4th International Conference on Power Engineering, Energy and Electrical Drives Istanbul, Turkey, 13-17 May 2013

MPPT Techniques for PV Systems

Dalila BERIBER Abdelaziz TALHA

LINS Laboratory, Faculty of Electronics and Computer, LINS Laboratory, Faculty of Electronics and Computer,
University of Sciences and Technology Houari Boumediene University of Sciences and Technology Houari Boumediene
Allergies, Algeria Allergies, Algeria
dberiber@yahoo.fr abtalha@gmail.com

Abstract—This paper propose a detailed comparative survey of

four maximum power tracking techniques: Perturb and Observe II. DESCRIPTION OF THE PHOTOVOLTAIC GENERATOR
(P&O), Incremental Conductance (InC), fuzzy logic based The electrical equivalent-circuit of a solar cell is shown in
tracking technique and a, less known, method using only the figure1. It is composed of a light-generated current source,
photovoltaic current measurement. The drawback of the three diodes, series resistance, and parallel resistance [1].
studied methods; P&O, InC and one sensor algorithm, is that at
steady state the operating point oscillate around the maximum Figure 1. Equivalent electrical circuit of a cell
power point, giving rise to the waste of the output panel’s
available energy. Simulation results show that the proposed fuzzy
logic controller (FLC) can provides faster and stable tracking
maximum power as compared to the other studied methods. Rs I
Id Ish
Keywords- MPPT, Buck-Boost, Perturb and Observe, Fuzzy Iph Vd Rsh V RLoad
logic controller, NPC-VSI, Grid.

I. INTRODUCTION
Characteristic equation for the current and voltage of a
Solar power is an alternative technology that will hopefully solar cell is given as fellows [1] [3]:
lead us away from our petroleum dependent energy sources.
The major problem with solar panel technology is that the q .( V + R s . I ) V + R s .I
efficiencies for solar power systems are still poor and the costs I = I ph − I sat .[exp( ) − 1] − (1)
per kilo-watt-hour (kwh) are not competitive, in most cases, to nkT R sh
compete with petroleum energy sources. Solar panels
themselves are quite inefficient (approximately 30%) in their where I denotes a current of a solar array (A), V denotes an
ability to convert sunlight to energy. However, the charge output voltage of a solar array(V), Iph denotes the light
controllers and other devices that make up the solar power generated current (A), Isat denotes a diode reverse saturation
system are also somewhat inefficient and costly. Our goal is to current (A), q denotes the electronic charge =1,6.10-19C, n
design a Maximum Power Point Tracker (MPPT), a specific denotes a dimensionless deviation factor from the ideal p–n
kind of charge controller that will utilize the solar panel to its junction diode, k is Boltzmann’s constant =1.3807.10-23 JK-1, T
maximum potential. denotes a cell temperature (K), Rs denotes a series resistance
The MPPT is a charge controller that compensates for the (Ÿ), and Rsh denotes a shunt resistance (Ÿ).
changing Voltage Current characteristic of a solar cell. The Figure2 gives the power–voltage (P–V) characteristics of a
MPPT fools the panels into outputting a different voltage and PV module respectively for different values of solar radiation
current allowing more power to go into the battery or batteries and temperature.
by making the solar cell think the load is changing when you It is seen that the output characteristics of the solar array is
really are unable to change the load [1]. The MPPT monitors nonlinear and vitally affected by the solar radiation,
the output voltage and current from the solar panel and temperature and load condition.
determines the operating point that will deliver that maximum In order to maximize the output power from a solar module,
amount of power available to the batteries. If our version of the it has to be operated at a unique point with specified voltage
MPPT can accurately track the always-changing operating and current values, or in other words, at a specified load
point where the power is at its maximum, then the efficiency of resistance. This requires a separate power converter circuit for
the solar cell will be increased. the MPPT. In our design, a boost type DC–DC converter is
Many algorithms have been developed for tracking employed to match the load to the PV array to extract the
maximum power point of a PV generator. These algorithms maximum power.
vary in effectiveness, complexity, convergence speed, sensors
required and cost [2]. Four MPPT methods are studied in this
paper; the P&O method, the Incremental Conductance method,
the fuzzy logic method and only current measurement method.

978-1-4673-6392-1/13/$31.00 ©2013 IEEE

POWERENG 2013 1437
4th International Conference on Power Engineering, Energy and Electrical Drives Istanbul, Turkey, 13-17 May 2013

120
IV. DIFFERENT ALGORITHM MPPT
E=1000 W/m² As is well known, the maximum power point (MPP) of
E=800 W/m²
100
E=600 W/m² photovoltaic power generation system depends on array
E=400 W/m² temperature and solar irradiation, so it is necessary to
80 E=200 W/m²
constantly track MPP of solar array. For years, research has
focused on various MPP control algorithms to draw the
P (W )

60
maximum power of the solar array. In this section, the
40 effectiveness of these four different control algorithm are
thoroughly investigated via numerical simulation.
20
A. Perturb and Observe method
0
5 10 15 20 25 30
Perturbation and Observation method has been widely used
V(V) due to its ease of implementation [6]. P&O algorithm will force
the PV system to approach to the maximum power point by
120
T= 0°C increasing or decreasing the PV panel-output voltage. Figure4
100
T= 25°C
T=50°C
presents the control flow chart of the P&O algorithm.
T= 75°C

80 Start
P (W )

60 Measure V(k-1) and I(k-1)

40 P (k ) = I (k ) × V (k )
Δ P = P (k ) − P (k − 1 )

20
No
ΔP > 0
0
0 5 10 15 20 25 30
V(V) Yes

No
Figure2. The effect of the irradiation and the temperature on PV generator V (k ) > V (k − 1) V (k ) > V (k − 1)
Yes

III. BOOST TYPE DC–DC CONVERTER Yes No

δ (k + 1) = δ (k ) − D δ (k + 1) = δ (k ) − D
In figure3 the schematic of the boost converter power stage
is given. It consists of the power switch K (MOSFET
transistor), boost inductor L, filter capacitor C2, output diode D δ (k + 1) = δ (k ) + D δ (k + 1) = δ (k ) + D

and load resistor RLoad [4].

Figure4. Flowchart of the P&O algorithm

Ii Ic1 L IL D Is Io In order to find the direction change for maximizing power,

Ic2 the P&O method perturbs the operating voltage of the PV
C1 K Vs C2 RLoad Vo panel; if power increases, then the operating voltage is further
Vi perturbed in the same direction, whereas if it decreases, then
the direction of perturbation is reversed. This process is
repeated periodically until the MPP is reached [6]. The system
then oscillated around the MPP. The duty cycle perturbation at
Figure3. MPPT Boost converter
time (t+1) can be decided on the basis of the following
relationship [7]:
The converter steady state waveforms in the continuous
conduction mode. When the switch K is in the on state, the
current in the boost inductor increases linearly, and at that d (t + 1) = d (t ) + (2 Sign − 1)D (3)
time, the diode is in the off state. When the switch K is turned
off, the energy stored in the inductor is released through the Where Sign is given by:
diode to the output RLoadC2 circuit. The pulsating current
produced by the switching action is smoothed by the Sign = ([P (t ) − P (t − 1)] > 0 ) ⊕ ([V (t ) − V (t − 1)] > 0 ) (4)
capacitive filter and a DC voltage is provided to the load. The
boost converter transfer function is obtained by considering its P(t) and V(t) are, respectively, power and voltage drawn
steady state operation [3] [5]. The DC voltage transfer from the PV panel.
function is:
The oscillation around the MPP can be minimized by
reducing the perturbation step-size D. however dynamic
Vo − δ (2) performance is hampered by smaller perturbation step-size
M (δ ) = =
Vi 1− δ (Figure5). This trade off requires careful tuning of the duty
cycle perturbation step-size.

POWERENG 2013 1438

4th International Conference on Power Engineering, Energy and Electrical Drives Istanbul, Turkey, 13-17 May 2013

100 Start
Available maximum power
V(k-1) , I(k-1)
80
Step-Size = 0,008 V(k) , I(k)
Panel's Power [W]

60 No Yes
ΔV = 0
Yes Yes
40 ΔI I ΔI ≤ ε
+ ≤ε
ΔV V
No
20 Yes
No ΔI > 0
Step-Size = 0,002
ΔI I Yes
0
0 5 10 15 20 25 + >0 No
ΔV V
Time [s]
No
Figure5. Perturbations Step-size effects on the performances of the P&O
algorithm δ (k + 1) = δ (k ) + D δ (k + 1) = δ (k ) + D

δ (k + 1) = δ (k ) − D δ (k + 1) = δ (k ) − D
B. Incremental Conductance method
Incremental conductance (IncCond) method is based on the
fact that the slope of PV panel power versus voltage curve is Figure6. The InC algorithm flowchart
zero at the MPP, positive on the left, and negative on the right
of the MPP [7]. The relationship between the instantaneous
conductance (I V) and the incremental conductance (ΔI ΔV) is
83.6

given by: Panel's Power [W]

83.58

83.56
­ ΔI I
° ΔV + V = 0 at MPP
° 83.54
° ΔI I (5)
® + > 0 left of MPP
° Δ V V 83.52

° ΔI I
° ΔV + V < 0 right of MPP 18 18.1 18.2 18.3 18.4 18.5 18.6 18.7 18.8
¯ Time [s]
İ=0

Because of the noise, of measurement’s faults and the 83.58

Available maximum power
quantification, the condition (ΔI ΔV ) + (I V ) = 0 is seldom 83.578
Panel's Power [W]

satisfied, therefore in steady state, the system oscillate around 83.576

the MPP. To overcome this drawback we introduce a new 83.574

parameter İ, as: 83.572

83.57
83.568
ΔI I
+ ≤ε (6)
83.566

ΔV V 83.564
20.39 20.4 20.41 20.42 20.43
Time [s]

The IncCond algorithm is shown in the flowchart figure6 İ = 0,01

[8]. Figure7. Stationary regime

The amplitude of the oscillations, around the MPP, is

controlled by the value of İ. It decreases with the increase of İ C. Current measurement method
(Figure7). The methods discussed so far require both current and
voltage measurement. Salas and al [10] proposed a new
However, for a relatively great value of İ, the operating algorithm for seeking maximum power point of a PV generator
point moves away from the true MPP. Hence, the parameter İ needing only the PV current value. In this paper, the studied
value is to be chosen carefully for improved performance of system is formed by a PV panel and the output section by a
the MPPT system [9]. 12V battery. A DC/DC boost converter is inserted between the
PV panel and the battery. The energy's conversion principle,
applied to the static converter, gives:

Pin = V pv I pv = Vbat I out (7)

POWERENG 2013 1439

4th International Conference on Power Engineering, Energy and Electrical Drives Istanbul, Turkey, 13-17 May 2013

Where Pin is the available power at the output of the PV The tow inputs of the FLC are the error E and, also, the
panel and Vbat is the battery voltage, which is assumed associated change of error CE, which are shown in formula10.
constant. For boost converter, output current is given by:
­ P(k ) − P(k − 1)
° E (k ) = V (k ) − V (k − 1)
I out = −
(1 − δ ) I (8) ® (11)
δ pv
°CE (k ) = E (k ) − E (k − 1)
¯
Formula (6) becomes: Where P(k) and V(k) refers to the output power and voltage
of PV panel at the sampling instant k. gE and gCD are the
V pv I pv = − V bat
(1 − δ ) I (9) inputs scaling factors, and gdD is the Defuzzification gain.
pv
δ While dD denotes the output of the fuzzy process.
The fuzzy logic controller consists of three functional
So we define an objective function P* as: blocks: fuzzification, Fuzzy rules and inference engine, and
finally Defuzzification.

P* =
(1 − δ ) I (10)
Fuzzification
pv
δ The fuzzy process requires that each variable used in
describing the control rules has to be expressed in terms of
It appears that the maxima of both Pin and its corresponding fuzzy set notations with linguistic labels [13]. Figure10 show
objective function P* will coincide. When P* is maximized the memberships functions of the input variables E(k) and
using the P&O algorithm, for example, it tracks the maximum CE(k) and the output variable dD(k). In which each
power closely and also respond to changes in atmospheric membership function is assigned with five fuzzy set, including
conditions efficiently (Figure8 and 9) [10] [11]. PB (Positive Big), PS (Positive Small), ZE (Zero Equivalent),
NS (Negative Small) and NB (Negative Big).
100 NG NP ZE PP PG
1

80 0.8
83.65
Degree of membership
Panel Power [W]

83.6 0.6

60
Panel Power [W]

83.55
0.4
83.5
40
83.45 0.2

83.4
20 0
83.35
22 22.2 22.4 22.6 22.8 -0.02 -0.015 -0.01 -0.005 -0,0025 0 0.0025 0.005 0.01 0.015 0.02
Time [s] E
0
0 2 4
Time [s]
6 8 10
(a)
Figure8. Algorithm performances in a constants atmospherics conditions 1
NG NP ZE PP PG

0.8

100
Degree of membership

1000W/m2 0.6

80
0.4
Panel's Power [W]

60 0.2

T = 25°C 0

40
-200 -160 -80 0 80 160 200
CE

20 200W/m2 200W/m2 (b)

NG NP ZE PP PG
0 1
0 10 20 30 40 50 60 70 80
Time [s]
0.8

Figure9. Algorithm performances in a variables atmospherics conditions

Degree of membership

0.6

D. Fuzzy logic controller method 0.4

Advances in microelectronic technology permitted to the 0.2

fuzzy logic control to become the most significant and fruitful 0

application for fuzzy logic theory. Fuzzy logic controllers, -0.02 -0.015 -0.01 -0.005 -0.0025 0 0.0025 0.005 0.01 0.015 0.02

based on fuzzy logic, provides a mathematical tool for

D

converting linguistic control rules in the form of (IF-THEN) (c)

statements into an automatic control strategy [12] [13] [14]. Figure10. Membership function

POWERENG 2013 1440

4th International Conference on Power Engineering, Energy and Electrical Drives Istanbul, Turkey, 13-17 May 2013

Fuzzy rules and inference engine 83.65

The kernel of fuzzy logic controller is the fuzzy inference Maximum available power = 83.577W
system. Fuzzy inference is the process of formulating the 83.6

mapping from a given input to an output using fuzzy logic. The 83.55

Panel Power [W]

mapping then provides a basis from which decisions can be
made. The proposed Mamdani-type inference system 83.5

endeavours to force the error function (E in formula 11) to 83.45 P&O algorithm
zero. Two cases are to consider [14]: InC algorithm
• First case: E is positive; working point is on the left 83.4 Fuzzy algoritm

of the MPP. If the change of error CE is positive, then the Current Only algorithm
83.35
working point converges toward the MPP. If CE is negative, 47.7 47.8 47.9 48 48.1 48.2 48.3 48.4 48.5 48.6
the inverse that occurs. Time [s]

• Second case: E is negative; working point is, Figure12. Comparing controllers performances in a constants atmospherics
therefore, on the right of the MPP. In this case if CE is positive, conditions-steady state
working point moves away of the MPP and vice versa if CE is
negative. 90

From that, we summarises, in table1, this process reasoning 80 1000W/m2

as a set of a fuzzy IF-THEN rules [14]. 70

Panel Power [W]

60
Table1. Inference Matrix 50 P&O algorithm
InC algorithm
CE 40
Fuzzy algoritm
NG NP ZE PP PG 30 Current Only algorithm
20
NG ZE ZE PG PG PG 200W/m2 200W/m2
10
NP ZE ZE PP PP PP
E ZE PP ZE ZE ZE NP
0
0 10 20 30 40 50 60 70 80
Time [s]
PP NP NP NP ZE ZE
Figure13. Comparing controllers performances in a variable atmospherics
PG NG NG NG ZE ZE conditions, T=25°C

Defuzzification
-15°C
The process of Defuzzification calculates the crisp output 100 -15°C 1000W/m2
of the FLC. It describes the mapping from a space of fuzzy
logic statement, corresponding to the inferred output, into a +45°C
80
Panel's Power [W]

non-fuzzy control action. In this paper the centre of gravity 89.55

87.05

Defuzzifier, which is the most common one, is adopted.

Panel's Power [W]

Panel's Power [W]

89.5 87
60 86.95
89.45

V. SIMULATION RESULTS
86.9

89.4 86.85

40
The four studied MPPT algorithms are compared in terms
86.8
34 34.05 34.1 34.15 34.2 34.25 34.3 34.35 34.4 108.9 108.95 109 109.05 109.1 109.15 109.2 109.25 109.3
Time [s] Time [s]
P&O algorithm
of their tracking capability at steady state (Figure 11 and 12) 20 InC algorithm
and variable environmental conditions (Figure 13 and 14). Fuzzy algoritm
Current Only algorithm
0
0 50 100 150
90 Time [s]
Figure14. Comparing controllers performances in a variable atmospherics
80
Maximum available Power = 83.755W conditions, E=1000W/m2
70

60
At standard conditions figure11 shows the transient
Panel Power [W]

responses of the tracked power obtained from the four MPP

50 controllers. It can be observed that the FLC reaches MPP faster
40
P&O algorithm compared to the other controllers. Steady state behaviour of the
InC algorithm
PV system (Figure12) using FLC is more stable than the other
30 FLC algoritm
Current Only algorithm
MPPT methods. Power’s waste is, here, considerably reduced.
20 Figure13 shows the performance of the PV system, using the
10
four MPPT methods, under constant temperature and a
changing irradiance, whereas, figure14 shows the performance
0
0 0.5 1 1.5 2 2.5 3 3.5 4 of the four controllers, under constant irradiance and variable
Time [s] temperature. In variable atmospheric conditions, the simulation
Figure11. Comparing controllers performances in a constants atmospherics results show that the performances of the four controllers are
conditions quite similar.

POWERENG 2013 1441

4th International Conference on Power Engineering, Energy and Electrical Drives Istanbul, Turkey, 13-17 May 2013

VI. CONCLUSION supplied PV system using fuzzy controller. IEEE Trans. Ind. Electron,
2003; 50: 749–758.
Perturb and observe controller is very simple and can be [5] Enrique JM, Duràn E, Sidrach-de-Cardona M, Andùjar JM. Theoretical
carried out easily. A drawback of P&O algorithm is that, at assessment of the maximum power point tracking efficiency of
steady state, the system’s operating point oscillates around the photovoltaic facilities with different converter topologies. Solar
MPP giving rise to the waste of the available power. The Energy,2007; 81: 31-38.
choosing of the perturbation step-size is very critical; the step- [6] Femia N, Petrone G, Spagnolo G, Vitelli M. Optimization of Perturb and
size determines how fast the MPP is reached, fast tracking can Observe Maximum Power Point Tracking Method. IEEE Trans. Power
Electron., 2005; 20: 963–973.
be achieved with bigger step-size, but the oscillations around
[7] Youngseok J, Junghun S, Gwonjong Y, Jaeho C. Improved Perturbation
the MPP will be raised. There is tradeoff between the dynamic and Observation Method (IP&O) of MPPT control for photovoltaic
and steady performance. The InC method, witch is more power systems. The 31st Photovoltaic Specialists Conference, Lake
complex than the P&O, permits a slight reduction in the Buena Vista, Florida, USA, 2005, pp. 1788–1791.
oscillation’s amplitude, but the system might not operate at the [8] Hussein KH, Mota I. Maximum photovoltaic power tracking: An
MPP. InC method suffers from the same problems associated algorithm for rapidly changing atmospheric conditions. in IEE Proc.
to P&O algorithm such as requirement of ad-hoc tuning Generation Transmiss. Distrib., 1995, pp. 59–64.
parameters, tradeoff between dynamics and steady state [9] Kobayashi K, Takano I, Sawada Y. A study on a two stage maximum
power point tracking control of a photovoltaic system under partially
performance. The major advantage of the single current sensor shaded insolation conditions. in IEEE Power Eng. Soc. Gen. Meet.,
technique is the fact that it uses the measurement of only one 2003, pp. 2612–2617.
variable: the photovoltaic current. The proposed FLC provides [10] Salas V, Olias E, Lazaro A, Barrado A. New algorithm using only one
faster and stable tracking of maximum power as compared to variable measurement applied to maximum power point tracker. Solar
the other MPPT methods studied in this paper. Energy Material and Solar Cells 87, 2005, pp. 675–684.
[11] D’Souza N S., Lopes LAC, Liu X. An intelligent maximum power point
REFERENCES tracker using peak current control. in Proc. 36th Annu. IEEE Power
Electron. Spec. Conf., 2005, pp. 172–177.
[1] Gergaud O, Multon B, Ben Ahmed H. Analysis and experimental
validation of various photovoltaic system models. 7th International [12] Hilloowala RM, Sharaf AM. A rule-based fuzzy logic controller for a
ELECTRIMACS Congress, Montréal, Canada, 2002, pp. 1-6. PWM inverter in photovoltaic energy conversion scheme. in Proc. IEEE
Ind. Appl. Soc. Annu. Meet., 1992, pp. 762–769.
[2] Esram T, Chapman P.L. Comparison of photovoltaic array maximum
power point tracking techniques. IEEE Trans. Energy Conversion, 2007; [13] Khaehintung N, Pramotung K, Tuvirat B, Sirisuk P.
22: 439–449. RISCmicrocontroller built-in fuzzy logic controller of maximum power
[3] Tafticht T, Agbossou K, Doumbia ML, Chériti A. An improved point tracking for solar-powered light-flasher applications. in Proc. 30th
Annu. Conf. IEEE Ind. Electron. Soc., 2004, pp. 2673–2678.
maximum power point tracking method for photovoltaic systems.
Renewable Energy, 2008; 33: 1508–1516. [14] Won CY, Kim DH, Kim SC, Kim WS, Kim H.-S. A new maximum
power point tracker of photovoltaic arrays using fuzzy controller. in
[4] Veerachary M, Senjyu T, Uezato K. Neural-network-based maximum-
Proc. 25th Annu. IEEE Power Electron. Spec. Conf., 1994, pp. 396–403.
power-point tracking of coupledinductor interleaved-boostconverter-

POWERENG 2013 1442