Online ISSN 1848-3380, Print ISSN 0005-1144

ATKAFF 55(1), 22–31(2014)

Aranzazu D. Martin, Jesus R. Vazquez

Backstepping Controller Design to Track Maximum Power in

Photovoltaic Systems
DOI 10.7305/automatika.2014.01.289
UDK 681.511.4-531.9:621.314.1; 621.311.243
IFAC 2.3.2; 3.1
Original scientific paper

This work presents a new control method to track the maximum power point of a grid-connected photovoltaic
(PV) system. A backstepping controller is designed to be applied to a buck-boost DC-DC converter in order to
achieve an optimal PV array output voltage. This nonlinear control is based on Lyapunov functions assuring the
local stability of the system. Control reference voltages are initially estimated by a regression plane, avoiding local
maximum and adjusted with a modified perturb and observe method (P&O). Thus, the maximum power extraction
of the generating system is guaranteed. Finally, a DC-AC converter is controlled to supply AC current in the point of
common coupling (PCC) of the electrical network. The performance of the developed system has been analyzed by
means a simulation platform in Matlab/Simulink helped by SymPowerSystem Blockset. Results testify the validity
of the designed control method.
Key words: Backstepping, Buck-boost converter, Maximum power point (MPP), Photovoltaic system

Sinteza backstepping regulatora za praćenje maksimalne proizvodnje energije u fotonaponskim susta-

vima. Ovaj rad predstavlja novu metodu upravljanja za sliježenje točke maksimalne snage fotonaponskog (PV)
sustava. Dana je sinteza backstepping regulatora za primjenu u silazno-uzlaznom DC-DC pretvaraču za postizanje
optimalnog izlaznog napona PV-a. Ova je nelinearna metoda upravljanja zasnovana na Ljapunovim funkcijama
osiguravajući tako lokalnu stabilnost sustava. Upravljačke reference napona prvo su estimirane korištenjem regre-
sijske ravnine izbjegavajući lokalne maksimume, a zatim podešene tzv. modificiranom perturbiraj i uoči metodom
(P&O). Prema tome, zagarantirano je maksimalno izvlačenje energije iz sustava proizvodnje. Naposlijetku, DC-AC
pretvaračem upravlja se na način da osigurava željena izmjenična struja u točki zajedničkog spoja (PCC) elek-
troenergetske mreže. Ponašanje razvijenog sustava analizirano je kroz simulacije provedene u Matlab/Simulink
okruženju uz korištenje SymPowerSystem biblioteke.

Ključne riječi: Backstepping, silazno-uzlazni pretvarač, točka maksimalne proizvodnje energije (MPP), fotona-
ponski sustav

1 INTRODUCTION

Nowadays, renewable energy sources are widely used

and, particularly, photovoltaic (PV) energy systems have
become widespread everywhere. The grid-connected PV
systems consist of an array of solar modules, a DC-DC
power converter, a DC-AC power converter and a control
system. The complete scheme is presented in Fig. 1.
The PV modules convert the solar energy to electrical
energy so as to be transferred to the electrical power sys-
tem [1-2]. The generated power depends on the environ-
ment conditions, such as temperature and solar irradiation.
Therefore, there is only one operating point with a Max- Fig. 1. Structure of the grid-connected PV system
imum Power Point (MPP) under particular conditions by
each PV module.

22 AUTOMATIKA 55(2014) 1, 22–31

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

A DC-DC power converter, in this case a buck- 2 PV SYSTEM MODEL

boost [3], which converts the DC power from one volt-
The model of the photovoltaic system is shown in this
age level to another higher or lower to the input voltage,
section. Firstly, it will be presented the model of the solar
has to be added at the output of the photovoltaic array to
cell, secondly, the DC-DC power converter modeling and,
achieve the optimum voltage and to implement the Maxi-
finally, the inverter model.
mum Power Point Tracking (MPPT).
The MPP can be tracked through different MPPT al- 2.1 Model of the PV array
gorithms that control the switching converter in order to
obtain the maximum power under all conditions [4-15]. The solar cell turns the light into electrical energy. Its
There are various methods, some of them are based on the well-known model consists of a current source Il that rep-
well-known principle of perturb and observe (P&O) [7-8], resents the current generated by the photons, (it will be
on sliding mode control method [9-10], Ripple Correlation constant if the radiation and the temperature are constants
Control (RCC) [11], artificial neuronal networks or fuzzy too), in A, an anti-parallel diode D1 , a shunt electrical re-
based algorithms [12-14], amongst others. sistance Rsh , in Ω, which represents the current leakage,
and a series resistance Rs , in Ω, which models the ohmic
In the first method mentioned above, P&O, the output
losses. The equivalent circuit is shown in Fig. 2.
power gets the equilibrium point but it has an oscillatory
behavior and that point is not always achieved, obtaining a
local maximum instead of a global maximum. The artificial
neuronal network has a better performance but presents an
involved structure, whereas fuzzy logic control does not
need a mathematical model. The methods have different
accuracy and complexity.
A new control method for MPPT is proposed in this pa-
per. A nonlinear backstepping controller [16-18] in a buck-
boost converter has been designed to track the maximum
power point with the help of a regression plane, which pro- Fig. 2. Equivalent circuit of a PV cell
vides the PV array output reference voltage helped by a
modified P&O method to achieve the MPP faster [19-20]. The equation that relates I, the solar cell output current
This case, the duty ratio of the switching converter of the in A, to V , the solar cell output voltage, in V , is shown in
buck-boost converter is controlled. This way, the robust- (1):
ness is increased, global asymptotic stability is guaranteed  (V +Rs I)
q
 V+R I
s
by means of Lyapunov and the MPP can be assured even I = Il − Io e nKTK − 1 − , (1)
Rsh
with changeable conditions. A nonlinear control has been
chosen due to the nonlinear, time variant nature and vari- where Il is the light generated current, in A, Io is the cell
able structure of the buck-boost, thus a linear control im- reverse saturation current, in A, q is the charge of an elec-
plies a model linearization that is simple but it cannot con- tron, in C, n is the ideal factor (dimensionless), K is the
trol the converter in a wide range. Boltzmann’s constant, in J/K and TK is the working tem-
In addition, the control of the DC-AC power inverter of perature of the cell, in K [10].
the PV system has been designed to inject electrical power Therefore, the solar cells are connected in series and in
to the electrical network by means of a PI control [21-22]. parallel in order to create a solar module depending on the
So, the global control makes possible extract the maximum capacity demands. The array used in this work has been
power of the PV system, inject active power and regulate modeled in a real installation and it has sixteen modules
the input voltage of the DC-AC power inverter. that consist of 72 solar cells in series and one solar cell in
The paper is organized as follows. Section 2 presents parallel. There are four modules shunt connected and four
the PV system model, describing the model of the PV array modules in parallel. The voltage will be obtained multi-
and the buck-boost converter. The proposed design of the plying the voltage of one solar cell by the number of cells
control to make the system track the maximum power point connected in series whereas the current will be calculated
is developed in Section 3, where the regression plane and multiplying the current of a solar cell by the number of
the backstepping controller are explained, this section also cells that are shunt connected.
includes the design of the control of the inverter. Section 4 Under standard conditions, 1000 W/m2 and 25 ◦ C, the
presents the simulation results and Section 5 describes the electrical parameters of the solar system are shown in Ta-
main conclusions. ble 1.

AUTOMATIKA 55(2014) 1, 22–31 23

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

Table 1. Electrical parameters of the solar module

Parameter Values
Maximum power 1555 W
Maximum power voltage 102.6 V
Maximum power current 15.16 A
Open-circuit voltage 165.8 V
Short-circuit current 17.56 A

There is only one operating point for a PV array with a

maximum output power that changes with the cell temper- Fig. 3. Buck-boost power converter
ature, T in ◦ C, and the solar radiation, G in W/m2 . The
MPPs with different irradiations and a cell temperature of
25 ◦ C are presented in Table 2. this case, the load is a resistance. Parameters R, L, C1 and
C (resistance in Ω, inductor in H and capacitors in F) are
Table 2. Maximum power points to 25 ◦ C and changeable constants.
irradiance The duty cycle of the converter is D = tON /tC , being
Radiation (W/m2 ) Maximum Power (W) tON the time which the switch is ON and tC is the switch-
200 383 ing period (0 < D < 1).
400 748 When the switch is ON, the PV array supplies energy
600 1066 to the inductor and the diode is inversely polarized and the
800 1332 charge is insulated from the source. Thus, the supplied en-
1000 1555 ergy is stored in the inductor. When the switch is OFF,
1200 1733 the diode conducts and the stored energy in the inductor
is transferred to the charge.
If the irradiance and temperature of the whole PV array The converter can work in two modes, Continuous
is homogeneous, the power curve of the system has an only Conduction Mode (CCM) and Discontinuous Conduction
maximum. In partial shading conditions, the P-V curve Mode (DCM), depending on the current of the inductor in
has several local maximum but one global maximum point the operation period. In CCM, the current of the inductor
[23]. In sections 3.1 and 3.2, the proposed control will is never zero, the current fluctuates between maximum and
be detailed in order to prove that the global maximum is minimum values based on the time that the switch is ON.
achieved helped by a plane regression that places the sys- In DCM, when the switch is OFF there is a time which
tem close to the maximum point. the current of the inductor is zero because tOF F , the time
which the switch is OFF, is greater than the time that the
2.2 Model of the buck-boost converter inductor can transfer energy. Thus, the inductor is com-
Figure 3 shows the topology of the buck-boost power pletely discharged.
converter used in this work. It consists of power electronic In this work, the DC-DC converter will work in CCM.
components such as capacitor and inductor elements con- In order to guarantee this mode, the minimum value of the
nected as is shown in the Fig. 3. In closed loop, the function inductor and the capacitor must be calculated as is shown
of the control is to regulate the voltage of the solar mod- in (2) and (3) [25]:
ules, in order to achieve the maximum power, by means of
(1 − Dmin )2 vo
the control of the duty cycle, D. Lmin = , (2)
2Io min fs
The ON/OFF commutation of the switch, controlled by Io max Dmax
means of Pulse Width Modulation (PWM) principle, all- Cmin = , (3)
∆vo fs
ows the charge and discharge of energy of the storage el-
ements, getting an output voltage higher or lower to the where ∆vo is the incremental output voltage of the buck-
input voltage. The transistor T1 and the diode D2 imply a boost converter, Dmin and Dmax are the minimum and
non-linear behavior of the converter. maximum duty cycle respectively, Io min and Io max are the
The output voltage of the DC-DC power converter is vo minimum and maximum output current of the DC-DC con-
in V, vP V is the input voltage of the converter or the output verter and fs is the frequency.
voltage of the solar array in V, iP V is the output current of Thus, the system ranges from short-circuit voltage to
the PV array, in A, and iL is the inductor current in A. In open-circuit voltage in order to obtain the I-V character-

24 AUTOMATIKA 55(2014) 1, 22–31

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

istics of the solar modules. The buck-boost converter pro- 3 CONTROL DESIGN
vides an output impedance that changes from zero to infi-
nite. The main objective of the proposed control is to adjust
the PV array output reference voltage, given by a regres-
The output voltage of the buck-boost converter has op-
sion plane, using a backstepping controller to regulate the
posite polarity from the input voltage. Besides, the out-
DC-DC input voltage and to guarantee the maximum en-
put/input ratio can be determined as it is shown in (4):
ergy extraction of the PV modules. Thereby, the local max-
vo D imum is avoided adjusting the reference voltage instead of
= . (4)
vP V 1−D the duty cycle in P&O method.
Using the state averaging method [24], the model can Then, the optimum voltage must be obtained modifying
be defined as are shown in (5), (6) and (7): the voltage around the reference or theoretical voltage ac-
cording to a backstepping controller instead of a standard
1 1 PI control. This control is required to achieve the voltage
v̇P V = iP V − iL D, (5)
C1 C1 that supplies the maximum power at the input of the con-
vo vP V − vo verter.
i̇L = + D, (6)
L L
iL vo iL A control strategy for the inverter has been proposed to
v̇o = − − + D. (7) supply active power to the electrical system.
C RC C
Duty cycle must be controlled so as to get the voltage 3.1 Maximum power reference voltage
with which it is obtained the maximum energy.
A regression plane gives the reference or theoretical
2.3 Model of the inverter voltage to achieve the MPP under any conditions of tem-
The DC-AC power inverter connects the PV system to perature and solar radiation. This way, the MPPT block
the electrical network. Its aim is to provide the maximum reaches the value of the voltage that supplies the maximum
active power to the network. power, and the peak of the power curve is achieved ini-
A three-phase inverter has been used in this work. It tially faster. Even if there are several local maximum (for
consists of a control circuit to calculate the reference cu- example, a partial shading case), the proposed control ap-
rrent and a power circuit to follow this reference and to proaches initially to the global maximum and it does not
inject the final current to the network. The power circuit stay at a local maximum.
includes six insulated-gate bipolar transistors (IGBT) de- Firstly, the characteristic curves of the system are ob-
vices with anti-parallel diodes to get the desired AC output tained. Secondly, the regression plane is calculated for a
with a specific sequence of commutation of the transistors. variety of cell temperatures and solar radiation. The regres-
Two capacitors have been included in order to regulate the sion plane is shown in Fig. 5.
voltage. The DC-AC topology is shown in Fig. 4. The out-
put of the DC-DC power converter will be connected with
the DC side.

Fig. 4. DC-AC power converter

The DC-AC outputs are connected to the network

through output reactance to establish a pulse-width mod-
ulation (PWM) current control. In this work, a hysteresis- Fig. 5. MPP Reference voltage depending on irradiance
band technique has been used [26]. and temperature

AUTOMATIKA 55(2014) 1, 22–31 25

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

Temperature T ranges from 0 ◦ C to 80 ◦ C whereas the Next step must study the behavior of the current error,
irradiance ranges from 200 W/m2 to 1000 W/m2 . A volt- z1 = iL − α1 , where the inductor current should reach
age matrix is obtained for the maximum power depending α1 to make the error vanish in order to achieve the control
on the environmental conditions. Thus, the reference volt- aim. The time derivative of this error is shown in (13):
age is calculated by linear interpolation and the backstep-
ping controller will take into account the reference voltage ż1 = i̇L − α̇1 . (13)
supplies by the plane regression.
Now, the time derivative of α1 , (12), replacing iL by
3.2 Buck-boost backstepping controller z1 + α1 , yields (14):
The aim of this control is to extract the maximum C1 k12 C1 r 1 Ḋ
power. In order to do this, a non-linear backstepping con- α̇1 = k1 z1 − e1 − v̈ + i̇P V − α1 . (14)
D D PV D D
troller is designed to control the duty cycle of the switch
of the buck-boost converter. This way, the output voltage Equation (14) with (6) gives the time derivative of z1 ,
of the PV modules can be regulated to track the reference as is shown in (15):
voltage.
vo vP V − vo
The backstepping control is used to design stable con- ż1 = + D− (15)
L L !
trols with a recursive methodology. It must stabilize the
origin of a system by means of feedback control laws and C1 k12 C1 r 1 Ḋ
k1 z1 − e1 − v̈ + i̇P V − α1 .
using Lyapunov functions to prove the stability of the sys- D D PV D D
tem.
In this case, next steps should be followed to design the Similarly to which it is done in the Lyapunov function
controller. V1 , another Lyapunov function should be defined with the
same characteristics, being as (16):
Firstly, the voltage error is defined as is shown in (8):
1 1 1
e1 = vP V − vPr V , (8) V2 = V1 + z12 = e21 + z12 . (16)
2 2 2
where vPr V is the reference output voltage of the PV mod- And its time derivative is (17), accounting for (9) and
ules and it must be reached by the control in order to en- (15), and replacing iL by z1 + α1 :
force vP V to its appropriate value. Deriving e1 with respect 
to time and accounting for (5), (9) is obtained: vo vP V − vo
V̇ 2 = −k1 e21 + z1 + D+
L L
1 1  
ė1 = v̇P V − v̇Pr V = iP V − iL D − v̇Pr V . (9) C1 k12 D C1 r
C1 C1 e1 − − k1 z1 + v̈ − (17)
D C1 D PV

In (9), iL behaves as a virtual control input. Now, a 1 Ḋ
Lyapunov function is selected. It must be positive definite i̇P V + α1 = −k1 e21 − k2 z12 .
D D
and radially unbounded for all x, and the time derivative
of Lyapunov function must be negative definite for all x Equation (17) will be negative when k2 is positive, be-
to assure the solution is locally asymptotically stable. The ing a constant, in order to ensure the stability of the system.
chosen function and its derivative are defined as: Therefore, the term between square brackets must be zero,
1 2 thus working out the value of the controller Ḋ, yields (18):
V1 = e , (10)  
2 1   1 h vo v P V − vo 2 D2
1 1 Ḋ = − D− D − e1 C1 k12 − +
V̇1 = e1 ė1 = e1 iP V − iL D − v̇Pr V = −k1 e1 . α1 L L C1
C1 C1 i
(11) z1 (k1 − k2 )D − C1 v̈Pr V + i̇P V , (18)

V̇1 will be negative if k1 is constant and positive. This way, where 0 < D < 1 and α1 6= 0.
the reference current for the control, α1 , so-called stabi- In order to prove that the controller is locally asymptot-
lization function, can be obtained working out the value of ically stable in the equilibrium point, a study of the stability
the iL from (11): will be carried out.
1 Firstly, the errors should be defined taking into account
α1 = (C1 k1 e1 + iP V − C1 v̇Pr V ) . (12) the references, as follows in (19), the error of the voltage,
D

26 AUTOMATIKA 55(2014) 1, 22–31

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

(20), the error of the current and, (21), the error of the con- reference voltage instead of the duty cycle value. The ad-
troller: dition of theoretical reference voltage and the incremental
reference voltage gives the final reference voltage used in
e1 = vP V − vPr V , (19) backstepping control. In order to implement this method, it
e2 = iL − irL , (20) is necessary to take into account vref,f , iP V , vP V , vo and
e3 = D − D . r
(21) iL . Hence, equation (18) is implemented so as to get the
duty cycle that controls the DC-DC converter.
This way, the time derivatives of the errors are (22),
(23) and (24), replacing vP V , iL and D by e1 +vPr V , e2 +irL
and e3 + Dr respectively, and using (5) and (6):
1 1
ė1 = iP V − (e2 + irL )(e3 + Dr ), (22)
C1 C1
vo e1 + vPr V − vo
ė2 = + (e3 + Dr ), (23)
L L
1 h vo
ė3 = − (e3 + Dr )−
α1 L
e1 + vPr V − vo Fig. 6. Control scheme of the DC-DC converter
(e3 + Dr )2 −
 L 
2 (e3 + Dr )2 3.3 Inverter control strategy
e1 C 1 k1 − +
C1 The designed control strategy supplies electrical power
i
z1 (k1 − k2 )(e3 + D ) −r
C1 v̈Pr V + i̇P V . (24) to the electrical system. This method has been proposed
because it is effective, simple and easy to implement with
a reduced computation time.
It must be known that the last two terms are zero be-
cause vPr V and iP V are constants. To obtain locally asymp- The PV power system injects AC current to the elec-
totically stability, the errors and its derivatives will be zero trical network, iAC in A. In this work, the aim of the in-
in the equilibrium point. Thus, the references values can be verter control is to inject current co-linear with the voltage
calculated as in (25), (26) and (27): in point of common coupling. Thus, the current reference
is shown in (28):
−vo + vo Dr
vPr V = , (25) iAC,ref = K vAC , (28)
Dr
iP V where K is an adjustment parameter and it can be adjusted
irL = r , (26)
D by a PI control, Fig. 7. and vAC is the PCC alternating volt-
vo
r
D = . (27) age. If the error between iAC,ref and the inverter output
vo − vPr V current, iAC is bigger than a hysteresis band [26], the state
of the appropriate semiconductors will change to follow
In (27), v0 < 0 and vPr V > 0, thus 0 < Dr < 1. Next the reference. The frequency of the switching depends on
step is to get the Jacobian in the equilibrium point, where current changes from the upper to lower limit or vice versa.
the errors are zero and vPr V is the reference voltage at the Equation (28) assures the power grid synchronization be-
input of the DC-DC converter, so as to achieve the eigen- cause the output current of the system and the voltage of
values to ensure the locally stability. Hence, the obtained the common connection point are in phase with each other.
reference values assure that the real part of the Jacobian
Note that the reference voltage at the input inverter is
eigenvalues is negative, thus the locally asymptotically sta-
set at 600 V, greater than the grid voltage so as to guarantee
bility is guaranteed.
correct working of the inverter [27]. Hence, the aim of the
In Section 4, a numerical example will be presented to control is to keep constant this voltage. This way, on the
check the proposed control. one hand, if the voltage of the PV system is greater than
Figure 6 shows the DC-DC converter control scheme. reference voltage, the energy of the solar modules is stored
On the one hand, with the regression plane and T and G in the capacitor and it is not injected to the AC system. On
values, the theoretical reference voltage is obtained. On the the other hand, if the voltage of the PV system is lower
other hand, taking into account iP V and vP V , a modified than 600 V, the load requires more power than the PV sys-
P&O method is applied. This proposed method is based on tem can supply, thus the source must supply the rest of the
the traditional P&O, but it provides an incremental value of active power that it is necessary.

AUTOMATIKA 55(2014) 1, 22–31 27

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

k1 and k2 are the parameters used in backstepping

method, in (18), L is the inductor of the DC-DC converter,
C1 and C are the capacitor at the input of the buck-boost
converter and the capacitor of the DC-DC converter re-
spectively whereas R is the load resistor at the AC side.
The system has been simulated for changeable envi-
ronmental conditions. Thus, a change of the solar radia-
tion of the whole system from 600 W/m2 to 1000 W/m2
Fig. 7. Control scheme of the DC-AC converter at 0.3 s was considered, as it is shown in Fig. 9.a, with
a temperature of 25 ◦ C, to prove that the MPP is always
achieved and to check the transient performance of the de-
The inverter control strategy is shown in Fig. 7. The
signed controller. This way, the robustness of the system
vref is 600 V, vo is the output voltage of the DC-DC con-
with changes of solar radiation is evaluated.
verter and u is the signal control of the inverter.
In Fig. 9.b and Fig. 9.c, the voltage and the power are
presented with their references values respectively where
4 RESULTS
it is shown the controller tracking behavior. The reference
A practical case has been simulated in Mat- voltage is reached by the voltage that supplies the PV array,
lab/Simulink, using SimPowerSystem blockset, to validate whereas the maximum power is achieved as well.
the proposed control. The PV system is connected to a
radial electrical network with a three-phase source and a
resistive load, Fig. 8. The PV array block models sixteen
PV modules according to Fig. 6. The DC/DC block, the
DC/AC block (without capacitors), the backstepping con-
trol block and the inverter control block were shown in fig-
ures 3, 4, 6 and 7 respectively.

Fig. 8. Simulink block diagram of the system Fig. 9. Results of backstepping controller (a) irradiance,
(b) DC voltage, (c) PV power
The simulation parameters and other element values
are presented in Table 3. According to the Table 2, the maximum power is
1066 W when the irradiance is 600 W/m2 and the temper-
ature is 25 ◦ C, and the maximum power for 1000 W/m2
Table 3. Simulation parameters
and 25 ◦ C is 1555 W. The reached powers are 1047 W and
Parameter Values
1531 W respectively, instead of the maximum power due
k1 60
to the losses in the buck-boost converter. However, there
k2 9000
are only slight losses and the performance of the system is
L 20e-3 H
98.46%, quite satisfactory. Besides, the transient response
C 48e-6 F of the system is satisfactory, because it is stabilized after
C1 1e-3 F a smooth transient response in less than two periods. This
R 50 Ω way, the validation of the system using backstepping con-

28 AUTOMATIKA 55(2014) 1, 22–31

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

troller is guaranteed and the maximum power point is al-

ways achieved with low losses.
Moreover, as it is shown in Fig. 10.a, the inverter con-
trol detailed in Fig. 7 allows to obtain a constant voltage
at the input of the DC/AC converter even though there are
changeable environmental conditions; in the studied case,
the maximum variation of voltage is 0.5% in transient pe-
riod, from 0.3 s to 0.317 s. Besides, figures 10.b and 10.c
show the voltage and the current at the output of the in-
verter. The alternating voltage at the output of the DC/AC
inverter, Fig. 10.b, is stable without transient perturbances
whereas the current at the output of the inverter, Fig. 10.c,
depends on the supplied power of the PV system.

Fig. 11. P-V curve with multiple maximum

Fig. 10. Simulation results of the inverter control. (a) in- Fig. 12. Output power with multiple maximum
verter input DC voltage, (b) inverter output AC voltage,
vAC , (c) inverter output AC current, iAC
Fig. 12.
In the previous case, a homogeneous irradiance of The maximum achieved is 1256 W and the local maxi-
the whole system has been considered. Thus, the PV mum is avoided.
curve shows an only peak. The proposed control allows to
achieve the global maximum power point even if there are 5 CONCLUSION
more than a peak. A partial shading case has been simu- In this work, a novel backstepping controller to track
lated in order to prove the correct working of the proposed the MPP of a PV system has been designed. This non-
controller when there are several peaks in the P-V curves. linear controller has been applied to a buck-boost converter
This case, there are two shaded modules in series and to achieve the PV array output voltage which obtains the
the rest of the modules are not shaded. The irradiance and maximum power for all environmental conditions, check-
temperature in shaded modules are 500 W/m2 and 25 ◦ C ing the robustness for different values of solar radiation.
respectively whereas the environmental conditions in not The control reference voltage has been obtained with the
shaded modules are 1000 W/m2 and 25 ◦ C. In these con- help of a regression plane. Furthermore, the global asymp-
ditions, the P-V curve has two maximums, but the global totic stability of the controller has been proved.
maximum is 1302 W with 110 V, Fig. 11. The simulation results show the performance of the PV
Therefore, the power at the output of the DC-DC con- system is 98.46% and the transient response is about 0.04 s
verter is close to the global maximum, as it is shown in with sudden changes in environmental conditions.

AUTOMATIKA 55(2014) 1, 22–31 29

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

The output power has been injected to an AC electri- [6] Doron Shmilovitz, Yoash Levron, “Distributed Maximum
cal system through a DC/AC converter. The control of this Power Point Tracking in Photovoltaic Systems – Emerging
inverter has been validated in the simulation platform. The Architectures and Control Methods”, Automatika, vol. 53,
inverter input voltage remains constant and the output volt- no. 2, pp. 142-155, 2012.
age is stationary for all conditions of the system. The in- [7] Villalva, M.G.; Gazoli, J.R.; Filho, E.R., “Analysis and sim-
jected currents depend on the available output power. Be- ulation of the P&O MPPT algorithm using a linearized PV
sides, the proposed method has been validated with two array model”, Power Electronics Conference, 2009. COBEP
peaks power curve to check that the local maximum is ’09. Brazilian, pp. 189-195, 2009.
avoided. [8] Elgendy, M. A.; Zahawi, B.; Atkinson, D. J., “Evaluation
To sum up, it has been proved the usefulness and the ef- of Perturbe and Observe MPPT Algorithm Implementation
fectiveness of the backstepping method to control a buck- Techniques”, 6th IET International Conference on Power
boost converter to extract the maximum power of a PV sys- Electronics, Machines and Drives (PEMD 2012), pp. 1-6,
2012.
tem, contrasting the results in a numerical simulation.
A possible extension of this work could be to improve [9] Miao, Z., W. Jie, et al., “The application of slide technology
the control of the DC-AC power inverter in order to add in PV maximum power point tracking system”, Fifth World
Congress on Intelligent Control and Automation, WCICA,
new functions to the PV system, such as electrical system
2004.
conditioning. The inverter could inject reactive power and
harmonic current to compensate the non-linear loads of the [10] M. I. Arteaga Orozco, J. R. Vázquez, P. Salmerón, “MPP
electrical system. This way, the Power Quality (PQ) of the Tracker of a PV System using Sliding Mode Control with
Minimum Transient Response”, International Review on
electrical network would be improved.
Modelling and Simulations (I.RE.MO.S.), December 2010.
Next research step will be to develop an experimental
platform where the whole system could be proved. This [11] T. Esram, J.W. Kimball, P.T. Krein, P.L. Chapman, P.
Midya, “Dynamic maximum power point tracking of photo-
paper has been made with this final objective, for this the
voltaic arrays using ripple correlation control”, IEEE Trans-
PV array model is a grid-connected real system available
actions on Power Electronics, vol. 21, No. 5, pp. 1282-1291,
by the authors in the research group installation. September 2006.

ACKNOWLEDGMENT [12] Ramaprabha, R.; Mathur, B.L.; Sharanya, M., “Solar Array
Modeling and Simulation of MPPT using Neural Network”,
This work was supported by the MEC (Ministerio de International Conference on Control, Automation, Commu-
Educación of Spain) under FPU grant (Formación de Pro- nication and Energy Conservation, pp. 1-5, 2009.
fesorado Universitario). [13] Syafaruddin, Karatepe, E., Hiyama, T., “Artificial neural
The authors are with the Electrical Engineering network-polar coordinated fuzzy controller based maximum
Department at Escuela Técnica Superior de Inge- power point tracking control under partially shaded condi-
niería in University of Huelva, Spain (e-mail: aran- tions”, IET Renewable Power Generation, vol. 3, no. 2, pp.
zazu.delgado@die.uhu.es; vazquez@uhu.es). 239-253, June 2009.
[14] Theodoros L. Kottas, Yiannis S. Boutalis and Athanassios
REFERENCES D. Karlis, “New Maximum Power Point Tracker for PV
[1] M. Buresch, “Photovoltaic Energy Systems Design and In- Arrays Using Fuzzy Controller in Close Cooperation With
stallation”, McGraw-Hill, New York, 1983. Fuzzy Cognitive Networks”, IEEE Transactions on Energy
[2] I. H. Altas1 and A.M. Sharaf, “A Photovoltaic Array Simu- Conversion, vol. 21, no. 3, September 2006
lation Model for Matlab-Simulink GUI Environment”, IEEE [15] Giovanni Petrone, Giovanni Spagnuolo, Massimo Vitelli,
International Conference on Clean Electrical Power, ICCEP “Distributed Maximum Power Point Tracking: Challenges
’07, pp. 341-345. and Commercial Solutions”, Automatika, vol. 53, no. 2, pp.
[3] Tasi-Fu Wu; Yu-Kai Chen, “Modeling PWM DC/DC Con- 128-141, 2012.
verters out of Basic Converter Units”, IEEE Transactions on [16] Youping Zhang, Barıs¸ Fidan, Petros A. Ioannou, “Back-
Power Electronics, vol. 13, no. 5, pp. 870-881, 1998. stepping Control of Linear Time-Varying Systems with
[4] A. Daoud, A. Midoun, “Simulation and Experimental Study Known and Unknown Parameters”, IEEE Transactions on
of Maximum Power Point Tracker Based on a DC/DC Buck Automatic Control, vol. 48, no. 11, November 2003.
Converter”, International Review of Electrical Engineering, [17] El Fadil, H.; Girt, F., “Backstepping Based Control of PWM
vo. 5. No. 2, Part A, pp. 514-520, April 2010. DC-DC Boost Power COnverters”, Symposium on Industrial
[5] Tirshan Esram and Patrick L. Chapman, “Comparison of Electronics, ISIE 2007, pp. 395-400, 2007.
photovoltaic array maximum power point tracking tech-
niques”, IEEE Transactions on Energy Conversion, vol. 22,
no. 2, June 2007.

30 AUTOMATIKA 55(2014) 1, 22–31

Backstepping Controller Design to Track Maximum Power in Photovoltaic Systems A. D. Martin, J. R. Vazquez

Aránzazu D. Martín was born in Bollullos del

[18] Fan Liping; Yu Yazhou; Boshnakov, K., “Adaptative Back-
Condado, Huelva, Spain, on December 21, 1983.
stepping Based Terminal Sliding Mode Control for DC-DC She received the degree in Industrial Engineering
Convertor”, International Conference on Computer Applica- from the University of Huelva, Spain, in 2008.
tion and System Modeling (ICCASM 2010), pp. V9-323-V9- She did a master in Control Engineering, Elec-
327, 2010. tronic Systems and Industrial Computer Science
in 2011. Since 2008, she has been collaborat-
[19] Michael Sokolov and Doron Shmilovitz, “A Modified ing with the Electrical Engineering Department
MPPT Scheme for Accelerated Convergence”, IEEE Trans. at University of Huelva. She is currently work-
On Energy Conversion, vol. 23, no. 4, December 2008, pp. ing as a researcher with the FPU grant. Now, her
1105-1107. research interests include renewable energy, dis-
tributed electrical generation, power quality, active power filter and fuzzy
[20] Sachin Jain and Vivek Agarwal, “A New Algorithm for
logic.
Rapid Tracking of Approximate Maximum Power Point in
Photovoltaic Systems”, IEEE Power Electronics Letters, vol.
2, no. 1, March 2004, pp. 16-19.
Jesús R. Vázquez was born in Huelva, Spain,
[21] Minaei, S.; Yuce, E.; Tokat, S.; Cicekoglu, O., “Simple Re- on December 24, 1967. He received the degree
alizations of Current-Mode and Voltage-Mode PID, PI and in electrical engineering from the University of
PD Controllers”, Proceedings of the IEEE Internationsl Sym- Seville, Spain, in 1995. He obtained the Ph.D.
posium on Industrial Electronics, ISIE, vol. 1, pp. 195-198, degree in 2004. For one year, he was with the
2005. electrical department of Nissan Motor Ibérica
S.A., Barcelona, Spain. Since 1996, he is with the
[22] Georgios Tsengenes, Georgios Adamidis, “Comparative Electrical Engineering Department at the Escuela
Evaluation of Fuzzy-PI and PI Control Methods for a Three Técnica Superior de Ingeniería of the University
Phase Grid Connected Inverter”, Proceedings of the 2011- of Huelva. He teaches Electric Circuits, Electri-
14th European Conference on Power Electronics and Appli- cal Power Quality and Photovoltaic Systems and
cations (EPE 2011), 2011. his research interests include power quality, active power filters, renew-
able energy, distributed electrical generation and artificial network appli-
[23] Paraskevadaki, E.V.; Papathanassiou, S.A., “Evaluation of cations.
MPP Voltage and Power of mc-Si PV Modules in Partial
Shading Conditions”, IEEE Transactions on Energy Conver-
sion, vol. 26, no. 3, pp. 923-932, 2011. AUTHORS’ ADDRESSES
[24] R. Ortega, A. Lorai, P. J. Niklasson and H. Sira-Ramirez, Aránzazu D. Martín, M.Sc.
“Passivity-based Control of Euler-Lagrange Systems”, Lon- Prof. Jesús R. Vázquez, Ph.D.
don: Springer-Verlag, pp. 168-171, 1998. Electrical Engineering Department,
[25] Rashid Muhammad H., “DC DC Convertes” in Power Elec- Escuela Técnica Superior de Ingeniería,
tronics Handbook, Ed. Academic Press, 2001, pp. 216-224. University of Huelva,
[26] Devaraj, D.; Sakthivel, S.; Punitha, K., “Modeling of Photo- Ctra. Palos de la Frontera s/n, Palos de la Frontera,
voltaic Array and Simulation of Adaptive Hysteresis Current ES-21819, Huelva, Spain
Controlled Inverter for Solar Application”, 3rd International email: aranzazu.delgado@die.uhu.es, vazquez@uhu.es
Conference on Electronics Computer Technology (ICECT),
vol. 6, pp. 302-306, 2011. Received: 2012-06-04
[27] P. Salmerón, J.; J. R.Vázquez; “Active Power Line Condi- Accepted: 2012-10-24
tioners” in A. Moreno, Power Quality, Mitigation Technolo-
gies in a Distributed Environment, Ed. Springer, London,
2007, pp. 231-291 19.

AUTOMATIKA 55(2014) 1, 22–31 31