978-1-6654-1741-9/22/$31.00 ©2022 IEEE
Using ANN based MPPT controller to increase PV
central performance
Elaid Bouchetob
Faculté des hydrocarbures et de la chimie, Laboratoire
d’électrification des entreprise industrielles, LREEI
Université de M’hamed Bouguerra,
Boumerdes, Algeria
e.bouchetob@univ-boumerdes.dz
Bouchra Nadji
Faculté des hydrocarbures et de la chimie, Laboratoire
d’électrification des entreprise industrielles, LREEI
Université de M’hamed Bouguerra,
Boumerdes, Algeria
b.nadji@univ-boumerdes.dz
Abstract—Recently, searches for renewable energy and
artificial intelligence have grown in popularity. Photovoltaic (PV)
technology is an ecologically beneficial and cost-effective source
of power, since it requires fewer components for maintenance
than other green energy sources. Slow reaction to rapid
variations in solar temperature and irradiance may render the
MPP incapable of tracking. In addition, as a consequence of the
deficiency of conventional approaches (P&O, IC, etc.), artificial
intelligence (AI) methods such as (ANN, FLC, ANFIS, etc.) are
becoming increasingly popular and are used to monitor the MPP
in PV systems. Each of these methods has benefits. Utilizing the
benefits of the DC-DC converter and the ANN-based MPPT
controller, our work proposes a novel research and design to
improve the performance and efficiency of the Ghardaia SKTM
installation (Module 1) by leveraging the DC-DC converter and
the MPPT controller.
Keywords— PV system; ANN; MPPT; DC-DC converter;
simulink.
I. INTRODUCTION
Investors and academics are developing renewable energy
technologies to balance the energy shortfall and reduce
pollution caused by fossil fuel emissions [1]. Photovoltaic (PV)
technology is a low-cost, ecologically beneficial, and low-
maintenance energy source [2]. Non-linear panel characteristics
are caused by variables like solar radiation, temperature, and
shading (I-V and P-V). In spite of this, the final set of curves
reveals just a single peak power point, designated MPP.
Several strategies and procedures may be used to increase the
PV tracking efficiency, but maximum power point tracking
(MPPT) has been rated the most effective and popular.
Additionally, P&O[5] is being developed as part of these
strategies. fractional open-circuit voltage[6], IC[7], NN[8], and
FLC[9].
Artificial intelligence offers a more exact and quick
reaction to changes in irradiance and temperature as compared
to traditional MPPT methods such as P&O (IC, etc.). [10] In
comparison, conventional MPPT approaches include: An
artificial neural network, often known as an ANN, is a piece of
software for computers that makes use of many forms of
intelligence to rapidly link Irradiance and Temperature to
output (Vmpp)[11]. The approach known as maximum power
point tracking (MPPT), which is based on artificial neural
networks (ANNs), makes use of the benefits offered by ANNs,
including their noise rejection capabilities and the fact that they
do not require any prior knowledge of physical factors that are
important to PV systems.
Module 1 is the primary focus of our study on this
paper. We designed an entirely new connection by making use
of artificial intelligence in order to increase the quantity of
energy that is supplied into the inverter. Our goal was to
increase the amount of power that was available. The output
power of an ANN-based MPPT controller was measured and
compared to the actual power generated on a cloudy winter day
for the purpose of this investigation.
II. MATERIAL AND METHODE
A. Present Ghardaia SKTM central
Sonelgaz Algeria started exploiting electric energy from
Ghardaia SKTM PV central in 2015; this last consists of 8
modules with 4 different technologies of PV panels (Mono-
crystalline, Multi crystalline, Amorphous, and Couch mince).
The modules (1, 5 and 7) use the technology "mono-
crystalline’’ with (105, 105, 255) KW of power production in
successive; the modules (2, 6 and 8) use the technology "Multi-
crystalline" with (98, 98, and 263) KW of power in successive;
and module 3 uses the "Thin layers" technology with 100 KW
of power. Finally, module 5 uses the "Amorphous" technology
with 100 KW of power. The output of those modules has
connected with eight inverters.
B. Proposed design
Module 1 has 420 panels. Every 20 panels are linked in
series in a channel, resulting in an output voltage of 607 V at
MPP. The parallel connection of 21 channels in a single
junction box with an output current of 173A at MPP yields a
module output power of 105 kW. However, we presented a
new matrix in which 10 panels are connected in series for each
channel (output at MPP is 303.5 V), and the junction box
contains 42 channels with 346 A at MPP. Figure 1 illustrates
both designs:
2022 2nd International Conference on Advanced Electrical Engineering (ICAEE) | 978-1-6654-1741-9/22/$31.00 ©2022 IEEE | DOI: 10.1109/ICAEE53772.2022.9961982
Authorized licensed use limited to: Consortium - Algeria (CERIST). Downloaded on December 24,2022 at 22:54:11 UTC from IEEE Xplore. Restrictions apply.