Applied Energy 185 (2017) 95–106

Contents lists available at ScienceDirect

Applied Energy

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

PCM thermal storage design in buildings: Experimental studies andapplications to solaria in cold climates

http://dx.doi.org/10.1016/j.apenergy.2016.10.0460306-2619/� 2016 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Dept. of Energy, Information Engineering andMathematical Models, University of Palermo, Viale delle Scienze, Building 9,90128 Palermo, Italy.

E-mail address: guarino@dream.unipa.it (F. Guarino).

Francesco Guarino a,⇑, Andreas Athienitis b, Maurizio Cellura a, Diane Bastien b

aUniversity of Palermo, Dept. of Energy, Information Engineering and Mathematical Models, ItalybConcordia University, Dept. of Building, Civil and Environmental Engineering, Canada

h i g h l i g h t s

� This paper analyzes the performance of a building-integrated thermal storage system.� A wall opposing a glazed surface serves as phase change materials thermal storage.� The study is based on both experimental and simulation studies.� Heat is stored and released up to 6–8 h after solar irradiation.� Yearly heating requirements are reduced by 17% in a cold climate.

a r t i c l e i n f o

Article history:Received 13 July 2016Received in revised form 5 October 2016Accepted 16 October 2016

Keywords:Phase change materialsPCMSolariaBuilding designBuilding simulationExperimental studiesCold climates applications

a b s t r a c t

As energy availability and demand often do not match, thermal energy storage plays a crucial role to takeadvantage of solar radiation in buildings: in particular, latent heat storage via phase-change material isparticularly attractive due to its ability to provide high energy storage density. This paper analyzes theperformance of a building-integrated thermal storage system to increase the energy performances ofsolaria in a cold climate. A wall opposing a highly glazed façade (south oriented) is used as thermal stor-age with phase change materials embedded in the wall. The study is based on both experimental andsimulation studies. The concept considered is particularly suited to retrofits in a solarium since thePCM can be added as layers facing the large window on the vertical wall directly opposite.Results indicate that this PCM thermal storage system is effective during the whole year in a cold cli-

mate. The thermal storage allows solar radiation to be stored and released up to 6–8 h after solar irradi-ation: this has effects on both the reduction of daily temperature swings (up to 10 �C) and heatingrequirements (more than 17% on a yearly base). Coupling of the thermal storage system with natural ven-tilation is important during mid-seasons and summer to improve the PCM charge-discharge cycles and toreduce overheating.Results also show that cooling is less important than heating, reaching up to 20% of the overall annual

energy requirements for the city of Montreal, Canada. Moreover, the phase change temperature range ofthe material used (18–24 �C) is below typical summer temperature levels in solaria, but the increase inthermal capacity of the room alone can reduce annual cooling requirements by up to 50%.

� 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Energy storage is expected to play a key role in moving to a low-carbon electricity system. It can supply more flexibility and balanc-ing to the grid, providing a backup to intermittent renewableenergy; locally, it can improve the management of distribution

networks, reducing costs and improving efficiency. In this way, itcan ease the market introduction of renewables, accelerate thedecarbonisation of the electricity grid, improve the efficiency ofelectricity transmission and distribution (reduce unplanned loopflows, grid congestion, voltage and frequency variations), stabilisemarket prices for electricity, while also ensuring a higher securityof energy supply [1–7].

As solar accessibility and demand often do not match, thermalenergy storage plays a crucial role to take advantage of solarradiation in buildings. Building-integrated thermal energy storage

Nomenclature

i node being modeledi + 1 adjacent node to interior of constructioni � 1 adjacent node to exterior of constructionj + 1 new time stepj time stepDt calculation time stepDx finite difference layer thickness (always less than con-

struction layer thickness)Cp specific heat of materialkw thermal conductivity for interface between i node and

i + 1 node

ke thermal conductivity for interface between i node andi � 1 node

q density of materialCV(RMSE) coefficient of variation of the root mean square errorMBE mean bias errorn number of measuresTm,i monitored temperatureTm;i mean monitored temperatureTs,i simulated temperature

96 F. Guarino et al. / Applied Energy 185 (2017) 95–106

systems [8,9] cover a wide range of materials, techniques anddesigns depending on the applications and aims. They howeverall have in common this underlining concept: being able to storeexcess energy for later use in order to reduce the time mismatchbetween energy availability and demand. Effective utilization ofthermal energy storage for ambient renewable energy (e.g. solarheat for heating and cool outdoor air for free cooling) with properdesign and control has proven promising in reducing peak demandand energy costs associated with space conditioning. Building-integrated thermal energy storage systems have recently attractedsignificant research interest. Savings in room space and material isachievable in comparison with conventional centralized and ther-mally isolated storage systems (e.g. water/ice tanks). One uniquecharacteristic of building-integrated thermal energy storage sys-tems is their thermal coupling with thermal zones due to largeexposed surface areas.

Latent heat storage via phase-change materials (PCMs) [10–16]is particularly attractive due to its ability to provide high energystorage density. Several studies have demonstrated that the useof PCMs in well-insulated buildings can reduce heating and coolingenergy in residential buildings by as much as 25% and obtainsimilar reductions in the peak power required for air conditioning[17–19]. Such applications are of interest since they can have lowerheating and cooling requirements for a given volume than mostsensible systems and may be used in contexts where the applica-tion of standard solutions would be difficult, such as in renovationof historical buildings. On the other hand, daily charge-dischargecycles must be carefully planned.

PCMs represent a potential solution for reducing peak heatingand cooling loads and heating, ventilation, and air conditioning(HVAC) energy consumption in buildings for both new buildingsand retrofits. The use of latent energy storage systems may beone of the solutions to the energy mismatches in Net-Zero EnergyBuildings [20–29] when renewable energy production and buildingenergy demand are out of phase. A building-integrated and dis-tributed thermal storage could shift and reduce part of the loadof residential air conditioners at peak to off-peak periods. As aresult, capital investment in peak power generation equipmentcould be reduced for power utilities and then the savings couldbe passed on to customers. In areas where power utilities are offer-ing time of day rates, building-integrated thermal storage couldenable customers to take advantage of lower utility rates duringoff peak hours.

Literature published on PCMs over the last two decades covers abroad area. The target of this paper is to study the use of wallboardincorporated PCMs to be used passively in high performance build-ings with high window-to-wall ratios in cold climates, particularlyin retrofit applications where the PCM is only applied to thesurface which receives most of the solar radiation. Some relevant

previous experiences on building-integrated PCMs are brieflydescribed below.

Most existing studies focus on the design and optimization ofPCM layers into vertical walls or in the roof mostly located onthe inside walls in various positions.

In [30], Chen et al. propose the modeling of a simple room, aim-ing at determining the best positioning of PCM in the envelope forimproving all-year performance. At the optimal locations, the peakheat flux reductions were 51.3% and 29.7% for the south wall andthe west wall, respectively. The maximum time delays in the peakheat flux were 6.3 h for the south wall and 2.3 h for the west wall.During winter, energy savings in comparison to non-PCM roomscan reach 10%.

In [31] a light envelope test cell was equipped with 25 mm thickPCM on all internal surfaces of a test cell (cubical, around 1 m on alldimensions) to increase the thermal inertia of the envelope andreduce indoor temperature fluctuations. An identical test cell wasbuilt and operated on the same weather conditions but withoutPCM in the envelope. This study showed that PCM allowed a reduc-tion of the indoor temperature range of approximately 20 �C in thetest-cell. The influence of the PCM wall thickness was also studiedthrough numerical simulation, from 10 mm to 35 mm. It was foundthat for a wall thickness higher than 20 mm the indoor tempera-ture variation amplitude would not significantly decrease further.

Kuznik and Virgone [32] carried out an experimental researchin a full-scale test cell under controlled thermal and radiative con-ditions to evaluate the performance of walls with and withoutPCMs during a summer day. In a subsequent study, authors usedthe same PCM composite [13] to show that PCM wallboards reduceair temperature fluctuations in a room and overheating. In order toassess the potential of a PCMwallboard with 60% of microencapsu-lated paraffin within a copolymer (the melting and freezing tem-peratures are 13.6 �C and 23.5 �C respectively), a renovated officebuilding in Lyon was monitored during one year by Kuznik et al.[18]. A room was equipped with PCM wallboards in the lateralwalls and in the ceiling, and another room, identical to the firstone, was not equipped with PCM wallboards but also monitored.The results showed that the PCM wallboards enhance the thermalcomfort of occupants during the whole year.

Evers et al. [33] evaluated the thermal performance of enhancedcellulose insulation with paraffin and hydrated salts for use inframe walls. The thermally enhanced frame walls were heatedand allowed to cool down in a dynamic wall simulator thatreplicated the sun’s exposure in a wall of a building on a typicalsummer day. The results showed that the paraffin-based PCM-enhanced insulation reduced the average peak heat flux by up to9.2% and reduced the average total daily heat flow up to 1.2%.

Other studies focused specifically on PCM application to thefloor. In [34], authors discuss the application of PCM below

F. Guarino et al. / Applied Energy 185 (2017) 95–106 97

different tile, wood and metal floor coverings aiming to absorb thesolar radiation energy in the daytime and release the absorbedheat in the evening. Therefore, in winter, the indoor climate canbe improved and the energy consumption for space heating maybe greatly reduced. It is found that: (1) for the purpose of narrow-ing indoor air temperature swing, for a given position or weathercondition, the suitable melting temperature of PCM is roughlyequal to the average indoor air temperature of sunny winter days;(2) the heat of fusion and the thermal conductivity of PCM shouldbe larger than 120 kJ/kg and 0.5 W/(m K), respectively; (3) thethickness of shape- stabilized PCM plate used under the floorshould not be larger than 20 mm; (4) the performance of PCMscovered by tile or metal floor cover is higher than for a woodenfloor cover; (5) the air-gap between PCM plates and the floorshould be as small as possible.

Passive solar solutions include several different applicationsthat can benefit as well from the use of thermal energy storagein the form of PCM. Many applications are available in literature,some of which are reported below.

In [35] De Gracia et al. investigated the effect of PCM on doubleskin façades. An experimental test was performed with PCM in theair channel of a double skin façade, during the heating season inthe Mediterranean climate. Two identical house-like cubicleslocated in Puigverd de Lleida (Spain) were monitored during win-ter 2012, and in one of them, a ventilated facade with PCM waslocated in the south wall. The experimental results conclude thatthe use of the ventilated facade with PCM improves significantlythe thermal behavior of the whole building.

In [36] Diarce et al. evaluated the thermal performances of anactive façade in Spain that included a phase change material inits outer layer. Experimental results showed that the phase changeprocesses that took place in the façade led to increased heatabsorption in comparison to similar architectural solutions with-out PCM. Simulation results showed that thermal inertia of thewall was higher in the PCM configuration than in all the standardno-PCM alternatives simulated.

Most literature discussed previously focused on a uniform dis-tribution of PCM panels, aiming to create a uniformly more inertialroom. Many studies focused on optimizating the PCM layer posi-tioning among the envelope layers or thickness while othersfocused on the use of PCM in integration with specific existingsolar design solutions. This study proposes PCM energy storageonly on one wall of the indoor environment, directly irradiatedby solar radiation, which is suitable option for retrofit applications.The idea is to maximize the energy storage potential by adoptingseveral layers of PCM facing large glazed openings on the southfaçade. The concept aims at applications mostly in the residentialsector with solaria.

The optimal applications would be solaria or, more in general,lightweight structures with high window-to-wall ratios with thepossibility of performing natural ventilation to ease the dischargeof the thermal storage. This solution would also be of interest forNet Zero Energy Buildings, as the potential to shave and delay ther-mal loads is paramount to align generation and loads. Also in retro-fit applications where thermal storage needs to be added it is moreeasily done on a vertical interior surface.

This paper describes the results of the experimental tests and isprimarily oriented to building engineers, mechanical engineers andarchitects.

Table 1Properties of the PCM modules [31].

Latent heat of fusion 70,000 J/kg (18–24 �C)Specific heat capacity (average) 2500 J/(kg �C)Density 855.5 kg/m3

Dimensions 1000 mm � 1198 mm � 5.2 mmThickness 5.26 mm

2. Experimental setup

This section describes the results of an experiment performed inthe Solar Simulator and Environmental Chamber (SSEC) researchfacility located at Concordia University Montreal, Canada [37].

The main objective of this study is to investigate the potential ofusing PCMs on interior walls of solaria or rooms with high solargains as a mean to save energy and reduce the room temperaturefluctuations.

The main concept of the design is to create a passive PCM wall,oriented towards a window and able to absorb solar radiation andslowly release it in the following hours. This application is mainlydirected towards cold climates, since it does not aim – like manyPCM applications – to ‘‘shield” the whole room from the exterior,but instead it aims to increase the effective use of solar gains byabsorbing them in the storage system for later use.

PCM layers were placed on the interior surface of a test roomwall facing a large window and were tested under different indoorand outdoor conditions. The installed PCM panels are identical tothose used in [32]; each panel measures 1.2 � 1.0 m with a5.2 mm thickness. Five layers are used to cover around 80% ofthe back wall surface area. The main specifications given by themanufacturer are listed in Table 1.

The test room (Figs. 1and 2a, c) is a raised parallelepiped withinterior dimensions of 2.80 m width � 1.30 m depth � 2.44 mheight. A large 2.2 m by 2.2 m window is installed on the frontfaçade. It has a U-value of 1.3 W/(m2 K) at the center and 1.9 W/(m2 K) for the frame. Its Solar Heat Gain Coefficient is 0.262.

The test room is located inside the climatic chamber withdimensions of 8.9 m � 7.3 m � 4.7 m to test the performance ofthe room under different conditions.

Solar radiation was simulated by means of a solar simulator(Fig. 2b) equipped with six special metal halide lamps as sourceof radiation. In combination with glass filters, the lamp system pro-vides a spectral distribution very close to natural sunlight, whichfulfills the specifications of the relevant standards EN12975:2006and ISO 9806-1:1994.

Monitoring was performed with thermocouples placed at dif-ferent heights in the back wall at every layer interface and at a dis-tance of 50 mm from the front surface and in the middle of theroom for measuring the air temperature. Readings from thermo-couples were stored every three minutes for each thermocouplefor the whole test.

The different experimental conditions that were tested arebriefly summarized next. The numbering and the nomenclatureadopted is described in Fig. 3.

2.1. Test n.1

Test n.1 aimed at obtaining monitored data to be used for per-forming numerical model validation.

The test lasted around 60 h; the environmental chamber wasset to a constant temperature equal to 16 �C while the indoor envi-ronment of the test room was heated at a constant rate (350 W).

The heating phase lasted for 20 h and 51 min; after that, the testroom was free floating inside the environmental chamber.

Fig. 4 presents the results for temperature registered at differ-ent depths on a mid-height section of the PCM wall. Readings forthermocouple 1 (front layer), 6 (back layer) and 3 and 4 are pre-sented together with the air temperature; they are considered rep-resentative of the whole dataset.

Fig. 1. Schematics of the test room with exterior dimensions.

98 F. Guarino et al. / Applied Energy 185 (2017) 95–106

The most important observations are as follows:

(1) Temperature will rise progressively slower, the deeper thePCM layer considered. The phase change is also identifiableat around 10–15 h in the air temperature where the slopeof the curves suddenly changes (circled in Fig. 4).

(2) The innermost PCM panel shows a clearer phase change.There is a significant delay of 5–6 h between the beginningof the slope change of the air temperature and the phasechange of the innermost PCM layer. This means that thephase change of the front layer would allow for a stronger

Fig. 2. (a) Back view of the test room; (b) Solar simula

and faster thermal response than the layer at the back. Themelting process lasts for around 7–8 h considering all thedifferent depths;

2.2. Test n�2

The same experimental setup was used to model conditionsrepresentative of sunny days in mid-seasons in moderately coldclimates.

Boundary conditions for test n.2 are:

– The temperature in the chamber was set to 10 �C at the begin-ning of the test;

– A sinusoidal chamber temperature is set up with maximum at20 �C (At 11 a.m.) and minimum at 10 �C after 12 h;

– The solar radiation set point for the solar simulator was500W/m2 and the average radiation incident on the glass wasmeasured as 496.11 W/m2.

The test has been running for 3 consecutive days (Fig. 5). Solarradiation is simulated initially for 3.5 h in the morning. During thefirst day, the solar simulator is switched on from 9:30 a.m. until1:00 p.m. allowing indoor air temperature to reach nearly 35 �Cand a surface temperature of the PCM of around 21 �C. Fig. 5includes data for each thermocouple, with the exception of 4 and5 since they had a trend intermediate between layers 3 and 6.

On the second test day after solar irradiation, PCM tempera-tures values were higher (around 19 �C) than the day before at sim-ilar time after the solar simulator was switched off, and still in thephase change range. The minimum temperature of the sinusoidalair temperature of the climatic chamber was set to 6 �C for the lastday trying to allow a better solidification of the PCM.

From the analysis of Fig. 5, the following observations arise:

(1) During the three and a half hours of solar radiation incidenton the PCMwall, the temperature of the front layer (1) reactsquickly, rising by 17 degrees. The temperature of the backlayer (6) grows with a much lower slope, reaching only a3 �C increase during the irradiation time;

tor working; (c) Front-side view of the test room.

Fig. 3. Schematic and nomenclature of the thermocouple placements in the PCMwall.

15

17

19

21

23

25

27

29

31

33

35

0 5 10 15 20 25 30 35 40 45 50

Tem

pera

ture

[°C

]

Hours

Air T 1 T 3 T 4 T 6

Fig. 4. Temperatures at different depths from the inside to the outside.

F. Guarino et al. / Applied Energy 185 (2017) 95–106 99

(2) When solar irradiation ends at around 15 h, the front PCMlayer and the air temperature drop fast even though the con-cavity of the curve of the PCM layer is reversed at around25 h due to the melting process. However, it is worth notingthat the temperature of the contiguous layers continue to

13

15

17

19

21

23

25

27

29

31

33

35

0 5 10 15 20 25 30 35 40 45 50

Tem

pera

ture

[°C

]

Hours

Fig. 5. Thermocouples tempera

grow for more than 4 �C during around ten hours. The max-imum temperature of the front layer is higher than 30 �C:therefore, heat would flow from the front layer to the deeperones. When the temperatures of the front and back layersreach an equilibrium, the slope of the curves changes andalso the temperature of the back layer starts to decrease;

(3) The melting process for the deeper layers occurs 5–6 h laterthan the front one. At this step, the melting process starts,however it would probably not be a complete melting, sincethe temperature of the PCM layer before the third irradiationphase is around 19 �C. This has several implications and maybe a starting point for further analysis: without active venti-lation, after a high radiation period, the system does notcomplete a charge-discharge cycle and therefore, the start-ing temperatures for the following irradiation cycle wouldbe higher by 4–5 �C than those of 24 h before;

(4) The second irradiation phase lasts for 3 h (half an hour lessthan the first one) since it is stopped as the temperature ofthe indoor air reaches the same maximum value of the pre-vious day. This time, however, although the external tem-peratures are nearly the same of the previous day, theinternal layers temperatures are nearly 5 �C higher (closeto 26 �C, above the melting range). This means that thePCM would not probably even start melting with the currentsetup. For this reason the lower limit of the sinusoidal envi-ronmental chamber temperature was fixed for day 2–6 �C.As expected the cooling process is much faster than on day 1.

3. Modeling

In order to perform further studies, a model was created inEnergyPlus [38] by using a customized weather file to model theenvironmental chamber conditions. The phase change materialsproperties employed are those declared by the manufacturer.Enthalpy and conductivity curves as a function of temperatureare described in [39]. The variability of conductivity with temper-ature was taken in consideration, ranging between 0.17 and0.22 W/(m K).

The simulation used the finite difference conduction model andThermal Analysis Research Program (TARP) convection calcula-tions. PCMs were modeled through the enthalpy and conductivity

5

7

9

11

13

15

17

19

21

23

25

55 60 65 70 75 80

Cha

mbe

r te

mpe

ratu

re [°

C]

T 1

T 2

T 3

T 6

Air

Chamber

ture readings for Test n.2.

100 F. Guarino et al. / Applied Energy 185 (2017) 95–106

curves described in [39,40]. A third order backward differencesolution for the room air heat balance was adopted.

EnergyPlus allows for modeling of PCM using implicit finite dif-ference scheme coupled with an enthalpy-temperature function toaccount for phase change energy accurately. The implicit formula-tion for an internal node is shown in Eq. (1).

CpqDxTjþ1t � T j

i

Dt¼ 1

2kw

ðTjþ1tþ1 � Tjþ1

i ÞDx

þ keðTjþ1

t�1 � Tjþ1i Þ

Dx

"

þ kwðT j

tþ1 � T ji Þ

Dxþ ke

ðT jt�1 � T j

i ÞDx

#ð1Þ

where all symbols are described in the nomenclature.Model validation was performed by comparing experimental

and simulated data (Test n.1). Different datasets have been ana-lyzed: air average temperatures and the PCM superficial tempera-tures at different depths. The chosen Heat Balance Methodmodeling is based on the assumption that air temperature is con-stant in a zone volume.

This specific feature has required the aggregation of some of themost detailed monitored data into less detailed averages. Air tem-perature was recorded on five different vertical heights, in the cen-ter of the room. Temperature difference between the top and thebottom level is always below 1 �C, as shown in Fig. 6, which justi-fies the uniform indoor temperature assumption.

The simulation time step was adjusted to match the timeframeof the temperature readings (3 min). Monitored results have beencompared with the simulated ones and the absolute errors are pre-sented in Table 2.

This table represents the distribution of the absolute errorbetween measurements and simulation results, the percentagevalues are the percentile of exceedance, or in other words, the per-cent of a distribution that is equal to or below it. Absolute error for

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 5 10 15 20

Ver

tical

tem

pera

ture

diff

eren

ce

H

Fig. 6. Vertical tempe

Table 2Absolute errors in simulation [�C].

Percentile e

Air 1 2

MIN 0.0016 0.056 0.001110% 0.041 0.188 0.1325% 0.223 0.426 0.19250% 0.638 0.52 0.5375% 0.78 0.743 0.79590% 0.912 0.88 0.872MAX 0.98 0.923 0.937

each timestep was calculated and arranged in ascending orderranging from theminimum to themaximum absolute error. Table 3includes some more statistical metrics to assess the error in mod-eling: Mean Bias Error (MBE) and the coefficient of variation of theroot mean square error (CV(RMSE)).

The mean bias error (MBE) of a sample of n measurements isdefined as in Eq. (2):

MBE ¼ 1n

Xnj¼1

ðts;i � tm;iÞ ð2Þ

where tm,i is monitored values and ts,i is simulated values at time i.The coefficient of variation of the root mean square error (CV

(RMSE)) of a sample of n measurements is defined as in Eq. (3):

CVðRMSEÞ ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1n

Pnj¼1ðts;i � tm;iÞ2

qtm

ð3Þ

where tm,i are monitored values and ts,i are simulated values at timej, while tm is the mean of the monitored data.

4. Parametric analysis: results

4.1. Configurations definition

In order to explore the energy performance of PCM applicationsin the indoor environment, a parametric analysis has been per-formed. The system investigated is the one discussed in Section 2.

A PCM wall composed of five layers of PCM is located on thewall opposite to the large south oriented glazed opening (60% win-dow to wall ratio - WWR). The main concept behind this design isto allow for a direct solar radiation on the latent energy storage inorder to maximize solar heat gains and store them for later useduring winter. Two scenarios are discussed in the following: the

25 30 35 40 45 50ours

rature difference.

rrors

3 4 5 6

0.0014 0.0021 0.0013 0.00130.068 0.259 0.087 0.0870.155 0.288 0.151 0.1510.511 0.702 0.539 0.5390.755 0.765 0.733 0.6320.859 0.812 0.844 0.7580.929 0.897 0.967 0.832

Table 3Mean bias errors and root mean square errors.

Air 1 2 3 4 5 6

MBE 0.3683 �0.0032 0.2679 0.2253 0.1594 �0.1250 �0.1383Cv(RMSE) 1.96% 0.49% 1.82% 1.66% 2.04% 1.51% 1.50%

-400

-300

-200

-100

0

100

200

300

400

500

600 0

1/01

01:

00:0

0 0

1/11

16:

00:0

0 0

1/22

07:

00:0

0 0

2/01

22:

00:0

0 0

2/12

13:

00:0

0 0

2/23

04:

00:0

0 0

3/05

19:

00:0

0 0

3/16

10:

00:0

0 0

3/27

01:

00:0

0 0

4/06

16:

00:0

0 0

4/17

07:

00:0

0 0

4/27

22:

00:0

0 0

5/08

13:

00:0

0 0

5/19

04:

00:0

0 0

5/29

19:

00:0

0 0

6/09

10:

00:0

0 0

6/20

01:

00:0

0 0

6/30

16:

00:0

0 0

7/11

07:

00:0

0 0

7/21

22:

00:0

0 0

8/01

13:

00:0

0 0

8/12

04:

00:0

0 0

8/22

19:

00:0

0 0

9/02

10:

00:0

0 0

9/13

01:

00:0

0 0

9/23

16:

00:0

0 1

0/04

07:

00:0

0 1

0/14

22:

00:0

0 1

0/25

13:

00:0

0 1

1/05

04:

00:0

0 1

1/15

19:

00:0

0 1

1/26

10:

00:0

0 1

2/07

01:

00:0

0 1

2/17

16:

00:0

0 1

2/28

07:

00:0

0

Hea

t sto

rage

cha

rge

and

disc

harg

e [W

]

Scenario 2 Scenario 1

Fig. 7. Heat storage charging and discharging power for the PCM – north wall.

F. Guarino et al. / Applied Energy 185 (2017) 95–106 101

test room with no PCM layers (Scenario 1, from now on S1) andwith PCM panels in the north wall (S2).

Natural ventilation is included in both models through the‘‘Zone: Wind and stack object” in EnergyPlus. Such scenarios useempirical formulations to correlate wind angle and speed, fenes-tration areas and opening factors to calculate the air change rate.In order to avoid overcooling in the room, natural ventilation isoperated through a minimum indoor temperature set-point(23 �C) and a maximum outdoor temperature set-point (18 �C).

Models are simulated as both free floating and conditioned,with 18–26 �C set-points for heating and cooling. These heatingset-point were chosen to be outside the phase change temperaturerange. Weather data used is for Montreal, Canada (latitude 45 N).

4.2. Results

In this section hourly and yearly data are presented.The main aim of the design is to increase the utilization of solar

gains by storing excess heat for later use.Fig. 7 clarifies this concept by showing the energy stored and

released in both S1 and S2 during the whole simulated year. Heatcharge and discharge in the storage wall is shown in both S1 andS2. It can be seen that the presence of PCMs significantly improvesthe energy storage potential.

The effectiveness in charging and discharging the wall is higherfor S2 from roughly September to April than in the summermonths, while it is maximum for S1 during summer.

This trend is mirrored in the yearly data set included in Fig. 8,showing the solar beam radiation that reaches the north wall dur-ing the year.

The highest values are reported for winter due to solar geome-try; with the other walls working as solar shadings in summer, thePCM wall does not receive direct solar radiation in this season.

More insight is gained when selecting the first 10 days of theyear taken as example in Fig. 9 and analyzing the trends in moredetail. Fig. 9 shows the first days of the simulated year in the caseof a conditioned building is. It demonstrates that in addition to themuch larger heat storage rate in the PCM application scenario com-pared to S1, the discharge phase is also prolonged 6–8 h after theend of solar irradiation.

Some more detailed trends are described in the following, con-sidering different weather conditions both in the case of a condi-tioned building and a free floating (passive) building.

Figs. 10a and 10b show trends for temperature and heatingloads on some winter days. The simulated results in Fig. 10a showclearly the impact that the PCM can have on a conditionedbuilding.

When solar gains are high, the system is able to store energyand give it back in the following hours, as seen during the nightbetween January 1st and 2nd. Peak heating load of S2 are lowerby nearly 100W during January, 2nd at night, in a range between5 and 15% of the total S1 peak. Overall, the use of the PCM reducesthe heating power in nearly all the hours shown.

In Fig. 10b, the building is free floating. Although the phasechange temperatures are not reached, the sensible storage of thePCM wall allows to reduce the air temperature variations from15� and �15 �C, to 7 �C and �7 �C.

Figs. 11a and 11b describe similar trends for the selected coldsunny and cloudy days of late March. The conditioned building(Fig. 13a) show results dependent on the solar radiation regime:during the cold cloudy days, scenario S2 reports only a slightreduction in heating loads, while during the last, sunny days, theevening heating requirements are reduced by around 40%. Coolingrequirements are also relevant during the high solar radiation daysand are roughly reduced by 50% in scenario S2. The particular con-figuration of the simulated room causes indoor free floating tem-

-300

-150

0

150

300

450

600

750

900

Hea

t sto

rage

cha

rge/

disc

harg

e [W

]

Scenario 2 Scenario 1 Windows transmitted solar radiation

Fig. 9. Heat storage charge and discharge rate for the PCM (S2) or North wall (S1).

0

100

200

300

400

500

600

700

-20-18-16-14-12-10

-8-6-4-20

01/01 01:00:00 01/02 02:00:00 01/03 03:00:00 01/04 04:00:00

Power [W

]Te

mpe

ratu

re [

°C]

Window Transmitted Solar Radiation Rate [W] Outdoor Air Drybulb Temperature [C]

S2 Heating [W] S1 Heating [W]

Fig. 10a. Cold sunny days, January in Montreal: conditioned buildings with setpoints (S2).

0

100

200

300

400

500

600

01/0

1 0

1:00

:00

01/1

1 1

6:00

:00

01/2

2 0

7:00

:00

02/0

1 2

2:00

:00

02/1

2 1

3:00

:00

02/2

3 0

4:00

:00

03/0

5 1

9:00

:00

03/1

6 1

0:00

:00

03/2

7 0

1:00

:00

04/0

6 1

6:00

:00

04/1

7 0

7:00

:00

04/2

7 2

2:00

:00

05/0

8 1

3:00

:00

05/1

9 0

4:00

:00

05/2

9 1

9:00

:00

06/0

9 1

0:00

:00

06/2

0 0

1:00

:00

06/3

0 1

6:00

:00

07/1

1 0

7:00

:00

07/2

1 2

2:00

:00

08/0

1 1

3:00

:00

08/1

2 0

4:00

:00

08/2

2 1

9:00

:00

09/0

2 1

0:00

:00

09/1

3 0

1:00

:00

09/2

3 1

6:00

:00

10/0

4 0

7:00

:00

10/1

4 2

2:00

:00

10/2

5 1

3:00

:00

11/0

5 0

4:00

:00

11/1

5 1

9:00

:00

11/2

6 1

0:00

:00

12/0

7 0

1:00

:00

12/1

7 1

6:00

:00

12/2

8 0

7:00

:00

Sola

r rad

iatio

n [W

]

Beam radiation reaching the PCM wall

Fig. 8. Beam solar radiation incident on the PCM wall during a simulated year.

102 F. Guarino et al. / Applied Energy 185 (2017) 95–106

0

100

200

300

400

500

600

700

-20

-15

-10

-5

0

5

10

15

01/01 01:00:00 01/02 02:00:00 01/03 03:00:00 01/04 04:00:00

Power [W

]Te

mpe

ratu

re [°

C]

Window Transmitted Solar Radiation Rate [W] S2 Zone Air Temperature [C]

Outdoor Air Drybulb Temperature [C] S1 Zone Air Temperature [C]

Fig. 10b. Cold sunny days, January in Montreal: free-floating temperatures (S2).

0

100

200

300

400

500

600

-10

-8

-6

-4

-2

0

2

4

03/25 09:00:00 03/26 10:00:00 03/27 11:00:00 03/28 12:00:00

Power [W

]Te

mpe

ratu

re [°

C]

Window Transmitted Solar Radiation Rate [W] Outdoor Air Drybulb Temperature [C]S2 Heating [W] S2 Cooling [W]S1 Heating [W] S1 Cooling [W]

Fig. 11a. Cold sunny/cloudy days, March in Montreal: conditioned building (setpoints: 18–26 �C).

0

100

200

300

400

500

600

-10

0

10

20

30

40

50

03/2509:00:00

03/2519:00:00

03/2605:00:00

03/2615:00:00

03/2701:00:00

03/2711:00:00

03/2721:00:00

03/2807:00:00

03/2817:00:00

03/2903:00:00

Power [W

]Te

mpe

ratu

re [°

C]

Window Transmitted Solar Radiation Rate [W] S2 Zone Air Temperature [C]

Outdoor Air Drybulb Temperature [C] S1 Zone Air Temperature [C]

Fig. 11b. Cold sunny/cloudy days, March in Montreal: free-floating temperatures.

F. Guarino et al. / Applied Energy 185 (2017) 95–106 103

peratures to rise up to 40 �C in S1 (Fig. 13b) even when externaltemperature is around 0 �C. The use of the PCM in S2 allows reduc-ing the indoor maximum temperature by 10 �C during the sunniestdays.

Figs. 12a and 12b present results during typical days in August.Some cooling loads reductions are obtained (Fig. 12a) when com-pared to S1 (a peak cooking load reduction of 120 W– around17%), while the already limited heating load in S1 is zero in S2.

0

100

200

300

400

500

600

700

800

0

5

10

15

20

25

08/1904:00:00

08/1914:00:00

08/1924:00:00

08/2010:00:00

08/2020:00:00

08/2106:00:00

08/2116:00:00

08/2202:00:00

08/2212:00:00

08/2222:00:00

08/2308:00:00

Power [W

]

Tem

pera

ture

[°C

]

Window Transmitted Solar Radiation Rate [W]Outdoor Air Drybulb Temperature [C]S2 Heating [W]S2 Cooling [W]S1 Heating [W]

Fig. 12a. Summer days in Montreal, August – conditioned buildings with the setpoints.

0

50

100

150

200

250

300

350

400

450

500

0

5

10

15

20

25

30

35

40

45

50

08/18 23:00:00 08/19 23:00:00 08/20 23:00:00 08/21 23:00:00 08/22 23:00:00

Power [W

]Te

mpe

ratu

re [°

C]

Window Transmitted Solar Radiation Rate [W] Zone Ventilation Air Change Rate [ach]

S1 Zone Air Temperature [C] Outdoor Air Drybulb Temperature [C]

S2 Zone Air Temperature [C] S2 PCM Temperature

Fig. 12b. Summer days in Montreal, August – Free floating.

104 F. Guarino et al. / Applied Energy 185 (2017) 95–106

Natural ventilation has a relevant role in reducing air and PCMtemperatures and in helping the freeze-melting cycle. With com-parable incident solar radiation as in the three days described inFig. 12b, it can be seen that the role of natural ventilation is veryimportant: during August 19th and 20th the zone temperaturereaches 30 �C, acceptable for a solarium. On the 21st, natural ven-tilation was not performed during the hottest hours of the day: thiscauses indoor temperature to reach 45 �C in the case of S1. The useof PCM allows to lower this value by nearly 7 �C.

However, the PCM system will not revert to solid after August,22th and thus its latent storage capabilities would not be used atits best. It is worth noting that the PCM melting range (18–24) isnot appropriate for cooling conditions optimization, since duringmildly hot external conditions the PCM superficial temperature isabove the phase change boundaries.

The highest temperature of the PCMwall would be reached afterthe irradiation phase: thismeans that during the night, the PCMwilloften have higher temperatures than the air and will contribute tothe overheating of the room. In summer PCM will need to be care-fully discharged during the night if outdoor temperatures allow it.

4.3. Parametric study, annual summary

The annual results of the study are presented in Fig. 13. Itdescribes monthly contributions to heating and cooling require-ments for the two scenarios.

Annual S1 heating requirements are the most relevant contribu-tions to the total, amounting to 1374 kW hwhile cooling is equal to326 kW h in the case of S1. Heating is lower in the case of S2 whencompared to S1 during the whole year, varying from around 5% inJanuary to 30% in April and reaching nearly 100% in summer. Cool-ing has a reverse trend during the year and shows savings that arenot lower than 30% when comparing S2 to S1.

Overall, heating is reduced by 17.4%, while cooling is lower inS2 by roughly 50% during the whole year. Furthermore, coolingcould be reduced even more if different ventilation strategieswould be adopted, by further lowering the maximum externaltemperature allowed indoor.

5. Discussion

The paper discusses experimental studies and a set of calibratedbuilding simulations that aim at proposing solutions and applica-tions to solaria environments with application to one wall thatreceives most of the solar radiation. The two tests performed withdifferent conditions represent two different situations in a typicalyear in Montreal, and as such can be representative of the climatein Northern Europe, Northern United States and Canada, as well asmany more cold weather sites in the world. The results in theircurrent state are still local and the findings reported cannot be gen-eralized at this stage for any other climate conditions.

-100%

-90%

-80%

-70%

-60%

-50%

-40%

-30%

-20%

-10%

0%0

50

100

150

200

250

300

350

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Hea

ting

and

cool

ing

ener

gy re

quire

men

ts

[kW

h]S1 Heating

S2 Heating

S1 Cooling

S2 Cooling

S2 Heating variation

S2 Cooling variation

Fig. 13. Monthly aggregated heating requirements for all scenarios.

F. Guarino et al. / Applied Energy 185 (2017) 95–106 105

The experiments have shown that five PCM layers placed on aback wall, directly irradiated by the sun, have a strong impact onthe temperature fluctuations of the indoor environment. The exactnumber of layers depends on the amount of stored energy that isdesired and the time period of the storage. The dynamics of theinter-layer heat exchange and temperature fluctuation allow fora delayed and persistent heat delivery effect after irradiation hadstopped. This defines this particular solution set as one suited forimplementation in colder climates (or at least with heating energystorage purposes) to store solar energy and re-deliver it later to theindoor environment. The experiments have shown that under mildexternal temperatures (10 �C) the indoor temperature would notbe lower than 18 �C when PCMs are present.

One of the tests represented some optimal conditions for thestorage to work effectively. External temperature was not toolow and solar radiation was high enough to reach the melting tem-perature of the PCM. The heat storage worked well since the tem-perature of PCM layers would remain inside the melting range upto 5–8 h after the end of the solar irradiation phase. Since the min-imum PCM temperature was not low enough to complete thesolidification process the environmental chamber minimum(external) temperature was decreased from 10 �C to 6 �C. This con-cept highlights the need for a good control of the room: either nat-ural ventilation, solar shadings or a combination of the two areneeded to ensure effective use of the PCM.

Cooling and heating energy and peak power are always reducedduring the whole year. The proposed design solution is practical,particularly for retrofit applications in cold climates since the stor-age system is able to discharge heat up to 8 h after irradiation hasstopped. The optimal thickness of the PCM layers however is vari-able and depends on the design of the room and on the thermalloads of the building.

The system is also effective for reducing cooling needs duringthe hottest months but it is only an indirect result of increasingthe inertia of the room: the PCM is often completely melted duringsummer due to the melting range (18–24 �C) being too low forcooling applications. Moreover, since the main aim of a bioclimaticdesign of a building in cooling dominated climates is to shield theenvelope from solar radiation, the solution itself, although itobtains good results during summer, is not the best for such appli-cations. For cooling dominated applications, it could be useful touse a uniform distribution of PCM on the other surfaces of theroom and materials with higher phase change temperature ranges.

The system needs a finely tuned control to work at its best.Indoor air temperature set-points should be carefully adjusted tomatch also dynamically the melting/freezing states and the out-door conditions to activate and discharge the storage system.

Moreover, compared to conventional sensible heat storage, PCMstorages allow for a high energy density operating at lower tem-perature variation ranges. Thus, PCMs are usually more compactthan sensible storages and can therefore be particularly effectivein the case of building retrofits.

Lastly, it is worth highlighting the importance of the passiveapproach largely discussed in this paper, from the design conceptto all the applications and simulations described. Sensible heatstorage systems were not reviewed in this paper. As a completelypassive solution set, the design described can be probably charac-terized by higher first costs, in comparison to active systems solu-tions, but it would eliminate need for maintenance and be suitablefor retrofit on walls.

6. Conclusions

This study has analyzed the impacts of a PCM thermal energystorage system on the performance of a solarium in a cold climate.All the results point towards a reduction in daily temperatureswings and of the overall heating requirements.

The experiment showed that the PCM wall worked as intendedunder moderate cold weather conditions, since the PCM could pro-vide an appreciable time lag with the solidification process thattook around 5–8 h. However, without ventilation and after a highirradiation period, the system would not undergo a completecharge-discharge cycle and therefore, the starting temperaturesfor the following irradiation cycle would be around 19 �C, insidethe melting range. This has implications for the use of such sys-tems during shoulder months in moderately cold climates: naturalventilation is advised and also active solutions (fans blowing airthrough the PCM panels) could be beneficial. In solaria and highlyglazed buildings, the room dynamics can be highly improvedaccording to the simulation results when using PCM as a thermalstorage:

– Temperature air swings were decreased by up to 10 �C for day-night cycles,

– Heating peak loads were lowered significantly (up to 40%) dur-ing high irradiation days in winter;

Simulations defined the PCM solaria solution as clearly benefi-cial for cold regions applications. A concentrated PCM thermalstorage proved able to reduce heating requirements by 17.4% in ayear, largely increasing the storage capability of the wall anddelaying heat discharge for 6–8 h during the night. It is advisableto include natural ventilation in the design of solaria, since it can

106 F. Guarino et al. / Applied Energy 185 (2017) 95–106

largely reduce the cooling requirements and help in the tempera-ture peaks during summer.

Using a heating setpoint temperature below the phase change ishelpful in improving the performance of the system: although low(18 �C), it still is inside the comfort range.

The proposed concept is particularly easy to implement on awall opposite a large window in a solarium with high solar gainsand can be implemented as a retrofit solution with the numberof layers of PCM chosen based on the goals of peak load reductionand time length of storage.

References

[1] Cabeza LF, Castell A, Pérez G. 13 – Life cycle assessment (LCA) of phase changematerials (PCMs) used in buildings. In: Pacheco-Torgal F, Cabeza LF, LabrinchaJ, Magalhães Ad, editors. Eco-efficient construction and buildingmaterials. Woodhead Publishing; 2014. p. 287–310.

[2] Martínez-Lera S, Ballester J, Martínez-Lera J. Analysis and sizing of thermalenergy storage in combined heating, cooling and power plants for buildings.Appl Energy 2013;106:127–42.

[3] Li C, Wang RZ. Building integrated energy storage opportunities in China.Renew Sustain Energy Rev 2012;16(8):6191–211.

[4] Pavkovic D, Hoic M, Deur J, Petric J. Energy storage systems sizing study for ahigh-altitude wind energy application. Energy 2014;76:91–103.

[5] Tatsidjodoung P, Le Pierrès N, Luo L. A review of potential materials for thermalenergy storage in building applications. Renew Sustain Energy Rev2013;18:327–49.

[6] Guarino F, Cassarà P, Longo S, Cellura M, Ferro E. Load match optimisation of aresidential building case study: a cross-entropy based electricity storage sizingalgorithm. Appl Energy 2015;154:380–91.

[7] Navarro L, de Garcia A, Solé C, Castell A, Cabeza LF. Thermal loads insidebuildings with phase change materials: experimental results. Energy Procedia2012;30:342–9.

[8] Chen Y, Athienitis AK, Galal KE. Frequency domain and finite differencemodeling of ventilated concrete slabs and comparison with fieldmeasurements: Part 2. Application. Int J Heat Mass Transf 2013;66:957–66.

[9] Chen Y, Athienitis AK, Galal KE. Frequency domain and finite differencemodeling of ventilated concrete slabs and comparison with fieldmeasurements: Part 1, modeling methodology. Int J Heat Mass Transf2013;66:948–56.

[10] Rahimi M, Ranjbar AA, Ganji DD, Sedighi K, Hosseini MJ, Bahrampoury R.Analysis of geometrical and operational parameters of PCM in a fin and tubeheat exchanger. Int Commun Heat Mass Transfer 2014;53:109–15.

[11] Zhang Y, Zhou G, Lin K, Zhang Q, Di H. Application of latent heat thermalenergy storage in buildings: state-of-the-art and outlook. Build Environ2007;42(6):2197–209.

[12] Marchi S, Pagliolico S, Sassi G. Characterization of panels containing micro-encapsulated phase change materials. Energy Convers Manage2013;74:261–8.

[13] Kuznik F, Virgone J, Noel J. Optimization of a phase change material wallboardfor building use. Appl Therm Eng 2008;28(11–12):1291–8.

[14] Diarce G, Urresti A, García-Romero A, Delgado A, Erkoreka A, Escudero C, et al.Ventilated active façades with PCM. Appl Energy 2013;109:530–7.

[15] Zhou G, Zhang Y, Lin K, Xiao W. Thermal analysis of a direct-gain room withshape-stabilized PCM plates. Renewable Energy 2008;33(6):1228–36.

[16] Wang W, Li H, Guo S, He S, Ding J, Yan J, et al. Numerical simulation study ondischarging process of the direct- contact phase change energy storage system.Appl Energy 2015;150(61):68.

[17] Dutil Y, Rousse D, Lassue S, Zalewski L, Joulin A, Virgone J, et al. Modelingphase change materials behavior in building applications: comments onmaterial characterization and model validation. Renewable Energy2014;61:132–5.

[18] Kuznik F, David D, Johannes K, Roux J-J. A review on phase change materialsintegrated in building walls. Renew Sustain Energy Rev 2011;15(1):379–91.

[19] Kuznik F, Virgone J, Johannes K. In-situ study of thermal comfort enhancementin a renovated building equipped with phase change material wallboard.Renewable Energy 2011;36(5):1458–62.

[20] Cellura M, Guarino F, Longo S, Mistretta M. Energy life-cycle approach in netzero energy buildings balance: operation and embodied energy of an Italiancase-study. Energy Build 2014;72.

[21] Hachem C, Athienitis A, Fazio P. Parametric investigation of geometric formeffects on solar potential of housing units. Sol Energy 2011;85(9):1864–77.

[22] Cellura M, Campanella L, Ciulla G, Guarino F, Lo Brano V, Cesarini DN, et al. Theredesign of an Italian building to reach net zero energy performances: a casestudy of the SHC Task 40 – ECBCS Annex 52. ASHRAE Trans 2011:331–9.Montreal, QC.

[23] Cellura M, Campanella L, Ciulla G, Guarino F, Lo Brano V, Cesarini DN, et al. Anet zero energy building in Italy: design studies to reach the net zero energytarget. In: 12th Conference of international building performance simulationassociation building simulation 2011, BS 2011, Sydney, NSW. p. 649–55.

[24] Hachem C, Athienitis A, Fazio P. Evaluation of energy supply and demand insolar neighborhood. Energy Build 2012;49:335–47.

[25] Cellura M, Guarino F, Longo S, Mistretta M. Different energy balances for theredesign of nearly net zero energy buildings: an Italian case study. RenewSustain Energy Rev 2015;45:100–12.

[26] Marszal AJ, Heiselberg P, Bourrelle JS, Musall E, Voss K, Sartori I, et al. Zeroenergy building – a review of definitions and calculation methodologies.Energy Build 2011;43(4):971–9.

[27] Mohamed A, Hasan A, Sirén K. Fulfillment of net-zero energy building (NZEB)with four metrics in a single family house with different heating alternatives.Appl Energy 2014;114:385–99.

[28] Attia S, Hamdy M, O’Brien W, Carlucci S. Assessing gaps and needs forintegrating building performance optimization tools in net zero energybuildings design. Energy Build 2013;60:110–24.

[29] Attia S, Hamdy M, O’Brien W, Carlucci S. Computational optimisation for zeroenergy buildings design: interviews results with twenty eight internationalexperts. In: 13th Conference of the international building performancesimulation association, BS 2013, Chambery. p. 3698–705.

[30] Chen C, Guo H, Liu Y, Yue H, Wang C. A new kind of phase change material(PCM) for energy-storing wallboard. Energy Build 2008;40:882–990.

[31] Ahmad M, Bontemps A, Sallee H, Quenard D. Thermal testing and numericalsimulation of a prototype cell using light wallboards coupling vacuumisolation panels and phase change material. Energy Build 2006;38:673–81.

[32] Kuznik F, Virgone J. Experimental investigation of wallboard containing phasechange material: data for validation of numerical modeling. Energy Build2009;41(5):561–70.

[33] Evers AC, Medina MA, Fang Y. Evaluation of the thermal performance of framewalls enhanced with paraffin and hydrated salt phase change materials using adynamic wall simulator. Build Environ 2010;45(8):1762–8.

[34] Xu X, Zhang Y, Lin K, Di H, Yang R. Modeling and simulation on the thermalperformance of shape-stabilized phase change material floor used in passivesolar buildings. Energy Build 2005;37(10):1084–91.

[35] de Gracia A, Navarro L, Castell A, Ruiz-Pardo Á, Alvárez S, Cabeza LF.Experimental study of a ventilated facade with PCM during winter period.Energy Build 2013;58:324–32.

[36] Diarce G, Urresti A, Garcìa-Romero A, Delgado A, Erkoreka A, Escudero C, et al.Ventilated active façades with PCM. Appl Energy 2013;109:530–7.

[37] Guarino F, Dermardiros V, Chen Y, Rao J, Athienitis A, Cellura M, et al. PCMthermal energy storage in buildings: experimental study and applications.Energy Procedia 2014;70:219–28.

[38] EnergyPlus. Version 8.4. Lawrence Berkeley National Laboratory, Berkeley,USA; 2015.

[39] Kuznik F, Virgone J, Johannes K. Development and validation of a new TRNSYStype for the simulation of external building walls containing PCM. EnergyBuild 2010;42(7):1004–9.

[40] Kuznik Frédéric, Virgone Joseph, Roux Jean-Jacques. Energetic efficiency ofroom wall containing PCM wallboard: A full-scale experimental investigation.Energy Build 2008;40(2):148–56. doi: http://dx.doi.org/10.1016/j.enbuild.2007.01.022. ISSN 0378-7788.