International Journal of Thermal Sciences 135 (2019) 285–297

Contents lists available at ScienceDirect

International Journal of Thermal Sciences

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

Unsteady investigation of the heat ventilation in a box prototype T

a,b,∗ a a b a b
Badis Bakri , Ahmed Ketata , Slah Driss , Hani Benguesmia , Zied Driss , Fareh Hamrit
a
Laboratory of Electro-Mechanic Systems (LASEM), National School of Engineers of Sfax (ENIS), University of Sfax, B.P. 1173, km 3.5 Road Soukra, 3038, Sfax, Tunisia
b
Mechanical Engineering Department, Faculty of Technologies, University of M'sila, Algeria

ARTICLE INFO ABSTRACT

Keywords: This paper investigates the unsteady state of the heat ventilation in a box prototype with two holes. The CFD
CFD simulations were conducted using ANSYS Fluent 17.0 software, which solves the Navier-Stokes equations in
Unsteady state conjunction with the standard k-ω turbulence model and by a finite volume discretization method. The pre-
Finite volume sented results, consisting of the distribution of the velocity fields, the temperature, the total pressure and the
Heat ventilation
turbulent characteristics, are very useful to determine the time required for the heating operation and to shrink
Box prototype
the energy consumption of the buildings. The comparison between the founded results affirms that the heating
time presents a straight effect on the velocity fields. However, for the temperature, the box prototype requires
more time and more energy to warm up. In our application, we confirm that the duration of 30 s is sufficient to
allow the heating of the box prototype. Indeed, the numerical results compared using the experimental data
developed in our laboratory confirms the validity of the numerical method. The good agreements validate the
considered computational method.

1. Introduction unsteady cases. Calay et al. [10] performed a numerical study on the
unsteady respiratory airflow dynamics within a human lung based on a
The energy use worldwide is increasing every year. For example, the three-dimensional asymmetric bifurcation model by computational
consumption has more than doubled in the last 40 years. The housing fluid dynamics method. The numerical results for the resting and
and service sector is one of the largest energy users on the world maximal exercise breathing conditions indicated that the airflow is
market. To shrink the energy consumption, the use of efficient systems strongly dependent on the geometry and Reynolds number. Orhan et al.
is very crucial [1,2]. In this context, Driss et al. [3] developed a new [11] solved the unsteady Navier-Stokes equations, governing the flow
system occupied by a solar patio. This system improved the micro- under Boussinesq approximation, with the vorticity-stream function
climate of the building. In other work, Driss et al. [4] proposed an formulation using the finite difference method. The development of the
outlining environment suitable building, to improve the thermal com- flow and temperature fields following these temperature changes are
fort. Rauf and Robert [5] studied the life cycle in a residential building determined numerically. Beak et al. [12] observed some fluctuation in
for a life range of 150 years. Premrov et al. [6] considered a timber- the overall heat transfer characteristics for a certain time. Ultimately,
frame house, taking into accounts the climate data for three different the radiation was found to augment the heat transfer rate, which led to
European cities. Zhang et al. [7] evaluated the overall performance of reduce the time required for the flow to reach the steady state. Zhu
eight prevalent and proposed models for simulating airflows in enclosed et al. [13] studied numerically transient laminar natural convection of
environments. Soni and Aliabadi [8] compared steady-state inspiratory air in a tall cavity. The availability of the numerical algorithm was also
and unsteady flows with an inlet Reynolds number of 319 at an idea- assessed by considering the natural convection of air in a square cavity
lized ten-generation bronchial tube model via large-scale CFD simula- which is differentially heated from side walls. Jun et al. [14] demon-
tions. Evergren et al. [9] studied both steady and unsteady flow through strated that the transient temperatures at the heaters may become
a three-generation system of non-symmetric bifurcations, and they higher than their steady-state values. The study emphasizes that the
confirmed that the steady-state solution is not representative of the transient-stage temperatures at the heaters can exceed the corre-

∗
Corresponding author. Laboratory of Electro-Mechanic Systems (LASEM), National School of Engineers of Sfax (ENIS), University of Sfax, B.P. 1173, km 3.5 Road
Soukra, 3038, Sfax, Tunisia.
E-mail address: [email protected] (B. Bakri).

https://doi.org/10.1016/j.ijthermalsci.2018.09.023
Received 2 April 2018; Received in revised form 4 August 2018; Accepted 17 September 2018
1290-0729/ © 2018 Published by Elsevier Masson SAS.
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Table 2
Standard k-ω turbulence model constants.
α0 α∞ α∗∞ Rω Rk σk σω

1/9 1.9 1.0 2.95 6.0 2.0 2.0

Fig. 1. Geometrical arrangement.

Fig. 3. Boundary conditions.

sponding steady-state values. Irfan et al. [15] study numerically a

passive heating room using values of hourly averaged radiation during
winter in Elazıg region in Turkey. The effect of the overall heat transfer
coefficient of glass on the Nusselt number was also investigated. It was
revealed that the overall heat transfer coefficient for low Rayleigh
number affected the average Nusselt number more than that of high
Rayleigh number. Inna et al. [16] solved numerically the unsteady
conjugate natural convection problem in a square enclosure filled with
a porous medium having finite thickness heat conducting solid walls at
the presence of the internal heat source with constant temperature in
conditions of convective–radiative heat exchange with an environment
on one of the external borders. Zhang et al. [17] analyzed the back heat
transfer and heat transfer resonance phenomena, and their relation-
ships with the time-periodic flow patterns and temperature distribu-
tions. The findings are helpful to the understandings of the fluid flow
and heat transfer mechanisms in the related enclosure configurations.
Semen et al. [18] developed a numerical analysis of transient laminar
natural convection and surface radiation in a closed cavity with heat-
Fig. 2. Anemometer type AM 4204. conducting solid walls of finite thickness and a local heat source of
constant temperature in convective heat exchange with an environ-
ment. A two-dimensional numerical analysis of combined heat transfer
Table 1 in an air-filled square enclosure having heat-conducting solid walls of
Characteristics of the anemometer. finite thickness and a local heat source in conditions of convective heat
Description Anemometer type AM 4204 exchange. Noh-Pat et al. [19] presented a numerical study of the un-
steady conjugate heat transfer for a square cavity with a semi-trans-
Manufacturer Lutron
parent (glass) wall. The cases with or without solar control film on the
Probe type Telescopic
Measurement parameters Air velocity, temperature, gaz flow semitransparent wall were studied from 8:00 to 18:00 h; a time with the
Resolution Air velocity 0,1 m s−1 multicellular flow than the cavity without film. Armengol et al. [20]
Temperature 0,1 °C studied the effects of variable properties of air in the transient problem
Precision Air velocity 5% of the differentially heated square cavity. The investigation is therefore
Temperature ± 0,8 °C
put forward covering additional effects under regard to the principle of
Measuring range Air velocity from 0,2–20 m s−1
Temperature from −20 °C to +70 °C energy conservation, such as the time evolution of the total energy and
its components for both transient regime and steady-state cases. Mikhail

286
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Fig. 4. Meshing of the computational domain.

Hakan et al. [24] studied numerically the conjugate heat transfer by

mixed convection and conduction in lid-driven enclosures with the
thick bottom wall with an enclosure which is heated from the iso-
thermally bottom wall. From the existing literature, it is shown that the
investigation of the unsteady CFD simulation of the heat ventilation is a
suitable field for further research. The time-dependent outcomes can
provide important information, while the temporal variation of indoor
temperature and airflow is directly connected to thermal comfort and
human life. Also, there are a few types of research on the design
method. Therefore, the study of a box prototype was investigated by
experimental survey and simulations. An unsteady-state model was
implemented using ANSYS software and the results were compared to
the experimental results. The present paper is organized as follows.
Initially, the examined problem is presented and described and the
corresponding conditions are reported. Then the CFD simulation is
analyzed and information about the design of the box prototype, the
governing equations, the boundary conditions, the numerical procedure
and the validation of the model are provided. Subsequently, the results
and discussion follow which examine the distribution of the tempera-
ture, the velocity fields, the total pressure and the turbulent char-
acteristics. Finally, a number of conclusions are derived providing
useful information for similar cases. The results are auspicious and
promising.

2. Box prototype system

Fig. 5. Visualization plane.

Fig. 1 presents the geometrical arrangement of the considered box

prototype. In this system, the height, the width, and the length are
et al. [21] analyzed natural convection of micropolar fluid in a right- equal to h = 0.22 m, w = 0.20 m and l = 0.30 m, respectively. In this
angled wavy triangular cavity. The obtained results were illustrated an system, two square holes, with a peak h3 = h4 = 0.02 m, are sited in the
essential effect of these governing parameters on velocity, temperature same face. From the floor of the box prototype, a first hole, sited at a
and vorticity patterns as well as an average Nusselt number at the hot distance h1 = 0.05 m, gives the heated air flow from the outside. From
wavy wall and fluid flow rate. Iker et al. [22] developed a numerical the base, a second hole, sited at a distance h2 = 0.18 m, allows the
study to obtain turbulent flow and heat transfer for double FFS with evacuation of the cooled air flow. In practice, the heated air flow comes
different aspect ratios and Reynolds number. The results reveal both from a solar air system or from a patio with transparent roofing. As
turbulent intensity and heat transfer are increased with the increasing presented by Driss et al. [3], this patio can store the heat energy which
of Reynolds number. Fatih et al. [23] investigated numerically the is transmitted to the box prototype. To follow the aerodynamic char-
natural convection of ferrofluid in a partially heated square cavity. The acteristics of the considered box prototype, we measured the tem-
effects of a magnetic dipole on the heat transfer enhancement and fluid perature and the air velocities in different controlled points. Particu-
flow characteristics have been studied in a partially heated cavity. larly, we have used the hot-wire anemometer type AM 4204 since it is

287
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Fig. 6. Velocity fields in the plane defined by x = 0.06 m.

simple and offers the greatest flexibility of use and high-resolution expressions into instantaneous equations and taking time average
(Fig. 2). The characteristics of the anemometer are summarized in yields, we can write in the Cartesian system:
Table 1. By dropping the probe in the appropriate position of the box
prototype, the velocity and the temperature values can be read directly + ( ui ) = 0
t xi (1)
from the digital screen.

3. Numerical work p ui uj 2 ui
( ui ) + ( ui uj ) = + µ + ij
t xj xi xj xj xi 3 xi
3.1. Governing equations
+ ( u i u j ) + Fi
xj (2)
The air flow modeling is governed by the continuity equation, the
momentum equations and the energy equation [25–28]. Substituting These equations are the RANS (Reynolds-averaged Navier-Stokes)

288
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

( )+ ( ui ) = +G Y +S
t xi xj xj (12)
Constants of the standard k-ω turbulence model are presented in
Table 2.
By using the concept of the Reynolds analogy [29], the energy
equation is written as follows:

c p µt T
( E) + [ui ( E + p)] = K+ + ui ( ij )eff + Sh
t xi xj Prt xj
(13)
Where E is the total energy, K is the thermal conductivity and ( ij ) eff is
the deviatoric stress tensor, defined as:

uj ui 2 uk
( ij ) eff = µeff + µ ij
xi xj 3 eff xk (14)

3.2. Boundary and initial conditions

Fig. 7. Variation of the maximum value of the magnitude velocity vs. the For the considered box prototype, a view of the boundary conditions
heating time.
is presented in Fig. 3. In the first hole, the heat flow supplies the box
prototype from the outside air heater, where the velocity inlet and the
temperature are equal respectively to V = 3.4 m s−1 and T = 310 K.
equations, which present the same form as the instantaneous equations. After that, the air flow is evacuated from the second hole of the box
The new terms introduce the Reynolds stresses Ui U j should be prototype, where the outlet pressure is equal to p = 101325 Pa. The
modeled to close the RANS equation. The used method applies the others side surfaces of the box prototype are assumed as walls sur-
Boussinesq hypothesis relating the Reynolds stresses with the mean rounding the computational domain. In these surfaces, Dirichlet
velocity gradients: boundary conditions are imposed and admit these values V = 0 m s−1
for the velocity and T = 290 K for the temperature. For the initial
ui uj 2 uk
u i u j = µt + k + µt ij
conditions, we have set all the variables to be equal V = 0 m s−1 for the
xj xi 3 xk (3) velocity, T = 290 K for the temperature, p = 101325 Pa for the pres-
sure and null for the turbulent characteristics. In our simulations, the
This hypothesis is undertaken in different turbulence models. This
time-step size is equal to ts = 1 s. However, the number of the time-
approach presents low computational cost compared to other numerical
steps is equal to Nts = 70 and the maximum iterations per time step are
methods [29]. By using the standard k-ω turbulence model, the tur-
equal to Nits = 20000.
bulent viscosity μt is defined by:
k 3.3. Meshing
µt =
(4)
For the correction of the low-Reynolds number, α∗ is calculated as Fig. 4 presents the meshing of the box prototype. In the considered
follows: application, the cells number and the nodes number are equal respec-
tively to 167400 and 64312. This choice is done after a comparison
0 + Ret / Rk between different meshing [30]. In fact, the numerical results are su-
=
1 + Ret / Rk (5) perimposed with the experimental results developed using our box
Where: prototype. The optimal mesh with good simulation results corresponds
to a minimum calculated time. This choice is adopted taking into ac-
k count the precision and the resolution time [25].
Ret =
µ (6)
Rk = 6 (7) 4. Numerical analysis

i The distribution of the velocity fields, the temperature, the total

=
0
3 (8) pressure, the turbulent kinetic energy, the turbulence eddy frequency
and the turbulent viscosity are studied over the time in the plane
= 0.072 (9)
i
x = 0.06 m of the box prototype, as presented in Fig. 5. In the con-
However, for the high-Reynolds number: sidered simulations, the Reynolds number is evaluated to be equal to
Re = 5100.
= =1 (10)
For the standard k-ω turbulence model, the equations of the tur- 4.1. Velocity fields
bulent kinetic energy k and the specific dissipation rate ω are written as
follows: Fig. 6 shows the velocity fields in the plane x = 0.06 m containing
the two holes at different instances equal to t = 1 s, t = 5 s, t = 10 s,
k
( k) + ( k ui ) = + Gk Yk + Sk t = 20 s, t = 30 s and t = 60 s. According to these results, it has been
t xi xi xj (11) observed that the average velocity at the hole inlet of the box prototype

289
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Fig. 8. Distribution of the temperature in the plane defined by x = 0.06 m.

is governed by the boundary conditions and reaches V = 3.4 m s−1. confirmed by Fig. 7 presenting the maximum value of the magnitude
Indeed, a discharge area appears in the hole inlet and invaded the re- velocity over the heating time. In these conditions, the correlated
verse wall. In this side, the velocity changes his direction and two axial equation for the considered box prototype is founded as follows:
flows have been observed. The first ascending flow is responsible for the
Vmax = 0.00001 t3 – 0.00098 t2 + 0.02859 t + 4.55 (15)
recirculation zone appeared in the whole area of the box prototype.
This movement continues until the exit of the air flow through the hole
outlet. The second descending flow is the cause of the dead zone ap-
peared in the down area. Globally, the averaged velocity value is about 4.2. Temperature
V = 1.3 m s−1 in the discharge area. Elsewhere, the averaged velocity
presents a very low value. The comparison between the founded results Fig. 8 shows the temperature distribution in the plane x = 0.06 m
affirms that the maximum value of the averaged velocity increases over containing the two holes at different instances equal to t = 1 s, t = 5 s,
the time from V = 4.58 m s−1 at t = 1 s to V = 4.79 m s−1 at t = 60 s t = 10 s, t = 20 s, t = 30 s and t = 60 s. According to these results, the
and reaches the steady state at this instance. These results are temperature of the hole inlet is governed by the boundary conditions,

290
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

value of the temperature over the heating time. In these conditions, the
correlated equation for the considered box prototype is founded as
follows:

Tmin = 0.0003 t3 – 0.0418 t2 + 1.6298 t + 297.99 (16)

4.3. Total pressure

Fig. 11 shows the total pressure distribution in the plane x = 0.06 m

containing the two holes at different instances equal to t = 1 s, t = 5 s,
t = 10 s, t = 20 s, t = 30 s and t = 60 s. From these results, a com-
pression zone emerges in the hole inlet of the box prototype. Through
the air advancement, the total pressure decreases softly in the expulsion
area, produced from the hole inlet and invaded until the reverse wall.
This fact can be explained by the recirculation zone appeared in the
whole area of the box prototype. After that, the total pressure increases
to close the opposite wall and obtains the maximum value. In the hole
outlet, a depression zone is observed. The comparison between the
founded results affirms that the time has not affected the total pressure.
In these conditions, the maximum value of the total pressure is equal to
p = 101342 Pa. However, the minimum value reaches p = 101323 Pa.

Fig. 9. Temperature profiles in the direction defined by z = −0.005 m and

4.4. Turbulent kinetic energy
x = 0.06 m.

Fig. 12 shows the turbulent kinetic energy distribution in the plane

x = 0.06 m containing the two holes at different instances equal to
t = 1 s, t = 5 s, t = 10 s, t = 20 s, t = 30 s and t = 60 s. From these re-
sults, it is noticed that the turbulent kinetic energy is stumpy in the box
prototype inlet. In the expulsion area, produced in the hole inlet and
invaded the reverse wall, it is observed a slight increase of the turbulent
kinetic energy. The maximum value of the turbulent kinetic energy
emerges in the discharge area near the reverse wall and the hole outlet
of the box prototype. This fact can be explained by the recirculation
zone appeared in the whole area of the box prototype. The comparison
between the founded results confirms that the maximum value of the
turbulent kinetic energy depends on the heating time. In fact, this value
decreases from k = 0.489 m2 s−2 at t = 1 s to k = 0.284 m2 s−2 at
t = 60 s and reaches the steady state at this instance.

4.5. Turbulent eddy frequency

Fig. 13 shows the turbulent eddy frequency distribution in the plane

x = 0.06 m containing the two holes at different instances equal to
Fig. 10. Variation of the minimum value of the temperature vs. the heating t = 1 s, t = 5 s, t = 10 s, t = 20 s, t = 30 s and t = 60 s. From these re-
time. sults, it is noticed that the turbulent eddy frequency is stumpy in the
box prototype inlet. In the expulsion area, produced in the hole inlet
and invaded the reverse wall, it is observed a slight increase of the
which is equal to T = 310 K. In the discharge area, appeared in the hole turbulent eddy frequency. The maximum value of the turbulent eddy
inlet and invaded until the reverse wall, it has been observed a slightly frequency emerges in the discharge area near the reverse wall and the
decrease of the temperature. Above the discharge area, the temperature hole outlet of the box prototype. The comparison between the founded
decreases further. This fact can be explained by the recirculation zone results confirms that the maximum value of the turbulent eddy fre-
appeared in the whole area of the box prototype. This cooling continues quency depends on the heating time. In fact, this value decreases from
until the hole outlet, where the temperature presents the minimum ω = 6197 s−1 at t = 1 s to ω = 6011 s−1 at t = 60 s and reaches the
value, depending on the heating time. For example, at t = 1 s this value steady state in this instance.
reaches T = 290 K. However, this value increases over the time and
reaches T = 310 K at t = 60 s, representing the steady state in this in- 4.6. Turbulent viscosity
stance. From these results and as indicated in Fig. 9, presenting the
temperature profile in the direction defined by z = −0.005 m and Fig. 14 shows the turbulent viscosity distribution in the plane
x = 0.06 m, it is clear that the box prototype requires more time and x = 0.06 m containing the two holes at different instances equal to
more energy to warm up. In these conditions, we confirm that the t = 1 s, t = 5 s, t = 10 s, t = 20 s, t = 30 s and t = 60 s. From these re-
duration of 30 s is sufficient to allow the heating of the box prototype. sults, it is noticed that the turbulent viscosity is stumpy in the box
These results are also confirmed by Fig. 10 presenting the minimum prototype inlet. In the expulsion area, produced in the hole inlet and

291
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Fig. 11. Distribution of the pressure in the plane defined by x = 0.06 m.

invaded the reverse wall, the turbulent viscosity increases immensely. defined by the intersection of the two planes z = −0.005 m and
The maximum values of the turbulent viscosity emerge in the discharge x = 0.06 m of the box prototype at different instances equal to t = 1 s,
area near the reverse wall and the hole outlet of the box prototype. This t = 5 s, t = 30 s and t = 60 s. From the obtained results, it is clear that
fact can be explained by the recirculation zone appeared in the whole the velocity presents the same profile. In these conditions, the max-
area. The comparison between the founded results confirms that the imum velocity value is equal to V = 3.7 m s−1. Unlike the temperature
maximum value of the turbulent viscosity depends on the heating time. profiles, time does not have a great influence on the velocity profiles
In fact, this value decreases from μt = 0.00119 kg m−1.s−1 at t = 1 s to variation. In fact, the velocity values slightly takedown at t = 1 s and
μt = 0.00066 kg m−1.s−1 at t = 60 s and reaches the steady state in this t = 5 s. Indeed, the founded results have been confronted to the ex-
instance. perimental data presenting the velocity profiles. Indeed, it has been
noted that the value founded with the numerical model is nearest to the
5. Comparison with experimental results experimental results. In these conditions, the gap calculated between
the computational and the experimental results, equal to 5%, confirms
Fig. 15 shows the evolution of the velocity profiles in the direction the validity of the numerical method.

292
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Fig. 12. Distribution of the turbulent kinetic energy in the plane defined by x = 0.06 m.

6. Conclusion heating continues until the hole outlet, where the temperature presents
the minimum value depending on the heating time. The comparison
In this work, the impact of the unsteady state of the heat ventilation between the founded results affirms that the heating time presents a
in a box prototype was investigated. The distribution of the velocity straight effect on the velocity fields. In fact, the maximum value of the
fields, the temperature, the total pressure and the turbulent character- averaged velocity increases over the time until reaching a steady value
istics were studied to characterize the aerodynamic structure of the box at t = 60 s. However, for the temperature, the box prototype requires
prototype. From these results, a discharge area appears in the hole inlet more time and more energy to warm up. In our application, we confirm
and invaded the reverse wall. In this side, the velocity changes his di- that the duration of 30 s is sufficient to allow the heating of the box
rection and two axial flows have been observed. The first ascending prototype. The comparison between the computational and the ex-
flow is responsible for the recirculation zone appeared in the whole perimental results, confirms the validity of the numerical method.
area of the box prototype. From the hole inlet, the temperature is in- These results can be used in the future projects to shrink the energy
vaded by the reverse wall, where it has been observed a slight increase. consumption of the buildings and drying units.
Above the discharge area, the temperature decreases further. This

293
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Fig. 13. Distribution of the turbulence eddy frequency in the plane defined by x = 0.06 m.

294
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

Fig. 14. Distribution of the turbulent viscosity in the plane defined by x = 0.06 m.

295
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

−1
.s−3)
Yk Turbulence dissipation of k (kg.m−1.s−3)
Yω Turbulence dissipation of ω (kg.m−1.s−3)
Z Cartesian coordinate (m)
α0 Constant of the k-ω turbulence model (dimensionless)
α∞ Constant of the k-ω turbulence model (dimensionless)
α∗∞ Constant of the k-ω turbulence model (dimensionless)
δij Kronecker delta function (dimensionless)
μ Dynamic viscosity (Pa.s)
μt Turbulent viscosity (Pa.s)
μeff Effective viscosity (Pa.s)
ω Specific dissipation rate (s−1)
ρ Density (kg.m−3)
βi Constant of the k-ω turbulence model (dimensionless)
σk Turbulent Prandtl number for k (dimensionless)
σω Turbulent Prandtl number for ω (dimensionless)
τij Viscous shear stress tensor (Pa)
(τij)eff Deviatoric stress tensor (Pa)
Φ Equivalence ratio (dimensionless)
Γk Effective diffusivity of k (Pa.s)
Γω Effective diffusivity of ω (Pa.s)
Ω Swirl number (dimensionless)
Ωij Rate of rotation tensor (s−1)

Fig. 15. Velocity profiles in the direction defined by Z = −0.005 m and

References
X = 0.06 m.

[1] L. Wang, N.H. Wong, Coupled simulations for naturally ventilated rooms between
Nomenclature building simulation (BS) and computational fluid dynamics (CFD) for better pre-
diction of indoor thermal environment, Build. Environ. 44 (1) (2009) 95–112.
[2] J.P. Varughese, M.M. John, Effect of emissivity of shading device and air flow inside
E Total energy (J) cavity of Double Skin Facade for energy saving and Thermal Comfort in buildings: a
Fi Force components on the i direction (N) CFD modeling, International Conference on Energy Efficient Technologies for
Gk Generation of the turbulent kinetic energy (kg.m−1.s−3) Sustainability (2016) 815–820.
[3] S. Driss, Z. Driss, I. Kammoun, Computational study and experimental validation of
Gb Generation of turbulence kinetic energy (kg.m−1.s−3) the heat ventilation in a living room with a solar patio system, Energy and building
Gv Production of turbulent viscosity (kg.m.s−2) 119 (2016) 28–40.
Gω Generation of the dissipation rate of the turbulent kinetic [4] S. Driss, Z. Driss, I. Kammoun, Numerical simulation and wind tunnel experiments
on wind-induced natural ventilation in isolated building with patio, Energy 90
energy (kg.m−1.s−3) (2015) 917–925.
H Height (m) [5] A. Rauf, R.H. Crawford, Building service life and its effect on the life cycle embo-
h Thermal enthalpy (J.kg−1) died energy of buildings, Energy 79 (2015) 140–148.
[6] M. Premrov, V.Z. Leskovar, K. Mihalic, Influence of the building shape on the en-
k Turbulent kinetic energy (m2.s−2)
ergy performance of timber-glass buildings in different climatic conditions, Energy
K Thermal conductivity (W.m−1. K−1) 108 (2016) 201–211.
L Length (m) [7] Z. Zhang, W. Zhang, Z.J. Zhai, Q.Y. Chen, Evaluation of various turbulence models
Nts Number of the time-steps (dimensionless) in predicting airflow and turbulence in enclosed environments by CFD: Part 2
Comparison with experimental data from literature, HVAC R Res. 13 (6) (2007)
Nits Number of maximum iterations per time step (dimensionless) 871–886.
P Pressure (Pa) [8] B. Soni, S. Aliabadi, Large-scale CFD simulations of airflow and particle deposition
Pr Prandt number (dimensionless) in lung airway, Comput. Fluids 88 (2013) 804–812.
[9] P. Evegren, J. Revstedt, L. Fuchs, Pulsating flow and mass transfer in an asymmetric
QH Heat source or sink per unit volume (kg.m−1.s−3) system of bifurcations, Comput. Fluids 49 (2011) 46–61.
qi Diffusive heat flux (J) [10] R.K. Calay, J. Kurujareon, A.E. Holdø, Numerical simulation of respiratory flow
Re Reynolds number (dimensionless) patterns within human lung, Respir. Physiol. Neurobiol. 130 (2002) 201–221.
[11] O. Aydm, Transient natural convection in rectangular enclosures heated from one
Rk Constant of the k-ω turbulence model (dimensionless) side and cooled from above, Int. Commun. Heat Mass Tran. 26 (1999) 135–144.
Rω Constant of the k-ω turbulence model (dimensionless) [12] S.W. Baek, T.Y. Kim, Effects of surface radiation on the unsteady natural convection
Si Mass-distributed (kg.m−2.s−2) in a rectangular enclosure, International Journal of Aeronautical and Space
Sciences 3 (2002) 95–104.
Sij Mean rate-of-strain tensor (s−1) [13] Z.J. Zhu, H.X. Yang, Numerical investigation of transient laminar natural convec-
Sω Source terms of the specific dissipation rate of the turbulent tion of air in a tall cavity, Heat Mass Tran. 39 (2003) 579–587.
kinetic energy (kg.m−1.s−3) [14] J.H. Bae, J.M. Hyun, Time-dependent buoyant convection in an enclosure with
discrete heat sources, Int. J. Therm. Sci. 43 (2004) 3–11.
Sk Source terms of the turbulent kinetic energy (kg.m−1.s−3)
[15] I. Kurtbas, A. Durmus, Unsteady heat transfer by natural convection in the cavity of
T Temperature (K) a passive heating room, Int. J. Therm. Sci. 47 (2008) 1026–1042.
T Time (s) [16] I.A. Aleshkova, M.А. Sheremet, Unsteady conjugate natural convection in a square
ts Time-step size (s) enclosure filled with a porous medium, Int. J. Heat Mass Tran. 53 (23–24) (2010)
5308–5320.
u Velocity components (m.s−1) [17] W. Zhang, C. Zhang, G. Xi, Conjugate conduction-natural convection in an en-
ui’ Fluctuating velocity components (m.s−1) closure with time-periodic sidewall temperature and inclination, Int. J. Heat Fluid
U Free-stream velocity (m.s−1) Flow 32 (2011) 52–64.
[18] S.G. Martyushev, M.A. Sheremet, Conjugate natural convection combined with
V Magnitude velocity (m.s−1) surface thermal radiation in an air filled cavity with internal heat source, Int. J.
xi Cartesian coordinate (m) Therm. Sci. 76 (2014) 51–67.
x Cartesian coordinate (m) [19] F. Noh-Pat, J. Xamán, G. Álvarez, M. Gijón-Rivera, I. Hernández-Pérez, J. Arce,
E. Villanueva-Vega, Unsteady-RANS simulation of conjugate heat transfer in a
y Cartesian coordinate (m) cavity with a vertical semitransparent wall, Comput. Fluids 117 (2015) 183–195.
YM fluctuating dilatation in compressible turbulence (kg.m [20] J.M. Armengol, F.C. Bannwart, J. Xamán, R.G. Santos, Effects of variable air

296
B. Bakri et al. International Journal of Thermal Sciences 135 (2019) 285–297

properties on transient natural convection for large temperature differences, Int. J. experimental validation of the turbulent flow around a small incurved Savonius
Therm. Sci. 120 (2017) 63–79. wind rotor, Energy 74 (2014) 506–517.
[21] M.A. Sheremet, I. Pop, A. Ishak, Time-dependent natural convection of micropolar [27] Z. Driss, O. Mlayah, S. Driss, D. Driss, M. Maaloul, M.S. Abid, Study of the bucket
fluid in a wavy triangular cavity, Int. J. Heat Mass Tran. 105 (2017) 610–622. design effect on the turbulent flow around unconventional Savonius wind rotors,
[22] I. Yılmaz, H.F. Öztop, Turbulence forced convection heat transfer over double Energy 89 (2015) 708–729.
forward facing step flow, Int. Commun. Heat Mass Tran. 33 (2006) 508–517. [28] Z. Driss, O. Mlayah, S. Driss, M. Maaloul, M.S. Abid, Study of the incidence angle
[23] F. Selimefendigil, H.F. Öztop, K. Al-Salem, Natural convection of ferrofluids in effect on the aerodynamic structure characteristics of an incurved Savonius wind
partially heated square enclosures, J. Magn. Magn Mater. 372 (2014) 122–133. rotor placed in a wind tunnel, Energy 113 (2016) 894–908.
[24] H.F. Oztop, C. Sun, B. Yu, Conjugate-mixed convection heat transfer in a lid-driven [29] ANSYS Fluent Theory Guide, ANSYS, Inc, Southpointe 2600, ANSYS Drive
enclosure with thick bottom wall, Int. Commun. Heat Mass Tran. 35 (2008) Canonsburg, PA, 2016, p. 15317.
779–785. [30] B. Bakri, S. Driss, A. Ketata, H. Benguesmia, F. Hamrit, Z. Driss, Study of the
[25] Z. Driss, O. Mlayeh, D. Driss, M. Maaloul, M.S. Abid, Numerical simulation and meshing effect on the turbulent flow in a building system with a k-ω turbulence
experimental validation of the turbulent flow around a small incurved Savonius model. International Conference on Mechanics and Energy (ICME’2016), December
wind rotor, Energy 74 (2014) 506–517. 22-24 2016, Hammamet, Tunisia.
[26] Z. Driss, O. Mlayeh, D. Driss, M. Maaloul, M.S. Abid, Numerical simulation and

297