se (4)

where H is the enthalpy, k is the thermal conductivity, and Se is thesource term.

The source term,

Se ¼ r

cp

vðfLÞvt

(5)

Enthalpy Hshould be calculated as:

H ¼ hþ DH (6)

where h is the sensible enthalpy, and it can be defined as:

h ¼ href þZT

Tref

cpdT (7)

DH ¼ ð1� f ÞL (8)

where href is the reference enthalpy, Cp is the specific heat, Tref isthe reference temperature, L is the latent heat and f is the liquidfraction. DH is the latent heat enthalpy.

The liquid fractionf is introduced as follows:

f ¼ 0 T < Ts

f ¼ 1 T > TL

f ¼ T � TsTL � Ts

Ts < T < TL

(9)

3.2.1.3. Momentum conservation equation. The forces acting on thefluid were considered to be pressure, viscosity, gravitational, andmomentum forces. The momentum source was defined as:

Fi ¼ð1� f Þ2�f 3 þ ε

� n!Amush (10)

where ε is a constant (0.001), which is used to avoid division byzero, and Amush is the mushy zone constant (1� 105) [27].

Considering the forces mentioned above, the momentumequation is expressed as:

v

vtðruiÞ þ V$ðr n! n!Þ ¼ �Vpþ V$

hm�V n!þ V n!T

� iþ p g!þ Fi

(11)

where t is the time, pis the pressure, m is the dynamic viscosity, r g!is the term representing gravitational force, and Fi is the termrepresenting momentum force.

3.2.1.4. Cooling charging capacity of TES unit. The charging capacityof multi-PCM in heat transfer tubes of the TES unit can bewritten asfollows:

Qi ¼ZTini

Ts;i

miCp;l;idT þmi,Li þZTs;i

Tfin

miCp;s;idT (12)

Fig. 4. Schematic diagram of the experimental apparatus. 1. Compressor; 2. Condenser;3. Receiver; 4. Filter; 5. Expansion valve; 6. Evaporator; 7. Pressure gauge; 8. Watertank; 9. Water pump; 10. Flow meter; 11. Electric heater; 12. SCR temperaturecontroller; 13. Temperature sensor; 14. Frequency converter; 15. Temperaturecontroller; 16. Storage tank; 17. PCM heat transfer tube; 18. Thermocouple extensionwire; 19. Data acquisition system; 20. Computer; 21. Electric heater.

Fig. 5. Photograph of the lab-scale TES system.

X.-Y. Li et al. / Energy 150 (2018) 591e600 595

Q ¼X3i¼1

Qi (13)

where Tini is initial temperature of PCM, Ts;i is the freezing tem-perature of PCM-i in the heat transfer tube, Tfin is the PCM's finaltemperature, Cp;l;i and Cp;s;i are the specific heat of liquid PCM-i andspecific heat of solid PCM-i in the heat transfer tube, respectively, Liis the latent heat of PCM-i in the heat transfer tube, and mi is themass of PCM-i in the TES unit.

The charging rate of TES unit can be calculated as follows:

q ¼ Q=t (14)

where Q is the charging capacity of TES unit, and t is the chargingtime.

3.2.2. Boundary and initial conditionsThe boundary and initial conditions must be considered to solve

the mathematical formulation. The following conditions regardingthe initial and boundary conditions. At the initial stage of chargingprocess, the liquid fraction is 1, while the temperatures of the threePCMs and HTF in the TES unit are set to be 12 �C. The heat flux onthe wall is considered to be zero, and is used as the boundarycondition. This condition can be justified as the wall of TES unit isconsidered to be adiabatic.

The inlet and outlet of HTF are prescribed at mass-flow inlet andoutflow. To investigate the influence of HTF inlet flow rate onmulti-PCM solidification process, the simulation is carried out at fourdifferent HTF mass flow rates of 0.1 kg/s, 0.3 kg/s, 0.5 kg/s, and0.7 kg/s. Based on equation 2, the Reynolds number of four massflows 0.1 kg/s, 0.3 kg/s, 0.5 kg/s, and 0.7 kg/s are 54, 163, 272, and380, respectively. These results show that the flow of HTF islaminar.

3.2.3. Mesh evaluationIn this study, uniform unstructured mesh with tetrahedral cells

was used. The mesh was modeled using the commercial pre-processor GAMBIT. The mesh independence was checked bycomparing the solidification time of PCM with five different meshsizes (see Table 2). In comparison to the mess size of 2,318,183, thedifferences in solidification times for mesh sizes of 3, 168, 910 and1,786,768 were 1.38% and 1.1%. Meanwhile, results of 1,309,954 and794,171 mesh sizes deviated up to 7.66% and 11.34%, compared to2,318,183 mesh sizes. The solidification time of PCM obviouslyincreased when the mesh sizes were between 1,786,768 and794,171. When the size was larger than 1,786,768, the solidificationtime changed a little. Therefore, a mesh of 2,318,183 cells wasadopted as the mesh size to carry out the numerical simulation. Inaddition, it was ensured that the mesh quality met therequirements.

4. Experiments

To validate the theoretical analyses, an experimental multi-PCMTES systemwas set up. The schematic diagram of the experimentalapparatus is shown in Fig. 4. Furthermore, Fig. 5 shows thephotograph of the test rig used for experiments. The apparatus iscomposed of a refrigeration system, a thermostat system, a

Table 2Solidification times at different mesh sizes.

Mesh size 3,168,910 2,318,183 1,786,768 1,309,954 794,171Solidification time (s) 5743.4 5824 5890 6270.5 6484.7

measurement system, a data acquisition system, and a multi-PCMTES unit. The TES unit is of 640mm height, with a240mm� 240mm cross-sectional area, and consists of 36 encap-sulated heat transfer tubes with multi-PCM. Each tube is of 30mmdiameter and 600mm length. The diffusers were installed at thebottom of tubes so that the HTF flowed uniformly. Thirty-six inletnozzles (having the radius of 5mm) were located at the diffusers.The outlet (with the radius of 10mm) was located at the side of theunit, and the height of the outlet above the lower part of the unitwas 625mm. The test section was coated with 50mm-thick poly-urethane layer.

The temperatures of HTF and PCM in the TES unit weremeasured using a Keithley 2700 data acquisition system with theOMEGA type-T thermocouple wires. The thermocouples in the TESunit were shown in Fig. 4. Tout and Tin were located at the HTF outletand inlet, while T1, T2, and T3 were located at the center of PCM-1,PCM-2, and PCM-3 (0.15m, 0.4m, and 0.55m far from the bottomof tube, respectively). The flow rate of the HTF was measured usingan ultrasonic flow meter. During the charging process, HTF at 2 �Centered the TES unit, exchanged heat with the PCM inside thetubes, and then, flowed out of the TES unit. All measurements werechecked for reproducibility. The reproducibility test showed

120

X.-Y. Li et al. / Energy 150 (2018) 591e600596

deviations were limited, within the range of measurement error ofthe various devices. The measured errors in the temperature andflow rate were ±0.3 �C and ±2%, respectively.

0 1000 2000 3000 4000 50000

30

60

90

0:1:0 1:1:11:2:3 1:3:22:1:3 2:3:13:1:2 3:2:1C

harg

ing

Cap

acit

y(kJ

)

Times(s)

Fig. 7. Charging capacity of different proportions of PCMs (under cooling chargecondition).

5. Results and discussion

5.1. Determination of the optimum proportion of multi-PCM

Before investigating the thermal performance of TES unit con-sisting of heat transfer tubes, which were filled with three PCMs,eight different proportions of multi-PCM were studied numericallyto determine the optimum proportion of the multi-PCM. The vol-ume ratios of PCM-3, PCM-2, and PCM-1 studied were 1:2:3, 1:3:2,2:1:3, 1:1:1, 2:3:1, 3:1:2, 3:2:1 and 0:1:0, respectively. Table 1presents the thermo-physical properties of the three PCMs. Theaverage phase change temperature of the three PCMs in heattransfer tubewas fitted to the conventional central air-conditioningsystems (typical chilled water supply temperature¼ 7 �C). Ac-cording to the minimum temperature limit of the chillers installedfor air-conditioning applications (ca. 1 �C), the temperature had tolie within the range of 6e8 �C [28]. This way, the eight differentproportions of the multi-PCM were studied numerically. Fig. 6shows the schematic view of the different proportions of PCMs inthe heat transfer tubes. The blue color (in the web version) in-dicates the tube section filledwith PCM-3, while the red (in thewebversion) represents the tube section filled with PCM-2. Addition-ally, the green color (in the web version) indicates the tube sectionfilled with PCM-1. Initially the PCM was in liquid state and thetemperature was 12 �C. The heat transfer tube had a constant walltemperature of 2 �C, and the value was used as the boundarycondition.

Fig. 7 shows the effects of different proportions of PCMs on thecharging capacity of all the PCMs in heat transfer tube. The chargingcapacity measures the thermal performance of heat transfer tube,while the slope of curve indicates the rate of charging process.According to the figure, the charging capacity of the heat transfertube using the single PCM-2 (phase change temperature suitablefor air-conditioning systems) is smaller than that of other pro-portions. Moreover, the slope of the curves with volume ratios of

1:1:1 1:2:3 1:3:2 2:1:3

Fig. 6. Schematic view of different PCM p

1:2:3, 1:3:2, 2:1:3, 1:1:1, 2:3:1, 3:1:2, and 3:2:1 were steeper thanthat of the 0:1:0. The steeper the slope of curve, the larger chargingrate of heat transfer tube. These results show that the charging rateof multi-PCM is larger than that of the single-PCM. When theproportions of PCMs were 0:1:0, 1:1:1, 1:2:3, 1:3:2, 2:1:3, 2:3:1,3:1:2 and 3:2:1, the time to reach the total charging capacity were4598 s, 4216.7 s, 4169.1 s, 4351 s, 4178 s, 4258.4 s, 4204 s, and4267 s, respectively. In addition, the total charging capacities weredetermined to be 76.32 kJ, 91.76 kJ, 102.32 kJ, 93.17 kJ, 100.81 kJ,82.62 kJ, 91.37 kJ and 84.28 kJ, respectively. Compared to thecharging capacity of 1:2:3 proportion, the charging capacities of1:1:1, 1:3:2, 2:1:3, 2:3:1, 3:1:2, and 3:2:1 proportions decreased byapproximately 10.32%, 8.94%, 1.47%, 19.25%, 10.7%, and 17.63%,respectively. A direct comparison between the charging capacityand charging rate of the three PCM in the different proportions hasbeen reported in Table 3. The results indicate that the proportion1:2:3 exhibited the largest charging capacity and the charging rate.

2:3:1 3:1:2 3:2:1 0:1:0

roportions in the heat transfer tube.

Table 3The charging capacity and charging rate of different proportions of PCM.

PCM-3:PCM-2:PCM-1 Time (s) Charging capacity (kJ) Charging rate (kJ/h)

0:1:0 4598 76.32 59.751:1:1 4216.7 91.76 78.341:2:3 4169.1 102.32 88.351:3:2 4351 93.17 77.082:1:3 4178 100.81 86.862:3:1 4258.4 82.62 69.843:1:2 4204 91.37 78.243:2:1 4267 84.28 71.11

Fig. 9. Temperature variations of the TES unit at different HTF flow rates.

X.-Y. Li et al. / Energy 150 (2018) 591e600 597

Therefore, the proportion 1:2:3 was selected as the optimum pro-portion, and was used to further investigate the dynamic charac-teristics of multi-PCM TES unit.

The average phase change temperature of the multi-PCM withthe volume ratio of 1:2:3 was 6.4 �C. The temperature was higherthan that of the ice (0 �C). Therefore, applying the multi-PCM to TESunit can increase the evaporation temperature of the chiller, andimprove the COP of conventional air-conditioning systems. Theapplication of multi-PCM TES system reduced the installed capacityof the air-conditioning system, and narrowed the gap between thepeak and off-peak loads of the electricity demand, thus savingoperational cost by shifting the electrical consumption from peakperiods to off-peak periods.

5.2. Effects of different HTF flow rates on charging process

Figs. 8e10 show the effects of HTF flow rate on the chargingprocess (multi-PCM with 1:2:3 proportion), while the HTF inlettemperature was 2 �C. From Fig. 8, the temperature fields of thecross sections with different HTF flow rates (at the charging time of5750 s) can be observed. There was no temperature region above10 �C in the TES tank, indicating that the PCM-3 (with the phasechange temperature of 10 �C) was completely frozen. ComparingPCM-3 sectionwith those of the PCM-1 and PCM-2, the PCM-3 wasthe fastest to completely freeze. This was due to the higher tem-perature difference between the HTF and PCM. During the sametime, a greater flow rate of HTF produced a faster temperature drop

W=0.1 kg/s W=0.3 kg/s

Fig. 8. Temperature profiles of the TES unit

in the TES unit. Fig. 9 shows the temperature variations of Tout, T1, T2,and T3 at different HTF flow rates. Tout were located at the HTFoutlet, while T1, T2, and T3 were located at the center of PCM-1,PCM-2, and PCM-3 (0.15m, 0.4m, and 0.55m far from the bottomof tube, respectively). During the initial period, apparent heat isdisplayed dominantly and the temperature of Tout, T1, T2, and T3decreased at almost the same speed. As time went on, the PCMstarted to freeze, which showed a nearly unchanged temperature ofT1, T2, and T3. Tout temperature decreased slowly. When the phasechange process was complete, the temperature of PCM starteddecreasing again. It also can be seen that, a greater flow rate of HTFproduced a faster temperature drop in the TES unit.

Fig. 10 indicates the multi-PCM liquid volume fraction variationduring the charging process in the TES unit. It can be seen that thecomplete charging time (liquid fraction of 0.02) reduced with theincrease in HTF flow rate. It took 2000 s to freeze the remaining 2%of the PCM. Therefore, when the volume fraction of liquid PCMwasreduced to 0.02, the charging process was considered to havecompleted. When the flow rates of HTF were 0.1, 0.3, 0.5, and 0.7,

W=0.5 kg/s W=0.7 kg/s

at different HTF flow rates (t¼ 5750 s).

Fig. 10. Variation of the liquid volume fraction of PCM at different HTF flow rates.

X.-Y. Li et al. / Energy 150 (2018) 591e600598

the complete charging times were 7656.6 s, 5824 s, 5422 s, and5250 s, respectively. Compared to the HTF flow rate of 0.1 kg/s, thecomplete charging times for the flow rates of 0.3 kg/s, 0.5 kg/s, and0.7 kg/s were shortened by approximately 23.9%, 29.1%, and 31.4%,respectively. The results indicate that the HTF flow rates cannot betoo slow; otherwise, the complete charging time will be too long,lowing energy storage efficiency. However, further increasing theHTF flow rate has little influence on the complete charging time,and would lead to unnecessary energy waste.

5.3. Effects of HTF flow rates on the charging capacity

Fig. 11 shows the effect of HTF flow rate on the charging capacityof the multi-PCM TES unit for the inlet temperature of 2 �C (multi-PCM with 1:2:3 proportion). These results indicate that thecharging capacity dramatically increased over time during theinitial period of charging process. Meanwhile, greater the HTF flowrate, higher is the charging capacity. For example, at 2000 s, byincreasing the HTF flow rate from 0.1 kg/s to 0.7 kg/s, the chargingcapacity increased from 1668 kJ to 2448 kJ, respectively. For theflow rate of 0.7 kg/s, the charging capacity increased by approxi-mately 46.76% compared to that for the flow rate of 0.1 kg/s. Thiswas due to the enhancement of heat transfer between the HTF and

Fig. 11. Charging capacity of the TES unit at different HTF flow rates.

PCMs. However, it can also be seen that there was not much dif-ference in the charging capacity when the HTF flow rate wasgreater than 0.3 kg/s. Moreover, further increasing the HTF flowrate has little influence on the total charging capacity. Therefore,the HTF flow rate of 0.3 kg/s has been selected for further analysis.

5.4. Effects of different HTF inlet temperatures on the volumefraction

The variation in the liquid volume fraction of multi-PCM atdifferent HTF inlet temperatures during the charging process isshown in Fig. 12 (multi-PCM with 1:2:3 proportion). The HTFentered at the flow rate of 0.3 kg/s while the inlet temperatures ofHTF studied were 1 �C, 2 �C, and 3 �C. When the liquid PCM reachedthe phase change temperature, it started to freeze, while the vol-ume fraction of liquid PCM gradually decreased. It took 1252 s,1536.3 s and 2006.4 s to freeze 50% of the PCM for the inlet HTFtemperatures of 1 �C, 2 �C, and 3 �C, respectively. At 3000 s, thecorresponding liquid fractions were 0.155, 0.23, and 0.34, respec-tively. Comparing HTF inlet temperature of 1 �C and 2 �C with thatof 3 �C, the corresponding liquid PCM volume fractions decrease byapproximately 54.41% and 32.35%. It can also be seen that thestorage was completely frozen at 2.7 h. When the inlet tempera-tures of HTF were 1 �C, 2 �C, and 3 �C, the complete charging times(with the liquid fraction of 0.02) were 4613 s, 5824 s, and 8040 s,respectively. The complete charging time for inlet temperatures of1 �C, and 2 �C decreased by approximately 42.62% and 27.56%,respectively, compared to the HTF inlet temperature of 3 �C. Thisindicates that the variations in inlet temperatures of HTF stronglyaffected the volume fraction of liquid PCM.

5.5. Effects of HTF inlet temperature on the charging capacity

Fig. 13 shows the effect of HTF inlet temperature on the chargingcapacity. The HTF flow rate of 0.3 kg/s (multi-PCM with 1:2:3proportion) indicated that the charging capacity increased morerapidly when the inlet temperature was smaller during thecharging process. In addition, the total charging capacity is a steadyvalue during the final period. Furthermore, the slope of the curve(representing the charging rate) increased with the decrease ininlet temperature. Therefore, the charging rate can be greatlyenhanced by decreasing the HTF inlet temperature. However,decreasing the inlet temperature has little influence on the total

Fig. 12. Variation in volume fraction of liquid PCM at different HTF inlet temperatures.

Fig. 13. Charging capacity of TES unit at different HTF inlet temperatures.

X.-Y. Li et al. / Energy 150 (2018) 591e600 599

charging capacity. For the inlet temperature of 2 �C, the chargingcapacity of multi-PCM TES unit with 1:2:3 proportion was largerthan that of the TES unit using a single PCM (PCM-2). From Fig. 12,the total charging capacity of TES unit with multi-PCM and single-PCM-2 were determined to be 3637.2 kJ and 2750.8 kJ, respectively.For the multi-PCM TES unit, the total charging capacity increasedby approximately 32.22% compared to that of the single-PCM-2.Moreover, the difference in time to reach the total charging ca-pacity between the multi-PCM and single-PCM-2 TES units was notobvious. Therefore, the multi-PCM TES unit attains better perfor-mance than the conventional single-PCM in the charging process.In the sensible TES unit, water was used to store the energy and nophase change occurred. When the same volume of water was

Fig. 14. Temperature curves of HTF and P

storied, the total charging capacity was determined to be 625.7 kJ.For the sensible TES unit, the total charging capacity decreased byapproximately 82.79% and 77.25% compared to that of multi-PCMTES unit and single-PCM-2 TES unit. Therefore, considering thesame volume, the latent TES unit can store more energy in com-parison with the sensible type.

5.6. Comparison between the experimental and numerical results

In the experiments, each heat transfer tube contained threePCMs, while the volume ratio of PCM-3, PCM-2 and PCM-1 was1:2:3. During the charging process, the HTF inlet temperature andinlet flow rate were 2 �C and 0.3 kg/s, respectively. The tempera-tures of thermocouples in the TES unit were also measured at thesame time. In order to validate the reliability of numerical model,Tout, T1, T2, and T3 were selected to make a comparison between thenumerical and experimental results. Fig. 14 shows the comparisonof numerical and experimental results. It can be seen that both thenumerical and experimental results have the same trend. However,the experimental results were higher than the numerical results,whereas the deviation between the results was within 6%. Thedifference between the numerical and experimental results mainlycauses by heat loss. As proposed by other author [29], the experi-mental process is not completely adiabatic. However, in the nu-merical simulation process, the heat loss was ignored. On the otherhand, there was fluctuation for the inlet temperature of HTF in theexperiments. The HTF inlet temperaturewas higher than 2 �C. It canalso be seen that the temperature measurements of the PCMdecreased quickly during the initial stage of the charging process,and became flat as the latent heat of the PCM started to be released.However, the comparison of numerical results with the experi-mental showed that both were in relatively good agreement, thusconfirming the validity of the proposed model.

CM for the cooling charge condition.

X.-Y. Li et al. / Energy 150 (2018) 591e600600

6. Conclusions

In this paper, a TES unit consisting of heat transfer tubes, whichwere filled with three kinds of PCMs, was proposed and analyzed.Numerical simulation and experimental investigations were car-ried out to study the performance of the proposed TES unit. Theeffects of volume ratio of multi-PCM, HTF inlet flow rate, HTF inlettemperature, on the performance of multi-PCM TES unit wereinvestigated. The following conclusions can be drawn:

(1) A novel inorganic PCM-1 (HS-W1) and an organic PCM-2(HS-W2) were developed and used in the heat transfertube TES unit. Using the DSC measurements, the phasechange temperatures of PCM-1 and PCM-2 are 5.3 �C and6.5 �C, respectively. The corresponding latent heats were271.2 kJ/kg and 226.2 kJ/kg, respectively. Thus, the PCMs canbe considered efficient cooling storage materials for multi-PCM TES unit.

(2) The simulation results showed the multi-PCM with 1:2:3proportions have the largest charging capacity amongvarious proportions of themulti-PCM. For themulti-PCM TESunit with 1:2:3 proportions, the total charging capacityincreased by approximately 32.22% compared to that of thesingle-PCM-2 TES unit. Therefore, the proportions betweenthe multi-PCM play an important role in the performanceimprovement of the TES unit.

(3) The simulation results showed that the HTF flow rate mustbe controlled within a certain range for practical applica-tions. During the charging process, it may be not necessary torequire a large HTF flow rate. Smaller the flow rate, longer istime interval for complete charging. Furthermore, smallerflow rate will cause lowing energy storage efficiency. How-ever, any further increasing the HTF flow rate has little in-fluence on the complete charging time, and would lead tounnecessary energy waste.

(4) Both the charging capacity and charging rate increased morerapidly when the HTF inlet temperature was lower at aconstant HTF flow rate. The time to reach complete chargingreduced with the decrease in HTF inlet temperature.Compared to 3 �C, the differences of 42.62% and 27.56% wereobserved for the inlet HTF temperatures of 1 �C and 2 �C,respectively. However, the HTF inlet temperature has littleeffect on the total charging capacity.

(5) The model developed in this paper has been validated usingexperimental data. The deviation between the numerical andexperimental results was within 6%. The comparison resultsshow that the heat loss demonstrates a certain influence onthe charging process. However, the result is satisfactoryconfirming the validity of the model.

Acknowledgements

The authors acknowledge the support provided by the NationalNatural Science Foundation of China (51476049).

References

[1] Ezan MA, Erek A, Dincer I. Energy and exergy analyses of an ice-on-coilthermal energy storage system. Energy 2011;36:6375e86.

[2] Medrano M, Yilmaz MO, Nogu�es M, Martorell I. Experimental evaluation ofcommercial heat exchangers for use as PCM thermal storage systems. ApplEnergy 2009;86:2047e55.

[3] Sanaye S, Fardad A, Mostakhdemi M. Thermoeconomic optimization of an icethermal storage system for gas turbine inlet cooling. Energy 2011;36:1057e67.

[4] Chan ALS, Chow TT, Fong SKF, Lin JZ. Performance evaluation of districtcooling plant with ice storage. Energy 2006;31:2750e62.

[5] Li XY, Zhao QQ, Qu DQ. Investigation on the dynamic characteristics of a directcontact thermal energy storage charging process for use in conventionalair-conditioning systems. Appl Therm Eng 2015;91:172e80.

[6] Barzin R, Chen JJJ, Young BR, Farid MM. Peak load shifting with energy storageand price-based control system. Energy 2015;92:505e14.

[7] Sun YJ, Wang SW, Xiao F, Gao D. Peak load shifting control using different coldthermal energy storage facilities in commercial buildings: a review. EnergyConvers Manag 2013;71:101e14.

[8] Li XY, Li L, Yu JW, Liu JQ, Wu YY. Investigation of the dynamic characteristics ofa storage tank discharging process for use in conventional air-conditioningsystem. Sol Energy 2013;96:300e10.

[9] Zhao YJ, Wang RZ, Wang LW, Yu N. Development of highly conductive KNO3/NaNO3, composite for TES (thermal energy storage). Energy 2014;70:272e7.

[10] Li TX, Lee JH, Wang RZ, Kang YT. Enhancement of heat transfer for thermalenergy storage application using stearic acid nanocomposite with multi-walled carbon nanotubes. Energy 2013;55:752e61.

[11] Calvet N, Py X, Oliv�es R, B�ed�ecarrats JP. Enhanced performances of macro-encapsulated phase change materials (PCMs) by intensification of the inter-nal effective thermal conductivity. Energy 2013;55:956e64.

[12] Wang JF, Ouyang YX, Chen GM. Experimental study on charging processes of acylindrical heat storage capsule employing multiple phase change materials.In: International Symposium Air Conditioning in High Rise Buildings; 2001.p. 439e47.

[13] Yang L, Zhang XS, Xu GY. Thermal performance of a solar storage packed bedusing spherical capsules filled with PCM having different melting points.Energy & Build 2014;68:639e46.

[14] Fang M, Chen GM. Effects of different multiple-PCMs on the performance of alatent thermal energy storage system[J]. Appl Therm Eng 2007;27(5e6):994e1000.

[15] Aldoss TK, Rahman MM. Comparison between the single-PCM and multi-PCMthermal energy storage design. Energy Convers Manag 2014;83:79e87.

[16] Peir�o G, Gasia J, Mir�o L, Cabeza LF. Experimental evaluation at pilot plant scaleof multiple-PCMs (cascaded) vs. single PCM configuration for thermal energystorage. Renew Energy 2015;83:729e36.

[17] Adine HA, Qarnia HE. Numerical analysis of the thermal behaviour of a shell-and-tube heat storage unit using phase change materials. Appl Math Model2009;33:2132e44.

[18] Mosaffa AH, Ferreira CAI, Talati F, Rosen MA. Thermal performance of amultiple-PCM thermal storage unit for free cooling. Energy Convers Manag2013;67:1e7.

[19] Mosaffa AH, Farshi LG, Ferreira CAI, Rosen MA. Energy and exergy evaluationof a multiple-PCM thermal storage unit for free cooling applications. RenewEnergy 2014;68:452e8.

[20] Wang P, Wang X, Huang Y, Li C, Peng ZJ, Ding YL. Thermal energy chargingbehaviour of a heat exchange device with a zigzag plate configuration con-taining multi-phase-change-materials (m-PCMs). Appl Energy 2015;142:328e36.

[21] Li XY, Qu DQ, Yang L, Li KD. Experimental and numerical investigation ofdischarging process of direct contact thermal energy storage for use in con-ventional air-conditioning systems. Appl Energy 2017;189:211e20.

[22] Xiao X, Zhang P. Numerical and experimental study of heat transfer charac-teristics of a shell-tube latent heat storage system: Part I e charging process.Energy 2015;79:337e50.

[23] Xiao X, Zhang P. Numerical and experimental study of heat transfer charac-teristics of a shell-tube latent heat storage system: Part II eDischarging pro-cess. Energy 2015;80:177e89.

[24] Johansson P, Kalagasidis AS, Jansson H. Investigating PCM activation usingtransient plane source method. Energy Procedia 2015;78:800e5.

[25] Ferrer G, Barreneche C, Sol�e A, Martorell I, Cabeza LF. New proposed meth-odology for specific heat capacity determination of materials for thermalenergy storage (TES) by DSC. J Energy Storage 2017;11:1e6.

[26] Saman W, Bruno F, Halawa E. Thermal performance of PCM thermal storageunit for a roof integrated solar heating system. Sol Energy 2005;78:341e9.

[27] Brent AD, Voller VR, Reid KJ. Enthalpy-porosity technique for modelingconvection-diffusion phase change: application to the melting of a pure metal.Numer Heat Transf Appl 1988:297e318. part an applications.

[28] Marcello DF, Mauro C, Alberto G. Performance analysis of an innovative PCM-based device for cold storage in the civil air conditioning. Energy Build2016;122:1e10.

[29] Mazzeo D, Oliveti G, Gracia AD, Coma J, Sol�e A, Cabeza LF. Experimentalvalidation of the exact analytical solution to the steady periodic heat transferproblem in a PCM layer. Energy 2017;140:1131e47.

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