Studies & Applications On Renewable Energy
Studies & Applications On Renewable Energy
Studies & Applications On Renewable Energy
59
ISSN : 1737-9934
Editor :
Dr. Ahmed Rhif (Tunisia)
ICID
International Centre for Innovation & Development
-PET- Vol. 59
Editor:
Dr. Ahmed Rhif (Tunisia)
Study of the Asymptotic Stability of a Motor Electronic Circuit under the Effect of Different Voltages. Page 5
Lassouani fatiha, A. Ghomri, D. Kherbouche.
Experimental Investigations and Model optimisation of Thermal Operated of Bubble Pump. Page 10
Raoudha Garma, Doniazed Sioud, Khlifa Mejbri, Ahmed Bellagi.
The impact of the transaction costs per hedging effectiveness ratio on renewable energy’s hedging Page 29
decisions.
Asma Abdallah, Ahmed Ghorbel.
Evaluation of Acceptance Angle for High Concentration Photovoltaics Based on Triple Page 53
Junction Solar Cells.
Nouamane Kellil, Abd Elkader Aissat, Sakina Atoui, Belkacem Bouzidi, Adel Mellit.
Performance Comparison of STATCOM and SVC Controllers for Static Voltage Stability Page 57
Enhancement and Reactive Power Compensation.
Mimi Belatel.
A new harmonic reduction approach based on signal filtering in solar inverter's controllers. Page 64
Walid Rahmouni, Ghalem Bachir, Michel Aillerie.
Effect of Catalysts On theproduction of Biodiesel from Used Domestic Waste Oils. Page 71
D.KHERBOUCHE, A. GHOMRI, F. LASSOUANI, BIBI TRIKI, BENMOUSSA.
Faults Detection and Isolation of DC-AC Static Converter Feeding Induction Machine. Page 75
K. Bedoud, T. Bahi, H. Merabet, D. Drici, B. Oudjani.
The influence of Mn in stucture of hydrotalcitederived catalysts on the activity of dry reforming Page 80
of methane.
F. Touahra, F. Bali, B. Djebarri, N. Aider, Z. Abdelsadek, K. Bachari, D. Halliche.
Design of a Permanent Magnet Synchronous Generator for Wind Turbines. Page 84
Zina Larabi, Kaci Ghedamsi, Djamel Aouzellag.
Characterization of Protease milk clotting Obtained from Turkey proventriculus (Meleagris gallopavo). Page 90
MEKHANEG B, GIRARDET J-M, BELLAL M-M.
Voltage Control of a Self Excited Induction Generator in a Wind Turbine: Implementation Page 93
and Simulation.
Rachid El Bachtiri, Imane Idrissi, Mohamed Salhi, Tariq Riouch.
Preparation and characterization of nickel-based catalysts deposited on ZrO2 for methane steam Page 99
reforming.
Akila BELHADI, Ouiza CHERIFI.
Environment and Technology for Architecture and Cultural Heritage. Page 103
Amri Khaoulaa, Alkama Djamel.
Preparation and characterization of nickel-based catalysts deposited on vanadium oxide for hydrogen Page 116
production by methane steam reforming.
A. Belhadi, A. Boudjemaa, S. Boumaza, M.Trari, O. Cherifi.
An air cooler system for retrofitted PV panels in a Mediterranean climate. Page 119
Amel F. Boudjabi, Abderrahmane Dib, Thabet Lebeche, Kamal Rezgui, Chafik Tamrabet.
Optimization of two level production distribution systems with stock and transportation costs. Page 128
Bensakhria Mohamed, Samir Abdelhamid.
Olive pomace steam gasification for production of syngas/hydrogen : sensitivity analysis. Page 134
Rim Tilouche, Raoudha Garma, Doniazed Sioud, Housam Binous, Ahmed Bellagi.
METHANE DECOMPOSITION over NICKEL SUPPORTED on ACID TREATED HZSM-5. Page 141
Lilia ALALGA, Aicha BENAMAR.
Hybrid Power System Optimization of the Emission Antenna of Beni Chograne FM-Radio Page 147
(Cherb-Errih,Mascara, Algeria).
Hamza SIYOUCEF, Benameur AFIF, HabibAllah BOUHAMIDA, Boualem MERABET.
A two dimensional numerical study of an evaporative cooler with a wet return surface. Page 152
Amel F. Boudjabi, Chadi Maalouf, Kamal Rezguia, Fabien Beaumont, Guillaume Polidori, Djamila Rouag.
Effects of the Cl-edge shape and the width on the structural, electronic and magnetic properties of Page 176
silicene nanoribbons.
M. T. Kadri, I. Djabri1, H. Belkhir.
Etanolysis of sunflower oil overhomogeneous and heterogeneous heteropolyacid catalysts. Page 181
Yasmina Idrissou, Tassadit Mazari, Chérifa Rabia.
Modeling of transport phenomena in laser welding of dissimilar metals Al/Mg. Page 188
Sabrine Ben Halim, Sana Bannour, Kamel Abderrazek, Wassim Kriaa, Michel Autric.
Feasibility study of a solar brackish water desalination system plant in Algeria. Page 193
Mohamed Abbas, Asma Adda, Djilali Tassalit, Hanane Aburdeh, Zahia Tgrine.
Modeling and Simulation of Low Voltage Power MOSFET for DC-DC Converter Applications. Page 197
Smail toufik, Belmili hocine.
Evaluation of wind park performances located At Algerian East Highlands. Page 203
A. Mezidi, H. Bendjebbas, K. Mohammedi.
Energy Flow Management for a Standalone Hybrid PV/Wind Power System. Page 208
Sabri Boulouma, Hachemi Rahmani, Boualem Bendib, Hocine Belmili, Naoual Seddaoui.
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Abstract—In the present paper, we study the single electron impact [15-21]. This formalism allows to gain
ionization of methane molecule by proton impact in high accuracy and computation time. In this formalism, the
incident energy range. In the theoretical model presented molecular target is described by Slater-type functions
centered at the carbon atom, according to the description
here, the initial state is composed from projectile and
proposed by Moccia [22]. The continuum of the ejected
methane wave functions. The target is described by means of
electron is described by Coulomb wave function. The
a molecular wave function centered on the heavy nuclei. The incidence and scattering state of the proton are described by
final state is constituted from the ejected electron and the plane wave functions.
scattered proton. This electron is described by Coulomb wave
function and the fast and heavy proton is described by a II. THEORY
plane wave function. The obtained results are compared to The single ionization of the CH4 molecule by proton
the available experimental data. impact can be presented as:
( ) ( ) ( ) (1)
Keywords—Cross section, methane molecule, 1CW model,
Born approximation. where is the projectile momentum, is the scattered
proton momentum and is the ejected electron
momentum. In atomic unit, we have
I. INTRODUCTION
(2)
Studying the ionization of atoms and molecules by charged
particles is important for a large number of areas, such as
is the mass of the proton. The double differential cross
plasma physics, radiation physics, astrophysics, radiation
therapy, and planetary atmospheres [1-6]. Cross sections, section (DDCS) of the ejection process can be deduced
differential in energy and emission angle, give important from the triple differential cross section (TDCS) by
information about the different ionization processes. This is integration over the scattering solid angle ̂ :
why interactions of charged particles with molecules have
actually important consideration, especially for the ( ) ( )
Copyright - 2019
Page 1
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
with (̂ ) ( ̂ ). The molecular wave function discrepancy between the experiment and the theory. One
of the effects is the distortion of the ejected electron that
given in equation (13) is defined in a molecular reference
can be deduced from the comparison between our results
frame. We need to make a transformation to the laboratory and that of the CDW-EIS model (see curves in the
reference frame thanks to the relationship intervals form 0° to about 30°), where a distorted wave
function has been used to describe the ejected particle. We
(̂ ) ∑ ( ) ( ̂ ) (15)
must not forget that we are studying a reaction between a
proton and molecular target. This reaction differs from the
where is a rotation operator, , and are the one where the projectile is an electron by the fact that the
Euler angles. Since the target is randomly oriented, the proton has a positive electric charge and an important mass
measured cross section is an average over all the possible compared to the electron. In this case, the interaction
orientations. This is why we need to average the between the proton and the active electron can induce the
theoretical cross section over all the Euler angles so-called capture phenomenon. This phenomenon is
considered as unimportant when the projectile comes with
( )̅ ( )
∫ (16) a very fast velocity. However, under certain energetic and
̂ ̂
geometric conditions, this phenomenon is going to have an
In the present formalism the ejected electron is described important effect. We believe that, this is why for a small
by a Coulomb wave: ejection energies (Ee = 20 and 11.3 eV), our theoretical
Figure (a). Double differential cross section of the single ionization of the CH4 molecule by 2 MeV proton impact. Solid-line
corresponds to our results. Dashed-line for the CDW-EIS model of Tachino et al [13]. Solid circles the experimental data of Senger
[21]. The left and right results are presented in logarithmic and linear scales, respectively.
results and those of Tachino et al [13] are in poor [5] R. Zellner, G. Weibring, Z. Phys. Chem. 161, 167, 1989.
agreement with the experimental data. Without forgetting [6] S.P. Khare, M. K. Sharma, S. Tomar, J. Phys. B: At. Mol. Opt.
Phys. 32, 3147, 1999.
that the more the ejection energy is small, the more the
[7] Z. Nikitaki et al. Free Radical Res. 50, S64–S78, 2016.
error on the experimental data becomes important. [8] S. Meylan, S. Incerti, M. Karamitros, N. Tang, M. Bueno, I.
Clairand and C. Villagrasa, Sci. Rep. 7, 11923, 2017.
IV. CONCLUSION [9] E. Alizadeh, T. M. Orlando and L. Sanche, Annu. Rev. Phys. Chem.
Single ionization of methane molecule by 2 MeV proton 66 379–398, 2015.
impact was considered for 6 ejection energies. Our results [10] L. Xianming, D. F. Shemansky, J. Geophys. Res. 111, A04303,
2006.
are compared to the theoretical results of Tachino et al
[11] Y. L. Yung, M. Allen and J. P. Pinto, Astrophys. J. Suppl. 55, 465–
[13] obtained from the CDW-EIS model and to the 506, 1984.
experimental data of Senger [21]. We find generally that [12] E. H. Wilson, and S. K. Atreya, J. Geophys. Res. 109, E06002,
our results are in well agreement with the experiment 2004.
better than those of Tachino et al [13]. [13] C. A. Tachino, J. M. Monti, Fojon O A, Champion C and Rivarola
R D, J. Phys: Conf. Ser.583 012020, 2015.
[14] N. B. Malhi, I. Ben-Itzhak, T. J. Gray, J. C. Legg, V. Needham, K.
This proves once more that our simple formalism can D. Carnes and J. H. McGuire, J. Chem. Phys.87 6502, 1987.
provide very accurate results, under certain geometric and [15] A. Boukhikhi, M. Sahlaoui, B. Lasri, A. Sekkal and M. Bouamoud,
energetic conditions. However, certain corrections are J. Phys B : At. Mol. Opt. Phys. 52, 015201, 2019.
needed to improve the accuracy of the formalism. As for [16] M. Sahlaoui, B. Lasri and M. Bouamoud, Can. J. Phys. 92, 136–
140, 2014.
example, the consideration of the distortion effect of the [17] M. Yavuz, N. Okumus, Z. N. Ozer, M. Ulu, M. Dogan, M.
ejected electron by the residual molecular ion. We may Sahlaoui, N. H. Benmansour, M. Bouamoud, J. Phys.: Conf. Ser.
also take into account the capture phenomenon. 488, 052031, 2014.
[18] M. Sahlaoui, M. Bouamoud, B. Lasri and M. Dogan, J. Phys. B: At.
V. REFERENCES Mol. Opt. Phys. 46, 115206, 2013.
[19] M. Sahlaoui and M. Bouamoud, J. Phys. B: At. Mol. Opt. Phys. 45,
[1] C.S. Enos, A.R. Lee, and A.G. Brenton, Int. J. Mass Spectrom. Ion
085201, 2012.
Process. 104, 137, 1991.
[20] M. Sahlaoui and M. Bouamoud, Can. J. Phys. 89, 723–727, 2011.
[2] K. Mitsuke, T. Takami, K. Ohno, J. Chem. Phys. 91, 1618J, 1989.
[21] M. Sahlaoui and M. Bouamoud, Can. J. Phys. 88, 905–910, 2010.
[3] D.H. Katayama, R.E. Huffman, C.L. O’Bryan, J. Chem. Phys. 59,
4309, 1973. [22] R. Moccia, J. Chem. Phys. 40, 2164, 1964.
[4] K. Tachibana, M. Nishida, H. Harima, Y. Urano, J. Phys. D: Appl. [23] B. Senger, Z. Phys. D: At. Mol. and Clusters. 9, 5631, 1988.
Phys. 17, 1727, 1984.
Abstract — Linearization methods play a very important role in The mathematical model of continuous current motor is
the study of nonlinear systems which are generally modeled by given by:
ordinary nonlinear differential equations. While many systems
may admit a linear behavior domain, linearity is always an
approximation of reality. The most classical approximation is
that determined by the Fréchet derivative of the nonlinear
equation.
In this paper, we applied the optimal derivative method to a
physical system of a motor electronic circuit. This system is
described by ordinary nonlinear differential equations. We are
interested in studying the asymptotic stability of this system.
The approximation obtained by optimal derivation gives
satisfactory results with respect to the exact results while
respecting the dynamics of the initial problem.
Fig.1 Motor electronic circuit
Keywords—Electric motor circuit, Ordinary differential
equation, Fréchet derivation, Optimal derivative, Asymptotic
stability. The application of Kirchof laws gives:
di(t )
I. INTRODUCTION U(t) = R i(t) + L + Ke (1)
In 1892, Lyapunov published his thesis (General Problem of dt
the Stability of the movement) where he introduced a very
rigorous definition of the problem of the stability of and
movement. This method is based on the search for a scalar d (t )
Im = Cm - Cf (2)
function of sign defined at real value. In some classes of
physical systems, it can be chosen as the energy of the system.
dt
Cm = Kc i Cf = Fm
Unfortunately, the difficulty of the research constitutes a real
Im: Moment of inertia of the motor shaft, Fm : Viscous
handicap since we do not know of a method for the
coefficient of friction, Cm : Engine couple, i: Induced current
construction in the general case; however, there are
and Cf : Viscous friction torque.
construction techniques applicable to particular cases. So, we
Equations (1) and (2) result in a differential system of two
can say that the study of stability in the case of nonlinear
equations whose variables are the induced current and the
systems poses a very difficult problem; this is why we use the
angular velocity .
most classical method based on linearization.
The electromechanical parameters of the motor are:
In this paper, we apply some concepts of stability for solving
and studying the stability of a nonlinear system. Then, we
searche a numerical method destined to modeling the
proposed nonlinear system. The optimal derivative can reveal
emerging behaviors.
R = 5 , L = 50 mH, Ke = 0.2 v/rad.s, Kc = 0.1 Nm/A,
II. CONTINUOUS CURRENT MOTOR Fm = 0.01 Nms/rad, Im = 5.10-2 kgm2.
Fig. 2 Variation of the angular velocity with respect to the induced current.
dx F x t
dt (8)
Fig. 4 Phase Plan. x0 x0
x R , F is a given function on an open subset R ,
n n
exists.
Copyright - 2019
Page 7
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
F y j t A y j t dt.
2
G j A A is obtained for initial conditions ( x0 , y0 )=(0.03,1.2). The
0
eigenvalues are
1 = -98.5256, 2= -0.2812. So, the system is asymptotically
stable, and makes equilibrium O appear as a stable node.
and assuming that ( y j) and invertible, we end up with:
Aj
0
F y t y t
j j
T
dt y j 1
n 2
EQ
xi ( t ) y i ( t )
i 1
or
x (t) represents the solution of the nonlinear system.
y (t) represents the solution of the linear system. Fig. 6 The solution ( x(t ),y (t ),z(t)) in the phase space fore the initial
conditions ( x0 , y0 )=(0.03,1.2).
IV. MAIN RESULTS OF THE DIGITAL STUDY
V. CONCLUSION GENERAL On the other hand, the calculation of errors shows that the
optimal derivative is a good approximation near the starting
Ordinary differential equations model a large family of point.
phenomena that manifests itself in several domains .
(electronics, biology [5], mechanics, etc.). They are often
REFERENCES
non-linear and innumerable and usually appear complex and
[1] T. Benouaz, S.M.A. Bekkouche, N.Gouali, Modelling of the Non
surprising effects. linear Systems by Optimal Derivative, Conmputational Procedure and
So, we can say that the study of stability in the case of Propreties, International Journal of Nonliear Science, V.1.15(2013,
nonlinear systems poses a very difficult problem; this is why N°3, pp. 271-283,ISSN 1749-3889 (print),1749-3897.
we use the most classical method based on linearization
[2] T. Benouaz, Optimal Derivative of a Non-Linear Ordinary
(optimal derivative). Differential Equation, Equadiff 99, International Conference on
It is an overall approximation, as opposed to the nonlinear Differential equations, Berlin (Germany) 1-7 August, Vol.2, pp.1404-
perturbation of a linear equation that differs from the classical 1407, World scientific Publishing Co. Pte. Ltd, 2000.
linear approximation near a point of equilibrium. The [3] T. Benouaz, O. Arino, Determination of the Stability of a Nonlinear
Ordinary Differential Equation by Least Square Approximation.
approach followed is of the least squares optimization type. Computational Procedure, Appl. Math. and Comp sci., N°1, Vol.5,
We also presented a study of stability using the concepts of pp.33-48, 1995.
stability for the study of physical linear systems. [4] T. Benouaz and Martin Bohner, “On the Relationship between the
Regarding the study of the behavior of the solutions of a Classical Linearization and Optimal Derivative”, Advances in
Dynamical Systems and Applications (ADSA), V. 2, N°1, pp. 41-57,
nonlinear equation (4) around a singular point where the 2007.
classical linearization does not make it possible to answer in [5] T.Benouaz, F.Lassouani, S.M.A.Bekkouche,M. Bohner, Application
this case. We therefore arrive at the conclusion that: of the optimal derivative to the study of a ratio-dependent model
The example presented shows the good agreement of the describing the evolution of HIV in Canada", Advances in Dynamical
Systems and Applications,( A.D.S.A.) Volume 6, Number 1, pp. 3–
optimal derivation with the nonlinear system (4) since the 12S, 2011.
layout of the solutions confirms the similarity of the two .
trajectories.
Copyright - 2019
ISSN 1737-9334 Page 9
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Abstract—Bubble pump is the key component for single pressure applied along all the tube length (pump with direct heating)
refrigeration systems. In order to estimate the performance of [11, 12] liquid is vaporized in the pumping tube. Bubbles
these systems, the performance of the bubble pump must first be increase in size and coalescence as they rise along the tube and
predicted. Successful analytical model of bubble pumps which finally push the liquid up.
better predict the required tube diameters, heat input, and
This apparently simple devise is quite complex from the
submergence ratio enhance the whole absorption refrigeration
system performance. In the present work, an experimental study point of view of analysis and modeling.
and a new theoretical model are presented to estimate the bubble Pfaff et al. [5] developed a mathematical model using the
pump mass flow rate for different heat flux inputs and then hydrostatic principle (manometer principle) to evaluate the
compared to the performance of experimental investigated bubble pump’s performance. In order to validate the
bubble pump for different tube diameters and submergence ratio. mathematical model, an experimental system was built and
tested.
Delano [13, 14] proposed an analytical model to evaluate
the bubble pump performance. The model was based on the
Keywords—Bubble pump, Experimental validation, Analytical
air-lift pump analysis developed by Stunning and Martin [15]
Model, Heat input, Tube diameter, Submergence ratio
which uses conservation equations of mass and momentum
I. INTRODUCTION and assigns a value to the slip ratio (S=VG/VL) between
phases velocities. The influence of the heat input, tube
The diffusion absorption technique is invented by the
diameter and submergence ratio (H/L) on the bubble pump
Swedish engineers von Platen and Minters [1] in the 1920s.
performance were investigated. To validate the analytical
The corresponding thermodynamic cycle is based on a
model, Delano conducted experiments with water, and it was
refrigerant/absorbent pair mixture as working fluids and an
shown experimentally to provide a good estimation of the heat
inert gas for pressure equalization. The most used working
input required to produce a desired liquid flow rate.
fluids for this called diffusion absorption refrigerator is the
A single pressure absorption heat pump analysis was
ammonia/water system with ammonia as refrigerant and water
presented by Schaefer [16]. She was also concerned by the
as absorbent. To maintain a uniform pressure throughout the
Einstein refrigeration generator design and in particular by the
machine, either hydrogen or helium is used as an auxiliary gas
bubble pump configuration. Her study builds on the work of
in the cold compartment [2]. Helium is proved to be preferable
Delano. She used a methodical approach to increase the
to hydrogen as auxiliary gas for the cycle [3].The circulation
efficiency of the Einstein cycle, therefore the relationship of
of the working fluids is carried out by a thermally driven
the diameter, submergence ratio, mass flow rate, and heat
bubble pump which can be powered by waste heat or solar
input have been analyzed for the bubble pump, and its physical
thermal energy. As a result of the absence of any mechanical
geometry has been altered to maximize its performance.
moving part, the refrigerator is silent and very reliable [4] in
White [17] performed experimental studies and analytical
addition to an economical and environmental friendly devise.
modeling of bubble pump using air-lift pump theory which
Thermal driven bubble pump simply consists of a vertical
operates on the same principals as a vapor bubble pump.
tube that is heated causing the formation of bubbles. The rising
The model (verified by the experimental data of the air-water
bubbles act as small gas pistons lifting liquid slugs to the
system) was then used to analyze the design considerations of
tube’s top. Heat can be supplied at the bottom of the tube [5-
a bubble pump using an ammonia-water mixture (Einstein
10] with a separate heating element where the pipe section is
cycle).
assumed to be insulated and vapor is supposed to be produced
only at the surface of the heating element and the pumping
tube serves simply as vapor/liquid riser. Heat may be also
Copyright - 2019
Page 10
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Pump characterisation
Pump tube diameter 8 mm
Pump tube length 1.4 m
Submergence ratio H/L 0.25/0.27
when the level rise from 35 to 38 cm the minimum required III. ANALYTICAL STUDY
heat is reduced from 100 to 75 W.
A. Mathematical Model
Experimental and theoretical studies on a bubble pump for a
diffusion-absorption refrigeration system was carried out by
Sathe [18] using Delano’s [13] methodology applied to the
Platen-Munters’ bubble pump, fig. 5. He concluded that a
correction of lose factor K is necessary for the model. The
correlation is found out by applying the experimental results
(methanol based bubble pump) to the analytical model for a
given pump tube diameter and a given submergence ratio.
Copyright - 2019
Page 12
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
V̇g ρG V̇G
ϵ̇ g = (29) ReG = 4 (46)
̇Vg +V̇f πd2 μG
E1 and E2 are given by: The correlation parameters in the two cases, water based
0.22
and methanol based bubble pump are summarized in the table
ρ 2.
E1 = 1.578Re−0.19
l0 ( f) (30)
ρg
TABLE2 VALUE OF COEFFICIENT OF CORRELATION K (WATER)
−0.08
ρ
E2 = 0.0273We ′
Re−0.51
l0 ( f) (31) Coef Value Coeff Value
ρ g
δ₁₀ -6.34168851E+00 δ₅₀ -9.43075199E+02
Where Rel0 and We′ are defined as:
ṁD δ₁₁ 9.52881153E+02 δ₅₁ 1.13622783E+05
Re = (32) δ₂₀ 1.51417591E+01 δ₆₀ -3.13213006E+02
μf
ṁ2 D δ₂₁ -2.07270761E+03 δ₆₁ 8.78401748E-03
We′ = (33)
σρf δ₃₀ 6.84549802E+00 δ₇₀ -1.39667124E+05
Where 𝑚̇ is the mass flow flux, D the equivalent diameter δ₃₁ -1.83113118E-04 δ₇₁ 1.01373745E+07
of the channel, 𝜇𝑓 the liquid phase viscosity, and 𝜎 the surface δ₄₀ 3.89222704E+02 δ₈₀ 3.04800171E+05
tension. δ₄₁ -5.43738808E+04 - -
IV. MODELLING RESULTS AND DISCUSSION
C. New model validation
B. Model optimization
The Model verification of bubble pump for is based in the
With the aim to optimize the analytical bubble pump model use of water in order to discuss the Dependence of mass flow
developed by Delano, based on parameters Regression, a new rate on submergence ratio.
correlation of the loss factor parameter K, eq.(34), is proposed The submergence ratio is one of the influencing parameter
based on the experimental investigation of the thermally on the bubble pump performance. For the same tube
operated vapor lift bubble pump carried out by Abduwadood diameter, increasing the submergence ratio leads to
and et al using water as working fluid and that conducted by increase the liquid flow rate. Therefore, the bubble pump is
sathe using methanol. The correlation is found out using more efficient when the submergence ratio increases.
TableCurve 2D by applying the experimental results to the
analytical model for a given tube diameter and a given
submergence ratio and inserted in the analytical model
equation (23). Next, the rectified model is developed using
FORTRAN program coding the ODRPACK algorithm. The
theoretical liquid mass flow is so calculated and then
compared with the experimental results.
The K is a function of (d/L) and (H/L) and it has the form:
B C
K = A + 0.5 + 1.5 (34)
Reg Reg
Where
H H 2
A = β1 + β2 ( ) + β3 ( ) (35)
L L
H H 2
B = β4 + β5 ( ) + β6 ( ) (36)
L L
H
C = β7 + β8 ( ) (37)
L
Where
d
β1 = δ10 + δ11 ( ) (38)
L
d
β2 = δ20 + δ21 ( ) (39) Fig. 6 Mass flow rate of the pumped liquid as a function of the driving power
L for the Ø 8mm tube and different submersion rates. Symbol: experimental data,
d 2
β3 = δ30 + δ31 / ( ) (40) Line: theoretical model corrected
L
d Figs. 6, 7 and 8 illustrate the comparison between the
β4 = δ40 + δ41 ( ) (41)
L improved analytical model and the experimental result of the
d
β5 = δ40 + δ51 ( ) (42) water based bubble pump performance when varying the
L
d 2 submergence ratio for fixed tube diameter. One can see that
β6 = δ60 + δ61 / ( ) (43) the predictions show the same behavior as the experimental
L
d results. In fact, the increasing of the heat input leads to a
β7 = δ70 + δ71 ( ) (44)
L
higher liquid flow rate until a maximum value is reached.
β8 = δ80 (45)
Heating the fluid beyond this limit will simply generates more
and
vapor and the produced high vapor velocity will change the
flow pattern from slug to annular flow which causes the 0.4 and 0.6), hence it can be suited to describe bubble pumps
expected decrease in pumping capacity. behavior.
Fig. 9 Comparison of the present model with Abduwadood 's model and
Fig. 7 Mass flow rate of the pumped liquid as a function of the driving power
experiments
for the Ø 10mm tube and different submersion rates. Symbol: experimental
data, Line: theoretical model corrected
V. CONCLUSIONS
Bubble pump is a crucial component of the Platen-Munter
[1] and Einstein-Szilard [22] single pressure absorption
refrigeration cycles. First, an experimental work has been
conducted to investigate bubble pump for different tube
diameters and submergence ratio. Second, as bubble pumps
have the same fundamental working principle as air-lift pumps,
a new theoretical model based on Delano work is presented to
estimate the bubble pump mass flow rate. In this model a new
correlation of the loss factor parameter K is proposed.
Compared to the experimental and theoretical investigated
bubble pump [20] for different tube diameters and
submergence ratio, the new developed model show an
excellent estimation of the bubble pump performances.
REFERENCES
[1] Von Platen BC, Munters CG. US Patent 1, 685,764; 1928.
Fig. 8 Mass flow rate of the pumped liquid as a function of the driving power [2] Jakob U, Eicker U, Schneider D, Taki AH, Cook MJ. Simulation and
for the Ø 10mm tube and different submersion rates. Symbol: experimental experimental investigation into diffusion absorption cooling machines
data, Line: theoretical model corrected for air conditioning applications. Applied Thermal Engineering 2007.
[3] Zohar A, Jelinek M, Levy A, Brode I. Numerical investigation of a
diffusion–absorption refrigeration cycle. International Journal
According to Fig. 6, the theoretical model shows a little of Refrigeration 2005;28(4):515–25.
[4] Koyfman A, Jelinek M, Levy A, Borde I. An experimental investigation
under-prediction of pumping capacity when decreasing the
of bubble pump performance for diffusion absorption refrigeration
bubble pump driving heat under 175 W. Nevertheless, one can system with organic working fluids. Applied Thermal Engineering
observe (cf. Figs. 7 and 8) that this model proves a very good 2003;23:1881–94.
agreement with the experimental points for different diameter [5] Pfaff M, Saravanan R, Maiya MP, Srinivasa M (1998) Studies on
bubble pump for a water–lithium bromide vapor absorption
and submergence ratio. So we can conclude that the model
refrigeration. Int J Refrig 21:452–462
describes very well the pump performance. [6] White SJ 2001 Bubble pump design and performance. M.Sc Thesis,
Georgia Institute of Technology
D. Model comparison [7] Koyfman A, Jelinek M, Levy A, Borde I (2003) An experimental
Referring to Fig.9, one can realize that the predicted investigation of bubble pump performance for diffusion absorption
refrigeration system with organic working fluids. App Ther Eng
maximum bubble pump capacity in the present work show a 23:1881–1894
great similarity whit the experimental data [19]. When [8] Vicatos G, Bennett A (2007) Multiple lift tube pumps boost
compared to abduwadood work, one can remark that refrigeration capacityin absorption plants. J Energy S Afr
experiments are better described by the present model (H/L = 18:49–57
[9] Siyoung Jeong, Sang-Kyun Lee, Kee-Kahb Koo: Pumping
characteristics of a thermosyphon applied for absorption refrigerators
Mass balance
LiBr-Mass balance
Energy balance
Fig. 3. Trans-criticalCO2 sub-cycle of configuration A in (𝑃 − ℎ)- Where 𝑚̇𝑖𝑛 and 𝑚̇𝑜𝑢𝑡 are the mass flow rates of inlet and
diagram. outlet streams. 𝑋 stands for salt mass fraction in the
LiBr/water solution,𝑄̇ is the heat transfer rate from/to the unit
component and 𝑊̇ , the work transfer rate to/from that
component, and ℎ , the specific enthalpy of refrigerant or
solution mixture.
Fig. 4. CO2 sub-cycle of configuration B in (𝑃 − ℎ)-diagram. 𝑄̇𝐸𝑉_𝑐𝑜𝑚𝑝 = 𝑚̇11 (ℎ16 − ℎ11 ) (10)
𝑄̇𝐶𝐷−𝑐𝑜𝑚𝑝 = 𝑚̇11 (ℎ41 − ℎ15 ) (11)
III. MATHEMATICAL MODEL 𝑄̇𝐶𝑜𝑜𝑙𝑒𝑟 = 𝑚̇11 (ℎ13 − ℎ14 ) (12)
During developing of mathematical models of 𝑊̇𝑐𝑜𝑚𝑝 = 𝑚̇11 (ℎ12 − ℎ11 ) (13)
configurations A et B following assumptions were made:
Analyses are made under steady state conditions. Model of absorption subunit of configuration B
Heat losses to the surroundings are assumed negligible
at generator, condenser absorber and evaporator. 𝑄̇𝐸𝑉−𝑎𝑏𝑠 = 𝑚̇9 (ℎ9 − ℎ10 ) (14)
Pressure losses in pipes and components are negligible. 𝑄̇𝐶𝐷_𝑎𝑏𝑠 = 𝑚̇9 (ℎ8 − ℎ7 ) (15)
Flows exiting condenser (liquid), evaporator (vapor), 𝑄̇𝐴𝐵 = (𝑚̇10 ℎ10 + 𝑚̇6 ℎ6 ) − 𝑚̇1 ℎ1 (16)
absorber (liquid)or generator (liquid & vapor) are 𝑄̇𝐺𝐸𝑁 = (𝑚̇4 ℎ4 + 𝑚̇7 ℎ7 ) − 𝑚̇3 ℎ3 (17)
saturated. 𝑄̇𝐸𝑉−𝑎𝑏𝑠 = 𝑄̇𝐶𝐷_𝑐𝑜𝑚𝑝 (18)
Solution and refrigerant valves are isenthalpic.
Evaporator energy of absorption cycle is equal to the Model of vapor-compression subunit of configurationB
condenser energy of compression cycle.
Effectiveness of heat exchangers is set to𝜖𝐻𝑋 = 0.7. 𝑄̇𝐸𝑉_𝑐𝑜𝑚𝑝 = 𝑚̇11 (ℎ14 − ℎ12 ) (19)
The isentropic efficiency of pumps is 𝜂𝑝𝑢𝑚𝑝𝑠 = 100%. 𝑄̇𝐶𝐷_𝑐𝑜𝑚𝑝 = 𝑚̇11 (ℎ11 − ℎ13 ) (20)
𝑊̇𝑐𝑜𝑚𝑝 = 𝑚̇11 (ℎ11 − ℎ12 ) (21)
Page 19
Copyright - 2019
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
The coefficient of performance (𝐶𝑂𝑃) a subunit and of the Table I summarizes the assumed operating conditions for
whole configuration is given by the ratio of the refrigeration configurations A et B.
load and the energy supplied to the cycle, i. e.
IV. RESULTS AND DISCUSSION
𝐶𝑂𝑃of an absorption subunit Simulation results for case A are shown in Tables II and III. It
should be mentioned that in order to accommodate the
𝑄̇𝐸𝑉−𝐴𝐵𝑆 enthalpy of the hot CO2 exiting the compressor to the energy
𝐶𝑂𝑃𝐴𝐵𝑆 = (22) needs of the absorption subunit, the pressure ratio must
𝑄̇𝐺𝐸𝑁 + 𝑊̇𝑝
exceeds 8. Otherwise, the energy of hot gas at (12) wouldn’t
be large enough to activate the generator of absorption sub-
𝐶𝑂𝑃of a vapor compression subunit cycle. In the present simulations, we set in a first approach the
high pressure at 12 to 100 bar instead of at the minimum of 96
𝑄̇𝐸𝑉−𝑐𝑜𝑚𝑝 bar. An optimization procedure would reveal the appropriate
𝐶𝑂𝑃𝐶𝑂𝑀𝑃 = (23) compression ratio to maximize the coefficient of
𝑊̇𝑐𝑜𝑚𝑝
performance,𝐶𝑂𝑃𝐴 . It must be noticed also that a better heat
integration scheme, e.g. using part of the rejected heat in the
𝐶𝑂𝑃of configuration A
cooler to preheat the salt solution coming from the absorber
and flowing to the generator, would enhance the 𝐶𝑂𝑃 of this
𝑄̇𝐸𝑉−𝑐𝑜𝑚𝑝 configuration.
𝐶𝑂𝑃𝐴 = (25)
𝑊̇𝑝 + 𝑊̇𝑐𝑜𝑚𝑝 It is interesting to notice that a refrigeration unit constituted
solely of the trans-critical system would achieve the same
𝐶𝑂𝑃 of configuration B COP, namely 1.35. But in that case, an exergy analysis would
reveal a major destruction of exergy due heat rejected to
𝑄̇𝐸𝑉−𝑐𝑜𝑚𝑝 environment represented by the process between points 12 and
𝐶𝑂𝑃𝐵 = (24) 13 in Fig. 3 (temperature decrease from 107°C to 15°C).
𝑄̇𝐺𝐸𝑁 + 𝑊̇𝑝 + 𝑊̇𝑐𝑜𝑚𝑝
Further, a cold source would be necessary to cool CO2 gas
from point 14 (35°C) to 15 (15°C) and that would again
For the numerical simulations, computer codes have been necessitate a second refrigeration cycle which then would
realized using the software Engineering Equation Solver, EES, reduce the 𝐶𝑂𝑃 of the combined system. To avoid the cooling
[29]. The refrigeration capacity of each configuration is set to between 14 and 15, one may lead directly flow 14 into the
100 kW for a low temperature evaporator of −35°C.The expansion valve. In this case however, the enthalpy change per
evaporation of the refrigerant water in the absorption kg CO2 associated to the evaporation is reduced and, by fixed
subsystem is assumed to take place at 10°C. The temperature cooling capacity, the mass flowrate of the refrigerant has to be
of CO2in that element (sub-cooler or condenser) is supposed to increased, reducing thus the thermal and exergetic
be 5°C higher, i.e., at 15°C. performance of the unit.
TABLE I TABLE II
OPERATING CONDITIONS FOR SIMULATIONS CASCADE A: STATE POINT RESULTS.
Parameters
ABSORPTION SUB-CYCLE Point
𝑃[kPa] 𝑇 [°C] 𝑚̇[kg/s] 𝑋[%]
Input variable Unit Value 1 1.2 35 0.067 61.64
Generator temperature 2 5.6 49 0.067 61.64
°C 83
(Configuration B) 3 5.6 60 0.079 52.19
Evaporator temperature °C 10 4 5.6 83 0.067 61.64
Condensation temperature °C 35 5 5.6 35 0.079 52.19
Absorber temperature °C 35 6 1.2 35 0.079 52.19
Solution heat exchanger efficiency, 𝜖𝐻𝑋 0.7 7 5.6 83 0.012 0
COMPRESSION SUB-CYCLE 8 5.6 35 0.012 0
9 1.2 10 0.012 0
Input variable Unit Value
10 1.2 10 0.012 0
Capacity kW 100 11 1202 -35 0.485 -
Cooler temperature °C 𝑇𝐶𝐷−𝑎𝑏𝑠 12 10000 164 0.485 -
Evaporator temperature °C -35 13 10000 107 0.485 -
Condensation temperature/Sub-cooler 14 10000 35 0.485 -
°C 15
outlet temperature
15 10000 15 0.485 -
Heat exchanger efficiency, 𝜖𝐻𝑋 0.7 16 1202 -35 0.485 -
Isentropic efficiency of compressor 0.68
Ikram.saafi@gmail.com
hmed.taieb@gmail.com
3
yosraisetkelibia@gmail.com
mejbri_khalifa@yahoo.fr
a.bellagi@enim.rnu.tn
*
Institut Supérieure des études technologiques de Kelibia, Route Oued el Khatf 8090
Abstract— This document proposes a simulation model of an Chen et al. [3], designed a new generator including a heat-
absorption-diffusion refrigerator using ammonia/water/ exchanger that reuses the waste heat. The new configuration
hydrogen as working fluids, developed to describe and of the cycle showed a slightly improvement of the COP (5%).
predict the behavior of the device under different operating Srikhirin et al. [4] carried out an experimental study on an
conditions. The system is supposed to be cooled in the NH3–H2O DAR cycle using helium as the auxiliary gas.
ambient air and operated with a free thermal source. Based
They developed a mathematical model to determine the
on the results obtained, a thermodynamic analysis of the
system is carried out in order to evaluate the contribution maximum performance for the appropriate operating
of the various components of the machine to the loss of conditions. They observed that the evaporator and absorber
machine performance. In addition, thermals conductances mass transfer performances have a strong effect on the system
of the internal and external heat flows that take place in the performance [5]. They also showed that the COP of the
machine are evaluated. The model is then modified by machine varies in the range 0.09 - 0.15.
replacing the thermal pinches of the different heat Maiya [6]; showed that helium is more advantageous than
exchangers with thermals conductances and used to hydrogen as inert gas although it requires a larger size of
perform a parametric study of the cycle. propulsion because of its greater viscosity. This study proved
also that a higher pressure of operation causes a reduction in
Keywords— Diffusion-absorption refrigerator. Model and the COP.
simulation. Water-ammonia. Zohar et al. [7] studied two configurations of a DAR with
and without condensate sub-cooling before the entry of the
I. INTRODUCTION evaporator. The results showed that the COP of the cycle
The Diffusion Absorption Refrigerator (DAR) introduced without subcooling is higher by approximately 14-20% than
by Platen and Munters in 1928 [1] has been recognized as one that of the cycle with subcooling. The best performances are
of the most encouraging sustainable technologies for obtained when the mass fraction of ammonia of the rich
production of cold. The cycle of the machine operates at a solution varies in the range (0.25-0.4).
constant total pressure level and uses ammonia as refrigerant, Ben Ezzine et al. [9] reported that the R124–DMAC
water as absorbent and hydrogen or helium as non-absorbable DAR gives a higher COP at lower driving temperatures
auxiliary inert gas. This inert gas is necessary to reduce the compared with the NH3–H2O system; they also
partial pressure of the refrigerant in the evaporator and the experimentally investigated a DAR using C4H10–C9H20
absorber and to allow the processes of evaporation and as the working fluid and helium as the auxiliary gas [10].
absorption to take place in the uniform pressure device. In the present paper a thermodynamic model for an
DAR operates only with thermal energy, no mechanical ammonia-water diffusion absorption refrigeration, with
and then no electric power is needed. This energy can be hydrogen as inert gas is developed. The performances of
provided by the fossil fuel combustion (gas, fuel, etc), but cycle are theoretically analyzed and evaluated. The effects of
also, for temperatures varying between 90 and 200°C, by the characteristics of bubble pump with various power inputs,
solar thermics, or the thermal discharges, etc. The growing heating temperatures and other parameters are investigated. In
concerns about worldwide energy and environmental addition, an exergy analysis is carried out to evaluate the
sustainability in recent years enlarge the development of percentages of irreversibilities created in the different parts of
DAR [2]. the machine
The DAR systems have been experimentally and
theoretically investigated by various researchers. II. DESCRIPTION OF THE CYCLE
The figure 1 shows a commercial low-power absorption-
diffusion machine that operates according to the diffusion-
Copyright - 2019
Page 23
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
absorption cycle. This cycle uses the refrigerant/absorbent: The rectifier is necessary to purify the ammonia vapor to
ammonia/water in combination with hydrogen as an inert gas. the desired composition y1. The corresponding
The absorption-diffusion cycle consists mainly of a temperature T1 is deduced from the liquid-vapor
generator, a rectifier, a condenser, an evaporator, a gas equilibrium. The flow rate of refrigerant vapor is deduced
exchanger, an absorber, a reservoir and a solution exchanger. from the molar flow rate of the gases flowing in the
gas circuit . The energy balance makes it possible to
determine the heat released to the rectifier.
B. CONDENSER
The condenser, shown in figure6, is cooled to
ambient air. It is characterized by a deviation of the
condensation temperature and air temperature: TC-air.
The liquid refrigerant at the outlet of the condenser is
assumed to be saturated. The mass, energy and entropy
balances are given by the following equations,
C. EXPANSION CHAMBER
The liquid / vapor equilibrium is estimated by
III. THERMODYMAMIC PROPERTIES
considering the equalities of the chemical potentials and
The properties of the ammonia / water binary mixture the energy balance. The molar flow rates of liquid and
(refrigerant / absorbent) are described by the PC-SAFT state vapor as well as the molar composition of the inert gas at
equation. The model obtained is in agreement with the the outlet are determined from the material balances.
constraints of the equality of the chemical potential in the Entropy creation is computed using the entropy balance.
liquid-vapor equilibrium. The inert gas is supposed to be
immiscible with the liquid phase and exists only in the gas
phase. This is considered an ideal solution of real gases. The
properties of hydrogen are taken from the NIST databank [11]
and have been put into an empirical model using the
Younglove state equation [12].
equivalent to a mass flow rate of 0.196 g/s. In the SHX,the continues its subcooling in the central tube of the evaporator
rich solution is preheated by the lean solution at a temperature to T10 = -17.4 °C.The gas flow at the inlet of the expansion
T12 = 126 °C, close to its saturation temperature equal to chamber is composed of almost pure hydrogen ( =
129.7 ° C. In the heating zone in the lower part of the bubble 0.99985). The circulation rate of the inert gas and the
pump, the power provided allows the evaporation of 8.6% refrigerant in the gas loop is r = 5.05.
of the liquid solution, the vapor produced for pumping the It should be noted that a significant amount of cold is
rest of the liquid. The vacuum ratio corresponds to a two- dissipated during the purification and subcooling of the inert
phase flow in the bubble pump tube is estimated with the gas in the GHX and evaporator (15.8W). On the other side,
CISE correlation [13] at about 0.283; this value corresponds and under the specified operating conditions, the degassing of
to a Slug flow. At the top of the bubble pump, the saturated the refrigerant in the generator is small, and therefore the
liquid and the vapor phase at temperature T 13 = 140.8 °C are flow rate of the rich solution is large compared to that of the
separated; the liquid is further heated in the boiler at a refrigerant (4.3 times). This means that the amount of heat
temperature T14 = 189.7 °C. The steam, with an average supplied to the generator is important in relation to the
ammonia molar composition of 0.634, is generated at 78.7% cooling power. For all these reasons, the COP of an
in the boiler and 21.3% in the bubble pump. In the rectifier, absorption-diffusion refrigerator is normally low: In this case,
the refrigerant vapor flow rate is = 0.00257mol/s (0.0438 it is equal to 0.126.
g/s) is purified to a molar ammonia composition of y1 = 0.99.
The liquid rich in water (x17 = 0.156) returns to the boiler TABLE 2: Basic simulation results.
with a flow rate =0.0019 mol/s. The power = 30.34 W Variable Value Variable Value
is rejected from the rectifier to the outside environment. The 20.7 0.012872
(bar) (mol/s)
condensing power released to the ambient air is = -21W.
The liquid refrigerant is at the temperature T 4 = 25 °C and (°C) 81.8 (mol/s) 0.011096
injected into the expansion chamber, where 34.3% of the (°C) 25.0 0.008522
(mol/s)
liquid is evaporated by reducing its hydrogen composition to
(°C) 27.4 (mol/s) 0.001900
= 0.9357. The two-phase flow is now at the very low
temperature T5 = 27.4 ° C. The expansion chamber is (°C) 5.0 0.086
equivalent to an isenthalpic expansion in a vapor compression
refrigerator. The liquid refrigerant continues its evaporation (°C) 34.0 0.283
in the annular space of the evaporator producing the useful 46.0 0.33427
(°C)
cold = 8.42W. The temperature at the outlet (6) of the
evaporator reaches T6 = 5 °C, 9.7% of refrigerant is still in the (°C) 23.6 0.13619
liquid state. In the annulus of the GHX, all the liquid (°C) 17.4 0.93570
refrigerant is evaporated and the superheated gas at T 7 =
34 °C. The absorber is equivalent to 4 theoretical stages. The (°C) 126.0 0.83456
ammonia-rich gas at the bottom of the absorber is
(°C) 140.8 0.03169
characterized by a molar composition of hydrogen =
0.8346 and a molar flow = 0.01545 mol / s. At its top is (°C) 189.7 0.02228
injected the liquid poor solution with a molar composition x15 85.2 0.94603
(°C)
= 0.1362 and at a temperature T 15 = 85.2 ° C and with a flow
rate = 0.00852mol / s. Due to the absorption process, a 5.05 0.99659
QE(W)
COP
condenser from the condenser is proportional, this latter 6,5
flow rate is equal to 5 times the first. 0,06
6,0
In figure 15, the COP is calculated under the basic 0,04 5,5
conditions, but by varying the motive power supplied to the 5,0
0,02
generator (boiler and bubble pump), . It is found that the 4,5
0,00 4,0
variation of the COP with has a maximum, but the 30 35 40 45 50 55 60 65 70
0,15
n1(mol/min)
0,80
0,35
0,14
0,75 0,30
xriche
0,13
xpauvre
xriche, xpauvre
0,70 0,25
0,12
0,20
0,65
0,11
35 40 45 50 55 60 65 70 0,15
Qm(W)
0,10
Figure 2: Variation of the molar flow rate of the refrigerant and the 156 159 162 165 168 171 174 177 180 183
stabilizes. 0,10
0,08
that the growth of the driving temperature has an effect on the 0,06
the absorber. On the other hand, it can also be noted that the 0,02
produces more ammonia vapor. However, above 190 °C, the 160 165 170 175 180 185 190 195 200 205 210 215
0,09
REFERENCES
16 18 20 22 24 26 28 30 32 34 36 38 40 [1] B.C. Von Platen, C.G. Munters, Refrigerator, US Patent 1 (1928)
Tamb(°C) 685-764.
[2] H.C. Steirlin, J.R. Ferguson, Diffusion absorption heat pump
[3] S. Mazzouz, Fonctionnement des cycles à mono-pression pour le
Figure 6: Variation of the experimental and theoretical COP as a froid solaire : étude théorique et expérimentale d'un pilote, PhD Thesis,
function of the ambient temperature. ENIM-University of Monastir, Tunisia, 2014.
[4] R. Garma, Machines frigorifiques à absorption à monopression :
Figure 6 shows the effect of the driving temperature on the Etude de la pompe à bulle, PhD Thesis, ENIM-University of Monastir,
COP. As can be noted, the minimum motive power found by Tunisia, 2015.
the simulation is lower than that measured experimentally. [5] R. ben Jemaa, PhD Thesis, Cycle de Platen et Munters : étude
expérimentale et théorique d’un pilote et analyse en dynamique de
Indeed, the generator is not ideally isolated and therefore the l’installation, ENIM-University of Monastir, Tunisia, 2014.
power supplied is actually acquired by the rich solution and is [6] J. Chen, KJ. Kim, KE. Herold. Performance enhancement of a
smaller than that measured experimentally. The maximum diffusion–absorption refrigerator. International Journal of
value of the theoretical COP is reached for a lower motive Refrigeration 1996; 19:208–18.
[7] P. Srikhirin, S. Aphornratana, Investigation of a diffusion
power. This maximum is 0.127 for a driving temperature of absorption refrigerator, Applied Thermal Engineering 22 (2002)
189 ° C. 1181–1193.
The various experiments carried out for different values of [8] L. S. Wilson, A. C. Carlos, A. T. Ednildo, C. S. Jacques, C. M.
the ambient temperature made it possible to compare the Evania, Exergoeconomic analysis of a single pressure absorption
refrigeration cycle, 20th International Congress of Mechanical
performances of the refrigerator found by the simulation and Engineering (2009)
those deduced from the experimental results. Figure 5 shows [9] A. Zohar, M. Jelinek, A. Levy, I. Borde, The influence of
the variation of the experimental and theoretical COP as a diffusion absorption refrigeration cycle configuration on the
function of the ambient temperature. Note that the two results performance, Applied Thermal Engineering 27 (2007) 2213–2219.
[10] A. Zohar, M. Jelinek, A. Levy and I. Borde, ‘Performance of
are in good agreement since the shape of the curves is Diffusion Absorption Refrigeration Cycle With Organic Working
identical in the common range of temperatures. As can be Fluids’, International Journal of Refrigeration, Vol. 32, N°6, pp. 1241
noted, the COP predicted by the model decreases – 1246, 2009.
continuously from 0.14 to 0.09 for ambient temperatures [11] N. Ben Ezzine, R. Garma, A. Bellagi, A numerical
investigation of a diffusion–absorption refrigeration cycle based on
between 15 and 40 °C. Indeed, this decrease is mainly due to R124–DMAC mixture for solar cooling, Energy 35 (2010) 1874–
the increase of the total pressure of the system when the 1883.
ambient temperature increases which makes the evaporation [12] N. Ben Ezzine, R. Garma, M. Bourouis, A. Bellagi,
of the ammonia vapor more difficult. Experimental studies on bubble pump operated diffusion
absorption machine based on light hydrocarbons for solar cooling,
Renewable Energy 35 (2010) 464–470.
[13] A. Benhmidene, B. Chaouachi, M. Bourouis, S. Gabsi, Effect of
VI. CONCLUSIONS operationg conditions on the performance of the bubble pump of
absorption-diffusion refrigeration cycles, Thermal science Vol. 15,
In this paper, we have carried out a theoretical simulation of N°3, pp. 793 – 806, 2011.
the absorption-diffusion machine. For this purpose, an [14] R. Garma, Y. Stiriba, M. Bourouis, A. Bellagi, Numerical
investigations of the heating distribution effect on the boiling flow in
advanced model of thermodynamic simulation is developed bubble pumps, International Journal of Hydrogen Energy, Vol. 39
to predict the performance and the operating conditions of the (2014) 15256-15260.
machine. This model takes into account the flows in the [15] B.A. Younglove, "Thermophysical Properties of Fluids. I. Argon,
bubble pump and the absorber-evaporator gas circuit. The Ethylene, Parahydrogen, Nitrogen, Nitrogen Trifluoride, and Oxygen,"
J. Phys. Chem. Ref. Data, Vol. 11, Suppl. 1, pp. 1-11, 1982
laws of conservation of mass and energy are established for
each organ.
2ahmed_isg@yahoo.fr
and the estimation of hedging effectiveness, but they forgot paper can be considered as the first paper that aims to fulfill
the effect of the tradeoff between transaction costs and this gap in the literature and analyzing, therefore, its value
effectiveness hedging on the portfolio decision. The main and implication on renewable hedging strategies [7, 12].
novelty of our paper is to examine firstly the time varying The reminder of this paper is laid as follow: Section 2
optimal hedging ratios from rolling window analysis among reviews briefly the previous empirical studies. Section 3
three MGARCH models (DCC, ADCC and GO-GARCH). provides data descriptions. Section 4 describes estimation
Then we investigate the impact of the tradeoff between methodology. Section 5 details the empirical results. Finally,
transaction costs and hedging effectiveness measured by the concluding remarks, implications and future research
ratio (TC/HE) on portfolio hedging decisions. The current opportunities are presented in the last section.
on clean energy stock prices. -After the outbreak of the 2008–2009 global
financial crises, they found risk-adjusted
returns.
- Detection of speculative bubbles, presented in
Germany’s renewable energy stocks, by the
ADF test.
Ortas and Moneva Measure the financial behavior -Clean techs indices yielded higher risk levels
[19] of 21 primary clean-technology State-space From 2002 to during market stability's period, resulting a clear
equity indices. market model 2011 and positive interaction between financial and
environmental performance.
Wen et al. [20] Documented return and -Results indicate that fossil fuel and alternative
volatility spillover effects An asymmetric From 2006 to energy stocks are considered as competing
between Chinese renewable Baba–Engle– 2012 assets.
energy stock pricesand fossil Kraft–Kroner -Investments in renewable energy are riskier
fuel companies. (BEKK) model than fossil fuels investments.
Inchauspe et al Examining the impact of oil State-space multi- -There is a positive connection between clean
[21] prices, technology stocks and factor model with From 2001 to energy and oil prices as well as a high
the MSCI World Stock Index on time-varying 2014 correlation with MSCI World Index and
renewable energy stocks. coefficients technology stock returns
Reboredo [22] Investigating systematic risk -Reported evidence of a significant association
and dependence structure Time-varying From 2005 to between oil and renewable energy stock prices.
between oil prices and various Copulas 2013 - Oil price dynamics contribute to nearly 30%
alternative energy sector equity and the CQ of clean energy stock price risk.
indexes. approach
Sanchez [10] -Calculating -Alternative energy hedge ratios vary
in-sample optimal hedge ratios Multivariate From 2002 to considerably over the sample period.
-Investigating volatility GARCH models 2015 -Volatility spillovers founding between clean
spillovers between oil prices and energy and technology stock prices are stronger
stock prices of alternative than those between renewable energy and oil
energy and technology. prices.
-The best hedge ratio for alternative energy is
providing by technology global markets.
Bondia et al. [23] Exploring the long-term -They find a significant short-run linkage
dependence structure between VECM From 2003 to between stock prices of alternative energy,
clean energy and technology (Vector Error 2015 technology companies, crude oil and US
stock prices, the returns of Correction Model) interest rate, while, in the long-run there is no
global oil prices and US interest significant relationship.
rate.
Analyzing dynamic correlation -Finding a weak short-term linkage between oil
Reboredo et al and causality in an alternative Wavelet approach From 2006 to prices and renewable energy stock prices, but in
[24] time-frequency setting between 2015 the long run the interaction is getting stronger.
international oil prices and new
energy stocks prices.
Testing the dynamic -There is a high interdependence structure
Ahmad [25] interdependence and investment The directional From 2005 to moving from technology to alternative energy.
performance between clean spillover approach 2015 -However, crude oil displays a restricted
energy, oil and technology stock and the Dynamic association renewable energy stocks and
prices. Conditional technology firms.
Correlation - Crude oil, when combined with clean energy
Models and technology indices, provides better
profitable hedge and portfolio investment
diversification.
Dutta A. [26] Investigating the impact of oil Employing three -Oil market uncertainty, emerged as a positive,
price uncertainty, as measured different range- From 2007 to statically and highly significant variable for
by the crude oil volatility index based estimators 2016 modeling, forecasting and predicting the
(OVX) on the variance of clean proposed by realized volatility of renewable energy stock
energy stocks. Parkinson (1980) returns, especially during the subprime crisis.
(henceforth RVP),
Rogers and
Satchell (1991)
(Hence forth
RVRS) and
Alizadeh et al.
(2002) (henceforth
RVABD).
Asia and Oceania and 27.1% for Europe, the Middle East and
Africa G. VVIX: Volatility VIX Index
B. CDS: Credit Default Swap Constructed at the aggregate market and represented by
Chicago Board Options Exchange (CBOE), we can look at
The Credit Default Swap (CDS) is a credit derivative contract the VVIX as the volatility of volatility index calculated from
between two counterparties which bring protection against a portfolio of VIX options (VVIX portfolio) through the same
credit losses. More precisely, the developed credit default algorithm used to measure the VIX.
swap (CDS) market allows CDS buyers to transfer Credit risk Moreover, the VIX index can be viewed as an important
to CDS sellers. indicator of market expectations regarding the future
More importantly, a CDS can also act as a hedge. distribution of the implied volatility.
C. Crude Oil H. Euro Stoxx 50: The European stock market index
Oil, conventional fossil fuel energy, is the most heavily Euro_Stoxx_50 index used for the Euro Area was introduced
traded physical commodity in the world. In this paper, oil on February 1998.
price returns (dollars per barrel) are measured by using the This index considered as Europe’s leading Blue-Chip index
average of the closing prices on the West Texas Intermediate specialist aims to provide a blue-Chip representation of super
(WTI) nearest Crude Oil futures contract which exchanges on sector leaders in the Eurozone.
the New York Mercantile Exchange (NYMEX). In principal, the Euro_Stoxx_50 is a composite index
represents the performance of the 50 most important
D. OVX: Crude Oil Volatility companies of up to 11 Eurozone countries (20 companies
from France, 14 from Germany, 5 each from Spain and
In 2008, Chicago Board Option Exchange (CBOE) Netherlands, 3 from Italy, and the remaining 3 are
introduced OVX as a new barometer to examine the respectively from Belgium, Finland and Ireland).
systematic behaviour of crude oil market uncertainty. As the It is the one of the most liquid European equity indices and
VIX, the idea of OVX is to measure the market’s expectation the most followed in the Eurozone.
of 30-day volatility of crude oil futures prices.
For Gold, options data are treated on Chicago Mercantile VSTOXX is a Measure of the implied volatility of Euro
Exchange 100 ounces Continuous futures contracts settlement Stoxx 50 in the Euro Area Market. Additionally, according to
price. Many previous studies Tully and Lucey [29]; Shahzad Zghal, R. et al. [31], the VSTOXX index helps to capture the
et al. [30] have shown that gold has been usually used as an equity risk as a whole, since it relies heavily on equity-based
efficient asset to store value and still treated as a significant options.
valuable metal in modern economies.
IV. METHODOLOGY: EMPIRICAL MODELS
F. VIX: Implied volatility of S&P500 on US Stock index
Recently, modelling the volatility dynamics and correlations
Introduced by the Chicago Board Options Exchange (CBOE) are highly relevant in finance.
in 1993, the VIX is used as a RISK-Neutral forward measure In this context, two models belonging to the DCC family
of the US stock market volatility. (DCC model of Engle [32] and ADCC model of Cappiello et
As such, the VIX is compiled from a portfolio of S&P500 al. [33] ) as well as the GO-GARCH model of Van der Weide
index options in order to measure the implied aggregate [34] have been applied for the purpose of modelling
volatility in options markets of the S&P500 index during the volatilities, conditional correlations and hedge ratios between
next 30-Calender day period and is commonly used as a NEX, CDS, Crude Oil, GOLD, Bond, Euro_Stoxx_50,
proxy. Based on previous findings, Higher values of the VIX VSTOXX, VIX, VVIX and OVX.
index denote a much riskier stock market, whilst, lower Let be a n x 1 vector of series of returns. The specification
values showed a less risky market. On a worldwide scale, it is of the multivariate GARCH models, with AR (1) process for
one of the most recognized measures of volatility.
conditional on the information set , is defined as or have different impacts on conditional standard deviations
follows: and correlations. Thus, it is described as follow:
= (1 – α – β) + λ + α +β +λ
| = +
Where the coefficient λ indicate the asymmetric effect or
Where the vector of residuals can be modelled as: “leverage effect” in the model. It tends to explain the role of
bad news in increasing volatility than do good news during
= ; ↝ iid (0, )
downturn period.
Represents the n x n conditional covariance matrix of , and are the unconditional matrices of and
is a n x 1 i.i.d random vector of errors and denotes an n respectively.
x n identity matrix. The variable defined as Hadamard product of an indicator
function and residuals is formally given by = I[ < 0]
. The indicator function which is expressed by I[ < 0] is
equal to one if the standardized residuals is negative, and 0
1) The DCC-GARCH model
otherwise.
The Engle [32] Dynamic Conditional Correlation (DCC) Leverage effects tend to explain the role of bad news in
model, generalization of CCC model, follows two step increasing volatility than do good news during downturn
procedures. In the first step, the GARCH parameters are period. Besides, both models DCC and ADCC are estimated
estimated followed by correlations in the second step such as: by a maximum likelihood estimator.
=
3) The GO-GARCH model
Where = ( ,…., ) is a diagonal matrix that includes
varying standard deviations on the diagonal and which
composed as follow is the conditional correlation matrix: Under the GO-GARCH model, the residual is modeled
= diag( diag( as follows:
Where is a n x n symmetric positive definite matrix given =A
by:
Where denotes a set of invisible independent factors (
= (1 – – ) + +
= ( , ,…, )). A is an invertible and time-invariant n x
Denotes the n x n unconditional correlation matrix of the
n and can be decomposed into an unconditional covariance
standardized residuals = / and is its conditional
matrix Σ and an orthogonal matrix U.
variance-covariance matrix of the residuals .
A= U
The parameters and are non-negative scalar parameters
The matrix A is composed into rows which represent the
satisfying + < 1 which implies that > 0.
factor weights assigned to each time series and columns of
Under the DCC specification, the time-varying conditional
correlation series are described by: representing the factors . The factors of can be specified
= as :
Finally, the conditional covariance matrix of the returns 5) The tradeoff between transaction costs and hedging
effectiveness.
- is: According to Chen and Sutcliffe [38], we can measure the
=A transaction cost (TC) as the sum of the absolute changes in
the dynamic hedge ratios. Then we calculate the TC/HE ratio
4) The hedging effectiveness as a measure of the tradeoff between hedging effectiveness
and transaction cost. A low TC/HE ratio indicates a better
The hedging effectiveness (HE) index (e.g. Ku et al. hedging instrument.
[36] and Chang et al. [37]), described by the following
equation, is used to evaluate the hedging performance of
hedge ratio and optimal portfolio. V. RESULTS AND DISCUSSION
NEX CDS Crude_Oil GOLD BOND Euro_Stoxx_ VSTOXX VIX VVIX OVX
50
Mean -0.0328 0.0325 -0.0044 0.0148 0.0014 -0.0099 -0.0065 -0.0077 0.0095 -0.0039
Median 0.0234 0.0000 0.0000 0.0000 0.0000 0.0000 -0.2374 -0.2919 -0.2473 -0.1811
Std. dev. 1.4851 2.8456 2.1678 1.1629 0.3850 1.4594 6.5320 7.4383 5.0225 4.7163
Min. -10.485 -32.1330 -16.7095 -9.8233 -2.7373 -9.0111 -43.4376 -35.0588 -23.6414 -43.9905
Max. 12.070 25.3664 17.9691 8.5889 3.5661 10.4376 47.0666 76.8245 37.3161 42.4968
Q1 -0.6530 -0.8733 -1.0734 -0.5029 -0.2012 -0.6685 -3.8065 -3.9941 -2.7529 -2.5016
Q3 0.6736 0.8438 1.0852 0.5772 0.2105 0.6760 3.1236 3.2537 2.2389 2.1002
Skewness -0.4680 0.0962 0.1864 -0.3865 -0.1202 -0.0498 0.5613 1.0685 0.9459 0.6585
Kurtosis 8.5689 14.5229 5.5904 7.5714 5.6903 6.0269 4.3152 7.3226 5.3219 10.1832
JB test 8746 2483 3695.1 6818.3 3818.2 4276.9 2340.2 6849.3 3755.1 1241
P-Value 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000*
Q (12) 135.79 29.954 9.9302 24.145 25.453 29.994 29.818 47.211 41.731 80.644
P-Value 0.000* 0.002* 0.622 0.019* 0.012* 0.002* 0.002* 0.000* 0.000* 0.000*
Q2 (12) 3496.9 405.06 590.65 276.08 308.48 1293.4 363.25 200.83 182.47 337.21
P-Value 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.000*
N obs 2825 2825 2825 2825 2825 2825 2825 2825 2825 2825
N.B:* denotes 5% significance level. JB test indicates Jarque-Bera statistics and Q(12) and Q(12)² are the Ljung–Box statistics.
The ARCH-LM test reports the LM-statistic.
CDS exhibits the highest average daily returns among the opportunities of positive (negative) returns. Kurtosis statistics
series (0.032), while NEX have the lowest average return (- suggest that all variables have kurtosis greater than 4, and as
0.032), while the lowest average return (-0.032) is observed we know that kurtosis for a normal distribution is 3, so we
in case of NEX. The mean daily return is positive for CDS can deduct that all series display of fat or heavy tails in their
(0.032), GOLD (0.0148), Bond (0.0014), and for VVIX distributions (leptokurtic). Our observations are confirmed by
(0.0095), whereas it is negative for NEX (-0.032), Crude Oil graphs of the time series and squared returns (Figure 1 and 2
(-0.0044), Euro Stoxx 50 (-0.0099), VSTOXX (-0.0065), VIX respectively).
(-0.0077) and OVX (-0.0039). Figure 1reveals that there is some heterogeneity in price co-
VIX shows the greatest volatility designed by its high movements of each index. For example, during the subprime
standard deviation (7,438), while Bonds have the lowest crisis (2007 – 2009) and 2011 – 2014 periods, Crude Oil and
standard deviation (0.3850). The nullity of normal NEX, Gold, Bond, Euro Stoxx 50 tends to move together
distribution is decisively rejected by the Jarque-Bera (JB) test with a strong trend. Although, CDS, VSTOXX, VIX, VVIX
for each one of the variables at the 5% significance level. and OVX show a similar time series patterns but display of a
Besides maximum and minimum values indicate that the little increase trend around the 2007 – 2009 financial crisis.
volatility of all sample series was similar in magnitude, with Moreover, visual inspection of figure 1 also reveals that NEX
the exception of GOLD, Bond, and Euro Stoxx 50 which had and Crude Oil commove jointly during higher and lower
lower volatility. phases of the latter one.
For the Ljung-Box Q-statistics on returns, we find that only Time series plots of the squared time series shown in Figure
the Crude Oil doesn’t exhibit significantly high serial 2 exhibit how volatility has changed over time. We can
correlation, unless on squared returns, Q-statistics indicate observe that all variables present a strong volatility clustering
that all sample variables present significant serial correlation around the Subprime Mortage crisis with the exception of
and strong volatility clustering effects. The Skewness values CDS, VIX and VVIX which show a little clustering effect at
are negative for returns of NEX, Gold, Bond, and Euro Stoxx the same period.
50; however, they are positive for the other series. This means
that negative (positive) Skewness denotes luck of higher
negative (positive) returns without corresponding
Page 36
Copyright - 2019
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Now, we turn to examine the unconditional correlation Table 4, show a positive correlation between NEX and each
between raw returns and squared returns which are squared asset return, where the strongest correlation occurs
summarized in table 3 and 4 respectively. for NEX/Euro Stoxx 50 (0,5176) while the pairwise
From table 3, we find that NEX correlates positively with NEX/CDS have the weakest correlation (0,0091).
CDS, Crude Oil, GOLD and Euro Stoxx 50 and negatively Overall, Table 4 denotes positive correlation among the
with the other indexes. The Euro Stoxx 50 tend to have a squared returns. However, the degree of correlation varies
stronger correlation (0,7194) than the other variables, widely among both raw returns (between 0,7 and 0,09), and
followed by the VSTOXX (-0,5269). Whereas, the lowest squared returns (between 0,5 and 0,009).
correlation is found between NEX and CDS (0,0955).
Table 3: Unconditional correlation between raw returns
NEX CDS Crude Oil GOLD BOND Euro Stoxx 50 VSTOXX VIX VVIX OVX
NEX 1 0.0955 0.3831 0.1208 -0.2800 0.7194 -0.5269 -0.5255 -0.3879 -0.3181
CDS 0.0955 1 0.0509 -0.0272 -0.0266 0.0967 -0.0859 -0.0238 -0.0200 -0.0354
Crude Oil 0.3831 0.0509 1 0.1888 -0.2005 0.3710 -0.2592 -0.1831 -0.1261 -0.2430
GOLD 0.1208 -0.0272 0.1888 1 0.1368 -0.0178 0.0121 0.0075 0.0145 -0.0190
BOND -0.2800 -0.0266 -0.2005 0.1368 1 -0.3475 0.2627 0.2871 0.2274 0.1620
Euro Stoxx 50 0.7194 0.0967 0.3710 -0.0178 -0.3475 1 -0.7415 -0.4850 -0.3739 -0.2903
VSTOXX -0.5269 -0.0859 -0.2592 0.0121 0.2627 -0.7415 1 0.5368 0.4382 0.3146
VIX -0.5255 -0.0238 -0.1831 0.0075 0.2871 -0.4850 0.5368 1 0.8132 0.4305
VVIX -0.3879 -0.0200 -0.1261 0.0145 0.2274 -0.3739 0.4382 0.8132 1 0.3431
OVX -0.3181 -0.0354 -0.2430 -0.0190 0.1620 -0.2903 0.3146 0.4305 0.3431 1
NEX CDS Crude Oil GOLD BOND Euro Stoxx 50 VSTOXX VIX VVIX OVX
0.1011
NEX 1 0.0091 0.1468 0.0145 0.0784 0.5176 0.2777 0.2761 0.1505
0.0843
Euro Stoxx 50 0.5176 0.0093 0.1377 0.0003 0.1207 1 0.5499 0.2353 0.1398
1
OVX 0.1011 0.0012 0.0590 0.0003 0.0262 0.0843 0.0989 0.1853 0.1177
Following, the table 5 (See annex) presents the estimated lower than the long-run persistence (β) which is agreed with
results of the DCC and ADCC models. First of all, for all DCC and ADCC results. Moreover, “The DCC model is
time series, the short-term persistence (α) and the long-term mean reverting as long as α+β<1”, based on the expression
persistence (β) are statistically significant and for each case α above, we calculate the sum of the persistence parameters (α
is less than β, their sums are close to unity providing evidence and β), we found that is less than one, which proof the mean-
of volatility clustering in all variables which is proves in fig2. reverting of volatility process.
Secondly, as known, if the shape parameters (λ), the
equivalence of the degrees of freedom in the distribution, tend
to infinity, the t-distribution tends to the normal distribution.
In this case, crude oil shows the highest estimated shape (7,26)
followed by NEX (6,84) and Bond (6,22), while the lowest
shape parameter is found by CDS (3,12).
Thirdly, we find that the estimated coefficients on and
are each positive and statistically significant at the 1%
significance level and the sum of both parameters is less than
one meaning that the dynamic conditional correlations are
mean-reverting.
GO-GARCH parameters
F1 F2 F3 F4 F5 F6 F7 F8 F9 F10
0.1681 0.0046 0.0091 0.0109 0.0018 0.0940 0.0027 0.0051 0.1398 0.1643
Omega
0.1652 0.0743 0.0613 0.1394 0.0229 0.1183 0.0410 0.0321 0.1081 0.1111
Alpha
0.6582 0.9203 0.9298 0.8595 0.9745 0.7867 0.9563 0.9622 0.7383 0.7304
Beta
-0.1771 0.0181 0.0157 -0.0667 0.2102 -0.1953 -0.0194 0.0958 0.2476 0.2755
Skew
1.7192 2.7872 2.4955 0.4475 1.5198 1.0692 2.3265 0.9274 0.9762 1.1700
Shape
Turning to the correlations between correlations reported in NEX/OVX and NEX/VVIX. In either case, OVX present a
table 7, we find that dynamic conditional correlations realized negative association with NEX which means that our model
by DCC and ADCC models are similar. Furthermore, for exhibits a higher level of interdependence in this case.
each pair of correlations, the correlations between DCC/GO- Summary findings of correlations between hedge ratios
GARCH or ADCC/GO-GARCH are significantly less estimated from three MGARCH models (DCC, ADCC and
correlated. Otherwise, the lowest correlation between GO-GARCH) are presented in table 8, which suggest that
DCC/GO-GARCH and ADCC/GO-GARCH is found among hedge ratios obtained from DCC and ADCC models show a
NEX/Crude Oil followed by NEX/GOLD and NEX/CDS. perfect high correlation. Beyond this,
However, we note a higher correlation for NEX/Bond,
DCC/ADCC 0.9094 0.9640 0.9904 0.9784 0.7206 0.9539 0.9435 0.9245 0.9570
DCC/GOGARCH 0.6535 0.4161 0.4623 0.9056 0.8557 0.8457 0.7952 0.8447 0.8472
ADCC/GOGARCH 0.6201 0.4147 0.4187 0.8824 0.5320 0.7660 0.6997 0.6993 0.7864
DCC/ADCC 0.9240 0.9968 0.9986 0.9979 0.9825 0.9561 0.9802 0.9816 0.9905
DCC/GOGARCH 0.1812 0.4677 0.5385 0.6160 0.4736 0.0924 0.2038 0.3640 -0.4857
ADCC/GOGARCH 0.0662 0.4601 0.5296 0.6186 0.4653 0.1267 0.2141 0.3544 -0.4669
Summary statistics of hedge ratio and hedging shows the following values of HE: 0.3006, 0.3010 and 0.3015
effectiveness are reported in table 9, in order to examine the respectively for the refits 10, 20 and 60 days. Thus, it means
robustness of our findings with the change in the number of that all hedging results are robust to model refits.
model refits. Our results find that, for each pair, hedging According to the table 9, Refit=10, the average value of the
effectiveness values estimated with a student distribution are hedge ratio between NEX and VIX is 0,40 for th GO-
extremely similar beyond all three model refits and for each GARCH model, this means that a $1 long position in NEX
GARCH model specification. As example, hedging can be hedged for 40 cents with a short position in the VIX
effectiveness values presented by the DCC model for the pair market.
NEX/VVIX are equal for all models (0.2276), the same thing Moreover, results show that ADCC hedge provide the
for the ADCC model, where HE values are equals to 0,2145. highest hedging effectiveness for CDS, Gold and Euro Stoxx
Taking another example, the case of NEX/CDS hedge, the 50 and GO-GARCH hedge provide the highest HE for the
hedging effectiveness values obtained by the DCC model are other indices with the exception of NEX/Bond pair series
0.0155, 0.0153 and 0.0145 for the refits 10, 20 and 60 days which imply that its highest HE is achieved with DCC model.
respectively. In the case of NEX/VIX, the ADCC model Opposed to what have been reported by Ahmad et al [7], we
Copyright - 2019
Page 40
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
find that both ADCC and GO-GARCH are chosen over DCC For each GARCH model specification, Hedging
model. Additionally, our analysis suggest that NEX/VIX has effectiveness, transaction cost, and the TC/HE ratio produced
the highest HE ratio (HE= 0,40) which means that VIX is the from model refits every 10days, 20 days and 60 days are
best hedge for clean energy stocks followed by VSTOXX summarized in table 10. Based on the above, our findings
(HE= 0,37) and Euro Stoxx 50 (HE= 0,36). Overall, as for suggest that the VSTOXX is the most appropriate hedging
investment and hedging their risk inside portfolio creation, instrument due its lowest TC/HE ratio= 0,09 of all variables
investors who are looking for higher returns from NEX with the DCC model followed by the VIX index
should join it with VIX. This finding is compatible with those (TC/HE=0,11). However, GOLD is the least suitable hedging
of Ahmad et al. [7] and Hood and Malik [39] who prove that instrument due its highest TC/HE ratio = 4,32.
VIX is the best hedge for ECO and US equities respectively.
Table 9: Summary statistics of hedge ratios (β) and hedge effectiveness (HE) for NEX investors – MVT
NEX/CDS
DCC 0.0717 0.0150 0.2325 0.0155 0.0713 0.0150 0.2325 0.0153 0.0686 0.0018 0.2325 0.0145
ADCC 0.0440 0.0121 0.1984 0.0158 0.0437 0.0121 0.1984 0.0158 0.0417 0.0088 0.1984 0.0148
GOGARH 0.0670 -0.0526 0.0836 0.0090 0.0671 -0.0526 0.0829 0.0090 0.0707 -0.0448 0.5459 0.0090
NEX/Crude
Oil
DCC 0.1048 -0.0387 0.3881 0.0891 0.1050 -0.0343 0.3881 0.0891 0.1057 -0.0343 0.3877 0.0892
ADCC 0.1141 -0.0419 0.3888 0.0957 0.1143 -0.0376 0.3888 0.0957 0.1151 -0.0376 0.3888 0.0959
GOGARH 0.7663 0.3585 1.4717 0.1489 0.7740 0.3585 1.4746 0.1554 0.7714 0.3608 1.5144 0.1516
NEX/GOLD
DCC 0.0485 -0.8271 0.6082 0.0463 0.0486 -0.8271 0.6082 0.0458 0.0490 -0.8211 0.6082 0.0455
ADCC 0.0655 -0.8638 0.8278 0.0466 0.0656 -0.8638 0.8193 0.0462 0.0655 -0.8586 0.8060 0.0459
GOGARH 0.2529 -0.2155 0.5371 0.0411 0.2586 0.0028 0.5348 0.0396 0.1467 0.0167 0.1796 0.0401
NEX/BOND
DCC -0.7984 -2.4314 0.3467 0.0714 -0.8000 -2.4314 0.3334 0.0714 -0.8054 -2.4463 0.3334 0.0718
ADCC -0.7848 -2.4459 0.4512 0.0679 -0.7860 -2.4459 0.4376 0.0679 -0.7903 -2.4611 0.4376 0.0683
GOGARH -0.1905 -0.7619 -0.0048 0.0542 -0.2491 -0.3208 -0.0250 0.0542 -0.2515 -0.3204 -0.0555 0.0544
NEX/Euro
Stoxx 50
DCC 0.4591 0.2458 0.7096 0.3445 0.4594 0.2497 0.7096 0.3446 0.4607 0.2570 0.7050 0.3459
ADCC 0.4934 0.2487 0.8101 0.3638 0.4936 0.2519 0.8129 0.3637 0.4952 0.2572 0.8084 0.3653
GOGARH 0.8448 0.7043 0.9421 0.3247 0.3888 0.1090 0.7547 0.3249 0.8374 0.6866 0.9255 0.3232
NEX/VSTOX
X
DCC -0.0709 -0.1730 -0.0227 0.2908 -0.0709 -0.1731 -0.0232 0.2896 -0.0713 -0.1731 -0.0232 0.2898
ADCC -0.0717 -0.1657 -0.0280 0.2787 -0.0717 -0.1655 -0.0280 0.2787 -0.0721 -0.1655 -0.0280 0.2793
GOGARH -0.3128 -0.5966 -0.1536 0.3703 -0.3130 -0.5956 -0.1537 0.3707 -0.3143 -0.5955 -0.1540 0.3712
NEX /VIX
DCC -0.0671 -0.1642 -0.0155 0.3187 -0.0672 -0.1642 -0.0155 0.3192 -0.0676 -0.1642 -0.0156 0.3198
ADCC -0.0680 -0.1786 -0.0208 0.3006 -0.0681 -0.1786 -0.0208 0.3010 -0.0684 -0.1786 -0.0211 0.3015
GOGARH -1.3149 -7.2498 -0.5511 0.4065 -1.3134 -7.2479 -0.5528 0.4073 -0.3409 -0.6472 -0.1192 0.4092
NEX /VVIX
DCC -0.0813 -0.1946 -0.0240 0.2276 -0.0814 -0.1946 -0.0237 0.2276 -0.0817 -0.1946 -0.0237 0.2276
ADCC -0.0824 -0.1962 -0.0232 0.2145 -0.0824 -0.1978 -0.0232 0.2145 -0.0826 -0.1944 -0.0232 0.2145
GOGARH -1.1675 -7.1408 -0.3309 0.2966 -1.1671 -7.1407 -0.3309 0.2965 -0.2823 -0.5352 -0.1129 0.2970
NEX /OVX
DCC -0.0641 -0.1828 -0.0035 0.1098 -0.0642 -0.1828 -0.0033 0.1099 -0.0645 -0.1828 -0.0037 0.1100
ADCC -0.0616 -0.1619 -0.0001 0.1011 -0.0617 -0.1619 0.0001 0.1010 -0.0618 -0.1619 -0.0003 0.1012
GOGARH -0.2375 -0.3846 -0.1746 0.1623 -0.2422 -1.1264 -0.1746 0.1640 -0.7089 -1.8406 -0.2874 0.1619
Table 10: The hedging effectiveness (HE), the transaction cost TC, and the TC/HE ratio under different model refits
Refit=10 Refit=20 Refit=60
HE TC HE TC HE TC
NEX/CDS
DCC 1.55% 4.15 2.67 1.53% 4.18 2.72 1.45% 4.34 2.98
ADCC 1.58% 3.95 2.49 1.58% 3.90 2.46 1.48% 3.85 2.60
GOGARH 0.91% 2.48 2.73 0.90% 2.47 2.72 0.90% 4.59 5.06
NEX/Crude
Oil
DCC 8.91% 5.10 0.57 8.91% 5.08 0.57 8.92% 5.08 0.56
ADCC 9.57% 6.10 0.63 9.57% 6.09 0.63 9.59% 6.08 0.63
GOGARH 14.89% 26.76 1.79 15.54% 27.62 1.77 15.16% 26.75 1.76
NEX/GOLD
DCC 4.63% 24.36 5.26 4.58% 24.31 5.29 4.55% 24.26 5.32
ADCC 4.66% 25.85 5.53 4.62% 25.79 5.57 4.59% 25.71 5.59
GOGARH 4.11% 17.80 4.32 3.96% 17.11 4.31 4.01% 2.69 4.67
NEX/BOND
DCC 7.14% 40.08 5.61 7.14% 40.05 5.60 7.18% 39.96 5.56
ADCC 6.79% 45.59 6.70 6.79% 45.55 6.70 6.83% 45.54 6.66
GOGARH 5.42% 13.61 2.50 5.42% 5.64 1.03 5.44% 5.52 1.01
NEX/Euro
Stoxx 50
DCC 34.45% 14.10 0.40 34.46% 14.07 0.40 34.59% 13.97 0.40
ADCC 36.38% 22.61 0.62 36.37% 22.56 0.62 36.53% 22.55 0.61
GOGARH 32.48% 6.98 0.21 32.49% 20.01 0.21 32.32% 6.65 0.20
NEX/VSTOXX
DCC 29.08% 2.83 0.09 28.96% 2.84 0.09 28.98% 2.84 0.09
ADCC 27.87% 3.52 0.12 27.87% 3.52 0.12 27.93% 3.54 0.12
GOGARH 37.03% 9.52 0.25 37.07% 9.55 0.25 37.12% 9.68 0.26
NEX /VIX
DCC 31.87% 3.81 0.11 31.92% 3.82 0.11 31.98% 3.85 0.12
ADCC 30.06% 4.55 0.15 30.10% 4.56% 0.15 30.15% 4.57 0.15
GOGARH 40.65% 153.60 3.77 40.73% 153.51 3.76 40.92% 17.55 0.42
NEX /VVIX
DCC 22.76% 5.31 0.23 22.77% 5.31 0.23 22.76% 5.30 0.23
ADCC 21.45% 6.08 0.28 21.46% 6.09 0.28 21.45% 6.08 0.28
GOGARH 29.66% 150.41 5.07 29.65% 150.16 5.06 29.70% 10.48 0.35
NEX /OVX
DCC 10.98% 3.41 0.31 10.99% 3.42 0.31 11.00% 3.40 0.30
ADCC 10.11% 3.76 0.37 10.10% 3.77 0.37 10.12% 3.75 0.37
GOGARH 16.23% 5.08 0.31 16.40% 6.58 0.40 16.19% 44.46 2.74
In order to investigate the robustness of our findings on hedged by CDS, Crude Oil, Bond, GOLD, Euro Stoxx 50,
different forecast length, we calculate hedge ratio from fixed VSTOXX, VIX, VVIX and OVX. Based on several
length rolling window analysis. To this end, we fix, firstly, multivariate GARCH models refitted every 10, 20 and 60
our rolling window and refit the DCC, ADCC and GO- observations, our findings suggest that the VIX is the best
GARCH models every 20 observations, results are shown in hedge ratio for renewable energy as it has the highest HE,
table 11(see Annex). The forecasts lengths chosen are of 500, followed by VSTOXX and Euro Stoxx 50 which robust
1000 and 1500. We estimate GARCH models (DCC and through the different forecast length.
ADCC) with a student distribution (MVT). The GO-GARCH While previous studies focuses on estimating time varying
estimated using a multivariate affine negative inverse hedge rat, our paper takes a new approach and analyzing the
Gaussian (MANIG) distribution. Results show that Euro effect of the tradeoff between transaction costs and
Stoxx 50 provides the most effective hedge for NEX only for effectiveness hedging on the portfolio decisions, which is
500 forecast length, however, for the 1000 and 1500 forecast considered as new insight into hedging strategies for clean
length, the VIX is the best hedge ratio. Additionally, for the energy investors. Our significantly different results show that
NEX/VIX hedge, the DCC model is preferred (largest HE the VSTOXX is the best hedging instrument for renewable
value) across all forecast horizons. The same case for the energy since it yields the lowest TC/HE followed by the VIX.
NEX/VSTOXX, NEX/VIX and NEX/Crude Oil hedges. For
the NEX/CDS hedge, the DCC model is preferred for
forecasts lengths of 500 and 1000, and ADCC is preferred for ACKNOWLEDGMENT
1500 forecast. For the pair NEX/GOLD hedge the DCC I gratefully thank an anonymous referee for the helpful
model is preferred for 500 and 1000 forecast length, while the comments and suggestions that have improved the quality of
GO-GARCH model is preferred for 1500 forecast length. the paper.
However, for the NEX/OVX, the DCC model is chosen only
for 500 forecast and GO-GARCH model for 1000 and 1500. REFERENCES
The table 12 (see Annex) report the transaction costs, the
hedging effectiveness and the TC/HE of the rolling window [1] Ahmad, W. (2017). On the dynamic dependence and investment
performance of crude oil and clean energy stocks. Research in
estimations with different forecasts length. Results show that International Business and Finance, 42, 376-389.
the VIX is the best hedge ratio for NEX in all cases. [2] Broadstock, D. C., Cao, H., & Zhang, D. (2012). Oil shocks and their
Moreover, in most situations the GO-GARCH model makes impact on energy related stocks in China. Energy Economics, 34(6),
the best hedge decision with the exception of VSTOXX, VIX, 1888-1895.
[3] Bamati, N., & Raoofi, A. (2019). Development level and the impact of
and VVIX where the DCC yields the best hedge ratio. technological factor on renewable energy production. Renewable
Energy.
[4] Kumar, S., Managi, S., & Matsuda, A. (2012). Stock prices of clean
energy firms, oil and carbon markets: A vector autoregressive
analysis. Energy Economics, 34(1), 215-226.
VI. CONCLUSIONS [5] Bouraiou, A., Necaibia, A., Boutasseta, N., Mekhilef, S., Dabou, R.,
Facing to the rapid growth of the world’s population Ziane, A., ... & Touaba, O. (2019). Status of renewable energy
industrialization, urbanization, the energy demand and carbon potential and utilization in Algeria. Journal of Cleaner Production,
119011.
dioxide emissions are expected to have an important increase [6] International Energy outlook 2017, U.S. Energy Information
over the next twenty years all over the world. Thus allowing, Administration, September 14, 2017. www.eia.gov/ieo.
the scientific community has reported that increasing the [7] Ahmad, W., Sadorsky, P., & Sharma, A. (2018). Optimal hedge ratios
usage of clean energy is the greatest solution on reducing for clean energy equities. Economic Modelling, 72, 278-295.
[8] Pham, L. (2019). Do all clean energy stocks respond homogeneously
world’s dependence on fossil fuels as well as fighting climate to oil price?. Energy Economics, 81, 355-379.
change by reducing greenhouse gas emissions. Based on the [9] Sadorsky, P. (2012). Correlations and volatility spillovers between oil
previous challenges, investors need to hedge their prices and the stock prices of clean energy and technology
investments against risk fluctuations of renewable energy companies. Energy Economics, 34(1), 248-255.
[10] Sánchez, E. R. (2015). VOLATILITY SPILLOVERS AMONG
assets. This paper, estimate time-varying optimal hedge ratio ALTERNATIVE ENERGY, OIL AND TECHNOLOGY GLOBAL
for the aim of whether an investment in clean energy can be
MARKETS (Doctoral dissertation, Universidad de Castilla-La [26] Dutta, A. (2017). Oil price uncertainty and clean energy stock returns:
Mancha). new evidence from crude oil volatility index. Journal of Cleaner
[11] Elie, B., Naji, J., Dutta, A., & Uddin, G. S. (2019). Gold and crude oil Production, 164, 1157-1166.
as safe-haven assets for clean energy stock indices: Blended copulas [27] Asma A, Ahmed G. Hedging Oil Prices with Renewable Energy
approach. Energy, 178, 544-553. Indices: A Comparison between Various Multivariate Garch Versions.
[12] Jitmaneeroj, B. (2018). The effect of the rebalancing horizon on the Biostat Biometrics Open Acc J. 2018; 6(3):
tradeoff between hedging effectiveness and transaction [28] http://www.wildershares.com/stock.php
costs. International Review of Economics & Finance, 58, 282-298. [29] Tully, E., & Lucey, B. M. (2007). A power GARCH examination of
[13] Henriques, I., & Sadorsky, P. (2008). Oil prices and the stock prices of the gold market. Research in International Business and Finance, 21(2),
alternative energy companies. Energy Economics, 30(3), 998-1010. 316-325.
[14] Sadorsky, P. (2012). Modeling renewable energy company [30] Shahzad, S. J. H., Raza, N., Shahbaz, M., & Ali, A. (2017).
risk. Energy Policy, 40, 39-48. Dependence of stock markets with gold and bonds under bullish and
[15] Ferstl, R., Utz, S., & Wimmer, M. (2012). The effect of the Japan bearish market states. Resources Policy, 52, 308-319.
2011 disaster on nuclear and alternative energy stocks worldwide: An [31] Zghal, R., Ghorbel, A., & Triki, M. (2018). Dynamic model for
event study. Business Research, 5(1), 25-41. hedging of the European stock sector with credit default swaps and
[16] Broadstock, D. C., Cao, H., & Zhang, D. (2012). Oil shocks and their EURO STOXX 50 volatility index futures. Borsa Istanbul
impact on energy related stocks in China. Energy Economics, 34(6), Review, 18(4), 312-328.
1888-1895. [32] Engle, R. (2002). Dynamic conditional correlation: A simple class of
[17] Managi, S., & Okimoto, T. (2013). Does the price of oil interact with multivariate generalized autoregressive conditional heteroskedasticity
clean energy prices in the stock market?. Japan and the World models. Journal of Business & Economic Statistics, 20(3), 339-350.
Economy, 27, 1-9. [33] Cappiello, L., Engle, R. F., & Sheppard, K. (2006). Asymmetric
[18] Bohl, M. T., Kaufmann, P., & Stephan, P. M. (2013). From hero to dynamics in the correlations of global equity and bond returns. Journal
zero: Evidence of performance reversal and speculative bubbles in of Financial econometrics, 4(4), 537-572.
German renewable energy stocks. Energy Economics, 37, 40-51. [34] Van der Weide, R. (2002). GO‐GARCH: a multivariate generalized
[19] Ortas, E., & Moneva, J. M. (2013). The Clean Techs equity indexes at orthogonal GARCH model. Journal of Applied Econometrics, 17(5),
stake: Risk and return dynamics analysis. Energy, 57, 259-269. 549-564.
[20] Wen, X., Guo, Y., Wei, Y., & Huang, D. (2014). How do the stock [35] Glosten, L. R., Jagannathan, R., & Runkle, D. E. (1993). On the
prices of new energy and fossil fuel companies correlate? Evidence relation between the expected value and the volatility of the nominal
from China. Energy Economics, 41, 63-75. excess return on stocks. The journal of finance, 48(5), 1779-1801.
[21] Inchauspe, J., Ripple, R. D., & Trück, S. (2015). The dynamics of [36] Ku, Y. H. H., Chen, H. C., & Chen, K. H. (2007). On the application
returns on renewable energy companies: A state-space of the dynamic conditional correlation model in estimating optimal
approach. Energy Economics, 48, 325-335. time-varying hedge ratios. Applied Economics Letters, 14(7), 503-509.
[22] Reboredo, J. C. (2015). Is there dependence and systemic risk between [37] Chang, C. L., McAleer, M., & Tansuchat, R. (2011). Crude oil
oil and renewable energy stock prices?. Energy Economics, 48, 32-45. hedging strategies using dynamic multivariate GARCH. Energy
[23] Bondia, R., Ghosh, S., & Kanjilal, K. (2016). International crude oil
Economics, 33(5), 912-923.
prices and the stock prices of clean energy and technology companies:
[38] Chen, F., & Sutcliffe, C. (2012). Better cross hedges with composite
Evidence from non-linear cointegration tests with unknown structural
hedging? Hedging equity portfolios using financial and commodity
breaks. Energy, 101, 558-565.
futures. The European Journal of Finance, 18(6), 575-595.
[24] Reboredo, J. C., Rivera-Castro, M. A., & Ugolini, A. (2017). Wavelet-
[39] Hood, M., & Malik, F. (2013). Is gold the best hedge and a safe haven
based test of co-movement and causality between oil and renewable
under changing stock market volatility?. Review of Financial
energy stock prices. Energy Economics, 61, 241-252.
Economics, 22(2), 47-52.
[25] Ahmad, W. (2017). On the dynamic dependence and investment
[40]
performance of crude oil and clean energy stocks. Research in
International Business and Finance, 42, 376-389.
Copyright - 2019
Page 44
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
ANNEX :
Table 11: Summary statistics of hedge ratios (β) and Hedging Effectiveness (HE) for NEX investors under alternative
forecast length
Forecast length= 500 Forecast length= 1000 Forecast length= 1500
mean min max HE mean min max HE mean min max HE
NEX/CDS
DCC 0.0485 -0.0963 0.2343 0.0096 0.0543 -0.0377 0.2343 0.0105 0.0619 -0.0271 0.2343 0.0119
ADCC 0.0330 -0.0250 0.2006 0.0095 0.0350 -0.0119 0.2006 0.0103 0.0389 -0.0105 0.2006 0.0120
GOGARH 0.0689 -1.2063 1.5832 0.0076 0.1088 -0.0137 0.6195 0.0081 0.0670 -0.0641 0.0828 0.0091
NEX/Crude
Oil
DCC 0.2301 -0.0668 1.0334 0.1408 0.1832 -0.0668 0.8780 0.1009 0.1140 -0.0668 0.3880 0.0736
ADCC 0.2283 -0.0618 0.9754 0.1455 0.1831 0.0103 0.7553 0.1074 0.1224 -0.0618 0.3899 0.0806
GOGARH 0.3701 0.1468 1.7304 0.1836 0.3405 0.1468 1.4826 0.1772 0.2701 0.1468 0.7834 0.1618
NEX/GOLD
DCC 0.1330 -1.0938 0.8862 0.0677 0.1133 -0.8265 0.8106 0.0571 0.0440 -0.8265 0.6126 0.0364
ADCC 0.1465 -1.1444 0.9811 0.0678 0.1263 -0.8630 0.9576 0.0571 0.0546 -0.8630 0.8328 0.0366
GOGARH 0.1520 -1.0978 0.2884 0.0646 0.1554 0.0502 0.1869 0.0476 0.2794 0.0167 0.8705 0.0451
NEX/BOND
DCC -0.9658 -2.5751 0.3289 0.0860 -0.9348 -2.4287 0.3289 0.0807 -0.7716 -2.4287 0.3289 0.0662
ADCC -0.920 -2.5475 0.4353 0.0793 -0.9065 -2.5475 0.4353 0.0759 -0.7463 -2.4430 0.4353 0.0615
GOGARH -0.2361 -1.1139 -0.0020 0.0707 -0.2686 -0.3406 -0.0271 0.0714 -0.2503 -0.3196 -0.0271 0.0521
NEX/Euro
Stoxx 50
DCC 0.5699 0.2506 1.1414 0.4275 0.5303 0.2506 0.9443 0.3876 0.4854 0.2506 0.8614 0.3501
ADCC 0.6140 0.2540 1.1822 0.4425 0.5792 0.2540 1.1647 0.4064 0.5286 0.2540 0.9909 0.3704
GOGARH 0.8550 0.5774 1.1323 0.3917 0.4365 0.1046 1.3443 0.3645 0.8442 0.6759 0.9389 0.3328
NEX/VSTO
XX
DCC -0.1056 -0.3288 -0.0216 0.3311 -0.0938 -0.3288 -0.0216 0.3140 -0.0766 -0.1730 -0.0216 0.2900
ADCC -0.1054 -0.3158 -0.0277 0.3174 -0.0947 -0.3118 -0.0277 0.3020 -0.0777 -0.1654 -0.0277 0.2818
GOGARH -0.3652 -1.1997 -0.1536 0.3829 -0.3592 -1.1999 -0.1536 0.3853 -1.2500 -4.5260 -0.4900 0.3780
NEX /VIX
DCC -0.0966 -0.2815 -0.0155 0.3548 -0.0870 -0.2463 -0.0155 0.3435 -0.0752 -0.1642 -0.0155 0.3397
ADCC -0.0972 -0.3267 -0.0207 0.3368 -0.0873 -0.3073 -0.0207 0.3245 -0.0759 -0.1784 -0.0207 0.3215
GOGARH -0.3775 -1.0042 -0.1178 0.4137 -0.3793 -0.9238 -0.1178 0.4278 -0.3459 -0.6472 -0.1178 0.4206
NEX /VVIX
DCC -0.1103 -0.3970 -0.0245 0.2212 -0.1006 -0.3970 -0.0245 0.2264 -0.0869 -0.1946 -0.0245 0.2327
ADCC -0.1105 -0.4311 -0.0232 0.2100 -0.1005 -0.3047 -0.0232 0.2143 -0.0877 -0.1970 -0.0232 0.2200
GOGARH -0.2807 -0.9639 -0.1129 0.2594 -0.2873 -0.9648 -0.1129 0.2827 -1.1917 -7.1704 -0.3309 0.2985
NEX /OVX
DCC -0.0870 -0.3836 -0.0033 0.1343 -0.0780 -0.3836 -0.0033 0.1246 -0.0667 -0.1827 -0.0033 0.1102
ADCC -0.0826 -0.3665 0.0035 0.1246 -0.0743 -0.3665 0.0001 0.1161 -0.0634 -0.1616 0.0001 0.1016
GOGARH -0.5673 -2.4910 -0.0826 0.1272 -0.6503 -2.4910 -0.2147 0.1517 -0.2403 -0.3843 -0.1748 0.1574
HE TC HE TC HE TC
NEX/CDS
DCC 0.96% 12.99 13.48 1.05% 9.26 8.81 1.19% 7.51 6.28
ADCC 0.95% 9.61 10.12 1.03% 5.86 5.64 1.20% 5.77 4.80
GOGARH 0.76% 21.54 28.08 0.81% 12.31 15.18 0.91% 1.56 1.71
NEX/Crude Oil
DCC 14.08% 18.60 1.32 10.09% 15.97 1.58 7.36% 9.01 1.22
ADCC 14.55% 18.93 1.30 10.74% 15.73 1.46 8.06% 9.76 1.21
GOGARH 18.36% 25.20 1.37 17.72% 21.40 1.20 16.18% 10.31 0.63
NEX/GOLD
DCC 6.77% 38.34 5.65 5.71% 30.66 5.36 3.64% 26.25 7.20
ADCC 6.78% 42.31 6.23 5.71% 33.75 5.90 3.66% 27.37 7.46
GOGARH 6.46% 12.83 1.98 4.76% 1.25 0.26 4.51% 18.71 4.14
NEX/BOND
DCC 8.60% 51.53 5.99 8.07% 39.63 4.90 6.62% 37.78 5.70
ADCC 7.93% 70.68 8.91 7.59% 47.83 6.30 6.15% 43.93 7.14
GOGARH 7.07% 21.93 3.09 7.14% 4.00 0.56 5.21% 5.14 0.98
NEX/Euro
Stoxx 50
DCC 42.75% 20.06 0.46 38.76% 17.66 0.45 35.01% 15.78 0.45
ADCC 44.25% 31.67 0.71 40.64% 30.46 0.74 37.04% 26.54 0.71
GOGARH 39.17% 16.78 0.42 36.45% 32.06 0.87 33.28% 7.71 0.23
NEX/VSTOXX
DCC 33.11% 6.08 0.18 31.40% 4.58 0.14 29.00% 3.94 0.13
ADCC 31.74% 7.68 0.24 30.20% 5.87 0.19 28.18% 4.81 0.17
GOGARH 38.29% 19.72 0.51 38.53% 16.42 0.42 37.80% 70.08 1.85
NEX /VIX
DCC 35.48% 4.99 0.14 34.35% 4.66 0.13 33.97% 4.69 0.13
ADCC 33.68% 6.42 0.19 32.45% 5.64 0.17 32.15% 5.49 0.17
GOGARH 41.37% 19.74 0.47 42.78% 18.50 0.43 42.06% 18.63 0.44
NEX /VVIX
DCC 22.12% 6.95 0.31 22.64% 5.74 0.25 23.27% 6.33 0.27
ADCC 21.00% 7.37 0.35 21.43% 5.98 0.27 22.00% 6.80 0.30
GOGARH 25.94% 11.15 0.43 28.27% 10.76 0.38 29.85% 132.62 4.44
NEX /OVX
DCC 13.43% 7.91 0.58 12.46% 4.41 0.35 11.02% 4.08 0.37
ADCC 12.46% 8.20 0.65 11.61% 4.53 0.39 10.16% 4.37 0.43
GOGARH 12.72% 18.56 1.45 15.17% 40.91 2.69 15.74% 6.00 0.38
Table 12: The Hedging Effectiveness, the Transaction Cost, and the TC/HE ratio for different forecast lengths
Abstract-Metronidazole is an antibiotic drug for the methods can be used to remove residues in
treatment of annerobic infections. When discharged wastewater [4]. Coagulation-floculation processes
in water after use, it reacts with living organisms thus have been mainly used for wastewater treatment to
causing adverse effects for their lives. This kind of separate suspended and/or fatty particles, the main
residue must be removed from wastewater. function of coagulants widely used like aluminum
In this work, the elimination of metronidazole is and iron salts is to flocculate colloidal particuls into
carried using coagulation-flocculation and
larger particles that can be removed by precipitation
electrocoagulation techniques in batch mode.
The first part of this work is devoted to the study of or flotation[5].
the degradation of metronidazole by coagulation- Currently, electrocoagulation is another
flocculation. The results obtained indicate that process drawing the attention from the scientific
degradation rates do not exceed 14.13% with the community due to the many advantages it offers, it
optimal dose of the coagulant Al2(SO4)3 but they is referred to as an environmental technology owing
increase to 73.84% for a pH 9. to the fact that it uses a clean reagent (electron) for
The second part of this work deals with the study of the abatement of organic and recalcitrant pollutants
the elimination of the considered antibiotic by [6].
electrocoagulation process. The coagulant provided
In this perspective, the present study
from aluminum electrodes led to a reduction of
77.12% with 1A of applied current, the optimal pH investigates the removal of metronidazole from
conditions increase this yield to 80.10%. synthetic wastewater by coagulatio-floculation and
An improvement in yields even with high initial electrocoagulation.
concentrations of drug in water was recorded with
electrocoagulation compared to coagulation. II. Experimental design, materials and methods
Overall, after conducting an economic study, the cost A. Chemicals
is lower in coagulation-flocculation than in -Pharmaceutical compound: Metronidazole
electrocoagulation. This trend is mainly related to solid with the properties given in table 1. 30mg
aluminum consumption which is higher in
was added to 1L of distilled water, the mixture
electrocoagulation syetem.
was homogenized by slight magnetic stirring
Keywords- Elimination, Metronidazole, Coagulation- for 24h.
flocculation, Electrocoagulation. – Al2(SO4)3· 16H2O was used as coagulant
- "SP6" is a cationic polymer was used as
flocculent.
I. Introduction TABLE I
The presence of pharmaceutical products in Physical and chemical properties of metronidazole [4]
our water resources has raised great concern among
scientists worldwide due to their persistence in Molecular formula C6H9N303
nature via different ways including modern Usage Antibiotic
medicine and veterinary industries, agricultural Molecular weight (g/mol) 171,2
runoff, direct discharge from refineries of urban Solubility in water (g/mol) 9,5
and hospital wastewater, human waste[1,2]. Due to pKa 2,55
their highly polluting nature, it is not possible to Melting point 159-163
discharge treated and untreated waste either into Molecular structure
water course or in the soil without causing great
damage [3]. In order to treat this pollution, several
Process Coagulation-flocculation
Parameter
Coagulant 100 120 150 175 200 250
dose(mg/l)
pH solution 1 3 5 9
Initial 10 20 30 40 50
[Metronidazole]
Fig .2 Electrocoagulation apparatus Process Electrocoagulation
Parameter
C. Experimental procedure Current intensity 0,25 0,5 0,75 1 1,25
(A)
1) Jar test coagulation: pH solution 1 3 5 9
Coagulation-flocculation treatment was carried out Initial 20 30 40 50 60
in a conventional jar test apparatus (model Velp [Metronidazole]
Scientifica C6F). The experimental procedure is as
follow: E. Calculations
(1) Dosing stage: Water sample was put into Evaluation of coagulation – flocculation/
the 1L beaker. After adjusting pH solution, electrocoagulation efficiency is given as
if needed, coagulant and flocculent were C0−Ce
R% = × 100 (1)
added then the mixture was stirred for C0
C0: Initial metronidazole concentration
15min with a speed of 120rpm
Ce: The equilibrium concentration.
(2) Coagulation stage: After 15 min rapid
The calculation of the operational cost of
mixing, the stirring speed was reduced to 45
the two water treatment processes takes into
rpm and continued to stir 45 min.
account the consumed aluminum sulphate and the
(3) Settling stage: After 30min stationary, the
electrical energy consumed by the flocculator.
water samples were taken to analysis by
According to the author [5], the calculation
spectrophotometry UV visible.
equations used are the following:
𝑜𝑝𝑒𝑟𝑎𝑡𝑖𝑜𝑛𝑎𝑙 𝑐𝑜𝑠𝑡 = 𝑎𝐶𝑒𝑛𝑒𝑟𝑔𝑦 + 𝑏𝐶𝐴𝑙3+
2) Electrocoagulation : ′𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑑𝑒𝑠)
Electrocoagulation runs were carried out in a (2)
batch mode (Figure1). 1,5 g of NaCl was added,
the pH was adjusted to the desirable values using C’ = b CAl3+Al2 (SO4)3.18 H2O (3)
Hcl and NaOH solutions. A pair of aluminum
electrodes of size 12 cm×3 cm×0.5 cm, separated Where, (a) and (b) are ratios for the price
by a space of 1 cm and dipped in the wastewater. calculation of the international market on energy
The current was provided by a GW GPR3030D, and chemicals for the year 2011, they are
3A-30 VDC power supply. The effluent under successively 0.05 US $/kWh, and 3.08 $/kg of
treatment was homogenized by gentle magnetic aluminum.
stirring at 300 rpm which allows the separation of For coagulation treatment, the value of
gasses formed from the solution, thus avoiding the aluminum consumed is calculated by the
formation of foam which can affect the course of knowledge of the Al3 + concentration introduced in
the batch process. After runs, samples are the form of Al2(SO4)3 salt. For electrocoagulation,
subjected to settling for 30 min. the theoretical concentration of dissolved aluminum
by unit of treated wastewater volume (KgAl/m3) is
The analysis of the samples is carried out given by Faraday law:
𝐼𝑡𝑀
before and after experimentation by UV-visible 𝐶𝐴𝑙 3+ (𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑑𝑒) = (4)
𝑍𝐹𝑉
spectrophotometry (Model 2300UV Where, M the aluminum molecular weight (27
spectrophotometer) at the absorption wavelength of g/mol), I the current intensity applied (A), t the
metronidazole (320 nm). The calibration line is mean residence time in the reactor, Z the number of
plotted by measuring the absorbances of a series of electrons involved in reaction of dissolution (Z = 3
standard solutions obtained by diluting the stock
solution of metronidazole.
for aluminum), F the constant of Faraday (96500 From Fig.3, we deduced that the pH variation
C/mol) and V the volume of the reactor (1L). has an important influence on the removal yield.
The consumed energy by unit of treated wastewater When de solution pH take the value 9, the
volume (KWh/m3), is given by Eq 5, removal yield take it’s high level of 73,84%.
𝑡
KWh 𝐼 ∫ 𝑈𝑑𝑡
𝑆𝐸𝐸𝐶 ( 3 ) = 0 3 (5)
𝑚 𝑉
90
U is the average cell voltage
R%
80
I is current intensity in ampere 70
T is time in hour 60
V is volume of the solution (m3) 50
40
30
III. Results 20
1. Elimination of metronidazole by 10
coagulation-floculation 0
1 3 5 7.2 9 PH
A) Effect of coagulant dose Fig.3 Effect of pH on the removal of Metronidazole
Coagulation-floculation tests are carried out on by coagulation-floculation.
synthetic wastewater containing 50g/l of
Metronidazole without regulating pH solution. The C). Effect of initial Metronidazole concentration
temperature is at 20 to250 C and the coagulant At room temperature, coagulant dose of 150
dosage is of 100 to 250 mg/l. mg / l and pH = 9, we calculated after coagulation
The results presented in fig.2 showed that treatment the residual concentrations of
increasing the dose of coagulant increases the metronidazole in solution in function of different
removal rate of metronidazole. However, when the initial concentrations. Obtained results are shown in
dosage is more than a certain value and continues to fig.4
70
be added, the removal efficiency will reduced. The
R %
100
R(%)
From Fig.7 that shows the removal yield of
80 metronidazole versus of initial concentration, the
initial concentration has a remarqueable effect on
60
Metronidazole removal which ranged from 87,73%
40 to 68,46 % going to high concentrations. Such
20 behavior was observed by authors [7, 8].
0
3. Optimization of treatment removal and
0 100 200 300 400
Current density (A/m2)
cost analysis
At the optimum conditions, calculations done with
equations mentioned abode are presented in table 2.
Fig.5 Effect of current density on the removal of TABLE II
Metronidazole by electrocoagulation Operational cost of coagulation and electroagulation
It can be seen from the fig.5 that in the range Consumed Consumed Cost Removal
69,44-277,77 A/m2 , the Metronidazole removal Al(g/m3) Energy(KWh/m3) DZD/m3 Efficiency
efficiency yield increased significantly achieving a (%)
89.53 %. The increase of density up to a value of CC 12,159 / 2918 79,95
277,77 A/m2 did not lead to a significant EC 335,54 18700 76142,12 89,53
improvement of Metronidazole reduction. From
data analysis it can be stated that the optimum IV. Conclusion
current density is 277,77 A/m2. The aim of this work was to eliminate a
B. Effect of pH pharmaceutical antibiotic "metronidazole" in a
The initial synthetic wastewater pH is a key synthetic effluent by means of two processes:
parameter of the electrocoagulation process. To coagulation-flocculation and electrocoagulation.
study the pH influence on Metronidazole removal, It can be observed that the application of
the pH of synthetic wastewater was adjusted to the coagulation-flocculation gave almost no reduction
desired value that ranges from 3to 9. Fig.6 shows with an optimal mass of the coagulant Al2 (SO4)3,
the results obtained, standing out that in the process knowing that the yield reached 22,03 % at real pH
stage at pH 6 and respectively 7 and 9, removal solution. This yield is enhanced from 32% to 89%
efficiency reached its maximum level. by electrocoagulation even with low applied current
100 values and without adjusting pH solution. For the
R(%)
90 effect of initial pH, it’s remarqueable in the case of
80 coagulation, the yield pass to 79,95% and without
70
60
significant effect on electrocoagulation.
50 It should be noted that the study of the
40 effect of initial Metronidazole concentration on
30
treatment processes is significantly better with
20
10
electrocoagulation than with coagulation-
0
pH
flocculation.
0 2 4 6 8 10
References
Fig.6 Effect of pH on the removal of Metronidazole [1] Kummerer, K, “Pharmaceuticals in the Environment:
by electrocoagulation. Sources, Fate, Effects, and Risks”, second ed. Springer, New
c) Effect of initial concentration on elimination of York, 2004.
metronidazole by electrocoagulation [2] Beausse, J, “Selected drugs in solid matrices: a review of
environmental determination, occurrence and properties of
The effect of varying initial concentration principal substances”, Trends Anal. Chem. 23, 753–761, 2004.
of Metronidazole (20 to 60 mg/L) on efficiency [3]Jjemba, P.K., “Excretion and ecotoxicity of pharmaceutical
removal was investigated using the following and personal care products in the environment”. Ecotoxicology.
operating parameters: current density of Environ. Saf. 63, 113–130, 2006.
[4] Homem V., Lúcia S, “Degradation and removal methods of
277,77A/m2 mA/cm2 and reel pH solution. antibiotics from aqueous matricese. A review”. Journal of
100 Environmental Management. 92 : 2304-2347, 2011.
R(%)
[5] A. Amirtharajah, C.R. O’Melia, “Coagulation process:
80
destabilization, mixing, and flocculation”, Chap. 4. In: Water
60
Quality and Treatment: A Handbook of Community Water
Supplies, 4th ed. AWWA, New York, 1990.
40 [6] Lia, T., Zhua, Z., Wanga, D., Yaob, C., Tang, H.,
“Characterization of floc size, strength and structure under
20 various coagulation mechanisms”. Powder Technol. 168, 104–
0
110, 2006.
0 10 20 30 40 50 60 70 [7] Alexandra raluca miron,Ana-maria Andreea Chivu and al,
Initial (Mteronidazole) (mg/l)
“Pharmaceutical industry wastewater treatment through
Fig.7 Effect of initial concentration on the removal electrocoagulation”; revista de chimie. 65(12):1399-1406, 2014
of Metronidazole by coagulation-floculation. [8] Benny Marie Band and al, “Removal of Pharmaceuticals
from Wastewater by Intermittent Electrocoagulation”; Water
9(2) 85; 2017.
Abstract— High Concentration Photovoltaics (HCPV), can be concentration levels in the sub-cell is based on the value of
considered as a new technology in the field of solar energies, is in the desired bandgap. Due to the complexity and to the high
continuous progress. The use of less expensive optical costs of manufacturing this type of solar cell, researchers
components in the design of HCPV systems reduce significantly have used optical concentrators, such as Fresnel lenses [4] or
the capital costs of the PV power plant installations and,
parabolic mirrors [5], which are less expensive and
subsequently, the levelized cost of electricity (LCOE). However,
the major disadvantage of these systems is the need of a high concentrate the flow of sunlight on a reduced active area of a
level tracking accuracy systems. In this paper, we focus on the high efficiency MJSC. CPVs generators convert direct normal
characteristic of Triple-Junction solar cell (TJSC) irradiation (DNI), provided by the sun, into electric power [6].
InGaP/InGaAs/Ge, with n-on-p polarity on Ge substrate at This means that a dual axis sun tracker must be mounted on
different angles of incidence of solar direct irradiation, relative the CPV module (MJSC + concentrator), to maintain the
to the plane of the concentrator surface for two groups of TJSC; concentrator surface and the DNI flux in permanent
with a secondary optic element (SOE) and without SOE., The perpendicularity. This is one of the disadvantages of the
solar simulator (SAV-HELIOS3198) of the University of Jaén, CPVs. The higher the concentration rate is, the more accurate
Spain was used to simulate the sun spectrum.
of the sun tracking systems are required. High precision is
expensive and requires additional maintenance costs, which
Keywords— Multijunction, CPV, Concentrator, Photovoltaic, affects the cost of energy production. Here, precision is
Generator, defined by the parameter of the acceptance angle. The main
objective of this work is to evaluate the impact of adding the
I. INTRODUCTION SOE to the concentrator on the electrical performance of the
CPV module. This is done by characterizing the current -
The MJSCs were developed in the 1990s for space voltage I - V of the MJSC solar cells at different angles of tilt.
applications [1]. The manufacturing cost of this type of solar .
cells is relatively, high (~10.000$/kg) [2]. Nowadays, MJSC
are gaining popularity in terrestrial photovoltaic II. MATERIAL AND RESULTS
applications [3]. The concentration photovoltaic systems
(CPV) are one of these applications. The MJSCs are formed A. Optical devices
by a stack of different sub-cells (usually 3), with different
bandgap each one, to absorb a broader range of wavelengths We used two TJSC type: 3C44A from AZUR SPACE®
of the solar spectrum for a high conversion into electrical (Fig.1), coated with a multi-layer antireflective giving low
energy. These sub-cells are formed of an alloy of semi- reflectance over wavelength range 0.3 to 1.8μm, and were
conductor elements (III and V family of the periodic table). optimized for HCPV applications, under concentrated
The choice of semiconductor materials with their incident illumination (up to 300 Suns).
Page 53
Copyright - 2019
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Cups Filters
No 00 50
Std X X X
Brilliant X - -
Transparent - X X
5,0
4,5
4,0
3,5
3,0
Fig. 2 CPV module with SOE.
2,5
I (A)
with SOE
1,0 without SOE
10%
0,8
MPP (normalized)
0,6
Fig.5 Acceptance angle for CPV module with SOE and without SOE
0,4
The different I-V curves of the TJSC characterization was
done in ambient temperature Tamb=18°C. According to the
0,2
calibration of spectral distribution tests done before (Fig.4.),
we choose DNI= 850W/m² (SMR=1) to characterize the two
CPV modules (with SOE and without SOE), at different 0,0
angles of tilt. Fig.6. and Fig.7. show the behaviour of the I-V 0° 0,2° 0,4° 0,6° 0,8° 1° 1,2° 1,4° 1,6° 1,8° 2,0° 2,2° 2,4° 2,6° --
DNI tilt angle
curves for the TJSC of the CPV module without SOE and the Fig.8. variation of the MPP for the two CPV modules at
CPV module with SOE, respectively. different tilt angle
0
4 0,2 1,0
0,4
0,6
0,8
3 0,8
1
Normalized values
1,2
1,4
I (A)
0,6
2
0,4
1 Isc[A]
Voc[V]
0,2 Pmp[W]
FF(ad)
0
Eff
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 (a)
V (V)
0,0
Fig.6. I-V curve for the TJSC without SOE at different tilt angle 0.0°0.2°0.4°0.6°0.8°1.0°1.2°1.4°1.6°1.8°2.0°2.2°4.4°2.6°
5
0°
1° 1,0
1,8°
4 2°
2,2° 0,8
2,4°
Normalized values
3 2,6°
0,6
I(A)
2
0,4
Voc
Isc
1 Pmp
0,2 E
FF (b)
0
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 0,0
0° 0,2° 0,4° 0,6° 0,8° 1° 1,2° 1,4°
V(V)
DNI tilt angle
Fig.7. I-V curves for the TJSC with SOE at different tilt angle Fig.9 . Behavior of different electrical parameters of TJSC, a) with SOE
(up) and b) without SOE (down) at different tilt angle (DNI =850W/m²)
3
All hypertext links and section bookmarks will be removed terrestrial III–V multijunction solar cells for concentrator
from papers during the processing of papers for publication. photovoltaic systems”. Sol. Energy Mater. Sol. Cells 2010,
If you need to refer to an Internet email address or URL in 94, 1314–1318.
your paper, you must type out the address or URL fully in
Regular font. [4] D. Rumyantsev, V. M. ANDREEV, A. W. Bett, F.
Dimroth, M. Hein, G. Lange, M. Z. Shvarts, and O.
III. CONCLUSION Sulima “, Progress in Development of All-Glas
Terrestrial Concentrator Modules Based on Composite
In this work, a CPV module with secondary optical
Fresnel Lenses and III-V Solar Cells” in Proceedings
element was characterized with a concentration of 650 suns,
of the 28th IEEE Photovoltaic Specialists Conference,
using a solar simulator with 1.5D AM spectrum. It has been
1169-1172 (2000).
shown that SOE can increase the acceptance angle up to 40%,
compared to the CPV module without SOE. This gives more
[5] M Dreger, Maike Wiesenfarth, Tobias Schmid and
tolerance to the tracking systems, which has a direct impact
Andreas W. Bett, “Analysis of high concentration
on the costs of electricity production using the CPV
passively cooled cpv module designs using mirror optics”,
generators.
Conference: 28th EU-PVSEC 2013
[3] Law, D.C.; King, R.; Yoon, H.; Archer, M.; Boca, A.;
Fetzer, C.; Mesropian, S.; Isshiki, T.; Haddad, M.;
Edmondson, K., “Future technology pathways of
Abstract— This paper compares the shunt controllers and contingency cases must be determined. Voltage instability is
considers the impacts of the Static Var Compensator (SVC) and mainly associated with reactive power imbalance [2-3].
Static Synchronous Compensator (STATCOM) on stability of Reactive Power, also known as use-less/watt-less power, is
voltage and reactive power compensation. The objective of this the power merely absorbed and returned in load due to its
work is to improve the dynamic voltage control specially with
reactive properties; in this case, energy is first stored and then
increasing use of nonlinear loads; this study compares the
voltage regulating system stability, the reactive power and the released in the form of magnetic or electrostatic field in
economical of this device. We present a modeling and simulation inductors and capacitors. It is measured in Volt-Ampere
of STATCOM and SVC in Matlab/Simulink dynamic voltage reactive [4]. Advances in power electronic technologies
and reactive power control. It is found that these controllers together with sophisticated electronic control methods made
behave better than simple shunt capacitors. These FACTS possible the development of fast static compensators namely
provide better behavior in terms of loss reduction; voltage profile Flexible AC Transmission Systems. The FACTS technology
and reactive power compensation but STATCOM have a much has become one of the most valuable compensation
better performance as compared to SVC. techniques, because it applies the latest advances in power
electronics to achieve additional and more effective control of
Keywords— Comparison; FACTS, STATCOM; SVC; Control; the parameters of the electrical systems. This represents the
Voltage stability; Power compensation. most efficient combination of conventional primary
equipment, high power semiconductor devices,
microelectronics and telecommunications equipment,
I. INTRODUCTION allowing a most flexible power electric system [5]. The role of
The Modern electric power utilities are facing many FACTS devices such as SVC and STATCOM in system
challenges problems due to ever increasing demand and performance improvement is specified: stability improvement
voltage is attending instability by increases in loss and various of network system, power swings damping, voltage
power quality problems [1]. In recent years, the increase in regulation, increase of power transmission and chiefly as a
peak load demand and power transfers between utilities has supplier of controllable reactive power to accelerate voltage
elevated concerns about system voltage security. Voltage recovery after fault occurrence, are considered as
collapse has been deemed responsible for several major improvement factors [6]. In this paper a comparison of
disturbances and significant research efforts are under way in performances between SVC and STATCOM controllers are
an effort to further understand voltage phenomena. Voltage presented. The objective is to compare the merits and demerits
stability is the ability of a power system to maintain adequate of two FACTS devices, namely, SVC and STATCOM. A
voltage magnitude so that when the system nominal load is large portion of this work is concentrated on the study of
increased, the actual power transferred to that load will voltage stability and reactive power compensation.
increase. The main cause of voltage instability is the inability
of the power system to meet the demand for reactive power. II. DESCRIPTION OF STATCOM AND SVC
Voltage instability is the cause of system voltage collapse, in One of the major causes of voltage instability is the
which the system voltage decays to a level from which it is reactive power limit of the system. Improving the system's
unable to recover. Voltage collapse may lead to partial or full reactive power handling capacity via Flexible AC
power interruption in the system. There are two types of transmission System (FACTS) devices is a remedy for
voltage stability based on simulation time; static voltage prevention of voltage instability [2].
stability and dynamic voltage stability. Static analysis
involves only the solution of algebraic equations and therefore A. Static Synchrounous Compensator (STATCOM)
is computationally less extensive than dynamic analysis. Static The STATCOM is a main member of the FACTS family
voltage stability is ideal for the bulk of studies in which a of power electronic based controllers. It has been studied for
voltage stability limit for many pre-contingency and post many years and probably the most widely used FACTS device
Page 57
Copyright - 2019
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
in today’s power system [7]. The STATCOM is shunt STATCOM is a generator whose capacitive or inductive
connected reactive power compensation device that is capable output current can be controlled independent of the AC
of generating and or absorbing reactive power and in which system voltage. It is a solid state switching converter capable
the output can be varied to control the specific parameters of to generate or absorb reactive power at its output terminals.
an electric power system [8]. The STATCOM is made up of a The steady state V-I characteristics of the STATCOM are
shunt transformer, a voltage source converter (VSC), a DC shown in Fig. 2: [7], [10].
capacitor and a controller. If there is no energy storage device
coupled to the DC link and the losses are neglected, neither
shunt converter is capable of absorbing or generating real
power so that only operating in the reactive domain is possible.
The reactive power exchange of STATCOM with the AC
system is controlled by regulating the output voltage
amplitude of VSC. If the amplitude is increased above that of
the AC system, the current flows through the shunt
transformer from the STATCOM to the AC system, and the
device generates reactive power (capacitive). If the amplitude
is decreased to a level below that of the AC system, then the
current flows from the AC system to STATCOM. The
capacitor is used to maintain DC voltage to the VSC which
itself keeps the capacitor charged to the required levels. Thus,
by controlling the VSC output voltage lead or lag with respect Fig. 2. Terminal characteristic of STATCOM
to the AC system voltage, the capacitor DC voltage can be
decreased or increased, respectively, to control the reactive B. Static Var Compensator (SVC)
power output of the device. The structure of a STATCOM
with phase and voltage controls is depicted in Fig 1 [9]. The Static Var Compensator (SVC), is a first generation
FACTS controller, it is a variable impedance device
where the current controlled using back to back connected
thyristor valves. The application of thyristor valve technology
to SVC is an offshoot of the developments in HVDC
technology. SVC provides an excellent source of rapidly
controllable reactive shunt compensation for dynamic
voltage control through its utilization of high-speed
thyristor switching/controlled devices. Typical TSC-TCR type
SVC configuration is shown in Fig 3.
Hence SVC can supply dynamic reactive power support for A. General evaluation of STATCOM and SVC
maintaining the constant voltage which will enhance the Evaluation of STATCOM and SVC devices is shown in
generator and voltage stability margins tremendously. table I below [4]:
2
V
Qc = (2) TABLE I. Evaluation of STATCOM and SVC
Xc
Dependency of
The SVC steady-state response will follow the voltage- the current
current (V-I) characteristic curve shown in Fig. 4. The output from the
voltage-current characteristic is used to illustrate the SVC voltage level in Non-Linear Linear
rating and steady-state performance with the typical steady the point of
state operating region [11-14]. connection
Dependency of
the reactive
power output
from the Linear Quadratic
voltage level in
the point of
connection
Reactive Power
generation
during
the three phase Moderate Low
short circuit
scenario
Overload Yes (±25% for 1s)
capability No
Maintenance
and
service Low High
requirements
Fig.4. Steady state volt-current (V-I) characteristic of a SVC
from $40 to $60
Cost $55 to $70 per kVar per kVar
B. Comparison of VI and VQ curves in STATCOM and SVC transmission line can be compensated by STATCOM and
Considering the curves which relate voltage magnitude to SVC [16-17].
current (VI) or reactive power (VQ) for the aim of voltage
support capabilities is common. A decrement in system load
level results in an increase in voltage magnitude at all system
nodes. Both the SVC and the STATCOM holds the voltage
magnitude by absorbing inductive current. On the other hand,
an increase in the system load level produces a decrease in
nodal voltage magnitudes. For this condition, the devices
maintain the voltage magnitude by injecting a capacitive
current. In Fig 5, there are the VI and VQ curves for the
STATCOM and the SVC.
1.05
1.04
1.03
Voltage reference [pu]
1.02
1.01
0.99
0.98
0.97
Fig. 5. (a) VI and (b) VQ curves for STATCOM and SVC 0.96
Time [sec]
1
0.8
Reactive power reference [pu]
0.6
0.2
-0.2
-0.4
voltages levels the STATCOM can continue to supply the full -0.6
-0.8
current is independent of the system voltage, whereas the Fig. 8. Dynamic references at the (STATCOM/SVC) bus
compensating current of the SVC decreases linearly with the
system voltage [15].
C. Simulation model of STATCOM and SVC 1.05
1
Vm (pu)
0.9
middle point of the transmission line. In that test model two 0.85
0.75
0.65
STATCOM
SVC
0.2
-0.2
-0.6
-1 STATCOM
SVC
by manually. After creating a fault the disturbances in Fig. 9. STATCOM and SVC dynamic responses (loads without fault)
1.02
1.015
Vabc [pu]
1.005
0.985
value indicates the reactive power generated. The two Figures 0.45
0.4
certify that the STATCOM absorbs reactive power when the 0.35
0.3
Iabc [pu]
AC system voltage rises and generates reactive power when
0.25
0.2
0.1
0.05
decreases, at that time SVC act as a capacitive and vice-versa. Fig. 13. Voltage and current in STATCOM (fault in phase A)
Vm (pu)
1.1
1.05
1 950
Voltage measured [pu]
945
0.95
940
0.9
935
P [MW]
0.85
930
0.8 925
STATCOM
SVC
0.75 920
Time [sec]
Qm (pu) 915
0.5
910
Time [sec]
Reactive power mesuread [pu]
0 -30
-40
-0.5 -50
Q [MVar]
-60
-1
-70
STATCOM
-80
SVC
-1.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-90
Time [sec]
Fig. 10. STATCOM and SVC dynamic responses (fault in phase A) -100
0 0.1 0.2 0.3 0.4 0.5
Time [sec]
0.6 0.7 0.8 0.9 1
1.1
1
Voltage measured [pu]
0.9
1.1
0.8
1
0.7
0.9
0.6
STATCOM 0.8
SVC
Vabc [pu]
0.5
0.7
Time [sec]
Qm (pu)
0.5 0.6
0.5
Reactive power mesurad [pu]
0 0.4
Time [sec]
-0.5
1.4
1.2
-1
1
STATCOM
SVC 0.8
Iabc [pu]
-1.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time [sec] 0.6
0.4
Fig. 11. STATCOM and SVC dynamic responses (fault in phases A, B) 0.2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time [sec]
1.1
0.8
0.7
0.6
0.5 1000
0.4 STATCOM
900
SVC
Time [sec]
Qm (pu) 800
0.6
P [MW]
0.4
700
Reactive power measured [pu]
0.2
600
0
-0.2
500
-0.4
-0.6 400
Time [sec]
-0.8
-20
-1 STATCOM
SVC -40
-1.2
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Time [sec] -60
-100
-120
-140
provides a better voltage profile compared to SVC. The best Fig. 16. P and Q in STATCOM (fault in phases A and B)
location for reactive power compensation as far as
improvement of static voltage stability margin is concerned.
Vm Vref (pu)
1.5 1.025
Vm
Vref
1.02
1
1.015
1.01
0
1.005
-0.5
1
-1 0.995
Time [sec] Time [sec]
Qm Qref (pu)
0.4
Qmes
0.35
1.2 Qref
0.3
1 0.25
0.15
0.6
0.1
0.4 0.05
0
0.2
-0.05
-0.1
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Time [sec] Time [sec]
Fig. 17. Voltage and current in STATCOM (fault in phases A, B and C) Fig. 20. STATCOM Voltage and reactive power (fault in phases A, B and C)
1000 (p )
1.1
900
1.05
800
1
700
500
0.85
400
0.8
300
0.75
200
0.7
100 STATCOM
0.65
SVC
0
Time [sec]
Time [sec]
Qm (pu)
20 0.5
-40
Q [MVar]
-60
-0.5
-80
-100
-1
-120
STATCOM
-140 SVC
-1.5
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Time [sec] Time [sec]
Fig. 18. P and Q in STATCOM (fault in phases A, B and C) Fig. 21. STATCOM and SVC dynamic responses (fault in phase A, B and C)
Figures 13 to 18, determine the response of the voltage, One of the major causes of voltage instability is the reactive
current, active and reactive power system and determine the power limit of the electric system. Improving the system's
stability of the network. Three studies of power network are reactive power handling capacity via STATCOM and SVC is
modeled for transient stability to obtain the operating a remedy for prevention of voltage instability and hence
performance of STATCOM during the disturbance with fault voltage collapse. Figures 20 and 21, compares the shunt
in phase A, in phases B and C and in the three phases. It is compensator, STATCOM and SVC in static voltage and
found that these controllers significantly increase the reactive power stability improvement in network presented in
load ability margin of power system. And it also shows Fig. 19 (expanded test network of the grid presented in Fig. 6
that STATCOM has the ability to provide more capacitive and Fig.7). Performance results are compared under different
power during fault situation over SVC. Hence, STATCOM operating system conditions for the second network test
exhibits faster response than SVC. system.
Fig. 19. Simulink model of (STATCOM/SVC) for voltage and reactive power regulation
IV. CONCLUSION [13] M.A. Kamarposhti and M. Alinezhad, “Comparison of SVC and
STATCOM in static voltage stability margin enhancement”,
The FACTS technology has become one of the most International Journal of Electrical and Electronics Engineering, Vol. 4,
valuable compensation techniques. A comparison study of Issue. 5, 2010, pp. 323-328.
STATCOM and SVC in static voltage stability improvement [14] B. R. Rajeev and G.N. Madhu, “Voltage enhancement assessment using
is presented. Various merits and demerits of the shunt SVC FACTS controller”, International Journal of Electrical,
Electronics and Data Communication, Vol. 2, Issue. 10, 2014, pp.
compensation devices are presented. STATCOM and SVC 40-45.
increase the static voltage stability margin and power transfer [15] E.B. Martínez and C.Á. Camacho, “Technical comparison of FACTS
capability. But these controllers are expensive when compared controllers in parallel connection”, Journal of Applied Research and
to the shunt capacitor. Simulations using Matlab/Simulink are Technology, Vol. 15, 2017, pp. 36-44.
carried out to verify the performance of the proposed [16] P. Kumar, A.S. Mane, P.P. Patil, R.M. Hatkar, N.P. Pangam and S.V.
Sawant, “Application of FACTS device via STATCOM and SVC with
controllers. We can see that STATCOM is better compared comparison”, Proceedings of National Conference on Emerging
with that the conventional SVC. The ability to provide more Technologies in Electrical and Electronics Engineering under TEQIPII,
reactive power during fault situations is one of the most February 2014, pp. 1-5.
important advantages of the STATCOM over the SVC, which [17] N.B. Kadandani and Y.A. Maiwada, “Simulation of static synchronous
helps for faster recovery of the system during faults situations. compensator (STATCOM) for voltage profile improvement”,
Innovative Systems Design and Engineering, Vol.6, No.7, 2015, pp. 1-8.
REFERENCES
[1] S.O.Farees, M.T.L. Gayatri and K. Sumanth, “Performance comparison
between SVC and STATCOM for reactive power compensation by
using fuzzy logic controller”, International Journal of Innovative
Technology and Reasearch, Vol. 2, Issue 3, 2014, pp. 991-994.
[2] M.A. Kamarposhti and M. Alinezhad, “Comparison of SVC and
STATCOM in static voltage stability margin enhancement”, World
Academy of Science, Engineering and Technology, International
Journal of Electrical and Computer Engineering, Vol.3, No.2, 2009,
pp. 297-302.
[3] A.S.Yome and N. Mithulananthan, “Comparison of shunt capacitor,
SVC and STATCOM in static voltage stability margin enhancement”,
International Journal of Electrical Engineering Education, Vol. 41,
N°. 2, 2004, pp. 158-171.
[4] A.S. Sengar, R. Chhajer, G. Fandi and F.O. Igbinovia, “Comparison of
the operational theory and features of SVC and STATCOM”, Poster
2015, Prague, May 2014, pp. 1-6.
[5] E.B. Martinez and C.A. Camacho, “Technical comparison of FACTS
controllers in parallel connection”, Journal of Applied Research and
Technology, Vol. 15, 2017, pp. 36-44.
[6] M. Sedighzadeh, A. Rezazadeh and M. Parayandeh, “Comparison of
SVC and STATCOM impacts on wind farm stability connected to
power system”, International Journal of Engineering and Applied
Sciences (IJEAS), Vol.2, Issue 2, 2010, pp.13-22.
[7] A. Bhargava, B. Das and V. Pant, “Load flow study of large power
systems with multiple STATCOMs”, National Power System
Conference, NPSC 2014, Indian Institue of Technology, Madras, 2014,
pp. 31-35.
[8] C. Nandi, S. Deb, M.D. Barma, and A.K. Chakraborty, “Study and
simulation of the SVC and STATCOM effect on voltage collapse and
critical fault clearing time”, International Journal of Modeling and
Optimization, Vol. 2, No. 4, 2012, pp. 462-465.
[9] E.K. Abharian, M. Karimi, and P. Farhadi, “STATCOM controller
design based on MLP for power flow control”, International Journal of
Modeling and Optimization, Vol. 1, No. 4, 2011, pp. 327-333.
[10] M. Laouer, A. Mekkaoui and M. Younes, “STATCOM and capacitor
banks in a fixed-speed wind farm”, The International Conference on
Technologies and Materials for Renewable Energy, Environment and
Sustainability, TMREES14, Published in Energy Procedia, Vol. 50,
2014, pp. 882-892.
[11] V. Yarlagadda, B.V.S. Ram and K.R.M. Rao, “Testing and control of
TSC-TCR type Static Var Compensator (SVC) using microcontroller”,
International Journal of Control and Automation, Vol. 5, N°. 3, 2012,
pp. 277-286.
[12] M. Biswas and K. Das, “Voltage level improving by using Static VAR
Compensator SVC”, Global Journal of Researches in Engineering,
Vol. 11, Issue 5, 2011, pp. 13-18.
Copyright - 2019
Page 63
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Abstract— Grid-connected photovoltaic technology has A natural idea to attenuate the effect of these noises is to
gained an increasing interest worldwide as a clean source of remove them before measured signals are used in the
energy. Such system has one major drawback which is the controller. This idea has been realized in PID controllers by
harmonics it generates. Therefore, optimal control is of utmost filtering methods where low-pass filters are employed to
importance for power quality improvement and low harmonic filter out the high frequency component of noise [12].
injection. The most widely used controllers are the
Proportional-Integral controllers based on synchronous In this paper, the filter based approach is extended to the
reference frame. These controllers are affected by the signal PI controllers used in photovoltaic systems. The effect of
noises and interferences which can generate undesired control filtering techniques is investigated over the Total Harmonic
activity. This in turns lowers the power quality. Current Distortion with Fast Fourier Transformation (FFT).
The proposed strategy focuses on the major input signals of
This paper proposes a simple harmonic reduction strategy the PI controller which are the set-point from the DC-link
based on signal filtering. A first order filter is implemented on voltage, the measured current and the grid voltage in feed-
the main inputs of the controller; which are the set-point, the forward. Simulations through Simulink-Matlab are carried
feedback current and the feed-forward voltage. An overview of out to assess the impact on harmonics generation.
the system’s main components is presented. Then, the validity
of the proposed method is proved through various simulations II. SYSTEM'S STRUCTURE
carried out on Simulink-Matlab. The Fast Fourier
Transformation is used for power quality assessment. Photovoltaic is the technology that converts solar energy
into electricity [13]. The system under consideration
The results obtained clearly show the validity of the represents a two-stage topology of the grid-tied PV system.
proposed strategy. The PV array converts solar irradiation into electrical energy.
The DC-DC boost converter with a Maximum Power Point
Keywords—FFT, PI Controller, Grid-tied PV system, Filter, Tracker (MPPT) optimizes the PV output. A DC-link
Harmonic reduction (key words) capacitor enhances the DC output voltage stability, and
therefore, reduces the effect of fluctuation on the AC output
I. INTRODUCTION [14]. A two-level three-phase inverter transfers the generated
Nowadays renewable energy sources become more and PV power from the boost converter to the main grid through
more important due to energy costs and environmental issues a passive filter [15].
[1]. Currently, special attention is given to the photovoltaic
Fig 1 depicts a typical transformer-less PV system.
technology. It has grown more than 40% per year since last
decades and it is expected to experience larger growth in the A. Photovoltaic array
future due to cost reduction and increasing efficiencies [2 3]. A PV array is mainly constituted of PV cells, which
Power converters play a vital role in photovoltaic power converts the sunlight falling on it directly into electricity by
generation systems [4]. An inverter based interface is usually photovoltaic effect [16]. The simplest equivalent circuit of a
implemented for grid connection [5 6]. Their control is often solar cell is a current source in parallel with a diode. The
based on an outer active-reactive power control loop and an output of the current source is directly proportional to the
inner Current control loop. A Phase-Locked Loop (PLL) light falling on the cell [17].
technique ensures the grid synchronization [7]. The The single diode model of PV cell is shown in fig 2.
Synchronous Reference Frame (SRF) based Proportional-
Integral (PI) controller is commonly used for the inverters’
current control [1]. It is the most popular type of controller
due to its ability to decouple the active and reactive
components of the current [8].
One serious power quality issue in photovoltaic Systems
is the harmonics it generates. Consequently, providing a high
quality sinusoidal output with minimal distortion is a major
requirement for the controllers [9]. However, Most of the Fig. 1. Grid connected PV system
measured converters’ signals are affected by switching
noises and other disturbances which can generate undesired
control activity. The performances are then reduced and the
harmonics are worsen [10 11].
Vd Vpv +Ipv Rs
Ish = =
Rsh Rsh
Vpv +(Ipv Rs )
Id = Isat [exp ( ) − 1] Fig. 4. Flowchart of P&O algorithm
A UT
The expression of the current supplied by the cell Ipv is C. DC/AC- Inverter
then defined by: An inverter is a power electronic device, which generates
a sinusoidal AC output with controllable frequency and
Vpv +(Ipv Rs ) Vpv +(Ipv Rs )
Ipv = IPH − Isat [exp (
A UT
) − 1] −
Rsh
magnitude from a DC input. The inverter in the grid-tied PV
system acts as an interface between energy sources and the
Knowing that : utility grid [14]. In our work a voltage source inverter (VSI)
is used, its general structure is shown in Fig 5.
Isat : Diode Current (A);
Iph : The photo-generated current; The power circuit of the three-phase inverter consists of
Rs: The series resistor; six bidirectional switches to connect the three-phase grid
Rsh : The shunt resistor; through a passive filter. The two switches of each inverter
UT: The thermal potential; leg must operate in a complementary mode to avoid the short
A : Ideality factor; circuit of the DC-Link [15].
B. DC/DC – Boost converter The switch is controlled by a PWM generator in order to
obtain the same waveform as the reference signal. The PWM
The power output of the Solar PV module varies with signal is obtained by comparing a triangular signal (carrier)
solar irradiation and solar cell temperature [16]. to the reference signal obtained from the controller as shown
To extract the maximum power from the panel, a DC-DC in Fig 6 [20].
converter is used. It's the interface that allows adaptation
between the PV panel and the load or the grid [19]. The
topology of a DC-DC boost converter is shown in Fig 3.
The most commonly used MPPT technique is the Perturb
and Observe algorithm (P&O), due to its ease of
implementation in its basic form [17]. Its flow chart is given
in Fig 4.
Cf = 0.05 Cb
Lg = r Li
Fig. 18. Transient response of the DC-link controller with Dc-link voltage
excessively filtered Fig. 20. FFT of the system with set-point filtering
Fig. 19. FFT of the system with DC-link voltage filtering Fig. 21. FFT of the system with input filtering
the controller’s output. From the simulation results it can be [12] Y. Zhu, B. Zhu, H. Hong-Tao Liu and K. Qin, "An Estimator-Based
seen that filtering the current signal is not suitable, possibly Feedback Control Approach for Measurement Noise Rejection," 2018
IEEE 14th International Conference on Control and Automation
because of the delay induced by the filers. (ICCA), Anchorage, AK, 2018, pp. 167-172.
doi: 10.1109/ICCA.2018.8444317.
We have shown that the system is affected by the noise
injected through the input signals. Filtering these signals is [13] L. Atik et al, "Maximum power point tracking algorithm based on
sliding mode and fuzzy logic for photovoltaic sources under variable
found to be a simple yet very effective way to reduce the environmental conditions", 2017. Available: 10.1063/1.4976282.
harmonics generated by photovoltaic systems. [14] M. Shayestegan, “Overview of Grid-Connected Two-Stage
Furthermore, future work would focus on optimization Transformer-Less Inverter Design” Journal of Modern Power Systems
and Clean Energy, vol. 6, no. 4, 29 Jan. 2018, pp. 642–655,
techniques to design the filters. In addition, more tests on 10.1007/s40565-017-0367-z.
complex power systems would give insights about the [15] A. Kihal, F.krim, B.Talbi, A. Laib, A.Sahli, “A Robust Control of
controller’s behavior with the proposed strategy Two-Stage Grid-Tied PV Systems Employing Integral Sliding Mode
implemented. Theory.” Energies, vol. 11, no. 10, 17 Oct. 2018, p. 2791,
10.3390/en11102791.
REFERENCES [16] A.Ahlawat, D. Gupta, S.K. Gupta, “Modeling of a pv array
[1] E. Avci, M. Ucar, “PR Control of grid-connected 3-level inverter with and implementation of an efficient MPPT based control
LCL filter”, V.International energy technologies conference, mechanism in stand-alone photovoltaic systems”, I-Manager’s
ENTECH’17, Istanbul, 2017. Journal on Circuits and Systems, vol. 5, no. 2, 2017, p. 51,
10.26634/jcir.5.2.13665.
[2] H. Khalfalla, S. Ethni, M. Al-Greer, V. Pickert, and M. Armstrong,
"An adaptive proportional resonant controller for single phase PV [17] N. Derbel and Q. Zhu, “Modeling, Identification and Control Methods
grid connected inverter based on band-pass filter technique," in in Renewable Energy Systems.” edited by. Green Energy and
Compatibility, Power Electronics and Power Engineering (CPE- Technology, Singapore, Springer Singapore, 2019.
POWERENG), 2017 11th IEEE International Conference on, 2017, [18] M. Fares, L. Atik, G. Bachir and M. Aillerie, "Photovoltaic panels
pp. 436-441. characterization and experimental testing", Energy Procedia, vol. 119,
[3] S. Alsadi and T. Khatib, "Photovoltaic Power Systems Optimization pp. 945-952, 2017. Available: 10.1016/j.egypro.2017.07.127.
Research Status: A Review of Criteria, Constrains, Models, [19] L. Atik, et al, "Comparison of Four MPPT Techniques for PV
Techniques, and Software Tools", Applied Sciences, vol. 8, no. 10, p. Systems", AIP Conference Proceedings 1758, 030047, 2016; doi:
1761, 2018. Available: 10.3390/app8101761. 10.1063/1.4959443
[4] K Sekar and V Duraisamy’s “Efficient Energy Management System [20] T. Z. Touhami, P. Petit, G. Bachir, M. Aillerie, “New Topology of
for Integrated Renewable Power Generation Systems” JSIR Vol 74, Photovoltaic Microinverter Based on Boost Converter”, Energy
June 2015, pp. 325-329 Procedia, vol. 119, July 2017, pp. 938–944,
[5] B. Kim, J. Cho, J. Kim, J. Lee and C. Won, "An enhanced control 10.1016/j.egypro.2017.07.106.
strategy for seamless transfer of grid-connected single-phase inverter [21] A.E.Kahlane, L.Hassaine, M.Kherchi, "LCL filter design for
in synchronous rotating frame," 2017 20th International Conference photovoltaic grid connected systems, Revue des Energies
on Electrical Machines and Systems (ICEMS), Sydney, NSW, 2017, Renouvelables SIENR’14 Ghardaïa, pp 227-232, 2014
pp.1-6. doi: 10.1109/ICEMS.2017.8056258 [22] J. Lettl, J.Bauer, L.Linhart, "Comparison of different filter types for
[6] M. Parvez, M. F. M. Elias, N. A. Rahim, "Performance analysis of PR grid connected inverter". Progress in Electromagnetics Research
current controller for single-phase inverters", 4th IET Clean Energy Symposium, pp 1426-1429, 2011.
and Technology Conference (CEAT 2016), pp. 1-8, 2016. [23] M. Chabakata M. Petit, F. Costa, R. Marouani, A. Mami, "Optimized
[7] Z. Ali, N. Christofides, L. Hadjidemetriou, E. Kyriakides, "An Design of an LCL Filter for Grid Connected Photovoltaic System and
Advanced Current Controller with Reduced Complexity and Analysis of the Impact of Neighbors’ Consumption on the System.
Improved Performance under Abnormal Grid Conditions" in Journal of Electrical Systems, 13, pp 618-632, 2017.
Proceedings IEEE PowerTech, Manchester, pp. 1-6, 2017. [24] Hägglund, Tore. “Signal Filtering in PID Control, ” IFAC
[8] M. W. Alzahlan, K. M. Alawasa and H. D. Al-Majali, "Performance Proceedings Volumes, vol. 45, no. 3, 2012, pp. 1–10,
Evaluation of Different Optimal-Tuned Current Controllers for 10.3182/20120328-3-it-3014.00002.
Voltage-Source Converter Connected to a Weak AC Grid," 2019 [25] F.Blaabjerg, R.Teodorescu, M.Liserre, A.V.Timbus, “Overview of
IEEE Jordan International Joint Conference on Electrical Engineering Control and Grid Synchronization for Distributed Power Generation
and Information Technology (JEEIT), Amman, Jordan, 2019, pp. 198- Systems”, IEEE Transactions on Industrial Electronics, vol. 53, no. 5,
203. doi: 10.1109/JEEIT.2019.8717376. Oct. 2006, pp. 1398–1409, 10.1109/tie.2006.881997.
[9] D. Zammit, C. Spiteri Staines, M. Apap, J. Licari, "“Design of PR [26] L.Hassaine,E.OLias, J.Quintero, V.Salas, “Overview of Power
current control with selective harmonic compensators using Inverter Topologies and Control Structures for Grid Connected
matlab", J. Elect. Syst. Inf. Technol., vol. 4, no. 3, pp. 347-358, Jan. Photovoltaic Systems.” Renewable and Sustainable Energy Reviews,
2017. vol. 30, Feb. 2014, pp. 796–807, 10.1016/j.rser.2013.11.005.
[10] S. Haitao, M.J. Mnati, M.N. Ibrahim, A. Van der Bossche, “A Tuning [27] S. Kim, “Electric motor control : DC, AC, and BLDC motors”,
Method for PID Controller for an Integrating System with Time Chapter 4 Modeling of alternating current motors and reference frame
Delay”, MATEC Web of Conferences, vol. 249, 2018, p. 03007, theory, Elsevier, pp.153-202, 2017.
10.1051/matecconf/201824903007. [28] A.M. Farayola, A.N. Hasan, A. Aili, “Comparison of Modified
[11] V.R. Segovia, T. Hägglund, K.J. Aström, “Measurement Noise Incremental Conductance and Fuzzy Logic MPPT Algorithm Using
Filtering for PID Controllers”, Journal of Process Control, vol. 24, no. Modified CUK Converter”, 8th International Renewable Energy
4, Apr. 2014, pp. 299–313, 10.1016/j.jprocont.2014.01.017. Congress (IREC), Mar. 2017, 10.1109/irec.2017.7926029.
B. The method :
containing the UFO with agitation conserved. To ensure a into two layers, biodiesel and impurities, and then we will
good homogenization of the mixture, magnetic stirrer is set at
decant two solutions (see Figure 3.3).
stirring level 3 and at a temperature of 50 ° C.
Table 2: Results of the influence of the mass ratio MeOH / FUO for KOH and
NaOH, ρ (biodiesel volumic mass) (g/ cm3), ρ'(glycerin volumic mass) (g/
cm3),ŋ (reaction yield)
1 2 3 4 5 6 7 8
KOH NaOH
Catalyst
/oil 1/4 1/5 1/6 1/7 1/4 1/5 1/6 1/7
Ratio
ρ
0.86 0.89 0.87 0.84 0.86 0.86 0.87 0.84
ρ'
0.13 0.86 2.55 0.85 0.13 0.85 2.55 0.85
ŋ
89 92 91 92 61 76 52 43
Copyright - 2019
Page 72
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
which leads to the best yields, indeed, for KOH ratios are
conserved between 89 and 92 which is a very good yield
compared to the yields obtained for in the case of NaOH used
as catalyst, ratios in this case are between 43 and 76. Our
results, reveals that even if we change the solvent/oil ration it
is the KOH catalyst which remain the better for getting a good
biodiesel yield.
Copyright - 2019
Page 73
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
REFERENCES
The obtained results show that the time reaction of two hours
is the better for the two catalysts. However, the KOH catalyst
gives the best reaction yield of 97% compared with the NaOH
with the reaction yield of 85%.
Conclusion:
In this work we studies the catalyst effect on the used oil
transesterification to produce biodiesel, the used frying oils
react with alcohol (ethanol/methanol) in different mass ratio
(MeOH / UFO or EtOH/UFO).
Our obtained results show that the use of KOH as a catalyst
gives a very good biodiesel yield regardless of the type of the
mass ratio and the temperature time. Indeed, the maximum
obtained yield of 97% of biodiesel is generated with a mass
ratio MeOH/ HFU of 1/7 and 1% by mass KOH with a
reaction time of 2 hours. We note also that the 95% biodiesel
Abstract— This paper focuses on the detection and isolation of faults quality and can cause potential secondary faults in other
that may appear on three-phase inverters. We are particularly components [11]. Since the protective equipment intervenes
interested on the stator currents acquisition to develop decision only at the last stage of defect, it is obvious that the investment
algorithms based on d-q Park’s vector components, for the in the field of the detection of malfunctions seems an
monitoring of latent faults that shorten the life of the installations
unavoidable solution. Currently, a large number of research is
and damage them in the short term. This Fault Detection and
Isolation (FDI) method, allows the detection and localization of oriented towards monitoring the state of static converters [12,
IGBT semiconductor open circuit fault using only motor phase 13]. An open switch fault diagnosis in three phase inverter using
currents. The drive chain modeling (Power Inverter-Machine) and diagnostic variable method is proposed by Mala R. Ubale and
the monitoring algorithm are developed under Matlab / Simulink all [14], which they used average current of three phase inverter
environment where the results are presented and analyzed. The for the detection of single as well as multiple faulty switches. In
simulation results obtained the effectiveness and validity of the [15], diagnosis methods for IGBT (Insulated Gate Bipolar
considered algorithm. Transistor) open switch fault applied to 3-phase AC/DC PWM
converter are given. Indeed, these methods are based on the
Keywords- fault detection, fault isolation, modeling, FDI, modification of established fault diagnostic methods for DC/AC
diagnosis, AC-DC, Converter inverters. The types of faults that can happen in converters and
a comparative study among different converters used in wind
I. INTRODUCTION energy conversion systems are presented in [16].
Faults detection and localization are two essential tasks in The work presented in this paper is based on the detection
the industrial field. They can detect a fault in an early manner and isolation of open switch faults of semiconductors (IGBT).
to avoid unexpected installation shutdowns, this, to ensure a The paper is organized as follows. Section 2, describes the
quick and timely maintenance avoiding expenses. The most modeling of drive chain. The principles of FDI approach will be
used methods are those based on the mathematical models. presented in section 3. In Section Conclusion, computer
These use the redundancy existing between the different simulation results obtained under Matlab/Simulink are shown
variables measured in terms of static or dynamic relations. Fig.1 and discussed. Finally, the reported work is concluded.
presents the structural diagram of 3-phase PWM inverter
induction motor drive. As shown, the induction machine (IM)
in the studied system is connected directly to the 3-phase PWM
inverter to allow an efficient control of IM. The converter is
𝑇11 𝑇21 𝑇31
controlled by Pulse Width Modulation (PWM). According to
the research that has been done on conversion system failures,
Power supply
power converters and its control circuits have a high failure rate 𝑣𝑎
Data measurement
Power supply
𝑣𝑎
𝑇𝑒𝑚
𝑣𝑏 IM
𝑣𝑐
𝑎𝑏𝑐
𝑖𝑑
𝑖𝑎𝑏𝑐 𝑖𝑑𝑞
𝑑𝑞 𝑖𝑞
𝑆1 𝑆2 𝑆3
Control algorithm
Using d-q transformation, we can detect open circuit IV. SIMULATION RESULTS
switches faulty situation of inverter. A normal operation is Fig. 3 presents the current waveform of three-phase current
represented by a circle centered at the origin. If an open switch 𝑖𝑎 , 𝑖𝑏 𝑎𝑛𝑑 𝑖𝑐 with switch open circuit of T 5, applied at t=1s. We
fault has occurred, there is a change in the phase current value can observed that when the fault is applied, two phases vary in
at the location of fault. Indeed, a change in the circle shape the same direction and the third varies in the opposite direction.
illustrates the occurrence of an open fault condition. Figure 2 Then, we deduce that there is a fault.
represents the structure of the diagnosis block diagram method
that we have developed in this work.
150 1000
ia ib ic
100
500
50
Currents phases (A)
0
id
-500
-50
-1000
-100
-150 -1500
0.9 0.95 1 1.05 1.1 1.15 -1500 -1000 -500 0 500 1000
Time (s) iq
Fig. 3. Three phase trajectories under open switch fault (switch T22) Fig. 4. d-q current trajectories ‘without fault’
T11 T12
. 2000 1000
1500
500
1000
0
500
id
id
-500
0
-1000
-500
-1000 -1500
-1500 -1000 -500 0 500 1000 1500 2000 2500 -1500 -1000 -500 0 500 1000
iq iq
Fig. 5. d-q current trajectories under open switch fault ‘switch T11’ Fig. 6. d-q current trajectories under open switch fault ‘switch T12’
T21 T22
1500 1500
1000 1000
500 500
0 0
id
id
-500 -500
-1000 -1000
-1500 -1500
-1600 -1400 -1200 -1000 -800 -600 -400 -200 0 200 -200 0 200 400 600 800 1000 1200 1400
iq iq
Fig. 7. d-q current trajectories under open switch fault ‘switch T21’ Fig. 8. d-q current trajectories under open switch fault ‘switch T22’
T31 T32
500 1500
0 1000
-500 500
d
id
-1000 0
-1500 -500
-1500 -1000 -500 0 500 1000 1500 2000 -2000 -1500 -1000 -500 0 500 1000
iq iq
Fig. 9. d-q current trajectories under open switch fault ‘switch T31’ Fig. 10. d-q current trajectories under open switch fault ‘switch T32’
REFERENCE
[1] Ribrant J,BertlingL.Survey of failures in wind power systems with focus on [10] Yang, S., Xiang, D., Bryant, A., et al.: ‘Condition monitoring for device
swedish wind power plants during1997–2005. IEEE Trans Energy Convers, vol. reliability in power electronic converters – a review’, IEEE Trans. Power
22, pp. 167-173, 2007. Electron., vol. 25, pp. 2734–2752, 2010.
[2]Hahn. B, Durstewitz. M, Rohrig. K. Reliability of wind turbines. In : Wind [11] Hongshan Zhao, Liangliang Cheng, Open-circuit faults diagnosis in back-
energy.Berlin,Heidelberg:Springer, pp. 329-32. ISBN978-3-540-33865-9, to-back converters of DF wind turbine, IET Renew. Power Gener., vol. 11, pp.
2007. 417-424, 2017.
[3] Lin Y, TuL, LiuH, LiW. Fault analysis of wind turbines in china. Renew [12] Lee, J.S., Lee, K.B., Blaabjerg, F.: ‘Open-switch fault detection method of
Sustain EnergyRev, vol. 55, pp. 482–90, 2016. a back-to-back converter using NPC topology for wind turbine systems’, IEEE
[4] FreireNMA, Estima JO, Marques Cardoso AJ. Open-circuit fault diagnosis Trans. Ind. Appl., vol. 51, pp. 325–335, 2015.
in PMSG drives for wind turbine applications. IEEE Trans Ind Electron, vol. [13] Jlassi, I., Estima, J.O., EI Khil, S.K., et al.: ‘Multiple open-circuit faults
60, pp. 3957–67, 2013. diagnosis in back-to-back converters of PMSG drives for wind turbine systems’,
[5] Freire N, Estima J, Cardoso A. Converters fault-diagnosis in PMSG drives IEEE Trans. Power Electron., vol. 30, pp. 2689–2702, 2015.
for wind turbine applications. In : Proceedings of the IECON 2010 – 36th annual [14] M. R. Ubale, R. B. Dhumale, S. D. Lokhande, Open switch fault diagnosis
conference on IEEE industrial electronics society, pp. 403-408, 2010. in three phase inverter using diagnostic variable method, International Journal
[6] Fuchs F. Some diagnosis methods for voltage source inverters in variable of Research in Engineering and Technology, vol. 02, 2013, http://www.ijret.org.
speed drives with induction machines-asurvey. In : Proceedings of the 29th [15] Won-Sang Im, J.S Kim, J.M. Kim, D.C. Lee, and K.B. Lee, Diagnosis
annual conference of the IEEE on industrial electronics society (IECON'03); Methods for IGBT Open Switch Fault Applied to 3-Phase AC/DC PWM
vol. 2,2003.p.1378–85. Converter Journal of Power Electronics, vol. 12, No. 1, 2012,
[7] H. Merabet, T. Bahi, K. Bedoud, and D. Drici, Real-Time Switches Fault http://dx.doi.org/10.6113/JPE.2012.12.1.120, pp. 120-127.
Diagnosis for Voltage Source Inverter Driven Induction Motor Drive, Int. J. [16] Blaabjerg, F., Ma, K.: ‘Future on power electronics for wind turbine
Elec. & Elecn. Eng. & Telcomm. vol. 8, n o. 2, pp. 103-107, 2019. doi: systems’, IEEE J. Emerging Sel. Top. Power Electron., vol. 1, pp. 139–152,
10.18178/ijeetc.8.2.103-107 2013.
[8] Sethom, H.B.A., Ghedamsi, M.A.: ‘Intermittent misfiring default detection [17] K. Bedoud, T. Bahi, S. Vaidyanathan and H. Merabet, Control of Matrix
and localization on a PWM inverter using wavelet decomposition’, J. Electr. Converter Fed Induction Motor Drive, vol.10, No.2, pp 881-890, 2017.
Syst., vol. 4, pp. 1–12, 2008.
[9] Yang, S., Bryant, A., Mawby, P., et al.: ‘An industry-based survey of
reliability in power electronic converters’, IEEE Trans. Ind. Appl., vol. 47, pp.
1441–1451, 2011.
Copyright - 2019
Page 79
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
1
Centre de Recherche Scientifique et Technique en Analyses Physico-chimiques (CRAPC), BP 384-Bou-Ismail-RP 42004,
Tipaza-Algérie.
1
ftouahra@gmail.com
6
bachari2000@yahoo.fr
2
Laboratory of Natural Gas Chemistry, Faculty of Chemistry (USTHB), BP 32 16111 Algiers, Algeria.
2
balifer@gmail.com
7
dhalliche@yahoo.fr
3
Department de Chimie, Faculty of sciences, University of M’hamed Bougara, Independence Avenue, 35000 Boumerdès,
Algeria.
3
djebaribaya@gmail.com
5
zoulisabrina@yahoo.fr
4
Département de Chimie, Faculté des Sciences (UMMTO), Tizi-ouzou, Algérie
4
n_aider@yahoo.fr
Abstract— The NiAl-HDL and NiAlMn-HDL hydrotalcite-like NiAlMn-LDH catalyst drived form hydrotalcite for dry
compounds, were successfully synthesized following co- reforming of methane.
precipitation method at pH = 12. A several characterization
techniques including XRD, N2 adsorption and desorption, H2- II. EXPERIMENTAL
TPR, SEM-EDX and TEM, were utilized to determine the
structure function relationship for the obtained catalysts. These
A. Catalysts preparation
catalysts were evaluated in the reforming of methane reaction The NiAl-LDH and NiAlMn-LDH hydrotalcite-derived
under continuous flow with CH4/CO2 ration equal to 1, at catalysts were prepared via co-precipitation method with
atmospheric pressure and a temperature 700 °C. The Mn
molar ratio (nNi2+/nAl3+) = 2, in order to induce the formation
addition to the NiAl-LDH showed improved resistance to coke
deposition.
of hydrotalcite structures. The aqueous solutions of nitrate
metals were added dropwise to a vigorously stirring solution
Keywords— Syngas, Hydotalcite, Mn, Co-precipitation, Methane
of NaOH (2M) at room temperature while pH was maintained
dry reforming.
at 12. The obtained slurry was heated at 80 °C, kept under
I. INTRODUCTION stirring for 15 hours for maturation, and then filtered, washed
with water and finally dried at 100 °C in oven overnight. The
Dry Reforming of Methane (DRM, CH4+ CO2 2CO + physico-chemical properties of the synthesized materials were
2H2), is nowadays considered as a perspective alternative for characterized using various techniques such as ICP, XRD,
the production of syngas (H2 + CO) involving the valorization BET, TPR-H2, SEM-EDX, TEM and TPO-O2.
of CO2, coming either from capture or naturally present in the
reactant gas, i.e. such in biogas. Furthermore, the H2/CO B. Catalytic activity
molar ratio in DRM-syngas, equal to 1, makes it applicable
Catalytic testing experiments were carried out in a continuous
for Fischer–Tropsch synthesis [1]. One of the main
flow system at atmospheric pressure using a fixed-bed tubular
drawbacks confronted during the process of CO 2 reforming of
quartz reactor. Prior to catalytic reaction, 100 mg of each
methane, is the poor stability of the catalyst over prolonged
sample was reduced in-situ under constant hydrogen flow at
period of times, this is mainly associated with carbon
700 °C for one hour. After that, the temperature was cooled
deposited on the catalyst’s Surface [2]. Several studies have
down to the initial reaction's temperature and a feed gas
been proposed to limit the formation of coke such as, the
mixture containing CH4:CO2: Ar in a ratio of 20:20:60 was
incorporation of the active phase in well-defined structures
passed through. The total gas flow rate was set to 20 mL/min.
such as perovskites, hydrotalcites, etc., maybe on stream or
The effluent gases were analyzed using a gas chromatograph
(pre)reduction, resulting in a higher dispersion of the active
(Delsi), equipped with a thermal conductivity detector (TCD).
phase species which leads to a good compromise between
stability and catalytic activity. In this context, the purpose of
the present work was to prepare efficient NiAl-LDH and
III. RESULTS AND DISCUSSION Table 1 shows the specific surface area values (SBET) of the
NiAl-LDH and NiAlMn-LDH solids before calcination and
A. X-ray diffraction analysis after calcination at 800 °C.
TABLE I
Fig. 1A shows the X-ray diffractograms of uncalcined
SPECIFIC SURFACES OF THE NiAl-LDH AMD NiAlMn-LDH
samples (NiAl-LDH and NiAlMn-LDH). The corresponding
X-ray diffractograms for NiAl-LDH, and NiMnAl-LDH are Samples BET Surfaces area (m2 /g)
consistent with those expected for hydrotalcite-like with Precursors Calcined Reduced Used
sharp and symmetric reflections for (003), (006), (110) and
(113) planes and broad asymmetric for (012), (015) and (018) NiAl-LDH 79 105 98 75
[JCPDS file 15-0087]. The results of XRD experiments for
the calcined samples (at 800°C) are presented in Fig. 2B. As NiAlMn-LDH 90 121 114 101
can be seen, after thermal treatment, the hydrotalcite structure
of the samples was destroyed. On the NiAl-cal sample, XRD Analyses of the obtained data revealed that the surface areas
patterns showed the formation of NiO species [JCPDS file of two calcined solids are relatively high. We also note that
47-1049] and NiAl2O4 spinel [JCPDS file 10-0339], an SBET of two materials have increased after calcination (Table
amorphous aluminum oxide phase should be also formed 1) associated with the removal of water, nitrate and carbonate
(not detected in XRD)[3]. The Mn-containing samples from the interlayer space of the hydrotalcite structures.
exhibited peaks corresponding to NiO, NiAl2O4 and Mn3O4
[JCPDS file 24-0734]. C. Temperature programmed reduction (TPR)
The TPR profiles obtained for all the samples are plotted in
Fig. 2. The TPR of NiAl-cal sample showed a profile with a
(A)
main peak at 530 °C, which can correspond to the reduction
(003) of free or NiO weakly interacting with Al2O3. A broad peak
was observed around 677 °C; it is well known that the species
Intensity (a.u)
NiAlMn-cal
470
H2- consomation (u. a)
780
NiAl-cal 650
NiMnAl-cal 530
677
10 20 30 40 50 60 70 80
2 Theta (degrees) NiAl-cal
Fig. 5 TEM for the reduced samples for the reduced samples (a) NiAl-cal
and (b) NiMnAl-cal.
F. Catalytic activity tests
Intensity (a.u)
NiAlMn-cal
The catalytic performances of the synthesized catalysts were
evaluated at 700°C in methane dry reforming reaction, after
an ‘‘in situ’’ H2 reduction at 750 °C for 1 h. The catalytic
performances and selectivity results are presented in Fig. 6,
NiAl-cal
including both conversions of CH4 and CO2, and H2/CO
molar ratio. As can be seen, the Mn-based catalysts showed
10 20 30 40 50 60 70 80
2Theta (degree)
higher catalytic activity than NiAl-cal. The conversion of
CH4 increased from 82 % NiAl-cal to 86 % NiAlMn-cal.
Fig. 3 XRD of reduced catalysts at 750oC. Mn3O4, NiO, Ni0, MnO whereas CO2 conversion increased from 86 % NiAl-cal to
et NiAl2O4.
88 % in the case of NiAlMn-cal. This result is due to the
incorporation of Mn in Ni-based catalyst which plays a good
D. BET surface of reduced samples role in promoting a synergistic effect that further stabilizes
Evaluation of the specific surface area of all catalysts after the catalyst. In addition, previous work by our group also
reduction was also studied. Indeed, data shown in Table 1 elucidated the possibility of deposited carbon being oxidized
demonstrated that the reduction step lead to a small decrease in the presence of manganese oxide (Mn3O4) to release
in the specific surfaces areas. carbon monoxide CO:
[4 ]. This was confirmed by the presence of Mn3O4 and MnO
E. SEM & TEM analyses after reduction species prior to and after reaction, respectively.
The morphology of NiAl-cal and NiAlMn-cal after reduction
was analyzed by SEM-EDX and TEM (Fig. 4 and Fig. 5, 100
(A)
90
respectively). Following reduction, catalysts demonstrated the
80
presence of smaller and evenly distributed Ni particles on the 70
CH4 Conversion (%)
surface of the catalysts. The Ni0 particle have been confirmed 60 NiAl-cal
NiAlMn-cal
by the EDX analysis. 50
40
30
20
10
0
0 5 10 15 20 25 30 35 40 45 50
Time (h)
100
(B)
90
80
70
CO Conversion (%)
60
NiAl-cal
50 NiAlMn-cal
40
30
2
20
10
0
0 5 10 15 20 25 30 35 40 45 50
Fig. 4 SEM-EDX for the reduced samples (a) NiAl-cal and (b) NiMnAl-cal. Time (h)
1.10
(C)
1.05 IV. CONCLUSIONS
1.00 The solids NiAl-LDH and NiAlMn-LDH have been prepared
0.95 by co-precipitation method. The synthesized catalysts were
H /CO ratio
NiAlMn-cal
0.80
catalyst exhibited high activity and selectivity, but also high
0.75 stability and good resistance to carbon deposition; this could
0.70 be related to the capacity of manganese oxide in oxidizing
0 5 10 15 20 25 30 35 40 45 50
Time (h) carbon deposition forming CO as by-product.
Fig. 6 (A) CH4 conversion, (B) CO2 conversion and (C) H2/ CO ratio
obtained during DRM.
ACKNOWLEDGMENT
G. Temperature programmed oxidation TPO-O2 The authors would like to thank Centre de Recherche
Scientifique et Technique en Analyses Physico-Chimiques
TPO-O2 curves of two catalysts after reaction are shown in (CRAPC), Algeria, for the financial support.
Fig. 7. The TPO-O2 profile of the spent catalyst shows the
presence of a broad peak between 450°C -700 °C in the case
of NiAlMn-cal catalyst. On the other hand, TPO of NiAl-cal REFERENCES
sample showed a broad peak with high intensity around
[1] A.S. Al-Fatesh, M.A. Naeem, A.H. Fakeeha and A.E. Abasaeed, Bull.
600 °C. The only oxidation peak appearing in the TPO-O2 Chem. Soc. Jpn., vol. 86, pp. 742, 2013.
profile of the spent catalysts was attributed to the oxidation of [2] M. Usman, W.W. Daud and H.F. Abbas, Renew. Sust. Energ. Rev., vol.
one type of carbonaceous species (filamentous carbon) on the 45, pp. 710, 2015.
surface of both samples [6]. As can be seen, the addition of [3] B. Djebarri, V. Gonzalez-Delacruz, D. Halliche, K. Bachari, A. Saadi, A.
Caballero, J.P. Holgado and O. Cherifi, React. Kinet., Mech. Cat., vol.111,
Mn within NiAl-LDH, prompted less carbon deposition pp. 259-275, 2014.
during the reaction of CO2 reforming of methane. [4] F. Touahra, M. Sehailia, D. Halliche, K. Bachari, A. Saadi and O. Cherifi,
Int. J. Hydrogen Energy, vol. 41, pp. 21140-21156, 2016.
[5] P. Arnoldy and J.A. Moulijn, J. catal., vol. 93, pp. 38-54, 1985.
[6] M. Rezaei, Iranian J. Hydrogen Fuel Cell., vol. 2, pp. 215-226, 2016.
TCD signal (a.u.)
NiAl-cal
NiAlMn-cal
100 200 300 400 500 600 700 800 900 1000
Temperature (°C)
Fig. 7 TPO profiles of catalysts after DRM reaction.
Abstract— The objective of this work is the design of a surface (DDPMG). The authors concluded that all of the geared
permanent magnet synchronous generator (SPMSG) with a generators have high losses in the gearbox which reduce the
large number of poles to be used in the conversion of wind energy yield, DDSG is the heaviest and most expensive
energy. The design of the SPMSG is performed by three steps. alternative, and that there is important copper losses in the
The first step is the geometrical design which is performed
stator and the rotor. DDPMG does not have a gearbox and
analytically according to the imposed technical specifications
and constraints. The second step is the verification and the electrical excitations which reduce the losses and improve the
examination of the electromagnetic behavior of the designed energy yield.
SPMSG using a finite elements method under Ansys Maxwell The authors in [5] compared the embedded magnets with
RMXpert software. Finally, with the third step, the main the surface mounted magnet machine. The study showed that
performances of the generator connected to an infinite bus are higher torque is achieved with surface mounted magnet
investigated and the steady state parameters of the SPMSG are machine and it needs less magnet material compared to
also calculated. embedded magnets. Five types of rotor structures are
presented and discussed in [6] to determine the suitable rotor
Keywords— Design; SPMSG; Ansys Maxwell RMXpert; Finite structure with high power for a PMSG. And it is concluded
Elements Method; Wind Energy that the surface permanent magnet (SPM) structure with a
magnet width/pole pitch of 90% is the most suitable rotor
I. INTRODUCTION
structure for high efficiency, low voltage, and high power
Among the renewable energy sources, the wind power has output.
notably increased in the last decade [1-2] and the main two According to different comparisons, discussions and
types of electrical machines used in the conversion of wind presentations of the direct drive generators given in the
energy are the doubly fed induction generators and literature, our choice is focused on the design of a surface
synchronous generators [1], but the tendency is the use of the permanent magnet synchronous generator (SPMSG). It has
direct drive generators with a better efficiency due to the various advantages, such as higher efficiency, larger power
elimination of the gearbox which is prone to complications density and higher reliability which reduce operation and
such as generation of noise, regular maintenance, increased maintenance costs [7-9].
losses, and is responsible for high percentage of outage time. The direct-drive generators are designed with a large
Several types of these generators and geared drives have diameter and small pole pitch to increase the efficiency, to
been presented, discussed and compared by many authors to reduce the active material and to keep the end winding losses
address the effects of each concept. small [10], [11]. The air-gap should not exceed a few
Among the conversion technologies presented in [3], it millimetres to avoid excessive magnetization requirements
appears that the direct-drive technology offers good [10]. A design procedure of a direct drive low speed surface
performance with respect to reliability, maintenance, energy mounted permanent magnet (SMPM) machine is presented in
extraction, and grid power quality. [12], The result showed that the higher the pole number, the
In [4], the authors compared five different generator lower the weight for the same amount of copper losses, and
systems, namely doubly-fed induction with three stages the concentrated winding, with a number of slots per pole per
(DFIG3G) and with single-stage gearbox (DFIG1G), phase q=0.5 are more advantageous than a distributed
permanent magnet generator with single-stage gearbox winding with q=1, the weight and the torque ripple are lower,
(PMSG1G), direct drive generator with electrical excitation and the torque is higher for the same amount of copper losses.
(DDSG) and direct drive permanent magnet generator
of the principal dimensions of SPMSG is performed. With: Js : The current density; K r : The stator slot fill factor;
Secondly, internal modelling concerning the electromagnetic K s is the slot proportion; Q s is the number of stator slots.
behaviour of the generator is simulated in Ansys Maxwell The height (𝑚ℎ ) and the width (𝑚𝑤 ) of the permanent magnet
RMXpert software. Finally, external modelling concerning are expressed by [18-20]:
the dynamic performance of the generator connected to an μ
mh = K c g B r r (7)
infinite bus is illustrated. Bg1
−1
Where rr1 is the ratio of the radius bore to the active length TABLE II
(L), Pn is the rated power, K w1 is the stator fundamental DESIGN RESULTS FOR THE SPMSG
winding factor, Bg1 is the fundamental air gap flux density, A1
Bore radius Rb 1.466 m
is the linear current density, Ω is the rated speed and cos φ is
the rated power factor. Active length 𝐿 0.293 m
The gap length (g), the stator yoke (ys ) and the rotor yoke Stator yoke ys 14.4 mm
(yr ) are expressed by [14], [16-17]:
Rotor yoke yr 14.4 mm
g = 0.001 + 0.003 R b L (3)
Height of slot 𝑠ℎ 56.7 mm
πR b B g 1
ys = yr = αm (4) Width of slot 𝑠w 24 mm
2p B y
Where: p : The number of pole pairs; 𝐵𝑦 : The flux density in Gap length 𝑔 3 mm
the iron core; αm : The rotor pole embrace. Height of permanent magnet 𝑚ℎ 8.7 mm
The height (𝑠ℎ ) and the width (𝑠𝑤 ) of the stator slot are given
Width of permanent magnet 𝑚𝑤 50 mm
by [14-15] and [18]:
A
sh = 1 (5)
Js K r K s
III. NUMERICAL DESIGN and lower cost of coils and the short pitching can be used to
reduce cogging and certain harmonics [16].
To verify and validate the analytical design obtained in the
section 2, it is necessary to perform a numerical design with 3.00
Ansys Maxwell RMXpert software.
The numerical design concerns the internal modeling of
B (tesla)
2.00
the SPMSG using a finite elements method to analyze its
electromagnetic behavior, and the external modeling of
SPMSG to analyze its dynamic performances. 1.00
model to small elements where equilibrium equations are Fig. 3 B-H curve of steel_1010
formulated. The combination of these equations gives us the
equation for the whole structure. The boundary conditions are
then imposed and the equations of equilibrium are solved.
In this study, a magnetodynamic analysis is considered, and
the basic equation is given by:
∂A
rot 1
μ
rot A + σ = J + rot M (9)
∂t
The flux lines path and the flux density distribution are
shown in fig.6 and fig.7, respectively. These results are
obtained at 0.4s which correspond to the position of the rotor
113.054 degrees, and at the rated speed 46.875 rpm.
ANSOFT
1000.00
Curve Info
Phase Current ia
Phase Current ib
500.00
Phase Current ic
(Amperes)
0.00
-500.00
-1000.00
0.00 125.00 250.00 375.00
Electric Degree
Fig. 9 Winding currents under load
Fig. 7 Flux density distribution 1.00
ANSOFT
Curve Info
Air-Gap Flux Density
C. Performance Analysis
The SPMSG is directly driven by the wind turbine torque 0.50
(mechanical input energy) which is converted into electrical
(Tesla)
output energy [22-23], and then delivered to the infinite bus. 0.00
The system is simulated in Ansoft Maxwell 2D RMXpert
software under the conditions given in the table 3.
-0.50
TABLE III
CONDITIONS OF THE ANALYSIS
Curve Info
Rated speed 46.875 rpm Air-Gap Power
8.00E+005
Operating temperature 75°C
6.00E+005
The main performances of SPMSG obtained in the Ansoft
(W)
0.00
-500.00
-1000.00
0.00 125.00 250.00 375.00
Electric Degree
TABLE V
FULL LOAD CHARACTERISTICS OF THE SPMSG
The results show that, the voltage and the current winding
Iron core loss 11.507e-5 kW
have the waveform with a root mean square (RMS) values of
400V for the phase voltage and 582 A for the current phase as
shown in fig.8 and fig. 9 respectively. The variation of the air Total loss 31.679 kW
gap flux density versus the electric degree is illustrated in the
fig. 10, it has a wave form with a peak value of 0.81 T. As Efficiency 95.421 %
shown in fig. 11, the air gap power versus power angle is
presented with a rated value of 660 kW which is reached at Rated torque 141 kN.m
47.11 deg of the power angle. The transient plots of the
moving torque and the current winding are illustrated in fig.
12 and fig. 13 respectively. After the transient period, the TABLE VI
current phase reaches its root mean square (RMS) value 582 A STEADY STATE PARAMETERS OF THE SPMSG
and the moving torque reaches its average 141 kN.m. D-Axis reactive inductance Lad 0.618 mH
Moreover, the no-load and the full load characteristics and
the steady-state parameters of the SPMSG are also calculated Q-Axis reactive inductance Laq 0.618 mH
in Maxwell RMXpert software. They are summarized in table
4, table 5 and table 6, respectively. D-Axis inductance L1+Lad 1.965 mH
TABLE IV
NO LOAD CHARACTERISTICS OF THE SPMSG Q-Axis inductance L1+Laq 1.965 mH
Stator teeth flux density 1.49 T
Armature leakage inductance L1 1.347 mH
Stator yoke flux density 1.51 T
Armature phase resistance R1 0.031 Ohm
Rotor yoke flux density 1.51 T
D and Q axis time constant 0.019 s
Air gap flux density 0.81 T
proves the effectiveness and the feasibility of the designed International Conference on Power Engineering, Energy and Electrical
Drives, Istanbul, 2013, pp. 760-763.
generator for the wind turbine.
[14] S. Belakehal ―Conception & Commande des Machines à Aimants
For the future work, it is desirable to comfort the obtained Permanents Dédiées aux Energies Renouvelables », Thèse de Doctorat
results in this paper with the experimental by realizing a en sciences, Electrotechnique, Université de Constantine, Algérie
prototype of the designed SPMSG. 2010.
[15] T. Wu, T. Camarano, J. Zumberge, M. Wolff, E. S. Lin, H. Huang, X.
REFERENCES Jia,‖Electromagnetic Design of Aircraft Synchronous Generator with
High Power-Density », 50th AIAA Aerospace Sciences Meeting
[1] T. P. M. Bazzo, J. F. Kolzer, R. Carlson, F. Wurtz, L. Gerbaud. including the New Horizons Forum and Aerospace Exposition, 09 - 12
―Multidisciplinary design optimization of direct-drive PMSG January 2012, Nashville, Tennessee.
considering the site wind profile‖. Electric Power Systems Research. [16] A. abdelli « Optimisation multicritère d’une chaîne éolienne passive »,
Volume 141. Pp 467-475, 2016. Thèse de Doctorat en génie électrique, Institue Polytechnique de
[2] T. P. M. Bazzo, J. F. Kolzer, R. Carlson, F. Wurtz, L. Gerbaud, Toulouse (INPT), France 2007.
« Multidisciplinary Design Optimization of a Direct-Drive PMSG [17] M. Belatel, H. Benalla « Etude par CAO d’une machine synchrone à
including the Power Converter Cost and Losses », IEEE, pp 1909-1914, aimant permanent pour la production d’énergie éolienne », Revue des
January 2016. Energies Renouvelables Vol. 11 N°2 (2008) 167 – 180.
[3] M. R. Dubois., H. Polinder, J. A. Fereira ―Generator Topologies for [18] S. Mouty « Conception de machines à aimants permanents à haute
Direct Drive Wind Turbines and adapted Technology for Turbines densité de couple pour les éoliennes de forte puissance », Thèse de
Running in Cold Climate‖, proceeding of the conference on wind Doctorat en génie électrique, Université de Franche – Comté, Ecole
energy in cold climate, Matane, Canada, 2001. doctorale Sciences pour l’ingénieur et microtechniques, France 2013.
[4] H. Polinder, F.F.A. van der Pijl, G.J. de Vilder, P. Tavner, "Comparison [19] L. Drouen, F. Scuiller, J.F. Charpentier, E. Smail, S. Clenet, « Modèle
of direct-drive and geared generator concepts for wind turbines", IEEE analytique de pré-dimensionnement de machines lentes à aimants de
Trans. Energy Conversion, Vol. 21, pp. 725-733, September 2006. grand diamètre et a grand nombre de poles », Conférence EF 2007,
[5] P. Salminen, J. Pyrhönen, M. Niemelä ―II-17. a comparison between ENSEEIHT, Toulouse, 6-7 Septembre 2007.
surface magnets and embedded magnets in fractional slot wound pm [20] S. Arslan, E. Kurt, O. Akizu, J. M. Lopez-Guede ―Design optimization
motors‖, ISEF 2003 – 11th International Symposium on study of a torus type axial flux machine‖, Journal of Energy Systems,
Electromagnetic Fields in Electrical Engineering Maribor, Slovenia, 2(2), 2018, pp. 43-56.
September 18-20, 2003. [21] B. Kundrotas, A. Petrovas, R. Rinkeviciene, A. Smilgevicius ―
[6] H. Haraguchi, S. Morimoto and M. Sanada, "Suitable design of a Research of Six-Phase Induction Motor Windings‖, Electronics and
PMSG for a small-scale wind power generator," 2009 International Electrical Engineering. Vol. 20, 2014, PP. 93-96.
Conference on Electrical Machines and Systems, Tokyo, 2009, pp. 1- [22] T. Z. Htet, Z. Zhao, Q. Gu, J. Li, ―Modelling and Analysis of Radial
6. Flux Surface Mounted Direct-Driven PMSG in Small Scale Wind
[7] P. Xu, K. Shi, Y. Sun, H. Zhua, « Effect of pole number and slot Turbine‖, Advances in Science, Technology and Engineering Systems
number on performance of dual rotor permanent magnet wind power Journal Vol. 2, No. 6, 94-99 (2017).
generator using ferrite magnets », AIP Advances 7, 056631, January [23] N. Fazli and J. Siahbalaee, "Direct torque control of a wind energy
2017. conversion system with permanent magnet synchronous generator and
[8] A. Hebala, W. A. M. Ghoneim, H. A. Ashour, ―Different Design Matrix Converter," 2017 8th Power Electronics, Drive Systems &
Approaches of Surface Mounted High Performance PMSG », IEEE, Technologies Conference (PEDSTC), Mashhad, 2017, pp. 166-171.
41458, ACCS’017& PEIT’017, Alexandria, Egypt 2017.
[9] K. Sindhya, A. Manninen, K. Miettinen, J. Pippuri, ―Design of a
Permanent Magnet Synchronous Generator using Interactive
Multiobjective Optimization », Transactions on Industrial Electronics,
DOI 10.1109/TIE.2017.2708038, IEEE.
[10] Msc. Zoran Ivanovski ‖ Direct - Drive Wind Turbines‖ International
Journal of Scientific & Engineering Research Volume 2, Issue 10,
Oct-2011, ISSN 2229-5518
[11] D. Bang, H. Polinder, G. Shrestha and J. A. Ferreira, "Promising
Direct-Drive Generator System for Large Wind Turbines," 2008 Wind
Power to the Grid - EPE Wind Energy Chapter 1st Seminar, Delft,
2008, pp. 1-10.
[12] F. Libert and J. Soulard, ―Design Study of Different Direct-Driven
Permanent-Magnet Motors for a Low Speed Application‖, in
Proceedings of the Nordic Workshop on Power and Industrial
Electronics (NORpie), Trondheim, Norway, June 2004.
[13] A. Lebsir, A. Bentounsi, R. Rebbah, S. Belakehal and M. E. H.
Benbouzid, "Comparative study of PMSM and SRM capabilities," 4th
2.5. Determination of the enzymatic activity the recovery of the enzyme. In fact, [8] obtained a clotting
activity for the crude extract of 2.43 U mL-1 for a
The milk clotting activity expressed as rennet unit concentration of 5% NaCl, whereas [11] obtained a clotting
(RU) was determined according to [7]. Briefly, 1 mL of crude activity of 7.83 U mL-1 for a concentration of 10% NaCl.
extract or fractions was added to 10 mL of Berridge substrate These results suggest that the appropriate method for the
(12 g low heat spray dried skim milk powder in 100 mL of extraction of coagulating enzymes greatly depends on the
0.01 M CaCl2), mixed thoroughly and incubated at 30 °C. The NaCl concentration.
time required for the first appearance of the particles was
discernible recorded. One unit of milk clotting activity was The purification was achieved by flocculation of
defined as the amount of enzyme required to clot 10 mL of clotting materials from the crude extract with ammonium
milk in 100 s at 30 °C. sulfate at 50% saturation, as the fraction precipitated at 25%
saturation of ammonium sulfate presented no clotting activity,
10 𝑉 and as the fraction precipitated at 75% showed only a low
= (1)
𝑇𝑐 𝑄
activity of 0.1 RU. In the present work, the extract precipitated
in 50% ammonium sulfate displayed a clotting activity of ca.
Where:
3.2 RU and specific activity of 1.5 RU mg-1 higher than the
initial specific activity of the crude extract of 1.2 RU mg -1
RU: Rennet Unit; V: volume of standard substrate (mL);
(purification factor of 1.26 and purification yield of 29%). [12]
Q: volume (mL) of crude gastric extract (CGE) or
obtained a clotting fraction precipitated by ammonium sulfate
chromatographic fraction; T c: clotting time (s).
between 40 and 60% saturation from the viscera of bolti fish
(Tilapia nilotica), with a specific activity of 0.25 U mg-1
III. RESULTS AND DISCUSSION
protein, a purification factor of 4.16, and a purification yield
of 42%. Similarly, [13] have prepared a clotting enzyme
Poultry viscera, especially turkey proventriculus, are
extract by ammonium sulfate precipitation between 20 to 60%
a source of milk coagulant proteases, currently under-
saturation from sea bream (Sparus latus Houttuyn) , with a
exploited and easy to obtain throughout the year and require a
specific activity of 3.0 U mg-1 protein, a purification factor of
better valuation in the cheese industry, unlike ruminant rennet,
1.4, and a purification yield of 64%.
the availability of which is lower and the cost is greater.
The chromatographic profile obtained by cation-
Pepsin is one of the animal origin that has been tested for
exchange FPLC (Fig. 1) showed 4 peaks, and clotting activity
potential use as substitutes of rennet [8]
was found in the collected fractions 12–16 eluted at 0.4–0.5 M
In this study, turkey proventriculus was used to purify milk NaCl and corresponding to materials the most strongly
clotting enzyme, the purification procedure developed in this adsorbed on the column (fourth peak).
study resulted in an active band with apparent molecular mass The active fractions were analyzed by SDS-PAGE (Fig. 1) and
of 36.5 kDa that is consistent with the molecular mass a single band with an apparent molecular mass of 36.5 kDa
reported by [9]. Both animal and microbial milk clotting was unveiled, and agreed well with the results of [14], who
proteases belong to a class of acid aspartate proteases and have found a same apparent molecular mass of 36 kDa for a clotting
molecular masses between 30000 to 40000 Da. However, the enzyme partially purified.
purification of the enzyme using ammonium sulfate is
IV. CONCLUSIONS
recommended for the enzyme preparation with excellent milk
clotting properties. The presence of one active band of purified
In the present study, a fraction of enzyme was
enzyme might be indicative of monomer protease. The
purified for the first time from proventriculus of turkey. A
specific activity in the crude extract increases from 1.2 to 4.6
fraction highly enriched in clotting enzyme was recovered
after FPLC purification; this is explained by the elimination of
with high purity by chromatography methods and turkey
the inactive proteins in the crude extract.
enzyme can advantageously replace calf rennet as it displayed
3.1. Purification of turkey’s milk clotting enzyme a milk clotting activity. Its apparent molecular mass was 36.5
kDa. Future investigations will focus on the molecular
Purification of the clotting protease from characterization of the coagulating enzyme.
proventriculus of turkey is summarized on Table 1. The
extraction of crude extract was performed by maceration of
the proventriculus in 5% (w/v) NaCl and displayed a clotting
activity of 11.1 RU.
Table 1 REFERENCES
Summary of purification steps of pepsin-like enzyme of turkey (Meleagris
[1] M.M. Grozdanovic, L.Burazer and M Gavrovic-Jankulovic,
gallopavo) Kiwifruit (Actinidia deliciosa) extract shows potential as a low-cost and
efficient milk-clotting agent. International Dairy Journal, 32, 46–52,
Purification Clotting Amount Specific Purific Purificati
2013.
step activity of total Activity ation on Yield [2] A.Kumar, J. Sharma, A Kumar Mohanty, S.Grover and V. Kumar Batish,
Purification and characterization of milk clotting enzyme from goat
(RU) protein (RU mg- Factor (%)
(Capra hircus). Comparative Biochemistry and Physiology Part B:
(mg mL- 1
) Biochemistry and Molecular Biology, 145, 108–113, 2006.
1 [3] M. Beigomi, M.A Mohammadifar, M Hashemi, M.G. Rohani, K Senthil
)
and M Valizadeh, Biochemical and Rheological Characterization of a
Crude 11.11 9.2 1.2 1 100 Protease from Fruits of Withania coagulans with a Milk-clotting activity.
Food Sci. Biotechnol, 23, 1805–1813, 2014.
extract [4] J. Tsouli, Etude comparée de l'activité enzymatique de 3 variétés
d'artichauts du genre Cinara Cardunculus L sur la coagulation du lait. Le
Ammonium Lait, INRA Editions, 54, 415–421, 1974.
sulfate 3.22 2.11 1.52 1.26 28.98 [5] U.K. Laemmli, Cleavage of structural proteins during the assembly of the
head of bacteriophage T4. Nature, 277, 680–685, 1970.
(50%) [6] M.M. Bradford, A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye
Cation-exchange FPLC binding. Anal Biochem, 72, 248–254, 1976.
Fraction 12 2.22 1.14 1.94 1.27 68.94 [7] N.J. Berridge, The purification and crystallization of rennin. Biochem J,
39, 179–187, 1945.
Fraction 13 2.38 0.95 2.5 1.28 107.2 [8] F.A. Benyahia-Krid, F.Adoui, O. Aissaoui-Zitoun, H. Boughellout, E-H.
Siar, A. Zikiou, and M.N. Zidoune, Effect of Sodium Chloride and
Fraction 14 2.77 1.08 2.56 1.02 116.38 Incision on the Chicken Pepsin Coagulant Activity Extracted from
Proventriculus,Dried Under Partial Vacuum. Advances in Dairy
Fraction 15 2.94 1.41 2.08 0.81 106.13 Research, 5:3, 2017.
Fraction 16 2.63 0.95 2.76 1.32 89.45 [9] M.B. Rao, A.M., Tanksale, M.S Ghatge and V.V.Deshpande, Molecular
and Biotechnological Aspects of Microbial Proteases. Microbiol Mol Biol
Rev, 62, 597-635, 1998.
[10] R.G. Beka, F.Krier, M. Botquin, V.D. Guiama, P.Donn, D.G
kDa
100
M 12 13 14 15 16 Libouga, C.M Mbofung, K.Dimitrov, M.C. Slomianny, D.Guillochon,
75
50
and D. Vercaigne-Marko, Characterization of a milk-clotting
Fig1 37 extract from Balanites aegyptiaca fruit pulp. International Dairy
25 Journal, 34, 25–31. 2014.
20
[11] H.Temiz , E.Okumus, U.Aykut, M., Dervisoglu and F.Yazici, Partial
15
Purification of pepsin from turkey proventriculus. World Journal of
Microbiology and Biotechnology, 24, 1851–1855, 2008.
[12] A.E. El-Beltagy, T.A. El-Adawy, E.H. Rahma and A.A. El-Bedawey,
Purification and characterization of an acidic protease from the viscera of
bolti fish (Tilapia nilotica). Food Chemistry, 86, 33–39. 2004.
[13] Q. Zhou, X.P Fu, L.J. Zhang, W.J. Su and M.J. Cao, Purification and
characterization of sea bream (Sparus latus Houttuyn) pepsinogens and
pepsins. Food Chemistry, 103, 795–801, 2007.
[14] H. Temiz, U. Aykut, E. Okumus, and S. Turhan, The partial purification
and properties of pepsin obtained from Turkey proventriculus.
Biotechnology and Bioprocess Engineering, 12, 450–456, 2007.
[8]
2imane.idrissi@usmba.ac.ma
3tariq.riouch@gmail.com
Abstract— This paper proposes an automatic solution to control has the line voltage at its terminals; while in the case of star
voltage and frequency, in spite of load variations, for an connection, each single capacitor has a phase voltage at its
induction machine used as a generator and excited by a capacitor terminals. However, it is well-known that delta connection of
bank. Our solution consists in developing a mathematical formula the capacitor bank gives more merits than the star one. Indeed,
that allows the calculation of the necessary capacity for a given it enables the same self-excitation conditions with a three times
resistive load. This formula has been determined based on the smaller value of the capacitance [1].
single-phase equivalent model of the generator. A test rig located The major drawback of the self-excited induction generator
in our laboratory has been used to carry out experimental is poor voltage and frequency regulations [2]. The variable
achievements. First tests allow us to determine electrical
frequency depends on the load level and the variable voltage is
characteristics of the induction machine, and to deduce
equivalent model parameters. Second tests allow us to determine
produced by the excitation capacitor value, magnetization
capacitance values under no-load and resistive load conditions. characteristics, the electrical load and its power factor [3].
Furthermore, the control curves of the generator are obtained by When a driven asynchronous generator is self-excited by
modeling and simulating the studied system in the Matlab & capacitors, a voltage is produced. However, if a resistive load is
Simulink software. Experimental and simulated results are connected to the terminals of this SEIG, its voltage drops.
compared and the similarity of results has been demonstrated. Therefore, the load cannot operate under its rated voltage. So,
in order to ensure that the load operates at its rated voltage, it is
necessary to compensate the drop. For practical use, it is
Keywords— Wind Turbine, Self-Excited Induction Generator, necessary to maintain the terminal voltage constant. That issue
Voltage Control, Magnetization is known as the excitation automation of the self-excited
generator. This automation consists of maintaining SEIG
I. INTRODUCTION
voltage constant, regardless of the load variations.
Nowadays, the asynchronous machine is increasingly The values of the capacitance required to excite the SEIG
utilized as a generator in isolated power systems, as in wind were analyzed by [4]. Several methods are proposed for
turbines installed in remote areas. In these regions, the calculation of minimum capacitance value, required for self-
electricity production is known to be difficult and expensive. excitation of an isolated three-phase self-excited induction
The extensive use of the asynchronous machine over generator (SEIG). In [5], the minimum capacitance value is
conventional alternators, is justified by its enormous determined under various conditions as variable load and
advantages such as: mechanical and electrical simplicity, variable rotor speed. This developed technique uses Adaptive
rugged construction, low cost, power generation capability at Particle Swarm Optimization (APSO) technique, devoted for
varying speed, and easy maintenance. However, to operate as a the non-linear equations. For Wind turbine driven Self-excited
generator, it requires reactive power to produce magnetizing induction generators, an algorithm of fixed capacitor selection
current responsible of its excitation. is proposed by [6], which is able to increase annual energy
When the reactive power is supplied by the electrical grid, captured by the Wind Energy Conversion System (WECS).
the induction generator is called Grid-Connected Induction In this paper, we propose an automatic solution which
enables the compensation of the voltage drop to maintain
Generator; while it is named Self-Excited Induction Generator
constant the terminal voltage in spite of load variations. It is
(SEIG) or the Off-grid Induction Generator, when the reactive based on a model we will establish. We will develop a
power is provided by a capacitor bank. mathematical formula to compute the value of the necessary
For the SEIG, a capacitor bank is star or delta connected capacity for a given resistive load value.
across the stator terminals, while the rotor is driven by the wind
turbine. In the case of delta connection, each single capacitor
This work is supported by the CNRST center, Rabat, Morocco (GISER Project)
The outline of this paper is given as follows: In section 2, a This later produces a flux which gives voltage ‘bc’ across stator
mathematical model of SEIG is presented. Then, the excitation terminals. Voltage ‘bc’ generates current ‘od’ which engenders
automation of the asynchronous generator is developed in the voltage ‘de’. This voltage buildup process continues till the
section 3. Section 4 shows the implementation of the system magnetization curve of the induction generator cuts the
and experimental results. Furthermore, the simulation results, capacitor load line. Intersection point is marked as ‘f’ in the
performed in the Matlab & Simulink software, are presented in Fig. 3.
section 5 and the conclusion is given in section 6.
(2)
(5)
From the equation (5), we can clearly notice that the load
variations cause the variation of capacitance value if the
constant output voltage (380V in our case of study). However,
this solution remains valid as long as the load power factor is
unity.
Copyright - 2019
Page 95
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
To measure the rotor resistance, the same manipulation , is consumed only at the level of the circuit stator and the
scheme as before is performed with a short circuited rotor this magnetic stator. Then, the following equation can be written:
time.
(14)
(15)
(16)
Fig.9. The performed manipulation scheme for Rr measurement
Therefore, Ls= 6,17 mH
Therefore, Rr= 0.55 Ω
Rotor leakage reactance (Lr) measurement:
Rμ and Lμ measurement: This parameter was determined from the blocked rotor test
The elements of the magnetic circuit (Rμ and Lμ) are of the asynchronous machine. This test consists of star-
calculated from the synchronous test of the asynchronous coupling the machine and then feeding it using a balanced
machine. three-phase voltage. We gradually increased the value of the
When the machine is operating in the synchronous mode, supply voltage until the rotor shaft began to rotate, at which
its rotor speed N is equal to the synchronism speed Ns. point we blocked the shaft with pliers to bring the speed back to
Therefore, the slip s equals zero and we have Rr / s tends to the 0. In this test, the reactive power, noted , is consumed by
infinity limit. For this reason, we have considered that the rotor both the rotor and at the stator because the current flowing
branch is open-circuit so we determined the value of Rμ and Lμ through the magnetic circuit is negligible. This allows us to
from calculation of joules losses and iron losses. write the equation below:
- Calculation of Rμ :
(17)
(7)
Fig.15 Output voltage by the generator with the nominal resistive load
From the figure, it is clear that the new value of the output
Fig.12 Simulink scheme of the no-load test of the asynchronous generator voltage is 350.4V and the frequency f is equal to 49.218 Hz.
However, these results do not meet our objectives, which are
We have fixed the rotor speed at 1510 rpm, the powers 380V and 50 Hz.
obtained from experimental test are P = 2870 W and S = To overcome this problem, we have drawn two control
curves. These curves allow the determination of the capacitance
values which excite continuously the generator to deliver a -Resistive load test C = 77.58 μF C = 77.58 μF
fixed voltage of 380 V and a constant frequency of 50 Hz. The -Constant load
R = 710.645 Ω U = 413.7 V U = 385 V
first control curve plots the capacity variation along with the
resistive load while fixing the rotor speed N of the generator.
For instance, when N is equal to 1503 rpm, we get this curve:
From this table, we conclude that the simulation results
consolidate the experimented ones.
VI. CONCLUSION
The use of induction machine as a generator is becoming
popular in renewable energy systems such as wind turbines.
The induction generator has the possibility to operate
Fig.16 Capacitance value variation as function of resistive load for N = 1503
independently from the grid utility since the reactive power
rpm necessary needed for its excitation is provided by an external
source, which is in most of the time, a capacitor bank
For N= 1528 rpm, we have: connected to the generator stator windings. However, the major
problem of this technology is the control of the output voltage
under variable loads and variable rotor speeds. To solve this
problem, an automatic solution of voltage drop compensation
so as to maintain constant voltage under load variations, is
proposed. This method consist sin developing a mathematical
formula, which determines the necessary capacitance value for
a given resistive load. To carry out the implementation of the
induction generator excitation, a test rig composed of wound
Fig.17 Capacitance value variation as function of resistive load for N = 1528 rotor asynchronous motor driven by DC motor with separated
rpm excitation, has been used. Firstly, we have determined the
electrical parameters of the induction generator equivalent
Therefore, we can conclude that the capacitance value is model. Secondly, the excitation of the induction generator has
inversely proportional to the resistive load. been ensured by means of connecting a capacitor bank to the
The second control curve plots the capacitance variation as generator stator windings, under no-load and resistive load
function of the rotor speed of the generator while maintaining conditions. The necessary capacitance value has been
the load constant (nominal load R = 32Ω). This scenario occurs determined in the two cases with the objective to obtain a
for variable speed turbines with fixed loads. The obtained curve constant output voltage of 380 V. Then, the control curves of
is presented below: the generator obtained from experimental results are compared
to those got by modeling and simulating the studied system in
the software Matlab & Simulink. The comparison established
has demonstrate the similarity between simulated and
experimented results.
REFERENCES
[1] Levi, E., & Liao, Y. W. (2000). An experimental investigation of self-
Fig.17 Capacitance value variation as function of rotor speed of the generator excitation in capacitor excited induction generators. Electric Power
Systems Research, 53(1), 59-65.
The obtained curve is a linear regression line, which shows [2] Shah, M., & Somkun, S. (2017). Efficiency Evaluation of Three Phase
and Single Phase C2C Self-Excited Induction Generator for Micro Hydro
that when the rotor speed decreased, the capacitance value Power Application. Energy Procedia, 138, 193-198.
increased. Furthermore, when the generator is operating at the [3 ] Borkowski, D. (2016). Voltage and frequency control of A stand-alone
synchronous speed of rotation, the capacitance value reaches its induction generator using sVPWM converter in a small reservoir
maximum. hydropower plant. Czasopismo Techniczne, 2016(Elektrotechnika Zeszyt
3-E 2016), 51-63.
The following table presents the comparison between the [4] Capelo, B., Pérez-Sánchez, M., Fernandes, J. F., Ramos, H. M., López-
simulation results and those obtained from the experimental Jiménez, P. A., & Branco, P. C. (2017). Electrical behaviour of the pump
tests: working as turbine in off grid operation. Applied energy, 208, 302-311.
[5] Chatterjee, H. S., & Mahato, S. N. (2016, January). Calculation of
TABLE III. SIMULATED AND EXPERIMENTED RESULTS COMPARISON minimum capacitance of isolated three phase induction generator using
Operating conditions Simulation results Experimental results adaptive PSO technique. In 2016 International Conference on
Microelectronics, Computing and Communications (MicroCom) (pp. 1-
-No load test C = 67.63 μF C = 65.65 μF 4). IEEE.
-Constant rotor speed [6] Fernandes, J. F., Pérez-Sánchez, M., da Silva, F. F., López-Jiménez, P.
N = 1510 rpm A., Ramos, H. M., & Branco, P. C. (2019). Optimal energy efficiency of
U = 380 V U = 385 V
isolated PAT systems by SEIG excitation tuning. Energy conversion and
management, 183, 391-405.
#
Laboratoire de chimie du Gaz Naturel, Faculté de Chimie, U.S.T.H.B, BP 32, Alger, Algérie.
1 sarakila@yahoo.fr
2
second.author@second.com
Abstract—The activity and stability of NiO (4wt%) catalysts calcination and reaction by BET, atomic absorption, X-ray
supported on ZrO2 were investigated for the methane steam diffraction (XRD).
reforming between 475 and 700 °C at atmospheric pressure. The
catalysts were prepared by the impregnation method. They have
been characterized by various physical methods like atomic II. EXPERIMENTAL
absorption, BET technique, X-ray diffraction. According to the
catalyst, the activity can be due, not only to the interaction A. Catalysts preparation
nickel-support but also to the physico-chemical properties of the The solids Ni/ ZrO2 were prepared by impregnation
support. of the support ZrO2 with Ni(NO3)2,6H2O (1 M) (Merck). The
solution was stirred during 2 h, and then evaporated at 80 °C
under agitation. The solid was dried at 80 °C (12 h), calcined
Keywords— Methane, steam reforming, nickel, coke, supports during 2 h at 300 °C and 2 h at 500 °C at a heating rate of
oxides 4 °C/mn, in air flow (1.2 L h-1). The catalysts were then
sieved to have a particle diameter less than 0.16 mm.
I. INTRODUCTION
Among the advanced research and the existing B. Catalysts characterization
processes in the area of chemical conversion of natural gas, The solids were analyzed by atomic absorption thanks
the classical methane steam reforming (MSR) is an indirect to a spectrometer type Perkin-Elmer 1100 B. The specific
conversion via synthesis gas and remains extensively used [1]. surface areas were determined by the BET method using
The hydrogen is a clean energy that could play a key role as an analyzer of surface type (Coultronics 2100E). The
energetic source for fuel cells and for the environmental crystal structures were identified by XRD with a
protection [2]. diffractometer (/2 CGF) provided with a proportional
Among the catalysts based on transition elements, detector and using Kα molybdenum anticathode
used in the steam reforming of methane, nickel is the most (λ=0.70930 Å). The X-ray patterns were obtained for 2θ
studied element [3]. It has been reported that nickel presents values ranging from 20° to 70°. The apparent size of
good catalytic performances when supported on Al2O3 doped nickel oxide (nickel) particles was calculated by the
by MgO and CaO [4]. Mixed oxides such as Al2O3-ZrO2, Scherrer formula, L = 0.9λ/βcos θ, where β is the width of
prepared by sol-gel method, or by grafting zirconium on the the most intense NiO peak at half-height, and θ is the
surface of the alumina. corresponding Bragg angle.
The major problem with nickel based catalysts, is
carbon deposition to form coke which deactivates the C. Reactivity
catalysts. To remedy this inconvenient, some authors The catalytic tests were carried out in a quartz
examined the use of the nickel promoted by an alkaline oxide reactor fixed bed (L= 65 cm, = 1 cm) under atmospheric
[5] in sweeter conditions. pressure. The catalyst (0.1 g) is pre-reduced at 500 °C
The aim of the present work is to compare the activity, under hydrogen flow (1.2 L h-1) before introduction of the
the selectivity and the carbon formation of two series of gas mixture (H2O/CH4 = 3.3) diluted in argon flow and
nickel based catalyst (4 wt. %) supported on ZrO2 during the obtained by crossing of a mixture CH4 (10 %)/Ar, in a
MSR reaction. The catalysts have been characterized after saturator with a thermostated bath (65 °C , 1.2 L h -1). The
reactants and the products (CO, CO2 and CH4) were After calcination at 500 °C, the XRD pattern of 4 wt%
analyzed by gas chromatography thanks to TCD and after MSR reaction, the presence of Ni° metallic species
chromatograph (Hewlett-Packard 5730) on a carbosieve B and that of support are visible in the patterns of Ni/ZrO2
column, 100-200 mesh, of 2 m of length, using hydrogen (Table 1), The absence of peaks corresponding to carbon in
as vector gas. presence of Ni/ZrO2 catalyst could be due to an amorphous
Before each analysis, the reactants and products pass form of carbon.
through a water-trap at 0 °C to remove water. The conversion
of CH4 and product selectivities are calculated using
following formulas: B. Methane Steam reforming reaction
n in
n out
TABLE I TABLE II
CHARACTERISTICS OF NI/SUPPORT SYSTEMS CATALYTIC ACTIVITIES OF NIO/ZRO2
100
50
4%Ni/ZrO2
700°C
Methane conversion (%)
40
80
650°C
600°C 30
60
550°C 20
17,4
Hydrogen production (%)
result is agreement with the literature, according to Nielsen [7] Temperature (°C)
who found that carbon deposition, on nickel catalyst has Fig.3 : Production of hydrogen according to the temperature on the nickel-
different temperature regimes: usually, ~ 550 °C, the based catalysts, with ratios of 4% NiO/ZrO2 at different temperatures, m=
activation energy of coke formation changes to a negative 0.1 g, Tred= 500 °C/H2/overnight, d= 1.2 L h-1, H2O/CH4= 3.3.
value.
The testing results show that catalytic performances of
Ni/ZrO2 are very sensitive to temperature from 500 to 700 °C, IV. CONCLUSIONS
and conversions varying between 22 - 82 % (Fig 2) and CO
selectivity varying between 12 - 41% (table II). We notice an The reactivity, of CH4/H2O reaction, for synthesized Ni-
important carbon deposition at 600 °C on catalyst (table II). based catalysts (4 % of NiO) supported on ZrO2 was
The best hydrogen production is at temperature between investigated at temperatures ranging from 475 to 700 °C with
650-700°C (17.5%) (fig 3). the aim of facilitating the formation of synthesis gas. It has
Moreover, we noticed no diffusional limitation (effect been shown that:
of mass, flow etc…). The experimental conditions of the The catalyst is more active. This result can be explained
reaction are chosen so that they are as optimal as possible. It by a better dispersion of the active phase for the lower
appears that the catalyst is more active. This is can be due to a charged catalysts.
better dispersion of the active phase for the weakly charged
catalyst, a result in agreement with the literature data.
REFERENCES
[1] Sanbing Zhang, Jiankang Wang, Xiaolai Wang, Journal of Natural
Gas Chemistry 2008, 17:179.
[2] Shinku Lee, Joongmyeon Bae, Sungkwang Lim, Joonguen Park, J [5] N. Martin, M. Viniegra, E. Lima, G. Espinosa, Ind. Eng. Chem. Res.,
Power Sources, 2008, 180:506. 2004, 43 : 1206.
[3] K. O. Christensen , D. Chen, R. Lødeng, A. Holmen. Appl Catal A: [6] A. Belhadi, O. Cherifi; J. Soc. Alger. Chim., 2009; 19(1), 49-62.
General, 2006, 314(1): 9. [7] J.R. Rostrup-Nielsen, in Catalysis Science and Technology, eds, J.R.
[4] Tadeusz Borowiecki, Wojciech Gac, Andrzej Denis, Appl. Catal. A: Anderson and M. Boudart (Springer–Verlag, New York), 1984, Vol.
Gen. 2004, 270:27. 5:1.
a
¹ PhD student, Laboratory ETAP:
Environment and Technology for Architecture
and Cultural Heritage, University Blida1,
Algeria; bProfessor, University of Guelma,
Algeria.
amrikhaoula12@gmail.com*corresponding author
Copyright - 2019
Page 103
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Copyright - 2019
ISSN 1737-9334 Page 104
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
The study is structured in two main parts. The 2.2. Geographical character:
geomorphological context of the region, as Algerian lower Sahara, which have long
acted as relays along an itinerary, have
limited to a single ground floor, forming a which excludes the use of wood, ashlar and
clay, only the exploited local materials are
compact and homogeneous structure. It is
used: louss (sand rose) and tafza (gypsum).
presented with blind walls, the continuity
The first is a very hard gypsum
of which is interrupted only by small
concretization, which in free form gives sand
openings placed at the top and a single
roses, but in continuous sedimentation gives a
intimate door. The first reading of the resistant slab; this material constitutes the
typical plan of the house allowed us to construction stones of walls and basements
observe that this unit takes up in its spatial .Figure.05.a.
organization the classical model of the The second is a lighter and more reliable
Maghreb house with the principle of coating which, once fired, gives a good
centrality, a courtyard in the center, plaster, once dried it constitutes a very
named (haouch) and the rooms around, resistant binder, found buried under the sand
in the form of parts or plates in the northern
called (diars). Nevertheless, there are
part of the region, this stone is removed and
some particularities, one linked to the
burned in incinerators. The development of
existence of a space particular to the
this industry to promote the creation of
region called (sabat), (Marc cote, 2005).
traditional ovens called (kusha). These burnt
Which is a relaxation space semi-open to
stones are struck with a kind of drumstick
the courtyard, the other concerns the until they become a powder. Figure.05.b.
denomination of the spaces. It also has a
(skiffa), a chicane space of great symbolic
value, used to keep the courtyard away
from outside view.Figure.4.
4. Investigation phase:
To assess the thermal performance and
examine the behaviour of dwellings in the Figure.6: The instrument used during the
face of climatic constraints, a campaign.
measurement campaign was carried out 4.2. Description of sample:
on two samples during the summer period The sample is presented by a vernacular
of 2018, precisely during the month of house, built of local materials: sand rose
August. Measurements were carried out stone and gypsum plaster. Overall, this
between 6am and 9pm using an unit retains its original appearance with a
anemometer measuring the three central courtyard and surrounding rooms,
parameters: indoor temperature "c°", 50 cm thick walls made of sand rose
relative humidity %", air velocity "m/s". stones and gypsum mortar, all covered by
[5]. as well as the outdoor temperature at low-profile vaults pierced by ventilation
street level. The two rooms chosen are: holes. Measurements were taken in the
the sabat room and the living room with a living room and in the sabat room.
one-hour interval between measurements. Figure.07.
Température Interieure
Température Extérieur
Humidité
50 50
48
46 45
Humidité (%)
40
35
38
36 30
34
32 25
30
Figure 07: view of the sample 28
20
26
15
6 8 10 12 14 16 18 20 22
Heure (H)
5. Result and discussion:
Figure.09. Evolution of the temperature and
Temperature measurements inside the
relative humidity inside the living room as
rooms:
well as the outside temperature at the street
The thermal investigation shows different level, 05 August 2018 , (Vernacular housing).
thermal behaviours during the 24 hours of The diagrams on figures .07.08. show
measurement, governed by variations in positive behaviour, especially during the
external temperatures, solar incident critical period, especially between 12:00
power and human activity. Indeed, and 16:00 hours. Measured temperature
considering that the type of thermal increases over time but it keeps a range of
transfer is carried out by convection 8.9 C° between the outside and the inside,
through the external walls as well as the because it reaches a maximum of 38.5 C°
roof. at 12.30h and 14h while outside the
maximum measured temperature is 44.9
Température Interieure
Température Extérieur
50 Humidité
50 C°. Overall, we can distinguish 3 decisive
48
46 45
periods:
Température Interieure (C°)
44
40
42
From 6am to 10am: a humidity level is
Humidité (%)
40
35
38
36 30
too high caused by the accumulation of
34
32 25 heat released during the night, (50%),
30
20
28
with the beginning of human activity and
26
15
6 8 10 12 14 16 18 20 22
the opening of windows the air begins to
Heure (H)
relative humidity inside the summer room(Sabat) slight increase in temperature inside.
as well as the outside temperature at the street From 12 noon: during this period the indoor
level, 05 August 2018,(Vernacular housing). temperature rises rapidly and reaches its peak
at 1:30 p.m. influenced by outside
temperature fluctuations and the change in
human behaviour from 12 a.m. onwards: all steady temperature inside the rooms. The
openings are closed except that the doors are shape and height of the roof also plays a
opened and protected by shades that very important role, the vaults serve to
encourage the renewal of air with the central
break the sun's rays and increase the shade
courtyard. Nevertheless, there is a slight
area by its longitudinal shape. This
difference between the two diagrams: the
reduces the intensity and amount of heat
summer chamber (Sabat) .Figure.08. which
exposure. From the inside, the hollowed
faces northwest has a uniform variation with
an interval of less than two degrees compared
shape is used to accumulate the hot air at
to the living room which faces southeast, the top to evacuate it through the
mainly due to the orientation effect and the ventilation holes. Roof ventilation has
number of walls in contact with the outside been shown to be the most effective
environment. because the roof surface is the warmest
From 4pm to 8pm: the temperature being the most exposed and therefore the
measured inside begins to gradually suction effect is more accelerated
decrease with a slight increase in humidity In addition, there are no openings that
caused by the watering of the sand lead directly to outside. All the rooms in
covering the rooms and the central the house open onto the inner courtyard,
courtyard which in turn serves regulate the
During the night: walls release the heat temperature of the rooms by the
stored during the day towards the interior ventilation ticket eventually with the
[7]and consequently the interior existence of a water source.
temperature increases, this justifies the
return of temperature increase from 22 h. 6. Conclusion
the thing that explains the nomadism of Vernacular architecture has proven its
the occupants during the night towards the performance, whatever the time and place of
yard to sleep. construction, an intuitive approach to notions
The interval of temperature between of sustainability and respect for places; it
inside and outside is mainly justified by represents a whole knowledge, acquired
the large mass and thickness of walls, through secular experience. At the end of this
which measures 45 cm, built of gypsum article, we will conclude that a set of
stone, which have good thermal information has been found around this
insulation, used to store heat throughout vernacular heritage and can be summarized as
the day and decrease its direct transfer. It follows:
insulates the interior and maintains a
Despite the natural obstacles and difficult 2.André Voisin, Le Souf monographie d’une
région saharienne, (El Walid, 1985), pp113-117.
climatic conditions, the scarcity of water
3.Jean Brunhes, Les Oasis du Souf et du M'zab
resources, the unavailability of varieties of comme types d'établissements humain. In :
materials, the need for housing and Bulletin de la société des géographes, tome V, one
Semestre, (Masson ET Cie éditeurs. Paris, 1902),
stabilization has led Soufi humain to acquire pp.7-14.
and perfect knowledge and skills. Subtle and 4. Organisme Technique de Mise en valeur des
Richesse du sous-sol Saharienne, présentation du
adapted to the local context. This acquisition Souf, (Le Porodon-Hydra. Alger. 1971), pp.02.
was manifested through different aspects: 5. Givoni B, L’homme, l’architecture et le climat,
access to water and exploration of the (Edition de Moniteur), pp.21.
1. References:
1.Cote Marc, La ville et le désert, le bas Sahara
algérien, (Karthala et Ireman. Paris, 2005), pp.141-
143.
Abstract— Over the last few years, scientists emphasized their II. MATERIAL AND METHODS
researches on biogas production harder than ever to encounter
The green filamentous macroalga was provided by a
the energy needs. In particular, biogas crowdsourced by algal
wastes became a tendency research field due to the high content treatment plant in the Refinery R1K (Skikda, Algeria). It was
of volatile solids in algae. The primary objective of the present removed from the sludge thickeners stored at room
study is to assess the biogas productivity from algal wastes and temperature and in the dark.
explore temperature and granulometry parameters effects on The brown microalgae (diatoms) were collected by brushing
biogas yield. However, the biogas was produced of freshwater some artificial substrates (glass plates of 10 cm2), immersed
macroalgae (Oedogonium sp.) and microalgae (Thalassiosira sp. in the same basin of decantation of the treatment plant. Two
and Chaetoceros sp.) wastes. It was found that algal biogas yields dominant microalgae were identified and growing in vitro in
are directly proportional to temperature and inversely fed-batch bioreactors, to obtain larger amounts for the
proportional to granulometry. Optimal conditions for biogas
production of biogas.
production from algal wastes were determined as the
temperature of 55 °C, a granulometry of 200 μm, a residence The batch reactors for anaerobic digestion experiments
time of 21 days and an alga–inoculum ratio of 1:2 (w:w). Highest were set up in 5L glass bottles (Fig.1). Two parameters were
biogas yield obtained under these conditions was found as monitored during the experiment that are Temperature and
342.59 cm3 CH4 g−1 volatile solids (VS) from Oedogonium sp., and blending provided with a heating magnetic agitator. The
287.65 cm3 CH4 g−1 VS from Thalassiosira sp. Followed by amount of biogas produced was measured volumetrically by
Chaetoceros sp. (254.37 cm3 CH4 g−1 VS). It may be concluded evacuation of the liquid.
that algal biomass is a good source for future biogas production,
albeit more studies are required to enhance biogas yield and
properties.
2
1 1- Flexible tube
2- Two valves for connection
Keywords— Biogas, algal wastes, biogas yield, temperature, 3- Glass bottle
granulometry. 4- Bicycle inner tube
3
I. INTRODUCTION
Algae that have come into prominence for biofuel
production in the last decade are considered a third generation 4
300
The main components of the produced biogas from 0.25mm
CO2% 15.91 32.05 27.31 Fig. 3 Produced biogas yields according to granulometry and residence time
CH4% 78 65.2 64.14 for different algae-inoculum wastes.
25°C
300 35°C From a technical point of view, it has been demonstrated that
45°C the production of biogas through the use of algal wastes is
200 55°C
both sustainable and ecological. However, the efficiency and
60°C
100 quality of produced biogas should be enhanced through
further studies. In this sense, biogas quality and efficiency
0
1:2 1:3 1:4 1:2 1:3 1:4 1:2 1:3 1:4
can be increased by various physical, chemical and biological
pretreatments and anaerobic co-digestion with different
Oedogonium sp. Thalassiosira sp. Chaetoceros sp.
Alga–inoculum ratio
wastes. Biogas production from the waste algal biomass
Fig. 4 Produced biogas yields according to experimental temperature should be developed in parallel to the algae bio-economic
conditions and algae–inoculum ratio for different algae wastes. market to support the economy and environment.
REFERENCES
Biogas yields (cm3 CH4 g−1 VS)
300
200 0.25mm
[1] J. Baeyens, Q. Kang, and L. Appels, “Challenges and opportunities in
0.55mm improving the production of bio-ethanol,” Progress in Energy and
100 0.75mm Combustion Science, vol. 47, pp. 60-88, 2015.
1.00mm [2] Q. Kang, L. Appels, and T. Tan, “Bioethanol from Lignocellulosic
0 biomass: Current findings determine research priorities,” The Scientific
1:2 1:3 1:4 1:2 1:3 1:4 1:2 1:3 1:4 World Journal, vol. 13, 2014.
Oedogonium sp. Thalassiosira sp. Chaetoceros sp. [3] D. Özçimen, B. İnan, and S. Akış, “Utilization alternatives of algal wastes
Alga–inoculum ratio for solid algal products,” In: Algal Biorefineries. Springer, pp. 393–418.
Fig. 5 Produced biogas yields according to granulometry and algae– 2015.
inoculum ratio for different algae-inoculum wastes. [4] N. CHAIB, S. DZIZI, F. NOUNE, and H. KADDECHE, “Production of
3rd generation biofuel from green microalgae : Oedogonium nodulosum,”
International Journal of Control, Energy and Electrical Engineering (CEEE),
Biogas yields obtained by using smaller granulometry vol. 9, pp. 10-14, 2019.
(0.25 mm) are higher than that produced by big grain size. All [5] J. Singh and S. Gu, “Biomass conversion to energy in India—A critique,”
correspond in our case to alga: inoculum mass ratio of 1:2. Renewable Suncesstaintanble Energy Review, vol. 14, pp.1367–1378, 2010.
The highest yields were obtained for Oedogonium sp. (342.59 [6] A. T. Koçer and D. Özçimen, “Investigation of the biogas production
potential from algal wastes,” Waste Management & Research, pp. 1-6, 2018
cm3 CH4 g-1 VS), then Thalassiosira sp. (287.65 cm3 CH4 g-1 [7] A. Karlsson, X. B. Truong, and B. H. Svensson, “Anaerobic treatment of
VS) followed by Chaetoceros sp. (254.37 cm3 CH4 g-1 VS). activated sludge from Swedish pulp and paper mills -Biogas production
potential and limitations,” Environmental Technology, DOI: [15] R. Rajkumar, Z. Yaakob, and M. S. Takriff, “Potential of the micro and
10.1080/09593330.2010.543932, Source: PubMed, 2011. macro algae for biofuel production: A brief review,” BioResources, vol. 9, pp.
[8] L. Appels, J. Lauwers, and J. Degreve, “Anaerobic digestion in global 1606–1633, 2013.
bioenergy production: potential and research challenges,” Renewable [16] D. Zhou, L. Zhang, and S. Zhang, “Hydrothermal liquefaction of
Sustainble Energy Review, Vol. 15, pp. 4295–4301, 2011. macroalgae enteromorpha prolifera to bio-oil,” Energy Fuels, vol. 24, pp.
[9] R. Ramaraj, Y. Unpaprom, and N. Dussadee, “Potential Evaluation of 4054–4061, 2010.
Biogas Production and Upgrading Through Algae,” International Journal of [17] J. Poullot, “Caractérisation des BIOGAZ Bibliographie : Mesures sur
New Technology and Research, vol. 2(3), pp. 128-133, 2016. sites,” INERIS DRC-02-27158-AIRE-n°316b-JPo, 2002.
[10] D. Hughes, M. S. Kelly and K. D. Black, “Biogas from Macroalgae: is it [18] J. De Vrieze, L. Raport, and B. Willems, “Inoculum selection influences
time to revisit the idea?” Biotechnology for Biofuels, vol. 5, pp. 86. 2012. the biochemical methane potential of agro-industrial substrates,” Microbial
[11] M. Hannon, J. Gimpel, and M. Tran, “Biofuels from algae: Challenges Biotechnology, vol. 8, pp. 776–786, 2016.
and potential,” Biofuels, vol. 1, pp. 763–784, 2010. [19] D. Yadav, L. Barbora, and L. Rangan, “Tea Waste and Food Waste as a
[12] Y. Chen, J. J. Cheng, and K. S. Creamer, “Inhibition of anaerobic Potential Feedstock for Biogas Production,” Environmental Progress &
digestion process: A review,” Bioresources Technology, vol. 99, pp. 4044– Sustainable Energy, DOI 10.1002/ep.12337, 2016.
4064. 2008. [20] A. Melis, “Green alga hydrogen production: progress, challenges and
[13] G. Migliore, C. Alisi, and A. Sprocati, “Anaerobic digestion of prospects,” International Journal of Hydrogen Energy, vol. 27, pp. 1217-
macroalgal biomass and sediments sourced from the Orbetello lagoon, Italy,” 1228, 2002.
Biomass and Bioenergy, vol. 42, pp. 69–77, 2012. [21] V. Alonso, A. Martin, and R. Borja, “Anaerobic treatment of wastewater
[14] C. H. Vanegas and J. Bartlett, “Green energy from marine algae: biogas produced in the manufacture of cellulosic pulp from wheat straw,”
production and composition from the anaerobic digestion of Irish seaweed Environmental Technology, vol. 14, pp. 1145-1153, 1993.
species,” Environmental Technology, DOI: 10.1080/09593330.2013.765922,
2013.
Abstract— In this work, we are interested in reaction (2) to produce synthesis gas according to the
the preparation of nickel-based materials following reactions:
deposited on vanadium oxide.Our solid are
elaborated by impregnation method and CH4 + H2O CO + 3H2H = 226 Kj/mol. (1)
characterized by different physico-chemical CO + H2O CO2 + H2H = -34.3 Kj/mol. (2)
analysis techniques such as: Atomic absorption,
BET, X-ray diffraction, Chemical analysis. The
results of XRD analyses carried out showed the
Due of the endothermic reaction, it operates that under
presence of the NiO phase for the solid. The severe conditions pressure and temperature (15 - 30 bar and
reaction of steam reforming of methane, in the 850-900°C) in order to increase the conversion to methane.
presence of nickel-based catalysts (4% nickel) The syngas (CO + H2) obtained has a defect in carbon
deposited on vanadium oxide, is carried out at monoxide. In these reactions, the supported catalysts are
atmospheric pressure between 500 and 650 ° C. largely used. The support is not necessarily inert, there are
Keywords—NiO, V2O4, hydrogen, steam reforming. metal-support interactions that exalt the activity of the
catalysts [1]. The reactivity and stability of Ni /SiO2 catalysts
I. INTRODUCTION are investigated in the steam reforming reaction at 565 °C. The
Today, alternative and sustainable solutions are proposed to results showed a great dependence on the steam-reforming
replace fossil fuels. Hydrogen is an alternative energy source according the ratio R = H2O / CH4.
for the future. It can be produced by electrolysis (4%) and Steam reforming of hydrocarbons is an industrial process of
essentially by reforming of fossil fuels (95%). Natural gas, high temperature during which a serious operational problem
consisting mainly of methane, propane and butane, is a fairly arises namely the deposit of carbon. To overcome this problem,
abundant and widely available raw material in the world. The Zhuang et al. [3] have shown that carbon deposition decreases
preferred route of its recovery is its transformation into on Ni catalysts promoted by cerium oxide CeO2, the latter
synthesis gas (CO + H2). It can be converted into many contributes to a better dispersion of Ni and increases its
products usually derived from petrochemistry (ammonia, activity. Thus, the addition of promoters (alkaline or alkaline-
methanol, higher alcohols, aldehydes, carboxylic acids, earth metal) to methane steam reforming catalysts generally
gasoline… suppresses carbon deposition [4].
Steam methane reforming is the most common and In this work, the search for catalysts that is more efficient,
economical way to make hydrogen. There are two primary available and less costly for used in CH4 + H2O reaction on
reactions: the reforming reaction (1) and the water gas shift nickel-based catalysts deposited on vanadium. Various
parameters that influenced on the catalytic activity are
investigated.
Copyright - 2019
Page 116
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
The solid Ni\V2O4 with a nickel percentage of 4 wt.% Xc(%)= [neCH4 – (nsCH4 + nsCO2 + nsCO)] x 100
was prepared by impregnation of the support V2O4with
Ni(NO3)2,6H2O (1 M) (Merck)[5-6]. The solution was stirred
during 2 h, and evaporated at 80 °C under agitation. Then solid III. RESULTS AND DISCUSSION
was dried at 80 °C (12 h), and calcinated at 300 °C (2 h) and
500 °C(6h) with a heating rate of 4 °C/mn (1.2 L h-1). The A. Materials characterization
solids were sieved to have a particle diameter less than 0.16 The physical characteristics of the solids are regrouped in
mm. the Table1
II.2. Catalysts characterization TABLE I
CHARACTERISTICS OF NI/V2O4 SYSTEMS
The solids were analysed by atomic absorption using
a Perkin Elmer 100 B spectrometer. The specific surface areas materials Ni (%) Surface area XRD
were determined by the BET method using an area analyzer (m2/g)
(Coultronics 2100E). The crystal structures were identified by
V2O4 - - 15 V2O4
XRD with a diffractometer (/2 CGF) provided with a
Ni/V2O4 4 3.95 5.8 NiO
proportional detector and using Kα molybdenum as
anticathode (λ=0.70930 Å). The X-ray patterns were obtained V2O4
for 2θ values ranging from 20 to 70°. The apparent size of
nickel particles was calculated by the Scherrer formula, The atomic absorption analysis shows that the Ni
L = 0.9λ/βcos θ, where β is the width of the most intense NiO composition is very close to the theoretical one.The specific
peak at half-height, and θ is the corresponding Bragg angle. surface areas of 4wt %Ni/V2O4 is slightly lower than the V2O4
support (See table 1).
II.3. Reactivity
The XRD pattern of 4 wt% Ni/V2O4 shows peaks assigned
The catalytic tests were carried out in a quartz reactor fixed
to NiO and V2O4 (Fig 1).
bed (L= 65 cm, = 1 cm) under atmospheric pressure. The
catalyst (0.1 g) is pre-reduced at 500 °C under hydrogen flow
(1.2 L h-1) before introduction of the gas mixture (H2O/CH4=
3.3) diluted in argon flow and obtained by crossing of a
mixture CH4 (10 %)/Ar, in a saturator with a thermostated 700
bath (65°C, 1.2 L h-1). The reactants and the products (CO, 600
CO2 and CH4) were analysed by gas chromatography using a
TCD chromatograph (Hewlett-Packard 5730) on a carbosieve 500
(%) (%)
60
44 17 05 3 78
Hydrogen(%)
500
50
550 52 16 21 5 63
40
600 61 11 31 8 58
30
650 60 11 67 13 22
20
The evolution of the CH4 conversion versus time shows that 500 550 600 650
Temperature (°C)
a stationary regime is reached after 3 h. The setting in mode
is accompanied by the formation of CO2 which strongly Fig 3: Production of hydrogen according to the temperature for
decreases with the profit of CO formation as one approaches 4%Ni/V2O4 nickel-based catalysts
the stationary regime. This indicates an incomplete reduction
of catalysts during the pre-treatment in H2 (500 °C) and thus The best hydrogen production is atlow temperature (500 ° C
the reactional mixture produces species on surfaces which are (78%)) because of the formation of carbon (fig 3).
more reduced than hydrogen itself [6]. Moreover, no diffusional limitation (effect of mass, flow
etc…) is observed. The experimental conditions of the reaction
are chosen after optimization. It appears that catalyst
60 elaborated with 4% of Nickel is active. It can be owed to a
better dispersion of the active phase for the weakly charged
50 catalyst. This result is in agreement with the literature data.
Conversion (% )
40 TT G (%)
Sel CO2(%) IV. CONCLUSIONS
30 The reactivity of CH4/H2O reaction for Ni/V2O4catalyst
was investigated at temperatures ranging from 500 to 650 °C
20
with the aim of facilitating the formation of synthesis gas. It
has been shown that: -The NiO phase is observed.
10
500 550 600 650 The catalyst remains stable for more than 7 hours, and the
Temperature (°C)
steady state is obtained after 3 hours, which suggests a synergy
between the nickel and the support. The best production of H2
in obtained at 650 ° C for the yield (≈ 60%).
Fig 2: Conversion of the methane and selectivity of CO2
versus temperature for 4%Ni/V2O4 nickel-based catalyst.
REFERENCES
The testing results show that catalytic performances
of Ni/V2O4 are very sensitive to temperature from 500 to 650 [1] S. Lee, J. Bae, S. Lim, J. Park, J Power Sources, 2008, 180:506.
[2] K O. Christensen, D. Chen, R. Lødeng, A. Holmen. Appl Catal A:
°C, and conversions varying between 44 and 60 % and CO General, 2006, 314(1): 9.
selectivity varying between 5 and 67%. We notice an [3] T. Borowiecki, W. Gac, A. Denis, Appl. Catal. A: Gen. 2004, 270:27.
important carbon deposition at 600 °C on catalysts. This result [4] N. Martin, M. Viniegra, E. Lima, G. Espinosa, Ind. Eng. Chem. Res.,
is in agreement with the literature, according to Nielsen [7] 2004, 43 : 1206.
[5] A. Belhadi, O. Cherifi; J. Soc. Alger. Chim., 2009, 19(1), 49-62.
who found that carbon deposition, on nickel catalyst has [6] A.BELHADI, L. BOUDJELLAL, S. BOUMAAZA , M. TRARI”
different temperature regimes: usually, ~ 550 °C. The Hydrogen production over the hetero-junction MnO2/SiO2 ” Int. J. Hyd.
activation energy of coke formation changes to a negative Energ; 2018 (43), 3418-3423.
value. [7] J.R. Rostrup-Nielsen, J.R. Anderson and M. Boudard (editors).
Catalysis- Science and Technology, Vol.5, Springer Verlag,
Heidellberg (1984), 1
1
Atlas Solaire Algérien, EISSN/ ISSN 1112-3850 ,2002. Fig.3. Solar radiation during measurement period
2
The panels are formed of 33 high-power monocrystalline Si solar cells.
A. Power conversion efficiency Then, the electrical efficiency in the case of the 50°
A photovoltaic system is determined according to its inclination angle is presented on Fig.6. It increased for the
electrical efficiency at standard conditions [2]. The linear ventilated panel with a rate of 0.54% on average. The
correlation between the temperature and electrical efficiency difference is very small and the ratio for the two cases is
is expressed in previous studies [8, 15] by the equation: almost the same after 60 minutes of operation. At the last
measure time, the ratio decreased of about 2%, this is due to
(1) the important temperature of the panel.
Where :
is the Power conversion efficiency ;
stc is the electrical efficiency in standard conditions ;
Β is the efficiency correction coefficient for temperature.
The value of is usually taken equal to 0.0045 1/°C for
crystalline Si cells [6]. For the current study, stc and were
estimated from measurements. The electrical efficiency at
standard conditions for the PV M75 is 14, 55% while was
deduced from the electrical efficiency results of the different
cases. It was estimated at 0, 0112 .
For the first angle of 20° as observed on Fig.4., the mean
rate of increase in the power conversion efficiency between Fig.6. Electrical power conversion efficiency
the retrofitted panel A and the conventional panel B was
about 0.88%. The ratio is nearly 12% for the panel A and
B. Panel temperature measurements
equal to 11-11,25% for the panel B.
During the tests, the temperature difference between
conventional and ventilated panels for the three inclinations is
shown in Fig. 7, Fig. 8 and Fig.9.
The temperature at the beginning of measurements was
greater for the "panel A". The fans were not turned on, so,
the module was more heated at the backside because of the
insulated box, but with the fans switched on, cooling started
and the panel temperature decreased in few minutes. For the
first case on Fig.7, the difference recorded was – 4°C at first
and 12° C as a maximum value, with a mean difference of 8
degrees. The temperature of the combined system was
approximately steady around 35° C and below the 40° C
Fig.4. Electrical power conversion efficiency
graduation.
The mean rate in the second case for an inclination angle
of 36°, is equal to 1,039 %. The ratio is about 11,5-12% in
panel A and close to the 11% value for panel B. The
difference between the two curves is quite steady.
For the last angle 50°, as it appears on Fig.12, the inlet air
temperature which was nearly the same of the ambient
temperature is high compared to the previous tests, the
difference values varied from 0.5 degrees to 2.5 degrees.
After the 90 first minutes, the cooling system became
inefficient.
Fig.9. Panel temperature at 50° inclination
ensures a quite steady cells temperature against an increasing [6] H.A. Zondag, D.W.De Vries, W.G.J. Van Helden, R. J.C. Van Zolingen,
A..A. Van Steenhoven. Solar Energy72 (2002) 113.
temperature for a conventional PV. The angle of inclination
[7] H.A. Zondag, D.W.De Vries, W.G.J. Van Helden, R. J.C. Van Zolingen,
equal or below the latitude of the site considered is more A..A. Van Steenhoven.. Solar Energy 74 (2003) 253.
favourable, where the performance turned out better for the [8] Stephan Krauter. Solar Energy Materials and Solar Cells,82(2004),131
measurement period. For the angle greater than the latitude, [9] Rossi C, Tagliafico LA, Scarpa F, Bianco V. 76 (2013)634
[10] Rossi C, De Rosa M, Bianco V, Scarpa F, Tagliafico LA. .
the electrical energy and conversion power efficiency
MECHANICS ’14: 10th International Conference on Applied and
declined due to the lower amount of solar energy absorbed by Theoretical Mechanics, Salerno, Italy, 2014,
the cells. [11] De Rosa Mattia, Romano Giorgio, Rossi Cecilia, Scarpa Federico,
Tagliafico Luca A., Energy Procedia 81(2015)345.
REFERENCES [12] Cătălin George Popovici, Sebastian Valeriu Hudișteanu, Theodor Dorin
Mateescu, Nelu-Cristian Cherecheș., Energy Procedia 85 (2016) 425.
[1] Renewables 2018 – Global status report, REN 21(2018).
[13] Lippin Pauly, L. Rekha, Christy V. Vazhappilly, Melvinraj C.R.
[2] Martin A. Green. Progress In photovoltaics Research And Applications
Procedia technology 24 (2016) 513.
17 (2009)183.
[14] S. Nizetic , F. Grubisic-Cabo, I. marinic-Ktagic, A.M. Papadopoulos..
[3] A. Goetzberger, V.U. Hoffman, Photovoltaic Solar Energy Generation,
Energy 111(2016)211.
Springer. Berlin Heidelberg, ed. New York ( 2005).
[4] Martin A. green, Yoshihiro Hishikawa, Ewan D. Dunlop, Dean H. Levi,
Jochen Hohl- Ebinger, Anita W. Y. Ho-Baillie. Wiley progress in ACKNOWLEDGMENTS
Photovoltaics, (2017). The authors wish to thank the laboratory technician for his collaboration.
[5] Wisam A. M. Al- Shohami, Raya Al-Dadah, Saad Mahmoud. Applied
thermal engineering 109 (2016) 475.
Unité de Développement des Equipements Solaires. UDES. /Centre de Développement des Energies Renouvelables .CDER. Bou-Ismail.42415.w.Tipaza. Algerie.
hbendjebbas@gmail.com
I. INTRODUCTION
Rapid growth of the solar industry is challenging engineers
and building officials to develop appropriate wind load
estimation and structural design approaches, given a current
lack of building code provisions directly applicable to solar
installations [1]. The need to understand and quantify wind
loading effects on individual and fields of tracking solar
collectors has emerged as an important design and
development issue. This understanding is important in the
cost effective design of both conventional concentrators and
innovative low-cost concepts which can be considerably less
robust than their conventional counterparts. Moreover, to aid Fig.1. LFR geometry
in the development of that understanding, methods for the
adequate modeling and simulation of wind loading on As illustrated in Fig.1, a simplified model of a Fresnel linear
individual concentrators within the actual field environment concentrator was used in this work with twelve mirrors 1m
are needed [2]. width and 10m length, the receiver is 7m from the ground and
It is important to determine the wind loads on individual the total length of the model is 15m.
modules within an array because the aggregate flow field is The numerical simulation consists in studying the effect of
influenced by adjacent modules. This is especially true if the wind on the LFR by varying the direction of the wind
mutual interactions between modules were such as to with respect to the LFR at 0 °, 30 °, 60 ° and 90 ° respectively
augment wind loads compared to those that would occur on a (Fig.2). The direction of 0 ° corresponds to a vertical
single concentrator at equivalent wind free-stream conditions. direction of the wind relative to the surface of the mirrors of
Some information is available for photovoltaic and parabolic- the LFR.
trough arrays [3].
The present study is a three-dimensional numerical
simulation of wind loads effect around linear Fresnel reflector
at different wind directions.
Winds of interest design have some characteristics near the III. NUMERICAL PROCEDURE
ground. Mean (time averaged) wind speed increases with
height above ground, and winds have gusts of short duration A. Governing equations
which are significantly higher than the mean and which define
peak wind loads. Both the variation of speed with height and The The k-ε turbulence model is the most widely used and
the turbulence must be accounted for in the design of LFR [6]. tested two-equation model, with many improvements
For the assessment of wind loads on structures, the power- incorporated over the years [4]. The formulation of this model
law profile has been used most widely because of its is as follow [5]: The Reynolds stresses ui' u 'j are related to the
simplicity [7] and it can be written as: shear stress of the flow ij by the following equation:
ij ui' u 'j
The value of the exponent “α” refers to the profile of the
mean velocity over uniform level surfaces equal to 1/7 in Where is the density of the fluid. We can find the value
open areas. is the mean wind velocity, is the for , and hence ui' u 'j from the following equation:
reference velocity at reference height z. Where is the density of the fluid. We can find the value
Three cases of reference velocity (reference height of 10m) for , and hence ui' u 'j from the following equation:
were considered: 30m/s, 40m/s and 50m/s respectively. u i u j 2
ij t k ij
x j xi 3
The result of a numerical calculation is dependent on the
mesh that is used. A too coarse mesh will give a high error in Where t is the turbulent viscosity. In the case of
the result and as the mesh size gets finer this error should k turbulence model, turbulence viscosity is defined as:
decrease. However, if the size of the mesh elements is small
enough so that the numerical result is close to the real k2
t C
solution, a further decrease of the element size should not
affect the solution significantly since the result is already
correct. Where C is the model constant [3].
TABLE I
VALUES OF k TURBULENCE MODEL CONSTANTS [3]
C C 1 C 2 k
COMSOL and 0.0 1.44 1.92 1.3 1.0
ANSY CFX 9
D( k ) k u i
t ij
Dt x j k x j
xj
D ( ) u i 2
t C 1 ij C 2
Dt x j x j
k x j k
Fig.3. Entire domain meshing and calculating domain
Down
Top
Fig.4. Pressure on LFR components (wind direction at 0°)
Fig.7. Pressure on the top and down of FLR in the case of 30°Wind direction
(30m/s).
Fig.6. Wind velocity vertical profiles Fig.8. Pressure with wind direction
A 3-D numerical simulation was performed to predict and The wind direction of 30° generates the most important
estimate the effect of wind loads on linear Fresnel pressures on the mirrors. Fig.8 representing the maximum
concentrator. Numerical computation was made for wind pressures exerted by the wind on the FLR mirrors with
speed of 30 m/s, 40m/s and 50m/s respectively by varying different directions shows that the maximum pressure is when
wind direction of 0°, 30°, 60° and 90° respectively. the wind speed is 50 and a wind direction of 30°
Abstract— In this paper we address the optimization of provided in section 5. Finally Section 6 provides the
production, distribution, and routing decisions to adapt conclusions drawn and future research lines.
retailers’ demand with the general objective of optimizing the
total cost. This work aims at implementing a solution that is II. STATE OF THE ART
related to production, replenishment, inventory allocation and
transportation for a combined optimization of stock and
In the research literature these four decisions can be found
transportation costs. The methodology adopted to tackle this in Bhatnagar et al. [8], Sarmiento and Nagi [9]. Regarding the
issue is based on two heuristics greedy and genetic algorithms. large number of studies in this area different tools and
The results demonstrate that the proposed heuristics are techniques have been used in past publication for the
effective and performs impressively in terms of computational optimization of the PRP problem. Mathematical techniques
efficiency and solution quality. which have been widely used in the past include heuristic
procedures. Several metaheuristics, such as Boudia using
Keywords— Optimization, Production-distribution systems, memetic algorithm with dynamic population management for
Routing decision, Inventory, Genetic Algorithms. an integrated production distribution problem [10], GRASP
[12], tabu search [13, 14], and ALNS [18]. Ettl et al [11] they
I. INTRODUCTION considered a queuing network model with constraint
This Over the recent years modeling and simulation of nonlinear optimization to minimize the inventory cost while
production routing problems (PRP) has been well satisfying customer service level requirements, the main
developed, given the huge role it plays in the analysis of the disadvantage of this approach is that it does not consider the
discrete behavior of this type of systems, this later combines production capacity, which is a key constraint in the PRP
lot-sizing decisions, inventory management and routing. In the systems. Chang et al. [15] discussed a serial production
PRP within each period production facility must determine system with finite storage space that follows a base-stock
whether or not to make the product and decide the policy. They developed a phase-type approximation to
corresponding quantity in which this product can be estimate the effective arrival rate of customers in which they
inventoried at the retailers, Considering production setup cost can determine the optimal base-stock level. Bard et al [17]
and variable cost, transportation and routing costs [4, 2]. they presented a comparative analysis of a series of heuristics
According to recent studies, integrated supply chain for an inventory routing problem (IRP) using branch-and-
planning problems consider the optimization of stock and price framework. Golden et al [16] were the first to
transport costs separately which is significant and cannot be investigate the interrelated problem of inventory allocation
neglected, Li et al. [1] a supply chain may be seen as a and vehicle routing. They developed a heuristic for designing
composed layers. Deliveries are made between each pair of an integrated delivery planning system aimed at comparing
layers. Each pair represents one level of the distribution the distribution rule used by the company with their approach.
network; transportation is considered the building block of Recently, new forms of relationships in the supply chain
any supply chain. Many shipment policies have been have been adopted. One of these is called vendor managed
proposed in literature for this problem. For instance, a current inventory (VMI), Archetti et al [7] have introduced a
trend in the application of freight distribution systems particular policy, known as deterministic order up to level
policy, the problem is to evaluate for each time period the
involving two or more levels where at first freight travels
quantity to ship to each retailer and the vehicle route, they
using ships and then is loaded on trucks to distribute it over
present a mixed integer linear programming model with
the warehouses or retailers [3]. The reminder of this paper additional valid inequalities to strengthen the linear relaxation.
organized as follows: section 2 presents the notation used Chitsaz et al [5] presented a general model for the assembly
throughout the paper and introduces the literature concept of routing problem (ARP). The problem formulated as a mixed-
the PRP. Mathematical formulation presented in section 3. integer linear program and they proposed a three-phase
Section 4 a two phase algorithm is presented to solve the decomposition matheuristic that relies on the iterative solution
proposed model and. The experimentation and results are of different subproblems. Leopoldo et al [6] developed a
Constraints (4) and (5) guarantee the upper bounds on Step 7: Solve production problem using retailers’ updated
inventory levels at production facility or retailers. Constraints costs using Greedy algorithm
(6) confirm that the inventory level of the production facility Step 8: Solve routing problem using Genetic algorithm
does not exceed the maximum level. Constraints (7) link the Step 9: Update sol if necessary
continuous variable to the binary variable which means if Until a stopping criterion in met
there is a delivery to a retailer in planning period t, then it Step 10: Calculate the sol
must be 1. Constraints (8) define the vehicle capacity. If (sol < best sol)
Constraints (9) ensure that each retailer is visited at most once Set best sol = sol
in a given period. Similarly, constraints (10) guarantee the End if
vehicle loading and routing restrictions used mostly once per Until a stopping criterion in met
planning period. Constraints (11) define number of trips End.
during planning horizon. Finally, constraints (12) and (13) are With the aim to simplify our solution, we set some goals
variable definitions. based on the retailer’s demand and the vehicles capacity, to
Initial attempts to solve instances of this size with CPLEX be more precise the sum of retailers’ demands in planning
8.1 were not encouraging. This was not surprising since the period t (Drt). The production unit model is demonstrated in
model is NP-hard, to deal with such a complex problem is to relations (14) to (19):
decompose it into smaller problems of reduced complexity
for which optimal or near optimal solutions can be
determined.
Given the fact that during the initial tests that we B. Genetic operators
performed to solve the production submodel problem a The mechanism for generating the initial population allows
greedy search algorithm attempts to generate a feasible the production of individuals’ population as the basis for
solutions, we found that this heuristic solutions are unlikely to future generations. The choice of the initial population is
perform well and even when they can be adapted, they may important because it can make more or less fast the
not be able to find good optimal feasible solutions in a convergence to the global optimum.
reasonable time. This implies the alternative of applying
extended heuristics, genetic algorithms. Hence the question 1) Selection and crossover
which needs to be asked is how to adapt two different The first step in genetic algorithm it starts with selection in
heuristics in an effective way which individuals are selected for crossover according to their
relative fitness, in this way the selection step implemented in
A. PROPOSED GENETIC ALGORITHM practice by the creation of a mating pool for the next
In the first phase chromosomes are created for an optimal generation. The choice of a selection technique in the
size that we describe the routing phase, in which corresponds suggested genetic algorithm is the roulette wheel operator
in calculating vehicle routes at each period. A chromosome is which used to select the initial population to create new
a sequence of nodes, indicating the order in which vehicles children, the probability of choosing an individual depends
must visit all retailers. In this case we consider a set of directly on its fitness.
retailers to be served by a fleet of identical vehicles in In the next step the crossover occurs when two members of
planning periods; we set a two dimensional matrix that each a population (chromosomes) are selected for reproduction, and
cell contains the retailer’s demand in a given time period. is usually based upon their relative fitness in solving the
The binary variables Urzt that imply if a vehicle z is used problem. Two parents P1 and P2 are selected randomly we use
in planning period t to serve a retailer r is the main issue to the order crossover (OX) with slightly modification, it starts
identify the other decision variables. The above procedure is by choosing two crossover points then applying sliding motion
repeated until a stopping criterion is achieved. The details of to fill up the left out holes by sending the mapped positions. It
the basic genetic algorithm are given below in algorithm 2 copies the portion of permutation elements between the
crossover points from the cut string directly to the offspring,
Algorithm 2 placing them in the same absolute position. An example of
Begin offspring generation is illustrated in Fig 1
[1] LI, J., SAVA, A., and XIE, X. Performance evaluation and [11] Ettl, M., Feigin, G. E., Lin, G.Y. and Yao. D.D. "A Supply Network
optimization of a two-stage production-distribution system with batch Model with Base-Stock Control and Service Requirements", Operations
orders and finite transportation time. The 16th world congress of IFAC, Research, Vol. 48/2, 2000. pp. 216-232.
Prague.2005 [12] Boudia M, Louly MAO,Prins C.Areactive GRASP and path relinking
[2] Thomas DJ, Griffin PM Coordinated supply chain management. Eur J for a combined production-distribution problem. Comput Oper Res;34
Oper Res 94(1):1–15. 1996 (11):3402–19. 2007
[3] Infante D, Paletta G,Vocaturo F.A ship-truck intermodal transportation [13] Armentano VA, ShiguemotoAL,Løkketangen A.Tabu search with path
problem. Marit Econ Logist; 11:247–59. 2009 relinking for an integrated production-distribution problem. Comput
[4] Geoffrion AM, Powers RF. 20 years of strategic distribution system Oper Res; 38(8):1199–209. 2011
design: an evolutionary perspective. Interfaces 25(5): 105–127. 1995 [14] Bard JF, Nananukul N.The integrated production-inventory-
[5] Masoud Chitsaz, Jean-François Cordeau, and Raf Jans. A unified distribution- routing problem. J Sched; 12(3):257–80. 2009
decomposition matheuristic for assembly, production, and inventory [15] Chang K-H, Lu Y-S Inventory management in a base-stock controlled
routing. INFORMS Journal on Computing, 31(1):134–152, 2019. serial production system with finite storage space. Math Comput Model
[6] Leopoldo Eduardo Cárdenas-Barrón, José Luis González-Velarde, 54:2750–2759. 2011
Gerardo TreviñoGarza, and Dagoberto Garza-Nuñez. Heuristic [16] Golden, B., Assad, A., Levy, L., and Gheysens, F. (1984). The fleet
algorithm based on reduce and optimize approach for a selective and size and mix vehicle routing problem. In Computers and Operations
periodic inventory routing problem in a waste vegetable oil collection Research, 11:49–66.
environment. International Journal of Production Economics, 211:44– [17] J. F. Bard and N. Nananukul, Heuristics for a multiperiod inventory
59, 2019. routing problem with production decisions, Computers and Industrial
[7] Archetti C, Bertazzi L, Laporte G, Speranza MG. A branch-and-cut Engineering, vol. 57, no. 3, pp. 713–723, 2009.
algorithm for a vendor managed inventory routing problem.TranspSci; [18] Adulyasak Y, Cordeau J-F, Jans R. Optimization based adaptive large
41(3):382–91. 2007 neighbor- hood search for the production routing problem. Transp
[8] Bhatnagar R, Chandra P, Goyal SK Models for multi-plant Sci;48(1):20-45. 2014
coordination. Eur J Oper Res 67(2):141–160. 1993 [19] Arkin E, Joneja D, Roundy R (1989) Computational complexity of
[9] Sarmiento AM, Nagi R. A review of integrated analysis of production- uncapacitated multi echelon production planning problems. OperRes
distribution systems. IIE Trans 11(3):1061–1074. 1999 Lett 8(2):61–66
[10] Boudia M,Prins C.A memetic algorithm with dynamic population
management for an integrated production-distribution problem. Eur J
Oper Res; 195 (3) :703–15. 2009
Abstract— This paper presents a thermodynamic simulation most efficient way of converting the energy embedded in
model of the olive pomace gasification process intended to biomass, and it is becoming one of the best alternatives for
produce hydrogen rich syngas. the reuse of waste solids [3].
The considered gasification process is simulated using the flow- Biomass gasification process produces very low levels of
sheeting software ASPENTECH.
particulates and very little amounts of NOx and SOx when
The performance of gasification process is investigated by
varying the gasification temperature, pressure and the steam to compared with fossil fuels [4]. It can also be used as a source
biomass ratio R in the feed. to produce various chemical species. The addition of steam in
Results show that for all investigated pressures, the produced the gasification process can enhance hydrogen production. [5,
hydrogen increases with temperature, reaches a maximum and 6]
then decreases. Extensive research has been carried out in biomass
We also noted that the parameter R has a large effect on the gasification using different reactor configurations, operating
H2/CO ratio in the exiting gas in the temperature range between condition and gasification agents.
900 and 1300°C. The H2/CO molar ratio is at its maximum with Thermodynamic models constitute valuable investigation
a value of two.
tools for prediction of produced gas composition under
Keywords— Olive pomace gasification, hydrogen production, various gasification operating conditions, in particular as it is
steam/biomass ratio, biomass, Aspen plus. the case in the present paper, for prediction of the expected
I. INTRODUCTION H2/CO ratio.
Olive oil extraction processes widely used by oil mills are
conventional pressure processes, either in batch mode (super
The increase in energy consumption in all economic press) or in continuous mode (centrifugation). These
sectors and the expected depletion of fossil fuel resources processes produce, apart from olive oil, by-products such
lead to a shortage of hundreds of tonnes of fossil fuels. This water as olive pomace.
insist the use of other alternative and renewable energy The thermal recovery of olive cake is the most relevant
resources such as solar photovoltaic, biomass, wind and alternative for the reduction of these olive by-products.
hydroelectric power generation. To meet this energy demand, Thermal processes convert these pomaces into usable energy
combustion processes, which are currently the most used for in different forms. The three main thermochemical
transport and industry, the tertiary sector or the residential conversions of olive pomace commonly developed
sector, will have to evolve in stages and be substantially correspond to combustion, pyrolysis, and gasification.
improved. They will have to diversify from fuel by The objective of this paper is to investigate the hydrogen
introducing a growing share of biomass, and waste products. and syngas production using steam gasification of olive
Also, as we face the problem of global warming and pomace.
greenhouse gas emission from conventional fuels via The study is carried out assuming the establishment of the
anthropogenic actions, extensive research and development thermodynamic equilibrium in the gasifier at the prevailing
focused on the use of biomass as auxiliary to fossil fuels. conditions of temperature and pressure set. The composition
Biomass is an attractive renewable energy source with of the gas is calculated by minimizing the free enthalpy of the
minimum negative environmental impact as compared to mixture, taking into account mass and energy balances. For
fossil fuels, because of its low sulphur content and its carbon the following simulations, the mass flow rate of olive pomace
dioxide neutral emissions [1, 2]. was set at 1000 kg/hr while the flow rate of water vapour was
Gasification converts biomass through partial oxidation varied from 100 to 1000 kg/hr. The temperature range
into an energy rich gaseous mixture, small quantities of char
and condensable compounds. It is considered as one of the
The whole gasification process (pyrolysis in the analysis data. This approach is based on the minimization of
decomposer then gasification in the gasifier) includes a set of the Gibbs energy of the reactanting mixture [12].
consecutive and parallel reactions summarized in Table II
2) Process modelling
[13].
The simulation are performed using ASPENTECH. In this
platform there is no especial gasification model ready for use.
TABLE II
CHEMICAL REACTIONS IN GASIFICATION PROCESS
Therefore the modelling of the whole process is decomposed
into blocks. The main unit operations involved in gasification
are drying, pyrolysis or decomposition, gasification,
combustion and finally, gas separating. These operations are
∆H25°C represented by blocks simulating their functions, namely
Reaction Reaction Name
(kJ/mol) reactor block RSTOIC, flash block FLASH2, reactor block
Combustion RYIELD, adiabatic reactor block RGIBBS and cyclone
C + 0.5O2 ↔ CO Partial oxidation -111 separator block SSPLIT.
As this study is based on an equilibrium approach, the
CO + 0.5O2 ↔ CO2 CO Oxidation -284 reaction kinetic parameters are not needed.
H2 + 0.5O2 ↔ H2O
Hydrogen
484
The following assumptions were contemplated in
combustion modelling the gasification process:
Reduction
C + CO2 ↔ 2CO
Boudourad
172
Steady state;
reaction Residence time for reactants sufficient to reach
C + H2O ↔ CO + H2 Steam reforming 131 chemical equilibrium;
Tar formation not considered;
Methane
C + 2H2 ↔ CH4 -74 Char is 100% carbon;
formation
CO + H2O ↔ CO2 + H2 Water gas shift -42
Biomass volatilization instantaneous and volatile
products consist mainly of H2, CO, CO2, CH4 and
Methane
CH4 + H2O ↔ CO + 3H2 206 H2O [14].
reforming
CO2 Reforming of
CH4 + CO2 ↔ 2CO + 2H2 247 The main objective of this paper is to investigate the effect
methane
of various operating parameters (Temperature, pressure, and
R, steam/biomass ratio) on the gasification process’s
performance.
B. Model development The effect of temperature on the syngas composition is
studied from 500 to 1500°C. Pressure is set respectively at 1,
1) Thermodynamic equilibrium model 50, 150, 250 bar. For R we consider just two values in the
Equilibrium models have mainly two approaches: simulation, R = 0.5 and R = 1.0.
stoichiometric and non-stoichiometric. The stoichiometric Gasification conditions summarized in Table III.
approach requires a clearly defined reaction set that
incorporates all chemical reactions and species involved. In
the non-stoichiometric approach, no particular reaction routes
considered in the numerical simulation. The only input
requested is the component set at equilibrium and the
elemental composition of feed, obtained from ultimate
B. Effect of pressure
Pressure is also an important factor which affects the
composition of syngas production and gasification process. Fig 3: Exit gas composition vs. temperature (R=0.5, P=50bar)
As the simulation results show, lower pressures for fixed
gasifier temperature favour the amount of H2 and CO.
Opposite effect is observed for the concentration of CO2 and
CH4 : they increase with pressure. Le Chatelier’s law may
explain these trends: chemical equilibrium at higher pressure
evolves in a direction that favours the decreasing of the gas
phase, enhancing so reactions consuming gas reactants.
C. Effect of steam/biomass ratio R
Fig 5: Exit gas composition vs. temperature (R=0.5, P=250bar) Fig 6: Exit gas composition vs. temperature (R=1, P=150bar)
C. Conclusion
A thermodynamic equilibrium model of olive pomace
gasification in a fixed bed gasifier is developed. Process
simulations using AspenPlus are performed in order to
investigate the effect of pressure, temperature and biomass to
steam ratio on the composition of produced gas. High
hydrogen amounts are observed by the highest considered
temperature. Lower pressure favours the production of H2.
The steam to biomass ratio R is to be optimized to ensure the
energy efficiency of the gasification process.
TABLE IV REFERENCES
SIMULATED CHARACTERISTICS VALUE OF MAXIMUM
[1] K. Maniatis, and G.Guiu, J.Riesgo, “Pyrolysis and gasification of
HYDROGEN PRODUCED biomass and waste, proceeding of an expert meeting”, The European
commission perspective in biomass and waste thermochemical
conversion., pp. 1–18, Nov. 2002.
R=0.5 [2] Lin Y, Tanaka S. “Ethanol fermentation from biomass resources:
current state and prospects,” Appl Microbiol Biotechnol, paper 42, pp.
Maximum 69-627, 2006,
Pressure Temperature
hydrogen [3] Maria Puig- Arnavat, Joan Carlos Bruno, Alberto Coronas, “Review
[bar] [°C]
produced and analysis of biomass gazification models,” Renewable and
Sustainable,Energy Reviews 14 pp.2841-2851, 2010.
[4] (2002) The IEEE website. [Online]. Available: http://www.ieee.org/
1 0.5 800 [5] M. Shell. (2002) IEEEtran homepage on CTAN. [Online]. Available:
50 0.5 1100 http://www.ctan.org/tex-
150 0.5 1400 archive/macros/latex/contrib/supported/IEEEtran/
[6] Moflih A.Adnan, Herri Susanto, Housam Binous, Oki Muraza. “Feed
250 0.5 1500 composition and gasification potential of several biomasses including
R=1 microalgae: a thermodynamic modelling approach”, International
journal of hydrogen energy 42, pp.17009-17019, 22 June 2017.
1 0.5 700
50 0.4 1100 [7] Mohammed Alae Bennini, Abdelghani Koukouch, Ilias Bakhtari,
150 0.4 1300 Mohamed asbik, Benoit Cagnon, Sylvie Bonnamy. “Characterization
and combustion of olive pomace in a fixed bed boiler: Effect of
250 0.4 1300 particle size” International journal of Heat and technology. paper 37,
pp. 229-238, Mach, 2019.
[8] Koukouch A, Idlimam A, Asbik M, Sarh B, Izrar B, Bostyn S, Bah A,
Ansari O, Zegaoui O, Amine A. “Experimental determination of the
effective moisture diffusivity and activation energy during convective
solar drying of olive pomace waste”. Renewable Energy 101: pp.565-
574, 2017.
http://doi.org/10.1016/j.renene.2016.09.006
Abstract: In this work, we have studied the effect of acid nanocarbon formed at reaction temperature 600 °C [1, 2].
treatment of a commercial HZSM-5 zeolite (HZ, Si/Al=11.5) on Metallic Ni° particles are the active species in the
the production of COx-free hydrogen and nano carbon material decomposition of CH4 and nano carbon growth [12]. It is well
deposited during the catalytic decomposition of methane (CDM)
known that the catalyst supporting material play an important
at 520 °C and 1 bar over pre-reduced 20 wt. % Ni-catalysts
supported on HZ and HZ-a sample resulting from the acid role in the dispersion of the active phase such as Ni
treatment. Both catalysts, prepared by impregnation, are crystallites, their size and surface areas and in the metal-
characterized by XRD, FTIR, N2-sorption, H2-TPR, SEM, EDX, support interactions which affect catalytic properties of Ni-
XPS, Raman, and potentiometric titration for acidity catalysts in the CDM [12]. Among Ni-catalysts supported on
measurement. The catalysts display close catalytic behaviour due conventional oxides like Al2O3, SiO2, TiO2, MgO, ZrO2, SiO2
to the slight structural and textural modifications of HZ-a. TEM is found as the best performing material [12] while Ni
images of both catalysts after 4 h of CDM show nano fibres of supported on HY zeolite is the most active and stable in the
carbon with Ni° at their tip, while XRD patterns indicate a better CDM compared to Ni supported on USY, SiO2, and SBA-15
crystallinity of the ZSM-5 phase and metallic Ni particles with
[13]. Ni-catalysts supported on HZSM-5 zeolites are found
HZ-a support.
promising catalysts in CDM process [4, 5]. The ordered
Keywords: CH4, CDM, Ni, ZSM-5 dealumination, H2, CNFs. mesoporous silicates (MCM-41) used as supports improve Ni
dispersion and carbon diffusion during methane
I. INTRODUCTION decomposition [2]. When HZSM-5 zeolites with different
Methane is the major component of natural gas, and this Si/Al ratios (40, 150, 300 and 485) are used as supports of 30
latter resource is the one’s world abundant. Hydrogen is wt. % Ni based catalysts, the one with a ratio of 150 exhibits a
considered as a large-scale carrier for renewable energy. The better H2 yield and longevity for the CDM carried out at 550
catalytic decomposition of methane (CDM) is an alternative °C under flow of pure methane. Best results are obtained
and promising route, compared to conventional processes, in (H2:171 Nm3 molNi) for 20 wt. % Ni/HZSM-5 (Si/Al=150)
the recent years to generate COx-free H2 for many [5]. The authors [5] related the superior performance of this
applications such as the fuel cells [1-11]. In addition, valuable catalyst to the higher Ni metal surface area and a direct
by-products such as carbon nanotube (CNT: SWCNT, correlation between Ni loading, H2 yield and Ni metal surface
MWCNT) and carbon nanofibers (CNF) which have area was deduced from their results. Awadallah et al. [4]
exceptional electrical, chemical and mechanical properties [1- compared catalytic performances of 40 wt. % Ni supported on
11]. The decomposition reaction is as follow: Z-25, Z-400 (Z=HZSM-5, number=Si/Al ratio) and
𝐶𝑎𝑡𝑎𝑙𝑦𝑠𝑡 amorphous silica (AS), and found that the texture,
CH4 (g) → 2 H2 (g) + Carbon (s) H° = 75.6 kJ mol-1. morphology, pore topology, crystalline structure of the
Without catalyst, cracking of the strong covalent C-H bonds support affect catalytic performances for the H2 production
occurs at temperatures 1200 °C and amorphous carbon is and nature of carbon deposited. The crystallite size and
obtained [11]. These temperatures are reduced by using dispersion of NiO are influenced by the crystalline structure
transition metal (Fe, Co and Ni) and especially Ni supported and pore geometry of the supports. Thus, Ni/Z400 exhibits
on conventional oxides, zeolites and ordered mesoporous higher metal Ni° dispersion, higher specific surface area,
silicates [1-11]. Ni-catalysts are the most active in CDM smaller Ni particles and better catalytic performance for the
reaction at temperature 600 °C. Ni metal contains partially CDM compared to Ni/Z-25 and Ni/AS.
filled 3d orbitals ([Ar] 3d8 4s2) which improve the In this work, we tried to modify a commercial HZSM-5
dissociation of CH4 molecules through partially accepting (HZ) with low Si/Al ratio of 11.5 by an acid treatment. The
electron [1]. But, the drawback of Ni-catalysts for CDM is resulted HZ-a sample along with HZ are used to prepare
their rapid deactivation due to sintering and encapsulation of Ni/HZ and Ni/HZ-a catalysts (Ni: 20 wt. %). The
nickel particles by the nano carbon. Addition of a second effectiveness of the post-synthesis treatment of the bare
metal such as Cu by creation of Ni-Cu alloys improves not support on H2 yield and nano carbon deposited during
only H2 production but also graphitization degree of
Copyright - 2019
Page 141
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
decomposition of methane at 520 °C is studied. Catalysts are 24.39° [14]. More detailed analysis of the XRD patterns
characterized by several techniques to see any correlation with reveals that all reflections of HZ-a are slightly shifted to
catalytic performances. higher values of 2 ( = 0.08-0.10°), indicating a very slight
dealumination, although Al rich zeolites (Si/Al20) are known
to be refractory to dealumination and desilication in acidic and
II. EXPERIMENTAL alkaline solutions, respectively [15]. This is confirmed by a
II.2.1 Catalysts preparation slight increase in the Si/Al molar ratio from 11.5 for HZ to
14.5 for HZ-a (EDX analysis) and by a crystallinity loss of
5 % by comparison to that of HZ taken as reference, hence
The support HZSM-5, denoted HZ, is prepared by 100 % crystalline. For the same Si/Al ratio, the topology of
calcination of a commercial NH4ZSM-5 (CBV 2314,
the zeolite plays an important role during a post-synthesis
Si/Al=11.5) at 500 °C for 5 h under air flow. A sample of HZ
treatment. Thus, BEA zeolite is more sensitive than HZSM-5.
is treated, under stirring and reflux at 80 °C for 5 h, in 2 M
As can be seen in Table 1, the crystallite size of HZ-a is
HCl aqueous solution (20 mL HCl/gzeolite).The resultant HZ-
slightly smaller than that of HZ with values of 52 and 55 nm,
a solid, after thorough washing, is oven-dried at 120 °C respectively. This indicates a slight contraction of the unit cell
overnight, and then calcined in air at 500˚C for 5 h. Ni/HZ and of HZ-a
Ni/HZ-a catalysts (Ni: 20 wt. %) are prepared by the wet
impregnation procedure using Ni (NO3)2.6H2O as precursor. Table I. Physico-chemical properties of the bare zeolites and
calcined and reduced 20 wt. % Ni-catalysts.
II.2.2 Catalyst characterization
Sample (Si/Al)a Cryst.b H+ c Particle size (nm) d
The crystalline phases are analysed by XRD in the 2 range (%) µmol/g Zeolite NiO Ni
of 10°-80°. The textural properties of the fresh catalysts are HZ 11.5 100 1061 55.0 - -
obtained from the data of N2 adsorption-desorption isotherms NiO/HZ 11.5 79 539 63.0 25 -
carried out at – 196 °C. The morphology of the samples is Ni/HZ 11.5 26 - 17.5 - 18
characterized by SEM and carbon formed by TEM, XRD and HZ-a 14.5 100 909 52.0 - -
Raman spectroscopy. The reducibility of catalysts and NiO/HZ- 14.5 90 212 64.0 23 -
interaction metal-support can be obtained from the H2-TPR a
profiles which are carried out in a home-made pulsed set-up. Ni/HZ-a 14.5 38 - 17 - 17
II.2.3 Decomposition of methane a: Framework Si/Al ratio.
b:Crystallinity (%) of zeolite = (I sample/I reference sample)x100.
Catalytic methane decomposition reaction is carried out at c: Number of Brønsted acid sites determined by potentiometric
520 °C and 1 bar in a fixed bed glass reactor with a catalyst titration.
d: particle size calculated via the Scherrer formula.
mass of 0.2 g and CH4 flow of 30 mL/min. Prior to the
reaction, the catalyst is reduced at 520 °C for 2 h with H2 (30
mL/min). The outflow gas is directly injected via a 6-way
valve (sample loop: 0256 mL) into the HP chromatograph
equipped with a 13 X column and TCD detector using Ar as a
+10000
formula:
H2 (mol. %) =
nH2
x 100
HZ-a
nH2+n unconsumed CH4
HZ
III. 3.1.Characterization results
Fig. 2 shows the XRD patterns of the fresh calcined 20 wt. % supported HZ material is the most affected. Reflections of
Ni supported on HZ and HZ-a samples which both exhibit metallic Ni particles and their crystallites size are slightly
diffraction reflections of crystalline NiO cubic phase at 2 = higher for reduced Ni/HZ-a compared to Ni/HZ (Table 1).
37.3°, 43.3°, 62.9°, 75.5° and 79.4° [JCPDS n° 01-1239]
along with clear reflexions of the HZSM-5 phase. Nickel
oxide such as Ni2O3 and NiSiO4 and Ni2Al2O4 phases are not
Ni
+ 8000
detected. So, calcined catalysts are denoted NiO/HZ and Ni/HZ-a
NiO/HZ-a while the reduced forms as Ni/HZ and Ni/HZ-a.
The crystallite size of NiO decreases from 42 nm with loaded
HZ to 35 nm for HZ-a (Table 1) while the crystallinity of HZ-
Intensity, a.u.
a support is slightly higher than that of supported HZ. A
drastic loss in the crystallinity of loaded zeolites, by
comparison to their unloaded counterparts, is expected after
Ni loading [4, 5], but here it is only of 10 and 21 % for HZ-a
Ni/HZ
and HZ samples, respectively. Large NiO particles located at
the outer surface of the supports and not interacting with them
could explain this relatively good crystallinity. But, SEM
images and Ni 2p XPS spectra indicate more NiO species at
the external surface of HZ-a. With a high Ni loading of 20 5 15 25 35 45 55 65 75
wt. %, a great part of NiO species will be located at the
2 Theta, degree
external surface of the zeolite support, and this for both
catalysts. Interaction between the support and NiO species can Fig. 3 XRD patterns of fresh reduced Ni- catalysts.
limit the aggregation of the latters. According to Awadallah et
al. [4], the use of HZSM-5 with well-defined porosity as
support limits aggregation of NiO species in 40 wt. % Ni- Fig. 4a and 4b show SEM images of the fresh calcined
catalysts by comparison to amorphous silica. NiO/HZ and NiO/HZ-a catalysts, respectively. The latter
catalyst seem to have more NiO species with smaller size at
the external surface of the HZ-a support. A greater
aggregation of NiO species seems to occur at the external
NiO/HZ-a NiO surface of HZ. Similar trend is observed by SEM for metallic
+ 8000
NiO/HZ
5 15 25 35 45 55 65 75
2 Theta, degree
Fig. 2 XRD patterns of fresh calcined NiO-catalysts.
Fig. 4 SEM images of fresh a) NiO/HZ and b) NiO/HZ-a catalysts.
P/P0=0.99. The micropore volume (Vmicro) is calculated via be assigned to the reduction of free NiO species having very
the t-plot method and the mesopore (Vmeso) is obtained by weak interaction with the supports and located on the external
subtracting Vmicro from Vp. As expected, after Ni loading of surface of the latters. The intense peak at 556-561 °C is due to
a zeolite, the specific surface of its loaded counterpart the reduction of bulk NiO. The weak shoulder reduction peaks
decreases by 37 and 14 % for HZ and HZ-a. This is in at T>561 °C could be assigned to the reduction of NiO
agreement with the XRD results. Table 1 shows clearly that deposited in the zeolite channels and also to large sized NiO
only Vmicro of the loaded supports is affected by a partial clusters [4, 5]. H2 consumed, given by the total area below
blocking by NiO species mainly located outside the pores. The the curves, is slightly lower with HZ-a support. Both catalysts
decrease in Vmicro leads to that of Brønsted acid sites located are totally reduced for T700 °C.
inside the pores by 50 and 70 % for HZ and HZ-a supports,
respectively.
III.3.2. Catalytic results
Table II. Characteristics of the calcined samples.
SBET Vpa Vmicrob Vmesoc Pore size The unloaded HZ and HZ-a samples are inactive in the
Sample catalytic decomposition of methane (CDM) at 520 °C under
(m2/g) (cm3/g) (nm)d
HZ 390 0.180 0.13 0.05 1.85 atmospheric pressure and pure flow of CH4. According to the
HZ-a literature [ ] and from our results, the active species for this
316 0.176 0.11 0.07 2.23
reaction are metallic Ni° particles [ ], and when NiO species
NiO/HZ 245 0.176 0.08 0.09 2.88 are not totally reduced, undesirable COx products are formed
NiO/HZa 271 0.160 0.07 0.07 2.36 in the H2 stream [ ]. Tested in the same operating conditions
and Ni loading (20 wt. %), the CDM reaction over pre-
a: Total pore volume determined at P/P 0=0.99. reduced Ni/HZ and Ni/HZ-a catalysts leads to COx-free
b: Micropore volume determined by the t-plot method.
c: Vmeso = Vp - Vmicro hydrogen and solid nano carbon according to the endothermic
d: Pore diameter calculated from 4 Vp/S
BET formula.
reaction: CH4 (g) H 2 (g) + C (s). So, if differences in
the catalytic performances are observed, they could be due to
the supports. According to Anjaneyulu et al. [5], there is a
direct relation between particle size and surface area of
400 metallic Ni particles and hydrogen yield. H2 formation is
561 °C
H2 consumption, µmol/g
Copyright - 2019
Page 144
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
min of reaction time with HZ-a support can be related to more other hand, the characteristic diffraction peaks of metallic Ni
accessible Ni° particles. However, more coke is deposited on phase along with those of ZSM-5 are clearly seen in Fig. 8,
Ni/HZ catalyst notably for the reacted Ni/HZ-a, indicating weak metal-
support interactions for this catalyst. Both Ni and ZSM-5
phases are drastically affected for Ni/HZ sample, indicating
45
likely a recovery by the carbon. A slight higher accumulation
of carbon is observed for the latter catalyst.
40
H2, mol %
Ni
35
+1200
Intensity, a.u.
30 Ni/HZ Ni/HZ-a
25
0 60 120 180 240 Graphitic carbon
Ni/HZ
Time on stream, min
REFERENCES
Fig. 7 TEM images of used catalysts in CDM at 520 °C for 4 h.
[3] Srilatha, K., Bhagawan, D., Shiva Kumar, S., Himabindu, V.,
Sustainable Fuel Production by Thermocatalytic Decomposition of
Methane – A Review, South African Journal of
Chemical Engineering, vol 24, pp 156-167, 2017.
[6] Ashik UPM, Wan Daud WMA, Abbas HF, Production of greehouse
gas free hydrogen by thermocatalytic decomposition of methane- a
review, Renew Sustain Energy Rev., vol. 44, pp. 2221-2256, 2015
1
siyoucef.hamza@univ-mascara.dz
2
b.afif@univ-mascara.dz, afifafif22@yahoo.fr
3
habib.bouhamida @univ-mascara.dz
4
boualem19985@yahoo.fr
*
LSTE Laboratory, M.S Mascara University, BP 305, Mascara 29000, Algeria
Abstract— Electricity originating from solar is nowadays one of used [4]. PV systems are useful in areas highly
the main contributors of electrical power to energy systems in the
future, but solar technology needs support to find its way to the
exposed to incident solar radiation [5]. Designing an
market. Since photovoltaic electricity always involves very high
costs compared to electricity from other sources, a striking
question is: how can photovoltaic achieve competitiveness? There
are different strategies for promoting applied photovoltaic
electricity generation around the world. In this work, we
simulated and designed a 40Kwp injected photovoltaic power
plant, evaluating the energy of the incident photon flux, the
energy produced by the photovoltaic panels and the energy
injected into the electricity grid. And this is to judge whether the
installation.
Energy sources like wind, solar or hydro became optimal hybrid (Wind/PV), a standalone system has
more and more popular mainly because they don’t been analysed, based on environmental and
produce emissions and are inexhaustible. PV energy economic aspects. When such design is accepted
is the fastest growing renewable source with a our findings are so as CO2 emission can be mostly
history dating since it has been first used as power reduced, compared to other existing Diesel-only
supply for space satellites. Currently, electricity is systems [6-8]. Different configurations types have
almost entirely supplied to cold rooms by been analysed, showing that based on the energy
conventional hydroelectric or thermal power plants cost, a hybrid PV-Diesel-Battery system is more
[1]. Greenhouse gas emissions and other cost-efficient than Diesel only and PV-Diesel
environmental pollution issues, for ex., are of without storage [9-11]. It has also shown that a rural
increasing concern [2]. Renewable energy village is economically best suited for a PV-Diesel
technologies, such as photovoltaic cells, are power generation system when energy storage
recommended for electricity production [1]. The systems are absent [12-14]. To identify an optimum
photovoltaic (PV) systems performances depend on model of PV-Diesel-Battery system in various
geographical locations and types of PV modules climatic areas, we developed other models (not
studied here) using Homer software.
Page 147
Copyright - 2019
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Discover 2VRE-
1600TG
Capital (DA) Replacement
(DA)
OM (DA) Salvage
(DA)
Total (DA)
thermal and electrical load served, respectively. The
Discover 2VRE-
1600TG
60000.00 89156.84 6534.18 -73827.27 81863.76
second term in the numerator is the portion of the
Homer Combined
Dispatch
160000.00 0.00 1966025.43 0.00 356025.43
annualized cost that results from serving the thermal
Solar world 250
SW
55084.32 0.00 473.59 0.00 55557.91
load. In systems, such as wind or PV, that do not
250 Mono
System Converter 556.12 784.46 0.00 -304.12 1000.45 serve a thermal load (Hthermal=0), this term is zero.
System 275640.43 89905.30 203033.21 -74131.39 494447.55 COE is a convenient metric with which to compare
systems, but HOMER does not rank systems based
Here, we have chosen the optimal choice for our on COE. Operating costs are annualized values of
study is the use of solar panels with batteries. The
all costs and revenues other than initial capital costs.
system architecture (SolarWord250SW 250 Mono
(0.145kW)) is so that 0.0371 kW was converted,
using Homer Combined Dispatch (Discover 2VRE-
1600TG (2.00 strings). Total costs (such as NPC)
are as follows :
Total NPC :494447,60DA,
Levelized COE : 102.83DA,
Fig. 4 Total NPC for hybrid PV/storage system.
Operating Cost : 6697.31DA.
The total net present cost (NPC) of a system is the
present value of all the costs the system incurs over The current net cost for 25 years is 494447.60 DA
its lifetime, minus the present value of all the (Fig. 4) and a COE is 102.83 DA, the battery takes
revenue it earns over its lifetime. Costs include first place, with a cost of 60 00000,00 DA, the
capital costs, replacement costs, O&M costs, fuel second place the control cost (Combined Dispatch
costs, emissions penalties, and the costs of buying CD) 160 000.00 DA; the PV system with a cost of
power from the grid. Revenues include salvage 55 084.32 DA. The last place converter with a cost
value and grid sales revenue. HOMER calculates of 556.12 DA. Economically, Fig. 5 gives the results
the total NPC by summing the total discounted cash
flows in each year of the project lifetime. The total
NPC is HOMER's main economic output, the value
by which it ranks all system configurations in the
optimization results, and the basis from which it Fig. 5 Economic results.
calculates the total annualized cost and the levelized
cost of energy. To calculate the COE, HOMER Different parameters originating from the simulation
divides the annualized cost of producing electricity are obtained: the present and annual worths are 465
(the total annualized cost minus the cost of serving € and 14€ respectively, the return on investment is ‒
the thermal load) by the total electric load served, 1.4%, internal rate of return (%) is n/a, simple and
using the following equation (1): discounted payback (a year) are respectively 13.50
and 13.39. Results of electrical production and
consumption are listed in Table VI, below :
TABLES VI. OUTPUT PARAMETERS
Production kWh/y %
Solar World 250 SW 250 Mono 221 100
Where, Cann, tot , Cboiler , Hserved and Eseved are the Total 221 100
total annualized and boiler marginal costs, and total Consumption kWh/y %
AC Primary Load 147 100 In this paper a solution to the lack of electricity
DC Primary Load 0 0 and energy conversion technologies is proposed.
Total 147 100 HOMER was found to be a very helpful tool for the
Quantity kWh/y %
microgrid planning and dispatching, so that
Excess Electricity 45.4 20.5 distributed sources of energy. Analysis reveal that
Unmet Electric Load 0.0342 0.0232 although a fully renewable-based microgrid, which
Capacity Shortage 0.0589 0.0400 has no carbon footprint, is the most preferred, the
net present cost (NPC) is higher. It is worth to
Quantity Value mention that when the microgrid is connected to the
Renewable Fraction 100 external grid, it is the most economically favorable
Max. Renew. Renetration 1089 option. Also, the most environmentally friendly
microgrid is the renewable energy microgrid, and
The mean monthly electricity production of our PV there is still much work to be done in terms of
system is shown in Fig. 5 and Tables VII, below. renewable energy and mixed system development,
because of their high initial capital and replacement
costs. Allowing a small amount of annual load to be
left unmet makes the microgrid more cost-effective.
Also, the break-even distance presented in this work
shows that for isolated microgrids, far away from
the external grid connectivity point, the mixed
microgrid, is the most economic optimal choice. In
this paper we analysed a hybrid system that can be
Fig. 5 Average monthly electricity production of the PV system.
used as a guideline on designing and implementing
grid-connected PV-based power systems in Algeria.
TABLES VII. OTHER OUTPUT PARAMETERS
Similar approaches could be applied to other regions
Capacity-based metrics Value Units
Nominal Renewable capacity divided total 100 % in this country, so that optimum results might be
nominal capacity different from the current ones depending on the
Usable Renewable capacity divided total 100 % local data.
capacity
REFERENCES
Energy-based metrics Value Units [1] http://magazine.ieee-pes.org/latest issue May/June 2019.
Total Renewable production divided by 150 % [2] V. Fthenakis, Point of View, Vol. 103, No. 3, March 2015 Proceedings
of the IEEE.
load
[3] I. M. Kirpichnikova, A. A. Maliugina, 2nd International ICIEAM
Total Renewable production divided by 100 % Conference, Chelyabinsk, Rev. Roum. Sci. Techn. – Électrotechn. Et
generation Énerg., doi: 10.1109/ICIEAM.2016.7911464, pp. 1–3 (2016).
One minus total nonrenewable production 100 % [4] K. Attari, A. Elyaakoubi, A. Asselman, Energy Reports, 2, 261-266
divided by load (November 2016).
[5] M. R. Maghami, H. Hizam, C. Gomes, M. A. Radzi, M. I. Radzi, S.
Hajighorbani, Renewable and Sustainable Energy Reviews, 59, 1307-
Peak values Value Units 1316 (June 2016).
Renewable output divided by load ( Homer 150 % [6] Cherechi Ndukwe, Tariq Iqbal “Sizing and dynamic modelling and
standard) simulation of a standalone PV based DC microgrid with battery storage
Renewable output divided by total 100 % system for a remote community in Nigeria” , Journal of Energy Systems,
Volume 3, Issue 2, Published:30.06.2019, pp 67-85
generation
[7] G. Alamri and T. Iqbal, “Sizing of a hybrid power system for a house in
One minus nonrenewable output divided by 100 % Libya,” in 2016 IEEE 7th Annual Information Technology, Electronics
total load and Mobile Communication Conference (IEMCON), pp. 1–6,
Vancouver, BC, Canada, 2016. View at Publisher · View at Google
Scholar · View at Scopus
[8] MD. NURUNNABI , NARUTTAM KUMAR ROY, EKLAS HOSSAIN
IV. CONCLUSIONS , HEMANSHU ROY POTA “Size Optimization and Sensitivity
Analysis of Hybrid Wind/PV Micro-Grids- A Case Study for
Bangladesh” Received September 12, 2019, accepted September 27,
2019, date of publication October 7, 2019, date of current version
boudjabi.af@gmail.com
*
ITheMM, SFR Condorcet FR CNRS 3417, University of Reims Champagne Ardenne
51687 Reims, Cedex 2, France
2
Chadi.maalouf@univ-reims.fr
Abstract—In this study, a two dimensional heat and mass m/s, channel gap less than 5 mm, channel height larger than 1
transfer numerical simulation based on species transport and m and working air ratio around 35–60%, a wet bulb
discrete phase model on CDF code Ansys/Fluent is presented for effectiveness greater than 100% is obtained. Lee et al. [5]
a dew point evaporative cooler. The air supply temperature have used a counter-current system with a wet return surface;
along the channel obtained is lower than previous numerical
they achieved the model using an aluminum plate supporting
data by 3.39% and 3.45% for the exhaust air temperature, while
the moisture content and relative humidity give a discrepancy of a wet media to activate heat transfer. For an air inlet
1.35% and 2.28% respectively. The parametric analysis allowed temperature of 30°C and a relative humidity of 50%, they
examining the effect of working air ratio where wet bulb found an air supply temperature of 22°C, 1.7 ° C lower than
effectiveness was greater than 120% for ratios up to 0.7 and the the wet bulb air temperature. A 3D numerical study of
dew point effectiveness reached more than 80% for this values. laminar airflow of a counter-current evaporative cooler has
For an increase of 15 °C in air inlet temperatures, the wet bulb also been proposed by Pengfei Liu and Xinyu Li [6]. The
effectiveness and the dew point effectiveness dropped of about authors used Ansys software to simulate a 200 mm length
12.69% and 25.26 %. system: a dry channel for the product air, a wet channel and a
dry channel for the working air with holes on the wet return
Keywords— dew point evaporative cooler, dry/wet channel, surface. They obtained a dew point efficiency of 71.2%. In a
product air, working air, wet bulb/dew point efficiency. 3D numerical simulation using, C++ language, Hamoon
Jafarian et al. [7] considered a modified boundary condition
I. INTRODUCTION on the separating wall surface (symmetry condition). Theirs
Nowadays, evaporative coolers’ investigations are trying results predicted the prototype performance to 3.53% and
to optimize the design of the cooler in order to reach lower air the authors deduced a maximum error between 2D and 3D
temperatures. Indeed, the usual evaporative systems (direct models less than 4.5%.
and indirect) have some disadvantages related to the humidity In an experimental and numerical work [8], the authors used a
for the direct systems and the limited efficiency for the new method based on NTU-LE-R model and a 2D numerical
indirect ones. In both cases, the supply air temperature cannot simulation on software COMSOL for a vertical dew point
be lower than the wet bulb temperature of the air [1]. In hot evaporative cooler. The discrepancy was about 6% between
and dry climate regions, the dew point evaporative coolers experimental and numerical data. Guangya Zoo et al. [9]
can be energy efficient and even with a low environmental identified the influential parameters on the Maisotsenko cycle
impact [2]. Ala Hasan [3] has carried out a 1D numerical exchanger and proposed a dew point effectiveness correlation.
study. He has considered four patterns of evaporative cooler Compared with numerical simulations and experimental data,
with one and two stages, co and counter current and cross the deviations were 2.35% and 6.75% respectively. Ali Pakali
flow systems. The author has obtained 116% wet bulb et al. [10] examinated the cooler using a 1D model and a 3D
efficiency for the case of one indirect stage with a wet return simulation with COMSOL multiphysics, which agree well
surface. with the experimental results within 8.5% and 10%
In another paper, Riangvilaikul and Kumar [4] have respectively.
conducted a numerical and experimental study of a vertical The aim of this paper is to investigate numerically a dew
configuration of dew-point evaporative system for different point evaporative cooler for the Mediterranean climate
conditions. Their results show that at: air velocity below 2.5 applicability. A 2D numerical model is developed in CFD
code ANSYS/ fluent and the results are discussed with - Steady –state of the airflow and heat and mass
reference to other works in the literature. transfer ;
- The water injection temperature is constant;
II. SYSTEM DESCRIPTION - Laminar airflow is in the longitudinal direction;
The counter current flow evaporative cooler has two - Air and water properties are set uniform.
channels separated by a return wet surface. A working air
ratio varying between 30 and 80% is applied. The product air B Mesh analysis
supplies the building as shown on Fig.1. The dry and wet A structured mesh with quadrilateral elements is generated.
channels have a gap H of 3.5 mm each, a system length of A refinement method is applied around the interfacial surface
500 mm and a plate of aluminium covered of water film (0.5 and in the return region. Several size elements of mesh were
mm height) is used for the heat and mass transfer between the tested in order to reach grid independence results: 42500,
dry product air and the wet working air. 54600, 66700, 90250, 152300 and finally 260400 cells. From
the mesh repartition of 66700 to 260400 elements, the results
were very close, therefore, the 90250 elements’ mesh was
Working air in the wet channel
chosen for the rest of the simulation.
H
Product air in the dry channel
III. MODEL VALIDATION
L1
A two dimensional numerical model is considered for a
L2
dew point evaporative cooler similar to the system used in
Hasan [3] and Maalouf et al. [11] papers. Table I shows the
Fig. 1 The dew point evaporative cooler design influential parameters of the airflow in the device.
TABLE I
COOLER PARAMETERS
A Mathematical model
Parameter Hasan[3] Current case
A numerical simulation of two-dimensionnal configuration Channel length (m) 0.5 0.5
is released using the commercial CFD code ANSYS/ fluent. Channel gap (m) 0.0035 0.0035
The transport equations of the air flow are resolved by finite Air inlet velocity (m/s) 0.676 0.676
volume method. The airflow is supposed laminar because of Working air ratio (-) 0.7 0.7, 0.2
the very thin gap compared to length [7,10]. 0.3, 0.5,0.8
For the water evaporation model, the species transport Air inlet temperature (°C) 30 30, 35, 40, 45
equation and the discrete phase model are used. The Inlet moisture content (g/kg air) 9 9
governing equations are:
Fig.3 shows the streamlines in the two channels. It
- The Continuity equation
illustrates a laminar air flow without recirculation zones.
. u 0 (1)
Where Twb and Tdp are air wet bulb and dew point
temperatures and wb, dp are wet bulb and dew point
effectiveness respectively.
ACKNOWLEDGMENT
This work is the result of a scientific cooperation between
France and Algeria. Authors wish to acknowledge the support
of the Hubert Curien Partnership (PHC TASSILI).
REFERENCES
[1] J.R. Camargo, C.D. Ebinumca, J.L Silveira.”Thermoeconomic
Fig.7 Effect of working air ratio on cooling effectiveness and air supply analysis of an evaporative dissicant air conditioning system”. Applied
temperature Thermal Enginnering, vol. 23 , pp. 1537-1549, 2003.
[2] N. Lechner, heating, Cooling, Lighting, Sustainable Design Methods
for Architects, 3rd ed. New jersey, U.S.A. Wiley, 2009.
B Effect of the air inlet temperature [3] A. Hasan. “ Indirect evaporative cooling of air to a sub-wet bulb
In the second analysis, the Fig. 8 illustrates the effect of temperature”, Appl. Therm. Eng., vol. 30, pp. 2460-2468, 2010.
[4] B. Riangvilaikul, S. Kumar. “ An experimental study of a novel dew
air inlet temperature on the cooler performance. The air outlet
point evaporative cooling system”. Energy Buid., vol. 42, pp. 637-644,
temperature increases from 17 °C to 26° when the air inlet 2010.
temperature rises of 15°C, so, the cooler loses about 12.7% [5] Pengfei Liu, Xinyu Li. Simulation on Three Dimensional Laminar
of its wet bulb effectiveness and 25.26 % of its dew point Properties of Dew Point Evaporative Cooling in Plate Heat Exchanger.
3rd International Conference on Mechanical Engineering and
effectiveness. These results are in good agreement with the
Intelligent Systems, ICMEIS, Singapore , 2015.
literature. [6] J. Lee, D.Y Lee. “Experimental study of a counter flow regenerative
evaporative cooler with finned channels”. Int.J.Heat Mass Transf., vol.
65 , pp. 173-179, 2013..
[7] Hamoon Jafarian, Hoseyn Sayyaadi, Farschad Torabi. “A numerical
model for a dew point counter-flowindirect evaporative cooler using a
modified boundary condition and considering effects of entrance
regions”. International Journal of Refrigeration, vol. 84, pp. 33-51,
2017.
[8] Yangda Wan, Jie Lin, Kian Jon Chua, Chengqin Ren.” A new method
for prediction and analysis of heat and mass transfer in the ounter-flow
dew point evaporative cooler under diverse climatic operating and
geometric conditions”. International Journal of heat and Mass transfer,
vol. 127, pp. 1147-1160, 2018.
[9] Guangya Zhu, Weiwei Chen, Shihua Lu.”Modelling of a dew point
effectiveness correlation for Maisotsenko cycle heat and mass
transfer”. Chemical Engineering & Processing: Process Intensification,
vol. 145, 107665, 2019.
[10] Ali Pakari, Saud Ghani.” Compari_son of 1D and 3D heat and mass
transfer models of a counterflow dew point evaporative cooling
Fig. 8 Effect of air inlet temperature on the air supply temperature and system: Numerical and experimental study”. Inetrnational Journal
cooler effectiveness. of refrigeration, vol. 99, pp. 114-125, 2019.
[11] Chadi maalouf, Tala Moussa, Amel Ferial Boudjabi, Djallel Abada,
Guillaume Polidori, Saffidine Djamila Rouag, Enernest Wultz.
“ Numerical study and design of a dew point evaporative cooler for
V. CONCLUSION buildings”. Conference EENVIRO 2018, Cluj Napoca, Romania,
This paper investigated numerically the thermal October 2018.
performance of a dew point cooler. The results are validated
by previous numerical data. It is found that a working intake
air ratio greater than 50% of the total inlet airflow, is more
efficient giving a wet bulb effectiveness above 100% and a
dew point effectiveness close to 80%. Considering outdoor
temperature, the system performance decline is not very
important for high temperatures which encourages using the
Abstract— Industrial heavy duty gas turbines (HDGT) are adopted GT model, there are two main quantities that
specially designed for power generation. They are specified with determine the turbine behavior, namely the output torque and
long life and higher reliability compared with other types of gas
the exhaust gas temperature [4, 5, 8,9]. These quantities come
turbines. Therefore, modeling and simulation of their behaviors,
under realistic operating conditions, play a key role for efficient along with two control loops: the speed governor control loop
design as well as reliable manufacturing practice. In this aspect, and the exhaust temperature control loop. The current study
several gas turbine models have been developed based on the focuses on two considered basic models: Rowen’s Model and
different parts of heavy duty gas turbine as well as their Frequency Dependent (FD) Model including time delays.
important functions in the gas turbine plants: the speed These time delays are introduced in feedback control loop.
governor, the fuel system, the combustion chamber, temperature They are articulated by exponential functions e-αs. In this case
control and the turbine. Recently there has been growing interest
in the study of heavy duty gas turbines models with different
study, a 265 MW single shaft HDGT is considered. For this
degrees of difficulty and merit. This paper presents modeling and purpose, brief descriptions of the above mentioned models are
analysis of the behavior of heavy duty gas turbine to understand given, along with the predicted simulation results. In order to
the power system problems. This is achieved by a complementary investigate the model performance at rated load, the
and comparative analysis of the response of different dynamic considered HDGT model is simulated under nominal
models published in different journals. In this paper, two models conditions operation without acceleration control loop. Speed
have been chosen using Matlab/Simulink simulation. The
reference is equal to one pu during all the simulations. In
frequency dependant model (FD) and the modified Rowen’s
model of heavy duty gas turbines have studied taking into order to apply the same conditions of simulation as Rowen’s
consideration the effectiveness and accuracy. In order to apply model, simple time delays are introduced in the FD model.
the same conditions of simulation as Rowen’s model, simple Three different situations of simulation are studied: a turbine
time delays are integrated in the FD model in our approach. step speed response, -3% turbine speed deviation, and 20%
load step increase. The parameters of 265 MW HDGT using
Keywords— Gas Turbine; Modeling and simulation; Mathematical
in the simulation models are defined in [6].
modeling; Electric Power plants.
II. ROWEN’S MODEL
I. INTRODUCTION
In 1983, W.I.Rowen provided a model of gas turbine that
Due to their specific characteristics, gas turbines (GT) are can be used for dynamic performance studies of power system
adopted worldwide for power generation, especially in [5, 6, 9, 10]. Rowen’s model consists of a set of algebraic
countries endowed with huge natural gas resources. On the equations describing the steady-state characteristics of gas
other hand, owing to their longer life cycle coupled with a turbine thermo-dynamics, simple time delays, and a few
higher availability factor, heavy duty gas turbines (HDGT) are related controls including temperature control, governor, and
well targeted for power generation on industrial scale. acceleration control. The model is suitable for a wide range of
However, an accuracy study in modeling and simulation of the ambient temperatures, and the influence of axial flow
gas turbines is necessary because of their dynamics which is compressor variable inlet guide vanes (IGVs) is included in
the most important in a power plant. A great number of the models as appropriate to the actual machinery
studies [1-9] has been published on gas turbines model. configuration [8-11].
Although there are no guidelines regarding the model The control system includes speed control, temperature
construction methodology, most of the published models stem control, axial flow compressor inlet guide vane control,
from the Rowen’s seminal model. This GT model account for and upper and lower fuel limits [11, 12]. The IGVs
the size as well as the characteristic features of the underlying controller is affected by exhaust gas temperature, while
power system as reported in [3]. Nonetheless, whatever the the minimum value of temperature and load / frequency
control leads to generation of the fuel demand signal. The the output power of the turbine and ambient temperature [9,
exhaust flow is proportional to a slightly non-linear function 10, 12]. As mentioned in [5, 15], a new formulation of the
of inlet guide vanes position multiplied by the rotor speed and gas turbine thermodynamic characteristics has been developed
corrected for ambient temperature [13, 14]. Exhaust flow is in order to reveal the effects of shaft speed. This model is
provided as an output from the model for use in heat recovery based on mathematical equations which are deduced from
system simulations. The duty of IGVs is to keep the output physical principles. The structure of the FD model, is given in
temperature at a design point which is held at a more or less [5, 14, 16].
constant value by regulating the flow of air mass into the
A. Air flow equations
compressor. The IGVs are in operation only during the startup
𝑠𝑖𝑛 (𝜃 𝐼𝐺𝑉 −𝜃0 )
and shut down of a simple operating gas turbine without any 𝑊𝑎 = 𝑞 𝑇𝑎, 𝑃𝑎 𝑈 ∆𝑁
heat recovery [5, 11,13]. Rowen’s model equations are given
𝑠𝑖𝑛 𝜃 𝑀𝑎𝑥 −𝜃0 (4)
𝑃𝑎
as follows: 𝑞 𝑇𝑎, 𝑃𝑎 = 𝑇𝑎0/𝑇𝑎
𝑃𝑎0 (5)
A. Turbine exhaust temperature calculation
1
𝑓1 = 𝑇𝑋 = [(𝑇𝑅 − 453 ( 𝑁 2 − 4.21𝑁 + 𝑈(∆𝑁) = 1 + ( 𝐴0 ∆𝑁) + (𝐴1 ∆𝑁 2 ) + (𝐴2 ∆𝑁 3 ) (6)
1+0.0050 15−𝑇𝑎
4.42 ) 0.82 ( 1 − 𝑊𝐹 )) + 722 ( 1 − 𝑁) + 1.94 ( 𝑀𝑎𝑥𝐼𝐺𝑉 −
𝐼𝐺𝑉 )] (1)
∆𝑁 = 𝑁( 𝑇𝑎0 /𝑇𝑎) − 1 (7)
B. Turbine torque calculation where: Ta=288K: The ambient temperature
𝑓2 = [1.16 𝑊𝑓 − 0.133 ]/𝑁 (2) Pa=1 atmosphere
C. Turbine exhaust flow calculation
519
𝑓3 = Wx = N (Ligv 0.257 ) (3) B. Compressor Pressure Ratio (CPR) equation
𝑇𝑎 +460
𝑃𝑎 0
where: 𝐶𝑃𝑅 = 𝐴5 𝑊𝑎 + 𝐴6 𝑊𝑓 + 𝐴7
𝑃𝑎 (8)
Ligv = Per Unit Inlet Guide Vane Actuator Stroke C. Exhaust temperature Equation
IGV = Actual Inlet Guide Vane Angle - Circular Degrees 𝐴3
𝑃𝑎
+𝐴4 𝑊 𝑓
𝑃𝑎0
Ta = Site Ambient Temperature (F) 𝑇𝑥 = 𝑇𝑎 +
𝑊𝑎 (9)
III. FREQUENCY DEPENDENT MODEL (FD MODEL)
The FD model presents a process that may enable
parameter estimation of the frequency dependency from
𝑃=
1−𝑊 𝑓0 (11) 700
FD Model
where, 600 Rowen's Model
500
turbine thermodynamics can be estimated from loading test Fig.3.The behavior of Exhaust temperature (step response)
data in the steady state characteristics, except for their
frequency dependency [14, 17]. 1.2
Rowen's Model
FD Model
Air flow may be regulated by Inlet Guide Vanes (IGVs) 1
1000
FD Model
1.4
Rowen's Model FD Model
900
Rowen's Model
800 1.2
700
1
Mesured Exhaust Temperature(K)
850
600
700
400 0.6
100 150
300
0.4
200
0.2
100
0 0
-100
0 20 40 60 80 100 120 140 160 180 200 -0.2
0 20 40 60 80 100 120 140 160 180 200
Time(sec)
Time(sec)
Fig.5. Exhaust temperature, for a droop speed by Fig.8. The behavior of the output power, for a 20% step
3% at 100 sec. increase of load at 100 sec.
1.6
As presented in the figures, the Rowen’s model dynamic
FD Model
Rowen's Model
responses are highly affected by each situation. First, we
1.4 noted that the exhaust temperatures and also the mechanical
output power follow the rated values, which are 783° K (we
noted 784° K ) and 1pu, respectively. In the second and third
1.2
600 800
and the IGV position are not activated during the simulation.
However, an increase of 20% in load would cause the
750
500
400
700
temperature control to be activated. Besides, it can be noted
100 150
300
that the mechanical output power remains around 1 pu.
200 V. CONCLUSION
100 The main aim of this paper is to present an overview of
0
existing gas turbine models and explore the relationship
between Rowen’s model and FD model. As mentioned earlier,
-100
0 20 40 60 80 100 120 140 160 180 200 simple time delays are included in the FD model, in order to
apply the same conditions of simulation as Rowen’s model.
Time(sec)
obtained simulations results highlight the usefulness and Transactions on Power Systems, Vol. 23, No. 1, 0885-8950, February
2008.
accuracy of the Rowen’s model compared to the frequency [9] M. R. Bank Tavakoli, B. Vahidi, and W. Gawlik, “An educational guide
dependent model that cannot be well tuned for a large amount to extract the parameters of heavy duty gas turbines model in dynamic
of disturbances. However, Rowen’s model, owing to its studies based on operational data”. IEEE Transactions on Power
Systems, Vol. 24, No. 3, 2009.
simplicity and accuracy, has been widely utilized for diverse [10] B. Vahidi, M. R. Bank Tavakoli, and W. Gawlik, “Determining
purposes and studies. parameters of turbine’s model using heat balance data of steam power
unit for educational purposes”. IEEE Transactions on Power Systems,
REFERENCES vol. 22, no. 4, pp. 1547–1553, 2007.
[11] H. Emam Shalan, M. A. Moustafa Hassan, a. B. G. Bahgat, “Parameter
[1] Khormali, I. Yousefi, H. Yahyaei, Sh. M. Aliyari. "Identification of an
estimation and dynamic simulation of gas turbine model in combined
industrial gas turbine based on Rowen's model and using multi-objective
cycle power plan based on actual operational data”. Journal of American
optimization method", International Conference On Robotics And
Science, 7(5), pp.303-310, 2011.
Mechatronics (ICROM), 2015.
[12] L. Meegahapola, D. Flunn, “Gas turbine modelling for power system
[2] N. Kakimoto, K. Baba, “Performance of gas turbine-based plants during
dynamic simulation studies”, Power Factory Applications for Power
frequency drops”. IEEE Transactions on Power Systems, vol. 18, No 3,
System Analysis, pp 175-195, 2014.
August 2003.
[13] S. O. Oyedepo, O. Kilanko, “ Thermodynamic analysis of a gas turbine
[3] L. Meegahapola, "Characterisation of gas turbine dynamics during
power plant modeled with an evaporative cooler”, International Journal
frequency excursions in power networks", IET Generation,
of Thermodynamics, vol. 17 (no. 1), pp. 14-20, 2014.
Transmission and Distribution, Vol. 8, (10) pp. 1733-1743, 2014.
[14] K. Kunitomi, A. Kurita, Y. Tada, S. Ihara, W. W. Price, l. M.
[4] J. Mantzaris, C. Vournas, “Modeling and stability of a single-shaft
Richardson, and g. Smith, “modeling combined-cycle power plant for
combined cycle power plant”. International Journal of Thermodynamics
simulation of frequency excursions”. IEEE Transactions on Power
Vol. 10, No. 2, pp. 1-9, June 2007.
Systems, vol. 18, no. 2, pp. 724–729, 2003.
[5] W.I ROWEN, “Simplified mathematical representation of heavy duty
[15] W. I. Rowen, “Simplified mathematical representations of single shaft
gas turbines”. Journal of Power, vol 105, pp. 865-869, October 1983.
gas turbines in mechanical drive service”. International Gas Turbine
[6] H. AsgarI, X. Q. Chen, R. Sainudiin , “ Modelling and simulation of
and Aeroengine Congress and Exposition Cologne, Germany, 1992.
gas turbines”, International Journal of Modelling, Identification and
[16] K. Kunitomi, A. Kurita, H. Okamoto, Y. Tada, S. Ihara, P. Pourbeik,
Control, Vol. 20, No.3, pp. 253-270,2013.
and W. W. Price, “ Modeling frequency dependency of gas turbine
[7] M. Khalilpour, K. Valipour, H. Shayeghi, N. Razmjooy, “Designing a
output”. IEEE Power. Eng. Soc, winter meeting, Columbus, vol. 2,
robust and adaptive PID controller for gas turbine connected to the
2001.
generator”, Research Journal of Applied Sciences, Engineering And
[17] H. E. Shalan, M. A. Moustafa hassan, A. B. G. Bahgat, “comparative
Technology, Vol. 5, N°. 5, pp. 1544-1551, 2013
study on modeling of gas turbines in combined cycle power plants”
[8] S. K. Yee, J. V. Milanovic, F. M. Hughes, “Overview and comparative
Paper ID 317, December 2010.
analysis of gas turbine models for system stability studies”, IEEE
Abstract— Mixed oxides such as SiO2-TiO2 were used by K. Fe2O3 / V2O5 have been deposited on glass and on silicon,
Chhor and all the V2O5-Fe2O3 in the photocatalytic degradation the thickness varies between 20 and 110 nm, measured by the
of phenol and salicylic acid as a function of the calcination profilometer D600
temperature [01]. The materials were prepared by sol-gel. The
amount of hydroxyl present on the surface of the materials the UV-Vis spectra of the samples indicate that in the near-UV
strongly modifies the adsorption of the molecular oxygen (and region (λ = 300-800 nm) the light reflectance spectrum of the
thus modifies the photocatalytic activity. modified samples tends to move towards the region of the
longest lengths of the light wave and the trend is very big. The
The sol-gel process has been widely used for depositing metal spectral lines of the specimens have an absorbance effect on
oxide films because of its many advantages, such as its low cost, the UV-Vis spectrum and a slightly less light absorption in
large area deposition and the ease of metal doping. shorter wavelength. This means that the samples are also able
Keywords : Mixed oxides, sol-gel, FTR, spectrometry UV.IR .
to absorb photons in the visible region and have lower energy
values of the band gap. However, the light absorption capacity
in the Visible and Ultraviolet region of the samples is 85% and
I. INTRODUCTION for samples that have undergone the heat treatment (heated
Generally, deposits of iron oxides and vanadium oxides are solution) at 100 °C is the same , and also for porous silicon
done by the electrochemistry or immersion coating and by Spin sample. Between (λ800 – 2500) nm, it is found that the
coating; the deposition is followed by annealing in an oven at a reflectance is almost 60% identical for the sample treated at
temperature of 300 to 600 ° C for several hours in order to 100 ° C and also for the porous silicon sample, we observed the
crystallize the layer. disappearance of interferences in porous silicon subtract due to
the deposited layer.
We already know that some types of materials (substrate)
are too sensitive to any temperature variation. Our work for the layer treated at 150 ° C a slight increase in the
consists of producing iron oxide and vanadium oxide layers, reflectance and decrease in absorbance in some UV regions
using the in situ dip coating (gel sol) technique by immersing and visible at 60% , for the IR region the reflectance decreases
the substrate in the solution for few second. The solution is and the absorbance increases at an average of 65% for the
heated at temperature of 100 and 150 and, which gave Fe2O3 / V2O5 / PSi structure
satisfactory results.
Figure. 2 avec le spectre de transmittance pour les structures We used the Fourier transform infrared spectroscopy "Thermo-
réalisées sur verre, nous pouvons calculer la bande interdite. La Nicolet", equipped with a detector DTGM KBr. All the spectra
progression de la bande interdite de 3,23 ev a été estimée pour were collected in a transmission mode in the area of the
l'échantillon non chauffé, les échantillons de 3,66 eV chauffés average infrared 4000-400 cm. The FTIR spectrum shown in
à 100 ° C et de 3,95 eV pour l'échantillon chauffé à 150 ° C. fig.5 represents the ferrite Fe2O4 with characteristic peak of
F2O3 at 527cm-1 and 601.25 cm-1, attributed to the Fe O bond
In FIG. 4 (a), (c) and (e) (d)) the morphology of the two layers modes.
of F2O3 and V2O5 of the sample is clearly observed with a
porous and granular layer (a) (c) and nanoparticles which is in The band centered at 3471cm-1 is characterized to the OH
the form of agglomerates (d) (c). These layers deposited vibration mode,
successively by dip coating technique (solution heated to 150 The band located at 1351.91 cm-1 corresponds to the
°C)., Fig 4 (b) Note an opaque and granulated layer of the deformation movement in the plane Fe H
sample solution heated to100°
The peak at about 1001 cm-1 is related to the stretching
band of Si- O- Si.
In the band in the range 990-1270 cm-1 is attributed to the
elongation vibrations of the Si-O-Si band, confirming the
formation of an oxide on the surface.
The elongation vibrations of the Si-H, Si-H2, Si-H3 bands
are centered around 2083cm -1.
.
(a)
(b)
(*310)
(* 220)
Fe2O3 structure Fe2O3/V2O5 / PSi heated to 150 ° C
250 Fe2O3
(* 221)
Fe2O3 V2O5
( #710)
Fe2O3
Fe2O3
( #711)
200
(#302)
V2O5
(* 410)
(* 311)
Intensité (u, a)
(* 330)
(* 410)
V2O5
(002)
(*510)
150
Fe2O3 (121)
100
50
(c) 25 30 35 40 45 50 55 60 65 70
0
2 ( degré)
REFERENCES
Abstract— This article presents the development of an optimal remains limited and ineffective as it has not solved the
lighting system for greenhouses, allowing uniform illumination problem of data recording and management as mentioned by
over the entire greenhouse surface and meeting the plants' needs [10-12].
in terms of PFFD. Our work aims to develop a control system Nowadays greenhouse control is done through the Internet
that is based on the use of a light sensor, a dimmable luminaire,
of Things (IoT).
a RASPBERRY PI 3 and a web application. The brightness
management is done in an autonomous and/or manual way The IoT is essentially composed of three main
according to the user's choice, the automatic control uses the technologies:
data from the lighting sensors, However, the user can opt for the (1) a sensor for data acquisition,
manual control to vary the luminous intensity of the luminaires. (2) a network for data communication between the
The system was designed, prototyped, and tested in Tunis connected objects and the computer.
province of Tunisia. Our contribution through this work (3) the computer that ensures the analysis and intelligent
represents the management of agriculture through digital processing of the various data to convert them into control
innovation (IoT) to optimize the lighting energy used in actions.
greenhouses through the use of the Daylight system control
The IoT has covered almost all areas of agriculture
technique and the combination of biological parameters that
define tomato lighting requirements and DLI to analyze the (irrigation, temperature control, humidity control, etc.).
measured data to choose the best control mode. All this has In irrigation, a system has been created to control the water
allowed us not only to supervise, control and reduce the energy pumps when the humidity level decreases in relation to the
cost of lighting, but also to improve the productivity and quality user's set point by using sensors for reading the humidity,
of tomatoes. relays for activating and deactivating the pumps [13].
Keywords— IOT; LED luminaire; Raspberry Pi; In temperature and humidity control, [14] has developed a
Smart Greenhouse; Web application. system using an Arduino board capable of monitoring and
controlling the ambient humidity, soil moisture, and
I. INTRODUCTION temperature of a greenhouse in the Valeriana jatamansi area.
Digitation and the technical evolution of systems have In the control and monitoring of air quality and CO2 [15]
made very remarkable progress during this decade, they have has implemented a system that automatically monitors and
quickly penetrated into our daily lives as well as into several controls air quality parameters and co2 inside greenhouses.
sectors of activity such as industrial activities, transport, All these systems have improved the efficiency of
medical, agriculture, etc... greenhouse control and monitoring, however, there are a few
Especially in the field of greenhouse agriculture, digitation parameters that haven’t been considered in the works of the
has enabled its users to improve the productivity and majority of author and which has, directly and indirectly,
profitability of their systems. impact on the energy balance of the greenhouse and which
Greenhouses are traditionally controlled by ON/OFF require more emphasis such as lighting, we have shown in a
buttons through the control panel which is usually placed previous article[16] that this parameter is essential to ensure
inside the greenhouse however this traditional method is not optimal plant growth and that we can make significant energy
efficient and requires the presence of the user to perform the savings if we manage to optimise the monitoring and control
control. of this factor and this is the subject of our study and
The technology subsequently evolved through the use of contribution.
time-delay systems [1-3] but unfortunately, this method does
not provide accurate, reliable and efficient control, nor does it
II. METHODOLOGY
allow data to be recorded in order to be able to analyse them
and detect more effective control methods. A. System Overview
The Greenhouse Control Mode underwent a radical We aim to design and implement a lighting control system
evolution a few years later using programmable controllers or using sensors spread over nine different locations in the
microcontrollers connected to sensors for data acquisition and greenhouse, to collect the different levels of illumination and
has relays for control [4-9]. Unfortunately, this method
transmit them to the RASPBERRY PI who will process this system, the interaction activity diagram of the control unit
data and the stocks in the web application database. /database is presented as shown in Figure.2.
The farmer will then be able to handle the greenhouse
remotely through a control panel designated to communicate
with the RASPBERRY PI, the control panel is a web
application accessible via the Internet or via the local network
of the greenhouse that offers several features for better
accessibility to the greenhouse,
we can present an overview of our system as shown in
Figure.1.
Interface I2C
Figure.15. Evolution of the measured DLI from 7h:00 to 12h:15 for the day
15-06-2019
greenhouse, altitude, latitude, etc. This difference has a direct The control of the luminaires allowed us to have a
effect on the consumption of fixtures and indirectly on the uniform illumination and distributed in the same
energy balance of the greenhouse as we have shown in a way in order to guarantee a better quality of
previous article [16]. This difference has a direct effect on the product and which has the same growth.
consumption of luminaires and an indirect effect on the Our system, allows us to control our greenhouse
greenhouse balance energy as we have shown in a previous manually and/or automatically, it eventually
article [16]. To compare the daily consumption of the lighting solved the data logging problem which will allow
system by adapting traditional lighting control systems with us to analyze the data in real-time and gives us the
our IoT control system, opportunity to improve our system in a
Traditional control systems generally use timers that adopt continuous way.
one of these two control strategies (control scenario 1 and Our contribution on the reduction of the energy
(control scenario 2) respectively: consumption of luminaires and hours of operation
Light one hour before the sunset until midnight [17] i.e. has allowed us to reduce the consumption of 12%
from 6:30 p.m. until midnight, which means a total of 5h:30 from one zone to another (zone 5 and zone 6
of additional lighting, Versus zone 2) and 58% compared to traditional
Light 4-5 hours of sunset until 1h after the sunrise [17], i.e. control mode (Zone 6 Vs control scenario 2).
from 23:30 to 5:30 am, which mean a total of 6h:30 of
additional lighting, ACKNOWLEDGEMENTS
In order to deduce the energy gain, we have superimposed I gratefully acknowledge the support and generosity of
the consumption curves relating to the traditional control BBL institute, without which the present study could not have
modes 1 and 2 as well as the groups of luminaires associated been completed.
with the sensor 2 and 6 as represented in figure.16.
REFERENCES
[1] O. J. Smith. (1957) Closer Control of Loops with Dead Time.
Chemistry Engineering Progress, Vol. 53, No. 5, pp. 217-219.
[2] MARSHALL, J.E. (1979) The Control of Time-Delay Systems, Peter
Peregrinus, I.E.E. Control Engineering Series.
[3] Marshall J.E. (1980) Identification strategies for time-delay systems.
In: Unbehauen H. (eds) Methods and Applications in Adaptive
Control. Lecture Notes in Control and Information Sciences, vol 24.
Springer, Berlin, Heidelberg.
[4] Ventzas, D. (1998). MC68HC11 Microcontroller Control of a
Greenhouse Environment as a Time Delay System. IFAC Proceedings
Volumes, 31(12), pp.239-244.
[5] van Straten, G. (1998). Acceptance of Optimal Operation and Control
Methods for Greenhouse Cultivation. IFAC Proceedings Volumes,
31(12), pp.11-18.
Figure.16. Ligting Power consumption [%] [6] Young, P., Price, L., Berckmans, D. and Janssens, K. (1998). Recent
According to Figure 16, we notice that the group of Developments in the Modelling and Control of Climate and
luminaires associated with sensor 6 consumes less than the Ventilation in Agricultural Buildings. IFAC Proceedings Volumes,
31(12), pp.3-10.
“1/5” of the energy required compared to the groups of [7] Hampel, J. and Wilkie, D. (1980). A MICROPROCESSOR BASED
luminaires associated with the sensor 2, we also note that PROCESS CONTROLLER FOR INTERACTIVE CONTROL
through our system, we can save up to 58% of energy APPLICATIONS. Digital Computer Applications to Process Control,
compared to traditional control mode. We have therefore not pp.457-463.
[8] Udink ten Cate, A. and van de Vooren, J. (1977). Digital adaptive
only succeeded in satisfying the DLI condition and control of a glasshouse heating system. IFAC Proceedings Volumes,
guaranteeing the same growth rate of the various plants but 10(16), pp.505-512.
also in optimising the energy consumption of the luminaires. [9] Ameur, S., Laghrouche, M. and Adane, A. (2001). Monitoring a
greenhouse using a microcontroller-based meteorological data-
III. CONCLUSION acquisition system. Renewable Energy, 24(1), pp.19-30.
[10] Wolfert, S., Ge, L., Verdouw, C. and Bogaardt, M. (2017). Big Data in
This research has shown that the supervision and control of Smart Farming – A review. Agricultural Systems, 153, pp.69-80.
lighting can improve the efficiency of our greenhouse in [11] Pham, X. and Stack, M. (2018). How data analytics is transforming
terms of consumption and quality of products. We have agriculture. Business Horizons, 61(1), pp.125-133.
[12] Kamilaris, A., Kartakoullis, A. and Prenafeta-Boldú, F. (2017). A
naturally drawn the following conclusions: review on the practice of big data analysis in agriculture. Computers
The average illumination differs from one area to and Electronics in Agriculture, 143, pp.23-37.
another and this has been shown through the [13] Karan Kansara et al “Sensor based Automated Irrigation System with
installation of nine sensors inside our greenhouse, IOT: A Technical Review”, International Journal of Computer Science
and Information Technologies, Vol. 6 Issue 6 , pp. 53315333, 2015.
this difference is essentially due to several [14] Thakur, D., Kumar, Y., Kumar, A., Kumar, P. and Singh, V. (2018).
parameters such as (greenhouse position, altitude, Real Time Monitoring of Valeriana Jatamansi Plant for Growth
latitude, geometric shape of the greenhouse, angle Analysis. Procedia Computer Science, 132, pp.507-517.
of inclination of the sun, etc.....) [16].
[15] Li, Y., Ding, Y., Li, D. and Miao, Z. (2018). Automatic carbon
dioxide enrichment strategies in the greenhouse: A review. Biosystems
Engineering, 171, pp.101-119.
[16] Belkadi, A., Mezghani, D. and Mami, A. (2019). Energy Design and
Optimization of a Greenhouse: A Heating, Cooling and Lighting
Study. Engineering, Technology & Applied Science Research, Vol. 9
issue 3, pp. 4235-4242.
[17] Ted Goldammer, 2018, “Greenhouse Management: A Guide to
Operations and Technology,” Apex Publishers, First Edition, ISBN
(13): 978-0-9675212-9-9.
[18] Damers D.A., Dorais M, C.Wien, A. Gosselin, 1998, ”Effects of
supplemental light duration on greenhouse tomato (Lycopersicon
esculentum Mill.) plants and fruits yields. Scientia Hort, 74, pp. 295-
306.
Abstract— This work presents a analysis of the high performance torque control of the doubly fed induction
performance of the proportional integral controller and machine (DFIM). DTC improves the induction machine
artificial neural network controller using for the control of controller dynamic performance and reduces the influence of
doubly fed induction machine speed. However, Direct Torque the parameter variation during the operation [4] .
Control is using to obtain high performance torque control but due to
the sensitivity of proportional integral regulators to parametric The Direct Torque Control (DTC) method was introduced,
variations because their coefficients are calculated according to these especially for asynchronous machines in 1985 by Takahashi
parameters. In this paper , we propose the replacement the speed PI and Depenbrock [5]. This control technique is based on a
regulator of double fed induction motor by a regulator based on
artificial intelligence especially neuronal controller to improve the
decoupling of the flux and the motor torque by orientation of
tuning performance. A complete simulation model for the the magnetic field of the stator and he is known to have a
proposed drive is developed using MATLAB/Simulink. The simple control structure with comparable performance
effectiveness of the proposed drive is verified at different to that of the field-oriented control (FOC) techniques
dynamic operating conditions by simulation results. Indeed , developed by [6] . The DTC have several advantages over
the simulation results showed that the proportional integral its competitor field oriented control. It is utilizes hysteresis
controller is sensitivity to motor parameter variations but the band controllers for both stator flux linkage and motor
artificial neural network controller ensures the best developed torque controls. Unlike FOC, the DTC scheme does
performances in rotor resistance and load variations and better not need any coordinate transformation, pulse width
dynamic characteristics modulation (PWM) and current regulators. Unlike FOC
methods, DTC techniques require utilization of hysteresis
band comparators instead of flux and torque controllers [7].
Keywords— Variable speed; Doubly fed induction machine; The fact that the DTC directly controls the switches without
DTC; controllers; artificial neural network; simulation. going through regulators, greatly improves its dynamic
performance compared to other controls.
Page 170
Copyright - 2019
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Copyright - 2019
ISSN 1737-9334 Page 171
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
3. p. s
Tem = (u s − s s )
1 − Lm 2.R s
(19)
L 0
L s Lr
0
Electromagnetic torque can be controlled by means of the
s
0 1 − Lm component of the stator voltage,under adequate decoupling of
0 the stator flux. DTC requires the estimation of stator flux and
L s L s Lr torque, which can be performed by means of two different
B= (12)
− Lm 1 phase currents, the state of the VSI and the voltage level in the
0 0 DC voltage bus. This work uses a DTC schemes for an
L s Lr Lr induction motor fed by two-level voltage source inverter (VSI).
Lm 1
0 0
L s Lr Lr
III. PI AND ARTIFICIAL NEURAL NETWORK CONTROLLER
The following torque expression: We chose the coefficients for =1 and response time
(Trep=20ms) at 5%.
a b
c V. CONCLUSION
REFERENCES
Abstract— Based on the density functional theory (DFT), the dimensional honeycomb structures formed by other elements,
structural, electronic and magnetic properties of bare silicene such as silicene [7-8]. Silicene is a two-dimensional
nanoribbons (SiNRs) with either zigzag edge (ZSiNRs) or hexagonal lattice of silicon (Si) atoms,The electronic
armchair edge (ASiNRs) and the effect of Cl-termination on the structure of the two-dimensional silicene is similar to that of
properties of the two types SiNRs are studied. The optimized
graphene, with zero gap and linear energy dispersion near the
structural parameters were in good agreement with the
experimental and theoretical values. The results show that both K point[9], but contrary to graphene, silicene has a buckled
bare zigzag and armchair are always metallic. The effect Cl- atomic structure(0.44 Å)[9].However, silicene also possesses
termination clearly on the electronic and magnetic properties of some unique properties which carbon has not [10-11]. For
SiNRs with either zigzag edge (ZSiNRs) or armchair edge example, the graphene is not easily compatible for the Si-
(ASiNRs).We observed a nearly degenerate flat edge-state band based electronics industry.
appears at the Fermi level Ef, which give rise to a very large In recent years, A new material, quasi one-dimensional
density of states (DOS up and DOS down) at Ef, demonstrating (1D) derivatives of silicene, called Silicene nanoribbons
magnetic properties for Cl-terminated 5-ZSiNR. For Cl- (SiNRs), (namely, graphene-like Si [12]) have been the
terminated 7-ASiNR the width between the top of the valance
subject of numerous experimental and theoretical studies[13-
band or the bottom of the conduction band representing a direct
semiconducting and nonmagnetic because of the σ-π mixing. The 14].The Silicene nanoribbons was successfully epitaxially
total density of states (DOS) and projected density of states grown on Ag surface [14-15-16-17]. The electronic and
(PDOS) analyses show that the top of the valence band or the magnetic properties of SiNRs with H-terminated and F-
bottom of the conduction band for the 7-ASiNR is mainly terminated have been synthesized and expected to possess
contributed by Cl atoms. The charge density distributions show unusual electronic and magnetic properties dependent on their
that there is covalent bonding between all sorts of the Si–Si width, edge shape and termination [18–19]. The first-
bonds, ionic bonding resulted between the edge Cl atom and the principle numerical methods showed all the H-terminated and
nearest Si atom. This because the electro-negativity of Si atoms F-terminate with the armchair edges ASiNRs are
is smaller than that of Cl atoms.
nonmagnetic semiconductors with a direct gap and the band
gaps present an oscillation behavior with a period of 3
Keywords— silicene nanoribbons, DFT calculations, Cl- dimers when ribbon width increases, and the H-terminated
termination , electronic properties and F-terminate with the zigzag edges (ZSiNRs) have
stable antiferromagnetic states.
I. INTRODUCTION In this paper, the structural, electronic and magnetic
Graphene is a form of carbon which has a two- properties of bare silicene nanoribbons (SiNRs) with either
dimensional honeycomb structure [1]. Due to their fascinating zigzag edge (ZSiNRs) or armchair edge (ASiNRs) and the
properties and potential applications in nanoelectronics such effect of Cl-termination on the properties of the two types
as zero band gap semi metal, very high carrier mobility [2], SiNRs are examined using the density functional calculations
graphene attracted much attention in recent years [3]. Quasi (DFT).This paper is organized as follows. The calculation
one-dimensional (1D) derivatives of graphene, called method and models of the two types of Cl-terminated SiNRs
graphene nanoribbons have been a subject of interest because are given in section 2, the energy band structures, the density
of thier electronic and magnetic properties dependent on their of states (DOS) and the charge density of the systems are
size [4], geometry [5], and edge-passivation pattern [6].The analyzed in section 3 and conclusions are presented in section
discovery of graphene also stimulates interest on two- 4.
II. CALCULATION METHODS AND MODELS III.1 STABILITY AND STRUCTURAL PROPERTIES
The present calculations have been performed by using the Taken 5-ZSiNR and 7-ASiNR with Cl-terminated as
Vienna ab initio simulation package (VASP) [20, 21], which example, we first make the full structure optimizations. We
is based on the spin-polarized density functional theory [22, find that the lattice constant of nanoribbons for 6-ZSiNR and
23].Ultrasoft Vanderbilt-type pseudopotentials [24] were used (b) 7-ASiNR are aZSiNR =3.88956Å and aASiNR=6.72357Å
to describe the interactions between ions and electrons. The respectively. The Si–Si lengths and Si–Cl bonds in 6-ZSiNR
exchange correlation functional is treated within the and 7-ASiNR with Cl-terminated are shown in Figure 2. (a)
generalized gradient approximation (GGA), in the form of and (b), respectively. For both SiNRs with Cl-terminated the
Perdew and Wang 91 (PW91) [25]. The cut-off energy for the length of the Si–Cl bond is always 2.06Å, but the Si–Si
plane waves is chosen to be 350 eV. The 2s2 2 p3, and lengths are varying between 2.22 Å and 2.29 Å.
electrons are taken as the valence electrons for Si and Cl
atoms, respectively. Supercells are used to simulate the
isolated nanostructures, and the distance between images is
chosen to be 15 Å in order to avoid interactions between the
ribbon and its periodic images.The sampled k points in the
Brillouin zone are generated by the Monkhorst–Pack scheme
[26] with -centred grids. The convergence of the total energy (a) 5-ZSiNR
is chosen to be 0.0001 ev and all atomic positions and unit
cells are optimized until the atomic forces are less than 0.02
eV/Å with a Gaussian smearing broadening of 0.2
eV.Pressures on the lattice unit cell are decreased to values
less than 1 kbar. We investigate for infinitely long single-
layer SiNRs and set the periodicity of the SiNR along the
ribbon axis. The width of ZSiNR is classified by the number
of the zigzag chains (Nz) across the ribbon width and the (b) 7-ASiNR
width of ASiNR is classified by the number of dimer lines
Fig. 2 Optimized lengths (A0) of Si–Si, and Si– Cl bonds in (a) 5-ZSiNR
(Na ) across the ribbon width. Thus we refer a ZSiNR with Nz and (b) 7-ASiNR terminated with Cl atoms. The mandarin and green
chains as Nz -ZSiNR and an ASiNR with Na dimer lines as Na balls represent Si and Cl atoms balls, respectively.
-ASiNR. Figure 1 presents the geometry for (a) 6-ZSiNR and
(b) 7-ASiNR with Cl atoms termination. The mandarin and The fully optimized structural parameters of 7-ASiNR and 5-
green balls represent Si and Cl atoms, respectively. ZSiNR are summarized in Table 1 and compared to
experimental values of Kara et al. [17] and theoretical
calculations of C. Lian et al. [27] and N. Gao et al [28]. The
Lattice constants aSiNR (in Å), the length of the Si–Cl bond
LSi-Cl and width of nanoribbons dwidth (in Å). We define the
width of nanoribbons dwidth = ddistance+ ddiameter , where
aZSiNR ddiameter is the diameter of the silicon atom, which is 2.2 Å and
ddistanceis the distance between the center of silicon atoms on
both edges . It shows that the 7-ASiNRs have 15% difference
and 5-ZSiNRs have17% difference with the experimental
ones. The deviations of our results with others are reasonable.
(a) 5-ZSiNR
Table 1 Geometric parameters of 5-ZSiNR and7-ASiNR,
compared to experimental data of Kara et al. [17] and calcul-
ations of Lian et al. [27] and Gao et al [28].
Parameters Present work Theorya,b Expc
aZSiN 3.88956 (3.868)a
aASiNR aASiNR 6.72357 (6.7)a
dwidth5-ZSiNR 15.72 (16.6)a 16
dwidth7-ASiNR 13.77 (14.7)a 16
LSi-Cl SiNR 2.06 (2.085)b
a
Ref. [27]
(b) 7-ASiNR b
Ref. [28]
c
Fig. 1 The geometry structures of (a) 5-ZSiNR and (b) 7-ASiNR Ref. [17]
terminated with Cl atoms. The mandarin and green balls represent Si and
Cl atoms balls, respectively. The area between two black lines represents III.2 ELECTRONIC PROPERTIES For getting insights
the lattice constant of SiNR. the electronic structure of bare 5-ZSiNR and 7-ASiNR, their
band structures are shown in Figure.3 (a) and (b), respectively.
It can be seen that the band structures the dangling bonds
III. RESULTS AND DISCUSSION introduce one or several bands crossing the Fermi level
Density(states/eV)
4
Energy(eV)
and 7-ASiNR are presented in Figure.4 (a) and (b), 2
-10
than of the total magnetic moment per supercell of bare 5- X -10 -5 0 5
1
Energy(eV)
0 0
Figure. 6. It can be seen that the bottom of the conduction
-1 band and the top of the valance band appeared at Γ point are
-1
separated. So that the Cl-terminated 7-ASiNR is nonmagnetic
semiconductors with a direct band gap approaches to 0.12 eV.
-2
-3 -2
X X
Fig. 3 The band structures of bare (a) 5-ZSiNR and (b) 7-ASiNR,
respectively. The Fermi level Ef is set to zero and is indicated by
the horizontal lines. 2
2 4
2
between non-bonding 3p electrons of Cl atoms and 3p state of
0 0
Si atoms, produce an important orbital mixing with the σ
-2 valence electrons. Such a phenomena was also observed by
-2 -4 S.Y.You et al. in the Si halides nanowires[34] and Y.-NYang
-6 et al in F-terminated ZSiNRs[19]. DFT calculations have
-4 -8 shown that the different terminations can affect the electronic
-10 properties of SiNRs .The Cl- terminated SiNRs has a
-4 -2 0 2 4
advantage because of its narrower band gap compared to the
-10 -8 -6 -4 -2 0 2 4
Energy(eV) Energy(eV)
Fig. 4 The spin polarized total density states of (a) 5-ZSiNR and (b) 7- SiNRs with H-terminated and F-terminated.
ASiNR, respectively. The Fermi level Ef is set to zero and is indicated
by the vertical lines.
20
0,4
edge dimer Si–Si bond with two adjacent Cl atoms, the
0,3
bond direction of the nanoribbon axis, the bond position from
0,2 the nanoribbon edge and the flatband that appeared in 5-
0,1 ZSiNR but does not appear in 7-ASiNR terminated with the
Cl atoms at the Fermi level Ef.
0,0
0,6
(c)
0,5
IV CONCLUSIONS The structural, electronic and
magnetic properties of the both bare SiNR with either zigzag
Density(States/eV)
0,4
0,3
NZ - ZSiNR or armchair NA - ASiNR edges with NZ=5 and
NA=7 were investigated based on the density functional
0,2
calculations under GGA. It has been demonstrated that, firstly,
0,1 the both bare 5-ZSiNR and 7-ASiNR having metallic
0,0
character with magnetic moment. It also agrees with the
-20 -15 -10
Energy(eV)
-5 0 5 experimental and theoretical results. Secondly, we examine
Fig. 7 The total density of states (DOS) (a),
the effect of Cl-termination on the electronic and magnetic
projected density of states (PDOS) on Si 3p atom properties of the two types SiNRs. The Cl-terminated 5-
(b) and Cl 3p atom (c) of Cl-terminated 7- ZSiNR having zero gap, also a nearly degenerate flat edge-
ASiNR.The Fermi level Ef is set as energy zero. state band appears at the Fermi level Ef , which give rise to
very large spin-polarized total density of states density of
The charge-densities contours of 5-ZSiNR and 7-ASiNR states ( up and down) at Ef.. This DOS peak demonstrating
terminated with the Cl atoms in the (0 0 1) crystallographic magnetic properties of Cl-5-ZSiNR. The value of the total
plane respectively, are presented in Figure 8. magnetic moment has declined. However, the Cl-terminated
7-ASiNR is nonmagnetic semiconductors with small direct
band gap. Because of the σ-π mixing effect, the interaction
between non-bonding 3p electrons of Cl atoms and 3p state of
Si atoms, produce an important orbital mixing with the σ
valence electrons. The DOS and PDOS analyses show that
the top of the valence band or the bottom of the conduction
band for the 7-ASiNR is mainly contributed by Cl atoms. The
bonding features are analyzed by using the charge-densities
contours in the (0 0 1) crystallographic plane. We found that
covalent bonding between all sorts of the Si–Si bonds, ionic
bonding resulted between the edge Cl atom and the nearest Si
atom.
(a) 5-ZSiNR
REFERENCES
[1] K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S.
V. Dubonos, I. V. Grigorieva, and A. A. Firsov, Science 306, 666
(2004).
[2] J.H. Chen, C. Jang, S. Xiao, M. Ishigami and M.S. Furher,
Nanotechnology 3, 206 (2008).
[3] A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov, and A.
K. Geim, Rev. Mod. Phys. 81, 109 (2009).
(b) 7-ASiNR [4] Y.W. Son, M.L. Cohen and S.G. Louie, Phys. Rev. Lett. 97, 216803
Fig. 8 The charge density contours of (a) 5-ZSiNR and (b) (2006).
7-ASiNR terminated with Cl atoms. [5] V. Barone, O. Hod and G.E. Scuseria, Nano Lett. 6, 2748 (2006).
[6] M. Wang and C.M. Li, Nanoscale 3, 2324 (2011). [20] G. Kresse and J. Furthmüller, Physical Review B 54, 11169 (1996).
[7] B. Aufray, A. Kara, S. Vizzini, H. Oughaddou, C. Landri, B. Ealet and [21] G. Kresse and D. Joubert, Physical Review B 59, 1758 (1999).
G. Le Lay, Appl. Phys. Lett. 96, 183102 (2010). [22] P. Hohenberg, and W. Kohn, Physical Review B 136, 864 (1964).
[8] A. Kara, H. Enriquez, A. P. Seitsonen, L. C. Lew Yan Voon, S. [23] W. Kohn and L.J. Sham, Physical Review A140, 1133, (1965).
Vizzini, B. Aufray, and H. Oughaddou, Surf. Sci. Rep. 67, 1 (2012). [24] D. Vanderbilt, Physical Review B 41, 7892, (1990).
[9] S. Cahangirov, M .Topsakal, E. AktÜrk, H. Sahin and S. Ciraci , Phys. [25] J.P. Perdew, J.A. Chevary, S.H. Vosko, K.A. Jackson, M.R. Pederson,
Rev. Lett.102, 236804 (2009). D.J. Singh and C.Fiolhais, Phys. Rev. B 46 , 6671(1992).
[10] M. Zhang, Y.H. Kan, Q.J. zang, Z.M. Su and R.S. Wang, Chem, Phys. [26] H J Monkhorst and J D Pack, Phys. Rev. B13, 5390 (1976).
Lett. 379 (2003) 81. [27] C. Lian and J.Ni, Physica B407, 4695–4699, (2012).
[11] H. Lee, Y. W. Son, N. Park, S. Han, and J. Yu, Phys. Rev. B 72,174431 [28] N. Gao, W. T. Zheng and Q. Jiang, Phys. Chem. Chem. Phys.14, 257–
(2005). 261. (2012).
[12] G.G. Guzman- Verri, and L.C. Lew Yan Voon, Phys. Rev. B 76, [29] Y. Miyamoto, K. Nakada and M. Fujita, Phys. Rev. B 59, 159858
075131(2007). (1999).
[13] Y. Ding and J. Ni, Appl. Phys. Lett. 95, 083115 (2009).
[14] A. Kara, S. Vizzini, C. leandri, B. Ealet, H. Oughaddou, B. Aufray, and [30] K. Nakada, M. Fujita, G. Dresselhaus, M.S. Dresselhaus, Phys. Rev. B
G. Le Lay, J. Phys. Condens. Matter 22, 045004 (2010). 54 ,17954,(1996).
[15] Y. Ding and J. Ni, Appl. Phys. Lett. 95, 083115, (2009). [31] P. De Padova, C.Quaresima, P.Perfetti, B.Olivieri, C.Leandri, B.Aufray,
[16] P. De Padova, C. Léandri, S. Vizzini, C. Quaresima, P. Perfetti, B. S.Vizzini and G.Le Lay, Nano Lett.8, 271 (2008).
Olivieri, H. Oughaddou, B. Aufray, and G. Le Lay, ibid. 8, 2299 [32] H. Motoyama, K Takeda and K Shiraishi,Pro. Mat. Res. Soc.symp,
(2008). 486 , 385 (1997).
[17] A. Kara, C. Lèandri, M. E. Dàvila, P. De Padova, B. Ealet, H. [33] D. B. Lu, Y L Song, Z X Yang and G Q Li, Appl. Surf. Sci.256, 6313
Oughaddou, B. Aufray, and G. Le Lay, J. Supercond. Novel Magn. 22, (2010).
259 (2009). [34] S. Y. You and Y Wang,Chin. J. Semiconductors 27, 1927. Chinese
[18] Y.L. Song, Y. Zhang, J.M. Zhang and D.B. Lu, Appl. Surf. Sci. 256, (2006).
6313-6317 (2010).
[19] Yan-Ni Yang, Jian-Min Zhang and Ke-Wei Xu, Physica E 57 , 21–27
(2014).
Abstract— H3PW12O40 and H4SiW12O40 were used as catalysts for the advantage of activating both esterification and
the sunflower oil ethanolysis as bulk and supported on silica, SiO2 transesterification of fatty acids and, in addition, the water
with (30wt%). The fatty acid ethyl esters, reaction products, were presence in the reaction mixture is not a barrier.
analyzed and quantified by gas chromatography. The influences of
Among the acid solid catalysts examined in the vegetal
ethanol/oil molar ratio (14,5/1 -58/1), temperature (35 or 70°C) and
oil transesterification as SO4–ZrO2 [7] and WO3–ZrO2 [8,9],
reaction time (30-180min) on biodiesel yield and moles number of
fatty acid ethyl esters formed were examined. emerge the tungsten based Keggin-type polyoxometalates
Both acids in bulk form are more effective for sunflower (POMs) in their acid or salt form. Their effectiveness is due
ethanolysis than those in supported form and the SiW based catalyst to their Bronsted acidity stronger than conventional acids as
is more effective than those based PW. H4SiW12O40 is the most H2SO4 and HNO3. Thus, the insoluble salts as (NH4)xH3-
performant catalyst with 55% of biodiesel yield at 70°C with a xPW12O40 were used for palmitic acid esterification [10],
reaction time of 3 hours and an alcohol/oil ratio of 29/1 and also, the (Cs)xH3-xPW12O40 for used vegetable oil transesterification
most active for the ethanolysis of palmitic acid. [11] and for Ag xH3-xPW12O40 castor oil transesterification
[12]. In its acid form, H3PW12O40 was supported on Nb2O5
Index Terms— Biodiesel; Ethanolysis; Sunflower oil; Ethyl esters;
[13], SiO2 [14] and MCM-41 [15] and tested for palmitic
Heteropolyacids.
acid esterification and transesterification of sunflower oil
and waste cooking, respectively [16].
I. I NTRODUCTION In the above-mentioned works, the most widely used
alcohol is methanol because of its low price; in recent years,
The greenhouse gas emission increase into the environment
however, the production increase of ethanol from
caused by the fossil fuel combustion has encouraged
agricultural biomass (beet, corn, sugar cane and starch) has
researchers to find an alternative. Biodiesel, composed of alkyl
led to lower prices. On the other hand, it can be noted that
esters, could be a promising fuel because it comes from a
ethanol is a renewable alcohol when it is derived from
renewable source, vegetable oils as sunflower, castor, soybean
biomass compared to methanol obtained from hydrocarbons.
….., or animal fats. In addition, the biodiesel has a very low
Ethanol in concentrated and hydrated form is used as a
toxicity, a biodegradability and a flash point lower than diesel
biofuel and also as an additive to fossil fuels. Few studies
from fossil fuel. On the other hand, contrarily to homogeneous
have focused on the ethanol use in the transesterification of
catalysis, the heterogeneous catalysis presents some
vegetable oils, although the ethyl esters are more
advantages as easy separation catalyst/reaction mixture,
biodegradable in water than methyl esters [17,18]
catalyst reuse and reducing wastes.
(Boehman, 2005; Makareviciene and Janulis, 2003). They
Several studies have reported biodiesel production from
present also a higher cetane index and heating power (Clark
vegetable oil methanolysis in heterogeneous systems using
et al., 1984) [19] and lower cloud, cold filter plugging, and
bases as calcium oxide [1], dolomite [2,3], Alumina/silica
pour points (Encinar et al., 2007). [20].
supported K2CO3 [4], sodium silicate [5] and hydrotalcite [6].
Production of fatty acid ethyl esters from rubber seed oil
Although, basic catalysis leads to high biodiesel yields, it
in hydrodynamic cavitation reactor in the presence of
causes a saponification reaction. To overcome this
sulfuric acid (4,0 wt.%) led to 92,5% of biodiesel yield,
disadvantage, authors have proposed an acid catalysis that has
Copyright - 2019
ISSN 1737-9334 Page 181
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
SHELL: HOW TO USE THE IEEETRAN LATEX CLASS 27
obtained with a 6:1 ethanol:oil molar ratio [21]. The soybean used as the carrier gas with a flow rate of 0,5 ml min-1. The
oil transesterification with ethanol using sulfonated activated samples were diluted with hexane. Initial oven temperature
carbon from corn cobs as a catalyst and a microwave-assisted was set at 90 °C for 5 min, then ramped to 280 °C at 4
process has proven to be a promising, less energy-intensive °C.min-1 and kept at the highest temperature for 5 min.
solution [22]. Injector temperature and injector volume were 250 °C and
In this context, we were interested in the sunflower oil 0,2 µl, respectively. The split ratio was 20:1. The ionization
ethanolysis in the presence of H3PW12O40 and H4SiW12O40 in source (electronic impact) temperature was kept at 230 °C
homogeneous and heterogeneous phases. The used support is and that of the interface at 280 °C.
silica, SiO2, with (30wt%) H3PW12O40 (noted S30) and The biodiesel yield was calculated from the content of
(30wt%) H4SiW12O40 (noted S'30). The preparation of the analyzed ethyl esters using the following equation [14]:
supported materials was previously reported [14]. The
influences of ethanol /oil molar ratio (14,5/1-58/1), Biodiesel Yield (%) = mol FAEEs number × 100/ 3 × mol
temperature (35 or 70°C) and reaction time (30-180min) on oil number.
biodiesel yield and moles number of fatty acid ethyl esters
were examined. III. RESULTS AND DISCUSSION
in heterogeneous system, S30 reacted and led to 2,39×10-4 against 4,93×10-4 – 10,25×10-4 for other ethanol/oil molar
moles of ethyl oleate and 11,78 × 10-4 moles of methyl oleate. ratios. C18:2 ethanolysis of toward ethyl linoleate was
Methyl linoleate (C18:2) was obtained with 35,37×10-4 and favored with a ratio of 29/1 and that of C16:0 to ethyl
34,01×10-4 moles and ethyl linoleate with 24,04×10-4 and palmitate, with a ratio of 43,5/1. Ethanol is very little active
13,45×10-4 moles in homogeneous and heterogeneous phases, with respect to the transesterification of C18:0 and C18:1
respectively. 5,63×10-4 mole of methyl palmitate and 4,02×10-4 fatty acids.
moles of methyl stearate and 5,51×10-4 moles of ethyl From these results, it can be observed that the highest
palmitate and 3,20×10 -4 moles for ethyl stearate in biodiesel yield (29%) was obtained with a molar ratio of
homogeneous phase were obtained. In heterogeneous phase, 43,5/1 and a reaction time of 180 min from sunflower oil
the formation of these products is less favoured with 3,80×10-4 ethanolysis carried out at 70°C. However, there is not a
mole of methyl palmitate and 4,02×10-4 moles of methyl significant difference between the obtained biodiesel yields
stearate and 2,47×10 -4 moles of ethyl palmitate and 2,18×10-4 (21-29%) for molar ratios of 29/1, 43,5/1 and 58/1. for this
moles for ethyl stearate. From these results, it can be observed reason, the ethanol/oil molar ratio will be fixed at 29/1 for
that whatever the fatty acid nature, methanolysis is more the followed study.
favoured than ethanolysis both in homogeneous and
heterogeneous systems. TABLE III
EEAGS MOLES NUMBER AND BIODIESEL YIELD (%) AS A FUNCTION OF
THE ALCOHOL/OIL MOLAR RATIO
TABLE II (CATALYST: S30, REACTION TEMPERATURE: 70°C, REACTION TIME: 3H)
EEAGS MOLES NUMBER AND BIODIESEL YIELD (%) OBTAINED
WITH METHANOLYSIS AND ETHANOLYSIS REACTION IN THE Molar ratio EEAGs moles number *104 Biodiesel
HOMOGENEOUS AND HETEROGENEOUS PHASE
C alcohol / oil yields (%)
C16:0 C18:0 C18:1 C18:2
. 14,5/1 4,08 0,67 1,39 4,93 11
4 Eff
EEAGs moles number *10
ect 29/1 2,47 2,18 2,39 13,45 21
Type of Type of C16:0 C18:0 C18 :1 C18 :2 Biodiesel
Catalysts Alcohol yield (%) of 43,5/1 13,33 1,61 2,06 10,25 29
te
HPW méthanol 5,63 4,02 11,41 35,37 63 mp 58/1 7,57 1,94 2,80 11,98 24
HPW éthanol 5,51 3,20 0,00 24,04 37 era
S30 méthanol 3,80 4,02 11,78 34,01 60 ture and time reaction
S30 éthanol 2,47 2,18 2,39 13,45 21
The figures 1 and 2 show the biodiesel yields and the
B. Effect of ethanol/oil molar ratio number of ethyl ester moles formed, as time reaction
function respectively, at 35 and 70°C. The sunflower oil
Table III shows the moles number of formed EEAGs and ethanolysis carried out at in the presence of S30 with a
biodiesel yield as function of ethanol/oil molar ratio. The molar ratio of 29/1. The variation in the reaction
sunflower oil ethanolysis was carried out at 70°C in the S30 temperature has a strong influence on both the biodiesel
presence with a reaction time of 3h. The results showed that yield and the number of ethyl ester moles formed. Thus,
biodiesel yield and ester ethyl formation are sensitive to whatever the reaction time, biodiesel yields increase from
ethanol/oil molar ratio variation. Thus, the biodiesel yield 12-27% to 11-46% with the reaction temperature from 35 to
increases from 11 to 29% with the increase of the ethanol/oil 70°C. The highest biodiesel yield (46%) was obtained after
molar ratio from 14,5/1 to 43,5/1 and then decreases from 29 150min of ethanolysis. It is noted that with a reaction time
to 24% for a ratio of 58/1. The decrease in biodiesel yield of 90min, the highest biodiesel yield (27%) was obtained at
observed with an excess alcohol could be explained, as in the 35°C and the lowest (11%) at 70°C. According with the
case of methanolysis, either by poor glycerol settling and observed results at 180min and 35°C, where a strong
consequently poor ester-glycerol separation and/or a shift in decrease of the biodiesel yield was obtained (from 24 to
the equilibrium towards the direction of triglyceride formation 12%), it is the same with 180min and 70°C (from 46 to
or to an emulsification effect during biodiesel washing that 21%).
would be favoured with an amount alcohol increase in the The results obtained highlight the importance of
reaction medium [24, 25]. temperature and reaction time on the displacement of the
The number of moles of C18:0 and C18:1 ester ethyl ethanolysis equilibrium of sunflower oil. Thus, a
(0,67×10-4 – 2,18×10-4 and 1,39×10-4 – 2,80×10-4, temperature increase of 35 to 70°C also makes it possible to
respectively) appears to be little affected by the rise ethanol/oil reach the activation energy necessary to initiate the reaction,
molar ratio from 14,5/1 to 58/1 unlike to C16:0 ester ethyl that to reduce the viscosity of the oil (immiscible reagent) and
display 13,33×10-4 moles for a ethanol/oil molar ratio of thus to make the mixing between the reagents more efficient
43,5/1 against 2,47×10-4 – 7,57×10-4 moles for other and consequently to displace the equilibrium towards the
ethanol/oil molar ratios. C18:2 ester ethyl that presents a mole formation of ethyl esters of fatty acids [26]. These results
number of 13,45×10-4 for a ethanol/oil molar ratio of 29/1 also show that 90 min are not enough to displace the
Copyright - 2019 Page 183
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
SHELL: HOW TO USE THE IEEETRAN LATEX CLASS 27
35
30
obtained in the presence of H3PW12O40, with 55 versus 37%,
25
suggesting that transesterification depends much more on
20 the proton number than on the acidity strength. It is known
15 that H3PW12O40 is a stronger acid than H4SiW12O40 [30-32].
10 This latter is also very active towards palmitic acid
5 ethanolysis compared to that of H3PW12O40 , with 27,78×10-
4
0 moles against 5,51×10-4 moles of ethyl palmitate, while
0 50 100 150 200 they have the same catalytic behavior towards linoleic acid
Time (min)
ethanolysis, with an ethyl linoleate moles number of
FIG. 1. BIODIESEL YIELD AS FUNCTION OF TEMPERATURE AT 35°C
AND 70°C (CATALYST : S30, ETHANOL/OIL : 29/1)
18,52×10-4 and a similar catalytic behavior towards stearic
acid ethanolysis, with an ethyl stearate moles number of
2,70×10-4 and 3,19×10-4.
In heterogeneous systems, S30 and S’30 led to biodiesel
yields inferior to those of H3PW12O40 and H4SiW12O40, bulk,
respectively, with 21 and 19% against 37 and 55%.
Expected result, given that acid sites active in homogeneous
phase are more accessible to reagents than those in
heterogeneous phase.
Supported on silica, both heteroployacids, were also
found to be less active towards the ethanolysis of C16:0,
C18:0, C18:1 and C18:2 fatty acids. The difference in the
number of moles of ethyl esters corresponding to C16:0 and
C18:2 acids is more pronounced in the case of the
H4SiW12O40 bulk use with 27,78×10-4 moles and 18,52×10-4
Copyright - 2019 moles against 8,16×10-4 moles and 7,86×10-4 moles,
ISSN 1737-9334 Page 184
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
SHELL: HOW TO USE THE IEEETRAN LATEX CLASS 27
respectively, for S’30. This observation seems to suggest that The bulk and supported heteropolyacids proved to be
the ethanolysis of both acids (C16:0 and C18:2) is more active catalysts in the transesterification of sunflower oil.
favored when the protons are more available. Regardless homogeneous (H3PW12O40) or heterogeneous
While, in the case of the H3PW12O40, similar behavior (H3PW12O40/SiO2) phase, the formation of methyl esters is
towards of the formation of ethyl esters regardless of fatty acid favored to that of ethyl esters, whatever the fatty acid
nature was observed for bulk and supported (S30). nature.
Supported heteropolyacids were tested at different Both acids in bulk form are more effective for sunflower
alcohol/oil molar ratios. Excepted for the ratio of 58/1, where ethanolysis than those in supported form and the SiW based
the biodiesel yield obtained with S30 is superior to that catalyst is more effective than those based PW.
obtained with S’30 (24 against16%), similar biodiesel yields The ethanolysis of palmitic acid (C16:0) is more favored
were observed for both supported catalysts, for each ratio ( 11- than those of C18:0 acids. Therefore, ethyl palmitate is the
12% for 14,5/1, 21-19 for 29/1, and 29 for 43,5/1). The main component in the biodiesel.
palmitic acid ethanolysis is more favored in the presence of
S’30 compared to S30, with 5,99×10-4 against 4,08×10-4 ,
8,16×10-4 against 5,51×10-4 and 16,31×10-4 against 13,33×10-4 ACKNOWLEDGMENT
of ethyl palmitate moles for alcohol/oil ratios of 14,5/1, 29/1 The author would like to thank Pr. Chérifa Rabia, for its
and 43,50/1, respectively. However, the opposite was observed help and support in making this work possible. This work is
for the alcohol/oil ratio of 58/1 (7,57×10-4 against 1,01×10-4 of carried out in the Natural Gas Laboratory, Faculty of
ethyl palmitate moles). Chemistry, University of Science and Technology Houari
Contrarily, to the ethyl palmitate formation, that of ethyl Boumediéne (USTHB), Algiers. Funded by the Minister of
linoleate is promoted by S30 catalyst, whatever the alcohol/oil Research.
ratio (4,9×10-4 -13,45×10 -4 against 3,75×10-4 -11,92×10-4 ethyl
linoleate moles). S’30 is inactive and S30 is few active V. REFERENCES
towards the oleic acid ethanolysis. S30 and S’30 have a similar
[1] Ming-Chien Hsiao b, Chin-Chiuan Lin a, Yung-Hung
catalytic behavior for stearic acid ethanolysis with ethyl
Chang. Microwave irradiation-assisted transesterification
stearate moles number inferior at 3,2×10-4.
of soybean oil to biodiesel catalyzed by nanopowder
TABLE IV calcium oxide, Fuel 90 (2011) 1963–1967.
EEAGS MOLES NUMBER AND BIODIESEL YIELD (%) AS A FUNCTION OF THE [2] Oguzhan Ilgen. Dolomite as a heterogeneous catalyst for
CATALYST COMPOSITION NATURE AND ALCOHOL/OIL MOLAR RATIO (REACTION
TEMPERATURE: 70°C, REACTION TIME: 3H)
transesterification of canola oil. Fuel Processing
Technology 92 (2011) 452–455.
EEAGs moles number ∗ 104
[3] Shuang Zhao, Shengli Niu, Hewei Yu, Yilin Ning,
Molar Biodiesel
ratio Catalyst C16:0 C18:0 C18:1 C18:2 yield (%) Xiangyu Zhang, Ximing Li, Yujiao Zhang, Chunmei Lu,
ethanol Kuihua Han. Experimental investigation on biodiesel
/oil
production through transesterification promoted by the
S30 4,08 0,67 1,39 4,93 11
14,5/1 S’30 5,99 0,56 0 3,75 12 La-dolomite catalyst. Fuel 257 (2019) 116092.
H3PW12O40 5,51 3,19 0 18,52 37 [4] I. Lukic´, J. Krstic´, D. Jovanovic´, D. Skala.
29/1 S30 2,47 2,18 2,39 13,45 21
H4SiW12O40 27,78 2,70 0,45 18,52 55 Alumina/silica supported K2CO3 as a catalyst for
S’30 8,16 1,16 0 7,86 19 biodiesel synthesis from sunflower oil. Bioresource
S30 13,33 1,61 2,06 10,25 29 Technology 100 (2009) 4690–4696.
43,5/1 S’30 16,31 2,12 0 7,37 29
S30 7,57 1,94 2,80 11,98 24 [5] Feng Guo a,b, Ning-Ning Wei a, Zhi-Long Xiu a , Zhen
58/1 S’30 1,01 1,72 0 11,92 16
Fang b is within the limits prescribed by the standard.
Transesterification mechanism of soybean oil to biodiesel
IV. CONCLUSION
catalyzed by calcinedsodium silicate. Fuel 93 (2012) 468–
472.
The biodiesel production, composed of alkyl esters, from [6] Amornmart Chantrasaa, Nisarat Phlernjaia, James G.
sunflower oil transesterification was carried out using Goodwin Kinetics of hydrotalcite catalyzed
H3PW12O40 and H4SiW12O40 as bulk and supported on silica transesterification of tricaprylin and methanol for
with 30wt% (noted S30 and S’30 respectively), as catalyst and biodiesel synthesis. Jr. Chemical Engineering Journal 168
ethanol as alcohol. The biodiesel yield and the ethanolysis (2011) 333–340.
activity of fatty acids that constitute the oil are very sensitive [7] Petchmala A, Laosiripojana N, Jongsomjit B, Goto M,
to the to temperature and time reaction an substrat/ethanol Panpranot J, Mekasuwandumrong O, et al.
ratio. For comparison, the methanolysis was studied bulk and Transesterification of palm oil and esterification of palm
H3PW12O40. fatty acid in near and super-critical methanol with SO4–
ZrO2 catalysts. Fuel 89 (2010) 2387–3239.
Copyright - 2019 Page 185
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
SHELL: HOW TO USE THE IEEETRAN LATEX CLASS 27
Copyright - 2019
ISSN 1737-9334 Page 187
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Abstract. The laser welding appears a promising process to the difference in thermal and physical properties, the formation
several industrial fields due to the possibility of assembly of of brittle intermetallic compounds extremely fragile in the
dissimilar materials while preserving the characteristics of this transition zone between the two materials and the presence of
technology: low heat input, narrow and deep weld bead, high cracks in the joint [7,8].
welding speed, etc. In this study, the transport phenomena (heat
transfer, fluid flow and species distributions) are numerically
Several dissimilar material welding techniques such as
modeling for the case of laser welding of dissimilar metals Al/Mg electron beam welding [9], MIG welding [10, 11], resistance
with the finite volume software, Ansys-Fluent. In order to spot welding [12], bonding diffusion [13, 14] have been
understand the complex physical phenomena involved, a three- studied. Other studies have focused on laser welding. For
dimensional transient model has been developed based on the example, Chen et al. [15] studied experimentally and
solution of the equations of mass, momentum, energy numerically the laser welding of Titanium / Aluminum joints.
conservation and solute transport in weld pool. Transient They have highlighted more particularly the inhomogeneous
temperature distribution and thermal cycles at different locations mixture which can be formed during the welding process. G.
are obtained. The distribution of alloy elements and composition Casalino et al. [16] studied numerically fiber laser welding of
profile (Al and Mg) are also present. It is found that the heat
transfer and the mass transport are two main phenomena in the
Ti6Al4V and AA5754-T40 materials using the Ansys
formation of a heterogeneous mixture in the fusion zone. Parametric Design Language (APDL). They studied the effect
Corresponding experiments are performed using the same of some welding conditions on thermal behavior.
parameters as in the computations, showing a good qualitative In this paper, numerical and experimental studies of laser
agreement between the two results. welding of dissimilar Aluminum / Magnesium alloys have
been carried out. For the simulation, a three-dimensional
Keywords: dissimilar metals, numerical modeling, heat transfer, model was developed using Ansys, Fluent software to predict
fluid flow, mass transport. weld bead dimensions and to simulate heat transfer and mass
transport during laser welding and then results were discussed.
It also gives an idea of the mass contents of Aluminum and
I. INTRODUCTION Magnesium in different areas of the weld.
Nowadays, several manufacturers in the automotive and
aeronautics fields have studied the use of light metal alloys in II. EXPERIMENTAL PROCEDURE
different welding processes. This would allow them to reduce
In this modeling work, the Aluminum alloy (A5754) and the
the weight of their mobile gear, or the reduction of energy
Magnesium alloy (AZ31) were used as base materials and
consumed, and therefore the emission of gaseous pollutants.
were arranged one on the other, Figure 1. The chemical
Recently, the use of magnesium and aluminum alloys has
composition and the thermo-Physical properties of the base
invaded the transportation industries thanks to their interesting
materials are shown in table I and table II, respectively. The
characteristics. Aluminum alloys are characterized by low
experiments were carried out with a continuous Yb: YAG laser
density, good corrosion resistance, high thermal and electrical
with a wavelength of 1032 nm. The radius of the laser beam is
conductivity [1, 2], potential recovery, durability, inexpensive
0.45 mm [17]. An Argon shielding gas with a flow rate of 20 l
maintenance [3] and recyclability [4]. For Magnesium alloys,
/ min was used to protect the weld bead from oxidation. The
they have high strength [5], good moldability, hot forming and
operating parameters adopted in this study are shown in Table
recycling [6]. However, it has been found that the assembly of
III.
dissimilar metals Al and Mg presents difficulties because of
A(1−β) 2
S⃗⃗ = − ( 3 ) v ⃗ (5)
β +b
V. BOUNDARY AND INITIAL CONDITIONS Figure 2 shows the distribution of the isotherms of the
surfaces of a dissimilar weld joint. These isotherms are
At time t = 0, the entire domain was considered to be in the connected to thermal fusion-solidification cycles characteristic
solid state with a temperature of 300 K and the gravitational of laser welding. From this figure, we distinguish the melt and
acceleration was assumed to be 9.81 ms−2, basic metal zones. The discontinuity of the physical properties
of the base materials such as the melting temperature, the
At the flat free surface of the liquid, shear force due to specific heat capacity and the thermal diffusivity affect the
surface tension (Marangoni force) is expressed as: dimensions of the weld pool, especially the depth of
penetration.
du ∂σ ∂T ∂σ ∂C
μ =β +β (6)
dz ∂T ∂x ∂𝐶 ∂x
dv ∂σ ∂T ∂σ ∂C
μ =β +β (7)
dz ∂T ∂y ∂𝐶 ∂y
Copyright - 2019
Page 190
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
VII. CONCLUSION
[5] Liu C C, Chen D L, Bhole S, Cao X and Jahazi M 2009 Mater. Charact.
60(5) 370-376.
[6] Carrino L., Squillace A., Paradiso V., Ciliberto S., Montuori M.,
Superplastic forming of friction stir processed magnesium alloys for
aeronautical applications: A modeling approach, Materials Science
Forum 735 (2013) 180-191.
[7] Liu C C, Chen D L, Bhole S, Cao X and Jahazi M 2009 Mater. Charact.
60 (5) 370-376
[8] Liu L M, Tan J H, Zhao L M and Liu X J 2008 Mater. Charact. 59(4)
479-483.
[9] C.-T. Chi, C.-G. Chao, T.-F. Liu, C.-H. Lee, Aluminium element effect
for electron beam welding of similar and dissimilar Magnesium–
Aluminium–Zinc alloys, Scr. Mater., 56 (2007) 733–736.
[10] M Ishak and M R Islam. Weldability of A7075-T651 and AZ31B
dissimilar alloys by MIG welding method based on welding
appearances, Journal of Physics, 2014.
[11] Zhang H T and Song J Q 2011 Mater. Lett. 65(21-22) 3292-3294.
[12] F. Hayat, The effects of the welding current on heat input, nugget
geometry, and the mechanical and fractural properties of resi stance
spot welding on Mg/Al dissimilar materials, Mater. Des., 32 (2011)
2476–2484.
[13] Jian Zhang, GuoqiangLuo, Yiyu Wang, QiangShen, Lianmeng Zhang,
An investigation on diffusion bonding of aluminum and magnesium
using a Ni interlayer, Materials Letters 83 (2012) 189–191.
[14] Carrino L., Paradiso V., Franchitti S., Squillace A., Russo S.,
Superplastic forming/diffusion bonding of a titanium alloy for the
realization of an aircraft structural component in multi-sheets
configuration, Key Engineering Materials 504-506 (2012) 717-722.
[15] Chen S, Li L, Chen Y, Huanga J. Joining mechanism of Ti/Al
dissimilar alloys during laser welding-brazing process, Journal of
Alloys and Compounds, 891–898, 2011.
[16] G. Casalino, M. Mortello, P. Peyre. FEM analysis of fiber laser welding
of Titanium and Aluminum, Procedia CIRP, 992 – 997, 2016.
[17] S. Bannour, K. Abderrazak, S. Mattei, J. E. Masse, M. Autric, H. Mhiri.
The influence of position in overlap joints of Mg and Al alloys on
microstructure and hardness of laser welds, journal of laser
applications, 2013.
Abstract—The southern regions in Algeria are suffering from technology dominates the industry of desalination and
fresh water shortage due to their climate severity and the brackish water treatment in Algeria and in all world
geographic situation. These areas have access to brackish countries. The Algerian government has launched a
groundwater and significant potential of solar energy. The huge desalination program to build 13 big desalination
combination of solar energy with desalination processes, such plants in several regions of northern Algeria and even
as reverse osmosis systems, presents possible solution to in the south for the brackish water treatment. For the
satisfy the fresh water needs for the population living in these moment, the desalination of seawater contributes, in
remote areas. This paper presents a technical and economic
Algeria, to 17% in the national supply of drinking
water and we will exceed 25% once the units in
evaluation of a small-scale brackish water reverse osmosis
progress put into service [1,2]. However, while
desalination unit powered by solar photovoltaic systems for
desalination techniques are now well controlled, their
fresh water production under Algerian arid climate
implementation requires large amounts of energy, in
conditions. In the first part of this paper, a mathematic model
the form of heat or electricity, costly and potential
has been developed for design and sizing the photovoltaic
sources of greenhouse gas emissions. In addition to the
field, taking into account two configuration; power system coastal areas, these features allow a very fast
with batteries and without batteries. The photovoltaic system installation and a perfect adaptation to the specific
model is coupled to reverse osmosis model to determine the needs of arid and semi-arid regions suffering from a
overall system performance (water production, Specific deficiency of drinking water, emphasizes. In the last
Energy Consumption…) and an optimization is performed in few years, the solar power desalination technology of
order to determine the suitable configuration. In the second brackish water being considered as one of the
part, the economic feasibility study of use solar energy has promising way for fresh water production in order to
been done using some economic parameters such as, satisfy the water demand growing. In this work, we
Levelized Water Costs (LCOW) Net Present Value (NPV) present a technical and economic evaluation of a small-
and Internal Rate of Return ( IRR). The investigation also scale brackish water reverse osmosis desalination unit
includes a discussion on the sensitivity of cost of unit of powered by solar photovoltaic systems RO/PV for
produced fresh water to some parameters such as the useful fresh water production under Algerian arid climate
life of reverse osmosis system, cost of solar photovoltaic conditions. The coupling of RO/PV using the
pannels and other associated parameters. The results of this MATLAB software has been modelled. The modelling
investigation can be used also as indicators to identify the has to be done, taking into account the types of systems
best locations for solar powered desalination plant used, as well as the territorial, environmental and
implementation for fresh water production in Algeria. economic stakes involved in their coupling.
Copyright - 2019
Page 193
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
* 10 BCM annually in the north, of which 7.4 BCM of The daily capacity of these plants ranges between
surface water and 2.6 BCM of groundwater (80% are 100,000 and 500,000 m3 each, in addition to the 21
mobilized by wells and boreholes) monobloc desalination plants with an daily average
* 5.37 BCM annually in the south, of which 0.37 BCM capacity of 2500 to 5000 m3 each and a total
of surface water and 5 BCM of groundwater. daily capacity of 57,500 m3.
Hereinafter, a brief overview of the available water It’s worth mentioning that the desalination plants in
resources is provided ( only desalination status): Algeria supply nearly 15% of total water consumption
per year, as indicated on figure 6. The total water
3. DESALINATION IN ALGERIA production is estimated at 8.24 Million per day or 3.3
In order to take advantages of its coastline Billion M3 per year.
stretching over 1200 km, Algeria has moved towards
mobilization of non-conventional water resources using
seawater desalination and brackish water
demineralization process.
Figure 1. Desalination plants Map in Algeria To determine the performance of the proposed
system (figure 3), a global model of the water
Since 2000, Desalination has become an imperative transformation (brackiqh water-drink water) and the
and inevitable solution for Algeria to overcome its solar field is performed in MATLAB in which the
current shortage of potable water. For this, more than system is divided into two components. The first is
11 seawater desalination Mega plants have been for reverse osmosis unit while the second one is
installed in the coaster region of the country with a dedicated to solar photovoltaic pannels and
total daily capacity of 21.1 million m3, see figure1. batteries.
PV pannels
RO unit
Solar batteries
7. REFERENCES
1. Algerian Ministry of Water Resources .
http://www.mree.gov.dz.
2. Z. Tigrine, H. Aburideh, M. Abbas, S. Hout,
N. Kasbadji Merzouk,
D. Zioui, M. Khateb: Membrane Desalination
Technology in Algeria: Reverse Osmosis for
Coastal Areas. Chapter In book: Exergy for A
Better Environment and Improved
Sustainability. Green Energy and Technology book
series (GREEN), 2. 197-218, 2018.
3. Ali Al-Karaghouli, Lawrence L. Kazmerski,
Energy consumption and water production cost of
conventional and renewable-energy-powered
desalination processes. Renewable and Sustainable
Energy Reviews, 24, 343-356, 2013.
4. Bilton, A.M. Wiesman, R. Arif, A. F. M.
Zubair, S. M. &Dubowsky, S. On the feasibility of
community-scale photovoltaic-powered reverse
osmosis desalination systems for remote locations.
Renewable Energy, 2011, 36, pp.3246–3256.
Abstract— In this paper, a new behavioral electro-thermal power devices. This is particularly significant at a high power
(ET) model for low voltage power MOSFET is improved and level when operating at low fundamental frequencies or with
implemented in the PSpice software environment. The (ET) large load variations when the junction temperature is at a level
model couples a thermal (RC) network model with the electrical close to the operating bounds. So, it is substantial to carry out a
model by the interaction between the junction temperature (Tj) thermal analysis of the devices, in order to determine if they
and the total power loss (Ptotal). The parameters of the (RC) are performing within the maximum allowed physical limits.
thermal network model are extracted from the data provided by Generally, three methods can be used to establish the thermal
the manufacturer’s datasheet using particle swarm optimization model for power electronic devices. The first approach is based
(PSO) method for constructing the transient thermal impedance
on analytical methods, which is achieved by solving analytical
(Zth). The (ET) model procedure is discussed and the
comparisons between simulations and with those provided by the
1D or 2D heat condition diffusion equation [3]. In this method,
manufacturer’s datasheet for the static and thermal the estimation of the heat spreading effect might be obtained
characteristics behaviors to validate the proposed model are inaccurately, involving a typical heat spreading angle of 45°
presented. Further, dynamic study of the (Tj) and the junction [4]. The second approach is based on numerical methods. It is
temperature fluctuation (∆Tj) is also performed by considering mainly based on the finite elements method (FEM) that uses a
the variation of the switching frequency (fsw), this process detailed model of the device, and at the same time, it takes into
demonstrates the good performance of the proposed model. In consideration the material proprieties and layers [5-
order to show the efficiency of the (ET) model in estimating the 6].Providing the exact solution is the success key advantage of
(Tj), a dc/dc buck converter is used as a case study. Finally, the this method, nevertheless, it is characterized by its low
power source obtained using dc/dc buck converter is efficient and computation speed in case where different system components
allowing the power loss reduction. with carious time scales are present. This results in a difficult
evaluation of the junction temperature evolution in real duty
Keywords—Electro-Thermal Model; Transient Thermal cycle or operating condition. The third approach, the most
Impedance ; MOSFET; Junction Temperature; Particle Swarm commonly used one, is the thermal resistor-capacitor (RC)
Optimization; dc/dc buck converter. network where circuit simulators are used to estimate the
instantaneous junction temperature [7-8]. As a result of
I. INTRODUCTION representing the process by a heat transfer coefficient value,
Power electronic DC-DC converters are used in many this method might not provide a good estimation of the heat
applications, such as, renewable energy generators, power convection from the cold plat to the coolant. The aim of this
systems and hybrid electric vehicles to manage power flow and paper is the improvement of electro-thermal (ET) model for a
to regulate voltage. With the development of distributed low voltage power MOSFET for estimating the junction
renewable energy generation, Small-scale inverter becomes a temperature (Tj). Firstly, a thermal model for the (ET) is
focus of research. Being able to convert all kinds of distributed improved on the basis of its transient thermal impedance
energy into a unified DC/DC power, the grid connected (Zthjc), the parameters of the (Zthjc) are extracted from the
inverter extensively broadens the application range of data provided by the manufacturer’s datasheet using particle
renewable energies, and thus acts as an indispensable role in swarm optimization (PSO) method. Thereafter, an electrical
the renewable energy generation system [1,2]. In renewable model of the (ET), which provides accurate predictions of the
energy applications, DC/DC converters can be used to obtain a junction temperature (Tj) dependent electrical characteristics,
well-regulated output voltage from low voltage power source, this (ET) model is implemented in PSpice circuit simulator.
such as, photovoltaic array and wind turbine. The DC/DC Furthermore, dynamic study of the junction and case
converter is based on MOSFET devices, the characteristics and temperature is performed by considering the variation of the
performance of a MOSFET can be considerably affected by switching frequencies (fsw) which shows the good electro-
junction temperature (Tj).Therefore, a fast and precise transient thermal performance of the proposed model. Finally, to better
thermal model based on the junction temperature dependence illustrate the capability and attractiveness of the electro-thermal
of the power loss is critical for proper thermal management of model in estimating the junction temperature, a dc/dc buck
converter is used as case study.
Copyright - 2019
Page 197
ISSN 1737-9334
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
A. Conduction Losses
The MOSFET conduction losses (Pco) are generated by the
on-state drain-source voltage drop VDS across the power
device and the instantaneous value of the current ID, which is Fig. 1. Diagram of the (ET) coupling simulation.
flowing through it, as shown in Fig. 2, are calculated as:
1 Toff Ton
Pco t VDS t .I D t dt
Ton Toff
Therefore, the instantaneous value of the MOSFET
conduction losses is:
1 Tsw
PcoM PcoM t dt RDSon .I Drms
2
Tsw 0
where IDrms is the rms value of the MOSFET on-state
current.
B. Switching losses
Switching loss (Psw) is the power dissipated during the
turn-On and turn-Off switching transitions. These losses are
related to the off-state blocking voltage VDD, the
instantaneous drain current ID, the switching frequency fsw
and the (Tj). They are calculated using the equation (5).
Psw Eon Eoff . fsw
where Eon and Eoff represent the turn-on and turn-off
energies respectively. The estimation of the Eon and Eoff is Fig.2. Current, voltage and Energy losses characteristic at turn-on and turn-
done by: off of the device
T j (t ) Tc (t )
Zth jc t
Pd
where Tj(t) is the junction temperature, Tc(t) the case
Fig. 2. Transfer characteristics comparison between the simulation results and temperature and Pd is the average dissipated power. When
datasheet at different temperatures. using the Foster network, the transient thermal impedance
curve can be fitted into a series consisting of a finite number of
exponential terms as given in (11).
n t
Z th jc t Ri 1 exp
i 1 i
with i is the thermal time constant.
i Ri Ci
Zthjc (°C/ W)
2 2 i 2
xi
K 1
xiK vi
K 1
1.5 Improved Model
where K is the constriction factor of the search, c1 and c2 Datasheet
1
are positive constants classed as acceleration coefficients, w is
the inertia weighting factor and rand1, rand2 are two functions
that generate the inherent randomness in the PSO. xi represents 0.5
the position of the ith particle, pbesti is the best previous
position of xi and gbest is the best previous position among the 0 -4 -3 -2 -1 0
members of the population chosen at random as informants. vi 10 10 10 10 10
is the velocity of particle xi. Time (s)
Fig. 5. The improved simulated transient thermal impedance curve and
D. Implementation of the PSO Algorithm the datasheet curve.
In the present work, the 4rd order thermal equivalent (RC)
The convergence characteristic of the proposed PSO is
Foster model pair parameters are extracted by fitting the step
shown in Fig 6. From this figure, it is can be observed that
response equation given in (11) to the Zth curves. The (PSO)
fitness function before 200 iterations is unstable; After 200
curve fitting algorithm is used for this purpose. The Ri and τi
iterations, no significant improvement is observed, thus the
parameters are determined (11) using PSO algorithm with set
optimized fitness function parameters can be obtained prior to
population size as 10; inertia weight as between 0.4 and 0.9;
the total training time taken (76.79 sec).
and acceleration factors c1 and c2 have been taken as equal to
2 with a maximum iteration set to 500. In order to estimate the 2
thermal impedance curve provided in the device datasheet
according to Fig.5 the results showing a very good matching
between the proposed method and the data provided by the
Fitness function value
1.5
manufacturer’s datasheet. The extracted parameters of the
Foster model are listed in Table 1.
1
TABLE I. LIST OF EXTRACTED PARAMETERS USING PSO ALGORITHM.
№ Ri (C / W) i ( s)
0.5
1 235.2314 10-6 76.3912 10-6
2 812.3754 10-3 1.7798 10-3
3 1.2408 17.5243 10-3 0
0 100 200 300 400 500
4 1.1465 6.8955 10-3 Iteration
Fig. 6. PSO convergence characteristic.
Copyright - 2019
ISSN 1737-9334 Page 200
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
Yes
If k <= Maxite ?
k = k+1
No
(b)
End Fig. 9. The power dissipated in the MOSFET (a), the MOSFET junction
temperature estimates estimated (b).
Fig. 10 shows that for a switching frequency (fsw) less than
Fig. 7 – Flowchart for particle swarm optimization algorithm.
50 kHz, the proposed model is not override the maximum
admissible ones of junction temperature. If the switching
VI. ESTIMATION OF JUNCTION TEMPERATURE frequency (fsw) is increased up to 70 kHz, the amplitude of the
To provide a realistic power loss waveform for junction temperature fluctuation becomes larger and the thermal
temperature evaluation, a dc/dc buck converter is constructed limitations of the device are exceeded, this involving device
in the PSpice software environment, using the electro-thermal failure. Fig. 11 presents the maximum junction temperatures
model, as shown in Fig. 8. The buck converter circuit, used (Tjmax) and swinging amplitudes (∆Tj) of the junction
during the simulation, has been designed to operate at the gate temperatures of the proposed (ET) model. Both the maximum
voltage VGS= 4.5V, the input voltage of buck converter junction temperatures and their swinging amplitudes (∆Tj)
Vin=15V, the Switching frequency of 10 kHz, duty ratio D= appear to increase linearly with increasing of the switching
50%, the inductance L = 50 µH, the output capacitance C = 50 frequency (fsw).
µF, and the electrical resistance R=4 ohm. The power loss 160
profile and the corresponding junction temperature variations
during the first 300 µs of simulation time are shown in Fig.9.
120
Tj(°C)
25 KHz
80 50 KHz
75 KHz
40
Fig. 8. Coupled electro-thermal simulation for a dc/dc buck converter.
0 0.01 0.02 0.03 0.04 0.05
The dynamic behavior of the junction temperature (Tj) and Time (s)
the corresponding power loss profile based on the proposed
(a)
electro-thermal model are shown in Fig. 9. It can be observed
Copyright - 2019
ISSN 1737-9334 Page 201
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
160
REFERENCES
[1] Faranda, R.S.; Hafezi, H.; Leva, S.; Mussetta, M.; Ogliari, E. The
150 optimum PV plant for a given solar DC/AC converter. Energies 2015, 8,
25 KHz
50 kHz pp.4853–4487, 2015.
75 KHz [2] Han, Y.; Chen, W.; Li, Q. Energy Management Strategy Based on
140
Tj(°C)
Copyright - 2019
ISSN 1737-9334 Page 202
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
#
Unité de Développement des Equipements Solaires-UDES/CDER
Bou Ismail, Tipasa- Algérie
2hbendjebbas@gmail.com
#
Unité de Recherche Matériaux Procédés et Environnement-U.M.B.B
Université M'Hamad Bougara Boumerdès- Algérie
3mohammedik@yahoo.com
Abstract— this work deals with a technico-economic study of the Tran and Chen (2016) [5]used data collected between 2005
production of the 5.1 MW wind farm located in Tebessa region. and 2011 (7-year data) to assess wind potential available on
The wind potential available at the Tebessa site used the 03-year Phuquoc Island in Vietnam and to calculate the theoretical
wind data from the measurements taken by the National energy production of a farm of 75 MW. Recently, Abdeslame
Meteorological Office (NOM) during the period 2014-2016. This
evaluation of the wind potential was carried out using the WAsP
et al. (2017), [6] used 10-year data to estimate electricity
OWC software, and a study of the wind potential assessment production wind farms installed on four windy sites in the
available in another region of the Algerian East Highlands Region Algerian highlands. Jourdier and Drobinski (2017), [7] used
was carried out. The results were used to predict the four-year observations to examine the Weibull distribution in
performances of wind farms that would be located in this selected assessments of wind resources and energy efficiency.
site. The obtained results were then used to calculate the cost of Ramadan (2017), [8] used two-year data to measure wind Firm.
the kilowatt-hour wind generated in this site. Energy and Resource Assessment in the Sinai Peninsula, Egypt.
Keywords— wind Park, WAsP, arid areas and economics Our vast country is subdivided into two distinct geographical
areas. The North Mediterranean is characterized by a coastline
I. INTRODUCTION of 1200 km, the two mountainous chains of the Atlas and the
Wind energy is one of the most developed renewable energy vast region of southern Algeria, characterized by the Saharan
resources. According to the Bloomberg New Energy Finance climate. Curb the rural exodus to the north and set populations.
Institute's (BNEF) report, cumulative global production of In view of solar energy potentials, geothermal and wind farms
solar and wind energy reached 1 TW in 2018, with a slight in southern Algeria, these regions are generally suitable for the
dominance of wind power by 54% . The installed wind implementation of energy systems based on renewable
capacity currently in the world is appreciable and reached 539 energies. The selected site is Tebessa and the wind data used
Gigawatt at the end of 2018. This growth is mostly due to the are records of 3 years of measurements carried out between
decrease of the cost of electricity from renewable sources, 2014 and 2016. The wind potential assessment is carried out
mainly electricity from solar PV and wind. Moreover, as using the WAsP OWC utility based on the statistical
reported by (Amine Akbi et al. 2016), it is recognized that the distribution of Weibull. The results were used to predict the
use of renewable energies can significantly reduce carbon performance of wind farms that would be located in this
dioxide or equivalent (CO2eq) emissions compared to their selected site. The obtained results were then used to calculate
fossil equivalents[2] . Therefore, wind energy is clean and the cost of the kilowatt-hour wind generated in this site.
sustainable; it does not release hazardous substances into the
environment and does not generate waste, wind energy can II. METHODOLOGY
contribute to mitigate climate change. a brief overview of the A. wind data analysis
main recent scientific work related to the analysis of
meteorological data and the determination of available wind The WAsP allows spatial interpolation from ground-based
resources in a site, study by Boudia and Guerri (2015) [3] and data while taking into account topography, roughness and
Belabes et al. (2015) [4] who evaluated the available wind obstacles. It has been used to establish wind atlas of several
potential in different regions of northern Algeria and countries and the European wind atlas [8] reported that the
Highlands using 10-year data. An estimate of the cost per prediction model of the wind resource implemented by the
kilowatt hour of wind was given in both books. wind speed f (V) given by Equation (1) [10] and [11].
k V
k 1
V k - Com is the operation, maintenance and repair costs (O&M
. f (v ) exp (1) costs)
A A A
- S is a scrap factor
account regardless of whether the money has been or will be Nominal power: Pn [KW] 850
paid or received in the past or in the future through deduction 52
Rotor diameter: D [m]
of accrued interest (discounting) of all payments flows to a
Pylon height: H [m] 55
common reference time [14]
Starting speed: Vd [m /s] 4
(4)
Where: IV. RESULTATS
- I is the CAPEX (Capital Expenditure) or the investment cost.
It comprises wind turbine price and other initial costs, A. WIND DATA STATISTICS
including provisions for civil work, land, infrastructure,
The results obtained at 10m height above the ground level are
installation and grid integration.
depicted fig. 1. It is found, that for the site of Tebessa, the
- r and i are the interest and inflation rate, respectively
average wind speed is of the order of 4.44 m/s with the north
- t is machine lifetime
and North-West as dominant sector and a power equal to
120w/m2.
B. WIND DATA STATISTICS AND TECHNICAL ANALYSIS Fig.2 Variation of the annual mean wind speed and density of power in E
Oued area, at 10m a.g.l
RESULTS
Fig.2 illustrates the annual change in mean wind speed
and power density at 10 m above ground level, for the Tebessa C. WIND MAP:
site. This figure shows that during three years, worth of data Fig 3 obtained by using the WAsP software illustrates the map
acquisition from 2014 to 2016, a decreasing of these quantities of the wind speed at 55 m above ground level (Wind turbine
was recorded. The average wind speed varies from a heght), of the wind turbine at the study site. The average
maximum of 5.56 m/s in 2014 to 4.32 m/s in 2016. With annual wind speed varies from 3.5 to 9 m/s.
regards to the power density, this figure shows that it varies
from 174 W / m2 with a maximum value in 2014 at 1120W/m2
in 2016.
Fig.1 Wind speed histogram (left) and wind rose (right) of the four area at 10 AEP Gross AEP Proportional Mean Power
m a.g.l. [GWh] net wake loss speed density
[GWh] [%] [m/s] [W/m2]
20.852 20.725 0.61 8.79 846
- The lifetime of wind turbines is 20 years Sustain. Energy Rev., vol. 51, pp. 1245–1255, 2015.
- Operating and maintenance costs are estimated at 25% the [5] V. T. Tran and T. H. Chen, “Wind energy resources
initial invest- ment cost (equipment cost/lifetime). on Phuquoc Island, Vietnam,” Energy Sources, Part
A Recover. Util. Environ. Eff., vol. 38, no. 11, pp.
- Inflation and interest rates are estimated at 8% and 6%
1612–1619, 2016.
respectively.
- The value of the parameter S is set to 10% of the [6] D. Abdeslame, N. Kasbadji Merzouk, S. Mekhtoub,
investment cost. M. Abbas, and M. Dehmas, “Estimation of power
The LCOE of wind energy produced in the site study varies generation capacities of a wind farms installed in
between 3.28 to 4.23 c$/kW windy sites in Algerian high plateaus,” Renew.
Energy, vol. 103, pp. 630–640, 2017.
2rhachemi8@yahoo.fr
3bendib1427@yahoo.fr
4belmilih@yahoo.fr
5nawelseddaoui@yahoo.com
Abstract— Hybrid power systems are effective and reliable energy electricity [3,5]. Instead of using power from wind energy
sources; they can better cover electricity demands for remote and alone or solar energy alone, combining both sources provides
small areas. However, the difficulty in energy management a better and more reliable system.
between different energy sources, the storage system and the load Generally speaking, HPS can be stand-alone or grid-tied.
is more challenging. This paper outlines the development of an
energy management control scheme for a standalone hybrid
In standalone systems, both sources together can supply power
photovoltaic/wind power system. The system under study is based and if any decrease in power occurs, in any one, the other
on a real experimental platform. Seen from the photovoltaic part, source can compensate for the first. Also, both systems
a sliding mode controller is developed to ensure an optimal together can supply higher loads if required. Whenever the
operation for the whole system under different possible generated power is in excess, it can also be stored in a battery.
generation and load profiles. The scenarios considered are when For reliable supply of power in remote locations or
the load demand is satisfied and when the demand is not satisfied, inaccessible rural areas, it is necessary to design and set up
in the first case the controller should provide PV energy to keep hybrid system, which combined the advantage of two different
the charging mode of battery optimal and to ensure energy energy technologies. These could be either two renewable
balance while in the second case, the controller should ensure
maximum power capture to remedy the insufficiency of the load
technologies or a renewable and a conventional energy or
demand. Simulation results using real meteorological data from fossil fuel technology [5,6].
the site and load profile from the laboratory are conducted to The main challenge in standalone PV/Wind HPSs is the
show the effectiveness of the proposed control strategy. management of the power flow between renewable sources,
the battery system and the load [2,4,5,6]. An efficient energy
Keywords— Hybrid PV/Wind, Distributed generation, Energy flow management strategy should be able to ensure maximum
management, Sliding mode, Buck converter. power delivery when there is a scarcity of production and large
load demand. Besides, it should ensure optimal battery
I. INTRODUCTION operation to optimize its lifecycle.
Electrical energy is an essential factor for the human The energy flow management in renewable HPS has been
development and welfare [1]. Conventional electrical energy studied by many researchers from the industrial and academic
generation is mainly based on fossil fuel resources, However fields as can be witnessed by the abundant research literature.
the exhaustive nature of these resources, their negative impact In fact, many research papers and deal with the control and
on the environment and their increasing prices have somehow energy management of hybrid standalone and grid-tied power
weakened their reliability [2,3]. In this context, the interest in systems, these methods are based on classical and nonlinear
renewable energy sources (RES: photovoltaic modules, wind control such as sliding mode [3,4] or based on soft computing
turbine, biomass, geothermal, etc.) has increased since these techniques such as neural networks such as in [2,5,6].
technologies can remedy the drawback of conventional energy This paper presents an energy flow management control
generation sources. One main drawback of renewable energy strategy for a standalone PV/Wind multi-source power system
sources is the intermittent nature of the resources such as wind with a 4kW PV array and a 2kW Wind subsystem. For this
speed and insolation [2,3,4]. purpose, the sliding mode methodology is applied to the
Hybrid power systems (HPS), which combine multiple photovoltaic side to ensure optimal operation for different
renewable and possibly conventional energy sources with scenarios (surplus or a deficit of energy). The proposed
storage elements (batteries, inertial wheels, fuel cells, etc.) are controller should be able de decide the operating mode and to
very interesting in terms sustainable and reliable electrical fulfil the control specifications for each mode. The paper is
energy production as well as the reduction of environmental organized as follows, in section II, the nonlinear model of the
impacts [3]. According to experts, a HPS that combines solar HES is developed and the control objective is outlined, in
and wind electric technologies offers several advantages over section III, the operational modes of the system and controller
either single system [3,4]. In other words, HPSs, can provide strategy developed, in section IV, simulation results using the
an economic, environment friendly and reliable supply for parameters of the real system along with real measured
meteorological data is carried out and the results are discussed, B. Photovoltaic subsystem modelling
section V presents conclusions on this work and perspectives The photovoltaic subsystem converts solar radiation from
for future works. the sun into electricity, the I-V characteristic of a solar cell is
given by [7,8]:
( )
II. HYBRID POWER SYSTEM DESCRIPTION AND MODELLING
q ( v + i R ) AKTc
In this section the dynamic state space model of the HPS i pv = I ph − I rs e pv pv s −1 (1)
(PV/Wind) is developed. The model is based on the PV/Wind where q is the electron charge in columns, A is the diode’s
platform installed on the roof of the SMS laboratory in the
UDES at Bou-Ismail in Algeria (Fig. 1). ideality factor, K is the Boltzmann’s constant and Tc is the
cell’s temperature, I ph is the photocurrent under a given
A. Hybrid Power System Description
insolation, and I rs is the cell reverse saturation current given
The HPS is built around a 48V DC bus topology, it is made
up of the following subsystems (Fig. 2): respectively by:
• Photovoltaic array based on STP-135 Suntech, STP 135 tb I ph = I sc + Ki (Tc − Tref ) G (
(2) )
modules, the array arranged into 5 parallel strings, each string ( )
I rs = I 0 (Tc Tref )
3 qEg 1 Tref −1 Tc KTc
consists of 6 series modules connected to the 48v DC bus via e (3)
Tristar DC/DC converters. where I sc is the cell’s short circuit current at reference
• Wind energy conversion system (WECS) based on 2
temperature and insolation, K i is the short-circuit temperature
Whisper 200 wind turbines (WT), these WT are based on a
PMSG with a rated power of 1kW, they are connected to the coefficient, Tc is the cell’s temperature in kelvin and G is the
48V DC bus via dedicated Whisper controllers. insolation measured in kW/m2.
• Storage subsystem, based on 24 series connected Classic The PV subsystem is our HES composed of N p parallel
OPzS 200 Lead Acid batteries with 2 V for each cell.
strings, with N ss series modules in each string, each module is
• Domestic and laboratory AC loads connected to the 48V DC
bus via a DC/AC converter (XTH-8000-48 Steca inverter). composed of N sc series cells, this makes a total of N ss N sc
series cells, thus the I-V curve for the whole system:
( )
In the following, mathematical models for each subsystem
q ( v + i R ) N N AKT
and the whole HPS will be derived. i pv = N p I ph − N p I rs e pv pv s cs ss c − 1 (4)
50
G=1000 G=800 G=600 G=400
40
Array current Ipv (A)
30
20
10
0
0 20 40 60 80 100 120 140
Array voltage Vpv (V)
Fig. 1 The experimental PV/Wind HPS (UDES, Bou-Ismail)
Fig. 3 Current vs. voltage curve of the HES PV array
5000
G=1000 G=800 G=600 G=400
4000
Array power Ppv (W)
3000
2000
1000
0
0 20 40 60 80 100 120 140
Array voltage Vpv (V)
And the corresponding power will be given by the wind subsystem is connected to the DC bus via whisper
( ( q ( v + i R ) N N AKT
p pv = v pv N p I ph − N p I rs e pv pv s cs ss c − 1 )) (5)
controller, the load is connected via an inverter and the PV
subsystem is connected via DC/DC buck converter as outlined
Typical I-V and P-V curves for the PV subsystem under in Fig. 5. The switched dynamic model for the buck converter
different conditions of insolation and a cell temperature of 20° can be written as:
are shown in figures Fig. 3 and Fig.4 respectively. i pv i0
v pv = − u pv (9)
C C
C. Wind subsystem modelling:
v v pv
The wind subsystem consists of 2 Whisper 200 wind is = − b + u pv (10)
turbines (WTs), each one has rated power of 1 kW. The power L L
generated by a WT is related to the wind speed by [1,9,10]: 1 1
vc = ib − vc (11)
1 Cb Cb Rb2
pwt = R 2C pV 3 (6)
2 where the different electric variables and parameters are
where V is the wind speed, C p is the WT power coefficient, illustrated in Fig.5. with their values in the appendix. The
is the air density and R is the rotor radius. The power battery voltage is given in (7).
Equations (7)-(11) can by assembled together to build the
coefficient C p describes the conversion efficiency of with
complete nonlinear state space model of the HES as follows
turbine, it has a theoretical limit of 0.69 called the Betz limit
x = f ( x) + g ( x)u (12)
[1]. Notice that in real WTs, this limit is never reached, typical
values for this limit are around 0.4 [1,9], for the case of our with:
Whisper 200 turbine, the variation of C p is shown in Fig.5, it v pv i pv C − is C
has a limit around 0.39, which is inferred from the x = is , f ( x ) = − vb L ( ) v pv L
, g x =
manufacturers datasheet. vc ib Cb − vc Cb Rb 0
Usually, wind turbines are usually characterized by their 2
power curves, the typical power curve for the Whisper 200
wind turbine is shown in Fig. 5. For our study, The WT model The control objective is to design a control law for the IGBT
consists of a look-up table generated from the power curves gate trigger u pv of the DC/DC buck converter in order to meet
given in the manufacturer’s data-sheet. an optimal energy balance between the storage system (battery
D. Battery and Converter System Modelling: bank), the load and sources (PV and Wind generators) [4].
The HPS is built around a 48V DC bus, the battery is III. ENERGY MANAGEMENT CONTROL DESIGN
directly connected to the 48V bus, for the purpose of our study, In this section the control law is design in order to ensure
the battery is modelled as a shown in Fig.6. ,it consists of a the energy balance between the different constituents of the
voltage sources Eb connected in series with a resistance Rb1 a HES, for this purpose the required operation specifications are
resistance Rb1 and capacitance Cb , the battery voltage is : [11] first outlined, then the control law is derived, notice the control
law a slightly modified version of that presented in [4].
vb = Eb + vc + ib Rb1 (7)
where ib is the battery current which is positive when the
battery is charging and negative when the battery is
discharging, it is given as follows:
ib = is + iw1 + iw2 − iL (8)
Fig. 5 Power curve and power coefficient of the Whisper-200 WT Fig. 7 Schematic of the DC bus of the HPS
p pv 1000
|mpp = 0 (16)
Insolation (w/m2)
v pv 900
( i pv v pv ) i pv 700
|mpp = v pv |mpp + i pv =0 (17)
v pv v pv mpp 600
500
Therefore, for the second mode of operation, the sliding 0 50 100 150 200 250 300
Time (s)
manifold that should be regulated around zero can be derived
from (19) as follows:
Fig. 10 Measured insolation curve for 5 minutes
30
29
Temperature ( C)
28
27
26
0 50 100 150 200 250 300
Time (s)
Fig. 11 Measured temperature variation for 5 minutes Fig. 14 Gate trigger signal variations
10
Wind speed (m/s)
0
0 50 100 150 200 250 300
Time (s)
Fig. 12 Measured wind speed for 5 minutes Fig. 15 Variations of the sliding surfaces s1 and s2
Fig. 13 Load power profile variations for 5 minutes Fig. 16 DC bus (battery) voltage variations
The simulation results using these measured data are shown Figures Fig.18 and Fig.19 shows the energy balance
in Figures Fig.14 – Fig.19. Fig.14 shows the switching signal between the produced energy from the PV and Wind systems
for the converters Gate, it can be seen that pulse width varies and the power consumption from the load and the power fed to
to ensure an optimal operation as required by the two modes of or from the battery. We see that our controller ensures an
operation stated above. energy balance in the first scenario where the production tracks
the needed power. While in the second mode of operation, even
Fig.15 shows the variations of the two sliding manifolds s1 though, the system is operated at its MPP, the energy balance
and s2 , we see that for the first mode of operation s1 is kept is not always satisfied when the needs for the load are very high.
There is always an energy deficit. This energy deficit can be
around zero while s2 is disregarded while in the second mode practically talked by a proper sizing of the whole HPS. Since
of operation, s2 is kept around zero and s1 is disregarded, in our study we used a theoretical profile for the load we cannot
which is in accordance with the control algorithm. Fig.16, conclude as if the system is property sized or note.
shows the battery voltage variation, when there is a surplus of In conclusion we see that the proposed control architecture
energy production, the battery will be charged to it its full state is elegant and efficient. and can ensure an optimal operation of
(58 V) and when there is a power deficit, it will be drawn from the system in both modes of operation, it has the ability to
the battery which lead to its discharge. Fig.17 really describes automatically decide the required operational mode and switch
the behaviour of our control strategy, we see that the operating to it so that the design requirements are satisfied. Besides, the
point for the PV system swings back and forth between the two simplicity of the proposed control scheme encourages its
modes of operation depicted by the two distinct regions in the experimental implementation on the real system existing on the
power curve. SMS lab in the UDES, Bou-Ismail.
Copyright - 2019
ISSN 1737-9334 Page 212
7ème Conférence Internationale des Energies Renouvelables (CIER-2019)
Proceedings of Engineering and Technology PET
REFERENCES
[1] I. Munteanu, A. I. Bratcu, N. A. Cutululis, and E. Ceanga, Optimal
Control of Wind Energy Systems: Towards a Global Approach,
Springer, 2008.
[2] S. Sumathi, L., A. Kumar and P. Surekha, Solar and Wind Energy
Conversion Systems, Springer, 2015.
[3] Jan Moravek, petr Mastny, “Hybrid Renewable Energy System-
Configuration and Control”, Recent Researches in Electric Power and
Energy Systems, Electrical Engineering Series. Chania, Crete Island,
Greece: p. 87-92, 2013.
[4] F. Valenciaga, P. F. Puleston and P. E. Battaiotto, “Power Control of a
Fig. 17 Variations of the PV operating point between the two modes Photovoltaic Array in a Hybrid Electric Generation System using
Sliding Mode Techniques”, IEEE Proceedings-Control Theory and
Applications, Vol. 148 Issue: 6, 2001.
[5] S. Zerkaoui, AK El Hajjaji, J. Bosche, “On-line Control Strategy for
instantaneous Power Management of Hybrid Power System Based on
Dynamic Fuzzy Logic Controller”, in the 7th IEEE Conference on
Industrial Electronics and Applications (ICIEA), 2012.
[6] A. Gligor, C. D., Dumitru and H. S., Grif, “Artificial intelligence
solution for managing a photovoltaic energy production unit”, Procedia
Manufacturing, Vol. 22, pp. 626-633, 2018.
[7] J. Khanam, S. Y. Foo, “Modeling of a Photovoltaic Array in Matlab
Simulink and Maximum Power Point Tracking using Neural Network”,
Electrical and Electronic Technology Open Access Journal, Vol. 2,
Issue 2, pp.40-46, 2018.
[8] H. Tsai, C. Tu, and Y. Su, “Development of Generalized Photovoltaic
Model Using MATLAB/SIMULINK”, in Proceeding of the World
Fig. 18 Produced power vs needed power Congress on Engineering and Computer Science, San Francisco, USA,
2008.
[9] N. Bounar, S. Labdai, A. Boulkroune, M. Farza and M. M’Saad,
“Adaptive Fuzzy Control Scheme for Variable-Speed Wind Turbines
Based on a Doubly Fed Induction Generator”, Iranian Journal of
Science and Technology-Transactions of Electrical Engineering,
October, 8th, 2019. https://doi.org/10.1007/s40998-0.19-00276-6.
[10] Jan. Moravek, Petr Mastny, “Model of PMSG Wind Turbine in Hybrid
Energy System”, in Proceedings of the 15th International Scientific
Conference on Electric Power Engineering, Czech Republic, 2014
[11] B. S. Borowy and Z. M. Salameh, “Dynamic Response of a Standalone
Wind Energy Conversion System with Battery Energy Storage to a
Wind Gust”, IEEE Transactions on Energy Conversion, Vol. 12, N. 1,
pp.73-78., 1997.
Copyright - 2019
ISSN 1737-9334 Page 213