Electric Power Systems Research 230 (2024) 110215

Contents lists available at ScienceDirect

Electric Power Systems Research

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

A three-stage reliability-centered framework for critical feeder

identification, failure modes prioritization, and optimal maintenance
strategy assignment in power distribution system
Sajjad Moradi-Sarvestani , Mohammad Reza Dehbozorgi , Mohammad Rastegar *
School of Electrical and Computer Engineering, Shiraz University, Shiraz, Iran

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

Keywords: In this paper, a three-stage reliability-centered maintenance (RCM) framework for critical feeder identification,
Critical feeder identification failure modes prioritization, and optimal maintenance strategy assignment is presented for the power distribu­
Distribution system tion system. This three-stage methodology addresses a notable gap in the literature by incorporating these three
Failure modes prioritization
stages together. The first stage ranks critical feeders using the BWM and TOPSIS methods. In the second stage, the
Optimal maintenance strategy assignment
Reliability-centered maintenance (RCM)
failure modes are prioritized using feeder rank, severity, occurrence, and detection factors and are assigned a
score representing their priority. Finally, in the third stage, a maintenance strategy assignment, formulated as a
mixed-integer linear programming (MILP) optimization problem is proposed. The optimization problem con­
siders three strategies, cold-line, hot-line, and run-to-failure (RTF), assigned to each failure mode. Besides, the
possibility of implementing cold-line repairs overlapping with the upstream grid’s scheduled outages and the
availability of several maintenance crews are taken into account. The objective function is to minimize the total
maintenance costs, which include operation, equipment, energy not supplied (ENS), expected energy not sup­
plied (EENS), and future corrective maintenance costs. The proposed approach is implemented on a real dis­
tribution system. The results indicate that the proposed RCM framework leads to a 26.32 % reduction in the
EENS compared to selecting the RTF strategy for all components. Besides, the proposed optimization-based
method leads to a 29 % lower objective function than the business-as-usual maintenance planning, and the
proposed three-stage RCM framework outperforms a previously published state-of-the-art method by decreasing
the objective function and the EENS.

1. Introduction [4], tries to extend the lifespan of equipment by implementing an

effective preservation plan [5–7]. Selecting the appropriate mainte­
1.1. Motivations nance strategy is crucial for achieving this goal and increasing the sys­
tem’s reliability while minimizing the associated costs, given the limited
Nowadays, due to the increasing demand for energy resources and budget of distribution companies and the restricted available human
societies’ ever-increasing reliance on electric energy, it is imperative to resources [8].
enhance the reliability of the power system. Improving reliability level is Various maintenance methods have been proposed and used in
especially important in distribution systems, as it is the final chain in the recent years, among which are risk-based maintenance, condition-based
conventional power system hierarchy [1]. Strategic asset management maintenance, preventive maintenance (PM), and corrective mainte­
offers a highly effective approach to attaining dependable reliability [2, nance (CM) [9]. The superior method should integrate the capabilities of
3]. Maintenance, as an approach based on mid-term asset management risk-based maintenance plans and take the financial aspects of different

Abbreviations: BWM, best-worst method; CM, corrective maintenance; ENS, energy not supplied; EENS, expected energy not supplied; FM, failure mode; FMEA,
failure mode and effect analysis; MCDM, multi-criteria decision-making; MILP, mixed-integer linear programming; MINLP, mixed-integer nonlinear programming;
PM, preventive maintenance; RCM, reliability-centered maintenance; RTF, run-to-failure; SAIFI, system average interruption frequency index; TOPSIS, technique for
order of preference by similarity to ideal solution; VoLL, value of lost load.
* Corresponding author.
E-mail address: [email protected] (M. Rastegar).

https://doi.org/10.1016/j.epsr.2024.110215
Received 4 October 2023; Received in revised form 26 January 2024; Accepted 31 January 2024
Available online 17 February 2024
0378-7796/© 2024 Elsevier B.V. All rights reserved.
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

strategies into account [10]. Reliability-centered maintenance (RCM) is [16–22].

one of the most effective methods of maintenance and asset manage­ Besides, most of the reviewed works in [16–22] that discuss RCM
ment that determines the optimal maintenance strategy for different planning have not specified the type of repair strategy, cold-line, hot-­
system components [11]. According to [6], RCM consists of two main line, or RTF. The severity, occurrence, and detection are three factors
steps; First, the most critical components are identified. After that, the commonly used to prioritize FMs in current studies. However, the crit­
optimal maintenance strategies for the most critical components and icality level of the feeder that contains the FM has not been addressed in
associated failure modes (FMs) are designated. previous works. Furthermore, the potential for conducting cold-line PM
overlapping with the upstream grid’s scheduled outages is not
1.2. Literature review mentioned in the existing literature. Additionally, the possibility of
performing simultaneous cold-line repairs by multiple maintenance
Multiple methods for critical component identification have been crews has not been explored in previous studies. The key factors of the
introduced in various studies [12–15]. In [12], a method for identifying previous works and a comparison with our proposed approach are
the critical components in distribution systems for maintenance sched­ presented in Table 1.
uling is proposed, using a combination of the analytical hierarchical
process and fuzzy sets theory. In [13], prioritizing components in a
power system based on the long-term costs incurred by their failures is 1.4. Contributions
proposed. In [14], a framework based on the fuzzy best-worst method
(BWM) and the technique for order of preference by similarity to ideal In this paper, a three-stage RCM framework, which consists of critical
solution (TOPSIS) is proposed. [15] proposes a method by using the feeder identification, FMs prioritization, and optimal maintenance
fuzzy BWM-VIKOR and simultaneous evaluation of criteria and alter­ strategy assignment is presented. The proposed method incorporates the
natives (SECA) approach to eliminate the need for the opinions of ex­ two steps of the RCM; critical feeder identification and FMs prioritiza­
perts, when ranking critical components. In summary, various tion as the first step and the optimal maintenance strategy assignment as
approaches that only identify the critical components are discussed in the second one. In the first stage, critical feeders are identified and
[12–15] as the first step in the RCM planning process. ranked using BWM and TOPSIS. BWM utilizes the opinions of the experts
Optimal maintenance strategy assignment is another crucial part of on the most and the least important criteria regarding the criticality of a
RCM planning, which has been investigated in several studies [16–22]. feeder. Some of the criteria are the feeder’s value of lost load (VoLL), the
A strategy refers to the action that operators undertake to address FMs. feeder’s outage rate, and the feeder’s number of outages. In the second
The maintenance crew may shut down the whole or a part of the dis­ stage, FMs are prioritized based on four factors: severity, occurrence,
tribution feeder or may conduct the PM while the feeder is in service. detection, and the feeder rank obtained in the first stage. The prioriti­
These strategies are called cold-line and hot-line maintenance, respec­ zation is performed using the BWM and TOPSIS methods to assign a
tively. The run-to-failure (RTF) is another approach in which the oper­ score to each FM, where the lower score means the higher priority of an
ators leave the component to fail, and after the failure, the CM will be FM. In the third stage, an optimal maintenance strategy assignment plan
conducted. Among the relevant works, in [16], the RCM planned to is proposed. The proposed scheme is formulated as an MILP optimiza­
minimize PM costs and maximize the reliability of the distribution sys­ tion problem, aiming to minimize maintenance costs. The objective
tem. In [17], a method to optimize the RCM budget allocation for dis­ function includes the manpower wages, the used equipment cost, the
tribution feeders is proposed to minimize the total reliability cost. In scheduled energy not supplied (ENS) cost during maintenance, the un­
[18], a maintenance budget allocation approach employs a diagnostic scheduled expected ENS (EENS) cost due to future failures, and future
importance factor as an index to prioritize components. In [19], a expected CM costs. The FMs scores obtained at the second stage are used
bi-level mixed-integer nonlinear programming (MINLP) model, to calculate the failure rates at this stage. Three maintenance strategies
comprising maintenance planning by microgrids’ owners and the sub­ including cold-line, hot-line, and RTF are considered. The proposed
sequent final planning by the distribution system operator, for the optimal maintenance strategy assignment for the FMs of a feeder is
microgrids’ maintenance planning is presented. A mixed-integer linear constrained by budget limitations, time limits for planned power out­
programming (MILP) model to optimize the long-term maintenance ages, the number of available maintenance crews, and the maximum
planning of medium voltage overhead lines in the distribution system is time of the upstream grid’s outage. This study considers the possibility
also proposed in [20]. [21] presents a condition-based maintenance of conducting cold-line repairs overlapping with the upstream grid’s
framework for the distribution system, incorporating opportunistic outages. Hence, it is possible to incorporate the scheduled outage time of
maintenance and Monte Carlo simulation. In [22], a study on optimal the upstream grid outages into the optimal maintenance strategy plan­
maintenance strategy assignment is presented, which involves deter­ ning to reduce the cost of ENS during cold-line operations. As a result,
mining the PM strategy, including cold-line, hot-line, and RTF. To sum the cold-line repair operations are divided into two categories: (1) cold-
up, RCM-based strategy assignment, which involves maintenance line repairs with a planned power outage, which incurs ENS cost, and (2)
budget allocation and selecting the optimal PM or CM strategy is mostly cold-line repairs overlapping with the upstream grid’s scheduled out­
formulated as an optimization problem, contained by various ages, which do not incur any ENS cost to the distribution utility.
constraints. Furthermore, the possibility of simultaneous cold-line repairs performed
by multiple maintenance crews is taken into account, leading to a
1.3. Research gaps reduction in the required outage time and ENS. Moreover, to evaluate
the effectiveness of the proposed method in this paper, a comparison is
As mentioned in the literature, the RCM comprises two steps; First, made with the SECA approach, which is an MCDM method that does not
the critical system components are selected and prioritized. Then, for require opinions of the experts.
each component, a maintenance strategy is designated to minimize the In summary, the main contributions of this paper are as follows:
costs associated with its maintenance. The first and most important
research gap is the hierarchical consideration of RCM steps. Many pa­ • Proposing a three-stage RCM framework for critical feeder identifi­
pers have specified how to rank critical components of the feeder and cation, FMs prioritization, and optimal maintenance strategy
they have not specified the maintenance strategy for them [12–15]. assignment;
Other papers, on the other hand, have not specified how the prioriti­ • Proposing a novel optimization formulation for the optimal main­
zation process works and the effect of feeder ranking and FMs priority tenance strategy assignment to minimize the sum of maintenance
have not been included in the optimal maintenance planning problem costs, which contains the repair operation costs, equipment cost, ENS

2
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

Table 1
Comparing previous works with the proposed approach of this paper for RCM planning of distribution systems.
Refs. Objective function Optimization Critical Using hot-line and RTF PM overlapping with the Using multiple
method components in maintenance strategy upstream grid’s scheduled maintenance crews to
identification outages conduct PM

[16] Min PM cost and Max reliability Genetic – – – –

Algorithm
[17] Min the total reliability cost MINLP – – – –
[18] – – ✓ – – –
[19] Min SAIFI MINLP – – – –
[20] Min CM and PM costs MILP – – – –
[21] – – ✓ – – –
[22] Min maintenance interruption MILP ✓ ✓ – –
and operation costs
Proposed Min maintenance interruption MILP ✓ ✓ ✓ ✓
approach cost, operation costs, EENS, and
future CM costs

cost during maintenance, EENS cost due to future failures, and ex­ 6) Maneuver switch density, which is equal to the number of maneuver
pected future CM costs; switches divided by the length of the feeder.
• Incorporating the duration of the upstream grid’s scheduled outages 7) Switch density, which is equal to the number of switches (breakers,
in the optimal maintenance strategy assignment of the three-stage reclosers, and disconnectors) of the feeder divided by the length of
RCM framework; the feeder.
• Modeling the effect of simultaneous cold-line maintenance operation 8) PM density, which is equal to the number of PMs performed on the
by multiple crews on the maintenance strategy assignment that may feeder in the year before the year under study, divided by the length
reduce the ENS of the cold-line maintenance in the proposed of the feeder. The mean value of previous years can also be used.
framework However, in this paper, due to the lack of data we use the PM density
of a year before the year under study.
The rest of this paper is organized as follows. In Section 2, the pro­
posed RCM framework is presented. The optimal maintenance strategy To determine the significance of each criterion in identifying the
assignment formulation is detailed in Section 3. Section 4 is devoted to most critical feeders, it is necessary to assign a weight to each of the
case studies, and finally, Section 5 presents the conclusions of the paper. eight criteria. For this purpose, the BWM method is employed, which is
known as one of the effective MCDM techniques [23]. Compared to
2. Proposed RCM framework other weighting methods, such as analytical hierarchical process, this
method requires fewer pairwise comparisons, and its output weights
This section presents the proposed three-stage RCM framework, usually have a good consistency ratio. The consistency ratio is used as an
which includes critical feeder identification, FMs prioritization, and indicator to evaluate the reliability of the obtained weights. If the con­
optimal maintenance strategy assignment. sistency ratio exceeds the specified value, pairwise comparisons should
be revised [23]. The BWM method is outlined in detail in [23]. In some
2.1. The first stage: critical feeder identification cases, the problem may have multiple optimal solutions. To address this
issue, a linear version of the BWM method was proposed in [24], which
Critical feeder component identification is the first stage of RCM is used in this paper.
planning. This stage of the proposed method seeks to rank feeders based After determining the weights of the eight criteria using the BWM
on their criticality and need for inspection. A multi-criteria decision- method, the feeders are ranked using the TOPSIS method. TOPSIS is a
making (MCDM) model, in line with [14], which uses the BWM and widely used MCDM technique that finds applications in various fields
TOPSIS methods, is utilized to identify critical feeders and rank them. [25]. This method employs the criteria weights obtained through the
The feeder criticality is evaluated based on eight criteria as follows: BWM method to rank the alternatives. Comprehensive explanations of
this approach are presented in [26].
1) Feeder load value, which is calculated based on the percentage of Therefore, the TOPSIS method assigns a score, called Si, to each
each type of customer (residential, commercial, etc.) and their VoLL. feeder and ranks them. A lower value of Si indicates a higher level of
2) Feeder equipment value, which is calculated based on the feeder’s feeder criticality and a higher rank on the critical feeders’ list. Hence,
equipment and its worth. the feeders ranking is obtained by sorting the feeders from the lower to
3) Outage rate, which is equal to the ratio of the feeder’s ENS to the sum higher Si values. Feeders with higher rankings should be given priority
of the total supplied energy and ENS. for inspection. Furthermore, in the second stage of the three-stage RCM
4) The number of outages (permanent or momentary). It should be framework, feeders’ ranking will be used to prioritize FMs.
noted that some outages are a result of natural causes or factors
beyond the control of operators, e.g., floods, bad weather, and the 2.2. The second stage: FMs prioritization
upstream grid’s unscheduled outages. This exclusion is done so that
critical feeders are identified with outages that can be prevented by FMs prioritization is the second stage of the RCM framework. This
maintenance. Outages caused by these roots are excluded from the prioritization is crucial in determining the level of risk that the FMs
outages list when calculating the outage rate and the number of impose on the system. To identify FMs with higher risk, a failure mode
outages. and effect analysis (FMEA) using MCDM is implemented [27]. In this
5) Climatic and pollution score, which represents the impact of weather paper, the BWM and TOPSIS methods are employed to calculate the
conditions and industrial pollution on feeders. This score is given by score of each FM and rank them based on specific factors. The FMs are
the distribution system operator on a scale of 1 to 5, where 1 prioritized using four factors as follows:
correspond to clean air with no pollutant and 5 represents feeders
with severely contaminated air.

3
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

1) Severity: Severity refers to the impact of the FMs on the system when ENS of cold-line repairs during the scheduled power outage imposed by
it fails. Three levels of severity are considered, (1) failures with high the distribution company, EENS, and the expected future CM costs. The
severity, (2) failures with moderate severity, and (3) failures with maintenance budget, power outage duration limitations, the sum of the
low severity. The level of severity is assigned to the FM using the duration of the upstream grid’s scheduled outages, and the number of
expertise of the distribution utility’s maintenance expert. maintenance crews are the constraints of the maintenance planning.
2) Occurrence: The FMs’ frequency of occurrence is extracted from the Moreover, the FMs scores obtained in the second stage are incorporated
distribution system inspection data. into the failure rate calculation. Details of the problem formulation of
3) Detection: The detection of FMs is rated on a scale of 1 to 5. A grade the optimal maintenance strategy assignment will be presented in Sec­
of 1 means that the FM is easily detectable when the PM crew tries to tion 3.
find it. A grade of 5, on the other hand, indicates that the FM is In conclusion, the flowchart of the proposed three-stage RCM
difficult to detect and requires special tools such as thermographic framework for distribution system maintenance planning is illustrated in
cameras. Fig. 1.
4) Feeder rank: The feeder rank is obtained in the first stage of the
proposed RCM framework. 3. Optimal maintenance strategy assignment formulation

To determine the importance of each factor for FMs prioritization, In this section, the mathematical model of the optimal maintenance
the BWM is utilized, which assigns a weight to each of the four factors. strategy assignment problem is proposed. As mentioned, three mainte­
Afterward, TOPSIS is applied to prioritize FMs. TOPSIS implementation nance strategies, consisting of cold-line (C), hot-line (H), and RTF are
results in assigning a score to each FM. The FM with a lower score has a considered. Additionally, since it is possible to carry out cold-line repairs
higher level of priority for maintenance. The FMs scores will be during the upstream grid’s outages, the cold-line repair strategy is
employed in the third stage of the RCM framework to assign the optimal divided into two categories: cold-line repair overlapping with the up­
maintenance strategy. stream grid’s scheduled outages (C1) and cold-line repair with a planned
power outage in the distribution system (C2). Therefore, four repair
strategies are considered to apply to each FM in each feeder. The
2.3. The third stage: optimal maintenance strategy assignment
problem is formulated as an MILP model, which is presented in the
following.
Optimal maintenance strategy assignment is the last stage of the
proposed three-stage RCM framework. In this paper, three maintenance
strategies, cold-line, hot-line, and RTF are considered and every FM is 3.1. The objective function
assigned an optimal maintenance strategy. The cold-line repair opera­
tions are divided into cold-line repairs with a planned power outage and The goal of the strategy assignment is to minimize the costs associ­
cold-line repairs overlapping with the upstream grid’s scheduled out­ ated with maintenance, which comprise the cost of repair operations,
ages. Optimal maintenance strategy assignment is formulated as an equipment, and ENS during cold-line repairs with a planned power
MLIP problem, where the aim is to minimize the objective function. The outage, EENS costs, and expected future CM costs. The objective func­
objective function includes the cost of repair operations, equipment, tion is defined as follows:

Fig. 1. Proposed three-stage RCM framework.

4
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

minOF = respectively. λf is the failure rate. TRf and TIf are the duration of the
∑ ( )
BC1 OC C
C Eq C Fault
+ ECfC cold-line repair and the duration of detecting and isolating the fault,
f f + f + EENSf V
f ∈ΩF respectively. LMf and Lf
Fault
are load level that experiences a power outage
∑ ( ) during the planned repair and the fault before detecting and isolating
+ BC2 OCfC + CfEq + ENSCf V M + EENSfC V Fault + ECfC
f
the fault, respectively.
f ∈ΩF (1)
∑ ( ) For each FM, if it is repaired using the cold-line or hot-line strategy,
+ BHf OCfH + CfEq + EENSHf V Fault + ECfH its failure rate will be reduced to the minimum value of its lifetime.
f ∈ΩF However, if no repair is carried out and the RTF strategy is chosen, the
∑ ( )
+ BRTF EENSfRTF V Fault + ECfRTF failure rate of the equipment will increase in proportion to the FM score
obtained at the second stage of the RCM framework. The score assigned
f
f ∈ΩF
to each FM, which is between 0 and 1, indicates its criticality, with a
ΩF and f are set and indices of FMs, respectively. OCCf and OCHf are lower score indicating a higher priority for maintenance. Eq. (6) de­
cold-line and hot-line operation costs, respectively. V M and V Fault are the scribes the effect of repair strategies on failure rates. As shown by this
VoLL during planned maintenance and equipment faults, respectively. equation, the failure rate of the components of the FMs with higher
CEq
f is equipment cost and ENSf is ENS of cold-line repair with a planned
C maintenance priority increases in proportion to their FM prioritization
power outage. ECCf , ECHf , and ECRTF are expected future corrective score.
f
( )
maintenance costs for selecting the cold-line, hot-line, and RTF strate­ λf = λmin BC1 C2 H
+ λmin RTF
[ ( )]
∀f ∈ ΩF (6)
f + Bf + Bf f Bf 1 + 1 − Scoref
gies, respectively. EENSCf , EENSHf , and EENSRTF are EENS associated with
f
f
selecting the cold-line, hot-line, and RTF strategies, respectively.
λmin
f is the minimum value of failure rate if the equipment is repaired and
The objective function consists of the sum of four equations, each
Scoref is FM score, obtained at the second stage of the RCM framework
defining the maintenance costs associated with a repair strategy. The
using the TOPSIS method.
first equation corresponds to the cold-line repair strategy overlapping
with the upstream grid’s scheduled outages (C1), and BC1 f is its binary
variable. As demonstrated by this equation, the company does not incur 3.2. Constraints
any ENS cost due to the execution of cold-line repairs during the up­
stream grid’s outages. The second equation corresponds to the cold-line The constraints of the proposed maintenance strategy assignment are
presented as follows:
repair strategy with a planned power outage (C2), where BC2f is the bi­
nary variable associated with it. This strategy includes all costs associ­
3.2.1. Maintenance budget constraints
ated with the first strategy (C1) and also considers the ENS cost resulting
The maintenance budget for both the cold-line and hot-line opera­
from the repair. The third equation takes into account the cost of
tions is limited and determined by the distribution company’s policies.
implementing the hot-line repair strategy (H), which does not involve
The maintenance budget limitations for the cold-line and hot-line repair
ENS cost for repair operations as it does not require a power outage, and
operations are expressed in constraints (7) and (8), respectively.
its binary variable is denoted by BHf . Finally, the fourth equation shows ∑( )
the cost of selecting the RTF strategy, with BRTF
f as its corresponding BC1
f + Bf
C2
OCfC ≤ BudjetC (7)
binary variable. In this strategy, no repairs are made, and the equipment f ∈ΩF

is allowed to function until it completely fails and stops working, after ∑

which, it is either repaired or replaced. Although there is no cost of BHf OCfH ≤ BudjetH (8)
f ∈ΩF
repair operations, equipment, and ENS resulting from the PM, this
approach leads to relatively high EENS and expected future CM costs. BudjetC and Budjet H are budgets for cold-line and hot-line repair
It is necessary to set only one binary variable associated with each operations, respectively.
FM equal to one. Hence, only one repair strategy should be selected for
each FM. This constraint is expressed in (2). 3.2.2. Power outage duration constraints
The duration of both the upstream grid’s scheduled outages and the
BC1 + BC2 + BHf + BRTF = 1 ∀f ∈ Ω F
(2)
f f f
maximum duration of planned power outages to perform cold-line re­
The ENS associated with the cold-line repair depends on two factors: pairs is limited. The constraint on the duration of the upstream grid’s
the duration of the repair and the load level during the planned power outage is indicated in (9), and the maximum allowable time for the
outage. Similar to the RTF’s future costs, other strategies (C1, C2, and H) planned power outage is defined in (10).
contain terms that describe the future costs of CM if the component fails. ∑
BC1 max
f TRf ≤ TOverlap (9)
However, due to the conducted PM, future costs of these strategies are
f ∈ΩF
substantially lower than those of the RTF. The EENS and expected future
CM costs are determined by considering the probability of equipment ∑
BC2 max
f TRf ≤ TPlanned (10)
failure in the future, which is dependent on their failure rate and how it f ∈ΩF
is affected by different maintenance strategies. The ENS of the cold-line
repair, EENS, and expected future CM costs of different strategies are TOverlap
max
and TPlanned
max
are the maximum duration of the upstream grid’s
defined in (3)-(5), respectively. scheduled outages and planned power outages, respectively.

ENSfC2 = TRf LfM ∀f ∈ ΩF (3) 3.2.3. Maintenance crews constraints

( ) The availability of multiple maintenance crews can decrease the
EENSfm = λf TRf LM Fault
f + TIf Lf ∀f ∈ ΩF , ∀m ∈ ΩM (4) required duration of cold-line repairs as simultaneous PMs can be con­
ducted when the planned outage is in effect. As a result, having several
( )
maintenance crews for performing simultaneous cold-line repairs re­
ECfm = λf OCfC + CfEq ∀f ∈ ΩF , ∀m ∈ ΩM (5)
duces both the power outage duration and the resulting ENS, in pro­
ΩM and m are set and indices of repair strategies (C1, C2, H, RTF), portion to the number of maintenance crews. Therefore, to model this
relationship, constraints (11) and (12) are defined in place of (9) and

5
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

(10), respectively. In addition, the impact of having multiple mainte­ Table 3

nance crews on the ENS of cold-line repairs is illustrated in constraint Top 10 critical feeders.
(13), which is used instead of Eq. (3). Rank Feeder Score Rank Feeder Score
/
∑ 1 F-149 0.1978618 6 F-115 0.5819017
Bf TRf K C ≤ TOverlap
C1 max
(11) 2 F-147 0.2329128 7 F-179 0.5858127
f ∈ΩF 3 F-238 0.5491322 8 F-270 0.5921164
4 F-266 0.5699045 9 F-366 0.5941110
/ 5 F-236 0.5744372 10 F-301 0.5946589
∑
BC2
f TRf K C ≤ TPlanned
max
(12)
f ∈ΩF

/ Table 4
ENSfC2 = TRf LfM K C ∀f ∈ ΩF (13) Weights of the four factors for FMs prioritization.
Factor Severity Occurrence Detection Feeder rank
KC is the number of the cold-line maintenance crew.
Weight 0.1995 0.2257 0.3454 0.2294

4. Case studies and results

of implementing the proposed method on the real distribution system
In this section, the proposed RCM framework is applied to a real are presented in detail, and the effectiveness of this method is analyzed
distribution system in Iran, which consists of 428 medium-voltage through different case studies. Furthermore, a comparative analysis is
feeders. The obtained data are from Mar. 2021 to Mar. 2022. The re­ undertaken, wherein the proposed method is compared with another
sults of the critical feeders ranking, FMs scores, and optimal mainte­ existing state-of-the-art approach. In this regard, the SECA approach is
nance strategy assignment are presented in the following. employed instead of BWM and TOPSIS methods within the three-stage
RCM framework.
4.1. Critical feeder identification results
4.3.1. Application to a simple test system
To identify critical feeders, first, the BWM method is employed and In this section, the proposed method for optimal maintenance
the resulting weights are presented in Table 2. The results indicate that strategy assignment is applied to the RBTS bus 2 distribution system.
among the eight criteria, the outage rate, feeder load value, and the This system comprises 22 load points, and its single-line diagram is
number of outages are the most important factors. The weights for these illustrated in Fig. 2. Comprehensive details about the RBTS bus 2 dis­
criteria are determined using input from 23 experts of the distribution tribution system are presented in [29]. The scenario assumes the exis­
company, which are provided in [28]. Experts’ opinions are based on tence of 14 failure modes across 14 feeder sections, as depicted in Fig. 2.
their practical experiences in addition to the utilities’ policies from Four cold-line maintenance crews, a two-hour upstream grid’s sched­
technical and economic aspects. uled outage, and a maximum four hours planned outages are considered.
After determining the weights of the eight criteria, the feeders are The proposed method is applied to this system, and the outcomes of the
ranked using TOPSIS. Table 3 shows the top 10 critical feeders. The selected strategy are presented in Table 6.
results show that feeder F-149 has the lowest score and should be given The results indicate that FMs close to the main grid are repaired
high priority for inspection. This feeder’s consumed annual load is either in a cold-line or hot-line manner. Specifically, four FMs in feeder
relatively low, and because of an outage that took several hours to sections 1, 2, 5, and 6 are repaired by cold-line strategy overlapping with
repair, the outage rate of this feeder from Mar. 2021 to Mar. 2022 the upstream grid’s scheduled outages. These FMs are at critical points
increased. As a result, given the outage rate’s weight and its effect on the where an occurrence of a fault would result in a significant ENS. The FM
critical feeders ranking, F-149 becomes the most critical feeder. in the feeder section 7 is addressed through a cold-line repair with a
planned power outage, while FMs 3, 8, 9, 11, and 12 are repaired in a
4.2. FMs prioritization results hot-line manner, since in the case of cold-line maintenance, loads with
high demand will lose power, which leads to a high ENS cost. Moreover,
To prioritize the FMs, first, considering the opinions of 8 experts of the RTF strategy is chosen for FMs at the end of feeders (4, 10, 13, and
the distribution company [28], the weights of the four factors using the 14), as these points are located at the end of the feeder and would not
BWM method are determined, which are presented in Table 4. The re­ cause a substantial ENS in case of a fault.
sults show that FM detection is the most important factor, while the
severity of the FM is the least important one. 4.3.2. Evaluation on the real distribution system
Based on the weight of each factor and using the TOPSIS method, the Based on the resulting scores from the FMs prioritization stage, the
FMs prioritization is carried out and the list of the top 20 FMs along with most critical FM and 4 out of the top 20 FMs belong to feeder F-006,
their scores is presented in Table 5. indicating a high priority for PM. Therefore, feeder F-006 is selected as a
candidate for the optimal maintenance strategy assignment problem.
The single-line diagram of feeder F-006 is depicted in Fig. 3. This feeder
4.3. Optimal maintenance strategy assignment results contains 150 distribution transformers, a circuit breaker, two dis­
connector switches, and four maneuver switches. In addition, the feeder
In this section, first, the proposed method for optimal maintenance has 65 m underground line and 72.142 km overhead line. The average
strategy assignment is applied to a simple test system. Then, the results load of this feeder is 5.9 MW. Besides, the feeder has 1110 customers.
The VoLLs during planned maintenance (VM), the VoLLs during equip­
Table 2 ment fault (VFault) [30], and the percentage of each type of customer in
Weights of the eight criteria for critical feeder identification. feeder F-006 are presented in Table 7. There is a total of 376 FMs
Criterion Weight Criterion Weight associated with this feeder from Mar. 2021 to Mar. 2022. The allocated
Feeder load value 0.1664 Climatic and pollution score 0.1189 budget for the cold-line and the hot-line repair operations of this feeder
Feeder equipment value 0.0915 Maneuver switch density 0.0550 is $680 each. Additionally, 10 cold-line maintenance crews are available
Outage rate 0.2580 Switch density 0.0540 to perform simultaneous cold-line repair operations on feeder F-006.
The number of outages 0.1640 PM density 0.0922
Furthermore, there is a 5 h upstream grid’s scheduled outage that can be

6
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

Table 5
Top 20 FMs with their scores.
Rank Feeder FM code Score Rank Feeder FM code Score

1 F-006 555 0.0014762 11 F-370 539 0.6706050

2 F-394 539 0.3313104 12 F-006 554 0.6710808
3 F-115 539 0.5739883 13 F-137 552 0.6730863
4 F-115 537 0.5799351 14 F-144 38,302 0.6750676
5 F-006 535 0.5805694 15 F-144 38,401 0.6750810
6 F-147 510 0.6177820 16 F-155 552 0.6787414
7 F-175 539 0.6267414 17 F-270 539 0.6810185
8 F-155 539 0.6634223 18 F-155 554 0.6814768
9 F-144 552 0.6656644 19 F-370 537 0.6814896
10 F-144 539 0.6675082 20 F-006 548 0.6828955

Fig. 2. The single line diagram of the RBTS bus 2 distribution system.

Table 6
The selected strategy for FMs in the RBTS bus 2 distribution system.

• Case 0: This case is the business-as-usual maintenance planning that

is implemented by the distribution system company. This approach
works by assigning the most critical FMs to the hot-line strategy until
the budget limit is met. After that, cold-line repairs are planned in the
same way, until the costs reach the budget. The rest of the FMs are
assigned as RTF.
• Case 1: This case represents the proposed method, where four repair
strategies are considered: cold-line repair overlapping with the up­
stream grid’s scheduled outage (C1), cold-line repair with a planned
power outage (C2), hot-line repair (H), and RTF strategy (RTF). Be­
sides, 10 cold-line maintenance crews are available.
• Case 2: This case is similar to Case 1, without considering the cold-
line repair during the upstream grid’s outage.
• Case 3: This case is similar to Case 1, except that only one cold-line
maintenance crew is available.
• Case 4: In this case, the RTF strategy for all FMs is considered. In
other words, no repairs are performed until the equipment fails and
stops working.
Fig. 3. The single line diagram of Feeder F-006.
• Case 5: This case is similar to Case 1, except that the SECA approach
instead of BWM and TOPSIS methods is utilized.
used for cold-line PMs. The maximum sum of allowed planned power
outages is 10 h.
In Tables 8 and 9, by comparing the outcomes of the proposed
To evaluate the effectiveness of the proposed optimal maintenance
approach in Case 1 with the conventional maintenance plan executed by
strategy assignment method, six case studies are defined as follows:
the company (Case 0), it is illustrated that the proposed optimization-

7
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

Table 7
VoLLs and percentage of each type of customer in feeder F-006.
Type of customer Residential Commercial Industrial Agricultural Lighting
M
V ($/kWh) 0.2 0.6 4 0.6 0.6
VFault ($/kWh) 0.6 1.8 12 1.8 1.8
Percentage of customers (%) 14 3 9 70 4

Table 8
Results of selected strategy, ENS, and EENS of Case 0 and Case 1.
Case 0 Case 1
Repair strategy C1 C2 H RTF C1 C2 H RTF

Number of repairs 0 50 182 144 25 10 155 186

Power outage (h) 0 10 0 0 5 2 – –
ENS (MWh) – 3.5284 – – – 0.0091 – –
EENS (MWh) – 0.5223 0.7666 2.1644 0.6886 0.0234 1.0187 1.5392
Total EENS (MWh) 3.4533 3.2699

Table 9
Costs of Case 0 and Case 1 ($).
Case 0 Case 1
Repair strategy C1 C2 H RTF C1 C2 H RTF

Repair operation – 79.9913 679.3193 – 42.8379 3.7114 549.8578 –

Equipment – 197.3087 133.2022 – 98.6723 7.3188 100.2676 –
ENS – 2822.7064 – – – 7.2521 – –
EENS – 1253.4385 1839.8644 5194.6718 1652.6684 56.2119 2444.7891 3694.0263
Expected future CM – 0.8874 0.6223 9.7169 1.9161 0.0342 0.5204 8.4564
Objective function 12,211.7292 8668.5407

based method yields a 29 % reduction in the objective function. Addi­ total number of cold-line repairs decreases from 35 to 4 as fewer crews
tionally, the EENS associated with the proposed method is lower by 5.3 are available at a given time. In addition, the total cost of Case 3 in­
% compared to the business-as-usual approach. The results show that in creases by $160.71 (1.8 %) compared to Case 1, because of the higher
Case 1, the entire duration of the upstream grid’s outage is used to ENS incurred during planned power outages. Furthermore, by
perform the cold-line repair, where 25 FMs are repaired in 5 h. However, comparing the outage time of Cases 1 and 3, it is observed that in Case 3,
no ENS cost is incurred for these 25 cold-line repairs. Additionally, 10 due to the shortage of cold-line maintenance crews, only 4 h of upstream
cold-line repairs are carried out by implementing a 2 h power outage. grid’s outages are used and only 2 FMs are repaired during this time,
The results indicate that the RTF strategy is the most commonly assigned while in Case 1, with the use of 10 maintenance crew, 25 FMs are
strategy to the FMs. While this strategy greatly increases EENS, it re­ repaired within the 5 h upstream grid’s outage. In Fig. 4, the effect of the
duces total maintenance costs by not performing any repairs until the number of cold-line maintenance crews on the objective function is
equipment fails. Following the RTF strategy, the hot-line repair strategy analyzed. The results illustrate a consistent decrease in the objective
is the second most frequently used strategy, where 155 FMs are repaired function as the number of maintenance crews increases. Therefore, the
using this strategy. Although the hot-line repair strategy has higher presence of multiple maintenance crews is beneficial for maintenance
operating costs compared to the cold-line repair, it does not incur ENS planning, as it allows for more efficient use of the available outage time
costs since no power outage is needed. and the repair of a higher number of FMs.
As results illustrated in Tables 10 and 11, in Case 2 which does not In Case 4, the RTF strategy is applied to all FMs. In this case, EENS
use the upstream grid’s outage (C1 strategy), selecting the RTF and hot- and objective function are 4.4379 MWh and $10662.5506, respectively.
line strategies happens for a higher number of FMs. Comparing the Although the cost of repair operations and equipment is not included in
objective function of Case 1 and Case 2 indicates that using the upstream Case 4, the EENS is significantly high. Comparing the results of Cases 1
grid’s outage in Case 1 reduces the total cost by $178.44 (2 %). On the and 4, it becomes evident that the objective function in Case 4 increases
other hand, by using the RTF strategy for more and more FMs, the total by $1994.01 compared to Case 1. Furthermore, the results show that by
EENS in Case 2 increases by 0.071 MWh compared to Case 1. Besides, implementing the proposed method for optimal maintenance strategy
the impact of the presence of multiple maintenance crews is demon­ assignment in Case 1, which leads to a reduction in the equipment
strated by comparing the results of Case 1 and Case 3. The results show failure rate, EENS is significantly lower compared to choosing the RTF
that by reducing the number of maintenance crews to one in Case 3, the strategy in Case 4. Specifically, the total EENS of Case 1 is approximately

Table 10
Results of selected strategy, ENS, and EENS of Case 2 and Case 3.
Case 2 Case 3
Repair strategy C1 C2 H RTF C1 C2 H RTF

Number of repairs – 10 164 202 2 0 164 210

Power outage (h) – 2 – – 4 – – –
ENS (MWh) – 0.0091 – – – – – –
EENS (MWh) – 0.0234 1.5341 1.7834 0.4109 – 1.1280 1.8137
Total EENS (MWh) 3.3409 3.3526

8
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

Table 11
Costs of Case 2 and Case 3 ($).
Case 2 Case 3
Repair strategy C1 C2 H RTF C1 C2 H RTF

Repair operation – 3.7114 659.7346 – 2.6601 – 628.0555 –

Equipment – 7.3188 139.7472 – 11.8565 – 129.3545 –
ENS – 7.2521 – – – – – –
EENS – 56.2119 3681.7456 4280.2416 986.2273 – 2707.1249 4352.9441
Expected future CM – 0.0342 2.1543 8.8265 1.5097 – 0.6514 8.8624
Objective function ($) 8846.9783 8829.2466

Table 12
Results of Case 5.
Repair strategy C1 C2 H RTF

Number of repairs 25 9 131 211

Power outage (h) 5 1.8 – –
ENS (MWh) – 0.0067 – –
EENS (MWh) 0.6842 0.0184 1.0272 1.6793
Total EENS (MWh) 3.4092
Costs ($)
Repair operation 39.4561 3.3403 479.5856 –
Equipment 102.6185 6.5869 117.1984 –
ENS – 5.3689 – –
EENS 1642.1762 44.2427 2465.2875 4030.3463
Expected future CM 1.9179 0.0308 0.5962 9.0244
Objective function ($) 8846.9783

optimal maintenance strategy assignment was formulated as an MILP

model to minimize the maintenance costs. In this problem, the upstream
Fig. 4. Effect of the number of cold-line maintenance crews on the objec­
grid’s scheduled outages were utilized to implement cold-line repairs,
tive function. which does not incur any ENS costs. The most important outcomes of
case studies are as follows:
26.32 % lower than that of Case 4. The results indicate the superiority of
• The proposed three-stage RCM framework using BWM and TOPSIS
the proposed method in improving the distribution system reliability by
approaches outperforms a recently published state-of-the-art
applying the optimal maintenance strategy while satisfying all the
method, the SECA approach, and leads to a 2 % and 4 % reduction
financial and operational constraints.
in the objective function and EENS, respectively.
To compare the proposed method in this paper with other ap­
• Utilizing the upstream grid’s outages for cold-line repairs can lead to
proaches, Case 5 employs the SECA method, as presented in [15]. SECA
a 2 % reduction in the objective function.
is one of the effective MCDM methods that does not require expert
• The proposed method results in a 29 % decrease in the objective
opinions and simultaneous evaluation of criteria and alternatives is
function and a 5.3 % reduction in the EENS compared to the
performed [31]. In this method, a multi-objective non-linear program­
business-as-usual approach.
ming model is introduced, aiming to maximize the overall performance
• The availability of multiple maintenance crews for simultaneous
of alternatives by leveraging variation information within and between
cold-line repairs demonstrated that presence of 10 maintenance
criteria in the decision matrix [31]. The overall performance scores of
crews instead of one crew can result in a 1.8 % decrement in the
alternatives and the objective weights of criteria are determined
objective function.
simultaneously by solving the multi-objective model. The important
aspect of this method is its independence from the opinions of experts, as
CRediT authorship contribution statement
it can find the weights on its own. The details of the SECA method and its
mathematical formulations are presented in [31].
Sajjad Moradi-Sarvestani: Writing – original draft, Methodology,
The results of the optimal maintenance strategy assignment in Case 5
Formal analysis, Conceptualization. Mohammad Reza Dehbozorgi:
are detailed in Table 12. A comparative analysis with the outcomes of
Writing – original draft, Validation, Methodology, Investigation,
Case 1 reveals a 2 % increase in the objective function for Case 5.
Conceptualization. Mohammad Rastegar: Writing – review & editing,
Furthermore, the results indicate an elevated preference for the RTF
Validation, Supervision, Conceptualization.
strategy in Case 5 compared to Case 1. As a result, the EENS in Case 5 is
0.14 MWh (4 %) higher than in Case 1. In summary, it is demonstrated
Declaration of competing interest
that the proposed method in this paper (Case 1) that considers experts’
opinion provides a solution with lower cost and EENS compared to the
The authors declare that they have no known competing financial
approach in Case 5 that ignores experts’ opinion.
interests or personal relationships that could have appeared to influence
the work reported in this paper.
5. Conclusions
Data availability
In this paper, a three-stage RCM framework for critical feeder iden­
tification, FMs prioritization, and optimal maintenance strategy
The data that has been used is confidential.
assignment was proposed. The BWM and TOPSIS methods were
employed to identify critical feeders and prioritize FMs. The FMs scores
were utilized to calculate the failure rate in the third stage, where the

9
S. Moradi-Sarvestani et al. Electric Power Systems Research 230 (2024) 110215

References [15] S. Bahrami, M. Rastegar, Security-based critical power distribution feeder

identification: application of fuzzy BWM-VIKOR and SECA, Int. J. Electr. Power
Energy Syst. 134 (2022) 107395.
[1] S. Kumar, R. Saket, D.K. Dheer, J.B. Holm-Nielsen, P. Sanjeevikumar, Reliability
[16] D. Piasson, A.A. Bíscaro, F.B. Leão, J.R.S. Mantovani, A new approach for
enhancement of electrical power system including impacts of renewable energy
reliability-centered maintenance programs in electric power distribution systems
sources: a comprehensive review, IET Gener. Transm. Distrib. 14 (10) (2020)
based on a multiobjective genetic algorithm, Electr. Power Syst. Res. 137 (2016)
1799–1815.
41–50.
[2] S. Wang, Y. Fan, S. Jin, P. Takyi-Aninakwa, C. Fernandez, Improved anti-noise
[17] H. Mirsaeedi, A. Fereidunian, S.M. Mohammadi-Hosseininejad, H. Lesani,
adaptive long short-term memory neural network modeling for the robust
Electricity distribution system maintenance budgeting: a reliability-centered
remaining useful life prediction of lithium-ion batteries, Reliab. Eng. Syst. Saf. 230
approach, IEEE Trans. Power Deliv. 33 (4) (2017) 1599–1610.
(2023) 108920.
[18] F. Keynia, M. Mirhosseini, A. Heydari, A. Fekih, A budget allocation and
[3] M. Mirhosseini, F. Keynia, Asset management and maintenance programming for
programming-based RCM approach to improve the reliability of power distribution
power distribution systems: a review, IET Gener. Transm. Distrib. 15 (16) (2021)
networks, Energy Rep. 8 (2022) 5591–5602.
2287–2297.
[19] S.A.A. Matin, S.A. Mansouri, M. Bayat, A.R. Jordehi, P. Radmehr, A multi-objective
[4] S.R. Khuntia, J.L. Rueda, S. Bouwman, M.A. van der Meijden, A literature survey
bi-level optimization framework for dynamic maintenance planning of active
on asset management in electrical power [transmission and distribution] system,
distribution networks in the presence of energy storage systems, J. Energy Storage
Int. Trans. Electr. Energy Syst. 26 (10) (2016) 2123–2133.
52 (2022) 104762.
[5] S. Wang, F. Wu, P. Takyi-Aninakwa, C. Fernandez, D.I. Stroe, Q. Huang, Improved
[20] M.A. Moghadam, S. Bagheri, A.H. Salemi, M.B. Tavakoli, Long-term maintenance
singular filtering-Gaussian process regression-long short-term memory model for
planning of medium voltage overhead lines considering the uncertainties and
whole-life-cycle remaining capacity estimation of lithium-ion batteries adaptive to
reasons for interruption in a real distribution network, Reliab. Eng. Syst. Saf.
fast aging and multi-current variations, Energy 284 (2023) 128677.
(2023) 109089.
[6] L. Bertling, R. Allan, R. Eriksson, A reliability-centered asset maintenance method
[21] Y. Wang, R. He, Z. Tian, Opportunistic condition-based maintenance optimization
for assessing the impact of maintenance in power distribution systems, IEEE Trans.
for electrical distribution systems, Reliab. Eng. Syst. Saf. 236 (2023) 109261.
Power Syst. 20 (1) (2005) 75–82.
[22] M. Enjavimadar, M. Rastegar, Optimal reliability-centered maintenance strategy
[7] M. Alimohammadi, J. Behnamian, Preventive maintenance scheduling of
based on the failure modes and effect analysis in power distribution systems,
electricity distribution network feeders to reduce undistributed energy: a case
Electr. Power Syst. Res. 203 (2022) 107647.
study in Iran, Electr. Power Syst. Res. 201 (2021) 107509.
[23] J. Rezaei, Best-worst multi-criteria decision-making method, Omega 53 (2015)
[8] M.I.H. Tusar, B.R. Sarker, Maintenance cost minimization models for offshore wind
49–57.
farms: a systematic and critical review, Int. J. Energy Res. 46 (4) (2022)
[24] J. Rezaei, Best-worst multi-criteria decision-making method: some properties and a
3739–3765.
linear model, Omega 64 (2016) 126–130.
[9] L.A. Neves, D.M. Frangopol, A. Petcherdchoo, Probabilistic lifetime-oriented
[25] M. Behzadian, S.K. Otaghsara, M. Yazdani, J. Ignatius, A state-of the-art survey of
multiobjective optimization of bridge maintenance: combination of maintenance
TOPSIS applications, Expert. Syst. Appl. 39 (17) (2012) 13051–13069.
types, J. Struct. Eng. 132 (11) (2006) 1821–1834.
[26] G.H. Tzeng, J.J. Huang, Multiple Attribute Decision Making: Methods and
[10] M. Ben-Daya, You may need RCM to enhance TPM implementation, J. Qual. Maint.
Applications, CRC Press, 2011.
Eng. 6 (2) (2000) 82–85.
[27] H.W. Lo, J.J. Liou, A novel multiple-criteria decision-making-based FMEA model
[11] J. Geisbush, S.T. Ariaratnam, Reliability centered maintenance (RCM): literature
for risk assessment, Appl. Soft Comput. 73 (2018) 684–696.
review of current industry state of practice, J. Qual. Maint. Eng. 29 (2) (2023)
[28] M.R. Dehbozorgi. The opinions of experts on the criteria of feeders’ criticality level
313–337.
and failure modes prioritization 2023 [Online] Available: https://github.com/re
[12] P. Dehghanian, M. Fotuhi-Firuzabad, S. Bagheri-Shouraki, A.A.R. Kazemi, Critical
zad77/RCM.
component identification in reliability centered asset management of power
[29] R.N. Allan, R. Billinton, I. Sjarief, L. Goel, K. So, A reliability test system for
distribution systems via fuzzy AHP, IEEE Syst. J. 6 (4) (2011) 593–602.
educational purposes-basic distribution system data and results, IEEE Trans. Power
[13] R. Ghorani, M. Fotuhi-Firuzabad, P. Dehghanian, W. Li, Identifying critical
Syst. 6 (2) (1991) 813–820.
components for reliability centred maintenance management of deregulated power
[30] M.J. Sullivan, D. Keane, Outage Cost Estimation Guidebook, Electric Power
systems, IET Gener. Transm. Distrib. 9 (9) (2015) 828–837.
Research Institute (EPRI), Palo Alto, 1995. CA (United States).
[14] S. Bahrami, M. Rastegar, P. Dehghanian, An fbwm-topsis approach to identify
[31] M. Keshavarz-Ghorabaee, M. Amiri, E.K. Zavadskas, Z. Turskis, J. Antucheviciene,
critical feeders for reliability centered maintenance in power distribution systems,
Simultaneous evaluation of criteria and alternatives (SECA) for multi-criteria
IEEE Syst. J. 15 (3) (2020) 3893–3901.
decision-making, Informatica 29 (2) (2018) 265–280.

10