You are on page 1of 8

Energy and Buildings 38 (2006) 258265

www.elsevier.com/locate/enbuild

A new PCM storage system for managing


simultaneously solar and electric energy
Zouhair Ait Hammou *, Marcel Lacroix
Faculte de Genie, Universite de Sherbrooke, Sherbrooke, Que., Canada J1K 2R1
Received 14 March 2005; received in revised form 25 May 2005; accepted 15 June 2005

Abstract
A new hybrid thermal energy storage system (HTESS), using phase change materials, is proposed for managing simultaneously the storage
of heat from solar and electric energy. Solar energy is stored during sunny days and released later at night or during cloudy days and, to smooth
power demands, electric energy is stored during off-peak periods and later used during peak periods. A heat transfer model of the HTESS is
developed and validated with experimental data. Simulations carried out for a period of 4 consecutive winter months indicate that, with such a
system, the electricity consumption for space heating is reduced by nearly 32%. Also, more than 90% of the electric energy is consumed
during off-peak hours. For electricity markets where time-of-use rate schemes are in effect, the return on the investment in such a thermal
storage system is very attractive.
# 2005 Elsevier B.V. All rights reserved.
Keywords: Thermal energy storage; Phase change materials; Solar energy; Electric energy; Packed bed

1. Introduction
In northern countries such as Canada, space heating
represents the main source of electricity consumption in new
homes. As a result, during harsh winter days, the demand for
electric energy is increasingly acute in the morning and in the
evening and quite often the distribution grid becomes
overloaded. This severe problem has created the need to shift
some of the on-peak demand to off-peak periods by making
use of electric storage systems. In these systems, electric
energy is conveniently converted into thermal energy and
stored in a material during the night and subsequently used the
next day. During on-peak periods, the current in the system is
automatically shut off and the storage unit discharges its heat
to the living space. Of course, to agree to bear the additional
cost involved in purchasing and installing a storage system, the
customer benefits from lower electricity rates during off-peak
periods allowing him to obtain a return on his investment.
* Corresponding author.
E-mail addresses: Zouhair.ait.hammou@USherbrooke.ca
(Z. Ait Hammou), Marcel.Lacroix@Usherbrooke.ca (M. Lacroix).
0378-7788/$ see front matter # 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.enbuild.2005.06.008

An increasing number of new homes are also equipped


with solar systems, passive or active, to further alleviate the
problem of electricity consumption for space heating. Since
the availability of solar energy is usually not coincident with
the demand, heat collected from solar radiation is also stored
in a thermal unit. For solar systems, however, heat is stored
during sunny days and released at night or later during
cloudy days. When both storage systems, electric and solar,
are employed simultaneously, conflicting situations may
arise resulting in overheating of the living space and/or poor
thermal performance of the units. Despite the fact that solar
systems have been the subject of many investigations in the
past [19], none of these studies has addressed the problem
of simultaneous thermal storage of solar and electric energy.
The objective of the present paper is therefore to propose
a system that can store and manage simultaneously and
efficiently heat from solar and electric sources. In this
system, called hybrid thermal energy storage system
(HTESS), solar energy is stored during sunny days and
recovered later at night or during cloudy days while thermal
electric energy is stored during off-peak periods and
recovered during on-peak periods.

Z. Ait Hammou, M. Lacroix / Energy and Buildings 38 (2006) 258265

Nomenclature
A
Cp
f
h
hsl
H
k
L
m
m

n
P
Pr
Qload
Qprov
Qrest
Qu
r
R
R0
Re
S00
t
T
Tref
T1
u
U
V
x
y

surface (m2)
heat capacity (J kg1 K1)
liquid fraction
convective heat transfer coefficient
(W m2 K1)
latent heat of fusion (J kg1)
bed height (m)
thermal conductivity (W m1 K1)
bed width (m)
mass of air (kg)
mass flow rate (kg s1)
number of spheres in a given control volume
bed depth (m)
Prandtl number
energy load (J)
energy provided (J)
electric energy needed to maintain the room
temperature at the set point (J)
useful energy output of the solar collector (J)
radial coordinate (m)
outside radius of capsule (m)
inside radius of capsule (m)
Reynolds number
absorbed solar radiation (W m2)
time (s)
temperature (K)
melting point (K)
outside temperature (K)
coolant velocity (m s1)
overall heat transfer coefficient (W m2 K1)
control volume in bed (m3)
transversal coordinate of HTESS (m)
axial coordinate of HTESS (m)

sp
w

259

set point
wall

Superscript
o
old

A mathematical model for predicting the thermal


performance of the HTESS is first presented. The model
is next validated with experimental data obtained from a
prototype. The thermal performance of the system is then
assessed for two different phase change materials (PCMs) by
conducting simulations that rely on meteorological data for a
period of 4 consecutive winter months.

2. Modelling the HTESS


A schematic of the HTESS is depicted in Fig. 1. The
HTESS consists chiefly of an interior wall of height H, depth
P and width L filled with capsules of radius R containing a
PCM. Heat is transferred via a stream of hot air, sometimes
referred as the coolant, coming from a solar collector and/or
from an electric heater. Heat is also collected from direct
solar radiation impinging on the wall. Openings at the
bottom and at the top of the wall allow the circulation of air
from the room. The mass flow rate inside the storage wall is
considered large enough so that heat transfer in the coolant is
convection dominated in the axial direction. On the other
hand, radial conduction heat transfer prevails inside the
capsules.

Greek symbols
e
porosity
u
temperature (K)
r
density (kg m3)
Subscripts
a
air
c
coolant
eff
effective
ex
exterior
in
inlet
j
layer number
l
liquid
n
capsule number in control volume
N
total number of horizontal layers
out
outlet
r
room
s
solid
Fig. 1. Schematic of the hybrid thermal energy storage system.

260

Z. Ait Hammou, M. Lacroix / Energy and Buildings 38 (2006) 258265

By dividing the bed of capsules into control volumes of


dimensions 2R  2R  P, the energy conservation equations
for the coolant, the PCM, the capsule wall and the adjacent
room may be stated as


@Tc x; y; t
@Tc x; y; t
u
coolant; erc Cpc
@t
@y
@2 Tc x; y; t
keff
hc Ac uw R; t  Tc x; y; t
(1)
@x2


@ur; t 1 @
@ur; t
@ f r; t
2
kr 2
;
PCM; rCp
 dH
@t
r @r
@r
@t
0 < r < R0

(2a)

where
dH rCpl  Cps ur; t  uref rhsl
capsule wall;



@uw r; t 1 @
2 @u w r; t
2
kw r
rw Cpw
;
@t
r @r
@r

R <r<R
mr Cpr

room;

N
X

(2b)

(3)
@Tr
mC
pr Tcout  Tr Uex Aex T1  Tr
@t

UA j Tc j x 0  Tr

(4)

j2

where Tc(x, y, t), u(r, t), uw(r, t) and Tr(t) are the unknown
coolant, phase change material, capsule wall and room
temperatures, respectively. u , in Eq. (1), is the mean coolant
velocity, e the porosity of the bed, keff the effective thermal
conductivity of coolant defined as keff = 0.5RecPrckc [10], Ac
2
the surface to volume ratio per capsule, i.e. Ac n4pR
Vc , and
hc is the external heat transfer coefficient estimated with an
empirical correlation proposed by Beek [11]. Tcout is the
average temperature at the outlet of the bed.
The first term of the right-hand side of Eq. (1) represents
the heat diffusion in the transversal direction (ox) while the
second term accounts for the heat transfer between the n
capsules of the control volume and the surrounding coolant.
The second term on the right-hand side of Eq. (2a) accounts,
via the liquid fraction f, for the solid/liquid phase change
that takes place inside the PCM. Finally, the second term on

the right side of Eq. (4) represents the heat losses from the
room to the outside atmosphere at temperature T1.
The boundary conditions for the above set of equations
are:
For the coolant,
Tc y; t Tcin t
keff

at y 0

@Tc
S00 UTr  Tc
@x

(5)
at x 0

@Tc
0 at x L
@x
For the PCM and the capsule wall,
@ur; t
0
@r
kw
(

at r 0

@uw r; t
@ur; t
k
@r
@r
uw r; t ur; t

(7)

(8)

@uw r; t
hc uw r; t  Tc at r R
@r

kw

(6)

at r R0

(9)
(10)

3. Numerical solution
The above set of equations (Eqs. (1)(4)) is solved using a
finite difference solution method. Using an upwind scheme
for the convection term and a fully implicit scheme for time
integration, discretization of Eq. (1) over each of the control
volumes depicted in Fig. 2 yields
0
acP TcP acE TcE acW TcW acN TcN a0cP TcP
bc

with

8
acP a0cP acN acE acW Dxhc Ac
>
>
>
>
Dx
>
>
a0cP erc Cpc
>
>
Dt
>
>
>
keff
>
>
< acE
Dxe
keff
>
>
> acW
>
>
Dx
w
>
>
>
Dx
>
>
a

er

>
cN
c Cpc u
>
>
Dy
:
bc Dxhc Ac uw R; t

Fig. 2. (ac) Schematic of the bed control volumes and the PCM capsules.

(11)

(12)

Z. Ait Hammou, M. Lacroix / Energy and Buildings 38 (2006) 258265

261

Eqs. (11) and (12) are applied to all internal control volumes.
For boundary conditions, the same equations are used except
for Eq. (12) which is modified as follows:
8
0
>
< acP acP acN acE acW Dxhc Ac U
atx 0
acW 0
>
:
bc Dxhc Ac uw R; t UTr S00
(13)

atx L

8
0
>
< acP acP acN acE acW Dxhc Ac U
(14)
acE 0
>
:
bc Dxhc Ac uw R; t

For a given temperature of the coolant at the inlet of the


storage wall (j = 1), and a given surface temperature of the
spherical capsules inside layer j = 2, the coolant temperature
distribution for layer j = 2 is obtained by solving the system
of Eqs. (11)(14) using a tri-diagonal matrix algorithm
(TDMA). The calculations then proceed to the next layer
until layer N is reached.
Eqs. (2a) and (2b) are discretized in the similar way
yielding,
aP uP aE uE aW uW aoP uoP Ss

(15)
Fig. 3. Overall calculation procedure.

with
ke ure2
kw urw2
;
aW
;
dre
drw
aP aE aW aoP

aoP r2 rCp u

aE

Dr
;
Dt
(16)

and
Ss r2 dHP f  f o

Dr
;
Dt

re rw
2

(17)

The system of Eq. (15) is also solved using a tri-diagonal


matrix algorithm.
The central feature of the present fixed grid phase change
method is the source term Ss. This source term keeps track of
the latent heat evolution, and its driving element is the liquid
fraction f. Its value is determined iteratively from [12]
fpm1 fpm v

ap
Dt
up  uref
Dr
r2 dHP

(18)

where v is an under-relaxation factor. The liquid fraction


update (Eq. (18)) is applied at every node after the (m + 1)th
solution of the linear system Eq. (15) for u. Since Eq. (18) is
not adequate for every node, the correction
!
0; if fpm1 < 0
m1
fp

(19)
1; if fpm1 > 1
is applied immediately after Eq. (18). Further details concerning the numerical implementation of the present fixed
grid phase change method may be found in the work of
Voller [12].

The integration of Eq. (3) is similar to that for Eqs. (2a)


and (2b) except for the source term. The effect of natural
convection that takes place inside the encapsulated molten
material is taken into account by using an effective thermal
conductivity given by [13].
Finally, integration of Eq. (4) yields
!
N
X
1
o
Tr BT1 CTcout D A j Tc j x 0
(20)
Tr
ga
j2
where
PN
m
Dt Uex Aex Dt U Dt j2 A j
ga 1

;
mr
mr Cpr
mr Cpr
m
Uex Aex Dt
U Dt
Dt
B
;
C
and D
mr
mr Cpr
mr Cpr

(21)

The overall calculation procedure is summarized in Fig. 3.

4. Experiment and validation


The above mathematical model was first validated with
analytical and numerical results available in the open
literature. Some of these validation results are reported in
reference [14].
An experimental rig was also set up to further check the
validity of the numerical model. The experimental prototype
of the HTESS, shown in Fig. 4, scale 1:6, consists of a water

262

Z. Ait Hammou, M. Lacroix / Energy and Buildings 38 (2006) 258265

Fig. 4. Schematic of the experimental rig: (1) cold water source; (2) hot water source; (3 and 4) tanks; (5 and 6) valves; (7) flow meter; (8) water jacket; (9) bath;
(10) ribbon heaters; (11 and 12) thermocouples; (13) DAQ; (14) computer; (15) flux meter; (16) PCM capsules. ()) Cold water flow in the water jacket; (!) hot
water flow in the storage wall.

bath 0.5 m high, 0.5 m deep and 0.15 m wide with an


adjoining 0.035 m thick storage wall made of aluminum.
Spherical capsules, 0.0285 m diameter filled with n-octadecane, are inserted into the storage wall cavity to simulate the
storage medium. The water bath represents the room. Thermal
energy is stored in the wall by: (1) hot water flowing through
the cavity of the storage wall, in order to simulate the effect of
heat transfer from a solar collector or from electric heaters,
and (2) by means of ribbon copper heaters covering uniformly
its external surface, in order to mimic the effect of direct solar
radiation. Furthermore, to represent heat losses through
windows, a water jacket made of aluminum is employed. The
electric power dissipated in the heaters is controlled by a

power supply. Temperatures are measured by means of 12


copperconstantan thermocouples (type E) located in the
water bath, in the water jacket and distributed in the storage
wall. Four flux meters and a flow meter are also used to
estimate the surface heat fluxes and water flow rate,
respectively. The thermocouples, the flux meters, the flow
meter and the electric heaters are all connected to a data
acquisition system (32 channel multiplexer amplifiers and 8
channel 12 bits A/D converter). The thermocouples, the flux
meters and the flow meter are accurate within 0.5 8C, 5%
and 5% of the maximum value, respectively.
As an example, an experiment lasting 1.6 h was conducted.
During the first 0.8 h, hot water at 333 K flows at a mass flow

Fig. 5. Predicted and measured temperatures of the water bath.

Fig. 6. Predicted and measured temperatures of the coolant 6.5 cm from the
inlet of the storage wall.

Z. Ait Hammou, M. Lacroix / Energy and Buildings 38 (2006) 258265

263

rate of m
0:1 kg s1 through the storage wall where heat is
transferred to the PCM capsules. For the last 0.8 h, the water
circulation is stopped and the ribbon copper heaters are turned
on releasing 500 W. For the entire duration of the experiment,
a flow of cold water (278 K < T < 283 K) is maintained
through the water jacket.
The above experiment was simulated with the numerical
model and the predictions were compared to the experimental
data (Figs. 5 and 6). Examination of these figures reveals that
the agreement between the experiment and the simulation is
excellent. The slight discrepancy shown in Fig. 6 is due to
the under-estimation of the predicted heat losses as adiabatic
boundaries were assumed in the numerical model.

5. Results and discussion


A series of numerical simulations was conducted in order
to investigate the effect of various storage materials and of
the operating conditions on the thermal behaviour and
performance of the HTESS. Since some of these results are
reported in reference [14], they will not be repeated here.
They will serve, however, as the starting point for the present
discussion.
The system under study consists of a room of 5 m 
5 m  3 m with a storage wall of 0.192 m thick comprising
spherical capsules, 0.064 m in diameter, made of polyethylene filled with capric acid or n-octadecane. The
thermo-physical properties of these materials are listed in
Table 1.
The simulations were carried out for a time period of four
consecutive winter months using meteorological data for the
city of Montreal, Canada [15]. These data are depicted in
Fig. 7.
The initial temperature of the HTESS was set equal to
294 K. The strategy retained for the operation of the HTESS
is as follows.
During off-peak hours, i.e. from 00:00 h to 07:00 h, the
room air follows path DEB in Fig. 1. It passes through the
auxiliary energy source at a mass flow rate m
0:08 kg s1
where its temperature increases from Tr to Tcin 333 K. It
then flows through the wall where its heat is transferred to
the storage material. The total amount of electric energy
consumed during off-peak hours is estimated as
X
X
mC
Qheat t
(22)
Qoff
pr Tcin  Tr Dt
off -peak
off -peak
During peak hours, i.e. from 07:00 h to 00:00 h, space
heating relies on direct solar radiation only. Moreover, if

Fig. 7. Hourly solar radiation on the storage wall and outside air temperature during 4 consecutive winter months in Montreal, Canada.

the temperature of the air flow emerging from the solar


collector is larger than that of the room temperature, then it
is allowed to pass through path AFC (Fig. 1). Otherwise, the
room air flows through the wall via path CB where it
retrieves some of the stored heat.
Sometimes, during sunny days, the energy stored during
the previous off-peak period can cause overheating of the
room. To avoid such a situation, the amount of energy that
needs to be stored during off-peak hours for the next day but
would not lead to overheating, Qneed, may be estimated from
the weather forecasts:
Qneed Qload  Qavail

24
24
X
X
Uex Aex Tsp  T1 Dt 
S00 A Qu Dt
1

(23)
where Qload represents the energy load needed to maintain
the set point temperature Tsp and Qavail is the available solar
energy for the following day assessed with the meteorological predictions.
When Qneed > 0, this amount of energy is stored over a
period of time estimated as
Dt

Qneed
mC
pr 333  Tsp

(24)

If Dt exceeds the off-peak period of 7 h, then the storage time


is limited to 7 h only.

Table 1
Physical properties of materials

n-Octadecane [14]
Capric acid [16]
Polyethylene [14]

Cpl (J kg1 K1)

Cps (J kg1 K1)

kl (W mK1)

ks (W mK1)

hsl (kJ kg1)

Tref (K)

r (kg m3)

2231
1720

1891
1950
2090

0.15
0.16

0.38
0.16
0.33

243.5
153.0

301.3
303.1

773.2
878.0
960.0

264

Z. Ait Hammou, M. Lacroix / Energy and Buildings 38 (2006) 258265

Fig. 8. Daily storage time.

Fig. 10. Total liquid fraction inside the HTESS.

Otherwise, when Qneed  0, no energy storage is


required.
Fig. 8 shows the duration of the daily storage period for
the 4 consecutive winter months. It is seen that, in order to
prevent overheating of the room during the following day,
24% of the off-peak periods have not been fully exploited. In
other words, the present control strategy, based on weather
forecasts, saves electric energy that would otherwise be
ventilated outside the next day as excess stored heat.
The total consumption of electricity Qelec may be
expressed as the sum of the electric energy consumed
during off-peak periods Qoff and the electric energy needed
to maintain the air temperature inside the room at the set
point Tsp, Qrest

Qrest is estimated as follows: if the room air temperature Tr(t)


happens to exceed the set point Tsp, heat must be evacuated
from the room and Qrest = 0. If, on the other hand, Tr(t)
becomes smaller than Tsp, an additional amount of electric
energy Qrest must be spent
X

Qrest

Qload t  Qprov t

(26)

4 days

where Qprov(t) is the energy provided by the HTESS:


Qprov t
"
mC
pr Tcout  Tr

#
UA j Tc j x 0  Tr Dt

Qelec Qoff Qrest

(25)

Fig. 9. Room temperature and outside temperature T1 for 10 consecutive


days.

(27)

Fig. 11. Temporal variation of Qload, Qprov and Qrest.

Z. Ait Hammou, M. Lacroix / Energy and Buildings 38 (2006) 258265


Table 2
Electricity consumption for space heating for 4 consecutive months
HTESS

Qelec
(off-peak hours)
(kWh)

Qelec
(peak hours)
(kWh)

Qelec
(total)
(kWh)

Capric acid
n-Octadecane
Without HTESS

2642
2617

251
282

2893
2899
4277

265

consecutive winter months. Results indicate that, by using a


HTESS, the electricity consumption for space heating is
reduced by 32%. Also, more than 90% of the electricity is
consumed during off-peak hours. For electricity markets
where time-of-use rate schemes are in effect, the return on
the investment in such a storage system is very attractive.

Acknowledgements

Table 3
Energy cost with time-of-use rates
HTESS

Cost
(off-peak hours)
(monetary unit)

Cost
(peak hours)
(monetary unit)

Cost (total)
(monetary unit)

Capric acid
n-Octadecane
Without HTESS

2642  1
2617  1

251  2
282  2

3144
3181
6340

As an example, Figs. 911 display the thermal behaviour of


the HTESS for a period of 10 consecutive days. Fig. 9
reveals that the thermal behaviour is nearly the same for
capsules filled with capric acid and for capsules filled with noctadecane. Fig. 10 shows, however, that latent heat storage
is, in the present case, better exploited with capsules filled
with n-octadecane as the melting point for this PCM is lower
than that of capric acid. It also suggests that if capric acid
capsules were to be used as the PCM, the width of the
storage wall L should be smaller. On the other hand, Fig. 11
exemplifies the merits of the present control system. Qprov
sometimes exceeds Qload. Sometimes, it is the other way
around. Therefore, to maintain the room temperature at the
set point, the control system must provide the right amount
of energy given by Qrest.
The total electric energy consumptions for space heating
over the 4 month period are summarized in Table 2. These
results show a reduction of 32% in the electric energy
consumption for a room with a HTESS. Also, more than
90% of the electricity is consumed during off-peak hours.
Table 3 translates these energy savings into monetary units.
Using a typical time-of-use rate scheme for which the peak
hour rates are twice those of off-peak hours and assuming a
flat rate of 1.5 monetary units when no heat storage is
employed, it is seen that the use of a HTESS can cut the
electricity bill by 50%.

6. Concluding remarks
A new hybrid thermal energy storage system was
proposed for managing simultaneously the storage of heat
from solar and electric energy. A heat transfer model of the
HTESS was developed and validated with experimental data
obtained with a prototype. Simulations were carried out
with two different PCMs using meteorological data for 4

The authors gratefully acknowledge the financial support


of the Ministe`re des Ressources naturelles du Quebec
(project PADTE) and of the Natural Sciences and
Engineering Research Council of Canada.

References
[1] D. Beckman, Solar Engineering of Thermal Processes, second ed.,
1991.
[2] M. Mbaye, E. Bilgen, Natural convection and conduction in porous
wall, solar collector systems without vents, Transactions of the ASME
114 (1992) 4046.
[3] S.S. Hsieh, J.T. Tsai, Transient response of the Trombe wall temperature distribution applicable to passive solar heating systems, Energy
Conversion and Management 28 (1988) 2125.
[4] M.F. Abd Rabbo, S.K. Adam, Trombe wall heat transfer analysis,
Journal of Solar Energy Research 6 (1988) 119.
[5] M. Telkes, Thermal storage for solar heating and cooling, in: Proceedings of the Workshop on Solar Energy Storage Subsystems for the
Heating and Cooling of Buildings, VA, USA, 1975.
[6] U. Stritih, Heat transfer enhancement in latent heat thermal storage
system for buildings, Energy and Buildings 35 (2003) 10971104.
[7] D.J. Close, R.V. Dunkle, K.A. Robeson, Design and performance of
thermal storage air conditioner system, Mechanical and Chemical
Engineering Transaction, Institute of Engineers, Australia, MC4, vol.
45, 1968.
[8] G.O.G. Lof, R.A. Tybout, Cost of house heating with solar energy,
Solar Energy 14 (1973) 253.
[9] B.L. Evans, S.A. Klein, Combined active collectionpassive storage
and direct gain hybrid space systems, in: ASME Meeting, 1984.
[10] K.A.R. Ismail, R. Stuginsky Jr., A parametric study on possible fixed
bed models for PCM and sensible heat storage, Applied Thermal
Engineering 19 (1999) 757788.
[11] J. Beek, Design of packed catalytic reactors, Advances in Chemical
Engineering 3 (1962) 203271.
[12] V.R. Voller, Fast implicit finite-difference method for the analysis of
phase change problems, Numerical Heat Transfer 17 (1990) 155169.
[13] G.D. Raithby, K.G.T. Hollands, A general method of obtaining
approximate solution to laminar and turbulent free convection problems, Advances in Heat Transfer 11 (1975) 266315.
[14] Z.A. Hammou, M. Lacroix, A hybrid thermal energy storage system
for managing simultaneously solar and electric energy, Energy Conversion and Management, in press.
[15] http://www.eere.energy.gov.
[16] Z. Yinping, J. Yi, J. Yi, A simple method, the T-history method, of
determining the heat of fusion, specific heat and thermal conductivity
of phase-change materials, Measurement Science and Technology 10
(1999) 201205.

You might also like