Electric Power Systems Research 187 (2020) 106432

Contents lists available at ScienceDirect

Electric Power Systems Research

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

Developing A Smart Power-Voltage Relay (SPV-Relay) with no T

Communication System for DC Microgrids
Eatmad W. Nahasa, Diaa-Eldin A. Mansoura, , Hossam A. Abd el-Ghanya, M.M. Eissab,c
⁎

a
Department of Electrical Power and Machines Engineering, Faculty of Engineering, Tanta University, Seperbay, Tanta 31511, Egypt
b
Department of Electrical Engineering, Faculty of Engineering, Helwan University, Helwan 11421, Egypt
c
Department of Electrical and Computer Engineering, College of Engineering, Sultan Qaboos University, Muscat, Oman

ARTICLE INFO ABSTRACT

Keywords: Microgrids are considered a prospective way towards improving electric service resiliency, reducing costs, and
Renewable energy sources (RESs) upgrading service reliability. DC microgrids offer many advantages rather than AC microgrids. In spite of the
DC microgrids numerous features and advantages of DC microgrids, their protection faces significant challenges such as self-
Fault characteristics limited current of photovoltaic (PV) systems, long time constant of wind energy systems, dependability on
Protection scheme
communication systems, etc. The paper introduces a new protection scheme for DC microgrids using the rate of
power (dP) and rate of voltage (dV) and mapping them as dP-dV profile. The new scheme is titled as smart
power/voltage relay (SPV-Relay). The proposed scheme is applicable for all types of renewable energy sources
(RESs) and energy storage systems independent of the power rating and configuration of the DC microgrid. The
proposed concept of SPV-Relay is described and the method used for fault discrimination is explained. The relay
characteristics are developed considering all the DC microgrid components. Three operating zones are identified
on dP-dV profile to discriminate between various fault types and locations. The sensitivity and stability of the
proposed relay are evaluated under different fault conditions as well as different control schemes and opera-
tional scenarios for DC microgrid.

1. Introduction integration of the DC RESs such as PV, fuel cell and wind energy, higher
efficiency in supplying the DC loads such as electric vehicles and LED
1.1. Motivation lights, and absence of synchronization for generators or connected
buses [7].
Recently, the development of microgrids is a growing global trend. However, DC microgrids are generally suffering from some chal-
Microgrids can help communities of utility users to achieve key policy lenges. Firstly, the stability of DC microgrids is still under investigation
goals and enhance the use of renewable energy. Moreover, the high [8,9]. In addition, the erection of DC microgrids lacks the required
penetration of renewable energy sources (RESs), such as wind energy standards. Currently, extensive efforts are exerted toward standardiza-
systems and photovoltaic (PV) systems, accelerated the deployment of tion of generation, transmission, and distribution of DC power. Finally,
microgrids. Microgrids can operate either in a non-autonomous way if the most critical and important challenge of DC microgrid operation is
interconnected to the main grid, or in autonomous way if isolated from the fault behavior and protection system as follows. First, the DC fault
the main grid [1]. The flexibility to transit among these modes makes current level has a high magnitude and a high rising rate [10] due to the
microgrids a reliable solution in case of lack of the main grid energy integration of power electronic converters [11] and energy storage
supply or in case of emergency. If the various sources in a microgrid devices [12]. This necessitates fast fault detection and isolation in order
were efficiently managed and coordinated, this will be reflected posi- to avoid possible damage for power electronic switches. Also, the fault
tively on the overall performance of the system [2]. According to the in DC microgrids is highly sensitive against fault resistance [13],
type of bus voltage, microgrids can be classified to DC microgrids, AC making resistive faults difficult to be detected or undetectable in some
microgrids and hybrid microgrids [3–5]. DC microgrids offered many cases. For example, in [14] the inductance estimation scheme failed in
benefits over AC ones [6]. These benefits include the less number of detecting resistive faults of 0.6 Ω near the DC bus, unless artificial line
energy conversion stages with a corresponding higher efficiency, simple inductance is added. In addition, response of RESs against faults is

⁎
Corresponding author.
E-mail address: [email protected] (D.-E.A. Mansour).

https://doi.org/10.1016/j.epsr.2020.106432
Received 31 December 2019; Received in revised form 18 April 2020; Accepted 24 May 2020
0378-7796/ © 2020 Elsevier B.V. All rights reserved.
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

challenging. For wind energy systems (WESs), their fault response is discriminate between various fault types and locations. Finally, dif-
greatly affected by their dynamic behavior [15]. For photovoltaic (PV) ferent case studies are presented to validate the effectiveness of the
systems, they have self-limited current making the current magnitude proposed relay.
not suitable for protection purposes. Another challenge in the protec- The novelty of the paper and main contributions can be summarized
tion of DC microgrids is the bidirectional nature of energy storage de- as follows:
vices making the reversal of current is not sufficient for identifying the
fault location. (1) A new relaying scheme based on local measurements with no
communication is proposed for DC microgrids using the features of
1.2. Review and lack of previous studies power and voltage derivatives.
(2) In addition to the absence of communication, the proposed relay
Previous studies on the protection of the DC microgrid ignored some offers fast, robust, and selective detection of faults regardless the
of abovementioned technical challenges that are critical in the protec- source behind it.
tion of the DC microgrid. In [16], non-unit protection scheme was (3) The proposed scheme is independent either on the adopted power
proposed for DC microgrid based on the first and second derivative of converter topology for interfacing RESs or on the operational mode.
the current. The main drawback in this protection scheme is the need of (4) Unlike other protection schemes [14,16,17,22], the proposed
adaptability for the different configuration of DC microgrid and the scheme acts well with resistive faults, self-limited sources such as
dependence of the current derivatives on the initial line current and photovoltaic (PV) system, and long time constant sources such as
fault resistance. Also, this scheme has high sensitivity to noisy currents. wind energy system (WES).
As a result, both the current derivative and current integral were used (5) The proposed scheme can be applied for other topologies and rat-
in [17] for ring-type DC microgrid to decrease the sensitivity against ings of DC microgrids, since it is based on power and voltage de-
noisy currents. However, both schemes in [16] and [17] will fail with rivatives rather than their absolute values.
RESs that have self-limited current such as PV systems or that have with
long time constant such as wind energy systems. Accordingly, these
2. System description
schemes cannot be used for radial feeders connecting such sources. A
protection scheme based on inductance estimation was proposed in
The DC microgrid is schematically shown in Fig. 1. It contains RESs,
[14], through which the observed inductance of each protective device
energy storage unit (ESU), DC bus, and loads. PSCAD/EMTDC computer
is estimated, and the faulted line exhibits the smaller inductance in
package was used for modeling purposes. The common RESs are WES
referring to the line inductance. In spite of inductance estimation based
and PV system, and thus, these sources are considered in the present
scheme prover proper selectivity, fast operation, and independence of
study. Regarding energy storage, Li-ion batteries were used due to their
microgrid topology, it requires artificial line inductance to be added at
high energy density. Controlled converters were used to interface the
both ends of each line resulting in an additional cost and a negative
previous sources with the DC bus. For WES and PV systems, the con-
impact on the system stability [9,18].
verters were controlled either for maximum power point tracking
Differential protection schemes were introduced in [19–21] for the
(MPPT) or for constant voltage depending on the operational conditions
DC microgrid. These schemes could achieve high selectivity. However,
[25]. For batteries; the converters were controlled for constant voltage.
they require synchronized measurements and reliable communication
The effective grounding system and DC bus configuration were taken
infrastructure. Unsynchronized measurements make these schemes
into consideration. The following subsections provide further details
vulnerable to false tripping in case of external faults. Also, dependence
about the modeling of individual components in the DC microgrid.
on communication for data transfer requires high bandwidth commu-
nication system adding further costs to the protection schemes and
implies further time delay. To avoid the time delay of differential 2.1. PV unit modelling
protection schemes, it was proposed to use the communication for only
blocking and tripping other relays to assist overcurrent protection The equation and the electrical equivalent circuit of the PV that
scheme [22]. However, this scheme is still based on overcurrent pro- identify the relationship between the PV output current and terminal
tection with the same aforementioned limitations. Also, a commu- voltage are well known [26]. In order to confirm the effectiveness of the
nication system is still used to transfer blocking and tripping signals.
In [23], the DC voltage and the direction of DC current were con- Ipv DCbus IBatt1
sidered. Using the current direction could overcome the synchroniza- PV Array
Vpv
DC-DC Bi-direction Li-ion
tion problem encountered in differential schemes. But, it will fail in case DC-DC
battery# 1
of energy storage devices due to their bidirectional nature. Moreover,
Vpv_ref
Voltage
the time constant of the current direction depends mainly on the feeder
MPPT control Edc Voltage and
control current control
inductance. Using both current and voltage derivatives was proposed in
Wind AC-DC IBatt2
[24] to protect battery energy storage system. But, this technique will uncontrolled DC-DC Bi-direction
Turbine #1 Li-ion
fail with the RES that has a self-limited current such as PV or that has rectifier DC-DC
battery# 2
higher time constant such as wind energy system. Optimum
speed MPPT control Edc Voltage and
current control
1.3. Paper organization and novelty

Voltage
In this paper, it is aimed to develop a Smart Power-Voltage Relay Wind AC-DC Loads
balancer
(SPV-Relay) without a communication system for selective, reliable, Turbine #4
uncontrolled DC-DC (P-P)&(P_G)
rectifier
and fast protection of DC microgrids. First, the main features of fault Edc/2
response in DC microgrids are investigated considering RESs, energy Optimum Voltage
speed MPPT control control
storage devices and their interfaced power electronic. Secondly, the
hypothesis behind the proposed SPV-Relay is presented and discussed.
DC Link
Then, the discrimination between normal and fault conditions based on
capacitor
SPV-Relay is clarified for various sources incorporated in the microgrid.
Next, zone classification for the proposed relay is determined to Fig. 1. The detailed model of the considered DC microgrid.

2
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

proposed protection scheme, two topologies with different control According to the proposed scheme, the discrimination and zone clas-
methods were used in the present study to interface PV unit to the DC sification of healthy operation and different fault conditions are illu-
bus. The first topology is based on using boost chopper controlled for strated. Finally, the SPV-Relay characteristics are determined according
MPPT to attain higher efficiency from the PV system. In this case, the to this classification.
reference voltage signal is the output from the MPPT technique as il-
lustrated in Fig. 1. The second topology is based on using boost chopper
controlled for constant voltage. In this case, the reference voltage signal 3.1. Fault features in DC microgrid
is the desired DC bus voltage.
For analyzing fault response of the DC microgrid, different fault
2.2. Wind energy unit modelling types are considered at different locations in the DC microgrid. There
are two types of faults in DC microgrids, pole to pole (P-P) fault and
The WES consists of a variable speed wind turbine and permanent pole to ground (P-G) fault. Fault features in DC microgrids can be
magnet synchronous generator due to its advantages compared to other summarized as follows.
wind turbine generators. Similar to PV unit, two topologies with dif- First, there are significant difference between fault current magni-
ferent control methods were used. The first topology is controlled for tudes in case of P-P fault and P-G one, where P-P faults has much higher
MPPT and consists of a three-phase uncontrolled rectifier followed by a current magnitudes compared to P-G ones. This makes it difficult to
buck-boost converter as shown in Fig. 1. This configuration enables the adjust the relay settings based on the current magnitude. Relay setting
simplicity of MPPT control with low number of controlled switches. The based on P-P faults will lead to a delay in the operation of protection
extracted power from wind (Pw) as a function of air density and wind scheme for P-G faults. On the contrary, relay setting based on P-G faults
speed is described in [27–29]. The output voltage from the generator is can cause mis-coordination for P-P faults. Second, ESU has a large
rectified through the uncontrolled rectifier and its output DC voltage is current contribution, but with a very high sensitivity to fault resistance
related to the optimum speed of wind turbine as in [30]. So, the MPPT [12] leading to a problem with overcurrent detection. This represent a
can be reached from the buck-boost converter that is controlling the big challenge in DC microgrids as a very fast tripping is required in such
rectifier output voltage. The second topology is controlled for constant microgrids compared to AC ones [36,37].
DC bus voltage through using three phase controlled rectifier. Third, fault response of RESs imposes difficulty with the overcurrent
relays as follows. For a fault on PV unit, the fault current is self-limited
2.3. Battery unit modelling by the short circuit current of PV, which is very close to the normal
current at maximum power point as illustrated in Fig. 2. Under solid
There are many forms for energy storage such as electrochemical fault condition, the current changed from the normal current at max-
battery, super capacitor, compressed air energy storage, super- imum power point, which was about 195 A, to the short circuit current
conducting magnetic energy storage, and flywheel energy storage of PV unit, which was 212 A. Thus, the change in current is only about
[31,32]. Owing to their high energy density, Li-ion batteries were taken 17 A, which is small compared to the rated PV current.
into account in the current study. Thus, their accurate model [33] has For fault on WES controlled for MPPT, the full-scale buck-boost
been built and connected to the DC bus through bidirectional converter converter is controlled so that the DC output voltage of the rectifier
as shown in Fig. 1. The main function of the battery is to tightly reg- corresponds to the optimum speed of wind turbine [30]. So, the MPPT
ulate the DC bus voltage through covering the energy mismatch be- control scheme forces the WES to continuously supply the maximum
tween the load requirements and RESs [34]. This role is carried out by power related to the wind speed, but with a relatively large time con-
using bidirectional buck-boost converter with proper control scheme. stant corresponding to the control action. Fig. 3 shows the fault re-
sponse of WES in this study when controlled for MPPT. The time con-
2.4. DC bus bar modelling stant was about 350 ms, and the fault current increased to only about
180 A in the first 10 ms compared to 110 A under normal case. As a
Selection of the DC operating voltage is an important factor in de- result of abovementioned challenges, not only the overcurrent protec-
termining the efficiency, cost and safety of a DC microgrid. In this tion will fail in fast fault detection for such renewable energy sources,
paper, 380 V was approved as the voltage level for standardized use to but also the first and second current derivatives will fail.
cover telecommunication and distribution system requirements. This The voltage magnitude and its derivative have the same limitations
level of the voltage necessitates effective grounding and appropriate of the current derivative regarding high sensitivity to fault resistance. In
protection schemes. addition, the voltage-based schemes cannot provide adequate in-
There are three types of grounding schemes in DC systems, one pole formation about the fault location, weather on the connecting line/
grounding, mid-point grounding and isolated grounding. The mid-point cable or on the DC bus.
grounding is the preferable one from the viewpoint of human safety, Owing to these challenges associated with the current and voltage
detection of the ground fault and capability of creating a three wire DC based protection schemes and also the limitation of their derivatives,
system [35]. So, three wire system with mid-point grounding is con- especially with RESs, a new decentralized protection scheme is
sidered in the present DC microgrid. Moreover, a half bridge voltage
balancer is implemented for balancing the voltage between the two
poles at ± 190 V.

3. The Proposed Protection Scheme for DC Microgrid

For selective and fast protection of a DC microgrid, investigation of

the fault response must be considered. The fault analysis of the DC
microgrid exhibited certain features that are essential in the design of
the proposed protection scheme. Accordingly, the fault response of the
DC microgrid is firstly identified with highlighting the protection
challenges associated with currently existing schemes. Then, the hy-
pothesis of the proposed protection scheme is presented and the moti-
vation behind using power and voltage derivatives is discussed. Fig. 2. The output current from PV system during fault.

3
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

1000 5
WES output current (A)
τ
800 4

Voltage (V)
600 3
400 2
200 1
0 0
1 2 3 4 5 20 40 60 80 100
Time (s) State of charge (%)
Fig. 3. The output current from wind energy system during fault. Fig. 5. Voltage characteristic of one Li-ion battery cell.

proposed based on local measurements of the power derivative (dP) and 3.3. Discrimination of normal and faults in dP-dV profile
the voltage derivative (dV) mapped as dP-dV profile.
As illustrated previously, all RESs in the DC microgrid can be clas-
sified into one of the aforementioned sources, according to their be-
3.2. Hypothesis of the proposed protection scheme havior. In this section, the attitude of each unit in the DC microgrid will
be studied under normal and fault conditions.
The theory of the new protection scheme is based on the dP-dV
profile. This profile discriminates between the faulty and healthy con- 3.3.1. PV unit
dition for different types of electrical sources. The electrical sources can It is well-known that the PV cell is considered a current source.
be categorized into three types which are constant voltage source When the PV array is connected to the DC bus via a converter with
(CVS), constant current source (CCS) and constant power source (CPS). MPPT control the whole unit acts as a constant power source related
The voltage-power characteristics of the different types are shown in directly to the PV radiation, regardless the value of the bus voltage. So,
Fig. 4. the relation between the output power and voltage can be described as
According to the constant voltage source, the voltage of the source is (1), where K is a constant representing the power at a certain radiation
maintained constant during the allowable operation modes, while the and the value of α turns to zero. The differentiation of the power with
power is varied with the supplied current. An example of the CVS is respect to the voltage results in:
batteries where their internal voltage is kept approximately constant
over the wide range of its state of charge as shown in Fig. 5, which dP
= lim
K×
depicts the voltage characteristics for a single battery cell. Moreover, dV 0 V1 (2)
the control of batteries could regulate its output voltage to be tightly
which can be written as follows:
regulated at 380 V. For the CCS, it has a constant current regardless of
the power. So, the voltage is varied proportionally with the output K×
dP = × dV
power. An example of CCS is PV cell under fault condition, where the V1 (3)
output current is constant regardless the fault resistance as shown in
Fig. 2. Finally, the CPS is a source in which the voltage and the current The value of the differential voltage dV is extremely small during the
are varied to maintain the output power constant. This category of normal operation according to the DC bus voltage regulation. Also, the
sources is represented clearly in RESs controlled for MPPT at certain factor linking the dP with dV is insignificant, where α in the numerator
environmental conditions. is very small.
The proposed power-voltage relationship by analogy with the load Therefore, all variations under normal operation will be located
equation [38] can be defined for the different types of sources as fol- around the origin of the dP-dV profile as shown in Fig. 6. But, during
lows: the fault condition, the PV unit converts to a constant current source

P = K *V (1) Normal operation

dP Fault condition
where K is constant and α is referred as voltage sensitivity parameter.
The parameter α will be zero for constant power source, where the
power is always constant regardless the value of the voltage. For con-
stant current and constant voltage sources, the parameter α becomes
equal to 1 and the constant K is replaced by the current I.

dV = -ve dV = +ve
dV
Voltage

Voltage
V oltage

Power Power Power

(a) Constant voltage (b) Constant current (c) Constant power

Fig. 4. Power-voltage characteristics of different types of sources Fig. 6. Zone classification for PV unit and wind unit on dP-dV profile.

4
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

and the power-voltage relationship is expressed as follows: Table 1

Evaluation of power derivative for battery during normal operation and fault
P = Isc × V (4)
condition.
where Isc is the PV short circuit current at a specific radiation and α is Normal operation
equal to unity. So, the differentiation of the power to the voltage will IdV I = ± Imax dVmax = ± 0.05
lead to: VdI V = Edc dI = ± 2Imax
Fault condition
dP = Isc × dV (5) IdV I = Isc |dVmax| >> 0.05
VdI V < Edc dI >> ± 2Imax
It is clear that the correlation factor between dP and dV becomes the
short circuit current of the PV array which is greater enough than the
normal state factor ( V1 ) . Moreover, the value of dV during fault
K×
Normal operation
condition is larger than that during normal state operation. It is clear dP Fault condition
that the change in power is significantly large compared to the change
in current that adopted in some of the previous schemes, making the

dP = +ve
proposed scheme more effective than both overcurrent scheme and
current derivative based scheme. The change in current becomes worse
with resistive faults or at low radiation. Also, when using fault current
limiters with microgrids [39], they suppress the spikes in the fault
current adding further challenge regarding current derivative based
scheme. On the other hand, the change in power keeps substantial dV = -ve dV = +ve
changes under these conditions. dV
According to the polarity of dV, the corresponding quarter on the
dP-dV profile can be determined. Under pole-pole (P-P) fault, the vol-
tage and power will drop exhibiting negative voltage derivative as well
as negative power derivative. Consequently, the P-P fault will lie in the

dP = -ve
third quarter as shown in Fig. 6. Under pole-ground (P-G) fault on one
pole, a temporary overvoltage occurs on the healthy pole causing a
positive voltage derivative. So, the P-G fault will be located in the first
or fourth quarter. It is evident that, the discrimination between the
normal operation and the fault condition became recognizable and Fig. 7. Zone classification for battery unit on dP-dV profile.
clear from the dP-dV profile.

3.3.2. Wind energy unit dP (pu)

The wind unit is considered a constant power source under MPPT
control. So, the relation between the output power and the voltage can
be described as (1), giving that K is a constant representing the power at
a certain wind speed and the value of α turns to zero. The differentia-
tion of the power with respect to the voltage results in similar equation
to that depicted in (3). Accordingly, the factor linking the dP with dV is Zone 2 0.5
insignificant and all normal variations are located around the origin of Zone 3 dEdc (pu)
the dP-dV profile similar to that previously discussed with PV unit. -0.5
During the fault condition, the wind unit continues as a constant
Zone 1

power source exhibiting a power-voltage relationship similar to that of

K×
the normal state. But, the magnitude of the factor ( V1 ) during the fault
becomes much greater than that during the normal state owing to the
small voltage value in the denominator. Also, the value of dV during
-2 2 ×10-3
fault is large. So, through using dP-dV profile, the fault condition can be
distinguished obviously from the normal state with similar locations to Fig. 8. The SPV-relay characteristics.
that applied with PV unit, as shown in Fig. 6.

3.3.3. Battery unit

As previously explained, the battery can be considered a constant Table 2
Zones classification for the SPV-relay.
voltage source due to the converter action in regulating the DC bus
voltage at a constant value through changing the battery current I. Zone 1
Accordingly, the relationship between the power and the voltage for PV unit Transmission line If fault in zone 1, the relay will trip
Wind energy unit instantaneously
batteries can be expressed as:
Battery unit
P=V×I (6) Zone 2
PV unit Pole-Pole on DC bus If fault in zone 2, the relay will trip
Consequently, the power derivative can be expressed as a function of Wind energy unit after a time delay of 0.2 ms
the voltage derivative and the current derivative according to the fol- Battery unit
lowing equation: Zone 3
PV unit Pole-Ground on DC If fault in zone 3, the relay will trip
dP = V × dI + I × dV (7) Wind energy unit bus after a time delay of 0.2 ms
Battery unit
During the normal state, the control of the battery tightly regulates
the DC bus voltage. So, the voltage derivative dV is very small. In ad-
dition, the value of battery current during the normal operation is

5
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

Measure all currents (IPV, IW, Ibatt,

Iload, and Ic) & DC bus voltage (Edc)
at certain sample k

Calculate Pk = Vk × Ik k=k+1

Calculate ∆P = Pk+1 – Pk
∆Edc = Edc k+1 – Edc k

Yes
∆P < -0.5 ?

No
∆Edc < -0.002 ? No

Yes Yes
∆P > 0.5 ?

No
Yes
∆Edc < -0.002 ?

No

Time delay ∆Edc > 0.002 ? Time delay

Trip this relay Trip all relays No Trip all relays

Fault on P-P fault on P-G fault on

Normal case
TL/cable DC bus DC bus

Zone 1 Zone 2 Zone 3

Fig. 9. Flowchart of the proposed SPV-relay operation.

limited to the maximum battery current (Imax). The evaluation of dP follow the same process described in (6) and (7).
value under normal operation is depicted in Table 1. The maximum
change in voltage equals to ± 5%, while the maximum change in cur- 3.3.4. Transmission line faults
rent equals to ± 2Imax, when the battery changes from charging mode For a fault on any transmission line (TL), the current transducer
to discharging mode or vice versa. It is clear that the maximum change beside this fault will measure the fault current contributed from all
of dP under normal operating conditions is small. On the other hand, other sources beside the DC link capacitor discharging current. This
under fault condition, the current derivative dI and voltage derivative results in a very high reversable power within very small time span.
dV will have high values as illustrated in Table 1, exhibiting higher Accordingly, large negative power difference will be detected by this
values of power derivative dP. transducer. This forces the operating point to move downwards along
During the P-P fault on the DC bus, the battery will contribute the the dP axis between the third and fourth quarters. In this case, current
fault with a large current magnitude. So, the current derivative will be transducers on other TLs will measure the currents contributed from
very large with a positive value. On the other hand, the voltage deri- only individual units, which are flowing in the positive direction and
vative will have a negative value. Consequently, one term of the right have an extremely lower values than that measured by the faulted line
side in (7) will be positive and the other term will be negative. Ac- transducer.
cording to the dominant effect, the power derivative can have a positive
or a negative value. Keeping that the voltage derivative is negative, this 3.4. Calculations of dP-dV
case of fault will be located either in the second quarter or in the third
quarter, as represented schematically in Fig. 7. Meanwhile, the P-G The scheme is based on sensing the sampled DC current (Ik) and
fault leads to positive voltage derivative and subsequent positive power voltage (Vk) and calculating the power at each sample (Pk) as follows:
derivative making this fault type located in the first or fourth quarter
Pk = Vk × Ik (8)
similar to PV unit and wind energy unit. It is important to point out
that, when PV system or WES is controlled for constant voltage, they where k is the sampling instant and the power derivative dP/dt can be

6
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

61
Cable/T.L. Fault PR Protective Relay
Without noise
PPV 60 With noise

PPV (pu)
B1 B2
PV Unit PR PW
F3
B4 B3 Wind 59
IPV PR
Unit
IW F3
58
PBatt1
Battery B5 B6
PR 57
Unit 1
F3 IBatt1 PBatt2 0.500 0.505 0.510 0.515 0.520
B8 B7 Battery Time (s)
PR
Unit 2
IBatt2 (a) PV power with and without noise
Edc Edc
F1 (P-P) PL_PP
B10 Loads 0.8
PR
B9 (pole-pole)
DC Link Iload_PP
PR 0.4
Capacitor

dPPV (pu)
Ic PL_PG Zone 2 Zone 3
B11 Loads
F2 (P-G) PR 0.0
(pole-ground)
Iload_PG
-0.4
Fig. 10. The configuration of the DC microgrid with the measurement points Zone 1
and fault locations. -0.8 -3
-10x10 -5 0 5 10
2.0 dEdc (pu)
60 PPV (b) The dP-dV profile at the terminals of PV unit.
Edc 1.5 Fig. 12. Voltage and power characteristics under noisy conditions in measured
PPV (pu)

Edc (pu)

40 PV power.
1.0

20 0.5 calculated using backward finite difference approximation as follows:

dP P (to + t ) P ( to ) P
= lim =
0 0.0 dt t 0 t t (9)
1 2 3 4 5
Time (s) where ΔP can be obtained as:
(a) The voltage and power at the terminals of PV unit. P = Pk + 1 Pk (10)

1.0 1.0 where Pk+1 and Pk are the power at the current sample and previous
sample, respectively. Similarly, the ∆Edc is given as follows:
dPPV
0.5 dEdc 0.5 Edc = Edck + 1 Edck (11)
dPPV (pu)

dEdc (pu)

In the present study, the sampling frequency was selected 20 kHz cor-
0.0 0.0
responding to time interval of 50 μs between samples. This sampling
time is practical due to the recent advances in digital technologies and
-0.5 -0.5
microprocessors. For increasing the robustness of the proposed scheme
against noise, a moving average window constituting of three succes-
-1.0 -1.0
sive samples was used [40]. Thus, the window size is 150 μs, which is a
1 2 3 4 5
very small time span compared to the required tripping time in DC
Time (s)
grids. The window moves with a step of one sample.
(b) The voltage and power derivatives at the terminals of PV unit.

3.5. SPV-relay characteristics

0.8
According to the previous discrimination of normal and faults in dP-
0.4
dV profile, the relay characteristics can be determined. These include
dPPV (pu)

Zone 3
Zone 2 zone classification and relay setting. For zone classification it can be
0.0
classified three different faulted zones in addition to the normal oper-
ating zone, which is located around the origin. The faulted zones re-
-0.4 present P-P fault on DC bus, P-G fault on DC bus, and faults on TLs.
Zone 1 Fig. 8 shows the proposed zone classification on dP-dV profile and
-0.8 -3 Table 2 shows their description. Regarding zone 1, if the fault were
-10x10 -5 0 5 10
dEdc (pu) detected in this zone by the dP-dV profile of any unit, this means that
the fault lies in the transmission line connecting this unit. Therefore, the
(c) The dP-dV profile at the terminals of PV unit.
relay of this unit will trip. For P-P fault at the DC bus, it will be detected
Fig. 11. Voltage and power characteristics for a change in PV radiation. in zone 2 by the dP-dV profile of PV unit, wind energy unit and battery
as explained previously. Finally, the P-G fault at the DC bus will be

7
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

200 1.6 80 1.6 80 1.6

Pbatt1 Pw PPV
150 1.4 60 1.4 60 1.4
Edc Edc Edc
Pbatt1 (pu)

PPV (pu)
Edc (pu)

Edc (pu)

Edc (pu)
PW (pu)
1.2 1.2 1.2
100 40 40
1.0 1.0 1.0
50 20 20
0.8 0.8 0.8
0 0.6 0 0.6 0 0.6
0.9 1.0 1.1 1.2 1.3 1.4 0.9
1.1 1.2 1.3 1.4 1.0 0.9 1.0 1.1 1.2 1.3 1.4
Time (s) Time (s) Time (s)
(a) The voltage and power at the terminals of each unit.

0.6 6 0.5 6 0.5 6

dPbatt1
0.4 3 0.0 3 0.0 3
dEdc
dPbatt1 (pu)

dPPV (pu)
dEdc (pu)

dEdc (pu)

dEdc (pu)
dPW (pu)

0.2
0 -0.5 0 -0.5 0
0.0
-3 -1.0 dPw -3 -1.0 dPPV -3
-0.2 dEdc dEdc
-3 -3 -3
-0.4 -6x10 -1.5 -1.5 -6x10 -6x10
0.96 1.00 1.04 1.00 0.96 1.04
0.96 1.00 1.04
Time (s) Time (s) Time (s)
(b) The voltage and power derivatives at the terminals of each unit.

1.0 1.0 1.0

0.05 ms 0.05 ms 0.05 ms
0.5 0.5 0.5
dPbatt1 (pu)

dPPV (pu)
dPW (pu)

0.0 0.0 0.0

-0.5 -0.5 -0.5

-1.0 -3
-1.0 -3
-1.0 -3
-10x10 -5 0 5 10 -10x10 -5 0 5 10 -10x10 -5 0 5 10
dEdc (pu) dEdc (pu) dEdc (pu)
(c) The dP-dV profile at the terminals of each unit.
Fig. 13. Voltage and power characteristics for 0.5 Ω resistive P-P fault on the DC bus.

1.0 1.0 1.0

0.05 ms 0.05 ms 0.05 ms
0.5 0.5 0.5
dPbatt1 (pu)

dPPV (pu)
dPW (pu)

0.0 0.0 0.0

-0.5 -0.5 -0.5

-1.0 -3
-1.0 -3
-1.0 -3
-20x10 -10 0 10 20 -20x10 -10 0 10 20 -20x10 -10 0 10 20
dEdc (pu) dEdc (pu) dEdc (pu)
Fig. 14. Voltage and power characteristics for solid P-P fault on the DC bus.

sensed in zone 3 by all units. For zone 1 operation the corresponding proposed relay operation and discrimination algorithm.
relay will trip instantaneously, while for zone 2 and zone 3 operation, The second step of determining relay characteristics is to set the
the relays were set with a time delay of 0.2 ms to keep selectivity and numerical threshold values in per unit (pu). The numerical threshold
prevent improper tripping in case of TL fault adjacent to the DC bus. It values depend on the sampling frequency (fs) and the maximum ex-
is important to point out that the delay time of 0.2 ms is highly suffi- pected change in the considered quantity with respect to time [16]. To
cient for zone 1 relay to trip since any TL fault will enter into zone 1 set the thresholds values, the following simplified equation is proposed:
directly after fault occurrence due to the large negative power differ-
ence, as discussed above. Fig. 9 illustrates the flowchart for the

8
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

5 8 80 2.0
0 dPPV 6 PPV
dIPV 60 1.5
Edc
dPPV (kW)

dIPV (A)

Edc (pu)
PPV (pu)
-5 4
40 1.0
-10 2
-15 0 20 0.5

-20 -2 0 0.0
1.000 1.002 1.004 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Time (s) Time (s)
(a) Current and power differences for PV system. (a) The voltage and power at the terminals of PV unit

5 0.8 4 10
0 0.6 PPV
5
dPW (kW)

Edc

dIW (A)

dPPV (pu)
2

dEdc (pu)
-5 0.4
0
-10 0.2
dPW 0
-15 dIW 0.0 -5
-20 -0.2 -2 -10x10
-3
1.000 1.002 1.004 0.99 1.00 1.01 1.02 1.03 1.04 1.05
Time (s) Time (s)
(b) Current and power differences for WES. (b) The voltage and power derivatives at the terminals of PV unit.
Fig. 15. Current and power differences for solid P-P fault on the DC bus.
0.4
Xmax
Xset = ± + SM 0.15 ms
fs × tmin (12) 0.2
dPPV (pu)

where ∆Xset is the setting of the defined quantity, ∆Xmax is the max- 0.0
imum change in the defined quantity, tmin is the minimum time to reach
the maximum change in the quantity, and SM is the safety margin. -0.2
For the considered DC microgrid the voltage derivative and power
derivative setting can be defined as follows. For the voltage derivative, -0.4
Eq. (12) can be rewritten as: -10x10
-3
-5 0 5 10
Vmax dEdc (pu)
Vset =± + SM
fs × tmin (13) (c) The dP-dV profile at the terminals of PV unit.

Fig. 16. Voltage and power characteristics for 0.5 Ω resistive P-G fault on the
The considered maximum change in the regulated voltage of the DC
DC bus.
bus is ± 5% of the rated value. In the considered DC microgrid, the
minimum time to reach ± 5% was measured 2.5 ms, which is close to
that obtained in [16]. Accordingly, the first term in (13) will investigated. The first case study is the operation under a change in PV
be ± 1×10−3. With keeping additional 1×10−3 as a safety margin, the radiation and the second one is the operation in the presence in noise.
threshold value of the voltage derivative was set as ± 2×10−3 pu. All results were presented in pu for a base voltage of 380 V and a base
For the power derivative, Eq. (12) can be rewritten as: power of 1 kW.

Pmax
Pset = ± + SM 4.1. Case 1: Change in PV radiation
fs × tmin (14)

The maximum change in the power can be obtained considering the In this case, a typical PV radiation was applied to the PV array as
daily load power in the DC microgrid [41]. In this study, the maximum shown in Fig. 11(a) and the battery energy storage could regulate the
change in the power for the time span of 2.5 ms was assumed DC bus voltage at 1 pu. The power derivative and the voltage derivative
about ± 50% of the rated load. This maximum change is considered with time is presented in Fig. 11(b). It is clear that the variation is
high to account for any variations in solar radiation, wind speed and/or within the controllable limits. So, the variation in dP-dV profile is re-
loading. By substitution in (14) and keeping acceptable safety margin, presented around the origin point within the normal zone as depicted in
the threshold value of the power derivative was set as ± 0.5 pu. The Fig. 11(c).
setting values for the voltage and power derivatives were added to
Fig. 8. 4.2. Case 2: Noise effect

4. Operation under Healthy Conditions The measured signals are vulnerable to noise, that can be originated
from high-frequency converter switching or from measurement equip-
The considered DC microgrid was replotted in Fig. 10 with in- ment. Accordingly, it is important to evaluate the effectiveness of the
dicating points of measurements, locations of protective relays and proposed protection scheme in the presence of noise. Since the DC link
possible faults. This section investigates the operation of the proposed capacitor suppresses the expected noise in measured voltages [42], the
protection scheme under healthy conditions. Two case studies are noise effect in the present study will be reflected on the measured

9
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

100 1.2 1.0

PPV
50 1.0 0.05 ms
Edc 0.5
PPV (pu)

Edc (pu)
0

dPPV (pu)
0.8
-50
0.0
-100 0.6

-150 0.4 -0.5

1.00 1.01 1.02 1.03 1.04 1.05
Time (s)
-1.0 -3
(a) The voltage and power at the terminals of PV unit -10x10 -5 0 5 10
dEdc (pu)
2 4
(a) The PV converter is operated in voltage control mode.
0 2
dPPV (pu)

dEdc (pu)
1.0
-2 0
0.05 ms
-4 -2 0.5

dPW (pu)
-3
-6 -4x10
0.996 0.998 1.000 1.002 1.004 0.0
Time (s)
(b) The voltage and power derivatives at the terminals of PV unit.
-0.5
6
-1.0 -3
3 -10x10 -5 0 5 10
dPPV (pu)

dEdc (pu)
0
(b) The wind energy converter is operated in voltage control mode.
-3 0.05 ms Fig. 18. The dP-dV profile for 0.5 Ω resistive P-P fault on the DC bus with
voltage control topology for RESs.
-6 -3
-10x10 -5 0 5 10
dEdc (pu) wind energy unit, the power exhibited a transient decrease due to the
(c) The dP-dV profile at the terminals of PV unit. limited current contribution in PV or the low response of contributed
Fig. 17. Voltage and power characteristics for 0.5 Ω resistive P-P fault at the current in wind energy unit. This was reflected as a decrease in the
mid-point of the TL connecting the PV unit. power derivative and voltage derivative with a consequent location in
the third quarter and zone 2 on dP-dV profile for PV and WES relays.
The indicated time on each dP-dV profile is the time taken to reach the
currents and corresponding powers. The noise level usually varies be-
trip zone. So, P-P fault was properly detected by all relays within 0.05
tween 30 dB and 60 dB signal to noise ratio [20]. The proposed scheme
ms. Fig. 14 depicts the corresponding dP-dV results for solid fault,
was evaluated for the worst possible noise level of 30 dB in measured
which indicated that the change in dP-dV profile becomes more pro-
powers. Fig. 12 shows an example for noisy condition in measured PV
nounced making it easy to detect the fault. To evaluate the current
power with the corresponding dP-dV profile. It is clear that power and
derivative based scheme, the difference current (dI) was obtained for
the voltage derivatives for noisy signals are still located within the
PV system and WES with the solid fault case, as shown in Fig. 15. Also,
normal zone.
the difference power was depicted on the same figure for the sake of
comparison. The time span between samples is 50 μs similar to that
5. Operation under Faulty Conditions used with current derivative based scheme [16]. The maximum dif-
ference currents in case of PV system and WES were 3.5 A and 0.34 A,
In this section, the proposed scheme was validated for different fault respectively, which is very far from the threshold values used in [16].
cases including different fault types and locations. For all cases, the This is attributed to the self-limited fault current in case of PV system
fault instant was set at t = 1 s and the fault resistance was considered and slow response fault current in case of WES. On the other hand, the
0.5 Ω, since fault resistance in DC microgrids usually has a maximum difference powers exhibited significant values either for PV system or
value of 0.6 Ω [13,16]. The presented cases include also the operation for WES.
of RESs with MPPT topology or with voltage control topology.

5.1. Case 1: P-P fault on the DC bus (F1) 5.2. Case 2: P-G fault on the DC Bus (F2)

Fig. 13 shows the obtained results for a resistive P-P fault on the DC During a P-G fault on the DC bus, the faulted pole voltage will de-
bus with a fault resistance of 0.5 Ω. During this fault, the voltage crease and the healthy pole voltage will exhibit a transient increase.
dropped and the battery power increased due to the increase in the Depending on this behavior, the voltage sensor on the healthy pole will
current contribution. So, the battery power derivative will have a po- detect an increase in the voltage with a corresponding increase in the
sitive value, while the voltage derivative will have a negative one. Thus, voltage derivative as shown in Fig. 16(a) and Fig. 16(b). While, the
the points of this fault on dP-dV profile for battery relay will be located whole DC bus voltage will be slightly affected causing negligible effect
in the second quarter, specifically in zone 2. But, for the PV unit and the in the powers of PV system and WES and slight effect in battery power.

10
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

Therefore, the points of this fault will be located in the first quarter, 6. Conclusion
specifically in zone 3 as shown in Fig. 16(c).
A new smart power-voltage relay (SPV-relay) has been developed
using localized measurements with no communication system. The
5.3. Case 3: P-P fault on transmission lines (F3)
scheme is built using the power and voltage derivatives and mapping
them as dP-dV profile. The proposed protection scheme has been tested
In this case, the fault was occurred at the mid-point of the TL con-
with various renewable energy sources as well as battery energy sto-
necting the PV unit. In this case, the current transducer on this TL rage. All the system components and power electronic interfaces have
measured a negative current contributed from all other sources beside
been modeled. For renewable energy sources, two different topologies
the DC link capacitor discharging current. This caused the power and have been considered. The first topology is controlled for MPPT and the
power derivatives at the terminals of PV unit to exhibit a significant
second topology is voltage-controlled topology for constant DC bus
decrease as shown in Fig. 17(a) and 17(b). This caused the power de- voltage. First, the fault features in DC microgrids have been discussed
rivative in Fig. 17(c) to enter into zone 1 on dP-dV profile directly at the
and the main challenges have been highlighted. Then, SPV-relay
first sample following the fault, i.e. after 50 μs. In this case, the relay at characteristics were built including zone classification and relay setting.
the terminals of PV unit will trip before tripping the relays on other TLs
The relay was evaluated under normal conditions and the operating
due to the coordination time delay of 0.2 ms. points were located properly in the normal zone on dP-dV profile. For
faulty conditions, different fault locations and fault types have been
5.4. Case 4: P-P fault on the DC bus (F1) with voltage control topology for investigated with solid and resistive faults. The proposed relay suc-
RESs ceeded in proper and fast detection of all faults within 0.2 ms after fault
occurrence preventing any damage for the microgrid components.
In some cases, RESs are integrated with the DC bus through con-
verter topologies controlled for constant voltage. In this mode, the CRediT authorship contribution statement
batteries are disconnected and either the PV converter or the wind
energy converter regulates the DC bus voltage, while the other con- Eatmad W. Nahas: Software, Validation, Formal analysis, Writing -
verter is kept operating in MPPT mode. Fig. 18(a) illustrates the dP-dV original draft. Diaa-Eldin A. Mansour: Conceptualization,
profile for 0.5 Ω resistive P-P fault on the DC bus, when the PV con- Methodology, Investigation, Formal analysis. Hossam A. Abd el-
verter is operated in voltage control mode and the wind energy con- Ghany: Investigation, Resources, Data curation, Writing - review &
verter is operated in MPPT mode. The drop in voltage forced the voltage editing. M.M. Eissa: Conceptualization, Methodology, Visualization,
derivative to enter into zone 2 enabling to successfully detect the fault Supervision.
after about 0.1 ms. Similarly, when the wind energy converter is op-
erated in voltage control mode and the PV converter is operated in Declaration of competing interests
MPPT mode, the points on dP-dV profile was located successfully into
zone 2 as shown in Fig. 18(b). None

Appendix

Table A1

Table A-1 Shows the parameters of all the components in the DC Microgrid
PV unit
PV array Prated=60 kW, Voc=375 V, Isc= 212 A
PV converter parameter C1=1 mF, L=5 mH, C2= 2 mF
Switching frequency=2 kHz
Wind unit
Wind turbine Prated=4 ×12 kW
Rotor radius=3.2 m
Rated wind speed=11 m/s
PMSG parameter Prated=10 kW, Vrated=380 V, frated=50 Hz
Stator resistance = 0.034 pu
Stator leakage inductance = 0.097 pu
Magnetic strength=1.1pu
LCL filter Lg= 0mH, Cf = 220 µF, Lf =35 mH
Wind converter parameter L= 10 mH, C=2 mF
Switching frequency=150 Hz
Battery 1 unit
Battery 1 rating 40 kWh, VBatt=350 V, IBatt= 115 Ah
Battery converter parameter L= 0.5 mH, C= 10 µF
Switching frequency= 2kHz
Battery 2 unit
Battery 2 rating 20 kWh, VBatt =300 V, IBatt =70 Ah
Battery converter parameter L= 0.3 mH, C= 5 µF
Switching frequency = 2 kHz
Voltage balancer
Parameter L= 8.5 mH, C=10 5 µF
Switching frequency = 3 kHz
DC link capacitor C= 20 mF, LT.L=5 mH, RT.L=0.05Ω

11
E.W. Nahas, et al. Electric Power Systems Research 187 (2020) 106432

References Trans. Power Del. 34 (3) (2019) 869–878.

[22] A. Shabani, K. Mazlumi, Evaluation of a communication-assisted overcurrent pro-
tection scheme for photovoltaic-based DC microgrid, IEEE Trans. Smart Grid 11 (1)
[1] W. Feng, et al., A review of microgrid development in the United States – A decade (2020) 429–439.
of progress on policies, demonstrations, controls, and software tools, Appl. Energy [23] A.A.S. Emhemed, K. Fong, S. Fletcher, G.M. Burt, A.L. Lvdc, Validation of fast and
228 (2018) 1656–1668. selective protection scheme for an LVDC distribution network, IEEE Trans. Power
[2] M.E. Meral, D. Çelík, A comprehensive survey on control strategies of distributed Del. 32 (3) (2017) 1432–1440.
generation power systems under normal and abnormal conditions, Annu. Rev. [24] E.W. Nahas, D.A. Mansour, H.A. Abd el-Ghany, M.M. Eissa, Accurate Fault Analysis
Control 47 (2019) 112–132. and Proposed Protection Scheme for Battery Energy Storage System Integrated with
[3] J.J. Justo, F. Mwasilu, J. Lee, J. Jung, AC-microgrids versus DC-microgrids with DC Microgrids, 2018 Twentieth International Middle East Power Systems
distributed energy resources: A review, Renew. Sust. Energy Rev. 24 (2013) Conference (MEPCON), Egypt, Cairo, Egypt, 2019, pp. 911–917.
387–405. [25] J. Rocabert, A. Luna, F. Blaabjerg, P. Rodríguez, Control of power converters in AC
[4] Y. Wang, Y. Li, Y. Cao, Y. Tan, L. He, J. Han, Hybrid AC/DC microgrid architecture microgrids, IEEE Trans. Power Electron. 27 (11) (2012) 4734–4749.
with comprehensive control strategy for energy management of smart building, Int. [26] S. Shongwe, M. Hanif, Comparative analysis of different single-diode PV modeling
J. Electr. Power Energy Syst. 101 (2018) 151–161. methods, IEEE J. Photovoltaics 5 (2015) 938–946.
[5] D. Çelik, M.E. Meral, Current control based power management strategy for dis- [27] A.O. Ibrahim, T.H. Nguyen, D.C. Lee, S.C. Kim, A fault ride-through technique of
tributed power generation system, Control Eng. Pract. 82 (2019) 72–85. DFIG wind turbine systems using dynamic voltage restorers, IEEE Trans. Energy
[6] M.M. Eissa, Medium-Voltage Direct Current Concept, Modeling, Operation, Control, Convers. 26 (3) (2011) 871–882.
Protection, and Management—An Extensive Article Review, in: M.M. Eissa (Ed.), [28] Z. Wu, X. Dou, J. Chu, M. Hu, Operation and control of a direct-driven PMSG-based
Medium Voltage Direct Current Grid, Academic Press, 2019, pp. 1–41. wind turbine system with an auxiliary parallel grid-side converter, Energies 6 (7)
[7] A.T. Elsayed, A.A. Mohamed, O.A. Mohammed, DC microgrids and distribution (2013) 3405–3421.
systems: An overview, Electr. Power Syst. Res. 119 (2015) 407–417. [29] D M Yehia, D A Mansour, W Yuan, Fault ride-through enhancement of PMSG wind
[8] Y. Yokomizu, D.M. Yehia, D. Iioka, T. Matsumura, Formulated representation for turbines with DC microgrids using resistive-type SFCL, IEEE Trans. Appl.
upper limitation of deliverable power in low-voltage DC distribution system, IEEJ Supercond. 28 (4) (2018) art. no. 5603105.
Trans. Power Energy 131 (4) (2011) 362–368. [30] N.A. Ahmed, M.R. Alrashidi, Development of an efficient utility interactive com-
[9] D.M. Yehia, Y. Yokomizu, D. Iioka, R. Watanabe, T. Matsumura, A novel approach bined wind / photovoltaic / fuel cell power system with MPPT and DC bus voltage
to deliverable power in low-voltage DC distribution system on the basis of voltage regulation, Electr. Power Syst. Res. 81 (5) (2011) 1096–1106.
stability, IEEJ Trans. Electr. Electron. Engineering 6 (2011) 395–402. [31] T.M.I. Mahlia, T.J. Saktisahdan, A. Jannifar, M.H. Hasan, H.S.C. Matseelar, A re-
[10] D. Salomonsson, L. Soder, A. Sannino, Protection of low-voltage DC microgrids, view of available methods and development on energy storage; technology update,
IEEE Trans. Power Del. 24 (3) (2009) 1045–1053. Renew. and Sust. Energy Rev. 33 (2015) 532–545.
[11] J. Yang, J.E. Fletcher, J. O'Reilly, Short-circuit and ground fault analyses and lo- [32] D.A. Mansour, Energy Storage Technologies in MVDC Microgrids, in: M.M. Eissa
cation in VSC-based DC network cables, IEEE Trans. Ind. Electron. 59 (10) (2012) (Ed.), Medium Voltage Direct Current Grid, Academic Press, 2019, pp. 189–207.
3827–3837. [33] M. Chen, G.A. Rincón-Mora, Accurate electrical battery model capable of predicting
[12] D.M. Yehia, D.A. Mansour, Modeling and analysis of superconducting fault current runtime and I-V performance, IEEE Trans. Energy Convers. 21 (2) (2006) 504–511.
limiter for system integration of battery banks, IEEE Trans. Appl. Supercond. 28 (4) [34] E.K. Belal, D.M. Yehia, A.M. Azmy, Adaptive droop control for balancing SOC of
(2018) art. no. 5603006. distributed batteries in DC microgrids, IET Gener. Transm. Distrib. 13 (20) (2019)
[13] D.A. Mansour, M.M. Eissa, Dynamic Modeling and Fault Analysis of Medium- 4667–4676.
Voltage Direct Current Microgrids, in: M.M. Eissa (Ed.), Medium Voltage Direct [35] D. Kumar, F. Zare, A. Ghosh, DC microgrid technology: system architectures, AC
Current Grid, Academic Press, 2019, pp. 127–149. grid interfaces, grounding schemes, power quality, communication networks, ap-
[14] M. Shamsoddini, B. Vahidi, R. Razani, Y. Mohamed, A novel protection scheme for plications, and standardizations aspects, IEEE Access 5 (2017) 12230–12256.
low voltage DC microgrid using inductance estimation, Int. J. Electr. Power Energy [36] S. Beheshtaein, R.M. Cuzner, M. Forouzesh, M. Savaghebi, J.M. Guerrero, DC mi-
Syst. 120 (2020) art. no. 105992. crogrid protection: A comprehensive review, IEEE J. Emerg. Sel. Top. Power
[15] M.M. Eissa, M.M.A. Mahfouz, G.M.A. Sowilam, A new developed smart grid pro- Electron. (2020) Early Access.
tection technique with wind farms based on positive sequence impedances and [37] M. Monadi, M.A. Zamani, J. Ignacio, A. Luna, P. Rodriguez, Protection of AC and
current angles, Electr. Power Syst. Res. 178 (2020) art. no. 106020. DC distribution systems embedding distributed energy resources: A comparative
[16] A. Meghwani, S.C. Srivastava, S. Chakrabarti, A non-unit protection scheme for DC review and analysis, Renew. and Sust. Energy Rev. 51 (2015) 1578–1593.
microgrid based on local measurements, IEEE Trans. Power Del. 32 (1) (2017) [38] D.M. Yehia, Y. Yokomizu, D. Iioka, T. Matsumura, Deliverable-power dependence
172–181. on distribution-line resistance and number of loads in low-voltage DC distribution
[17] A. Meghwani, R. Gokaraju, S.C. Srivastava, S. Chakrabarti, Local measurements- system, IEEJ Trans. Electr. Electron. Engineering 7 (1) (2012) 23–30.
based backup protection for DC microgrids using sequential analyzing technique, [39] L. Chen, et al., Application and design of a resistive-type superconducting fault
IEEE Syst. J. 14 (1) (2020) 1159–1170. current limiter for efficient protection of a DC microgrid, IEEE Trans. Appl.
[18] A. Tah, D. Das, An enhanced droop control method for accurate load sharing and Supercond. 29 (2) (2019) art. no. 5600607.
voltage improvement of isolated and interconnected DC microgrids, IEEE Trans. [40] M. Elgeziry, M. Elsadd, N. Elkalashy, T. Kawady, A. Taalab, M.A. Izzularab, Non-
Sust. Energy 7 (3) (2016) 1194–1204. pilot protection scheme for multi-terminal VSC–HVDC transmission systems, IET
[19] M. Monadi, C. Gavriluta, A. Luna, J.I. Candela, P. Rodriguez, Centralized protection Renew. Power Gener. 13 (16) (2019) 3033–3042.
strategy for medium voltage DC microgrids, IEEE Trans. Power Deliv. 32 (1) (2017) [41] K. Sayed, A.G. Abo-Khalil, A.S. Alghamdi, Optimum resilient operation and control
430–440. DC microgrid based electric vehicles charging station powered by renewable energy
[20] S. Dhar, P.K. Dash, Differential current-based fault protection with adaptive sources, Energies 12 (22) (2019) art. no. 4240.
threshold for multiple PV-based DC microgrid, IET Renew. Power Gener. 11 (2017) [42] Q. Liu, S. Wang, A.C. Baisden, F. Wang, D. Boroyevich, EMI suppression in voltage
778–790. source converters by utilizing dc-link decoupling capacitors, IEEE Trans. Power
[21] V. Nougain, S Mishra, A.K. Pradhan, MVDC microgrid protection using a centralized Electron. 22 (4) (2007) 1417–1428.
communication with a localized backup scheme of adaptive parameters, IEEE

12