2018 Twentieth International Middle East Power Systems Conference (MEPCON), Cairo University, Egypt

A Review of Partial Discharge Detection

Techniques in Power Transformers

Shady S. Refaat Mahmoud Sayed Mohammed A. Shams Amira Mohamed

Electrical and Computer Electrical Power & Machines Electrical Power & Machines Texas A&M University at
Engineering Dep., Texas Dep. Dep. Qatar, Doha, Qatar
A&M University at Qatar, Cairo University Cairo University Texas A&M University
Doha, Qatar Giza, Egypt Giza, Egypt College Station, US
[email protected] [email protected] [email protected] [email protected]

Abstract— Power Transformer is the heart of the electric insulation condition in electrical power equipment. Partial
power system. One of the main causes of power transformers’ discharge occurs when the electric field strength in a region
insulation failure is Partial discharge (PD). Partial discharges of the dielectric exceeds the breakdown strength of the
are electrical sparks which occur within the insulation material
of power transformers. Degradation of insulation material is dielectric material. Compromises in the insulation of a power
indicated by the presence of partial discharges. PD detection in transformer will render it prone to failures arising due to
transformers is an intricate task due to the inductive behavior stresses in the power system including short circuits,
of the transformer windings. The two main categories of PD switching transients and lightning strikes[1].
detection are on-line detection techniques and off-line detection Nowadays, many researches have focused on monitoring,
techniques. Off-line PD detection techniques incur high costs detecting, and locating partial discharges in transformer by
due to the downtime of power transformers alongside the
utilization of expensive instrumentation. In contrast, on-line PD developing of several techniques with low sensitivity to
detection techniques are less expensive and can be performed on various PD activities to ensure high reliability and continuous
energized power transformers. Due to the variable nature of PD operation of power system, reduce progressive deterioration
activity in power transformers, continuous on-line monitoring of the insulating materials and avoid catastrophic breakdown.
of PD activity is an optimal solution for gaining deeper insights Various PD measurement, activity monitoring and
into the characteristics of the PD activity. This paper provides detection techniques for power transformer reported in
an overview of PD activity in power transformers including its
mechanism and mathematical modeling. Off-line and on-line PD literature can be broadly categorized to Electrical,
detection techniques are reviewed and compared. Electromagnetic , Acoustic Emission, Gas Presence, Optical-
Based, Combinational Methods[3]. Modern PD detection
Keywords — corona, electric breakdown, partial discharges, techniques use a combination of electromagnetic sensors for
power transformers. reliable detection of PD activity. Meanwhile, PD detection
and measurement in power transformers is complicated by
the inductive nature of the transformer. Inductive mutual
I. INTRODUCTION coupling between the windings and resonant inductive
ower transformer role is critical for the operation of electrical
power system. They are exposed to various stress conditions Effects
during service operation, such as electrical, thermal,
mechanical, chemical, dynamic, environmental stresses[1].
These stresses cause the transformer electrical insulation Optical
system degradation, aging and eventual failure of electrical
insulation. Partial discharge (PD) is the initial stage of
insulation degradation, and cannot be avoided. Partial Mechanical
discharges are electrical discharges or sparks that occur
within an insulation material due to contamination, defects,
Electrical
poor insulation material, harsh operational condition,
moisture absorption, aging, and deterioration of insulation
material during service[2]. PD in power transformers can Chemical
occur in solid, liquid and gaseous insulation materials. In
power transformers, PD often forms in a gap of solid
Heat
insulation, air bubbles existing in liquid insulation, between
different insulation layers with varying dielectric properties,
or on sharp edges or metallic surface protrusions. PD testing Fig. 1. Physical Effects of partial discharge in Power Transformer.
is a non-destructive and sensitive tool for assessing the
1020

978-1-5386-6654-8/18/$31.00 ©2018 IEEE

coupling are some factors affecting PD detection in power terminals of the transformer. Partial discharge activity causes
transformers[4]. The two main types of PD detection are high-frequency signals that can be filtered and distinguished
namely the on-line detection method and the off-line from the voltage frequency. Partial discharge measurement
detection method. Given the variable nature of PD activity, often suffers from interference to accurately measure partial
continuous on-line monitoring is an optimal method for discharge severity. The sources of interference may be
gaining deeper insights into the characteristics of the PD classified into two categories, internal interferences, and
activity. Technological advancements in digital signal external interferences. The internal inference sources include
processing have significantly reduced the limitations winding imbalance, thermal noise and stray capacitance to
associated with on-line PD detection. ground imbalance. The external inference sources include
The contribution in this paper is to give a general insight mechanical vibrations and harmonic signal. The external
and discuss the various techniques in partial discharge within interference can be minimized through electrical isolation
the context of the power transformer. The effect of partial between the measuring circuit and the transformer control
discharge on the operation of power transformers will be unit.
discussed. Comparison of different recent techniques used in
Partial discharge is quantized using apparent charge q, this
the measurement, detection and location of partial discharge
including both the on-line and off-line techniques will be charge refers to the amount of charge when electric charge is
discussed. In addition, the advantages and limitations of the injected directly into the test terminals would give the same
detection partial discharge activity methods will be analyzed, effect.
while the technical challenges for each detection techniques
will be discussed. III. OFF-LINE PARTIAL DISCHARGE DETECTION
TECHNIQUES
II. PARTIAL DISCHARGE MECHANISM IN POWER
It is desirable to detect and locate the PD in the power
TRANSFORMERS
transformer before progressive deterioration of the insulating
Partial discharges in power transformer are classified into materials occurs. Offline partial discharge measurements
two categories, external partial discharges, and internal techniques require the transformer to be out of service, and
partial discharges. The external partial discharge refers to the disconnected from the power network. It can be performed
electrical discharge that happened outside the transformer during regular maintenance schedule or doubtful failure. It is
such as a discharge which occurs on surface such as static generally more expensive due to the down time incurred by
electrification, surface tracking and corona discharges. the transformers alongside the long duration of the test. Off-
The internal partial discharge refers to the electrical line partial discharge measurements techniques for
discharge that happened internally in a closed system such as transformer include direct detection circuit and dissipation
a discharge which occurs inside voids, voids produced in factor tip-up.
glue, bushings, bubbles, cavities in liquid or solid dielectrics,
bad connection, and fixed metallic particles during A. Direct Detection Circuit
transformer manufacturing process. Partial discharge PD detection was implemented using analog instruments,
activities in transformer insulation can be detected in visualizing PD pulses using an oscilloscope. Interpretation of
different forms including electromagnetic emissions, the analog measurements of PD activity including charge
acoustic emissions, electrical, mechanical, optical and transfer and phase relationship between PD pulses and the
gaseous forms as illustrated in Fig.1. applied voltage was done manually by observation. It is
Partial discharges are electrical sparks that occur in the important to note that liquid dielectric media in power
event of insulation failure. Partial discharges arise when the transformers tolerate higher levels of PD than solid dielectric
electric field strength in a region of the dielectric exceeds the media[4].
breakdown strength of the dielectric material. In the event of PD detection and measurement in power transformers is
a PD, electrons flow between the ends of the cavity in which inherently intricate due to the inductive nature of
the PD occurs[3]. Due to the flow of electrons in a minuscule transformers. The PD pulse measured at the transformer
region at the velocity of light, pulses of extremely short terminals has to pass through the transformer windings which
duration are detected. The motion of partial discharges can be modeled as a network of capacitors and inductors.
containing electric charge results in a current pulse in the Subsequently, the LC network filters most of the high-
void. In addition to the flow of electrons, the ionization of the frequency components of the partial discharge pulse thereby
gas in the dielectric will precipitate positive ions moving in
the opposite direction. Furthermore, the currents due to the C
Measuring
1
PD activity create electric pulses which emanate from the C Device
cavity. Analysis of the current pulses in the frequency domain C 2
indicate that the frequency spectrum of the current pulses Low Voltage 3
span across several hundred megahertz [4]. Winding High Voltage Winding
Star or Delta
Partial discharge activities in transformer insulation come Transformer Connected Zm Zm Zm
in the form of impulses. Those impulses can be detected
during the induced voltage test through the external circuit to
the transformer terminals. Fig. 2 shows an example for this Fig. 2. Circuit diagram of the partial discharge measuring circuit for a three-
circuit that is used to detect PDs by connection to the HV phase Power transformer.

1021
distorting the signal. In addition, the coupling between have shown that the location of partial discharge activity in
windings due to resonance can distort the detected PD pulses the transformer windings can be determined using the
further. frequency spectrum of the measured partial discharge
PD testing of power transformers can be implemented activity. The ratio of currents at the HV bushing impedance
using the induced test and the independent power frequency and neutral terminal is plotted against the percentage of the
voltage source test. The induced test applies a voltage stress winding length using frequency as the argument. The
across the low voltage windings of the transformer. On the intersection of this line with the axis corresponding to the
other hand, the power frequency voltage test is generally winding length determines the PD site or the site with
applied to smaller transformers; the voltage is applied across maximum PD activity. In the following sections of the paper,
on-line PD detection techniques will be discussed.
the high voltage winding of the transformer thereby creating
a voltage stress between the windings and the ground using a
IV. ON-LINE PARTIAL DISCHARGE DETECTION
coupling capacitor[4].
TECHNIQUES
The specifications for PD testing in power transformers
prescribe frequencies below 300 kHz [5]. Meanwhile, Online partial discharge measurements have long been
previous studies show that 40 to 200 kHz is the optimal used as an effective way to assess the condition of the
frequency range to detect partial discharges in power transformers. Majority of transformers used in power systems
transformers [6]. facing a lot of problems which affects the whole system
because of the uncertainty of their insulation condition.
In the power frequency voltage method of PD detection, a
Therefore, online partial discharge detection is useful for an
filter is connected with the detection impedance in order to
insulation diagnosis and continuity of operation during
separate the PD pulses from the power frequency component
measurements with the aim to optimize both maintenance and
and its harmonics along with other interfering signals. PD
life risk management. It is necessary to implement technical
pulses corresponding to up to 500 pC are filtered using analog
methods to identify endangered equipment before
filters; extreme levels of noise are filtered using digital filters
catastrophic failures occur. Fig. 4 highlights the necessary
that utilize the fast Fourier transform to remove noise in the
steps to be taken in order to successfully detect PD on-line.
frequency domain[4].
The online partial discharge measurement consists of a two
In order to analyze a PD pulse beyond the inception
stages; partial discharge activity signal detection stage and
voltage and extinction voltage, partial discharge pulse
interpretation stage. The on-line partial discharge
distribution measurement systems should be utilized. These
measurements techniques for transformer include magnetic
systems can characterize the type of PD fault in addition to
sensing like Transient earth voltage and High frequency CT,
determining the PD fault location. The classic partial
UHF sensing, Acoustic and Ultrasonic sensing.
discharge behavior is denoted by the appearance of positive
discharges along the negative half cycle and negative A. Acoustic sensing for Partial Discharge Measurement
discharges along the positive half cycle of the applied Partial discharges in oil create ultra-sonic acoustic waves
voltage[4]. and electromagnetic waves in the UHF range in addition to
the electrical pulses[10], [11]. Acoustic and electromagnetic
B. Dissipation Factor Tip-Up
techniques of measuring PD have advantages for on-site and
The most popular method of electrical PD location in on-line applications which distinguish them from electrical
transformers is based on the principle of comparing PD pulse sensors used in the IEC 60270 standard of partial discharge
magnitudes which are measured at the winding terminals of a testing[1]. Acoustic and electromagnetic sensors can be used
transformer. The initial peak of the oscillating PD signal on-line to locate partial discharges in three-dimensional space
corresponding to each phase is measured using an
using acoustic arrival times of partial discharges[11]. The
oscilloscope. Direct capacitive coupling is related to the
transformer tank is grounded; therefore, the sensors can be
initial peak of the signal while the remaining parts of the
placed on any wall of transformer tank. An acoustic signal
signal are related to the inductance of the transformer
windings[4]. from mechanical vibration of PD can be detected by
piezoelectric transducers as shown in Fig. 3 piezoelectric
(PZT) sensors based on ferroelectric are mounted on the
∆𝑉 ∆𝑉 𝑒𝑥𝑝 𝑛 (1) transformer tank in order to detect the acoustic waves
generated by partial discharges[1].
Equation 1 describes the amplitude of the PD pulse peak. The detection of PD activity using acoustic and
∆𝑉 is the peak voltage of the PD pulse, 𝐶 and 𝐶 are the electromagnetic techniques are performed with external
shunt capacitance associated with each transformer winding piezoelectric (PZT) sensors, PZT converts acoustic energy
and the series capacitance between successive windings into a useful electrical signal. It usually mounted on the
respectively. n corresponds to the number of windings from transformer tank wall; It has narrowband detection at 150
the location of partial discharge activity to the terminal at kHz.
which partial discharge is being measured. Consequently, the The drawbacks of the acoustic and electromagnetic
transformer terminal at which maximum partial discharge is techniques are suffering from the complex nature of the
detected is assumed to be closest to the location of maximum acoustic wave propagation. The waves don’t travel in perfect
partial discharge activity. Assuming that the windings of a path because the
transformer behave similar to uniform transmission lines in a
particular frequency spectrum, Harrold and Sletten[8], [9]
1022
 Transformer signals based on the observation that sound signals of poor
Tank quality have a higher variance in energy than frequency.
B. UHF Sensor for Partial Discharge Measurement
PD Source Dissolved Gas Analysis (DGA) is a PD diagnostic
PZT technique which can be applied to power transformers in
order to characterize and quantify the partial discharge
activity; however, DGA cannot be used to locate PD
activity[13]. UHF based PD detection can be used for
Oil
periodic tests of transformers as well as continuous on-line
monitoring of transformers[13].
Ultra-high frequency-based PD detection is most often
PZT used for onsite diagnosis of partial discharges in gas insulated
switchgear. The UHF sensors are distributed around the body
Oil of a power transformer and operate at the resonant frequency
between 20 and 500 kHz [12]. Reference [13] proposes the
use of a dielectric window in power transformers in order to
Fig. 3. PZT sensors on a transformer tank. better detect PD activity using UHF sensors. In addition, the
A source of PD is so small which cause that the waves can sensors mounted externally can be installed and removed
be seen as an acoustic point source. In addition, the needed with little to no downtime of the transformer. Furthermore,
sensitivity due to the weakness and the low intensity of the the signals from the various UHF sensors are managed using
received acoustic signals which needs very sensitive devices a multiplexer which enables switching between signals from
to any small changes in signal amplitude in order to detect a multiple sensors[12]. Meanwhile, the analysis of the UHF
PD. signals corresponding to the PD activity in power
Two types of sensors are used in the acoustic techniques; transformers is implemented using a spectrum analyzer. The
the accelerometers sensors and acoustic emission sensors. spectrum analyzer is capable of analyzing both the spectra
Fiber optic (FO) sensors has been used recently to internal and the point on wave (POW) of the ultra-high frequency
placed inside the transformer to increased signal amplitude signals pertaining to PD activity[12].
and elimination of multipath interference but the position UHF based PD detection involves the amplification and
sensor is a challenge, the sensitivity of those sensors is filtering of signals at the initial stage. Analog to digital
moderate resolution and they show dependent on integration conversion is done at earlier stages given the high bandwidth
of the detection instruments available at a reasonable
technology.
cost[13]. Digital signal processing techniques are used to
The primary advantages of the acoustic detection
characterize partial discharge activity from the detected PD
technique include the capability of on-line PD detection and pulses[13]. The analysis of UHF PD frequency spectra is
the immunity from electrical interference[12]. Nevertheless, accomplished using different methods including averaging
the accuracy of PD location using acoustic sensors is much the frequency spectra, subtraction between combinations of
lower due to the intricate nature of acoustic signals. frequency spectra, division between combinations of
Traversing different media, the velocity of propagation of the frequency spectra, and analyzing the max-hold spectra and
signal varies from the source of partial discharge to the various statistical information about the frequency
piezoelectric sensor. Meanwhile, other factors including spectra[12]. The POW frequency analysis is carried out by
reflection, refraction and attenuation due to the influence of
different media on the acoustic signals render acoustic signal
processing into a complicated task. Acoustic sensors are
mounted on the surface of a transformer between the two ends
of a winding in order to determine the time difference
between the acoustic signals detected by each sensor.
Subsequently, the length traversed by the acoustic signal and
the PD location can be determined[12].

𝐷 𝑣𝑇 (2)

PD detection in power transformers using acoustic sensors

involves two main signal processing steps: signal de-noising
and arrival time determination. Arrival time determination
techniques monitor changes in either the energy or the
frequency content of the signal[1]. The arrival time of the
acoustic signal is calculated using equation 2 as mentioned in
[1] eq. (4). Di is the radius of the sphere used for PD location,
v and T are the velocity of sound and the measured arrival
time respectively. On the other hand, the energy criterion is a Fig. 4. Partial Discharge Online Measurement Diagram in Transformer.
method of determining the arrival times of acoustic PD
1023
choosing a frequency that is much higher than the noise of This makes it easier to distinguish between the pure PD single
the measured UHF PD signal. Meanwhile, PD location is and the noise; which can be eliminated easily. The advantage
accomplished using the fact that high frequency components of FFT is that it uses less number of computations than
of the UHF PD signals are attenuated faster than the low traditional discrete Fourier transform. Its shortcoming,
frequency components[12]. however, is that converting the signal from time to frequency
domain causes the loss of important signal time information.
C. Electromagnetic Sensing
C. Wavelet Technique (WT)
Magnetic sensing can be done using two methods;
transient earth voltage (TEV); and high frequency current WT is one of the strongest techniques used in PD signal
transformer (HFCT). de-noising. It involves using small pulse shaped waveform –
TEV sensor is a capacitive coupling sensor that can be called mother wavelet – that can be shifted, stretched or
attached to the wall of the transformer to detect the compressed in order to fit the shape of the PD signal. The
electromagnetic waves in the form of voltage pulses signal is then passed through a high-pass filter and low-pass
propagating through the insulation to the tank walls. filter, producing two signals called detail and approximation
Meanwhile, the HFCT is an inductive coupling sensor that is respectively. The approximation is passed again through a
fitted around the transformer tank connection to ground. This high-pass and a low-pass filter yielding a level 2 detail and
allows it to detect high frequency current pulses that are approximation. The process is repeated again for a number of
passing thought this ground wire. times called decomposition level. Thresholding is then
While electromagnetic sensing is generally more accurate performed to remove the noise part of the signal leaving the
and sensitive than other detection methods; it is prone to pure signal intact. Fig. 5 shows a block diagram describing
electromagnetic interference and noises present in the field. the process of de-noising using WT.

TABLE I
COMPARISON OF PD DETECTION TECHNIQUES
Criteria UHF Acoustic Electromagnetic
Detection Detection detection
Electrical Prone to Immune Prone to interference
Interference interference from
interference
Location More Less Accurate
Accuracy accurate accurate
On-line Supports on- Supports Supports on-line
monitoring line on-line monitoring Fig. 5. Discrete wavelet transform block diagram.
monitoring monitoring

Table I compares the acoustic PD detection and UHF PD D. Artificial neural network (ANN)
detection techniques. On the other hand, off-line PD detection ANN is a system of computing that resembles the human
techniques incur considerable costs due to the downtime of neural system. It consists of interconnected nodes called
the electrical equipment alongside the expensive neurons that are connected to inputs and outputs through
instrumentation utilized. weights. The efficiency of ANN depends on three
parameters; the network architecture; the neuron’s function
V. SIGNAL DE-NOISING AND CLASSIFICATION and the learning algorithm. A commonly used architecture in
PD de-noising is the multilayer feed-forward neural network
Acquiring the PD signal –specifically online– is usually
(MLPFNN).
not enough to judge the severity of the defect. This is due to
the presence of various types of noise that are superimposed
VI. CHALLENGES OF PARTIAL DISCHARGE DETECTION
on the PD signal. That noise can be[14]:
IN POWER TRANSFORMER
 Discrete spectral interference (DSI): form radio
transmission and communication. Partial discharge is an effective method for evaluating the
 Stochastic noise: form sparking or arcing. insulation performance of power transformers windings
 White noise: form the ambiance. insulation. This paper describes the significant challenges
This requires performing some signal processing techniques and clarifies the complexities of partial discharge detection
in order to de-noise the signal. The most common techniques for power transformers as follows;
can be summarized as follows:  PD detection methods aims to distinguish between a
number of events including PD, noise and corona based on
A. Digital filtering the signal which attenuated due to the shielding effect of
Digital filtering can be used to de-noise the signal form uniformly distributed current pulse. This makes it harder
certain types of noises that happen to exist in a different for inductive coupling devices to detect the PD activity.
frequency band than the pure PD signal. However, digital  One of the most important challenges in PD
filtering fails to de-noise certain noise types which require detection is to compromise the integrity of comparing the
more sophisticated techniques. magnitude of the pd pulse with the calibration pulse to
B. Fast Fourier Transform (FFT) obtain its “true” magnitude which known as calibrating
problem.
FFT is a discrete Fourier transform algorithm used to  The sensitivity of the detection equipment is very
convert the signal form time domain to frequency domain. important to determine the optimum position of the PD
1024
 whereas the emitted signals from PD may be of low [13] A. S. Kumar, R. P. Gupta, K. Udayakumar, and A. Venkatasami,
“Online partial discharge detection and location techniques for
magnitude.
condition monitoring of power transformers: A review 2008,” in
 PD signals are very superposed by noise pulses, International Conference on Condition Monitoring and Diagnosis,
which cause higher susceptibility to interference from 2008, pp. 927–932.
electronic noise. That makes PD analysis very difficult for [14] S. S. Refaat and M. A. Shams, “A Review of Partial Discharge
Detection, Diagnosis Techniques in High Voltage Power Cables,”
software systems.
in IEEE 12th International Conference on Compatibility, Power
Electronics and Power Engineering (CPE-POWERENG), 2018.
VII. CONCLUSION
Partial discharges are one of the main hazards which can
compromise the insulation of power transformers. PD
activity in power transformers are measured using both on-
line and off-line techniques. Given the variable nature of
partial discharge activity, continuous on-line monitoring of
partial discharges gives the most precise evaluation of partial
discharge activity in power transformers. According to [12]
none of the on-line detection methods compared above can
detect transformer oil degradation. PD activity below 1000
pC can only be detected using the electrical method[12].
Acoustic detection of partial discharged is only possible for
partial discharges above 10000 pC[12]. Given the advantages
and disadvantages of the partial discharge detection
techniques, reliable detection of partial discharge activity will
require a combination of the techniques discussed.

ACKNOWLEDGMENT
This publication was made possible by NPRP grant
[NPRP10-0101-170085] from the Qatar National Research
Fund (a member of Qatar Foundation). The statements made
herein are solely the responsibility of the authors.

REFERENCES
[1] J. M. Abdallah, “Power transformers insulation materials
investigations: partial discharge,” Int. J. Energy Power Eng., vol.
9, no. 11, 2015.
[2] S. M. Markalous, S. Tenbohlen, and K. Feser, “Detection and
location of partial discharges in power transformers using acoustic
and electromagnetic signals,” IEEE Trans. Dielectr. Electr. Insul.,
vol. 15, no. 6, 2008.
[3] R. Bartnikas, “Partial discharges. Their mechanism, detection and
measurement,” IEEE Trans. Dielectr. Electr. Insul., vol. 9, no. 5,
2002.
[4] D. M. Robinson, Dielectric Phenomena in HV Cables. London:
Chapman and Hall, 1936.
[5] “IEEE Guide for Diagnostic Field Testing of Fluid-Filled Power
Transformers, Regulators, andReactors,” IEEE Std C57.152-2013,
pp. 1–121, 2013.
[6] “IEEE Guide for PD Measurement in Liquid-Filled Power
Transformers and Shunt Reactors,” IEEE Std C57.113-1991, 1992.
[7] A. Alsheikhly, H. Guzman, and H. G. Kranz, “A new diagnostic
tool through computer simulation calculation using expanded
partial discharge equivalent circuit,” in Proceedings of the 4th
International Conference on Conduction and Breakdown in Solid,
1992.
[8] G. H. Vaillancourt., R. Malewski, and D. Train, “Comparison of
three techniques of PD measurements in power transformers,”
IEEE Trans. Power Appar. Syst., pp. 900–909, 1985.
[9] H. R. T and A. M. Sletten, “Corona location in transformers by
radio frequency spectrum analysis, Part I: theory,” IEEE Trans.
Power Appar. Syst., pp. 1584–1590, 1970.
[10] W. R. Rutgers and Y. H. Fu, “UHF PD-Detection in a Power
Transformer,” in 10th Intern. Sympos. High Voltage Engineering,
1997, pp. 219–222.
[11] M. D. Judd, B. M. Pryor, S. C. Kelly, and B. F. Hampton,
“Transformer Monitoring Using the UHF Technique,” in 11th
Intern. Sympos. High Voltage Engineering, 1999, pp. 362–365.
[12] S. M. Markalous, “Detection and Location of Partial Discharges in
Power Transformers using Acoustic and Electromagnetic Signals,
Ph.D. thesis,” University of Stuttgart, Germany, 2006.

1025