Energy and Buildings 75 (2014) 342357

Contents lists available at ScienceDirect

Energy and Buildings

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

Optimal management proposal for hybrid water heating system

Oussama Ibrahim a,c , Farouk Fardoun a , Rac Younes b, , Hasna Louahlia-Gualous c
a

University Institute of Technology, Department GIM, Lebanese University, Saida, Lebanon

Faculty of Engineering, Lebanese University, Beirut, Lebanon
c
Universit de Caen Base Normandie, LUSAC, 120 rue de lexode, 50000 Saint L, France
b

a r t i c l e

i n f o

Article history:
Received 16 May 2013
Received in revised form 26 January 2014
Accepted 7 February 2014
Keywords:
Hybrid water heating systems
Numerical modeling
Heat pump water heater
Solar collector
Energy optimization
Management

a b s t r a c t
This paper presents the dynamic modeling of a domestic hybrid water heating system. The system is
composed of a solar collector, a heat pump water heater, a wind turbine, a battery and a hot water storage tank. Both air-source and geothermal-heat pumps are investigated. Detailed mathematical dynamic
models of the individual components are presented and validated. Simulations for typical days in summer
and winter for Beirut and Cedars, two Lebanese locations with different meteorological and demographic
conditions, are conducted using Matlab software. A renewable coverage factor (RCF) is dened, representing the renewable energy share with respect to the total delivered energy. Results reveal that the
proposed hybrid system is capable of securing all hot water needs in all case studies and RCF is always
above 63%. Furthermore, an energetic, economic optimal management model is developed for the proposed hybrid system. It is applied to the considered case studies, where results illustrate the optimum
size of each of the system components as well as the optimum energy-ow distributions among them
over two investigated time periods, ve and ten years. It is noticed that the initial cost of the hybrid
system is acceptable and important annual savings are obtained.
2014 Elsevier B.V. All rights reserved.

1. Introduction
Life is facing three major problems, which are negatively affecting health, peace and security. They are: (1) the dramatic increase
in the energy demand corresponding to the global population evolution as well as the development of living standards, (2) the great
consumption increase of the current primary energy source-the
non-renewable fossil fuels and (3) the aggravation of global warming and environmental pollution as a result of the preceding two
stated problems.
Lebanon is a developing country that meets the majority of its
energy needs from oil imports. The country suffers a huge shortage in the electricity generation which is compensated through
thousands of backup self-generators that are estimated to represent up to 30% of all electricity generated [1]. Thus, end use
electrical energy management is dispensable in this country,
because it can participate in minimizing the energy consumption

Corresponding author.
E-mail addresses: oibrahim@etu.unicaen.fr (O. Ibrahim), ffardoun@ul.edu.lb
(F. Fardoun), ryounes@ul.edu.lb (R. Younes), hasna.gualous@unicaen.fr
(H. Louahlia-Gualous).
http://dx.doi.org/10.1016/j.enbuild.2014.02.023
0378-7788/ 2014 Elsevier B.V. All rights reserved.

and consequently, reduce the operating expenditure and the local

environmental pollution.
Water heating is a major energy consumer all around the world.
For instance, its share of the total residential energy consumption
is about 11% in USA [2], 14% in Europe [3], 22% in Canada [4], 25% in
Australia [5], 25% in Russia [6], 30% in Japan [7], 29% in Mexico [8],
27% in China [9], 32% in South Africa [10], etc. Although no clear
statistics for energy consumption corresponding to the domestic
water heating in Lebanon, it is known that the country almost has
$4010 per capita GNI (Gross National Income) [11], which reects
the good standards of living and consequently the large energy
demand. This fact together with the illustrated worldwide statistics reveal that the residential hot water production constitutes an
important portion of the total consumed energy in Lebanese homes.
Furthermore, 75% of residents rely on electric heaters for domestic
hot water production [12], which add an unsustainable burden on
the Lebanese electricity sector. Therefore, an important portion of
the residential electrical consumption may be saved by the proper
choice and management of the domestic water heating system.
This paper presents the detailed numerical modeling of a proposed hybrid water heating system. The performance of this system
is investigated for two case studies in Lebanon, where one is located
in the capital-Beirut, a coastal metropolitan city, and the other
is in Cedars, a mountainous town. The main reason for choosing

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

Nomenclature
A
a
C
Cb
COP
Cp
D
dt
E
E
g
g(t)
h
H
h(t)
I
k
L
Lc
m

m
P
P
Q
Q
R
r
T
t
Ut
Ubp
Ubw
Upw
u
V
W
x, y, z
Z0
a









area [m2 ]
thermal diffusivity [m2 /s]
battery capacity [Wh]
bond conductivity [W/mK]
coefcient of performance
specic heat [J/kgK]
diameter [m]
time step [s]
electrical energy [kWh]
electrical capacity [W]
gravitational acceleration [m/s2 ]
hot water ow rate from storage tank [kg/s]
enthalpy [J/kg]
height/ground heat exchanger depth[m]
domestic hot water demand prole [kg/s]
solar radiation [W/m2 ]
thermal conductivity [W/mK]
length [m]
characteristic length [m]
mass [kg]
mass ow rate [kg/s]
pressure [Pa]
power [W]
heat energy [kWh]
heat capacity [W]
thermal resistance [m2 K/W]
radius [m]
temperature [K or C]
time
overall loss heat transfer coefcient from the top of
collector [W/m2 K]
loss heat transfer coefcient from the bottom of the
absorber plate [W/m2 K]
loss heat transfer coefcient from the bottom of the
tube wall [W/m2 K]
overall heat transfer coefcient between absorber
plate and tube-wall [W/K]
velocity [m/s]
volume [m3 ]
width [m]
spatial coordinates
surface roughness length [m]
absorptance of coating
convective heat transfer coefcient [W/m2 K]
radiative heat transfer coefcient [W/m2 K]
combined-convective and radiative-heat transfer
coefcient [W/m2 K]
coefcient of thermal expansion [1/K]
thickness [mm]
emittance
dynamic viscosity [Pa s]
density [kg/m3 ]
StefanBoltzmann constant
transmittance of glass cover

Subscripts
1,2,3
1st, 2nd, 3rd region
a
air
b
borehole
battery
bat
c
condenser
consume consumed water

d
cs
e
eq
gr
HP
HTF
i
L
l
o
p
r
s
st
tap
u
w
WB
wat

343

downward
cross section
evaporator
equivalent
grout
heat pump
heat transfer uid
in, inside
load
lower part
out
absorber plate
refrigerant
soil
storage
tap water
upper part/upward
wall
wind-battery system
water in storage tank

Dimensionless numbers
Symbol

Explanation

Relation

Gr

Grashof number

Gr =

Nu

mean Nusselt number

Nu =

Pr
Ra
Re

Prandtl number
Rayleigh number
Reynolds number

Pr =
Ra = Gr Pr
uD
Re = 

Lc3 g2 |T |

D
k
Cp
k

2

these case studies is the evaluation of the proposed system in different climatic and demographic conditions, which represent the
extreme conditions of the country. Furthermore, an optimal management model for the presented system is suggested and solved
for these two case studies. The main objective of this model is the
optimization of: the energy-ow distributions among the system
components, the size of each component and system cost.
2. Literature review
A Taiwanian research group presented and studied an integraltype solar assisted heat pump water heater (ISAHP) [1317]. In
this design, the storage tank and the Rankine cycle units were integrated together to make a more compact size and a thermosyphon
loop was used to transfer heat from the condenser to the water
storage tank. Furthermore, the condenser was of tube-in-sheet,
unglazed type and the solar collector was itself the evaporator.
The ISAHP absorbs energy from solar radiation and ambient air
simultaneously. It was found that the thermal performance of
an ISAHP is marginally inuenced by the variation of ambient
temperature. Guoying et al. [18] performed a simulation study of
solar-air source heat pump water heater with a specially designed
at-plate heat collector/evaporator with spiral-nned tubes. The
designed collector/evaporator was exposed directly to both ambient air and solar radiation and hence, it could collect solar energy
through the plate and ambient-air energy through the natural
convection between the tubes and the ambient. It was proved that
this system could produce 55 C-hot water all around the year,
overcoming the problem of direct expansion solar assisted heat
pump (DX-SAHP) that fails in rainy days. Anderson and Morrison
[19] studied a solar-boosted heat pump heater with at unglazed
aluminum solar evaporator panels to absorb solar and ambient
energy. A wrap-around condenser coil on the outside of the water

344

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

tank was adopted. They proved that the system performance could
be improved by concentrating the condenser coils in the lower
portion of the tank. Nuntaphan et al. [20] studied the performance
of an indirect solar assisted heat pump water heater using the
refrigerant mixture R22/R124/R152a, where hot water leaving the
solar collector is circulated into a water-heated evaporator. It was
noticed that the hot water temperature is 40% greater than that
obtained from conventional solar water heating systems. Li and
Yang [21] developed a simplied mathematical model of solar
assisted air source heat pump system which mainly consists of
two loops: the solar collector loop and the airwater heat pump
unit. The effects of various parameters, including circulation ow
rate, solar collector area, solar collector tilt angle and initial water
temperature in the preheating solar tank, are investigated. The
results show that the system performance is governed strongly by
the change of circulation ow rate, solar collector area and initial
water temperature in the preheating solar tank. Huang et al. [22]
designed, built and tested a heat-pipe enhanced solar-assisted
heat pump water heater (HPSAHP). It is a heat pump with dual
heat sources that combines the performance of conventional solar
assisted heat pump and solar heat pipe collector. The HPSAHP can
operate in two modes: the heat-pump mode when solar radiation
is low and the heat-pipe mode, without electricity consumption,
when solar radiation is high and can thus achieve high energy
efciency. Experimental results have shown that COP of the hybridmode is greater by 28.7% compared to the heat-pump mode. Biaoua
and Bernier [23] performed a TRNSYS (Transient System Simulation
program) simulation study of four alternative means for domestic
hot water production in zero net energy homes (ZNEH) for two
climates (Montreal and Los Angeles). The studied systems were: (i)
a regular electric hot water tank; (ii) the desuperheater of a ground
source heat pump (GSHP) with electric backup; (iii) thermal solar
collectors with electric backup; and (iv) a heat pump water heater
(HPWH) indirectly coupled to a space conditioning GSHP. Results
revealed that the third alternative provided the best performance
and that the fourth alternative is better than the second from
an energy point of view. Trillat-Berdal et al. [24] presented an
experimental study of a ground-coupled heat pump combined
with solar thermal collectors which can meet domestic hot water
and heating-cooling building energy needs. Solar heat is used as a
priority for domestic hot water heating and when the preset water
temperature is reached, excess solar energy is injected into the
ground via boreholes. Experiments over 11-month period showed
that the heat pump COP in heating mode had an average value of
3.75. Kjellsson et al. [25] studied the combination of solar collectors
with GSHP systems for heating and domestic hot water, using
TRNSYS. The study focused on the comparison among four different
system designs: (1) a base system with no solar heating (conventional GSHP); (2) a system where all solar heat recharges the
borehole; (3) all solar heat is used for domestic hot water; and (4)
all solar heat recharges the borehole in November-February, and is
used for domestic hot water during the rest of the year. It was recommended that the optimal design is the fourth one. Bakirci et al.
[26] constructed an experimental set-up of a solar assisted-GSHP
system. The experimental results obtained from October 2008 to
May 2009 revealed that COP of the heat pump and the whole system were in the range of 33.4 and 2.73, respectively. It is worth
to mention that the temperature of the condenser water outlet was
between 43 and 73 C which is very good for oor heating. Although
the study did not mention water heating, it is obvious that the system could be used for instantaneous or storage domestic hot water
production.
This overview reveals that research concerning hybrid domestic
hot water systems could be dealt with in more details (load prole,
losses, coverage portion from each energy source, renewable share,
system optimization). Moreover, most investigations studied the

design and performance of solar, air dual source HPWHs or examined the solar assisted ground source heat pump water heaters. In
this study, a detailed investigation of a domestic hybrid water heating system is presented, for different climates and annual periods,
using specied numerical simulation models. Additionally, optimal management of the proposed system has been suggested and
examined.
3. System description
The objective of this study is to propose a domestic hybrid water
heating system, model it and study its optimal management in
different meteorological conditions.
The proposed hybrid water heating system consists of a solar
collector, a heat pump water heater either of air source (ASHPWH)
or of ground source (GSHPWH) type, a wind turbine, battery storage
and a hot water storage tank. Fig. 1 illustrates a schematic of the
proposed system.
The HPWH (air source or ground source) is connected in parallel
with the solar collector to the storage tank. Both systems can operate simultaneously to transfer heat to water. The solar collector is
initiated when the temperature difference between the uid at its
outlet and the water in the tank exceeds 8 K. On the other hand, the
operation is denied when this temperature difference is less than
2 K. The wind turbine generates electricity from the available wind
energy and stores it in the battery. In its turn, the battery feeds the
heat pump with the necessary electrical load; however, when its
capacity is minimal, the electricity needed is secured from the grid.
The collector is supposed to transfer heat to water as long as solar
energy is sufcient and the temperature of the heat transfer uid is
greater than the water-tank temperature by a certain amount. On
the other hand, the HPWH operates when the average temperature
of water in the storage tank is below 48 C and stops when this temperature exceeds 55 C. This functioning procedure is illustrated in
Fig. 2.
4. Mathematical formulation
The mathematical model of each of the presented individual
systems is illustrated in the following subsections.
4.1. Flat-plate solar thermal collector
A at-plate solar thermal collector is modeled in this section. The
collector consists of parallel rows of copper-tubes connected to two
headers; a bottom header and an upper header, which collects the
HTF before being transferred into a water tank, either directly or via
a heat exchanger. The tubes are bonded under the absorber plate.
In addition, a glass cover is situated at a certain distance above the
absorber plate.
Finite volume numerical method is used, where the absorber
plate as well as each single tube is divided into a certain number of
control volumes. The following assumptions are considered:
Pressure drop along the tubes is neglected and thus the momentum equation is not needed.
During a time step, the density of HTF is constant. In addition, the
mass ow rate is always constant.
Same mass ow rate per each of the parallel tubes is considered.
HTF ow through the tubes is one-dimensional.
Axial conduction of HTF is neglected.
Conduction in the glass cover is neglected.
Absorber-plate area is assumed to be the same as that of the glasscover area.

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

345

Fig. 1. Schema of the hybrid water heating system.

One-dimensional heat ow through the back insulation and

through the glass-cover is considered.
Adiabatic boundary condition at the edges is considered.

kp

Tp
|x=0,L = 0;
x

kw
Considering the above stated assumptions, modeling the atplate solar collector could be simplied by dividing the absorber
plate in the direction perpendicular to the tubes (y-direction), such
that each tube is covered by one division of the absorber. This division is, in turn, bordered by two adjacent divisions that have no
tube beneath. Hence, taking advantage of the obtained symmetrical conguration, the collector could now be modeled only by a
single tube as shown in Fig. 3.
The energy balance for the absorber plate, the tube wall and the
HTF are given by Eqs. (1)(3), respectively.
(VCp )p

Tw
|x=0,L = 0;
x

+ Upw (Tw Tp ) + Ut dxdy(Ta Tp ) + Ubp Do (Ta Tw )

Tw
2 Tw
+ i Di (THTF Tw )
= (kAcs )w
t
x2
Upw
(Tp Tw ) + Ubw Do (Ta Tw )
dx

THTF
THTF
p )HTF
(Acs Cp )HTF
= i Di (Tw THTF ) (mc
t
x

(2)
(3)

The initial and boundary conditions are considered as follow:

THTF (x, 0) = Tw (x, 0) = Tp (x, y, 0) = Ta ;

THTF (0, t) = Ti

The equivalent transmittanceabsorptance product (a)eq is

evaluated using Eq. (4) [27]:
(a)eq = 0.96(a)beam

(4)

where (a)beam = 1.02a is the transmittance-absorptance product

related to the beam radiation only, noting that the calculation of
the transmittance of the cover  is given in [27].

Upw =

(1)

Tp
|y=0,W = 0;
y

4.1.1. Heat transfer coefcients

The heat transfer coefcient for the heat ow between the
absorber plate and the tube-wall is given by Eq. (5):

Tp
2 Tp
2 Tp
+ (kV )p
= (a)eq Idxdy + (kV )p
2
t
x
y2

(Acs Cp )w

kp

1
(p /(kp Dwo dx)) + (1/Cb dx)

(5)

The term (p /kp Dwo dx) is very small and thus, it could be
neglected. Hence, Upw = Cb dx.
The HTF-side heat transfer coefcient is computed from the following Nusselt numbers (Eq. (6)) [28,29]:

Nu =

3.66,

for laminar forced ow

0.023Re0.8 Pr 0.4

Ra
16

for turbulent forced ow

)3/4 )

[1 e(16(1/2Ra

where Ra * = Ra(ri /L)

],

for natural convection

(6)

346

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

Fig. 2. Flow chart of the functioning procedure of the proposed hybrid water heating system.

The overall loss heat transfer coefcient for the top part of the
collector (Ut ) is given by Eq. (7):

where the radiative heat transfer coefcients are calculated as

follow [27]:

1
1
1
=
+
Ut
pcover + apcover
wind + acoversky

p-cover =

(7)

2
)(Tp + Tcover )
(Tp2 + Tcover

(1/p ) + (1/cover ) 1

(8)

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

347

Fig. 3. Discretization scheme of a single tube of the solar collector.

2
2
coversky = cover (Tcover
+ Tsky
)(Tcover + Tsky )

(9)

The sky temperature is determined from Eq. (10) [30]:

Tsky = 0.0552Ta1.5

(10)

The heat transfer coefcient for the natural convection between

the absorber plate and the glass cover is evaluated using equation
(11) [31]:

Nu = 1 + 1.44 1

Ra cos
5830

1708[sin(1.8)
Ra cos

1/3

1.6



1708
Ra cos

+
1

(11)

where the exponent + means that only positive values of the

terms in the square brackets are to be used (i.e., zero value is used
if the term is negative).
The wind heat transfer coefcient is calculated using the following Nusselt numbers (Eq. (12)) when the wind velocity is high
[32]:

Nu =

0.664Re0.5 Pr 1/3
0.037Re0.8 Pr 1/3

for laminar ow
(12)
for turbulent ow

However, if the wind velocity is very small, then natural convection will have a higher inuence than forced convection. This
phenomenon is detected when Gr/Re2 > 10, and the wind heat
transfer coefcient will be calculated using the Nusselt number for
an inclined heated surface with natural convection (Eq. (13)) [32]:
Nu = 0.14[(GrPr)1/3 (Grcr Pr)1/3 ] + 0.56(GrPr sin(slope))1/4

(13)

where Critical Grashof number (Grcr ) is dened in [32].

The thermal resistance between the bottom insulation and the
ambient is neglected, because its magnitude is very small compared
with the thermal resistance through the insulation [27]. Hence,
the heat transfer coefcients for the heat losses from the bottom
of the absorber plate (Ubp ) and from the bottom of the tube-wall
(Ubw ) are evaluated by considering the heat conduction through
the insulation only.
4.1.2. Validation of the model
An experimental test performed by the Testing and Laboratories Division, Florida Solar Energy Center (FSEC) was employed to
validate the developed dynamic model for the at-plate solar collector [33]. Simulations are conducted using the test conditions and
collector conguration illustrated in [33]. The simulated instantaneous collector efciencies are compared to experimental ones.

Fig. 4. Predicted and experimental efciency with error analysis of the investigated
solar collector.

Fig. 4 shows good agreement between the predicted and experimental results, where the absolute error is less than 3%. Hence, the
developed model is considered to be quite accurate.
4.2. ASHPWH
A typical ASHPWH is modeled in this section. Static models are
chosen to simulate the actuator components since the dynamics
of compressors and expansion devices are generally much faster
than those of the heat exchangers [34,35], where the compressor
is considered as an adiabatic rotary one and the throttling device
is considered as a thermostatic expansion valve. Concerning the
heat exchangers, lumped parameter, moving boundary dynamic
models are chosen. The modeled evaporator is a tube and ns heat
exchanger type, while the condenser is a smooth copper helical
coiled tube, immersed inside the water tank.
4.2.1. Heat exchangers
The dynamics of a vapor compression system are assumed to
be dominated by the dynamics of the heat exchangers, since those
of the actuator components are much faster. The heat exchanger
dynamic models are generally classied into three groups: lumped
parameter models, nite volume/nite difference models, and
moving boundary models. In this work, the moving boundary
approach is chosen to model both the evaporator and condenser,
principally because of its capability to handle all transient load
changes [36] as the load in our case-the hot water-may change
a lot in its quantity and temperature. Furthermore, this approach
is capable to capture important dynamics due to the complex heat
exchanger behavior [37] and perform fast computer simulations
[36,38,39], while preserving the simplicity of lumped parameter
models [37] and the accuracy of nite volume/nite difference
models [36].

348

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

Rasmussen and Alleyne [37] have developed a moving boundary model for the multi-phase heat exchangers in the air to air
heat pump. In this work, this approach will be applied for air to
water heat pump system. The mathematical formulations of the
heat exchanger models are developed, where each is divided into
a number of control volumes (regions) according to the phase of
refrigerant and thus, the evaporator is divided into two regions
(two-phase and superheated) while the condenser is divided into
three (superheated, two-phase and subcooled). Considering several simplifying assumptions [37], the governing partial differential
equations (PDEs) are simplied to one dimensional PDEs illustrated
in Eqs. (14) and (15). In addition, the tube wall energy balance is
given by Eq. (16) [37]:

Fig. 5. Predicted and experimental Pe and Pc vs. tank water temperature (Ta = 5 C)
with error analysis.

(Acs )
(m)
+
=0
t
z

(14)

(Acs h Acs P)
(mh)
+
= Di i (Tw Tr )
t
z

The heat transfer coefcient of the two-phase region inside the

evaporator is given by Eq. (20) [41]:

(15)

Nu = 102 .(Re2 Kf )

(Cp A)w

Tw
= Di i (Tr Tw ) + Do o (Ta Tw )
t

(16)

Eqs. (14)(16) are then integrated along each control volume to

obtain the governing equations which are independent of spatial
parameters. After that, dening the state vector xe = [L1 Pe ho Tw1
Tw2 ]T for the evaporator and algebraically combining the derived
governing equations, the governing equation of evaporator dynamics reduces to the following compact state space form Eq. (17):

11

Y12

f/8 (Re 1000) Pr

f/8 (Pr 2/3 1)

(18)

where f = 1/(0.79 ln (Re) 1.64)2

The heat transfer coefcient of the two-phase region inside the
condenser is calculated by the correlations given in Eq. (19) [40]:


Nu =

Nu =

0.6 + (0.387 Ra1/6 )

(1 + (0.559/Pr)

9/16 8/27

5.03Re1/3 Pr 1/3 ,
0.0265Re0.8 Pr 1/3 ,

(21)

(17)

o Ao (Ta Tw2 ) i2 Ai (Tw2 Tr2 )

Y55

4.2.2. Heat transfer coefcients

The single-phase refrigerant side heat transfer coefcient is
computed from the well-known Gnielinski correlation (Eq. (18)):

1 + 12.7

where Kf = (h/L g)
The water-side natural heat transfer coefcient is calculated
using the following correlation for Nusselt number (equation 21)
[42]:

where L1 is the length of the two-phase region (rst region), Tw1

and Tw2 are the temperatures of the evaporator tube wall in the rst
region and the superheated region (second region), respectively.
Expressions of all elements in matrix [Y] are given in [37].
Similar derivation is carried out for the condenser dynamics resulting in equation of the form [B].x c = f , with state vector
xc = [Lc1 Lc2 Pc hco Tcw1 Tcw2 Tcw3 ]T , where B is a matrix, f is a vector, Lc1 and Lc2 are the lengths of the superheated region (rst
region) and two-phase region (second region), respectively, Tcw1 ,
Tcw2 and Tcw3 are the temperatures of the condenser tube wall in
the rst region, second region and subcooled region (third region),
respectively.

Nu =

(20)

m (h h ) + A (L /L )(T T )
g
w1
r1
in in
i1 i 1 total

0
m out (hg hout ) + i2 Ai (L2 /Ltotal )(Tw2 Tr2 )

in m
o
m
0
x e =

o Ao (Ta Tw1 ) i1 Ai (Tw1 Tr1 )

0

Y
21 Y22 Y23 0

Y31 Y32 Y33 0

0
0
0 Y44
Y51

0.4

Re < 55, 000

(19)

The air-side heat transfer coefcient is given by Eq. (22) [43]:

= 61.67u0.63
a

(22)

4.2.3. Validation of the model

The proposed ASHPWH dynamic simulation model is validated
using the experimental investigations presented by Guo et al. [44].
The experimental setup was composed of a rotary compressor,
thermostatic expansion valve, nned and tube air cooled evaporator, and a condenser, which was made up of a smooth copper helical
coiled tube immersed in an insulated water tank. Using the physical parameters stated in [44], the proposed simulation program
was executed to study the water heating process from an initial
temperature of 15 C to the set temperature (55 C) at 5 C ambient
temperature. Figs. 5 and 6 show respectively, the simulated variations of pressures (condensing and evaporating) and COP versus
tank water temperature with the corresponding experimental data
as taken from Guo et al. [44] as well as the absolute error analysis
for each. It could be noticed from both gures that the simulation
results are in quite good agreement with the experimental results
as the maximum absolute error, considering all studied parameters,
is about 7%. Thus, the developed model could be used in further
studies concerning the illustrated ASHPWH.
4.3. Ground heat exchanger

Re55, 000

where Re = 0.5.G . [1 + (f /g )1/2 ]di /f and G = 4m/d

i

The dynamic numerical modeling of a vertical ground heat

exchanger that can be coupled to a heat pump is presented in this

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

(mcp )HTF,u

349

THTF,d THTF,u
THTF,u
Tb THTF,u
+
=
Rub
Rdu
t

THTF,u
(24)
z
THTF,d Tb
THTF,u Tb
2ks dz(Ts (1, j) Tb )
T
(mcp )gr b =
+
+
Rdb
Rub
t
ln(rm (1)/rb )
(25)
p )HTF,u
(mc

The resistances Rd-b , Ru-b and Rd-u are described in [46], where:
Fig. 6. Predicted and experimental COP vs. tank water temperature (Ta = 5 C) with
error analysis.

section. The heat exchanger is a U-tube embedded inside a borehole, where the two pipes (upward and downward) of the tube are
symmetric with respect to the borehole axis and are separated by
a distance 2d.
The borehole and the ground eld surrounding it are divided
into m divisions in the vertical direction. In addition, each vertical division of the ground is divided into n annular divisions,
where the thickness of one annular division is 1.2 times that of the
previous division. Fig. 7 shows a cross section of the system with
the adopted gridding.
Several simplifying assumptions suggested in [45] are considered. In addition, the axial conduction of HTF, pressure drops
along the pipes and thermal capacity of the pipe wall are
neglected. Moreover, the ow of HTF is one dimensional and
only heat transfer by conduction occurs along both the annular and the vertical divisions in the ground surrounding the
borehole.
The energy balance equations of the descending and ascending
HTF as well as that of the borehole are given by Eqs. (23)(25),
respectively:
(mcp )HTF,d

THTF,d
t

THTF,d
z

Additionally, the energy balance for ground temperature is

given by Eq. (26):
(cp )s dzAcs (i, j)

Ts (i 1, j) Ts (i, j)
Ts (i + 1, j) Ts (i, j)
TS (i, j)
=
+
R(i 1, j)
R(i, j)
t
+ ks dzAcs (i, j)

2 Ts
z 2

(26)

The thermal capacity of the annular region is lumped at the

barycentric radius rm (i), which is equivalent to (Eq. (27)):

rm (i) =

r 2 (i) r 2 (i 1)
2

(27)

The values of the thermal resistances R(i,j) are calculated using

equation (28):
R(i,j) =

ln(rm (i + 1)/rm (i))

2kdz

(28)

The initial and boundary conditions are considered as follow:

Tb THTF,d
THTF,u THTF,d
+
Rdb
Rdu
p )HTF,d
(mc

Rd-b : thermal resistance between the downward-circulating HTF

and the borehole wall;
Ru-b : thermal resistance between the upward-circulating HTF and
the borehole wall;
Rd-u : thermal resistances between the pipes

(23)

THTF,u (z, 0) = THTF,d (z, 0) = Tb (z, 0) = Ts (x, z, 0) = Tground (z); THTF,d (0, t) = Ti
Tb (0, t) = Ts (x, 0, t) = Tground (0); Ts (x, H + dz/2, t) = Tground (H + dz/2);
Ts (rmax , z, t) = Tground (z)

Fig. 7. Cross section of the ground heat exchanger, borehole, and ground with the adopted gridding.

350

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

4.5. Battery
A lead-acid battery is used as a part of the studied hybrid system.
It stores energy produced by the wind turbine and supplies electrical energy to the load; the electricity needed by the compressor
of the heat pump. At a given time, the excess energy generated by
the wind turbine is stored in the battery that undergoes a charging
state. In this case, the available battery capacity is calculated by Eq.
(32) [51]:

Cbat (t) = Cbat (t 1) + PwT (t)

AC/DC

Fig. 8. Simulated and experimental inlet and outlet temperatures of HTF in the
ground heat exchanger vs. time with error analysis.

The undisturbed ground temperature (Tground ) is determined

from equation (29) [47]:

Tground = Tmean Tamp exp Depth

cos

 2  
365


365 as

0.5 

tnow tshift Depth/2 (365/as )

0.5



(29)
where, Tmean = mean surface temperature (average air temperature); Tamp = amplitude of surface temperature (maximum air
temperature minus minimum air temperature); Depth = depth
below the surface; as = thermal diffusivity of the ground (m2 /day);
tnow = current time (day); tshift = day of the year of the minimum
surface temperature.
4.3.1. Validation of the model
The experimental results found in [48] are used to validate the
developed dynamic model of the ground heat exchanger, where the
values of the geometric and thermo-physical properties illustrated
in the considered reference are used as inputs for the developed
model. Simulations are conducted over a period of 12.5 h, where
the temperatures of the heat transfer uid at the inlet and outlet
of the ground heat exchanger are recorded. These temperatures
are then compared to those obtained from the above mentioned
experimental test.
Fig. 8 shows the simulated variation of the HTF-temperature at
the inlet and outlet of the ground heat exchanger versus time with
the corresponding experimental data as taken from [48]. It also
shows that the absolute percentage error of the simulated results is
less than 4%. Hence, the simulation model is quite reliable and could
be used for further studies concerning the ground heat exchanger.
4.4. Wind turbine
A small wind turbine is modeled by its power output as a function of the wind speed. This power is evaluated using equation (30)
[49]:

PWT =

 2


P
(u
u2cut in )/(u2rated u2cut in )

WT,rated wind

PWT,rated

ucut
if

in

ln(Hhub /zo )
ln(Hdata /zo )

ch dt

(32)

On the other hand, when the output energy from the wind turbine is less than the load energy, the battery is in the discharging
state and its capacity is computed by Eq. (33) [51]:

Cbat (t) = Cbat (t 1)

PL (t)

DC/AC

PwT (t)
AC/DC dt

dch

(33)

It is worth to note that when neither the wind turbine energy

output nor the battery storage capacities are sufcient, the electrical load is satised by grid electrical energy.
The charging and discharging efciencies (
ch and
dch ) of the
battery are considered 75% and 100%, respectively [52]. Furthermore, the converter efciencies (
AC/DC and
DC/AC ) are considered
in this study as 92%.
The available battery capacity is bounded between minimum
and maximum allowable storage capacity values. In this study, the
maximum battery charge quantity is set as the nominal capacity
of the battery and the minimum capacity is determined using the
maximum allowable depth of discharge (DOD) as shown in Eq. (34)
[53]:
Cbat,minimum = (1 DOD) Cbat,nominal

(34)

According to manufacturers specications, the maximum battery lifetime can be attained using a 3050% DOD [53]. In this study,
a value of 50% is considered.
4.6. Water storage tank
The water storage tank stores heated water, ready for being supplied to the demand site. It has a cylindrical shape and is surrounded
by a mantle heat exchanger connected to the solar collector. The
helical, coiled tube-water condenser of the HPWH is placed inside
the tank to release heat to water by natural convection. The storage tank is installed horizontally and uniform temperature for the
water inside is considered. The energy balance for the tank model
is given by Eq. (35):
d(mCp T )wat
= (Q HP + Q solar Q L Q loss )
dt

urated < uwind < ucut

where ucut in , ucut off and urated are the cut in, cut off and rated wind
speeds that characterize the wind turbine system.
Using the reference wind speed data at a given height, the wind
speed (uwind ) at a certain hub height is calculated from Eq. (31) [50]:
uwind = udata

(35)

< uwind < urated

uwind < ucut

PL (t)

DC/AC

(31)

in

(30)

off

or uwind > ucut

off

Hot water owing out of the tank is mixed with the tap water
before reaching the end user and thus, its ow rate (g(t)) depends on
the domestic hot water prole (h(t)). This ow rate is determined
from the following energy conservation Eq. (36):
g(t)[(Cp T )wat (Cp T )tap ] = h(t)[(Cp T )consume (Cp T )tap ]

(36)

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

Fig. 9. Ambient temperature for typical winter and summer days in Beirut and
Cedars.

Fig. 10. Solar radiation for typical winter and summer days in Beirut and Cedars.

In this study, the temperature of the consumed hot water

(Tconsume ) is considered 45 C. As for the tap water temperature
(Ttap ), it is determined using Eq. (37) [54].
Ttap = 1.28 + 0.19Tmean1 + 0.79Tmean2

(37)

where, Tmean1 : daily average outdoor air temperature on the

same day ( C), Tmean2 : average outdoor air temperature for past
seven days ( C)

Fig. 12. Hot water consumption proles for typical winter and summer days in
Lebanon [57].

two locations. These data are obtained using meteonorm software

[56]. In addition, Fig. 12 shows the daily, domestic hot water consumption proles for typical summer and winter days in Lebanon
[57].
It has been previously stated that one of the hybrid system components is a heat pump water heater, which is either an air-source
or a ground-source type. The choice of the HPWH type depends
on two parameters: the average coefcient of performance (COP)
and the surface availability for a ground heat exchanger. In fact,
Beirut is an urban city, where almost all people live in apartments of
high-story buildings. As a result, the needed surface area to install a
ground heat exchanger as a part of the GSHPWH appears to be very
difcult. Hence, ASHPWH is used as one of the hybrid system components in Beirut. As for Cedars, it is a rural town, where a ground
heat exchanger could be easily installed. Thus, to choose the appropriate HPWH type in Cedars, the average COP for heating 220 L
water volume from 48 C to 55 C, which is the operational interval
considered in this study, is evaluated for both types. Fig. 13 clearly
shows that the average COP of GSHPWH is signicantly greater
than that of ASHPWH for the winter season, while it is slightly
lower in summer at the corresponding mean seasonal temperatures
(Tmean summer and Tmean winter ). Therefore, GSHPWH is employed for
the proposed hybrid system in Cedars.
The average COP is calculated, according to equation (38):

 Tfinal

5. Results and discussion

The hybrid system, described in Section 3, is simulated in this
part. Matlab software [55] is utilized to develop the individual subsystems models and then to integrate them in a global model
that represents the proposed hybrid system. Two case studies are
selected to investigate the performance of the proposed system:
one is in Beirut city and the other is in Cedars town. The main reason
for selecting these sites is the evaluation of the hybrid water heating
system in diverse climatic and demographic conditions. Figs. 911
respectively illustrate the hourly ambient temperature, solar radiation and wind speed for typical summer and winter days in the

Fig. 11. Wind speed for typical winter and summer days in Beirut and Cedars.

351

COPaverage =

initial
 tfinal

tinitial

(mcp dT )wat

Pconsumed dt

(38)

Subsequently, simulations for the hybrid water heating system in Beirut and Cedar are carried out for typical days of winter
and summer, starting from an average water temperature of 55 C.
Table 1 illustrates values of important parameters used in simulations.
Fig. 14 shows the variation of tank-water temperature during typical summer and winter days in Beirut and Cedars. It is
noticed that the water temperature does not reach its inferior

Fig. 13. Average COP for ASHPWH and GSHPWH in winter and summer seasons of
Cedars.

352

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

Fig. 14. Variation of tank-water temperature during typical winter and summer days in Cedars and Beirut.

set temperature (48 C) in summer for both cities, while this

phenomena is observed several times in winter. This is mainly
interpreted by: (1) higher hot water consumption prole in winter,
(2) higher heat losses to the ambient in winter, due to low air temperatures, and (3) low tap-water temperature in winter. Moreover,
the proposed hybrid water heater is capable of providing all hot
water needs in Beirut and Cedars throughout the year without the
need of any other energy sources.
Fig. 15 shows the useful energy transferred to the water tank
from the solar and heat pump systems in all studied cases. It is
shown that the heat pump delivers no energy to water during summer, while both systems participate in delivering energy to the hot
water tank in winter. It is noticed that the collected solar energy
and the period of solar system operation are approximately the
same for both locations in summer; however, in winter, they are
signicantly greater for Beirut. When the solar system is started,
an initial peak value for the useful collected energy is observed due
to high temperature difference between HTF at the collector outlet
and the water in the tank (>8 C). In fact, when the solar system is off
and there is no mass ow rate, the heat transfer from the absorber
plate to HTF is low, because it is done by natural convection and
thus, the plate will have signicant higher temperature than that
of HTF due to large heat storage. Afterwards, when the system is
on, HTF gains heat from both collected solar energy and pre-stored
heat in the absorber plate and thus, an initial peak energy collection
is observed.
Fig. 16 illustrates the total electrical energy consumed during
typical winter day in Beirut and Cedars. In summer, no electrical energy is needed in both case studies, because all necessary
thermal energy is secured by the solar system. As for winter, solar
energy is not sufcient and thus, electrical energy is needed to
drive the HPWH in both locations. For Beirut, it is noticed that
the HPWH intervenes in the morning (7:449:07 AM) and in the
evening (6:167:53 PM), where the electrical energy share from
battery represents about 18% and 65%, respectively. For Cedars, the
HPWH operates in the morning (7:069:34 AM), in the afternoon
(3:235:14 PM) and in the evening between (6:339:29 PM), where
the electrical energy share from battery is marginal due to low wind
speeds.
Fig. 17 illustrates the loss heat transfer rate from the water in
the tank during typical summer and winter days. It shows that heat
loss is higher in winter than it is in summer. This is because winter
has lower ambient temperatures which lead to higher temperature difference between the tank-water and ambient air and thus,
higher heat losses. Similarly, heat loss in Beirut is lower than it is
in Cedars that has lower ambient temperatures. Furthermore, it is
noticed that heat loss varies widely according to the state of the

system components. For example, heat lost from tank-water during a typical winter day in Cedars increases sharply in zones 2, 6 and
8 (Fig. 17); because the heat pump is operating and the water temperature is increasing (Fig. 14). Consequently, the water-ambient
temperature difference increases and thus, heat loss increases. In
zones 1, 3, 5, 7 and 9, neither the heat pump nor the solar collector
are operating and hence, heat loss varies according to the variation
of water-ambient temperature difference. In zone 4, the solar system is on, and thus, hot HTF penetrates through the mantle gap and
transfers heat to the water in the tank. In this case, thermal losses
from the water are only from the circular end-edges of the tank and
this interprets the extremely low heat-loss values.
An indicator for the renewable energy share of the system is
dened as the renewable coverage factor (RCF), which is the ratio
of the renewable thermal energy to the total energy delivered to
the water tank. It is calculated from equation (39):

tfinal

RCF =

100 (

where Qtotal =

tinitial QHP

tfinal
tinitial

tfinal

tinitial Qsolar

Qtotal

QHP +

tfinal

tinitial

tfinal

tinitial EgridHP )

(39)

Qsolar .

Calculating RCF for all studied cases, it is found that the value
is 100% for a typical summer day in both investigated locations,
while it is about 86% and 63% for a typical winter day in Beirut
and Cedars, respectively. Therefore, as the studied cases represent
the extreme climatic conditions in Lebanon, it could be concluded
that RCF of this hybrid water heating system is very high all around
the country, year round. Hence, using such a system is supposed to
signicantly reduce the residential energy consumption and operating costs as well as to decrease the greenhouse gas emissions due
to the reduction of grid-electrical energy demand.
6. Management problem
In this section, an energetic-economic optimal management
model is introduced for the proposed hybrid water heating system.
Two objectives are aimed out of this model, where one of these
is energetic tending at minimizing the total grid-electrical energy
consumption and consequently maximizing the renewable energy
share. The other objective is economic aiming at obtaining a positive net present value (NPV) for the proposed hybrid system, which
is also greater than that of conventional electric water heaters over
a pre-specied period of time.
Solving for these objectives will optimize the number of individual systems that compose the proposed hybrid water heating
system as well as determine the optimum energy-ow distributions among different system components. This model is applied for

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

353

Fig. 15. Useful energy transferred to the water tank from the solar and heat pump systems during typical summer and winter days in Beirut and Cedars.

Fig. 16. Electrical energy consumed during a typical winter day in Beirut and Cedars.

the case studies presented in Section 5 (Beirut and Cedars), where

the objective function f is dened as follows:
f =

j +

j x1 x2

(40)

Therefore, minimization f tends to maximize the number of

solar units as well as the energy delivered by them, whereas this
minimizes the number of heat pump units and the energy delivered by them. Furthermore, concerning the electricity fed to HPWH,
minimizing f tends to maximize the electricity delivered by the

Fig. 17. Loss heat transfer rate from hot water tank during typical summer and winter days in Beirut and Cedars.

354

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

Table 1
Values of parameters used in simulation.
Flat-plate solar collector
Slope [ ]
Length, L [m]
Width, W [m]
Area, A [m2 ]
Thermal conductivity (absorber and tube), kp /kw [W/mK]
Density (absorber and tube), p /w [kg/m3 ]
Specic heat (absorber and tube) (Cp )p /(Cp )w [J/kgK]
Number of tubes
HTF [kg/s]
Total mass ow rate of HTF, m

34
1.92
0.96
1.84
401
8940
390
8
0.015

HPWH
Refrigerant

R134a

Compressor
Rated input power, Prated [W]
Swept volume, Vk [m3 ]

920
16.5 106

Condenser (smooth copper coil)

Length, Lc [m]
Inner diameter, Dci [m]
Outer diameter, Dco [m]

50
0.00915
0.0099

Evaporator (n and tube, air-heated type)

Front surface area, Afrontal [m2 ]
Air-side area, Aair-side [m2 ]
Inner diameter of tube, Dei [m]
Outer diameter of tube, Deo [m]
Pitch of traverse tubes, pttraverse [mm]
Pitch of longitudinal tubes, ptlongitudinal [mm]
Fin thickness, n [mm]
Pitch of ns, ptn [mm]

0.3445
6.17
0.00902
0.00952
25.1
25.1
0.13
1.34

Evaporator (concentric tube, liquid heated type)

Length, Le [m]
Inner diameter of tube, Dei [m]
Outer diameter of tube, Deo [m]

10
0.00902
0.00952

Borehole
Depth, Hb [m]
Radius, rb [m]
Number of vertical divisions, m

110
0.055
20

U-tube, ground heat exchanger

Inner radius, rwi [m]
Outer radius, rwo [m]
Spacing (center-to-center), 2d [m]
Thermal conductivity of wall, kw [W/mK]
HTF [kg/s]
Mass ow rate of HTF, m

0.013
0.016
0.06
0.4
0.25

Grout
Thermal conductivity, kgr [W/mK]
Thermal diffusivity, agr [m2 /s]

1.3
3.3 107

Soil (ground)
Thermal conductivity, ks [W/mK]
Density, s [kg/m3 ]
Specic heat (Cp )s [J/kgK]
Thermal diffusivity, as [m2 /s]
Number of annular divisions, n

3.5
2700
830
1.56 106
20

Wind turbine
Rated power, PWT, rated [kW]
Cut in wind speed, ucut in [m/s]
Cut off wind speed, ucut off [m/s]
Rated wind speed, urated [m/s]
Hub height in Beirut/Cedars, HWT [m]
Surface roughness length in Beirut/Cedars, Z0 [m]

1
3
25
10
30/20
2/0.55

Battery
Initial capacity, Ci [Wh]

Water storage tank

Inner diameter, Di [m]
Outer diameter, Do [m]
Length, Ltank [m]
Heat transfer coefcient between tank wall and HTF [W/m2 K]
Loss heat transfer coefcient between mantle cover and
ambient air [W/m2 K]
Volume of water [L]
Initial water temperature, Twat,i [ C]

0.46
0.57
1.38
94
0.83
220
55

wind-battery units and consequently maximize the number of

these units, while it minimizes the grid electricity consumption.
In addition, the dened objective function is minimized such that
the following constraints are satised (Eqs. (41)(43)):
j

tfinal

tinitial

QHPj + j

tfinal

tinitial

Qsolarj

tfinal

tinitial

QLj +

tfinal

tinitial

QLossj
(41)

tfinal
tinitial

Egrid-HPj + j

tfinal
tinitial

Est-HPj = j

tfinal
tinitial

EHPj

NPVhybrid system NPVEWH

(42)
(43)

where, j , j ,  j and j are real positive numbers used to determine

the optimum energy-ow distributions among the system components in a season (j) typical day. x1 and x2 are the number of solar
collectors and wind-battery base-units, respectively. QHPj and Qsolarj
are the amount of useful energy delivered to the water in the storage tank by the heat pump and solar systems in a season (j) typical
day, respectively. QLj is the energy load of the hot water consumption prole in a season (j) typical day. QLoss j is the energy loss from
the hot water tank in a season (j) typical day. Egrid-HPj and Est-HPj , are
the amount of electrical energy delivered to the heat pump by the
grid electrical network and the battery storage system in a season
(j) typical day, respectively. EHPj is the electrical energy load of the
heat pump in a season (j) typical day. NPV is the net present value.
EWH is the conventional electric water heater.
The net present value (NPV) is calculated using equation (44):
NPV =

[1 (1 + DR)n ]
AS IC
DR

(44)

where, DR is the discount rate; it is considered 10% in this study,

AS is the annual savings, IC is the initial capital cost, n is the time
period of the study.
The annual savings are considered null for EWH, since it is
considered as the conventional water heater to which the hybrid
system is compared. On the other hand, the annual savings of the
hybrid system are evaluated using Eq. (45).
AShybrid system =

ASj

Given that:

tfinal

tinitial QLj ) +

ASj =

(45)

tfinal

tinitial QLossj

EWH
j

tfinal
tinitial


Egrid HPj

costgrid

electricity

Nd

where Nd is the number of days of the studied season (j).

The initial capital cost of the hybrid water heating system is
calculated from Eq. (46):
IChybrid system = cos ttank + x1 cos tsolar + x2 cos tWB + x3 cos tHP

(46)

where, costtank , costHP , costsolar , costWB and costEWH are the average capital costs for one unit of each of the hot water tank, heat
pump, solar collector, wind-battery system and conventional electric water heater, respectively.
Simulations of the developed hybrid conguration are carried
out with small base-units for the renewable components, solar and
wind systems. They are considered to operate as maximum as possible. Hence, if the optimization solution indicates a higher energy
ow from either of these systems, then this means that a larger
system is needed. This larger system is expressed as a multiplication of the studied base unit. This multiplication is dened, for
the solar unit, by x1 and is computed as a rounding-down of the
maximum between j values to the nearest integer. Similarly,
rounding-down the maximum between j values to the nearest

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

355

Fig. 18. Optimum energy-ow distributions among system components during typical winter and summer days in Beirut and Cedars for time periods of ve and ten years.

integer, represents the multiple (x2 ) for the case of the wind-battery
unit.
Concerning the HPWH, a small unit is chosen such that it has a
heating capacity sufcient enough to overcome the load capacity
in addition to the heat losses. In this study, a 920 W unit is adopted.
Accordingly, the optimum number of heat pump systems (x3 ) will
be one unless all values of j are zero, then x3 will be also zero.
For special cases that would exist in the optimal management
of the hybrid water heating system, a modied objective function
should be dened. A special case could exist when all values of  j
are zero, given that at least one of the values of j is not zero. In
this case, the model tends to maximize the value of x2 as possible.
In addition, j values will be equal to x2 and thus, neither the
optimum number of wind-battery units nor the optimum energy
ow from them is obtained. In such cases, to solve this issue, the
model is resolved with another modied objective function f1
(equation 47), with x2 being now the rounding-up of the maximum
between j values to the nearest integer.
f1 =

j x1 + x2 +

(47)

Another special case could also exist when all values of j are
zero. In this case, the values of  j will also be zero and the model
tends to maximize the value of x1 as possible. In addition, j
will be equal to x1 and thus, neither the optimum number of solar
units nor the optimum energy ow from them is obtained. To solve
this issue, the model is resolved with another objective function
f2 (Eq. (48)) with x1 being now the rounding-up of the maximum
between j values to the nearest integer.
f2 = x1 +

(48)

The presented optimal-management model is applied for the

two investigated case studies, Beirut and Cedars, where it is solved
using lingo optimization software [58]. The following assumptions
are considered:
Each year is considered to be divided equally between summer
and winter.
The investigated typical winter and summer days represent all
days of winter and summer, respectively.
The maximum number of solar collector units that could be
installed for one residential unit is two in Beirut and four in Cedars
(x1 = 2 (Beirut); x1 = 4 (Cedars)).
The maximum number of wind-battery units that could be
installed for one residential unit is two in Beirut and four in Cedars
(x2 = 2 (Beirut); x2 = 4 (Cedars)).
An efciency of 90% is considered for the conventional EWH [59].
The mean electricity-tariff is considered to be 0.24 UScent/kWh.
This value is obtained from the average value of the real cost of
grid electricity (about 17 UScent/kWh [60]) and that of the electricity produced by private generators (about 0.31 UScent/kWh

Table 2
Results for the optimum number of system components as well as capital cost and
annual savings of the hybrid system.
No. of units

Time-period
5 (years)

x1
x2
x3
Capital cost (US$)
NPV (US$)
Annual savings (US$)

10 (years)

Beirut

Cedars

Beirut

Cedars

2
0
1
1550
525.5
547.5

1
0
1
2400
532
618

2
1
1
3550
39.5
584

4
0
1
3600
1104
669.5

[61]), because both electricity sources (grid and private generators) constitute the electricity sector in almost all Lebanese zones.
The average initial capital cost of one unit for each of the hot water
tank, solar collector, wind-battery system and conventional EWH,
as obtained from the Lebanese market, are 400, 400, 2000 and 190
US$, respectively.
The average initial capital costs of an ASHPWH and a GSHPWH
are 350 and 1600 US$, respectively. These values are taken from
international markets, because they are not found, to the best of
the authors knowledge, in Lebanon.
The energetic, economic optimal management model is investigated for two time periods: ve and ten years. Fig. 18 shows the
optimal energy-ow distributions among system components for
the two case studies during typical winter and summer days for the
two studied time periods. In addition, Table 2 illustrates, for each
case, the optimal number of units for each system component, as
well as the overall initial capital cost, the annual savings and the
NPV. It is shown that: (1) the initial cost of the hybrid system for
the investigated cases is between 1550$ and 3600$, which could
be considered acceptable, especially if incentive loans were found;
(2) the annual savings range between 574.5$ and 669.5$ and (3)
the greatest NPV of the hybrid system is obtained in cedars and the
lowest in Beirut over 10 year time period.
7. Conclusions
In this study, a domestic hybrid water heating system is proposed and modeled. The system is composed of a solar collector,
air-source/geothermal heat pump water heater, wind turbine, battery and a hot water storage tank. Numerical dynamic models of
each of the individual components is presented and validated. Simulations for typical days in summer and winter of two Lebanese
cities, Beirut and Cedars, are conducted using small base units
(1.84 m2 solar collector, 920 W HPWH, 1 kW wind turbine). Results
show that the solar system alone is sufcient for Beirut and Cedars
in the summer season. As for winter, both heat pump and solar
systems participate in delivering the needed energy for the water
tank. For Beirut, it is noticed that the HPWH intervenes in the

356

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

morning (7:449:07 AM) and in the evening (6:167:53 PM), where

the electrical energy share from battery represents about 18% and
65%, respectively. For Cedars, the HPWH operates in the morning (7:069:34 AM), in the afternoon (3:235:14 PM) and in the
evening between (6:339:29 PM), where the electrical energy share
from battery is very marginal due to low wind speeds. In addition, a renewable-share indicator, the renewable coverage factor
(RCF), is dened and evaluated for all case studies. It is shown
that RCF, in Lebanon, is always above 63% and could reach 100%,
as for the case of summer in Beirut and Cedars. Furthermore, an
energetic-economic optimal management model is suggested for
the proposed hybrid water heating system. This model tends to
optimize the size of the hybrid system, where a multiplication of
the base units for each component is to be selected. In addition,
the optimal energy-ow distributions among the system components are also obtained. Economically, the model is solved for a
net present value (NPV) of the hybrid system being greater than
that of a conventional electric water heater. The developed model
is applied for the two considered case studies (Beirut and Cedars),
where two time periods, ve and ten years, are investigated and an
average electricity price is adopted. In Beirut, for the time period
of ve years, the results yield two solar collectors and one heat
pump base units without any wind-battery units. For ten years
time period, same results are obtained with an additional one windbattery unit. In Cedars, for the time period of ve years, the results
yield one solar collector and one heat pump base units. For ten years
time period, four solar collector and one heat pump base units are
obtained. It is noticed that the optimization results yield no windbattery units for Cedars. The main results of the energetic-economic
optimal management model show that: (1) the initial cost of the
hybrid system for the investigated cases is between 1550$ and
3600$, which could be considered acceptable, especially if incentive
loans were found, (2) the annual savings range between 574.5$ and
669.5$ and (3) the greatest net present value for the hybrid system
is obtained in cedars and the lowest in Beirut over 10 year time
period. Finally, the proposed optimal management model yields
reliable results and could be generalized to other locations.

Acknowledgment
The authors would like to thank the Azm research center at
EDST-Lebanese University as well as the Lebanese National Council
for Scientic Research (CNRS) for their support.

References
[1] LEBANON, Social Impact Analysis For the Electricity and Water Sectors. The
World Bank, MENA Knowledge and Learning. Quick notes series, Number 14,
November 2009.
[2] M. Levine, D. rge-Vorsatz, K. Blok, L. Geng, D. Harvey, S. Lang, G. Levermore, A.
Mongameli Mehlwana, S. Mirasgedis, A. Novikova, J. Rilling, H. Yoshino, Residential and commercial buildings, in: B. Metz, O.R. Davidson, P.R. Bosch, R. Dave,
L.A. Meyer (Eds.), Climate Change 2007: Mitigation. Contribution of Working
Group III to the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change, Cambridge University Press, Cambridge, United Kingdom/New
York, NY, USA, 2007.
[3] Energy Efciency Trends and Policies in the Household & Tertiary sectors
in the EU 27.Lessons from the ODYSSEE/MURE project. ADEME-Agence de
lEnvironnement et la Maitrise de lEnergie, November 2009.
[4] C. Aguilar, D.J. White, D.L. Ryan, Domestic Water Heating and Water Heater
Energy Consumption in Canada, Canadian Building Energy End-Use Data and
Analysis Center-CBEEDAC 2005-RP-02.
[5] State of Australian Cities 2010, Published by the Australian government, Infrastructure Australia, Major Cities Unit under the number: INFRA-09173, March
2010.
[6] Energy Efciency in Russia: Updated Reserves. The World Bank Group report,
2008.
[7] M. Aoshima, Energy Consumption Survey in Japans residential sector,
EDMC/IEEJ The Energy Data and Modeling Center/The Institute of Energy
Economics, February 2010.

[8] J.A. Rosas-Floresa, D. Rosas-Floresa, D.M. Glvez, Saturation energy consumption, CO2 emission and energy efciency from urban and rural households
appliances in Mexico, Energy and Buildings 43 (2011) 1018.
[9] N. Zhou, M.A. McNeil, D. Fridley, J. Lin, L. Price, S. de la Rue du Can, J. Sathaye,
M. Levine, Energy Use in China: Sectoral Trends and Future Outlook. Ernest
Orlando Lawrence Berkeley National Laboratory, Environmental Energy Technologies Division-LBNL 61904, January 2007.
[10] Natural resource accounts, Energy accounts for South Africa, Statistics South
Africa (SSA), Discussion document-April 2005, 19952001.
[11] A. Houri, Renewable Energy Resources in Lebanon: Practical Applications,
ISESCO Science and Technology Vision 1 (2005) 6568.
[12] Lebanon Social Impact Analysis-Electricity and Water Sectors, The World Bank,
Report No. 48993-LB, 2009.
[13] B.J. Huang, J.P. Chyng, Integral-type solar-assisted heat pump water heater,
Renewable Energy 16 (1999) 731734.
[14] B.J. Huang, J.P. Chyng, Performance characteristics of integral type solar assisted
heat pump, Solar Energy 71 (6) (2001) 403414.
[15] J.P. Chyng, C.P. Lee, B.J. Huang, Performance analysis of a solar-assisted heat
pump water heater, Solar Energy 74 (2003) 3344.
[16] B.J. Huang, C.P. Lee, Long-term performance of solar-assisted heat pump water
heater, Renewable Energy 29 (2003) 633639.
[17] B.J. Huang, C.P. Lee, Performance evaluation method of solar-assisted
heat pump water heater, Applied Thermal Engineering 27 (2007)
568575.
[18] X. Guoying, Z. Xiaosong, D. Shiming, A simulation study on the operating
performance of a solar-air source heat pump water heater, Applied Thermal
Engineering 26 (2006) 12571265.
[19] T.N. Anderson, G.L. Morrison, Effect of load pattern on solar-boosted
heat pump water heater performance, Solar Energy 81 (2007) 1386
1395.
[20] A. Nuntaphan, C. Chansena, T. Kiatsiriroat, Performance analysis of solar water
heater combined with heat pump using refrigerant mixture, Applied Energy 86
(2009) 748756.
[21] H. Li, H. Yang, Study on performance of solar assisted air source heat pump
systems for hot water production in Hong Kong, Applied Energy 87 (2010)
28182825.
[22] B.J. Huang, J.P. Lee, J.P. Chyng, Heat-pipe enhanced solar-assisted heat pump
water heater, Solar Energy 78 (2005) 375381.
[23] A.L. Biaoua, M.A. Bernier, Achieving total domestic hot water production with
renewable energy, Building and Environment 43 (2008) 651660.
[24] V. Trillat-Berdal, B. Souyri, G. Fraisse, Experimental study of a ground-coupled
heat pump combined with thermal solar collectors, Energy and Buildings 38
(2006) 14771484.
[25] E. Kjellsson, G. Hellstrom, B. Perers, Optimization of systems with the combination of ground-source heat pump and solar collectors in dwellings, Energy
35 (2010) 26672673.
[26] K. Bakirci, O. Ozyurt, K. Comakli, O. Comakli, Energy analysis of a solar-ground
source heat pump system with vertical closed-loop for heating applications,
Energy 36 (2011) 32243232.
[27] J.A. Dufe, W.A. Beckman, Solar Engineering of Thermal Processes, 3rd edition,
John Wiley & Sons, New Jersey, 2006.
[28] Y.A. Cengel, Heat Transfer, A Practical Approach, 2nd edition, 2002.
[29] J. Padet, Convection thermique et massique, Nombre de Nusselt: partie 1, 2005.
[30] H.M.S. Hussein, M.A. Mohamad, A.S. El-Asfouri, Transient investigation of a
thermosyphon at-plate solar collector, Applied Thermal Engineering 19 (1999)
789800.
[31] K.G.T. Hollands, T.E. Unny, G.D. Raithby, L. Konicek, Free convection heat transfer across inclined air layers, Transactions of the ASME Journal of Heat Transfer
98 (1976).
[32] J.P. Holman, Heat Transfer, McGraw-Hill, New York, 1986.
[33] Solar Collector Test Report on MSC-32 American Energy Technologies Inc.
Florida Solar Energy Center, Testing And Laboratories Division, 1988.
[34] F. Fardoun, O. Ibrahim, A. Zoughaib, Dynamic Modeling of an Air Source Heat
Pump Water Heater, in: 10th International Energy Agency (IEA) Heat Pump
Conference, Tokyo, Japan, 2011.
[35] F. Fardoun, O. Ibrahim, A. Zoughaib, Quasi-steady state modeling of an air source
heat pump water heater, Energy Procedia 6 (2011) 325330.
[36] S. Bendapudi, J.E. Braun, E.A. Groll, A comparison of moving-boundary and
nite-volume formulations for transients in centrifugal chillers, International
Journal of Refrigeration 31 (2008) 14371452.
[37] B.P. Rasmussen, A.G. Alleyne, Dynamic Modeling and Advanced Control of Air
Conditioning and Refrigeration System, ACRC: TR-244, 2006.
[38] B. Li, A.G.A. Alleyne, Full Dynamic Model of a HVAC Vapor Compression Cycle
Interacting with a Dynamic Environment, in: American Control Conference,
Hyatt Regency Riverfront, St. Louis, MO, USA, June 1012, 2009.
[39] T.L. McKinley, A.G. Alleyne, An advanced nonlinear switched heat exchanger
model for vapor compression cycles using the moving-boundary method, International Journal of Refrigeration 31 (2008) 12531264.
[40] L. Fu, G. Ding, C. Zhang, Dynamic simulation of air-to-water dual-mode
heat pump with screw compressor, Applied Thermal Engineering 23 (2003)
16291645.
[41] E. Granryd, I. Ekroth, P. Lundqvist, . Melinder, B. Palm, P. Rohlin, Refrigerating
Engineering, Royal Institute of Technology, KTH, Stockholm, 2005.
[42] S.W. Churchill, H.H.S. Chu, Correlating equations for laminar and turbulent free
convection from a horizontal cylinder, International Journal of Heat and Mass
Transfer 18 (1975) 10491053.

O. Ibrahim et al. / Energy and Buildings 75 (2014) 342357

[43] N.Z. Jiang, G.Q. Yao, Q.J. Zhou, Technologies and Applications of Heat Pump
for Air-conditioning, Mechanical Industry Press, Beijing, 1997, pp. 125126,
195196 (in Chinese).
[44] J.J. Guo, J.Y. Wu, R.Z. Wang, S. Li, Experimental research and operation optimization of an air-source heat pump water heater, Applied Energy 88 (2011)
41284138.
[45] C.K. Lee, H.N. Lam, Computer simulation of borehole ground heat exchangers
for geothermal heat pump systems, Renewable Energy 33 (2008) 12861296.
[46] G. Hellstrom, Ground Heat Storage, Thermal Analysis of Duct Storage Systems
(Doctoral thesis), Department of Mathematical Physics, University of Lund,
Lund, Sweden, 1991.
[47] T. Kasuda, P.R. Archenbach, Earth Temperature and Thermal Diffusivity at
Selected Stations in the United States, ASHRAE Transactions (1965) 71
(part 1).
[48] G.A. Florides, P. Christodoulides, P. Pouloupatis, An analysis of heat ow through
a borehole heat exchanger validated model, Applied Energy 92 (2012) 523533.
[49] R. Pallabazzer, Evaluation of wind generator potentiality, Solar Energy 55
(1995) 4959.
[50] A. Ouammia, H. Dagdougui, R. Sacile, Mimet, A. Monthly, seasonal assessment
of wind energy characteristics at four monitored locations in Liguria region
(Italy), Renewable and Sustainable Energy Reviews 14 (2010) 19591968.
[51] G. Notton, S. Diaf, L. Stoyanov, Hybrid photovoltaic/wind energy systems for
remote locations, Energy Procedia 6 (2011) 666677.

357

[52] S. Diaf, D. Diaf, M. Belhamel, M. Haddadi, A. Louche, A methodology for optimal sizing of autonomous hybrid PV/wind system, Energy Policy 35 (2007)
57085718.
[53] B. Ai, H. Yang, H. Shen, X. Liao, Computer-aided design of PV/wind hybrid
system, Renewable Energy 28 (2003) 14911512.
[54] M. Nabeshima, Techniques of Statistical Data Analysis to Offer Suggestive
Observation for the Facility Designer (PhD thesis), Osaka City University, 1998.
[55] Matlab software, 2009.
[56] Meteonorm, Global Meteorological Database, version 7.
[57] A. Houri, H. Salloum, A. Ali, A.K. Abdel Razik, L. Houri, Assessment of Energy and
Financial Performance of a Solar Hot Water System in a Single Family Dwelling:
Case study from Marjeyoun-South Lebanon, LCEC-The Lebanese Center for
Energy Conservation, 2009.
[58] Lingo users guide, Lindo systems Inc., 2011 www.lindosystems.org
[59] Domestic Water Heating Fuel Cost Comparisons, Nova Scotia Department of
Energy-Publications & Reports, Canada, February 2011.
[60] MEW-Ministry of Energy and Water. The minister: Jebran Bassil. Policy paper
for the electricity sector, June 2010.
[61] Energy Ministry sets price for generator services, The Daily Star, November
2012, Available at: http://dailystar.com.lb/Business/Lebanon/2012/Nov-02/
193610-energy-ministry-sets-price-for-generator-services.
ashx#axzz2SWuN7Y6L