[Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

Active Disturbance Rejection Control Applied

to a Three-Phase Grid-Connected Photovoltaic
System
Ali BOUKHRISS
SAEDD, High School of Technology of Essaouira, Caddy Ayad University Morocco
a.boukhriss@uca.ma

Abstract—The objective of this work is to propose an active In addition to conventional MPPT methods, soft
disturbance rejection control (ADRC) for a two-stage grid- computing techniques are increasingly used in PV systems,
connected photovoltaic (PV) array. ADRC combined with such as fuzzy logic control FLC, artificial neural network
incremental conductance (ADRC_IncCond) is used to extract
ANN, genetic algorithm GA etc.
the maximum power from the PV array and compared to the
fuzzy logic control (FLC) and the perturb and observe (P&O) FLC is very robust, it does not require accurate
algorithm. A two-stage inverter is also controlled by ADRC to mathematical models. It has a fast response and less
keep the DC bus voltage constant and provide a unity power oscillation under changing atmospheric conditions [15-18].
factor to the power supplied to the grid. The performance of However, it remains complex and its effectiveness depends
the control is evaluated, under varying atmospheric conditions, on the performance of the designer and the accuracy of the
by simulations in the Matlab Simulink environment.
rules.
Index Terms—fuzzy logic control, maximum power tracking,
Simulations show that the ANN variable step-size has a
photovoltaic system, reactive power control, voltage control. fast tracking speed, less oscillation and good efficiency
compared to the fixed ANN step-size which has the same
I. INTRODUCTION drawbacks as the P&O technique [19-20], nevertheless, it
remains very complex and expensive and requires a lot of
The number of new renewable energy plants, namely
information on PV parameters.
wind and solar photovoltaic, is increasing worldwide. The
In addition of the MPPT control, the inverter must be
global trend is to reduce the impact of polluting fossil fuels
controlled. In the literature, PI controllers are widely used
[1-2]. Despite their intermittent nature, the diversification of
due to their simple design. However, they have limitations
sites and the nature of the production process used: solar or
and weaknesses with respect to internal system disturbances.
wind, these production plants can participate in the
To avoid problems related to non-linearities and internal
secondary regulation of the network frequency, or even in
disturbances in systems, non-linear control methods are
the primary regulation, thanks to the reliability of current
widely used such as sliding mode control, back-stepping
weather forecasts [3-4].
control etc. [21-22].
In this work, a grid-connected photovoltaic system is
In this work, active disturbance rejection control ADRC
studied. Two configurations are widely used. The first one
will be applied. This technique allows real-time rejection of
has a single stage, i.e. the PV panel is directly connected to
disturbances, whether internal or external, and does not
the grid via an inverter [5-6], while the second one has an
require detailed knowledge of the mathematical model of
intermediate stage (DC/DC converter) between the PV panel
the system. Its linear form (linear active disturbance
and the inverter. The two-stage system keeps the DC bus
rejection control LADRC) is simple in design and
voltage constant at the input of the inverter, thanks to the
application for system control [23-25].
MPPT control of the boost converter [7].
This technique will be applied to the control of the
The main objective is to inject the generated electrical
inverter in order to maintain a constant voltage at the DC
energy into the electrical grid, guaranteeing a unit power
bus and to ensure a unity power factor for the power fed into
factor, a minimum distortion rate and a maximum power
the grid. On the other hand, a combination of incremental
extraction from the PV array. In this regard, DC/DC
conductance and active disturbance rejection control
converter is controlled in MMPT mode to extract the
ADRC_IncCond will be used to control the DC/DC
maximum power from the PV. Several techniques are used
converter to ensure the maximum power point MPP from
in the literature. Perturb and observe technique P&O is
PV array.
widely used for its simplicity of implementation, but it
The rest of this article is organized as follows. The design
exhibits oscillations around the maximum power point when
of the system structure is described in Section 2. In Section
environmental conditions change rapidly [8-10].
3, the overview of the mathematical model of the system is
Incremental conductivity IncCond is also one of the most
presented. Section 4 describes the ADRC control technique.
widely used MPPT techniques, overcoming some of the
In Section 5, ADRC_IncCond, FLC and P&O techniques
drawbacks of the P&O strategy. Many research studies have
are applied to the first stage, while the ADRC is applied to
shown that IncCond outperforms P&O in terms of tracking
the second stage. Simulations results are presented in
efficiency, rise time, fall time and dynamic response [11-
Section 6 and finally a conclusion in Section 7.
14].

Digital Object Identifier 10.4316/AECE.2022.01010

87
1582-7445 © 2022 AECE
 [Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

II. DESIGN OF THE SYSTEM STRUCTURE switch.

The structure of the model considered consists of a two- Output voltage is expressed as:
stage PV system connected to the grid. The PV array is V
VDC  PV (4)
connected to a DC/DC converter controlled to track the 1 D
maximum power point MPP. The grid is connected to the where, VPV and VDC are respectively the input and output
inverter via a filter and a DY transformer whose secondary voltage. D is the duty cycle value located between 0 and 1.
winding (high voltage) is connected to the grid. A filter
consisting of a resistor Rg in series with an inductor Lg is
used to reduce harmonics generated by the use of power
electronics semiconductors (IGBTs-Diodes). The PV system
is based on a combination of modules providing 250 kW
under standard test conditions (1000 W/m2 and 25°C).
A 250 kW load is inserted between buses 1 and 2. Fig. 1
Figure 3. Boost converter model
shows the schematic diagram of the model studied.
C. DC/AC Converter Model
A two level inverter model is used in this paper as shown
in Fig. 4. Each arm consists of two complementary switches
modeled by an associated IGBT-Diode transistor.

Figure 1. PV system schematic diagram

III. SYSTEM MODELING

A. PV modeling
Photovoltaic panel is modeled by considering an
association of NPP strings in parallel. Each string is formed Figure 4. Inverter model
by NSS modules in series. The module consists of NS cells in D. Filter and Transformer Design
series. One diode model of the PV module is widely used in
A series resistance and inductance filter is connected to a
the literature (Fig. 2). This model consists of a photonic
DY transformer. Equivalent model in per unit for DY
current source Iph, a junction diode D and a resistor Rp
transformer in the positive sequence is described in Fig.
connected in parallel. The whole assembly is connected in
5,where: VL and VH denote low and high voltage
series with a resistor Rs.
respectively, z1 and z2 are the per-unit leakage complex
impedance for primary and secondary winding refereed to
the primary winding, G and Bm represent respectively losses
resistances and magnetizing current. The DY transformer
includes a 30° phase shift in the high winding voltage.

Figure 2. Model of PV module

Photovoltaic module delivers IPV current and VPV voltage

at its output. The output current is expressed as follow:
  V  Rs I PV   VPV  Rs I PV
I PV  I ph  I 0  exp  PV   1  (1)
  N SVT    RP Figure 5. Per unit transformer model in positive sequence
I sc  K I ΔT
I0  (2) Using Kirchhoff voltage law KVL for the filter associated
 V  KV ΔT 
exp  oc   1 to transformer (we assume that winding resistance and
 N S VT  exciting current are neglected), voltage equations are given
G in per unit (time is kept in seconds) and in dq axis
I ph   I sc  K I ΔT  (3) representation system (Park transformation) as:
Gn
where, Voc is the open-circuit voltage, Isc is the short-circuit  1 dI gd
Vcd  Vgd  RTot I gd  LTot  LTot I gq
current, KV is the open-circuit voltage/temperature   b dt
 (5)
coefficient,KI is the short circuit current/temperature V  V  R I  1 L dI gq  L I
coefficient, VT thermal voltage and α is the ideality factor.  cq gq Tot gq
 b Tot dt Tot gd

B. DC/DC Converter Model where:

As shown in Fig. 3, the boost converter consist of an o Vc andVg (in dq-axis) are converter and grid voltage
association of inductance L, capacitor C, diode D and an respectively and ωb is the base angular speed;
electronic switch designed by IGBT-Diode transistor. A o  RTot  Rg  RT ; LTot  Lg  LT 
pulse width modulation PWM signal controls the electronic

88
 [Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

o (RT, LT) represent the total resistances and inductances, this purpose, an extended state vector is constructed with an
in per unit, of the transformer windings referred to the additional state, namely the function f assumed to be
low voltage. differentiable and defined as follows:
t
X   x1 xn1    y y  y n1 f  (12)
t
E. DC Bus Voltage Mode x2  xn
The power flow equations through the DC bus voltage are Equation (9) is written as:
given as follows:
 x1  x2
dV 3 
CDCVDC DC  PPV  Vgd I gd (6)  x  x  bu
dt 2  n n 1 (13)
Equation (6) in per unit becomes:  xn 1  h ; h  f
 y  x1
dV 3
CDCVDC DC  b  PPV  Vgd I gd  (7) or in matrix form:
dt 2
or in the linear form: X  AX  Bu  Eh (14)
dw 3 where:
 b  PPV  Vgd I gd  (8)
dt CDC 0 1  0  0 0
A     B    E    
where: PPV is the photovoltaic power and w  VDC 2 . 0 0  1 b 0
0 0  0   n 1 n 1 0   n 11 1   n 11
We assume that losses power in the boost converter,
inverter, filter and transformer are neglected. The extended state observer ESO is then constructed for
(14) as:
IV. ADRC CONTROL
The ADRC structure is becoming more and more popular, 
Z  AZ  Bu  G  y  y 
y  CZ
(15)
several authors have been interested in its description and
 Z   z z  z t
analysis [24], [26-28].  1 2 n 1

To illustrate its concept, we first consider the general where:  G   g 1 g 2  g n 1 

t

model of an nth-order nonlinear time-varying (NLTV) C  1 0  0

dynamical system with a single input u and a single output y 
described by the following equation: In its original form, as reported in [29-30], the functions
 
y (t ) n   f y (t ) n 1 , y (t ) n  2 , , y  t  , d  t   bu  t  (9) gi with (i=1…n+1) are non-linear functions. To facilitate its
construction and the determination of its parameters, Z. Gao
where: proposed to replace the functions by constant gains [31],
o d(t) represents the external disturbances and b is a i.e.:
constant parameter;
G   g1 g 2  g n 1   0 1 0 2  2  0 n 1  n 1  (16)

o f  f y (t ) n 1 , y (t ) n  2  , , y  t  , d  t   represent the
The pole placement technique is adopted to set gains so
unknown NLTV dynamics of the system. It describes the
totality of the disturbances, whether they are internal or
 
that s n 1  1s n   n s   n 1 is Hurwitz. To simplify

external. We note that for this system, only the order and the the setting, i are chosen to have:
parameter b are given. s n 1  1 s n     n s   n 1  ( s  1) n 1 (17)
Since the detailed mathematical system model is
where:
unknown, the ADRC provides an alternative way to control
this kind of problem by estimating and rejecting in real time  n  1!
i  i  1, 2..n  1 (18)
the totality of the perturbations f. i ! n  1  i  !
We denote by f the estimate of the totality of The characteristic polynomial of (15) then becomes:
  s   sI   A  GC    s  0 
n
disturbances. By choosing a control law in the form: (19)
 f  u0 This reduces the ESO setting to a single parameter 0 .
u (10)
b
With a well-tuned observer, the original NLTV system (9) B. Control Low
is reduced to an equivalent cascaded integrator system The control law u allows the system to be reduced to a
which is easily controlled (where u0 is the new control input cascade of integrators, which can be controlled by a
signal): generalized PD controller. We define:
 f  u0 r   yr , y r , , yr     r1 , r2 , , rn 
n 1
(20)
y   f  b
n
 u0 (11)  
b
where:
Nevertheless, such an approach is only justified if an o r1 denotes the input bounded reference signal, which the
estimation of perturbations is possible. This is the purpose output should follow according to the control law;
and role of the extended state observer (ESO). o ri its (i-1)th derivative for i = 2…n.
A. Extended State Observer We then write:
n
The ADRC relies on the proper operation of the ESO and u0  gi  ri  zi  (21)
its ability to establish a good estimate of perturbations. For i 1

89
 [Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

where:gi are non-linear functions. the PV generator. In case of low photovoltaic production,
NB: in linear form, functions gi are replaced by constants the missing energy will be taken from the grid.
coefficients ki. In order to control the DC/AC converter, two cascaded
We define: loops are used. The slower one controls the DC bus voltage


ei  ri  xi
xi  xi  zi
(i  1..n)
(i  1..n  1)
(22)
VDC. The other, faster one, controls direct and quadratic
currents (Id&Iq) flowing through the RL filter.
The state equation for the dynamic error is written as: A. Current Control (Intern Loop)
e  Ae  Bx (23)
Equation (5) of the filter associated to the transformer is
where: written in ADRC canonical form as:
 dI gd
 0 1 0  0    f d  I gd , d , t   bo _ d ud

A
       dIdt (26)
0  0 1 0   gq  f q  I gq , d , t   bo _ q uq
 k1  kn  2 kn 1 kn  nn  dt
b
 0  0 0  
f d I gd , d , t  
Vgd  RTot I gd  LTot I gq 
B  
    LTot
0  0 0 where:
 
 k1  kn 1 n n 1   b  bo _ d  Vcd
 LTot 
 e1   x1  
e     ; x      
f q I gq , d , t  b
LTot

Vgq  RTot I gq  LTot I gd 
 en  n1  xn 1   n 11 o
 
The ki terms are chosen so that the polynomial   b  bo _ q  Vcq
n n 1
 LTot 
( s  kn s    k1 ) is of type Hurwitz. As before, to o ud  Vcd and uq  Vcq
simplify the design of the controller, we take:
o bo _ d and bo _ q are two parameters to estimate and they
s n  kn s n 1  k1   s  c 
n
(24)
are typically chosen equal. A first approximation might
where: 
n! be: bo _ d  bo _ q  b
ki  cn 1i (25) LTot
 
i  1 ! n  1  i  !
Note:
and ωc>0 represents the controller bandwidth, which make o Quadratic current reference I q _ ref is set to be zero, thus
the closed loop characteristic polynomial to have n poles at
only active power is flowing through filter and reactive
- ωc.
power is null.
In practice, ω0 is often chosen in the range of 3~7ωc,
o Direct current reference I d _ ref is generated by the DC
which allows to simplify the tuning of the linear ADRC to a
single parameter ωc. voltage controller.
For the first order system (n=1): o The shift angle induced by DY transformer will be take
o The ESO setting is: into account to evaluate the reference value U abc _ ref

z
 2  2

 1       2 
Z   1  ;  1   2 ;  1    1 02    20 
z g
 2    20   0 
g
used for the inverter PWM block to generate the
appropriates IGBTs-Diodes switching signals (Fig. 8).
o The control low setting is: B. DC Bus Voltage Control
r   yr  ; k1  c ; u0  k1  yr  z1  Equation (8) can be writing into ADRC canonical form
 
 f  z ; u  u0  f  u0  z2 as:
 2
b0 b0
  dw
  f w  w, d , t   bw 0 uw
ADRC schematic diagram is illustrated in Fig. 6.  dt (27)
uw  I gd
where:
3
o f w  w, d , t   b  PPV  Vgd I gd   bw 0 I gd
CDC
o bw 0 is the parameter to estimate.

Figure 6. A first order system ADRC structure Fig. 8 represents the schematic control diagram applied to
the inverter.
V. INVERTER CONTROL
The DC/AC inverter can supply 250 KW to a three-phase
load connected to the grid. Depending on the weather
conditions, the load will be supplied totally or partially by

90
 [Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

Figure 7. Simulation system model

Figure 9. ADRC_IncCond MPPT Simulink model

B. FLC MPPT technique

Fuzzy logic controllers are widely used in several
industrial applications, and moreover, this type of controller
does not require detailed knowledge of the model. It based
on three blocks: Fuzzification, rule base table operation, and
Figure 8. Inverter control schematic diagram Defuzzification. The input variables are converted into
linguistic variables used by the membership functions. For
VI. BOOST CONTROL MPPT control, the input variables typically used are:
The DC/DC converter is controlled to extract the o Error E
maximum power from the PV array for each condition such P P  k   PPV  k  1
E  PV  PV (31)
as atmospheric changes (irradiation and temperature) or load VPV VPV  k   VPV  k  1
variations. In this work, two techniques are investigated, the o Change of error CE
ADRC control based on the incremental conductance
CE  E  k   E  k  1 (32)
ADRC_IncCond and a fuzzy logic control FLC, which will
be compared to the classical perturb and observe P&O Error E checks for the MPP while change of error CE
control. indicates the direction of voltage change. The output
variable is ratio change in duty cycle dD used for DC/DC
A. ADRC_IncCond for MPPT Control converter.
The output PV power is PPV=VPV x IPV, where IPV and VPV Membership functions and fuzzy rules for input and
are PV array output current and voltage respectively. output variables are listed respectively in Fig. 10 and Table
At the maximum power point MPP we have: I.
dPPV dI The three strategies for controlling the maximum power
 I PV  VPV PV  0 (28) point MPP will be evaluated in Matlab Simulink according
dVPV dVPV
to the schematic diagram (Fig. 11) with consideration of
Considering the time derivatives of the voltage and
their efficiency  defined as:
current, the error e can be written as follows:
I PV PPV
 (33)
PMP
e  I PV  VPV t (29)
VPV where: PPV and PMP are respectively power extracted from
t PV and theoretical maximum power.
The main objective is to keep the error e at zero to
achieve a MPP tracking. Since the ADRC does not require TABLE I. FUZZY LOGIC RULES
CE
detailed knowledge of the system, a first order model will be E
NB NM NS ZE PS PM PB
adopted as a first approximation. NB NB NB NB NB NM NS ZE
de NM NB NB NB NM NS ZE PS
 f  e, d , t   bo u (30) NS NB NB NM NS ZE PS PM
dt ZE NB NM NS ZE PS PM PB
where: u  D (D duty cycle) and b0 a parameter to estimate. PS NM NS ZE PS PM PB PB
Fig. 9 shows Simulink diagram control of MMPT based PM NS ZE PS PM PB PB PB
PB ZE PS PM PB PB PB PB
on ADRC_IncCond technique.

91
 [Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

power curves at step change (Fig. 13), we concluded that

ADRC_IncCond and FLC have a better dynamic
performance than P&O (faster response and no overshoot).
In terms of quality, the PV power produced by the
ADRC_IncCond control is very stable and shows few
oscillations compared to the FLC and P&O controls. The
latter shows more oscillations due to the classical algorithm
used.
Table V shows the efficiency of the control strategies
under standard test conditions, taking into account the
magnified curves described in Fig. 13. It can be seen that
ADRC_IncCond performs well compared to the other two
controls.
As shown in Fig. 14, the DC bus voltage is kept constant
and overshoots are less than 5% (less than 30V) even if
external disturbances occur.

Figure 10. FLC input and output variables

Figure 12. Temperature and irradiation variations

260
255
250 254
253
240
0.25 0.3 0.35

230 210
205 240
220 200
195
Figure 11. Overview of the MPPT schematic diagram 210 235
0.42 0.44 1.22 1.24 1.26 1.28

200 IC-ADRC
FLC

VII. DISCUSS AND RESULTS 190 PO

0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

The performance of the proposed control was tested by Time (s)

simulations under Matlab simulink software. The grid- Figure 13. Output PV power
connected photovoltaic system consists of 135 strings of 5 600
modules in series to cut 250 KW. The characteristics of the
U dc (V)

PV array are listed in Table II. 580

The simulation model for the whole system is shown in
Fig. 7, which is used to check out the performance of the 560

proposed control strategies. 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
Three control schemes are applied to the boost converter Time (s)
to track the maximum power point MPP under atmospheric Figure 14. DC bus voltage
variations. We first applied the ADRC_IC control, then FLC
and finally the classical P&O algorithm. While the converter It should be noted, as discussed in previous studies [23-
was kept under control, for all three strategies, using the 24], that ADRC is one of the most robust control methods. It
ADRC control for both the external loop (DC bus voltage) is able to provide good control even if the system is subject
and internal loop (direct and quadratic currents in dq-axis to variations in its internal or external parameters. Indeed, its
reference). design does not require exact knowledge of the system.
PV system and control parameters used in the simulations The current waveform at bus 2 is practically sinusoidal
are listed respectively in Table III and Table IV. and has a zero phase shift with respect to the voltage at the
To evaluate the effectiveness of the control, the system is output of the transformer (Fig. 15 and Fig.16), which means
subjected to a step change in irradiance from 1000 W/m2 to that no reactive power flow is exchanged with the grid, as
800 W/m2 at time t=0.4s. A second step change is applied expected by the inverter control (Fig. 17).
from 800 W/m2 to 1000 W/m2 at time t=1.2s. The powers in buses 1 and 2 are presented in Fig. 17, it
The temperature is also varied from 25 oC to 45 oC at time can be seen that the 250 KW load absorbs all the energy it
t=0.8s, followed by a variation from 45 oC to 25 oC at time needs from the PV generator as long as the standard
t=1.6s (Fig. 12). conditions are met (1000 W/m2 and 25°C). If this is not the
As shown in Fig. 13, maximum power is extract from PV case, the energy deficit is supplied by the grid. Similarly, if
array for the three MPPT strategies (254 KW in standard there is no load or if the load is less than 250 KW, the PV
test conditions 1000 W/m2 and 25°C). By zooming in on the array injects the excess energy directly into the grid.

92
 [Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

using the extended state observer which estimates the

overall uncertainties in real time, since the detailed
V a (KV) & I a (A)

mathematical model is not required.

MMPT techniques for photovoltaic arrays are analyzed
and evaluated. The simulation results showed the good
performance of ADRC_IncCond compared to FLC and
P&O in terms of efficiency, response time and stability.

Figure 15. Current and voltage in bus 2 REFERENCES

[1] G. Carrington, J. Stephenson, “The politics of energy scenarios: are
V a (KV) & I a (A)

international energy agency and other conservative projections

hampering the renewable energy transition?,” Energy research &
social science, vol. 46, pp. 103-113, 2018.
doi:10.1016/j.erss.2018.07.011
[2] E. Asmelash, G. Prakash, R. Gorini, D. Gielen, “Role of IRENA for
global transition to 100% renewable energy,” In: Uyar T. (eds)
Accelerating the Transition to a 100% Renewable Energy Era.
Lecture Notes in Energy, vol. 74, pp. 51-71, 2020, Springer, Cham.
Figure 16. Zoom of current and voltage in bus 2 doi:10.1007/978-3-030-40738-4_2
[3] H. Bevrani, A. Ghosh, G. Ledwich, “Renewable energy sources and
frequency regulation: survey and new perspectives,” IET Renewable
Power Generation, vol. 4, no. 5, pp. 438, 2010. doi:10.1049/iet-
rpg.2009.0049
[4] E. Jahan et al., “Primary frequency regulation of the hybrid power
system by deloaded PMSG-based offshore wind farm using
centralised droop controller,” The Journal of Engineering, vol. 2019,
no. 18, pp. 4950-4954, 2019. doi:10.1049/joe.2018.9326
[5] F. J. Lin, K. C. Lu, B. H. Yang, “Recurrent fuzzy cerebellar model
articulation neural network based power control of a single-stage
Figure 17. Active and reactive powers in buses 1 & 2 three-phase grid-connected photovoltaic system during grid faults,”
IEEE Transactions on Industrial Electronics, vol. 64, no. 2, pp. 1258–
TABLE II. PV ARRAY CHARACTERISTICS 1268, 2017. doi:10.1109/tie.2016.2618882
Parameter Value [6] A. B. Rey-Boué, N. F. Guerrero-Rodríguez, J. Stöckl, T. I. Strasser,
Maximum output power PMAX 414.801 W “Modeling and design of the vector control for a three-phase single-
Open circuit voltage VOC 85.3 V stage grid-connected PV system with LVRT capability according to
short-circuit current ISC 6.09 A the Spanish grid code,” Energies, vol. 12, no. 15, pp. 2899, 2019.
Voltage at maximum power point VMP 72.9 V doi:10.3390/en12152899
Current at maximum power point IMP 5.69 A [7] N. Aouchiche, “Meta-heuristic optimization algorithms based direct
Number of cells in series NS 128 current and DC link voltage controllers for three-phase grid connected
Number of modules in series per string 5 photovoltaic inverter,” Solar Energy, vol. 207, pp. 683–692, 2020.
Number of parallel string 123 doi:10.1016/j.solener.2020.06.086
[8] T. Selmi, M. Abdul-Niby, L. Devis, A. Davis, “P&O MPPT
implementation using MATLAB/Simulink,” Ninth International
TABLE III. PV SYSTEM PARAMETERS
Conference on Ecological Vehicles and Renewable Energies (EVER)
Parameter Value IEEE, pp. 1-7, 2014. doi:10.1109/ever.2014.6844065
CPV = 1.7 mF ; CDC = 17 mF ; LBOOST = 1 mH [9] E. Mamarelis, G. Petrone, G. Spagnuolo, “A two-steps algorithm
Boost converter
VDC = 580V ; PWM = 5 KHz improving the P&O steady state MPPT efficiency,” Applied Energy,
Filter Rg = 1.5 m ; Lg = 1 mH vol. 113, pp. 414-421, 2014. doi:10.1016/j.apenergy.2013.07.022
T1 : (300 KVA – 280V / 25 KV- Y ) [10] F. Liu, Y. Kang, Y. Zhang, S. Duan, “Comparison of P&O and hill
Transformer
T2 : (50 MVA – 25 KV / 120 KV- Y) climbing MPPT methods for grid-connected PV converter,” In 2008
Simulink solver Ode23t – Step size 5s 3rd IEEE Conference on Industrial Electronics and Applications, pp.
804-807, June 2008. doi:10.1109/ICIEA.2008.4582626.
TABLE IV. ADRC CONTROL PARAMETERS [11] S. Z. Mirbagheri, S. Mekhilef, S. M. Mirhassani, “MPPT with
Parameter Value Inc.Cond method using conventional interleaved boost converter,”
MPPT bo = 1000 ; k1 = 350 ; o = 1750 Energy Procedia, vol. 42, pp. 24-32, 2013.
doi:10.1016/j.egypro.2013.11.002
DC bus voltage bo = -580 ; k1 = 50 ; o = 150
[12] K. Ishaque, Z. Salam, G. Lauss, “The performance of perturb and
Current loop bo = 561 ; k1 = 500 ; o = 2500 observe and incremental conductance maximum power point tracking
method under dynamic weather conditions,” Applied Energy, vol.
TABLE V. THE EFFICIENCY OF MPP TRACKING CONTROLS 119, pp. 228-236, 2014. doi:10.1016/j.apenergy.2013.12.054
Standard test [13] C. H. Lin, C. H. Huang, Y. C. Du, J. L. Chen, “Maximum
ADRC_IncCond FLC P&O
conditions photovoltaic power tracking for the PV array using the fractional-
1000 W / m2, 25°C 99.88 % 99.52 % 99.48 % order incremental conductance method,” Applied Energy, vol. 88, no.
12, pp. 4840-4847, 2011. doi:10.1016/j.apenergy.2011.06.024
VIII. CONCLUSION [14] A. Safari, S. Mekhilef, “Simulation and hardware implementation of
incremental conductance MPPT with direct control method using cuk
In this paper, a model of a two-stage grid-connected converter,” IEEE Transactions on Industrial Electronics, vol. 58, no.
photovoltaic system is developed. The model is subjected to 4, pp. 1154-1161, 2011. doi:10.1109/tie.2010.2048834
[15] H. Bounechba, A. Bouzid, K. Nabti, H. Benalla, “Comparison of
varying atmospheric conditions in order to highlight the perturb & observe and fuzzy logic in maximum power point tracker
performance of the controllers used. An active disturbance for PV systems,” Energy Procedia, vol. 50, pp. 677-684, 2014.
rejection control is presented and applied to control the doi:10.1016/j.egypro.2014.06.083
[16] A. M. Noman, K. E. Addoweesh, H. M. Mashaly, “A fuzzy logic
whole system. control method for MPPT of PV systems,” IECON 2012 - 38th
The performance and robustness of the ADRC is Annual Conference on IEEE Industrial Electronics Society, pp. 874-
examined under disturbances for MPPT and inverter control, 880. doi:10.1109/iecon.2012.6389174

93
 [Downloaded from www.aece.ro on Sunday, July 31, 2022 at 17:36:30 (UTC) by 41.141.159.25. Redistribution subject to AECE license or copyright.]

Advances in Electrical and Computer Engineering Volume 22, Number 1, 2022

[17] B. Bendib, F. Krim, H. Belmili, M. F. Almi, S. Boulouma, “Advanced unbalanced grid voltage,” International Review on Modelling and
fuzzy MPPT controller for a stand-alone PV system,” Energy Simulations (IREMOS), vol. 7, no. 1, pp. 95-105, 2014.
Procedia, vol. 50, pp. 383-392, 2014. doi:10.15866/iremos.v7i1.265
doi:10.1016/j.egypro.2014.06.046 [25] A. Boualouch, A. Essadki, T. Nasser, A. Frigui, A. Boukhriss,
[18] H. Rezk, A. M. Eltamaly, “A comprehensive comparison of different “ADRC performance in power control of DFIG used in wind turbine
MPPT techniques for photovoltaic systems,” Solar Energy, vol. 112, during grid voltage fault,” International Journal of Control and
pp. 1-11, 2015. doi:10.1016/j.solener.2014.11.010 Automation, vol. 11, no. 9, pp. 43–56, 2018.
[19] A. K. Rai, N. D. Kaushika, B. Singh, N. Agarwal, “Simulation model doi:10.14257/ijca.2018.11.9.05
of ANN based maximum power point tracking controller for solar PV [26] Q. Zheng, L. Q. Gaol, Z. Gao, “On stability analysis of active
system,” Solar Energy Materials and Solar Cells, vol. 95, no. 2, pp. disturbance rejection control for nonlinear time-varying plants with
773-778, 2011. doi:10.1016/j.solmat.2010.10.022 unknown dynamics,” Proceedings of the 46th IEEE Conference on
[20] S. Messalti, A. Harrag, A. Loukriz, “A new variable step size neural Decision and Control, 2007, pp. 3501-3506.
networks MPPT controller: review, simulation and hardware doi:10.1109/CDC.2007.4434676
implementation,” Renewable and Sustainable Energy Reviews, vol. [27] S. Shao, Z. Gao, “On the conditions of exponential stability in active
68, pp. 221-233, 2017. doi:10.1016/j.rser.2016.09.131 disturbance rejection control based on singular perturbation analysis,”
[21] A. Boualouch, A. Essadki, T. Nasser, A. Boukhriss, A. Frigui, “Power International Journal of Control, vol. 90, no. 10, pp. 2085-2097, 2017.
control of DFIG in WECS using backstipping and sliding mode doi:10.1080/00207179.2016.1236217
controller,” International Journal of Electrical and Computer [28] B. Z. Guo, Z. L. Zhao, “On convergence of non-linear extended state
Engineering, 2015, vol. 9, no 6, p. 612-618, 2015. observer for multi-input multi-output systems with uncertainty,” IET
doi.org/10.5281/zenodo.1108312 Control Theory & Applications, vol. 6, no. 15, pp. 2375–2386, 2012.
[22] Y. Errami, M. Ouassaid, M. Cherkaoui, M. Maaroufi, “Sliding mode doi:10.1049/iet-cta.2012.0123
control scheme of variable speed wind energy conversion system [29] Z. Gao, Y. Huang, J. Han, “An alternative paradigm for control
based on the PMSG for utility network connection,” Studies in system design,” Proceedings of the 40th IEEE Conference on
Computational Intelligence, pp. 167-200, 2014. doi:10.1007/978-3- Decision and Control (Cat. No.01CH37228), pp. 4578-4585 vol.5,
319-11173-5_6 2001. doi:10.1109/CDC.2001.980926
[23] A. Boukhriss, A. Essadki, A. Bouallouch, T. Nasser, “Maximization [30] J. Han, “From PID to active disturbance rejection control,” in IEEE
of generated power from wind energy conversion systems using a Transactions on Industrial Electronics, vol. 56, no. 3, pp. 900-906,
doubly fed induction generator with active disturbance rejection March 2009, doi:10.1109/TIE.2008.2011621
control,” Second World Conference on Complex Systems (WCCS), [31] Z. Gao, “Scaling and bandwidth-parameterization based controller
pp. 330-335, 2014. doi:10.1109/ICoCS.2014.7060907 tuning,” Proceedings of the 2003 American Control Conference, pp.
[24] A. Boukhriss, T. Nasser, A. Essadki, A. Boualouch, “Active 4989-4996, 2003, doi:10.1109/ACC.2003.1242516
disturbance rejection control for DFIG based wind farms under

94