Energy xxx (2015) 1e10

Contents lists available at ScienceDirect

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

Optimization of heat pump system

bert Sa

nta a, *, La

szlo

Garbai b, Igor Frstner a
Ro
a
b

Subotica Tech e College of Applied Sciences, Subotica, Serbia

Department of Building Service and Process Engineering, Budapest University of Technology and Economics, Hungary

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 23 January 2015
Received in revised form
10 July 2015
Accepted 11 July 2015
Available online xxx

The purpose of this study is to produce a mathematical model to describe the operation of a water-towater heat pump system for steady-state condition. The set-up model is deterministic. It consists of
distributed as well as lumped parameters. The proposed mathematical models of heat exchangers were
described by coupled differential equations, while the models of the compressor and the expansion valve
are of lumped parameters. The RungeeKutta and the AdamseMoulton predictor-corrector methods were
applied for the numerical solution of differential equations, i.e. the equation systems. The developed
mathematical model is validated with 118 tests using R134a as a working uid. The results show that an
average difference between the modeled and experimental results for the coefcient of performance is
1.73%, which means that the proposed mathematical model can be used to determine the optimum
operating point of a heat pump system for a given heat demand for heating, by determining the
maximum value of the coefcient of performance.
2015 Elsevier Ltd. All rights reserved.

Keywords:
COP (coefcient of performance)
Heat pump
Mathematical model
R134a
Optimization

1. Introduction
Early 2008, the European Commission adopted a futureoriented proposal [1], intended to reduce greenhouse gases still
causing increases of the greenhouse effect, and at the same time to
increase the application of renewable energy resources to a 20%
ratio within the total of energy use by 2020. Hungary, as an EU
Member State, incorporated the EU decision on such major increase
of renewable energy resources in its legal system, and its objective
is to raise the proportions of renewable energy use from the current
5e6% to14.65%. In the authors' view, the spreading of heat pump
technology, the most energy-efcient method for heating and
cooling buildings, is the best solution for Hungary, having in mind
the given natural conditions.
In order to improve the energy efciency of heat pumps and to
increase the quality of operation, it is unavoidable to strive to
describe, as precisely as possible, heat pump operation and the
processes therein. Therefore, on the one hand, the theoretical basis
of the present study is the development and renement of the
mathematical model of the heat pump, while on the other hand, it
is the optimization of the system's operation.
A thorough literature review reveals that the mathematical
modeling of heat pumps was discussed extensively in the past.
Koury et al. [2] proposed a model for a refrigeration system with
* Corresponding author. Tel.: 36 309556449.
E-mail address: santa@vts.su.ac.rs (R. S

anta).

distributed parameters model for heat exchangers. Numerical

simulations were carried out to verify the possibility of controlling
the refrigeration system and the superheating of the refrigerant in
the evaporator outlet by varying the compressor speed and the
throttling valve position. Choi et al. [3] also carried out numerical
simulations for a multi-type heat pump, to examine the performance of the system in transient conditions. Hermes et al. [4]
presented a mathematical model for transient and steady simulations of the thermal and uid-dynamic behavior of the refrigerant
ow through adiabatic and non-adiabatic capillary tubes. YoubieIdrissi et al. [5] developed a rather simple simulation model,
which allows the highlighting of the behavior of a water-to-water
heat pump operating under steady-state conditions, from its heat
exchangers geometry and compressor characteristics.
In their paper Sheng et al. [6] investigated the energy saving
potential of HTHP&DW system compared to the reference technologies based on vapor compression (CVC (conventional vapor
compression) system) and desiccant cooling (HDC (hybrid desiccant cooling) system).
Zhenjun et al. [7] presented the biogas engine-driven heat pump
air conditioner in a new-style system which includes a biogas engine driven heat pump, primary heat exchanger, second heat
exchanger, sprayed room and fans, pumps, etc. The authors investigated the inuence of various factors including the outdoor air
temperature and humidity both in summer and winter.
The mathematical model of the heat pump enables the design of
optimal systems, searching for more economical solutions,

http://dx.doi.org/10.1016/j.energy.2015.07.042
0360-5442/ 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

assessment of operational characteristics, as well as lifetime and

cost planning. The energy optimum of a heat pump is realized
when the COP (coefcient of performance) takes up the maximum
value.
In the design phase, the nominal state of the maximum energy
efciency can be set by selecting the optimal system parameters, in
an effort to jointly minimize the total cost of system installation and
operation. In the case that the system had already been designed
and manufactured, the coefcient of performance can only be
improved by reducing the mechanical work invested. Therefore,
the aim is to set the optimum operating point by aligning the rpm
of the motors running the compressor and the pumps as well as by
other decisions detailed later on, which, in turn, can be realized by
setting up a mathematical model to enable the optimization of
operation.
The literature review shows that numerous researchers have
discussed the optimization and analysis of heat pump operation.
Nyers et al. [8] discussed the COP value of the heat pump
condenser, by varying the mass ow of the heated media and the
refrigerant. Sarkar et al. [9] presented the experimental results for a
heat pump system for simultaneous water cooling and heating. The
cooling and heating capacities and the COP have been studied for
various operating conditions (water mass ow rates and water inlet
temperatures of both evaporator and gas cooler) and expansion
valve openings. Primal et al. [10] detected an important inuence of
the evaporator conditions on the charge, when a water-to-water
heat pump under different evaporation temperatures was tested.
They found that the optimal charge decreased with the evaporating
temperature. Zhang et al. [11] presented a system optimization of
air source heat pump water heater, including calculation and
testing. Zhao et al. [12] analyzed the steady state performance of
the heat pump system, showing that the overall system is a two
time-scale dynamic system.
Based on the presented literature review it can be concluded
that the researchers carried out a deep and thorough analysis of
numerous specic heat pump systems. The present study aims at
contributing to this eld in order to ll a gap in literature by discussing the scientic design of the steady state operation of the
water-to-water heat pump system that was missing or just partially
discussed in literature. Thus, the objective is to describe, more
precisely than before, the thermodynamic heat transfer and uid
mechanics processes within the system components in order to
realize the heat pump's cycle, as well as to be able to identify
changes of state at each point e and otherwise at discretionary
points e of the cycle.
The proposed mathematical model allows the optimization of
the system operation for different heating demands; that is, it examines how to satisfy a given heat demand by using a minimum
amount of electric power. Therefore, the aim is to maximize COP as
the target function of the system by examining what decision parameters can be used to set it, both on the water side of the
evaporator and the condenser e and what values thereof e, taking
into consideration the behavior of the compressor and the expansion valve at operating points other than nominal ones. The goal
function of system operation is:

Qcon


/max!:
COP


W

(1)

Heating capacity:

Qcon Qeva W:

Cooling capacity:



Qeva f m_ cw ; Tcw;i; Tcw;o :

(3)

Compression power:



W f m_ ref ; peva;i; pcon;o ; h :

(4)

2. Design of the mathematical model for the heat pump

system
The physical model of the major elements of the heat pump
system is shown in Fig. 1. This physical model, which is described
mathematically, consists of an evaporator, a compressor, a
condenser, and an expansion valve.
The presented study focuses on shell and tube heat exchangers.
The test model developed for the entire heat pump system also
includes investigations of plate heat exchangers; however, the
characteristics of different types of heat exchangers e the relationship of input and output proles e are specied by the
manufacturers.
In the present case, the refrigerant is R134a; owing in the tubes
of heat exchangers, while the primary and secondary uid-water,
ows in the shell side. In the shell side of heat exchangers bafes
are applied primarily for supporting the tubes and for inducing
cross ow over the tubes, resulting in improved heat transfer performance. The bafes are perpendicular to the tube bundle. The
examined heat exchangers have segmental bafes.
The mathematical model of the heat pump system is set up for
steady state behavior. It is deterministic, and of distributed parameters, meaning that the relations between variables can be
uniquely dened. They are independent of time, and the parameters with their values are taken into account according to location.
The mathematical model of the heat exchangers consists of basic
equations e governing equations in a different terminology e and
auxiliary equations, while the models of the compressor and the
expansion valve are of lumped parameters.
The following assumptions have been made regarding the
physical and mathematical models:
 refrigerant ow in the evaporator and the condenser is onedimensional and permanent,
 refrigerant liquid and vapor phases are in thermodynamic
equilibrium,
 in heat exchangers, only axial ows are taken into account.
In the following subsections, basic equations and conditional
equations are stated separately for each system component. By
solving them, changes of state of the refrigerant can be specied at
certain points of the cycle, namely in the evaporator, at the inlet and
outlet points of the compressor, in the condenser, and in the
expansion valve, thereby making it possible to set control parameters according to discretional consumer demands, which in turn
can be used for specifying the optimal operating point of the
system.
2.1. Heat exchangers

(2)

2.1.1. Basic equations

For the mathematical model of heat exchangers, a two zone
model is utilized. The evaporator is divided into two zones

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

Fig. 1. Schematic diagram of a water-to-water heat pump.

(evaporation and superheating zones), and the condenser is also

divided into two zones (superheating and condensation zones).
Each zone is considered as a control volume. The model is established by applying the energy conservation equation for each heat
exchanger control volume (refrigerant, secondary uid and pipe
wall), as well as the mass and momentum conservation equations
(refrigerant). This procedure generated the following set of differential equations.
Mass conservation equation of the refrigerant

vr$w
0 /r$w Gref const:
vz

Pipe wall energy conservation equation



aw $Tw  Twall  aref $ Twall  Tref 0:

(8)

Water energy conservation equation

(5)





m_ w $cw $ Tw;i  Tw;o  aw $Aout $ Tw  Twall 0:

(6)

2.1.2. Auxiliary equations

In addition to the afore-mentioned basic energy equations,
several auxiliary equations are required in the simulation of the
heat pump. These include the heat transfer, pressure drops, thermodynamic property relations, and other thermo-physical data.

(9)

Momentum conservation equation of the refrigerant



v r$w2
vp fwall 2

$w $r 0:
vz 2din
vz
Energy conservation equation of the refrigerant



v
w2
K
_
h
 q$
Gref $
0:
vz
Ain
2

 Heat transfer

(7)

Since phase change of the refrigerant occurs as it ows in the

system, the correlations for heat transfer coefcient for the refrigerant should be selected considering its phase.

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

4
Table 1
Refrigerant side heat transfer coefcients.
Region

Heat transfer coefcient

Single phase
l
av 0:023$Re0:8 $Pr n $ :
din

(10)

Evaporation
Co4:1

aeva 12:6$

Xtt4:8

$af 0:43$1  x$anb :

Condensation
5

acon 28:6$e3:15$10

$Ree

 
l
$Pr 1:11 $
:
din

(11)

aw 1:66$Re

$Pr

0:038

l
$ :
De

The thermodynamic and transport properties of R134a are

determined by the P-R state equation [19]:

(13)

Refrigerant pressure drop in the heat exchangers has also been

introduced into the model. The single-phase (vapor) ow friction
factor was calculated using the Darcy [18] equation, while the shell
(water) side of the heat exchangers is determined from the correlation by Santa [17], (Eq. (14)).

fwall $

 2
Dz
w
3:09 $ $r:
De
2

(15)

 Thermodynamic and Thermo-physical Properties

 Pressure drops

Dp

Dp
Dx
f
G2ref $vv $  wall $G2ref $vv $x:
Dz
Dz 2$din

(12)

For the single phase regions in the evaporator and condenser,

the DittuseBoelter correlation [13], (Eq. (10)) was selected. The
Santa correlation [14], (Eq. (11)) was used for the prediction of the
heat transfer coefcient in ow boiling. The proposed heat transfer
coefcient of evaporation was partly based on the GungoreWinterton correlation [15]. The Santa correlation [14], (Eq.
(12)) was applied to obtain the condensation heat transfer coefcient, which is modied by the Akers correlation [16].
Internal heat transfer coefcients for both evaporator and
condenser are shown in Table 1.
In the DittuseBoelter (Eq. (10)) correlation, the exponent of Pr
should n 0.4 for heating and n 0.3 for cooling, so the correlation
is different when it is used in the evaporator and condenser.
The external (water-side) heat transfer coefcient for horizontal
tubes is given by Santa [17] (Eq. (13)).
0:43

drop is the sum of the momentum pressure drop and the frictional
pressure drop. The total pressure drop in the evaporator and the
condenser for the two-phase refrigerant and superheated vapor
has been adopted by using the simple correlation proposed by
Santa [17].

(14)

The total pressure drop of the two-phase refrigerant depends

on the variation of kinetic and potential energy of the refrigerant
and on the friction on the ow channel wall. The total pressure

R$T
aT

:
v  b v$v b b$v  b

(16)

with

s!#2
"
R2 $Tc2
T
aT 0:45724$
$ 1 k$ 1 
:
Tc
pc

k 0:37464 1:54226$u  0:26992$u2 :

R$Tc
:
b 0:07780$
pc
where a and b are energy and size parameters, respectively. R is the
universal gas constant. T, p and v are temperature, pressure and
molar volume, respectively. u 0.3268 is the acentric factor.
Subscript c indicates the critical value.
The above equations, supplemented by auxiliary equations, and
together with the equations of state on the refrigerant and the
water, and the equations describing heat exchanger characteristics,
comprise the steady state mathematical model of distributed
characteristics of the evaporator and the condenser. The systems of
equations (5)e(7) are suitable for solution by the RungeeKutta and
AdameMoulton methods. By knowing the initial characteristics,
and by progressing at step intervals of Dz (Fig. 2), the equation
system can be used to dene the values of refrigerant pressure and
enthalpy, as well as temperature changes of the cooled water
owing in the shell side.

Fig. 2. Changes of working medium parameters in a shell and tube evaporator in function of length.

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

2.2. Compressor

3. Solution of the mathematical model of the heat pump

system

The compressor as a system component can be described by the

enthalpy change specied by the compression ratio and the internal efciency of the compressor. The output demand of the
compressor is:


 1
W m_ ref $ hcon;in  heva;o $ :
h

(17)

2.3. Expansion valve

The expansion valve (proportioning valve) is intended to supply
appropriate quantities of refrigerant to the evaporator in function
of overheating. Isenthalpic throttling is assumed in operation. The
mass ux of the refrigerant owing through the expansion valve is:

q
Gref C$ Dp$rref :

(18)

The complexity of the developed mathematical model of the

system required the simulation to be performed by solving the
equations of the model through an iterative process, which is
shown as a general owchart in Fig. 3. The simulation algorithm is
developed for steady state behavior. In the simulation, each variable of the model was made discrete, meaning that each function of
independent variables was represented by discrete numbers.
The developed simulation algorithm is coded in C.
The input data into the simulation is organized in three subgroups:
 Transport properties (thermodynamic and physical properties
of the refrigerant, physical properties of water),
 Thermodynamic parameters (water mass ow rate, refrigerant
mass ow rate, cold water inlet temperature, temperature of the
superheated refrigerant vapor),
 Geometric properties (dimensions of the evaporator and
condenser).
The simulation process consists of four sub-processes, which are
based on the mathematical models of the compressor, condenser,
expansion valve and evaporator. First, the compressor's mathematical model runs, followed by the condenser's and expansion
valve's mathematical models. The computation results for the mass
ow of the refrigerant in the compressor (m_ ref ;comp ) and the mass
ow of the refrigerant in the expansion valve e (m_ ref ;tev ) are
compared. The temperature of the superheated refrigerant vapor is
adjusted (Tsh) until the resulted difference between computation
results is greater than the dened acceptable error. Following this
procedure, the evaporator's mathematical model runs. The
computation result output enthalpy from the condenser (hcon,o) is
compared to the input enthalpy to the evaporator (heva,in). Until the
resulted difference between enthalpies is greater than the dened
acceptable error, the evaporator's temperature is adjusted (Teva),
and the procedure is repeated. Following this, an output is generated, which contains the required data for the analysis of the heat
pump system.
In terms of the numerical resolution of differential equations/
equation systems, application of the RungeeKutta method e
among the great number of such methods e represents a
compromise between simplicity and accuracy. This method has
been made more precise by applying the AdamseMoulton
predictor-corrector method.
The method essentially requires the rst 4 iterations to be performed using RungeeKutta's method in order to provide an input
for the AdamseMoulton method. This latter method can be used
for the calculation of solutions starting from iteration 5.
AdamseMoulton's is an iterative method: the desired level of accuracy can be achieved in the given step, usually by performing
iteration several times (n < 10).
The function that is sought F(z) (p(z), h(z), T(z)) is a vector
function with the following components: p(z) e refrigerant pressure, h(z) e refrigerant enthalpy, T(z) e water temperature, each of
them is a function of the distance z from the starting point of the
tube (z 0). This means that the calculation of F actually involves
the calculation of three quantities. The equation F0 G (p,h,T,z) is to
be solved.
4. Experimental results and model validation

Fig. 3. Flowchart of the simulation program.

The experimental setup consists of the heat pump and a system

of measuring instruments. The specication of the heat pump is

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

6
Table 2
Specications of the main components of the heat pump system.
Name

Remarks

Compressor
Evaporation

Reciprocating compressor, R134a, rated input power: 904 W, piston displacement: 32.7 cm3.
Shell and tube heat exchanger.
Shell inner diameter Din 32 mm. The tubes are made of copper and have a staggered layout.
Number of tubes: 5. Tube inner diameter din 6 mm. Number of bafes: 15. Bufe cut 30%. Length of evaporator z 3 m
Shell and tube heat exchanger.
Shell inner diameter Din 32 mm. The tubes are made of copper and have a staggered layout.
Number of tubes: 5. Tube inner diameter din 6 mm. Number of bafes: 15. Bufe cut 30%. Length of condenser z 3 m
Internally-equalized type

Condensation

Expansion valve

Fig. 4. Installation and arrangement of measuring points.

presented in Table 2. The measuring system is fully equipped with

sensors to measure key variables such as pressures, temperatures,
volumetric and mass ow rates, and compressor energy consumptions, which is presented in Fig. 4. The parameters of the
Table 3
Measured parameters and uncertainty.
Measuring instruments

Name

Acc.

Thermocouples

Dallas
DS18S20
Meters Prematlak
0e400 kPa,
Mihajlo Pupin
Transducers MP-1M2
Wattmeter, 0e3 kW,
tip el2, Iskra
Volumetric ow meters:
INSA 4.115
BMET 55643811
Coriolis ow meter:
Krohne Optimas 6400

0.3 K

Pressure gauge and sensor

Power
Flow meters

1%
0.5%

0.5
0.6%, 0.2%

0.1%

measuring instruments are presented in Table 3. Measuring instruments were placed along the shell and tube evaporator and the
condenser, the length of which was z 3 m. Measuring points were
installed at 10 discrete points in the heat exchangers. The distance
between measuring points was Dz 30 cm (Fig. 5). At the
measuring points thus installed, the temperature and pressure of
the refrigerant as well as of the cooled and the heated medium
were to be measured. Meters were applied to measure the volumetric ow of the cooled and the heated mediums, and a mass ow
meter was used for measuring the mass ow of the refrigerant.
To validate the mathematical model of the heat pumps, a
comparison between experimental data and model outputs was
carried out.
Below, the following measurement and mathematical conditions are given:
Working medium:
Mass velocity of refrigerant:
Reynolds number range of refrigerant:
Volumetric ow of water
Reynolds number range of water

R134a and water

G 106, 114, 135 [kg/m2s].
3500 < Re < 5000 [].
V_ 1; 1:5; 2 m3 =h
2000 < Re < 8500 [].

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

Fig. 5. Installation and arrangement of measuring points.

Fig. 6. Prediction of the water temperature in the shell side evaporator.

The temperature proles of the primary and secondary uid in

the shell side of the evaporator and the condenser are shown in
Fig. 6 and Fig. 11.
Fig. 6 reects the prediction of the primary water temperature
of the evaporator, where the maximum prediction error is 0.37  C,

Fig. 7. Prediction temperature of the two-phase refrigerant in the tube of evaporator.

Fig. 8. Prediction of the cooling capacity.

while Fig. 11 shows the maximum discrepancy of the values of the

measured values, where the prediction error is 0.46  C.
Figs. 7 and 12 present the operating temperatures predicted by
the model compared with the measured values. The model prediction error in the evaporating temperatures is within 0.18  C. For

Fig. 9. Prediction of the pressure gradient in the tube of evaporator.

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042

nta et al. / Energy xxx (2015) 1e10

R. Sa

Fig. 10. Prediction of the compressor power consumption.

Fig. 13. Prediction of the pressure gradient in the tube of condenser.

Fig. 11. Prediction of the water temperature in the shell side condenser.

Fig. 14. Prediction of the condenser capacity.

the condensation temperature, the prediction error is within

0.79  C.
Fig. 8 shows the predicted values for the cooling capacity. The
predicted values are within an error margin of 18.48%. The relative
error predictions for the condenser capacity are about 26.98%. They
are shown in Fig. 14.
Fig. 10 shows the prediction for the compressor power consumption against the power measured by the digital wattmeter.
The prediction error is about 6.45%.

Figs. 9 and 13 show the prediction of pressure gradient in the

tube of evaporator and condenser. The predicted values are within
an error margin of 16.37%. The predicted values of the condenser
are within an error margin of 19.21%.
Finally, the coefcient of performance, COP, of the facility is
embodied in Fig. 15. The model predicts that 21 of the performed
tests are within the acceptable limits with a relative error of
11.48%.

Fig. 12. Predicted temperature of the two-phase refrigerant in the tube of condenser.

Fig. 15. Prediction of the coefcient of performance COP.

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

5. Maximization of the coefcient of performance of the

cycle, setting the optimal operating point in the case of a
given consumer heat demand
The methodology for maximizing the coefcient of performance
is presented in respect to the case of a given heat demand, together
with the results yielded for an example worked up. In order to
adjust an optimal operating point e the maximum COP, the intervening features are as follows:
 mass ow of cooled water m_ cw ,
 mass ow of refrigerant m_ ref ,
 mass ow of heated water m_ hw.

Basic data:






consumer heat demand: QH 3.5 kW,

thermal conductance: kA 330.6 W/K,
temperature of water returning from consumer: Thw 30  C,
temperature of primary water: Tcw 13  C,
surface of condenser and evaporator: Ain 0.376 m2.
Results are shown below in the optimization matrices.

6. Conclusions

The inlet temperature of cold water at the evaporator is a given

invariable value. The optimal operating point is set to a given heat
demand, i.e., to the given heat transfer surfaces at the consumer.
This should be associated with an also known value of medium
temperature difference at the heat transfer surfaces.
The optimal operating point is to be determined in the
following manner. The value of one of the intervening parameters is xed and the values of the other two parameters are
changed along the rows and columns of the matrix. The matrix
components are to also include the obtained COP values. At the
function of one of the columns and one of the rows of the matrix,
the maximum COP appears among the matrix elements. A new
COP matrix is calculated by changing the xed parameter value,
whereby the absolute maximum COP value can be obtained
through the set of maximum COP values. The course of the
calculation is presented through the data of an example in
Table 4.
Table 4
Optimization matrices: with matrix elements containing COP values.
QH 3.5 kW

_ cw 0:27 kg=s
m

_ ref g=s
m

_ hw kg=s
m
0.368

_ hw kg=s
m
0.481

_ hw kg=s
m
0.602

_ hw kg=s
m
0.715

17.19
17.14
17.08
16.98

4.81
4.78
4.68
4.62

4.80
4.77
4.65
4.60

4.78
4.76
4.64
4.59

4.77
4.74
4.63
4.57

QH 3.5 kW

_ cw 0:42 kg=s
m

_ ref g=s
m

_ hw kg=s
m
0.368

_ hw kg=s
m
0.481

_ hw kg=s
m
0.602

_ hw kg=s
m
0.715

17.81
17.76
17.70
17.66

5.02
4.97
4.85
4.77

5.01
4.96
4.84
4.76

4.99
4.95
4.83
4.75

4.98
4.94
4.82
4.74

QH 3.5 kW

_ cw 0:52 kg=s
m

_ ref g=s
m

_ hw kg=s
m
0.368

_ hw kg=s
m
0.481

_ hw kg=s
m
0.602

_ hw kg=s
m
0.715

18.40
18.21
18.10
18.07

5.24
5.20
5.10
5.02

5.23
5.19
5.09
5.01

5.22
5.18
5.08
4.99

5.21
5.17
5.07
4.98

QH 3.5 kW

_ cw 0:64 kg=s
m

_ ref g=s
m

_ hw kg=s
m
0.368

_ hw kg=s
m
0.481

_ hw kg=s
m
0.602

_ hw kg=s
m
0.715

18.87
18.69
18.55
18.40

5.42
5.34
5.23
5.15

5.41
5.33
5.22
5.14

5.40
5.32
5.22
5.14

5.40
5.31
5.21
5.13

Bold values present the maximum values of the COP.

This paper presents a steady-state model for a vapor compression heat pumps using R134a as working uid. The experimental
tests for the model validation were performed in a wide range of
operating conditions, allowing checking the robustness of the
model.
The solution of the mathematical model presented above can be
used for the determination of several changes of state: of the
refrigerant in the cycle, and of water owing in the shell side of the
evaporator and the condenser. The solution is given for a steadystate condition. Also, the impact on the coefcient of performance of the cycle can be qualied. Through the solution of the
mathematical model the main objective was to dene and set an
operating point for each of the various heat demands whereby the
cycle and the coefcient of performance (COP) of the system is
maximal. At the same time, the results yielded by the solution of
the mathematical model also conrm the correctness of the model
setup as well as its suitability for meeting the objective.
It has been stated that a maximum COP value can be achieved in
the course of satisfying a given heat demand by minimizing the
mass ux of the heated medium while maximizing the mass ux of
the cooled medium.
Based on the analysis and discussions of the calculating results,
some conclusions can be drawn, as listed below:
a.) When solving the basic equations, the accuracy of the results
is primarily determined by the accuracy of the input parameters, the heat transfer and the pressure drop correlations. The proposed correlations describe the heat transfer
coefcient value as a function of the Re number and the
vapor quality with smaller standard deviation and relative
error. Also, the refrigerant's pressure drop correlations are
also determined more accurately.
b.) The developed mathematical model is validated with 118
tests using R134a as a working uid. The results show that an
average difference between the modeled and experimental
results for the coefcient of performance is 1.73%.
c.) By establishing and solving the proposed mathematical
model, the optimum operating point of a heat pump system
for a given heat demand for heating can be determined, by
determining the maximum value of the coefcient of
performance.
d.) The optimum operating point is determined by using optimization matrices (Table 4). Testing was performed using the
following data:
 Mass ow of cooled water m_ cw 0:27  0:64kg=s ,
 Mass ow of refrigerant m_ ref 16:98  18:97 g=s;
 Mass ow of heated water m_ hw 0:368  0:715kg=s :
The value of one parameter, the mass ow of the cooled water
was xed, while the other two parameters' values were changed
along the rows and columns of the matrix. The values of the COP
were written into cells. The procedure was repeated for different

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042


nta et al. / Energy xxx (2015) 1e10

R. Sa

10

values of cooled water mass ows, and the maximum COP was
determined.
e.) The testing example was performed for a QH 3.5 kW heat
demand. The maximum COP of 5.42 was determined for inlet
cold water temperature of Tcw 13  C, if the mass ow of the
cold water was maximized m_ cw 0:64 kg=s, while the mass
ow
of
the
heated
medium
was
minimized
m_ hw 0:368 kg=s.

References
[1] European Commission. Directorate-General Communication. Combating
climate change: the EU leads the way. Cat number: NA-AB-08-128-EN-C.
Luxembourg: EUR-OP; 2008.
[2] Koury RNN, Machado L, Ismail KAR. Numerical simulation of a variable speed
refrigeration system. Int J Refrig 2001;24:192e200.
[3] Choi JW, Lee G, Kim MS. Numerical study on the steady state and transient
performance of a multi-type heat pump system. Int J Refrig 2011;34:429e43.
[4] Hermes CJL, Melo C, Goncalves JM. Modeling of non-adiabatic capillary tube
ows. A simplied approach and comprehensive experimental validation. Int J
Refrig 2008;31:1358e67.
[5] Youbi-Idrissi M, Bonjour J, Meunier F. Local shifts of the uid composition in a
simulated heat pump using R-407C. Appl Therm Eng 2005;25:2827e41.
[6] Sheng Y, Zhang Y, Zhang G. Simulation and energy saving analysis of high
temperature heat pump coupling to desiccant wheel air conditioning system.
Energy 2015;83:583e96.
[7] Zhenjun X, Huaizhi W, Meiling W. Energy performance and consumption for
biogas heat pump air conditioner. Energy 2010;35:5497e502.
[8] Nyers J, Garbai L, Nyers A. A modied mathematical model of heat pump's
condenser for analytical optimization. Energy 2015;80:706e14.
[9] Sarkar J, Bhattacharyya S, Ramgopal M. Experimental investigation of transcritical Co2 heat pump for simultaneous water cooling and heating. Therm Sci
2010;14(1):57e64.
[10] Primal F, Palm B, Lundqvist P, Granryd E. Propane heat pump with low
refrigerant charge: design and laboratory tests. Int J Refrig 2004;27:761e73.
[11] Zhang J, Wang RZ, Wu JY. System optimization and experimental research on
air source heat pump water heater. Appl Therm Eng 2007;27:1029e35.
[12] Zhao L, Zaheeruddin M. Dynamic simulation and analysis of a water chiller
refrigeration system. Appl Therm Eng 2005;25:2258e71.
[13] Dittus FW, Boelter LMK. Heat transfer in automobile radiators of the tubular
type. Publ Eng 1930;2:443.
[14] Santa R, Garbai L. Measurement testing of heat transfer coefcients in the
evaporator and condenser of heat pumps. J Therm Anal Calorim 2015;119:
2099e106.
[15] Gungor KE, Winterton RHS. A general correlation for ow boiling in tubes and
annuli. Int J Heat Mass Transf 1986;29(3):351e8.
[16] Akers WW, Deans HA, Crosser O. Condensing heat transfer within horizontal
tubes. Chem Eng Program Symp 1959;55:171e6.
[17] Santa R. Simulation and optimisation of compressor driven heat pumps for
building services. Budapest: University of Technology and Economics Faculty
of Mechanical Engineering; 2014. p. 69e74.
[18] Darcy H. Recherches Experimentales Relatives au Mouvement de L'Eau dans
les Tuyaux. Mallet-Bachelier 1857;2:268.

[19] Alexander DJ, Libretto SE. An overview of the toxicology of HFA-134a (1,1,1,2tetrauoroethane). Hum Exp Toxicol 1995;14(9):715e20.

Nomenclature
A: area (m2)
d: diameter (m)
De: hydraulic diameter (m)
x: vapor quality ()
K: perimeter (m)
G: mass ux (kg/sm2)
_ mass ow rate (kg/s)
m:
T: temperature (K)
T: average uid temperature (K)
f: friction factor ()
p: pressure (Pa)
Dp: pressure drop (Pa)
_ heat ux (Wm/2)
q:
w: velocity (m/s)
H: enthalpy (J/kg)
z: tube length (m)
_ volumetric ow rate (m3/s)
V:
v: specic volume (m3/kg)
C: characteristic constant of the TEV valve ()
Re: Reynolds number
Pr: Prandtl number
Xtt: Martinelli parameter
Co: convection number
Greek letters

a: heat transfer coefcient (W/m2K)

l: thermal conductivity (W/mK)
r: density (kg/m3)
h: efciency ()
Subscripts
eq: equivalent
tp: two-phase
sh: superheated
ref: refrigerant
c: cold
h: hot
i: inlet
o: outlet
in: inside
out: outside
l: liquid phase
v: vapor phase
w: water
wall: tube wall
eva: evaporation
con: condensation
comp: compressor
tev: expansion valve

Please cite this article in press as: S
anta R, et al., Optimization of heat pump system, Energy (2015), http://dx.doi.org/10.1016/
j.energy.2015.07.042