Download Insulation condition assessment of medium voltage power

Insulation condition assessment of medium voltage power cables using on-site PD detection and
analysis techniques
Edward Gulski, Bjorn R. Hamerling, Frank J. Wester, Johan J. Smit, Edwin Groot(1), Peter Schikarski(2)
Delft University of Technology, High Voltage Laboratory, The Netherlands
(1)
InfraCore, NUON, Alkmaar, The Netherlands
(2)
Haefely Test AG, Tettex Instruments Division, Dietikon, Switzerland
SUMMARY
In the last ten years, to determine the insulation
condition of the distribution power cable grid,
several on-site diagnostic measuring methods
have been developed. As a result, on the basis of
on-site testing and measurements of partial
discharges, important information is generated
about discharging insulation defects, see figure 1.
In particular, applying off-line the PD diagnosis at
oscillating voltage waves has shown that PD
knowledge rules can be defined to support
condition-based maintenance, see figure 2.
Figure 1. PD mapping made using OWTS®
method for a 840m long 10kV PILC power
cable The concentration of discharges at the
120m location originates from a oil-filed splice.
To obtain maximum efficiency in applying
advanced on-site diagnosis to a large number of
different cable sections, a certain pre-selection is
necessary. This pre-selection could be done in
different ways, e.g. using failure history, position in
the network or on the basis of an on-line diagnosis,
see figure 3. In particular, to support the off-line
diagnosis by prioritising particular cable sections
on the one hand, and after an on-site inspection to
trend the discharge process in a suspicious cable
section the on-line PD diagnostics is becoming
more and more interesting. Unfortunately, in
contrast to off-line PD diagnosis, the sensitivity of
on-line PD detection as well as the interpretation of
measuring results is strongly dependent on local
external disturbances and operation conditions of
the cable section. Finally, based on systematic
laboratory and filed investigations between off-line
and on-line PD diagnosis transferability rules have
to be developed supporting field application of both
techniques.
To contribute to this investigation, this paper will
firstly describe the measurement method of the online PD diagnostic, secondly it will present
laboratory measurement results and thirdly it
handles the first application of the diagnostic in
practice and compares the results with the OWTS
measuring method. Finally there will be drawn
some conclusions with respect to the applicability
of the on-line PD discharge detection diagnostic.
Figure 2. OWTS® PD system during off-line
diagnosis of a 10 kV PILC cable section;
location Alkmaar, The Netherlands.
spectrum analyser
10 kV PILC cable
terminations and the
HFCT sensor
Figure 3. VHF PD detection system during online PD detection of a 10 kV PILC cable
section. The HFCT sensor is temporally
installed in the earth circuit of the 10 kV PILC
cable termination; location Purmerend, The
Netherlands.
Insulation condition assessment of medium voltage power cables using on-site PD detection and
analysis techniques
Edward Gulski, Bjorn R. Hamerling, Frank J. Wester, Johan J. Smit, Edwin Groot(1), Peter Schikarski(2)
Delft University of Technology, High Voltage Laboratory, The Netherlands
(1)
InfraCore, NUON, Alkmaar, The Netherlands
(2)
Haefely Test AG, Tettex Instruments Division, Dietikon, Switzerland
ABSTRACT
some conclusions with respect to the applicability
of the on-line PD discharge detection diagnostic.
In this contribution fundamental aspects of on-site
partial discharge detection on distribution power
cables as a tool for insulation condition
assessment is discussed. In particular, based on
laboratory and field tests the combination of off-line
and on-line PD diagnostics is presented.
HV Supply
Inductor
Cable
under
test
HV Switch
1. INTRODUCTION
Computer unit
In the last ten years, to determine the insulation
condition of the distribution power cable grid,
several on-site diagnostic measuring methods
have been developed [1-2]. As a result, on the
basis of on-site testing and measurements of
partial discharges, important information is
generated about discharging insulation defects. In
particular, applying off-line the PD diagnosis at
oscillating voltage waves has shown that PD
knowledge rules can be defined to support
condition-based maintenance [3-5].
To obtain maximum efficiency in applying
advanced on-site diagnosis to a large number of
different cable sections, a certain pre-selection is
necessary. This pre-selection could be done in
different ways, e.g. using failure history, position in
the network or on the basis of an on-line diagnosis
[6]. In particular, to support the off-line diagnosis
by prioritising particular cable sections on the one
hand, and after an on-site inspection to trend the
discharge process in a suspicious cable section
the on-line PD diagnostics is becoming more and
more interesting. Unfortunately, in contrast to offline PD diagnosis, the sensitivity of on-line PD
detection as well as the interpretation of measuring
results is strongly dependent on local external
disturbances and operation conditions of the cable
section. Finally, based on systematic laboratory
and filed investigations between off-line and online PD diagnosis transferability rules have to be
developed supporting field application of both
techniques.
To contribute to this investigation, this paper will
firstly describe the measurement method of the online PD diagnostic, secondly it will present
laboratory measurement results and thirdly it
handles the first application of the diagnostic in
practice and compares the results with the OWTS
measuring method. Finally there will be drawn
HV + PD
detection
V
A
t
q
t
A
Figure 1. Principals of off-line PD diagnosis of
distribution power cables using oscillating wave
test system OWTS® : Upper photo: 36 kV HV
Supply, PD analyzer (30kg), Lower photo:
inductor, HV Switch, HV+PD detection (65kg).
Spectrumanalyser & laptop
To termination
HFCT
High voltage
Earthshield
power cable
Figure 2. Principals of on-line VHF PD analysis
of distribution power cables using 50 MHz CT, a
Spectrum Analyzer controlled by a computer.
2. PD MEASURING SYSTEMS
2.1. Oscillating Wave Test System OWTS®
(off-line)
With this method, see figure 1, the cable sample is
charged with a DC power supply over a period of
just a few seconds to the usual service voltage [25]. Then a specially designed solid state switch
connects an air-core inductor to the cable sample
Figure 3. Measurement results from a 10kV
2870 m long paper/oil power cable, installed in
1960:
a) PD pattern at inception voltage;
b) PD pattern at 2*U0.
The air core inductor has a low losses factor
design, so that the resonant frequency lies close to
the range of power frequency of the service
voltage: 50Hz to 1 kHz. Due to the fact that the
insulation of power cables usually has a relative
low dissipation factor, the Q of the resonant circuit
remains high depending upon the cable: 30 to
more than 100. As a result, a slowly decaying
oscillating waveform (decay time of 0.3 to 1
second) of test voltage is applied to energise the
cable sample.
During tens of power frequency cycles the PD
signals are initiated in a way similar to 50 (60) Hz
inception conditions, see figure 3. Since the
oscillating frequency represents the AC conditions
of the power line frequency, the measurement
bandwidth of the PD circuit has been chosen in
accordance with IEC 60270 recommendations. For
the purpose of location by travelling waves, the
bandwidth is increased up to several MHz. In
combination with a 100 MHz digitiser and
depending upon the cable type, e.g. XLPE or
paper-oil, the detection and location of PD pulses
remains sensitive for cable lengths of few
kilometres, see figure 4.
PD spectrum
Noise spectrum
3-core PILC; 945m; 1991
Cable A (PD activity upto U0)
8000
PD magnitude [pC]
7000
6000
5000
4000
3000
2000
1000
0
0
190
380
570
760
950
cable length [m]
Figure 4. PD phase-resolved pattern and the
PD mapping of discharge activity in the PILC
cable insulation due to mechanical damage at
85m location.
in a closure time of <1µs. Now series of voltagecycles starts to oscillate with the resonant
frequency of the circuit: f =1/( 2π * √L*C ) ,
where L represents the fixed inductance of 0.8 H of
the air core and C represents the capacitance of the
cable sample.
PD activity
Figure 5. Laboratory results of VHF PD
measurement of discharge activity in a XLPE
cable termination: artificial defect.
The PD signals, which are ignited during one or
more oscillating waves and are detected by the
system, can be processed for two purposes. Firstly,
each of the PD pulses can be analysed for
reflections using travelling wave analysis (figure 3).
Statistical evaluation of PD signals obtained after
several oscillating waves can be used to evaluate
the location of discharge sites in the power cable
(figure 4).
ms finally resulting in a phase-resolved PD pattern,
known from off-line PD detection and recognition.
2.2. VHF PD Analyzer (on-line)
In contrast to off-line PD diagnosis, where the
cable sample is energized by an external power
supply by applying an on-line diagnosis the cable
sample is energized by the power line frequency of
the network.
To detect PD pulses and to suppress the external
disturbances present in the network a VHF PD
detection system can be used [7]. In particular, a
high frequency split-core current transformer
(HFCT) connected to the earth of the cable section
has been used. This 5 MHz HFCT is connected
through a coax-cable to the spectrum analyzer
(HP-8590L). To control the SA and to process the
data, the SA was connected via a serial data-cable
to a computer, see figure 2.
Using spectrum analysis [8] the distinction
between external disturbances and cable section
internal PD have been obtained, see figures 5-6.
In particular, using difference spectrum and by
tuning the SA to frequency ranges where the
presence of disturbances was minimal, typical
phase-resolved PD pattern could be measured and
analyzed for the presence of internal discharges,
see figures (5-7). For this purpose, the SA which
Figure 7. Laboratory results of a PD phase resolved PD pattern measured at a PD free
cable section.
coil (HFCT)
termination
high voltage
cable
splice
spectrumanalyser
Test
transformer
IEC 60270 coupling
capacitor 1nF
PD spectrum
Noise spectrum
Figure 8. The measuring set-up as used in the
laboratory for VHF PD detection system. Using
5m and 3m long XLPE cable samples, different
artificial defects (typical examples of bad
workmanship) have been investigated.
3. LABORATORY MEASUREMENT RESULTS
PD activity
Figure 6. Laboratory results of VHF PD
measurement of discharge activity in a XLPE
cable splice: artificial defect.
has to been set to the ‘zero-span’ mode on a
certain center frequency (CF), triggers to the 50 Hz
voltage wave of the cable and makes sweeps of 20
The PD activity of two different cable samples was
measured in the way as described above (figure
8):
1. a bad installed field grading in a splice
2. interruption between the semi-conductive layer
and the field grading of the termination.
Using IEC60270 PD detection at test voltage <
10kV the presence of discharges was confirmed.
Using the VHF PD detection system the discharge
activity has been also confirmed. In figure 5-6 the
frequency spectra as well as the phase-resolved
PD patterns of the PD activity are shown. It follows
from these results that:
1. the PD-activity only appeared in the frequency
spectrum range from about 10 kHz to 10 MHz
2. the best signal to noise ration was gained
when the SA resolution bandwidth was set to
100 kHz.
To be sure that the increase (in the range of about
1 MHz to 5 MHz) in the frequency spectrum is
caused by internal PD activity, the phase-resolved
PD pattern has been used. As a result, cable
internal PD-activity shows stable pattern occurring
at several frequencies, see figures 5-6.
To check the sensitivity of the VHF PD detection a
10 kV XLPE cable containing discharges was
stressed at different voltage levels. The PD-level
was detected by means of an IEC 60270 PDmeasurement system. As a result, using VHF PD
detection a sensitivity starting at 200pC…400pC
has been obtained. These values of 200 to 400 pC
is obtained by doing several measurements and
calculating the mean.
Figure 10. OWTS PD system during off-line
diagnosis of a 10 kV PILC cable section;
location Alkmaar, The Netherlands.
spectrum analyser
For of-line measurements the set-up as shown in
figure 10 has been used. For this purpose the
cable was disconnected from the network and the
OWTS unit consisting HV coil was connected to
the cable section.
The results of testing several cable sections using
both measuring methods confirm that when the online VHF PD detection system did not detect any
PD activity also using off-line OWTS method no
PD activity has been observed.
On the contrary, the on-line presence of PD phaseresolved patterns could also be confirmed by the
off-line OWTS method. As an example, the results
of an 849m long PILC 10kV cable section
consisting discharges in a oil-filled joint at 120m
location are shown.
1. On-line using VHF PD detection system; for
measuring results example see figure 11
2. Off-line using OWTS® 25 analyzer; for
measuring results example see figure 12
It follows from these results that the presence of a
particular discharge site, e.g. oil-filled joint
produced discharges can be detected using both
methods. In particular, using on-lie detection rough
indication about discharging site is provided
whereas using off-line detection the location as
well as the severity of the discharges can be
determined.
10 kV PILC cable
terminations and the
HFCT sensor
Figure 9. VHF PD detection system during online PD detection of a 10 kV PILC cable
section. The HFCT sensor is temporally
installed in the earth circuit of the 10 kV PILC
cable termination; location Purmerend, The
Netherlands.
4. PRACTICAL MEASUREMENT RESULTS
For the on-line PD VHF field measurements
detection circuit as shown of figure 2 has been
used. In figure 9 the on-lie connection of the PD
detection system to the cable section is shown. In
particular, the location of the HFCT sensor was of
importance to have sensitive PD coupling path.
Figure 11. VHF PD on-line detection measuring
results obtained on an 840m long 10kV PILC
cable showing discharges in a oil-filed joint, see
also figure 12.
(a)
(b)
As a result, combining the less sensitive on-line PD
detection and high sensitive off-line diagnosis e.g.
OWTS can be of benefit by insulation condition
assessment of distribution power cable networks.
In particular, in combination with failure history and
the position in the network, the on-line PD
detection could be useful by prioritising particular
cable sections for an off-line in-depth analysis.
To complete this work, more systematic
investigation is necessary resulting in the following
knowledge.
1. Transferability rules for both methods e.g.
voltage
and
momentary
power
load
dependence of discharging sites, and their online and off-line detection ability.
2. On-line detection criticality limits for applying
off-line diagnosis.
3. Based on 1) and 2) PD off-line diagnosis
knowledge rules for condition assessment of
cable insulation.
6. REFERENCES
Figure 12. OWTS off-line detection measuring
results obtained on an 840m long 10kV PILC
cable, see also figure 11:
a) PD phase-resolved pattern obtained at 10kV
b) PD mapping showing PD concentration of
7.000 pC at the 120m location (oil-filled splice).
5. CONCLUSIONS
In this study two diagnostic methods for distribution
power cables in service have been investigated.
An on-line PD detection has been studied and
compared with an off-line PD diagnostic method.
Based on this study with regard to on-line PD
detection the following can be concluded:
1. To detect on-line PD activity in distribution
power cable networks, the VHF detection
technique through the earth circuit of the
power cable is possible.
2. To suppress disturbances and noises present
during on-line PD measurements spectrum
analysis can be used.
3. To detect PD signals of discharging defects
the phase-resolved patterns can be used.
4. The sensitivity of an on-line VHF PD detection
circuit is in the range of few hundreds of pC’s.
Comparing both methods provides the following
conclusions:
1. Using both methods no PD have been
detected in discharge free cable sections.
2. Using both methods discharges could be
detected in cables containing discharging
sites: e.g. bad workmanship in XLPE
accessories, aging effect in PILC accessories.
[1] R. Plath, W. Kalkner, I. Krage, Vergleich von
Diagnosesystemen zur Beurteilung des
Alterungszustandes
PE/VPE-isolierter
Mittelspannungskabel, Elektrizitätswirtschaft,
Jg. 96 (1997)
[2] E. Gulski, J.J. Smit, P.N. Seitz, J.C. Smit, PD
Measurements On-site using Oscillating Wave
Test System, IEEE International Symposium on
EI, Washington DC, USA, June 7 - 10, 1998
[3] E. Gulski, F.J. Wester, J.J. Smit, P.N. Seitz, M.
Turner, Advanced PD Diagnosis of MV Power
Cables using Oscillating Wave Test System,
IEEE EI Magazine March/April 2000, Vol. 16
No.2
[4] E. Gulski, F.J. Wester, P. Schikarski, Insulation
Condition Assessment of MV Power cables
Using PD Detection And Analysis At Oscillating
Voltages, paper 2.3 at ERA HV Plant Life
Extension Conference, Brussels, 23-24
November, 2000
[5] F.J. Wester, E. Gulski, J.J. Smit, Condition
Based Maintenance of MV Power Cable
Systems on Basis of Advanced PD diagnostics,
paper in this proceedings
[6] C. Walton, Regulatory and Business Drivers For
Reducing Underground Cable Failures, paper
3.6 at ERA HV Plant Life Extension Conference,
Brussels, 23-24 November, 2000
[7] J.P. Zondervan, E. Gulski, J.J. Smit, T. Grun, M.
Turner, A New Multipurpose Partial Discharge
Analyzer For On-site and On-line Diagnosis of
HV Components, 11th ISH, London, 23-27
August, 1999
[8] S. Meijer, E. Gulski, J.J. Smit, R. Brooks, Digital
Analysis of Multiple Faults in GIS, 1999 IEEE
Iner. Symp. On EI, Arlington, USA, June 7-10,
1998