energies

Article
Fuzzy Logic-Based Perturb and Observe Algorithm
with Variable Step of a Reference Voltage for Solar
Permanent Magnet Synchronous Motor Drive System
Fed by Direct-Connected Photovoltaic Array
Mohamed Redha Rezoug 1, *, Rachid Chenni 2 ID
and Djamel Taibi 1
1 Departement of Electrical Engineering, Kasdi Merbah University Ouargla, Ouargla 30000, Algeria;
taibi.djamel@yahoo.fr
2 MoDERNa Laboratory Mentouri, University of Constantine1, Constantine 25000, Algeria;
rachid.chenni@gmx.fr
* Correspondence: redha77.rmr@gmail.com; Tel.: +213-077-271-0414

Received: 17 January 2018; Accepted: 14 February 2018; Published: 22 February 2018

Abstract: Photovoltaic pumping is considered to be the most used application amongst other
photovoltaic energy applications in isolated sites. This technology is developing with a slow
progression to allow the photovoltaic system to operate at its maximum power. This work introduces
the modified algorithm which is a perturb and observe (P&O) type to overcome the limitations of
the conventional P&O algorithm and increase its global performance in abrupt weather condition
changes. The most significant conventional P&O algorithm restriction is the difficulty faced when
choosing the variable step of the reference voltage value, a good compromise between the swift
dynamic response and the stability in the steady state. To adjust the step reference voltage according
to the location of the operating point of the maximum power point (MPP), a fuzzy logic controller
(FLC) block adapted to the P&O algorithm is used. This allows the improvement of the tracking pace
and the steady state oscillation elimination. The suggested method was evaluated by simulation using
MATLAB/SimPowerSystems blocks and compared to the classical P&O under different irradiation
levels. The results obtained show the effectiveness of the technique proposed and its capacity for the
practical and efficient tracking of maximum power.

Keywords: photovoltaic pumping; MPPT; fuzzy logic control; modified P&O algorithm; variable step

1. Introduction
The electrical energy use is an important demand of our daily life that keeps on growing. Thus,
the increase in the electrical energy production is linked to the value of the consumption and the
increase in consumption cannot be completely stopped. To remedy that, we boost the systems’
efficiency. In addition, the solar energy production via photovoltaic (PV) cells is a very interesting
renewable source that should be used to meet the demand for electrical energy.
The abundance and risk-free nature of photovoltaic energy has increased its widespread use until
it arrived at an annual growth rate of 35–40% [1].
Recently, photovoltaic energy has caught a lot of interest because its benefits include relatively
modest and low complexity structural requirements that are harmless to our planet and renewable
compared to standard energy sources. The use of photovoltaic systems allows us to improve the living
conditions in remote areas while preserving a clean environment. Last year, photovoltaic systems
were installed all around the world. As an example, in 1993, around 10,000 photovoltaic systems
were installed for water pumping, and after five years, the photovoltaic systems increased in number
six-fold to almost 60,000 systems [2].

Energies 2018, 11, 462; doi:10.3390/en11020462 www.mdpi.com/journal/energies

Energies 2018, 11, 462 2 of 15

The photovoltaic system is constituted of elements that are interconnected and created to achieve
the specific purpose and to provide the required electricity by the use of a device related to the
load conditions.
The efficiency range may decrease during the fluctuation in: solar radiation, panel temperature
conditions, and load. For this, many scientists have examined different maximum power point tracking
(MPPT) techniques that were discussed by designers to study control methods and power converters
for the tracking of the MPP of a photovoltaic module and extract their disadvantages and advantages
so as to understand better their influence on the system’s performance.
Lately, for the automatic identification and exploitation at a MPPT, various methods and
techniques have been published that differ in many degrees (complexity, yield, cost, material,
popularity, speed of convergence, etc.).
A comparison between several aspects of the principal MPPT techniques (CV, P&O, INC, SC,
and OV) was discussed in [3–5]. Among the drawbacks of the CV technique (constant voltage) [3]
is the negligence of its purpose, which is the tracking of the MPP that was visible when it set the
reference voltage to a better fixed voltage and kept it constant under any operating condition, but this
technique is also advantageous compared to the P&O (perturb and observe) and INC techniques
(incremental conductance), since it overcomes them in efficiency when it is a photovoltaic network of
a feeble irradiation and a low implementation cost.
The fact that the OV (fractional open-circuit voltage) and SC (fractional short-circuit current)
techniques use an additional static switch to perform the open-circuit and short-circuit conditions
makes them more moderate and efficient in energy gains compared to the CV technique. The triviality
of the static switches made the authors in [3,5] prefer the P&O and INC techniques for their lower
relative cost. In reference [5], the author shows that the P&O techniques implantation cost is more
economical than the INC technique’s despite the similarity of their performances.
In [6–8] the authors deduced that the P&O method is the most commonly used technique in
practice, due to its ease of implementation of a low-cost device. The P&O method has a high MPP
tracking performance compared to other techniques. However, it loses this characteristic during
quick changes of irradiation and temperature. When weather conditions change slowly or constantly,
the algorithm also oscillates around or near the MPP, wasting an amount of the available energy [9].
In this work, a drive system of variable speed consists of a set of interacting elements such as:
the photovoltaic panels, the voltage inverter, the permanent magnet synchronous motor (PMSM),
and the centrifugal pump. The motor, the space vector pulse width modulation (SVPWM) control,
the PV model, and the P&O algorithm will be explained in detail. To improve the drive system in both
dynamic response and steady-state, a variable step of reference voltage with P&O algorithm can be
implemented using a fuzzy logic controller (FLC).
The collaboration of the system’s various elements (motor, P&O algorithm, FLC, and drive system)
will be analyzed in terms of performance under different operating conditions such as the irradiation
and variation. The results are acquired using MATLAB/Sim-Power-System blocks.

2. Modeling of the Photovoltaic Cells

The modeling of photovoltaic cells resembles the modelling of a current generator mounted in
parallel with a diode, as shown in Figure 1. The cell behaves like a diode when there isn’t any light to
generate the current. As the incident light intensity increases, the cell generates a current.
 Energies 2018, 11, 462 3 of 15
Energies 2018, 11, x FOR PEER REVIEW 3 of 15

Figure 1. The electrical modeling of the photovoltaic (PV) cell.

Figure 1. The electrical modeling of the photovoltaic (PV) cell.

The characteristic equation can be deduced directly by using the Kirchhoff law: [10,11]
The characteristic equation can be deduced directly by using the Kirchhoff law: [10,11]
I PV = I ph − I d − I SH (1)
IPV = I ph − Id − ISH (1)
The expansion of the equation connecting the current delivered by a PV module, which consists
The
of expansion
putting of in
NS cells theseries
equation
and connecting
the voltage the current
at its delivered
terminals by a PV
and several module,
variables (aswhich consistsin
summarized of
putting
Table NS cellsgives
1 which in series and the voltage
the simplified circuitat model),
its terminals and severalbelow
is represented variables
and (as
in summarized
the associatedin
Table 1 which
equations: gives the simplified circuit model), is represented below and in the associated equations:

 ( RS(·RISPV ) )    R R. IS · IPV
     
S PV + V PV 
+V
. I PV +PV
V PV + VPV
IPV =I I ph=−I IS −exp
PV ph I S exp Vt V  −  − 
− 1 −
Rsh 
(2)
(2)
  t    R sh 
( RS · IPV + VPV ) RS · IPV + VPV
     
Id = IS exp −1 and ISH = (3)
  (RSV.It PV + VPV )    RS .I PV +RVshPV 
I d = I S exp  − 1 and I SH =   (3)
  Vt    NSR·nsh·k· T 
I ph = Np · I ph,cell , IS = Np · IS,cell and Vt = (4)
q
N S .nand
.k .T R and R represent
where n is defined as the = N p .Iideality
I phdiode’s ph, cell
factor = N p .I S ,cellbetween
, I S (typically and V = 2),
1t and S SH (4)
the series and shunt resistors. Manufacturers often tend to supply the R value q in the datasheet of the
S
product
where n is[11]. If notasprovided,
defined the diode’stheideality
equation of R(typically
factor S can be extracted
between 1by differentiating
and the
2), and RS and Equation
RSH (1)
represent
then
the rearranging
series and shunt it in the RS phrase.
resistors. Manufacturers often tend to supply the RS value in the datasheet of the
product [11]. If not provided, the equationof RS can be extracted by differentiating the Equation (1)
dVPV + RS ·dIPV

VPV + RS · IPV
then rearranging it in the RdI PV = 0 − IS · q
S phrase. · e q( n·k· T ) (5)
n·k· T
 VPV + RS .I PV 
 dV + R .dI ·( T/q q) n.k .T 
dI PV  PVPV− S n·kPV
=R0S −=I S−.qdI .e  (5)
(6)
 PV n.kI .T·eq( VPV n+·Rk·ST· IPV )
dV
S
The solar cells’ efficiency is decreased by the power
dI ( q )
n.k . T dissipation through the internal resistors
− PV − resistances can be done as that of a parallel shunt
RS = exploitative
during operation. The modeling of these  V + R S . I PV 
dV q  PV too  (6)
resistor (RSH ) and a series resistor (RS ). If the RPVSH decreases  n.k .T much, the VOC decreases, while the RS
 
I S .ethe series values and shunt resistors, R and R ,
increase may drop the ISC . It is doable to approximate S SH
fromThethesolar
slopes of the I/V curve respectively to the V OC and the ISC . However,
cells’ efficiency is decreased by the power dissipation through the internal resistors the series resistance
RS of the
during PV module
operation. Theismodeling
determined of from
these the the I-V characteristics
1/slope ofresistances
exploitative can be doneatasopen thatcircuit voltage
of a parallel
V , and the shunt resistance R
OC resistor (RSH) and a series resistor
shunt SH is equal to the inverse of the slope of the I-V characteristics
(RS). If the RSH decreases too much, the VOC decreases, while at the
the RS increaseISC
short-circuit . After
may drop that,
the the
ISC. measurements of the resistances
It is doable to approximate the series to the ISC and
values andthe VOCresistors,
shunt will thenRbeS
taken and noted.
and RSH, from the slopes of the I/V curve respectively to the VOC and the ISC. However, the series
resistance RS of the PV module is determined from the 1/slope of the I-V characteristics at open
circuit voltage VOC, and the shunt resistance RSH is equal to the inverse of the slope of the I-V
characteristics at the short-circuit ISC. After that, the measurements of the resistances to the ISC and
the VOC will then be taken and noted.
Energies 2018, 11, 462 4 of 15

Table 1. PV Cell Parameters.

Hints PV Cell Parameters

IS Diode saturation current for Np
IS ,cell Cell diode saturation current
Iph Photocurrent for Np
Iph ,cell Cell photocurrent
Ipv The output current
Id The diode current
ISH The shunt resistance current
Vt Thermal voltage for Ns
n The ideality factor of diode (1.1007)
k The Boltzmann’s constant (1.38 × 10−23 J/K)
T The p–n junction’s temperature
q The electron’s charge (1.6 × 10−19 ◦ C)

Under Matlab/Simulink, the IPV equation was used to define the electrical characteristics of the
PV module (Appolo-solar-energy ASEC-205 G6M), as well as to simulate the characteristics of the
P-V and I-V for various levels of temperature and irradiation, which are shown in Table 2 by respecting
the standard STC’s three factors (irradiation, air mass, cell temperature).

Table 2. Manufacturer Specifications for a PV Module.

Parameter Value
Cell Type Polycrystalline
Rated Power 205 ± 3.0% (W)
Voltage at Maximum Power Vmp 25.87 (V)
Open Circuit Voltage Voc 33.58 (V)
Short-Circuit Current Isc 8.61 (A)
Diode saturation current Id 2.3682 × 10−9 (A)
Current at Maximum Power Imp 7.93 (A)
Temperature Coefficient of Voc −0.42 (%/◦ C)
Temperature Coefficient of Isc 0.19 (%/◦ C)
Cells number per module 6×9
Series resistance RS 0.43947 (Ω)
Shunt Resistance RSH 116.9171 (Ω)

Figure 2 shows the variation of the I-V and P-V output’s characteristics of the simulated PV module
as a function of the irradiation and temperature variation. These agree with the characteristics provided
by the module manufacturer.
Energies 2018,
Energies 11,11,
2018, 462x FOR PEER REVIEW 5 of5 15
of 15

Array type: Apollo Solar Energy ASEC-205G6M; Array type: Apollo Solar Energy ASEC-205G6M;
1 series modules; 3 parallel strings 1 series modules; 3 parallel strings
30 30
1 kW/m2
25 25
0.8 kW/m2
20 20
Current (A)

Current (A)
0.6 kW/m2
15 0.5 kW/m2 15
70 oC o o25 oC
5040C C
10 10

5 5

0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Voltage (V) Voltage (V)

700
600 1 kW/m2
600 25 oC
500 0.8 kW/m2 500
Power (W)

Power (W)
400 o
0.6 kW/m2 400 o 40 C
70 oC50 C
300 0.5 kW/m2
300
200 200
100 100

0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Voltage (V) Voltage (V)

(a) (b)
Figure
Figure 2. (a)
2. (a) P–VP–V and
and I–VI–V curves
curves forfor different
different irradiance
irradiance andand constant
constant T (b)
T; (b) P–V
P–V and
and I–VI–V curves
curves forfor
different temperatures and constant
different temperatures and constant G. G.

3. Fuzzy Logic-Based P&O Algorithm with Variable Step of a Reference Voltage

3. Fuzzy Logic-Based P&O Algorithm with Variable Step of a Reference Voltage
The P&O algorithm’s basic structure and easy implementation caused it to be considered as the
The P&O algorithm’s basic structure and easy implementation caused it to be considered as
most commonly used MPPT algorithm for all techniques. It is built on the notion that the
the most commonly used MPPT algorithm for all techniques. It is built on the notion that the
power–voltage curve dP/dV reaches zero at the curve’s top.
power–voltage curve dP/dV reaches zero at the curve’s top.
Decrementing or incrementing periodically the photovoltaic generator’s voltage values or the
Decrementing or incrementing periodically the photovoltaic generator’s voltage values or the
current output, the corresponding output power comparison of the photovoltaic generator P (n + 1),
current output, the corresponding output power comparison of the photovoltaic generator P (n + 1),
and the previous disturbance P (n) is the P&O’s operating principle.
and the previous disturbance P (n) is the P&O’s operating principle.
The disturbance has to be maintained in the same initial direction if the disturbance of the
The disturbance has to be maintained in the same initial direction if the disturbance of the
terminal voltage causes an increase in the power (dP/dV > 0), otherwise it will be moved in the
terminal voltage causes an increase in the power (dP/dV > 0), otherwise it will be moved in the
converse direction. This process is redone until the maximum power reaches the point dP/dV = 0. We
converse direction. This process is redone until the maximum power reaches the point dP/dV = 0.
prefer the P&O algorithm for its advantages, but let us not forget its disadvantage, which is the
We prefer the P&O algorithm for its advantages, but let us not forget its disadvantage, which is the
unstoppable oscillation around the power even when it reaches the MPP, which causes extra losses
unstoppable oscillation around the power even when it reaches the MPP, which causes extra losses
of power. This can be remedied by implementing either a direct duty cycle control where the power
of power. This can be remedied by implementing either a direct duty cycle control where the power
is measured at each PWM cycle [12–14], or a reference voltage control where a reference voltage is
is measured at each PWM cycle [12–14], or a reference voltage control where a reference voltage is
manipulated as a disturbance parameter and a proportional integral (PI) controller is required to
manipulated as a disturbance parameter and a proportional integral (PI) controller is required to
accommodate the duty cycle [15–17].
accommodate the duty cycle [15–17].
To eliminate the limitation that exists in the conventional P&O algorithm’s implementations,
To eliminate the limitation that exists in the conventional P&O algorithm’s implementations,
we use a P&O algorithm based on the modified conventional algorithm, in which we implemented a
we use a P&O algorithm based on the modified conventional algorithm, in which we implemented
fuzzy logic controller block to provide a variable step of reference voltage. The suggested modified
a fuzzy logic controller block to provide a variable step of reference voltage. The suggested modified
P&O algorithm’s flowchart is shown in Figure 3.
P&O algorithm’s flowchart is shown in Figure 3.
Energies 2018, 11, 462 6 of 15
Energies 2018,
Energies 2018, 11,
11, xx FOR
FOR PEER
PEER REVIEW
REVIEW 66 of
of 15
15

Figure 3.
Figure 3. Flowchart
Flowchart of
of the
the modified
modified perturb
perturb and
and observe
observe (P&O)
(P&O) algorithm.
algorithm.
Figure 3. Flowchart of the modified perturb and observe (P&O) algorithm.

The modified
The modified power power values values ΔP ΔPPV as well
PV as well asas the
the modified
modified current
current values
values ΔI ΔIPV of the
PV of the photovoltaic
photovoltaic
The modified
generator
generator present the
present power
the inputs
inputsvaluesof ∆PPV
of the
the as well
fuzzy
fuzzy logic
logic ascontroller
the modified
controller block.
block.current values ∆IPV of
The adjustments
The adjustments of the
of the photovoltaic
the controller to
controller to
generator
the variable present
step the
ΔV inputs
of the of
P&O the fuzzy
depend logic
on controller
the current block.
solar The adjustments
temperature
the variable step ΔVref of the P&O depend on the current solar temperature and irradiation. To lower
ref and of the
irradiation. controller
To lowerto
the
the variable
the cost andstep
cost and ∆Vref of the
complexity
complexity of P&O
of the
the FLC
FLC depend
block’s
block’s on implementation,
the current solar temperature
implementation, it is
it is implemented
implementedand irradiation.
in aa way
in way so To that
so lower
that it
it
the cost
shares and
the complexity
same sensors of the
with FLC
the block’s
P&O thatimplementation,
measure the it is
voltages
shares the same sensors with the P&O that measure the voltages and currents of the PV generator.implemented
and currents in a
ofway
the so
PV that it shares
generator.
the same
The sensors
The reference
reference with the P&O
voltage
voltage is that
is deemed
deemedmeasure as aathedisturbance
as voltages and
disturbance and
andcurrents
control
control of the PV generator.
element
element of the
of the proposed
proposed
The
algorithm, reference
algorithm, so so the
the ΔVvoltage
ΔVref is deemed
is disturbed
ref is disturbed by as a disturbance
by subtracting
subtracting or and
or addingcontrol
adding the element
the variable
variable step of the
step in proposed
in accordance
accordance with algorithm,
with the
the
so the ∆V
variation of
ref is
thedisturbed
acquired byPV subtracting
output or
power. adding
This the
step variable
is step
calculated
variation of the acquired PV output power. This step is calculated by the adapted FLC block, so it has in
by accordance
the adapted with
FLC the variation
block, so it of
has
the
the acquired
the role of
role PV output
of adjusting
adjusting thepower.
the value of
value This
of thestep
the stepisin
step incalculated
accordance
accordance by with
the adapted
with FLC block,
the operating
the operating pointsoposition.
point it has the
position. The
Therole of
step
step
adjusting
value changes
value the value of
changes proportionally the step
proportionally with in accordance
with the with
the interval the operating
interval between
between the point
the operating position.
operating point The
point and step
and the value
the MPP. changes
MPP. ThisThis
proportionally
process continues with the
until interval
the MPP between
is attained, the operating
whose FLC point
block and
sets
process continues until the MPP is attained, whose FLC block sets the step value as zero. Therefore, the
the MPP.
step This
value process
as zero. continues
Therefore,
until
this the MPP
this ensures
ensures aa is attained,
fast
fast dynamic
dynamic whose FLC and
response
response block
and sets the step
eliminates
eliminates value as zero.
the oscillations
the oscillations Therefore,
around
around the MPP
the MPPthiswhen
ensures
when a fast
aa steady
steady
dynamic
state is
state response
is reached.
reached. and eliminates the oscillations around the MPP when a steady state is reached.

4.
4. Fuzzy
Fuzzy Logic
Logic Controller
Controller
4. Fuzzy Logic Controller
The
The fuzzy
fuzzy logic controller
logic controller proposed is
controller proposed
proposed is based
basedononan aninitial
initialknowledge
knowledgeofofthe thesystem.
system.It It
is is
usedusedto
The fuzzy logic is based on an initial knowledge of the system. It is used to
to modify
modify and and control
and control
control the the
the stepstep
step of of reference
of reference voltage.
reference voltage.
voltage. A A functional
A functional diagram
functional diagram
diagram of of a FLC
of aa FLC block
FLC block
block cancan
can bebe
be
modify
structured
structured by
by collaborating
collaborating between
between four
four sections:
sections: fuzzification,
fuzzification, inference,
inference, defuzzification,
defuzzification, and
and rule
rule
structured by collaborating between four sections: fuzzification, inference, defuzzification, and rule
base,
base,asasillustrated
illustratedinin
inFigure
Figure 4 where thethe
modified values of theof photovoltaic generator power ∆PPVΔP and
base, as illustrated Figure 44 where
where modified
the modified values
values of the photovoltaic
the photovoltaic generator
generator power
power ΔPPV PV
those of
and those the
those of current
of the ∆I
the current PV
current ΔI present
ΔIPV the
present the inputs
the inputs and
inputs and ∆V
and ΔVref presents
ΔVref the
presents the output
the outputthat
output that we
that we will
we willsend
will send
send toto the
to the
the
and PV present ref presents
P&O
P&O algorithm.
algorithm.
P&O algorithm.

Rule
Rule
∆VPV
∆V PV ∆VPV
∆V PV

× ∆Vref
∆V ref
Fuzzification
Fuzzification Inference
Inference Defuzzification
Defuzzification
∆IPV
∆I PV
∆IPV
∆I PV

Figure 4.
Figure
Figure 4. General
4. General diagram
General diagram of
diagram of aaa fuzzy
of fuzzy logic
fuzzy logic controller.
logic controller.
controller.

The input
The input variables,
variables, ΔP
ΔPPV
PV,, ΔI
ΔIPV
PV,, and
and output
output ΔV
ΔVref of the
ref of the fuzzy
fuzzy logic
logic controller
controller proposed
proposed can
can be
be
calculated by the use of these equations:
calculated by the use of these equations:
Energies 2018, 11, 462 7 of 15

The input variables, ∆PPV , ∆IPV , and output ∆Vref of the fuzzy logic controller proposed can be
calculated by the use of these equations:

∆V
Energies 2018, 11, x FOR PEER REVIEW PV
(k) = (VPV (k) − VPV (k − 1)) 7 of 15
(7)

∆IPV (kΔ)V=(k()I=PV(V(k)(k−
) − VIPV((k k− 1−)) 1)) (7)
(8)
PV PV PV

∆PPV (ΔkI) = ∆VPV (k) × ∆IPV (k) (9)

PV (k ) = (I PV (k ) − I PV (k − 1)) (8)
Using Equations (7) and (8), the power variation values ∆PPV and the output current variation
ΔPPV (k ) = ΔVPV (k ) × ΔI PV (k ) (9)
∆IPV of the PV generator can be stabilized before the fuzzification process to achieve the control
Using Equations (7) and (8), the power variation values ΔPPV and the output current variation
calculation’s simplification.
ΔIPV of the PV generator can be stabilized before the fuzzification process to achieve the control
In this model, the membership function of the output and input variables used as presented
calculation’s simplification.
in Figure 5, allInmembership
this model, the functions
membership(∆P PV , ∆I
function of PV ∆Vref ),and
the, output areinput
expressed
variables with
used asa presented
triangular in function.
Five linguistic
Figureterms are defined:
5, all membership NB as(ΔP
functions negative big,
PV, ΔIPV, ΔV ref),NS as negative
are expressed withsmall, ZZ as
a triangular zero, PS
function. Fiveas positive
small, andlinguistic terms arebig.
PB as positive defined: NB as negative big, NS as negative small, ZZ as zero, PS as positive
small, and PB as positive big.

(a)

(b)

(c)

Figure 5. (a,b) Membership functions of the input variables (∆PPV , ∆IPV ) and (c) the output variable ∆Vref .

All information of the monitored parameters are available in the fuzzy rules base that is a set
of verification laws (IF THEN). We possess two inputs and each one has five membership functions,
Energies 2018, 11, 462 8 of 15

which gives a 25 rule constitution in the FLC inference rules, as given in the Table 3. The function of
the latter is to define the variable step of reference voltage so as to convey it to the P&O algorithm to
track the PPM and stop iteration once that point is reached.

Table 3. Fuzzy rules base.

∆PPV
∆IPV
NB NS ZZ PS PB
NB NB NS NS ZZ ZZ
NS NS ZZ ZZ ZZ PS
ZZ ZZ ZZ ZZ PS PS
PS ZZ PS PS PS PB
PB PS PS PB PB PB

A fuzzy controller output is always a fuzzy set, so it must be filtered by a defuzzification process
to define the net output of the proposed fuzzy control.
Diverse defuzzification techniques were proposed by different authors. In which the defuzzification
method of the center of gravity (COG) or centroid [18–20] is the most common, where the shredder
determines the gravity center to use its value as an FLC output.

5. Description of Solar Permanent Magnet Synchronous Motor (PMSM) Drive System

5.1. Vector Control of PMSM

One single source composed of three parallel panels is enough to power the PMSM and increase
its efficiency. The Kirchhoff voltage law can be utilized so as to express the PMSM voltage equation,
as given in the equation:
dψd (t)
Vd (t) = Rid (t) + − ωe ψq (t) (10)
dt
dψq (t)
Vq (t) = Riq (t) + − ωe ψd (t) (11)
dt
where R is the winding resistance of the stator, ω e is the electrical rotor frequency, and, respectively,
ψd and ψq are the linkage fluxes of the two axis (d-axis and q-axis) when:

ψd = Ld id + ψm (12)

ψq = Lq iq . (13)

The ψm is the linkage of the flux due to the rotor magnets effects. The electromagnetic torque of
a PMSM in the axis d and q can be given by:

2  
Te = p ψd id − ψq iq . (14)
3
The sign p shows the number of pairs of poles. Based on the authors’ approach [21], the pole
placement method can design the PI controller’s parameters. In Figure 6, a global schematic
presentation of the PMSM controlled by a vector control powered by an inverter is shown. The desired
voltage Vref will be created according to the power generated from three PV panels to consume all the
power generated from the panel.
Energies 2018, 11, 462 9 of 15
Energies
Energies2018,
2018,11,
11,xxFOR
FORPEER
PEERREVIEW
REVIEW 99ofof15
15

Figure
Figure6.6.Global
Figure Globalschematic
schematic
6. Global ofofthe
schematic of permanent
the permanent
the magnet
magnet
permanent synchronous
synchronous
magnet motor
motor
synchronous (PMSM)
(PMSM)
motor controlled.
controlled.
(PMSM) controlled.

Effectively,
Effectively,after
aftercomparing
comparingthe themeasured
measuredvoltage
voltageVVPVPVand
andthe
thedesired
desiredVVrefref, ,the
thePI
PIcontroller
controllerisis
Effectively, after comparing the measured voltage VPV and the desired Vref , the PI controller
added
addedininthetheouter
outerloop
loopsosoas astotogenerate
generatethe
thedesired
desiredΩΩrefrefspeed.
speed.This
Thislatter
latterspeed speedsignal
signalisisalso
also
is added in the outer loop so as to generate the desired Ωref speed. This latter speed signal is also
compared
comparedwith withthe
themeasured
measuredone oneas asan
aninner
innerloop.
loop.This
Thiserror
errorsignal
signalwill
willbe
besent
senttotothe thePI
PIcontroller,
controller,
compared with the measured one as an inner loop. This error signal will be sent to the PI controller,
which
which generates
generates transversal
transversal stator
stator current
current asas aa reference
reference one.
one. Because
Because the
the magnets
magnets produce
produce aa
which generates transversal stator current as a reference one. Because the magnets produce a constant
constant
constant flux,
flux, the
the current
current IdId must
must be be controlled
controlled atat zero
zero current.
current. PMSM
PMSM parameters
parameters are are cited
cited inin
flux, the current Id must be controlled at zero current. PMSM parameters are cited in Appendix A.
Appendix
AppendixA. A.
5.2. Space Vector Pulse Width Modulation SVPWM
5.2.
5.2.Space
SpaceVector
VectorPulse
PulseWidthWidthModulation
ModulationSVPWM SVPWM
The SVPWM is used in modern drive of electric motors so as to obtain arbitrary waveforms.
It willThe
be SVPWM
The SVPWM
studied onisisaused
used ininmodern
three-phasemodern drive
driveof
inverter. ofelectric
The electric
conceptmotors
of theso
motors soas
SVPWM astotoobtain
obtain arbitrary
consistsarbitrary waveforms.
waveforms.
of reconstructing ItIt
the
will
willbebe studied
studied on
on a a three-phase
three-phase inverter.
inverter. The
The concept
concept of
ofthe
the SVPWM
SVPWM
voltage vector Vmoy from eight voltage vectors. Each one of these vectors corresponds to an association consists
consists of
of reconstructing
reconstructing the
the
voltage
voltage
of vector
the statesvector VVmoy
of the moy from
from eight
three-phase eight
voltagevoltage
voltage vectors.
vectors.
inverter Each one
one ofof these
Each[20].
switches these vectors
vectors corresponds
corresponds toto an an
association
association ofofthe
A reference states
statesofof
thevoltage the
thethree-phase
vector three-phase
is calculated voltage
voltage inverter
globallyinverter switches
switches[20].
and approximated [20]. over a modulation period
A A reference
reference voltage
voltage vector
vector isis calculated
calculated globally
globally and
and
T by an average voltage vector Vmoy . The latter is elaborated by the application approximated
approximated over
overaof
amodulation
the adjacentperiod
modulation period
voltageTT
by an average
by an average voltage
and thevoltage vector V
vectorVVmoy
moy . The latter is elaborated by the application of the
. The latter is elaborated by the application of the adjacent voltage adjacent voltage
vectors null vectors 0 and V 7 . A combinatorial analysis of all the possible states of the
vectors
vectors and
and the
the null
null vectors
vectors
switches makes it doable to calculate the VV0 0andand VV7.7voltage
.AAcombinatorial
vector (VSαanalysis
combinatorial , VSβ ). ofofall
analysis allthe
thepossible
possiblestates
statesofofthe
the
switches
switches makes
makes it
itdoable
doable to
to calculate
calculate the
the voltage
voltage vector
vector (V
(V ,
SαSα V
, V ).
SβSβ).
The eight voltage vectors redefined by the association of the switches are represented in the plane
(α, β)The eight
eightvoltage
TheFigure
by voltage
7, [22–25]. vectors
vectorsredefined
redefinedby bythe
theassociation
associationofofthe theswitches
switchesare arerepresented
representedininthe
the
plane
plane(α,(α,β)β)by
byFigure
Figure7,7,[22–25].
[22–25].

Figure 7.7.Representation
Figure7.
Figure Representation
Representationof of
inverter status
ofinverter
inverter and reference
status
status and voltagevoltage
andreference
reference vectors in
voltage the stationary
vectors
vectors ininthe reference.
thestationary
stationary
reference.
reference.

5.3.
5.3.Modeling
Modelingofofthe
theCentrifugal
CentrifugalPump
Pump
Energies 2018, 11, 462 10 of 15

5.3. Modeling of the Centrifugal Pump

A pump
Energies has
2018, 11, a maximum
x FOR suction (aspiration) capacity, which is the value of the vacuum
PEER REVIEW it can
10 of 15
produce. This characteristic differs depending on the type and the pump’s technical design.
A pump has pump’s
The centrifugal a maximumwork suction (aspiration)
involves capacity, which
three parameters: is the
the valuethe
height, of the vacuum
flow, it can
and the speed.
produce. This characteristic differs depending on the type and the pump’s technical
To find the solution of this system of equation, we consider one of the three constant parameters; design.
in generalTheit iscentrifugal
the speedpump’s
which work involvesWe
is constant. three
haveparameters:
tested ourthesystem
height,using
the flow, and where
a load the speed.
the To
torque
find the solution of this system of equation, we consider one of the three constant parameters; in
increases with the speed squared, such as a pump centrifugal, fan, etc.
general it is the speed which is constant. We have tested our system using a load where the torque
increases with the speed squared, such as a pump centrifugal,
2 fan, etc.
Tr = k × Ω (15)
Tr = k × Ω2 (15)
where k is the constant.
where k is the constant.
6. Simulation Results
6. Simulation Results
The simulation has been done using MATLAB/Sim-Power-Systems blocks as clarified in Figure 8.
The simulation has been done using MATLAB/Sim-Power-Systems blocks as clarified in Figure
The modified P&O block requires
8. The modified P&O block the V
requires PVVand
the I IPVof
PV andPV the PV
ofthe arraytotoestimate
PV array estimatethethe reference
reference voltage
voltage
shown in Figure
shown 8. 8.
in Figure TheTheVVrefref is
is compared
compared with withVPVVPV
andand regulated
regulated by a by a PI controller
PI controller to definetothe
define
the reference
reference speed ofthe
speed of thepump
pump motor.
motor. The sampling
The time time sampling step
step of the of theofmodels
models of the (control
the two blocks two blocks
blocks
(control and and
blocks power blocks)
power is taken
blocks) is as 5.14 as
taken μs.5.14 µs.

Figure
Figure 8. The
8. The simulationblock
simulation blockdiagram
diagram of
ofthe
thesolar
solardrive
drivesystems.
systems.

In the first part, a comparative study on the conventional P&O and the P&O based on fuzzy
In theisfirst
logic donepart, a comparative
to show the performancestudyofon thetechniques
both conventionalunder P&O andconditions.
similar the P&O based on fuzzy logic
The comparative
is donestudy takes into consideration two important characteristics: the maximum power point’s trackingstudy
to show the performance of both techniques under similar conditions. The comparative
takes speed
into consideration two important
and the oscillation in the steady characteristics:
state. To display the
themaximum
practice ofpower point’sFLC-based
the suggested tracking speed
P&O and
algorithm,inthis
the oscillation thealgorithm
steady state.was To
tested withthe
display twopractice
referenceof voltage steps: 0.01
the suggested and 0.05P&O
FLC-based under two
algorithm,
irradiationwas
this algorithm changes.
testedThe
with irradiation was changed
two reference voltagefrom 1000
steps: 0.01to and
500 W/m at 0.2 s,two
0.05 2under then augmented
irradiation to
changes.
1000 W/m 2 at 0.6 s as shown in Figure 9a. 2
The irradiation was changed from 1000 to 500 W/m at 0.2 s, then augmented to 1000 W/m at 0.6 s as 2

shown in Compared
Figure 9a.to the P&O with a fixed step of 0.01, the P&O with a fixed step of 0.05 shows a good
dynamic performance. It can converge more rapidly to the steady state, but the oscillation is much
Compared to the P&O with a fixed step of 0.01, the P&O with a fixed step of 0.05 shows a good
higher. The dynamic performance of the P&O algorithm can be further improved with a larger step size.
dynamic performance. It can converge more rapidly to the steady state, but the oscillation is much
Figure 9b represent variation of voltage reference step size at 0.2 s and 0.6 s. Figure 9c shows that
higher. The dynamic performance
the steady-state oscillation has been oftotally
the P&O algorithm
eliminated in thecan
casebe further
of the improved
FLC-based P&O and withwhen
a larger
step size.
the output power of the PV generator is greater than 615 W. In addition, the FLC-based P&O offers a
Figure 9b represent
faster dynamic variation
response than theofP&O
voltage reference
of 0.01 step size at 0.2 s and 0.6 s. Figure 9c shows that
and 0.05.
In the second
the steady-state part, the
oscillation hasresults of the simulation
been totally eliminated were obtained
in the to the
case of show the performance
FLC-based P&O and of the
when
fuzzy logic-based P&O algorithm under different conditions such as quick
the output power of the PV generator is greater than 615 W. In addition, the FLC-based P&O offers changes of irradiation.
Thedynamic
a faster results ofresponse
the drive than
system ofP&O
the the solar pumping
of 0.01 motor are obtained to operate at 25 °C and 1000
and 0.05.
W/m2. They are given in Figure 10a–d. The power characteristic of the photovoltaic generator is
In the second part, the results of the simulation were obtained to show the performance of the
shown in Figure 10a with a maximum power point determined as 615 W. The PV generator’s voltage
fuzzy logic-based P&O algorithm under different conditions such as quick changes of irradiation.
Energies 2018, 11, 462 11 of 15

The results of the drive system of the solar pumping motor are obtained to operate at 25 ◦ C11and
Energies 2018, 11, x FOR PEER REVIEW of 15
1000 W/m2 . They are given in Figure 10a–d. The power characteristic of the photovoltaic generator is
shown in Figure
and current 10a with as
are changed a maximum power 10b.
shown in Figure pointItdetermined
should be knownas 615 W. thePV
thatThe PVgenerator’s
generator’s voltage
voltage
and current are changed as shown in Figure 10b. It should be known that the
is adjusted to the MPP of about 26 V. The three-phase stator currents are given in Figure 10c. The PV generator’s voltage is
adjusted to the MPP of about 26 V. The three-phase stator currents are given
motor speed is changed, as illustrated in Figure 10d, which is limited to the maximum available in Figure 10c. The motor
speed
speed isofchanged,
about 3000 as rpm.
illustrated in Figure
As a second 10d,the
work, which is limited
irradiation to the maximum
is adjusted to 500 W/m available speedare
2. The results of
about 3000 rpm. As a second work, the irradiation is adjusted to 500 W/m 2 . The results are given in
given in Figure 10e–h for this adjusted condition. In this circumstance, the MPP of the PV generator
Figure 10e–h at
is obtained forabout
this adjusted
315 W,condition.
as presented In this
in circumstance,
Figure 10e. The the MPP of the PVgenerator’s
photovoltaic generator iscurrent
obtainedis
at about 315 W, as presented in Figure 10e. The photovoltaic generator’s
decreased from 24 A to 12 A, as illustrated in Figure 10f. Depending on this current decrease, current is decreased from
the
24 A to 12 A, as illustrated in Figure 10f. Depending on this current decrease,
stator’s currents and the motor speed are decreased as shown in Figure 10g–h. The motor speed the stator’s currents andis
the motor speed
decreased from are decreased
2980 rpm to as shown
2250 rpm.inThe
Figure 10g–h.
results The that
show motor speed
the motoris decreased
operates at from
the2980 rpm
point of
to 2250 rpm.
maximum power. The results show that the motor operates at the point of maximum power.

(a)

(b)

Figure 9. Cont.
Energies 2018, 11, 462 12 of 15
Energies 2018, 11, x FOR PEER REVIEW 12 of 15

Energies 2018, 11, x FOR PEER REVIEW 12 of 15

(c)
Figure 9.9. Steady-state (c)
Figure Steady-state and
and dynamic
dynamic response
response comparison
comparison of the proposed
of the proposed fuzzy
fuzzy logic
logic controller
controller
(FLC)-based
(FLC)-based P&Owith
P&O withthe
Figure 9. Steady-state the
andP&OP&O algorithm
algorithm
dynamic withwith
response step step of reference
of reference
comparison of voltagevoltage of
of 0.05 and
the proposed 0.05
0.01.
fuzzy and 0.01. (a)
(a) Changes
logic controller
Changes of
of(FLC)-based irradiation;
irradiation; (b)
P&O (b)
Variable Variable
StepP&O
with the Step of
of a Referencea Reference
algorithmVoltage; Voltage;
(c) PV
with step (c)
of array PV array
output
reference output power.
power.of 0.05 and 0.01. (a)
voltage
Changes of irradiation; (b) 2Variable Step of a Reference Voltage; (c) PV array output power.
25°C, 1000 W/m
700 25 °C, 500 W/m2
350
25°C, 1000 W/m
2 1000 W/m2
600
700 25 °C, 500 W/m2
300
350 500 W/m2
1000 W/m2
500
600
250
300
500 W/m2
Power (W)

400
500
Power (W)

200
250
Power (W)

300
400
Power (W)

150
200

200
300 100
150

100
200 50
100

0
100 050
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Voltage (V) Voltage (V)
0 0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
(a)
Voltage (V)
(e)
Voltage (V)
30 (a) 30 (e)
2530 2530

25 25
20 20
Vpv , Ipv

Vpv , Ipv

1520 1520
Vpv , Ipv

Vpv , Ipv

15 15
10 10

510 510

5 5
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time [S] Time [S]
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
(b)
Time [S]
(f)
Time [S]

(b) (f)

Figure 10. Cont.

Energies 2018, 11, 462 13 of 15
Energies 2018, 11, x FOR PEER REVIEW 13 of 15

30 20

15
20
10
10
5
Iabc [A]

Iabc [A]
0 0

-5
-10
-10
-20
-15

-30 -20
0.2 0.205 0.21 0.215 0.22 0.2 0.205 0.21 0.215 0.22
Time [S] Time [S]

(c) (g)
3000

2000
2500

2000 1500
Speed [rpm]

Speed [rpm]
1500
1000

1000

500
500

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
Time [S] Time [S]

(d) (h)
Figure10.10.
Figure Simulation
Simulation results
results of FLC-based
of FLC-based P&O algorithm.
P&O algorithm. OperatingOperating at 1000
at 25 ◦ C and 25 °C and2 conditions;
W/m 1000 W/m2
(a)conditions;
P-V curve (a) P-Varray,
of PV curve(b)of the
PV array, (b) the
PV array’s PV array’s
current current(c)
and voltage, and voltage,
Stator (c) Stator
currents, currents,
(d) Motor (d)
speed,
Motor speed, ◦ operating at 252 °C and 500 W/m 2 conditions; (e) the PV array’s P-V curve, (f) the
operating at 25 C and 500 W/m conditions; (e) the PV array’s P-V curve, (f) the parameters (I, V) of the
PVparameters (I, V) currents,
array, (g) Stator of the PV(h)
array,
Motor(g)speed.
Stator currents, (h) Motor speed.

7.7.Conclusions
Conclusions

InInthis
thispaper,
paper,ananimprovement
improvementof ofthe
theP&O
P&O algorithm
algorithm with
with the
the Fuzzy
Fuzzy Logic
Logic isis developed
developed for for a
permanent magnet synchronous motor (PMSM) drive system to increase the
a permanent magnet synchronous motor (PMSM) drive system to increase the efficiency of the system efficiency of the system
withoutthe
without theenergy
energystorage
storagebybya DC/DC
a DC/DCconverter
converterand andbatteries.
batteries.The
Thesystem
systemisisconnected
connecteddirectly
directlytoto
the photovoltaic panels. To ensure a nominal power supply of 26 volts for the motor, a minimum ofof
the photovoltaic panels. To ensure a nominal power supply of 26 volts for the motor, a minimum
threepanels
three panelsininparallel
parallelisissufficient.
sufficient.
Fora adirect
For directconnection
connectiontotoPV PVpanels
panelsthat
thatcan
cangenerate
generatea alow
lowvoltage
voltagelevel,
level,a alow-voltage
low-voltagePMSMPMSM
is used. To reduce the torque ripple, we use six poles. Because of the need
is used. To reduce the torque ripple, we use six poles. Because of the need for a 3000 rpm speed, for a 3000 rpm speed, the
nominal
the nominal frequency
frequency of of
thethemotor
motoris is
equal
equalto to
150
150Hz.
Hz.
In addition, so as to maximize the power
In addition, so as to maximize the power flow between flow betweenthethePMSM
PMSMand andthethePVPVarray
arrayininsituations
situations
of vigorous radiation and temperature changes, the FLC-based P&O algorithm
of vigorous radiation and temperature changes, the FLC-based P&O algorithm has been developed. has been developed.
InIna achanging
changingcondition
conditionenvironment,
environment,the theanalysis
analysisofofthe
theperformance
performanceofofthe thedeveloped
developedFLC-based
FLC-based
P&O algorithm, low voltage PMSM, motor drive, and
P&O algorithm, low voltage PMSM, motor drive, and solar system using MATLAB/SimPowerSystems solar system using
MATLAB/SimPowerSystems blocks is performed. The results of the simulation
blocks is performed. The results of the simulation show that at its peak capacity, the pump motor show that at its peak
capacity,
works the pump
perfectly evenmotor
underworksrapidlyperfectly
changing even under
solar rapidlyThis
radiation. changing
provessolar
that radiation.
the MPPs Thishaveproves
been
that the MPPs have been accurately measured
accurately measured with the developed P&O algorithm. with the developed P&O algorithm.

AuthorContributions:
Author Contributions:Mohamed
MohamedRedha
RedhaRezoug,
Rezoug,Rachid
RachidChenni,
Chenni,and
andDjamel
DjamelTaibi
Taibiconceived
conceivedand
anddesigned
designedthe
the
experimentsand
experiments and performed
performed them;
them; Djamel
Djamel TaibiTaibi and Mohamed
and Mohamed RedhaRedha
RezougRezoug
analyzed analyzed
the data;the data;Chenni
Rachid Rachid
contributed reagents/materials/analysis
Chenni contributed tools; Mohamed
reagents/materials/analysis Redha Rezoug
tools; Mohamed Redha wrote thewrote
Rezoug paper.the paper.
Conflicts
Conflictsofof
Interest: The
Interest: authors
The declare
authors nono
declare conflict ofof
conflict interest.
interest.
Energies 2018, 11, 462 14 of 15

Appendix A.
PMSM Parameters;
V dc = 26 V; R = 0,002 Ω; Ld = 786.10−5 H; Lq = 102.10−4 H; p = 3; Tn = 1.8 Nm; ψm = 0.058 Wb; Pn = 415 W;
In = 20 A; Ωn = 2980 rpm; Un = 15 V.

References
1. Tyagi, V.V.; Rahim, N.A.A.; Rahim, N.A.; Selvaraj, J.A./L. Progress in solar PV technology: Research and
achievement. Renew. Sustain. Energy Rev. 2013, 20, 443–461. [CrossRef]
2. Sahin, A.Z.; Rehman, S. Economical Feasibility of Utilizing Photovoltaics for Water Pumping in Saudi Arabia.
Int. J. Photoenergy 2012, 2012, 9. [CrossRef]
3. Faranda, R.; Leva, S. Energy comparison of MPPT techniques for PV Systems. WSEAS Trans. Power Syst.
2008, 3, 446–455.
4. Salas, V.; Olías, E.; Barrado, A.; Lázaro, A. Review of the maximum power point tracking algorithms for
stand-alone photovoltaic systems. Sol. Energy Mater. Sol. Cells 2006, 90, 1555–1578. [CrossRef]
5. Hohm, D.; Ropp, M.E. Comparative study of maximum power point tracking algorithms.
Prog. Photovolt. Res. Appl. 2003, 11, 47–62. [CrossRef]
6. Piegari, L.; Rizzo, R.; Spina, I.; Tricoli, P. Optimized Adaptive Perturb and Observe Maximum Power Point
Tracking Control for Photovoltaic Generation. Energies 2015, 8, 3418–3436. [CrossRef]
7. Cabal, C.; Alonso, C.; Cid-Pastor, A.; Estibals, B.; Seguier, L.; Leyva, R.; Schweitz, G.; Alzieu, J.
Adaptive digital MPPT control for photovoltaic applications. In Proceedings of the 2007 IEEE International
Symposium on Industrial Electronics, Vigo, Spain, 4–7 June 2007; pp. 2414–2419.
8. Liu, X.; Lopes, L.A. An improved perturbation and observation maximum power point tracking algorithm
for PV arrays. In Proceedings of the 2004 IEEE 35th Annual Power Electronics Specialists Conference,
Aachen, Germany, 20–25 June 2004; pp. 2005–2010. [CrossRef]
9. Emilio, M.; Giovanni, P.; Giovanni, S. Two-steps algorithm improving the P&O steady state MPPT efficiency.
Appl. Energy 2014, 113, 414–421. [CrossRef]
10. Walker, G. Evaluating MPPT converter topologies using a MATLAB PV model. J. Electr. Electron. Eng. 2001,
21, 49–56.
11. Villalva, M.G.; Gazoli, J.R.; Filho, E.R. Comprehensive Approach to Modeling and Simulation of Photovoltaic
Arrays. IEEE Trans. Power Electron. 2009, 24, 1198–1208. [CrossRef]
12. Liu, F.; Kang, Y.; Zhang, Y.; Duan, S. Comparison of P&O and hill climbing MPPT methods for grid-connected
PV converter. In Proceedings of the 2008 3rd IEEE Conference on Industrial Electronics and Applications,
Singapore, 3–5 June 2008.
13. Tan, C.W.; Green, T.C.; Hernandez-Aramburo, C.A. Analysis of perturb and observe maximum power point
tracking algorithm for photovoltaic applications. In Proceedings of the IEEE 2nd International Power and
Energy Conference, Johor Bahru, Malaysia, 1–3 December 2008.
14. Aashoor, F.A.O.; Robinson, F.V.P. A Variable Step Size Perturb and Observe Algorithm for Photovoltaic
Maximum Power Point Tracking. In Proceedings of the 47th International Universities Power Engineering
Conference (UPEC), London, UK, 4–7 September 2012.
15. Piegari, L.; Rizzo, R. Adaptive perturb and observe algorithm for photovoltaic maximum power point
tracking. IET Renew. Power Gener. 2010, 4, 317–328. [CrossRef]
16. Weidong, X.; Dunford, W.G. A modified adaptive hill climbing MPPT method for photovoltaic power systems.
In Proceedings of the 2004 IEEE 35th Annual Power Electronics Specialists Conference, Aachen, Germany,
20–25 June 2004.
17. Chihchiang, H.; Chihming, S. Study of maximum power tracking techniques and control of DC/DC
converters for photovoltaic power system. In Proceedings of the 29th Annual IEEE Power Electronics
Specialists Conference, Fukuoka, Japan, 22 May 1998.
18. Alajmi, B.N.; Ahmed, K.H.; Finney, S.J.; Williams, B.W. Fuzzy-logic-control approach of a modified hill-climbing
method for maximum power point in microgrid standalone photovoltaic system. IEEE Trans. Power Electron. 2011,
26, 1022–1030. [CrossRef]
Energies 2018, 11, 462 15 of 15

19. Musa, S.; Mohd-Radzi, M.A.; Hizam, H.; Abdul-Wahab, N.I.; Hoon, Y.; Mohd-Zainuri, M.A.A. Modified synchronous
reference frame based shunt active power filter with fuzzy logic control pulse width modulation inverter. Energies
2017, 10, 758. [CrossRef]
20. Deniz, E.; Aydogmus, O.; Aydogmus, Z. Implementation of ANN-based selective harmonic elimination
PWM using hybrid genetic algorithm-based optimization. Measurement 2016, 81, 32–42. [CrossRef]
21. Wang, Q.G.; Zhang, Z.; Astrom, K.J.; Cheka, L.S. Guaranteed dominant pole placement with PID controllers.
J. Process Control 2009, 19, 349–352. [CrossRef]
22. Benlarbi, K.; Mokrani, L.; Nait-Said, M.S. A fuzzy global efficiency optimization of a photovoltaic water
pumping system. Sol. Energy 2004, 77, 203–216. [CrossRef]
23. Mohammed, Y. Design and Implementation of a Robust Current-Control Scheme for a PMSM Vector Drive
with a Simple Adaptive Disturbance Observe. IEEE Trans. Ind. Electron. 2007, 54, 1981–1988. [CrossRef]
24. Hilloowala, R.M.; Sharaf, A.M. A rule-based fuzzy logic controller for a PWM inverter in a stand alone wind
energy conversion scheme. IEEE Trans. Ind. Appl. 1996, 32, 57–65. [CrossRef]
25. Farhat, M.; Barambones, O.; Sbita, L. Efficiency optimization of a DSP-based standalone PV system using
a stable single input fuzzy logic controller. Renew. Sustain. Energy Rev. 2015, 49, 907–920. [CrossRef]

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).