21, rue dArtois, F-75008 PARIS

http : //www.cigre.org

B1-201

CIGRE 2012

RECENT EXPERIENCES WITH AC HIPOT & PD COMMISSIONING TESTING OF

XLPE CABLE SYSTEMS RATED 69KV AND ABOVE
Mark. FENGER, Mark CREDLAND, Howard SEDDING
Kinectrics Inc
CANADA

SUMMARY
Over the last decade and a half, thousands of kilometres of installed XLPE HV and EHV cable
systems have been subjected to after-laying commissioning testing prior to energization.
The
commissioning test usually consists of a combination of AC HiPot & Partial Discharge testing. Partial
discharge testing, in particular, gained acceptance as a valid diagnostic tool for condition assessment
of cable insulation. It is well known and understood that the results obtained from a partial discharge
test depend not only on the conditions under which the test was performed but also on the test
equipment itself including the type of sensor used and its location. The issues related to actual
discharge, induced charge and measured apparent charge are also well understood. Yet, for
commissioning testing of solid dielectric, test specifications often reference magnitude levels in pC of
apparent charge as the only pass/fail acceptance criterion. The practice is based on factory acceptance
testing of individual cable components. This paper outlines the difference between PD testing
performed on individual components in the factory and PD testing performed on installed systems in
the field with respect to magnitude calibration of PD pulses. The paper also proposes how PD tests
can be applied as a valuable tool for commissioning tests despite issues related to non diagnostic
nature of PD magnitudes. The paper further provides statistical summary of tests performed on more
than 2000km of HV and EHV cable systems including failure rates of accessories and PD occurrence
rates. The paper also provides several case studies of PD detected in cable accessories during AC
HiPot commissioning testing.

KEYWORDS
Partial Discharges, Pulse Propagation, Calibration, Sensitivity Assessment.

mark.fenger@kinectrics.com

INTRODUCTION
Partial discharges occur in the bulk of high voltage insulation materials where local electrical field
conditions are sufficiently high to sustain PD activity. In the case of extruded cables (EPR or XLPE
cables) partial discharges typically occur in cavities at the conductor shield, cavities in the insulation
due to shrinkage or gas-formation, near defects in the insulation shield, near loosely bound solid
particles in the insulation, at protrusions, at splinters or fibers or near contaminants in the insulation
shield. In cable joints or terminations, partial discharges typically occur along dielectric interfaces,
along stress interfaces, in cavities near the conductor or insulation shield due to, for instance,
misalignment during installation or thermal movement as a result of normal operation. Finally, partial
discharges may also occur within the cable insulation itself around mechanically degraded spots and or
impurities resulting in the formation of electrical trees.

ACTUAL, INDUCED AND APPARENT CHARGE

Partial discharges are a high frequency phenomenon. Fundamentally, whenever a partial discharge
occurs internal to a cable section or a cable joint, a charge and consequently a high frequency current
- is induced on the cable conductor (high voltage electrode) and the cable shield (grounded electrode).
The magnitude of a measured partial discharge signal depends partly on the magnitude of the partial
discharge current itself, i.e. the higher the actual partial discharge current the higher the induced
currents, and partly on the radial proximity of the partial discharge location relative to the cable
conductor, i.e. the closer to the cable conductor, the higher the induced current on the conductor [2, 3].
The relationship between the induced charge on the high voltage electrode and the actual discharge it
self may be evaluated by the - function [2, 3]. The rise-time of the induced partial discharge current
is similar to the frequency of the actual partial discharge current itself whereas the fall time of the
induced current is dependent on the impedance of the measuring system itself [2]. In this case, the
impedance of the cable system is defined not just by the series impedance of a PD sensor and monitor
but also in the impedance of the part of the cable system a given PD pulse must travel through prior to
being detected. The rise-time of the partial discharge current itself depends on path and velocity of the
partial discharge (avalanche) and, consequently, the frequency depends primarily on (1) the strength of
the electrical field in the void (the higher the strength of the electrical field the higher the velocity of
the avalanche itself, the faster the rise time of the PD current and the higher the frequency of the PD
current) and (2) the size of the void relative to the direction of the electrical field (the longer the void,
the longer the duration of the PD pulse, the longer the rise time of the PD current and the lower the
frequency of the PD current). Also, local conditions in the void such as gas pressure, temperature,
void surface conductivity has an impact on the partial discharge behavior.
In addition, as the induced PD currents propagate through the cable towards the cable ends, they are
subjected to primarily attenuation and, to a lesser extent, dispersion [10]. In addition, it should be
noted that the magnitude of high frequency currents propagating along the coaxial cable line further
deteriorates as a result of impedance mismatch due to joints and cross bonding.
It can thus be intuitively seen that for shorter cable runs induced currents as a result of partial
discharge activity may be readily detected via a terminal measurement, i.e. via a capacitive or
inductive sensors connected to the conductor or shield at the end of a cable. For longer cable runs,
attenuation will prevent the measurement of inducted currents related to PD activity occurring from
the opposite cable and thus a distributed PD measurement must be performed. A distributed PD
measurement refers to the scenario where sensors are connected to joints and splices throughout the
length of the cable. Often, external clamp-on radio frequency current transformers are placed around
the bonding lead connecting a joint to the link box (cross bonding box). Also, in some instances,
joints may be equipped with integrated partial discharge sensors.
Furthermore, it was previously concluded that a framework for performing a meaningful sensitivity
assessment on a field installed transmission class cable would fulfill the following criteria [11]:

1. Does not attempt to correlate discharge pulses measured in mV or mA to charge associated

with the partial discharge current itself.
2. Does not rely on averaging to increase the signal-to-noise ratio.
3. Does not rely on narrow band-pass measurements to increase the signal-to-noise ratio.
4. Does inject pulses that have rise times and magnitudes similar to the currents induced by
typically partial discharge pulses.

THE MEASUREMENT OF APPARENT CHARGE

It is well understood that IEC270 applies to a narrow-band, low frequency partial discharge
measurement performed on a HV component acting as a lumped capacitance [10]. Voltage signals
may be injected via a coupling capacitor thus providing a relationship between actual, detected pulse
magnitude as measured in mV and apparent charge as measured in pC. When injecting a voltage of
magnitude Vinj through a coupling capacitor of value Cinj, the charge injected, Qinj, into the system can
be calculated via the well known capacitor formula
Qinj Vinj C inj

(1)

with the calibration constant K being calculated as

K

Qinj
V meas

V inj
V meas

C inj V inj C inj

1
V meas

(2)

This approach involves three major assumptions:

1. The test objective is a pure, lumped capacitance
2. The duration of the injected pulses is such full polarization is achieved (for the injection capacitor)
3. All of the charge on the capacitor is transferred to the terminal of the cable under test.
With respect to (1): No cable objective is purely capacitive. Even for shorter lengths of cable systems,
for instance subjected to type tests in the laboratory, the inductive component plays a factor with
respect to HF pulse propagation.
With respect to (2): If a capacitor is driven with a time-varying voltage that changes rapidly enough,
then the polarization of the dielectric cannot follow the signal. As an example of the origin of this
mechanism, the internal microscopic dipoles contributing to the dielectric constant cannot move
instantly, and so as frequency of an applied alternating voltage increases, the dipole response is limited
and the dielectric constant diminishes. Thus, at higher frequencies, the change in dielectric constant,
referred to as dielectric dispersion, should be considered.

Qinj Vinj Cinj ( f )

(3)

With respect to (3): If the impedance of the cable under test is significantly higher than the impedance
of the injection capacitor, nearly full charge transfer is achieved. If not, the equivalent circuit acts as a
voltage divider and the charge division needs to be considered accordingly.
In addition, it should be understood that an IEC 60270 style calibration relates only to the apparent
charge on the cable terminal for a pure, lumped capacitive test object. Due to attenuation of high
frequency pulses traveling through coaxial conductors, the apparent charge on the sheath or conductor
at the discharge spot higher than the measured apparent charge on the terminal [10, 11].

Lastly, an installed sectionalized transmission class cable constitutes a distributed capacitive load and
not a lumped capacitive load. For a lumped capacitance, it can be assumed that the difference in
apparent charge as measured on the terminal and apparent charge at the location of the PD source is
minimal as only attenuation along of a shorter length of cable needs to be considered. For a
distributed capacitance, the difference between apparent charge as measured on the terminal and
apparent charge at the location of the PD source will be further amplified due as a result of signal
losses due to stray capacitances, grounding etc. Thus, for distributed capacitive systems, such as field
installed cables, to ensure optimal sensitivity, PD measurements should be performed at each
accessory. Thus, the apparent charge may be measured at or very close to the PD source location
itself. Should a system allowing for measurement and storage of individual signal waveforms the
apparent charge of a PD signal may be evaluated by integrating up the appropriate area underneath the
PD signal

QPD

t2

R
t1

V PD (t )

(4)

PD

Where RPD is the restive input impedance of the PD monitor (typically 50 Ohms) and t1 and t2 define
the beginning and end time of the peak associated with the apparent PD pulse itself.
To preserve the waveform of a PD signal thus preserving information on the type and nature of a PD
source, modern partial discharge systems make use of broadband acquisition technology and not
narrow band technology. Consequently, IEC270 does not apply for field measurements performed on
installed transmission class cables with broad band partial discharge measurement systems.

t1 t2

QP

Figure 1: Example evaluation of apparent charge

Previously, it was shown that any a assessment on any system provide only for a normalization of the
sensor and partial discharge monitor as defined by H PD.Sensor ( s) H PD.Monitor ( s ) as loaded into the test
impedance and does not provide any meaningful information on the sensitivity of the PD measurement
itself [12]. Furthermore, it was shown that for broadband PD measurements, the frequency response
of the system should be characterized for the entire frequency domain for which the PD sensor and PD
instrument is able to detect signals. In other words, the injection pulses used for the sensitivity
assessment should match the frequency response of the sensor, the frequency response of the partial
discharge monitor used and match the frequency content of partial discharge pulses from known
partial discharge sources [12].
Therefore, taking these factors into account PD magnitudes are not diagnostic of nature and a given
discharge level as evaluated in mV or pC does not correlate directly to the health of the insulation
system under test. As a consequence, from a scientific point of view, an acceptance criterion for cable
system should not be based on a PD level alone but on a number of additional criteria see Section
IV.

TESTING PROCEDURE & ACCEPTANCE CRITERIA

For the 15 years or so, a combination of AC HiPot & Partial Discharge Testing has gained ground as
an after-laying acceptance test for XLPE cable systems. Cables are subjected to an AC HiPot
withstand in according to IEC 60840 and IEC 62067. Often, during the AC HiPot test, a partial
discharge test is performed. For transmission class cable systems, due to the issues related to
attenuation, reflection at joints and signal loss due EM coupling, a distributed PD test is often
performed.
Table 1 provides a quantative summary of AC HiPot & PD tests performed on a XLPE cable
population rated 115kV and above containing approximately 1,715 km of XLPE cable rated 69kV and
above, 312 individual phases, 2,654 joints and 624 terminations is given.
For all the cable systems tested no break-down of a cable termination was observed during the AC
HiPot test.
Rated Voltage
[kV]
66/69/72
115
132/138/161
220/230
345/400

No of Terminations
44
106
42
108
100

No of
Joints
156
30
335
1,317
972

Joint Failure
Rate
0.00%
3.33%
0.60%
1.81%
0.35%

Joint
PD*
1.28%
0.00%
0.90%
0.84%
0.31%

Termination
PD*
0.00%
0.94%
4.7%
7.4%
1.00%

Table 1: Failure & PD Occurrence Rates (*but no failure)

For this study, 27 individual phases rated 115kV (81 terminations and zero joints) as well as 9
individual phases rated 138kV (18 terminations, 156 joints) were not subjected to PD measurements.
All accessories for phases rated 161kV and above were subjected to PD measurements. For the
remaining phases, PD testing was performed during the AC HiPot on 32 individual phases
immediately following a successful AC HiPot and at a reduced voltage level of 1.2U0. For the
remaining phases, partial discharge testing was performed on all accessories during the AC HiPot.
For 132/138kV and 220/230kV class cables, AC HiPot testing was performed at 1.7U0 whereas for the
345kV/400kV class cables AC HiPot testing was performed at 1.25U0 and 1.2U0 respectively where
1U0 corresponds to rated line-to-ground voltage. As can be seen, the failure rates for the 132/138kV
and 220/230kV class cables are somewhat comparable. The failure rates for the 345kV/400kV class
cables are significantly lower. This is likely related to the comparatively lower test voltage level for
these circuits.
Due to the issues outlined in Sections II and III of this paper, the PD tests on all cable circuits were
performed as distributed PD measurements using broadband PD monitors having a frequency span
from DC to 400MHz. Furthermore, when PD testing was performed, the testing followed the
decision schedule outlined in Figure 2. In the absence of clearly defined PD pass/fail criteria, when
PD is detected, the decision process with respect to component replacement involves discussion
between the various stakeholders (end client, installer, equipment manufacture, test service provider
and third party consultants). Therefore, it is important general guidelines have been agreed to prior to
testing. Due to the issues related to PD magnitudes an evaluation of the severity of a PD source is not
determined on the measured PD level alone. As can be seen from Figure 2, when a PD source is
detected, the Partial Discharge Inception Voltage (PDIV) and Partial Discharge Extinction Voltage
(PDEV) are measured. In addition, phase plots of the source are obtained so the PD magnitude and
intensity can be quantified. Furthermore, the AC HiPot is allowed to continue and thus the dynamic
behavior of the PD source can be assessed and in most cases quantified.

START

AC HiPot
250kV

PD
Activity

NO

PROC. C2

YES

Increase
Voltage to
250k V
YES
Cont.
Remaining
AC HiPot

END

Notify CLIENT of
PD Activity, PD
Level, PD Stability

YES

PROC. A
Cont.
AC HiPot
END OF ACTION

NO

MAYBE

Internal
PD

Additional Testing

YES

PROC. B
Evaluate PDIV, PDEV
Evaluate PD Stability
Evaluate Pattern

Meas
PDEV
PDIV
Notify Client of
o PDIV & PDEV
o PD Level

PD Meas.
@ PDEV5kV
PD Meas.
@ PDEV5kV

NO

Lower AC
Voltage by
500v/sec

END
Repeat START
Adjacent
Joints

END

Evaluate PD Results
(On-Site)

Kinectrics
Memo
Report

Increase
Voltage to
PDIV +
5kV

Deliver to Client)

DECISION:
CLIENT

PROC. C1

+
Figure 2: PD Decision Tree

Though no standard for PD pass/fail criteria exist the guiding principle for the tests performed on the
cable systems as part of this study has been that the accessories should be PD free to the sensitivity
level of the PD test at the voltage test level. However, for the 220/230kV cable systems, in three out
of ten instances were PD activity was detected in joint or termination, the accessory were not replaced
but conditionally accepted by the end client. In all of these cases, the PDIV was higher than 1.4U0 and
the PDEV was higher than 1.2U0. Furthermore, in these instances, follow up maintenance testing was
mandated by the end client 3-5 years after commissioning. In all other instances where PD was
detected, the accessory affected was replaced or rebuilt and a full AC HiPot & PD retest was
performed.

Segment Wav ef orm

150

150

50

Ch 3

100

Volts (mV)

V olt s (mV)

Segment Waveform

Ch 3

100

-50
-100
-150
0 1

6 7
8 9 10 11 12 13 14 15
Time uSec

50
0
-50
-100
-150
0 1

2 3

4 5

6 7 8 9 10 11 12 13 14 15 16
Time uSec

Figure 3: Example of PD detected on a 345kV Cable Termination

An example of a PD source detected from a 345kV termination during an AC HiPot test is given in
Figure 3. The AC HiPot test level was set to 250kV as per IEC 62067. As can be seen negative and
positive polarity discharge pulses having magnitudes of approximately -52mV and +62mV can be
found centered at 45 and 225 phase angle with reference to the phase-to-ground test voltage. The
frequency content of the 1st moment for the signal activity measured was approximately 2MHz. The
PD source occurred approximately 22 minutes into the AC HiPot and was hereafter present. The
PDEV was measured to approximately 220kV. A standard calibration as per IEC 60270 was
performed on this termination yielding a calibration constant of 0.81pC/mV thus yielding an apparent
charge of 42pC and 51pC respectively. However, evaluating the apparent charge based on the
waveform captured yields a discharge level of approximately -1,100pC and +1,500pC. Similar
differences have been observed on 220/230kV cables systems. The significant difference relates to the
issues outlined in Section III.
Another example of PD detected in a 220kV XLPE joint is provided in Figure 4. In this particular
case, partial discharge measurements were performed following a successful AC HiPot at 216kV
(1.7U0) for 60 minutes. Partial discharge measurements were performed at 153kV (1.2U0) following
conditioning of the cable system at 178kV (1.4U0) for 15 seconds.

Figure 4: Example of PD Measurements on a 220kV XLPE joint at 153kV (1.2U0)

As can be seen from Figure 4, negative and positive polarity PD pulses can be found centered around
45 and 225 degree phase angle with reference to the phase to ground voltage. The PDIV and PDEV
were measured to 153kV and 145kV respectively. Thus, PD activity may occur intermittently during
on-line conditions in the event of a transient voltage wave being applied to the system. A standard
calibration as per IEC 60270 was performed as per the end-clients test protocol on this joint. The
results yielded a calibration constant of 0.19pC/mV thus yielding an apparent charge of -3pC and
+4.2pC respectively thus passing the end-clients 5pC acceptance level and the joint was left in place.
The joint has been in service for approximately 24 months and no in-service failure has been observed.
Lastly, of the cable systems involved with this study, four failures have occurred within 5 years of
passing an AC HiPot PD test. Three failures have occurred on 220/230kV rated systems. Of these
failures, two failures were related to third party external damage occurring during operation whereas
the root cause of the third failure was not positively identified but evidence pointed towards thermomechanical movement of a stress cone within a joint. One in-service failure has occurred on a 400kV
rated cable system operated at 345kV line voltage. The failure occurred within a cable section itself.
The cause of the failure has not been disclosed though third party damage was ruled out.

CONCLUSIONS
Previously, the framework for PD Sensitivity Assessment Procedure for broadband PD measurements
performed on transmission class cable systems was outlined [12]. The procedure has been used on a
population of cable system for the past 5 years. In the absence of a standard for PD testing of field
installed HV & EHV Cable systems a PD measurement procedure as outlined in Figure 1 has been in
use. To date, PD the procedure has been effective in avoiding in-service failures. However, for
345kV/400kV rated cable systems, the accessory failure rates and PD occurrence rates reported during
commissioning testing as compared to accessory failure rates for 69kV through 230kV rated systems
suggest the test levels of 250kV and 260kV respectively may not be sufficiently high to identify all
installation related defects during commissioning testing.
Activities for tracking the service performance of the cable systems being part of this study is ongoing and it is expected an update will be produced in 5 years which will then involve 10 years of
operational experience a population of cable systems subjected to a combination of AC HiPot & PD
acceptance testing.

BIBLIOGRAPHY
[1] L.A. Dissado & J.C. Fothergill, Electrical Degradation and Breakdown in Polymers, IEE
Publications, Peter Peregrinus Ltd. London, United Kingdom
[2] A. Pedersen, G. C. Crichton and I. W. McAllister, The Theory and Measurement of Partial
Discharge Transients, IEEE Trans on Dielectrics and Electrical Insulation, Vol 26, No. 3, pp.
487-497, 1991
[3] Pedersen, G.C. Crichton and I. W. McAllister, PD Related Field Enhancement in The Bulk
Medium, Gaseous Dielectrics VII, Plenum Press, New York, 1994.
[4] A. Pedersen, G. C. Crichton and I. W. McAllister, PD-Related Stresses In The Bulk Dielectric
And Their Evaluation, Annual Record CEIDP, pp.474-480, 1993
[5] A. Contin, A. Cavallini, G.C. Montanari, F. Puletti, A novel technique for the identification of
defects in stator bar insulation systems by partial discharge measurements, IEEE ISEI, pp. 501505, Anaheim, USA, April 2000.
[6] F. Puletti, F. Ombello, M. Albertini, A. Cavallini, G.C. Montanari, Improved diagnostic tools
for cable accessories by digital PD detection systems", Conf. On High Voltage Plant Life
Extension, pp. 2.2.1-2.2.9, Linkebeek, Belgique, November 2000.
[7] Z. Nadolny, J. M. Braun, & R. J. Densley, Investigation of partial discharge pulse shapes
occurring at interfaces in model joints, Proceedings of the 1998 IEEE 6th International
Conference on Conduction and Breakdown in Solid Dielectrics, pp. 119 122, 1998, 1998
[8] S. Boggs & R. J. Densley, Fundamentals of partial discharge in the context of field cable
testing IEEE Electrical Insulation Magazine, Vol. 16, No. 5, pp. 13 18, Sept.-Oct. 2000
[9] IEC 60270, High Voltage Test Techniques- Partial Discharge Measurements December 2008
[10] N. Oussalah, Y. Zebboudj & S. A Boggs, Partial Discharge Pulse Propagation in Shielded
Power Cable and Implications for Detection Sensitivity, IEEE Electrical Magazine, Vol 23.
Issue 6, pp.. 5 10, Nov/Dec 2007.
[11] M. Fenger, H. G. Sedding, Sensitivity Assessment for PD Measurements on Transmission Class
Cables, Proceedings of the 2007 JiCable Conference, A7.1, 2007.
[12] M. Fenger, Sensitivity Assessment for PD Measurements on Transmission Class Cables,
Proceedings of the 2008 ISEI, 2008