CIRED 20th International Conference on Electricity Distribution
Paper 0551
FAILURES OF MEDIUM VOLTAGE CABLE JOINTS IN RELATION TO THE AMBIENT
TEMPERATURE
Rogier JONGEN Edward GULSKI
Delft University of Technology Delft University of Technology
Netherlands Netherlands
r.a.jongen@tudelft.nl e.gulski@tudelft.nl
ABSTRACT
The failure occurrence of a typical resin cable joint used in
the medium voltage cable network in the Netherlands shows
to have a correlation with the ambient temperature. This
behaviour is especially seen when the failures occurred
during the summer months are considered. Especially in the
years 2003 and 2006 high temperatures in the months July
or August were recorded and the number of failures
increased significantly in these months. This ambient
temperature has an influence on the soil temperature
through the year and can even have an effect on daily basis.
It is observed that the failures follow the rise and fall of the
average air temperature during the different days of the
month with increasing and decreasing number of failures
with some time delay. The change of the soil temperature
around the joint results in a higher temperature within the
joint, thus contributing to the failure of a joint, which is
already in its ageing stage. The presence of a hotspot in
combination with a rise of the soil temperature can result
into breakdown.
INTRODUCTION
In the 1970’s a specific type of resin cable joint was used in
three phase medium voltage cables in the Netherlands.
These joints now contribute considerably to the outage time
because of breakdown of the resin insulation. Since the past
5 years an overall failure database is available from where
failure data can be obtained. These data can be used to
perform statistical failure analysis. This analysis shows that
the joints are in their wear-out life period and that ageing is
the main reason for failures. This means that an increasing
failure rate with increasing age is visible [1, 2].
However the number of expected failures resulting from the
analysis shows to differ with the occurring failures. This is
especially the case in several summer months. This raised
the question if the higher ambient temperature has an
influence on the occurrence of failures.
An important factor on the temperature of a cable or cable
joint is the current through the cable. To determine if the
current had a contribution in the failure occurrence, it was
analysed if an increase in the current or an abnormality
could be found in cable circuits with a failure in he joint for
some days till a few minutes before the failure occurrence.
From this study it could be concluded that no significant
changes could be seen, which could be related to the failure
of the joint.
The temperature of the soil surrounding the joints changes
during the year. This affects the temperature distribution in
CIRED2009 Session 1 Paper No 0551
Prague, 8-11 June 2009
Johan SMIT
Delft University of Technology
Netherlands
j.j.smit@tudelft.nl
the joint, which can cause a breakdown in the joint
insulation. Even on depths most cables are located (0.5 – 1
meter of depth) temperature changes throughout the year.
In this paper, the thermal properties of soil are discussed
together with the influence of the ambient temperature on
the soil temperature throughout the year. The effect the
temperature rise can have on the cable and joint is
described. Finally the correlation of the ambient
temperature and the number of occurring failures regarding
the resin cable joints is shown.
THERMAL PROPERTIES OF SOIL
The temperature distribution in the soil is dependant on
various parameters. These include the thermal properties of
soil, which can be described by three typical parameters [3]:
• Volume heat capacity, C; the change in heat content of
a unit bulk volume of soil per unit change in
temperature
• Thermal conductivity, κ; quantity of heat transferred
through a unit area of the conducting body in a unit
time under a unit temperature gradient
• Thermal diffusivity Dh; the change in temperature
produced in a unit volume by the quantity of heat
flowing through the volume in unit time under a unit
temperature gradient. The thermal diffusivity can be
described as the product of the thermal conductivity
and the volume heat capacity.
The thermal properties are very dependant on the soil type
and the composition of the soil. The presence of e.g. water,
air or organic material results in a significant change of the
thermal properties. Due the continuously changing
concentration of water and air in the soil the thermal
conductivity is also time variable. Therefore the thermal
conductivity changes with time and depth. The water
content of the soil and the heat distribution are dependent
processes. These are interacting with and influencing each
other. This can result in the movement of water as well as of
heat in the soil, which results in different thermal properties
of the soil.
A heat source, e.g. online cable, can influence the
parameters, due to the movement of water from the location
with higher temperature to a place with lower temperature.
Consequently the (local) temperature distribution in the soil
changes and may result in a change of the moisture gradient.
These effects influences and entail each other.
CIRED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0551

DEVELOPMENT OF SOIL TEMPERATURE
Besides the thermal properties of the soil, is the soil
temperature dependant on the local meteorological
circumstances. This means that the changes of temperature
of day and night, as well the changes of seasons, highly
affect the condition of the soil. Weather conditions, like e.g.
rain, soil changing properties, surface coverage (e.g. forest
or road) and human management also influence the thermal
regime of the soil, which makes modelling complex.
The temperature development in the soil can be modelled.
The annual as well the daily variations of soil temperature
same depth. The effect for a depth of 5 cm50 and 100 cm is
shown in Figure 1. To obtain this figure the average thermal
parameters are obtained from [5]. By using the ambient air
temperatures in the model an underestimation of the soil
temperature is made. Also the daily fluctuation disappears
in the model when the depth increases. When a comparison
is made between the real measured soil temperatures
throughout the year with the calculated temperatures this
effect can be seen [6].
30
25

Soil Temperature [°C]

can be described by a sinusoidal function as a function of 20

time and depth [3, 4]: 15

10
T ( z , t ) = Ta + Az sin (ω t + φ ( z ) ) (1)
5

0
Where T(z,t) is the temperature at depth z as a function of
time t. Ta is the average temperature of the surface and Az is -5
0 50 100 150 200 250 300 350
the temperature fluctuation at depth z, ω is the radial
z = 5 cm z = 50 cm z = 100 cm Day of the year
frequency and φ ( z ) is the time lag as a function of z.
Figure 1 Soil temperature development at various depths throughout the
The effect of the depth can be added in this equation: year based on average daily ambient temperature

  z  − z The temperature distribution at various depths can vary

T ( z , t ) = Ta + A0 sin  ω t −   e d (2)
  d  throughout the year, depending on the season. This effect
can be seen in Figure 1 were an example is shown for the
with A0 is the amplitude fluctuation at the surface. temperature development starting in winter season.

The constant d is the damping depth. At this depth the

temperature amplitude has decreased with a factor 1/e with
respect to the amplitude at the soil surface, A0. The damping
depth is described by:
1 1
 2κ  2
 2D  2
d =  = h  (3)
 Cω   ω 
As can be seen, the damping depth is inversely proportional
to the square root of the frequency. This means that the
damping depth is 19 times larger for the annual variation
compared to the diurnal variation.
The effect of annual and diurnal fluctuation can be included
in equation 2. Also a phase constant can be introduced
which represents the arbitrary zero point for the time scale,
(2) then becomes:

  z  − d
z

T ( z, t ) = Ta , y + Ay sin  ω y t + φ y −   e y
  d y  
 (4)
  z  −
z
Figure 2 The behaviour of temperature at different depths at different times
+ Ad sin  ωd t + φd −   e dd of a summer day from [3]
  d d  
The diurnal variation of the temperature shows a more
Where the subscripts y and d denote the yearly and daily
dynamic character, especially for small depths. An example
variations.
of an illustration of the variation of the temperature at
The daily changes depending on the depth are super
different depths throughout the day for a day in the summer
imposed on the annual fluctuation of the temperature at the

CIRED2009 Session 1 Paper No 0551

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

Paper 0551

is shown in Figure 2, taken from [3]. In this figure also the failures increase in the months June to August, especially in
heat flow in the soil is shown. It can be seen that at small July. In these months the average temperature is often the
depths the temperature follows almost directly the ambient highest.
temperature, while the effect on greater depths is lagging the
temperature rise and fall. 25 75

Number of failures
20 60
CABLE TEMPERATURE AND HEAT FLOW

Temperature [°C]
The thermal properties of soil surrounding the cable are 15 45

very important for the thermal behaviour of the cable and 10 30

have to be taken into account when calculating the
maximum load on the cable [7]. The soil contributes in the
heat flow between the cable and the ambient environment.
The thermal resistivity (the inverse of the thermal
conductivity) of the soil is of great importance for the
thermal behaviour of the soil. The thermal resistance can
vary dependant on the depth, the presence of adjacent
cables and the soil type [8]. The soil resistivity can be
further influenced by [9]:
1. Air, which has a high thermal resistivity. The soil
around the cable has to be tightly packed to
achieve lower thermal resistances
2. Replacement of the air with water results in a
better thermal resistivity, but water is still not a
good conductor of heat
3. Organic materials in the soil have a very high
resistivity even when they are very wet.
4. Filling materials used for covering the cable duct
with quartz material will have the lowest resistivity
5 15
0 0
1 2 3 4 5 6 7 8 9 10 11 12
Average temperature 2002-2006
Average Number of failures 2002-2006 Month
Figure 3 Average ambient temperature per month and average number of
failures in the period 2002-2006
To see the effect more clearly, the number of failures and
the average ambient temperature per month for 2006 are
shown in Figure 4. In 2006, a rise in the total number of
failures was seen after a decrease in the previous years.
Especially in the month July ambient temperatures were
higher than normal. In this month the number of failures was
relatively high, as can be seen from Figure 4.
25 125
20 100

Number of failures
The above mentioned points show that the thermal
Temperature [°C]

properties of the soil can vary a lot even along the length of 15 75
a cable. As mentioned before, the water content can play an
important role in determining the thermal properties of the 10 50
soil. As long as the water content is high enough, the
thermal resistivity is rather stable with increasing or 5 25

decreasing water content. However, when the critical water

content is reached the thermal resistivity increases
dramatically with decreasing water content. The soil around
the cable gets to dry and the cable heat results in less water
content, which can result in additional heating. This can
finally result in a thermal runaway condition and thermal
instability of the cable. In general it can be stated that if the
cable is located deeper in the soil, the more stable the
thermal environment of the cable will be.
JOINT FAILURES AND AMBIENT AIR
TEMPERATURE
In previous sections the thermal properties of the soil were
discussed and the influence this can have on the cable or
cable joint. A specific type of 10 kV resin cable joint
applied during the 1970s shows typical failure behaviour
throughout the year. A striking aspect is that most failures
occurred in the summer months. The average failure
developments per month in the past 5 years together with
the average monthly ambient temperature [10] are shown in
Figure 3. From the figure it can be seen that the number of
0 0
1 2 3 4 5 6 7 8 9 10 11 12
Month
Average temp Number of failures
Figure 4 Average ambient temperature and number of failures per month in
2006
When the ambient temperature is correlated with the
number of failures that occurred in the months July and
August of the past 5 years, it can be seen that the correlation
coefficient is around 0.88 for both months. This could
indicate that there is a relation between ambient temperature
and failures. If the correlation is taken from the total number
of failures and ambient temperature throughout the year, as
shown in Figure 3 for the year 2006, the coefficient is 0.64.
To get a closer view of the effect of the ambient
temperature, in Figure 5 the ambient temperature is shown
per day for the end of June till the beginning of August
2006, together with the number of failures per day.

CIRED2009 Session 1 Paper No 0551

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

Paper 0551

30 18
0.75 July
25 15 0.65

Number of failures
Average temp [°C]

0.55
20 12
0.45

15 9 0.35

correlation
0.25
10 6
0.15

5 3 0.05

-0.05 2002 2003 2004 2005 2006

0 0
-0.15
23 26 29 2 5 8 11 14 17 20 23 26 29 1 4 7
June July August -0.25
Day year
Average temp. per day Number of failures
no delay 1 day delay 2 days delay 3 days delay 4 days delay 5 days delay
Figure 5 Average daily ambient temperatures from the end of June till the 7

beginning of August of 2006 0.75

0.65 August
In this figure it can be seen that the majority of the failures 0.55

occur after a period with higher ambient temperatures. 0.45

0.35

correlation
0.25
There is some delay between the moment of the high
0.15
ambient temperature and the moment of failure. This delay
0.05
is around 1 to 3 days, which can be seen in Figure 6. In the
-0.05 2002 2003 2004 2005 2006
case for the month July in 2006 the correlation between the
-0.15
ambient temperature and the number of occurred failures is
-0.25
0.49. When a delay of one day is taken into account it can year

be seen that the correlation coefficient increases to 0.57. For 8 no delay 1 day delay 2 days delay 3 days delay 4 days delay 5 days delay

other years it can be seen that a delay of 2 to 3 days results Figure 6 Correlation of the occurrence of failures and ambient temperature
in a higher correlation coefficient. including a delay of 0-6 days, the line shows the critical correlation value.

CONCLUSIONS
REFERENCES
Statistical analysis of the lifetime data of a population of
[1] R. A. Jongen, P.H.F. Morshuis, J.J. Smit, A.L.J. Janssen, E.
resin cable joints presented in previous work showed that Gulski, “Statistical Analysis of In-service Failed Resin Cable Joints”,
the joints are in their wear-out stage of life. Ageing was Proceedings International Conference on Condition Monitoring and
denoted as the main reason for failure. An increase in the Diagnosis, Korea, 2006.
number of failures was observed last year, which was higher [2] D. Hillel, Environmental Soil Physics, Academic Press, San
Diego, USA, 1998.
than expected. For this reason, the relation between ambient [3] R.E. White, Principles and Practice of Soil Science: the Soil as
temperature and failures was investigated. The following a Natural Resource, fourth edition, Blackwell Publishing, Malden, USA,
can be concluded: 2006.
• Soil temperature is highly dependant on various factors, [4] D.A. De Vries, “Thermal properties of soil In Physics of Plant
Environment”, W. R. van Wijk (ed.) North Holland Pub. Co., Amsterdam ,
like water content, soil type and ambient air 1963
temperature. The loading of the cable can influence the [5] J. Wu, D. L. Nofziger, “Incorporating temperature effects on
soil properties as well, which can result in a change of pesticide degradation into a management model”. Journal of Environmental
the (local) soil temperature. Quality, Vol. 28, pp. 92-100, 1999.
[6] IEC 287-2-1, Electric Cables, Calculation of Current Rating,
• An increase of failures in the summer months can be 1994.
clearly seen. Especially in the warmer months June- [7] C.L. Cooper, M.L. Dyer, G.G. Karady, “Investigation of the
August an increase is visible. Seasonal Variation of Ground Thermal Conductivity on High Voltage Cable
Ampacity”, Large Engineering systems Conference on Power Engineering,
• The peaks in the average daily ambient temperatures July 2004, pp. 108-112.
are often followed by an increasing number of failures. [8] G. Campbell, K. Bristow, “Soil thermal resistivity”. Australian
This can be caused by the high thermal stress in the Power Transmission and Distribution. Chapel Hill, Qld :PTD Publications:
cable joint as a result of the rise of the soil temperature 46-48, 2002.
[9] KNMI, Royal Dutch Meteorological Institute, the Netherlands,
together with the aged insulation of the joint. http://www.knmi.nl.

CIRED2009 Session 1 Paper No 0551