1

Electro-thermal Simulation of ZnO Arresters for

Diagnosis Using Thermal Analysis
E. T. Wanderley Neto and E. G. da Costa, UFCG
M. J. A. Maia, T. C. L. Galindo and A. H. S. Costa, CHESF

equipments present an air gap between varistors and housing

Abstract – Surge arresters are equipments of greater importance (Fig. 1). They also present a greater mechanical and electric
for the protection of electrical systems. Their choice for a specific resistivity. Polymeric insulated arresters usually do not present
substation line or equipment must follow a rigid selection and an air gap between ZnO and housing (Fig 2). In most cases, the
testing. The maintaining of these equipments by the realization of
regular monitoring also plays a major rule for the system
polymeric material is molded directly on the varistor columns.
integrity – an arrester failure may result in an its explosion or
system turn off. Thermographic inspection is one of the most used
techniques for monitoring high voltage devices. It registers the
temperature gradient along the equipments surface, indicating
overheat. For zinc oxide (ZnO) arresters, this technique presents
a limitation. Because of the low thermal conductivity of the
porcelain or polymeric housing there is no direct relation between
the temperatures on the housing surface and the temperatures on
varistors. A solution for this limitation is the development of
computational routines to make this correlation. This paper
presents a program based on finite difference techniques that
makes the heat transfer over all the arrester indicating the
temperature of any desired point. It considers both ceramic and
polymeric housed arresters and can be used to simulate electrical
tests defined in technical normative and to estimate the inner Fig. 1 – Constructive design of a porcelain insulated arrester [9].
arrester temperatures when the outside temperatures are
measured. A comparison between thermal dissipation for
porcelain and polymeric insulated arresters is also presented.

Index Terms — Arrester, monitoring, simulation,

thermograph, varistor, zinc oxide.

I. INTRODUCTION

A RRESTERS are protection equipments connected

between phase and ground. They limit the voltage level in
equipments like transformers, avoiding them to be submitted Fig. 2 – Constructive design of a polymeric insulated arrester [9].
to a voltage inadequate to their operation. They are
characterized by a non-linear current flow through their II. NORMATIVE TESTS AND THERMOGRAPHIC INSPECTION
terminals, presenting a low leakage current for continuous
A. Normative Tests
operation and a high level of conduction when submitted to
overvoltages – in this case, the energy related to the high Arresters must be submitted to tests defined by norms so
current conduction must be dissipated by joule effect. that it can be considered adequate for the use in a power
ZnO arresters present a very simple structure. They are substation [13,14]. These tests usually measure the capacity of
composed basically of an insulating housing, made of the arrester to stand short-circuit requests with the correct
porcelain or polymeric material (usually silicone rubber), and operation of the pressure-relief device (in the case of ceramic
an inner active column, composed of the ZnO varistors and housing) or with a fragmentation previously evaluated (in the
thermal dissipating elements. The varistors are the main case of polymeric housing devices).
components of the equipment. They provide the desired non- Another test usually applied to arresters is the use of a
linear characteristics and present a strong relation with stressing operation cycle for a long time interval (5,000 hours)
temperature, which is associated with its energy absorption called accelerated weather ageing cycle under operation
capacity [1,2]. The main differences between polymeric and voltage. This test evaluates the behavior of the equipment by
porcelain insulated arresters are the constructive designs and the simulation of the environmental conditions for which it
the housing mechanical properties. Porcelain insulated would be submitted during its lifetime.
 2

The execution of these tests demands a previous study and a values can be obtained by the execution of tests with actual
significant cost. It is necessary the utilization of a new sample high voltage arresters.
arrester and the participation of specialized technicians in a For the normative tests, a sequence of different voltage
laboratory. A digital model is able to execute these normative entrances is used so that in digital simulation, equations
tests presenting an estimation of the final results. The representing different functions for the applied voltage are
presentation of these previous results represents a saving in developed – periodic waveforms, steps, impulses, voltage
resources and time, accelerating the realization of the real tests surges, allowing the usage of a wide variety of power entrance
in the sample arrester. for the arrester model.
It is important to notice that the arrester also absorbs
B. Thermographic Evaluation
thermal energy by other means like solar radiation and
proximity with heat sources. Although they represent only a
Different monitoring techniques can be used with surge small portion of the total energy absorbed by the arrester, they
arresters. One of these techniques is the thermography. must be correctly represented in the modeling so that it will be
Thermographic inspection is usually used with most the closest possible to real systems.
equipments in substations and with transmission lines to detect
overheating patterns which may indicate failures in the 2) Energy dissipation
equipment [3]. Besides, any changing in the arrester behavior
results in temperature changing [5]. Meantime, using The electrical energy absorbed by the arrester is converted
thermographic inspection results in the analysis of only the in thermal energy by joule effect. This energy can now be
external surface of the equipments. In the case of arresters, it dissipated to the environment by means of conduction,
can indicate some kind of problem, although it is not possible convection and radiation. Each one of these heat transfer
to know what is happening to its varistor column, which processes are used for different regions of the arrester. Inside
represents its functional part. solid elements, the conduction plays a major role, while
The knowledge of the varistors behavior inside the convection and radiation takes place in regions where air or
insulating housing is important to detect which kind of other gaseous substances are present.
problem is happening to the arrester. One alternative to One of the most desired characteristics of an arrester is its
analyze the arrester behavior is the usage of electro-thermal thermal stability. It has to return to its normal operation
models that represent the whole equipment structure and the conditions after the occurrence of any electrical request so that
physical phenomena involved. With the aid of an adequate it is able to face another possible subsequent surge [1]. When a
model, it is possible to estimate the temperatures in the ZnO varistor is degraded, it looses its thermal stability reaching a
elements starting from the outside arrester surface process of thermal runaway, characterized by the continuous
temperatures obtained from the thermographic inspection. increase of current and temperature resulting in the varistor
Besides, this knowledge and the employment of simulation are cracking or even the arrester explosion. Thermal runaway can
also extremely useful for the realization of electrical tests that also happen to equipments improperly chosen to specific
usually precedes arrester acquisition by electrical companies. operations. A degraded or improper arrester must be replaced
before compromising the integrity of the electrical system in
III. THE ARRESTER ELECTRO-THERMAL MODEL which it is connected [9], that is the importance of preliminary
ZnO arresters can be modeled in different ways – by testing and regular monitoring.
analogy with electrical circuitry, by section models or by The heat transfer flow for dissipation depends mainly on
computational simulation. This last one is the most used the kind of materials used. ZnO varistors presents a high
modeling, being more accurate, flexible and fast [11]. thermal conductivity level, so that the energy dissipated in this
Digital simulation makes use of mathematical equations that elements are easily conducted to their borders. Leaving the
represents the electric and thermal phenomena related to the ZnO column, the kind of housing used will now determine how
arrester operation. These phenomena include mainly the the heat transfer happens. Porcelain housed arresters present a
energy absorption in the active components of the arrester and very low heat transfer duty. It happens because they have a
the dissipation of this same energy by heat transfer. large air gap between the ZnO column and the housing. As the
air presents a very low conductivity, the heat transfer happens
by means of convection, which is a slow process. Besides that,
A. Fundamentals
the porcelain also presents a lower value of conductivity when
1) Energy absorption compared to silicon rubber.
An arrester absorbs electrical energy transforming it into
thermal energy. The varistors power absorption can be
calculated by the measurement of the voltage applied to the B. Finite Differences Method
arrester and the current running through its terminals. If this The finite differences technique allows the analysis of the
power is maintained for a certain period of time, the total temperature variation in a bulk solid in both radial and axial
energy absorbed can be also calculated. In laboratory, these directions. An arrester model using this method can be used to
 3

determine the heat flow and thermal stability of an arrester IV. LABORATORY TESTS AND SIMULATION
submitted to different operational states [7,8].
The method is based on the construction of a net A. Experimental Set
representing the desired arrester. Each element of the net is An experimental set with actual arresters samples was used
considered as an elementary unity for which all the thermal, so that the results estimated in digital simulation could be
mechanic and electric proprieties are valid. compared with the results measured in laboratory. The sample
The efficiency of this method depends on the quality of de arrester, a high voltage source, a resistive divider, a
generated net. The most complex is the geometry of the compensating capacitor, a digital oscilloscope and a digital
analyzed object, the less efficient will be the generated net. To computer compose this set. A thermographer and contact
avoid this problem, the arrester representation is made by thermometers are also used for the tests. A schematic diagram
approximating the round borders like the ones of the housing is shown in Fig. 5.
sheds by orthogonal borders. It allows an optimized use of the The resistive divider provides a voltage signal proportional
finite differences technique but results in an accuracy loss to system voltage so that the oscilloscope can be used. The
because of the approximation done for the sheds. A simple capacitor branch is used to compensate the capacitive current
example of the finite elements is shown in Fig. 4 [11,6]. generated by the arrester. The oscilloscope registers the values
of voltage and current and a computational program developed
in laboratory stores and processes these values and shows the
amount of energy absorbed by the arresters for each changing
in the value of the applied voltage. At the same time, thermal
measurements are made with thermographs so that the values
of the temperatures obtained can be compared with the values
calculated in the simulation.
Fig. 3 – Example of finite elements with heat transfer by conduction
mechanism.

In Fig. 3, the element I transfer its heat by conduction to the

elements I+1, I-1, I+M and I-M. Another possibility is the one
shown in Fig. 4. In this picture there is an air gap between the
elements I and I+1, representing part of the ZnO varistor and
part of the porcelain housing respectively. For this
configuration, the heat transfer made between these two
elements is made by convection and radiation.

Fig. 5 – Experimental set.

B. Methodology
For the tests realized in laboratory, a continuous
overvoltage is applied to the arrester resulting in overheating
Fig. 4 – Example of finite elements with heat transfer by conduction, of the ZnO varistors. Contact bimetallic thermometers are used
convection and radiation mechanisms. in different parts of the arrester, including the inner column.
The oscilloscope registers the applied voltage and the leakage
For the complete object being studied, some elements will current, sending these values to the computer. Termographic
behave as heat sources, so that the heat transfer flow can be pictures are obtained for each ten minutes so that the
calculated for all the other elements. For the arrester, two evaluation of the arrester external behavior is also analyzed.
different options are possible. In the first case, the sources are When the tests are concluded, the values of voltage, current
the elements receiving the electric power that is converted do and temperature are used for the electrothermal simulation.
thermal power. In the second case, the sources are the finite Different cases can be executed, according to the users
elements containing the housing surface, for which the requirements. Voltage and current can be used as entry data so
temperatures are measured with the thermograph. The finit that a thermal evaluation of the whole arrester can be estimated
differences method can now be used to analyze the behavior of or, the values of the measured surface temperatures can be
the arrester when submitted to a voltage level, or to estimate used to anticipate the inner temperatures.
the inner temperatures for monitoring purposes.
 4

V. RESULTS
140

Different test cases were executed with the simulation 120

Temperature (°C)
program in order to analyze the results. The first case tested 100
represents the arrester cooling after an overvoltage. The
80
second case represents the execution of normative tests usually
made for the electrical companies to determine the feasibility 60
of the arrester usage in their substations. The third case 40
represents the estimative of the arrester inner temperatures
20
when the outer temperatures are fixed (e.g., when the outer
temperatures are measured with an thermographer). For all the 0
cases, the results involving porcelain housed arrester could be 4 8 12 16 20 24
compared with the results obtained with real arresters in Time (min)
laboratory tests. For the porcelain cases, the simulation and
laboratory tests results match with a very small error. For the Fig. 7 - Cooling curve for a varistor in a polymeric housing arrester when
simulation involving polymeric arresters, additional tests submitted to an overvoltage.
should be made in laboratory although the results shown here
presents the behavior expected for these equipments. 60

55
A. Arrester cooling after an overvoltage 50

Temperature (°C)
For this case, an overvoltage is applied to the arrester until 45
its inner temperature increases to a previously determined 40
level. The voltage source is disconnected and the arrester cools 35
down. The heat transfer can be seen in pictures bellow. The 30
Fig. 6 and 7 show the temperature on the surface of a varistor
25
in the middle of the active columns after the temperature of
20
127°C is reached. An important point to be noticed is that the
0,0 0,2 0,4 0,6 0,8 1,0 1,2 1,4
cooling for the polymeric housed arrester is faster because of
Lenght (m)
the internal air gap absence.
Fig. 8 and 9 show the temperature along the length of the ZnO Surface Porcelain internal surface Porcelain external surface

arresters when the maximum temperature is reached on the Fig. 8 – Comparison between ZnO column surface temperature and porcelain
housing surface temperature.
ZnO surface. Because of the air gap, the difference between
the ZnO surface and the porcelain surface is extremely large
when compared with the polymeric case. This difference is the 28,6

main obstacle for a correct evaluation of termographic images 28,4

28,2
of porcelain-housed arresters. For the polymeric case, the
Temperature (°C )

28
internal housing surface is in direct contact with the ZnO 27,8
column surface, so there is no distinction for the temperatures 27,6
on these surfaces. 27,4
27,2
140 27

120 26,8
26,6
Temperature (°C )

100 0 0,2 0,4 0,6 0,8

80 Lenght (m)

60 ZnO surface Polymeric external surface

40 Fig. 9 – Comparison between ZnO columns surface temperature and

polymeric housing surface temperature.
20

0
B. Normative tests
0 100 200 300 400 500
For these simulations, a sequence of pulses and
Time (min)
overvoltages defined by technical norms are applied to the
Fig. 6 – Cooling curve for a varistor in a porcelain housing arrester when arrester and its thermal behavior is analyzed. This is the kind
submitted to an overvoltage. of test usually made with real arrester samples to evaluate its
 5

viability for a specific site. Fig. 10 and 11 present the results simulation software for the analysis of the arrester when a
for a normalized ANSI test. Fig. 10 depicts the temperature for thermographer is used for equipment inspection.
points on ZnO surface and porcelain internal and external As shown in Fig. 12, even when a porcelain outside
surface at 54 cm from the base. In Fig. 11, the points are about temperature of 30°C is measured by a thermographer, the
60 cm from the base. varistor already reached about 53°C emphasizing the absence
of control of the inner arrester temperature.
180
60
150
55
Temperature (°C)

120 50

Temperature (°C)
45
90
40
60
35

30 30

25
0
0 100 200 300 400 500 20
Time (min) 0,2 0,4 0,6 0,8 1 1,2 1,4
Arrester lenght (m)
ZnO Surface Porc. Internal surface Porc. External surface
Estimated temperature Fixed temperature
Fig. 10 – ANSI test for a porcelain housed arrester.
Fig. 12 – Temperature evaluation on varistor column surface of a porcelain
housed arrester when the outer surface temperature is fixed.
100

90
VI. CONCLUSION
80
Temperature (°C)

70
Arresters are equipments of fundamental importance to
substations and transmission systems. Being connected directly
60
between electric system and ground they are submitted to an
50
electrical leakage current which increases when facing an
40
electrical surge, overvoltage or when the arrester varistors are
30
degraded.
20 The main consequence of increasing the leakage current is
0 100 200 300 400 500 600 700 800
Time (min)
the elevation of the varistors temperature, what can lead to
their complete degradation and cracking. To avoid the
ZnO surface Housing external surface
problems related above, arresters behavior must be studied
Fig. 11 – ANSI test for a polymeric housed arrester. before their installation in the site, determining their feasibility
to that specific use. They must also be submitted to preventive
Results emphasize the better thermal dissipation of inspection from time to time when already installed. The
polymeric housing arrester. For the used ANSI test, the analysis of thermographic images is one of the most used
temperature pattern in the chosen point of polymeric housed inspection procedures in substations.
arrester follows narrowly the applied voltage pattern. For the An electrothermal simulation computational program was
porcelain case, the difficult to transfer energy is visible in Fig. presented. This program is capable of simulating the heating
10. The temperature doesn’t follow the pattern of the applied over all the arrester when a sequence of voltage waveforms is
voltage, demonstrating a high thermal resistivity of the applied on its terminals – what is useful to analyze the arrester
porcelain housing – the temperature increases and decreases in behavior when submitted to extreme electrical solicitations or
a very low duty. norm tests. It is also useful to estimate the inner arrester
Other interesting observation is the proximity between temperature when the temperatures in the arrester housing
temperatures in silicon rubber surface and varistors surface for surface are known – this is a tool for predicting the
polymeric case and the distance between temperatures on temperatures in the varistors surface when a thermal image is
porcelain surface and varistors surface for porcelain cases. analyzed.
These results also indicate a better dissipation for polymeric Simulations of different cases were executed showing the
case mainly due to the absence of internal air gap. usefulness of the program. The simulation results also compare
C. Inner temperature estimation the behavior of porcelain housing and polymeric housing
arresters when facing similar electrical conditions, showing
For this case, the temperature on the housing surface is fixed
that arresters with polymeric housing are less affected by
in about 30°C so that the temperature in the ZnO varistors can
overheating.
be estimated. It could represent the application of the
 6

VII. ACKNOWLEDGMENT Campina Grande since 1978. E-mail: [email protected].

Authors would like to thank Max Norat from CHESF for M. J. de A. Maia received his B.Sc. degree in electrical engineering in 1978
being always ready to help with technical information and from the Universidade Federal da Paraíba. Since 1978, he has been with the
Companhia Hidro Elétrica do São Francisco, Recife, PE, Brazil. His research
equipment lending.
interests include power systems, surge arresters, transmission and distribution
of electric energy. E-mail: [email protected].
VIII. REFERENCES

[1] W. G. Carlson, T. K. Gupta, A. Sweetana. “A Procedure for Estimating

the Lifetime of Gapless Metal Oxide Surge Arresters”. IEEE Trans. on
Power Systems, vol. 1, n. 2, p. 67-73, April, 1986.
[2] T. K. Gupta. “Application of zinc oxide varistors”. J. Am. Ceram.
Soc., Vol. 73, N° 7, p. 1817-1840, 1990.
[3] Z. Korendo and M. Florkowski, “Thermography-based diagnostics of
power equipment”, IEE Power Engineering Journal, pp. 33-42, Feb.
2001.
[4] M. V. Lat. “Thermal properties of metal oxide surge arresters”. IEEE
Transactions on Power Apparatus and Systems, Vol. 102, N° 7, pp.
2194-2202, July 1983.
[5] C. Heinrich, V. Hinrichsen. “Diagnostics and Monitoring of Metal-
Oxide Surge Arresters in High-Voltage Networks - Comparison of
Existing and Newly Developed Procedures”. IEEE Transactions on
Power Delivery, Vol. 16, N° 1, January 2001.
[6] E. G da Costa, S. R. Naidu, A. G de Lima. “Electrothermal model for
complete metal-oxide surge arresters”. IEE Proceedings – Generation,
Transmission and Distribution, Vol. 148, N° 1, pp. 29-33, January
2001.
[7] C. R. Maliska, “Transferência de Calor e Mecânica dos Fluidos
Computacional”. LTC, Rio de Janeiro, RJ, 1995.
[8] J. P. Holman, “Transferência de calor”. Mc Graw-Hill do Brasil, São
Paulo, 1983.
[9] ABB Switchgear, “Physical properties of zinc oxide varistors”. ABB
Power Technology Products AB, 2001
[10] V. Hinrichsen, “Metal Oxide Surge Arresters – Fundamentals”, Siemens
– Power Transmission and Distribution Power Voltage Division, Berlin,
2001
[11] E. T. Wanderley Neto. “Pára-Raios de Óxido de Zinco: Fundamentos,
Caracterização e Monitoramento”. Research project presented at
Federal University of Campina Grande, February 2003.
[12] E. G. da Costa, “Análise do Desempenho de Pára-Raios de Óxido de
Zinco”, Ph.D Dissertation, Electrical Engineering Department, Federal
University of Paraíba, 1999.
[13] IEEE standard for metal oxide surge arresters for ac power circuits.
ANSI/IEEE C62.11-1987.
[14] Metal-oxide surge arresters without gaps for a.c. systems. IEC 99-4,
SURGE ARRESTERS Part 4. First edition, 1991-11.

IX. BIOGRAPHIES

Estácio Tavares Wanderley Neto was born in

Campina Grande, Brazil, on September 15, 1977. He
graduated from the Federal University of Paraíba,
Campina Grande, and obtained his Masters degree
from the Federal University of Campina Grande in
2003. His experience includes a professional
apprenticeship at the Companhia Hidrelétrica do São
Francisco – CHESF in 2000. Estácio is now applying
for his Ph.D. degree from the Federal University of
Campina Grande in the area of zinc oxide arresters
electrothermal studies. E-mail: [email protected].

Edson G. da Costa was born in Brazil, 1954. He

graduated as Electrical Engineer in 1978, received his
Masters degree in 1981 and his Ph.D. title in 1999, at
Federal University of Paraíba. His interests are power
systems, electric fields, partial discharges, arresters
and insulators. His is professor at the Electrical
Engineering Department of Federal University of