Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)

ISSN: 2278-0181
ETRASECT - 2016 Conference Proceedings

Solar MPPT Charge Controller with ANN

Controller
Dr. B. L. Mathur
Akshay Kumar Atrey HOD(PG)
Department of Electrical engineering Jodhpur Institute of Department of Electrical engineering Jodhpur Institute of
Engineering and Technology Engineering and Technology
Jodhpur, Rajasthan, India Jodhpur, Rajasthan, India.

Abstract— Photovoltaic generation is the technique which uses The Hill Climbing (HC) and the Perturb and Observe (P&O)
photovoltaic cell to convert solar energy to electric energy. are the most known and commercially used techniques. Other
Nowadays, PV generation is developing increasingly fast as a
modified methods such as the incremental Conductance
renewable energy source. However, the disadvantage is that it
has very low efficiency and PV generation is intermittent (INC) technique, the neural network (ANN) technique, and
because it depends considerably on weather conditions. This fuzzy logic controller technique, have been also reported to
paper proposes an intelligent control methods for the maximum improve the performance of these techniques. In HC-MPPT
power point tracking (MPPT) of a photovoltaic system under technique, the duty cycle is directly incremented or
various conditions. In this paper, a simulation study of the decremented in fixed steps depending on the panel voltage
maximum power point tracking (MPPT) for a photovoltaic and power values until the maximum power point (MPP) is
system using an artificial neural network is presented. The reached. The P&O technique shares the same HC concept of
system simulation is elaborated by combining the models operation, but with an additional PI control loop. In the P&O,
established of solar PV module and a SEPIC or CUK converter.
the converter input reference voltage is the perturbed variable
Finally the performance comparison between artificial neural
network controller and Perturb and Observe method has been and the duty cycle is computed through an additional PI
carried out which has shown the effectiveness of artificial control loop. The additional control loop results in an
neural networks controller to draw much energy and fast increase in the P&O efficiency.
response against change in working conditions.
II. IMPORTANCE OF ANN CONTROLLER
Keywords— Solar Energy, Photovoltaic, MPPT, P&O,CUK
Converter, Artificial Neural Network. The P&O method is commonly used because of its simplicity
and ease of implementation. Furthermore, P&O (with a small
I. INTRODUCTION step size) in nominal conditions can have MPPT efficiencies
mostly the same like other complex techniques, and still
Significant progress has been made over the last few years in easier implementation. However, the drawback of this
the research and development of renewable energy systems technique is that the operating point of the PV array oscillates
such as wind, sea wave and solar energy systems. Among around the MPP. Therefore, the power loss may increase.
these resources, solar energy is considered nowadays as one Furthermore, when the sun insolation changes rapidly, the
of the most reliable, daily available, and environment friendly P&O method probably fails to track the MPP. Another
renewable energy source. However, solar energy systems possible disadvantage is that the MPPT may not be able to
generally suffer from their low efficiencies and high costs. In locate the MPP as the amount of sunlight decreases, because
order to overcome these drawbacks, maximum power should the PV curve flattens out. Recently intelligent control based
be extracted from the PV panel using MPPT techniques to control schemes MPPT have been introduced. In this paper,
optimize the efficiency of overall PV system. MPPT is a real- an intelligent control technique uses artificial neural network
time control scheme applied to the PV power converter in control is associated to an MPPT controller in order to
order to extract the maximum power possible from the PV improve energy conversion efficiency. The simulation can
panel. The MPPT working principle is based on the generate two different solutions for the control of converter
maximum power transfer theory. The power delivered from system; one is P&O controller and the other one is ANN
the source to the load is maximized when the input resistance controller. The circuit diagram of the energy conversion
seen by the source matches the source resistance. Therefore, system is shown in Fig.1.
in order to transfer maximum power from the panel to the
load the internal resistance of the panel has to match the
resistance seen by the PV panel. For a fixed load, the
equivalent resistance seen by the panel can be adjusted by
changing the power converter duty cycle .There are various
MPPT techniques based on different topologies and with
varying complexity, cost, and overall produced efficiency.

Fig.1. Schematic diagram of the proposed power conversion PV array.

Volume 4, Issue 12 Published by, www.ijert.org 1

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ETRASECT - 2016 Conference Proceedings

The system consists of photovoltaic panel, a DC/DC CUK I = Iph -Is(e Akt - R
-1) sh (1)
converter, a control unit and a resistive load. The first stage of where, Iph is photocurrent; Is is diode saturation current; q is
the system is solar panel. The I-V characteristic of a panel coulomb constant (1.602e-19C); k is Boltzman‘s constant
depends on the temperature and solar irradiance. The three (1.381e-23 J/K); T is cell temperature (K); A is P-N junction
most important characteristics of PV panel are the short ideality factor; Rs and Rsh are intrinsic series resistances.
circuit current, open circuit voltage and the MPP which is a Photocurrent is the function of solar radiation and cell
function of panel temperature and solar irradiance. The power temperature described as
stage is the CUK converter which duty cycle is regularly S
adjusted to track the maximum power point that can be S
re
delivered by the PV panel. The proposed MPP tracker, which I
ph =( f )[Iph,ref +CT(T -Tref)] (2)
is based on artificial neural networks control, has the
where, S is the real solar radiation (W/m2); Sref , Tref Iph,ref
objective to draw as much power as possible from the PV
is the solar radiation, cell absolute temperature, photocurrent
module by adjusting continuously the duty cycle of the
in standard test conditions respectively; CT is the
DC/DC converter. This point corresponds to the maximum
temperature coefficient (A/K).
power point (MPP) on the PV curve. The main content of this
paper is organized into several sections. Section two is
The relationship of the voltage and current in PV array is:
dedicated to the study of the characteristics of solar panels.
IR
Section three is devoted to the study of the maximum power
q V s Np V IRs
point tracking (MPPT) and section four discusses the
modeling of DC/DC converters. Section five presents the A N N
different MPPT algorithms (P & O algorithm and ANN I =NphIph -NpIs(e kT (s + p ) -1) -Rsh (Ns + Np ) (3)
algorithm) to make a comparison between these two Where, NS and NP are cell numbers of the series and parallel
algorithms. Section six is dedicated to the simulation of the cells respectively.
two methods (P & O and ANN). Finally, a general conclusion
finished the paper. Considering different temperatures and solar irradiations,the
simulated output characteristics of the PV array are depicted
III. BASIC IDEA OF PV CELL in fig.3. and fig.4.

Photovoltaic cell is the most basic generation part in PV

system. Single-diode mathematic model is applicable to
simulate silicon photovoltaic cells, which consists of a
photocurrent source Iph, a nonlinear diode, internal
resistances Rs and Rsh, as shown in Fig.2.

Fig. 2. Single-diode mathematic model of a PV cell

The mathematic relationship for the current and voltage in the

single-diode equivalent circuit can be described as
Fig.3.Characterstics curve of Solar PV module with different solar
irradiations
q(V
+IRs) V +IRs

Volume 4, Issue 12 Published by, www.ijert.org 2

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ETRASECT - 2016 Conference Proceedings

IV. CUK CONVERTER WORKING

Fig.5 shows the electrical circuit of a Cuk converter.The

power switch ‗S‘ is used to modulate the energy transfer
from the input source to the load by varying the duty cycle D.
The relationship between input and output voltages of Cuk
converter operating at steady state condition is given by:

Fig.4.Characterstic curve of Solar PV module with different solar irradiations

A dynamic tracking method is necessary to extract the
Fig.5. Basic circuit of DC/DC Cuk converter
maximum power from the PV cells. Many researches has
been developed concerning the different algorithms for the
The control strategy lies in the manipulation of the duty cycle
maximum power point tracking (MPPT) considering the
of the switch which causes the voltage change. When the
variations of the system parameters and weather changes,
switch is off, the inductor currents iL1 and iL2 flow through
such as perturb and observe method, open and short circuit
the diode. Capacitor C1 is charged through the diode. The
method, incremental conductance algorithm,Tempreture
circuit is shown in Figure: 4.8, Capacitor C1 is charged
method, Tempreture perametric method, fuzzy logic and
through the diode by energy from both the input and L1.
artificial neural network. The block diagram in Fig.2 presents
Current iL1 decreases because VC1 is larger than Vs. Energy
a PV generator with MPPT . The load or the battery can be
stored in feeds the output. Therefore iL2 also decreases.
charged from a PV panel using a MPPT circuit with a specific
controller to track the peak power generated by the PV panel.

Other protection devices can be added. The control circuit

takes voltage and current feedback from the battery, and
generates the duty cycle D. This latter defines the output
voltage of the Cuk converter. Many MPTT control techniques
have been conceived for this purpose these last decades. They
can be classified as:

Voltage feedback based methods which compare the PV Fig.6 when switch S is OFF, Capacitor charging
operating voltage with a reference voltage in order to
generate the PWM control signal of the DC/DC When the switch is on, VC1 reverse biases the diode. The
converter. inductor currents iL1 and iL2 flow through the switch as
shown in Figure: 4.9. Since VC1 >Vo, C1 discharges through
Current feedback based methods which use the PV the switch, transferring energy to the output and L2.
Therefore iL2 increases the input feeds energy to L1 causing
module short circuit current as a feedback in order to iL1 to increase.
estimate the optimal current corresponding to the
maximum power.

Power based methods which are based on iterative

algorithms to track continuously the MPP through the

current and voltage measurement of the PV module. In
this category, one of the most successful used method is
perturbation and observation (P&O) technique.

Fig.7 when Switch S is ON

The inductor currents iL1 and iL2 are assumed to be
continuous. The voltage and the current expressions in steady
state can be obtained in two different ways. If we assume the
capacitor voltage VC1 to be constant,

Volume 4, Issue 12 Published by, www.ijert.org 3

 Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ETRASECT - 2016 Conference Proceedings

perturbation cycle. If a given perturbation leads to increase

(or decrease) the output power of the PV, the successive
then equating the integral of the voltages across L1 and L2 perturbation is generated in the same (or opposite) direction,
over one time period to zero yields. on Fig.10. We consider that the maximum power point MPP.
If the operating point C is on the left of MPP, the duty cycle
The output voltage can be derived by: must be decreased until the MPP is reached. If the operating
point is on the right of the MPP, the duty cycle is increased to
If D=0.5 the output is the same as the input. reach the MPP. The results for different tests using the P&O
Vo = D algorithm are presented and compared to those obtained with
If 0.5<D<1 the output is larger than the input. the artificial neural network (ANN) MPPT controller in sect.
Vs D -1
V. DIFFERENT ALGORITHM MPPT B. ARTIFICIAL NEURAL NETWORK CONTROLLER
It is necessary to constantly track the MPP of a solar panel. For METHOD
the past years, research has focused on various MPP control
algorithms to draw the maximum power of the solar array. In The MPPT strategy proposed here consists of a combination
this section, the effectiveness of two different control of an artificial neural network and the MPPT technique in
algorithms are thoroughly investigated using a numerical order to implement of the duty cycle regulator. When solar
simulation. irradiation changes slowly, the system controls the DC-DC
converter using the P&O, and the neural network learns
A. P&O controller method: Its relationship to the duty
cycle(D) is: simultaneously the MPP found by the P&O. However if the
solar radiation varies too rapidly, the neural network
If 0<D<0.5 the output is smaller than the input. controller tracks the MPP rapidly and adjusts the duty cycle
of the DC-DC converter. Neural networks usually require
independent and identically distributed samples to ensure
successful on-line learning. Here, however, similar training
samples are used by the artificial neural network.

Fig.9. Proposed MPPT system with ANN controller

To deal with these training samples, we have used an MLP in

Fig.8. flow chart P&O method. order to ensure fast and correct learning. The main idea is that
the neural network learns each sample online because it is
The P&O algorithm is the most commonly used in PV systems difficult to store all learning samples in small devices. In Fig.
applications due to its ease of implementation and simplicity. It 11, the ANN learning technique is a memory-based one and
is an iterative method for obtaining the MPP. Whereas, it allows estimating at any instant the required optimal duty
measures a PV module current and voltage, then perturbs the cycle‗D‘. Even with sparse data in a multidimensional
operating point of a PV module to determine the change measurement space, the algorithm provides smooth
direction. Fig.8 shows the flow chart of the classical P&O transitions from one estimated value of D to another. The
algorithm. The P&O algorithm has been broadly used because ANN consists of an input layer (Ppv), a pattern layer, a
of its practical implementation. The MPP tracker operates by summation layer an output layer. The output of the ANN is
periodically incrementing or decrementing the solar panel the duty cycle D(x) as follow:
voltage, current or the duty cycle comparing to the PV output
power with that of the previous

Volume 4, Issue 12 Published by, www.ijert.org 4

Special Issue - 2016 International Journal of Engineering Research & Technology (IJERT)
ISSN: 2278-0181
ETRASECT - 2016 Conference Proceedings

weights of the neurons are changed subsequently to the

presence and severity of a perturbation. They are maintained
constants in stable working condition.

VI. CONCLUSION

This paper discusse Artificial neural network-based MPPT,

Under variation in atmospheric conditions. By using artificial
neural network, point of maximum power is specified fast
Fig.10.MPPT with ANN controller Modeling in MATLAB and precisely. Another advantage of the artificial neural
network in PV maximum power-point tracking is its better
The MPPT technique proposed differs from other techniques dynamic performance in comparison with any the other
in that the duty cycle of the switching of the DC/DC boost methods. Also the maximum power point is tracked by dc-dc
converter is optimally calculated on-line. The algorithm of cuk converter. So the maximum power solar energy and the
the three-point weights comparison is run periodically by best efficiency are obtained.
perturbing the solar array terminal voltage and comparing the
PV output power on three points of the P-V curve. The three VII. ACKNOWLEDGEMENT
points are the current operation point (A), a point B,
perturbed from point A, and a point C, with doubly perturbed This research paper is made possible through the help and
in the opposite direction from point B. Fig.13 depicts the support from everyone, including: parents, teachers, family,
three possible cases. In these cases, for the points A and B, if friends, and in essence, all sentient beings. Especially, please
the power corresponding to the point B is greater than or allow me to dedicate my acknowledgment of gratitude
equal to that of point A, the status is assigned a positive toward the advisor and contributor: I would like to thanks Dr.
weighting. Otherwise, the status is assigned a negative B. L. Mathur sir for his most support and encouragement. He
weighting. Amongst the three measured points, if two are kindly read my paper an offered me valuable advice.
positively weighted, the duty cycle of the converter should be
increased. On the contrary, when two are negatively REFERENCES
weighted, the duty cycle of the converter should be
decreased. In the other cases with one positive and one [1] Christopher J. Lohmeier"Highly Efficient Maximum Power Point
negative weighting, the MPP is reached or the solar radiation Tracking Using a Quasi-Double-Boost DC/DC Converter for
Photovoltaic Systems"University of Nebraska-Lincoln,
has changed rapidly and the duty cycle must not be changed.
[email protected].
Fig.13 shows the idea of the MPP detection algorithm. By [2] San Luis Obispo"DESIGN AND SIMULATION OF
monitoring voltage V and current I, the P&O algorithm PHOTOVOLTAIC WATER PUMPING SYSTEM" 2005.
determines whether generated power has increased. If so, the [3] Alberto Dolara, Francesco Grimaccia, Sonia Leva, Marco Mussetta
and Emanuele Ogliari"A Physical Hybrid Artificial Neural Network
next change in voltage should be the same as the last V. If
for Short Term Forecasting of PV Plant Power Output".
not, the next change in voltage should be negative P&O [4] Shusmita Rahman, Nadia Sultana Oni,Quazi Abdullah Ibn
achieves the function of an MPPT easily, Masud"Design of a Charge Controller Circuit with Maximum Power
Point Tracker (MPPT) for Photovoltaic System”.
[5] Mahmoud A. YOUNIS, Tamer KHATIB, Mushtaq NAJEEB, A Mohd
ARIFFIN"An Improved Maximum Power Point Tracking Controller
for PV Systems Using Artificial Neural Network".
[6] Module 3 DC to DC converter Version 2 EE IIT,Kharagpur.

Fig.11. Characterstic curve with the working of P&O method

but it cannot track MPP rapidly when solar radiation changes

quickly. In order to eliminate this drawback, our MPP
technique utilizes an ANN to achieve learning and maximum
power point tracking. The MPP tracker operates by
periodically incrementing or decrementing the estimated solar
panel voltage Vpv. If a perturbation occurs on the PV output,
then the subsequent perturbation is generated in the opposite
direction. The

Volume 4, Issue 12 Published by, www.ijert.org 5