An Approach of The Analytic Hierarchy Process

to Acquire The Weighting Factor of High Voltage
Circuit Breaker Health Index
Achmad Setiawan Wind Adiati Rahman Azis Prasojo
PT PLN (Persero)Headquarter PT PLN (Persero) Department of Electrical Engineering
Jakarta, Indonesia East Java & Bali Transmission Central Unit Politeknik Negeri Malang
[email protected] Surabaya, Indonesia Malang, Indonesia
[email protected] [email protected]

Abstract— The circuit breaker (CB) health condition greatly weighting factor since it is a data-intensive activity, in which
affects the reliability of CB, which eventually will affect the data uncertainty is possible to occur that finally can reduce the
reliability of the power system. A method that can be used to degree of confidence of the HI overall score. Then, the
determine CB health conditions based on multi criteria is the Analytic Hierarchy Process (AHP) method transforms the
calculation the health index. The critical issue in developing expert judgment into a number for every criterion that
health index for a circuit breaker is the weighting factor. There composes the HI, and these number is defined as the
are many techniques to formulate the weighting factor, this Weighting Factor (WF).
paper proposes a new model of the circuit breaker health index
weighting factor determination by utilizing the Analytical The application of AHP to determine CB HI has been
Hierarchy Process (AHP). The weighting factor of the proposed widely researched in literature. Depending on the type of
circuit breaker health index model is generated using the AHP circuit breaker voltage operation, the utilization of the AHP is
based on experts' interviews and surveys. The result presents concentrated on the determination of HI for an old high voltage
that the condition assessment factor is judged to be more circuit breaker [3]. The combination of the fuzzy set theory,
important than the profile assessment factor to reflect the circuit the analytical hierarchy process, and evidential reasoning were
breaker health index. The contact resistance is considered as the proposed [4] to obtain a circuit breaker condition assessment
most significant compared to the insulation quality, leakage model. Reference [5] used the Principal Component Analysis
current, Open Close Timing on CB, and insulation resistance
(PCA) and AHP to create ranking in-service transformer based
parameters. The AHP approach has proven to be able to capture
on their insulation condition data. It was demonstrated that
both subjective and objective evaluation measures of the circuit
breaker health index weighting factor.
PCA and AHP could be an effective supplement or
replacement to a track-record proven empirical formula
Keywords—Analytic Hierarchy Process (AHP), Circuit through evaluation of the large size of in-service transmission
Breaker, Health Index, Weighting Factor. transformers measured for multiple oil test parameters.
Inspired by the possibility of AHP to capture CB expert
I. INTRODUCTION judgment, this paper adopts AHP with its capability on
One of the PLN core businesses is to ensure the availability resuming the expert's interviews to calculate the weights.
and reliability of electric transmission from the power plant to Then, scoring and weighting techniques are applied to
the customers. Therefore, in asset management, maintenance determine the overall CB Health Index.
of power apparatus, including High Voltage Circuit Breaker
(CB), plays a vital role in preserving both of reliability and the II. THE FRAMEWORK OF THE CIRCUIT BREAKER
availability of power systems. PLN operates thousands of CB HEALTH INDEX
in various operation voltage with a wide variety of CB The concept of HI comes from such an approach, having
insulation types installed in many substations. As part of asset the purpose of providing a direct and unequivocal indication
management, PLN has a division within its transmission main about the conditions of electrical apparatus, considering
office unit to manage all the circuit breakers in addition to history, stresses, critically, and output of diagnostic
other substation equipment. Due to a large number of CB measurements [6] [7]. The CB HI is gathered from the offline
installed in the grid, with different operational characteristics, measured data, visual inspection, maintenance completeness,
environmental conditions, ages, and aging mechanisms, age, operational stresses, and the failure rate of the equipment.
utilities agree to implement a condition monitoring on CB as Thus, HI can be employed as a fundamental tool for asset
well as other grid apparatus. management, able to provide diagnostic information and
Thereby, applying condition monitoring in CB will allow minimize the intervention of the expert operators that have to
the asset manager to establish health monitoring and fault download, process, and analyze the whole flow of diagnostic
diagnosis. The Health Index (HI) of a CB is primarily affected data from the data-mining database [8].
by the overall condition monitoring. A health index is assessed A. Fundamental of Circuit Breaker Health Index
by quantifying the results of field inspections, site testing into
The most common health index calculation procedure is to
an objective and quantitative index, providing the overall
perform offline diagnostic testing, which aims to collect
health of the assets. The use of Asset HI in managing assets
offline measurement data. However, this cannot describe the
and prioritizing investments into capital and maintenance
overall health condition of CB due to sudden changes in
programs is very helpful [1] [2]. The weighting and scoring
conditions caused by overloads or transients during operation.
methods are applied throughout the process of health index
In addition to that, offline testing does not consider real
calculation. Expert judgment is proposed to determine the
operation stresses. Another procedure is to use online

 
Authorized licensed !""##$!
use limited %&
to: Dedan Kimathi 317 on December 18,2020 at 20:42:56 UTC from IEEE Xplore. Restrictions apply.
University of Technology. Downloaded
 measurement and data mining, e.g., measurement of leakage
current, operational stresses, equipment failure rate, age, and
anomaly findings.
There are two parameters needed to calculate the CB health
index; namely parameter score and weighting factor, as
mentioned in (1). The parameter score arises from the criteria
of each test [9]. Utilize reference based on international
standards, recommendations, or use local guidelines that are
applied in corresponding utility. The Weighting Factor of each
parameter is derived by using the AHP. This results in a score
Si. The next step is to multiply each score to its predetermined
weighting factors Wi. Fig. 1 illustrates the calculation of
transformer HI using n-parameters.
ூ
` ‫ ܫܪ‬ൌ ෍ ܹ௜ ܵ௜ (1)
௜ୀଵ

Si = Score of each parameter; Wi = Weighting Factor.

B. Weighting Factor Determination Fig. 1. An overview of the CB health index procedure.
The relative importance of each criterion is determined by
allocating weights. The weighting factor is one of the a. Identifying the parameters.
important parts of the overall health index calculation [10].
According to the problem analysis, the CB
This factor represents how significant each parameter to derive
condition is best reflected from 2 factors that are
a single health index value from various parameters. The most
condition assessment factor and profile
common practice in adjusting the weighting factor of a health
assessment factor. There are nine parameters
index is using the experience and interview from field
that compose those factors. The parameter is
engineers. While the experts engage in many years of
fulfilled by online/offline measurement and data
experience, this approach has a strong possibility of subjective
mining.
results when not done properly.
b. Evaluating the parameters.
III. METHODOLOGY
Taking into account that the validity of the
The AHP is applied to obtain the weighting factor, being parameters from measurement data is subject to
one of the approaches methods to accommodate the years' accuracy and in need to comparing the critically
experiences from the utility asset managers or field engineers. of the parameters, so it is necessary to create
A. Analytical Hierarchy Process other parameters to comparing each other. These
AHP is a multi-criteria decision-making method that parameters are Measurement Reliability (MR)
combines qualitative and quantitative analysis. AHP consists and Criticality (CR). The MR aims to measure
of a hierarchy to get the importance of a set of alternatives to the reliability of the results on each parameter,
achieve a particular goal [11]. It has the capability to digitizes and CR is intended to measure the criticality of
experts' thinking processes of complex systems and judge the parameter compared to the others.
humans subjectively as the primary qualitative analysis. While (3) Create a pairwise comparison matrix to indicate the
the quantitative aspect of AHP, it has a widely used method to contribution or impact of each element to the criterion
determine the index weights by numerically differentiating at a higher level. Either performing an interview with
between various judgment factors to maintain the consistency the expert or collecting statistical or finding facts
of the thinking process. from other consideration to fulfill the pairwise
comparison matrix.
B. Subjective Weight Calculation Based on AHP
The proposed AHP model to acquire CB HI is built based (4) Perform the consistency check for the result.
on the following steps: (5) Repeat steps 3 and 4 for all levels and criteria in the
(1) Setup the objective goal, which is nothing else to hierarchy until the last parameter finished.
determine the weighting factors for the circuit (6) Calculate the weighting factor by summing each
breaker Health Index. weighted priority from the lowest level of the
(2) Build the hierarchy model based on the root cause proposed AHP hierarchy.
analysis of the problem and the evaluation factors.
The hierarchy of AHP proposed is shown in Fig. 2.

318 on December 18,2020 at 20:42:56 UTC from IEEE Xplore. Restrictions apply.
Authorized licensed use limited to: Dedan Kimathi University of Technology. Downloaded
 Fig. 2. AHP model for CB HI determination

The pairwise comparisons give an insight of the relative

importance of two factors by putting the scales. The important IV. RESULTS AND DISCUSSION
of pairwise comparisons is to translate subjective point of view CIGRE Working Group has conducted surveys on the
into exact numbers using the scales which shown in Table I. It failure recently, and the survey results said that mostly CB
is necessary to use balance scale rather than linear scale, since failures found to be triggered due to the malfunction of
the unevenly disperse local weights could ruin the sensivity of operating mechanism and control circuit and in that order
the scales when comparing closed elements [12]. compared to other CB assemblies [14] [15]. By number, major
CB failures, which caused by aging, wear, and corrosion, are
The form (2) provides the n x n comparison matrix, which the most common causes (about 50%) then followed by the
uses the balance scale. Since it is matrix pairwise, the diagonal manufacturing faults, design faults, and incorrect maintenance
values are 1, as the element is mirrored by diagonal of the as much as 15%.
matrix. Therefore, the value of the lower and higher triangle is
mutually reversed. Inspired by the recent CB failure surveys which have been
performed by CIGRE Working Group, this paper uses two
ͳ
ͳ kinds of factors to evaluate the condition of the CB, and these
‫ͳ ۇ‬ ‫ݑ ڮ‬௡ଵ ‫ۊ‬ are Condition Assessment Factor (CAF) and Profile
‫ݑ‬ଶଵ ‫ۋ ͳ ڮ‬
‫ ܥ‬ൌ ൣ‫ݑ‬௜௝ ൧ ൌ ‫ݑ ۈ‬ଶଵ ͳ (ʹ) Assessment Factor (PAF). The CAF consists of 5 parameters,
‫ۈ‬ ‫ݑ‬௡ଶ ‫ۋ‬ which are insulation quality test, leakage current, CB Open
‫ڭ‬ ‫ڭ‬ ‫ڰ‬ ‫ڭ‬ Close timing, contact resistance, and insulation resistance. All
‫ݑۉ‬௡ଵ ‫ݑ‬௡ଶ ‫ی ͳ ڮ‬ these parameters are collected from direct measurement when
there is either offline activity maintenance or online
Where uii=1 and uij=1/uji, i and j=1,2,3,…,n, respectively. measurement. While, the PAF is composed by the anomalies
finding, age, operational stresses, and failure rate.
Then, the next step is to calculate the weighting factor
using equation (3). A. Condition Assessment Factor (CAF)
೙
This factor is the main factor to indicate the performance
ඥς௡௠ୀଵ ‫ܥ‬௜௡ degradation of the CB. As mention above, CAF consists of 5
‫ݓ‬௜ ൌ (3) parameters; each type of parameter has a different importance
σ௡௜ୀଵ൫ ೙ඥς௡௠ୀଵ ‫ܥ‬௜௡ ൯
level on its status evaluation. One of the most important
Where wi=weighting factor for parameter i, Cin=pairwise parameters in CAF is the insulation quality test. There are three
matrix, and n=the size of the matrix. types of insulation used in the CB asset, which studied in this
paper. The insulation type is SF6 gas insulation (almost 70%),
TABLE I. SCALES FOR PAIRWISE COMPARISON [13] vacuum interrupters (about 20%), and oil, respectively. The
Definition Linear Scale Balanced Scale quality of the insulation within a CB is necessary to ensure the
Equal 1 1 ability and the preparedness of the CB in extinguishing the arc.
Weak 2 1.22
Moderate 3 1.5 The steps in examining the CAF are begun by creates and
Moderate Plus 4 1.86 completes the pairwise comparison of the parameters using a
Strong 5 2.33 balanced scale, as shown in Table I. A parameter in the
Strong Plus 6 3 comparison matrix generates one unique local weight. With
Very Strong 7 4 assistance from the expert experience, the pairwise
Very-very Strong 8 5.67 comparison matrix is acquired. Table II gives the comparison
Extreme 9 9 result of the analysis on how critical each parameter compared

319 on December 18,2020 at 20:42:56 UTC from IEEE Xplore. Restrictions apply.
Authorized licensed use limited to: Dedan Kimathi University of Technology. Downloaded
 to the other parameter in the condition assessment factor. This TABLE IV. RESULT OF THE COMPARISON ON MR & CR TO CRITERIA 1 IN
CONDITION ASSESSMENT FACTOR
factor is responsible for has the weight output (individual
weight) as MRxCAF, CRxCAF, and wCAF. Each weighting
Criteria MR CR Cri.1xCAF
factor provides weighting factor information for every
parameter in the CAF, as displayed in Table II and Table III. MR 1.000 0.922 0.480
The individual weights provide some importance or priorities CR 1.084 1.000 0.520
for assessing and fixing the CB.
B. Profile Assessment Factor
The next step demonstrates how to derive the weight of
MR and CR in the aspect of CAF. This comparison gives the To acquire the overall CB health index, there is one factor
importance level of accuracy aspect (MR) and criticality aspect left to be completed, that is PAF. Before PAF could be
when deciding parameters in CAF. Detail comparison results finished, data mining was employed to identify and collect the
are provided in Table IV. data needed to fulfill this factor. The main purpose of this
In order to find the final weighting factor, it would be section was to collect input data to measure the relative
helpful if we multiply the weighting factor of the parameters importance of any parameter over the parameter, and it
(MR x CAF and CR x CAF) with the corresponding weighting involved extensive interviews with the expertise. For data
factor of the Parameter Criteria 1 (MR and CR), respectively. collection, the decision-makers performed a pairwise
Fig. 3 reveals the overall weighting factor results in a condition comparison, producing a set of comparison matrixes that, in
assessment factor. In this figure, the contact resistance is the turn, resulted in an assessment instrument that reflected the
most significant factor affecting the condition assessment weight of each item (individual weight) in the hierarchy. Table
factor in the CB. The aim of the contact resistance test is to V and VI reports the comparison matrix result of profile
detect changes in contact geometry. It yields to imperfect assessment factor related to MR and CR. In the aspect of
performance of CB. Furthermore, the deformation in the accuracy, anomaly findings, and CB life are the 2-top
contact geometry will increase the risk for excess heat inside parameter in measurement reliability (MR). The deviation of
the CB, then elevate the possibility of reaching high those two parameters is close to each other. Otherwise, the
temperature than the CB maximum temperature design for, failure rate is the least accurate parameter in the aspect of
and thereby the probability of explosion of the breaker [16]. accurateness.
While, next after contact resistance, the insulation quality is
considered as the second factor that has an important effect to
depict the condition assessment factor. Unfortunately,
insulation quality measurement has less accuracy than the
contact resistance test. That's why it needs to do this test more
than once to assure the repeatability and the validity of the test
result.Comparison Result Matrix of How Accurate Each
Parameter on Condition Assessment Factor
TABLE II. COMPARISON RESULT MATRIX OF HOW ACCURATE EACH
PARAMETER ON CONDITION ASSESSMENT FACTOR
Resistance

Resistance
Insulation

Insulation
Leakage
Current

CB O-C

Contact
Timing
Test

Parameter MRxCAF

Insulation Fig. 3. Overall Weighting Factor Result in CAF

1.000 1.007 0.871 0.712 1.597 0.197
Test
Leakage
0.993 1.000 0.982 0.807 1.083 0.191 TABLE V. COMPARISON RESULT MATRIX OF HOW ACCURATE EACH
Current
CB O-C PARAMETER ON PROFILE ASSESSMENT FACTOR
1.149 1.018 1.000 0.924 1.400 0.214
Timing
CB Life (Age)

Operationaal

Contact
Failure Rate

1.405 1.239 1.083 1.000 1.695 0.249

Anomaly
Findings

Stresses

Resistant
Insulation Parameter MRxPAF
0.626 0.924 0.714 0.590 1.000 0.149
Resistance
TABLE III. COMPARISON RESULT MATRIX OF HOW CRITICAL EACH
PARAMETER ON CONDITION ASSESSMENT FACTOR Anomaly
1.000 0.843 1.336 1.882 0.297
Findings
Resistance

Resistance
Insulation

Insulation
Leakage

CB Life (Age) 1.187 1.000 1.219 1.252 0.286

Current

CB O-C

Contact
Timing
Test

Parameter CRxCAF Operational

0.748 0.821 1.000 0.990 0.218
Stress
Failure Rate 0.531 0.799 1.010 1.000 0.199
Insulation
1.000 1.513 1.635 0.873 1.454 0.249
Test
Leakage
0.661 1.000 0.903 0.866 0.818 0.167
Current
CB O-C
0.612 1.108 1.000 0.852 0.939 0.175
Timing
Contact
1.146 1.155 1.174 1.000 0.957 0.214
Resistant
Insulation
0.688 1.222 1.065 1.044 1.000 0.195
Resistance

320 on December 18,2020 at 20:42:56 UTC from IEEE Xplore. Restrictions apply.
Authorized licensed use limited to: Dedan Kimathi University of Technology. Downloaded
 TABLE VI. COMPARISON RESULT MATRIX OF HOW CRITICAL EACH C. Consistency Test
PARAMETER ON CONDITION ASSESSMENT FACTOR
This section presents the method to evaluate the
consistency of the matrix pairwise comparison, which is

CB Life (Age)

Operationaal

Failure Rate
caused by expert estimation. In fact, the expert view is

Anomaly
Findings

Stresses
Parameter CRxPAF demonstrably inconsistent in making a comparative judgment,
and this could happen when deals with intangibles criteria that
have no objective scales of measurement. Therefore, AHP
provides built-in systematic assessments for consistency
Anomaly
1.000 1.331 1.522 1.172 0.307 judgments. It is one of the greatest advantages of the AHP.
Findings
CB Life (Age) 0.751 1.000 0.853 0.925 0.217 AHP employs a test statistic, which called as the
Operational consistency index (CI), to measure consistency in the result of
0.657 1.172 1.000 0.879 0.225
Stress the expert estimation of the parameters. Then, the CI is
Failure Rate 0.853 1.081 1.138 1.000 0.251 compared to values of the same index for a randomly generated
matrix that is RI. Table VIII presents the RI values for the
TABLE VII. RESULT OF THE COMPARISON ON MR & CR TO CRITERIA n=1,2,…,9th (n is the number of the elements) order judgment
1 IN PROFILE ASSESSMENT FACTOR matrix pairwise comparison.
Criteria MR CR Cri.1xPAF The ratio of CI to the average RI for the same order matrix
is called the consistency ratio or CR. The formula for
MR 1.000 0.916 0.478
CR 1.092 1.000 0.522
calculating the consistency index (CI) and the consistency ratio
(CR) is as follows:
Table VII shows a detailed comparison of MR and CR
ɉ௠௔௫ Ȃ ݊
relatively compared to each other in the aspect of the profile ‫ ܫܥ‬ൌ (4)
assessment factor. The result of the aggregation of the expert ݊െͳ
interview represented that CR has a more important level than ‫ܫܥ‬
‫ ܴܥ‬ൌ  (5)
MR to consider the criteria 1 in the PAF. Comparing to the ܴ‫ܫ‬
same comparison in Table IV, the result between both tables is
not much different. It implies that experts have the same If the result of the consistency ratio is zero, then it means
agreement to put the same critical level on CR than MR. that the comparisons are perfectly consistent; higher ratios
indicate lower consistencies. It should be noted that each set of
According to the Table V, Table VI, and Table VII, the judgment needs to have one inconsistency ratio. It is important
overall weighting factor of PAF is derived by multiplying the to emphasize that the objective is to make "good" decisions,
individual weighting factor of the parameters (MR x PAF and not to minimize the inconsistency ratio. As previously
CR x PAF) with the corresponding weighting factor of the mentioned, the zero-inconsistency ratio means perfect
Parameter Criteria 1 (MR and CR). In the stacked bar, as consistency, but if the inconsistency ratio is more than zero,
shown in Fig. 4, it describes the overall weighting factor results there is some inconsistency. A ratio of 0.1 or less is generally
in the profile assessment factor. The head to head analysis of considered a tolerable inconsistency. The number 0.1 presents
each parameter on the profile assessment factor provides a as a common guideline for a maximum acceptable level, and
unique perspective that helps asset manager to find the most this number is applied in this study.
appropriate parameter for reflecting the health condition of the
circuit breaker. Anomaly finding is considered as the most The result of the consistency test can be seen in Table IX.
significant factor affecting the profile assessment factor in the For each of the 4-weighting factor, the inconsistencies were
circuit breaker. Interestingly, anomaly finding is also 0.003, 0.006, 0.008, and 0.002, respectively. There is no doubt
categorized as the most accurate parameter in CAF. It means that all values are less than 0.1 indicating acceptable levels of
the asset manager must perform well defining the anomaly inconsistency.
finding into a number, thus can be used in calculating the HI
of the circuit breaker. TABLE VIII. CONFORMITY INDICATORS RI [12]
Matrix Order n RI
1 0
2 0
3 0.58
4 0.90
5 1.12
6 1.24
7 1.32
8 1.41
9 1.45

TABLE IX. CONSISTENCY TEST ON THE COMPARISON MATRIX

Weighting Consistency
max CI CR
Factor (Y/N)
MR x CAF 5.018 0.005 0.003 Y
Fig. 4. Overall Weighting Factor Result in PAF CR x CAF 5.035 0.009 0.006 Y
MR x PAF 4.035 0.012 0.008 Y
CR x PAF 4.009 0.003 0.002 Y

321 on December 18,2020 at 20:42:56 UTC from IEEE Xplore. Restrictions apply.
Authorized licensed use limited to: Dedan Kimathi University of Technology. Downloaded
 TABLE X. RESULT OF THE COMPARISON ON CRITERIA 2 RELATIVELY TO REFERENCES
CRITICALITY

Criteria CAF PF WF [1] T. Hjartarson and S. Otal, "Predicting Future Asset Condition Based
on Current Health Index and Maintenance Level," in IEEE 11th
CAF 1.000 2.131 0.681 International Conference on Transmission & Distribution
PAF 0.469 1.000 0.319 Construction, Operation and Live-Line Maintenance, Oct 2006.
[2] N. Dominelli, "Equipment Health Rating of Power Transformers," in
TABLE XI. GLOBAL WEIGHTING FACTORS IEEE International Symposium on Electrical Insulation, Indianapolis-
USA, September 2004.
Factor Parameter WP WF
[3] C. S, T. Suwanasri and C. Suwanasri, "Health index determination of
Insulation Quality 0.224 aged high voltage substation in distribution system," in IEEE ECTI-
Leakage Current 0.178 CON, 2019.
Condition Assessment
CB O-C Timing 0.194 0.681 [4] P.-C. Lin, J.-C. Gu and M.-T. Yang, "Intelligent maintenance model
Factor
Contact Resistance 0.231 for condition assessment of circuit breakers using fuzzy set theory and
Insulation Resistance 0.173 evidential reasoning," IET Generation, Transmission & Distribution,
Anomaly Findings 0.303 vol. 8, no. 7, pp. 1244-1253, 2014.
Profile Assessment CB Life (Age) 0.250
0.319 [5] S. Tee, Q. Liu and Z. Wang, "Insulation condition ranking of
Factor Operational Stresses 0.221
transformers through principal component analysis and analytic
Failure Rate 0.226 hierarchy process," The Institution of Engineering and Technology
(IET) Generation, Transmission and Distribution, vol. 11, no. 1, pp.
D. CAF and PAF Weighting Factor 110-117, 2017.
AHP now has an opportunity to calculate the CAF and PAF [6] A. Jahromi, R. Piercy, S. Cress and W. Fan, "An approach to power
weighting factor by using matrix pairwise comparison. The transformer asset management using health index," IEEE Electrical
result presents that the condition assessment factor has the Insulation Magazine,, vol. 25 , no. no.2, pp. pp.20,34, March-April
most significant value of the weighting factor. 2009.
[7] L. Fornasari, A. Cavallini and G. Montanari, "A Smart Operational
E. Global Weighting Factor Monitoring Approach for High Voltage Electrical assets," in IEEE
CATCON, Kolkata, India, December 2013.
To acquire the global weighting factor, it needs to
[8] G. Montanari, "Condition Monitoring and Dynamic Health Index in
aggregate weights from Table II up to Table X. The result of Electrical Grids," in IEEE International Conference on Condition
the aggregation is shown in Table XI. The global weights Monitoring and Diagnosis, Xi'an - China, 2016.
create two sets of values; WP is weighting factors for [9] C.Neumann, "Optimization of Maintenance, Diagnostic Technique
parameter level, and WF is weighting factors for factor level. and Design of Substation Equipment based on Analysis of Service
The values coming from this global weighting factor could be Experience," in CIGRE B3-301, 2007.
used as the weighting factor to determine the circuit breaker [10] R. A. Prasojo, A. Setiawan, Suwarno, N. U. Maulidevi and B. A.
health index. Soedjarno, "Development of Analytic Hierarchy Process Technique in
Determining Weighting Factor for Power Transformer Health Index,"
V. CONCLUSIONS in 2nd International Conference on High Voltage Engineering and
Power Systems (ICHVEPS), Denpasar, Bali, Indonesia, 2019.
This paper mainly focused on the method to determine the
[11] S. Liu, Z. Liu, Y. Zhang, Y. Cao, J. Li and C. Hou, "Application of
weighting factor of the circuit breaker health index using AHP. AHP and Grey Correlation Analysis in the Study of Electric Life of
The analytical hierarchy process was appli ed to aggregate Vacuum Circuit Breaker," in 2018 28th International Symposium on
the input data to measure the relative importance of each Discharges and Electrical Insulation in Vacuum (ISDEIV),
parameter over other parameters on the circuit breaker health Greifswald, Germany, 2018.
index. The survey and interview were performed on five [12] A. Ishizaka and A. Labib, "Review of the main developments in the
experts with experience in operation and maintenance of the analytic hierarchy process," Expert System Application, vol. 38, no. 11,
circuit breaker for at least five years. p. 14336–14345, 2011.
[13] T. L. Saaty, "A scaling method for priorities in hierarchical structures,"
The experts come to an agreement that the condition J. Math. Psychol., vol. 15, no. 3, pp. 234-281, 1977.
assessment factor is more important than the profile [14] W. G. A3.12, "Failure survey on circuit breaker control systems
assessment factor to reflect the circuit breaker health index. summary report for Electra.," CIGRE, Electra , vol. 216 , Oct 2004.
The contact resistance is considered as the most significant [15] A. Janssen, D. Makareinis and C. E. Solver, "International surveys on
among the five parameters in the condition factor. This also circuit breaker reliability data for substation and system studies," IEEE
applies to the anomaly findings parameter that has the biggest Trans. Power Del., vol. 29, no. 2, April 2014.
weighting factor in the profile assessment factor. [16] M. Khoddam, J. Sadeh and P. Pourmohamadiyan, "Performance
Evaluation of Circuit Breaker Electrical Contact Based on Dynamic
The AHP model developed in this study can serve as a Resistance Signature and Using Health Index," in IEEE Transactions
basis for developing the health index model of the circuit on Components, Packaging and Manufacturing Technology, 2016.
breaker. One of the advantages of the AHP model is easy to
reproduce by using the different health index structure.
ACKNOWLEDGMENT
The author would like to thank PT PLN (Persero),
especially for PLN UIT JBTB, for providing the data and the
deep discussion.

322 on December 18,2020 at 20:42:56 UTC from IEEE Xplore. Restrictions apply.
Authorized licensed use limited to: Dedan Kimathi University of Technology. Downloaded