PD MEASUREMENTS ON POWER CABLE TERMINATIONS

Mohan Watwe
Doble Engineering Company

ABSTRACT
Cable network systems in the power system are one of the most important part but also the
part most vulnerable to failure. More than half of cable failure stems from electrical reason
and the rest of them are due to external non-electrical inference. The main reasons for
failures on newly installed cables are-bad workmanship or damage caused during
manufacture, transportation or installation. Most defects observed in the cable accessories
results in partial discharges (PD) under the AC stress. Sensitive PD measurements can give
better understanding to assess the health of cable system

There are many diverse measurement methods which can be used to assess the condition of
high voltage cable at site. Each one of them has its own advantages and disadvantages and it
depends on the actual condition and defect which method is appropriate.

This paper describes the UHF and Acoustic PD measurement method to measure PD in the
cable accessories for EHV-XLPE cable systems in on line condition at site.

INTRODUCTION

Cable failure probability usually follows the so called bathtub curve where the failure
probability is high at the beginning and at the end of cable life. The rise of failure probability at
the end of cable’s life is mainly caused by cable aging. From statistics it is found that most of
the faults happen in cable joints and terminations. The addition of terminations and joints can,
if not properly installed, introduce PD into the cable system.

Most of the defects observed in the cable accessories results in partial discharges (PD) under
the AC stress. Sensitive PD measurements can give better understanding to assess the
health of cable system.

All power cables with polymer insulation are tested in the factory according to IEC60840 and
IEC62067 under ideal conditions (no environmental influences) and with calibrated PD
measuring systems in a shielded room. Similar conditions do not exist at site and
electromagnetic disturbances become a major issue to carry out measurements as per IEC
60270 at site.

Non-conventional partial discharge methods have since been established for condition
assessment of power cable insulation. This paper discusses applications that employ UHF
and Acoustic sensors, suitable for HV XLPE cable terminations like transformer sealing ends,
cable terminations for GIS /metal-clad substations and outdoor sealing ends based on field
experiences during investigations at site.

UHF-Partial Discharge measurement

The test procedures recommended in the relevant standards IEC 60270 and IEC 60885-3
(conventional PD measurements) are not practical in an energized substation. The required
sensitivity is lost in the background noise due to the low signal-to-noise (S/N) ratio. This is
because the upper measuring frequency is limited to 500 kHz and below by IEC standards
and this is where substation noise is most prevalent.

From a physical point of view, the S/N ratio can be improved by increasing the measuring
frequency significantly higher than 500 kHz and by using either a frequency selective signal
processing or ultra wide band signal processing. Therefore the UHF PD measuring
technology is increasingly being used as an alternative

© 2012 Doble Engineering Company -10th IndiaDoble Power Forum

All Rights Reserved
However, there is a direct relation between frequency and signal attenuation. Due to the
strong attenuation of the higher frequency spectrum of PD pulses, the PD sensor needs to be
placed in close proximity to the PD source. This means that the PD decoupling device needs
to be installed in location where the likelihood of PD occurring is the greatest. In a solid
dielectric cable system, those locations are cable accessories (terminations and splices).
Therefore, the UHF method is an appropriate method to check for correctly assembled power
cables accessories such as joints and terminations.

The UHF-Partial Discharge Measuring (PDM-) system is consisting of the ultra-high

frequency measuring instrument LDS-6/UHF or PD Smart UHF, the pre-amplifier LDA-5/GIS
and the UHF-PD current sensor (UHF-PD-coupler).

A schematic diagram of the test set-up is reported in figure 1.

PDM SYSTEM
LDS-6 UHF
OR
PD SMART-UHF

Fig. 1: PD testing circuit acc. UHF-PD measuring method (non-conventional).

The attachment of the UHF-PD-coupler to the cable termination is evident from figure 2.

Fig. 2: The UHF-PD-Current Coupler connected to the cable termination

(Barrier: Housing - Power Cable)
Specs of UHF Coupler

Lower cut-off frequency: < 100 MHz

Upper cut-off frequency: > 1 GHz
Insertion loss (mean): -10 dB (@ 50 Ohm source and load)

© 2012 Doble Engineering Company -10th IndiaDoble Power Forum

All Rights Reserved
In order to capture the very fast electromagnetic transients of PD events the lower section
of the cable termination has been bridged by a measuring loop which is attached to the
UHF-PD-coupler. Therefore PD defects close to the PD decoupling loop are well
detected, whereas noises and even PD events far from the decoupling point are strongly
attenuated, this ensures an excellent S/N ratio.

This arrangement can be considered as a short circuit between the grounded parts of the
GIS enclosure and the cable termination, for the lower frequency range. In the UHF range,
however, the resulting impedance becomes about 50 Ω, which ensures an excellent S/N ratio
and thus a high PD detection sensitivity even under noisy on-site condition.

The output of the UHF-PD-coupler is connected via a coaxial cable to a preamplifier installed
close to the UHF sensor. Via a coaxial cable the signal is fed to the PD detector.

For PD pulse processing the PDM instrument is equipped with an UHF-Processing unit with
functionality similar to a spectrum analyzer. The measuring frequency in the zero span mode
of the UHF Processing unit is adjustable with a frequency bandwidth of 8 MHz in the
frequency range of 110 MHz to 900-1700 MHz

For the measurements in this mode various measuring frequencies having better S/N ratio
and signal strength are selected. The acquisition as well as the storage and the visualization
of the PD data are performed by means of the computer-based PD measuring system.

Voltage Synchronization

The decoupled PD signals from the PD coupler can be synchronized to the main power
supply of the measuring device (50 Hz line frequency).

PD Signal Evaluation (Non-Conventional)

In the recorded measuring files all necessary information is available. This permist all display
and diagnosis facilities from a rough overview about the PD magnitude up to the derivation
of expert system analysis. The PD Signal evaluation strategy is schematically shown
in figure 3.

Fig. 3: PD signal evaluation strategy: Φ- phase, u - test voltage, q – apparent charge,

H - number of pulses

In figure 4 the basic signals for the PD analysis resp. PD pattern evaluation is shown

Fig. 4: Phase Resolved PD classification

© 2012 Doble Engineering Company -10th IndiaDoble Power Forum
All Rights Reserved
Test Procedure

• Setup the UHF Partial discharge measuring (PDM) system (LDS-6/UHF)

• Voltage synchronization is achieved by triggering on the internal power supply of the
PD detector
• PD detector measuring frequency is adjusted to obtain a better signal to noise ratio
• Function of the measuring circuit including PDM instrument is verified by functional
check procedure: injecting a voltage signal of 500 mV (or 300mV) into each PD
current coupler
• Disconnecting the signal generator, capturing and recording of the measurements

Test Sequence

The above steps were carried out at 5 different measuring frequencies per measuring point.
All feeders were in online state, i.e. the test was carried out with normal operating voltage.
A measurement was done with the PD current sensor without connecting to the cable
termination was done to record the ambient noise at one location.

Finally the evaluation of measured data using analysis software (PRPD Pattern Recognition
and Replay-Mode) was carried out.

Measuring Results- Transformer End

The measuring recordings are listed below as the PRPD Pattern (Phase Resolved Partial
Discharge Pattern) - diagrams. The recordings are quantified in mV.

R PHASE Y PHASE B PHASE

Transformer I- fc 560 MHz

R PHASE Y PHASE B PHASE

Transformer II- fc 560 MHz

R PHASE Y PHASE B PHASE

Transformer III- fc 560 MHz
Fig. 5: Phase Resolved PD pattern
© 2012 Doble Engineering Company -10th IndiaDoble Power Forum
All Rights Reserved
Observation:

For the cable sealing ends at transformer I, no typical PD patterns were observed despite of
many signals captured. There are no significant differences between the phases neither in
phase position nor in magnitude.

At transformer II, no typical PD patterns were detected. For different measuring frequencies
partly different patterns were observed, whereas magnitudes and patterns were comparable
between the phases.

At the sealing ends of transformer III the detected signals are with lower magnitude. The
behavior of all phases is also comparable. No dedicated PD patterns have been observed.

Measuring Results- GIS End

R PHASE Y PHASE B PHASE

GIS -I fc 560 MHz

R PHASE Y PHASE B PHASE

GIS –II- fc 560 MHz

R PHASE Y PHASE B PHASE

GIS –III- fc 560 MHz
Fig. 6: Phase Resolved PD pattern

The overall signal activity in the GIS substation was found less than at the transformers
side.

For feeder 1 there was partly different behavior through all measuring frequencies.
The signal magnitudes of the three phases are comparable and no typical PD
patterns have been recorded.

© 2012 Doble Engineering Company -10th IndiaDoble Power Forum

All Rights Reserved
At feeder 2 the activity was found equal or less with signal phase position at the single
frequencies comparable to feeder 1. Measurements are comparable between phases and did
not reveal typical PD patterns.

Even fewer noise signals were detected at feeder 3. No PD patterns became apparent.

There was no indication for dangerous partial discharges found on the sealing ends in the
GIS substation.

Acoustic -Partial Discharge measurement

The DFA100 instrument analyses the electric signals from an ultrasonic sensor. This sensor l
detects the acoustic signals emitted from partial discharges in cable terminations and joints.
The patterns of the acoustic signal indicate PD activity in the cable termination. Partial
discharges in the termination are discovered when there is distinct phase correlation of the
acoustic signal (phase correlation to 50Hz and 100Hz).

The instrument gives results that show the instants where the discharges appear relative
to the phase of the If a synchronizing signal from a 50 Hz power supply in the station is used,
the power frequency voltage. Thus, amplitude vs. phase plot is
obtained as illustrated in Figure 1.2. In the given example partial discharges appear
more or less concentrated in two parts of the power cycle, i.e. in both half cycles.

Phase plot measured on a cable termination with the DFA100

Figure 7

Below is a picture where piezoelectric (acoustic) sensor was placed in phase R. cable
termination on a 110 KV GIS and relevant acoustic patterns captured by DFA100 instrument.
The DFA100 instrument was synchronized to power grid by using the Synchronization Box
during the survey.

Placement of acoustic sensor on cable termination phase R – point 1

Figure 8
© 2012 Doble Engineering Company -10th IndiaDoble Power Forum
All Rights Reserved
 DFA100 plot - Phase synchronization mode for point 1
Figure 9

Measurement at 110 KVGIS

Cable End

Placement of acoustic sensor on cable termination phase R – point 2

Figure 10

DFA100 plot - Phase synchronization mode for point 2

Figure 11

© 2012 Doble Engineering Company -10th IndiaDoble Power Forum

All Rights Reserved
Figure 9 and figure 11 presents Phase Synchronization plots captured by the DFA100
instrument. Both plots show two “clouds” of dots which occurrences is related to power
frequency. This is a clear indication of PD source inside the termination. The magnitude of the
signal shows that the PD energy is relative low. However, acoustic signal could be attenuate
inside cable termination and only fraction of its energy reaches enclosure.

During on-site test, cable terminations in two left phases Y and B were also tested.
Neither of them show acoustic signal indicating PD. For phase R it was recommend to repeat
the acoustic test after some time (2, 3 months) in order to look for trends of magnitude of the
acoustic signal and to assess the risk of failure.

CONCLUSION

As can be seen from the different case studies as above, the unconventional PD monitoring
technology at power cable accessories has been successful as a valuable monitoring and
diagnostic tool for assessing the health of HV- and EHV power cable terminations.

It can be seen that undesirable failures can be prevented if suitable measuring methods are
applied for commissioning- and inspection tests at power cable systems.

REFERENCES

[1] Eberhard Lemke “Practical Experience in PD Sensing of Power Cable Accessories in the
VHF/UHF Frequency Range” ICC Fall 2010 Meeting Scottsdale / Fountain Hills, ( 2010)
[2] Instruction Manual, The Doble Type MEU 2500-Volt Portable Insulation Test Set,
Copyright, 1987, Doble Engineering Company
[3] Partial discharge measurements. IEC 60 270 (2001)
[4]Test methods for partial discharge measurements of extruded power cables. IEC 885-3
(1989)
[5] E. Lemke, H. Elze, W. Weissenberg: Experience in PD diagnosis tests of HV cable
terminations in service using the ultra-wide band PD probing. 14th ISH Deft (2003)
[6] E. Lemke, T. Strehl, W. Weissenberg, S. Markalous: Ultra-Wide-Band PD Diagnostics of
Cable Terminations in Service, IEEE Transactions on Dielectrics and Electrical Insulation
Magazine, 2008
[7] Matthias Boltze, Sacha. Michel Markalous, Alain Bolliger, Omar Ciprietti, Javis Chiu” On-
Line Partial Discharge MonitoringAnd Diagnosis at Power Cables; 76th Annual International
Doble Client Conference-2009

© 2012 Doble Engineering Company -10th IndiaDoble Power Forum

All Rights Reserved