Received: 3 April 2022 Revised: 19 May 2022 Accepted: 14 June 2022 IET Generation, Transmission & Distribution

DOI: 10.1049/gtd2.12539

ORIGINAL RESEARCH

Transient stability enhancement in multiple-microgrid networks

by cloud energy storage system alongside considering protection
system limitations

Hossein Ebrahimi1 Amin Yazdaninejadi2 Sajjad Golshannavaz1

1
Electrical Engineering Department, Urmia Abstract
University, Urmia, Iran
The gas turbine synchronous generators (GTSGs) are widely deployed as distributed gen-
2
Electrical Engineering Department, Shahid Rajaee erations (DGs) in countries with massive natural gas production owing to their low prices.
Teacher Training University, Tehran, Iran
However, due to the low inertia time constants of these synchronous-based DGs, they
are more susceptible to power grid faults which stands as a transient stability issue in net-
Correspondence
Sajjad Golshannavaz, Nazloo campus of Urmia works with multiple microgrids (MGs). On the other hand, the cloud energy storage system
University, 13th kilometer of Serow road, Urmia, (CESS) is a new concept that centralizes the individual distributed energy storage of one or
West Azerbaijan, Iran.
more MGs. Here, the employment of CESS with synchronverter grid connections creates
Email: [email protected]
a suitable opportunity for improving the transient stability of the network by providing
higher inertia. This is while the fault current contribution of the synchronverter-based
CESS imperils protection constraints in these networks. Therefore, a proper protection
coordination index (PCI) is considered in the conducted study to identify the optimal size
of the synchronverter-based CESS through a two-stage optimization algorithm that pre-
serves the protection constraints among protective devices. Finally, the transient stability
of the network with synchronverter-based CESS is assayed by calculation of the critical
clearing time (CCT) for faults. Numerical studies are carried out on the IEEE 33-bus test
system. Results are discussed in depth.

1 INTRODUCTION sient stability. Therefore, the penetration of these low-inertia

DGs in the network jeopardizes their transient stability [3].
The ever-increasing growth in electricity consumption forces The stability of a system has a wide area of aspects each of
power system planners and operators to provide proper which are distinct field of study. According to [4], the dynamic
provisions for meeting the generation-consumption balance. stability or small-signal stability of a system is its ability to return
However, today, considering the economical, technical, and to the normal operating condition following small disturbances.
environmental constraints, the construction of bulk power On the other hand, the transient stability of the system is related
plants is not a completely viable option. Therefore, the deploy- to the ability of the synchronous machines to remain synchro-
ment of distributed generations (DGs), renewable energy nism following a large disturbance such as three-phase faults.
sources (RESs), energy storage systems (ESSs), and demand- The dynamic or small-signal stability of the system is assessed
side management (DSM) programs are the available alternatives through eigenvalue analysis of the power system’s model. How-
[1]. Among the various technologies of the DGs, small-scale ever, the assessment of the transient stability of a system is based
gas turbine synchronous generators (GTSGs) are dominant in on time-domain simulations. Regarding this context, the single-
countries with massive natural gas production due to the low phase microgrids (MGs) controlled in the stationary frame are
price of natural gas. These small-scale GTSGs are susceptible modelled and their dynamic stability is assessed in [5]. The
to the disturbances, such as faults, owing to their low inertia [2]. maximum allowable penetration level of inverter-based gener-
The inertia of the electric network directly influences its tran- ations and the associated small-signal stability are explored in

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the
original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
© 2022 The Authors. IET Generation, Transmission & Distribution published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.

IET Gener. Transm. Distrib. 2022;1–11. wileyonlinelibrary.com/iet-gtd 1

2 EBRAHIMI ET AL.

[6]. Moreover, in [7], a control strategy of energy storage units is different types of converters that are working in grid-following
proposed to preserve the dynamic stability of DC MGs. Addi- and grid-forming states in the vicinity of each other. For a sim-
tionally, a robust control scheme is proposed to improve the ilar mission, the synchronverter-based HVDC link is employed
small and large-signal stability of independent MGs in [8]. The to adjust the transfer capability and transient stability of the
transient stability of MGs considering the presence of both interconnection point of weak AC zones. By the way, none of
synchronous generators and inverter-based generators is inves- the reviewed literature neither considered the connection of the
tigated in [9]. Protective schemes are explored in [2], [10] for CESS with synchronverter type of inverters to bring up the
preserving the transient stability of MGs and active distribu- opportunity of providing higher inertia to the network nor the
tion networks, respectively. Moreover, various researchers have protection constraints of the protective devices are taken into
intended to improve the transient stability of electric systems account for the deployment of CESS.
by employing virtual synchronous generators as virtual inertia The low inertia time constant of the synchronous-based DGs
sources [11–19]. makes them more prone to failure in response to grid faults
Meanwhile, a new concept of energy storage namely cloud which is an issue of transient stability for multiple-MG net-
energy storage system (CESS) is trending in the literature which works. Meantime, CESSs as a new concept for centralizing
is used in MGs to centrally store the electrical energy for the the energy storage management of MGs opens up a novel
customers [20]. The communication and technological infras- technological field to be explored and benefitted from. Employ-
tructures needed for implementation of the CESS besides its ing synchronverters as the grid-connecting interface of CESSs
economic model are described and discussed in [20]. The can be useful in gaining the inertia of the network. Here, the
communication safety and security requirements of the CESS connection of the CESS by synchronverter to the connection
system are met by [21] through designing and developing a node points of multiple MGs of a highly GTSG-penetrated distribu-
controller which controls the connection of the public user to tion network is proposed for improving the transient stability
the CESS. Based on [22], the cost of battery degradation and of the connection points. The employed synchronverter-based
uncertainty of the price should be considered in the design and CESS system provides high inertia against severe disturbances,
decision-making processes of the CESS in a long-term man- such as three-phase faults in the vicinity of the connection
ner. Moreover, the decisions of investment and operation of points of MGs. In order to deal with the high fault current
the CESS operator and the consumers are considered in [23] contribution of the synchronverter-based CESS, a protection
to express the benefits of this type of storage system. Further- coordination index (PCI) is considered in a two-stage optimiza-
more, the contribution of end-users in the demand response tion algorithm that determines the optimal collaboration of the
programs (DRPs), the uncertainties of the load demand, and synchronverter-based CESS in the relaying process. Afterwards,
the energy storage loss are considered for the configuration the critical clearing time (CCT) of the network is calculated
of the CESS in [24]. Authors of [25] carried out a research with and without the employment of the synchronverter-based
study on the aggregation of the domestic battery storage of CESS to validate the performance of the proposed method
different technologies to provide grid-scale services including in improving the transient stability of the network. The main
energy arbitrage, reserve market, and frequency regulation in the contributions of this paper could be listed in the following:
UK. They concluded that current regulatory obstacles do not
allow this aggregation to be economically beneficial, but the fre- ∙ A synchronverter-based CESS is proposed to take advantage
quency regulation and the fast response reserves yield enough of the characteristics of synchronous generators;
annual profits to cover up the losses. These ancillary services ∙ The synchronverter-based CESS is employed to pose high
require technical amendments and devices to provide proper inertia for improving the transient stability of MGs in the
infrastructure. distribution network;
One of the auxiliary and supplementary devices for provid- ∙ A two-stage optimization algorithm is proposed for the allo-
ing the grid-scale services is the inverter of the CESS system. cation of the CESS. This optimization process uses the PCI
According to the inverter type of the CESS, it can provide to prevent violation of selectivity tasks by installing the CESS.
special services. Synchronverters are a special type of inverter
that mimic the behaviour of a synchronous generator; but, The rest of this paper is organized as follows. Section 2
unlike synchronous generators, the parameters of synchron- describes the proposed method. In Section 3 the results of
verters can be modified to obtain the required services [26]. numerical studies are presented and discussed. Finally, Section
A controller design for a self-synchronizing synchronverter is 4 concludes the paper.
presented in [27]. Moreover, an analysis method of asymmetric
fault current employing the instantaneous symmetric compo-
nent method is presented in [28] by setting up the equivalent 2 MATERIALS AND METHODS
model of the synchronverter. Furthermore, the authors of [29]
have designed and developed a shunt active filter based on syn- In this section, first, the CESS concept is expressed. Then,
chronverter to enhance its performance. Additionally, a control the proposed structure of the synchronverter-based CESS
approach of the synchronverters in MGs with high penetration is described. Finally, an optimization algorithm is proposed
of inverter-based DGs is proposed in [30] to improve the net- for allocating the synchronverter-based CESS which does not
work security. Another study in [31] investigates the stability of endanger the protection of the distribution network.
EBRAHIMI ET AL. 3

FIGURE 1 The data flow for the deployment of CESS in a distribution network with multiple MGs

2.1 CESS operation concept

Conventional ESSs provide a wide range of services for resi-

dential, commercial, and industrial customers along with many
utility-side opportunities. However, employing this system can-
not completely satisfy the economical requirements of all parties
of the electric network owing to the high investment costs,
maintenance requirements, and management issues in the case
of end-user-scale ESSs. The novel concept of the CESS reduces
the investment, maintenance, and management problems of
the scattered ESS employment by integrating them into one
centrally managed unit.
CESS is a grid-scale energy storage system that provides a
shared capacity for storing and withdrawing energy for the dis-
tribution network and MG users. According to the scale of the
CESS, it can present grid-wide technical and economic services.
Employment of a CESS calls the need for telecommunication
infrastructure to establish communication between the end-
users, CESS operator, and grid operator. Figure 1 shows the data
flow for the deployment of a CESS in a distribution network FIGURE 2 Two parts of the synchronverter: (a) Control part, (b) power
with multiple MGs. part
According to Figure 1, after gathering and processing the
data from end-users of the MG, the MG operators transmit the
required information such as the charging/discharging schedule type of machine can be found in various references [32, 33].
to the CESS operator and the CESS operator sends the needed Synchronverter is a type of inverter that mimics the behaviour
signals such as charging/discharging price to the MG operator. of a synchronous generator. Figure 2 shows a synchronverter
Considering the employed DSM programs in MGs, MG opera- that includes two parts: (a) The power part, and (b) the control
tors optimize the required charging/ discharging schedule based part. The control part is designed in a way that the system can
on the demanded rates by the CESS operator. In addition to the express the characteristics of a synchronous generator. On the
economic aspect of employing the CESS, it can provide techni- other hand, the power part is the interface of the synchronverter
cal benefits such as frequency and voltage ancillary services. with the rest of the grid.
In the power part, the DC voltage is supplied by a CESS
through a DC link. R f , L f , and C f are the resistance, induc-
2.2 The proposed model of tance, and capacitance of the interfacing filter, respectively. Rg ,
synchronverter-based CESS Lg , and vg are the equivalent resistance, inductance, and volt-
age of the grid, respectively. The back electromotive force e
Synchronous generators are reluctant to quick alterations of out- stands for the constant field current state of the synchronous
put power owing to their intrinsic inertia. The model of this generator.
4 EBRAHIMI ET AL.

By considering that the synchronverter mimics the behaviour By deployment of the synchronverter-based CESS with a
of the synchronous generator, the back electromotive force of high value of inertia in the distribution network feeding several
the synchronverter is as follows: MGs, the inertial response of the network and consequently the
transient stability of the MGs in event of disturbances such as
̇ f i f s˜
e = 𝜃M in𝜃 = E s˜
in𝜃 (1) faults gets improved. In this regard, the inertia of the synchron-
verter can be adjusted by adjusting the inertia time constant and
In which: the damping constant in (4). This feature of the synchronverter
distinguishes its performance from the synchronous generators.
1
𝜃̇ = (Tm − Te + D p (𝜃̇ ∗ − 𝜃))
̇ (2)
Js
1 ( ) 2.3 The proposed two-stage optimization
Mf if = Qset − Q + Dq (E ∗ − Vt ) (3) method
Ks

J = D p 𝜏 fr (4) Directional overcurrent relays (DOCRs) are widely used for

protecting power systems, especially for active distribution net-
works due to the bi-directional current flow in these networks.
̇ q 𝜏v
K = 𝜃D (5) The operation function of a DOCR (t) is expressed in the
following:
M f , i f , s˜
in𝜃 and are the virtual mutual inductance, the vir-
( )
tual field current, and a three-phase Sine function, respectively. t = f TMS , Isc , IP , A, B (10)
E is the magnitude of back electromotive force e. Moreover,
and Vt represent the angular frequency and terminal voltage As can be seen, the tripping time of a DOCR is a function
of the synchronous 𝜃̇ generator with 𝜃̇ ∗ and E ∗ reference of its time multiplier setting (TMS ), pickup current (I p ), and
values, respectively. Tm and Te are the mechanical and elec- the short circuit current flowing through the relay (Isc ). In this
tromagnetic torques of the synchronverter. Qset represents the equation, A and B are constant values. The PCI is defined as
reactive power setpoint determined by the grid codes based the maximum changing rate of DG penetration level (ΔP) with
on the reactive current requirements. Furthermore, D p and Dq respect to the changing rate of coordination interval (ΔCTI )
stand for the frequency and voltage droop coefficients, respec- [34].
tively. J and K also denote the virtual moment of inertia and
reactive power regulating constant, respectively. Equations (2) ΔP
PCI = − (11)
calculate the angular frequency and Equation (3) expresses the ΔCTI
induced voltage in the rotor windings by the current of stator
windings. According to (4) and (5), J is dependent on the fre- The higher the value of PCI is, the less the impact of installing
quency droop time constant 𝜏 fr and K is calculated based on the DG in that bus on the protection system would be. The unit
the voltage droop time constant 𝜏v . The mechanical and elec- of the PCI is MVA/s. Employment of the PCI for allocating the
tromagnetic torques generated by the synchronverter can be synchronverter-based CESS helps to determine the best candi-
calculated by the following equations. date bus for its installation. As shown in Figure 3, the method
in this paper is based on a two-stage optimization algorithm
Pset in which the first stage is a simple optimization for obtaining
Tm = (6)
𝜃̇ protection settings without the presence of any synchronverter-
⟨ ⟩ based CESS is done. Then, in the second stage, an optimization
Te = M f i f i, s˜in𝜃 (7) is performed for each candidate bus based on the maximiza-
tion of the synchronverter-based CESS penetration in each CTI
In (6), Pset is the active power setpoint which can be obtained value. Afterward, the PCI calculation is performed and the pri-
from the DC voltage regulating loop in Figure 2a. Additionally, ority of the buses for installing the synchronverter-based CESS
in (7),⟨.,.⟩ denotes the conventional inner product in ℝ3 space. is determined.
The reactive power of the synchronverter Q and its setpoint Qset The optimization model employed in the first stage is as
are calculated by (8) and (9), respectively. follows:
∑( )
̇ f i f ⟨i, cos𝜃⟩
Q = −𝜃M ˜ (8) minimize T = t p + tb (12)

subject to:
Qset = −Vtd Iq∗ (9)
tb − t p ≥ CTI (13)
where Vtd and Iq∗ are the d-axis component of Vt and the
q-axis component of the reference current to generate Qset ,
respectively. I pmin ≤ I p ≤ I pmax (14)
EBRAHIMI ET AL. 5

subject to (13)–(16) and the limitation of the synchronverter-

max
based CESS size Ssyn−CESS as follows:

0 ≤ Ssyn−CESS
k
≤ Ssyn−CESS
max
(18)

In (17), pen is the penetration level of the synchronverter-

k total
based CESS at bus k. Moreover, Ssyn−CESS and S pre−installed are
the size of the synchronverter-based CESS (MVA) and the total
pre-installed power (MVA) in the system. After obtaining the
optimal size of the synchronverter-based CESS for each CTI
value in candidate buses, the value of the PCI should be cal-
culated based on (11) to determine the priority of the buses in
installing the synchronverter-based CESS.
In the proposed two-stage method, the bus impedance matrix
of the network (Zbus ) plays a vital role. In the first stage, the
Zbus is constructed to calculate the short circuit currents for
obtaining the optimal settings. As it is known, by including
a new component in the electrical system, the Zbus would be
altered. Therefore, in the second stage of the optimization, the
Zbus matrix should be reconstructed based on new components
added to the network. Then, the short circuit currents are recal-
culated based on the new Zbus to perform the optimization of
the second stage for each specified value of the CTI and newly
added synchronverter-based CESS at the candidate bus. This
procedure is illustrated in the flowchart in Figure 4.

2.4 CCT evaluation

FIGURE 3 The two-stage optimization method

Small-scale synchronous generators which are considered to be

GTSGs in this study are more prone to instability due to their
low inertia time constants. Therefore, it is possible that during
TMS min ≤ TMS ≤ TMS max (15) disturbances of the network such as faults, the protection system
cannot clear the fault before their critical clearing time which
0 ≤ t p ,tb ≤ t max (16) causes the instability of the network. The CCT of the network
is the minimum CCT of all synchronous generators:
in which t p and tb are the operation times of the main and { }
CCTnet = min CCTsg1 , CCTsg2 , … (19)
backup protective devices. TMS min and TMS max demonstrate the
lower and upper bounds of the TMS , respectively. Moreover, I p
Employment of high-inertia synchronverter-based CESS can
is restrained by its minimum I pmin and maximum I pmax values.
be helpful in improving the inertia of the network and con-
Equation (12) defines the objective function of the optimization
sequently increasing the CCT. The CCT of a synchronous
problem which is the sum of the operation time of the main and
generator CCTsgi for a fault in the near bus is calculated by the
backup devices. Inequality (13) is included in the model to pre-
following equation [2]:
vent the violation of selectivity tasks. Additionally, constraints
(14), (15), and (16) determine the restrictions of I p , TMS , and √
the operation times of the protective devices, respectively. 4H (𝛿cr − 𝛿0 )
CCTsgi = (20)
After performing the first stage of the algorithm in which 𝜔Pm
the associated protection settings are obtained, these settings
are considered as constant values for the second stage. The opti- where H is the inertia time constant of the generator. 𝛿cr is
mization process of the second stage maximizes the penetration the critical rotor angle that crossing it causes the synchronous
level of the synchronverter-based CESS by altering the value of generator to get unstable. Moreover, 𝛿0 is the rotor angle in
the CTI. The objective function of this optimization stage is: which the synchronous generator supplies the electric power
Pm before the faulty situation. Finally, 𝜔 denotes the angular
k
Ssyn−CESS velocity of the rotor. In this study, by calculating the PCI for
maximizepen = total
(17) each candidate bus and extracting the installation priority list for
S pre−installed
6 EBRAHIMI ET AL.

FIGURE 4 The algorithm of the two-stage optimization method

the synchronverter-based CESS, a CTT evaluation is performed fault. The maximum value of the fault clearing time is taken
to have more insight into the inertial response of the network 5 s in this study.
with and without the employment of the synchronverter-based
CESS. The procedure of calculating the CCT of the network is
performed by the following steps: 3 TEST SYSTEM SPECIFICATIONS
AND RESULTS
∙ First, a load flow calculation is performed;
∙ Then, the minimum fault clearing time is set to a predefined In this section, the specification of the test system under
minimum value of 0.05 s; study, assumptions, and the obtained results from numerical
∙ If the system is not stable, the CCT of the network is the simulations are expressed.
predefined minimum value. But, if it stays stable in that pre-
defined minimum value, the fault clearing time would be
increased gradually by 0.01 s steps; 3.1 Test system and assumptions
∙ By gradually increasing the fault clearing time the transient
stability of the network is assessed until it gets unstable in an The numerical studies are carried out on the IEEE 33-bus test
individual time which is the CCT of the network; system which is illustrated in Figure 5. As can be seen, the
∙ The algorithm defines a maximum value for the fault clearing deployed DGs in this test system are GTSGs. The bus num-
time after which the system would stay stable forever for that bers and capacities of these GTSGs are tabulated in Table 1
EBRAHIMI ET AL. 7

based on the conducted study in [35]. Moreover, the place and

direction of the DOCRs of the network are shown in Figure 5.
Besides, buses 14–18 are considered to be the connection points
of MGs in this study. The upstream grid supplies 10 MVA and
the total supplied power by the upstream grid and SGTGs is
total
S pre−installed = 17.8125 MVA. The base MVA of the network is
100 MVA and the nominal voltage is 12.66 kV which results in a
base current equal to 7898.9 A. It should be noted that accord-
ing to [10] if the transient stability of a system is preserved for
three-phase faults, it can also be assured that other fault con-
ditions and disturbances cannot disturb the transient stability of
the system. Therefore, in this study, three-phase fault as a severe
case are considered to evaluate the transient stability.
The values of CTI in the first stage of the optimization, A,
and B are 0.3, 0.14, and 0.02. The maximum and minimum
boundaries of TMS are considered 0.1 and 1 s, respectively.
Moreover, the value of I p is restrained by 1.1–1.5 times the
normal-state current. The maximum operation time of the
relays is considered 2 s. All constant parameters of the relay
characteristics are extracted from IEC 60255-151:2009 standard
[36]. All parameters of the synchronverter are taken from [26]
except for the inertia J which is considered 73.44 for imposing
high inertia from the synchronverter-based CESS. Furthermore,
the value of CTI in the second stage of the optimization is con-
templated to change from 0.27 to 0.33 s in 0.01 s steps. The
maximum size of the synchronverter-based CESS is also taken
max
Ssyn−CESS = 62.5 MVA. The numerical simulations are carried
out in MATLAB simulation environment.
It is worth mentioning that the critical condition of the sys-
tem is considered in this study to obtain robust results. In
this regard, the capacity and place of the synchronverter-based
CESS are determined based on the PCI within which the CESS
installation would have the least impact on the selectivity of the
protection system. Moreover, the transient stability of the sys-
tem is assessed based on the worst-case scenario of the system
i.e., the three-phase fault. Therefore, the obtained results would
be robust against the other single and two-phase faults.
FIGURE 5 Modified IEEE 33-bus test system

TABLE 1 Bus numbers and capacities of GTSGs [35] 3.2 Simulation results
GTSG Bus no. Capacity (MVA)
In order to explore the performance of the proposed
GTSG 1 8 0.3125 synchronverter-based CESS in inertia enhancement of the
GTSG 2 11 0.9375 multiple-MG distribution network, two cases are considered.
GTSG 3 12 0.3125 In the first case, for all MG-connected buses (buses 14–18) of
GTSG 4 15 0.625 the test system, the PCI method is evaluated for determining
the maximum synchronverter-based CESS penetration level
GTSG 5 19 0.3125
in a manner that protection limitations are met. Then, in
GTSG 6 23 1.25
the second case, two scenarios are considered to compare
GTSG 7 25 1.25 the effect of the proposed synchronverter-based CESS and
GTSG 8 26 0.3125 conventional inverter-based CESS on inertia enhancement.
GTSG 9 28 1.25 Since employment of the conventional inverter-based CESS
GTSG 10 30 0.3125 cannot improve transient stability of the system meaning-
fully, the results of this scenario would be the same as those
GTSG 11 31 0.625
results when the CESS is not employed. Therefore, the results
GTSG 12 32 0.3125
associated with them are reported as the same scenario. In
8 EBRAHIMI ET AL.

the abovementioned scenarios, to compare the effect of the

proposed synchronverter-based CESS and the conventional
inverter-based CESS (or without CESS), the transient stability
improvement is evaluated based on the CCT of the system.

3.2.1 Case 1
As stated in Section 2.3, a two-stage optimization procedure
is performed to assess the impact of installing synchronverter-
based CESS in different buses on the protection of the
distribution network. From the first stage, the relays’ settings are
obtained. In the second stage, by employing the obtained relay
settings, the optimal capacity of the synchronverter-based CESS
is obtained through gradually increasing the CTI from 0.27 to
0.33 in 0.01 steps where the contribution of synchronverter-
based CESS does not violate selectivity tasks of relays. This
optimization calculates the optimal capacity for installation of
the synchronverter-based CESS at each value of CTI. The
obtained results are depicted in Figure 6 based on the pene-
tration of the synchronverter-based CESS in buses 14–18 of
the testbed. In this figure, the maximum allowable penetra-
tion levels of the synchronverter-based CESS for CTI of 0.3
s are shown. Moreover, by moving toward the end of the
network, the value of the maximum installable synchronverter-
based CESS increases. According to Figure 6, it is possible to
increase the penetration level of synchronverter-based CESS in
buses 14, 15, 16, 17, and 18 up to 13.6%, 13.7%, 13.8%, 14.3%,
and 14.4% without violating the protection selectivity. The slope
of these curves shows the value of the PCI at each bus.
Here, the minimum, maximum, and average values of the PCI
at each bus are reflected in Table 2 which indicates that installing
synchronverter-based CESS in bus 18 has the lowest impact on
the protection system of the testbed. As can be seen, the min-
imum, maximum, and average PCI values for bus 18 are 3.16,
33.97, and 12.04 s which are the highest values among all. On
the contrary, the installation of the synchronverter-based CESS
in bus 14 poses the most effect on the protection system of
the testbed. The values of the minimum, maximum, and average
PCI for bus 14 are 3.15, 15.22, and 7.62 s as the lowest values
among all.

3.2.2 Case 2
Based on reported results in the previous case, Table 3 shows
the maximum allowable installation at each candidate bus with-
out jeopardizing the protection system in CTI = 0.3 s by
considering the total installed capacity of the test system and
the allowable increment in synchronverter-based CESS pene-
tration. In this case, two scenarios are considered to compare
the effect of the proposed synchronverter-based CESS and
conventional inverter-based CESS on inertia enhancement. As
pointed out before, since the employment of the conven-
tional inverter-based CESS cannot improve transient stability FIGURE 6 Variation of synchronverter-based CESS penetration with
of the system, instead testbed with conventional inverter-based CTI in (a) Bus 14, (b) Bus 15, (c) Bus 16, (d) Bus (17), and (e) Bus 18
CESS and without CESS, the same scenario is considered.
EBRAHIMI ET AL. 9

TABLE 2 The average values of PCI at each bus

Minimum Average Maximum

Bus 14 3.15 7.62 15.22

Bus 15 3.15 8.05 16.87
Bus 16 3.15 8.6 19.12
Bus 17 3.16 10.79 28.22
Bus 18 3.16 12.04 33.97

TABLE 3 The maximum allowable synchronverter-based CESS capacity

at each bus

Capacity (MVA)

Bus 14 2.420
Bus 15 2.440
Bus 16 2.457
Bus 17 2.542
Bus 18 2.567

Therefore, the first scenario considers the testbed with conven-

tional inverter-based CESS (or without CESS) and the second
scenario explores the testbed with synchronverter-based CESS
in different candidate buses.
In the pointed scenarios, the transient stability improvement
is evaluated based on the CCT of the system. Figure 7 illus-
trates the inertial response in the MGs’ connection points of
the test system by performing the CCT calculation in each sce-
nario. The synchronverter-based CESS is allocated according to
the capacities in Table 3 in each bus. Moreover, disturbances are
considered to be by midway faults in lines 13–14, 14–15, 15–
16, 16–17, and 17–18. In Figure 7, results associated with the
first scenario are reported with black lines while results asso-
ciated with the second scenario are depicted with red lines.
According to this figure, in the second scenario, installing the
synchronverter-based CESS in none of the candidate buses
does not reduce the CCT of the network. However, when the
synchronverter-based CESS is installed in bus 14, for a fault
in line 14–15, the test system still is not stable. The reason
is that a bolted severe fault occurs in the vicinity of the low-
inertia GTSG4 and the high-inertia synchronverter-based CESS
is located on the other side of the fault location which cannot be
helpful in the stability of the GTSG4. Moreover, Figure 7 shows
that installation of the synchronverter-based CESS nearby the
GTSG4 is more useful in improving the transient stability of the
network than installing it in far buses. Furthermore, it is obvious
that by moving towards the end of the network, the impact of
the faults on the CCT of the network is getting lower and lower.
Table 4 shows the average CCT values of the network for
faults explored in Figure 7 to determine the best candidate bus
from a transient stability improvement viewpoint. According to
Table 4, the average value of CCTs in the test system is improved FIGURE 7 CCT values of the network for faults in different lines when
by 22.03%, 29.85%, 31.06%, 24.31%, and 24.60% by installing the synchronverter-based CESS is installed in (a) Bus 14, (b) Bus 15, (c) Bus 16,
the synchronverter-based CESS in buses 14, 15, 16, 17, and (d) Bus (17), and (e) Bus 18
10 EBRAHIMI ET AL.

TABLE 4 The average CCT values of the network considering installing distribution networks with multiple MGs has been explored.
the synchronverter-based CESS in candidate buses According to the ability of the synchronverters in emulating
Average CCT (s) the characteristics of synchronous generators, they were con-
templated to impose a high amount of inertia in response to
No CESS 1.407
different disturbances such as severe three-phase faults in net-
Bus 14 1.717
works with high penetration of low-inertia GTSGs. Assessment
Bus 15 1.827 of the PCI and CCT of the network was considered as two cases
Bus 16 1.844 of this study for evaluating the performance of the network
Bus 17 1.749 with synchronverter-based CESS in different places in view-
Bus 18 1.753 points of protection system and transient stability, respectively.
In the first case, the PCI has been calculated by a two-stage
optimization algorithm in which the optimal capacity of the
synchronverter-based CESS was obtained in different buses for
different values of the CTI between 0.27 and 0.33 s. Then, in
the second case, for the maximum allowable capacity of the
synchronverter-based CESS at different candidate buses with
CTI equal to 0.3 s, the CCT value of the network has been com-
puted. The computed CCTs showed that employment of the
synchronverter-based CESS at different buses yields enhance-
ment in transient stability. Moreover, results indicated that the
installation of the synchronverter-based CESS in bus 18 has
the least impact on the protection system of the network. Fur-
thermore, installing the synchronverter-based CESS in bus 16
presents the maximum improvement in CCT and transient
stability of the multi-MG test system.
FIGURE 8 Transient response of GTSG 2 to a three-phase fault in line
13–14
FUNDING INFORMATION
The author(s) received no specific funding for this work.
18. Therefore, bus 16 is the best candidate for the installation
of the synchronverter-based CESS to bring transient stability CONFLICT OF INTEREST
enhancement in the MGs of the test system. The authors certify that they have no affiliations with or involve-
For more clarification regarding the performance of the ment in any organization or entity with any financial interest
synchronverter-based CESS in improving the transient stability (such as honoraria; educational grants; participation in speak-
of the system in response to disturbances, Figure 8 illustrates ers’ bureaus; membership, employment, consultancies, stock
the transient response of the GTSG2 to a three-phase fault ownership, or other equity interest; and expert testimony or
in line 13–14 for both scenarios where synchronverter-based patent-licensing arrangements), or non-financial interest (such
CESS is/is not at bus 16. According to Figure 8, a three-phase as personal or professional relationships, affiliations, knowledge
fault is injected at time 0.1 s and then cleared after 0.25 s at or beliefs) in the subject matter or materials discussed in this
time 0.35 s. Figure 8 shows that the rotor angle of GTSG2 goes manuscript.
to the unstable area when the synchronverter-based CESS is
not installed at the first scenario. The reason lies in Figure 7c. DATA AVAILABILITY STATEMENT
According to Figure 7c, for the fault at line 13–14, the CCT of Data sharing not applicable to this article as no datasets were
the system is 0.05 s when the synchronverter-based CESS is not generated or analyzed during the current study.
installed. That is, clearing fault at time 0.35 s cannot preserve the
transient stability of the system. This is while; the CCT of the REFERENCES
system when the synchronverter-based CESS is installed at bus 1. Ebrahimi, H., Yazdaninejadi, A., Golshannavaz, S.: Demand response pro-
16 is equal to 0.746 s for a fault incepted at line 13–14. Since this grams in power systems with energy storage system-coordinated wind
CCT is more than the time of fault clearance, the systems should energy sources: A security-constrained problem. J. Clean. Prod. 335,
130342 (2022) doi: https://doi.org/10.1016/j.jclepro.2021.130342.
stay stable. Therefore, in the second scenario by installing the
2. Yazdaninejadi, A., Nazarpour, D., Talavat, V.: Optimal coordination
synchronverter-based CESS in the second scenario at bus 16, of dual-setting directional over-current relays in multi-source meshed
the rotor angle of GTSG 2 does not show instability. active distribution networks considering transient stability. IET Gener.
Transm. Distrib. 13(2), 157–170 (2019) https://doi.org/10.1049/iet-gtd.
2018.5431.
4 CONCLUSION 3. Patil, K.R., Karnik, S.R., Raju, A.B.: Impacts of distributed generation on
power system stability. In: 2021 International Conference on Intelligent
In this study, the contribution of synchronverter-based CESS Technologies (CONIT), Hubli, India, pp. 862–867 (2021) https://doi.org/
10.1109/CONIT51480.2021.9498452
on transient stability and selectivity tasks of the protection in
EBRAHIMI ET AL. 11

4. Anaya-Lara, O., Hughes, F.M., Jenkins, N., Strbac, G.: Influence of wind- 21. Yan, T., Liu, J., Niu, Q., Chen, J., Xu, S., Niu, M.: Network security pro-
farms on power system dynamic and transient stability. Wind Eng. 30(2), tection technology for a cloud energy storage network controller. Glob.
107–127 (2006) https://doi.org/10.1260/030952406778055018. Energy Interconnect. 3(1), 85–97 (2020) https://doi.org/10.1016/j.gloei.
5. Ashtiani, N.A., Khajehoddin, S.A., Karimi-Ghartemani, M.: Modeling and 2020.03.007.
stability analysis of single-phase microgrids controlled in stationary frame. 22. Li, X., Chen, G., Dong, Z.Y.: Co-optimisation model for the long-term
IEEE Trans. Power Electron. 37(7), 7759–7774 (2022) https://doi.org/ design and decision making in community level cloud energy storage sys-
10.1109/TPEL.2022.3147157. tem. IET Renewable Power Gener. 14(17), 3518–3525 (2020) https://doi.
6. Systems, L., et al.: Understanding small-signal stability of low-inertia org/10.1049/iet-rpg.2020.0612.
systems. IEEE Trans. Power Syst. 36(5), 3997–4017 (2021) 23. Liu, J., Zhang, N., Kang, C., Kirschen, D.S., Xia, Q.: Decision-making mod-
7. Ghorashi Khalil Abadi, S.A., Habibi, S.I., Khalili, T., Bidram, A.: A model els for the participants in cloud energy storage. IEEE Trans. Smart Grid.
predictive control strategy for performance improvement of hybrid energy 9(6), 5512–5521 (2018) https://doi.org/10.1109/TSG.2017.2689239.
storage systems in DC microgrids. IEEE Access. 10, 25400–25421 (2022) 24. Xia, Y., Xu, Q., Zhao, J., Yuan, X.: Two-stage robust optimisation of user-
https://doi.org/10.1109/ACCESS.2022.3155668. side cloud energy storage configuration considering load fluctuation and
8. Baghaee, H.R., Mirsalim, M., Gharehpetian, G.B., Talebi, H.A.: A general- energy storage loss. IET Gener. Transm. Distrib. 14(16), 3278–3287 (2020)
ized descriptor-system robust H∞ control of autonomous microgrids to https://doi.org/10.1049/iet-gtd.2019.1832.
improve small and large signal stability considering communication delays 25. Rappaport, R.D., Miles, J.: Cloud energy storage for grid scale applications
and load nonlinearities. Int. J. Electr. Power Energy Syst. 92, 63–82 (2017) in the UK. Energy Policy. 109, 609–622 (2017) https://doi.org/10.1016/
https://doi.org/10.1016/j.ijepes.2017.04.007. j.enpol.2017.07.044.
9. Chen, X., Pei, W., Tang, X.: Transient stability analyses of micro-grids 26. Zhong, Q.C., Weiss, G.: Synchronverters: Inverters that mimic syn-
with multiple distributed generations. In: 2010 International Conference chronous generators. IEEE Trans. Ind. Electron. 58(4), 1259–1267 (2011)
on Power System Technology, Zhejiang, China, pp. 1–8, (2010) https:// https://doi.org/10.1109/TIE.2010.2048839.
doi.org/10.1109/POWERCON.2010.5666120 27. Dong, S., Jiang, J., Chen, Y.C.: Analysis of synchronverter self-
10. Teimourzadeh, S., Davarpanah, M., Aminifar, F., Shahidehpour, M.: An synchronization dynamics to facilitate parameter tuning. IEEE Trans.
adaptive auto-reclosing scheme to preserve transient stability of micro- Energy Convers. 35(1), 11–23 (2020) https://doi.org/10.1109/TEC.2019.
grids. IEEE Trans. Smart Grid. 9(4), 2638–2646 (2018) https://doi.org/ 2945958.
10.1109/TSG.2016.2615096. 28. He, L., Shuai, Z., Zhang, X., Liu, X., Li, Z., Shen, Z.J.: Transient
11. Cheema, K.M., et al.: Virtual synchronous generator: Modifications, stabil- characteristics of synchronverters subjected to asymmetric faults. IEEE
ity assessment and future applications. Energy Rep. 8, 1704–1717 (2022) Trans. Power Delivery 34(3), 1171–1183 (2019) https://doi.org/10.1109/
https://doi.org/10.1016/j.egyr.2021.12.064. TPWRD.2019.2906766.
12. Cheema, K.M., Mehmood, K.: Improved virtual synchronous genera- 29. França, B.W., Aredes, M., Da Silva, L.F., Gontijo, G.F., Tricarico, T.C.,
tor control to analyse and enhance the transient stability of microgrid. Posada, J.: An enhanced shunt active filter based on synchronverter con-
IET Renewable Power Generation 14, 495–505 (2020) https://doi.org/ cept. IEEE J. Emerg. Sel. Top. Power Electron. 10(1), 494–505 (2022)
10.1049/iet-rpg.2019.0855. https://doi.org/10.1109/JESTPE.2021.3103836.
13. Hajiakbari Fini, M., Hamedani Golshan, M.E.: Determining optimal vir- 30. Teimourzadeh, S., Aminifar, F., Davarpanah, M., Shahidehpour, M.:
tual inertia and frequency control parameters to preserve the frequency Adaptive control of microgrid security. IEEE Trans. Smart Grid 9(4),
stability in islanded microgrids with high penetration of renewables. Electr. 3909–3910 (2018) https://doi.org/10.1109/TSG.2018.2807987.
Power Syst. Res. 154, 13–22 (2018) https://doi.org/10.1016/j.epsr.2017. 31. Rosso, R., Engelken, S., Liserre, M.: Robust stability investigation of the
08.007. interactions among grid-forming and grid-following converters. IEEE J.
14. Liu, J., Yang, D., Yao, W., Fang, R., Zhao, H., Wang, B.: PV-based vir- Emerging Sel. Top. Power Electron. 8(2), 991–1003 (2020) https://doi.
tual synchronous generator with variable inertia to enhance power system org/10.1109/JESTPE.2019.2951091.
transient stability utilizing the energy storage system. Prot. Control Mod. 32. Grainger, J.J., Stevenson, W.D.: Power system analysis McGraw-Hill, New
Power Syst. 2(1), 39 (2017) https://doi.org/10.1186/s41601-017-0070-0. York (1994), 1st edn.
15. Cheema, K.M., Milyani, A.H., El-Sherbeeny, A.M., El-Meligy, M.A.: Mod- 33. Fitzgerald, A.E., Kingsley, J.C., Umans, S.D.: Electric Machinery, McGraw-
ification in active power-frequency loop of virtual synchronous generator Hill, New York (1985)
to improve the transient stability. Int. J. Electr. Power Energy Syst. 128, 34. Zeineldin, H.H., Mohamed, Y.A.R.I., Khadkikar, V., Pandi, V.R,: A protec-
106668 (2021) https://doi.org/10.1016/j.ijepes.2020.106668. tion coordination index for evaluating distributed generation impacts on
16. Pramanik, M.A.H., Roy, T.K., Anower, M.S., Ghosh, S.K., Mahmud, M.A.: protection for meshed distribution systems. IEEE Trans. Smart Grid 4(3),
Partial feedback linearizing backstepping excitation controller for a syn- 1523–1532 (2013) https://doi.org/10.1109/TSG.2013.2263745.
chronous generator in a simple power system to improve the transient 35. Fatemi, S.S., Samet, H.: Considering DGs voltage protection in optimal
stability. In: 2021 IEEE Industry Applications Society Annual Meeting coordination of directional overcurrent relays to minimize the energy not
(IAS), Vancouver, BC, Canada, pp. 2–7 (2021) https://doi.org/10.1109/ supplied. IEEE Syst. J. 15(3), 4037–4045 (2020) https://doi.org/10.1109/
IAS48185.2021.9677294 jsyst.2020.3001378.
17. Xiong, X., Wu, C., Cheng, P., Blaabjerg, F.: An optimal damping design of 36. IEC and Standard.: Measuring relays and protection equipment - Part
virtual synchronous generators for transient stability enhancement. IEEE 151 : Functional requirements of over/under current protection. IEC
Trans. Power Electron. 36(10), 11026–11030 (2021) https://doi.org/10. 60255-151, http://ci.nii.ac.jp/naid/20001650550/en/ (2009). Accessed
1109/TPEL.2021.3074027. 17 January 2022
18. Wei, S., Zhou, Y., Huang, Y.: Synchronous motor-generator pair to enhance
small signal and transient stability of power system with high penetration
of renewable energy. IEEE Access. 5, 11505–11512 (2017) https://doi.
org/10.1109/ACCESS.2017.2716103.
19. Hou, X., Han, H., Zhong, C., Yuan, W., Yi, M., Chen, Y.: Improvement of How to cite this article: Ebrahimi, H., Yazdaninejadi,
transient stability in inverter-based AC microgrid via adaptive virtual iner- A., Golshannavaz, S.: Transient stability enhancement in
tia. In: 2016 IEEE Energy Conversion Congress and Exposition (ECCE), multiple-microgrid networks by cloud energy storage
Milwaukee, WI (2016) https://doi.org/10.1109/ECCE.2016.7855195 system alongside considering protection system
20. Liu, J., Zhang, N., Kang, C., Kirschen, D., Xia, Q.: Cloud energy storage for
residential and small commercial consumers: A business case study. Appl.
limitations. IET Gener. Transm. Distrib. 1–11 (2022).
Energy. 188, 226–236 (2017) https://doi.org/10.1016/j.apenergy.2016.11. https://doi.org/10.1049/gtd2.12539
120.