IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO.

4, AUGUST 2021 5069

Generalized IDA-PBC Control Using Enhanced

Decoupled Power Sharing for Parallel Distributed
Generators in Standalone Microgrids
Nidhal Khefifi , Azeddine Houari , Mohamed Machmoum, Abdelhakim Saim, and Malek Ghanes

Abstract— The development of an advanced modular control C f i , L f i and R f i Filter inductance, capacitance, and
strategy for distributed generation-based islanded MicroGrids equivalent resistance of the DG i ,
(MGs) is proposed in this article. This control strategy aims at
achieving robust performances and accurate load power sharing L L_i Inductive impedance “i ” of the con-
in spite of system architecture. This strategy is based on the necting filters to the PCC.
interconnection and damping assignment passivity-based control i ai , i bi , i ci abc-inverter-side currents of DG “i .”
(IDA-PBC), which provides sufficient conditions to ensure the v ai , v bi , v ci abc-inverter-side output voltages of
system modularity and stability. The design methodology of
DG “i .”
the proposed method is declined into three important steps.
The whole system is modeled using the port-controlled Hamil- i di , i qi dq-components of the DG “i ”
tonian (PCH) formalism, the Hamiltonian function is mini- inverter-side currents (i ai , i bi , and
mized to synthesize the corresponding control laws, and finally, i ci ).
the stability of the synthesized control laws is verified. In this v di v qi dq-components of the DG “i ”
work, the Hamiltonian function is augmented with an enhanced
decoupled droop (E2D) control in order to guarantee the stability
inverter-side output voltages (v ai ,
of the whole system and ensure accurate power sharing when v bi , and v ci ).
multiple DG units are interconnected. The effectiveness and i Ldi , i Lqi dq-components of the output current
modularity of the proposed modular IDA-PBC control with the of DG“i.”
E2D technique are evaluated and compared with a recent control ωi Angular frequency of DG “i.”
strategy using an inner proportional-integral control with a
decoupled droop technique. Experimental results and discussions v Cai , v Cbi v Cci abc-filter capacitor voltages.
are provided under resistive–inductive and nonlinear loading v Cdi , v Cqi dq-components of the abc-voltages
conditions. across capacitor filter C f i .
Index Terms— Decoupled droop control, distributed genera- Pi and Q i Measured active and reactive powers
tion, load power sharing, microgrid, passivity-based control, port- at the output of DG “i.”
controlled Hamiltonian (PCH). Pi∗ and Q ∗i Reference active and reactive powers
N OMENCLATURE at the output of DG “i .”
MG MicroGrid. E ∗ and ω∗ Desired nominal voltage and angular
IDA-PBC Interconnection and damping assign- frequency.
ment passivity-based control. ωmax and E max Maximum angular frequency and
DG Distributed generator. voltage. amplitude variations.
PCH Port-controlled Hamiltonian. P and Q Active and reactive power deviations.
PCC Point of common coupling. K pi and K qi Frequency and voltage droop gains.
E2D Enhanced decoupled droop control. Z vi Virtual impedance “i.”
Ng Number of DGs. Rvi and L vi Resistive and inductive components
of the virtual impedance Z vi .
Manuscript received April 8, 2020; revised September 16, 2020; accepted
October 12, 2020. Date of publication October 28, 2020; date of current λi and ξi Reactive and active coefficients of
version July 30, 2021. This work was supported by the IREENA RFI WISE the decoupled droop.
Fund. Recommended for publication by Associate Editor Dehong M. Xu. ο pi and οqi Time constant of the imitated inertia.
(Corresponding author: Nidhal Khefifi.)
Nidhal Khefifi, Azeddine Houari, Mohamed Machmoum, and Abdel- S Laplace operator.
hakim Saim are with the Institut de Recherche en Energie Electrique X State variables vector
de Nantes Atlantique, University of Nantes, 44602 Saint-Nazaire, France U Control inputs vector.
(e-mail: [email protected]; [email protected];
[email protected]; [email protected]). H Hamiltonian function.
Malek Ghanes is with the École Centrale de Nantes-LS2N, UMR CNRS d External perturbation.
6004, 44321 Nantes, France (e-mail: [email protected]). J and R Interconnection and damping
Color versions of one or more of the figures in this article are available
online at https://ieeexplore.ieee.org. matrices.
Digital Object Identifier 10.1109/JESTPE.2020.3034464 g Output matrix.
2168-6777 © 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
5070 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO. 4, AUGUST 2021

n and m Number of state and inputs vari- requirements are generated close to end users, which offers
ables for a single generator. more flexibility, increases energy efficiency, and improves
xi DG “i ” state variables vector. system reliability by integrating multiple DG units [8]–[10].
ui DG “i ” control inputs vector. The use of decentralized droop control methods is widely
Hi DG “i ” energy. adopted to control the active and reactive power sharing in
gi DG “i ” input matrix. MGs essentially when multiple DG units are interconnected.
Ji and Ri DG “i ” interconnection and damp- These techniques allow controlling the power participation
ing matrices. of each DG unit by regulating the voltage and the frequency
di DG “i ” external perturbations. locally without communication. The conventional droop
Hdi DG “i ” desired energy function. method with (P-f) and (Q − E) is applicable when the
Jdi and Rdi DG “i ” desired interconnection system shows predominant inductive impedance, with
and damping matrix. small R/X ratio as it is the case for high- and medium-
x i∗ DG “i ” desired state-space vector voltage systems. However, in case of low-voltage MGs, line
at the equilibrium point. impedances are mainly resistive with high R/X ratio or even
Jai and Rai DG “i ” new assigned interconnec- a complex impedance with equivalent resistive–inductive
tion and damping matrix. behavior, which implies that the conventional droop
βi DG “i ” applied control. technique is no longer effective in low-voltage MGs [11].
E iref and ωiref DG “i ” voltage and frequency ref- In this situation, the coupling between active and reactive
erences. powers is strengthened, which makes their control more
∗ ∗
i di and i qi DG “i ” current references. difficult, resulting in inaccurate power-sharing and stability
∗ ∗
v Cdi and v Cqi DG “i ” voltage reference. problems [12]. To solve these problems and improve
ai j Parameters for the IDA-PBC con- power-sharing accuracy, several droop control techniques are
troller. proposed in the literature. A power transformation droop
NL Nonlinear load. technique in which modified active and reactive power
RL Resistive–inductive load. expressions are adopted instead of the measured ones [13].
DTS Decoupled trigonometric saturated Although this technique shows interesting power sharing
droop control. performances, it increases control complexity. Alternatively,
G V and G C PI transfer functions of the outer the use of virtual impedances through the addition of virtual
and inner loops of the compared passive elements is widely employed in order to enforce the
controller. desired behavior and mitigate the existing coupling between
k pV and ki V Proportional and the integral PI active and reactive powers [14], [15]. Even though the addition
control gains of the voltage loop. of virtual passive elements considerably improves the desired
k pC and kiC Proportional and the integral PI impedance predominance without inflicting additional power
control gains of the current loop. losses, the output voltage amplitude can drop considerably.
Kpfmax and KIfmax Proportional and the integral coef- In the same line, an advanced droop control technique with
ficient related to the frequency additional decoupling signals and scheduled decoupling gains
and voltage of virtual power sag have been proposed in [16]. This technique allows achieving
limit of the decoupled trigonomet- accurate power sharing, but the addition of supplementary
ric control. signals can affect the steady-state error. To solve this problem,
δ f and δ E Coefficients fixing the variation distributed cooperative control techniques have been proposed
limits of voltage and frequency of in [17]. This technique achieves satisfactory performances,
the trigonometric function. but it requires advanced communications limiting its
μ p and μq Concavity coefficients of the exploitation. In [18] and [19], a DTS droop controller is
trigonometric function. proposed achieving stable power sharing and proper dynamic
decoupling between active and reactive powers irrespective
of lines impedances behavior. However, this technique suffers
I. I NTRODUCTION
from relatively high computing time as the control complexity

T ODAYS’ energy transition challenges make the

emergence of MGs with renewable energy resources
integration as a key solution to meet with numerous climate,
is increased requiring multiple trigonometric operations. All
these techniques require considerable knowledge on the
MG system architecture and parameters, which cannot
technoeconomic, and energy-resiliency requirements [1]. be guaranteed when the MG is subject to arbitrary loads
In order to meet with these requirements, many sustainable and DGs’ connection and disconnection. Therefore, when
energy technologies are exploited, including solar photovoltaic considering scalable MGs, the system and control laws
panels, wind turbines, and energy storage systems [2], [3]. parameters become no longer convenient, and the system
This concept integrates multiple distributed generators (DGs) stability will not be assured. In other side, and in order to
that operate near to the point of power consumption to maintain the output voltage amplitude and frequency within
supply end users either in islanded or in grid-connected acceptable ranges, some research works propose adding a
operation mode [4]–[7]. In the islanded mode, the electricity higher control level commonly known as a secondary control.

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
KHEFIF et al.: GENERALIZED IDA-PBC CONTROL USING ENHANCED DECOUPLED POWER SHARING 5071

This communication-based control level provides means to

manage the interaction between DG units, restore voltage and
frequency values, and ensure in consequence more accurate
power sharing while preserving the system stability. Similarly,
secondary control techniques have been proposed using either
stochastic control [20] or neural network techniques [21], [22].
These controllers have some limitations essentially related to
their complex structure and their dependence on high-speed
communication. Conversely, the use of decentralized control
approaches can be more interesting as they do not require
communication. In addition, the use of lower control levels
is required to maintain the output voltage of each DG
unit within standardized electrical boundaries [23], [24].
At this level, many control techniques can be found in
the literature, including classical control techniques [25],
[26], nonlinear control techniques [15], [27], learning-
based control methods [28], [29], and predictive control
techniques [30], [31]. These methods can ensure satisfactory
control performances but suffer from their dependence on the
system parameters. This model dependence may reduce their
effectiveness in the case of scalable MG applications. Indeed,
in such applications, the system architecture may evolve,
and thus, the control laws no longer guarantee the desired
performances and system stability. At this level, the use Fig. 1. Steps for IDA-PBC control synthesis.
of energetic approaches, such as the IDA-PBC method,
is recommended. This control method can be interesting for both voltage and power-sharing controllers with the objective
MG applications as it can synthesize robust control laws to ensure high-voltage quality and guarantee accurate active
and provide sufficient conditions to guarantee the stability and reactive power sharing between interconnected DG units.
of complex and multiphysical systems despite the existing The effectiveness and modularity of the proposed IDA-PBC
coupling between them [32], [33]. This method has been control with the E2D technique are evaluated and compared
particularly used in distributed systems since it allows the with a recent control strategy proposing the use of an inner
synthesis of modular control laws [34]. proportional–integral control with a saturated trigonometric
In this article, the development of an advanced modular droop control technique.
control strategy based on the IDA-PBC is proposed. The This article is organized as follows. A short overview of the
objectives of the proposed strategy are to ensure the system IDA-PBC method and its theoretical synthesis is presented in
modularity and stability while achieving robust performances Section II. Section III details the design methodology of the
and accurate load power sharing in spite of system architec- IDA-PBC control for the studied MG system. Section IV intro-
ture. For design purposes, the whole system is first modeled duces a recent decoupled droop control used as a comparative
using the PCH formalism before the Hamiltonian function technique method. To demonstrate the effectiveness of the
is minimized to synthesize the control laws, and finally, proposed IDA-PBC controller with the E2D control, detailed
the stability of the synthesized control laws is verified. The comparative experimental results are presented in Section V.
Hamiltonian function refers to the system physical energy Finally, this article ends with a conclusion in Section VI.
and is assimilated to the Lyapunov function, which allows
guaranteeing the stability of the synthesized controllers. The II. IDA-PBC M ETHODOLOGY
PCH formalism is characterized by an interesting symplectic
This section presents the essential knowledge and synthesis
geometrical form with a diagonal skew-symmetric matrix that
steps of the control laws to understand the proposed modular
allows synthetizing modular control laws. The resulting con-
IDA-PBC control and the Hamiltonian formalism. The detailed
trol laws guarantee the system modularity and stability even in
control synthesis steps are presented in Fig. 1 and described
the case of system reconfiguration and expansion. To further
as follows.
exploit the advantages offered by the Hamiltonian formalism
Step1 (System Modeling on the PCH Form): To synthesize
and achieve accurate power sharing when multiple DG units
the IDA-PBC controller, the system is first modeled in the PCH
are interconnected, the Hamiltonian function is augmented
form described in expression (1) [34]. The use of this model
with an E2D power-sharing controller. The proposed E2D
form is motivated not only by the use of an energy approach
controller uses an additional filter that improves the system
necessary to synthesize the desired control but also by its
response against transients by introducing virtual inertia, while
modular structure allowing the modeling of several connected
the Hamiltonian form is maintained. This modeling formalism
parallel DGs units to the PCC
allows guaranteeing a generic control synthesis based on
the IDA-PBC method. This technique is applied to design ẋ = [J (x) − R(x)]∇ H (x) + g(x)u + d. (1)

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
5072 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO. 4, AUGUST 2021

The modular form of the previous PCH formula is detailed in describes the external perturbations. Ri (x i ) and Ji (x i ) are,
the following expressions where: respectively, the damping and interconnection matrices that
⎡ ⎤ ⎡ ⎤ ⎡ ⎤ verify Ri (x i ) = RTi (x i ) ≥ 0 and Ji (x i ) = −JiT (x i ).
x1 u1 d1
Step 2 (IDA PBC Law Design): In order to synthesize
⎢ .. ⎥ ⎢ .. ⎥ ⎢ .. ⎥
⎢ . ⎥ ⎢ . ⎥ ⎢ . ⎥ the IDA-PBC controller consisting to shape the energy of
⎢ ⎥ ⎢ ⎥ ⎢ ⎥
x =⎢ x
⎢ i ⎥
⎥ ; u(x) = ⎢
⎢ i ⎥ u ⎥ ; d = ⎢ di ⎥;
⎢ ⎥
the system by controlling its dissipation, the steps presented
⎢ . ⎥ ⎢ . ⎥ ⎢ . ⎥ in Fig. 1 are to follow.
⎣ .. ⎦ ⎣ .. ⎦ ⎣ .. ⎦
First of all, designers need to identify the desired energy
x ng u ng dng function of the studied system noted generally by Hdi (x i ).
⎡ ⎤
g1 · · · 0 · · · 0 Then, the control “u” is set to assure not only the PCH form
⎢ .. . . .. .. ⎥ but also the desired dynamic performances expressed in the
⎢ . . . . ⎥
⎢ ⎥ following equation:
g(x) = ⎢
⎢ 0 g i 0 ⎥
⎥
⎢ . . . . ⎥
⎣ .. .. . . .. ⎦ ẋ i∗ = [Jdi (x i ) − Rdi (x i )]∇ Hdi (x i ). (4)
0 · · · 0 · · · gng Expression (4) describes the desired state-space representation
⎡ ⎤
J1 (x 1 ) · · · 0 ··· 0 equations at the equilibrium point x i∗ in the port Hamiltonian
⎢ .. .. .. .. ⎥ formulation, where Hdi (x i ) is the energy function that charac-
⎢ . . . . ⎥
⎢ ⎥ terizes the local minimum at the desired equilibrium point. Jdi
J (x) = ⎢
⎢ 0 J (x
i i ) 0 ⎥
⎥
⎢ .. . . ⎥ and Rdi represent, respectively, the desired interconnection
⎣ . .. .. ⎦ and damping matrix. These matrices satisfy, respectively:
0 ··· 0 · · · Jng (x ng ) Jdi (x i ) = −JdiT (x i ) and Rdi (x i ) = RTdi (x i ) ≥ 0.
⎡ ⎤
R1 (x 1 ) · · · 0 · · · 0 The calculation of the control laws is done by solving the
⎢ .. .. .. .. ⎥ matching equation defined by (5). This equation links the
⎢ . . . . ⎥
⎢ ⎥ expression of the closed loop (4) to that of the open loop (3)
R(x) = ⎢⎢ 0 R (x
i i ) 0 ⎥
⎥
⎢ . . . ⎥ [Ji (x i ) − Ri (x i )]∇ Hi (x i ) + gi (x i )βi (x i ) + di
⎣ .. .. .. ⎦
0 · · · 0 · · · Rng (x ng ) = [Jdi (x i ) − Rdi (x i )]∇ Hdi (x i ). (5)
⎡ ⎤
H1 · · · 0 · · · 0 Solving the matching equation (5) requires the use of new
⎢ .. . . . .. ⎥ assigned matrices Jai (x i ) and Rai (x i ) and a new vector called
⎢ .
⎢ . .. . ⎥ ⎥ K i (x i ) [35]. These components are fixed by the designer to
H (x) = ⎢
⎢ 0 H i 0 ⎥;
⎥
⎢ .. .. . . .. ⎥ facilitate the synthesis of the applied control defined by u i =
⎣ . . . . ⎦ βi (x i ).
0 · · · 0 · · · Hng This matching equation can be expressed differently by the
1 T −1 following equation:
where H i = x i Q i x i . (2)
2
[(Ji (x i ) + Jai (x i )) − (Ri (x i ) + Rai (x i )]K i (x i )
In these expressions, x ∈ Rng is the state vector. u ∈
= −[Jai (x i ) − Rai (x i )]∇ Hi (x i ) + gi (x i )βi (x i ) + di . (6)
Rng is the control input vector, where “ng” represents the
number of parallel DG units. g(x) ∈ Rng is the input matrix. The determination of the elements that constitute the matri-
H (x) ∈ Rng×ng is a function defining the system energy. ces Rai (x i ) and Jai (x i ) depends on the control objectives.
“d” is a vector corresponding to the external perturbation. The designer is free in his choice. Often, in electrical power
J (x) ∈Rng×ng and R(x) ∈Rng×ng are, respectively, the inter- systems, the interconnection matrix is set to avoid the coupling
connection and damping matrices. These two matrices verify between voltage and current where the damping matrix is
R(x) = RT (x) ≥ 0 and J (x) = −J T (x). selected to enhance the dynamic convergence.
This modular formulation allows the synthesis of a global Step 3 (Stability Verification): This last step is mandatory in
control for all systems based on only one elementary DG order to prove the stability of the system around its equilibrium
indexed by “i.” Following the previous model expression in point x i∗ . In this phase, the designer must verify that the
(1), the elementary model of each DG can be expressed by resulting closed-loop system keeps the PCH form and shows
the following expression: the desired dynamics described in (4). Therefore, the selected
damping and interconnection matrix Rai (x i ) and Jai (x i ) need
ẋ i = [Ji (x i ) − Ri (x i )]∇ Hi (x i ) + gi (x i )u i + di . (3)
to fulfill the following five conditions [36].
In this model, x i ∈ Rn represents the state vector and 1) Structure Preservation: Rdi (x i ) = Ri (x i ) + Rai (x i ) =
u i ∈ Rm depicts the input control vector in which “n” and RTdi (x i ) ≥ 0 and Jdi (x i ) = Ji (x i )+Jai (x i ) = −JdiT (x i ).
“m” represent, respectively, the number of states and inputs 2) Integrability: K i (x i ) is the gradient of a scalar function:
variables for a single generator and guarantee m < n. Hi (x i ) ∇ K i (x i ) = [∇ K i (x i )]T .
is the energy of the elementary DG i system, gi (x i ) is the input 3) Equilibrium Condition: K i (x i ) at x i∗ satisfies: K i (x i∗ ) =
matrix of the elementary DG i system, and di is a vector that −∇ Hi (x i∗ ).

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
KHEFIF et al.: GENERALIZED IDA-PBC CONTROL USING ENHANCED DECOUPLED POWER SHARING 5073

Fig. 2. Structure of the studied system.

4) Lyapunov Stability: The Jacobian of K i (x i ) at x i∗ satisfies

the bound: ∇ K i (x i∗ )> −∇ 2 Hi (x i∗ ).
5) Invariance of the Equilibrium Point
[37]:−[∇ Hdi ]T Rdi (x i∗ )∇ Hdi ≤ 0.
Condition 5) guarantees that the solution is bounded for the
desired function Hdi .
Checking the five previous conditions proves that the control
Fig. 3. Classical droop control for MG with inductive behavior.
of the closed-loop system u i = βi (x i ) will be a port-controlled
where K i (x i ) = ∇ Hai (x i ) and Hdi (x i ) = Hi (x i ) + Hai (x i ).
dv Cqi
C fi = i qi − i Lqi − C f i ωi v Cdi. (7)
III. A PPLICATION OF THE IDA-PBC TO THE S TUDIED MG dt
This section presents the application of the proposed method In these expressions, “i ” designates the index of each DG. v di
to the studied MG by considering the design steps that have and v qi and i Ldi and i Lqi represent, respectively, the output
already been presented in Fig. 1. voltages of the inverter “i ” and the load currents in each DG
unit in the dq-frame reference. i di and i qi and v Cdi and v Cqi
A. Step1: System Modeling on the PCH Form represent separately the dq-components of the line currents
(i ai , i bi , and i ci ) and the voltage through the capacitors C f i .
The studied islanded MG is shown in Fig. 2. It is composed Finally, ωi denotes the angular frequency in DG i .
of “ng” parallel three-phase inverters that supply a local load 2) Interaction Model: The modeling of parallel DG units
connected to the (PCC) via LCL filter. The local load is sup- needs to take into consideration the interaction between them
plied by power provided from primary sources. In this work, to avoid the problem of circulating currents [38]. This inter-
the primary sources are supposed constant and are presented action is modeled by droop equations. In fact, systems based
by a dc link. In this scheme, R f i , C f i , and L f i represent, on renewable energies have low inertia. The idea of using the
respectively, the internal per-phase resistance, capacitor, and droop equations consists of conferring to the system a kind of
inductance of the inverter output filter. The link between DGs virtual inertia so that it behaves like a conventional generator in
and the PCC is modeled by an impedance characterized by an front of disturbances [39]. The use of these equations permits
inductance L L_i . to ensure the power sharing without the need for expensive
In each DG unit, a low-level control is implemented to communications protocol [9]. A brief review of the traditional
regulate at the same time the output voltage across the droop control is presented in this section. In fact, in most
capacitors C f i and the power sharing using a droop control low-voltage DG systems, distances between DGs are modeled
technique. This level guarantees load power sharing between by inductors, which promotes the inductive behavior of the
the paralleled DGs units. To obtain the individual control laws, system. In this case, the use of the classical droop control,
each DG system is separately modeled by an internal and an presented in (8), is justified. It permits as presented in Fig. 3 to
interaction model. control the active power by acting on the frequency component
1) Internal Model: The internal model of each DG unit that and controlling the reactive one by managing the voltage
corresponds to the inverter output LC filter is described in the amplitude [38]
rotating dq reference by the following equations:
di di δ̇i = ωi
L fi = −R f i i di + L f i ωi i qi + v di − v Cdi ωi = ω∗ − K pi (Pi − Pi∗ )
dt
L fi
di qi
= −R f i i qi − L f i ωi i di + v qi − v Cqi E i = E ∗ − K qi (Q i − Q ∗i ) (8)
dt
dv Cdi where Pi and Q i are, respectively, the measured active and
Cfi = i di − i Ldi + C f i ωi v Cqi reactive power at the filter output , Pi∗ and Q ∗i denote,
dt
Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
5074 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO. 4, AUGUST 2021

respectively, the desired active and reactive power references The new formula of the E2D control is written as follows:
of DGi , ω∗ ∈ R is the desired nominal frequency, and E ∗ is dδi
the desired nominal voltage amplitude. Finally, frequency and = ωi
dt
voltage droop gains are designed, respectively, by K pi and K qi .  
dωi
These coefficients are defined by the following equations: ο pi = K pi Pi∗ − Pi + λi E i∗ − E i +(ωi − ωi∗ )
dt
ωmax d Ei 
K pi = οqi = K qi Q ∗i − Q i + ξi (ωi∗ − ωi ) + (E i − E i∗ ) (12)
P dt
E max where ο pi and οqi correspond to the time constant of the
K qi = . (9)
Q low-pass filter used to mimic the inertia of the real synchro-
In this equation, ωmax is the maximum tolerable pulsation nous generator.
deviation related to the active power variation P. In the In regard to the planned control that contains the local
same way, E max is the maximum authorized voltage variation and the interconnected model composed by the enhanced
associated with the reactive power change Q. droop control, the virtual impedance, and the individual
In practice, it is difficult to know the impedance value DGi model, the candidate Hamiltonian function can be
connecting each DG unit to the PCC. Under such condi- taken as the sum of all the energy stored in the used
tions, the conventional droop alone cannot achieve the desired filters
reactive and active power sharing. The interest in adding a 1 1 1
Hi (x i ) = L f i i di
2
+ L f i i 2qi + C f i v 2Cdi
virtual impedance to the primary control structure allows first 2 2 2
to support the inductive behavior of the output impedance 1 1 1
and second to avoid the additional cost of installing real + C f i v 2Cqi + ο pi ωi2 + οqi E i2
2 2 2
physical impedances [4]. 1 1
The added virtual impedance control law expression for + L vi i Ldi
2
+ L vi i Lqi2
. (13)
2 2
each (DG) is described in the following equation [40]:
For the proposed energy function, the chosen state variable
di Ldi vector is designed as
L vi = v Cdi − Z vi i Ldi
dt T
di Lqi xi = x 1i x 2i x 3i x 4i x 5i x 6i x 7i x 8i
L vi = v Cqi − Z vi i Lqi (10) T
dt = Q i i di i qi v Cdi v Cqi i di i qi i di i qi
where Z vi = Rvi + L vi ωi , with Rvi and L vi , respectively, where Q i = diag{L f i L f i C f i C f i ο pi οqi L vi L vi } and
the value of resistance and inductance of the used virtual Hi (x) = (1/2)x iT Q −1
i xi .
impedances. Using (8), (10), and (12), the modular system modeling
Note that in case, in the case of low-voltage systems, the fre- in the PCH form of a DGi system is derived as follows
quency variation influences widely both the active and reactive considering ∇ Hi (x i ) = [i di i qi v Cdi v Cqi ωi E i i Ldi i Lqi ]T .
power and, consequently, the voltage amplitudes. Moreover, In this model, we consider that λi = −ξi
the value of the virtual impedance is limited by the droop
voltage and system stability. This means that the coupling Ji (x i )
⎡ ⎤
cannot be effectively eliminated. To guarantee the decoupling 0 L f i ωi −1 0 0 0 0 0
between the reactive and the active power, decoupling control ⎢ ⎥
⎢ −L f i ωi 0 0 −1 0 0 0 0 ⎥
terms are introduced to the classical droop to ensure the ⎢ ⎥
⎢ 1 0 0 C f i ωi 0 0 −1 0 ⎥
compensation of the voltage and frequency deviations. The ⎢ ⎥
⎢ ⎥
expression of the previous classical droop control (8) becomes ⎢ 0 1 −C f i ωi 0 0 0 0 −1 ⎥
=⎢
⎢
⎥;
as follows: ⎢ 0 0 0 0 0 −λi 0 0 ⎥
⎥
⎢ ⎥
⎢ 0 0 0 0 −ξ i 0 0 0 ⎥
δ̇i = ωi ⎢ ⎥
 ⎢ 0 0 1 0 0 0 0 0 ⎥
ωi = ω∗ − K pi Pi − Pi∗ + λi (E ∗ − E i ) ⎣ ⎦
 0 0 0 1 0 0 0 0
E i = E ∗ − K qi Q i − Q ∗i + ξi (ω∗ − ωi ) (11) ⎡ ⎤
0
where λi and ξi present, respectively, the decoupling reactive ⎢ ⎥
⎢ 0 ⎥
and active droop coefficients. ⎢ ⎥
⎢ 0 ⎥
The studied system is based on renewable energy systems. ⎢ ⎥
⎢ ⎥
These systems are characterized by weak inertia. To improve ⎢ 0 ⎥
⎢
di = ⎢ ∗ ∗ ∗⎥
⎥
the behavior of these systems in front of current disturbances ⎢ ω + K pi Pi + λi E i ⎥
and to achieve high-voltage and frequency regulation, addi- ⎢ ∗ ⎥
⎢ E + K qi Q ∗i + ξi ωi∗ ⎥
tional virtual inertia is introduced to expressions (11). This ⎢ ⎥
⎢ 0 ⎥
additional inertia, modeled by a low-pass filter, improves the ⎣ ⎦
stability of each DG system. 0

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
KHEFIF et al.: GENERALIZED IDA-PBC CONTROL USING ENHANCED DECOUPLED POWER SHARING 5075

Ri (x i ) can be expressed as
⎡ ⎤
Rfi 0 0 0 0 0 0 0 Jdi (x i )
⎢ ⎡ ⎤
⎢ 0 Rfi 0 0 0 0 0 0 ⎥⎥ 0 0 a13i 0 0 0 0 0
⎢ 0 0 0 ⎥ ⎢ 0 ⎥
⎢ 0 0 0 0 0 ⎥ ⎢ 0 0 a24i 0 0 0 0 ⎥
⎢ ⎥ ⎢ a31i ⎥
⎢ 0 0 0 0 0 0 0 0 ⎥ ⎢ 0 0 0 0 0 a37i 0 ⎥
=⎢
⎢
⎥; ⎢ 0 a48i ⎥
⎢ 0 0 0 0 1 0 0 0 ⎥⎥ =⎢
a42i 0 0 0 0 0 ⎥
⎢ ⎢ 0 0 ⎥
1 0 0 ⎥ ⎢ 0 0 0 0 0 0 ⎥
⎢ 0 0 0 0 0 ⎥ ⎢ 0 0 ⎥
⎢ ⎥ ⎢ 0 0 0 0 0 0 ⎥
⎣ 0 0 0 0 0 0 Zv 0 ⎦ ⎣ 0 0 a73i 0 0 0 0 0 ⎦
0 0 0 0 0 0 0 Zv 0 0 0 a84i 0 0 0 0
gi (x i ) Rdi (x i )
⎡ ⎤ ⎡ ⎤
1 0 0 0 a11i 0 0 0 0 0 0 0
⎢0 ⎢ 0 ⎥
⎢ 1 0 0 ⎥ ⎥ ⎢ a22i 0 0 0 0 0 0 ⎥
⎢0 ⎡ ⎤ ⎢ 0 ⎥
0 ⎥ ⎢ 0 a33i 0 0 0 0 0 ⎥
⎢ 0 0 ⎥ v di ⎢ 0 ⎥
⎢ ⎥ ⎢ v qi ⎥ =⎢
0 0 a44i 0 0 0 0 ⎥
⎢0 0 0 0 ⎥ ⎢ 0 ⎥ (16)
=⎢ ⎥; u i = ⎢
⎢
⎥
⎥ (14) ⎢ 0 0 0 a55i 0 0 0 ⎥
⎢0 0 − K pi 0 ⎥ ⎣ Pi ⎦ ⎢ 0 ⎥
⎢ ⎥ ⎢ 0 0 0 0 a66i 0 0 ⎥
⎢0 −K qi ⎥ ⎣ 0 ⎦
⎢ 0 0 ⎥ Qi 0 0 0 0 0 a77i 0
⎢ ⎥
⎣0 0 0 0 ⎦ 0 0 0 0 0 0 0 a88i
0 0 0 0 where ai j = −a j i and aii ≥ 0 with i, j = {1, . . . , 8} for i = j .
 ∗ Rdi (x i ) = Rdi T
(x i ) ≥ 0 and Jdi (x i ) = −JdiT (x i ) are,
Hdi x i respectively, the desired damping and interconnection matrices
1  ∗ 2 1  ∗ 2 certifying a13i = −a 31i = −1,a24i = −a 42i = −1,a73i =
= L f i i di − i di + L f i i qi − i qi
2 2 −a 37i = −1, a84i = −a 48i = −1, and a j j i ≥ 0, where
1  1  j = {1, . . . , 8} and a55i = a66i = 0.
∗ 2 ∗ 2
+ C f i v Cdi − v Cdi + C f i v Cqi − v Cqi The implemented control laws are obtained by solving the
2 2
1  1  matching equation (5) and are based on the four following
+ ο pi ωi − ωi∗ + οqi E i − E i∗
2 2
control loops.
2 2 1) Enhanced Decoupled Droop Control Loop: This loop is
1  ∗ 2 1  ∗ 2 used to generate automatically the reference voltage (E iref and
+ L vi i Ldi − i Ldi + L vi i Lqi − i Lqi
2 2 ωiref ) according to the measured powers
1  1   
= x 1i − x 1i∗ +
2
x 2i − x 2i∗
2
ωiref = ω∗ − K pi Pi − Pi∗ + λi E i∗ − E i + a55i (ωi − ω∗ )
2L f i 2L f i   
E iref = E ∗ − K qi Q i − Q ∗i +ξi ωi∗ − ωi +a 66i E i − E ∗ .
1  1 
x 3i − x 3i∗ + x 4i − x 4i∗
2 2
+ (17)
2C f i 2C f i
Both of the reference voltage amplitudes and the reference pul-
1  1 
x 5i − x 5i∗ + x 6i − x 6i∗
2 2
+ sations of the E2D control are composed of the classical droop
2ο pi 2οqi control, the decoupling active and reactive power expressions
1  2
1  and an additional compensation term for the used references.
x 7i − x 7i∗ x 8i − x 8i∗ .
2
+ + (15) E iref and ωiref are used to provide the three voltage references
2L vi 2L vi
for each DG system in the following equation:
V1ref(i) (t) = E iref sin ∈ (ωiref (t))
 
2π
V2ref(i) = E i sin ωi (t) −
ref ref
3
B. Step2: IDA PBC Laws Design  
2π
This section is dedicated to the synthesis of the proposed V3ref(i) = E iref sin ωiref (t) + . (18)
3
control illustrated in the synoptic scheme presented in Fig. 4. These voltage references will be expressed in a stationary
The desired Hamiltonian function has a local minimum at framework to obtain the dq-axis voltage references defined
the desired equilibrium point x i∗ , which is defined as presented ∗
by Vcdi ∗
and Vcqi .
in (15). 2) Virtual Impedance Loop: Improving the system response
In these expressions, the designer considers the open-loop and the power sharing needs to take into consideration the
Hamiltonian function defined in (3) and the desired control output impedance. At this stage of control, a virtual impedance
objective (4) that allows the system to reach minimal energy is enforced as follows:
at the desired equilibrium point x i∗ . Then, with regard to ∗ ∗
the initial form of the interconnection and damping matrices v Cdi−Zv = (Rvi + L vi ωi )i Ldi + a77i (i Ldi − i Ldi )
∗ ∗
(Ji (x i ) and Ri (x i )) presented in (14), the new desired matrices v Cqi−Zv = (Rvi + L vi ωi )i Lqi + a88i (i Lqi − i Lqi ). (19)

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
5076 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO. 4, AUGUST 2021

Fig. 4. Structure of the used global control.

3) Voltage Loop Control: The third loop is the voltage loop Rdi (x i )
⎡ ⎤
control; it is used to generate the dq-currents reference to the a11i 0 0 0 0 0 0 0
currents loop ⎢ 0 a22i 0 0 0 0 0 0 ⎥
⎢ ⎥
⎢ 0 0 a33i 0 0 0 0 0 ⎥
∗
v Cdi−Zv ∗
= (Rvi + L vi ωi )i Ldi + a77i (i Ldi − i Ldi ) ⎢ ⎥
⎢ 0 0 0 a44i 0 0 0 0 ⎥
∗ ∗
v Cqi−Zv = (Rvi + L vi ωi )i Lqi + a88i (i Lqi − i Lqi ). (20) =⎢
⎢ 0
⎥
⎥
⎢ 0 0 0 a55i 0 0 0 ⎥
⎢ 0 0 0 0 0 a66i 0 0 ⎥
4) Current Loop Control: The inner loop that controls ⎢ ⎥
⎣ 0 0 0 0 0 0 a77i 0 ⎦
the currents to generate the voltage PWM references of the
inverter 0 0 0 0 0 0 0 a88i
∗
 ∗ = [Rdi (x i )] > 0
T
(23)
v Cdi = −a11i i di − i di + v di − R f i i di + L f i ωi qi
∗
 ∗
v Cqi = −a22i i qi − i qi + v qi − R f i i qi − L f i ωi di . (21) where a13i = −a 31i , a24i = −a 42i , a37i = −a 73i , a48i = −a 84i ,
and a ppi ≥0 etk I > 0, given that p = {1, . . . , 8}.
C. Step 3: Stability Verification This proves that the first structural condition 1 is verified.
a) Integrability:
This section presents the verification of the necessary
 
conditions to guarantee the local stability at the point of Hai x i∗ = Hdi x i∗ − Hi (x i )
equilibrium x ∗ by applying the proposed controller based 1  1 
x 1i − x 1i∗ + x 2i − x 2i∗
2 2
on the passivity. The five conditions already presented in =
2L f i 2L f i
Section II are checked. 1  1 
x 3i − x 3i∗ + x 4i − x 4i∗
2 2
1) Structure Preservation: +
2C f i 2C f i
⎡ ⎤ 1  1 
0 0 a13i 0 0 0 0 0 + x 5i − x 5i∗ +
2
x 6i − x 6i∗
2
⎢ 0 0 0 a24i 0 0 0 0 ⎥ 2ο pi 2οqi
⎢ ⎥ 
⎢a 0 0 0 a37i 0 ⎥ 1  2
1 
⎢ 31i 0 0 ⎥ + ∗
x 7i − x 7i + x 8i − x 8i∗
2
⎢ ⎥
⎢ 0 a42i 0 0 0 0 0 a48i ⎥ 2L vi 2 2L vi
Jdi (x i ) = ⎢
⎢ 0
⎥ 1 x 1i 1 x 2i2
⎢ 0 0 0 0 0 0 0 ⎥⎥ − +
⎢ 0 0 ⎥
2 L fi 2 L fi
⎢ 0 0 0 0 0 0 ⎥ 1 x 3i2 1 x 4i2 1 x 5i2 1 x 6i2
⎢ ⎥
⎣ 0 0 a73i 0 0 0 0 0 ⎦ + + + +
2 Cfi 2 Cfi 2 ο pi 2 οqi
0 0 0 a84i 0 0 0 0 
1 x 7i2 1 x 8i2
+ + (24)
= −[J di (x i )]T (22) 2 L vi 2 L vi

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
KHEFIF et al.: GENERALIZED IDA-PBC CONTROL USING ENHANCED DECOUPLED POWER SHARING 5077

∂ Ha (x i )
K (x) = Lasalle’s theorem
∂ xi ⎡ a11i  ⎤
 2
1 ∗ 1 ∗ 1 ∗ 1 ∗ − x 1i − x 1i∗
= − x 1i , − x 2i , − x 3i , − x , ⎢ L fi ⎥
L fi L fi Cfi C f i 4i ⎢ a22i  ⎥
⎢− − x 2i∗
2
⎥
 ⎢ L f i x 2i ⎥
1 1 1 ∗ 1 ∗ T ⎢ a  ⎥
− x 5i∗ , − x 6i∗ , − x 7i , − x 8i (25) ⎢ 33i ∗ 2 ⎥
ο pi οqi L vi L vi ⎢− x − x i3i ⎥
⎢ C f i 3i ⎥
⎢ a44i  ⎥
⎢− − x 4i ⎥
∗ 2
⎢ C x 4i ⎥
where K (x) = [K 1 , K 2 , K 3 , K 4 , K 5 , K 6 , K 7 , K 8 ]T −[∇ Hd (x i )] Rd (x i )∇ Hd (x i ) = ⎢
T fi ⎥≤
⎢ a55i  ∗ 2 ⎥
⎢− x − x 5i ⎥
⎢ ο pi 5i ⎥
∂ K 1 (x i ) ∂ K 2 (x i ) ∂ K 3 (x i ) ∂ K 4 (x i ) ⎢ a66i  ⎥
= 0; = 0; = 0; =0 ⎢− ∗ 2 ⎥
− x 6i ⎥
∂ xi j ∂ xi j ∂ xi j ∂ xi j ⎢ x 6i
⎢ οqi ⎥
∂ K 5 (x i ) ∂ K 6 (x i ) ∂ K 7 (x i ) ∂ K 8 (x i ) ⎢ a77i  2 ⎥
= 0; = 0; = 0; =0 ⎢− − x 7i∗ ⎥
⎢ L vi x 7i ⎥
∂ xi j ∂ xi j ∂ xi j ∂ xi j ⎣ a77i  ⎦
∗ 2
(26) − x 8i − x 8i
L vi
(31)
where j = {1, . . . , 8}.
This proves that the second condition of integrability where a ppi ≥ 0. This further proves that the equilibrium point
∇ K i (x i ) = [∇ K i (x i )]T is verified. is an invariant point and the chosen Hdi (x) ensures that the
solution is bounded.
2) Equilibrium Assignment: Using (25) and considering the
To better show the stability of the system, eigenvalues
equilibrium point x i∗ = (x 1i∗ , x 2i∗ , x 3i∗ , x 4i∗ , x 5i∗ , x 6i∗ , x 7i∗ , x 8i∗ ), the
of the system have been studied in what follows. For that,
following expression (27) is obtained:
the whole microgrid system is modeled. This small-signal
 model corresponds to the linearization of the system model
 ∂ Ha x i∗
K x i∗ = around its operating point obtained by calculating its Jacobian
∂ xi matrix. This technique presented in [41] is mathematically

1 ∗ 1 ∗ 1 ∗ 1 ∗ formulated by building a single model: the DGs and their
= − x 1i , − x 2i , − x 3i , − x ,
L fi L fi Cfi C f i 4i controls, the interconnections between DGs, and the supplied
 used local load.
1 1 1 ∗ 1 ∗
− x 5i∗ , − x 6i∗ , − x 7i , − x . (27) The small-signal model of the studied MG is presented
ο pi οqi L vi L vi 8i
using the following state-space model:
From another side x˙mg = Amg x mg (32)

 1 ∗ 1 ∗ 1 ∗ 1 ∗ where Amg represents the global matrix of all the MG in small
−∇ H x i∗ = − x ,− x ,− x ,− x
L f i 1i L f i 2i C f i 3i C f i 4i signal and x mg represents their corresponding state variables.
 By solving (33), the eigenvalues of the studied MG can
1 1 1 ∗ 1 ∗
− x 5i∗ , − x 6i∗ , − x 7i , − x 8i . (28) be obtained. These values represent the solution of the char-
ο pi οqi L vi L vi
acteristic equation that characterizes which is related to the
different frequencies and damping existing in the studied
From (27) and (28), it can be noted that the equilibrium system allowing to prove and study the stability of the system
condition 3 is well verified. 
3) Lyapunov Stability: det Amg − λI = 0. (33)
 The resulting eigenvalue spectrum of the studied system
 1 ∗ 1 ∗ 1 ∗ 1 ∗
K x i∗ = − x 1i , − x 2i , − x 3i , − x is presented in Fig. 5. It can be seen that the obtained
L fi L fi Cfi C f i 4i eigenvalues are depending on the frequency of the system.

1 1 1 ∗ 1 ∗ Indeed, the high-frequency ones corresponding to the eigen-
− x 5i∗ , − x 6i∗ , − x 7i , − x
ο pi οqi L vi L vi 8i values presented in cluster “III” are sensitive to the state
= [0, 0, 0, 0, 0, 0, 0, 0] (29) variables of LCL filter block of the used inverters and its
 ∗
 ∂ 2
H d x corresponding line currents. The eigenvalues in cluster “II”
−∇x2i Hd x i∗ = − i
are largely sensitive to the state variables of the output filter,
∂ x i2
  voltage controller, and current controller. The low-frequency
2 2 1 1 2
=− + + + + < 0 (30) dominant relative to the eigenvalues regrouped in cluster “I”
L fi C f i ο pi οqi L vi is largely sensitive to the state variables of the power-sharing
controller. This obtained result illustrates that by using the
where L f i , C f i , ο pi ,οqi , and L vi are positive coefficients. proposed control, the obtained eigenvalues of the closed-loop
4) Invariance of the Equilibrium Point: The fifth condition system are characterized by negative real parts, which proves
is to verify the invariance of the equilibrium point using that the proposed control ensures the closed-loop stability

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
5078 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO. 4, AUGUST 2021

the trigonometric function, respectively, and μ p and μq are the

control coefficients that set the concavity of the trigonometric
function, respectively.
Kpfmax and KIfmax are, respectively, the proportional and
the integral coefficient related to the frequency and voltage of
virtual power sag limit control [19].
The parameters of the inner and the outer loop of the
compared control technique are calculated by applying the
pole placement technique. (k pV = 2ξ C f ωV , ki V = C f ω2V )
and (k pC = 2ξ L f ωC , kiC = L f ωC2 ), where ξV and ωV are
the desired damping factor and bandwidth of the voltage loop
and ξC and ωC are separately the desired damping factor and
bandwidth of the current loop, respectively. In order to escape
Fig. 5. Complete eigenvalues of the studied microgrid systems. the interaction issues between the two loops, the damping
of the system. Moreover, it can be seen that applied to factor ξ is set equal to 0.7. whereas the bandwidth of the
the studied system, the proposed control guarantees a large current loop (ωC ) is set to 2000 rad/s, ten times higher than
stability margin with good decoupling between the eigenvalues that of the voltage loop (ωV ).
allowing to better guarantee the stability of the system.
V. C OMPARATIVE E XPERIMENTAL R ESULTS
IV. C OMPARATIVE C ONTROLLER
The effectiveness of the proposed control structure is veri-
To evaluate the performance of the proposed controller, fied on a lab-scale MG constituted of three DGs controlled by
a comparison test with the DTS droop control is per- using a dSPACE 1007 rapid prototyping. The experimental
formed [19]. In this technique, the voltage regulation loop is test bench is shown in Fig. 6. The performances of the
composed of two nested loops based on the conventional PI proposed modular IDA-PBC-E2D control strategy are verified
controller. using extensive comparisons with the PI-DTS control strategy
The control laws of the inner and the outer loop are detailed in Section IV. The corresponding control and circuit
expressed on the following expression: parameters are listed in the Appendix.

i d∗ = G V (s) v Cd − v Cd∗
+ i Ld − C f ωv Cq
∗
 ∗
i q = G V (s) v Cq − v Cq + i Lq + C f ωv Cd A. RL Load

v d∗ = G C (s) i d∗ − i d + R f i d − L f ωi q + v Cd A thorough comparative study has been conducted in the

v q∗ = G C (s) i q∗ − i q + R f i q + L f ωi d + v Cq (34) lab-scale MG, shown in Fig. 6, in order to evaluate the
performances of the proposed IDA-PBC-E2D compared to the
where G V (s) = k pV +(ki V /s) and G C (s) = k pC + (kiC /s) PI-DTS in terms of power-sharing performances and control
present, respectively, the PI transfer functions of the outer and flexibility. The resulting active and reactive power participation
inner loops. The decoupled trigonometric droop control laws of each DG and its corresponding current waveforms is shown,
are expressed as follows: respectively, in Figs. 8 and 9. In this test, the performances
⎡ ⎤
  2δ f 2δ f   of the proposed IDA-PBC-E2D are evaluated and compared
⎢ π − (arctan(μ p (P ∗ − P)))
f
= ⎣ 2δ π ⎥ to the PI-DTS when the MG is supplying a resistive inductive
E E 2δ E ⎦ (arctan(μq (Q ∗ − Q))) linear load of 3.2 kW and 2.2 kVAr as it is shown in Fig. 7.
π  π    ∗ The experimental test has been conducted in three main steps
f f f where, in each step, an additional DG unit is connected
+ + + (35)
E max E min E∗ to test the control performances of the compared methods
with and their efficacy in maintaining stable performances and
⎡ ⎤ accurate power sharing between DGs. The RL test considers
  Kpfmax + KIfmax  
f ⎢ ⎥ Pmax P the following scenario.
= ⎣ KpV + s
E max KIVmax ⎦ Q max Q (36)
1) From t = 0 to t = 4s: Only DG1 supplies the RL load
max
s 2) From t = 4 to 12 s: Both DG1 and DG2 supply the RL
and load.
⎡ ⎤ 3) From t = 12 to 20 s : DG1, DG2, and DG3 supply the
  Kpfmin + KIfmin  
f ⎢ s ⎥ Pmin P RL load.
=
E min ⎣ KpVmin + KIVmin ⎦ Q min Q
(37)
As it can be appreciated from Fig. 8, the proposed
s IDA-PBC-E2D method shows almost similar power-sharing
where E and f are, respectively, the operating voltage and performances compared to the PI-DTS method. However,
frequency of the inverter, f ∗ and E ∗ are, respectively, the fre- the PI-DTS method presents considerable power oscillations
quency and voltage references, δ f and δ E denote the coeffi- compared to the proposed method. These oscillations increase
cients that fix the variation limits of voltage and frequency of when interconnecting more DG units.

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
KHEFIF et al.: GENERALIZED IDA-PBC CONTROL USING ENHANCED DECOUPLED POWER SHARING 5079

Fig. 6. Experimental test bench of the studied system.

Fig. 7. Configuration of the proposed experimental test.

Fig. 9. Experimental comparison between the proposed IDA-PBC-E2D

controller (bottom) and the PI-DTS controller before (top) and after (middle)
resetting: DG#1, DG#2, and DG#3 active and reactive power sharing for the
RL load scenario.

TABLE I
V OLTAGE THD R ATES —RL L OAD

Fig. 8. Experimental comparison between the proposed IDA-PBC-E2D

controller (bottom) and the PI-DTS controller before (top) and after (middle)
resetting: DG#1, DG#2, and DG#3 active and reactive power sharing for the
RL load scenario. adapted for each configuration. On the other side, the proposed
IDA-PBC-E2D achieves accurate power sharing with fewer
The output current waveforms illustrated in Fig. 9 show oscillations even when interconnecting multiple DG units. This
that the proposed IDA-PBC-E2D method achieves accurate confirms the modular functionality of the proposed IDA-PBC-
current sharing compared to the PI-DTS method that presents E2D technique that can be designed to include multiple DGs,
some current-sharing inaccuracy that can introduce circulating which is of importance to preserve the scalability of MGs.
currents.
The measured THD rates are summarized in Table I for
the different test steps. The proposed IDA-PBC-E2D method B. Nonlinear Load
ensures satisfactory harmonic performances by maintaining the
total voltage harmonic distortions in small rate around 0.62% In this section, the effectiveness of the proposed IDA-PBC-
independently of the MG configuration. This is not the case E2D control strategy is verified when the MG is supplying
when applying the PI-DTS controller, which presents a high the nonlinear load shown in Fig. 10. This load consists of a
THD rates with large variations that go from 1.2% to 1.52% three-phase diode rectifier with 20% of current harmonics.
depending on the number of the used DG. The obtained The test consists of the following scenario.
results demonstrate the superiority of the proposed IDA-PBC- 1) From t = 0 to t = 2 s (Z1): Only DG1 supplies the
E2D in maintaining stable performances and accurate power load (3.3 kW and 300 VAr).
sharing compared to the PI-DTS control that shows some 2) From t = 2 to 8 s (Z2): Both DG1 and DG2 supply the
weakness to accommodate MG expansion as it needs to be load (3.3 kW and 300 VAr).

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
5080 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO. 4, AUGUST 2021

Fig. 10. Nonlinear load.

Fig. 11. Experimental comparison between the proposed IDA-PBC-E2D con-

troller (right) and the PI-DTS controller (left): DG#1, DG#2, and DG#3 active
and reactive power sharing for the nonlinear load.

3) From t = 8 to 14 s (Z3): Only DG1 and DG2 supply

the load that increases from (3.3 kW and 300 VAr) to
(6 kW and 600 VAr).
4) From t = 14 to 20 s (Z4): DG1, DG2, and DG3 supply
the load (6 kW and 600 VAr).
The obtained active and reactive powers are shown
in Fig. 11 to compare between the proposed IDA-PBC-E2D
control technique and the PI-DTS. It can be noted that when
supplying the nonlinear load, both control techniques show
similar active and reactive power-sharing performances. How-
ever, the proposed IDA-PBC-E2D offers higher performances Fig. 12. Experimental comparison between the proposed IDA-PBC-E2D con-
troller (right) and the PI-DTS controller (left): DG#1, DG#2, and DG#3 output
characterized by lower oscillations than the PI-DTS under currents (Phase-a), and the PCC voltage under a nonlinear load when one DG
nonlinear load changes either when two or three DGs are (Z1), two DGs (Z2), additional nonlinear step load (Z3), and three DGs (Z4)
connected. are supplying the system.
The obtained results confirm the advantages of the proposed TABLE II
control in maintaining the system stability even under MG V OLTAGE THD R ATES —N ONLINEAR L OAD
expansion. The proposed control achieves stable and modular
control independently of the MG architecture, i.e., the number
of interconnected DG units and loads.
Moreover, Fig. 12 shows the waveforms of the PCC voltage
and the output current of each DG unit phase “a.” It can be
appreciated from this figure that the proposed control strategy
allows maintaining accurate load current sharing between the
interconnected DGs. By following the obtained voltage, THD the PI-DTS. Moreover, it is shown that the proposed method
rates are summarized in Table II. It can be seen that the allows maintaining stable performances even when additional
proposed IDA-PBC-E2D allows obtaining better performances DGs and loads are connected to the MG, which confirms the
than the PI-DTS, which can be noticed when comparing the modular functionality that characterizes the proposed control
measured voltage THD rates. Indeed, the superiority of the technique. Indeed, the design methodology of the proposed
proposed method is highlighted as it obtained smaller THD control strategy guarantees the stability of the system as
rates between 3.14% and 3.58% than the PI-DTS that obtained well as its modularity even under system reconfiguration and
values between 4.17% and 5.91%. Note that the maximum expansion, which sustains the plug and play characteristic in
allowable voltage THD value is fixed to 5% by the IEEE 1547- MGs.
2018 Standard [42].
Following the two experiment tests, it can be concluded VI. C ONCLUSION
that the proposed IDA-PBC-E2D control ensures better power This article has proposed a modular control strategy based
quality and more accurate power-sharing performances than on passivity to guarantee proper voltage control and accurate

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
KHEFIF et al.: GENERALIZED IDA-PBC CONTROL USING ENHANCED DECOUPLED POWER SHARING 5081

power sharing in islanded MGs with multiple DG units. This [10] Y. Du, X. Lu, J. Wang, B. Chen, H. Tu, and S. Lukic, “Dynamic
control is based on an interesting modeling formalism called microgrids in resilient distribution systems with reconfigurable cyber-
physical networks,” IEEE J. Emerg. Sel. Topics Power Electron., early
the “PCH” that permits not only to synthesize a stable control access, Mar. 19, 2020, doi: 10.1109/JESTPE.2020.2981921.
but also to extend the study to a system with several DG [11] P. Zhang, H. Zhao, H. Cai, J. Shi, and X. He, “Power decoupling
units due to the modularity of the used energetic model. The strategy based on ’virtual negative resistor’for inverters in low-voltage
microgrids,” IET Power Electron., vol. 9, no. 5, pp. 1037–1044, 2016,
system modeling in the Hamiltonian form and the controller doi: 10.1049/iet-pel.2015.0137.
synthetizer are comprehensively detailed. Extensive compar- [12] U. B. Tayab, M. A. B. Roslan, L. J. Hwai, and M. Kashif, “A review of
isons between the proposed E2D control using the IDA-PBC droop control techniques for microgrid,” Renew. Sustain. Energy Rev.,
vol. 76, pp. 717–727, Sep. 2017, doi: 10.1016/j.rser.2017.03.028.
and an advanced control technique based on a decoupled droop [13] B. K. Unnikrishnan, M. S. Johnson, and E. P. Cheriyan, “Small signal
control are presented. The carried experimental tests show that stability improvement of a microgrid by the optimised dynamic droop
control method,” IET Renew. Power Gener., vol. 14, no. 5, pp. 822–833,
the proposed approach ensures satisfactory voltage quality and Apr. 2020, doi: 10.1049/iet-rpg.2019.0428.
accurate active and reactive power sharing even under system [14] Z. Peng et al., “Droop control strategy incorporating coupling com-
expansion, which confirms the modular functionality of the pensation and virtual impedance for microgrid application,” IEEE
Trans. Energy Convers., vol. 34, no. 1, pp. 277–291, Mar. 2019,
proposed control technique. doi: 10.1109/TEC.2019.2892621.
[15] A. Saim, A. Houari, M. A. Ahmed, A. Djerioui, M. Machmoum,
and J. M. Guerrero, “Adaptive reference trajectory for power qual-
A PPENDIX ity enhancement in three-phase four-wire standalone power supply
systems with nonlinear and unbalanced loads,” IEEE J. Emerg. Sel.
System Parameters: L f i = 3 mH,r f i = 0.1 ,C f i = Topics Power Electron., vol. 8, no. 2, pp. 1593–1603, Jun. 2020,
44 μF, L L1 = 1 mH, L L2 = 2 mH, U DCi = 450 V, F s = doi: 10.1109/JESTPE.2020.2966923.
[16] A. Haddadi, A. Yazdani, G. Joos, and B. Boulet, “A gain-
10 kHz, and Rated voltage : 110 Vrms, 60 Hz. scheduled decoupling control strategy for enhanced transient per-
Control Parameters: a11i = a22i = 1,a33i = a44i = formance and stability of an islanded active distribution network,”
1, a 13i = −a 31i = 1, a 24i = −a 42i = 1, a 55i = a66i = IEEE Trans. Power Del., vol. 29, no. 2, pp. 560–569, Apr. 2014,
doi: 10.1109/TPWRD.2013.2278376.
0, a 77i = a88i = 0, k pV = 0.024, k i V = 2.82,a37i = [17] Y. Xu, Q. Guo, H. Sun, and Z. Fei, “Distributed discrete
−a 73i = 0,, a 48i = −a 84i = 0, λi = 8,ξi = −8,K pi = robust secondary cooperative control for islanded microgrids,” IEEE
Trans. Smart Grid, vol. 10, no. 4, pp. 3620–3629, Jul. 2019,
K qi = 1e−5,kiC = 16922, and k pC = 14.15. doi: 10.1109/TSG.2018.2833100.
[18] T. Wu, Z. Liu, J. Liu, S. Wang, and Z. You, “A unified virtual power
decoupling method for droop-controlled parallel inverters in microgrids,”
R EFERENCES IEEE Trans. Power Electron., vol. 31, no. 8, pp. 5587–5603, Aug. 2016,
doi: 10.1109/TPEL.2015.2497972.
[1] V. Krishnamurthy and A. Kwasinski, “Effects of power electronics, [19] A. E. M. Bouzid, P. Sicard, H. Chaoui, A. Cheriti, M. Sechilariu,
energy storage, power distribution architecture, and lifeline dependencies and J. M. Guerrero, “A novel decoupled trigonometric saturated
on microgrid resiliency during extreme events,” IEEE J. Emerg. Sel. droop controller for power sharing in islanded low-voltage micro-
Topics Power Electron., vol. 4, no. 4, pp. 1310–1323, Dec. 2016, grids,” Electr. Power Syst. Res., vol. 168, pp. 146–161, Mar. 2019,
doi: 10.1109/JESTPE.2016.2598648. doi: 10.1016/j.epsr.2018.11.016.
[2] N. Khefifi, A. Houari, Mohamed. Machmoum, and M. Ghanes, “Inter- [20] J. Lai, X. Lu, X. Yu, and A. Monti, “Stochastic distributed sec-
connection and damping assignment passivity for the control of ondary control for AC microgrids via event-triggered communication,”
PV/battery hybrid power source in islanded microgrid,” Int. J. Renew. IEEE Trans. Smart Grid, vol. 11, no. 4, pp. 2746–2759, Jul. 2020,
Energy Res. (IJRER), vol. 9, no. 4, pp. 1790–1802, Dec. 2019. doi: 10.1109/TSG.2020.2966691.
[21] H. R. Baghaee, M. Mirsalim, G. B. Gharehpetan, and H. A. Talebi,
[3] V. Mohan, R. Suresh, J. G. Singh, W. Ongsakul, and N. Madhu,
“Nonlinear load sharing and voltage compensation of microgrids based
“Microgrid energy management combining sensitivities, interval and
on harmonic power-flow calculations using radial basis function neural
probabilistic uncertainties of renewable generation and loads,” IEEE
networks,” IEEE Syst. J., vol. 12, no. 3, pp. 2749–2759, Sep. 2018,
J. Emerg. Sel. Topics Circuits Syst., vol. 7, no. 2, pp. 262–270, Jun. 2017,
doi: 10.1109/JSYST.2016.2645165.
doi: 10.1109/JETCAS.2017.2679030.
[22] H. R. Baghaee, M. Mirsalim, and G. B. Gharehpetian, “Power cal-
[4] F. Zandi, B. Fani, I. Sadeghkhani, and A. Orakzadeh, “Adaptive complex
culation using RBF neural networks to improve power sharing of
virtual impedance control scheme for accurate reactive power sharing
hierarchical control scheme in multi-DER microgrids,” IEEE J. Emerg.
of inverter interfaced autonomous microgrids,” IET Gener., Transmiss.
Sel. Topics Power Electron., vol. 4, no. 4, pp. 1217–1225, Dec. 2016,
Distrib., vol. 12, no. 22, pp. 6021–6032, Dec. 2018, doi: 10.1049/iet-
doi: 10.1109/JESTPE.2016.2581762.
gtd.2018.5123. [23] M. Hamzeh, S. Emamian, H. Karimi, and J. Mahseredjian, “Robust
[5] Z. Wang, W. Wu, and B. Zhang, “A distributed quasi-Newton method control of an islanded microgrid under unbalanced and nonlinear load
for droop-free primary frequency control in autonomous microgrids,” conditions,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 4, no. 2,
IEEE Trans. Smart Grid, vol. 9, no. 3, pp. 2214–2223, May 2018, pp. 512–520, Jun. 2016, doi: 10.1109/JESTPE.2015.2459074.
doi: 10.1109/TSG.2016.2609422. [24] M. Pichan and H. Rastegar, “Sliding-mode control of four-leg inverter
[6] R. Bhattarai, N. Gurung, and S. Kamalasadan, “Dual mode control of with fixed switching frequency for uninterruptible power supply appli-
a three-phase inverter using minimum variance adaptive architecture,” cations,” IEEE Trans. Ind. Electron., vol. 64, no. 8, pp. 6805–6814,
IEEE Trans. Ind. Appl., vol. 54, no. 4, pp. 3868–3880, Jul. 2018, Aug. 2017, doi: 10.1109/TIE.2017.2686346.
doi: 10.1109/TIA.2018.2826469. [25] M. Shahparasti, M. Mohamadian, A. Yazdian, A. A. Ahmad, and
[7] Y. Kwon, A. Kwasinski, and A. Kwasinski, “Coordinated energy M. Amini, “Derivation of a stationary-frame single-loop controller
management in resilient microgrids for wireless communication net- for three-phase standalone inverter supplying nonlinear loads,” IEEE
works,” IEEE J. Emerg. Sel. Topics Power Electron., vol. 4, no. 4, Trans. Power Electron., vol. 29, no. 9, pp. 5063–5071, Sep. 2014,
pp. 1158–1173, Dec. 2016, doi: 10.1109/JESTPE.2016.2614425. doi: 10.1109/TPEL.2013.2287906.
[8] X. Hou, Y. Sun, H. Han, Z. Liu, W. Yuan, and M. Su, “A fully [26] S. J. Williamson, A. Griffo, B. H. Stark, and J. D. Booker, “A controller
decentralized control of grid-connected cascaded inverters,” IEEE for single-phase parallel inverters in a variable-head pico-hydropower
Trans. Sustain. Energy, vol. 10, no. 1, pp. 315–317, Jan. 2019, off-grid network,” Sustain. Energy, Grids Netw., vol. 5, pp. 114–124,
doi: 10.1109/TPWRD.2018.2816813. Mar. 2016, doi: 10.1016/j.segan.2015.11.006.
[9] H. Han, X. Hou, J. Yang, J. Wu, M. Su, and J. M. Guerrero, “Review [27] Z. Li, C. Zang, P. Zeng, H. Yu, S. Li, and J. Bian, “Control of a grid-
of power sharing control strategies for islanding operation of AC micro- forming inverter based on sliding-mode and mixed H2 /H∞ control,”
grids,” IEEE Trans. Smart Grid, vol. 7, no. 1, pp. 200–215, Jan. 2016, IEEE Trans. Ind. Electron., vol. 64, no. 5, pp. 3862–3872, May 2017,
doi: 10.1109/TSG.2015.2434849. doi: 10.1109/TIE.2016.2636798.

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.
5082 IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN POWER ELECTRONICS, VOL. 9, NO. 4, AUGUST 2021

[28] W. Lu, K. Zhou, D. Wang, and M. Cheng, “A general parallel structure Azeddine Houari received the Engineering degree
repetitive control scheme for multiphase DC–AC PWM converters,” in electrical engineering from the University of
IEEE Trans. Power Electron., vol. 28, no. 8, pp. 3980–3987, Aug. 2013, Bejaia, Béjaïa, Algeria, in 2008, and the Ph.D.
doi: 10.1109/TPEL.2012.2229395. degree in electrical engineering from the University
[29] S. Jiang, D. Cao, Y. Li, J. Liu, and F. Z. Peng, “Low-THD, fast- of Lorraine, Lorraine, France, in 2012.
transient, and cost-effective synchronous-frame repetitive controller for Since 2014, he has been an Assistant Professor
three-phase UPS inverters,” IEEE Trans. Power Electron., vol. 27, no. 6, with the University of Nantes, Saint-Nazaire, France,
pp. 2994–3005, Jun. 2012, doi: 10.1109/TPEL.2011.2178266. where he carried out research activities at the Insti-
[30] H. T. Nguyen, E.-K. Kim, I.-P. Kim, H. H. Choi, and tutde Recherche en Energie Electrique de Nantes
J.-W. Jung, “Model predictive control with modulated optimal Atlantique (IREENA). His current research deals
vector for a three-phase inverter with an LC filter,” IEEE Trans. with the power quality and the stability issues in
Power Electron., vol. 33, no. 3, pp. 2690–2703, Mar. 2018, stationary and embedded dc and ac microgrids.
doi: 10.1109/TPEL.2017.2694049.
[31] J. Zhang, L. Li, D. G. Dorrell, M. Norambuena, and J. Rodriguez,
“Predictive voltage control of direct matrix converters with improved
output voltage for renewable distributed generation,” IEEE J. Emerg.
Sel. Topics Power Electron., vol. 7, no. 1, pp. 296–308, Mar. 2019,
doi: 10.1109/JESTPE.2018.2874275. Mohamed Machmoum received the Engineering
[32] S. Pang, B. Nahid-Mobarakeh, S. Pierfederici, Y. Huangfu, degree from the Institut Supérieur Industriel, Liège,
G. Luo, and F. Gao, “Towards stabilization of constant power loads Belgium, in 1984, and the master’s and Ph.D.
using IDA-PBC for cascaded LC filter DC/DC converters,” IEEE degrees from the l’Institut National Polytechnique
J. Emerg. Sel. Topics Power Electron., early access, Oct. 3, 2019, de Lorraine (INPL), Lorraine, France, in 1985 and
doi: 10.1109/JESTPE.2019.2945331. 1989, respectively, all in electrical engineering.
[33] R. Ortega, M. W. Spong, F. Gomez-Estern, and G. Blankenstein, “Sta- In 1991, he joined the École Polytechnique de
bilization of a class of underactuated mechanical systems via intercon- l’Université de Nantes, Nantes, France, as an Assis-
nection and damping assignment,” IEEE Trans. Autom. Control, vol. 47, tant Professor, where he has been a Full Professor
no. 8, pp. 1218–1233, Aug. 2002, doi: 10.1109/TAC.2002.800770. since September 2005. He is currently the Head of
[34] S. Sanchez, R. Ortega, R. Grino, G. Bergna, and M. Molinas, “Condi- the Institut de Recherche en Energie Electrique de
tions for existence of equilibria of systems with constant power loads,” Nantes Atlantique (IREENA), University of Nantes, Saint-Nazaire, France.
IEEE Trans. Circuits Syst. I, Reg. Papers, vol. 61, no. 7, pp. 2204–2211, His main area of interest includes power electronics, power quality, wind and
Jul. 2014, doi: 10.1109/TCSI.2013.2295953. tidal energy conversion systems, and microgrids. In these fields, he has super-
[35] N. Khefifi, A. Houari, M. Machmoum, M. Ghanes, and M. Ait-Ahmed, vised more than 25 Ph.D. students and published one book related to modeling
“Control of grid forming inverter based on robust IDA-PBC for power of electrical machines, several chapter books, and nearly 240 journals and
quality enhancement,” Sustain. Energy, Grids Netw., vol. 20, Dec. 2019, conferences in his areas of expertise.
Art. no. 100276, doi: 10.1016/j.segan.2019.100276.
[36] R. V. Meshram, M. Bhagwat, S. Khade, S. R. Wagh,
A. M. Stankovic, and N. M. Singh, “Port-controlled phasor Hamiltonian
modeling and IDA-PBC control of solid-state transformer,” IEEE Trans.
Control Syst. Technol., vol. 27, no. 1, pp. 161–174, Jan. 2019,
doi: 10.1109/TCST.2017.2761866. Abdelhakim Saim received the B.S. and M.S.
[37] M. Hilairet, M. Ghanes, O. Béthoux, V. Tanasa, J.-P. Barbot, and degrees in electronics and control engineering from
D. Normand-Cyrot, “A passivity-based controller for coordination of Blida University, Blida, Algeria, in 2010 and 2012,
converters in a fuel cell system,” Control Eng. Pract., vol. 21, no. 8, respectively, and the Ph.D. degree in control engi-
pp. 1097–1109, Aug. 2013, doi: 10.1016/j.conengprac.2013.04.003. neering from the University of Tizi Ouzou, Tizi
[38] J. M. Guerrero, L. Hang, and J. Uceda, “Control of distributed unin- Ouzou, Algeria, in 2017.
terruptible power supply systems,” IEEE Trans. Ind. Electron., vol. 55, Since 2017, he has been an Assistant Professor
no. 8, pp. 2845–2859, Aug. 2008, doi: 10.1109/TIE.2008.924173. with the Department of Control and Instrumenta-
[39] A. Zurfi and J. Zhang, “Investigation of the line frequency for demand- tion, University of Sciences and Technology Houari
side primary frequency control using behind-the-meter home batteries,” Boumediene, Algiers, Algeria. He is currently with
Int. J. Renew. Energy Res. (IJRER), vol. 8, no. 2, pp. 784–796, Jun. 2018. the Institut de Recherche en Energie Electrique de
[40] C. Li, S. K. Chaudhary, M. Savaghebi, J. C. Vasquez, and J. M. Guerrero, Nantes Atlantique (IREENA), University of Nantes, Saint-Nazaire, France.
“Power flow analysis for low-voltage AC and DC microgrids considering His research deals with the power quality and stability of microgrids.
droop control and virtual impedance,” IEEE Trans. Smart Grid, vol. 8,
no. 6, pp. 2754–2764, Nov. 2017, doi: 10.1109/TSG.2016.2537402.
[41] N. Pogaku, M. Prodanovic, and T. C. Green, “Modeling, analysis
and testing of autonomous operation of an inverter-based microgrid,”
IEEE Trans. Power Electron., vol. 22, no. 2, pp. 613–625, Mar. 2007,
doi: 10.1109/TPEL.2006.890003. Malek Ghanes received the M.Sc. and Ph.D.
[42] IEEE Standard for Interconnection and Interoperability of Distributed degrees in applied automatic and informatics from
Energy Resources with Associated Electric Power Systems Interfaces, IRCCyN (now LS2N), Centrale Nantes (CN),
IEEE Standard 1547-2018 (Revision of IEEE Standard 1547-2003), Nantes, France, in 2002 and 2005, respectively.
Apr. 2018, pp. 1–138, doi: 10.1109/IEEESTD.2018.8332112. From September 2005 to September 2006, he held
a post-doctoral position at GREYC Laboratory.
From September 2006 to April 2016, he was with
the ECS-Lab, Quartz, ENSEA, Cergy, France, where
he was an Associate Professor and the Head of
the Department of Automatic and Electrical Engi-
Nidhal Khefifi received the Electrical Engineer- neering. In May 2016, he was promoted as a Full
ing degree from the École nationale supérieure Professor at LS2N-ECN. He has published more than 100 journals articles and
d’ingénieurs de Tunis (ENSIT), Tunis, Tunisia, conference papers. He holds several patents. His research interests include
in 2016, and the M.S. degree in information and observation and control of nonlinear systems with applications to electric
system sciences from Aix-Marseille Université, Mar- systems.
seille, France, in 2017. He is currently pursu- Prof. Ghanes was a recipient of the Best Paper Award from the Control
ing the Ph.D. degree with a focus on the power Engineering Practice (CEP) journal in 2013 and the Applied Research Award
quality and stability of microgrids with the Insti- from the Federation of Electronic, Electrical and Communications Industries
tutde Recherche en Energie Electrique de Nantes (FIEEC) in 2015. He is currently holding the Chair on Electric Vehicle
Atlantique (IREENA), University of Nantes, Nantes, performances between Renault and CN. He has served as an Associate Editor
France. for the IFAC CEP journal from 2012 to 2020.

Authorized licensed use limited to: UNIVERSIDAD DE GUADALAJARA. Downloaded on March 20,2024 at 04:03:45 UTC from IEEE Xplore. Restrictions apply.