CIRED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0055

CHANGES IN INSULATION CONDITION OF DRY CURED XLPE CABLES

OVER A 6-YEAR PERIOD

Hans Jørgen JØRGENSEN

Danish Energy Association – Denmark
hjj@danskenergi.dk

ABSTRACT The measurement program also included a test of the

The service experience with dry cured XLPE cables is good
and their internal failure rate is much lower than is the case
for the older steam cured cables. However, water tree aging
also occurs in dry cured cables, unless the insulation is
protected against moisture by water barriers.
Since 2001, a series of diagnostic measurements has been
carried out on in total 58 dry cured XLPE cable sections
without water barrier. The cables were installed in 10 kV to
20 kV networks in the period between 1979 and 1996. The
objective of these measurements has been to get an
assessment of the insulation condition and the rate of
change in insulation condition for these cable sections and
the results are presented in the paper.
integrity of the sheath. A DC voltage of 2 kV (for PVC
sheaths) or 5 kV (for PE sheaths) was applied between the
cable screen and ground and the current flowing through the
sheath was measured. The measured values were compared
with limit values for the sheath currents through intact
sheaths determined in a joint Nordic project [3]. These limit
values are:
• I0=500 µA/km at 2 kV for cables with PVC sheaths
• I0=10 µA/km at 5 kV for cables with PE sheaths.
CABLES INCLUDED IN THE INVESTIGATION
Cables from at least 5 different cable manufacturers were
included in the test program, but the majority of the cables
had been produced by one of the manufacturers.

INTRODUCTION Most of the cables had been installed in the 1980’ies as

shown in Figure 1.
The steam cured XLPE cables installed in the 1960’es and
10
1970’ies started to fail due to water treeing after 10 to 15
9
years and many of them have now been replaced. 8

7
Number of cables

Dry cured cables without water barrier were introduced in 6

Denmark in 1978 and until now no increase in failure rate 5
has been observed for these cables. However, in 1998 and 4
1999 diagnostic measurements revealed signs of ageing also 3

in dry cured cables and this led to the initiation of a 6-year 2

measurement program as described in [1]. 1

0
1978 1980 1982 1984 1986 1988 1990 1992 1994 1996
DIAGNOSTIC METHOD Year of installation

The measurements were carried out using the method called
dielectric spectroscopy [2]. VLF voltages in the range
between 0.5 U0 and 1.5 U0 and with frequencies typically
between 0.03 Hz and 0.3 Hz were applied, and the dielectric
losses in the insulation were measured.
Based on the variation of the dielectric losses with voltage
and frequency, the cables were classified into four
categories, corresponding to increasing levels of ageing:
• Good cables with low losses
• Aged cables with voltage dependent losses
• Significantly aged cables with transition to leakage
current at the highest applied voltage level
• Bad cables with leakage current occurring at normal
operating voltage.
Figure 1. Year of installation for the cables included in the
project.
Unfortunately, it was not possible to assess the insulation
condition of all these cables. Around 25 % of the cable
sections had been equipped with cable accessories that had
high losses and prevented the measurement of the losses in
the cable insulation.
If a network owner wants to be able to carry out diagnostic
measurements on XLPE cables, only cable accessories with
high DC resistance at voltages of 1.5 U0 to 2 U0 should be
installed. So far, measurement of this DC resistance is not
included in the type tests for accessories, but the customers
should ask the manufacturers to give this information and
choose among the types with high DC resistance.

CIRED2009 Session 1 Paper No 0055

CIRED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0055

SHEATH CONDITION
When testing the integrity of the cable sheaths it was found
that the sheath currents in some of the cables were close to
the limit values of I0=500 µA/km (PVC) and I0=10 µA/km
(PE) given above. Therefore, the cables were divided into
three categories
• 44 % of the cables had sheath currents below I0,
indicating an intact sheath
• 21 % of the cables had sheath currents between I0 and
10 I0 and may have a damaged sheath
• 35 % of the cables had sheath currents above 10 I0 and
certainly have a damaged sheath.
cables with leakage current 37 dry cured cables were
checked at least twice. In Figure 3, the results of the first
and the final assessment of the insulation condition of these
cables are shown. A significant deterioration of the
insulation condition is seen in the figure. 16 out of the 37
cables were found to have a poorer insulation condition at
the last measurement than at the beginning of the 6-year
period.
25
20
First assessment
Final assessment

Number of cables
15
In Figure 2 the sheath condition is shown for three groups of
cables with different time in operation. It is seen that there 10

are only small differences in sheath condition between the

three groups despite the difference in installation period. A 5

simple explanation for this could be that the sheath

damaging mainly occurs during installation or shortly 0
Good Aged Significantly aged Bad
thereafter. But other explanations including changes in Insulation condition
installation practice, soil movements and excavation work
Figure 3. Changes in insulation condition for 37 dry cured
can not be excluded.
cables within the 6-year project period.
100%
The most significant change in insulation condition was
90%
observed for a cable from 1983. In 2001 the cable was
80%
found to be in good condition with low dielectric losses but
Percentage of cables

70%

60%
in 2007 transition to leakage current was measured at 1.5 U0
Intact
50% Possibly damaged
(9 kV) and the leakage current response was maintained
40%
Damaged after lowering the voltage to U0 (6 kV) as shown in
30% Figure 4.
20%

10%

0%
1979-85 1986-90 1991-96
Year of installation

Figure 2. Percentage of cables with damaged or possibly

damaged sheath for three different installation periods.

INSULATION CONDITION
In the years 2002 to 2007 some of the cables have been
selected for re-measurements. The selection was based on
the result of the first diagnostic measurements so that the
cables showing the clearest signs of ageing at the first
measurements in 2001 were investigated more often than
the cables with good insulation condition.

In 2001, some of the cables were categorised as having a

very poor insulation condition as a leakage current response
was measured. Later diagnostic tests of these cables have
revealed that the leakage current was caused by the
accessories and not the insulation. The actual insulation
condition of these cables can therefore not be determined.
A few cables were only measured once but apart from the
Figure 4. Dielectric losses (dissipation factor) measured in
2007 on a 10 kV cable from 1983.
So far, this is the only time a response like the one in
Figure 4 has been registered for a dry cured cable. For a
steam cured cable a similar response, which is an indication
of the presence of long water trees, will mean a high risk of

CIRED2009 Session 1 Paper No 0055

CIRED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0055

failure in the case of an overvoltage on the system e.g. in the

case of a single-phase earth fault at another location in the
network . This probably also is true of a dry cured cable.
However, the number of water trees in the insulation of a
dry cured cable is much lower than in a steam cured cable
and it is not yet known how much this fact reduces the risk
of failure. The cable in question has been in operation for
more than a year since the measurement in Figure 4 was
carried out and has not failed yet.

The poorest insulation condition was found in one of the

oldest cables and the results shown in Figure 5 confirm that
the oldest cables also show the most pronounced signs of
ageing. But even a cable with only around 15 years of
service can be significantly aged.

100%

90%

80%
Percentage of cables

70%

60% Good
Aged
50%
Significantly aged
40% Bad

30%

20%

10%

0%
1979-85 1986-90 1991-96
Year of installation

Figure 5. Assessment of insulation condition for cables

installed in three different installation periods.

The condition of the sheath was expected to have an

influence on the insulation condition, and the initial
measurements presented in [1] also seemed to support this
expectation. However, based upon the final assessment of
the cables the overall insulation condition surprisingly
seems to be best for the cables with damaged sheaths as
shown in Figure 6. It is not logical that a damaged sheath
should prevent ageing, so it must be a matter of
coincidence.

100%

90%

80%
Percentage of cables

70%

60% Good
Aged
50%
Significantly aged
40% Bad

30%

20%

10%

0%
Intact Possibly damaged Damaged
Sheath condition Figure 7. Diagnostic measurements on a 10 kV cable in
Figure 6. Assessment of insulation condition for cables with three consecutive years: 2005 (top), 2006 (middle) and
varying sheath condition. 2007 (bottom).

CIRED2009 Session 1 Paper No 0055

CIRED 20th International Conference on Electricity Distribution
Several possible explanations for the fact that a damaged
sheath does not necessarily lead to poor insulation condition
could be given. As the cables have no radial water barrier,
water with time will diffuse into the insulation, even if the
sheath is intact, and local variations in soil humidity will
probably have a larger influence on the ageing process than
a hole in the sheath.
Furthermore, the effects of a damaged sheath may be very
local and may only contribute moderately to the average
losses for the whole cable section.
INFLUENCE OF SOIL HUMIDITY
It is known from laboratory measurements that if the
insulation dries out, an XLPE cable with water trees does
not any longer give the typical water tree response in
diagnostic measurements, as the water trees become less
conductive.
In 2006 this phenomenon was observed for underground
cables still in service. The summer in Denmark in 2006 was
extremely hot and dry, and for some of the investigated
cables this led to a large reduction in losses as shown in
Figure 7 on the previous page.
A year later, the losses for most of the cables had returned
to a level close to the values from 2005 (or earlier).
However, a few still had low losses when the final
measurements were carried out in 2007.
As the growth of water trees is also stopped when the cables
dry out, the dry summer of 2006 may have added a year or
more to the expected service life of some of the cables.
CONCLUSION
The measurements have shown that 50 % of the cables are
still in good condition. 25 % show signs of slight ageing and
around 23 % are significantly aged. One cable (around 2 %)
was found to have bad insulation condition with high risk of
failure in the case of an overvoltage occurring on the
system.
In the 6-year project period the insulation condition was
found to deteriorate for more than 40 % of the cables. The
CIRED2009 Session 1 Paper No 0055
Prague, 8-11 June 2009
Paper 0055
greatest change was observed for a cable installed in 1983.
In 2001 this cable was found to be in good condition but in
2007 it showed signs of long water trees.
Of the tested cable sections at least 35 % had damaged
sheaths. There seems to be no correlation between the
number of years in service and the occurrence of sheath
damage, so probably the damaging of the sheaths mainly
occurs during the installation of the cables.
From the measurements, the insulation conditions for cables
with intact sheaths seem to be poorer than for cables with
damaged sheaths. This must, however, be a matter of
coincidence, and there may be some local deterioration of
the insulation which does not contribute enough to the total
dielectric losses of the insulation to change the assessment
of the insulation condition.
In 2006 the summer in Denmark was extremely warm and
dry and this led to a drying of some of the cable sections.
The dielectric losses in several of the investigated cables
were reduced and their insulation condition could not be
assessed correctly. Measurements in 2007 on the same
cables showed that most of them again had obtained their
“normal” dielectric losses, but some still seemed to be dry.
None of the tested cable sections has failed yet so it is still
unknown how fast the last part of the degradation of the
insulation will proceed. To get information on this the cable
sections will be followed also in the coming years.
REFERENCES
[1] H. J. Jorgensen and J. S Christensen, 2003,
"Experience from diagnostic testing of MV XLPE
cables", Proceedings CIRED 2003, Session 1, Paper
No. 75
[2] S. Hvidsten, P. Werelius and J. Christensen, 2001,
“Evaluation of on-site dielectric response methods for
non-destructive testing of water treed MV XLPE
cables”, Proceedings CIRED 2001, vol. 1, paper 1.48.
[3] J. T. Benjaminsen, 2001, Sheath Testing of MV
Polymer Cables. Background and Handbook,
TR A5484, SINTEF Energy Research, Trondheim,
Norway.