See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/269279453

A comprehensive comparison of different MPPT techniques for photovoltaic

systems

Article  in  Solar Energy · February 2015

DOI: 10.1016/j.solener.2014.11.010

CITATIONS READS

369 7,883

2 authors:

Ali M. Eltamaly Hegazy Rezk

Mansoura University Minia University (Egypt)& Prince Sattam Univeristy (KSA)
231 PUBLICATIONS   4,136 CITATIONS    261 PUBLICATIONS   5,760 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

HVDC Systems View project

DFIG fo wind energy system applications View project

All content following this page was uploaded by Ali M. Eltamaly on 21 December 2017.

The user has requested enhancement of the downloaded file.

 Available online at www.sciencedirect.com

ScienceDirect
Solar Energy 112 (2015) 1–11
www.elsevier.com/locate/solener

A comprehensive comparison of different MPPT techniques

for photovoltaic systems
Hegazy Rezk a,⇑, Ali M. Eltamaly b,c
a
Electrical Engineering Dept., Faculty of Engineering, Minia University, Minia, Egypt
b
Sustainable Energy Technologies Center, King Saud University, Riyadh, Saudi Arabia
c
Electric Engineering Department, Al-Mansoura University, Egypt

Received 1 April 2014; received in revised form 1 November 2014; accepted 8 November 2014

Communicated by: Associate Editor Igor Tyukhov

Abstract

This paper aimed to study the behavior of different maximum power point tracking (MPPT) techniques applied to PV systems. In this
work, techniques such as hill climbing (HC), incremental conductance (INC), perturb-and-observe (P&O), and fuzzy logic controller
(FLC) are assessed. A model of PV module and DC/DC boost converter with the different techniques of MPPTs was simulated using
PSIM and Simulink software. Co-simulation between PSIM and Simulink software packages is used to establish FLC MPPT technique.
The co-simulation is done to take advantage of each program to handle certain part of the system. The response of the different MPPT
techniques is evaluated in rapidly changing weather conditions. The results indicate that, FLC performed best among compared MPPT
techniques followed by P&O, INC, and, HC MPPT techniques in both dynamic response and steady-state in most of the normal oper-
ating range.
Ó 2014 Elsevier Ltd. All rights reserved.

Keywords: Maximum power point trackers; Incremental conductance; Perturb-and-observe; Hill climbing; Fuzzy logic; MPPT

1. Introduction efficiency of electric power generation is low, and the

amount of electric power generated by solar arrays changes
Renewable energy sources such as photovoltaic (PV) continuously with weather conditions (Berrera et al., 2009;
power play a crucial role in electric power generation, Sreekanth and Raglend, 2012). Moreover, because of non-
and become essential these days due to shortage and envi- linear I–V and P–V characteristics of PV systems, their out-
ronmental impacts of conventional fuels. In the future, put power is always changing with weather conditions, i.e.,
PV energy will gain more importance due to the shortage solar radiation, atmospheric temperature and also nature
of fossil fuels and their environmental effects. More than of load connected (Fangrui et al., 2008; Emad and
45% of necessary energy in the world will be generated by Masahito, 2010). Maximum power point tracking (MPPT)
PV arrays (Faranda et al., 2008). Unfortunately, PV gener- is essential as there is a probable mismatch between the
ation systems have two major problems: the conversion load characteristics and the maximum power points
(MPPs) of the PV module in order to ensure optimal utili-
⇑ Corresponding author. zation of solar cells (Xiao and Dunford, 2004; Ashish et al.,
E-mail addresses: [email protected] (H. Rezk), eltamaly@ 2007). Using of MPPT leads to reduce the cost of energy
ksu.edu.sa (A.M. Eltamaly). generated by PV panels (Hohm and Ropp, 2000). In recent

http://dx.doi.org/10.1016/j.solener.2014.11.010
0038-092X/Ó 2014 Elsevier Ltd. All rights reserved.
2 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11

years, a large number of techniques have been proposed for carried out using MPPT based on different techniques;
tracking the MPP of PV systems. There are many tech- HC, INC, P&O and, FLC. The system submodels are
niques available in the literature such as fractional open- explained in the following sections:
circuit voltage and short-circuit current (Fangrui et al.,
2008), the Artificial Neural Network (ANN) technique
2.1. PV solar module
(El Sayed, 2013), and the fuzzy logic control (Eltamaly,
2010; Eltamaly et al., 2010). Also, it was demonstrated in
The PV module used in this study consists of 36 poly-
Ibrahim and Houssiny (1999) technique to track the maxi-
crystalline silicon solar cells electrically configured as two
mum power using look-up table in the microcomputer.
series strings of 18 cells each. Its main electrical specifica-
Another common approach is to use the array power as
tions are shown in Table 1. The equivalent circuit model
the feedback. The popular tracking methods based on this
for a PV module is addressed in Khaehintung et al.
approach are widely adopted in PV power systems
(2006), Abouobaida and Cherkaoui (2012), Emad and
(Faranda et al., 2008) which include but not limited to, per-
Masahito (2011), Christy et al. (2014), Gokmen et al.
turb and observe method (P&O) (Koutroulis et al., 2001;
(2013).
Ioan and Marcel, 2013), the incremental conductance
method (INC) (Xiao and Dunford, 2004; Esram and
Chapman, 2007) and the hill climbing method (HC). These 2.2. DC/DC boost converter
techniques are widely applied in the MPPT controllers due
to their simplicity and easy implementation. Several differ- According to maximum power transfer theory, maxi-
ent MPPT methods have been proposed, but there has been mum power is being transferred from source to load when
no comprehensive comparison between different techniques source impedance is equal to the load impedance (load
and their tracking efficiencies under varying weather condi- matching). The load matching can be done by adjusting
tions. The objective of this work was to bridge this gap. In the duty cycle of the DC/DC converter. The duty cycle is
this work, the attention will be focused on simulation com- the ratio between the switching on time of switch to the
parison study between these widely applied MPPT tech- switching period. In order to track MPP the converter must
niques, considering solar radiation variation in order to be operated with duty cycle corresponding to it. With vary-
understand which technique has the best performance in ing atmospheric conditions the duty cycle of the DC/DC
fast changing weather conditions. converter has to be adjusted to extract maximum power
This paper is organized as follows; section II descries the from PV module (RezaReisi et al., 2013). There are several
PV system modeling, illustrates basic operation principles architectures of DC/DC conversion circuits which can be
for HC, INC, P&O and FLC techniques respectively. Sim- used for this purpose. In the present work the boost config-
ulation results, analysis and discussion are illustrated in uration is chosen due to its wide spread use and high reli-
Section 3. Finally, conclusions are given in Section 4. ability with respect to other more complex configurations
(Berrera et al., 2009). The complete power device scheme
is shown in Fig. 1. The diode D1 is provided to protect
2. PV system modeling
the PV module against negative current which could dam-
age it. C1 placed at boost input to limit the high frequency
The solar PV generation system consists of a PV mod-
harmonic components (Berrera et al., 2009).
ule, DC/DC boost converter and a battery as shown in
Fig. 1. Radiation (R) is incident on the PV module. It gen-
erates a voltage (V) and current (I). The temperature of the 2.3. MPPT techniques
module is measured at T. The negative terminal of the bat-
tery and the module are connected and are also connected In this work, four MPPT techniques have been selected
to ground. The simulation of the PV system has been for the purpose of comparison; hill climbing (HC), incre-

Fig. 1. Block diagram of a general PV system.

 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11 3

Table 1 duty cycle should be increased in order to make dD > 0.

The electrical specifications of MSX-60 PV module. If dP < 0, the duty cycle is then reduced to make dD < 0.
Maximum power (Pmax) 60 W The main problem associated with this technique is because
Voltage @ Pmax (Vmp) 17.1 V of the tradeoff between the stability of the system in con-
Current @ Pmax (Imp) 3.5 A
Short-circuit current (Isc) 3.8 A
stant radiation period and lack of fast response in rapidly
Open-circuit voltage (Voc) 21.1 V changing radiation. The constant radiation period needs
very small value of change in duty cycle, DD to prevent
high oscillation of power around the maximum power
mental conductance (INC), perturb and observe (P&O) point which reduces the energy captured from PV. On
and fuzzy logic controller (FLC). the other hand, the rapidly changing radiation needs higher
value of duty cycle for fast tracking the maximum power.
(1) Hill climbing (HC) technique The flowchart of this technique is shown in Fig. 2. This
technique has been simulated in the PSIM software pack-
The advantage of the hill climbing MPPT technique is age as shown in Fig. 3.
its simplicity. It uses the duty cycle of boost converter as
the judging parameter when the task of the maximum (2) Incremental conductance (INC) technique
power point tracking is implemented. When the condition
dP/dD = 0 is accomplished, it represents that the maximum Among all the MPPT strategies, the incremental con-
power point has been tracked (Xiao and Dunford, 2004). ductance technique is widely used due to the high tracking
The flow diagram of the HC algorithm is shown in accuracy at steady state and good adaptability to the rap-
Fig. 2. The duty cycle in every sampling period is deter- idly changing atmospheric conditions (Fangrui et al.,
mined by the comparison of the power at present time 2008). This technique employs the slope of the PV array
and previous time. If the incremental power dP > 0, the power characteristics to track MPP. The slope of the PV
array power curve is zero at the MPP, positive for values
of output voltage smaller than the voltage at MPP, and
negative for values of the output voltage greater than the
start
voltage at MPP. The derivative of the PV module power
is given as in (1), and the resultant equation for the error
Sense V(i), I(i) e is as in (2) (Abouobaida and Cherkaoui, 2012; Emad
and Masahito, 2011; Chen et al., 2014).
P(i) = V(i) * I(i) dP dðV  IÞ dI
¼ ¼I þV ¼0 ð1Þ
dV dV dV
No
P(i)-P(i-1) >0
Yes Also, dI
dV
þ VI ¼ VIðiÞIði1Þ
ðiÞV ði1Þ
þ VIðiÞ
ðiÞ
¼0
IðiÞ  Iði  1Þ IðiÞ
e¼ þ ð2Þ
D(i)=D(i)-Δ D(i) =D(i)+Δ V ðiÞ  V ði  1Þ V ðiÞ

Therefore tracking the MPP requires the following pro-

cedure as shown in Fig. 4. It can be implemented by a sim-
ple discrete integrator with the error signal e as the input,
Return and a scaling factor k. The function of the scaling factor
k is to adapt the error signal e to a proper range before
Fig. 2. State flow chart of HC MPPT technique. the integral compensator. As the operating point

Fig. 3. Implementation of HC MPPT control technique using PSIM software.

4 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11

Start Start

Sense V, I
Sense V, I
P(i) = V(i) * I(i)

I (i ) − I (i − 1) I (i )
е= + Yes
V (i ) − V (i − 1) V (i ) P(i)-P(i-1) =0

No
No Yes
No Yes P(i)-P(i-1) >0
e>0
Yes No No Yes
V(i)-V(i-1) >0 V(i)-V(i-1) >0
=0

D(i)=D(i-1)-k*e D(i)=D(i-1) D(i)=D(i-1)+k*e

V(i)=V(i)-Δ V(i) =V(i)+Δ V(i) =V(i)-Δ V(i) =V(i)+Δ

return
return

Fig. 4. State-flow chart of INC MPPT technique.

Fig. 6. State flowchart of P&O MPPT technique.
approaches the MPP, the error signal e becomes smaller,
resulting in an adaptive and smooth tracking
(Abouobaida and Cherkaoui, 2012).
To improve, both, the MPPT speed and accuracy simul-
taneously a modified dynamic change in step size for INC
is introduced (Fangrui et al., 2008). This technique
improves the performance of INC technique but at a cost
of increased complexity of the control system.
The flow chart for INC technique is shown in Fig. 4
which was simulated in the PSIM software package as
shown in Fig. 5.

(3) Perturb and observe (P&O) technique

P&O is the most frequently used technique to track the Fig. 7. P&O MPPT control technique using PSIM software.
maximum power due to its simple structure (Ioan and
Marcel, 2013). This technique operates by periodically per-
turbing the PV module terminal voltage and comparing the this technique is that the PV module terminal voltage is
PV output power with that of the previous perturbation perturbed every MPPT cycle; therefore when the MPP is
cycle (Faranda et al., 2008). As shown in Fig. 6 if the PV reached, the output power oscillates around the maximum,
module operating voltage changes and power increases resulting in power loss in the PV system. To remedy this
the control system moves the operating point in that direc- problem, a modified P&O technique has been introduced
tion; otherwise the operating point is moved in the opposite in Al-Diab and Sourkounis (2010) by multiplying the
direction. The flowchart of this technique is shown in change in the duty cycle by dynamic constant depending
Fig. 6. PSIM software package has been used to simulate on the previous change in the extracted power as shown
this technique as shown in Fig. 7. A common problem in in (1). Another technique (Amrouche et al., 2007) used

Fig. 5. INC MPPT control technique using PSIM software.

 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11 5

ANN to predict this multiplying constant. These tech- Table 2

niques increase the complexity of the system and may cause Rules for a fuzzy system.
more oscillations in stable weather conditions. DE
E NB NM NS ZE PS PM PB
(4) Fuzzy logic controller (FLC) technique for PV MPPT NB NB NB NB NB NM NS ZE
NM NB NB NB NM NS ZE PS
FLC has been introduced in many researches as in NS NB NB NM NS ZE PS PM
Gounden et al. (2009), Ben Salah et al. (2008), Altasa ZE NB NM NS ZE PS PM PB
PS NM NS ZE PS PM PB PB
and Sharaf (2008), Khaehintung et al. (2004), Karlis PM NS ZE PS PM PB PB PB
et al. (2007), Veerachary et al. (2003), Rajesh and Mabel PB ZE PS PM PB PB PB PB
(2014) to force the PV to work around MPP. FLCs have
the advantages of working with imprecise inputs, not need-
ing an accurate mathematical model, and handling nonlin-
earity. The details of using FLC in MPPT of PV system are output, which is typically a change in duty cycle, DD of
shown in Eltamaly (2010). The error signal can be calcu- the power converter, can be looked up in a rule base given
lated as shown in (3). The value of DE is calculated as in Table 2. A triangular membership function can be used
shown in (4). The model of the proposed system has been for both inputs and output variables, as it can easily be
simulated in co-simulation between PSIM and Simulink implemented on the digital control system. The linguistic
software packages. The idea behind using co-simulation variables assigned to DD for the different combinations of
is the easy simulation of FLC in Simulink and power cir- E and DE are based on the power converter being used
cuit in PSIM. and also on the knowledge of the user.
These linguistic variables of input and output MFs are
P ðiÞ  pði  1Þ
EðnÞ ¼ ð3Þ then compared to a set of pre-designed values during aggre-
V ðiÞ  V ði  1Þ gation stage. The proper choice of If-then rules or fuzzy
DEðiÞ ¼ EðiÞ  Eði  1Þ ð4Þ inference is essential for the appropriate response of the
FLC system. The inference used in this work is tabulated
The detailed logic, theory and implementation of this in Table 2. Some researches proportionate these variables
model can be found in Eltamaly (2010) and Eltamaly to only five fuzzy subset functions as in (Eltamaly et al.,
et al. (2010). The output power from the PV system and 2010). Table 2 can be translated into 49 fuzzy rules or
the voltage are used to determine the E and DE based on IF-THEN rules to describe the knowledge of control as
(4) and (5). The inputs to a FLC are usually E and DE. follows:
The range of E and DE are fixed judiciously based on trial
and error. These variables are expressed in terms of linguis- R25 : If E is NM and DE is PS then DD is NS
tic variables or labels such as PB (Positive Big), PM (Posi- R63 : If E is PM and DE is NS then DD is PS
tive Medium), PS (Positive Small), ZE (Zero), NS ...
(Negative Small), NM (Negative Medium), NB (Negative
R51 : If E is PS and DE is NB then DD is NM
Big) using basic fuzzy subset. Each of these acronyms is
described by a mathematical membership functions, MF Defuzzification is for converting the fuzzy subset of con-
as shown in Fig. 8. Once E and DE calculated and con- trol form inference back to values. As the plant usually
verted to the linguistic variables based on MF, the FLC required a nonfuzzy value of control, a defuzzification
stage is needed. Defuzzification for this system is the height
method. The height method is both very simple and very
NB NM NS ZE PS PM PB
fast method. The height defuzzification method in a system
Error, E of rules formally given by (6):
MFs !,
Xm Xn
DD ¼ cðkÞ  W k WK ð5Þ
k¼1 k¼1
NB NM NS ZE PS PM PB
Change of where DD = change of duty cycle; c(k) = peak value of
Error, each output; Wk = height of rule k.
MFs
In the defuzzification stage, FLC output is converted
from a linguistic variable to a numerical variable. This pro-
NB NM NS ZE PS PM PB vides an analog signal which is DD of the boost converter.
Change in This value is subtracted from previous value of D to get its
Duty Ratio new value.
MFs
PSIM model showing the calculating E and DE and the
inputs to Simulink is shown in Fig. 9. The simulation
Fig. 8. A fuzzy system with two inputs, 1 output and 7 MFs each. model showing the output signal from FLC in Simulink
6 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11

Fig. 9. PSIM model showing the calculating E and DE and the inputs to
Simulink.
Fig. 12. Chang in the solar radiation (W/m2).

radiation profile used in simulation. A step change in radi-

ation is applied to study the response of each technique
during disturbance.
The MPPT techniques are compared in terms of
dynamic response and the percentage reduction in energy.
The percentage reduction in energy can be obtained from
the following equation:
Emax  E
%Ered ¼  100 ð6Þ
Emax
Fig. 10. Simulation model of output signal from FLC in Simulink to
PSIM and the switching circuit. where Emax: the maximum energy that can be generated
from PV; E: the generated energy generated from PV sys-
tem during simulation time.
to PSIM and the switching circuit that control the switch is To compare the performance of the different MPPT tech-
shown in Fig. 10. The simulation model of the proposed niques with change in duty cycle of boost converter, the sim-
system in Simulink is shown in Fig. 11, where SLINK1 is ulations were performed under exactly the same conditions.
the name that used in Simulink for the input to PSIM part The sampling time is the inverse of the sampling frequency.
of simulation. Sampling frequency defines the number of samples per sec-
ond. The sampling time for the MPPT is chosen to be
3. Analysis and discussion of the simulation results 0.01 sec. The duty cycle is therefore updated every
0.01 sec. The dynamic response of the PV module output
The whole system shown in Fig. 1 has been simulated power in case of using HC MPPT technique with different
using co-simulation of PSIM and Simulink software pack- increment in the duty cycle, DD = 2%, 4%, 6%, 8% and
ages. It consists of a 60 W solar PV module, a DC/DC 10% respectively. Figs. 13 and 14 shows the dynamic
boost converter operating at switching frequency of response of HC MPPT in starting and steady state respec-
30 kHz, an input inductance of 1 mH, an output capacitor tively. It is clear from Fig. 13 that the starting time is inver-
of 47 lF and a 24 V battery. Four MPPT techniques are sely proportional to the incremental value of duty cycle DD.
implemented in simulation as a controller to study and The maximum starting time to reach the maximum power
compare the dynamic response of the PV system under rap- for 800 W/m2 radiation occurs at the lower value of DD
idly radiation changes. All simulations are performed with (DD = 2%) which is about 0.16 sec, which does not have a
step changes in solar radiation. Fig. 12 indicates the solar
PV Output Power (W)

Error

SLINKI

Δ Error Fuzzy
Controller

PV-System

ΔD

Memory
Fig. 13. Starting period of HC–MPPT for DD = 2%, 4%, 6%, 8% and
Fig. 11. Simulink simulation model of FLC MPPT. 10%.
 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11 7

PV Output Power (W)

PV Output Power (W)

Fig. 14. Dynamic response of the PV module output power with HC–
MPPT for DD = 2%, 4%, 6%, 8% and 10%.
Fig. 16. Dynamic response of the PV module output power with INC–
MPPT for DD = 2%, 4%, 6%, 8% and 10%.
serious effect on the performance of the PV system. The
minimum starting time to reach the maximum power for fast change in radiation especially in tropical environment
800 W/m2 radiation occurs at the higher value of DD this MPPT will take longer time to follow the maximum
(DD = 10%) which is about 0.03 sec. This proves that the power point which will reduce the energy captured from
higher the DD the quicker will be tracking of the maximum the system. This proves that the higher the DD the quick
power. Fig. 13 shows that the oscillation of the generated reach of the maximum power. Fig. 16 shows that the oscil-
power is directly proportional to DD. It is clear from this lation of the generated power is directly proportional to the
technique that it has difficulty in providing good perfor- DD. It is clear from this technique that it has difficulty in
mance in both dynamic and steady-state response together providing good performance in both dynamic and steady-
because of the tradeoff between the steady state perfor- state response together because of the tradeoff between
mance and transient stability. As seen in Fig. 14, in steady the steady state performance and transient stability. It is
state (constant radiation) the better performance can be clear that in steady state (constant radiation) the better per-
achieved with lower values of DD but the fast response in formance can be achieved with lower values of DD but the
rapidly changing radiation can be achieved with higher val- fast response in rapidly changing radiation can be achieved
ues of DD. So, a smart compromising between the effects of with higher values of DD. Also, it is clear that INC tech-
DD on the operation of the HC–MPPT in steady state and nique has better performance than HC technique in both
transient stability should be considered. the transient and steady state response.
The dynamic response of the PV module output power The dynamic response of the PV module output power
in case of using INC MPPT technique with different incre- in case of using P&O MPPT technique with different incre-
ment in the duty cycle, DD = 2%, 4%, 6%, 8% and 10% is ment in the duty cycle, DD = 2%, 4%, 6%, 8% and 10% is
shown in Fig. 15 (starting/transient) and Fig. 16 (steady shown in Fig. 17 (starting/transient) and Fig. 18 (steady
state), respectively. It is clear from Fig. 15 that the starting state), respectively. The maximum starting time (about
time is inversely proportional to the incremental value of 0.16 sec) to reach the maximum power, at 800 W/m2 radi-
duty cycle DD. The maximum starting time to reach the ation, occurs at the lower value of DD (2%). This proves
maximum power for 800 W/m2 radiation occurs at the that the higher the DD the quick reach of the maximum
lower value of DD which is about 0.15 sec, which does power.
not have a serious effect on the performance of the PV sys-
tem especially on low change in radiation. But, in case of
PV Output Power (W)

Fig. 15. Starting period of INC–MPPT for DD = 2%, 4%, 6%, 8% and Fig. 17. Starting period of P&O–MPPT for DD = 2%, 4%, 6%, 8% and
10%. 10%.
8 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11

2% 4% 6% 8% 10% Pmax

54

PV Output Power (W)

PV Output Power (W)

52

50

48

46
1.79 1.8 1.81 1.82 1.83
Time (s)
Fig. 20. Dynamic response of the PV module output power with FLC–
Fig. 18. Dynamic response of the PV module output power with P&O– MPPT for DD = 2%, 4%, 6%, 8% and 10%.
MPPT for DD = 2%, 4%, 6%, 8% and 10%.
maximum power for 800 W/m2 radiation occurs at the
It is clear from results shown in Figs. 13–18 that the lower value of DD which is about 0.6 sec. Fig. 20 shows
starting current for HC, INC, and P&O is almost the same that the oscillation of the generated power is directly pro-
for all values of DD. Fig. 18 shows that the oscillation of portional to the DD but for a very short time after the tran-
the generated power for P&O is almost constant with all sient occurs, after that the oscillations with all values of DD
values of DD. It is clear from the simulation results of this are almost similar. It is clear from the simulation results of
technique that it can work with almost same oscillation for this technique that it has better performance than other
all values of DD. So, high value of DD can be chosen for techniques in both dynamic and steady-state response
fast response in rapidly changing radiation and this value together because of the dynamic change of the DD depend-
will not considerably affect the performance of steady state ing on the situation required.
(constant radiation). Hence, with P&O technique it is pos- Fig. 21 shows three traces showing the performance of
sible to achieve the satisfactory performance in both tran- FLC–MPPT in starting period for 2%, 4%, 6%, 8% and
sient and steady state which is its major benefit over HC 10% as a maximum value of DD. In the upper trace of this
and INC techniques. figure, the power generated with respect to the maximum
The simulation of the FLC for MPPT has been carried power. The middle trace of Fig. 21 shows the variation
out by co-simulation between PSIM and Simulink simula- of DD for each case. The lower trace shows the duty cycle
tion packages as shown in Figs. 9–11. The motivation for during starting period. It is clear from this figure that the
using the co-simulation is that the PSIM is more effective starting time is inversely proportional to the maximum
and simple for power electronics PV module development value of the duty cycle. Note that the value of DD becomes
and Simulink is better in handling fuzzy logic control. the same for all values of maximum allowable value of DD
The simulation of FLC technique with different increment after the starting period and any sudden change in the radi-
in the duty cycle, DD = 2%, 4%, 6%, 8% and 10% respec- ation which prove that the value of maximum allowable
tively was performed. Figs. 19 and 20 show the dynamic DD affects only in the transient periods.
response of FLC MPPT technique in starting and steady
state respectively. It is clear from Fig. 19 that the starting
time is inversely proportional to the incremental value of
Power (W)
PV Output

duty cycle DD. The maximum starting time to reach the

PV Output Power (W)

ΔD
Duty ratio, D

Fig. 19. Starting power of FLC–MPPT for DD = 2%, 4%, 6%, 8% and Fig. 21. The performance of FLC–MPPT in starting period for 2%, 4%,
10%. 6%, 8% and 10% as a maximum value of DD.
 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11 9

Fig. 22 shows three traces showing the performance of Pmax HC INC P&O FLC
60

PV Output Power (W)

FLC–MPPT in normal operating period for 2%, 4%, 6%,
50
8% and 10% as a maximum allowable value of DD. In
40
the upper trace of this figure, the power generated with
30
respect to the maximum power. In the middle trace of
20
Fig. 22 it shows the variation of DD for each case. The
10
lower trace shows the duty cycle in starting period. It is 0
clear from this figure that the starting time is inversely pro- HC INC P&O FLC
portional to the maximum value of the duty cycle. Also it is 0.4
notable that the value of DD becomes the same for all val-

Duty ratio, D
0.3
ues of maximum allowable value of DD after the normal
operating period and any sudden change in the radiation 0.2

which prove that the value of maximum allowable DD 0.1

affects only in the transient periods.
0
Fig. 23 shows the variation of the generated power of 0 0.05 0.1 0.15 0.2 0.25 0.3
PV in the upper trace and duty cycle, D in the lower trace Time (s)
during starting period for HC, INC, P&O, and FLC MPPT
Fig. 23. The variation of the generated power of PV and duty cycle, D
techniques. The best value for DD in each technique has respectively during starting period for all techniques under study.
been chosen for each technique as has been discussed
before. So, the change in duty cycle, DD is chosen to be
2% for HC, INC, and P&O and 10% for FLC technique. Pmax HC INC P&O FLC
It is clear from this figure that there is no much difference PV Output Power (W) 60
on starting time between FLC and the other MPPT tech- 58
niques under study, where all of them reach the maximum
56
power within 0.2 sec. approximately. The slowest technique
54
is FLC MPPT technique and the fastest technique is the
52
INC technique. The limited difference in starting time does
not affect the generated energy considerably. HC INC P&O FLC
Fig. 24 shows the variation of the generated power of
0.36
PV in the upper trace and duty cycle, D in the lower trace
Duty ratio, D

0.34
during rapid change of radiation period for HC, INC, 0.32
P&O, and FLC MPPT techniques. The best value for DD 0.3
0.28
in each technique has been chosen for each technique as
0.26
has been discussed before. So, the change in duty cycle, 0.24
0.2 0.4 0.6 0.8
DD is chosen to be 2% for HC, INC, and P&O and 10%
Time (s)
for FLC technique. It is clear from this figure that the fuzzy
controller technique has the best tracking for the maximum Fig. 24. The variation of the generated power of PV and duty cycle, D for
all techniques under study.

55
2% 4% 6% 8% 10%
power than the other techniques. Also, it can be observed
Power (W)
PV Output

50
that oscillation around maximum power is limited with
FLC technique followed by P&O, INC, and HC
45
techniques.
40
Fig. 25 shows the starting time required for each MPPT
2% 4% 6% 8% 10%
0.1 technique under study to reach the maximum power corre-
0.05 sponding to 1 kW/m2 solar radiation along with the maxi-
ΔD

0 mum allowable change in duty cycle DD. It is clear from

-0.05
this figure that the starting time is inversely proportional
-0.1
to the allowable change in duty cycle DD in all MPPT tech-
2% 4% 6% 8% 10%
niques under study. Also, it is clear from this figure that the
Duty ratio, D

0.4
0.35
starting time variations for HC, INC and P&O techniques
0.3 are almost the same and are shorter than the time associ-
0.25
0.2 ated with the FLC technique. So, HC, INC and P&O tech-
0.8 1 1.2 1.4
niques are better than FLC with respect to the starting
Time (s)
time.
Fig. 22. Dynamic response of the PV module of output power, DD, D Fig. 26 shows the percentage reduction of energy along
with FLC–MPPT for DD = 2%, 4%, 6%, 8% and 10%. with the maximum allowable change in the duty cycle, DD
10 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11

1.4 response of PV system. The simulation results show that,

FLC technique has the lower percentage reduction in the
1.2
generated energy compared to all the other techniques
under study in the entire range of duty cycle. It is
1
Starting time (Sec.)

FLC followed by the INC technique, which has the lower per-
0.8 centage reduction in the generated energy than the P&O
and HC for change in duty cycle in the range lower than
0.6 5%.
HC, INC, and P&O
On the transient performance front, FLC performed
0.4 worst, with highest settling time compared to the other
techniques under study. On the basis of transient perfor-
0.2
mance, the rank of the MPPT techniques under study is
0 arranged from best to worst as follows: FLC, P&O, INC,
0 0.05 0.1 0.15 0.2 and HC. Nevertheless, settling time has lower importance
Max. allowable change in duty ratio than the percentage reduction in the generated energy.
Fig. 25. The starting time required for each MPPT technique under study. To summarize, when overall performance compared, the
FLC and P&O are superior with respect to other MPPT
techniques as they can effectively improve the tracking
30 speed and minimize steady state error simultaneously.
HC
25 INC
P&O Acknowledgments
% Reduction in energy

INC
FLC
20
The authors acknowledge the College of Engineering
15 HC Research Center and Deanship of Scientific Research at
King Saud University in Riyadh, Saudi Arabia, for the
10
P&O FLC
financial support to carry out the research work reported
5 in this paper.

0 Appendix A
0 0.05 0.1 0.15 0.2
Max. alllowable change in duty ratio Description of PSIM blocks used in this paper
Fig. 26. The percentage reduction of energy along with the maximum Zero-order hold (the ZOH samples the input at
allowable change in the duty cycle, DD.
the beginning of a clock cycle, and holds the
sampled value until the next clock cycle)
for four MPPT techniques under study. These curves have Unit delay block (a unit delay block delays the
been drawn for the simulation of normal operating range input by one sampling period)
without rapid change in radiation to ideally compare the Comparator (the output of a comparator is high
generated energy. It is clear that INC technique has the (value = 1) when the non-inverting input is
lower percentage reduction in the generated energy com- higher than the inverting input. When the
pared to the P&O and HC in the range of maximum allow- positive input is lower, the output is zero. If the
able change in duty cycle for range D < 5%. However, the two inputs are equal, the output is undefined
FLC technique has the lower percentage reduction in the and it will keep the previous value)
generated energy compared to other MPPT under study NOT gate (the output is the inversion of the
in all the range of maximum allowable change in duty ratio input logic signal)
which makes it the best technique. But, FLC is the worst On-off switch controller (the on-off switch
starting time compared to the other techniques under controller interfaces between the control circuit
study. But, starting time has lower importance than the and the power circuit)
percentage reduction in the generated energy. Simple bidirectional switch (a simple switch
From the simulation results depicted in Figs. 25 and 26 conducts current in both directions. It is on
MPPT techniques under study can be ranked from best to when the gating is high, and is off when the
worst as: FLC, P&O, INC, and HC. gating is low)
Sign function block (the output of the sign
function block is the sign of the input. When the
4. Conclusions
input is positive, the output is 1. When the input
is 0, the output is 0. When the input is negative,
In this paper, four MPPT techniques have been com-
the output is 1)
pared on the basis of, both, transient and steady-state
 H. Rezk, A.M. Eltamaly / Solar Energy 112 (2015) 1–11 11

References Faranda, R., Leva, S., Maugeri, V., 2008. MPPT techniques for PV
systems: energetic and cost comparison. Power and Energy Society
Abouobaida, H., Cherkaoui, M., 2012. Comparative study of maximum General Meeting – Conversion and Delivery of Electrical Energy in the
power point trackers for fast changing environmental conditions. In: 21st Century, 20–24 July, pp. 1–6.
Multimedia Computing and Systems (ICMCS), International Confer- Gokmen, N., Karatepe, E., Ugranli, F., Silvestre, S., 2013. Voltage band
ence, 10–12 May 2012, pp. 1131–1136. based global MPPT controller for photovoltaic systems. Sol. Energy
Al-Diab, A., Sourkounis, C., 2010. Variable step size P&O MPPT 98, 322–334.
algorithm for PV systems. In: Optimization of Electrical and Gounden, N.A., Peter, Sabitha A., Nallandula, Himaja, Krithiga, S.,
Electronic Equipment (OPTIM), 12th IEEE International Conference, 2009. Fuzzy logic controller with MPPT using line-commutated
pp. 1097–1102. inverter for three-phase grid-connected photovoltaic systems. Renew.
Altasa, I.H., Sharaf, A.M., 2008. A novel maximum power fuzzy logic Energy J. 34, 909–915.
controller for photovoltaic solar energy systems. Renew. Energy 33, Hohm, D.P., Ropp, M.E., 2000. Comparative study of maximum power
388–399. point tracking algorithms using an experimental, programmable,
Amrouche, B., Belhamel, M., Guessoum, A., 2007. Artificial intelligence maximum power point tracking test bed. In: Photovoltaic Specialists
based P&O MPPT method for photovoltaic systems. Revue des Conference, Anchorage, AK, pp. 1699–1702.
Energies Renouvelables ICRESD-07 Tlemcen, 11–16. Ibrahim, H.E., Houssiny, F.F., 1999. Microcomputer controlled buck
Ashish, P., Nivedita, D., Ashok, K.M., 2007. Comparison of various regulator for maximum power point tracker for DC pumping system
MPPT algorithms for cost reduction. IEEE Trans. Power Electron. 22 operates from photovoltaic system. In: IEEE International Fuzzy
(2), 698–700. Systems Conference, August 22–25, vol. 1, pp. 406–411.
Ben Salah, C., Chaabenea, M., Ben Ammara, M., 2008. Multi-criteria Ioan, V.B., Marcel, I., 2013. Comparative analysis of the perturb-and-
fuzzy algorithm for energy management of a domestic photovoltaic observe and incremental conductance MPPT methods. In: 8th Inter-
panel. Renew. Energy 33, 993–1001. national Symposium on Advanced Topics in Electrical Engineering,
Berrera, M., Dolara, A., Leva, S., 2009. Experimental test of seven widely May 23–25, Bucharest, Romania.
adopted MPPT algorithms. In: IEEE Bucharest Power Tech Confer- Karlis, A.D., Kottas, T.L., Boutalisb, Y.S., 2007. A novel maximum
ence, June 28th – July 2nd, Bucharest, Romania. power point tracking method for PV systems using fuzzy cognitive
Chen, Y.-T., Lai, Z.-H., Liang, R.-H., 2014. A novel auto-scaling variable networks (FCN). Electr. Power Syst. Res. 77 (3–4), 315–327.
step-size MPPT method for a PV system. Sol. Energy 102, 247–256. Khaehintung, N., Wiangtong, T., Sirisuk, P., 2006. FPGA implementa-
Christy, J.S., Raj, M., Jeyakumar, A.E., 2014. A two stage successive tion of MPPT using variable step-size P&O algorithm for PV
estimation based maximum power point tracking technique for applications. In: International Symposium on Digital Object Identifier,
photovoltaic modules. Sol. Energy 103, 43–61. ISCIT, pp. 212–215.
Eltamaly, A.M., 2010. Modeling of fuzzy logic controller for photovoltaic Khaehintung, N., Pramotung, K., Tuvirat, B., Sirisuk, P., 2004. RISC
maximum power point tracker. In: Solar Future 2010 Conf. Proc., microcontroller built-in fuzzy logic controller of maximum power
Istanbul, Turkey, February, pp. 4–9. point tracking for solar-powered light-flasher applications. In: 30th
Eltamaly, A.M., Alolah, A.I., Abdulghany, M.Y., 2010. Digital imple- Annu. Conf. IEEE Ind. Electron. Soc., pp. 2673–2678.
mentation of general purpose fuzzy logic controller for photovoltaic Koutroulis, E., Kalaitzakis, K., Voulgaris, N.C., 2001. Development of a
maximum power point tracker. In: Power Electronics Electrical Drives microcontroller-based photovoltaic maximum power point tracking
Automation and Motion (SPEEDAM), 2010 International Sympo- control system. IEEE Trans. Power Electron. 16, 46–54.
sium on Digital Object Identifier, pp. 622–627. Rajesh, R., Mabel, M.C., 2014. Efficiency analysis of a multi-fuzzy logic
Emad, M.A., Masahito, S., 2010. Modified adaptive variable step-size controller for the determination of operating points in a PV system.
MPPT based-on single current sensor. In: TENCON 2010 – IEEE Sol. Energy 99, 77–87.
Region 10 Conference, 21–24 November, pp. 1235–1240. RezaReisi, A., Moradi, M.H., Jamasb, S., 2013. Classification and
Emad, M.A., Masahito, S., 2011. Stability study of variable step size comparison of maximum power point tracking techniques for photo-
incremental conductance/impedance MPPT for PV systems. In: Power voltaic system: a review. Renew. Sustain. Energy Rev. 19, 433–443.
Electronics and ECCE Asia (ICPE & ECCE), 2011 IEEE 8th Sreekanth, S., Raglend, I.J., 2012. A comparitive and analytical study of
International Conference, May 30 2011 – June 3, pp. 386–392. various incremental algorithms applied in solar cell. In: International
El Sayed, M., 2013. Modeling and simulation of smart maximum power Conference on Computing, Electronics and Electrical Technologies
point tracker for photovoltaic system. Minia J. Eng. Technol. (MJET) [ICCEET], 21–22 March, pp. 452–456.
32 (1). Veerachary, M., Senjyu, T., Uezato, K., 2003. Neural-network-based
Esram, T., Chapman, P.L., 2007. Comparison of photovoltaic array maximum-power-point tracking of coupled-inductor interleaved-boost
maximum power point tracking techniques. IEEE Trans. Energy converter-supplied PV system using fuzzy controller. IEEE Trans. Ind.
Convers. 22 (2), 439–449. Electron. 50 (4), 749–758.
Fangrui, L., Shanxu, D., Fei, L., Bangyin, L., Yong, K., 2008. A variable Xiao, W., Dunford, W.G., 2004. A modified adaptive hill climbing MPPT
step size INC MPPT method for PV systems. IEEE Trans. Ind. method for photovoltaic power systems. In: 35th Annual IEEE Power
Electron. 55 (7), 2622–2628. Electronics Specialist Conference, vol. 3, pp. 1957–1963.

View publication stats