Increasing Reliability of XPLE-Insulated Cables:

FEA-Based Ampacity Calculation and Online

2021 IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineering (ElConRus) | 978-1-6654-0476-1/20/$31.00 ©2021 IEEE | DOI: 10.1109/ElConRus51938.2021.9396638

Monitoring of Insulation Degradation

Mikhail Doronin Simon Dubitsky,
Higher school of HV eng Higher school of HV eng
Peter the Great polytechnic university Peter the Great polytechnic university
St. Petersburg, Russia St. Petersburg, Russia
mvdoronin93@gmail.com dubitskij_sd@spbstu.ru

George Greshnyakov Artem Egupov

Higher school of HV eng Peter the Great polytechnic university
Peter the Great polytechnic university St. Petersburg, Russia
St. Petersburg, Russia egupov_a@spbstu.ru
g.greshnyakov@gmail.com

Abstract— The length of service of a HV power cable line is of electromagnetic and thermal field. The possibility of using
determined by accurate prediction of its temperature state during equivalent circuits relies to some assumptions and
normal and emergency operation. In addition to calculating line simplifications about the nature of the magnetic and thermal
ampacity according to IEC 60278 standard, it is proposed to use field patterns. The standard includes rather complicate
FEM simulation to account for mutual electromagnetic and procedures for considering the electromagnetic interaction of
thermal influence of several cable lines. neighboring cables. The full-blown analysis according to the
During the operation of the line it is necessary to monitor IEC 60287 in case of several parallel cable lines appears so
online the parameters reflecting the degradation of the cable complex that there is the need of dedicated software, such as
insulation. Gradual degradation of cable insulation is a complex CymCap, ETAP [2] and others. A big number of real-life
physical and chemical process. It is reflected in some electrically knowledge and experience are incorporated to the IEC 60287 by
measured parameters available for long-term monitoring without means of empirical factors and terms.
interrupting normal operation.
The frequency dependence of loss factor tan δ is proposed as a An alternative approach is the direct numerical simulation of
parameter that is sensitive to cable degradation. Proposed the way electromagnetic and thermal field by a general purpose Poisson
of estimation of residual resource of cable line by analyzing the equation solver. The geometrical nature of most cable lines
rate of change of frequency dependency of loss factor. allows to use a simple 2D analysis due to translational symmetry
of a cable line. A modern easy to use 2D solver allows to
Keywords—cable ampacity; multiphysics FEA simulation; AC consider many significant factors, such as full details of the cable
cable losses; sheath losses; cable shield bonding; residual life design, exact cable line formation, including but not limited to
prediction; dissipation factor monitoring
the spaced flat and touched trefoil formations, thermal backfill,
I. AMPACITITY CALCULATION concrete or steel cable tray, e.t.c.. The circulating currents in
cable shield and/or metallic sheaths can be handled by the
The design of a cable power lines includes an assessment of coupling of electromagnetic field equations with Kirchhoff
their current carrying capacity. The primary goal of ampacity equations of grounding circuit.
calculation is predicting the cable temperature by normal
operation with rated current. Such analysis is performed in The field simulation approach to the cable ampacity
steady-state mode. Also, transient thermal analysis is used for calculation not only works well for standard steady state
estimation of how long the overcurrent can be allowed until the condition with symmetric 3-phase current system, but also easily
cable core reaches the temperature limit. consider the asymmetric current load, and higher harmonics of
the current.
Normally the cable line ampacity is calculated according to
the IEC standard 60287-2009. The core of the standard describes Generally, the ampacity calculation procedure is an inverse
procedures for 1) refinement of AC losses due to skin effect and problem of finding the source current by its thermal effect. It is
proximity effect, additional losses in metallic shield and practically treated by repetitive solving of a number of direct
conductive sheath with one-side and both sides shield bonding, problems, which calculate the cable temperature by given
and 2) calculation the temperature of each cable core with given current. Each direct problem is a multiphysic analysis [3],
losses and given cooling conditions. includes at least two kind of solutions: the AC magnetic analysis
for refining the AC losses in all conductive shells of a cable, and
The standard is mainly based on the J. Neher and the heat transfer analysis that calculates the space distribution of
M. McGrath approach [1], developed in the late 50s of the 20th temperature by given losses. More elaborated approach my
century. This method employs equivalent circuit representation

1248 978-1-6654-0476-1/21/$31.00 ©2021 IEEE

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on June 19,2021 at 23:19:19 UTC from IEEE Xplore. Restrictions apply.
include also CFD analysis to refine the mass and heat transfer, where U is the potential difference between branch
for example, between a cable and a pipe. It is worth to note that terminals, R is the DC resistance of the branch (Ohm). The
the CFD analysis is normally much more complicated than both integration is performed over the conductor’s cross-sectional
magnetic and thermal analyses and should be avoided when area Ώ. The numerical solution of coupled equations (1) and (2)
possible by replacing the real gas or liquid media with thermally results in the distribution of the current density over all
equivalent solid media. conducting elements of the system (core, shield, armor, and
adjacent conductive structures).
An important side effect of a calculation of AC magnetic
field around a cable line is detailed knowledge of the magnetic The second stage involves solving the heat transfer equation
field pattern around a cable line that can affect the human health. in the same computational domain on the same finite element
When the magnetic field above an underground cable lines does mesh:
not fulfill the level set by national safety regulation, there is the
need to effectively control the external magnetic field Error! ∂  ∂T  ∂  ∂T  ∂T
Reference source not found.. That task can be also done with  λx  +  λy  = −q − c ρ
the same magnetic field simulation. ∂x  ∂x  ∂y  ∂y  ∂t
(3),
At the first stage, equations are solved for a quasi-stationary
alternating magnetic field of an industrial frequency of 50 Hz, where T – temperature, K; t – time, s; λ - specific thermal
taking into account eddy currents. The equations are written with conductivity, W/(m∙K); q – specific heat source, W/m3, s -
respect to the complex vector magnetic potential (MVP) A. In specific heat, J/(kg∙K), ρ is the mass density, kg/m3.
the case of translational symmetry there is only one nonzero The IEC 60287 electromagnetic and thermal models appears
component of MVP: A = Az [6] quite accurate in case of a single cable line, but may result in a
loss of accuracy for a cable assembly containing a number of
∂  1 ∂A  ∂  1 ∂A  closely located cable lines. To test this hypothesis, a series of
 +   = − jext + iωσ A ampacity calculations were carried out using two methods: IEC
∂x  µ ∂x  ∂y  µ ∂y 
(1) 60287 with help of the CymCap software, and the coupled FEA
simulation of AC magnetic field and heat transfer using the
where μ is permeability (H/m), σ– specific conductivity (S/m), QuickField software [5]. The objects for comparative analysis
ω– cyclic frequency (rad/s), jext – external current density are:
(A/m2).
a) Three-phase underground line with single-phase cable
The geometric model includes cross-sections of all cables, laying in a spaced flat formation and in a touched trefoil
thermal backfill, concrete tray, e.t.c. and a pretty amount of formations.
surrounding soil and air, limited by artificial far boundary with
zero field boundary condition. For cable ampacity calculation b) Concrete cable block 110/330 kV - for cable entry to the
there is no need to include to the model all cable layers, such as substation with 10 cable lines organized in 4x10 pipes. Each
semi conductive electric screens, water block layers, bedding cable is laid in a separate polyethylene pipe, the fourth pipe in
and so on. The cable homogenization procedure [4] includes each layer is left unoccupied for future maintenance. The
merging of neighboring layers with similar thermal electromagnetic and thermal interaction of big amount of
conductivity, as well as special correcting of electric and thermal individual cables is difficult to assess, whereas it affects
conductivity of a cable conductor and electrical shield. The significantly the temperature of individual cables.
conductivity is scaled down proportional to the ratio of The result of comparative study of IEC 60287 and
geometrical cross-section of the region to the rated copper cross- Multiphysics FEA leads to the following conclusion: both
section. method give almost same temperatures for individual cable lines
With both side bonding of the cable shield a closed loop is of conventional design [6], whereas with the complex block of
formed through which the induced current flows. In case of one- 10 cable lines the FEA-based calculation gives significantly
point shield bonding there is no closed loop, but induced eddy higher temperatures than the IEC 60287 analysis. Repetitive
currents still create non-zero losses in the shield. The drawback calculation with varying distance between cables shows clearly
of one-point shield bonding is induced voltage at the open end that the main reason of higher temperature is the strong
of the shield that can be hazardous to personnel. The induced proximity effect, that can be handled by the standard in only a
voltage is also calculated with an AC magnetic field analysis. To limited way.
correctly take into account the grounding scheme, it is necessary Numerical experiments clearly show that the ten closely
to solve the magnetic field equations (1) together with Kirchhoff located cable lines demonstrate much stronger electromagnetic
equations of the grounding circuit. For the current in a branch and thermal interaction that the IEC 60287 standard predicts. It
containing a solid conductor in a magnetic field, the following can be concluded that numerical modeling can provide the
equation applies: electrical designer with more complete data than the IEC 60287
standard does.
U
I= − γ  iωA ⋅ dS
R Ω (2)

1249

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on June 19,2021 at 23:19:19 UTC from IEEE Xplore. Restrictions apply.
 II. REMAINING LIFE SPAN ESTIMATION BY ONLINE
MONITORING
The choosing of an easy measuring parameter that is
sensitive to the overall condition of a cable insulation - a
performance criterion -is directly related to the study of the
physical mechanisms that lead to insulation aging. One of the
main factors here is the manufacturing process [7]. It should be
noted that the decrease in the operational life of cables with
XLPE insulation is largely due to structural changes in the
process of thermal aging [9].
Heat (thermal) aging is a complex process that is a
combination of factors. These includes: changes in the sub-
molecular organization of electrical insulation; structural
disturbances due to thermo-oxidative reactions; diffusion
exchange between insulation materials and polymer electrically Fig. 1. Dependence on temperature and frequency for cuts from insulation: 1-
conductive shields, other elements of the cable structure, as well initial, after aging at 358K, 2- 500h, 3- 2000h, 4- 3500h.
as the environment.
It should be noted that all experimental curves illustrating
Taking into account the fact that XLPE insulation contains thermal aging correspond to studies of RF cable sections with
antioxidants, it can be assumed that dielectric strength is affected insulation from thermoplastic low-density polyethylene
to a greater extent than thermal-oxidative degradation, composition 107-01K. The cable sections thermal aging tests of
diffusional transfer, and changes in morphology, namely, the RK75-24-17, RK50-44-17 brands were carried out in a heat
secondary crystallization, accompanied by an increase in the size chamber without applying voltage. It has been established that
of spherulite nuclei. the dependencies illustrated in Fig. 1 are preserved when
Based on the facts described, in particular, in [9, 10], it can measuring tan δ on pressed plates, cuts from insulation, outer
be reasonably estimated that thermal aging under operating sheath on cable segments [9].
conditions entails a decrease in dielectric strength by about one Taking into account all of the above, it can be argued that the
third, compared with the initial level. position of the maximum tanδ on the temperature or frequency
The choice of the criterion should assume the possibility of dependence corresponds to a certain structure polymer, and the
measurements along the entire cable length, as well as during ΔT and Δf values are the essence of the deviation due to aging
operation (online monitoring). This excludes the possibility of from the initial location of the maximum tanδ value in its
using directly structural and thermophysical parameters, since temperature and frequency dependences, respectively, should be
for their measurement it is necessary to violate the integrity of recognized as a measure of quantitative assessment of the aging
the cable. In other words, the choice of the structure-sensitive process.
parameter is limited by such characteristics, for the measurement When choosing a criterion (which one is ΔT or Δf), it is
of which it is not necessary to destroy the cable under test. necessary to take into account the measurement conditions. The
In the cross-linked polyethylene (XPLE), as well as in a measurement of the temperature deviation must be carried out
thermoplastic, there are two types of dielectric losses: dipole- with uniform сable heating, which impossible in practice. The
segmental (α-process) and dipole-group (β and γ-processes). For parameter Δf is more convenient for measurement, since when
XLPE, as well as for thermoplastic PE, it is important to study determining it, the sample is at a fixed operating temperature of
the α-process [9]. It was found that as the time of heat aging the cable, or it is de-energized, and it is only necessary to remove
increases, the intensity of relaxation processes (β and γ- the frequency dependence of tan δ (f). showed that for
relaxation regions) decreases, which leads to a decrease in tan δ vulcanized polymers the regions of maxima are wider, i.e. the
at the maxima of the temperature and frequency dependences , spectrum of relaxation times is wider.
but the position of the maxima remains at the same temperature The temperature range of the appearance of dipole-
or frequency. segmental polarization depends on the chemical structure of the
At the same time, as the aging time increases in the α- catalysts and vulcanization agents, the obtained density of the
relaxation region (the movement of polar groups (plasticizers, cross-linked polymer network: the greater the network density,
antioxidants, catalysts, etc.) over the surface of PE crystallites), the higher the temperature of the maximum loss for the dipole-
tan δ increases at the maximum of the frequency and segmental process.
temperature dependences, in addition, the maxima shift: The plasticizers introduction lowers the glass transition
frequency dependence in the low frequencies region, and temperature and therefore the maximum loss temperature.
temperature dependence in the high temperatures region of (Fig. Taking into account the noted regularities, it can be assumed that
1). for XLPE insulation the general rules will be valid, according to
which losses for dipole-segmental processes in thermoplastic PE
insulation appear.

1250

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on June 19,2021 at 23:19:19 UTC from IEEE Xplore. Restrictions apply.
 Assume that the parameter sensitive to the aging processes The water sorption also is a very important factor. So, the
of XLPE insulation is the dielectric loss tangent, tan δ. The sorption of 0.01-0.1% of the mass of H2O leads to an increase in
angle δ for a dielectric is the complement to 90⁰ of the σrest from 102 to 104 times. The degree of water dissociation
displacement angle φ between current and voltage in the increasing, residues of catalysts, dyes, etc. in polymer insulation
dielectric. Thus, if the instantaneous values of current and leads to a sharp increase in σrest. On the other hand, an increase
voltage and the displacement angle between them are fixed at in the degree of crystallinity leads to a decrease in σrest by two or
each harmonic, then it is possible to determine the angle δ of its more orders of magnitude.
tangent for each harmonic, which means that it is possible to
obtain the dependence of tan δ on the frequency online. The It should be noted that there is anisotropy in electrical
angle φ can be measured as follows (Fig. 2) conductivity in oriented polymers, in particular, linear PE. For
example, σrest is reducing along the direction of extrusion. It can
be argued that the polymer insulation aging and a decrease in its
resource leads to a significant increase in σrest.
Summing up, we can conclude that along with tan δ, the
electrical conductivity of the insulation and its dielectric
constant ε (capacitance of the insulating layer) can be used as
structure-sensitive aging parameters of a polymer insulation,
including the XLPE-insulation.
REFERENCES
[1] J. H. Neher, M. H. McGrath, “Calculation of the temperature rise and load
capability of cable systems,” AIEE Trans., Vol. 76, Part 3, 1957, pp. 755-
772.
[2] IEC Standard-Electric Cables – Calculation of the Current Rating – Part
2: Thermal Resistance – Section 1: Calculation of the Thermal Resistance,
IEC Standard 60287-2-1, 1994–12
Fig. 2. Measurement of tan δ (f) through phase voltage and current in isolation
[3] Korovkin N., Greshnyakov G., Dubitsky S. Multiphysics approach to the
boundary problems of power engineering and their application to the
On the fig.2 I is the total current through the insulation to analysis of load-carrying capacity of power cable line //Electric Power
earth, U0 is the phase to ground voltage Quality and Supply Reliability Conference (PQ), 2014. – IEEE, 2014. –
С. 341-346.
Place the dielectric current sensor at the grounding point of [4] S. Dubitsky, G. Greshnyakov, N. Korovkin Comparison of Finite
the shield (armor). In this case, the sensor that fixes the voltage Element Analysis to IEC 60287 for Predicting Underground Cable
vector (phase) is located on the capacitive tap (ring) of the socket Ampacity.- 2016 IEEE International Energy Conference (ENERGYCON)
part of the plug ("plug-in", socket-type) of the indoor 4-8 April 2016. DOI: 10.1109/ENERGYCON.2016.7514107
termination [5] S.D. Dubitsky, G.V. Greshnyakov, N.V. Korovkin, Management of the
magnetic field of a buried cable power line, St. Petersburg polytechnic
Another option for obtaining information about the phase university journal of engineering sciences and technology, vol. 23(03)
voltage vector is to position the sensor directly on the winding (2017) pp. 88–100, DOI: 10.18721/JEST.230308
of the supply transformer. Measurements should be made for [6] QuickField Finite Element Analysis System. Version 6.4 User’s Guide.
Tera Analysis Ltd., 2020. p. 251 [online resource]. URL:
each harmonic. The cable shield must only be grounded at one https://quickfield.com/free_doc.htm. (checkind date 30.11.2020)
side and at one point.
[7] Tadzhibaev A.I., Kaniskin V.A., Kostenko E.M. Non-destructive method
A completely informative characteristic for a non- for life time assessment of electrical cables with polymer insulation in
service conditions // Electrichestvo.– 1995.– #5.– pp. 19–23. (in Russian)
destructive resource assessment method could be directly the
[8] Peshkov I.B., Shuvalov M.Ju. Electrochemical aging of polymer extruded
polymer insulation electrical conductivity. The concentration insulation of power cables: research, diagnosis, resource assessment //
and mobility of charge carriers is influenced by a number of Izvestia RAN. Energetika.- 2008, - #1- pp.70-88. (in Russian)
factors. The main ones should be considered the physical [9] Shuvalov M.Ju. Research of high voltage cables, development of
structure of the polymer, the presence of impurities and special advanced electrical calculation methods and microdiagnostics. // Dr. Sci
additives. The rest conductivity σrest is a function of many Thesis. -Moscov: VIIKP, 2000, 342 p.
arguments. Refinery procedure LDPE from low molecular [10] Globus E.I., Kabanova N.G., Kostenko E.M., Lobanov A.M., Sazhin B.I.
weight impurities, can lead to a decrease in res by 1-3 orders of Assesment of cable polyethilene insulation by the method of dielectric
magnitude. losses. // Plastic masses, 1989.-# 2, pp.77-78. (in Russian)

1251

Authorized licensed use limited to: BOURNEMOUTH UNIVERSITY. Downloaded on June 19,2021 at 23:19:19 UTC from IEEE Xplore. Restrictions apply.