CEAI, Vol.18, No.1 pp.

86-94, 2016 Printed in Romania

An Improved Sliding Mode Control for Maximum Power Point Tracking in

Photovoltaic Systems
Abdelhakim Belkaid,1,2,3 Jean-Paul Gaubert,2
Ahmed Gherbi 3

1
Department of Electromechanics, University of Bordj Bou Arreridj, El-Anasser 34030,
Bordj Bou Arreridj, Algeria (Tel: 33-(0)751282309; e-mail: [email protected]).
2
Laboratory of Information Technology and Automatic control for the Systems- LIAS ENSIP, Bat B25, 2 rue
Pierre Brousse, BP 633, 86022 Poitiers, France (e-mail: [email protected])
3
Automatic Laboratory of Setif (LAS), University of Setif 1, El Maabouda, Street of Bejaia, 19000 Setif,
Algeria (e-mail: [email protected])

Abstract: Tracking the maximum power point (MPP) has a great interest in the study of photovoltaic
(PV) systems. This task is very difficult due to the non linearity of PV current-voltage characteristics
which are dependent on the temperature and irradiation conditions. The sliding mode control (SMC)
based MPPT with two different step sizes designed for Boost-type DC/DC converter method is
investigated in this paper. The robustness of the proposed controller is tested under rapidly changing
solar radiations. The SMC based MPPT is compared to perturb and observe (P&O) method and to
incremental conductance (IncCond) method. The PV-MPPT system is simulated by Matlab/Simulink
environment and verified by practical implementation within DS1104 R&D controller board. The
simulation and experimental results are satisfactory and demonstrate that the new SMC can follow the
PV peak power at different operating conditions with best performance.
Keywords: Photovoltaic; dSPACE controller board; Maximum power point tracking; Sliding mode
control based MPPT; Boost converter.
optimal voltage by open circuit voltage. The last two methods
1. INTRODUCTION
are very simple, but they have a weaker and less accurate
Direct conversion of the suns radiation into electrical energy performance. Reference (Petreus, 2010) compares between
using semiconducting materials is known under the name of four MPPT algorithms which are: fixed step P&O, variable
photovoltaic effect. Materials presently used for photovoltaic step P&O, incremental conductance and fuzzy logic.
include mono-crystalline silicon, polycrystalline silicon, and
The SMC is mostly used to control the power electronic
amorphous silicon. Unfortunately, these materials exhibit low
converters which constitute a case of variable structure
efficiency in energy conversion. The PV output power
systems (Guldemir, 05). Recently, this method is used in
generation is influenced by climatic conditions (e.g.
photovoltaic systems (Bianconi, 2011; Bianconi, 2013,
irradiance, panel temperature) and load variation.
Komurcugil, 2012) but it is often used to regulate a reference
Therefore a maximum power point tracking (MPPT) grandeur obtained with another technique such as P&O or
technique which is intended to control the DC/DC converter IncCond which make the system more complex. This paper
duty cycle is required to ensure an optimal operation of the proposes SMC based MPPT not to control but to optimize the
PV array at different operating conditions (Bratcu, 2008). An PV array yield. It is organized as follows, section 2 and 3
overview of more than thirty of these MPPT techniques is provides a modeling of PV array and boost converter,
done in (Palavee, 2013). Other review articles (Eltawil, 2013; respectively. In section 4, basics of MPPT algorithms are
Esram, 2007; Salas, 2006; Subudhi, 2012) give the detailed illustrated: P&O, IncCond, and SMC. Thereafter, section 5
review of the existing MPPT methods. Among these and 6 depicts the results of simulations and experiments
methods, perturb and observe (P&O) (Femia, 2005; Azzouzi, respectively. Finally, some conclusions are presented in the
2013) and incremental conductance (IncCond) (Dhar, 2013; end of the paper.
Safari, 2011) are widely used in the literature, but they fail
2. MATHEMATICAL MODEL OF THE PV ARRAY
under fast varying climatic conditions (Ghassami, 2013). The
P&O method consists on perturbing the PV output voltage Solar cell or PV cell is the elementary component which
and observing the PV output power to determine the direction converts the energy of light directly into electricity by the PV
of the peak power. The IncCond method tracks the MPP by effect. It can produce approximately 2 Watts under 0.5 Volts.
comparing the instantaneous to the incremental conductance. Different models based on the PV cell equivalent electrical
There are also other techniques such as, fractional short circuit are used to explain this effect. The one-diode model
circuit current method (Kollimalla, 2013) which estimates the (Azzouzi, 2013; Ahmad, 2014; Gergaud, 2002; Jamri, 2010)
optimal current by short circuit current and fractional open or the two-diode model (Petcut, 2010; Dragomir, 2010) are
circuit voltage method (Murtaza, 2012) which estimates the usually considered. The single diode model (fig. 1) is the
CONTROL ENGINEERING AND APPLIED INFORMATICS 87

most classical model described in the literature, it includes a

current generator to model the incident luminous flux, one
diode for the polarization phenomena of the cell, a series
resistance representing the various contacts and connections
resistances, and a shunt resistance characterizing the various
leakage currents due to the diode and to the edge effects of
the junction.
Based on experiments, similar models are used also for
arrays. Hence, in this paper the mathematical PV array model
given by equation (1) (Yu, 2004) will be used.
Fig. 1. PV cell electrical model.
exp 1
From these characteristics, it can be seen that the increase of

the irradiation or the decrease in temperature increases the
(1) power of the panel.
Figure 4 illustrates the experimental identification of the
where and are the number of PV panels coupled in power-voltage characteristic of three modules coupled in
series and parallel, respectively; is the number of PV cells series. The experimental characteristic is obtained by discrete
connected in series in one branch, is the PV array output measurement points. These points are achieved successively
current, is the PV array output voltage, and are by varying a resistance load connected directly to the PV
source as in (Petcut, 2010). The identifying results coincide
respectively the PV panel series and parallel resistances, is
totally with the simulation characteristic derived from the
the ideality factor.
mathematical model adopted.
The photo-current depends on the sunshine and
3. MATHEMATICAL MODEL OF THE DC/DC
temperature as given in the following formula: CONVERTER

(2) The studied system consists on solar energy source coupled
to adaptation converters (maximum power point trackers,
with is the short circuit current temperature coefficient, MPPT). These adapters are DC/DC converters with various

is the short circuit current at standard test conditions topologies. The MPPT DC-DC converter delivers power to
(STC) which are solar irradiance of 1000 W/m2 (1 sun) at DC bus and follows the peak power of photovoltaic panel.
spectral distribution of AM 1.5 and cell temperature of They accept a DC input voltage and output a DC voltage
298 Kelvin. higher, lower or the same as the input voltage. Most MPPT
Moreover, the reverse saturation current is calculated using trackers are based on buck converter (step-down), boost

converter (step-up), buck-boost converter or cuk converter.
(3) These converters use inductors and capacitors to control the
energy flow from the solar module to the load by
where is the open circuit voltage at STC conditions, is continuously opening and closing a switch. The switch is
the open circuit voltage temperature coefficient, and usually an electronic device (MOSFET or IGBT) that
/ is the thermal voltage, 1.38065 10 / operates in two states: in the conduction mode (on) or in the
is the Boltzmanns constant and 1.60218 10 is the cut-off mode (off). The non-isolated boost dcdc converters
electronic charge. are widely used in stand-alone photovoltaic power systems
because of their simplicity and efficiency (Xiao, 2007).
Fig. 2 depicts the PV module model in SIMULINK
environment. Table 1. MSX 60 solar panel specifications (Ishaque,
2012).
The MSX-60 Solar Panel from Solarex Manufacturer with
specifications presented in table I is used for simulations and STC Power Rating 60 W
experiments. Fig. 3 shows the power-voltage and the current- Open Circuit Voltage 21.1 V
voltage characteristics of the PV module under different Short Circuit Current 3.8 A
irradiation levels with fixed temperature (25C). These 17.1 V
Voltage at Maximum Power
characteristics are achieved using MATLAB simulations with
the following parameters 0.357 , 151 , Current at Maximum Power 3.5 A
36, 1,
3.8 ,
21.1 , Temperature coefficient of -0.08 V/C
0.08 / , 0.003 %/ . Temperature coefficient of 0.003 %/C
88 CONTROL ENGINEERING AND APPLIED INFORMATICS

with ,

4
a)
3

simulation characteristic
2
experimental identification

0
0 10 20 30 40 50 60 70
Fig. 2. SIMULINK model of the PV module.

60
4 1000 W/m2 200
simulation characteristic
experimental identification b)
3 750 W/m2 150
40

2 500 W/m2
100

20
1 250 W/m2 50

0 0 0
0 5 10 15 20 0 10 20 30 40 50 60 70


Fig. 3. PV power and current characteristics for different
irradiances. Fig. 4. Experimental identification: a) current-voltage
characteristic b) power-voltage characteristic.
The non-isolated DC/DC boost converter is inserted to
interface the PV output to the rest of the conversion chain. 4. MPPT TECHNIQUES
This converter is widely used in stand-alone PV power
systems, It is characterized by its cyclic ratio (0 1), 4.1 The perturb and observe method
which helps to express the mean values of the output
parameters in function of the input ones. If the chopping The P&O method is most used in the literature and especially
frequency is sufficiently high compared to the characteristic in practice because of its ease of implementation. It works by
frequency of the system, the equations governing the step up periodically perturbing (incrementing or decrementing) the
type DC-DC converter topology using mean values can be panel terminal voltage and comparing the PV output
written as follows: (Keshri, 2014) power with that of the previous perturbation cycle. If the
perturbation leads to an increase (decrease) in panel power,
(4) the succeeding perturbation is made in the same (opposite)
direction. In this way, the MPP tracker continuously seeks
where is the output capacitor voltage. the peak power condition as described by the algorithm of the
In state space form, the dynamic model is obtained as Fig.6.
(Radjai, 2014)
4.2 The Incremental conductance method

(5) The IncCond follows the MPP by comparing the
instantaneous conductance ( / ) with the opposite of
incremental conductance ( / ) of the PV array
where is the inductor current. The control input is the (Radjai, 2014). At the MPP the two items are equal. Fig. 7
switch position; it is set to 0 when the switch is open and it is gives the flowchart of the IncCond method.
set to 1 when the switch is closed.
If we set with matrix 4.3 Improved sliding mode control method
transpose, then the above expression can be written as:
The objective of the sliding mode control is firstly to design
, , (6) switching surface. Then, the second stage consists in
conceiving a control law which is responsible to force the
CONTROL ENGINEERING AND APPLIED INFORMATICS 89

system trajectories towards this area of state space and will Start
maintain it in this one (Bianconi, 2011). On sliding surface,
the system has the desired behavior and is insensitive to
parameter variations and external disturbances. Measure ,

As shown in fig. 3, when the PV panel is operating in its

MPP, the slope of P-V characteristics is null Calculate

0 (7) Yes No
>0
Then, the switching function can be chosen as the slope of
P-V characteristics 0 0
No Yes Yes
(8) No

The sliding mode exists on the switching surface 0. The
general control law combine two terms, a non-linear Fig. 6. Flowchart of the perturb and observe method
component , and an equivalent control .
Start
(9)
to ensure the attractiveness of the control variable to the Measure ,
switching surface; to maintain the operation point in
switching surface and displace it to the origin. No Yes
0
These components determined by considering the Lyapunov
function satisfying the control objective both in attractively
0 0
mode and in sliding mode.
(10) No No

To ensure the attractiveness of the control variable to the 0 0

switching surface, the time derivative of must be negative
No Yes Yes No
definite.

0 0 (11)
The latest expression is known under the name of reaching Fig. 7. Flowchart of the Incremental conductance method
condition.
exp exp (15)

From the expression (1), and cannot be established
from generally available information. For this, we made some We replace (15) in (14):
assumptions that are generally valid for silicon cells:
exp (16)

- In an ideal case, the series resistance is neglected
and the parallel resistance approaches infinity, where is the open circuit voltage, is the short circuit
- exp 1. current. And thus,

By application of the first assumption, the expression (1)

becomes: exp

(17)

exp 1 (12) The derivative is given by:

If the panel is in a short circuit & 0, we 2 exp

(18)

can write:
When 0, the system operates on the left of the MPP,
(13)
the voltage must be increased to attain the MPP ( 0),
And with second assumption, we can write:
replacing in (18) it follows that 0, and hence
exp (14) 0.

When 0, the system functions in right, the voltage
If the panel is in an open circuit 0 & , we
can write: must be decreased ( 0), which implies that 0,
therefore 0.
90 CONTROL ENGINEERING AND APPLIED INFORMATICS

Thus the system could reach global stability, regardless of the IncCond methods. A trapezoidal profile was chosen for
location of the operating point in left or in right of MPP. varying irradiation between two levels from 500 W/m2
(maximum power 30 , optimum voltage
To satisfy reaching condition (11), a constant rate reaching
law can be selected to the non-linear component 8,55 , optimum current 1,75 ) to 1000 W/m2
(maximum power 60 , optimum voltage
(19) 17,1 , optimum current 3,5 ) for a duration time of
where (positive constant), is the scaling factor which is 1.2 s . All simulations and experiments tests are made using
tuned at the design time to adjust the step size. the switching frequency and the following
parameters: . , , and
The equivalent control proposed by Filippov (Slotine, . In simulations, the load is fixed at
1991) characterize the system dynamics on the sliding .
surface, is determined by using the invariance conditions
(Komurcugil, 2012): The obtained simulation results for tracking MPP (maximum
power, optimal current, optimal voltage) using P&O,
0 and 0 (20) IncCond and SMC are illustrated in fig. 9, fig. 10 and fig. 11,
respectively. For example, one can read from fig. 11 at 0.7
(21) second which correspond to standard test conditions (STC)
the following values: maximum power 59.6 ,
From (13), it finds: optimum voltage 34,23 17.115 ,
0 & 0 (22) optimum current 27,8 3.475 .
Combination of formulas (20), (21) and (22), implies 0, Fig. 12 depicts the comparison of tracking MPP between
thus, we obtain: these three methods. It can be noticed that the P&O and the
IncCond methods cannot track the true MPP especially when
1 (23) increasing or decreasing the irradiance. Fig. 13 compares the
instantaneous efficiency between the new SMC, IncCond and
After designing the sliding mode via the design of switching P&O methods. The yield of PV panel by using the proposed
function and the reaching mode, it is then possible to express algorithm is close to unity from 0.027 s until the end of the
the overall control law by combining (19) and (23) irradiance profile. At STC (irradiance of 1000 W/m2,
1 (24) temperature of 25 C), the MSX 60 module offers 60 W
under optimal voltage value of 17.1 V and optimal current
This expression is applied when the operating point moves in value of 3.5 A.
the good direction towards the MPP, but if the operating
point displaces far from the MPP the parameter k must be The results demonstrate that the current of MPP is
doubled. From this principle the method takes name of SMC dramatically affected by fast varying of irradiance, unlike to
with two different step sizes. the MPP voltage which is only slightly affected. Also, it can
be seen that the SMC approach has good dynamic response
5. SIMULATION RESULTS time at start-up; it is ten time faster than the two others
This study proposes a novel MPPT method to optimizing the methods. From figures of the simulation results, one can
yield of a photovoltaic module. The approach is based on conclude that the proposed MPPT method has better
sliding mode control with two different step sizes applied to performance compared to P&O and IncCond methods during
Boost DC/DC converter as shown in fig. 8. both steady state and dynamic state conditions. In other
terms, the proposed MPPT system can track the desired
The work aims to verify the performance of the proposed behavior of the maximum power rapidly and with fewer
MPPT tracker. The new method is compared to P&O and fluctuations which imply good conversion efficiency.
L iL
I pv
D
G
u 1
G V pv K
C1 C2 Vo R
T P
u 0
V

V pv

I pv MPPT PWM
Vo

Fig. 8. The proposed MPPT photovoltaic system configuration.

CONTROL ENGINEERING AND APPLIED INFORMATICS 91

Fig. 13. Comparison of the instantaneous efficiency for the

Fig. 9. Tracking MPP by using P&O method. three methods under a trapezoidal irradiation profile.
6. EXPERIMENTAL RESULTS
A test bench has been designed as shown in figure 14. The
prototype which was built consists of a TDK-Lambda GEN
300-11 programmable DC voltage source as a photovoltaic
emulator, a diode in order to block reverse currents in the PV
source, DC/DC boost converter to step up the voltage level
and resistance load. The LEM PR30 current probe is used for
sensing the PV output current. The differential sensor ST
1000 2-way is used to measure the input and output voltage
of the converter. Operational Amplifier is used to amplify the
control signals (0-5 V) generated by the dSpace card. Fig. 15
shows clearly the operation of the test bench.
Fig. 10. Tracking MPP by using IncCond method.
DC/DC Converter
PC Ampli Op.

DS1104 PV Emulator Resistance

Fig. 14. Experimental test bench.

Fig. 11. Tracking MPP by using new SMC method.

Fig. 12. Tracking MPP for the three methods under a

trapezoidal irradiation profile. Fig. 15. Implementation of the proposed MPPT algorithm.
92 CONTROL ENGINEERING AND APPLIED INFORMATICS

Tektronix oscilloscope is used to view the different signals.

The MPPT algorithms have been implemented in a DS1104
R&D controller board to generate the PWM signal for acting
the IGBT gate of the chopper. Control Desk software is used
to supervise the displacement of the MPP on the P-V
characteristics under fixed or varying climatic conditions.

Fig. 20. Tracking the PV peak power using IncCond method

Fig. 16. Experimental measurement of the PV array power,

voltage and current using the P&O method.

Fig. 21. Tracking the PV peak power using the proposed
MPPT method
The PV emulator is programmed for replacing three MSX 60
panels connected serially feeding a resistance of 72 , the
temperature is kept fixe at 25 C and the irradiance is
Fig. 17. Experimental measurement of the PV array power, changed with a trapezoidal profile between 500 W/m2 and
voltage and current using the IncCond method. 1000 W/m2. Figures 16, 17 and 18 show the waveforms of
output current , voltage and power for
the P&O, IncCond and SMC respectively. The measurement
scale of these grandeurs is as follows: 50 W/div, 500 mA/div,
10 V/div. The results confirm that the optimal current is
dramatically affected by rapid-changing irradiance, unlike the
optimal voltage which is only slightly affected. In fig. 18 one
can see the oscillations in current and voltage curves. These
oscillations are acceptable because the ratio of their width to
the average value is less than 5%. The experimental results
were very similar to the results obtained by simulation. These
results either in simulation or in experimental demonstrate
that the SMC with two different step sizes is able to track the
Fig. 18. Experimental measurement of the PV array power, PV array MPP under constant or rapidly varying of insolation
voltage and current using the proposed MPPT method. with best performance in both steady and dynamic states
compared to P&O and IncCond methods.
Using the Control Desk software, it can be draw Fig. 19, Fig.
20, Fig. 21. These figures present the practical behaviors of
tracking MPP during the trapezoidal profile by using P&O,
IncCond and improved SMC, respectively. They show the
displacement of the peak power (MPP) when the irradiance

profile change from 500 W/m2 to 1000 W/m2 (rising edge)

and from 1000 W/m2 to 500 W/m2 (falling edge). The best
method is that which takes a more rapid course. From the
results, it can be concluded that the proposed MPPT present
Fig. 19. Tracking the PV peak power using the P&O method
CONTROL ENGINEERING AND APPLIED INFORMATICS 93

better performance in time response or in conversion International Conference on Automation Quality and
efficiency than the P&O and IncCond methods. Testing Robotics (AQTR),1-6
Eltawil M. A., Zhao Z., (2013), MPPT techniques for
6. CONCLUSIONS
photovoltaic applications, Renewable Sustainable
In this paper, a high performance MPPT based on sliding Energy Rev. 25, 793-813.
mode control applied to a boost converter has been presented. Esram, T., Chapman, P.L., (2007), Comparison of
The proposed MPPT system by using two different step sizes photovoltaic array maximum power point tracking
is capable to track the maximum power under constant and techniques, IEEE Trans. Energy Convers., 22, 439449
fast changing of solar radiation conditions with best Femia N.,Petrone G.,Spagnuolo G.,Vitelli M., (2005),
performance in both steady and dynamic states. A Optimization of perturb and observe maximum power
compromise between accuracy and rapidity is tacking point tracking method. IEEE Transactions on Power
account when selecting the increments. The new MPP tracker Electronics; 20 (4): 96373.
has been compared to P&O and IncCond methods. It is Gergaud B, Multon B, Ben Ahmed H (2002) Analysis and
designed, simulated and digitally implemented on the experimental validation of various photovoltaic system
DS1104 R&D controller board. So, the aim of this work is models. In: Proceedings of the 7th international
attained and it can say that the proposed control technique ELECTRIMACS congress, Montral, 1-6.
may be considered as an interesting solution in PV systems Ghassami A.A., Sadeghzadeh S.M., Soleimani A., (2013), A
control area. high performance maximum power point tracker for PV
Systems, Electr. Power Energy Syst. 53, 237-243.
ACKNOWLEDGMENT Guldemir H. (2005), Sliding mode control of DCDC boost
The work reported in this paper has been developed within converter, Journal of Appleid Sciences 5 (3), 588-592.
the Laboratory of Information Technology and Automatic Ishaque, K., Salam, Z., Amjad, M., Mekhilef, S., (2012), An
control for the Systems (LIAS), Poitiers University, France. improved Particle Swarm Optimization (PSO)-based
And it has been supported by the CNEPRU research project MPPT for PV with reduced steady-state oscillation.
No. J0201220110016, conducted at the department of IEEE Trans. Power Electron. 27, 36273638.
electrical engineering, University of Setif, Algeria. This Jamri M. S. and Wei T. C., (2010), Modelling and Control of
support is gratefully acknowledged. a Photovoltaic Energy System Using the State-Space
Averaging Technique, American Journal of Applied
REFERENCES Sciences 7 (5): 682-691,
Ahmad H. Besheer, Ahmed M. Kassem, Almoataz Y. Keshri R., Bertoluzzo M., Buja G., (2014), Integration of a
Abdelaziz, (2014). Single-diode model based Photovoltaic panel with an electric city car, Electric
Photovoltaic module: Analysis and comparison Power Components and Systems, 42:5, 481-495.
approach, Electric Power Components and Systems, Kollimalla S.K., Mishra M.K., (2013), A new adaptive P&O
42:12, 1289-1300. MPPT algorithm based on FSCC method for
Azzouzi M. (2013). Optimization of Photovoltaic Generator photovoltaic system, International Conference on
by Using P&O Algorithm under different weather Circuits, Power and Computing Technologies, ICCPCT,
conditions, CEAI, Vol.15, No.2 pp. 12-19. 21st -23rd Mars 2013, Nagercoil, India, 406-411.
Bianconi E., Calvente J., Giral R., Petrone G., Andres Komurcugil H., (2012), Adaptive terminal sliding-mode
Ramos-Paja C., Spagnuolo G., Vitelli M., (2011),A fast control strategy for DCDC buck converters, ISA
current-based MPPT technique based on sliding mode Transactions No.51 (2012), pp.673-681.
control, IEEE International Symposium on Industrial Murtaza A.F., Sher H.A., Chiaberge M., Boero D., De
Electronics (ISIE), 27-30 June, Gdansk, Poland, 59-64. Giuseppe M., Addoweesh K.E., (2012), A novel hybrid
Bianconi E., Calvente J., Giral R., Mamarelis E., Petrone G., MPPT technique for solar PV applications using perturb
Andres Ramos-Paja C., Spagnuolo G., Vitelli M., (2013), & observe and fractional open circuit voltage techniques,
Perturb and observe MPPT algorithm with a current 15th International Symposium on Mechatronika, 5th-7th
controller based on the sliding mode, Electr. Power December, Prague, Czech, pp. 1-8.
Energy Syst. 44, 346-356. Pallavee Bhatnagar A. and Nema B. R. K., (2013),
Bratcu A. I., Munteanu I., Bacha S., Raison B., (2008). Conventional and global maximum power point tracking
Maximum Power Point Tracking of Grid-connected techniques in photovoltaic applications: A review, j.
Photovoltaic Arrays by Using Extremum Seeking renew. sustainable energy 5, 032701 (2013).
Control, CEAI, Vol.10, No.4 pp. 3-12. Petcu F., Dragomir T.L., (2010), Solar Cell Parameter
Dhar S., Sridhar R., Mathew G., (2013), Implementation of Identification Using Genetic Algorithms, CEAI, vol. 12,
pv cell based standalone solar power system employing No. 1, 30-37
incremental conductance mppt algorithm, IEEE Petreus, D. ; Moga, D. ; Rusu, A. ; Patarau, T. ; Daraban, S.,
International Conference on Circuits, Power and (2010), A Maximum Power Point tracker for
Computing Technologies (ICCPCT), 21st -23rd Mars photovoltaic system under changing luminosity
2013, Nagercoil, India, pp. 356-361. conditions, 2010 IEEE International symposium
Dragomir, T.L. ; Petreus, D.M. ; Petcut, F.M. ; Ciocan, I.C., Industrial Electronics (ISIE), 556 561
(2010), Comparative analysis of identification methods Radjai T., Rahmani L., Mekhilef S., Gaubert J-P., (2014),
of the photovoltaic panel characteristics, IEEE Implementation of a modified incremental conductance
94 CONTROL ENGINEERING AND APPLIED INFORMATICS

MPPT algorithm with direct control based on a fuzzy Subudhi B., Pradhan R., (2012), A comparative study on
duty cycle change estimator using dSPACE, Solar maximum power point tracking techniques for
Energy, 110 (2014) 325-337. photovoltaic power systems, IEEE Trans. Sust. Energy,
Safari, A., Mekhilef, S., (2011), Simulation and hardware 8998
implementation of incremental conductance MPPT with Xiao Weidong, Ozog Nathan, Dunford William G. (2007),
direct control method using cuk converter, IEEE Trans. Topology Study of Photovoltaic Interface for Maximum
Ind. Electron., 2011, 58, pp. 11541161 Power Point Tracking, IEEE transactions on industrial
Salas, V., E. Olas, A. Barrado and A. Lzaro, (2006), electronics, vol. 54, no. 3, june 2007, 1696-1704.
Review of the maximum power point tracking Yu G.J., Jung Y.S., Choi J.Y., Kim G.S., (2004), A novel
algorithms for stand-alone photovoltaic systems. Solar two-mode MPPT control algorithm based on
Energy Materials and Solar Cells, 90, 1555-1578. comparative study of existing algorithms, Sol. Energy 76
Slotine J-J.E., Li W., (1991), Applied Nonlinear Control, 455-463.
Prentice-Hall, Englewood Cliffs, New Jersey 07632,
p.459.