Arabian Journal for Science and Engineering

https://doi.org/10.1007/s13369-021-05490-1

RESEARCH ARTICLE-MECHANICAL ENGINEERING

Numerical Simulation of a Single‑compartment Household

Refrigerator During Start‑up Transient and Stationary Cyclic Operation
with Different Refrigerant Charges
Matheus dos Santos Guzella1,2   · Luben Cabezas‑Gómez2   · Luiz Gustavo Monteiro Guimarães3

Received: 30 June 2020 / Accepted: 18 February 2021

© King Fahd University of Petroleum & Minerals 2021

Abstract
A numerical study of the start-up transient and cyclic stationary operation of a single-compartment household refrigerator
and an analysis of performance quantities at different refrigerant charges are performed. The working fluid is HFC-134a.
The numerical model was developed using the commercial software GT-SUITE®. The individual models available in the
software were applied for each component of the refrigeration system (compressor, evaporator, condenser, capillary tube
and refrigerated compartment). The numerical model considers a solution of mass, momentum and energy principles using
a distributed formulation in space and implicit in time. The ambient temperature and refrigerant charge of the refrigeration
system are input data. The inside air temperature, power consumption and discharge and suction pressure were compared
against literature experimental data for the start-up at different ambient temperatures (32 °C and 43 °C) and cyclic opera-
tions at 32 °C. Results for different refrigerant charges are presented considering both ambient temperatures. The numerical
results showed reasonable agreement in comparison with literature experimental data, even for the different refrigerant
charges. The maximum error for the inside air temperature is 4 °C and 2.5 °C during pull-down tests at, at 32 °C and 43 °C,
respectively. The model failed to represent the pressure and power consumption peak during start-up, due to the simplicity
of the compressor model. However, the steady-state values obtained by the numerical model are in good agreement with the
experimental data. The same can be stated for the discharge and suction pressures.

Keywords  Capillary tube · Heat exchanger · Household refrigerator · Refrigeration charge · Transient numerical simulation
Abbreviations D Tube diameter
A Constant (−) E Constant (−)
Ai Internal tube cross-sectional area F Constant (−)
As Internal tube superficial area g Gravity (m/s2)
B Constant (−) G Mass velocity (kg/m2 s)
Cf Friction coefficient (−) h Specific enthalpy (J/kg)
Cp Pressure coefficient (−) H Height (m)
dz Control volume length (m) m Mass (kg)
ṁ Mass flow rate (kg/s)
nrpm Compressor speed (rpm)
* Matheus dos Santos Guzella P Pressure (Pa)
[email protected] u Average velocity (m/s)
1
Institute of Science and Technology, Federal University
t Time (s)
of Jequitinhonha and Mucurí Valleys, UFVJM, Diamantina, T Temperature (°C)
Minas Gerais, Brazil T Average temperature (°C)
2
Heat Transfer Research Group, Department of Mechanical V Volume ­(m3)
Engineering, Engineering School of São Carlos, University VD Compressor displacement ­(m3)
of São Paulo, EESC-USP, São Carlos, São Paulo, Brazil UAgas Global conductance of the gasket (W/K)
3
Department of Thermal and Fluid Sciences, x Refrigerant quality (−)
Federal University of São João del Rei, UFSJ,
São João del‑Rei, Minas Gerais, Brazil

13
Vol.:(0123456789)
 Arabian Journal for Science and Engineering

Q̇ Heat transfer rate (W) household refrigerators. Hermes et al. [9] developed a
Ẇ Power (W) first-principles model to simulate transient start-up and
cycling operations of a frost-free 440-l top-mount refrig-
Greek symbols
erator. Cycling transient operation was controlled consid-
ε Void fraction (-)
ering the freezer temperature, while a thermo-mechani-
ℏ Convection heat transfer coefficient (W/m2.K)
cal damper was used to control fresh food compartment
ρ Density (kg/m3)
temperature. Experimental data were used to calibrate
η Efficiency (−) 
the compressor model and validate the proposed model.
Subscripts Results of system energy consumption agreed within 10%
1 Inlet of experimental data, while air temperatures presented a
2 Outlet maximum deviation of ± 1%. In the recent years, similar
a Air models for simulating household refrigerators have been
amb Ambient published in the literature [9–13].
cab Cabinet Borges et al. [14] developed a first-principles simulation
comp Compressor model based on a semi-empirical quasi-steady approach.
e External The model was applied to predict the cycling behaviour of
g Global household refrigerators and aiming an energy consumption
gas Gasket study. Numerical results were confronted with experimental
h Homogeneous data and showed a maximum deviation of ± 2%. A sensitivity
i Internal analysis was also carried out to identify potential method
iso Isentropic for energy savings. In a subsequent work, Borges et al. [15]
l Liquid developed a quasi-steady-state simulation model for predict-
mod Modified ing the transient behaviour of a two-door household refriger-
v Vapour ator subjected to periodic door opening and evaporator frost-
vol Volumetric ing. The predicted power consumption and air temperature
r Refrigerant resulted in deviations within ± 10% and ± 2 K, respectively.
t Tube The individual components of household refrigerators
w Wall have also been analysed in some research studies. In this
case, a common approach is to consider the homogeneous
equilibrium model for two-phase condition. Hermes et al.
[16] proposed a numerical model for plate-type, roll-bond
1 Introduction evaporators based on the finite-volume method. Guzella
et al. [17] developed a finite-volume model for the simu-
Domestic refrigerators are refrigeration systems character- lation of wire-on-tube condensers. The authors performed
ized by low cooling capacities (50–250 W) and low quanti- numerical simulations with the developed model and com-
ties of charge (20–200 g) when compared with air condi- pared its numerical results with other numerical models
tioning systems [1, 2]. These systems are composed of a from the literature. The authors reported good agreement.
hermetically sealed compressor, a condenser, an evaporator For capillary tubes, Guzella et al. [18] proposed a slight
and a capillary tube as expansion device, which is placed in modification in Hermes et al. [19] in order to account for the
contact with the compressor suction line, forming an addi- possibility of choked flow at the outlet. The authors analysed
tional heat exchanger. different empirical correlations for the computation of the
In the past years, several numerical models were devel- thermal effectiveness of the capillary tube-suction line heat
oped for the simulation of domestic refrigerators as well as exchangers. The results were confronted against different
for their individual components. Melo et al. [3] proposed one databases and the numerical results showed good agreement
of the first domestic refrigerators simulation model, applied when compared with experimental data.
to a two-compartment domestic refrigerator operating with Gedik et al. [20] performed an experimental investiga-
CFC-12. The model was applied to reproduce pull-down tion of the performance of a household refrigerator using a
tests, and experimental data were used for validation and chimney-type condenser. The working fluid was HFC-134a
good agreement between results was observed. Later, simi- and the refrigerant charge was 139 g. Three condensers with
lar research studies focusing in household refrigerators have different sizes were tested and the results showed that the
been published [4–8]. efficiency and energy consumption of the refrigerator vary
Recent robust and sophisticated numerical models with the chimney height. The best performance was obtained
were developed for simulating the transient operation of for the condenser with height equal to 170 cm.

13
Arabian Journal for Science and Engineering

Harun-Or-Rashid [21] compared the performance of Typically, numerical simulations are carried at fixed refrig-
a wire-on-tube condenser and a louvre fin micro-channel erant charge [9]. This is due to the difficulty in numerically
condenser trough numerical simulations. The results were matching the refrigerant degree of superheating at the outlet
compared against experimental data and a good agreement of the evaporator with the refrigerant inventory. This main
was reported. The authors found that the micro-channel difficulty is associated with the strong nonlinearity associ-
condenser occupied less space and required less refriger- ated with this numerical solution, leading to time-consuming
ant charge than the wire-on-tube condenser, for the same calculations and also to convergence issues.
thermal performance. Hence, this paper advances a simulation model developed
Some research studies focused on different refrigera- in GT-SUITE® platform capable to simulate the transient
tion technologies, such as absorption refrigeration systems operation of a single-compartment household refrigerator
[22–24]. Ahmadi et al. [22] performed an optimization of an operating with HFC-134a at different refrigerant charges.
absorption refrigerator with four-temperature-levels using an GT-SUITE® allows the development of simulation models
ecological function. The optimization goal was to maximize of specific refrigeration systems, including household refrig-
the coefficient of performance, the ecological function and erators. In this case, the simulations are performed using the
the thermoeconomic criterion. The results were compared geometrical and performance parameters as input data for
on the basis of error analysis and a good agreement was the components and the ambient temperature and refriger-
observed. Ahmadi [23] also performed a multi-objective ant charge in the system as initial condition. Experimental
optimization study, focusing on a three-heat-source irre- results from the literature are used to validate the numeri-
versible refrigerator. The NSGAII optimization algorithm cal model for pull-down simulation tests and for cycling
was chosen. The numerical results were compared with operation. The numerical results of inside air temperature,
results from other studies and a reasonable agreement was compressor discharge and suction pressures and compressor
observed. Ghorbani et al. [24] investigated an integrated power consumption are confronted against experimental data
cryogenic natural gas plant, focusing on the reduction in at different ambient temperatures, namely 32 °C and 43 °C.
energy consumption by using an absorption refrigeration Furthermore, results for these same variables are compared
system rather than a compression refrigeration system. The with experimental data for different refrigerant charges.
authors report a reduction in the specific power, capital cost
and prime cost of product by 38.94%, 31.9% and 15.31%,
respectively. 2 Mathematical Model
Other studies focused on more specific systems, such
as an integrated cryogenic natural gas process based on A schematic view of the refrigeration system of the house-
coupling of absorption–compression refrigeration system, hold refrigerator is presented in Fig. 1. All the geometrical
organic Rankine cycle and solar parabolic trough collectors characteristics of this system can be seen in Hermes [30].
[25] and a self-refrigeration system for a new process for The global simulation of the household refrigerator was
recovery of natural gas liquids from a treated feed gas [26]. developed in the commercial software GT-SUITE®. The
From the point of view of simulation analysis in these heat exchangers are modelled by the numerical solution of
equipments, although numerical models are increasingly
being developed to perform simulations, the use of com-
mercial software is still not common to perform household
refrigerator simulations. Using GT-Suite®, the resulting
model can be easily modified for household refrigerators
with different characteristics. One of the most important
advantages of GT-Suite® is the possibility to simulate the
household refrigerator at a given refrigeration charge and an
ambient temperature. In fact, some research studies focused
on the impact of the refrigerant charge over the performance
of refrigeration systems and heat pumps, such as [2, 27–29].
However, studies that present a complete transient simula-
tion model that allows the numerical analysis of a domes-
tic refrigerator with a certain refrigerant charge appear to
be scarce in the open literature. Thus, the application of a
numerical model to simulate the performance of a household
refrigerator during transient operation at different refriger-
ant charges is a desired contribution to the open literature. Fig. 1  Schematic representation of the refrigeration system

13
 Arabian Journal for Science and Engineering

mass, momentum and energy equations using a one-dimen- The energy equation is expressed as a function of the
sional distributed formulation in space and using an implicit refrigerant mass flow rate, specific enthalpy and pressure
formulation in time. Hence, the condenser, evaporator and [31].
capillary tube-suction line heat exchanger circuitries are ∑
d(𝜌Vh) dP ( )
divided into control volumes. The scalar variables (pressure, = mh
̇ +V + 𝜆i As,i Tr − Tw (4)
temperature, density and others) are assumed uniform in the dt boundaries
dt
control volumes. The vector variables (mass flux, velocity
In Eq. (4), V is the volume of the control volume, λi is
and others) are computed in the inlet and outlet sections of
the internal heat transfer coefficient, As,i is the internal heat
the control volumes. A schematic view of the spatial dis-
transfer surface area, Tr is the refrigerant average tempera-
cretization for scalar and vector variables is shown in Fig. 2
ture and Tw is the wall average temperature. It should be
[31]:
noticed that the presented model takes into account the pres-
sure drop in the heat exchangers, which is neglected in some
2.1 General Model for the Heat Exchangers
models in the literature, for example Hermes and Melo [9].
Similarly, an energy equation can be written for the heat
The mathematical model is based on the first-principles
exchanger external surface:
equations of mass, momentum and energy and is presented
( )
as shown in GT-SUITE® Manual [31]. The mass and d 𝜌Vw cp Tw ( ) ( )
momentum conservation equations are given by: = 𝜆i As,i Tr − Tw − 𝜆e As,e Tw − Ta (5)
dt
dm ∑
= ṁ (1) In Eq. (5), cp is the wall specific heat capacity, λe is the
dt boundaries external side heat transfer coefficient, A s,e is the external
heat transfer surface area and Ta is the external air average
� �
Ai dP +
∑
̇ − 4Cf 𝜌u�u�
(mu)
Ai dz
− Cp 𝜌u�u�
Ai temperature. For the condenser, it is assumed that Ta is the
dṁ boundaries
2 Di 2 ambient temperature, since the heat transfer mechanism is
=
dt dz natural convection. For the evaporator, Ta is the air tempera-
(2) ture inside the thermal cabinet.
In Eq. (1), m represents the refrigerant mass inside the Empirical correlations are applied to predict heat transfer
control volume, t is the time and ṁ is the mass flow rate. coefficients on the refrigerant side, λi, and air side, λe. For
In Eq. (2), Ai is the internal tube cross section area, P is the the refrigerant side, single-phase and two-phase correlations
refrigerant pressure, u is the average velocity between con- are considered. The empirical correlations provide closure
trol volume inlet and outlet, Cf is the Fanning friction factor, relations for the numerical model and are summarized in
Cp is a pressure drop coefficient and dz is the length of the Table 1:
control volume. In Eq. (2), both flow regimes, laminar and In GT-SUITE®, two-phase flow is modelled as homoge-
turbulent, are covered [31]. neous and a void fraction correlation is applied to estimate
The Fanning friction factor is computed using the correla- the average density. The homogeneous model is a simpli-
tion proposed by Serghides [32] for single-phase flow and fied model that considers the two phases as a homogeneous
the one proposed by Friedel [33] for two-phase flow. mixture flowing with same velocity [39]:
According to [31], the pressure drop coefficient is com- [ ( )( 𝜌 )]−1
puted as: 1−x
(6)
v
𝜀= 1+
x 𝜌l
P1 − P2
Cp = 1 (3) The subindexes l and v represent the saturated liquid and
𝜌 u2
2 1 1 vapour properties, and x is the vapour quality of the refrig-
In Eq. (3) the subindexes 1 and 2 represent the inlet and erant mixture. For two-phase flow, the mixture density is
outlet variables, respectively. computed using the void fraction:

Fig. 2  Discretization methodol-
ogy in space for the variables of
the numerical model

13
Arabian Journal for Science and Engineering

Table 1  Empirical heat transfer correlations employed in the numerical model

Single-phase internal convection heat transfer coefficient for the condenser model Dittus e Boelter [34]
Two-phase internal convection heat transfer coefficient for the condenser model Dobson and Chato [35]
Single-phase internal convection heat transfer coefficient for the evaporator model Dittus e Boelter [34]
Two-phase internal convection heat transfer coefficient for the evaporator model Klimenko [36]
External convection heat transfer coefficient for the condenser model Tanda and Tagliafico [37]
External convection heat transfer coefficient for the evaporator model Churchil and Chu [38]

The model only allowed the prescription of constant heat

𝜌 = (1 − 𝜀)𝜌l + 𝜀𝜌v (7)
transfer coefficients for external and internal interactions
between external air and external wall and internal air and
2.2 Compressor Model internal wall. These parameters were set as 7 and 5 W/m2-
K, respectively, based upon numerical values obtained from
In this paper, a simplified compressor model in GT-SUITE® steady-state operation.
is considered, in which the volumetric and global efficiencies For global simulation of the household refrigeration, the
were specified and assumed as constant values. The mass flow simulation model requires the geometrical parameters of the
rate displaced by the compressor is calculated as follows, based condenser, the evaporator, capillary tube and compressor,
on the refrigerant density in the compressor suction line, ρinlet: as well as the refrigerated compartment geometrical data.
These parameters were obtained from [30]. As input data,
𝜌1 VD nrpm 𝜂vol
ṁ comp = (8) the ambient temperature and refrigerant charge are applied.
60
In Eq. (8), VD and nrpm are the compressor displacement and
the nominal speed, respectively. 3 Results and Discussion
The electrical power consumption Ẇ comp is computed tak-
ing into account the global efficiency, which is defined as the The numerical methodology applied in GT-SUITE® applies
product of the isentropic, mechanical and electrical efficien- a fully implicit scheme in time. The time integration step
cies. Hence, the compressor power consumption is given by: was 0.1 s and the convergence criteria were set to 1­ 0−4 for
(
h2,iso − h1
) all variables involved. These are standard values suggested
̇
Wcomp = ṁ comp (9) in [31]. Experimental data from [30] were obtained using
𝜂g
software Engauge Digiter Version 4.1, available in http://
The volumetric and global efficiencies, ηvol and ηg, are www.fftw.org.
assumed as constant quantities and equal to 85% and 70%,
respectively. 3.1 Simulation of Start‑up Transient

2.3 Cabinet Model The numerical simulations were performed at different

ambient temperatures, namely 32 °C and 43 °C. The stand-
A standard model available in GT-SUITE® is employed to ard refrigerant charge is 80.9 g of HFC-134a, which resulted
simulate the refrigerated compartment. In order to account in optimal operation of the refrigerator [30].
for additional heat transfer rate from the gasket and improve First, pull-down tests were simulated. In this case, the
the model available in the software, the quasi-static empirical action of the thermostat was disregarded. These tests allow
expression obtained by [30] was applied: the analysis of the temporal evolution of pressure, tempera-
( ) ture, power consumption, among other performance param-
Q̇ gas = UAgas Tamb − T cab (10) eters of the refrigeration system. Hence, these simulations
are associated with the transient start-up until the system
The global conductance, UAgas , of the gasket is given by reaches operation at steady-state condition. In Fig. 3, the air
[30]: temperature obtained from the numerical model is presented
( ) during the pull-down simulation, as well as experimental
UAgas = 0.011039 Tamb − T cab − 0.196766 (11) data from [30]. It is observed a good agreement between
numerical results and experimental data, especially for the
first hour of test. However, the inflexion of the curve of air

13
 Arabian Journal for Science and Engineering

namely 32 °C and 43 °C. The compressor model is unable

to represent the correct transient characteristics of the com-
ponent during the start-up. For both cases, the numerical
predictions show practically constant quantities. However,
the steady-state value is correctly represented, especially for
case b. It should be mentioned that the volumetric and global
efficiencies are also constant in the present model, which is
also a limitation of the employed model.
For both ambient temperatures, the coefficient of perfor-
mance (COP) was computed after the system reached steady-
state operation: 0.87 and 0.88, respectively, for Tamb = 32 °C
and ­Tamb = 43 °C.
In Fig. 5, the compressor discharge and suction pressures
obtained from numerical simulations of pull-down test are
compared with experimental data from [30]. Again, due to
Fig. 3  Model validation in the start-up transients at different ambient
the simplicity of the compressor model, results are unable to
temperatures
represent the peak associated with the discharge and suction
pressures during the first minutes of simulation, since it is
temperature is delayed for the numerical results for both based on a quasi-static approach. For both cases, the numeri-
ambient temperatures. cal predictions show again practically constant quantities.
In Fig. 4, the compressor power obtained from numeri-
cal simulations of the pull-down tests is compared with
experimental data from [30], for both ambient temperatures,

Fig. 4  Comparison of numerical and experimental data of compres- Fig. 5  Comparison of numerical and experimental data of discharge
sor power: a Tamb = 32 °C and b Tamb = 43 °C and suction pressure: a Tamb = 32 °C and b Tamb = 43 °C

13
Arabian Journal for Science and Engineering

However, the steady-state value is correctly represented,

especially for the suction pressure for both cases.

3.2 Simulation of the Cycling Operation

In this subsection, results of the cycling operation of the

household refrigerator are presented considering the ambient
temperature equal to 32 °C. These results were obtained con-
sidering a thermostatic valve acting as an on–off control over
the compressor. Following [30], the temperature set-point is
5 °C with a hysteresis of 2 °C. Since the experimental and
numerical cycles were different, the results were scaled by
the bounded time interval considered in the comparison of
the results [9]. Results of air temperature (Fig. 6), discharge
and suction pressures (Fig. 7), and compressor power con-
sumption (Fig. 8) are compared with the experimental data
from [30].
The results in Fig. 6 show that the numerical model is
capable to capture reasonably the maximum and minimum
temperatures associated with cycling operation. However,
the numerical model predicts a cycle which is shifted with
respect to the experimental data. This can be explained by
the fact that the air temperature experimentally obtained
is based on an average value calculated between different
measured points. The numerical model assumes a lumped
model for the thermal cabinet.
Regarding Fig. 7, the numerical model is able to accu-
rately predict, with reasonable agreement, the pressure
Fig. 7  Comparison of numerical and experimental during cycling
distribution at the inlet and at the outlet of the compres-
operation: a Discharge pressure and b Suction pressure
sor (Fig. 7). For the discharge pressure, it can be noticed
that the off-period obtained from the numerical simulations
is smaller than the experimental off-period. Theoretically, on a quasi-static approach. In this case, the model fails to
the power consumption curve should decrease instantly, as represent the behaviour of the mass flow rate in the com-
noticed in the experimental data. These effects can prob- pressor during the off-period. Also, an explanation might
ably be explained by the compressor model, which is based

Fig. 6  Comparison of numerical and experimental data of the air Fig. 8  Comparison of numerical and experimental data of the com-
temperature during cycling operation (Tamb = 32 °C) pressor power consumption during cycling operation (Tamb = 32 °C)

13
 Arabian Journal for Science and Engineering

be associated with the numerical convergence difficulties of

the system of equations during the off-period.
Finally, in Fig. 8, it is presented the compressor power
consumption during the cycling operation. Considering the
results of the compressor discharge and suction pressures,
the off-period of the power consumption obtained from the
numerical simulations is smaller than the experimental. This
numerical results predicted a small on–off cycle. For the
power consumption, this could result in an over-prediction
of the energy consumption.

3.3 Variation in the Refrigerant Charge

In this section, numerical results of performance parameters

of the refrigerator are presented by varying the refrigerant
charge in the system. In this case, the action of the thermo-
stat is not considered, in order to obtain steady-state per-
formance quantities. An unprecedented comparison with
experimental data of numerical results of discharge and
suction pressure, compressor power consumption and air
temperature at different refrigerant charge is performed. The
results are presented only for ambient temperature equal to
43ºC, since experimental data from [30] are available only
for this value.
In Fig. 9, it is analysed the behaviour of the compres-
sor power with increasing refrigerant charge. As expected,
the compressor power increases with the refrigerant charge,
since the discharge and suction pressures increase. It is noted Fig. 10  Comparison of numerical and experimental data of discharge
a good agreement between numerical results and experimen- a and suction b pressure at different refrigerant charges (Tamb = 43 °C)
tal data, especially for low refrigerant charge values. At high
refrigerant charges (higher than 85 g), the discrepancies are
more evident, although within ~10%. represent the correct behaviour of this quantity when the
In Fig. 10, the numerical results of discharge and suc- refrigerant charge is increased. Good agreement is observed.
tion pressures are compared with experimental data. For For charges inferior to 80 g, the numerical model underes-
the discharge pressure, the numerical model is capable to timates the discharge pressure, while the opposite occurs
for charges higher than 80 g. For the suction pressure, the
agreement between experimental and numerical data is even
more pronounced. The numerical model overestimates the
suction pressure for almost all the range of refrigerant charge
considered. Only for 95 g, it is observed an underestimation
of the suction pressure, although only a slight difference.
Finally, in Fig. 11, a comparison of the air temperature
obtained from numerical simulations and experimental data
from [30] is presented. Good agreement is observed in the
range of 70–90 g. For all range of refrigerant charge, the
numerical model tends to underestimate the air temperature.
This could be associated with the simplicity of the cabinet
model, which considers only one mesh point to represent
the domain conformed by the air. In this case, air tempera-
ture is obtained by an average of the measured air temper-
ature obtained at different positions in the interior of the
Fig. 9  Comparison of numerical and experimental data of compres- cabinet [30]. Despite the observed discrepancies and due to
sor power at different refrigerant charges (Tamb = 43 °C) the simplicity of the numerical model, it can be stated that

13
Arabian Journal for Science and Engineering

ongoing investigation aiming at making the use of house-

hold refrigerators more optimized from the energy con-
sumption point of view.

Funding  This study was funded by the Brazilian National Council for
Scientific and Technological—CNPq (processes 306675/2014–5 and
304972/2017–7).

Declarations 

Conflict of interest  The authors declare that they have no conflict of

interest.

Fig. 11  Comparison of numerical and experimental data of air tem-

References
perature at different refrigerant charges (Tamb = 43 °C)
1. Guzella, M.S.; Cabezas-Gómez, L.: Numerical simulation of
the performance of a household refrigerator during start-up
and cycling operations. In: Proceedings of IV Journeys in Mul-
the results obtained by varying the refrigerant charge are tiphase Flows (JEM 2017), paper 0047
satisfactory. 2. Bjork, E.; Palm, B.: Performance of a domestic refrigerator
under influence of varied expansion device capacity, refriger-
ant charge and ambient temperature. Int. J. Refrig. 29, 789–798
(2006)
4 Conclusion 3. Melo, C.; Ferreira, R.T.S.; Pereira, R.H.; Negrão, C.O.R.:
Dynamic behavior of a vapour compression refrigerator: a theo-
The numerical model developed in GT-SUITE® predicts retical and experimental analysis. In: Proceedings of International
Refrigeration and Air Conditioning Conference at Purdue, paper
well the transient behaviour of a single-compartment house-
52 (1988)
hold refrigerator. First, numerical results were compared 4. Lunardi, M.A.: Numerical simulation of the dynamic behavior of
against experimental data for pull-down tests. The compari- household refrigerators. MSc thesis, Federal University of Santa
son reveals that the numerical model can represent reason- Catarina, Florianópolis-SC, Brazil (in Portuguese) (1991)
5. Chen, Z.J.; Lin, W.H.: Dynamic simulation and optimal match-
ably the inside air temperature. The model failed to represent
ing of a small-scale refrigeration system. Int. J. Ref. 14, 329–335
the inlet and outlet pressures of the compressor, as well as (1991)
the power consumption during the start-up transient, due to 6. Yuan, X.; Chen, Y.; Xu, D.G.; Yian, L.X.: A computer simula-
the simplicity of the compressor model. The steady-state tion and experimental investigation of the working process of a
domestic refrigerator, p. 1198–1202. IIR International Congress
values were in good agreement with the experimental data. of Refrigeration, Montreal, Canada (1991)
Later, the numerical results for the cycling operation are 7. Jansen, M.J.P.; de Wit, J.A.; Kuijpers, L.J.M.: Cycling losses in
compared again with experimental data. In this case, the domestic appliances: an experimental and theoretical analysis. Int
comparison reveals a fair agreement. The numerical model J Refrig 15(3), 152–158 (1992)
8. Jakobsen, A.: Energy optimisation of refrigeration systems–the
tends to predict smaller cycles, leading to delayed on–off domestic refrigerator–a case of study. Ph. D Thesis, The Technical
periods. In the end, the authors carried simulations at dif- University of Denmark (DTU) (1995)
ferent refrigerant charges. The numerical predictions were 9. Hermes, C.J.L.; Melo, C.: A first-principles simulation model for
in reasonable agreement with the experimental trends for the start-up and cycling transients of household refrigerators. Int
J Refrig 31, 1341–1357 (2008)
almost all the range of the refrigerant charge considered. The 10. Li, B.; Alleyne, A.G.: A dynamic model of a vapor compres-
results obtained with these simulations are rarely published sion cycle with shut-down and start-up operation. Int J Refrig 33,
in literature articles of household refrigerators. Hence, the 538–552 (2010)
presented model can be used as a tool to obtain performance 11. Tulapurkar, C.; Kandelwal, R.: Transient lumped parameter
modeling for vapour compression cycle based refrigerator. In:
parameters of household refrigerators at different refrigerant Proceedings of International Refrigeration and Air Conditioning
charges. This can be important during equipment analyses Conference at Purdue, paper 1129 (2010)
and during design processes. 12. Lin, E.; Ding, G.; Zhao, D.; Liao, Y.; Yu, N.: Dynamics model
The realization of further investigations considering a for multi-compartment indirect cooling household refrigerator
using Z-transfer function based cabinet model. Int J Therm Sci
more robust compressor model, as well as other relations 50, 1308–1325 (2011)
for void fraction and system thermal inertia computation, 13. Tagliafico, L.A.; Scarpa, F.; Tagliafico, G.A.: A compact dynamic
should contribute to increase the knowledge of household model for household vapor compression refrigerated systems.
refrigerators operation. This alternative is the focus of an Appl. Therm. Eng. 35, 1–8 (2012)

13
 Arabian Journal for Science and Engineering

14. Borges, B.N.; Hermes, C.J.L.; Gonçalves, J.; Melo, C.: Transient refrigeration system, organic Rankine cycle and solar parabolic
simulation of household refrigerators: a semi-empirical quasi- trough collectors. J. Therm. 1–29 (2020)
steady approach. Appl. Energy. 88, 748–754 (2011) 26. Mehrpooya, M.; Vatani, A.; Sadeghian, F.; Ahmadi, M.H.: A
15. Borges, B.N.; Hermes, C.J.L.; Melo, C.: Transient simulation of a novel process configuration for hydrocarbon recovery process
two-door frost-free refrigerator subjected to periodic door opening with autoerefrigeration system. J. Nat. Gas. Sci. Eng. 42, 262–270
and evaporator frosting. Appl. Energy. 147, 386–395 (2015) (2017)
16. Hermes, C.J.L.; Melo, C.; Negrão, C.O.R.: A numerical simula- 27. Zhang, Z.; Wu, Q.B.; Li, L.P.; Kong, X.Q.: Effects of refrigerant
tion model for plate-type, roll-bond evaporators. Int. J. Refrig. 31, charge and structural parameters on the performance of a direct-
335–347 (2008) expansion solar-assisted heat pump system. Appl. Therm. Eng.
17. Guzella, M.S.; Cabezas-Gómez, L.; Silva, J.A.; Maia, C.B.; Han- 73, 522–528 (2014)
riot, S.M.: Numerical modeling of the thermal–hydraulic behav- 28. He, Y.; Liang, X.; Cheng, J.; Shao, L.; Zhang, C.: Approaching
ior of wire-on-tube condensers operating with HFC-134a using optimum COP by refrigerant charge management in transcritical
homogeneous equilibrium model: evaluation of some void frac- ­CO2 heat pump water heater. Int. J. Refrig. 118, 161–172 (2020)
tion correlations. Heat. Mass. Transf. 52, 183–195 (2016) 29. Wang, Y.; Ye, Z.; Song, Y.S.; Yin, X.; Cao, F.: Energy, exergy,
18. Guzella, M.S.; Cabezas-Gómez, L.; Guimarães, L.G.M.; Tibiriçá, economic and environmental analysis of refrigerant charge in
C.B.: A modified approach for numerical simulation of capillary air source transcritical carbon dioxide heat pump water heater.
tube-suction line heat exchangers. Appl. Energy. 102, 283–292 Energy. Convers. Manag. 223, 113209 (2020)
(2016) 30. Hermes C.J.L.: Development of mathematical models for numeri-
19. Hermes, C.J.L.; Melo, C.; Knabben, F.T.: Algebraic solution cal simulation of domestic refrigerators in transient operations.
of capillary tube flow part II: capillary tube suction line heat M.Sc Dissertation, Federal University of Santa Catarina (in Por-
exchangers. Appl. Energy. 30, 770–775 (2010) tuguese) (2000)
20. Gedik, E.; Kılıçaslan, E.; Acar, B.; Ergun, A.; Ozbas, E.: Experi- 31. GT-SUITE: Flow theory manual v7.3. Gamma Technologies Inc,
mental investigation of a household refrigerator performance Westmont (2012)
using chimney-type condenser. Arab. J. Sci. Eng. 41, 1691–1697 32. Serghides, T.K.: Estimate friction factor accurately. Chem. Eng.
(2016) J. 91(5), 63–64 (1984)
21. Harun-Or-Rashid, M.; Jeong, J.H.: Replacement of present con- 33. Friedel, L.: Improved friction pressure drop correlations for hori-
ventional condenser of household refrigerator by louver fin micro- zontal and vertical two-phase pipe flow. In: European Two-Phase
channel condenser. Arab. J. Sci. Eng. 44, 753–761 (2019) Flow Meeting, paper E2 (1979)
22. Ahmadi, M.H.; Ahmadi, M.; Mehrpooya, M.; Hosseinzade, H.; 34. Dittus, F.W.; Boelter, L.M.K.: Heat transfer in automobile radia-
Feidt, M.: Thermodynamic and thermo-economic analysis and tors of the tubular type. Int. Commun. Heat. Mas. 12, 3–22 (1930)
optimization of performance of irreversible four-temperature-level 35. Dobson, M.K.; Chato, J.C.: Condensation in smooth horizontal
absorption refrigeration. Energy. Convers. Manag. 88, 1051–1059 tubes. ASME. J. Heat. Transf. 120, 193–213 (1998)
(2014) 36. Klimenko, V.V.: A generalized correlation for two-phase forced
23. Ahmadi, M.H.; Ahmadi, M.A.: Multi objective optimization of flow heat transfer. Int. J. Heat. Mass. Tran. 31, 541–552 (1988)
performance of three-heat-source irreversible refrigerators based 37. Tanda, D.W.; Tagliafico, L.: Radiation and natural convection
algorithm NSGAII. Renew. Sustain. Energy. Rev. 60, 784–794 heat transfer from wire-and- tube heat exchangers in refrigeration
(2016) appliances. Int. J. Refrig. 20(7), 461–469 (1997)
24. Ghorbani, B.; Mehrpooya, M.; Shimohammadi, R.; Hamed, M.: 38. Churchill, S.W.; Chu, H.H.S.: Correlation equation for laminar
A comprehensive approach toward utilizing mixed refrigerant and and turbulent free convection for a vertical plate. Int. J. Heat.
absorption refrigeration systems in an integrated cryogenic refrig- Mass. Tran. 18, 1323–1329 (1975)
eration process. J. Clean. Prod. 179, 495–514 (2018) 39. Rice, C.K.: The effect of void fraction correlation and heat flux
25. Ghorbani, B.; Ebrahimi, A.; Skandarzadeh, F.; Ziabasharhagh, assumption on refrigerant charge inventory predictions. ASHRAE.
M.: Energy, exergy and pinch analyses of an integrated cryogenic Trans. 93, 341–367 (1987)
natural gas process based on coupling of absorption–compression

13