View metadata, citation and similar papers at core.ac.
uk brought to you by CORE
provided by El Servicio de Difusión de la Creación Intelectual
The 20th International Symposium on High Voltage Engineering, Buenos Aires, Argentina, August 27 – September 01, 2017
A STUDY OF OIL-PAPER INSULATION VOLTAGE DEPENDENCY
DURING FREQUENCY RESPONSE ANALYSIS
D. M. Robalino1*, J. Cheng2, P. Werelius3 and R. Alvarez4
1Megger, USA
2Megger, China
3Megger, Sweden
4Universidad Nacional de La Plata, Argentina
*Email: <[email protected]>
Abstract: People often believe that power factor/dissipation factor testing at power
frequency (50/60 Hz) usually exhibits a flat response as a function of test voltage if the
insulation is in good condition. Dielectric Frequency Response, DFR is the extension of
power factor testing except that the measurement is performed from 1 kHz down to
typically 1 mHz. It is a very useful tool for evaluating the moisture content in solid
insulation of HV and EHV components such as power transformers, bushings, instrument
transformers and PILC cables. The voltage dependent phenomenon also called “the
Garton effect”, caused by paper absorbing electric charges in oil is investigated. The
application of DFR in HV and EHV substations required a conceptual analysis of the
phenomenon to better interpret the condition of the insulation system while increasing the
signal to noise ratio to minimize the effect of surrounding interference. As a result of this
work, authors provide practical recommendations regarding test voltages and frequency
ranges to be used under high interference environments. The wide application of the
method is supported with experimental field data.
1. INTRODUCTION the solid insulation and the conductivity () or %DF
of the liquid insulation can be obtained.
Line-frequency power factor/dissipation factor
testing is one of the most common routine Nevertheless, due to the wide spectrum of
practices in factory and field to determine the frequencies used for the measurement sequence,
condition of the insulation in complex MV, HV and the applied voltage in most cases does not exceed
EHV electrical equipment. 140 Vrms. This voltage level, much lower than the
traditional 10 kV line-frequency measurement,
Manufacturers, researchers, service companies faces a new challenge: the signal-to-noise ratio is
and field operators have been using capacitance extremely low, especially at low frequencies. The
(C) and power factor/dissipation factor (PF/DF) way to overcome this challenge is to increase the
metering technology for almost 80 years looking at signal-to-noise ratio by boosting up the test
results obtained at line-frequency values voltage. The effect of such approach is discussed
(50/60Hz). The instrumentation used for this test throughout this document.
has improved gradually as power electronic
components became more efficient and lighter, but 2. THE DIELECTRIC RESPONSE
the testing practices in the field faced challenges of
increasing voltages and higher electromagnetic The dielectric response might be obtained in the
interference (EMI) environments at the substations. time or frequency domains.
In the time domain, the most common test is the
In this paper, authors focus particularly on the insulation resistance (IR) test. For IR test, the
advanced method to diagnose oil-paper insulation instrument applies a DC voltage and leakage
systems based on the dielectric response. The current is measured. Only one minute is required
dielectric response method involves measurement for IR test and 10 minutes for polarization index
of dielectric parameters such as capacitance (C), (PI) test. The PDC method is an extension of IR
permittivity (), power factor (PF) and/or dissipation testing except that the test time is longer, typically
factor (DF) as a function of frequency. The 10,000 (ten thousand) seconds.
frequency range of the dielectric response starts at In the frequency domain, line-frequency power
1000 Hz and typically stops at 1 mHz, depending factor (PF) or dissipation factor (DF or tan )
on the temperature of the insulation under test. instrumentation allows for a quick test at voltages
One of the advantages of this method is the up to 12 kV (for portable field application). The test
capability to discriminate between the properties of is intended to measure dielectric losses in the
the cellulose and the liquid insulation [1]. As a insulation and not to overstress the insulation.
result, the percentage moisture content (%mc) of Therefore, test voltages should not exceed the
specimen’s rated voltages.
An application of the line-frequency PF/DF test humidity, the DC and low frequency interference is
is the tip-up test or line-frequency power factor at more active as compared to dry environments.
different voltages. The power factor tip-up is The problem in hand is almost negligible when
defined as the difference in the power factor testing inter-winding power transformer insulation
measured at two voltages. Its application has been using a UST (ungrounded specimen test) mode.
well investigated specially in dry type HV insulation Although, a full DFR test will involve GST
where the increase of PF/DF was originally (grounded specimen test) and/or GST-guard mode
attributed to the increase of the electrical where the measurement is more sensitive to the
conductivity of the walls of the cavities subjected to noise influence.
partial discharges and also of the resistivity of the The technology available nowadays makes use
solid insulation material [2]. The power factor of noise suppression algorithms and filters to
component arising from the dielectric losses mitigate the effect of noise. These algorithms
generally changes very little with voltage; however, certainly improve the quality of measurements but
with some defects in the solid insulation, such as the challenge in the field increases when low
uncured resin sections or contamination due to capacitance specimens (i.e., bushings, HV
ionic impurities, significant space charge losses instrument transformers and HVDC transformers)
may arise leading to an increasing or decreasing require advanced diagnostics. This is easy to
PF/DF value with voltage [3]. visualize because the capacitive current is a direct
Dielectric Frequency Response (DFR) also function of voltage and frequency as described
known as Frequency Domain Spectroscopy (FDS) below:
is an advanced application of the line-frequency
PF/DF method. During DFR test, the insulation
system is subjected to a low sinusoidal voltage of I c C V 2 f C V
(1)
140Vrms and measurements are carried out at
different frequencies starting from 1 kHz down to
typically 1mHz if the temperature of the insulation Figure 2 depicts the influence of noise on the
is around 20°C. Figure1 is an example of the DFR dielectric response measurement on C1 of three
response of a liquid filled power transformer. bushings mounted on a two-winding power
transformer with a capacitance value of
approximately 300 pF.
Figure 1: Dielectric frequency response of the inter- Figure 2: Low Voltage DFR measurement on C1
winding insulation of a liquid-filled two-winding for
4. three POC type
VOLTAGE bushings
DEPENDENCE
transformer
As observed in the example of Figure 2, the
influence of noise is not constant (not even of the
3. SUBSTATION INTERFERENCE same polarity). The test on the C1 capacitance of
H1 and H2 show greater influence of the
As part of the study presented in [4], DFR surrounding interference as compared to H3. The
readings, taken in 88 randomly selected very low frequencies below 10 mHz are the ones
substations, provided average reference values of mainly affected. All frequencies above 10 mHz
AC and DC interference. The line-frequency AC show a smooth response.
interference was found to range between 0 – 100
A, and DC interference ranged between 0 – 50 4. VOLTAGE DEPENDENCE
nA. The main concern during DFR testing is the
effect of the low frequency interference and DC The voltage dependence phenomenon in solid
interference, affecting specifically the low insulation during line-frequency PF/DF test is well
frequency DFR measurements. As a matter of fact, described in [3] and [5]. The losses measured in
environmental conditions will also influence the the insulation system when a line-frequency
measurement. Under high ambient relative (50/60Hz) signal is applied, is a composite of
dielectric losses which are constant with voltage In Figure 3(b) the tip-up test shows the
and power loss due to discharges. Mathematically discharge inception voltage at about 2.5 kV. This
this can be expressed as the total conductance of effect is also observed using the so called
the system in (2) percentage voltage dependence factor (%VDF)
representing the distortion of the measured current
1
f with respect to its fundamental. The distortion in
2f
G C0 tan 0
Vm
Q sin t
0
i i
(2)
the measured current signal, as represented in
Figure 4, implies that the insulation system has lost
its linearity.
Based on (2), if no discharges occur, the amount
of charge increase on the electrodes or conductors
as a result of internal discharges (Qi) equals zero.
To visualize this effect, a narrow band DFR test at
different voltage levels carried out on epoxy type
MV equipment is presented in Figure 3.
Figure 4: Applied Voltage and measured current
through a voltage dependent insulation system
Therefore, the %VDF can be calculated in a
similar way as the total harmonic distortion (THD)
of the measured current signal:
(a) k
I
h2
2
h
%VDF 100% (3)
I1
Transformer liquids behave like weakly
dissociated electrolytes and display a linear ohmic
response when exposed to low voltages.
Considering the use of voltages greater than 140
Vrms, users should take into account the effect that
this may have on the low frequencies of the
dielectric response. In [6] Garton describes the
effect of different magnitudes of an alternating
electric field on the dielectric loss of films of
(b) insulating liquids in porous impregnated insulation.
The Garton effect is described as the charge
accumulation occurring at the interface between
Figure 3: (a) DFR obtained from MV epoxy-type paper and oil insulation due to the differences of
insulation specimen at different voltages; (b) Tip-up electrical properties of each material. The
test and measured % voltage dependence factor. accumulation speed is proportional to the field
strength that is perpendicular to the interface and
Figure 3(a) shows the effect of the applied the ion accelerating time. In the frequency domain,
voltage on the solid insulation along a frequency the alternating electric field changes its polarity
sweep. Important to notice here that even in a solid regularly and when the electric field strength is
insulation system, the DFR at 0.5, 1 and 2 kV high and the charging period is short, the charge
overlap perfectly one on top of the other. accumulation at the interface could be neglected.
As a result, when the frequency response the %mc is verified with the modelling software
reaches very low frequencies and the charging and results are summarized in Table 1.
period increases, the charge accumulation will take
place and the non-linearity effect is observed.
5. THE EFFECT ON THE FREQUENCY
DOMAIN RESPONSE
First of all, the analysis of the dielectric response
in the frequency domain (as mentioned in section
2), provides the following information:
• Percentage moisture concentration in the
solid insulation;
• Conductivity and/or dissipation factor of the
liquid insulation;
• Individual temperature correction of PF or DF Figure 6: DFR measurement at different
from any temperature value to a normalized voltages of two similar OIP bushings R & B
comparative temperature based on the unique tested simultaneously in factory (~20°C) without
and individual condition of the insulation interference.
system; and,
• Identification of non-typical responses, Table 1. DFR analysis on bushings R & B
resulting from contamination of the insulation
system.
Test Voltage 140Vrms 1400Vrms
Figure 5 describes the regions within the %DF %DF
frequency spectrum of the DFR measurement Dielectric
@60Hz & %mc @60Hz %mc
that correspond to each of the materials and Characteristics 20°C & 20°C
mathematical modelling parameters. Bushing R 0.299 0.44 0.299 0.47
Bushing B 0.3 0.41 0.3 0.42
Table 1 shows no discrepancy in the analysis of
the conditions of R & B bushings tested with LV
and HV DFR.
Finally, the work presented in [8] summarizes
the experience of the authors testing EHV
bushings mounted on 745 kV reactors. The use of
LV DFR yield only distorted curves and the only
alternative was to boost up the signal-to-noise ratio
using a DFR amplifier with 1400 Vrms output signal.
Figure 5: Frequency response of liquid filled
power transformer showing the different areas of
influence of paper, liquid and XY geometrical
modelling parameters
From the experimental work presented in [7],
capacitance and loss factor are influenced
mainly by the increase of electric field in the
frequencies between 1 and 5 mHz. Frequencies
between 1 kHz down to 5 mHz have none or
negligible effect. This particular statement is
verified testing OIP bushings in factory at
140Vrms and 1400Vrms without the influence of Figure 7: DFR test on OIP 800 kV bushing at
external interference, and results are presented 500Vrms and 1400Vrms in the field
in Figure 6.
Two high voltage bushings are tested
simultaneously in factory using the same The only way to obtain a realistic and accurate
excitation signal. The nonlinear effect is response of the EHV bushings required the use of
observed to influence the frequency domain HV DFR. Without it, the response will be
response at ~ 10 mHz. Down to this frequency, misleading and the effect of the AC and DC
interference in the substation could create an [6] C. G. Garton, “Dielectric loss in thin films of Insulating
Liquids” Proceeding Institution Electrical engineering,
uncertain bias in the modelling reference curve that Vol. 88, 1941, p. 103-120.
would yield to erratic diagnostics. [7] J. Cheng, P. Werelius, D. Robalino, D. Wang, “The
nonlinear behaviour of the transformer insulation system”,
6. CONCLUSIONS IEEE Proceedings of the Conference on Electrical
insulation and Dielectric phenomena, 2013
[8] I. Güner, D.M. Robalino, P. Werelius, “HV and EHV
The complex insulation system of liquid filled bushing condition assessment – Field experience” CIGRE
power transformers and bushings is of organic – IEC 2016 Colloquium on EHV and UHV (AC & DC),
origin and thus, susceptive to aging and Montreal, 2016.
degradation. DFR has demonstrated in the last 20
years (first field portable unit for DFR was brought
to the market in 1997) the capability to provide
valuable information not observed before with line-
frequency PF/DF test.
This paper clearly identified the effect of the
magnitude of the voltage applied to the insulation
system under test and the possible sources of non-
linearity.
In the higher frequency range, power losses due
to discharges will affect the linearity of the
response.
In the frequencies below 5 mHz, the dielectric
loss of films of insulating liquids in porous
impregnated insulation affects the linearity of the
response.
In that regard, authors recommend to:
- Perform DFR UST measurements on power
and distribution transformers on the inter-
winding capacitance at low voltage (140Vrms)
in the frequency range between 1 kHz and
typically 1 mHz (at ~20°C). For lower
temperatures, lower frequencies may be
needed.
- Perform DFR UST measurements on low
capacitance specimens (i.e. bushings C1)
with LV DFR in factory down to 2 mHz (at
~20°C). In the field, it may be necessary the
use of HV DFR and test at 1400Vrms to
overcome the effect of noise. In order to
avoid the effect of non-linearity at very low
frequencies, the HV DFR test should run
down to 5 mHz or when the harmonic
distortion of the measured signal is
observed.
- HV and EHV Bushings do not require very
low frequencies to be analysed as the
dominant material is paper.
REFERENCES
[1] S. M. Gubanski, et al., “Dielectric Response Diagnoses for
Transformer Windings,” CIGRE 2010
[2] R. Goffaux, “On the nature of dielectric loss in high
voltage insulation”, IEEE Transactions Electrical
Insulation, Vol. EI-13, No. 1, 1978.
[3] IEEE Std. 286-2000 (R2012). “IEEE Recommended
practice for measurement of power factor tip-up of Electric
machinery Stator Coil insulation”.
[4] J. Skoldin, P. Werelius, M. Ohlen, “Dielectric Frequency
Response Measurement Technology for Measurements in
High Interference AC and HVDC Substations”. 2011
TechCon Asia-Pacific, Sydney, Australia
[5] T. W. Dakin, “The Relation of Capacitance Increase with
High Voltages to Internal Electric Discharges and
Discharging Void Volume”. AIEE Transactions. Part III,
Issue 3, 1959
