Available online at www.sciencedirect.

com

Energy Procedia 30 (2012) 225 234

SHC 2012

Thermal energy storage by PCM-air heat exchangers:

temperature maintenance in a room
Pablo Doladoa*, Ana Lazaro, Monica Delgado, Conchita Pealosa, Javier Mazo,
Jose Maria Marin, Belen Zalba
Aragn Institute of Engineering Research (I3A), Betancourt Bld. Maria de Luna s/n, Zaragoza, 50018, Spain

Abstract
This paper describes how to design a PCM-air heat exchanger once an application is specified. Free-cooling and
temperature maintenance in rooms with special requirements possess high potential for PCM application in different
countries according to their climate. Until now, the low thermal conductivity of PCM and air hindered the
development of suitable heat exchangers. The overall objective here is to apply methodologies to study PCM-air heat
exchangers that allow the development of applications with technical and economical viability. In particular, using
the combined technique of design of experiments and numerical simulations, the feasibility of the possible
application of this type of equipment is studied and optimized for temperature maintenance in rooms.

2012 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
2012 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of PSE AG
Selection and/or peer-review under responsibility of PSE AG
Keywords: Thermal energy storage; phase change material; design; optimization

1. Background
Lets think about a telecom shelter, somehow far from civilization. The place is full of electronic
equipment, continuously generating heat due to its operation. The chiller unit, plugged in to the electrical
grid, is running avoiding the room air temperature to increase. But what if a short circuit, blackout or
other similar problem happens? If the chiller stops running, the temperature will definitely rise and
problems with the electronic devices will occur... maybe, even the functionality of the telecom shelter
could crash! Electricity supply would be reset once technicians reach the place and fix the problem, but in

* Corresponding author. Tel.: +34 976 761000 ext. 5258; fax: +34 976 762616.
E-mail address: [email protected].

1876-6102 2012 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
Selection and/or peer-review under responsibility of PSE AG
doi:10.1016/j.egypro.2012.11.027

226

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

the meantime, specific equipment is needed to ensure the temperature will not overcome certain safety
threshold value.
Nomenclature
cp [J/(kgK)], specific heat capacity at constant pressure
eair [m], thickness of the air gap between two PCM plates
eplate [m], thickness of the PCM plate
Etmod [kJ], total stored thermal energy in 1 module
h [W/(m2K)], air convection coefficient
i, time step in the numerical scheme
m [kg], building enclosure mass

m [kg/s], mass flow

MPCM [kg], PCM mass
N, number of PCM plate walls (inside the TES unit)
Qdemand [kW], internal cooling demand

Toutside [C], outdoors air temperature

Troomi [C], room temperature at i instant
Tsurface [C], average surface temperature of PCM plates

Tairinlet/outler [C], air temperature at storage unit inlet or at the outlet

tadd, T=38C [s], elapsed time to reach 38C in the room
tadd, T=44C [s], elapsed time to reach 44C in the room

V [m3/h], volumetric air flow

V [m3], volume
w [m], width
%Melt, ratio of PCM melted, percentage
T=Tairinlet-Tairoutlet [K], temperature difference in the TES unit
p [Pa], pressure difference
t [s], time interval
[W/(mK)], thermal conductivity
eff [W/(mK)], effective thermal conductivity
[kg/m3], density
Cair

Uair Vair c p ,air / (2N ) [J/(sK)], heat capacity

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

NTU

 h'xw

Cair , number of transfer units

Bi , Biot number
Re, Reynolds number
DOE, Design of Experiments
HX, Heat Exchanger
PCM, Phase Change Material
TES, Thermal Energy Storage

2. Methodology
Currently, due to the feedback between the simulations and experimental studies, to complement the
performance of physical experiments with numerical simulations is becoming increasingly common in
many sectors. Some of these numerical experiments are conducted to select as well as the tools as the
optimal process parameters that directly allow developing products that reduce to some extent the cost as
well as time required for the corresponding physical experimentation [1, 2]. In other cases the aim is to
obtain a deeper understanding of the effect of varying certain process parameters (sensitivity studies)
towards the optimization of the process itself [3]. Del Coz Daz et al. [4] apply the DOE to numerical
simulations with the double aim of limiting the number of runs, significantly reducing the computational
cost on the one hand, and on the other hand, allowing influence analysis of each parameter on the
responses of interest. Whitt [5] indicates that as the simulations require a statistical analysis of their
results, it is appropriate to consider the perspective of experimental design. However, Salazar and Baena
Zapata [6] point out that the numerical studies based on the methodology of DOE are scarce in the
literature. These authors invite researchers working in the area of statistics (although extensible to any
other area) to implement the methodology of design and analysis of experiments to the numerical
simulation studies as they consider that this type of analysis improves the quality of the findings (being
these more convincing) as well as the presentation of the simulation analysis results. In addition,
numerical simulation studies considered under this type of approaches allow analyzing many factors at
once so optimizing the time spent in the execution of simulations. Further information on DOE concepts
applicable to numerical simulations can be found in the work by Gunasegaram et al. [7] where advantages
and limitations are reported by the authors. This kind of approach focused on product design rather than
on process, is really scarce in scientific literature dealing with thermal equipment.
As a starting point we state the case brought by Lazaro [8], which provides that, for proper running of
the electronic equipment, the maximum air temperature in the room should be between 38 C and 48 C,
in particular the threshold temperature is established at 44 C. The idea behind this system is that after a
failure of the conventional cooling system, the TES unit is intended to smooth the evolution of the
temperature of the room so that it extends the time to reach a certain threshold temperature value. The aim
is this period to be about two hours, so technicians have sufficient time to reach the place where the room
is located and to repair the damage of the cooling system without having to stop the electronic equipment.
A series of restrictions were put on the TES system, regarding dimensions and electrical power
consumption, as well as on pressure drop.

227

228

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

2.1. Empirical model: simulations of a case study and modular design

An empirical model was built from the experimental results gathered from the experimental setup (Fig.
1). The aim was to simulate the thermal behavior of the tested heat exchanger in different cases. These
simulations were used to evaluate the technical viability of the application. The model describes the
temperature evolution of a room with an internal cooling demand, where the PCM-air heat exchanger is
operating. The room has also a ventilation system. The enclosure temperature was considered to be the
average between the outside temperature and the room temperature (the enclosure is the typical
lightweight construction of telecom shelters: insulation material layer (glass fiber) between two plate
metal sheets). A diagram of the room is shown in Fig. 2. Eq. (1) shows the energy balance applied to the
air inside the room. The heat generation of the electronic equipment is 5 kW. For the temperature
evolution inside the room, an energy balance was stated with the following assumptions: 1) the cooling
effect of the terrain was not considered. The ground floor area is supposed to be occupied by the
equipment; 2) exterior ventilation is introduced only when it is favorable, and considering that the
environment outside the room is 40 C (as boundary condition specified by the company). T is obtained
at each instant as a function of T surface and the inlet air temperature (at instant i equal to Troomi); and the
T surface at instant i is obtained from the stored energy evolution. The real-scale PCM-air heat exchanger
tested was constituted of 18 parallel PCM plates-wall. A module is constituted by a metallic PCM
container between two air channels modules. The pressure drop is the same for each module, and the air
distribution through the air channels can be considered uniform. The unitary air flow through a module is
mair HX divided by 18. Since the geometry and the air flow were kept identical, total thermal energy
stored in one module between two temperatures is the stored energy for the real-scale PCM-air heat
exchanger between the two temperatures divided by 18. The total melting time depends on that stored
energy as well as on Qdemand . In order to apply a modular design, the relation between the experimental
heat rate provided by the electrical resistances to simulate the thermal load and the heat rate demanded in
a potential application is established according to the experimental number of PCM modules (18) and the
ones required for design. Details on empirical model and modular design can be found in Lazaro [8, 9].

Fig. 1. Experimental setup to test real scale PCM-Air heat exchangers

229

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

Enclosure

T surface

Qdemand

V ttilation
Ven
Tairinlet

Toutside

Troomi

Tairoutlet

= Tairiinlet - Tairiinlet

Fig. 2. Schematic diagram of the room (telecom shelter) in which the temperature is evaluated

i-1
T i

Troom
+Tooutside
i
i-1
room +Toutside
t id
t id = Vc
mcpp,enclosure
enclosure
p,air Troom Troom
r

air
2
2

i
i 1
i-1
= m ventilation
cpp,air
Troom +Q
Qddemandd -m
mair
cp,air
, 
il i c
air
i Toutside
id -T
i HX c
air
i T t  i,i-1

(1)

2.2. Theoretical model: DOE

O applied to simulations, improving design
The empirical model can give a very fast approach of relevant design parameters such as the PCM
average phase change temperature. However, if we want to analyze the behavior of the equipment when
modifying any other parameter or variable, or if we need to improve/optimize the design, we have to
move to a numerical model [10, 11]. Furthermore, because the simulation itself is not a design tool, the
combined methodology of DOE applied to numerical simulations is proposed to size the equipment. This
technique greatly reduces the time spent in performing the simulations required to find the optimal
equipment as well as and a potential cost saving on the experimental if the prototype-model similarity
relations are met. Moreover, contrary to a sequential analysis, it is reasonable to use a mathematical and
statistical methodology that allows planning the sequence of experiments on the philosophy of maximum
information with minimum effort. For the implementation of DOE a series factors (mass of PCM, air
flow, air channel width, thickness of the PCM plate, finishing of the plates) and responses (additional
time for the air to reach a temperature of 44 C in the room, pressure drop, initial investment) were
considered. To meet the requirements of the proposed study case, the application of the DOE
methodology has been posed for a wider range rather than the experimental validity. Due to the specific
requirements of the study case, the similarity relationships between the experimental equipment and the
equipment proposed for the application will not be adjusted in all cases (model-prototype similarity).
Thus, calculation of the dimensionless relations of interest that characterize the equipment operation is
made for both the experimental equipment and for each of the simulated runs. So the degree of

230

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

dimensionless similarity of the proposed equipment with the experimental one is checked. Therefore the
need for additional experimentation of the proposed unit can be stated, in the case of not being within the
experimental validity range. The experimental validity range has been established for a series of
relationships that characterize the heat exchange process in the TES unit. Specifically, the following
ranges of values have been determined: the Reynolds number; the number of transfer units; the Biot
number; and the thermal conductivities ratio (eff/), that quantifies the effect of natural convection within
the PCM inside the plate. As this relation goes beyond 1 the effect of natural convection is more
substantial. The determination of all these parameters is detailed in [11], pp 115-8.
The series of restrictions put on the TES system are:
x Dimensions limitation due to the telecommunications shelter: the maximum length of the system is
limited to 2.5 m (height of the shelter) which limits the section of the PCM to 1.25 m. Likewise, the
width of the unit is also limited to 5 m due to the wall dimensions;
x Electrical power consumption limitation of the fan: so the fan can be electrically supplied by batteries
without being essential a connection to the grid. Pressure drop should be less than 30 Pa.
x Operating conditions are values to be considered inside the corresponding factor domains.
For the implementation of DOE the following factors and responses were considered:
x Factors (listed in table 1 along with their domain): mass of PCM, air flow, air channel width, thickness
of the PCM plate, finishing of the plates (related to rugosity or to bulges in the surface of the plates).
x Responses: melting ratio in 3 hours, additional time for the air to reach a temperature of 38 C
(compared with the evolution of temperature without unit TES) in the room, additional time for the air
to reach a temperature of 44 C (compared with the evolution of temperature without TES unit) in the
room, pressure drop; initial investment (mainly depending on the amount of PCM, the installed fan,
the casing, and whether or not the plates have bulges on its surface).
2.3. Response optimization
Response surface design methodology is often used to refine models after important factors have been
determined using factorial designs, especially if curvature in the response surface is suspected. Response
optimization helps to identify the combination of input variable settings that jointly optimize a single
response or a set of responses. This is useful when there is a need to evaluate the impact of multiple
inputs on a response. Given that the main objective of the TES unit is to extend the time period during
which the room temperature is below a certain temperature limit, the highest importance has set to that
response. Table 2 lists the input parameters in the optimization. It has been considered that the most
important requirement is to get the unit to extend as much as possible the time to reach the temperature
limit of the air in the room, assigning the greatest importance to the maximum temperature limit (44 C)
and considering also important, but lesser, the time to reach the first temperature limit (38 C) as well as
the p (in order to be as lower as possible so that the electrical power consumption of the corresponding
fan will be reduced). Also the investment and the melting ratio are interesting responses considered in the
study, as they are related to economical and technical feasibility respectively. Once the objectives are
defined, each variable is assigned a weight (between 0.1 and 10) and an importance (also between 0.1 and
10). In this approach to the optimization, each of the values of the responses is transformed using a
desirability function. The weight defines the shape of this function for each response and is related to the
emphasis on achieving the target (a value greater than one emphasizes the importance of achieving the
goal; a unit value gives equal importance to the objective and the limits; a value less than one puts less
emphasis on the goal). After calculating the desirability for each response, the composite desirability is
calculated (weighted geometric mean of the single ones) that allows to obtain the optimal solution. In this
case, the same weight is set to each of the answers assuming a unit value. This will set the target as

231

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

important as any value within the limits for the corresponding answer. On the other hand, assigning a
value to the importance of each answer is related to the importance given to each of the answers, and if
any of these responses is more important than the others (most important is a 10, less important is a 1).
Table 1. List of factors and their corresponding domain
MPCM [kg]
Domain

V

[m3/h]

eplate [mm]

eair [mm]

Finish

10

1.5

20

Level (-)

100

1000

Level (+)

200

2000

14

Table 2. Optimization parameters

Response variable

Objective

Weight

Importance

tadd T=44C

Maximize

10

minimize

tadd T=38C

Maximize

Investment

minimize

%Melt

Maximize

2.4. Model-prototype similarity

To check the degree of similarity between the experimental equipment and the different cases studied
in the numerical simulations, the following dimensionless groups that characterize the process were
calculated: Re, NTU, Bi, and the eff/ ratio.
3. Results and discussion
3.1. Empirical model results
Technical feasibility of incorporating the TES unit in the telecom shelter for temperature maintenance
is checked first with the empirical model. Fig. 3 shows that, taken into account the boundary conditions
shown in table 3, a reduction of the air temperature in the room can be provided with the TES unit,
increasing significantly the time to reach the threshold temperature (nearly 1 hour).
3.2. Theoretical model results: DOE & numerical simulations
Once the usefulness is checked, the theoretical model comes into play. The operating conditions of the
unit proposed by Lazaro, as well as the main simulation results obtained with the theoretical model
running with those conditions are shown in table 4.
3.3. Optimization
Then, the combined technique is applied in order to optimize the design using Minitab 15 [12]. The
Minitab Response Optimizer tool is used to show how different experimental settings affect the selected
responses for the response surface design. Minitab calculates an optimal solution and draws the
optimization plot. The optimal solution serves as the starting point for the plot (the settings can be

232

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

Temperature (C)

modified interactively to see how different setting affect responses). What is interesting of the optimized
results (Fig. 4) is the value of composite desirability as well as its trend according to each of the factors
considered. The composite desirability obtained in this case (0.919) indicates that the values determined
by the optimization nearly fulfill the requirements of the response variables. The trends of composite
desirability for each factor allow to adjust their value (usually due to physical or technological
constraints) while keeping high desirability values. However, at least there are two drawbacks to use this
configuration: first, it does not respect the width limitation (this unit has a width of more than 10 meters),
and secondly, when manufacturing the TES unit it will be more feasible to use a PCM thickness higher
than 0.5 mm (proposed in the optimization). Thus, moving in the optimization plot to a greater value of
PCM thickness without reducing too much the composed desirability and rounding parameters, a value of
2.5 mm in thickness is selected (which also meets the width restriction). Table 5 shows the results of the
corresponding simulation. The results of the last proposed unit are somewhat unfavorable compared to the
optimized unit, but the proposed thickness of PCM is much more realistic than the optimized one. Yet the
responses provided by the proposed unit represent a storage that improves the very first one every aspect
consider but the volume.
48
46
44
42
40
38
36
34
32
30
28
26
24
22
20

w TES
TTroom
room w TES
TTroom
roomw/o
w/oTES
TES

time (h)

Fig. 3. Empirical model results of Lazaros case

Fig. 4. Optiization plot results

233

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

Table 3. Boundary conditions Lazaros case

Generated
heat
5000 W

Room
temperature
27 C

Ventilation
airflow
2500 m3/h

Ambient
temperature
40 C

Envelope total
mass
2180 kg

cp envelope
(glass fiber)
880 J/(kgK)

Initial room
temperature
20 C

Table 4. Operating conditions and Main results of the simulation with Lazaros case
Results

Operating conditions
MPCM
[kg]
132

3
V [m /h]

eplate [mm]

eair [mm]

Finishing

%Melt

Investment []

1340

6.5

12

69.47

3924

tadd T=38C
[min]
36

tadd T=44C
[min]
63

p
[Pa]
36

Table 5. Main results of the proposed and optimized units

Unit
Proposed
Optimized

%Melt
92.64
100

Investment []
3489
3234

tadd, T=38C [min]

37
60

tadd, T=44C [min]

73
96

p [Pa]
5
3

3.4. Model-prototype similarity

As shown in Fig. 5, dimensional analysis of these units show that the natural convection within the
PCM is not going to be significant in any of the 2 units (proposed and optimized), being the heat transfer
process by pure conduction for the second unit, and the ratio eff/ within the range of experimental
validity for the other one [11]. Furthermore, since both Re, Bi, and NTU are within the experimental
validity range, the units can be used for design purposes.

Fig. 5. Re numbers (up left), Bi numbers (up right), NTU numbers (down left), and eff/ ratio (down right) for each simulation and
experimental validity range

234

Pablo Dolado et al. / Energy Procedia 30 (2012) 225 234

4. Conclusions
The combined methodology of Design of Experiments applied to the numerical simulations seems to
be a valid tool for improving the design of this kind of heat exchangers. When applied to the case study of
temperature maintenance in a room and compared to the very first design based on an empirical model:
time to reach the maximum air temperature in the room was increased (19.7%), the initial investment was
reduced by 11% and the PCM melting ratio was improved by 23.2%, as a drawback, the volume occupied
by the unit was increased around 3 times. When applied to the study case, this methodology shows an
improvement of the proposed design compared to the one initially posed: the time to reach the threshold
air temperature has been extended, the initial investment has been reduced, and the melting degree of the
system has been improved; however, as a drawback, the TES unit volume has been increased.
Furthermore, the proposed methodology of applying the DOE to the numerical simulations has
demonstrated to be a powerful tool to: reduce the number of numerical simulations runs and the invested
time; improve the design; find an optimal point of operation; and find the factors settings that enable to
fulfill with the requirements.
Acknowledgements
The authors would like to thank the Spanish Government for the partial funding of this work within the
research project ENE2011-28269-C03-01. Special thanks are extended to Mr. Miguel Zamora, CIAT
R&D Manager. Thanks also to the fruitful collaboration within the Annex 24-Task 42 of the ECES
Implementing Agreement of the IEA, as well as in the Cost Action TU0802 (NeCOEPCM).
References
[1] Gunasegaram DR, Smith BJ. MAGMASOFT helps assure quality in a progressive Australian iron foundry. Proceedings of
the 32nd National Convention of the Australian Foundry Institute 200; Fremantle (Australia), 99104.
[2] Calise F, Palombo A, Vanoli L. Maximization of primary energy savings of solar heating and cooling systems by transient
simulations and computer design of experiments. Appl Energ 2010;87:524.
[3] Gupta A, Ding Y, Xu L, Reinikainen T. Optimal parameter selection for electronic packaging using sequential computer
simulations. J Manuf Sci Eng, ASME 2006;128:70515.
[4] Del Coz Diaz JJ, Garcia Nieto PJ, Lozano Martinez-Luengas A, Suarez Sierra JL. A study of the collapse of a WWII
communications antenna using numerical simulations based on design of experiments by FEM. Eng Struct 2010;32,7:1792-800.
[5] Whitt W. Analysis for the design of simulation experiments. New York: Columbia University; 2005.
[6] Salazar JC, Baena Zapata A. Analysis and design of experiments applied to simulation studies. Dyna 2009;76,159:249-57.
[7] Gunasegaram DR, Farnsworth DJ, Nguyen TT. Identification of critical factors affecting shrinkage porosity in permanent
mold casting using numerical simulations based on design of experiments. J Mater Process Technol 2009;209,3:120919.
[8] Lazaro A. Thermal energy storage with phase change materials. Building applications: materials characterization and
experimental installation to test PCM to air heat exchanger prototypes. Thesis, University of Zaragoza, Spain; 2009.
[9] Lazaro A, Dolado P, Marin JM, Zalba B. PCM-air heat exchangers for freecooling applications in buildings: empirical model
and application to design. Energy Conv Manag 2009;50:444-9.
[10] Dolado P, Lazaro A, Marin JM and Zalba B. Characterization of melting and solidification in a real scale PCM-air heat
exchanger: Numerical model and experimental validation. Energy Conv Manag 2011;52:1890-907.
[11] Dolado P. Thermal Energy Storage with phase change. Design and modelling of storage equipment to exchange heat with
air. Thesis University of Zaragoza, Spain; 2011.
[12 ] Minitab 15 Users guide. Access date 2012/06/21 at: http://www.minitab.com/en-US/support/