244 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 59, NO.

1, JANUARY/FEBRUARY 2023

Protection of DC Microgrids Based on Complex

Power During Faults in On/Off-Grid Scenarios
Satyavarta Kumar Prince , Shaik Affijulla , Senior Member, IEEE, and Gayadhar Panda , Senior Member, IEEE

Abstract—The DC microgrid is an effective platform for in- for power electronic converters and communication infrastruc-
tegration of renewable energy sources, energy storage systems, ture with integration of distributed generation (DG) and energy
and smart electronic loads. However, the integration of distributed storage sources (ESS) to maintain high/good quality power for
generators can result weak fault currents with change in its direc-
tion during fault conditions, which lead to failure of conventional the electric customers [1], [2]. Further, no synchronization or
over-current relays with poor coordination. The above scenario reactive power control requirements are necessary for the DC
may exhibits relay mal-operation and force the DC microgrid into power generated by the PV array, fuel cell, and ESS, with the
blackout mode which is extremely undesirable. Thus, this paper feeders being connected to DC mains via fewer conversion
proposes an complex power concept (real and imaginary), which stages.
can be extracted using Fast Fourier Transform (FFT) to construct
the effective protection schemes for rapid short-circuit fault de- Moreover, most distributed sources and energy storage de-
tection and fault isolation in a DC microgrid. To achieve fault vices naturally have DC outputs with easy to control. Therefore,
isolation, solid-state DC circuit breakers are used in conjunction DC microgrid is a better choice for integrating distributed energy
with proposed real and imaginary power, which is extracted from sources and storage devices. DC microgrid can provide a cost
the total FFT power signal (i.e. by using bus voltage and line effective platform (i.e. reducing energy losses by decreasing the
current). Further, the relay trip threshold value for real and imag-
inary power is also determined under various pole-pole (P-P) and required number of AC–DC conversion stages) for the power
pole-ground (P-G) fault scenarios in the DC microgrid. The pro- electronics application industry and electric vehicles (EVs).
posed protection scheme is validated on simple and modified IEEE Generally, the IEEE Standard 1547.4 recommends that the
9-bus DC microgrids under various P-P and P-G fault scenarios microgrid can able to operate either in On-Grid or Off-Grid
during On/Off-Grid modes through MATLAB/Simulink software. scenarios [3]. During above microgrid operations, the magnitude
The simulation results reveal that the proposed protection scheme
based on real and imaginary power has accurately identified the of fault currents are very high and low in the DC microgrid under
fault in the simulated DC microgrids. Thus, the proposed complex On-Grid and Off-Grid scenarios respectively. This behavior of
power based fault identification approach can be quite effective for DC microgrid can greatly impact the selectivity, sensitivity, and
protection of DC microgrids during On/Off-Grid scenarios. response time of the DC protection schemes [4], [5]. Further-
Index Terms—DC microgrids, DC faults, fast Fourier transform, more, DC fault currents may increase very transiently during
power converters, protection scheme, relays. short-circuit conditions due to low cable impedance and no zero-
crossings [6]. Thus, the development of an effective protection
I. INTRODUCTION scheme for DC microgrid is one of the most significant task for
the utility/industry, which is the backbone to showcase the grid
HE development of medium/low voltage direct current
T (DC) distribution systems for the successful deployment
of distributed generators and DC loads are prompting nowadays
stability [7], [8]. In contrast to AC microgrids, DC microgrids
have different protection needs due to wide variation of fault
current magnitude, quick transients and sharp discharge of DC
to attain the benefits in terms of losses and stability. Thus, DC bus capacitors.
microgrid technologies are proven to be an efficient platform In literature, various fault identification methodologies are
discussed to protect the DC microgrid against pole-pole (P-P)
Manuscript received 25 March 2022; revised 10 July 2022; accepted 5 and pole-ground (P-G) faults. The fault detection based on the
September 2022. Date of publication 12 September 2022; date of current version voltage waveform and its characteristics is discussed in [6],
19 January 2023. Paper 2022-PSEC-0237.R1, presented at the 2020 3rd Interna-
tional Conference on Energy, Power and Environment: Towards Clean Energy [9]. Specifically, the harmonic content of the distorted voltage
Technologies, Shillong, Meghalaya, India, 05-07 March 2021, and approved waveform is used to detect faults in the DC microgrid. In [10], the
for publication in the IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS by concept of voltage drop is considered for identification of faults
the Power Systems Engineering Committee of the IEEE Industry Applications
Society. This work was supported by the RECTPCL-CSR Funded Project by the DC protective relays. However, the above techniques
under Grant RECTPCL/CSR/2016-17/693. (Corresponding author: Satyavarta may not entirely reliable for DC protection scheme because
Kumar Prince.) the voltage drop can be attributed to other changes in the DC
The authors are with the Department of Electrical Engineering, Na-
tional Institute of Technology Meghalaya, Meghalaya 793003, India microgrid. Further, the presence of distortion in the current
(e-mail: [email protected]; [email protected]; gayadhar. waveform has been utilized to detect faults in the DC microgrid
[email protected]). [11]. In contrast, the use of sequence components has been found
Color versions of one or more figures in this article are available at
https://doi.org/10.1109/TIA.2022.3206171. useful only for AC systems, which require complex calcula-
Digital Object Identifier 10.1109/TIA.2022.3206171 tions and are not been suitable for protection of DC systems

0093-9994 © 2022 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
PRINCE et al.: PROTECTION OF DC MICROGRIDS BASED ON COMPLEX POWER DURING FAULTS IN ON/OFF-GRID SCENARIOS 245

[12]. Moreover, various protection schemes e.g. over-current

[13], directional over-current [14], parameter estimation [15],
traveling wave [16], and current differential [17] are suggested
for fault identification in the DC microgrid. The main challenges
of above protection schemes are synchronization, comparison,
and transmission of the trip signal within the required time frame
of current measurements under high rate of raise of DC currents
[18]. Furthermore, multiple sample differential protection is
discussed to address the time synchronisation error on a DC
distribution system; which demands a register shifted data stor-
age to compare multiple samples for effective relaying operation
[19]. This may leads to requirement of advanced communication
and synchronous measurement tools [20], which is undesirable
for economic and effective protection schemes.
Thus, the fault detection and isolation strategy based on
extracted real and imaginary power is proposed in this paper for
protection of DC microgrid. The bus voltage and line current
parameters are extracted through IED (intelligent electronic de-
vice) measurement devices to compute real and imaginary power
by utilizing FFT of bus power signal. In brief, the contribution
of this paper is summarized as,
1) Developed a DC microgrid protection scheme based on
extracted real and imaginary power through FFT of bus Fig. 1. Schematic of 4-bus low voltage DC microgrid with test faults.
voltage and line current parameters.
2) Proposed a relay tripping parameter i.e. ‘ξ’ to detect the
fault irrespective of whether a DC microgrid is in Off-Grid
shown in Fig. 1 is considered in this article, which comprises
or On-Grid mode.
of photovoltaic and energy storage battery connected to a util-
3) Validated the proposed DC protection scheme simple and
ity grid and supplies power to the load through a DC–DC
modified IEEE 9-bus DC microgrids test model under
converter. Moreover, the utility grid is connected to the DC
various fault scenarios.
microgrid through an medium/low voltage transformer and a
4) Rigorous analysis have been performed by varying dis-
voltage source converter (VSC). Usually, the photovoltaic (PV)
tance and resistance of fault to showcase the efficiency
operates in reference to the specified PV curve characteristics.
and robustness of proposed DC protection scheme.
Further, the switch SW is utilized to slide the microgrid from
5) Simulations have been executed under different modes of
On-Grid to Off-Grid mode and vice versa. During Off-Grid
operation i.e. On/Off-Grid operations to reveal the efficacy
mode, DC load is fed only from renewable energy sources and/or
of proposed DC protection scheme.
battery storage via a bi-directional converter, whereas the utility
6) The competency of proposed DC protection scheme is
grid and/or other DC microgrid sources can fed the DC load in
compared with other existing protection schemes.
On-Grid mode.
The contributions of this paper are extended version of authors
Generally, the converter of battery energy storage can operates
work [21] related to fault detection in DC microgrid by using
in buck and boost modes during On-Grid and Off-Grid scenarios
differential current approach.
respectively [13] and the interfacing converter for PV system
The remainder of the paper is organized as follows: Section II
can operates only in the boost mode. Furthermore, pole-pole
describes the architecture of On/Off Grid DC microgrid with
(P-P) and pole-ground (P-G) faults are simulated at various
faults. The proposed DC protection scheme based on real and
locations under On/Off-Grid scenarios by measuring the voltage
imaginary power for fault identification and relay trip decision
and current signals through intelligent electronic devices i.e.
algorithm is discussed in Section III. Section IV includes the
IEDs.
simulation results of various fault scenarios i.e. pole-pole (P-P)
A solar cell is composed of a P-N semiconductor junction,
and pole-ground (P-G) faults under On/Off-Grid modes by vary-
which generates a direct current (DC) when exposed to light
ing different fault parameters to reveal the accuracy of proposed
and linearly varies with solar radiation. The I–V characteristics
DC protection scheme. Further, the comparative assessment is
of the PV model can be mathematically modelled as
performed with the existing protection schemes to showcase the
efficacy of proposed DC protection scheme. Finally, conclusive
remarks are included in Section V.  q(VP V − IP V Rs ) 
IP V = IL − IO e (nkT ) −1
II. DC MICROGRID ARCHITECTURE
To demonstrate the performance of proposed DC protection VP V + I P V R s
scheme, a low voltage direct current 4-bus DC microgrid as − (1)
Rsh

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
246 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 59, NO. 1, JANUARY/FEBRUARY 2023

On rearranging (3), the fault current is expressed as

VDC
I(s) = 2 L , (4)
s + 2αs + ω 2

R∗ 1
where, α = and ω =
L LC
By applying inverse Laplace transform on (4), the general
expression for current in time domain is expressed as
Fig. 2. Equivalent electric circuit of DC microgrid with fault. i(t) = A1 eβ1 t + A2 eβ2 t (5)
where, A1 and A2 are current coefficients, which depend on
initial conditions; β1 and β2 are the roots of the characteristic
Ns nkT
VP V = equation i.e. denominator expression of (4) and can be deter-
q mined mathematically as.
 
ISC + KI (T − Tref )G + IO − IP V + NP ) 
ln β1,2 = −α ± α2 − ω 2 (6)
I O NP
NS On applying Euler theorem to (5) with initial conditions,
− RS IP V (2) the fault current i.e. i(t) response in DC microgrid can be
NP
mathematically expressed as
where, IP V , IL and IO represents solar cell, light-generated VDC  −β t
and diode currents respectively; n is diode’s ideality factor, q i(t) = e 1 − e−β2 t
L(β2 − β1 )
is charge of electron, k is the Boltzmann constant and VP V is 
solar cell output voltage; Rs and Rsh are solar cell series and −β1 e−β1 t + β2 e−β2 t
+ IL (0) (7)
shunt resistances respectively; (NS ) and (NP ) are the number of (β2 − β1 )
series and parallel connected PV arrays respectively to enhance The contribution of fault current from the converter is ignored
the voltage and current parameters respectively. in this paper due to current limiting behavior. Similarly, the fault
Thus, a maximum power point tracking (MPPT) technique is voltage v(t) in the DC microgrid can be derived from the above
implemented on photovoltaic systems to ensure for extraction RLC circuit analysis.
of highest feasible power despite the presence of insulation and Further, the fault voltage v(t), fault current i(t) are trans-
temperature variations by utilizing (1) and (2). Several MMPT formed into Fourier domain to extract the fault nature and
approaches are also discussed in the literature [22], [23], [24], mathematically expressed as
[25]. In this paper, an incremental conductance approach based
MPPT [23] is employed to extract the maximum power from √ N −1 2πink
2
solar panel arrays, boost the PV voltage as desired and stabilize I(k) = i(n)e N (8)
N n=0
the DC bus voltage i.e. VDC . The protection scheme to identify
faults on the DC microgrid is explained in the subsequent √ N −1 2πink
section. 2
V (k) = v(n)e N , (9)
N n=0
III. PROPOSED PROTECTION SCHEME where, I(k) and V (k) are Fourier domain fault current and
The natural response of RLC circuit as shown in Fig. 2 is voltage respectively, N is total number of samples, k is Fourier
presented to showcase the characteristics of a fault behavior in spectrum index value. In this paper, sample window of N = 20
DC microgrid. The strength of fault is modelled in terms of fault is considered.
resistance i.e. Rf in reference to the circuit excitation, VDC i.e. By utilizing (8) and (9), the fault complex power i.e. S(k) in
voltage across the DC bus capacitor. This approach appropriately DC microgrid is computed as
represents the electrical circuit with DC excitation to determine S(k) = I(k)V (k) (10)
fault current for protection of DC microgrid equipments.
On applying Kirchhoff’s law, the voltage characteristics of To accurately extract the fault nature in DC microgrid, the
Fig. 2 with P-P fault can be mathematically expressed in fre- authors have proposed real and imaginary FFT powers i.e. RFFT
quency domain (i.e. Laplace transform) as and IFFT respectively, which are mathematically computed as

VDC 1 RF F T = real_part(S) (11)

= LsI(s) + R∗ I(s) + I(s), (3)
s Cs IF F T = imaginary_part(S) (12)
where, VDC is the initial voltage across the capacitor C, Various simulations have been performed on 4-bus DC mi-
R∗ = R + Rf , R and L represents the line resistance and in- crogrid as shown in Fig. 1 under different fault scenarios i.e.
ductance parameters respectively of the cable. P-P and P-G faults. Interestingly, it is observed that the response

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
PRINCE et al.: PROTECTION OF DC MICROGRIDS BASED ON COMPLEX POWER DURING FAULTS IN ON/OFF-GRID SCENARIOS 247

Fig. 3. Response of fault current and voltage during (a) 0.1 Ω, (b) 2 Ω, (c) 10 Ω and (d) 100 Ω P-P fault on bus-C in On-Grid mode.

of real and imaginary FFT powers i.e. RFFT and IFFT powers
respectively are quite sensitive to the faults in the DC microgrid.
Thus, authors have considered the RFFT and IFFT powers to
effectively detect and classify the DC microgrid fault scenarios
respectively.
Moreover, a relay trip threshold can be computed based
on IFFT and RFFT powers to execute the relay trip decision
algorithm. Therefore, a complex (i.e. real and imaginary) power
based protection scheme is proposed to effectively detect and
classify the electric faults in DC microgrid. The computation
procedure for proposed relay trip threshold and relay decision
algorithm is explained in the subsequent subsections.
Fig. 4. Response of FFT Power, Real FFT Power, and IFFT Power during P-P
Fault near bus C with fault resistance of 0.1 Ω in On-Grid mode.
A. Computation of Relay Trip Threshold
This section introduces, a threshold value for the relay setting
parameter ‘ξ’ (fault tripping) to improve the relay’s sensitivity all lines is defined as ‘ξ’ to detect the fault in the DC microgrid.
and selectivity for identifying the faulty locations. The off-line Mathematically, ‘ξ’ can be defined as:
procedure steps to compute ‘ξ’ parameter for a DC microgrid
are given below: ξ = min min(max(IF F Til )) , (14)
Υ Ψ
Step 1: A simulation of DC microgrid with high resistance
faults such as P-P and P-G faults for each line ‘l’ connected where, Ψ is a set of different faults like P-P and P-G on a line l
to ith bus in the DC microgrid at a distance of 10%, 30%, at locations 10%, 30%, 50%, 70% and 90% from a bus i. Υ is a
50%, 70% and 90% from ith bus with the events are simulated set of all lines in a microgrid.
independently. At control center, the above discussed procedural steps (i.e. 1,
Step 2: For each fault on the ‘l’ line, the IF F Til is computed 2 and 3) are executed, where the configuration of DC microgrid
at the interval of 20 ms. The maximum of the calculated IF F Til is available and communicate the computed ‘ξ’ setting to the
is given by ‘max(IF F Til )’. Similarly, the process is repeated respective relays in the DC microgrid.
for each type and location of faults on line ‘l’. The minimum of
them is given by B. Algorithm for Relay Tripping Decision
min(max(IF F Til )) (13) The algorithm for tripping the relay under a fault scenario
Ψ
is based on ‘ξ’ irrespective of whether a DC microgrid is in
for a line ‘l’. Off-Grid or On-Grid mode. In the proposed algorithm, a trip
Step 3: Similarly, the process in (Step 2) is repeated for each signal by a relay is generated through the following steps:
line ‘l’ in the microgrid. Then a minimum value of (13) among Step a: Sample the voltage and current parameters of a line

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
248 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 59, NO. 1, JANUARY/FEBRUARY 2023

TABLE I
RESPONSE OF THE IFFT POWER DURING P-P FAULTS AT A VARIETY OF FAULT RESISTANCES IN ON/OFF GRID MODE

Fig. 5. Response of FFT power, real FFT power, and IFFT power during P-P Fig. 7. Response of extracted imaginary FFT Power during P-P Fault near bus
fault near bus C with fault resistance of 2 Ω in On-Grid mode. C of 0.1 Ω fault resistance in On-Grid mode.

Fig. 6. Response of FFT Power, Real FFT power and IFFT power during P-P
fault near bus C with fault resistance of 10 Ω in On-Grid mode. Fig. 8. Response of extracted imaginary FFT power during P-P fault near bus
C of 2 Ω fault resistance in On-Grid mode.

Step b: Calculate the complex power and transform into

Step e: Initiate the backup protection based on an area clas-
Fourier domain as per (10).
sified according to the exiting and traditional DC microgrid
Step c: Extract the imaginary FFT power i.e. IF F T by using
protection schemes for final trip decision. Proceed to Step a.
(12)
Step d: If IF F T > ξ then the fault has occurred in the DC
microgrid and trip signal will send to circuit breaker by the IV. SIMULATION RESULTS
activated relay. Solar and battery based DC microgrid with both On-Grid and
Due to malfunction, circuit breaker fails to clear the fault, Off-Grid mode is considered in this paper to evaluate the effec-
proceed to next step. tiveness of the proposed protection scheme. The DC microgrid

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
PRINCE et al.: PROTECTION OF DC MICROGRIDS BASED ON COMPLEX POWER DURING FAULTS IN ON/OFF-GRID SCENARIOS 249

TABLE II
RESPONSE OF THE IFFT POWER DURING P-G FAULTS AT A VARIETY OF FAULT RESISTANCES IN ON/OFF GRID MODE

30%, 50% and 70% and resistance (from 0.1Ω to 25Ω). The
performance of proposed complex power i.e. real and imagi-
nary powers response is depicted in the simulation time frame
between 0.95 s and 1.15 s. The fault is initiated at t = 1.0 s with
a fault duration of 120 ms during all simulation scenarios of this
paper.
The proposed a relay tripping parameter i.e. ‘ξ’ is computed
on applying the procedure as discussed in Section III A for
various DC microgrids to detect the fault irrespective of whether
a DC microgrid is in Off-Grid or On-Grid mode. The ‘ξ’ for
4-bus and modified IEEE 9-bus DC microgrids is ‘80’ and ‘400’
respectively.

Fig. 9. Response of extracted IFFT Power during P-P fault near bus C of 10 Ω A. Response of voltage/current Under Varying Fault
fault resistance in On-Grid mode.
Resistances
TABLE III A P-P fault i.e. F_3 as shown in Fig. 1 is simulated on the line
EVALUATION OF VARIOUS DC MICROGRID PROTECTION SCHEMES IN A ‘C-PV’ under fault resistances of 0.1 Ω, 2.0 Ω, 10 Ω and 100 Ω.
COMPARATIVE FRAMEWORK
The response of bus voltage and line current measurements
under above fault scenarios is depicted in Fig. 3. It is observed
from Fig. 3 that the fault current and voltage varies in proportion
with resistance in the DC microgrid. Further, the fault current is
very low for short duration during high fault resistance scenario
and may not be detectable by a conventional DC protection
scheme due to weak sensitivity.
Thus, it is necessary to develop an effective protection al-
as depicted in Fig. 1 is implemented on MATLAB/Simulink en- gorithm for detecting a fault in DC distribution. In this paper,
vironment. A DC supply is connected to the active load through authors have tried to showcase the efficacy of proposed DC
the bus bar with total installed capacity of 144 kW and faults protection scheme during above fault scenarios.
are assessed at various points along the DC bus. In Fig. 1, the
DC/AC source is activated by using a power converter, allowing
it to operate in either On or Off-Grid modes. The DC bus uses the B. Performance Under On-Grid Scenario
DC load, which may comprise of converter-interface renewable The proposed protection scheme is executed on DC microgrid
energy resources or energy storage system. The utility gird is in On-Grid mode as shown in Fig. 1. To showcase the efficacy,
connected to the DC bus-A; bus-C and bus-D are connected a P-P fault near bus C with varying fault resistance is simulated.
with 500 V DC bus. The response of complex FFT power (SFFT), real FFT (RFFT)
The proposed fault identification algorithm based on complex power, and imaginary FFT (IFFT) power under 0.1 Ω, 2.0 Ω and
power is validated under various faults i.e. pole-pole (P-P) and 10 Ω fault resistances is depicted in Figs. 4, 5, and 6 respectively.
pole-ground (P-G) faults under On/Off-Grid modes by varying Further, Figs. 7, 8, and 9 show the IFFT power at different fault
different fault parameters e.g. location of fault (LOF) i.e. 10%, resistance values extracted over a period of 120 ms.

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
250 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 59, NO. 1, JANUARY/FEBRUARY 2023

Fig. 10. Response of IFFT power during P-P fault near bus C of (a) 0.1 Ω, (b) 2 Ω, and (c) 10 Ω fault resistance in Off-Grid mode.

Fig. 11. Response of extracted RFFT power during (a) P-P and (b) P-G fault
near bus C with fault resistance of 2 Ω in On-Grid mode.

If no fault conditions exist, the extracted value of the IFFT(Q)

power yields zero, whereas the Q value is not equal to zero
when fault conditions exist. When the fault is removed from the
DC microgrid, the value of IFFT(Q) begins to decay. Similarly,
the IFFT values obtained at fault resistances of 2 Ω and 10 Ω
decrease during the fault; it should be noted that even in these
two cases, the value of IFFT(Q) does not equal zero during the
fault.
The simulation results reveal that the value of extracted IFFT
power is greater than proposed relay setting parameter i.e.
ξ > 80 under all P-P faults by varying fault resistances.

C. Performance Under Off-Grid Scenario

Fig. 12. Proposed fault identification scheme for DC microgrid.
This section addresses the proposed IFFT power based protec-
tion mechanism under Off-Grid mode. Similarly, a P-P fault is
initiated on bus C with varying fault resistances. The response of to zero during a pre-fault event. The simulation results reveal
imaginary FFT (IFFT) power under 0.1 Ω, 2.0 Ω and 10 Ω fault that the value of extracted IFFT power is greater than proposed
resistances is depicted in Fig. 10(a) , (b), and (c) respectively. relay setting parameter i.e. ξ > 80 under different pole-pole
The extracted value of IFFT is more considerable during a fault (P-P) faults by varying fault resistances in Off-Grid mode of
(i.e. Q is not equal to 0), but the extracted value of IFFT is equal DC microgrid.

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
PRINCE et al.: PROTECTION OF DC MICROGRIDS BASED ON COMPLEX POWER DURING FAULTS IN ON/OFF-GRID SCENARIOS 251

TABLE IV
TESTING THE PROTECTION SCHEME FOR MODIFIED IEEE 9-BUS DC MICROGRID SYSTEM UNDER VARIOUS LOADS

and multi-sample differential [19] protection strategy is used for

identification of DC microgrid faults.
The proposed and above existing protection approaches
are executed during various faults e.g. pole-pole (P-P) and
pole-ground (P-G) faults by varying different parameters i.e.
fault resistance and location under On/Off-Grid scenarios. The
efficacy of proposed protection strategy with other methods is
tabulated in Table III. Thus, it is revealed from Table III that the
proposed protection scheme based on imaginary FFT power is
quite effective, fast and robust to identify the faults in the DC
microgrid during On/Off-Grid scenarios.

E. Classification of Faults Based on Proposed RFFT Power

Further, a real FFT (RFFT) power i.e. from (11) is utilized to
classify the DC microgrid faults i.e. pole-pole (P-P) and pole-
ground (P-G) faults. A rigorous simulations are performed under
above fault scenarios to exhibit the characteristics of proposed
RFFT power.
Due to the page limit, the response of RFFT power under
under P-P and P-G faults with fault resistance of 2 Ω only is
shown in Fig. 11(a) and (b) respectively. The simulation results
reveal that the RFFT power is more negative under P-P fault,
whereas the RFFT power is close to zero during P-G fault during
Fig. 13. Schematic of modified IEEE 9 bus DC microgrid with test faults. faulted period. Thus, a threshold i.e. ξ ∗ is proposed for classi-
fication of faults in DC microgrid. The value of ξ ∗ is typically
depends on system configuration and its installed capacity. In
this paper, the value of ξ ∗ = −10000 is considered for 4-bus
Further, rigorous simulations are performed on 4-bus DC
DC microgrid. The detailed flow chart for classification of fault
microgrid during On/Off-Gird modes by varying fault resistance
in DC microgrid during On/Off-Grid scenarios is depicted in
and location of fault under pole-pole (P-P) and pole-ground
Fig. 12.
(P-G) faults. The asterisk (*) mark indicates that the fault is
occurred near to that bus in the DC microgrid. The maximum
value imaginary power extracted during fault period by installed F. Validation on Modified IEEE 9-Bus DC Microgrid
relays i.e. RA , RB , RC and RD is comprehensively tabulated in The modified IEEE 9-bus DC microgrid [26] as shown in
Tables I and II under P-P and P-G fault scenarios respectively. Fig. 13 is modelled to validate the proposed protection scheme
Thus, the proposed protection scheme based on imaginary on the MATLAB/Simulink platform. The scenarios such as
power has accurately detected the faults in Off/On-Grid scenar- pole-pole (P-P) fault, pole-ground fault, fault between source
ios in the DC microgrid. and converter and sudden change in system load are simulated
to reveal the efficacy of proposed IFFT power based on DC
microgrid protection scheme. The performance of proposed
D. Comparative Assessment of Protection Schemes protection scheme under above scenarios is provided in the
To showcase the strength, the performance of proposed pro- subsequent sections.
tection scheme based on IFFT power is assessed with other 1) Performance Under P-P Fault Scenario: In this scenario,
existing i.e. [7], [17] and [19] DC microgrid protection ap- a P-P fault is initiated between bus B-I and between bus H-PV
proaches. A resistance estimation method has been utilized in as shown in Fig. 13. To reveal the effectiveness of the proposed
[7] for detection of faults in DC microgrid. A differential [17] protection scheme, SFFT, RFFT and IFFT are calculated as per

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
252 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 59, NO. 1, JANUARY/FEBRUARY 2023

Fig. 14. Response of FFT power, real FFT power, & IFFT power during P-P fault near bus B with fault resistance of (a) 0.1 Ω, & (b) 2 Ω in On-Grid mode.

Fig. 15. Response of FFT power, real FFT power, & IFFT power during P-P fault near bus H with fault resistance of (a) 0.1 Ω, & (b) 2 Ω in On-Grid mode.

(10), (11), and (12) and IFFT threshold limits are determined extracted IFFT power of the faulty line is qualified the proposed
according to Section III A. threshold criterion i.e. ξ = 400 in the event of a P-G fault.
The response of SFFT, RFFT and IFFT powers at 0.1 Ω and Further, the extracted IFFT power shows that the pre-fault
2 Ω, as well as the zoomed-out IFFT power extracted in each mean value is ‘0.85,’ the fault peak value is ‘4000,’ and the
scenario are shown in Figs. 14(a), (b) and 15(a), (b) respectively. post-fault average value is ‘0.08’. Thus, the proposed protection
The simulation results reveal that the extracted value of IFFT scheme is also effective and robust under P-G fault scenarios in
power does not equal zero during fault and exceeds the pre- the DC microgrid.
scribed threshold limit i.e. ξ = 400, while becomes zero during 3) Performance Under a Fault Between Source and DC–DC
pre/post fault scenarios. Converter: In this case, the P-G/P-P fault is initiated between
Thus, the designated protective relays are capable of detecting photovoltaic (PV) and DC–DC converter near bus C with fault
both types of fault under different fault resistances as per the resistances of 0.1 Ω and 2 Ω. The response of proposed IFFT
proposed threshold value of IFFT power and proposed protection power during above P-G/P-P fault scenarios is depicted in
scheme is proved to be a effective and robust for DC microgrid. Figs. 17 and 18.
2) Performance Under P-G Fault Scenario: Similarly, a P–G The simulation results reveal that the above fault event can
fault scenario is initiated and the response of SFFT, RFFT, and be effectively detected by using proposed IFFT power, which is
IFFT powers are shown in Fig. 16(a) and (b) respectively. The shown in Fig. 17 (under P-G fault) and Fig. 18 (under P-P fault).
peak value of IFFT power is observed as ‘4000’ and ‘1810’ Further, the IFFT power based threshold value i.e. ξ = 400 is
during the fault period at a fault resistance of 0.1 Ω and 2 Ω realized during above fault scenarios. Although the extracted
respectively. The simulation results reveal that the maximum IFFT power can detect the fault in a short time, if the protection

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
PRINCE et al.: PROTECTION OF DC MICROGRIDS BASED ON COMPLEX POWER DURING FAULTS IN ON/OFF-GRID SCENARIOS 253

Fig. 16. Response of FFT power, real FFT power, & IFFT power during P-G fault near bus B with fault resistance of (a) 0.1 Ω, & (b) 2 Ω in On-Grid mode.

Thus, the proposed protection scheme based on IFFT power

is effective and able to detect the P-P/P-G fault between source
and DC–DC converter in the DC microgrid.
4) Performance Under Sudden Change in Load: In this case,
the DC load is varied from base load of 250 kW to sudden
change in load of 450 kW to showcase the performance of
proposed protection scheme for DC microgrid. The faults such as
pole-pole (P-P) fault, pole-ground fault are simulated at different
loading scenarios.
The efficacy of proposed protection scheme in terms of IFFT
power is tabulated in Table IV. The simulation results reveal that
the proposed IFFT power under sudden change in DC load is
less than the designated relay trip threshold i.e. ξ = 400. Thus,
Fig. 17. Response of FFT power, real FFT power, & IFFT power during P-G the proposed protection scheme based on imaginary FFT power
fault between PV and DC–DC converter with fault resistance of 2 Ω in On-Grid
mode. is insensitive to sudden change in DC load of microgrid.
Furthermore, the average computational burden of the pro-
posed protection scheme is approximate less than 0.1 s for
execution of relay trip decision algorithm.

V. CONCLUSION
In this article, an effective fault identification strategy based
on real and imaginary FFT powers is presented to suite for
protection of DC microgrids. The imaginary FFT power based
protection scheme uses voltage and current information from
the faulted bus. The proposed algorithm can detect and clas-
sify the fault in bus segment by utilizing complex FFT power,
real FFT (RFFT) power, and imaginary FFT (IFFT) power at
various fault locations in DC microgrid. Further, a relay trip
threshold is proposed to accurately identify the pole–pole (P–P)
and pole–ground (P–G) faults under various fault resistance
Fig. 18. Response of FFT power, real FFT power, & IFFT power during P-P
fault between PV and DC–DC converter with fault resistance of R = 0.1 Ω, in
and location during On/Off-Grid scenarios. The efficacy of the
On-Grid mode. proposed protection scheme is rigorously validated on 4-bus
and modified IEEE 9-bus DC microgrids. The simulation results
reveal that the IFFT and RFFT protection scheme has effectively
system does not detect the fault quickly, the power converter detected classified the faults under all abnormal conditions.
may fail. Thus, the proposed complex power based protection scheme

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.
254 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 59, NO. 1, JANUARY/FEBRUARY 2023

is quite suitable for protection of DC microgrids in On/Off-Grid [20] R. Mohanty, S. Sahoo, A. K. Pradhan, and F. Blaabjerg, “A cosine
modes. similarity-based centralized protection scheme for DC microgrids,” IEEE
Trans. Emerg. Sel. Topics Power Electron., vol. 9, no. 5, pp. 5646–5656,
Oct. 2021.
REFERENCES [21] S. K. Prince, S. Affijulla, and G. Panda, “Fault detection in IEEE 9-bus
DC microgrid system using differential current method,” in Proc. 3rd Int.
[1] F. Nejabatkhah and Y. W. Li, “Overview of power management strategies Conf. Energy, Power and Environ.: Towards Clean Energy Technol., 2021,
of hybrid AC/DC microgrid,” IEEE Trans. Power Electron., vol. 30, no. 12, pp. 1–6.
pp. 7072–7089, Dec. 2015. [22] M. Veerachary and V. Khubchandani, “Analysis, design, and control of
[2] J. Zhao and Z. Xu, “Ramp-limited optimal dispatch strategy for switching capacitor based buck–boost converter,” IEEE Trans. Ind. Appl.,
PV-embedded microgrid,” IEEE Trans. Power Syst., vol. 32, no. 5, vol. 55, no. 3, pp. 2845–2857, May/Jun. 2019.
pp. 4155–4157, Sep. 2017. [23] G. A. Raiker et al., “Current control of boost converter for PV interface with
[3] R. Ahmadi and M. Ferdowsi, “Improving the performance of a line reg- momentum-based perturb and observe MPPT,” IEEE Trans. Ind. Appl.,
ulating converter in a converter-dominated DC microgrid system,” IEEE vol. 57, no. 4, pp. 4071–4079, Jul./Aug. 2021.
Trans. Smart Grid, vol. 5, no. 5, pp. 2553–2563, Sep. 2014. [24] N. Kumar, I. Hussain, B. Singh, and B. K. Panigrahi, “Self-adaptive InC
[4] Y. Yang, C. Huang, and Q. Xu, “A fault location method suitable for LV algorithm for swift and ripple-free maximum power harvesting from PV
DC line,” IEEE Trans. Power Del., vol. 35, no. 1, pp. 194–204, Feb. 2020. array,” IEEE Trans. Ind. Inform., vol. 14, no. 5, pp. 2031–2041, May 2018.
[5] F. Liu, W. Liu, X. Zha, H. Yang, and K. Feng, “Solid-state CB snubber [25] H. Toodeji and S. Aghaei, “Domestic PV system with feedback
design for transient overvoltage suppression at bus fault interruption linearization-based control strategy for module-level MPPT under partial
in LV DC microgrid,” IEEE Trans. Power Electron., vol. 32, no. 4, shading condition,” J. Modern Power Syst. Clean Energy, vol. 9, no. 6,
pp. 3007–3021, Apr. 2017. pp. 1530–1539, 2021.
[6] A. Shekhar, L. Ramírez-Elizondo, S. Bandyopadhyay, L. Mackay, and [26] X. Li and K. Li, “Analysis and design of connection between solar farm
P. Bauera, “Detection of series arcs using load side voltage drop for and IEEE 9 bus system,” in Proc. IEEE Asia-Pacific Conf. Image Process.,
protection of low voltage DC systems,” IEEE Trans. Smart Grid, vol. 9, Electron. Comput., 2021, pp. 797–801.
no. 6, pp. 6288–6297, Nov. 2018.
[7] G. K. Rao and P. Jena, “Fault detection in DC microgrid based on
the resistance estimation,” IEEE Syst. J., vol. 16, no. 1, pp. 1009–1020,
Mar. 2022.
[8] C. Yuan, M. A. Haj-ahmed, and M. S. Illindala, “Protection strategies for Satyavarta Kumar Prince received the M.Tech. de-
medium-voltage direct-current microgrid at a remote area mine site,” IEEE gree in electrical engineering from the National In-
Trans. Ind. Appl., vol. 51, no. 4, pp. 2846–2853, Jul./Aug. 2015. stitute of Technology Meghalaya, Shillong, India, in
[9] H. F. Habib, M. M. Esfahani, and O. A. Mohammed, “Investigation of 2018, where he is currently working toward the Ph.D.
protection strategy for microgrid system using lithium-ion battery dur- degree in protection and control in a DC microgrid.
ing islanding,” IEEE Trans. Ind. Appl., vol. 55, no. 4, pp. 3411–3420, His research interests include the DC fault analysis,
Jul./Aug. 2019. estimation of line parameter during fault and fault
[10] L. Mackay, A. Shekhar, B. Roodenburg, L. Ramírez-Elizondo, and P. detection, and location algorithms for DC microgrid
Bauer, “Series arc extinction in DC microgrids using load side voltage systems.
drop detection,” in Proc. IEEE First Int. Conf. DC Microgrids, Georgia,
USA, 2015, pp. 239–244,.
[11] D. Paul, “Phase-ground fault current analysis and protection of a high-
resistance grounded power system,” IEEE Trans. Ind. Appl., vol. 56, no. 4, Shaik Affijulla (Senior Member, IEEE) received the
pp. 3306–3314, Jul./Aug. 2020. M.Tech. degree in electrical engineering from the
[12] A. Rahmati and R. Adhami, “A fault detection and classification technique National Institute of Technology Hamirpur, Hamir-
based on sequential components,” IEEE Trans. Ind. Appl., vol. 50, no. 6, pur, India, in 2011 and the Ph.D. degree in electrical
pp. 4202–4209, Nov./Dec. 2014. engineering from the National Institute of Technology
[13] S. Chakraborty, M. D. Weiss, and M. G. Simoes, “Distributed intelligent Meghalaya, Shillong, India, in 2018. He is currently
energy management system for a single-phase high-frequency AC micro- an Assistant Professor with the Department of Elec-
grid,” IEEE Trans. Ind. Electron., vol. 54, no. 1, pp. 97–109, Feb. 2007. trical Engineering, the National Institute of Technol-
[14] H. M. Sharaf, H. H. Zeineldin, and E. El-Saadany, “Protection coordina- ogy Meghalaya. His research interests include the
tion for microgrids with grid-connected and islanded capabilities using cyber security of smart grids, Big Data analytics,
communication assisted dual setting directional overcurrent relays,” IEEE phasor measurement units, wide area monitoring and
Trans. Smart Grid, vol. 9, no. 1, pp. 143–151, Jan. 2018. protection, and power system dynamics.
[15] Y. Liu, A. P. Sakis Meliopoulos, Rui Fan, and L. Sun, “Dynamic state
estimation based protection of microgrid circuits,” in Proc. IEEE PES
Gen. Meeting, 2015, pp. 1–5.
[16] K. A. Saleh, A. Hooshyar, and E. F. El-Saadany, “Ultra-high-speed
traveling-wave-based protection scheme for medium-voltage DC micro- Gayadhar Panda (Senior Member, IEEE) received
grids,” IEEE Trans. Smart Grid, vol. 10, no. 2, pp. 1440–1451, Mar. 2019. the Ph.D. degree in electrical engineering in 2007
[17] S. Dhar, R. K. Patnaik, and P. K. Dash, “Fault detection and location of from Utkal University, Bhubaneswar, India. In 2013,
photovoltaic based DC microgrid using differential protection strategy,” he joined the National Institute of Technology (NIT)
IEEE Trans. Smart Grid, vol. 9, no. 5, pp. 4303–4312, Sep. 2018. Meghalaya, Shillong, India, as an Associate Profes-
[18] O. C. Heredia et al., “Electronic differential based on active disturbance sor. He is currently a Professor with the Electrical
rejection control for a four in-wheel drive EV,” in Proc. IEEE Int. Power Engineering Department, NIT. His research interests
Renewable Energy Conf., 2021, pp. 1–6. include automatic generation control, stability im-
[19] C. Li, P. Rakhra, P. J. Norman, G. M. Burt, and P. Clarkson, “Multi-sample provements using flexible alternating current trans-
differential protection scheme in DC microgrids,” IEEE Trans. Emerg. Sel. mission system devices, power quality, power elec-
Topics Power Electron., vol. 9, no. 3, pp. 2560–2573, Jun. 2021. tronic converters, and distributed power generation.

Authorized licensed use limited to: Toronto Metropolitan University Library. Downloaded on August 26,2023 at 20:52:54 UTC from IEEE Xplore. Restrictions apply.