You are on page 1of 14

Renewable Energy 80 (2015) 538e551

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

Numerical investigation about the improvement of the thermal


potential of an Earth-Air Heat Exchanger (EAHE) employing the
Constructal Design method
Michel Kepes Rodrigues a, Ruth da Silva Brum b, Joaquim Vaz c,
Luiz Alberto Oliveira Rocha a, b, Elizaldo Domingues dos Santos a, c,
 Isoldi a, c, *
rcio Andre
Lie
a
Programa de Pos-Graduaça~o em Modelagem Computacional (PPGMC), Universidade Federal do Rio Grande (FURG), Rio Grande, Rio Grande do Sul, Brazil
b
Programa de Pos-Graduaça~o em Engenharia Meca ^nica (PROMEC), Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Rio Grande do Sul,
Brazil
c
Escola de Engenharia, Universidade Federal do Rio Grande (FURG), Rio Grande, Rio Grande do Sul, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: The Earth-Air Heat Exchanger (EAHE) is a device used to improve the thermal condition of built envi-
Received 17 October 2014 ronments, allowing the reduction of electrical energy consumption of traditional air conditioner systems.
Accepted 19 February 2015 Fundamentally, its operational principle is based on fluid mechanics and heat transfer, areas in which
Available online 11 March 2015
Constructal Design has been widely used to seek for the optimal geometries, i.e., which leads to the best
performances. In spite of this fact, the employment of Constructal Design for improvement of the EAHE
Keywords:
thermal potential has not been performed into literature. Therefore, the main purpose of this work is to
Earth-Air Heat Exchanger (EAHE)
perform a numerical investigation on different geometrical configurations of an EAHE using the Con-
Solar energy
Numerical simulation
structal Design to obtain the highest thermal potential. Results indicated that, for the same area occupied
Constructal Design by the ducts and fixed mass flow rate of air, the increase of the number of ducts (complexity of geometry)
improved the EAHE thermal performance up to approximately 73% for cooling and 115% for heating.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction has been successfully used to define the best geometrical shapes,
i.e., the engineering system configuration that conducts to an
One of the main purposes of engineering is the improvement of optimized performance. Several examples of its application can be
systems performance to rationalize the available energetic re- found in literature. Into the heat transfer subject, Constructal
sources. In the past, the man-made designs were governed by Design has conducted the optimization of cooling cavities intruded
practice and intuition. In spite of the system performance be into solids with internal heat generation. Several shapes for cavities
analyzed and evaluated on a scientific basis, system design was have been studied. These studies comprise from geometries with a
kept at the level of art [1]. Nowadays, the advent of computational few degrees of freedom, C-shaped, elliptical, triangular and T-sha-
approach has allowed the evaluation of complex flow architectures ped [2,3], even those with complex patterns as H, Y, X, T and T-Y-
with several degrees of freedom. shaped cavities [4e8]. Constructal Design has also been employed
In this context, the search for the best geometries which lead to for improvement of simple and complex arrange of fins, e.g., T-
a superior performance must be included among the goals of any shaped [9], Y-shaped [10], twice Y-shaped [11], T-Y-shaped [12] and
engineering design. Concerning the heat transfer and fluid me- T-Y-shaped fins with lateral extended surfaces [13]. Constructal
chanics areas, the Constructal Design, a relatively novel method, Design has been also devoted to improve the performance in fluid
mechanics problems, for instance references [14e16] showed that
the pressure drop for flows inside ducts with round cross-section
* Corresponding author. Programa de Po s-Graduaç~ ao em Modelagem Computa- are minimal in comparison with those reached for other regular
cional (PPGMC), Universidade Federal do Rio Grande (FURG), Rio Grande, Rio polygonal cross-sections. Examples of vascular tree-shaped flow
Grande do Sul, Brazil. Tel./fax: þ55 5332336620, þ55 5391097356 (mobile).
E-mail address: liercioisoldi@furg.br (L.A. Isoldi).
architectures for laminar and turbulent flows can be seen in

http://dx.doi.org/10.1016/j.renene.2015.02.041
0960-1481/© 2015 Elsevier Ltd. All rights reserved.
M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551 539

Refs. [17e19]. Concerning the study of convection heat transfer,


Constructal Design has been employed for achievement of the best
arranges of channels in internal flows and cylinders in external
flows [20e22].
In spite of these applications and several others widely
described in literature, the employment of the Constructal Design
method specifically related with the harnessing of the thermal
energy stored in the subsoil, generated by the direct incidence of
the solar radiation over the soil surface, is a subject that has not
been sufficiently reported by the scientific community. The evi-
dence of this fact is that only one work was found, where the
Constructal Design was applied to determine the thermal coupling
between a heat pump and the soil during the annual temperature
cycle, indicating among other conclusions that for small volume
fractions of the heat sink the best shape is the most slender
possible; and for higher volume fractions the best shape evolves to
an approximately square shape [23].
Therefore, we believe that the use of Constructal Design method
to determine the optimal arrangement geometry that conducts to
the best thermal performance of an Earth-Air Heat Exchanger
(EAHE) is an original contribution of the present work, once no
reference covering this topic was found in the literature. To do so, a
tridimensional numerical model previously verified and validated
[24,25] was employed to analyze the thermal and fluid dynamic
behavior of five different arrangements for the buried ducts that
compose the EAHE. By means of Constructal Design, the objective
function, degrees of freedom and constraints were adequately
defined for each arrangement, allowing the performance compar-
ison among them. The effect of some geometric parameters over
the average monthly thermal potential (TP) of EAHE (which rep-
resents the average temperature difference between the ambient
external air and the EAHE outlet air) was also evaluated, allowing
the achievement of theoretical recommendations about future
constructions of device.
After this introductory section, a brief overview about EAHEs
with focus in its main operational principle is presented. Following,
more details about the computational model used are commented,
showing its mesh independence test as well as its validation.
Thereafter an explanation about the Constructal Design method Fig. 1. EAHE behavior during (a) warm, (b) cold and (c) intermediate periods.
followed by its application in the arrangements of the EAHE is
realized. In sequence the results for the five proposed EAHE con-
figurations are presented and discussed. Finally, the conclusions (Fig. 1b). However, during periods in which the external air presents
obtained during this research are reported. intermediate temperature magnitudes, the use of an EAHE is not so
effective (Fig. 1c).
2. Earth-Air Heat Exchanger It is worth to mention that as the depth increases, the amplitude
of thermal variation decreases exponentially and phased out in
The Earth-Air Heat Exchanger (EAHE) is a device that can be time in comparison with the variation on the surface as a result of
defined as an underground heat exchanger, composed by one or the thermal inertia of the soil. Therefore, in cold periods, the subsoil
more buried ducts, that can to harness (by capturing and/or dissi- does not seem as cold, and during warm periods the subsoil does
pating) the thermal energy of the superficial layers of soil, through not seem so hot, featuring a performance milder and more stable
the insufflated air flow inside it. Commonly, EAHEs are used on the compared to the variation that occurs in the external environment,
improvement of the thermal condition of built environments, being this characteristic thermal behavior which enables the use of
promoting a meaningful reduction in the electrical energy con- EAHEs.
sumption of traditionally used air conditioning equipment. These There are several studies involving experimental, analytical and
devices have a simple and efficient main operational principle: the numerical investigations about EAHEs, as well as review papers
external air is propelled to pass within buried ducts, exchanging about this subject. Apart those already mentioned in this work, it is
heat with the surrounding soil, and coming out of the ducts with a also possible to highlight the works presented in Refs. [27e37]. In
milder temperature which allows its use to enhance the thermal addition, there are publications where the effect of operational and/
condition inside the buildings (Fig. 1). As explained in Ref. [26], the or constructive parameters over the EAHE performance are evalu-
EAHEs can operate because the solar radiation that reaches the soil ated. In Refs. [25,26] different computational models developed in
surface is absorbed and stored as thermal energy. Due to its FLUENT software were used to determine the optimal installation
immense mass and insulating properties, relative to heat flux, the depth for an EAHE composed by one duct, being defined 3 m as
superficial layers of soil present lower temperatures than the ideal depth when constructive and operational parameters are
external air during warm periods (Fig. 1a), likewise, during cold taking into account. In Ref. [38] an one-dimensional analytical
seasons the subsoil temperature is warmer than the external air method was used to analyze the influence of the design parameters
540 M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551

of the EAHE on the thermo-hydraulic performance, indicating that


the augmentation of the number of ducts in parallel combined with
the reduction of the ducts diameter causes an improvement in the
EAHE performance. In Ref. [39] a quasi-steady state three-
dimensional model was developed in CFX software to simulate
the EAHE behavior, allowing to perform a parametric analysis to
examine the effect of duct length, duct radius, air flow velocity and
installation depth on the EAHE performance, concluding that the
longer duct of smaller diameter buried at higher depth and having
lower air flow velocity provide an increase in the EAHE perfor-
mance. However, publications involving the improvement of the
EAHE performance employing a geometric optimization technique
were not found.

Fig. 2. Computational domain.

3. Computational modeling
with tetrahedral cells) and its boundary conditions is presented in
The numerical simulations carried out in this work were per- Fig. 2.
formed using a computational modeling developed in the FLUENT The prescribed temperature functions, at the upper soil surface
soil
(Tsurface ambient ), are defined, respectively, by
) and at the inlet air (Tair
software, which is based on the Finite Volume Method (FVM). To do
so, initially a mesh independence test was developed. After that, (in  C):
using the independent mesh, the validation of the computational  
model was carried out.
soil
Tsurface ðtÞ ¼ 18:70 þ 6:28 sin 1:72  102 t þ 26:24 (1)
Taking into account the experimental research developed in the
city of Viam~ ao (geographic coordinates: 30 040 5100 S, 51 010 2400 W and
and an altitude of 111 m) in the southern Brazil [24,26], in the  
present work the EAHE computational domains were constructed ambient
Tair ðtÞ ¼ 23:18 þ 6:92 sin 1:72  102 t þ 26:42 (2)
and discretized in the GAMBIT software, being composed by one or
more ducts inserted in a clay soil portion. It is important to mention being t the time in seconds. A time step of 3600 s, with a maximum
that the thermal resistance of the duct is strongly lower than the of 200 interactions per time step, for a total numerical simulation
thermal resistances of the convective flow inside the duct and the time of 6.3072  107 s were adopted in all simulated cases.
soil. Moreover, the construction of mesh in the duct becomes The turbulent air flow is taken into account by means the Rey-
extremely complex when the duct walls are taken into account. As a nolds Stress Model (RSM). The Coupled algorithm was used for the
consequence, the simplification in which the air flows directly in a treatment of transient pressure and velocity fields, while the Up-
cylindrical perforation in the soil portion is assumed for all nu- wind scheme was employed to tackle with the advective terms of
merical simulations performed in the present study, as already conservation equations of momentum and energy, as well as, for
adopted in previous works [24,25,33,34,36]. In spite of this equations of the closure model. Besides, as in Ref. [24], the calcu-
simplification, the numerical method used here led to results of lations were considered converged when the residuals for mass,
temperature as function of time very similar to those predicted momentum, and energy between two consecutives iterations were
experimentally. A temperature function representing the annual lower than 103, 103, and 106, respectively. It is worth to mention
solar radiation which reaches the soil surface [24,25,36], obtained that in Refs. [25,26] a convergence condition of 106 was adopted
from experimental data [26], is imposed as boundary condition in for mass, momentum and energy, respectively, but no significant
the upper surface of the computational domain. Likewise, the differences were observed for the prediction of EAHE behavior if
annual air temperature variation is considered in the inlet of the compared with results generated with the convergence criterion
ducts, where also a prescribed velocity of 3.3 m/s is imposed as used in the present work; occurring however a relevant increase in
boundary condition [24,25,36]. The atmosphere pressure is processing time, which is undesirable in a geometric optimization
considered in the outlet of the ducts and the other surfaces of the study. More details about this computational model, including
computational domain are defined as thermally insulated. As initial boundary and initial conditions, material properties and numerical
condition the temperature of all computational domains is parameters, can be encountered in Refs. [24e26,36].
considered equal to the average temperature of the soil (18.70  C).
Besides, the thermo-physical properties of the air that flows inside
the EAHE and of the surrounding clay soil are shown in Table 1. 3.1. Mesh independence test
A schematic representation of the computational domain of an
EAHE composed by four ducts, its independent mesh (generated Fig. 2 presents an independent mesh generated with tetrahedral
computational cells in accordance with the pattern defined by the
mesh independence study performed. To define this pattern of
Table 1
Thermo-physical properties of soil and air. mesh generation, it was considered the numerical model presented
in Ref. [25] and a computational domain with only one duct. The
Material Property
duct and soil lengths are equal to 26 m. The duct has a diameter
Density (kg/m3) Thermal Specific Absolute viscosity d ¼ 200 mm while the height and width of soil are, respectively,
conductivity heat (J/kg K) (kg/m s)
15 m and 10 m. The duct is installed at a depth of 3 m from the
(W/m K)
superior soil surface. It is important to highlight that this compu-
Air 1800 2.1 1780 1.798  105 tational domain represents the Installation 1 of the present work,
Soil 1.16 0.0242 1010 e
as will be seen later in Sections 5 and 6.
M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551 541

Table 2 numerically simulated, as already been made in Ref. [25]. Hence, for
Mesh independence test. the validation process, the duct A is delimited by the blue marks,

T T duct B by the green marks, and duct C by the red marks, on Fig. 3.
Mesh refinement EAHE outlet temperature ( C) Relative difference j Tj jþ1

Duct Soil
Then, a computational domain similar to that depicted in Fig. 2,
but being composed by a triangular arrangement of the ducts A, B,
d/3 3d 25.40 1.97  103
and C was employed (Fig. 4). The clay soil portion has a length of
4d/17 12d/17 25.45 0.39  103
5d/26 15d/26 25.46 e L ¼ 25.77 m, a width of Ws ¼ 10 m, and a height of Hs ¼ 15 m, where
ducts with diameters of dA ¼ dB ¼ 110 mm and dC ¼ 100 mm were
installed. The ducts A and B are installed at 1.60 m from superior
Table 2 shows the numerical results for the EAHE outlet tem- soil surface, while the duct C is buried in a depth of 0.50 m from
perature obtained at final simulation time of 6.3072  107 s for superior soil surface, having this triangular installation a vertical
different mesh refinements for the duct and surrounding soil. spacing of Sv ¼ 1.10 m and a horizontal spacing Sh ¼ 0.60 m and
In Table 2 the variable d represents de duct diameter, Tj is the being at an average depth of Dave ¼ 1.05 m from the superior soil
EAHE outlet temperature defined with a current mesh refinement, surface (see Fig. 6c). Considering the indication of the indepen-
and TJþ1 is the EAHE outlet temperature obtained with the next dence mesh test described above, this computational domain was
mesh refinement. Therefore, the independent mesh was deter- discretized with 39,549 tetrahedral cells for the duct A (maximum
mined by successive refinements (h-refinement), increasing the cell size of 36.67 mm, i.e. dA/3), 39,549 tetrahedral cells for the duct
number of computational cells from the current mesh refinement B (maximum cell size of 36.67 mm, i.e. dB/3), 42,118 tetrahedral cells
to the next mesh refinement until to achieve a relative difference for the duct C (maximum cell size of 33.34 mm, i.e. dC/3), and
between two consecutive mesh refinements of 2.00  103. 1,539,859 tetrahedral cells for the soil (maximum cell size of
Must be informed that it was impossible to generate meshes in 300 mm, i.e. 3$dC), totalizing 1661075 computational cells.
the duct with computational cells larger than d/3. Eq. (1) was imposed as boundary condition at the upper soil
Then, observing Table 2 one can note that the independent surface, while the annual temperature variation for the inlet air in
mesh was defined with refinements of d/3 and 3d for the duct and the ducts A, B, and C were, respectively, prescribed by:
soil, respectively. This pattern for the mesh generation was already
adopted in previous works [24,25,36] and will be used for all nu- dA
 
merical simulations of the present research.
Tin air
ðt Þ ¼ 20:50 þ 5:66 sin 1:78  102 t þ 0:98 (3)

 
3.2. Validation of the computational model dB
Tin air
ðt Þ ¼ 21:79  5:96 sin 1:83  102 t þ 5:40 (4)
Experimental data of [40] were adopted to validate the
computational model employed in this work. This experiment was dC
 
2
conducted in city of Viama ~o during the year 2007. The external Tin air
ð t Þ ¼ 22:62 þ 6:08 sin 1:83  10 t þ 0:95 (5)
ambient air was circulated with the aid of a ventilator in three
buried polyvinyl chloride (PVC) ducts (A, B and C) to exchange heat emphasizing that Eqs. (3)e(5) were statistically adjusted from the
with the surrounding soil, featuring an EAHE. The outlet air of the experimental data measured at the beginning of the straight stretch
EAHE was used to improve the thermal conditions of a building (see Fig. 3) of the EAHE studied in Refs. [24,26,40], with Pearson's R
constructed for the study (called Casa Ventura) [24]. The temper- correlation coefficients of 0.87, 0.88, and 0.86 respectively. To
ature of the air flow in the ducts was monitored by four sensors promote all experimental data adjusts, i.e. Eqs. (1)e(5), a numerical
(with an accuracy of ±0.1  C) in each duct, located at the inlet of the code developed in the software MATLAB was implemented. This
duct, at the beginning of the straight stretch, at the end of the code uses the least squares method for the best fit of periodical
straight stretch, and at the outlet of the duct [24,26]. curves to the data presented in Refs. [26,40]. However, due the
As one can note in Fig. 3, the geometric configuration of this nonlinearities of the equations system generated, it was necessary
experimental EAHE is very complex. Therefore, for the validation of to obtain its approximate solution by means the Newton method.
the computational model a simplification was adopted: only the More details about the numerical code developed to fit the exper-
straight stretch of the experiment presented in Refs. [24,36,40] was imental data used in the present paper are available in Ref. [41].
In addition, prescribed velocities at the inlet air of 3.3 m/s,
3.6 m/s, and 2.5 m/s in the ducts A, B, and C were, respectively,
adopted. The other boundary and initial conditions, as well as the
numerical procedures, are the same earlier mentioned.
So, the annual variation of the EAHE outlet air temperature in
each duct was taken into account to compare the numerical results
obtained with the computational model used in the present work
with the experimental data of [26,40], as can be seen in Fig. 5. The
fitted curve to the experimental data has a Pearson's R correlation
coefficient of 0.94 (Fig. 4a), 0.95 (Fig. 4b) and 0.93 (Fig. 4c) for the
ducts A, B and C, respectively.
One can note a close agreement between numerical results and
experimental data for the annual air temperature variation in the
EAHE outlet. It is important to emphasize that the numerical results
were obtained from periodically measures, carried out every
21600 s, by a numerical probe located at the outlet of each duct of
the EAHE. From Fig. 5aec the absolute average differences around
1.6  C, 1.3  C, and 1.5  C, were encountered, respectively, for the
Fig. 3. Complex geometric configuration of the experimental EAHE. ducts A, B and C, representing absolute average differences of
542 M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551

Fig. 4. EAHE installation used for the numerical model validation process.

approximately 7.2%, 6.2%, and 6.9%, validating and showing the principle from which geometric shape and structure are deduced,
effectiveness of the computational model employed in the present but also an engineering method for optimizing the paths for flows
work. These average differences were evaluated comparing the through finite-size open systems [44,45].
temperature values numerically obtained, at the outlet of ducts A, B This theory indicates that if a system is free to morph under
and C, with those defined by the fitted experimental data repre- global constraints, the better flow architecture is the one that
sented by the follow curves, respectively: minimizes the global flow resistances, or maximizes the global flow
  access. A basic outcome of the Constructal Theory is that system
dA
Tout air
ðt Þ ¼ 21:02 þ 4:68 sin 1:82  102 t þ 0:71 (6) shape and internal flow architecture do not develop by chance, but
they result from the permanent struggle for better performance
  and therefore must evolve in time. As in engineered systems, in
dB
Tout air
ðt Þ ¼ 21:00 þ 4:57 sin 1:83  102 t þ 0:67 (7) nature the competition is permanent (e.g., river basins, global cir-
culations, trees and animals morph and improve in time under
dC
  changing constraints) [1,45].
Tout air
ðt Þ ¼ 20:52 þ 5 þ 95 sin 1:82  102 t þ 0:88 (8) The Constructal Design method guides the designer (in time)
toward flow architectures that have greater and greater global
Therefore, every 21600 s the relative difference (RD) between performance for the specified flow access conditions (fluid flow,
numerical and experimental values of temperature was calculated heat flow, flow of stresses, etc.). Constructal Design is about strat-
as: egy, the strategy learned from seeing and applying the Constructal
Law in basic flow configurations. It is about the compact lessons of
100½TnðtÞ  TeðtÞ
RDðtÞ ¼ (9) optimal shape and structure, which are fundamental and univer-
TeðtÞ sally applicable [16,45].
In other words, the Constructal Design is a method based on
where Tn(t) and Te(t) are the numerical and experimental tem- objectives and constraints (local and global) used to apply the
peratures as function of time, respectively. After that, the absolute Constructal theory. Constructal theory is the visualization that
average difference was defined considering all RD values for each generation of flux systems configurations is a physical phenome-
duct. non [14]. This physical principle is the Constructal Law which can
be intended as a generalization of a tendency of all things to flow
4. Constructal Design method along paths of minimal resistance. In this sense, this law has been
used for several applications in all the domains of design genera-
The Constructal Design method is based on Constructal Theory. tion and evolution in several fields: biology, physics, technology
This theory was created by Adrian Bejan in 1997, when a new evolution, social organization, sustainability and engineering
geometric solution philosophy was applied to the conductive [14,16,17].
cooling of electronics devices [14,42]. These works have a funda- According to Bejan and Zane [46], everything that has some kind
mental importance because they are considered as the starting of movement, animate or inanimate, is a flow system. These sys-
point for the employment of Constructal Theory to problems in tems generate shape and structure in time with the purpose to
engineering and other branches of science [16,43,45]. Ever since, facilitate their movement across a landscape filled with resistances.
the Constructal Theory has been adopted to explain deterministi- In this sense, the designs seen in nature are not result of chance.
cally the generation of shapes in nature [14,45]. They arise spontaneously because they enhance access, explaining
According to Constructal Theory the geometric patterns of all deterministically how configurations in nature are generated.
flux systems are ruled by a physical principle: the Constructal Law. This theory indicates, as already mentioned, that if a system is
Constructal Law states that: “for a flow system to persist in time (to free to morph under global constraints, the better flow architecture
survive) it must evolve in such a way that it provides easier and is the one that minimizes the global flow resistances, or maximizes
easier access to the currents that flow through it”. It is not only a the global flow access. In this sense, in the present work it is defined
M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551 543

Constructal Design method was applied by means of two volume


fractions parameters: the air volume fraction (f) and the installa-
tion volume fraction (j), mathematically expressed, respectively,
by:

Va npd2i L npd2i
f¼ ¼ ¼ (10)
Vs 4Ws Hs L 4Ws Hs

and

Vi Ai L Ai
j¼ ¼ ¼ (11)
Vs Ws Hs L Ws Hs

where Va is the internal volume of the ducts that form the EAHE, i.e.,
the volume where the air flows; Vs is the total volume of the soil
portion; Vi is the volume occupied by the EAHE installation; n is the
number of ducts of the installation; p is the mathematical constant;
di is the duct diameter of the installation; L is the soil length, as well
as, the duct length; Ws is the soil width; Hs is the soil height; and Ai
is the area occupied by the EAHE (dashed lines in Fig. 6cee). This
area can be defined for the Installation 3 (Fig. 6c) and Installation 4
(Fig. 6d) by:

Sh Sv
Ai ¼ (12)
2

and for Installation 5 (Fig. 6e) by:

Ai ¼ Sh Sv (13)

being Sh the horizontal spacing between ducts and Sv the vertical


spacing between ducts, as indicated in Fig. 6. For the Installations 1
(Fig. 6a) and 2 (Fig. 6b) the parameter j is not valid due its
configurations.
These volume fractions defined in Eqs. (10) and (11) can be
understood as constraints of the problem and they are an important
part of the Constructal Design method in the searching for the
superior performance of a system. In the present simulations the
increase of number of ducts will led to a decrease of the diameters
of ducts. Moreover, for each installation, the ducts have the same
diameter.
It is also important to take into account additional local con-
straints for the installations studied here: Sv must be smaller than
(2Dave e di); Sv and Sh must be larger than di; Sh must be smaller
than (Ws e di); and Dave must be larger than di/2.
Fig. 5. Numerical model validation: (a) duct A, (b) duct B and (c) duct C. In addition, the degree of freedom (DOF) assumed during the
application of the Constructal Design is the ratio between vertical
and horizontal spacing of the ducts (Sv/Sh) in each installation.
the following constraints: the volume of ducts where the air flows Obviously that in Installation 2 (Fig. 6b) this degree of freedom is
and the installation volume of EAHE. The main purpose here is to restricted to the Sh spacing, while in Installation 1 (Fig. 6a) it is not
maximize the TP of the EAHE for several studied geometries. The applicable because there is only one duct. The DOF Sv/Sh can vary
geometries are varied by changing the geometric degrees of freely however the above mentioned constraints must be taken
freedom of the problem. In this sense, several geometric possibil- into account.
ities for each configuration (with two, three and four ducts) are Finally, the average monthly thermal potential (TP) of the EAHE
generated and numerically solved in order to evaluate which con- is considered as the objective function for the Constructal Design
figurations lead to the highest thermal performance. According to method application. As already mentioned, the TP defines the
the constructal principle of optimal distribution of imperfections, it average temperature difference between the ambient external air
is expected that the most distributed temperature field will lead to and the EAHE outlet air.
the best thermal performance. To facilitate the understand about the application of Constructal
Design method, it is possible to visualize all of its stages in Fig. 7.
5. Constructal Design applied to the EAHE
6. Results and discussions
As already mentioned, five different configurations for the
installation of the EAHE (see Fig. 6) are numerically investigated in To generate the results the dimensions of the soil portion are
this work. To ensure an adequate and consistent comparison based in those proposed in the work of [25], being: L ¼ 26 m;
among the thermal performance of these installations, the Ws ¼ 10 m; and Hs ¼ 15 m, as well as, the average installation depth,
544 M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551

Fig. 6. Configurations of the EAHE: (a) Installation 1, (b) Installation 2, (c) Installation3, (d) Installation 4, and (e) Installation 5.

Dave ¼ 3 m (see Fig. 6). Besides, the air volume fraction is assumed a ReD ¼ 23,531. For all studied cases the fluid flow is turbulent, i.e.,
constant value of f ¼ 2:534  104 which is obtained considering with the same phenomenological fluid dynamic and thermal
the follow diameter for Installations 1 to 5, respectively: 220 mm; behavior. Besides, even considering a fixed value for f, three
155 mm; 127 mm; 127 mm and 110 mm (being the latter employed distinct values for j are considered: j ¼ 0.003; 0.007 and 0.010.
in the works of [24,25,36]), ensuring the same mass flow rate in all These values were defined assuming, respectively, the follow values
studied EAHE configurations. In order to keep fixed the mass flow for the area occupied by the EAHE: Ai ¼ 0.5 m2; 1.0 m2 and 1.5 m2, as
rate, the Reynolds number in each installation is, respectively, given defined by Eq. (12) or Eq. (13).
by: ReD ¼ 47,063; ReD ¼ 33,158; ReD ¼ 27,168; ReD ¼ 27,168 and In order to investigate how the ducts arrangement can affect the
TP of an EAHE and to define which configuration conducts to its
superior performance, several numerical simulations are carried
out. To do so, considering the volume fractions f and j and varying
the DOF Sv/Sh, a total of 151 cases were simulated. These simula-
tions are stratified in the following way: 1 (Installation 1); 15
(Installation 2); 45 (Installation 3); 45 (Installation 4) and 45
(Installation 5). In this sense, several sub-sections are inserted to
present and discuss the results for each installation. To summarize
a comparison among the installations is performed. In addition, the
use of EAHE is recommended only in warm and cold periods, so the
results are focused from May to July when TP for heating occurs, as
well as, in January, February, October, November and December
when a TP for cooling is generated by means the EAHE.
For the TP evaluation the inlet and outlet air temperatures in
each EAHE configuration were numerically monitored, with a
periodicity of 21600 s totalizing four measures per day, during the
year numerically simulated. These data were monthly grouped
enabling to define the average thermal potential.

6.1. Installation 1

Fig. 7. Schematic representation of the Constructal Design method employed to the In the Installation 1 (see Fig. 6a) the concept of installation
EAHE optimization developed in the present work. volume fraction can not be applied, as already commented. Only
M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551 545

one case was numerically simulated and the thermal potentials


obtained are presented in Table 3.
Table 3 indicates that the higher TP for cooling was reached in
December and the higher TP for heating was obtained in June.
These values and the other TPs presented in Table 3 will be used as
reference in the searching for a superior performance of the EAHE.

6.2. Installation 2

As occurred in Installation 1, the concept of installation volume


fraction neither can be applied for the Installation 2 (see Fig. 6b).
Hence, in Fig. 8, the thermal potentials of the EAHE are investigated
through the variation of the horizontal spacing between the ducts
(Sh).
One can note in Fig. 8 a very stable behavior for the TPs in
Installation 2. It is noticed an increase of TP as the value of Sh in-
creases, until a stabilization of TP magnitude. For the lowest values
of Sh the heat exchange between the fluid flow and the soil is
smoothed due to the interference between the ducts. The thermal
field in the portion of soil between the ducts is affected, leading to a
reducing of thermal potential of the soil in this specific region. As
the distance Sh increases, the interference of one duct in the other
decreases and TP increases, until a superior limit where the thermal Fig. 8. Thermal potential variation for Installation 2.
field in the portion of soil between the ducts is not affected by the
presence of the two ducts, i.e., the thermal field around both ducts June and July, being this ratio the optimized shape which generates
behaves similarly to that one presented with the insertion of only the highest TPs. However, the effect of the ratio Sv/Sh over the
one duct buried in the soil. The optimal configuration is achieved thermal potential (TP) is almost insensitive, i.e., the geometry
for (Sh)o ¼ 5.00 m that conducts to the superior performance of the variation does not cause a significant effect over the EAHE
EAHE, achieving the maximum TPs for heating and cooling. Thus, performance.
this spacing between the two ducts can be defined as the best If the results for j ¼ 0.007 in Fig. 9 were observed, it is possible
shape for Installation 2, allowing to reach maximum TPs for cooling to identify that the maximum values of TP are obtained for (Sv/
of 6.23  C and for heating of 1.68  C. Actually, there is no relevant Sh)o ¼ 1.00 for the months of May, June, July, October, November
difference in the TP for Sh > 5.0 m, however (Sh)o ¼ 5.00 m was and December, while for the months of January and February the
defined as the optimal geometry for the EAHE once it is the small EAHE geometry does not causes a significant change in the TP. This
horizontal spacing that maximizes the TP. It is important to same trend is noticed in Fig. 9 for j ¼ 0.010, i.e., for all analyzed
mention that a stabilization of the thermal potential (TP) from a months (Sv/Sh)o ¼ 1.00, being February the exception.
specific value of Sh, in this case for Sh  5.0 m, was expected. Be- For the design of this installation the arrangement area
sides, Fig. 8 was plotted up to Sh ¼ 7.00 m in order to avoid the constraint led to two extreme situations. In one hand, for the lowest
influence of the boundary condition imposed to the lateral surfaces ratios of Sv/Sh, it is possible to suppress the interactions among the
of the computational domain (which are prescribed as adiabatic). thermal fields of ducts in the soil. For this situation, the three ducts
tend to be aligned in the same depth of the soil, i.e., with the same
6.3. Installation 3 thermal potential. On the other side, for the highest ratios of Sv/Sh,
the two ducts of the bottom line are placed in a deeper region of the
For the Installation 3 (see Fig. 6c) the concept of installation soil than that proposed for lower ratios of Sv/Sh. However, the
volume fraction is applicable, hence the three values of j earlier interaction between the thermal fields of the two ducts can affect
mentioned were adopted, being the results showed in Fig. 9. As the thermal field of the soil, smoothing the heat transfer between
considered in Fig. 8, to eliminate the boundary condition influence the soil and duct (similarly to the behavior noticed for the lowest
of the computational domain lateral surfaces only values for Sv/ value of Sh in Installation 2). For this installation, results show that
Sh  1.00 were showed in Fig. 9. This procedure was repeated for the non-interaction between the ducts led to the best thermal
the Installations 4 and 5. performance of the EAHE. This behavior is more evident for designs
In Fig. 9, when j ¼ 0.003, the EAHE superior performances for with higher volume fraction of ducts arrangement, i.e., when the
heating were achieved for the smallest ratio (Sv/Sh)o ¼ 1.00 for May, ducts have more freedom.
In spite of the occurrence of a stable behavior for TP for several
situations depicted in Fig. 9, in general, it is achieved a theoretical
Table 3
recommendation about the best geometrical arrange of ducts for
Thermal potentials for Installation 1.
Installation 3. In order to evaluate the effect of parameter j over the
Month TP ( C) once maximized thermal potentials, (TP)m, and their respective
May 0.52 optimal geometries, (Sv/Sh)o, for Installation 3 the optimal results
June 1.10 achieved in Fig. 9 are summarized in Fig. 10. As expected, the
July 0.95 maximum values of TP for heating and cooling were generated for
January 4.19
February 3.35
j ¼ 0.010, except in the month of May when the maximized TP
October 2.65 occurs for j ¼ 0.007. However, the effect of j over (TP)m seems
November 3.79 almost insensitive, especially in periods of year where the soil
December 4.34 thermal potential is low as the months of May and October.
546 M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551

Fig. 10. Effect of j over (TP)m and (Sv/Sy)o for Installation 3.

Fig. 10 also illustrates that for all studied values of j the same
optimal ratios of (Sv/Sh)o ¼ 1.00 was obtained for months of May,
June, October and November. On the other hand, for months of
January, February and December the j parameter has an influence
in the definition of the optimized EAHE shape.

6.4. Installation 4

In Fig. 11 the numerical results for the Installation 4 of the EAHE


(see Fig. 6d) are depicted. One can note that for j ¼ 0.003 the
behavior of TP as a function of Sv/Sh presents an intermediate value
of (Sv/Sh)o ¼ 6.86 as the optimal shape. The exception occurs in July,
being its optimized shape defined by (Sv/Sh)o ¼ 1.00. For j ¼ 0.007
and j ¼ 0.010 the best shapes for the EAHE are achieved for the
lowest ratios of (Sv/Sh)o studied, (Sv/Sh)o ¼ 1.00, similar to the
Fig. 9. Thermal potential variation for Installation 3.
behavior noticed for the month of July and for j ¼ 0.003. For
j ¼ 0.003, there is one specific region where the increase of
M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551 547

Fig. 12. Effect of j over (TP)m and (Sv/Sy)o for Installation 4.

interaction between the two ducts of above line are compensated


by the thermal potential gained by the sinking of the two ducts.
From this point (for values of Sv/Sh > 6.86) the interaction between
the ducts is large enough to smooth again the thermal potential.
This behavior was not noticed for the other values of j, and it can be
related with the reduction of interaction between the ducts for the
values of j ¼ 0.007 and 0.010.
Fig. 12 shows how the j variation affects the values of (TP)m and
(Sv/Sh)o. In the cold period, the (TP)m for heating presents slightly
higher values for j ¼ 0.007, while during the warm period the
superior performance of the EAHE was achieved for j ¼ 0.010.
Moreover, for j ¼ 0.007 and j ¼ 0.010 the EAHE best shapes
were defined by (Sv/Sh)o ¼ 1.00. However, for j ¼ 0.003 the optimal
ratio of (Sv/Sh)o ¼ 1.00 was found only for July, being (Sv/Sh)o ¼ 6.86
for the other studied months.
Fig. 11. Thermal potential variation for Installation 4.
Comparing the optimal arranges achieved with Installation 3
and 4 shows that the kind of arrange has influence over its optimal
548 M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551

shape, i.e., does not exist one universal shape that maximizes the
thermal performance of EAHE. In this sense, for each proposed
arrange it is necessary to evaluate which geometry leads to the
highest TP.

6.5. Installation 5

Fig. 13 presents, for the Installation 5 (see Fig. 6e), the TP vari-
ation relative to the DOF Sv/Sh, for the three studied installation
volume fraction. The ratio Sv/Sh is changed from rectangular
arrangement with a large horizontal side (for the lowest ratios of Sv/
Sh) towards rectangular arrangement with large vertical side (for
the highest ratios of Sv/Sh). These configuration variations are also
indicated in Fig. 13.
In Fig. 13 it is possible to note that for j ¼ 0.003 there is two
optimal points for Sv/Sh ¼ 1.50 and Sv/Sh ¼ 4.50 where the TP is
maximized. This trend was observed for heating as well as for
cooling. Therefore, for values of Sv/Sh < 1.50 and Sv/Sh > 4.50 almost
constant values for TP were achieved for heating and cooling in
every studied months. Moreover, for the range 1.50 < Sv/Sh < 4.50
the EAHE has a poor performance in comparison with the optimal
points, reaching to magnitudes of TP similar to that reached for
geometries with highest values of Sv/Sh. Also in Fig. 13, one can note
that for j ¼ 0.007 and j ¼ 0.010 there is an optimal intermediate
ratio of Sv/Sh which maximizes the TP for the rectangular arrange-
ment of the EAHE.
So, from Fig. 13, the maximized thermal potentials (TP)m and the
correspondent optimized ratios (Sv/Sh)o can be defined, allowing
the analysis of the effect of installation volume fraction over these.
One can observe in Fig. 14 that for the months of May, June and
July the (TP)m for heating is always achieved for j ¼ 0.003, while
during the months of January, February, October, November and
December the (TP)m for cooling is always reached when j ¼ 0.010.
Besides, still observing Fig. 14, for j ¼ 0.007 and j ¼ 0.010 an
unique value of (Sv/Sh)o was found in each case, being, respectively,
6.25 and 2.67, for all analyzed months. However, for j ¼ 0.003
different values of (Sv/Sh)o were obtained: 1.00 for May, June and
July, 1.62 for October, 4.50 for November and December and 8.00 for
January and February.

6.6. Comparison among installations

The Constructal Design method ensures, by means the instal-


lation (j) and air (f) volume fractions, a consistent performance
comparison among the different configurations for the ducts
arrangement of EAHEs proposed in this work, as showed in Fig. 15.
The same trend was noticed in all proposed installations: the
higher TPs for heating and cooling were achieved, respectively, in
June and December.
For j ¼ 0.003, in Fig. 15, it is possible to identify that the
Installation 5 achieves a superior performance for cooling and
heating, among all studied installations. Therefore, one can define
that the Installation 5 is the optimal arrangement configuration for
the ducts of the EAHE. Furthermore, the Installation 5 with
j ¼ 0.010 can be considered the global optimal geometry for
cooling, i.e., its performance is better than all other analyzed cases.
Similarly, the global optimal geometry for heating was obtained
when j ¼ 0.003 in Installation 5. The results also clearly revealed
that the employment of configurations with n  2 leads to a sig-
nificant augmentation of thermal performance of the EAHE in
comparison with the use of only one duct (n ¼ 1), this difference is
especially noted for cooling purposes.
An interesting behavior also can be noticed in Fig. 15 if the (TP)m
Fig. 13. Thermal potential variation for Installation 5.
for cooling of Installations 3 and 4 are compared: for the months of
October, November and December the Installation 4 presents a
M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551 549

Fig. 14. Effect of j over (TP)m and (Sv/Sy)o for Installation 5.

better performance; for the month of January there is no relevant


difference between these installations performances; and for the
month of February the Installation 3 has a superior performance.
This behavior was observed for all analyzed installation volume
fractions (j). It was noticed that the decrease of TP from January to
February was lower for Installations 1, 2 and 3, than for Installations
4 and 5. This fact indicates that the reduction of thermal perfor-
mance of the soil can leads to a requisition of more simple geom-
etries for situations where the TP of soil is decreasing.
Finally, Fig. 15 indicates that the increase of number of ducts,
and hence the decrease of the duct diameter to maintain a constant
value for the air volume fraction, causes an improvement in the
EAHE performance. Similar trend was previously observed in
Refs. [38,39]. This fact can be explained by the division of the air
flow which facilitates the heat transfer between the air flowing in
the duct and the surrounding soil of the duct. When there is more
Fig. 15. Comparison among installations.
ducts with smaller diameter the heat exchange is more effective
550 M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551

than when only one duct with higher diameter is used, for the same [9] Bejan A, Almogbel M. Constructal T-shaped fins. Int J Heat Mass Transf
2000;43:2101e15.
flow regime conditions. This same trend was already observed in
[10] Lorenzini G, Rocha LAO. Constructal Design of Y-shaped assembly of fins. Int J
other researches where the Constructal Design method was Heat Mass Transf 2006;49:4552e7.
employed and is in agreement with the constructal principle of [11] Xie ZH, Chen LG, Sun FR. Constructal optimization of twice Y-shaped assem-
“optimal distribution of imperfections” [6,8,47e49]. blies of fins by taking maximum thermal resistance minimization as objective.
Sci China Technol Sci 2010;53:2756e64.
[12] Lorenzini G, Rocha LAO. Constructal Design of T-Y assembly of fins for an
7. Conclusions optimized heat removal. Int J Heat Mass Transf 2009;52:1458e63.
[13] Lorenzini G, Correa RL, dos Santos ED, Rocha LAO. Constructal Design of
complex assembly of fins. J Heat Transf 2011;133:081902 (7 pages).
In the present work an original application of the Constructal [14] Bejan A. Shape and structure, from engineering to nature. Cambridge: Cam-
Design method was proposed: the improvement of the average bridge University Press; 2000.
[15] Bejan A. Constructal theory of pattern formation. Hydrology Earth Syst Sci
thermal potential (TP) of an Earth-Air Heat Exchanger (EAHE). The 2007;11:753e68.
obtained results demonstrate that the Constructal Design method [16] Bejan A, Lorente S. Design with constructal theory. Hoboken: Wiley; 2008.
can be successfully applied in the search for a superior performance [17] Bejan A. Advanced engineering thermodynamics. Hoboken: Wiley; 2006.
[18] Reis AH. Constructal view of scaling laws of river basins. Geomorphology
of an EAHE. 2006;78:201e6.
The maximum TP for cooling was (TP)m ¼ 7.51  C, generated by [19] Cetkin E, Lorente S, Bejan A. Natural constructal emergence of vascular design
Installation 5 (see Fig. 6e). If this value is globally compared with with turbulent flow. J Appl Phys 2010;107:114901 (9 pages).
[20] Bello-Ochende T, Meyer JP, Ogunronbi OI. Constructal multiscale cylinders
the maximized TP of the other configurations, there is an increase in
rotating in cross-flow. Int J Heat Mass Transf 2011;54:2568e77.
the EAHE performance of 5.93% (Installation 4, Fig. 6d), 9.48% [21] Kim Y, Lorente S, Bejan A. Constructal multi-tube configuration for natural and
(Installation 3, Fig. 6c), 20.39%, (Installation 2, Fig. 6b) and 72.63% forced convection in cross-flow. Int J Heat Mass Transf 2010;53:5121e8.
[22] Rocha LAO, Lorente S, Bejan A. Tree-shaped vascular wall designs for localized
(Installation 1, Fig. 6a). On the other hand, the maximum TP for
intense cooling. Int J Heat Mass Transf 2009;52:4535e44.
heating was (TP)m ¼ 2.38  C (Installation 5) which promotes an [23] Rocha LAO, Lorente S, Bejan A, Anderson R. Constructal Design of under-
improvement of 15.53%, 20.20%, 41.67%, and 115.39% when ground heat sources or sinks for the annual cycle. Int J Heat Mass Transf
compared with the maximum TP achieved by Installation 4, 2012;55:7832e7.
[24] Vaz J, Sattler MA, dos Santos ED, Isoldi LA. Experimental and numerical
Installation 3, Installation 2, and Installation 1, respectively. analysis of an earth-air heat exchanger. Energy Build 2011;43:2476e82.
As in all proposed installations the TP for cooling always was [25] Brum RS, Rocha LAO, Vaz J, dos Santos ED, Isoldi LA. Development of
higher than the TP for heating, being this trend already observed in simplified numerical model for evaluation of the influence of soil-air heat
exchanger installation depth over its thermal potential. Int J Adv Renew En-
Refs. [24,25,36], and taking into account the parameters and con- ergy Res 2012;1:505e14.
straints considered in the present paper, one can recommend the [26] Vaz J, Sattler MA, Brum RS, dos Santos ED, Isoldi LA. An experimental study
arrangement proposed in Installation 5, with four ducts in a rect- on the use of Earth-Air Heat Exchangers (EAHE). Energy Build 2014;72:
122e31.
angular array, employing a ratio Sv/Sh ¼ 2.67 for a j ¼ 0.010 to [27] Pfafferott J. Evaluation of earth-to-air heat exchangers with a standardized
ensure the use of an EAHE with the best shape that provides its method to calculate energy efficiency. Energy Build 2003;35:971e83.
superior performance. [28] Sodha MS, Buddhi D, Sawhney RL. Thermal performance of underground air
pipe: different earth surface treatments. Energy Convers Manag 1991;31:
In future works other air and installations volume fractions, as
85e104.
well as, other arrangements will be investigated. [29] Mihalakakou G, Santamouris M, Asimakopoulos DN. Modelling the thermal
performance of earth-to-air heat exchangers. Sol Energy 1994;53:301e5.
[30] Santamouris M, Mihalakakou G, Balaras CA, Lewis JO, Vallindras M, Argiriou A.
Acknowledgments Energy conservation in greenhouses with buried pipes. Energy 1996;21:353e60.
[31] Gauthier C, Lacroix M, Bernier H. Numerical simulation of soil heat exchanger-
E. D. dos Santos thanks to FAPERGS (Process: 12/1418-4). E. D. storage systems for greenhouses. Sol Energy 1997;60:333e46.
[32] Florides G, Kalogirou S. Ground heat exchangers e a review of systems,
dos Santos, L. A. O. Rocha, L. A. Isoldi, and R. da S. Brum (doctoral models and applications. Renew Energy 2007;32:2461e78.
scholarship) thanks to CNPq. [33] Bansal V, Misra R, Agrawal GD, Mathur J. Performance analysis of earty-pipe-
air heat exchanger for winter heating. Energy Build 2009;41:1151e4.
[34] Bansal V, Misra R, Agrawal GD, Mathur J. Performance analysis of earty-pipe-
Appendix A. Supplementary data air heat exchanger for summer cooling. Energy Build 2010;42:645e8.
[35] Ascione F, Bellia L, Minichiello F. Earth-to-air heat exchangers for Italian cli-
mates. Renew Energy 2011;36:2177e88.
Supplementary data related to this article can be found at http:// [36] Brum RS, Vaz J, Rocha LAO, dos Santos ED, Isoldi LA. A new computational
dx.doi.org/10.1016/j.renene.2015.02.041. modeling to predict the behavior of earth-air heat exchangers. Energy Build
2013;64:395e402.
[37] Chiesa G, Simonetti M, Grosso M. A 3-field earth-heat-exchange system for a
References school building in Imola, Italy: monitoring results. Renew Energy 2014;62:
563e70.
[1] Bejan A, Lorente S. The constructal law (La Loi Constructale). Int J Heat Mass [38] De Paepe M, Janssens A. Thermo-hydraulic design of earth-air heat ex-
Transf 2006;49. 445e445. changers. Energy Build 2003;35:389e97.
[2] Biserni C, Rocha LAO, Bejan A. Inverted fins: geometric optimization of [39] Bisoniya TS, Kumar A, Baredar P. Parametric analysis of earth-air heat
the intrusion into a conducting wall. Int J Heat Mass Transf 2004;47: exchanger system based on CFD modelling. Int J Power Renew Energy Syst
2577e86. 2004;1:36e46.
[3] Rocha LAO, Montanari GC, dos Santos ED, Rocha AS. Constructal design [40] Vaz J. Estudo Experimental e Nume rico sobre o Uso do Solo como Reservatorio
applied to the study of cavities into a solid conducting wall. Therm Eng de Energia para o Aquecimento e Resfriamento de Ambientes Edificados.
2007;6:41e7. Doctoral Thesis. Universidade Federal do Rio Grande do Sul, Programa de Po  s-
[4] Biserni C, Rocha LAO, Stanescu G, Lorenzini E. Constructal H-shaped cavities Graduaç~ ao em Engenharia Civil; 2011.
according to Bejan's theory. Int J Heat Mass Transf 2007;50:2132e8. [41] Brum RS, Ramalho JVA, Rocha LAO, Isoldi LA, dos Santos ED. A Matlab code to
[5] Lorenzini G, Biserni C, Isoldi LA, dos Santos ED, Rocha LAO. Constructal Design fit periodic data. Revista Brasileira de Computaça ~o Aplicada 2015;7:1e10.
applied to the geometric optimization of Y-shaped cavities embedded in a [42] Bejan A. Constructal-theory network of conducting paths for cooling a heat
conducting medium. J Electron Packag 2011;133:8. 041008. generating volume. Int J Heat Mass Transf 1997;40:799e816.
[6] Lorenzini G, Biserni C, Garcia FL, dos Santos ED, Isoldi LA, Rocha LAO. [43] Ghodoossi L. Conceptual study on constructal theory. Energy Convers Manag
Constructal Design of isothermal X-shaped cavities. Therm Sci 2014;18: 2004;45:1379e95.
349e56. [44] Bejan A, Lorente S. Constructal Theory of generation of configuration in nature
[7] Lorenzini G, Biserni C, Link FB, Isoldi LA, dos Santos ED, Rocha LAO. Constructal and engineering. J Appl Phys 2006;100:041301.
Design of T-shaped cavity for several convective fluxes imposed at the cavity [45] Isoldi LA, Real MV, Correia ALG, Vaz J, dos Santos ED, Rocha LAO. Flow of
surfaces. J Eng Thermophys 2013;22:309e21. stresses: constructal design of Perforated Plates Subjected to Tension or
[8] Lorenzini G, Rocha LAO. Geometric optimization of T-Y-shaped cavity ac- Buckling. In: Rocha LAO, Lorente S, Bejan A, editors. Constructal law and the
cording to Constructal Design. Int J Heat Mass Transf 2009;52:4683e8. unifying principle of design. New York: Springer; 2013. p. 295e321.
M. Kepes Rodrigues et al. / Renewable Energy 80 (2015) 538e551 551

[46] Bejan A, Zane JP. Design in nature: how the constructal law governs evolution [48] Lorenzini G, Biserni C, Estrada E, Isoldi LA, dos Santos ED, Rocha LAO. Con-
in biology, physics, technology, and social organization. New York: Double- structal design of convective Y-shaped cavities by means of genetic algorithm.
day; 2012. J Heat Transf 2014;136:071702.
[47] Lorenzini G, Rocha LAO, Biserni C, dos Santos ED, Isoldi LA. Constructal Design [49] Lorenzini G, Biserni C, Estrada E, dos Santos ED, Isoldi LA, Rocha LAO. Genetic
of cavities inserted into a cylindrical solid body. J Heat Transf 2012;134. algorithm applied to geometric optimization of isothermal Y-shaped cavities.
071301e1-071301-6. J Electron Packag 2014;136:031011.

You might also like