Detection And Classification Of Magnetizing Inrush And Interturn

Fault Current In Transformer : Signal Processing Approach.

ABSTRACT threshold value. Mainly sine cosine

correlations,rectangular transforms ,discrete
Transformer protection has always been a Fourier transform (DFT), least square method,
challenging problem for protection engineers. Walash functions, Haar functions and Kalman
Various faults are occurs in the transformer, filtering techniques, are prominent.
such as open-circuit faults, earth faults, phase to
phase faults, over heating faults and inter-turn In normal condition, the excitation current is
faults. In this paper TEO based technique is small, so the transformer can escape the
implemented to detect the inrush current and influence of excitation current by the setting
interturn fault currents of transformer. This value. However, inrush current is likely to occur
paper presents a new strategy to enhance the when the transformer is closed with no load. The
security of transformer by using TEO. TEO is a inrush current is so big that it may reach about
nonlinear operator, which has been developed by 6-8 times of rated current. The big current can
Teager. Teager's energy operators are defined in cause the malfunction of differential protection
both the continuous and discrete domains and
are very useful 'tools' for analyzing single Internal faults in power transformer windings
component signals from an energy point-of- may include second harmonic. Also due to
view. It can track the energy and identify the advanced power electronics techniques used
instantaneous frequency and instantaneous power transformer also may produce
amplitude of mono-component signal. It is also magnetizing inrush current with second
used as a contrast enhancer of gray level images. harmonic components. So, this technique is not
useful to discriminate the inrush current. Other
INTRODUCTION techniques like artificial neural network voltage
and flux restraint and fuzzy logic are also not
Power transformer is a crucial part of any power appropriate as they require complex algorithms
system. For the continuous supply, reliable and and depends on the parameters of the power
efficient performanceof any power system it is transformer.
needed to protect it from different kinds of
faults. Mainly differential protection scheme is
used to protect the power transformers. Relay is The TEO was proposed by H. M. Teager while
an important and key element of any protection work on non-linear speech production
system. One of the main problem in differential modelling; in this method detection of inrush
protection of power transformer is to avoid the current and interturn fault current of transformer
wrong operation of relay due to inrush currents.
is taken place by calculating the energy of
High magnitudes of inrush currents can
misguide relay to operate falsely. There are signal. The new approach is used in this paper
many research works done to discriminate the by calculating the statistical parameter and
inrush currents from internal fault currents. thresholding method.
Mostly technique used by researchers is the
harmonic analysis approach to discriminate the Standard deviation (STD) is calculated at the
inrush currents from internal fault currents. The starting of the event from detection flag. The
main idea of the harmonic restraint differential results of STD range are tabulated in TABLE I.
relay is to extract the fundamental (1st), the
second(2nd) and sometimes the fifth (5th) The rest of the paper is organized as follows.
harmonics and to compare the ratio of the 2nd The phenomenon of inrush current and interturn
and 5th harmonics with 1 st to a predefined
fault current are in section II and section III. The
TEO mathematical modelling and algorithm are
in section IV. The experimental setup is
Inter-turn fault results in circulating currents
presented in section V and section VI contains among the short circuited turns. This results in
the result and discussion of inrush current and higher I2R losses, which deteriorates the
interturn fault current investigations. Finally, winding insulation. If the short circuited turns
some conclusions are summarized in section happen to be close to the core, an earth fault
VII. may arise from the turns to the core.

II.INTERTURN FAULT CURRENT III.INRUSH CURRENT

Interturn short circuit faults are the leading cause When a power transformer is switch on from
primary side, with keeping its secondary circuit
of power transformer’s failure. If not detected at
open, it acts as a simple inductance. When
incipient stage, these faults usually develop into electrical power transformer runs normally, the
more severe faults that would result in flux produced in the core is in quadrature with
irreversible damage to the transformer, applied voltage as shown in the figure below.
unexpected outages and the consequential That means, flux wave will reach its maximum
losses. Therefore, it is very important to detect value, ¼ cycle or π/2 angle after, reaching
incipient faults such as interturn short circuit maximum value of voltage wave. Hence as per
the waves shown in the figure, at the instant
fault at their developing stage, so that these
when,the voltage is zero, the corresponding
transformers could be safely removed from steady state value of flux should be negative
service. This avoids the chance of sudden failure maximum. But practically it is not possible to
and improves the reliability of power systems. It have flux at the instant of switching on the
is necessary to develop on-line diagnostic supply of transformer. This is because, there will
system that would aid in transformer be no flux linked to the core prior to switch on
maintenance. Different methods are used in the supply. The steady state value of flux will
only reach after a finite time,depending upon
these areas of interturn fault detection. Most of how fast the circuit can take energy. This is
the methods used are offline and having some because the rate of energy transfer to a circuit
limitations. cannot be infinity. So the flux in the core also
will start from its zero value at the time of
Inter-turn faults are short circuits between coil switching on the transformer.
turns on a single winding. According to Faraday's law of electromagnetic
induction the voltage induced across the winding
is given as
e = dφ/dt.
Where φ is the flux in the core. Hence the flux
will be integral of the voltage wave.

Voltage induced across the winding

e = E.sinwt = dφ/dt

therefore, φ = ∫e.dt = E ∫ sinwt.dt

If the transformer is switched on at the instant of

voltage zero, the flux wave is initiated from the
same origin as voltage waveform, the value of
flux at the end of first half cycle of the voltage
waveform will be
φm’= 2 φm

Where φm is the maximum value of steady state

flux. time of energizing the transformer. This current
is transient in nature and exists for few
The transformer core is generally saturated just milliseconds. The inrush current may be up to 10
above the maximum steady state value of flux. times higher than normal rated current of
But in our example, during switching on the transformer. Although the magnitude of inrush
transformer the maximum value of flux will current is so high but it
jump to double of its steady state maximum generally does not create any permanent fault in
value. As, after steady state maximum value of transformer as it exists for very small time. But
flux, the core becomes saturated, the current still inrush current in power transformer is a
required to produced rest of flux will be very problem, because it interferes with the operation
high. So transformer primary will draw a very of circuits as they have been designed to
high peaky current from the source which is function.
called magnetizing inrush current in transformer Some effects of high inrush include nuisance
or simply inrush current in transformer. fuse or breaker interruptions, as well as arcing
and failure of primary circuit components, such
as switches. High magnetizing inrush current in
transformer also necessitate oversizing of fuses
or breakers. Another side effect of high inrush is
the injection of noise and distortion back into the
mains.

IV.METHOD

A.TEO Mathematical Modelling.

The continuous time signal

𝑥(𝑡) = 𝐴𝐶𝑜𝑠(𝑤𝑐 𝑡 + ∅) TEO, ψ[x(t)]

Is defined as
̈ =̇̈ 𝐴2 𝑤𝑐 2
̇ ]2 − 𝑥(𝑡) 𝑥(𝑡)
𝜑[𝑥(𝑡)] = [𝑥(𝑡)
……. (1)

When Eq. (1) is transferred to an equivalent

discrete time form by using an approximation of
Waveforms showing inrush current
𝑥[𝑛]−𝑥[𝑛−1]
Magnetizing inrush current in transformer is the 𝑥̇ [𝑛] ≈ 𝑇
, for small T, it becomes
current which is drown by a transformer at the
Ψ[𝑥(𝑛)] = ([𝑥(𝑛)]2 − 𝑥(𝑛 − 1)𝑥(𝑛 + 1))/𝑇 2 From (4), the constant amplitude of signal is
determined as,
…….(2)
𝜓[𝑥(𝑛)]
𝐴 = 𝑠𝑞𝑟𝑡(𝑠𝑖𝑛2 (Ω ))………(5)
Where T is the sampling period between two 𝑐

adjacent samples, 𝑥(𝑖)and 𝑥(𝑖 + 1). In many

The TEO amplitude of the signal is obtained by
cases, an assumption of T = 1 is very acceptable,
using equation (5), then a combined TEO and
because the digital frequency in radian/sample
threshold algorithm used for inrush current and
has the information of T. As a result, we obtain a
interturn fault current detection , where a simple
final discrete form of TEO
TEO calculation is only used to detect an
Ψ[𝑥(𝑛)] = [𝑥(𝑛)]2 − 𝑥(𝑛 − 1)𝑥(𝑛 + 1) instance when an event occurs.

…….(3) In statistics, the STD (also represented by the

Greek letter sigma σ or the Latin letter s) is a
This result gives us a very meaningful measure that is used to quantify the amount of
significance that the magnitude and frequency variation or dispersion of a set of data values. A
components of a time varying sinusoidal signal low standard deviation indicates that the data
can be calculated by only three consecutive points tend to be close to the mean (also called
sampled data. For tracking of individual the expected value) of the set, while a high
amplitude and frequency of a signal, the energy standard deviation indicates that the data points
operator ψ[x[n]] needs to be separated into two are spread out over a wider range of values. The
parts, amplitude (A) and frequency (Ωc), by equation of STD is as,
using energy separation algorithm.
1 𝑁
𝜎𝑒2 = ∑ (𝑋 − 𝜇𝑐 )2 ………(6)
𝑁 𝑗=1 𝑗
TEO can approximately estimate the amplitude
envelope of amplitude modulated (AM) signals
and the instantaneous frequency component of
frequency modulated (FM) signals. It can be
also applied to AM-FM signals which mean the
B. TEO Threshold Based Algorithm.
signals with variable amplitude and variable
frequency.
A rule based algorithm is used for the detection
Consider a signal with constant amplitude and
of inrush current and interturn fault current of
constant Frequency which is expressed as
transformer. The beginning of the signal is used
𝑥(𝑛) = 𝐴 𝐶𝑜𝑠(Ω𝑐 𝑛 + ∅) for setting the threshold value for the stationary
background signal called ‘‘reference segment’’
Where and must be transient free. The length of the
reference segment can be set differently due to
𝜑[𝑦[𝑛]] the characteristics of signal. The remaining part
Ω𝑐 = cos−1(1 − )
2𝜑[𝑥[𝑛]] of the signal is called as ‘‘test segment’’. By
applying TEO on reference segment we get the
So, from equation (3) the discrete time energy average and threshold value, where as threshold
operatorBecomes, value is the product of experimentally set value
and average value which is variable. Then
𝜓[𝑥(𝑛)] = 𝐴2 𝑠𝑖𝑛2 (Ω𝑐 )……..(4) comparison of them takes place. If it is lower
than the threshold, then it is called as
background. However, if it is greater than the 5.“Teager Energy Operator and its Application
threshold, then it is a transformer fault current, in the Study of Induction Motor Rotor Broken
Bars Fault”,Shihua Yin, Niaoqing Hu, Ling
then comparison is takes place with next sample
Chen, Lei Hu.
average value, if it is higher than it has continues Science and Technology on Integrated Logistics
transformer fault current, if lower then assumed Support Laboratory,
as end of event. The transformer fault currents National University of Defense Technology
further classified from their STD ranges; hence ,Changsha Hunan, China. 2015 1EEE.
using this algorithm estimated duration of
transformer fault currents, detection and 6.“Single-Phase Grid Voltage Frequency
Estimation Using Teager Energy Operator Based
classification is done.
Technique”Md. Shamim Reza, Mihai Ciobotaru,
and Vassilios G. Agelidis,
The flowchart of TEO threshold based method.
10.1109/JESTPE.2015.2405094, IEEE.
EXPERIMENTAL SETUP
7.“Signal processing using the Teager Energy
Operator and other nonlinear operators”,
University of Oslo, Department of Informatics.
Cand. Scient Thesis, Eivind Kvedalen,May
RESULT AND DISSCUSSION 2003.

CONCLUSIONS

REFERENCE

1.“Detection of generated and measured

transient power quality events using Teager
Energy Operator”,Abdulhamit Subasi , A.
Serdar Yilmaz , Kadir Tufan. Science Direct,
November 2010.
2. “A new generalization of the discrete teager-
kaiser energy operator - Application to
biomedical signals”,meryem jabloun ,Prisme
laboratory of the orleans university, orleans.
IEEE2017.

3.“Combining Continuous Wavelet Transform

and Teager-Kaiser Energy Operator for ECG
Arrhythmia Detections”. Nady, M. Moness, M.
Massoud,R. R. Gharieb, Department of Electical
and Electronic Engineering Faculty of
Engineering, Assiut University Assiut, Egypt,
2016 IEEE.

4.“Teager-Kaiser Energy operator in Hilbert

space for heart sound segmentation”
L. N. Sharma,Department of Electronics and
Electrical Engineering,Indian Institute of
Technology Guwahati, India2016 IEEE.