Electric Power Systems Research 191 (2021) 106900

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

Active resonance damping and harmonics compensation in distributed T

generation based islanded microgrids
Abdelhakim Saima,b, , Azeddine Houaria, Mourad Ait-Ahmeda, Mohamed Machmouma,
⁎

Josep M. Guerreroc
a
IREENA Laboratory, University of Nantes, 44600, Saint Nazaire, France
b
LSEI Laboratory, University of Sciences and Technology Houari Boumedien, 16111 Bab Ezzouar, Algeria
c
Center for Research on Microgrids (CROM), 9220 Aalborg, University, Aalborg, Denmark

ARTICLE INFO ABSTRACT

Keywords: The development of distributed generation based microgrids with high penetration of electronically interfaced
Distributed generation systems shows great interest in various applications. These systems commonly use LC-L filters with inherent
Resonance damping resonance characteristics, which increases the risk of resonance amplification and propagation in microgrids.
Harmonics compensation These resonances lead to current and voltage harmonics amplifications resulting in major power quality and
Power quality
stability issues. For this aim, an effective active resonance damping method is proposed to dampen out the
Islanded microgrid
undesired resonance amplifications. The proposed method uses a filter-based approach with inherent stability
characteristics to reconstruct and compensate for the undesired resonance harmonics. This method uses an
external control level that prevents control bandwidth limitation as it directly superposes the harmonic com­
pensating signal to the control signal. This method offers simplicity of design and implementation without the
need for additional measurements or prior knowledge of system parameters, which is of importance in scalable
microgrid applications. Experimental results and discussions are provided to evaluate the effectiveness of the
proposed strategy.

1. Introduction
The development of distributed generation (DG) based microgrids
(MGs) shows great promises to service a variety of stationary and em­
bedded loads [1]. These systems encourage on-site generation based on
low-emission energy technologies and a high share of distributed en­
ergy resources and energy storage systems [2]. Moreover, this concept
represents an economic and eco-friendly solution to extend the elec­
trical coverage to remote locations in villages, towns and cities without
the need to upgrade the existing electrical facilities [3]. These small-
scale grid configurations connect multiple DG units, storages and loads
to form single controllable entities [4]. Those entities can operate either
in grid-connected or in islanded mode in order to provide end-users
with reliable service even during contingencies [5].
In islanded mode of operation, these systems may present some
weaknesses to handle disturbances due to their reduced rotational in­
ertia and short-circuit power capacity [6]. This is mainly due to the
increasing integration of power electronics based DG units and loads
appliances with high share of intermittent energy resources [7]. More
specifically, these systems can face multiple disturbances either at low
or high frequency ranges [8]. Low frequency disturbances can be in­
troduced by load changes and sources intermittency [9]. Besides, high
frequency disturbances can appear as the circuit configuration of MGs is
prone to resonances phenomena [10]. These phenomena are due to the
use of high order passive filters with inherent resonance characteristics
such as LC and LCL filters [11]. These resonances result in harmonics
amplification, which increases current and voltage distortion [12].
Moreover, interactions could happen among the interconnected DG
units and the background system, which can result in multiple re­
sonances [13]. The risk of resonances is augmented by the increasing
usage of nonlinear and unbalanced electronic load appliances [14]. The
characteristic frequencies of these resonances may vary and additional
resonances may appear once more DG units and loads are connected to
the MG [15]. These resonances can propagate and affect the whole MG
as DG units and loads are coupled through small line impedances [16].
This situation represents an important matter in MGs since it affects
power quality, increases losses, damages electronics, and can even lead
to system instability [17].
The damping of such resonances is still a challenging task that re­
quires the development of advanced analysis and damping methods.

⁎
Corresponding author at: Université de Nantes - Laboratoire IREENA 37 Boulevard de l’Université BP 406 44602 SAINT-NAZAIRE CEDEX, France.
E-mail addresses: Abdelhakim.Saim@univ-nantes.fr, asaim@usthb.dz (A. Saim).

https://doi.org/10.1016/j.epsr.2020.106900
Received 1 July 2020; Received in revised form 11 September 2020; Accepted 18 October 2020
0378-7796/ © 2020 Published by Elsevier B.V.
A. Saim, et al.
Fig. 1. Typical microgrid configuration.
Several researches contributing to resonance analysis and damping in
MGs are proposed in the literature [18]. These methods use either
passive or active damping approaches. Passive damping approaches are
based on the filter circuit reconfiguration through the addition of en­
ergy dissipative elements [19, 20]. Although the passive damping
shows robust damping performances, the use of active damping is
preferred since it allies efficiency and flexibility, which is of importance
to achieve scalable MG functionality. The active damping uses either
filter-based or feedback-based control methods. The feedback based
methods consider the use of complementary control loops in order to
obtain a wider damping region [21]. These methods feedback either the
capacitor current through a high pass filter, or similarly the capacitor
voltage through a low pass filter to achieve the desired damping per­
formances [22]. Similarly, the filter inductor current can be used for
damping purpose [23]. Although these methods achieve satisfactory
damping performances, the control complexity is increased and addi­
tional measurements are required [24]. This results in extra costs and
increases the computational effort, which limits their practical im­
plementation.
Besides, the filter based methods incorporate additional filters to the
control system with the aim to mitigate the undesired dynamic [25].
These methods commonly use lead-lag compensators, high pass filters,
or notch filters achieving satisfactory damping performances [26].
However, these methods remain sensitive to parameters uncertainties
and their efficiency depends on the accuracy of the resonance frequency
estimate. Self-commissioning filters with complex structures have been
proposed in the literature to achieve adaptive damping solutions [27].
These methods can affect the control dynamic, which limits their ap­
plicability in MGs wherein hierarchical strategies with multi-levels
multi-loops controllers are usually adopted. In order to achieve re­
sonances damping and preserve the system dynamic, the introduction
of an external active damping control loop has been proposed in [28].
Although this method achieves suitable damping performances while
preserving the system dynamic, it involves a complex design procedure
with multi-parameters setting.
In this paper, a hierarchical control strategy with harmonics com­
pensation and active resonance damping is proposed to control islanded
MGs operation and dampen out the undesired resonance amplifications.
The proposed method uses a washout filter with inherent stability
characteristics to reconstruct and compensate for resonance harmonics.
It uses an external control level that superposes the resulting harmo­
nics-compensating signal directly on the control signal, which preserves
the dynamic of the inner control levels. This method offers simplicity of
design and implementation with the ability to achieve resonance
damping even for uncertain MG configurations. More specifically, and
considering the context of islanded MGs, this method presents the fol­
lowing advantages:
• Straightforward to design and implement without the need for ad­
ditional measurements.
• Preserves the dynamic of the inner control levels since it acts
2
Electric Power Systems Research 191 (2021) 106900
directly on the voltage control signal.
• Robust against system parameters uncertainties and frequency var­
iation, which is of importance when considering scalable MGs.
This paper is organized as follows; In Section 2, the modeling of
islanded MGs is described. Sections 3 describes the main control levels
of the proposed control strategy and active resonance damping method.
The performances of the proposed active resonance damping method
are analyzed in Section 4, while the system stability is assessed in
Section 5. Experimental test results and discussions are presented in
Section 6. Finally, Section 7 underlines the main contributions of this
work.
2. System description and modeling
2.1. System description
A typical low voltage MG configuration is shown in Fig. 1. It is
mainly composed of multiple DG units with distributed energy re­
sources and loads. Each DG unit consists of a locally controlled inverter
with an output LC-L filters. The LC-L filter is employed to filter out
switching harmonics and connect the DG inverter to the load bus.
2.2. System modeling
Considering the coupling among the DG units and the background
system, the development of a generalized model of MGs is necessary to
analyze the main resonance characteristics and instability issues in
MGs.
Firstly, given that the switching frequency is sufficiently high, each
DG unit is modeled based on the dynamic behavior of its corresponding
output LC-L filter as follows[32]:
1 L1, i s + L2, i s (Ci s 2 + 1)
vPCC = vinv, i iDG, i
L1, i Ci s 2 + 1 L1, i Ci s 2 + 1 (1)
where, vinv, i and vPCC represent the inverter voltage and the PCC voltage,
respectively, while iDG, i represents the output current. L1, i and L2, i denote
the inverter-side and the line-side filter inductances, respectively, while
Ci represents the filter capacitance.
Eq. (1) corresponds to two Thévenin terminals consisting of an
equivalent voltage source GDG, i vinv, i and an output impedance ZDG, i . The
equivalent representation of DG units is illustrated in Fig. 2(a),
vPCC = GDG, i vinv, i ZDG, i iDG, i (2)
Similarly, an equivalent load model is defined considering an LC
filter interfaced load.
1 CL . j s
iL . j = ij + vPCC
LL . j CL . j s 2 + 1 LL . j CL . j s 2 + 1 (3)
where, LL . j and CL . j represent the load filter inductance and capacitance,
respectively. i j and iL . j represent the load side and the load filter side
A. Saim, et al. Electric Power Systems Research 191 (2021) 106900

Fig. 2. Equivalent representations: Distributed generators (a), Loads (b), Microgrid (c).

current, respectively. and resonant gains of the inner and the outer loop controllers; and 0 is
As illustrated in Fig. 2(b), the load circuits can be represented with the angular frequency.
two Norton terminals consisting of an equivalent current source Note that, the use of an inner current control loop contributes to
GL, j i j and output admittance YL, j = 1/ZL, j . resonances attenuation. However, to achieve the desired attenuation,
the proportional gain kpc, i needs to be increased too much. This in fact
iL . j = GL . j i j + YL . j vPCC (4)
enhances the control dynamic and improves disturbances rejection, but
At this level, given the parallel representation shown in Fig. 2(c), can cause resonance interaction between DG units.
and based on (2) and (4). the MG PCC's voltage can be expressed as
follows [29]: 3.2. Level 2 : power sharing control
N M
GDG, i YDG, i vinv, i GL, j i j
vPCC =
i=1 j=1
The interconnection of multiple DG units can lead to large circu­
N M
i=1
YDG, i + j=1
YL, j (5) lating currents due essentially to the unbalanced system configuration
between DG units, line impedance mismatches and so on. The presence
where, YDG, i = 1/ZDG, i represents the ithDG unit admittance, while N and of these currents will increase power losses and damage MG compo­
M represent the number of DG units and loads, respectively. nents. Therefore, the control of the power participation of each DG unit
Eq. (5) shows that the circuit configuration of MGs presents sig­ is required to achieve accurate power sharing and avoid circulating
nificant coupling between the DG units and the background system. currents.
This situation increases the risk of resonance interactions and propa­
gation. For this aim, the development of advanced active resonance
3.2.1. Droop control loop
damping methods is important to preserve MGs power quality and
The second control level includes a well-established power sharing
stability.
method based on droop control [30]. Assuming that the DG unit output
impedance is predominantly inductive, the following droop control
3. Proposed active damping method
expressions are used to control the participation of each DG unit.

As shown in Fig. 3, the proposed control strategy adopts a hier­ 0,i = 0 mi Pi m d,i dP/dt
i (8)
archical control structure with multi-levels multi-loops control. The
first and second control levels include respectively a cascaded voltage- Ei = E0 ni Qi n d,i dQ/dt
i (9)
current controller and a communication-less power sharing controller.
Then, the voltage reference vDG
* , i is expressed as:
The third control level is proposed to mitigate resonance phenomena
and preserve the stability of islanded MGs. * , i = Ei sin (
vDG 0, i t) (10)

where, mi , ni , md, i and nd, i represent the proportional and derivative

3.1. Level 1 : voltage and current control
droop slope coefficients. Ei and 0, i denote the droped voltage amplitude
and frequency.
The output voltage vDG, i of each DG unit is controlled via the control
These expressions include a proportional term that is synthetized to
of the filter line current iinv, i .
improve the power sharing accuracy, whereas the derivative term in­
More specifically, a cascaded voltage and current controllers with
troduces a damping ratio that attenuates transient, avoids large starting
selective harmonics compensation is designed using multi-resonant
current oscillations and enhances the plug and play characteristic of the
controllers. The use of resonant control offers the possibility to achieve
corresponding DG unit.
selective harmonics compensation, which is important to enhance
system disturbances rejection capabilities. The voltage and current
control expressions are given as follows: 3.2.2. Virtual impedance loop
In low voltage MGs, the output impedance of DG units presents a
7
s complex behavior with low X/R ratio. This situation may strengthen the
Gc, i (s ) = kpc, i + kc, i
h = 1, odd
s 2 + ( 0 h) 2 (6) coupling between the active and reactive power and entail the efficacy
of the droop controller, expressed in (8) and (9). For this aim, a virtual
7
s impedance loop is added in order to enhance the inductive pre­
Gv, i (s ) = kpv, i + k v, i
s 2 + ( 0 h)2 dominance of the output impedance. The virtual impedance is enforced
h = 1, odd (7)
by dropping the ith DG voltage reference proportionally to the output
where kpc, i and kpv, i , kc and k v represent, respectively, the proportional current and its derivative.

3
A. Saim, et al. Electric Power Systems Research 191 (2021) 106900

Fig. 3. Proposed hierarchical control strategy with resonance active damping.

* , i = Ei sin (
vDG 0, i t) ZD, i iDG, i (11) frequencies, which allows dampening the undesired resonance ampli­
th
fication. A Washout filter with inherent stability and unity gain char­
where, ZD, i = RD, i + LD, i s , represents the i virtual impedance, with acteristics is used to reconstruct the undesired resonance harmonic
RD, i and LD, i the virtual resistance and inductance, respectively. disturbances [31]. The washout filter serves as a high pass filter that
passes unchanged the undesired high frequency harmonic components
3.3. Level 3 : proposed active damping method and blocks out low frequency components.
Considering the context of MGs, this method presents the following
The circuit configuration of islanded MGs is prone to experience advantages:
severe resonance phenomena. These phenomena result in harmonics
amplification, which affects power quality and increases the risk of • Dynamic: Preserves the dynamic of the inner controllers as it acts on
instability. Moreover, the amplification of harmonic disturbances re­ the control signal.
duces the efficiency and reliability of MGs since it results in the increase • Robustness: Does not require prior knowledge of system parameters
of power losses and damages MG components. Thus, and as shown in and frequencies.
Fig. 3, a third control level is added to the conventional DG control • Simplicity: Simple to design and implement using only the output
scheme in order to mitigate resonance harmonics amplification and voltage measure.
propagation. The proposed control level acts directly on the voltage
control signal in order to compensate for the resulting harmonic dis­ 4. Resonance damping performances
turbances and overcome issues related to resonance phenomena. The
proposed resonance damping control law is given in (12). 4.1. Resonance damping model
s
* , i = vinv, i
vinv kd vDG, i In order to evaluate the efficacy of the proposed damping method,
s+ w (12) the expression given in (2) and (5) are rearranged to integrate the
where vinv
* , i represents the resulting inverter control voltage, vDG, i is the proposed method. The damping model of DG units is obtained by
output voltage, and w is the washout filter cut-off frequency. The value substituting (12) in (2) given that vDG, i = vPCC + Z L2, i iDG, i as follows:
of w is non-critical as long as it is chosen sufficiently lower than the GDG, i ZDG, i + kd GDG, i W (s ) Z L2, i
switching frequency. The term kd is the damping coefficient that can be vPCC = vinv, i iDG, i
1 + kd GDG, i W (s ) 1 + kd GDG, i W (s ) (13)
used to scale up the damping ratio (see, Section 5).
From Eq. (12), it can be seen that the voltage control signal is where, W (s ) = s /(s + w ) represents the Washout filter transfer func­
dropped proportionally to the undesired voltage harmonic dis­ tion.
turbances. This introduces a resistive behavior around the resonance The obtained expression can be expressed as:

4
A. Saim, et al. Electric Power Systems Research 191 (2021) 106900

Fig. 4. Bode diagram: (a) voltage gain and (b) output impedance of as single distributed generator before (blue) and after damping for kd [1 10].

d
vPCC = GDG , i vinv, i
d
ZDG , i iDG, i (14) Similarly, the effectiveness of the proposed method is verified when
considering a typical MG configuration with two DG units and a
where, d
GDG , i and
d
ZDG , i represent, the
damped voltage gain and output common load. The obtained bode diagrams are illustrated in Fig. 5. It
th
impedance expressions of the i DG unit, respectively. can be seen that the MG is subject to two main resonance peaks with up
Similarly, the damping model of islanded MGs considering multiple to 130 dB of magnitude at 384 Hz and 1100 Hz, and a resonance peak
DG units and loads is obtained by substituting (12) in (5) as follows: due to the load circuit of 44 dB at 876 Hz. The introduction of the
N d M proposed damping method allows reducing the resonances peaks to less
GDG , i YDG, i vinv, i GL, j i j
vPCC =
i=1 j =1
than 0 dB. This damping method shows strong resonance mitigation
N d M
i=1
Y DG ,i + j=1
YL, j (15) performances even in presence of multiple resonances.
th
where d
Y DG ,i = d
1/ZDG , i represents the damped i DG unit admittance.
5. Stability analysis and parameters setting
4.2. Resonance damping evaluation
This section presents a general stability analysis in order to assess
the effectiveness of the proposed damping method and its capability for
The effectiveness of the proposed resonance damping method is
attenuating resonance amplifications considering different values
evaluated based on the frequency response of a single DG unit and a MG
of kd and w . This is done by examining how the roots change with the
with multiple DG units and loads. The obtained diagrams are illustrated
variation of the damping coefficient kd and the washout filter
in Figs. 4-5 using the parameters listed in the Appendix.
frequency w . For the sake of simplicity, only the current and voltage
As illustrated in Fig. 4a, the frequency response of a single DG unit
proportional control gains are considered since the resonant gains do
highlights the performances of the proposed damping for increasing
not affect the system dynamics, whereas the dynamic of the droop
values of the damping coefficient kd . This figure shows the presence of
controller is much slower than that of the inner control loops. The
a single resonance peak with 153 dB at 876 Hz that is significantly
closed-loop dynamics is derived by substituting the following control
reduced with the introduction of the proposed method to less than 0 dB.
law into (2).
The proposed method shows satisfactory damping performances alle­
viating the resonance peak introduced by the LC-Lfilter without af­ * ,i
vinv, i = ((vDG vDG, i ) Cv (s ) iinv, i ) Cc (s ) (16)
fecting its harmonics attenuation performances. The DG unit output
impedance shown in Fig. 4b tends to have a resistive behavior around being iinv, i = iDG, i + vDG, i / Zc, i the filter inductor current, and the filter
the resonance peak, which improves resonance harmonics attenuation. capacitor voltage vDG, i = Z L2,i iDG, i + vPCC ; Zc, i = 1/(Ci s ) and Z L2, i = L2, i s

Fig. 5. Bode diagram of the DG side transfer functions (vpcc / vinv,i ) (a) and the load side transfer function (vpcc / i j) (b) before (blue) and after damping for 0 kd 10 .

5
A. Saim, et al. Electric Power Systems Research 191 (2021) 106900

The closed loop transfer function is given as follows: the stable region as the washout frequency increases
The parameters of the proposed damping method can be adjusted
GDG, i Cc, i Cv, i
vPCC = * ,i
vDG flexibly to improve the system stability and achieve the desired
1 + GDG, i Cc , i (Cv, i Zc , i + 1)/ Zc, i
damping performances.
ZDG, i + GDG, i Cc , i ((Cv, i Zc , i + 1) Z L2,i + Zc, i )/ Zc, i
iDG, i
1 + GDG, i Cc, i (Cv, i Zc, i + 1)/Zc , i (17) 6. Experimental tests
The obtained equation can be simplified to:
An experimental lab-scale islanded MG with two parallel DG units
* ,i
vPCC = GCP, i vDG ZCP, i iDG, i (18) was built in the IREENA Laboratory in order to verify the effectiveness
of the proposed resonance active damping control strategy. The ex­
where, GCP, i and ZCP , i represent the closed loop voltage gain and output perimental test bench comprises two SEMIKRON inverters with similar
impedance. LC-L output filters supplying linear and rectifier type nonlinear loads.
Considering the damping method, the closed loop transfer function The proposed hierarchical control strategy is implemented on a dSPACE
becomes: Micro Auto Box control prototyping system. The implementation of the
d
GDG control requires 39.2 µs whereas the proposed damping method re­
, i Cc, i Cv, i
vPCC = d
* ,i
vDG quires 2.1 µs, based on 100 µs sampling.
1+ GDG, i Cc, i (Cv, i Zc, i + 1)/ Zc, i
d
ZDG , i + GDG, i Cc, i ((Cv, i Zc, i + 1) Z L2, i + Zc , i )/ Zc, i 6.1. Case 1 – linear load
d
iDG, i
1 + GDG , i Cc, i (Cv, i Zc, i + 1)/ Zc, i (19)
The efficacy of the proposed hierarchical control strategy is eval­
This equation can be simplified as: uated when both DG units are supplying a common linear load. The
d * ,i d load circuit is shown in Fig. 7. The measured voltage and current wa­
vPCC = GCP , i vDG ZCP , i iDG, i (20)
veforms are shown in Fig. 8 before and after damping. The comparison
being d
GCP , i and ,i,
the damped closed loop voltage gain and output
d
ZCP between the DG#1, DG#2 and the PCC voltage and current waveforms
impedance. shows similar resonance characteristics with high frequency harmonic
Based on (20), the system stability is evaluated using root locus distortions. Moreover, and as it can be appreciated from the harmonic
diagrams. Fig. 6 shows the root locus diagram for increasing values of spectrum illustrated in Fig. 10(a), the whole MG is affected by at least
kd and w . It can be seen, that the system is initially marginally stable three resonance peaks around 1020 Hz. This situation indicates that the
with an oscillatory behavior since it presents a pair of complex con­ resonance harmonic disturbances may propagate and affect the whole
jugate poles (P1 and P2) near at the imaginary axis. These poles are MG performances even when introduced by a single part. It is shown
lightly dampened, which makes the system response prone to re­ that the introduction of the proposed active damping method allows
sonances oscillations. As it can be observed from Fig. 6. (a), the in­ alleviating the undesired resonance harmonics. As reported in Table 1,
troduction of the proposed method moves the poles P1 and P2 toward the introduction of the proposed damping method reduces noticeably
the stable region. By increasing the coefficient kd , the real part of these the amplitude of the resonance harmonics that varies from 4.97 V and
poles increases and the damping ratio is improved, which results in a 12.7 V before damping, to 0.17 V and 2.12 V after damping. The output
more dampened system. P1 and P2 moves to higher frequencies when voltage quality is considerably enhanced by reducing the total har­
kd is up to 5, which facilitates their dampening. monic distortion content from 6.6% before damping to 1.1% after
Similarly, Fig. 6. (b) shows that by increasing the washout filter damping.
frequency, the poles P1 and P2 are attracted to the real axis making the
system more dampened. However, when the washout filter frequency 6.2. Case 2 – nonlinear load
increases beyond the switching frequency, the poles P1 and P2 move
toward the imaginary axis making the system less damped and more The efficacy of the proposed control strategy is evaluated when the
oscillatory. Moreover, the proposed damping method guarantees stable nonlinear load shown in Fig. 7 is supplied. This highly nonlinear load
performances as it is based on a washout filter with inherent stability stands for a widespread of load applications in MGs. The measured
characteristics. This filter introduces a stable pole P3 that moves toward DG#1, DG#2 and load voltage and current waveforms are shown in

Fig. 6. Root-locus diagram without (encircled) and with damping for 0≤kd≤ 10 when ωw=5k rad.s−1 (a), and for 1k≤ωw≤ 20k rad.s−1 when kd=1 (b).

6
A. Saim, et al. Electric Power Systems Research 191 (2021) 106900

Fig. 7. Loads circuit representation.

Fig. 8. Measured DG #1, DG #2 and PCC voltage (left) and current (right) waveforms, and the damping signal before and after damping – Case 1: Linear load[x-axis:
20 ms/div; y-axis: 10A/div - 100 V/div].

Table 1. amplitude of the resonance harmonics that varies from 4.83 to 12.5 V
Resonances damping performances. before damping, and from 0.99 V to 1.48 V after damping. Moreover,
Linear load Nonlinear load this dampening method allows enhancing the voltage waveform quality
Frequency Before After Before After by eliminating the undesired lower frequency harmonics that in­
troduces the supplied nonlinear load at 300 Hz and 660 Hz. The output
900 Hz 4.97 V 2.12 V – –
voltage quality is considerably enhanced by reducing the total har­
1020 Hz 12.7 V 1.77 V 5.81 V 1.48 V
1140 Hz 5.47 V 0.17 V 12.5 V 0.99 V
monic distortion content to 1.9%, while it was 6.9% before damping.
1260 Hz – – 4.83 V 1.44 V It is worth mentioning, that the proposed active resonance damping
control level achieves suitable damping performances attenuating the
undesired resonance amplifications without affecting the performances
Fig. 9 before and after damping. High frequency voltage and current of the inner control levels.
oscillations are noticed due mainly to resonances excitation. As shown
in the harmonic spectrum illustrated in Fig. 10(b), these resonances 7. Conclusion
oscillations appear at high frequencies around 1140 Hz. The introduc­
tion of the proposed active damping method allows mitigating the An advanced hierarchical control strategy with active resonance
undesired harmonic disturbances and considerably reduces the damping was designed to control DG based MGs operations and

Fig. 9. Measured DG #1, DG #2 and PCC voltage (left) and current (right) waveforms, and the damping signal before and after damping – Case 2: Nonlinear load[x-
axis: 20 ms/div; y-axis: 10A/div - 100 V/div].

7
A. Saim, et al.
Fig. 10. PCC's voltage harmonic spectrum before and after damping – Case 1: Linear load (a), Case 2: Nonlinear load (b).
dampen out the undesired resonance amplifications. This control
strategy comprises three control levels including voltage and current
control, power sharing control, and active resonance damping control.
The proposed active resonance damping control uses a filter based
approach to reconstruct and compensate for the undesired resonances
harmonics. The use of a Washout filter with inherent stability char­
acteristics has been proposed to extract the undesired resonances har­
monics content. The effectiveness of the proposed active damping
method has been verified and its stability investigated using a com­
prehensive numerical analysis considering variable damping coefficient
values. The experimental tests have shown that the proposed control
strategy demonstrates effective performances for resonances damping,
compensating for multiple resonances without affecting the system
dynamic.
Declaration of Competing Interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ­
ence the work reported in this paper.
Aknowlegement
This work has been supported by the Regional Fund and FEDER
under the framework of RFI Electronique WISE program
Appendix
Power stage parameters: E0 =√2 × 110 V, 0 =2π60 rad.s-1,
Vdc =250 V, Fs =10 kHz, L1, i =1 mH, L2, i =0.2 mH, Ci =33 μF.
Control parameters: Level 1: kpc =5.21, kc =60, kpv = 2.58, k v =70.
Level 2: mi =0.003 rad.s-1.W-1, ni =0.005 V.Var-1,
md, i =0.00004 rad . W 1, nd, i =0.00003 V . s. VAr 1, Ld, i =500 μH. Level 3:
kd =1, w =2π1000 rad. s-1.
References
[1] A. Gupta, S. Doolla, K. Chatterjee, Hybrid AC-DC Microgrid: systematic Evaluation
of Control Strategies, IEEE Trans. Smart Grid 9 (4) (2018) 3830–3843 Jul..
[2] S. Kumar Tiwari, B. Singh, P.K. Goel, Design and Control of Microgrid Fed by
Renewable Energy Generating Sources, IEEE Trans. Ind. Appl. 54 (3) (2018)
2041–2050 May.
[3] M. Nasir, H.A. Khan, A. Hussain, L. Mateen, N.A. Zaffar, Solar PV-based scalable DC
microgrid for rural electrification in developing regions, IEEE Trans. Sustain.
Energy 9 (1) (2018) 390–399 Jan..
[4] M. Meral, D. Çelik, A comprehensive survey on control strategies of distributed
generation power systems under normal and abnormal conditions, Annu. Rev.
Control 47 (2019) 112–132.
[5] D. Çelik, M. Meral, A novel control strategy for grid connected distributed gen­
eration system to maximize power delivery capability, Energy 186 (2019) 115850.
[6] M. Hajiakbari Fini, M.E. Hamedani Golshan, Determining optimal virtual inertia
and frequency control parameters to preserve the frequency stability in islanded
microgrids with high penetration of renewables, Electr. Power Syst. Res. 154 (2018)
13–22 Jan..
[7] Y.S. Kim, E.S. Kim, S. Il Moon, Frequency and voltage control strategy of standalone
microgrids with high penetration of intermittent renewable generation systems,
IEEE Trans. Power Syst. 31 (1) (2016) 718–728.
[8] J. He, X. Liu, M. Lei, C. Wang, A Broad Frequency Range Harmonic Reduction for
Cascaded Power Cells Based Islanded Microgrid with Lumped PCC Filter, IEEE
Electric Power Systems Research 191 (2021) 106900
Trans. Power Electron. (2020) 1 Jan..
[9] D. Sharma, S. Mishra, Disturbance-Observer-Based Frequency Regulation Scheme
for Low-Inertia Microgrid Systems, IEEE Syst. J. 14 (1) (2020) 782–792 Mar..
[10] A. Akhavan, H.R. Mohammadi, J.M. Guerrero, A comprehensive control system for
multi-parallel grid-connected inverters with LCL filter in weak grid condition,
Electr. Power Syst. Res. 163 (2018) 288–300 Oct..
[11] F. Wu, J. Ye, X. Luo, Z. Zhang, Y. Li, Resonance characteristic analysis and damping
control for LCL-filter-based voltage source converter during bus transfer, Electr.
Power Syst. Res. 150 (2017) 144–151 Sep..
[12] N. Rashidirad, M. Hamzeh, K. Sheshyekani, E. Afjei, High-Frequency Oscillations
and Their Leading Causes in DC Microgrids, IEEE Trans. Energy Convers. 32 (4)
(2017) 1479–1491 Dec..
[13] A. Akhavan, H.R. Mohammadi, J.C. Vasquez, J.M. Guerrero, Coupling effect ana­
lysis and control for grid-connected multi-microgrid clusters, IET Power Electron.
13 (5) (2020) 1059–1070. Apr..
[14] F. Yazdi, S.H. Hosseinian, Variable cost model predictive control strategies for
providing high-quality power to AC microgrids, IET Gener. Transm. Distrib. 13 (16)
(2019) 3623–3633. Aug..
[15] Q. Ye, R. Mo, H. Li, Multiple resonances mitigation of paralleled inverters in a solid-
state transformer (SST) Enabled AC Microgrid, IEEE Trans. Smart Grid 9 (5) (2018)
4744–4754 Sep..
[16] M. Lu, X. Wang, P.C. Loh, F. Blaabjerg, Resonance Interaction of Multiparallel Grid-
Connected Inverters With LCL Filter, IEEE Trans. Power Electron 32 (2) (2017)
894–899 Feb..
[17] A. Luo, Q. Xu, F. Ma, Y. Chen, Overview of power quality analysis and control
technology for the smart grid, J. Mod. Power Syst. Clean Energy 4 (1) (2016) 1–9
Jan..
[18] Y. Han, et al., Modeling and Stability Analysis of $LCL$ -Type Grid-Connected
Inverters: a Comprehensive Overview, IEEE Access 7 (2019) 114975–115001. Aug..
[19] M. Routimo and H. Tuusa, “LCL type supply filter for active power filter -
Comparison of an active and a passive method for resonance damping,” in PESC
Record - IEEE Annual Power Electronics Specialists Conference, 2007, pp.
2939–2945.
[20] A.K. Balasubramanian, V. John, Analysis and design of split-capacitor re­
sistiveinductive passive damping for LCL filters in grid-connected inverters, IET
Power Electron 6 (9) (2013) 1822–1832.
[21] Y. Tang, W. Yao, P.C. Loh, F. Blaabjerg, Design of LCL Filters with LCL Resonance
Frequencies beyond the Nyquist Frequency for Grid-Connected Converters, IEEE J.
Emerg. Sel. Top. Power Electron. 4 (1) (2016) 3–14 Mar..
[22] M.A. Awal, H. Yu, L. Della Flora, W. Yu, S. Lukic, I. Husain, Observer based ad­
mittance shaping for resonance damping in voltage source converters with LCL
filter, in 2019 IEEE Energy Conversion Congress and Exposition, ECCE 2019, 2019,
pp. 4455–4462.
[23] J. Dannehl, F.W. Fuchs, S. Hansen, P.B. Thøgersen, Investigation of active damping
approaches for PI-based current control of grid-connected pulse width modulation
converters with LCL filters, IEEE Trans. Ind. Appl. 46 (4) (2010) 1509–1517 Jul..
[24] Y. Guo, L. Chen, X. Lu, J. Wang, T. Zheng, S. Mei, Region-Based Stability Analysis
for Active Dampers in AC Microgrids, IEEE Trans. Ind. Appl. 55 (6) (2019)
7671–7682 Nov..
[25] D. J. L. M. F. F, Filter-Based Active Damping of Voltage Source Converters With LCL
Filter, IEEE Trans. Ind. Electron. 58 (8) (2011) 3623–3633.
[26] W. Yao, Y. Yang, X. Zhang, F. Blaabjerg, P.C. Loh, Design and Analysis of Robust
Active Damping for LCL Filters Using Digital Notch Filters, IEEE Trans. Power
Electron 32 (3) (2017) 2360–2375 Mar..
[27] J. Yuan, A. Al Durra, E. El-Sadaany, Adaptive digital notch filter based on online
grid impedance estimation for grid-tied LCL filter systems, Electr. Power Syst. Res.
172 (2019) 183–192 Jul..
[28] A. Saim, A. Houari, J.M. Guerrero, A. Djerioui, M. Machmoum, M.A. Ahmed,
Stability analysis and robust damping of multiresonances in distributed-generation-
based islanded microgrids, IEEE Trans. Ind. Electron. 66 (11) (2019) 8958–8970
Nov..
[29] J. He, Y.W. Li, D. Bosnjak, B. Harris, Investigation and Active Damping of Multiple
Resonances in a Parallel-Inverter-Based Microgrid, IEEE Trans. Power Electron 28
(1) (2013) 234–246 Jan..
[30] A. Saim, R. Mellah, A. Houari, M. Machmoum, A. Djerioui, Adaptive resonant based
multi-loop control strategy for parallel distributed generation units in standalone
microgrid application, Electr. Power Syst. Res. 143 (2017).
[31] J.H. Chow and R.J. Thomas, Systems and control theory for power systems. 1995.
[32] Abdelhakim Saim, Azeddine Houari, Josep.M Guerrero, Ali Djerioui, Mourad Ait
Ahmed, Mohamed Machmoum, Modeling of complex resonances in islanded
Microgrids, 2018 7th International Conference on Renewable Energy Research and
Applications (ICRERA) (2018), https://doi.org/10.1109/ICRERA.2018.8566906.