Anda di halaman 1dari 10

# Applied Thermal Engineering 78 (2015) 268e277

## Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Research paper

## Underground electric cables a correct evaluation of the soil

thermal resistance
Ferdinando Salata*, Fabio Nardecchia 1, Andrea de Lieto Vollaro 2, Franco Gugliermetti 3
DIAEE, Area Fisica Tecnica, Sapienza University of Rome, Via Eudossiana, 18, 00184 Rome, Italy

h i g h l i g h t s
 Optimal installation of electrical buried cables.
 Proper evaluation of the thermal resistance of the soil.
 Inuence of the geometric parameters of the trench.
 Mutual inuence of multiple conductors in the same excavations.
 Correction of design formulas provided by regulations.

a r t i c l e i n f o

a b s t r a c t

Article history:
Accepted 22 December 2014
Available online 3 January 2015

Nowadays companies supplying medium/high voltage electric energy, near residential areas, tend to use
underground cables laying. Hence the design engineer is required to estimate the thermal resistance
around the underground cable to perform a right dimensioning of the cables; as a matter of fact what it
should be avoided is the overheating determined by a bad heat dissipation due to the Joule effect.
IEC rules provides a formula for the evaluation of the soil thermal resistance which is easy to apply. But
from an experimental point of view, as the bibliography shows, it was discovered it tends to underestimate the problem when dealing with very dry soils in particular.
Thanks to an experimental system, some useful data were collected for the validation of a 2D FEM
model of an excavated area with a linear heat source reproducing the underground conduit. The numerical model presents a variation of both the geometrical parameters of the excavated area and the
distance characterizing other cables (hence heat sources) located in the same site of installation
examined. In this way two dimensionless coefcients, useful to correct the values of the thermal
resistance furnished by the current regulations, were determined.

Keywords:
Underground cable
Linear heat source
Dry soil
IEC 287
Design method
CFD model

1. Introduction
It can be noticed how, in the past few years, near residential
areas in particular, there was a tendency to use underground power
cables supplying energy through medium/high voltage alternating
current. The data [1] conrm that this solution is not widely used
yet: when the voltage is very high (380 kV) the percentage of the

## * Corresponding author. Tel.: 39 0644585661; fax: 39 064880120.

E-mail addresses: ferdinando.salata@uniroma1.it (F. Salata), fabio.nardecchia@
uniroma1.it (F. Nardecchia), andrea.delietovollaro@uniroma1.it (A. de Lieto
Vollaro), franco.gugliermetti@uniroma1.it (F. Gugliermetti).
1
Tel.: 39 0644585685; fax: 39 064880120.
2
Tel.: 39 0644585720; fax: 39 064880120.
3
Tel.: 39 0644585429; fax: 39 064880120.
http://dx.doi.org/10.1016/j.applthermaleng.2014.12.059

## underground electric network, respect to the total value, presents

the following variations: 0.4% for countries such as United States
and Spain, 0.3% in Germany and Holland, 0.1% in Sweden and
Canada. Other countries, France and Switzerland, are characterized
by lower values, 0.2%. In Italy the underground electrical power
grid, with a voltage of 380 kV, equals to 25 km out of 10,700 km
(0.2% of the total), whereas the one with a voltage of 220 kV and
132e150 kV is 950 km out of 52,000 km (1.8% of the total).
There are many reasons why there is a tendency to prefer underground cables, one is the possible electromagnetic pollution
determined by overhead conduits. Underground cables imply extra
costs which, according to the valuation performed by European
managers grid, is 10e13 times more the price of a traditional
132e150 kV overhead line and 12e17 times more the very high
voltage power lines. Such difference is due to the fact that with

## 123e150e220 kV power lines, underground cables are able to

transfer the same amount of power of overhead cables, whereas for
380 kV power lines we need a doubling of the line. These costs are
determined by the necessary equipments for the exertion of underground cables requiring stations at regular intervals. Other
costs, caused by underground lines management, are determined
by the maintenance of the cables during their exertion, which is
more complicated and expensive. From a statistical point of view
the unavailability of the power line, in those sections affected by a
damage, varies from the typical values characterizing an overhead
line (about a few hours) to values that can reach 25 days (for underground cables) like in Refs. [2e4]. The fact that the cables are
installed underground does not mean that they are immune from
damages and malfunctioning, actually we have the opposite situation. As a matter of fact the problems caused by the disposal of
heat produced by the cables, due to the Joule effect, are wellknown.
Thus the cables laying is a solution which can be easily applied
to those sections where the power line has a high voltage, but not
very high, especially when there are external factors, such as areas
located near highly urbanized zones. Thanks to the environmental
benets of the exposed population, it is still possible to have underground cables at 132e150e220 kV with reasonable costs [5]; if
we are dealing with peripheral lines in particular, since they do not
belong to the group of the main high-voltage power lines, and
where failure may be more easily tolerated by the electrical system.
During the planning phase of the long-distance power line, it is
necessary to estimate the soil thermal resistance to verify the heat
disposal. Such value is provided by the IEC 287-2-1 regulation
through Formula (1), which is easy to apply:

RIEC


p
1
rT ln u u2  1
2p

(1)

269

## for the geometric conguration, taken as a point of reference and

which corresponds to the experimental one (with just one
conductor cable), the value of the resistance estimated has an underestimation of the 70% respect to the one measured experimentally. These results were later conrmed by the numerical
simulations performed through a validated FEM model [11]. The
regulations IEC 287-2-1, to make the estimation easier, provide the
Equation (1) to calculate the soil thermal resistance. It does not take
into consideration though that for very dry soils this equation underestimates the problems of heat disposal caused by the underground electric cable due to the Joule effect demonstrated in Ref.
[12].
In this work, it was also taken into consideration how the shape
of the excavated area and the positioning of the cable in the excavated area can nullify the value calculated through the formula
provided by the current regulations. The regulations [13], concerning underground cables installation (called the M1 type),
dene the minimum depth of the cable laying based on its category: category zero and 1:0.5 m; category 2:2:0.6 0.8 m; category
3:1.0 1.2 m. As a matter of fact the depth, width and the bed of
sand where the electric cable must be positioned and the backll of
the excavated area are all factors affecting the thermal resistance of
the soil thus presenting different values respect to those assumed
by the regulations [14].
Even the presence of many cables in the same excavated area
determines a thermal eld affecting, in a negative way, the heat
disposal of those cables installed individually [15]. In order to make
a more accurate evaluation of the actual thermal resistance of the
soil the choice was to suggest corrective coefcients of the Formula
(1) provided by the IEC regulation [16]. The purpose was to identify
two corrective factors (f1 and f2, dimensionless) in order to have a
better approximation of the actual value of the soil resistance as
shown in Refs. [17e19] and reach this type of relation:

where:

## RFEM f1 \$f2 \$RIEC

 r: soil thermal resistivity (from 0.7 K m/W, for very humid soils,
to 3.0 K m/W, for very dry soils);
 u: (2L/D);
 L: distance from the center of the cable respect to the ground
level [mm];
 D: outside diameter of the underground cable [mm].

where:

## It is clear that an underestimated valuation can cause serious

problems during the exertion of an underground conductor. A bad
disposal of the heat generated by the surrounding soil determines
phenomena which quicken the aging process of the insulation layer
of the cable. They are deeply affected by the maximum temperatures of exertion that the electric cable must endure even for short
time intervals as shown by Ref. [6]. Hence, furnishing the design
engineers with something reliable about soil resistance [7], in order
to avoid damages and malfunctioning which always lead to a lower
reliability and higher extraordinary maintenance costs respect to

Hence the FEM software modeled different geometrical congurations [20] of the excavated area, and afterward different
positioning for two more cables installed near the disturbed cable,
in the same installation area where the cables are positioned. The
FEM model was validated experimentally thanks to a scale model of
an excavated area reproduced in a laboratory [9,21].

## 2. Material and method

In this case study the focus is on the soil thermal resistance of
very dry soils [8] which present hard conditions of exertion for
underground conduits.
From the experimental analysis performed in the laboratory,
through the equipment here described in the paragraph 2.1 (whose
results were previously published [9,10]), it can be noticed how the
value of the resistivity experimentally measured in the rst attempt
is higher than the value provided by the Equation (1). In particular,

(2)

##  f1: dimensionless factor affected by the variation of the

geometrical parameters of the excavated area;
 f2: dimensionless factor affected by the presence of other cables
in the installation area.

## 2.1. Experimental measurements

In order to be able to study the thermal effects of an underground power cable in the surrounding soil it was reproduced a
scale model (Fig. 1) in a laboratory. A wooden box (whose plant
presents the following dimensions: length 1.9 m, width 1.5 m and
depth 0.35 m) was insulated from the surrounding environment
through polystyrene panels. The inside was covered with a waterproof enamel to avoid humidity exchanges between the soil and the
surrounding environment. Then it was realized a trench (width:
0.008 m), representing the section of the excavated area with the
underground cable, with the axis of symmetry of the longest side
included between the ground level and a depth of 0.15 m. Inside
this trench, at its midpoint, at a depth of 0.135 m and above a sand
bed, it was positioned a pipe made of steel. The pipe, which on scale

270

## represents the transit of a power line, has a diameter of 0.005 m, a

depth of 0.25 mm and a length of 1.5 m. In the zone right above the
cable, being the shortest section for the thermal ux generated by
the underground cable because of the Joule effect, there is a heat
exchange with the outside environment (Fig. 2).
This is the reason why between the ground level and the cable,
ve type of K thermocouples (Chromel (NieCr) ()/Alumen (NieAl)
()) were positioned, each placed at a distance of 25 mm from the
other. The cold junction of the thermocouples is immersed in a
thermostated bath with an accuracy of 0.1 K. They permit to
monitor the trend of the temperatures of the material covering the
electrical cable. Even the steel pipe, simulating the electrical cable,
has a thermocouple similar to what it was just described measuring
the reached temperature determined by the Joule effect. A stabilized constant current generator, with a voltage varying from 2 V to
15 V, supplies the pipe with electricity. The voltage drop DV at the
extremities of the pipe has a range of 1.8 2.8 V. The power owing
through the pipe was estimated by measuring the voltage drop at
the extremities of a resistance accuracy (rp 0.1 U 0.00001 U).
The laboratory, where the whole experimental equipment was
set, is a thermostated environment with the possibility to assess
the indoor temperature with an accuracy of 1 K.
The box is lled with argillaceous and compact soil (expanded
and sifted clay) reproducing an undisturbed dry soil. The trench is

lled with river sand around the cable up to the ground level. The
sand, before being placed in the trench, was subject to a drying
process of 8 h at a temperature of 80  C (353 K).
The thermophysical properties of the backll materials are
showed in Table 1 [22,23]. Their density and thermal conductivity
were measured respecting the requirements of the IEE Standard
442-1981-1996 [24] by using specic tools (MAE A5000T Thermal
conductivity probe: MAE CTS-45) with a maximum accuracy of 5%.
All the signals sampled are the result of a HP Agilent 34970A
DATA multimeter model. The measurement error of the multimeter, during the measuring phase of the output signal of the
thermocouples, has a maximum value of 4 mV. Every thermocouple furnished values with an accuracy of 0.2 K, with the
temperatures ranging between 273 K and 373 K. The total measurement error (calibration error included) is of 4 K.
In order to determine, from an experimental point of view, the
soil resistance 4 measurements were carried out waiting for the
system to be in stationary conditions. The system was supplied
with electric power which determined, for every linear meter of the
pipe, a disposal of a heat q respectively of 8, 10, 13, 18 W/m. The
stationary condition was considered satisfying after 6 h the thermocouple, registering the pipe temperature, presented variations
higher than 0.2 K. The soil resistance was estimated as follows:

Rsper

Tt  Tair Tt  Tair rp It

q
Vr Vt

(3)

## The maximum differential error, based on the measurement

chain, is respectively:

## dRsper dTt  dTair drp dIt dVr dVt

Rsper
Tt  Tair
rp
It
Vr
Vt


6
4
0:033 10 3:6\$10 0:0035 y3:7%

(4)

where:
 dTt 0:6 K is the error generated by the type of thermocouple
used for the calculation of both the soil temperature and multimeter used for the surveying;
 dTair 0:4 K is the error generated by the type of thermocouple used for the calculation of the air temperature;
 Tt  Tair 30 K is the temperature difference maintained
during the experimental phase;
 drp =rp 106 is the accuracy of the resistor used by the experimental equipment;
 dIt =It 0:0005=1:4 is the error determined by the measuring
instrument for the reading of the electric power owing
through the cable;
 dVr =Vr ydVt =Vt 0:0035 is the error generated by the
measuring instrument for the reading of the potential drop
concerning the extremities of the accuracy resistor and the
cable;
Such value is acceptable because of the accuracy required by the
experimental measurements.

Table 1
Thermophysical properties of the materials used for the experimental equipment.

## Fig. 2. Measurement chain of the experimental equipment in the laboratory.

Material

Density d [kg/m3]

Specic heat cp
[kJ/(kg K)]

Thermal conductivity
l [W/(m K)]

Sand
Clay

1700
380

837.2
840

0.35
0.29

271

## 2.2. Simulations through fem modeling

A mathematical model was implemented to reproduce the same
phenomena examined in the laboratory by using the experimental
data obtained in the laboratory. Such model was realized thanks to
a numerical prediction software which later was adjusted and
validated through the comparison of the outputs furnished by the
simulations performed with the experimental data. The software
used was a certied FEM produced by the software house Ansys.
Thanks to the simulation software it was possible to modify both
the geometry of the excavated area (with the simulated underground conduit) together with the number and the positioning of
the cables, with the certainty that the equations set to solve the
problem, the thermophysical properties of the simulated materials
and the conditions set for the surrounding area respect the physical
phenomena actually examined, thus obtaining valid results from
the simulations performed.
The commercial software used solves the energy equation as
follows:

0
1

X ! 
v
!
!
rE V\$ y rE p V\$@keff VT 
hj J j teff \$ y A Sh
vt

## Fig. 3. Detail of the mesh realized around the underground cable.

(5)
where: r is the density [kg/m3], E is the energy [J], !
y is the speed
[m/s], p is the pressure [Pa], keff is the effective conductivity, T is the
!
temperature [K], h is the enthalpy [H], J j is the diffusion ux of
species j, teff is the effective stresses tensor. The rst three terms on
the right side of the equation are respectively the energy transferred by conduction, species diffusion and the viscose dissipation.
Sh is the heat of the chemical reactions and other volumetric heat
sources which, in this case study, are absent. For a closer examination refer to Ref. [25].
Governing equation of the system is solved by nite-element
method employing semi-implicit method for pressure-linked
equations (SIMPLE) algorithm. The second order upwind scheme
is used for discretization of the equations. The convergence criteria
for all the dependent variables are specied as 1015.
While trying to reproduce the geometry to examine, one of the
main problems faced was the correct generation of the mesh
constituting the grid where the software would solve the equations. A structured mesh helps avoiding problems caused by the
error propagation. As a matter of fact we chose this kind of mesh,
preferring the Quad Map type. Due to the peculiar geometry
taken into consideration, it was assumed the case of a indenitely
extended cable and it was examined its transversal section in a
two-dimensional reproduction. Even the spacing and number of
cells can compromise the good quality of the calculus, hence the
numerical results obtained. What it should not happen is an excess
of cells, where estimation is performed, in order to avoid an
excessive load on the hardware resources of the machinery performing the calculation. This is why a mesh sensitivity was
conducted.
The cable is represented by a circumference divided into 4 arcs,
each of 90 . The area surrounding the cable is formed by a square
mesh representing the sand bed. The center of the circumference,
representing the cable, is placed in the point of intersection of the
diagonals of the square. The mesh connecting both the circular and
square geometry is the area requiring a higher level of accuracy to
avoid a loss of information while solving those equations describing
the thermal exchange phenomena inside the cable. This is the
reason why the more we distance from the area representing the
excavated area, the denser the section with the structured mesh
will be. Fig. 3 shows a detail of the mesh inside the cable.

It can be noticed how every arc forming the cable is divided into
22 equidistant nodes. The side characterizing the square geometry
surrounding the cable is divided into the same number of nodes as
well. Each node is connected, through a segment, to the corresponding node on the arc of the circumference. Each segment is
divided into 15 nodes with a ratio of 1.05. The rest of the excavated
area is divided into mapped quadrilaterals with cells characterized
by a regular shape. Outside the excavated area the ground is
mapped with cells characterized by a rectangular geometry where
the more we distance from the area taken into consideration, the
less dense the section will be; the distance between each node
increases with a ratio of 1.02. The mesh is then formed by about the
whole of 5.2 105 cells and it will be called, Mesh 1X.
The properties of the materials supplied as inputs to the simulation software are reported in Table 1. As for the experimental
equipment, the section of the mesh representing the excavated area
was virtually lled with dry river sand, while the simulation of soil
outside the trench was performed by using expanded clay.
The conditions of the area surrounding the cable furnished to
the software are: a temperature of 333 K on the surfaces of the
underground cable, a temperature of the outside environment of
313 K and an adiabatic condition on the other limit surfaces of the
excavated area.
To perform mesh sensitivity analysis two more meshes were
realized: one which is less dense (Mesh 0.5X) composed by a
number of cells of 1.8 105 and another which is denser (Mesh 2X)
with a number of cells of 1.3 106. In order to determine the good
quality of the number, while solving the thermal exchange equations of the three meshes, the difference between the heat exiting
from the circular surface (representing the electric cable) and the
dispersion of heat through the upper surface of the mesh (representing the soil) were compared. Table 2 reports these differences.

Table 2
Percentage difference between the heat generated, for each linear meter, by the
underground cables and the one dispersed from the soil (at ground level) for the
mesh tested with reference to 1X Mesh.

D%

Mesh 0.5X

Mesh 1X

Mesh 2X

32.58%

2.21%

272

## Table 2 shows how 1X Mesh solves the Equation (5) with a

higher accuracy, concerning the calculation, than 0.5X Mesh which
is less dense; we can notice how 2X Mesh (with a higher number of
nodes) does not present signicant improvements (for what concerns the accuracy of calculation) if compared to 1X Mesh. Once a
good level of numerical accuracy is reached, we can continue with
other simulations by using geometry of 1X Mesh, without the
necessity to make estimations with a higher number of nodes.
2.3. Numerical model validation
In order to make a comparison of the output of the numerical
simulation performed thanks to the experimental data, and determine their good quality, it was taken into consideration the value of
the soil thermal resistance inside the cable and the trend of the
temperature provided by the FEM model respect to the experimental data of the temperatures supplied by the thermocouples
placed in the excavated area above the cable.
During the numerous experimental tests, the extremities of the
pipe simulating the underground cable were supplied with electric
energy and heated, thanks to the Joule effect, until the pipe reached
a temperature of 61  C (334 K), whereas the environment of the
laboratory presented a temperature of 17  C (290 K). Once the
system reached a state of stability, these conditions were kept for a
time interval of 6 h. The complete model was then simulated, at a
stationary speed, with a scenario characterized by the same surrounding conditions. While making an estimation of the soil thermal resistance, the simulations presented an overestimation of
3.69% of the value measured in the laboratory.
Fig. 4 shows, right in the section between the heated pipe and
the ground level, the trend of the temperatures (these temperatures were normalized if compared to the air temperature of the
environment) of the numerical model respect to the values estimated through the thermocouples.
The maximum deviation ratio between the simulated and
experimental temperatures is 0.6%. The results reported lead to
consider the approximations of the numerical model as satisfying
results for its validation.
3. Results
3.1. How the geometry of the excavated area affects the soil thermal
resistance

## underground cable as a geometrical parameter, but do not take into

consideration the shape and the materials lling the cable [16].
Sometimes, during the phase of realization, it is hard to respect the
instructions provided by the plan to create an optimal geometry of
the excavated area. Moreover realizing an excavated area is
expensive and during their realization the regulations are the only
parameters to respect; if there is any precise regulation the tendency is to carry out the simplest solution on the eld. Sometimes it
happens that once the conductor is laid over the sand bed, the
trench is not lled, due to economical reasons, with sand but with
the soil turned over to dig the trench. In this way the underground
electric cable is surrounded by different conditions, if compared to
those assumed by the regulations conceived to give instructions for
the estimation of the soil thermal resistance.
To examine these elements, while using the validated numerical
model, the geometrical parameters (Fig. 5) of the excavated area
were varied slightly and it was estimated the soil resistance for each
simulated case.
The geometrical parameters of the excavated area are dened as
follows:






## h: depth of the underground cable respect to the ground level;

d: thickness of the sand bed above the cable;
c: whole thickness of the sand bed;
b: width of the excavated area;
D: diameter of the underground cable (0.011 m).

## The following gures represent the extreme positions of the

geometry characterizing the excavated area. On the left there is the
geometrical conguration of the excavated area considered as a
point of reference, in the middle there is the displacement of the
cable or the corners of the geometry characterizing the trench and
on the right the nal conguration. Among the extreme positions
represented, some intermediate geometrical congurations, subject to numerical simulations, were realized.
Here is the description of the 4 cases:
1. Variations of the depth values characterizing the underground
cable laying (Fig. 6): the underground cable, located in an
excavated area, completely lled with sand, was shifted to the
minimum values (h value included between 0.5 m and
1.2 m) according to the values established by the regulations.
With a 0.1 m variation of the position of the cable, it is

## For the estimation of the soil thermal resistance, through

Equation (1), the current regulations consider the depth of the

Fig. 4. Trend of the temperatures: simulated values against experimental values according to the distance from the pipe.

## F. Salata et al. / Applied Thermal Engineering 78 (2015) 268e277

273

Fig. 6. Variation of the depth values characterizing the underground cable laying in a trench lled with sand.

## possible to examine 8 different geometries (congurations:

1.1, 1.2, 1.8).
2. Variation of the thickness of the ll, sand or dirt (Fig. 7): as
previously said, the excavated area is seldom covered until the
ground level with sand only; due to economical reasons the ll
is often made of a layer of sand right above the conductor, followed then by the covering of the rest of the space with the dirt,
previously dug, up to the ground level. While keeping the cable
at a depth h (whose value is taken as a point of reference), 11
intermediate congurations, with a 0.1 m variation of the
geometrical parameter d (congurations: 2.1, 2.2, 2.11),
were examined.
3. Variation of the dimensions of the excavated area with the same
installation depth (Fig. 8): while keeping the underground cable
at the same depth h, and the dimension of the excavated area
being equal, the geometrical parameters c and b were varied.
In this way 8 different geometrical congurations (congurations: 3.1, 3.2, 3.8) were examined.
4. Variation of both the dimensions of the excavated area and
installation depth (Fig. 9): with the dimension of the excavated
area being equal, the geometrical parameters c and b were
varied in the same way as in the previous case, but at the same
time the depth, h, of the underground cable was subject to a

variation. 8 different geometrical congurations (congurations: 4.1, 4.2, 4.8) were examined.
The geometrical parameters of every conguration examined
are reported in Table 3. Every conguration is identied through an
ID composed of two gures divided by a point. The rst gure refers
to the case, the second represents the geometrical conguration.
The geometry considered as a point of reference is the same for the
4 cases and the second gure is number 1 (1.1 2.1 3.1 4.1).
The last column of Table 3 reports the ratio between the soil
thermal resistance estimated through the output data of the FEM
simulations respect to the values calculated (in accordance with the
regulations) through Equation (1).
While examining the results reported in the last column of
Table 3 and the graph in Fig. 10, it can be noticed how, in every case
analyzed, the values of the soil resistance are higher than those
provided by the regulations, in accordance with what was said in
the previous paragraph. This represents a demonstration of the
validity of the superheating problems affecting the exerted cables
planned by taking into consideration the data provided by the
regulations.
Hence it is possible to determine a dimensionless multiplicative
factor f1 (function of the geometrical parameters considered)

## Fig. 7. Variation of the thickness of the ll, sand or soil.

Fig. 8. Variation of the dimensions of the excavated area with the same installation depth.

274

## F. Salata et al. / Applied Thermal Engineering 78 (2015) 268e277

Fig. 9. Variation of the dimensions of the excavated area and installation depth.

## correcting the soil thermal resistance value estimated through

Formula (1) provided by the regulations:



 c a  d b  h g
RFEM
c d h
A\$
f1 ; ;
\$
\$
b D D
b
D
D
RIEC

(6)

## The adimensionalized and independent geometrical parameters

are:
 c/d: ratio between the thickness of the excavated area made of
sand and its section above the underground cable;
 d/D: ratio between the portion of the excavated area made of
sand above the underground cable and its diameter;

Table 3
Geometrical parameters of the different geometries examined through the FEM
model and the ratio between the thermal soil resistance, used for the simulations,
and the one estimated by the regulations.
RFEM
RIEC

Conguration

h

## 1.1 2.1 3.1 4.1

1.2
1.3
1.4
1.5
1.6
1.7
1.8
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
3.2
3.3
3.4
3.5
3.6
3.7
3.8
4.2
4.3
4.4
4.5
4.6
4.7
4.8

1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5

1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
1.1
1.0
0.9
0.8
0.7
0.6
0.5

0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.7500
0.8082
0.8764
0.9568
1.0358
1.1726
1.3214
1.5068
0.8082
0.8764
0.9568
1.0358
1.1726
1.3214
1.5068

1.3875
1.3875
1.3875
1.3875
1.3875
1.3875
1.3875
1.3875
1.2875
1.1875
1.0875
0.9875
0.8875
0.7875
0.6875
0.5875
0.4875
0.3875
1.2875
1.1875
1.0875
0.9875
0.8875
0.7875
0.6875
1.2875
1.1875
1.0875
0.9875
0.8875
0.7875
0.6875

1.711
1.636
1.567
1.501
1.434
1.369
1.305
1.214
1.870
1.993
2.103
2.209
2.313
2.415
2.522
2.640
2.765
2.903
1.829
1.915
1.991
2.064
2.138
2.217
2.304
1.607
1.508
1.416
1.329
1.252
1.187
1.135

 h/D: ratio between the depth of the underground cable and its
diameter.
Thanks to the exertion of the statistical function nlint of the
MatLab software it was possible to determine the constant values of
the relation (6). The values estimated are:
A 0.1218;

a 0.0443;
b 0.3273;
g 0.9003.
The design engineer facing these kind of problems can use the
formula (a formula which can be applied easily) provided by the
regulations (1) corrected through the multiplicative factor f1, in
this way the result is a soil thermal resistance whose value reects
the reality. It must be stressed that this is a precautionary value
since it refers to very dry soils [26e28]. The good quality of the
relation (6), used to determine the corrective coefcient f1, is
reported in Fig. 11 showing a maximum deviation of 8% respect to
the values used for the estimation (they are reported in the last
column of Table 3). The corrective coefcient f1 is valid in those
values characterizing the geometric parameters reported through
Table 3.
3.2. How the presence of other cables in the excavated area around
the underground cable affects the soil thermal resistance
Sometimes it happens that, for technical and economical reasons, many cables are installed in the same excavated area. The

Fig. 10. Ratio between the simulated soil thermal resistance and the one of the regulations for each geometric conguration examined.

## F. Salata et al. / Applied Thermal Engineering 78 (2015) 268e277

275

Table 4
Geometrical parameters of the congurations with the 3 cables examined through
the FEM model and ratio between the resistance of the soil for the simulations and
the one of the conguration with just one cable taken as a point of reference
assumed by the regulations.
Conguration

Fig. 11. Trend of the soil resistance estimated through the correlation (6) respect to the
value provided by the simulations.

## presence of other linear heat sources around the cable determines a

higher thermal stress, hence more problems caused by heat
disposal [29,30]. To examine these issues, we assumed the presence
of two more cables (with the same diameter) positioned horizontally (case A), vertically (case B) or a triangular mesh installation
(case C) around the cable, thus presenting more difculties for what
concerns the heat transfer (in bold, Fig. 12).
The geometrical parameters, useful for the characterization of
every different case, are:
 i: axle spacing between the cables;
 e: distance between the ground level and the underground cable with the lowest value of depth;
 h: depth of the cable laying respect to the ground level;
 D: diameter of the underground cable.
The diameter of the underground cable (D 0.011 m) and its
installation depth (h 1.2 m) present constant values for every
conguration examined.
While using the FEM model, for each case were analyzed 5
different congurations (.a, .b, , .e). For case A and case B, in
particular, the axle spacing i was subject to a progressive increase
of its value for every conguration (2D case.a; 4D case.b;
8D case.c; 12D case.d; 24D case.e). Instead in case C,
while still using, during the simulations, the structured mesh previously described, the axle spacing was subject to the same increase
of its horizontal and vertical projection.
The geometrical parameters of every conguration examined
are summarized in Table 4. Each conguration analyzed is identied through an ID presenting two letters divided by a dot. The rst

A.a
A.b
A.c
A.d
A.e
B.a
B.b
B.c
B.d
B.e
C.a
C.b
C.c
C.d
C.e

## Geometrical parameters [m]

i

0,0221
0,0441
0,0883
0,1324
0,2648
0,0221
0,0441
0,0883
0,1324
0,2648
0,0312
0,0624
0,1248
0,1873
0,3745

1,2000
1,2000
1,2000
1,2000
1,2000
1,1559
1,1117
1,0235
0,9352
0,6704
1,1779
1,1559
1,1117
1,0676
0,9352

RFEM
RIEC

1,9175
1,8460
1,7715
1,7283
1,6688
1,8955
1,8011
1,6818
1,5923
1,3785
1,8805
1,7933
1,6897
1,6190
1,4773

letter is in capital letters and represents the case, the second letter
is in lowercase and represents the geometrical conguration. The
geometry, taken as a point of reference for all 3 cases (A, B, C), is the
one presenting just one cable and the same adopted in the previous
paragraph.
While analyzing the values reported in the last column of
Table 4 and the graph in Fig. 13, the soil resistance seems to be, in
every case, higher than the one furnished by the regulations. This
conrms the superheating problems characterizing the cables if
installed in the same excavated area.
As in the previous paragraph, it is necessary to determine a
dimensionless multiplicative factors f2 (function of the geometrical parameters considered) able to correct the value of the soil
thermal resistance estimated through Formula (1) provided by the
regulations.


 d     z

RFEM
i i e
i
i
e
; ;
B\$
f2
\$
\$
D h i
D
h
i
RIEC

(7)

## The adimensionalized and independent geometrical parameters

are:
 i/D: ratio between the axle spacing of the cables and the
diameter of the underground cable examined;
 i/h: ratio between the axle spacing and the installation depth of
the underground cable examined;

Fig. 12. Geometrical parameters of the distance characterizing the three cables in case A (horizontal disposition), case B (vertical disposition), case C (triangular disposition)
inside the excavated area.

276

## F. Salata et al. / Applied Thermal Engineering 78 (2015) 268e277

Fig. 13. Ratio between the soil thermal resistance simulated and the one taken as a
point of reference for every geometrical conguration examined.

Fig. 14. Trend of the soil resistance estimated through correlation (7) respect to the
value obtained with the simulations.

##  e/i: ratio between the installation depth of the closest cable to

the ground level and the axle spacing between the cables.
The constants in the relation (7) can now be determined:
B 0,0096;
d 1,0544;
1143;
z 0,0019.
In this case the design planner can use the formula provided by
the regulations (1) corrected through the multiplicative factor f2.
The good quality of the relation (7), useful to determine the
corrective coefcient f2, is reported in Fig. 14 showing a
maximum deviation of 10% respect to the values used for its
estimation (reported in the last column, Table 4). The corrective
coefcient f2 is valid in those values characterizing the geometrical
parameters reported through Table 4.

## When dealing with underground electric cables installation, the

estimation of the soil thermal resistance is vital to avoid a superheating of the cables, caused by the Joule effect, that can lead to a
drastic decrease of the service life of the insulation material [6]. The
maintenance of underground conduits is hard to perform and
hence it is economically heavier than other types of underground
installations. This is the reason why thermal problems must be
taken into consideration during the planning phase [13]. An underestimation of the value of the soil thermal resistance, especially
for extremely dry soils, can determine several problems. The
possible presence of humidity in the soil increases the capacity of
heat exchange, thus helping the disposal of excessive heat of the
cables, but it is necessary to perform a dimensioning of the cables
(assuming the worst condition possible, that is a dry soil [26]).
The IEC rules furnish [16], for the estimation of the soil resistance around the cables, a value that tends to underestimate the dry
soil thermal resistance in the excavated area. In this study it was
carried out the estimation of two dimensionless factors f1 (6) and
f2 (7), multiplicative of the value provided by (1), providing a
more accurate calculation of the resistance values useful for the
cables dimensioning. This is why it was reproduced, in a laboratory,
an experimental equipment permitting a validation of a bidimensional numerical model. Through this numerical model it was
possible to study the thermal problem which varies with the
changing conditions of the cable installation; the geometrical parameters of the excavated area were subject to variation to determine the corrective coefcient f1 and the interaction according to
the distance separating the cable examined from other cables
(installed in the same area), in order to calculate the coefcient f2.
The expressions (2), (6) and (7) can help design engineers to
estimate the soil thermal resistance value for both the dimensioning of underground electric conduits and other situations
requiring the science of engineering, such as: planning of
geometrical heat exchangers, underground pipelines for the
transportation of hydrocarbons and every time there is a soil heat
exchange.
The benet of using the corrective coefcients here suggested is
a correct determination of the soil thermal resistance surrounding
the electric cables. This can help the right dimensioning of the
cables while keeping in mind the problem of the heat dissipation
generated by the joule effect; in this way malfunctioning during the
regular exertion of the cables can be avoided. On the other hand, in
order to use correctly the formula here suggested, some information concerning the geometry of the excavated area where the cables are placed, are necessary.
The results obtained in this case study are not interesting for the
scientic community only, but they can represent a guideline for
this type of industry as well. The corrective factors can be used
directly by designers of underground power lines.
This research will have a further development through the study
of the parameters variability impact, such as: cables diameter and
the surrounding conditions (thermophysical properties of the
materials, temperatures, etc).
Acknowledgements

4. Conclusions
The study here presented is an approach to improve the standard formula for the estimation of the thermal resistance of dry
soils surrounding the underground electric cables. Experimentally
it was demonstrated that the regulations tend to underestimate the
thermal issue of heat dissipation towards the soil causing malfunctioning problems affecting the cables which turn out to be
underpowered. Corrective coefcients were estimated and discussed by examining the numerical results obtained.

## This research received no specic grant from any funding

agency in the public, commercial, or not-for-prot sectors. A special
thanks to Mrs. Flavia Franco for the help she provided in the
preparation of this paper.
Nomenclature
b
c

## the width of the excavation [m]

the total thickness of the sand backll [m]

c/b

## ratio of the size of part of the excavation composed of the

cable's sand bed [dimensionless]
d
the thickness of the sand backll on top of the cable [m]
D
external diameter of the cable [mm]
d/D
ratio of the thickness of the layer of sand on top of the
cable and the cable diameter [dimensionless]
e
distance between the underground cable with the lowest
value of depth and the ground level [m]
e/i
ratio between the installation depth of the closest cable to
the ground level and the axle spacing between the
cables [dimensionless]
f1
correctional factor affected by the variation of the
geometrical parameters of the excavated area
[dimensionless]
f2
correctional factor affected by the presence of other
cables in the installation area [dimensionless]
h
depth at which a cable is installed according to the works
plan [m]
h/D
ratio of the depth at which the cable was installed
compared to ground level and diameter of the
cable [dimensionless]
i
axle spacing between the cables [m]
i/h
ratio between the axle spacing and the depth of the
examined underground cable [dimensionless]
It
electric power [A]
L
distance of the axis of the cable from the surface of the
soil [mm]
RFEM
numerical value of the soil resistivity [m K/W]
RIEC
value provided by the regulations concerning the soil
resistivity [m K/W]
Rsper
experimental value of the soil resistivity [m K/W]
q
thermal power for each linear meter of the cable [W/m]
rp
value of the resistance of the precision resistor [U]
Tair
air temperature [K]
Tt
temperature of the cable in the experimental section [K]
u (2L/D) [dimensionless]
Vr
voltage drop produced on precision resistor [V]
Vt
voltage drop at the end of cable [V]
r
resistivity of the soil [K m/W]
References
 Statistics of AC Underground Cables in Power Networks, Technical
[1] CIGRE,
Brochure Ref. No.: 338 CE/SC: B1, 2007. Paris-France.
 Remaining Life Management of Existing AC Underground Lines,
[2] CIGRE,
Technical Brochure. Technical Brochure Ref. No. : 358 CE/SC: B1, 2008. ParisFrance.
[3] N. Kovac, N. Grulovic-Pavljanic, A. Kukavica, Generated heat within power
cable sheaths per unit time and volume, Appl. Therm. Eng. 52 (2013) 90e96.
[4] V. Chatziathanasiou, P. Chatzipanagiotou, I. Papagiannopoulos, G. De Mey,
B. Wiecek, Dynamic thermal analysis of underground medium power cables
using thermal impedance, time constant distribution and structure function,
Appl. Therm. Eng. 60 (2013) 256e260.
 Update of Service Experience of HV Underground and Submarine Cable
[5] CIGRE,
Systems, Technical Brochure Ref. No.: 379 CE/SC: B1, 2009. Paris-France.

277

## [6] X. Chen, L. Hu, Y. Xu, X. Cao, S. Gubanski, Investigation of temperature effect

on electrical trees in XLPE cable insulation, in: 2012 IEEE Annual Report
Conference on Electrical Insulation and Dielectric Phenomena (CEIDP) Montreal, Canada, 2012, pp. 612e615.
[7] D.N. Singh, K. Devid, Generalized relationships for estimating soil thermal
resistivity, Exp. Therm. Fluid Sci. 22 (2000) 133e143.
[8] G.J. Groeneveld, A.L. Snijders, G. Koopmans, J. Vermeer, Improved method to
calculate the critical conditions for drying out sandy soils around power cables, Proc. IEE 131 (2) (1984) 42e53 part C.
[9] R. De Lieto Vollaro, F. Salata, A. Vallati, Experimental study of the thermal
behavior of various types of soil in the presence of a heat source (in Italian), in:
Proceedings, 65 National Congress ATI e Domus de Maria (CA), 2010.
[10] R. de Lieto Vollaro, L. Fontana, A. Vallati, Experimental study of thermal eld
deriving from an underground electrical power cable buried in nonhomogeneous soils, Appl. Therm. Eng. 62 (2014) 390e397.
[11] R. de Lieto Vollaro, L. Fontana, A. Vallati, Thermal analysis of underground
electrical power cables buried in non-homogeneous soils, Appl. Therm. Eng.
31 (2011) 772e778.
[12] L. Exizidis, I. Papagiannopoulos, V. Chatziathanasiou, G. De Mey, B. Wiecek,
Evaluation of a buried power cable's thermal behavior using phase diagrams
and calculation of the phase difference between temperature and power,
Appl. Therm. Eng. 70 (2014) 770e775.
[13] IEC 60287-3-1/2 Electric Cables e Calculation of the Current Rating e Sections
on Operating Conditions, 1999. Geneva-Switzerland.
[14] M.A. El-Kady, Calculation of the sensitivity of power cable capacity to variations of design and environmental parameters, IEEE Trans. Power App. Syst.
PAS-103 (8) (1985) 2043e2050.
[15] K. Mochlinski, Assessment of the inuence of soil thermal resistivity on the
ratings of distribution cables, Proc. IEE 123 (1) (1976) 60e72.
[16] IEC 287-2-1: 1994, Electric Cables e Calculation of the Current Rating e Part
2: Thermal Resistance e Section 1: Calculation of Thermal Resistance, 1994.
Geneva-Switzerland.
[17] D.A. De Vries, Thermal properties of soils, in: W.R. van Wijk (Ed.), Physics of
Plant Environment, John Wiley, New York-USA, 1963.
 Current ratings of cables buried in partially dried out soil. Part 1:
[18] CIGRE,
simplied method that can be used with minimal soil information, Electra 104
(1986) 11e22.
 Determination of a value of critical temperature rise for a cable backll
[19] CIGRE,
material, Electra 145 (1992) 15e29.
[20] C.C. Hwang, Y.H. Jang, Extension of the nite elements method for thermal
analysis of underground cable systems, Electr. Power Syst. Res. 64 (2003)
159e164.
[21] A. De Lieto Vollaro, R. De Lieto Vollaro, F. Salata, A. Vallati, Implementation of a
CFD model to study the thermal eld within the soil in proximity of electrical
cables (in Italian), in: Proceedings, 66 National Congress ATI e Rende
(Cosenza), 2011.
[22] UNI 10351, Building Materials. Thermal Conductivity and Vapor Permeability,
1994. Milan-Italy.
[23] UNI EN ISO 10456, Materials and Building Products. Properties Hygrometric e
Tabulated Values Design and Methods for the Determination of Thermal
Values Declared and Designed, 2008. Milan-Italy.
[24] IEEE 442-1981, Guide for Soil Thermal Resistivity Measurements, 1981. New
York-USA.
[25] ANSYS FLUENT, Theory Guide, 2009. Canonsburg-USA.
[26] A.N. Arman, D.M. Cherry, L. Gosland, P.M. Hollingsworth, Inuence of soil
moisture migration on power rating of cables in H.V. transmission, Proc. IEE
111 (1964) 1000e1016.
[27] F. Donnazi, E. Occhini, A. Seppi, Soil thermal and hydrological characteristics
in designing underground cables, Proc. IEE 126 (6) (1979).
[28] H. Brakelmann, Physical Principles and Calculation Methods of Moisture and
Heat Transfer in Cable Trenches, EU Report 19, VDE Verlag - Germany, Berlin,
1984.
[29] R. de Lieto Vollaro, L. Fontana, A. Quintino, A. Vallati, Improving evaluation of
the heat losses from arrays of pipes or electric cables buried in homogeneous
soil, Appl. Therm. Eng. 31 (17e18) (2011) 3768e3773.
[30] F. Freschi, M. Tartaglia, Power lines made of many parallel single-core cables:
a case study, IEEE Trans. Ind. Appl. 49 (4) (2013) 1744e1750.