Engineering Failure Analysis 55 (2015) 182–192

Contents lists available at ScienceDirect

Engineering Failure Analysis

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

Failure analysis of power transformer for effective maintenance

planning in electric utilities
Raji Murugan a,⇑, Raju Ramasamy b,1
a
Department of Electrical & Electronics Engineering, College of Engineering Guindy, Anna University, Chennai 600 025, Tamil Nadu, India
b
Department of Industrial Engineering, College of Engineering Guindy, Anna University, Chennai 600 025, Tamil Nadu, India

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, power transformer failures are analyzed and their root causes are systemat-
Received 14 March 2015 ically investigated in Tamil Nadu Transmission Corporation Limited (TANTRANSCO)/Tamil
Received in revised form 29 May 2015 Nadu Generation and Distribution Corporation Limited (TANGEDCO) electric utilities,
Accepted 1 June 2015
based on 196 failure cases from the year 2009–2013. Failure analysis is conducted in
Available online 6 June 2015
two phases. Initially, voltage level, geographical zone and power transformer components
based failure analysis are performed through statistical analysis. Secondly, the most signif-
Keywords:
icant factors that cause power transformer failures are identified by using root cause anal-
Power transformers
Failure analysis
ysis (RCA). Finally, current maintenance practice is reviewed and an effective maintenance
Statistical analysis planning has been proposed for implementation in order to prevent these failures and to
Root cause analysis maintain the power transformers in good operating condition during their life cycle. This
Maintenance study provides a practical guidance to help maintenance personnel for the best utilization
of the power transformer in electric utilities.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Power transformers are the most expensive and strategic components of electric power system [1]. It plays an important
role by interconnecting in every stage of power transmission and distribution system [2]. Practically, the power transformer
is one of the high risk equipment in electric power system due to its harsh operating condition at various circumstances such
as high temperature, emergency overloading and continuous operation in outdoor environment. These conditions lead to
unexpected failure of power transformer. The failure of power transformer directly affects the reliability of the whole net-
work [3]. Failures in power transformers may cause disturbances to operating systems, resulting in unprepared outages and
power delivery problems. The power transformer failures in electric utilities can be very expensive and may take long time
for renewal or replacement of damaged components [4]. The failure not only impacts the cost-effective factors but also
causes image of the electric utilities lowered due to customer’s dissatisfaction [2].
It has been reported in several literature that there are large number of power transformer failures in various electric util-
ities across the world. CIGRÉ working group [5] has conducted a survey on failures in large power transformers and found
that about 41% of failures were due to on-load tap changers (OLTC) and about 19% were due to the windings, 13% were due to
leakage, bushing with 12%, 12% were others and 3 percent were core failures. Minhas et al. [6] conducted power transformers
failure analysis in Eskom network at South Africa during the period 1985–1995. Six failure modes were identified which

⇑ Corresponding author. Mobile: +91 9176763230.

E-mail addresses: raji.murugan12@gmail.com (R. Murugan), rraju@annauniv.edu (R. Ramasamy).
1
Mobile: +91 9444904300.

http://dx.doi.org/10.1016/j.engfailanal.2015.06.002
1350-6307/Ó 2015 Elsevier Ltd. All rights reserved.
 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192 183

includes lightning, core, tap-changer failures, ageing, and short circuit. In similar outage analysis study, Jagers et al. [7] found
that Bushings, tap changers and windings represent about 79% of outage causes and the contribution of core related failures
was only 2%. Jongen et al. [8] found that tap changer and bushing failure were the dominant causes of outage for transform-
ers having a rated voltage ranging from 110 to 150 kV. Abdelfatah et al. [9] conducted an outage data analysis for 220 kV
power transformers in Egypt Electricity Company, over the period 2002–2009 and found that outage causes were related
to Transformer related outage category – buchholz and pressure relief, protection system fails, breakdown and damage, fire
fighting systems, hot spots, oil leakage, flash over; Power System related outage category – outage of incomers, bus bar pro-
tection failures; Human factors related – operational and maintenance mistakes, etc. Thanapong suwanasri et al. [10] studied
the failure statistics of transformers in Thailand at a rated voltage of 230/115/22 kV and the failure analysis showed that
unknown failures was the highest with 41.3% followed by 31.7% of bushing, 17.5% of tap changer, 7.9% of leakage, and
1.6% of winding. McElroy [11] reported four single-phase EHV autotransformer failures due to transformer winding reso-
nance in an American utility. Previous studies reported only transformer component related failures through statistical anal-
ysis but associated root causes were not explored. This paper not only identifies the failure mode, the root causes of the
failures are also investigated.
This study is carried out based on 196 failure cases collected from the year 2009 to 2013 in Tamil Nadu Transmission
Corporation Limited (TANTRANSCO)/Tamil Nadu Generation and Distribution Corporation Limited (TANGEDCO) electric util-
ity context in Tamil Nadu. The major problem in TANTRANSCO/TANGEDCO is the frequent failure of power transformers,
resulting in fire, oil spillage, electric system outage and complete damage of equipment thereby increasing the unforeseen
repair or replacement costs. This also leads to loss of revenue to electric utilities. The main objective of this study is to ana-
lyze the failures of power transformer and, to minimize those failures by appropriate maintenance planning. Two methods
are employed in this study for failures analysis. First, statistical analysis is used in determining the various indices of power
transformer failures. Second, root cause analysis (RCA) is used in identifying the root causes of failures in power transform-
ers. Finally, condition based maintenance (CBM) is proposed in order to predict the current condition of the power trans-
former for preventing failures.
The structure of the paper is organized as follows. Section 2 discusses the methodology used in this study, Section 3
describes the data collection, Section 4 discusses on the statistical analysis performed to identify power transformer failures,
Section 5 discusses on the root cause analysis performed, Section 6 reviews the current maintenance practice in electric util-
ities, Section 7 describes the effective maintenance planning for preventing power transformer failures, Section 8 provides
the results and discussion and Section 9 concludes the paper.

2. Methodology

The study is conducted based on the failure data gathered from the TANTRANSCO & TANGEDCO electric utilities. The
methodology for the power transformers failure analysis is shown in Fig. 1. The power transformer failure analysis is con-
ducted in two phases. In the first phase, statistical failure analysis of power transformers is conducted. The failure data col-
lected has been split up into three different groups as voltage, geographical zone and power transformer component based
failures. The purpose of this grouping is to identify the highest failure impacts that occurred in each of the population.
In the second phase, power transformer component based failures are further analyzed in order to identify the root causes
for failures. Root cause analysis (RCA) is an essential step to identify the causes for failures of power transformer. RCA is facil-
itated by the use of various standard techniques such as check sheets, pareto analysis, brainstorming, reliability and main-
tainability circles, benchmarking, fail safe methods and cause and effect diagrams. These techniques are often used in
continuous and they are effective tools in identifying and solving maintenance and reliability problems [12].
A distinctive advantage of cause and effect diagrams is a graphical tool that is used to identify all potential failure causes
for each failure and effects in a systematic way [13]. The cause and effect diagram, also known as the fishbone diagram
because of its fish like shape and Ishikawa diagram in honor of its developer [14]. In this study, cause and effect diagram
is constructed to identify the factors influencing failures in power transformer components.
Finally, Time Based Maintenance (TBM), the current practice in TANTRANSCO & TANGEDCO is reviewed and Condition
Based Maintenance (CBM) is developed for effective maintenance implementation in electric utilities. The motive, advan-
tages, CBM basis condition assessment techniques, interpretation criteria and CBM decision model for effective maintenance
planning are described. This would enhance the power transformer operation for the best utilization in electric utilities.

3. Data collection

The state owned TANGEDCO is responsible for generation, purchase of bulk power and distribution of electric power for
low voltage consumers whereas the state owned TANTRANSCO is responsible for transmission and selling bulk power to the
distribution companies, high voltage (HV) and extra high voltage (EHV) consumers. The TANTRANSCO/TANGEDCO consists
of 842 numbers of substations in various voltage levels. Among these substations, large numbers of electrical equipments are
failing every year due to various factors including poor maintenance, overloading of electric network and aging reasons. The
power transformer is one among the major frequent failure component.
184 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192

Transformer failure data

from the year 2009 to 2013

Geographical zone based failures

Voltage based failures (Chennai north, Chennai
Power transformer component
(400KV,230KV,110KV,66KV south,Trichy,Covai,
based failures
and 33KV) Madurai,Vellore,
Tirunelveli,Erode,Vilupuram)

Failure data analysis

(Statistical analysis)

Identification of causes for

failures
(Root cause analysis)

Review current maintenance

practice

Effective maintenance planning

Fig. 1. Methodology for the power transformers failure analysis.

The failure data of transformers are collected from TANTRANSCO & TANGEDCO for five years from 2009 to 2013, for 196
transformers in voltage populations ranging from 33 kV to 400 kV and MVA rating from 5 MVA to 315 MVA. The collected data
include the transformer location, date and time, transformer failure duration, the disruptions in terms of power supply
interruptions, protection accomplishment and transformer repair time etc. From the total failures of 196 units, the age of
the population is in range of 2–30 years and the average age is 15 years. The number of failures of power transformers per volt-
age subpopulation during 2009–2013 and their total numbers are tabulated in Table 1. Table 2 shows the number of power
transformer failures in various geographical zones during 2009–2013. The power transformer components based failures dur-
ing 2009–2013 are shown in Table 3. Based on the failure data, various indices of power transformer failures are analyzed. The
failures of power transformers fall under different failure categories. These failure categories are related to transformer
component, protection, electric network, working environment conditions, human error, design/manufacturing/installation
failures and unclassified failures (others) [9]. But, the data availed from TANTRANSCO/TANGEDCO electric utilities, deals with
the voltage, geographical zone and power transformer component related failures.

4. Statistical analysis for power transformer failures

Over 196 power transformer failure cases are collected for analysis. The voltage subpopulation, geographical zone and
power transformer component based analysis of failures are presented below.

4.1. Voltages based failure analysis

The distribution of failures for various voltage categories for five years from 2009 to 2013 is shown in Table 1. The voltage
population of power transformers consists of 11 kV, 33 kV, 66 kV, 110 kV, 230 kV and 400 kV transformers. The rated power
is within the range of 5 MVA up to 315 MVA. There were 196 failures in five years. In Fig. 2 the percentage of failures versus
various voltage based failures are given in this study period. It can be found that there were 36% of failures in 33/11 kV, 22%
 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192 185

Table 1
Number of power transformer failure per voltage population during 2009–2013.

Years/subpopulation 400– 230– 110– 110– 110– 110– 66– 33– Total no. of
(kV) 230 kV 110 kV 66 kV 33 kV 22 kV 11 kV 11 kV 11 kV failures
2009 0 0 0 0 3 7 0 11 21
2010 0 1 1 9 9 6 0 15 41
2011 0 1 0 10 8 11 0 14 44
2012 0 0 0 10 7 8 0 16 41
2013 0 2 1 13 8 9 1 15 49
Total 0 4 2 42 35 41 1 71 196

Table 2
Number of power transformer failure in various geographical zones during 2009–2013.

Zone 2009 2010 2011 2012 2013 Total no. of failures

Chennai north 3 1 6 3 8 21
Chennai south 1 2 3 3 9 18
Villupuram 3 3 5 1 3 15
Vellore 2 10 5 5 9 31
Trichy 6 8 8 12 5 39
Madurai 1 3 4 5 3 16
Tirunelveli 2 3 6 5 7 23
Covai 6 3 4 3 16
Erode 3 5 4 3 2 17
Total 21 41 44 41 49 196

Table 3
Power transformer component based failures during 2009–2013.

Category wise failures 2009 2010 2011 2012 2013 Total no. of failures
Winding 7 5 13 3 28
Core 4 1 4 6 15
Bushing 2 9 9 6 26
OLTC 1 5 6 3 5 20
Tank 2 1 3 6
Coolant 3 2 4 9
Insulation 11 10 20 19 20 80
Others 7 1 1 1 2 12
Total 21 41 44 41 49 196

40%
35%
Percentage of failures

30%
25%
20%
15%
10%
5%
0%
230/110 KV 110/66 KV 110/33 KV 110/22 KV 110/11 KV 66/11 KV 33/11 KV
Failures of power transformer in various voltage populaons

Fig. 2. Failure statistics of power transformers voltage based failures.

in 110/33 kV and 21% in 110/11 kV voltage ratios. The voltage ratios 33/11 kV, 110/33 kV and 110/11 kV were the major fail-
ures encountered across nine zones in TANTRASCO/TANGEDCO electric utilities. The results of the analysis for the voltage
based population are most useful, when deciding about spare parts or spare transformers and repair or replacement decision
making for the electric network.
186 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192

4.2. Geographical zone based failure analysis

The distribution of power transformer failures based on geographical zone is shown in Table 2. Fig. 3 represents the per-
centage failures versus the geographical zone based failures. It can be noted that 20% of failures occurred in Trichy zone, 16%
of failures occurred in Vellore zone, 12% in Tirunelveli zone and 11% in Chennai north zone. The causes of failure for trans-
formers vary from one geographical zone to others. It is mainly due to various factors for failures across these zones which
includes working environment conditions, network loading conditions during summer, aging factor, poor maintenance of
power transformer and lack of training of maintenance personnel, etc. Therefore, the results of analysis for the geographical
zone wise are most useful to find the critical zones according to their percentages of failures. This will help to support spare
part management and effective maintenance activity planning in electric utilities.

4.3. Power transformer component based failure analysis

The power transformer component related failure is shown in Table 3. From Fig. 4 it can be found that insulation failures
are the predominant causes of failures, which contributes to about 41% from the total failures. The subsequent major con-
tributors are winding failures with 14% and bushings with 13%. OLTC contributes to 10% and core contributes to 8% of fail-
ures. The components with high percentage of failures are considered as critical and should be carefully focused. Therefore,
these failures are further analyzed in order to identify its root cause in the following section.

5. Root causes of failure for power transformer

The result of failure analysis reveals that predominant failure occurs on various components of power transformer.
Therefore, the transformer components related failures are further explored in order to identify its root causes. The insula-
tion is the most leading failure followed by winding and bushing failures. The remaining failures are OLTC, coolant and tank,
etc.
To illustrate cause and effects, a cause and effect diagram is developed and is shown in Fig. 5. The Cause and effect dia-
gram is used to describe the complete set of power transformer potential causes for failures. The diagram is structured with
the topic of interest of ‘‘power transformer failure’’ attached at the right-hand end. It has 8 main causes including insulation,
winding, bushing, OLTC, core, coolant, tank, others and 27 primary causes. The representation of failures on the diagram is
made in the order of importance. The most key causes are placed at the beginning (Insulation and winding). Placements of
the main causes are shown in the upper zone and left part of the axis whereas the less important causes are shown in lower
zone and right part of the axis.
The identified potential failure modes and their causes are (i-a) Solid insulation – mechanical damage, Over load causing
over heating insulation and fault in insulation material (i-b) Liquid insulation-oxidization of oil, thermal decomposition of oil
and contamination from moisture (ii) Winding – electrical (lightning surge, switching surge, connection fault and over volt-
age), mechanical (shipping damage, vibration, electromechanical forces which includes the hoop buckling of the innermost
winding, conductor tipping, conductor telescoping, spiral tightening, failure of clamping system and displacement of connec-
tion leads), thermal (overloading of the transformer, failure of the cooling system, low oil quality, operation of the trans-
former under excessive ambient temperature conditions and operation of the transformer in an overexcited condition)
and Insulation failure (local over heating of the winding insulating material, Cooling system fail and insulation containment
fail). (iii) Bushing-fault in material, damage of the porcelain and insulation failure. (iv) Tap changer – diverter switch or tap
selector failure, control device and drive mechanism fails. (v) Core – electrical (over voltage), mechanical (displacement of
the core steel during the construction or DC-magnetism) and insulation failure (poor insulation of the tightening screws of
the core, an obstructed oil-cooling duct and ungrounded core system that cause excessive heating of the core).(vi) Others –
operational errors, lack of maintenance, lack of skills for operators, aging of transformers. (vii) Cooling System – Pump, fan,
radiator failure and temperature gauge and control circuit and (viii) tank – leakage and rupture due to internal overpressure.
From this, the causes of power transformer component based failures are identified. Therefore, it is very important for elec-
tric utilities to prevent a transformer from these failures by developing effective maintenance, which is discussed in the fol-
lowing section.

6. Review of the current maintenance practice in electric utilities

The maintenance of power transformer plays a vital role in order to achieve maximum availability during their life cycles
in electric utilities. There are various maintenance schemes such as corrective, preventive, time based maintenance, but
TANTRANSCO/TANGEDCO is currently practicing time based maintenance (TBM). The authors reviewed the current mainte-
nance practice to understand the practicality and their issues. Fig. 6 shows the flowchart of TBM practice in
TANTRANSCO/TANGEDCO power sector. These maintenance works are performed at fixed time intervals regardless of equip-
ment condition. The time intervals are chosen based on maintenance schedule for annual, monthly and weekly basis using
equipment manufacturers’ specification or based on the regulations and instructions framed by the electric utilities. Once
the need for maintenance is decided, it is carried out based on a weekly work schedule. Prior to performing weekly
 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192 187

25%

Percentage of failures
20%

15%

10%

5%

0%
CHN N CHN S VPM VLR TRY MDU TIN CBE ERD
Failure of power transformer in zone wise

Fig. 3. Failure statistics of power transformer geographical zone based failures.

Others Winding
6% 14%
Core
8%
Insulaon
41%
Bushing
13%
OLTC
Coolant Tank 10%
5% 3%

Fig. 4. Failure statistics of power transformer component based failures.

maintenance schedule, maintenance activities, availability of labor and spare parts are prepared. Then, power transformer to
be serviced must be switched off (outage) to perform maintenance works. If outage is unfeasible, works are rescheduled for
some other time.
Time based maintenance activities in power transformer include visual inspection, cleaning bushings, oil level checking,
repair or replacement of failed components, oil replacement, tightening of tank gasket bolts and testing. Upon completion of
maintenance works, the equipment will be put into operation and their test reports will be gathered and filed. Then, the

INSULATION
INSULATION WINDING
WINDIN G BUSHING
BUSHING TAP
TAPCHANGER
CHANGER

Solid Insulation

1.Mechanical
1.Electrical 1.Fault in material
damage
2.Mechanical 2.Damage of
2.Fault in material
3.Thermal porcelain
3.Overheating
4.Insulation 3.Insulation failure 1.Diverter switch
4.Aging
2.Control device
3.Drive mechanism
Liquid Insulation
1.Oxidization of oil
2.Thermal decomposition
3.Contamination
POWER
POWER
TRANSFORMER
TRAN SFORMER
FAILURES
FAILURE S

1.Electrical
2.Mechanical
1.Leakage 1.Pump
3.Insulation
2.Internal rupture 2.Fan
3.Raditor failure
1.Operational errors
2.Lack of maintenance

TANK
TANK COOLANT
COOLANT OTHERS
OTHERS CORE
CORE

Fig. 5. Cause and effect diagram for power transformer failures.

188 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192

TANTRANSCO/TANGEDCO
time based maintenance regulation
and instructions

Yearly maintenance schedule

Monthly maintenance schedule

No

Is maintenance needed?

Yes

List maintenance work

activities

Prepare spare parts list Determine labor availability

Determine spare parts Weekly maintenance schedule

availability (Including outage request)

Perform maintenance work as

Test reports/Documentation
per the schedule

Fig. 6. Flowchart of a time based maintenance in TANTRANSCO/TANGEDCO.

performed maintenance works are registered in database as per the maintenance schedule laid down in the utility mainte-
nance regulations. Experience with TBM as scheduled maintenance has many drawbacks. The usage of TBM practice is min-
imized recently due to high maintenance costs, unnecessary outage of network system, labor intensive and time consuming.
Therefore, there is a need for effective maintenance in TANTRANSCO/TANGEDCO electric utility.

7. Effective maintenance planning for preventing power transformer failures

The Statistical, root cause analysis of failures and issues in current maintenance practice of power transformers constitute
an important basis for establishing the condition based maintenance (CBM) as effective maintenance in
TANTRANSCO/TANGEDCO electric utility. The aim of effective maintenance is to keep the power transformers in good work-
ing condition in its life cycle thereby extending the lifetime of equipment and reducing the failure probability in electric util-
ity. The essential need of the maintenance activity is to deploy at acceptable costs with no adverse environmental impact
[15]. CBM is the most widely employed strategy in various industries [16]. CBM is performed according to the actual con-
dition of the equipment.
The most important task of electric utility is to reduce failure occurrence and cost of maintenance of power transformers
in order to provide high reliability and availability.
The proposed CBM approach for TANTRANSCO/TANGEDCO is shown in Fig. 7. The CBM concept comprises of different ele-
ments, in which online condition monitoring and offline diagnostic test constitute the core of the CBM strategy. The first step
is monitoring which focuses on identifying symptoms of possible failure modes or abnormal condition in power transformer
components. The techniques should be cost efficient and possible to develop models for anomaly detection of any potential
problem at an early stage using sensors and data acquisition system. If an anomaly is detected, the own monitoring system
 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192 189

Sensors
Online
monitoring

Data acquisition system,

Graphical user interface

Anomaly detection,
1 Offline diagnostic test
Evaluation
Diagnostic test (International Maintenance decision
2
standards)

Fig. 7. CBM approach for power transformer.

will be able to diagnose, but normally, further offline diagnostic tests are required. The second step is the offline condition
diagnostic where an anomalous situation is explored to determine the type of faults and the severity of the problem. The
condition diagnostic test needs to be performed only on certain units are deemed abnormal condition. Finally, the mainte-
nance activities can be scheduled according to the condition of the assets.
To achieve this, condition based maintenance decision model for effective maintenance of power transformer in
TANTRANSCO/TANGEDCO power sector has been developed which is shown in Fig. 8. The steps involved in the CBM based
decision making are:

7.1. Step 1. Assessment of current condition of power transformer

CBM involves continuous or intermittent collection of online critical data from power transformers. But normally, further
offline measurement test has to be performed followed by online critical indication. The measurement data has to be inter-
preted based on the operating condition of the power transformer in order to determine current condition or failure mode if
any. The current condition of the power transformer is predicted through the following measurements:

 Online direct measurements (continuous condition monitoring the power transformer variables and component condition)

Online measurement is performed regularly and preferably continuously during the service condition of the power trans-
former. On-line monitoring system serves to detect in the form of an early warning system due to chemical/dielectric/ther-
mal or mechanical impact in power transformer. The list of online measurements includes hot spots, top oil and bottom oil
monitoring, winding temperature monitoring, localized faults/defects, gases monitoring, on-load tap changer mechanism
monitoring, vibration monitoring and cooling system monitoring, etc. The use of sensor, data acquisition and measurement
techniques for on-line applications together with a mechanism for generation of alarms when an anomaly is predicted aids
maintenance decision.

 Offline condition diagnostic measurements (standard diagnostic tests and measurements)

Offline diagnostic techniques should be applied in order to determine the current condition of power transformers. There
are several standard diagnostics techniques used to assess the current condition of power transformer. For effective utiliza-
tion, these techniques are prioritized as (1) visual inspection – inspection of oil conservator, oil level, tank external condition,
gaskets and oil sampling valves, assessment of radiators, coolers, gaskets, fans and pumps, oil temperature, winding temper-
ature, oil leakage checks, bushings external condition, bushing oil levels, pressure relief devices, buchholz relays, displays
alarm and trip signals indicators and inspection of dehydrating breather; (2) chemical test – dissolved gas analysis (DGA)
and physical chemical and electrical analysis of oil quality (PCEA); (3) dielectric test – insulation resistance/polarization
index, power factor, capacitance test and (4) electrical test – turns ratio, winding resistance, leakage reactance, excitation
current, magnetic balance test. The priorities of condition diagnostic techniques can distinguish which diagnostic parameter
affects the transformer’s current condition. These tests must be applied in the order of priorities as recommended by the
authors to identify and estimate the actual power transformer condition and its performances. The high priority chemical
based DGA and oil quality tests are performed initially. The dielectric tests are performed when there are faults detected
from oil analysis. Finally, electrical tests are performed when there are faults detected from dielectric analysis.
In addition to the standard diagnostic tests, there are some advanced tests including Polarization–Depolarization Current
(PDC), Return voltage measurement (RVM), Frequency domain spectroscopy (FDS), Frequency response analysis (FRA),
Frequency response stray loss (FRSL) Frequency response dissipation factor (FRDF), Partial discharges (PD), thermal imaging,
vibration analysis of transformer components, and degree of polymerization. These monitoring tests may detect problems
such as moisture in insulation paper and ageing of paper, conductivity, winding deformation, mechanical faults, local partial
discharge, hot spot at connectors, as well as insulation degradation.
190 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192

Predicting current condition by

online monitoring and offline
diagnosis testing

Evaluation
process

Does the current condition reach the No

failure limit or not?

Yes

Maintenance before
1,4 Maintenance decision 3 Maintenance after term
term

Replacement

Fig. 8. CBM decision model for effective maintenance of power transformer.

7.2. Step 2. Evaluation process

Evaluation process activities are helpful to identify the existing faults or weaknesses and also give some indication of
expected service reliability and remaining life of power transformers. These are performed based on online monitoring
and offline diagnostic measurement results which provide guidance on the current condition of the transformer. Online con-
dition monitoring evaluation processes consists of limit checking and trend analysis. Limit checking consists of comparing
actual measurements with configured limit values. A notification (alarm) is generated when the limit values exceed.
Trend analysis consists of discerning whether the level of a measured variable has increased or decreased over time, and
if it has, how quickly or slowly the increase or decrease has occurred. Offline measurement results are evaluated from the
recommended standards i.e. International Electro technical Commission (IEC) or Institute of Electrical and Electronics
Engineers (IEEE) or those provided by institutions i.e. International Council on Large Electric Systems (CIGRÉ), Electric
Power Research Institute (EPRI) or guidelines of diagnostic instrument manufacturer in order to predefine the failure limit.
In general, these interpretation criteria are applicable to all kinds of transformers. On the other hand, the criteria are appli-
cable when an actual test is compared to a reference test such as factory test or previous test. In case the transformer is not in
a satisfactory state, diagnostic measurement results are compared with allowable condition limit (inline to the standards) to
determine the current condition of the power transformer. If the current condition exceeds allowable limit, proceed for
maintenance.

7.3. Step 3. Maintenance decision process

The evaluation process outcome is important for maintenance decisions in electric utilities. The power transformer main-
tenance decision is determined based on inspection, analysis of measurement results, failure analysis and condition of the
equipment. This can be in four conditions. These four conditions are considered to be independent. Further it has been dis-
cussed as follows:

 If power transformer is not in an acceptable operational performance condition and if failures have occurred, mainte-
nance can be performed ahead of schedule if the components are repairable type (condition 1).
 If power transformer is not in an acceptable operational performance condition and if failures have occurred, replacement
can be performed ahead of schedule if the components are not repairable type (condition 2).
 If power transformer is in a good operating state, maintenance can be deferred for a specified time, unless it reaches the
failure limit or is about to exceed the failure limit in the mean time (condition 3).
 If the operational performance of power transformer is marginally acceptable but the timing is remote enough for the
condition to become an unacceptable operational performance, and if failures have occurred, maintenance can be per-
formed ahead of schedule (condition 4).
 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192 191

The power transformer after maintenance or replacement, proceed to Step 1 as a routine process. The maintenance deci-
sions in electric utilities enable operation and maintenance personnel in priority of work, initiation of work order and what
repair (maintenance) or replacement is to be performed. For example, if the equipment’s current condition level reaches or
exceeds the failure limit, the power transformer will be prepared for maintenance or replacement. Otherwise, the power
transformer is presumed to be in good condition and can still be used.

8. Result and discussions

Failure analysis of power transformer is very important in electric utilities. In this study, the failure analysis of power
transformer in TANTRANSCO/TANGEDCO electric utility was investigated and the following results were obtained.

(1) Transformer voltage level failures revealed the voltage levels having highest failures and their associated causes across
nine zones of electric utilities. Voltage level 110/11 kV, 110/33 kV and 33/11 kV encountered the major failures due to
network overloading, cyclic loading and network transient faults.
(2) Geographical zone wise power transformer failures indicated the critical zone according to their highest percentage of
failures. The main reasons for failures were (i) most of the transformers were located in adverse working environment
conditions and no special attention was given to the operating conditions, (ii) maintenance personnel shortage in their
respective zones, (iii) ageing factor, (iv) deterioration of insulating oil due to varying load fluctuations.
(3) Power transformer component based failures reveal that insulation, winding, bushing, on-load tap changer have the
highest failures during the study period. The various root causes for failures against each component were identified
for in-depth understanding of the causes for failures.
(4) These failure results can be utilized to support maintenance activity in electric utilities. Therefore, condition based
maintenance has been proposed for implementation. But there are some issues in implementing CBM approach in
power transformer as follows:
(i) Technological factors associated with CBM. Behind the CBM strategy, information communication and technology
(ICT) challenges are to be solved within power transformer applications. On one side, it refers to the advances in
sensors and on the other side, to the advances in the field of communications. Other technological advances are
still in their infancy like information data quality, frequency, noise and level of details of data availability. It means
that there are some limitations in ensuring the accuracy of diagnostics and prognostics.
(ii) Economical factors associated with CBM. The initial cost of CBM can be high. CBM requires sophisticated instru-
mentation and usage of various online monitoring devices. Also, to implement the CBM, in addition to investment
in software cum hardware, training of personnel for operation and maintenance makes it expensive.

9. Conclusion

The power transformer must be reliable for continuing operation due to its key role in the functioning of the electric
power system network. Through the collection of failure data of power transformers, voltage level, geographical zone and
component based operational failures were presented over the study period using statistical analysis. Subsequently, primary
causes were identified in power transformer component related failure using root cause analysis. The outcome of failure
analysis highlights the various causes of the failures providing crucial input data to decide on effective maintenance in elec-
tric utilities. Until recently, TBM practices in TANTRANSCO/TANGEDCO electric utilities experienced many drawbacks.
Therefore, CBM has been proposed with a view to reduce the failures based on the current condition of equipment instead
of scheduled time based maintenance. For effective maintenance implementation, CBM decision model was developed. The
CBM decision model predicts the current condition of the transformer by online monitoring parameters followed by offline
diagnostic measurement, evaluation criteria and maintenance decision. With the proposed condition based maintenance
practices, power transformers can be effectively maintained according to its actual condition by reducing failures tremen-
dously. This will enhance the operating conditions, extend the lifetime of power transformer and reduce maintenance costs.
Finally, the proposed method can also be further applied to other kinds of electrical equipments in the electric utility.

Acknowledgments

The authors acknowledge Tamil Nadu Generation and Distribution Corporation (TANGEDCO) & Tamil Nadu Transmission
Corporation (TANTRANSCO) for their valuable discussions, data, and support. This work has been supported through Anna
Centenary Research Fellowship, Anna University, Chennai. The Grant number is M.H.No:14/CR/ACRF/2014.

References

[1] Metwally IA. Failures monitoring, and new trends of power transformers. IEEE Potentials 2011:36–43.
[2] Wardani NUA, Purnomoadi AP, Septiani HI, Arifianto I, Cahyono B. Condition assessment of 500/150 kV power transformer based on condition based
maintenance. Int Conf Electr Eng Inf 2011:14–7.
[3] Abdelfatah M, EL Shimy M, Ismail HM. Outage data analysis of utility power transformers based on outage reports during 2002–2009. Int J Electr Power
Energy Syst 2013;47:41–51.
192 R. Murugan, R. Ramasamy / Engineering Failure Analysis 55 (2015) 182–192

[4] Mijailovic V. Probabilistic model for planning keeping of power transformer spare components with general repair time distribution. Electr Power Syst
2013;97:109–15.
[5] CIGRE Working Group 12.05. An international survey on failures in large power transformers in service. Electra 1983;88:21–48.
[6] Minhas MSA, Reynders JP, De Klerk PJ. Failure in power system transformers and appropriate monitoring techniques. Presented at the 11th
International Symposium on High Voltage Engineering London; 1999.
[7] Jagers JN, Khosa J, De Klerk PJ, Gaunt CT. Transformer reliability and condition assessment in a South African Utility. Presented in XV International
Symposium on High Voltage Engineering Slovenia; 2007.
[8] Jongen R, Peter Morshuis, Gulski E, Smit J. Statistical analysis of power transformer component life time. 8th International Power Eng Conf Singapore;
2007. p. 1273–7.
[9] Abdelfatah M, El-Shimy M, Ismail HM. Reliability analysis of 220 kV power transformers in Egypt. Ain Shams Eng J 2011;2:183–94.
[10] Suwanasri Thanapong, Chaidee Ekkachai, Adsoongnoen Cattareeya. Failure statistics and power transformer condition evaluation by dissolved gas
analysis technique. International Conference on Condition Monitoring and Diagnosis China; 2008.
[11] McElroy AJ. On the significance of recent EHV transformer failures involving winding resonance. IEEE Trans Power Ap Syst 1975;94:1301–6.
[12] Madu CN. Competing through maintenance strategies. Int J Qual Reliab Manage 2000;17:937–48.
[13] Mariajayaprakash A, Senthilvelan T. Failure detection and optimization of sugar mill boiler using FMEA and Taguchi method. Eng Fail Anal
2013;30:17–26.
[14] Jayswal A, Li X, Zanwar A, Lou HH, Huang Y. A sustainability root cause analysis methodology and its application. Comput Chem Eng 2011;35:2786–98.
[15] Plavsic T. The contribution of failure analyses to transmission network maintenance preferentials. Eng Fail Anal 2013;35:262–71.
[16] Abu-elanien AEB, Salama MMA. Asset management techniques for transformers. Int J Electr Power Energy Syst 2010;80:456–64.