Energy Reports 13 (2025) 5931–5951

Contents lists available at ScienceDirect

Energy Reports
journal homepage: www.elsevier.com/locate/egyr

Research paper

Optimal coordination of non-directional overcurrent relays for active

distribution network using new voltage-current relay characteristic
Nattapol Haupala , Sillawat Romphochai , Krischonme Bhumkittipich *
Department of Electrical Engineering, Faculty of Engineering, Rajamangala University of Technology Thanyaburi (RMUTT), Pathum Thani 12110, Thailand

A R T I C L E I N F O A B S T R A C T

Keywords: The variation in short-circuit current levels caused by the distributed generation (DG) installation leads to the
Overcurrent relays coordination failure of the overcurrent relay in the protection system. This paper presents a new voltage-
Relay coordination controlled overcurrent characteristic of an overcurrent relay. The relay characteristic proposed in this
Interval linear programming
research is designed to improve the performance of overcurrent relays by using a pickup voltage multiplier. The
Distribution network
pickup voltage multiplier consists of three key parameters: pickup voltage, parameter D, and time multiplier
Distributed generation
setting (TMS). This article presents a method for determining the optimal values for pickup voltage parameters D
and TMS using nonlinear programming and the interval linear programming solution to minimize the total
tripping times and maintain the coordination of the primary and backup relays under short-circuit conditions in
the radial system with and without DG integrations under different sizes and locations. The proposed relay
characteristic is verified on the IEEE 13-bus and IEEE 33-bus radial distribution system with synchronous-based
DG (SBDG) and wind turbine-based DG (WTDG). In addition, the simulation results of the proposed relay
characteristic are compared with the conventional standard relays and the overcurrent relay characteristics in
the literature. The simulation results of the proposed relay characteristic demonstrate its capability for fast short-
circuit clearance, reduction in the total operation time of the protection scheme, and maintaining coordination
between the primary and backup relays under increased short-circuit current conditions in power systems with
SBDG and WTDG installation.

1. Introduction increased short-circuit current. Limiting the power rating of DG and

appropriately selecting the installation location within the power system
The installation of Distributed Generation (DG) in a distribution by considering the fault current levels and the operating times of pro­
system (DS) has several aims, such as increasing the reliability of the tection relays is a commonly used method to reduce short-circuit current
system, improving the power quality of the system, and reducing dis­ in the system, to avoid impacts on the existing protection scheme (Zhan,
tribution losses (Barker and De Mello, 2000). However, the increasing 2016; Lee et al., 2011). Most traditional protection systems in electrical
penetration of DG in the system, along with the higher power ratings of distribution networks are non-communication-based systems. There­
DG and the distance between the fault location and the DG installation fore, one way to address the coordination failure between the primary
point, contribute to higher short-circuit currents in the power system. and backup relays is by pre-setting appropriate relay settings based on
This affects the coordination of protection devices within the system the system’s load current and short-circuit current (JunYing et al.,
(Sa’ed et al., 2013). Moreover, the increase in short-circuit current due 2024). Another method is to pre-plan the relay settings in anticipation of
to the growing penetration of DG has led to an incorrect perception of future increases in short-circuit currents due to the penetration of DG
fault current levels by protection devices, failing proper relay coordi­ (Huchel and Zeineldin, 2016). These approaches help the traditional
nation (Girgis and Brahma, 2001). Solutions to the problem of protec­ protection system accommodate changes in short-circuit currents when
tion relay coordination failure caused by increased short-circuit currents there is increased DG penetration in the electrical network.
in power systems with DG installations include reducing the Enhancing the operating characteristics of overcurrent protection
short-circuit current to avoid changes in the relay operating character­ relays is another method to address the coordination failure between
istics and improving the relay operating characteristics to match the primary and backup relays caused by the increase in short-circuit

* Corresponding author.
E-mail address: [email protected] (K. Bhumkittipich).

https://doi.org/10.1016/j.egyr.2025.05.053
Received 26 January 2025; Received in revised form 28 April 2025; Accepted 19 May 2025
Available online 23 May 2025
2352-4847/© 2025 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

current in power systems with DG installations. This approach involves

developing from the traditional non-communication-based protection
systems, to meet user requirements, such as reducing the overall oper­
ating time of all protection relays in the system (Hong et al., 2021),
minimizing violations of the Fault Ride Through (FRT) requirements for
doubly-fed induction generator (DFIG)-based wind turbine DG (WTDG)
(Lim and Lim, 2020), and coordinating with superconducting fault
current limiters (SFCLs) to reduce short-circuit currents in the power
system (Saleh et al., 2015). Although the proposed method can achieve
the fault current reduction, the FRT capability and the grid code satis­
faction, the cost effective is still very high. For solving both an economic
and grid code requirements, modifying the operating characteristics of
overcurrent protection relays can be done easily because modern over­
current relays have been developed as one of the functions within
microprocessor-based protective relays. These relays can be designed
and programmed with greater flexibility to meet user requirements, to
the extent that they can be referred to as Protection Programmable Logic
Controllers (PPLC) (Leschert et al., 2021). Improving the operating
characteristics of overcurrent relays by using voltage quantities is a
commonly adopted method because microprocessor-based protective
relays can simultaneously measure current values from CTs and voltage
values from PTs in the electrical system without the need for additional
equipment. In Saleh et al. (2015), a time-current-voltage characteristic
of the overcurrent relay was proposed to reduce the tripping time of
protective relays and to address the coordination failure between pri­
mary and backup relays in protection systems by improving the tradi­
tional inverse time characteristic using an exponential function of the
voltage magnitude. To solve the coordination issues of the proposed
relay, the researchers determined the optimal setting parameters using
the Sequential Quadratic Programming (SQP) method and applied this
specific relay characteristic to an overcurrent protection scheme for
distribution systems with the penetration of DG, considering both
synchronous-based DG (SBDG) and inverter-based DG (IBDG) sources.
In Keil and Jager (2008), a non-standard overcurrent characteristic was
proposed to improve relay coordination in electrical distribution sys­
tems. This relay characteristic was modified from the inverse time
characteristic of IEEE C37.112- 1996 using a logarithmic function of the
short-circuit voltage along with parameters A, B, and C. These param­
eters were optimized using a Genetic Algorithm (GA) to address the
coordination issues of the proposed relay. The researchers applied this
relay characteristic to an overcurrent protection scheme for distribution
systems with the penetration of SBDG. Both research works selected
synchronous-based DGs to highlight the impact of fault currents on relay
Fig. 1. The problem statement diagram of the relay coordination problem.
coordination. Although the relay characteristics proposed in these two
studies demonstrated faster operation of the primary relays and a
reduction in the total operating time of all protection relays in the sys­ radial distribution systems equipped with SBDG and DFIG-based WTDG
tem, they did not present the coordination time interval (CTI) between to evaluate the fault current sensitivity due to the DG penetration.
the primary and backup relays, which is essential for thoroughly Optimal coordination for the proposed non-directional overcurrent
addressing relay coordination issues. Additionally, they did not show the protection scheme is achieved using Interval Linear Programming (ILP)
performance of the protection system under dynamic system changes optimization combined with Nonlinear Programming (NLP) optimiza­
during the fault occurences. Moreover, both studies set a condition that tion, by setting inequality constraints on the Coordination Time Interval
the CTI should not be less than 0.2 seconds, which resulted in CTI values (CTI) in an interval form, ensuring that the CTI between the primary and
between relay pairs exceeding 0.2 seconds and even surpassing 0.5 sec­ backup relays remains within the range of 0.2–0.5 seconds. Further­
onds. This led to the delayed operation of the backup relays due to more, the proposed protection scheme results are compared with the
increased tripping times. traditional protection scheme and the time-current-voltage protection
For this reason, this research proposes a new operating characteristic schemes proposed in references Saleh et al. (2015) and Keil and Jager
of a voltage-controlled overcurrent relay (VCOCR), inspired by the (2008). Thus, the highlighted contributions of this paper are as follows:
electromechanical induction disk-type overcurrent relay characteristic.
The performance of the relay characteristic is enhanced by introducing a 1. We propose a new operating characteristic for a voltage-controlled
voltage multiplier and D parameter, enabling the proposed relay to overcurrent relay, inspired by the electromechanical induction
reduce tripping time while maintaining coordination between the pri­ disk-type overcurrent relay characteristic.
mary and backup relays and CTI within an acceptable value in the 2. The proposed method considers the inequality constraints on the CTI
overcurrent protection scheme for distribution systems, both with and value in interval form to resolve the optimal coordination failures
without the presence of DG. To test and verify the effectiveness of the between primary and backup relays.
proposed relay characteristic, the proposed method has been applied to 3. The performance results of the proposed protection scheme in a
the overcurrent protection schemes of the IEEE 13-bus and IEEE 33-bus distribution system are verified under both steady-state and dynamic

5932
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 2. The LVRT requirement of FERC Order No.661 (Zavadil et al., 2005).

conditions, considering dynamic changes in power output from DGs,

which demonstrate the effective operation of the proposed relay
characteristics and fault current sensitivity, thereby satisfying a grid
code requirement.

2. Relay coordination problem and solution

The second section presents the issue of coordination failures in

traditional relays and proposed solutions. The increase in short circuit
currents caused by the installation of DG in power systems, which is the Fig. 3. The radial distribution system with DG.
root cause of the problem, is discussed in Subsection 1. The issue of
coordination failures in overcurrent relays in power systems with DG et al., 2016). During a fault in the distribution system, the control system
installations is addressed in Subsection 2. The improvement of voltage- of the rotor-side converter is used to control the power output of the
controlled overcurrent relay characteristics and the optimization using DFIG and the rotor’s short-circuit current, and the stator short-circuit
ILP, which are the methods employed in this article, are presented in current is limited by the stator impedance. As a result, the short cir­
Subsections 3 and 4, respectively. All sections are interconnected with cuit current of DFIG depends on the control strategy of the rotor-side
the proposed overcurrent relay characteristics and the procedures for converter and the stator impedance (Jun, 2021). Furthermore, the
determining optimal settings, as illustrated in the problem statement DFIG is designed to ride through voltage dips caused by the faults in the
diagram in Fig. 1. distribution system. The rotor-side converter control system maintains
the generator’s output voltage and frequency within acceptable values
of the grid code’s low-voltage ride-through (LVRT) requirement
2.1. SBDG and DFIG-based wind turbine in DS (Monika et al., 2022). The example of the LVRT requirement of FERC
Order No.661 is shown in Fig. 2 (Zavadil et al., 2005).
Distributed generation (DG) improves the distribution system’s ef­ Installing SBDG or DFIG in a radial power distribution system can
ficiency, reliability, and resilience. The SBDG refers to an electrical increase the short-circuit current in the system. This can be explained by
generator typically used in a distributed energy system to generate performing a short-circuit analysis of the simple radial distribution
electricity. A prime mover, such as a diesel engine, gas turbine, or steam system, as shown in Fig. 3. When a three-phase short circuit occurs at
turbine, normally drives the SBDG. It is often preferred for the distri­ point F before the installation of the DG, Thevenin’s equivalent circuit of
bution system because of its high efficiency and ability to respond the radial distribution system is shown in Fig. 4(a), and the short circuit
quickly to changes in load. However, it is less efficient than other DG current IF is expressed as:
technologies, such as inverter-based distributed generators. In addition, [ ] [ ][ ]
the SBDG is unsuitable for applications requiring variable output power, V1 Z11 Z12 0
= (1)
such as renewable energy systems. During a short circuit in a distribu­ V2 Z21 Z22 + ZF IF
tion system, the SBDG acts as a source of electrical energy, and the
magnitude of short circuit current contribution by SBDG depends on its V2
IF = (2)
sub-transient reactance and the voltage at the generator terminals Z22 + ZF
(Calderaro et al., 2009). Wind power is an alternative energy source for
Where V2 is a Thevenin voltage at bus2, Z22 is the Thevenin impedance
electricity generation to reduce the environmental impact of electricity
at the fault location, and ZF is the fault impedance at point F. After the
generation associated with the carbon neutrality policy. The DFIG-based
DG installation at bus 2, the three-phase fault equivalent circuit presents
WTDG is widely used in wind turbine generation systems because it can
as Fig. 4(b) and the short circuit current IF is described by the following.
operate at variable speeds, which allows it to generate power smoothly
derived from the wind. Moreover, using power electronic converters, Z21 Z12
Z22(new) = Z22 − (3)
consisting of the rotor-side and the grid-side converter, can enhance the Z11 + ZDG
power output control capability, which can maintain the stability and
reliability of the distribution system (Omar and Moayed, 2024; Ghulam

5933
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 4. Thevenin equivalent circuit of the radial distribution system (a) without DG (b) with DG.

Fig. 5. The time coordination curve for the OCR coordination.

[ ] [ ][ ]
V1 Z11(new) Z12(new) 0 enhance the relay characteristics by incorporating voltage magnitude.
= (4)
V2 Z21(new) Z22(new) + ZF IF The radial system is as in Fig. 3; the OCRs R1 and R2, used for feeder
protection, operate based on the IEC255–3 characteristic equations as
V2 written by Eq. (6) (IEEE, 2019).
IF = (5)
Z22(new) + ZF
A
tij = TMSi × ( )B (6)
Where Z22(new) is the Thevenin impedance at the fault location after Ifij
the DG installation as Eq.(3). As shown in Eq. (5), adding ZDG in the IPi
− 1
equivalent circuit causes a decrease in the overall impedance, which
increases the short circuit current IF at point F. Therefore, it may affect Where tij is the tripping time of ith relay for jth fault location, TMSi is the
the coordination and operation of relays R1 and R2, as shown in Fig. 3. time multiplier setting of ith relay, Ifj is the short circuit current of ith
For example, suppose the fault current exceeds the optimal setting relay for jth fault location, IPi is the pickup current of ith relay, A and B are
values of relays R1 and R2. In that case, the relays may cause circuit characteristic parameters. Suppose that a short circuit occurs on feeder 2
breakers CB1 and CB2 to trip simultaneously unnecessarily, leading to at point F while the radial system is not equipped with DG. This case is
unintentional interruptions and undesired downtime. considered the coordination of relay R1 and relay R2. The coordination
Additionally, an increase in fault current can impact the duration and curve of relays R1 and R2 are shown in Fig. 5
severity of voltage dips, affecting the safe and reliable operation of the The primary relay R2 operates depending on the multiple of current
DG system. Protective devices are designed to quickly isolate faults and value M2 and can trip within operating time t2. The operation point of
limit the duration and severity of voltage dips. However, if the fault relay R2 is shown as point A on the operating curve of relay R2. In case of
current reaches or exceeds the interrupting current rating of the circuit the misoperation of relay R2, the backup relay R1 is operated by the
breaker CB1 and CB2, these devices may take longer to operate or may multiple of current M1 and trips within operating time t1. The operating
fail to operate entirely. This could result in prolonged voltage dips and point of relay R1 is shown as point B on the curve of relay R1. The CTI
potential damage to the DG system and other equipment. Therefore, between relays R1 and R2 is between points A and B, where the coor­
relays R1 and R2 should be designed to quickly detect faults and mini­ dination time ranges from 0.2 to 0.5 sec. After DG is installed at bus 1,
mize the duration and severity of voltage dips to prevent disconnection the sum of the short-circuit currents of DG and the substation can in­
of the DG. crease the short-circuit current at point F. Hence, the multiple of current
of relay R2 changes from M2 to M2′ point C. The multiple of the current
2.2. OCR coordination problem in the radial system of relay R2 varies from M2 to M2′. As a result, the operating point of relay
R2 is shifted from point A to point C and the operating point of relay R1
The level of short circuit current significantly impacts the over­ is shifted from point B to point D, the CTI of relay R1 and R2 is the time
current protection of feeder lines after the installation of DG in a radial interval between t1′ and t2′ which reduced to less than 0.2 s. If the CTI
distribution system. This can result in a lack of coordination between the values of the relays R1 and R2 are less than 0.2, then the R1 cannot be
primary and backup overcurrent relays. To address the coordination coordinated with R2, which is the effect of higher short-circuit currents
issues that arise from DG installations, a thorough study and analysis of resulting from the DG installation in the distribution system (Meskin
their effects on OCR operation have been conducted. This subsection et al., 2020).
addresses the relay coordination problem and presents a solution to

5934
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Since the operation characteristic curves of relays R1 and R2 are

influenced by their respective TMS values, the solution to maintaining 2.4. ILP optimization for the relay coordination problem
coordination between them is to optimize the TMS values. The optimi­
zation ensures that the desired coordination time is maintained within This subsection presents the overview of ILP optimization for the
0.2–0.5 seconds. The optimization of the TMS values for both relays relay coordination problem. The ILP method is more effective for solving
results in the shifting of the curves of relay R1 and relay R2 to green and the relay coordination problem than LP because ILP significantly re­
blue dotted lines, respectively. This adjustment ensures that the coor­ duces the number of coordination constraints compared to LP. This
dination time interval of both relays remains within the desired range of reduction allows ILP to solve the relay coordination problem in fixed
0.2–0.5 seconds. However, an unintentional consequence of improving network topologies with superior performance. Additionally, ILP results
the TMS values is a reduction in the operating sensitivity of relays R1 can be converted to LP results, providing flexibility in utilizing ILP
and R2 (Meskin et al., 2020). It can affect the fault isolation capability outcomes for LP-based analyses. In Noghabi et al. (2010), the ILP
within the relay R1 and R2 zones, especially when dealing with high problem’s general form is Eqs. (9) and (10)
short-circuit current values.
Minimize f(x) = (CI )T X (9)

2.3. The time-current voltage characteristics for OCR Subjectto : AI X ≤ bI (10)

I I I
Where A , b are the interval matrix, C is the coefficient matrix and X
The improvement of the relay’s sensitivity by using the voltage
is the uncertain parameter. The result of ILP is converted to the result of
magnitude was proposed in (Saleh et al., 2015). The proposed method
standard LP; thus, the ILP form can be written by the form of standard LP
improved the tripping time by multiplying the voltage exponential
as Eqs. (11)-(13):
function with the IEC standard equation as written by
T
( )ki Minimize f(x) = (C orCT )X (11)
1 TMSi ⋅A
tij = 1− Vfij × ( )B (7)
e Ifij
− 1 Subject to AX1 − AX2 ≤ b (12)
IPi

X1 ≥ 0, X2 ≥ 0 (13)
where the first term is the exponential function of the voltage magni­
tude, which consists of the short-circuit voltage magnitude of the ith Where Aand Aare the bound of matrix A. CandC are the bound of matrix
relay at jth fault location (Vfij) and the constant parameter ki. The second C. X1 and X2 are the uncertain parameters. In the rest of this paper, the
term is the standard IEC characteristic equation. As described in Eq. (7), coordination problem of OCRs is defined as an ILP problem and solved
the tij depends on the value of TMSi and the exponential function of the by utilizing the theorems demonstrated earlier.
fault voltage magnitude. Therefore, the characteristic curve of Eq. (7) is
a relationship of three parameters, including the operating time of relay
3. Proposed overcurrent characteristic and coordination
(t),the pickup current multiplier (If/Ip) , and the short-circuit voltage
problem formulation
magnitude (Vf). The during-fault voltage magnitude in the protection
zone of the relay that is being reduced to a very low level allows the
3.1. Proposed OCR characteristic
exponential behavior in Eq. (6) to act as a flexible TMS value, which is
sensitive to the during-fault voltage magnitude. Consequently, the relay
The proposed relay characteristic equation has been improved based
operation with the time-current-voltage characteristic becomes more
on the fundamental equation of the electromechanical disc relay. In
sensitive than conventional overcurrent relays. This enhanced sensi­
(IEEE, 2019), the rotating induction disc equation of the current-voltage
tivity enables effective fault isolation within the protection zone of the
relay can be express as:
relay. By controlling the operating time of the relay based on the fault
voltage magnitude, this relay characteristic can reduce the misoperation dθ
K1 VI = Kd + τs (14)
problem of the protection relay in systems with DG installation. How­ dt
ever, as the sizing or penetration of DG increases in the distribution Where I is the current from a current transformer, V is the voltage
system, then the short-circuit current also increases. This ongoing in­ from a voltage transformer, θ is the angle of disc rotation, τs is the spring
crease in short-circuit current impacts the interaction of time-current- torque, K1 is a torque conversion constant and Kd is a damping torque
voltage relays, necessitating a solution that involves optimizing both constant. The K1VI is the driven torque of current-voltage magnitude
the TMS and the K parameter of the exponential function. By adjusting (Mason, 2011). The driven torque is related to the multiples of pickup
these values, the operating time of the relay can be aligned with the current and the multiples of pickup voltage. Eq. (13) can be solved by
increasing short-circuit current, ensuring adequate protection and co­ referring to conventional standard characteristic equations and
ordination within the system. substituting V––VpMV and I– –IpMI. Consequently, the solution for Eq.
The overcurrent relay characteristic model in (Keil and Jager, 2008)
(15) is derived as follows(Nattapol and Krischonme, 2020):
is another example of the development of standard overcurrent relay
characteristics into a non-standard form for use in distribution systems A
t= × TMS (15)
connected to Distributed Generation (DG). The researchers proposed a MDV MBI − 1
non-standard overcurrent relay equation, as shown in Eq.7. This equa­
Where MV is the multiples of pickup voltage, D is the exponential
tion retains the inverse operating characteristic of the traditional form
parameter, MI is the multiples of pickup current, A and B are the pa­
but adds a logarithmic function of the short-circuit voltage (Vfij) to help
rameters depend on the conventional standard characteristic and TMS is
reduce the relay’s tripping time (tij). Additionally, the researcher can
the time multiplier setting. The implementation of proposed relay
optimize the parameters A, B, and C to ensure proper coordination be­
characteristic on the over current relay at ith bus can be explained as
tween the primary and backup relays.
follows: the MI depends on the fault current at jth location and the pickup
( )
log Vfij + Ai current of ith relay then the MI = Ifij/Ipi and the MV relies on the normal
tij = ( )Bi + Ci (8)
Ifij
voltage magnitude of ith bus (Vnom_i), the fault voltage magnitude of ith
IPi
− 1 bus at jth location (Vfij) and the pickup voltage of ith relay (VPi) then the
MV= (Vnom_i-Vfij)/(Vnom_i-Vpi). The proposed relay characteristic equation

5935
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

of ith relay at jth fault location is expressed as:

A
tij = TMSi × ( )Di ( )B (16)
Vnom i − Vfji Ifij
Vnom i − VPi IPi
− 1

Fig. 6 illustrates the concept of the pickup voltage multiplier. When a

short circuit happens within the zone of the ith relay near the bus where
the WTDG is installed, the voltage at the ith relay installation point varies
according to the orange curve depicted in Fig. 6. Comparing Fig. 6 with
Fig. 2, the normal voltage magnitude is Vnom_i, the restoration voltage
magnitude is VPi, and the fault voltage magnitude is Vfi. For this result,
the value of (Vnom_i - Vfji) is greater than the value of (Vnom_i - VPi), and the
ratio (Vnom_i - Vfji) / (Vnom_i - VPi) is greater than 1.0 pu. Fig. 7 illustrates
the relationship between VPi and the pickup voltage multiplier (MV),
where Vfji is set to zero.
The Di parameter in Eq. (16) is the index value used to control the
pickup voltage multiplier of the ith relay. Fig. 8 shows the relationship
between the D parameter and the pickup voltage multiplier value when
Fig. 6. The concept of pickup voltage multiplier value.
A= 0.14, B= 0.02, Vnom_i= 1.0 pu., VPi = 0.95pu., Vfij= 0pu., TMS= 1,
and D ranging from 0 to 1. The pickup voltage multiplier has the greatest
effect on relay operating time when D is set to 1 and the least effect when
D is set to 0. For this reason, parameter D must be set properly to control
the relay’s operating characteristics.

3.2. Formulation of coordination problem

The application of the proposed characteristic in Eq. (16) for the

feeder protection system of the distribution system shown in Fig. 9 can
be explained as follows:
Suppose relay R1 is responsible for protecting against short circuits
at position 1, using the proposed characteristics in Eq. (16). The oper­
ating time of the primary protection (tPij) is
tP11 = CP11 ⋅TMS1 (17)

Fig. 7. The relationship between the pickup voltage and the pickup Where the subscript P is the primary protection mode. The TMS1 is the
voltage multiplier. time multiplier setting of R1, and the coefficient CP11 is
A1
CP11 = ( )D1 ( )B (18)
Vnom 1 − Vf11 If11
Vnom 1 − Vp1 Ip1
− 1

Vf11 is the voltage at bus 1 during a short circuit at position 1, which

relay R1 detects through the PT (Potential Transformer), and If11 is the

Fig. 8. The relationship between the D parameter and the pickup voltage multiplier value.

5936
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 9. The example of the feeder protection system of the distribution system.

short-circuit current at position 1, which relay R1 detects through CT

(Current Transformer) installed on bus 1. Additionally, relay R1 func­
tions as backup protection for relay R2 to protect against short circuits at
position 2. Therefore, the operating time equation for the backup pro­
tection of relay R1 is
tB122 = CB122 ⋅TMS1 (19)

Where the subscript B is the backup protection and the coefficient CB122
is
A1
CB122 = ( )D1 ( )B (20)
Vnom 1 − Vf 122 If122
Vnom 1 − Vp1 Ip1
− 1

Vf122 is the voltage at bus 1 during a short circuit at position 2 in the

zone of relay R2, which relay R1 detects via the PT (Potential Trans­
former), If122 is the short-circuit current at position 2 in the zone of relay
R2, which relay R1 detects via CT (Current Transformer) installed on bus
1. For the example distribution system shown in Fig. 9, if the system
includes the relay Ri and relay Rh at the end system, the operating time
of the relay Ri to protect against a short circuit at position j and the
operating time of relay Ri as backup protection for relay Rh to protect
against a short circuit at position k is given by:
tPij = CPij ⋅TMSi (21)

tBihk = CBihk ⋅TMSi (22)

The operating time of the relay Rh to protect against a short circuit at
position k is
tPhk = CPhk ⋅TMSh (23) Fig. 10. The flow chart of determine D parameter.

Finding the optimal TMS values for relay R1 through relay Rh to is set to the maximum load current of bus 1. The pickup voltage Vp1 can
achieve the minimum total operating time is the objective in solving the be set as the nominal voltage specified in the grid code for the small
coordination problem for all relays in the protection system. Therefore, distribution system. However, if the distribution system is more com­
the objective function in Eq. (11) can be applied to find the minimum plex, the value of Vp1 should be determined using a method to solve the
total operating time as follows: nonlinear equation. The subsequent relays up to relay Rh in the system
⎡ ⎤T ⎡ ⎤ use this same configuration for calculating the coefficients for both
CP11 + CB122 TMS1
⎢ ⋮ ⎥ ⎢ ⎥ primary and backup protection functions for each relay. In Eq. (18) and
Minimize T = ⎢ ⎥ ⎢ ⋮ ⎥ (24)
⎣ CPij + CBihk ⎦ ⎣ TMSi ⎦ Eq.(20), the relationship between D1 and coefficients of R1 (CP11, CB122)
CPhk TMSh is nonlinear. Therefore, the Newton-Raphson method is employed to
determine the feasible value of D1 as Eq. (29), and the flow chart for
Additionally, the constraints Eq.(12) and Eq.(13) can be used to finding the D parameter is illustrated in Fig. 10. All relays in the system
define the constraint equations for the problem as follows: use the same method for setting the D parameter as relay R1.
CBihk TMSi − CPhk TMSh ≥ CTImin (25)
f(Dki )
Dk+1
i = Dki − (29)
CBihk TMSi − CPhk TMSh ≤ CTImax (26) fʹ(Dki )

Where the Dk+1i is the estimated value obtained from the initial es­
TMSmin ≤ TMSi ≤ TMSmax (27) timate Dki , finding the optimal TMS values for each relay to solve the
coordination problem within the protection system and minimize the
TMSmin ≤ TMSh ≤ TMSmax (28)
total operating time of all relays can be represented as a flow chart, as
The values of CTImin and CTImax in Eq. (25) and Eq.(26) are set to 0.2 shown in Fig. 11, with each relay operating based on the presented
and 0.5 sec.(Hong et al., 2021), respectively. The values of TMSmin and equations.
TMSmax in Eq. (27) and Eq.(28) are set to 0.05 and 3 (Hong et al., 2021),
respectively. To calculate the coefficient of R1 (CP11, CB122) using Eq. 4. System and simulation setups
(18) and Eq.(20), in addition to the short-circuit current and voltage at
position 1 (If11, Vf11) and the short-circuit current and voltage at position The power distribution system protection simulation using the pro­
2 (If122, Vf122), the value of Vnom_1 is set equal to the pre-fault voltage, Ip1 posed protective relay is divided into two systems: the simulation of

5937
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Table 2
Sizes and Types of DGs connected at Bus 2 and Bus 7 for each scenario.
Scenario Bus 2 Bus 7

1 600 kVA SBDG 600 kVA SBDG

2 1000 kVA SBDG 1000 kVA SBDG
3 600 kVA SBDG 1000 kVA WTDG
4 1000 kVA SBDG 1000 kVA WTDG

4.1. 4.1 IEEE13bus system setup

The proposed OCR has been set up on the feeder protection scheme
in a modified IEEE 13 bus radial system, as illustrated in Fig. 12. The
tested system consists of ten feeders. Each feeder is assumed to be a
balanced three-phase feeder, and the feeder impedance values are
derived from (IEEE Distribution System Analysis Subcommittee). The
substation at bus 1 is a 115/4.16 kV, 5000 kVA substation. The DGs in
this simulation consist of a 600 kVA SBDG, 1000 kVA SBDG, and
1000 kVA WTDG. The SBDG supplies power to the system via a
480 V/4.16 kV step-up transformer with a transient reactance of 10 %.
The WTDG is the DFIG wind turbine, which delivers the power to the
distribution system through a 575 V/4.16 kV step-up transformer. Each
DG operates at a unity power factor. The short-circuit locations are
defined as nodes on each feeder, designated as F1-F10. Each node rep­
resents a three-phase fault simulated at five different locations: 0 %,
25 %, 50 %, 75 %, and 100 % of the feeder distances, and the fault
impedance is assumed to be zero to evaluate the proposed method under
the bolt fault incidence.
The proposed overcurrent protection scheme is applied to ten over­
current relays (R1-R10) and demonstrates the performance of the pro­
posed protection scheme in comparison with the protection scheme
Fig. 11. The flow chart of the optimal TMS values for each relay. proposed in (Saleh et al., 2015) and the conventional protection scheme.
Thus, the overcurrent protection schemes are divided into four schemes,
including traditional overcurrent protection schemes based on the relay
characteristic as Eq. (6) (scheme-A), an overcurrent protection scheme
based on the relay characteristic of Eq. (7) (scheme-B), overcurrent
protection scheme based on the relay characteristic as Eq.(8) (sche­
me-C). The proposed overcurrent protection scheme is based on the
relay characteristic as Eq. (16) (scheme-D). The IPi for the four schemes
is assumed to be a fixed value based on the maximum load demand. For
scheme-D, Vnom_i is defined as 1.0 per unit, and VPi is 0.9 per unit ac­
cording to the FERC ORDER No.661 grid code requirement (Zavadil
et al., 2005). For the nonlinear variables as ki of scheme-B, Ai, Bi, Ci of
scheme-C, and Di of scheme-D, the Newton-Raphson method is used to
determine the feasible values of these variables. The TMSi values for
scheme-A, scheme-B, and scheme-D are obtained by solving the optimal
coordination problem as the flow chart in Fig. 11.
Each protection scheme demonstrates its effectiveness under
different cases; the testing cases are divided into 4 cases as follows:
Case-1 demonstrates the simulation results of a three-phase fault at
each location in the system without DGs.
Case-2 shows the simulation results of a three-phase fault at each
location in the system with DGs connected at bus 2. The tested DGs
include a 600 kVA SBDG, 1000 kVA SBDG, and 1000 kVA WTDG.
Fig. 12. IEEE 13 bus radial system under study.
Case-3 involves simulating a three-phase fault at each location in the
system with DGs connected at Bus 7. The tested DGs are the same as
feeder protection on the IEEE 13-bus system and the feeder protection
those in Case2.
on the IEEE 33-bus system. Both simulations demonstrate the proposed
Case-4 simulates a three-phase fault at each location in the system by
protective relay’s application in protection systems for small-scale
connecting DGs at Bus 2 and Bus 7. The installation of the DG at both
power distribution systems with slight variations in short-circuit cur­
buses involves adjusting the size and type of DG as shown in Table 2:
rent levels and large-scale power distribution systems with significant
The results of scheme-D are compared with those of scheme-A,
variations in short-circuit current levels. The results of these simulations
scheme-B and scheme-C in all cases to demonstrate the superior per­
illustrate the proposed overcurrent protective relay’s performance
formance of scheme-D in comparison with scheme-A, scheme-B and
efficiency.
scheme-C.

5938
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 13. IEEE 33 bus radial system under study.

Table 3
Sizes and Types of DGs connected in the system for each scenario.
Scenario DG Type Location

1 SBDG 2 MW Bus 13
WTDG 6 MW Bus 14
2 SBDG 2 MW Bus 13
WTDG 9 MW Bus 14

4.2. IEEE 33bus system setup

The IEEE 33-bus system is a radial distribution system with a voltage

rating of 12.66 kV. The total power demand of the loads in the system is
5.686 MVA. The IEEE 33bus test system details are described in (Baran
and Wu, 1989). In this simulation, 32 non-directional overcurrent relays
have been installed to protect all feeders in the system, as shown in
Fig. 13. Each relay protects the upstream section by detecting
short-circuit currents within its protection zone. Additionally, the relays
serve as backup protection for the next relay in sequence. For example,
as illustrated in Fig. 13, relay R16 acts as the primary protection against
short circuits at the middle point on feeder L17–18. It also functions as a
backup to support the operation of relay R17.
The feeder protection system tested is divided into schemes: Scheme-
A, Scheme-B, Scheme-C, and Scheme-D, similar to the protection system
testing of the IEEE13-bus system. The pickup current of all relays in each
scheme is set to the maximum load current of each bus. Since the IEEE
33-bus system exhibits significant variations in short-circuit current
levels and voltages across different buses, it is necessary to determine
the pickup voltage (Vp) and parameter D for Scheme-D using the graph
in Fig. 7 and Fig. 8 for the initial value setting and calculate the value
appropriate values through the Newton-Raphson method, as illustrated
in the flow chart in Fig. 14.
The TMS (Time Multiplier Setting) parameters for Scheme-A,
Scheme-B, and Scheme-D are optimized using a linear optimization
method, as outlined in Section 3.2. The TMS range is set between 0.05
and 3, as specified in references (Hong et al., 2021). The relay param­
eters A, B, and C for Scheme-C are also determined using the
Newton-Raphson method.
The installation of DG in the IEEE 33-bus system to test the operation
of protection relays involves deploying a DG unit of the SBDG type and
WTDG type. The SBDG supplies power to the system via a 480 V/
12.66 kV step-up transformer with a transient reactance of 10 %. The
WTDG is the DFIG wind turbine, which delivers the power to the dis­
tribution system through a 575 V/12.66 kV step-up transformer. Each
DG operates at a unity power factor. The power rating and installation
location of the DG are specified in Table 3. In each scenario, simulations
Fig. 14. The flow chart of determine the pickup voltage (Vp) and parameter D. are performed for the mid-point short-circuit locations on feeder
L17–18. These simulations evaluate the coordination between the pri­
mary and backup relay pairs in protecting against feeder short circuits.

5939
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Table 4
Optimal setting values of each scheme.
Relay Ip(p.u.) Scheme A Scheme B Scheme C Scheme D

TMS K TMS A B C D TMS

R1 0.00342 0.228 0.463 0.264 1.008 0.134 0.024 0.061 0.593

R2 0.0005 0.158 0.387 0.200 1.25 0.945 0.190 0.022 0.241
R3 0.00045 0.170 0.455 0.227 1.035 0.091 0.110 0.023 0.251
R4 0.00025 0.056 0.223 0.052 1.040 1.000 0.048 0.017 0.050
R5 0.0025 0.154 0.386 0.188 1.067 0.428 0.495 0.041 0.420
R6 0.00082 0.050 0.640 0.055 1.217 0.997 0.186 0.029 0.062
R7 0.00033 0.172 0.416 0.233 1.250 0.396 0.342 0.020 0.240
R8 0.00018 0.056 0.202 0.052 1.057 1.000 0.194 0.016 0.050
R9 0.00015 0.058 0.207 0.054 1.234 0.996 0.0968 0.016 0.050
R10 0.00478 0.050 1.662 0.070 1.114 1.000 0.0972 0.046 0.070
Computation time (sec.) 0.270 0.361 0.574 0.437

The IEEE-13 and IEEE-33 buses with the relay models were verified
on DIgSILENT PowerFactory software to simulate the steady-state and
the dynamic conditions. The optimal relay’s parameters using the in­
terval linear programming approach were calculated by MATLAB soft­
ware. The simulations were carried out on a personal computer with the
following specifications: an HP VICTUS Laptop with an Intel(R) Core
(TM) i5–11400CPU running at 2.70 GHz (12CPUs), 16.0 GB of RAM and
a 64-bit operating system. The performance evaluation of relays pre­
sented in Scheme-D for the IEEE 33-bus and IEEE 13-bus distribution
systems is based on the total operation time for short-circuit protection
and the coordination time between primary and backup relay pairs. This
evaluation is conducted for both without DG and with DG installation.
The performance of the proposed relay scheme (Scheme-D) is compared
with the results of Scheme-A, Scheme-B, and Scheme-C. The simulation
Fig. 15. The convergence curve between the relative error and iteration
results and comparative analysis are presented in the following section. number of the optimal setting value determination.

5. Simulation result and discussion

Table 5
This section presents the test results of the proposed overcurrent Optimal primary and backup relay operating time of scheme-A.
relay characteristics, which were tested on the IEEE 13-bus and IEEE 33- Feeder F.L. Operating time in sec.
bus power distribution systems. The proposed characteristics’ test re­
Rp Rb1 Rb2 Rb3
sults are compared with those of three overcurrent relay characteristics
from the literature review: Scheme-A, Scheme-B, and Scheme-C. The L1–2 F1 R1:0.597 - - -
L2–3 F2 R2:0.259 R1:0.719 - -
first subsection presents the test results on the IEEE 13-bus system, while
L2–5 F3 R3:0.272 R1:0.709 - -
the second subsection presents the results on the IEEE 33-bus system. L5–6 F4 R4:0.080 R3:0.283 R1:0.761 -
L2–7 F5 R5:0.438 R1:0.751 - -
5.1. Simulation result of IEEE 13bus L8–9 F6 R6:0.104 R5:0.492 R1:0.858 -
L7–10 F7 R7:0.277 R5:0.491 R1:0.856 -
L10–11 F8 R8:0.080 R7:0.282 R5:0.510 R1:0.896
This section presents the proposed overcurrent protection scheme L10–12 F9 R9:0.081 R7:0.285 R5:0.520 R1:0.917
(Scheme-D) simulation results compared to three schemes in the liter­ L7–13 F10 R10:0.233 R5:0.497 R1:0.868 ​
ature (Scheme-A, Scheme-B, Scheme-C, and Scheme-D) on the IEEE
13bus test system. The analysis of four testing cases in the previous
section is categorized into four subsections to evaluate the performance 5.1.1. Simulation results of the relays’ performance in IEEE 13 bus system
comparison of the schemes under various operated conditions. Firstly, to without DG
verify the protective relay’s schemes for the base-case operation sce­ This subsection presents the results of a base case scenario (case-1).
narios without DGs, the results are demonstrated in subsection A, which Case-1 is defined as a protective coordination problem determining the
shows that the primary and backup protections are considered during optimal setting value of three protection schemes. The optimal setting
the different fault locations. Subsection B shows the simulation results of values of ten relays for three schemes are shown in Table 4. The last row
the protection schemes with DG integrations under different scheduled of Table 4 shows the computation time values for determining the
powers and locations. The simulation results of the coordination per­ optimal setting values of each scheme. The computation time for setting
formance between the primary and backup relay of each protection adjustments in Scheme D is comparable to that of Scheme B, as both
scheme with the penetrations of the DGs are compared in subsection C. have the same number of parameters (two). However, the computation
Finally, the fault clearing time of the relays in the radial system con­ time for Scheme D is longer than for Scheme A since Scheme A has only
nected with WTDG considering the LVRT time is present in subsection D. one parameter. Conversely, the computation time for Scheme C exceeds
All simulation results show the advantages of scheme-D, which is better that of Scheme D because Scheme C has three parameters. The results of
than scheme-A, scheme-B and scheme-C with various DG scheduled the computation time using the proposed method indicate that an
powers and locations. increasing number of parameters leads to longer computation times.
The convergence graph in Fig. 15 illustrates an example of the
relative error converging to zero in determining the optimal settings for
relays R1, R3, R5, R7, and R9 in Scheme D. This demonstrates that the

5940
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Table 6 Table 10
Optimal primary and backup relay operating time of scheme-B. The total relay operating times considering the fault on each feeder.
Feeder F.L. Operating time in sec. Feeder Operating time in sec.

Rp Rb1 Rb2 Rb3 Scheme A Scheme B Scheme C Scheme D

L1–2 F1 R1:0.500 - - - L1–2 3.028 2.557 1.422 2.236

L2–3 F2 R2:0.226 R1:0.664 - - L2–3 4.915 4.487 3.691 3.865
L2–5 F3 R3:0.241 R1:0.662 - - L2–5 5.010 4.614 3.985 3.935
L5–6 F4 R4:0.061 R3:0.268 R1:0.742 - L5–6 5.725 5.447 5.463 4.561
L2–7 F5 R5:0.386 R1:0.710 - - L2–7 6.019 5.569 5.826 4.865
L8–9 F6 R6:0.062 R5:0.465 R1:0.855 - L8–9 7.294 6.953 8.022 5.942
L7–10 F7 R7:0.250 R5:0.464 R1:0.853 - L7–13 8.040 7.098 8.887 6.047
L10–11 F8 R8:0.060 R7:0.266 R5:0.493 R1:0.907 L7–10 8.147 7.880 8.222 6.906
L10–12 F9 R9:0.061 R7:0.275 R5:0.510 R1:0.937 L10–11 8.863 8.669 9.046 7.541
L7–13 F10 R10:0.067 R5:0.473 R1:0.868 - L10–12 9.084 9.028 10.715 7.875
Total 66.126 62.302 65.279 53.772

Table 7
Optimal primary and backup relay operating time of schemec-C. Table 11
The total relay operating times for the case-2.
Feeder F.L. Operating time in sec.
Detail of DG Total of Operating Time (sec)
Rp Rb1 Rb2 Rb3
Scheme A Scheme B Scheme C Scheme D
L1–2 F1 R1:0.278 - - -
L2–3 F2 R2:0.193 R1:0.537 - - 600 kVA SBDG at Bus 2 62.906 61.241 55.074 54.213
L2–5 F3 R3:0.227 R1:0.551 - - 1000 kVA SBDG at Bus 2 63.017 61.300 55.240 54.259
L5–6 F4 R4:0.048 R3:0.353 R1:0.679 - 1000 kVA WTDG at Bus 2 62.854 60.771 54.487 53.520
L2–7 F5 R5:0.542 R1:0.603 - -
L8–9 F6 R6:0.190 R5:0.593 R1:0.813 -
L7–10 F7 R7:0.367 R5:0.592 R1:0.810 - overcurrent relays based on the proposed characteristics exhibit faster
L10–11 F8 R8:0.083 R7:0.233 R5:0.576 R1:0.883
operation than the conventional overcurrent relay and the relay pro­
L10–12 F9 R9:0.186 R7:0.324 R5:0.633 R1:0.926
L7–13 F10 R10:0.206 R5:0.597 R1:0.829 - posed in (Saleh et al., 2015).
Table 10 illustrates the summary of each scheme for primary and
backup relay operating times under five fault locations on each feeder.
For example, for the five fault locations as 0 %, 25 %, 50 %, 75 %, and
Table 8
100 % of feeder length on feeder L10–12, the summation of the relay
Optimal primary and backup relay operating time of scheme-D.
operating times of primary and backup with the scheme-D is 7.875 s.
Feeder F.L. Operating time in sec.
The operating time of the primary and backup relays with the proposed
Rp Rb1 Rb2 Rb3 relay model and the operating times are less than the results obtained by
L1–2 F1 R1:0.437 - - - Scheme-A, B, and C. As a result, demonstrated in the final row of
L2–3 F2 R2:0.238 R1:0.529 - - Table 10, the total sum of the primary and backup relay operating times
L2–5 F3 R3:0.243 R1:0.536 - - of the scheme-D for all fault locations on all feeders is 53.772 seconds,
L5–6 F4 R4:0.05 R3:0.258 R1:0.600 -
which is reduced by 18.68 %, 13.70 %, and 17.62 %, in comparison
L2–7 F5 R5:0.400 R1:0.560 - -
L8–9 F6 R6:0.062 R5:0.453 R1:0.667 - with the scheme-A the scheme-B and the scheme-C, respectively, with
L7–10 F7 R7:0.243 R5:0.452 R1:0.665 - coordinated protection capability.
L10–11 F8 R8:0.050 R7:0.266 R5:0.474 R1:0.711 The simulation results of the base case scenario, as illustrated in
L10–12 F9 R9:0.048 R7:0.271 R5:0.489 R1:0.743 Tables 6, 7, and 8, reveal that the relay tripping times of Scheme-D are
L7–13 F10 R10:0.068 R5:0.457 R1:0.675
lower than those of Scheme-A, Scheme-B, and Scheme-C, resulting in a
​

reduced total tripping time for Scheme-D compared to Scheme-A

Scheme-B and Scheme-D as shown in Table 10. This demonstrates that
Table 9 scheme-D can reduce the total tripping times for the radial distribution
The operating time of Relay R9 of each scheme.
system in the base case that the tripping performance is higher than
Feeder Operating time of Relay R9 in sec. scheme-A, scheme-B, and scheme-C.
Scheme A Scheme B Scheme C Scheme D
5.1.2. Performances of the proposed Relay Model in IEEE 13 bus system
L10–12 0.081 0.061 0.186 0.048
with DGs
This subsection demonstrates the testing results of the three pro­
optimization process as the flow chart in Fig. 11 can effectively be used tection schemes, including scheme-A, scheme-B, scheme-C, and scheme-
to determine the optimal relay settings. D with DGs, called case-2. For the evaluation of the performances of the
As a result, the relay operating times for each scheme are presented proposed protection scheme, the different DG’s output powers, loca­
in Tables 5, 6, 7, and 8, corresponding to fault locations (F.L.) F1 to F10. tions, and types are considered for analyzing the results of the protection
As shown in Table 4, for a fault occurring at 50 % of the feeder length on schemes. In this case-2, the short circuit current direction of the DG in
each feeder, the coordination time of all relay pairs in each protection the feeder is determined, as well as the short circuit current direction
scheme can operate within the coordination time interval (CTI) of derived by the main substation. For this reason, the back-feeding short-
0.2–0.5 seconds. In addition, the relay tripping time of Scheme-D is circuit current in the feeder is not considered. Therefore, in the case of
lower than that of Scheme-A, Scheme-B, and Scheme-C. For example, in the radial system connecting with DG at bus2, the short circuit location
Table, when a fault occurs at location F9 on feeder L10–12, the primary at feeder L1–2 is not carried out.
relay R9 of Scheme-A operates at 0.081 s, the operating time of relay R9 Table 11 shows the total of all relay operating times of each pro­
of Scheme-B decreases to 0.061 s. In comparison, the operating time of tection scheme for the faults on nine feeders. As a result, the proposed
relay R9 of Scheme-C decreases to 0.061 s. This result demonstrates that scheme’s relay operating times are less than the total relay operating

5941
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Table 12 L10–11, and L10–12.

The summary of relay operating times for the case-3. As a result in Table 12, for the short-circuit on five feeders in the
Detail of DG Total Operating Time (sec) system connected with DG at bus 7, despite the change in DG locations
from bus 2 to bus 7, the total relay operating times of the proposed
Scheme A Scheme B Scheme C Scheme D
protection scheme is still less than the total relay operating times
SBDG 600 kVA at Bus 7 41.668 40.316 40.822 35.302 derived from the scheme-A, scheme-B, and scheme-C. As an example, for
SBDG 1000 kVA at Bus 7 41.589 40.225 40.116 35.250
WTDG 1000 kVA at Bus 7 41.449 39.979 39.141 34.906
the fault on five feeders in the system connecting with 1000 kVA SBDG
at bus 7, the total relay operating times of the proposed protection
scheme is 35.250 s, which is reduced by 15.24 % of the total operating
times compared with the scheme-A, 14.11 % when it is compared with
Table 13
the scheme-B and 12.12 % when it is compared with the scheme-C.
The summary of relay operating times for the case-4.
Moreover, for the fault on five feeders in the system connecting with
Scenario Total of Operating Time (sec) 1000 kVA WTDG installed at bus 7, the total relay operating times using
Scheme A Scheme B Scheme C Scheme D the proposed protection scheme is less than the total relay operating
1 39.747 38.920 36.882 35.180 times of scheme-A, scheme-B, and scheme-C. Therefore, the entire
2 39.641 38.800 36.727 35.165 operating times are reduced by 15.91 %, 12.85 %, and 10.82 %
3 39.333 38.166 35.816 34.355 compared to schemes A, B, and C, respectively.
4 39.298 38.044 35.598 34.145
To achieve a wide range of short-circuit current values and demon­
strate the proposed protection scheme’s performances, this test has
times of scheme-A, scheme-B, scheme-C, and scheme-D. For the short carried out the cases of simultaneously installing DG at bus 2 and bus 7
circuit on nine feeders in the system connected with 600 kVA SBDG at and included the change in DG’s rated power and technologies. For case-
bus 2, the total relay operating times of the proposed protection scheme 4, the short circuit occurs only on the five feeders, including L8–9,
is 54.213 s. The proposed method can decrease the full operating times, L7–10, L7–13, L10–11, and L10–12. The summary of the total operating
which are 13.81 %, 11.47 %, and 1.56 % compared with schemes A, B, time of relays R6, R7, R8, R9, and R10 of each scheme is demonstrated in
and C, respectively. Furthermore, for the fault on feeders in the system Table 13. As the test result in Table 13, for the fault on five feeders, the
connecting with 1000 kVA WTDG, despite the change in DG type from total operating times of seven relays using the proposed protection
SBDG to WTDG, the total relay operating times using the proposed scheme are less than the total relay operating times of scheme-A,
protection scheme remain less than the total relay operating times of scheme-B, and scheme-C. Furthermore, with the short circuit on
scheme-A and scheme-B, with 14.85 %, 11.93 %, and 1.77 % reduction, feeders in the system with two 600 kVA SBDGs at bus 2 and bus 7, the
respectively. total relay operating times of the proposed protection scheme is 35.18 s.
Due to the non-directional overcurrent protection scheme and to The entire operating times with the proposed relay model are decreased
prevent the backfeeding short circuit currents on the feeder in case-3, by 11.49 %, 9.6 %, and 4.61 %, compared with schemes A, B, and C,
the faults are assumed only on the five feeders: L8–9, L7–10, L7–13, respectively.

Table 14
The clearing time of the primary relay (R8) and backup relay (R7) for the three fault locations on feeder L10-11.
Scenarios Fault Location Tripping Time (sec)

Scheme A Scheme B Scheme C Scheme D

Rp Rb Rp Rb Rp Rb Rp Rb

Without DG connection 0% R8: R7: R8: R7: R8: R7: R8: R7:
0.079 0.281 0.064 0.264 0.097 0.327 0.048 0.261
50 % R8: R7: R8: R7: R8: R7: R8: R7:
0.079 0.282 0.065 0.270 0.097 0.334 0.050 0.265
100 % R8: R7: R8: R7: R8: R7: R8: R7:
0.080 0.285 0.067 0.280 0.097 0.327 0.050 0.270
600 kVA SBDG at Bus 2 0% R8: R7: R8: R7: R8: R7: R8: R7:
0.075 0.264 0.063 0.255 0.097 0.297 0.045 0.251
50 % R8: R7: R8: R7: R8: R7: R8: R7:
0.076 0.267 0.064 0.257 0.097 0.301 0.047 0.256
100 % R8: R7: R8: R7: R8: R7: R8: R7:
0.077 0.270 0.066 0.263 0.097 0.304 0.049 0.259
1000 kVA SBDG at Bus 2 0% R8: R7: R8: R7: R8: R7: R8: R7:
0.075 0.266 0.062 0.255 0.097 0.299 0.045 0.252
50 % R8: R7: R8: R7: R8: R7: R8: R7:
0.076 0.270 0.064 0.259 0.097 0.310 0.047 0.257
100 % R8: R7: R8: R7: R8: R7: R8: R7:
0.077 0.272 0.065 0.262 0.097 0.300 0.049 0.264
1000 kVA WTDG at Bus 2 0% R8: R7: R8: R7: R8: R7: R8: R7:
0.073 0.264 0.063 0.256 0.097 0.297 0.044 0.252
50 % R8: R7: R8: R7: R8: R7: R8: R7:
0.074 0.267 0.064 0.258 0.097 0.305 0.046 0.257
100 % R8: R7: R8: R7: R8: R7: R8: R7:
0.075 0.270 0.064 0.262 0.097 0.309 0.047 0.262
1000 kVA SBDG at Bus 2 &1000 kVA WTDG at Bus 7 0% R8: R7: R8: R7: R8: R7: R8: R7:
0.070 0.259 0.061 0.241 0.097 0.297 0.045 0.246
50 % R8: R7: R8: R7: R8: R7: R8: R7:
0.071 0.262 0.063 0.250 0.097 0.300 0.045 0.250
100 % R8: R7: R8: R7: R8: R7: R8: R7:
0.072 0.267 0.064 0.262 0.097 0.300 0.046 0.257

5942
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Table 15 coordination, demonstrating the superior performance of the proposed

The clearing time interval of the primary relay (R8) and backup relay (R7) for protection scheme (scheme D). For conciseness, the results in Table 15
the three fault locations on feeder L10-11. show the CTI of the primary relay (R8) and the backup relay (R7) for
Case Study Fault CTI (sec) three fault locations, assumed to be at 0 %, 50 %, and 100 % of the
Location
Scheme Scheme Scheme Scheme
feeder length on feeder L10–11 across five scenarios. The results indicate
A B C D that despite variations in DG power ratings, types, and locations, relays
R8 and R7 with the proposed protection scheme still maintain coordi­
With out DG 0% 0.201 0.200 0.231 0.213
connected 50 % 0.203 0.205 0.237 0.215 nation times, with CTI values exceeding 0.2 s. However, in protection
100 % 0.205 0.213 0.246 0.220 schemes A and B, the CTI values fall below the minimum acceptable
600 kVA SBDG at 0% 0.189 0.192 0.207 0.206 threshold for coordination. Therefore, the results in Table 15 demon­
Bus 2 50 % 0.191 0.193 0.204 0.209 strate the effectiveness of the proposed protection scheme in dis­
100 % 0.193 0.197 0.200 0.210
1000 kVA SBDG 0% 0.191 0.193 0.202 0.207
tinguishing between the operations of the backup relay and the primary
at Bus 2 50 % 0.193 0.195 0.213 0.210 relay in all simulation cases with varying DG ratings and installation
100 % 0.195 0.197 0.203 0.215 locations.
1000 kVA WTDG 0% 0.191 0.193 0.200 0.208
at Bus 2 50 % 0.193 0.194 0.208 0.211
5.1.4. D. The proposed protection scheme considering a radial system with
100 % 0.195 0.198 0.212 0.215
1000 kVA SBDG 0% 0.189 0.180 0.200 0.201 WTDG
at Bus 2 & 50 % 0.191 0.187 0.202 0.205 This subsection presents the performances of the proposed protection
1000 kVA WTDG 100 % 0.195 0.198 0.200 0.211 scheme (scheme-D) on the radial systems with WTDG. The results show
at Bus 7 the tripping times and CTI of the primary and backup relays of three
schemes for the fault on the radial system with WTDG. The LVRT time
In the case of faults on five feeders in the radial system with according to the FERC Order No.661 grid code requirement (Zavadil
1000 kVA WTDG at bus 7 and 600 kVA SBDG at bus 2, despite the dif­ et al., 2005) is shown in the second column of Table 16 and Table 17.
ferences in DG type and DG-rated power, the total operating times of The results are compared with the tripping time of the primary relay (Rp)
seven relays of the scheme-C remain to be less than the total of relay and backup relay (Rb) in the column of each protection scheme. As
operating times of scheme-A and scheme-B, which are reduced by shown in Table 16, when a short-circuit occurs on the feeder L5–6 at 0 %
12.65 % of the total operating times of scheme-A, 11.09 % of the full of the line length on the radial system with 1000 kVA WTDG at bus 2,
operating times of scheme-B and 4.26 % of the entire operating times of the tripping time of the primary relay (R4) of the proposed protection
scheme-C. The results in the case of faults on feeders in the radial system scheme is 0.047 s, which satisfies the FERC Order No.661 grid code.
connecting with DG clearly show that the performance of the proposed Thus, the tripping times of the proposed scheme are reduced by
protection scheme is superior when we compare it to scheme A, scheme 37.33 %, 16.07 %, and 6 %, respectively.
B, and scheme C. In the case of a short circuit incidence on the feeder L10–11 at
Table 14 presents the fault-clearing times on feeder L10–11 for relays location 0 % of feeder length on the radial system with 1000 kVA WTDG
R8 (primary relay) and R7 (backup relay) under each protection scheme, at bus 7 and 1000 kVA SBDG at bus 2, the simulation results, as shown in
comparing scenarios without DG installation and the four scenarios in Table 17, show that the primary relay (R8) of the proposed protection
Table 2. The results in Table 14 show that the primary relay operation of scheme can perform within 0.045 seconds, which does not exceed the
R8 in the proposed protection scheme (Scheme D) is able to clear faults LVRT time according to the requirements of FERC Order No.661. It is
faster than relay R8 in Schemes A, B, and C in all simulation cases with also less than the tripping time of the primary relay of scheme-A,
varying DG ratings and installation locations. Additionally, the opera­ scheme-B, and scheme-C, with 35.71 %, 13.46 %, and 50 % tripping
tion of relay R7 in the proposed protection scheme (Scheme D) is also time reductions, respectively.
capable of providing sufficiently fast backup protection in all cases. Fig. 16 illustrates the voltage profile at bus 2 of the 1000 kVA WTDG
Therefore, the results in Table 14 demonstrate the responsiveness and location caused by a three-phase short circuit at 0 % of the feeder L5–6
superior performance of both the primary and backup relays in the length, with the operation of the primary relay (R4) under protection
proposed protection scheme compared to Schemes A, B, and C. scheme A (black line), scheme B (green line), scheme C (green line), and
scheme D (red line), respectively, according to the results shown in
5.1.3. C. The proposed protection scheme considering primary and backup Table 16. Scheme D is capable of isolating the short circuit rapidly,
relay coordination with DG allowing the voltage magnitude at bus 2 to recover to above 0.9 p.u.
This subsection presents the results of the relay coordination of faster than the other cases when compared to protection schemes A, B,
schemes A, B, C, and D on the radial system connected with the DGs. The and C. This fast short circuit isolation capability of scheme D enhances
results of relay coordination for each scheme are indicated by the the Fault Ride-Through (FRT) capability of the WTDG, maintains the
clearing time interval (CTI) between the primary and backup relays for connection between the WTDG and the power system under short circuit
faults on feeders. This section’s analysis considers the effects of DG lo­ conditions, and also helps maintain the stability of the power system
cations, rated powers, and types to evaluate the CTI of relay during fault conditions.

Table 16
The tripping time of the relay piars (R4 and R3) for the fault location on feeder L5-6 considering with LVRT time.
Fault Location Time of LVRT Operating Time (sec.)
(sec.)
Scheme A Scheme B Scheme C Scheme D

Rp Rb Rp Rb Rp Rb Rp Rb

0% 1.000 R4: R3: R4: R3: R4: R3: R4: R3:

0.075 0.263 0.056 0.247 0.06 0.350 0.047 0.250
50 % 1.150 R4: R3: R4: R3: R4: R3: R4: R3
0.076 0.266 0.057 0.250 0.061 0.383 0.048 0.254
100 % 1.358 R4: R3: R4: R3: R4: R3: R4: R3:
0.077 0.271 0.060 0.258 0.06 0.433 0.049 0.259

5943
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Table 17
The tripping time of the relay piars (R8 and R7) for the three fault locations on feeder L10-11 considering with LVRT time.
Fault Location Time of LVRT Operating Time (sec.)
(sec.)
Scheme A Scheme B Scheme C Scheme D

Rp Rb Rp Rb Rp Rb Rp Rb

0% 1.000 R8: R7: R8: R7: R8: R7: R8: R7:

0.070 0.263 0.061 0.232 0.0972 0.297 0.045 0.246
50 % 1.150 R8: R7: R8: R7: R8: R7: R8: R7:
0.072 0.263 0.063 0.240 0.0973 0.299 0.046 0.251
100 % 1.358 R8: R7: R8: R7: R8: R7: R8: R7:
0.075 0.270 0.064 0.254 0.0976 0.296 0.047 0.258

Fig. 16. The voltage profile at bus2 of the 1000 kVA WTDG location.a: The convergence curve between the relative error and iteration number of the optimal setting
value determination for relay pairs R1-R2–19 b: The convergence curve between the relative error and iteration number of the optimal setting value determination
for relay pairs R5-R6–26, R5-R6c: The convergence curve between the relative error and iteration number of the optimal setting value determination for relay pairs
R16-R17 d: The convergence curve between the relative error and iteration number of the optimal setting value determination for relay pairs R31-R32.

5944
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 16. (continued).

5.2. The simulation result of IEEE33bus of Scheme A and Scheme B, as Scheme A has only one parameter and
Scheme B has two parameters, respectively. The convergence graphs in
The simulation of feeder protection systems on the IEEE 33-bus Figs. 16(a), 16(b), 16(c), and 16(d) illustrate an example of relative error
distribution system is divided into two subsections. Subsection A pre­ converging to zero while determining the optimal settings for the pri­
sents the simulation results of the feeder protection system on a distri­ mary and backup relay pairs in Scheme D. This demonstrates that the
bution network without DG installation. Subsection B presents the optimization algorithm, as shown in the flow chart in Fig. 15, can
simulation results of the protection schemes with distributed generation effectively be utilized to determine the optimal relay settings for the
(DG) integration under specified power levels and various locations, proposed protection scheme. The tripping times of the relays for each
along with the coordination performance results between the primary scheme are outlined in Table 19. This table provides an example of the
and backup relays of each protection scheme in the presence of DG. tripping times, specifically at the midline short-circuit point for the pair
Additionally, Subsection B presents the voltage characteristics during of primary and backup relays. The column tp represents the operating
the short-circuit clearance process for each scheme on a power system time of the primary relay, while the column tb indicates the operating
connected to WTDG, considering the LVRT time and the simulation of time of the backup relay for each scheme.
the proposed protection relay operation under the dynamic changes of For example, regarding the short circuit at the midpoint of feeder
renewable energy sources in the distribution system. L1–2, relay R1 in Scheme-D can clear a short circuit within
0.239 seconds. For a short circuit on feeder L2–3, relay R1 can still clear
5.2.1. Simulation result of relay on IEEE 33-bus without DG installation the fault in 0.488 seconds while serving as backup for relay R2. In both
This section presents the simulation results of the protection systems instances, the fault clearing time of relay R1 is faster than that of relay
for Scheme A, Scheme B, Scheme C, and Scheme D, where the system R1 in Scheme-A, Scheme-B, and Scheme-C. The results in Table 19
operates without the installation of Distributed Generation (DG), serving indicate that most primary and backup relays in Scheme-D have lower
as the baseline case. The voltage and current values of the short-circuit tripping times than those in Scheme-A, Scheme-B, and Scheme-C. For
at the midpoint on each feeder are used as input data to determine the instance, the tripping time of R17 in Scheme-D is 0.049 seconds, less
relay settings for each protection scheme. The relay setting values for than the tripping times of R17 in Scheme-A, Scheme-B, and Scheme-C,
each scheme are shown in Table 18, and the last row of Table 18 presents which are 0.1 seconds, 0.099 seconds, and 0.150 seconds, respectively.
the calculation time required to determine the optimal setting values for Additionally, the last row of Table 18 presents the total operation
each scheme. The computation time for the optimization algorithm of times of the relays for each scheme in both primary and backup pro­
Scheme D is similar to that of Scheme C since both have three param­ tection modes. The results show that the total operation time for primary
eters. However, the computation time for Scheme D is longer than that protection in Scheme D is 4.69 seconds, which is a reduction of 98.44 %

5945
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Table 18
Optimal setting values of each scheme for IEEE 33-bus.
Relay Ip(p.u.) Scheme A Scheme B Scheme C Scheme D

TMS K TMS A B C Vp D TMS

R1 0.0916 1.134 1.642 0.889 1.250 0.069 0.110 0.272 0.323 0.257
R2 0.0816 1.109 1.755 0.772 1.087 0.186 0.372 0.4706 0.962 0.676
R3 0.0590 1.220 1.703 0.984 1.053 0.104 0.208 0.036 0.135 0.05
R4 0.0558 0.731 1.708 1.000 1.082 0.090 0.070 0.296 0.836 0.115
R5 0.0546 0.596 1.702 0.989 1.089 0.112 0.141 0.493 0.827 0.953
R6 0.0254 0.713 1.075 0.975 1.117 0.076 0.029 0.184 1.000 0.078
R7 0.0210 0.679 1.068 0.870 1.003 0.075 0.194 0.258 1.000 0.208
R8 0.0162 0.657 1.020 0.955 1.000 0.074 0.227 0.4274 0.747 0.647
R9 0.0148 0.595 1.000 0.968 1.022 0.084 0.329 0.2073 0.967 0.221
R10 0.0136 0.533 0.098 0.395 1.006 0.041 0.330 0.036 0.991 0.077
R11 0.0122 0.419 1.231 0.973 1.064 0.040 0.022 0.097 0.602 0.057
R12 0.0110 0.404 1.394 0.945 1.005 0.044 0.186 0.265 1.000 0.357
R13 0.0092 0.348 1.156 0.779 1.036 0.022 0.010 0.086 0.931 0.153
R14 0.0063 0.307 1.334 0.828 1.146 0.040 0.130 0.848 0.139 1.539
R15 0.0048 0.234 1.696 0.890 1.058 0.088 0.500 0.800 0.1441 0.960
R16 0.0034 0.153 1.942 0.762 1.022 0.250 0.392 0.809 0.406 1.689
R17 0.0022 0.058 0.828 0.126 1.019 1.000 0.149 0.943 0.497 1.200
R2–19 0.0078 2.067 1.994 0.825 1.250 0.103 0.500 0.576 0.341 0.378
R19 0.0062 2.067 1.968 0.560 1.250 0.341 0.500 0.672 0.723 0.492
R20 0.0040 1.941 1.974 0.862 1.000 1.000 0.349 0.239 1.000 0.135
R21 0.0022 1.958 1.876 0.389 1.002 1.000 0.150 0.922 0.476 0.896
R3–23 0.0210 1.250 1.987 0.613 1.077 0.074 0.168 0.405 1.000 0.207
R23 0.8160 1.063 1.964 0.414 1.042 0.129 0.311 0.438 0.871 0.163
R24 0.0096 1.163 1.533 0.204 1.203 0.999 0.297 0.950 0.419 0.907
R6–26 0.0284 0.522 1.028 0.646 1.188 0.074 0.072 0.163 1.000 0.294
R26 0.0270 0.482 1.631 0.851 1.028 0.090 0.264 0.177 0.473 0.09
R27 0.0262 0.343 1.999 0.899 1.048 0.097 0.154 0.405 0.995 0.643
R28 0.0248 0.279 1.971 0.992 1.010 0.068 0.176 0.180 0.794 0.104
R29 0.0222 0.224 1.860 0.843 1.071 0.063 0.145 0.268 0.878 0.447
R30 0.0100 0.213 1.932 0.883 1.055 0.069 0.212 0.194 0.996 0.193
R31 0.0066 0.142 1.842 0.686 1.091 0.085 0.263 0.085 1.000 0.128
R32 0.0018 0.068 1.809 0.204 1.038 1.000 0.149 0.920 0.464 0.883
Computation time(s) 0.786 1.195 1.370 1.552

Table 19
The operating time of the primary and backup relays for each scheme.
Fault Scheme-A Scheme-B Scheme-C Scheme-D
Location tp(sec) tb(sec.) tp(sec.) tb(sec.) tp(sec.) tb(sec.) t(s) t(s)

L1–2 R1:3.232 - R1:0.517 - R1:0.711 - R1:0.239 -

L2–3 R2:2.807 R1:3.392 R2:0.499 R1:0.730 R2:0.540 R1:1.077 R2: 0.147 R1:0.488
L3–4 R3:2.539 R2:3.093 R3:0.565 R2:0.764 R3:0.408 R2:0.681 R3:0.162 R2:0.361
L4–5 R4:2.020 R3:2.454 R4:0.582 R3:0.781 R4:0.343 R3:0.614 R4:0.053 R3:0.506
L5–6 R5:1.826 R4:2.259 R5:0.740 R4:0.939 R5:0.494 R4:0.711 R5:0.235 R4:0.434
L6–7 R6:1.751 R5:2.041 R6:0.893 R5:1.092 R6:0.367 R5:0.748 R6:0.057 R5:0.442
L7–8 R7:1.644 R6:1.853 R7:0.870 R6:1.069 R7:0.199 R6:0.571 R7:0.094 R6:0.298
L8–9 R8:1.552 R7:1.768 R8:0.914 R7:1.112 R8:0.426 R7:0.637 R8:0.194 R7:0.393
L9–10 R9:1.467 R8:1.678 R9:0.964 R8:1.164 R9:0.509 R8:0.713 R9:0.152 R8:0.351
L10–11 R10:1.324 R9:1.526 R10:0.892 R9:1.091 R10:0.402 R9: 0.643 R10:0.133 R9:0.333
L11–12 R11:1.141 R10:1.343 R11:0.709 R10:0.908 R11:0.336 R10:0.535 R11:0.074 R10:0.273
L12–13 R12:0.992 R11:1.203 R12:0.665 R11:0.865 R12:0.525 R11:0.753 R12:0.172 R11:0.371
L13–14 R13:0.844 R12:1.049 R13:0.623 R12:0.822 R13:0.537 R12:0.884 R13:0.200 R12:0.400
L14–15 R14:0.666 R13:0.870 R14:0.469 R13:0.695 R14:0.677 R13:0.980 R14:0.570 R13:0.770
L15–16 R15:0.480 R14:0.683 R15:0.357 R14:0.556 R15:0.617 R14:0.876 R15:0.396 R14:0.707
L16–17 R16:0.293 R15:0.499 R16:0.244 R15:0.443 R16:0.421 R15:0.744 R16:0.228 R15:0.427
L17–18 R17:0.12 R16:0.303 R17:0.099 R16:0.299 R17:0.15 R16:0.447 R17:0.050 R16:0.249
L2–19 R2–19:2.921 R1:3.327 R2–19:0.178 R1:0.619 R2–19:0.687 R1:0.902 R2–19:0.120 R1:0.320
L19–20 R19:3.020 R2–19:3.243 R19:0.239 R2–19:0.438 R19:0.557 R2–19:0.916 R19:0.0932 R2–19:0.293
L20–21 R20:2.804 R19:3.312 R20:0.203 R19:0.402 R20:0.349 R19:0.584 R20:0.057 R19:0.260
L21–22 R21:2.579 R20:2.941 R21:0.099 R20:0.298 R21:0.15 R20:0.350 R21:0.050 R20:0.252
L3–23 R3:23:2.265 R2:2.997 R3–23:0.205 R2:0.801 R3–23:0.435 R2:0.697 R3–23:0.05 R2:0.418
L23–24 R23:2.152 R3–23:2.475 R23:0.172 R3–23:0.371 R23:0.482 R3–23:0.769 R23:0.049 R3–23:0.250
L24–25 R24:2.11 R23:2.386 R24:0.099 R23:0.298 R24:0.300 R3–23:0.658 R24:0.055 R23:0.249
L6–26 R6–26:1.385 R5:1.985 R6–26:0.603 R5:1.013 R6–26:0.459 R5:0.698 R6–26:0.187 R5:0.384
L26–27 R26:1.218 R6–26:1.421 R26:0.449 R6–26:0.649 R26:0.369 R6–26:0.574 R26:0.092 R6–26:0.294
L27–28 R27:0.926 R26:1.318 R27:0.439 R26:0.639 R27:0.430 R26:0.629 R27:0.173 R26:0.357
L28–29 R28:0.812 R27:1.021 R28:0.475 R27:0.674 R28:0.407 R27:0.687 R28:0.104 R27:0.420
L29–30 R29:0.695 R28:0.858 R29:0.413 R28:0.612 R29:0.435 R28:0.653 R29:0.190 R28:0.389
L30–31 R30:0.480 R29:0.694 R30:0.340 R29:0.541 R30:0.458 R29:0.734 R30:0.132 R29:0.331
L31–32 R31:0.297 R30:0.499 R31:0.224 R30:0.423 R31:0.413 R30:0.620 R31:0.128 R30:0.327
L32–33 R32:0.102 R31:0.304 R32:0.052 R31:0.252 R32:0.150 R31:0.483 R32:0.049 R31:0.249
Total time 48.444 54.793 14.818 21.359 13.74 21.57 4.69 11.50

5946
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 17. The operating time of primary protection of all relays in each protection scheme.

Fig. 18. The operating time of backup protection of all relays in each protection scheme.

compared to the total operating times of Scheme A, 68.35 % of the total solid black line indicates the variation of CTI values for the relay pairs in
operating times of Scheme B, and 65.86 % of the total operating times of Scheme-D. The CTI values for Scheme-D consistently remain within the
Scheme C. Meanwhile, the total operation time for backup protection in range of 0.2–0.5 across all primary and backup relay pairs. This con­
Scheme D is 11.50 seconds, which is also less than the tripping times of trasts with the CTI graph for Scheme-A, where the CTI exceeds the
the primary relays of Scheme A, Scheme B, and Scheme C, with re­ permissible range for the pairs R1-R2, R3-R4, R3–23-R23, and R5-
ductions of 79.011 %, 46.158 %, and 46.685 %, respectively. R6–26. The difference in CTI values between Scheme-B and Scheme-D is
Fig. 17 shows that the primary operation of all scheme-D relays is significant for the pair R2-R3–23. In Scheme-B, the CTI for this pair is
faster than that of Scheme-A, Scheme-B, and Scheme-C. This is illus­ 0.597 seconds, whereas in Scheme-D, it is 0.368 seconds. The CTI graph
trated by the solid black line in the graph, which is lower than the for Scheme-C exhibits greater fluctuations than that of Scheme-D
tripping timelines of Scheme-A, Scheme-B, and Scheme-C. Meanwhile, because the tripping time of all the relays in Scheme-C depends on a
the backup protection operation of all relays in Scheme D is also faster variable C and is not controlled by the TMS value as it is in Scheme-D. In
than the backup protection operation of relays in Schemes A, B, and C. Scheme-C, the CTI exceeds 0.5 seconds for the pair R1-R2. The results
This is shown in Fig. 18, where the backup protection tripping time displayed in the graph demonstrate that the proposed relay character­
graph for Scheme D (solid black line) is lower than the tripping time istics effectively maintain the CTI values between the primary and
graphs of Scheme A, Scheme B, and Scheme C. backup relays within an efficient range in the protection system.
Fig. 19 illustrates the graph of CTI levels for primary and backup
relay pairs within each protection scheme. The horizontal axis of the 5.2.2. Simulation result of relay on IEEE 33-bus with DG installation
graph represents the pair numbers of the primary and backup relays, The simulation of the proposed protection scheme on the IEEE 33-
while the vertical axis displays the CTI values for these relay pairs. The bus system with distributed generation (DG) installation demonstrates

5947
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 19. The CTI of all relay pairs for the IEEE 33bus system.

the adequate performance of the proposed relay characteristics.

Table 20
Table 20 and Table 21 show the tripping times of the primary relay R17
The tripping times of each relay for all cases with and without DG installation in
and the backup relays R16, R15, and R14 for each scheme in the event of
the IEEE 33 bus system.
a short circuit on feeder L17–18. The results in the tables compare
Scenarios Fault Scheme Tripping Time of Relays scenarios with and without DG (Distributed Generation) installation,
current
(A)
Rp Rb1 Rb2 Rb3 based on the scenarios outlined in Table 3. The results in Table 20
R17 R16 R15 R14 indicate that, in cases where the short-circuit current level changes due
Without DG 471.97 A 0.101 s 0.303 s 0.518 s 0.741 s to the installation of DG in the system, the primary relay R17 operating
installation B 0.098 s 0.298 s 0.525 s 0.760 s under the proposed protection scheme (Scheme D) is able to clear the
C 0.149 s 0.446 s 0.845 s 1.255 s fault faster than relay R17 in Schemes A, B, and C. Furthermore, the
D 0.050 s 0.249 s 0.458 s 0.681 s
With SBDG 814.72 A 0.087 s 0.260 s 0.437 s 0.617 s
operation of relays R16, R15, and R14 under the proposed protection
2 MW at Bus B 0.087 s 0.324 s 0.587 s 0.828 s scheme (Scheme D) can provide backup protection quickly enough to
13 &WTDG C 0.149 s 0.460 s 0.926 s 1.484 s support the operation of the primary relay. They can also maintain the
6 MW at Bus D 0.050 s 0.278 s 0.510 s 0.753 s coordination time between the primary and backup relays within the
14
defined range of 0.2–0.5 seconds, as shown in Table 21. Therefore, the
With SBDG 964.11 A 0.084 s 0.248 s 0.417 s 0.586 s
2 MW at Bus B 0.084 s 0.344 s 0.635 s 0.828 s results in Table 20 and Table 21 demonstrate both the sensitivity of the
13 &WTDG C 0.150 s 0.466 s 0.957 s 1.574 s primary relay’s operation and the effectiveness of the proposed pro­
9 MW at Bus D 0.050 s 0.296 s 0.548 s 0.804 s tection scheme in distinguishing between the roles of backup and pri­
14 mary relays.
Fig. 20 illustrates the voltage signal at bus 14 in a system where the
DG capacity has been increased to 11 MW, featuring 9 MW WTDG
Table 21 installed at bus 14 and 2 MW SBDG at bus 13, as detailed in the second
The CTI of relay pairs for all cases with and without DG installation in the IEEE scenario in Table 3. For example, a short circuit is simulated at the
33 bus system. midpoint of feeder L17–18. However, the primary relay, R17, and
Scenarios Scheme Fault CTI of relay pairs backup relays, R16 and R15, fail to trip. Consequently, the backup relay
current
Rp-Rb1 Rb1-Rb2 Rb2-Rb3 R14 activates to isolate the fault. The tripping times of relay R14 for
(A)
R17- R16- R15- each scheme are presented in Table 20 and are indicated in Fig. 20 as tA,
R16 R15 R14 tB, tC, and tD. As illustrated in this figure, the backup relay of scheme D
Without DG installation A 471.97 0.202 s 0.215 s 0.223 s can isolate the fault within 0.804 seconds, enabling the voltage at Bus 14
B 0.200 s 0.226 s 0.234 s to recover within 3 seconds, as mandated by the FERC grid code.
C 0.297 s 0.398 s 0.409 s In contrast, the backup relay of Scheme-C operates more slowly than
D 0.200 s 0.209 s 0.222 s
Scheme-D because Relay R14 of Scheme-C has a CTI value that differs
With SBDG 2 MW at Bus A 814.72 0.172 s 0.177 s 0.179 s
13 &WTDG 6 MW at B 0.237 s 0.262 s 0.241 s from R15 by 0.61 seconds. This results in a tripping time of
Bus 14 C 0.311 s 0.465 s 0.558 s 1.547 seconds, which exceeds the time limit specified by the grid code.
D 0.228 s 0.232 s 0.243 s The graph shows that the blue line encroaches on the LVRT graph
With SBDG 2 MW at Bus A 964.11 0.164 s 0.168 s 0.169 s (dashed red line) at point A. After the voltage drop, the WTDG is
13 &WTDG 9 MW at B 0.260 s 0.291 s 0.250 s
Bus 14 C 0.316 s 0.491 s 0.617 s
potentially disconnected from the system. Scheme-B’s backup relay can
D 0.246 s 0.252 s 0.256 s clear the fault within 0.886 seconds, staying within the time limit
specified by the grid code. For the operation of the backup relay in

5948
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 20. The voltage profile at bus 14 of the 9MVA WTDG location.

Fig. 21. The operation of the proposed relay under current variations due to a 9 MW WTDG connected with the system.

Scheme-A, although it operates faster than the backup relay of Scheme- bus 14 to a value greater than 0.9 pu, as the grid code requires. Addi­
D, during a fault, the coordination between the primary and backup tionally, it maintains coordination between primary and backup relays
relays drops below 0.2 seconds. This may lead to malfunction or mis­ within a range of 0.2–0.5 seconds. These results indicate that the pro­
coordination between the primary and backup relays. This simulation posed protection scheme operates effectively as backup protection. The
demonstrates that the proposed protection scheme (Scheme-D) can simulation results demonstrate that Scheme D can maintain the coor­
quickly isolate the fault enough to restore the voltage of the WTDG at dination time between the primary relay and the backup relay within the

5949
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Fig. 22. The operation of the proposed relay under voltage variations due to a 9 MW WTDG connected with the system.

0.2–0.5 seconds range, even with an increase in short-circuit current due scheme put forth in (Keil and Jager, 2008).
to DG installation in the system. This capability helps reduce the The simulation of the IEEE 13-bus and IEEE 33-bus systems under
occurrence of cascade tripping in a radial power system. varying generation capacities, different types, and DG installation lo­
The operation of the proposed protection relay during fluctuations in cations has demonstrated that the coordination between primary and
renewable energy supply is analyzed, as illustrated in Figs. 21 and 22. backup relays within the proposed characteristics can maintain the CTI
For example, a 9 MW wind power plant is suddenly connected and value in the 0.2–0.5 seconds range. This differs from the conventional
disconnected at Bus 14, causing voltage oscillations and a temporary characteristics and those referenced in Saleh et al. (2015) and Keil and
voltage drop to 0.824 per unit, which is below the pickup voltage of Jager (2008), which are influenced by changes in generation capacity
relay R14. However, relay R14 does not activate because the maximum and DG installation locations, resulting in a CTI value for primary and
current flowing through line L14–15 is 0.0042 per unit, which does not backup relay pairs falling below 0.2 seconds. The ability to sustain co­
exceed the pickup current of relay R14. Consequently, the voltage and ordination between primary and backup relay pairs in the proposed
current fluctuations from this event do not impact the operation of relay protection scheme aids in minimizing malfunctions of the protection
R14. The simulation results demonstrate that the proposed relay char­ system, such as cascade tripping in the distribution network. This im­
acteristics can function effectively in a system with dynamic changes in proves the stability and reliability of power distribution in the electrical
electrical power from renewable energy sources. This capability en­ system. Furthermore, the simulation of voltage fluctuations and
hances the stability and reliability of power distribution in a radial short-circuit currents in a radial power distribution system with DG
power system. installation, along with the simulation of power variations from DG
within the distribution system, illustrates that the proposed protection
6. Conclusion scheme functions effectively in a distribution system experiencing dy­
namic electrical power changes from DG. This capability improves the
The coordination issue of overcurrent protection relays, caused by stability and reliability of power distribution in a radial power system.
the rise in short-circuit currents within the radial distribution system
featuring DG, has resulted in new operating characteristics for over­ Funding
current relays. This paper presents a computational technique involving
the operational features of a voltage-controlled overcurrent protection This work was supported by the National Science, Research and
relay. The relay’s operating time can be influenced by altering the D Innovation Fund (NSRF), and Thailand Science Research and Innovation
parameter, which adjusts the voltage multiplier value. The setting (TSRI) through Rajamangala University of Technology Thanyaburi
values for four overcurrent protection schemes were established using (Grant No.: FRB680045/0168, Project Code: FRB68E0707).
an optimal setting method under three-phase short-circuit conditions in
a radial distribution system without DG installation. Simulation results CRediT authorship contribution statement
highlighted the superior performance of the proposed overcurrent pro­
tection scheme compared to three others: the conventional overcurrent Bhumkittipich Krischonme N.: Writing – review & editing, Vali­
protection scheme, the scheme introduced in (Saleh et al., 2015), and dation, Supervision, Methodology, Investigation, Conceptualization.
the one proposed in (Keil and Jager, 2008). In the IEEE 13-bus test Sillawat Romphochai: Writing – review & editing, Validation, Super­
system, the proposed protection scheme reduced the total operating vision, Software, Resources, Methodology, Investigation, Data curation,
time of all relays while maintaining the coordination time interval be­ Conceptualization. Nattapol Haupala: Writing – original draft, Soft­
tween primary and backup relays under short-circuit conditions in the ware, Resources, Project administration, Methodology, Formal analysis,
radial distribution system with varying sizes and locations of DG in­ Data curation, Conceptualization.
tegrations. When a short circuit occurs in a test system with SBDG and
WTDG installation, the proposed overcurrent protection scheme can Declaration of Competing Interest
swiftly isolate the short circuit within the timeframe stipulated by the
grid code, allowing the WTDG to remain connected to the power system The authors declare that they have no known competing financial
during such conditions. In testing on the IEEE 33-bus system, the pro­ interests or personal relationships that could have appeared to influence
posed protection scheme yields a combined operating time for both the work reported in this paper.
primary and backup protections that is less than that of the total oper­
ating time for the three other schemes: the conventional overcurrent
protection scheme, the scheme proposed in (Saleh et al., 2015), and the

5950
N. Haupala et al. Energy Reports 13 (2025) 5931–5951

Data availability Lee, H.J., Son, G., Park, J.-W., Aug. 2011. "Study on wind-turbine generator system sizing
considering voltage regulation and overcurrent relay coordination. IEEE Trans.
Power Syst. 26 (3), 1283–1293. https://doi.org/10.1109/TPWRS.2010.2091155.
The authors do not have permission to share data. Leschert, D., Sommerstad, K., Tremblay, E., Bennett, A., 2021. "Microprocessor-based
protective relay configurations: effective documentation in industrial applications
References (July-Aug). IEEE Ind. Appl. Mag. 27 (4), 25–35. https://doi.org/10.1109/
MIAS.2021.3063091 (July-Aug).
Lim, S.T., Lim, S.H., Jun. 2020. "Analysis on protective coordination between over-
"IEEE Standard for Inverse-Time Characteristics Equations for Overcurrent Relays. vol., current relays with voltage component in a power distribution system with SFCL.
no IEEE Std C37. 112-2018 (Revis. IEEE Std C37. 112-1996), 5 Feb. 2019 1–25. IEEE Trans. Appl. Supercond. 30 (4), 1–6. https://doi.org/10.1109/
https://doi.org/10.1109/IEEESTD.2019.8635630. TASC.2020.2968252.
Baran, M.E., Wu, F.F., Apr. 1989. ‘‘Network reconfiguration in distribution systems for Mason, C.R., Oct. 2011. Art & Science of Protective Relaying, Chapter 2, GE Consumer &
loss reduction and load balancing. IEEE Trans. Power Del. 4 (2), 1401–1407. Electrical. Retrieved, pp. 14–35.
Barker, P., De Mello, R., Jul. 2000. "Determining the impact of distributed generation on Meskin, M., Domijan, A., Grinberg, I., Nov. 2020. "Impact of distributed generation on
power systems. I. Radial distribution systems. Proc. IEEE Power Eng. Soc. Summer the protection systems of distribution networks: analysis and remedies–review
Meeting, Seattle, WA, USA, pp. 1645–1656. https://doi.org/10.1109/ paper. IET Gener. Trans. Distrib. 14 (24), 5944–5960. https://doi.org/10.1049/iet-
PESS.2000.868775. gtd.2019.1652.
Calderaro, V., Milanovic, J.V., Kayikci, M., Piccolo, A., Jan. 2009. The impact of Monika, Y., Nitai, P., Devender, S.K., Dec. 2022. Low voltage ride through capability for
distributed synchronous generators on quality of electricity supply and transient resilient electrical distribution system integrated with renewable energy resources.
stability of real distribution network. Elec Power Syst. Res. 79 (1), 134–143. https:// Energy Rep. 9, 833–858. https://doi.org/10.1016/j.egyr.2022.12.023.
doi.org/10.1016/j.epsr.2008.05.022. Nattapol, H., Krischonme, B., 2020. "Voltage-based non-standard inverse time for over
Ghulam, S.K., Jie, W., Mazhar, H.B., Aug. 2016. Active and reactive power control of the current relay in distribution system connected dfig-based wind turbines. GMSARN
double fed induction generator based on wind energy conversion system. Energy Int. J. 15 (2), 113–120.
Rep. 2, 194–200. https://doi.org/10.1016/j.egyr.2016.08.001. Noghabi, A.S., Mashhadi, H.R., Sadeh, J., July 2010. "Optimal coordination of directional
Girgis, A., Brahma, S., Jul. 2001. "Effect of distributed generation on protective device overcurrent relays considering different network topologies using interval linear
coordination in distribution system. Proc. LESCOPE’01 Large Eng. Syst. Conf., programming. IEEE Trans. Power Deli 25 (3), 1348–1354. https://doi.org/10.1109/
Halifax, NS, Can. 115–119. https://doi.org/10.1109/LESCPE.2001.941636. TPWRD.2010.2041560.
Hong, L., Rizwan, M., Wasif, M., Ahmad, S., Zaindin, M., Firdausi, M., Apr. 2021. "User- Omar, A., Moayed, M., Feb. 2024. Performance improvement of rotor current controller
Defined Dual Setting Directional Overcurrent Relays with Hybrid Time Current- in doubly fed induction generation wind turbine with artificial intelligence methods.
Voltage Characteristics-Based Protection Coordination for Active Distribution Energy Rep. 11, 2236–2254. https://doi.org/10.1016/j.egyr.2024.01.064.
Network. IEEE Access 9, 62752–62769. https://doi.org/10.1109/ Sa’ed, J.A., Favuzza, S., Ippolito, M.G., Massaro, F., Jun. 2013. "Investigating the effect of
ACCESS.2021.3074426. distributed generators on traditional protection in radial distribution systems
Huchel, L., Zeineldin, H.H., May 2016. "Planning the coordination of directional (Grenoble, France). Proc. IEEE Grenoble Conf. 1–6. https://doi.org/10.1109/
overcurrent relays for distribution systems considering DG. IEEE Trans. Smart Grid 7 PTC.2013.6652100.
(3), 1642–1649. https://doi.org/10.1109/TSG.2015.2420711. Saleh, K.A., Zeineldin, H.H., Al-Hinai, A., El-Saadany, E.F.Optimal Coordination of
"IEEE Standard for Inverse-Time Characteristics Equations for Overcurrent Relays. vol., Directional Overcurrent Relays Using a New Time–Current–Voltage Characteristic,"
no IEEE Std C37. 112-2018 (Revis. IEEE Std C37. 112-1996), 5 Feb. 2019 1–25. IEEE Trans on Power Deli, vol. 30, no. 2, pp. 537-544, Apr. 2015. doi: 10.1109/
https://doi.org/10.1109/IEEESTD.2019.8635630. TPWRD.2014.2341666..
IEEE Distribution System Analysis Subcommittee, IEEE 13 Node Test Feeder, [Online]. Saleh, K.A., El Moursi, M.S., Zeineldin, H.H., Dec. 2015. "A new protection scheme
Available: 〈https://cmte.ieee.org/pes-testfeeders/resources/〉. considering fault ride through requirements for transmission level interconnected
Jun, Y., Aug. 2021. An improved fault current calculate method and protection scheme of wind parks. IEEE Trans. Ind. Inf. 11 (6), 1324–1333. https://doi.org/10.1109/
Doubly-fed Induction Generator. Energy Rep. 7, 5168–5174. https://doi.org/ TII.2015.2479583.
10.1016/j.egyr.2021.08.107. Zavadil, R., Miller, N., Ellis, A., Muljadi, E., 2005. "Making connections [wind generation
JunYing, W., ChiaKwang, T., Rahim, N.A., Rodney, H.G.Tan, Sook-Chin, Y., Dec. 2024. facilities (Nov.-Dec). IEEE Power Energy Mag. 3 (6), 26–37 (Nov.-Dec).
Communication-less adaptive overcurrent relay coordination for service restoration Zhan, H., et al., Jan. 2016. "Relay protection coordination integrated optimal placement
in distribution systems. Energy Rep. 13, 256–263. https://doi.org/10.1016/j. and sizing of distributed generation sources in distribution networks. IEEE Trans.
egyr.2024.12.015. Smart Grid 7 (1), 55–65. https://doi.org/10.1109/TSG.2015.2420667.
Keil, T., Jager, J., Jan. 2008. "Advanced coordination method for overcurrent protection
relays using nonstandard tripping characteristics. IEEE Trans. Power Del. 23 (1),
52–57. https://doi.org/10.1109/TPWRD.2007.905337.

5951