Electromagnetic and Thermal Analysis of High

Voltage Three-Phase Underground Cables Using

Finite Element Method

Stefania Conti, Member, IEEE Emanuele Dilettoso and Santi Agatino Rizzo
Dip. di Ingegneria Civile e Architettura Dip. di Ingegneria Elettrica Elettronica e Informatica
University of Catania University of Catania
[email protected] [email protected], [email protected]

Abstract— In this paper a three core high-voltage section III the details and results of electrical analysis are
underground cable used for urban power networks is analyzed given. Section IV explains the coupled electromagnetic-
by means of Finite Element Method. The electrical analysis is thermal analysis performed and shows some remarkable results
performed in order to investigate the electric stress within the of simulation. The author’s conclusions follow in section V.
cable, also in presence of defect in the insulation layer. The
electro-thermal analysis permits the full description of the cable II. THREE CORE HV UNDERGROUND CABLE SPECIFICATION
behavior not only in terms of electric and magnetic
performances, but also regarding the effect of the Joule heating
Underground cables are generally required to deliver 3-
on the surrounding ambient. phase power. Although for the purpose either three-core cable
or three single core cables may be used, the former solution
Keywords— High Voltage; Underground Three Phase Cables; presents certain advantages in terms of costs and power losses.
Finite Element Method; Electromagnetic Field Analysis; Thermal The considered cable is inspired to the CityCable of
Analysis NKTcables®, construction code 2X(FL)2YVF ST2Y [7]
developed especially for high voltage (up to 132KV)
I. INTRODUCTION installations in urban areas as well as in industrial compounds.
Nowadays the three phase underground cables have a It provides compact and strong three cores design, fast and
substantial role in the power delivery distribution network and, cheap retrofitting of existing installation and no risk of
consequently, it is essential a complete knowledge of the cable environmental pollution. It consists of different layers as shown
behavior in order to evaluate performance and to prevent in Fig.1. The material used for the three circular phase
ageing, degradation and fails. Moreover, installation of High conductors is stranded copper; the insulating material is the
Voltage (HV) cables in urban area is non-trivial. In order to fit Cross Linked Polyethylene (XLPE). In order to smooth the
well in urban area they often are hidden in underground field distortion, a semi-conductor layer is used both around the
tunnels, with preexistent installations, like gas pipes, water copper and the XLPE insulation. Each single core is
pipes, telecommunication cables [1]. Hence, it is necessary to surrounded by a coated aluminum foil and a polyethylene over-
evaluate the impact of electromagnetic fields and heating sheath: this combination, called aluminum polyethylene
produced by power cables both on other installations and live laminated (APL) sheath, ensures high short circuit capability
organisms. Nevertheless, public urban MV and LV and water-tightness combined with lowest, weight and smallest
installations tend to have complex implementation especially in dimensions. The group of the three cables is sheathed by
the so called Smart Grids, characterized by continuously polyethylene.
increasing loading of distribution systems, also due to the Both wire armor and external pipe are made of steel. Finally,
presence of distributed generators injecting power into the the external covering of the pipe is a polyethylene layer.
networks [3]. Geometrical data of the cable are given in Table I. The laying
Finite Elements Method (FEM) is an effective tool, depth of the cable is p=1.2m, i.e. the standard depth reported in
characterized by good accuracy and reliability that permits the technical data [7]. Voltages and currents of the phases are a
study of cables behavior under different aspects, such as the balanced 3-phase set at frequency f=50Hz. Electrical, magnetic
electromagnetic, thermal and mechanical, also by means of and thermal properties of the materials employed in this model
coupled analyses [3-5]. The computational cost, generally the are given in Table II.
major drawback of applying FEM, is, in this kind of The FEM analysis was performed in COMSOL Multiphisycs
application, quite irrelevant, being possible to perform the [8]. The FEM mesh used consists of 15930 triangular elements;
study of cables by using e simple 2D analysis. the minimum element quality is 0.07321 and the average
In this paper FEM is used for electromagnetic and thermal quality is 0.8103. The chosen mesh represents a good trade-off
study of three core HV cable. The paper is organized as between computational costs and solution accuracy. Fig.2 plots
follows. In section II the examined cable is described. In the mesh in the cable area.

978-1-5386-5186-5/18/$31.00 ©2018 IEEE

 TABLE I. GEOMETRICAL DATA

Parameter Values[mm]
Diameter of conductor 18.4

Thickness of inner semi-conductor 1.4

Thickness of XLPE insulation 11

Thichness of outher semi-conductor 1.4

Thickness of APL sheath 2

Thickness of Polyethylene sheat 1

Thickness of Steel wire armour 2

Thickness of Steel pipe 4

Thickness of Polyethylene covering 1

Overall diameter dc 135

Fig. 1. Geometry of the cross section of the cable [m]

III. ELECTRICAL ANALYSIS

A 2D in-plane current conservation problem in the
frequency domain is achieved in order to investigate the
electric stress within the cable. Under the assumptions of quasi-
static time-harmonic steady-state behavior, supposing that the
electric field is curl-free, the problem can be conveniently
analyzed in terms of electric potential, by solving the
differential equation:

-∇∙ σ+jωε ∇V =0 (1)

where V is the electric potential, σ is the electric conductivity,

ω=2πf is the angular frequency and ε=εrε0 is the electric
permittivity. The boundary value problem is completed by the
following Dirichlet boundary conditions on the three phase
conductors
2π 2π
Fig. 2. FEM mesh in the cable area
V1 =V0 , V2 =V0 e-j 3 , V3 =V0 e+j 3 , (2)
combined with Vg=0 on ground.
Note that supposing the electric field E curl-free implies: TABLE II. MATERIALS PROPERTIES

∇× E = − jB=0 (3) εr μr
σ κ
electrical thermal
As ω isn't zero, and considering in this model the electric relative relative
Material conductivity conductivity
permittivity permeability
fields only in plane xy, the z-component Bz of the magnetic [S/m] [W/(m·K)]
flux density B must be equal to zero. Subsequently the solution Copper 1 1 5.99e7 400
of (1), the electric field E can be obtained by: Steel 1 4 4.7e6 50.2
E = − ∇V (4) Aluminum 1 1 3.77e7 237

The current density J is calculated using the following Polyethylene 2.25 1 1.0e-18 0.46
constitutive relations: Semiconductor 2.25 1 2 10
J=σ E (5) XLPE 2.5 1 1.0e-18 0.46
In the case investigated the RMS value of the line voltage Soil (dry) 1 1 28 0.4
is 132KV, then V0 =132/√3)KV.
 An Intel® CoreTM i5-6200 CPU (2.3GHz) was used for the
analysis; the linear solver found the solution in about 4s. The
voltage distribution in the cable is shown in fig.3; as can be
seen the voltage at the steel sheath is approximately zero. Fig.4
plots the electric field stress along the radius of a phase. The
result agrees with the cable technical data, that report a
maximum field strength at conductor screen and at core screen
of 10.3 KV/mm and 4.9 KV/mm respectively [7].
It is well known that a stronger and inhomogeneous
electrical field can be caused by voids, bubbles, or defects in
the XLPE insulation [6]. Fig. 5 shows the results of inserting a
void-defect in the XLPE layer of phase 2; the void defect
causes a distortion and an increase of the electric field.
Moreover, this effect becomes stronger as the distance d from Fig. 5. Electric field [V/m] distribution in presence of a void-defect
the conductor decreases. Fig. 6 plots the electric field stress
through the defect when d=5mm (a) and d=2mm (b): as might
be seen, in the latter case the value of the electric field exceeds
the maximum values expected in the technical data.

(a)

Fig. 3. Electric Potential plot [V/m] in the cable area

(b)

Fig. 6. Electric field stress [KV/mm] in radial direction in presence of a

void defect in the XLPE layer (a) d=5mm (b) d=2 mm

IV. COUPLED ELECTROMAGNETIC THERMAL ANALYSIS

The aim of this analysis is the evaluation of per-unit-
length parameters and the study of the behavior of the cable at
operating temperature. At first the electromagnetic is solved,
and the solution thus obtained is used to compute the heating
power density inside the conductors, from which a steady-state
thermal analysis is performed by considering the same mesh.
Fig. 4. Electric field stress [KV/mm] of the cable in radial direction The electrical conductivity of the conductor is assumed to be
temperature-dependent and so, although the electromagnetic
problem is linear, an iterative procedure needs to be used to where Is is the current coming from an outside source and Acoil
solve the coupled problem [9]. is the cross sectional surface of the core. The 3-phase set of
currents in the coils is:
The scope of magnetic analysis is the solution of Maxwell–
2π 2π
Ampère's law in the frequency domain, using the z-component
I1 =I0 , I2 =I0 e-j 3 , I3 =I0 e+j 3 (14)
of the vector potential A as unknown. Indeed, being the
magnetic flux density B a solenoidal vector field, it is possible being I0=406A the maximum value of current, as in data sheet.
to define A as: The ambient temperature considered in the thermal analysis
is T=15°C as indicated in technical data of the cable. The
B=∇×A (6) iterative non-linear solver found an accurate solution in five
The equation to solve in the FEM domain is the following iterations and about 25s. The distribution of the norm of the
curl-curl equation: magnetic flux density in the cable is shown in Fig. 7. Fig. 8
plots the temperature distribution in the cable. The maximum
−ω2 εA+jωσA+∇× μ-1 ∇×A =Js (7) temperature of the cable is 77.35 °C, that is the operating
temperature of real cables (about 70 to 90°C). Fig. 9 shows the
where μ=μrμ0 is the magnetic permeability and Js is the applied temperature in the space around the cable: note that the when
source current density. The values of other fields can be the cable carries the maximum current value the soil surface
derived from A by means of (6), Maxwell equations and temperature raise up from 15°C to about 35°C.
constitutive relations. In particular, the conduction current J is
calculated by:
J=σE= − jωσA (8)
The solution of (8) is used to compute the heating power
density Q inside the conductors,
2
J
Q= 2σ (9)

from which a steady-state thermal analysis is performed by

solving the Poisson differential equation:
-∇∙ ∇T =Q (10)
where T is the temperature and κ is the thermal conductivity. It
is faithful to assume that the thermal response time is much
larger than the transient time of the electromagnetic fields. Fig. 7. Magnetic flux density [T] in the cable
The FEM domain for the electromagnetic analysis is a circular
domain with a diameter equal to 5dc, and on its boundaries it is
imposed the condition:
n×A=0 (11)
being n the outward unit vector to the boundary.
The electrical conductivity varies with the temperature as:
σ0
σ T = (12)
1+α(T-Tref )
where σ0 is the conductivity at temperature Tref =20°C
=293.15K and α is the electrical resistivity temperature
coefficient. The values of σ0 are those in Table II; α is equal to
0.00386K-1 for the copper and 0.00390K-1 for the aluminum,
whereas the variation of the conductivity of steel with
temperature is negligible. The coils are modelled as multi-
conductors (stranded) coils, i.e. are not subject to skin effect
and, in this case, the source current density Js is equal to:
Is
= (13) Fig. 8. Cable temperature [°C]
Acoil
Fig. 9. Soil and air temperature [°C] Fig. 11. Steel armor current density [A/m2]

The per-unit-length parameters at operating temperature

were calculated and are reported, together with technical data,
in Table III. There is a good matching between the declared
data and the results of FEM simulation. Table IV shows the
undesired power losses for sheath, armor and pipe. The
corresponding current densities are shown in Fig. 10, Fig. 11
and Fig. 12 respectively.

TABLE III. CABLE PER-UNIT-LENGTH PARAMETERS

Conductor AC
Inductance
resistance
[mH/km]
[Ω/km]
FEM simulation 0.1001 0.35
Technical data
0.0975 0.38
(operating at 90°)
Fig. 12. Steel pipe current density [A/m2]
TABLE IV. CABLE LOSSES

Power losses [kW/km] V. CONCLUSION

Aluminum sheath Steel armour Steel pipe
A three Core HV underground cable used for urban power
4.18 1.03 0.603 network was analyzed by means of Finite Element Method.
Electrical analysis showed correctness of technical data, it was
demonstrated as the presence of defect in insulation layer could
cause relevant peaks of the electric field strength, higher than
the maximum values declared. Thanks to the electro-thermal
analysis was performed the calculation of per-unit-length
parameters at operative temperature; undesired power losses
were also evaluated. It was pointed out that when the cable
carries high currents, the temperature of the soil rises
considerably.
REFERENCES
[1] M. A. Prsa, N. D. Mucalica, K. K. Kasas-Lazetic, “Determination of
electromagnetic field in a three-phase system with three single-core
power electric cables 110 KV”, procedings of IEEE EUROCON 2009,
May 2009.
[2] J. Cho, K. Sim, J. Bae, H. Kim, J. Kim, K. Seong, H. Jang, C. Lee and
D. Koh, “Design and Experimental Results of a 3 Phase 30 m HTS
Power Cable”, Ieee Transactions On Applied Superconductivity, Vol.
16, No. 2, June 2006, pp. 1602-1605
Fig. 10. Aluminum sheath current density [A/m2] [3] S. Conti, S. Raiti, G. Tina, U. Vagliasindi, “Distributed Generation in
LV distribution networks: voltage and thermal constraints”, in Proc. of
 2003 IEEE PowerTech Conference, Vol.2, 23 June 2003, Bologna, pp. [6] M Alsharif , P. A. Wallace, D. M- Hepburn and C. Zhou, “FEM
413-418. Modelling of Electric Field and Potential Distributions of MV XLPE
[4] C. Kocatepe, C.F. Kumru an T. Ozturk, “Analysis of eccentric axial high Cables Containing Void Defect”, Proceedings of the 2012 COMSOL
voltage underground cable”, 2011 7th International Conference on Conference in Milan, Milan (Italy), 2012.
Electrical and Electronics Engineering (ELECO), Bursa, Turkey, [7] http://www.nkt.com/industries/power-transmission
December 2011 [8] https://www.comsol.com/documentation
[5] M. Kumar, M. S. Bhaskar,S. Padmanaban, P. Siano, F. Blaabjerg and Z. [9] G. Aiello, S. Alfonzetti, E. Dilettoso, N. Salerno, "Transient Thermal
Leonowicz, “Electric field analysis of extra high voltage (EHV) Analysis of an Eddy-Current Heated Conductor Applying FEM-DBCI",
underground cables using finite element method.”, Proceedings of the IEEE Transactions on Magnetics, vol. 49, n. 5, 2013, pp. 1861 - 1864
2017 IEEE International Conference on Environment and Electrical
Engineering and 2017 IEEE Industrial and Commercial Power Systems
Europe (EEEIC / I&CPS Europe). IEEE Press, 2017.