Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Seminar Report Final

Download as pdf or txt
Download as pdf or txt
You are on page 1of 28

A

SEMINAR REPORT
ON

DETECTION OF WINDING
DEFORMATION IN POWER
TRANSFORMER
Submitted in partial fulfillment of the requirements for the degree of
MASTER OF TECHNOLOGY
(Power Systems)
By
PISE SUMIT GOVIND
(P13PS012)
: Supervisor:
Dr. H. G. Patel

DEPARTMENT OF ELECTRICAL ENGINEERING


SARDAR VALLABHBHAI NATIONAL INSTITUTE OF TECHNOLOGY
SURAT 395007

November 2014.

SURAT-395 007, GUJARAT, INDIA


DEPARTMENT OF ELECTRICAL ENGINEERING

CERTIFICATE
This is to certify that the seminar report entitled DETECTION OF WINDING
DEFORMATION IN POWER TRANSFORMER submitted by PISE SUMIT GOVIND,
P13PS012 is a record of bonafide work carried out by him in partial fulfillment of the
requirement for the award of the degree of MASTER OF TECHNOLOGY IN
ELECTRICAL ENGINEERING (Power Systems).

Date: 20/11/2014
Place: SURAT

Dr. H. G. Patel
(Faculty Supervisor)

Examiners

PG In charge
Prof. H. R. Jariwala

Head of Department
Dr. (Ms.) Anandita Chowdhury

SVNIT

SARDAR VALLABHBHAI NATIONAL INSTITUTE OF TECHNOLOGY

Acknowledgements
I take this opportunity to express my gratitude and deep regards to my guide Dr. Hiren
G. Patel, Assistant Professor, EED, SVNIT for his exemplary guidance, monitoring and
constant encouragement throughout the course of this seminar. The blessing, help and
guidance given by him time to time shall carry me a long way in the journey of life on which
I am about to embark.
Further I would like to thank Prof. H. R. Jariwala, P.G. In-charge, POWER SYSTEM
ENGINEERING, SVNIT, & Dr. (Ms.) Anandita Chowdhury HoD, EED, SVNIT for their
cordial support, valuable information and guidance, which helped me in completing this task
through various stages.
I wish to express my gratitude towards my parents, for keeping faith in my abilities and
encouraging me to improve at every step. Last but not the least, I am thankful to my friends
who helped me forget my worries and have a relaxed mind.

Pise Sumit Govind


Sardar Vallabhbhai National Institute of Technology
Surat

Abstract
Power transformers are the most expensive and critical component of our power system.
Its failure would be a costly event. Winding deformation is caused mainly due to short circuit
events, other reasons may be ageing of insulation, and mechanical stresses developed due to
transportation. Although, deformation may not result in immediate failure, however the
dielectric strength of the winding is greatly reduced. And a further untoward short circuit
faults may result in complete rupture of insulation which will result in complete damage of
the winding. To prevent permanent damage of the transformers, a routine diagnosis is
necessary for detection of winding deformation.
This seminar report presents FRA which measures the impedance or transferred
impedance of the transformer winding. It is measured as function of the frequency by
applying a low voltage sinusoidal test signal with variable frequency in the range from 100
Hz to 10 MHz. The signals are measured at discrete frequencies to determine amplitude and
phase of the transfer function for the desired full frequency range. Final aging or degradation
assessment of the test is based on verification of repeated recorded signals with original
recorded signals during installation. The interpretations of the results remain a grey area and
are usually left to experts to determine the type and location of fault.
This report also discusses a novel method for deformation diagnostics of transformer
windings using terminal measurements. The shortcomings of existing practices in locating
deformation and determining its extent are discussed. But the method based on Deformation
Coefficient which requires fewer terminal measurements to determine location and extent of
deformation.

ii

Table of Contents
Acknowledgements ................................................................................................................... i
Abstract..................................................................................................................................... ii
List of Figures ........................................................................................................................... iv
List of Tables ............................................................................................................................ iv
Chapter 1 .................................................................................................................................... 1
INTRODUCTION ..................................................................................................................... 1
1.1.
1.2.

Overview ................................................................................................................. 1
Detection of Winding Deformation ......................................................................... 1
1.2.1.
Reactance Comparison Method ....................................................................... 1
1.2.2.

Frequency Response Analysis (FRA) .............................................................. 2

1.2.3.

Using Deformation Coefficient: ....................................................................... 2

1.3.
Motivation ............................................................................................................... 3
1.4.
Report Organization ................................................................................................ 3
Chapter 2 .................................................................................................................................... 5
MODELLING A TRANSFORMER WINDING ...................................................................... 5
2.1.
Introduction ............................................................................................................. 5
2.2.
Lumped Parameter Model ....................................................................................... 5
Chapter 3 .................................................................................................................................... 7
FREQUENCY RESPONSE ANALYSIS .................................................................................. 7
3.1.
3.2.
3.3.

Principle of FRA Test .............................................................................................. 7


FRA Measurement Setup......................................................................................... 7
Fault Diagnosis using FRA...................................................................................... 8
3.3.1.
Types of Analysis ............................................................................................. 8

3.4.
Simulation Results ................................................................................................... 9
Chapter 4 .................................................................................................................................. 13
DEFORMATION COEFFICIENT .......................................................................................... 13
4.1.
4.2.

Introduction ........................................................................................................... 13
Procedural Steps for Proposed Method ................................................................. 14
4.2.1.
Measurements ................................................................................................ 14
4.2.2.

Determination of Cs and Cg from Fingerprint Readings ................................ 15

4.3.
Uniqueness of Deformation Coefficient ................................................................ 16
4.4.
Fault diagnosis using deformation coefficient (DC) ............................................. 19
Chapter 5 .................................................................................................................................. 21
CONCLUSION ........................................................................................................................ 21
References ................................................................................................................................ 22

iii

List of Figures
Figure 2.1 Transformer Winding Model.................................................................................... 6
Figure 3.1 Experiment setup of the system................................................................................ 8
Figure 3.2 Driving point impedance characteristic of a winding with terminals 22' shorted .. 10
Figure 3.3 Driving point impedance characteristic of a winding with terminals 22' open
circuited.................................................................................................................................... 10
Figure 3.4 Driving point characteristic (open circuited farther end 22') ................................. 11
Figure 4.1 DC v/s p.u. change in Cg ........................................................................................ 17
Figure 4.2 DC v/s p.u. change in Cs......................................................................................... 17
Figure 4.3 Change in capacitance across terminals 1 and 2 as a function of change in
sectional series capacitance...................................................................................................... 18

List of Tables
Table 2.1 Parameters of the winding ......................................................................................... 6
Table 3.1 Frequency and amplitude of natural frequencies ..................................................... 12
Table 4.1 Coefficients .............................................................................................................. 19

iv

Chapter 1
INTRODUCTION
1.1.

Overview

Power transformers play a central role within the electrical transmission and distribution
network. They are generally involved in energy transfer in power transmission and
distribution networks, converting ac voltage or current from one level to another level
depending on the receiver end. During its lifetime, it is subjected to mainly electrical,
mechanical and thermal stresses. The main reason behind the stresses can be due to ageing of
insulation, during transportation, and the most critical reason is a short circuit event. During
short circuit, the current in the transformer winding is enormous as compared to normal
current. The high magnitude of short circuit current has detrimental effect on the transformer
and other equipment associated with it.
The interaction of short circuit current and the mutual flux (linking both the primary and
secondary windings of transformer) gives rise to large amount of electromagnetic forces
which are responsible for the deformations occurring in the transformer winding. The
winding is subjected to mainly two types of deformations, axial deformation due to the action
of axial force and radial deformation due to the action of radial forces on it. The deformations
in the winding may lead to permanent failure of the transformer, if undiagnosed for long
period of time. The short circuit withstand capability is drastically reduced due to
deformation in the winding. Future short circuit faults and high mechanical or dielectric stress
may completely rupture the insulation and may result in permanent damage of the
transformer. Since, the failure of a transformer will be dangerous and costly for the utility,
proper diagnosis is necessary to prevent its overall collapse.

1.2.

Detection of Winding Deformation

Commonly used diagnostic techniques are:


1.2.1. Reactance Comparison Method
In this method, the reactance of the transformer under test is measured before and
after a short circuit test. According to IEC-60075 standard, the reactance should not exceed

1% for 100 MVA and above rated power transformers. However, this method is not
applicable for detecting winding displacements which has negligible impact on the reactance.
1.2.2. Frequency Response Analysis (FRA)
FRA is a generally well-known testing technique within the industry to determine a
transformer winding deformation, e.g. coils, turns, layers, HV leads, etc., owning to shortcircuit currents (faults), impact during transportation and aging. Dick and Erven were the first
to use the FRA method to detect the transformer winding deformation in 1978[1]
It is based on the principle that every transformer has its unique signature. FRA
signature is the transfer function output/input for an entire range of frequency starting from
few Hz to MHz of a transformer. The signature can be obtained either by injecting an impulse
across the transformer terminals at low voltages, or by making a frequency sweep of a
sinusoidal signal. The former approach is known as low voltage impulse (LVI) method and
the latter is sweep frequency response analysis (SFRA). The obtained measurements are
compared with a reference set. The reference can be a previous measurement during healthy
condition in time based comparison. If the references are drawn from separately tested legs of
transformer, such comparison is known as construction based comparison. The comparison
done with identically constructed transformer is known as type based comparison. If the
comparisons show deviations in the nature of the curves, it indicates deformation or
displacement of the winding.
SFRA is a powerful and highly sensitive technique. The main advantages of SFRA
over LVI are:
High Signal to noise ratio because of the filtering of the broadband noise by the
network analyzer.
Less measuring equipment needed for the purpose
Finer frequency resolution can be achieved even at low frequencies which was
not possible in LVI approach.
Wider range of frequencies can be injected.
1.2.3. Using Deformation Coefficient:
This is an elegant method recently proposed [2]that can be applied to any current
carrying coil or winding, with a simple procedure for deducing the status of the winding. In
2

this novel method, instead of a sweep frequency analysis, only three measurements at
selected high frequency and one at selected low frequency are required. In, application of this
method is explained for a single transformer winding through simulation studies. There are
two principal innovative aspects of this method: measurements from both the winding ends
and Deformation Coefficient (DC), which help in determining the location and extent of a
deformation. The procedure used to calculate the DC makes it lie in a very narrow range for
all winding sections irrespective of the extent of deformation.

1.3.

Motivation
The demand for reliable supply of electrical energy has grown over few decades. The

role played by power transformer towards power distribution without unwanted outages is of
the highest importance. To prevent uneconomical outages, and maintain the reliability of
power, routine diagnosis of the transformer becomes necessary to maximize its uptime. One
of the reasons of failure of the transformers is the deformations in the winding. The major
cause of deformations in the winding is due to the effect of electromagnetic forces arising as
an outcome of a short circuit fault.
Various diagnostic techniques like reactance comparison method, dissolved gas
analysis, furan analysis, low voltage impulse methods, etc. have been used for many decades.
Still, the need of the time was a highly sensitive technique that could detect even minor
deformation occurring in the winding structure. Sweep frequency analysis approach is a
highly sensitive technique which has several advantages over other techniques. The locating
of deformations and estimating their extent is very difficult task. Recently a generalized offline approach based on Deformation Coefficient has been suggested which requires fewer
terminal measurements [3] to determine location and extent of deformation.

1.4.

Report Organization

Chapter 1st consists of a brief overview of the topic, detection of deformation techniques,
motivation and report organization.
Chapter 2nd deals with the modelling of transformer winding for the analysis.
Chapter 3rd discusses Frequency Response Analysis Method for detection of winding
deformation.

Chapter 4thdiscusses a novel method for deformation diagnostics of transformer


windings using terminal measurements. The shortcomings of existing practices in
locating deformation and determining its extent are also discussed.

Chapter 2
MODELLING A TRANSFORMER
WINDING
2.1.

Introduction
Various analytical methods have been employed in recent decades for transformer

modelling. Some of the methods have drawbacks in terms of detailed depiction of real
transformer geometry but have been utilized due to their simplicity for analyzing deformed
windings.
For modelling, the frequency range of interest is from few Hz to about 1 MHz
High frequency behavior of windings is identified by their resonances (i.e. transfer
function maxima and minima). The introduction of mechanical displacement will alter these
characteristic properties. From an operational requirement, the type and location of faults are
imperative. These relations can be established by measurements on the power transformer or
through the use of appropriate model of the transformer for simulation.
Transformer windings are generally represented by a lumped parameter equivalent
circuit [3]. In this chapter, the winding model shown in Figure 2.1 is most appropriate for
description of a transformer winding over a wide range of frequencies.

2.2.

Lumped Parameter Model

The lumped parameter model has been used in many works and is shown that:
A complete lumped-element model of the transformer can be built by means of a
series n-stage ladder network as illustrated in Figure 2.1.
The equivalent circuit of a transformer winding (with terminals of other windings
grounded) consists of a finite number of sections having elements like Cg, Cs, Lii, and
Lij, which are the sectional shunt capacitance, sectional series capacitance, selfinductance and mutual inductance, respectively.
Each section of the winding is represented by a pi () model with Cg/2 as its two legs.
Parameters of the winding are listed in table 2.1
5

L1n
L2n
L22

L11

Cs

Cs
Cg/2

Lnn

Cg

Cs
Cg

Cg/2

Cg
n

1'

2'

Figure 2.1 Transformer Winding Model

Table 2.1 Parameters of the winding

Number of sections = 8
Number of turns in each section = 20
18 SWG (Standard Wire Gauge)
Mean diameter of winding = 185 mm
Height of winding = 225 mm
Sectional series capacitance Cs = 2.2 nF
Sectional ground capacitance Cg = 1.0 nF
Inductance values (H)
L

M1

M2

M3

M4

M5

M6

M7

117

58.1

31.4

19.14

12.55

8.6

5.95

4.45

(The measured self and mutual inductances are given in the table. Here, L is the selfinductance of the sections, and M1, M2, etc., are the mutual inductances between the sections.
The suffix indicates sectional separation. For example, mutual inductance between sections
1-4 or 2-5 or 3-6 or 4-7 is M3 and that between sections 1-8 is M7.)

Chapter 3
FREQUENCY RESPONSE ANALYSIS
3.1.

Principle of FRA Test


The principle of FRA test is quite simple. As already discussed, the transformer

winding can be represented by lumped parameter model. It can generate a unique signature
when tested at discreet frequencies and plotted as a curve. The distance between conductors
of the transformer forms a capacitance. Any movement of the conductors or windings will
change this capacitance. This capacitance being a part of complex network, any change in
this capacitance will be reflected in the curve or signature.
An initial FRA test is carried out to obtain the signature of the transformer frequency
response by injecting various discreet frequencies. This reference is then used for future
comparisons. A change in winding position, degradation in the insulation, etc. will result in
change in capacitance or inductance thereby affecting the measured curves. Tests are carried
out periodically or during major external events like short circuits and results compared
against the initial signature to test for any problems. FRA test reveals if the transformers
mechanical or electrical integrity has been compromised.

3.2.

FRA Measurement Setup


Currently the FRA testing is done when the transformer is in offline mode. A good

deal of preparation is needed before testing can be carried out safely. The transformer needs
to be disconnected from the rest of the network. The recommendation is that FRA testing is
performed when the transformer is under maintenance or after refurbishing commissioning
following a major fault. The test configuration for such work is shown in Figure 3.1. If
testing were carried in the field, the measurement equipment needed include a spectrum
analyzer, properly shield 50 ohms coaxial cable and a bushing capacitance (220pF). The
bushing capacitance acts as an insulation device for a safe passage of electrical energy
through an earth field.
The most common three FRA measurement techniques used currently are: the end to
end voltage ratio, the input impedance and the transfer voltage ratio measurement.

CB

Impedance
Measurements
Z

SPECTRUM
ANALYZER

PC

Figure 3.1 Experiment setup of the system

3.3.

Fault Diagnosis using FRA


The sensitivity of FRA has been broadly tested in numerous works [4, 5] as a tool for

fault simulation experiments and real case studies of transformer examinations. Numerous
types of faults can be detected by FRA, this include winding movement, winding
deformation, loss of clamping pressure at the end terminals of windings, inter-turn faults and
multiple core grounding etc.
There has been consensus between experts that major faults that are caused by large
movements of the core or winding can be identified in the low frequency range. While the
minor faults which include inter-turn faults are identified in the high frequency range.
The disadvantage of the FRA as a diagnostic method is that there are no standard
processes to analyze and deduce the measurement. Regularly, the diagnosis work is
performed by specialists through the aid of visual inspection and mathematical parameters.
The examination depends on factors such as the type of recording used for comparison, the
topographies extracted from the frequency response etc. These factors are account below:
3.3.1. Types of Analysis
The SFRA method is built on the analysis of frequency response recordings taken
during the lifespan of the transformer. There are two prospects:

Analysis with known reference recordings. It is assumed that a set of historical


copy offering a healthy state of the transformer is accessible.

Analysis without reference recordings. If reference recordings are unavailable


of the transformer there are two possibilities:
8

Analysis using recordings which belong to different phases of the same


transformer. As part of asymmetric character of the transformer,
various phases exhibit different characteristics. This methodology is
beneficial in that measurement is made under the equivalent
circumstances.
On the other hand, analysis can be done using a twin transformer. The
comparison is made on the basis of recording from an identical
transformer either new or old with same representation. This present
issues since construction characteristic of a transformer that matches
operational conditions are difficult to find.

3.4.

Simulation Results
The frequency response curve of a laboratory scale winding having eight sections,

with values of inductances (self and mutual) and capacitances of various sections as listed in
table 2.1, is shown in Figure 3.2 and Figure 3.3. The driving point impedance characteristics
with other terminal 2 grounded and terminal 2 open are displayed in Figure 3.2 and Figure
3.3 respectively.
Figure 3.4 shows driving point characteristics of the model winding with ground
capacitance change effected for sections 4, 6 and 8 (done one at a time), along with the case
of no deformation.

Impedance (ohms)

10

10

10

10

Impedance

10

-5

5
Frequency (Hz)

10
x 10

Phase
100

Phase (deg)

50
0
-50
-100

5
Frequency (Hz)

10
x 10

Figure 3.2 Driving point impedance characteristic of a winding with terminals 22' shorted

Impedance (ohms)

10

10

10

10

Impedance

10

-5

5
Frequency (Hz)

10
x 10

Phase
100

Phase (deg)

50
0
-50
-100

5
Frequency (Hz)

10
x 10

Figure 3.3 Driving point impedance characteristic of a winding with terminals 22' open circuited

10

Table 3.1 shows amplitude and frequency of the poles and zeros of the driving point
impedance characteristics for the various deformations. One can easily observe the
complexity in differentiating the curves in Figure 3.4. Also the changes in natural frequencies
are diminutive. To detect these changes, one needs costly instrumentation in addition to
expertise for deducing conclusions about the winding status. Methods based on synthesis,
IFD (Index for Frequency Deviation), IMD (Index for Magnitude Deviation) and integrals of
magnitude difference curve are suggested in the literature [6, 7]. Furthermore frequency
dependence of inductance and resistance makes the analysis more complicated.

Impedance (ohms)

10
10
10
10
10

Impedance

10

No deformation
Section 4 radially deformed by 5%
Section 6 radially deformed by 15%
Section 8 radially deformed by 30%

-5

-10

Frequency (Hz)

10
x 10

Phase
100

Phase (deg)

50
0
-50
-100

5
Frequency (Hz)

Figure 3.4 Driving point characteristic (open circuited farther end 22')

11

10
x 10

Table 3.1Frequency and amplitude of natural frequencies

Comparative study of shift in natural frequencies due to radial deformation at different sections
Section

Pole
Frequency

Zero
Log(Z)

No deformation
Section 4 by 5%
Section 6 by 15%
Section 8 by 30%
Pole/zero number

96.75
96.8
97.55
103

No deformation
Section 4 by 5%
Section 6 by 15%
Section 8 by 30%
Pole/zero number

186
186.6
186.8
193

No deformation
Section 4 by 5%
Section 6 by 15%
Section 8 by 30%
Pole/zero number

257.5
257.7
259.3
261.5

No deformation
Section 4 by 5%
Section 6 by 15%
Section 8 by 30%
Pole/zero number

322
322.5
323
324.6

No deformation
Section 4 by 5%
Section 6 by 15%
Section 8 by 30%
Pole/zero number

380
380.5
381
383

No deformation
Section 4 by 5%
Section 6 by 15%
Section 8 by 30%
Pole/zero number

432
432.5
433
434.2

No deformation
Section 4 by 5%
Section 6 by 15%
Section 8 by 30%

472
472
473.5
473.5

Log(Z)

kHz

kHz
Pole/zero number

Frequency

1
4.66
4.663
4.6548
4.685

41.1
41.2
42.2
44

2
4.55
4.545
4.563
4.545

138
138.2
138.7
146

0.63
0.632
0.625
0.635
3

4.325
4.329
4.32
4.305

228
228
230
234

1.015
1.016
1.01
1.04
4

4.035
4.03
4.03
4.025

303
304
304.5
306.25

1.3
1.3
1.31
1.31
5

3.68
3.68
3.67
3.67

369
369.5
370.5
372

1.595
1.595
1.595
1.6
6

3.32
3.325
3.32
3.305

425
425.5
426.2
427.9

1.86
1.86
1.865
1.87
7

2.92
2.91
2.92
2.905

12

0.34
0.34
0.335
0.325

468
468.2
469.6
469.9

2.2
2.21
2.21
2.21

Chapter 4
DEFORMATION COEFFICIENT
4.1.

Introduction
The most important task of any winding diagnosis method is that it should be capable

of locating and finding extent of deformation. A winding deformation changes the inductance
and capacitance associated with its corresponding part (section). It is suggested in [8]that
minor axial and radial deformations can be correlated with changes in Cs and Cg,
respectively. Hence, the proposed method makes use of changes in capacitance values for the
diagnostics purpose. At low frequencies the terminal impedance (between 1 and 1) is
capacitive when terminals 2 and 2 are open (Figure 3.3); this is obvious because at low
frequencies the inductive reactance is very small. Further, referring to Figure 2.1, the terminal
capacitance at a low frequency is merely the sum of all sectional ground capacitances. This
simple measurement can be used to detect any change in ground capacitance, which can be
correlated to the extent of radial deformation of the winding. The selection of low frequency
is discussed in the following section.
The next obvious task is to locate the deformed section. After working on different
possibilities, with primary focus being simplicity, the authors have identified another
important feature of the winding. It is well-known that the driving point impedance
characteristics for any transformer winding show that the terminal impedance is capacitive at
high frequencies, as also evident from Figure 3.3. If terminal capacitance is measured at such
a high frequency, then it is clear from Figure 2.1 that the terminal capacitance is series and
parallel combination of the capacitances (Cs and Cg) in the capacitive ladder network. The
leading question to be answered was: Can terminal capacitance measurement at such a high
frequency from both the ends give additional information about location of deformation? The
approach developed requires that, at the selected high frequency, terminal impedance
(capacitance) measurements be done at both ends of the winding. The first measurement
needs to be carried out between terminals 1 and 1 (termed as C1H) and the second between
terminals 2 and 2 (termed as C2H) with reference to Figure 2.1. While performing the
measurement at one end, the other end of the same winding has to be open-circuited and
terminals of all other windings should be grounded. Any deformation leads to a change in the
measured capacitive reactance. After the terminal capacitances are measured at both winding
13

ends, their deviations from the fingerprint values can be obtained, and then the ratio of the
deviations at the two ends can be computed. Any non-limiting function of this ratio is defined
as Deformation Coefficient (DC). For example, DC can be obtained as:
C C1' H
DC log10 1H

'
C2 H C2 H

(4.1)

Where, C1H and C2H are the fingerprint (intact) values of measured terminal
capacitances at the selected high frequency and C1H and C2H are the terminal capacitance
values after deformation at the same frequency. Using this coefficient the deformed section
can be located effectively, since its value is distinct and is in a narrow range for a given
section of the winding, irrespective of the amount of practical deformation, as demonstrated
in the following section.
The next question to be answered was: How to determine the extent of axial
deformation, if present? Intuitively, it was clear that the capacitance between terminals 1 and
2 should reflect change in sectional capacitance Cs. With an additional measurement at the
selected high frequency between the two ends of the winding (between terminals 1 and 2 in
Figure 2.1), the change can be determined. This can be correlated to the extent of axial
deformation of the winding.

4.2.

Procedural Steps for Proposed Method

With this background, the proposed method can be outlined as consisting of the following
step[8, 9].
4.2.1. Measurements
After isolating the transformer and removing electrical connections of the winding
under test with other windings and connecting all terminals of other windings to ground, the
following four capacitance measurements need to be done:

The capacitance (C1H) between winding terminals 1 and 1 (Figure 2.1) at


selected high frequency (FH): The high frequency needs to be selected only
once initially, and is such that the impedance offered by the winding remains
capacitive beyond that frequency. This can be easily ensured by observing the
phase of the impedance.

14

The capacitance (C2H) across winding terminals 2 and 2 at the same high
frequency.

The capacitance (C12H) across winding terminals 1 and 2 at the same high
frequency.

The capacitance (CL) between winding terminals 1 and 1 or terminals 2 and


2 at FL (a low frequency, say about 50 to 100 Hz): The low frequency is
selected such that the measured impedance value is predominantly decided by
the parallel combination of sectional ground capacitances.

Only these four measured capacitances and two selected frequencies need to be
preserved as the fingerprint or reference values for the winding under test (for the purpose of
future diagnostics).
4.2.2. Determination of Cs and Cg from Fingerprint Readings
One can determine the values of Cs and Cg from the reference measurements as
follows. Here an eight section model of a uniform winding is used for the demonstration
purpose.
First, sectional ground capacitance is determined from the measured value of CL:
C
Cg L
(4.2)
8
Now, ratio of C12H to Cg is (this is obtained by standard network reduction technique):
C12 H 1 2k 4 16k 3 20k 2 8k 1

Cg
4
4k 3 10k 2 6k 1

(4.3)

Where, k Cs

Cg

Using the measured value of C12H and previously calculated value of Cg, k and hence
Cs can be determined.

15

4.3.

Uniqueness of Deformation Coefficient


Once the reference values are obtained, charts or tables can be prepared for values of

DC by changing sectional ground capacitance to simulate radial deformation (by changing


both Cg/2 capacitance values for a given section in pi () model representation) or by
changing sectional series capacitance to simulate axial deformation. A table or chart can also
be prepared for C12H (= C12H - C12H) for various sectional series capacitance changes. These
tables and charts form the set of references to be used in future diagnostic tests. Charts in
Figure 4.1 and Figure 4.2 show DC values for various % changes in sectional ground and
series capacitance respectively. Similarly, Figure 4.3 shows variation of C12H with % changes
in series capacitance for different sections.
From Figure 4.1 and Figure 4.2, it is evident that the values of DC for a particular
section lie in a narrow range. Thus, each section has an almost constant and distinct DC value
irrespective of the extent of practical deformation. For the eight sections winding model
considered in this work the mathematical expressions can be obtained for DC and C12H.[10]
The expressions for DC as a function of p. u. change in sectional ground capacitance are
given below. For sections at the extreme ends, the DC expression is different from the rest of
the sections and which is of the form (prefix s stands for positive/negative sign in the DC
expressions):

x a x 2 bx c
DC s log10 k g

x d x e

(4.4)

Whereas, for the rest of the sections:

x a x 2 bx c
DC s log10 k g
x d x 2 ex f

(4.5)

The DC as a function of p.u. change in sectional series capacitance is:

yg
DC s log10 k s

yh

(4.6)

Similarly, the expression for C12H is of the form


C12H k

16

y
y i

(4.7)

Sectional ground capacitance change


4
First section
Section section
Third section
Fourth section

3.5

Deformation Coefficient

2.5

1.5

0.5

0.05

0.1

0.15
p.u. deformation

0.2

0.25

0.3

Figure 4.1 DC v/s p.u. change in Cg


Sectional series capacitance change
5
First section
Section section
Third section
Fourth section

4.5
4

Deformation Coefficient

3.5
3
2.5
2
1.5
1
0.5
0

0.05

0.1

0.15
p.u. deformation

0.2

Figure 4.2 DC v/s p.u. change in Cs

17

0.25

0.3

Change in C12H v/s sectional deformation


60
Section 1 or 8
Section 2 or 7
Section 3 or 6
Section 2 or 7

Change in C12H (in pF)

50

40

30

20

10

0.05

0.1

0.15
0.2
p.u. change in sectional Cs
(Base Cs=2.2nF, Cg=1.0nF)

0.25

0.3

Figure 4.3 Change in capacitance across terminals 1 and 2 as a function of change in sectional series
capacitance

18

In the expressions, x indicates p. u. change in sectional ground capacitance and y


indicates p. u. change in sectional series capacitance. Table 4.1 gives coefficients of above
expressions for different sections. For the fifth to eighth sections, the DC expressions are
negative of the fourth to first sections, respectively, due to the assumed symmetrical nature of
the winding.
Table 4.1Coefficients

Coefficient

Sec-1

Sec-2

Sec-3

Sec-4

Sec-5

Sec-6

Sec-7

Sec-8

kg

-1337.1

954.4

56.0

3.8

3.8

56.0

954.4

-1337.1

10.8

10.798

10.794

10.777

10.777

10.794

10.798

10.8

13.931

15.747

16.545

16.977

16.977

16.545

15.747

13.931

26.77

42.302

49.947

52.702

52.702

49.947

42.302

26.77

8.985

9.898

10.502

10.755

10.755

10.502

9.898

8.985

8.53

19.32

17.529

17.15

17.15

17.529

19.32

8.53

-----

72.745

57.402

54.164

54.164

57.402

72.745

-----

ks

74083

1146.1

59.08

3.8387

3.8387

59.08

1146.1

74083

1.1830

1.3805

1.445

1.4644

1.4644

1.445

1.3805

1.1830

1.9386

1.5703

1.4945

1.4743

1.4743

1.4945

1.5703

1.9386

-196.1

-55.0

-17.6

-8.3

-8.3

-17.6

-55.0

-196.1

1.4734

1.4734

1.4734

1.4734

1.4734

1.4734

1.4734

1.4734

4.4.

Fault diagnosis using deformation coefficient (DC)


Reference tables/charts can be prepared as explained previously for a transformer

winding along with reference values for terminal capacitances, and selected high and low
frequencies. During an actual diagnostic test (after transport, some field use or a short-circuit
event), measurements need to be taken as per the procedure explained in section 4.2.1. Three
measurements at the selected high frequency (C1H, C2H, C12H) and one at the selected low
frequency (CL) are required for the winding under test.

19

Any deviation in the values of C1H and C2H from the reference values indicates a
deformation, and a difference between the values of CL and CL indicates a change in the
sectional ground capacitance. This can be correlated to the extent of radial deformation as
explained earlier. Comparison of the calculated DC value with the values in the reference
chart corresponding to ground capacitance changes (Figure 4.1) gives the deformed section
number. No appreciable difference between the values of C L and CL indicates a series
capacitance change. Comparison of calculated DC value with the reference chart
corresponding to series capacitance changes (Figure 4.2) gives the deformed section number.
Finding the difference between C12H and C12H and comparing it with the values
corresponding to the deformed section number in the reference chart as in Figure 4.3 gives
the extent of axial deformation.

20

Chapter 5
CONCLUSION

FRA based simulation studies for driving point function show that it is difficult to
deduce conclusions regarding deformed section and extent of deformation if it occurs
towards the farther (neutral) end of windings.

One requires lengthy algorithms, advanced resources and possibly, the analysis by an
expert to deduce the conclusions about the location and extent of deformation.
Moreover, these methods are not generalized in the sense that, diagnostic conclusions
for one winding cannot be applied directly for diagnostic studies of another winding.

Thus a novel and elegant method for winding deformation diagnostics is described.
Unlike FRA diagnostic studies, the proposed method does not require frequency
sweep to deduce conclusions about winding state. Also, the need of interpretation by
an expert is not essential. The method requires three high frequency and one low
frequency terminal capacitance measurements to determine the location and extent of
deformation.

Any non-limiting function of the ratio of change in terminal capacitance at one end
of the winding to the corresponding change at the other end is defined as
Deformation Coefficient (DC). The DC is found to be an effective indicator for
locating a winding deformation. It is observed that the DC for each section is distinct
and lies within a narrow range of values for any practical extent of deformation.

In the case of a deformation leading to a ground capacitance change, the extent of


deformation can be determined from the low frequency terminal capacitance
measurement. The capacitance, as measured across the two terminals at the selected
high frequency, is found to give the extent of series capacitance change.

The features including simplicity of procedure and a reduction in the number of


parameters to be measured and more importantly, deskilled analysis make the
proposed method a serious contender for deformation diagnostics. The method has a
potential to detect multiple section deformations.

Hence the method of deformation coefficient is the most suitable method for
detecting transformer winding deformation.

21

References
[1]

[2]
[3]

[4]
[5]

[6]
[7]

[8]

[9]

[10]

E. Al-Ammar, G. G. Karady, and O. P. Hevia, "Improved technique for fault detection


sensitivity in transformer maintenance test," in Power Engineering Society General Meeting,
2007. IEEE, 2007, pp. 1-8.
P. Joshi and S. Kulkarni, "A diagnostic method for determining deformations in a transformer
or reactor winding," Indian Patent Application No, 1893.
P. M. J. a. S. V. Kulkarni, "Use of Deformation Coefficient for Transformer Winding
Diagnostics," International Journal of Emerging Electric Power Systems, vol. Vol. 9, p. Art. 7,
October 2008.
E. Dick and C. Erven, "Transformer diagnostic testing by frequuency response analysis,"
Power Apparatus and Systems, IEEE Transactions on, pp. 2144-2153, 1978.
V. Venegas, J. L. Guardado, S. G. Maximov, and E. Melgoza, "A computer model for surge
distribution studies in transformer windings," in EUROCON 2009, EUROCON'09. IEEE, 2009,
pp. 451-457.
S. Kulkarni and S. Khaparde, "Transformer engineering: design and practice, ser," Power
engineering. New York, NY: Marcel Dekker, Inc, vol. 25, 2004.
K. Ragavan and L. Satish, "Localization of changes in a model winding based on terminal
measurements: Experimental study," Power Delivery, IEEE Transactions on, vol. 22, pp. 15571565, 2007.
M. Wang, A. J. Vandermaar, and K. Srivastava, "Improved detection of power transformer
winding movement by extending the FRA high frequency range," Power Delivery, IEEE
Transactions on, vol. 20, pp. 1930-1938, 2005.
E. Rahimpour, J. Christian, K. Feser, and H. Mohseni, "Transfer function method to diagnose
axial displacement and radial deformation of transformer windings," Power Delivery, IEEE
Transactions on, vol. 18, pp. 493-505, 2003.
P. Joshi and S. Kulkarni, "Transformer winding diagnostics using deformation coefficient," in
Power and Energy Society General Meeting-Conversion and Delivery of Electrical Energy in
the 21st Century, 2008 IEEE, 2008, pp. 1-4.

22

You might also like